YOSHIYA NISHIMURA, 1 LEWIS H. ROMER, 1-3 AND JOHN J. LEMASTERS 1

Transcription

1 Mitochondrial Dysfunction and Cytoskeletal Disruption During Chemical Hypoxia to Cultured Rat Hepatic Sinusoidal Endothelial Cells: The ph Paradox and Cytoprotection by Glucose, Acidotic ph, and Glycine YOSHIYA NISHIMURA, 1 LEWIS H. ROMER, 1-3 AND JOHN J. LEMASTERS 1 We investigated mechanisms underlying death of cultured rat liver sinusoidal endothelial cells exposed to chemical hypoxia with KCN (2.5 mmol/l) to simulate the adenosine triphosphate (ATP) depletion and reductive stress of anoxia. During chemical hypoxia, acidotic ph prevented cell death. Glucose ( mmol/l) also prevented cell killing. Cytoprotection by glucose but not acidosis was associated with prevention of ATP depletion. After 4 hours of chemical hypoxia at ph 6.2 (simulated ischemia), rapid cell death occurred when ph was restored to ph 7.4 with or without washout of KCN (simulated reperfusion). This ph-dependent reperfusion injury (ph paradox) was prevented after KCN washout at ph 6.2. Glycine (0.3-3 mmol/l) also prevented the ph paradox, but glucose did not. The initial protection by acidotic ph and glycine during simulated reperfusion was lost when ph was later restored to 7.4 or glycine was subsequently removed. Mitochondria depolarized during chemical hypoxia. After washout of cyanide, mitochondrial membrane potential ( ) did not recover in cells that subsequently lost viability. Conversely, those cells that repolarized after cyanide washout did not subsequently lose viability. The actin cytoskeleton and focal adhesions became severely disrupted during chemical hypoxia at both ph 6.2 and 7.4 and did not recover after cyanide washout under any condition. Glucose during chemical hypoxia prevented cytoskeletal disruption. In conclusion, endothelial cell damage during simulated ischemia/reperfusion involves mitochondrial dysfunction, ATP depletion, and ATP-dependent cytoskeletal disruption. Glycine and acidotic ph prevented cell killing after reperfusion but did not reverse mitochondrial injury or the profound disruption to the cytoskeleton. (HEPATOLOGY 1998;27: ) Abbreviations: ATP, adenosine triphosphate; GBSS, Gay s balanced salt solution; KRH, Krebs-Ringer-Hepes buffer; TMRM, tetramethylrhodamine methylester;, membrane potential. From the Departments of 1 Cell Biology and Anatomy, 2 Pediatrics, and 3 Anesthesiology, University of North Carolina at Chapel Hill, Chapel Hill, NC. Received August 1, 1997; accepted December 17, Supported in part by grants DK-37034, HL-03299, and DK from the National Institutes of Health and a Grant-In-Aid from the American Heart Association. Address reprint requests to: Dr. John J. Lemasters, Department of Cell Biology and Anatomy, University of North Carolina at Chapel Hill, Campus Box 7090, 236 Taylor Hall, Chapel Hill, NC Fax: (919) Copyright 1998 by the American Association for the Study of Liver Diseases /98/ $3.00/ Ischemia/reperfusion injury is an important problem in liver surgery and systemic shock. Postoperative graft dysfunction and failure after liver transplantation also involve ischemia/reperfusion injury. 1-3 In previous studies, sinusoidal endothelial cells were shown to be a critical target of this reperfusion injury. 4-6 However, the underlying mechanisms of reperfusion-induced endothelial injury remain unclear. During ischemia, tissue ph becomes acidotic due to anaerobic glycolysis and adenosine triphosphate (ATP) hydrolysis. Acidosis protects cells from injury during hypoxia, 7-12 as does glycolytic metabolism generating ATP. 10,13-15 However, when extracellular ph recovers to physiological levels after hypoxia, rapid cell death occurs, a phenomenon called the ph paradox. 9,11,12,16,17 Because correction of acidosis occurs rapidly after reperfusion, the ph paradox may be a major cause of reperfusion injury. Glycine also strongly protects against hypoxic and toxic injury to many cell types. 12,18-22 Previously, we reported that glycine prevented reperfusioninduced endothelial cell killing after liver preservation for tissue transplantation surgery and improved graft function and graft survival after liver transplantation in rats. 23,24 The cytoskeleton regulates cell motility, the intracellular distribution of organelles, and overall cell shape and topography. In vascular endothelium, the cytoskeleton is essential for maintenance of barrier function and the regulation of transendothelial permeability During cold ischemic liver storage for transplantation, sinusoidal endothelial cells round up and detach from their underlying attachments. 1,28 Such changes suggest disruption of the cytoskeleton. In other cell types, ATP depletion also causes perturbation of the cytoskeleton. 29,30 However, the relation of cytoskeletal disruption to ATP depletion and cell injury in sinusoidal endothelial cells is not well studied. Because perturbation of structure and loss of viability of endothelial cells are initial events leading to liver graft failure from preservation injury, our aim was to characterize various protective strategies against cell killing and cytoskeletal disruption using a model of ischemia/reperfusion injury to cultured sinusoidal endothelial cells. MATERIALS AND METHODS Isolation of Sinusoidal Endothelial Cells. Sinusoidal endothelial cells of male Sprague-Dawley rats ( g) were isolated and purified by collagenase perfusion and centrifugal elutriation by modification of the procedures described by Braet 31 and Knook. 32 Under pentobarbital anesthesia, the abdomen was opened, a catheter was inserted into the portal vein, and the liver was perfused with Ca 2 -free buffer

