Promotion of Eosinophil Survival by Human Bronchial Epithelial Cells and its Modulation by Steroids Gerard Cox, Takayuki Ohtoshi, Carlo Vancheri, Judah A. Denburg, Jerry Dolovich, Jack Gauldie, and Manel Jordana Department of Pathology, Molecular Virology and Immunology Program, McMaster University, Hamilton, Ontario, Canada
Accumulation of eosinophils in the bronchial tissue occurs in a variety of inflammatory disorders of the human airway. We asked whether airway epithelial cells released factors that could influence eosinophil survival and thus contribute to accumulation of these cells in the tissues. Using conditioned medium (CM) generated from cultured human bronchial epithelial cells (HBEC), we examined the in vitro survival of eosinophils isolated from human peripheral blood. When cultured in control medium, more than 90% of the eosinophils were dead by day 4. In contrast, culture in HBEC-CM resulted in dose-dependent survival at day 6 of 69 ± 9.4%,40.5 ± 5.9%, and 25 ± 2% viability with 2,0.5, and 0.1% HBEC-CM, respectively (n = 4). Granulocyte/macrophage colony-stimulating factor (GM-CSF) was detected in the HBEC-CM by enzyme-linked immunosorbent assay at levels of 22 to 48 pg/ml. Furthermore, preincubation of the HBECCM with a neutralizing monoclonal antibody to human GM-CSF completely inhibited this increased survival of eosinophils. Because corticosteroids are potent eosinopenic agents, we also examined the effects of the synthetic steroid budesonide on this system. Budesonide inhibited both spontaneous and interleukin-l (IL-l)-induced GM-CSF production by cultured HBEC. In addition, preincubation of eosinophils with budesonide caused marked abrogation of the survival induced subsequently with either HBEC-CM or recombinant human GM-CSF. In summary, HBEC can support eosinophil survival via the elaboration of GM-CSF and thus may contribute to the local control of inflammatory cell accumulation. Steroids may modulate this process both by inhibiting cytokine production from HBEC and by a direct effect on eosinophils, preventing their response to cytokines.
Eosinophils accumulate in the respiratory mucosa in a number of inflammatory diseases such as asthma, allergic rhinitis, and nasal polyposis and have been implicated in the pathogenesis of these conditions (1, 2). Typically, in the bloodstream, eosinophils are thought to be short-lived cells (3), and the mechanisms underlying their accumulation at sites of inflammation are incompletely understood. As eosinophils are frequently within or close to the epithelial layer in inflamed tissue (2, 4), we asked whether human bronchial epithelial cells (HBEC) release products that support the (Received in original form August 1, 1990 and in revised form November 27, 1990) Address correspondence to: Dr. M. Jordana, Department of Pathology, McMaster University, 1200 Main Street West, Hamilton, Ontario, L8N 3Z5 Canada. Abbreviations: conditioned medium, CM; enzyme-linked immunosorbent assay, ELISA; granulocyte/macrophage colony-stimulating factor, GMCSF; human bronchial epithelial cell(s), HBEC; Hanks' balanced salt solution, HBSS; serum-free hormonally defined Ham's F-12medium, HD-Fl2; interleukin, IL; phosphate-buffered saline, PBS; recombinant human, rho Am. J. Respir. Cell Mol. BioI. Vol. 4. pp. 525-531, 1991
survival of eosinophils, thereby contributing to their local accumulation. Our results demonstrate that conditioned medium (CM) from HBEC markedly prolongs the in vitro survival of eosinophils, and that this effect is mediated by granulocyte/macrophage colony-stimulating factor (GM-CSF) in the CM. In addition, we demonstrate that the synthetic steroid budesonide inhibits both unstimulated and interleukin (IL)-l-stimulated GM-CSF production by HBEC, and moreover that budesonide has a direct effect on eosinophils as it prevents the ability of these cells to respond to cytokines such as GM-CSF.
Materials and Methods Reagents Budesonide was kindly provided by Drs. B. Anderson and M. Vangielzehm of AB Draco (Lund, Sweden). A stock solution of 10-4 M was made in 30% ethanol/70% phosphatebuffered saline (PBS), and further dilutions were made using the same diluent, which was also used at equivalent concentrations in all control wells. In addition, a dexamethasone
526
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 4 1991
(Organon, Toronto, Canada) stock solution of 10-4 M diluted in PBS was prepared. Cytokines Recombinant human (rh) GM-CSF, (sp act, 9.3 x lQ6 U/mg) and IL-3 (sp act, 4 x 106 U/mg) were obtained from Genetics Institute (Boston, MA). rhIL-5 (sp act, 109 U/mg) and a rat IgG 2a neutralizing monoclonal antibody to human GM-CSF was kindly provided by Dr. 1. Abrams (DNAX Research Institute, Palo Alto, CA); l ug/ml of this antibody neutralizes 2 ng/ml of human GM-CSF in a K61 bioassay. GM-CSF was assayed by an enzyme-linked immunosorbent assay (ELISA) (Genzyme Corp., Boston, MA). The linear portion of the standard curve corresponds to GMCSF levels in the 10 to 250 pglml range.
added to a concentration of 10-6 M. Untreated cultures were exposed to an equal volume of diluent. The supernatants were handled as above until the assay for GM-CSF was performed. In experiments involving cytokine neutralization by antibody, the HBEC-CM was preincubated with the antibody at 37° C for 1 h prior to its addition to the eosinophil suspension. Eosinophil Isolation After informed consent, eosinophils were isolated by a previously described method (7), from the peripheral blood of atopic patient volunteers with eosinophil differentials of 4 to 10%. Ten times-concentrated PBS was added to Percoll solution (Pharmacia, Uppsala, Sweden) to achieve an osmolarity of 310 to 320 mosm/kg. This was further diluted with regular PBS to 75 and 65 % solutions. Discontinuous gradients were made using 15 ml of each solution layered in a 50-ml conical tube (Coming Glassworks, Coming, NY). Heparinized venous blood was incubated with 10-6 M N-formylmethionylleucylphenylalanine (FMLP) at 37° C for 15 min. Fifteen milliliters of blood were then layered on top of the Percoll gradient and centrifuged at 400 X g (1,500 rpm) at 22° C for 25 min. The plasma and the neutrophil/mononuclear cell band were discarded, and 10 ml was collected from the 65 to 75% interface (density, > 1.085 g/ml). This was mixed with 10 ml HBSS without calcium and centrifuged at 530 X g (2,000 rpm) for 10 min. The cells were washed once more in HBSS, and erythrocytes were removed by cold hypotonic lysis. The cells were then resuspended in supplemented RPM!. Purity as assessed on cytospins stained with Diff-Quikf (American Hospital Supply Corp., McGaw Park, IL), a modification of WrightGiemsa staining, was greater than 90 % and viability was greater than 99% on trypan blue dye exclusion in all experiments.
HBEC Cultures Specimens of human bronchi were obtained from organs removed at surgery. All patients had a primary pulmonary neoplasm requiring lobectomy or pneumonectomy. A portion of the bronchus at a site distant from the lesion requiring resection was taken and transported to the laboratory in minimal essential medium. HBEC were dispersed as previously described (5). Briefly, tissue was rinsed 3 times in Hanks' balanced salt solution (HBSS) and incubated overnight at 4° C in 0.1% protease (Sigma Chemical Co., St. Louis, MO) solution in Ham's F-12 medium. Heat-inactivated fetal bovine serum (GIBCO, Grand Island, NY) was added to 10% vol/vol, and cells were detached by gentle agitation. The suspension was then filtered through a 60-l-'m Nitex mesh (Cellector; Sargent-Welch, Toronto, Canada) and centrifuged at 270 X g (1,000 rpm) for 10 min. The cell pellet was resuspended in serum-free hormonally defined Ham's F-12 medium (HD-Fl2) containing 1% penicillin/streptomycin, 5 I-'glml insulin (GIBCO), 5 I-'glml transferrin (GIBCO), 25 I-'g/ml epidermal growth factor (Collaborative Research, Lexington, MA), 15 I-'g/ml endothelial cell growth supplement (Collaborative Research), 2 x 10-10 M triiodothyronin (GIBCO), and 10-7 M hydrocortisone (GIBCO). Fiftythousand cells in 2 ml HD-FI2 were plated onto collagencoated wells, and cells were cultured in a humidified atmosphere at 37° C with 5% CO 2 , The medium was changed at day 1 and on alternate days thereafter. The monolayer achieved confluency at days 7 to 12. Using this method in upper airway tissues, we have previously shown that by immunohistochemistry ~ 95% of the cells in the culture stain positive for keratin, an extremely small number of cells stain positive for vimentin, and in addition, HBEC cultures established in this manner do not stain with antibody to the monocyte CD-14 antigen (6).
Eosinophil Culture Eosinophils were incubated in 24-well tissue culture plates (Nunclon; GIBCO, Burlington, Canada) at a concentration of 250,000 cells in 1 ml. Survival was assessed at days 2,4, and 6 and was calculated using the formula: 100 X [(number of eosinophils recovered) X (percentage of cells viable by trypan blue exclusion)]/number of eosinophils originally delivered (8). In all experiments, duplicate wells of each condition were assessed. In experiments involving steroid treatment, either budesonide or dexamethasone, at 10-6 M, was added to the eosinophil suspensions 1 h prior to the addition of HBEC-CM or recombinant human cytokines. Wells with untreated cells received the same volume of diluent.
Generation of HBEC-CM When a confluent monolayer was achieved, the medium was changed to RPMI supplemented with 10% FBS, 1% penicillin/streptomycin, and 1% Hepes buffer. At confluence, there were 270,000 to 340,000 cells/plate. HBEC-CM was generated with supplemented RPMI from these monolayers over 48 h and then pooled, filtered (0.45 I-'m; Millipore Corp., Bedford, MA), and stored at -20° C until use. In the treatment of HBEC cultures, budesonide or dexamethasone was
Eosinophil Density After culture, eosinophil density was assessed after centrifugation on continuous Percoll gradients as previously described (9). Continuous gradients were generated by placing 10 ml of Percoli solution (1.084g/ml) in 15-ml polypropylene tubes (Falcon, Oxnard, CA) and centrifuging at 20,000 X g for 15 min at 22° C in a fixed-angle rotor. One milliliter of eosinophil suspension was subsequently placed on top and centrifuged at 100 X g for 20 min at 22° C. Control gra-
Cox, Ohtoshi, Vancheri et al.: Steroid Modulation of HBEC-supported Eosinophil Survival
100
527
100
~ HBEC-CM •
~
,.,
75
75 ..J
e.'"
...J
!i!
~
sa:
50
sII:
::l
50
. (/)
~
~
Ul
~
HBEC-CM + Antl-GM-CSF
25
o
25
0.1
0.5
2
0+---
10
EPITHELIAL CELL-CONDITIONED MEDIUM (0/0)
Figure1. Dose response of eosinophil survival in human bronchial epithelial cell-conditioned medium (HBEC-CM). HBEC-CM was generated over 48 h from various primary cultures. Freshly isolated eosinophils were cultured in RPMI enriched with increasing concentrations of HBEC-CM, and survival (corrected for cell number, see MATERIALS AND METHODS) was assessed at day 6 by trypan blue dye exclusion. Resultsshown are the means (± SD) offour independent experiments.
dients were spun with I rnl RPM!. Density of the fractions of the control gradients was measured by density meter (DMA40; Anton Paar, Graz, Austria). Calibration of this instrument was done on each day of experiments. Statistics The statistical significance of differences in result of the experiments was evaluated using a Student's t test for unpaired data except in Table 1, where a t test for paired data was employed.
DAY 4
DAY 6
Figure 2. Effect of anti-granulocyte/macrophage colony-stimulating factor (anti-GM-CSF) on eosinophil survival. The HBEC-CM was incubated with anti-GM-CSF at 1:100 dilution for 1 hat 37° C prior to adding to the eosinophil culture. Results shown are the means (± SD) of four independent experiments. The differences between the anti-GM-CSF-treated and the untreated HBEC-CM survivals at both time points are statistically significant (P < 0.01).
GM-CSF Content of HBEC-eM and Effect of Steroid Treatment The individual GM-CSF content of five different primary HBEC cultures was 22, 27, 31, 37, and 48 pglml (Table lA). Addition of budesonide at 10-6 M to these HBEC cultures consistently decreased GM-CSF production from 33 ± 10 to 17.2 ± 11.9 pg/rnl (P< 0.05). Treatment of the HBEC cultures with IL-l (125 Vlml) stimulated GM-CSF production from mean 27 to 38.9 pglrnl (P < 0.05) (Table lB). Budesonide treatment of the HBEC cultures inhibited ILl-induced GM-CSF production (n = 3). Similar to the effect
Results Effect of HBEC-CM on Eosinophil Survival and Density The survival of human peripheral blood eosinophils in supplemented RPMI, in vitro, was always less than 10% at day 4. In contrast, survival at day 4 was promoted when eosinophils were incubated with 2 % HBEC-CM: 72.3 ± 5.5% versus 4.3 ± 2.3 % (n = 4; P < 0.01). This effect to promote survival is dose dependent: survival at day 6 was 69 ± 9.4 %, 40.5 ± 5.9%, and 25 ± 2% in the presence of 2, 0.5, and 0.1% HBEC-CM, respectively (n = 4) (Figure 1). The differences in survival in 0.5 % and 0.1% are statistically significantly different from survival with 2 and 10% HBEC-CM. Preincubation of the HBEC-CM with a neutralizing monoclonal antibody to human GM-CSF resulted in complete inhibition of eosinophil survival. Figure 2 shows that eosinophil survival decreased from 72 to 4 % at day 4 and from 69 % to 0 at day 6 (n = 4), with anti-GM-CSF. In separate experiments, HBEC-CM was preincubated with antibody to IL-3 and no difference in eosinophil survival at day 6 was detected compared with untreated HBEC-CM: 46.6 ± 5.2 % versus 45.5 ± 5.1%, respectively (n = 3). At the time of isolation, all the eosinophils were harvested from the fraction of the Percoll gradient with density> 1.085 glrnl. In contrast, 60 ± 6% of the viable eosinophils had density less than 1.080 g/liter (n = 3) after culture with HBEC-CM for 6 d.
TABLE 1
Granulocyte/macrophage content in human bronchial epithelial cell-conditioned medium* A. No.1 No. 2 No. 3 No.4 No.5 Mean B.
No.5· No. 6 No. 7 No. 8 Mean
+ Budesonide
RPMI
22 27 31 37 48 33
+ Dexamethasone
11
8.5 8.5 22 36 17.2
RPMI
+ IL-I
48 19 33 8 27
63 27.5 53 12 38.9
8.4 24
Budesonide + IL-I 49 10
7.5
Dexamethasone + IL-I 48 20 8.5
* Values are given as pg/m!. A. Confluent human bronchial epithelial cell monolayers were maintained in RPMI supplemented with 10% fetal calfserum andtreated with budesonide or dexamethasone at 10-6 M or an equal volume of diluent. The differences in granulocyte/macrophage colony-stimulating factor content between control and budesonide-treated cultures are statistically significant (P < 0.05). B. IL-I (125 V/m!) was added to cultures I h aftereither steroid (10" M) ordiluent treatment. The differences between control andIL-l-treated cultures are statistically significant (P < 0.05).
528
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 4 1991
TABLE 2
TABLE 3
Eosinophil survival with diluent or budesonide*
Effect of steroid treatment of the eosinophils on cytokine-induced survival *
Eosinophil Survival (%)
RPM! Diluent Budesonide
Day 2
Day 4
84 ± 6.7 88 ± 6 79 ± 5.6
6 ± 3.6 4 ± 3.1 5 ± 2.7
* Freshly isolated eosinophils were cultured in RPMI alone or RPMI containing either budesonide at 10-0 M or an equal volume of the ethanol/phosphate-buffered saline diluent solution. Results are the mean ± SO of three independent experiments. The differences are not statistically significant. of budesonide, dexamethasone (10-6 M) inhibited both basal and IL-l-induced output (n = 2) (Table lA).
Eosinophil Survival (%)
Control Budesonide Dexamethasone
100
iii'
75
co e.'" ..J
~ a:
50
:;: ::::l
C/)
al
25
o UNTREATED
DILUENT
BUDESONIDE
Figure 3. Effect of pretreatment of eosinophils on their survival in HBEC-CM. Freshly isolated eosinophils were treated with budesonide at 10- 6 M for 1 h prior to adding the HBEC-CM. The ethanol/phosphate-buffered saline diluent was added in volume equal to that delivered to the budesonide-treated samples. Survival was assessed at day 6. Results are expressed as means (± SD) of three independent experiments. The survival of the budesonidetreated samples is statistically significantly less than in the untreated and diluent-treated samples (P < 0.01).
IL-3
IL-5
69.5 ± 6.5 22.2 ± 6.8t 49.9 ± 4.7*
64.5 ± 9.5 16.8 ± 9t 51.6 ± 0.7
55.9 ± 6 23.6 ± 4.8t 46 ± 3.45
Definition of abbreviations: GM-CSF = granulocyte/macrophage colonystimulating factor; IL-3 and IL-5 = interleukin-3 and -5, respectively. * Freshly isolated eosinophils were pretreated with budesonide or dexamethasone at 10-0 M I h prior to the addition of cytokine. Control samples received an equal volume of the ethanol/phosphate-buffered saline diluent. GM-CSF was added at 10 U/rnl at day O. IL-3 was added at 10 U/rnl on days 0 and 3. IL-5 was added at 100 U/rnl on days 0 and 3. Survival was assessed at day 6. Results are the mean ± SO of three independent experiments. t
Effect of Steroid Treatment on Eosinophil Survival Eosinophil survival with CM generated in the presence of corticosteroid was assessed. In these experiments, different wells of the primary HBEC culture were treated with either budesonide or dexamethasone. While untreated 2 % HBECCM supported eosinophil survival of 60 ± 2.9 % at day 6, HBEC-CM generated in the presence ofbudesonide or dexamethasone at 10-6 M induced survival of 4.3 ± 3.2 % and 8.1 8 ± 2.7%, respectively (n = 3; P < 0.05). The addition ofbudesonide, or the diluent solution to eosinophil cultures, did not affect eosinophil survival in the absence of HBEC-CM or cytokines (Table 2). However, eosinophils that had been exposed to steroid prior to the addition of HBEC-CM showed a marked reduction in survival, from 69% to 10% (Figure 3). As shown in Table 3, the addition of recombinant human (rh) GM-CSF resulted in eosinophil survival at day 6 of 69 ± 6.5% (n = 5), and this increased survival was similarly inhibited by steroid pretreatment of the eosinophils. The effect of budesonide was dose dependent; eosinophil survival with 10 U/ml rhGMCSF was 22.2 ± 2.3%, 54.5 ± 3.8%, and 59.4 ± 6.1% with
GM-CSF
z c o.m.
*p < 0.05.
budesonide at 10-6 , 10-8 , and 10-10 M, respectively (n ~ 3; P < 0.05 for 10-6 and 10-8 M budesonide). Because IL-3 and IL-5 have also been shown to promote eosinophil survival, we examined whether steroids would also prevent their survival-enhancing effects. Eosinophils incubated with rhIL-3 (10 U/m!) and rhIL-5 (100 U/m!) had prolonged survival of 64.5 ± 9.5% and 55.9 ± 6% at day 6. Eosinophil survival was reduced to 16.8 ± 9% and 23.6 ± 4.8% at day 6 (n = 3) (Thble 3), when the eosinophils were treated with budesonide prior to the addition of these cytokines. With all cytokines, treatment with dexamethasone also caused reduced survival, but the effect was less marked than that seen with budesonide (Table 3).
