INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

There is consensus among international asthma treatment guidelines for the early introduction of inhaled corticosteroid (ICS) disease-modifying treatment, in addition to short-acting ₂ agonists as needed for relief, in all patients but those with the mildest disease. However, in many patients low-dose ICS does not lead to adequate control of asthma, and it has become a matter of debate whether in this situation it is better to increase the dose of ICS or to introduce a regular long-acting ₂ agonist, such as salmeterol or eformoterol. Taking the latter course, i.e., combining a long-acting, inhaled ₂ agonist with the current dose of ICS, has been demonstrated in international multicenter clinical trials to lead to a greater improvement in symptoms and lung function than increasing the dose of ICS (e.g., see References 1 and 2). However, there are few data on the impact of such a regimen on the status of the underlying disease process, and there have been highly publicized concerns regarding regular and long-term treatment with inhaled ₂ agonists in patients with asthma.

One suggestion has been that chronic dilatation of the airways would lead to excessive allergen deposition from the atmosphere, and as a result increased acute inflammation (3). Consistent with this has been the report that regular and sustained use of ₂ agonists may cause a worsening of asthma control and severity compared with as needed use (4), although this is controversial. There have been epidemiological studies linking use of  agonists to excess asthma mortality (5, 6), although it has been difficult to exclude confounding by disease severity (7).

There is also evidence that regular use of long-acting, inhaled ₂ agonists is associated with loss of bronchoprotection to induced bronchoconstriction (8). Such tolerance has been shown not to be reversed or preventable by ICS (9). Furthermore, underlying sensitivity to allergen challenge may actually increase on regular use of  agonists (10), interpreted as an upregulation of a mast cell-mediated acute inflammatory response.

On the other hand, there is evidence from in vitro studies and animal models to suggest that long-acting ₂ agonists could potentially exert some antiinflammatory action in addition to bronchodilatation that could contribute to their beneficial effects in asthma control (11). However, in human asthma, limited published bronchoscopic studies have for the most part shown no significant influence on airway inflammation in patients either with stable asthma or after allergen challenge (12). These were all small studies with poor power, and mainly gave bronchoalveolar lavage (BAL) data on patients with very mild disease whose potential for demonstrating either pro- or antiinflammatory effects of long-acting, inhaled ₂ agonists was limited. In another study, also of patients with mild asthma receiving as needed ₂ agonist only, Wallin and co-workers demonstrated a reduction in airway eosinophils and mast cells with inhaled eformoterol, most marked in those with initially abnormal, elevated cell numbers (15).

We have undertaken a double-blind, randomized, parallel-group, placebo-controlled study of subjects with asthma, who were already using low-dose ICS, but still symptomatic, to determine the effects of 12-wk treatment with salmeterol on "allergic" inflammation of the airways, as well as clinical status, in a clinically relevant group. In this report we have assessed especially the inflammatory components thought to be most conspicuous in asthma, i.e., mast cells, eosinophils, lymphocytes, and macrophages in BAL and bronchial biopsies. The rationale for studying patients already on low-dose ICS was to allow for potential deterioration in the salmeterol-treated group in partially treated individuals. At the same time, any antiinflammatory action could be contrasted with a positive control arm in which ICS therapy was increased with addition of inhaled fluticasone. We have included adequate numbers of subjects to give optimal power to such an intervention study (16). Our null hypothesis was that salmeterol would have no effect on airway cells, either pro- or antiinflammatory.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Asthma volunteers were 20 to 70 yr of age, all nonsmokers, with diagnosed asthma treated for at least 12 mo with ICS in a dose of up to 500 µg of beclomethasone dipropionate (BDP) or budesonide (BUD) per day. The FEV₁ at baseline had to be at least 60% of its predicted normal value. Patients who had suffered from acute respiratory tract infection during the previous 4 wk, who had any change in asthma medication, or who had been admitted to hospital with airway disease in the 4 wk before the study were excluded. All subjects gave informed, written consent before commencing the study, which was approved by the Alfred Campus Joint Ethics Committee. Eleven normal, nonsmoking, nonatopic volunteers provided control endobronchial biopsy (EBBx) data and 28 provided BAL data.

Study Design

The study was a double-blind, randomized, parallel-group study with three treatment arms. There was a 2- to 6-wk run-in period during which patients received their usual dose of ICS (up to 500 µg of BDP or BUD per day) plus inhaled albuterol (200 µg) as needed. At the end of the run-in period, patients were eligible for randomization if they remained "actively asthmatic" as defined by at least one of the following criteria: (1) had a symptom score of more than 2 on seven of the last 14 d; (2) required the use of rescue inhaled albuterol on more than seven of the last 14 d; (3) had a variation of > 15% in peak expiratory flow (PEF) over a 24-h period on at least seven of the last 14 d, plus some degree of symptoms and rescue medication use during that time.