2 1040 NISHIMURA, ROMER, AND LEMASTERS HEPATOLOGY April 1998 (in mmol/l: NaCl 118, NaHCO 3 25, KCl 5, KH 2 PO 4 1.2, HEPES 25, EGTA 0.5 [ph 7.4]) at 20 ml/min for 10 minutes, followed with Gay s balanced salt solution (GBSS) (in mmol/l: NaCl 136, NaHCO 3 2.6, KCl 5, CaCl 2 1.5, MgCl 2 1, Na 2 HPO 4 0.8, KH 2 PO 4 0.2, MgSO 4 0.3, glucose 5.6, HEPES 25 [ph 7.4]) containing 0.05% collagenase and 5% fetal bovine serum at 10 ml/min for 20 minutes. Fetal bovine serum was used to inactivate nonspecific proteases. Each solution was warmed in a water bath at 37 C and bubbled with 95% O 2 /5% CO 2. After perfusion, the liver was excised and placed into a 10-cm culture dish on ice. Glisson s capsule and other connective tissue were picked up by tweezers and discarded. The suspension was then filtered through 60-µm Nylon mesh, GBSS was added to a final volume of 50 ml, and the suspension was centrifuged and resuspended three times in GBSS at 80g for 2 minutes. The nonparenchymal cells in the pooled supernatants were pelleted by centrifugation at 1,000g for 5 minutes and resuspended in 10 ml GBSS. Ten milliliters of a Percoll solution (mixture of 1 volume of 10 phosphate-buffered saline [in mm: NaCl 138, KCl 2.5, Na 2 HPO 4 8, KH 2 PO 4 1.5] and 9 volumes of Percoll) were then added with thorough mixing. The resulting cell suspension in 45% Percoll was added to two 15-mL centrifugal tubes, and 2 ml of phosphatebuffered saline was gently layered onto the top of each tube. After 10 minutes of centrifugation at 1,000g, the endothelial cell rich layer at the Percoll phosphate-buffered saline interface was carefully aspirated and diluted into 50 ml GBSS. The resulting suspension was centrifuged at 1,000g for 5 minutes, and the pellet was resuspended in 10 ml GBSS. This suspension was further purified by a centrifugal elutriation at 4 C (JE-6 elutriation rotor in J-21-B centrifuge, Beckman Instruments, Palo Alto, CA). At a rotor speed of 2,500 rpm, the cell suspension was loaded at 11 ml/min for 20 minutes. Flow rate was then increased to 18 ml/min, and 150 ml of effluent was collected. This final cell suspension was centrifuged at 1,000g for 5 minutes and resuspended in RPMI 1640 with 20% fetal FIG. 2. Protection by glucose against cyanide-induced endothelial cell killing. Endothelial cells were incubated in KRH at ph 7.4 with 2.5 mmol/l KCN and the indicated concentration of glucose. Viability was monitored, as described in Fig. 1. Data points are means SEM of 8 to 12 experiments from three to four cell isolations. After 240 minutes and longer, cell viability was significantly increased with 0.3 to 10 mmol/l glucose versus no glucose (P.01). FIG. 1. Protection by acidotic ph against cyanide-induced endothelial cell killing. One-day cultured endothelial cells were incubated in KRH with 2.5 mmol/l KCN at various values of ph. Cell viability was monitored by propidium iodide fluorometry, as described in Materials and Methods. Data points are means SEM of 8 to 12 experiments from three to four cell isolations. After 120 minutes and longer, data points for ph 6.2, 6.6, 7.0, and 7.8 were significantly different from ph 7.4 (P.01). bovine serum, 2 mmol/l glutamine, 1 µmol/l phorbol myristate acetate, and antibiotics. Cell viability was assessed by Trypan blue exclusion and was routinely greater than 90%. Phorbol myristate acetate was used to increase viability in culture without altering sinusoidal endothelial cell structure or functions. 33 Cells were seeded on either type I rat tail collagen-coated 24-well multiwell plates or glass coverslips. After 90 minutes of incubation in 5% CO 2 /air at 37 C, the medium was changed to wash out unattached cells, and the cells were further cultured overnight. Purity of endothelial cells was greater than 95% as assessed by ovalbumin uptake. 34 Chemical Hypoxia and Simulated Reperfusion. To simulate the ATP depletion and reductive stress of hypoxia/ischemia, overnight cultured endothelial cells were incubated in Krebs-Ringer-HEPES buffer (KRH) (in mmol/l: NaCl 115, KCl 5, CaCl 2 2, KH 2 PO 4 1, MgSO 4 1.2, HEPES 25 [ph 7.4]) and exposed to 2.5 mmol/l KCN, a reversible inhibitor of mitochondrial cytochrome oxidase (chemical hypoxia). 10,35-37 These conditions of chemical hypoxia were reversed by replacing the cyanide-containing buffer with several changes of cyanide-free KRH containing 10 mmol/l glucose (simulated reperfusion). In some experiments, ph during simulated hypoxia/ischemia and/or reperfusion was varied between ph 7.8 and ph 6.2 by adjusting with HCl. Assessment of Cell Viability. Viability of cells in 24-well multiwell plates was assessed by propidium iodide fluorometry. 38 During incubation with 30 µmol/l propidium iodide, fluorescence intensity was read with a fluorescence scanner (CytoFluor 2300, Millipore, Bedford, MA) at an excitation wavelength of 546 nm (40-nm band pass) and an emission wavelength of 620 nm (50-nm band pass). At the end of experiments, 375 µmol/l digitonin was added to cause 100% cell death. Cell viability was then calculated as: 100 (X A)/(B A) 100

3 HEPATOLOGY Vol. 27, No. 4, 1998 NISHIMURA, ROMER, AND LEMASTERS 1041 Measurement of Cellular ATP. Cellular ATP was measured by a bioluminescence assay, as described previously. 40 Briefly, cultured cells were treated with 1 ml of cold 11.7 mol/l perchloric acid and detached with a rubber scraper. The extract was centrifuged for 2 minutes at 4,000g in a table-top centrifuge. The supernatant was removed, neutralized with 5 mol/l KOH/0.5 mol/l imidazole, and centrifuged again. ATP in the supernatant was then measured using a luciferase assay kit (Promega, Madison, WI) and a luminometer (Optocomp I, MGM Instrument, Hamden, CT). 41,42 Cytoskeletal Staining. Cells were seeded on 12-mm coverslips and incubated under various conditions as described above. The cells were fixed with 3.7% formaldehyde in 0.1 mol/l sodium phosphate buffer for 20 minutes at ph 7.4 or ph 6.2 according to the experimental condition before fixation, rinsed in phosphatebuffered saline, and permeabilized for 2 minutes with triton X-100 in Tris-buffered saline (in mmol/l: NaCl 150, Tris-HCl 50, NaN 3 15 [ph 7.6]) before staining for immunolabeling. The coverslips were then incubated for 60 minutes at 37 C with a monoclonal antivinculin antibody (7F9, gift of Dr. Alexey Belkin, University of North Carolina at Chapel Hill), rinsed with Tris-buffered saline, and stained with a combination of rhodamine-conjugated, affinitypurified, donkey anti-mouse immunoglobulin G antibody (Chemicon, Temecula, CA) and fluorescein-conjugated phalloidin (Molecular Probes, Eugene, OR) in Tris-buffered saline for 60 minutes at 37 C. Following the antibody incubations, the coverslips were washed in Tris-buffered saline, rinsed in deionized water, and mounted with Mowiol (Calbiochem, San Diego, CA). Coverslips were viewed on an Olympus BX60 microscope equipped for FIG. 3. ph paradox of reperfusion injury to endothelial cells. Sinusoidal endothelial cells were incubated in KRH with 2.5 mmol/l KCN at ph 6.2 for 4 hours. Subsequently, cyanide was washed out with cyanide-free KRH containing 10 mmol/l glucose at ph 6.2 or 7.4. In one group, ph of the washout buffer was changed to 7.4 without removal of KCN. Viability was monitored as described in Fig. 1. Data points are means SEM of 11 experiments from four cell isolations. *P.01 vs. ph 7.4. where X is fluorescence intensity at each time point, A is initial fluorescence, and B is fluorescence after digitonin. Cell viability determined by this assay is equivalent to that assessed by Trypan blue labeling and lactate dehydrogenase release. 