INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In patients with asthma who remain symptomatic in spite of low to moderate doses of inhaled corticosteroid (ICS) asthma guidelines now suggest either an increase in the dose of ICS or addition of a long-acting inhaled ₂-agonist (1). A number of studies have indicated that the choice of adding a long-acting inhaled ₂-agonist gives better clinical control than increasing ICS (2) and may also be better at decreasing exacerbation rates (3). However, there is little pathological data on the relative effects of these alternative regimens on the airway inflammation that underlies asthma, although some is now appearing in the literature (4, 5), and none at all on their implications for the long-term "remodeling" process that occurs in the airway wall (6).

Long-acting ₂-agonists were designed primarily to promote airway smooth muscle relaxation, but they may also potentiate the antiinflammatory effects of ICS on inflammatory cells (5), as well as having beneficial effects on microvascular leakage (7), which are all characteristic of asthma. On the other hand, there have been concerns that regular ₂-agonist use may worsen asthma control (8), increase mortality (9), and merely suppress symptoms, which could just disguise underlying deterioration in airway inflammation (10, 11).

Because airway remodeling, that is, structural changes in airway wall components, is thought to result from airway inflammation in asthma (5), worsening inflammation could hypothetically lead to more long-term remodeling and the potential for worse fixed airway narrowing and bronchial hyperreactivity (BHR) (11).

ICSs, such as beclomethasone dipropionate (BDP), budesonide (BUD), or fluticasone propionate (FP), are able to decrease acute airway inflammation (12). FP has been shown to have high topical potency and low clinical systemic activity at a moderate dose (13). In particular, FP has approximately twice the potency of BDP in terms of the McKenzie skin vasoconstriction assay in humans (14). There are no data that we are aware of on its effects on long-term structural changes in the airways, and relatively few data on effects of ICS in general on remodeling changes, despite their importance in asthma prophylaxis (15).

Previously, in two cross-sectional studies, our group showed that airways in subjects with asthma untreated with ICS have more vessels and a greater percentage area of vasculature in the lamina propria immediately below the epithelium than normal (6, 18). The increased vascularity indices in asthma were found to be related to functional changes in terms of lung function, BHR, and ICS dose. ICS treatment was associated with a relatively normalized airway wall vascularity in subjects with asthma at least in terms of the percentage area of airway wall occupied by vessels, although the absolute number of vessels per unit area remained high (6). Increased vascularity may therefore be a relatively easily quantifiable index of the more generalized airway remodeling process in the airway wall, and its response to medication (19).

The aims of this study were to quantify the relative effects of adding the long-acting ₂-agonist salmeterol, compared with adding FP, on the vascularity of the subepithelial lamina propria, as an index of remodeling, in symptomatic subjects with asthma already on low to moderate doses of ICS. This is the group in which long acting ₂-agonist treatment is currently indicated in asthma guidelines. It is also the group in which potential negative as well as a potential positive effects are most likely to be demonstrable, that is, partially treated inflammatory changes in the airway wall could potentially get worse as well as improve further. Our null hypothesis was that salmeterol would have no effect on airway vessel remodeling as evidenced by angiogenic parameters in airway biopsies; our concern was that salmeterol may actually increase airway wall vascularity (due to its vasodilatory action) and so worsen airway reactivity.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The study was approved by the Alfred Hospital Ethics Committee and written informed consent was obtained from all subjects prior to investigations.

Subjects and Protocol

All patients and control volunteers were current nonsmokers, none had smoked within 12 mo, and none had more than a 10 pack-year history.

Fifty patients with asthma from our volunteer database agreed to be part of this placebo-controlled, double blind, parallel group study. The full details of this clinical study have been previously published (5). At recruitment and run-in, all patients continued to have symptoms on most days in spite of being treated with 200-500 µg /day BDP or 200-400 mg/d BUD unchanged for at least the previous 12 mo. All had measurable airway hyperresponsiveness to methacholine (provocative dose [PD₂₀] FEV_1.0 methacholine < 2 mg) but their FEV₁ at baseline was at least 60% of their predicted normal value.

Forty-five of the 50 completed the study, and underwent bronchoscopy and airway endobronchial biopsy before and after 3 mo of study treatment, the last dose of study medication being taken the night before the procedure. We were able to obtain adequate paired biopsy material for sectioning and staining for vascularity in only 34 of these patients. In this study subgroup, there were 20 males and 14 females, with demographic data shown given in Table 1. For the purpose of this study we have evaluated the clinical details as well as the airway wall vascularity indices in these patients only.

