Anesthetic Preconditioning Improves Adenosine Triphosphate Synthesis and Reduces Reactive Oxygen Species Formation in Mitochondria after Ischemia by a Redox Dependent Mechanism

The investigation conformed to the Guide for the Care and Use of Laboratory Animals (US National Institutes of Health No. 85–23, revised 1996) and was approved by the institutional review board. The preparation has been described in detail previously. 2–4Hearts from guinea pig (n = 100) were perfused at 37°C, at constant pressure (55 mmHg) with an oxygenated Krebs–Ringer's solution of the following composition (in mm): Na⁺138, K⁺4.5, Mg²⁺1.2, Ca²⁺2.5, Cl⁻134, HCO³⁻14.5, H²PO⁴⁻1.2, glucose 11.5, pyruvate 2, mannitol 16, probenecid 0.1, EDTA 0.05, and insulin 5 U/L.

Left ventricular pressure (LVP) was measured isovolumetrically using a transducer connected by stiff saline-filled tubing to a latex balloon placed in the left ventricle through an incision in the left atrium. Measured characteristics of LVP were systolic and diastolic LVP; calculated characteristics were developed (systolic–diastolic) LVP and maximal and minimal first derivative of LVP (dLVP/dt^maxand dLVP/dt^min, respectively). Coronary inflow (CF) was measured by an ultrasonic flowmeter (Transonic T106X, Ithaca, NY). Atrial and ventricular bipolar leads were used to measure spontaneous heart rate. Coronary inflow (a) and coronary venous (v) Na⁺, K⁺, Ca²⁺, Po², pH, and Pco²were measured off-line with an intermittently self-calibrating analyzer (Radiometer ABL 505, Copenhagen, Denmark). Coronary sinus Po²tension (PvO²) was also measured continuously on-line with a Clark electrode (model 203B, Instech; Plymouth Meeting, PA). Myocardial O²consumption (Mvo²) was calculated as CF/heart weight (g) · (Pao²–PvO²) · 24 ml O²/mL at 760 mmHg.

Global ischemia was achieved by clamping the aortic inflow line. If ventricular fibrillation occurred on reperfusion and did not convert within 30 s a bolus of lidocaine (250 μg) was given. At the end of 120 min reperfusion hearts were removed, cut into six transverse sections and stained with 1% 2,3,5-triphenyltetrazolium chloride in 0.1 m KH²PO⁴buffer (pH 7.4, 38°C) for 10 min as described previously. 3,4Infarct size was expressed as a percentage of total heart weight.

The ROS formation was measured by the fluorescent dye DHE (Molecular Probes, Eugene, OR). The DHE enters cells and upon oxidation by ROS, with relative selectivity for superoxide 14(O^2·-), is converted to ethidium bromide; in this form it intercalates with DNA and causes the nucleus to exhibit a right shift in fluorescence. To measure fluorescent intensity of oxidized DHE, each experiment was carried out in a light-blocking Faraday cage, as we have described in detail previously, for measurements of cytosolic Ca²⁺, 3,15mitochondrial Ca²⁺, 12,15Na⁺, 15and NADH 13,15in isolated beating hearts. Briefly, the distal end of a trifurcated fiberoptic cable (optical surface area 3.85 mm²) was placed against the LV free wall through a hole in the tissue bath. Netting was applied around the heart for optimal contact without impeding relaxation. The fiberoptic cable was connected to a modified spectrophotofluorometer (SLM Aminco-Bowman II; Spectronic Instruments, Urbana, IL). Fluorescent emissions (λ^em) at 590 nm (bandwidth 4 nm) were amplified by a photomultiplier tube (700 V) and recorded following excitation (λ^ex) with a 150 W xenon arc lamp filtered through a 540 nm monochromator (bandwidth 4 nm). The excitation wavelength penetrates, with decreasing intensity, through the whole 4 mm of the guinea pig left ventricular wall.

