Heart and Circulatory Physiology

Ecto-5′-nucleotidase is not required for ischemic preconditioning in rabbit myocardium in situ

Takayuki Miki, Tetsuji Miura, Rolf Bünger, Katsuo Suzuki, Jun Sakamoto, Kazuaki Shimamoto


This study tested the hypothesis that cardiac ecto-5′-nucleotidase (ecto-5′-NT) activity during ischemic preconditioning (PC) contributes to augmented tolerance against ischemia, thereby reducing infarct size in the rabbit heart in situ. The effects of α,β-methylene-adenosine diphosphate (AOPCP), a selective inhibitor of ecto-5′-NT, on cardiovascular responses to AMP were measured to establish in vivo activities of the enzyme and its inhibitor. Left atrial infusion of AOPCP (0.75 mg ⋅ kg−1 ⋅ min−1) raised AOPCP plasma levels to 138 μM; under these conditions negative chronotropic and inotropic effects of AMP were blocked, demonstrating essentially full inhibition of ecto-5′-NT in the heart in situ. This AOPCP-blocked heart in situ model was used to examine the proposed contribution of ecto-5′-NT in ischemic PC. Myocardial infarction caused by 30-min ischemia was followed by 3-h reperfusion. Infarct size (IS) was measured and expressed as a percentage of the size of the area at risk (%IS/AR). In untreated controls, %IS/AR was 38.1 ± 3.8%; PC (5-min ischemia, 5-min reperfusion) markedly reduced %IS/AR to 10.0 ± 2.0%. Essentially identical IS reductions by PC were observed in AOPCP-blocked animals (%IS/AR = 13.8 ± 2.2 and 13.3 ± 1.8% in rabbits receiving AOPCP at 0.75 and 1.50 mg ⋅ kg−1 ⋅ min−1, respectively); here plasma AOPCP levels were established before and during PC but not during the subsequent prolonged ischemia. As expected, AOPCP also did not affect %IS/AR in non-PC controls (%IS/AR = 35.5 ± 3.7%). In contrast but as predicted, adenosine-receptor blockade by 8-phenyltheophylline (10 mg/kg iv) substantially attenuated IS reduction by PC in both AOPCP-blocked and control hearts (%IS/AR = 25.2 ± 4.3 and 21.8 ± 2.2%, respectively;P < 0.05 vs. PC alone). The results demonstrate that cardiac ecto-5′-NT is not required for ischemic PC against infarction in the rabbit.

  • adenosine
  • adenosine 5′-monophosphate
  • myocardial infarct size
  • in situ heart
  • α,β-methylene-adenosine diphosphate
  • 8-phenyltheophylline

preconditioning the myocardium with a brief transient ischemia enhances its tolerance against ischemic damage. A number of recent studies support the hypothesis that cardioprotection by ischemic preconditioning is mediated by adenosine receptors in various species, with the exception of the rat (6, 18,33). The exact metabolic mechanism of adenosine accumulation in the cardiac interstitium during preconditioning remains controversial. Under normoxic conditions, adenosine is thought to be derived in part from S-adenosyl homocysteine but mainly from dephosphorylated cytosolic AMP via cytosolic 5′-nucleotidase (3, 4, 19, 20, 26). In ischemic or hypoxic myocardium, adenosine is generated predominantly via the cytosolic 5′-nucleotidase pathway. The ecto-5′-nucleotidase may also contribute to net adenosine production (11, 20, 29, 30). The relative contributions of the extra- and intracellular pathways to interstitial adenosine accumulation during preconditioning are not known. Kitakaze et al. (14-16) recently proposed that stimulation of ecto-5′-nucleotidase by α-adrenoceptor activation may be importantly involved in cardioprotection by preconditioning. These investigators reported that the enzymatic activity of cardiac ecto-5′-nucleotidase was increased by ischemic preconditioning and also by α-adrenoceptor stimulation using methoxamine (15). It was also found that α,β-methylene adenosine diphosphate (AOPCP), a strong inhibitor of both native and isolated ecto-5′-nucleotidase (20), and prazosin, an α-adrenoceptor blocker, given before preconditioning attenuated the infarct size-limiting effect of preconditioning in the dog heart in situ (14). However, the requirement for ecto-5′-nucleotidase in ischemic preconditioning has not been established for other models of myocardial infarction, including rabbit and pig heart in situ. In addition, the possibility has been raised that the requirement of this enzyme for preconditioning might differ among species. In the rabbit heart, for example, α-adrenergic blockade by phentolamine (11), BE-2254, a specific α-receptor antagonist (32), and prazosin (T. Miura, unpublished observations) all failed to abolish or alleviate the protection by ischemic preconditioning against infarction. Thus the putative α-adrenoceptor-mediated stimulation of ecto-5′-nucleotidase may not represent a species-independent, fundamental mechanism of cardioprotection by ischemic preconditioning.