10,38 Results are expressed as means SEM. The significance of differences between groups was analyzed with ANOVA, followed by Bonferroni s test. Imaging of Mitochondrial Membrane Potential and Loss of Cell Viability. Mitochondrial membrane potential was visualized by uptake of tetramethylrhodamine methylester (TMRM), as previously described. 39 One-day cultured cells on collagen-coated glass coverslips were loaded with 500 nmol/l TMRM in culture medium for 15 minutes at 37 C for 15 minutes. After two washes with GBSS, the coverslips were transferred into a perifusable temperature-controlled chamber (Bioptechs Inc., Butler, PA) and mounted on the microscope stage. The chamber was filled with KRH containing 5 µmol/l propidium iodide and 100 nmol/l TMRM, the latter to maintain the equilibrium distribution of the cationic TMRM between mitochondria and the buffer. Nuclear labeling of propidium iodide indicated cell death. To simulate hypoxia/ischemia, the chamber was infused with KRH containing 2.5 mmol/l KCN, 100 nmol/l TMRM, and 5 µmol/l propidium iodide. Reperfusion was simulated by infusing cyanide-free KRH containing 10 mmol/l glucose, 500 nmol/l TMRM, and 5 µmol/l propidium iodide. TMRM was increased to 500 nmol/l during simulated reperfusion because virtually all TMRM was released during chemical hypoxia. Fluorescence of TMRM and propidium iodide was imaged with a Zeiss LSM 410 laser scanning confocal microscope (Thornwood, NY) using a 568-nm excitation light and a 590-nm long pass emission filter. FIG. 4. Protection by glycine against ph-dependent reperfusion injury to endothelial cells. Endothelial cells were incubated in KRH with 2.5 mmol/l KCN at ph 6.2 for 4 hours. Subsequently, KCN was washed out with KRH containing 10 mmol/l glucose at ph 7.4 and the indicated concentration of glycine. Viability was monitored as described in Fig. 1. Data points are means SEM of 10 to 11 experiments from four cell isolations. At all time points after cyanide washout, cell viability was significantly increased in the presence of 0.3 to 3 mmol/l glycine versus no glycine (P.01).

4 1042 NISHIMURA, ROMER, AND LEMASTERS HEPATOLOGY April 1998 epifluorescence. Fluorescence micrographs were taken on T-max 400 film (Eastman Kodak Co., Rochester, NY). RESULTS Protective Effect of Acidosis Against Cyanide-Induced Cell Killing to Sinusoidal Endothelial Cells. Viability of sinusoidal endothelial cells exposed to 2.5 mmol/l KCN (chemical hypoxia) was measured using propidium iodide fluorometry. At ph 7.4, cell viability began to decrease after 3 hours and fell to about 10% after 6 hours (Fig. 1). Decreased ph strongly protected sinusoidal endothelial cells against KCN-induced killing. At a ph of 6.6 or less, cell killing was completely blocked up to 6 hours. In contrast, more alkaline buffer (ph 7.8) accelerated cell killing (Fig. 1). Protective Effect of Glucose Against Cyanide-Induced Cell Killing to Sinusoidal Endothelial Cells. To determine whether glycolytic ATP production can prevent KCN-induced killing of sinusoidal endothelial cells, glucose was added during chemical hypoxia at ph 7.4. As shown in Fig. 2, glucose protected against KCN-induced cell killing in a dose-dependent manner. At a glucose concentration of 1 mmol/l or more, viability was greater than 80% after 6 hours. ph-dependent Reperfusion Injury to Sinusoidal Endothelial Cells. To simulate the ATP depletion, reductive stress, and acidosis of ischemia, endothelial cells were incubated with 2.5 mmol/l KCN at ph 6.2 for 4 hours. To simulate the reversal of respiratory inhibition and recovery of ph during reperfusion, KCN was washed away at ph 7.4 with glucose-containing buffer. Although little cell killing occurred during simulated ischemia, cell viability decreased to less than 30% within 2 hours of simulated reperfusion at ph 7.4 (Fig. 3). By contrast, when KCN was washed out at ph 6.2, virtually no cell killing occurred. Moreover, when buffer ph was changed to ph 7.4 without washing out KCN, cell death occurred to the same extent as in the absence of KCN (Fig. 3). By comparison, when ph remained at ph 6.2 in the presence of KCN, little cell killing occurred (Fig. 1). These data indicated that reperfusion injury in this model was dependent on the recovery of ph to 7.4 rather than on the washout of KCN. Protective Effect of Glycine Against ph-dependent Reperfusion Injury to Sinusoidal Endothelial Cells. To assess protection by glycine against the ph-dependent reperfusion injury, we exposed endothelial cells to KCN at ph 6.2 for 4 hours and FIG. 5. Mitochondrial polarization and loss of viability in endothelial cells during simulated ischemia and reperfusion. Endothelial cells were loaded with TMRM (500 nmol/l) and incubated in KRH with 100 nmol/l TMRM and 5 µmol/l propidium iodide at ph 7.4. After collecting a baseline confocal fluorescence image, the buffer was changed to KRH containing 2.5 mmol/l KCN, 100 nmol/l TMRM, and 5 µmol/l propidium iodide at ph 6.2. After 2 hours, cyanide was replaced with washes of fresh KRH containing 10 mmol/l glucose, 500 nmol/l TMRM, and 5 µmol/l propidium iodide at ph 7.4. In the baseline image, punctate labeling by TMRM showed normal polarized mitochondria (upper left). After cyanide addition, mitochondrial fluorescence decreased substantially after 15 minutes (upper middle) and disappeared almost completely after 120 minutes (upper right). However, none of the cells lost viability, as shown by the absence of nuclear labeling with propidium iodide. After cyanide washout and restoration of ph 7.4 (simulated reperfusion), some cells recovered mitochondrial polarization (mitochondrial TMRM fluorescence) almost immediately (lower left). Cells that did repolarize lost viability (arrowheads, lower panels).

