INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

For the first time in 20 yr, a new class of drugs has been developed to treat asthma (1). These novel compounds act by either inhibiting the synthesis of leukotrienes or by blocking their biologic activity. Thus, leukotriene inhibition may offer a more specific mechanism to block inflammatory responses in asthma than currently available therapeutic interventions. To date, clinical trials have shown a therapeutic benefit in patients with asthma triggered by numerous factors, including cold air (2), exercise (3), aspirin (4), and exposure to antigen (5). Although support exists for an anti-inflammatory effect of leukotriene inhibition in asthma (6), the mechanisms of these clinical benefits are not completely understood. Investigators have found indirect evidence for anti-inflammatory effects in the attenuation of the late bronchoconstrictor response to inhaled antigens after treatment with leukotriene inhibitors (5). This response has been observed in some but not all subjects, suggesting that there may be a subpopulation of asthmatics for whom leukotrienes are an important factor in stimulating airway inflammation (10).

We reasoned that the variable responsiveness to leukotriene inhibitors in asthmatic patients may reflect intrinsic differences in either allergen-induced leukotriene generation or airway responsiveness to leukotrienes. In previous studies there were no identifiable subpopulations of allergic asthmatics that failed to show hyperresonsiveness to inhaled LTC₄ and LTD₄ (11) or that failed to release sulfidopeptide leukotrienes in the airways immediately after segmental allergen challenge (12).

The objectives of this study were to determine whether there were differences in leukotriene response to allergen in allergic asthmatics and to investigate the effects of inhibiting the 5-lipoxygenase pathway on leukotriene generation and inflammation within their airways. We measured sulfidopeptide leukotrienes (LTC₄/D₄/E₄), LTB₄, inflammatory cytokine mediators, and inflammatory cells in bronchoalveolar lavage fluid (BALF) before and 24 h after segmental antigen challenge in 18 allergic subjects with mild to moderate asthma at baseline. Interim analysis of BALF mediators before initiation of a 6-wk treatment with the 5-lipoxygenase inhibitor zileuton or placebo revealed that only nine of the 18 asthmatic subjects had at least a 4-fold increase in BALF LTC₄/D₄/E₄ levels after segmental antigen challenge, whereas the remaining subjects had < 2.2-fold increase in BALF LTC₄/D₄/E₄ levels. The high-LT group also had higher levels of inflammatory cytokines and cells in their post-antigen BALF than did the low-LT group. Subsequent post hoc analysis of the high- and low-LT subgroups revealed that treatment with the 5-lipoxygenase inhibitor zileuton reduced several measures of antigen-induced indices of inflammation, but only in those subjects with a leukotriene response to antigen challenge at baseline. These data suggest a possible explanation for the variable clinical response to leukotriene-modifying drugs in asthmatic patients and demonstrate that these drugs may have potent anti-inflammatory effects in a subpopulation of asthmatics.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patient Selection

Twenty nonsmoking allergic subjects with mild-to-moderate asthma were recruited from the local community. After informed consent was obtained, the subjects completed an asthma questionnaire and underwent preliminary evaluation, which included physical examination, allergy skin testing, and bronchial responsiveness testing with methacholine. Subjects were required to have clinical asthma based on the NAEP (NHLBI) criteria, which include a history of asthma symptoms (cough, dyspnea, or wheeze), bronchial hyperresponsiveness (BHR), and prior treatment with asthma medications (13), and had a positive response to intradermal injection of short ragweed antigen (Greer Laboratories; Lenoir, NC). A range of ragweed antigen dilutions from 1:20 to 1:1,000,000 was tested along with 1.8 mg/ml histamine and saline vehicle, which were included as positive and negative controls. A positive skin test was defined as a wheal  5 mm in diameter. Baseline spirometry was performed using a SensorMedics spirometer model 2400 (SensorMedics; Yorba Linda, CA). Methacholine bronchial responsiveness was tested using standard techniques by having the subjects inhale doubling concentrations of methacholine chloride (from 0.15 to 25 mg/ml) for 0.6 s during each of five breaths from a DeVilbiss nebulizer (DeVilbiss Co., Somerset, PA) until the FEV₁ decreased 20% or more (14). Bronchial responsiveness was expressed as the provocative dose of methacholine that caused a 20% drop in FEV₁ (PD₂₀FEV₁). Spirometry was repeated 5 min after inhalation of either the diluent or each concentration of methacholine. The cumulative methacholine dose required to produce a 20% drop in FEV₁ (PD₂₀) was calculated by interpolation from the dose-response curve, and a cutoff of 80 cumulative breath units was used as the criterion for bronchial hyperresponsiveness.

