Disruption of the A3 Adenosine Receptor Gene in Mice and Its Effect on Stimulated Inflammatory Cells*

The A3 adenosine receptor (A3AR) is one of four receptor subtypes for adenosine and is expressed in a broad spectrum of tissues. In order to study the function of A3AR, a mouse line carrying a mutant A3 allele was generated. Mice homozygous for targeted disruption of the A3AR gene,A3AR −/−, are fertile and visually and histologically indistinguishable from wild type mice. The lack of a functional receptor in the A3AR −/− mice was confirmed by molecular and pharmacological analyses. The absence of A3AR protein expression in the A3AR −/− mice was demonstrated by lack ofN 6-(4-amino-3-[125I]iodobenzyl)adenosine binding to bone marrow-derived mast cell membranes that were found to express high levels of A3AR in wild type mice. InA3AR −/− mice, the density of A1and A2A adenosine receptor subtypes was the same as inA3AR +/+ mice as determined by radioligand binding to brain membranes. Additionally, A2B receptor transcript expression was not affected by ablation of theA3AR gene. A3AR −/− mice have basal heart rates and arterial blood pressures indistinguishable fromA3AR +/+ mice. Functionally, in contrast to wild type mice, adenosine and the A3AR-specific agonist, 2-chloro-N 6-(3-iodobenzyl)-adenosine-5′-N-methyl-carboxamide (2-Cl-IB-MECA), elicit no potentiation of antigen-dependent degranulation of bone marrow-derived mast cells fromA3AR −/− mice as measured by hexosaminidase release. Also, the ability of 2Cl-IB-MECA to inhibit lipopolysaccharide-induced tumor necrosis factor-α productionin vivo was decreased in A3AR −/−mice in comparison to A3AR +/+ mice. The A2A adenosine receptor agonist, 2-p-(2-carboxyethyl)phenylamino)-5′-N-ethylcarboxamidoadenosine, produced inhibition of lipopolysaccharide-stimulated tumor necrosis factor-α production in both A3AR −/− andA3AR +/+ mice. These results show that the inhibition in vivo can be mediated by multiple subtypes, specifically the A3 and A2A adenosine receptors, and A3AR activation plays an important role in both pro- and anti-inflammatory responses.

Adenosine is a naturally occurring nucleoside that exhibits diverse and potent physiological responses in the central nervous system and the cardiovascular, renal, pulmonary, and immune systems. The actions of adenosine are mediated through G-protein-coupled receptors classified into four subtypes, A 1 , A 2A , A 2B , and A 3 , on the basis of their affinity order profiles for agonists and antagonists (1)(2)(3). The A 3 adenosine receptor (A3AR) 1 is the most recent subtype identified, since it had remained pharmacologically obscure until its gene was cloned in the early 1990s. The A3AR has been cloned from multiple species including rat (4,5), human (6,7), sheep (8), canine (9), and rabbit (10).
The physiological importance of the A 3 adenosine receptor has not been established. A3AR is expressed in a broad spectrum of tissues. The most abundant expression in human is found in lung and aorta (7). The evaluation of the functional role of the A3AR has been hampered by the presence of multiple adenosine receptor subtypes in target tissues and lack of selective A3AR agonists and antagonists. In addition, the unique species difference between rat and human receptors in their ability to bind xanthine antagonists and differences in tissue distribution of expression has contributed to the difficulty in selecting an appropriate model to study the function of the A3AR (3). Despite these obstacles, a number of studies have been reported in which the A3AR has been linked to a variety of physiological processes including degranulation of antigen-stimulated rat basophilic leukemic cell line, RBL-2H3 (11), mast cell-dependent constriction of hamster cheek pouch arterioles (12), mast cell mediation of hypotension in rats (13,14), induction of apoptotic cell death (15), inhibition of TNF␣ production from LPS-stimulated murine (16) and human (17) macrophage-like cell lines, inhibition of platelet-activating factor-induced chemotaxis of human eosinophils (18), neural (19) and cardiac (20,21) protection, and depression of locomoter activity in mice (22).

