Choice of Anesthetic Combination Determines Ca²⁺Leak after Ischemia-Reperfusion Injury in the Working Rat Heart: Favorable versusAdverse Combinations

• Volatile anesthetics, propofol, and opioids may elicit cardioprotection in hearts exposed to ischemia–reperfusion injury

• The choice of anesthetic combination determines the postischemic sarcoplasmic Ca²⁺leak and intracellular Ca²⁺overload after ischemia–reperfusion injury in the working rat heart

META-analyses on the antiischemic cardioprotective properties of general anesthetic agents in high-risk surgical patients remain elusive. Even for volatile anesthetics, a highly promising class of cardioprotective drugs, the currently available evidence is still controversial.1Some studies show impressive postischemic functional improvement and favorable changes in biomarkers associated with cardioprotection,2,3whereas other studies could only find mediocre or no signs of protection.4Nonetheless, the American Heart Association recommends in its current Guidelines for Perioperative Management of High-Risk Surgical Patients Undergoing Noncardiac Surgery the preferential use of volatile anesthetics in these patients.5 

In sharp contrast to these disparate clinical findings, there is overwhelming evidence of cardioprotection in the experimental setting, not only by volatile anesthetics but also by other anesthetics6,7such as propofol8and opioid receptor agonists.9We and others have previously speculated that this discrepancy between experimental and clinical observations might be explained, in part, by the fact that diseased diabetic and/or infarct-remodeled hearts of patients may be less amenable to pharmacologic protection against ischemia–reperfusion (IR) injury.10Another possible explanation could be potential antagonistic interactions between individual anesthetics when used in combination resulting in a loss of cardioprotection. In fact, it is well known that some frequently used perioperative drugs enhance (nitroglycerin, statins), whereas others abolish (sulfonylureas, cyclooxygenase-2 inhibitors, β-adrenoceptor antagonists) the most effective endogenous mechanism of cardioprotection, namely preconditioning.11One recent study in rabbit hearts demonstrated inhibitory effects of propofol on desflurane-induced preconditioning,12whereas another study showed that morphine could enhance cardioprotection by isoflurane preconditioning.13However, so far no study has systematically explored the interaction between common clinically used anesthetics on cardioprotection. The question of whether and how different combinations of anesthetics might antagonize or potentiate their effects on cardioprotection is essential for the clinician, because a general anesthetic procedure consists of a mixture of drugs, and the most favorable cardioprotective combination of anesthetics should be preferentially used in at-risk patients. To gain insight into such interactions, we designed a study using a standardized experimental model, i.e. , the working rat heart, in which the net effects of three combinations of anesthetics, namely sevoflurane, propofol, and remifentanil, were examined on the recovery of postischemic left ventricular work. In addition, to obtain a mechanistic understanding of potential interactions, we used this experimental model to measure drug-induced alterations in Ca²⁺overload and sarcoplasmic Ca²⁺leak. We hypothesized that certain combinations would show antagonism, whereas others might show synergy in cardioprotection.

The investigation conforms to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85–23, revised 1996) and the experimental protocol used in this investigation was approved by the University of Alberta Animal Policy and Welfare Committee (Edmonton, Alberta, Canada).

One hundred eleven male Sprague-Dawley rats (350–400 g) were anesthetized with pentobarbital (150 mg/kg, intraperitoneally). Hearts were rapidly removed and perfused initially in a nonworking Langendorff mode with Krebs-Henseleit solution for 15 min. The working mode perfusion was established (11.5 mmHg preload, 80 mmHg afterload, 5 Hz) with a recirculating perfusate of 100 ml (37°C, pH 7.4) gassed with 95% O²/5% CO²mixture that consisted of a modified Krebs-Henseleit solution containing (mM): KCl (4.7), NaCl (118), KH²PO⁴(1.2), MgSO⁴(1.2), CaCl²(2.5), NaHCO³(25), glucose (11) palmitate (1.2, prebound to 3% bovine serum albumin) and insulin 100 mU/l. Hearts were paced at 5 Hz.