In an attempt to delineate the possible role of cardiac ecto-5′-nucleotidase during ischemic preconditioning, the native enzyme was examined in the heart in situ using the rabbit model (9, 21,33). To inhibit ecto-5′-nucleotidase in situ, we infused a large dose of AOPCP (0.75 mg ⋅ kg−1 ⋅ min−1) into the left atrium before and during preconditioning. Inhibitory potency of AOPCP in vivo was verified in two ways. First, we measured the effects of AOPCP on hemodynamic responses to bolus injections of AMP, the immediate substrate of ecto-5′-nucleotidase. The hemodynamic effects of bolus AMP appear to be mainly caused by dephosphorylation to adenosine by ecto-5′-nucleotidase (7, 27). Second, we measured the level of free AOPCP in plasma and compared it with the known inhibition constant (K i) of ecto-5′-nucleotidase. Finally, to test whether ecto-5′-nucleotidase is required for ischemic preconditioning against infarction in vivo, the effects of intravenous AOPCP before and during preconditioning were quantitated in terms of infarct sizes and area at risk, using computer-assisted planimetry in combination with standard infarct staining techniques.


Experiment 1: Efficacy of Intravenous AOPCP as Inhibitor of Native Ecto-5′-Nucleotidase in In Situ Rabbit Hearts

Surgical preparation.

Male albino rabbits (Japanese White) weighing 2.2–2.6 kg were anesthetized with pentobarbital sodium (30 mg/kg iv) and ventilated with a Harvard respirator (model 683). Respiratory rate, tidal volume, and inspired oxygen were adjusted to maintain arterial blood gases within the physiological range. The chest was opened by left thoracotomy, and the heart was exposed. A saline fluid-filled catheter was placed in the right carotid artery and connected to a Nihon-Kohden SCK-580 transducer to monitor arterial blood pressure. Catheters also were inserted into the jugular vein and the left atrium for infusing drugs. A catheter-tip manometer (Nihon-Kohden CTD-0960; diameter 2 mm) was advanced into the left ventricle (LV) via the left atrial appendage to monitor LV pressures and the first time derivative of LV pressure (LV dP/dt) obtained by electronic differentiation using a Nihon-Kohden EQ-601G unit. In some rabbits (group 3), a plastic T tube (3.0-mm ID) was inserted into the descending thoracic aorta with a side arm connecting to a reservoir (volume 2 l) by a Tygon tube (3.5-mm ID) to externally control systemic blood pressure. The reservoir was filled with Ringer solution at 37°C.

Specific protocols.

The effects of intravenous AMP on hemodynamic parameters were assessed in three groups. In group 1(n = 4 rabbits), rabbits received bolus injections of AMP (0.034 mg/kg) before and after intravenous 8-phenyltheophylline (8-PT, 10 mg/kg), which nonselectively inhibits adenosine receptors (10). The hemodynamic effects of AMP were of short duration and fully reversible within 30 s; 8-PT was given 5 min after the first AMP bolus. The second bolus of AMP was administered 10 min after the 8-PT injection. In group 2(n = 6 rabbits), AMP boluses were administered before and during AOPCP infusion. Five minutes after the first bolus of AMP, 0.75 mg ⋅ kg−1 ⋅ min−1of AOPCP was infused into the left atrium, and five minutes later the second injection of AMP was given. The AOPCP infusion was continued until the hemodynamic parameters returned to pre-AMP levels. Ingroup 3(n = 5 rabbits), the AMP-induced decreases in blood pressure were minimized by instrumenting the rabbits with the aortic pressure reservoir (see Surgical preparation). The hemodynamic responses to AMP (in 0.034 and 0.068 mg/kg doses) were measured in the absence and presence of intra-atrial AOPCP. The hydrostatic height of the pressure reservoir was adjusted to match the mean arterial blood pressure immediately before the reservoir line was opened. When the maximum negative chronotropic effect of AMP was recorded, the reservoir line was closed to avoid excessive entry of Ringer solution into the aorta during the AMP-induced vasodilatation. Although the aortic reservoir could not completely clamp the arterial pressure level, it markedly attenuated the blood pressure fall by AMP-induced vasodilatation (seeresults). Heart rate, mean and phasic arterial blood pressures, LV pressures, and LV dP/dt were continuously recorded using an eight-channel direct recorder (Nihon-Kohden WT-687G). Additional rabbits (group 4,n = 6) were instrumented to assess plasma AOPCP level during AOPCP infusion at 0.75 mg ⋅ kg−1 ⋅ min−1. In these rabbits, AOPCP was infused into the left atrium at 0.75 mg ⋅ kg−1 ⋅ min−1, and 1.5 ml arterial blood samples were taken before and 8 and 16 min after AOPCP infusion was started. Plasma was separated from blood cells by centrifugation at 4°C and processed for AOPCP assay.

Determination of free AOPCP in plasma water.

