Extract
Note: Please read the complete
full text with Figures and Tables at
Introduction
Ischemia and reperfusion (I/R) induces myocardial injury. Cytosolic
Ca2+ ([Ca2+]c) is believed to be a precipitating factor.
There is increasing evidence, however, showing that mitochondria, which has a huge capacity to
accumulate[1_3] and buffer
Ca2+ under condition of high
[Ca2+]c
loading[4], plays a critical role in the maintenance of cellular
Ca2+ homeostasis in health and
disease[5]. When the heart is subjected to hypoxia or ischemia,
[Ca2+]c increases. When
[Ca2+]c increases, mitochondria take
up Ca2+ from cytosol. Accumulation of
Ca2+ in mitochondria
([Ca2+]m) is believed to be the cause of irreversible ischemic
injury[6], therefore attenuation of
[Ca2+]m should be cardioprotective. In support of this, a nitric oxide (NO) donor has been
shown to protect the heart against I/R injury accompanied by a reduction in
[Ca2+]m, suggested by the authors to be a likely
mechanism for NO-induced protection[7]. It has also been reported that opening of the ATP-sensitive potassium channel,
which confers cardioprotection induced by ischemic insults, is accompanied by attenuation of mitochondrial
Ca2+ overload and reduction in mitochondrial membrane
potential[8]. The observation, which was based on a study of correlative nature,
also supports that cardioprotection might be the consequence of attenuation of
Ca2+ overload and/or reduction in
mitochondrial membrane potential. In fact it was shown that in the heart subjected to ischemia, recovery in contractile function during
reperfusion is improved after ruthenium red is given, which significantly attenuated the
[Ca2+]m, but not
[Ca2+]c,
overload[9]. This evidence is based on a cause_effect study indicating that
[Ca2+]m plays a crucial role in cardiac injury and recovery after
I/R. Unfortunately, myocardial injury was not measured in the study.
In the present study we determined myocardial injury induced by I/R
in vivo and in vitro with manipulations that
attenuated the overload of either
[Ca2+]c and
[Ca2+]m, or
[Ca2+]m alone. We also studied the contractile function
in vivo. BAPTA-AM, a Ca2+ chelator, was used to attenuate the overload of both
[Ca2+]c and
[Ca2+]m, whereas ruthenium red, an inhibitor of
the uniporter of Ca2+ transport across the mitochondrial membrane, was used to attenuate the overload of
[Ca2+]m. Results are unequivocal evidence that
[Ca2+]m, but not
[Ca2+]c, is the immediate cause of myocardial injury induced by myocardial I/R.
Materials and methods
Isolated perfused heart preparation The protocols of this study were approved by the Committee on the Use of
Experimental Animals for Teaching and Research, The University of Hong Kong (Hong Kong SAR, China). Male Sprague-Dawley
rats of 250_300 g body weight were anesthetized with sodium pentobarbitone (60 mg/kg, ip) and given heparin (200 IU, iv).
Hearts were excised rapidly and placed in ice-cold Krebs-Henseleit (K-H) perfusion buffer before being mounted on a
Langendorff apparatus. Hearts were perfused at 37 °C with K-H buffer at a constant pressure (100 cm
H2O) and equilibrated with 95%
O2/5% CO2. The buffer contained (mmol/L): NaCl 118.0, KCl 4.7,
CaCl2 1.25,
KH2PO4 1.2, MgSO4 1.2,
NaHCO3 25.0 and glucose 11.0. For hearts subjected to regional ischemia, a silk suture was placed around the left coronary artery to form
a snare. The coronary artery was occluded by pulling the snare to produce ischemia. Reperfusion was achieved by releasing
the occlusion. In the present study, the isolated heart was subjected ischemia for 30 min followed by reperfusion for 120 min,
known to induce myocardial injury. A balloon was inserted through the left atrium into the left ventricle and the left
ventricular end diastolic pressure (LVEDP) was adjusted to between 4 and 8 mmHg. Cardiac parameters, namely, heart rate
(HR), left ventricular developed pressure (LVDP) and velocity of contraction and relaxation
(±dP/dtmax), were monitored
continuously. Coronary flow, expressed in mL/min, was measured by timed collection of effluent at regular intervals, using
a calibrated tube.
