Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Hemolytic anemia and methemoglobinemia are well recognized dose-limiting side effects in the therapeutic use of arylamine drugs, such as primaquine and dapsone (Beutler, 1969). Because these compounds are not hemotoxic when incubated with red cells in vitro, it has long been appreciated that metabolites are responsible for the onset of the hemolytic response. In the cases of aniline (Harrison and Jollow, 1986), dapsone (Grossman and Jollow, 1988), and phenacetin (Jensen and Jollow, 1991), we have shown that the hemolytic metabolites are their N-hydroxy derivatives. Primaquine metabolism, on the other hand, is relatively more complex, and the metabolites that mediate the hemotoxic responses have not been identified. A variety of known and putative phenolic metabolites of primaquine are redox-active and therefore have the potential to mediate primaquine hemotoxicity (Strother et al., 1984; Baird et al., 1986; Fletcher et al., 1988; Agarwal et al., 1991); however, direct evidence for their hemolytic activity is lacking. We have recently explored an alternative hypothesisthat primaquine hemotoxicity is mediated by an N-hydroxylated metabolite, 6-methoxy-8-hydroxylaminoquinoline (MAQ-NOH).

Although the mechanism underlying the damage and removal of red cells by hemolytic N-hydroxylamines remains unclear, oxidative stress has long been considered to play a prominent role in the process (for review, see Beutler, 1971). This concept is based on the well known association of hemotoxicity with oxidation of erythrocytic GSH (to glutathione disulfide and glutathione protein-mixed disulfides), with methemoglobin formation, and with enhanced sensitivity to the hemolytic effect of these agents in individuals deficient in glucose-6-phosphate dehydrogenase activity.

Considerable evidence generated by Kiese (1974) suggests that the oxidative stress provoked by N-hydroxylamines is due to a cyclic oxidation-reduction reaction involving the arylhydroxylamine oxyhemoglobin and molecular oxygen, which yields the nitrosoarene methemoglobin and partially reduced forms of oxygen, respectively. This interaction has been shown to produce greater than stoichiometric amounts of reactive oxygen species (i.e., hydrogen peroxide, superoxide anion radical, and hydroxyl radical) and sulfur-centered free radicals (i.e., glutathione and hemoglobin thiyl radicals) and, thus, has been proposed to generate reactive species capable of causing cellular injury (Rostorfer and Cormier, 1957; Maples et al., 1990; Bradshaw et al., 1995; Bradshaw et al., 1997).

We have reported recently that an N-hydroxy metabolite of primaquine, MAQ-NOH, is a direct-acting hemolytic agent in rats and, hence, may be a contributor to primaquine-induced hemolytic anemia (Bolchoz et al., 2001). Although the hemolytic response induced by this metabolite was generally similar to that of other structurally related hydroxylamines, we observed some marked differences in the oxidative activity of this metabolite in rat red cells (Bolchoz et al., 2002). In particular, when hemolytic concentrations of MAQ-NOH were added to rat red cells, the depletion of cellular GSH and the formation of methemoglobin were significantly less than that seen with equally hemolytic concentrations of other arylhydroxylamines, implying a quantitative difference in the extent of oxidant stress and cellular injury. Furthermore, this metabolite induced lipid peroxidation without evidence of protein oxidation. When red cell GSH was depleted by titration with diethyl maleate (DEM) before MAQ-NOH exposure, however, there was a marked enhancement of hemolytic activity without a corresponding increase in lipid peroxidation. The enhanced hemotoxicity was accompanied instead by the appearance of protein oxidation in the form of disulfide-linked hemoglobin-skeletal protein adducts. We have hypothesized that free radicals, derived either from molecular oxygen or MAQ-NOH itself (i.e., a compound-centered free radical), are responsible for the damage to membrane lipids and the skeletal proteins that mark the injured red cells for premature removal from the circulation by macrophages in the spleen.