EXPERIMENTAL PROCEDURES
Generation of A3R Ϫ/Ϫ Mice-The mouse genomic DNA utilized to construct the targeting vector was obtained by polymerase chain reaction screening (5Ј-GCTGCCATCGGGCTCTGTG and 5Ј-GCAGGCATA-GAAGTGCATCT) of a P1 129 Mouse ES Library (Genome Systems, Inc.). The targeting vector was introduced into the embryonic stem (ES) line E14Tg2a (23) by electroporation, and G418-resistant colonies were identified by established protocols (24). Two A3AR-targeted ES cell clones were injected into 3.5-day-old blastocysts from C57BL/6 mice and reimplanted into pseudopregnant foster mothers. Resultant progeny were screened initially by coat color chimerism as the indicator of ES cell contribution. Chimeras were bred to B6D2 and C57BL/6 mice and heterozygote mice determined by genomic Southern blot of tail biopsy DNA. Heterozygote mice were intercrossed to produce homozygote A3AR Ϫ/Ϫ animals and wild type (A3AR ϩ/ϩ ) littermate controls.
Southern Blot Analysis-Genomic DNA was isolated from ES cell clones and tail biopsies by modifications of standard procedures (25) and analyzed by genomic Southern blot for targeted recombination. Ten micrograms of genomic DNA was digested with the restriction enzyme BamHI, electrophoresed on a 0.7% agarose gel, transferred to Hybond-N, and UV cross-linked according to the manufacturer's suggestion (Amersham Pharmacia Biotech). Filters containing the immobilized DNA were incubated with 2 ϫ 10 6 cpm/ml of radiolabeled probe in RapidHyb (Amersham Pharmacia Biotech) solution for 3 h at 65°C. The A3 targeting probe corresponds to a 0.75-kb XhoI-HindIII fragment located immediately upstream of the short arm of the targeting construct. Filters were washed at 65°C in a solution containing 0.1ϫ SSC and 0.1% SDS and analyzed by autoradiography.
Purification of Bone Marrow-derived Mast Cell (BMMCs)-BMMCs were isolated from the femurs of 8 -12-week-old A3AR Ϫ/Ϫ and A3AR ϩ/ϩ mice according to established procedures (26). BMMCs were cultured in the presence of interleukin-3 and passed weekly to eliminate adherent cells as described (26). Experiments were carried out within 4 -8 weeks of culture initiation.
Northern Blot Analysis-Total brain RNA was isolated from A3AR Ϫ/Ϫ and A3AR ϩ/ϩ mice using the Trizol Reagent (Life Technologies, Inc.). Poly(A) ϩ mRNA was isolated by subjecting total brain RNA to two cycles of oligo(dT)-cellulose chromatography. Total RNA was isolated from A3AR Ϫ/Ϫ , A3AR ϩ/Ϫ , and A3AR ϩ/ϩ cultured BMMCs using RNAzol (Tel-Test, Inc.) according to the manufacturer's recommendations. Fifteen micrograms of total BMMC RNA or 5 g of poly(A) ϩ brain RNA was fractionated on a 1% agarose/formaldehyde gel (27), transferred to Hybond-N, and hybridized in RapidHyb solution containing 2 ϫ 10 6 cpm/ml radiolabeled probe. Mouse A 1 , A 2A , and A 2B cDNA probes were a kind gift from Dr. Diana Marquardt (University of California, San Diego). The mouse A 3 probe corresponds to a 431-base pair SacI fragment containing part of the 2nd extracellular loop and stopping 18 amino acids upstream of the stop codon. Filters were washed at 65°C in a solution containing 0.1ϫ SSC and 0.1% SDS and analyzed by autoradiography.
