Dexamethasone reduces cell surface levels of CD11b on human eosinophils

Overnight incubation of human eosinophils (Eøs) with the glucocorticoid hormone dexamethasone (DEX; 0.1 μM) resulted in lower expression of the CD11b, but not CD49d, antigen on their plasma membrane, as assessed by flow cytometry. DEX produced a consistent inhibitory effect (ranging from 16% to 20%) when tested at a concentration of 0.1 μM. Eø stimulation with 100 ng/ml eotaxin produced an increase in CD11b (+26%), but not CD11c, levels and concomitantly a reduction (–25%) on CD62L expression. The inhibition exerted by DEX upon CD11b levels was also evident following eotaxin upregulation, with a degree of inhibition similar to that seen on basal levels. These data highlight a novel mechanism of action by which glucocorticoid hormones may be effective in reducing Eø accumulation during allergic inflammation in man.


Introduction
In ltration of eosinophils (Eøs) into sites of allergic in ammation is a characteristic feature of diseases such as asthma. 1 The products of activated Eøs have been shown to play a role in manifesting the symptoms of asthma. 2 Although many in ammatory mediators (including lipid mediators, cytokines and chemokines) are potent activators of Eøs, very few act speci cally on this cell type. Eotaxin, which is a member of the CC family of chemokines, was isolated from the bronchoalveolar lavage after allergen challenge in sensitized guinea pigs. 3 It is a potent, and more importantly speci c, activator of Eøs.
Activation of leukocytes by in ammatory mediators results in shedding of the selectin adhesion molecule CD62L from the cell surface and upregulation of the b 2 integrin CD11b/ CD18. 4,5 This complex plays an important role in adhesive and activation processes of Eøs 6 -8 and eotaxin increases levels of the CD11b subunit (Mac-1) on the cell surface of human Eøs. 9 Glucocorticoid hormones are used routinely to treat allergic disorders, and indeed treatment with natural or synthetic glucocorticoids is effective in reducing the symptoms of these diseases, including asthma. 10 Although little is known about their mechanism of action, it appears that part of their anti-in ammatory actions may be mediated indirectly by reducing the number of circulating Eøs thereby attenuating Eø in ltration into allergic tissues. 11,12 More recently glucocorticoid hormones, in particular dexamathasone (DEX), have been shown to reduce adhesion molecule expression on selected target cells. For example, DEX was active on endothelial cells where levels of CD62E and CD54 were increased with cytokines or lipopolysaccharide. 13,14 Such an effect of DEX on the activated endothelium could contribute to the inhibition exerted by this drug on leukocyte accumulation during the host in ammatory response. 13 Much less is known about the effect of DEX upon adhesion molecules expressed by circulating leukocytes. A study with bovine neutrophils reported that administration of DEX in vivo reduced basal expression of CD62L and CD18 levels on this polymorphonuclear leukocyte. 15 To date, no studies have investigated an effect of glucocorticoids on the adhesion molecules expressed on Eø plasma membrane. In the present study we have tested whether DEX affected either basal or eotaxin-upregulated CD11b levels on human Eøs.

Material and Methods
Isolation of human blood Eù s Blood (80 ml) was obtained, out of the hayfever season, from individuals who suffered from allergic rhinitis, and it was collected into sodium citrate. After dextran sedimentation ( nal concentration 1.2% dextran; Pharmacia, Uppsala, Sweden) the buffy coat was layered onto histopaque gradients (density 1.077; Sigma, Poole, UK). After centrifugation at 400 3 g for 25 min the leukocyte pellet was collected and all procedures were carried out at 48 C from this point. The remaining red blood cells in the granulocyte pellet were lysed hypotonically. Eøs were separated from neutrophils by incubation with anti-CD16 MACS microbeads (Miltenyi Biotec GmbH, Germany) at a concentration of 40 ml microbeads per 1 3 10 8 granulocytes followed by elution of the Eøs on a CS-type MACS column and magnetic separator (Milltenyi Biotec GmbH, Germany). The Eøs obtained were greater than 96% pure with contaminating cells constituting of mononuclear cells. Cells were either incubated overnight with DEX or stimulated with eotaxin immediately (see below).