The present studies were conducted to determine whether oxygen- and/or sulfur-centered free radicals were present in rat red cell suspensions exposed to hemolytic concentrations of MAQ-NOH and, thus, have a role in MAQ-NOH-induced hemolytic injury. MAQ-NOH was added to rat red cell suspensions, and the incubates were analyzed by EPR spectroscopy using 2-ethoxycarbonyl-2-methyl-3,4-dihydro-2H-pyrrole-1-oxide (EMPO) as a spin trap. We report that MAQ-NOH was able to generate hydroxyl radicals and ferryl heme species under hemolytic conditions. Furthermore, formation of the hydroxyl radical was constant for 20 min and dependent on the presence of erythrocytic GSH. These data support the concept that oxygen-derived free radicals are involved in the process underlying MAQ-NOH-induced hemolytic damage.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Chemicals. Diethylenetriaminepentaacetic acid (DTPA; metal chelator) and DEM (GSH delpetor) were obtained from Sigma-Aldrich (St. Louis, MO). EMPO (EPR spin trap) was purchased from Oxis Research (Portland OR). MAQ-NOH was synthesized as described previously (Bolchoz et al., 2001). All other reagents were of the best grade commercially available.

Animals. Male Sprague-Dawley rats (75-100 g) were purchased from Harlan Labs (Indianapolis, IN) and were maintained on food and water ad libitum. Animals were acclimated to a 12-h light/dark cycle for 1 week before their use. Blood was collected from the descending aorta of anesthetized rats into heparinized tubes and washed three times in isotonic phosphate-buffered saline (pH 7.4) supplemented with 10 mM D-glucose (PBSG). Following removal of the plasma and buffy coat, the cells were resuspended (40% hematocrit) and used the same day they were collected.

EPR Studies. To determine whether free radicals could be detected in MAQ-NOH-exposed red cell suspensions, spin trapping experiments were performed in incubations containing hemolytic concentrations of MAQ-NOH. Reaction mixtures (2 ml) contained 10 mM EMPO and 0.1 mM DTPA in a red cell suspension (40%) in PBSG at room temperature under aerobic conditions. Experiments that used hemolyzed red cells were conducted by suspending washed red cells in ice-cold deionized water. Reactions were initiated by addition of MAQ-NOH dissolved in DMSO (10 µl). EPR spectra were recorded on a Bruker ELEXYS E-500-10 spectrometer system (Bruker Instruments, Inc., Billerica, MA) operating at 20 mW with a microwave frequency of 9.77 GHz, a receiver gain of 69 dB, a time constant of 0.164 s, a modulation amplitude of 1 G, a modulation frequency of 100 Hz, and a scan time of 41.94 s. For experiments to determine the dependence of radical adduct generation on red cell GSH, red cell suspensions were titrated with diethyl maleate to deplete GSH (by >95%), as described previously (Bolchoz et al., 2002). The DEM-treated red cells were then washed once and resuspended to 40% hematocrit before their use in the EPR experiments.

MAQ-NOH Stability Studies. Stability of MAQ-NOH in buffer and in red cell suspensions was determined using an HPLC assay developed previously to detect MAQ-NOH formation in liver microsomes (Bolchoz et al., 2001).

Spectrophotometric Detection of Hemoglobin Oxidation. Formation of methemoglobin and ferrylhemoglobin in MAQ-NOH-treated red cell suspensions was determined as described previously (Harrison and Jollow, 1987; Bradshaw et al., 1997). The amount of ferrylhemoglobin was measured in red cell incubates pretreated with sodium sulfide (2 mM), which irreversibly converts the unstable ferryl heme species to sulfhemoglobin (Giulivi and Davies, 1990).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Formation of EMPO Radical Adducts in Rat Red Cells. To establish whether free radicals could be detected in rat red cells exposed to MAQ-NOH under previously determined hemolytic conditions (Bolchoz et al., 2001), rat red cell suspensions (40% hematocrit) containing 10 mM EMPO were exposed to a TC₅₀ concentration (median toxic concentration; i.e., the concentration required to produce 50% of the maximal response) of MAQ-NOH (350 µM). As shown in Fig. 1A, the spectrum recorded 3 min after the addition of MAQ-NOH showed a four-line EMPO radical adduct signal pattern. The signal was dependent on the presence of MAQ-NOH (Fig. 1B) and EMPO (Fig. 1C) and was generated equally well in lysed red cells (Fig. 1E). The hyperfine splitting constants (Table 1) were identical to those of an EMPO hydroxyl radical adduct (Olive et al., 2000). As shown in Fig. 2, formation of EMPO-OH was concentration-dependent and occurred over the range of hemolytic concentrations of MAQ-NOH. No evidence for an EMPO-thiyl radical adduct or a compound-centered free radical was observed in these incubates.

