INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

ADENOSINE PRODUCES POTENT vasodilation in a majority of mammalian vascular beds (12, 24, 26, 38). The cardiovascular effects of adenosine are mediated by activation of four known cell surface receptors (adenosine A₁, A_2A, A_2B, and A₃), and are dictated by receptor subtype distribution, agonist affinity, and efficiency of cell signaling mechanisms (6, 8, 25, 33). It is now established that vasodilatory effects of adenosine and its analogs are mediated by A₂ receptors in different species, including bovine, canine, porcine, rat, guinea pig, and humans (1, 9, 16, 21, 27, 34). However, the relative contribution of A₂ adenosine receptor subtypes (A_2A and A_2B) in modulating coronary and peripheral vascular responses is not fully understood.

Because adenosine receptors are widely distributed throughout the body, understanding differences in regional vascular responses mediated by each receptor subtype will be necessary for the development of future adenosinergic therapies. Characterization of adenosine receptor-mediated vascular responses and their specific subcellular mechanisms will be best achieved by combining a traditional pharmacological approach with newly available gene-modified animal models. Indeed, a murine model with targeted deletion of A_2A adenosine receptor has been developed and is known to be hypertensive, but the effects of this deletion on the coronary circulation are not yet known (15). Before investigating effects of adenosine in this and other yet-to-be-developed gene-modified models, it is necessary to examine adenosine-induced effects on the coronary and aortic vasculature in mice, because there is limited information in this species (11). The isolated perfused mouse heart is a well-established model to examine pharmacological and physiological responses in vitro (37). Moreover, although resistance vessels are more directly related to the regulation of systemic blood pressure, aortic preparations are routinely used to investigate the functional effects of vasodilator substances and their receptor-mediated mechanisms of action in vitro (5, 7, 32). Therefore, parallel experiments were performed in isolated hearts and aortic rings to assess adenosine-induced responses in these physiologically distinct vascular beds.

A selective A_2A adenosine receptor agonist 2-p-(2-carboxyethyl)phenethylamino-5'-N-ethylcarboxamidoadenosine (CGS-21680) and a selective antagonist 5-amino-7(-phenylethyl)-2-(8-furyl)pyrazolo-[4,3-e]-1,2,4-triazolo[1,5-c]pyrimidine (SCH-58261) were used to directly examine A_2A receptor-mediated vascular effects in this study. CGS-21680 is the most potent and selective A_2A adenosine receptor agonist known, causing coronary vasodilation at low nanomolar concentrations (39) and virtually no effect at A_2B receptors with inhibitor constant (K_i) values at A_2A and A_2B receptors of 15 nM and >100 µM, respectively (6, 8). SCH-58261 is a potent and highly selective competitive antagonist at adenosine A_2A receptors both in vivo and in vitro at nanomolar concentrations, and it has little or no affinity (up to the micromolar range) for adenosine A_2B and A₃ receptors (30, 31, 41). A_2B receptor-mediated responses have been demonstrated by using a nonselective agonist 5'-(N-ethyl-carboxamide)-adenosine (NECA) in concert with a putative A_2B receptor antagonist benzo[g]pteridine-2,4(1H,3H)-dione (alloxazine) (17, 36). NECA has affinity for both A_2A and A_2B receptors with K_i values of 3 to 20 nM and 0.5 to 5 µM, respectively (6, 8). Likewise, alloxazine is the only nonxanthine antagonist reported to have ninefold greater selectivity for A_2B receptors than A_2A receptors (4, 6).

The purpose of this study was, therefore, twofold: 1) to characterize the responses to adenosine and its analogs in the coronary and aortic vasculature, and 2) to determine which adenosine receptor subtype(s) are involved in these effects using different antagonists. We hypothesized that the vasodilatory response to adenosine would differ in the coronary and aortic vasculature in mice.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

All of the experimental protocols were performed according to the guidelines of the Animal Care and Use Committee at East Carolina University.

