Abstract
Phagocyte recognition, uptake, and nonphlogistic degradation of neutrophils and other leukocytes undergoing apoptosis promote the resolution of inflammation. This study assessed the effects of anti-inflammatory glucocorticoids on this leukocyte clearance mechanism. Pretreatment of “semimature” 5-day human monocyte-derived macrophages (Mφ) for 24 h with methylprednisolone, dexamethasone, and hydrocortisone, but not the nonglucocorticoid steroids aldosterone, estradiol, and progesterone, potentiated phagocytosis of apoptotic neutrophils. These effects were specific in that the potentiated phagocytosis of apoptotic neutrophils was completely blocked by the glucocorticoid receptor antagonist RU38486, and glucocorticoids did not promote 5-day Mφ ingestion of opsonized erythrocytes. Similar glucocorticoid-mediated potentiation was observed with 5-day Mφ uptake of alternative apoptotic “targets” (eosinophils and Jurkat T cells) and in uptake of apoptotic neutrophils by alternative phagocytes (human glomerular mesangial cells and murine Mφ elicited into the peritoneum or derived from bone marrow). Importantly, methylprednisolone-mediated enhancement of the uptake of apoptotic neutrophils did not trigger the release of the chemokines IL-8 and monocyte chemoattractant protein-1. Furthermore, longer-term potentiation by methylprednisolone was observed in maturing human monocyte-derived Mφ, with greater increases in 5-day Mφ uptake of apoptotic cells being observed the earlier glucocorticoids were added during monocyte maturation into Mφ. We conclude that potentiation of nonphlogistic clearance of apoptotic leukocytes by phagocytes is a hitherto unrecognized property of glucocorticoids that has potential implications for therapies aimed at promoting the resolution of inflammatory diseases.
Inflammatory responses, which evolved to eliminate invading microorganisms and repair damaged tissues, become undesirably persistent in a number of disease states. Although there is strong evidence that leukocytes can mediate tissue injury in inflammatory disorders 1 , little is known of the mechanisms that promote the resolution of inflammation by eliminating leukocytes from tissues. An important factor in the successful resolution of inflammation is the recognition, uptake, and degradation by phagocytes of intact leukocytes undergoing deletion by apoptosis 2, 3, 4, 5, 6, 7, 8, 9 . Not only are tissues protected from the noxious contents of leukocytes, but apoptotic cells are also rapidly phagocytosed and degraded by professional phagocytes (macrophages (Mφ)4) and by semiprofessional phagocytes (e.g., glomerular mesangial cells, fibroblasts) without inciting proinflammatory secretory responses 6, 10, 11, 12 .
Although several molecular pathways by which phagocytes recognize apoptotic cells have been identified (reviewed in Refs. 13 and 14), the mechanisms that control phagocytic capacity for the clearance of apoptotic cells are less well understood. The modulation of phagocyte capacity for apoptotic cell clearance represents a potential therapeutic target in the control of inflammatory disease. Although CD44-mediated increases in the Mφ uptake of apoptotic neutrophils may hold some therapeutic promise 15 , it is unlikely that there is clinical utility in the observation that pro-inflammatory cytokines promote Mφ ingestion of apoptotic cells 16 . Therefore, we sought to investigate further the regulatory mechanisms that might control phagocyte clearance of leukocytes undergoing apoptosis.
It is well established that glucocorticoids are powerful anti-inflammatory agents that suppress many phlogistic responses including inflammatory cell recruitment and activation 17, 18, 19, 20 . However, there have been relatively few studies of the effects of these steroids on the resolution phase of inflammation. We and others have shown that glucocorticoids delay constitutive apoptosis in neutrophils, whereas eosinophil apoptosis is accelerated 21, 22, 23 , providing an attractive explanation for the therapeutic efficacy of glucocorticoids in eosinophilic inflammation. However, implicit in this observation is the need for nonphlogistic clearance of an increased tissue load of apoptotic eosinophils, implying that glucocorticoids might also up-regulate the phagocytic capacity for apoptotic granulocytes so that greater numbers are safely removed. Moreover, although glucocorticoids delay neutrophil apoptosis, these agents are effective in suppressing inflammatory responses characterized by intense infiltration of tissues with neutrophils 24 that will eventually undergo apoptosis, implying that glucocorticoids may also up-regulate the phagocyte capacity to clear apoptotic neutrophils.
