INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Airways inflammation plays a key role in asthma and is responsible for epithelial destruction, muscular hypertrophy, basement-membrane thickening, and increased mucus production. The airway inflammatory exudate in mild asthma (A) consists mainly of eosinophils, T-cells, and mast cells (1). The pathologic changes associated with status asthmaticus (SA) differ from those in A in the severity of the lesions (4). Moreover, in SA patients requiring mechanical ventilation (MV), the inflammatory exudate in bronchial lavage fluid (BLF) is mainly composed of neutrophils (about 80% of total cells), and so differs from that observed in A (4). Nevertheless, bronchial neutrophilia in SA was clearly associated during the first 48 h after the beginning of MV with an influx of eosinophils associated with a large release of eosinophil cationic protein (ECP) (8).

Several CC chemokines have been implicated in eosinophil recruitment (9). Eotaxin seems specific for eosinophil activation (10), whereas regulated on activation, normal T-cell expressed and secreted (RANTES) and monocyte chemotactic protein (MCP)-3 are active on eosinophils as well as on other leukocytes (11). Macrophage inflammatory peptide (MIP)- 1 mainly recruits monocytes and basophils, as well as inducing a moderate recruitment and activation of eosinophils (16). MCP-1 is a potent chemoattractant factor for mononuclear cells and basophils (17). The cytokine interleukin (IL)-5 is also involved in eosinophil differenciation, maturation, and migration (18). These different chemokines and IL-5 have recently been implicated in A (12).

Because of the specificity of inflammation associated with SA during MV, we suspected some differences in the profiles of secretion of CC chemokines and IL-5 between patients with SA and those with A. In the present study we compared the concentrations of CC chemokines and IL-5 in BLF from patients with SA, patients with A, and controls consisting of ventilated controls (V) and healthy volunteers (C). Moreover, we measured eosinophil chemotactic activity in these subjects' BLF and we evaluated the roles of eotaxin, MCP-3, RANTES, and IL-5, using specific neutralizing antibodies. We found that BLF from SA patients contained increased levels of CC chemokines and IL-5, and increased chemotactic activity for eosinophils, as compared with BLF from the other groups.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

The study completed data already published on the cellular influx and inflammatory activity of BLF in SA (from nine patients) (8, 21). Characteristics of the patients with SA, those with A, nonasthmatic patients undergoing mechanical ventilation (V), C are summarized on Table 1. One patient in the SA group was a current smoker and two were ex-smokers, whereas two patients in the V group were active smokers. There were no smokers or ex-smokers in the A and C groups.

Asthma was defined according to the American Thoracic Society (ATS) criteria (22). In patients with A and SA, skinprick tests were performed to common aeroallergens (Dermatophagoides pteronyssinus, D. farinae, cat and dog danders, pollens, mold). Total serum IgE was determined with the Phadebas paper radioimmunosorbent test (Pharmacia Diagnostics, Uppsala, Sweden). Atopy was defined by the occurrence of two or more positive skinprick tests. All patients with SA and A had a reversible obstructive pattern with an FEV₁ of 20% after 200 µg of albuterol and/or bronchial hyperreactivity to metacholine (provocative concentration causing a 20% decrease in FEV₁ [PC₂₀] < 8 mg/ml). Asthma severity was evaluated with the asthma and allergy severity score (AAS) on a scale from 1 to 5 (23). The severity of asthma did not differ between patients with A and patients with SA when studied at times distant from acute episodes.

