J.-S. K. and S. O. contributed equally to this work.
Liver Biology and Pathobiology (Pathobiology)
Nitric oxide protects rat hepatocytes against reperfusion injury mediated by the mitochondrial permeability transition
Article first published online: 27 MAY 2004
Copyright © 2004 American Association for the Study of Liver Diseases
Volume 39, Issue 6, pages 1533–1543, June 2004
How to Cite
Kim, J.-S., Ohshima, S., Pediaditakis, P. and Lemasters, J. J. (2004), Nitric oxide protects rat hepatocytes against reperfusion injury mediated by the mitochondrial permeability transition. Hepatology, 39: 1533–1543. doi: 10.1002/hep.20197
- Issue published online: 27 MAY 2004
- Article first published online: 27 MAY 2004
- Manuscript Accepted: 11 FEB 2004
- Manuscript Received: 7 AUG 2003
- National Institutes of Health. Grant Numbers: 5-RO1 AG07218, 1 PO1 DK59340, 5-P30-DK34987, 1-P50-AA11605
We investigated the effects of nitric oxide (NO) on hepatocellular killing after simulated ischemia/reperfusion and characterized signaling factors triggering cytoprotection by NO. Cultured rat hepatocytes were incubated in anoxic Krebs-Ringer–HEPES buffer at pH 6.2 for 4 hours and reoxygenated at pH 7.4 for 2 hours. During reoxygenation, some hepatocytes were exposed to combinations of NO donors (S-nitroso-N-acetylpenicillamine [SNAP] and others), a cGMP analogue (8-bromoguanosine-3,5-cGMP [8-Br-cGMP]), and a cGMP-dependent protein kinase inhibitor (KT5823). Cell viability was determined by way of propidium iodide fluorometry. Inner membrane permeabilization and mitochondrial depolarization were monitored by confocal microscopy. SNAP, but not oxidized SNAP, increased cGMP during reperfusion and decreased cell killing. Other NO donors and 8-Br-cGMP also prevented cell killing. Both guanylyl cyclase and cGMP-dependent kinase inhibition blocked the cytoprotection of NO. However, 5-hydroxydecanoate and diazoxide— mitochondrial KATP channel modulators—did not affect NO-dependent cytoprotection or reperfusion injury. During reoxygenation, confocal microscopy showed mitochondrial repolarization, followed by depolarization, inner membrane permeabilization, and cell death. In the presence of either SNAP or 8-Br-cGMP, mitochondrial repolarization was sustained after reperfusion preventing inner membrane permeabilization and cell death. In isolated rat liver mitochondria, a cGMP analogue in the presence of a cytosolic extract and adenosine triphosphate blocked the Ca2+-induced mitochondrial permeability transition (MPT), an effect that was reversed by KT5823. In conclusion, NO prevents MPT-dependent necrotic killing of ischemic hepatocytes after reperfusion through a guanylyl cyclase and cGMP-dependent kinase signaling pathway, events that may represent the target of NO cytoprotection in preconditioning. (HEPATOLOGY 2004;39:1533–1543.)
Naturally occurring acidosis is highly protective against ischemic cell death.1–3 However, a return to physiological pH during reperfusion is an important factor that precipitates cell killing in a variety of cell types.1–6 A key mechanism underlying pH-dependent cell killing after reperfusion is onset of the mitochondrial permeability transition (MPT).3, 5, 6 Permeability transition (PT) pore opening abruptly increases mitochondrial permeability to small molecules with a molecular mass of up to 1,500 Da, which causes mitochondrial depolarization and uncoupling of oxidative phosphorylation.7 Uncoupling and activation of the uncoupler-stimulated mitochondrial adenosine triphosphate (ATP)ase then cause ATP depletion and necrotic cell death. Onset of the MPT can also initiate apoptotic cell death.5, 8–12 Because the MPT is the common event leading to both apoptosis and necrosis, cyclosporin A, a PT pore blocker, prevents both forms of cell death after ischemia/reperfusion.5
Nitric oxide (NO) is a diffusible signaling gas that both promotes and prevents cell injury, depending on the cell type and the pathological setting.13–21 In hepatocytes, NO prevents development of apoptosis by tumor necrosis factor α and Fas ligand.14, 15 NO signaling has also been implicated in protective ischemic preconditioning to myocardium.22 The target of NO cytoprotection in preconditioning is not known and may include stimulation of cyclic guanosine monophosphate (cGMP) formation, s-nitrosylation of degradative enzymes and caspases, reaction with injurious reactive oxygen species, and effects on mitochondrial respiration.14, 15, 23–25 Of particular interest are recent studies showing that NO can promote mitochondrial biogenesis in various cells26 and induce ischemic preconditioning in liver.25 However, the effects of NO on pH- and MPT-dependent cell death after reperfusion are unknown. Accordingly, the goal of the present study was to investigate the role of NO in MPT-dependent ischemia/reperfusion injury to cultured rat hepatocytes and to characterize the signaling mechanisms underlying NO-mediated cytoprotection. Our results indicate that increased levels of cGMP after exposure of hepatocytes to NO stimulates cGMP-dependent protein kinase (PKG), which in turn blocks onset of the MPT and subsequent necrotic cell death.
