Studies with primaquine in vitro: superoxide radical formation and oxidation of haemoglobin.

1. The production of superoxide radicals from primaquine diphosphate in aqueous solution has been demonstrated, using as indicator the reduction of cytochrome C with inhibition of the reaction by superoxide dismutase. 2. Primaquine-mediated oxidation of haemoglobin to methaemoglobin was reduced by the addition of catalase and increased by superoxide dismutase. Mannitol, a hydroxyl radical scavenger, abolished the increase in methaemoglobin observed in the presence of superoxide dismutase. EDTA reduced the oxidation of haemoglobin with and without superoxide dismutase. 3. Although the oxidation of haemoglobin in the presence of primaquine includes the effects of hydrogen peroxide, superoxide and hydroxyl radicals and metal ions, the results indicate that hydrogen peroxide, rather than the superoxide radical, is the main oxidizing species. The increase in haemoglobin oxidation occurring with superoxide dismutase may result from the augmented rate of hydrogen peroxide formation from superoxide radicals.     

Effects of an NADPH-generating system on primaquine degradation by hamster liver fractions

1. Primaquine (PQ) often causes severe anaemia in individuals with glucose 6-phosphate dehydrogenase (G6PD) deficient erythrocytes, and metabolites have been implicated as the toxic substance. These studies present data identifying additional metabolites of PQ. 2. Two metabolites of primaquine (PQ) previously identified in human studies, namely, 6-methoxy-8-aminoquinoline (MAQ) and 8-(3-carboxy-1-methylpropylamino)-6-methoxyquinoline (PQC) were also formed on incubation of PQ with hamster liver fractions for up to 24 h without an NADPH-generating system. 3. The alcohol (PQAOH) and lactam (PQLT) derivatives of PQ were also formed on incubation with hamster liver fraction used in these studies. 4. The microsomal metabolism of PQ was decreased in presence of an NADPH-generating system, but not by SKF-525A or glutathione (GSH) indicating that the oxidative reactions were probably not due to the cytochrome P-450 system or free radical mechanisms.     

Effects of an Nadph-Generating System on Primaquine Degradation by Hamster Liver Fractions

1.Primaquine (PQ) often causes severe anaemia in individuals with glucose 6-phosphate dehydrogenase (G6PD) deficient erythrocytes, and metabolites have been implicated as the toxic substance. These studies present data identifying additional metabolites of PQ.

2.Two metabolites of primaquine (PQ) previously identified in human studies, namely, 6-methoxy-8-aminoquinoline (MAQ) and 8-(3-carboxy-1-methylpropylamino)-6-methoxyquinoline (PQC) were also formed on incubation of PQ with hamster liver fractions for up to 24 h without an NADPH-generating system.

3.The alcohol (PQAOH) and lactam (PQLT) derivatives of PQ were also formed on incubation with hamster liver fraction used in these studies.

4.The microsomal metabolism of PQ was decreased in presence of an NADPH-generating system, but not by SKF-525A or glutathione (GSH) indicating that the oxidative reactions were probably not due to the cytochrome P-450 system or free radical mechanisms.     

Understanding the mechanisms for metabolism-linked hemolytic toxicity of primaquine against glucose 6-phosphate dehydrogenase deficient human erythrocytes: evaluation of eryptotic pathway

