Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

The cardioprotective effects of the purine nucleoside adenosine have been well documented. Adenosine, when administered before ischemia, delays the onset of ischemic contracture, improves postischemic ventricular function and reduces infarct size (Lasley et al., 1990, 1995; Lasley and Mentzer, 1992; Randhawa et al., 1995; Thornton et al., 1992). There are numerous reports that infusing adenosine during reperfusion does not improve postischemic recovery of function (Reibel and Rovetto, 1979; Ambrosio et al., 1989; Randhawa et al., 1995; Sekili et al., 1995) and conflicting reports on whether increasing myocardial adenosine levels during reperfusion reduces infarct size in in vivo preparations (Zhao et al., 1993; Vander Heide and Reimer, 1995). These findings indicate that adenosine treatment is most effective as a cardioprotective agent when administered before ischemia. The beneficial effects of adenosine when administered before ischemia appear to be mediated via its effects on the cardiac myocytes. This hypothesis is supported by numerous reports of the cardioprotective effects of adenosine A₁ receptor agonists and the blockade of adenosine's protection with the adenosine A₁ receptor antagonist DPCPX (Lasley and Mentzer, 1992; Thornton et al., 1992; Yao and Gross, 1993; Grover et al., 1996). There are additional reports that adenosine may protect the ischemic heart via the activation of myocyte A₃ receptors (Auchampach et al., 1997). The preischemic administration of adenosine A₂ receptor agonists does not protect the ischemic heart (Lasley and Mentzer, 1992; Thornton et al., 1992; Yao and Gross, 1993).

By virtue of their discrete locations, these receptors are exposed to adenosine concentrations in two different extracellular compartments. Adenosine A₂ receptors, located primarily on vascular and endothelial cells, are in contact with adenosine in the vascular space. Myocyte A₁ and A₃ receptor activations are determined by the concentration of adenosine in the ISF. The adenosine levels in these two compartments can be quite different due to the rapid uptake and metabolism of adenosine by red blood cells and endothelial cells (Nees et al., 1985; Deussen et al., 1986). The results of radiolabeled adenosine studies indicate that exogenous adenosine concentrations 1 µM are incorporated into endothelial cells (Nees et al., 1985). The coronary endothelial metabolic barrier to exogenous and endogenous adenosine has been reported in various isolated heart models during normoxia and hypoxia (Fenton and Dobson, 1987; Heller and Mohrman, 1988; Matherne et al., 1990; Tietjan et al., 1990).

The hypothesis that adenosine protects the ischemic heart via the activation of an adenosine receptor located primarily on the cardiac myocytes requires that adenosine accumulate in the compartment in contact with the A₁ (and A₃ receptor) (i.e., the ISF compartment). We previously reported that treatment of in situ canine myocardium with the adenosine deaminase inhibitor EHNA increases endogenous ISF adenosine levels during regional ischemia and attenuates postischemic stunning (Dorheim et al., 1991). In the same stunning model, intracoronary adenosine infusion increased preischemic ISF adenosine and improved regional function (Randhawa et al., 1995). However, there have been no dose-response studies determining the gradient between vascular and ISF adenosine levels and their associations with the cardioprotective effect of adenosine. Therefore, we tested the hypothesis that exogenous adenosine must be infused at a concentration sufficient to overcome the coronary endothelial barrier and accumulate in the ISF compartment to improve postischemic ventricular function. This hypothesis was tested in the isolated perfused rat heart preparation in which ISF adenosine levels were estimated with the cardiac microdialysis technique.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Isolated perfused heart preparation. Male Wistar rats (300-350 g) were heparinized (500 U i.p.) and anesthetized with sodium pentobarbital (65 mg/kg i.p.). Hearts were then rapidly excised, arrested in ice-cold Krebs-Henseleit buffer and placed on a Langendorff perfusion apparatus. Retrograde aortic perfusion was initiated with Krebs' buffer with the following composition (in mM): NaCl 117, KCl 4.7, MgSO₄ 1.2, KH₂PO₄ 1.2, CaCl₂ 1.25, NaHCO₃ 25.0, glucose 11.0 and sodium pyruvate 1.0. The perfusate was gassed with 95% O₂/5% CO₂ resulting in a pH 7.35 to 7.45, PCO₂ 35 to 40 mm Hg and PO₂ 560 to 620 mm Hg. The perfusion circuit was maintained at 37°C with a water bath circulator, and myocardial temperature was maintained at 37°C by submersing the heart into a water-jacketed chamber filled with Krebs-Henseleit buffer. Hearts were paced at 300 beats/min (4 msec, 5 V) via electrodes placed on the right ventricle. Ventricular function was assessed by measuring LVDP with a fluid-filled latex balloon, connected via a polyethylene catheter to a pressure transducer (model P23XL; Gould, Cleveland, OH). The balloon was inserted into the left ventricle via the left atrium and was inflated to yield an end-diastolic pressure of 5 mm Hg. Once the left ventricular balloon volume was set during the preischemic perfusion, it was maintained constant during ischemia and reperfusion. Coronary flow rate was determined via timed collections of effluent overflow from the heart bath.