Overnight incubation of Eù s with DEX
Cells were resuspended at a concentration of 5 3 10 6 Eøs per ml in RPMI 1640 medium containing 10% heat-inactivated fetal bovine serum, 1% penicillin-streptomycin, 0.5% gentamycin and 2 mM L-glutamine (all obtained from Sigma, Poole, UK). One hundred ml of the cell suspension was added to at-bottomed 96-well plates (in duplicate for each treatment). Either 100 ml buffer or different concentrations of DEX (1 ±0.001 mM nal concentration; David Ball Laboratories, Warwick, UK) was added to the wells and the cells incubated at 378 C with 5% CO 2 in air for 18 h. After the incubation, the cells were washed and either (i) resuspended in phosphate buffered saline (PBS) containing 0.2%bovine serum albumin (BSA) for the measurement of cell surface adhesion molecules or, (ii) resuspended in RPMI 1640 containing 0.1% BSA for stimulation with mediators.

Stimulation with mediators
Either fresh cells or cells which had been incubated overnight with DEX or buffer were incubated in a shaking water bath at 378 C for 30 min in the presence of buffer, human recombinant (hr) eotaxin (100 ng/ ml nal concentration; PeproTech EC Ltd, London, UK) or 0.1 mM nal concentration platelet activating factor (PAF; C16 form, Sigma, Poole, UK). The reaction was stopped by placing the samples on ice for 5 min. The cells were then pelleted and resuspended in PBS containing 0.2% BSA for the measurement of cell surface adhesion molecules.

Measurement of cell surface adhesion molecules
The presence of CD11b, CD11c, CD62L and CD49d (a 4 subunit of the b 1 heterodimer CD49d/ CD29) were measured on the cell surface using FACS analysis as previously described. 16 For this purpose, the following monoclonal antibodies (mAb) were used: mouse-anti human CD11b (clone 44; Serotec, Oxford, UK); mouse-anti human CD11c (clone 3.9; Serotec); mouse anti-human CD62L (clone FMC46; Serotec), and mouse anti-human CD49d (MAX68P; kindly gifted by Dr M. Robinson, Celltech Therapeutics, Slough, UK). MAX68P is a well characterized monoclonal antibody which has been shown to bind to Chinese hamster ovary cells transfected with CD49d (personal communication, Dr M. Robinson, Celltech Therapeutics). Brie y, cells were placed in at-bottomed 96-well plates (2.5 3 10 5 Eøs per well, 20 ml). Non-speci c binding sites were blocked with 20 ml human IgG (15 mg/ ml) before the addition of 20 ml of the appropriate mAb (giving a nal concentration of 10 ±20 mg/ ml). Control wells were incubated with similar amounts of mouse IgG (Sigma). Plates were incubated at 48 C for 1 h prior to washing and a further incubation with 40 ml of FITC-conjugated sheep anti-mouse IgG antibody (1:50 dilution of stock; Sigma, Poole, UK) for 30 min at 48 C. The plates were washed again and FACS analysis performed using a FACScan II analyser (Becton Dickinson, CA, USA) with aircooled 100 mW argon ion laser tuned to 488 nm and Consort 32 computer running Lysis II software. The Eø population could be clearly identi ed because of the side and forward scatter characteristics (con rming in this way the high degree of purity) and the value of the mean uorescence intensity (MFI) associated with it was measured at 548 nm with a band width of 10 nm (FL-1 channel). Results are reported as the MFI units for each sample and have been corrected for the control (mouse IgG) value.

Statistics
Data are mean SEM of n experiments performed in duplicate. Statistical analysis was performed on raw data using the Student's paired t-test.

Results
Effect of eotaxin on cell surface adhesion molecule levels When human Eøs were incubated with eotaxin for 30 min, there was a signi cant increase (21%) in the levels of CD11b molecules on the cell surface with a concomitant 32% reduction in the CD62L levels (Fig. 1). In comparison, we found no effect on the levels of CD11c. We also examined the effect of another mediator, PAF, on the levels of CD11b and CD62L. From a typical experiment where the basal levels of CD11b was 220 MFI units, after eotaxin or PAF exposure this increased to 277 (26% increase) and 303 (38% increase), respectively. Similarly, basal levels of CD62L was 131 MFI, and this was reduced to 117 and 74 MFI units after eotaxin and PAF stimulation, respectively.

