IFN-γ Prevents Adenosine Receptor (A2bR) Upregulation To Sustain the Activation Response

This information is current as Heather B. Cohen, Amanda Ward, Kajal Hamidzadeh, Katya of October 1, 2021. Ravid and David M. Mosser J Immunol 2015; 195:3828-3837; Prepublished online 9 September 2015; doi: 10.4049/jimmunol.1501139 http://www.jimmunol.org/content/195/8/3828 Downloaded from

Supplementary http://www.jimmunol.org/content/suppl/2015/09/09/jimmunol.150113 Material 9.DCSupplemental http://www.jimmunol.org/ References This article cites 31 articles, 8 of which you can access for free at: http://www.jimmunol.org/content/195/8/3828.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists by guest on October 1, 2021

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2015 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

IFN-g Prevents Adenosine Receptor (A2bR) Upregulation To Sustain the Macrophage Activation Response

Heather B. Cohen,*,†,1 Amanda Ward,*,† Kajal Hamidzadeh,*,† Katya Ravid,‡ and David M. Mosser*,†

The priming of with IFN-g prior to TLR stimulation results in enhanced and prolonged inflammatory production. In this study, we demonstrate that, following TLR stimulation, macrophages upregulate the adenosine 2b receptor (A2bR) to enhance their sensitivity to immunosuppressive extracellular adenosine. This upregulation of A2bR leads to the induction of macrophages with an immunoregulatory phenotype and the downregulation of inflammation. IFN-g priming of macrophages selectively prevents the induction of the A2bR in macrophages to mitigate sensitivity to adenosine and to prevent this regulatory transition. IFN-g–mediated A2bR blockade leads to a prolonged production of TNF-a and IL-12 in response to TLR ligation. The pharmacologic inhibition or the genetic deletion of the A2bR results in a hyperinflammatory response to TLR Downloaded from ligation, similar to IFN-g treatment of macrophages. Conversely, the overexpression of A2bR on macrophages blunts the IFN-g effects and promotes the development of immunoregulatory macrophages. Thus, we propose a novel mechanism whereby IFN-g contributes to host defense by desensitizing macrophages to the immunoregulatory effects of adenosine. This mechanism over- comes the transient nature of TLR activation, and prolongs the antimicrobial state of the classically activated macrophage. This study may offer promising new targets to improve the clinical outcome of inflammatory diseases in which macrophage activation is dysregulated. The Journal of Immunology, 2015, 195: 3828–3837. http://www.jimmunol.org/

xposure of macrophages to TLR ligands initiates the IFN-g represents a key component in chronic inflammatory production of inflammatory and mediators that environments (2). Activated NK cells, invariant NKT, and TH1 E can influence immune responses (1). The production of cells secrete IFN-g, which primes macrophages to increase their these cytokines and mediators must be tightly regulated because production of inflammatory cytokines and mediators (6). In this their overproduction can result in damage to surrounding tissue way, IFN-g plays vital roles in cell-mediated immunity and host (2). We recently described a negative feedback mechanism defense against intracellular pathogens (7). However, macro- whereby macrophages terminate the response to TLR stimuli to phages exhibiting an “IFN signature” are observed in rheumatoid

prevent immunopathology (3). After TLR stimulation, macro- arthritis, multiple sclerosis, and many other autoimmune diseases, by guest on October 1, 2021 phages alter their metabolism and release low levels of ATP, indicating that IFN-g can have a pathogenic role during chronic which is rapidly converted into immunosuppressive adenosine via macrophage activation (8, 9). Although the ability of IFN-g to CD39-mediated hydrolysis. Adenosine promotes a transition to- enhance the inflammatory potential of TLR-activated macro- ward an immunoregulatory macrophage phenotype (4, 5). We also phages is a well-known phenomenon, how IFN-g affects the demonstrated that a deficiency of CD39 on macrophages resulted intrinsic regulation of macrophage activation remains to be in severe immunopathology in a mouse model of septic shock, determined. In this study, we investigated the effect of IFN-g on because of the inability of these cells to undergo this transition to TLR-activated macrophages, and we reveal that IFN-g sustains regulatory macrophages. Thus, the generation of adenosine by inflammatory macrophage activation by attenuating their sensi- macrophages is an important mechanism that prevents overactive tivity to extracellular adenosine. Extracellular adenosine can sig- inflammatory responses. nal through four different adenosine receptors, designated A1, A3, A2a, and A2b (10). The relative expression of these receptors varies with cell type, anatomic location, and activation state (11). We demonstrate that IFN-g treatment selectively blocks the in- *Department of Cell Biology and Molecular Genetics, University of Maryland, Col- duction of the adenosine 2b receptor (A2bR). We also show that lege Park, MD 20742; †Maryland Pathogen Research Institute, College Park, MD 20742; and ‡School of Medicine, Boston University, Boston, MA 02118 IFN-g inhibition of TLR-induced A2bR prevents completion of an important negative feedback loop, thus rendering macrophages 1Current address: ImmunoOncology, EMD Serono Research and Development Insti- tute, Billerica, MA. hyperinflammatory. Finally, we demonstrate that overexpression Received for publication May 18, 2015. Accepted for publication August 4, 2015. of A2bR on macrophages rescues their ability to respond to This work was supported in part by National Institutes of Health Grant R01 adenosine. Thus, we reveal the therapeutic potential of targeting GM102589. macrophage-specific A2bR to reduce chronic inflammation. Address correspondence and reprint requests to Dr. David M. Mosser, Cell Biology and Molecular Genetics, 3102 Bioscience Research Building, 4066 Campus Drive, Materials and Methods University of Maryland, College Park, MD 20742. E-mail address: [email protected] Mice and macrophage isolation The online version of this article contains supplemental material. Abbreviations used in this article: BMDM, bone marrow–derived macrophage; eATP, C57BL/6 mice were purchased from the Charles River Laboratories. These extracellular ATP; NECA, N-ethyl carboxamidoadenosine; Poly(I:C), polyinosinic- studies were reviewed and approved by the University of Maryland In- polycytidylic acid; qPCR, quantitative PCR; WT, wild type. stitutional Animal Care and Use Committee. Bone marrow–derived mac- rophages were prepared from 6–8-wk-old female C57BL/6, A2ar2/2 2 2 2 2 Copyright Ó 2015 by The American Association of Immunologists, Inc. 0022-1767/15/$25.00 (Jackson Labs), A2br / (12), or Stat1 / (Taconic) mice and their www.jimmunol.org/cgi/doi/10.4049/jimmunol.1501139 The Journal of Immunology 3829 Downloaded from http://www.jimmunol.org/ by guest on October 1, 2021

FIGURE 1. IFN-g prolongs the macrophage activation state. (A and C) Cytokine production by LPS-stimulated unprimed (black squares, solid line) or IFN-g–primed (gray circles, solid line) BMDMs over time. (A) TNF-a and (C) IL-12/23p40 was detected with ELISA. The means 6 SEM of three in- dependent experiments are shown. (B and D) Unprimed (black squares, solid line) or IFN-g–primed (gray circles, solid line) BMDMs were stimulated with LPS and then washed with PBS (break in y-axis) and replaced with LPS-free media and monitored over the next 1–6 h for (B) TNF-a and (D) IL-12/23p40. (E) BMDMs were stimulated with LPS alone or in the presence of increasing doses of POM-1. Eight hours after stimulation, cytokine production was assessed with ELISA for TNF-a (black bars), IL-12/23p40 (dark gray bars), and IL-10 ( gray bars). The means 6 SEM of three experiments are shown. (F–I) Unprimed BMDMs (black bars) and IFN-g–primed BMDMs (gray bars) were stimulated with LPS for 4 h. (F) Il-10,(G) Arg1,(H) Il-33,and(I) Hmox-1 mRNA levels were determined with qPCR. The means 6 SEM of at least three independent experiments are shown. (J) WT (black bars) or STAT12/2 BMDMs (striped bars) were left unprimed or primed with IFN-g and then stimulated with LPS. TNF-a (left) and IL-12/23p40 (right) were measured in the supernatants 8 h after stimulation. Protein was measured by ELISA. The means 6 SD of triplicates are shown. **p , 0.01, ***p , 0.001. 3830 THE DOWNREGULATION OF A2bR BY IFN-g respective littermate controls, as described previously (13) and differen- addition of 10 ng/ml Ultra-Pure LPS with or without 108 cells/ml heat- tiated in 20% L929 (LC14) conditioned media unless noted otherwise. killed L. monocytogenes,0.5mg/ml Pam3CSK4, 0.5 mg/ml Poly(I:C) low RAW264.7 cells (TIB-71) were obtained from American Type Culture m.w. and high m.w., 0.5 mg/ml ST-FLA, or 0.5 mg/ml FSL-1 with or without Collection (Manassas, VA). 10 mM ATP or adenosine unless indicated otherwise. Reagents Gene expression analysis ATP and adenosine were purchased from Sigma-Aldrich (St. Louis, MO). Total RNA was isolated by using TriZol (Invitrogen, Life Technologies) Recombinant mouse IFN-g was purchased from R&D Systems (Minne- and converted to cDNA using the ThermoScript Kit (Invitrogen, Life apolis, MN) and reconstituted according to manufacturer’s directions. Technologies) according to the manufacturer’s protocol. Quantitative PCR Ultra Pure LPS Escherichia coli K12, Pam3CSK4, heat-killed Listeria (qPCR) analysis was performed using a LightCycler 480 Real-Time PCR monocytogenes, polyinosinic-polycytidylic acid [Poly(I:C)] low m.w. and System (Roche Diagnostics Corporation-Roche Applied Science; Indian- high m.w., ST-FLA, and FSL-1 were purchased from InvivoGen (San apolis, IN) and GoTaq qPCR Master Mix (Promega, Madison, WI). Primer Diego, CA). MRS1754, CGS21680, N-ethyl carboxamidoadenosine (NECA), pairs used to amplify specific gene products are listed in Supplemental and BAY 60-6583 were purchased from Tocris Bioscience (R&D Systems, Table I. Relative expression levels were calculated using the DDCt method Minneapolis, MN). (14). Gapdh was used as the housekeeping gene for normalization. Stimulation conditions Flow cytometry Abs and analysis All in vitro experiments were performed at 1–2 3 106 cells/ml in tissue Anti-mouse CD16/32 was purchased from AnaSpec (Freemont, CA). Anti- culture–treated plates. Mouse macrophage studies were performed in mouse CD39-AF647 (24DMS1) was purchased from eBioscience (San DMEM/F12+GlutaMax (Life Technologies, Grand Island, NY) supple- Diego, CA). Anti-mouse F4/80-pacific blue (BM8) was purchased from mented with 10% FBS, 1% penicillin–streptomycin, and 1% glutamine BioLegend (San Diego, CA). Anti-mouse CD11b-FITC (M1/70) was (Life Technologies) unless otherwise indicated. For all experiments, purchased from BD Biosciences (San Jose, CA). Anti-mouse TNFa-APC macrophages were primed for 16 h with 100 U/ml IFN-g followed by the (MP6-XT22) was purchased from BD Biosciences. Cell fixation and Downloaded from http://www.jimmunol.org/ by guest on October 1, 2021

FIGURE 2. IFN-g blocks ATP-induced immunosuppression downstream of eATP hydrolysis. (A and B) Unprimed (black bars) and IFN-g primed (gray bars) BMDMs were stimulated with LPS in the absence or presence of 10 mM ATP for 8 h. (A) TNF-a and (B) IL-12/23p40 production was measured by ELISA. The means 6 SEM of three independent experiments are shown. (C and D) Unprimed (black bars) and IFN-g–primed (gray bars) BMDMs were stimulated with LPS in the absence or presence of 10 mM adenosine for 8 h. (C) TNF-a and (D) IL-12/23p40 production was measured by ELISA. The means 6 SEM of three independent experiments are shown. (E–G) Unprimed BMDMs (black bars) and IFN-g–primed BMDMs (gray bars) were stimulated with LPS and adenosine (Ado) for 4 h, and mRNA levels were determined with qPCR. The means 6 SEM of at least three independent experiments are shown. **p , 0.01, ***p , 0.001. The Journal of Immunology 3831 permeabilization for intracellular staining was performed using the BD N1-A2bR plasmid (Addgene, Cambridge, MA). The CMV promoter drives Cytofix/Cytoperm fixation/permeabilization kit according to the manu- the expression of both cDNAs. Prewarmed media (500 ml) was immedi- facturer’s protocol. Samples were analyzed with a BD FACSCanto (BD ately added to the cuvette and cells transferred to tissue culture plates. Biosciences). Data were analyzed using FlowJo software (Tree Star, Cells were allowed to recover for 24 h, and media was replaced prior to Ashland, OR). stimulation. For intracellular staining experiments, cells were stimulated for 4 h in the presence of Golgi-Plug per the manufacturer’s protocol (BD Secreted cytokine detection Biosciences). Cytokine production was determined with sandwich ELISA using the Statistical analysis following capture and biotinylated detection Ab pairs: anti-mouse TNF- purified (G281-2626), TNF-biotin (MP6-XT3), IL-12/23p40–purified Data analysis was performed using GraphPad Prism software (GraphPad (C15.6), IL-12/23p40–biotin (C17.8), IL-10–purified (JES5-2A5), IL-10– Software, La Jolla, CA) and analyzed using the Student t test. The statistical biotin (JES5-16E3). differences between groups, with the p values, are indicated in the figures. ATP hydrolysis assay Results Extracellular ATP (eATP) hydrolysis was detected using the ATPlite Luminescence ATP Detection Assay System (PerkinElmer, Waltham, IFN-g sustains inflammatory macrophage responses MA) according to the manufacturer’s protocol but omitting the cell-lysis Priming bone marrow–derived murine macrophages with IFN-g step (3). prior to TLR stimulation with LPS results in a heightened Inorganic phosphate release assay proinflammatory response, characterized by the secretion of high levels of TNF-a (Fig. 1A, 1B) and IL-12/23p40 (Fig. 1C, 1D). Adenine nucleotide hydrolysis was detected using the BIOMOL Green Reagent (Enzo Life Sciences, Farmingdale, NY) according to the manu- IFN-g increased not only the magnitude but also the duration of Downloaded from facturer’s protocol. inflammatory cytokine production by TLR-activated macrophages (Fig. 1B, 1D). To demonstrate this, unprimed or IFN-g–primed Macrophage transfection macrophages were stimulated with LPS, and the LPS was re- RAW264.7 cells were transfected by nucleofection using the Amaxa Mouse moved after 2 h. The macrophages were recultured in fresh LPS- Macrophage Nucleofector Kit (Lonza Cologne, Cologne, Germany) free media, and cytokine production was assessed 1–6 h later. according to manufacturer’s protocol. Briefly, 1 3 106 macrophages were resuspended in 100 ml of Nucleofector solution (Lonza Cologne) and Unprimed macrophages quickly stopped producing TNF-a nucleofected using the Nucleofector I Device and program Y-01. Cells (Fig. 1B) and IL-12/23p40 (Fig. 1D) following the removal of http://www.jimmunol.org/ were transfected with 2 mg pmaxGFP (Lonza Cologne) or with pEYPF- the activating stimuli. In contrast, IFN-g–primed macrophages by guest on October 1, 2021

FIGURE 3. LPS-induced ATP release and hydrolysis is intact in the presence of IFN-g.(A) Levels of ATP released from LPS-stimulated unprimed (black line) and IFN-g primed (gray line) RAW264.7 cells were measured over time. The means 6 SEM of two independent experiments are shown. (B) Flow cytometry analysis of CD39 surface expression (black histograms) on unprimed (left) and IFN-g primed (right) BMDMs compared with isotype controls (gray histograms). Data are representative of three independent experiments. (C) The hydrolysis of 500 mM ATP by unprimed BMDMs (black squares, solid line) over time compared with IFN-g–primed BMDMs (gray circles, solid line). Percent of hydrolysis was compared with levels of ATP detected in the absence of cells and determined by the ATPlite assay. The means 6 SEM from three independent experiments are shown. 3832 THE DOWNREGULATION OF A2bR BY IFN-g continued to secrete TNF-a and IL-12/23p40 with little reduction primed macrophages show reduced levels of transcripts for Il-10, in the amounts secreted after the LPS had been removed from Arg1, Il-33, and Hmox-1 (Fig. 1F–I). IFN-g is well known to macrophage cultures. This prolonged cytokine production by signal through STAT1 (16). To determine whether STAT1- macrophages following IFN-g priming was similar to a phenotype dependent signaling was necessary for enhanced cytokine pro- we recently observed when macrophages lacking CD39 (an duction, macrophages from Stat1-deficient mice were examined ectonucleoside triphosphate diphosphohydrolase) were stimulated and compared with wild type (WT) mice. Priming with IFN-g with TLR ligands (3). Treating macrophages with POM-1, a CD39 failed to prolong inflammatory cytokine responses and did not inhibitor, resulted in a dose-dependent increase in the production result in an increase in TNF-a or IL-12/23p40 (Fig. 1J) in Stat1- of TNF and IL-12 and a decrease in IL-10 production (Fig. 1E), deficient mice. Collectively, these results suggest that IFN-g analogous to what we observed following IFN-g priming. delays the normal transition that macrophages make from an inflam- In the absence of IFN-g priming, TLR-stimulated macrophages matory TLR-stimulated (M1) macrophage to an anti-inflammatory gradually transition from an inflammatory (M1) phenotype into regulatory macrophage. an anti-inflammatory, macrophage capable of regulating innate immune responses. These macrophages upregulate a number of ATP hydrolysis is intact in IFN-g–primed macrophages immunoregulatory transcripts not expressed by M1 macrophages In view of the above findings, we hypothesized that IFN-g priming (3, 15). The priming of macrophages with IFN-g prior to TLR might inhibit a component of the macrophage purinergic autor- stimulation prevents the production of regulatory transcripts that egulatory mechanism we described recently (3). To determine accompany this transition to regulatory macrophages. IFN-g– whether IFN-g prevented macrophage responses to ATP, macro- Downloaded from http://www.jimmunol.org/ by guest on October 1, 2021

FIGURE 4. IFN-g preferentially inhibits LPS-induced A2bR expression in macrophages. (A) mRNA expression of A1r (dotted bars), A2ar (horizontally striped bars), A2br (diagonally striped bars), and A3r (solid black bars) in unprimed and IFN-g–primed BMDMs stimulated for 4 h. The means 6 SEM of three independent experiments are shown. (B and D) Unprimed BMDMs (black bars) and IFN-g–primed BMDMs (gray bars) were stimulated with the indicated TLR agonists for 4 h. (B) A2bR and (D) A2aR mRNA levels were determined by qPCR. The means 6 SEM of at least three independent determinations are shown. (C) Unprimed BMDMs (black bars) and IFN-g–primed BMDMs (gray bars) were stimulated with LPS over time. A2bR mRNA levels were determined by qPCR. The means 6 SEM of two independent experiments are shown. (D) Unprimed BMDMs (black bars) and IFN-g primed BMDMs (gray bars) were stimulated with the indicated TLR agonists for 4 h. A2aR mRNA levels were determined by qPCR. The means 6 SEM of at least three independent determinations are shown. (E) WT BMDMs (black bars) or STAT12/2 BMDMs (striped bars) were left unprimed or IFN-g primed, and then stimulated with LPS for 4 h. A2bR mRNA levels were determined by qPCR. The means 6 SEM of at least three independent determinations are shown. *p , 0.05, **p , 0.01, ***p , 0.001. The Journal of Immunology 3833 phages were stimulated with LPS in the presence or absence of regulatory macrophages. Priming macrophages with IFN-g exogenous ATP, and cytokine production was measured the fol- prevented this induction. Transcripts for IL-10 (Fig. 2E), IL-33 lowing day. ATP (10 mM) significantly reduced the amount of (Fig. 2F), and Arg1 (Fig. 2G) remained low in IFN-g–primed TNF-a (Fig. 2A) and IL-12/23p40 (Fig. 2B) secreted by macro- macrophages, indicating that under these condition the cells phages stimulated with LPS alone. IFN-g priming of macrophages failed to transition to an immunoregulatory activation state. negated the immunosuppressive effect of ATP. These cells con- The release of ATP and its hydrolysis to bioactive adenosine tinued to secrete high levels of TNF-a (Fig. 2A) and IL-12/23p40 was then examined in IFN-g–primed macrophages. Only negligible (Fig. 2B) despite the presence of ATP. Similar studies were per- amounts of ATP were detected in the supernatants of primary formed to determine whether IFN-g renders macrophages less macrophages following LPS stimulation, because these cells ex- sensitive to extracellular adenosine. Macrophages were stimulated press high levels of CD39, which rapidly hydrolyzes ATP to AMP with LPS in the presence or absence of adenosine (10 mM), and (3). Therefore, we performed this analysis on RAW264.7 cells that LPS-induced TNF-a (Fig. 2C) and IL-12/23p40 (Fig. 2D) pro- express low levels of CD39 (3). RAW264.7 cells released ATP duction by macrophages was measured. Similar to extracellular following TLR stimulation, and the amount of ATP released was ATP, the attenuation of inflammatory cytokine secretion by aden- essentially unchanged by priming with IFN-g (Fig. 3A). Next, the osine was decreased by priming cells with IFN-g. We previously level of CD39 expressed on the surface of primary macrophages demonstrated that the addition of adenosine to resting macrophages exposed to LPS alone was compared with expression on cells resulted in the rapid induction of several transcripts specific to following IFN-g priming. CD39 remained highly expressed by Downloaded from http://www.jimmunol.org/ by guest on October 1, 2021

FIGURE 5. Specific antagonism of the A2bR recapitulates IFN-g–mediated adenosine insensitivity. (A and B) Unprimed (black bars) and MRS1754- primed (gray striped bars) BMDMs were stimulated with LPS in the absence or presence of 10 mM adenosine for 8 h. (A) TNF-a and (B) IL-12/23p40 production was measured by ELISA. The means 6 SEM from at least three independent experiments are shown. (C) Unprimed (black squares, solid line), IFN-g–primed (gray circles, solid line), and MRS1754-primed (10 mM; gray open squares, dotted line) BMDMs were stimulated with LPS in the presence or absence of adenosine at various concentrations. Eight hours after stimulation, TNF-a levels were detected by ELISA. The percent of TNF-a produced is based on TNF-a levels produced by LPS alone for each condition. The means 6 SEM from at least three independent experiments are shown. (D and E) Unprimed (black bars) and IFN-g–primed (gray bars) BMDMs were stimulated with LPS in the absence or presence of 1 nM of the A2bR-specific agonist (D) NECA or (E) BAY 60-6583. TNF-a production was measured 8 h after stimulation by ELISA. Percent of TNF-a produced is based on TNF-a levels produced by LPS alone for each condition. The means 6 SEM from three independent experiments are shown. **p , 0.01, ***p , 0.001. 3834 THE DOWNREGULATION OF A2bR BY IFN-g macrophages regardless of culture conditions and was actually diminished (Fig. 4E). Thus, IFN-g signals through STAT1 to in- slightly upregulated by IFN-g (Fig. 3B). We next examined the hibit the expression of A2bR selectively in macrophages exposed effect of IFN-g on ecto-ATPase activity and observed similar rates to TLR stimuli. These data suggest that IFN-g may interfere with of eATP hydrolysis between unprimed and IFN-g primed macro- macrophage purinergic autoregulation by decreasing the sensitivity phages (Fig. 3C). These data indicate that the ability of TLR- of macrophages to extracellular adenosine. stimulated macrophages to release and hydrolyze eATP remains IFN-g inhibits A2bR-dependent immunosuppression intact in the presence of IFN-g. Together, these data suggest that the inability of IFN-g–primed macrophages to transition readily We hypothesized that IFN-g inhibits the induction of A2bR to to an anti-inflammatory (regulatory) state occurs downstream of maintain an inflammatory activation state. To address the func- eATP hydrolysis and is caused by a decreased ability of these tional significance of A2bR during macrophage activation, mac- macrophages to respond to adenosine. rophages were treated with the A2bR-specific antagonist MRS1754 (17) and then exposed to LPS in the presence or absence of aden- IFN-g inhibits LPS-induced A2bR expression in macrophages osine. Unprimed macrophages were highly sensitive to adenosine, Adenosine receptor expression on LPS-stimulated macrophages and they exhibited marked reduction in LPS-induced TNF-a was examined in the presence or absence of IFN-g priming. In (Fig. 5A, black bars) and IL-12/23p40 production (Fig. 5B, black unprimed macrophages, LPS dramatically upregulated A2bR bars). The effect of adenosine was completely dependent on the transcript levels relative to naive unstimulated macrophages A2bR, because macrophages treated with MRS1754 demonstrated (Fig. 4A). This upregulation was specific to A2br transcripts, be- no significant difference in TNF-a or IL-12/23p40 levels regard- cause transcripts for the other three adenosine receptors were not less of adenosine being present in their extracellular milieu significantly increased in response to LPS stimulation (Fig. 4A). (Fig. 5A,B, striped bars). Thus, A2bR antagonism by MRS1754 Downloaded from This phenomenon was also observed in macrophages stimulated abrogates adenosine responsiveness, to levels similar to those with a variety of other TLR agonists, including Pam3CSK, observed after IFN-g treatment. The sensitivity of macrophages HKLM, Poly(I:C), and bacterial flagellin, suggesting that the in- to adenosine was directly measured by exposing unprimed or duction of A2br in macrophages is not specific to TLR4 signaling IFN-g primed macrophages to increasing doses of adenosine. (Fig. 4B). On unprimed cells, adenosine suppressed LPS-induced TNF-a

The induction of the A2bR by LPS stimulation was severely in a dose-dependent manner, indicating that macrophages are http://www.jimmunol.org/ inhibited by priming cells with IFN-g (Fig. 4A). IFN-g specifically sensitive to low levels of extracellular adenosine (Fig. 5C). inhibited A2br expression on inflammatory macrophages and did Macrophages treated with MRS1754 or primed with IFN-g were not affect expression of the other adenosine receptors. Transcripts less sensitive to adenosine, decreasing cytokine secretion by less for A2br were not simply delayed by IFN-g priming, but inhibited than 30% in the presence of 10 mM adenosine (Fig. 5C), indi- for the entire 8-h observation period (Fig. 4C). In contrast to A2br, cating that IFN-g impairs adenosine-mediated immunosuppres- the expression of A2ar transcripts encoding the A2aR were not sion via A2bR signaling blockade. decreased after IFN-g priming, regardless of the stimulation To examine the role of A2bR-dependent signaling in modulat- (Fig. 4D). Adenosine receptor expression on LPS-stimulated ing cytokine production by macrophages, cells were exposed to the macrophages after priming with IFN-g was examined in WT A2bR-specific agonist BAY 60-6583 and the A2-type adenosine by guest on October 1, 2021 macrophages and compared with macrophages lacking STAT1. receptor selective agonist NECA. TNF-a was measured. Unprimed The treatment of Stat1-deficient macrophages with IFN-g did not macrophages stimulated with LPS and either of the A2bR agonists result in a significant decrease in A2br expression, whereas A2bR secreted 70–80% less TNF-a relative to macrophages exposed to receptor expression on WT macrophages was substantially LPS alone (Fig. 5D, 5E). Macrophages treated with IFN-g how-

FIGURE 6. IFN-g inhibits A2bR-de- pendent immunosuppression in macro- phages. (A and B) Unprimed (black bars) and IFN-g–primed (gray bars) WT (A)or A2br2/2 (B) BMDMs were stimulated with LPS in the absence or presence of 10 mM adenosine for 8 h. TNF-a pro- duction was measured by ELISA. The means 6 SEM from at least three inde- pendent experiments are shown. (C and D) Unprimed (black bars) and IFN-g– primed (gray bars) WT (C)orA2ar2/2 (D) BMDMs were stimulated with LPS in the absence or presence of 10 mM adenosine (Ado) for 8 h. TNF-a pro- duction was measured by ELISA. The means 6 SEM from three indepen- dent experiments are shown. *p , 0.05, ***p , 0.001. The Journal of Immunology 3835 ever, largely retained (70%) their ability to secrete TNF-a, even in deficient cells, and TNF-a production returned to levels compa- the presence of NECA (Fig. 5D) or BAY 60-6583 (Fig. 5E). rable to LPS stimulation alone (Fig. 6, gray bars). Together, these The ability of adenosine to attenuate TNF-a production in bone data indicate that IFN-g hinders the induction of the A2bR, the marrow–derived macrophages (BMDMs) from A2br-deficient major adenosine receptor responsible for regulating inflammatory mice was examined. Unprimed macrophages from WT mice were cytokine production in TLR-stimulated macrophages. sensitive to 10 mM adenosine, decreasing TNF-a by ∼50% (Fig. 6A, black bars). Unprimed A2br-deficient BMDMs, in con- A2bR overexpression permits the development of trast, were not responsive to adenosine. Stimulation of these cells immunosuppressive macrophages in the presence of IFN-g with LPS in the presence of adenosine resulted in TNF-a pro- TNF-a production was measured with intracellular staining and duction that was comparable to that from stimulation in the ab- flow cytometry in RAW264.7 cells in which the A2bR was sence of adenosine (Fig. 6B, black bars). As controls, unprimed overexpressed. Resting, nonstimulated cells produced minimal A2ar-deficient BMDMs resembled their WT littermate counter- TNF-a, regardless of their transfection status (Fig. 7A, left two parts (Fig. 6C, black bars) and were sensitive to adenosine- panels). Priming cells with IFN-g and stimulating them with LPS mediated suppression of TNF-a production (Fig. 6D, black resulted in high TNF-a production. IFN-g–primed cells that were bars). Priming macrophages with IFN-g decreased A2bR expres- transfected with control plasmid failed to respond to exogenous sion on WT macrophages and decreased adenosine sensitivity. adenosine and continued to make high levels of TNF-a (Fig. 7A, Thus, after IFN-g priming, WT macrophages resembled A2br- right upper panel). In contrast, IFN-g–primed macrophages Downloaded from http://www.jimmunol.org/ by guest on October 1, 2021

FIGURE 7. A2bR overexpression permits the development of immunosuppressive macrophages in the presence of IFN-g.(A) BMDMs were transfected with plasmids containing GFP (top row) or A2bR-EYFP (bottom row) cDNA. Twenty-four hours later, cells were either left unprimed or primed with IFN-g followed by stimulation with LPS alone or in the presence of 10 mM adenosine (Ado) and BFA. After 4 h of stimulation, TNF-a production was assessed with intracellular FACS. Data are representative from three independent experiments. (B) Histograms representing TNF-a production within transfected (GFP/EYFP+) cells. (C) Quantification of TNF-a levels within stimulated transfected macrophages as determined by intracellular FACS. Data are mean 6 SEM of three independent experiments. **p , 0.01. 3836 THE DOWNREGULATION OF A2bR BY IFN-g transfected with pA2bR were more sensitive to adenosine. The types have pointed to a link between IFN-g and adenosine sig- transfected cells (EYFP+) produced lower levels of TNFa in re- naling (27); however, the molecular mechanism governing IFN-g– sponse to exogenous adenosine, whereas the EYFP2 nontrans- mediated A2bR suppression in macrophages remains elusive. It fected cells continued to make comparable amounts of TNF-a has been shown, for example, that IFN-g downregulates A2bR (Fig. 7A, bottom right panel). The degree of TNF-a attenuation signaling in intestinal epithelial cells, leading to an inhibition of correlated with the level of A2bR expression. Macrophages over- CREB phosphorylation and protein kinase A activation. It has also expressing the A2bR secreted, on average, 40–50% less TNF-a been shown that smooth muscle cells can signal through the A2bR compared with macrophage transfected with the GFP containing to induce the downregulation of IFN-g–induced major histocom- plasmid alone (Fig. 7B, 7C). These results demonstrate that the patibility class II expression (27). A2bR signaling appears to overexpression of A2bR in macrophages improves their ability to promote anti-inflammatory macrophages; however, other studies sense extracellular adenosine and to suppress inflammatory cyto- have associated A2bR with inflammatory roles in disease models kine production. These data reveal that IFN-g interferes with the such as asthma, thus demonstrating diverse and perhaps tissue- autoregulation of macrophages after TLR activation by specifically specific roles for A2bR in the regulation of inflammation (28). attenuating A2bR-mediated responses, resulting in prolonged in- It is interesting that we observed increased A2br levels in flammatory responses. macrophages exposed to TLR ligands or in response to exogenous purinergic nucleotides (data not shown). This observation suggests Discussion that adenosine signaling itself may positively regulate the ex- Although macrophage activation is a necessary component of host pression of the A2bR. Indeed, putative CREB, IFN-g–activated defense, if left uncontrolled it can lead to tissue damage, organ sequence (GAS), and specificity protein 1 (Sp1) binding sites can failure, and even death. Prolonged macrophage activation has been be found in close proximity to the transcriptional start site of A2br. Downloaded from associated with an extensive list of chronic inflammatory diseases, Both CREB and Sp1 have been shown to play roles in the in- such as inflammatory bowel disease, rheumatoid arthritis, and duction of many regulatory macrophage-associated genes, in- a variety of macrophage activation syndromes (9, 18). However, cluding HB-EGF and IL-10 (22, 29). Other potential mechanisms the molecular mechanisms governing the regulation of macro- of A2bR regulation are not excluded (30). For example, in smooth phage activation under these pathologic conditions remains in- muscle cells, TNF-a induced by LPS stimulation upregulates the completely understood. In this study, we reveal a novel mechanism A2br gene promoter, suggesting an effect via NF-kB sites within http://www.jimmunol.org/ wherein IFN-g interferes with the macrophage’s intrinsic ability to this gene promoter (25). It is not clear however, how IFN-g control its activation state, and demonstrate that IFN-g actively inhibits A2br gene expression induced by TLR. Regardless of the desensitizes macrophages to immunosuppressive adenosine, thereby transcriptional mechanism of control, it is interesting to speculate sustaining inflammatory responses. that the A2bR may be included as a novel genetic marker of This work demonstrates that IFN-g inhibits a recently described immunoregulatory macrophages. purinergic autoregulatory loop that is initiated upon TLR ligation This work points to the therapeutic potential of manipulating of macrophages. We previously demonstrated that macrophages macrophage A2bR signaling to influence inflammation. Strategies secrete ATP and rapidly convert ATP to adenosine to autoregulate to overexpress A2bR on macrophages may represent a novel ap- by guest on October 1, 2021 their activation state (3). That work suggested that TLR stimulation proach to treat chronic inflammatory diseases, whereas A2bR of macrophages was a transient event, followed by a transition to downregulation may enhance host defense to intracellular infec- an immunoregulatory phenotype. The deletion or inhibition of the tions. It will be valuable to assess the effects of macrophage-A2bR ecto-ATPase, CD39, prevented this transition and rendered macro- expression in vivo. High TNF-a and IFN-g levels often accompany phages hyperinflammatory because of an inability to convert ATP chronic inflammatory disorders (31). Moreover, extracellular into adenosine. In this study, we reveal that this autoregulatory loop adenosine levels have been demonstrated to increase 100-fold depends on the adenosine A2bR and that IFN-g blocks the transi- during chronic inflammatory conditions, such as rheumatoid ar- tion to an immunosuppressive state by preventing A2bR signaling. thritis (32). Thus, our results showing that CD39 activity is intact Extracellular adenosine can signal through four receptors, in the presence of IFN-g are consistent with these observations. In designated A1, A3, A2a, and A2b (19). All adenosine receptors are this study, we show that IFN-g potentiates the release of these G protein–coupled receptors; however, only the A2aR and A2bR inflammatory cytokines from macrophages by blocking induction are coupled to Gas, thus inducing the generation of intracellular of the anti-inflammatory receptor, A2bR. We demonstrate that cAMP upon stimulation (20, 21). We have previously demon- preserving or overexpressing the A2bR signaling in macrophages strated that the accumulation of intracellular cAMP can drive the may reverse the inflammatory phenotype of macrophages exposed development of anti-inflammatory macrophages (22). Because of to IFN-g. Thus, targeting A2bRs on macrophages may enhance the lower affinity of the A2bR to adenosine compared with the our current arsenal of treatment options for chronic inflammatory A2aR (23), it has also been proposed that the A2bR may be diseases. preferentially used by cells during pathophysiologic diseases when extracellular adenosine levels increase 10–100-fold above Disclosures homeostatic conditions. Indeed, our work demonstrating the se- D.M.M. is the chief executive officer of a company that is developing lective upregulation of the A2bR upon TLR stimulation is con- therapeutics to mitigate macrophage inflammatory responses. The other sistent with other reports showing that inflammatory TNF-a and authors have no financial conflicts of interest. hypoxia can induce the A2bR (24–26). Thus, these data collec- tively support a role for A2bR in mediating a negative feedback loop to prevent overactive inflammation. References The mechanisms by which the A2bR is regulated at the tran- 1. Maisonneuve, C., S. Bertholet, D. J. Philpott, and E. De Gregorio. 2014. Unleashing the potential of NOD- and Toll-like agonists as vaccine adjuvants. scriptional level in macrophages remain poorly understood. Our Proc. Natl. Acad. Sci. USA 111: 12294–12299. current study shows that IFN-g inhibits the development of reg- 2. Mosser, D. M., and J. P. Edwards. 2008. Exploring the full spectrum of mac- rophage activation. Nat. Rev. Immunol. 8: 958–969. ulatory macrophages after TLR activation by inhibiting A2bR 3. Cohen, H. B., K. T. Briggs, J. P. Marino, K. Ravid, S. C. Robson, and upregulation following TLR ligation. Previous studies in other cell D. M. Mosser. 2013. TLR stimulation initiates a CD39-based autoregulatory The Journal of Immunology 3837

