Glucocorticoids promote survival of anti-inflammatory macrophages via stimulation of adenosine receptor A3

Glucocorticoids (GC) are very effective anti-inflammatory and immunosuppressive agents widely used in the treatment of many autoimmune and allergic diseases.¹  However, long-term therapy with GCs is associated with substantial side effects that limit their use in many clinical conditions. Thus, novel insights in the molecular mechanisms of GCs action on immune cells are a prerequisite for the development of more specific anti-inflammatory therapies.²  The anti-inflammatory effect of GCs on immune cells was believed to be primarily mediated by interaction with transcriptional factors essential for expression of proinflammatory genes (transrepression). GC-mediated gene expression (transactivation), on the other hand, has been supposed to be responsible for many GC side effects.³  With respect to monocytes, however, we and other groups demonstrated that GC-mediated gene transcription is crucial for differentiation of an active anti-inflammatory monocytic phenotype.^4-6 

Monocytes and macrophages are central components of the immune system. During infection and inflammation, they play a pivotal role in the generation of inflammatory mediators and regulation of innate and adaptive immune responses. They also contribute to resolution of inflammation, not only by producing anti-inflammatory cytokines but also by removal of proinflammatory pathogens, cellular debris and apoptotic cells present at later stages of inflammation.^7,8  Thus, monocytes and macrophages do not represent a homogenous cell population, but rather comprise specific subsets with proinflammatory or anti-inflammatory properties depending on their stage of differentiation as well as on distinct mechanisms of activation.^9-11 

It was generally believed that anti-inflammatory actions of GCs on monocytes and macrophages are mainly due to down-regulation of proinflammatory functions.^12,13  However, analyzing the general GC-induced expression pattern by microarray technology, we have previously shown that GC-treatment did not cause general suppression of monocytes, but induced specific differentiation with an anti-inflammatory phenotype, which seem to play a pivotal role in resolution of inflammation.^4,5  We have also shown that GCs induce a survival signal in these cells.⁴ 

Generally, monocytes are relatively short living cells, which die spontaneously in the absence of appropriate stimuli. Apoptosis is a critical process that regulates removal of monocytes thus controlling turnover of this cell population. Proinflammatory factors like lipopolysaccharide (LPS), interleukin-1β (IL-1β), or tumor necrosis factor-α (TNFα) are known to enhance survival of proinflammatory monocytes.¹⁴  Thus, proinflammatory signals released from cells within the inflammatory site result in local accumulation of a population of activated macrophages. During the later course of inflammation, these cells are replaced by a population of macrophages that promote the resolution phase of inflammation. However, much less is known about the molecular mechanisms promoting survival of these anti-inflammatory subpopulations of monocytes.

In the present study, we identified a novel mechanism of GC-induced anti-inflammatory effects. GC-treatment up-regulated expression of A3 adenosine receptor (A3AR), and stimulation of A3AR was responsible for activation of extracellular signal regulated kinase (ERK)1/2 phosphorylation, which in turn led to protection of monocytes from apoptosis. Our results therefore demonstrate that adenosine-induced survival of anti-inflammatory monocytes is a novel mechanism by which GCs are capable of down-regulating inflammatory processes.

Approval was obtained from the ethics committee of the Medical Faculty of Münster for these studies. Monocytes were isolated and cultured as described earlier.⁴  Dexamethasone (DEX), methylprednisolone (MP), and triamcinolone (T) were obtained from Sigma-Aldrich.

To induce apoptosis, monocytes (1 × 10⁶ cells/mL) cultured in polystyrene tubes (Becton Dickinson) were treated with different stimuli: 200nM staurosporine (STS; Axxora), 1 μg/mL actinomycin D (ActD), 1μM cycloheximide (CHX; both from Sigma-Aldrich), and 1 μg/mL mouse monoclonal anti-Fas antibody (clone CH-11; Beckman Coulter). To investigate changes in susceptibility to spontaneous and serum withdrawal-induced apoptosis, unstimulated monocytes were cultivated in medium with and without serum, respectively.