Therapeutic utility of primaquine, an 8-aminoquinoline antimalarial drug, has been limited due to its hemolytic toxicity in population with glucose 6-phosphate dehydrogenase deficiency. Recent investigations at our lab have shown that the metabolites generated through cytochrome P(450)-dependent metabolic reactions are responsible for hemotoxic effects of primaquine, which could be monitored with accumulation of methemoglobin and increased oxidative stress. The molecular markers for succeeding cascade of events associated with early clearance of the erythrocytes from the circulation were evaluated for understanding the mechanism for hemolytic toxicity of primaquine. Primaquine alone though did not induce noticeable methemoglobin accumulation, but produced significant oxidative stress, which was higher in G6PD-deficient than in normal erythrocytes. Primaquine, presumably through redox active hemotoxic metabolites generated in situ in human liver microsomal metabolism-linked assay, induced a dose-dependent methemoglobin accumulation and oxidative stress, which were almost similar in normal and G6PD-deficient erythrocytes. Primaquine alone or in presence of pooled human liver microsomes neither produced significant effect on intraerythrocytic calcium levels nor affected the phosphatidyl serine asymmetry of the normal and G6PD-deficient human erythrocytes as monitored flowcytometrically with Annexin V binding assay. The studies suggest that eryptosis mechanisms are not involved in accelerated removal of erythrocytes due to hemolytic toxicity of primaquine.     

Cytochrome P450-dependent toxic effects of primaquine on human erythrocytes

Primaquine, an 8-aminoquinoline, is the drug of choice for radical cure of relapsing malaria. Use of primaquine is limited due to its hemotoxicity, particularly in populations with glucose-6-phosphate dehydrogenase deficiency [G6PD(−)]. Biotransformation appears to be central to the anti-infective and hematological toxicities of primaquine, but the mechanisms are still not well understood. Metabolic studies with primaquine have been hampered due to the reactive nature of potential hemotoxic metabolites. An in vitro metabolism-linked hemotoxicity assay has been developed. Co-incubation of the drug with normal or G6PD(−) erythrocytes, microsomes or recombinant cytochrome P₄₅₀ (CYP) isoforms has allowed in situ generation of potential hemotoxic metabolite(s), which interact with the erythrocytes to generate hemotoxicity. Methemoglobin formation, real-time generation of reactive oxygen intermediates (ROIs) and depletion of reactive thiols were monitored as multiple biochemical end points for hemotoxicity. Primaquine alone did not produce any hemotoxicity, while a robust increase was observed in methemoglobin formation and generation of ROIs by primaquine in the presence of human or mouse liver microsomes. Multiple CYP isoforms (CYP2E1, CYP2B6, CYP1A2, CYP2D6 and CYP3A4) variably contributed to the hemotoxicity of primaquine. This was further confirmed by significant inhibition of primaquine hemotoxicity by the selective CYP inhibitors, namely thiotepa (CYP2B6), fluoxetine (CYP2D6) and troleandomycin (CYP3A4). Primaquine caused similar methemoglobin formation in G6PD(−) and normal human erythrocytes. However, G6PD(−) erythrocytes suffered higher oxidative stress and depletion of thiols than normal erythrocytes due to primaquine toxicity. The results provide significant insights regarding CYP isoforms contributing to hemotoxicity and may be useful in controlling toxicity of primaquine to increase its therapeutic utility.     

Studies on the mechanism of toxicity of the mycotoxin sporidesmin. 2--Evidence for intracellular generation of superoxide radical from sporidesmin

Formation of hydrogen peroxide, the dismutation product of superoxide radical, has been demonstrated in erythrocytes incubated with the mycotoxin sporidesmin. Erythrocytic thiols, both non-protein and protein-bound, were depleted in the presence of sporidesmin, whilst haemoglobin was oxidized to methaemoglobin. Irreversible haemoglobin oxidation also occurred in these cells, shown by the formation of Heinz bodies; purified haemoglobin likewise suffered oxidative damage when incubated with sporidesmin in the presence of glutathione. Sporidesmin has previously been shown to generate superoxide radical in vitro; the erythrocytic changes induced by the mycotoxin, which are characteristically produced by compounds which generate 'active oxygen' species, suggest that it is also capable of generating this radical intracellularly.     