View larger version (30K): [in this window] [in a new window]   Fig. 1.   Detection of EMPO-radical adduct signals in rat erythrocyte suspensions exposed to MAQ-NOH. A, EPR spectrum recorded 3 min after the addition of MAQ-NOH (350 µM) in DMSO to a 40% red cell suspension (2 ml) in PBSG containing 10 mM EMPO and 0.1 mM DTPA. B, as in A, except DMSO alone (10 µl) was added. C, as in A, except EMPO was omitted. D, as in A, except red cells were omitted. E, as in A, except lysed red cells were used. Instrument settings: power, 20 mW; microwave frequency, 9.77 GHz; receiver gain, 69 dB; time constant, 0.164 s; modulation amplitude, 1 G; scan time, 41.94 s.

View larger version (22K): [in this window] [in a new window]   Fig. 2.   Effect of concentration of MAQ-NOH on EMPO-OH radical adduct formation in rat erythrocytes. EPR spectrum recorded 3 min after the addition of 150 µM (A), 350 µM (B), 750 µM (C), or 1.5 mM MAQ-NOH (D) in DMSO (10 µl) to a 40% red cell suspension (2 ml) in PBSG containing 10 mM EMPO and 0.1 mM DTPA. Instrument settings: power, 20 mW; microwave frequency, 9.77 GHz; receiver gain, 69 dB; time constant, 0.164 s; modulation amplitude, 1 G; scan time, 41.94 s.

Time Dependence of Hydroxyl Radical Formation. Previous EPR studies using the spin trap 5,5-dimethyl-pyrroline-N-oxide (DMPO) have demonstrated that glutathione thiyl radical adducts are generated in rat red cell suspensions exposed to hemolytic concentrations of the arylhydroxylamine phenylhydroxylamine (Maples et al., 1990; Bradshaw et al., 1995). These experiments also revealed that at higher drug concentrations and longer incubation times, the glutathione thiyl radical adduct signal was replaced by a hemoglobin thiyl radical adduct signal. In the present study, when rat red cells containing EMPO were exposed to MAQ-NOH (350 µM) and the spectrum recorded for 30 min, no change in the shape of the hydroxyl radical signal was detected (Fig. 3). The signal intensity, however, began to decline after 20 min due most likely to decay of the radical generating species.

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Time dependence of MAQ-NOH-induced EMPO-radical adduct formation in rat erythrocytes. MAQ-NOH (350 µM) was added to a 40% red cell suspension in PBSG containing 10 mM EMPO and 0.1 mM DTPA. EPR spectra were recorded 1 min (A), 3 min (B), 6 min (C), 10 min (D), 15 min (E), and 30 min (F) after the addition of MAQ-NOH. Instrument settings: power, 20 mW; microwave frequency, 9.77 GHz; receiver gain, 69 dB; time constant, 0.164 s; modulation amplitude, 1 G; scan time, 41.94 s.

View larger version (25K): [in this window] [in a new window]   Fig. 4.   Effect of time of EMPO addition on the MAQ-NOH-induced EPR signal. EMPO (10 mM) was added to a 40% red cell suspension in PBSG containing 0.1 mM DTPA 1 min (A), 5 min (B), 10 min (C), 20 min (D), 30 min (E) after the addition of MAQ-NOH (350 µM). EPR spectrum recorded 3 min after the addition EMPO. Instrument settings: power, 20 mW; microwave frequency, 9.77 GHz; receiver gain, 69 dB; time constant, 0.164 s; modulation amplitude, 1 G; scan time, 41.94 s.

