Enhancement of inosine-mediated A_2AR signaling through positive allosteric modulation

Abstract

Inosine is an endogenous nucleoside that is produced by metabolic deamination of adenosine. Inosine is metabolically more stable (half-life 15h) than adenosine (half-life <10 s). Inosine exerts anti-inflammatory and immunomodulatory effects similar to those observed with adenosine. These effects are mediated in part through the adenosine A_2A receptor (A_2AR). Relative to adenosine inosine exhibits a lower affinity towards the A_2AR. Therefore, it is generally believed that inosine is incapable of activating the A_2AR through direct engagement, but indirectly activates the A_2AR upon metabolic conversion to higher affinity adenosine. A handful of studies, however, have provided evidence for direct inosine engagement at the A_2AR leading to activation of downstream signaling events and inhibition of cytokine production. Here, we demonstrate that under conditions devoid of adenosine, inosine as well as an analog of inosine 6-S-[(4-Nitrophenyl)methyl]-6-thioinosine selectively and dose-dependently activated A_2AR-mediated cAMP production and ERK1/2 phosphorylation in CHO cells stably expressing the human A_2AR. Inosine also inhibited LPS-stimulated TNF-α, CCL3 and CCL4 production by splenic monocytes in an A_2AR-dependent manner. In addition, we demonstrate that a positive allosteric modulator (PAM) of the A_2AR enhanced inosine-mediated cAMP production, ERK1/2 phosphorylation and inhibition of pro-inflammatory cytokine and chemokine production. The cumulative effects of allosteric enhancement of adenosine-mediated and inosine-mediated A_2AR activation may be the basis for the sustained anti-inflammatory and immunomodulatory effects observed in vivo and thereby provide insights into potential therapeutic interventions for inflammation- and immune-mediated diseases.

1. Introduction

The endogenous purine nucleoside, inosine is formed through the metabolic conversion of adenosine by the enzyme adenosine deaminase (ADA). It is produced both extracellularly as well as intracellularly during normal cell metabolism. Inosine has a longer half-life (15 h; [1]) relative to adenosine (< 10 s; [2]) and consequently, the basal level of inosine in the interstitial fluid can be 2–7 times higher than that of adenosine. In agreement with this observation, in pathological conditions, there is an increase in tissue inosine levels [3–6].

Resembling adenosine, inosine exerts a wide range of anti-inflammatory and immunomodulatory properties. These include inhibition of proinflammatory cytokine and chemokine production [7–9], induction of anti-inflammatory cytokine production [7], improvement of islet transplant survival [10], reduced multiorgan inflammation and prolonged survival in scurfy mice [11] as well as alleviation of clinical signs of myelin oligodendrocyte glycoprotein-induced experimental autoimmune encephalomyelitis [12], allergic lung inflammation [13], streptozotocin-induced non-obese type 1 diabetes [14], TNBS-induced colitis [15] and glycodeoxycholic acid-induced acute pancreatitis [16].

Inosine exerts anti-inflammatory and immunomodulatory effects through specific membrane bound G protein-coupled receptors (GPCRs) termed P₁-purinoceptors, also known as adenosine receptors (AR). There are four AR subtypes termed A₁R, A_2AR, A_2BR and A₃R. Among them, A_2AR plays a critical nonredundant role in down-regulating inflammation [17]. A_2AR is coupled to the stimulatory G protein Gαs [18]. Adenosine engagement at the A_2AR leads to an increase in intracellular cAMP levels as well as phosphorylation of signal-regulated kinase-1 and -2 (ERK1/2). Utilizing a combination of label-free, cell-based, and membrane-based functional assays in conjunction with an equilibrium agonist-binding assay we have recently provided in vitro evidence for direct inosine engagement at the A_2AR and subsequent induction of downstream cAMP production and ERK1/2 phosphorylation [19].

GPCRs initiate signaling upon binding of cognate ligands at evolutionarily conserved sites termed orthosteric sites. In addition, GPCRs also contain allosteric sites that are topologically distinct from the orthosteric sites. Therefore, structural determinants of ligand binding at the orthosteric and allosteric sites are inherently different. Unlike orthosteric ligands, allosteric ligands have little or no intrinsic ability to activate GPCRs upon engagement at the allosteric site. They only modulate orthosteric ligand-mediated receptor function through conformational changes that manifest as altered affinity and/or efficacy of the receptors towards orthosteric ligands, hence they preserve the endogenous orthosteric ligand-mediated physiological responses. Allosteric ligands that enhance the orthosteric ligand-mediated responses are termed positive allosteric modulators (PAMs).

To examine the potential of allosteric enhancement of the A_2AR function to alter inflammatory immune responses in vitro and in vivo, we developed a series of compounds with PAM activity. One of these compounds, AEA061, is a small molecule that meets the stringent criteria of a PAM of the A_2AR [20]. AEA061 has no intrinsic activity towards either rat or human A_2ARs but enhances the affinity and maximal response of the receptors to adenosine. AEA061 is selective towards A_2AR. It does not affect potency or efficacy of A₁R or A₃R but increases the potency of A_2BR by two-fold without altering efficacy (unpublished data). Positive allosteric modulation of the A_2AR with AEA061 inhibits inflammatory cytokine and chemokine production in vitro and reduces circulating plasma TNF-α and MCP-1 levels and increases plasma IL-10 in endotoxemic A_2AR intact, but not in A_2AR deficient, mice [20].

