A Redox-Based Mechanism for Cardioprotection Induced by Ischemic Preconditioning in Perfused Rat Heart
Abstract Recent studies have suggested that mild redox alterations can regulate cell function. Therefore, we tested the hypothesis that alteration in the thiol redox state might be responsible for the cardioprotective effects conferred by ischemic preconditioning in the perfused rat heart. We find that preconditioning with four 5-minute periods of ischemia, each separated by 5 minutes of reflow, is associated with a significant loss of glutathione (3.98±0.32 μmol/g dry wt, n=8) compared with no preconditioning (6.38±0.24 μmol/g dry wt, n=14). We further find that the addition of N-acetylcysteine (NAC, a glutathione precursor and antioxidant) during the preconditioning protocol not only blocks the loss of glutathione (5.60±0.31 μmol/g dry wt, n=9) but also blocks the protective effects of preconditioning. It is observed that after 20 minutes of ischemia followed by 20 minutes of reflow, untreated hearts recover 38±7% (n=5) of their initial preischemic contractile function, whereas preconditioned hearts recover 91±11% (n=7). Hearts preconditioned in the presence of NAC recover 24±3% (n=7) of their preischemic function. Similarly, the addition of NAC reverses the protective effect of preconditioning on creatine kinase release. On reflow after 60 minutes of ischemia, creatine kinase release from control hearts was 271±20 IU · 20 min−1 · g dry wt−1 (n=5), whereas preconditioned hearts release only 170±26 IU · 20 min−1 · g dry wt−1 (n=6), and hearts preconditioned in the presence of NAC release 361±30 IU · 20 min−1 · g dry wt−1 (n=5). We also find that hearts preconditioned in the presence of NAC have less attenuation of the decline in pHi than hearts preconditioned in the absence of drug. Thus, a redox-sensitive mechanism may be involved in the protection afforded by ischemic preconditioning.
It has become increasingly clear that redox regulation plays an important role in a variety of cell functions. The transcription factors, AP-1 and NF-κB, and the activity of several kinases and phosphatases are subject to redox regulation.1 2 3 4 5 6 TNF-α reduces cellular levels of glutathione and inactivates a redox-sensitive protein phosphatase.7 Similarly, protein tyrosine kinase, an endoplasmic reticulum–associated tyrosine kinase, has been shown to be affected by redox regulation; diamide, a thiol-oxidizing agent, activates this kinase.8 A redox-based mechanism has also been invoked to account for the differential effect of nitric oxide on neuroprotection and neurodestruction.9 Also, many ion transporters and channels, such as the Na+-Ca2+ exchanger and Na+ and Ca2+ channels, have been shown to be sensitive to cellular redox state.10 11
In view of the wealth of data suggesting that mild redox alterations can affect cell function, we tested the hypothesis that a redox change might be responsible for the cardioprotective effect conferred by ischemic preconditioning, the phenomenon whereby one or more brief intermittent periods of ischemia, each separated by brief periods of reflow, can reduce the area of necrosis, reduce the severity of arrhythmias, and improve functional recovery after a subsequent longer period of ischemia.12 13 14 15 16 17 We report in the present study that preconditioning with four 5-minute periods of ischemia, each separated by 5 minutes of reflow, is associated with loss of glutathione in Langendorff-perfused rat hearts and that this loss is prevented by the addition of NAC, a glutathione precursor and antioxidant, during the preconditioning protocol. We further report that the addition of NAC during preconditioning blocks or diminishes the protective effects of preconditioning: the decrease in acidification during ischemia, the improvement in functional recovery on reflow after 20 minutes of ischemia, and the decrease in creatine kinase release on reflow after 60 minutes of ischemia. Thus, a redox-sensitive mechanism may be involved in the protection afforded by ischemic preconditioning.
Materials and Methods
Isolated Rat Heart Preparation
In the present study, all rats received humane care in accordance with the “Guide for the Care and Use of Laboratory Animals” published by the US National Institutes of Health (publication No. [NIH] 8523, revised 1985). Male Sprague-Dawley rats (200 to 320 g) were anesthetized with intraperitoneal pentobarbital (≈25 mg). The animals were heparinized (200 U IV), the heart was rapidly excised, and the aorta was cannulated. Retrograde perfusion was begun under constant pressure (90 cm H2O). The nonrecirculating perfusate was a Krebs-Henseleit buffer containing (mmol/L) NaCl 120, KCl 4.7, MgSO4 1.2, KH2PO4 1.2, CaCl2 1.25, NaHCO3 25, and glucose 11. The buffer was maintained at pH 7.4 by bubbling with a mixture of 95% O2/5% CO2 and at a temperature of 37°C.