Langendorff-perfused heart preparation. Hearts were isolated from male Balb/c mice (Charles River) of 10 to 12 wk of age as previously described (10, 37). Briefly, mice were deeply anesthetized with pentobarbital sodium (100 mg/kg ip), a thoracotomy was performed, and hearts were quickly excised en bloc with the mediastinum and placed in heparinized ice-cold buffer to arrest cardiac contraction. After lungs, trachea, and extraneous connective tissues were removed, the aorta was carefully tied to an aortic cannula made from a 20-gauge blunted needle. Hearts were retrogradely perfused at a constant pressure of 80 mmHg with warmed Krebs-Henseleit buffer in standard Langendorff fashion and allowed to beat spontaneously. The composition of the modified Krebs-Henseleit buffer was composed of (in mM) 118 NaCl, 4.7 KCl, 1.2 MgSO₄, 1.2 KH₂PO₄, 25 NaHCO₃, 2.5 CaCl₂, 0.5 Na₂EDTA, 11 glucose, and 2.0 pyruvate. The buffer was prefiltered to particle size of <0.22 µm and bubbled continuously with 95% O₂-5% CO₂ at 37°C (pH 7.4). The left ventricle was vented with a small polyethylene apical drain, and a water-filled balloon made of plastic wrap was inserted into the left ventricle across the mitral valve through a left atriotomy. The balloon was connected to a fluid-filled pressure transducer by polyethylene tubing for continuous measurement of left ventricular developed pressure (LVDP). Hearts were then immersed in perfusate maintained at 37°C and the ventricular balloon was inflated to yield a left ventricular end-diastolic pressure of 2 to 5 mmHg. Coronary flow was continuously measured using an ultrasonic flow probe (T106, Transonic Systems; Ithaca, NY) placed in the aortic perfusion line, and aortic pressure was recorded via a pressure transducer attached to the sidearm of an aortic cannula. All of the transducers and an ultrasonic flowmeter were coupled to a PowerLab/4sp data acquisition system (ADInstruments; Castle Hill, Australia) and functional data were recorded on an Apple G4 Power Mac computer using PowerLab Chart 3.5.6 software (ADInstruments). Baseline coronary flow, LVDP, and heart rate (derived from the ventricular pressure tracing) were monitored for an initial 30-min equilibration period. Hearts with persistent arrhythmias or poor LVDP (<50 mmHg) during equilibration were excluded from the study.

Protocol for isolated heart experiments. When hearts reached a steady-state coronary flow, increasing concentrations of adenosine and its analogs were infused by a Harvard infusion pump (Harvard Apparatus) into the aortic cannula immediately above the heart at a rate of 1% of the basal flow to achieve the desired concentration in the perfusate. All agonist concentration-response curves (CRCs) were constructed noncumulatively and one CRC was performed on each heart. The concentration of agonist was tested in steps of 0.5-log units. Infusion of each agonist concentration was maintained until coronary flow demonstrated a new steady state (typically within 5 min), and a washout period of at least 5 min was allowed before the administration of the next (higher) concentration. It is reported (14) that successive addition of agonist in the same heart did not have any blunted response on repetition in the rat, and we also did not observe any blunted response on repetition of agonist concentration in the same heart (data not shown). Therefore, the influence of an antagonist was investigated on the same heart in a paired manner for only one agonist concentration. When responses to antagonists are examined, the effect of agonist was first determined in the absence of the antagonist (control). After complete washout of control response (when coronary flow returned to baseline value), the antagonist was infused into the perfusion line and allowed to equilibrate for at least 10 min before adding the same dose of the agonist in the perfusion. The antagonist remained present during agonist administration until steady-state response was achieved. Changes in coronary flow, heart rate, and LVDP were expressed as percent change from predrug baseline value.

Preparation of isolated aortic rings. Male Balb/c mice of 8 to 10 wk old were used in this study. The preparation of isolated mouse aorta was similar to that described by Pomerleau et al. (32). Under deep anesthesia with pentobarbital sodium (100 mg/kg ip), a thoracotomy was performed and the thoracic aorta was gently removed. After removal of fat and connective tissues, the aorta was cut transversely into 2 to 3 rings of 3 to 4 mm in length. The rings were mounted vertically between two stainless steel wire hooks with extreme care not to damage the endothelium, and suspended in 10-ml organ baths containing Krebs-Henseleit solution continuously gassed with 95% O₂-5% CO₂ (37°C, pH 7.4). Aortic rings were allowed to equilibrate for 90 min with an initial resting tension of 1 g (the length-tension relationship determined separately), and the bathing solution was changed at 15-min intervals. Composition of the Krebs-Henseleit solution was (in mM) 118 NaCl, 4.8 KCl, 1.2 MgSO₄, 1.2 KH₂PO₄, 25 NaHCO₃, 2.5 CaCl₂, and 11 glucose. Changes in isometric tension were measured with a fixed-range precision force transducer (model TSD 125C, BIOPAC Systems) connected to a differential amplifier (model DA 100B, BIOPAC Systems). Data were recorded on a Dell computer using MP 100 WSW BIOPAC Systems digital acquisition system and analyzed using Acqknowledge 3.2 Software (BIOPAC Systems).

