Does Ischemic Preconditioning in the Human Involve Protein Kinase C and the ATP-Dependent K+ Channel?
Studies of Contractile Function After Simulated Ischemia in an Atrial In Vitro Model
Abstract Protein kinase C (PKC) and the ATP-dependent K+ channel (KATP channel) have been implicated in the mechanism of ischemic preconditioning in animal models. This study investigated the role of KATP channels and PKC in preconditioning in human myocardium and whether KATP channels are activated via a PKC-dependent pathway. Right atrial trabeculae were superfused with Tyrode’s solution and paced at 1 Hz. After stabilization, muscles underwent one of nine different protocols, followed by simulated ischemia (SI) consisting of 90 minutes of hypoxic substrate-free superfusion paced at 3 Hz and then by 120 minutes of reperfusion. Preconditioning consisted of 3 minutes of SI and 7 minutes of reperfusion. The experimental end point was recovery of contractile function after SI, presented here as percentage recovery (%Rec) of baseline function. %Rec was significantly improved by preconditioning by the KATP channel opener cromakalim (CK), and by the PKC activator 1,2-dioctanoyl-sn-glycerol (DOG) compared with nonpreconditioned controls when these treatments were given before the SI insult (control group, 29.5±3.6%; preconditioned group, 63.5±5.4%, CK-treated group, 52.9±3.1%; and DOG-treated group, 48.0±3.5%; P<.01). The effects of CK could be blocked by the KATP channel blocker glibenclamide (%Rec, 17.8±3.5%). Preconditioning could be blocked by the PKC antagonist chelerythrine (%Rec, 24.1±5.0%) and the KATP blocker glibenclamide (%Rec, 24.8±3.1%). The effects of DOG could also be blocked by glibenclamide (%Rec, 23.1±2.3%). These findings show that protection against contractile dysfunction after SI can be induced by activation of PKC and by the opening of the KATP channel and that the protection induced by PKC activation and preconditioning can be blocked by blocking the KATP channel. This suggests that the mechanism of preconditioning in humans may act via PKC and rely on the action of the KATP channel as the end effector.
A brief period of ischemia followed by reperfusion renders the heart more resistant to injury from a subsequent longer ischemic insult instead of accentuating the injury.1 This endogenous and highly protective phenomenon is known as ischemic preconditioning. Although ischemic preconditioning is now established in a variety of mammalian models, there is only limited evidence that the effect can also be induced in humans.
Evidence that preconditioning occurs in humans comes partly from studies during therapeutic PTCA. These studies show that the severity of myocardial ischemia is less during repeat balloon inflations than during the first inflation and are thought to demonstrate a preconditioning effect.2 3 4 However, these studies must be interpreted with caution because of the short periods of ischemia involved (<2 minutes) and the confounding effect of the possible opening up of collateral vessels. Other studies in humans have retrospectively examined whether a period of angina 24 to 48 hours before an acute myocardial infarction confers a protective effect,5 6 and these also lend support to the theory that preconditioning occurs in humans. The only in vivo prospective study using a model of global ischemia to investigate preconditioning in humans is that of Yellon et al,7 who examined whether a preconditioning protocol before cross-clamp fibrillation during coronary artery bypass surgery protects the myocardium from prolonged ischemia. Their study showed that preconditioning ultimately leads to a preservation of ATP levels in preconditioned human hearts in contrast to nonpreconditioned hearts. These findings are in keeping with the metabolic changes seen in animal models8 and add weight to the theory that preconditioning does occur in humans.
The mechanism of preconditioning is incompletely understood. The most favored current hypothesis for preconditioning suggests that endogenous ligands such as adenosine initiate an intracellular pathway by acting on G protein–linked receptors, which leads to the activation of PKC via diacylglycerol.9 Activated PKC then phosphorylates a secondary effector protein, which is thought to induce protection.
