Effects of Nitrous Oxide on the Rat Heart In Vivo : Another Inhalational Anesthetic That Preconditions the Heart?

The study was performed in accord with the regulations of the German Animal Protection Law and was approved by the Bioethics Committee of the District of Düsseldorf (Düsseldorf, Germany). A total of 40 male Wistar rats were used (200–250 g; 7 per group for infarct size measurement and 4 per group for Western blot).

Monoclonal mouse anti–α-tubulin and rabbit polyclonal anti–actin antibody were purchased from Sigma (Taufkirchen, Germany). The enhanced chemoluminescence protein detection kit was purchased from Santa Cruz (Heidelberg, Germany). Total PKC-ϵ rabbit polyclonal and phospho PKC-ϵ rabbit antibody were from Upstate (Charlottesville, VA). Peroxidase-conjugated goat anti-rabbit and donkey anti-mouse antibodies were from Jackson Immunolab (Dianova, Hamburg, Germany). Anti–phospho Src (Tyr 416) and phospho Src (Tyr 527) rabbit polyclonal antibodies were from Cell Signaling (Frankfurt/M, Germany). All other materials were purchased from either Sigma or Merck-Eurolab (Munich, Germany).

Male Wistar rats (200-250 g) were anesthetized by intraperitoneal S -ketamine injection (150 mg/kg). It has been demonstrated previously that the S -enantiomer of ketamine does not influence myocardial preconditioning. Animals had free access to food and water at all times before the start of the experiments.

All animals were left untreated for 10 min before the start of the respective preconditioning protocol ().

Further preparation and infarct size measurement by triphenyltetrazolium chloride (TTC) staining were performed as described previously. In summary, after tracheal intubation, the lungs were ventilated with oxygen-enriched air and a positive end-expiratory pressure of 2–3 cm H²O. Respiratory rate was adjusted to maintain partial pressure of carbon dioxide within physiologic limits. Body temperature was maintained at 38°C by the use of a heating pad. The right jugular vein was cannulated for saline and drug infusion, and the left carotid artery was cannulated for measurement of aortic pressure. Anesthesia was maintained by continuous α-chloralose infusion. A lateral left-sided thoracotomy followed by pericardiotomy was performed, and a ligature (5-0 Prolene®; Ethicon GmbH, Norderstedt, Germany) was passed below the main branch of the left coronary artery. The ends of the suture were threaded through a propylene tube to form a snare, and the coronary artery was occluded by tightening the snare. Successful coronary artery occlusion was verified by epicardial cyanosis.

After 120 min of reperfusion, the heart was excised and mounted on a modified Langendorff apparatus for perfusion with ice-cold normal saline via  the aortic root at a perfusion pressure of 40 cm H²O to wash out intravascular blood. After 5 min of perfusion, the coronary artery was reoccluded, and the remainder of the myocardium was perfused through the aortic root with 0.2% Evans blue in normal saline for 10 min. Intravascular Evans blue was then washed out by perfusion for 5 min with normal saline. This treatment identified the area at risk as unstained. The heart was then cut into transverse slices, 2 mm thick. The slices were stained with 0.75% TTC solution for 10 min at 37°C and fixed in 4% formalin solution for 6 h at room temperature. The area of risk and the infarcted area were determined by planimetry using SigmaScan Pro 5® computer software (SPSS Science Software, Chicago, IL) and corrected for dry weight.

Rats were divided into four groups ():

Rats received 20% oxygen plus 80% nitrogen during three 5-min periods before they were subjected to 25 min of left coronary artery occlusion, followed by 2 h of reperfusion. During the washout times, rats received 50% oxygen and 50% nitrogen.

Rats received 60% nitrous oxide (equivalent to 0.4 minimum alveolar concentration [MAC] in rats) plus 20% oxygen and 20% nitrogen for three 5-min periods. The preconditioning periods were interspersed with two 5-min washout periods. After the last preconditioning, a 10-min final washout was performed before the 25-min coronary artery occlusion. Ischemia was followed by 2 h of reperfusion. During the washout times, rats received 50% oxygen and 50% nitrogen.

Rats received 1.4% isoflurane (1 MAC in rats) plus 20% oxygen and 78.6% nitrogen for three 5-min periods. The preconditioning periods were interspersed with two 5-min washout periods. After the last preconditioning, a 10-min final washout was performed before the 25-min coronary artery occlusion. Ischemia was followed by 2 h of reperfusion. During the washout times, rats received 50% oxygen and 50% nitrogen.

