Positive allosteric modulation of the adenosine A_2a receptor attenuates inflammation

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.

PC-12 cells (ATCC) were grown in DMEM supplemented with 5% FBS, 10% horse serum and amphotericin B (0.25 mg/ml). CHO-hA_2aR cells were grown in DMEM/F-12 (1:1) supplemented with 10% FBS, 2 mM glutamine and G418 (0.2 mg/ml). All cells were maintained at 37°C in a 5% CO₂ incubator.

PC-12 cells were washed with and resuspended in Hanks’ balanced salt solution (HBSS), transferred to 96-well plates (1.4×10⁶ cells per well) and incubated with rolipram (50 μM), adenosine, CGS 21680 and AEA061 at indicated concentration(s) for 15–30 min at 37°C. CHO cells expressing human A_2aR (CHO-hA_2aR; [16]) were seeded in the absence of G418 in 96-well plates (2×10⁴ cells/well) 20 h prior to assay. Assay conditions were essentially the same as above with the exception of inclusion of adenosine deaminase (ADA; 3U/ml) and adenosine 5′ [α,β-methylene] diphosphate (50 μM). Intracellular cAMP levels were quantified using Parameter ELISA kits (R&D Systems).

The binding affinity of adenosine to hA_2aR was evaluated using the [³H]-labeled A_2aR-selective agonist CGS 21680. CHO-hA_2aR cell membranes were prepared as described by Gao et al., [17]. Membranes (10 mg/well) in binding buffer (50 mM Tris pH 7.4, 10 mM MgCl₂, 1 mM EDTA, 1.5 U/ml ADA) were incubated with AEA061 (10 μM) or vehicle (DMSO) in a 96-well WGA FlashPlate at 25°C for 15 min. After addition of [³H]CGS 21680 (0.0-110 nM), plates were incubated for an additional 75 min followed by washing with cold PBS containing AEA061 (10 μM) and [³H]CGS 21680. Binding was quantified using a TopCount scintillation counter. Specific binding was calculated by subtracting counts obtained in the presence of NECA (50 μM) from total counts.

CHO-A_2aR membranes were incubated in assay buffer (75 mM Tris (pH 7.5), 12.5 mM MgCl₂, 1 mM EDTA, 100 mM NaCl, 3 μM GDP and 0.5 % BSA) containing CGS 21680 (0.0 – 1000 nM), ADA (2U/ml) and GTPγ³⁵S (0.4 nM) with or without AEA061 (10 μM) in a 96-well WGA FlashPlate for 3 h at 25°C. Non-specific binding was estimated using GTP-γ-S (10 μM). Plates were washed twice with cold PBS and radioactivity was quantified using a TopCount scintillation counter.

Human monocytes were isolated from buffy coat preparations (Stanford University blood bank) by depleting other cell types using MACS monocyte isolation kit II (Miltenyi Biotech) as suggested by the manufacturer. To isolate mouse splenocytes, spleens were aseptically removed, punctured and rinsed with RPMI 1640 medium to yield a cell suspension. Following red blood cell lysis with ACK buffer (Invitrogen Corporation), cells were seeded in 96-well plates in the same medium. Splenic monocytes/macrophages were isolated by plastic adherence after incubation for 2 h at 37°C and 5% CO₂, followed by washing twice with PBS to remove nonadherent cells. Monocytes/macrophages were seeded in 96-well plates (2x10⁵ cells per well) and stimulated with phorbol 12-myristate 13-acetate (PMA; 5 ng/ml) or lipopolysaccharide (LPS; 5 – 50 ng/ml) in RPMI 1640 containing 1% FBS for 12 h in the presence of indicated concentration(s) of either AEA061 or NECA. For experiments evaluating the effects of A_2aR antagonists, cells were pretreated for 15 min with the antagonist ZM 241385 at 37°C before stimulation with PMA. Splenocytes were seeded in 96-well plates (2×10⁵ cells per well) and stimulated with LPS (50 ng/ml) in DMEM containing 1% FBS for 18 h in the presence of indicated concentrations of AEA061. Cytokine/chemokine levels in the culture supernatants were quantified using Quantikine ELISA kits (R&D Systems) and/or bead-based multiplex immunoassays (Eve Technologies).