Plasma separated by centrifugation was loaded onto an ultrafiltration cartilage (Minicent-10, mol wt cutoff = 10,000, TOSOH) and centrifuged again at 1,500 g for 30 min to obtain a deproteinized plasma water sample. AOPCP concentration in the sample was determined by HPLC. The HPLC system consisted of an EP-300 pump, an Eicompack MA-5ODS column (Eicom), and a Soma UV-VIS Detector/S-3702 (Soma-Kogaku). The mobile phase was 0.1 M KH2PO4buffer (pH = 3.5) containing 1% acetonitrile. The ultrafiltrate was diluted with nine parts of Ringer solution, and 20 μl of the diluted sample were injected into the reverse-phase column, which was maintained at 25°C by a column thermostat (Eicom AT-300 Column Oven). AOPCP eluted as a sharp peak at ∼3.4 min. A calibration curve, based on 100- to 1,000-ng AOPCP injections, demonstrated linearity between AOPCP concentration and measured peak absorbance (r = 0.988,n = 18 samples).

Experiment 2: Effects of AOPCP on Interstitial Adenosine as Assessed by In Vivo Cardiac Microdialysis

To examine the possibility that AOPCP per se might alter extracellular adenosine concentrations, we measured the effects of AOPCP on dialysate adenosine levels in the heart.

Surgical preparation.

We instrumented four rabbits as inexperiment 1 without placing the catheter-tip manometer into the LV. Details of the microdialysis method and its validity for detecting changes in interstitial adenosine in the rabbit were described previously (22). Using a 26-gauge needle, we inserted the microdialysis probe (Eicom OP-100–05, Eicom, Japan), consisting of a 5-mm-long dialysis fiber (220-μm OD, molecular size cutoff = 50,000 Dalton) glued between two polyethylene tubes of 0.28-mm ID, into the midmyocardium in the territory of the left marginal artery. The dialysis probe was perfused at 2 μl/min with Ringer solution (in mM: 147 Na+, 4 K+, 4.5 Ca2+, 155.5 Cl) containing 10 U/ml of low-molecular-weight heparin (Dalteparin; mol wt ∼5,000). A 2-h stabilization period after insertion of the probe allowed the dialysis purines to reach a stable level with adenosine in the submicromolar range.

Specific protocol.

At baseline dialysate was collected for 10 min, and then infusion of AOPCP into the left atrium at the rate of 0.75 mg ⋅ kg−1 ⋅ min−1was begun. Fifteen minutes later the second dialysate sample was collected while intravenous AOPCP infusion continued. The dialysate samples were frozen at −20°C and assayed within 14 days.

Assay of dialysate purines.

Purines (adenosine, inosine, hypoxanthine, and xanthine) in the dialysate were analyzed by the same HPLC system and mobile phase as those in experiment 1. Of the 20 μl of dialysate collected during the 10-min sampling period, 15 μl were injected into the chromatography column. Absorbance was measured at 260 nm, and the dialysate purines were identified by retention times (22). The recovery of purine using in vitro dialysis of a known standard solution (containing 1 ng/l of each compound) using the Eicom OP-100–05 dialysis probe at a perfusion rate of 2 μl/min and 37°C was 17.9 ± 0.8% for adenosine, 19.4 ± 0.9% for inosine, and 26.9 ± 1.1% for hypoxanthine (n = 6 samples).

Experiment 3: Effect of Intravenous AOPCP on Preconditioning Against Infarction in In Situ Heart

Surgical preparation and infarct model.

Surgical preparation and instrumentation of the experimental myocardial infarction model was detailed previously (9, 21, 33). In brief, male albino rabbits weighing 2.0–2.7 kg were anesthetized, ventilated, and thoracotomized as in experiment 1. Catheters were placed in the right carotid artery, the jugular vein, and the left atrium. A silk thread (4–0 silk) was passed around a large marginal branch of the left coronary artery to be used as a snare for coronary occlusion. Arterial blood pressure and bipolar electrocardiogram were monitored.

Specific protocols.


Rabbits were divided into five groups, all of which were subjected to 30 min of coronary occlusion followed by 3 h of reperfusion (Fig.1). The control group (n = 11) were untreated; the PC group (n = 10) were preconditioned with 5 min of coronary occlusion followed by 5 min of reperfusion before the 30-min ischemia-reperfusion. In the AOPCP group (n = 7) without preconditioning, 0.75 mg ⋅ kg−1 ⋅ min−1of AOPCP was continuously infused into the left atrium for 16 min; infusion was started 15 min before the 30-min ischemia. In the AOPCP-treated preconditioned groups, the AOPCP-PC1 group (n = 7) and the AOPCP-PC2 group (n = 6) received AOPCP at 0.75 and 1.50 mg ⋅ kg−1 ⋅ min−1, respectively. Timing and duration of the AOPCP infusion in these later groups were identical to those in the AOPCP group, i.e., AOPCP was infused for 16 min commencing 5 min before preconditioning. This AOPCP infusion protocol was selected to primarily inhibit ecto-5′-nucleotidase activity during preconditioning but not that during the reperfusion period. In pilot experiments, the attenuating effect of AOPCP (at 0.75 mg ⋅ kg−1 ⋅ min−1) on the hemodynamic response to AMP almost disappeared 15 min after discontinuation of AOPCP infusion, suggesting that the inhibitory effect of AOPCP quickly diminishes within this time frame.