Study Protocol

Twenty subjects entered this placebo-controlled, double-blind study (Table 1). After skin testing and methacholine challenge, each subject underwent an initial 2-d segmental antigen challenge study according to NHLBI guidelines (15). On Day 1, spirometry was performed followed by investigational bronchoscopy. After mild sedation with fentanyl and midazolam and local anesthesia with 2% lidocaine, bronchoscopy was performed. The right middle lobe was lavaged with six 20-ml aliquots of 0.9% NaCl at 37°C. Four milliliters of 0.9% NaCl were instilled into a subsegment of the lingula. The bronchoscope was then repositioned and 4 ml of 0.9% NaCl containing 480 protein nitrogen units ragweed antigen (ALK Laboratories, Milford, CT) was instilled into a subsegment of the right upper lobe. The bronchoscope was then withdrawn and the subject was observed for at least 2 h. After 24 h the subject returned and spirometry and investigational bronchoscopy were repeated. The second bronchoscopy included lavages of the lingula and the right upper lobe.

Subjects were then randomized to receive 6 wk of therapy with either zileuton 600 mg or identical placebo capsules orally four times per day. Subjects were allowed to use albuterol and theophylline, but refrained from inhaled corticosteroids, nedocromil, cromolyn, aspirin, or nonsteroidal anti-inflammatory medications. Spirometry and methacholine BHR testing were repeated after 22, 34 to 38, and 41 d on study drug. On study Day 42, the subjects received a physical examination, repeat laboratory testing, and underwent a second 2-d study, including spirometry, investigational bronchoscopy, and segmental antigen challenge, which was identical to the predrug segmental antigen challenge study. Compliance with study medication was verified by patient diary and by counting medication tablets. The experimental protocol used in this study was approved by the institutional review board for human research at the University of Maryland, Baltimore.

Processing of BALF

The six aliquots returned from each segment of the lung were pooled, quantified, and poured through sterile gauze to remove mucus. The cells were separated from the fluid by centrifugation at 500 × g for 10 min at 4° C. An aliquot of the cell-free BALF was immediately frozen at 80° C for measurement of leukotrienes, interleukins IL-6, IL-8, IL-1, and tumor necrosis factor- (TNF-). The remainder of the supernatant was concentrated 10- to 20-fold by spin-cell ultrafiltration (5 kD cutoff; Amicon Corp., Danvers, MA) and stored at 80° C to measure IL-4, IL-5, and granulocyte/macrophage-colony-stimulating factor (GM-CSF). Cells were counted using a hemacytometer, viability was assessed by trypan blue dye exclusion, and differential cell counts were obtained from cytopreparations stained with Diff-Quik (Baxter Scientific Products, Miami, FL). At least 200 cells were counted on each slide. Separate slides were stained with alcian blue and counterstained with safranin to count metachromatic cells. All cell counts were performed by two investigators (J.H. and P.W.), and the means of the two counts were recorded for each slide.

Reagents

Unless otherwise stated, all chemicals were obtained from Sigma Chemical (St. Louis, MO).