Externalization of phosphatidylserine was measured by flow cytometric staining with fluorescein isothiocyanate (FITC)–conjugated annexin V (BD PharMingen), and fragmentation of DNA was assessed by the method of Nicoletti as described previously.^15,16  Flow cytometric analysis was performed using FACSCalibur (BD Biosciences) and CellQuest Pro Version 4.0.2 (BD Biosciences) software.

For measurement of caspase-3, -8, and -9 activity, cells were lysed in caspase lysis buffer (20mM HEPES [N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid], pH 7.3, 84mM KCl, 10mM MgCl₂, 0.2mM EDTA [ethylenediaminetetraacetic acid], 0.2mM EGTA[ethyleneglycoltetraacetic acid], 0.5% NP-40, 5mM dithiothreitol [DTT] supplemented with protease inhibitors) for 10 minutes on ice and then centrifuged at 7000g at 4°C for 20 minutes. Cellular protein (20 μg) was incubated at 37°C with 50μM of the fluorogenic caspase-3 substrate Ac-DEVD-AMC (N-acetyl-Asp-Glu-Val-Asp-aminomethylcoumarin), caspase-8 substrate Ac-IEPD-AMC (N-acetyl-Ile-Glu-Pro-Asp-aminomethylcoumarin), or caspase-9 substrate Ac-LEHD-AMC (N-acetyl-Leu-Glu-His-Asp-aminomethylcoumarin; all from Bachem) in 200 μL buffer containing 50mM HEPES, pH 7.3, 100mM NaCl, 0.1% CHAPS, 10% sacharose, and 10mM DTT. The release of AMC was measured by fluorometry (HTS7000; PerkinElmer). The activity of caspases was expressed as a fluorescence increase (ΔF) per microgram of cellular protein.

Cell lysates were prepared for Western blot analysis as described earlier.¹⁷  Proteins were detected with antibodies against total and phospho-ERK1/2 (Thr202/Tyr204), total and phospho-SAPK/JNK (Thr183/Tyr185), total and phospho-p38 (Thr180/Tyr182), phospho-c-Raf (Ser338), phospho-MEK1/2 (Ser217/221), phospho-p90RSK (Ser380; all from Cell Signaling Technology), and adenosine A3 receptor (C-17; Santa Cruz Biotechnology).

Monocytes were isolated on 2 consecutive days. Monocytes isolated on day 1 were stimulated for 16 hours with 100nM DEX or left untreated then washed and resuspended in fresh medium without any additional stimuli. After 24 hours, conditioning medium was harvested and immediately used in experiments. In this order, monocytes isolated on day 2 were harvested, washed, and one part of them was resuspended in fresh medium and stimulated with DEX or ethanol as control. The remaining cells were resuspended in conditioned medium from GC-treated monocytes or control monocytes and subsequently cultured for 16 hours. Finally, cells were harvested and lysates for analysis of ERK1/2 phosphorylation were prepared.

For real-time reverse transcription polymerase chain reaction (RT-PCR), RNA from control monocytes and monocytes, which were stimulated with DEX for 4 and 16 hours, was analyzed in duplicate. cDNA was synthesized from 1 μg total RNA using SuperScriptII RNase-H reverse transcriptase (Invitrogen). The primers used for PCR analysis were as follows: A3AR forward, 5′-GGGCATCACAATCCACTTCT-3′; A3AR reverse, 5′-TGTGAGTGGTGACCCTCTTG-3′; c-Myc forward, 5′-TCAAGAGGTGCCACGTCTCC-3′; c-Myc reverse, 5′-CTTGACCCTCTTGGCAGCAG-3′; GAPDH forward, 5′-TGCACCACCAACTGCTTAGC-3′; GAPDH reverse, 5′-GGCATGGACTGTGGTCATGAG-3′; RPL forward, 5′-AGGTATGCTGCCCCACAAAAC-3′; RPL reverse, 5′-TGTAGGCTTCAGACGCACGAC-3′. Real-time RT-PCR was performed using the QuantiTect SYBR Green PCR kit (QIAGEN) as described previously.⁴ 

For inhibition studies, cells were pretreated with PD 98 059 (10μM), GF 109203X (10μM), PP2 (10μM; all from Merck), MRS 1220 (0.5μM), Mifepristone (RU-486, 3μM; all from Sigma-Aldrich), or Boc-FLFLF (Boc-Phe-Leu-Phe-Leu-Phe, 100μM; Bachem), for 30 minutes before stimulation with DEX (100nM). After 16 hours of culture, cells were harvested and stimulated with STS as described above. In parallel, cell lysates for analysis of ERK1/2 phosphorylation status were prepared.

