Effect of A_2B Adenosine Receptor Gene Ablation on Adenosine-Dependent Regulation of Proinflammatory Cytokines

There is growing evidence that adenosine plays an important role in the regulation of inflammation. Adenosine is an intermediate product of adenine nucleotide metabolism, and it also serves as a signaling molecule that can modulate cellular functions via binding to cell surface G-protein-coupled receptors of the P1 purinergic family comprising A₁, A_2A, A_2B, and A₃ adenosine receptor subtypes. Among adenosine receptors, the A_2B subtype has the lowest affinity to adenosine. Although A_2B receptors are likely to remain silent under normal physiological conditions (Fredholm et al., 2001b), they can be activated during inflammation when interstitial adenosine concentrations are increased as a result of cell stress, injury, and tissue hypoxia (Fredholm et al., 2001a).

Many studies suggest that A_2B receptors are involved in adenosine-dependent regulation of proinflammatory paracrine factors. We have shown previously that stimulation of A_2B receptors in human mast cell line HMC-1 increased production of proinflammatory cytokines interleukin (IL)-1β, IL-3, IL-4, IL-8, and IL-13 (Feoktistov and Biaggioni, 1995; Ryzhov et al., 2004). Studies in human primary cell cultures demonstrated that A_2B receptors increased monocyte chemoattractant protein-1 and IL-6 release from airway smooth muscle cells and fibroblasts, thus mediating the putative proinflammatory and profibrotic actions of adenosine in asthma (Zhong et al., 2004, 2005). Pharmacological inhibition of A_2B receptors significantly reduces elevations in proinflammatory cytokines as well as mediators of fibrosis and airway destruction induced by high adenosine levels in the lungs of adenosine deaminase-deficient mice (Sun et al., 2006b). Stimulation of proinflammatory cytokines via A_2B receptors is not limited to the lung; it also has been observed in various cell types of different origins. A_2B receptors stimulated IL-8 production in human microvascular endothelial (Feoktistov et al., 2002) and glioblastoma (Zeng et al., 2003) cell lines. A_2B receptors were implicated in the stimulation of IL-6 release in osteoblasts (Evans et al., 2006), intestinal epithelial cells (Sitaraman et al., 2001), pituitary folliculostellate cells (Rees et al., 2003), astrocytes (Schwaninger et al., 1997), astrocytoma cells (Fiebich et al., 2005), and astroglioma cells (Fiebich et al., 1996). For this and other reasons, A_2B receptors have been suggested to mediate proinflammatory actions of adenosine.

Contrary to this body of evidence, a recent report on a mouse phenotype resulting from deletion of the A_2B receptor gene suggested that the A_2B receptor protects against inflammation and excessive vascular adhesion (Yang et al., 2006). This conclusion was based on low-grade inflammation observed in the A_2B adenosine receptor knockout (A_2BKO) mice at rest that was manifested by increased leukocyte adhesion to the vascular wall and increased expression of adhesion molecules in vascular endothelium as a consequence of elevated tumor necrosis factor (TNF)-α plasma levels. Lipopolysaccharide (LPS)-induced elevations of cytokine levels were also exaggerated in A_2BKO mice. Bone marrow-derived cells, particularly macrophages, were identified as a primary source of increased cytokine production in A_2BKO mice. Interpretation of these results as evidence of anti-inflammatory actions of A_2B receptors seems surprising not only because it contradicts previous pharmacological evidence but also because a baseline inflammatory phenotype in A_2B receptor knockout animals assumes tonic stimulation of A_2B receptors, which is unlikely given their low affinity for adenosine (Linden, 2006). Therefore, alternative explanations could be considered. For example, the observed changes could be interpreted as a result of developmental adaptation to the genetic removal of the A_2B receptor, or they could reflect the loss of a previously unrecognized function of the A_2B receptor protein independent of adenosine signaling.

In this context, it should be noted that the effects of direct agonist stimulation of adenosine receptors on proinflammatory cytokine production have not been studied in A_2BKO mice. Without this information, it is difficult to determine whether the reported inflammatory phenotype in A_2BKO mice indeed reflects the loss of adenosine-dependent signaling functions of A_2B receptors or points toward the alternative interpretations outlined above. In the current study, we chose two well recognized immunomodulatory actions of adenosine and investigated the effects of the A_2B receptor gene ablation on adenosine-dependent stimulation of IL-6 release and suppression of LPS-induced TNF-α secretion (Haskó et al., 1996; Ritchie et al., 1997) by directly activating adenosine receptors in wild-type and A_2BKO mice.