In preliminary experiments (n = 4) background autofluorescence (no DHE) was determined after initial perfusion and equilibration, and for each experimental protocol. All subsequently recorded values of DHE fluorescent intensity were adjusted for the minimal change in background fluorescence with any maneuver or drug. DHE was first dissolved in 1 ml dimethyl sulfoxide containing 16% (weight/vol) Pluronic I-127 (Sigma Chemical, St. Louis, MO). The perfusate contained probenecid (100 mm) to retard cellular leakage of DHE. Hearts were loaded with 10 μm DHE in Krebs-Ringer's solution for 25 min followed by washout of residual DHE with standard perfusate for 15 min. Loading of DHE increased diastolic LVP approximately 8% and increased coronary flow approximately 10%; the effect on diastolic LVP was nearly reversed and flow returned to baseline values on washout. DHE loading increased fluorescence intensity from 0.03 ± 0.01 before loading to 2.7 ± 0.2 arbitrary units (a.u.) after washout. Each washout value was adjusted to 0 a.u. to normalize fluorescence intensity values for each experiment.

Mitochondrial ATP synthesis and ROS formation were determined in cardiac mitochondria in a parallel series of experiments using the same treatment protocols except that reperfusion function and infarct size were not determined. Briefly, mitochondria were isolated by differential centrifugation as described by Solem and Wallace. 16Mitochondria prepared by this methodology have been shown to be metabolically active with respiratory control ratios of 3.5–5.0 with the substrate succinate and 8.0–10.0 with substrates glutamate/malate giving corresponding ratios of ATP to oxygen of 1.5–1.7 and 2.5–2.7. 16Treated hearts were removed at 5 min reperfusion and immediately placed in buffered physiologic solution at 4°C. Next the atria and right ventricle were trimmed away and the left ventricle was dissected, weighed, and minced into 1-mm pieces in 1 ml of 4°C isolation buffer (200 mm mannitol, 50 mm sucrose, 5 mm KH²PO⁴, 1 mm EGTA, 5 mm 3-(n-morpholino)propanesulfonic acid, and 0.1% bovine serum albumin; pH 7.15 adjusted with potassium hydroxide). The tissue was rinsed clear of blood and light debris with isolation buffer and transferred to a glass Potter–Elvehjem homogenizing vessel placed on ice. Immediately after adding 2.5 mg protease in 2.5 ml isolation buffer, the tissue was gently homogenized on ice with a Teflon (DuPont, Wilmington, DE) pestle for 30 s. Then 17.5 ml cold isolation buffer was added to dilute the protease and homogenization was continued for 60 s. Exposure to the concentrated protease was limited to 30 s to maintain mitochondrial integrity and yield. Untreated control hearts (Sham) were similarly isolated.

The tissue suspension was centrifuged at 8,000 g  for 10 min to remove protease. The pellet was resuspended in 3.5 ml isolation buffer, and an additional 25 ml isolation buffer was added and the suspension centrifuged (700 g  for 10 min) to remove cellular debris. The supernatant containing the mitochondrial fraction was further centrifuged (8,000 g  for 10 min), and the pellet was washed twice by resuspension in 3.5 ml isolation buffer without EGTA and centrifuged at 8,000 g  for 10 min. The final mitochondrial pellet was resuspended to 1 g of the original tissue (weight/ml) in cold isolation buffer without EGTA. All procedures were performed at 4°C, and total protein concentration was determined using the dye binding method of Bradford 17with bovine serum albumin as a standard.

Mitochondria (10 μg mitochondrial protein/ml) were preincubated for 10 min in a medium containing 180 mm sucrose, 45 mm KH²PO⁴, 10 mm Mg-acetate, 1 mm EDTA, 1 mm pyrophosphate, 1 g/l bovine serum albumin, and 150 mm ATP in a 25°C water bath. ATP synthesis was initiated by adding respiratory substrates (1 mm pyruvate + 1 mm malate) to drive electron flow through complex I of the electron transport chain (ETC). Aliquots of 50 ml were taken from the incubations at 4 min intervals for 12 min, and reactions were halted by adding the samples to 500 ml 2.5% (vol/vol) trichloroacetic acid. 18Samples were neutralized with 100 ml of 1 mm Tris base. The neutralized supernatant was assayed for ATP by luciferin–luciferase luminometry using a modification of the luminescence method of Strehler 19and components of the Sigma Bioluminescent Somatic Cell Assay Kit (catalog no. FL-AA, Sigma Chemical Co., St. Louis, MO). Mitochondrial protein concentration was determined using the Bradford assay, 17and the rate of mitochondrial ATP synthesis was calculated and expressed as μmol ATP · min⁻¹· mg mitochondrial protein⁻¹.