The 10 SA patients were admitted to the intensive care unit (ICU) of the University Hospital of Lille for acute respiratory failure linked to asthma, and required tracheal intubation and positive-pressure MV. Dates of recruitment for the patients were from January 1995 to September 1998. The triggering event was identified in five SA patients as a viral infection in two patients, psychologic conflict in one, and inhalation of nonspecific irritants in two cases. As defined by the interval between the first symptom of the acute attack of asthma and respiratory failure requiring MV for less or more than 3 h, respectively, the onset was acute for three patients and progressive for seven patients. At the time of admission, treatment consisted of systemic methylprednisolone (2 mg/kg/d), a ₂-agonist (5 to 20 µg/kg/d), curare, and benzodiazepine, according to ATS guidelines. Hypoventilation with various degrees of permissive hypercapnia was necessary in all patients (peak airway pressure < 50 cm H₂O, decreased inspiratory-to-expiratory [I:E] ratio). The fraction of inspired oxygen (FI_O2) was adapted to obtain a Pa_O2 above 60 mm Hg. Patients with pneumonia were excluded from the study.

Eleven patients with A were included in the study. Exclusion criteria were evidence of infection, treatment with systemic or inhaled steroids, or hospitalization during the previous 6 wk. FEV₁ (% predicted) was 87.2 ± 6.4% (mean ± SEM) in this group of patients. The PC₂₀ metacholine was 1.5 ± 0.6 mg/ml.

Four patients with benzodiazepine overdoses requiring mechanical ventilation, who presented with atelectasis, served as the V group. These patients were admitted to an ICU, were nonasthmatic, and had no history of atopy, no respiratory disease, and no infection in the preceding 6 wk.

Eight healthy nonsmoking volunteers (C group) were included in the study. None had a history of asthma, atopy, or tobacco use. Lung function was within the normal range. The study was approved by the Ethics Committee of the University Hospital of Lille.

Fiberoptic Bronchoscopy and Bronchoalveolar Lavage

Bronchial lavage (BL) was performed in patients with SA only when therapeutic benefits could be expected (atelectasis or refractory SA despite optimal medical treatment) (23, 24). Nineteen BL procedures were performed in the 10 SA patients; six of these patients had more than one BL procedure. Indications for bronchoscopy were atelectasis in eight of 19 BL procedures and attempts to remove distal mucus plugs in 11 procedures. In nine cases, significant improvement in clinical parameters and/or of the chest radiograph was noted after BL. Fiberoptic bronchoscopy was performed through an endotracheal tube adaptor designed to minimize air leak (Model 514900; Rüsch AG, Kernen, Germany). FI_O2 was increased to 1.0 for 15 min before and 15 min after the procedure. After wedging of the tip of the bronchoscope into different segmental bronchi (guided by a chest radiograph in cases of atelectasis), BL was performed by infusing two or three 15-ml aliquots of sterile 0.9% saline solution at room temperature. Each aliquot was immediatly aspirated and the recovered fractions were pooled. Characteristics of the sites of BL are reported in Table 2.

Controls in the V group who presented with typical features of atelectasis were lavaged with a similar protocol.

In patients with A and in healthy volunteers, fiberoptic bronchoscopy was performed after local anesthesia with lidocaine 2% applied to the upper respiratory tract. BL was performed in a segmental bronchus of the right middle lobe by slow infusion of two 15-ml aliquots of sterile 0.9% saline solution. The different fractions of BLF were pooled. During bronchoscopy, oxygen was available and the patient had an intravenous infusion to provide venous access if needed.

BLF Processing and Cell Evaluation

An aliquot of 5 ml of each BL was sent for quantitative and qualitative cultures (the diagnosis of lower respiratory tract infection was defined as > 10⁶ pathogens/ml). Among the 19 BL procedures performed in SA patients, three were positive: one on Day 9 with Pseudomonas aeruginosa at 10⁸ cfu/ml; one with Streptococcus pneumoniae at 10⁶ cfu/ml on Day 4, and one with Proteus mirabilis 10⁷ cfu/ml on Day 1. All other BL were found to be sterile. Total and differential cell counts were performed on all BL. At least 400 cells in each cytospin preparation were read by two investigators blinded to the study groups, and the results were averaged. Results are expressed as the percentages and numbers of cells per milliliter of recovered fluid. Cell-free supernatants (BLF) were collected by centrifugation, aliquoted, and kept frozen at 70° C until the day of biologic assay.