Materials and Methods
Hepatocyte Isolation and Culture.
Animals received humane care according to protocols approved by the Institutional Care and Use Committee of the University of North Carolina. Hepatocytes were isolated and cultured, as previously described.27 For cell killing and cGMP assays, aliquots (1 mL) of 1.5 × 105 cells were plated onto 24-well plates (Falcon, Lincoln Park, NJ). For confocal microscopy, 4.5 × 105 cells were cultured on 40-mm round glass coverslips in 60-mm culture dishes. All experiments were performed in Krebs-Ringer–hydroxyethylpiperazine-N-2 ethanesulfonic acid (HEPES) buffer (KRH) containing 115 mM NaCl, 5 mM KCl, 2 mM CaCl2, 1 mM KH2PO4, 1.2 mM MgSO4, and 25 mM HEPES (pH 6.2 or 7.4) at 37°C.
Simulation of Ischemia/Reperfusion.
To simulate ischemia, hepatocytes were incubated in KRH at pH 6.2 in an anaerobic chamber (Coy Laboratory Products, Ann Arbor, MI) for 4 hours. To simulate the reoxygenation and return to physiological pH of reperfusion, anaerobic KRH at pH 6.2 was replaced with aerobic KRH at pH 7.4. In some experiments, hepatocytes were incubated with the NO donors S-nitroso-N-acetylpenicillamine (SNAP), 1-[N-(2-aminoethyl)-N-(2-aminoethyl)amino]diazen-1-ium-1,2-diolate (DETA NONOate), or spermine 1,2-diolate (spermine NONOate) added at the beginning of reoxygenation. SNAP and NONOate were freshly prepared for each experiment in dimethyl sulfoxide and 0.1 mM NaOH, respectively, and were protected from light. Oxidized SNAP was prepared by exposing SNAP solution (0.1 M in dimethyl sulfoxide) to light at room temperature for 48 hours. In other experiments, hepatocytes were treated with 10 μM of the guanylyl cyclase inhibitor 1H-(1,2,4)-oxadiazolo-(4,3)quinoxalin-1-one (ODQ),14 50 μM of the cGMP analogue 8-Br-cGMP,28 5 μM of the cGMP-dependent kinase inhibitor KT5823,29 0.1 to 1 mM of the mitochondrial KATP channel blocker 5-hydroxydecanoic acid (5-HD),30 or 1 to 50 μM of the mitochondrial KATP channel activator diazoxide31 at 20 minutes before and then continuously after reperfusion. SNAP or oxidized SNAP were added only at the time of reperfusion.
Assay for Cell Death.
Cell death was assessed using propidium iodide fluorometry, as previously described.32
Measurement of cGMP.
Hepatocellular cGMP was determined by way of an acetylation method using a commercial kit according to the manufacturer's instructions (Amersham Pharmacia Biotech, Piscataway, NJ).
Hepatocyte Loading With Fluorescent Indicators.
Hepatocytes cultured on glass coverslips were coloaded with 300 nM of tetramethylrhodamine methyl ester (TMRM) and 1 μM of calcein-AM in KRH to monitor mitochondrial membrane potential and onset of the MPT, respectively, as previously described.3
The TMRM and calcein fluorescence were imaged using an inverted Zeiss 510 laser scanning confocal microscope equipped with a 63 × N.A. 1.4 oil-immersion planapochromat lens. Calcein and TMRM were excited with 488-nm and 543-nm light, respectively. Emission was divided by a 545-nm dichroic mirror directed through 500 to 530 band pass and 560-nm long-pass red barrier filters. Pinholes were set to Airy units of 1.0 in both channels.