Measurement of the area of risk For determination of infarct size, the coronary artery was re-occluded at the end of
reperfusion and a solution with 2.5% Evans blue was perfused to determine the area of risk. Hearts were then frozen and cut
into slices, which were then incubated in a sodium phosphate buffer containing 1%
(w/v) 2,3,5-triphenyl-tetrazolium chloride for 15 min to visualize the unstained infarcted region. Infarct and risk zone areas were determined by planimetry with the
software Image/J from NIH. The area at risk was expressed as a percentage of the left ventricle. The infarct size was expressed
as a percentage of the risk zone. The risk zones in different groups were similar to each other (Figure 1, 3).
Preparation of isolated ventricular
myocytes Single ventricular myocytes were prepared from the hearts of male
Sprague-Dawley rats by enzymatic
dissociation[10]. The heart was perfused using a Langendorff apparatus with a 100% oxygenated,
non-recirculating Ca2+-free Tyrode¡¯s solution. The perfusion solution was switched to a 100% oxygenated recirculated, low
Ca2+ (50 µmol/L) Tyrode¡¯s solution containing 0.03% collagenase and 1% bovine serum albumin (BSA) for 10 min. The
ventricles were cut, minced, and gently triturated with a pipette in the low
Ca2+ Tyrode¡¯s solution containing BSA at 37 °C for
10 min. The cells were filtered through 200 mm nylon mesh and re-suspended in the Tyrode¡¯s solution, in which the
Ca2+ concentration was gradually increased to 1.25 mmol/L in 40 min. Only rod-shaped cells with clear cross-striations were used.
For ischemic insults, myocytes were incubated for 10 min with a solution containing 10 mmol/L
2-deoxy-D-glucose and 10 mmol/L sodium dithionite
(Na2S2O4), that induce metabolic inhibition and anoxia
(MI/A)[11,12], two consequences of ischemia.
This was followed by perfusion with normal K-H solution for 10 min.
Cell viability Trypan blue exclusion was used as an index of the viability of the
myocytes[13,14]. After cells were incubated
with 0.4% trypan blue dye for 3 min, they were counted in a hemocytometer chamber under a light microscope. Dead cells are
not able to exclude trypan blue and thus appear blue. The cell morphology was determined by microscopic
examination[12]. Only rod-shaped (length:width, >3:1) cells were used for data collection.
Intracellular Ca2+ recording
Intracellular Ca2+ and its transient were determined by a spectrofluorometric method using
the sensitive dye Fura-2 as Ca2+ indicator. Loading of cells with Fura-2/AM was carried out as described
previously[10]. After stabilization, isolated myocytes were incubated with 1 mmol/L Fura-2/AM at room temperature for 30 min. The loaded cells
were washed 3 times with fresh K-H buffer solution containing 1% BSA to wash out the unincorporated Fura-2/AM. The
myocytes were kept at room temperature for approximately 30 min to allow complete hydrolysis of acetoxymethyl ester
groups and generate Ca2+-sensitive Fura-2 free anion. The
Ca2+-dependent signal of Fura-2 was obtained by illuminating at
340 and 380 nm and recording the emitted light at 510 nm. The fluorescence ratio
(F340/F380) of Fura-2/AM loaded myocytes represents resting
[Ca2+]c level.
Measurement of [Ca2+]m
The freshly isolated ventricular myocytes were incubated with 10 mmol/L
Rhod-2/AM[15] in normal Tyrode¡¯s solution for 2 h at 37 °C. Myocytes were then washed with 1.25 mmol/L
Ca2+ Tyrode¡¯s solution to get rid of extracellular Rhod-2/AM. The myocytes were kept at room temperature for approximately 30 min to allow complete
hydrolysis of acetoxymethyl ester groups of mitochondrial dye. Confocal images were acquired using an Olympus 1X71 inverted
confocal microscope (Olympus, Japan) with a 40× water immersion objective. Rhod-2 was excited by a 543 nm helium-neon
laser and emission fluorescence was captured at >560 nm. Images were acquired in the frame scan mode with 512/512 pixels
once every min before and during MI/A and reperfusion (MI/AR). The background fluorescence was estimated by taking an
image in the absence of myocytes. The Rhod-2 fluorescence was determined by taking an image in the presence of myocytes.
The background fluorescence was then subtracted from the Rhod-2 fluorescence images. The relative Rhod-2 fluorescence
was quantified as F/F0, where
F was the fluorescence intensity of myocytes, and
F0 was the fluorescence intensity before
MI/A.
