Heart and Circulatory Physiology

Myocardium tolerant to an adenosine-dependent ischemic preconditioning stimulus can still be protected by stimuli that employ alternative signaling pathways

David A. Liem, Maaike te Lintel Hekkert, Olivier C. Manintveld, Frans Boomsma, Pieter D. Verdouw, Dirk J. Duncker


Clinical studies on cardioprotection by preinfarct angina are ambiguous, which may involve development of tolerance to repeated episodes of ischemia. Not all preconditioning stimuli use identical signaling pathways, and because patients likely experience varying numbers of episodes of preinfarct angina of different degrees and durations, it is important to know whether myocardium tolerant to a particular preconditioning stimulus can still be protected by stimuli employing alternative signaling pathways. We tested the hypothesis that development of tolerance to a particular stimulus does not affect cardioprotection by stimuli that employ different signaling pathways. Anesthetized rats underwent classical, remote or pharmacological preconditioning. Infarct size (IS), produced by a 60-min coronary artery occlusion (CAO), was determined after 120 min of reperfusion. Preconditioning by two 15-min periods of CAO (2CAO15, an adenosine-dependent stimulus) limited IS from 69 ± 2% to 37 ± 6%, but when 2CAO15 was preceded by 4CAO15, protection by 2CAO15 was absent (IS = 68 ± 1%). This development of tolerance coincided with a loss of cardiac interstitial adenosine release, whereas two 15-min infusions of adenosine (200 μg/min iv) still elicited cardioprotection (IS = 40 ± 4%). Furthermore, cardioprotection was produced when 4CAO15 was followed by the adenosine-independent stimulus 3CAO3 (IS = 50 ± 8%) or the remote preconditioning stimulus of two 15-min periods of mesenteric artery occlusion (IS = 49 ± 6%). In conclusion, development of tolerance to cardioprotection by an adenosine-dependent preconditioning stimulus still allows protection by pharmacological or ischemic stimuli intervention employing different signaling pathways.

  • infarct size
  • remote preconditioning

ischemic preconditioning (IPC) is the most powerful means of endogenous cardioprotection against irreversible cell injury in the experimental animal (26, 31). However, clinical studies on infarct size (IS) limitation by brief anginal episodes preceding acute myocardial infarction are ambiguous (3, 4, 16, 27, 28, 44); such ambiguity has been attributed to a loss of cardioprotection by ischemic preconditioning in the aging (1, 2, 3, 19) or pathological (9, 12, 15, 18) heart. Another confounding factor could be development of tolerance to IPC, i.e., the loss of cardioprotection when the same preconditioning stimulus is repetitively applied (6, 14, 32). For example, Cohen et al. (6) demonstrated that in rabbits the cardioprotection produced by a single 5-min coronary artery occlusion (CAO) followed by 10 min of reperfusion (1CAO5) was lost when the 5-min CAO stimulus was applied at 30-min intervals for 8 h during 3 days.

In recent years, it has become apparent that not all preconditioning stimuli employ the same signaling pathway to exert their cardioprotective action (7, 10, 23, 24, 34). For instance, in the rat, cardioprotection by a single 15-min CAO followed by 10 min of reperfusion (1CAO15) is adenosine dependent but does not involve reactive oxygen species (ROS), whereas cardioprotection by three cycles of 3 min of CAO interspersed by 5 min of reperfusion (3CAO3) depends on ROS (24) but does not involve adenosine (22, 23). The major aim of the present study was therefore to investigate whether tolerance that develops when the same IPC stimulus is applied repetitively also implies tolerance to a stimulus that employs a different signal transduction pathway. Hence, in the first part of the study, we investigated whether tolerance to a particular (adenosine-dependent) preconditioning stimulus also affects cardioprotection by a stimulus that employs an alternative (adenosine-independent) pathway. Myocardium can be preconditioned by local myocardial ischemia, as well as by brief ischemia in noncardiac tissue such as the small intestine, kidneys, and skeletal muscle (5, 11, 25, 30), which, at least for the small intestine, involves a neurogenic pathway (11, 25). Hence, in the second part of the study, we investigated whether cardioprotection by remote preconditioning via a 15-min mesenteric artery occlusion (MAO15) is affected by the development of tolerance to a classical IPC stimulus.

