Functional characterization of coronary vascular adenosine receptors in the mouse

Abstract

1. Coronary responses to adenosine agonists were assessed in perfused mouse and rat hearts. The roles of nitric oxide (NO) and ATP-dependent K⁺ channels (K_ATP) were studied in the mouse.

2. Resting coronary resistance was lower in mouse vs rat, as was minimal resistance (2.2±0.1 vs 3.8±0.2 mmHg ml⁻¹ min⁻¹ g⁻¹). Peak hyperaemic flow after 20–60 s occlusion was greater in mouse.

3. Adenosine agonists induced coronary dilation in mouse, with pEC₅₀s of 9.4±0.1 for 2-[p-(2-carboxyethyl)phenethylamino]-5′-N-ethyl carboxamidoadenosine (CGS21680, A_2A-selective agonist), 9.3±0.1 for 5′-N-ethylcarboxamidoadenosine (NECA, A₁/A₂ agonist), 8.4±0.1 for 2-chloroadenosine (A₁/A₂ agonist), 7.7±0.1 for N⁶-(R)-(phenylisopropyl)adenosine (R-PIA, A₁/A_2B selective), and 6.8±0.2 for adenosine. The potency order (CGS21680=NECA>2-chloroadenosine>R-PIA>adenosine) supports A_2A adenosine receptor-mediated dilation in mouse coronary vessels. 0.2–2 μM of the A_2B-selective antagonist alloxazine failed to alter CGS21680 or 2-chloroadenosine responses.

4. pEC₅₀s in rat were 6.7±0.2 for CGS21680, 7.3±0.1 for NECA, 7.6±0.1 for 2-chloroadenosine, 7.2±0.1 for R-PIA, and 6.2±0.1 for adenosine (2-chloroadenosine>NECA=R-PIA>CGS21680> adenosine), supporting an A_2B adenosine receptor response.

5. NO-synthase antagonism with 50 μM N^G-nitro L-arginine (L-NOARG) increased resistance by ∼25%, and inhibited responses to CGS21680 (pEC₅₀=9.0±0.1), 2-chloroadenosine (pEC₅₀=7.3±0.2) and endothelial-dependent ADP, but not sodium nitroprusside (SNP). K_ATP channel blockade with 5 μM glibenclamide increased resistance by ∼80% and inhibited responses to CGS21680 in control (pEC₅₀=8.3±0.1) and L-NOARG-treated hearts (pEC₅₀=7.3±0.1), and to 2-chloroadenosine in control (pEC₅₀=6.7±0.1) and L-NOARG-treated hearts (pEC₅₀=5.9±0.2).

6. In summary, mouse coronary vessels are more sensitive to adenosine than rat vessels. A_2A adenosine receptors mediate dilation in mouse coronary vessels vs A_2B receptors in rat. Responses in the mouse involve a sensitive NO-dependent response and K_ATP-dependent dilation.

Introduction

Adenosine receptor sub-types mediating coronary vasodilation appear to differ between species, and controversy exists regarding mechanisms contributing to adenosine-mediated dilation in different (and also within) vascular beds. For example, while there is evidence of NO-dependent components to adenosine responses in coronary vessels (Newmann et al., 1988; Leipert et al., 1992; Vials & Burnstock, 1993; Zanzinger & Bassenge, 1993; Abebe et al., 1994; Kuo & Chancellor, 1995; Hein et al., 1999; 2000), this is not universally observed (Sabouni et al., 1990; Kirkeboen et al., 1994; Lewis & Hourani, 1997; Kemp & Cocks, 1999; Lew & Kao, 1999). There is only preliminary data available regarding the receptor sub-type mediating coronary dilation in the mouse (Morrison et al., 2001), and there are no data regarding mechanisms involved. Therefore, the primary goal of the present study was to functionally characterize the adenosine receptor sub-type mediating coronary dilation in mouse heart. To identify the receptor, coronary responses to adenosine (the endogenous and non-selective agonist), CGS 21680 (A_2A selective agonist), NECA (A₁/A₂ agonist), 2-chloroadenosine (A₁/A₂ agonist), and R-PIA (A₁/A_2B agonist) were examined, together with the effects of A_2B-selective antagonism with alloxazine. The potential contributions of NO and K_ATP channel activation to adenosine receptor responses were assessed via competitive inhibition of NO-synthase with L-NOARG and inhibition of K_ATP channels with glibenclamide. Coronary vascular function in mouse was also compared to that in the more thoroughly characterized rat heart.