There is some evidence in humans that adenosine is the endogenous mediator of protection,10 but there is no evidence as yet that PKC is involved in human preconditioning, despite much evidence that PKC is essential in several animal models of preconditioning.9 11 12 The role of PKC is not entirely certain, and there is some conflicting evidence in a pig model that inhibition of PKC may be beneficial.13 It seems likely that the end-effector protein in the hypothesis described above induces myocardial protection by slowing cell metabolism and thus reducing ATP consumption.14 This would lead to the preservation of cells during a prolonged period of ischemia. It has been suggested that the KATP channel may be this end effector, since the opening of these channels causes an influx of K+ that shortens the cardiac action potential and limits ATP depletion and Ca2+ influx, and there is much support for this theory (eg, Auchampach et al15 and Gross and Auchampach16 ). There is experimental evidence in favor of the KATP channel being involved in preconditioning in larger mammals such as the pig17 and the dog,15 16 but the evidence in smaller animal models is less convincing, with both positive and negative studies in the rabbit18 19 and negative studies in the rat.20 21 A recent clinical study in humans involving PTCA22 found that blockade of the KATP channel with GB prevented the preconditioning effect normally seen in this angioplasty model during repeat balloon inflation, suggesting that the KATP channel may also be important in preconditioning in humans.
The aims of the present study were to investigate whether PKC plays a central role in preconditioning in human myocardium and whether the opening of the KATP channel is essential for preconditioning to occur. This study also set out to test the hypothesis that the KATP channel may be the end effector in humans by investigating whether this channel is activated via a PKC-dependent pathway. We examined isolated, superfused, and contracting right atrial trabeculae and measured differences in contractile recovery after SI. The roles of PKC and the KATP channel were investigated pharmacologically.
Materials and Methods
Experiments were performed on trabeculae derived from human atrium. Specimens of right atrial appendage were obtained from the right atrial cannula insertion site in patients undergoing cardiac bypass before coronary artery surgery. All subjects had chronic stable angina. Patients were excluded if they had atrial arrhythmias or right ventricular failure or were taking antiarrhythmic drugs or oral hypoglycemic agents (for diabetes mellitus). Prior ethical approval had been obtained from the Middlesex Hospital Clinical Investigations Panel. The specimens were kept oxygenated in modified Tyrode’s solution at 4°C. A single atrial trabecula (diameter, ≤0.9 mm; length, ≥3 mm) was identified under magnification, tied at one end with a 7/0 silk suture, and then dissected out with a small wedge of atrial wall at the free end. The silk suture was used to attach one end of the trabecula to a fixed post in an organ bath, and the other end was attached to a force transducer (Gould Statham UCT2) by snaring the wedge of atrial wall. The muscles were suspended horizontally in the organ bath through which there was a continuous flow of superfusate oxygenated with a 95% O2/5% CO2 gas mixture. The organ bath was covered to prevent gas exchange with the atmosphere. The Po2, Pco2, and pH in the organ bath were monitored by intermittent analyses of the effluent by using an automated blood gas analyzer (model AVL 993, AVL Medical Instruments). The pH was kept between 7.35 and 7.45, and the temperature, monitored by a thermocouple in the bath, was maintained at 37°C. Flow through the bath was maintained at 8 mL/min. During the SI, the superfusate was free of substrate and was bubbled with 95% N2/5% CO2 (pH 7.24 to 7.34) so that it was hypoxic.
Once suspended, trabeculae were paced by field stimulation at 1 Hz. The platinum pacing electrodes were positioned within the bath on either side of the trabecula, and pacing was driven by an isolated stimulator (Digitimer DS2) triggered by a computerized clock. The pulse width was fixed at 5 ms, and the pulse amplitude was set at twice threshold (6 to 8 V). The output of the force transducer was recorded on paper (Gould RS3400 chart recorder). The width and length of specimens were measured at the end of the experiment with an eyepiece graticule in an overhead microscope (Prior), and all specimens were then weighed.
The Tyrode’s solution consisted of (mmol/L) NaCl 118.5, KCl 4.8, NaHCO3 24.8, KH2PO4 1.2, MgSO4 · 7H2O 1.44, CaCl2 · 2H2O 1.8, glucose 10.0, and pyruvic acid 10.0. In the substrate-free Tyrode’s solution, choline chloride (7 mmol/L) was substituted for glucose and pyruvic acid to maintain constant osmolarity. All reagents were Analar grade (BDH Chemicals) except for the pyruvic acid (Sigma Chemical Co). The selective PKC blocker CHE (LC Laboratories) was dissolved in distilled water and added to the Tyrode’s solution to give a 10 μmol solution. The diacylglycerol analogue DOG (Sigma), a selective PKC activator, was initially dissolved in dimethyl sulfoxide (10 mg in 200 μL), and this was added to the Tyrode’s solution to give a final concentration of 30 μmol. The KATP channel opener CK was dissolved in dimethyl sulfoxide initially (5.16 mg in 200 μL) and then added to the Tyrode’s solution to give a final concentration of 30 μmol. The KATP channel blocker GB was made up to a concentration of 1 μmol in 100 mL Tyrode’s solution.