5 HEPATOLOGY Vol. 27, No. 4, 1998 NISHIMURA, ROMER, AND LEMASTERS 1043 then washed out KCN at ph 7.4 in the presence of various concentrations of glycine. As shown in Fig. 4, glycine prevented lethal reperfusion injury in a dose-dependent fashion. At a glycine concentration of 1 mmol/l or more, virtually no cell killing occurred after restoration of ph. Half-maximal protection occurred at about 0.3 mmol/l glycine. Notably, protection against reperfusion injury did not require incubation with glycine before or during chemical hypoxia. Mitochondrial Membrane Potential During Simulated Ischemia/ Reperfusion. Mitochondrial was monitored by TMRM uptake. TMRM is a cationic fluorophore that is taken up electrophoretically by mitochondria in response to their negative. 43 During incubation in the absence of KCN, TMRM produced punctate labeling of mitochondria observed by confocal microscopy. This labeling indicated the intracellular distribution and polarization of mitochondria (Fig. 5, upper left). TMRM labeling disappeared during chemical hypoxia at ph 6.2, which indicated loss of mitochondrial. Mitochondrial fluorescence decreased markedly within 15 minutes (Fig. 5, upper middle) and was virtually absent after 120 minutes of chemical hypoxia (Fig. 5, upper right). When after 30 minutes, the cyanide-containing buffer was washed out with fresh buffer at ph 7.4, mitochondrial TMRM fluorescence was promptly restored in nearly all cells (data not shown). This observation demonstrated the reversibility of respiratory inhibition by cyanide in agreement with earlier studies in hepatocytes and cardiac myocytes. 36,37 By contrast, when cyanide was removed after 2 hours of exposure, recovery of was heterogeneous and correlated with cell viability (Fig. 5, lower panels). Cells whose mitochondrial did not recover after cyanide washout subsequently lost viability and were labeled with propidium iodide within 60 minutes. Oppositely, in cells that retained viability, mitochondrial recovered, as indicated by the return of punctate TMRM fluorescence. This return of punctate TMRM fluorescence was rapid and occurred within 2 minutes after KCN washout (Fig. 5, lower left panel). Overall, in experiments with cultured sinusoidal endothelial cells from four isolations, recovery of occurred in 65 of 142 cells after cyanide washout at ph 7.4. Of these cells, 64 retained viability after 60 minutes. By contrast, of 41 cells that lost viability, 40 did not recover mitochondrial immediately after simulated reperfusion. In other experiments, KCN was washed out at ph 6.2. At ph 6.2, lethal cell injury was prevented, but mitochondrial did not recover in most cells (Fig. 6, upper right and lower left panels). When ph was subsequently increased to 7.4 after another hour, most cells that had not recovered mitochondrial lost viability, whereas most cells that had regained mitochondrial retained viability (Fig. 6, lower right panel). Similarly, when KCN was washed out and replaced with KRH containing glycine at ph 7.4, cell killing was prevented (Fig. 7, upper right and lower left panels). Glycine protected against lethal ph-dependent reperfusion injury, but the amino acid did not support recovery of mitochondrial in the majority of cells (Fig. 7, upper right and lower left panels). When glycine was subsequently removed, those cells that had not recovered mitochondrial rapidly lost viability (Fig. 7, lower right panel). After FIG. 6. Mitochondrial repolarization and cell viability after reperfusion at acidotic ph. TMRMloaded endothelial cells were incubated in KRH with 2.5 mmol/l KCN, 100 nmol/l TMRM, and 5 µmol/l propidium iodide at ph 6.2 for 2 hours. Subsequently, cyanide was removed with washes of fresh KRH containing 10 mmol/l glucose, 500 nmol/l TMRM, and 5 µmol/l propidium iodide at ph 6.2. One hour later, the ph of the experimental buffer was raised to ph 7.4. Mitochondrial and cell death were monitored as described in Fig. 5. At the end of 2 hours of chemical hypoxia at ph 6.2, all punctate mitochondrial TMRM fluorescence was lost (upper left). After washout at ph 6.2, mitochondrial polarization (TMRM labeling) returned in only one cell in the field (upper right, arrow) within 2 minutes. However, no loss of viability occurred even after 60 minutes (lower left). After raising ph to 7.4, almost all cells lost viability within 15 minutes, as revealed by nuclear propidium iodide staining, and these cells had not recovered mitochondrial polarization. By contrast, the cell that had recovered mitochondrial polarization retained viability (lower right, arrow).

6 1044 NISHIMURA, ROMER, AND LEMASTERS HEPATOLOGY April 1998 simulated reperfusion at ph 6.2 or in the presence of glycine, 12 of 109 cells from three isolations recovered mitochondrial, and these cells regaining mitochondrial depolarization also retained viability even after removal of cytoprotection (change of ph to 7.4 or removal of glycine). Of 97 cells that did not repolarize when cyanide was washed out at ph 6.2 or at ph 7.4 with glycine, 69 cells lost viability within 20 minutes. Thus, the association remained constant between recovery of after protective reperfusion and the retention of viability after the protective factor was removed. Cells cultured on glass coverslips were exposed to 2 hours of chemical hypoxia at ph 6.2, after which lethal reperfusion occurred. In contrast, 4 hours of chemical hypoxia at ph 6.2 were needed for subsequent ph-dependent reperfusion injury to occur in cells cultured on plastic. Thus, cells cultured on glass were somewhat more vulnerable to injury than cells cultured on plastic. Lack of Protection by Cyclosporin A Against Reperfusion Injury to Sinusoidal Endothelial Cells. Cyclosporin A is reported to prevent cell death from reperfusion injury and oxidative stress in various cell types by blocking onset of the mitochondrial membrane permeability transition. 20,39,44-46 Accordingly, we assessed the ability of cyclosporin A to protect against ph-dependent reperfusion injury to sinusoidal endothelial cells. As shown in Fig. 8, cyclosporin A did not prevent lethal cell injury in the concentration range studied here (0.2-2 µmol/l). ATP Depletion During Chemical Hypoxia. During basal incubations, ATP content of cultured sinusoidal endothelial cells was 288 pmol/10 6 cells. This ATP concentration is consistent with amounts previously reported for other cells. 