Male 8–10 weeks old A_2aR null mice (C;129-Adora2atm1Jfc/J; Jackson laboratories) and receptor competent BALB/cJ control mice (Jackson laboratories; n = 10-14 per group) were dosed intravenously with AEA061 (0.1, 1 and 10 mg/kg) 10 min prior to i.p. injection of LPS (20 mg/kg). Two h after LPS dosing, mice were terminally bled under anesthesia using cardiac puncture technique. Plasma was isolated using lithium heparin tubes and stored at -80°C. Plasma cytokine/chemokine levels were quantified using Quantikine ELISA kits (R&D Systems) and/or bead-based multiplex immunoassays (Eve Technologies).

Data were analyzed using GraphPad Prism Software. Dose response curves were generated by non-linear regression with a variable slope. Synergism between adenosine and AEA061 was analyzed using one-way ANOVA with the Bonferroni correction.

PC-12 cells, derived from a pheochromocytoma of rat adrenal medulla, are known to exhibit high Gαs-coupled AR activity. Agonist-mediated activation of the ARs in these cells leads to a significant rise in intracellular cAMP levels. We took advantage of this quantifiable cellular response to screen derivatives of a small molecule of plant origin for the ability to potentiate adenosine-mediated activation of endogenous ARs in intact PC-12 cells. To rule out the possibility that the elevated levels of cAMP were due to inhibition of phosphodiesterase 4 (PDE4) by the compounds, we utilized a saturating dose of a prototypical PDE4 inhibitor, rolipram. Using this screening paradigm, we have identified a family of non-agonist small molecules that modulate AR activity. One member of this family, AEA061 [15], enhanced adenosine-mediated cAMP production by PC-12 cells in a dose-dependent manner (Figure 1A). The synergism between AEA061 and adenosine was significant at 1 μM of the compound with adenosine levels ≥1 μM (p <0.05). Higher concentrations of AEA061 produced significant synergism at all concentrations of adenosine (p <0.05). On the basis of the EC₅₀ values, AEA061 (10 μM) potentiated adenosine-mediated cAMP production by 11.5 fold. This AEA061-mediated shift in agonist EC₅₀ was saturable, as 100 μM of AEA061 did not yield a greater effect. In addition, AEA061 increased the maximal response by 1.9, 2.4 and 4.2 fold at 1, 10 and 100 μM respectively. These results demonstrate that the affinity as well as the efficacy of adenosine at the endogenous Gαs-coupled adenosine receptors is enhanced by AEA061.

There are two Gαs-coupled adenosine receptor subtypes (A_2aR and A_2bR) whose activation leads to an increase in intracellular cAMP levels [18]. Although there are reports of A_2bR expression by PC-12 cells, A_2aR appears to be the predominant adenosine receptor subtype. To evaluate the role of A_2aR and/or A_2bR in AEA061-mediated enhancement of cAMP production, PC-12 cells were treated with the A_2aR antagonist ZM 241385 and the A_2bR antagonist MRS 1754. As shown in Figure 1B, adenosine increased cAMP production by 4.3 fold over the control and AEA061 (10 μM) enhanced cAMP levels by an additional 11.7 fold. As expected, ZM 241385 (1 nM and 10 nM) blocked adenosine-mediated cAMP production. In addition, ZM 241385 at 10 nM, but not 1 nM, blunted the adenosine plus AEA061-mediated cAMP production. In contrast, MRS 1754, an A_2bR-specific antagonist, failed to inhibit adenosine- or adenosine plus AEA061-induced cAMP production. These results support the notion that enhanced-cAMP production by PC-12 cells in the presence of AEA061 is due to potentiation of rA_2aR but not rA_2bR activity.

To further investigate the observation that AEA061-dependent potentiation of cAMP production by PC-12 cells is due to enhanced-activation of the endogenous rA_2aR, we evaluated the effects of AEA061 on cAMP production stimulated by the A_2aR-selective adenosine analog CGS 21680. As shown in Figure 1C, CGS 21680 increased cAMP production in a dose-dependent fashion while the addition of AEA061 significantly enhanced it. Synergism between CGS 21680 and AEA061 was evident at 1 nM of CGS 21680 and above (p <0.05 – p <0.001), confirming our observation that the AEA061-dependent increase in intracellular cAMP levels is mediated through the endogenous rA_2aR. Of note, AEA061, even at 100 μM, did not induce cAMP production in the absence of the agonist, indicating that the compound has no intrinsic activity and that AEA061-mediated enhancement of rA_2aR activity is due to allosterism rather than agonism at the receptor.

Since the human A_2aR (hA_2aR) is the most pharmacologically characterized A_2aR [19] relative to other species, we investigated the hA_2aR for potential allosteric modulation. As A_2aR is evolutionarily conserved in terms of sequence, structure and function, new knowledge gained here will serve as a guide to the A_2aRs of other species. To establish whether the hA_2aR is amenable to functional enhancement, we evaluated the effects of AEA061 on cAMP production by CHO-hA_2aR cells. As shown in Figure 2A, AEA061 dose-dependently increased cAMP production by these cells in the presence but not in the absence of adenosine. These results demonstrate that AEA061 has no intrinsic activity towards the hA_2aR, but can allosterically augment agonist-mediated hA_2aR activity. On the basis of EC₅₀ values, AEA061 enhanced the adenosine-mediated response by 2.8 fold (Figure 2B). Moreover, AEA061 also increased the maximal response by 2.1 fold. These results establish that, similar to the rA_2aR, positive allosteric modulation of the hA_2aR causes enhanced adenosine potency and efficacy at the receptor.

The A_2aR couples to two different classes of Gα subunits. A_2aR coupling to the Gαs and Gαolf subunits that belong to the Gs class stimulates adenylate cyclase to increase intracellular cAMP upon agonist engagement at the receptor. In contrast, Gα15 and Gα16 subunits that belong to the Gq class couple the A_2aR receptor to phospholipase C to stimulate the production of inositol phosphates upon receptor activation. A_2aR-mediated activation of Gα subunits can be measured by incorporation of a radiolabeled non-hydrolysable GTP analog, GTPγ³⁵S, into membrane preparations bearing the hA_2aR. To investigate whether AEA061 enhances hA_2aR-coupled Gα activation, agonist-induced increases in GTPγ³⁵S incorporation were quantified in the presence or absence of the compound. As shown in Figure 3A, AEA061 dose-dependently enhanced CGS 21680-induced GTPγ³⁵S incorporation. Analysis of the agonist dose response indicated that AEA061 potentiated agonist-mediated Gα activation by 3.7 fold. These results demonstrate that the responsiveness of agonist-mediated hA_2aR-coupled Gα activation is enhanced by AEA061 and thus links AEA061 directly to A_2aR activation.

In general, as a functional consequence of binding, allosteric modulators alter dissociation kinetics of the agonist. To investigate whether enhanced A_2aR activation by AEA061 in the presence of agonist was due to altered agonist binding at the receptor, equilibrium binding studies were performed using the radiolabeled A_2aR-selective agonist CGS 21680 and membrane preparations containing hA_2aR. An enhancement in receptor affinity (1.8 fold) and B_max (3 fold) was observed with AEA061 (Figure 4). Taken together, these results demonstrate that AEA061 augments the affinity as well as the number of agonist binding sites at the hA_2aR.