Fig. 1.

Experimental protocols in experiment 3. Inprotocol 1, preconditioning (PC) was performed by 5-min coronary occlusion and 5-min reperfusion before 30-min sustained ischemia. Infusion of α,β-methylene-adenosine diphosphate (AOPCP) into left atrium was started at 15 min before sustained ischemia and continued for 16 min. Infusion rate of AOPCP was 0.75 mg ⋅ kg−1 ⋅ min−1in AOPCP and AOPCP-PC1 groups and 1.50 mg ⋅ kg−1 ⋅ min−1in AOPCP-PC2 group. In protocol 2, both PT-PC and PT/AOP-PC groups received intravenous injection of 10 mg/kg 8-phenyltheophylline (PT; indicated by arrows) and then PC before sustained ischemia. Saline (i.e., vehicle) was infused in PT-PC group, and AOPCP was administered in PT/AOP-PC groups through catheter inserted into left atrium. Timing of intra-atrial infusion was same as that inprotocol 1.

After 3 h of coronary reperfusion, the rabbits were heparinized (2,000 U iv) and killed with a pentobarbital overdose (60 mg/kg). The heart was excised and immediately processed for postmortem analysis.


Although AOPCP failed to prevent preconditioning inprotocol 1 (seeresults), the possibility remained that adenosine rather than AMP contributed to the observed preconditioning effect in presence of AOPCP. To test this hypothesis, 8-PT was used to nonselectively inhibit cardiomyocyte adenosine receptors, which were previously implicated in cardioprotection (6). We previously showed that 10 mg/kg of 8-PT alone did not change infarct size in normal control animals but significantly inhibited, albeit not completely, preconditioning against infarction in open-chest, anesthetized rabbits (33). Accordingly, 8-PT-treated rabbits were divided into two groups, a PT-PC group (n = 9) and a PT/AOP-PC group (n = 9), as shown in Fig. 1. All animals were subjected to the 30-min ischemia-reperfusion regimen and also preconditioned during the adenosine receptor blockade with 8-PT. As requisite new controls, AOPCP (0.75 mg ⋅ kg−1 ⋅ min−1) and its vehicle (physiological saline) were infused into these methylxanthine-treated groups using the left atrial catheter. 8-PT was injected 10 min before preconditioning or 20 min before sustained ischemia-reperfusion. Infusion of AOPCP or vehicle was started 15 min before ischemia and continued for 16 min.

Postmortem quantitation of infarct size and area at risk.

Hearts were excised and immediately immersed in physiological saline to avoid air embolism in the coronary artery. The heart was mounted on a Langendorff apparatus and perfused at 60 mmHg with saline for 30 s to wash out blood. The coronary snare was retightened, and a suspension of fluorescent particles (3- to 30-μm diameter, Duke Scientific, Palo Alto, CA) was infused into the aortic cannula to negatively mark the ischemic region. The heart was then frozen in a freezer (−20°C) and subsequently sectioned into 2-mm-thick slices parallel to the atrioventricular plane. The slices were incubated in 100 mM sodium phosphate buffer with 1% triphenyltetrazolium (pH = 7.4) at 37°C for 20 min to visualize the infarcts. The macroscopic outlines of infarct and area at risk were traced on a clear acetate sheet using room light and ultraviolet light, respectively. The areas delineated were measured using a computer-assisted image analysis system, PIAS II (PIAS, Osaka, Japan). The actual sizes or volumes of infarct and of the risk region were obtained by multiplying the known slice thickness (2 mm) by the measured areas.


AMP, AOPCP, and 8-PT were purchased from Sigma (St. Louis, MO). Low-molecular-weight heparin (Dalteparin) was from Kissei Pharmaceutical (Tokyo, Japan). AMP and AOPCP were dissolved in 0.9% saline to prepare 0.5 mg/ml and 10 mg/ml solutions, respectively, before use. For a 8-PT injection, 25 mg of 8-PT were dissolved in 0.5 ml of dimethyl sulfoxide and 0.5 ml of 1 N NaOH, which was then diluted with saline to 5 ml.

Statistical Analysis

Data are presented as single values or as means ± SE. Hemodynamic function before and after treatment with 8-PT or AOPCP was analyzed by paired t-test. Infarct sizes and hemodynamics between groups were compared using one-way ANOVA and repeated-measures ANOVA, respectively. Linear correlations between infarct size and the size of the risk area were obtained using ordinary least-square methods. To test for differences in the regression lines between the study groups, analysis of covariance was performed, using infarct size as the dependent variable and the size of area at risk as the independent covariate.