Eligible patients were randomly assigned, using computer-generated numbers in balanced blocks, to receive one of the following treatments for a period of 12 wk, added to their continued background ICS and inhaled albuterol (200 µg) as needed: salmeterol (50 µg twice daily), fluticasone propionate (100 µg twice daily) (i.e., at least doubling the effective ICS dose in these subjects), or placebo. All intervention dry powder diskhalers were identical, and all interventions consisted of 1 disk twice per day.

Outcome Measures

Diary-card data: Patients completed daily diary cards during the run-in and treatment periods, recording the best of three PEFs measured on mini-Wright peak flow meters (Clement-Clarke, Harlow, UK) in the morning and evening before and after medication; symptoms of asthma during the night or daytime (according to a five-point scale, with 0 indicating no symptoms and 4 indicating incapacitating symptoms); number of awakenings due to asthma; and use of albuterol for rescue therapy in relief of symptoms.

Clinical indices: On the screening visit, atopic status was assessed by skin prick testing to a panel of seven common aeroallergens and was defined by a response of  3 mm to one or more.

FEV₁ and bronchial responsiveness to methacholine were measured before and toward the end of the period of intervention. These measurements were performed in the morning, after a bronchodilator-free period of at least 8 h about 1 wk before starting study medication, and again at the same time of day approximately 12 h after a dose of study medication. A Vitalograph wedge bellows spirometer (Vitalograph, Bucks, UK) was used for measurement of FEV₁ according to ATS criteria. Bronchial hyperresponsiveness (BHR) to methacholine was assessed by a previously established technique (17) and expressed as the cumulative dose required to provoke a 20% decrease in FEV₁ (PD₂₀), using linear interpolation from a dose-response plot.

Bronchoscopy and Bronchial Sampling Procedures

Bronchoscopy was performed a minimum of 2 d after methacholine challenge, before and at the end of treatment periods for patients with asthma. The last dose of study medication was taken 12 h before bronchoscopy. One bronchoscopy only was performed for normal controls. Subjects were premedicated with intravenously administered atropine (0.4 mg) and 5-15 mg of midazolam. Inhaled albuterol (200 µg) was given 15 min before bronchoscopy. Lignocaine spray (4%) was applied to the pharynx and larynx and 2% lignocaine was applied below the cords in 2-ml aliquots as required up to a maximum dose of 6 ml. Subjects were monitored by pulse oximetry and administered oxygen at 4 L/min during the procedure.

Bronchoalveolar lavage (BAL) of the right middle lobe was performed with three 60-ml aliquots of phosphate-buffered saline (PBS) and aspirated after minimal dwell time. Endobronchial biopsies (EBBx) were taken from the segmental subcarinae of the right lower lobe of each patient, using alligator forceps (Olympus, Tokyo, Japan). EBBx were transported on ice, embedded in O.C.T., and snap frozen in a liquid nitrogen-cooled isopentane slurry within 15 min and stored at 80° C.

BAL processing. Raw BAL fluid was maintained at 4° C and the volume measured. Total cell counts were performed on the unfiltered BAL fluid using a modified Neubauer hemocytometer. Cytocentrifuge preparations were made with 200 µl of unfiltered BAL aspirate (Cytospin III, 82 × g, 10 min; Shandon, Runcorn, UK). Differential cell counts were performed by counting 500 cells on duplicate slides, stained with Diff-Quik stain (Lab Supply, Melbourne, Australia). After fixation in Carnoy's fluid, toluidine blue staining was used to enumerate mast cells through metachromasia, with 5,000 cells being assessed.