Cytokine and Total Protein Assays

All assays were performed using commercial enzyme-linked immunosorbent assay (ELISA) kits from R&D Systems (Minneapolis, MN). The sensitivities for these assays were as follows: TNF-, 0.5 pg/ml; IL-1, 3.9 pg/ml; IL-4, 31 pg/ml; IL-5, 7.8 pg/ml; IL-6, 0.15 pg/ml; IL-8, 31 pg/ml; GM-CSF, 7.8 pg/ml. For all preantigen and postantigen samples, mixing studies were performed to exclude interference with each ELISA. In all cases, recovery of samples spiked with recombinant cytokine was at least 80% of predicted. Total protein was measured using a commercial reagent (Bio-Rad, Hercules, CA) and concentration was calculated using a standard curve constructed with bovine serum albumin (Sigma). All results were expressed as simple concentration in neat BALF.

Leukotriene Assays

Heparinized venous blood was stimulated to produce eicosanoids by the addition of 50 µM calcium ionophore A23187 followed by incubation for 30 min at 37° C. The blood was then chilled on ice and the plasma was collected by centrifugation at 1,000 × g for 10 min and stored at 20° C until analysis. Samples were extracted by the addition of four volumes of acetonitrile: methanol 1:1 and centrifuged at 1,500 × g for 10 min at 4° C, and the resultant supernatants were diluted 1:20 in enzyme immunoassay (EIA) buffer (0.1 M potassium phosphate at pH 7.4, 0.01% wt/vol sodium azide, 0.4 M NaCl, 1 mM EDTA, and 0.01% wt/vol bovine serum albumin). BALF (1 ml) was extracted in duplicate by addition of four volumes of ethanol, followed by incubation at 4° C for 2 h then cleared by centrifugation for 15 min at 1,000 × g. The supernatants were dried under a nitrogen stream at ambient temperature and resuspended in 250 µl EIA buffer. The LTB₄ levels in the extracted plasma and BALF samples were measured using a validated competition enzyme immunoassays (EIA) after modification of published methods (16). For LTB₄ assays, EIA plates (Nunc; VWR Scientific, Chicago, IL) were precoated overnight with mouse antirabbit IgG antibody (Cayman Chemical; Ann Arbor, MI), and blocked for 24 h with 0.1 M potassium phosphate, 0.03% sodium azide, 0.4 M NaCl, 1 mM EDTA, and 3% wt/vol bovine serum albumin. After washing the EIA plates with wash buffer (0.01 M potassium phosphate, 0.05% vol/vol Tween 20); 50-µl aliquots of extract, standard (BioMol Chemical; Plymouth Meeting, PA), or QC control, 50 µl of LTB₄-acetylcholinesterase tracer (Cayman Chemical), and 50 µl of anti-LTB₄ (PerSeptive Biosystems, Cambridge, MA) were added to each well in duplicate. After overnight incubation and washing, the plates were developed by the addition of 200 µl Ellman's reagent per well. LTC₄/D₄/E₄ levels of BALF extracts were determined using a commercially available LTC₄/D₄/E₄ EIA Kit (PerSeptive Diagnostics) modified as described below. Briefly, 50 µl of sample, standard, or QC control and 50 µl of antisulfidopeptido leukotriene antibody were pipetted into each well in duplicate and incubated overnight at 4° C. LTC₄-alkaline phosphatase tracer (50 µl) was added to each well and the plate was incubated for 3 h at 4° C. The plate was washed and 200 µl of paranitrophenyl phosphate substrate were added and further incubated for 4 h at ambient temperature. For either assay the absorbance at 405 nm of the EIA plate was measured (max EIA plate reader; Molecular Devices, Menlo Park, CA) and the leukotriene concentrations of the study samples were calculated from the standard curve using a four-parameter logistic model (Softmax; Molecular Devices). The lower detection limits for LTB₄ in plasma and BALF were 1.56 ng/ml and 14.5 pg/ml, respectively. The lower detection limits for LTC₄/D₄/E₄ in neat BALF was 7.8 pg/ml. For BALF samples in which LTC₄/D₄/E₄ was undetectable, the lower detection limit, 3.9 pg/ml, was assigned. The coefficients of variation were < 10% for both the LTB₄ and the sulfidopeptido leukotriene assays.

Statistical Analysis

All data were expressed as mean ± SE in the text and in the figures. Statistical analyses were performed using Statview II (Abacus Concepts; Berkeley, CA). Differences in mediator levels and cell numbers among prechallenge, saline-challenged control, and postantigen segments were tested using Friedman's test, and differences between pairs of data were tested using Wilcoxon's sign rank test. The effect of therapy on mediator and cellular composition of postantigen BALF, FEV₁, and PD₂₀ were tested by calculating the post-therapy:pretherapy ratio of each parameter and comparing each of these values in the zileuton and placebo groups using a Mann-Whitney test. Correlations were analyzed using Spearman's rank test. A post hoc analysis of 5-lipoxygenase inhibitor effects on antigen-induced inflammation was performed by dividing the subjects into low- and high-LT subsets based on the measured increase in BALF LTC₄/D₄/E₄ levels after the first segmental antigen challenge. Based on a coefficient of variation of 10% for the leukotriene assays and at least 2-fold variability in dilution introduced during segmental antigen challenge, we established a priori a 2.5-fold increase in BALF LTC₄/D₄/E₄ concentration after segmental antigen challenge as the criterion for a significant antigen-induced LTC₄/D₄/E₄ response. The low- and high-LT subsets were each analyzed separately and were compared using the Mann-Whitney test. Statistical significance was defined as p < 0.05. Data are presented graphically as box plots in which the top, bottom, and line through the middle of each box correspond to the top quartile, bottom quartile, and median, and the error bars correspond to the bottom and top decile. The data are reported in the text as mean ± SE.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The characteristics of the 20 subjects are summarized in Table 1. The subjects assigned to receive zileuton had lower FEV₁% predicted values and tended to have higher methacholine PD₂₀ values than the placebo group. The two groups were otherwise similar. Eighteen of 20 subjects completed the study. Two subjects randomized to zileuton therapy withdrew from the study after the first segmental antigen challenge and before initiation of drug because of severe coughing during investigational bronchoscopy in one subject and onset of a postbronchoscopy productive cough in the other subject. Investigational bronchoscopy and segmental antigen challenge were well tolerated in the rest of the subjects. Bronchoalveolar lavage returned similar proportions of the instilled volume in the prechallenged (41.3 ± 2.8%), saline-challenged (45.8 ± 2.7%), and antigen-challenged (38.4 ± 2.7%) segments. 5-lipoxygenase inhibition was well tolerated; there were no differences in reported adverse effects between the zileuton and the placebo groups.

BALF Cytokine and Cell Composition

Baseline BALF contained only slightly higher percentages of neutrophils (1.31 ± 0.52%), eosinophils (0.84 ± 0.02%), and metachromatic cells (1.02 ± 0.17%) than we have found in healthy nonasthmatic subjects in our laboratory (17). BALF collected 24 h after segmental antigen challenge contained higher concentrations of total protein, IL-1, IL-4, IL-5, IL-6, IL-8, GM-CSF, and TNF compared with prechallenge levels (Figure 1A). There were greater numbers of macrophages, lymphocytes, neutrophils, eosinophils, and metachromatic cells (Figure 1B) compared with baseline (prechallenge) and control (saline-challenged) BALF. Saline challenge stimulated slight increases in levels of IL-4, IL-6, and IL-8, and numbers of neutrophils and eosinophils in BALF.

View larger version (59K): [in this window] [in a new window]   View larger version (46K): [in this window] [in a new window]   Figure 1.   Composition of BALF before and 24 h after segmental antigen challenge or saline challenge (control). Concentrations of cytokines, leukotrienes, and total protein (A) and numbers of cells (B) in BALF obtained before challenge (Pre) and after either saline control (Ctr) or antigen challenge (Ag) were measured in all subjects prior to drug treatment as described in METHODS. The data are expressed as box plots. The first p value listed for each graph compares postantigen with prechallenge data. If significant, a second p value is listed comparing saline-challenged control with prechallenge values. Mediator concentrations are pg/ml except for total protein (mg/ml).

