Suppression of transformed phenotypes of human fibrosarcoma cells by overexpression of recombinant fibronectin.

Loss of fibronectin (FN) from the cell surface has been shown to be closely associated with malignant transformation of cells. To elucidate the role of the FN matrix in the modulation of malignant phenotypes, we overexpressed a full-length cDNA encoding plasma-type FN in HT1080 human fibrosarcoma cells. The cells overexpressing FN adopted a more flattened morphology and deposited a moderately developed FN matrix both in vitro and in vivo, although the level of expression of integrin alpha5beta1 remained unchanged. FN-overexpressing cells exhibited a reduced cell motility on the substratum and grew poorly when injected s.c. into nude mice. Overexpression of FN also suppressed the ability of the tumor cells to proliferate in soft agar, whereas the suppression was reversed by inclusion in soft agar of the Arg-Gly-Asp (RGD)-containing peptide and adhesion-blocking antibodies against the central cell-binding domain of FN. Neither cell motility nor growth potential was altered by overexpression of a truncated form of FN lacking the central cell-binding domain. These results, taken together, indicate that increased deposition of FN in the pericellular matrix per se can suppress the motility and growth potential of tumor cells through interaction with RGD-recognizing integrins, most likely alpha5beta1.

Involvement of CPP32/Yama(-like) proteases in Fas-mediated apoptosis.

Fas (Apo-1/CD95) belongs to the tumor necrosis factor/nerve growth factor receptor family and transmits apoptotic signals by binding to its ligand. Interleukin-1beta-converting enzyme (ICE), which shows substantial homology to the product of the cell death gene, ced-3, of Caenorhabditis elegans, is reported to be involved in Fas-mediated apoptosis. Using two human carcinoma-derived cell lines with undetectable levels of ICE, we found that an agonistic antihuman Fas antibody induces the activation of CPP32/Yama(-like) proteases that are ICE(-like) protease family members, and that a tetrapeptide inhibitor of CPP32/Yama protease, DEVD-CHO, inhibits the Fas-mediated activation of the proteases, Fas-mediated apoptosis, and CPP32/Yama(-like) proteolytic activities in vitro. Fas-mediated apoptosis is inhibited by the CPP32/Yama inhibitor DEVD-CHO, but not by the ICE inhibitor YVAD-CHO, suggesting a dominant role for the CPP32/Yama(-like) proteases and not ICE itself in Fas-mediated apoptosis of the human carcinoma cell lines.

Induction of apoptosis as well as necrosis by hypoxia and predominant prevention of apoptosis by Bcl-2 and Bcl-XL.

The molecular mechanism of cell death due to hypoxia has not been elucidated. Our recent observations that overexpression of the anti-apoptotic proto-oncogene bcl-2 and a bcl-2-related gene, bcl-x, prevents hypoxic cell death suggest that hypoxia induces apoptosis. Using electron microscopy and confocal and nonconfocal fluorescence microscopy, we show here that hypoxia does, in fact, induce both necrosis and apoptosis, and that the proportion of these two modes is highly dependent on the cell type. Overexpression of Bcl-2 or Bcl-Xl blocks hypoxia-induced apoptosis in a dose-dependent manner.

Bcl-2 expression prevents activation of the ICE protease cascade.

The Bcl-2 family and the ICE family of cysteine proteases play important roles in regulating cell death. We show here that induction of cell death by a Ca2+ ionophore or hypoxia results in increased levels and activity of active ICE(-like) proteases and the subsequent activation of CPP32/Yama(-like) proteases, and that inhibition of these protease activities reduces the extent of cell death. Overexpression of the anti-apoptotic proteins Bcl-2 or Bcl-xL inhibits the cell death and the activation of ICE(-like) and CPP32/Yama(-like) proteases, indicating that Bcl-2 and Bcl-xL act upstream of these proteases. We also show that specific inhibition of ICE(-like) proteases in vivo prevents activation of CPP32/Yama(-like) proteases, whereas inhibition of CPP32/Yama(-like) proteases does not prevent activation of ICE(-like) proteases, suggesting the existence of a protease cascade in vivo that requires ICE(-like) proteases for activation of CPP32/Yama(-like) proteases. Induction of necrotic cell death by KCN also induces activation of ICE(-like) proteases but not of CPP32/Yama(-like) proteases, and Bcl-2 and Bcl-xL inhibit the activation and the cell death, suggesting that the functional site of Bcl-2 and Bcl-xL is also upstream of ICE(-like) proteases in at least some forms of necrosis.

Retardation of chemical hypoxia-induced necrotic cell death by Bcl-2 and ICE inhibitors: possible involvement of common mediators in apoptotic and necrotic signal transductions.

Inhibition of the respiratory chain reaction by cyanide, rotenone or antimycin A (chemical hypoxia) induces necrotic cell death characterized by apparently intact chromatin, remarkable mitochondrial swelling with loss of crista structure, and loss of plasma membrane integrity. The treatments induce no apoptotic cell death, as defined by fragmented nuclei with condensed chromatin, fragmented or condensed cytoplasm. The anti-apoptotic proteins Bcl-2 and Bcl-xL effectively retard the chemical hypoxia-induced necrotic cell death. The necrotic cell death is also retarded by inhibitors of ICE(-like) proteases, including interleukin-1beta converting enzyme (ICE), which are common mediators of apoptosis. These results indicate that Bcl-2/Bcl-xL and ICE(-like) proteases modulate apoptotic and at least some forms of necrotic cell death. Both cell death pathways appear to involve some common mediators; however necrotic or apoptotic cell death signals might be transduced through multiple pathways, because Bcl-2/ Bcl-xL or inhibitors of ICE(-like) proteases are relatively less potent in blocking necrotic cell death than in preventing apoptosis.

Bcl-2 blocks loss of mitochondrial membrane potential while ICE inhibitors act at a different step during inhibition of death induced by respiratory chain inhibitors.

Bcl-2, Bcl-xL, CrmA and tetrapeptide ICE inhibitor reduce the extent of necrotic cell death induced by cyanide, which primarily damages mitochondria. Although none of them affects the drastic decrease in ATP levels induced by cyanide, Bcl-2 and Bcl-xL but not CrmA or ICE inhibitor inhibit the cyanide-induced decrease in mitochondrial membrane potential. A similar blocking effect is observed on necrotic cell death induced by other respiration inhibitors, rotenone and antimycin A, and on apoptotic cell death induced by etoposide or calcium ionophore. These results indicate that Bc1-2 and Bcl-xL protect mitochondria against the loss of function during both apoptosis and at least some forms of necrotic cell death. The ICE family proteases act at a different step other than the loss of mitochondrial membrane potential.

Bcl-2 and Bcl-xL block apoptosis as well as necrosis: possible involvement of common mediators in apoptotic and necrotic signal transduction pathways.

The proto-oncogene bcl-2 and a bcl-2-related gene bcl-x prevent apoptotic cell death induced by various treatments. Although a mechanism has been proposed that involves Bcl-2 activity on reactive oxygen species (ROS), we find that expression of Bcl-2 or Bcl-xL prevents cell death induced by withdrawal of oxygen (hypoxia) and that the cell death does not involve ROS, suggesting that Bcl-2 or Bcl-xL exerts an anti-cell death function by a mechanism other than through regulation of ROS activity. Using electron microscopy, and confocal and non-confocal fluorescence microscopy, we show that hypoxia induces both necrosis and apoptosis. Overexpression of Bcl-2 or Bcl-xL blocks hypoxia-induced apoptosis and, although to a lesser extent, necrosis. The anti-apoptotic proteins Bcl-2 and Bcl-xL effectively inhibit KCN-induced cell death which is characterized by necrotic features including apparently intact chromatin, remarkable mitochondrial swelling with loss of crista structure and loss of plasma membrane integrity. The necrotic cell death is also inhibited by inhibitors of ICE (interleukin-1 beta converting enzyme)(-like) proteases, the common mediators of apoptosis. These results indicate that Bcl-2/Bcl-xL and ICE(-like) proteases modulate both apoptotic and at least some forms of necrotic cell death, suggesting that both cell death pathways involve some common mediators.

Correlation between cellular ATP level and bile excretion in the rat liver.

The influence of the cellular level of adenosine triphosphate (ATP) in the liver on bile excretion was studied in rats. In ischemia, the cellular ATP level decreased rapidly--and, concomitantly, bile flow stopped within 5 min. Administration of L-ethionine i.p. to rats reduced the bile flow rate with decrease in the cellular ATP level. The correlation between the bile flow rate and the cellular ATP level was confirmed in a liver perfusion system. On anoxic perfusion, the ATP level and bile flow rate changed in the same manner as in ischemia. The recovery rates of both on reoxygenation decreased with increase in the anoxic perfusion period. During perfusion under oxygenated conditions, decrease in cellular ATP to various levels by infusion of various concentrations of potassium cyanide, an inhibitor of respiration, resulted in corresponding and concomitant suppression of bile excretion. Kinetic analysis of the bile flow rate revealed a Michaelis-Menten-type curve for the cellular ATP level. The apparent Kms for ATP of bile flow rate in L-ethionine-treated rat liver and liver perfused with potassium cyanide were 1.0 and 1.6 mM, and their Vmax values were 4.1 and 2.5 microliter/min/g liver, respectively. The concentrations of main bile components, such as phospholipids, cholesterol, and taurocholate increased, but their total outputs decreased with decrease in the ATP level, and returned to the normal range with recovery of the ATP level. Thus, it was shown experimentally that the extent of hepatic injury can be assessed simply by monitoring the bile flow rate, which reflects the cellular level of ATP.

Levels of purine compounds in a perfusate as a biochemical marker of ischemic injury of cold-preserved liver.

Biochemical markers of ischemic injury of rat liver were studied in an extracorporeal perfusion system. During anoxic perfusion, purine compounds appeared in the perfusate as soon as they were formed in the liver and their recovery in the perfusate balanced the loss of adenine nucleotides from the liver. In contrast, cytosolic aspartate aminotransferase did not appear in the perfusate at slow rates of liver perfusion or during hypothermic perfusion. The production of purine compounds was further investigated in hypothermically preserved liver in connection with the restoration of some metabolic functions of liver. The amount of purine compounds released into the perfusate was found to be closely related to the degrees of damage of the hepatic functions of gluconeogenesis, ureogenesis, and mitochondrial respiration on reperfusion. These results indicate that release of purine compounds into the perfusate is a good marker of ischemic damage.

Beneficial effects of cyclosporine on reoxygenation injury in hypoxic rat liver.

The effect of CsA on hypoxia-reoxygenation injury was studied in perfused rat livers. CsA did not attenuate hypoxic injury, as assessed by the release of lactate dehydrogenase and mitochondrial aspartate aminotransferase. During reoxygenation, the release of lactate dehydrogenase was also not affected by CsA. However, the release of mitochondrial aspartate aminotransferase into the cytosol, which indicates mitochondrial injury, was significantly reduced by CsA. The effect of CsA on mitochondrial function during hypoxia-reoxygenation was also investigated. CsA administration increased both the respiratory control ratio and the adenine nucleotide content after reoxygenation in both isolated mitochondria and perfused livers. In addition, glucose production by perfused livers after reoxygenation was increased by CsA. We conclude that the beneficial effect of CsA on hypoxia-reoxygenation injury may be partly due to protection of the mitochondria against reoxygenation injury.

Adenine nucleotide metabolism and its relation to organ viability in human liver transplantation.

The relationship between adenine nucleotide metabolism and ischemic damage was studied in human liver. Thirty transplanted grafts were divided into two groups according to their functional outcome. Cellular adenine nucleotide levels were assayed by high-performance liquid chromatography. During cold ischemia, the adenosine triphosphate (ATP) level was not correlated with graft function, but two grafts with low total adenine nucleotides (TAN) levels showed poor function after transplantation. After recirculation, the ATP level showed good recovery in grafts that functioned satisfactorily (n = 24, 5.47 +/- 1.51 mumol/g dry weight), but remained low in poorly functioning grafts (n = 6, 3.30 +/- 1.68 mumol/g dry weight) (P less than 0.01). The level of recovery of ATP was inversely related to the period of warm ischemia during implantation (P less than 0.01). Bile production, used as a parameter of initial function, was observed shortly after implantation in 17 of 24 grafts that functioned satisfactorily, but in only 1 of 6 poorly functioning grafts. It is concluded that loss of adenine nucleotides and lack of bile production during transplantation are good markers of damaged grafts in human liver transplantation.

Restoration of immune abnormalities in diabetic BB rats after pancreas transplantation. I. Macrochimerism of donor-graft-derived RT6+ T cells responsible for restoration of immune responsiveness and suppression of autoimmune reaction.

Diabetes-prone (DP) BB rats (RT1(u), RT6.1) spontaneously develop insulin-dependent diabetes mellitus (IDDM) and the disease manifestation resembles that in human IDDM. DP rats are immunodeficient with severe T lymphocytopenia due to the absence of T cells expressing the RT6 differential alloantigen, which have immunoregulatory functions. MHC- and non-MHC-compatible Wistar Furth (WF; RT1(u), RT6.2) pancreases were transplanted into DP rats. WF pancreas grafts were destroyed by IDDM recurrence (insulitis), but not by rejection, with a mean survival time of 65.3 +/- 21.7 days. To prevent the recurrence of IDDM in the grafts, monoclonal antibodies to intercellular adhesion molecule-1 and leukocyte function-associated antigen-1 were administered. WF pancreas grafts were indefinitely accepted (>108.0 +/- 26.8 days) in monoclonal antibody-treated DP recipients. The number of T cells was increased and cellular immune responses restored only in the DP rats that had accepted grafts. The increased number of T cells was due to the peripheral appearance of donor-type RT6.2+ T cells, which represented 34.3 +/- 7.0% of total splenic T cells. The cytotoxicity of splenic T cells to WF islet cells was suppressed in the presence of RT6+ T cells in vitro. These findings demonstrated that stable macrochimerism of donor-derived RT6+ T cells could restore the immune responses and prevent the recurrence of IDDM in the DP recipients.

Attenuation of warm ischemic injury of rat lung by inflation with room air--assessment of cellular components and the surfactant in the bronchoalveolar lavage fluid in relation to changes in cellular adenosine triphosphate.

Studies were made on the effects in rat lungs of aerobic and anaerobic conditions on the intracellular levels of adenosine triphosphate and its related metabolites, the releases of intracellular enzymes, and the secretion of pulmonary surfactant. After warm ischemia for 120 min, the ATP content of lungs inflated with air was significantly higher (8.0 +/- 1.2 mumol/g dry weight) than those of deflated lungs and lungs inflated with nitrogen (0.8 +/- 0.7 mumol/g dry weight and 2.0 +/- 0.7 mumol/g dry weight, respectively; P < 0.001). The amounts of intracellular enzymes, such as lactate dehydrogenase, cytosolic and mitochondrial aspartate aminotransferase, and protein in the bronchoalveolar lavage fluid (BALF) of air-inflated lungs were significantly less than those in BALFs of deflated and nitrogen-inflated lungs (P < 0.001). The BALF-contents of dipalmitoyl phosphatidylcholine (DPPC), the main component of alveolar surfactant of aerobic and anaerobic ischemic lung were, however, similar. During 120-min warm ischemia after lavage, air-inflated lungs secreted significantly more DPPC into the alveolar space than nitrogen-inflated lungs did (P < 0.001). We conclude that cell membranes in the lungs are damaged under anaerobic conditions, but that inflation of ischemic lungs with air is effective for protecting them from cell injury and for maintaining the intracellular level of ATP and the ability of the cells to secrete pulmonary surfactant.

Generation of nitric oxide as a rejection marker in rat pancreas transplantation.

In clinical pancreas transplantation, no reliable marker for the early diagnosis of acute rejection has been reported. This is one reason why the graft survival rate of pancreas transplantation alone is much lower than that of other organs, such as hearts, livers, and kidneys. We designed an experiment to investigate acute rejection of pancreas allografts in hyperglycemic rats by measurement of blood glucose levels and nitric oxide (NO) products (nitrite plus nitrate, and nitrosyl hemoglobin). As recipients, Lewis rats were rendered hyperglycemic by intravenous injection of streptozotocin before transplantation. F344 rats were used as donors of pancreas allografts. Lewis rats were also used as donors of syngeneic pancreas grafts. After transplantation, the blood glucose level returned to a normal level and rejection was defined as the recurrence of hyperglycemia. The mean survival time of pancreas allografts was 14 +/- 0.7 days. The plasma level of nitrite plus nitrate in allografted rats peaked on postoperative day 7. Electron spin resonance spectra of NO bound to hemoglobin were detected in the blood from allografted rats with a peak on postoperative day 7, whereas NO bound to hemoglobin was not detected in the blood from recipients of syngeneic grafts at any sampling time. The results show that NO was synthesized in the earlier period than the elevation of the blood glucose level during rejection after pancreas transplantation in rats.

Enzyme release from mitochondria during reoxygenation of rat liver.

Reoxygenation-induced release of mitochondrial aspartate aminotransferase (mAST) into the cytosol was studied using perfused rat liver. As the absolute activity of mAST in the perfusate did not indicate the degree of mitochondrial enzyme release, the following 3 methods were applied: measurement of the mAST to total AST ratio in the efferent perfusate, the digitonin infusion method, and measurement of mAST activity in the cytosolic compartment isolated from perfused livers. The results by all 3 methods were consistent and showed that mitochondrial injury occurs on reoxygenation. The mitochondrial Ca2+ content was proportional to the extent of mAST release during reoxygenation, indicating involvement of Ca2+ in the enzyme release. CsA, a potent inhibitor of Ca(2+)-induced increase in permeability of the mitochondrial membrane, completely prevented mAST release on reoxygenation. We conclude that during reoxygenation of hypoxic liver, mAST leaks into the cytosol in a Ca(2+)-dependent, CsA-sensitive manner.

Donor-specific tolerance by perioperative intrathymic injection of bone marrow cells in the rat cardiac allograft model: use of FK506 can shorten the necessary duration of pretransplant intrathymic conditioning.

BACKGROUND: Many strategies of tolerance induction by intrathymic (IT) injection of donor alloantigens have been reported to date; however, the timing of IT injection is usually 1-3 weeks before transplantation. METHODS: To apply IT injection to cadaveric organ transplantation, 1 x 10(8) fully allogeneic bone marrow cells (BMC) of Buffalo (BUF; RT1b) rats were intrathymically injected into Wistar Furth (WF; RT1u) rats at the time of BUF cardiac allografting with short-course therapy of antilymphocyte serum (ALS) and FK506 in our experimental model. RESULTS: Allogeneic IT injection of BUF BMC with ALS and FK506 indefinitely prolonged graft survival (mean survival time > 210 days) in all WF rats. On day 130 after grafting, tolerant WF rats accepted donor BUF skin grafts (> 120 days) but not third-party Lewis skin grafts. In control groups, syngeneic IT injection of WF BMC or intravenous injection of donor BUF BMC in combination with ALS/FK506 therapy failed to induce tolerance. In vivo testing was performed during induction (1 month) or during maintenance (6 months of tolerance. In the mixed lymphocyte reaction (MLR), spleen T cells of tolerant rats at 1 month after grafting displayed hyporesponsiveness after stimulation with donor cells. The addition of interleukin (IL)-2 to MLR culture did not restore T-cell responsiveness. Tolerant rats had a significantly decreased frequency of T cytotoxic cell precursors (fTcp) of 1:4,926, and frequency of IL-2-producing T helper cell precursors (fThp) of 1:23,925, compared with naive rats (1: 2,158 and 1:4,266, respectively). By 6 months after grafting, however, the anti-donor MLR proliferative responses of tolerant rats had been restored to the levels of naive splenic T cells. These tolerant rats displayed restoration of the (fTcp) of 1:2,842 and of the (fThp) of 1:5,630, which were comparable frequencies of naive rats. Suppressor T cells did not contribute in this model. In cardiac grafts of tolerant rats induced by IT injection, expression of both Th1 (interferon-gamma and IL-2) and Th2 (IL-4 and IL-10) cytokines was detected in the early phase; thereafter, expression was completely inhibited, except for interferon-gamma in the chronic phase. CONCLUSIONS: Perfect donor-specific tolerance was obtained by IT injection of donor BMC at the time of transplantation, while alloimmune responses were maintained at levels similar to those of naive rats.

The rejection mechanism of rat pancreaticoduodenal allografts with a class I MHC disparity.

The present study has demonstrated for the first time that PVG.R1 (RT1.AaBcDcCc) pancreatic grafts are rejected by so-called "low"-responder PVG (RT1.AcBcDcCc) recipients with an isolated class I MHC disparity (mean survival time; MST: 21.4 +/- 1.8 days, n = 5), whereas PVG.R1 heart grafts are able to survive indefinitely (MST: > 100 days, n = 5). Splenic CD4+ T cells but not CD8+ T cells from the PVG recipients of PVG.R1 pancreatic grafts show a remarkable proliferative response against donor class I RT1.Aa alloantigens, while only a minimal proliferation is observed in the PVG recipients of PVG.R1 heart grafts or naive PVG rats. Naive PVG rats display an extremely low frequency of IL-2-producing helper T cell precursors (fThp) of 1/40,609 +/- 15,441 against class I RT1.Aa alloantigen. The PVG recipients of PVG.R1 heart grafts have a slightly greater fThp of 1/17,326 +/- 6822. On the other hand, the PVG recipients that rejected PVG.R1 pancreatic grafts show a significantly increased fThp of 1/5030 +/- 3396 compared with those of PVG.R1 heart grafts (P < 0.05) or naive PVG rats (P < 0.01). The frequency of cytotoxic T cell precursors (fTcp) increases slightly in the PVG recipients of PVG.R1 pancreatic grafts (1/1848 +/- 330) compared with those of PVG.R1 heart grafts (1/2215 +/- 2131) or naive PVG rats (1/2476 +/- 585). The size of cytotoxic T cell clones alone does not adequately account for a proliferation sufficient to complete the rejection of pancreatic grafts. The PVG recipients of PVG.R1 pancreatic grafts, but not heart grafts, demonstrate a strong cytotoxic alloantibody response to donor class I RT1.Aa alloantigens. In the study of alloantibodies, IgM is detected mainly in the early phase and IgG in the late phase during the course of pancreatic rejection. It is determined that in blocking studies by FACS analysis these antibodies target class I MHC antigens. These results suggest that cytotoxic T cells do not appear to be responsible for the rejection of PVG.R1 pancreatic grafts in PVG recipients. Rather, the rejection is mediated by CD4+ T cells and complement-fixing antibodies directed at class I MHC antigens.

Regulation of complement-mediated swine endothelial cell lysis by a surface-bound form of human C4b binding protein.

BACKGROUND: Human C4b-binding protein (C4bp) functions as a cofactor for factor I in the degradation of C4b and C3b and, in addition, accelerates the rate of decay of the C4b2a complex. METHODS: In this study, we constructed a surface-bound form of human C4b-binding protein (C4bp-PI) consisting of a short consensus repeat 1-8 of the alpha-chain of C4bp and a glycosyl phosphatidylinositol (GPI) of the decay-accelerating factor (CD55) and established stable swine endothelial cell (SEC) lines expressing C4bp-PI by transfection of cDNA. Amelioration of complement-mediated lysis by the transfectant molecules was tested as an in vitro hyperacute rejection model of swine to human discordant xenograft, using the lactate dehydrogenase assay. RESULTS: Flow cytometric profiles of the stable SEC lines with C4bp-PI showed a high level of expression of this molecule. The cell lysate of the SEC line with C4bp-PI showed strong cofactor activity in not only C4b but also C3b, whereas the activity of plasma C4bp to bind to C3 was very weak. Approximately 150 x 10(4) molecules of C4bp-PI per SEC blocked human complement-mediated cell lysis by approximately 75%. CONCLUSIONS: The results suggest that the surface-bound form of C4bp will be very useful in clinical xenotransplantation.

Involvement of Ca2+ release and activation of phospholipase A2 in mitochondrial dysfunction during anoxia.

During anoxic incubation, depletion of mitochondrial ATP was followed by release of Ca2+ with concomitant increase in the rate of state 4 respiration due to disruption of the diffusion barrier against protons. The external addition of ATP and its non-metabolizable analog, beta,gamma-methylene adenosine 5'-triphosphate, prevented both the release of Ca2+ and increase in the rate of state 4 respiration. Addition of EGTA, which did not prevent release of the ion, resulted in little increase in the respiration rate. Addition of an inhibitor of mitochondrial phospholipase A2, such as quinacrine, dibucaine, or chlorpromazine, also prevented increase in the respiration rate without affecting Ca2+ release from mitochondria during anoxic incubation. Non-esterified polyunsaturated fatty acids were also found to be liberated from anoxic mitochondria. External addition of the ATP-analog, EGTA, and inhibitors of phospholipase A2 suppressed the liberation of non-esterified polyunsaturated fatty acids. Melittin and Ca2+, which activate phospholipase A2, increased the rate of state 4 respiration and the liberation of fatty acids. These findings support the hypothesis proposed previously that the following sequence changes occurs in mitochondria during anoxia; depletion of ATP, liberation of free calcium from mitochondria, and disruption of the diffusion barrier against H+ of the inner membrane. The results also indicate another event; activation of phospholipase A2 by release Ca2+ which results in H+ leakiness of the inner membrane.