eNOS-derived NO has been identified as a critical trigger of cardioprotection that is recruited by endogenous pathways [28
] and by the volatile anesthetic isoflurane [19
]. Studies investigating protection against myocardial ischemia and reperfusion injury have focused primarily on NO derived from a NOS isoform in CM, but our data indicate that EC are an underappreciated and important paracrine source of NO during cardioprotection. Cardiac endothelial cell–myocardial signaling has been described in the context of heart development, modulation of contractility, control of rhythmicity, and in the pathogenesis of heart failure [1
], however, little is known regarding the specific contribution of EC to CM protection against ischemia and reperfusion injury. Cardiac EC outnumber CM in vivo by 3:1, although, the mass ratio of these two cell types is .04–.05 [29
]. This abundance of EC in the heart provides a diffusion radius between EC and CM sufficient for effective EC–CM-NO signaling even considering the short biological half life of NO. Thus, EC–CM interactions are likely to play a critical role that influences outcome during pharmacological protection against myocardial injury, and the current results support the concept that this action occurs through distinct signaling events in EC versus CM.
The transcription factor HIF1α has been demonstrated to play an important role during hypoxic- and ischemic-preconditioning in the myocardium and brain [30
], and is involved in anesthetic-induced organ protection [21
]. HIF1α is a heterodimeric protein composed of a constitutively expressed β-subunit and an α-subunit under regulatory control. Functional control of the concentration of HIF1α in tissue occurs by changes in both expression and post-translational stabilization of the protein [32
]. Hypoxia is the predominant stimulus for production of HIF1α; however non-hypoxic stimuli including reactive oxygen species (ROS), insulin, thrombin, growth factors, and cytokines can modify HIF1α expression [33
]. Primary regulation of HIF1α under normoxic conditions occurs via specific hydroxylation and subsequent degradation by prolyl hydroxylase (PHD) within the oxygen-dependent domain [36
]. Although the activity of PHD is oxygen-dependent, this enzyme is also regulated by ROS and NO [37
]. Thus, deactivation of PHD by NO could suggest the presence of a positive feedback loop between NO and HIF1α activation. Hydroxylation prevents transactivation of HIF1α to the nucleus under normal oxygen tension. However, during hypoxia HIF1α is not subjected to proteasomal degradation, but rather, is translocated to the nucleus where it induces expression of proteins (e.g. VEGF, erythropoietin, hemoxygenase and GLUT1 transporter) associated with protection against hypoxic injury [38
]. Increases in the expression of HIF1α have been observed in cells under normal oxygen tension, suggesting that alternate signaling pathways may also regulate the expression of this important transcription factor.
The current results confirm and extend previous findings by demonstrating that isoflurane enhanced CM survival in the presence of EC by differentially regulating HIF1α and modulating NO production in EC versus CM. Pharmacological stimulation with isoflurane produced an increase in cytosolic and nuclear expression of HIF1α in EC, but not in CM, and this action was blocked by inhibition of ERK with either of two distinct MEK inhibitors. Isoflurane has previously been shown to increase release of signaling ROS by mitochondria [10
] which could result in ERK activation via redox regulation of mitogen activated protein kinase phosphatases [40
]. Although we previously observed increased HIF1α expression in the myocardium, isoflurane did not increase HIF1α in isolated CM in the present investigation [21
]. This finding might be explained by a lack of discrimination between proteins expressed in specific cell types in the whole heart using Western blotting previously, or because of species-dependent (i.e. rat versus rabbit) differences in the magnitude or time-dependence of cell signaling events. For example, isoflurane stimulated the expression of HIF1α in Hep3B cells in a time-dependent manner [22
]. However, maximal expression of HIF1α occurred between 4 and 8 h after exposure to isoflurane in Hep3B cells, whereas, our results in EC indicate a more rapid (1 h) increase in expression. During co-culture of EC and CM, isoflurane produced significant protection of CM against hypoxia and reoxygenation injury, but, this beneficial effect was abolished by suppressing HIF1α expression in EC with lentiviral shRNA. Thus, the findings support the contention that signals initiated in the endothelium by pharmacological agents such as volatile anesthetics result in the transference of factors that mediate protection of CM in a paracrine fashion.
VEGF is a recognized downstream target regulated by HIF1α [41
] and our previous findings demonstrated that isoflurane administered prior to coronary artery occlusion in rats resulted in enhanced myocardial expression of HIF1α and VEGF [21
]. HIF1α expression appeared to be essential for induction of VEGF mRNA and expression of VEGF protein, actions that were inhibited by HIF1α knockdown with siRNA in basophils [43
]. VEGF is also known to stimulate an increase in NO production [44
], however, it is unknown if increases in VEGF in EC were specifically responsible for HIF1α-mediated increases in NO production and subsequent protection of CM against injury observed during isoflurane in the present study. This possibility is currently under investigation.
The non-selective NOS inhibitor l
-NAME abolished the protective effect of isoflurane on CM in co-culture when administered either throughout experimentation or only during reoxygenation. In contrast, the presence of l
-NAME only before hypoxia had no effect on CM protection. These results suggest that NO is an important mediator of pharmacological protection by isoflurane, consistent with previous findings in eNOS–/–
]. Furthermore, silencing of HIF1α resulted in a significant decrease in EC-derived eNOS and phosphorylated eNOS that likely explains the dramatic reduction in NO production. CM failed to express eNOS protein at any timepoint during the experiment supporting the concept of endothelial derived NO in our co-culture model. These data suggest that expression of HIF1α during normoxic conditions is important for sustained expression and phosphorylation of eNOS and thus production of NO. Importantly, HIF1α may also be important for maintaining eNOS expression during hypoxia and reoxygenation as silencing of HIF1α resulted in a further decrease in eNOS expression under these conditions.
The targets of EC-derived NO within CM that are related to cytoprotection have been incompletely elucidated, although, evidence suggests that NO may have direct effects on specific mitochondrial proteins. For example, direct or indirect effects of NO on the mPTP have been implicated in mechanisms responsible for cardioprotection [45
]. Pharmacological protection against ischemia and reperfusion injury with isoflurane has been shown to be dependent on inhibition of this somewhat elusive protein complex [19
]. The present results confirm earlier findings and further demonstrate that the presence of EC delays opening of the mPTP in CM subjected to laser induced injury. Remarkably, these beneficial actions were augmented by isoflurane. NO released by EC may produce favorable effects on CM by binding to heme-containing proteins such as cytochrome c oxidase: an effect that inhibits its activity. This action would be expected to modulate terminal electron transfer to molecular oxygen resulting in the attenuation of cellular respiration and potentially decreasing the formation of damaging ROS [46
] that promotes mPTP opening.
Notably, the current findings support the existence of crosstalk between EC and CM that contributes to cardioprotective signaling, and indicate that NO is an important mediator during cell–cell communication. Isoflurane stimulated an increase in NO production from EC under baseline conditions, but, did not significantly increase NO production in co-culture. The lack of increase of NO observed under these conditions may have been related to fewer total EC in co-culture (1:12 EC–CM) as compared with EC cultured alone. Alternatively, myoglobin is capable of binding NO to form iron-nitrosyl myoglobin and CM have been shown to scavenge free NO [47
]. Hypoxia profoundly decreased NO production in EC, CM and in EC–CM co-culture, however, NO concentrations were nearly two fold higher during hypoxia in isoflurane-treated compared to untreated-cells in co-culture. This finding suggests that volatile anesthetics may preserve the mechanisms that are responsible for regulating NO production by NOS [17
] or by influencing nitrite/nitrate metabolism during hypoxia. Interestingly, NO concentrations remained depressed after reoxygenation in EC or CM alone, but, were restored to baseline values in co-culture. This observation supports the idea that CM may reciprocally influence EC by secreting additional paracrine factors that impact NO production by EC. For example, the alarmin cytokine High Mobility Group Box 1 has been shown to be released by necrotic or injured CM [48
] and this protein is a ligand for the EC receptor for advanced glycosylation end-products (RAGE) [49
]. The latter has been shown to be upregulated during hypoxia and also negatively regulates NO production [50
]. The role of alarmins or of other candidate proteins (e.g. neuregulin, PDGF-B or angiopoietin-1) [51
] to modulate EC–CM crosstalk and influence CM survival after hypoxia and reoxygenation remains to be elucidated. Although isoflurane did not further improve NO production by EC and CM after hypoxia and reoxygenation, isoflurane did enhance CM survival in co-culture. This finding may suggest that NO production during hypoxia or the initial period of reoxygenation is the most critical in eliciting isoflurane-induced cardioprotection.
The results of the current investigation support the contention that paracrine and autorcrine factors contribute to protection of CM against hypoxia and reoxygenation injury. However, interpretation of the data is subject to certain limitations. We used neonatal CM rather than adult CM because neonatal CM can be readily cultured over several days and retain a beating phenotype. Although the expression of neonatal isoforms of certain proteins may not fully recapitulate the expression profile in adult CM, the protection of neonatal CM by isoflurane is consistent with that found in the adult heart. In addition, adult CM lose their ability to beat spontaneously when in culture and undergo morphological (decrease in cell size), electrophysiological (changes in cardiac ion channel profile and cell capacitance) and structural changes (myofibrillar damage, substantial decrease in T-tubule density), that might alter cellular signaling events [53
In conclusion, the results demonstrate that pharmacological induction of HIF1α in EC by a volatile anesthetic agent promotes cardioprotection of CM against hypoxia and reoxygenation injury, and this action is NO-dependent and associated with inhibition of mPTP opening in CM. The results further suggest that crosstalk between EC and CM contributes to cardioprotection.