Stability of MAQ-NOH in Presence and Absence of Red Cells. To determine the relative stability of MAQ-NOH in buffer (PBSG) versus red cells at 37°C, the concentration of MAQ-NOH was measured in aliquots taken as a function of time of incubation by HPLC-electrochemical detection. When incubated in buffer alone, MAQ-NOH was relatively stable (Fig. 5), disappearing with a half-life of over 30 min. In marked contrast, when incubated in the presence of red cells, MAQ-NOH was highly unstable and disappeared with a half-life of about 2 to 3 min (Fig. 5).

View larger version (16K): [in this window] [in a new window]   Fig. 5.   Stability of MAQ-NOH in blood versus buffer. MAQ-NOH (350 µM) was added to red cell suspensions or PBSG alone and allowed to incubate aerobically at 37°C. Aliquots were removed at designated intervals and assayed for MAQ-NOH concentration using HPLC with electrochemical detection. Values for peak height (millimeters) are means ± S.D. (n = 3).

Effect of GSH Depletion on Hydroxyl Radical Formation. Recent work in this laboratory has shown that MAQ-NOH hemolytic activity is markedly enhanced in rat red cells depleted of GSH (Bolchoz et al., 2002). To examine the effect of GSH depletion on hydroxyl radical formation, rat red cells were depleted of GSH (95% depletion) by titration of the suspension with DEM before their exposure to MAQ-NOH. As shown in Fig. 6B, hydroxyl radical adducts were not detected in red cell suspensions that lacked GSH, even when using concentrations of MAQ-NOH as high as 750 µM (Fig. 6D).

View larger version (27K): [in this window] [in a new window]   Fig. 6.   Effect of removal of GSH from rat erythrocytes on the MAQ-NOH-generated EPR signal. A, EPR spectrum recorded 3 min after the addition of 150 µM MAQ-NOH to 40% red cell suspension in PBSG containing 10 mM EMPO and 0.1 mM DTPA. B, as in A, except red cells were treated with DEM to deplete GSH (>95%). C, as in A, except 750 µM MAQ-NOH was added. D, as in C, except red cells were treated with DEM to deplete GSH (>95%). Instrument settings: power, 20 mW; microwave frequency, 9.77 GHz; receiver gain, 69 dB; time constant, 0.164 s; modulation amplitude, 1 G; scan time, 41.94 s.

Ferrylhemoglobin Formation in Rat Red Cells. Ferrylhemoglobin is known to be generated in red cells by the interaction of hemoglobin and H₂O₂ and is considered to be a potent cytotoxic oxidant capable of peroxidation of unsaturated fatty acids (Kanner and Harel, 1985a,b; Galaris et al., 1990). Since hemolytic concentrations of MAQ-NOH induce lipid peroxidation in rat red cells (Bolchoz et al., 2002), it was of interest to determine whether ferrylhemoglobin was formed under the conditions used to detect free radical generation by EPR. Spectrophotometric scans of rat red cells exposed to hemolytic concentrations of MAQ-NOH did not reveal the presence of ferrylhemoglobin, as measured by the increase in absorbance at 545 and 580 nm in samples treated with ferricyanide to remove interference by the methemoglobin peak at 630 nm. When the red cells were pretreated with Na₂S to trap the ferryl heme species as sulfhemoglobin (620 nm) (Berzofsky et al., 1971), however, the presence of ferrylhemoglobin could be demonstrated (Fig. 7A). The fact that ferrylhemoglobin was detected only by conversion to sulfhemoglobin suggests a steady low-level production of this oxidant species during the incubation period (Bradshaw et al., 1997). Production of ferryl heme was dependent on MAQ-NOH concentration (Fig. 7B) and on the duration of incubation (Fig. 7C).

View larger version (16K): [in this window] [in a new window]   Fig. 7.   A, spectrophotometric detection of ferrylhemoglobin in rat erythrocytes exposed to MAQ-NOH (750 µM). Rat erythrocytes (40% suspension) were preincubated with sodium sulfide (2 mM) to trap the ferryl heme species as sulfhemoglobin (max 620 nm). Potassium cyanide (10%, 20 µl) was added to remove any interference due to methemoglobin (max 630 nm). B, concentration dependence of MAQ-NOH-induced ferrylhemoglobin formation. C, time dependence of MAQ-NOH-induced ferrylhemoglobin formation. Values are means ± S.D. (n = 3).