Isolated Rat Liver Mitochondria.
Rat liver mitochondria were isolated in 250 mM of sucrose and 2 mM of K+-HEPES buffer, pH 7.4, as previously described.33
Preparation of Liver Cytosolic Extract.
A portion of fresh rat liver used for mitochondrial isolation was placed in ice-cold phosphate-buffered saline supplemented with a protease inhibitor cocktail [2.08 mM 4-(2-aminoethyl)benzenesulfonyl fluoride, 1.6 μM aprotinin, 80 μM bestatin, 30 μM pepstatin A, 28 μM E-64, and 40 μM leupeptin] and minced with scissors. Phosphate-buffered saline was added at a final dilution of 3 g liver/20 mL. The tissue was then homogenized in the cold at high speed with a Potter-Elvehjem tissue grinder utilizing pestle A (10 strokes) and then pestle B (3 strokes) mounted to a Fisher Steadi-Speed Stirrer. The homogenate was then centrifuged for 10 minutes. The pellet was discarded, and the supernatant was centrifuged for 1 hour at 100,000g. The fatty top layer was aspirated, and the supernatant was collected and stored on ice until used. All supernatants were prepared fresh and used within 5 hours of preparation.
Measurement of Ca2+-Induced Mitochondrial Swelling.
Mitochondrial swelling was assessed by decreased absorbance at 540 nm at 25°C, as previously described.33 In some experiments, mitochondrial suspensions were treated with 100 μM of 8-(4-chlorophenylthio)-cGMP (8-pCPT-cGMP), a cGMP analogue (Biolog, Bremen, Germany), 2.5 μg of cytosolic extract, and/or 50 μM of ATP 2 minutes prior to addition of 250 μM of Ca2+.
Differences between means were compared using the Student's t test using a level of significance of P < .05. Data were expressed as means ± SEM. All experiments are representative of at least three different cell or mitochondrial isolations.
Protection by SNAP Against Ischemia/Reperfusion-Induced Necrosis in Cultured Rat Hepatocytes.
After 4 hours of simulated ischemia, reperfusion was simulated by replacing anaerobic KRH at pH 6.2 with aerobic KRH at pH 7.4 to reoxygenate and restore physiological pH. After this reoxygenation, necrotic cell killing increased to more than 65% within 2 hours (Fig. 1A). When SNAP—an NO donor with a half-life of 6 to 10 hours at 37°C—was added at the time of reoxygenation at pH 7.4, cell killing was markedly diminished in a dose-dependent manner (Fig. 1A). SNAP at 50 to 200 μM protected against cell death, whereas higher concentrations became toxic (Fig. 1A).
To further investigate whether or not cytoprotection by SNAP was due to generation of NO, hepatocytes after 4 hours of ischemia were reperfused with 200 μM of oxidized SNAP that had lost its NO-generating capability.15 Reoxygenation at pH 7.4 of hepatocytes with oxidized SNAP did not protect against hepatocyte killing, which supports the conclusion that protection by SNAP against cell death induced by simulated ischemia/reperfusion was due to NO generated by SNAP (Fig. 1B). Cell killing in this model of reperfusion injury was strongly pH-dependent, because reoxygenation at pH 6.2 in the absence of SNAP prevented nearly all cell killing (Fig. 1B), as has been shown previously.3, 5
DETA NONOate is an NO donor that releases NO with a half-life of 20 hours at 37°C.34 To further support a protective role of NO in reperfused hepatocytes, cells were reoxygenated at pH 7.4 in the presence of various concentrations of DETA NONOate added at the time of reoxygenation (Fig. 2A). DETA NONOate also prevented cell killing after reperfusion in a dose-dependent manner. As with SNAP, high concentrations of DETA NONOate became toxic. Similarly, reoxygenation with a third NO donor—spermine NONOate with a half-life of about 40 minutes34—also decreased cell killing in a dose-dependent fashion (Fig. 2B).
Role of Guanylyl Cyclase and cGMP in NO-Dependent Cytoprotection.