BALF Leukotriene Concentrations

Baseline BALF LTC₄/D₄/E₄ and LTB₄ levels were narrowly distributed in these subjects (22.5 ± 7.4 pg/ml for LTC₄/D₄/E₄ and 38.9 ± 3.8 pg/ml for LTB₄; mean ± SE) (Figure 1). Instillation of saline into control segments had no effect on leukotriene levels in BALF recovered 24 h later. The relative change in leukotriene levels before and 24 h after antigen challenge varied widely among the asthmatic subjects. In eight subjects there was  2.2-fold change (1.14 ± 0.22; mean ± SE) in LTC₄/ D₄/E₄ levels after antigen challenge, whereas LTC₄/D₄/E₄ levels increased at least 4.2-fold (234 ± 102) after antigen challenge in the other nine subjects (Figure 2). The former group will be referred to as the low-LT group and the latter as the high-LT group. In one subject, pre-challenge LTC₄/D₄/E₄ was undetectable and the post-antigen LTC₄/D₄/E₄ level (16 ng/ml) was less than the mean prechallenge level of the low-LT group. Therefore this subject was assigned to the low-LT group. As expected postantigen BALF LTC₄/D₄/E₄ and LTB₄ levels correlated significantly (p < 0.001) (Figure 2C). In the nine subjects who received placebo and completed both segmental antigen studies, antigen-induced BALF LTC₄/D₄/E₄ levels measured 6 wk apart were similar (rho = 0.667; p = 0.059 by Spearman's rank test), suggesting a reproducible response.

View larger version (21K): [in this window] [in a new window]   View larger version (21K): [in this window] [in a new window]   View larger version (19K): [in this window] [in a new window]   Figure 2.   Leukotriene concentrations in BALF after segmental antigen challenge. Concentrations of LTC4/D4/E4 (A) and LTB4 (B) in BALF obtained before and 24 h after segmental antigen challenge were measured in all subjects prior to drug treatment as described in METHODS. The subjects were assigned to a high-LT group if they demonstrated at least a 2.5-fold increase in LTC4/D4/E4 level after antigen challenge. The medians are indicated and the p values listed compare the antigen-induced increases between the low- and the high-LT groups. The correlation between LTB4 and LTC4/D4/E4 concentrations (C) in postantigen BALF were analyzed using Spearman's rank test.

We reasoned that the two groups of subjects identified by the BALF LTC₄/D₄/E₄ response to antigen may represent distinct subpopulations of asthmatics. To compare the intensity of antigen-induced airway inflammation, we compared the cytokine and cellular composition of the post-antigen BALF in these two groups of subjects. Although there were no detectable differences in the composition of BALF between the low- and the high-LT groups at baseline, the postantigen BALF contained higher levels of TNF-, IL-5, and IL-6 (Figure 3A) in the high-LT group compared with the low-LT subjects. The mean total protein concentration in the postantigen BALF was 8.4-fold higher in the high-LT group, indicating greater protein extravasation compared with the low-LT group. The number of eosinophils in postantigen BALF was 9.6-fold higher in the high-LT than in the low-LT subjects (Figure 3B).

View larger version (52K): [in this window] [in a new window]   View larger version (38K): [in this window] [in a new window]   Figure 3.   Composition of postantigen BALF in the low- and the high-LT subsets of asthmatic subjects before treatment. Low- and high-LT group assignments were assigned as described in METHODS. Concentration of cytokines, leukotrienes, and total protein (A) and numbers of cells (B) were measured in post-antigen BALF from the baseline study as described in METHODS and are displayed as box plots. The p value listed for each plot compares the values for the low- and the high-LT groups. Mediator concentrations are pg/ml except for total protein (mg/ml).