Experiment 1: Effects of 8-PT and AOPCP on Hemodynamic Responses to Systemic AMP

Table 1 summarizes cardiac and systemic effects of AMP injected before and during adenosine receptor blockade (by 8-PT) or during AOPCP infusion. Under control conditions, AMP did not affect heart rate but reduced mean arterial blood pressure by 18 mmHg and maximum LV dP/dt (LV dP/dt max) by 475 mmHg/s. Adenosine receptor blockade by 8-PT nearly completely abolished the AMP-induced hypotension and hence the fall in LV dP/dt max. These findings suggested that the hypotension caused by AMP was mediated at least in part by adenosine; this is the case because adenosine is the 5′-nucleotidase-catalyzed dephosphorylation product of AMP and because AMP is a much less potent vasodilator than adenosine (2). To ascertain that ecto-5′-nucleotidase was mediating the hemodynamic responses to injected AMP, AOPCP, a selective blocker of the ectoenzyme in intact tissues (20), was infused (at 0.75 mg ⋅ kg−1 ⋅ min−1). HPLC analysis of plasma in group 4 (n= 6) showed that this regimen established free plasma AOPCP levels of 138 ± 19 and 147 ± 26 μM, 8 and 16 min after the onset of infusion, respectively; these levels were well above the 50 μM level known to inhibit the native ectoenzyme by >90% (20, 30). During the AOPCP infusion mean blood pressure, but not heart rate or LV dP/dt max, was slightly reduced from 94 to 80 mmHg. Injection of AMP still induced arterial hypotension, albeit by only 8 mmHg instead of the 22 mmHg in the control group (Table 1, group 2). Thus AOPCP inhibition of ecto-5′-nucleotidase attenuated AMP-induced hypotension by 63%, the remainder of the hypotensive effect likely reflecting the direct vasodilator action of AMP (2), although we cannot fully rule out that some residual ecto-5′-nucleotidase activity is responsible, in part, for the small hypotension after AMP injection in the AOPCP-treated rabbit. AOPCP per se decreased mean aortic pressure slightly (Tables 1 and 2). However, it is unlikely that this effect rather than blockade of ecto-5′-nucleotidase caused attenuation of the AMP hypotensive response; indeed, the hypotensive effect of AMP was virtually identical at blood pressure between 94 (control ingroup 2) and 81 (control ingroup 1) mmHg.

View this table:
Table 1.

Effects of 8-phenyltheophylline and AOPCP on hemodynamic response to AMP injection

View this table:
Table 2.

Effects of AOPCP on hemodynamic response to AMP injection in rabbits with an aortic pressure reservoir (group 3)

Table 2 summarizes the data from AMP injections into rabbits whose blood pressure change was minimized by the aortic reservoir technique (see materials and methods). In the rabbits instrumented with the aortic reservoir (group 3), the AMP-induced hypotension was only ∼50% of the change seen without the reservoir (groups 1 and2 in Table 1). Interestingly, this reduction in AMP-induced hypotension unmasked the expected but small negative chronotropic effect of AMP (Table 2). This negative chronotropism by AMP was also inhibited by AOPCP, indicating mediation by ecto-5′-nucleotidase and hence adenosine.

Experiment 2: Effect of AOPCP on Cardiac Dialysate Adenosine

In the control period, dialysate adenosine, inosine, and hypoxanthine levels were low (0.08 ± 0.03, 0.22 ± 0.02, and 0.80 ± 0.08 μM, respectively). Purine levels in the dialysate did not significantly change during AOPCP infusion (adenosine, 0.06 ± 0.03; inosine, 0.18 ± 0.04; hypoxanthine, 0.80 ± 0.10 μM). Although these purine levels in the dialysate are lower than in earlier studies using rabbits and dogs (17, 35), this difference in the dialysate adenosine may be caused by different dialysis probe technologies and animal species, as discussed in our recent study (22). Indeed, the interstitial adenosine level estimated using our measured recovery was 0.44 μM. This level is quite similar to the value of 0.24 μM reported by others for rabbit hearts (17).

Experiment 3: Effect of Preconditioning, AOPCP, and 8-PT on Infarct Size

Effects of AOPCP.

Heart rate and blood pressures were comparable in all experimental groups under baseline conditions (Table 3). AOPCP again reduced blood pressure by ∼15–20 mmHg, regardless of whether the dose was 0.75 (AOPCP and AOPCP-PC1 groups) or 1.50 (AOPCP-PC2 group) mg ⋅ kg−1 ⋅ min−1. However, there was no significant intergroup difference in rate-pressure products before and during coronary occlusion, suggesting that myocardial oxygen consumption during coronary occlusion was similar among the groups. In the PC group, infarct size expressed as a percentage of the area at risk (%IS/AR) was 10.0 ± 2.0% (Table4), which was nearly fourfold smaller than the %IS/AR in the control group (38.1 ± 3.8%). Table 4 also shows that, as expected, AOPCP per se did not change %IS/AR. Furthermore, AOPCP did not inhibit or abolish protection against infarction by preconditioning, because measured infarct size did not increase because of AOPCP in preconditioned rabbits (AOPCP-PC1 and AOPCP-PC2 groups). These results were obtained using AOPCP doses that were shown byexperiment 1 to fully inhibit ecto-5′-nucleotidase and hence AMP-dependent interstitial adenosine formation in situ. Nevertheless, %IS/AR may be influenced by the size of the risk area in the rabbit model of infarction (38). Thus we assessed the effect of AOPCP on preconditioning also in terms of the infarct size-risk area size relationship. As shown in Fig.2 A, the slope of the regression line relating infarct size and risk area was markedly decreased by preconditioning (P < 0.05). Clearly, our preconditioning protocol shifted infarcts toward smaller sizes for a given area at risk, a hallmark of myocardial protection by preconditioning. Figure 2 B shows that AOPCP, which did not reduce interstitial adenosine (see above), also did not alter the slope of the infarct size-risk size relationship. Finally, Fig. 2 C demonstrates that AOPCP also failed to increase the slope of the regression in preconditioned hearts.