Because tolerance to IPC has not been investigated in the rat, we first established a model for the development of tolerance on the basis of our experience with the adenosine-dependent stimulus 1CAO15 in this species. Capitalizing on the observations by Vogt et al. (41), who showed in pigs that progressive loss of adenosine production rendered myocardium tolerant to protection by 10-min CAO but still responsive to exogenous adenosine, we also investigated whether loss of adenosine release also contributes to development of tolerance in the rat heart and whether exogenous adenosine still induces protection once tolerance has developed.



Experiments were performed in ad libitum-fed male Wistar rats (300–380 g) in accordance with the Guide for the Care and Use of Laboratory Animals (NIH Publication No. 86-23, revised 1996) and with approval of the Erasmus University Rotterdam Animal Care Committee.

Surgical and Experimental Procedures

Pentobarbital sodium-anesthetized (60 mg/kg ip) rats were intubated for positive-pressure ventilation with oxygen-enriched room air. Through the carotid artery, a PE-50 catheter was positioned in the thoracic aorta for measurement of arterial blood pressure and heart rate (11, 40). In the inferior caval vein, a PE-50 catheter was placed for infusion of Haemaccel (Hoechst) to compensate for blood loss during surgery and to maintain central venous pressure during the experimental protocol and for drug infusion during the experiments. After thoracotomy, via the left third intercostal space, the pericardium was opened, and a silk 6-0 suture was looped under the left coronary artery for later CAO. A catheter was positioned in the abdominal cavity to allow intraperitoneal administration of pentobarbital sodium for maintenance of anesthesia. Rectal temperature was continuously measured and maintained at 36.5–37.5°C (11, 40). After completion of surgery, a 30-min stabilization period was allowed before experimental protocols were carried out. Rats that fibrillated were allowed to complete the protocol, provided that conversion to normal sinus rhythm occurred spontaneously within 1 min or that defibrillation via gentle thumping on the thorax was successful within 2 min after onset of fibrillation. Occlusion and reperfusion were visually verified. In 13 additional rats, a microdialysis probe (model CMA/20, Carnegie Medicine, Stockholm, Sweden; 4 × 0.5 mm membrane, 20-kDa cutoff) was implanted into the myocardial area at risk (AR) to determine myocardial interstitial adenosine levels (17). Samples were collected during each 15-min CAO at a rate of 2 μl/min. At the conclusion of each experiment, probe recovery was determined ex vivo with a stock solution containing 100 μM adenosine and found to be 14 ± 1% (percentage of adenosine concentration in the stock solution recovered in the probe samples). All samples were stored at −50°C for later analysis. The adenosine concentrations in dialysate samples were determined by reverse-phase high-performance liquid chromatography (36).

IS Analysis

IS was determined as previously described (11, 40). Briefly, after 120 min of reperfusion, the left coronary artery was reoccluded, and 10 ml of trypan blue (0.4%; Sigma Chemical) were immediately infused intravenously into the femoral vein to stain the normally perfused myocardium dark blue and delineate the nonstained AR. Subsequently, hearts were excised, rinsed in cold saline, and cut into 2-mm-thick slices from apex to base. From each slice, the right ventricle was removed and the left ventricular AR (nonstained) was dissected from the remaining left ventricular tissue. The AR was then incubated for 10 min in 37°C nitro blue tetrazolium (Sigma Chemical; 1 mg/ml Sorensen buffer, pH 7.4), which stains viable tissue purple but leaves infarcted tissue unstained. After the infarcted area (IA) was isolated from the non-IA, the different areas of the left ventricle were dried and weighed separately. Myocardial IS was computed as IA expressed as a percentage of AR (11, 40).

Experimental Design

Rat hearts were preconditioned with one or multiple 15-min CAOs separated by 15 min of reperfusion (nCAO15, adenosine-dependent IPC stimuli), a sequence of three 3-min CAOs interspersed by 5 min of reperfusion (3CAO3, adenosine-independent stimulus), or two 15-min mesenteric artery occlusions (MAOs) separated by 15 min of reperfusion (2MAO15, remote myocardial preconditioning stimulus). Pharmacological cardioprotection was produced by multiple 15-min infusions of adenosine (200 μg/min iv) separated by 15 min of washout (nADO15). Myocardial infarcts were produced by a 60-min CAO (index ischemia), and IS was determined after 120 min of reperfusion (35).