37,40 After exposure to KCN at either ph 6.2 or 7.4, ATP decreased to less than 5% of baseline within 15 minutes and less than 1% after 120 minutes (Fig. 9). ATP depletion was virtually identical at ph 7.4 and 6.2. Glucose (10 mmol/l) retarded the depletion of ATP. In the presence of glucose, ATP gradually decreased to about 20% of the baseline after 4 hours (Fig. 9). After 4 hours of chemical hypoxia at ph 6.2, ATP did not recover after washout at ph 7.4, ph 6.2, or at ph 7.4 with glycine (data not shown). Thus, ATP did not recover even when reperfusion injury was prevented by acidotic ph or glycine. Cytoskeletal Changes During Chemical Hypoxia and Simulated Reperfusion. Cultured sinusoidal endothelial cells formed monolayers with closely contiguous cell borders (Fig. 10A and 10C). The cells exhibited well-developed filamentous actin arrays at the cortical rim and dense peripheral bands. Multiple vinculin-containing focal adhesions and filamentous actin stress fibers were observed at the ventral surface. Chemical hypoxia at ph 7.4 severely disrupted cytoskeletal organization. Dense peripheral bands, stress fibers, and focal adhesions disappeared, leaving only fragments of cortical actin. Additionally, cellular retraction caused large gaps to appear between cells with a loss of monolayer continuity (Fig. 10B and 10D). These changes occurred within 2 hours. The marked deterioration of cytoskeletal structure caused by cyanide was prevented by inclusion of 10 mmol/l glucose in the buffer (Fig. 11A and 11C). By contrast, acidotic ph had only a modest protective effect against cytoskeletal disruption during chemical hypoxia, and the majority of discrete focal adhesions, dense peripheral bands, and stress fibers were lost or distorted (Fig. 11B and 11D). Compared with chemical FIG. 7. Mitochondrial repolarization and loss of viability after reperfusion with glycine. Endothelial cells were incubated in KRH with 2.5 mmol/l KCN, 100 nmol/l TMRM, and 5 µmol/l propidium iodide at ph 6.2 for 2 hours. Subsequently, cyanide was removed with washes of fresh KRH containing 10 mmol/l glucose, 500 nmol/l TMRM, 5 µmol/l propidium iodide, and 3 mmol/l glycine at ph 7.4. One hour later, the buffer was replaced with glycine-free solution that was otherwise identical. Mitochondrial polarization and cell death were monitored as described in Fig. 5. After 2 hours of chemical hypoxia at ph 6.2, virtually all TMRM fluorescence was lost, but no labeling with propidium iodide had occurred (upper left). After cyanide washout in the presence of glycine, a few cells recovered mitochondrial TMRM fluorescence within 2 minutes (arrows, upper right). All cells retained viability in the presence of glycine even after 60 minutes of simulated reperfusion (lower left). When glycine was subsequently removed, those cells that had not recovered mitochondrial polarization went on to lose viability, as shown by nuclear labeling with propidium iodide (lower right). By contrast, those cells that had repolarized retained viability (lower right, arrows).

7 HEPATOLOGY Vol. 27, No. 4, 1998 NISHIMURA, ROMER, AND LEMASTERS 1045 FIG. 8. Lack of protection by cyclosporin A against ph-dependent reperfusion injury to endothelial cells. Endothelial cells were incubated in KRH with 2.5 mmol/l KCN at ph 6.2 for 4 hours. Subsequently, KCN was removed with washes of KRH containing 10 mmol/l glucose at ph 7.4 and the indicated concentration of cyclosporin A. Viability was monitored as described in Fig. 1. Data points are means SEM of nine experiments from three cell isolations. triggered lethal reperfusion injury in cultured sinusoidal endothelial cells. This observation is consistent with previous reports in perfused livers, cultured hepatocytes, and other cells of a phenomenon called the ph paradox, whereby the return of ph from acidotic to normal after reperfusion, rather than reoxygenation, can precipitate lethal cellular reperfusion injury. 9,11,12,16,17 After exposure to KCN, the mitochondria of cultured sinusoidal endothelial cells depolarized as mitochondrial respiration became inhibited (Fig. 5). Simultaneously, ATP levels decreased profoundly (Fig. 9). Although cell killing was prevented at a ph of 6.6 or less, virtually complete ATP depletion and mitochondrial depolarization still occurred. After washout of KCN at ph 7.4, most cells failed to recover their mitochondrial. These cells subsequently lost viability (Fig. 5). When reperfusion occurred at ph 6.2 or with buffer containing glycine, lethal cell injury was almost completely prevented, but mitochondrial polarization recovered in only a few cells. Moreover, when extracellular ph was subsequently increased from ph 6.2 to 7.4 or glycine was removed, cells that had not recovered mitochondrial function lost viability rapidly (Figs. 6 and 7). By contrast, cells that had recovered mitochondrial polarization retained viability after restoration of ph 7.4 or removal of glycine. Thus, recovery of mitochondrial polarization was strongly predictive of retention of cell viability, whereas lack of mitochondrial recovery was predictive of cell death when ph was ultimately returned to ph 7.4 or glycine was removed. The continuing absence of mitochondrial polarization after cyanide washout suggests that the mitochondria had become uncoupled. The fact that glucose did not protect against cell killing after reperfusion supports a role for uncoupling, because glycolytic substrate hypoxia at ph 7.4, however, ph 6.2 did preserve some focal adhesions and dense peripheral bands. After cyanide washout following 4 hours of chemical hypoxia at ph 6.2, cytoskeletal organization did not recover, even when washout was at ph 6.2 or in the presence of glycine, conditions in which cell viability was retained (data not shown). DISCUSSION In the study described here, we simulated the reductive stress, ATP depletion, and acidosis of ischemia by incubating cultured sinusoidal endothelial cells with KCN at ph 6.2. To simulate the reversal of respiratory inhibition and the recovery of ph after reperfusion, we washed out cyanide at ph 7.4. Cyanide is a reversible respiratory inhibitor. In cultured hepatocytes and myocytes, washout of cyanide leads to recovery of structural and functional changes due to cyanide, such as bleb formation and ATP depletion. 36,37 Similarly, cyanide inhibition of sinusoidal endothelial cells was reversible, because washout of cyanide after 30 minutes of exposure led to prompt recovery of mitochondrial. Our results showed that acidosis ( ph 6.6) protected against cyanide-induced killing of sinusoidal endothelial cells assessed by nuclear labeling of propidium iodide (Fig. 1), as shown previously in other cell types, notably hepatocytes. 10,47 However, recovery to ph 7.4 with or without washout of KCN rapidly induced cell killing (Fig. 3). These results showed that restoration of ph after simulated ischemia FIG. 9. ATP during chemical hypoxia and reperfusion. Endothelial cells incubated in KRH were exposed to 2.5 mmol/l KCN at ph 7.4, at ph 6.2, or at ph 7.4 with 10 mmol/l glucose. ATP was measured as described in Materials and Methods. Data are means SEM from three cell isolations. *P.05 vs ph 7.4 without glucose.

8 1046 NISHIMURA, ROMER, AND LEMASTERS HEPATOLOGY April 1998 protects against cell killing after respiratory inhibition but not after uncoupling. 47 This is supported by our observation that ATP did not recover after washout of cyanide, even when cell killing was prevented by acidotic ph or glycine. Support for the importance of ATP depletion to cell killing at ph 7.4 is the observation that glucose retarded ATP depletion during chemical hypoxia at ph 7.4 and protected cells against cyanide-induced cell death (Figs. 2 and 9). Previously, glucose was reported to retard lethal injury to cultured sinusoidal endothelial cells in cold buffer with or without cyanide. 14 Compared with these findings during cold hypoxia, the protective effect of glucose was more complete during warm hypoxia studied here. Taken together, the results of the present study strongly suggest that killing of sinusoidal endothelial cells during warm ischemic injury at ph 7.4 is ATP-dependent, as shown previously for cyanideinduced killing of hepatocytes. 48 Recently, a permeability transition of mitochondrial inner membrane was implicated in cell injury due to ischemia/ reperfusion and oxidative stress. 20,39,44-46 The mitochondrial permeability transition is caused by opening of a nonspecific high-conductance pore in the mitochondrial inner membrane. 49 Prevention of the mitochondrial permeability transition is implicated as the mechanism underlying cytoprotection by acidotic ph in reperfusion injury, because acidotic ph blocks the opening of this pore. 50,51 Previously, Fujii et al. reported that pretreatment of cultured sinusoidal endothelial cells with cyclosporin A, an inhibitor of permeability transition pore, protected against anoxia/reoxygenation injury. 52 By contrast, cyclosporin A did not prevent sinusoidal endothelial cell killing after reperfusion of livers stored for transplantation. 24 In the present study, cyclosporin A failed to prevent reperfusion injury (Fig. 8), which is consistent with data after cold storage. The time of hypoxia/ischemia stress and the consequent degree of injury may be the factor responsible for different findings with cyclosporin A in different studies. Fujii et al. 52 used a relatively short period of hypoxia (30 minutes) compared with 4 hours in the present study and 24 hours of cold ischemic storage used by Currin et al. 24 In this study, impaired mitochondrial function and loss of cell viability after reperfusion were closely related, suggesting an important role for mitochondrial dysfunction in reperfusion injury to sinusoidal endothelial cells, despite the lack of protection by cyclosporin A. If the mitochondrial membrane permeability transition was the basis for this injury, then the permeability change was so severe that cyclosporin A could not prevent it, which is consistent with observations in isolated mitochondria where cyclosporin A provided only temporary protection against onset of the mitochondrial permeability transition. 53 Glycine and acidotic ph prevented cell killing without FIG. 10. Actin cytoskeleton and focal adhesion during chemical hypoxia. Filamentous actin (A, B) and vinculin (C, D) were localized by fluorescein-conjugated phalloidin and immunofluorescence, respectively, as described in Materials and Methods. Untreated endothelial cells showed well-defined dense peripheral actin bands and stress fibers (A), and vinculin-containing focal adhesions (C). After 4 hours of exposure to KCN at ph 7.4, both polymerized actin filaments were disorganized, and focal adhesions had largely lost (B, D). Bar 20 µm.

9 HEPATOLOGY Vol. 27, No. 4, 1998 NISHIMURA, ROMER, AND LEMASTERS 1047 supporting mitochondrial function or restoring cellular ATP levels after reperfusion. These results indicate that glycine and acidotic ph act after mitochondrial dysfunction, as illustrated in Fig. 12. This interpretation is supported by the observation that subsequent removal of glycine or increasing ph to 7.4 after reperfusion quickly precipitated killing of sinusoidal endothelial cells. This is in agreement with previous reports showing that protection by glycine occurs at a late stage of cell injury without preventing ATP depletion Extracellular acidosis is also cytoprotective against cell injury caused by ischemia/reperfusion and various toxicants. 7-12,16,17,47,57 The mechanism of cytoprotection by extracellular acidosis remains incompletely understood, but protection is mediated by cytosolic acidification, which inhibits activation of degradative enzymes such as proteases and phospholipases. 58,59 Endothelial cell killing is a prominent feature of reperfusion injury to livers stored for transplantation. 1-6 Reperfusion with mildly acidic and glycine-containing solutions, such as Carolina rinse solution, improves endothelial cell viability and the function and survival of liver grafts after transplantation. 23,24,60 The results of the present study suggest that the cytoprotective effects of acidosis and glycine against lethal endothelial cell injury participates in graft improvement with Carolina rinse solution. Chemical hypoxia caused marked cytoskeletal disruption characterized by degradation of actin microfilaments, disappearance of focal adhesions, retraction of the cytoplasm, and the appearance of gaps between the cells. These cytoskeletal changes were related to ATP depletion, because glucose increased cellular ATP during cyanide exposure and prevented cytoskeletal disruption, whereas acidotic ph, which prevented cell death but not ATP depletion, produced only modest protection against cytoskeletal disruption. Moreover, when ph-dependent reperfusion injury was blocked by acidic ph or glycine, essentially no improvement of cytoskeletal organization could be observed. The cytoskeleton is essential to the barrier function of endothelium. Thus, although preventing cell death, glycine and acidotic ph did little to restore normal endothelial structure and function. Similarly, in livers stored ischemically for transplantation, rounding and cytosolic retraction of sinusoidal endothelial cells occurs. 1,28 Based on our observations in cultured endothelial cells, these changes in cold-stored livers may also involve an ATP-dependent disruption of the cytoskeleton. FIG. 11. Effect of glucose and acidotic ph on cytoskeletal disruption during chemical hypoxia. Endothelial cells were incubated for 4 hours in KRH with 2.5 mmol/l KCN and 10 mmol/l glucose at ph 7.4 (A, C) or with 2.5 mmol/l KCN at ph 6.2 (B, D). Cells were stained for filamentous actin and vinculin, as described in Fig. 10. In the presence of glucose, actin fibers and vinculin-containing focal adhesions were well preserved (A, C) compared with untreated cells (see Fig. 10A and 10C). By contrast, during chemical hypoxia at ph 6.2, preservation of cytoskeletal architecture was incomplete, despite the protection of acidosis against loss of cell viability (B, D). Bar 20 µm.

10 1048 NISHIMURA, ROMER, AND LEMASTERS HEPATOLOGY April 1998 permeability barrier that culminates in outright cell death. Glycine and acidotic ph prevent this penultimate event but do not reverse ATP depletion, cytoskeletal disruption, or mitochondrial depolarization. Thus, there are limitations to the cytoprotection of glycine and acidotic ph. Although both prevent reperfusion-induced cell killing, neither promotes restoration of the normal barrier and metabolic functions of the endothelium. Acknowledgment: The authors thank Natalie V. McLean for expert technical assistance. REFERENCES FIG. 12. Scheme of cellular injury to sinusoidal endothelial cells during chemical ischemia/reperfusion. See text for details. Our scheme of injury to sinusoidal endothelial cells during simulated ischemia and reperfusion is summarized in Fig. 12. Chemical hypoxia causes ATP depletion, which in turn produces cytoskeletal disruption. Glucose, by providing an alternative source of ATP, blocks these cytoskeletal changes and prevents other events, leading ultimately to cell death. In tissue ischemia, ph drops substantially due to ATP hydrolysis and anaerobic glycolysis of endogenous glycogen. This naturally occurring acidosis protects strongly against the onset of cell death. However, an apparent uncoupling of mitochondria nonetheless develops during ischemia that prevents these organelles from repolarizing when respiratory inhibition is removed. When cells are reperfused at ph 7.4, this mitochondrial injury promotes cell killing, because cells whose mitochondria regain their also retain viability after reperfusion. The rapid restoration of physiological ph after reperfusion leads to breakdown of the plasma membrane 1. Caldwell-Kenkel JC, Currin RT, Tanaka Y, Thurman RG, Lemasters JJ. Kupffer cell activation and endothelial cell damage after storage of rat livers: effect of reperfusion. HEPATOLOGY 1991;13: Lemasters JJ, Thurman RG. Reperfusion injury after liver preservation for transplantation. Annu Rev Pharmacol Toxicol 1997;37: Ikeda T, Yanaga K, Kishikawa K, Kakizoe S, Shimada M, Sugimachi K. Ischemic injury in liver transplantation: difference in injury sites between warm and cold ischemia in rats. HEPATOLOGY 1992;16: Otto G, Wolff H, David H. Preservation damage in liver transplantation: electron-microscopic findings. Transplant Proc 1984;16: McKeown CMB, Edwards V, Phillips MJ, Harvey PRC, Petrunka CN, Strasberg SM. Sinusoidal lining cell damage: the critical injury in cold preservation of liver allografts in the rat. Transplantation 1988;46: Caldwell-Kenkel JC, Thurman RG, Lemasters JJ. Selective loss of nonparenchymal cell viability after cold ischemic storage of rat livers. Transplantation 1988;45: Pentilla A, Trump BF. Extracellular acidosis protects Ehrlich tumor cells and rat renal cortex against anoxic injury. Science 1974;185: Bing OH. Brooks WW. Messer JV. Heart muscle viability following hypoxia: protective effect of acidosis. Science 1973:180: Bond JM, Herman B, Lemasters JJ. Protection by acidotic ph against anoxia/reoxygenation injury to rat neonatal cardiac myocytes. Biochem Biophys Res Commun 1991;179: Gores JG, Nieminen A-L, Fleishman KE, Dawson TL, Herman B, Lemasters JJ. Extracellular acidosis delays onset of cell death in ATP-depleted hepatocytes. Am J Physiol 1988;255:C315-C Currin RT, Gores GJ, Thurman RG, Lemasters JJ. Protection by acidotic ph against anoxic cell killing in perfused rat liver: evidence for a ph paradox. FASEB J 1991;5: Zagar RA, Schimpf BA, Gmur DJ. Physiological ph: effect on posthypoxic proximal tubular injury. Circ Res 1993;72: Anundi I, King J, Owen DA, Schneider H, Lemasters JJ, Thurman RG. Fructose prevents hypoxic cell death in liver. Am J Physiol 1987;253: G390-G Rauen U, Hintz K, Hanssen M, Lauchart W, Becker HD, de Groot H. Injury to cultured liver endothelial cells during cold preservation: energy-dependent versus energy-deficiency injury. Transpl Int 1993;6: Callahan DJ, Engle MJ, Volpe JJ. Hypoxic injury to developing glial cells: protective effect of high glucose. Pediatr Res 1990;27: Bond JM, Chacon E, Herman B, Lemasters JJ. Intracellular ph and Ca 2 homeostasis in the ph paradox of reperfusion injury to neonatal rat cardiac myocytes. Am J Physiol 1993;265:C129-C Lemasters JJ, Bond JM, Chacon E, Harper IS, Kaplan SH, Ohata H, Trollinger DR, et al. The ph paradox in ischemia-reperfusion injury to cardiac myocytes. In: Karmazyn M, ed. Myocardial Ischemia: Mechanisms, Reperfusion, Protection. Basel, Switzerland: Birkhäuser Verlag, 1996: Weinberg JM, Davis JA, Abarzua M, Rajan T. Cytoprotective effects of glycine and glutathione against hypoxic injury to renal tubules. J Clin Invest 1987;80: Dickson RC, Bronk SF, Gores GJ. Glycine cytoprotection during lethal hepatocellular injury from adenosine triphosphate depletion. Gastroenterology 1992;102: Qian T, Nieminen A-L, Herman B, Lemasters JJ. Role of ph i and Na in reperfusion injury to rat hepatocytes: protection by cyclosporin A and glycine. Am J Physiol 1997;273:C1783-C Wetzels JF, Wang X, Gengaro PE, Nemenoff RA, Burke TJ, Schrier RW. Glycine protection against hypoxic but not phospholipase A 2 induced injury in rat proximal tubules. Am J Physiol 1993;264:F94-F99.

11 HEPATOLOGY Vol. 27, No. 4, 1998 NISHIMURA, ROMER, AND LEMASTERS Carini R, Bellomo G, Grazia de Cesaris M, Albano E. Glycine protects against hepatocyte killing by KCN or hypoxia by preventing intracellular Na overload in the rat. HEPATOLOGY 1997;26: Bachmann S, Peng X-X, Currin RT, Thurman RG, Lemasters JJ. Glycine in Carolina rinse solution reduces reperfusion injury, improves graft function, and increases graft survival after rat liver transplantation. Transplant Proc 1995;27: Currin RT, Caldwell-Kenkel JC, Litchman SN, Bachmann S, Takei Y, Kawano S, Thurman RG, et al. Protection by Carolina rinse solution, acidotic ph, and glycine against lethal reperfusion injury to sinusoidal endothelial cells of rat livers stored for transplantation. Transplantation 1996;11: Andries LJ, Brutsaert DL. Endocardial endothelium in the rat: cell shape and organization of the cytoskeleton. Cell Tissue Res 1993;273: Romer LH, McLean N, Turner CE, Burridge K. Tyrosine kinase activity, cytoskeletal organization, and motility in human vascular endothelial cells. Mol Biol Cell 1994;5: Thurston G, Baldwin AL. Endothelial actin cytoskeleton in rat mesentery microvasculature. Am J Physiol 1994;266:H1896-H Momii S, Koga A, Eguchi M, Fukuyama T. Ultrastructural changes in rat liver sinusoids during storage in Euro-Collins solution. Virchows Archiv B 1989;57: Bershadsky AD, Gelfand VI, Svitkina TM, Tint IS. Destruction of microfilament bundles in mouse embryo fibroblasts treated with inhibitors of energy metabolism. Exp Cell Res 1980;127: Hinshaw DB, Armstrong BC, Burger JM, Beals TF, Hyslop PA. ATP and microfilaments in cellular oxidant injury. Am J Pathol 1988;132: Braet F, De Zanger R, Sasaoki T, Baekeland M, Janssens P, Smedsrød B, Wisse E. Methods in laboratory investigation. Assessment of a method of isolation, purification, and cultivation of rat liver sinusoidal endothelial cells. Lab Invest 1994;70: Knook DL, Sleyster ECH. Separation of Kupffer and endothelial cells of the rat liver by centrifugal elutriation. Exp Cell Res 1976;99: De Zanger R, Braet F, Arnez Camacho MR, Wisse E. Prolongation of hepatic endothelial cell culture by phorbol myristate acetate. In: Wisse E, Knook DL, Balabaud C, eds. Cells of the Hepatic Sinusoid. Volume 6. Leiden: The Kupffer Cell Foundation, 1997: Blomhoff R, Smedsrød B, Eskild W, Granum PE, Berg T. Preparation of liver endothelial cells and Kupffer cells in high yield by means of an enterotoxin. Exp Cell Res 1984;150: Lemasters JJ, DiGuiseppi J, Nieminen A-L, Herman B. Blebbing, free Ca 2 and mitochondrial membrane potential preceding cell death in hepatocytes. Nature 1987;325: Herman B, Nieminen A-L, Gores GJ, Lemasters JJ. Irreversible injury in anoxic hepatocytes precipitated by an abrupt increase in plasma membrane permeability. FASEB J 1988;2: Bond JM, Herman B, Lemasters JJ. Recovery of cultured rat neonatal myocytes from hypercontracture after chemical hypoxia. Res Commun Chem Pathol Pharmacol 1991;71: Nieminen A-L, Gores GJ, Bond JM, Imberti R, Herman B, Lemasters JJ. A novel cytotoxicity screening assay using a multiwell fluorescence scanner. Toxicol Appl Pharmacol 1992;115: Nieminen A-L, Saylor AK, Tesfai SA, Herman B, Lemasters JJ. Contribution of the mitochondrial permeability transition to lethal injury after exposure of hepatocytes to t-butylhydroperoxide. Biochem J 1995;307: Nieminen AL, Gores GJ, Dawson TL, Herman B, Lemasters JJ. Toxic injury from mercuric chloride in rat hepatocyte. J Biol Chem 1990;265: McElroy WD, Deluca MA. Firefly and bacterial luminescence: basic science and applications. J Appl Biochem 1983;5: Lundin A, Thore A. Analytical information obtainable by evaluation of the time course of firefly bioluminescence in the assay of ATP. Anal Biochem 1975;66: Ehrenberg B, Montana V, Wei M-D, Wuskell JP, Loew LM. Membrane potential can be determined in individual cells from the nernstian distribution of cationic dyes. Biophys J 1988;53: Dauchen MR, McGuinness O, Brown LA, Crompton M. On the involvement of a cyclosporin A sensitive mitochondrial pore in myocardial reperfusion injury. Cardiovasc Res 1993;27: Pastorino JG, Snyder JW, Serroni A, Hoek JB, Farber JL. Cyclosporin A and carnitine prevent the anoxic death of cultured hepatocytes by inhibiting the mitochondrial permeability transition. J Biol Chem 1993;268: Shimizu S, Kamiike W, Hatanaka N, Miyata M, Inoue T, Yoshida Y, Tagawa K, et al. Beneficial effects of cyclosporine on reoxygenation injury in hypoxic rat liver. Transplantation 1994;57: Nieminen A-L, Dawson TL, Gores GJ, Kawanishi T, Herman B, Lemasters JJ. Protection by acidotic ph and fructose against lethal injury to rat hepatocyte from mitochondrial inhibitors, ionophores and oxidant chemicals. Biochem Biophys Res Commun 1990;167: Nieminen A-L, Saylor AK, Herman B. Lemasters JJ. ATP depletion rather than mitochondrial depolarization mediates hepatocyte killing after metabolic inhibition. Am J Physiol 1994;267:C67-C Szabo I, Zoratti M. The mitochondrial megachannel is the permeability transition pore. J Bioenerg Biomembr 1992;24: Petronilli V, Carmen C, Bernardi P. Modulation of the mitochondrial cyclosporin A sensitive permeability transition pore. J Biol Chem 1993;268: Gunter TE, Pfeiffer DR Mechanisms by which mitochondrial transport calcium. Am J Physiol 1990;258:C755-C Fujii Y, Johnson ME, Gores GJ. Mitochondrial dysfunction during anoxia/reoxygenation injury of liver sinusoidal endothelial cells. HEPA- TOLOGY 1994;20: Broekemeier KM, Pfeiffer DR. Inhibition of the mitochondrial permeability transition by cyclosporin A during long time frame experiments: relationship between pore opening and the activity of mitochondrial phospholipases. Biochemistry 1995;34: Miller GW. Lock EA. Schnellmann RG. Strychnine and glycine protect renal proximal tubules from various nephrotoxicants and act in the late phase of necrotic cell injury. Toxicol Appl Pharmacol 1994;125: Venkatachalam MA. Weinberg JM. Patel Y. Saikumar P. Dong Z. Cytoprotection of kidney epithelial cells by compounds that target amino acid gated chloride channels. Kidney Int 1996;49: Nichols JC, Bronk SF, Mellgren RL, Gores GJ. Inhibition of nonlysosomal calcium-dependent proteolysis by glycine during anoxic injury of rat hepatocytes. Gastroenterology 1994;106: Penttila A. Trump BF. Studies on the modification of the cellular response to injury. III. Electron microscopic studies on the protective effect of acidosis on p-chloromercuribenzene sulfonic acid-(pcmbs) induced injury of Ehrlich ascites tumor cells. Virchows Archiv B Cell Pathology 1975;18: Bronk SF, Gores GJ. ph-dependent nonlysosomal proteolysis contributes to lethal anoxic injury of rat hepatocytes. Am J Physiol 1993;264:G744- G Harrison DC, Lemasters JJ, Herman B. A ph-dependent phospolilase A 2 contributes to loss of plasma membrane integrity during chemical hypoxia in rat hepatocytes. Biochem Biophys Res Commun 1991;174: Gao WS, Takei Y, Marzi I, Lindert KA, Caldwell-Kenkel JC, Currin RT, Tanaka Y, et al. Carolina rinse solution a new strategy to increase survival time after orthotopic liver transplantation in the rat. Transplantation 1991;52: