Effects of Budesonide and Formoterol on Eosinophil Activation Induced by Human Lung Fibroblasts

Transcription

1 Effects of Budesonide and Formoterol on Eosinophil Activation Induced by Human Lung Fibroblasts FOKJE M. SPOELSTRA, HENK F. KAUFFMAN, HESSEL HOVENGA, JACOBIEN A. NOORDHOEK, JAN G. R. de MONCHY, and DIRKJE S. POSTMA Departments of Allergology and Pulmonology, University Hospital Groningen, Groningen, The Netherlands Budesonide and formoterol are extensively used in current asthma therapy. Budesonide is known as potent antiinflammatory agent and formoterol also appears to have some antiinflammatory properties. We investigated inhibitory effects of these drugs on eosinophil activation in vitro as induced by fibroblast-conditioned medium (FCM). We measured the modulation of expression of clonal designator (CD)11b and L-selectin with flow cytometry after 4 h or 16 h of culture of eosinophils when budesonide or formoterol was applied either directly to the eosinophils while they were stimulated with FCM (direct method) or when each drug was applied to lung fibroblasts from which conditioned medium was then administered to eosinophils (indirect method). In the direct method, budesonide (10 8 M) inhibited the modulation of CD11b (44 [25th to 75th percentiles: 26 to 66]% of control) and L-selectin (30 [ 13 to 48]% of control) only after 16 h, and not after 4 h. Formoterol did not directly inhibit the modulation of eosinophil CD11b and L-selectin expression. In the indirect method, both budesonide and formoterol inhibited lung fibroblast activation, resulting in diminished eosinophil activation after 4 h. Budesonide or formoterol at 10 8 M inhibited upregulation of CD11b to 26 [15 to 40]% and 38 [23 to 46]%, respectively, and inhibited L-selectin shedding to 14 [ 3 to 50]% and 27 [2 to 62]%, respectively, of control values. These results show that budesonide inhibits eosinophil activation primarily through effects on lung fibroblasts, presumably by inhibiting production of granulocyte macrophage colony-stimulating factor. After longer incubation periods, budesonide also directly inhibits eosinophil activation. In contrast, formoterol can inhibit eosinophil activation only via inhibitory effects on lung fibroblasts. We did not observe an additional effect of formoterol, beyond the effects induced by budesonide under any circumstance studied. Lung fibroblasts, in addition to eosinophils, may serve as important target cells for antiinflammatory treatment in asthma. (Received in original form November 17, 1999 and in revised form April 19, 2000) Supported by a grant from AstraZeneca, The Netherlands. Correspondence and requests for reprints should be addressed to Dr. H. F. Kauffman, Ph.D., Laboratory of Allergology, Clinic for Internal Medicine, University Hospital Groningen, Hanzeplein 1, 9713 GZ Groningen, The Netherlands. Am J Respir Crit Care Med Vol 162. pp , 2000 Internet address: Eosinophils are thought to be the most important effector cells in asthma. These cells infiltrate lung tissue, especially during exacerbations caused by allergen inhalation, and thereupon degranulate and produce bronchoconstrictive substances (1, 2). When activated, eosinophils show increased expression of clonal designator (CD)11b and decreased expression of L-selectin in bronchoalveolar lavage fluid (BALF) after allergen challenge in asthmatic individuals (3, 4). Activation of eosinophils can be induced by many different factors of various cellular origins. The lung fibroblast is a resident cell that mediates activation of eosinophils in vitro, as shown by upregulation of CD11b expression and shedding of L-selectin (5). This eosinophil activation is mainly mediated by granulocyte-macrophage colonystimulating factor (GM-CSF) produced by lung fibroblasts (5). Asthma treatment currently consists primarily of antiinflammatory strategies involving corticosteroids, whereas 2 - agonists, through their smooth-muscle relaxing capacity, relieve asthma symptoms. In vivo studies have shown that the corticosteroid budesonide inhibits eosinophil infiltration and activation in the airways of asthmatic patients (6, 7). The longacting 2 -agonist formoterol has a small inhibitory effect on eosinophil infiltration (8) and serum levels of eosinophil cationic protein (ECP) (9), but it is less potent than budesonide. Patients treated with formoterol in addition to 200 g budesonide had significantly fewer exacerbations of asthma (expression of acute inflammation) than did patients receiving 200 g budesonide, but more exacerbations than patients receiving 800 g budesonide (10). It is not clear whether these