Hemodynamic Measurements-Mice were anesthetized with ketamine, xylazine mixture (3.8 mg/ml ketamine, 0.46 mg/ml xylazine at 0.1 ml per 15 g mouse body weight) and placed on a circulating water heating pad. A micro-ranethane catheter, filled with heparinized saline, was inserted into the right carotid artery and held in place with suture and veterinary adhesive. The catheter was sealed and then tunneled subcutaneously and exteriorized between the scapulae. The catheter was protected by a small plastic tube glued to the mouse back. After the mouse recovered from the anesthesia, it recovered overnight with food and water, ad libitum. All surgeries were done in accordance with AICUC standards. On day of the experiment, the arterial catheter was connected by a pressure transducer to a data acquisition system (MI 2 , Malvern, PA). After a quieting down period, control heart rate and blood pressure measurements were made every 15 min for a period of an hour. The final reading is reported as the basal heart rate and mean arterial pressure value. Means (Ϯ S.E.) were done for each group (n ϭ 16).
Hexosaminidase Release Assay-BMMCs were incubated overnight at 37°C with a monoclonal IgE directed against human dinitrophenyl albumin (Sigma) at a concentration of 100 ng/ml/10 6 cells to give 100% occupancy of IgE receptors. After removal of unbound IgE by washing twice in a glucose/saline, Pipes-buffered solution (119 mM NaCl, 5 mM KCL, 25 mM Pipes, 5.6 mM glucose, 1 mM CaCl 2 , 0.4 mM MgCl 2 , 0.1% bovine serum albumin), 5 ϫ 10 5 cells were transferred to 96-well microtiter plates. Cells were incubated in the presence and absence of adenosine (10 nM to 100 M) or the adenosine analogue 2-Cl-IB-MECA (2-chloro-N 6 -(3-iodobenzyl)-adenosine-5Ј-N-methyl-carboxamide) (0.1 nM to 1 M) for 1 min prior to stimulation with antigen (dinitrophenyl albumin, 10 -100 ng/ml) at 37°C. Reactions were terminated after 20 min by centrifugation at 2000 ϫ g for 5 min at 4°C. The extent of mast cell degranulation was determined by comparing levels of hexosaminidase activity in the supernatant and cell pellets. Hexosaminidase activity was determined by incubating supernatant and cell lysate with 1 mM p-nitrophenyl-N-acetyl-B-D-glucosaminidine (M r 342.3, Sigma) for 1 h at 37°C. The reaction was terminated by adding 0.1 M Na 2 CO 3 / NaHCO 3 , and the absorbance was measured at 410 nm. Hexosaminidase release was expressed as a percentage of the total amount of hexosaminidase present in the cells as described previously (29).
Measurement of TNF␣ Production-Mice had access to food and water ad libitum. At zero time, mice were injected intravenously with 0.2 ml of saline vehicle with or without test compound, followed immediately by intraperitoneal injection of 0.5 ml of saline containing 10 g of LPS (lipopolysaccharide B from Escherichia coli 0111:B4, Difco) and 800 mg/kg D-galactosamine. Mice were sacrificed by CO 2 asphyxiation 90 min later, and heparinized blood was obtained by cardiac puncture. Plasma was prepared by centrifugation at 1000 ϫ g for 15 min, and plasma TNF␣ levels were determined by enzyme-linked immunosorbent assay as described in Molineaux et al. (30). TNF␣ levels are expressed as the mean Ϯ S.E. Results were compared statistically using the Student's t test, and differences were considered to be significant at or below p ϭ 0.05.
Materials-Adenosine was purchased from Roche Molecular Biochemicals. CGS21680 was purchased from RBI (Natick, MA).

RESULTS
Cloning of the Mouse A3AR Gene and Generation of Targeting Plasmid-The mouse genomic DNA encoding the A 3 adenosine receptor was cloned by polymerase chain reaction screening from a P1 129 mouse ES library. The mouse A3AR coding sequence lies on two exons separated by an approximately 2.1-kb intron interrupting the second intracellular loop. The deduced amino acid sequence encodes a 319-amino acid protein and exhibits 89% overall identity to the rat (4, 5), 76% to the sheep (8), and 73% to the human (7) A3ARs. A targeting construct designed to disrupt the A 3 adenosine receptor gene was generated containing 1.5 kb of mouse genomic DNA immediately upstream of the A3AR coding sequence, a neomycin resistance gene replacing the initiation codon through the third transmembrane domain and part of the intron, and 7.5 kb of genomic DNA downstream containing the remaining intron and the remainder of the A3AR coding sequence (Fig. 1A). This DNA construct was electroporated into embryonic stem cells (line E14Tg2a), and transfectants were selected with G418.
