Effect of oxygen and endotoxin on lactate dehydrogenase release, 5-hydroxytryptamine uptake, and antioxidant enzyme activities in endothelial cells

We compared the effects of 95% O2 (hyperoxia) alone, endotoxin (20 ng/ml) alone, and 95% O2 plus endotoxin on the release of lactate dehydrogenase (LDH), uptake of 5-hydroxytryptamine (5-HT), and antioxidant enzyme activities in porcine pulmonary arterial and aortic endothelial cells in monolayer culture. Hyperoxia increased LDH release and decreased 5-HT in both endothelial cell types. Hyperoxia also caused a decrease in catalase (CAT) activity and an increase in total superoxide dismutase (SOD) and glutathione reductase (GSH-Red) activities in both cell types. Endotoxin alone had no effect on LDH release, 5-HT uptake, or antioxidant enzyme activities. However, endotoxin prevented the hyperoxic increase in LDH release and the hyperoxic decrease in 5-HT uptake. Endotoxin plus 95% O2 had no consistent effect on the antioxidant enzyme profile in pulmonary artery or aortic endothelial cells. These results indicate that (1) hyperoxia injures both pulmonary artery and aortic endothelial cells in culture and causes changes in the antioxidant enzyme profile that are similar in the two cell types; (2) hyperoxia-induced decreases in CAT activity and increases in SOD activity may be responsible for increased sensitivity of endothelial cells to O2 toxicity; and (3) endotoxin protects against hyperoxic injury to endothelial cells in vitro, but increases in antioxidant enzyme activities are not the mechanism for this protection.     

Butyrate increases catalase activity and protects rat pulmonary artery smooth muscle cells against hyperoxia

A protective effect of butyrate against hyperoxia was found with adult rat pulmonary artery smooth muscle cells. Butyrate (5mM) when added just prior to the hyperoxic exposure (95%) markedly decreased lactate dehydrogenase release from cells during 68 hours of exposure (22% release with butyrate versus 98% without). The uptake and reduction of a tetrazolium compound as another index of cell viability also showed similar improvement with butyrate. Butyrate was associated with a striking increase of catalase to three times the control in the air exposed group while GSH content and the activities of superoxide dismutase and glutathione peroxidase were not significantly changed. In the groups exposed to hyperoxia alone, both enzyme activities were decreased compared to the air exposed controls. When butyrate was present with hyperoxia, the superoxide dismutase was maintained closer to the air exposed control values and the catalase activity remained nearly twice as high as the air exposed control cells. These results suggest that butyrate protects rat pulmonary artery smooth muscle cells from hyperoxia by increasing catalase activity which may help to preserve superoxide dismutase activity. This may be a good model to determine the biological significance of catalase and its interrelationships with other antioxidant systems within the cell.     

Temporal Association Between Pulmonary Inflammation and Antioxidant Induction Following Hyperoxic Exposure of the Preterm Guinea Pig

The time course and nature of the pulmonary inflammatory and antioxidant responses, both during and after hyperoxic-induced acute lung injury were studied in the preterm guinea pig. Three-day preterm (65 days gestation) guinea pigs were randomly exposed to either 21% O₂ (control) or 95% O₂ (hyperoxia) for 72 hours. All pups were then maintained in ambient conditions for up to a further 11 days, during which time lung damage was monitored. In animals exposed to hyperoxia, evidence of acute lung injury and inflammation was characterized by a marked increase in microvascular permeability and elevated numbers of neutrophils in bronchoalveolar lavage fluid. Protein concentration, elastase-like activity and elastase-inhibitory capacity in lavage fluid were at a maximum at the end of the 72 hours hyperoxic exposure. Four days later, all values had returned to control levels. In contrast, increased numbers of neutrophils, macrophages and lymphocytes were recovered in the lavage fluid during this early recovery period. Coinciding with the influx of inflammatory cells, there was a significant increase in glutathione peroxidase, manganese superoxide dismutase and catalase activities in immature lung. Lung copper/zinc superoxide dismutase activity remained unchanged during both experimental periods. The strong temporal relationship between the influx of inflammatory cells to the lung and the induction of pulmonary antioxidant enzyme defences suggests that a common mechanism underlies both responses. These findings have led us to regard inflammation in the hyperoxic-injured immature lung as a beneficial event and not, as previously suggested, as part of the injurious process.     