Trabeculae were suspended in the bath and paced unstretched for 30 minutes. They were then gradually stretched in a stepwise manner for ≈15 minutes until the maximum force of contraction was achieved. All muscles were then allowed to equilibrate for at least 45 minutes. All groups eventually underwent a period of SI followed by reperfusion, which consisted of 90 minutes of hypoxic substrate-free superfusion and rapid pacing (3 Hz) followed by reperfusion for 120 minutes with normal Tyrode’s solution and pacing at 1 Hz. The preconditioning protocol consisted of 3 minutes of hypoxic substrate-free superfusion with rapid pacing at 3 Hz followed by reperfusion with oxygenated normal Tyrode’s solution and pacing at 1 Hz. Samples were used for one protocol only. Fig 1⇓ shows the experimental protocols for the nine groups: (1) control group, stabilization period then SI (n=6); (2) PC group, preconditioning protocol then SI (n=6); (3) C+DOG group, 10 minutes of superfusion with DOG (30 μmol) before SI (n=6); (4) PC+CHE group, protocol as for the PC group with CHE (10 μmol) present during the 7-minute reperfusion period (n=6). (5) CK group, 5 minutes of superfusion with CK (30 μmol) and 7-minute washout before SI (n=6); (6) CK+GB group, protocol as for CK group with GB (1 μmol) present from 10 minutes before the administration of CK until SI (n=4); (7) PC+GB group, protocol as for PC group with GB (1 μmol) present from 10 minutes before the preconditioning protocol until SI (n=4); (8) C+GB group, 10 minutes of GB (1 μmol) before SI (n=4); and (9) DOG+GB group, 10 minutes of superfusion with DOG (30 μmol) before SI with GB (1 μmol) present from 20 minutes before SI (n=5).
Data are expressed as group mean±SEM. Statistical differences between groups were evaluated by one-way ANOVA with Fisher’s protected least-significant difference post hoc test.23 A value of P≤.01 was considered significant.
Samples were obtained from patients with stable ischemic heart disease (40 men and 7 women; age range, 36 to 80 years; mean age, 62 years). There were no exclusions in this study; all samples completed the protocol to which they were randomly assigned. There were 47 patients in total, and each of the 47 experimental samples came from a different patient and underwent one protocol only. Baseline characteristics including resting force and peak developed force were similar in all groups (see Table 1⇓); therefore, experimental data on developed force are presented graphically as a percentage of baseline developed force.
Figs 2⇓ and 3⇓ show the results for developed force as a percentage of baseline. Time 0 indicates the beginning of SI; therefore, reperfusion with oxygenated Tyrode’s solution extends from 90 to 210 minutes. As shown in the diagram of the protocols (see Fig 1⇑), SI is preceded in all groups by a stabilization period and various pretreatments (indicated on the graphs by drug/preconditioning). As noted above, there was no significant difference in developed force between the nine groups at the end of the stabilization period, and this baseline function is denoted as 100% developed force. Both graphs include the data for the control and PC groups for comparison.
Fig 2⇑ illustrates the results for those groups specifically relevant to the investigation of the KATP channel. The preconditioning protocol caused a significant reduction in developed force to 41.8±10.5% in the PC group, which recovered to baseline before entering SI. A similar reduction in function occurred in the PC+GB group, which also returned to baseline before SI. CK treatment caused a significant reduction in function in the CK and CK+GB groups (see Fig 2⇑). Reperfusion before the long hypoxia in these groups did not result in the full recovery of developed force to baseline values, although the CK+GB group did partially recover (47.1±13.1%). There was no difference in function between groups during the 90-minute period of hypoxia. On reperfusion with oxygenated Tyrode’s solution, there was significantly greater recovery of function in the PC and CK groups compared with the control group (control, 29.5±3.6%; PC, 63.5±5.4%; and CK, 52.9±3.1%; P<.01). Recovery in the PC group was not significantly different from that in the CK group. There was no difference between control and the other groups (CK+GB, 17.8±3.5%; PC+GB, 24.8±3.1%; and C+GB, 29.7±3.8%), which all recovered poorly. GB appears to have no independent detrimental effect on atrial tissue, since recovery of function in the C+GB group was the same as in group C.