Patients were randomly given one of the following treatments in addition to continuing their "background" dose of ICS: placebo (n = 11), 50 µg salmeterol twice a day (n = 12), or 100 µg FP twice a day (n = 11) in identical dry powder diskhalers for a 3-mo period. The dose of FP would have effectively doubled the mean daily dose of ICS. During the study period subjects continued to take albuterol as required. Randomization was via a computer-generated list of random numbers.

Twenty-eight normal volunteers without asthma were recruited concurrently; there were 22 males and 6 females and 14 of these were atopic (Table 1). None of these normal subjects had a history of asthma and all had a PD₂₀ methacholine > 3.9 mg. They underwent the same bronchoscopy procedures with airway biopsy, but on a single occasion only.

Clinical and Physiological Measurements

Atopy was defined as a response to one or more common aeroallergens (house dust mix, southern grass, rye grass, cat fur, feather, Cladosporium, and Aspergillus) on skin prick test with a wheal diameter of > 3 mm in the presence of a positive histamine and negative control reaction.

FEV₁ and airway responsiveness to methacholine were measured before and toward the end of the study period, in the morning after a bronchodilator-free period of at least 8 h. The second set of tests was done 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. Airway reactivity to methacholine was assessed by a previously established technique (20) and expressed as the cumulative dose required to provoke a 20% decrease in FEV₁ (the PD₂₀ methacholine, PD₂₀M) using linear interpolation from a dose-response plot.

Bronchoscopy and Sampling

All patients and control subjects at bronchoscopy were sedated with intravenous 5-15 mg midazolam and 0.4-0.6 mg intravenous atropine. If required for symptoms or if FEV₁ predicted was less than 70%, 200 µg inhaled albuterol was also given. Approximately 45% of the patients with asthma in each group had albuterol on one or both occasions. Otherwise no medication was taken by patients since the night before bronchoscopy. Eight of the 28 of the control subjects were also given albuterol prior to bronchoscopy in order to detect any acute effect of albuterol on vascularity indices as a potential confounder. Pulse oximetry was performed throughout the procedure and supplemental oxygen was provided via nasal cannulas for all subjects. The airways were anaesthetized with topical lidocaine4% above the vocal cords and 2% below in 2 ml aliquots to a total dose of  24 mg. Endobronchial biopsies were obtained with alligator forceps (Olympus code no. FB15C) from segmental subcarinal areas of the right lower lobe.

Immunohistochemical Staining of Vascular Structures

Endobronchial biopsies were snap frozen in ornithine carbamyltransferase (OCT) in an isopentane-liquid N₂ slurry immediately after bronchoscopy and stored at 80° C until indirect immunoperoxidase staining was performed.

A three-layered immunoperoxidase staining method was used as previously described (6) using a mouse anti-human collagen type IV monoclonal antibody (Dako, Denmark) to outline the endothelial basement membrane of vascular structures. Each staining run included a negative isotype immunoglobulin G₁ (IgG₁) control slide and nasal polyp positive control slide.

Slides were coded and assessed by one experienced blinded observer (B.E.O.). Sections were analyzed by a method described previously (18). In brief, sections were cut in duplicate and the one least affected by tears and holes was selected, and analyzed to a depth of 150 µm below the epithelial basement membrane using a computerized image analyzer (Video Pro 32; Leading Edge Pty. Ltd., Adelaide, Australia). Smooth muscle and glands were excluded. All structures internal to vessel endothelial basement membrane were included in the evaluated vascular area. For each patient at least five nonoverlapping consecutive high power (×40 objective) fields of one slide were assessed and results added. For each, the total number of vessels/mm² of lamina propria was determined by dividing the number of vessels counted by the total area measured. In addition, the area occupied by vessels was expressed as a percentage of the total area of lamina propria assessed. If there were less than five fields available, the section was discarded. In some individuals up to 10 fields were available, and then all were scored to avoid any selection bias.

Intraobserver and interobserver reproducibility of measurements have been previously published for this method (6).

Statistical Analysis

All vascularity indices and clinical data were presented as means and standard deviations, except for PD₂₀M, which was presented as geometric means and ranges. Clinical indices, FEV₁, and PD₂₀M both within and between groups were compared using Student's t test (two tailed). p Values  0.05 were considered as significant.

Comparison of changes that occurred to vascularity indices (vessels/mm² of lamina propria and percentage area of lamina propria occupied by vessels) after treatment within a group and between groups was assessed using an analysis of covariance model (general linear model), which included the potential confounders: the difference in total area of lamina propria measured to obtain vascularity (that is, total area of lamina propria measured in each biopsy after treatment minus total area measured before treatment), sex, age, atopy status, bronchodilator (albuterol) premedication prior to bronchoscopy, baseline FEV₁, and PD₂₀M.