Effects of 5-lipoxygenase Inhibition on BALF Inflammatory Indices

5-lipoxygenase inhibition caused significant reductions in BALF LTC₄/D₄/E₄ levels in the antigen-challenged lung segments (Table 2), but there were no statistically significant effects of this intervention on any other mediator or on the cellular composition of BALF when all subjects were analyzed. However, a separate analysis of the high-LT subset revealed consistent reductions in postantigen BALF cell numbers after 6 wk of treatment with zileuton (Figure 4B). 5-lipoxygenase inhibition, but not placebo therapy, reduced the eosinophil count in the postantigen BALF of the high-LT subjects by 68% (Figure 4B). The same therapy did not alter the composition of postantigen BALF in the low-LT subjects (Figure 4A).

View larger version (31K): [in this window] [in a new window]   Figure 4.   Effect of 5-lipoxygenase inhibitor therapy on postantigen BALF composition in the low- and the high-LT subsets of asthmatic subjects. Low- and high-LT group assignments were assigned as described in METHODS. A 2-d investigational bronchoscopy/segmental antigen challenge protocol was performed before and after 6 wk of therapy with the 5-lipoxygenase inhibitor zileuton or placebo. The total numbers of eosinophils, neutrophils, and metachromatic cells were determined as described in METHODS. For each subject in the low-LT (A) and high-LT (B) groups, the ratio of postantigen cell count measured after 6 wk of therapy: cell count measured before therapy was calculated and the values for subjects receiving active drug and those receiving placebo were displayed as box plots and were compared using a Mann-Whitney test. The p values are indicated.

Correlation of BALF and Blood Leukotriene Levels

To determine whether the leukotriene response in the bronchoalveolar compartment reflected the intrinsic capacity of each subject's mononuclear phagocytes to synthesize leukotrienes, we compared venous blood LTB₄ generation in the low- and high-LT subjects. LTB₄ levels were measured after a 30-min incubation with or without the leukotriene inducer A23187. There were no differences between the low- and high-LT groups in basal (33.7 ± 9.4 versus 24.1 ± 3.8 ng/ml in the low- and high-LT subjects) or PMA-induced (216 ± 32.5 versus 210 ± 26 ng/ml) LTB₄ levels or in the effect of a 6-wk zileuton treatment on LTB₄ generation (86.5 ± 4.8 versus 89.4 ± 3.2% reduction in PMA-induced LTB₄ levels in the low- and high-LT subjects). Furthermore, there was no correlation between postantigen BALF LTC₄/D₄/E₄ levels and unstimulated LTB₄ (rho = 0.152; p = 0.54) or stimulated LTB₄ (rho = 0.132; p = 0.60) generation in blood.

Comparison of Asthma Severity in the High- and Low-LT Groups

There were no differences in prestudy spirometry, PD₂₀ (methacholine) values or skin test sensitivity between the high- and the low-LT groups. There were no differences in sex, age, race, or asthma duration. At baseline FEV₁ (% predicted) was 3.16 ± 0.34 L (76 ± 4.1%) versus 3.38 ± 0.22 L (76.4 ± 3.1%) in the low- and high-LT groups. The methacholine PD₂₀ FEV₁ was 4.64 ± 3.0 and 8.1 ± 3.3 in the low- and high-LT subjects. The ragweed antigen titer that induced a detectable wheal and flare was 613 ± 258 in the high-LT group and 574 ± 264 in the low-LT group.

Clinical Response to Zileuton Treatment

There was a small improvement in BHR, but no change in FEV₁ after a 6-wk treatment with the 5-lipoxygenase inhibitor zileuton (Table 3). FEV₁ improved in each of the four high-LT subjects during 5-lipoxygenase inhibitor therapy, and BHR tended to improve in the low-LT group, but neither difference reached statistical significance.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Leukotriene generation in the airways increased 24 h after segmental antigen challenge in a subset of subjects with mild-to-moderate asthma. This subset of asthmatic subjects also had higher cytokine levels and greater numbers of inflammatory cells in postantigen BALF than did those asthmatics without a prominent leukotriene response, suggesting more intense antigen-induced airway inflammation in the high-LT subjects. Administration of a 5-lipoxygenase inhibitor, zileuton, reduced some indices of inflammation in the postantigen BALF in the high-LT, but not the low-LT asthmatic subjects.