beneficial clinical effects of budesonide and formoterol are caused mainly by direct effects on eosinophil activation, or whether the actions of other cells necessary for mediating eosinophil activation and migration (e.g., fibroblasts, epithelial cells) are suppressed by these drugs. Glucocorticoids are known to inhibit eosinophil survival, CD11b upregulation, superoxide production, degranulation, and chemokine production in vitro (11 15). Evidence for indirect inhibition of eosinophil function by corticosteroids has been obtained by incubation of epithelial cells and monocytes with these drugs. The conditioned media so produced were less potent in inducing either an increase in eosinophil survival (16) or eosinophil chemotaxis (17), owing to inhibition of cytokine production by epithelial cells and monocytes. Short-acting 2 -agonists such as isoproterenol can inhibit the modulation of CD11b and L-selectin expression on eosinophils (18), as well as the production of eosinophil peroxidase and leukotriene C 4 (LTC 4 ) (19), but failed to show marked antiinflammatory effects in clinical studies. Long-acting 2 -agonists, such as salmeterol, inhibit formyl-met-leu-phe (fmlp)- induced neutrophil adhesion (20). Direct effects of formoterol on eosinophil activation have not yet been investigated. With respect to indirect effects, we found that formoterol, like budesonide, can inhibit GM-CSF production by lung fibroblasts (unpublished results). The aim of the present study was to determine whether eosinophil activation as induced by fibroblast-conditioned medium (FCM) and measured by modulation of CD11b and L-selectin expression can be inhibited by budesonide and formoterol via a direct action on eosinophils or indirectly via the mediatory action of lung fibroblasts. METHODS Study Subjects Eight nonallergic volunteers were included for the isolation of peripheral blood eosinophils. Inclusion criteria were an age of 18 to 40 yr; no history of respiratory symptoms, asthma or allergy; a total plasma IgE 100 IU/ml; and no specific IgE against cat, dog, house dust mite, tree, or grass pollen allergens. All participants gave their informed consent, and the study was approved by the Medical Ethics Committee of the University Hospital Groningen.

2 1230 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL Eosinophil Isolation Eosinophils were isolated from peripheral blood anticoagulated with ethylenediamine tetraacetic acid (EDTA) by negative selection with anti-cd16-magnetic beads, using a modification of the method of Hansel and colleagues (21). Briefly, blood was diluted 1:1 with 0.9% NaCl solution and centrifuged for 20 min at 1,000 g over isotonic Percoll (Pharmacia LKB, Uppsala, Sweden) with a density of g/ ml to separate granulocytes from other leukocytes. Granulocytes were collected from the bottom of the tube and were treated twice with cold ammonium chloride (155 mm NH 4 Cl, 10 mm KHCO 3, 0.1 mm Na 2 EDTA) to lyse contaminating erythrocytes. Cells were resuspended in 2% fetal calf serum (FCS; Bodinko, Alkmaar, The Netherlands) in phosphate-buffered saline (PBS) (2%) at a concentration of 10 8 ml, and 100 l of immunomagnetic beads (Miltenyi Biotec, Bergisch Gladbach, Germany) per milliliter of cell suspension were added. For immunomagnetic separation, a magnetic-activated cell sorter (MACS)-type CS column (Miltenyi Biotec) was used. Eluted fractions contained 98 2% eosinophils with a viability of 98 1%. Eosinophils were suspended in RPMI-1640 (Gibco BRL, Paisley, UK) supplemented with 10% FCS, 2 mm L-glutamine (Biowhittaker, Verviers, Belgium), 125 g/ml streptomycin (Radiumfarma-Fisiopharma, Milan, Italy), and 125 U/ml penicillin (Yamanouchi Pharma, Leiderdorp, The Netherlands), hereafter referred to as RPMI complete medium. Lung Fibroblasts Pulmonary parenchyma was obtained from bilobal lung resection material (derived from healthy lobe) from a nonasthmatic individual after oncologic surgery. Small pieces of tissue were cultured and fibroblasts were allowed to grow out (explant technique). Fibroblasts were cultured in Ham s F12 medium (BioWhittaker) supplemented with 10% FCS (Bodinko BV), 125 U/ml Na-penicillin G (Yamanouchi Pharma), and 125 g/ml streptomycin sulfate (Radiumfarma-Fisiopharma), hereafter referred to as Ham s complete medium. Fibroblasts were passaged by trypsinization with trypsin-edta (BioWhittaker). They were passaged in a 1:4 ratio, grown to confluence (Passage 5) in 6 to 7 d (microscopic examination) in 24-well culture plates (Costar Europe Ltd., Badhoevedorp, The Netherlands), and used for experiments. Fibroblasts were characterized with antibodies