The results of the microarray analysis have been published previously.⁴  In short, monocytes from 4 individual donors were stimulated for 16 hours with 10nM fluticasone propionate (equivalent to 100nM DEX) or vehicle control. RNA preparation, sample preparation, and hybridization to Human Genome 133 A Gene Chip arrays (Affymetrix) for microarray analysis were performed as described previously.¹⁷  Raw data were processed by MicroArray Suite (MAS) Version 5.0 software (Affymetrix). For more sophisticated statistical analysis, genes with a consistent change call and an expression over background in at least 3 of 4 experiments were retained and further analyzed using Expressionist Suite Version 3.1 software (GeneData). We retained only genes with a fold-change of more than or equal to 2.0 or less than or equal to −2.0 and a P value of less than .05 (paired t test).

To identify transcription factors with statistically overrepresented binding sites in the promoter regions of regulated genes, we used the computational tool CARRIE (Computational Ascertainment of Regulatory Relationships Inferred from Expression) with the implemented promoter sequence analysis tool ROVER (Relative OVERabundance of cis-elements) with minor modifications.¹⁸  Promoter sequences were defined as the region of 1000 bases upstream to 50 bases downstream of the transcription start site and obtained using the computational tool PromoSer.¹⁹  To identify statistically overrepresented binding sites in the promoter regions of genes induced by GC,⁴  an equally sized group of control genes with stable expression but no detectable regulation by GC is used. As additional control, a group of randomly chosen genes out of the bulk of genes analyzed by microarray technology was used to exclude underrepresentation of regulatory sequences within the group of nonregulated genes with stable expression (presumably many so called housekeeping genes). Regulatory sequences with stable overrepresentation within the group of selected genes independent from the chosen control genes were retained.

Monocytes were stimulated for 16 hours with 100nM DEX or left untreated. Monocytes (7 × 10⁶) were used per investigated antibody. The cells were fixed with 1% formalin for 10 minutes. Reaction was stopped by addition of glycine to a final concentration of 0.125 M. Cells were washed, lysed, and sonicated. The DNA-protein complexes were immunoprecipitated with antibody against c-Myc (N-262; Santa Cruz Biotechnology) and extracted by protein A/G sepharose beads. After several washing steps, the cross-link between DNA and protein was reversed at 65°C, followed by RNA and protein digestion with RNase and proteinase K and subsequent extraction of DNA (Jetquick PCR Product Purification Spin kit; Genomed). The DNA was amplified with primers flanking the SAP30 (Sin3A-associated protein) and Myc promoters including the binding site for c-Myc: SAP30 forward, 5′-TTCCGTCTCAACCACATTTCAAGT-3′; SAP30 reverse, 5′-CCTCCTTAAGCTGTGTCCA-3′; Myc forward, 5′-AAGGGCAGGGCTTCTCAGAG-3′; Myc reverse, 5′-GCGAGTTAGATAAAGCCCCGAA-3′. Real-time PCR was performed using the QuantiTect SYBR Green PCR kit (QIAGEN). The relative proportion of immunoprecipitated fragments were determined using comparative Ct method. ΔCt was obtained by subtracting Ct value of input control from Ct value of gene of interest. Results are presented as fold enrichment (2^ΔΔCt, where ΔΔCt value was calculated by subtracting ΔCt for unstimulated monocytes from ΔCt for GC-stimulated monocytes).

All values given through are expressed as means plus or minus SEM from at least 3 independent experiments. Statistical analysis was performed using t test. Differences were considered as significant when P was less than or equal to .05.