Phenol and catechol induce prehemolytic and hemolytic changes in human erythrocytes

The toxic potency of two industrially used compounds (phenol and catechol) was studied in human blood cells in vitro. Catechol was found to be a more harmful toxin than phenol, since it provokes statistically significant changes in the function of erythrocytes even at low doses. Most of the changes was statistically significant for the doses of 50 ppm of catechol and 250 ppm of phenol. Both compounds induced methaemoglobin formation, glutathione depletion and conversion of oxyhaemoglobin to methaemoglobin, which is associated with superoxide anion production and lead to formation of ferryl hemoglobin, hydrogen peroxide or hydroxyl radicals. It is known that oxidation of catechol leads to formation of semiquinone radicals. Semiquinones are able to bind to nucleophilic residues like -SH or -NH2 of proteins and these macromolecules may undergo inactivation. We observed among especially susceptible to action of catechol are catalase (CAT) (100 ppm) and superoxide dismutase (SOD) (250 ppm). Decrease of the activity of catalase and SOD by catechol induced radical species formation. This lead to inhibition of another protective enzymes such as glutathione-S-transferase (500 ppm), glutathione reductase (1000 ppm), glucose-6-phosphate dehydrogenase activity (1000 ppm). Cytotoxicity of phenol or catechol was noted as hemolysis. Haemoglobin liberated from erythrocytes in this process may further generate oxygen free radicals and subsequently initiate enzymes damage. It seems to be essential that in phenol and catechol toxicity special role play damages of heme proteins and other proteins molecule, and damages of lipids are not so important.     

Protection of wheat bran feruloyl oligosaccharides against free radical-induced oxidative damage in normal human erythrocytes

The present work assessed the protective effect of water-soluble feruloyl oligosaccharides (FSH), ferulic acid ester of oligosaccharides from wheat bran, against in vitro oxidative damage of normal human erythrocytes induced by a water-soluble free radical initiator, 2,2′-azobis-2-amidinopropane dihydrochloride (AAPH). In the whole process of AAPH-initiated oxidation, hemolysis occurred quickly after the lag time. The rate of hemolysis is correlated dose-dependently with AAPH concentration. Significant decrease in reduced glutathione (GSH) levels of erythrocyte with concomitant enhancement in oxidized gluthione (GSSG) levels was noticed. It was also observed that lipid and protein peroxidation of erythrocytes induced by AAPH was significantly increased, and scanning electron microscopy observations showed that AAPH induced obvious morphological alteration in the erythrocytes from a smooth discoid to an echinocytic form. FSH suppressed depletion of GSH, lipid peroxidation, and methaemoglobin and protein carbonyl group formation of erythrocytes in concentration- and time-dependent manners, remarkably delayed AAPH-induced hemolysis. Morphological changes to erythrocyte caused by AAPH were effectively protected by FSH. It was also observed that FSH could work synergistically with endogenous antioxidants in erythrocytes. These results indicated that FSH efficiently protected normal human erythrocytes against oxidative stress, and they could be used as a potential source of natural antioxidants.     

Oxidative activity of hydroxylated primaquine analogs. Non-toxicity to glucose-6-phosphate dehydrogenase-deficient human red blood cells in vitro

The individual effects of two putative metabolites of primaquine (5,6-dihydroxyprimaquine and 5,6-dihydroxy-8-aminoquinoline) on the hexose monophosphate shunt (HMS) and on the ATP-dependent proteolytic system which rapidly degrades oxidized erythrocyte protein were measured in intact red blood cells in vitro from two blood donors. In red cells treated with nitrite (1-40 mM) or phenylhydrazine (0.01-10 mM), proteolytic activity was detected only with concentrations (7.5 mM NaNO2 and 0.25 mM phenylhydrazine) causing greater than 15-fold elevation of HMS activity, and glucose-6-phosphate dehydrogenase (G6PD)-deficient (25% of normal activity) red cell suspensions thus treated showed approximately 30% greater proteolysis. G6PD-normal and deficient red cells treated with the primaquine analogs, however, did not experience proteolysis with concentrations (0.25 mM) in excess of those causing 17-fold elevation of HMS activity. Stimulation of the HMS by the primaquine analogs thus appears unrelated to an erythrotoxic oxidative stress. Methylene blue is known to cause an elevation of HMS activity through direct and diaphorase II-dependent oxidation of reduced nicotinamide adenine dinucleotide phosphate (NADPH) which is independent of injurious oxidative stress. It was found that the putative primaquine metabolites also caused direct and diaphorase II-dependent oxidation of NADPH in dilute hemolysate, thus suggesting that the putative primaquine metabolites have a methylene blue-like redox disposition in red blood cells. Results obtained in this study suggest that the hemolytic toxicity of primaquine may be unrelated to processes which lead to oxidative deterioration of red cell protein.     