Flow cytometry. BAL fluid was centrifuged at 100 × g for 15 min. BAL cell pellets were resuspended to a cell concentration of 2 × 10⁶/ ml in PBS, and 100-µl aliquots were then incubated with 30 µl of monoclonal antibodies, directly conjugated to either phycoerythrin (PE; red fluorochrome), fluorescein isothiocyanate (FITC; green fluorochrome), or peridinin chlorophyll protein (PerCP; orange fluorochrome) as a third color. The antibody combinations used were CD45/ CD14 (Leucogate; allowing the identification of lymphocytes), CD4/ CD8/CD3 (helper/inducer, suppressor/cytotoxic, and total T cells). CD4/CD25/CD3, CD3/CD4/HLA-DR (CD25 and HLA-DR on T lymphocytes as markers of activation), and an isotype control (all antibodies from Becton Dickinson, Mountain View, CA). After staining for 20 min at 4° C, red blood cells were lysed for 10 min, using a commercially available lysing buffer (Ortho mune; Ortho Diagnostic Systems, Raritan, NJ). Cells were then washed in 2 ml of PBS and centrifuged (15 min, 100 × g), and immediately analyzed with a flow cytometer equipped with an argon ion laser (FACScan; Becton Dickinson). Lymphocytes were gated with the Leucogate tube on forward and side light scatter, verified by the characteristic staining pattern of lymphocytes with CD45 (common leukocyte antigen), and lack of staining with CD14 (a monocyte/macrophage marker). Lysis II software (Becton Dickinson) was used for quadrant analysis and results expressed as the percentage of CD3-positive lymphocytes with the isotype-matched controls used to set negative cutoff.

Immunohistochemistry for inflammatory cells in endobronchial biopsies. Seven-micron cryostat sections were fixed in paraformaldehyde-lysine-periodate (PLP) for 10 min at 4° C in a humidified chamber. Monoclonal antibodies used were AA1 (to mast cells; Dako, Glostrup, Denmark), anti-EG1 and anti-EG2 (thought to recognize the preformed and cleaved form of eosinophil cationic protein, respectively; Pharmacia Diagnostics, Uppsala, Sweden), anti-CD3, anti-CD4, anti-CD8, anti-CD25 (lymphocytes; Dako), anti-HLA-DR (Dako), anti-CD68 (macrophages; Dako), as well as a negative isotype control. Sections were incubated overnight at 4° C. After washing in TRIS-buffered saline (TBS) for 10 min at room temperature, positive binding was amplified with a biotinylated secondary and tertiary antibody linked to horseradish peroxidase (Vectastain Elite ABC kit; Vector Laboratories, Burlingame, CA). Application of metal-enhanced diaminobenzidine (DAB; Dako) facilitated visualization of the positive signal as a brown precipitate contrasting with blue Harris hematoxylin counterstain. Positive staining in the lamina propria was assessed by light microscopy, by experienced observers, using a Video Pro 32 image analysis system. Two quality slides from one of the best tissue blocks from each subject were coded and scored blind as to treatment and disease status. These scores were averaged. Results were expressed as the number of positive cells per millimeter of epithelial basement membrane in quantifiable lamina propria up to a depth of 150 µm, excluding vessels, mucosal glands, and smooth muscle.

Statistical Analysis

The two-way Mann-Whitney U test was used for comparison of biopsy indices between asthmatics (at baseline) and normal controls.

Mean 24-h symptom scores, mean 24-h use of rescue medication, and mean morning peak flow over 2 wk before the end of run-in and treatment periods were used for comparison of clinical effects.

Generalized linear modeling analyses were used to compare the differences after treatment (i.e., post-pre) within and between groups. The potentially confounding variables included in the model were age, sex, atopic status, background ICS dosage and log baseline PD₂₀, and percent predicted FEV₁.

The Spearman rank correlation was used to test any association between inflammatory markers and clinical indices.

Statistical analyses was performed using the SAS version 6.12 package (18).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Seventy-one patients with asthma were enrolled in the study, 50 of whom were eligible to be randomized after the run-in period. Forty-five subjects completed the 12-wk study and only their data were used for subsequent analyses. Baseline demographic and clinical indices are presented in Table 1. Of the five withdrawals after randomization, two related to a patient decision not to continue after the first bronchoscopy, three were due to asthma exacerbation (one in the placebo group after 11 wk and two in the salmeterol group after 1 and 6 wk of therapy).

Physiological and Clinical Indices

The changes in clinical and physiological indices after treatment are shown in Figures 1 and 2. Improvement tended to occur with both salmeterol and fluticasone, but the changes were more striking and significant for salmeterol.

View larger version (28K): [in this window] [in a new window]   Figure 1.   Clinical indices after 12 wk of treatment. Data are presented as mean values with standard error bars. Sal = salmeterol group (n = 13); Pla = placebo group (n = 16); Flu = fluticasone group (n = 16).

View larger version (19K): [in this window] [in a new window]   Figure 2.   PD20 methacholine before and after 12 wk of treatment.

The geometric mean PD₂₀ improved in all three groups after 12 wk of treatment, but only the change in the fluticasone and salmeterol groups reached significance (Figure 2).

BAL

Overall, when the whole group of patients with asthma (n = 45) was compared with unaffected control subjects, there was an "asthmatic" signal for an increase in eosinophil, mast cell, and lymphocyte numbers (see Table 2).

