Intramyocardial Injections and Protection Against Myocardial Ischemia
An Attempt to Examine the Cardioprotective Actions of Adenosine
Background Although adenosine has been proposed to be a cardioprotective agent, direct examination of such protection is confounded by its short half-life and hemodynamic effects. We attempted to avoid these problems by injecting adenosine directly into cardiac muscle.
Methods and Results We gave four adenosine injections (each 0.15 mL, 5 mg·mL−1 saline) into the left ventricular wall of rat hearts before a 60-minute occlusion. Although infarcts were smaller in adenosine-treated hearts (29±6%) than in controls (52±5%; P<.05), injection of saline also reduced infarct size (29±7%). Infarcts in hearts subjected to needle insertion but no fluid injection differed neither from control nor from fluid-treated hearts (38±4%). Adenosine reduced ectopic beats and the incidence of ventricular tachycardia during occlusion. In contrast, saline injection prolonged the duration of arrhythmias. To examine the spatial relationship between protection and the injection site, we gave 18 saline injections (each 0.15 mL) into canine myocardium before a 60-minute occlusion. Infarcts were smaller in saline-treated hearts than in controls (P<.01). Because infarcts in four hearts occupied <3% of the risk region, we concluded that fluid injection did not itself cause appreciable necrosis and speculated that muscle was protected in the vicinity of the injection site. Previous work indicated that muscle can be protected by stretch. We examined this hypothesis by adding gadolinium chloride (a stretch-activated channel blocker) to the saline (0.008 g·mL−1) injection in rat hearts. We again found small infarcts after saline injection (26±5%); however, gadolinium blocked protection (50±7%; P<.03).
Conclusions Although we were only partially successful in documenting adenosine-mediated cardioprotection, we found evidence for myocyte protection via a stretch-activated mechanism.
Adenosine has been proposed to be a crucial mediator of cardiac protection against ischemia and has been implicated to be responsible for the protection provided by ischemic preconditioning. This suggestion was based on studies in which pharmacological blockade of adenosine receptors was associated with the loss of protection provided by ischemic preconditioning.1 In addition, intravenous administration of adenosine before ischemia was found to protect rabbit hearts from ischemia,2 and intracoronary delivery was found to both attenuate myocardial stunning3 and reduce infarct size in dogs.4 In contrast, other in vivo studies have failed to demonstrate protection either by intravenous adenosine infusion in rats or rabbits5 6 or by intracoronary infusion in rabbits.7
One confounding factor in all of the adenosine infusion studies is its very short half-life. In human plasma, the half-life is ≈1.5 seconds, but in dogs it may be as long as 3 minutes.8 We speculated that such rapid removal of adenosine would severely limit the amount able to reach the proposed site of action for cardiac protection, the A1-adenosine receptors on the myocytes. Our hypothesis was that to test the potential cardioprotective effect of adenosine, it would be necessary to deliver adenosine directly to the proposed site of action. We aimed to achieve this objective by injecting adenosine directly into the heart muscle. Unexpectedly, however, we found that “placebo” injection of saline alone into the myocardium reduced infarct size. As a result, we expanded our initial objectives and sought to determine the mechanism of this saline-mediated protection, whether the injection site corresponded to the necrotic or the protected tissue, and whether the protection could be duplicated in more than one species.
The protocols used in this study were approved by the hospital’s Animal Care and Use Committee and conform to the principles of the American Physiological Society. The Heart Institute, Good Samaritan Hospital is accredited by the American Association for Accreditation of Laboratory Animal Care.
Fifty-nine female, retired-breeder Sprague-Dawley rats (body mass range, 310 to 485 g) were anesthetized with ketamine (≈100 mg/kg IM) and xylazine (≈40 mg/kg IM). Additional anesthetic was administered in the same concentrations as given above as required throughout the experiment. We performed a tracheostomy and ventilated the lungs with room air. Catheters were inserted into the left femoral artery and vein to measure blood pressure and to administer fluids. A thoracotomy was performed via the fourth intercostal space to expose the basal region of the heart. A stitch was taken through the myocardium with a C-1 taper needle and 5-0 polypropylene suture from the atrioventricular groove to the region of the pulmonary cone to allow occlusion of the left coronary artery. We tied additional sutures to each arm of the stitch suture to permit release of the occlusion knot.9
Rats were randomized to four groups: (1) control (C), (2) adenosine (A), (3) saline (S), and (4) needle (N). In the control group, there was no intervention and the hearts were subjected to 1 hour of coronary artery occlusion followed by 5 hours of reperfusion. In the adenosine group, we injected adenosine (5 mg·mL−1 dissolved in saline) directly into the region of the left ventricular wall that would become ischemic when the left coronary artery was occluded. We gave four injections, 0.15 mL each, at 1-minute intervals with a 27-gauge needle. In the saline group, four saline injections (0.15 mL each) were given in the same manner as for the adenosine group. In the needle group, no fluid was injected, but the needle was inserted into the myocardium according to the above regimen. Two minutes after the last injection or needle stick, we occluded the left coronary artery by tying a knot in the suture that had been stitched through the myocardium.