The methaemoglobin forming and GSH depleting effects of dapsone and monoacetyl dapsone hydroxylamines in human diabetic and non-diabetic erythrocytes in vitro

The respective methaemoglobin forming and GSH depleting capabilities of monoacetyl dapsone hydroxylamine (MADDS-NHOH) and dapsone hydroxylamine (DDS-NHOH) were compared in human diabetic and non-diabetic erythrocytes in vitro with a view to select the most potent agent for future oxidative stress and antioxidant evaluation studies. Administration of both metabolites to non-diabetic erythrocytes over the 20min period of the study resulted in significantly more methaemoglobin formation at all four time points compared with the diabetic erythrocytes (P<0.0001). At all four time points, significantly more methaemoglobin was formed in response to MADDS-NHOH in non-diabetic cells compared with the effects of DDS-NHOH on diabetic erythrocytes (P<0.0001). At the 5 and 10min time points, significantly more methaemglobin was formed in non-diabetic cells in the presence of MADDS-NHOH compared with DDS-NHOH (P<0.05). At the 5min time point only, significantly more methaemoglobin was formed in the presence of MADDS-NHOH in diabetic cells compared with that of DDS-NHOH (P<0.01). However, compared with diabetic control GSH levels, the presence of DDS-NHOH caused a significant depletion in GSH at 5, 10 and 20min time points in diabetic cells (P<0.001). In addition, the presence of DDS-NHOH caused a significant reduction in GSH levels in diabetic cells in comparison with those of non-diabetics at the 5, 10 and 20min, (P<0.005). DDS-NHOH was also associated with a significant depletion of GSH levels in diabetic cells compared with those of non-diabetic control erythrocytes (P<0.0001). The presence of MADDS-NHOH in diabetic erythrocytes led to a significant reduction in GSH levels at the 20min time point compared with those of non-diabetics (P<0.001), but there were no significant differences at the 5, 10 and 15min points. Due to its greater GSH-depleting action, DDS-NHOH will be selected for future use in the oxidative stress assessment in diabetic erythrocytes.     

The interaction of phosphine with haemoglobin and erythrocytes.

1. Phosphine progressively converts oxyhaemoglobin to methaemoglobin and hemichrome species, with the product formed being time- and concentration-dependent. 2. The reaction of phosphine with oxyhaemoglobin leads to the formation of phosphite and phosphate. 3. Incubation of rat erythrocytes with various concentrations of phosphine results in the progressive uptake of phosphine by the erythrocytes in a temperature-dependent first-order process. 4. Uptake of phosphine by erythrocytes causes crenation, but conversion of oxyhaemoglobin to methaemoglobin and hemichrome could not be demonstrated.     

The interaction of phosphine with haemoglobin and erythrocytes

1. Phosphine progressively converts oxyhaemoglobin to methaemoglobin and hemichrome species, with the product formed being time- and concentration-dependent.

2. The reaction of phosphine with oxyhaemoglobin leads to the formation of phosphite and phosphate.

3. Incubation of rat erythrocytes with various concentrations of phosphine results in the progressive uptake of phosphine by the erythrocytes in a temperature-dependent first-order process.

4. Uptake of phosphine by erythrocytes causes crenation, but conversion of oxyhaemoglobin to methaemoglobin and hemichrome could not be demonstrated.     