To test the hypothesis that glucocorticoids potentiate phagocyte capacity for the nonphlogistic clearance of apoptotic leukocytes, we have undertaken the first study of the effects of glucocorticoids upon this process, which is important in the resolution of inflammation. We report that glucocorticoids specifically promote safe clearance by various phagocyte types of apoptotic leukocytes from different lineages, establishing that glucocorticoids have hitherto unrecognized but beneficial regulatory effects upon phagocytes, which may promote the safe termination of inflammatory responses.
Materials and Methods
Materials
All reagents were obtained from Sigma (St. Louis, MO) unless otherwise stated. Culture media (Iscove’s modified DMEM, RPMI 1640, and HBSS) and supplements (100 U/ml penicillin, 100 58/ml streptomycin, 2 mM glutamine, and FCS) were obtained from Life Technologies Laboratories (Paisley, U.K.). Brewer’s thioglycollate medium was obtained from Difco (Detroit, MI). Methylprednisolone was obtained from Upjohn (Birmingham, U.K.); dexamethasone, hydrocortisone, aldosterone, progesterone, and estradiol were obtained from Sigma (Poole, U.K.). RU38486 was kindly provided by Dr. Ian Hall (University of Nottingham). Percoll was obtained from Pharmacia Fine Chemicals (Piscataway, NJ). Sterile tissue culture plasticware was obtained from Falcon Plastics (Cockeysville, MD).
Granulocyte isolation and induction of apoptosis
Human neutrophils (>98% pure on May-Giemsa-stained cytopreparations) were isolated from fresh, citrated blood of healthy volunteers by dextran sedimentation and discontinuous plasma-Percoll (Pharmacia Fine Chemicals) density gradient centrifugation 2, 25 . The neutrophils were aged in tissue culture at 37°C and 5% CO2 for 24 h in Iscove’s DMEM with 10% autologous platelet-rich, plasma-derived serum to undergo apoptosis verified by typical morphology 2 . Only aged neutrophils with a viability (assessed by trypan blue dye exclusion) >98% were used.
Human eosinophils (95–98% pure on May-Giemsa-stained cytopreparations) were isolated from the blood of mildly eosinophilic but healthy human volunteers by using a combination of discontinuous plasma-Percoll density gradient centrifugation to obtain eosinophil-enriched granulocyte populations followed by immunomagnetic depletion of neutrophils using 3G8 CD16 mAb, essentially as described 6, 26 . Purified eosinophils were then cultured for up to 72 h in conditions identical to those for neutrophils to undergo apoptosis verified by typical morphology 26 . Only aged eosinophils with viability >98% by trypan blue dye exclusion were used.
Preparation of human monocyte-derived Mφ
Human monocytes were prepared by well-established methods from the mixed mononuclear cell band of the discontinuous plasma-Percoll gradients used to prepare neutrophils 2 . Mononuclear cells were suspended in Iscove’s DMEM at 4 × 106/ml, and 100 μl was added to each well of a 96-well plate, which was then incubated at 37° for 1 h. Nonadherent cells, including contaminating lymphocytes, were then washed off, and adherent monocytes were cultured in Iscove’s DMEM with 10% autologous platelet-rich, plasma-derived serum for 5 days to mature into monocyte-derived Mφ.
Preparation of murine Mφ populations
Ten-week-old BALB/c mice were purchased from the University of Nottingham Biomedical Facility and were housed and treated under U.K. Home Office-approved conditions. Thioglycollate-elicited inflammatory peritoneal Mφ and bone marrow-derived Mφ were prepared as described 27 . Briefly, inflammatory peritoneal exudate Mφ were elicited into the peritoneal cavity of 10-wk-old female BALB/c mice by injection of 3% Brewer’s thioglycollate and harvested 5 days later after humane killing by peritoneal lavage with DMEM medium alone. Exudate cells were adhered to 96-well plates for 1 h at 0.1 × 106 cell/well, and nonadherent cells were washed off and replaced with DMEM containing 10% FCS plus supplements as above. Peritoneal Mφ were used within 48 h after isolation. Bone marrow-derived Mφ were isolated from femurs of humanely killed mice, cut at both ends, and the bone marrow expelled by flushing with a 25-gauge needle with DMEM with 10% FCS supplements as above and 10% conditioned medium from L929 cells as a source of monocyte-CSF. Mφ were employed after 7–10 days culture.