To monitor contractility, a latex balloon connected to a Statham pressure transducer was inserted into the left ventricle. The balloon was inflated to give an end-diastolic pressure of 5 to 10 cm H2O. Global ischemia was created by cross-clamping the perfusate inflow line. To decrease the “no-reflow” region at the end of the ischemic period, the balloon was collapsed, and the heart was reperfused. After a few minutes of reperfusion, the balloon was reinflated to an end-diastolic pressure of ≈5 to 10 cm H2O to assess recovery of contractile function.
31P Nuclear Magnetic Resonance
Rat hearts were bathed in the perfusate for optimal magnetic susceptibility matching. Hearts were placed in a 20-mm NMR tube, and the perfusate was switched to phosphate-free Krebs-Henseleit buffer so that the inorganic phosphate resonance detected by 31P NMR would be composed of only intracellular phosphate. 31P NMR spectra were obtained at 146.1 MHz by using a Nicolet wide-bore NT-360 spectrometer. The temperature of the heart was maintained at 37±0.5°C by the variable temperature unit of the Nicolet NT-360. We shimmed on the proton signal from the heart and routinely obtained a nonspinning line width at one-half height of ≈0.2 ppm. Spectra were acquired every 5 minutes by using a 2-second interval between scans with a pulse of 70° (35 microseconds). The spectral width was ±3425 Hz, and 4000 data points were collected. The free-induction decay was multiplied by an exponential function corresponding to 40-Hz line broadening before Fourier transformation. The pHi determination was made from the chemical shift difference between the intracellular inorganic phosphate and phosphocreatine peaks.18
Glutathione was measured from 100 to 200 mg of heart tissue that was homogenized with 3 mL of 6% (vol/vol) perchloric acid in liquid nitrogen, and the supernatant, obtained after centrifugation at 9000g for 10 minutes, was neutralized with 2 mol/L K2CO3. Total glutathione (GSH+GSSG) in the sample was determined by measuring the rate of 5,5′-dithiobis-2-nitro-5-thiobenzoic acid reduction by GSH, detected at 412 nm.19 The reaction mixture included 0.6 mmol/L 5,5′-dithiobis-2-nitro-5-thiobenzoic acid, 0.21 mmol/L NADPH, and 0.5 U glutathione reductase in 125 mmol/L phosphate buffer (pH 7.5) in a total volume of 1 mL, and the assay was performed at 25°C. The concentration in the sample was evaluated by comparison with the standard curve. The tissue glutathione content is expressed as micromoles per gram dry weight.
Creatine kinase activity was measured spectrophotometrically at 25°C20 by using a creatine kinase kit containing a sulfhydryl-reducing agent, NAC, in the assay mixture.
As illustrated in Fig 1⇓, three sets of experiments were performed: to evaluate function and metabolism (protocol A), to measure creatine kinase release (protocol B), and to assay glutathione content (protocol C).
Protocol A: Function, Energy Metabolism, and pHi
These studies were performed in the NMR spectrometer, and as shown in Fig 1A⇑, pHi and high-energy phosphates were measured throughout the experimental procedure. LVDP was also monitored continuously; we report recovery of LVDP measured at 20 minutes of reflow. The protocol consisted of a 25-minute control/equilibration period, a 40-minute treatment period, a 20-minute global normothermic ischemia period, and a 20-minute reperfusion period. The differences in treatment period are summarized as follows: In group I (control, n=5), hearts were perfused with phosphate-free Krebs-Henseleit buffer. In group II (NAC, n=7), 4 mmol/L NAC (Sigma Chemical Co) was added to the perfusate during the 40-minute treatment period. In group III (PC, n=7), hearts were preconditioned with four cycles of 5 minutes of ischemia each separated by 5 minutes of reflow. Group IV (PC+NAC, n=7) was identical to group III except that 4 mmol/L NAC was added to the perfusate 10 minutes before preconditioning and was present throughout the preconditioning protocol. During the final 20-minute reperfusion period, all hearts were perfused with Krebs-Henseleit buffer without NAC.