View this table:
Table 3.

Hemodynamic data in experiment 3

View this table:
Table 4.

Infarct size data in experiment 3

Fig. 2.

Scatter plots of infarct size vs. size of area at risk.A: control (○) vs. PC (•). Regression lines: y = 0.55x − 0.12,r = 0.82, standard errors (SE) of slope and intercept = 0.13 and 0.10, respectively, in control group;y = 0.06x + 0.03,r = 0.30, SE of slope and intercept = 0.07 and 0.06, respectively, in PC group. Regression line of PC group was significantly less steep than control regression line.B: control (○) vs. AOPCP (•) group. Regression line in AOPCP group:y = 0.62x − 0.22,r = 0.89, SE of slope and intercept = 0.14 and 0.13, respectively. Regression lines of AOPCP group are not statistically different from control regression line.C: control (○) vs. AOPCP-PC1 (•) and AOPCP-PC2 (▵) groups. Regression lines:y = 0.10x + 0.03,r = 0.36, SE of slope and intercept = 0.12 and 0.12, respectively, in AOPCP-PC1 group;y = 0.24x − 0.10,r = 0.69, SE of slope and intercept = 0.13 and 0.12, respectively, in AOPCP-PC2 group. Slopes of regression lines of both AOPCP-PC1 and AOPCP-PC2 groups were significantly less steep than that of control group.

Effects of 8-PT.

There were no significant differences in heart rate, blood pressure, and rate-pressure products between the PT-PC and PT/AOP-PC groups under baseline conditions (Table 3). Preconditioning and AOPCP infusion slightly reduce mean blood pressure and rate-pressure products. However, the between-group difference in rate-pressure products was not statistically significant. Mean blood pressure during coronary occlusion was slightly lower in the PT/AOP-PC group than in the PT-PC group, but heart rates and rate-pressure products were not different between groups. The %IS/AR values were increased similarly (i.e., approximately doubled) in both theophylline groups (PT-PC, 21.8 ± 2.2%; PT/AOP-PC 25.2 ± 4.3%) relative to the PC control group, in which %IS/AR near 10% were obtained (Table 4). Clearly, infarct size reduction by preconditioning was similarly attenuated by adenosine-receptor blockade in the PT-PC and PT/AOP-PC groups. Alleviation of the protective preconditioning effect was also obvious from the risk area-infarct size relationship; the slope of the linear regression in both the PT-PC and the PT/AOP-PC groups was significantly shifted upward (Fig. 3).

Fig. 3.

Scatter plots of infarct size vs. size of area at risk. In addition to data of study groups in protocol 2, those of control and PC groups inprotocol 1 are presented for comparison. ○, Control; •, PC; □, PT-PC; ▵, PT/AOP-PC. Regression lines:y = 0.27x − 0.04,r = 0.95, SE of slope and intercept = 0.17 and 0.15, respectively, in PT-PC group;y = 0.32x − 0.03,r = 0.74, SE of slope and intercept = 0.15 and 0.12, respectively, in PT/AOP-PC group. Regression line for risk area size-infarct size relationship was not statistically different between PT-PC and PT/AOP-PC groups. However, upward shift of their regression lines to regression in PC group was statistically significant. Regression lines in PT-PC and PT/AOP-PC groups were statistically different from that in control group as well.


The findings of the present study in the anesthetized rabbit indicate that 1) large doses of intra-atrial AOPCP with proven blockade of native cardiac ecto-5′-nucleotidase fail to prevent infarct size reduction by ischemic preconditioning;2) however, intravenous 8-PT markedly lessens infarct size reduction by preconditioning both in control and AOPCP-blocked animals. These observations virtually rule out a requirement for cardiac ecto-5′-nucleotidase during ischemic preconditioning as far as protection against myocardial infarction in the rabbit is concerned. The data, however, confirm a substantial role for endogenous adenosine and the cardiac adenosine receptor system in the mediation of preconditioning in the rabbit heart in situ.