Pilot experiments to develop a model for tolerance to classical IPC by an adenosine-dependent IPC stimulus.

Because there were no previous studies on tolerance to IPC in the rat heart, we first established 1) the number and timing of CAO15 required to elicit tolerance to IPC (Fig. 1). On the basis of these experiments (see results), 4CAO15 interspersed by 15 min of reperfusion and applied between 175 and 70 min before the 60-min index ischemia was selected to induce tolerance, whereas 2CAO15 separated by 15 min of reperfusion was used as the preconditioning stimulus.

Fig. 1.

A: time window of protection by 15-min episodes of coronary artery occlusion (CAO15). B: model for development of tolerance. IPC, ischemic preconditioning; Rep, reperfusion; IA, infarct area; AR, area at risk. *P < 0.05 vs. corresponding control.

Adenosine and development of tolerance to preconditioning.

We first established whether the cardioprotection by 2CAO15, similar to 1CAO15 (23, 24), depends on adenosine receptor activation, but not on ROS generation (Fig. 2A). For this purpose, we used the adenosine receptor antagonist 8-sulfophenyl theophylline (8-SPT, 50 mg/kg iv) (23) and the ROS scavenger mercaptopropionyl glycine (MPG, 1 mg·kg−1·min−1 iv) (24). Subsequently, we investigated whether loss of adenosine signaling could have contributed to development of tolerance to 2CAO15 (Fig. 2B). For this purpose, we measured interstitial adenosine levels during 4CAO15 and determined whether an exogenous adenosine infusion indeed reinstates protection in myocardium that has become tolerant to 4CAO15 by replacing the 2CAO15 by two episodes of ADO15 (4CAO15 + 2ADO15). Finally, we subjected rats to one (1ADO15) or six (6ADO15) episodes of 15-min intravenous infusion of 200 μg/min adenosine (Fig. 2C) to determine whether repeated administration of exogenous adenosine leads to tolerance to its cardioprotection (8, 13, 39).

Fig. 2.

A: role of adenosine in cardioprotection by the classical IPC stimulus 2CAO15. 8-SPT, 8-S-sulfophenyl theophylline (hatched bars); MPG, mercaptopropionyl glycine (open bars). B: role of adenosine in development of tolerance to the classical IPC stimulus 2CAO15. Numbers in protocol for group 14 represent interstitial adenosine concentrations at baseline and during each of the 4 periods of CAO15 interspersed and followed by 15 min of reperfusion. C: development of tolerance to repeated infusions of exogenous adenosine (hatched bars). *P < 0.05 vs. control (group 1) or 4CAO15 + 70-min Rep (group 6). †P < 0.05 vs. 2CAO15 + 10-min Rep. ‡P < 0.05 vs. baseline concentration. ¶P < 0.05, 6AD015 vs. 1AD015.

Cross tolerance between adenosine-dependent and other IPC stimuli.

To investigate whether the cardioprotection by the adenosine-independent preconditioning stimulus 3CAO3 or remote preconditioning is also lost after myocardium has become tolerant to the adenosine-dependent stimulus (2CAO15), we replaced the 2CAO15 by the adenosine-independent classical stimulus 3CAO3 (4CAO15 + 3CAO3; Fig. 3A) or by remote preconditioning with two episodes of MAO15 (4CAO15 + 2MAO15; Fig. 3B).

Fig. 3.

Classical adenosine-independent IPC stimulus (3CAO3; A) and remote preconditioning stimulus (2MAO15; checkered bars, B) can still afford protection when myocardium has become tolerant to 2CAO15. *P < 0.05 vs. control (group 1) or 4CAO15 + sham. †P < 0.05 vs. corresponding stimulus without preceding 4CAO15 + 70-min Rep (group 6).

Data Analysis and Presentation

IS was analyzed by one-way ANOVA followed by Student-Newman-Keuls test. Hemodynamic variables were compared by two-way ANOVA for repeated measures followed by Dunnett's test. Statistical significance was accepted when P < 0.05. Values are means ± SE.