Contributions of superoxide, hydrogen peroxide, and transition metal ions to auto-oxidation of the favism-inducing pyrimidine aglycone, divicine, and its reactions with haemoglobin

The influence of O2-, H2O2 and metal ions on the auto-oxidation of divicine, a pyrimidine aglycone, was studied. In air at pH 7.4, the hydroquinonic form oxidized within a few minutes. Superoxide dismutase (SOD) markedly decreased the initial rate, giving a lag phase followed by rapid oxidation. Although catalase or diethylenetriamine-penta-acetic acid (DTPA) alone had little effect, each in the presence of SOD further slowed the initial rate and increased the lag. H2O2 decreased the lag time, as did Cu2+, Fe2+ or haemoglobin. GSH substantially increased the lag phase, but it eventually reacted with the divicine to form a 305 nm-absorbing adduct. These results indicate that an O2(-)-dependent mechanism of divicine auto-oxidation normally predominates. Auto-oxidation can also occur by a mechanism involving H2O2 and transition metal ions or haemoglobin, and if both these reactions are prevented by SOD and DTPA or catalase, a third mechanism, requiring build-up of an autocatalytic intermediate, becomes operative. Oxyhaemoglobin did not react directly with divicine, but reacted with the H2O2 produced by divicine auto-oxidation to give mainly an oxidized derivative presumed to be ferrylhaemoglobin. Divicine was shown to reduce ferylhaemoglobin to methaemoglobin, and this reaction was probably responsible for the acceleratory effect of haemoglobin on divicine oxidation. These results indicate that O2 rather than oxyhaemoglobin is likely to initiate divicine oxidation in the erythrocyte. Haemolytic crises, which are thought to result from this oxidation, occur only sporadically in glucose-6-phosphate dehydrogenase deficient individuals following ingestion of fava beans. A characteristic of the crises is acute depletion of erythrocyte GSH, and the vulnerability of these cells could relate to the ability of GSH, in combination with SOD, to protect against the autocatalytic mechanism of divicine auto-oxidation. Our demonstration of a variety of auto-oxidation pathways also suggests possible areas of individual variation.     

Antioxidative effects of isoflavon genisteine-8-C-glycoside in vitro and in vivo

Bioflavonoids (polyhydroxyphenols) are ubiquitous components of plants, fruits and vegetables; these compounds are efficient scavengers of free oxygen radicals and peroxides. The aim of this study was to investigate the antioxidative effects of genistein-8-C-glycoside (G8CG) isolated from the flowers of Lipinus luteus L. G8CG dose-dependently inhibited membrane lipid peroxidation and prevented GSH oxidation in human red blood cells and rat liver homogenates under tert-butylhydroperoxide-induced oxidative stress and single whole-body gamma-irradiation (1 Gy) of rats.     

Effects of the antioxidants dihydrolipoic acid (DHLA) and probucol on xenobiotic-mediated methaemoglobin formation in diabetic and non-diabetic human erythrocytes in vitro(1)

The antioxidant effects of dihydrolipoic acid (DHLA) and probucol were investigated in a human erythrocytic in-vitro model of diabetic oxidative stress, where xenobiotics were used to form methaemoglobin. 4-Aminophenol mediated haemoglobin oxidation in non-diabetic erythrocytes was not affected by the presence of either DHLA or probucol. However, with diabetic cells, there were significant increases (P<0.01) in 4-aminophenol-mediated haemoglobin oxidation in the presence of DHLA. Methaemoglobin formed by nitrite in non-diabetic and diabetic cells was not altered by either DHLA or probucol except at one time point in diabetic cells. In non-diabetic as well as diabetic cells, methaemoglobin formed by MADDS-NHOH was significantly reduced at all three time points in the presence of DHLA (P<0.0001) but unaffected by probucol. In the presence of DHLA only, methaemoglobin formed by the products of rat microsomal oxidation of both 4-aminopropiophenone and benzocaine was markedly reduced for both xenobiotics in diabetic and non-diabetic cells (P<0.0001) compared with cells incubated in the absence of DHLA. There were no significant differences between total cellular thiol levels determined between diabetic and non-diabetic erythrocytes, nor did DHLA or probucol affect resting thiol levels. MADDS-NHOH caused a significant thiol depletion in diabetic cells, which was restored in the presence of DHLA. A further study is required to determine how DHLA attenuates the potent REDOX reactions that occur during hydroxylamine-mediated methaemoglobin formation.     