Methods

Langendorff perfused heart model

Hearts were isolated from mice and rats as described by us in detail previously (Headrick, 1996; Headrick et al., 2000). Specifically, adult male C57/B16 mice (7–12 weeks old, 26.3±0.3 g body weight, 118±8 mg blotted heart weight) or male Wistar rats (12–16 weeks old, 337±15 g body weight, 1.22±0.19 g blotted heart weight) were anaesthetized with 50 mg kg⁻¹ sodium pentobarbitone, a thoracotomy performed and hearts rapidly excised into ice-cold perfusion fluid. The aorta was cannulated and hearts initially perfused in a retrograde fashion at a hydrostatic pressure of 100 mmHg with modified Krebs bicarbonate buffer containing (in mM): NaCl, 120; NaHCO₃, 25; KCl, 4.7; KH₂PO₄, 1.2; CaCl₂, 2.5; Mg₂SO₄ 1.2; glucose, 15; and EDTA, 0.6. Perfusate was equilibrated with 95% O₂, 5% CO₂ at 37°C, giving a pH of 7.4. Perfusate temperature was maintained at 37°C and hearts were constantly bathed in perfusate within a small water jacketed chamber maintained at 37°C. The left ventricle was vented with a polyethylene drain to prevent accumulation of Thebesian drainage. Coronary perfusion was constantly monitored via a cannulating ultrasonic flow-probe (Transonic Systems Inc., Ithaca, NY, U.S.A.) in the aortic perfusion line and aortic perfusion pressure was monitored using a P23XL pressure transducer (Viggo-Spectramed, Oxnard, CA, U.S.A.) connected to a MacLab data acquisition unit (ADInstruments, Castle Hill, Australia). Aortic pressure was held constant at 100 mmHg. After an initial 20 min of stabilization at intrinsic heart rate all hearts were electrically paced via silver electrodes using a Grass stimulator (Grass S9 stimulator, Quincy, MA, U.S.A.). Mouse hearts were paced at a rate of 400 beats min⁻¹ while rat hearts were paced at 300 beats min⁻¹ (0.5 ms square pulses 20% above threshold, typically 1–4 V). Hearts were allowed to stabilize for an additional 10 min before experimentation. Hearts were then switched to constant flow perfusion at a rate giving an aortic pressure of 100 mmHg prior to drug infusion (13.2±0.7 ml min⁻¹ g⁻¹ in mice, 10.6±0.8 ml min⁻¹ g⁻¹ in rats). This permitted analysis of coronary dilation without the complication of changes in coronary flow rate and drug delivery. Flow was delivered by a Gilson MiniPuls 2 peristaltic pump (Gilson Inc., Middleton WI, U.S.A.).

Experimental protocol

Stabilized hearts were treated with a dilatory agonist (2-chloroadenosine, R-PIA, CGS21680, NECA, adenosine, SNP, or ADP) infused in incremental concentrations. Agonists were infused at each concentration for 1–3 min period during which vascular responses stabilized. Only one concentration-response curve was acquired per heart. Changes in aortic pressure were measured, and coronary flow was constantly monitored to verify a constant flow rate.

To examine the effects of A_2B antagonism in mouse hearts, responses to CGS21680 and 2-chloroadenosine were acquired in the presence of 200 nM alloxazine (n=7 for both groups). Responses to 2-chloroadenosine were also acquired in a sub-set of hearts treated with a higher 2 μm concentration of alloxazine (n=4). Alloxazine infusion was initiated 10 min prior to acquisition of agonist concentration-response curves.

For NO-synthase inhibition in mouse hearts, infusion of 50 μM L-NAME (final concentration) was initiated after stabilization. After a further 10 min period concentration-response curves for CGS21680 (n=10), 2-chloroadenosine (n=8), ADP (n=6), or SNP (n=7) were acquired, as described above. To further examine mechanisms of adenosine-mediated dilation, K_ATP channel inhibition was studied in mouse hearts. Infusion of 5 μM glibenclamide was initiated alone or in conjunction with 50 μM L-NAME. After a further 10 min period, concentration-response curves for CGS21680 and 2-chloroadenosine were acquired (n=6 in all groups). In preliminary experiments 5 μM glibenclamide was shown to fully block dilatory responses to 0.1–3.0 μM of the K_ATP opener minoxidil (data not shown).