against vimentin, cytokeratin, desmin, smooth-muscle actin, and fibronectin, using fluorescence microscopy. Fibroblast purity was more than 98%, the only contaminating cells being smooth-muscle cells. Drugs Budesonide (AstraZeneca, Zoetermeer, The Netherlands) was obtained from a Pulmicort inhaler (Astra) and dissolved in 96% ethanol in a concentration of 10 2 M. Subsequently, solutions of to 10 7 M budesonide were prepared in RPMI complete medium. The vehicle control for 10 8 M budesonide was ethanol in the corresponding dilution. Formoterol fumarate dihydrate (Astra Draco AB, Lund, Sweden) was dissolved in dimethylsulfoxide (DMSO) in a concentration of 10 2 M. Working solutions (10 10 to 10 7 M) were prepared in RPMI The vehicle control for 10 8 M formoterol was DMSO in the corresponding dilution. Fibroblast-Conditioned Media Fibroblasts grown in 24-well plates (Costar, Cambridge, MA), reached confluence and were preincubated with different concentrations of budesonide or formoterol for 30 min in Ham s complete medium or in Ham s complete medium without drugs. They were then stimulated with 10 U/ml interleukin (IL)-1 (Boehringer Mannheim, Mannheim, Germany) for 4 h. The wells were washed in 1 ml RPMI, and fibroblasts were allowed to release their products in 1 ml of RPMI complete medium for 4 h. The conditioned media were harvested, pooled, centrifuged, and aliquoted in 1-ml portions before storage at 80 C. Eosinophil Stimulation To evaluate direct effects of budesonide and formoterol on eosinophil activation, we preincubated isolated eosinophils for 30 min with different concentrations of the two drugs, and subsequently cultured these cells for 4 h or 16 h with FCM in the presence of the same concentrations of the drugs. To evaluate indirect effects of budesonide and formoterol, we suspended eosinophils in conditioned medium derived from fibroblasts that had or had not been incubated with different concentrations of budesonide and/or formoterol and cultured them for 4 h. Flow Cytometric Analysis of Receptor Expression and Survival After culture, eosinophils were harvested and the wells in which they had been cultured were washed once with 1 ml fluorescence-activated cell sorting buffer (PBS, 2% FCS, 0.01% [wt/vol] NaN 3 ). Eosinophils were incubated for 30 min at 4 C with 10 l anti-leu15 (CD11b) or anti-leu8 (L-selectin) antibody (Becton Dickinson, San Jose, CA) or with IgG 2a (CLB, Amsterdam, The Netherlands), as well as 10 l AB serum (20% in fluorescence-activated cell sorting buffer) to avoid nonspecific binding. Samples were washed twice with cold fluorescence-activated cell sorting buffer and incubated for 15 min at 4 C in the dark with 4 l (GAM-FITC) (Becton Dickinson) and 46 l 5% AB serum. Cells were washed once with fluorescence-activated cell sorting buffer and resuspended in 135 l of this buffer. Samples were kept on ice until fluorescence-activated cell sorter (FACS) analysis, and before analysis, 15 l of 10 g/ml propidium iodide (Sigma Chemical Co., St. Louis, MO) was added to allow discrimination between viable and dead cells. Only viable cells were assayed for adhesion-molecule expression. Fluorescence was measured with a FACS ELITE cell sorter (Becton Dickinson). Data analysis was done with WinList software (Verity Software House, Inc., Topsham, ME). Mean fluorescence intensity (MFI) with the control antibody was subtracted from MFI for CD11b and L-selectin expression, yielding specific mean fluorescence values (SMF). The SMF of eosinophils cultured in RPMI complete medium alone was calculated as a basal value (0%), and the SMF of eosinophils cultured in FCM was calculated as maximal activation (100%). GM-CSF Levels GM-CSF levels in fibroblast-conditioned media were measured with an enzyme-linked immunosorbent assay (ELISA) kit for GM-CSF (R&D Systems, Abingdon, UK). Pooled conditioned media were monitored regularly during the study while stored at 80 C, to ascertain stable GM-CSF levels. Statistical Methods Results are presented as median percentages of adhesion molecule expression, median cytokine concentrations, and 25th to 75th percentiles. To evaluate dose-dependent effects, results were tested with Friedman s test. For further evaluation of differences between samples, we used the one-tailed Wilcoxon s signed ranks test for related samples and the Mann Whitney U test for intergroup comparisons. Correlations were tested with Spearman s rank correlation test. Differences were considered significant when p RESULTS Eosinophil Activation by FCM Eosinophils cultured in FCM showed significantly increased expression of CD11b and decreased expression of