Cardiac microdialysis. Myocardial ISF nucleoside levels were estimated with the cardiac microdialysis technique, which has been described in detail previously (Van Wylen et al., 1990, 1992). Briefly, silica tubing (170 µm outer diameter) was inserted into both ends of a microdialysis fiber (200 µm inner diameter) and sealed in place within the dialysis fiber with cyanoacrylic glue, such that a 0.8-cm length of dialysis fiber remained exposed. The dialysis fiber is then inserted into the left ventricular wall and pulled through the myocardium until the dialysis "window" rests entirely within the muscle. The inflow silica tubing was then connected via a larger silica tube to a gas-tight glass syringe filled with Krebs-Henseleit buffer previously gassed with 95% N₂/5% CO₂ so that PO₂, PCO₂ and pH were ~30 mm Hg, ~40 mm Hg and ~7.35, respectively. The low PO₂ value was chosen to minimize the supply of oxygen locally by the dialysis probe. Dialysate samples were collected from the outflow silica tube into small plastic vials, diluted with 100 µl of distilled H₂O and immediately frozen at 80°C until analysis. Once the fiber was implanted, the heart was allowed a 60-min recovery/stabilization period before initiation of the experimental protocol. This time is necessary to allow nucleoside levels, which are initially elevated due to tissue trauma during fiber implantation, to stabilize at a low basal level (Van Wylen et al., 1992). The fibers were perfused with Krebs' buffer at a flow rate of 0.75 µl/min. In vitro studies (n = 6) performed at 37°C indicated that 8-mm fibers perfused at this flow rate yielded a recovery rate of 69 ± 2% for 10 µM adenosine.

Nucleoside measurements. Dialysate and coronary effluent adenosine and inosine levels were determined by reverse-phase HPLC as described previously (Lasley et al., 1995). Peaks were monitored by absorbance at 254 nm, identified by retention time of external standards and quantified by peak area based on computer-generated standard curves. Computer analysis was performed using the Waters Associates MAXIMA software.

Protocols. Control untreated hearts (n = 8) were compared with hearts treated for 10 min immediately before ischemia with either 1, 10 or 100 µM adenosine (n  6 per group). Preliminary studies indicated that dialysate adenosine levels had equilibrated/stabilized 60 min after fiber implantation. Preischemic and pretreatment base-line dialysate samples were then collected for a 15-min period (min 60-75); samples from two hearts were combined before analysis. Pooling of samples was necessary since nucleoside levels in one 15-min collection in a normally perfused heart were below the limits of detection in our HPLC assay. Dialysate samples were also collected during the 10-min adenosine pretreatments. After 5 min of adenosine infusion, pacing was temporarily ceased to record spontaneous heart rate. During the last min of the 10-min preischemic dialysate collection, a coronary effluent sample was obtained and immediately frozen. After collection of dialysate and venous samples and recording of preischemic hemodynamic data, hearts were submitted to 30 min of complete global normothermic ischemia. Five 6-min dialysate samples were collected during ischemia, after which the hearts were reperfused for 30 min. Dialysate samples were not collected during reperfusion because preliminary results and previous studies (Randhawa et al., 1995; Lasley et al., 1995) indicated no effect of adenosine pretreatment on postischemic adenosine levels. At the conclusion of the experiment, the hearts were inspected to ensure that the dialysis fiber was positioned within the muscle wall.

Statistics. Results are expressed as mean ± S.E. Recovery of function and coronary flow are expressed as a percent of preischemic values. Data were analyzed by a one-way analysis of variance (ANOVA) with statistical significance between control and treatment groups determined by Dunnett's test. A value of P < .05 was considered statistically significant.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

The base-line and preischemic hemodynamic data of hearts in the four groups are shown in table 1. All hearts had similar coronary flow rates and LVDPs before any interventions. As shown in table 1, all three doses of adenosine increased coronary flows to similar values, but only the two higher adenosine concentrations (10 and 100 µM) were associated with reduced spontaneous heart rates. Adenosine had no effect on LVDP before ischemia.