Both arterial blood pressure and lead I of the ECG were monitored throughout the experiment and recorded on a chart recorder. The paper speed of the chart recorder was set at 5 mm·s−1 during both the injections and the first 35 minutes of coronary artery occlusion so that ectopic beats could be easily identified. For each rat, we recorded the time of the first and last ectopic beats, from which we calculated the duration of cardiac arrhythmias. The total number of ectopic beats and their temporal distribution were determined. In addition, we recorded the incidence of ventricular tachycardia (defined here as a run of at least four successive ectopic beats) and ventricular fibrillation.
After 60 minutes of occlusion, the knot was released and the artery reperfused. The incidence and number of ectopic beats during the first 3 minutes of reperfusion were recorded. After 5 hours of reperfusion, the coronary artery was briefly reoccluded, and ≈0.5 mL of blue pigment (Unisperse Blue, Ciba-Geigy Corp) was injected into the circulation via the left femoral vein. The area not perfused by the pigment represents the tissue supplied by the occluded vessel and is referred to as the AR. Approximately 15 seconds after the pigment injection, with the animal under deep anesthesia, 3 mL of a saturated solution of potassium chloride was injected into the heart to induce cardiac arrest. The hearts were cut into 5 or 6 slices parallel to the atrioventricular groove and photographed. The slices were then immediately incubated in a 1% solution of TTC at 37°C for 15 minutes and rephotographed. TTC stains viable muscle red, whereas necrotic muscle does not stain and so appears pale. We used planimetry to determine the size of both the AR and the AN, and then from the masses of the heart slices, we calculated the mass of the AR (expressed as a percentage of the total left ventricular mass) and the mass of necrosis (expressed as a percentage of the mass of the AR). Rats with an AR that occupied <20% of the left ventricle were excluded from further analysis.
The purpose of this protocol was to determine the spatial relationship between the injection sites and muscle necrosis and also to determine whether the protective effect of fluid injections could be duplicated in another species. Nine female dogs (15 to 26 kg) were anesthetized with sodium pentobarbital (30 mg/kg), intubated, and ventilated with room air. After the left jugular vein (for administration of fluids and supplemental anesthesia) and the left carotid artery (for measurement of heart rate and arterial pressure) were cannulated, the heart was exposed through a left lateral thoracotomy and suspended in a pericardial cradle. A fluid-filled catheter was positioned in the left atrium for later injection of radiolabeled microspheres (141Ce, 103Ru, or 95Nb) for measurement of RMBF. A segment of the LAD was isolated, usually distal to its first major diagonal branch, for later placement of occlusive vascular clamps.
After baseline hemodynamic measurements had been obtained, all dogs received 18 intramyocardial injections of saline (0.15 mL each) into the soon-to-be-ischemic LAD territory with a 21-gauge needle. The needle was marked 2 cm from its tip so that the injections could be made at approximately the same depth in the tissue and so that the injection sites would be close to the tissue likely to become necrotic (that is, the subendocardium). The injections were made over a period of ≈5 minutes throughout the region that we expected to become ischemic and were spaced as evenly as possible. Five minutes after the last injection and after prophylactic administration of lidocaine (0.15 mg/kg to control the incidence of lethal arrhythmias), all dogs underwent 1 hour of LAD occlusion followed by 4 hours of reperfusion.
Heart rate and arterial pressure were measured at baseline, after the intramyocardial injections, throughout LAD occlusion, and after reperfusion. In addition, the severity of ischemia was assessed in all dogs by measurement of RMBF at 30 minutes into the LAD occlusion.
At the end of the protocol, the LAD was ligated at the site of the previous occlusion, and Unisperse Blue pigment (0.25 to 0.5 mL/kg) was injected into the coronary vasculature via the left atrial catheter to delineate the in vivo extent of the AR. Under deep anesthesia, cardiac arrest was induced by intracardiac injection of potassium chloride solution. The hearts were rapidly excised, cut into 5 to 7 transverse slices, and photographed for later measurement of AR. The heart slices were incubated for 10 minutes in TTC at 37°C to distinguish necrotic from viable myocardium, rephotographed for later calculation of AN, and stored in formalin.
After fixation, right ventricular tissue was trimmed from each heart slice, and the remaining left ventricular tissue was weighed. Photographic images of the heart slices were projected and traced at magnifications of approximately ×2 to ×4. The extent of AR and AN in each heart slice was quantified by computerized planimetry, corrected for the mass of the tissue slice, and summed for each heart.
