Amphotericin B-induced apoptosis and cytotoxicity is prevented by the Na+, K+, 2Cl(-)-cotransport blocker bumetanide

Amphotericin B is the most commonly used antifungal drug although it exhibits poor effectiveness and considerable toxicity during treatment. It acts as a ionophore inducing cellular potassium efflux. The efflux of potassium, which is necessary for cell shrinkage during apoptosis, is counteracted by increased inward pumping of potassium ions. Modulation of potassium pump activity could therefore interact with programmed cell death depending on the nature of the disruption of cellular potassium homeostasis and subsequently affect the cytotoxicity of various drugs. We explored the role of apoptosis in amphotericin B-induced cytotoxicity in a mesothelioma cell line (P31) and investigated the role of K+ influx inhibitors of Na+, K+, ATPase and Na+, K+, 2Cl(-)-cotransport in these processes. Clone formation was used to determine the cytotoxicity of amphotericin B, ouabain (Na+, K+, ATPase blocker), and bumetanide (Na+, K+, 2Cl(-)-cotransport blocker), alone or in combination. Apoptosis was estimated by quantifying free nucleosomes. Amphotericin B (3.2 micromol/L, 3 mg/L) per se reduced the percentage of surviving clones to 64% and increased the number of nucleosomes by 31% compared to untreated control. When ouabain (100 micromol/L) was added to amphotericin B a less than additive effect on clone formation was seen but no reduction of nucleosomes was noted. Bumetanide (100 micromol/L) per se was not cytotoxic but increased cellular nucleosome expression. Bumetanide eradicated amphotericin B-induced reduction of formed clones and generated nucleosomes. In conclusion, the induction of apoptosis seems to be of significant importance in amphotericin B-induced cytotoxicity. Amphotericin B-induced cytotoxicity and apoptosis was eradicated by the Na+, K+, 2Cl(-)-cotransport inhibitor bumetanide. The changes of cellular K+ fluxes induced by bumetanide combined with amphotericin B needs further elucidation. Bumetanide could possibly be used in antifungal therapy to increase amphotericin B effectiveness doses without increasing its adverse effects.     

Protective effect rendered by human ceruloplasmin on erythrocytes in hepatocerebral dystrophy

In order to elucidate the protective effect of human ceruloplasmin (CP) on erythrocytes in patients with hepatocerebral dystrophy (HCD), the parameters reflecting the interaction of CP from the blood of healthy donors (n-CP) and of HCD patients (h-CP) with erythrocytes from healthy donors (n-ER) and from HCD patients (h-ER) were estimated. The protective effects of n-CP and h-CP on n-ER and h-ER during the Cu2+-induced lysis were compared. It was shown that the ability of h-CP to prevent the human ER breakdown upon Cu2+-induced lysis is much lower (approximately 3-fold) than that of n-CP. The differences in the protective effect of n-CP and h-CP are manifested in a greater degree during the n-ER lysis than during the h-ER lysis.     

Effect of amphotericin B on cholesterol-containing liposomes of egg phosphatidylcholine and didocosenoyl phosphatidylcholine. A refinement of the model for the formation of pores by amphotericin B in membranes

(1) Binding and K+-permeability measurements were performed on egg and 22 : 1c/22 : 1c-phosphatidylcholine liposomes with or without cholesterol. (2) Amphotericin B binds specifically to cholesterol in both types of liposome despite the difference in bilayer thickness. (3) Addition of amphotericin B to one side of the cholesterol-containing egg phosphatidylcholine bilayers induces a fast K+ efflux from the outermost compartment of the liposomes. In contrast, the total K+ content of sonicated unilamellar cholesterol-containing egg phosphatidylcholine vesicles is released by amphotericin B. (4) Amphotericin B addition to one side of the cholesterol-containing 22 : 1c/22 : 1c-phosphatidylcholine liposomes does not cause a change in K+ permeability. The presence of amphotericin B on both sides of the bilayer, however, induces an increase in K+ permeability. (5) A model is proposed which accounts for the effect of bilayer thickness on the amphotericin B-induced permeability changes in membranes.     

