Reactive Oxygen Species Precede Protein Kinase C-δ Activation Independent of Adenosine Triphosphate–sensitive Mitochondrial Channel Opening in Sevoflurane-induced Cardioprotection

The study was performed in accordance with the guidelines of the Institutional Animal Care and Use Committee of the VU University Medical Center (Amsterdam, The Netherlands). Male Wistar rats (weight, 250–400 g; Harlan Netherland, Horst, The Netherlands) were anesthetized with sodium pentobarbital (80 mg/kg intraperitoneally; Ceva Sante Animale BV, Maassluis, The Netherlands). After 1,000 U heparin (Leo Pharma, Breda, The Netherlands) was intravenously injected, the heart was quickly removed and subsequently retrogradely perfused through the aorta with Tyrode buffer (95% O²and 5% CO², pH 7.4) at room temperature. Standard solution contained 120 mm NaCl, 1.22 mm MgSO⁴.7H²O, 1.99 mm NaH²PO⁴, 27.0 mm NaHCO³, 5.0 mm KCl, 1 mm CaCl², and 10 mm glucose. During dissection, 30 mm 2,3-butanedione monoxime and 15 mm KCl was added to the solution to protect the myocardium. 24Subsequently, a suitable right ventricular trabecula (length, 2–5 mm; diameter, < 0.2 mm), 25running from the free ventricle wall to the tricuspid valve, was carefully dissected.

Muscles were mounted between a force transducer (AE801; SensoNor, Horten, Norway) and a micromanipulator in an airtight organ bath. After mounting, superfusion was started with normal Tyrode buffer at 27°C with a flow of 2 ml/min, and trabeculae were stimulated with two platinum field electrodes (duration, 5 ms; stimulation frequency, 0.5 Hz). The stimulation voltage was set at two times the stimulation threshold. Subsequently, the length of individual trabeculae was set at 95% of the maximal length as determined by a force–length relation, followed by a stabilization period of 40 min. The stimulation frequency and temperature were then decreased to 0.2 Hz and 24°C, respectively, followed by another 20 min of stabilization. After this stabilization period, the initial developed contractile force (F^dev,start) and maximal force (F^max,start), as determined by a postextrasystolic potentiation protocol, were recorded. Postextrasystolic potentiation determines the contractile reserve of trabeculae by maximally filling the sarcoplasmic reticulum with Ca²⁺. 26Trabeculae that did not stabilize, spontaneously contracted, or had a saturation (= F^dev/F^max) less than 25% or greater than 75% were excluded. The contractile forces measured by the force transducer were digitized and recorded with a sampling rate of 500 Hz.

Trabeculae were randomly assigned to distinct experimental groups as represented in figure 1. Trabeculae (except for time controls) were subjected to simulated ischemia (SI) by switching to superfusion with hypoxic glucose-free Tyrode (95% N²and 5% CO², partial pressure of oxygen [Po²]≤ 4 mmHg) and by increasing the stimulation frequency to 1 Hz. Ischemia decreased F^devto zero followed by an increase in passive force after starting SI. When the passive force increased to 50% of F^max,start(= start of rigor), trabeculae were subjected to SI for another 40 min and were subsequently reperfused for 60 min with normal oxygenated buffer solution. After reperfusion, recovery of F^dev(F^dev,rec) and F^max(F^max,rec) were determined and expressed as percentages of F^dev,startand F^max,start, respectively. Except for the time and inhibitor control experiments, trabeculae were preconditioned for 15 min with normal Tyrode saturated with 3.8 vol% vaporized sevoflurane (Sevorane; Abbott B.V., Hoofddorp, The Netherlands) 30 min before the onset of SI. The volume percentage of sevoflurane in the gas phase above the Tyrode was continuously monitored by a calibrated infrared anesthetic monitor (Capnomac Ultima; Datex, Helsinki, Finland). Sevoflurane was washed out 15 min before the onset of SI.