Damage of cell membrane and antioxidative system in human erythrocytes incubated with microcystin-LR in vitro.

The effect of the exposure of human erythrocytes to different concentrations of microcystin-LR were studied. Lipid peroxidation, membrane fluidity, cell morphology, haemoglobin oxidation and changes in the activity of antioxidant enzymes were investigated. Human erythrocytes were incubated with microcystin-LR at concentrations of 1-1000 nM for 1, 6, 12 and 24 h. We observed that microcystin-LR induces a significant increase of the level of thiobarbituric acid reactive substances (TBARS), formation of echinocytes, haemolysis, conversion of oxyhaemoglobin to methaemoglobin, decrease of membrane fluidity on the level of 16 carbon atom fat acids. The compound also changed antioxidative enzymes activities: catalase, superoxide dismutase and glutathione reductase and formation of reactive oxygen species (ROS). All of the observed changes point out that 100 nM of Microcistin LR is the liminal (threshold) toxic dose for human erythrocytes. This dose caused most of the described changes. Observed damages of erythrocytes membrane and antioxidative enzymes may be the result of direct covalent binding of microcystin-LR with -SH residues of proteins and indirectly be related with reactive oxygen species formation.     

Effects of the antioxidants dihydrolipoic acid (DHLA) and probucol on xenobiotic-mediated methaemoglobin formation in diabetic and non-diabetic human erythrocytes in vitro

The antioxidant effects of dihydrolipoic acid (DHLA) and probucol were investigated in a human erythrocytic in-vitro model of diabetic oxidative stress, where xenobiotics were used to form methaemoglobin. 4-Aminophenol mediated haemoglobin oxidation in non-diabetic erythrocytes was not affected by the presence of either DHLA or probucol. However, with diabetic cells, there were significant increases (P<0.01) in 4-aminophenol-mediated haemoglobin oxidation in the presence of DHLA. Methaemoglobin formed by nitrite in non-diabetic and diabetic cells was not altered by either DHLA or probucol except at one time point in diabetic cells. In non-diabetic as well as diabetic cells, methaemoglobin formed by MADDS-NHOH was significantly reduced at all three time points in the presence of DHLA (P<0.0001) but unaffected by probucol. In the presence of DHLA only, methaemoglobin formed by the products of rat microsomal oxidation of both 4-aminopropiophenone and benzocaine was markedly reduced for both xenobiotics in diabetic and non-diabetic cells (P<0.0001) compared with cells incubated in the absence of DHLA. There were no significant differences between total cellular thiol levels determined between diabetic and non-diabetic erythrocytes, nor did DHLA or probucol affect resting thiol levels. MADDS-NHOH caused a significant thiol depletion in diabetic cells, which was restored in the presence of DHLA. A further study is required to determine how DHLA attenuates the potent REDOX reactions that occur during hydroxylamine-mediated methaemoglobin formation.     

The antioxidant protection of CELLFOOD® against oxidative damage in vitro

CELLFOOD® (CF) is an innovative nutritional supplement containing 78 ionic/colloidal trace elements and minerals combined with 34 enzymes and 17 amino acids, all suspended in a solution of deuterium sulfate. The aim of this study was to investigate, for the first time, the antioxidant properties of CF in vitro in different model systems.Three pathophysiologically relevant oxidants were chosen to evaluate CF protection against oxidative stress: hydrogen peroxide, peroxyl radicals, and hypochlorous acid. Both biomolecules (GSH and plasmid DNA) and circulating cells (erythrocytes and lymphocytes) were used as targets of oxidation.CF protected, in a dose-dependent manner, both GSH and DNA from oxidation by preserving reduced GSH thiol groups and supercoiled DNA integrity, respectively. At the same time, CF protected erythrocytes from oxidative damage by reducing cell lysis and GSH intracellular depletion after exposure to the oxidant agents. In lymphocytes, CF reduced the intracellular oxidative stress induced by the three oxidants in a dose-dependent manner.The overall in vitro protection of biomolecules and cells against free radical attacks suggests that CF might be a valuable coadjuvant in the prevention and treatment of various physiological and pathological conditions related to oxidative stress, from aging to atherosclerosis, from neurodegeneration to cancer.     