Mediator Levels in BLF

Concentrations of IL-5 (Coulter-Immunotech; Marseille, France), MCP-1, MIP-1, RANTES (R&D Systems, Abingdon, UK), MCP-3, and eotaxin (Pharmingen, San Diego, CA) were measured with enzyme-linked immunosorbent assays (ELISAs) according to the respective manufacturers' protocols, and the results were expressed in pg/ml. The sensitivities of the assays were 1, 5, 7, 2.5, 10, and 20 pg/ml for IL-5, MCP-1, MIP-1, RANTES, MCP-3, and eotaxin, respectively. The assays used in the study did not show cross-reactivity or interference from any other cytokines or CC chemokines, including MCP-1, -2, -3, and -4; MIP-1 and -; eotaxin; or eotaxin-2. One-to-one dilutions of BLF with sample diluent were made for testing. The sensitivity of the assays was not affected by the addition of BLF from the different groups of patients. The percentage recovery of a dose of recombinant cytokine was evaluated after dilution with BLF (1:2 [vol/vol]) from SA patients. The percentage recoveries were 91%, 70%, 82%, 80%, 85%, and 77% for IL-5, MCP-1, MCP-3, RANTES, MIP-1, and eotaxin, respectively, with similar results at high or low concentrations of the respective recombinant cytokines. These percentages were not significantly different from those obtained with BLF from controls or from A patients.

Chemotactic Activity of BLF

Human eosinophils were purified from peripheral blood of asthmatic patients with a magnetic cell separation (MACS) system as previously described (25). Briefly, granulocytes (eosinophils and neutrophils) were separated from blood and eosinophils were then purified by neutrophil depletion with CD16 microbeads (Miltenyi Biotec, Bergish Gladbach, Germany). The negative immunomagnetic fraction yielded highly purified eosinophils (98.1 ± 0.4%). Contaminating cells corresponded to 1.6 ± 0.4% lymphocytes and 0.3 ± 0.2% neutrophils. Eosinophils were then resuspended in Hanks' balanced salt solution (HBSS) (Life Technologies, Buckinghamshire, UK) supplemented with 0.1% of fetal calf serum (HBSS-FCS) at 1.5 × 10⁶ cells/ml.

The chemotaxis assay was developed with a 48-well microchemotaxis assembly (Neuro Probe; Cabin John, MD) with a 5-µm-pore polycarbonate filter (Nuclepore Corp., Pleasanton, CA). After incubation at 37° C for 1.5 h, the number of eosinophils migrating through the filter was determined microscopically with ×1,000 magnification. Results were expressed as the difference between eosinophil number per high-power field (hpf) with native sample or positive control minus the eosinophil number obtained with medium alone (10.9 ± 1.7 eosinophils/hpf). In preliminary experiments, we determined that a 1:20 (vol/vol) dilution of BLF was optimal in the chemotaxis assay. Eotaxin (200 ng/ml) (R&D Systems, Abingdon, UK) was used as a positive control and induced an increase in migration of 42 ± 6.1 eosinophils/hpf.

To determine the role of IL-5 and CC chemokines in the BLF- dependent eosinophil chemotactic activity, BLF specimens were preincubated for 1 h at 37° C with neutralizing anti-IL-5, anti-eotaxin, anti-MCP-3, anti-RANTES (R&D Systems) mouse monoclonal antibodies (mAbs) or with an isotype control mouse mAb (Pharmingen, San Diego, CA) at final concentrations of 10 µg/ml. The eosinophil chemotactic activity of treated BLF was then evaluated as described, and was compared with the corresponding control (diluted BLF with an irrelevant antibody). Results were expressed as percent inhibition, defined as: 100 × ([BLF  baseline level]  [BLF with antibody  baseline level])/(BLF  baseline level). The specific antibodies neutralized about 80% of the eosinophil chemotactic activity of the corresponding human recombinant cytokines.