Culture of human mesangial cells
Mesangial cells were prepared using standard methods of serial culture/trypsinization in tissue culture flasks of adherent outgrowth cells from glomeruli obtained by sieving diced human normal renal cortex as described previously 3, 11 . Cells were cultured in RPMI 1640 medium with 10% FCS and 5 ml/500 ml of insulin/selenium/transferrin growth supplement from Life Technologies Laboratories (Grand Island, NY) and were used between passages four and six after subculture into 96-well plates. Great care was taken to verify the purity and phenotype of mesangial cells, as previously described 3, 11 . Cells were uniformly smooth muscle actin-positive and CD45-negative (by immunofluorescence). Cells did not take up acetylated low-density lipoprotein or opsonized zymosan particles (excluding Mφ contamination) and were cytokeratin and factor VIII-related Ag-negative (by immunofluorescence), excluding contamination with glomerular epithelial or endothelial cells.
Culture of Jurkat T cells and induction of apoptosis
Cells of the Jurkat T cell line (a gift from C. Gregory, University of Nottingham) were grown in suspension culture in RPMI 1640 with 5% FCS and supplements. To induce morphologically verified apoptosis, cells were deprived of serum with cycloheximide at 50 μg/ml for 4 h before being washed and employed in interaction assays with Mφ.
Interaction assays
A microscopically quantified phagocytic assay of Mφ phagocytosis of aged polymorphonuclear leukocytes (PMNs) was used, which has been described and illustrated in detail before 2, 28 . Minor adaptations for Mφ cultured in 96-well plates have been described 16 . Briefly, PMNs aged for 24 h in culture, to undergo apoptosis, were washed once in HBSS, suspended in Iscove’s DMEM, and 0.5 × 106 aged PMNs in 50 μl of medium were added to each washed well of Mφ. After interaction for 30 min at 37°C in 5% CO2 atmosphere, the wells were washed in saline at 4°C to remove noningested apoptotic PMNs, fixed with 2% glutaraldehyde in 0.9% saline, stained for myeloperoxidase (MPO; present in PMNs but not Mφ) using hydrogen peroxide and dimethoxybenzidine as substrate as previously described 2, 29 , and then the proportion of Mφ ingesting neutrophils were counted by inverted light microscopy, exactly as described 2, 28 . Because it was also important to determine whether the phagocytic capacity of individual Mφ had been increased, the number of PMNs within 100 randomly selected Mφ in each well was counted in some experiments. These counts were possible because the duration of the phagocytic assay was sufficiently short for ingested PMNs to remain intact 2, 28, 29 . The Mφ uptake of aged apoptotic eosinophils, 0.1 × 106 in 50 μl medium for each well of Mφ or of IgG-opsonized human erythrocytes (EIgG), 0.1 × 106 in 50 μl medium for each well of Mφ, was determined by similar means, as described 26, 28 , because these cell types also stain for peroxidase. The uptake of apoptotic Jurkat T cells was assessed by staining with Haemalum (BDH, Poole, U.K.), as described 4 .
To quantify mesangial cell phagocytosis of apoptotic PMNs, we used a previously described, reproducible, microscopically quantified assay in which the uptake of apoptotic PMN was shown by electron microscopy and susceptibility to inhibition by colchicine to be a result of active phagocytosis 3 . Culture medium was aspirated from the mesangial cells, and 50 μl of a 20 × 106/ml suspension of neutrophils in RPMI 1640 with 10% FCS was added to each well together with 50 μl of Iscove’s DMEM. The two cell types were cocultured in 5% CO2 at 37°C for 3 h, and the interaction was then stopped by the addition of cold normal saline and the removal of nonadherent cells by manual washing with a Pasteur pipette. The mesangial cell monolayer was then trypsinized, and a separate cytocentrifuge preparation was prepared for each well. These were fixed with 2% glutaraldehyde and stained for MPO and finally counterstained with Haemalum. The proportion of mesangial cells containing brown MPO-positive globules wholly within the outline of the cell was then counted by light microscopy with a minimum of 500 cells per slide being counted.
Effect of steroid hormones
Before interaction with apoptotic leukocytes, Mφ or mesangial cells were preincubated for varying durations of time in medium plus supplements/serum, as described above, containing the desired final concentration of steroid hormone. Control wells were left undisturbed in medium plus supplements/serum. However, preliminary experiments (not shown) demonstrated that the replacement of medium after aspiration at the preincubation times used did not, in itself, alter the phagocytic signal.