Experimental protocols Rats were anesthetized with pentobarbital, then injected with BAPTA-AM at 0.5 mg/kg, 1 mg/kg,
1.5 mg/kg or 2.5 mg/kg, iv, ruthenium red (10 mg/kg, iv) or dimethylsulfoxide
(Me2SO) as the solvent control through cannulated femoral veins. After one h, the time it takes for the drug to take
effect[16], the rats were killed and the hearts
removed.
Ventricular myocytes were subjected to continuous perfusion (the vehicle control) MI/A followed by reperfusion with or
without drug treatment. Two drugs, namely BAPTA-AM
(1 µmol/L, 2.5 µmol/L, or 25 µmol/L) and ruthenium red (50
mmol/L), were given during ischaemia for 10 min, then during reperfusion for another 10 min.
Drugs and chemicals Eagle¡¯s mininal essential medium, type I collagenase, BSA,
2-deoxy-D-glucose, Fura-2/AM, BAPTA-AM and trypan blue were purchased from Sigma Chemical (St Louis, MO, USA). Ruthenium red was purchased from
TOCRIS. Sodium dithionite was purchased from Merck and Tocris Cookson. Rhod-2/AM was purchased from Molecular
Probes. All chemicals were dissolved in distilled water or Krebs solution, except Fura-2/AM, Rhod-2/AM, calcein/AM and
BAPTA-AM, which were dissolved in Me2SO at a final
Me2SO concentration of <0.1%, which itself had no
effect[17]. The concentration of BAPTA-AM was chosen according to previous
studies[16,18]. Ruthenium red (10 mg/kg, iv) was injected
according to a previous study of Malinowska et
al[19]. The present study showed that ruthenium red at this concentration did
not affect the contractile function measured in isolated perfused rat heart, which is in agreement with the observation of the
previous study[19]. Ruthenium red 50 µmol/L was chosen for experiments in the isolated ventricular myocytes preparation
based on the concentration used in systemic administration, distributed in the extracellular fluid compartment volume of
approximately 20% of the body weight. In our preliminary experiments we found that at this concentration ruthenium red had
no significant effect on the electrically-induced
[Ca2+]i transient.
Statistical analysis Values are presented as means±SEM. Statistical comparisons were made using one-way
anova and the Newman-Keuls test. P<0.05 was considered statistically significant.
Results
Effects of BAPTA-AM on infarct size in isolated rat hearts subjected myocardial I/R
The infarct sizes of the isolated hearts after 30 min ischaemia and 2 h reperfusion
after injection of 0.5 mg/kg, 1 mg/kg, 1.5 mg/kg and 2.5 mg/kg
BAPTA-AM were 35.92%±1.57%, 31.83%±1.11%, 30.95%±1.88% and 31.09%±1.97%, respectively (Figure 1).
The infarct sizes of the groups receiving 1 mg/kg, 1.5
mg/kg and 2.5 mg/kg BAPTA-AM were significantly smaller than
that of the control with injection of
Me2SO only (37.6%±2.34 %). The observation indicates that BAPTA-AM reduces cardiac injury when given
in vivo.
Effects of MI/AR on percentage of non-blue cells in ventricular myocytes treated with BAPTA-AM
The percentage of non-blue (live) cells was 44.55%±1.41% at the end of
reperfusion, significantly lower than that of the vehicle control (61.69%±2.55 %) (Figure 2), indicating that MI/AR
increases cell death.
In the groups treated with 1, 2.5 and 25 µmol/L BAPTA-AM, the percentages of non-blue cells at the end of reper- fusion
were 45.6%±2.39%, 48.4%±1.14%, and 53.04%±
2.05 %, respectively (Figure 2). The values were significantly higher than in the group subjected to MI/AR only
(44.55%±1.41%). The results indicate that BAPTA-AM at
2.5 and 25 µmol/L reduces cell death.
Effects of myocardial I/R on infarct size
In the hearts receiving 50 µmol/L ruthenium red, the infarct size was significantly
reduced compared with the vehicle control group (Figure 3), indicating ruthenium red attenuates injury induced by
myocardial ischemia and reperfusion.