Effect of DEX on basal and stimulated CD11b levels
Next, we investigated the effect of DEX on the levels of CD11b. For this series of experiments, Eøs were incubated for 18 h with or without the glucocorticoid hormone after which either (i) cell surface CD11b levels were measured or, (ii) a 30 min stimulation with eotaxin was performed prior to quanti cation of cell surface levels of CD11b. Incubation of Eøs overnight (in the absence of DEX) did not affect CD11b expression when compared to levels of the integrin on freshly isolated Eøs. Further, overnight incubation did not alter the extent of eotaxin-stimulated CD11b increase when compared to fresh Eøs (data not shown).
Following incubation with DEX there was a signi cant reduction on basal CD11b expression on Eø plasma membrane (Fig. 2). A similar reduction was also measured after Eø stimulation with eotaxin. The effect of DEX on basal CD11b levels was investigated further, nding a similar inhibition at all doses tested (a typical experiment is shown in the insert of Fig. 2). We nally investigated whether this effect of DEX was the result of a non-speci c suppression of Eø genomic activity, or rather, was speci c for CD11b. The data shown in Table 1  the lack of effect of DEX incubation on basal CD49d expression on the plasma membrane of Eøs.

Discussion
In the present study we have investigated the effect of DEX on the expression of adhesion molecules on the surface of human Eøs. We examined the effect of DEX on basal, as well as eotaxin-stimulated, CD11b levels. We have shown, to our knowledge for the rst time, that DEX is able to reduce the expression of CD11b on human Eøs. We used the CC chemokine eotaxin to stimulate CD11b upregulation on Eøs. The concentration of eotaxin used was shown to be maximal for upregulating CD11b on Eøs in previous studies. 9,17 Different levels of CD11b upregulation has been achieved between studies. One reason for the difference could be the source of Eøs; a previous study used Eøs from non-atopic donors 9 whereas we have used Eøs from atopic subjects. Therefore, the Eøs in the present study might be desensitized to certain mediators through exposure in vivo. It also appears that the cells were more responsive to PAF than eotaxin which is in contrast to previous ndings. 9 Again, this may be due to the desensitization status of the cells.
Together with cytokines and chemokines, adhesion molecules play a pivotal role in the process of leukocyte recruitment. Recently adhesion molecules have been shown to be a target (although not too sensitive) for glucocorticoid hormone action. 13 Whereas DEX effect on CD62E or CD54 upregulation on endothelial cells, broblasts and epithelial cells is well accepted, much less is known of the potential effect of this steroid on leukocyte adhesion molecules. Few studies have reported an ability of glucocorticoid hormones to reduce basal expression of selected adhesion molecules, though some species speci city has also been found. Prolonged (. 3 days) exposure to DEX reduced basal levels of CD62L and CD18 on bovine neutrophils; 15 CD62L and CD11a, but not CD11b, on rat neutrophils; 18 and CD54, but not CD11b, on rat monocytes. 16 In man, treatment with methyl prednisolone reduced upregulated, but not basal, CD11b expression on neutrophils after cardiopulmonary bypass, with no effect upon CD11a and CD11c. 19 No studies have addressed this question on Eøs.
In the present study DEX was found to reduce both basal CD11b expression on Eø cell surface, and this reduction explained the lower levels measured after cell activation with eotax-in. Therefore, the glucocorticoid affected the basal level of the antigen expressed on the Eø plasma membrane but not the cellular responsiveness to eotaxin. DEX effect was not due to an alteration of Eø viability. Further, it was not secondary to a non-speci c suppression of Eø genomic activity as levels of a different integrin (CD49d) were not modi ed with Eø incubation with DEX. However, increased avidity, without alterations in total expression, of CD49d has been reported following exposure to chemokines. 20 It would therefore be of interest to investigate whether DEX exerted an effect on b 1 avidity states.
From these initial data we cannot identify the molecular mechanism by which DEX reduces CD11b expression on human Eøs. It could be the result of reduced gene transcription (which may be the case for CD54 as its promoter region contains glucocorticoid response elements (GRE) 21 or, an effect on mRNA or protein stability. The former hypothesis appears to be unlikely since GRE in the promoter region of the human CD11b gene has not been described. 22 However, an indirect mechanism of DEX through transcription factors or on CD11b vesicular turnover cannot be excluded.
Eø recruitment in experimental models of allergic in ammation relies on CD11b function as demonstrated in vitro on Eø transendothelial migration, 6 and in vivo on eotaxin-induced Eø accumulation. 23 Based on the data presented here, the possibility that DEX may affect Eø accumulation via, at least in part, an effect on CD11b should be taken into account, especially when long-lasting dose regimens are used. This can be particularly true for human diseases, where glucocorticoid hormones are given for prolonged periods for the treatment of asthmatic and allergic rhinitis. 10 In conclusion, we believe the results from this study are important in understanding the anti-in ammatory mechanism(s) of actions exerted by glucocorticoid hormones in diseases of allergic in ammation.