Effect of glucocorticoid receptor antagonist RU38486
The effect of glucocorticoid receptor antagonist RU38486 was assessed using final concentrations of RU38486 or ethanol-based vehicle control >10-fold those employed for glucocorticoids. Medium was aspirated from phagocytes 24 h before assay of phagocytosis and was replaced by 180 μl of RU38486 at 11.1-fold final steroid concentration in medium plus supplements/serum as above. Twenty minutes later, 20 μl of medium plus supplements/serum containing no steroid (as a control) or steroid at 10-fold final concentration was added, and the phagocytes were incubated for the desired period until interaction with apoptotic leukocytes.
Effect of cycloheximide
The protein synthesis inhibitor cycloheximide was included at 2.5 μM for 12 h in medium plus supplements/serum ± steroid before assay of phagocytosis of apoptotic leukocytes.
Assessment of Mφ and mesangial cell response following the ingestion of apoptotic PMNs
Monocyte-derived Mφ in 24-well plates were either incubated in Iscove’s DMEM alone or interacted with apoptotic PMNs or opsonized zymosan. After 30 min, Mφ were carefully washed, and 250 μl Iscove’s DMEM was added to each well. After 24 h coculture with Mφ, the medium was aspirated, centrifuged, and stored at −80°C. Mφ phagocytosis of neutrophils or zymosan was quantified by fixing three of the wells on each plate with 2% glutaraldehyde and, in the case of neutrophils, by staining for MPO as outlined previously. The proportion of Mφ ingesting apoptotic neutrophils or opsonized zymosan was counted by inversion light microscopy 28 .
Mesangial cells in 96-well plates were either cultured in the presence of RPMI 1640 with 10% FCS with or without 10 ng/ml IL-1 for 6 h or interacted for 3 h with apoptotic PMNs as previously outlined. Mesangial cells were then carefully washed with normal saline, and the medium was replaced with 50 μl of RPMI 1640 without FCS and incubated at 37°C. The medium was aspirated after 24 h, centrifuged, and stored at −80°C. Cytocentrifuge preparations from 4 wells on each plate were used to quantify mesangial cell phagocytosis of apoptotic PMNs.
Measurement of IL-8 and monocyte chemoattractant protein-1 (MCP-1) levels in tissue culture supernatants
IL-8 and MCP-1 levels were measured by ELISA using a peroxidase conjugate for detection. Briefly, assay plates were coated with protein A-purified goat polyclonal Ab against pure natural IL-8 30 or rabbit polyclonal Ab against recombinant MCP-1 3132 , granulocyte chemotactic protein-2, ENA-78, GROa, inflammatory protein-10, NAP-2 33 , or cytokines (IL-1, IFN-γ, IFN-β) were not detectable. The sensitivity of the MCP-1 and IL-8 ELISA were <1 ng/ml and <100 pg/ml, respectively.
Statistical methods
All values are expressed as mean ± SEM of the indicated number of experiments. Statistical significance (defined as p < 0.05) was evaluated using Student’s t test and where appropriate by one-way ANOVA with comparison between groups using the Newman-Keuls procedure.
Results
Glucocorticoids specifically promote phagocytosis of apoptotic neutrophils by semimature monocyte-derived Mφ
To facilitate the detection of possible potentiating effects upon Mφ phagocytosis of apoptotic neutrophils 16 , we studied “semimature” cultures of human monocyte-derived Mφ matured for 5 days in adherent culture on plastic because maximal capacity for the uptake of apoptotic cells generally requires maturation for around 7 days 29 . Pretreatment of 5-day Mφ with the glucocorticoid methylprednisolone (Fig. 1⇓) or dexamethasone (data not shown) for different periods of time up to 24 h caused a progressive increase in the proportion of Mφ that were capable of phagocytosing apoptotic neutrophils and in the number of apoptotic neutrophils ingested per Mφ. For example, methylprednisolone at 200 nM for 24 h augmented the number of apoptotic neutrophils ingested per 100 phagocytically active Mφ from 138.9 ± 4.1 to 209.6 ± 5.1 (p < 0.01; n = 9). Short preincubation times (<3 h) were without effect (Fig. 1⇓), excluding the remote possibility that there was a “carry over” effect of glucocorticoid treatment of Mφ upon added apoptotic neutrophils. The specificity of glucocorticoid effects upon Mφ phagocytosis was further investigated by comparison of the effects of pretreatment of Mφ for 24 h with a number of different steroids. These studies demonstrated that methylprednisolone, dexamethasone, and hydrocortisone acted in a concentration-dependent manner to augment phagocytosis (Fig. 2⇓), whereas the mineralocorticoid aldosterone at 5 nM, a concentration 10-fold above that inducing maximal effects on sodium transport in vitro 34 , and the sex steroids progesterone (1 μM) and estradiol (1 μM) were without effect (Table I⇓). The specificity of the observed augmentation of Mφ phagocytosis of apoptotic neutrophils by glucocorticoids was further demonstrated by the failure of glucocorticoid pretreatment to induce either binding or phagocytosis of nonapoptotic neutrophils (data not shown) and the lack of effect of methylprednisolone on the phagocytosis of EIgG; in a series of nine experiments, pretreatment for 24 h with methylprednisolone at 200 nM increased the proportion of 5-day Mφ ingesting apoptotic neutrophils from 14.1 ± 1.5% to 29.6 ± 3.2% (mean ± SE, p < 0.001) but had no significant effect on the proportion of 5-day Mφ ingesting EIgG (untreated 5-day Mφ 52.7 ± 6.4% vs methylprednisolone-treated 5-day Mφ 54.3 ± 7.1%).