Analysis of previous efficacy studies of leukotriene inhibition suggests that responsiveness to these agents may be heterogeneous (18). Our results suggest that there may be two subpopulations in asthma based on BALF LTC₄/D₄/E₄ levels 24 h after segmental antigen challenge. Only a single concentration of ragweed antigen was used for segmental antigen challenge. Thus, the heterogeneity in antigen-induced LT expression may reflect variability in sensitivity to antigen rather than a decreased capacity for LT generation. Two lines of evidence argue against this explanation. First, the single antigen dose used, 480 PNU, was a high concentration, which should exceed the threshold of all the asthmatic subjects. Second, there was no difference between the high- and the low-LT groups in threshold for antigen skin test responsiveness. Diaz and colleagues (19) found a similar heterogeneity in bronchoalveolar LTC₄ levels 6 h after whole lung antigen challenge. They reported that the high-LTC₄ producers could be distinguished from the low-LTC₄ subjects by the presence of a late bronchoconstrictor response to inhaled antigen. Elevated LTC₄/D₄/E₄ levels can be detected in BALF as early as 5 min after antigen challenge and also 24 and 48 h later (12, 20). Whereas mast cells are a predominant source of the early allergen-induced LTC₄ generation, eosinophils and macrophages may contribute to the late LTC₄ response. Because macrophages have a much greater capacity to generate LTB₄ than do eosinophils (21, 22), the coexpression of bronchoalveolar LTB₄ and LTC₄/D₄/E₄ after antigen challenge in our study suggests that bronchoalveolar macrophages are also a source of both leukotrienes. The reproducibility of the post-BALF LTC₄/D₄/E₄ measurements performed 6 wk apart in the placebo group suggests that the leukotriene response to antigen challenge in the bronchoalveolar compartment is stable over this period of time. The lack of correlation between LTB₄ generation in blood and LTC₄/D₄/E₄ levels in postantigen BALF suggests that the latter does not reflect the intrinsic capacity of mononuclear phagocytes for leukotriene production. However, leukotriene generation in blood was stimulated by calcium ionophore rather than antigen, a potent nonspecific stimulus that bypasses early signaling pathways.

Although zileuton treatment reduced LTC₄/D₄/E₄ levels in postantigen BALF, we did not detect a similar effect on BALF LTB₄ levels. This result was surprising because zileuton blocks synthesis of a common precursor of LTC4/D4/E4 and LTB4. In fact, we did find a marked reduction in LTB₄ generation in blood obtained from subjects treated with zileuton. The discrepancy in the response of BALF LTB₄ and LTC₄/D₄/E₄ levels to zileuton therapy may reflect the timing of bronchoscopy after antigen challenge since leukotriene levels were analyzed 24 h after segmental antigen challenge. Although antigen induces persistent generation of LTC₄ lasting at least 48 h (20), LTB₄ generation may be less sustained, as suggested by the lower levels of LTB₄ than of LTC₄/D₄/E₄ in the postantigen BALF. Diaz and colleagues previously reported that LTC₄ was elevated, but LTB₄ was undetectable 6 h after antigen challenge.