Targeted Disruption of the Mouse A3AR Gene-Homologous recombination with the endogenous A3AR gene results in disruption of the A3AR gene by replacement of the 5Ј-half of the receptor with a neomycin resistance gene. Southern blot analysis was used to screen BamHI digested genomic DNA prepared from 60 G418-resistant ES cell clones for targeted re-combination. Four ES cell clones were positive for targeted recombination as determined by a decrease in the size of the BamHI fragment from 5.5 kb in wild type to 3.1 kb in the mutant allele (Fig. 1B). Two of the targeted clones were injected into 3.5-day-old blastocytes and produced highly chimeric mice. Male chimeras were bred to B6D2 and C57BL/6 mice, and transmission of the mutant allele was determined by coat color and genomic Southern blot of DNA obtained from tail biopsies.
Animals heterozygous for the mutant allele were intercrossed, and the progeny of these matings yielded wild type mice, heterozygotes, and mice homozygous for the targeted allele with the expected Mendelian frequency (Fig. 1B). No difference was observed between the A3AR-deficient and wild type littermates by visual inspection or routine histological analysis of all organs. In addition, despite high levels of the A3AR expression in the testes (4), both male and female A3AR Ϫ/Ϫ mice demonstrated normal fertility. Body mass was identical between A3AR ϩ/ϩ and A3AR Ϫ/Ϫ mice. The cardiovascular effects produced by adenosine, coronary vasodilation, hypotension, and bradycardia, are well documented (32). A3AR transcripts are expressed in heart; however, the involvement of the A3AR in maintaining basal blood pressure or heart rate has not been explored. The mean arterial blood pressure and resting heart rate was evaluated in A3AR ϩ/ϩ and A3AR Ϫ/Ϫ mice and found to be virtually identical. Values measured for basal mean arterial pressure were 116.71 Ϯ 4.61 mm Hg for A3AR ϩ/ϩ and 115.87 Ϯ 5.03 mm Hg for A3AR Ϫ/Ϫ mice and for basal heart rate, 537.59 Ϯ 26.06 beats/min for A3AR ϩ/ϩ and 578.20 Ϯ 21.71 beats/min for A3AR Ϫ/Ϫ mice.
mRNA Expression of Adenosine Receptor Subtypes in A3AR ϩ/ϩ and A3AR Ϫ/Ϫ Mice-It has been previously shown that A 2A and A 2B adenosine receptor subtype mRNAs, but not A 1 adenosine receptor subtype mRNAs, are expressed in murine BMMCs (33). A3 transcript expression was not determined in this initial analysis of BMMCs. In cultured BMMCs isolated from A3AR ϩ/ϩ mice, high levels of A3AR transcripts were detected ( Fig. 2A). In comparison, no A3AR mRNA was detected in cultured BMMCs isolated from A3AR Ϫ/Ϫ mice. To determine whether the expression of A 1 , A 2A , and A 2B adenosine receptor subtypes were affected by the elimination of A3AR expression, mRNA levels were determined by Northern blot analysis. No significant change in levels of either A 2A or A 2B adenosine receptor transcript was seen in cultured BMMCs from wild type and A3AR Ϫ/Ϫ mice (Fig. 3). Consistent with previous reports (33), no expression of A 1 receptor transcript was detectable in cultured BMMC mRNA from wild type and A3AR Ϫ/Ϫ mice (data not shown). In order to evaluate A 1 receptor expression in A3AR ϩ/ϩ and A3AR Ϫ/Ϫ mice, RNA was isolated from whole brain and analyzed by Northern blots. Similar to A 2A and A 2B expression in BMMCs, no differences were observed between A3AR ϩ/ϩ and A3AR Ϫ/Ϫ mice (Fig. 3).