There was no significant change in BAL profiles after any treatment (Table 2), but there was a decrease in lymphocyte activation with fluticasone, significant for HLA-DR (p = 0.05) (Figure 3) with a strong trend for CD25 (p = 0.1) (Table 2).

View larger version (21K): [in this window] [in a new window]   Figure 3.   BAL lymphocyte activation status before and after 12 wk of treatment in our three treatment groups.

Bronchial Biopsies

From the 45 completed patients, 40 pairs of biopsies were found to be adequate for analyses (14 for the placebo group, 12 for the salmeterol group, and 14 for the fluticasone group). This is in keeping with our previous experience of the "loss to follow-up" rate with airway biopsies.

In general, there were no marked changes in inflammatory cells in the bronchial lamina propria after treatment with either placebo, salmeterol, or fluticasone (Table 3). However, salmeterol was associated with a significant reduction in EG1-positive cells from a median of 18.3 to 7.6 cells/mm (p = 0.01) (Figure 4). There were no significant changes in EG1 numbers after treatment with either fluticasone (from a median of 8.2 to 5.6 cells/mm, p = 0.77) or placebo (from a median of 12.6 to 8.0 cells/mm, p = 0.68). However, when changes between groups of EG1-positive cells were compared, there was no significant difference (p = 0.33 for salmeterol versus placebo; p = 0.30 for salmeterol versus fluticasone). For EG2-positive cells, there was an apparent trend for reduction after treatment with salmeterol (from a median of 4.1 to 1.8 cells/mm) and fluticasone (from a median of 4.3 to 1.3 cells/mm) compared with placebo treatment (from a median of 2.3 to 3.3 cells/mm), but these changes were not significant (p = 0.28, p = 0.47, and p = 0.77, respectively).

View larger version (18K): [in this window] [in a new window]   Figure 4.   Total eosinophil numbers (identified by EG1 immunostaining) in the lamina propria before and after 12 wk of treatment in patients with asthma versus those in our three treatment groups. Values are also included for our unaffected control group (n = 11).

Relationship between Inflammatory Markers and Clinical Indices

There were no significant correlations between airway inflammatory cell numbers and clinical indices at baseline, or between changes in cell numbers and clinical indices with treatments.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Overall, our null hypothesis was sustained, i.e., the effect of salmeterol on allergic inflammation was largely neutral. This was true for mast cells, lymphocytes, and macrophages. However, we did find a significant reduction in airway eosinophilic infiltrate in biopsies as assessed by immunostaining with EG1 and to a lesser extent with EG2 after 12-wk treatment with salmeterol in those patients already treated with a low dose of ICS. Interestingly, increasing ICS therapy by adding fluticasone to their routine low-dose ICS did not show any significant reduction in either EG1 or EG2 cell numbers, although there was an effect on lymphocyte activation in BAL, suggesting that there was subtle additional antiinflammatory potential for an increase in ICS. Addition of salmeterol gave significant improvement in 24-h symptom scores and rescue albuterol use, morning peak flow, and PD₂₀, which were generally more marked than with fluticasone.

Airway inflammation is thought be an important pathogenic mechanism in asthma. We chose eosinophils, mast cells, and lymphocytes as the main indices of airway inflammation in this study because they have been proposed to be major effector cells in asthma, both allergic and nonallergic (19). In particular, eosinophils are increased in number in asthmatic airways (20), and are reduced after effective treatment with ICS (19). Correlations between eosinophils in the airway and severity of disease have previously been reported (21).

EG1 is a mouse monoclonal antibody that recognizes ECP, one of the major granule products of eosinophils. In contrast to EG1, which recognizes both the stored and secreted forms of ECP (22), EG2 is thought to recognize only active and secreted forms of ECP and has been taken to identify activated eosinophils (22), although this is controversial (23). In the current study, EG2-positive cells were consistently fewer in number than EG1-positive cells, indicating that these two markers are not exactly synonymous, and our data suggest that salmeterol had more effect on total eosinophils than on EG2-positive cells.

Fluticasone did not influence eosinophil numbers, probably owing to prior modification by the previous ICS. It may well be that eosinophils in asthmatic airways, especially the EG2-positive cells, are sensitive to ICS and the preexisting low-dose of ICS had already achieved maximal antieosinophil effect. Most previous bronchoscopic studies of ICS were in patients with mild asthma treated with  agonist only. Any residual eosinophilic inflammation may be relatively resistant to ICS, and may remain evident even with 2,000 µg of fluticasone per day (19). Salmeterol may have a beneficial effect on eosinophils through an alternative antiinflammatory mechanism or in some way may make the eosinophils more sensitive to ICS action. Such a potentiating antiinflammatory action of salmeterol has been demonstrated (24), and has also been shown for the effect of ICS on eosinophil apoptosis (25). Whatever the mechanism, what is clear from our study is a definite absence of deterioration in airway eosinophilia on administration of salmeterol, which was one of the central questions being addressed, and that had been widely feared (26).

