INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Morphological changes in airways of subjects with asthma include acute inflammatory changes in combination with structural alterations. This latter component is termed airway remodeling and includes epithelial cell changes, subepithelial collagen deposition, changes in the composition of the extracellular matrix (ECM), in addition to smooth muscle hypertrophy and/or hyperplasia (1). Mathematical models highlight the potential contribution of airway remodeling to altered airway behavior in asthma, mainly relating various components of the remodeling process to hypersensitivity and hyperreactivity of the airways (2).

The effect of current treatment modalities on structural airway changes remains to be fully established. Remodeling is thought to result from inefficient repair of tissue damage caused by allergic airway inflammation. Hence, it can be anticipated that antiinflammatory agents, in particular inhaled corticosteroids (ICS), might influence this phenomenon. In vitro data clearly illustrate the responsiveness of structural tissue components including fibroblasts and smooth muscle cells to steroids (3). However, it remains to be fully established to what extent ICS at currently advocated doses can prevent or reverse the structural alterations occurring in vivo. Most of the human data available to date relate to the effect of ICS on subepithelial fibrosis as a parameter of airway remodeling. Comparison of these studies is hampered by differences in the steroids used, dosing regimen, or treatment duration, but the overall effect, if any, seems limited (7). A similar variable effect on growth factor expression in biopsies has also been reported (11, 13). The response to steroids on other components of remodeling, including blood vessels, the smooth muscle layer, and the adventitial ECM composition, is even less known. In a recent study, fluticasone propionate was reported to decrease mucosal blood flow, probably without affecting neovascularization (14).

The question has also been raised whether early introduction of ICS could prevent airway remodeling from occurring, and whether this would not be more effective than reversing established changes. This hypothesis is supported by open studies illustrating that postponing introduction of ICS results in a loss of the maximal achievable improvement in forced expiratory volume in 1 s and bronchial hyperresponsiveness (15- 18). Animal models of airway remodeling might provide useful information in this respect, as they allow us to evaluate the effect of ICS on various components of remodeling and relate the morphological changes to alterations in airway function. We have developed an in vivo rat model of airway remodeling (19). In this model, we have investigated the effect of treatment with fluticasone propionate on the development of airway remodeling and on established structural changes.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals

Specific pathogen-free, male Brown Norway (BN) rats (weight 200- 250 g) were obtained from Harlan CPB (Zeist, The Netherlands). After arrival the rats were kept in standard animal research facilities and received food and water ad libidum. The rats were aged between 14 and 16 wk and weighed between 250 and 350 g at the time of testing.

Study Design

All animals were sensitized by intraperitoneal injection on Day 0 and Day 7 with 1 mg ovalbumin (OA) (Grade III; Sigma Chemical Co., Poole, Dorset, UK) and 200 µg Al(OH)₃ in 0.5 ml saline.

The rats were treated as follows: To evaluate the effect of FP on the development of structural airway changes, the animals were exposed to aerosolized 1% (wt/vol) OA (Grade III; Sigma Chemical Co., Poole, Dorset, UK) for 30 min, every 2 d from Day 14 to Day 28. Thirty minutes prior to OA exposure, the rats were treated with aerosolized fluticasone propionate (FP) (Glaxo Wellcome R&D, Stockley Park, Uxbridge, UK) (10 mg in 30 ml of 0.1% ethanol/phosphate-buffered saline [PBS]) or placebo. The dose was derived from pilot experiments, showing that this dose had a limited effect on thymus weight, similar to that of 3 mg/kg prednisolone intraperitoneally. It was estimated that this dose would result in a lung deposition of approximately 4 µg/kg body weight. This would amount to daily treatment of a 70 kg adult with approximately 2,000 µg/d.

The effect of FP treatment on established structural airway changes was evaluated by challenging the rats with 1% OA for 30 min, every 2 d from Day 14 to Day 27. From Day 28 to Day 42 the animals were treated every 2 d with FP (10 mg in 30 ml of 0.1% ethanol/PBS) or placebo.