Protocol for aortic relaxation experiments. After equilibration, the responsiveness and stability of each individual ring was checked by the successive administration of a submaximally effective concentration of l-phenylephrine hydrochloride (phenylephrine) (1 µM). The integrity of the vascular endothelium was assessed pharmacologically by acetylcholine-induced relaxation of phenylephrine-precontracted rings. Tissues that did not elicit reproducible and stable contraction with phenylephrine, and did not relax >50% to 10 µM acetylcholine were discarded from the study. Preparations were then washed several times with drug-free Krebs-Henseleit solution, and allowed to relax fully for 30 min before the experimental protocol began. To determine the vasodilator responses to adenosine receptor agonists, the aortic rings (with baseline resting tension of 1 g) were precontracted with phenylephrine. The CRCs for aortic relaxation by adenosine and its analogs were obtained by cumulative addition of agonist in the organ bath of rings precontracted with 1 µM phenylephrine. The concentration of agonist in the organ bath was increased in steps of either 0.5- or 1-log units. In all cases, agonists were added to yield the next higher concentration only when the response to the lower dose reached a steady state. One CRC was constructed for each ring. To examine the effect of an antagonist, it was added 15 min before contraction of the tissue with phenylephrine, and was present throughout the experiments. Antagonist experiments were performed in parallel using two rings from the same aorta, one serving as control (without antagonist) and one serving as treated (with antagonist). The vasodilator (relaxant) responses were expressed as percent decrease of phenylephrine-induced precontraction where contraction produced by 1 µM phenylephrine in each ring from its initial resting tension (1 g) was considered as 100%. In experiments with 2-chloradenosine (CAD), the phenylephrine precontracted rings relaxed >100% by higher concentration of CAD, and this was due to a decrease of phenylephrine-induced contraction in excess to the initial resting tension of 1 g. In the current study, an increase in the coronary flow was considered as vasodilation to reflect qualitatively the vasorelaxation of aortic rings. The terms vasodilation and vasorelaxation have been interchangeably used throughout the manuscript.

Data analysis. Experimental values are presented as means ± SE. For each CRC to adenosine and its analogs, the concentration required to produce a 50% response (EC₅₀) in coronary flow was obtained by graphic analysis of individual curve. Significant differences were estimated by Student's t-test for paired data from the same experiment and unpaired data from different experiments. The difference between two CRCs was estimated by two-way ANOVA for repeated measures. A P value of <0.05 was considered significant.

Chemicals. Adenosine, CAD, phenylephrine, and acetylcholine were purchased from Sigma (St. Louis, MO). NECA, CGS-21680, and alloxazine were purchased from RBI (Natick, MA). All other chemicals were of the highest grade available and were purchased from Sigma. NECA, CGS-21680, and adenosine antagonists were dissolved in 100% DMSO as 10 mM stock solution, followed by serial dilutions in 50% DMSO and distilled water. All other chemicals were dissolved in distilled water.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Baseline functional parameters in isolated mouse hearts. Baseline functional parameters for male Balb/c mice were recorded at the end of the 30-min equilibration period for each heart before beginning the experimental protocol. The coronary flow (normalized to the wet weight of each heart), heart rate, and LVDP at equilibrium were 9.42 ± 0.23 ml · min¹ · g¹, 353 ± 3.5 beats/min, and 87.5 ± 3.1 mmHg, respectively (n = 61).