Materials and Methods

Reagents. LPS (from Escherichia coli, serotype 055:B5), 5′-N-ethylcarboxamidoadenosine (NECA), and MRS 1754 were purchased from Sigma (St. Louis, MO). 3-Isobutyl-8-pyrrolidinoxanthine (IPDX) was synthesized as described previously (Feoktistov et al., 2001). Dimethyl sulfoxide (DMSO) was purchased from Sigma. When used as a solvent, final DMSO concentrations in all assays did not exceed 0.1%, and the same DMSO concentrations were used in vehicle controls.

Mice. Six- to 8-week-old age- and sex-matched mice were used. A_2BKO mice were obtained from Deltagen (San Mateo, CA), and wild-type (WT) C57BL/6 mice were purchased from Harlan World Headquarters (Indianapolis, IN). Genotyping protocols for A_2BKO were described previously (Csoka et al., 2007). All of the A_2BKO mice used in these studies were backcrossed to the C57BL/6 genetic background for 10 generations.

Real-Time Reverse Transcription-Polymerase Chain Reaction. Real-time reverse transcription-polymerase chain reaction (RT-PCR) analysis was performed as described previously (Ryzhov et al., 2007). Total RNA was isolated from cells using the RNeasy Mini Kit (QIAGEN, Valencia, CA). Real-time RT-PCR was carried out on an ABI PRISM 7900HT Sequence Detection System (PE Applied Biosystems, Foster City, CA). Primer pairs and FAM-labeled probes for murine adenosine receptors and β-actin were provided by Applied Biosystems. RT-PCR was performed under conditions recommended by the manufacturer. A standard curve for each amplicon was obtained using serial dilutions of total RNA. The results from triplicate polymerase chain reactions for a given gene at each time point were used to determine the mRNA quantity relative to the corresponding standard curve. The relative mRNA quantity for a given gene measured from a single reverse-transcription reaction was divided by the value obtained for β-actin to correct for fluctuations in input RNA levels and varying efficiencies of reverse-transcription reactions.

Measurement of cAMP Accumulation. cAMP accumulation was measured as described previously (Feoktistov et al., 2002). Cells in 12-well plates were preincubated in 150 mM NaCl, 2.7 mM KCl, 0.37 mM NaH₂PO₄, 1 mM MgSO₄, 1 mM CaCl₂, 5 g/l d-glucose, 10 mM HEPES-NaOH, pH 7.4, and 1 U/ml adenosine deaminase containing the cAMP phosphodiesterase inhibitor papaverine (1 mM) for 15 min at 37°C. Adenosine agonists, forskolin, or their vehicle (DMSO) were added to cells, and the incubation was allowed to proceed for 10 min at 37°C. The reaction was stopped by the addition of one-fifth volume of 25% trichloroacetic acid. The extracts were washed five times with 10 volumes of water-saturated ether. cAMP concentrations were determined using a cAMP assay kit (GE Healthcare, Little Chalfont, UK). In parallel measurements, total protein in macrophages was determined using a Coomassie Plus Bradford assay (Pierce, Rockford, IL), and intracellular cAMP levels were expressed as picomoles per milligram of protein.

Analysis of IL-6 and TNF-α Secretion from Macrophages. Freshly isolated macrophages were incubated in the presence or absence of LPS and in the presence or absence of NECA and the A_2B receptor antagonists in RPMI 1640 medium with 10% calf serum, 1× antibiotic antimycotic mixture (Sigma), and 1 U/ml adenosine deaminase for 16 h at 37°C. After collection of culture media, the cells were lysed with 0.4 N NaOH and assayed for total protein using a Coomassie Plus Bradford assay (Pierce). IL-6 and TNF-α concentrations in culture media were measured using ELISA kits (R&D Systems, Minneapolis, MN). Released cytokines were expressed as picograms per milligram of protein.

Animal Procedures. All studies were conducted in accordance with the Institute of Laboratory Animal Resources (1996) as adopted and promulgated by the United States National Institutes of Health.