In four additional groups, trabeculae were superfused with the PKC catalytic site inhibitor chelerythrine (6 μm; Sigma-Aldrich Chemie BV, Zwijndrecht, The Netherlands), the specific inhibitor of PKC-δ rottlerin (1 μm; Sigma-Aldrich Chemie BV), the ROS scavenger n-(2-mercaptopropionyl)-glycine (MPG; 300 μm; Sigma-Aldrich Chemie BV), or the mito K^+ATPchannel inhibitor 5-hydroxydecanoic acid sodium (5-HD; 100 μm; Sigma-Aldrich Chemie BV) 10 min before preconditioning with sevoflurane until the onset of SI (fig. 1). The intrinsic effects of the distinct inhibitors on F^dev,recafter SI were studied in individual inhibitor control experiments.

Intracellular translocation and/or activation of distinct signaling molecules was visualized by immunofluorescent staining and subsequent three-dimensional digital imaging microscopy. Isolated trabeculae were subjected to similar protocols as compared with the force measurement protocols including the application of inhibitors. However, trabeculae were immediately embedded in gelatin, rapidly frozen in liquid nitrogen, and stored in −80°C until further use after the inhibitor washout period. For immunofluorescent staining, sections were paraformaldehyde fixated and incubated with a primary antibody (rabbit) raised against rat PKC-δ (662–673) or PKC-ε (728–737; both obtained from Research & Diagnostic Antibodies, Benicia, CA). Sections were then incubated with a fluorescein isothiocyanate–labeled swine anti-rabbit secondary antibody (F0205; DakoCytomation BV, Heverlee, The Netherlands). Simultaneously, sections were counterstained for the sarcolemma with 10% (vol/vol) wheat germ agglutinin (W-7024; Molecular Probes, Leiden, The Netherlands). Finally, nuclei were stained with 4′,6 diamidino-2-phenylindole (DAPI)–containing mounting medium (H1200, Vectashield; Vector Laboratories, Burlingame, CA). Separate sections were incubated with a primary antibody (rabbit) raised against rat nitrotyrosine (A-21285; Molecular Probes) and then incubated with a secondary antibody as described above.

Sections were qualitatively analyzed with use of an inverted digital imaging microscopy workstation (Axiovert 200 Marianas; Carl Zeiss, Sliedrecht, The Netherlands) equipped with a motorized stage and multiple fluorescent channels. A cooled charge-coupled device camera (Cooke Sensicam; Cooke Co., Eugene, OR) was used to record images. Exposures, objective, montage, and pixel binning were automatically recorded and stored in memory. Dedicated imaging and analysis software (SlideBook, version 3.11) was obtained from Intelligent Imaging Innovations (Denver, CO) and included advanced deconvolution techniques.

The sample size of each experimental group was 6, except for the inhibitor control groups (n = 5) and the groups for immunofluorescent analysis (n = 3). Data were evaluated for normal distribution, and a one-way analysis of variance was performed, followed by a Tukey post hoc  analysis, to determine differences between the experimental groups. A P  value less than 0.05 was considered to reflect a significant difference. All values are represented as mean ± SEM.

Table 1shows the initial general trabeculae characteristics for each experimental group. Minor variation was detected in trabeculae cross-sectional areas among groups. F^dev,start/F^max,startvalues were similar between experimental groups, implying that trabeculae had on average comparable contractile force reserves. Furthermore, passive force at 95% of the maximal length expressed as a percentage of F^max,startwas similar between groups before the onset of SI. SI resulted in an immediate decrease of F^devto zero, followed by an increase in passive force, ultimately leading to a rigor state. On average, the time to onset of the rigor state was 28 min. MPG increased the time to rigor, but we found an insignificant correlation between time to rigor and F^dev,recor F^max,rec.