A control group consisted of sensitized animals that were exposed to PBS during the whole experiment (Day 14 to Day 42). The aerosol was generated with an ultrasonic nebulizer (Sirius Nova; Carl Heyer GmbH, Bad Ems, Germany) and drawn into a perspex exposure chamber containing awake animals. The output of the nebulizer was 3.0 ml/ min and the median particle size was 3.2 µm according to specifications of the manufacturer.

Outcome Measures

Twenty-four hours after the last aerosol exposure, the rats were anesthetized (60 mg/kg pentobarbital [Sanofi, Libourne, France], intraperitoneally and ventilated. Assessment of airway responsiveness to carbachol (carbamyl choline hydrochloride; Merck, Darmstadt, Germany) was performed as previously described (19). In brief, increasing concentrations (0.1-0.5-1-5-10 mg/ml) of carbachol were inhaled for 30 s at 5-min intervals. Cumulative dose-response curves for the changes in lung resistance (RL), expressed as the percentage change from baseline, were constructed. Immediately after measuring airway responsiveness, bronchoalveolar lavage was performed and blood was collected for measurement of OA-specific immunoglobulin E (IgE) in serum as previously reported (19).

Lung Histology

Following lavage, lungs were fixed by intratracheal instillation of 4% paraformaldehyde. The lungs were removed and sections were embedded in paraffin and cut in 2-µm-thick sections. To evaluate peribronchial infiltration with eosinophils, Congo Red (0.5% in 50% ethanol) staining was performed. The total number of eosinophils in the bronchial wall was determined. Goblet cells were quantified using Periodic Acid-Schiff (PAS) staining. The results were expressed as number of goblet cells per millimeter basement membrane. To avoid observer bias, all microscopic slides were coded prior to analysis by one observer and read blind.

Immunohistochemistry

Cell proliferation. To assess cell proliferation in the epithelium, animals were treated with 5'-bromo-2'-deoxyuridine (BrdU) (Sigma Chemical Co., Poole, Dorset, UK). BrdU was injected at a dose of 20 mg/kg on Days 22, 25, 27, and 28 (Experiment 1) or twice a week from Day 14 through Day 41 (Experiment 2). Immunohistochemical staining for BrdU with mouse anti-BrdU (Nycomed Amersham Plc., Buckinghamshire, UK) was performed using the streptavidin-biotin method, as previously described (19). The number of BrdU-positive cells in airway epithelium is expressed per millimeter basement membrane.

Fibronectin. Immunohistochemical staining of fibronectin was carried out as previously described (19). Briefly, after blocking the sections for endogenous peroxidase activity and nonspecific binding, they were incubated for 1 h with goat anti-rat fibronectin (Calbiochem-Novabiochem GmbH, BadSoden, Germany) and for 30 min with biotinylated rabbit anti-goat (DAKO A/S, Glostrup, Denmark). After incubation for 30 min with the streptavidin-biotinylated horseradish peroxidase complex (Nycomed Amersham Plc., Buckinghamshire, UK), the chromogenic substrate 3,3'-diaminobenzidine (DAB) (DAKO A/S) was added for 10 min, giving a brown reaction product. Sections were counterstained with hematoxylin. As a negative control, the primary antibody was substituted with normal goat serum.

Analysis of Airway Morphometry and Fibronectin Deposition

Morphometric analysis of the tissue sections stained with Congo Red was performed by light microscopy and measurements were performed using a Zeiss Axiophot microscope (Zeiss, Oberkochen, Germany) at magnification of ×400. For each treatment group three tissue sections of 7 to 16 animals were analyzed. Only large airways cut in reasonable cross sections (defined by a ratio of minimal internal diameter to maximal internal diameter less than 0.5) were examined. Airways with a length of basement membrane (P_bm) > 2 mm were defined as being large (20). A mean of 27 large airways (range 18 to 37) per experimental group was analyzed. The following parameters were measured: length of the basement membrane of the epithelium (P_bm), the perimeter of the outer border of the smooth muscle (P_mo), and the outer adventitial perimeter (P_o); the areas defined by these parameters (A_bm, A_mo, and A_o) were also measured. The following parameters were calculated based on the measured values: the total bronchial wall area (WA_t = A_o  A_bm), the inner wall area (WA_i = A_mo  A_bm), and the outer wall area (WA_o = A_o  A_mo). The airway wall area was normalized to the squared length of basement membrane.