Effects of MI/AR on percentage of viable cells in ventricular myocytes treated with ruthenium red
In the myocytes treated with ruthenium red at 50 µmol/L, the percentage of non-blue cells was significantly greater than that of the group
subjected to MI/AR only (Figure 4), indicating a protective effect of ruthenium red against ischemic insult.
Effects of MI/AR on ventricular function and coronary flow
Ischemic insults resulted in marked decreases in LVDP, the
rate-pressure product (LVDP×HR),
±dP/dtmax and coronary flow and a significant increase in LVEDP during reperfusion in the
isolated perfused rat heart (Table 1). Prior injection of either 2.5 mg/kg BAPTA-AM or 50 µmol/L ruthenium red attenuated the
effects of ischemic insults on all these parameters (Table 1). Ischemic insult also reduced the heart rate slightly at the end of
reperfusion. Neither BAPTA-AM nor ruthenium red attenuated the effect of ischemic
insult. Interestingly, in the BAPTA-AM treated group the heart rate was further reduced after 60 min reperfusion (Table 1).
Effects of MI/AR on fluorescence of
F340/F380 in ventricular myocytes treated with BAPTA-AM or ruthenium red
During MI/A, the fluorescence ratio of
F340/F380, which represents the
[Ca2+]c level, was gradually increased. Immediately after
reperfusion, the ratio dropped, but was still higher than the original basal level. The fluorescence ratio during MI/AR was
significantly higher than that of the vehicle control (Figure 5). This result indicates that
[Ca2+]c was increased during
reperfusion.
When myocytes were treated with BAPTA-AM at 1 mmol/L,
2.5 µmol/L and 25 µmol/L, the
F340/F380 ratios at the end of MI/A
were 1.04±0.008, 1.02±0.007 and 1±0.007, respectively, significantly lower than the group subjected to MI/AR only
(1.07±0.006). The fluorescence ratios after 2 min of reperfusion were 1.01±0.05, 1.01±0.037 and 0.99±0.019, respectively, also
significantly lower than that of the group subjected to
MI/AR only (1.03±0.008) (Figure 5). This shows that BAPTA-AM attenuates
[Ca2+]c overload induced by MI/AR.
However, in myocytes treated with ruthenium red at 50 µmol/L, the fluorescence ratios at the end of MI/A and after 2 min
of reperfusion were the same as those in the group subjected to MI/AR (Figure 5), indicating that ruthenium red does not
affect [Ca2+]c.
Effects of MI/AR on
F/F0 of Rhod-2 The
F/F0 of Rhod-2, which represents the
[Ca2+]m level, did not change during
MI/A, but increased significantly during reperfusion (Figure
6).
As the effect of BAPTA-AM was the greatest at 25
µmol/L on [Ca2+]i, we determined the
F/F0 in myocytes treated with 25
µmol/L BAPTA-AM. From 6 min of MI/A to the end of reperfusion, the
F/F0 was significantly lower than that of the group
subjected to MI/AR (Figure 6). Interestingly, it was even lower than that of the control group without any treatment (Figure
6). This indicates that BAPTA-AM abolishes
[Ca2+]m overload induced by MI/AR. Ruthenium red at 50 µmol/L also
attenuated the F/F0 at the end of MI/A and after 2 min of reperfusion (Figure 6).
Discussion
A previous study has shown that treatment with ruthenium red, that attenuates
[Ca2+]m overload, improves the recovery
of contractile functions, indicating that attenuated
[Ca2+]m, but not
[Ca2+]c, is responsible for recovery of contractile
functions in vivo[19]. In agreement with this result, we also found in the present study that attenuation of
[Ca2+]m, but not
[Ca2+]c, overload with 50 µmol/L ruthenium red improved the recovery of contractile functions impaired by
I/R in the isolated perfused rat heart. More importantly, we observed that ruthenium red reduced myocardial injury induced by
ischemic insult and reperfusion in both the whole heart and isolated ventricular myocyte preparations. The observations are
based on a cause-effect study that showed
[Ca2+]m is crucial in cardiac
injury/protection[20]. The observations are also in
agreement with the finding that opening of the mitochondrial ATP-sensitive potassium channel confers cardioprotection and
reduces mitochondrial Ca2+ overload during I/R, suggesting that a reduction in mitochondrial
Ca2+ overload might be responsible for
cardioprotection[21]. They are also in agreement with a recent finding that
[Ca2+]m accumulation is crucial in burn
injury-induced myocardial inflammation and
function[22] .
In the present study we found that BAPTA-AM reduced injury induced by ischemic insult and reperfusion. This is in
agreement with the previous finding that BAPTA-AM reduces programmed myocyte cell death induced by angiotensin
II[23]. We also found that BAPTA-AM attenuated the overload of both
[Ca2+]c and
[Ca2+]m. Ruthenium red over-load, as shown in
the present study, reduced injury induced by I/R, therefore the cardioprotective effect of the
Ca2+ chelator must be due to attenuation of
[Ca2+]m, but not
[Ca2+]c, overload. So
[Ca2+]c overload induced by I/R leads to
[Ca2+]m overload, which
precipitates myocardial injury.
A previous study has shown that an NO donor protects the cardiomyocytes of neonatal rats against I/R-induced injury,
which is accompanied by attenuation of
[Ca2+]m[7]. The authors suggested that the attenuation of
[Ca2+]m was a likely cause
of NO-induced protection. Our study provided strong support to the suggestion.
It has also been suggested that hypercontracture, a predominant feature of reperfusion injury, is a cause of I/R-induced
injury as control of hypercontracture reduces the extent of injury. One of the causes of hypercontracture might be high
[Ca2+]c when energy recovery is
rapid[7]. This does not seem to be in total agreement with the finding that attenuation of
[Ca2+]m, but not
[Ca2+]c, is responsible for cardio-protection and improved contractile recovery. Further study is needed.
Ruthenium red at a concentration range up to 100 µmol/L has been shown to inhibit the efflux of
Ca2+ by way of the ryanodine receptor of the sarcoplasmic reticulum of both skeletal and cardiac
muscles[18]. This might reduce the
[Ca2+]c, which might in turn reduce
[Ca2+]m. In the present study we observed that ruthenium red at 50 mmol/L did not affect
[Ca2+]c at the end of MI/A and after 2 min of reperfusion, indicating that the inhibitory effect of ruthenium red on
Ca2+ efflux by way of the ryanodine receptor did not affect the effects of MI/A or reperfusion on
[Ca2+]c. The inhibitory effect of
Ca2+ efflux might also increase the
Ca2+ content in the sarcoplasmic reticulum. There is, however, no evidence of the direct
influence of Ca2+ content in the sarcoplasmic reticulum on cardiac function.
In the present study we observed that abolition of both mitochondrial and cytosolic
Ca2+ overload with BAPTA-AM resulted in small reductions in myocardial infarct size than abolition of the mitochondrial and cytosolic
Ca2+ overload with ruthenium red. This is not due to the degree of attenuation in
Ca2+ overload, as the mitochondrial
Ca2+ level was lower after treatment with BAPTA-AM than with ruthenium red. One possible explanation is that ruthenium red might itself have
cardioprotective action in addition to its inhibitory effect on
Ca2+ transport across the mitochondrial membrane. Further
study is warranted.
In conclusion, the present study has provided unequivocal evidence based on a cause-effect study that, in addition to
improved recovery of contractile function, attenuation of
[Ca2+]m, but not
[Ca2+]c, is responsible for reduced injury induced
by I/R.
Acknowledgement
We thank Mr CP MOK for assistance.
References
1 Borgers M, Shu LG, Xhonneux R, Thone F, Van Overloop P. Changes in ultrastructure and
Ca2+ distribution in the isolated working rabbit
heart after ischemia. A time-related study. Am J Pathol 1987; 126: 92_102.
2 Miller TW, Tormey JM. Subcellular calcium pools of ischaemic and reperfused myocardium characterised by electron probe. Cardiovasc
Res 1995; 29: 85_94.
3 Griffiths EJ, Ocampo CJ, Savage JS, Rutter GA, Hansford RG, Stern MD,
et al. Mitochondrial calcium transporting pathways during
hypoxia and reoxygenation in single rat cardiomyocytes. Cardiovasc Res 1998; 39: 423_33.
4 Carafoli E. The homeostasis of calcium in heart cells. J Mol Cell Cardiol 1985; 17: 203_12.
5 Duchen MR. Contributions of mitochondria to animal physiology: from homeostatic sensor to calcium signalling and cell death. J Physiol
1999; 516 ( Pt 1): 1_17.
6 Shen AC, Jennings RB. Kinetics of calcium accumulation in acute myocardial ischemic injury. Am J Pathol 1972; 67: 441_52.
7 Rakhit RD, Mojet MH, Marber MS, Duchen MR. Mitochondria as targets for nitric oxide-induced protection during simulated ischemia and
reoxygenation in isolated neonatal cardiomyocytes. Circulation 2001; 103: 2617_23.
8 Murata M, Akao M, O'Rourke B, Marban E. Mitochondrial ATP-sensitive potassium channels attenuate matrix Ca(2+) overload during
simulated ischemia and reperfusion: possible mechanism of cardioprotection. Circ Res 2001; 89: 891_8.
9 Miyamae M, Camacho SA, Weiner MW, Figueredo VM. Attenuation of postischemic reperfusion injury is related to prevention of
[Ca2+]m overload in rat hearts. Am J Physiol 1996; 271: H2145_53.
10 Dong H, Sheng JZ, Lee CM, Wong TM. Calcium antagonistic and antiarrhythmic actions of CPU-23, a substituted tetrahydro-isoquinoline.
Br J Pharmacol 1993; 109: 113_9.
11 Ho JC, Wu S, Kam KW, Sham JS, Wong TM. Effects of pharmacological preconditioning with U50488H on calcium homeostasis in rat
ventricular myocytes subjected to metabolic inhibition and anoxia. Br J Pharmacol 2002; 137: 739_48.
12 Zhang WM, Wong TM. Suppression of cAMP by
phosphoino-sitol/Ca2+ pathway in the cardiac kappa-opioid receptor. Am J Physiol
1998; 274: C82_7.
13 Hiebert L, Ping T. Protective effect of dextran sulfate and heparin on adult rat cardiomyocytes damaged by free radicals. J Mol Cell
Cardiol 1997; 29: 229_35.
14 Zhou X, Zhai X, Ashraf M. Direct evidence that initial oxidative stress triggered by preconditioning contributes to second window of
protection by endogenous antioxidant enzyme in myocytes. Circulation 1996; 93: 1177_84.
15 Ishida H, Hirota Y, Genka C, Nakazawa H, Nakaya H, Sato T. Opening of mitochondrial K(ATP) channels attenuates the ouabain-induced
calcium overload in mitochondria. Circ Res 2001; 89: 856_8.
16 Billman GE, McIlroy B, Johnson JD. Elevated myocardial calcium and its role in sudden cardiac death. FASEB J 1991; 5: 2586_92.
17 Wu S, Li HY, Wong TM. Cardioprotection of preconditioning by metabolic inhibition in the rat ventricular myocyte. Involvement of
kappa-opioid receptor. Circ Res 1999; 84: 1388_95.
18 Xu L, Tripathy A, Pasek DA, Meissner G. Ruthenium red modifies the cardiac and skeletal muscle Ca(2+) release channels (ryanodine
receptors) by multiple mechanisms. J Biol Chem 1999; 274: 32680_91.
19 Malinowska B, Kwolek G, Gothert M. Anandamide and methanan-damide induce both vanilloid VR1- and cannabinoid CB1
receptor-mediated changes in heart rate and blood pressure in anaesthetized rats. Naunyn Schmiedebergs Arch Pharmacol 2001; 364: 562_9.
20 Matsunaga M, Saotome M, Satoh H, Katoh H, Terada H, Hayashi H. Different actions of cardioprotective agents on mitochondrial
Ca2+ regulation in a Ca2+ paradox-induced
Ca2+ overload. Circ J 2005; 69: 1132_40.
21 Wang Y, Ashraf M. Role of protein kinase C in mitochondrial
KATP channel-mediated protection against
Ca2+ overload injury in rat myocardium. Circ Res 1999; 84: 1156_65.
22 Maass DL, White J, Sanders B, Horton JW. Role of cytosolic vs mitochondrial
Ca2+ accumulation in burn injury-related myocardial
inflammation and function. Am J Physiol Heart Circ Physiol 2005; 288: H744_51.
23 Cigola E, Kajstura J, Li B, Meggs LG, Anversa P. Angiotensin II activates programmed myocyte cell death in vitro. Exp Cell Res 1997;
231: 363_71.
24 Piper HM, Meuter K, Schafer C. Cellular mechanisms of ischemia-reperfusion injury. Ann Thorac Surg 2003; 75: S644_8.
|