Protocol B: Creatine Kinase Release
The same protocol as described above was followed except that to allow measurable creatine kinase release, the duration of global ischemia was extended to 60 minutes (Fig 1B⇑). The coronary effluent was collected during 20 minutes of reperfusion, and creatine kinase activity was measured. The same four groups of hearts were studied: group I (control, n=5), group II (NAC, n=5), group III (PC, n=6), and group IV (PC+NAC, n=5). Creatine kinase release during 20 minutes of reperfusion is expressed as IU per 20 minutes per gram dry weight (IU · 20 min−1 · g dry wt−1).
These studies were also performed in the NMR spectrometer to allow 31P NMR monitoring. Measurements of pHi and ATP during the preischemic period and 20 minutes of ischemia include data from protocols A and B, whether the total duration of ischemia was 20 minutes or 60 minutes, but the reflow data do not include the hearts subjected to 60 minutes of ischemia. Therefore, for the pHi and ATP measurements, the number of hearts per group was greater than that for the recovery of LVDP (only from protocol A). ATP and pHi experiments are as follows: group I (control, n=10), group II (NAC, n=12), group III (PC, n=13), and group IV (PC+NAC, n=12).
Protocol C: Glutathione Content
The same protocol as above was followed except that at the end of the treatment period (ie, immediately before sustained ischemia), the heart was freeze-clamped by tongs precooled with liquid nitrogen (Fig 1C⇑). Four groups of hearts were studied: group I (control, n=14), group II (NAC, n=6), group III (PC, n=8), and group IV (PC+NAC, n=9).
Values are expressed as mean±SEM. Statistical analysis (for glutathione content, creatine kinase release, and LVDP recovery) was performed by a Systat 5 program using fully factorial (M)ANOVA. The time-dependent ATP and pHi data were subjected to ANOVA for repeated measurements. When ANOVA demonstrated that significant differences existed, a post hoc (Tukey) test was performed. The level of statistical significance was P<.05.
Hemodynamics Before Sustained Ischemia
As shown in the Table⇓, there are no differences in LVDP, heart rate, or coronary flow rate among the four groups at the end of the 25-minute control/equilibration period. In hearts treated with 4 mmol/L NAC (group II), coronary flow was increased to 139% of the initial value at the end of the 40-minute treatment period, while LVDP and heart rate were unchanged. Preconditioning with four cycles of 5 minutes of ischemia and 5 minutes of reflow resulted in myocardial stunning; LVDP was reduced to 72% of the initial value at the end of the fourth reflow, immediately before the sustained period of ischemia. The changes in LVDP in hearts preconditioned in the presence of NAC (group IV) were similar to those with preconditioning alone (group III) at the end of the preconditioning protocol.
Role of Thiol Redox State in Preconditioning
It has been well documented that compared with nonpreconditioned perfused rat hearts, hearts preconditioned with four cycles of 5 minutes of ischemia and 5 minutes of reperfusion show improved recovery of LVDP (measured on reflow after 20 to 30 minutes of sustained ischemia) and reduced release of the intracellular enzyme creatine kinase during reflow after 60 minutes of sustained ischemia.14 15 16 17 To determine if a redox change occurs during the preconditioning protocol, we measured glutathione levels, which provide an index of thiol oxidation. As shown in Fig 2⇓, preconditioning leads to a reduction in glutathione levels to 3.98±0.32 μmol/g dry wt compared with no preconditioning (6.38±0.24 μmol/g dry wt). Furthermore, addition of NAC during preconditioning blocks this decline in glutathione (5.60±0.31 μmol/g dry wt). NAC alone did not significantly alter glutathione levels.
