Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Migration of eosinophils into specific tissue sites is a hallmark of allergic diseases such as asthma and allergic rhinitis. Eosinophil recruitment can contribute to perpetuation of the condition by releasing tissue-damaging agents such as major basic protein, eosinophil peroxidase, lipid mediators, and multipotent cytokines (1). Therefore, it is important to have a better understanding of the mechanisms involved in the selective migration of eosinophils from the blood to allergic tissues and the further activation of these cells in the tissue in response to specific stimuli.

The movement of eosinophils from the blood into tissues is controlled by a series of adhesive interactions first between the leukocyte and the endothelium and subsequently with the surrounding connective tissue. CD11b/ CD18 (also known as Mac-1, _M2, and C3bi; a member of the ₂ integrin family of heterodimeric adhesion molecules) and very late antigen-4 (VLA-4) (also called CD49d/CD29 and ₄₁; a member of the ₁ integrin family of adhesion molecules) play important roles in eosinophil extravasation. Neutralization of the CD11b and CD49d subunits, using monoclonal antibodies (mAbs), inhibits human eosinophil adhesion (2, 3), transendothelial migration in vitro (3, 4), and murine eosinophil infiltration into extravascular tissue in response to the chemokine eotaxin (5) or after antigen challenge in vivo (6, 7).

Glucocorticoid hormones are now used as first-line drugs for the treatment of asthma and have wide application in the therapeutic control of other allergic and inflammatory diseases. Inhibition of cytokine-stimulated eosinophil survival is one of the important actions of glucocorticoids (8). Any direct effect(s) they may have on adhesion molecule expression and function remains controversial (9). We have recently demonstrated an effect of the glucocorticoid dexamethasone (DEX) on CD11b expression on the cell surface of human eosinophils (10). Overnight incubation with DEX resulted in reduction of basal CD11b levels as well as inhibition of its upregulation after stimulation with eotaxin. Similar findings have been described previously in interleukin (IL)-3-stimulated upregulation of CD11b expression (11). So far, no studies have addressed the functional importance of these observations in the migration of eosinophils.

In this study we have examined the effect of prolonged exposure of murine eosinophils to DEX on CD11b and CD49d protein as well as on chemotaxis. We have also investigated the possible molecular mechanisms of DEX on integrin expression and function. We propose that the integrins CD49d and CD11b are novel targets for glucocorticoid action in allergic inflammation.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animals

CBA/Ca mice overexpressing the murine IL-5 gene were used for all experiments. The mice were obtained from the National Heart and Lung Institute in the Imperial College of Medicine, London UK (12, 13) and were bred in-house. Mice were maintained on a standard chow pellet diet with tap water ad libitum, using a 12-h light/dark cycle, and were housed in an air-conditioned animal facility at 20 ± 2°C with a relative humidity of 55 ± 10%. Experimental work was performed according to Home Office regulations (Guidance on the Operation of Animals, Scientific Procedures Act 1986).

Isolation of Mouse Blood Eosinophils

The protocol for murine blood eosinophil isolation (> 95% purity) is based on a protocol described previously (12, 13). Briefly, blood was obtained by cardiac puncture and the red blood cells were sedimented with 1.25% Dextran (T500; Pharmacia, Sollentuna, Sweden). The resulting leukocytes from the buffy coat were layered onto discontinous Percoll (Sigma, Poole, UK) gradients (densities of 1.070, 1.075, 1.080, and 1.085 g/ml). The gradients were centrifuged (1,500 × g, 25 min, 20°C) and eosinophils and lymphocytes were collected from the 1.080/1.085 interface. The lymphocytes were removed by a 20-min incubation with 7.5 µg/ ml anti-CD2 mAb (10 µg/ml) together with anti-B220 (CD45R) mAb (7.5 µg/ml). Cells were then washed and incubated with goat antirat immunoglobulin (Ig)G microbeads for 20 min before being run through a MACS BS column (Miltenyi Biotec, Inc., Surrey, UK). The eosinophil-rich effluent was collected, and the populations obtained were over 95% pure eosinophils (using Kimura stain) and greater than 98% viable (using Trypan blue exclusion).

Incubation of Eosinophils with DEX

Preliminary studies and previous publications have shown that eosinophils exhibit poor viability after long-term (> 24 h) culture and that DEX enhances eosinophil cell death (8). Therefore, murine recombinant (mr) IL-5 was added to the culture medium at a concentration that prolonged eosinophil survival (14). Table 1 illustrates the viability of eosinophils when cultured in media with and without mrIL-5 as well as the detrimental effect of DEX on eosinophil survival. Different concentrations of the cytokine were assessed in viability studies (data not shown), and it was found that 1 ng/ml of mrIL-5 (R&D Systems, Abingdon, UK) maintained > 84% viability of eosinophils at 48 h in the presence of 1 µM DEX (David Ball Laboratories, Warwick, UK), as measured by Trypan blue exclusion test. In a few experiments, eosinophil apoptosis was also monitored after 48 h coincubation using fluorescein isothiocyanate (FITC)-annexin V (15 min at room temperature), performing the procedure as recommended by the manufacturer (R&D Systems).

  For the overnight(s) incubation, eosinophils were washed and resuspended in RPMI 1640 (containing 1 ng/ml mrIL-5, 10% fetal calf serum, 1% penicillin/streptomycin, and 1% L-glutamine; Sigma) at a concentration of 6.7 × 10⁵/ml. Cells were plated, in duplicate for each treatment, into 24-well plates (Costar, Cambridge, MA) in a final volume of 1.5 ml. Eosinophils were incubated in the presence or absence of DEX (10⁶ to 10⁸ M) for 24 or 48 h at 37°C with 5% CO₂ at a constant humidity. In a separate set of experiments, the glucocorticoid receptor antagonist RU486 (mifepristone; a generous gift of Roussel Uclaf, Romainville, France) was used to assess glucocorticoid receptor mediation. Eosinophils were incubated for 48 h with 10 µM RU486 (in 0.1% dimethyl sulfoxide [DMSO]) with or without 1 µM DEX. The incubation conditions used were the same as those described earlier.

Eosinophil Activation with Eotaxin

In separate sets of experiments, 48 h after eosinophil incubation with mrIL-5 in the absence or presence of 1 µM DEX, cells were stimulated with eotaxin (0 to 100 ng/ml) for 30 min at 37°C before determination of CD11b expression by flow cytometry as detailed later.

Incubation of Eosinophils with Other Steroids

The effect of the glucocorticoids hydrocortisone and prednisolone in modulating integrin expression was assessed and compared with the sex steroids testosterone and progesterone. Budesonide (generously supplied by Dr. P. Del Soldato, NicOx S.A., Sophia-Antipolis, France) was tested at 0.1 to 10 µM, whereas hydrocortisone 21-hemisuccinate (Sigma), prednisolone 21-hemisuccinate (Sigma), testosterone (Sigma), or progesterone (Sigma) hormones were added to purified eosinophils at a concentration of 10 µM in 0.1% DMSO and incubated for 48 h in the same media and conditions as described earlier. Control eosinophils were incubated with 0.1% DMSO.

In Vivo Treatment with DEX

Mice (n = 4-6) were treated daily with vehicle (100 µl subcutaneously) or DEX (0.3 or 1 mg/kg subcutaneously) for 3 d. Blood was collected by cardiac puncture under halothane anesthesia 24 h after the last injection, eosinophils were prepared as detailed earlier, and cell-surface expression of CD11b and CD49d was measured by flow cytometry as described later.

Cell-Surface and Total CD11b and CD49d Expression by Flow Cytometry

The levels of CD11b and CD49d on the cell surface of eosinophils were measured by flow cytometry. Freshly isolated or cultured eosinophils were plated into 96-well plates (0.2 to 0.3 × 10⁶ per well) together with 20 µl blocking human IgG (15 mg/ml; Sigma) and 20 µl of either rat antimouse CD49d mAb (5 µg/ml, clone CRL1911; a kind gift from Dr. M. Robinson, Celltech Therapeutics, Slough, UK) or CD11b mAb (10 µg/ml, clone 5C6; Serotec, Oxford, UK). After incubation for 45 min at 4°C, the wells were washed with phosphate-buffered saline (PBS) containing 0.2% bovine serum albumin (BSA), and the cells were incubated (30 min, 4°C) with 30 µl of diluted FITC-conjugated rabbit antirat IgG (Serotec). The cells were washed and either analyzed immediately or fixed with an equal volume of 2% paraformaldehyde and analyzed within 1 wk.

  For quantification of total cell CD11b levels, eosinophils were fixed in 2% paraformaldehyde (4°C, 30 min) before immunostaining and maintained in 0.02% saponin (Saponaria Sponges; Sigma) to enable permeabilization (15) throughout the entire experiment.

  In all cases, flow cytometry was performed using a FACScan II analyzer (Becton Dickinson, Oxford, UK) with an air-cooled 100-mW argon ion laser tuned to 488 nm and a Consort 32 computer running Lysis II software. Data were recorded as median fluorescence intensity (MFI) units in the FL1 (green) channel gating only viable eosinophils. For every experiment, data were calculated as the means of the MFI value for each treatment in duplicate.

Eosinophil Chemotaxis Assay

The chemotaxis assay is based on a previously described protocol (16). After incubation of eosinophils in the absence or presence of DEX, the cells were washed and resuspended in RPMI (supplemented with 0.1% BSA, pH 7.4) at a concentration of 4 × 10⁶ eosinophils/ml. The role of CD11b or CD49d in eosinophil chemotaxis was assessed after cell incubation with 10 µg/ml anti-CD11b mAb alone, 10 µg/ml anti-CD49d mAb alone, or with both mAbs together for 15 min at room temperature. Control cells were incubated with an identical concentration of rat IgG (Sigma). The cells were washed and resuspended at 4 × 10⁶ eosinophils/ml and used in the chemotaxis assay. Neuroprobe ChemoTx 96-well disposal chemotaxis plates (5-µm pore size; Receptor Technologies Ltd., Adderbury, UK) were used. Chemoattractant (eotaxin, 1 to 100 ng/ml; platelet-activating factor [PAF], 0.01 to 1 µM) or vehicle (27 µl of RPMI-BSA) was placed in triplicate in the bottom wells, the polycarbonate filter was placed on top, and 25 µl of the cell suspension was placed on top of the filters. The plates were incubated for 2 h in a humidified incubator at 37°C with 5% CO₂. The cells remaining on top of the filters were absorbed off, and the filter tops were carefully washed to ensure removal of all nonmigrated cells. The plates were then centrifuged (312 × g, 5 min) to pellet all cells on the undersides of the filters. The filters were removed, and cells in the bottom wells were counted by light microscopy. Data are reported as migration index (MI), calculated as follows: (number of cells migrating to chemoattractant)/(number of cells migrating to vehicle). In preliminary experiments the chemotactic response to eotaxin was controlled because of the finding that no cell migration above vehicle wells was obtained if the chemotactic gradient was disrupted (i.e., 10 ng/ml eotaxin present both in the upper and bottom well) (data not shown).

Quantitation of Blood Eosinophil Counts

Blood was collected into heparinized syringes by cardiac puncture under halothane anesthesia. Total leukocyte counts were performed using Turk's stain, and differential cell counts were made on blood smears after staining with May-Grumwald-Giemsa. Under these conditions, eosinophils were well distinguished from neutrophils and mononuclear cells. In mrIL-5 transgenic mice, eosinophils constituted 56.8 ± 2.2% of the total blood leukocyte count (n = 12).

Detection of CD11b Messenger RNA by Reverse Transcriptase/Polymerase Chain Reaction Analysis

The cells were washed using sterile PBS, and sterile tubes, pipettes, and gloves were used at all times. Total RNA was extracted using a commercially available kit purchased from Promega (RNAgents Total RNA Isolation System, Madison, WI). Per the manufacturer's instructions, eosinophils were lysed in an appropriate volume of denaturing solution consisting of guanidine thiocyanate and a buffer containing citrate, sarcosine, and -mercaptoethanol (Promega). The pH was brought down with 2 M sodium acetate (pH 4), and contaminating DNA and proteins were removed using the suggested volumes of phenol/chloroform/isoamyl alcohol mixture. The mixture was centrifuged at 10,000 × g for 10 min in a refrigerated minicentrifuge, and the aqueous RNA phase was collected. The RNA was precipitated with isopropanol for 30 min at 20°C and centrifuged, after which the RNA pellet was washed with 75% ethanol (in diethylpyrocarbanate [DEPC]-treated water). The pellet was air-dried and resuspended in DEPC-treated water. The yield and purity of the RNA were estimated spectrophotometrically at 260- and 280-nm wavelengths (Pharmacia LKB Ultrospex III spectrophotometer) and an absorbance_260-280 ratio of 1.6 to 2.0 was desired.

For each sample, 1 µg of total RNA was converted to first-strand complementary DNA (cDNA) using Moloney murine leukemia virus reverse transcriptase (RT) (Pharmacia Biosystems Europe, St. Albans, UK) and oligo(dT) primers (Pharmacia Biosystems). Polymerase chain reaction (PCR) amplification reactions were performed on aliquots of the cDNA. The primers for murine CD11b were 5'-CAG-ATC-AAC-AAT-GTG-ACC-GTA-TGG-3' and 5'-CAT-CAT-GTC-CTT-GTA-CTG-CCG-C- 3' (forward and reverse), which amplified a fragment of 497 base pairs (bp) in length (OligoExpress Ltd., Middlesex, UK). Mouse glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a housekeeping gene: 5'-ACC-ACA-GTC-CAT-GCC-ATC-AC-3' and 5'-TCC-ACC-ACC-CTG-TTG-CTG-TA-3' (forward and reverse) (OligoExpress Ltd.). All PCR reactions were performed with one PCR bead (Pharmacia Biosystems) per tube in a final volume of 25 µl, containing 1.5 U of Taq polymerase, 10 mM Tris-HCl (pH 9 at room temperature), 50 mM KCl, 1.5 mM MgCl₂, 0.2 mM of a mixed preparation of deoxynucleotides (Pharmacia Biosystems), and 10 µM primers and stabilizers, including BSA. The PCR cycling profile for CD11b was: an initial denaturation for 2 min at 94°C; 35 cycles of denaturation at 94°C (45 s), annealing at 56°C (45 s), and extension at 72°C (1 min); followed by a final holding cycle at 72°C (7 min). For GAPDH, similar conditions were used, apart from the annealing temperature (60°C). In selected experiments PCR products formed after 20, 25, 30, or 35 cycles were determined. Amplification products were visualized using ethidium bromide fluorescence in agarose gels. Band intensities were transformed into arbitrary units using densitometry using XX software running on a Apple Macintosh Power PC computer.

Statistics

Data are expressed as means ± standard error of the mean (SEM) of n mice per group. When comparing either unpaired or paired sets of data, unpaired or paired Student's t test (two-tailed) was used, respectively. Comparisons between multiple treatment groups from the same experiment were made using repeated-measures analysis of variance for paired comparisons. When significant differences were found between groups, Dunnett's test was used to test significance. In all cases a P value < 0.05 was accepted as significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Effects of mrIL-5 and DEX on Eosinophil Survival

To maintain viability of the eosinophils during prolonged incubations (i.e., > 24 h) with DEX the cells required the presence of mrIL-5. This was confirmed by Trypan blue exclusion assay (Table 1). No changes in the degree of cell apoptosis were seen in the 48-h incubation protocol with 2 ± 0.5% of FITC-annexin V-positive cells exposed to mrIL-5 alone and 2.5 ± 0.4% when 1 µM DEX was added (n = 3, not significant).

Incubation of eosinophils with mrIL-5 for 24 h resulted in upregulation of cell-surface CD11b protein levels (~ 63% increase, from 334 ± 53 MFI units without mrIL-5 to 543 ± 55 MFI units in the presence of 1 ng/ml mrIL-5, n = 6-8 mice per group). It was not possible to determine CD11b levels on eosinophils cultured in the absence of mrIL-5 for 48 h because of poor cell viability. However, the levels of CD11b remained elevated at this time point when cultured with mrIL-5, though lower than at 24 h after incubation (379 ± 104 MFI units, n = 6). Although the effect of DEX was mainly investigated in the presence of mrIL-5, DEX was able to reduce basal levels of the integrin in eosinophils cultured in the absence of cytokine (220 ± 42 MFI units at the 24-h time point, n = 6; P < 0.05 versus the control value reported earlier). In all further experiments eosinophils were cultured in the presence of 1 ng/ml mrIL-5, and all data are expressed as paired observations.

Effect of DEX on Surface Integrin Levels

The time course of DEX effects on cell-surface integrin levels was established. Initial experiments were carried out at 24 h; however, a more consistent inhibitory effect of the steroid was seen after a 48 h incubation. A representative fluorescence histogram appears in Figure 1 depicting the reduction in both CD11b (Figure 1A) and CD49d (Figure 1B) after 48 h culture with DEX. After 48 h incubation, exposure of eosinophils to 1 µM DEX resulted in a significant reduction in both CD11b (36 ± 4% inhibition) and CD49d (37 ± 6% inhibition) cell-surface levels (Figure 1C). For this set of experiments the MFI units for CD11b and CD49d in untreated eosinophils were 445 ± 142 and 199 ± 10, respectively.

View larger version (25K): [in this window] [in a new window]   Figure 1.   Concentration-related inhibition of cell-surface CD11b and CD49d levels by 48 h treatment with DEX in vitro. Histograms represent flow cytometric analysis of CD11b (A) and CD49d (B) expression on viable gated eosinophils cultured with 1 µM DEX or vehicle (0.1% DMSO). Control histograms for vehicle (grey) and DEX (black) are superimposed. (C) Cumulative data for DEX concentration responses after 48 h culture with 1 ng/ml mrIL-5. Eosinophil CD11b expression in vehicle (0.1% DMSO)-treated cells was 407.4 MFI values. Data are expressed as means ± SEM of n = 5-7 per group. *P < 0.05 versus control as calculated on original values.

In a different set of experiments, DEX induced a 20% inhibition of CD11b surface expression, with a similar reduction in CD49d expression (Figure 2A). Incubation with RU486 (10 µM) 1 h before the addition of DEX abrogated this inhibitory effect, verifying that the action of the glucocorticoid is mediated through the glucocorticoid receptor.

View larger version (39K): [in this window] [in a new window]   Figure 2.   Activation of the glucocorticoid receptor is required to inhibit integrin expression. (A) RU486 (10 µM) or 0.1% DMSO was added to eosinophils 1 h before vehicle or DEX (1 µM) and incubated further for 48 h. Eosinophils were then harvested and surface levels of CD11b and CD49d measured using flow cytometry. Data are means ± SEM percent of n = 6 mice per group. *P < 0.05 versus vehicle-treated cells as calculated on original values. (B) Eosinophils were incubated for 48 h with the reported steroids (10 µM in all cases) and surface levels of CD11b and CD49d measured using flow cytometry. Data are means ± SEM of n = 5 mice per group. *P < 0.05 versus untreated cells. (C) As in B, but cells were incubated with the reported concentrations of budesonide. Data are means ± SEM of n = 5 mice per group. *P < 0.05 versus untreated cells.

Effects of Other Steroid Hormones on Surface Integrin Levels

There was no significant difference in CD11b or CD49d expression between control eosinophils and cells treated with either testosterone or progesterone (Figure 2B). However, incubation of eosinophils with hydrocortisone resulted in a significant reduction of CD11b (35% inhibition) but not CD49d surface levels. Prednisolone (10 µM) treatment produced a significant reduction in both CD11b and CD49d eosinophil surface expression by 32.6 ± 5.8 and 35.6 ± 7.7%, respectively (Figure 2B). A higher degree of activity was measured with budesonide. This steroid was effective in reducing CD11b expression with a significant inhibition at 0.1 and 1 µM (Figure 2C).

Effect of DEX on Stimulated Expression of CD11b

Eosinophil incubation (48 h) with 1 µM DEX also reduced CD11b expression after upregulation of this adhesion molecule with eotaxin. A significant inhibition was measured after cell incubation with 10 and 30 ng/ml eotaxin (Figure 3).

View larger version (19K): [in this window] [in a new window]   Figure 3.   DEX inhibits eotaxin-induced CD11b upregulation. Eosinophils were incubated for 48 h with vehicle (0.1% DMSO) or with 1 µM DEX before 30 min stimulation with eotaxin (0 to 100 ng/ ml). Surface CD11 levels were measured by flow cytometry. Data are means ± SEM of n = 4 mice per group. *P < 0.05 versus untreated cells.

Involvement of CD11b and CD49d in Eosinophil Chemotaxis

The functional relevance of the reduction of cell-surface integrin levels by DEX was determined by assessing eosinophil migration in vitro. First, the relevance of CD11b and CD49d in mediating eosinophil chemotaxis in our experimental system was addressed using neutralizing mAbs directed against the CD11b or CD49d integrin subunits (Figure 4). With 1 ng/ ml eotaxin as the chemoattractant, eosinophil migration was inhibited by 55 ± 10 and 55 ± 6% when cells were incubated with either anti-CD11b or -CD49d mAbs, respectively (P < 0.05 in both cases) (Figure 4A). Interestingly, when the two mAbs were added together, no further inhibition of eosinophil chemotaxis was observed.

View larger version (25K): [in this window] [in a new window]   Figure 4.   Role for CD11b and CD49d in mediating eosinophil migration to eotaxin and PAF. Freshly isolated eosinophils were incubated with either anti-CD11b mAb alone, anti-CD49d mAb alone, or both mAbs together for 15 min and then used in the chemotaxis assay. Control cells were incubated with rat IgG. The chemoattractants used were 10 ng/ml mr eotaxin (A) and 0.1 µM PAF (B). Data are means ± SEM of the MI of 7 to 13 mice per group. Dashed line indicates basal cell migration. *P < 0.05 and **P < 0.01 versus control group using paired t test.

To confirm the importance of these two integrins in mediating eosinophil migration in vitro, another class of chemoattractant was tested. Eosinophil chemotaxis to 0.1 µM PAF was inhibited by 77 ± 9 and 76 ± 10% when mAbs to CD11b or CD49d were used, respectively (Figure 4B). Once again, blocking both integrins did not increase the inhibition of migration (n = 2 mice, data not shown).

Functional Relevance of the Effect of DEX

The effect of DEX on eosinophil migration was then investigated. A concentration-dependent eotaxin-mediated increase in migrated eosinophils was observed. Eosinophils incubated for 48 h with 1 µM DEX exhibited a significant reduction in chemotaxis to eotaxin. The percent inhibitions of the migration to 100 and 10 ng/ml eotaxin were 40 ± 7 and 56 ± 6, respectively (Figure 5A). An extent of inhibition similar to that of PAF was achieved, with percent inhibitions of 53 ± 10 and 40 ± 10 for concentrations of 0.1 and 0.01 µM PAF, respectively (n = 6; data not shown). It was noted that incubation with DEX did not alter basal eosinophil migration, i.e., migration to vehicle alone (values are means ± SEM; × 10³ cells): 20.4 ± 3.6 cells for untreated eosinophils, and 23.4 ± 5.2 cells after treatment with 1 µM DEX (n = 9; not significant; paired t test).

View larger version (40K): [in this window] [in a new window]   Figure 5.   Inhibition of eosinophil chemotaxis by DEX, and reversal by RU486. Eosinophils were cultured with 1 ng/ml mrIL-5 alone or together with 0.01 to 1 µM DEX for 48 h. The cells were harvested and prepared for use in the cell chemotaxis assay. Different concentrations of eotaxin were used as chemoattractant on cells cultured with vehicle and 1 µM DEX (A). Different concentrations of DEX were used against 10 ng/ml eotaxin as chemoattractant (B). In some cases, cells were incubated with 1 µM DEX plus 10 µM RU486. Data are means ± SEM of the migration index of 7 to 13 mice per group. Dashed line indicates basal cell migration. *P < 0.05 and **P < 0.01 versus control group (paired t test) at the same concentration of chemoattractant.

Culture of eosinophils with DEX (0.01 to 1 µM) for 48 h induced a significant inhibition of eosinophil chemotaxis to 1 ng/ml eotaxin (Figure 5B). A significant inhibition was obtained with 0.1 µM DEX, a concentration even more effective than 1 µM DEX, which was the only concentration found to be effective in modulating integrin cell surface levels (see Figure 1). A similar reduction in chemotaxis was observed with PAF (data not shown, n = 6).

Treatment of eosinophils with RU486 (10 µM) alone did not change the characteristics of cell chemotaxis (n = 4, data not shown). In these experiments, incubation with 1 µM DEX produced around 30% reduction in the number of eosinophils recovered from the bottom wells. However, when eosinophils were coincubated with RU486 and DEX, the reduction in cell chemotaxis was abolished (Figure 5B).

Effect of DEX after In Vivo Treatment

The in vivo relevance of DEX effect on eosinophil integrin expression was assessed after a subchronic treatment protocol. A 3-d treatment with DEX significantly reduced CD11b and CD49d expression on circulating eosinophils only at the dose of 1 mg/kg, but not of 0.33 mg/kg (Figure 6). The degree of inhibition of both CD11b and CD49d expression was again in the range of 30 to 40%, comparative to that measured in the in vitro experiments.

View larger version (48K): [in this window] [in a new window]   Figure 6.   Effect of in vivo DEX on eosinophil integrin expression. Mice were treated daily with vehicle (100 µl) or with the reported doses of DEX administered subcutaneously for 3 consecutive days. Circulating eosinophils were prepared 24 h after the last drug administration and tested for CD11b and CD49d expression by flow cytometry. Data are means ± SEM of n = 4-6 mice per group. *P < 0.05 versus control group (unpaired t test).

As a positive control, we monitored the effects of DEX on circulating white blood cells. Three-day treatment of mice with DEX reduced circulating eosinophils from 24.3 ± 4.5 × 10⁶ cells/ml of blood to 6.3 ± 1.1 × 10⁶ cells/ml of blood (n = 5; P < 0.05). The values for neutrophils and mononuclear cells were (× 10⁶ cells/ml blood): from 1.82 ± 0.63 to 3.05 ± 0.67 (for neutrophils, n = 5; P < 0.05), and from 26.7 ± 3.5 to 17.4 ± 1.2 (for mononuclear cells, n = 5; P < 0.05) in the PBS- and DEX-treated groups of mice, respectively.

Indications for the Molecular Mechanism of DEX

Inconsistent results were obtained when CD11b messenger RNA (mRNA) was determined in cultured eosinophils (data not shown). Hence, to assess whether DEX could modulate the expression of the CD11b gene, RT-PCR was performed on peripheral blood eosinophils of mice treated for 3 d with 1 mg/kg DEX. Results from three representative animals for each treatment are shown in Figure 7A. Total RNA was extracted from purified eosinophils, and CD11b mRNA was expressed in cells taken from control animals with a PCR product size of 497 bp. This basal expression was not altered after the 3-d systemic treatment with DEX (Figure 7A). Treatment of animals with DEX did not alter CD11b mRNA also seen when PCR analysis was conducted with different numbers of cycles (Figure 7B).

View larger version (31K): [in this window] [in a new window]   Figure 7.   Comparison of expression of CD11b mRNA in eosinophils pretreated with saline and DEX. Total RNA was extracted from circulating eosinophils of mice treated for 3 d with 1 mg/kg DEX and saline subcutaneously. A total of 1 µg of RNA was converted to cDNA, and GAPDH and CD11b (A) PCR products were expressed on agarose gels and visualized using ethidium bromide. Data are from three separate mice per group. (B) Ratio of CD11b/GAPDH products as measured with different cycle numbers (n = 3).

To investigate whether DEX had an effect on the CD11b cellular distribution, we performed a series of experiments in which both intracellular and cell-surface expression of this adhesion molecule were measured. Table 2 illustrates this data and reports that incubation with 1 µM DEX reduced only cell-surface CD11b levels (35% inhibition, n = 9; P < 0.05), with no modification on intracellular CD11b protein levels. A similar effect was seen on CD49d levels, with a good degree of inhibition on the cell-surface expression in this set of experiments as well, with no significant alterations in intracellular protein expression (Table 2).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In the present study we propose that glucocorticoid hormone inhibition of cell-surface integrin expression on eosinophils may be another mechanism by which these potent anti-inflammatory drugs regulate tissue eosinophilia. Such an effect seems to be the result of an interference with the integrin externalization process, and it is functionally linked to a reduced eosinophil migratory response.

Glucocorticoid hormones attenuate tissue accumulation of eosinophils in allergic reactions by a combination of mechanisms. First, steroids potently reduce the number of circulating eosinophils by interfering with their release from the bone marrow (17); second, this class of drugs inhibits the expression of proinflammatory cytokines involved in the eosinophil trafficking process (18, 19). These two mechanisms of action may also be linked together, as in the case of the cytokine IL-5. IL-5 is known to be the most potent cytokine for the maturation and mobilization of eosinophils from the bone marrow (20), and nanomolar concentrations of glucocorticoids block the synthesis of this cytokine at the level of gene transcription (21). However, neither mechanism produces a direct effect on the eosinophil itself. To date, the major direct effect of glucocorticoid hormones on this type of leukocyte is a drastic reduction in cell survival associated with an increased rate of apoptosis (8). Such a proapoptotic effect can clearly be responsible for the known ability of glucocorticoid hormones to suppress tissue eosinophil numbers.

Attempts to investigate potential direct effects of DEX on eosinophil integrin expression and consequent eosinophil adhesion have been partially unsuccessful (22). Eosinophil incubation with 0.1 µM DEX modestly modified the extent of CD11b expression on the cell surface of human eosinophils cultured with granulocyte macrophage colony-stimulating factor (GM-CSF). We speculated that the reason for this inefficacy could reside in the long half-life of this specific protein and/or in the concentration of the steroid used. In fact, 24-h incubations of human eosinophils with 1 µM DEX reduced both basal and eotaxin-stimulated CD11b expression on the cell surface (10). The present study was undertaken to investigate this finding further by analyzing the molecular mechanisms responsible for this effect using murine eosinophils.

In initial experiments, vital precautions were taken to ensure that culturing conditions would not affect eosinophil survival, inasmuch as DEX-induced eosinophil apoptosis would hamper the assessment of integrin levels on these cells. In addition, the eosinophils obtained from these IL-5 transgenic mice have been exposed to high serum levels of the cytokine (23). Hence, mrIL-5 was added to eosinophils as a survival aid in the majority of the experiments. As expected (24), addition of mrIL-5 to the culture medium maintained eosinophil survival to an acceptable level up to 48 h (longer time points were not investigated). More importantly, coaddition of mrIL-5 greatly attenuated the toxic effect of DEX as measured both in terms of cell viability and apoptosis. Overall, these were found to be the best experimental conditions for investigating the potential effect of DEX on eosinophil integrin expression.

We chose to examine the effect of cell incubation DEX on VLA-4 (by monitoring its  subunit, also termed CD49d) and CD11b/CD18 (by monitoring its  subunit, CD11b) because of (1) the pivotal role that these integrins play in the eosinophil recruitment process in vitro and in vivo (25- 27), and (2) our previous study with human eosinophils in which DEX reduced cell-surface expression of CD11b (10). In the present study, eosinophil incubation with DEX produced a dose-dependent reduction in cell-surface levels of both CD49d and CD11b. A similar degree of inhibition (between 20 and 40%) was achieved by the glucocorticoid for both integrin subunits, and a significant effect was seen only at the highest concentration tested, 1 µM. The DEX inhibition of CD49d expression appears to be a peculiarity of mouse eosinophil, inasmuch as our previous study failed to report such an effect with human eosinophils (10). Nonetheless, the length of incubation may have a role to play here (24 versus 48 h) and may possibly be related to the rate of protein turnover on the eosinophil plasma membrane (see later discussion). However, in line with our study on human eosinophils (10), the steroid also reduced CD11b expression after upregulation with eotaxin.

The specificity of the DEX effect on integrin expression was controlled in two distinct ways. First, an involvement of the glucocorticoid receptor was demonstrated by using the receptor antagonist RU486 (mifepristone) (28). Eosinophils have long been reported to express the glucocorticoid receptor (29). In the presence of ligand, RU486 stabilizes the association of the glucocorticoid receptor with heat shock protein 90 (30), with consequent block of the transcription of genes containing glucocorticoid response elements, thus preventing the agonist from producing its effects. We demonstrated here that RU486 reverses the inhibitory effect of DEX on cell-surface integrin expression. Second, the inhibitory action of DEX was mimicked by two distinct glucocorticoid molecules, hydrocortisone and prednisolone. Both agents suppressed integrin expression on the eosinophil cell surface (with the yet-unclear exception of hydrocortisone against CD49d) when added at 10 µM. In contrast, eosinophil incubation with an equimolar concentration of progesterone and testosterone failed to affect the extent of either CD11b or CD49d expression. Importantly, the potent glucocorticoid budesonide, largely used in the clinic, reduced CD11b expression on mouse eosinophils with optimal concentrations of 0.1 and 1 µM. Overall, these data are congruent with the involvement of the glucocorticoid receptor in the actions of steroid hormones in inhibiting integrin levels on eosinophils.

Next, we tried to address the mechanisms responsible for this reduction in integrin expression. As reported by other authors (31), measurement of mRNA levels in eosinophils is a difficult task because the culture conditions somehow interfere with the purity and stability of the extracted RNA. In preliminary in vitro experiments we were unable to obtain a reliable result on the effect of DEX on CD11b mRNA as measured by RT-PCR (data not shown). However, this technical problem was circumvented by measuring CD11b mRNA in eosinophils extracted from animals treated with DEX in vivo. In this case, we could detect the correct PCR product with reproducibility, seeing no consistent difference in the bands after 3-d treatment with 1 mg/kg DEX.

Protein expression was then investigated as a possible molecular mechanism for glucocorticoid action. Total CD11b protein levels were measured after a validated permeabilization protocol (15). In vitro treatment with DEX for 48 h failed to reduce the intracellular level of this protein. Thus, the effect of DEX on surface integrin levels is not due to a reduction in expression of mRNA or protein. Slightly less than 10% of CD11b or CD49d was expressed on the plasma membrane, and it is this fraction that was selectively targeted by cell incubation with DEX. This observation would suggest that the glucocorticoid may be interfering with the process of externalization for CD11b and also for CD49d. However, it is unclear which specific subcellular organelles contain these integrins, whether they are different, and whether they follow a similar pattern of externalization on the plasma membrane. Despite the absence of a specific ultrastructural study, this is very likely to be the case, not only because of the analogy with the neutrophil (32) but also because eosinophils are able to rapidly upregulate CD11b levels after activation with eotaxin (33, and this study) and other chemoattractant agents, including PAF (22). Consequently, our current working hypothesis for the molecular mechanism underlying this set of observations is an inhibitory effect of the glucocorticoid on the externalization process or exocytosis for both CD11b and CD49d, which in basal conditions occurs in blood leukocytes (34).

An important aspect of the study was to establish functional relevance for this DEX-induced inhibition of CD11b and CD49d cell-surface levels. We therefore developed a chemotaxis assay using blood-derived mouse eosinophils. Both of these integrins are believed to be important in eosinophil chemotaxis, as demonstrated by cell incubation with specific mAbs. CD11b and CD49d integrins appeared to produce eosinophil chemotaxis through a similar mechanism of action, with a similar level of inhibition. Interestingly, cell coincubation with both antibodies did not result in an additive inhibitory effect. The reason for this mutually exclusive phenomenon between CD11b- and CD49d-mediated cell migration is obscure and requires further investigation. Because engagement of cell-surface integrins produce an outside/inside signal that leads to subsequent cell activation (35), it is possible that in the present experimental conditions both CD11b and CD49d activate the same signaling pathway operating in the migrating eosinophil.

Eosinophil incubation with DEX resulted in a reduced cell migration when chemotaxis was stimulated with either eotaxin or PAF. Interestingly, this inhibitory effect was seen not only at the 1 µM concentration (which inhibited cell-surface integrin expression and eosinophil chemotaxis to a similar degree), but also when 0.01 and 0.1 µM DEX were used. The inhibitory effect produced by eosinophil incubation for 48 h with the latter concentrations of DEX is clearly unrelated to an effect on CD11b and CD49d expression. Therefore, eosinophil chemotaxisand by extension, eosinophil migration to an inflamed tissueis more sensitive to glucocorticoid effect than eosinophil integrin levels. These data show that eosinophil migration is clearly a sensitive target for DEX and likely other glucocorticoid hormones: because we used direct-acting chemoattractants such as eotaxin and PAF, the steroid is likely to be interfering with the eosinophil activation process required to achieve migration. Potential targets for DEX effect at low concentrations may be the receptor for urokinase-type plasminogen activator (uPAR) and/or cytoskeletal proteins (36). A recent study has demonstrated that integrin activation and leukocyte adhesion require uPAR-mediated signaling (37). Interestingly, uPAR expression is sensitive to the inhibitory action of glucocorticoid, at least on cells of the myelomonocytic lineage (38), and this receptor is also expressed on human eosinophils (39). Similarly, the integrin CD11d/CD18 (31), recently described on human eosinophils, may be the target for lower concentrations of glucocorticoid hormones. Finally, we cannot exclude the possibility that the steroid might reduce the expression of the receptors specific for eotaxin and PAF, or of signaling molecules (e.g., G protein) associated with chemoattractant receptors. These are clearly speculative hypotheses at the moment, and they warrant future investigation.

Finally, an in vivo emphasis was given to our observations. Subchronic treatment of mice with DEX produced a dose-dependent inhibition of both CD11b and CD49d levels on circulating eosinophils. This inhibitory effect was again in the range of that seen after in vitro cell incubation with the glucocorticoid, i.e., around 30 to 40%. A lower dose of DEX was not sufficient to cause a reduction in surface integrin expression. It is worth noticing that (1) a single dose of DEX did not alter integrin expression on circulating eosinophils (data not shown), and (2) the effective dose of 1 mg/kg subcutaneously is able to reduce eosinophil trafficking in experimental models of allergic inflammation (40). In this context we cannot exclude the possibility that the reduction in adhesion molecule expression seen after in vivo treatment with DEX could be secondary to an inhibitory effect on certain cytokines (41), but whatever the mechanism(s), we show that glucocorticoid inhibition of eosinophil integrin expression is a phenomenon that can be observed in vivo.

Besides a series of studies that have reported glucocorticoid inhibitory effects on endothelial adhesion molecules, it is now clear that these drugs can also affect the extent of adhesion molecule expression on blood leukocytes. This effect is not only species-specific but also cell-specific. For instance, repeated but not single administrations of DEX to rats reduced basal intercellular adhesion molecule-1 levels on circulating monocytes but not basal or upregulated CD11b levels on neutrophils (42). In addition, DEX has been shown to reduce basal CD11a, but not CD11b, expression on rat neutrophils and basal CD18 expression on bovine neutrophils (43), and stimulated downregulation of CD62L again on rat neutrophils (44). Human neutrophils downregulate anesthesia-induced CD11b expression with systemic treatment with glucocorticoids (45). In preliminary experiments performed with asthmatic and atopic volunteers (n = 2) under steroid treatment, cell-surface CD11b expression was negligible when compared with that in healthy volunteers, and these low levels were not further reduced by cell incubation with DEX (L. H. K. Lim and A. M. Das, unpublished data).

In conclusion, with this study we demonstrate that mouse eosinophils' migratory capacity is greatly affected by DEX, and that this is partly related to a reduced cell-surface translocation of integrins on the leukocyte plasma membrane. This and the data obtained after treatment with DEX in vivo prompt the proposition that DEX inhibition of eosinophil integrin expression may have clinical relevance and may contribute to the overall potent action that glucocorticoid hormones exert on eosinophil trafficking as seen in asthma and other allergic pathologies.