Cardiac output (ml/min) and aortic flow (ml/min) were measured using ultrasonic flow probes (Transonic T206, Transonic Systems Inc., Ithaca, NY) placed in the left atrial inflow and the aortic outflow lines.14Left ventricular work (ml/min · mmHg) was calculated as left ventricular work = cardiac output · (aortic systolic pressure - preload). Coronary flow (ml/min) was calculated as the difference between cardiac output and aortic flow.

The measurements of intracellular [Ca²⁺] ([Ca²⁺]ⁱ) were performed as described previously.15The hearts were loaded with 5 μM fluorescent Ca²⁺indicator indo-1 AM (TEFLabs, Austin, TX) dissolved in dimethyl sulfoxide (final concentration less than 0.25%) for 25 min during the working mode perfusion. After a washout period of 5 min in the Langendorff mode, the working mode was reestablished with fresh modified Krebs-Henseleit solution using a second perfusion circuit and the hearts were assigned to the treatment groups. Indo-1 fluorescence was measured from a small area (approximately 0.3 cm²) of the epicardial surface of the left ventricle free wall using a spectrofluorometer (Photon Technology International, London, Ontario, Canada) (see Supplemental Digital Content 1, https://links.lww.com/ALN/A814). Signals were acquired at 500 Hz, and the ratio of indo-1 fluorescence emitted at 405 nm and 485 nm (F405/F485 ratio) was calculated to provide an index of [Ca²⁺]ⁱ. Diastolic [Ca²⁺]ⁱ, systolic [Ca²⁺]ⁱ, and Ca²⁺transient amplitude were determined.

After 15 min of perfusion in the working mode, all hearts were subjected to 20 min of 37°C zero-flow ischemia and 30 min of reperfusion (see Supplemental Digital Content 1, https://links.lww.com/ALN/A814). All treatments were initiated 15 min before ischemia and continued until the end of reperfusion. The duration of 20 min of global ischemia induced marked left ventricular dysfunction. Longer periods of global ischemia would have resulted in no functional recovery at reperfusion, which would have biased measurements of [Ca²⁺]ⁱ. Hearts were assigned to one of the following 15 groups: (1) time-matched perfusion without treatment (IR) (n = 8), (2) sevoflurane (Abbott Canada, Saint-Laurent, Québec, Canada) (SEVO+IR) 2 vol.-% (n = 8), (3) propofol 10 μM (n = 6) (PROP+IR) in the formulation of Diprivan® 1% (AstraZeneca Canada, Mississauga, Ontario, Canada), (4) remifentanil (Abbott Canada) 3 nM (REMI+IR) (n = 9), (5) N -2-mercaptopropionyl-glycine (MPG, Sigma-Aldrich, Oakville, Ontario, Canada) 10 μM (MPG+IR) (n = 5), (6) sevoflurane 2 vol.-% plus propofol 10 μM (SEVO/PROP+IR) (n = 9), (7) remifentanil 3 nM plus propofol 10 μM (REMI/PROP+IR) (n = 7), (8) sevoflurane 2 vol.-% plus remifentanil 3 nM (SEVO/REMI+IR) (n = 9), (9) sevoflurane 2 vol.-% plus MPG 10 μM (SEVO/MPG+IR) (n = 5), (10) sevoflurane 2 vol.-% plus Intralipid 10% (Sigma-Aldrich) (SEVO/INTRA+IR; corresponding to Intralipid control of propofol 100 μM) (n = 6), (11) remifentanil 3 nM plus MPG 10 μM (REMI/MPG+IR) (n = 6). Additional groups (12–14) were studied to determine the concentration-response relationship for the sevoflurane 2 vol.-% - propofol interaction (0.1, 1, and 100 μM) (n = 7 for each concentration) and (15) to test whether Intralipid alone affects functional recovery or Ca²⁺overload after IR (n = 6). Hearts with time-matched aerobic perfusion served to determine baseline phosphoprotein levels (n = 6). At the end of the perfusions, all hearts were immediately frozen in liquid nitrogen with Wollenberger clamps and stored at −80°C for subsequent molecular analyses.

