Adenosine receptor subtypes and vasodilatation in rat skeletal muscle during systemic hypoxia: a role for A₁ receptors

Abstract

1. In anaesthetized rats we tested responses evoked by systemic hypoxia (breathing 8% O₂ for 5 min) and adenosine (i.a. infusion for 5 min) before and after administration of a selective adenosine A₁ receptor antagonist DPCPX (8-cyclopentyl-1,3-dipropylxanthine), or a selective adenosine A_2A receptor antagonist ZM 241385. Arterial blood pressure, (ABP), heart rate (HR), femoral blood flow (FBF) and femoral vascular conductance (FVC: FBF/ABP) were recorded together with the K⁺ concentration in arterial blood ([K⁺]_a) and in venous blood of hindlimb muscle ([K⁺]_v) before and at the 5th minute of hypoxia or agonist infusion.

2. In 12 rats, DPCPX reversed the fall in ABP and HR and the increase in FVC evoked by the selective A₁ agonist CCPA (2-chloro-N⁶-cyclopentyladenosine; i.a. infusion for 5 min). DPCPX also reduced both the increase in FVC induced by hypoxia and that induced by adenosine; the control responses to these stimuli were comparable in magnitude and both were reduced by ∼50%.

3. In 11 rats, ZM 241385 reversed the fall in ABP and increase in FVC evoked by the selective A_2A agonist CGS 21680 (2-p-(2-carboxyethyl)-phenethylamino-5′-N-ethylcarboxamidoadenosine hydrochloride; i.a. infusion for 5 min). ZM 241385 also reduced the increase in FVC induced by adenosine by ∼50%, but had no effect on the increase in FVC induced by hypoxia.

4. In these same studies, before administration of DPCPX, or ZM 241385, hypoxia had no effect on the venous-arterial difference for K⁺ ([K⁺]_v-a), whereas after administration of either antagonist, hypoxia significantly reduced [K⁺]_v-a suggesting an increase in hypoxia-induced K⁺ uptake, or a reduction in K⁺ efflux.

5. These results indicate that both A₁ and A_2A receptors are present in hindlimb muscle and can mediate vasodilatation and that A₁ and A_2A receptors contribute equally to dilatation induced by infused adenosine. However, they suggest that endogenous adenosine released during systemic hypoxia induces dilatation only by acting on A₁ receptors. Given previous evidence that adenosine can stimulate receptors on skeletal muscle fibres that are coupled to ATP-sensitive K⁺ (K_ATP) channels so promoting K⁺ efflux, our results allow the proposal that K_ATP channels may be coupled to both A₁ and to A_2A receptors and may be stimulated to open by adenosine released during hypoxia, but indicate that, during systemic hypoxia, K⁺ efflux caused by either receptor subtype makes a very minor contribution to the muscle vasodilatation.

It is widely accepted that adenosine is released within tissues under hypoxic conditions. Accordingly, several studies in which adenosine receptor antagonists or adenosine deaminase have been used indicate that locally released adenosine makes important contributions to the respiratory and cardiovascular responses that are induced by systemic hypoxia. For example, adenosine has been shown to contribute to the central respiratory depression of systemic hypoxia, by actions within the brain, and to play a part in the bradycardia and myocardial depression by local actions on the sino-atrial node and ventricular muscle (see Neylon & Marshall, 1991; Thomas et al. 1994; and references therein). In these ways the actions of adenosine tend to counteract the hyperventilation and tachycardia that are the reflex, neurally mediated responses to systemic hypoxia (Marshall, 1994). In addition, experiments on the rat have indicated that locally released adenosine is responsible for at least 50 % of the vasodilatation that occurs in skeletal muscle during systemic hypoxia and which overcomes the reflex vasoconstriction evoked by sympathetic nerve activation (Mian & Marshall, 1991; Neylon & Marshall, 1991; Thomas et al. 1994). Such experiments have indicated that adenosine not only acts directly on the blood vessels to induce dilatation (Mian & Marshall, 1991; Skinner & Marshall, 1996), but also that it produces dilatation in a more indirect way. Thus, from measurements of the potassium concentration in the venous efflux of hindlimb muscle and from the effects of an adenosine receptor antagonist and glibenclamide, which is an antagonist of ATP-sensitive K⁺ (K_ATP) channels, we have argued that adenosine acts on receptors on the skeletal muscle fibres that are coupled to K_ATP channels so causing a release of K⁺ which then acts on the vascular smooth muscle to cause vasodilatation (Marshall et al. 1993). The idea that there are K_ATP channels on skeletal muscle fibres that are coupled to adenosine receptors and can be opened by adenosine has been supported by direct electrophysiological evidence (Barrett-Jolly et al. 1996).

Over recent years it has become apparent that adenosine acts by at least three major types of receptor (A₁, A₂ and A₃), the A₂ receptor type being divided into two subtypes (A_2A and A_2B; see Tucker & Linden (1993) for review). Of these, the A₁ and A_2A receptors seem to be of greater functional importance in respiratory and cardiovascular control. To date, the cardioinhibitory effects of adenosine have generally been attributed to an action on A₁ receptors (Evans & Schenden, 1982). However, there is much less certainty about the vasodilator actions of adenosine and specifically about its vasodilator actions during systemic hypoxia. Thus, in the older literature, adenosine-mediated dilatation was generally attributed to stimulation of the A₂ receptor (Olsson & Pearson, 1990). Indeed, Poucher (1996) recently showed that the component of functional hyperaemia in skeletal muscle of the cat that can be ascribed to adenosine is mediated by stimulation of A_2A receptors. However, in experiments on porcine coronary arteries in vitro, adenosine was found to cause dilatation by acting on A₁ receptors (Merkel et al. 1992) and in rabbit hearts in vitro, coronary dilatation induced by hypoxic perfusion was attributed to stimulation of A₁ receptors (Nakhostine & Lamontagne, 1993). Further, in a recent study on the microcirculation of diaphragm muscle of the rat in vivo, exogenous adenosine was shown to induce dilatation via A₁ receptors (Danialou et al. 1997).