Interconnection between the structure and protective action of normal and pathological ceruloplasmin preparations in copper-induced erythrocyte lysis]

It was found that the differences in the protective effects of ceruloplasmin (CP) isolated from the blood of healthy donors and of the ceruloplasmin-like protein (pat-CP) isolated from the blood of patients with hepatovertebral dystrophy (HCD) during Ca(2+)-induced lysis of erythrocytes (RBC) result from significant changes in the carbohydrate fragment of pat-CP, the bulk of which (65%) is devoid of mannose and acetylglucosamine residues. According to the data from lentil-lectin Sepharose chromatography, only 4% of pat-CP molecules contain the [formula; see text] fragment necessary for the binding to ER receptors. The curves reflecting the Cu2+ accumulation in healthy donor ER and in pat-CP during the Cu(2+)-induced lysis were found to differ significantly. The ability of pat-CP to prevent the accumulation of Cu2+ in ER and pat-ER was markedly decreased compared with CP. Besides, CP prevented the diminution of reduced glutathione (GSH) in ER in a greater degree than pat-CP, whereas pat-ER, in contrast with CP, had no effect on the GSH concentration in pat-ER. It is suggested that the reactions occurring in the cell during Cu(2+)-induced lysis of ER and pat-ER are different.     

Internalization of Voltage-Dependent Sodium Channels in Fetal Rat Brain Neurons: A Study of the Regulation of Endocytosis

Abstract: In fetal rat brain neurons, activation of voltage-dependent Na⁺ channels induced their own internalization, probably triggered by an increase in intracellular Na⁺ level. To investigate the role of phosphorylation in internalization, neurons were exposed to either activators or inhibitors of cyclic AMP- and cyclic GMP-dependent protein kinases, protein kinase C, and tyrosine kinase. None of the tested compounds mimicked or inhibited the effect of Na⁺ channel activation. An increase in intracellular Ca²⁺ concentration induced either by thapsigargin, a Ca²⁺-ATPase blocker, or by A23187, a Ca²⁺ ionophore, was unable to provoke Na⁺ channel internalization. However, Ca²⁺ seems to be necessary because both neurotoxin- and amphotericin B-induced Na⁺ channel internalizations were partially inhibited by BAPTA-AM. The selective inhibitor of Ca²⁺/calmodulin-dependent protein kinase II, KN-62, caused a dose-dependent inhibition of neurotoxin-induced internalization due to a blockade of channel activity but did not prevent amphotericin B-induced internalization. The rate of increase in Na⁺ channel density at the neuronal cell surface was similar before and after channel internalization, suggesting that recycling of internalized Na⁺ channels back to the cell surface was almost negligible. Pretreatment of the cells with an acidotropic agent such as chloroquine prevented Na⁺ channel internalization, indicating that an acidic endosomal/lysosomal compartment is involved in Na⁺ channel internalization in neurons.     

Concentration- and time-dependence of amphotericin-B induced permeability changes across ergosterol-containing liposomes

The effect of amphotericin B on the permeability properties of liposomes prepared by reverse-phase evaporation was examined by using an osmotic method. This study has revealed that the magnitude and type of the alterations in permeability induced by amphotericin B in liposomes made of egg phosphatidylcholine and ergosterol depend not only on the amphotericin B concentration in the external aqueous solution but also on the time elapsed after mixing. Thus, low amphotericin B concentrations (from 0.2 to 1.2 microM) led to, (1) an small increment of the total extent of shrinkage of liposomes suspended in non-electrolytes such as urea or salts like KNO3, (2) an enhancement of urea and salt permeabilities at the same time scale at which volume changes were measured (ms to s), (3) a maximal blocking by tetraethylammonium of amphotericin B-induced urea permeability and (4) an enhancement of glucose permeability but only after liposomes were incubated with amphotericin B for some minutes before mixing. The high amphotericin B concentration regime (beyond 1.2 microM) led to, (1) a decrease of the total extent of shrinkage of liposomes immediately after rapid mixing of liposomes with urea solutions containing amphotericin B and (2) a 50% reduction of the tetraethylammonium blocking of amphotericin B-induced urea permeability. These results are explained by assuming that amphotericin B may form in ergosterol-containing liposomes two types of active channel differing in internal diameter.     

Oxygen requirement for cupric ion induced hemolysis

The Cu(II) induced lysis of rabbit erythrocytes occurs in two phases. First there is a lag period of 60 min where few cells lyse, followed by a period of rapid lysis where most of the remaining cells undergo hemolysis. Lysis is effectively inhibited if the incubation is conducted in nitrogen or if the erythrocytes are pre-saturated with carbon monoxide. These results suggest that oxygenated hemoglobin is necessary for lysis. It has been reported that copper binds to oxyhemoglobin and releases superoxide ion. This additional oxidative stress can initiate lipid peroxidation which leads to cell hemolysis.     