L-selectin as compared with eosinophils cultured in RPMI complete medium alone after both 4 h and 16 h incubation (Table 1). FCM-cultured eosinophils also showed significantly greater expression of CD11b after 4 h and 16 h than did freshly isolated eosinophils. A decrease in CD11b and L-selectin expression on RPMI complete medium cultured eosinophils as compared with freshly isolated eosinophils was observed after 16 h, whereas comparable expressions were observed after 4 h. Direct Effect of Budesonide and Formoterol on FCM-Induced Eosinophil Activation All tested concentrations of budesonide significantly (p 0.05) inhibited modulation of eosinophil CD11b and L-selectin expression in the presence of FCM, with a small but significant dose-dependent effect (Figure 1A). The vehicle control for

3 Spoelstra, Kauffman, Hovenga, et al.: Effect of Budesonide and Formoterol on Eosinophil Activation 1231 TABLE 1 CD11b AND L-SELECTIN EXPRESSION ON EOSINOPHILS BEFORE AND AFTER CULTURE 4 h 16 h Sample CD11b L-Selelectin CD11b L-Selectin FIE 62 [57 79] 34 [28 38] 62 [57 79] 34 [28 38] RPMI 59 [71 88] 33 [25 43] 54 [51 59] 24 [15 28] FCM 131 [ ]* 19 [9 21]* 94 [86 105]* 14 [10 19]* B [67 154] 17 [12 17] 124 [ ] 20 [11 23] B9 102 [69 145] 17 [11 22] 97 [85 114] 28 [18 32] B8 98 [67 146] 16 [11 18] 85 [77 102] 30 [20 34] B7 77 [72 93] 31 [21 38] F [70 154] 15 [10 16] 128 [ ] 21 [10 23] F9 107 [58 155] 16 [10 22] 105 [96 118] 26 [17 31] F8 104 [73 145] 15 [10 20] 90 [80 109] 29 [18 35 F7 85 [75 106] 30 [19 34] BF [68 138] 16 [11 21] 127 [ ] 20 [10 22] BF9 94 [72 142] 15 [10 22] 89 [78 95] 28 [18 32] Definition of abbreviations: B10 through B7 budesonide at concentrations of M to 10 7 M; BF10 budesonide and formoterol each at a concentration of M; BF9 budesonide and formeterol each at a concentration of 10 9 M; FCM fibroblast conditioned medium (representing eosinophils after 4 h or 16 h culture in FCM [interleukin- 1 -stimulated fibroblasts]); FIE freshly isoated eosinophils; F10 through F7 formoterol at concentrations of M to 10 7 M; RPMI eosinophils after culture (4 h or 16 h) in RPMI complete medium. Results represent median values [25th to 75th percentiles] of specific mean fluorescence (n 6). * p 0.05 compared with RPMI values. p 0.05 compared with FIE values (Wilcoxon s signed ranks test). budesonide did not significantly inhibit CD11b upregulation (104 [81 to 113]%), but did inhibit L-selectin shedding (70 [52 to 83]%; not shown). This inhibitory effect was significantly smaller than the effect of the corresponding concentration of budesonide (43 [26 to 66]% and 30 [ 13 to 48]%, respectively). Incubation of eosinophils with formoterol for 16 h in the presence of FCM did not inhibit modulation of their CD11b or L-selectin expression (Figure 1B). We observed an increase in eosinophil activation by formoterol, which became significant at 10 9 M and 10 8 M with regard to L-selectin shedding. The vehicle control did not influence modulation of CD11b or L-selectin (105 [95 to 118]%, and 96 [83 to 110]%, respectively; not shown). The effects of the combination of budesonide and formoterol at M or 10 9 M did not differ from those of budesonide alone at M or 10 9 M, respectively (Figure 1A). Effects of both budesonide and formoterol on CD11b and L-selectin expression on eosinophils were negative at an incubation time of 4 h (results not shown). Indirect Effect on FCM-Induced Eosinophil Activation via Fibroblasts Preincubated with Budesonide and Formoterol Conditioned media from fibroblasts that had been incubated with different concentrations of budesonide before IL-1 stimulation caused less eosinophil activation than did conditioned medium from fibroblasts stimulated with IL-1 alone (FCM) after 4 h (Figure 2A). There was a significant (p 0.05) dosedependent decrease in modulation of CD11b and L-selectin expression on eosinophils incubated with FCM derived from fibroblasts preincubated with M to 10 7 M budesonide. The vehicle control for budesonide did not inhibit CD11b upregulation (91 [69 to 101]%; not shown). The vehicle control did inhibit L-selectin shedding as compared with FCM (84 [79 to 93]%), but the effect was smaller than the effect induced by the corresponding concentration of budesonide (14 [ 3 to 50]%). Incubation of fibroblasts with formoterol during IL-1 stimulation resulted in conditioned media that were able to Figure 1. Direct effects of M to 10 8 M (A) budesonide (B10, B9, B8) and combined budesonide and formoterol (BF10, BF9), and of (B) formoterol (F10, F9, F8) on fibroblast-conditioned medium (FCM)- induced colonal designator (CD)11b upregulation (solid circles) and