Cyclosporine A-induced changes to erythrocyte redox balance is time course-dependent

Cyclosporine A-treated transplant recipients develop pronounced cardiovascular disease and have increased oxidative stress and altered antioxidant capacity in erythrocytes and plasma. These experiments investigated the time-course of cyclosporine A-induced changes to redox balance in plasma and erythrocytes. Rats were randomly assigned to either a control or cyclosporine A-treated group. Treatment animals received 25 mg/kg of cyclosporine A via intraperitoneal injection for either 7 days or a single dose. Control rats were injected with the same volume of the vehicle. Three hours after the final injections, plasma was analysed for total antioxidant status, alpha-tocopherol, malondialdehyde, and creatinine. Erythrocytes were analysed for reduced glutathione (GSH), alpha-tocopherol, methaemoglobin, malondialdehyde, and the activities of superoxide dismutase, catalase, GSH peroxidase, and glucose-6-phosphate dehydrogenase (G6PD). Cyclosporine A administration for 7 days resulted in a significant increase (P<0.05) in plasma malondialdehyde, methaemoglobin, and superoxide dismutase and catalase activities. There was a significant decrease (P<0.05) in erythrocyte GSH concentration and G6PD activity in cyclosporine A animals. There were no significant differences (P>0.05) between groups following a single dose of cyclosporine A in any of the measures. In summary, cyclosporine A alters erythrocyte redox balance after 7 days administration, but not after a single dose.     

Oxidative effects in human erythrocytes caused by some oximes and hydroxylamine

Both oximes and hydroxylamine (HYAM) are compounds with known oxidative capacity. We tested in vitro whether acetaldoxime (AAO), cyclohexanone oxime (CHO), methyl ethyl ketoxime (MEKO) or HYAM affect haemoglobin oxidation (into HbFe³⁺), formation of thiobarbituric acid reactive substances (TBARS), and glutathione (GT) depletion in human haemolysate, erythrocytes or blood. All these parameters are known to be related to oxidative stress. Glutathione S-transferase (GST) activity was measured as it may be affected by oxygen radicals. All three oximes caused a low degree of HbFe³⁺ accumulation in erythrocytes. This was higher in haemolysates indicating that membrane transport may be limiting or that protective mechanisms within erythrocytes are more effective. HbFe³⁺ accumulation was lower for the oximes than for HYAM. AAO and HYAM caused TBARS formation in blood. For HYAM this was expected as free radicals are known to be generated during HbFe³⁺ formation. Free radical generation by AAO and HYAM in erythrocytes was confirmed by the inhibition of GST. For the other two oximes (CHO and MEKO) some special effects were found. CHO did inhibit erythrocyte GST while it did not cause TBARS formation. MEKO was the least potent oxime as it caused no TBARS formation, little HbFe³⁺ accumulation and little GST inhibition in erythrocytes. However, GT depletion was more pronounced for MEKO than for the other oximes, indicating that glutathione conjugation occurs. TBARS formation, GT depletion and GST modulation caused by the oximes and HYAM were also tested in rat hepatocytes. However, no effects were found in hepatocytes. This suggests that a factor present in erythrocytes is necessary for free radical formation. Studies with proposed metabolites of the oximes (i.e. cyclohexanone, acetaldehyde or methylethyl ketone) and addition of rat liver preparations to the erythrocyte incubations with oximes, suggest that metabolism is not a limiting factor in erythrocyte toxicity. Received: 18 August 1997 / Accepted: 21 October 1997