Binding Analysis of Tissues from A3AR ϩ/ϩ and A3AR Ϫ/Ϫ Mice-The lack of functional A3AR protein expression was confirmed by binding assays on BMMC membranes utilizing the A3AR agonist [ 125 I]ABA (Fig. 2B). BMMC membranes were utilized for the binding analysis as they express high levels of A3AR. Membranes prepared from A3AR ϩ/ϩ mice exhibited specific [ 125 I]ABA binding. Heterozygote mice exhibited a 50% decrease in specific [ 125 I]ABA binding. In comparison, no specific [ 125 I]ABA binding was measured on membranes from A3AR Ϫ/Ϫ mice (Fig. 2B). The density of A 1 and A 2A adenosine receptors expressed in whole brain from A3AR ϩ/ϩ and A3AR Ϫ/Ϫ mice was measured by radioligand binding of [   Mice-Adenosine has been shown to potentiate mediator release murine from BMMCs after stimulation with antigen or calcium ionophore (33). The A 2A adenosine receptor agonist, CGS21680, failed to enhance mediator release from BMMCs, and it was postulated that the A 2B receptor and/or another receptor subtype modulated the secretory process. To determine which adenosine receptor subtype mediates the response in BMMCs, degranulation experiments were carried out in BMMCs from A3AR Ϫ/Ϫ and A3AR ϩ/ϩ mice. The ability of a selective A3AR agonist, 2-chloro-N 6 -(3-iodobenzyl)-adenosine-5Ј-N-methyl-carboxamide (2-Cl-IB-MECA), to potentiate mast cell degranulation was evaluated on antigen-stimulated BM-MCs isolated from A3AR Ϫ/Ϫ and A3AR ϩ/ϩ mice. BMMCs were preincubated with varying concentrations of 2-Cl-IB-MECA (0.1 nM to 1 M) prior to antigen challenge. BMMCs from wild type mice demonstrated a dose-dependent increase in hexosaminidase release (81.5 Ϯ 11.5% at 1 M, p ϭ 0.05), whereas BMMCs from A3AR Ϫ/Ϫ mice showed no change over antigen alone (Fig. 4). In comparison BMMCs from A3AR Ϫ/Ϫ mice exhibited no difference to wild type mice in the potentiation of antigen-induced degranulation by prostanoid prostaglandin E 2 (115 Ϯ 35 and 113 Ϯ 28%, respectively). To determine whether activation of the A 2A or A 2B adenosine receptor expressed on BMMCs could also potentiate mast cell degranulation, BMMCs from A3AR ϩ/ϩ and A3AR Ϫ/Ϫ mice were preincubated with varying concentrations of adenosine prior to antigen challenge. BMMCs from wild type mice displayed a dose-dependent potentiation of hexosaminidase release with adenosine preincubation, whereas A3AR Ϫ/Ϫ mice showed no potentiation over antigen alone (Fig. 5). Mean percent hexosaminidase release with antigen alone in A3AR ϩ/ϩ and A3AR Ϫ/Ϫ was 27 Ϯ 3 and 27 Ϯ 5%, respectively. Preincubation with 1 ϫ 10 Ϫ4 M adenosine prior to antigen challenge resulted in an increase in hexosaminidase release by 67.3 Ϯ 11% (p ϭ 0.028) in wild type mice, whereas A3AR Ϫ/Ϫ mice showed no change (28 Ϯ 5%) over antigen alone. These results support the key role of A3AR activation in enhancing mast cell mediator release.

