|Home | About | Journals | Submit | Contact Us | Français|
Macrophages are activated by IFN-γ, a proinflammatory and proatherogenic cytokine that mediates its downstream effects primarily through STAT1. IFN-γ signaling induces phosphorylation of two STAT1 residues: Tyr701 (Y701), which facilitates dimerization, nuclear translocation, and DNA binding; and Ser727 (S727), which enables maximal STAT1 transcription activity. Immunosuppressive molecules such as adenosine in the cellular microenvironment can reduce macrophage inflammatory and atherogenic functions through receptor-mediated signaling pathways. We hypothesized that adenosine achieves these protective effects by interrupting IFN-γ signaling in activated macrophages. This investigation demonstrates that adding adenosine to IFN-γ-stimulated murine RAW 264.7 and human THP-1 macrophages results in unique modulation of STAT1 serine and tyrosine phosphorylation events. We show that adenosine inhibits IFN-γ-induced STAT1 S727 phosphorylation by >30% and phosphoserine-mediated transcriptional activity by 58% but has no effect on phosphorylation of Y701 or receptor-associated JAK tyrosine kinases. Inhibition of the adenosine A3 receptor with a subtype-specific antagonist (MRS 1191 in RAW 264.7 cells and MRS 1220 in THP-1 cells) reverses this adenosine suppressive effect on STAT1 phosphoserine status by 25–50%. Further, RAW 264.7 A3 receptor stimulation with Cl-IB-MECA reduces IFN-γ-induced STAT1 transcriptional activity by 45% and STAT1-dependent gene expression by up to 80%. These data suggest that A3 receptor signaling is key to adenosine-mediated STAT1 modulation and anti-inflammatory action in IFN-γ-activated mouse and human macrophages. Because STAT1 plays a key role in IFN-γ-induced inflammation and foam cell transformation, a better understanding of the mechanisms underlying STAT1 deactivation by adenosine may improve preventative and therapeutic approaches to vascular disease.
Activated macrophages are central to inflammatory and atherogenic processes through their secretion of proinflammatory factors and uptake of modified lipoproteins within the artery wall. The cytokine IFN-γ has a well-documented role in the activation of macrophages and their promotion of atherosclerosis. IFN-γ signaling mediates macrophage immunological function and lipid metabolism by up-regulating the expression of proinflammatory mediators (1–3), scavenger receptors (4), and intracellular cholesterol-trafficking genes (5). IFN-γ also inhibits expression of genes involved in reverse cholesterol transport (6, 7), resulting in reduced cholesterol efflux and the transformation of macrophages into lipid-laden foam cells. The ability of IFN-γ to induce both macrophage activation and cholesterol imbalance suggests that this cytokine may serve as a critical link between vascular inflammation and development of the earliest atherosclerotic lesions (8).
Cell surface binding of IFN-γ induces dimerization of its receptor subunits (IFNGR1 and IFNGR2) and subsequent activation of the receptor-associated JAK kinases 1 and 2 (9). Activated JAKs phosphorylate the intracellular domain of IFNGR1, creating a docking site that recruits STAT1 to the receptor (10, 11). STAT1 is phosphorylated on Tyr701 (Y701) and then undergoes dimerization through reciprocal Src homology-2-phosphotyrosine interactions (11, 12). STAT1 homodimers translocate to the nucleus and regulate gene expression by binding γ-activated sequence (GAS)3 elements in the promoters of IFN-γ-responsive genes (13). During early activation, a second, independent phosphorylation event occurs at the STAT1 Ser727 (S727) motif. STAT1 S727 phosphorylation is required for nearly 80% of IFN-γ-induced transcriptional activity but does not mediate DNA binding or nuclear translocation (9, 14, 15). Because a significant portion of the IFN-γ-induced biological response results from signaling through STAT1, inhibition of this transcription factor may represent a promising therapeutic strategy to reduce macrophage activation and its role in atherogenesis (12, 16).
The purine nucleoside adenosine has emerged as an important endogenous regulator of macrophage activation and function. Under conditions of stress and inflammation, local extra-cellular concentrations of adenosine rise as a result of ATP catabolism and cell secretion (17, 18). Most of the known immunomodulatory effects of adenosine are mediated through its interaction with specific cell surface G protein-coupled receptors. Macrophages have been reported to express all four of the adenosine receptor subtypes, A1, A2A, A2B, and A3 (19, 20). Ligation of one or more of these receptors suppresses the production of proinflammatory factors, stimulates expression of reverse cholesterol transport proteins, and inhibits macrophage foam cell formation (20–22). Previous studies using bacterial LPS as a stimulus have shown that adenosine can exert its anti-inflammatory effects by suppressing activation of NF-κB and ERK1/2 pathways (23, 24). However, the mechanism by which adenosine reduces macrophage activation in IFN-γ-stimulated cells has not yet been elucidated.