Monocytes were exposed to 100nM DEX for 48 hours or left untreated. Cell death was assessed by annexin V staining and Nicoletti assay. GC-treated monocytes showed higher spontaneous survival rates in comparison with untreated cells (Figure 1A-B). In addition, after 6 hours of serum withdrawal, a higher proportion of apoptotic cells was detected in untreated monocytes, whereas GC-treated cells were protected from apoptosis. These differences were even more pronounced after 18 hours (Figure 1C-D). We also tested death receptor–induced apoptosis using anti-Fas antibody. After 18 hours of incubation, an increased proportion of annexin V–positive cells was detectable in control monocytes, while the apoptosis rate in GC-treated cells was significantly diminished (Figure 1E). There was a clear dose-response relation of GC concentrations and the degree of resistance to apoptosis of monocytes challenged by CHX and ActD, inhibitors of transcription and protein biosynthesis, in the range of 10 to 0.1μM DEX (Figure 1F-G). These results demonstrated that during the stimulation with GCs, some common mechanism is activated, which increases survival of monocytes in various stress condition. This antiapoptotic effect is also shared by other GCs. Monocytes, which were cultured for 2 days in the presence of 100nM MP or T and subsequently challenged to apoptosis induced by STS, were also significantly protected from cell death, as assessed by annexin V staining (Figure 1H).

To identify the GC-induced survival mechanisms, we investigated susceptibility of GC-treated monocytes to apoptosis induced by STS in greater detail. Similar to the results obtained with CHX and ActD, significant protection from STS-induced apoptosis was achieved when monocytes were cultured in the presence of 10 to 0.1μM DEX (Figure 1A). Next, cells were cultured for 2 days in the presence or absence of 100nM DEX and subsequently stimulated to undergo apoptosis with STS. Amounts of apoptotic cells were measured after 3 and 6 hours of treatment with STS. We detected a significantly lower amount of apoptotic cells already after 3 hours in GC-treated monocytes compared with untreated monocytes. These differences were even more pronounced after 6 hours (Figure 2B-C). Caspase-9 is an essential component of the mitochondrial pathway of apoptosis and initiates activation of downstream-located caspases like the executioner caspase-3. After 3 and 6 hours, both caspase-9 and caspase-3 activity was diminished in lysates of GC-treated cells compared with controls as shown by cleavage of caspase-specific peptide substrates (Figure 2D-E). Once activated caspase cascades can lead to proteolytic cleavage of caspase-8, which in turn serves as a positive feedback loop for mitochondrial activation. Accordingly, kinetic of caspase-8 activation in course of STS-induced apoptosis was slightly delayed in time, but patterns of caspase-8 activation resembled those observed for caspase-3 and -9 (Figure 2F).

Different members of the mitogen-activated protein (MAP)–kinase family are known to modulate apoptotic mechanisms. We therefore determined the time course of activation of MAP kinases in cell lysates of GC-treated and control monocytes using antibodies specific for phosphorylated ERK1/2, SAPK/JNK, and p38 MAPK by Western blot analysis. There was no activation of JNK1/2 in response to GCs. We detected only basal phosphorylation of p38 MAPK, but no visible differences between GC-treated monocytes and control. No changes were detected in total amounts of JNK1/2 and p38 MAPK. However, we observed a delayed but prolonged phosphorylation of ERK1/2 in the course of GC treatment. Significant phosphorylation was observed first after 12 hours of incubation in the presence of GCs, increased with time, and remained sustained up to 48 hours. The ERK-phosphorylation correlated well with the effect of GCs on apoptosis, which also occurred after prolonged stimulation with GCs. Total ERK1/2 expression did not change throughout the time course of GC treatment (Figure 3A). To clarify mechanism responsible for activation of ERK/MAPK signaling pathway, we performed medium transfer experiments. In this case, monocytes were cultured for 24 hours in conditioned medium from GC-treated or control monocytes. We observed that conditioned medium from GC-treated monocytes was also able to induce ERK/MAPK activation in monocytes, and amounts of phopshorylated ERK1/2 were comparable with those detected in GC-treated monocytes (Figure 3B). These results suggest involvement of some soluble factor and a receptor-mediated pathway in induction of ERK activation. To demonstrate that GC-induced antiapoptotic effect in monocytes is mediated via GC receptor, we involved in our studies steroid receptor blocker RU-486. Pretreatment of monocytes in the presence of 3μM RU-486 30 minutes before stimulation with DEX completely abrogated GC-induced antiapoptotic effect (Figure 3C) and inhibited ERK1/2 phosphorylation in response to GC treatment (Figure 3D).

ERK1/2 phosphorylation is frequently linked to protection from apoptosis. We therefore checked whether there is any relationship between activation of ERK and resistance of GC-treated cells to apoptosis using PD 98 059, a specific MEK inhibitor. Monocytes were preincubated for 30 minutes with 10μM PD 98 059 and subsequently cultured for additional 16 hours in the presence or absence of GCs. Simultaneously, apoptosis was induced with STS for 6 hours. Pretreatment with PD 98 059 significantly inhibited ERK phosphorylation in response to GC treatment. Importantly, suppression of ERK activation also restored susceptibility of GC-treated monocytes to apoptosis (Figure 4A).

We took advantage of highly selective protein kinase C (PKC) inhibitor GF 109203X to check whether ERK phosphorylation in response to GC-treatment is PKC-dependent. Monocytes were preincubated for 30 minutes in the presence of 10μM GF 109203X, subsequently stimulated with GCs, and cultured for additional 16 hours. As assessed by Western blot analysis and annexin V staining, PKC inhibitor neither inhibited ERK phosphorylation nor affected susceptibility of GC-treated cells to STS-induced apoptosis (Figure 4B). Thus, PKC activation is not involved in GC-induced ERK phosphorylation.

Next we examined the phosphorylation status of molecules located upstream or downstream of ERK using antibodies directed against the phospho-form of c-Raf, MEK1/2, and 90-kDa ribosomal S6 kinase (p90RSK). We observed that ERK1/2 activation was accompanied by phosphorylation of c-Raf and MEK1/2, and their phosphorylation pattern mostly resembled that observed for ERK1/2. Also p90RSK, located downstream of ERK1/2, was phosphorylated in cell lysates of GC-treated cells (Figure 5A).

Because the Src family of nonreceptor tyrosine kinases is involved in ERK activation in some biologic systems, we evaluated the role of Src in induction of ERK phopshorylation during GC treatment. PP2 (10μM), a selective inhibitor of the Src family of protein tyrosine kinases, was added 30 minutes before GC stimulation. Cells were subsequently cultured for 16 hours. ERK1/2 phosphorylation in response to GC treatment was significantly suppressed by Src kinase inhibitor (Figure 5B). In parallel, PP2 inhibitor equalized apoptosis rates of control and GC-treated cells in response to STS (Figure 5C-D). These data, as well as results from transfer experiments, point to a receptor driven process underlying the GC-induced antiapoptotic effects.

To identify potential receptor pathways involved in GC-mediated antiapoptotic effects, we checked our previous gene expression microarray analysis of GC-stimulated monocytes for up-regulated transmembrane signaling receptors. In this group, A3AR was the most up-regulated molecule. By using quantitative real-time PCR, we confirmed up-regulated expression of A3AR in monocytes after 4- and 16-hour treatment with GCs (Figure 6A). To investigate the contribution of this receptor to phosphorylation of ERK/MAPK and resistance to apoptosis of GC-treated monocytes, we used MRS1220, a highly selective antagonist for human A3AR. Monocytes were first preincubated for 30 minutes with 0.5μM MRS1220 and subsequently stimulated with GCs for 16 hours. MRS1220 reduced ERK1/2 phosphorylation in GC-treated monocytes substantially even at very low concentrations (Figure 6B). Further, we noticed that A3AR antagonist did not alter susceptibility of control monocytes to apoptosis, but significantly inhibited the antiapoptotic activity of GC treatment (Figure 6C).

In our previous work, we found that GC treatment induces enhanced expression of another transmembrane molecule, formyl peptide receptor (FPR). To verify a contribution of FPR in GC-mediated antiapoptotic effect and activation of ERK1/2, we included Boc-FLFLF, an antagonist that preferentially inhibits activity triggered through the FPR in our studies. Monocytes were pretreated for 30 minutes in the presence of Boc-FLFLF and subsequently stimulated with DEX for 16 hours. As assessed by annexin V staining and Western blot analysis, Boc-FLFLF neither altered the resistance of GC-treated monocytes to STS-induced apoptosis nor inhibited ERK1/2 phosphorylation (Figure 6D-E).

Activated ERK/MAPK signaling pathway is known to phosphorylate and activate an array of downstream targets, including transcription factors, and thereby modulate gene expression. One of the transcription factors activated by Ras/Raf/MEK/ERK cascade and playing an important role in controlling cell growth and vitality is c-Myc. Our gene expression microarray analysis has shown up-regulation of c-Myc expression in response to GCs. Moreover, we detected a significant overrepresentation of c-Myc/Max binding sites (P < .001) among the genes that we identified as up-regulated by 16 hours of GC treatment using Carrie promoter analysis software.¹⁸  Applying quantitative real-time RT-PCR we were able to confirm that expression of c-Myc was significantly increased after 4 hours of treatment with GCs (Figure 6F). To verify the functional significance of increased c-Myc expression for the GC-induced gene-expression pattern we used chromatin immunoprecipitation (ChIP) to detect binding of the c-Myc protein to promoter regions of GC-regulated genes that contain a putative c-Myc binding site. Analyzing promoter regions of GC-induced genes (SAP30 and c-Myc⁴ ) by ChIP, we could demonstrate that GC treatment indeed induced enhanced binding of c-Myc to the SAP30 and Myc promoter in vivo (Figure 6G).

The exact immunosuppressive effects of GCs on different cells of the immune system, particularly monocytes and macrophages, are still not clear. Analyzing the GC-induced expression pattern in human monocytes by genome-wide expression profiling has previously demonstrated that, in monocytes, the main GC effect is not suppression of proinflammatory mediators, but rather induction of an hitherto undescribed anti-inflammatory monocytic phenotype,^4,5  which can actively participate in resolution of inflammation. These GC-induced monocytes produce anti-inflammatory molecules (IL-10, CD163, IL-1R2) and can specifically accumulate at sites of inflammation because of higher migratory capacity to chemotactic stimuli. GC-treated monocytes are able to limit tissue damage because of antioxidative properties and high capacity for phagocytosis of proinflammatory stimuli (ie, microbial agents, particles, and cellular debris) and thus can actively participate in resolution of inflammation.

We now unravel a GC-induced survival mechanism that may result in accumulation of these anti-inflammatory monocytes at sites of inflammation. GC effects on immune cell survival are known to be cell type-specific. GCs induce apoptosis in lymphocytes and eosinophils, while GCs delay neutrophil apoptosis.^20,21  The increased resistance of GC-treated monocytes to apoptosis is crucial for their anti-inflammatory function. Monocytes are generally short-living cells, which even in bloodstream are only present for 2 to 3 days, after which they migrate into the tissue, where they differentiate into macrophages or die spontaneously by apoptosis. Similarly under in vitro culture conditions monocytes undergo rapidly apoptosis when left unstimulated. This form of cell death can be inhibited by addition of growth factors (macrophage colony-stimulating factor [M-CSF], granulocyte macrophage colony-stimulating factor [GM-CSF]) or proinflammatory stimuli, such as LPS, IL-1β, TNFα thus resulting in sustained proinflammatory activities of monocytes.¹⁴  Correspondingly anti-inflammatory monocytes must be protected from apoptosis to complete their differentiation and to be efficient in down-regulation of inflammation.