Fibronectin deposition was analyzed by light microscopy and quantitative measurements were performed using a computerized image analysis system (LEICA Q500MC; Leica Cambridge Ltd., Cambridge, UK) as previously described (19). For each experimental group three lung sections of 5 to 11 rats were examined. The amount of fibronectin was measured in a mean of 20 large airways (range 19 to 21). The airways cut in cross sections according to the previously mentioned criteria were analyzed. The area stained for fibronectin is expressed per millimeter basement membrane.

Data Analysis

Reported values are expressed as mean ± standard error of the mean (SEM). The cumulative dose-response curves of bronchial responsiveness were compared through analysis of variance (ANOVA). Cellular composition of bronchoalveolar lavage fluid (BALF), OA-specific IgE, goblet cells, and BrdU-positive cells in the epithelium and peribronchial eosinophils for the different experimental groups were compared via Kruskal-Wallis test for multiple comparisons. When significant differences were observed, pairwise comparisons were made using a Mann-Whitney U test with Bonferroni corrections. For the analysis of the morphometry and the measurement of fibronectin deposition in the airway wall the data of the different rats were pooled together. To ensure that a similar range of airway sizes was being compared, the frequency distribution of P_bm of the airways of each experimental group was compared using the Kolmogorov-Smirnov test. The mean values of WA_t normalized to (P_bm)² and WF normalized to P_bm were compared between the experimental groups using one-way ANOVA with post-hoc (LSD and Scheffé) tests. p Values less than 0.05 were considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The Effect of FP Treatment on the Development of Airway Structural Changes

Inflammatory cell infiltrate in BALF. The total cellularity of BALF recovered from sensitized and OA-exposed animals was not significantly different from PBS-exposed animals. Differential cell count revealed a significantly increased percentage of eosinophils and neutrophils and a decrease in the percentage of macrophages in the OA-exposed animals (p < 0.0001). Treatment with FP decreased the percentage of eosinophils (p < 0.0001) but did not change the percentage of neutrophils and macrophages. The percentage of eosinophils was still significantly increased versus the PBS-exposed animals (p < 0.005) (Table 1).

OA-specific IgE. OA-specific IgE levels in serum were elevated in sensitized rats challenged with OA when compared to PBS-exposed rats (p < 0.0001). This was unaffected by treatment with FP (Table 2).

Lung histology. Lungs of OA-exposed animals showed patchy inflammatory infiltrates that were located predominantly around the airways. In addition to mononuclear cells, these infiltrates contained increased numbers of eosinophils (p < 0.05 versus PBS) (Table 2). There was a thickening of the airway wall, associated with increased fibronectin deposition. There was no apparent denudation of the epithelium. OA-exposure also induced goblet cell hyperplasia (p < 0.0001 versus PBS) and epithelial cell proliferation (p < 0.01), shown by BrdU incorporation. In our model, we did not observe thickening of the smooth muscle layer either by morphometric analysis or BrdU incorporation (cf. 19).

Treatment with FP inhibited both the increase in the number of goblet cells and proliferating BrdU⁺ cells per millimeter basement membrane in the airway epithelium (p < 0.01 and 0.005 versus Pl/OA) (Table 2). The number of infiltrating eosinophils in the bronchial wall was decreased by treatment with FP, but this did not reach significance, although the number of infiltrating eosinophils was no longer different from PBS-exposed control animals (Table 2).

Morphometry and fibronectin deposition. The total airway wall area, normalized to the squared length of basement membrane, was significantly increased when animals were repeatedly exposed to OA (0.0685 ± 0.0038 versus 0.0247 ± 0.0023 for PBS-exposed animals; p < 0.0001). Treatment with FP partly inhibited this increase (0.0474 ± 0.0033; p < 0.0001), but the total area was still significantly thickened compared with that of control animals (p < 0.0001) (Figure 1A).

View larger version (14K): [in this window] [in a new window]   Figure 1.   Effect of concomitant treatment with placebo (striped bar) and FP (solid bar) on WAt (A) and amount of fibronectin (B) in large airways during 2 wk of OA exposure versus PBS exposure (open bar) (*p < 0.01 versus PBS, + p < 0.0001 versus PBS, ++ p < 0.01 versus Pl/OA, § p < 0.0001 versus Pl/OA).

Fibronectin deposition in large airways was significantly increased after 2 wk of allergen exposure (3.688 ± 0.462) in comparison with PBS (0.933 ± 0.173; p < 0.0001). This increase was partly inhibited by FP treatment (2.274 ± 0.283; p < 0.01), but fibronectin deposition was still significantly enhanced compared with PBS-exposed animals (p < 0.01) (Figure 1B).

Airway responsiveness to carbachol. OA-exposure for 2 wk increased airway responsiveness to aerosolized carbachol in sensitized BN rats when compared with PBS-exposed animals. This is illustrated by a leftward shift of the dose-response curve (p < 0.0001) (Figure 2A). Treatment with FP partly prevented this increase in reactivity (p < 0.01 versus placebo-treated OA-exposed animals). However, the dose-response curve of the FP-treated OA-exposed animals remained significantly different from the PBS-exposed control group (p < 0.005).

View larger version (9K): [in this window] [in a new window]   Figure 2.   Effect of concomitant treatment (A) and posttreatment (B) with placebo (closed squares) and FP (closed triangles) on OA-induced hyperresponsiveness in sensitized BN rats versus PBS (open squares) (ANOVA Pl/OA versus PBS, p < 0.0001; FP/OA versus PBS, p < 0.005; FP/OA versus Pl/OA, p < 0.01; OA-Pl versus PBS, p < 0.001; OA-FP versus PBS, p < 0.0001).

The Effect of FP on Established Airway Remodeling

Inflammatory cell infiltrate in BALF. After interrupting OA exposure for 2 wk, total and differential cell counts in BALF were back to control levels both in the FP and placebo-treated animals (Table 1).

OA-specific IgE. OA-specific IgE levels in serum decreased when the animals were no longer allergen exposed for 2 wk, but remained significantly higher than in the PBS-exposed group (p < 0.0001). Again there was no difference between placebo and FP treatment (Table 2).

Lung histology. The number of eosinophils in the peribronchial area and the number of goblet cells in placebo-treated OA-exposed animals were no longer increased compared with the control group (Table 2). In the FP-treated OA-exposed animals the number of eosinophils was significantly lower than in the control group (p < 0.01). No effect was observed on the number of goblet cells (Table 2).

The amount of BrdU-positive cells in the epithelium was still significantly increased in the placebo-treated OA-exposed animals (p < 0.05 versus PBS). FP treatment caused a significant increase of epithelial cells with mitogenic activity (p < 0.0001 versus PBS and p < 0.005 versus OA-Pl) (Table 2).

Morphometry and fibronectin deposition. The increase in total airway wall area of large airways induced by OA persisted at the end of the 2-wk allergen-free interval (0.0437 ± 0.0035 versus 0.0247 ± 0.0023 in control animals; p < 0.001). This was unaffected by FP (0.0460 ± 0.0026) (Figure 3A).

View larger version (12K): [in this window] [in a new window]   Figure 3.   Effect of posttreatment with placebo (striped bar) and FP (solid bar) on WA t (A) and amount of fibronectin (B) in large airways after 2 wk of OA exposure versus PBS exposure (open bar) (*p < 0.01 versus PBS, + p < 0.001 versus PBS, ++ p < 0.0001 versus PBS).