Reactive hyperaemic responses were assessed in a separate group of mouse (n=7) and rat hearts (n=8) which were perfused at a constant aortic pressure of 100 mmHg after pacing was initiated. After a further 10 min stabilization hearts were subjected to 20 s of total coronary occlusion followed by reperfusion for 15 min, before studying the response to 60 s of coronary occlusion and reperfusion. Flow responses were monitored and the peak hyperaemic response determined (in absolute units and as per cent of baseline resistance), together with per cent flow-debt repayment. Flow-debt repayment was calculated as:

where flow-debt is the product of baseline flow (ml min⁻¹ g⁻¹)×period of occlusion (min), and excess flow during hyperaemia was calculated as the volume of flow during the period of the hyperaemic response minus the baseline flow. The period of the hyperaemic response was defined as the time from onset of reperfusion to the point where flow fell to the pre-occlusion level.

Data analysis

Responses to agonists were compared between treatment groups using a multi-way analysis of variance for repeated measures, followed by Tukeys HSD post-hoc test for multiple comparisons when differences were detected. pEC₅₀ values were obtained from concentration-response data (expressed as absolute units) by fitting the following four-parameter logistic equation to data for individual experiments:

where A is the response at zero dose (i.e. the pre-infusion value), and B is the response at infinite dose. The equation was fit to data using the Statistica program (Statsoft, Tulsa, OK, U.S.A.), and individual pEC₅₀ values derived from each fit. pEC₅₀s were compared between treatment groups by one-way analysis of variance with Tukeys HSD post-hoc test. In all statistical tests P<0.05 was considered indicative of statistical significance. Dilatory responses to agonists were calculated as a per cent of resting resistance, or were scaled (as a per cent) to the maximal dilatory response observed. All values are reported as mean±s.e.mean.

Chemicals

2-Chloroadenosine, adenosine, CGS21680, R-PIA, NECA, ADP, SNP, alloxazine, L-NOARG, glibenclamide, and minoxidil were all purchased from Sigma/RBI (Sigma Chemical, Castle Hill, Australia). All other chemicals used were of analytical grade or better.

Results

Coronary vascular function in mouse vs rat heart

Baseline coronary tone, peak reactive hyperaemic responses, and minimal vascular tone observed during infusion of the relatively non-selective A₂ adenosine receptor agonist 2-chloroadenosine are shown in Table 1. Resting coronary resistance was lower in mouse vs rat hearts (∼7.5 vs 9.5 mmHg ml⁻¹ min⁻¹ g⁻¹). The minimal resistance achieved during dilation with 2-chloroadenosine was also lower in mouse (2.2 mmHg ml⁻¹ min⁻¹ g⁻¹) vs rat heart (3.8 ml⁻¹ min⁻¹ g⁻¹) (Table 1). The peak reactive hyperaemic response and extent of flow repayment following 60 s occlusion was significantly greater in mouse vs rat (Table 1). Similarly, the peak response following a shorter 20 s occlusion was greater in mouse (35.8±2.3 ml⁻¹ min⁻¹ g⁻¹ at ∼5 s of reperfusion) vs rat (24.3±1.2 ml⁻¹ min⁻¹ g⁻¹ at ∼15 s), as was repayment (136±9% in mouse vs 98±6% in rat). Collectively, data show that resting and minimum coronary resistances are lower in mouse. However, the dynamic range over which coronary vessels dilate is similar in both species (∼5.5 mmHg ml⁻¹ min⁻¹ g⁻¹).

Table 1.

Coronary vascular functional parameters in mouse and rat hearts

Effects of adenosine receptor agonists in mouse vs rat heart

All adenosine receptor agonists concentration-dependently dilated the coronary circulation in mouse and rat hearts (Figure 1). Response magnitude, expressed as per cent change in resistance, was similar in both species. CGS21680 was the most potent agonist in mouse, whereas it was one of the less potent in rat. The rank order of potencies was CGS21680=NECA>2-chloroadenosine>R-PIA>adenosine for mouse. Statistical analysis of pEC₅₀s yielded P values of 0.879 for NECA vs CGS21680, and less than 0.001 for 2-chloroadenosine vs NECA and CGS21680, R-PIA vs 2-chloroadenosine, and adenosine vs R-PIA. This rank order of potencies supports A_2A-mediated dilation, and contrasted with that in the rat: 2-chloroadenosine>NECA=R-PIA>CGS21680>adenosine (Table 2). Statistical analysis of pEC₅₀s yielded P values of 0.001 for NECA vs 2-chloroadenosine, 0.835 for R-PIA vs NECA, and less than 0.005 for CGS21680 vs R-PIA and adenosine vs CGS21680. This rank order of potencies supports an A_2B-mediated response in rat. In addition, sensitivity to all adenosine agonists was higher in mouse vs rat. Furthermore, as shown in Table 2 (and apparent in curves depicted in Figure 1), the steepness or slope of concentration-response curves is greater in mouse vs rat. Slope factors for all adenosine agonists were significantly higher in mouse (Table 2).