Mice received i.p. injections with LPS (10 mg/kg), NECA (0.5 mg/kg), or its vehicle (DMSO) in sterile phosphate-buffered saline (PBS). NECA was injected 10 min before LPS. In another set of experiments, mice received i.p. injections with NECA (0.5 mg/kg) or its vehicle (DMSO) in PBS without subsequent LPS injection. Blood was collected from the retro-orbital vein using heparinized Natelson blood collecting tubes (Fisher Scientific, Pittsburgh, PA) 1.5 h after LPS injection for TNF-α assay or 3 h after NECA injection for IL-6 assay, and cytokine levels were determined from the isolated plasma. These time points were chosen from ancillary studies that showed that the maximal plasma levels of TNF-α and IL-6 were reached at 1.5 h after LPS injection and 3 h after NECA injection, respectively.

Isolation of Mouse Peritoneal Macrophages. Mice received i.p. injections with 2 ml of 3% thioglycollate. After 4 days, peritoneal cells collected by lavage were seeded onto 24-well plates in RPMI 1640 medium with 10% calf serum and 1× antibiotic antimycotic mixture (Sigma) for 2 h to allow the macrophages to adhere to the plates. Nonadherent cells were subsequently removed by washing with RPMI 1640 medium, and the adherent macrophages were used for experiments immediately after isolation.

Statistical Analysis. Data were analyzed using the GraphPad Prism 4.0 software (GraphPad Software Inc., San Diego, CA) and presented as mean ± S.E.M. Comparisons between animal groups (wild-type versus A_2BKO mice) were performed using two-tailed paired and unpaired t tests. Comparisons between groups treated with A_2B receptor antagonists and a control (vehicle) group were performed using one-way analysis of variance (ANOVA) followed by Dunnett's post-tests. In studies comparing the effects of stimulation of adenosine receptors on different groups of animals, we used a two-factor ANOVA followed by Bonferroni's post-tests. Factor A was the intervention (stimulation with NECA or no stimulation), fixed, and two levels. Factor B was the animal group (wild-type versus A_2BKO mice), fixed, and two levels. Of particular interest was the interaction between the two factors; i.e., whether the animal group modified the response to NECA. A p value <0.05 was considered significant.

Results

Effect of Adenosine A_2B Receptor Gene Ablation on Adenosine Receptor mRNA Expression in Peritoneal Macrophages. Real-time RT-PCR analysis of wild-type macrophages revealed preferential expression of mRNA encoding A_2A receptor subtype (0.96 ± 0.18% β-actin), with lower expression of A_2B and A₃ receptor subtypes (0.43 ± 0.11 and 0.3 ± 0.05% β-actin, respectively) and no detectable levels of A₁ receptor transcripts (Fig. 1). As expected, we did not detect the expression of A_2B receptor mRNA in A_2BKO macrophages. We also documented that A_2B receptor gene ablation had no significant effect on A_2A and A₃ receptor mRNA expression in A_2BKO macrophages (0.88 ± 0.2 and 0.25 ± 0.05% β-actin, respectively).

Effect of Adenosine A_2B Receptor Gene Ablation on cAMP Levels in Peritoneal Macrophages. Adenosine receptor subtypes were initially characterized by their effects on adenylate cyclase activity; A_2A and A_2B receptors stimulate adenylate cyclase via coupling to G_s proteins, whereas A₁ and A₃ receptors inhibit stimulation of this enzyme via coupling to G_i proteins (Fredholm et al., 2001a). Therefore, we questioned whether A_2B receptor gene ablation had any effect on cAMP levels in peritoneal macrophages. As seen in Fig. 2, comparative analysis revealed no significant difference in basal cAMP levels between macrophages isolated from wild-type and A_2BKO mice (0.563 ± 0.06 and 0.689 ± 0.036 pmol/mg, respectively; Fig. 2). Likewise, we found no significant difference in adenosine receptor-mediated cAMP accumulation between macrophages isolated from wild-type and A_2BKO mice. Stimulation of adenosine receptors with the stable adenosine analog NECA (10 μM) increased cAMP levels in wild-type and A_2BKO macrophages by 56 ± 7 and 41 ± 9%, respectively. These effects were relatively small compared with more robust effects of the adenylate cyclase stimulator forskolin (1 μM), which was used as a positive control (Fig. 2). Thus, our results suggest that A_2B receptor gene ablation has negligible effects on the regulation of cAMP levels in mouse peritoneal macrophages.