This investigation addresses the potential of adenosine to modulate IFN-γ-induced signal transduction and macrophage activation. In this study, we show that adenosine treatment reduces the expression of many IFN-γ-regulated genes implicated in inflammation and atherogenesis through the modulation of STAT1 activation. We demonstrate that adenosine signaling reduces STAT1 serine phosphorylation but has no effect on STAT1 tyrosine phosphorylation status or on the activation of JAK tyrosine kinases. Distinct modulation of the two STAT1 phosphorylation sites suggests that adenosine inhibits IFN-γ-induced macrophage activation by blocking the S727-controlled STAT1 transcriptional activity rather than the Y701-regulated DNA binding. Our data indicate that these downstream effects result from adenosine signaling through the A3 receptor subtype. These findings point to an important immunosuppressive role for adenosine in macrophage-mediated inflammation. A better understanding of the cross-talk between adenosine and JAK-STAT signaling pathways may provide guidance in the design of novel preventative or therapeutic interventions to vascular disease.
Adenosine, erythro-9-(2-hydroxy-3-nonyl)adenine (EHNA), 2-chloro-N6-cyclopentyladenosine (CCPA), 2-p-(2-carboxyethyl)phenethylamino-5′-N-ethylcarboxamidoadenosine (CGS21680), N-ethylcarboxamidoadenosine (NECA), 2-chloro-N6-(3-iodobenzyl)adenosine-5′-N-methyluronamide (Cl-IB-MECA), 8-cyclopentyl-1,3-dipropylxanthine (CPX), 5-amino-7-(β-phenylethyl)-2-(8-furyl)pyrazolo(4,3-3)-1,2,4-triazolo(1,5-c)pyrimidine (SCH 58261), alloxazine, 3-ethyl-5-benzyl-2-methyl-4-phenylethynyl-6-phenyl-1,4-(±)-dihydropyridine-3,5-dicarboxylate (MRS 1191), 9-chloro-2-(2-furanyl)-5-((phenylacetyl)amino)-[1,2,4]triazolo[1,5-c]quinazoline (MRS 1220), and PMA were purchased from Sigma-Aldrich. Mouse rIFN-γ was obtained from Calbiochem. Human rIFN-γ was obtained from R&D Systems.
The mouse macrophage cell line RAW 264.7 was obtained from the American Type Culture Collection (ATCC) and grown in DMEM supplemented with 10% FBS, penicillin (100 U/ml), and streptomycin (100 U/ml). Cells were cultured at 37°C in a humidified atmosphere of 5% CO2.
RAW 264.7 cells were treated for 4 h with medium (untreated control), IFN-γ (1000 U/ml), adenosine (300 μM), IFN-γ plus adenosine, IFN-γ plus one of the selective adenosine receptor agonists (30 μM), or IFN-γ plus adenosine plus MRS 1191 (a selective A3 adenosine receptor antagonist, 50 μM). EHNA (45 μM) was added in combination with all adenosine treatments to inhibit adenosine degradation by ectodeaminases on the macrophage cell surface. Receptor-specific agonists were added to cells 30 min before treatment with IFN-γ, and MRS 1191 was added to cells 20 min before treatment with IFN-γ plus adenosine. In experiments using both an A3 receptor agonist and antagonist, MRS 1191 was added to cells 20 min before administration of Cl-IB-MECA (A3 receptor-specific agonist) and 50 min before treatment with IFN-γ. A DMSO control was included for all experiments and showed no effect on the responses studied.
The human monocyte-macrophage cell line THP-1 was obtained from ATCC and grown in RPMI 1640 supplemented with 10% FBS, 0.05 mM 2-ME, 10 μg/ml gentamicin, and 0.25 μg/ml amphotericin B. Cells were cultured as monocytes in suspension at 37°C in a humidified atmosphere of 5% CO2.
Before use in experimentation, THP-1 monocytes were differentiated into macrophages via treatment with PMA (100 nM) for 24 h followed by an additional 24-h incubation period in fresh medium. THP-1 monocyte-derived macrophages were then treated with medium (untreated control), human rIFN-γ (1000 U/ml), IFN-γ plus adenosine (100 μM), IFN-γ plus adenosine (300 μM), or IFN-γ plus adenosine (100 μM) plus one of the selective adenosine receptor antagonists (10 μM) for 4 h. MRS 1220, a selective human A3 receptor-specific antagonist, was used in place of MRS 1191 for THP-1 treatment. Receptor antagonists were added to cells 30 min before stimulation with IFN-γ plus adenosine. EHNA (45 μM) was added in combination with all adenosine treatments.