A univariant correlation test (Pearson; two-tailed) was used to determine the relationship between clinical indices at baseline (PD₂₀M and FEV₁) and vascularity indices.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The details of the full clinical, physiological, and airway inflammatory profiles before and after 3 mo of study medication for the whole study population of 45 have been presented in detail elsewhere (5). The trends for physiological and clinical change in this subgroup of 34 in which we are able to obtain vascularity data were very similar to the whole group, that is, an improvement in physiological and clinical indices on both salmeterol and ICS over 3 mo of therapy, most marked for salmeterol (Table 1).

Total area of lamina propria in the biopsies available for analysis was the only significant confounder shown to have a potential independent effect on vascularity indices, that is, an increasing area of lamina propria tended to have a "diluting" effect on the vascularity quantified. This is related to the vascularity being more dense just below the epithelial basement membrane. However, there was no difference in this between groups or within groups when comparing before and after treatment. In addition, there was no difference in area of lamina propria assessed between control subjects without asthma and subjects with asthma at baseline (Table 2). Thus this potential confounder affected vascularity assessment equally between groups and before and after treatment. There was no effect on vascularity indices on whether individuals had albuterol premedication prior to bronchoscopy.

When compared with control subjects without asthma, the patients with asthma when grouped together at baseline had a significant increase in the number of vessels/mm² of lamina propria (524 ± 137 vessels/mm², n = 34 versus 425 ± 130 vessels/mm², n = 28; p = 0.004) (Table 2). However, no statistical difference was detected when comparing the percentage area of lamina propria occupied by vessels between these groups (Table 2).

There was a decrease in the number of vessels/mm² of lamina propria after treatment only within the salmeterol group compared with baseline (before, 535 ± 153 vessels/mm² versus after, 400 ± 142 vessels/mm², n = 12; p = 0.04) (Figures 1 and 2 and Table 2). The placebo and FP-treated groups did not show even a trend toward a change with treatment. The only trend seen between groups was the change in the placebo group compared with that of the salmeterol group (p = 0.15) (Figure 2 and Table 2). There was no significant change in percentage area of lamina propria occupied by vessels within any of the groups. There was no relationship demonstrated between changes in vascularity indices and physiological changes over the 3 mo of the study.

View larger version (106K): [in this window] [in a new window]   Figure 1.   Representative photomicrographs of anticollagen type IV monoclonal antibody staining in endobronchial biopsies showing vessels (arrows) in the subepithelial lamina propria of a patient (a) treated with low dose ICS and (b) after 3 mo treatment with supplementary salmeterol.

View larger version (21K): [in this window] [in a new window]   Figure 2.   Number of vessels/mm2 in lamina propria endobronchial biopsy specimens from the three asthmatic groups studiedbefore and after 3 mo of treatment.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The patients with asthma in this study had an increase in density of vascularity at baseline, that is, in total number of blood vessels per unit area of the lamina propria sampled, in spite of being on a low dose of inhaled corticosteroid. However, they did not demonstrate an increase in the percentage of lamina propria occupied by vessels. We have previously shown in a cross-sectional study that long-term ICS treatment is associated with a down regulation of the increased airway vascularity found in patients with asthma and this was more marked on total area of airway vascularity rather than on vessel number (6). Our current data are, therefore, consistent with such a differential effect, and confirm the previous findings. Importantly, this study has demonstrated for the first time that neither 3 mo treatment with salmeterol, nor FP, when added to low dose ICS treatment in patients asthma was associated with adverse affects on airway vascularity, which we have used as an index of remodeling of the airway lamina propria in asthma.

There has been controversy regarding the potential adverse effects of regular ₂-agonist therapy. Concerns included the possibility of a worsening of airway inflammation, and increased long-term damage through excessive airway remodeling, thereby enhancing airway hyperresponsiveness. In terms of airway vascularity no adverse effect was detected in this study. It should also be noted that the subjects involved in this study were typical of those in whom the addition of a long-acting ₂-agonist would be considered according to current international guidelines (1), that is, symptomatic in spite of regular inhaled corticosteroid. This gives particular relevance to the outcome of the study.