The effects of increasing concentrations of exogenous adenosine on coronary venous and dialysate adenosine are shown in figure 1, A and B. Coronary venous adenosine concentration in control hearts was 0.04 ± 0.007 µM. Infusions of 1, 10, and 100 µM adenosine increased effluent adenosine levels to 0.30 ± 0.06, 7.2 ± 0.4 and 74.7 ± 5.3 µM, respectively. Base-line dialysate adenosine concentrations were 0.28 ± 0.06, 0.21 ± 0.07, 0.31 ± 0.05 and 0.30 ± 0.05 0.28 ± 0.06, respectively. Adenosine levels in the 10-min dialysate sample collected during the 1 µM adenosine treatment were not detectable, but adenosine concentrations increased to 2.4 ± 0.3 and 26.6 ± 1.6 µM, respectively during the 10 and 100 µM infusions. Base-line dialysate inosine levels in the control and 1 to 100 µM adenosine groups were 0.24 ± 0.04, 0.20 ± 0.03, 0.20 ± 0.06 and 0.31 ± 0.08 µM, respectively. The infusions of 10 and 100 µM adenosine increased dialysate inosine levels to 1.12 ± 0.17 and 4.13 ± 0.45 µM, respectively.

View larger version (15K): [in this window] [in a new window]   Fig. 1.   The effects of exogenous adenosine on preischemic coronary venous (A) and dialysate (B) adenosine concentrations in the isolated perfused rat heart. ADO was infused for 10 min at the indicated concentrations immediately before ischemia. Dialysate collections (10 min at 0.75 µl/min) were started immediately on initiating the ADO infusion, and effluent samples were collected after 5 min infusion. Mean values of effluent and dialysate ADO are shown above each bar. *, P < .05 vs. control values, n  5 per group.

Dialysate adenosine concentrations during the period of global ischemia are shown in figure 2. As expected, dialysate adenosine levels increased rapidly with the onset of ischemia. During the first two dialysate collection times (0-6 and 6-12 min) during ischemia dialysate adenosine levels in control hearts increased to 0.76 ± 0.11 and 8.12 ± 1.14 µM, respectively. Dialysate levels during the same time period in hearts treated with 1 µM adenosine were comparable to control hearts (0.46 ± 0.06 and 8.86 ± 1.54 µM, respectively). In 10 µM adenosine-pretreated hearts, dialysate adenosine concentrations were significantly elevated for the first 6 min of ischemia (4.73 ± 1.91 µM, P < .05) but were comparable to control hearts during the subsequent 6-min collection (11.57 ± 1.04 µM). Dialysate adenosine levels in hearts pretreated with 100 µM adenosine were significantly greater than those in control hearts during the first 12 min of ischemia: 23.91 ± 3.05 and 17.07 ± 2.34 µM, respectively, P < .05). After the first 12 min of ischemia, there were no differences in dialysate adenosine concentrations among the groups. Dialysate levels increased throughout the ischemic period and attained concentrations of ~40 µM during the 24- to 30-min collection.

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Dialysate ADO concentrations during 30 min of global normothermic ischemia. Five 6-min dialysate samples (at 0.75 µl/min) were collected during ischemia. The data points for each sample are plotted at 3, 9, 15, 21, and 27 min on the x-axis. The ADO treatments were administered as described in figure 1. Values are mean ± S.E., n  5 per group, *, P < .05 vs. control hearts.

The effects of global ischemia on dialysate inosine concentrations are illustrated in figure 3. inosine levels in the control hearts during the first 6-min ischemia were 1.30 ± 0.27 µM, a value comparable to that in hearts treated with 1 µM adenosine (0.82 ± 0.14 µM). Significantly greater dialysate inosine concentrations were seen in hearts treated with 10 µM adenosine (2.48 ± 0.29 µM) and 100 µM adenosine (8.84 ± 1.00 µM). Dialysate inosine levels in the latter two groups were also greater than control levels after 12-min ischemia. Thereafter there were no differences in inosine concentrations for the remainder of ischemia.

View larger version (16K): [in this window] [in a new window]   Fig. 3.   Dialysate INO concentrations during global ischemia in control and ADO-treated hearts. Dialysate samples were collected as described in figure 2. Values are mean ± S.E., n  5 per group, *, P < .05 vs. control hearts.