The total amount of fibronectin in the bronchial wall, normalized to the length of the basement membrane, remained elevated after the 2-wk allergen-free period (2.371 ± 0.422 versus 0.933 ± 0.173 in control animals; p < 0.01). Treatment with FP did not change the fibronectin deposition (2.370 ± 0.376) (Figure 3B).

Airway responsiveness to carbachol. At the end of the 2-wk allergen-free interval (Day 43), OA-exposed animals maintained increased airway responsiveness to carbachol in comparison with the control group (p < 0.001) (Figure 2B). The airway responsiveness was unaffected by treatment with FP (p < 0.0001 versus PBS).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In addition to inflammatory cell infiltration, structural alterations, the so-called airway remodeling, are observed in airways of subjects with asthma. It has been hypothesized that airway remodeling has important consequences on airway function. Accelerated decline in lung function, lack of airway distensibility, and especially bronchial hyperresponsiveness have been related to airway remodeling. Preventing or reversing airway remodeling is therefore considered an important therapeutic target in asthma. However, the exact pathophysiological mechanisms underlying remodeling and the functional consequences remain to be fully established. Hence, when evaluating the effect of treatment on aspects of remodeling, it is important to evaluate a range of histological changes in relation to indices of airway function. Evaluation of remodeling in bronchial biopsies is largely restricted to measuring the degree of subepithelial fibrosis. Obvious ethical and technical limitations prohibit a more complete assessment that includes thickness of the smooth muscle layer, adventitial changes, and morphometry of the total airway wall. In vivo animal models can provide useful additional information in this respect. Most of the currently developed in vivo models of allergen-induced airway inflammation have focused on acute inflammatory changes, accompanied by a transient increase in airway responsiveness. Small laboratory animal models that display more chronic changes are sparse, but are increasingly being developed (20- 23). In our model, OA-exposure of sensitized rats induced, in addition to an influx of eosinophils around the airways, goblet cell hyperplasia, epithelial cell proliferation, increased fibronectin deposition, and an increase in total wall area. These histological alterations were accompanied by airway hyperresponsiveness (AHR) to inhaled carbachol. Importantly, after a 2-wk allergen-free period, goblet cell hyperplasia and airway eosinophil infiltration regressed, but the other histological changes as well as AHR were still observed. We believe that the persistence of these alterations adds to its value as a representative in vivo model of human asthma.

The effect of ICS on airway remodeling in asthma remains to be further explored. In the present study we addressed the effect of inhaled steroids both in preventing airway remodeling and reversing established changes. Currently available biopsy studies illustrate a variable effect on subepithelial fibrosis or expression of growth factors (7). One of the features of remodeling in asthma is increased fibronectin deposition beneath the basement membrane (24, 25). To the best of our knowledge, the effect of steroids on this phenomenon has not been directly addressed. In vitro studies yield conflicting data. Some studies mention an inhibition of fibronectin production by glucocorticoids (26, 27), whereas in other experiments an increase is observed (27). In the present in vivo study, established enhanced fibronectin deposition induced by a 2-wk allergen exposure is not influenced by treatment with FP. However, simultaneous treatment partly inhibits the OA-induced increase. It has been postulated that fibronectin in the airways, in addition to the contribution of microvascular leakage associated with local inflammation, is mainly produced by fibroblasts (28) and injured epithelial cells (29), in response to a wide range of mediators, cytokines, and especially growth factors (30, 31). Biopsy studies indicate a variable limited effect of ICS on growth factor expression (11). The present data would also indicate that steroids do not fully inhibit the allergen-induced growth factor synthesis resulting in fibronectin deposition. However, this needs to be addressed further.