Fig 3⇑ includes the data for the experimental groups specifically relevant to the investigation of the role of PKC. The PC+CHE group underwent a reduction in function similar to that in the PC group during the preconditioning protocol. The addition of DOG (C+DOG group) caused some reduction in function before SI, but this was not significantly different from the control group. On reperfusion, there was a significantly greater percentage of functional recovery in the C+DOG group (48.0±3.5%), which was similar to the recovery seen in the PC and CK groups (see Fig 2⇑). The PC+CHE group had the same poor functional recovery as the control group (PC+CHE, 24.1±5.0%; control, 29.5±3.6%). Finally, in the DOG+GB group, where the KATP channel blocker GB was given in addition to DOG, the protective effect of DOG was lost, and this group had the same recovery as the control group (DOG+GB, 23.1±2.3%; control, 29.5±3.6%).
Table 2⇓ shows the data for time to onset of contracture, defined as the time from the onset of the 90-minute hypoxia to an increase of 0.05 g in resting force above baseline, and time to peak contracture, defined as the time taken for the baseline force to rise to maximum. There was no significant difference between the experimental groups and the control group for either the time to onset of contracture or time to peak contracture.
The present study confirms that human atrium can be protected from an ischemic insult by activation of the KATP channel with CK. The protection brought about by both preconditioning and CK can be blocked by the KATP channel blocker GB, suggesting that the protection of preconditioning may result from KATP channel opening.
Activation of PKC with the diacylglycerol analogue DOG results in a degree of myocardial protection similar to that seen with preconditioning or treatment with CK (P<.01); therefore, it seems likely that the mechanism of preconditioning in humans involves a PKC-dependent pathway, in keeping with the majority of evidence from other species.9 11 In addition, the finding that the protective effect of activating PKC with DOG is abolished by blocking the KATP channel with GB suggests that the KATP channel may be activated via PKC and that this channel may be the end effector in the mechanism of preconditioning in humans. CK treatment resulted in a decrease in developed force before the 90-minute SI insult. However, the protective action of CK is unlikely to be due to cardioplegia, because GB abolished the protective effect of treatment with CK but not the preischemic cardiodepression that it caused.
The present study is the first to provide direct evidence that the KATP channel is involved in preconditioning in human myocardium. The first study to demonstrate a possible role for this channel in preconditioning in humans was that of Tomai et al,22 who used a clinical angioplasty model. Their evidence is indirect but highly persuasive. They showed that ischemic preconditioning during brief repeated coronary occlusions is completely abolished by pretreatment with GB, suggesting that preconditioning is mediated by KATP channels in humans. There are several clinical studies suggesting that preconditioning occurs in humans,2 3 4 5 6 7 but as yet, the only direct evidence that we have comes from two in vitro studies. The study of Ikonomidis et al24 used human ventricular cardiomyocyte cell cultures; they were able to show that these isolated myocytes could be protected against a 90-minute period of SI by a preconditioning protocol. The other study providing direct evidence comes from our own department10 and uses the same model as the present study. We were able to show that human atrial trabeculae that had been preconditioned had a significantly better postischemic recovery of contractile function than nonpreconditioned trabeculae. Importantly, this study also demonstrated that adenosine was involved in human preconditioning, since we were able to show that an adenosine agonist, 8-p-sulfophenyltheophylline, was able to induce the same degree of protection against ischemia as preconditioning, providing the first evidence in favor of adenosine being the endogenous ligand that “switches on” preconditioning in humans.
The present study used isolated human atrial trabeculae in an attempt to examine the mechanism of preconditioning in human myocardium. The advantages of an isolated human preparation are that one can avoid the problem of variable collateral flow experienced with in vivo models and that there are no ethical considerations involved in the use of experimental drugs in vitro. Atrial tissue was chosen because it was available and because sampling was part of the routine procedure for coronary artery bypass grafting. Atrial specimens make stable preparations and are generally disease free. There are differences between atrial and ventricular tissue, and results from one may not be applicable to the other; however, early experiments in our laboratory (authors’ unpublished data, 1995) suggest that an identical preconditioning effect can be induced in ventricular tissue if a shorter ischemic insult is used (60 minutes instead of 90 minutes). Adenosine receptors are present in both atrium and ventricle,25 as are KATP channels,26 although there may be differences in the density of these receptors. For example, Tung et al27 have shown in the guinea pig that the density of KATP channels is significantly lower in atrial than ventricular tissue. In view of these similarities in both the response to experimental ischemia and the presence of relevant channels and receptors, it seems probable that human atrial and ventricular tissue will respond in a similar way to ischemic preconditioning.