Statistical Analysis

Group data are expressed as mean ± SEM. Because most of the data were not normally distributed, differences between the four subject groups was assessed with the nonparametric Kruskal-Wallis test. When this test indicated significant differences, the difference between two groups was tested for significance with a Mann-Whitney U test. The correlations between cytokines, total number of eosinophils, and ECP levels were evaluated with the nonparametric Spearman's rank correlation coefficient test. A value of p < 0.05 was considered statistically significant. Data are expressed as box plots. Horizontal lines represent the median, squares the 25th and 75th percentiles, and error bars the 10th and 90th percentiles. Isolated circles represent outliers.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cellularity in BL

Statistical analysis of BL cellularity was reevaluated in comparison with previously published results, since two new patients with SA were added (three BL procedures) (8, 21). The total number and percentage of neutrophils were significantly higher in the SA than in the V (p < 0.05), A (p < 0.01), and C (p < 0.001) groups. The total number and percentage of eosinophils were higher in the SA than in the V (p < 0.01) or C (p < 0.05) groups. Compared with that in the A group, the total number of eosinophils was significantly increased in the SA group (p < 0.01), although the percentage was not significantly different (Figure 1). In the group of SA patients, when we compared the values for the 10 BL procedures performed in the first 48 h after the beginning of MV with the nine procedures performed after this 48-h period, neither the number nor percentage of neutrophils were different. In contrast, the total number as well as the percentage of eosinophils were significantly increased in the first 48 h of MV as compared with those after the first 48 h (p = 0.007 and p = 0.01, respectively) (Figure 1).

View larger version (15K): [in this window] [in a new window]   Figure 1.   (A) Total eosinophil number in BLF of the SA, A, V, and C groups. (B) Total number of eosinophils in BLF from SA patients before and after 48 h following the beginning of MV. Data are expressed as box plots. Horizontal lines represent the median, squares the 25th and 75th percentiles, and error bars the 10th and the 90th percentiles. Circles represent outliers below the 10th and above the 90th percentiles. *p < 0.05 compared with V patients; #p < 0.05 compared with A and C groups

Levels of Chemokines (MCP-1, MCP-3, MIP-1, RANTES, and Eotaxin) and IL-5 in BLF

The concentrations of MCP-1 were significantly higher in BLF from SA patients than in those from other groups (Figure 2). MCP-1 levels were significantly higher in the A and V groups than in the C group. A nonsignificant trend toward higher levels of MCP-3 (Figure 2) in the SA as compared with the V, A, and C, groups was observed (p = 0.08; p = 0.09; p = 0.1, respectively). Concentrations of MIP-1 and RANTES (Figure 2) were significantly greater in the SA than in the V, A, or C groups. There was no difference for either of these chemokines among the V, A, and C groups. Levels of eotaxin (Figure 2) were increased in the SA as compared with the V and C groups, but the differences did not reach statistical significance (p = 0.06 and p = 0.1, respectively). Similar levels of eotaxin were observed in BLF from the A and SA groups, and this chemokine was significantly increased in the A as compared with the V group (p = 0.02). IL-5 concentrations were significantly higher in the SA than in the V, A, and C groups (Figure 2); a significant difference was also observed between the A and C groups.

View larger version (22K): [in this window] [in a new window]   View larger version (24K): [in this window] [in a new window]   View larger version (16K): [in this window] [in a new window]   View larger version (28K): [in this window] [in a new window]   View larger version (24K): [in this window] [in a new window]   View larger version (21K): [in this window] [in a new window]   Figure 2.   Concentrations of CC chemokines and IL-5 in BLF of SA, A, V, and C groups. Levels of MCP-1 (A), MIP-1 (B  ), RANTES (C ), MCP-3 (D), eotaxin (E ), and IL-5 (F  ) are shown. Data are expressed as box plots. Horizontal lines represent the median, squares the 25th and 75th percentiles, and error bars the 10th and 90th percentiles. Circles represent outliers below the 10th and above the 90th percentiles. *p < 0.05 compared with V patients; #p < 0.05 compared with A and C groups.