Thus, the objectives of the present study were to establish the adenosine receptor subtypes by which adenosine can induce vasodilatation in skeletal muscle of the rat and to establish their functional importance in mediating muscle vasodilatation during systemic hypoxia. To this end we recorded the changes in femoral vascular conductance (FVC) and in other cardiovascular variables brought about by selective agonists of A₁ and A_2A receptors (CCPA (2-chloro-N⁶-cyclopentyladenosine) and CGS 21680 (2-p-(2-carboxyethyl)-phenethylamino-5′-N-ethylcarboxamidoadenosine hydrochloride), respectively), and also tested the effects of the selective A₁ and A_2A receptor antagonists DPCPX (8-cyclopentyl-1,3-dipropylxanthine) and ZM 241385, respectively, upon the responses induced by adenosine and by systemic hypoxia. In view of our previous findings on the role of adenosine in causing K⁺ efflux from skeletal muscle (Marshall et al. 1993), we monitored the effects of the A₁ and A_2A receptor agonists and antagonists upon the concentrations of K⁺ in the arterial blood and venous drainage of hindlimb muscle.

Some of the results of the present study have been described in brief previously (Bryan & Marshall, 1996a, b).

METHODS

Experiments were performed on male Wistar rats using techniques that were very similar to those described previously (Neylon & Marshall, 1991; Marshall et al. 1993). Briefly, anaesthesia was induced with an oxygen-halothane mixture (3.5 % halothane) and maintained by continuous infusion of Saffan (Pitman-Moore Ltd, Uxbridge, Middlesex, UK) delivered at 7–12 mg kg⁻¹ h⁻¹i.v. during surgery and at 5–8 mg kg⁻¹ h⁻¹i.v. during the experimental protocol (see below). During the experimental period, the level of anaesthesia was adjusted so that pinching the paw evoked a sluggish withdrawal reflex and a transient rise in arterial pressure of 10–15 mmHg. The trachea was cannulated with a T-shaped cannula so that the animal could breathe air, or a hypoxic mixture that was blown from a gasbag across the side-arm by an air pump. The left brachial artery was cannulated so that arterial blood pressure (ABP) could be recorded; heart rate (HR) was derived from the pressure signal. All branches of the left iliac and femoral artery and vein that did not supply skeletal muscles were ligated. A cuff-type electromagnetic flow probe connected to a meter was then placed on the left femoral artery so that femoral blood flow (FBF) to the muscles of the hindlimb could be recorded continuously. The right femoral arteries and veins were cannulated such that the cannulae tips lay at the bifurcation of the dorsal aorta and vena cava, respectively. Blood samples (130 μl) were removed via the femoral arterial cannula for analysis of blood gases via a Nova Stat Profile analyser (Stat 3; VA Howe, MA, USA) and of arterial K⁺ concentration ([K⁺]_a), while the femoral vein cannula allowed analysis of the concentration of K⁺ in the blood draining skeletal muscle ([K⁺]_v) For analysis of plasma [K⁺], whole blood (0.1 ml) was centrifuged and 25 μl of plasma was extracted. The remaining blood was immediately returned to the animal, the volume of plasma having been replaced with saline. The plasma samples were stored in a refrigerator and analysed for potassium by flame photometry (CIBA, Corning Diagnostics Ltd), samples from the same experimental group being assayed together. In addition, the ventral tail artery was cannulated retrogradely such that the cannula tip lay close to the bifurcation of the dorsal aorta; it was used for infusion of agonists, preferentially into the hindlimb from which blood flow was being recorded, by means of an infusion pump.

In experimental Group 1, all recorded variables were displayed on an 8-channel pen recorder. In Groups 2 and 3, the variables were recorded on an Apple Power Mac computer (7100/66) by Maclab/8s (ADInstruments, Hastings, West Sussex, UK) at a sampling frequency of 40 Hz. In both cases FVC was computed on-line as FBF/ABP.

All animals were allowed to stabilize for at least 45 min following surgery at the experimental level of anaesthesia (see above) before the experimental protocol began.

Experimental protocols

Group 1

In 12 rats (body weight, 305 ± 4 g; mean ±s.e.m.), recordings were made of responses evoked by a 5 min period of breathing 8 % O₂ in N₂, by a 5 min infusion of adenosine (1.2 mg kg⁻¹ min⁻¹i.a.) into the hindlimb via the tail artery and then by a 5 min infusion of the selective adenosine A₁ receptor agonist CCPA (0.35 μg kg⁻¹ min⁻¹i.a.) also via the tail artery. CCPA is ∼10 000-fold selective for adenosine A₁vs. A₂ receptors (see Tucker & Linden, 1993). At least 10 min was allowed between the periods of hypoxia and adenosine infusion for stabilization of all variables. The dose of adenosine was chosen, on the basis of preliminary experiments, to induce cardiovascular responses of similar magnitude to those induced by breathing 8 % O₂. The dose of CCPA was chosen from preliminary experiments as a dose which would induce a substantial increase in FVC (muscle vasodilatation) without reducing the HR to less than ∼250 beats min⁻¹ (see Results). When the responses induced by CCPA had stabilized, usually within 2 min of the end of the CCPA infusion, the selective A₁ receptor antagonist DPCPX was given (0.1 mg kg⁻¹i.v.) This dose of DPCPX has been shown to be effective and selective for adenosine A₁ receptors (Kellet et al. 1989). DPCPX is reported to be 700-fold selective for A₁vs. A₂ receptors in binding studies and has a high affinity for A₁ receptors in functional studies (pA₂, 8.2; where pA₂ is -log₁₀ of the concentration of the antagonist that produces a 2-fold shift of the dose-response curve; Coates et al. 1994; and references therein). Approximately 10 min after administration of DPCPX, when the baseline values were stable, the three experimental stimuli, breathing 8 % O₂, infusion of adenosine and of CCPA, were repeated as described above. Blood samples were taken for analysis of arterial blood gases, [K⁺]_a and [K⁺]_v during air breathing immediately before the period of hypoxia and in the 5th minute of the hypoxic period both before and after administration of DPCPX. Blood samples for analysis of [K⁺]_a and [K⁺]_v were also taken before and in the 5th minute of CCPA infusion both before and after administration of DPCPX.