We reasoned that if a redox mechanism were responsible for the protection induced by preconditioning, then inhibition of the redox change during preconditioning should block these beneficial effects. As shown in Figs 3⇓ and 4⇓, NAC treatment by itself has no effect on recovery of LVDP or creatine kinase release after ischemia in nonpreconditioned hearts (group II). The addition of NAC during the preconditioning period, however, blocks the preconditioning-induced improvement in postischemic contractile function and decrease in enzyme release. As shown in Fig 3⇓, recovery of LVDP after 20 minutes of ischemia was 91±11% in PC hearts compared with 24±6% (P<.05) in PC+NAC hearts. Recovery of LVDP in hearts treated with NAC alone (22±7%) was not significantly different than that observed in control hearts (38±7%, P>.05). Furthermore, the recovery of LVDP in control hearts (group I), NAC hearts (group II), and PC+NAC hearts (group IV) did not differ significantly. Similarly, as shown in Fig 4⇓, creatine kinase release during 20 minutes of reflow after 60 minutes of ischemia was significantly less in PC hearts (170±26 IU · 20 min−1 · g dry wt−1) than in PC+NAC hearts (361±30 IU · 20 min−1 · g dry wt−1), control hearts (271±20 IU · 20 min−1 · g dry wt−1), or NAC hearts (261±34 IU · 20 min−1 · g dry wt−1). Thus, hearts preconditioned in the presence of NAC had levels of creatine kinase release and recovery of LVDP that were not significantly different from those found in nonpreconditioned hearts but were significantly different from those in hearts preconditioned without NAC.
Ischemia has been shown to cause a large decline in pHi, and preconditioning has been shown by several groups to attenuate the decline in pHi during the sustained period of ischemia.16 17 This reduced acidification could be beneficial in reducing calcium overload during ischemia and reflow.16 If a change in thiol redox state is the mediator of preconditioning and reduced ionic derangements are involved in preconditioning, then it might be expected that blocking thiol redox changes by perfusion with NAC would block the ability of preconditioning to reduce acidification. Interestingly, as shown in Fig 5⇓, the effect of preconditioning on the decline in pHi during ischemia is attenuated by the addition of NAC. We find that NAC treatment by itself in nonpreconditioned hearts did not alter the time course or extent of decline in pHi during ischemia. In NAC hearts (group II), pHi decreased to 6.03±0.08 during 15 to 20 minutes of ischemia, similar to that in untreated (control) hearts (5.96±0.12, P>.05). However, PC+NAC hearts showed pHi values (6.29±0.03) during 15 to 20 minutes of sustained ischemia that were intermediate between the NAC hearts (6.03±0.08) and the PC hearts (6.43±0.03); pHi values (at 15 to 20 minutes) in these three groups were significantly different from each other (P<.05).
The time course of changes in ATP during the preischemic period, 20 minutes of ischemia, and 20 minutes of reflow in all groups is depicted in Fig 6⇓. There were no significant differences in the decline in ATP content during ischemia. Upon reperfusion, ATP recovered to ≈30% in all groups. In addition, recovery of phosphocreatine content (≈90% of the initial value) and coronary flow (≈80% of the initial value) was not significantly different among the groups, indicating that the differences in recovery of LVDP were not due to differences in perfusion.
Effects of NAC on Nonpreconditioned Hearts
NAC, a thiol-containing low-molecular-weight compound, has been used as an antioxidant to prevent depletion of intracellular glutathione stores in several disease processes.21 22 Although it has been shown that 150 minutes of perfusion with NAC can increase tissue glutathione content in perfused rabbit hearts,23 in our experiments, tissue glutathione content was not increased after 40 minutes of aerobic perfusion with 4 mmol/L NAC. This presumably is due to the shorter duration of perfusion with NAC. Glutathione is synthesized by the consecutive actions of γ-glutamylcysteine synthase and glutathione synthase.24 25 The first step is controlled by feedback inhibition by glutathione. This negative-feedback mechanism seems normally to determine the upper level of intracellular glutathione, even in the presence of NAC.
In the present study, postischemic functional recovery and creatine kinase release in NAC-treated nonpreconditioned hearts did not differ significantly from those in untreated hearts, indicating that NAC treatment by itself is not detrimental. In fact, in the previous studies, NAC treatment was found to improve postischemic functional recovery and to decrease creatine kinase release in the perfused rabbit heart subjected to 60 minutes of low-flow ischemia (1 mL/min),23 when 1 μmol/L NAC was given throughout the experimental protocol (including before, during, and after ischemia). In other studies, NAC, also administered throughout the experiment, has been shown to reduce myocardial stunning, but NAC administration failed to limit infarct size in an in vivo canine model26 and failed to reduce creatine kinase release in an in vitro rat model.27 These authors attributed the protective effects of NAC to the increased thiol availability and/or direct scavenging of oxygen-derived free radicals, which presumably were produced during the early reperfusion period, by NAC. In the present study, when NAC was perfused only before the sustained period of ischemia, we found that NAC was not protective.