Total RNA was prepared from cells using the RNeasy kit (Qiagen) and subsequently quantified using a Nanodrop ND-1000 spectrophotometer (NanoDrop Technologies). Total RNA was pooled from two independent experiments and GeneChip analyses were performed as previously reported (25). An 8-μg aliquot of total RNA from each pooled sample was used for preparation of biotin-labeled RNA, as described in the Affymetrix One Cycle Sample Preparation protocol (Affymetrix). Biotin-labeled RNA samples were hybridized to mouse genome 430A 2.0 high-density oligo-nucleotide arrays. The hybridization, washing, labeling, and scanning of the GeneChips was performed as described in the Affymetrix protocols by the Microarray Core Facility in the UC Davis Genome and Biomedical Sciences Facility. The data were analyzed by GeneChip Operating System 1.4 (Affymetrix). The upper p value limit for statistically reliable detection of an mRNA was 0.05, independent of its signal intensity (usually >10), given that the detected mRNA can be confirmed by an independent method such as quantitative real-time PCR (qRT-PCR). Complete microarray data are available at http://www.ncbi.nlm.nih.gov/geo/ (Gene Expression Omnibus accession number GSE14612).
Total RNA was prepared and quantified as described above. Total RNA (5 μg) was used to generate cDNA following the protocol for Superscript II (Invitrogen). Equal amounts of cDNA were used in duplicate and amplified with the SYBR Green I Master Mix (Applied Biosystems). Real-time detection of PCR was performed using the GeneAmp 7700HT Sequence Detection System (Applied Biosystems). Primer sequences are as follows: IFN regulatory factor (IRF) 1 forward, 5′-AT TCAGGCCATTCCTTGTGC-3′, and IRF1 reverse, 5′-GCAAGAACGG GTCAGAGACC-3′; inducible NO synthase (iNOS) forward, 5′-GATG GTCCGCAAGAGAGTGC-3′, and iNOS reverse, 5′-AACGTAGACC TTGGGTTTGCC-3′; CD36 forward, 5′-TCCAGCCAATGCCTTTGC-3′, and CD36 reverse, 5′-TGGAGATTACTTTTCAGTGCAGAA-3′; GAPDH forward, 5′-GCAACAGGGTGGTGGACCT-3′, and GAPDH reverse, 5′-GGATAGGGCCTCTCTTGCTCA-3′.
Whole-cell lysates were prepared on ice using lysis solution (20 mM HEPES (pH 7.9), 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 10% glycerol, 2× protease, and 2× phosphatase inhibitor mixtures). Protein concentrations were determined with the BCA Protein Assay (Pierce). Lysate protein was subjected to 7.5% SDS-PAGE and electrophoretically transferred to a nitrocellulose membrane. After blotting, membranes were probed with primary Ab at 4°C overnight. The following primary Abs were used: anti-phospho-STAT1 (S727; Cell Signaling Technology); anti-phospho-STAT1 (Y701; Santa Cruz Biotechnology); anti-phospho-JAK1 (Y1022/1023; Santa Cruz Biotechnology); and anti-phospho-JAK2 (Y1007/1008; Cell Signaling Technology). Membranes were incubated with HRP-conjugated secondary anti-rabbit Ab (Amer-sham) for1hat room temperature. Blots were developed with the ECL detection system according to the manufacturer's instructions (Amer-sham). Blots are representative of three separate experiments. To ensure equal loading of protein, all blots were stripped with Restore Buffer (Pierce), washed, blocked, and probed with anti-STAT1 (Santa Cruz Biotechnology), anti-JAK1 (Santa Cruz Biotechnology), or anti-JAK2 (Cell Signaling Technology) as described.
RAW 264.7 macrophages were transfected with a STAT1 homodimer reporter vector (GAS-Luc), negative control, or positive control construct from the Cignal GAS Reporter (luc) kit (SA Biosciences) using Attractene transfection reagent (Qiagen) according to the manufacturer's instructions. After 24 h, the transfection solution was aspirated and replaced with fresh DMEM. Cells were then treated with IFN-γ, adenosine, or adenosine receptor agonists as described above, and luciferase activities were measured at various time points using the Dual-Glo Luciferase assay system (Promega). Promoter activity was normalized to Renilla luciferase activity and expressed as fold change from control.
Nuclear protein fractions were isolated from treated RAW 264.7 cells according to the protocol outlined in the Active Motif Nuclear Extract Kit (Active Motif) and protein concentrations determined as described above. Nuclear extract (12.5 μg) from each sample was used in the TransAm STAT Family Transcription Factor Assay Kit (Active Motif). COS-7 (IFN-γ-stimulated) nuclear extract (Active Motif) was used as a positive control for STAT1α as recommended by the manufacturer. The binding, washing, and colorimetric reaction steps were performed as described in the Active Motif protocol. STAT1 DNA binding activation is represented as OD measurement at 450 nm with a reference wavelength of 655 nm.
Values in the figures and text are expressed as mean ± SEM. Data were analyzed by one- or two-way ANOVA followed by the Student-Newman-Kuels post hoc test. Statistical significance of paired data was determined by Student's t test. Differences were considered to be statistically signifi-cant at p < 0.05.