L-selectin shedding (open circles) by eosinophils after 16 h. Results are presented as median percentages (and separate experiments) of adhesion molecule modulation (n 6) as compared with stimulation with FCM alone (100%). *p 0.05 compared with FCM alone. significantly inhibit modulation of CD11b and L-selectin eosinophils (Figure 2B) in a dose-dependent manner. This effect on CD11b modulation was lost with M formoterol. The vehicle control for formoterol did not significantly inhibit CD11b expression (86 [73 to 105]%); however, L-selectin shedding was significantly inhibited (74 [49 to 90]%; not shown), although still significantly less so than by the corresponding concentration of formoterol (27 [2 to 62]%). The effect of the combination of M or 10 9 M budesonide and formoterol did not differ from that of budesonide alone at M or 10 9 M, respectively. GM-CSF Production by Fibroblasts Fibroblasts produced significantly smaller quantities of GM- CSF when they were incubated with increasing concentrations of budesonide and/or formoterol before IL-1 stimulation (Table 2). The ethanol and DMSO vehicle controls also had an inhibitory effect on GM-CSF production, but this effect was smaller than the effect of the corresponding concentrations of budesonide and formoterol. The dose-dependent indirect effects of fibroblast-conditioned media on eosinophil activation after incubation of fibroblasts with budesonide or formoterol reflects the dose-dependent in-

4 1232 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL TABLE 2 LEVELS OF GRANULOCYTE MACROPHAGE COLONY-STIMULATING FACTOR IN FIBROBLAST-CONDITIONED MEDIUM FROM FIBROBLASTS INCUBATED WITH BUDESONIDE AND/OR FORMOTEROL, VEHICLE, OR MEDIUM (CONTROL) Incubation with Budesonide GM-CSF (pg/ml) Incubation with Formoterol GM-CSF (pg/ml) FCM 271 FCM 271 B F B9 56 F9 54 B8 31 F8 34 B7 25 F7 32 E8 180 D8 145 BF9 44 BF Definition of abbreviations: B10 through B7 budesonide at concentrations of M to 10 7 M; BF9 budesonide and formoterol each at a concentration of 10 9 M; BF10 budesonide and formoterol each at a concentration of M; DMSO dimethylsulfoxide; E8 ethanol at 10 8 M; FCM fibroblast conditioned medium; FIE freshly isoated eosinophils; GM-CSF granulocyte-macrophage colony-stimulating factor. Results of one representative GM-CSF determination in pooled media from lung fibroblasts incubated with (10 U/ml) IL-1 for 4 h (fibroblast-conditioned media) in the presence of M to 10 7 M budesonide (B10, B9, B8, B7) or M to 10 7 M formoterol (F10, F9, F8, F7) or both (BF10, BF9), or of ethanol (E8) and DMSO (D8) vehicle controls. After 4 h, the cells were washed and left in RPMI-1640 for 4 h before media were harvested. Figure 2. Indirect effects of (A) budesonide (B10, B9, B8, B7) and combined budesonide and formoterol (BF10, BF9), and of (B) formoterol (F10, F9, F8, F7) at M to 10 7 M on fibroblast conditioned media-induced CD11b upregulation (solid circles) and L-selectin shedding (open circles) of eosinophils after 4 h. Results are presented as median percentages (and separate experiments) of adhesion molecule modulation (n 6) as compared with stimulation with control FCM without previous use of drugs (100%). *p 0.05 compared with control FCM. hibition of GM-CSF levels in these media (Table 2). Eosinophil CD11b expression correlated positively (r 0.763) and eosinophil L-selectin expression correlated negatively (r 0.518) with GM-CSF levels in the different fibroblast-conditioned media (p 0.001). DISCUSSION We have previously shown that conditioned medium from IL-1 stimulated lung fibroblasts activates eosinophils (CD11b upregulation and L-selectin shedding) after 4 h of incubation (5). The present in vitro study extends this observation, in that eosinophil activation was indirectly inhibited (as mediated by inhibition of fibroblast function) both by budesonide and formoterol. Direct inhibition of eosinophil activation by budesonide was observed after 16 h but not after 4 h. In contrast, formoterol did not directly inhibit eosinophil activation at either time point. Thus, our results show that budesonide can reduce eosinophil activation both by direct effects on eosinophils and indirectly via fibroblast modulation, whereas formoterol can reduce eosinophil activation only via its effect on fibroblasts. Our results also suggest that the indirect inhibitory effect of budesonide and formoterol on eosinophils is mediated by an effect on GM-CSF production by lung fibroblasts. That budesonide can inhibit eosinophil activation directly and via fibroblast modulation is in accordance with the observation that glucocorticoids inhibit eosinophil