Time course of potentiation by 200 nM methylprednisolone of human monocyte-derived Mφ phagocytosis of apoptotic PMNs. Note increase in methylprednisolone-treated Mφ (filled bars) vs control Mφ (open bar) as early as 3 h. In all conditions, phagocytosis was assayed at the same time, with differing preincubation times. Data are mean ± SEM, n = 17. ∗, p < 0.05 and ∗∗, p < 0.001 vs control values.
Concentration-related potentiation of human monocyte-derived Mφ phagocytosis of apoptotic PMNs by 24-h treatment with glucocorticoids. Methylprednisolone, dexamethasone, and hydrocortisone exhibited similar effects. Logistic considerations required a separate series of experiments for each glucocorticoid explaining variation in controls for each compound. Data are mean ± SEM, n = 12 for each compound. ∗, p < 0.05 and ∗∗, p < 0.001 vs control values.
Effect of 24-h treatment with nonglucocorticoid steroid hormones on human monocyte-derived Mφ uptake of apoptotic PMNsa
Glucocorticoid promotion of Mφ uptake of apoptotic neutrophils is specific for the glucocorticoid receptor and inhibitable by cycloheximide
To confirm that glucocorticoids were acting via Mφ steroid receptors, we assessed the effect of the specific steroid receptor antagonist RU38486 upon glucocorticoid-promoted 5-day Mφ phagocytosis of apoptotic neutrophils (Fig. 3⇓). RU38486 used at a 10-fold greater concentration than the glucocorticoids completely blocked increased phagocytosis stimulated by 24-h treatment with methylprednisolone, dexamethasone, and hydrocortisone without exhibiting any independent inhibitory effect upon Mφ phagocytosis of apoptotic cells (Fig. 3⇓). Together with the lack of effect upon Mφ phagocytosis of EIgG cited above, these data argue strongly against a nonspecific “membrane” effect on Mφ. Furthermore, methylprednisolone failed to promote phagocytosis of apoptotic neutrophils by Mφ pretreated for 12 h with the protein synthesis inhibitor cycloheximide (a reversible blocker of mRNA translation) at 2.5 μM, a concentration previously shown to inhibit >95% of protein synthesis in myeloid cells and to inhibit Mφ phagocytosis of apoptotic cells without toxic effects on the Mφ, as evidenced by retained ability to ingest opsonized erythrocytes 35, 36 . Thus, in a series of seven experiments in which 15.7 ± 1.7% (mean ± SE) of 5-day Mφ ingested apoptotic neutrophils under control conditions and 12-h pretreatment with methylprednisolone at 200 nM increased recognition to 38.3 ± 4.2% (p < 0.001 vs control), such glucocorticoid pretreatment was unable to reverse the inhibitory effect of concomitant 12-h pretreatment with cycloheximide at 2.5 μM, which significantly reduced the proportion of Mφ ingesting apoptotic neutrophils whether Mφ had been treated with methylprednisolone (4.9 ± 0.9%, p < 0.001) or not (5.7 ± 1.4%, p < 0.001).