To determine the effect of adenosine on macrophage activation, we performed gene expression profiling in RAW 264.7 cells after a 4-h treatment with medium, IFN-γ, or IFN-γ plus adenosine. Total RNA was pooled groupwise and processed for microarray analysis. We found stimulation with IFN-γ alone caused a ±≥2-fold change in the expression of 430 genes compared with medium control. These results are consistent with previous work that showed >400 IFN-γ-sensitive genes in the RAW 264.7 mouse macrophage cell line (26). When we compared results from IFN-γ vs IFN-γ plus adenosine treatment groups, we found that ~2700 genes were differentially expressed by ±≥2-fold, and 97% of these genes (2688 genes) were repressed in the IFN-γ plus aden-osine group compared with IFN-γ treatment alone. Genes suppressed by the addition of adenosine included inflammatory mediators, scavenger receptors, and lipid storage enzymes, as well as transcription factors and other signaling molecules (Table I).
To confirm results from the microarray analysis, we performed qRT-PCR on select genes known to be involved in atherogenesis and regulated by STAT1 activity. Fig. 1 illustrates the effects of a 0 –1000 μM concentration range of adenosine in IFN-γ-activated RAW 264.7 cells. Our results indicate that adenosine suppresses IFN-γ-induced gene expression in a concentration-dependent manner. Genes such as IRF1 and iNOS contain a STAT1 binding sequence in their promoter region and exhibit altered gene expression profiles depending on whether or not activated STAT1 is present (27–30). The transcription factor IRF1 was induced nearly 50-fold after treatment with IFN-γ alone (0 μM adenosine) and was significantly reduced 30 – 65% with increasing concentrations of adenosine ( p < 0.05 to p < 0.001; Fig. 1A). iNOS, known to be strongly expressed in activated macrophages and macrophages isolated from advanced atherosclerotic plaque (28), showed a robust response to IFN-γ (59 ± 9-fold increase over control) and a similar but less pronounced adenosine repressive effect with concentrations >150 μM (Fig. 1B). Although the scavenger receptor gene CD36 does not contain a known consensus STAT1 binding motif, previous work has shown that STAT1 can regulate CD36-mediated foam cell formation by an indirect mechanism (31). In Fig. 1C, we demonstrate a dose-dependent reduction in CD36 expression in IFN-γ-stimulated cells ranging from 33% (adenosine, 50 μM; p < 0.05) to 68% (adenosine, 1000 μM; p < 0.001). Because IFN-γ-induced expression of these genes is regulated by STAT1 activity, it is possible that adenosine exerts its effects by modulating JAK-STAT1 signaling rather than by altering gene transcription directly.
Full activation of STAT1 by IFN-γ requires phosphorylation at both Y701 and S727. Y701 phosphorylation is essential for STAT1 dimerization, nuclear translocation, and DNA binding (15). S727 phosphorylation enhances the transcriptional activity of tyrosine-phosphorylated STAT1 by up to 80% (15). To determine whether adenosine modulates these two distinct phosphorylation events differently, we performed immunoblot analysis on RAW 264.7 whole-cell lysates with phosphoserine- and phosphotyrosine-specific STAT1 Abs. The overall level of serine phosphor-ylation was significantly reduced ( p < 0.01) in macrophages treated with IFN-γ plus adenosine compared with cells stimulated with IFN-γ alone (Fig. 2). In whole-cell lysates from IFN-γ-treated cells, a 16% increase in STAT1 serine phosphorylation above baseline levels (Fig. 2, dotted line) was observed after 10 min, reaching 41% over control at 20 min and 70% over control at 240 min poststimulation (Fig. 2B). In contrast, the phosphoserine band intensity measured in IFN-γ- plus adenosine-treated cells increased by significantly less, only 12 and 37% over baseline at 20 and 240 min, respectively.
As shown in Fig. 3, adenosine treatment had no effect on whole-cell Y701 phosphorylation status. IFN-γ stimulation resulted in robust tyrosine phosphorylation of STAT1 above baseline levels at 10 min and at subsequent measurement points, but there was no significant difference between treatment groups at any time (Fig. 3). The lack of adenosine effect on STAT1 Y701 phosphorylation levels suggests that an adenosine deactivation target is independent of the receptor-associated JAK tyrosine kinase pathway. To confirm these results, we subjected whole-cell lysates to immunoblot analysis using phospho-specific Abs to JAK1 and JAK2. As expected, we found that IFN-γ led to a rapid rise in both JAK1 and JAK2 phosphorylation band density above baseline levels, and this IFN-γ-induced JAK activation was not altered by adenosine treatment at any time point (data not shown). Adenosine had no effect on STAT1 or JAK phosphorylation levels in unstimulated cells, and no change in the total amount of protein in either treatment group was detected. Taken collectively, our data imply that any adenosine effect on STAT1 must occur downstream of the JAK-receptor complex.
To investigate whether the adenosine-mediated reduction in STAT1 S727 phosphorylation from Fig. 2 has functional consequences, we measured STAT1 transcriptional activity in RAW 264.7 macrophages transfected with a GAS reporter construct. The GAS reporter included a 40:1 mixture of an inducible STAT1-responsive luciferase gene and a constitutively expressing Renilla luciferase gene used as an internal control. As shown in Fig. 4, an IFN-γ challenge increased expression of the GAS-luciferase construct by 2-fold over control cells starting at 60 min poststimulation. Including adenosine with IFN-γ treatment delayed any measureable increase in STAT1 activation by >60 min, and led to significantly reduced STAT1 activity at 60, 120, and 240 min post-stimulation compared with cells treated with IFN-γ alone ( p < 0.05; Fig. 4). In fact, maximal STAT1 activity observed at 240 min poststimulation was 58% less in IFN-γ- plus adenosine-treated cells compared with cells exposed to only IFN-γ (6.49 ± 0.85 vs 15.57 ± 1.13, respectively). Our data suggest that adenosine blocks IFN-γ-induced STAT1 transcriptional activity, providing a functional correlate for its effect on phosphorylation of STAT1 at S727.
The similar overall STAT1 pY701 levels in both IFN-γ and IFN-γ plus adenosine treatment groups suggest that adenosine has little impact on phosphotyrosine-regulated events such as STAT1 dimerization, nuclear translocation, and DNA binding. To determine whether adenosine alters the DNA binding function of activated STAT1, we used nuclear extract from IFN-γ- and IFN-γ-plus adenosine-treated RAW 264.7 cells in a transcription factor ELISA assay. As expected, STAT1 DNA binding activity measured in extract-only samples was significantly greater in all cells treated with IFN-γ (±adenosine) compared with cells treated with medium or adenosine alone ( p < 0.05; Fig. 5). The observed pattern in both IFN-γ and IFN-γ plus adenosine treatment groups is consistent with results from the phosphotyrosine-specific immunoblot analysis described previously. These data confirm that STAT1 DNA-binding ability is not disrupted by adenosine treatment. Furthermore, these results suggest that adenosine does not impede STAT1 dimerization because only the stable dimer is capable of subcellular localization to the nucleus and binding DNA (9, 32).
To confirm the specificity of the assay, we repeated the experiment with the same nuclear extract in the presence of competitor oligonucleotides containing wild-type STAT1 consensus binding sites. As shown in Fig. 5, competitive oligos reduced STAT1 DNA binding in nuclear extract from IFN-γ- and IFN-γ- plus adenosine-stimulated cells by 80 –90% at all measured time points ( p ≤ 0.001). These results suggest that nuclear extract measurements represent specific STAT1 binding to its consensus site rather than nonspecific Ab artifact or incomplete wash steps.
Extracellular adenosine exerts its physiological actions primarily by occupation of one or more cell surface receptors (17, 33). To determine through which receptor adenosine modulates STAT1 activation, RAW 264.7 cells transfected with a GAS reporter construct were pretreated with one of the adenosine receptor-specific agonists and subsequently exposed to IFN-γ for 4 h. Our results show that an A3, but not an A1 or A2, adenosine receptor-specific agonist attenuated STAT1 activity in IFN-γ-treated cells. Pretreatment with CCPA (A1 receptor agonist), NECA (non-specific A1 and A2 receptor agonist), and CGS21680 (A2A receptor agonist) all resulted in a 14- to 15-fold increase in STAT1 activity over control cells, levels comparable to what was observed in cells treated with IFN-γ alone (Fig. 6, A–C). In contrast, cells stimulated with IFN-γ plus the A3 receptor agonist Cl-IB-MECA, showed a 45% reduction in STAT1 activity compared with cells treated with IFN-γ alone (8.9 ± 0.95- vs 16.1 ± 1.8-fold increase from control, respectively; p < 0.05; Fig. 6D). A DMSO vehicle control was also included and exhibited no effect on IFN-γ-induced STAT1 activation (data not shown). These results suggest that adenosine-mediated suppression of STAT1 transcriptional activity occurs through the A3 receptor.
To further explore a role for the A3 receptor in STAT1 modulation, we exposed RAW 264.7 macrophages to the A3 receptor-specific antagonist, MRS 1191, for 20 min before treatment with adenosine and IFN-γ. After 4 h, we collected whole-cell lysates for immunoblot analysis with phosphoserine- and phosphotyrosine specific STAT1 Abs. The IFN-γ-induced increase in STAT1 S727 phosphorylation band intensity was reduced by 30% with adenosine treatment ( p < 0.05; Fig. 7A). MRS 1191 significantly reversed this adenosine suppressive effect ( p < 0.05), resulting in similar STAT1 serine phosphorylation band intensity levels (2.02 ± 0.31-fold increase over control) to those from cells treated with IFN-γ alone (2.31 ± 0.08-fold increase over control). These results suggest that A3 receptor signaling plays a key role in mediating the inhibition of STAT1 S727 phosphorylation by adenosine.
As shown in Fig. 7B, adenosine signaling had no effect on whole-cell STAT1 Y701 phosphorylation status. Tyrosine phosphorylation of STAT1 increased significantly above control levels in all cells stimulated with IFN-γ ( p < 0.05), including those cells treated with adenosine or adenosine plus MRS 1191. The absence of an adenosine effect on STAT1 Y701 phosphorylation status provides further evidence that any A3 receptor-mediated adenosine action is uniquely targeted to the STAT1 S727 residue.
Finally, we measured the expression of two STAT1-dependent genes in activated RAW 264.7 macrophages following A3 receptor-specific stimulation and inhibition. Our results show that stimulation of the A3 adenosine receptor subtype with Cl-IB-MECA 30 min before an IFN-γ challenge reduced expression of IRF1 by 18% (from a 43.5 ± 1.59-fold to a 35.6 ± 1.12-fold increase over control; p < 0.001; Fig. 8A) and iNOS by 80% (from a 32.7 ± 2.69-fold to a 6.58 ± 0.33-fold increase over control; p < 0.001; Fig. 8B). Pretreating cells with MRS 1191 reversed this effect and restored expression of IRF1 and iNOS to levels comparable with those measured in cells treated with IFN-γ alone. These results, obtained using both an A3 receptor-specific agonist and antagonist, suggest that A3 receptor signaling is both necessary and sufficient to mediate suppression of these STAT1-dependent genes by adenosine.
To test whether the adenosine-mediated reduction in STAT1 S727 phosphorylation is mouse or cell type specific, we performed immunoblot analysis on whole-cell lysates from the human THP-1 cell line. Immediately before each experiment, we differentiated THP-1 monocytes into macrophages via PMA treatment (100 nM) for 24 h followed by a 24-h media washout period. Fig. 9 illustrates the effects of a 100 and 300 μM dose of adenosine on STAT1 phosphorylation in IFN-γ-activated THP-1 macrophages. Our results indicate that adenosine suppresses IFN-γ-induced STAT1 S727 phosphorylation in a concentration-dependent manner but has no effect on STAT1 phosphotyrosine status. The 2.44 ± 0.11-fold increase in STAT1 phosphoserine band intensity induced by IFN-γ stimulation was decreased by 34.2 and 48.1% with 100 and 300 μM adenosine treatments, respectively ( p < 0.05; Fig. 9A). Because the 100 μM dose of adenosine affected such a significant suppressive response on STAT1, we used this lower adenosine concentration for all future experiments in THP-1 cells. In contrast to STAT1 phosphoserine status, neither dose of adenosine altered the IFN-γ-induced rise in STAT1 phosphotyrosine levels. All three treatments with IFN-γ (IFN-γ, IFN-γ plus 100 μM adenosine, and IFN-γ plus 300 μM adenosine) triggered a >2.5-fold increase in STAT1 Y701 band intensity over untreated THP-1 cells ( p < 0.05; Fig. 9B).
We next investigated the adenosine receptor subtype responsible for mediating STAT1 deactivation in human macrophages by exposing THP-1 cells to adenosine receptor-specific antagonists (10 μM) for 30 min before treatment with adenosine (100 μM) and IFN-γ (1000 U/ml). After 4 h, we collected whole-cell lysates for immunoblot analysis with phosphoserine- and phosphotyrosinespecific STAT1 Abs. As observed previously, the IFN-γ-induced increase in STAT1 S727 phosphorylation band intensity was reduced by 36% (from 3.01 ± 0.23-fold over control to 1.92 ± 0.17-fold over control; p < 0.05) with adenosine treatment (Fig. 10). We observed similar inhibition of STAT1 S727 phosphorylation in cells treated with A1,A2A, and A2B receptor-specific antagonists (CPX, SCH 58261, and alloxazine, respectively), suggesting that these three receptor subtypes do not play an important role in adenosine-mediated STAT1 deactivation (Fig. 10). In contrast, the addition of a human A3 receptor-specific antagonist, MRS 1220, significantly reversed the suppressive effect of aden-osine. A3 receptor inhibition enabled STAT1 phosphoserine band intensity levels to reach 2.91 ± 0.11-fold over control, similar to levels measured with IFN-γ alone and 30 –50% higher than levels measured in cells treated with IFN-γ plus adenosine with or without one of the other receptor-specific antagonists. These results suggest that A3 receptor signaling plays a role in the adenosine-mediated suppression of STAT1 S727 phosphorylation in human, as well as mouse, macrophages.
As shown in Fig. 11, adenosine signaling had no effect on whole-cell STAT1 Y701 phosphorylation status. Tyrosine phosphorylation of STAT1 increased significantly above control levels in all cells stimulated with IFN-γ ( p < 0.001) despite the addition of adenosine or adenosine plus receptor-specific antagonists. The lack of adenosine effect on STAT1 Y701 phosphorylation supports our previous findings in both RAW 264.7 and THP-1 cells, suggesting that any A3 receptor-mediated adenosine action is STAT1 serine site specific.