survival induced by epithelial-cell conditioned medium both directly (22) and indirectly (16). Comparing the extent of direct and fibroblastmediated indirect effects on eosinophils, it is clear that budesonide produces its largest inhibitory effect on fibroblasts and not on eosinophils after 4 h of stimulation, the only exception to this being that M budesonide produced more inhibition directly (16 h) than indirectly (4 h). This means that indirect inhibition via fibroblasts is faster. Indirect inhibition of eosinophil activation by budesonide showed a clear dosedependent relationship, whereas dose-dependency at 16 h after direct application of budesonide to eosinophils was less prominent. The observed differences in direct and indirect inhibition by budesonide and formoterol on eosinophil CD11b and L-selectin expression are probably due to different effects on regulatory mechanisms in eosinophils compared with fibroblasts, in which they act on GM-CSF production. Some studies have contributed to insight into the mechanisms regulating CD11b expression. It is known that GM-CSF stimulates CD11b upregulation on eosinophils (12), and the study in which this was seen also found indications that CD11b upregulation was dependent on RNA and protein synthesis after 24 h. The same study showed inhibition of IL-3 induced and GM- CSF induced CD11b upregulation on eosinophils by dexamethasone. Surface expression of CD11b can be inhibited by antisense oligodeoxynucleotide to p65, a subunit of nuclear factor (NF)- B, but CD11b mrna expression cannot be so inhibited. This suggests modulation of CD11b cell-surface expression by another factor that is regulated by NF- B (25). Therefore, we think that direct inhibition of CD11b upregulation on eosinophils by budesonide seen in the present study may be modulated by NF- B. Formoterol, however, is most likely unable to intervene through this mechanism, since no direct inhibition by formoterol of CD11b upregulation on eosinophils was found in the present study. GM-CSF is also known to induce L-selectin shedding by

5 Spoelstra, Kauffman, Hovenga, et al.: Effect of Budesonide and Formoterol on Eosinophil Activation 1233 granulocytes (26). Enzyme activity (surface metalloproteinase) and probably a conformational change in the L-selectin molecule are necessary for extracellular cleavage of L-selectin (27). The extracellular metalloproteinase has not yet been characterized, and no further information now exists about regulation of its production and activation. This report is the first to point out that formoterol can inhibit eosinophil activation via modulation of fibroblast activity. Formoterol is clearly unable to directly inhibit eosinophil activation, both after 4 h and after 16 h. In contrast, formoterol at concentrations of 10 9 M and 10 8 M resulted in increased L-selectin shedding, suggesting eosinophil activation. The latter observation is apparently in contrast to the reported inhibitory direct effects of formoterol on eosinophil activation as reflected by increased chemotaxis and degranulation (28), and also in contrast to the inhibition by isoproterenol of eosinophil adhesion molecule expression (18). However, an increased respiratory burst of eosinophils has been described after prolonged (4 to 24 h) incubation with short-acting 2 -agonists (13). These discrepancies may be due to different mechanisms of action of short- and long-acting 2 -agonists on cells, to the different stimuli used in the different studies, to the types of responses measured, or to the duration of incubation. The significant inhibitory effects of the ethanol and DMSO vehicle controls on L-selectin shedding by eosinophils via FCM, can be explained by the decrease in GM-CSF levels observed in the vehicle control samples. That CD11b upregulation was not significantly inhibited by the vehicle controls might have been due to the higher sensitivity of L-selectin shedding to GM-CSF. However, the direct effect on L-selectin shedding after 16 h of exposure to ethanol, although also small, was significantly inhibitory, suggesting that L-selectin shedding is more sensitive to stimuli than is CD11b upregulation. The inhibitions effected by the vehicle controls were small, and always significantly smaller than inhibition by the corresponding concentration of budesonide or formoterol. Therefore, inhibition by the drugs was evident, but might have been slightly overestimated in our results. There are strong indications that inhibition of GM-CSF production by fibroblasts is responsible for the decreased effects of FCM modulated by budesonide and formoterol on eosinophil