DISCUSSION
In this study, mice deficient in the expression of A3AR were generated. The A3AR Ϫ/Ϫ mice were determined to be healthy and fertile and exhibited no differences in the development of any organ systems. In addition, the expression of the other adenosine receptor subtypes, A 1 , A 2A , and A 2B , was not affected by the ablation of the A3AR gene. The physiological consequence of the loss of A3AR expression was evaluated by measuring the response of antigen-stimulated BMMC degranulation to adenosine and inhibition of LPS-stimulated TNF␣ production in vivo in A3AR Ϫ/Ϫ and A3AR ϩ/ϩ mice. Our results support the essential role of the A3AR as a mediator in these processes as both functions were found to be eliminated in mice lacking A3AR. The ablation of adenosine or 2-Cl-IB-MECA potentiation of antigen-stimulated BMMC degranulation and decrease in 2-Cl-IB-MECA inhibition of LPS-stimulated TNF␣ production are the first phenotypic differences to be reported for the A3AR Ϫ/Ϫ mice.
Adenosine has been reported to potentiate mediator release from multiple types of mast cells including human lung mast cells (35), rat mast cells (36), murine BMMCs (33), and the rat basophilic leukemic cell line, RBL-2H3 (11,37), after stimulation with antigen or calcium ionophore. Adenosine has been implicated to contribute to the pathophysiology of asthma as inhaled adenosine provokes bronchoconstriction in asthmatic but not normal patients (38). The response to adenosine is believed to be predominantly mediated through mast cell activation (39). Attempts to elucidate definitively the adenosine receptor subtype that mediates mast cell activation has been pharmacologically difficult due to lack of agonists and antagonists that display sufficient selectivity to discriminate between multiple adenosine receptor subtypes present on these cells. Recent studies employing the rat mast cell-derived basophilic leukemia cell line, RBL-2H3, a model for mast cells, have suggested that the A3AR is the adenosine receptor subtype that mediates mast cell degranulation.
The relationship between A3AR activation and mast cell degranulation was first postulated by the detection of A3AR on RBL-2H3 cells, using the A3AR agonist 125 I-labeled aminophenylethyladenosine N 6 -2-(4-aminophenyl) (APNEA) (11). It was further shown that enhancement of antigen-induced secretion from RBL-2H3 cells correlated with an affinity order profile for adenosine receptor agonists similar to that established for the cloned rat A3AR (11). Additional support for the A3AR role in enhancement of antigen-induced degranulation of RBL-2H3 cells came from the observation that dexamethasone increased the amount of 125 I-APNEA binding, possibly due to an increase in A3AR expression, and the A3AR-dependent release of hexosaminidase (37).
A correlation between A3AR activation and nonimmunologically dependent mast cell mediator release has recently been demonstrated in vivo. The hypotensive response to A3AR activation by APNEA in the anesthetized rat is qualitatively similar to the mast cell degranulation agent compound 48/80. Additionally, the hypotensive response to APNEA was suppressed by mast cell degranulation inhibitors, and depletion of mast cell mediators with compound 48/80 decreased the effectiveness of APNEA-induced hypotension (14). Van Schaick et al. (40) have shown that intravenous administration of the A 3 agonist, 2-Cl-IB-MECA, in conscious rats resulted in a hypotensive response coincident with an increase in plasma histamine concentrations indicating the involvement of the A3AR in mediator release from mast cells in vivo.
In this report, we demonstrate high levels of A3AR expression on murine BMMCs. This is the first report of A3AR expression on native mast cells in contrast to established cell lines such as RBL-2H3 cells. In BMMCs from A3AR Ϫ/Ϫ mice, the potentiation of antigen-dependent degranulation by either adenosine or the A3AR agonist, 2-Cl-IB-MECA, was found to be eliminated. The observation of the absence of adenosine or 2-Cl-IB-MECA potentiation of BMMC degranulation in A3AR Ϫ/Ϫ mice supports the conclusions drawn from previous studies with RBL-2H3 cells and in rats in vivo in which the A3AR was implicated in mediating mast cell degranulation.