Similar to its metabolic effects, adenosine pretreatment was associated with dose-dependent effects on postischemic ventricular function (fig. 4). Control hearts recovered 32 ± 3% and 60 ± 3% of preischemic LVDP after 10- and 30-min reperfusion, respectively. Hearts treated with 1 µM adenosine recovered similarly (34 ± 9% and 46 ± 7% at 10 and 30 min). Hearts pretreated with higher concentrations of adenosine exhibited significantly greater postischemic function. At 10-min reperfusion, hearts pretreated with 10 and 100 µM adenosine recovered 56 ± 5% and 83 ± 8%, respectively (both P < .05 vs. control and 1 µM adenosine). After 30-min reperfusion, recovery in these two groups was 80 ± 5% and 90 ± 4%, respectively. After 30-min reperfusion, LVEDP in control hearts was 60 ± 5 mm Hg, similar to that in 1 µM adenosine-treated hearts (56 ± 7 mm Hg). In contrast, LVEDP values after 30-min reperfusion in hearts treated with 10 and 100 µM adenosine were 45 ± 3 and 28 ± 7 mm Hg, respectively.

View larger version (19K): [in this window] [in a new window]   Fig. 4.   Postischemic functional recovery in control and ADO-treated isolated hearts. Data are expressed as percent of preischemic (PreI) LVDP. Values are mean ± S.E., n  5 per group, *, P < .05 vs. control hearts.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

The results of this study performed in the isolated rat heart indicate that adenosine must be administered at a dose sufficient to overcome the coronary endothelial metabolic barrier to enhance postischemic ventricular function. Increasing doses of adenosine (1, 10, 100 µM) increased preischemic coronary flows to similar extents, but only the two higher doses of adenosine decreased spontaneous heart rate, increased dialysate adenosine concentrations and attenuated postischemic dysfunction. In addition, dialysate adenosine levels were elevated only immediately before ischemia and for the first 5 to 6 min of ischemia. Because the microdialysis technique samples the ISF compartment, which bathes the cardiac myocytes, these results suggest that preischemic adenosine treatment exerts its protective effect on the cardiac myocytes.

Although the cardioprotective effects of ADO have been acknowledged for several years, its exact mechanism of action remains unknown. One aspect of adenosine cardioprotection that has been appreciated is the need to administer high concentrations of adenosine to achieve beneficial effects in ischemic myocardium. Initial experiments on adenosine's effects in ischemic and postischemic myocardium in isolated hearts typically used 50 to 100 µM adenosine (Reibel and Rovetto, 1979; Ambrosio et al., 1989). The working hypothesis was that high concentrations of adenosine would stimulate postischemic ATP resynthesis via the purine salvage pathway resulting in improved ventricular function. However, it is now thought that adenosine does not improve function via this mechanism, nor does reperfusion ATP content correlate with ventricular function (Neely and Grotyohann, 1984; Mallet and Bünger, 1993; Lasley et al., 1998b).

It is also now well recognized that adenosine is rapidly metabolized by endothelial cells and red blood cells resulting in its short half-life. In isolated perfused hearts, exogenous adenosine at a concentration of 1 µM is nearly entirely taken up and metabolized by coronary endothelial cells (Nees at al., 1985). With the advent of techniques to sample ISF metabolites, it became apparent that this metabolic barrier could produce significantly different adenosine concentrations in the ISF and vascular compartments under a variety of conditions (Fenton and Dobson, 1987; Heller and Mohrman, 1988; Matherne et al., 1990; Tietjan et al., 1990; Herrmann and Feigl, 1992). Despite these reports, there are few cardioprotection studies in which extracellular adenosine levels have been measured. To our knowledge, this is the first study to perform a dose-response curve to exogenous adenosine in which both ISF and vascular adenosine levels were measured and effects on postischemic function were determined. All three doses increased vascular adenosine levels and coronary flow, but only 10 and 100 µM adenosine increased preischemic ISF adenosine levels, estimated with the cardiac microdialysis technique, and decreased spontaneous heart rate. We observed nearly identical results with 1 and 10 µM adenosine on ISF adenosine concentrations as reported by Heller and Mohrman (1988), who used myocardial transudate fluid to estimate ISF adenosine.