Figure 1.

Concentration-response curves for adenosine agonist mediated vasodilation in (A) mouse, and (B) rat hearts. Responses to adenosine (n=8 for mouse, n=7 for rat), 2-chloroadenosine (n=7 for mouse, n=7 for rat), R-PIA (n=6 for mouse, n=8 for rat), NECA (n=7 for mouse, n=7 for rat), and CGS 21680 (n=8 for mouse, n=7 for rat) were studied. Responses are shown as per cent of baseline coronary resistance. All values are means±s.e.mean.

Table 2.

Concentration-response data for adenosine agonists in mouse and rat hearts

A_2B-selective antagonist alloxazine, when infused at 200 nM concentration, failed to alter responses to either A_2A-selective CGS21680 (pEC₅₀=9.2±0.1) or non-selective 2-chloroadenosine (pEC₅₀=8.2±0.1) (Figure 2). A 10 fold higher concentration of alloxazine also failed to alter responses to 2-chloroadenosine in mouse heart (Figure 2B).

Figure 2.

Concentration-response curves for (A) CGS21680 and (B) 2-chloroadenosine in mouse hearts in the presence and absence of alloxazine. Responses were obtained for CGS21680 alone (n=8) or in the presence of 200 nM alloxazine (n=7), and for 2-chloroadenosine alone (n=7) or in the presence of 200 nM (n=7) or 2 μm (n=4) alloxazine. Control data for CGS21680 and 2-chloroadenosine are taken from Figure 1. Responses are shown as per cent coronary resistance. All values are means±s.e.mean. *P<0.05 vs values in untreated hearts (P<0.05).

Effects of NO-synthase inhibition in mouse heart

Treatment of mouse hearts with 50 μM L-NOARG significantly increased baseline coronary resistance by ∼25% to 9.5±0.6 mmHg ml⁻¹ min⁻¹ g⁻¹ (P<0.05). Responses to the adenosine receptor agonists CGS21680 and 2-chloroadenosine were significantly reduced by L-NOARG (Figures 3 and 4). Inhibitory effects were present whether responses were expressed relative to resting resistance (Figure 3) or scaled to maximal dilation (data not shown). Effects of L-NOARG were primarily observed at low to moderate agonist concentrations, with no depression of responses to high agonist levels (i.e. above 10⁻⁸ M CGS21680 and 10⁻⁷ M 2-chloroadenosine) (Figure 3). Coronary sensitivities to both CGS21680 and 2-chloroadenosine (reflected by pEC₅₀s) were significantly depressed by L-NOARG treatment, with a greater shift in 2-chloroadenosine sensitivity (13 fold) vs CGS21680 sensitivity (2.5 fold) (Table 2). Representative traces from hearts treated with 2-chloroadenosine and CGS21680±L-NOARG, are shown in Figure 4.

Figure 3.

Concentration-response curves from mouse hearts in the presence and absence of 50 μM L-NOARG. Responses were obtained for (A) CGS21680 (n=10), (B) 2-chloroadenosine (n=8), (C) SNP (n=6 in both groups), and (D) ADP (n=7 in both groups). Control responses for CGS21680 and 2-chloroadenosine are taken from the data shown in Figure 1. Responses are shown as per cent of baseline coronary resistance. All values are means±s.e.mean. * P<0.05 vs values in untreated hearts (P<0.05).

Figure 4.

Representative tracings from mouse hearts treated with (A) CGS21680, (B) CGS21680+50 μM L-NOARG, (C) 2-chloroadenosine, and (D) 2-chloroadenosine+50 μM L-NOARG. The trace shown in (E) is from a heart displaying phasic oscillations in coronary tone during infusion of 50 μm L-NOARG.

L-NOARG failed to reduce responses or alter sensitivity to the endothelial-independent dilator SNP (Figure 3C, Table 2). There were no significant inhibitory effects when responses were expressed either as per cent of resting coronary resistance (Figure 3C), or scaled to the maximal dilation observed (data not shown). Indeed, due to slightly increased resting resistance with L-NOARG, there was an insignificant trend towards slightly enhanced responses to higher levels of nitroprusside (when expressed relative to the initial resting resistance). Dilatory responses to the endothelial-dependent dilator ADP were significantly inhibited by L-NOARG (Figure 3D, Table 2).