activation. First, in a previous study, we showed that GM-CSF is the most important FCM component in upregulating CD11b and L-selectin shedding on eosinophils in vitro (5). Second, in the present study budesonide and formoterol reveal a dose-dependent inhibition of GM-CSF in the different fibroblast-conditioned media, as shown in Table 2, which mirrors the dose-dependent inhibition of CD11b and L-selectin expression. Furthermore, the detected levels of GM-CSF in the fibroblast-conditioned media correlate significantly with the corresponding degrees of eosinophil CD11b and L-selectin expression. Moreover, there is in vivo evidence of reduced GM-CSF levels in sputum of patients using glucocorticoid therapy (29), and GM-CSF levels in BALF correspond with the activation state of BALF eosinophils (2). The results of the present study imply that for the daily practice of asthma therapy, the use of budesonide has beneficial effects, since it decreases eosinophil activation both directly and indirectly, via lung fibroblasts. This may lead to an additive effect in the in vivo situation. Formoterol may also contribute to inhibition of eosinophil activation via lung fibroblasts, but has no direct effect on eosinophil activation. These findings agree with the in vivo results achieved by Wallin and coworkers, who found a small antiinflammatory effect of formoterol on eosinophil infiltration in airway walls of asthmatic individuals, which was smaller than the effect of budesonide in the same setting (8). The finding also agree with the observation that treatment of asthma patients with 800 g budesonide resulted in a larger decrease in the severe exacerbation rate than did addition of formoterol (10). Because we found no significantly different effects of the combination of budesonide and formoterol (10 9 M, M) than of budesonide alone on eosinophil activation, we conclude that formoterol can be given in combination with budesonide without counteracting the inhibitory actions of budesonide on eosinophil activation. In summary, budesonide inhibits the modulation of CD11b and L-selectin expression on eosinophils via its effects on fibroblasts. At longer incubation times, budesonide can also inhibit eosinophil activation directly. Formoterol exerts its inhibitory effect on eosinophil activation via inhibition of fibroblast-produced soluble factors only. We conclude that resident cells in the lungs, such as fibroblasts, may be important targets for the antiinflammatory actions of budesonide and formoterol, eventually leading to less infiltration by and activation of inflammatory cells such as eosinophils. Acknowledgment : The authors thank Dr. M. Boorsma for critically reading this manuscript. References 1. Hamid, Q., Y. L. Song, T. C. Kotsimbos, E. M. Minshall, T. R. Bai, and J. C. Hogg Inflammation of small airways in asthma. J. Allergy Clin. Immunol. 100: Woolley, K. L., E. Adelroth, M. J. Woolley, R. Ellis, M. Jordana, and P. M. O Byrne Granulocyte-macrophage colony-stimulating factor, eosinophils and eosinophil cationic protein in subjects with and without mild, stable, atopic asthma. Eur. Respir. J. 7: Mengelers, H. J., T. Maikoe, L. Brinkman, B. Hooibrink, J. W. Lammers, and L. Koenderman Immunophenotyping of eosinophils recovered from blood and BAL of allergic asthmatics. Am. J. Respir. Crit. Care Med. 149: Kroegel, C., M. C. Liu, W. C. Hubbard, L. M. Lichtenstein, and B. S. Bochner Blood and bronchoalveolar eosinophils in allergic subjects after segmental antigen challenge: surface phenotype, density heterogeneity, and prostanoid production. J. Allergy Clin. Immunol. 93: Spoelstra, F. M., H. Hovenga, J. A. Noordhoek, D. S. Postma, and H. F. Kauffman Changes in CD11b and L-selectin expression on eosinophils are mediated by human lung fibroblasts in vitro. Am. J. Respir. Crit. Care Med. 158: Lim, S., A. Jatakanon, M. John, T. Gilbey, B. J. O Connor, K. F. Chung, and P. J. Barnes Effect of inhaled budesonide on lung function and airway inflammation: assessment by various inflammatory markers in mild asthma. Am. J. Respir. Crit. Care Med. 159: Gauvreau, G. M., J. Doctor, R. M. Watson, M. Jordana, and P. M. O Byrne Effects of inhaled budesonide on allergen-induced airway responses and airway inflammation. Am. J. Respir. Crit. Care Med. 154: Wallin, A., T. Sandstrom, M. Soderberg, P. Howarth, B. Lundback, G. Della-Cioppa, S. Wilson, M. Judd, R. Dukanovic, S. Holgate, et al The effects of regular inhaled formoterol, budesonide, and placebo on mucosal inflammation and clinical indices in mild asthma. Am. J. Respir. Crit. Care Med. 159: Wong, B. J., J. Dolovich, E. H. Ramsdale, P. O Byrne, L. Gontovnick, J. A. Denburg, and F. E. Hargreave Formoterol compared with beclomethasone and placebo on allergen-induced asthmatic responses. Am. Rev. Respir. Dis. 146: Pauwels, R. A., C. G. Löfdahl, D. S. Postma, A. E. Tattersfield, P. O Byrne, P. J. Barnes, and A. Ullman Effect of inhaled formoterol and budesonide on exacerbations of asthma. N. Engl. J. Med. 337: Walsh, G. M., and A. J. Wardlaw Dexamethasone inhibits prolonged survival and autocrine granulocyte macrophage colony-stimulating factor production by human eosinophils cultured on laminin or tissue fibronectin. J. Allergy Clin. Immunol. 