Rather than the “true” ischemia of classic preconditioning, the present study used a period of hypoxic superfusion in combination with rapid pacing to simulate ischemia, and there is a great deal of evidence in animal models of both regional and global hypoxia and in cell culture that hypoxia is as effective as ischemia in inducing preconditioning.28 29 30 31 We know that preconditioning causes improved recovery of contractile function as well as reduced infarct size. For example, a recent in vitro study by Jenkins et al,32 who measured both infarct volume and functional recovery after ischemia, showed that the improved recovery of global left ventricular function produced by preconditioning is proportional to a reduction in infarction, and Cohen et al33 and Przyklenk et al34 were able to correlate improved recovery of systolic shortening with reduced infarct size in in vivo models of regional ischemia. There is no direct evidence in atrial tissue that recovery of global function is proportional to infarct volume, but this is likely to be the case. It is clear from the evidence of these studies that the enhanced recovery of contractile function due to preconditioning, which follows a prolonged period of ischemia, is due to a reduction in infarct size and is not due to a reduction in stunning. Our model uses a 90-minute period of SI as the “ischemic insult,” a period much more likely to lead to cell death than stunning, which is induced by shorter periods of 5 to 15 minutes of ischemia. This suggests that our model is indeed one of preconditioning. However, although unlikely, we cannot entirely exclude the possibility that our model involves stunning, and Auchampach et al15 have demonstrated that K+ channel openers are effective in inducing myocardial protection in canine models of both stunning and infarction. In summary, there is evidence that hypoxia can be as effective as ischemia in preconditioning, that recovery of contractile function is a good correlate for the degree of myocyte necrosis, and that this experimental model is one of preconditioning. Therefore, although the findings of the present study strictly relate to hypoxic preconditioning, they should be relevant to ischemic preconditioning also.
In conclusion, the present study demonstrates for the first time that in human myocardium the protective effects of preconditioning can be induced by activation of PKC and the opening of the KATP channel. It also suggests that the KATP channel is activated via PKC, making this channel a likely candidate for the “end effector” in human preconditioning.
Selected Abbreviations and Acronyms
|C||=||control (used in association with drug for specimen groups)|
|KATP channel||=||ATP-dependent K+ channel|
|PC||=||preconditioning (only when used to refer to specimen groups)|
|PKC||=||protein kinase C|
|PTCA||=||percutaneous transluminal coronary angioplasty|
Dr M.E. Speechly-Dick was funded by a Junior Research Fellowship from the British Heart Foundation. Additional funding was provided by The Hatter Foundation.
- Received May 17, 1995.
- Accepted July 26, 1995.
- © 1995 American Heart Association, Inc.
Murry CE, Jennings RB, Reimer KA. Preconditioning with ischemia: a delay of lethal cell injury in ischemic myocardium. Circulation. 1986;74:1124-1136.
Deutsch E, Berger M, Kussmaul WG, Hirshfeld JW, Herrmann HC, Laskey WK. Adaptation to ischemia during percutaneous transluminal coronary angioplasty: clinical, hemodynamic, and metabolic features. Circulation. 1990;82:2044-2051.
Kloner RA, Shook T, Przyklenk K, Davis VG, Junio L, Matthews RV, Burstein S, Gibson M, Poole K, Cannon CP, McCabe CH, Braunwald E. Previous angina alters in-hospital outcome in TIMI-4. Circulation. 1995;91:37-47.
Ottani F, Galvani M, Ferrini D, Sorbello F, Limonetti P, Pantoli D, Rusticali F. Prodromal angina limits infarct size: a role for ischemic preconditioning. Circulation. 1995;91:291-297.
Murry CE, Richard VJ, Reimer KA, Jennings RB. Ischemic preconditioning slows energy metabolism and delays ultrastructural damage during a sustained ischemic episode. Circ Res.. 1990;66:913-931.
Ytrehus K, Liu Y, Downey JM. Preconditioning protects ischemic rabbit heart by protein kinase C activation. Am J Physiol.. 1994;266:H1145-H1152.