The microdialysis technique has been used in numerous preparations to estimate ISF metabolites, such as adenosine. One of the disadvantages of this technique is that placement of the fiber induces local tissue trauma, which may alter the microenvironment from which the fiber samples ISF, and the dialysate fluid does not completely equilibrate with the adjacent ISF. The fiber perfusion rate determines the absolute and relative recovery rates of the metabolites of interest, and the in vivo recovery can be influenced by changes in coronary flow rates. Given the large increase in coronary flow with 1 µM adenosine in this study, it is possible that any adenosine accumulating in the ISF compartment could have been washed out faster than we could sample it. In addition, we were comparing instantaneous effluent measurements with dialysate measurements collected over a 10-min period. It must also be pointed out that estimates of ISF adenosine levels determined with cardiac microdialysis are greater than those determined via other techniques, so the vascular to ISF gradient reported here may be underestimated. Despite these limitations, the microdialysis technique does provide reliable directional changes in ISF adenosine levels and is the only technique currently available to estimate ISF metabolites during myocardial ischemia.

Given the potent coronary endothelial metabolic barrier (Nees et al., 1985) and the relatively low vascular adenosine concentration (~0.3 µM) achieved with the 1 µM adenosine infusion, it was expected that there would be little, if any, increase in dialysate adenosine. The fact that vascular adenosine levels increased 10-fold but dialysate adenosine levels did not change is consistent with the high capacity of the coronary endothelium to metabolize adenosine. Increasing the adenosine concentrations to 10 and 100 µM did not further increase coronary flow but significantly increased dialysate adenosine concentrations. During these two infusions, vascular adenosine levels were only 3-fold greater than dialysate adenosine concentrations. The likely explanation for these findings is that the adenosine transport and metabolism pathways were saturated, resulting in the paracellular diffusion of adenosine from the vascular to ISF compartments. This hypothesis is supported by the observation that dialysate inosine concentration during the 100 µM adenosine infusion was only 4-fold greater than that during the 10 µM adenosine infusion. In the presence of red blood cells, which rapidly transport and metabolize adenosine, this gradient would be expected to be larger. We have reported that in the intact pig, a 50 µg/kg/min intracoronary adenosine infusion was associated with a coronary venous adenosine level of 15.2 ± 1.8 µM, whereas the dialysate adenosine concentration was only 2.0 ± 0.1 µM, a 7-fold gradient (Lasley et al., 1998a).

The increases in ISF adenosine with the 10 and 100 µM doses and the effects they produced in this study are consistent with at least two adenosine A₁ receptor-mediated effects. These two doses produced dose-dependent decreases in spontaneous heart rate, an effect consistent with adenosine's negative chronotropic effects which are mediated by the activation of myocyte A₁ receptors (Belardinelli et al., 1989). In addition, these two doses of adenosine improved postischemic function similar to that produced by treatment with adenosine A₁ receptor agonists. Preischemic administration of adenosine exerts its beneficial effects on postischemic function and infarct size by activating adenosine A₁ and/or A₃ adenosine receptors located on the cardiac myocytes (Lasley and Mentzer, 1992; Thornton et al., 1992; Yao and Gross, 1993; Grover et al., 1996; Auchampach et al., 1997). In contrast, the preischemic infusion of adenosine A₂ receptor agonists does not attenuate postischemic ventricular dysfunction (Lasley and Mentzer, 1992; Yao and Gross, 1993).

The observation in the present study that pretreatment with 10 µM adenosine improved postischemic function is at odds with two previous studies using comparable doses. Grover and Sleph (1994) reported that a 10-min pretreatment with 30 µM adenosine neither reduced spontaneous heart rate nor protected the globally ischemic isolated perfused rat heart. It is not clear why they observed no decrease in heart rate with this dose of adenosine because we saw a decrease with a 3-fold lower dose. Regarding the lack of protection reported by Grover and Sleph (1994), although the ischemia time in their study (25 min) was similar to that in the present study (30 min), control hearts in their study recovered only 16% of preischemic rate-pressure product compared with 60% recovery in our study. Hohlfeld et al. (1989) reported that 15 µM adenosine administered for 2 hr before ischemia did not attenuate postischemic ventricular dysfunction in the isolated perfused rat heart. However, the adenosine infusion was terminated 5 min before ischemia, and this adenosine preconditioning regimen does not improve postischemic function (Asimakis et al., 1993; Sekili et al., 1995).