Mortality and Exclusions

Of the 190 rats that entered the study, 21 were excluded because of sustained ventricular fibrillation during CAO or pump failure (≤3 rats per group) and 6 were excluded because of an AR <10% of the left ventricle.

Heart Rate and Arterial Blood Pressure

Table 1 shows the hemodynamic data for the various experimental groups. Importantly, there was no correlation between the rate-pressure product at the onset of the 60-min CAO and IS (r2 = 0.003, P = 0.55).

View this table:
Table 1.

HR and MAP

Area at Risk

There were no intergroup differences in AR (34 ± 1% of left ventricle, P = 0.09) between the experimental groups.

Development of the Model for Tolerance to Classical IPC

Figure 1A shows that the protection by 1CAO15 was not affected when the reperfusion period between 1CAO15 and the 60-min index ischemia period was extended from 10 to 90 min but was lost when the index ischemia period was further extended to 175 min. The protection by 1CAO15, when applied 70 min before the 60-min CAO, was abolished, however, when this stimulus was preceded by three additional episodes of CAO15 (4CAO15 + 70-min Rep; Fig. 1B). Figure 1 also illustrates that 2CAO15 tended to be slightly more protective (IS = 37 ± 6%) than 1CAO15 (IS = 43 ± 5%, P = NS) and that the protection was abolished when preceded by the tolerance-inducing 4CAO15. This loss of protection by 2CAO15 was not due to cumulative necrosis induced by the preceding 4CAO15 (10 ± 4%), because the combined IS of 4CAO15 (10 ± 4%) and 2CAO15 + 60-min CAO (37 ± 6%) amounted to 47 ± 7%, which was still significantly less than the IS produced by the 60-min CAO alone (69 ± 2%; Fig. 1A) and the IS produced by 6CAO15 followed by the 60-min CAO (68 ± 1%; Fig. 1B).

Adenosine and Development of Tolerance

The protection by 2CAO15 was virtually abolished by 8-SPT but not by MPG (Fig. 2A), demonstrating the critical role of adenosine in mediating the cardioprotection by 2CAO15 and its independence of ROS generation.

During the first CAO15, the average interstitial adenosine concentrations increased sevenfold; during the second, third, and fourth CAO15, however, the adenosine concentrations were no longer different from baseline (Fig. 2B). When 4CAO15 was followed by 2ADO15, IS produced by the 60-min CAO was limited to 40 ± 4%. Because 10 ± 4% of the AR had already become necrotic after 4CAO15, the additional irreversible damage produced by the 60-min CAO was (40 ± 4) − (10 ± 4%), which equals 30 ± 5% of the AR. If we take into account that only 90 ± 4% [100% − (10 ± 4%)] of the AR was viable at the onset of the 60-min CAO, the percentage of the AR that became infarcted during the 60-min CAO amounted to 33 ± 6% of the viable AR [(30 ± 5%)/(90 ± 4%)]. This degree of protection was not different from the IS limitation by 2ADO15 alone (IS = 35 ± 8%; Fig. 2B). These findings indicate that exogenous adenosine still produces cardioprotection at a time that the myocardium has become tolerant to protection by the adenosine-dependent stimulus 2CAO15.

Although the cardioprotection by exogenous adenosine was unperturbed in hearts tolerant to cardioprotection by 2CAO15, IS limitation by 6ADO15 was less than that by 1ADO15 (Fig. 2C), indicating that repeated adenosine infusions caused a blunting of its cardioprotective actions.

Cross Tolerance Between Adenosine-Dependent and -Independent Classical IPC Stimuli

When 4CAO15 preceded the adenosine-independent (22, 23) but ROS-dependent (24) 3CAO3 stimulus, IS limitation was still present, although it was less (IS = 50 ± 8%, P < 0.05 vs. control) than the protection by 3CAO3 alone (IS = 29 ± 5%; Fig. 3A). Taking into account that IS was 10 ± 4% after 4CAO15 alone, we calculated (see above) that 44 ± 9% of the AR that was viable at the onset of 3CAO3 became infarcted (P = NS vs. 3CAO3 alone). These findings indicate that myocardium that has become tolerant to protection by an adenosine-dependent IPC stimulus can still be protected by a classical adenosine-independent IPC stimulus.