Speechly-Dick ME, Mocanu MM, Yellon DM. Protein kinase C: its role in ischemic preconditioning in the rat. Circ Res.. 1994;75:586-590.
Li Y, Kloner RA. Does protein kinase C play a role in ischemic preconditioning in rat hearts? Am J Physiol.. 1995;268:H426-H431.
Vogt A, Barancik M, Weihrauch D, Arras M, Podzuweit T, Schaper W. Protein kinase C inhibitors reduce infarct size in pig hearts in vivo. Circulation. 1994;90(suppl I):I-647. Abstract.
Reimer KA, Vander Heide RS, Murry CE, Jennings RB. Role of altered energy metabolism in ischemic preconditioning. In: Przyklenk K, Kloner RA, Yellon DM, eds. Ischemic Preconditioning: The Concept of Endogenous Cardioprotection. Norwell, Mass: Kluwer Academic Publishers; 1994:75-103.
Auchampach JA, Grover GJ, Gross GJ. Blockade of ischemic preconditioning in dogs by the novel ATP-dependent potassium channel antagonist sodium 5-hydroxydecanoate. Cardiovasc Res.. 1992;26:1054-1062.
Gross GJ, Auchampach JA. Blockade of ATP-sensitive potassium channels prevents myocardial preconditioning in dogs. Circ Res.. 1992;70:223-233.
Schulz R, Rose J, Heusch G. Involvement of activation of ATP-dependent potassium channels in ischemic preconditioning in swine. Am J Physiol.. 1994;267:H1341-H1352.
Toombs CF, Moore TL, Shebuski RJ. Limitation of infarct size in the rabbit by ischaemic preconditioning is reversible with glibenclamide. Cardiovasc Res.. 1993;27:617-622.
Thornton JD, Thornton CS, Sterling DL, Downey JM. Blockade of ATP-sensitive potassium channels increases infarct size but does not prevent preconditioning in rabbit hearts. Circ Res.. 1993;72:44-49.
Liu GS, Downey JM. Ischemic preconditioning protects against infarction in rat heart. Am J Physiol.. 1992;263:H1107-H1112.
Grover GJ, Dzwonczyk S, Sleph PG, Sargent CA. The ATP-sensitive potassium channel blocker glibenclamide (glyburide) does not abolish preconditioning in isolated ischemic rat hearts. J Pharmacol Exp Ther.. 1993;265:559-564.
Tomai F, Crea F, Gaspardone A, Versaci F, de Paulis R, de Peppo AP, Chiariello L, Gioffré PA. Ischemic preconditioning during coronary angioplasty is prevented by glibenclamide, a selective ATP-sensitive K+ channel blocker. Circulation. 1994;90:700-705.
Snedecor GW, Cochran WG. Statistical Methods. Ames, Iowa: Iowa State University Press; 1980:215-237.
Ikonomidis JS, Tumiati LC, Weisel RD, Mickle DAG, Li RK. Preconditioning human ventricular cardiomyocytes with brief periods of simulated ischaemia. Cardiovasc Res.. 1994;28:1285-1291.
Böhm M, Pieske B, Ungerer M, Erdmann E. Characterization of A1 adenosine receptors in atrial and ventricular myocardium from diseased human hearts. Circ Res.. 1989;65:1201-1211.
Heidbüchel H, Vereecke J, Carmeleit E. Three different potassium channels in human atrium: contribution to the basal potassium conductance. Circ Res.. 1990;66:1277-1286.
Tung R, Shen WK, Kurachi Y. Comparison of ATP-sensitive K channel in atrial vs ventricular myocytes in guinea-pig. J Am Coll Cardiol. 1992;19(suppl A):245A. Abstract.
Lasley RD, Anderson GM, Mentzer RM. Ischemic and hypoxic preconditioning enhance postischemic recovery of function in the rat heart. Cardiovasc Res.. 1993;27:565-570.
Cohen MV, Liu GS, Downey JM. Preconditioning causes improved wall motion as well as smaller infarcts after transient coronary occlusion in rabbits. Circulation. 1991;84:341-349.
Przyklenk K, Bauer B, Ovize M, Kloner RA, Whittaker P. Regional ischemic ‘preconditioning’ protects remote virgin myocardium from subsequent coronary occlusion. Circulation. 1993;87:893-899.