Correlations between chemokines or IL-5 and total number of eosinophils or ECP are reported in Table 3. A significant correlation between ECP levels and eosinophil numbers was observed in asthmatic subjects. Concentrations of MCP-1 were correlated with ECP but not with the number of eosinophils. There was a significant relationship between total number of eosinophils and MCP-3. Levels of MIP-1, RANTES, and IL-5 were correlated with total number of eosinophils and ECP levels. The concentration of eotaxin was correlated with eosinophils but not with ECP.

Chemotactic Activity of BLF on Eosinophils

Chemotactic activity of BLF was significantly greater in the SA (14.3 ± 2 eosinophils/hpf) than in the A (5.7 ± 0.7 eosinophils/hpf), V (3.3 ± 0.7 eosinophils/hpf), and C (2.9 ± 1.2 eosinophils/hpf) groups (p < 0.01) (Figure 3). Chemotactic activity was also significantly greater in the A than in the C group, whereas the difference from the V group was not significant (p = 0.08). In the SA group, anti-RANTES, anti-MCP-3, anti-eotaxin, and anti- IL-5 neutralizing mAbs blocked BLF-dependent eosinophil migration by 24.6 ± 14.2%, 46.3 ± 38.7%, 48.9 ± 33.5%, and 49.2 ± 37%, respectively, as compared with the corresponding controls (Figure 3). Levels of MCP-3 and eotaxin were positively correlated with the eosinophil chemotactic activity of BLF (r = 0.7, p = 0.01; and r = 0.65, p = 0.03, respectively). IL-5 and RANTES did not correlate, although a trend toward this was observed (r = 0.5, p = 0.1).

View larger version (18K): [in this window] [in a new window]   Figure 3.   (Left panel ) Eosinophil chemotactic activity in BLF from SA, A, V, and C groups. (Right panel ) Inhibition percentages of eosinophil chemotactic activity observed after the addition of anti-RANTES, anti-MCP-3, anti-eotaxin, and anti-IL-5 neutralizing antibodies to BLF from SA patients (n = 7). This percentage is calculated as described in METHODS. Data are expressed as box plots. Horizontal lines represent the median, squares the 25th and 75th percentiles, and error bars the 10th and the 90th percentiles. Circles represent outliers below the 10th and above the 90th percentiles. *p < 0.05 compared with V patients; #p < 0.05 compared with the A and C groups.

Modification of Cytokine Secretion during the Course of Therapy in SA

Levels of chemokines and IL-5 were compared in BLF from BL performed within 48 h (10 BL procedures) and BL performed after 48 h of MV (nine BL procedures) in patients with SA (Figure 4). Concentrations of eotaxin and IL-5 were significantly higher in BLF from BL performed before 48 h (p = 0.006, p = 0.02, respectively) than from BL performed after 48 h of MV. Levels of MIP-1, RANTES, and MCP-3 were not significantly different in BLF from BL performed before and after 48 h of MV. Although its levels tended to be higher in BLF from BL performed after 48 h, no significant difference was oberved for MCP-1.

View larger version (20K): [in this window] [in a new window]   Figure 4.   Concentrations of CC chemokines and IL-5 in BLF of SA patients from BL performed in the first 48 h of MV, and after 48 h. Levels of MCP-1, MCP-3, MIP-1, RANTES, eotaxin, and IL-5 are shown. Data are expressed as box plots. Horizontal lines represent the median, squares the 25th and 75th percentiles, and error bars the 10th and the 90th percentiles. Circles represent outliers below the 10th and above the 90th percentiles. &p < 0.05: statistical significance of the difference between < 48 h and > 48 h.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The main result reported in this study is an increased level of CC chemokines (MCP-1,-3, MIP-1, RANTES, and eotaxin) and IL-5 in the BLF of patients with SA. Among these cytokines, MCP-3, eotaxin, IL-5, and to a lesser extent RANTES were associated with eosinophil chemotactic activity in these patients' BLF. Although SA patients had higher levels of MCP-3 and eotaxin than did the other groups, a statistically significant difference was not found, probably because of the small number of samples.

Previous studies by our group showed that the inflammatory exudate in SA was mainly neutrophilic (8, 21). However, a significant influx of eosinophils was also observed in the airways of SA patients, in association with high concentrations of ECP, particularly in the first 48 h after the beginning of treatment (high doses of glucocorticoids and MV). Eosinophils are of major importance in the bronchial inflammatory process observed in mild and severe forms of asthma. Therefore, the identification of mechanisms and molecules implicated in eosinophil recruitment might help in shedding light on the pathophysiology of this disease and in the design of specific therapies for it. CC chemo-kines and IL-5 are involved in eosinophil migration (9, 11, 12, 20). CC chemokines such as eotaxin, MCP-3, and RANTES bind to a CCR3 receptor, a chemokine receptor highly expressed on eosinophils, and are potent eosinophil chemotactic factors, suggesting a key role for CCR3 in eosinophil-associated inflammation (11). Furthermore, eotaxin is a better eosinophil chemoattractant after IL-5 priming (26). MIP-1 has also been found to be involved in eosinophil migration and activation, but with a lower activity than CCR3 ligands (16). In mild atopic and nonatopic asthma, Humbert and colleagues showed a marked increase in expression of MCP-3, RANTES, and IL-5 messenger RNA (mRNA) (27). In atopic asthma, eotaxin protein levels and mRNA expression were found to be increased and localized in the epithelium and the submucosa of airways; moreover, an eosinophil chemotactic activity dependent on eotaxin was detected in BLF (28). Another finding was that MCP-1, RANTES, and MIP-1 levels were increased in asthmatic patients as compared with controls (29), and an eosinophil chemotactic activity, inhibited mainly by anti-RANTES and anti-MCP-3 neutralizing antibodies, was detected in these patients' BLF. In allergic asthmatic patients, MIP-1, RANTES, and MCP-1 levels in BLF further increased at 4 h after local allergen challenge (30). RANTES and IL-5 expression has been reported in the airways of patients with A and increased after local or natural allergen challenge (31).

The role of CC chemokines has also been evaluated in animal models. In a guinea pig model of allergy, IL-5 and eotaxin appeared to cooperate in mediating eosinophil migration from the bone marrow into the lung in response to an allergen challenge, IL-5 being involved in eosinophil maturation and mobilization into the blood, whereas eotaxin induced the recruitment of eosinophils toward the lung (9). In ovalbumin-sensitized mice, Stafford and coworkers showed an upregulation of MCP-3 mRNA expression after ovalbumin challenge. In this model, anti-MCP-3 antibodies significantly inhibited eosinophil recruitment into the airways (13). Furthermore, Gonzalo and associates showed in a murine model of allergic inflammation that several CC chemokines had a coordinated action in lung inflammation as well as for eosinophil recruitment (32).

In the present study, high levels of CC chemokines such as RANTES, eotaxin, and MCP-3, and of IL-5 were detected in BLF from SA patients. The numbers of eosinophils and/or concentrations of ECP correlated closely with levels of IL-5, and to a lesser extent with those of the other chemokines. The relatively low level of correlation between chemokines and eosinophils and/or ECP may be explained in several ways. Recruitment of eosinophils implicates the expression by endothelial and epithelial cells of adhesion molecules, whose expression we did not evaluate in SA. Variations in CC chemokine receptor expression, such as of CCR1 and CCR3 in eosinophils from SA patients, might account for this lack of relationship. Nevertheless, IL-5 and chemokines are involved in eosinophil chemotactic activity of the BLF of SA patients, as demonstrated by the inhibitory activity of specific neutralizing antibodies to these proteins. Moreover, levels of MCP-3 and eotaxin were correlated with the eosinophil chemotactic activity of this BLF, and a trend toward such corrleation was observed for IL-5 and RANTES. The sum of the percent inhibition obtained with the different antibodies was found to be largely above 100%, indicating probable redundancy in this system. Taken together, these data suggest that eosinophil migration in the airways during mild or severe asthma was dependent on the coordinated action of IL-5 and of CC chemokines active on CCR1 and/or CCR3.

Cell sources of CC chemokines and IL-5 were not identified in this study. Because of the high level of nonspecific binding of antibodies and nuclear probes to cells present in BLF from SA patients, immunocytochemistry and in situ hybridization can hardly be performed in such a setting. However, various cell types could potentially be involved in the production of CC chemokines and IL-5, including epithelial, smooth-muscle, endothelial, and inflammatory cells (11, 12).

In patients with SA, BL was performed at different times after the beginning of treatment. The difference in levels of chemokines and IL-5 in BL performed before and after 48 h following the beginning of treatment might be explained by the use of glucocorticoids or by the rapid resolution of airway inflammation in some patients. Glucocorticoids inhibit cytokine-mediated eosinophil recruitment by different mechanisms, including the induction of eosinophil apoptosis (33). This may explain the decrease in eosinophil number and lower level of IL-5, produced in part by these cells, found in our study. Supporting this hypothesis was the finding by Humbles and coworkers that dexamethasone suppressed eosinophil accumulation measured 24 h after an allergen challenge, whereas this treatment did not affect eotaxin levels (9). In our study, levels of eotaxin were significantly lower in BLF sampled after 48 h of treatment. We cannot exclude a direct inhibitory effect of glucocorticoids on the local production of eotaxin, as was previously demonstrated in vitro (10). In contrast, although glucocorticoids inhibited the release of MIP-1, MCP-3, and RANTES by epithelial bronchial cells in vitro (34), no significant effect was observed in vivo in SA patients for these chemokines. These results may suggest that cells implicated in the production of these chemokines were different and/or less sensitive glucocorticoids in SA patients than were cells producing eotaxin and IL-5. Surprisingly, in contrast to those of the other tested cytokines, and despite the in vitro inhibition by glucocorticoids of its macrophage-derived production, the levels of MCP-1 tended to increase during the course of treatment in SA patients. One hypothesis for this would be that MCP-1 induction is related to mechanical forces, such as those observed during MV (35), counteracting the effects of glucocorticoids.

The role of MCP-1 and MIP-1 in SA requires discussion. Their production during the allergic reaction, and their properties, suggest their involvement in mononuclear cell recruitment and in basophil migration and degranulation (16, 17, 36). The implication of their involvement in SA is confirmed in animal models of allergic reaction, particularly in terms of leukocyte recruitment and development of bronchial hyperreactivity (32). In addition, it has been shown that MCP-1 and MIP-1 modulate the immune response: MCP-1 can shift helper T-cells toward a Th₂-like cell role, whereas MIP-1 upregulates a Th₁-like immune response (37). Moreover, MCP-1 is also involved in fibrotic processes, inducing activation and production of transforming growth factor (TGF)-₁ and collagen by fibroblasts (38). Since high levels of TGF-₁ were detected in BLF from SA patients (21), this mechanism may be involved in lung remodelling during SA.

In conclusion, the results of our study show that CC chemokines and IL-5 are present at high levels in bronchi from SA patients, and could participate in eosinophil recruitment. These results also support the hypothesis that CC chemokines and IL-5 have a coordinate action not only on eosinophil recruitment, but also on other inflammatory processes associated with SA. Inhibition of CC chemokine or IL-5 effects might be beneficial during the acute phase of severe asthma attacks.