Vascular effects of adenosine and its analogs on isolated hearts and aortic rings. Adenosine and its analogs CAD, NECA, and CGS-21680, applied noncumulatively, produced concentration-dependent increases in the coronary flow (vasodilation) in isolated mouse hearts perfused at constant pressure (Fig. 1A). CGS-21680 and NECA displayed greater potency for coronary vasodilation than CAD and adenosine. The percent increase in coronary flow caused by 100 nM concentration of adenosine, CAD, NECA, and CGS-21680 was 128.98 ± 4.87, 185.55 ± 4.46, 398.92 ± 21.19, and 441.74 ± 21.98%, respectively. Comparing the EC₅₀ values for coronary vasodilation (Table 1), the relative order of agonist potency was CGS-21680 = NECA CAD  adenosine.

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Concentration-dependent vascular effects of adenosine, 2-chloradenosine (CAD), 5'-(N-ethyl-carboxamido)- adenosine (NECA), and CGS-21680 in isolated perfused mouse heart and mouse aorta. Vehicle control experiments for DMSO were included only in aorta. Each symbol with a vertical bar represents the mean ± SE of 6-12 experiments. Concentration-response curves (CRCs) for coronary flow (A) and aortic relaxation (B) were constructed noncumulatively and cumulatively, respectively. y-axis, Changes in the coronary flow, expressed as percent change from immediate control value, which was considered as 100%, and the aortic relaxation as a percent decrease of l-phenylephrine hydrochloride, (phenylephrine)-induced contraction; x-axis, molar concentration of agonists on a logarithmic scale.

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Comparison of the vascular effects of adenosine, CAD, and NECA in isolated perfused mouse heart and mouse aorta. To facilitate the comparison of vascular responses between heart and aorta, data in Fig. 1 were normalized (maximal response for each agonist was considered as 100%) and plotted for NECA (A), CAD (B), and adenosine (C). Each symbol with a vertical bar represents the mean ± SE of 6-12 experiments. A: CRCs for NECA in the heart and aorta; B: CRCs for CAD in the heart and aorta; and C: CRCs for adenosine in the heart and aorta. y-axis, Responses are expressed as percentage of maximum response; x-axis, molar concentration of agonists on a logarithmic scale.

Influence of A_2A adenosine receptor blockade on adenosine-induced vascular responses in isolated mouse hearts and aortic rings. The reproducibility of responses induced by CGS-21680 (100 nM) and NECA (100 nM) on the coronary flow was assessed in a subset of three hearts for each agonist (data not illustrated). Each heart was exposed to the same agonist concentration three times (each time for 5 min), separated by at least 10 min of agonist-free perfusion. The increases in coronary flow in three 5-min CGS-21680 (100 nM) infusions were 435.44 ± 8.40, 429.21 ± 17.99, and 421.70 ± 17.99% of baseline, respectively. NECA (100 nM) also produced qualitatively similar and reproducible effects on coronary flow with repeated exposure, and the percent increase in coronary flow for three 5- min NECA infusions was 396.86 ± 9.61, 407.97 ± 23.74, and 421.54 ± 12.03% of baseline, respectively. Thus it is evident that adenosine-induced coronary vascular responses in mouse hearts do not exhibit tachyphylaxis with successive dosing as has been reported (14) in isolated rat hearts.

View larger version (13K): [in this window] [in a new window]   Fig. 3.   Influence of selective blockade of adenosine A2A receptor with SCH-58261 on CGS-21680 (A and B) and NECA (C) induced increases in the coronary flow. Control responses to CGS-21680 (10 nM and 100 nM) and NECA (100 nM) were first determined in the absence of SCH-58261. After a 10- to 12-min washout, hearts were pretreated with SCH-58261 (100 nM) for 10 min before a second exposure to CGS-21680 and NECA. Bars represent the means ± SE of 4-6 experiments from different hearts for each concentration of agonist. y-axis, Increases in the coronary flow are expressed as percent change from baseline or washout value that was assigned as 100%; x-axis, molar concentration of agonists. *P < 0.05 vs. respective control value.

View larger version (15K): [in this window] [in a new window]   Fig. 4.   Influence of selective blockade of adenosine A2A receptor with SCH-58261 on NECA- and CAD-induced relaxation of mouse thoracic aorta precontracted with 1 µM phenylephrine. Each symbol with a vertical bar represents the mean ± SE of 7-12 experiments. A: CRCs for NECA, control, and SCH-58261 (100 nM). B: CRCs for CAD, control, and SCH-58261 (100 nM). For other details, see Fig. 1.