Specificity of glucocorticoid enhancement of human monocyte-derived Mφ phagocytosis of apoptotic neutrophils; abrogation by competitive antagonism of steroid receptor. Note RU38486 at 10 times the glucocorticoid concentration completely inhibited potentiation by methylprednisolone at 200 nM, dexamethasone at 1 μM, and hydrocortisone at 1 μM without exerting an independent inhibitory effect on phagocytosis; controls included RU38486 vehicle. Data are mean ± SEM, n = 10. ∗∗, p < 0.001 vs control values.
Glucocorticoids also promote phagocytosis in other models of phagocyte clearance of apoptotic cells
In view of the observation that ligation of Mφ CD44 specifically promotes phagocytosis of apoptotic neutrophils, having little effect upon uptake of apoptotic cells of other lineages 15 , it was of interest to determine whether the prophagocytic effect of glucocorticoids upon Mφ was restricted to uptake of apoptotic neutrophils. This proved not to be the case in that methylprednisolone also promoted phagocytosis of apoptotic Jurkat cells from 13.4 ± 1.4% of Mφ under control conditions to 24.2 ± 1.5% after 24-h culture with 200 nM methylprednisolone (mean ± SE, n = 6, p < 0.001). Furthermore, 24-h incubation of Mφ with 1 μM dexamethasone increased Mφ phagocytosis of apoptotic eosinophils from 49.0 ± 3.9% to 67.8 ± 5.1% (mean ± SE, n = 6, p < 0.001).
We were also concerned to establish whether the prophagocytic effect of glucocorticoids was limited to the human monocyte-derived Mφ, because previous work has established that other phagocyte populations recognize apoptotic cells by mechanisms distinct from the αVβ3 vitronectin receptor/thrombospondin/CD36 system characterized in the human monocyte-derived Mφ (reviewed in 13 . Human mesangial cells, “semiprofessional” glomerular phagocytes, which usually recognize apoptotic neutrophils by a CD36-independent, αVβ3-mediated mechanism 11 also exhibited increased phagocytosis of apoptotic neutrophils when treated with glucocorticoids (Fig. 4⇓). Furthermore, this effect was not limited by phagocyte species in that glucocorticoids exhibited a similar promotion of phagocytosis of apoptotic neutrophils by murine bone marrow-derived Mφ populations and murine thioglycollate-elicited peritoneal inflammatory Mφ (Fig. 5⇓), which normally employ αVβ3-independent phosphatidylserine receptors in the uptake of apoptotic cells 27 .
Time course of glucocorticoid potentiation of human glomerular mesangial cell (HMC) phagocytosis of apoptotic neutrophils. Methylprednisolone, dexamethasone, and hydrocortisone, all at 200 nM, exerted similar effects. Under all conditions, HMC phagocytosis of apoptotic PMNs was assayed at the same time; preincubation times varies as shown. Data are mean ± SEM, n = 6. ∗, p < 0.05; ∗∗, p < 0.001 vs control values.
Concentration-related potentiation of murine Mφ phagocytosis of human apoptotic neutrophils by 24-h treatment with methylprednisolone. A, Murine bone marrow-derived Mφ (BMMφ). B, Thioglycollate-elicited peritoneal Mφ. Data are mean ± SEM, n = 6. ∗∗, p < 0.001 vs control values.
Glucocorticoid-promoted uptake of apoptotic cells does not promote proinflammatory secretory responses by phagocytes
Growing evidence supports the concept that phagocyte clearance of leukocytes undergoing apoptosis is an injury-limiting leukocyte disposal mechanism likely to promote the resolution of inflammation (reviewed in 37 . Strong support is provided by data that demonstrate that Mφ and mesangial cells phagocytosing apoptotic cells are not stimulated to secrete proinflammatory eicosanoids, granule enzymes, or cytokines 6, 10, 11 . The wide-ranging anti-inflammatory effects of glucocorticoids suggested that the promotion of phagocyte clearance of apoptotic cells would be unlikely to provoke proinflammatory responses from ingesting phagocytes, but it was clearly necessary to test this assumption. We found that increases in Mφ uptake of apoptotic neutrophils induced by methylprednisolone pretreatment did not invoke the release of the proinflammatory chemokines IL-8 and MCP-1 by Mφ (Table II⇓) or mesangial cells (Table III⇓). However, methylprednisolone at 200 nM failed to abrogate IL-8 and MCP-1 secretion triggered by Mφ phagocytosis of opsonized zymosan particles (Table III⇓), indicating that glucocorticoids were not acting to suppress production of these chemokines by Mφ.