Effect of GSH Depletion on Ferrylhemoglobin and Methemoglobin Formation. Since the enhancement of MAQ-NOH hemolytic activity caused by depletion of cellular GSH was accompanied by loss of the hydroxyl radical signal, it was of interest to determine the effect of GSH depletion on ferrylhemoglobin formation. GSH-normal and -depleted rat red cells were treated with a range of hemolytic concentrations of MAQ-NOH and assayed for ferrylhemoglobin, as described above. As shown in Fig. 8A, depletion of cellular GSH before addition of MAQ-NOH had no effect of the formation of ferryl heme in the red cells.

View larger version (43K): [in this window] [in a new window]   Fig. 8.   Effect of GSH depletion on MAQ-NOH-induced ferrylhemoglobin formation (A) and methemoglobin formation (B). Ferrylhemoglobin and methemoglobin were determined 10 min after addition of the indicated concentrations of MAQ-NOH to the red cell incubates. Data points are means ± S.D. (n = 4).

View larger version (34K): [in this window] [in a new window]   Fig. 9.   Effect of Na2S pretreatment on MAQ-NOH-induced methemoglobin formation. Red cell suspensions were incubated with or without 2 mM sodium sulfide for 5 min before the addition of MAQ-NOH (750 µM). Methemoglobin formation was measured 10 min after addition of MAQ-NOH. Data points are means ± S.D. (n = 3).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In previous studies on primaquine-induced hemolytic anemia, we have demonstrated that its N-hydroxy metabolite MAQ-NOH is directly hemolytic in rats and that this hemotoxicity is associated with lipid peroxidation without significant protein oxidation (Bolchoz et al., 2001, 2002). When GSH is depleted before MAQ-NOH exposure, however, there is a pronounced exacerbation of MAQ-NOH hemolytic activity, and this exacerbation is accompanied by the appearance of protein oxidation without change in lipid peroxidation. The present studies were undertaken to determine whether radical species are produced in rat red cells exposed to hemolytic concentrations of MAQ-NOH, and if so, whether the enhancement of hemolytic activity by GSH depletion is accompanied by alterations in free radical formation.

Experimentally, rat red cells were incubated with MAQ-NOH and analyzed by EPR using EMPO as the free radical spin trap. EMPO was used instead of DMPO because the superoxide, hydroxyl, and thiyl radical adducts are more readily distinguished and because its superoxide radical adduct is more stable and does not spontaneously decay to a hydroxyl radical adduct signal (Olive et al., 2000; Zhang et al., 2000) as is observed with the DMPO-superoxide adduct (Thornalley, 1986; Makino et al., 1990; Hanna et al., 1992). We report that MAQ-NOH generated an EMPO-hydroxyl radical adduct (Fig. 1) that was dependent on the presence of red cells (intact or lysed), EMPO, and MAQ-NOH. The EMPO-hydroxyl radical adduct signal was stable for 20 min (Fig. 3), and its intensity was proportional to MAQ-NOH concentration (Fig. 2). Addition of the EMPO reagent at intervals after exposure of the red cells to MAQ-NOH (Fig. 4) showed the radical concentration to be constant for at least 20 min.

As noted above, depletion of red cell GSH before the addition of MAQ-NOH significantly enhances the hemolytic activity of MAQ-NOH. Paradoxically, the hydroxyl radical was not only not enhanced but also was lost completely (Fig. 6). Furthermore, the increased toxicity was not accompanied by an increase in ferryl heme formation (Fig. 8A) and, by implication, not by increased levels of hydrogen peroxide.