The results presented in this study support a role for adenosine in suppressing IFN-γ-stimulated inflammation and macrophage activation and, further, provide a novel mechanism behind these adenosine-mediated effects. To our knowledge, this is the first demonstration that adenosine treatment can modulate IFN-γ-induced gene expression by reducing STAT1 serine phosphorylation and phospho-S727-mediated transcriptional activity. We show that inhibition of STAT1 phosphorylation occurs only at the serine residue, whereas adenosine has no effect on tyrosine phosphorylation status or function. Furthermore, our data are the first to illustrate that the A3 receptor subtype plays a principal role in mediating the STAT1 modulation and anti-inflammatory action of adenosine following an IFN-γ challenge.
IFN-γ regulates macrophage activation and intracellular cholesterol accumulation by inducing the expression of genes involved in inflammation and lipid uptake. Each of the genes selected for analysis in this study has been implicated in the pathogenesis of atherosclerosis as either an inflammatory mediator (iNOS), a scavenger receptor contributing to foam cell formation (CD36), or a transcription factor (IRF1) important for sustaining secondary IFN-γ transcriptional responses. Our results showing that adeno-sine reduces expression of these genes suggest a beneficial role for this nucleoside in suppressing IFN-γ-regulated inflammation and macrophage activation. These data are consistent with many other studies demonstrating that exogenous adenosine has significant anti-inflammatory and antiatherogenic activity (22, 34–36).
The marked decrease in IRF1 and iNOS mRNA implies that adenosine mediates its anti-inflammatory effects through a pre-translational mechanism (19). For this reason, we investigated the effect of adenosine on STAT1, one of the key mediators of IFN-γ signaling. IRF1 (30) and iNOS (27) both contain a STAT1-binding sequence in their promoter regions, making these genes likely candidates for direct STAT1-mediated control. Maximal IFN-γ-induced gene transcription occurs from full STAT1 activation following the separate phosphorylations of specific tyrosine and serine residues. The differential effects resulting from these independent phosphorylation events prompted our detailed analysis of the mechanism underlying the observed adenosine suppression of IFN-γ-regulated genes. With the JAK tyrosine kinase activity intact and robust STAT1 Y701 phosphorylation detected in both IFN-γ and IFN-γ plus adenosine treatment groups, we conclude that adenosine has little impact on the initial STAT1 recruitment event or phosphotyrosine-mediated functions. This conclusion is supported by our transarray data from RAW 264.7 cells showing similar STAT1 DNA binding activity across all time points in both treatment groups. In contrast, we show that adenosine treatment both delays STAT1-responsive promoter activity and markedly attenuates the maximal STAT1 activity in IFN-γ-stimulated macrophages, providing a functional correlate to the reduction in STAT1 S727 phosphorylation observed in adenosine-treated cells.
Previous studies using STAT1 S727 mutants (S7273 → A727) or serine kinase inhibitors have emphasized a critical role for serine phosphorylation in mediating the biological impact of IFN-γ (9, 15). It is likely that adenosine exerts its immunosuppressive effects through inhibition of a serine kinase upstream of STAT1 rather than through enhanced phosphatase activity because no alteration in STAT1 Y701 phosphorylation was observed (29). However, it is difficult to determine which serine kinase is affected given the complex regulation of STAT family serine phosphorylation. Inconsistent results in the literature suggest that each signaling pathway involved is stimulus and cell type specific (9). Several serine kinases, including calcium/calmodulin-dependent protein kinase II (CaMKII; 37), p38 MAPK (38, 39), and PI3K/Akt pathways (11) have been implicated in the stress-induced phosphorylation of STAT1 at S727 in various cell types. Although recent work in macrophages suggests that IFN-γ-induced serine phosphorylation is p38 independent (40), increasing evidence supports a role for CaMKII (37, 41). Activation of CaMKII results from an IFN-γ-induced Ca2+ flux and occurs independently from IFN-γ-stimulated JAK tyrosine kinase activation (41). These distinct pathways involved in downstream kinase activation permit the unique modulation of serine and tyrosine phosphorylation events such as we observed in this study.
The data resulting from our investigation suggest that adeno-sine-mediated inhibition of STAT1 S727 phosphorylation occurs via the adenosine A3 receptor. Expression of the A3 receptor on both RAW 264.7 and THP-1 macrophages has been previously confirmed using radioligand binding and immunocytochemistry techniques (24) or gene expression analysis (42), and the protective effects of this adenosine receptor subtype against an LPS challenge are well documented (24, 43, 44). However, to our knowledge, our results are the first to show an A3 receptor-mediated inhibition of IFN-γ-induced macrophage activation and STAT1 signaling. We demonstrate a significant reduction in STAT1-responsive promoter activity upon introduction of the A3 receptor-specific agonist Cl-IB-MECA to IFN-γ-stimulated RAW 264.7 cells. Our data indicate that this effect is unique to the A3 receptor subtype given that neither specific nor nonspecific agonists of A1 and A2 receptors altered promoter activity in IFN-γ-treated cells. Immunoblot results from IFN-γ-stimulated mouse and human macrophages support this conclusion by showing reduced suppression of STAT1 S727 phosphorylation with A3 receptor inhibition. Specific stimulation of the RAW 264.7 A3 receptor with Cl-IBMECA and the complete reversal of its suppressive effect on IRF1 and iNOS gene expression with MRS 1191 (an A3 receptor-specific antagonist) suggest that A3 receptor signaling is both necessary and sufficient to mediate the suppression of these STAT1-dependent genes by adenosine.
Our data, obtained using both selective agonists and antagonists of all four adenosine receptor subtypes, clearly indicate that A3 receptor signaling is central to the inhibition of STAT1 S727 phosphorylation and phosphoserine-mediated transcriptional activity in IFN-γ-stimulated macrophages. Although an abundance of experimental work has been dedicated to elucidating the signaling mechanism following A3 receptor activation in macrophages, this signal transduction pathway remains largely enigmatic (45). Stimulation of the A3 receptor in macrophages does not appear to involve either of the two classical pathways: a Gi-mediated inhibition of adenylyl cyclase and concomitant decrease in cAMP; or a Gq-mediated activation of phospholipase C and concomitant increase in intracellular Ca2+ (35, 45, 46). Previous research in RAW 264.7 cells using an LPS challenge demonstrates that A3 receptor activation may in fact have an opposite effect from the classical Gq pathway and instead block intracellular Ca2+ accumulation (24). Thus, it is possible that inhibition of an IFN-γ-mediated intracellular Ca2+ flux by the A3 receptor could account for reduced STAT1 serine phosphorylation levels such as we observed in this study.
It has been shown in previous investigations that higher concentrations of adenosine analogs are required to affect functional responses in RAW 264.7 cells, potentially due to differences in postreceptor signaling processes in this cell type (24). The 300 μM adenosine concentration used for RAW 264.7 cell treatment and the 100 μM adenosine concentration used for THP-1 cell treatment in this investigation were selected due to a strong, consistent inhibition of IFN-γ-induced gene expression or STAT1 phosphoserine band intensity, respectively, in initial dose-response experiments. We have found that both the 100 μM and 300 μM concentrations are within the range selected in other investigations (23, 35, 36, 43, 47) and are not uncommon physiologically (48, 49). As such, the present results offer new insight into the immunosuppressive action in macrophages of adenosine in the context of inflammatory vascular disease and, therefore, provide a basis for further exploration in other experimental models.
In summary, the current study reveals a novel mechanism by which adenosine modulates macrophage activation in IFN-γ-stimulated human and mouse cell lines. We demonstrate that adenosine signaling reduces serine phosphorylation of STAT1, resulting in significant loss of its transcriptional activity. This adenosine-mediated suppression of genes central to macrophage scavenging and immunological functions could lead to a disruption in the atherogenic cycle of inflammation and cholesterol accumulation in the injured arterial wall. Deciphering the role of adenosine in modulating STAT1 signaling may improve therapeutic strategies to prevent the initiation and progression of inflammatory vascular disease.
We thank Dr. Anne Knowlton and Robin Altman for their helpful discussions and advice.
1This work was supported by a Howard Hughes Medical Institute Integrating Medicine into Basic Science Fellowship, an Achievement Rewards for College Scientists Scholarship, and National Institutes of Health Grant HL55667.
3Abbreviations used in this paper: GAS, γ-activated sequence; IRF, IFN regulatory factor; iNOS, inducible nitric oxide synthase; CaMKII, calcium/calmodulin-dependent protein kinase II; EHNA, erythro-9-(2-hydroxy-3-nonyl)adenine; CCPA, 2-chloro-N6-cyclopentyladenosine; CGS21680, 2-p-(2-carboxyethyl)phenethylamino-5′-N-ethylcarboxamidoadenosine; NECA, N-ethylcarboxamidoadenosine; Cl-IB-MECA, 2-chloro-N6-(3-iodobenzyl)adenosine-5′-N-methyluronamide; CPX, 8-cyclopentyl-1,3-dipropylxanthine; SCH 58261, 5-amino-7-(β-phenylethyl)-2-(8-furyl)pyrazolo(4,3-3)-1,2,4-triazolo(1,5-c)pyrimidine; MRS 1191, 3-ethyl-5-benzyl-2-methyl-4-phenylethynyl-6-phenyl-1,4-(±)-dihydropyridine-3,5-dicarboxylate; MRS 1220, 9-chloro-2-(2-furanyl)-5-((phenylacetyl)amino)-[1,2,4]triazolo[1,5-c]quinazoline; qRT-PCR, quantitative real-time PCR.
The authors have no financial conflict of interest.