No evidence of oxidant events in amphotericin B cytotoxicity versus L. infantum promastigotes.

Amphotericin B is used for the treatment of systemic mycoses and visceral leishmaniasis. The objective of our study was to evaluate the impact of catalase, ascorbic acid and ketoconazole on the amphotericin B toxicity towards Leishmania promastigotes membrane by two flow cytometric tests, the membrane potential assay using a cationic dye, [DiOC5(3)], and the membrane permeability test using propidium iodide. The collapse of membrane potential appeared at amphotericin B concentrations weaker than those assessed by the membrane permeability test. The binding of amphotericin B to membrane sterol was not modified by catalase or ascorbic acid whereas amphotericin B-induced growth inhibition could be modulated by these products. The permeabilizing effect of amphotericin B on parasite membrane was strongly reduced in the presence of ketoconazole. These results confirmed the pore hypothesis of amphotericin B action and suggested that flow cytometric methods constituted a valuable alternative to conventional methods for assessing the effect of drugs on cellular membrane and evaluating parasite susceptibility to polyene antibiotics.     

Adenylate cyclase toxin from Bordetella pertussis. The relationship between induction of cAMP and hemolysis.

Bordetella pertussis produces a calmodulin-activated adenylate cyclase (AC) that exists in several forms. Only one form of AC, of apparent 200 kDa, is a toxin that penetrates eukaryotic cells and generates uncontrolled levels of intracellular cAMP. Recombination studies in transposon Tn5-insertion mutants of B. pertussis and amino acid sequence homology with alpha-hemolysin of Escherichia coli suggested that AC toxin may also have a hemolytic activity. Here, we demonstrate that only the toxic form of B. pertussis AC possesses hemolytic activity. Immunoblotting of membranes from sheep erythrocytes throughout the process of cell lysis detects the presence and accumulation of only the 200-kDa form of B. pertussis AC. cAMP generation induced by AC toxin in sheep erythrocytes is immediate whereas appearance of hemolysis is delayed by about 1 h and requires a higher level of AC toxin activity. Addition of exogenous calmodulin to sheep erythrocyte incubation medium potentiates the hemolytic activity of AC toxin but blocks cAMP generation. Extracellular Ca2+ at mM concentrations is absolutely required for cAMP generation but not for hemolysis. However, binding of AC toxin to sheep erythrocytes in the absence of exogenous Ca2+ followed by reincubation of cells in a toxin-free buffer containing Ca2+ leads to an immediate rise in intracellular cAMP. Human erythrocytes bind AC toxin and generate cAMP but are resistant to lysis. These results show that binding of AC toxin to erythrocytes can cause both cAMP generation and hemolysis or only one of these depending on conditions applied and cell type used.     

A tormentor in the quest for plant p53-like proteins

Pretreatment with salicylic acid (SA), an inducer of plant disease resistance, enhanced the capacity of parsley cells for the induction of a rapid K⁺/pH response and the subsequent coumarin (phytoalexin) secretion. In SA-primed cells, a low elicitor dose induced these two responses to a similar extent as did a high elicitor dose in non-primed cells. These observations suggest that the SA-mediated augmentation of the early K⁺/pH response may contribute to the enhancement of subsequent coumarin secretion. As the amphotericin B-induced K⁺/pH response was not enhanced in SA-primed cells, it is concluded that signaling components that are improved by priming are located between elicitor signal perception and the plasma membrane transporters mediating the K⁺/pH response.     

Spin-labeled amphotericin B: synthesis, characterization, biological and spectroscopic properties

A biologically active spin-labeled derivative of amphotericin B has been synthesized by the nucleophilic addition of amphotericin B to 4-(2-iodoacetamido)-2,2',6,6'-tetramethylpiperadine-N-oxyl in dimethyl-sulphoxide at 40 degrees C. The derivative is a moderately water-soluble compound which displays the same biological activity of the parental compound against the sensitive organism Leishmania mexicana; also, the rates of proton-cation exchange induced by the two compounds in large unilamellar liposomes are indistinguishable. The ESR spectra of spin-labeled amphotericin B in lipid vesicles indicate a high degree of motion, very similar to that encountered for the compound in aqueous solutions at neutral pH and in deoxycholate micelles, and suggest that the structures formed by the antibiotic in membranes are composed by a small number of molecules. In contrast, the spectra of the labeled antibiotic in ethanol, diethyl ether and dimethylformamide indicate restricted motion and exchange interactions, probably resulting from the micellar aggregation induced in these media. Ascorbate at 10 mM is able to reduce completely the nitroxide group of the labeled antibiotic in lipid vesicles in less than 30 s, indicating that an asymmetric disposition of the antibiotic molecules across the membrane is capable of inducing its biological and ionophoric properties. Ni2+ and Cu2+ produce moderate exchange broadening of the ESR signal of spin-labeled amphotericin B in lipid vesicles; the comparison of this phenomenom with the exchange broadening produced by the same ions in the ESR spectrum of 2,2',6,6'-tetramethylpiperidine-N-oxyl in water solution suggests an specific Cu2+-amphotericin B interaction in membranes.     

The mechanism of the hemolytic activity of polyene antibiotics

The kinetics of the filipin-, amphotericin B- and nystatin-induced hemolysis of human erythrocytes were investigated. Filipin-induced hemolysis is of the damage type. It is an all-or-none process, partly inhibited by Ca2+ or Ba2+ but not by Mg2+, Na+ or SO42-. The hemolytic activity of filipin is explained by the formation of large aggregates within the erythrocyte membrane in the form of large perforations, permeable to substances of low molecular weight as well as to macromolecules, including hemoglobin. In isotonic KCl solution, both amphotericin B and nystatin, at low concentrations, form smaller aggregates within the membranes. As a result, the permeability of the membranes to KCl increases and hemolysis occurs. However, the kinetics of the hemolysis induced by the two polyenes is complex. The process shows some features of the permeability type and some of the damage type. It is suggested that amphotericin B and nystatin may simultaneously form a number of transport systems, differing in their molecular organisation and hemolytic activity. Their participation in erythrocyte membrane permeability can be modified by small changes in membrane organisation and the chemical composition of the incubation medium. In isotonic solutions of divalent cation chlorides, and at higher antibiotic concentration, additional aggregates, allowing divalent cations to permeate, appear. These structures do not permit SO4(2-) to permeate.     

Nucleated cell killing by complement: effects of C5b-9 channel size and extracellular Ca2+ on the lytic process

For C5b-9 channels to mediate cytolysis of a nucleated cell, a sufficient number of channels must be formed in the plasma membrane to override the compensatory mechanisms that nucleated cells might employ to survive. It is well known that nucleated cells are relatively resistant to lysis by complement in comparison to erythrocytes, and it is now evident that this resistance is due, in part, to the ability of nucleated cells to rapidly eliminate C5b-9 from the cell surface. The ability of nucleated cells to eliminate complement complexes is related to physiochemical properties of the complex, such as channel diameter, which in turn affect Ca2+ fluxes that stimulate metabolic processes involved in the elimination process. Paradoxically, these same channel properties that stimulate the defense response may also be responsible for the lethal effects of complement. To further study the role of channel size on cytolysis of nucleated cells by C5b-9, we examined the lytic efficiency of larger C5b-9 channels containing several C9 molecules in comparison with smaller C5b-9 channels containing fewer C9. We have obtained data to indicate that although the larger channels were more cytolytically potent, the channel size had little influence on the rate of cell death. In contrast, the rate of lysis of erythrocytes was substantially slower when smaller C5b-9 channels were present. In evaluating the effect of the extracellular Ca2+ concentration, [Ca2+]o, on nucleated cell lysis in the presence of a lytic number of C5b-9 complexes, it was observed that when the [Ca2+]o was increased the rate of cell death also increased. These findings suggest that lysis of nucleated cells by C5b-9, unlike erythrocytes, may not be entirely due to colloid osmotic deregulation.     

Clotrimazole enhances lysis of human erythrocytes induced by t-BHP

Clotrimazole (CLT) is an antifungal and antimalarial agent also effective as a Gardos channel inhibitor. In addition, CLT possesses antitumor properties. Recent data provide evidence that CLT forms a complex with heme (hemin), which produces a more potent lytic effect than heme alone. This study addressed the effect of CLT on the lysis of normal human erythrocytes induced by tert-butyl hydroperoxide (t-BHP). For the first time, it was shown that 10 μM CLT significantly enhanced the lytic effect of t-BHP on erythrocytes in both Ca²⁺-containing and Ca²⁺-free media, suggesting that the effect is not related to Gardos channels. CLT did not affect the rate of free radical generation, the kinetics of GSH degradation, methemoglobin formation and TBARS generation; therefore, we concluded that CLT does not cause additional oxidative damage to erythrocytes treated with t-BHP. It is tempted to speculate that CLT enhances t-BHP-induced changes in erythrocyte volume and lysis largely by forming a complex with hemin released during hemoglobin oxidation in erythrocytes: the CLT–hemin complex destabilizes the cell membrane more potently than hemin alone. If so, the effect of CLT on cell membrane damage during free-radical oxidation may be used to increase the efficacy of antitumor therapy.     

Lysis of erythrocytes from stored human blood by phospholipase C (Bacillus cereus).

The ability of phospholipase C (Bacillus cereus) to lyse erythrocytes from human blood that had been stored under Transfusion Service conditions for up to 16 weeks has been examined. When incubated at 20 degrees C with enzyme (0.03 mg/ml, 55 units/ml) for up to 1 h fresh erythrocytes were not lysed. After about 4 weeks of storage a population of very readily lysed erythrocytes appeared. The morphological changes in erythrocytes from blood stored up to 16 weeks were examined by scanning electron microscopy. The proportion of very readily lysed erythrocytes correlated well with the proportion of spheroechinocytes I. This morphological form was shown to be preferentially removed by phospholipase C and before lysis a transient appearance of smooth spheres occurred. The decrease in blood ATP concentrations on storage was measured and found to correlate with the disappearance of discoid erythrocyte forms, but not directly with the increased susceptibility of the erythrocytes to lysis by the enzyme. However, erythrocytes of up to at least 15 weeks of age could be made less susceptible to lysis by pre-incubation in a medium designed to cause intracellular regeneration of ATP. During the lysis of spheroechinocytes I by electrophoretically pure recrystallized phospholipase C a rapid degradation of phosphatidylcholine, phosphatidylethanolamine and phosphatidylserine + phosphatidylinositol) occurred together with a slower degradation of sphingomyelin.     

Trypanosoma cruzi: Involvement of proteolytic activity during cell fusion induced by epimastigote form

Human erythrocytes were fused by Trypanosoma cruzi from 7 and 14 day old culture (stationary and declination phases, respectively) while only lysis was induced by "4 day old culture parasite (exponential phase). Lysis and erythrocyte fusion were studied by phase contrast microscopy, measuring of hemolysis and gel electrophoresis. The fusogenicity is Ca²⁺-dependent while lysis is delayed in the absence of exogenous Ca²⁺. The proteolysis of erythrocyte protein bands 1, 2, 2.1, 2.3 and 3 are common features of both fusion and lysis processes. Nevertheless the breakdown rate of ankyrin (band 2.1) and band 3 are different in fused or in lysed cells. The lysis process is associated with a faster degradation of band 2.1 and increase of band 2.3 than in the case of the fusion process. By contrast, degradation of band 3 occurs faster in the fusion than in the lytic event. Treatment of fusogenic parasites but not erythrocytes with TPCK, soybean trypsin inhibitor or FCS inhibited to some extent the fusion process and the decrease of bands 1, 2, 2.1, 2.3 and 3. The results suggest that proteases from fusogenic parasites may be directly or indirectly involved in the proteolysis of band 2.1 in a way related to induction of fusion.Abbreviations TPCK Tos-Phe-CH₂Cl,1-chloro-4 phenyl-3-L-tosylamidobutan-2-one - SDS-PAGE Sodium Dodecyl Sulphate-Polyacrylamide Gel Electrophoresis - FCS Fetal Calf Serum     

The action of sphingomyelinase from Bacillus cereus on ATP-depleted bovine erythrocyte membranes and different lipid composition of liposomes