Influence of A_2B adenosine receptor blockade on adenosine-induced vascular responses in isolated mouse hearts and aortic rings. To examine whether adenosine A_2B receptors are involved in regional vascular responses, we investigated the influence of a relatively selective A_2B receptor antagonist, alloxazine (4), on adenosine-mediated coronary flow and aortic relaxation using a similar protocol. Alloxazine is the only nonxanthine antagonist reported to have ninefold greater selectivity for A_2B receptors than A_2A receptors (4, 6) and has been successfully used to identify a role for A_2B receptors in chick heart cells and rat pial artery (17, 36). We tested the concentration-dependent effects of alloxazine on baseline coronary flow and on adenosine-induced increases in coronary flow. Alloxazine had direct effect on coronary flow in isolated mouse hearts. At 1 µM, alloxazine decreased the coronary flow to 89.44 ± 2.61% of baseline (n = 5, P < 0.05 vs. baseline), whereas at 10 µM, it increased the coronary flow to 117.92 ± 5.78% of baseline (n = 5, P < 0.05 vs. baseline). Presence of 1 µM alloxazine did not influence NECA-induced increases in the coronary flow (Fig. 5A). However, in the presence of 10 µM alloxazine coronary flow was significantly attenuated to 334.23 ± 23.98% from 415.91 ± 28.38% of predrug baseline value (P < 0.05 vs. respective control, Fig. 5B).

View larger version (23K): [in this window] [in a new window]   Fig. 5.   Influence of alloxazine, a putative A2B receptor blocker, on NECA-induced increases in the coronary flow. Control responses to NECA (100 nM) were first determined in the absence of alloxazine. After 10- to 12 min washout, hearts were pretreated either with 1 µM alloxazine (A) or 10 µM alloxazine (B) for 10 min before second exposure to NECA. Bars represent the means ± SE of 5 experiments from different hearts. *P < 0.05 vs. respective control value. For other details, see Fig. 3.

View larger version (19K): [in this window] [in a new window]   Fig. 6.   Influence of alloxazine, a putative A2B receptor blocker, on NECA-induced relaxation of mouse thoracic aorta precontracted with 1 µM phenylephrine. Each symbol with a vertical bar represents the mean ± SE of 5-12 experiments. CRCs for NECA, control, and alloxazine (1 µM and 10 µM). For other details see Fig. 1. * P < 0.05 vs. respective control value.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

In the present study of adenosine and its analogs (CAD, NECA, and CGS-21680), in isolated mouse hearts and aortic rings, we observed that: 1) in isolated hearts, adenosine agonists produced concentration-dependent increases in coronary flow with similar maximal response; 2) in aortic preparations, all agonists (except CGS-21680) produced marked concentration-dependent relaxation of phenylephrine-precontracted rings; 3) adenosine and its analogs were ~100-fold more potent in increasing coronary flow than in producing aortic relaxation; and 4) adenosine-induced coronary vasodilation was more susceptible to selective A_2A receptor blockade than was aortic relaxation. To our knowledge, this is the first study to investigate in parallel the vascular responses of adenosine and its analogs in the mouse heart and aorta.

In isolated mouse hearts, the CRCs for adenosine agonists on coronary flow were bell shaped and similar maximal responses were reached with each agonist. In comparison, the CRCs for aortic relaxation by adenosine agonists were steeper and a plateau was not reached with any agonist even at concentrations as high as 1 mM. In isolated hearts, all of the adenosine agonists increased coronary flow at low nanomolar concentrations, whereas micromolar concentrations were necessary to induce aortic relaxation (Fig. 1). This differential sensitivity of adenosine agonists between coronary flow and aortic relaxation suggests the possible involvement of different receptor subtype(s) in coronary and aortic relaxation.

A₂ adenosine receptors mediate vasorelaxation in most vascular beds (2, 3, 6, 8, 19, 27, 28, 33, 34, 39). In isolated hearts, CGS-21680 and NECA were equally potent at producing coronary vasodilation (Fig. 1A) and had lower EC₅₀ values for this response than that of both CAD and adenosine (Table 1). This suggests that murine coronary vasodilation is predominantly mediated by adenosine A_2A adenosine receptors. Our observation that selective A_2A receptor blockade with SCH-58261 equally antagonized the coronary vasodilation caused by CGS-21680 and NECA (Fig. 3, B and C) further supports this conclusion. SCH-58261 is a potent and highly selective competitive antagonist at adenosine A_2A receptors both in vivo and in vitro, and it has little or no affinity up to the micromolar range for adenosine A_2B and A₃ receptors (30, 31, 41). Although a detailed analysis of the relationship between the concentration of CGS-21680 and SCH-58261 was not carried out, it appears that the antagonism was of a competitive nature, because no complete inhibition of CGS-21680-induced coronary flow was noted. Our findings are in agreement with the reported results in guinea pig aorta and porcine coronary vessels where adenosine A_2A receptors are abundant, and SCH-58261 (100 nM) has been shown to competitively antagonize CGS-21680-induced increases in coronary conductance and vasodilation (3, 30). Both adenosine A_2A and A_2B receptors have been reported to be present in human and porcine coronary artery endothelial cells (29). The suggestion that A_2B adenosine receptors may coregulate coronary flow was initially made by Makujina et al. (18), and this has recently been confirmed in human small resistance-like coronary arteries (13). In the present study, coronary flow changes caused by NECA could be partially reduced using the nonselective antagonist alloxazine (Fig. 5B).

Among nonselective adenosine agonists (NECA, adenosine, and CAD), NECA has been used to characterize A_2B receptor-mediated effects in vascular and cardiac preparations (6, 13, 17, 19, 20, 35, 36). Although NECA is a relatively potent A_2B receptor agonist among these nonselective agents (K_i values at A_2B receptors: NECA = 0.5-5 µM; adenosine and CAD = 5-20 µM), it also possesses a 100-fold greater affinity for A_2A receptors than A_2B receptors (8). Similarly, alloxazine is only ninefold more potent at A_2B than A_2A receptors (4, 6). Because these available pharmacological agents are nonselective, it is difficult to quantitate the effect of A_2B receptor activation in the murine coronary circulation. Here it is noteworthy that in isolated hearts from adenosine A_2A receptor knockout mice, CGS-21680 produced no effect on coronary flow and dose-response curves for adenosine-induced coronary vasodilation were significantly shifted to the right with attenuated maximal response compared with wild-type hearts, indicating the involvement of another adenosine receptor subtype in the regulation of coronary flow (23).

The findings in aortic ring preparations were markedly different from those observed in isolated hearts. All nonselective adenosine agonists produced marked relaxation of aortic rings, whereas the selective A_2A receptor agonist CGS-21680 was ineffective (Fig. 1B). Agonist-induced aortic relaxation occurred at a higher concentration (micromolar) than coronary vasodilation (nanomolar). In contrast to coronary vasodilation, selective blockade of adenosine A_2A receptors with SCH-58261 in aortic rings had no influence on agonist-induced relaxation (Fig. 4A), whereas blockade of A_2B receptors with alloxazine significantly inhibited NECA-induced relaxation (Fig. 6). The current observation that CGS-21680 is a substantially weaker aortic vasodilator than NECA in mice is consistent with findings from other models, including guinea pig aorta (19), rat renal artery (20), rat mesenteric arterial bed (35), and human small coronary arteries (13). Moreover, in guinea pig aorta, SCH-58261 failed to antagonize NECA-induced relaxation leading to the conclusion that aortic vasodilation results from adenosine A_2B receptor activation (41). We observe similar findings in mouse aorta where SCH-58261 failed to antagonize NECA- and CAD-induced relaxation. In mouse aorta, alloxazine caused a much greater inhibition of NECA-induced vasorelaxation than SCH-58261, which is consistent with findings in rat pial arteries in which NECA-induced vasodilation was blocked by alloxazine but not by an A_2A receptor antagonist (36).