Activation of A_2aR has been shown to inhibit TNF-α production by monocytes/macrophages [20]-[22]. We therefore investigated the effect of allosteric enhancement of A_2aR activity on TNF-α production mediated by two different stimuli acting through two distinct signaling pathways: stimulation of the TLR4 receptor by LPS, and protein kinase C (PKC)/mitogen-activated protein kinase (MAPK) activation by PMA. Consistent with the role of allosteric enhancement of the hA_2aR, AEA061 inhibited TLR4-mediated TNF-α production by freshly-isolated human and mouse monocytes in a dose-dependent manner with IC₅₀ values of 1.42 μM (Figure 5A) and 1.6 ± 0.44 μM (Figure 5B) respectively. AEA061 also dose-dependently inhibited TNF-α production mediated by PKC/MAPK activation in freshly isolated human monocytes (IC₅₀ = 1.85 ± 0.9 μM, Figure 5C). In addition, we evaluated the effects of A_2aR positive allosteric modulation on the production of cytokines and chemokines by LPS-stimulated splenocytes. Of interest, the production and/or release of KC, IL-1α, MIP-1α, MIP-1β, MIP-2, RANTES and TNF-α were consistently inhibited by allosteric enhancement of the A_2aR in a dose-dependent manner (Figure 6) while additional cytokines and chemokines remained unchanged (data not shown). A standard battery of preclinical safety pharmacology tests on AEA061 did not reveal any toxicity or side effects, hence ruling out cytotoxicity as a possible cause for the inhibition of cytokine/chemokine production. Taken together, these results suggest that positive allosteric modulation of the A_2aR of monocytes and splenocytes is sufficient to enhance receptor responsiveness to endogenous adenosine and inhibit production of multiple pro-inflammatory cytokines and chemokines.

To examine whether the inhibition of TNF-α production by AEA061 is mediated through selective allosteric enhancement of the A_2aR, we evaluated the effects of AEA061 in the presence and absence of the A_2aR-specific antagonist, ZM 241385, on TNF-α production by human monocytes. As a control, we used the non-selective high affinity AR agonist, NECA, in parallel experiments. NECA inhibited PMA-stimulated TNF-α production by freshly isolated human monocytes/macrophages and the A_2aR-selective antagonist, ZM 241385, reversed the NECA-mediated inhibition in a dose-dependent manner (Figure 7A). Similarly, ZM 241385 blunted the inhibition of PMA-stimulated TNF-α production enhanced by AEA061 in a dose-dependent manner (Figure 7B). The A_2aR antagonist, ZM 241385, at 1 μM restored PMA-stimulated TNF-α responsiveness in both NECA- and AEA061-treated cells. These results support the notion that selective allosteric enhancement of the A_2aR of monocytes renders them more responsive to endogenous adenosine thereby further enhancing adenosine’s anti-inflammatory activity.

To further investigate the notion that allosteric enhancement significantly augments endogenous adenosine-mediated A_2aR activation and subsequent inhibition of inflammation in vivo, we utilized a mouse model of endotoxemia. A_2aR null mice and receptor competent BALB/cJ control mice were injected intraperitoneally with LPS, and after 2 h plasma was collected for the measurement of cytokine and chemokine levels. In LPS-stimulated receptor competent BALB/cJ mice, allosteric enhancement of the A_2aR with AEA061 at 1 and 10 mg/kg significantly reduced plasma TNF-α and MCP-1 levels compared to vehicle treatment (Figure 8A and B). Although AEA061 administration reduced plasma levels of IL-12, p40, MIP1-α, MIP-2, MIG and M-CSF relative to that of the untreated animals, the level of reduction did not reach statistical significance (data not shown). In contrast, AEA061 significantly enhanced the plasma levels of the anti-inflammatory cytokine, IL-10, at 0.1 and 10 mg/kg (Figure 8C). These results demonstrate that positive allosteric modulation of the A_2aR enhances receptor responsiveness to endogenous adenosine and down-regulates the inflammatory cytokine cascade in vivo. The group of A_2aR null mice that received vehicle alone had higher circulating plasma TNF-α (Figure 8D) and lower MCP-1 (Figure 8E) and IL-10 (Figure 8F) levels relative to their BALB/cJ receptor-competent counterparts. AEA061 at any concentration did not alter plasma levels of TNF-α, MCP-1 or IL-10 in LPS-treated A_2aR null mice. These results confirm that selective allosteric enhancement of the A_2aR by AEA061 is responsible for the anti-inflammatory effects observed in vivo. It is unclear why AEA061 enhanced IL-10 production at 0.1 and 10 kg/ml in receptor competent BALB/cJ mice yet failed to enhance IL-10 production at 1 mg/kg while the same dose of 1 mg/kg significantly affected TNF-α and MCP-1 production. More in depth studies are needed to clarify this observation.