The effect of steroids on epithelial cell function and the ensuing consequences on the remodeling process also need to be fully established. It has been postulated that one of the basic elements of remodeling in asthma consists of reactivation of the epithelial-mesenchymal trophic unit. Increased activity of the epithelium, which fails to go through a normal repair process following injury, could lead to continuous stimulation of the underlying myofibroblasts and therefore be one of the driving factors behind remodeling (32). Steroids are traditionally considered to delay wound healing and cause epithelial atrophy through inhibition of growth factor production (33). However, these effects are generally seen with high doses of steroids in tissue that is not continually exposed to inflammatory mediators (34). Biopsies from steroid-treated subjects with asthma, on the other hand, show increased expression in epithelial cells both of the antiapoptotic marker Bcl-2 and proliferating cell nuclear antigen (PCNA). This indicates that under circumstances that prevail in airways of subjects with asthma, steroids increase both the survival and proliferation of cells, suggesting that they can help in inducing epithelial repair (35). The observations in the present study that treatment with FP increases BrdU incorporation in the epithelial layer support this hypothesis. On the other hand, concomitant treatment with FP inhibits epithelial cell proliferation. This could probably be due to the antiinflammatory effect of FP through which epithelial injury by inflammatory mediators could be avoided, and the need for reepithelialization be obviated.

In the present study, the number of goblet cells after a 2-wk allergen-free period was no longer significantly higher than in the control animals. Therefore we could not evaluate a possible effect of treatment with FP on goblet cell hyperplasia. This is in contrast with the experiment of Mathur and coworkers (36) who reported that treatment of sensitized and allergen-exposed mice with dexamethasone reduced established goblet cell hyperplasia. This apparent discrepancy could at least in part be explained by differences in the design of the study. Mathur and coworkers evaluated the number of goblet cells 10 d after discontinuing allergen exposure, whereas in our study 14 d had elapsed. During this additional time interval, goblet cell hyperplasia could have spontaneously regressed. This is supported by a study by Blyth and coworkers (37) showing that at 11 d but no longer at 14 d after challenge a significant difference in goblet cell number persisted between allergen-exposed and control mice. Importantly, species differences could in part explain these discrepancies, as in contrast to rat airways, murine airways are largely lined with Clara cells that may simply transform into goblet cells without involvement of damaging processes.

The present study also supports the hypothesis that airway wall thickness is linked to bronchial responsiveness. Fourteen days after the last OA-exposure, the rats still showed increased AHR to inhaled carbachol. This was not affected by treatment with FP. This correlates with the results of the morphometric analysis. The total airway wall was, albeit less, still significantly thickened compared with PBS-exposed animals. On the other hand, simultaneous treatment with FP largely inhibited allergen-induced increase in airway wall thickness and AHR. The mechanisms underlying the thickening of the airway wall remain to be further examined, but it is tempting to speculate that this is largely due to changes in the ECM composition. In our study, the effect of posttreatment and concomitant treatment with FP on airway wall thickness correlated to a large extent with the effects on fibronectin deposition. Of interest also is that AHR persisted despite the disappearance of airway eosinophil infiltration. This observation is in concordance with data showing that reducing airway eosinophils with anti-interleukin-5 (IL-5) in already sensitized animals does not reduce antigen-induced AHR (36) and the finding that allergen-induced airway hyperresponsiveness can develop in the absence of eosinophilic airway inflammation (38). These and other observations both in animal models and in humans further illustrate that eosinophils and BHR are not always causally linked.

Finally, our observations also support early intervention with inhaled corticosteroids as an approach to prevent airway remodeling. It has been reported in a few open clinical studies that delayed introduction of inhaled corticosteroids results in a decrease of the maximal improvement in lung function and airway hyperresponsiveness (15, 39). Assuming that these functional alterations relate to airway remodeling, this suggests the presence of an enhanced degree of airway remodeling, irresponsive of steroids. This hypothesis remains to be fully proven in humans, but our study underlines this concept by illustrating that established airway wall thickening and fibronectin deposition respond very poorly to steroids, whereas the same dose given simultaneously largely prevents these alterations. These morphological differences correspond strikingly to the different effect on airway responsiveness to carbachol.

In conclusion, in the present study, allergen-induced structural alterations could not be reversed by treatment with inhaled corticosteroids, but concomitant treatment could partly prevent these changes and this was accompanied by an improved airway responsiveness. This supports the hypothesis that early intervention could prevent airway remodeling in asthma.