The availability of a selective A_2A adenosine receptor agonist (CGS-21680) and antagonist (SCH-58261) has allowed direct examination of A_2A receptor-mediated effects (3, 30, 41), but characterizing A_2B receptor-mediated responses is limited by the lack of selective A_2B ligands. A common strategy to circumvent this problem is the indirect determination of A_2B effects using compounds that are potent and selective for other adenosine receptor subtypes (6). Gurden et al. (9) have demonstrated that the relative potencies of CGS-21680 and NECA can be used as a reference to differentiate A_2A- from A_2B-mediated responses. CSG-21680 is virtually ineffective at A_2B receptors but is as potent as NECA in activating A_2A receptors, and both agonists possess an EC₅₀ value in the low nanomolar range (6, 8, 25). That is, when the effects of CGS-21680 are as potent as NECA, it implies that the response is mediated by A_2A receptor activation. Conversely, when the effects of CGS-21680 are less potent than NECA, it implies the response is mediated by A_2B receptor activation. In the present study, both CGS-21680 and NECA demonstrated the same sensitivity and maximal effect on coronary vasodilation (Fig. 1A), and these responses were equally inhibited by A_2A selective antagonism by SCH-58261 (Fig. 3, B and C). Thus adenosine-induced coronary vasodilation in mice is predominantly mediated through A_2A receptor activation. In the murine aorta, CGS-21680 caused much less relaxation than NECA (Fig. 1B) implying that this response is mediated through A_2B receptors. Furthermore, the putative A_2B receptor antagonist alloxazine inhibited NECA-induced relaxation (Fig. 6), whereas the selective A_2A antagonist SCH-58261 had no effect on NECA- and CAD-induced relaxation (Fig. 4, A and B). Thus adenosine-induced aortic relaxation in mice is predominantly mediated through A_2B receptor activation.

In the present study, all of the adenosine agonists (NECA, adenosine, and CAD) had ~100 times greater sensitivity for coronary vasodilation than aortic relaxation (Fig. 2). A similar pattern of relative potencies has been observed in other species where the coronary circulation was more sensitive to adenosine than the peripheral circulation (19, 28). The 100-fold difference in potency between the two vascular beds in the present study (Fig. 2) is mirrored by the known 100-fold greater affinity of adenosine at A_2A receptors than at A_2B receptors (8, 25). Recently, Mattig and Deussen (22) have shown a difference between the two vascular beds where porcine coronary vascular smooth muscle cells have 63% greater adenosine metabolism than aortic endothelial cells, despite their quantitatively similar flux rates under identical experimental conditions. In addition, Zhang et al. (40) have shown a role of extracellular adenosine on bovine coronary artery vasodilation by using adenosine deaminase (ADA). We have performed preliminary experiments examining the effect of ADA on adenosine-induced relaxation in isolated mouse hearts and aortic rings, and observed that ADA markedly attenuated vascular responses in both preparations (data not shown). Although the metabolism differs between the vascular beds, the current observation of a 100-fold difference in sensitivity to exogenous adenosine between the coronary vasculature and aortic rings is more likely related to differential affinity for adenosine receptors, because adenosine itself possesses a higher affinity for A_2A (1-20 nM) than for A_2B (5-20 µM) receptors (8). Shryock et al. (39) recently reported that the high potency of adenosine and its analogs to cause coronary vasodilation in guinea pig heart could be explained by the presence of a large adenosine A_2A receptor reserve in the coronary circulation. Although we did not determine whether an A_2A receptor reserve exists in our preparations, the different sensitivity of adenosine agonists in producing coronary and aortic vasodilation supports the possibility that different A₂ receptor subtypes mediate these effects.

In summary, the present study provides the first functional evidence that adenosine-induced vascular responses in mouse heart and aorta are quite different with respect to concentration dependence and response characteristics. In addition, we have demonstrated compelling evidence for the differential involvement of adenosine receptor subtypes in mouse coronary vasodilation and aortic relaxation. We conclude that murine coronary vasodilation is mediated predominantly by activation of A_2A adenosine receptors, and aortic relaxation is likely predominated by activation of A_2B receptors. A more definitive characterization of adenosine receptors in the regulation of coronary vasodilation awaits studies in A_2A and A_2B receptor knockout mice and/or the use of selective A_2B receptor agents. Understanding the relative contributions to these effects will have important consequences for the development of future adenosinergic therapy.