We now investigated the detailed molecular mechanisms underlying GC-induced resistance of monocytes to apoptosis. Here we demonstrate that monocytes that were incubated in the presence of GCs were markedly protected from apoptosis. This GC-mediated effect was mediated by the nuclear steroid receptor. GC-treated monocytes showed resistance to apoptosis induced by various factors acting both via the mitochondrial, such as STS and drugs, and the death receptor pathway like CD95/Fas. This indicates that GCs activated a common protective mechanism that eventually inhibited activation of the caspase cascade. The activation of both caspase-9, the most apical caspase in the mitochondrial pathway of apoptosis, and caspase-3 was significantly inhibited in GC-treated monocytes. Several studies have shown that phosphorylation of caspase-3 and -9 inhibited their activity and in this manner protected the cells form apoptosis. Different isoforms of PKC, as well as ERK/MAPK and PI3K, were already described to be capable of phosphorylating caspases thereby exhibiting antiapoptotic effects.^22,23  Our studies applying PKC inhibitor GF109203X failed to inhibit GC-induced resistance to apoptosis. In contrast, we identified the ERK/MAPK pathway as principle intracellular target of GC-mediated protection from apoptosis. We demonstrated that GCs induced delayed and sustained activation of Raf/MEK/ERK/p90RSK signaling pathway, but not p38 and SAP/JNK. Activated Raf/MEK/ERK signaling cascade mediated protection of GC-treated monocytes from apoptosis, since its disruption restored monocyte susceptibility to this form of cell death. Raf/MEK/ERK is known to play an important role in modulating growth and survival of hematopoetic cells and is also frequently mutated or inappropriately activated in several cancers.²⁴  The Raf/MEK/ERK pathway has been found to enhance cell survival in response to various apoptotic stimuli including mitochondrial and death receptor-mediated apoptosis and survival factor removal.²⁵  The pro-survival action of this MAP kinase cascade seems to be mediated by dual mechanisms. A transcription-independent pathway involves phosphorylation of proapoptotic Bad and caspase-9 at serine or threonine residues resulting in inhibition of apoptosis.^26,27  Pro-survival activity of the MAP kinase cascade also involves transcription-dependent mechanisms which comprise, among others, activation of 90-kDa ribosomal S6 kinase (p90RSK). The presence of phosphorylated form of p90RSK was sufficient to prevent cells from apoptosis by up-regulation of transcription of genes coding for antiapoptotic proteins.²⁸  The ERK1/2 pathway affects multiple targets that may mediate its pro-survival activity. These include transcription factors that can be either up-regulated or phosphorylated by different members of MAP kinase cascade and promote cell survival. c-Myc is one of the transcription factors localized downstream of ERK1/2. Its enhanced expression and activation in normal tissues correlates with cell growth and protection against apoptosis.^29,30  In the present work we show that GC treatment led to up-regulated transcription of c-Myc and also to its activation. Moreover, c-Myc binding sites were overrepresented among genes induced by GC treatment, indicating an involvement of c-Myc in GC-induced gene expression. Consistently, we have shown enhanced binding of c-Myc to the promoter regions of GC-induced genes with putative c-Myc binding sites. Further studies are required to systematically identify target genes, which are under control of c-Myc or other transcription factors activated by ERK/MAPK and could be involved in protection of GC-treated monocytes against apoptosis. In conclusion, activation of Raf/MEK/ERK/p90RSK in course of GC treatment led to enhanced survival of monocytes in stress conditions. This represents a novel mechanism promoting prolonged survival of these anti-inflammatory cells.

Looking for the trigger responsible for antiapoptotic activities of GCs, we could also show for the first time that treatment of monocytes with GCs resulted in up-regulated expression of A3AR and that phosphorylation of ERK/MAPK and enhanced resistance of monocytes to apoptosis was a consequence of ligand binding by A3AR. In the present study, we clarified the signaling pathway induced by ligation of A3AR in monocytes. A3AR generally couples to the G_i and G_o class of heterotrimeric G proteins, thereby inhibiting cAMP level.³¹  A3AR mediates cell signaling via release of βγ subunits of G-proteins and subsequent Src nonreceptor kinases activation, followed by phosphorylation of the adapter molecules Shc, its interaction with Grb2 and stimulation of Ras/Raf/MEK/ERK signaling pathway.³²  A3AR has also been linked to cell activation via phospholipase C (PLC) and phospholipase D (PLD) and subsequent activation of PKC.³³  PKC is known to trigger ERK/MAPK independently of Ras through direct interaction with Raf.³⁴  In our studies, GC-induced activation of ERK/MAPK pathway was abrogated only by Src inhibitor PP2, whereas PKC inhibitor influenced neither ERK/MAPK phosphorylation nor cell resistance to apoptosis. Therefore we proposed that ERK1/2 activation through A3AR in course of treatment of monocytes with GCs is mediated only by Src-dependent mechanisms, but not by PLC-dependent activation of PKC (Figure 7).