Rats received a mixture of 60% nitrous oxide, 20% oxygen, and 18.6% nitrogen to which 1.4% isoflurane was added for three 5-min periods. The preconditioning periods were interspersed with two 5-min washout periods. After the last preconditioning, a 10-min final washout was performed before the 25-min coronary artery occlusion. Ischemia was followed by 2 h of reperfusion. During the washout times, rats received 50% oxygen and 50% nitrogen.

Rats were divided into three groups ():

Rats received 20% oxygen plus 80% nitrogen during three 5-min periods. The preconditioning periods were interspersed with two 5-min washout periods. After the last preconditioning, a 10-min final washout was performed, and the hearts were excised and shock frozen in liquid nitrogen. During the washout times, rats received 50% oxygen and 50% nitrogen.

Rats received 60% nitrous oxide (equivalent to 0.4 MAC in rats) plus 20% oxygen and 20% nitrogen for three 5-min periods. The preconditioning periods were interspersed with two 5-min washout periods. After the last preconditioning, a 10-min final washout was performed, and the hearts were excised and shock frozen in liquid nitrogen. During the washout times, rats received 50% oxygen and 50% nitrogen.

Rats received a mixture of 60% nitrous oxide, 20% oxygen, and 19.4% nitrogen to which 0.6% isoflurane was added for three 5-min periods. The preconditioning periods were interspersed with two 5-min washout periods. After the last preconditioning, a 10-min final washout was performed, and the hearts were excised and shock frozen in liquid nitrogen. This isoflurane concentration corresponds to 0.4 MAC in rats. To compare the results with those of our previous study using a similar preconditioning protocol with xenon and isoflurane, we used the same isoflurane concentration for the investigations of the molecular targets. This isoflurane concentration was shown to induce a strong PKC phosphorylation and translocation in our previous study. During the washout times, rats received 50% oxygen and 50% nitrogen.

After each preconditioning protocol, the end-tidal concentration of the anesthetic was measured during the final washout period to guarantee a sufficient washout of the respective anesthetic. End-tidal concentration of the anesthetic returned to zero within 2 min during the last washout.

Aortic pressure and electrocardiographic signals were digitized using an analog-to-digital converter (PowerLab/8SP; ADInstruments Pty. Ltd., Castle Hill, Australia) at a sampling rate of 500 Hz and were continuously recorded on a personal computer using Chart for Windows version 5.0 (ADInstruments Pty. Ltd.).

Tissue specimens were prepared for protein analysis to investigate PKC-ϵ activation and distribution (membrane and cytosolic fraction) within the myocytes. The excised hearts were frozen in liquid nitrogen. Subsequently, a tissue fractionation was performed that was adapted from the literature.19,,20 The frozen tissue was pulverized and dissolved in lysis buffer containing Tris base, EGTA, NaF, and Na³VO⁴ (as phosphatase inhibitors), a freshly added protease inhibitor mix (aprotinin, leupeptin, and pepstatin), and dithioerythriol. The solution was vigorously homogenized on ice (Homogenizer; IKA®, Staufen, Germany) and then centrifuged at 1,000g , 4°C, for 10 min. This centrifugation at low speed allows a raw separation between the cytosolic fraction that still contains cellular organelles and their membranes and the membrane fraction still containing nuclear particles. The supernatant, containing cytosolic fraction, was centrifuged again at 16,000g , 4°C, for 15 min to clean up this fraction and to separate the mitochondria and other cellular organelles from this fraction. The remaining pellet was resuspended in lysis buffer containing 1% Triton X 100, incubated for 60 min on ice and vortexed. The solution was centrifuged at 16,000g , 4°C, for 15 min, and the supernatant was collected as membrane fraction.

After protein determination by the Lowry method, equal amounts of protein were mixed with loading buffer containing Tris-HCl, glycerol, and bromphenol blue. Samples were vortexed and boiled at 95°C before being subjected to SDS-PAGE. Samples were loaded on a 7.5% (PKC-ϵ) or a 10% (Src) SDS electrophoresis gel, respectively. The proteins were separated by electrophoresis and then transferred to a polyvinylidenfluorid membrane by tank blotting. Unspecific binding of the antibody was blocked by incubation with 5% fat dry milk powder solution in Tris-buffered saline containing Tween (TBS-T) for 2 h. Subsequently, the membrane was incubated overnight at 4°C with the respective first antibody at indicated concentrations. After washing in fresh, cold TBS-T, the blot was subjected to the appropriate horseradish peroxidase–conjugated secondary antibody for 2 h at room temperature. Immunoreactive bands were visualized by chemoluminescence detected on x-ray film (Hyperfilm ECL; Amersham, Freiburg, Germany) using the enhanced chemoluminescence system from Santa Cruz. The blots were quantificated using a Kodak Image station® (Eastman Kodak Co., Rochester, NY). Equal loading of the protein to the SDS-PAGE gel was ensured by Coomassie blue staining (Coomassie brilliant blue®; Serva electrophoresis GmbH, Heidelberg, Germany) of each gel, and the results are presented as ratio of phosphorylated protein to total protein, as ratio of phosphorylated protein to actin, or as ratio of phosphorylated protein to α-tubulin (as indicated in the respective figure legend) after multiplication of the average light intensity by 10 to facilitate the presentation of an X-fold increase.

Sample sizes were calculated using the program GraphPad StatMate version 1.01 (GraphPad Software, San Diego, CA). For infarct size with and α error of α < 0.05 and a β error < 10%, a group size of n = 7 is necessary to detect a minimal difference in infarct size of 30% with a power of 95%.

For Western blot assay with and α error of α < 0.05 and a β error < 10%, a group size of n = 4 is necessary to detect a minimal difference in phosphorylation state of 45% with a power of 95%.

Data are expressed as mean ± SD. Group comparisons were analyzed by Student t  test with Welch modification (Graph Pad Prism version 4.00) followed by Bonferroni correction for multiple comparisons.

Statistical analysis of the hemodynamic variables was performed by two-way repeated-measures analysis of variance for time and treatment (experimental group) effects. If an overall significance was found, comparisons between groups were performed for each time point using one-way analysis of variance followed by Dunnett post hoc  test with the control group as the reference group. Time effects within each group were analyzed by repeated-measures analysis of variance followed by Dunnett post hoc  test with the baseline value as the reference time point. P  < 0.05 was considered statistically significant.

shows the time course of heart rate and mean aortic pressure during the experiments. No significant differences in heart rate and aortic pressure were observed between the experimental groups during baseline, ischemia, and reperfusion periods. Isoflurane and isoflurane plus nitrous oxide transiently reduced mean aortic pressure during preconditioning cycles, but the values recovered to baseline during the washout periods. Isoflurane significantly reduced heart rate during the preconditioning periods compared with baseline conditions (preconditioning 1: 402 ± 37, preconditioning 2: 382 ± 45, preconditioning 3: 380 ± 37 vs.  baseline: 459 ± 36) and with the control group (control: preconditioning 1: 459 ± 34, preconditioning 2: 451 ± 37, preconditioning 3: 460 ± 33). For isoflurane plus nitrous oxide, a difference of heart rate compared with the control group was only found during the third preconditioning cycle (preconditioning 3: 399 ± 37) and disappeared after the final washout. The respective changes disappeared after the final washout.

Nitrous oxide had no effect on infarct size compared with controls (59.1 ± 15.2% of the area at risk vs.  51.1 ± 10.9%; ). In contrast, isoflurane significantly reduced infarct size to 30.9 ± 10.6%; P  < 0.01). A similar infarct size reduction could be observed for the combined application of isoflurane and N²O (28.7 ± 11.8%; P  < 0.01 vs.  controls).

Direct influence of nitrous oxide administration on PKC-ϵ was determined by the use of a phospho-specific antibody against PKC-ϵ. Nitrous oxide had no effect on phosphorylation of PKC-ϵ (2.3 ± 1.8 vs.  2.5 ± 1.7 in controls; ). Total PKC-ϵ protein amount was equal in the different slots because the Western blot (, middle Western blot) showed a uniform distribution. Equal loading of protein samples was confirmed by detection of α-tubulin on the respective blot (, lowest Western blot).

Protein kinase C is activated not only by phosphorylation, but also by translocation to the membrane. Both activation steps can occur in parallel and also independently of each other. Therefore, the absence of an effect of nitrous oxide on PKC-ϵ phosphorylation does not necessarily exclude a possible effect of nitrous oxide on PKC-ϵ translocation to membrane regions.

Western blot assay of fractionated tissue () clearly revealed that nitrous oxide had no effect on translocation of PKC-ϵ to membrane regions (ratio membrane to cytosol PKC-ϵ: 7.0 ± 4.3 vs.  7.4 ± 5.2 in controls; ). α-Tubulin detection served as loading control (, lower Western blot).

Nitrous oxide had no effect on phosphorylation of Src-kinase (Tyr 416/527) (4.6 ± 3.9 vs.  5.0 ± 3.8 and 3.2 ± 2.2 vs.  3.5 ± 3.0; ). Actin detection served as internal standard (, lower Western blot), and the ratio of phosphorylated protein to actin is shown. In addition, protein on the electrophoresis gels were stained with Coomassie blue to show equal protein loading on the respective gel ().

We could recently show that the volatile anesthetic isoflurane (0.4 MAC) enhanced phosphorylation as well as translocation of PKC-ϵ and that blockade of PKC in vivo  by calphostin C completely abolished isoflurane-induced cardioprotection. Therefore, we aimed to clarify whether this increased phosphorylation and translocation of a key enzyme in isoflurane-induced preconditioning is influenced in the presence of nitrous oxide in an equianesthetic dose (0.4 MAC). The combined administration of isoflurane plus nitrous oxide (0.4 MAC/0.4 MAC) induced PKC-ϵ phosphorylation (5.4 ± 1.9 vs.  2.8 ± 1.5; P  < 0.001; ) and translocation (ratio membrane to cytosol PKC-ϵ: 13.8 ± 13.0 vs.  6.7 ± 2.0 in controls; P  < 0.05; ) similar to that seen in the previous study with isoflurane (0.4 MAC) alone,3,,4 demonstrating that PKC-ϵ is still activated by isoflurane in the presence of nitrous oxide.

We investigated whether nitrous oxide preconditions the rat heart in vivo  and whether nitrous oxide affects the two key mediator enzymes of anesthetic-induced preconditioning, PKC and PTKs. We also investigated whether isoflurane-induced preconditioning and activation of PKC isoform ϵ is influenced by nitrous oxide.

The new findings of the current study are as follows: First, nitrous oxide is the first inhalational anesthetic that does not precondition the heart in vivo . Second, nitrous oxide did not influence the activation (phosphorylation and translocation) of PKC-ϵ or Src tyrosine kinase. Third, isoflurane-induced preconditioning and activation of PKC-ϵ by isoflurane were not affected by nitrous oxide.

The results obtained from the current study were surprising because not only all volatile anesthetics but also the inert gas xenon (0.4 MAC) had a strong cardioprotective effect on the rat heart in vivo . Moreover, the preconditioning effect was for a long time regarded to be a specific action of inhalational anesthetics. Therefore, nitrous oxide seems to be an exception to this rule.

Because nitrous oxide is commonly used in combination with volatile anesthetics such as isoflurane, a potential interaction between both anesthetics might influence the cardioprotection of volatile anesthetics. However, we could show that the presence of nitrous oxide had no effect on isoflurane-induced preconditioning, because the combined administration of nitrous oxide and isoflurane (0.4 MAC/1 MAC) reduced infarct size up to 30%. These data suggest that there is no interaction between both anesthetics with regard to the preconditioning effect of the volatile anesthetic.

The data from infarct size measurement were strongly supported by the findings on the molecular target PKC-ϵ, which has been identified as mediator of isoflurane-induced preconditioning in our in vivo  rat model.3,,4 Isoflurane-induced phosphorylation and translocation of PKC-ϵ were not influenced by nitrous oxide. Moreover, nitrous oxide alone did not affect the phosphorylation and translocation state of this enzyme.

The phenomenon of anesthetic-induced preconditioning has raised a lot of attention in the past years, and preconditioning by different inhalational anesthetics has been demonstrated in vitro  and in vivo  in different animal species and in humans.

Ischemic and anesthetic-induced preconditioning share several steps in the signaling cascade. One of these steps is the activation of different PKC isoforms by phosphorylation and translocation. PKC plays a central role in both preconditioning forms. We previously demonstrated that PKC-ϵ phosphorylation was increased after three 5-min periods of xenon (0.4 MAC) and isoflurane (0.4 MAC) inhalation and that the PKC inhibitor calphostin C abolished both the anesthetic-induced infarct size reduction and the increased phosphorylation and translocation of PKC isoform ϵ. Similarly to xenon and isoflurane-induced preconditioning, desflurane preconditioning involves the ϵ isoform of PKC. Moreover, the activation of PKC-ϵ seems to be initiated only by low doses of isoflurane because 0.4 MAC but not 1 MAC significantly increased PKC-ϵ phosphorylation and translocation. This is in contrast to findings of Ludwig et al. , who found that 1 MAC isoflurane given for 30 min before index ischemia induced translocation of PKC-ϵ. However, there are two aspects that should be taken into account regarding the discrepancy between these results. First, Ludwig et al.  used a different preconditioning model in comparison to our study. They induced preconditioning by 30 min of isoflurane administration followed by a 15-min washout phase. In our study, 15 min of isoflurane (1 MAC) and a 10-min washout were used. Second, in the former study, hearts for tissue preparation and immunohistochemistry were excised directly after the preconditioning phase and not after the washout phase. As it is know from ischemic preconditioning and another study of our group, the translocation and phosphorylation of PKC follow a sensitive time course pattern. Therefore, these differences in time course pattern and preconditioning duration might explain the different results.

However, identification of the PKC isoform mainly responsible for the cardioprotection by anesthetics is still a matter of controversy. For example, Uecker et al.  identified PKC-δ activation rather than PKC-ϵ as a key step in isoflurane-induced preconditioning. They did not find an increased phosphorylation of PKC-ϵ after 15 min of isoflurane (1.5 MAC) administration in the isolated rat heart. However, as in our previous studies, they reported an increased translocation of PKC-ϵ to membrane regions. In contrast to these two studies, Stowe et al.  identified PKC-ϵ as key mediator of preconditioning with sevoflurane in guinea pigs. These divergent data clearly indicate that there is a complex signaling cascade of anesthetic preconditioning depending on the choice of anesthetic, the species, and the preconditioning protocol. Moreover, there is obviously a pronounced discrepancy between the in vitro  and in vivo  situations.

In recent preconditioning studies using the same experimental model as in the current study, PKC-ϵ could be identified as the central mediator of anesthetic-induced preconditioning in vivo .2,,3,,4 Therefore, we focused on PKC-ϵ as one potential target of nitrous oxide and demonstrated that this inhalational agent does not influence PKC-ϵ activation.

Besides the activation of PKC, a variety of additional kinases are involved in ischemic as well as in anesthetic preconditioning. The activation of different PKC isoforms is closely related to PTK phosphorylation. Evidence has been reported that PTKs can be upstream of,25,,26 parallel to,17,,27,,28 or downstream of PKC.29,,30 Regarding the signal transduction cascade of ischemic preconditioning, the blockade of PTK has been shown to inhibit cardioprotection by ischemic preconditioning.29,,31 Moreover, PTKs were identified as triggers rather than mediators of ischemic preconditioning. Vahlhaus et al.  demonstrated in pigs that a combined inhibition of PKC and TK is needed to sufficiently block cardioprotection of ischemic preconditioning. Fryer et al.  partially blocked ischemic preconditioning by the use of genistein and lavendustin A. The same group reported a successful blockade of cardioprotection induced by one 5-min cycle of preconditioning ischemia using genistein. However, genistein did not abrogate preconditioning by three 5-min cycles of ischemia in rats. In contrast to these studies, Kitakaze et al.  could not show an implication of PTKs in the infarct size limiting effect of ischemic preconditioning in the canine heart. In summary, the role of PTK activation and its relation to PKC seem to depend on species and the choice of the ischemic preconditioning protocol.

Concerning anesthetic-induced preconditioning, there is much more controversy regarding the implication of PTKs in cardioprotection. Stadnicka et al.  showed in an in vitro  model of guinea pig ventricular myocytes that PTK may modulate the responsiveness of sarcolemmal adenosine triphosphate–sensitive potassium channels to isoflurane. In contrast to the sarcolemmal adenosine triphosphate–sensitive potassium channels, the mitochondrial adenosine triphosphate–sensitive potassium channels are activated independently of PKC and PTKs by isoflurane as reported by a later study of the same group. These in vitro  data are supported by a recent study of Ludwig et al. , who found that the translocation of PKC-ϵ and -δ and activation of Src PTK mediate isoflurane-induced preconditioning in vivo . They used an in vivo  rat model and the PTK blocker lavendustin A and PP1 in combination with 30 min of isoflurane (1 MAC) preconditioning. A recent study of our group using an in vivo  rabbit model and a 15-min desflurane preconditioning stimulus demonstrated that the administration of two structurally different PTK inhibitors could not block cardioprotection by desflurane.

These data again indicate that the implication of PTKs strongly depends on the choice of the anesthetic agent, the intensity of the preconditioning stimulus, and the species. In addition, a parallel pathway of PTKs and PKC, dependent on the force of the stimulus as already suggested for ischemic preconditioning by Fryer, seems a possible explanation for the above described discrepancies.

Nevertheless, besides PKC and adenosine triphosphate–sensitive potassium channels, PTKs seem to be one of the most supposable targets of preconditioning phenomena. Hence, we choose Src kinase phosphorylation as a further cellular target to evaluate effects of nitrous oxide on preconditioning mediators. In agreement with the results from PKC-ϵ phosphorylation, we found no effect of nitrous oxide on either phosphorylation site of Src kinase.

Preconditioning indicates that the cardioprotective stimulus is not present during ischemia. Nitrous oxide has a low blood:gas partition coefficient of only 0.47 and shows an extremely rapid onset and offset of its action. A sufficient washout of nitrous oxide and isoflurane was guaranteed by a high fresh gas flow during the washout period. Moreover, the end-tidal concentration of anesthetics was measured continuously and returned to zero within 2 min during the last washout.

In the current study, we could not detect remarkable differences in the hemodynamic variables between the groups. During administration of the anesthetics, there was a transient reduction in mean aortic pressure and heart rate. Hemodynamics rapidly recovered to baseline conditions during the washout period. Most important, there were no significant differences between the groups during ischemia , and this would have been the only time period of critical importance in our experimental setting because major changes in this period might modify the ischemic injury. Regarding discussion of hemodynamic effects in our experimental setting, we would like to remark that the classic endpoint of cardioprotection by preconditioning is lethal cell damage, i.e. , infarct size which was measured directly in our study.

Hemodynamics were measured as a potential confounder of this target variable (i.e. , major changes of hemodynamics during ischemia might modify ischemic injury) but not as an endpoint (i.e. , stunning).

Major changes in hemodynamics cannot be expected from the salvage of tissue if one takes into account the total mass of the infarcted tissue; it seems very unlikely that one can see changes in hemodynamic variables during resting conditions, i.e. , when all compensatory mechanisms can still be used to keep the cardiac output at its normal resting value.

This is also what we had observed in all previous studies and in different animal models ranging from rat and rabbit to dogs. The overall size of the ischemic area is only around 25–31% of the total left ventricular contractile mass. In this area, 60–29% are infarcted. Even a reduction of the infarct size by 50% thus only gives 12–15% more of left ventricular contractile mass. With regard to healthy animals with a low resting cardiac output during anesthesia, such a small change in contractile mass is far away from bringing the left ventricle out of its compensatory range.

It should also be considered that we used only one concentration of nitrous oxide, and our results must be limited to this concentration. However, we aimed to investigate the most clinically relevant dose of both anesthetics alone and in combination. Therefore, we choose 1 MAC isoflurane and 0.4 MAC (corresponding to 60 vol%) nitrous oxide.

In general, several doses, preferably different by the factor of 10 in the form of a dose–response curve, should be investigated. Therefore, investigating a relevant further increase in the dose would result in severe technical problems because hypoxic conditions occur in the in vivo  rat model. The only alternative would be to repeat all experiments under hyperbaric conditions to further increase the nitrous oxide dose, but this would not lead to clinically relevant results.

In summary, 60% nitrous oxide had no effect on two important mediators of anesthetic-induced preconditioning, PKC-ϵ and the PTK Src. In accord with these results, on the molecular level, nitrous oxide alone had no preconditioning effect on the rat heart in vivo . Therefore, in contrast to all other inhalational anesthetics, nitrous oxide is the first one that has no preconditioning effect. Because nitrous oxide is commonly used in combination with volatile anesthetics, it is of special importance that isoflurane-induced preconditioning and PKC-ϵ activation after isoflurane are not influenced by nitrous oxide.

The authors thank Claudia Dohle and Yvonne Grüber (Technicians, Department of Anesthesiology, University Hospital of Düsseldorf, Düsseldorf, Germany) for excellent technical support.