Effect of 24-h treatment with 200 nM methylprednisolone upon IL-8 and MCP-1 release from human monocyte-derived Mφ ingesting particulate stimulia
Effect of 24-h treatment with 200 nM methylprednisolone upon IL-8 and MCP-1 release from stimulated human mesangial cellsa
During monocyte maturation, earlier and more prolonged exposure to glucocorticoid results in greater potentiation of Mφ phagocytic capacity at 5 days
Although human monocyte-derived Mφ are able to ingest senescent neutrophils undergoing apoptosis, freshly isolated human peripheral blood monocytes lack this capacity, which is progressively acquired over several days as adherent monocytes cultured with autologous serum differentiate into Mφ 29 . Because the mechanisms responsible are very poorly understood, we considered it possible that early exposure of maturing monocytes to glucocorticoids might undesirably disrupt acquisition of the capacity to ingest apoptotic neutrophils. In fact, the earlier maturing monocytes were exposed to 200 nM prednisolone throughout subsequent culture, the greater (compared with control) the potentiation of the uptake of apoptotic neutrophils at 5 days (Fig. 6⇓).
Progressively earlier inclusion of methylprednisolone at 200 nM in cultures of maturing monocytes leads to greater potentiation of 5-day Mφ phagocytosis of apoptotic neutrophils. Compare potentiation by 1-day treatment with that for 4-day treatment (i.e., methylprednisolone included from 1 day of monocyte (Mφ culture). Data are mean ± SEM, n = 6. ∗, p < 0.05; ∗∗, p < 0.001 vs control values.
Discussion
This report establishes the capacity of glucocorticoids to promote nonphlogistic phagocytosis of cells undergoing apoptosis. This effect was specific in that in human monocyte-derived Mφ treated for 24 h from 4 days’ maturation it was mediated via the RU48686-inhibitable steroid receptor, not induced by the mineralocorticoid aldosterone or by sex steroids, and not observed for FcR-mediated phagocytosis. Furthermore, there was evidence of a generally relevant promotion of phagocyte clearance of apoptotic cells in that glucocorticoids promoted Mφ phagocytosis of apoptotic Jurkat T lymphocytes and apoptotic eosinophils as well as promoting the uptake of apoptotic neutrophils by both human glomerular mesangial cells and two different populations of murine Mφ. Finally, the promoting effect of glucocorticoids was not limited to “semimature” 4-day Mφ because earlier exposure to glucocorticoid during Mφ differentiation resulted in a still greater capacity for phagocytosis of apoptotic cells at 5 days.
The data provide new insights into the anti-inflammatory effects of glucocorticoids, demonstrating a hitherto unrecognized mechanism by which these agents may ameliorate inflammatory injury (the promotion of safe phagocytic clearance of leukocytes being deleted by apoptosis). Data from in vitro studies emphasize that glucocorticoids may coordinately promote the deletion of certain leukocyte types by engaging apoptosis 22, 23, 38 , while at the same time up-regulating the capacity for phagocytic clearance (this study). Recently published observations that steroid treatment of asthma is associated with the resolution of eosinophilic inflammation and increased evidence of eosinophil phagocytosis by broncho-alveolar Mφ provide evidence that similar mechanisms may operate in vivo in man 8 . Furthermore, although glucocorticoids slow constitutive apoptosis in neutrophils 21, 23 , glucocorticoid-treated neutrophils eventually undergo apoptosis and therefore require phagocytic clearance in vivo to prevent secondary necrosis with the resultant release of toxic granule contents. Further studies will be required to establish whether these findings can be extrapolated to inflammatory conditions in different tissues in vivo. However, it appears likely that the observed increases in both the proportion of Mφ ingesting apoptotic neutrophils and in the number of ingested cells per phagocytically active Mφ will equate to a significant increase in clearance capacity in vivo.
Nevertheless, our observation that glucocorticoid-mediated enhancement of Mφ phagocytosis of apoptotic cells was not achieved by a costly loss of the teleogically appropriate lack of proinflammatory response by the phagocyte is of fundamental importance. Methylprednisolone-enhanced phagocytosis of apoptotic cells failed to stimulate IL-8 and MCP-1 release from either Mφ or mesangial cells. By contrast, under the conditions employed, methylprednisolone did not suppress Mφ release of IL-8 and MCP-1 after phagocytosis of opsonized zymosan. These findings were underscored by comparable observations when mesangial cells were employed as the phagocyte, where IL-1 was employed as a positive control stimulus 11, 30 instead of opsonized zymosan. Therefore, failure to elicit a proinflammatory response did not merely reflect a general suppression of chemokine synthesis by glucocorticoids, which has been reported in other in vitro systems (reviewed in 24 .