We sought to further establish that inosine directly, and not through its conversion to adenosine via salvage pathways, activates the A_2AR. To this end, we examined the ability of inosine as well as the inosine analog 6-S-[(4-Nitrophenyl)methyl]-6-thioinosine (NBMPR) to activate the A_2AR in cell-based and cell-free assays in the presence of ADA and AEA061. We demonstrated that inosine as well as NBMPR activates the A_2AR in the presence and in the absence of ADA. Moreover, both inosine and NBMPR inhibit pro-inflammatory cytokine and chemokine production by A_2AR intact, but not by A_2AR deficient mouse splenic monocytes. We next sought to determine if inosine-mediated A_2AR activation is amenable to allosteric modulation. We now present data supporting the hypothesis that positive allosteric modulation of the A_2AR enhances inosine-mediated A_2AR activation as demonstrated by increased cAMP production, ERK1/2 phosphorylation and inhibition of pro-inflammatory cytokine and chemokine production.

2. Materials and Methods

2.1. Mice

Male BALB/cJ mice and A_2AR null mice (C;129S-Adora2atm1Jfc/J; Jackson Laboratories) were housed at 68–72 °F with a 12 h light/dark cycle, fed normal rodent chow and water ad libitum and were kept in a pathogen-free environment. A protocol approved by the Animal Care and Use Committee of the Molecular Medicine Research Institute was used in this study.

2.2. Materials

CGS 21680, NBMPR, and ZM 241385 were purchased from Tocris Biosciences. Growth media and adenosine deaminase were obtained from Lonza and Worthington Biochemical Corporation respectively. Rolipram, adenosine, inosine, adenosine 5′-[α,β-methylene] diphosphate and LPS (E. coli O111:B4) and all the reagents (unless otherwise stated) were purchased from Sigma-Aldrich.

2.3. Cell culture

CHO-K1 cells stably expressing human A_2AR (CHO-hA_2AR; [21]) were grown in DMEM/F-12 (1:1) supplemented with 10% FBS, 2mM glutamine and G418 (0.2 mg/ml). All cells were maintained at 37 °C in a 5% CO₂ incubator.

2.4. Cell-based cAMP assay

CHO-hA_2AR [21] cells were seeded in 96-well half-area white plates (Greiner bio-one; 5 × 10³ cells/well) in the absence of G418 20 h prior to assay. Cells were either pretreated or left untreated (control) prior to stimulation. For the pretreatment, CHO-hA_2AR cells were washed twice with Hanks’ balanced salt solution (HBSS) and incubated in HBSS containing adenosine deaminase (ADA; 3 U/ml) and adenosine 5′[α,β-methylene] diphosphate (50 μM) for 15 min at 37°C. Both control and pretreated cells were washed twice with HBSS and incubated with rolipram (50 μM), adenosine 5′-[α,β-methylene] diphosphate (50 μM), adenosine, inosine and NBMPR at indicated concentration(s) in the presence or in the absence of ZM 241385 (100 nM) for 10 min at 37 °C. Pretreated cells were also incubated with the same assay components in the presence of ADA (3 U/ml) for 10 min at 37 °C. Intracellular cAMP levels were quantified using an HTRF assay kit (Cisbio).

2.5. Cell-free membrane-based cAMP assay

HEK293-hA_2AR cell membranes (PerkinElmer) were incubated in HBSS, containing adenosine 5′-[α,β-methylene] diphosphate (50 μM) and ADA (3 U/ml) at 37 °C for 20 min. Membranes were washed twice with 33 mM HEPES containing 0.1% Tween 20 and stimulated with the same buffer containing 100 μM ATP, 2 μM GTP, 10 μM GDP, 2 μM MgCl₂, 150 mM NaCl, 50 μM adenosine 5′-[α,β-methylene] diphosphate, 50 μM rolipram, ADA (3 U/ml) and NBMPR (0–300 μM) or CGS 21680 (100 nM) in the presence and in the absence of ZM 241385 (100 nM) in half-area white plates (Greiner bio-one; 4.5 μg protein/well) for 30 min at 37 °C. cAMP levels were quantified using an HTRF assay kit (Cisbio).

2.6. ERK1/2 phosphorylation assay

CHO-hA_2AR cells [21] were seeded in 96-well plates (Greiner bio-one; 2.5 × 10⁴ cells/well) in the absence of G418 20 h prior to assay. The medium was replaced with medium lacking serum and incubated for an additional 3 h. Cells were either pretreated or left untreated (control) prior to stimulation. For the pretreatment, CHO-hA_2AR cells were washed twice with HBSS and incubated in HBSS containing ADA (3 U/ml) and adenosine 5′-[α,β-methylene] diphosphate (50 μM) for 15 min at 37 °C. Control as well as pretreated cells were washed with warm HBSS to remove ADA and incubated with adenosine 5′-[α,β-methylene] diphosphate (50 μM), inosine and AEA061 at indicated concentration(s) in the presence or in the absence of ZM 241385 (100 nM) for 10 min at 37 °C. Pretreated cells were also incubated with the same assay components in the presence of ADA (3 U/ml) for 10 min at 37 °C. The assay was terminated by aspirating the assay buffer and incubating cells with lysis buffer (50 μl/well) at room temperature with shaking for 10 min. Phospho ERK1/2 levels were detected using an Alphascreen Surefire kit (PerkinElmer) according to the manufacturer’s suggested protocol. Briefly, 10μl of the lysate was transferred to a ProxiPlate-384 (PerkinElmer) and incubated with 10 μl of assay detection mixture at room temperature in the dark for 2 h. Fluorescent emissions were quantified using an EnSpire multimode plate reader (PerkinElmer).