Curiously, a significant number of hearts displayed phasic oscillations in coronary tone during treatment with L-NOARG, rendering it difficult to acquire concentration-response data in those hearts. Representative traces from various experiments, including the latter hearts displaying oscillations in tone, are shown in Figure 4.

Effects of K_ATP channel inhibition in mouse heart

Treatment of mouse hearts with 5 μM glibenclamide significantly increased baseline coronary resistance by ∼35% to 9.8±0.5 mmHg ml⁻¹ min⁻¹ g⁻¹ (P<0.05). Treatment with 5 μM glibenclamide+50 μM L-NOARG further increased resting resistance to 12.6±0.7 mmHg ml⁻¹ min⁻¹ g⁻¹ (P<0.05). Responses to CGS21680 and 2-chloroadenosine were significantly inhibited by glibenclamide, and were further inhibited by co-treatment with glibenclamide+L-NOARG (Figure 5). The degree of inhibition observed with glibenclamide was considerably greater than that with L-NOARG alone. The pEC₅₀s for CGS21680 and 2-chloroadenosine were shifted to higher concentrations by an order of magnitude in the presence of glibenclamide (Table 2). The substantial inhibitory effects of glibenclamide were additive to those for L-NOARG. The pEC₅₀s for CGS21680 and 2-chloroadenosine were further shifted by at least an order of magnitude in hearts treated with glibenclamide+L-NOARG together, relative to pEC₅₀s for hearts treated with either glibenclamide or L-NOARG alone (Table 2).

Figure 5.

Concentration-response curves from mouse hearts in the presence and absence of 5 μM glibenclamide or 5 μM glibenclamide+50 μM L-NOARG. Responses were obtained for (A) CGS21680 (n=10), and (B) 2-chloroadenosine (n=8). Control responses and responses in the presence of L-NOARG alone are also shown (from Figure 3). Responses are shown as per cent of baseline coronary resistance. All values are means±s.e.mean. *P<0.05 vs values in untreated hearts (P<0.05). †P<0.05 vs values in L-NOARG treated hearts (P<0.05).

Discussion

Little information exists regarding adenosine receptor sub-types mediating coronary dilation in the mouse, or mechanisms of adenosine-mediated coronary dilation in this species. Our data indicate that mouse heart is more sensitive to adenosine agonists than rat heart, and show that different receptor sub-types mediate dilation in these species. Adenosine-mediated coronary dilation in the mouse involves NO-dependent and K_ATP-dependent components.

Coronary vascular function in mouse vs rat heart

Baseline and maximally dilated coronary resistances were lower in mouse vs rat heart (Table 1). Additionally, reactive hyperaemic responses were higher in the mouse (Table 1). The mouse circulation therefore appears to operate at lower resting and dilated resistances than in the rat, likely reflecting correlations between vascular density, mitochondrial capacity, metabolic rate and body mass (Pietschmann et al., 1982; Rakusan & Tomanek, 1986). Owing to greater metabolic rate in mouse (i.e. mouse heart rate is 550–600 beats min⁻¹ vs ∼350 beats min⁻¹ in the rat), the murine heart requires a much greater O₂ delivery. This can be accomplished by limited adaptations in O₂ carrying capacity of blood and O₂ transport from blood to mitochondria, and/or it can be met by increased coronary vascularization and coronary flow, as shown here. Higher hyperaemic responses and lower coronary resistance in mouse heart permits greater resting and maximal O₂ delivery.

The adenosine receptor mediating coronary dilation in mouse and rat hearts

Endogenous adenosine may play an important role in modulating coronary vascular resistance during and following pathophysiological stimuli (Berne, 1980; Olsson & Pearson, 1990), although its role in regulation of coronary tone under more physiological conditions in vivo is questionable (Tune et al., 2001). We show that murine coronary vessels are highly sensitive to adenosine agonists. Sensitivities to all agonists, including adenosine, are higher in mouse than in rat (Figure 1, Table 2). Importantly, the rank order of agonist potencies differs between mouse and rat (Table 2). As noted initially by Gurden et al. (1993), two differing orders of agonist potencies are observed for A₂ receptor mediated vascular responses. They observed an order of potency of CGS21680⩾NECA>R-PIA in canine coronary vessels, reflecting an A_2A receptor-mediated response, and NECA>R-PIA>CGS21680 in guinea-pig aorta, reflecting an A_2B receptor-mediated response. Kull et al. (1999) documented a functional potency profile of NECA⩾CGS21680>R-PIA⩾2-chloroadenosine>adenosine for rat and human A_2A receptors. We obtained an order of CGS21680=NECA>2-chloroadenosine>R-PIA>adenosine in the mouse which is comparable to that for canine (Gurden et al., 1993), human and rat A_2A receptors (Kull et al., 1999). In contrast we obtained a profile of 2-chloroadenosine > NECA=R-PIA> CGS21680>adenosine in rat, supporting a primarily A_2B response.

To test for A_2B receptors we also infused the A_2B selective antagonist alloxazine (Figure 2). While there is a lack of potent and highly selective A_2B antagonists, alloxazine exhibits ∼10 fold selectivity for murine A_2B receptors vs A_2A receptors (Brackett & Daly, 1994) and completely inhibits A_2B receptors at 2 μM. We chose a lower concentration in order to minimize antagonism at the A_2A receptors which appear to be present. At 200 nM alloxazine failed to alter responses to A_2A selective CGS21680 and non-selective 2-chloroadenosine (Figure 2). Hearts receiving the maximally effective 2 μM concentration also displayed unaltered responses to 2-chloroadenosine (Figure 2B). These data verify A_2A mediated responses to both receptor agonists, and indicate an absence of functional A_2B receptor in mouse coronary vessels.

Previous studies indicate that A_2A receptors mediate coronary dilation in dog (Glover et al., 1996), pig (Lew & Kao, 1999; Hein et al., 1999), and guinea-pig (Belardinelli et al., 1998). Functional identification of A_2A receptors in the pig (Lew & Kao, 1999; Hein et al., 1999) has recently been supported by molecular analysis (Hein et al., 2001). In contrast, A_2B receptors are of primary importance in mediating coronary dilation in humans (Kemp & Cocks, 1999) and in the rat (Lewis & Hourani, 1997; Rose'meyer et al., 1999). A preliminary study by Morrison et al. (2001) indicates that gene deletion of the A_2A receptor eliminates CGS21680-mediated coronary dilation in the mouse. While findings from genetically modified knock-out animals are not directly applicable to wild-type tissues, and although we have only assessed receptor identity in a single strain of mice, and rodents strains can show variability in cardiovascular responses, our data together with the findings of Morrison et al. (2001) support A_2A-mediated coronary dilation in the murine heart.

It is worth noting that the slope of concentration-response curves is greater for all agonists in rat vs mouse (Table 2). This suggests that the efficacy of adenosine agonists at murine A_2A receptors may be greater than at rat A_2B receptors, and/or supports heterogeneity in adenosine responses in rat. Lewis & Hourani (1997) provide evidence that both A_2A and A_2B receptors may mediate dilation in rat heart, and there is support for an unidentified, potentially intracellular receptor in rat vessels (Prentice & Hourani, 1996; Prentice et al., 1997) in addition to mouse aorta (Prentice et al., 2001). This multiplicity of effector mechanisms may contribute to a broader concentration range over which agonists act in rat. The narrow range and steep slope in mouse is consistent with a single coronary receptor sub-type.

In assessing the identity of the adenosine receptor we chose to assess the rank order of agonist potencies, and study effects of a selective antagonist. This conventional pharmacological approach remains a key method for identification of adenosine receptors (Belardinelli et al., 1998; Brackett & Daly, 1994; Gurden et al., 1993; Hein et al., 1999; Kemp & Cocks, 1999; Kull et al., 1999; Lew & Kao, 1999; Prentice et al., 1996; 1997; Rose'meyer et al., 1999), and was adopted despite alternate methodologies, including epigenetic approaches (targeted gene knock-out and anti-sense oligonucleotides) (Nyce, 1999). Despite selectivity of epigenetic techniques, these methods possess limitations. Anti-sense studies in intact organs are largely restricted to brain and lungs due to difficulties in selectively introducing oligonucleotides, and these studies are limited by poor penetration into cells and nuclei. With respect to gene knock-out, an A_2A knockout mouse has been assessed in terms of neurophysiological responses (Ledent et al., 1997), and there are preliminary data regarding adenosine and CGS21680 mediated coronary dilation in this model (Morrison et al., 2001). While these findings support an important A_2A response in murine vessels, interpretation of gene knock-out data are limited since they assess effects of a genes long-term absence rather than the normal role of the gene product itself. Absence of the gene may lead to unknown developmental, morphological and functional changes, together with unpredicted compensatory responses. Receptor identification remains best achieved via complimentary pharmacological and molecular approaches.

Role of NO in adenosine receptor-mediated coronary dilation in mouse

There is evidence that adenosine is a mixed endothelial-dependent/independent dilator, activating both endothelial and smooth muscle receptors within the same vessel (Headrick & Berne, 1990; Moritoki et al., 1990; Rose'meyer & Hope, 1990; Headrick et al., 1992; Vials & Burnstock, 1993; Hein et al., 1999). To test the potential involvement of NO in adenosine responses, we studied effects of NO-synthase inhibition with L-NOARG. L-NOARG significantly increased baseline resistance and inhibited responses to endothelial-dependent ADP but not SNP, verifying selective inhibition of NO-dependent responses, and supporting a role for endogenous NO in control of resting tone in murine coronary vessels (Table 1, Figure 3). Curiously, a number of hearts displayed phasic oscillations in coronary resistance during L-NOARG treatment (Figure 4). Though the mechanism of these oscillations is unclear, it may reflect competition between constriction due to reduced NO release and relaxation due to locally released dilators. Importantly, while L-NOARG did not alter responses to SNP, it significantly attenuated responses to 2-chloroadenosine and CGS 21680, substantially increasing EC₅₀s and threshold concentrations at which agonists induced dilation (Figure 3, Table 2).

While our data agree with studies supporting partial NO- or endothelial-dependent coronary responses to adenosine in guinea-pig (Newmann et al., 1988; Leipert et al., 1992; Vials & Burnstock, 1993), dog (Rubanyi & Vanhoutte, 1985; Zanzinger & Bassenge, 1993), and pig (Abebe et al., 1994; Kuo & Chancellor, 1995; Hein et al., 1999; 2000), they contrast with studies in human and porcine coronary vessels (Sabouni et al., 1990; Kirkebeon et al., 1994; Kemp & Cocks, 1999; Lew & Kao, 1999), and in rat heart (Lewis & Hourani, 1997). Varying observations from different species supports pronounced species differences in the mechanisms of adenosine-mediated coronary dilation. Contradictory observations made within a single species (e.g. Abebe et al., 1994; Kirkebeon et al., 1994; Kuo & Chancellor, 1995; Hein et al., 1999; Lew & Kao, 1999) demonstrates a need for further research. In this respect, both A_2A and A_2B receptors may mediate coronary dilation in the same species and there appear to be differences in transduction mechanisms for these sub-types. There is a greater weight of unequivocal data supporting NO-dependence of adenosine A_2A responses (Newmann et al., 1988; Leipert et al., 1992; Vials & Burnstock, 1993) whereas there is controversy regarding NO-dependence of coronary A_2B responses (Kirkebeon et al., 1994; Kemp & Cocks, 1999; Lewis & Hourani, 1997; Lew & Kao, 1999).

Functional antagonism vs NO-synthase inhibition

There may be two effects of L-NOARG in coronary vessels–enhanced functional antagonism increasing the resistance over which dilatory agonists can induce relaxation, and direct antagonism of NO-synthase dependent responses. Lewis & Hourani (1997) recently concluded that apparent antagonism of adenosine via NO-synthase inhibitors may reflect functional antagonism (i.e. vasoconstriction). They found that L-NOARG reduced adenosine responses in rat, but also reduced responses to NO-synthase independent SNP. However, responses to SNP were quite low in their study and L-NOARG only modestly reduced these in contrast with 2–10 fold differences in adenosine-response magnitude (Lewis & Hourani, 1997). We show that L-NOARG increases resting coronary resistance (Table 1), and this should alter dilatory response amplitude. If ‘pre-constrictors' do not directly inhibit mechanisms of action of a dilator, enhanced functional antagonism and resting tone simply increases the amplitude (but not sensitivity) of dilatory responses, as noted by Lew (1995). Functional antagonism is useful in increasing response amplitude for dilators, removing constraints imposed by degree of pre-constriction, and permitting analysis of agonist efficacies (Broadley & Nicholson, 1979; Lew, 1995). These factors are exemplified by responses to SNP: L-NOARG tends to enhance response amplitude to high levels of SNP when responses are expressed relative to the higher initial tone (Figure 3C). It is necessary to normalize such responses to make meaningful comparisons, and to remove differences which are not relevant (Lew, 1995). These minor differences are absent when SNP responses are scaled to maximal dilation (data not shown). Importantly, in direct contrast to SNP, responses to low levels of adenosine agonists are substantially reduced by L-NOARG while maximal responses are unaltered (Figure 3), the threshold effective concentration of 2-chloroadenosine is increased ∼10 fold (Figure 3), and coronary sensitivity to 2-chloroadenosine and CGS21680 is significantly reduced (Table 2). Selective effects of L-NOARG on A₂-mediated responses (and on ADP responses) supports significant NO-dependence of adenosine-mediated dilation in mouse heart.