Discussion Allergic rhinitis, nasal polyposis, and asthma are diseases of the respiratory tract characterized by chronic inflammation in which the accumulation of eosinophils in the tissue is a prevalent feature (1). These cells have also been found in the bronchoalveolar lavage of patients with idiopathic pulmonary fibrosis, and their presence has been associated with a poorer prognosis (10). It has been suggested that the contribution of these cells to the disease process resides in their ability to release products that can cause tissue damage (11-14). The generation of tissue eosinophilia is a complex process, only partly understood, which involves a series of steps from bone-marrow production and exit to the bloodstream, adherence to the endothelium, migration into and through tissues, and finally activation with release of granule contents. Because eosinophils in the bloodstream are shortlived cells (3), the extent to which they survive in the tissue likely depends on locally generated signals and represents an important factor in the accumulation of these cells at a particular site. It has become increasingly clear that within inflamed tissues a number of resident cells have the ability to release a variety of cytokines that modulate the activity of other cells. Specifically, tissue structural cells such as fibroblasts and endothelial cells have recently been shown capable of producing cytokines.including IL-6 (15), IL-8 (16, 17), and the colony-stimulating factors (18-21). In addition, Smith and associates recently demonstrated that rat tracheal epithelial cells could produce GM-CSF (22). Because in the
Cox, Ohtoshi, Vancheri et al.: Steroid Modulation of HBEC-supported Eosinophil Survival
clinical conditions alluded to above, eosinophils are frequently seen in close relation to the epithelium, we asked whether HBEC release cytokines that could directly interact with human peripheral blood eosinophils. The objectives of the current work were twofold. First, to study the interaction between human airway epithelial cells and eosinophils with regard to eosinophil survival. Second, to investigate the mechanisms by which steroids, which have been shown to be effective in vivo reducing both blood (23) and tissue eosinophilia (2), might modulate this interaction. Our data show that CM from primary cultures of HBEC supports the survival of human peripheral blood eosinophils in vitro (Figure 1). This effect occurs in a dose-dependent manner, and concentrations of HBEC-CM as low as 0.1% elicit detectable activity. In addition to showing prolonged survival, eosinophils cultured in HBEC-CM became hypodense and on electron microscopy showed reduced granularity and increased vacuolation (data not shown). These changes in density and appearance are features consistent with an activated phenotype of the eosinophil (24-26) and are similar to the changes we have previously documented after exposure of eosinophils to CM from human lung fibroblasts (27). By means of a specific immunoassay, we detected GM-CSF in the HBEC-CM at levels ranging between 22 and 48 pglml. We then proceeded to examine whether inhibition fo the GM-CSF present in our HBECCM with a specific monoclonal neutralizing rat anti-human GM-CSF antibody affected eosinophil survival. As shown in Figure 2, preincubation of the HBEC-CM with this antibody resulted in complete abrogation of HBEC-CM-induced eosinophil survival at an antibody concentration of 1:100. Three cytokines, GM-CSF, IL-3, and IL-5, have been shown to cause prolonged survival and activation of human eosinophils (27-29). IL-5 was previously known as eosinophil differentiating factor - a murine factor capable of inducing both human and murine colonies to survive and undergo eosinophilic differentiation (29). Begley and colleagues (30) and Yamaguchi and associates (29) showed that IL-5 was active on mature human eosinophils, causing increased survival. In addition, Lopez and co-workers (31) demonstrated that IL-5 was a selective activator of human eosinophil function. Rothenberg and associates (28) examined the effects of IL-3 on human eosinophils and found it had similar actions to IL-5 with regards to survival, density, and activation. Begley and colleagues first documented the survival-promoting effects of recombinant GM-CSF (30), and this was later confirmed and expanded by Owen and co-workers (32). Interestingly, while the effect of GM-CSF on the viability of eosinophils was enhanced by the presence of 3T3 fibroblasts, the 3T3 feeder layer alone had no effect on eosinophil survival (32). We found, however, that GM-CSF derived from human lung fibroblast monolayers supported eosinophil survival and induced density and morphologic changes consistent with activation (27). Rothenberg and associates (8) have previously shown that eosinophils co-cultured with either human umbilical vein or bovine aortic endothelial cells had increased survival and functional activities. Lamas and associates (33) also used human umbilical vein endothelial cells as the source of eosinophil survival factor and showed its activity to be increased following IL-l treatment of the endothelial cell cultures. It is likely that GM-CSF mediated
529
this effect as endothelial cells produce increased amounts of this cytokine in response to stimulation by IL-I or tumor necrosis factor (19, 21, 34). While Danner and Luger (35) have shown that human keratinocytes can produce IL-3, Churchill and co-workers (36) found no IL-3 activity in supernatants of cultured human tracheal epithelial cells. Given that the source of IL-5 appears primarily restricted to cells oflymphoid origin (37), and that we can totally abrogate the survival effect with anti-GM-CSF, our data demonstrate that HBEC-derived GM-CSF is the cytokine causing prolonged survival of human eosinophils in our in vitro system. Corticosteroids are potent anti-inflammatory agents that are effective in the treatment of inflammatory disorders of the respiratory tract associated with tissue eosinophilia (38). Corticosteroids affect a number of steps in the inflammatory response (23,38-41) likely including the local accumulation of eosinophils. However, the specific mechanisms by which steroids elicit their effects are not well understood. In these studies, we used budesonide, a synthetic steroid available for clinical use in inhaled form in Canada and Europe, as well as dexamethasone. Budesonide is a potent corticosteroid, having 2 to 10 times the potency of dexamethasone, depending on the assay employed (42). We first examined the effect ofbudesonide and dexamethasone on GM-CSF production by HBEC. Corticosteroid treatment markedly reduced the basal production of GMCSF by HBEC, and CM generated from budesonide-treated HBEC elicited minimal eosinophil survival compared to CM from untreated HBEC cultures. This is in agreement with the findings of Lamas and associates, who showed that dexamethasone treatment of endothelial cell cultures markedly inhibited eosinophil survival induced by supernatant generated by those cells (33). As it is likely that during the course of inflammation in vivo, activated macrophages release cytokines that amplify the inflammatory response, we examined the effect of IL-l and budesonide on HBEC cultures. Similar to its action on fibroblasts and endothelial cells (18, 34), IL-l markedly stimulated the production ofGM-CSF by HBEC, and this was substantially prevented by treatment of the HBEC cultures with either budesonide or dexamethasone (Table 1). Eosinophils are known to have high-affinity steroid-binding sites (43), and direct effects of corticosteroids on eosinophils such as inhibition of adherence and chemotaxis (39,44) and inhibition of antibody-dependent cellular cytotoxicity (45) have been documented previously. Thus, we proceeded to examine whether budesonide had any direct effect on eosinophil survival. As shown in Table 2, neither budesonide nor the diluent directly altered eosinophil survival. However, eosinophils treated with budesonide showed a markedly reduced survival upon subsequent exposure to either HBEC-CM (Figure 3) or rhGM-CSF (Table 3). Dexamethasone had a similar but smaller effect. This inhibition of survival is not specific for GM-CSF, as budesonide also prevented IL-3- and IL-5-induced eosinophil survival (Table 3). Lamas and associates examined the direct effect of dexamethasone on eosinophil survival induced by endothelial cell supernatants (33) and found a small effect that was not statistically significant. Differences in the drug used and, more importantly, in the method of treatment might explain this apparent discrepancy. We exposed eosinophils to the ste-
530
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 4 1991
roid for 1 h prior to the addition of either GM-CSF or CM because we had found the corticosteroid effect to be considerably less when added either simultaneously with or 1 h after the cytokine or CM (data not shown). In addition, whereas Lamas and associates used dexamethasone at a dose of 10-7 M, we found our optimal effect with budesonide, which is in itself more potent than dexamethasone, at a dose of 10-6 M. The mechanism underlying the direct effect of corticosteroids on the capacity of eosinophils to respond to cytokines is unknown and might involve interference with membrane receptors, modulation of intracellular second messenger signals, or direct nuclear effects. Regardless of the specific mechanism involved, our data demonstrate two pathways by which the topical administration of corticosteroids might modulate the accumulation of eosinophils at a tissue site. In summary, we have presented evidence that HBEC markedly support survival of eosinophils in vitro and that this effect is mediated by GM-CSF. In addition, we have shown that corticosteroids interfere with this epithelial cell-eosinophil interaction by two independent mechanisms: one on the effector cell involving GM-CSF production by HBEC, and another on the target cell, the eosinophil, affecting the responsiveness of this cell to cytokines that are likely present in the microenvironment during chronic inflammation. Acknowledgments: The writers thank Drs. W. Bennett and J. E. M. Young, Department of Surgery, and Prof. J. M. Kay, Department of Pathology, St. Joseph's Hospital, Hamilton, Ontario, Canada, for their cooperation in providing access to surgically removed specimens. This work was supported by the Medical Research Council of Canada, the Council for Tobacco Research, and AB Draco, Lund, Sweden. Dr. Cox was supported by a Glaxo Research Fellowship. Dr. Jordana is a Research Fellow of the Parker B. Francis Research Foundation.