One action of NO is to activate guanylyl cyclase and increase cellular cGMP.35, 36 To determine whether or not guanylyl cyclase was involved in NO-mediated cytoprotection, hepatocytes were exposed to 4 hours of anoxia at pH 6.2 and treated with 10 μM of ODQ, a guanylyl cyclase inhibitor, beginning 20 minutes before and then continuously after reoxygenation at pH 7.4. SNAP was then added at the beginning of reperfusion. ODQ alone did not affect cell viability after reoxygenation, indicating that ODQ was not promoting hepatocellular toxicity; however, ODQ completely abrogated the protective effects of SNAP during reoxygenation, and cell killing after reoxygenation with SNAP plus ODQ was virtually the same as after reoxygenation without either agent (Fig. 3A).
Reversal by ODQ of NO cytoprotection after simulated ischemia/reperfusion suggested a role of guanylyl cyclase–generated cGMP in the prevention of necrosis. To assess directly whether or not NO was promoting cGMP formation, we measured cGMP in hepatocytes before, during, and after simulated ischemia/reperfusion. cGMP in normoxic hepatocytes was 83.7 ± 6.7 fmol/106 cells (Fig. 4), similar to previously reported values.28 At the end of 4 hours of anoxia, cGMP remained close to normoxic levels; however, during the first 2 hours of reoxygenation it decreased progressively by 54%. By contrast, during reperfusion with SNAP, cGMP after 2 hours became 2.6-fold greater than without SNAP (Fig. 4). These findings support the conclusion that NO release by SNAP activated cGMP formation during reoxygenation. Cell killing was also assayed after reoxygenation of hepatocytes at pH 7.4 in the presence of 50 μM of 8-Br-cGMP, a soluble cGMP analogue. 8-Br-cGMP produced cytoprotection to nearly the same extent as SNAP (Fig. 3B).
Involvement of PKG in NO-Mediated Cytoprotection Against Ischemia/Reperfusion Injury to Hepatocytes.
PKG plays an important role in cGMP signaling pathways.36 To investigate the role of PKG, hepatocytes were reoxygenated in the presence of 5 μM of KT5823, a selective PKG inhibitor.14 KT5823 alone did not alter cell viability after simulated ischemia/reperfusion (Fig. 3C). By contrast, reperfusion with KT5823 reversed SNAP cytoprotection. Similarly, KT5823 reversed the cytoprotection of 8-Br-cGMP during reperfusion (data not shown). Taken together, these results support the conclusion that NO release by SNAP promotes guanylyl cyclase activation, cGMP formation, and cGMP-dependent PKG activation to lead to protection against reperfusion injury.
Lack of Involvement of Mitochondrial KATP Channels in NO-Dependent Cytoprotection Against Ischemia/Reperfusion to Hepatocytes.
NO may activate mitochondrial KATP channels, leading to ischemic preconditioning in the heart.37, 38 To determine involvement of mitochondrial KATP channels in NO-dependent cytoprotection, hepatocytes were subjected to simulated ischemia and treated with 5-HD (100–300 μM), a mitochondrial KATP channel blocker, beginning 20 minutes before and then continuously after reoxygenation at pH 7.4. SNAP was then added at the beginning of reperfusion. Reoxygenation with 5-HD plus SNAP did not alter NO-dependent cytoprotection (Fig. 5 and data not shown). 5-HD higher than 300 μM was cytotoxic by itself (data not shown). In other experiments, hepatocytes were reoxygenated with diazoxide (1–10 μM), a mitochondrial KATP channel opener, in the absence of SNAP, but no protection against cell death after ischemia/reperfusion was observed (Fig. 5 and data not shown). Diazoxide higher than 10 μM was cytotoxic (data not shown). These data did not support the conclusion that NO-mediated cytoprotection against ischemia/reperfusion injury to hepatocytes is associated with mitochondrial KATP channel opening.
Blockade of the MPT by SNAP After Simulated Ischemia/Reperfusion of Hepatocytes.
To examine whether or not cytoprotection by NO was linked to inhibition of the MPT, mitochondrial membrane potential and inner membrane permeability were monitored using laser scanning confocal microscopy. Hepatocytes were coloaded with TMRM and calcein. TMRM is a red cationic fluorescent dye that accumulates electrophoretically into mitochondria in response to the negative mitochondrial membrane potential.3 Calcein, a green fluorophore with a molecular weight of 623 Da, loads into the cytosol at 37°C and is excluded by mitochondria that have closed PT pores. In normal hepatocytes, polarized mitochondria take up red-fluorescing TMRM while simultaneously excluding green fluorescing calcein.3 Mitochondrial exclusion of calcein results in the appearance of many small, dark, round voids in images of green calcein fluorescence where each void is a single mitochondrion.