2.7. Cytokine assays

Splenic monocytes/macrophages were isolated from 8–10 weeks old male BALB/cJ mice and A_2AR null mice by plastic adherence after incubation for 2 h at 37 °C and 5% CO₂, followed by washing with warm media to remove nonadherent cells. Monocytes/macrophages were seeded in 96-well plates (2 × 10⁵ cells per well) and stimulated with lipopolysaccharide (LPS; 50 ng/ml) in RPMI 1640 containing 1% heat inactivated FBS for 4 h in the presence and in the absence of ADA (3 U/ml) and indicated concentration(s) of inosine, CGS 21680 and AEA061. Cytokine/chemokine levels in the culture supernatants were quantified using ELISA Max kits (BioLegend) and/or bead-based multiplex immunoassays (Eve Technologies).

2.8. Data analysis

Data were analyzed using GraphPad Prism Software. Dose response curves were generated by non-linear regression with a variable slope. Statistical comparisons of the two groups were compared using a paired t-test. Comparisons of multiple groups were performed using one-way ANOVA followed by Tukey’s multiple comparisons test.

3. Results

3.1. Inosine dose-dependently induces A_2AR-mediated cAMP production

Utilizing a combination of label-free, cell-based, and membrane-based functional assays in conjunction with an equilibrium agonist-binding assay we have demonstrated that inosine is a low-affinity agonist at the A_2AR [19]. We have also provided evidence to dismiss the notion that exogenous inosine influences extracellular adenosine levels through the equilibrative nucleoside transporter 1 (ENT1) and ENT2 to activate the A_2AR in cellular assays [19]. To further establish that inosine directly activates the A_2AR, we evaluated inosine-mediated cAMP production by CHO-K1 cells stably transfected with human A_2AR (CHO-hA_2AR) in the presence of adenosine deaminase (ADA), an enzyme that eliminates adenosine by catalyzing the hydrolytic deamination of adenosine to inosine. To this end, we utilized a previously described approach that isolates and decreases background signaling and thereby enhances the cAMP signal generated by exogenous inosine [19]. In CHO-hA_2AR cells, adenosine induced cAMP production (Fig 1A). The A_2AR inverse agonist ZM 241385 blocked the adenosine-mediated cAMP production indicating that the adenosine effects are mediated through the A_2AR. Addition of ADA at 3 U/ml (Fig 1A) and 6 U/ml (data not shown) both significantly reduced baseline cAMP levels and completely abolished adenosine-mediated cAMP production by CHO-hA_2AR cells suggesting that 3 U/ml of ADA is sufficient to rapidly convert high-affinity agonist adenosine to low-affinity agonist inosine. In the presence of the same concentration of ADA, exogenous inosine increased cAMP production by CHO-hA_2AR cells that were blocked by ZM 241385 demonstrating that inosine directly engages the A_2AR and is a bonafide agonist at this receptor. Moreover, dose response analysis indicated that ADA increased the EC₅₀ without altering E_max for inosine-mediated cAMP production in CHO-hA_2AR cells (Fig 1B).

Fig. 1.

Inosine directly and dose-dependently stimulates hA_2AR-mediated cAMP production. CHO-hA_2AR cells were incubated with adenosine (control) or inosine with and without ADA (3 U/ml) and in the presence/absence of the hA_2AR-selective inverse agonist ZM 241385 for 10 min (A). Mean intracellular cAMP levels ± SEM of a representative experiment are shown (n=3; ***, < 0.001 vs DMSO; θθθ, p < 0.001 vs without ADA; ###, p < 0.001 vs DMSO with ADA; τττ, < 0.001 vs without ADA and ZM 241385; ^^, p < 0.001 vs with ADA but without ZM 241385). Dose response of hA_2AR-mediated cAMP production to inosine (B). Mean intracellular cAMP levels ± SEM of a representative experiment are shown (n=3; ***, < 0.001 vs ADA).

3.2. Inosine analog 6-S-[(4-Nitrophenyl)methyl]-6-thioinosine induces A_2AR-mediated cAMP production

The inosine analog 6-S-[(4-Nitrophenyl)methyl]-6-thioinosine (NBMPR) is a high- and low-affinity inhibitor of ENT1 and ENT2 respectively [22]. ENT1 and ENT2 are major transporters of adenosine and inosine across cell membranes [23]. Therefore, NBMPR has been used in cell-based AR activation assays to reduce the potential impact of cellular uptake, intracellular production and transport of adenosine and inosine. As NBMPR structurally resembles inosine, we examined whether NBMPR induces cAMP production by CHO-hA_2AR cells. As shown in Fig 2A, both adenosine as well as NBMPR activated A_2AR-mediated cAMP production by CHO-hA_2AR cells. To rule out the possibility that NBMPR-dependent activation of the A_2AR is not due to an increase in extracellular adenosine upon NBMPR addition via some hitherto unknown mechanism(s), we included ADA in the assay. The addition of ADA at 3 U/ml reduced basal as well as adenosine-induced cAMP production by CHO-hA_2AR cells (Fig 2A). In addition, both adenosine- and NBMPR-mediated cAMP production were inhibited by the A_2AR inverse agonist ZM 241385 in the presence and in the absence of ADA indicating selectivity of the two agonists. These results suggest that the inosine analog NBMPR stimulates the A_2AR resembling the activation seen with inosine.

Fig. 2.

Inosine analog NBMPR dose-dependently induces hA_2AR-mediated cAMP production in cellular and cell-free membrane assays.

CHO-hA_2AR cells were incubated with adenosine (control) or NBMPR with and without ADA (3 U/ml) and in the presence/absence of the hA_2AR-selective inverse agonist ZM 241385 for 10 min (A). Mean intracellular cAMP levels ± SEM of a representative experiment are shown (n=3; **, < 0.01 vs DMSO; ***, < 0.001 vs DMSO; θθθ, P < 0.001 vs without ADA; ###, p < 0.001 vs DMSO with ADA; τττ, < 0.001 vs without ADA and ZM 241385; ^^^, p < 0.001 vs with ADA but without ZM 241385). CHO-hA_2AR cell membranes were incubated with indicated concentrations of NBMPR and CGS 21680 (CGS) in the presence and in the absence of the A_2AR-selective inverse agonist ZM 241385 for 30 min (B). Mean cAMP production ± SEM of representative experiments are shown (n=6; ** and ***, p < 0.05 and p < 0.001 vs DMSO respectively; ^^ and ^^^, p < 0.05 and p < 0.001 vs without ZM 241385 respectively).

We demonstrated previously that inosine activates cAMP production by HEK293-hA_2AR cytosol-free membrane preparations indicating that inosine directly activates the A_2AR [19]. This cell-free assay is completely devoid of potential confounding variables such as influx, intracellular production and efflux of adenosine and inosine. To further establish inosine agonism at the A_2AR we evaluated the efficacy of the inosine analog NBMPR on A_2AR activation in HEK293-hA_2AR membrane preparations. As shown in Fig 2B, NBMPR dose-dependently increased cAMP production and the A_2AR inverse agonist ZM 241385 blocked NBMPR-induced cAMP production by HEK293-hA_2AR cytosol-free membrane preparations. These results indicate that the inosine analog NBMPR specifically and directly activates the A_2AR.

3.3. Positive allosteric modulation enhances inosine-mediated hA_2AR activation

AEA061, a small molecule that meets the stringent criteria of a positive allosteric modulator (PAM) of the A_2AR, enhances adenosine-mediated A_2AR activation and exerts pharmacological activity in a mouse model of endotoxemia [20]. We sought to examine whether inosine-mediated A_2AR activation is also amenable to functional enhancement by AEA061. To this end, we evaluated the effects of AEA061 on inosine-mediated cAMP production by CHO-hA_2AR cells, doing so in the presence of ADA to eliminate adenosine. As shown in Figure 3A, inosine induced cAMP production (p < 0.001) and the A_2AR inverse agonist ZM 241385 blocked this induction (p < 0.001). AEA061 by itself did not induce cAMP production. However, it enhanced inosine-mediated cAMP production (p < 0.05) that was again blocked by the A_2AR inverse agonist ZM 241385 (p< 0.01). These results demonstrate that AEA061 has no intrinsic activity towards hA_2AR-mediated cAMP production, but can allosterically augment inosine-driven hA_2AR-mediated cAMP production.

Fig. 3.

PAM of the A_2AR AEA061 enhances inosine-inducible, A_2AR-mediated cAMP production. CHO-hA_2aR cells were incubated with inosine (Ino; 1000 μM) and AEA061 (10 μM) in the presence and in the absence of ZM 241385 (100 nM) for 10 min (A). Mean cAMP levels ± SEM of a representative experiment is shown (n=3; ***, p < 0.001 vs DMSO; ###, < p < 0.001 vs inosine; &&&, p < 0.001 vs without ZM 241385). AEA061 increases E_max (B) and alters EC₅₀ (C) of inosine-mediated A_2AR activation. CHO-A_2AR cells were incubated with AEA061 and varying concentrations of inosine for 10 min. Mean E_max (B) and EC₅₀ (C) values of four experiments with SEM are shown (**, < 0.01 vs DMSO; # and ##, p < 0.05 and 0.01 vs without ADA pretreatment; ^^, p < 0.01 vs ADA pretreatment without subsequent ADA addition respectively).

ADA is a multifunctional protein with intracellular and extracellular localization. In addition to converting adenosine to inosine, ADA physically interacts with the A_2AR and allosterically modulates adenosine binding to the A_2AR [24, 25]. To fully understand the effect of AEA061 on inosine-mediated A_2AR activation in the context of the natural cellular milieu, we examined the dose response of inosine at the hA_2AR in CHO-hA_2AR cells under several assay conditions. CHO-hA_2AR cells were pretreated with ADA and adenosine 5′-[α,β-methylene] diphosphate, an inhibitor of ecto-5′-nucleotidase to prevent exogenous production of adenosine, then washed and stimulated in the presence or in the absence of ADA with inosine. Regardless of the treatment condition, AEA061 increased the E_max (maximal response) for inosine-induced hA_2AR-mediated cAMP production (Fig 3B). Relative to the control (without the pretreatment), both ADA pretreatment and ADA pretreatment plus subsequent ADA addition increased the EC₅₀ of inosine-mediated cAMP production (decreased affinity of the hA_2AR; Fig 3C). AEA061 reduced the EC₅₀ of inosine (increased hA_2AR affinity to inosine) in ADA pre-treated CHO-hA_2AR cells (p < 0.01) but not in cells without the ADA pretreatment alone or with ADA pretreatment and subsequent ADA addition. These results suggest that the positive allosteric modulator AEA061 exhibits differential effects on the affinity and efficacy of the A_2AR to inosine. AEA061 enhances efficacy of the A_2AR towards inosine both in the presence and absence of membrane-bound adenosine and ADA. However, AEA061 enhances A_2AR affinity to inosine only in the absence of membrane-bound adenosine and ADA.