To our knowledge, this is the first published report of an antiinflammatory action of a long-acting ₂ agonist based on bronchial biopsies from patients with stable asthma who are already receiving inhaled ICS, although one report showed similar but more extensive changes with eformoterol in steroid-naive patients (15). Our current finding is in contrast to our previous limited placebo-controlled, crossover study, which found no change in BAL cell counts and mast cell tryptase after 8 wk of salmeterol treatment of patients with asthma who are already taking 400 to 1,000 µg of BDP (13). However, no bronchial biopsies were obtained and the BAL cell data from the current study were comparable. Another small placebo-controlled, crossover BAL study was also negative (14) and the study by Wallin and co-workers (15) also showed changes only in biopsies, and not BAL, with eformoterol. This implies that the two modalities of bronchoscopic sampling provide rather different signals.

There are clinical studies both for and against a potential antiinflammatory action of long-acting inhaled ₂ agonists. Twentyman and co-workers found that salmeterol given before allergen challenge ablated both the early and late phase of allergen-induced bronchoconstriction, and although these aspects may have been confounded by change in baseline airway caliber, there was also inhibition of the expected increase in BHR, suggested as due to an antiinflammatory action (27). Long-acting ₂ agonists have also been shown to inhibit the rise in serum ECP with allergen challenge (28), and clinical studies have confirmed that long-acting ₂ agonists decrease exacerbation rates (2, 29) and may have ICS-sparing potential (3), again consistent with a disease-modifying action.

Conversely, Soler and colleagues (31) showed that salmeterol provided bronchoprotection against directly acting histamine, but not against adenosine 5'-monophosphate (AMP), an "indirect" bronchoconstrictor acting via mast cell activation. Wempe and co-workers found no effect of the oral long-acting -agonist bambuterol on the circadian increase in blood eosinophils and ECP levels in nocturnal asthma (32). Other studies found that salmeterol did not inhibit allergen-induced eosinophilia in peripheral blood (33) or rise in urinary leukotriene E4 level (34). One study has shown that the effect of salmeterol on asthmatic symptom control might be at the expense of masking a preexacerbation rise in sputum eosinophilia, although in this study that was in the context of a marked reduction in ICS dose (35).

Because almost every cell type is able to express the ₂ adrenoceptor, including inflammatory leukocytes, it is possible that prolonged ₂-receptor stimulation could influence some of the pathways contributing to airway inflammation in asthma. Salmeterol is able to reduce the adherence of neutrophils, eosinophils, monocytes, and lymphocytes in mucosal blood vessels after antigen challenge in sensitized rats through its action on ₂ receptors (36). Furthermore, Eda and co-workers demonstrated that eformoterol was able to inhibit platelet-activating factor (PAF)-induced eosinophil chemotaxis and degranulation in vitro (37). In guinea pigs pretreatment with salmeterol significantly inhibited eosinophil accumulation in the airway lumen induced by PAF (38). Unique non-₂ receptor-mediated membrane stabilization has been suggested as one mechanism for salmeterol downregulation of inflammation induced by PAF and other bioactive lipids (39).

Consistent with the literature, addition of salmeterol in the current study resulted in improvement in clinical indices equivalent to, if not better than, that occurring after addition of fluticasone. The improvement in PD₂₀ methacholine with salmeterol is a novel finding. This may well have been due predominantly to prolonged functional antagonism at the 12-h time point postdose, but there is also now the intriguing possibility that the demonstrated antieosinophil actions may have contributed. There was some improvement in BHR in all groups, which may reflect improvement in compliance with background ICS medication, although this was not reflected in analyzed inflammatory indices.

In conclusion, this study has shown that mild to moderate inflammation was still present in the airways of patients with symptomatic asthma who had been receiving long-term low-dose ICS. Importantly, administration of salmeterol to these patients did not cause deterioration of residual eosinophilic "allergic" inflammation, but in contrast was associated with a fall in eosinophil numbers not seen when increasing the dose of ICS with fluticasone. There was a parallel improvement in the clinical status in these patients, more marked with salmeterol than with increasing ICS. Overall, our data give histological support to current therapeutic guidelines that recommend the use of regular long-acting ₂ agonists in combination with ICS, when ICS are not adequately controlling symptoms.