Although a hydroxyl radical has previously been shown to be generated in red cells exposed to other hemolytic N-hydroxylamines, such dapsone hydroxylamine (Bradshaw et al., 1997), the mechanism by which the radical is generated is unclear. It is well known that oxyhemoglobin exists in the red cell as an equilibrium mixture that includes a superoxide-ferriheme component. Dissociation of this complex is believed to be responsible for the constant, low-level oxidant stress that is present in unchallenged red cells. Kiese (1974) proposed an interaction between arylhydroxylamines and oxyhemoglobin that may be regarded as an exacerbation of this normal process. Thus, they described the occurrence of a cyclic oxidation-reduction reaction between arylhydroxylamines and oxyhemoglobin that yields nitrosoarene, methemoglobin, and a reactive oxygen species (such as superoxide or hydrogen peroxide). Subsequent reduction of methemoglobin by NADH-dependent methemoglobin reductase and of the nitrosoarene by a NADPH-dependent diaphorase would support a greater than 1:1 stoichiometry in regard to active oxygen production. Superoxide/hydrogen peroxide would in turn generate hydroxyl radicals by way of an iron-catalyzed Fenton reaction. Although there is controversy over whether hemoglobin is able to catalyze a Fenton reaction (Gutteridge, 1986; Puppo and Halliwell, 1988; Dong Mao et al., 1994), Comporti and colleagues (Ferrali et al., 1992; Ciccoli et al., 1994; Ciccoli et al., 1999) have presented evidence to suggest that the oxidation of hemoglobin to methemoglobin facilitates the release of heme and/or free iron in a diffusible redox-active form.

Although the MAQ-NOH-induced hydroxyl radical of the present studies could be generated by this mechanism, the loss of the hydroxyl radical when GSH was depleted from the red cells strongly suggests that GSH is involved in the process. An alternate hypothesis, proposed by Winterbourn and colleagues (Munday and Winterbourn, 1989; Winterbourn, 1993), is that superoxide formation is dependent on GSH. In this concept, GSH acts as a direct radical scavenger of a diverse range of radical species, accepting the unpaired electron from the radical to form a glutathione thiyl radical intermediate. The thiyl radical electron is then transferred to molecular oxygen to form superoxide anion radical. Superoxide is converted to hydrogen peroxide by superoxide dismutase, creating a GSH/superoxide radical sink. In the present studies, an excessive production of superoxide/hydrogen peroxide in the presence of kinetically free iron derived from methemoglobin could account for the observed hydroxyl radical adduct signal in red cells exposed to hemolytic concentrations of MAQ-NOH.

It is apparent that although the GSH/superoxide radical sink concept explains satisfactorily the dependence of the hydroxyl radical signal on cellular GSH levels, it does not provide an explanation for the initiation of the radical series or how methemoglobin formation is provoked by MAQ-NOH. The relative stability of MAQ-NOH in buffer (t_1/2, >30 min) compared with its instability in the red cells (t_1/2, ca. 2 min) suggests that a direct reaction with molecular oxygen to generate the initial radical species is unlikely and that, in some fashion, hemoglobin or other cellular constituent(s) are needed to initiate radical formation. Similarly, comparison of methemoglobin levels in the presence and absence of Na₂S (Fig. 9) indicates that the comproportionation of ferryl heme is not a major contributor to MAQ-NOH-induced methemoglobin formation and, hence, that methemoglobin must arise by a different pathway.

Although neither postulate alone provides a satisfactory explanation, a combination appears to be compatible with the current data. Thus, the rapid disappearance of MAQ-NOH in red cells (compared with that seen with buffer alone) and the rapid formation of methemoglobin (Bolchoz et al., 2001) suggest that the interaction of MAQ-NOH and hemoglobin could generate methemoglobin, a compound-centered free radical, such as a quinoline hydronitroxide radical and hydrogen peroxide (Fig. 10, reaction 1). The generation of this radical species may be analogous to the formation of the phenylhydronitroxide radical form of phenylhydroxylamine described by Mason and colleagues (Maples et al., 1990). Of interest, delocalization of the radical character into the heterocyclic ring system could stabilize the radical nature of the species and offer an explanation for the continued generation of the hydroxyl radical signal over a 20 min period (Fig. 4) even though the MAQ-NOH has disappeared (Fig. 5). Once formed, the nitroxide free radical could react with GSH to form a thiyl radical (Fig. 10, reaction 2), which in turn could generate a superoxide radical and lead to additional hydrogen peroxide formation via superoxide dismutase (Fridovich, 1986). As noted above, kinetically-free iron derived from methemoglobin could catalyze hydroxyl radical formation.

View larger version (20K): [in this window] [in a new window]   Fig. 10.   Working hypothesis for MAQ-NOH-induced hemolytic injury. See the text for description. GPx, glutathione peroxidase.

Although this combined schema permits rationalization of the available data, it should be noted that there are major discrepancies. First, we have been unable to detect a compound-centered radical derived from MAQ-NOH in the red cell incubates. Of interest, aerobic incubation of MAQ-NOH in PBSG buffer alone (without the iron chelator DTPA) gives rise to a radical species that can be detected by EPR (L. Bolchoz, A. Gelasco, D. Jollow, D. McMillan, unpublished observations). The identity of this species and its relevance to hemotoxicity are currently under investigation.

Second, we have been unable to detect a glutathione thiyl radical in MAQ-NOH incubates. This is of concern since previous analogous studies with phenylhydroxylamine demonstrated the presence of thiyl radicals (Maples et al., 1990; Bradshaw et al., 1995). Failure to detect thiyl radicals in MAQ-NOH red cell incubates raises questions about the intermediacy of this radical species in the proposed reactions and requires further study.

As noted above, there is a lack of correlation between the very rapid disappearance of MAQ-NOH and the apparent steady production of hydroxyl radical for at least 20 min after addition of the MAQ-NOH to the red cells. The fact that hydroxyl radical continues to be generated long after MAQ-NOH has disappeared argues strongly that MAQ-NOH induces the formation of an as yet undetected more stable intermediate. We have not, however, been able to detect this species even in the GSH-depleted/high-MAQ-NOH-concentration experiment in which the loss of the hydroxyl radical leaves a background signal (Fig. 6D). On the other hand, additional support for the existence of such a species has been gathered in an MAQ-NOH-treated, GSH-depleted red cell experiment in which the addition of the lipid soluble sulfhydryl-donating compound cysteamine led to the trapping of a cysteamine thiyl radical by EMPO (Bolchoz et al., unpublished observations). Thus, even in red cells in which the hydroxyl radical is not detectable, a precursor species that is more stable than MAQ-NOH must be present.

The role of hydrogen peroxide in the hemolytic process is also unclear. It is well known that hydrogen peroxide reacts with hemoglobin to form ferryl heme and, hence, that ferryl heme levels in the red cells reflect the generation of hydrogen peroxide in excess of its metabolic clearance (Kanner and Harel, 1985a,b; Harel and Kanner, 1988). Thus, the formation of ferrylhemoglobin in these red cells may be taken as evidence for the formation of excess hydrogen peroxide under MAQ-NOH-induced hemolytic conditions. It has long been thought, however, that glutathione peroxidase plays a major role in the removal of hydrogen peroxide from the red cell (Cohen and Hochstein, 1963). Thus, in the present studies, depletion of GSH from the red cells should have decreased hydrogen peroxide elimination by glutathione peroxidase, resulting in enhanced hydrogen peroxide levels and enhanced ferryl heme formation. Experimentally, depletion of red cell GSH had no effect on the production of ferryl heme (Fig. 8A), implying no significant change in hydrogen peroxide levels. Since ferryl heme appears to play little or no role in MAQ-NOH-induced methemoglobin formation (Fig. 8B), the lack of enhancement in ferryl heme levels is not due to enhanced removal by comproportionation to methemoglobin. The data are consistent with the proposal by Eaton (1991) that catalase plays a more significant role in controlling red cell peroxide levels than has previously been considered.

In summary, the present studies clearly demonstrate that under hemolytic conditions, MAQ-NOH generates three active oxygen species in rat red cells: hydroxyl radical, hydrogen peroxide, and ferryl heme, each of which has the chemical potential to inflict the initial oxidant injury to the red cell that ultimately leads to its premature sequestration by the spleen. Enhancement of the susceptibility of the cells by prior depletion of cellular GSH, however, did not increase the levels of any of these species, and hence, these data alone do not allow us to conclude which, if any, of these oxidants could be causal in MAQ-NOH hemotoxicity. Of particular interest, the marked discrepancy between the rapid disappearance of MAQ-NOH from the red cell incubate and the sustained oxidant stress for at least 20 min suggests the presence of an as yet undetected pro-oxidant species derived from MAQ-NOH under hemolytic conditions.