3.4. Positive allosteric modulation enhances inosine-mediated hA_2AR-dependant ERK1/2 activation

We demonstrated previously that A_2AR activation by both adenosine and inosine leads to ERK1/2 phosphorylation [19]. To determine whether the PAM of the A_2AR, AEA061, potentiates inosine-inducible, A_2AR-mediated ERK1/2 activation, we evaluated the effects of AEA061 on ERK1/2 phosphorylation in CHO-hA_2AR cells. Consistent with our previous observations, inosine increased ERK1/2 phosphorylation (Fig 4A) indicating that inosine effects are mediated through the A_2AR. Moreover, AEA061 increased both basal and inosine-induced ERK1/2 phosphorylation (Fig 4A). The A_2AR inverse agonist ZM 241385 reversed inosine-, AEA061- and inosine plus AEA061-mediated ERK1/2 phosphorylation. Dose response analyses indicated that AEA061 alone enhanced the E_max (maximal response; Fig 4B) regardless of the assay conditions. AEA061 reduced EC₅₀ (increased affinity) to inosine with and without the ADA pre-treatment (Fig 4C). Although not statistically significant, AEA061 lowered the EC₅₀ to inosine in CHO-hA_2AR cells with ADA pre-treatment and subsequent ADA addition. These results collectively suggest that both AEA061- and inosine-induced ERK1/2 phosphorylation are mediated through the A_2AR and that AEA061 functions both as an agonist as well as a PAM with respect to ERK1/2 activation, a sharp contrast to its effect on cAMP activation where AEA061 functions solely as a PAM.

Fig. 4.

AEA061 potentiates inosine-inducible, A_2AR-mediated ERK1/2 phosphorylation. CHO-hA_2AR cells were incubated with inosine (Ino; 1000 μM) and AEA061 (10 μM) in the presence and in the absence of ZM 241385 for 10 min (A). Phospho ERK1/2 levels were quantified using a FRET-based detection kit. Mean FRET ratios ± SEM of a representative experiment are shown (n=3; ***, p < 0.001 vs DMSO; ##, < 0.01 vs AEA061; && and &&&, p < 0.01 and 0.001 vs without ZM 241385 respectively). AEA061 enhances E_max (B) and reduces EC₅₀ (C) of inosine-mediated ERK1/2 phosphorylation. CHO-A_2aR cells were pre-treated and incubated with AEA061 and varying concentrations of inosine for 10 min. Mean E_max (B) and EC₅₀ (C) values of three independent experiments with SEM are shown (* and ***, < 0.05 and 0.001 vs DMSO respectively).

3.5. Inosine suppresses the production of pro-inflammatory cytokines through the A_2AR

Inosine exerts anti-inflammatory effects through inhibition of the production of pro-inflammatory cytokines and chemokines [7, 26–28]. Using pharmacological tools, Haskó et al. [7] demonstrated that inosine-mediated suppression of TNF-α production is mediated through both A₁R and A_2AR in vitro. To further assess the role of A_2AR in inosine-mediated suppression of cytokine production, we investigated inosine’s effects on cytokine production by monocytes isolated from A_2AR intact and deficient mice. To distinguish adenosine effects from those of inosine, we utilized ADA which converts adenosine to inosine. Consistent with the role of the A_2AR, activation of the receptor with the selective agonist CGS 21680 in the absence and in the presence of ADA inhibited TNF-α production by A_2AR-intact but not A_2AR-deficient monocytes (Fig 5A and 5B). In the absence of ADA, inosine at 30 μM and 100 μM inhibited LPS-stimulated TNF-α production by 18.6 ± 3.7% (p < 0.01) and 30.1 ± 1.3% (p < 0.001) by A_2AR intact monocytes respectively (Fig 5A). The same concentrations of inosine inhibited LPS-stimulated TNF-α production by 5.2 ± 1.3% and 16.3 ± 0.7% (p < 0.001) in A_2AR deficient monocytes. These results indicate that inosine-mediated inhibition of TNF-α production is reduced by 46% (p < 0.01) in A_2AR-deficient monocytes compared with receptor-intact monocytes suggesting that inosine-mediated inhibition of TNF-α at least in part mediated through the A_2AR. The inclusion of ADA in the assay prevents A_2AR activation by adenosine as it converts extracellular adenosine produced by monocytes to inosine. When the confounding effects of adenosine were eliminated, inosine-mediated inhibition of LPS-stimulated TNF-α production continued to be reduced in both A_2AR intact and A_2AR deficient monocytes (Fig 5B). In the presence of ADA, inosine at 30 μM and 100 μM inhibited LPS-stimulated TNF-α production by 13.2 ± 1.9% (p < 0.01) and 16.8 ± 1.6% (p < 0.01) in A_2AR intact monocytes respectively (Fig 5B). The same concentrations of inosine inhibited LPS-stimulated TNF-α production by 5.2 ± 2.2% and 12.4 ± 1.2% (P < 0.01) in A_2AR deficient monocytes, indicating a 2.5- and a 1.4-fold reduction in the inhibition of TNF-α production in A_2AR deficient monocytes in comparison with A_2AR-intact monocytes at 30 μM and 100 μM inosine respectively. These results indicate that inosine-mediated A_2AR activation leads to inhibition of TNF-α production as is the case for adenosine-mediated A_2AR activation. Inhibition of TNF-α production in A_2AR deficient monocytes by inosine and not by A_2AR selective agonist CGS 21680 suggests that inhibitory effects of inosine are also mediated through other adenosine receptor subtype(s).

Fig. 5.

Inosine-mediated A_2AR activation inhibits TNF-α production by mouse splenic monocytes. A_2AR intact and deficient mouse splenic monocytes were stimulated with LPS (50 ng/ml), inosine (Ino) and CGS 21680 (CGS) in the absence (A) and in presence (B) of ADA (3 U/ml) for 4 h. Mean % inhibition of TNF-α production relative to DMSO (control) of a representative experiment performed in triplicate with SEM are shown (n=3; ** and ***, p < 0.01 and 0.001 vs A_2AR^+/+ DMSO; ## and ###, p < 0.01 and 0.001 vs A_2AR^−/− DMSO; ^, ^^ and ^^^, p < 0.05, 0.01 and 0.001 vs A_2AR^−/− respectively).

3.6. PAM of the A_2AR enhances inosine-mediated suppression of pro-inflammatory cytokines production

To examine whether allosteric enhancement of inosine-mediated A_2AR stimulation leads to increased suppression of cytokine production, we evaluated the effects of inosine in the presence and in the absence of AEA061 on TNF-α production by LPS-stimulated monocytes. To assess the potential for A_2AR activation by adenosine produced via the salvage pathway, we performed the assays with and without ADA. Consistent with the anti-inflammatory role of the A_2AR, selective activation of this receptor with CGS 21680 in the absence of ADA inhibited TNF-α production by LPS-stimulated A_2AR-intact monocytes by 39% and the PAM of the A_2AR enhanced this CGS 21680-mediated inhibition to 49% (Fig 6A; p < 0.001). Under the same conditions, CGS 21680 alone and in combination with AEA061 reduced LPS-stimulated production in A_2AR deficient monocytes by only 6.7% and 14% respectively (p < 0.001; Fig 6A) demonstrating the inhibitory effects of these agents are mediated through the A_2AR. Similarly, both inosine and AEA061 as single agents dose-dependently inhibited TNF-α production by LPS-stimulated A_2AR intact monocytes in the absence of ADA. In combination, they produced even greater inhibition than either agent alone indicating allosteric enhancement (Fig 6A). Although the higher concentration of AEA061 (0.3 μM) both alone and in combination produced a greater percent inhibition of TNF-α production by A_2AR intact monocytes than AEA061 at 0.1 μM, the difference in percent inhibition of TNF-α production between A_2AR intact and deficient monocytes had narrowed. This suggests loss of selectivity of AEA061 towards the A_2AR at higher concentrations. In the presence of ADA (Fig 6B), inosine and AEA061 produced the same general pattern of inhibition of TNF-α production by A_2AR intact and deficient monocytes. However, the overall percent inhibition of TNF-α production was somewhat lower in the presence of ADA relative to the absence of ADA. Inhibition of production of the pro-inflammatory chemokine CCL3 by inosine and AEA061 in the absence of ADA (Fig 6C) closely resembles that of TNF-α (Fig 6A) in A_2AR intact and deficient monocytes under the same conditions. However, in the presence of ADA, the difference in CCL3 inhibition (Fig 6D) between A_2AR intact and deficient monocytes was more prominent. Inosine and AEA061 also individually inhibited CCL4 production by monocytes (Fig 6E) and the combination of these two agents produced greater inhibition than each agent alone suggesting allosteric potentiation of the A_2AR. The addition of ADA reduced the inhibition of CCL4 production under the same conditions, but increased the difference in inhibition between A_2AR intact and deficient monocytes (Fig 6F). Collectively, these results indicate that positive allosteric potentiation of inosine-mediated A_2AR activation inhibits pro-inflammatory cytokine/chemokine production by monocytes.

Fig. 6.

PAM of the A_2AR potentiates inosine-mediated inhibition of TNF-α in an A_2AR dependent manner. Splenic monocytes isolated from A_2AR intact and deficient mice were stimulated with LPS (50 ng/ml), inosine (Ino), CGS 21680 (CGS) in the absence (A, C, E) and in presence (B, D, F) of ADA (3 U/ml) for 4 h. Mean percent inhibition of TNF-α (A & B), CCL3 (C & D) and CCL4 (E & F) production relative to DMSO (control) of a representative experiment performed in replicates of four with SEM are shown (n=4; *, ** and ***, p < 0.05, 0.01 and 0.001 vs DMSO; ## and ###, p < 0.01 and 0.001 vs A_2AR^+/+ AEA061; ^^^, p < 0.001 vs CGS 21680; φ, φφ and φφφ, < 0.05, 0.01 and 0.001 vs A_2AR^+/+ respectively).