The presence of cholesterol or phosphatidylethanolamine in sphingomyelin liposomes enhanced 2- to 10-fold the breakdown of sphingomyelin by sphingomyelinase from Bacillus cereus. On the other hand, the presence of phosphatidylcholine was either without effect or slightly stimulative at a higher molar ratio of phosphatidylcholine to sphingomyelin (3/1). In the bovine erythrocytes and their ghosts, the increase by 40-50% or the decrease by 10-23% in membranous cholesterol brought about acceleration or deceleration of enzymatic degradation of sphingomyelin by 50 or 40-50%, respectively. The depletion of ATP (less than 0.9 mg ATP/100 ml packed erythrocytes) enhanced K+ leakage from, and hot hemolysis (lysis without cold shock) of, bovine erythrocytes but decelerated the breakdown of sphingomyelin and hot-cold hemolysis (lysis induced by ice-cold shock to sphingomyelinase-treated erythrocytes), either in the presence of 1 mM MgCl2 alone or in the presence of 1 mM MgCl2 and 1 mM CaCl2. Also, ATP depletion enhanced the adsorption of sphingomyelinase onto bovine erythrocyte membranes in the presence of 1 mM CaCl2 up to 81% of total activity, without appreciable K+ leakage and hot or hot-cold hemolysis. These results suggest that the presence of cholesterol or phosphatidylethanolamine in biomembranes makes the membranes more susceptible to the attack of sphingomyelinase from B. cereus and that the segregation of lipids and proteins in the erythrocyte membranes by ATP depletion causes the deceleration of sphingomyelin hydrolysis despite the enhanced enzyme adsorption onto the erythrocyte membranes.     

Distinction of two components of passive Ca2+ transport into human erythrocytes by Ca2+ entry blockers

The nature of downhill Ca2+ net-transport into human erythrocytes was investigated using the experimental models of Ca2+ pump inhibition by vanadate and of intracellular chelation of Ca2+ by quin2. Ca2+ uptake by erythrocytes loaded with 0.5 mM vanadate and suspended in 145 mM Na+ -5 mM K+ media was reduced by about 60% when medium K+ was raised to 80 mM. Organic and inorganic Ca2+ entry blockers such as nifedipine (10(-5) M), verapamil (10(-4) M), diltiazem (10(-4) M), Co2+ (1.5 mM) and Cu2+ (0.1 mM) as well as the K+ channel blocker quinidine (1mM) inhibited Ca2+ uptake in 145 mM Na+ -5 mM K+ media by 60-75%. Flunarizine was less effective. In vanadate-loaded cells suspended in 70 mM Na+ -80 mM K+ media, in contrast, flunarizine exerted a dose-dependent inhibition of Ca2+ uptake by up to 80% at 10(-5) M, the other blockers being ineffective (except for verapamil at 10(-4) M). A similar pattern of inhibition was seen in quin2-loaded erythrocytes. The different susceptibility towards inhibitors may indicate that passive Ca2+ uptake by vanadate-loaded erythrocytes suspended in 145 mM Na+ -5 mM K+ media, on the one hand, and by vanadate-loaded erythrocytes suspended in 70 mM Na+ -80 mM K+ media as well as by quin2-loaded erythrocytes, on the other hand, is mediated by two different transport components.     

Liposome oxidation and erythrocyte lysis by enzymically generated superoxide and hydrogen peroxide.

Xanthine oxidase, acting on acetaldehyde under aerobic conditions, produces a flux of O2- and H2O2 which attacks artificial liposomes and washed human erythrocytes. The liposomes were peroxidized and the erythrocytes suffered oxidation of hemoglobin followed by lysis. The oxidation of hemoglobin followed by lysis. The oxidation of hemoglobin, within the exposed erythrocytes, could be largely prevented by prior conversion to carbon monoxyhemoglobin, without preventing lysis. Hemolysis thus appeared to be a consequence of direct oxidative attack on the cell stroma. The enzyme-generated flux of O2- and of H2O2 also inactivated the xanthine oxidase. Superoxide dismutase or catalase, present in the suspending medium, protected the liposomes against peroxidation, the erythrocytes against lysis, and the xanthine oxidase against inactivation. Scavengers of O2('deltag), such as histidine or 2,5-dimethylfuran, which do not react with O2- or H2O2, also prevented peroxidation of liposomes and lysis of erythrocytes when present at low concentrations. In contrast a scavenger of OH-, such as mannitol was ineffective at low concentrations and provided significant protection only at much higher concentrations. It is proposed that O2- and H2O2 cooperated in producing OH- and O2('deltag), which were the proximate causes of lipid peroxidation and of hemolysis.