Differential Role of Calcium/Calmodulin-dependent Protein Kinase II in Desflurane-induced Preconditioning and Cardioprotection by Metoprolol: Metoprolol Blocks Desflurane-induced Preconditioning

All experimental procedures and protocols used in this investigation were reviewed and approved by the Animal Care and Use Committee of the local authorities (Government of Unterfranken, Würzburg, Germany) and conformed to the regulations of the German animal protection law. Furthermore, all conformed to the Guiding Principles in the Care and Use of Animals  of the American Physiologic Society and were in accordance with the Guide for the Care and Use of Laboratory Animals .9 

General preparation was performed as previously described.4Briefly, male New Zealand White rabbits were anesthetized with sodium pentobarbital (30 mg/kg intravenous bolus, followed by an infusion of 20–30 mg · kg⁻¹· h⁻¹) via  the left marginal auricular vein. Sodium pentobarbital was chosen because of its negligible effects on preconditioning.10No opioids or neuromuscular blocking agents were used throughout the investigation. Depth of anesthesia was verified by recurrent testing of palpebral reflexes and hind paw withdrawal throughout the experiment. After tracheotomy and tracheal cannulation, animals were artificially ventilated (Cicero®; Dräger, Lübeck, Germany) using positive pressure with an air and oxygen mixture (70%/30%). Arterial blood drawn from the auricular artery was analyzed using an ABL 505 blood gas analyzer (Radiometer, Copenhagen, Denmark), and blood gases were maintained within a normal physiologic range by adjusting the respiratory rate or tidal volume. End-tidal concentration of desflurane was measured at the tip of the endotracheal tube by an infrared anesthetic gas analyzer that was calibrated with known standards before and during experimentation. The rabbit minimum alveolar concentration (MAC) of desflurane used in the current investigation was 8.9%.11Left ventricular (LV) pressure and the maximum increase of LV pressure (+dP/dt^max) were measured with a saline-filled PE 50 polyethylene catheter inserted into the left ventricle via  the right carotid artery. Mean arterial pressure was monitored by insertion of a 2.5-French microtipped catheter (Millar Instruments Inc., Houston, TX) via  the right femoral artery into the descending aorta. Rectal body temperature was maintained at 38.5°± 0.5°C12by a servo-controlled heating pad (Föhr Instruments, Seeheim, Germany). After a left fourth thoracotomy and pericardiotomy, the left heart was exposed and suspended in a pericardial cradle. A silk ligature (2-0) was placed halfway between the base and the apex of the heart around a prominent branch of the left anterior descending coronary artery to form a snare. By tightening the snare, a coronary artery occlusion was produced, and reperfusion was instituted by loosening the snare. Each rabbit received 300 U/kg heparin 5 min before coronary artery occlusion for anticoagulation. Coronary artery occlusion was verified by epicardial cyanosis, regional dyskinesia in the ischemic zone, and electrocardiographic changes. Adequate reperfusion was confirmed by epicardial hyperemic response and reversion of electrocardiographic changes. Hemodynamic parameters, body temperature, and electrocardiogram were continuously recorded and analyzed using a personal computer (Hewlett Packard, Palo Alto, CA) and hemodynamic data acquisition and analysis software (Notocord® hem 3.5; Croissy sur Seine, France). Data were digitized at a sampling rate of 1,000 Hz.

The experimental protocol used in this investigation is illustrated in figure 1. Baseline systemic hemodynamics were recorded following a 30-min equilibration period after completion of instrumentation and calibration. All rabbits were subjected to 30 min of coronary artery occlusion followed by 3 h of reperfusion. Rabbits were randomly assigned to one of the study groups by opening a sealed envelope containing information about the study group after completion of the preparation of each animal. After 8 rabbits were randomized to each group and a preliminary data analysis had been performed, 2 more rabbits were randomized to each group to achieve a group size of 10 per group. To investigate the interaction between desflurane and metoprolol, rabbits received either vehicle (0.9% saline [control]), 1.0 MAC desflurane, or metoprolol at 0.2, 1.0, 1.75, or 2.5 mg/kg. Metoprolol at 0.2, 1.0, 1.75, and 2.5 mg/kg was coadministered in four separate groups with desflurane (n = 10 per group). Desflurane and metoprolol were administered continuously for 30 min and discontinued 30 min before coronary occlusion. In two separate groups, a bolus of KN-93 (300 μg/kg; Sigma-Aldrich, Munich, Germany) solved in dimethyl sulfoxide, a specific inhibitor of CaMK II, was administered directly into the left ventricle in the presence or absence of desflurane (n = 6 per group).