Effect of normobaric hyperoxia on antioxidant defenses of HeLa and CHO cells

The effect of increased intracellular oxygen activation on cellular antioxidant defenses in CHO and HeLa cells was studied. In both cell types, hyperoxic exposure (up to 4 days, 600-700 mm Hg O2) and in CHO cells menadione (up to 3 days, 15 microM) failed to affect the enzymatic antioxidant defenses Mn-containing superoxide dismutase (Mn-SOD), CuZn-SOD, catalase and glutathione peroxidase. The markedly increased antioxidant enzyme activities observed in a recently obtained oxygen-tolerant CHO variant persisted under normoxia. These data suggest that the synthesis of antioxidant enzymes is constitutive. Glutathione levels of HeLa cells did not respond to hyperoxia whereas in CHO cells hyperoxia and menadione exposure resulted in a 2- and 7-fold increase in glutathione contents, respectively. However, considering the large variations in glutathione contents observed under normal culture conditions, it is uncertain whether this increase is to be considered as a true adaptive response.     

Protection against pulmonary oxygen toxicity by interleukin-1 and tumor necrosis factor: Role of antioxidant enzymes and effect of cyclooxygenase inhibitors

Rats injected with interleukin-1 (10 g) and tumor necrosis factor (10 g) and then exposed continuously to hyperoxia (> 99% O₂, 1 atm) survived longer, had increased lung reduced/oxidized glutathione ratios, smaller pleural effusions, less pulmonary hypertension and improv+++ed arterial blood gases. The percentage of animals surviving for 72 hours in hyperoxia increased from 8% to 94%. Although relatively small increases in glutathione redox cycle enzymes occurred four and sixteen hours following cytokine injection, dramatic increases in all major antioxidant enzymes including superoxide dismutase, glucose-6-phosphate dehydrogenase, glutathione reductase, glutathione peroxidase, and catalase had occurred following 72 hours of exposure to hyperoxia. The protective effect of IL-1 + TNF against lethal pulmonary O₂ toxicity could be partially inhibited by pre-injection of lysine acetylsalicylate or, less effectively, of ibuprofen.Recent studies have suggested that both IL-1 and TNF can induce manganese (mitochondrial) superoxide dismutase mRNA and protein synthesis in a variety of cell types. Preliminary studies suggest that IL-1 alone, in ample dosage, can provide protection against lethal pulmonary O₂ toxicity. Future studies should be directed toward identification of acute phase changes in lung antioxidant enzymes, surfactant proteins and/or lipid components, enzymes needed for synthesis of surfactant phospholipids, and/or other protective proteins. Additional work also needs to be done in identifying the lung cell types in which early enzyme induction occurs. These studies should provide a better understanding of mechanisms whereby protection against pulmonary O₂ toxicity can occur. An understanding of the molecular mechanisms inducing protective proteins should lead to more precise pharmacologic control of these processes.     

Cytokines increase rat lung antioxidant enzymes during exposure to hyperoxia

Pretreatment with the combination of tumor necrosis factor/cachectin (TNF/C) and interleukin 1 (IL-1) increased glucose-6-phosphate dehydrogenase (G6PDH), glutathione reductase (GR), glutathione peroxidase (GPX), catalase (CAT), and superoxide dismutase (SOD) activities in lungs of rats continuously exposed to hyperoxia for 72 h, a time when all untreated rats had already died. Pretreatment with TNF/C and IL-1 also increased, albeit slightly, lung G6PDH and GR activities of rats exposed to hyperoxia for 4 or 16 h. By comparison, no differences occurred in lung antioxidant enzyme activities of TNF/C and IL-1- or saline-pretreated rats exposed to hyperoxia for 36 or 52 h; the latter is a time just before untreated rats began to succumb during exposure to hyperoxia. The results raise the possibility that TNF/C and IL-1 treatment can increase lung antioxidant enzyme activities and that increased lung antioxidant enzymes may contribute to the increased survival of TNF/C and IL-1-pretreated rats in hyperoxia for greater than 72 h.     

Effects of various oxygen concentrations on antioxidant enzymes and the quantity of tissue phospholipid fatty acids in the carp

Analyses were made of the phsopholipid fatty acids and the antioxidant enzymes in the carp (Cyprinus carpio morpha) at three different oxygen concentrations, corresponding to hyperoxia, hypoxia and anoxia. Variations of the oxygen concentration were found to influence the quantities of phsopholipid fatty acids, as well as the antioxidant enzyme activities. In hyperoxia and hypoxia the amount of polyunsaturated fatty acids in carp liver was higher than in anoxia, but in other tissues there was no significant differences. As to the antioxidant enzyme system, the glutathione peroxidase activity and the lipid peroxidation value increased significantly with decrease of the oxygen concentration, while the total superoxide dismutase activity decreased on lowering of the oxygen level.     

Expression of nitric oxide synthase isoforms (NOS II and NOS III) in adult rat lung in hyperoxic pulmonary hypertension

Breathing air with a high oxygen tension induces an inflammatory response and injures the microvessels of the lung. The resulting development of smooth muscle cells in these segments contributes to changes in vasoreactivity and increased pulmonary artery pressure. This in vivo study determines the temporal and spatial expression of endogenous endothelial nitric oxide synthase (NOS III) and inducible NOS (NOS II), important enzymes regulating vasoreactivity and inflammation, in the adult rat lung during the development of experimental pulmonary hypertension induced by oxidant injury. We analyzed the cellular distribution of these NOS isoforms, using specific antibodies, and assessed enzyme activity at baseline and after 1-28 days of hyperoxia (FIO₂ 0.87). The number of NOS III-immuno-positive endothelial cells increased early in hyperoxia and then remained high. By day 28, the relative number of these cells had increased from 40% in proximal vessels and 13-16% in distal alveolar vessels of the normal lung to 73-86% and 40-59%, respectively, in hyperoxia. Pulmonary alveolar macrophages (PAMs), normally few in number and only weakly immunopositive for NOS II or III in the normal lung, increased in number in hyperoxia and were strongly immunopositive for each isoform. These morphological data were supported by a temporal increase in total and calcium-independent NOS activity. Thus NOS expression and activity significantly increased in hyperoxia as pulmonary hypertension developed, and NOS III expression increased selectively in vascular endothelial cells, while both NOS isoforms were expressed by the PAM population. We conclude that this increase in expression of a potent vasodilator, an antiproliferative agent for smooth muscle cells, and an antioxidant molecule represents an adaptive response to protect the lung from oxidant-induced vascular and epithelial injury.     

Expression of antioxidant enzymes in rat kidney during ischemia-reperfusion injury

The effect of ischemia-reperfusion on activity, protein and m-RNA levels of catalase, copper-zinc and manganese containing superoxide dismutases and glutathione peroxidase, the enzymes that are involved in free radical detoxification was studied in rat kidney. Ischemia alone did not alter either the activities or protein levels of superoxide dismutase and glutathione peroxidase. However, catalase activity was found to be inhibited to 82% of control. The inhibition of catalase was due to the inactivation of the enzyme as there was no significant change in enzyme protein level. Reperfusion following ischemia, however, led to a significant decrease in both the activities as well as the protein levels of all the antioxidant enzymes. The observed overall decrease in total superoxide dismutase activity was the net effect of a decrease in copper-zinc superoxide dismutase while manganese superoxide dismutase activity was found to be increased following reperfusion. This observed increased manganese superoxide dismutase activity was the result of its increased protein level. The mRNA levels for catalase, superoxide dismutases, and glutathione peroxidase were observed to be increased (100–145% of controls) following ischemia; reperfusion of ischemic kidneys, however, resulted in a significant decrease in the levels of mRNAs coding for all the enzymes except manganese superoxide dismutase which remained high. These results suggest that in tissue, the down regulation of the antioxidant enzyme system could be responsible for the pathophysiology of ischemia-reperfusion injury.     

Augmentation of antioxidant enzymes in vascular endothelium

The endothelium is a key site of injury from reactive oxygen species that can potentially be protected by the antioxidant enzymes superoxide dismutase and catalase. Large proteins, such as superoxide dismutase and catalase, do not readily penetrate cell membranes, which limits their efficacy in protecting cells from cellular reactions involving both intracellularly and extracellularly generated reactive oxygen species. Two methods are described that promote enzyme delivery to cultured endothelial cells and confer increased resistance to oxidative stress. The first method is to entrap the antioxidant enzymes within liposomes, which then become incorporated by endothelial cells and can increase enzyme specific activities by as much as 44-fold within 2 h. The second method involves covalent conjugation of polyethylene glycol (PEG) to superoxide dismutase and catalase, a technique that increases circulatory half-life and reduces protein immunogenicity. Conjugation of PEG to superoxide dismutase and catalase increased cellular-specific activities of these enzymes in cultured endothelial cells (but at a slower rate than for liposome entrapped enzymes) and rendered these cells more resistant to oxidative stress. Both liposome-mediated delivery and PEG conjugation offer an additional benefit over native superoxide dismutase and catalase because they can increase cellular antioxidant activities in a manner that can provide protection from both intracellular and extracellular superoxide and hydrogen peroxide.     

Immunolocalization of antioxidant enzymes in adult hamster kidney

Summary Immunoperoxidase and immunogold techniques were used to localize the following antioxidant enzyme systems in the adult hamster kidney at the light and ultrastructural levels: superoxide dismutases, catalases, peroxidases and glutathione S-transferases. Each cell type in the kidney showed specific patterns of labelling of these enzymes. For example, proximal and distal tubular and transitional epithelial cells showed significant staining for all of these enzymes, while glomerular cells and cells of the thin loop of Henle did not show significant staining at the light microscope level. In addition, high levels of glutathione peroxidase were found in smooth muscle cells of renal arteries. At the ultrastructural level, each enzyme was found in a specific subcellular location. Manganese superoxide dismutase was found in mitochondria, catalase was localized in peroxisomes, while copper, zinc superoxide dismutase and glutathione S-transferase (liver and placental forms) were found in both the nucleus and cytoplasm. Glutathione peroxidase was found to have a broad intracellular distribution, with localization in mitochondria, peroxisomes, nucleus, and cytoplasm. Microvilli of tubular cells were labelled by antibodies to catalase, copper, zinc superoxide dismutase, glutathione peroxidase, and glutathione S-transferases. Cell types that were negative by light microscopy immunoperoxidase studies showed definite labelling with immunogold post-embedding ultrastructural techniques (glomerular cells and cells of the loop of Henle), demonstrating the greater sensitivity of the latter technique. These observations demonstrate that there are large variations in the levels of antioxidant enzymes in different cell types, and that even within a distinct cell type, the levels of these enzymes vary in different subcellular locations. Our results demonstrate for the first time the overall antioxidant enzyme status of individual kidney cell types, thereby explaining why different cell types have differing susceptibilities to oxidant stress. Possible physiological and pathological consequences of these findings are discussed.     

Proposed mechanism for neonatal rat tolerance to normobaric hyperoxia

Induction of two forms of superoxide dismutase, catalase and glutathione peroxidase, occurs very rapidly in neonatal rat lung tissue upon exposure of these animals to 94 + % normobaric oxygen. No such oxygen-mediated enzyme induction occurs in the lungs of adult rats. The aged-dependent pattern of enzyme induction correlates with the well-established age-dependent tolerance of neonatal rats to hyperoxia. Enzyme induction occurs in the lungs of neonates in only those species known to be resistant to oxygen-provoked lung damage. Compromise of oxygen-mediated enzyme induction predisposed the neonatal rats to pulmonary oxygen toxicity. These data have formed the basis of the proposal that oxygen induction of the superoxide dismutases catalase and glutathione peroxidase provides a vital part of the defense mechanism against oxygen toxicity. A biochemical mechanism of oxygen-provoked pulmonary damage has been elaborated to explain the role of each enzyme in the protection against oxygen and free radical toxicity.     

Expression of bovine and mouse endothelial cell antioxidant enzymes following TNF-alpha exposure

Endothelial cells are primary targets for injury by reactive oxygen species. Endothelial catalase, copper-zinc superoxide dismutase (CuZnSOD), and manganous superoxide dismutase (MnSOD) provide potential antioxidant enzymatic defenses against oxidant-induced cellular damage. Previous studies in vivo and in vitro have demonstrated that in certain cell types exposure to oxidants may increase the expression of one or more of these antioxidant enzymes, thus providing greater intracellular potential to withstand oxidant-induced cell stress. To test whether endothelial antioxidant enzyme expression is influenced by similar oxidant-induced stresses in vitro, we have exposed endothelial cells to tumor necrosis factor-alpha (TNF-alpha) and have measured levels of catalase, CuZnSOD and MnSOD mRNA, and protein. Our results demonstrate a selective increase of MnSOD mRNA, with coordinate increases of both MnSOD protein and enzyme activity in endothelial cells treated for 24/h with TNF-alpha. In contrast, levels of catalase and CuZnSOD mRNA and protein remained unchanged in these cells after TNF-alpha treatment. These observations were made in microvessel endothelial cells derived from murine and bovine sources. Our results indicate that TNF-alpha can act specifically to increase enzymatic antioxidant potential in endothelial cells by induction of a particular antioxidant enzyme encoding mRNA species. These data demonstrate the capacity of endothelial cells to mount an antioxidant defense in response to exposure to an inducer of oxidative damage.     

Pulmonary oxygen toxicity: Lack of tolerance of mice of various ages

Prolonged continuous exposure of adult (3–4 months) and old (21 months) mice to hyperoxia did not lead to significant changes in the activities of superoxide dismutase and catalase in liver or blood. Lung superoxide dismutase activity increased by 25% during initial exposure to 100% O₂, but then fell progressively to below control level. Exposure of mice to 60% or 80% O₂ increased their susceptibility to further exposure to 100% O₂. The results clearly show that both adult and old mice are incapable of coping with the high oxygen environment and that antioxidant enzyme induction and the associated partial protection from pulmonary O₂ toxicity are not the general rule in mammalian lung exposed to subtoxic oxygen levels.     

Wood smoke extract induces oxidative stress-mediated caspase-independent apoptosis in human lung endothelial cells: role of AIF and EndoG

Although a link between toxic smoke and oxidant lung vascular injury has been indicated, the cellular mechanisms of smoke-induced injury to lung endothelial cells are unknown. We investigated oxidative stress and apoptosis induced by wood smoke extract (SE) in human pulmonary artery endothelial cells (HPAECs) and delineated their relationship. We found that SE increased intracellular reactive oxygen species (ROS), depleted intracellular glutathione, and upregulated Cu/Zn superoxide dismutase and heme oxygenase-1 (2 antioxidant enzymes), but it failed to alter the expression of catalase and glutathione peroxidase. In addition, SE promoted apoptosis as indicated by the external exposure of membrane phosphatidylserine, the loss of mitochondrial membrane potential, an increase in the level of Bax (a proapoptotic protein), and enhanced DNA fragmentation. This apoptosis was associated with mitochondrial-to-nuclear translocation of apoptosis-inducing factor (AIF) and endonuclease G (EndoG) (2 apoptogenic proteins) but was independent of caspase cascade activation. Whereas N-acetylcysteine (an antioxidant) effectively reversed the SE-induced increase in ROS and depletion of glutathione, it also suppressed SE-induced nuclear translocation of either AIF or EndoG and prevented the enhanced DNA fragmentation that would have resulted from this. We conclude that 1) although SE upregulates Cu/Zn superoxide dismutase and heme oxygenase-1, it nevertheless increases intracellular oxidative stress in HPAECs, and 2) SE promotes oxidative stress-mediated caspase-independent HPAEC apoptosis that involves mitochondrial-to-nuclear translocation of AIF and EndoG. Thus modulations of the expression of antioxidant enzymes and the caspase-independent apoptotic pathway are possible target choices for potential therapeutic regimes to treat smoke-induced lung injury.     

Retinol supplementation induces oxidative stress and modulates antioxidant enzyme activities in rat sertoli cells

Recent intervention studies revealed that supplementation with retinoids resulted in a higher incidence of lung cancer. Recently the causal mechanism has begun to be clarified. We report here that retinol caused cellular oxidative stress and modulated superoxide dismutase, catalase and glutathione peroxidase activities. Retinol (7 μM) significantly increased TBARS, conjugated dienes, and hydroperoxide-initiated chemiluminescence in cultured Sertoli cells. In response to retinol treatment superoxide dismutase, catalase and glutathione peroxidase activities increased. TBARS content and catalase activities were decreased by a free radical scavenger. These findings suggest that retinol may induce oxidative stress and modulate antioxidant enzyme activities in Sertoli cells.     

Free radical damage to cultured porcine aortic endothelial cells and lung fibroblasts: Modulation by culture conditions

Summary Culture conditions modulating cell damage from xanthine plus xanthine oxidase-derived partially reduced oxygen species were studied. Porcine thoracic aorta endothelial cells and porcine lung fibroblasts were maintained in monolayer culture. Cells were prelabeled with⁵¹Cr before xanthine plus xanthine oxidase exposure. Endothelial cells showed 30 to 100% more lysis than fibroblasts and thus seemed more sensitive to this oxidant stress. The effect of cell culture age, as indicated by population doubling level (PDL), was examined. Response of low PDL endothelial cells and fibroblasts subjected to oxidant stress was compared with the response of PDL 15 cells. Both low PDL endothelial cells and fibroblasts responded differently to the lytic effect of xanthine oxidase-derived free radicals than did higher PDL cells. Specific activities of the antioxidant enzymes catalase, managanese superoxide dismutase, copper-zinc superoxide dismutase, glutathione peroxidase, and glucose-6-phosphate dehydrogenase were measured in both low and high PDL fibroblasts and endothelial cells. Antioxidant enzyme specific activities could only partially explain the differences in response to oxidant stress between fibroblasts and endothelial cells and between low and high PDL cells. Cell culture medium composition modulated the rate of production, and relative proportions of xanthine plus xanthine oxidase-derived partially reduced species of oxygen, i.e. superoxide, hydrogen peroxide, and hydroxyl radical. Serum content of medium was important in modulating free radical generation; superoxide production rates decreased 32%, H₂O₂ became undetectable, and hydroxyl radical generation decreased 54% in the presence of 10% serum. The medium protein and iron content also modulated free radical generation. The data suggest that cell culture media constituents, cell type, and cell culture age greatly affect in vitro response of cells subjected to oxidant stress. Research supported by American Lung Association Fellowship Training Grant and Research Training Grant, the R. J. Reynolds Corporation, and National Institutes of Health Grants HL29784 and 1 HL 23805.     