The protection seen with 10 µM adenosine in the present study is consistent with at least two other studies. Fralix et al. (1993) reported that 20 µM adenosine decreased intracellular Ca⁺⁺ and H⁺ accumulations during global ischemia in the isolated rat heart and increased postischemic function. More recently Woolfson et al. (1996) performed an adenosine preconditioning dose-response curve in the isolated perfused rabbit heart. These investigators reported that doses of 6 to 100 µM adenosine (including 10 µM adenosine) dose-dependently reduced infarct size. Although we did not measure infarct size in the present study, it is likely that irreversible injury occurred in our 30-min ischemia protocol, as control hearts exhibited severe increases in end-diastolic pressures during ischemia and reperfusion. Thus, the improvement in postischemic function in this study was most likely due to reduction of both stunning and infarction. We did not perform as extensive an adenosine dose-response curve as Woolfson et al. (1996), but we did observe dose-dependent cardioprotective effects of adenosine. Hearts treated with 100 µM adenosine exhibited a more rapid recovery of postischemic LVDP than did hearts treated with 10 µM adenosine. Hearts treated with the higher dose also exhibited lower end-diastolic pressures during reperfusion. These dose-dependent cardioprotective effects of 100 µM adenosine are consistent with greater increases in preischemic ISF adenosine and a more pronounced negative chronotropic effect (i.e., greater extent of adenosine A₁ receptor activation).

An important observation in the present study is that the elevations in ISF adenosine with the 10 and 100 µM ADO pretreatments persisted only before ischemia and for the initial 6 min of ischemia. We observed similar findings in the anesthetized, open chest pig where dialysate adenosine levels in hearts treated with intracoronary adenosine (50 µg/kg/min) before coronary artery occlusion remained elevated for the first 5 min of ischemia but not thereafter (Lasley et al., 1998b). These results suggest that maintaining elevated ISF adenosine levels during the early ischemic period plays an important role in adenosine cardioprotection. However, in two additional studies in our laboratory (one in dogs, one in pigs) using the microdialysis technique, adenosine cardioprotection was not associated with elevated ISF adenosine levels during the early ischemic period (Randhawa et al., 1995; Martin et al., 1997). The common observation in all of our adenosine cardioprotection studies has been the increase in ISF adenosine concentration before the onset of ischemia. These results suggest that the preischemic elevation in ISF adenosine levels maintains the cardiomyocyte adenosine A₁ (and A₃) receptor in an activated state at the onset of ischemia to initiate adenosine's cardioprotective effect. This requirement may also explain why adenosine preconditioning, in which the adenosine infusion is terminated before ischemia, does not improve postischemic function (Asimakis et al., 1993; Sekili et al., 1995).

Although it is likely that the 10 and 100 µM adenosine concentrations activated multiple adenosine receptors, we think that the cardioprotective effect of adenosine in this acute model of ischemia/reperfusion is due to adenosine A₁ receptor activation. The 1 µM adenosine dose increased coronary flow (A₂ receptor activation) to the same extent as the two higher doses but did not alter spontaneous heart rate (A₁ receptor activation). Further evidence for the lack of adenosine A₂ receptor involvement is provided by several reports that adenosine A₂ receptor agonists do not attenuate postischemic dysfunction (Lasley and Mentzer, 1992; Yao and Gross, 1993). The beneficial effect of adenosine on postischemic function in the rat heart is mimicked by adenosine A₁ receptor agonists (Lasley and Mentzer, 1992; Grover et al., 1996), and the cardioprotective effects of adenosine in the isolated rat heart are blocked by the selective adenosine A₁ receptor antagonist DPCPX (Lasley and Mentzer, 1992; Grover and Sleph, 1994). The antistunning effect of adenosine in the dog is also blocked by DPCPX (Yao and Gross, 1993). Because the rat adenosine A₃ receptor is very insensitive to DPCPX (Zhou et al., 1992), this finding suggests that the adenosine A₃ receptor is not involved in this cardioprotective effect. Finally, although there have been numerous reports that adenosine A₃ receptor agonists reduce infarct size in the rabbit, to date there is only one report, also in the rabbit, that the adenosine A₃ receptor may play a role in adenosine-induced improvement of postischemic function (Auchampach et al., 1997).

In summary, the results of this study indicate that exogenous adenosine must be infused at a dose sufficient to overcome its rapid and extensive metabolism by the coronary endothelium to accumulate in the ISF compartment. The cardioprotective effect of adenosine was associated elevated ISF adenosine only at the onset of ischemia and for the first 6 min of ischemia. Because we have previously reported that the protective effect of high dose adenosine (100 µM) in this same preparation was blocked by the A₁ receptor antagonist DPCPX (Lasley and Mentzer, 1992), this suggests little involvement of adenosine A₃ receptors in adenosine cardioprotection in the rat.