References 1. Spry, C. J. F. 1988. Eosinophils in disease: respiratory tract diseases. In: Eosinophils. Oxford United Press, New York. 193-212. 2. Beasley, R., W. R. Roche, J. A. Roberts, andS. T. Holgate. 1989. Cellular events in the bronchi in mild asthma and after bronchial provocation. Am. Rev. Respir. Dis. 139:806-817. 3. Spry, C. J. F. 1971. Mechanisms of eosinophilia. VI. Eosinophil mobilization. Cell Tissue Kinet. 4:365-374. 4. Krajina, Z., and A. Zirdum. 1987. Histochemical analysis of nasal polyps. Acta Otol. 103:435-440. 5. Wu, R., and M. M. J. Wu. 1986. Effects ofretinoids on human bronchial epithelial cells: differential regulation of hyaluronate synthesis and keratin protein synthesis. J. Cell. Physiol. 127:73-82. 6. Ohtoshi, T., C. Vancheri, G. Cox et al. 1991. Monocyte-macrophage differentiation induced by human upper airway epithelial cells. Am. J. Respir. Cell Mol. Bio!' 4:255-263. 7. Roberts, R. L., andJ. I. Gallin. 1985. Rapid method for isolation ofnorma1 human peripheral blood eosinophils on discontinuous percoll gradients and comparison with neutrophi1s. Blood 65:433-440. 8. Rothenberg, M. E., W. F. Owen, Jr., D. S. Silberstein, R. J. Soberman, K. F. Austen, and R. L. Stevens. 1987. Eosinophils cocultured with endothelial cells have increased survival and functional properties. Science 237:645-647. 9. Frick, W. E., J. B. Sedgwick, and W. W. Busse. 1988. Hypodense eosinophils in allergic rhinitis. J. Allergy Clin. Immunol. 82:119-125. 10. Peterson, M. W., M. Monick, and G. W. Hunninghake. 1987. Prognostic role of eosinophils in pulmonary fibrosis. Chest 92:51-56. 11. Gleich, G. J. 1990. The eosinophil and bronchial asthma: current understanding. J. Allergy Clin. Immuno!' 85:422-436. 12. Motojima, S., E. Frigas, D. A. Loegering, and G. J. Gleich. 1989. Toxicity of eosinophil cationic proteins for guinea pig tracheal epithelium in vitro. Am. Rev. Respir Dis. 139:801-805. 13. Venge, P., R. Dahl, K. Fredens, and C. G. B. Peterson. 1988. Epithelial
injury by human eosinophils. Am. Rev. Respir. Dis. 138:S54-S57. 14. Ayars, G. H., L. C. Altma, M. M. McManus eta!' 1989. Injurious effects of the eosinophil peroxide-hydrogen peroxide-halide system and major basic protein on human nasal epithelium in vitro. Am. Rev. Respir. Dis. 140:125-131. 15. Van Damme, J., S. Cayphas, J. Van Snick et al. 1987. Purification and characterization of human fibroblast-derived hybridoma growth factor identical to T-cell derived B-cell stimulatory factor-2 (interleukin-6). Eur. J. Biochem. 168:543-550. 16. Strieter, R. M., S. H. Phan, H. J. Showell et al. 1989. Monokine-induced neutrophil chemotactic factor gene expression in human fibroblasts. J. Bioi. Chem. 264:10621-10626. 17. Strieter, R. M., S. L. Kunkel, H. J. Showell, and R. M. Marks. 1988. Monokine-induced gene expression of a human endothelial cell-derived neutrophil chemotactic factor. Biochem. Biophys. Res. Commun. 156: 1340-1345. 18. Kaushansky, K., N. Lin, and J. W. Adamson. 1988. Interleukin-1 stimulates fibroblasts to synthesize granulocyte-macrophage and granulocyte colony-stimulating factors. J. Clin. Invest. 81:92-97. 19. Bagby, G. C., E. McCall, K. A. Bergstrom, and D. Burger. 1983. A monokine regulates CSA production by vascular endothelial cells. Blood 62:663-668. 20. Bagby, G. C., G. Shaw, and G. M. Segal. 1989. Human vascular endothelial cells, granulopoiesis, and the inflammatory response. J. Invest. Dermatol. 93:48S-52S. 21. Sieff, C. A., C.A. Niemeyer, and D. V. Faller. 1987. The production of hematopoietic growth factors by endothelial accessory cells. Blood Cells 13:65-74. 22. Smith, S. M., D. K. P. Lee, J. Lacy, and D. L. Coleman. 1990. Rat tracheal epithelial cells produce granulocyte/macrophage colony-stimulating factor. Am. J. Respir. Cell Mol. Bioi. 2:59-68. 23. Gordon, A. S. 1965. Some aspects of hormonal influences upon the leukocytes. Ann. NY Acad. Sci. 59:907-927. 24. DeSimone, C., G. Donell, D. Meli et al. 1982. Human eosinophils and parasitic diseases. II. Characterization of two cell fractions isolated at different densities. Clin. Exp. Immunol. 48:249-255. 25. Peters, M. S., G. J. Gleich, S. L. Dunnette, and T. Fukirda. 1988. Ultrastructural study of eosinophils from patients with the hypereosinophilic syndrome: a morphologic basis of hypodense eosinophils. Blood 71:780-785. 26. Prin, L., J. Chamo, M. Capron et al. 1984. Heterogeneity of human eosinophils. II. Variability of respiratory burst activity related to cell density. Clin. Exp. Immunol. 57:735-742. 27. Vancheri, C., J. Gau1die, J. Bienenstock et al. 1989. Human lung fibroblast-derived granulocyte-macrophage colony stimulating factor (GM-CSF) mediates eosinophil survival in vitro. Am. J. Respir. Cell Mol. Bioi. 1:289-295. 28. Rothenberg, M. E., W. F. Owen, Jr., D. S. Silbersteinetal. 1988. Human eosinophils have prolonged survival, enhanced functional properties and become hypodense when exposed to human interleukin-3. J. Clin. Invest. 81:1986-1992. 29. Yamaguchi, Y., T. Suda, J. Suda et al. 1988. Purified interleukin-5 supports the terminal differentiation and proliferation of murine eosinophilic precursors. J. Exp. Med. 167:43-56. 30. Begley, C. G., A. F. Lopez, N. A. Nicola et al. 1986. Purified colonystimulating factors enhance the survival of human neutrophils and eosinophils in vitro: a rapid and sensitive microassay for colony-stimulating factors. Blood 68:162-166. 31. Lopez, A. F., C. J. Sanderson, J. R. Gamble, H. D. Campbell, I. G. Young, and M. A. Vadas. 1988. Recombinant human interleukin-5 is a selective activator of human eosinophil function. J. Exp. Med. 167:219224. 32. Owen, W. F., Jr., M. F. Rothenberg, D. S. Silbersteinetal. 1987. Regulation of human eosinophil viability, density, and function by granulocytemacrophage colony-stimulating factor in the presence of 3T3 fibroblasts. J. Exp. Med. 166:129-141. 33. Lamas, A. M., G. V. Marcotte, and R. P. Schleimer. 1989. Human endothelial cells prolong eosinophil survival: regulation by cytokines and glucocorticoids. J. Immunol. 142:3978-3984. 34. Fibbe, W. E., M. R. Daha, P. S. Heimstra et al. 1989. Interleukin-1 and polytrlj-polytrC) induce production of granulocyte CSF, macrophage CSF, and granulocyte-macrophage CSF by human endothelial cells. Exp. Hematol. 17:229-234. 35. Danner, M., and T. A. Luger. 1987. Human keratinocytes and epidermoid carcinoma cell lines produce a cytokine with interleukin 3-like activity. J. Invest. Dermatol. 88:353-361. 36. Churchill, L., B. Friedman, R. P. Schleimer, and D. Proud. 1990. Granulocyte-macrophage colony-stimulating factor (GM-CSF) production by cultured human tracheal epithelial cells. J. Allergy Clin. Immuno!' 85:A358. (Abstr.) 37. Niemeyer, C. M., C. A. Sieff, B. Mathay-Prevot et al. 1989. Expression of human interleukin- 3 (multi-CSF) is restricted to human lymphocytes
Cox, Ohtoshi, Vancheri et al.: Steroid Modulation of HBEC-supported Eosinophil Survival
and T-cell tumor lines. Blood 73:945-951. 38. Schleimer, R. P. 1988. Glucocorticoids: their mechanism of action and use in allergic diseases. In Allergy Principles and Practice. 3rd ed. E. Middleton, Jr., C. E. Reed, E. F. Ellis, N. F. Adkinson, Jr., and J. W. Yunginger, editors. Mosby, Washington, DC. 739-765. 39. Kajita, T., Y. Yui, H. Mitaetal. 1981. Effects of corticosteroids oneosinophil chemotaxis and adherence. J. Clin. Invest. 67:28-36. 40. Bochner, B. S., B. K. Rutledge, and R. P. Schleimer. 1987. Interleukin-I (IL-I) production by human lung tissue. II. Inhibition by antiinflammatory steroids. J. Immunol. 139:2303-2307. 41. Rebuck, J. W., R. W. Smith, and R. R. Margulis. 1951. The modification of leukocytic function in human windows by ACTH. Gastroenterology 19:644-657.
531
42. Clissold, S. P., and R. C. Heel. 1984. Budesonide. A preliminary review of its pharmacodynamic properties and therapeutic efficacy in asthma and rhinitis. Drugs 28:485-518. 43. Peterson, A. P., L. C. Altman, J. S. Hill, K. Gosney, and M. E. Kadin. 1981. Glucocorticoid receptors in normal human eosinophils: comparison with neutrophils. J. Allergy Clin. Immunol. 68:212-217. 44. Wegner, C. D., R. H. Gundel, P. Reilly, N. Haynes, L. G. Letts, and R. Rothlein. 1990. Intercellular adhesion molecule-I (ICAM-I) in the pathogenesis of asthma. Science 247:456-459. 45. Hallam, C., D. I. Pritchard, S. Trigg, and R. P. Eady. 1982. Rat eosinophil-mediated antibody-dependent cellular cytotoxicity: investigations of the mechanisms of target lysis and inhibition by glucocorticoids. Clin. Exp. Immunol. 48:641-648.
Accumulation of eosinophils in the bronchial tissue occurs in a variety of inflammatory disorders of the human airway. We asked whether airway epithelial cells released factors that could influence eosinophil survival and thus contribute to accumulation of these cells in the tissues. Using conditioned medium (CM) generated from cultured human bronchial epithelial cells (HBEC), we examined the in vitro survival of eosinophils isolated from human peripheral blood. When cultured in control medium, more than 90% of the eosinophils were dead by day 4. In contrast, culture in HBEC-CM resulted in dose-dependent survival at day 6 of 69 ± 9.4%,40.5 ± 5.9%, and 25 ± 2% viability with 2,0.5, and 0.1% HBEC-CM, respectively (n = 4). Granulocyte/macrophage colony-stimulating factor (GM-CSF) was detected in the HBEC-CM by enzyme-linked immunosorbent assay at levels of 22 to 48 pg/ml. Furthermore, preincubation of the HBECCM with a neutralizing monoclonal antibody to human GM-CSF completely inhibited this increased survival of eosinophils. Because corticosteroids are potent eosinopenic agents, we also examined the effects of the synthetic steroid budesonide on this system. Budesonide inhibited both spontaneous and interleukin-l (IL-l)-induced GM-CSF production by cultured HBEC. In addition, preincubation of eosinophils with budesonide caused marked abrogation of the survival induced subsequently with either HBEC-CM or recombinant human GM-CSF. In summary, HBEC can support eosinophil survival via the elaboration of GM-CSF and thus may contribute to the local control of inflammatory cell accumulation. Steroids may modulate this process both by inhibiting cytokine production from HBEC and by a direct effect on eosinophils, preventing their response to cytokines.
Eosinophils accumulate in the respiratory mucosa in a number of inflammatory diseases such as asthma, allergic rhinitis, and nasal polyposis and have been implicated in the pathogenesis of these conditions (1, 2). Typically, in the bloodstream, eosinophils are thought to be short-lived cells (3), and the mechanisms underlying their accumulation at sites of inflammation are incompletely understood. As eosinophils are frequently within or close to the epithelial layer in inflamed tissue (2, 4), we asked whether human bronchial epithelial cells (HBEC) release products that support the (Received in original form August 1, 1990 and in revised form November 27, 1990) Address correspondence to: Dr. M. Jordana, Department of Pathology, McMaster University, 1200 Main Street West, Hamilton, Ontario, L8N 3Z5 Canada. Abbreviations: conditioned medium, CM; enzyme-linked immunosorbent assay, ELISA; granulocyte/macrophage colony-stimulating factor, GMCSF; human bronchial epithelial cell(s), HBEC; Hanks' balanced salt solution, HBSS; serum-free hormonally defined Ham's F-12medium, HD-Fl2; interleukin, IL; phosphate-buffered saline, PBS; recombinant human, rho Am. J. Respir. Cell Mol. BioI. Vol. 4. pp. 525-531, 1991
survival of eosinophils, thereby contributing to their local accumulation. Our results demonstrate that conditioned medium (CM) from HBEC markedly prolongs the in vitro survival of eosinophils, and that this effect is mediated by granulocyte/macrophage colony-stimulating factor (GM-CSF) in the CM. In addition, we demonstrate that the synthetic steroid budesonide inhibits both unstimulated and interleukin (IL)-l-stimulated GM-CSF production by HBEC, and moreover that budesonide has a direct effect on eosinophils as it prevents the ability of these cells to respond to cytokines such as GM-CSF.
Materials and Methods Reagents Budesonide was kindly provided by Drs. B. Anderson and M. Vangielzehm of AB Draco (Lund, Sweden). A stock solution of 10-4 M was made in 30% ethanol/70% phosphatebuffered saline (PBS), and further dilutions were made using the same diluent, which was also used at equivalent concentrations in all control wells. In addition, a dexamethasone
526
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 4 1991
(Organon, Toronto, Canada) stock solution of 10-4 M diluted in PBS was prepared. Cytokines Recombinant human (rh) GM-CSF, (sp act, 9.3 x lQ6 U/mg) and IL-3 (sp act, 4 x 106 U/mg) were obtained from Genetics Institute (Boston, MA). rhIL-5 (sp act, 109 U/mg) and a rat IgG 2a neutralizing monoclonal antibody to human GM-CSF was kindly provided by Dr. 1. Abrams (DNAX Research Institute, Palo Alto, CA); l ug/ml of this antibody neutralizes 2 ng/ml of human GM-CSF in a K61 bioassay. GM-CSF was assayed by an enzyme-linked immunosorbent assay (ELISA) (Genzyme Corp., Boston, MA). The linear portion of the standard curve corresponds to GMCSF levels in the 10 to 250 pglml range.
added to a concentration of 10-6 M. Untreated cultures were exposed to an equal volume of diluent. The supernatants were handled as above until the assay for GM-CSF was performed. In experiments involving cytokine neutralization by antibody, the HBEC-CM was preincubated with the antibody at 37° C for 1 h prior to its addition to the eosinophil suspension. Eosinophil Isolation After informed consent, eosinophils were isolated by a previously described method (7), from the peripheral blood of atopic patient volunteers with eosinophil differentials of 4 to 10%. Ten times-concentrated PBS was added to Percoll solution (Pharmacia, Uppsala, Sweden) to achieve an osmolarity of 310 to 320 mosm/kg. This was further diluted with regular PBS to 75 and 65 % solutions. Discontinuous gradients were made using 15 ml of each solution layered in a 50-ml conical tube (Coming Glassworks, Coming, NY). Heparinized venous blood was incubated with 10-6 M N-formylmethionylleucylphenylalanine (FMLP) at 37° C for 15 min. Fifteen milliliters of blood were then layered on top of the Percoll gradient and centrifuged at 400 X g (1,500 rpm) at 22° C for 25 min. The plasma and the neutrophil/mononuclear cell band were discarded, and 10 ml was collected from the 65 to 75% interface (density, > 1.085 g/ml). This was mixed with 10 ml HBSS without calcium and centrifuged at 530 X g (2,000 rpm) for 10 min. The cells were washed once more in HBSS, and erythrocytes were removed by cold hypotonic lysis. The cells were then resuspended in supplemented RPM!. Purity as assessed on cytospins stained with Diff-Quikf (American Hospital Supply Corp., McGaw Park, IL), a modification of WrightGiemsa staining, was greater than 90 % and viability was greater than 99% on trypan blue dye exclusion in all experiments.
HBEC Cultures Specimens of human bronchi were obtained from organs removed at surgery. All patients had a primary pulmonary neoplasm requiring lobectomy or pneumonectomy. A portion of the bronchus at a site distant from the lesion requiring resection was taken and transported to the laboratory in minimal essential medium. HBEC were dispersed as previously described (5). Briefly, tissue was rinsed 3 times in Hanks' balanced salt solution (HBSS) and incubated overnight at 4° C in 0.1% protease (Sigma Chemical Co., St. Louis, MO) solution in Ham's F-12 medium. Heat-inactivated fetal bovine serum (GIBCO, Grand Island, NY) was added to 10% vol/vol, and cells were detached by gentle agitation. The suspension was then filtered through a 60-l-'m Nitex mesh (Cellector; Sargent-Welch, Toronto, Canada) and centrifuged at 270 X g (1,000 rpm) for 10 min. The cell pellet was resuspended in serum-free hormonally defined Ham's F-12 medium (HD-Fl2) containing 1% penicillin/streptomycin, 5 I-'glml insulin (GIBCO), 5 I-'glml transferrin (GIBCO), 25 I-'g/ml epidermal growth factor (Collaborative Research, Lexington, MA), 15 I-'g/ml endothelial cell growth supplement (Collaborative Research), 2 x 10-10 M triiodothyronin (GIBCO), and 10-7 M hydrocortisone (GIBCO). Fiftythousand cells in 2 ml HD-FI2 were plated onto collagencoated wells, and cells were cultured in a humidified atmosphere at 37° C with 5% CO 2 , The medium was changed at day 1 and on alternate days thereafter. The monolayer achieved confluency at days 7 to 12. Using this method in upper airway tissues, we have previously shown that by immunohistochemistry ~ 95% of the cells in the culture stain positive for keratin, an extremely small number of cells stain positive for vimentin, and in addition, HBEC cultures established in this manner do not stain with antibody to the monocyte CD-14 antigen (6).
Eosinophil Culture Eosinophils were incubated in 24-well tissue culture plates (Nunclon; GIBCO, Burlington, Canada) at a concentration of 250,000 cells in 1 ml. Survival was assessed at days 2,4, and 6 and was calculated using the formula: 100 X [(number of eosinophils recovered) X (percentage of cells viable by trypan blue exclusion)]/number of eosinophils originally delivered (8). In all experiments, duplicate wells of each condition were assessed. In experiments involving steroid treatment, either budesonide or dexamethasone, at 10-6 M, was added to the eosinophil suspensions 1 h prior to the addition of HBEC-CM or recombinant human cytokines. Wells with untreated cells received the same volume of diluent.
Generation of HBEC-CM When a confluent monolayer was achieved, the medium was changed to RPMI supplemented with 10% FBS, 1% penicillin/streptomycin, and 1% Hepes buffer. At confluence, there were 270,000 to 340,000 cells/plate. HBEC-CM was generated with supplemented RPMI from these monolayers over 48 h and then pooled, filtered (0.45 I-'m; Millipore Corp., Bedford, MA), and stored at -20° C until use. In the treatment of HBEC cultures, budesonide or dexamethasone was
Eosinophil Density After culture, eosinophil density was assessed after centrifugation on continuous Percoll gradients as previously described (9). Continuous gradients were generated by placing 10 ml of Percoli solution (1.084g/ml) in 15-ml polypropylene tubes (Falcon, Oxnard, CA) and centrifuging at 20,000 X g for 15 min at 22° C in a fixed-angle rotor. One milliliter of eosinophil suspension was subsequently placed on top and centrifuged at 100 X g for 20 min at 22° C. Control gra-
Cox, Ohtoshi, Vancheri et al.: Steroid Modulation of HBEC-supported Eosinophil Survival
100
527
100
~ HBEC-CM •
~
,.,
75
75 ..J
e.'"
...J
!i!
~
sa:
50
sII:
::l
50
. (/)
~
~
Ul
~
HBEC-CM + Antl-GM-CSF
25
o
25
0.1
0.5
2
0+---
10
EPITHELIAL CELL-CONDITIONED MEDIUM (0/0)
Figure1. Dose response of eosinophil survival in human bronchial epithelial cell-conditioned medium (HBEC-CM). HBEC-CM was generated over 48 h from various primary cultures. Freshly isolated eosinophils were cultured in RPMI enriched with increasing concentrations of HBEC-CM, and survival (corrected for cell number, see MATERIALS AND METHODS) was assessed at day 6 by trypan blue dye exclusion. Resultsshown are the means (± SD) offour independent experiments.
dients were spun with I rnl RPM!. Density of the fractions of the control gradients was measured by density meter (DMA40; Anton Paar, Graz, Austria). Calibration of this instrument was done on each day of experiments. Statistics The statistical significance of differences in result of the experiments was evaluated using a Student's t test for unpaired data except in Table 1, where a t test for paired data was employed.
DAY 4
DAY 6
Figure 2. Effect of anti-granulocyte/macrophage colony-stimulating factor (anti-GM-CSF) on eosinophil survival. The HBEC-CM was incubated with anti-GM-CSF at 1:100 dilution for 1 hat 37° C prior to adding to the eosinophil culture. Results shown are the means (± SD) of four independent experiments. The differences between the anti-GM-CSF-treated and the untreated HBEC-CM survivals at both time points are statistically significant (P < 0.01).
GM-CSF Content of HBEC-eM and Effect of Steroid Treatment The individual GM-CSF content of five different primary HBEC cultures was 22, 27, 31, 37, and 48 pglml (Table lA). Addition of budesonide at 10-6 M to these HBEC cultures consistently decreased GM-CSF production from 33 ± 10 to 17.2 ± 11.9 pg/rnl (P< 0.05). Treatment of the HBEC cultures with IL-l (125 Vlml) stimulated GM-CSF production from mean 27 to 38.9 pglrnl (P < 0.05) (Table lB). Budesonide treatment of the HBEC cultures inhibited ILl-induced GM-CSF production (n = 3). Similar to the effect
Results Effect of HBEC-CM on Eosinophil Survival and Density The survival of human peripheral blood eosinophils in supplemented RPMI, in vitro, was always less than 10% at day 4. In contrast, survival at day 4 was promoted when eosinophils were incubated with 2 % HBEC-CM: 72.3 ± 5.5% versus 4.3 ± 2.3 % (n = 4; P < 0.01). This effect to promote survival is dose dependent: survival at day 6 was 69 ± 9.4 %, 40.5 ± 5.9%, and 25 ± 2% in the presence of 2, 0.5, and 0.1% HBEC-CM, respectively (n = 4) (Figure 1). The differences in survival in 0.5 % and 0.1% are statistically significantly different from survival with 2 and 10% HBEC-CM. Preincubation of the HBEC-CM with a neutralizing monoclonal antibody to human GM-CSF resulted in complete inhibition of eosinophil survival. Figure 2 shows that eosinophil survival decreased from 72 to 4 % at day 4 and from 69 % to 0 at day 6 (n = 4), with anti-GM-CSF. In separate experiments, HBEC-CM was preincubated with antibody to IL-3 and no difference in eosinophil survival at day 6 was detected compared with untreated HBEC-CM: 46.6 ± 5.2 % versus 45.5 ± 5.1%, respectively (n = 3). At the time of isolation, all the eosinophils were harvested from the fraction of the Percoll gradient with density> 1.085 glrnl. In contrast, 60 ± 6% of the viable eosinophils had density less than 1.080 g/liter (n = 3) after culture with HBEC-CM for 6 d.
TABLE 1
Granulocyte/macrophage content in human bronchial epithelial cell-conditioned medium* A. No.1 No. 2 No. 3 No.4 No.5 Mean B.
No.5· No. 6 No. 7 No. 8 Mean
+ Budesonide
RPMI
22 27 31 37 48 33
+ Dexamethasone
11
8.5 8.5 22 36 17.2
RPMI
+ IL-I
48 19 33 8 27
63 27.5 53 12 38.9
8.4 24
Budesonide + IL-I 49 10
7.5
Dexamethasone + IL-I 48 20 8.5
* Values are given as pg/m!. A. Confluent human bronchial epithelial cell monolayers were maintained in RPMI supplemented with 10% fetal calfserum andtreated with budesonide or dexamethasone at 10-6 M or an equal volume of diluent. The differences in granulocyte/macrophage colony-stimulating factor content between control and budesonide-treated cultures are statistically significant (P < 0.05). B. IL-I (125 V/m!) was added to cultures I h aftereither steroid (10" M) ordiluent treatment. The differences between control andIL-l-treated cultures are statistically significant (P < 0.05).
528
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 4 1991
TABLE 2
TABLE 3
Eosinophil survival with diluent or budesonide*
Effect of steroid treatment of the eosinophils on cytokine-induced survival *
Eosinophil Survival (%)
RPM! Diluent Budesonide
Day 2
Day 4
84 ± 6.7 88 ± 6 79 ± 5.6
6 ± 3.6 4 ± 3.1 5 ± 2.7
* Freshly isolated eosinophils were cultured in RPMI alone or RPMI containing either budesonide at 10-0 M or an equal volume of the ethanol/phosphate-buffered saline diluent solution. Results are the mean ± SO of three independent experiments. The differences are not statistically significant. of budesonide, dexamethasone (10-6 M) inhibited both basal and IL-l-induced output (n = 2) (Table lA).
Eosinophil Survival (%)
Control Budesonide Dexamethasone
100
iii'
75
co e.'" ..J
~ a:
50
:;: ::::l
C/)
al
25
o UNTREATED
DILUENT
BUDESONIDE
Figure 3. Effect of pretreatment of eosinophils on their survival in HBEC-CM. Freshly isolated eosinophils were treated with budesonide at 10- 6 M for 1 h prior to adding the HBEC-CM. The ethanol/phosphate-buffered saline diluent was added in volume equal to that delivered to the budesonide-treated samples. Survival was assessed at day 6. Results are expressed as means (± SD) of three independent experiments. The survival of the budesonidetreated samples is statistically significantly less than in the untreated and diluent-treated samples (P < 0.01).
IL-3
IL-5
69.5 ± 6.5 22.2 ± 6.8t 49.9 ± 4.7*
64.5 ± 9.5 16.8 ± 9t 51.6 ± 0.7
55.9 ± 6 23.6 ± 4.8t 46 ± 3.45
Definition of abbreviations: GM-CSF = granulocyte/macrophage colonystimulating factor; IL-3 and IL-5 = interleukin-3 and -5, respectively. * Freshly isolated eosinophils were pretreated with budesonide or dexamethasone at 10-0 M I h prior to the addition of cytokine. Control samples received an equal volume of the ethanol/phosphate-buffered saline diluent. GM-CSF was added at 10 U/rnl at day O. IL-3 was added at 10 U/rnl on days 0 and 3. IL-5 was added at 100 U/rnl on days 0 and 3. Survival was assessed at day 6. Results are the mean ± SO of three independent experiments. t
Effect of Steroid Treatment on Eosinophil Survival Eosinophil survival with CM generated in the presence of corticosteroid was assessed. In these experiments, different wells of the primary HBEC culture were treated with either budesonide or dexamethasone. While untreated 2 % HBECCM supported eosinophil survival of 60 ± 2.9 % at day 6, HBEC-CM generated in the presence ofbudesonide or dexamethasone at 10-6 M induced survival of 4.3 ± 3.2 % and 8.1 8 ± 2.7%, respectively (n = 3; P < 0.05). The addition ofbudesonide, or the diluent solution to eosinophil cultures, did not affect eosinophil survival in the absence of HBEC-CM or cytokines (Table 2). However, eosinophils that had been exposed to steroid prior to the addition of HBEC-CM showed a marked reduction in survival, from 69% to 10% (Figure 3). As shown in Table 3, the addition of recombinant human (rh) GM-CSF resulted in eosinophil survival at day 6 of 69 ± 6.5% (n = 5), and this increased survival was similarly inhibited by steroid pretreatment of the eosinophils. The effect of budesonide was dose dependent; eosinophil survival with 10 U/ml rhGMCSF was 22.2 ± 2.3%, 54.5 ± 3.8%, and 59.4 ± 6.1% with
GM-CSF
z c o.m.
*p < 0.05.
budesonide at 10-6 , 10-8 , and 10-10 M, respectively (n ~ 3; P < 0.05 for 10-6 and 10-8 M budesonide). Because IL-3 and IL-5 have also been shown to promote eosinophil survival, we examined whether steroids would also prevent their survival-enhancing effects. Eosinophils incubated with rhIL-3 (10 U/m!) and rhIL-5 (100 U/m!) had prolonged survival of 64.5 ± 9.5% and 55.9 ± 6% at day 6. Eosinophil survival was reduced to 16.8 ± 9% and 23.6 ± 4.8% at day 6 (n = 3) (Thble 3), when the eosinophils were treated with budesonide prior to the addition of these cytokines. With all cytokines, treatment with dexamethasone also caused reduced survival, but the effect was less marked than that seen with budesonide (Table 3).
Discussion Allergic rhinitis, nasal polyposis, and asthma are diseases of the respiratory tract characterized by chronic inflammation in which the accumulation of eosinophils in the tissue is a prevalent feature (1). These cells have also been found in the bronchoalveolar lavage of patients with idiopathic pulmonary fibrosis, and their presence has been associated with a poorer prognosis (10). It has been suggested that the contribution of these cells to the disease process resides in their ability to release products that can cause tissue damage (11-14). The generation of tissue eosinophilia is a complex process, only partly understood, which involves a series of steps from bone-marrow production and exit to the bloodstream, adherence to the endothelium, migration into and through tissues, and finally activation with release of granule contents. Because eosinophils in the bloodstream are shortlived cells (3), the extent to which they survive in the tissue likely depends on locally generated signals and represents an important factor in the accumulation of these cells at a particular site. It has become increasingly clear that within inflamed tissues a number of resident cells have the ability to release a variety of cytokines that modulate the activity of other cells. Specifically, tissue structural cells such as fibroblasts and endothelial cells have recently been shown capable of producing cytokines.including IL-6 (15), IL-8 (16, 17), and the colony-stimulating factors (18-21). In addition, Smith and associates recently demonstrated that rat tracheal epithelial cells could produce GM-CSF (22). Because in the
Cox, Ohtoshi, Vancheri et al.: Steroid Modulation of HBEC-supported Eosinophil Survival
clinical conditions alluded to above, eosinophils are frequently seen in close relation to the epithelium, we asked whether HBEC release cytokines that could directly interact with human peripheral blood eosinophils. The objectives of the current work were twofold. First, to study the interaction between human airway epithelial cells and eosinophils with regard to eosinophil survival. Second, to investigate the mechanisms by which steroids, which have been shown to be effective in vivo reducing both blood (23) and tissue eosinophilia (2), might modulate this interaction. Our data show that CM from primary cultures of HBEC supports the survival of human peripheral blood eosinophils in vitro (Figure 1). This effect occurs in a dose-dependent manner, and concentrations of HBEC-CM as low as 0.1% elicit detectable activity. In addition to showing prolonged survival, eosinophils cultured in HBEC-CM became hypodense and on electron microscopy showed reduced granularity and increased vacuolation (data not shown). These changes in density and appearance are features consistent with an activated phenotype of the eosinophil (24-26) and are similar to the changes we have previously documented after exposure of eosinophils to CM from human lung fibroblasts (27). By means of a specific immunoassay, we detected GM-CSF in the HBEC-CM at levels ranging between 22 and 48 pglml. We then proceeded to examine whether inhibition fo the GM-CSF present in our HBECCM with a specific monoclonal neutralizing rat anti-human GM-CSF antibody affected eosinophil survival. As shown in Figure 2, preincubation of the HBEC-CM with this antibody resulted in complete abrogation of HBEC-CM-induced eosinophil survival at an antibody concentration of 1:100. Three cytokines, GM-CSF, IL-3, and IL-5, have been shown to cause prolonged survival and activation of human eosinophils (27-29). IL-5 was previously known as eosinophil differentiating factor - a murine factor capable of inducing both human and murine colonies to survive and undergo eosinophilic differentiation (29). Begley and colleagues (30) and Yamaguchi and associates (29) showed that IL-5 was active on mature human eosinophils, causing increased survival. In addition, Lopez and co-workers (31) demonstrated that IL-5 was a selective activator of human eosinophil function. Rothenberg and associates (28) examined the effects of IL-3 on human eosinophils and found it had similar actions to IL-5 with regards to survival, density, and activation. Begley and colleagues first documented the survival-promoting effects of recombinant GM-CSF (30), and this was later confirmed and expanded by Owen and co-workers (32). Interestingly, while the effect of GM-CSF on the viability of eosinophils was enhanced by the presence of 3T3 fibroblasts, the 3T3 feeder layer alone had no effect on eosinophil survival (32). We found, however, that GM-CSF derived from human lung fibroblast monolayers supported eosinophil survival and induced density and morphologic changes consistent with activation (27). Rothenberg and associates (8) have previously shown that eosinophils co-cultured with either human umbilical vein or bovine aortic endothelial cells had increased survival and functional activities. Lamas and associates (33) also used human umbilical vein endothelial cells as the source of eosinophil survival factor and showed its activity to be increased following IL-l treatment of the endothelial cell cultures. It is likely that GM-CSF mediated
529
this effect as endothelial cells produce increased amounts of this cytokine in response to stimulation by IL-I or tumor necrosis factor (19, 21, 34). While Danner and Luger (35) have shown that human keratinocytes can produce IL-3, Churchill and co-workers (36) found no IL-3 activity in supernatants of cultured human tracheal epithelial cells. Given that the source of IL-5 appears primarily restricted to cells oflymphoid origin (37), and that we can totally abrogate the survival effect with anti-GM-CSF, our data demonstrate that HBEC-derived GM-CSF is the cytokine causing prolonged survival of human eosinophils in our in vitro system. Corticosteroids are potent anti-inflammatory agents that are effective in the treatment of inflammatory disorders of the respiratory tract associated with tissue eosinophilia (38). Corticosteroids affect a number of steps in the inflammatory response (23,38-41) likely including the local accumulation of eosinophils. However, the specific mechanisms by which steroids elicit their effects are not well understood. In these studies, we used budesonide, a synthetic steroid available for clinical use in inhaled form in Canada and Europe, as well as dexamethasone. Budesonide is a potent corticosteroid, having 2 to 10 times the potency of dexamethasone, depending on the assay employed (42). We first examined the effect ofbudesonide and dexamethasone on GM-CSF production by HBEC. Corticosteroid treatment markedly reduced the basal production of GMCSF by HBEC, and CM generated from budesonide-treated HBEC elicited minimal eosinophil survival compared to CM from untreated HBEC cultures. This is in agreement with the findings of Lamas and associates, who showed that dexamethasone treatment of endothelial cell cultures markedly inhibited eosinophil survival induced by supernatant generated by those cells (33). As it is likely that during the course of inflammation in vivo, activated macrophages release cytokines that amplify the inflammatory response, we examined the effect of IL-l and budesonide on HBEC cultures. Similar to its action on fibroblasts and endothelial cells (18, 34), IL-l markedly stimulated the production ofGM-CSF by HBEC, and this was substantially prevented by treatment of the HBEC cultures with either budesonide or dexamethasone (Table 1). Eosinophils are known to have high-affinity steroid-binding sites (43), and direct effects of corticosteroids on eosinophils such as inhibition of adherence and chemotaxis (39,44) and inhibition of antibody-dependent cellular cytotoxicity (45) have been documented previously. Thus, we proceeded to examine whether budesonide had any direct effect on eosinophil survival. As shown in Table 2, neither budesonide nor the diluent directly altered eosinophil survival. However, eosinophils treated with budesonide showed a markedly reduced survival upon subsequent exposure to either HBEC-CM (Figure 3) or rhGM-CSF (Table 3). Dexamethasone had a similar but smaller effect. This inhibition of survival is not specific for GM-CSF, as budesonide also prevented IL-3- and IL-5-induced eosinophil survival (Table 3). Lamas and associates examined the direct effect of dexamethasone on eosinophil survival induced by endothelial cell supernatants (33) and found a small effect that was not statistically significant. Differences in the drug used and, more importantly, in the method of treatment might explain this apparent discrepancy. We exposed eosinophils to the ste-
530
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 4 1991
roid for 1 h prior to the addition of either GM-CSF or CM because we had found the corticosteroid effect to be considerably less when added either simultaneously with or 1 h after the cytokine or CM (data not shown). In addition, whereas Lamas and associates used dexamethasone at a dose of 10-7 M, we found our optimal effect with budesonide, which is in itself more potent than dexamethasone, at a dose of 10-6 M. The mechanism underlying the direct effect of corticosteroids on the capacity of eosinophils to respond to cytokines is unknown and might involve interference with membrane receptors, modulation of intracellular second messenger signals, or direct nuclear effects. Regardless of the specific mechanism involved, our data demonstrate two pathways by which the topical administration of corticosteroids might modulate the accumulation of eosinophils at a tissue site. In summary, we have presented evidence that HBEC markedly support survival of eosinophils in vitro and that this effect is mediated by GM-CSF. In addition, we have shown that corticosteroids interfere with this epithelial cell-eosinophil interaction by two independent mechanisms: one on the effector cell involving GM-CSF production by HBEC, and another on the target cell, the eosinophil, affecting the responsiveness of this cell to cytokines that are likely present in the microenvironment during chronic inflammation. Acknowledgments: The writers thank Drs. W. Bennett and J. E. M. Young, Department of Surgery, and Prof. J. M. Kay, Department of Pathology, St. Joseph's Hospital, Hamilton, Ontario, Canada, for their cooperation in providing access to surgically removed specimens. This work was supported by the Medical Research Council of Canada, the Council for Tobacco Research, and AB Draco, Lund, Sweden. Dr. Cox was supported by a Glaxo Research Fellowship. Dr. Jordana is a Research Fellow of the Parker B. Francis Research Foundation.
References 1. Spry, C. J. F. 1988. Eosinophils in disease: respiratory tract diseases. In: Eosinophils. Oxford United Press, New York. 193-212. 2. Beasley, R., W. R. Roche, J. A. Roberts, andS. T. Holgate. 1989. Cellular events in the bronchi in mild asthma and after bronchial provocation. Am. Rev. Respir. Dis. 139:806-817. 3. Spry, C. J. F. 1971. Mechanisms of eosinophilia. VI. Eosinophil mobilization. Cell Tissue Kinet. 4:365-374. 4. Krajina, Z., and A. Zirdum. 1987. Histochemical analysis of nasal polyps. Acta Otol. 103:435-440. 5. Wu, R., and M. M. J. Wu. 1986. Effects ofretinoids on human bronchial epithelial cells: differential regulation of hyaluronate synthesis and keratin protein synthesis. J. Cell. Physiol. 127:73-82. 6. Ohtoshi, T., C. Vancheri, G. Cox et al. 1991. Monocyte-macrophage differentiation induced by human upper airway epithelial cells. Am. J. Respir. Cell Mol. Bio!' 4:255-263. 7. Roberts, R. L., andJ. I. Gallin. 1985. Rapid method for isolation ofnorma1 human peripheral blood eosinophils on discontinuous percoll gradients and comparison with neutrophi1s. Blood 65:433-440. 8. Rothenberg, M. E., W. F. Owen, Jr., D. S. Silberstein, R. J. Soberman, K. F. Austen, and R. L. Stevens. 1987. Eosinophils cocultured with endothelial cells have increased survival and functional properties. Science 237:645-647. 9. Frick, W. E., J. B. Sedgwick, and W. W. Busse. 1988. Hypodense eosinophils in allergic rhinitis. J. Allergy Clin. Immunol. 82:119-125. 10. Peterson, M. W., M. Monick, and G. W. Hunninghake. 1987. Prognostic role of eosinophils in pulmonary fibrosis. Chest 92:51-56. 11. Gleich, G. J. 1990. The eosinophil and bronchial asthma: current understanding. J. Allergy Clin. Immuno!' 85:422-436. 12. Motojima, S., E. Frigas, D. A. Loegering, and G. J. Gleich. 1989. Toxicity of eosinophil cationic proteins for guinea pig tracheal epithelium in vitro. Am. Rev. Respir Dis. 139:801-805. 13. Venge, P., R. Dahl, K. Fredens, and C. G. B. Peterson. 1988. Epithelial
injury by human eosinophils. Am. Rev. Respir. Dis. 138:S54-S57. 14. Ayars, G. H., L. C. Altma, M. M. McManus eta!' 1989. Injurious effects of the eosinophil peroxide-hydrogen peroxide-halide system and major basic protein on human nasal epithelium in vitro. Am. Rev. Respir. Dis. 140:125-131. 15. Van Damme, J., S. Cayphas, J. Van Snick et al. 1987. Purification and characterization of human fibroblast-derived hybridoma growth factor identical to T-cell derived B-cell stimulatory factor-2 (interleukin-6). Eur. J. Biochem. 168:543-550. 16. Strieter, R. M., S. H. Phan, H. J. Showell et al. 1989. Monokine-induced neutrophil chemotactic factor gene expression in human fibroblasts. J. Bioi. Chem. 264:10621-10626. 17. Strieter, R. M., S. L. Kunkel, H. J. Showell, and R. M. Marks. 1988. Monokine-induced gene expression of a human endothelial cell-derived neutrophil chemotactic factor. Biochem. Biophys. Res. Commun. 156: 1340-1345. 18. Kaushansky, K., N. Lin, and J. W. Adamson. 1988. Interleukin-1 stimulates fibroblasts to synthesize granulocyte-macrophage and granulocyte colony-stimulating factors. J. Clin. Invest. 81:92-97. 19. Bagby, G. C., E. McCall, K. A. Bergstrom, and D. Burger. 1983. A monokine regulates CSA production by vascular endothelial cells. Blood 62:663-668. 20. Bagby, G. C., G. Shaw, and G. M. Segal. 1989. Human vascular endothelial cells, granulopoiesis, and the inflammatory response. J. Invest. Dermatol. 93:48S-52S. 21. Sieff, C. A., C.A. Niemeyer, and D. V. Faller. 1987. The production of hematopoietic growth factors by endothelial accessory cells. Blood Cells 13:65-74. 22. Smith, S. M., D. K. P. Lee, J. Lacy, and D. L. Coleman. 1990. Rat tracheal epithelial cells produce granulocyte/macrophage colony-stimulating factor. Am. J. Respir. Cell Mol. Bioi. 2:59-68. 23. Gordon, A. S. 1965. Some aspects of hormonal influences upon the leukocytes. Ann. NY Acad. Sci. 59:907-927. 24. DeSimone, C., G. Donell, D. Meli et al. 1982. Human eosinophils and parasitic diseases. II. Characterization of two cell fractions isolated at different densities. Clin. Exp. Immunol. 48:249-255. 25. Peters, M. S., G. J. Gleich, S. L. Dunnette, and T. Fukirda. 1988. Ultrastructural study of eosinophils from patients with the hypereosinophilic syndrome: a morphologic basis of hypodense eosinophils. Blood 71:780-785. 26. Prin, L., J. Chamo, M. Capron et al. 1984. Heterogeneity of human eosinophils. II. Variability of respiratory burst activity related to cell density. Clin. Exp. Immunol. 57:735-742. 27. Vancheri, C., J. Gau1die, J. Bienenstock et al. 1989. Human lung fibroblast-derived granulocyte-macrophage colony stimulating factor (GM-CSF) mediates eosinophil survival in vitro. Am. J. Respir. Cell Mol. Bioi. 1:289-295. 28. Rothenberg, M. E., W. F. Owen, Jr., D. S. Silbersteinetal. 1988. Human eosinophils have prolonged survival, enhanced functional properties and become hypodense when exposed to human interleukin-3. J. Clin. Invest. 81:1986-1992. 29. Yamaguchi, Y., T. Suda, J. Suda et al. 1988. Purified interleukin-5 supports the terminal differentiation and proliferation of murine eosinophilic precursors. J. Exp. Med. 167:43-56. 30. Begley, C. G., A. F. Lopez, N. A. Nicola et al. 1986. Purified colonystimulating factors enhance the survival of human neutrophils and eosinophils in vitro: a rapid and sensitive microassay for colony-stimulating factors. Blood 68:162-166. 31. Lopez, A. F., C. J. Sanderson, J. R. Gamble, H. D. Campbell, I. G. Young, and M. A. Vadas. 1988. Recombinant human interleukin-5 is a selective activator of human eosinophil function. J. Exp. Med. 167:219224. 32. Owen, W. F., Jr., M. F. Rothenberg, D. S. Silbersteinetal. 1987. Regulation of human eosinophil viability, density, and function by granulocytemacrophage colony-stimulating factor in the presence of 3T3 fibroblasts. J. Exp. Med. 166:129-141. 33. Lamas, A. M., G. V. Marcotte, and R. P. Schleimer. 1989. Human endothelial cells prolong eosinophil survival: regulation by cytokines and glucocorticoids. J. Immunol. 142:3978-3984. 34. Fibbe, W. E., M. R. Daha, P. S. Heimstra et al. 1989. Interleukin-1 and polytrlj-polytrC) induce production of granulocyte CSF, macrophage CSF, and granulocyte-macrophage CSF by human endothelial cells. Exp. Hematol. 17:229-234. 35. Danner, M., and T. A. Luger. 1987. Human keratinocytes and epidermoid carcinoma cell lines produce a cytokine with interleukin 3-like activity. J. Invest. Dermatol. 88:353-361. 36. Churchill, L., B. Friedman, R. P. Schleimer, and D. Proud. 1990. Granulocyte-macrophage colony-stimulating factor (GM-CSF) production by cultured human tracheal epithelial cells. J. Allergy Clin. Immuno!' 85:A358. (Abstr.) 37. Niemeyer, C. M., C. A. Sieff, B. Mathay-Prevot et al. 1989. Expression of human interleukin- 3 (multi-CSF) is restricted to human lymphocytes
Cox, Ohtoshi, Vancheri et al.: Steroid Modulation of HBEC-supported Eosinophil Survival
and T-cell tumor lines. Blood 73:945-951. 38. Schleimer, R. P. 1988. Glucocorticoids: their mechanism of action and use in allergic diseases. In Allergy Principles and Practice. 3rd ed. E. Middleton, Jr., C. E. Reed, E. F. Ellis, N. F. Adkinson, Jr., and J. W. Yunginger, editors. Mosby, Washington, DC. 739-765. 39. Kajita, T., Y. Yui, H. Mitaetal. 1981. Effects of corticosteroids oneosinophil chemotaxis and adherence. J. Clin. Invest. 67:28-36. 40. Bochner, B. S., B. K. Rutledge, and R. P. Schleimer. 1987. Interleukin-I (IL-I) production by human lung tissue. II. Inhibition by antiinflammatory steroids. J. Immunol. 139:2303-2307. 41. Rebuck, J. W., R. W. Smith, and R. R. Margulis. 1951. The modification of leukocytic function in human windows by ACTH. Gastroenterology 19:644-657.
531
42. Clissold, S. P., and R. C. Heel. 1984. Budesonide. A preliminary review of its pharmacodynamic properties and therapeutic efficacy in asthma and rhinitis. Drugs 28:485-518. 43. Peterson, A. P., L. C. Altman, J. S. Hill, K. Gosney, and M. E. Kadin. 1981. Glucocorticoid receptors in normal human eosinophils: comparison with neutrophils. J. Allergy Clin. Immunol. 68:212-217. 44. Wegner, C. D., R. H. Gundel, P. Reilly, N. Haynes, L. G. Letts, and R. Rothlein. 1990. Intercellular adhesion molecule-I (ICAM-I) in the pathogenesis of asthma. Science 247:456-459. 45. Hallam, C., D. I. Pritchard, S. Trigg, and R. P. Eady. 1982. Rat eosinophil-mediated antibody-dependent cellular cytotoxicity: investigations of the mechanisms of target lysis and inhibition by glucocorticoids. Clin. Exp. Immunol. 48:641-648.