Assuming that responses to low levels of 2-chloroadenosine in the presence of L-NOARG can be considered largely NO-synthase independent, we can estimate NO-synthase dependent dilation by subtraction of NO-synthase independent responses from control (mixed) responses (Headrick & Berne, 1990; Headrick et al., 1992). As shown in Figure 6 the NO-synthase dependent L-NOARG sensitive response is important at low levels of A₂ receptor activation. Dilation with 1–30 nM 2-chloroadenosine is almost entirely NO-dependent (Figure 6). At higher levels of activation NO-synthase independent dilation predominates. This is consistent with previous observations indicating that NO-dependent adenosine responses are sensitive yet small in amplitude (Headrick & Berne, 1990; Moritoki et al., 1990; Rose'meyer & Hope, 1990; Headrick et al., 1992; Vials & Burnstock, 1993). While small amplitude might be considered evidence for a minimal functional role (Vials & Burnstock, 1993), high sensitivity (Figure 6) ensures that it is quantitatively important at low, physiologically relevant levels of receptor activation.

Figure 6.

NO-synthase dependent and NO-synthase independent responses to 2-chloroadenosine in perfused mouse heart. The raw data is derived from the results shown in Figure 3. The NO-synthase independent response is represented by the dilatory response in the presence of 50 μM L-NOARG. The NO-synthase dependent response is calculated as the difference between the control response (in the absence of L-NOARG) and the NO-synthase independent response (in the presence of L-NOARG). Responses are scaled to per cent of maximal dilation observed.

Role of K_ATP channels in adenosine receptor-mediated coronary dilation in mouse

In addition to release of NO, activation of K_ATP channels is implicated in A_2A adenosine receptor-mediated coronary dilation (Kuo & Chancellor, 1995; Randall, 1995; Hein et al., 1999; 2001). In contrast, A_2B receptors may mediate coronary dilation in human vessels via NO and K_ATP-independent mechanisms (Kemp & Cocks, 1999). Our data reveal that K_ATP blockade with glibenclamide, at a level which fully antagonizes responses to minoxidil, markedly inhibits responses to CGS21680 and 2-chloroadenosine (Figure 5). Importantly, effects of K_ATP channel inhibition are additive to (and greater than) those for NO-synthase inhibition via L-NOARG. Considering the 120–300 fold reduction in sensitivity to CGS21680 and 2-chloroadenosine with combined NO-synthase and K_ATP inhibition, these two pathways appear to be the primary mechanisms contributing to A_2A-mediated dilation. The more pronounced inhibitory effects of K_ATP channel blockade and additivity with the effects of L-NOARG suggest that A_2A receptors trigger NO-mediated dilation via mechanisms distinct from activation of smooth muscle and/or endothelial K_ATP channels.

Concluding remarks

In conclusion, the present study demonstrates that peak coronary flows, reactive hyperaemia and adenosine sensitivity are all substantially greater in mouse vs rat heart. Adenosine agonist potency profiles and effects of alloxazine support A_2A adenosine receptor-mediated coronary dilation in mouse vs A_2B adenosine receptor-mediated dilation in rat. Our data demonstrate a significant NO-dependent and K_ATP channel-dependent components to adenosine mediated coronary dilation in the mouse. A_2A-mediated dilation appears more strongly dependent on K_ATP channels than NO formation. However, although NO-independent dilation predominates over NO-dependent dilation at moderate to high agonist levels, the high-sensitivity NO-dependent response may play an important role under physiological conditions when adenosine concentrations and the level of A_2A receptor activation are relatively low.

Abbreviations

CGS21680

2-[p-(2-carboxyethyl)phenethylamino]-5′-N-ethyl carboxamidoadenosine

K_ATP

ATP-dependent K⁺ channel

L-NOARG

N^G-nitro L-arginine

NECA

5′-N-ethylcarboxamidoadenosine

NO

nitric oxide

R-PIA

N⁶-(R)-(phenylisopropyl) adenosine

SNP

sodium nitroprusside