NAC in Preconditioned Hearts
Numerous studies have documented that small changes in thiol redox potential can exert signaling functions that can be blocked by high levels of thiol antioxidants. Previous studies have shown that transient thiol oxidation results in a modest reduction in glutathione levels, which is critical for the activation of NF-κB by TNF-α, since NAC blocks both the decrease in glutathione and the activation of NF-κB.3 The present study and other studies1 2 3 4 5 6 7 8 show that thiol redox potential can exert a regulatory role in cellular metabolism in a manner comparable to phosphorylation and dephosphorylation.
The present study shows that during preconditioning, there is thiol oxidation during the brief cycles of ischemia and reperfusion, resulting in loss of glutathione. This is prevented by NAC, either because NAC serves as an antioxidant that minimizes glutathione oxidation or because NAC serves as a substrate to replenish glutathione when tissue glutathione is decreased and the synthase becomes active. The data in the present study illustrate that a redox-sensitive mechanism can play a role in the protective effects of ischemic preconditioning in the heart. Although prevention of thiol oxidation is associated with protection against oxidative injury induced by toxic drugs and other stresses, during preconditioning, thiol oxidation is associated with adaptive responses that protect cardiac myocytes from subsequent ischemic injury.
The data in the present study are also consistent with the hypothesis that free radicals are mediators of preconditioning28 29 and that NAC either directly or via glutathione negates this mediator. Previous work on the effect of antioxidants in ischemic preconditioning has produced conflicting results. Richard et al30 found that perfusion with MPG before and during preconditioning in the in vivo rat did not affect the ability of preconditioning to reduce infarct size. Conversely, Tanaka et al29 found that perfusion with MPG or superoxide dismutase diminished the protective effect of preconditioning on infarct size in the in vivo rabbit heart. However, neither of these studies measured glutathione; therefore, the effectiveness of the antioxidants is difficult to assess. In the present study, perfusion with NAC prevented the loss of glutathione and reversed the protective effect of preconditioning on postischemic recovery of contractile function and on enzyme release.
Previous investigators have tested the hypothesis that the protective effect of preconditioning might be mediated by an increase in antioxidant defenses, but experimental data suggest that antioxidant defenses are not increased in preconditioned rabbit myocardium.31 Contrary to the expectation that preconditioning might be associated with increased protection against oxidation, the present study suggests that thiol oxidation during the preconditioning protocol may actually be protective and that elimination of thiol oxidation during the preconditioning protocol, by treatment with NAC, blocks the protective effects of preconditioning in rat heart.
Thus, we have shown that preconditioning is associated with a more oxidized cellular redox state, as measured by the loss of glutathione. If we block this more oxidized redox state by the addition of NAC during preconditioning, we block the protective effects of preconditioning. Hearts preconditioned in the presence of NAC have a greater decline in pHi during ischemia than hearts preconditioned in the absence of NAC, and hearts preconditioned in the presence of NAC show no improvement in recovery of function and no diminution in enzyme release compared with nonpreconditioned hearts. These data support the hypothesis that redox control is important in the protection afforded by preconditioning.
Selected Abbreviations and Acronyms
|control group||=||rat hearts perfused with phosphate-free Krebs-Henseleit buffer|
|GSH+GSSG||=||reduced plus oxidized glutathione|
|LVDP||=||left ventricular developed pressure|
|NAC group||=||rat hearts treated with NAC|
|NF-κB||=||nuclear factor κB|
|NMR||=||nuclear magnetic resonance|
|PC group||=||rat hearts treated with preconditioning|
|PC+NAC group||=||rat hearts treated with preconditioning in the presence of NAC|
|TNF-α||=||tumor necrosis factor-α|
This study was supported in part by National Institutes of Health grant R01-HL-39752 (Dr Steenbergen).
- Received March 9, 1995.
- Accepted April 19, 1995.
- © 1995 American Heart Association, Inc.
Schenk H, Klein M, Erdbrugger W, Droge W, Schulze-Osthoff K. Distinct effects of thioredoxin and antioxidants on the activation of transcription factors NF-κB and AP-1. Proc Natl Acad Sci U S A. 1994;91:1672-1676.