Effect of Adenosine A_2B Receptor Gene Ablation on IL-6 Secretion. Incubation of quiescent macrophages for 16 h resulted in accumulation of low levels of IL-6 in cell culture media (Fig. 3A). Adenosine is known to stimulate IL-6 production in various cells, including peritoneal macrophages (Ritchie et al., 1997). In agreement with previously published data, 10 μM NECA stimulated IL-6 generation in wild-type macrophages, resulting in a 6-fold increase in IL-6 production from 61 ± 14 to 361 ± 57 pg/mg. Although A_2B receptor gene ablation had no effect on basal release of IL-6, it almost completely abrogated the NECA-induced IL-6 secretion from A_2BKO macrophages (Fig. 3A). Two-way ANOVA demonstrated significant interaction (p = 0.0003) between adenosine A_2B receptor gene ablation and NECA effects. In vivo, i.p. injection of 0.5 mg/kg NECA to wild-type mice produced significantly higher IL-6 plasma levels (171 ± 17 pg/ml) compared with IL-6 plasma levels (14 ± 2 pg/ml) observed in the control group of wild-type mice that received injections with vehicle (0.1% DMSO in PBS). Although A_2B receptor gene ablation had no significant effect on IL-6 plasma levels in control mice that received injections with vehicle, it dramatically reduced the ability of NECA to raise IL-6 plasma levels in A_2BKO mice (Fig. 3B). Two-way ANOVA confirmed significant (p = 0.004) interaction between these factors. Thus, we concluded from these studies that A_2B receptors are essential for adenosine-dependent upregulation of IL-6.

Effect of Adenosine A_2B Receptor Gene Ablation on TNF-α Secretion. In addition to stimulation of IL-6 production, adenosine is known to suppress TNF-α release induced by LPS (Haskó et al., 1996; Ritchie et al., 1997; Haskó and Cronstein, 2004; Kreckler et al., 2006). Our results showed that quiescent A_2BKO macrophages secreted higher levels of TNF-α compared with macrophages isolated from wild-type mice (72 ± 4 and 43 ± 11 pg/mg, respectively; p < 0.05). Stimulation of adenosine receptors with 10 μM NECA had no significant effect on basal TNF-α release from both wild-type and A_2BKO macrophages (Fig. 4A). In contrast, NECA significantly suppressed TNF-α release from both wild-type and A_2BKO macrophages activated by 10 ng/ml LPS (Fig. 4B). However, two-way ANOVA revealed no interaction between A_2B receptor gene ablation and the effects of NECA (p = 0.5). In other words, the absence of A_2B adenosine receptors did not change the suppression of LPS-activated TNF-α release in response to stimulation of macrophage adenosine receptors by NECA. In vivo, i.p. injection of 0.5 mg/kg NECA to wild-type mice before challenge with 10 mg/kg LPS significantly attenuated TNF-α plasma levels compared with those induced by LPS alone. This inhibitory effect of NECA was even greater in A_2BKO mice (Fig. 4C), indicating that A_2B receptors are not involved in adenosine-dependent suppression of LPS-induced TNF-α release.

Effect of Adenosine A_2B Receptor Antagonists on IL-6 and TNF-α Secretion. In a complementary approach, we chose two structurally unrelated selective A_2B antagonists IPDX and MRS 1754 (Feoktistov et al., 2001; Ji et al., 2001) to analyze the role of A_2B receptors in the regulation of inflammatory cytokines in peritoneal macrophages. We initially confirmed that both IPDX and MRS 1754 significantly inhibited NECA-induced IL-6 secretion, but they had no effect on the basal cytokine release in the absence of NECA (Table 1). In contrast, both IPDX and MRS 1754 failed to antagonize NECA-dependent suppression of LPS-induced TNF-α secretion. Albeit not statistically significant, there was a tendency for both selective A_2B antagonists to attenuate LPS-induced TNF-α release in the absence or presence of NECA (Table 2).

Discussion

Adenosine released under inflammatory conditions can exert anti-inflammatory or proinflammatory actions, depending on the specific adenosine receptor subtype involved. The low-affinity A_2B receptor was known to mediate proinflammatory effects of adenosine by up-regulating cytokines and growth factors. This view has been supported by a large body of evidence provided by pharmacological analysis of adenosine-dependent cytokine and growth factor secretion in various cells, tissues, and organs (Feoktistov and Biaggioni, 1995; Fiebich et al., 1996, 2005; Schwaninger et al., 1997; Feoktistov et al., 2002; Rees et al., 2003; Zeng et al., 2003; Ryzhov et al., 2004; Zhong et al., 2004, 2005; Evans et al., 2006; Sun et al., 2006b). This concept has been recently challenged by findings that A_2BKO mice exhibit moderate inflammation primarily caused by elevated basal and LPS-stimulated plasma TNF-α (Yang et al., 2006). This led the authors to suggest that the A_2B receptor attenuates inflammation. However, it is unclear whether these results actually reflect the loss of the putative taming of proinflammatory cytokine production via activation of A_2B receptors by endogenous adenosine.

We revisited these conclusions by studying adenosine-dependent regulation of cytokine secretion in A_2BKO and wild-type mice. We conducted studies in peritoneal macrophages previously identified as a major source of elevated plasma TNF-α in A_2BKO mice (Yang et al., 2006) and in vivo using previously established models of adenosine receptor-dependent IL-6 secretion (Weiterová et al., 2007) and suppression of LPS-induced TNF-α production (Parmely et al., 1993; Salvatore et al., 2000; Gomez and Sitkovsky, 2003).

Our data show that engagement of A_2B receptors by agonist leads to proinflammatory events that can be blocked by selective A_2B antagonists or by genetic ablation of A_2B receptors. We determined that A_2B receptors mediate stimulation of IL-6 production, but we found that they are not involved in suppression of LPS-induced TNF-α secretion by adenosine. Stimulation of adenosine receptors with the stable adenosine analog NECA up-regulated IL-6 and suppressed LPS-induced TNF-α in wild-type mice. Two structurally different selective A_2B antagonists IPDX and MRS 1754 both inhibited NECA-induced IL-6 release, but they had no significant effect on the ability of NECA to suppress LPS-induced TNF-α secretion from peritoneal macrophages. In this respect, our results correlate well with a recent work of Kreckler et al. (2006), which demonstrated that the potency and efficacy of NECA to inhibit LPS-induced TNF-α secretion from wild-type mouse peritoneal macrophages are not influenced by blocking A_2B receptors with MRS 1754.

Genetic ablation of A_2B receptors completely abrogated NECA-induced increase in IL-6 release from peritoneal macrophages and dramatically reduced the ability of NECA to raise IL-6 plasma levels in vivo. In contrast, the absence of A_2B adenosine receptors did not affect adenosine receptor-dependent suppression of LPS-activated TNF-α release in peritoneal macrophages, nor did it reduce the ability of NECA to suppress LPS-induced increase in TNF-α plasma levels in vivo. Thus, our results obtained in A_2BKO mice confirm the previous concept stemming from pharmacological studies that A_2B receptors mediate proinflammatory effects of adenosine. However, our data do not support the hypothesis that the A_2B receptor mediates tonic anti-inflammatory effects of adenosine in wild-type mice by suppressing TNF-α secretion.

Mouse peritoneal macrophages express A_2A, A_2B, and A₃ adenosine receptors. The dominant role of A_2A receptors in inhibiting LPS-induced TNF-α production in these cells has been established previously (Kreckler et al., 2006). The anti-inflammatory role of A_2A receptors also has been shown in vivo, because selective A_2A agonists attenuate the increase in plasma TNF-α levels induced by LPS in wild-type mice but have no effect in A_2A adenosine receptor knockout (A_2AKO) animals (Gomez and Sitkovsky, 2003). Our data show that ablation of the A_2B receptor gene had no effect on the expression of mRNA encoding A_2A receptors in peritoneal macrophages and on NECA-induced suppression of LPS-activated TNF-α release. Although both A_2A and A_2B receptors can stimulate adenylate cyclase, we found that NECA-induced cAMP accumulation was similar in peritoneal macrophages obtained from wild-type and A_2BKO animals, indicating the dominant role of A_2A receptors in this process. Of interest, elevation of cAMP in murine macrophages attenuates LPS-induced TNF-α secretion (Mauel et al., 1995), but it has no effect on basal IL-6 release (Martin and Dorf, 1991; Tang et al., 1998). Because A_2B receptors can couple to other intracellular pathways, e.g., to G_q-phospholipase C pathway not activated by A_2A receptors (Feoktistov and Biaggioni, 1995; Ryzhov et al., 2006), it is possible that the differential regulation of TNF-α and IL-6 secretion by A_2A and A_2B receptors in mouse peritoneal macrophages can be explained by the coupling of these receptors to distinct intracellular pathways. Additional studies are needed to delineate the signaling pathways linking activation of adenosine receptors to modulation of IL-6 and TNF-α production.

Our study confirmed the previous observation (Yang et al., 2006) that A_2BKO mice have an increased basal and LPS-stimulated TNF-α release from macrophages and higher TNF-α plasma levels. However, our findings are in disagreement with the interpretation that this reflects activation of A_2B receptors by endogenous adenosine suppressing TNF-α release in wild-type mice. Several lines of evidence argue against this possibility. 1) Our experiments in macrophages were conducted in the presence of adenosine deaminase, thus excluding a possibility of receptor activation by endogenous adenosine. Nevertheless, we observed greater basal and LPS-stimulated TNF-α release from A_2BKO macrophages compared with their wild-type control. 2) The A_2BKO phenotype was not mimicked by selective A_2B antagonists IPDX or MRS 1754 in wild-type mice, because they did not induce an increase in basal or LPS-stimulated TNF-α secretion from peritoneal macrophages; on the contrary, these selective antagonists blocked adenosine-induced IL-6 release. 3) If it is assumed that tonic activation of A_2B receptors has anti-inflammatory effects, a similar pattern should be observed in A_2AKO mice, given the overwhelming evidence of the anti-inflammatory role of this receptor subtype and its much higher affinity to adenosine (Fredholm et al., 2001a,b). Paradoxically, it has been found that LPS-induced release of TNF-α from peritoneal macrophages (Kreckler et al., 2006) or serum TNF-α levels induced by i.p. LPS challenge in vivo (Gomez and Sitkovsky, 2003) were decreased in A_2AKO mice. A graphic representation of these findings is shown in Fig. 5. It would be erroneous to conclude that A_2A receptors are proinflammatory simply based on the observation that LPS has less effect on TNF-α release in A_2AKO mice. This would disregard the overwhelming evidence of the anti-inflammatory role of this receptor subtype (for review, see Haskó and Cronstein, 2004). Likewise, we argue that the increase in inflammatory indices observed in A_2BKO mice in the absence of stimulation of adenosine receptors cannot be interpreted as indicative of putative anti-inflammatory effects of activation of the A_2B receptor by adenosine.

Although our experiments argue against an anti-inflammatory effect resulting from activation of A_2B receptors by adenosine, they do not exclude the possibility that these receptors are still associated with adenosine-independent regulation of TNF-α production. An alternative explanation of this phenomenon could be that the A_2B receptor protein interacts with other signaling pathways unrelated to adenosine. Recent evidence suggests that adenosine receptors play a role in the assembly of multiprotein signaling complexes (Sitaraman et al., 2002; Gsandtner et al., 2005; Pacheco et al., 2005; Milojevic et al., 2006; Sun et al., 2006a). Rearrangement of proteins normally coupled to the A_2B receptor as a result of the A_2B knockout may affect regulation of TNF-α synthesis or release. Future studies that define the full makeup of A_2B receptor complexes with associated signaling components are needed to validate this hypothesis. However, this phenomenon seems to be irrelevant to potential adenosine-based therapeutic approaches to inflammation. It has been speculated that the proinflammatory environment observed in A_2BKO mice could be induced with A_2B receptor antagonists (Hua et al., 2007). However, our current study does not support this hypothesis. It is possible that the phenomenon of enhanced basal and endotoxin-stimulated TNF-α release has functional significance in human cases of A_2B mutations or polymorphisms, but this hypothesis remains very speculative.

In conclusion, our study in A_2BKO mice confirmed the concept drawn from previous pharmacological data that stimulation of A_2B receptors with adenosine leads to proinflammatory events based on the observation that A_2B receptors mediate adenosine-dependent secretion of IL-6, a cytokine generally considered to be proinflammatory especially during chronic inflammation (Gabay, 2006). Our data argue against a role of A_2B receptors in suppression of LPS-stimulated TNF-α production by adenosine. An increase in basal and LPS-stimulated TNF-α release observed in A_2BKO mice is not directly related to adenosine signaling through the A_2B receptor, but it may represent a loss of other unidentified functions of this protein or reflect a developmental adaptation to the A_2B receptor gene ablation.