Tissue homogenate was used for immunoblotting of calcium/calmodulin-dependent protein kinase type IIδ (CaMKIIδ) in 8% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and for phospholamban in 10% SDS-PAGE. The following primary antibodies were used: rabbit anti-CaMKIIδ (1:500) (Abcam, Cambridge, MA), rabbit antipThr286-CaMKIIδ (1:1000) (Abcam), antiphospholamban (1:500) (Abcam), antipThr17-phospholamban, and antipSer16-phosphlamban (1:300) (Santa Cruz Biotechnology, Inc., Santa Cruz, CA). For ryanodine receptor-2 (RyR2) immunoblots, sarcoplasmic reticulum fractions were prepared. Frozen heart tissue was homogenized in ice-cold buffer containing 10 mM Hepes (pH 7.4), 1.5 mM MgCl², 10 mM KCl, 0.3 M sucrose, and 0.05% Nonidet P-40 and centrifuged at 500 g for 10 min at 4°C. The resulting supernatant was centrifuged at 100,000 g for 45 min and the obtained pellet was suspended in buffer containing 50 mM Tris (pH 7.4), 2 mM EDTA, 0.15 M NaCl, 1% Nonidet P-40, 0.1% SDS, and 5 mM 1,4-dithiothreitol supplemented with protease and phosphatase inhibitor cocktail mix (Sigma-Aldrich). There were 60 μg of protein subjected to electrophoresis on a 5% gel and the following primary antibodies anti-RyR2 (1:1000) (Abcam) and antipSer2814-RyR2 (1:1000) (Badrilla Ltd., Leeds, United Kingdom) were used. Protein concentration was measured (Bradford assay) to ensure that equal amounts of protein were subjected to electrophoresis. Immunoreactivity was visualized by horseradish peroxidase-conjugated antibodies using a peroxidase-based chemiluminescence detection kit (ECL) (PerkinElmer, Woodbridge, Ontario, Canada). The intensity of the bands was quantified using ImageJ**software. Phosphorylation was normalized to the expression of the corresponding total protein. Actin (I-19; Santa Cruz) and/or α-tubulin (clone B-5–1-2; Sigma-Aldrich) were used as loading controls.

Values are given as mean (SD) or median (interquartile range) except for figures 1and 2where standard errors of the mean (SEM) are given to avoid confusion with multiple overlapping lines. With an expected difference of 50% between group means, 20% SD of the means, significant levels of α = 0.05 and β = 0.8, a sample size of five hearts per group was necessary.15Time-dependent measurements were analyzed by two-way repeated measures analysis of variance (ANOVA) followed by multiple comparison procedures (Student-Newman-Keuls method). The significance of differences in hemodynamic and calcium measurements among groups was determined by unpaired two-tailed Student t  test (two groups) or by one-way ANOVA followed by the Holm-Sidak method or by Kruskal-Wallis one-way ANOVA on ranks followed by the Dunn method for post hoc  analysis, depending on the underlying data distribution. Differences are considered significant if P < 0.05. Statistical analyses were performed using Sigmaplot Version 11.0 (Systat Software, Inc., Chicago, IL).

Periischemic sevoflurane (2 vol.-%) and remifentanil (3 nM) treatment similarly improved recovery of left ventricular work during reperfusion when compared with untreated hearts (fig. 1). These beneficial actions were accompanied by higher peak systolic pressures, stroke volumes, and coronary flows (see Supplemental Digital Content 2, https://links.lww.com/ALN/A815). Interestingly, the amplitudes of the Ca²⁺transients in hearts exposed to remifentanil (3 nM) were lower, but left ventricular work was higher than in untreated hearts; the lack of positive correlation suggests that remifentanil prevented ischemia-induced Ca²⁺desensitization. Remifentanil and sevoflurane comparably reduced postischemic diastolic [Ca²⁺]ⁱ, a sensitive marker of ischemic damage. Peak ischemic [Ca²⁺]ⁱwas lower in remifentanil-treated than sevoflurane-treated hearts (P = 0.008). Consistent with previous results,15treatment with propofol 10 μM did not improve functional recovery or decrease postischemic Ca²⁺overload. Intralipid alone or in combination with sevoflurane did not affect functional recovery or Ca²⁺overload (see Supplemental Digital Content 3, https://links.lww.com/ALN/A816).

Concomitant administration of propofol 10 μM with sevoflurane 2 vol.-% did not reduce left ventricular work under aerobic conditions, but completely abolished sevoflurane-mediated improvement in the recovery of postischemic left ventricular work (fig. 2). Peak ischemic [Ca²⁺]ⁱand postischemic Ca²⁺transient amplitudes were not altered by the combination treatment when compared with sevoflurane alone. However, the addition of propofol to sevoflurane prevented the beneficial effects of sevoflurane on postischemic diastolic [Ca²⁺]ⁱand values were increased to levels measured in untreated hearts. In contrast, administration of propofol 10 μM with remifentanil 3 nM did not interfere with remifentanil-induced cardioprotection. Examination of the concentration-response relationship for propofol-mediated inhibition of sevoflurane-mediated protection indicates that 1 μM propofol reduced functional recovery elicited by sevoflurane by 33% (P = 0.004) (fig. 3A) and increased diastolic [Ca²⁺]ⁱby 6% compared with protected hearts (fig. 3, B and C). Only the lowest concentration of propofol (0.1 μM) did not diminish sevoflurane-induced protection. Our demonstration that the different combinations of anesthetics exhibit disparate interactions is likely due to distinct mechanisms of cardioprotection.

To demonstrate the ability of MPG, a known scavenger of reactive oxygen species (ROS) similar to propofol, to mimic the antagonistic actions of propofol, MPG 10 μM was concomitantly administered to sevoflurane 2 vol.-%. Although MPG 10 μM alone did not affect postischemic functional recovery or Ca²⁺overload compared with unprotected hearts, MPG completely abolished sevoflurane-mediated protection similar to propofol (fig. 4). This confirms previous experiments showing the essential role of ROS in sevoflurane preconditioning.16,17Finally, when remifentanil 3 nM was added to sevoflurane 2 vol.-%, functional recovery was similar to sevoflurane alone or to remifentanil alone. Unexpectedly, this combination markedly increased peak ischemic [Ca²⁺]ⁱ(P < 0.001) and Ca²⁺transient (P = 0.02) compared with remifentanil alone, but did not increase postischemic Ca²⁺overload (fig. 2B). Coadministration of MPG to remifentanil did not block remifentanil-induced cardioprotection (see Supplemental Digital Content 4, https://links.lww.com/ALN/A817).

Because CaMKIIδ activity is known to play a causal role in postischemic Ca²⁺overload and Ca²⁺leak from the sarcoplasmic reticulum,18,19CaMKIIδ phosphorylation at Thr286 was determined in aerobic hearts as well as in postischemic hearts that were either untreated or that were exposed to single and combination treatments. These experiments show that CaMKIIδ phosphorylation was very low in aerobic hearts, but was significantly increased in untreated postischemic hearts. In hearts exposed to cardioprotective single or combination treatments, CaMKIIδ phosphorylation was inhibited, whereas unprotected hearts exhibited markedly higher phosphorylation levels (fig. 5). Similarly, examination of ratios of phospho-RYR2 to total RYR2 indicates that RYR2, a downstream target of CaMKIIδ, was significantly more phosphorylated in the nonprotective combination treatment, sevoflurane plus propofol, as opposed to the protective combo-treatments, sevoflurane plus remifentanil, or propofol plus remifentanil (fig. 6). Phosphorylation of phospholamban at position Thr17, which is selectively mediated by CaMKIIδ, and at position Ser16, which is selectively mediated by protein kinase A, was not different between unprotected and protected hearts (see Supplemental Digital Content 5, https://links.lww.com/ALN/A818) implying that the postischemic sarcoplasmic Ca²⁺leak in this model is not caused by phospholamban phosphorylation but is due to selective phosphorylation of RyR2 by activated CaMKIIδ.

Our study systematically investigated the interactions of commonly applied classes of anesthetics on cardioprotection. For this purpose, the isolated working rat heart model perfused with glucose and palmitate as metabolic substrates was used to measure left ventricular work and beat-to-beat oscillations in intracellular Ca²⁺during IR injury. The anesthetics or their combinations were administered before, during, and after the ischemic insult and therefore mimic clinical protocols. We report the following salient findings. First, when the anesthetics were administered as single treatments remifentanil or sevoflurane, but not propofol, markedly improved the recovery of left ventricular mechanical function and decreased diastolic Ca²⁺overload. This is consistent with previous findings showing no protection with propofol at concentrations as high as 100 μM.12,15Second, in the combination treatments propofol had no effect on remifentanil-induced cardioprotection, but antagonized sevoflurane-induced cardioprotection in a concentration-dependent manner. Antagonism was clearly demonstrable at low sedative concentrations of propofol (1 μM). Conversely, the combination of the two protective drugs, sevoflurane (2 vol.-%) and remifentanil (3 nM), preserved protection but did not show additive or synergistic interactions. Finally, our experiments demonstrate that postischemic diastolic Ca²⁺overload in the myocardium exposed to the various anesthetic combinations correlates closely with CaMKIIδ activation and Ca²⁺leak from the sarcoplasmic reticulum as reflected by hyperphosphorylation of the RyR2. The clinical consequences of such interactions, particularly the antagonistic action of low concentrations of propofol on sevoflurane-induced protection, need careful evaluation.

Ca²⁺is vital for modulating cellular functions, serves as activator of the contractile apparatus, and regulates the release and uptake of Ca²⁺from the sarcoplasmic reticulum.20At higher intracellular concentrations, however, Ca²⁺is highly toxic and impairs left ventricular mechanical function and causes irreversible cell injury and death. It is therefore a sensitive and reliable marker of cellular dysfunction during and after IR injury. In quiescent cardiomyocytes, cytosolic Ca²⁺concentrations reach approximately 1−2 × 10⁻⁷M, which is 10,000-fold lower compared with the extracellular space. Elevation of average intracellular Ca²⁺concentration enhances binding of Ca²⁺to calmodulin, the intracellular Ca²⁺sensor, which subsequently interacts with more than 100 proteins, one of which is CaMKIIδ, the predominant isoform of CaMKII expressed in the heart.20The Ca²⁺-calmodulin complex promotes CaMKIIδ autophosphorylation at Thr286, which in turn enhances its affinity to Ca²⁺-calmodulin ensuing additional autophosphorylation at Thr305 and Thr306. Phosphorylated CaMKIIδ critically affects the release of Ca²⁺from the sarcoplasmic reticulum via  phosphorylation of RyR2 at Ser2814. CaMKIIδ further regulates sarcoplasmic Ca²⁺uptake by phospholamban phosphorylation at Thr17, diminishing its inhibitory effects on sarcoplasmic reticulum Ca²⁺ATPase, cardiac isoform (SERCA2a), which pumps Ca²⁺from the cytoplasm back into the sarcoplasmic reticulum against a 10,000-fold concentration gradient.20Thus, measurements of the phosphorylation status of CaMKIIδ and its downstream targets RyR2 and phospholamban serve as useful indices of the mechanisms that regulate Ca²⁺leak and Ca²⁺sequestration, respectively. Our finding that CaMKIIδ phosphorylation was increased during reperfusion suggests that the reduced left ventricular work and enhanced intracellular Ca²⁺overload occurred in response to CaMKIIδ-mediated facilitation of Ca²⁺release from the sarcoplasmic reticulum by the RyR2. In support, increased CaMKIIδ activity has been reported in failing human hearts21and rabbit hearts after coronary artery ligation.22Multiple studies have also shown a close relation between CaMKIIδ activity and apoptosis.23,24Moreover, CaMKIIδ activation per se  exerts direct detrimental effects. Transgenic overexpression of CaMKIIδ causes hyperphosphorylation of RyR2 with subsequent diastolic Ca²⁺leakage from the sarcoplasmic reticulum, leading to arrhythmias or heart failure.25In contrast, inhibition of CaMKIIδ, by transgenic overexpression of an inhibitory peptide, abolishes cardiac deterioration after infarction.26Despite the differential CaMKIIδ activation in protective versus  nonprotective anesthetic combinations in our experiments, phospholamban phosphorylation was not different among the treatments. This is similar to previous observations in rat cardiomyocytes where selective hyperphosphorylation of the RyR2 at Ser2814 without simultaneous alterations in phospholamban phosphorylation was shown.27Collectively, CaMKIIδ activation with subsequent hyperphosphorylation of RyR2 reflects a common feature of impaired postischemic Ca²⁺handling in hearts exposed to nonprotective anesthetic combinations. On the other hand, protective anesthetic combinations prevent hyperphosphorylation of RyR2 and thereby enable improved Ca²⁺homeostasis and enhanced functional recovery after IR.

Sevoflurane- and opioid-induced cardioprotection share many signaling components such as PKC, mitochondrial K^ATPchannels, and ROS.17,28While ROS may be essential in sevoflurane-induced protection,16,17other mechanisms, such as δ- and κ-opioid or N -methyl-D-aspartate receptor activation may be more important in remifentanil protection.9,29It is noteworthy that Ca²⁺transient amplitudes in remifentanil-treated hearts were lower than in sevoflurane-treated hearts even during preischemic aerobic conditions. This points to additional differences in the mechanism of action of these two agents on Ca²⁺homeostasis and Ca²⁺sensitivity, and suggests that their cardioprotective mechanisms may be distinct. This dissociation between Ca²⁺transient amplitude and cardiac contractility, indicative of enhanced Ca²⁺sensitization, was published previously based on studies in isolated rat cardiomyocytes—a model where contractility was not measured,30and isolated perfused rat hearts,31where remifentanil had marked positive inotropic action. Together, these data indicate that there is an increase in Ca²⁺sensitization in the presence of remifentanil. This is consistent with less Ca²⁺loading and may contribute to its cardioprotective effectiveness. The participation of ROS in sevoflurane-induced but not remifentanil-induced protection is supported by the observation that the ROS scavenger MPG, with properties similar to propofol, markedly reduces the salutary effects of sevoflurane but not remifentanil. This mechanism also explains the lack of inhibitory effects of the ROS scavenger propofol on remifentanil-induced protection. Although a previous study showed inhibitory effects of propofol on permeability transition opening in mitochondria during reperfusion,8which is likely due to the scavenging effects of propofol, we and others were unable to demonstrate any protection by propofol, even at high concentrations.12,15 

According to the commonly accepted threshold theory of preconditioning, which implies that a certain degree of stimulation is required to reach the level where a cell or organ is able to activate its endogenous protection program, it would be conceivable that the application of two well-defined protection stimuli should induce a more consistent and reliable overall cell protection. Indeed, morphine enhances isoflurane preconditioning and postconditioning in hearts subjected to coronary artery ligation in vivo .13,32However, in our experiments, we did not observe any synergistic or additive protection in hearts treated with a combination of sevoflurane (2 vol.-%) and remifentanil (3 nM). It is possible that the maximum preconditioning trigger stimulus has been already reached with 2 vol.-% sevoflurane alone or 3 nM remifentanil alone, making the protection elicited by the second drug redundant. In addition, we cannot exclude that a more sustained ischemic injury might uncover additive cardioprotective effectiveness in hearts exposed to a sevoflurane-remifentanil combination. Nonetheless, 2 vol.-% sevoflurane and 3 nM remifentanil are relatively low concentrations compared with those previously used to elicit effective cardioprotection.3,9In addition, the extent of improvement in functional recovery and attenuation of Ca²⁺overload in our working rat heart model by sevoflurane or remifentanil still left room for further improvements.

The potential for interactions between different protection strategies including pharmacologic agents and/or ischemic conditioning has been largely neglected in experimental and more so clinical studies in the past. However, the question of whether and how different anesthetic agents might interact to either antagonize or potentiate the beneficial effects of ischemia- or drug-induced cardioprotection is of great importance for the clinician, because a general anesthetic procedure consists of a mixture of drugs, and the most favorable cardioprotective combination of anesthetics should be preferentially used in at-risk patients. Our results from this study in rat hearts imply that the outcome in cardioprotection for combined therapies is highly context-sensitive and cannot be inferred or extrapolated from the assessment of single cardioprotective therapies. Indeed, our results suggest that the lack of standardization in the use of anesthetics (choice of combination, timing, and concentrations) may have contributed to ambiguous and contradictory results arising from both basic and clinical studies concerning anesthetic-induced cardioprotection, as well as preconditioning or postconditioning.

In conclusion, this study shows that the choice of anesthetic combination determines the functional recovery and postischemic Ca²⁺homeostasis in working rat hearts. We anticipate that results from these experiments may encourage the careful design of clinical studies so that perioperative cardioprotection can be optimized.