Adenosine was dissolved in physiological saline. The vehicle for CCPA (Research Biochemicals Inc.) was 3 % dimethylsulphoxide (DMSO) in saline and that for DPCPX (Research Biochemicals Inc.) was 10 % DMSO-0.1 M NaOH (50/50 v/v) in saline.

Group 2

In 11 rats (308 ± 4 g), the same protocol as for Group 1 was used, but the selective ligands used were the adenosine A_2A receptor agonist CGS 21680 (1.2 μg kg⁻¹ min⁻¹i.a.) and the A_2A receptor antagonist ZM 241385 (0.05 mg kg⁻¹i.v.). CGS 21680 is ∼170-fold selective for A_2Avs. A₁ receptors (see Tucker & Linden, 1993), while ZM 241385 has a high affinity for the A_2A receptor (pA₂, 9.02), a 100-fold lower affinity for the A_2B receptor (pA₂, 7.06) and a 1000-fold lower affinity for the A₁ receptor (pA₂, 5.95) (Poucher et al. 1995; Keddie et al. 1996). The dose of CGS 21680 was chosen from preliminary experiments such that it would induce an increase in FVC similar to that induced by breathing 8 % O₂, while the dose of ZM 241385 was chosen from these experiments as the dose which was able to completely reverse the increase in FVC induced by the test dose of CGS 21680.

The vehicle for CGS 21680 (Research Biochemicals Inc.) was the same as that used for CCPA (3 % DMSO in saline). The vehicle for ZM 241385, which was a kind gift from Zeneca Pharmaceuticals, was 3 % polyethylene glycol 400/0.1 M NaOH (50/50 v/v) in saline.

Group 3

In eight rats (317 ± 7 g) a similar protocol to that used in Groups 1 and 2 was performed to assess the effects of time and the vehicle for CCPA, CGS 21680 and DPCPX. To this end, responses evoked by 8 % O₂, adenosine and the vehicle for CCPA and CGS 21680 were recorded before and after administration of the vehicle for DPCPX. The vehicle for ZM 241385 has already been shown not to affect cardiovascular variables (see Neylon & Marshall, 1991).

At the end of the experiment, the animals were killed with an overdose of Saffan.

Statistical analyses

All results are expressed as means ±s.e.m. To analyse the responses evoked by hypoxia (8 % O₂) before and after administration of the antagonist (or vehicle), the values recorded for each variable before, at 45 s and at 1, 2 and 5 min of the period of hypoxia were compared by two-way ANOVA followed by Scheffé‘s test. Values were recorded at 45 s and at 1 min because it was apparent from previous experiments that some antagonists preferentially affect the early part of the response (Marshall et al. 1993). In practice, the values used for analysis were the mean of those recorded over a 5 s period on either side of the stated time. The responses evoked by adenosine and by CCPA and CGS 21680 were analysed by comparing the differences between the values recorded before and at the 5th minute of infusion, both before and after the antagonist, using Student's paired t test. As both the concentration and infusion rates of the agonists were adjusted in individual animals to achieve the same final dose per body weight, no attempt was made to compare the changes at different time points during the infusion; we simply used the end-point of 5 min. Blood gas values and pH before and after administration of the antagonists were also analysed using Student's paired t test. [K⁺]_a, [K⁺]_v and [K⁺]_v-a (venous-arterial difference) were compared using Wilcoxon signed-rank test since they were not normally distributed. In all cases, P < 0.05 was considered statistically significant.

RESULTS

Effects of the A₁ receptor antagonist DPCPX

In Group 1, systemic hypoxia (breathing 8 % O₂) induced a decrease in ABP and an increase in FVC (P < 0.001 and P < 0.05, respectively) indicating vasodilatation in hindlimb muscle (Fig. 1), as we have described previously (e.g. Skinner & Marshall, 1996). There was no significant change in HR but it tended to increase initially and then fall below the baseline by the 5th minute of hypoxia, as we have shown previously (see Thomas & Marshall, 1994). Meanwhile arterial O₂ pressure (P_a,O2) and arterial CO₂ pressure (P_a,CO2) fell, indicating hyperventilation (Table 1). Infusion of adenosine evoked a similar pattern of response to hypoxia in that ABP fell and FVC increased, but there was also a substantial fall in HR (Fig. 2), as we have described previously (Skinner & Marshall, 1996).

Figure 1. Cardiovascular responses evoked by systemic hypoxia (8 % O₂) for 5 min before and after administration of the A₁ receptor antagonist DPCPX.

Means ±s.e.m. recorded at the times indicated; hypoxia began at time 0. ▾, control values; •, values after administration of DPCPX (0.1 mg kg⁻¹i.v.). Abbreviations: ABP; arterial blood pressure; HR, heart rate; FBF, femoral blood flow; FVC, femoral vascular conductance. ††P < 0.01, †P < 0.05, significant difference between values recorded before and after DPCPX.

Table 1.

Arterial blood gas and arterial pH (pH_a) values recorded during air-breathing and at the 5th minute of hypoxia (8% O₂) before and after administration of the A₁ or A_2A receptor antagonist or vehicle

Experimental group	Pa,O2(mmHg)	Pa,CO2(mmHg)	pHa
Group 1
Before DPCPX
Air	82.5 ± 3	41.5 ± 2	7.39 ± 0.01
8% O2	35.2 ± 1	36.6 ± 2	7.43 ± 0.01
After DPCPX (1 mg kg−1)
Air	88.2 ± 2**	41.0 ± 2	7.39 ± 0.01
8% O2	36.5 ± 1	34.2 ± 2*	7.43 ± 0.01
Group 2
Before ZM 241385
Air	81.8 ± 2	39.2 ± 1	7.40 ± 0.01
8% O2	39.3 ± 3	32.2 ± 2	7.46 ± 0.02
After ZM 241385 (0.05 mg kg−1)
Air	82.3 ± 1	38.9 ± 1	7.39 ± 0.01
8%O2	39.9 ± 3	31.1 ± 2	7.46 ± 0.01
Group 3
Before vehicle
Air	80.5 ± 2	37.7 ± 3	7.42 ± 0.01
8% O2	33.5 ± 2	31.0 ± 2	7.44 ± 0.01
After vehicle
Air	82.9 ± 2	39.6 ± 1 7	42 ± 0.01
8% O2	32.9 ± 2	32.3 ± 1	7.43 ± 0.01

P_a,O2, arterial O₂ pressure; P_a,CO2, arterial CO₂ pressure. Values are means ± S.E.M.

^**P < 0.01

^*P < 0.05, significant difference between values recorded before and after antagonist or vehicle.

Figure 2. Cardiovascular responses evoked by infusion of adenosine (1.2 mg kg⁻¹ min⁻¹i.a.) before and after administration of the A₁ receptor antagonist DPCPX.

Means ±s.e.m. recorded before (0) and at the 5th minute of infusion, as indicated. □, control; , after DPCPX (0.1 mg kg⁻¹i.v.). ***P < 0.001, **P < 0.01, significant change from baseline (0 min). †††P < 0.001, ††P < 0.01, significant difference between change recorded before and after DPCPX. Abbreviations as in Fig. 1.

The responses evoked by infusion of the adenosine A₁ receptor agonist CCPA are shown in Figs 3 and 4. As can be seen, CCPA evoked a long-lasting fall in ABP and increase in FVC, indicating muscle vasodilatation, accompanied by a pronounced fall in HR. As indicated above, this bradycardia limited the dose of CCPA we could infuse in that higher concentrations and infusion rates led to such large falls in HR that ABP fell dramatically. In practice, therefore, we did not induce an increase in FVC with CCPA that was equivalent in magnitude to that induced by hypoxia or adenosine.

Figure 3. Original traces showing cardiovascular changes evoked by infusion of the A₁ receptor agonist CCPA and the effect of the A₁ receptor antagonist DPCPX.

Infusion of CCPA (0.35 μg kg⁻¹ min⁻¹i.a.) began at the time indicated by the arrow and continued for 5 min. DPCPX (0.1 mg kg⁻¹i.v.) was given ~2 min after the end of the CCPA infusion as indicated by the arrow. Abbreviations as in Fig. 1.

Figure 4. Cardiovascular changes evoked by the A₁ receptor agonist CCPA and by subsequent administration of the A₁ receptor antagonist DPCPX.

Means ±s.e.m.□, control values; ▪, values recorded at the 5th minute of CCPA infusion (0.35 μg kg⁻¹ min⁻¹i.a.); , values recorded after DPCPX (0.1 mg kg⁻¹i.v.). ***P < 0.001, **P < 0.01. Abbreviations as in Fig. 1.

The A₁ receptor antagonist DPCPX, given at 0.1 mg kg⁻¹i.v., reversed the effects of CCPA; indeed FVC and FBF fell to levels that were significantly below the original baseline values (Fig. 4). By the time the responses evoked by hypoxia and adenosine were re-tested after administration of DPCPX, the baseline level of FVC was still slightly, but significantly, reduced (P < 0.05; see Figs 1 and 2) and P_a,O2 was significantly increased (Table 1). Under these conditions, the increase in FVC induced by hypoxia was significantly smaller than that before DPCPX (Fig. 1). P_a,O2 fell to a comparable level to that recorded during hypoxia before administration of DPCPX, but P_a,CO2 fell to a level that was lower than that before DPCPX (Table 1). The increase in FVC evoked by adenosine infusion was also significantly smaller after DPCPX; in addition, the fall in ABP was reduced and the bradycardia reversed to a tachycardia (Fig. 2). Infusion of CCPA at the end of the experiment had no detectable effects on the cardiovascular variables (data not shown).

The effects of DPCPX on [K⁺]_a and [K⁺]_v recorded during CCPA infusion and hypoxia are shown in Fig. 5A and B. Infusion of CCPA had no effect on [K⁺]_a or [K⁺]_v, either before or after administration of DPCPX (Fig. 5A). Before DPCPX, hypoxia tended to increase both [K⁺]_a and [K⁺]_v (P= 0.06 and 0.1, respectively); these changes achieved statistical significance after administration of DPCPX. The venous-arterial difference for [K⁺] during air breathing and hypoxia ([K⁺]_v-a), did not change significantly during hypoxia, or infusion of CCPA, before DPCPX (Fig. 5B). However, after administration of DPCPX [K⁺]_v-a became more negative during hypoxia, suggesting that DPCPX induced a reduction in the K⁺ efflux or increase in the K⁺ uptake caused by hypoxia (see Discussion).

Figure 5. Effects of systemic hypoxia (8 % O₂), and infusion of the A₁ and A_2A receptor agonists CCPA and CGS 21680, respectively, on [K⁺]_a, [K⁺]_v and [K⁺]_v-a, before and after administration of the A₁ or A_2A receptor antagonists DPCPX and ZM 241385, respectively.

Means ±s.e.m. CCPA, 0.35 μg kg⁻¹ min⁻¹i.a.; CGS 21680, 1.2 μg kg⁻¹ min⁻¹i.a.; DPCPX, 0.1 mg kg⁻¹i.v.; ZM 241385, 0.05 mg kg⁻¹i.v. In A and C: □, [K⁺]_a; ▪, [K⁺]_v. In B and D, all columns indicate [K⁺]_v-a. Samples were taken before and at the 5th minute of hypoxia or agonist infusion as indicated. **P < 0.01, *P < 0.05, significant difference between [K⁺]_a or [K⁺]_v values recorded at 0 and 5 min. ²²P < 0.01, ²P < 0.05, significant difference between [K⁺]_v-a values at 0 and 5 min.

Effects of the A_2A receptor antagonist ZM 241385

In Group 2, the cardiovascular responses and changes in blood gases evoked by hypoxia were comparable to those seen in Group 1: ABP decreased (P < 0.001) while FVC increased (P < 0.01) (cf. Figs 1 and 6, and Table 1). The cardiovascular responses evoked by adenosine infusion were also comparable to those of Group 1 except that in Group 2, HR did not change significantly (cf. Figs 2 and 7).

Figure 6. Cardiovascular responses evoked by systemic hypoxia (8 % O₂ for 5 min) before and after administration of the the A_2A receptor antagonist ZM 241385.

Means ±s.e.m. recorded at the times indicated; hypoxia began at time 0. ▾, control values; •, values after ZM 241385 (0.05 mg kg⁻¹i.v.). Abbreviations as in Fig. 1.

Figure 7. Cardiovascular responses evoked by infusion of adenosine (1.2 mg kg⁻¹ min⁻¹i.a.) before and after administration of the A_2A receptor antagonist ZM 241385.

Means ±s.e.m. recorded before (0) and at the 5th minute of infusion as indicated. □, control; , after ZM 241385 (0.05 mg kg⁻¹i.v.). ***P < 0.001, **P < 0.01, *P < 0.05, significant change from baseline (0 min). ††P < 0.01, significant difference between change recorded before and after ZM 241385. Abbreviations as in Fig. 1.

Infusion of the A_2A receptor agonist CGS 21680 induced a substantial and long-lasting increase in FVC and fall in ABP comparable to that induced by adenosine, with no change in HR (Fig. 8). When the A_2A antagonist ZM 241385 was given, the changes in FVC and ABP induced by CGS 21680 were reversed, such that ABP and FBF reached levels that were significantly higher than the baseline levels before CGS 21680 (Fig. 8). However, by the time the responses to hypoxia and adenosine were re-tested after administration of ZM 241385, none of the cardiovascular variables nor the blood gas values were different from those recorded before the original test stimuli (see Figs 6 and 7, and Table 1). Under these conditions, the cardiovascular responses induced by hypoxia, including the fall in ABP and increase in FVC (P < 0.001 and P < 0.01, respectively), were comparable to those induced before administration of ZM 241385 and there was no change in the levels to which P_a,O2 and P_a,CO2 fell during hypoxia (Table 1). By contrast, ZM 241385 greatly reduced the increase in FVC that was evoked by adenosine (Fig. 7), but had no effects on the other components of the response to adenosine.

Figure 8. Cardiovascular changes evoked by the A_2A agonist CGS 21680 and by subsequent administration of the A_2A receptor antagonist ZM 241385.

Means ±s.e.m.□, control values; ▪, values recorded at the 5th minute of CGS 21680 infusion (1.2 μg kg min⁻¹i.a.); , values recorded after ZM 241385 (0.05 mg kg⁻¹i.v.). ***P < 0.001, *P < 0.05. Abbreviations as in Fig. 1.

The effects of ZM 241385 on the changes in [K⁺]_a and [K⁺]_v induced by CGS 21680 and hypoxia are shown in Fig. 5C and D. Before administration of ZM 241385, hypoxia increased [K⁺]_a and tended to increase [K⁺]_v (P= 0.09) while infusion of CGS 21680 induced an increase in both [K⁺]_a and [K⁺]_v (Fig. 5C). After ZM 241385, hypoxia still increased both [K⁺]_a and [K⁺]_v, but infusion of CGS 21680 had no significant effect on either [K⁺]_a or [K⁺]_v. Neither hypoxia nor CGS 21680 had a significant effect on [K⁺]_v-a before administration of ZM 241385. However, after ZM 241385 [K⁺]_v-a became more negative during hypoxia (Fig. 5D) indicating that ZM 241385 reduced hypoxia-induced K⁺ efflux and/or increased hypoxia-induced K⁺ uptake (see Discussion).

Control experiments

In Group 3, the cardiovascular responses induced by hypoxia and by adenosine infusion and the changes in P_a,O2, P_a,CO2 and arterial pH induced by hypoxia were fully comparable with those seen in Groups 1 and 2 (Table 1). Infusion of the vehicle for CCPA and CGS 21680 had no effects on any of the variables recorded. Moreover, injection of the vehicle for DPCPX had no effect on the baseline values of the recorded variables, while the responses induced by hypoxia and by adenosine were fully comparable both qualitatively and quantitatively with those recorded before administration of the vehicle. The results of all the analyses performed to test for effects of the vehicles were not significant (P > 0.05); for the sake of brevity, the cardiovascular data are not shown.

DISCUSSION

The main new findings of the present study are firstly, that in hindlimb muscle of the rat, vasodilatation can be induced by infusion of either a selective adenosine A₁, or a selective A_2A, receptor agonist and that these responses can be completely reversed by a selective A₁ or A_2A receptor antagonist, respectively. Secondly, the muscle vasodilatation evoked by infusion of adenosine can be reduced by ∼50 % by either an A₁ or an A_2A receptor antagonist indicating that it is partly mediated by stimulation of A₁ receptors and partly by stimulation of A_2A receptors. Thirdly, the muscle vasodilatation induced by systemic hypoxia, which we know from previous experiments (e.g. Skinner & Marshall, 1996) is partly mediated by adenosine, can be substantially reduced by a selective A₁ receptor antagonist, but not by a selective A_2A receptor antagonist. This indicates that the component of the hypoxia-induced muscle dilatation that is mediated by adenosine is due to its action on A₁ receptors.

Receptors stimulated by exogenous adenosine

In the present study, systemic hypoxia or intra-arterial infusion of adenosine into the hindlimb induced a fall in systemic arterial pressure and substantial vasodilatation in hindlimb muscle, such that FVC increased by at least 100 %, as we have described before (see Skinner & Marshall, 1996). Similarly, intra-arterial infusion of the selective adenosine A₁ receptor agonist CCPA induced a fall in ABP and an increase in FVC. This change in FVC was not as large as that induced by systemic hypoxia or adenosine, but we purposely chose to limit the concentration and infusion rate of CCPA to limit the magnitude of the bradycardia that CCPA also evoked. Since the changes induced by CCPA were long lasting, we were concerned that profound bradycardia and the associated hypotension would lead to a gradual deterioration of the animal and so prevent us from completing the protocol effectively. The fact that CCPA evoked muscle vasodilatation is consistent with the observation that stimulation of A₁ receptors by topical application of an A₁ agonist evoked arteriolar dilatation in muscle of the diaphragm (Danialou et al. 1997).

Since the vehicles for CCPA and the A₁ receptor antagonist DPCPX had no obvious effects on the cardiovascular system, and yet DPCPX given with its vehicle completely reversed the responses induced by CCPA, it appears that the changes induced by CCPA were due to specific stimulation of A₁ receptors within muscle and that the dose of DPCPX we used did produce effective, functional blockade of the A₁ receptors (see Kellet et al. 1989). In fact, as the increase in FVC induced by CCPA was not only reversed by DPCPX, but caused FVC to reach a level that was significantly lower than that before infusion of CCPA, this suggests a tonic dilator influence of adenosine acting via A₁ receptors within the muscle. Since baseline P_a,O2 was raised after application of DPCPX, it appears that adenosine was also exerting a tonic inhibitory influence on respiration. This is consistent with previous evidence that the adenosine receptor antagonist 8-phenyltheophylline (8-PT), which is non-selective between receptor subtypes, caused an increase in baseline tidal volume (V_T) and P_a,O2 (Thomas & Marshall, 1994), and that exogenous adenosine can inhibit respiration by an action on A₁ receptors within the central nervous system (Wessberg et al. 1985).

Intra-arterial infusion of the A_2A receptor agonist CGS 21680 also induced an increase in FVC and a fall in ABP and in this case the magnitude of these changes was fully comparable with those induced by adenosine and by systemic hypoxia. Since neither the vehicle for CGS 21680 nor ZM 241385 had any effects when given alone, and given that the responses induced by CGS 21680 were reversed by ZM 241385, which is a highly potent and selective A_2A receptor antagonist (see Poucher et al. 1995), it seems that the agonist acted specifically via A_2A receptors to produce the muscle vasodilatation and that the dose of ZM 241385 was sufficient to selectively block these receptors. Since FVC returned to a level comparable to the original after administration of ZM 241385, any tonic dilator influence of adenosine within skeletal muscle was not exerted by A_2A receptors. However, the fact that ABP increased to a level that was significantly higher than the original baseline suggests that there may have been a tonic stimulation of A_2A receptors by adenosine in vascular beds other than muscle. As ZM 241385 had no effect on the baseline blood gas values, although stimulation of central A_2A receptors can depress respiration (Wessberg et al. 1985), endogenously released adenosine appears not to have a tonic influence on respiration via these receptors.

Putting these results together, we can conclude that both A₁ and A_2A receptors are present within hindlimb skeletal muscle of the rat and that, when stimulated separately, each receptor subtype induces vasodilatation. Since both DPCPX and ZM 241385 when given separately reduced the muscle dilatation evoked by infusion of adenosine by ∼50 %, it seems likely that the A₁ and A_2A receptor subtypes contributed equally to that response, although further experiments would be required to fully test this proposal. Nevertheless, the present results contrast with previous evidence that adenosine-induced dilatation in skeletal muscle is mediated by A₂ receptors (Olsson & Pearson, 1990). They also contrast with recent observations on the diaphragm muscle of the rat, which showed that arteriolar dilatation evoked by adenosine was totally abolished by the A₁ antagonist DPCPX, and that the A_2A receptor agonist CGS 21680 failed to evoke arteriolar dilatation (Danialou et al. 1997).

Adenosine receptors stimulated during systemic hypoxia

In our previous studies, 8-PT virtually abolished the increase in FVC evoked by infused adenosine, but only reduced the hypoxia-induced increase in FVC by ∼50 % (Marshall et al. 1993). This suggested that ∼50 % of the hypoxia-induced muscle dilatation is mediated by adenosine. Thus, the present finding that DPCPX reduced the hypoxia-induced muscle vasodilatation by ∼50 % strongly suggests that all of the adenosine-mediated component of this response is mediated by A₁ receptors. The observation that ZM 241385 had no effect on the response is fully consistent with this view, while the fact that ZM 241385 reduced the muscle vasodilatation evoked by infused adenosine is not incompatible with it. Since the dilatation evoked by infused adenosine was equal in magnitude to that evoked by hypoxia, whereas adenosine is only responsible for 50 % of the hypoxia-induced dilatation, it must be the case that the adenosine concentration achieved during infusion was substantially higher than that achieved during hypoxia. Thus, as adenosine has an 80-fold greater potency for A₁ than for A_2A receptors (Ueeda et al. 1991), it is likely that adenosine only reached the A_2A receptors in sufficiently high concentrations to induce dilatation during infusion and not during hypoxia. The fact that P_a,CO2 fell to a lower level during hypoxia after administration of DPCPX than before, suggests that adenosine released within the central nervous system reduced the hypoxia-induced increase in ventilation by acting on central A₁ receptors that limit the increase in V_T, (see Wessberg et al. 1985; Thomas & Marshall, 1994; and above).

In previous studies, the adenosine A₁ receptor was implicated in the coronary dilatation that occurred in the rabbit heart in vitro when the perfusing medium was equilibrated with 5 % O₂ rather than 95 % O₂ (Nakhostine & Lamontagne, 1993). However, in hindlimb muscle of the cat in vivo, the adenosine-mediated component of the dilatation induced by muscle contraction was entirely attributed to stimulation of A_2A receptors; the dilatation was reduced by ∼27 % by either 8-PT or ZM 241385 (Poucher, 1996). As far as we are aware, the present study provides the first evidence that A₁ receptors in skeletal muscle are functionally important in mediating a physiological response, namely the dilator response to systemic hypoxia.

The apparent disparity between our results and those of Poucher (1996) might reflect a difference in the relative preponderance of A₁ and A_2A receptors in skeletal muscle of the rat and cat. However, our finding that both A₁ and A_2A receptors are present in rat skeletal muscle and can mediate dilatation and yet only the A₁ receptors are involved in hypoxia-induced dilatation suggests the explanation is not that straightforward. Since the potency of adenosine for A₁ receptors is substantially greater than that for A_2A receptors (Ueeda et al. 1991), it might be that the concentration of adenosine reached in muscle at the level of hypoxia that we induced was not high enough to stimulate the A_2A receptors to evoke dilatation. However, if this were the case, unless there is a species difference, both A₁ and A_2A receptors would have been expected to mediate the dilatation induced in cat muscle by muscle contraction. A more likely possibility is that the A₁ and A_2A receptors are spatially separated within skeletal muscle and that the adenosine released during systemic hypoxia is released at sites that are close enough to A₁ receptors to predominantly stimulate them, whereas adenosine released during muscle contraction is released at sites that are close enough to A_2A receptors to predominantly stimulate them.

In glycolytic muscle fibres, the predominant muscle fibre type of cat hindlimb, 5′-nucleotidase, which generates adenosine from AMP, occurs in much lower concentrations than in oxidative fibres, but is localized to the boundaries of muscle fibres close to the precapillary arterioles and to the endothelium (Rubio et al. 1973). This prompted Poucher (1996) to propose that, during muscle contraction, adenosine is produced in high concentrations locally, at the boundaries of the muscle fibres, and produces dilatation by acting on A_2A receptors on the precapillary arterioles. It could be that these A_2A receptors are on the abluminal surface of the vascular smooth muscle. This would be compatible with the observation that adenosine that is applied topically to skeletal muscle, and therefore to the abluminal surface, produces dilatation in an endothelium- and nitric oxide (NO)-independent manner (Baker & Sutton, 1993).

On the other hand, our evidence from experiments on the rat suggests that, during systemic hypoxia, adenosine is predominantly released from the endothelium rather than from the skeletal muscle fibres. Systemic administration of adenosine deaminase, which does not readily cross the vascular endothelium (Poucher, 1996), reduced the hypoxia-induced muscle vasodilatation to the same extent as 8-PT, which crosses readily (see Marshall et al. 1993; Thomas et al. 1994). Further, the adenosine-mediated component of the muscle dilatation of systemic hypoxia is almost entirely NO dependent (Skinner & Marshall, 1996). Thus, adenosine released during systemic hypoxia apparently acts via the endothelium even though the endothelium represents a functional barrier to the movement of adenosine from interstitial space to plasma (Olsson & Pearson, 1990). It is therefore unlikely that the adenosine is generated by the skeletal muscle fibres (for further discussion see Mian & Marshall, 1991; Skinner & Marshall, 1996). As noted above, 5′-nucleotidase is localized to the endothelial cells of the arterioles as well as to the muscle fibre boundaries (Rubio et al. 1973), suggesting they are capable of generating adenosine from AMP. Moreover, in the heart, adenosine that was pre-loaded into the endothelial cells was shown to be released from them during hypoxic perfusion (Deussen et al. 1986).

There is the possibility that the adenosine that acts on skeletal muscle vasculature during systemic hypoxia reaches it by means of the blood stream, having been released elsewhere. However, this seems intuitively unlikely given that adenosine is rapidly broken down in the blood stream by adenosine deaminase and given that arterioles that are dichotomous branches of the same feeding arteriole may respond in directionally opposite manners to systemic hypoxia even though they show comparable dilator responses to exogenous adenosine (Mian & Marshall, 1991). On balance therefore, we propose that the majority of the adenosine that induces muscle vasodilatation during systemic hypoxia is released from the endothelium of skeletal muscle and that it acts by stimulating A₁ receptors on that endothelium.

Adenosine receptors, potassium and muscle vasodilatation

The proposals made so far do not preclude adenosine also being released from skeletal muscle fibres and making a contribution to the muscle dilatation of systemic hypoxia by acting directly on adenosine receptors on the vascular smooth muscle, although from the different effects of DPCPX and ZM 241385 these would also have to be A₁ receptors. Further, they do not preclude adenosine acting in a more indirect way, by stimulating adenosine receptors on the skeletal muscle fibres that are coupled to K_ATP channels and causing the release of K⁺ which then induces dilatation (see Marshall et al. 1993). In fact, in view of the latter findings, it is relevant to consider whether the adenosine receptors that might mediate such K⁺ efflux are of the A₁ or A_2A subtype.

In our previous study (Marshall et al. 1993), systemic hypoxia caused an increase in both [K⁺]_a and in the concentration of K⁺ in blood leaving the hindlimb, [K⁺]_v, both being measured at the 5th minute of hypoxia; the increase in [K⁺]_v was abolished by 8-PT, or by glibenclamide, the K_ATP channel inhibitor, even though neither drug affected the increase in [K⁺]_a. Although we did not calculate [K⁺]_v-a, in our previous study it tended to be negative both during air breathing and systemic hypoxia, indicating a net uptake of K⁺ by the skeletal muscle. Thus, the effects of 8-PT and glibenclamide on the hypoxia-induced increase in [K⁺]_v were most readily explained if these drugs blocked a stimulus for K⁺ efflux (adenosine receptors that open K_ATP channels) that was counteracting the effect of a stimulus for K⁺ uptake by muscle. The latter can be attributed to the ability of circulating adrenaline to stimulate the Na⁺-K⁺ pump on skeletal muscle fibres by acting on β-adrenoreceptors (Mian et al. 1990). We additionally concluded that systemic hypoxia induces K⁺ efflux from tissues other than muscle, which is independent of adenosine and K_ATP channels and which explains the persistent increase in [K⁺]_a (Marshall et al. 1993).

In the present study, hypoxia tended to increase [K⁺]_a and [K⁺]_v in Group 1 (P= 0.06 and 0.1, respectively) and tended to increase [K⁺]_v in Group 2 (P= 0.09); only the increase in [K⁺]_a in Group 2 reached statistical significance. We have no explanation for why these trends were weaker in the present than in our previous study (Marshall et al. 1993). Nevertheless, after application of DPCPX or ZM 241385, both [K⁺]_a and [K⁺]_v increased during hypoxia. Thus, neither the A₁ antagonist DPCPX, nor the A_2A antagonist ZM 241385, when given separately, prevented [K⁺]_v from increasing at the 5th minute of systemic hypoxia. Since small effects of either antagonist on K⁺ efflux from muscle might have been masked by a hypoxia-induced increase in [K⁺]_a and which by itself might have led to an increase in [K⁺]_v, we took the step of calculating [K⁺]_v-a; such calculations should reflect more accurately events specific to skeletal muscle than in our previous study, because we had taken additional precautions to isolate the venous drainage of hindlimb muscles from the other tissues of the hindquarters (cf. Marshall et al. 1993). In fact, the balance between K⁺ uptake and K⁺ efflux was apparently different in Groups 1 and 2 in that [K⁺]_v-a tended to be positive in Group 1, but negative in Group 2, both in normoxia and hypoxia; again, we do not know why this difference arose. Nevertheless, the difference between [K⁺]_v-a in normoxia and hypoxia became statistically significant both after administration of DPCPX (Group 1) and ZM 241385 (Group 2). Since the change in each case was in the direction of [K⁺]_v-a becoming more negative during hypoxia after administration of the antagonist, i.e. in the direction of greater K⁺ uptake, this is consistent with the idea that both A₁ and A_2A receptors on the skeletal muscle fibres contributed to K⁺ efflux by opening K_ATP channels. There is already electrophysiological evidence that stimulation of A₁ receptors on rat skeletal muscle fibres increases the open probability of K_ATP channels (Barrett-Jolly et al. 1996); to our knowledge, the effect of an A_2A receptor agonist has not been tested.

If both A₁ and A_2A receptor stimulation can open skeletal muscle K_ATP channels, then both CCPA and CGS 21680 might have been expected to cause net K⁺ efflux. In fact, at the 5th minute of CCPA infusion there was no change in [K⁺]_a, [K⁺]_v or [K⁺]_v-a either before or after administration of DPCPX. Infusion of CGS 21680 did cause an increase in [K⁺]_a and [K⁺]_v, which was abolished by ZM 241385, but [K⁺]_v-a across skeletal muscle was not significantly changed by CGS 21680, either before or after ZM 241385. This suggests that the A_2A agonist may have stimulated K⁺ efflux from tissues other than skeletal muscle, but allows no conclusion for skeletal muscle at least at the 5th minute of agonist infusion.

Thus, on balance, our findings on the effects of A₁ and A_2A receptor agonists and antagonists on K⁺ handling are equivocal. They suggest that any effect adenosine may have on K⁺ efflux in muscle, at the 5th minute of systemic hypoxia, by acting on either the adenosine A₁ or A_2A receptor subtypes, makes a minor contribution to the observed increase in FVC given that both DPCPX or ZM 241385 when given alone affected [K⁺]_v-a, but only DPCPX reduced the hypoxia-induced increase in FVC. It may be that A₁ and A_2A receptors must both be stimulated or blocked together for the change in K⁺ efflux to significantly affect FVC. However, we would have to acknowledge that instantaneous measurements of [K⁺] in arterial or venous blood may not be the best way of following hypoxia-induced changes in K⁺ handling; continuous measurements of [K⁺] in interstitial space with K⁺ electrodes might be preferable.

In summary, the present study has provided the first evidence that both A₁ and A_2A adenosine receptors are present in hindlimb skeletal muscle of the rat and can mediate muscle vasodilatation, but indicates that the adenosine-mediated component of the vasodilatation that occurs in muscle during systemic hypoxia is mediated by adenosine A₁ receptors. We propose that such A₁ receptors are on the endothelium and are stimulated by adenosine that is released by the endothelium. The cellular mechanisms by which A₁ and A_2A receptors in skeletal muscle induce vasodilatation are explored in the accompanying paper (Bryan & Marshall, 1999).