This study indicates that there was a positive effect on the vascular aspect of airway remodeling following treatment with salmeterol, with a significant decrease in number of vessels/ mm² of the lamina propria within the group and a strong trend for a difference when these changes with salmeterol were compared with spontaneous changes in the placebo group. There was no overall change in the salmeterol group in the percentage of the area of the lamina propria occupied by vessels. This may have been due to an overall dilatation of these vessels from a mean of 249 ± 74 µm² to 303 ± 105 µm², a change with salmeterol treatment that did not reach significance (p = 0.07). These figures were derived for each individual from the measurement of vessel number and the area they occupied. This is likely to represent a direct pharmacological effect of salmeterol on relaxing vascular smooth muscle. The improvement in airway responsiveness that we observed with salmeterol is therefore unlikely to be related to any change in vascularity, and more likely was due to physiological antagonism at the level of airway smooth muscle.

Increased airway vascularity may be a contributor to increased airway hyperresponsiveness and fixed airflow obstruction, which are characteristic of asthma (6). It has been shown that even a small increase in airway wall thickening, due for example to edema or vascular engorgement, may lead to excessive narrowing of the airway lumen as the smooth muscle contracts (21, 22). If long-acting ₂-agonists can reverse or stop the increase in thickness of the lamina propria, that is, inner wall thickness (20), from progressing further, and especially if this effect is additive to that of ICS, then this could be a potentially important aspect of their combined therapeutic action in the long-term management of asthma. However, the fact that we found that the vascularity as a percentage of area did not change (as above) may serve to limit the benefits of such an effect. It would be important to compare airway vascularity and methacholine responsiveness at 24 h after the last dose of ₂-agonist when any direct effect of the drug might be absent at both vascular and smooth muscle layer level, in any future study.

We did not demonstrate any further benefit to increasing the dose of ICS with addition of FP, although our data would suggest that the background dose of ICS that the patients were on was likely to already have had some effect in modifying the total volume of lamina propria occupied by vessels (6). In this previous study from our laboratory there was only a slight dose-response effect above 500 µg of BDP on area of vessels, and no significant dose-response effect of ICS on vessel number. Our current data are consistent with this. The possible beneficial effect of salmeterol at least on vessel number over 3 mo may therefore suggest a complimentary effect of long-acting ₂-agonists with ICS on this manifestation of airway remodeling, as we have previously shown on the cellular inflammatory pathology in the airways of patients with asthma (5, 23). Again, the coincidental vasodilatation that we saw with salmeterol may temper this potential synergistic benefit at least as long as the patients are under the active direct pharmacological influence of the drug.

Any effect of salmeterol on vessel number is likely to be related to an antiangiogenic effect on endothelial hyperplasia. There are few data available on the mechanisms involved in these pathological remodeling processes, although growth factors with angiogenic properties such as vascular endothelial growth factor (VEGF) and basic fibroblast growth factor (bFGF) are quite likely to be released in excessive amounts in the airways of patients with asthma (24). We have shown that interleukin (IL)-8, which is another growth factor with recently described angiogenic properties (25), is indeed increased in concentration in bronchoalveolar lavage (BAL) fluid in this same asthmatic sample population (23). Furthermore, salmeterol treatment was shown to be associated with a decrease in these levels (23). Thus direct effects of salmeterol on angiogenic growth factors may be an important aspect of its long-term effects in asthma, which requires further investigation.

₂-agonists may also have potential antiinflammatory properties that are complementary to the effects of ICS. The antiinflammatory actions of the glucocorticoids follow the activation and translocation of ubiquitously expressed cytoplasmic glucocorticoid receptors (GR) to the nuclei of cells. In recent in vitro studies, ₂-agonists, including salmeterol, caused a propranolol-sensitive, cAMP-dependent activation and nuclear translocation of GR in human lung fibroblasts and vascular smooth muscle cells (28). This provides a potential mechanism for complementary antiinflammatory cellular effects of ICS and ₂-agonists.

In conclusion, we have demonstrated that following treatment with salmeterol in patients with asthma already on ICS there was no worsening of parameters of airway vascular remodeling, which confirmed our null hypothesis. Our current data, combined with those in previous publications (6, 18), indicate a modifying effect by low doses of ICS on airway vascularity and a likely additive effect on vessel density (number per unit area) following salmeterol treatment complementary to the effects of ICS. We would hypothesize that these effects may be via modifications to angiogenic growth factors such as IL-8. We believe that this is the first time that the potential influence of salmeterol treatment on airway remodeling has been studied. Further studies are required to confirm these data and to provide insights into mechanisms of such effects, whether they are sustained over a longer time scale and how (if at all) they are translated into clinical and symptomatic benefit.