Increased expression and triggering of A3AR is the main, but not the only mechanism underlying GC-induced antiapoptotic effect. In our previous work, we have shown that molecules with antioxidative functions, especially genes involved in glutathione metabolism, are up-regulated in GC-treated monocytes. Consistently, production of reactive oxygen species in course of STS-induced apoptosis was almost completely inhibited in GC-treated monocytes and this effect was at least partly due to the fact that GC-treated monocytes have a higher capacity to replenish loss of intracellular glutathione promoting cell survival.⁴ 

In our previous work, we have also identified up-regulated expression of FPR1 in response to GC treatment. Interestingly, annexin A1, which is one of the most important mediators of GC-mediated effects,^35-37  has been suggested to signal through receptors belonging to the FPR family.^38,39  However, downstream signaling after annexin A1 binding to its receptor involves rapid and transient activation of ERK1/2 and p38, which stays in strict contrast to the prolonged ERK-activation responsible for the antiapoptotic effect described in our study.^37,40  Moreover, only an N-terminal peptide of annexin A1 (Ac1-25) has been shown to activate all FPR receptors, binding of the physiologically relevant annexin A1 has been reported only for FPR2, which was not regulated by GCs in our model.^37,41  Finally, annexin A1 has been shown to promote apoptosis in neutrophils showing a surprising converse effect to that of GCs.^37,42  These published data point against a role of annexin A1 in GC-induced antiapoptotic mechanism on monocytes and support the hypothesis that annexin A1 mediates a very specific subset of the whole spectrum of anti-inflammatory GC-effects.³⁷  Accordingly, inhibition studies using the inhibitor Boc-FLFLF did not influence GC-induced resistance of monocytes to apoptosis.

Adenosine has been described as a potent endogenous anti-inflammatory agent.^31,43  Adenosine binding to the A3AR enhances migration but inhibits major histocompatibility complex (MHC) class II expression, TNFα production, and oxidative burst in phagocytes,^43-48  which is pretty much similar to the effects described by our group for GC treatment of monocytes.⁴  High levels of adenosine are present in inflamed tissues. Thus, promoting the survival of a potent anti-inflammatory monocyte phenotype via A3AR is an exciting new mechanism by which adenosine can exert anti-inflammatory actions. The clinical relevance of our findings is underlined by the fact that several potent anti-inflammatory drugs like tacrolimus, sulfasalazin, and especially methotrexate (MTX) have been demonstrated to exert at least some of their immunomodulating effects via induction of adenosine.^43,44  There is convincing evidence that adenosine receptor antagonists selective for A3AR are potential therapeutics for asthma and chronic lung disease.⁴⁹  Accordingly, data obtained from A3AR −/− mouse confirm that this receptor plays an important role in anti-inflammatory responses.⁵⁰  Thus, the spontaneous phenotype of the A3AR −/− mouse is in good accordance with our data, but unfortunately precludes use of this model for confirmation of our experiments in vitro in experimental inflammation in this murine model. The A3AR and its effect on survival of anti-inflammatory monocytes may thus be a common mechanistic explanation for several anti-inflammatory actions of different immunosuppressive drugs. Thus promoting survival of anti-inflammatory monocytes may be a physiologically relevant mechanism whenever high amounts of adenosine are present, either by endogenous mechanisms or due to pharmacologic intervention.

In summary, we deciphered a novel molecular pathway responsible for GC-mediated anti-inflammatory effects. Promoting the survival of anti-inflammatory monocytes via specific activation of A3AR or its downstream signaling pathways may thus be a novel therapeutic strategy to modulate undesirable inflammation in autoimmune disorders via induction of resolution of inflammation rather than unspecific suppression of disease activity.

We thank A. Dick for excellent technical assistance.

This study was supported by research funding from the Interdisciplinary Center for Clinical Research (project Ro2/012/06) and the German Ministry for Research to J.R. (BMBF/DLR Fkz: 01KI07100) and from Innovative Medical Research of the University of Muenster to J.E. and D.V. (projects EH120605 and VI220708).

Contribution: K.B. and J.E. designed and performed the research and wrote the paper; K.T., M.A., and J.K. performed the research; D.V. analyzed the data; and J.R. designed the research and wrote the paper.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Prof Johannes Roth, Institute of Immunology, University of Muenster, Roentgenstr 21, D-48149 Muenster, Germany; e-mail: rothj@uni-muenster.de.