Figure 2depicts the effects of sevoflurane on recovery of force after SI. In time controls, F^devand F^maxwere decreased after approximately 170 min to 80 ± 9 and 81 ± 2% of F^dev,startand F^max,start, respectively. SI reduced the recovery of force (F^dev,rec) to 49 ± 3% (P < 0.01 vs.  time control). In addition, the recovery of F^maxwas decreased to 65 ± 4% (P < 0.05 vs.  time control), whereas SI did not affect passive force. Sevoflurane itself induced negative inotropy and reduced F^devto 50% of F^dev,start. However, this negative inotropy was completely reversed after the washout of sevoflurane, 15 min before SI. Preconditioning with sevoflurane improved the ischemic injury–related decrease in F^dev,recto 74 ± 7% (P < 0.05 vs.  SI) but did not improve F^max,rec(56 ± 4%; P > 0.05 vs.  SI).

Figures 3 and 4represent functional and immunohistochemical data, respectively, on the role of PKC-δ in sevoflurane-induced cardioprotection. Figure 3shows that chelerythrine (a) and rottlerin (b) both abolished sevoflurane-induced preservation of F^dev,rec(36 ± 4 and 46 ± 4% for chelerythrine and rottlerin, respectively; both P < 0.01 vs.  sevoflurane). Figure 4shows the translocation pattern of PKC-δ in response to sevoflurane in both longitudinal (a and d) and cross-sectional (b, c, and e–h) sections. In control sections, a cytosolic localization of the PKC-δ isoform can be distinguished (a–c). In addition, the red sarcolemmal counterstain is visualized in panel c. In contrast, in trabeculae exposed to sevoflurane, PKC-δ is present as a striation-like translocation pattern in the longitudinal sections (d, white arrows ). In cross-sectional sections, PKC-δ is visualized in a patchy distribution following distinct linear membrane structures of the cardiomyocytes (e, white arrows ). The sarcolemmal counterstain colocalizes with PKC-δ (f), which is indicated by the yellow color. Interestingly, the sevoflurane-induced translocation of PKC-δ to the sarcolemma was blocked by both chelerythrine (g) and rottlerin (h).

Figure 5shows the distribution of PKC-ε in control (a and b) and sevoflurane-exposed trabeculae (c and d). In all panels, green represents specific PKC-ε staining, and blue indicates the nuclear DAPI staining. In panels b and d, the sarcolemma has been counterstained by wheat germ agglutinin (red color). In control trabeculae, PKC-ε is localized in the perinuclear area of both cardiomyocytes and interstitial cells, as is shown by the green color surrounding the nuclei of cardiomyocytes. Sevoflurane exposure results in a reduction of perinuclear PKC-ε in cardiomyocytes but not in interstitial cells.

The role of mito K^+ATPchannels and ROS production in sevoflurane-induced cardioprotection in recovery of force and PKC-δ translocation is shown in figures 6 and 7, respectively. Inhibition of mito K^+ATPchannels by 5-HD (fig. 6a) and ROS scavenging by MPG (fig. 6b) both abolished sevoflurane-induced cardioprotection. Neither inhibitor established modulation of F^devor intrinsic cardioprotective properties. Although 5-HD inhibits sevoflurane-induced cardioprotection, figure 7shows that 5-HD does not abolish sevoflurane-induced PKC-δ translocation to the sarcolemma (a and b). In contrast, ROS scavenging by MPG inhibits sevoflurane-induced translocation of PKC-δ (c and d). Figure 8represents the sevoflurane-induced formation of nitrotyrosine, as indicated by green fluorescence. In time controls, nitrotyrosine was undetectable (a and b), whereas sevoflurane clearly induced sarcolemmal nitrosylation (c and d). The involvement of ROS in the peroxynitrite modification of sarcolemmal tyrosine residues was shown by the absence of nitrotyrosine when trabeculae were exposed to sevoflurane in the presence of MPG (e and f).

In this study, we demonstrated that sevoflurane-induced cardioprotection in functional cardiac tissue depends on activation of the PKC-δ isoform. Furthermore, exposure of trabeculae to sevoflurane induces translocation of PKC-δ but not of PKC-ε from the cytosol to the cardiomyocyte sarcolemma. PKC-δ translocation acts via  the production of ROS but not the opening of mito K^+ATPchannels. Finally, sevoflurane induces formation of sarcolemmal nitrotyrosine, which represents the reaction of peroxynitrite with tyrosine residues. To our knowledge, this is the first study showing the involvement of PKC-δ and its relative order compared with ROS and opening of mito K^+ATPchannels in sevoflurane-induced cardioprotection.