Cross Tolerance Between Classical IPC and Remote IPC Stimuli

Remote myocardial preconditioning by 2MAO15 limited IS to 49 ± 6% vs. 69 ± 2% in control rats (P < 0.05; Fig. 3B). When 4CAO15 preceded 2MAO15 (4CAO15 + 2MAO15), IS was limited to 49 ± 6% (43 ± 7% of AR that was viable at the onset of 2MAO15), which was not different from the cardioprotection by 2MAO15 alone (IS = 49 ± 6%).


The present study was undertaken to assess whether the development of tolerance to a particular IPC stimulus also affects the cardioprotection by stimuli that employ different mechanisms. The major findings can be summarized as follows: 1) IPC by 2CAO15 resulted in potent cardioprotection against subsequent 60-min index ischemia in an adenosine-dependent manner. However, when 4CAO15 preceded 2CAO15, the myocardium had become tolerant to the protection by 2CAO15. 2) Development of tolerance coincided with loss of myocardial interstitial adenosine release. Although repeated infusion of adenosine was capable of producing tolerance as well, the loss of adenosine release, in conjunction with the finding that exogenous adenosine still afforded protection after 4CAO15, is consistent with previous observations in pigs (41) that loss of adenosine release contributes to the development of tolerance. 3) Myocardium that had become tolerant to 2CAO15 could still be protected by the adenosine-independent 3CAO3 stimulus and by the remote preconditioning stimulus 2MAO15.

Development of Tolerance to Preconditioning in Rat Heart

Cohen et al. (6) demonstrated in conscious rabbits that 5-min CAOs at 30-min intervals for 8 h during 3 days resulted in tolerance. Iliodromitis et al. (14) showed in anesthetized rabbits that myocardial tolerance already started to develop after four cycles of 5 min of CAO and 10 min of reperfusion. The present study shows that, also in the rat heart, tolerance develops after a limited number of brief CAOs. Because in our study we used 15-min, rather than 5-min, CAOs, it could be argued that the loss of protection by 2CAO15 after 4CAO15 was caused by cumulative necrosis. This is, however, highly unlikely inasmuch as the combined IS of 4CAO15 (10 ± 4%) and 2CAO15 followed by 60 min of CAO (37 ± 6%) was significantly less (47 ± 7%) than IS after 6CAO15 followed by 60 min of CAO (68 ± 1%). Moreover, the cardioprotection by 1CAO15 + 70 min of reperfusion (IS = 45 ± 8%) was prevented when 1CAO15 was preceded by three additional episodes of CAO15 (4CAO15 + 70 min of reperfusion, IS = 69 ± 2%). Nor can this loss of protection be explained by cumulative necrosis of 3CAO15 (which, in view of the IS by 4CAO15, must have been smaller than 10 ± 4%), and the 1CAO15 + 70 min of reperfusion + 60 min of CAO (IS = 45 ± 8%), which was still significantly less than <69 ± 2% in the group subjected to 4CAO15 + 70 min of reperfusion + 60 min of CAO.

We established that, similar to 1CAO15 (23), cardioprotection by 2CAO15 involves activation of adenosine receptors, whereas ROS do not play a role in 1CAO15 (24) or 2CAO15 (present study). In view of the similarly prominent role of endogenous adenosine in cardioprotection by 1CAO5 in rabbits (31) and 1CAO10 in swine (34), a reduced adenosine receptor responsiveness (13, 39) and a progressive loss of adenosine production during repeated occlusions (14, 41) have been proposed as mechanisms underlying the development of tolerance. Although we found that repeated adenosine infusions are capable of blunting adenosine's cardioprotection, the observation that during the second, third, and fourth CAO15, the myocardial interstitial levels of adenosine were no longer different from baseline is consistent with the hypothesis that a progressive loss of myocardial adenosine release contributes to the development of tolerance (41). Moreover, similar to the findings of Vogt et al. (41), we observed that intravenous infusion of adenosine could reinstate cardioprotection, suggesting that cardiac responsiveness to adenosine was maintained after 4CAO15.