Menon SD, Qin S, Gu GR, Tan YH. Differential induction of nuclear NF-κB by protein phosphatase inhibitors in primary and transformed human cells: requirement for both oxidation and phosphorylation in nuclear translocation. J Biol Chem. 1993;268:26805-26812.
Staal FJ, Roederer M, Herzenberg LA, Herzenberg LA. Intracellular thiols regulate activation of nuclear factor κB and transcription of human immunodeficiency virus. Proc Natl Acad Sci U S A. 1990;87:9943-9947.
Abate C, Patel L, Rauscher J, Curran T. Redox regulation of fos and jun DNA-binding activity in vitro. Science. 1990;249:1157-1161.
Hayashi T, Ueno Y, Okamoto T. Oxidoreductive regulation of nuclear factor κB: involvement of a cellular reducing catalyst thioredoxin. J Biol Chem. 1993;268:11380-11388.
Guy GR, Cairns J, Ng SB, Tan YH. Inactivation of a redox-sensitive protein phosphatase during early events of tumor necrosis factor/interleukin-1 signal transduction. J Biol Chem. 1993;268:2141-2148.
Bauskin AR, Alkalay I, Ben-Neriah Y. Redox regulation of a protein tyrosine kinase in the endoplasmic reticulum. Cell. 1991;66:684-696.
Reeves JP, Bailey CA, Hale CC. Redox modification of sodium-calcium exchange activity in cardiac sarcolemmal vesicles. J Biol Chem. 1986;561:4948-4955.
Chiamvimonvat N, O’Rourke B, Kamp TJ, Kallen RG, Hofmann F, Flockerzi V, Marban E. Functional consequences of sulfhydryl modification in the pore-forming subunits of cardiovascular Ca2+ and Na+ channels. Circ Res. 1995;76:325-334.
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.
Liu Y, Downey JM. Ischemic preconditioning protects against infarction in rat heart. Am J Physiol. 1992;263:H1107-H1112.
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.
Cave AC, Collins CS, Downey JM, Hearse DJ. Improved functional recovery by ischemic preconditioning is not mediated by adenosine in the globally ischemic isolated rat heart. Cardiovasc Res. 1993;27:663-668.
Steenbergen C, Perlman ME, London RE, Murphy E. Mechanism of preconditioning: ionic alterations. Circ Res. 1993;72:112-125.
Kida M, Fujiwara H, Ishida M, Kawai C, Ohura M, Yabuuchi Y. Ischemic preconditioning preserves creatine phosphate and intracellular pH. Circulation. 1991;84:2495-2503.
Jacobus WE, Pores IH, Taylor GJ, Nunnaly RL, Hollis DP. Tight coupling of intracellular pH and ventricular performance. J Mol Cell Cardiol. 1978;10(suppl 1):S39. Abstract.
Bergmeyer HU. Methods of Enzymatic Analysis. 2nd ed. New York, NY: Academic Press Inc; 1974.
Bernard GR, Lucht WD, Niedermeyer ME, Snapper JR, Olgetree ML, Brigham KL. Effects of N-acetylcysteine on pulmonary response to endotoxin in the awake sheep and upon in vitro granulocyte function. J Clin Invest. 1984;73:1772-1784.
Forman MB, Puett DW, Cates CU, McCroskey DE, Beckman JK, Greene HL, Virmani R. Glutathione redox pathway and reperfusion injury: effect of N-acetylcysteine on infarct size and ventricular function. Circulation. 1988;78:202-213.
Murry CE, Richard VJ, Jennings RB, Reimer KA. Preconditioning with ischemia: is the protective effect mediated by free radical-induced myocardial stunning? Circulation. 1988;78(suppl II):II-77. Abstract.
Tanaka M, Fujiwara H, Yamasaki K, Sasayama S. Superoxide dismutase and N-2-mercaptopropionyl glycine attenuate infarct size limitation effect of ischemic preconditioning in the rabbit. Cardiovasc Res. 1994;28:980-986.
Richard V, Tron C, Thuillez C. Ischemic preconditioning is not mediated by oxygen derived free radicals in rats. Cardiovasc Res. 1993;27:2016-2021.
Turrens JF, Thornton J, Barnard ML, Snyder S, Liu G, Downey JM. Protection from reperfusion injury by preconditioning hearts does not involve increased antioxidant defenses. Am J Physiol. 1992;262:H585-H589.