Infarct size and area at risk (AAR) were gravimetrically determined according to standard procedures.13Briefly, at the end of each experiment, the coronary artery was reoccluded and the AAR was determined by infusion of 2 ml patent blue (0.1 g/ml; Sigma-Aldrich, Taufkirchen, Germany). The rabbits were then killed with a lethal dose of pentobarbital, and the heart was rapidly excised. The heart was cut into five slices from apex to base, and the nonstained red myocardium (AAR) was separated from the nonischemic blue-stained LV normal areas. The samples of ischemic and nonischemic regions were incubated at 37°C for 20 min in 1% 2,3,5-triphenyltetrazolium chloride in 0.1 m phosphate buffer adjusted to pH 7.4. After overnight storage in 10% formaldehyde, infarcted (pale) and noninfarcted (brick-red) myocardium within the AAR was carefully separated and weighed. Infarct size was expressed as a percentage of the AAR. Rabbits with an AAR less than 15% of LV mass and those with intractable ventricular fibrillation or LV pump failure were excluded from the study. Infarct size was determined by an investigator blinded to the study protocol.

In a separate set of experiments, the effects of saline (control), 1.0 MAC desflurane, or 2.5 mg/kg metoprolol and KN-93 in the presence or absence of desflurane on CaMK II and phospholamban content and phosphorylation of phospholamban was investigated by Western immunoblotting. An interim analysis was performed after 5 animals per group had been studied. As a result, group size was adjusted to 7 animals per group to rule out a possible type I error (n = 7 per group). Five minutes after cessation of desflurane or metoprolol, the hearts were rapidly excised, and the left ventricle was shock frozen in liquid nitrogen and stored at −80°C until further treatment. In the KN-93 group, the hearts were excised at the corresponding time point. The samples were homogenized in ice-cold RIPA buffer (1× phosphate-buffered saline, 1% Igepal CA-630, 0.5% sodium deoxycholic acid, 0.1% sodium dodecylsulfate polyacrylamide, 20 mm sodium fluoride, 1 mm sodium orthovanadate, containing a protease-inhibitor cocktail [Roche, Grenzach-Wyhlen, Germany]) and centrifuged at 12,000g . Cytosolic and particulate cell fractions were left unseparated for further analysis. Proteins were separated on 15% polyacrylamide sodium dodecylsulfate polyacrylamide gels and subsequently transferred electrophoretically on nitrocellulose membranes (Whatman, Maidstone, United Kingdom). After transfer, nonspecific background was blocked using 5% nonfat milk powder in phosphate-buffered saline–Tween 20 (1 h at room temperature). Membranes were then incubated for 1 h at room temperature with the following specific antibodies: mouse antiphospholamban 1:3,000 (Affinity BioReagents, Golden, CO); mouse anti–glyceraldehyde-3-phosphate dehydrogenase 1:3,000 (Santa Cruz Biotechnology, Santa Cruz, CA); and rabbit anti–αB-crystallin (Assay Designs, Ann Arbor, MI) in 5% nonfat milk in phosphate-buffered saline–Tween 20 or incubated overnight (rabbit anti–p-phospholamban 1:1,000 (Badrilla, Leeds, United Kingdom) and goat anti–CaMK II 1:200 (Santa Cruz Biotechnology) at 4°C, respectively. Horseradish peroxidase–linked antibodies against mouse, rabbit, or goat 1:5,000 (Santa Cruz Biotechnology) were used as secondary antibodies. The protein bands were detected by ECL® detection reagent (GE Healthcare, Buckinghamshire, United Kingdom) and visualized using an x-ray film. The films were scanned and the band optical density was analyzed using Scan pack Software 3.0 (Bio-Rad Laboratories, Munich, Germany). Gel loading was normalized to αB-crystallin or glyceraldehyde-3-phosphate dehydrogenase expression, respectively.

Power analysis revealed a group size of n = 8 to detect a difference in means of 15% with a power of 0.8 at an α level of 0.05. Statistical analysis of hemodynamic data were performed using an overall 12 (control vs.  desflurane vs.  0.2 mg/kg metoprolol vs.  1.0 mg/kg metoprolol vs.  1.75 mg/kg metoprolol vs.  2.5 mg/kg metoprolol vs.  0.2 mg/kg metoprolol + desflurane vs.  1.0 mg/kg metoprolol + desflurane vs.  1.75 mg/kg metoprolol + desflurane vs.  2.5 mg/kg metoprolol + desflurane vs.  KN-93 vs.  KN-93 + desflurane) × 7 (baseline vs.  preconditioning vs.  memory vs.  coronary artery occlusion vs.  reperfusion 1 vs.  reperfusion 2 vs.  reperfusion 3) analysis of variance with repeated measures. In case of significant main effects of interactions, post hoc  one-way analyses of variance were conducted for each group and for each time. Statistical analysis for body weight, LV, AAR, AAR/LV, infarct size/AAR, and densitometry was performed using one-way analysis of variance (control vs.  desflurane vs.  0.2 mg/kg metoprolol vs.  1.0 mg/kg metoprolol vs.  1.75 mg/kg metoprolol vs.  2.5 mg/kg metoprolol vs.  0.2 mg/kg metoprolol + desflurane vs.  1.0 mg/kg metoprolol + desflurane vs.  1.75 mg/kg metoprolol + desflurane) and post hoc  Duncan test where appropriate. Statistical analysis of data was performed on a personal computer using SPSS 15.0 software (The Apache Software Foundation, Forest Hill, MD). All data are expressed as mean ± SEM.

One hundred thirty rabbits were instrumented to obtain 112 successful experiments (table 1). Three rabbits were excluded because of intractable ventricular fibrillation during the experimental protocol (one control, one 1.0 mg/kg metoprolol, and one 2.5 mg/kg metoprolol), 11 due to LV pump failure (one desflurane, two desflurane + 0.2 mg/kg metoprolol, two desflurane + 2.5 mg/kg metoprolol, two 0.2 mg/kg metoprolol, two 1.0 mg/kg metoprolol, one 1.75 mg/kg metoprolol, and one 2.5 mg/kg metoprolol), and 4 because LV AAR was less than 15% of the LV mass (one desflurane + 1.75 mg/kg metoprolol, one 1.0 mg/kg metoprolol, and two 2.5 mg/kg metoprolol).

There were no differences in hemodynamic parameters between experimental groups at baseline (table 2). Metoprolol given alone had no effect on mean arterial pressure, but decreased heart rate at doses of 1.0 mg/kg and higher. This reduction was sustained throughout the study protocol. Metoprolol had no effect on +dP/dt^maxat any dose given. Desflurane had no effect on heart rate. However, desflurane given alone or coadministered with different doses of metoprolol reduced mean arterial pressure, rate–pressure product, and +dP/dt^maxduring its application. After discontinuation of desflurane, mean arterial pressure and +dP/dt^maxreturned to baseline values. LV end-diastolic pressure was not affected by any of the study drugs. KN-93, given alone, reduced mean arterial pressure.

Myocardial infarct size (infarct size/AAR) was 60 ± 3% in control experiments. Desflurane at 1.0 MAC for 30 min and discontinued 30 min before coronary artery occlusion reduced infarct size to 33 ± 2%. The specific inhibitor of CaMK II, KN-93, had no effect on infarct size that was 55 ± 4% (n = 6). However, desflurane-induced preconditioning was blocked by KN-93 to an infarct size of 57 ± 3% (n = 6) (fig. 2). Metoprolol at 0.2 and 1.0 mg/kg did not affect infarct sizes that were 55 ± 3% and 53 ± 3%, respectively. Metoprolol at 1.75 and 2.5 mg/kg reduced infarct size to 48 ± 4% and 39 ± 5%. Desflurane-induced preconditioning was dose-dependently blocked by metoprolol at 0.2, 1.0, 1.75, and 2.5 mg/kg to infarct sizes of 46 ± 5%, 51 ± 3%, 52 ± 3%, and 55 ± 3%, respectively (n = 10 in all groups) (fig. 3).

Total CaMK II protein expression (fig. 4) and total phospholamban protein expression (fig. 5A) were not affected by any of the study drugs. Protein kinase A (PKA)–dependent phosphorylation of phospholamban at serine 16 was completely blocked by metoprolol but was left unaffected by desflurane and KN-93 (fig. 5B). Phosphorylation of phospholamban by CaMK II at threonine 17 was blocked by the specific CaMK II inhibitor KN-93 and by metoprolol, given alone or in combination with desflurane (fig. 5C). Desflurane given alone had no impact on the phosphorylation of phospholamban at threonine 17.

The first major finding of this study is that blockade of CaMK II activity by the specific inhibitor KN-93 abolishes desflurane-induced preconditioning. Ca²⁺is a pivotal second messenger in myocardial cells. Many of its intracellular actions are mediated by Ca²⁺/calmodulin-dependent protein kinases. CaMK II is the predominant isoform of Ca²⁺/calmodulin-dependent protein kinases in the heart and modulates voltage-gated L-type Ca²⁺channels and Ca²⁺release and uptake from the sarcoplasmatic reticulum via  sarcoplasmatic reticulum Ca²⁺adenosine triphosphatase and ryanodine receptors.6Therefore, CaMK II is important for the regulation of intracellular Ca²⁺homeostasis. Because cytosolic Ca²⁺overload is a major determinant of ischemia and reperfusion (I/R) injury,14a reduction of cytosolic Ca²⁺levels might beneficially affect the consequences of I/R injury. There is clear evidence that improved Ca²⁺handling during reperfusion is one mechanism of cardioprotection of preconditioning, because ischemic preconditioning prevents the depression in sarcoplasmatic reticulum Ca²⁺uptake and Ca²⁺pump adenosine triphosphatase activity triggered by I/R injury.8A recent study in isolated rat hearts demonstrated that blockade of CaMK II by KN-93 abrogated the beneficial effects of ischemic preconditioning on the recovery of LV contractile function after simulated ischemia.7Like ischemic preconditioning, sevoflurane-induced preconditioning reduces Ca²⁺overload and improves ventricular function.15The results of the current investigation now demonstrate that desflurane-induced preconditioning is abrogated by specific blockade of CaMK II by KN-93. These findings suggest that CaMK II is part of the signal transduction pathway of desflurane-induced preconditioning. Moreover, the results of the Western blot analysis demonstrated that CaMK II is effectively blocked by KN-93, because CaMK II–dependent phospholamban phosphorylation at threonine 17 was inhibited, whereas PKA-dependent phospholamban phosphorylation at serine 16 was not.16In contrast, metoprolol blocked phospholamban phosphorylation at both sites. It is possible that this difference in inhibition of phosphorylation sites of phospholamban accounts for the differences in cardioprotection between desflurane and metoprolol observed in the current study. While CaMK II inhibition and hence phosphorylation of phospholamban at threonine 17 blocks desflurane-induced preconditioning, the effective inhibition of CaMK II and PKA and hence phosphorylation of phospholamban at both sites might be necessary to confer cardioprotection by metoprolol. This contribution is contradictory only at first sight. It should be borne in mind that long and sustained β-adrenergic stimulation, as occurs in chronic heart failure, results in cardiomyocyte apoptosis induced by PKA-independent CaMK II activation and not by the classic cyclic adenosine monophosphate–PKA pathway.17,18Furthermore, increased CaMK II activity after myocardial infarction or from excessive β-adrenergic stimulation results in maladaptive remodeling and arrhythmias.19This observation can be explained by the finding that during reperfusion phospholamban is mainly phosphorylated by CaMK II.20Phospholamban phosphorylation results in increased activity of sarcoplasmatic reticulum Ca²⁺adenosine triphosphatase with consecutively higher Ca²⁺load in and release from the sarcoplasmatic reticulum with subsequent cytosolic Ca²⁺overload. Because ablation and phosphorylation of phospholamban exert the same effects on sarcoplasmatic reticulum Ca²⁺adenosine triphosphatase, phospholamban-deficient mice exhibit increased ischemic injury compared with wild-type mice.21Therefore, the prevention of phospholamban phosphorylation by metoprolol might contribute to its cardioprotective effects during reperfusion. On the other hand, short-lived β-adrenergic stimulation is known to exert positive effects on I/R injury and to induce preconditioning.22,23Therefore, CaMK II is necessary to induce ischemic and anesthetic preconditioning, while sustained and long-termed blockade of CaMK II by β-adrenergic receptor blockers might prevent the deleterious effects of adrenergic stimulation and hence prolonged CaMK II activation.

The second major finding of this study is that metoprolol reduced infarct size only, when the highest doses of metoprolol of 1.75 and 2.5 mg/kg were given before I/R. Metoprolol at 1.0 mg/kg did not confer cardioprotection. In another study using the rabbit myocardial infarction model, metoprolol at a dose of 1.0 mg/kg effectively reduced infarct size.24However, in that investigation, metoprolol was not given before ischemia but only during reperfusion. This is likely to result in higher plasma levels during reperfusion, achieved in the current protocol only at doses of metoprolol exceeding 1.0 mg/kg. As a third major finding, the current results further demonstrate that metoprolol dose-dependently attenuated or blocked desflurane-induced preconditioning. This finding is in accord with previously published investigations. Desflurane-induced improvement of the force of contraction after simulated ischemia is abolished by the nonselective β-adrenergic receptor blocker propranolol.5Moreover, β¹-selective receptor blockade by esmolol or inhibition of the downstream PKA by H-89 abolishes desflurane- and sevoflurane-induced preconditioning against myocardial infarction in the rabbit heart in vivo .4However, whether β-adrenergic blockade interferes with anesthetic-induced preconditioning in the clinical setting has not yet been demonstrated. The mechanism underlying the observation that metoprolol-induced infarct size reduction is not maintained in the presence of desflurane cannot fully be elucidated by the current results. However, cardioprotection by β-blockers aims at the prevention of ischemic episodes that is thought to be conferred by an improved myocardial oxygen supply and demand ratio, mainly induced by heart rate reduction.25It is possible that metoprolol-induced infarct size reduction after experimental ischemia depends on alternative mechanisms. This hypothesis is corroborated by the fact that in this and other studies on β-blocker–induced infarct size reduction26or restoration of LV function after ischemia,27beneficial effects of β-blockers were independent of heart rate reduction. Alternative effects of β-adrenergic receptor blockers include CaMK II inhibition, as demonstrated in this investigation, membrane stabilizing effects,28,29inhibition of phospholipase A,30and the inhibition of β oxidation of fatty acid accumulation.31These alternative effects of metoprolol might be impaired by the effects of desflurane-induced preconditioning. However, further investigations are needed to elicit the exact mechanisms underlying the infarct-size reducing properties of metoprolol and the resulting interaction with desflurane.

The current results should be interpreted within the constraints of several potential limitations. The LV AARs for myocardial infarction and coronary collateral blood flow are important determinants of myocardial infarct size. However, AAR was not different among experimental groups (table 3). Coronary collateral blood flow was not determined in this study. However, coronary collateral blood flow is negligible in rabbits.32Therefore, neither differences in AAR nor collateral blood flow can be attributed to differences in infarct size. Metoprolol reduced infarct size only at higher doses compared with doses used clinically in humans. The use of these high doses might explain the occurrence of LV pump failure in 10 of the rabbits receiving metoprolol. However, metoprolol at 0.2 mg/kg, a dose comparable to that used in humans, did not reduce infarct size or heart rate throughout ischemia and reperfusion. Heart rate reduction was used as a physiologic marker of effective β-adrenergic blockade, and this was only achieved at high doses. Therefore, as in the study by Feuerstein et al. ,24high doses of metoprolol had to be used to induce the intended effect of reducing heart rate. It has been demonstrated that I/R injury induces compartmentalization of CaMK II during transient ischemia with subsequent redistribution during reperfusion.33CaMK II translocation was not tested in our study. However, it is possible that CaMK II translocation might play a role in desflurane-induced preconditioning or cardioprotection by metoprolol. KN-93 is a selective competitive inhibitor of calmodulin binding to CaMK II and CaMK II autophosphorylation and has no inhibitory effect on other protein kinases.34However, KN-93 is known to directly inhibit L-type Ca²⁺channels and K⁺currents.35It cannot be excluded that inhibition of these ion fluxes might have influenced the results.

In conclusion, the results of the current study demonstrate that desflurane-induced preconditioning is mediated by CaMK II–dependent phospholamban phosphorylation, whereas metoprolol-induced cardioprotection during reperfusion is mediated by the blockade of PKA– and CaMK II–dependent phosphorylation of phospholamban. Furthermore, metoprolol reduced infarct size only at high doses but dose-dependently abrogated desflurane-induced preconditioning. These results suggest important negative interactions between desflurane-induced preconditioning and β-adrenergic receptor blockade that warrant further investigations on the underlying mechanisms.