Tissue ATP concentration was measured in segments of the left ventricle. These were rapidly dissected, frozen in liquid nitrogen, and stored at −70°C for later analysis. Frozen tissue samples (20 mg) were homogenized in 500 ml 2.5% trichloroacetic acid. After centrifugation, 400 ml aliquots of supernatant were neutralized with 80 ml 1 m Tris. 18–20The ATP concentration was expressed as nmol ATP · mg mitochondrial protein⁻¹.

Mitochondrial ROS was measured from the same mitochondrial pellet used to measure ATP synthesis. Determination of ROS was accomplished by a slight modification of the procedure of Garcia-Ruiz et al.  21Briefly, production of ROS was monitored with a fluorescent probe, dichlorohydro-fluorescein diacetate (DCFDA), which is oxidized in the presence of hydrogen peroxide (H²O²) to DCF. 22Guinea pig heart mitochondria (0.25–0.5 mg/ml) were incubated at 25°C with 2 μm DCFDA (Molecular Probes, Eugene, OR) in the absence or presence of 100 μm antimycin A (Sigma Chemical) to stimulate maximal ROS. Antimycin A profoundly enhances mitochondrial superoxide production by inhibiting at the ubiquinone site of complex III of the ETC. 21Superoxide is rapidly converted to hydrogen peroxide by mitochondrial MnSOD. The mitochondrial suspension was gently agitated and incubated at room temperature for measurements taken at 0, 30, 60, and 120 min. Fluorescence was determined at λ^em529 nm (bandwidth 10 nm) and λ^ex503 nm (bandwidth 5 nm), according to the spectral characteristics of DCF. The ROS formation was expressed as arbitrary units of fluorescence intensity of DCF.

Scavengers selected to block triggering by ROS during sevoflurane exposure were tiron (4,5-dihydroxy-1,3-benzene-disulphonic acid, 100 μm), and MnTBAP (Mn(III) tetrakis (4-benzoic acid) porphyrin chloride, 40 μm). There were six identical treatment groups for both isolated heart and isolated mitochondrial experiments (fig. 1). These were ischemia/reperfusion alone (I/R), sevoflurane before I/R (APC), with or without the ROS scavengers tiron (APC + TIR, TIR) or MnTBAP (APC + TBAP, TBAP). In addition, there was an isolated heart group in which hearts were perfused for 200 min but not subjected to ischemia or treatments (Time Control group; n = 4, not shown) and a Sham group (n = 24, 4 hearts/group) in which mitochondria were isolated from hearts neither subjected to treatments nor to IR injury. Isolated heart studies began 30 min after equilibration. Hearts were assigned randomly to each of the seven groups. Parallel experiments were performed for each experimental group, each on separate hearts: (1) isolated hearts for global cardiac function and ROS formation, and for infarct size measured at 120 min reperfusion (n = 48, 8 hearts/group); (2) isolated hearts as above but with removal of the heart at 5 min reperfusion to isolate cardiac mitochondria for ATP and ROS assays (n = 24, 4 hearts/group). A 5-min reperfusion point was selected to examine for maximal differences in mitochondrial function shortly after ischemia and to rule out any effects on these measures by nonviable cells. We have reported that ROS generated during IR and measured in coronary effluent are no longer detectable after 5 min reperfusion. 4 

The APC consisted of two 2-min periods of sevoflurane exposure delivered by vaporizer (3.2 vol%); these periods were separated by a 5 min-period of perfusion without sevoflurane. Sevoflurane was detected by gas chromatography (GC-8AIF, Shimadzu, Kyoto, Japan) as described previously. 2,4Inflow sevoflurane concentration was 0.42 ± 0.09 mm, which is equivalent to 2.98 ± 0.35 vol% for a minimal alveolar concentration (MAC) of approximately 1.42 ± 0.09. Sevoflurane was not detectable in the effluent at the end of the 15 min washout period before ischemia. Each scavenger of ROS was given beginning 5 min before sevoflurane, during sevoflurane exposures, and for 5 min after the second sevoflurane exposure. The scavengers were administered alone (without APC) to rule out any direct effects of these drugs.

All data were expressed as mean ± SE. Within group (time effect) for a given variable were compared to a baseline control (at 15 min) by Student-Newman-Keul test whenever univariate analysis of variance for repeated measures was significant (Super ANOVA 1.11 software for Macintosh from Abacus Concepts, Berkeley, CA). P < 0.05 (two tailed) was considered statistically significant.

There were no differences among groups in baseline functional variables: systolic or diastolic left ventricular pressure (LVP), dLVP/dt^max, dLVP/dt^min, coronary flow, heart rate, and myocardial oxygen consumption. This was true for hearts destined for functional studies and those destined for mitochondrial studies. In the time control group, no measured mechanical or metabolic variables changed over the 200 min of continuous perfusion (data not shown).

At 120 min reperfusion (200 min), developed LVP was reduced by 47 ± 2% compared to baseline in the APC group; this was significantly less than the decrease in the I/R group (62 ± 4%) (fig. 2). The improvement in developed LVP was due mostly to improved diastolic LVP as shown in the Table. This protection was completely abolished by either tiron or MnTBAP pretreatment during the preconditioning pulses. The Tablealso shows that dLVP/dt^maxand dLVP/dt^minwere improved in the APC group and that tiron or MnTBAP returned these levels to levels similar to those observed in the I/R group. For each of these variables, tiron or MnTBAP given without APC before ischemia had no effect different from those of the I/R group.

Both coronary flow and myocardial oxygen consumption improved significantly in the APC group on reperfusion and these effects were again abrogated by bracketing APC with either tiron or MnTBAP (Table). For each of these variables, tiron or MnTBAP given without APC before ischemia had no effect different from those of the I/R group.

DHE fluorescence intensity at 5 and 120 min reperfusion, respectively, was increased by 44 ± 5% and by 25 ± 6% above the baseline control in the I/R group, and by only 20 ± 3% and 4 ± 3% in the APC group. The non-APC groups did not differ significantly from the I/R group. Infarct size after 120 min reperfusion was decreased in the APC group compared to all other groups (P < 0.05) and there was no difference among the non-APC groups (fig. 3).

ATP synthesis and ROS formation were assessed in mitochondria isolated at 5 min reperfusion. Figure 4shows that mitochondrial ATP synthesis rate in Sham hearts, at the equivalent time-point as the standard (Sham, 11.2 ± 2.6 μmol ATP · min⁻¹· mg mitochondrial protein⁻¹, 100%), was reduced by 71 ± 3% (P < 0.05) in the I/R group, whereas APC returned ATP synthesis to near the Sham group values. Tiron or MnTBAP completely abrogated the effect of APC to restore ATP synthesis and neither tiron nor MnTBAP alone had any direct effect to alter ATP synthesis compared to the I/R controls. ATP concentration in the Sham group (6.7 nmol ATP · mg mitochondrial protein⁻¹) was not significantly different compared to the other groups.

Figure 5shows that ROS formation was increased by 48 ± 3% (P < 0.05) in the I/R group compared to the Sham group (a.u., 100%), but was not increased above the Sham group in the APC group. Tiron or MnTBAP given in the presence of APC completely abrogated the effect of APC on reducing ROS formation.

APC is manifested by attenuated cardiac injury consequent to IR in several animal models, 1,2,23and there is evidence of its occurrence in humans. 24Although APC is well known to improve global cardiac structure and function after ischemia, our experiments in isolated mitochondria demonstrate two mechanisms by which APC preserves mitochondrial energy production and mitochondrial function during early reperfusion after global ischemia. The APC phenomenon was evidenced on initial reperfusion not only by nearly normalized oxidative phosphorylation, but also by decreased ROS formation, to levels found in nonischemic hearts.

These results confirm and extend our previous observation in intact hearts that APC results in reduced coronary effluent ROS concentrations during initial reperfusion after ischemia. 4These studies, and our associated studies in intact hearts on the effects of APC on mitochondrial NADH and Ca²⁺, 11–13together indicate that APC leads to improved mitochondrial bioenergetics during IR. This suggests two key roles of mitochondrial–induced ROS formation in APC: 1) ROS are generated by mitochondria during anesthetic exposure leading to triggering of APC; 2) protection afforded by APC occurs, at least in part, because of a decrease in mitochondrial ROS formation during subsequent IR. Thus, the early component of ROS generation (likely a minute level) is involved in triggering the signaling pathways that induce APC. This subsequently leads to a decrease in the later component (the large excess of ROS normally formed during reperfusion after ischemia), effecting structural and functional preservation.

IR injury profoundly disrupts mitochondrial energy metabolism. It is well known that mitochondria isolated from hearts subject to IR exhibit decreased mitochondrial membrane potential (Δψ^m), ETC dysfunction, and ATP synthesis. 25,26IPC has been shown to improve mitochondrial electron transport function 27,28and ATP synthesis on reperfusion after ischemia, and it is suggested to open K^ATPchannels. 20,29In contrast, there is little information on the effect of APC on mitochondrial bioenergetics, although alterations in ROS generation, NADH concentrations and mCa²⁺have been documented. 4,12,13 

We postulate that there are common mitochondrial triggering mechanisms induced by IPC and by APC that lead to improved mitochondrial bioenergetics as an important factor in augmenting myocardial function after ischemia. We observed that a profoundly attenuated rate of ATP synthesis in mitochondria isolated shortly after global ischemia in nonpreconditioned hearts was effectively reversed by APC. We also observed that ATP synthesis rates in APC hearts were comparable to ATP synthesis rates in nonischemic hearts. Moreover, since we observed a significant reduction in ROS production in mitochondria isolated from preconditioned hearts, it is likely that APC also acts to attenuate the postischemic generation of mitochondrial ROS. Although we cannot exclude the possibility that APC also influences extra-mitochondrial ROS generating systems, the demonstrated role of the ETC in oxidant stress during ischemia, 30and our previous finding of normalized NADH 11and mitochondrial [Ca²⁺]12in sevoflurane-treated hearts, strongly suggests that the mitochondria is a key target for these protective effects. In support of this, it was evidenced nearly 30 yr ago that volatile anesthetics can inhibit NADH oxidoreductase. 31 

Our results suggest that reduced ROS formation is involved in mediating mitochondrial protection after global ischemia. Interestingly, our results also demonstrate a crucial role of ROS in triggering APC before global ischemia. This was demonstrated by the effects of two ROS scavengers given only briefly before, during and after APC, to abrogate, not only the protective effects on global function and infarct size, but also to reverse APC-induced mitochondrial protection. This was evidenced by return of ROS generation and depression of ATP synthesis to levels observed in nonpreconditioned hearts.

There is ample evidence that ROS are involved in triggering preconditioning. Exogenous administration of O^2·-32and H²O²33induces preconditioning. Furthermore, scavengers of ROS are known to inhibit preconditioning induced by various stimuli including brief ischemia 32and opioids. 34Our finding that the triggering of APC was inhibited by ROS scavengers is in agreement with our previous results 4and those of Mullenheim et al. , 5although the latter study was based indirectly on a ROS scavenging effect since ROS formation was not assessed.

We propose that O^2·-is generated from a mitochondrial ETC source during sevoflurane exposure, although one or more of its downstream products (e.g. , H²O²or hydroxyl radical) could be responsible for triggering the eventual preconditioning effect. We deliberately used two individual agents with highly different structures and ROS-depleting mechanisms. MnTBAP is a manganese containing porphyrin that acts as a superoxide dismutase mimetic, converting O^2·-to H²O²and H²O; however there is also evidence that it can act directly on H²O²35and peroxynitrite, 36both of which may have preconditioning effects. Tiron is a cell permeable, nonenzymatic ROS scavenger, also with a high specificity for O^2·-. 37Although it is an acceptable antioxidant, tiron does chelate a variety of metal ions and may interfere with metal ion binding of several intracellular proteins. 38,39We specifically utilized two mechanistically dissimilar but potent antioxidants with relative O^2·-selectivity to exclude possible non-O^2·-depleting effects of these agents that could result in APC attenuation or interference with ATP production. Either agent, when given with sevoflurane, had comparable effects on ROS production in mitochondria isolated during early reperfusion. Neither agent alone had direct effects on LVP, coronary flow, or infarction in agreement with previous reports that the drugs were washed out before the global ischemia. 40,41Indeed, our findings with MnTBAP and tiron support our earlier study that IR after APC led to reduced dityrosine fluorescence, an indicator of peroxynitrite which is primarily a product of O^2·-and nitric oxide (NO⁺). 4 

The mechanism by which ROS formed during the preconditioning phase leads to improved ATP synthesis on reperfusion is unclear. ROS are ubiquitous intracellular messengers that can initiate changes in mitochondrial function, or conversely, they can be formed as a result of altered mitochondrial function. During ischemia ROS, 42NADH and mitochondrial Ca²⁺accumulate 12,15,43and Δψ^mdecreases. 44The rise in mitochondrial Ca²⁺is likely secondary to a rise in cytosolic Ca²⁺consequent to inhibition of sarcolemmal Na⁺and Ca²⁺pumps, to Na⁺and Ca²⁺influx due to Na-H⁺exchange, and to reversed or slowed Na⁺-Ca²⁺- exchange. The decrease in Δψ^mfavors influx of Ca²⁺from the cytosol to the mitochondrial matrix; this could increase the probability of opening the membrane permeability transition pore that disables the mitochondrion by dissipating Δψ^m. 45 

Exogenous ROS (H²O²) are known to activate mitochondrial K^ATPchannels, 46possibly through activation of a specific isoform of protein kinase C (εPKC). 47This action could lead to a small depolarization of Δψ^mand a decrease in the electrochemical gradient for Ca²⁺entry. 48In support of this hypothesis is a reported role for the membrane permeability transition pore in preconditioning 49; preconditioning was shown to inhibit mitochondrial Ca²⁺loading and to inhibit the pore opening, and the putative mitochondrial K^ATPchannel opener diazoxide (a known preconditioning agent) was shown to induce a low conductance through the membrane permeability transition. 50,51 

Possible limitations of this study are that our Langendorff preparation is crystalloid perfused, and the presence of neutrophils and physiologic antioxidants normally present in blood might modify some of the responses we observed. Furthermore, our isolated heart preparation lacks autonomic influences, which could modify ischemia/reperfusion injury.

In conclusion, brief sevoflurane exposure improved mitochondrial bioenergetics and decreased mitochondrial ROS formation. This was associated with improved global cardiac function and reduced infarct size after ischemia. We demonstrated that APC leads to improved ETC function during IR, so that ATP availability is increased and ROS production is decreased, and it is tempting to speculate that it is these changes in mitochondria that underlie the global cardiac structural and functional preservation that is seen after reperfusion. The protective effects of APC were fully reversible by ROS scavengers given only during sevoflurane exposure. Our results support a model in which sevoflurane first acts on the mitochondrion via  a redox signaling mechanism, which may then elicit an effect to open mitochondrial K^ATPchannels by a ROS dependent mechanism. Knowledge of the mechanism of APC may help us to modify or mimic useful cardiac protective effects of anesthetics in the clinical setting.

The authors thank James Heisner, B.S., Research Technologist, Steve Contney, M.S., Research Associate, and Mary Ziebell, Research Technologist, for their technical assistance, and Mary Lorence-Hanke, Administrative Assistant, and Anita Tredeau, Administrative Assistant for their administrative assistance (all from the Department of Anesthesiology, Medical College of Wisconsin, Milwaukee, Wisconsin).