After 4 hours of simulated ischemia (anoxia at pH 6.2), mitochondria of hepatocytes were depolarized, and mitochondrial TMRM fluorescence was barely detectable (Fig. 6A). However, the depolarized mitochondria continued to exclude the green fluorescence of calcein, which indicated that PT pores were remaining closed (Fig. 6A). When anoxic hepatocytes were reoxygenated at pH 7.4, their mitochondria began to repolarize and take up TMRM within 5 minutes. Subsequently, after approximately 20 minutes of reperfusion, the mitochondria of reperfused hepatocytes lost their TMRM fluorescence (Fig. 6A). As depolarization occurred, calcein redistributed from the cytosol into the mitochondria, and the dark mitochondrial voids began to disappear. These events were indicative of onset of the MPT. After several more minutes, virtually all cellular calcein fluorescence was lost, which indicated failure of the plasma membrane permeability barrier and onset of necrotic cell death.
To assess whether or not NO blocks the MPT after reperfusion, TMRM and calcein fluorescence were monitored as hepatocytes were reoxygenated at pH 7.4 in the presence of SNAP (Fig. 6B). Soon after reperfusion with SNAP, repolarization of mitochondria again occurred, as judged by TMRM uptake (Fig. 6B). In contrast to reperfusion in the absence of SNAP, mitochondrial TMRM fluorescence persisted over the 2-hour period of reperfusion. Moreover, calcein did not redistribute from the cytosol into the mitochondria in the presence of SNAP, although partial redistribution seemed evident at later time points. The results indicated that NO substantially but perhaps not fully inhibited PT pore opening after simulated reperfusion (Fig. 6B). Similarly, reoxygenation with 8-Br-cGMP promoted sustained mitochondrial repolarization and inhibited inner membrane permeabilization (Fig. 6C).
Inhibition of the Mitochondrial Permeability Transition by cGMP in the Presence of Cytosolic Extract and ATP in Isolated Mitochondria.
When mitochondria were incubated in Ca2+-free assay medium, almost no decrease in absorbance was observed (data not shown). A virtually identical absorbance recording was obtained when mitochondria were incubated in Ca2+-free assay medium supplemented with ATP, 8-pCPT-cGMP (a cGMP analogue), and cytosolic extract (Fig. 7, trace a). By contrast, when 250 μM of Ca2+ was added to the unsupplemented assay medium, absorbance decreased more than 0.6 units, indicating large amplitude mitochondrial swelling and onset of the MPT (Fig. 7, trace b). Cyclosporin A (3 μM) blocked this swelling (data not shown). However, when Ca2+ was added to assay medium supplemented with ATP, 8-pCPT-cGMP, and cytosolic extract, swelling was markedly decreased (Fig. 7, trace c). This protection against onset of the MPT required the combination of ATP, 8-pCPT-cGMP, and cytosolic extract, because protection was lost if ATP, 8-pCPT-cGMP, or cytosolic extract was individually removed from the combination (Fig. 7, traces d, e, and f). KT5823 (1 μM), the PKG inhibitor, reversed protection against the MPT by the combination of ATP, 8-pCPT-cGMP, and cytosolic extract (Fig. 7, trace g), but KT5823 in the absence of Ca2+ did not cause mitochondrial swelling (data not shown). Taken together, these data strongly suggest that the combination of PKG (from the cytosolic extract), cGMP, and ATP acts to inhibit the MPT, most likely by a PKG-mediated protein phosphorylation reaction.
The naturally occurring acidosis of ischemia protects cells against anoxic killing, but restoration of normal pH during reperfusion precipitates cell death.1–4 Onset of the MPT is a causative event initiating both pH-dependent necrosis and apoptosis in reperfused hepatocytes.3, 5 We have shown that NO through a cGMP/PKG-dependent signaling pathway prevents necrotic cell death after simulated ischemia/reperfusion by inhibiting MPT-dependent mitochondrial inner membrane permeabilization (Fig. 8).
After 4 hours of simulated ischemia at pH 6.2, necrotic cell death occurred after reoxygenation at pH 7.4 but not at pH 6.2, confirming the pH-dependence of cell killing after ischemia/reperfusion to hepatocytes.3, 5 Reperfusion with SNAP, DETA NONOate, and spermine NONOate—NO donors with half-lives between 39 minutes and 20 hours—prevented this cell killing (Figs. 1A and 2). Because oxidized SNAP did not prevent cell killing after reperfusion (Fig. 1B), cytoprotection was attributed to NO. Pretreatment of hepatocytes with ODQ, a guanylyl cyclase inhibitor, and with KT5823, a PKG inhibitor, completely reversed cytoprotection by NO, suggesting a critical role of guanylyl cyclase and PKG in NO-mediated protection (Fig. 3A and C). The importance of the guanylyl cyclase was further supported by our findings that: (1) 8-Br-cGMP, a cGMP analogue, mimicked the protective effect of NO donors (Fig. 3B) and (2) SNAP increased intracellular cGMP levels during reoxygenation (Fig. 4). Furthermore, confocal microscopy of TMRM and calcein revealed that NO and 8-Br-cGMP blocked onset of the MPT and mitochondrial depolarization after reperfusion (Fig. 6). We also showed in isolated rat liver mitochondria that the combination of cGMP, cytosolic extract, and ATP inhibited the Ca2+-induced MPT, an effect that was reversed by the PKG inhibitor KT5823 (Fig. 7). By contrast, the mitochondrial KATP channel modulators 5-HD and diazoxide did not affect cell death after reperfusion (Fig. 5). These findings support the conclusion that NO activates guanylyl cyclase, which in turn increases hepatocellular cGMP and subsequently activates PKG, leading to inhibition of the MPT and necrotic cell death (Fig. 8). NO-stimulated PKG would appear, therefore, to exert a direct effect on the MPT by phosphorylating serine/threonine protein residues, possibly in the mitochondria themselves, but the exact target of PKG-mediated phosphorylation and its relation to PT pore function is unknown.
One action of NO is to bind to the heme prosthetic domain of the guanylyl cyclase to activate the conversion of GTP to cGMP.36 Cytoprotective effects of cGMP are reported in a variety of cell types and pathological conditions.14, 18, 29, 39, 40 In the liver, cGMP suppresses tumor necrosis factor α–induced apoptosis by decreasing caspase 3 activity.14 The present study shows that exogenous NO exposure during simulated reperfusion inhibits PT pore opening in mitochondria by a guanylyl cyclase–, cGMP- and PKG-dependent cascade. Previously cGMP and PKG were shown to inhibit mitochondrial depolarization, cytochrome c release, and apoptosis in rat astrocytes and the MPT in isolated rat brain mitochondria induced by hydrogen peroxide.18 Additionally, depending on NO concentration, NO suppresses PT pore opening in isolated liver mitochondria due to respiratory inhibition, mitochondrial depolarization, and inhibition of mitochondrial Ca2+ accumulation.23
In the present study, we show that cGMP also inhibits the MPT in isolated rat liver mitochondria, provided that ATP and a small amount of cytosolic extract are also present (Fig. 7). This extract presumably contains PKG, because cGMP- and ATP-dependent protection was reversed by KT5823. These results directly support the conclusion that PKG blocks the MPT in liver mitochondria.
Ischemic preconditioning was first described in the myocardium but occurs in many other organs, including the liver, brain, muscle, and small intestine.22, 41–45 Myocardial protection through ischemic preconditioning is biphasic, with an early phase developing within minutes and a late phase occurring after 12 to 24 hours.22 Whether or not a similar biphasic phenomenon occurs in the liver is not known. A major mechanism in ischemic preconditioning is NO signaling, especially in the late phase of myocardial preconditioning.46–48 In the liver, ischemic preconditioning leads to adenosine A2A receptor activation and subsequent kinase signaling.45, 49 As in the heart, NO is implicated in hepatic preconditioning. One effect of NO is vasodilatative, to improve hepatic reoxygenation after reperfusion.50 Other recent data show that treatment of hepatocytes with methylamine hexamethylene methylamine NONOate—an NO donor with a half-life of 1 minute—prior to ischemia mimics hepatic preconditioning through guanylyl cyclase/PKG-dependent activation of p38 MAP kinase.25 Future studies will be needed to determine a possible role of p38 MAP kinase in the guanylyl cyclase/PKG-mediated blockade of the MPT.
Here, we show that NO protects against simulated ischemia/reperfusion injury when used after ischemia only during the reperfusion phase rather than prior to onset of the ischemic stress. This protection applied at reperfusion was mediated through guanylyl cyclase/PKG-dependent inhibition of the MPT. Although activation of mitochondrial KATP channels by NO has been suggested to trigger ischemic preconditioning in the heart,38 our data did not support involvement of mitochondrial KATP channels in NO-mediated cytoprotection of hepatocytes against ischemia/reperfusion injury, because the KATP channel blocker, 5-HD, did not reverse NO-dependent cytoprotection and the KATP channel opener, diazoxide, did not mimic NO protection (Fig. 5).
Whether NO protects or promotes tissue injury in different pathological conditions is the subject of ongoing controversy.14, 21, 23, 48 Treatment of hepatocytes with NO donors and overexpression of NO synthases block apoptosis induced by tumor necrosis factor α and Fas ligand by way of caspase inactivation, blockade of cytochrome c release from mitochondria, and prevention of mitochondrial depolarization.14 By contrast, NO formation can be cytotoxic in various tissues.51–53 The cytoprotective and cytotoxic roles of NO appear to depend on NO concentration, with lower, more physiological concentrations promoting cytoprotection and higher concentrations causing toxicity. In the present study, NO effects were also concentration-dependent. For the NO donor SNAP, cytoprotection was observed at concentrations between 50 and 200 μM, whereas higher concentrations were cytotoxic (Fig. 1A). For DETA NONOate, concentrations between 100 and 800 μM were cytoprotective, but a concentration of 2000 μM was cytotoxic (Fig. 2A). Similarly, the short-lived NO donor spermine NONOate conferred cytoprotection only at low concentrations (5–10 μM) (Fig. 2B). The actual concentration of NO generated by different NO donors depends on temperature, donor concentration, half-life, and other factors. Published data with the NO donors used here indicate that free NO concentrations associated with cytoprotection were in the range of 1 to 2.5 μM,16 which represents a high physiological amount.
NO release by SNAP activated cGMP formation. In the absence of SNAP, cGMP declined after reperfusion, whereas cGMP increased in the presence of SNAP (Fig. 4). A NO concentration of approximately 250 nM causes half maximal stimulation of guanylyl cyclase.54 Because maximal stimulation requires about five times the half-maximal concentration, our estimated free NO concentration of 1 to 2.5 μM during cytoprotection is consistent with NO-dependent guanylyl cyclase activation. However, much lower NO concentrations causing half-maximal activation of guanylyl cyclase have recently been reported.55 Because hepatocytes have numerous heme-containing proteins that bind NO, intracellular free NO may be lower than the extracellular NO in our experiments. Moreover, formation of reactive oxygen species increases after reperfusion. Superoxide reacts rapidly with NO, which might further decrease intracellular NO relative to extracellular NO. Although intracellular free NO concentration cannot be determined precisely, NO-dependent cytoprotection seems most likely mediated through guanylyl cyclase activation, because (1) cGMP increased after reperfusion with an NO donor (Fig. 4), (2) the soluble cGMP analogue, 8-Br-cGMP, mimicked the cytoprotection of NO (Fig. 3B), (3) the guanylyl cyclase inhibitor, ODQ, reversed NO cytoprotection (Fig. 3A), and (4) cGMP-dependent inhibition of the MPT was observed in isolated rat liver mitochondria (Fig. 7).
NO also acts directly to inhibit the MPT by way of a mechanism involving respiratory inhibition, membrane depolarization, and inhibition of mitochondrial Ca2+ uptake.23 Such a direct effect of NO on mitochondria might also contribute to the cytoprotection of NO, because elevation of mitochondrial Ca2+ is a key event inducing the MPT in oxidative stress to hepatocytes.56 In the present study, however, cGMP/PKG-dependent signaling appears to be the major mechanism of cytoprotection, because soluble cGMP (Fig. 3B) mimicked the cytoprotection of NO and NO cytoprotection was completely reversed by both guanylyl cyclase and PKG inhibition (Fig. 3A and C).
In conclusion, NO prevents pH- and MPT-dependent cell death after ischemia/reperfusion by a guanylyl cyclase/cGMP/PKG-mediated signaling cascade (Fig. 8). In this pathway, NO activates the guanylyl cyclase, leading to elevation of intracellular cGMP. Increased cGMP subsequently stimulates PKG, leading to inhibition of PT pores. Since onset of the MPT is a causative factor inducing cell death after reperfusion, inhibition of the MPT by this signaling cascade prevents hepatocellular death after reperfusion.