Several studies showed specific translocation patterns of various PKC isoforms in response to ischemic as well as pharmacologic preconditioning. Inagaki et al.  27studied PKC translocation patterns of the cardioprotective agent JTV519 and demonstrated that this compound induced a translocation of PKC-δ but not of PKC-ε to the sarcolemma. Recently, Julier et al.  23demonstrated that sevoflurane itself results in translocation of PKC-δ to the sarcolemma in human atrial tissue, in addition to translocation of PKC-ε to the intercalated discs and mitochondria. However, these sevoflurane-treated patients received significantly more phenylephrine compared with control subjects to maintain blood pressure. Because α-adrenergic stimulation induces PKC translocation 17and cardioprotection, 28the sevoflurane-induced effects were biased by the phenylephrine-induced changes. We demonstrate with immunofluorescent analysis that sevoflurane itself induces translocation of PKC-δ to the sarcolemma. In addition, our functional experiments show that PKC-δ activation may be important for sevoflurane-induced cardioprotection. This is in agreement with observations showing that PKC-δ is primarily involved in pharmacologic preconditioning with opioids, 14adenosine 18and phenylephrine. 17However, we do not rule out that other PKC isoforms, such as PKC-ε, are involved in sevoflurane-induced preconditioning as well.

Interestingly, a functional role for PKC-ε was recently established during sevoflurane-induced cardioprotection in guinea pig hearts in addition to studies showing that ischemic preconditioning is mediated by PKC-ε. 22,29–31Some studies even indicate that PKC-δ inhibition is cardioprotective, as was demonstrated by the increased resistance against ischemia–reperfusion damage in mice hearts overexpressing a δ V1 inhibitor peptide. 32Nevertheless, the latter study also established the necessity of an active form of the PKC-δ because overexpression of the δ V1 inhibitor peptide at intermediate or very high doses resulted in cytoskeletal deformations, a lethal phenotype and less protection to ischemia–reperfusion, respectively. Liu et al.  19demonstrated that ischemic preconditioning resulted in a translocation of both PKC-δ and PKC-ε, whereas on pharmacologic preconditioning with opioids, only PKC-δ translocated. This is in accordance with our observations that sevoflurane itself does not cause sarcolemmal translocation of PKC-ε in trabeculae. In addition, Fryer et al.  14,15showed in isolated rat hearts that opioid-induced cardioprotection but not ischemic preconditioning was abolished by pharmacologic inhibition of PKC-δ. However, in contrast, Uecker et al.  21showed PKC-δ translocation to the mitochondria and PKC-ε to the sarcolemma in response to isoflurane. These observations indicate that distinct signal transduction pathways are involved not only in ischemic preconditioning and pharmacologic preconditioning, but also between different pharmacologic preconditioning stimuli.

Recently, a profound role for ROS in the signal transduction of volatile anesthetic–induced cardioprotection has been established. 3,7,33Our results indicate that translocation of PKC-δ occurs downstream of ROS because PKC-δ translocation could be inhibited by MPG. This is in agreement with observations of Zhang et al. , 34who showed that opioid-induced protection was mediated by ROS, which could not be blocked by rottlerin. In addition, Wang et al.  35showed that diazoxide-induced protection in isolated rat hearts was abolished by chelerythrine. In contrast, other studies showed that ROS-mediated cardioprotection could not be blocked by PKC inhibition. 36This suggests that ROS may also be involved in a signaling pathway independent from PKC. That sevoflurane indeed results in the production of ROS in our model was demonstrated by the positive nitrotyrosine staining of the cardiomyocyte sarcolemma on exposure to sevoflurane. Nitrotyrosine is induced when nitric oxide and superoxide form peroxynitrite, which can modify tyrosine residues to form nitrotyrosine. Others also demonstrated the involvement of ROS and nitrogen species as triggers of anesthetic preconditioning in isolated guinea pig hearts. 37 

It has been hypothesized that ROS production arises from the mitochondria as volatile anesthetics inhibit complex I of the electron transport chain. Whether opening of the mito K^+ATPchannels is essential for ROS production has not been completely elucidated. However, it was recently demonstrated that 5-HD does not inhibit sevoflurane-induced ROS production. 23In contrast, during isoflurane-induced preconditioning, opening of the mito K^+ATPchannels was essential for ROS production triggering cardioprotection. 33We demonstrate in functional trabeculae that opening of the mito K^+ATPchannels is essential for inducing the cardioprotective response but that opening of mito K^+ATPchannels is not essential for PKC-δ activation. With respect to the published literature, this supports that mito K^+ATPchannel opening is involved in sevoflurane-induced cardioprotection but is downstream of PKC-δ translocation. These data suggest that ROS production precedes PKC-δ activation and that PKC-δ activation precedes the opening of mito K^+ATPchannels in the signaling pathway of sevoflurane-induced cardioprotection.

Possible target substrates of PKC-δ during sevoflurane-induced cardioprotection remain speculative. Targets might include proteins involved in Ca²⁺homeostasis, as we previously demonstrated by showing that preconditioning with norepinephrine attenuated Ca²⁺overload and preserved Ca²⁺reuptake activity of the sarcoplasmic reticulum. 28,38In addition, PKC has been reported to modify the function of Ca²⁺-handling proteins, such as L-type Ca²⁺channels, 39the ryanodine receptor, 40sarcoplasmic Ca²⁺adenosine triphosphatase, 41and the Na⁺–Ca²⁺exchanger. 42Furthermore, anesthetic preconditioning might also preserve mitochondrial function. 43Future experiments in our laboratory will be directed toward unraveling the potential end-effector proteins of PKC-δ.

In our laboratory, the in vitro  model of isolated rat trabecula has been extensively used in previous studies regarding ischemia and reperfusion injury and preconditioning in functional cardiac tissue. 3,28,38Our experiments are performed under relatively hypothermic conditions (24°C) to ensure stability of trabeculae over time. This could influence signaling processes, such as translocation of PKC isoforms. However, it was reported recently that during cold cardioplegia, translocation of PKC isoforms still occurred and that the induced cardioprotection is mediated by protein kinase C. 44–46The negative inotropic effect of sevoflurane, when applied before SI, was completely restored after the washout period. There was no difference in F^devafter washout between groups. In studies regarding ischemic and anesthetic preconditioning, chelerythrine, rottlerin, MPG, and 5-HD are commonly used inhibitors for studying PKC, PKC-δ, ROS, and mito K^+ATPchannels, respectively. We used concentrations comparable with data found in the literature. Although rottlerin is reported to be a specific inhibitor of the PKC-δ isoform, one should be cautious of unspecific inhibitory effects. 47,48However, it has been demonstrated by Keenan et al.  49that PKC-δ is more potently inhibited by rottlerin than classic PKC isoforms. In addition, we show by immunofluorescent microscopy that rottlerin selectively inhibits sevoflurane-induced translocation of PKC-δ.

In summary, PKC-δ, mito K^+ATPchannels, and ROS are equally involved in sevoflurane-induced cardioprotection in isolated rat trabeculae. ROS production precedes PKC-δ translocation on stimulation with sevoflurane, while the opening of mito K^+ATPchannels occurs downstream of this translocation. Furthermore, sevoflurane exposure induces formation of peroxynitrite, which modifies tyrosine residues in the sarcolemma. We propose that sevoflurane directly induces ROS and peroxynitrite formation, which might result in recruitment of PKC-δ to the cardiomyocyte sarcolemma. Furthermore, ROS and peroxynitrite might also affect tyrosine residues in the PKC-δ isoform via  tyrosine kinase, leading to activation of downstream signaling molecules, including mito K^+ATPchannels. The downstream localization of mito K^+ATPchannels in the signaling pathway of sevoflurane-induced cardioprotection suggests a role for mitochondria as crucial organelles for cardiac preservation.