After we had measured the mass of the slices, tissue blocks were cut from the center of the previously ischemic bed and from the remote, normally perfused circumflex bed and divided into subendocardial, midmyocardial, and subepicardial segments. RMBF was then quantified.10
Previous experiments from our laboratory demonstrated that saline infusion, via the left atrium, dilated the heart and protected it against subsequent coronary artery occlusion.11 We speculated that the protection provided by the fluid injection might result from a similar mechanism, that is, local tissue stretch at the injection site. Furthermore, since gadolinium (a blocker of stretch-activated ion channels) prevented the protective effect of the saline infusion, our hypothesis was that it might do the same for the fluid injections. We tested this hypothesis in eight rats (body mass range, 295 to 390 g). All rats received saline injections as in protocol 1; however, we dissolved gadolinium chloride (0.008 g/mL) in the saline injected into four of the hearts. The rest of the procedure was the same as used in protocol 1. In addition to AR and infarct size measurement, we also measured the number, distribution, and time course of ectopic beats.
Protocol 1. Differences between groups were compared by one-way ANOVA followed by pairwise comparisons according to Tukey’s method. The effect of the intramyocardial injections on heart rate and blood pressure in each group were examined by paired t tests, and differences in rate-pressure product between groups were assessed by a two-factor repeated-measures ANOVA. We used Fisher’s exact test, with corrections made for multiple comparisons, to compare the incidence of ventricular tachycardia, ventricular fibrillation, and reperfusion arrhythmias between groups.
Protocol 2. Infarct sizes in dogs receiving intramyocardial saline injections were compared with data obtained from a complete cohort of nine control animals subjected to a comparable 1-hour episode of LAD occlusion.12 Specifically, ANCOVA (with mean transmural blood flow measured during occlusion as the covariate) was used to determine whether the relationship between infarct size and collateral blood flow differed between the treated dogs vs the historical controls.
Protocol 3. We used two-tailed t tests to compare the means of the two groups. All values are given as mean±SEM unless stated otherwise. In all cases, values were considered to be significantly different if the probability was <.05.
Mortality and exclusions. Four rats died during occlusion (2 controls, 1 saline, and 1 adenosine-treated). Another 5 rats were excluded because the AR occupied <20% of the left ventricle, and 2 were excluded as technical failures. We excluded 8 rats (7 controls and 1 saline-treated) because the occlusion knot had loosened and, if tightened, could have resulted in ischemic preconditioning. We made the decision to exclude these animals before we saw the infarct. Each group contained 10 rats in the final analysis.
AR and infarct size. There was no difference in AR between the groups (C, 53±3%; A, 42±5%; S, 51±5%; and N, 53±4%). In contrast, infarct size (expressed as a percentage of the AR) in both the adenosine (29±6%) and saline (29±7%) groups was smaller than that in the hearts subjected to occlusion alone (52±5%, P<.05; Fig 1⇓). In addition, infarcts in the adenosine and saline groups were patchy, with areas of necrosis separated by areas of viable tissue (Fig 2⇓). Although infarct size in the needle group (38±4%) was less than in the control group, the difference was not significant. However, infarcts in needle-treated hearts did not differ significantly from those in the adenosine and saline groups.
Hemodynamics. Both the fluid injections and the needle sticks resulted in immediate decreases in arterial pressure (Fig 3⇓). Although mean arterial pressure in the adenosine group recovered to baseline values before occlusion, the pressure in both the saline and needle-treated groups was lower than baseline just before occlusion (S, 86±4 versus 72±4 mm Hg, P<.03; N, 104±7 versus 90±6 mm Hg, P<.02). Fluid injections also resulted in decreases in heart rate (A, 140±12 versus 83±11 bpm, P<.005; S, 143±12 versus 106±9 bpm, P<.02). However, there was no statistically significant difference in the rate-pressure product between the groups at any time point (Table 1⇓). In addition, we found no correlation between the product of heart rate and mean arterial pressure (at any time point) and infarct size (data not given). Thus, the differences in infarct size were not a consequence of differences in heart rate or blood pressure.
Arrhythmias. The time course of ectopic events is shown in Fig 4⇓. Adenosine treatment reduced the average duration of the period of arrhythmias versus both the saline (P<.01) and needle (P<.05) groups. Arrhythmias in the saline-treated hearts lasted significantly longer than in all of the other groups (P<.05 versus control and needle, P<.001 versus adenosine). In addition, adenosine significantly reduced the average total number of ectopic beats versus both the control and saline groups (P<.01, Fig 5⇓). All of these changes resulted in differences in the shape of the temporal distribution of ectopic beats (Fig 6⇓). The control group exhibited a peak in the distribution at 10 to 15 minutes after occlusion. The distribution of ectopic beats in the needle group was similar to that of the controls, but with lower absolute values. In the saline group, the distribution peak was shifted to the period 15 to 20 minutes after occlusion and, as reported above, the distribution extended for a longer period of time than in the other groups. The small numbers of ectopic beats observed in the adenosine-treated hearts were relatively evenly distributed. Fig 7⇓ shows the incidence of ventricular tachycardia, ventricular fibrillation, and reperfusion arrhythmias in each group. Adenosine reduced the incidence of ventricular tachycardia versus the other three groups (P<.05). Reperfusion arrhythmias were most frequent in the saline group.
Saline injections did not affect either blood pressure or heart rate (data not given). Two dogs died as a result of ventricular fibrillation: one at 55 minutes after occlusion and the other at reperfusion. The remaining saline-treated and control hearts were equally ischemic during LAD occlusion; the mean transmural blood flow was 0.12±0.04 and 0.18±0.04 mL·min−1·g−1, respectively. The relationship between infarct size and blood flow is shown in Fig 8⇓; for any given blood flow, infarct size in the saline-treated hearts was less than in the control group (P<.01). Four of the saline-treated hearts had infarcts that occupied <3% of the AR. This observation indicates that the injections or injection sites were not themselves associated with appreciable necrosis. Thus, intramyocardial saline injections also elicited a protective effect in canine hearts.
AR and infarct size. There was no difference in AR between the saline (52±8%) and gadolinium (44±3%) groups. In contrast, infarct size was significantly higher in the gadolinium-treated (50±7%) than in the saline-treated hearts (26±5%, P<.03; Fig 9⇓). Infarct size in the saline group was comparable to that found with saline treatment in protocol 1 (29±7%), whereas infarct size in the gadolinium-treated hearts was similar to that found in hearts subjected to occlusion alone in protocol 1 (52±5%). Thus, the protective effect of the saline injection was blocked by the addition of gadolinium chloride to the solution.
Hemodynamics. As in protocol 1, the fluid injections caused immediate reductions in blood pressure (mean reduction in systolic pressure: saline, 15 mm Hg; saline+gadolinium, 13 mm Hg); however, in both groups, there was no statistically significant difference between baseline and preocclusion values (Table 2⇓). Similarly, heart rate before occlusion was not significantly different from baseline (Table 2⇓).
Arrhythmias. There was no difference in the total number of ectopic beats between the two groups. The time course of arrhythmias seen in the saline-treated hearts was similar to that found in the corresponding group in protocol 1, and the time course in the hearts treated with gadolinium was similar to that seen in the control group of protocol 1 (Fig 10⇓).
We found that injection of fluid into both rat and dog hearts before ischemia reduced the amount of necrosis produced by 60 minutes of coronary artery occlusion. This protective effect of saline injection in rat hearts could be prevented by the addition of gadolinium chloride, implicating stretch-activated ion channels as a potential mechanism of action. Although saline and adenosine injections reduced infarct size to the same extent in rat hearts, their effects on ischemia-induced arrhythmias were divergent. Adenosine injection significantly reduced the number of ectopic beats and the incidence of ventricular tachycardia, whereas saline injection extended the duration of ectopy without increasing the total number of ectopic beats.
Infarcts in rat hearts that received fluid injections were not only smaller than those found in the control hearts but also had a different appearance. In contrast to the homogeneous confluent infarcts in control hearts, fluid injections were characterized by patches of necrotic tissue separated by areas of viable tissue within the AR (Fig 2⇑). Such localized protection indicated that either the muscle salvage was confined to the tissue around the injection site or the injection caused necrosis but protected tissue distant from the site. We attempted to resolve these possibilities by microscopic examination of histological sections. However, the combination of a small number of injections given with a small needle into a small rat heart prevented us from accurately relating the histological location of the injection site to the photographs of the TTC-stained slices. We therefore tried a similar experiment in canine hearts but gave more saline injections (18 versus 4) and used a larger-diameter needle (21- versus 27-gauge). Again, saline injection reduced infarct size compared with historical controls. Although we know that the insertion of a hypodermic syringe needle into the myocardium does result in a small amount of muscle necrosis immediately adjacent to the insertion site,13 necrotic tissue in four of the fluid-treated dog hearts occupied <3% of the AR, a finding inconsistent with the concept that the injections caused appreciable cell death. Thus, we propose that the muscle protected by the fluid injections was not confined to the vicinity of the injection site.
Mechanism of Infarct Size Reduction
There are several possible mechanisms to explain the muscle protection observed after fluid injection: (1) local ischemia, (2) stimulated release of a protective substance, (3) a temperature-mediated effect, and (4) local stretching of muscle.
Ischemia. It is known that short periods of ischemia before a longer period of ischemia will reduce the amount of necrosis. It is possible that the injections were similar to this phenomenon of so-called ischemic preconditioning because the injection of fluid may have caused local ischemia by compression of blood vessels. In fact, we did observe some blanching of the tissue at the injection site immediately after the injection; however, this was not accompanied by any changes in the ECG. Nevertheless, the injections could have “ischemically preconditioned” small regions of the heart, protected them against the subsequent 60 minutes of coronary occlusion, and hence produced the “patchy” infarcts.
Stimulated release of a protective substance. The biochemical and metabolic response to the injections is unknown, and so it is possible that the injections stimulated the production of one or more substances, which in turn resulted in cardioprotection. For example, the injections almost certainly triggered release of catecholamines, agents that have been suggested to play a role in myocardial protection.14 In addition, Van Wylen et al15 demonstrated that the insertion of a microdialysis probe (≈300 μm in diameter) into a canine heart resulted in a rapid and substantial increase in adenosine. This elevation in interstitial adenosine concentration was as high as that found after 10 minutes of coronary artery occlusion and persisted for as long as 60 minutes. Thus, it appears likely that the insertion of a needle into the heart will produce a similar increase in adenosine. Alternatively, the trauma associated with the injection might serve as a stimulus, perhaps activating preexisting cardiac heat shock proteins, which are known to be capable of exerting cardioprotective effects.16
Temperature changes. Consideration of heat shock proteins leads us to the possible effect of injection-mediated temperature changes. This issue is relevant because infarct size in rabbit hearts has been found to correlate with right atrial temperature over the range 35°C to 42°C; the lower the temperature, the smaller the infarct.17 In addition, in isolated hearts, a modest 4°C reduction in temperature protected against reperfusion arrhythmias and modulated calcium homeostasis.18
To examine this issue, we placed a thermocouple into the myocardium of two rats and then made saline injections into a region immediately adjacent to the thermocouple. We chose to inject more fluid (0.2 mL) to increase the likelihood that the amount of fluid at the thermocouple was comparable to that at the center of the injection site in the infarct-size studies. In a total of seven injections (the temperature of the saline was 23°C), the minimum cardiac temperature was between 25°C and 28°C. In each case, the temperature returned to baseline within 30 seconds, and there was no cumulative effect of multiple injections. We conclude that the effect of the fluid injections on cardiac temperature was transient, which is in contrast to the previous studies in which the reduced temperature was maintained for 10 minutes18 or during the entire 4-hour protocol.17 However, we cannot rule out the possibility that even such transient changes protect via the same mechanism(s) as more prolonged periods of hypothermia. It is interesting to note that because coronary artery occlusion will reduce the temperature of the ischemic tissue (Warner et al19 found a statistically significant drop of ≈0.4°C after occlusion in dog hearts), the repetitive reductions in temperature caused by fluid injections will mimic, to some degree, the temperature changes that occur during the periods of ischemia required to elicit ischemic preconditioning. It is therefore possible that repetitive hypothermia may be a trigger for cardioprotection.
Local stretch of muscle. We had proposed, and subsequently demonstrated, that ischemic preconditioning can protect tissue distant from that subjected to the preconditioning regimen.20 21 We also proposed that myocardial stretch was a potential mechanism for triggering such distant protection. This hypothesis was subsequently supported by a study in which myocardium was found to be protected against ischemia by dilation of the heart with a saline infusion.11 Furthermore, this stretch-mediated protection was prevented by gadolinium chloride, a blocker of stretch-activated ion channels.22 23 It also appears that there is a link between stretch and heat shock protein production. Myocardial stretch, induced by either brief aortic cross-clamp, insertion of a cannula through the apex of the heart, or intraventricular balloon, has been shown to result in a rapid increase in mRNA for heat shock protein 70 and also increased amounts of the protein itself within 60 minutes.24
In the present study, we speculated that fluid injections might result in localized myocardial stretching in the vicinity of the injection site. To test this hypothesis, we included gadolinium in the injection solution in protocol 3. Saline injections that included gadolinium no longer reduced infarct size. In fact, infarct size in hearts that received gadolinium was almost identical to those in the control group in protocol 1. Thus, although we cannot rule out a potential role for the protective mechanisms 1 through 3 discussed above, our data are consistent with our previous observations that myocyte stretch could be an important mediator of cardiac protection.
Adenosine and Infarct Size Reduction
Most of the evidence supporting the cardioprotective effect of adenosine has been indirect. For example, agents that block adenosine receptors have been shown to block the protective effects of ischemic preconditioning.25 The indirect nature of such studies does not conclusively prove the involvement of adenosine, because it is possible that the effects of the blockers extended beyond the A1 receptors. Attempts to directly assess the protection by adenosine are complicated by its hemodynamic effects. Several in vivo studies have failed to demonstrate any protective effect of intravenously administered adenosine in either rabbits6 or rats,5 and even intracoronary administration of adenosine did not reduce infarct size in rabbits.7 These results are probably not surprising, because the half-life of adenosine in blood is so short that little, if any, adenosine would reach the A1 receptors. In fact, it has been found that the adenosine concentration must exceed 10−6 mol/L to cross the barrier provided by the endothelial cells.26 Nevertheless, some studies have reported reduced infarct size with intravenous2 and intracoronary adenosine.4
The effect of intramyocardial adenosine injections initially on heart rate and subsequently on arrhythmias provides circumstantial evidence that adenosine did reach the cardiac myocyte A1 receptors. However, because infarct size in saline-treated hearts was the same as that in the adenosine group, any protective effect of adenosine was not additive to that of saline-induced protection. This could indicate that adenosine provides no protection against ischemia-induced muscle necrosis. In support of this concept, Silva et al27 found that although the adenosine deaminase inhibitor pentostatin increased the amount of interstitial adenosine by more than threefold before coronary occlusion in dogs and by as much as 100-fold during occlusion, it failed to reduce infarct size. Lasley et al28 found that rabbits that received an infusion of adenosine (140 μg·kg−1·min−1 IV) before coronary artery occlusion achieved a myocardial interstitial adenosine concentration that was the same as that produced by a 5-minute period of “preconditioning” ischemia. Nevertheless, infarct size assessed after 45 minutes of ischemia was significantly lower in the “ischemically preconditioned” group than in the adenosine-treated group (adenosine-treated hearts did have smaller infarcts than the control group). Although by no means definitive, such results raise questions about the purported protective actions of adenosine in ischemia and ischemic preconditioning. An additional possibility is that adenosine and stretch protect via the same signal transduction mechanism, and so addition of adenosine to the saline injection did not add to the reduction of infarct size.
Another potentially important consideration is the duration of ischemia used, because it is known that prolonged ischemia can overwhelm even the powerful protective effect of ischemic preconditioning. In a series of pilot experiments, we found that adenosine injections before a 90-minute occlusion resulted in an infarct size of 52±10% of the AR. Because this value was only slightly less than the 60% to 70% infarcts typically obtained in control rat hearts subjected to 90 minutes of occlusion, we speculated that a shorter ischemic period might provide greater resolution of the treatment effect that we thought was apparent in our pilot experiments. This speculation proved to be correct. Thus, we conclude that the protective effects of fluid injection decrease as the duration of occlusion increases. It is therefore possible that an even shorter duration of occlusion might allow resolution of separate contributions to cardiac protection by adenosine and fluid injection; however, this possibility has yet to be examined.
Our original aim was to test the hypothesis that adenosine reduced infarct size; however, adenosine also has the potential to exert a favorable effect on cardiac arrhythmias. Indeed, adenosine injections significantly reduced the number of ectopic beats and the incidence of ventricular tachycardia versus control hearts, protection that was not achieved by intravenous delivery in rats.5 In contrast, hearts that received saline injections had a number of ectopic beats and an incidence of ventricular tachycardia similar to those of control hearts even though they also had smaller infarcts. In fact, injection of saline prolonged the time course of ectopy during coronary occlusion, perhaps reflecting the patchy and inhomogeneous infarcts in this group.
The protective effect of adenosine against arrhythmias is known, and similar data obtained in rats have been reported previously.29 However, this previous study gave a continuous intravenous infusion of adenosine starting before coronary artery occlusion and continuing throughout the ischemic period. In our study, we essentially gave four bolus doses of adenosine between 2 and 4 minutes before the onset of ischemia. Interestingly, the antiarrhythmic effect of this dosing regimen was long-lived and may have extended to protecting against reperfusion arrhythmias (Fig 7⇑). A similar persistent reduction in ischemia-induced arrhythmias was found after infusion of adenosine into the left ventricular lumen of dogs before coronary occlusion.30
Mechanism of Adenosine Protection Against Arrhythmias
The mechanism of protection by adenosine against arrhythmias in our study is unknown. Reductions in either blood pressure or heart rate might be expected to decrease the severity of arrhythmias31 ; however, these variables do not appear to have been important factors in our study. All three groups receiving injections experienced transient hypotension immediately after each injection, and yet a positive effect on arrhythmias was evident only in the adenosine-treated group. Even though the pressure drop after injection was greatest in the adenosine group, the mean arterial pressure before occlusion was not different from the baseline value. In contrast, the pressure before occlusion was significantly lower than baseline in both saline and needle groups. The ability of adenosine to reduce heart rate has been attributed to activation of A1-adenosine receptors in supraventricular tissue.32 Even though we injected the adenosine into the ventricular tissue, there was a significant decrease in heart rate. This observation suggests that adenosine reached the supraventricular tissue via diffusion through the interstitial space or that there was sufficient entry of adenosine into the coronary circulation. We do not believe that this is the mechanism of protection against arrhythmias, because the negative chronotropic effect of adenosine was not evident 5 minutes after occlusion. Although heart rate was significantly reduced by adenosine injection before occlusion, there was also a reduction in the heart rate of the saline-treated hearts. Furthermore, the heart rate 5 minutes into occlusion was similar in the adenosine (104±8 bpm) and saline (108±11 bpm) groups. Thus, it appears unlikely that the transient changes in blood pressure or heart rate were responsible for the observed differences in arrhythmias.
Adenosine is a potent coronary vasodilator; for example, Wainwright and Parratt30 reported a 200% increase in coronary arterial flow with a ventricular lumen infusion of 10 μg·kg−1·min−1 in dogs. However, the manner in which we administered adenosine would limit the possibility of the drug entering coronary arteries. Even if adenosine did enter arteries within the AR, the lack of a significant collateral circulation in rat hearts means that after occlusion of the left coronary artery, blood flow within the AR would be reduced to zero with or without the presence of adenosine.
Our dismissal of these various “indirect” mechanisms leaves us with a consideration of the direct cardiac effects of adenosine. For example, in ventricular myocytes, adenosine is known to antagonize the stimulatory effect of catecholamines on calcium inward currents,33 which would be consistent with an antiarrhythmic effect. This protection, coupled via guanine nucleotide–binding regulatory proteins, is achieved via inhibition of adenylyl cyclase activity, which in turn prevents an increase in cAMP. We speculate that this mechanism or other signal transduction–mediated mechanisms are more consistent with our data than a mechanism involving altered hemodynamic parameters or increased coronary flow.
Implications and Limitations of the Study
We had initially thought that the intramyocardial injections would be an effective way to test the action of adenosine and other agents, because it would effectively bypass any vascular effect of the drugs. Although we did manage to avoid some of the problems associated with intravenous adenosine administration, we encountered an unexpected complication: the injection itself resulted in myocardial protection. This finding emphasizes that we cannot assume that any intervention, even if seemingly trivial, does not affect the outcome of the experiment; ie, it is not possible to observe a system without perturbing it.
The main shortcoming of our attempt to examine the cardioprotective effect of adenosine is that we did not measure tissue adenosine levels and so do not know how much adenosine we were able to deliver or how long it remained. Although the observed effect of injected adenosine on arrhythmias provides indirect evidence that receptors were affected, the interstitial concentration of adenosine and/or the degree of receptor activation required for infarct size reduction could be higher.
Because we used only mature female animals in our study, we cannot discount the possibility that age and sex may have influenced the results. However, analysis of the effect of 1 hour of coronary artery occlusion in canine hearts determined that sex did not influence acute myocardial ischemia and infarction.34 Furthermore, our laboratory has found a similar lack of sex-related differences in rat hearts.35 How age may affect the results that we obtained is unknown.
It is tempting to label any intervention applied before coronary artery occlusion that subsequently reduces infarct size as either “preconditioning” or a “preconditioning mimetic,” the implicit suggestion being that the interventions act via the same mechanism as “ischemic preconditioning.” There is some evidence to link our apparent stretch-activated protection to the protection provided by ischemic preconditioning; however, we cannot be certain that they share a common mechanism. Although the possible link between the two phenomena merits further consideration and study, additional examination of injection-mediated protection itself is also warranted.
We conclude that because of its profound antiarrhythmic effects, intramyocardial injection of adenosine proved to be an effective means of targeting the A1-adenosine receptors. Unexpectedly, however, intramyocardial injection of saline alone decreased infarct size, with no additive effect of adenosine. We demonstrated this protective effect of intramyocardial fluid injection in both rats and dogs and, because protection was blocked by gadolinium chloride, propose that it is mediated by stretch-activated ion channels.
Selected Abbreviations and Acronyms
|AN||=||area of necrosis|
|AR||=||area at risk of infarction|
|LAD||=||left anterior descending coronary artery|
|RMBF||=||regional myocardial blood flow|
- Received September 11, 1995.
- Revision received December 4, 1995.
- Accepted December 21, 1995.
- Copyright © 1996 by American Heart Association
Downey JM, Liu GS, Thornton JD. Adenosine and the anti-infarct effects of preconditioning. Cardiovasc Res. 1993;27:3-8.
Toombs CF, McGee DS, Johnston WE, Vinten-Johansen J. Myocardial protective effects of adenosine: infarct size reduction with pretreatment and continued receptor stimulation during ischemia. Circulation. 1992;86:986-994.
Yao Z, Gross GJ. A comparison of adenosine-induced cardioprotection and ischemic preconditioning in dogs: efficacy, time course, and role of KATP channels. Circulation. 1994;89:1229-1236.
Li Y, Kloner RA. The cardioprotective effects of ischemic ‘preconditioning’ are not mediated by adenosine receptors in rat hearts. Circulation. 1993;87:1642-1648.
Liu GS, Thornton J, Van Winkle DM, Stanley AWH, Olsson RA, Downey JM. Protection against infarction afforded by preconditioning is mediated by A1 adenosine receptors in rabbit heart. Circulation. 1991;84:350-356.
Hale SL, Bellows SD, Hammerman H, Kloner RA. An adenosine A1 receptor agonist, R(-)-N-(2-phenylisopropyl)-adenosine (PIA), but not adenosine itself, acts as a therapeutic preconditioning-mimetic agent in rabbits. Cardiovasc Res. 1993;27:2140-2145.
Schrader J. Formation and metabolism of adenosine and adenine nucleotides in cardiac tissue. In: Phillis JW, ed. Adenosine and Adenine Nucleotides as Regulators of Cellular Function. Boca Raton, Fla: CRC Press; 1991:55-69.
Himori N, Matsuura A. A simple technique for occlusion and reperfusion of coronary artery in conscious rats. Am J Physiol. 1989;256:H1719-H1725.
Domenech RJ, Hoffman JE, Noble MM, Saunders KB, Henson RJ, Subyanto S. Total and regional coronary blood flow measured by radioactive microspheres in conscious and anesthetized dogs. Circ Res. 1969;23:581-596.
Ovize M, Przyklenk K, Kloner RA. Stretch preconditions the canine myocardium. Am J Physiol. 1994;266:H137-H146.
Przyklenk K, Sussman MA, Simkhovich BZ, Kloner RA. Does ischemic preconditioning trigger translocation of protein kinase C in the canine model? Circulation. 1995;92:1546-1557.
Whittaker P, Rakusan K, Kloner RA. Transmural channels can protect ischemic tissue: assessment of long-term myocardial response to laser- and needle-made channels. Circulation. 1996;93:143-152.
Bankwala Z, Hale SL, Kloner RA. α-Adrenoceptor stimulation with exogenous norepinephrine or release of endogenous catecholamines mimics ischemic preconditioning. Circulation. 1994;90:1023-1028.
Van Wylen DGL, Willis J, Sodhi J, Weiss RJ, Lasley RD, Mentzer RM Jr. Cardiac microdialysis to estimate interstitial adenosine and coronary blood flow. Am J Physiol. 1990;258:H1642-H1649.
Chien GL, Wolff RA, Davis RF, Van Winkle DM. ‘Normothermic range’ temperature affects myocardial infarct size. Cardiovasc Res. 1994;28:1014-1017.
Gambassi G, Cerbai E, Pahor M, Capogrossi MC, Carbonin P, Mugelli A. Temperature modulates calcium homeostasis and ventricular arrhythmias in myocardial preparations. Cardiovasc Res. 1994;28:391-399.
Warner KG, Khuri SF, Marston W, Sharma S, Butler MD, Assousa SN, Saad AJ, Siouffi ST, Lavin PT. Significance of the transmural diminution in regional hydrogen ion production after repeated coronary artery occlusions. Circ Res. 1989;64:616-628.
Przyklenk K, Bauer B, Ovize M, Kloner RA, Whittaker P. Regional ischemic ‘preconditioning’ protects remote virgin myocardium from subsequent sustained coronary artery occlusion. Circulation. 1993;87:893-899.
Yang XC, Sachs F. Block of stretch-activated ion channels in xenopus oocytes by gadolinium and calcium ions. Science. 1989;243:1068-1071.
Hansen DE, Borganelli M, Stacy GP Jr, Taylor LK. Dose-dependent inhibition of stretch-induced arrhythmias by gadolinium in isolated canine ventricle: evidence for a unique mode of antiarrhythmic action. Circ Res. 1991;69:820-831.
Knowlton AA, Eberli FR, Brecher P, Romo GM, Owen A, Apstein CS. A single myocardial stretch or decreased systolic fiber shortening stimulates the expression of heat shock protein 70 in the isolated, erythrocyte-perfused rabbit heart. J Clin Invest. 1991;88:2018-2025.
Downey JM, Liu Y, Ytrehus K. Adenosine and the antiinfarct effects of preconditioning. In: Przyklenk K, Kloner RA, Yellon DM, eds. Ischemic Preconditioning: The Concept of Endogenous Protection. Boston, Mass: Kluwer Academic Publishers; 1993:137-152.
Lasley RD, Konyn PJ, Hegge JO, Mentzer RM Jr. Effects of ischemic and adenosine preconditioning on interstitial fluid adenosine and myocardial infarct size. Am J Physiol. 1995;269:H1460-H1466.
Boachie-Ansah G, Kane KA, Parratt JR. Is adenosine an endogenous myocardial protective (antiarrhythmic) substance under conditions of ischemia? Cardiovasc Res. 1993;27:77-83.
Belardinelli L, Shryock JC, Song Y, Wang D, Srinivas M. Ionic basis of the electrophysiological actions of adenosine on cardiomyocytes. FASEB J. 1995;9:359-365.
Lerman BB, Belardinelli L. Cardiac electrophysiology of adenosine: basic and clinical concepts. Circulation. 1991;83:1499-1509.