The modulation of phagocyte capacity for uptake of apoptotic cells by glucocorticoids exhibits interesting differences from previous reports on the regulation of phagocytosis of these and other particles. In contrast to CD44-mediated enhancement of Mφ uptake of apoptotic neutrophils 15 , the prophagocytic effect of glucocorticoids required several hours of treatment and was unable to reverse inhibition by cycloheximide, consistent with a requirement for new protein synthesis. Moreover, unlike CD44 ligation, prophagocytic effects were not restricted to apoptotic target cells of the neutrophil lineage. Furthermore, in other phagocytic systems glucocorticoids have been reported to inhibit rat alveolar Mφ ingestion of carbon particles 39 and murine Mφ phagocytosis of heat-killed Saccharomyces cerevisiae 40, 41, 42, 43 . Nevertheless, freshly isolated human monocyte phagocytosis of β-glucan particles prepared from the same species of yeast was enhanced by glucocorticoids such as dexamethasone at 200 nM 44 , and there are reports that glucocorticoids up-regulate Mφ receptors for cytokines, including those for granulocyte-Mφ CSF 45 , which can promote Mφ uptake of apoptotic cells 16 . Consequently, the current data are not necessarily inconsistent with previous studies of the glucocorticoid regulation of phagocytosis.
Further studies will be needed to define mechanisms by which glucocorticoids potentiate phagocytosis of apoptotic cells. Our preliminary immunofluorescence flow cytometry studies (data not shown) on 5-day human monocyte-derived Mφ treated with 200 nM methylprednisolone for 24 h have shown no detectable difference in Mφ surface expression of components of the αvβ3 vitronectin receptor/thrombospondin/CD36 recognition mechanism 35 and a modest down-regulation of CD14 consistent with previous reports 46 , suggesting that up-regulated expression of these surface receptors is unlikely to account for the potentiating effects of glucocorticoids. This work on surface receptor expression will need to be expanded, and possible effects of glucocorticoids on cytoskeletal elements 47 should also be considered.
Lastly, although hydrocortisone and dexamethasone might appear to be equipotent in our studies, comparisons of concentrations at which different glucocorticoids exert a similar effect in vitro should be made with considerable caution because ABC transporters such as MDR1A can expel synthetic glucocorticoid substrates such as dexamethasone from within leucocytes 48 , with the result that the effective concentration of glucocorticoid at the intracellular receptor is lower than the extracellular concentration applied to the cell.
In conclusion, we report that glucocorticoids specifically enhance the nonphlogistic phagocytic uptake of apoptotic leukocytes of differing lineages by human and murine Mφ populations and also “semiprofessional” phagocytes, glomerular mesangial cells. The data add new emphasis to the anti-inflammatory properties of glucocorticoids and suggest that further work could identify novel therapeutic approaches toward promoting the safe resolution of inflammatory conditions.
Acknowledgments
We thank Prof. M. Horton (St Bartholomew’s Hospital, London, U.K.), Dr. Ian Hall, and Prof. Chris Gregory (University of Nottingham, U.K.) for the gifts of reagents. We also thank Judith Hayes and Dorothy May for expert secretarial assistance.
Footnotes
↵1 This work was supported by the Wellcome Trust (047273 and 039108), the Medical Research Council (G9016491), and the University of Nottingham. Y.L. was supported by the National Kidney Research Fund, J.M.C. was supported by the University of Edinburgh, and A.G.R. was supported by the Medical Research Council.
↵2 Y.L. and J.M.C. contributed equally to this work and are joint first authors.
↵3 Address correspondence and reprint requests to Prof. John Savill, Department of Medicine, Royal Infirmary of Edinburgh, Lauriston Place, Edinburgh HE3 9YW, U.K. E-mail address: J.Savill{at}ed.ac.uk
↵4 Abbreviations used in this paper: Mφ, macrophage(s); EIgG, IgG-opsonized E; MPO, myeloperoxidase; PMNs, polymorphonuclear leukocytes; MCP, monocyte chemoattractant protein.
- Received August 5, 1998.
- Accepted December 15, 1998.
- Copyright © 1999 by The American Association of Immunologists
References
In this issue