Discussion

Inosine, a metabolite of adenosine, exerts anti-inflammatory and immunomodulatory effects in vivo [28]. These effects are at least in part mediated through the A_2AR, a member of a class of purinergic G protein-coupled receptors. Inosine is less potent than adenosine in activating the A_2AR [29]. Hence, the prevailing notion is that the inosine effects observed in vivo are not generated by direct inosine engagement at the A_2AR but rather mediated indirectly through receptor activation by adenosine produced from the metabolic conversion of inosine. However, a handful of studies have provided initial circumstantial evidence for direct activation of A_2AR by inosine [7, 13, 27]. Recently, utilizing a combination of label-free, cell-based, and membrane-based functional assays in conjunction with an equilibrium agonist-binding assay we demonstrated that inosine directly engages the A_2AR and activates signaling events downstream of the receptor leading to cAMP production and ERK1/2 phosphorylation [19]. In the present study, we sought to provide additional proof for A_2AR activation through direct engagement of inosine at the receptor as well as determine if inosine-mediated engagement of the receptor was amenable to allosteric modulation.

We employed several approaches to rule out indirect and confounding effects of adenosine on A_2AR activation. We pretreated cells with adenosine deaminase (ADA) to rid the cells of endogenously produced adenosine and adenosine 5′-[α,β-methylene] diphosphate, an inhibitor of ecto-5′-nucleotidase, to halt endogenous/exogenous production of adenosine and inosine. Moreover, we utilized ADA during the treatment phase of the assay to eliminate adenosine produced during the short duration of the A_2AR activation assay. Under these conditions, exogenous inosine as well as the inosine analog NBMPR but not adenosine stimulated activation of the A_2AR. This inosine- and NBMPR-mediated A_2AR activation was dose-dependent and was inhibitable by the A_2AR inverse agonist ZM 241385 providing strong evidence for inosine agonism at this receptor.

Our data indicate that inosine inhibits pro-inflammatory cytokine and chemokine production in an A_2AR-dependent manner in vitro. The level of inhibition of pro-inflammatory cytokine production achieved with inosine in mouse monocytes is consistent with that published by Haskó et al. [7] and is also in agreement with the EC₅₀ values for inosine-mediated A_2AR activation in CHO-hA_2AR cells reported here. These results strongly support the contention that the activation of the A_2AR by inosine leads to a reduction in the production of pro-inflammatory cytokines and chemokines. Our results also indicate that ADA in general reduced the inhibition of pro-inflammatory cytokine and chemokine production in both A_2AR intact and deficient monocytes. This is consistent with the fact that extracellular adenosine produced by monocytes/macrophages during the assay is rapidly deaminated by exogenous ADA to inosine that has lower affinity than adenosine at the A_2AR, hence lower inhibition of TNF-α production by A_2AR intact monocytes. The reduced inhibition of TNF-α production in A_2AR-deficient monocytes in the presence of ADA suggests that inosine is also less potent than adenosine in activating A_2AR-independent pathways [30, 31] leading to inhibition of TNF-α production.

On the basis of crystal structure predictions and site-directed mutagenesis studies, the critical residues that form the adenosine binding pocket in the A_2AR have been identified [32, 33]. The nature of the interactions of inosine with these residues may explain the observed low potency of the receptor to inosine [19, 34]. To better understand the molecular mechanism of A_2AR recognition by adenosine and inosine, Deganutti et al. [35] utilized supervised molecular dynamics simulation, a computational method that allows identification and characterization of multiple stable receptor conformations such as orthosteric, allosteric and meta-binding states. They reported an overlap of meta-stable states predicted for binding of inosine and adenosine to the A_2AR suggesting that both agonists share a common molecular mechanism for receptor activation [35]. Their findings are in agreement with our previous [19] as well as present experimental evidence for functional agonism of inosine at the A_2AR and explain the molecular mechanism of inosine-mediated receptor activation.

Inosine and the inverse agonist ZM 241385 both bind to the same site at the A_2AR with different affinities (inosine has a lower affinity). Consistent with these observations, ZM 241385 at 100 nM completely reversed the effects of 100 μM inosine in CHO-hA2AR cells. However, the same concentration of ZM 241385 (100 nM) was not able to fully reverse the effects mediated by inosine at a higher concentration (1000 μM). This incomplete inhibition is most likely due to mass action as the parental CHO-K1 cells do not express any adenosine receptor subytpes and in addition, the ZM 241385 concentration used was 10,000–fold less than that of inosine. Partial inhibition of NBMPR-mediated A_2AR activation by ZM 241385 may be explained also by the law of mass action.

The molecular mechanism of A_2AR activation is consistent with conformational selection where binding of agonists and inverse agonists/antagonists stabilizes the receptor in active and inactive conformations respectively [36]. The active conformation(s) engage Gs and adenylate cyclase to produce cAMP whereas the inactive conformation(s) do not. Binding of a PAM such as AEA061 to the A_2AR facilitates this process as indicated by enhanced efficacy and potency of the agonist at the receptor [20]. Since AEA061 does not activate A_2AR-mediated cAMP production in the absence of an agonist [20], AEA061 by itself does not stabilize A_2AR conformation(s) that engage Gs. Depending on the cell type, agonist engagement at the A_2AR leads to Gs-mediated, cAMP-dependent and small G protein p21^ras-mediated, cAMP-independent ERK1/2 phosphorylation [37, 38]. Therefore, it is conceivable that AEA061 by itself, in the absence of an agonist, stabilizes an A_2AR conformation(s) that is capable of engaging p21^ras to activate ERK1/2, thereby exhibiting an agonist-like behavior.

Previously we demonstrated that the synthetic small molecule AEA061 does not activate the A_2AR by itself but augments adenosine-mediated A_2AR activation, and thus bears the hallmarks of positive allosteric modulation [20]. Our data indicate that AEA061 also enhances inosine-mediated A_2AR responsiveness in the base-line state as well as under conditions that isolate and eliminate the confounding effects of adenosine. Although the PAM of the A_2AR, AEA061, enhances the maximal response of inosine-mediated cAMP production and ERK1/2 phosphorylation under all three assay conditions, it exhibits differential effects with respect to EC₅₀ for cAMP production and ERK1/2 activation. AEA061 reduces the EC₅₀ for inosine-mediated ERK1/2 phosphorylation regardless of the assay condition. In contrast, AEA061 reduces the EC₅₀ for inosine-mediated cAMP production with ADA pretreatment and subsequent ADA addition. It is possible that ADA, in itself an allosteric modulator of the A_2AR [24, 25], differentially influences AEA061 effects on cAMP and ERK1/2 pathways.

Our data indicate that positive allosteric modulation of the A_2AR by AEA061 enhances inosine-mediated inhibition of pro-inflammatory cytokine and chemokine production by splenic monocytes/macrophages. Inhibition of cytokine and chemokine production by AEA061 alone in the absence of an exogenously added agonist is consistent with the fact that cells produce extracellular adenosine and inosine during normal cell metabolism in amounts sufficient to engage the A_2AR allowing allosteric enhancement with AEA061. As AEA061 exhibits no allosteric effects towards A₁R and A₃R and weak activity towards A_2BR (unpublished data), the A_2AR-independent inhibition of cytokine production by AEA061 at higher concentrations may be mediated through the low-affinity A_2BR. Although inosine does not activate A_2BR in stable cell lines overexpressing the receptor in vitro [19, 29], A_2BR is the principal mediator of inosine activity in the bladder [39]. Therefore, it is conceivable that the inhibition of cytokine and chemokine production observed in A_2AR deficient monocytes with AEA061 in the presence of exogenously added ADA is due to enhancement of extracellularly produced inosine mediating A_2BR activation. Further investigation will be necessary to rigorously examine this possibility.

In summary, the major outcomes of the present investigation are the demonstration of inosine agonism at the A_2AR, the amenability of inosine-mediated A_2AR activation to allosteric modulation and the effect of inosine- and PAM- mediated A_2AR activation on pro-inflammatory cytokine and chemokine production. Our data indicate that, as is the case for adenosine, positive allosteric modulation of the A_2AR with AEA061 potentiates inosine-mediated cAMP production, ERK1/2 phosphorylation and inhibition of pro-inflammatory cytokine and chemokine production. We previously proposed that both adenosine and inosine are natural agonists of the A_2AR and are important to mount a robust and effective immunomodulatory response. Adenosine with a shorter half-life (<10 sec) initiates A_2AR activation and upon conversion to metabolically more stable inosine (half-life 15 h) sustains A_2AR signaling to prolong anti-inflammatory and immunomodulatory responses. The PAM of the A_2AR AEA061 potentiates and preserves the natural temporal and physical pattern of both adenosine-mediated and inosine-mediated A_2AR activation to harness the full potential of the endogenous A_2AR-dependent mechanisms tasked to effectively downmodulate inflammation. Given the intricacies of GPCR signaling, it is conceivable that potentiation of adenosine- and inosine-mediated A_2AR activation through a PAM may have qualitatively and quantitatively different immunomodulatory responses as opposed to A_2AR activation through a high-affinity agonist. If proven, positive allosteric modulation of the A_2AR may provide new insights into potential therapeutic interventions for immune- and inflammation-mediated diseases.

Highlights.

• Under conditions devoid of adenosine, inosine dose-dependently activates the A_2AR.

• Inosine-mediated A_2AR activation increases cAMP and phospho-ERK1/2 levels.

• Inosine inhibits TNF-α, CCL3 and CCL4 production by monocytes via A_2AR activation.

• PAM of the A_2AR potentiates inosine-mediated receptor signaling.

• PAM of the A_2AR inhibits TNF-α, CCL3 and CCL4 production.

Abbreviations

ADA

adenosine deaminase

Ado

adenosine

AR

adenosine receptor

A₁R

adenosine A₁R

A_2AR

adenosine A_2A receptor

A_2BR

adenosine A_2B receptor

A₃R

adenosine A₃ receptor

cAMP

cyclic adenosine monophosphate

CGS 21680

4-[2-[[6-Amino-9-(N-ethyl-β-D-ribofuranuronamidosyl)-9H-purin-2-yl]amino]ethyl] benzenepropanoic acid hydrochloride

ENT1

equilibrative nucleoside transporter 1

ENT2

equilibrative nucleoside transporter 2

ERK1/2

extracellular signal-regulated kinase-1 and -2

GPCR

G protein-coupled receptor

HBSS

Hanks’ balanced salt solution

HEPES

4-(2-Hydroxyethyl) piperazine-1-ethanesulfonic acid, N-(2-Hydroxyethyl)piperazine-N′-(2-ethanesulfonic acid

HTRF

homogeneous time resolved fluorescence

Ino

inosine

LPS

lipopolysaccharide

NBMPR

6-S-[(4-Nitrophenyl)methyl]-6-thioinosine

PAM

positive allosteric modulator

PBS

phosphate buffered saline

SD

standard deviation

SEM

standard error of the mean

TNBS

2,4,6-trinitrobenzenesulfonic acid

ZM 241385

4-(2-[7-Amino-2-(2-furyl)[1,2,4]triazolo[2,3-a][1,3,5]triazin-5-ylamino]ethyl)phenol