100: Hartnell, A., A. B. Kay, and A. J. Wardlaw Interleukin-3-induced up-regulation of CR3 expression on human eosinophils is inhibited by dexamethasone. Immunology 77: Nielson, C. P., and N. E. Hadjokas Beta-adrenoceptor agonists block corticosteroid inhibition in eosinophils. Am. J. Respir. Crit. Care Med. 157:

6 1234 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL Miyamasu, M., Y. Misaki, S. Izumi, T. Takaishi, Y. Morita, H. Nakamura, K. Matsushima, T. Kasahara, and K. Hirai Glucocorticoids inhibit chemokine generation by human eosinophils. J. Allergy Clin. Immunol. 101: Fujisawa, T., A. Terada, J. Atsuta, K. Iguchi, H. Kamiya, and M. Sakurai IL-5 as a strong secretagogue for human eosinophils. Int. Arch. Allergy Immunol. 114(Suppl. 1): Roca-Ferrer, J., J. Mullol, E. Lopez, A. Xaubet, L. Pujols, J. C. Fernandez, and C. Picado Effect of topical anti-inflammatory drugs on epithelial cell-induced eosinophil survival and GM-CSF secretion. Eur. Respir. J. 10: Lantero, S., O. Sacco, C. Scala, M. C. Morelli, and G. A. Rossi Eosinophil locomotion and the release of IL-3 and IL-5 by allergenstimulated mononuclear cells are effectively downregulated in vitro by budesonide. Clin. Exp. Allergy 26: Spoelstra, F. M., C. Berends, B. Dijkhuizen, J. G. de Monchy, and H. F. Kauffman Effect of theophylline on CD11b and L-selectin expression and density of eosinophils and neutrophils in vitro. Eur. Respir. J. 12: Munoz, N. M., A. J. Vita, S. P. Neeley, K. McAllister, S. M. Spaethe, S. R. White, and A. R. Leff Beta adrenergic modulation of formylmethionine-leucine-phenylalanine-stimulated secretion of eosinophil peroxidase and leukotriene C 4. J. Pharmacol. Exp. Ther. 268: Bloemen, P. G. M., M. C. van den Tweel, P. A. J. Henricks, F. Engels, M. H. A. Kester, P. G. F. Van de Loo, F. J. Blomjous, and F. P. Nijkamp Increased camp levels in stimulated neutrophils inhibit their adhesion to human bronchial epithelial cells. Am. J. Physiol. (Lung Cell. Mol. Physiol.) 272:L580 L Hansel, T. T., I. J. De Vries, T. Iff, S. Rihs, and M. Wandzilak An improved immunomagnetic procedure for the isolation of highly purified human blood eosinophils. J. Immunol. Methods 145: Mullol, J., E. López, J. Roca Ferrer, A. Xaubet, L. Pujols, J. C. Fernandez-Morata, J. M. Fabra, and C. Picado Effects of topical antiinflammatory drugs on eosinophil survival primed by epithelial cells: additive effect of glucocorticoids and nedocromil sodium. Clin. Exp. Allergy 27: Tobler, A., R. Meier, M. Seitz, B. Dewald, M. Baggiolini, and M. F. Fey Glucocorticoids downregulate gene expression of GM-CSF, NAP- 1/IL-8, and IL-6, but not of M-CSF in human fibroblasts. Blood 79: Masquilier, D., and C. P. Sassone Transcriptional cross-talk: nuclear factors CREM and CREB bind to AP-1 sites and inhibit activation by Jun. J. Biol. Chem. 267: Sokoloski, J. A., A. C. Sartorelli, C. A. Rosen, and R. Narayanan Antisense oligonucleotides to the p65 subunit of NF-kappa B block CD11b expression and alter adhesion properties of differentiated HL- 60 granulocytes. Blood 82: Griffin, J. D., O. Spertini, T. J. Ernst, M. P. Belvin, H. B. Levine, Y. Kanakura, and T. F. Tedder Granulocyte-macrophage colonystimulating factor and other cytokines regulate surface expression of the leukocyte adhesion molecule-1 on human neutrophils, monocytes, and their precursors. J. Immunol. 145: Bennett, T. A., E. B. Lynam, L. A. Sklar, and S. Rogelj Hydroxamate-based metalloprotease inhibitor blocks shedding of L-selectin adhesion molecule from leukocytes functional consequences for neutrophil aggregation. J. Immunol. 156: Eda, R., H. Sugiyama, R. J. Hopp, C. Okada, A. K. Bewtra, and R. G. Townley Inhibitory effects of formoterol on platelet-activating factor induced eosinophil chemotaxis and degranulation. Int. Arch. Allergy Immunol. 102: Inoue, H., H. Aizawa, S. Fukuyama, S. Takata, K. Matsumoto, M. Shigyo, H. Koto, and N. Hara Effect of inhaled glucocorticoid on the cellular profile and cytokine levels in induced sputum from asthmatic patients. Lung 177:53 62.