Assessing Bioactivity of Cardiac Ecto-5′-Nucleotidase and Efficacy of AOPCP

The activity of ecto-5′-nucleotidase determined in vitro cannot be directly extrapolated to that of the native enzyme in vivo, in which pH and activators such as Mg2+ can rapidly change. Accordingly, we bioassayed the effects of AOPCP as an inhibitor of in situ ecto-5′-nucleotidase on the cardiovascular effects of AMP, the immediate substrate of the enzyme. Infusion of AOPCP at 0.75 mg ⋅ kg−1 ⋅ min−1into the left atrium significantly attenuated the negative chronotropic and inotropic effects of exogenous AMP in the rabbit. This dose of AOPCP produced plasma AOPCP levels near 140 μM, well above the reported K i of ecto-5′-nucleotidase (i.e., 6 nM in rat heart plasma membrane and 19 nM in chicken gizzard smooth muscle membrane; Ref. 37). Intracoronary AMP is rapidly dephosphorylated by ecto-5′-nucleotidase to adenosine and further degradatives (20,26, 27). Intravenous injection of AMP caused a fall in blood pressure and subsequent reduction of LV dP/dtin the rabbit, and these effects were also blocked by 8-PT, a nonselective adenosine-receptor blocker, and AOPCP. These findings suggest that injected AMP was locally dephosphorylated to adenosine by native ecto-5′-nucleotidase and that adenosine- rather than AMP-mediated vasodilatation was responsible for observed hypotension.

The observed negative chronotropic and inotropic effects of exogenous AMP could underestimate the true pharmacological potency of AMP in vivo, because acute systemic hypotension results in reduction in LV afterload and also baroreflex sympathetic activation. Indeed, when the reduction in aortic pressure caused by AMP was minimized by the arterial pressure reservoir, the negative chronotropic effect of AMP became obvious (Table 2). Again, AOPCP completely blunted the negative chronotropic effect of AMP, implying that adenosine is the mediator of AMP-induced bradycardia and demonstrating in vivo efficacy of AOPCP as an inhibitor of ecto-5′-nucleotidase (Table 2). Measured plasma levels of AOPCP were ∼140 μM, which is a concentration about threefold higher than that required to inhibit >80% of ecto-5′-nucleotidase in isolated guinea pig hearts (20, 30). Even a small dose of 0.3 mg ⋅ kg−1 ⋅ min−1of AOPCP fully inhibits ecto-5′-nucleotidase activity in the membrane fraction prepared from preconditioned rabbit myocardium (M. Kitakaze, personal communication). Therefore, in agreement with the blunted cardiovascular responses to injected AMP, the present greater than twofold larger AOPCP dose (0.75 mg ⋅ kg−1 ⋅ min−1) clearly was adequate to virtually fully block native ecto-5′-nucleotidase in the rabbit heart in situ. However, this and higher AOPCP doses did not measurably impair preconditioning protection against infarction (Fig. 2).

Cardiac Ecto-5′-Nucleotidase and Preconditioning in Rabbit

The inability of AOPCP to inhibit ischemic preconditioning in rabbit heart could principally be caused by increased adenosine accumulation. However, this is not supported by present data and those in the literature. 1) We observed that cardiac dialysate adenosine tended to decrease, not increase, during AOPCP infusion (experiment 2).2) Other investigators reported that AOPCP and pentoxifylline, another 5′-nucleotidase inhibitor, preserve cardiac ATP rather than stimulating its degradation during both normoxia and ischemia (20, 28, 31).3) AOPCP per se in doses that block ecto-5′-nucleotidase does not raise epicardial transudate levels of adenosine or the release of adenosine plus inosine (20, 23, 30).

The cardiac adenosine receptor system is a known mechanism for preconditioning against infarction also in the rabbit (Fig. 3; Refs. 6,33). Thus cardiac ecto-5′-nucleotidase could plausibly contribute to preconditioning by producing adenosine in the vicinity of the A1 receptor of cardiomyocytes. The catalytic unit of the membrane-bound enzyme faces extracellularly (39), and its AMP substrate may originate from adenine nucleotides of sympathetic nerve terminals (i.e., ATP coreleased with norepinephrine) (36) and possibly also from endothelial cells (5). Also, necrotic (or severely ischemic) cardiomyocytes with leaky cell membranes (4) provide interstitial adenine nucleotides and hence adenosine near the adenosine A1 receptor.

Nevertheless, the exact contribution of ecto-5′-nucleotidase to total interstitial adenosine in the myocardium has been difficult to determine. Headrick et al. (12), using 50 μM AOPCP, reduced adenosine levels in both epicardial transudate and coronary effluent during severe hypoxia by ∼20% in isolated guinea pig hearts. A more marked reduction of adenosine release into transmyocardial effluent by AOPCP in guinea pig hearts was reported by Imai et al. (13). In the dog, intracoronary AOPCP (80 μg ⋅ kg−1 ⋅ min−1) attenuated elevation of the adenosine level in coronary venous blood during a 5-min reperfusion period following 5 min of preconditioning ischemia by up to 50% (14). These results, however, do not rule out that during the ischemia-reperfusion protocols plasma membrane became permeable to AOPCP, a methylene derivative of 5′-ADP; AOPCP would then diffuse into the intracellular compartment, where it could inhibit cytosolic 5′-nucleotidase. Thus the use of AOPCP under conditions of myocardial ischemia-reperfusion damage raises the possibility that AOPCP becomes an inhibitor of both cytosolic and ecto-5′-nucleotidase. Obviously, such a mechanism would effectively reduce the cytosolic 5′-nucleotidase pathway that is predominant during oxygen deficiency (29). Consistent with this interpretation and as mentioned above, in normoxic hearts with intact cell membranes there were no measurable changes in adenosine or adenosine plus inosine releases when AOPCP inhibited >85% of native ecto-5′-nucleotidase (20, 23,30).

The current failure of AOPCP to measurably attenuate preconditioning combined with the previous recognition that adenosine receptor blockade abolishes preconditioning (6, 33) provides strong albeit pharmacological evidence that in mildly ischemic myocardium adenosine can be produced independently of ecto-5′-nucleotidase in amounts sufficient to trigger the preconditioning cascade. Our finding that 8-PT markedly inhibits infarct size reduction by preconditioning even in AOPCP-blocked hearts (Fig. 3) supports the role of an adenosine mechanism not related to ecto-5′-nucleotidase. Consequently, ecto-5′-nucleotidase is not required for ischemic preconditioning against infarction, at least in the rabbit heart in situ. This fact does not necessarily exclude a contribution of the enzyme in the absence of AOPCP.

Disparate Effects of AOPCP Concerning Ischemic Preconditioning of Rabbit Versus Dog Hearts In Situ

Our rabbit heart findings directly contradict reports by Kitakaze et al. (14-16) on the dog model of infarction. In their studies, AOPCP when applied during preconditioning alone partially blocked (40%) the infarct size-limiting effect of ischemic preconditioning. Even more striking, protection by preconditioning was fully abolished when AOPCP infusion begun before preconditioning was maintained throughout ischemia plus 60 min of reperfusion (14). Whether such a protocol can prevent or attenuate preconditioning in the rabbit is not known. Clearly, one explanation for the disparate results from rabbits versus dogs could be protocol related. Another possibility is a true species difference in ecto-5′-nucleotidase contributions to interstitial adenosine accumulation during the mild ischemia imposed by preconditioning. Electron microscopic immunohistochemistry (1) revealed species-specific sites and contents of ecto-5′-nucleotidases. Histochemical ecto-5′-nucleotidase is located on pericytes, not on cardiomyocytes, in both the dog and the rabbit. Enzyme contents appear to be higher in dog than rabbit hearts and even higher in rat and guinea pig hearts (1). In vitro assays also indicate that the enzyme activity is ∼20% higher in dog than rabbit hearts (M. Kitakaze, personal communication). However, it is questionable whether such small differences in ecto-5′-nucleotidase activities could explain the marked disparities in the effects of AOPCP on preconditioning in rabbits versus dogs.

Another potential explanation for the conflict between our data and those of Kitakaze is the fact that preconditioning mechanisms are redundant and not necessarily identical between species. Although protein kinase C appears to play a major role in the cardioprotection of preconditioning in the rabbit, dog, and rat (6), the mechanism upstream to protein kinase C may not be necessarily the same between different species. Adenosine-unrelated mediators such as bradykinin (8) and angiotensin II (24) can contribute to protein kinase C activation in preconditioning. Such potential alternate preconditioning triggers could be produced in greater quantities in the AOPCP-blocked rabbit heart compared with dog heart, thereby replacing and/or bypassing in the rabbit the ecto-5′-nucleotidase mechanism proposed for the dog.

Problem of Timing AOPCP and Adenosine for Preconditioning Studies

Inhibition by AOPCP of preconditioning in dog heart was marked yet incomplete when the drug was given only during preconditioning. To achieve full blunting of infarct protection by preconditioning, AOPCP infusion was required throughout preconditioning, during ischemia, and continuing during reperfusion (16). Olafsson et al. (25) reported infarct size reduction by exogenous adenosine when only infused during reperfusion. In marked contrast, Goto et al. (9) and Vander Heide and Reimer (34) recently failed to reduce infarct size by adenosine administered during reperfusion only. Adenosine and A1-receptor agonists were only effective when they were injected before ischemic insult (6, 33). Such substantial disparities, which may be related to protocol and apparently to animal model or species, are difficult to accept and still await plausible explanations. Nevertheless, the present results argue against a fundamentally important role of ecto-5′-nucleotidase in the production of adenosine for initiating the preconditioning mechanism.


The authors thank Dr. Masao Itoya for technical assistance and Kaoru Kido for technical advice and maintenance of HPLC.


  • Address for reprint requests: T. Miura, Second Dept. of Internal Medicine, Sapporo Medical Univ. School of Medicine, South-1 West-16, Chuo-ku, Sapporo, 060-8556 Japan.

  • This study was supported in part by Grants-in-Aid for Research from the Ministry for Education and Culture, Japan (04670547 and 08670812 to T. Miura).

  • This study was presented in part at the 67th Scientific Sessions of the American Heart Association, Dallas, TX, November 1994.


View Abstract