Retinol supplementation induces oxidative stress and modulates antioxidant enzyme activities in rat Sertoli cells

Recent intervention studies revealed that supplementation with retinoids resulted in a higher incidence of lung cancer. Recently the causal mechanism has begun to be clarified. We report here that retinol caused cellular oxidative stress and modulated superoxide dismutase, catalase and glutathione peroxidase activities. Retinol (7 μM) significantly increased TBARS, conjugated dienes, and hydroperoxide-initiated chemiluminescence in cultured Sertoli cells. In response to retinol treatment superoxide dismutase, catalase and glutathione peroxidase activities increased. TBARS content and catalase activities were decreased by a free radical scavenger. These findings suggest that retinol may induce oxidative stress and modulate antioxidant enzyme activities in Sertoli cells.     

Difference in proangiogenic potential of systemic and pulmonary endothelium: role of CXCR2

The systemic vasculature in and surrounding the lung is proangiogenic, whereas the pulmonary vasculature rarely participates in neovascularization. We studied the effects of the proangiogenic ELR+ CXC chemokine MIP-2 (macrophage inflammatory protein-2) on endothelial cell proliferation and chemotaxis. Mouse aortic, pulmonary arterial, and lung microvascular endothelial cells were isolated and subcultured. Proliferation ([3H]thymidine uptake) and migration (Transwell chemotaxis) were evaluated in each cell type at baseline and upon exposure to MIP-2 (1-100 ng/ml) without and with exposure to hypoxia (24 h)-reoxygenation. Baseline proliferation did not vary among cell types, and all cells showed increased proliferation after MIP-2. Aortic cell chemotaxis increased markedly upon exposure to MIP-2; however, neither pulmonary artery nor lung microvascular endothelial cells responded to this chemokine. Assessment of CXCR2, the G protein-coupled receptor through which MIP-2 signals, displayed no baseline difference in mRNA, protein, or cell surface expression among cell types. Exposure to hypoxia increased expression of CXCR2 of aortic endothelial cells only. Additionally, aortic cells, compared with pulmonary cells, showed significantly greater protein and activity of cathepsin S, a proteolytic enzyme important for cell motility. Thus the combined effects of increased cathepsin S activity, providing increased motility and enhanced CXCR2 expression after hypoxia, both contribute to the proangiogenic phenotype of systemic arterial endothelial cells.     

Modulation of oxidant lung injury by using liposome-entrapped superoxide dismutase and catalase

Increased cellular generation of partially reduced species of oxygen mediates the toxicity of hyperoxia to cultured endothelial cells and rats exposed to 95-100% oxygen. Liposomal entrapment and intracellular delivery of superoxide dismutase (SOD) to cultured porcine aortic endothelial cells increased the specific activity of cellular SOD up to 15-fold. The liposome-mediated augmentation of SOD activity persisted in cell monolayers and rendered these cells resistant to oxygen-induced injury in a cell SOD activity-dependent manner. Addition of free SOD to culture medium had no effect on cell SOD activity or resistance to oxygen toxicity. SOD and catalase-containing liposomes injected i.v. into rats increased lung-associated enzyme specific activities two- to fourfold. Liposome entrapment of both SOD and catalase significantly increased the circulating half-lives of these enzymes and was critical for prevention of in vivo oxygen toxicity. Free SOD and catalase injected i.v. in the absence or presence of control liposomes did not increase corresponding lung enzyme activities or survival time in 100% oxygen. These studies show that O2- and H2O2 are important mediators of oxygen toxicity and that intracellular delivery of oxygen protective enzymes can reduce tissue injury owing to overproduction of partially reduced oxygen species.