Previous reports have implicated the A 2B adenosine receptor to mediate BMMC degranulation through demonstration of an inability of pertussis toxin to affect adenosine-induced calcium mobilization (41). In other species, specifically, canine and human, the A 2B receptor has also been reported to mediate mast cell degranulation (42)(43)(44). This discrepancy in identifying which adenosine receptor subtype mediates mast cell degranulation may be due to a number of variables including differences in the levels of adenosine receptor subtype expression between cells selected for evaluation, the tissue source for mast cells, species differences or whether mast cells are immunologically primed prior to degranulation measurements. For exam- ple, RBL-2H3 cells and mouse BMMCs express high levels of A3ARs as measured by both transcript expression and radiolabeled binding. BMMCs are most similar phenotypically to the rat RBL-2H3 tumor cell line in which the A3AR has been reported to predominate the potentiation of mast cell mediator release. In contrast, dog BR mastocytoma cells express low levels of A3ARs. Degranulation of non-immunologically stimulated BR mastocytoma cells in response to adenosine receptor agonists was found to exhibit an affinity order profile most consistent with the A 2B adenosine receptor (42). In the human mast cell line, HMC-1, A 2B receptor activation has been shown to couple to mobilization of intracellular calcium and interleukin-8 secretion (43,44). The adenosine receptor subtype that mediates these responses has not been defined on HMC-1 cells and warrants further study with selective agonists, such as 2Cl-IB-MECA, to determine if the A3AR is involved as we have shown in the mouse. Recently, it has been reported that calcium mobilization in HMC-1 cells is stimulated with a potency order of 5Ј-N-ethylcarboxamidoadenosine Ͼ N 6 -(2-iodo)benzyl-5Ј-N-methylcarboxamidoadenosine (IB-MECA) consistent with activation of A 2B receptors (45). It is important to note that native human mast cells have not been evaluated for A3ARmediated responses.
In this report, we also evaluated the potential role of A3AR activation in the inhibition of LPS-stimulated TNF␣ production through a comparison of the responses of A3AR ϩ/ϩ and A3AR Ϫ/Ϫ mice. A number of published reports have linked A3AR activation with inhibition of LPS-stimulated TNF␣ production in vitro from cultured human monocytes and mouse and human macrophage-like cell lines. The possible involvement of A3AR was first postulated when adenosine receptor agonists were shown to inhibit TNF␣ production in endotoxinstimulated human monocytes with a pharmacological profile uncharacteristic of either A 1 or A 2A adenosine receptor subtypes (46). In J774.1 cells, a murine macrophage-like cell line, LPS-induced TNF␣ secretion and TNF␣ gene expression were inhibited by adenosine receptor agonists with an affinity order profile similar to the cloned rat A 3 receptor (16). In human U937 cells, micromolar concentrations of the A 3 agonist, IB-MECA, decreased LPS-stimulated TNF␣ production, and it was concluded that A3AR activation mediated this response (17). Recently, inhibition of LPS-stimulated TNF␣ production by high concentrations (1 M) IB-MECA was confirmed on human monocytes; however, the inhibition could be completely blocked with an A 2A adenosine receptor-selective antagonist indicating that IB-MECA was acting via A 2A and not A 3 receptor activation (47). The availability of the A3AR Ϫ/Ϫ mouse provided the opportunity to evaluate specific receptor activation in vivo. Our results suggest that activation of either A 3 or A 2A adenosine receptors can mediate inhibition of LPS-stimulated TNF␣ production in vivo in mice. These results are consistent with published reports obtained with murine cell lines where A3AR activation has been implicated to mediate the inhibition of LPS-stimulated TNF␣ production (16) and with in vivo studies in BALB/c mice where A 2A receptor activation was demonstrated to mediate inhibition of TNF␣ production (34).
In summary, our characterization of the A3AR Ϫ/Ϫ mouse has shown that A3AR activation can produce both pro-inflammatory and anti-inflammatory effects. Specifically, we have demonstrated that the A3AR is the adenosine receptor subtype mediating potentiation of antigen-induced degranulation in murine BMMCs. We have also shown that A 3 and A 2A receptor activation inhibits LPS-stimulated increases in TNF␣ in vivo. Further characterization of the A3AR Ϫ/Ϫ mouse may provide important information about the functional role of the A3AR in more complex inflammatory responses.