mechanism that limits macrophage inflammatory responses. Blood 122: 1935– 18. Ito, R., M. Shin-Ya, T. Kishida, A. Urano, R. Takada, J. Sakagami, J. Imanishi, 1945. M. Kita, Y. Ueda, Y. Iwakura, et al. 2006. Interferon-gamma is causatively in- 4. Hasko´, G., D. G. Kuhel, J. F. Chen, M. A. Schwarzschild, E. A. Deitch, volved in experimental inflammatory bowel disease in mice. Clin. Exp. Immunol. J. G. Mabley, A. Marton, and C. Szabo´. 2000. Adenosine inhibits IL-12 and 146: 330–338. TNF-[alpha] production via adenosine A2a receptor-dependent and independent 19. Koupenova, M., and K. Ravid. 2013. Adenosine, adenosine receptors and their mechanisms. FASEB J. 14: 2065–2074. role in glucose homeostasis and lipid metabolism. J. Cell. Physiol. DOI: 10.1002/ 5. Hasko´, G., and P. Pacher. 2012. Regulation of macrophage function by adeno- jcp.24352. sine. Arterioscler. Thromb. Vasc. Biol. 32: 865–869. 20. Fredholm, B. B. 2007. Adenosine, an endogenous distress signal, modulates 6. Schoenborn, J. R., and C. B. Wilson. 2007. Regulation of interferon-gamma tissue damage and repair. Cell Death Differ. 14: 1315–1323. during innate and adaptive immune responses. Adv. Immunol. 96: 41–101. 21. Ohta, A., and M. Sitkovsky. 2001. Role of G-protein-coupled adenosine recep- 7. Scharton, T. M., and P. Scott. 1993. Natural killer cells are a source of interferon tors in downregulation of inflammation and protection from tissue damage. gamma that drives differentiation of CD4+ subsets and induces early Nature 414: 916–920. resistance to Leishmania major in mice. J. Exp. Med. 178: 567–577. 22. Edwards, J. P., X. Zhang, and D. M. Mosser. 2009. The expression of heparin- 8. Hu, X., S. D. Chakravarty, and L. B. Ivashkiv. 2008. Regulation of inter- binding epidermal -like growth factor by regulatory macrophages. feron and Toll-like receptor signaling during macrophage activation by J. Immunol. 182: 1929–1939. opposing feedforward and feedback inhibition mechanisms. Immunol. Rev. 23. Fredholm, B. B., E. Irenius, B. Kull, and G. Schulte. 2001. Comparison of the 226: 41–56. potency of adenosine as an agonist at human adenosine receptors expressed in 9. Gordon, R. A., G. Grigoriev, A. Lee, G. D. Kalliolias, and L. B. Ivashkiv. Chinese hamster ovary cells. Biochem. Pharmacol. 61: 443–448. 2012. The interferon signature and STAT1 expression in rheumatoid arthritis 24. Sun, Y., Y. Duan, A. S. Eisenstein, W. Hu, A. Quintana, W. K. Lam, Y. Wang, synovial fluid macrophages are induced by a and Z. Wu, K. Ravid, and P. Huang. 2012. A novel mechanism of control of NFkB counter-regulated by the synovial fluid microenvironment. Arthritis Rheum. activation and inflammation involving A2B adenosine receptors. J. Cell Sci. 125: 64: 3119–3128. 4507–4517. 10. Hasko´, G., J. Linden, B. Cronstein, and P. Pacher. 2008. Adenosine receptors: 25. St Hilaire, C., M. Koupenova, S. H. Carroll, B. D. Smith, and K. Ravid. 2008. therapeutic aspects for inflammatory and immune diseases. Nat. Rev. Drug TNF-alpha upregulates the A2B adenosine receptor gene: The role of NAD(P)H Discov. 7: 759–770. oxidase 4. Biochem. Biophys. Res. Commun. 375: 292–296. 11. St Hilaire, C., S. H. Carroll, H. Chen, and K. Ravid. 2009. Mechanisms of in- 26. Kong, T., K. A. Westerman, M. Faigle, H. K. Eltzschig, and S. P. Colgan. 2006. duction of adenosine receptor genes and its functional significance. J. Cell. HIF-dependent induction of adenosine A2B receptor in hypoxia. FASEB J. 20: Downloaded from Physiol. 218: 35–44. 2242–2250. 12. Yang, D., Y. Zhang, H. G. Nguyen, M. Koupenova, A. K. Chauhan, M. Makitalo, 27. Xu, Y., K. Ravid, and B. D. Smith. 2008. Major histocompatibility class II M. R. Jones, C. St Hilaire, D. C. Seldin, P. Toselli, et al. 2006. The A2B transactivator expression in smooth muscle cells from A2b adenosine receptor adenosine receptor protects against inflammation and excessive vascular adhe- knock-out mice: cross-talk between the adenosine and interferon-gamma sig- sion. J. Clin. Invest. 116: 1913–1923. naling. J. Biol. Chem. 283: 14213–14220. 13. Johnson, C. R., D. Kitz, and J. R. Little. 1983. A method for the derivation and 28. Belikoff, B. G., L. J. Vaickus, M. Sitkovsky, and D. G. Remick. 2012. A2B continuous propagation of cloned murine bone marrow macrophages. J. Immu- adenosine receptor expression by myeloid cells is proinflammatory in murine

nol. Methods 65: 319–332. allergic-airway inflammation. J. Immunol. 189: 3707–3713. http://www.jimmunol.org/ 14. Edwards, J. P., X. Zhang, K. A. Frauwirth, and D. M. Mosser. 2006. Biochemical 29. Hu, X., J. Chen, L. Wang, and L. B. Ivashkiv. 2007. Crosstalk among Jak-STAT, and functional characterization of three activated macrophage populations. J. Toll-like receptor, and ITAM-dependent pathways in macrophage activation. J. Leukoc. Biol. 80: 1298–1307. Leukoc. Biol. 82: 237–243. 15. Fleming, B. D., P. Chandrasekaran, L. A. Dillon, E. Dalby, R. Suresh, A. Sarkar, 30. St Hilaire, C., D. Yang, B. M. Schreiber, and K. Ravid. 2008. B-Myb regulates N. M. El-Sayed, and D. M. Mosser. 2015. The generation of macrophages with the A(2B) adenosine receptor in vascular smooth muscle cells. J. Cell. Biochem. anti-inflammatory activity in the absence of STAT6 signaling. J. Leukoc. Biol. 103: 1962–1974. 98: 395–407. 31. Michetti, P., M. Stelle, P. Juillerat, M. Gassull, F. J. Heil, E. Stange, C. Mottet, 16. Schroder, K., P. J. Hertzog, T. Ravasi, and D. A. Hume. 2004. Interferon-gamma: J.-J. Gonvers, V. Pittet, J.-P. Vader, et al. 2009. Appropriateness of therapy for an overview of signals, mechanisms and functions. J. Leukoc. Biol. 75: 163–189. active Crohn’s disease: Results of a multidisciplinary international expert 17. Ji, X., Y. C. Kim, D. G. Ahern, J. Linden, and K. A. Jacobson. 2001. [3H]MRS panel-EPACT II. J. Crohn’s Colitis 3: 232–240. 1754, a selective antagonist radioligand for A(2B) adenosine receptors. Biochem. 32. Sottofattori, E., M. Anzaldi, and L. Ottonello. 2001. HPLC determination of

Pharmacol. 61: 657–663. adenosine in human synovial fluid. J. Pharm. Biomed. Anal. 24: 1143–1146. by guest on October 1, 2021