The 68% reduction in postantigen BALF eosinophil count and the consistent trend toward lower numbers of neutrophils and metachromatic cells in the high-LT subjects receiving 5-lipoxygenase inhibitor therapy suggests that leukotrienes may also contribute to late antigen-induced inflammation. Other groups have also reported modest reductions in BALF inflammatory indices in asthmatic subjects during treatment with leukotriene-modifying agents. Wenzel and colleagues (8) reported a decrease in the percentage of eosinophils in BALF and blood in subjects with nocturnal asthma during treatment with zileuton. Only six of 10 subjects had a reduction in BALF eosinophil percentage, and eight of 12 subjects showed an improvement in FEV₁, but segmental antigen challenge and postantigen leukotriene analysis were not performed in this study. Kane and colleagues (6) reported that the number of eosinophils in BALF obtained 24 h after segmental antigen challenge in a mixed group of subjects with atopy or mild allergic asthma were reduced during zileuton therapy. In this study, urinary LTE₄ levels were reduced during zileuton therapy in all 10 subjects, but leukotriene levels in BALF were not reported. Calhoun and colleagues (9) found that the treatment of mild asthma with the LTD₄ receptor antagonist, zafirlukast, eliminated metachromatic cells, produced small reductions in TNF- levels in BALF obtained 48 h after segmental antigen challenge, and decreased the capacity of BALF macrophages for superoxide release, but leukotriene levels were not reported in this study. These data suggest that leukotriene-modifying agents may have anti-inflammatory effects in asthmatics.

The data from this study further suggest that these agents may have potent anti-inflammatory effects in a subpopulation of asthmatics with leukotriene-dependent inflammation. Although we failed to show an effect of 5-lipoxygenase inhibition in the low-LT group, the small number of subjects does not provide sufficient power to exclude anti-inflammatory effects in persons without a prominent bronchoalveolar leukotriene response to antigen.

In this study, subjects treated with zileuton showed improvement in bronchial hyperresponsivenes, but not in FEV₁. Failure to demonstrate an increase in FEV₁ is consistent with the mild baseline airway obstruction in these subjects and/or the short duration of therapy. Israel and colleagues (23) reported a 13.4% improvement in FEV₁ during zileuton therapy in subjects with moderate asthma with baseline FEV₁ between 40 and 75%, whereas Fish and colleagues (24) failed to show any significant effect of treatment with the LTD₄ receptor antagonist zafirulukast on FEV₁ in subjects with milder asthma with mean FEV₁ that was 78% of predicted. Other methodologic factors may explain the lack of clinical improvement in the zileuton group, including: the use of a parallel design, suboptimal randomization producing a zileuton group that may have had more severe asthma at baseline, and small group size with only four subjects in each of the high- and low-LT groups. Considering the mild baseline obstruction in our subjects, the improvement in FEV₁ in each of the high-LT subjects who received zileuton therapy suggests that leukotriene-modifying agents may be of greater benefit in the subset of asthmatics who generate leukotrienes in response to allergen exposure.

It is important to consider the limitations of the segmental antigen challenge model of asthma. This model utilizes high concentrations of antigen presented in soluble form. This challenge induces higher levels of most inflammatory mediators and more inflammatory cells than is found even during naturally occurring asthma exacerbations (25). Sampling is limited to once or twice after antigen challenge and may therefore miss peaks in some mediators after stimulation with antigen. Nonetheless, the segmental antigen challenge model appears to provide important clues about mechanisms that contribute to airway inflammation in asthma. Although our data offer a possible biologic explanation for the heterogeneous clinical response of asthma to leukotriene modifier therapy, additional studies are necessary before extrapolating these results to clinical asthma. The responsiveness to anti-inflammatory effects of 5-lipoxygenase inhibition could not be predicted from the baseline studies performed. Although the high-LT group had much higher levels of most inflammatory cytokines and more inflammatory cells in the postantigen BALF, the composition of the baseline BALF was comparable in the low- and high-LT subjects. Furthermore, there were no differences in baseline spirometry or bronchial responsiveness between the two groups. Therefore, the best method for evaluating the responsiveness of an asthmatic patient to leukotriene-modifying agents remains a careful assessment of clinical response to a therapeutic trial with these agents.

In summary, segmental antigen challenge stimulated persistent bronchoalveolar generation of leukotrienes and higher levels of cytokines and inflammatory cells in nine of 18 asthmatic subjects studied. Treatment with a 5-lipoxygenase inhibitor consistently reduced levels of some inflammatory indices, but only in the high leukotriene producers.