Cross Tolerance to Other IPC Stimuli

The primary aim of the present study was to investigate whether jeopardized myocardium that has become tolerant to a particular preconditioning stimulus can still be rescued by an ischemic stimulus that operates via a different mechanism. Cross tolerance to remote preconditioning of the heart did not occur, inasmuch as the cardioprotection by brief intestinal ischemia, such as 2MAO15, was not affected when this stimulus was preceded by 4CAO15. We previously showed that MAO15 elicits cardioprotection via activation of a neurogenic pathway during early reperfusion of the mesenteric bed (11). We obtained evidence that adenosine receptor activation downstream of the neurogenic pathway, possibly in the myocardium (25), contributes to remote IPC by 1MAO15. However, in two additional rats, we did not observe an increase in myocardial interstitial adenosine levels during (1.6 ± 0.2 μM) or after (1.5 ± 0.1 μM) MAO15 compared with “baseline” adenosine levels measured after the preceding 4CAO15 (2.2 ± 0.9 μM). It must be emphasized that the myocardial adenosine concentrations represent the average concentrations of the 15-min microdialysis sampling period. Therefore, we cannot exclude the possibility that a brief transient increase in myocardial adenosine concentration during early mesenteric artery reperfusion was masked. Alternatively, other mediators of remote preconditioning, including bradykinin (33), calcitonin-gene related peptide (37), and opioids (29, 42), may also have contributed to the cardioprotection by remote IPC of hearts that have become tolerant. Studies are needed to further investigate the role of these other mediators in the cardioprotection by remote IPC of myocardium made tolerant by 4CAO15 to the cardioprotection by 2CAO15.

Tolerance by 4CAO15 did also not abolish the cardioprotection by 3CAO3. If one assumes that the 15-min ischemia episode encompasses the signaling cascade triggered by the 3-min episode, one would expect a complete loss of cardioprotection by 3CAO3 in myocardium that had become tolerant to CAO15. However, we and others showed that, in contrast to its involvement in CAO15 (23), adenosine is not involved in the cardioprotection by 3CAO3 (22, 23). Conversely, we showed that the ROS scavenger MPG attenuated the protection by 3CAO3 (24) but left the protection by CAO15 unaffected. Hence, the partial loss of protection by 3CAO3 in myocardium tolerant to CAO15 is difficult to explain. Future studies, involving other triggers and mediators, are required to determine the molecular basis for this partial cross tolerance to other classical preconditioning stimuli. Nonetheless, our data suggest that myocardium that has become tolerant to the protection by a stimulus employing a particular signal transduction pathway might still benefit from an IPC stimulus employing a different signal transduction pathway.

Clinical Relevance

Abundant evidence has been presented that IPC also occurs in humans with use of end points other than IS (20, 21, 38, 43). However, clinical studies on IS limitation by preinfarct angina are discordant (4, 16, 27, 28, 44). This has, at least in part, been ascribed to loss of preconditioning in the aging (1, 2, 3, 19) and pathological (9, 12, 15, 18) hearts. We hypothesized that development of tolerance might also contribute to the equivocal clinical findings, inasmuch as multiple brief episodes of abrupt ischemia in the hours to days preceding a myocardial infarction render animal hearts tolerant to the cardioprotective effects of preconditioning. However, rather than the repetitive bouts of brief ischemia of identical duration and severity that occur in the laboratory setting, patients are more likely to experience episodes of varying severity and duration of ischemia. The present study suggests that these patients could be less susceptible to the development of tolerance as a result of recruitment of different signal transduction pathways by distinct stimuli. Our study also indicates that, without a detailed knowledge of the number, severity, and duration of the preinfarct episodes of myocardial and/or remote organ ischemia, it is impossible to classify patients as preconditioned or tolerant. Finally, the observation that administration of exogenous adenosine is still protective in hearts that have become tolerant to IPC suggests that, in patients with unstable angina, administration of pharmacological agents that mimic preconditioning can still afford cardioprotection, at least in the (sub)acute setting (8, 13, 39).


The present study was supported by The Netherlands Heart Foundation Grant NHS 99.143. D. J. Duncker is the recipient of an The Netherlands Heart Foundation Established Investigator Stipend 2000T038.


The authors gratefully acknowledge the technical assistance of Liz Keijzer.


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract