- Email: [email protected]
Formation of Free Radicals and Protein Mixed Disulfides in Rat Red Cells Exposed to Dapsone Hydroxylamine
Free Radical Biology & Medicine, Vol. 22, No. 7, pp. 1183–1193, 1997 Copyright q 1997 Elsevier Science Inc. Printed in the USA. All rights reserved 0891-5849/97 $17.00 / .00
PII S0891-5849(96)00542-4
Original Contribution FORMATION OF FREE RADICALS AND PROTEIN MIXED DISULFIDES IN RAT RED CELLS EXPOSED TO DAPSONE HYDROXYLAMINE Timothy P. Bradshaw,*1 David C. McMillan,* Rosalie K. Crouch,† and David J. Jollow* Departments of *Cell and Molecular Pharmacology and Experimental Therapeutics, and †Ophthalmology, Medical University of South Carolina, Charleston, SC 29425, USA (Received 9 August 1995; Revised 19 June 1996; Re-revised 23 September 1996; Accepted 30 September 1996)
Abstract—The hemolytic activity of dapsone is well known to reside in its N-hydroxylamine metabolites. Addition of dapsone hydroxylamine (DDS–NOH) to red cell suspensions causes damage such that when reintroduced into the circulation of isologous rats, the injured cells are rapidly removed by the spleen. Hemolytic activity is associated with the extensive formation of disulfide-linked hemoglobin adducts on red cell membrane skeletal proteins. To determine if free radicals could be involved in this process, rat red cells were incubated with DDS–NOH in the presence of the spin trap, 5,5*-dimethyl-1-pyrroline-N-oxide (DMPO) and subjected to EPR analysis. Addition of DDS–NOH (25–50 mM) to a red cell suspension gave rise to a four-line (1:2:2:1) EPR spectrum with coupling constants identical to those of a DMPO-hydroxyl radical adduct (DMPO–OH) standard. No other radicals were detected; however, preincubation of red cells with cysteamine caused the DDS–NOH-generated DMPO–OH signal to be replaced by a cysteamine thiyl radical adduct signal. DDS–NOH-treated red cells were also found to contain ferrylhemoglobin, indicating the presence of hydrogen peroxide. Furthermore, DDS–NOH was found to stimulate salicylate hydroxylation in red cell suspensions, confirming the presence of oxygen radicals. These data support the hypothesis that oxygen radicals are involved in the mechanism underlying dapsone-induced hemolytic anemia. q 1997 Elsevier Science Inc. Keywords—Dapsone, Dapsone hydroxylamine, Erythrocytes, Rat, Hemolytic anemia, Hydroxyl radical, Cysteamine, Ferrylhemoglobin
sone administration in amounts sufficient to account for the toxicity of the parent drug.4 The hemolytic activity of arylamine drugs in humans has long been associated with a drug-induced oxidative stress within the red cell as evidenced by oxidation of cellular reduced glutathione (GSH) to oxidized glutathione (GSSG) and glutathione-protein mixed disulfides (protein-SSG), and by the enhanced sensitivity of individuals deficient in erythrocytic glucose-6-phosphate dehydrogenase activity.3,5,6 Oxidative stress within red cells is thought to result from a cyclic oxidation-reduction reaction that occurs between the arylhydroxylamine metabolite and oxyhemoglobin, yielding the nitrosoarene and methemoglobin, respectively.7 This reaction has been shown to produce hydrogen peroxide,8,9 and thus, is widely considered to generate other active oxygen species (i.e., superoxide anion
INTRODUCTION
Hemolytic anemia and methemoglobinemia are wellknown dose-limiting side effects in the therapeutic use of the arylamine drug dapsone.1–3 Because dapsone is not hemotoxic when incubated with red cells in vitro, it has long been recognized that toxicity is mediated by metabolites rather than the parent compound. We have previously shown that the hemolytic metabolites of dapsone are its N-hydroxy derivatives, dapsone hydroxylamine (DDS–NOH) and monoacetyl dapsone hydroxylamine. These metabolites are equipotent and direct-acting hemotoxicants, and are formed after dapAddress correspondence to: David J. Jollow, Ph.D., Department of Pharmacology, Medical University of South Carolina, 171 Ashley Avenue, Charleston, SC 29425-2251. 1 Present address of Timothy Bradshaw: Oligomer Development Department, Glaxo-Wellcome Research Institute, 5 Moore Drive, Research Triangle Park, NC 27709. 1183
/ 2b29 2433 Mp 1183 Thursday Mar 13 10:16 PM EL–FRB (vol.27#6 ’97) 2433
1184
T. P. BRADSHAW et al.
and hydroxyl radicals), which are capable of causing cellular injury. In studies on the effects of DDS–NOH in rat red cells, we have observed a rapid loss of GSH with concomitant formation of protein-SSG, disulfide-linked hemoglobin polymers, and disulfide-linked hemoglobin adducts on certain membrane skeletal proteins.10 We have hypothesized that oxygen-centered free radicals induce formation of both glutathione and hemoglobin thiyl radicals, and that the latter react with skeletal proteins of the cell membrane to form hemoglobin-skeletal protein adducts. This ‘‘damage’’ to critical membrane skeletal proteins is considered to initiate premature removal of the injured red cells by the spleen.6 An alternate hypothesis is that ‘‘compoundcentered’’ (e.g., arylhydronitroxide) free radicals are responsible for thiyl radical generation and subsequent hemoglobin-skeletal protein adduct formation.11–14 The present studies were undertaken to determine whether oxygen and/or compound-centered free radicals could be detected in rat red cells exposed to hemolytic concentrations of DDS–NOH and, thus, have a role in the mechanism underlying DDS–NOH-induced hemolytic injury. DDS–NOH was added to red cell suspensions in the presence and absence of a spin trap, and the incubates analyzed by EPR spectroscopy. The data indicate that in vitro treatment of rat red cells with concentrations of DDS–NOH at and below those necessary to induce a hemolytic response results in the generation of active oxygen species, including superoxide, hydrogen peroxide, hydroxyl radical, and ferryl heme species. Moreover, free radicals generated within red cells were able to transform the exogenous thiolcontaining compound, cysteamine, to a thiyl radical adduct. Addition of cysteamine to red cells also suppressed protein–SSG formation with concomitant production of cysteamine-protein mixed disulfides (protein–SSCysNH2). No evidence was observed that indicated the presence of a dapsone hydronitroxide free radical under hemolytic conditions. These data support the hypothesis that oxygen radicals, formed in the red cell as a consequence of DDS–NOH redox cycling, generate hemoglobin thiyl radicals, which in turn, attack membrane skeletal proteins.
MATERIALS AND METHODS
Chemicals Diethylenetriaminepentaacetic acid (DTPA), GSH, cysteamine, sodium salicylate, 2,3-dihydroxybenzoic acid, and 3,5-dihydroxybenzoic acid were obtained from Sigma Chemical Co. (St. Louis, MO). DMPO was purchased from Aldrich Chemical Co. (Milwaukee,
WI). DDS–NOH was synthesized as described previously.4 All other chemicals and reagents were of the best commercially available grade. Animals Male Sprague–Dawley rats (130–150 g) were purchased from Camm Research, Inc. (Wayne, NJ), and maintained on food and water ad lib. Animals were acclimated for 1 week to a 12-h light-dark cycle prior to their use. Red cells were collected from the descending aorta of ether-anesthetized rats into heparinized tubes, washed three times in isotonic phosphate-buffered saline (pH 7.4) supplemented with 10 mM D-glucose (PBSG), and resuspended in PBSG prior to their use. EPR studies Incubation mixtures (1 ml) contained 100 mM DMPO, 0.1 mM DTPA and red cells (5–20% suspension) in PBSG at 47C under aerobic conditions. Following a 5-min preincubation on ice, reactions were initiated by the addition of DDS–NOH dissolved in acetone (5 ml). EPR spectra were then recorded at ambient temperature on a Varian E-4 spectrometer operating at 10 mW power with a microwave frequency of 9.45 GHz and a field set of 3380 G as described previously.15 Additional spectra were recorded under the conditions described above following a 15-min preincubation at 377C in the presence and absence of 20 mM cysteamine or GSH. DMPO–adduct standards for hydroxyl radical and GSH thiyl radical were generated photolytically as described previously.16,17 HPLC analysis of salicylate hydroxylation in red cells Hydroxyl radical formation was detected with the use of salicylate as a trapping agent.18 Rat red cells (1.0 ml, 40% suspension) were preincubated with sodium salicylate (2.5 mM) for 5 min prior to the addition of DDS–NOH (180 mM). After a 30-min incubation at 377C, a 200 ml aliquot of the cell suspension was hemolyzed by a single freeze-thaw cycle. The hemolysate was then centrifuged (13,000 1 g for 3 min), and an aliquot of the supernatent (20 ml) was injected onto the HPLC. Products of hydroxyl radical attack on salicylate (2,3- and 2,5-dihydroxybenzoic acid) were separated on a Waters reverse-phase C18 column (Resolve) cartridge. The dihydroxybenzoic acids were eluted with an aqueous mobile phase consisting of monochloroacetic acid (75 mM), octanesulfonic acid (1.5 mM), EDTA (0.7 mM), and acetonitrile (5%) at a flow rate of 1.3
/ 2b29 2433 Mp 1184 Thursday Mar 13 10:16 PM EL–FRB (vol.27#6 ’97) 2433
Dapsone hydroxylamine in rat red cells
1185
ml/min.19 The dihydroxybenzoic acids were detected using a BAS LC-4 amperometric detector (Bioanalytical Systems, West Lafayette, IN) controlled at 5 nA full scale with a filter of 0.1 Hz in the reduction mode. A dual glassy-carbon electrode was operated in series; the upstream (generator) electrode was set at /750 mV, and the downstream electrode (detector) was set at /200 mV vs. Ag/AgCl. HPLC analysis of red cell sulfhydryls Following a 15-min preincubation in the presence and absence of cysteamine, DDS–NOH dissolved in acetone (10 ml) was added slowly to red cell suspension (2 ml, 40% suspension) and allowed to incubate in PBSG for 30 min at 377C. At the end of the incubation period, aliquots of the suspension (200 ml) were lysed and analyzed for protein–SSG and cysteamine-protein mixed disulfide (protein–SCysNH2) content by HPLCEC as described previously.10 The concentration of GSH or cysteamine released from protein following sodium borohydride-catalyzed reduction of disulfide bonds was determined from standard curves generated using known amounts of GSH and cysteamine treated identically to the experimental samples. Spectrophotometric detection of ferrylhemoglobin Rat red cells (1 ml, 40% suspension) were exposed to various concentrations of DDS–NOH. Aliquots (75 ml) were removed after designated intervals, lysed in 5 ml of hemolysis buffer (5 mM Na2HPO4, 0.5% Triton X-100, pH 7.5), and analyzed in a Shimadzu UV-160A double-beam UV-visible recording spectrophotometer. The amount of ferrylhemoglobin was measured in red cell incubations treated with sodium sulfide (2 mM), which converts ferrylhemoglobin to sulfhemoglobin (l max 620 nm).20 RESULTS
Formation of DMPO radical adducts in red cells Previous studies in our laboratory have shown that when rat 51Cr-labeled red cells are incubated in PBSG containing DDS–NOH for 2 h at 377C and then administered intravenously to isologous rats, the damaged red cells are removed rapidly from the circulation by the spleen. The hemolytic response is concentration dependent, with an EC50 of approximately 150 mM.4 To determine whether free radicals could be detected under these experimental conditions, rat red cells (5% suspension) were preincubated with 100 mM DMPO and then exposed to DDS–NOH. As shown in Fig. 1A,
Fig. 1. Detection of DMPO-radical adduct signals in rat erythrocytes exposed to DDS–NOH. (A) EPR spectrum recorded 5 min after addition of DDS–NOH (25 mM) to a 5% red cell suspension in PBSG containing 100 mM DMPO and 0.1 mM DTPA. (B) As in A, except DMPO was omitted. (C) As in A, except DDS–NOH was omitted. (D) As in A, except red cells were omitted. Receiver gain, 32,000; modulation amplitude, 2.5 G; time constant, 3.0.
addition of DDS–NOH (25 mM) to red cells generated a four-line (1:2:2:1) EPR signal. The appearance of the radical adduct signal was dependent on the presence of DMPO (Fig. 1B) and DDS–NOH (Fig. 1C). When red cells were excluded from the incubation mixture, a weak four-line signal was detected (Fig. 1D). At concentrations of DDS–NOH greater than 75 mM, the radical adduct signal deteriorated markedly and was not replaced by any new signal, even at concentrations of DDS–NOH as high as 200 mM (data not shown). In addition, no radical adduct signals were detected in lysed red cell incubates, which indicated that generation of this radical adduct is dependent on the the intact red cell. Based on the superficial structure of the signal, the EPR spectrum obtained from DDS–NOH-treated red cells (Fig. 1A) could be assigned as either a hydroxyl or thiyl free radical adduct signal.21 To discriminate between these two possibilities, authentic hydroxyl radical–DMPO (DMPO–OH) and glutathione thiyl radical-DMPO (DMPO–SG) adducts were prepared and
/ 2b29 2433 Mp 1185 Thursday Mar 13 10:16 PM EL–FRB (vol.27#6 ’97) 2433
1186
T. P. BRADSHAW et al.
their hyperfine splitting constants compared against those of the experimental sample. As shown in Table 1, the nitrogen (aN) and b-hydrogen (aH) coupling constants of the experimental sample and the DMPO-OH adduct standard were identical.
Hydroxylation of salicylate in red cells Although the formation of DMPO–OH has been cited frequently as evidence for the existence of hydroxyl radicals in a variety of in vitro systems, this radical adduct could have been formed as consequence of superoxide radical attack on the spin trap followed by reduction of the spin adduct to DMPO–OH.22 Thus, we utilized an alternative method for hydroxyl radical detection that exploits the ability of hydroxyl radicals to attack salicylate with subsequent formation of hydroxylated products that can be measured by HPLCEC.19 As shown in Fig. 2, addition of DDS–NOH (180 mM) to red cells containing salicylate (2.5 mM) resulted in the the formation of 2,3- and 2,5-dihydroxybenzoic acid (Fig. 2D). No peaks corresponding to 2,3- and 2,5dihydroxybenzoic acid were observed in red cells that contained DDS–NOH (Fig. 2B) or salicylate (Fig. 2C) alone.
Effect of cysteamine on radical adduct formation Rat red cells have been shown to lose GSH rapidly and extensively upon exposure to hemolytic concentrations of DDS–NOH, and this loss of GSH is associated with the formation of mixed disulfides between glutathione and the soluble protein of the red cell.10 Although glutathione thiyl radicals were not detected by EPR in DDS–NOH-treated red cells, their role as intermediates in the oxidation of GSH to GSSG and mixed disulfides has been well established.12 Given that GSH is lost rapidly from the red cell upon exposure to DDS–NOH, we examined the possibility that a thiyl radical signal could be observed in red cell suspensions
Table 1. Hyperfine Splitting Constants of DMPO Adducts Spin Adduct DMPO–OH standarda DMPO–SG standardb Red cell DMPO adductb
aN (G)
aH (G)
Reference
14.9 15.4 14.9
14.9 15.8 14.9
21 17
a Radical adduct standards formed by UV photolysis of hydrogen peroxide (DMPO–OH) and GSSG (DMPO–SG). The splitting constants for the adduct standards were determined experimentally. b Splitting constants determined from the EPR spectrum shown in Fig. 1A. Precise measurements were made from the original spectrum.
Fig. 2. Detection of salicylate hydroxylation in rat erythrocytes exposed to DDS–NOH. HPLC-EC elution profile of (A) 2,5- and 2,3-dihydroxybenzoic acid (DHBA) standards; (B) an erythrocyte suspension containing DDS–NOH (180 mM); (C) an erythrocyte suspension containing salicylate (2.5 mM); and (D) an erythrocyte suspension containing both DDS–NOH and salicylate. Erythrocytes (40% suspension) were preincubated for 5 min with salicylate prior to addition of DDS–NOH. The cells were then incubated for 30 min at 377C, and analyzed by HPLC-EC as described in Materials and Methods.
containing a high concentration of an exogenous thiolcontaining compound. Rat red cell suspensions were preincubated in the presence of the lipophilic thiol, cysteamine (20 mM), for 15 min prior to the addition of 50 mM DDS–NOH.
/ 2b29 2433 Mp 1186 Thursday Mar 13 10:16 PM EL–FRB (vol.27#6 ’97) 2433
Dapsone hydroxylamine in rat red cells
As shown in Fig. 3B, the presence of cysteamine resulted in replacement of the DMPO–OH control signal (Fig. 3A) with a stronger signal with different structure and coupling constants (Table 2). This new spectrum was generally similar in appearance to that of the photolytically generated cysteamine–DMPO radical adduct (DMPO–SCysNH2; Fig. 3C). However, the observed nitrogen coupling constants (aN) were not identical (Table 2). This reason for this difference is not clear; however, a mixture of DMPO radical adducts (i.e., DMPO–OH plus DMPO–SCysNH2) in the experimental sample could account for this difference. The possibility that the free radical species respon-
1187
Table 2. Hyperfine Splitting Constants of DMPO Adducts Spin Adduct DMPO–SG standarda DMPO–SCysNH2 standardb Red cell DMPO adductc
aN (G)
aH (G)
Reference
15.4 15.3 15.1
15.8 16.3 16.3
17
a Radical adduct standard formed by UV photolysis of GSSG (DMPO–SG). The splitting constants for the adduct standard were determined experimentally. b Radical adduct standard formed by UV photolysis of cysteamine (DMPO–SCysNH2). Splitting constants determined from the EPR spectrum shown in Fig. 2C. Precise measurements were made from the original spectrum. c Splitting constants of red cell DMPO adduct determined from the EPR spectrum shown in Fig. 2B. Precise measurements were made from the original spectrum.
sible for generating DMPO–SCysNH2 was capable of diffusing out of the red cell was examined by preincubating red cells with GSH, which cannot penetrate the plasma membrane, before addition of DDS–NOH. When GSH (20 mM) was substituted for cysteamine in DDS–NOH-treated red cell suspensions, no change in the DMPO–OH EPR spectrum was observed (data not shown), which indicated that the attacking free radical species was formed and remained within the intact red cell.
Effect of cysteamine on mixed disulfide formation
Fig. 3. Effect of addition of cysteamine to rat erythrocytes on the EPR signal induced by DDS–NOH. (A) EPR spectrum recorded 8 min after addition of DDS–NOH (50 mM ) to a 5% red cell suspension in PBSG containing 100 mM DMPO and 0.1 mM DTPA. (B) As in A, except cysteamine (20 mM) was added 15 min before DDS– NOH. (C) Photolytically generated DMPO–cysteamine adduct standard. Receiver gain, 32,000 (A), 3200 (B and C); modulation amplitude, 2.5 G; time constant, 3.0 (A), 1.0 (B and C).
To determine if the cysteamine-thiyl radicals generated on addition of cysteamine to DDS–NOH-treated rat red cells could suppress the formation of glutathione-protein mixed disulfides, red cells were pretreated with cysteamine (20 mM) 30 min prior to addition of various concentrations of DDS–NOH. After a 30-min incubation in the presence of DDS–NOH, an aliquot of the incubation mixture was removed for quantitation of protein–SSG and protein–SCysNH2 formation. Our laboratory has previously developed a highly sensitive HPLC-EC assay for determination of GSH, GSSG, and protein–SSG content in red cell suspensions.10 To demonstrate that this assay could also be used to quantitate protein–SCysNH2 formation, cysteamine, and GSH were added to red cell lysates, and the samples were then processed and analyzed by HPLC-EC. Figure 4A shows a typical HPLC elution profile of a sample from a red cell lysate treated with cysteamine and GSH. The data indicated that both thiols could be measured using the same electrochemical detector conditions, and were sufficiently well separated to permit quantitation of both thiols in the same experimental samples. A standard curve generated using known amounts of GSH and cysteamine (Fig. 4B) was used to quantitate mixed disulfide content in ex-
/ 2b29 2433 Mp 1187 Thursday Mar 13 10:16 PM EL–FRB (vol.27#6 ’97) 2433
1188
T. P. BRADSHAW et al.
Fig. 4. (A) HPLC-EC elution profile of GSH and cysteamine (CysNH2) standards. (B) Standard curve generated by addition of known amounts of GSH (j) and CysNH2 (l) to PBSG. Standards were treated in the same manner as the red cell samples.
perimental samples. Recovery of both GSH and cysteamine from incubation mixtures was ú95%. In agreement with previous studies,10 addition of DDS–NOH to rat red cells induced a concentrationdependent increase in protein–SSG formation (Fig. 5). Inclusion of cysteamine (20 mM) in the incubation mixture resulted in greater than 80% suppression of protein–SSG formation. This response was accompanied by a concentration-dependent formation of protein–SCysNH2 (Fig. 6). Collectively, these data indicate that cysteamine can substitute for GSH as a nucleophilic reactant in red cells, and support the hypothesis that thiyl radicals originate via oxygen-centered radical attack on cellular thiol groups.
Ferrylhemoglobin formation in red cells Hydrogen peroxide generated within red cells can undergo a variety of reactions, one of which is with oxyhemoglobin to form ferryl heme species ([HbFeIV – OH]30).20 Because ferryl heme species are potent oxidants capable of contributing to intracellular damage,12,23,24 it was of interest to determine whether ferrylhemoglobin is formed in rat red cells under hemolytic conditions. Addition of DDS–NOH to rat red cells induced a concentration-dependent increase in the absorbance at 630 nm, consistent with the formation of methemoglobin (HbFe3/). Addition of sodium cya-
nide to the cuvette to form cyanomethemoglobin abolished the peak at 630 nm, confirming its identity as methemoglobin.25 However, removal of the methemoglobin peak did not reveal a spectrum indicative of the presence of ferrylheme species (i.e., increase in absorbance at 545 and 580 nm).20 To examine the possibility ferrylhemoglobin was formed as an intermediate in DDS–NOH-treated rat red cells, sodium sulfide (2 mM) was added to the incubates before addition of DDS–NOH. Sodium sulfide has been used to trap the ferryl oxidation state of myoglobin23 and hemoglobin20 in cellular systems by converting it irreversibly to sulfhemoglobin.26 Addition of hemolytic concentrations of DDS–NOH to red cell incubates containing Na2S resulted in the appearance of a new peak at 620 nm, consistent with the formation of sulfhemoglobin.26 Addition of cyanide to the cuvette had no effect on the appearance of the spectrum. In Na2S-pretreated red cells ferrylhemoglobin was observed to accumulate with increasing incubation time, and this accumulation was concentration dependent (Fig. 7 inset), with an EC50 of about 30 mM DDS– NOH. Under the experimental conditions described above, the magnitude of the peak height at 620 nm represented the amount of ferrylhemoglobin that had accumulated (as sulfhemoglobin) in the incubation mixtures as a function of incubation time. To determine whether ferryl hemoglobin was accumulating during the incuba-
/ 2b29 2433 Mp 1188 Thursday Mar 13 10:16 PM EL–FRB (vol.27#6 ’97) 2433
Dapsone hydroxylamine in rat red cells
Fig. 5. Effect of pretreatment with cysteamine on DDS–NOH-induced protein–SSG formation in rat erythrocytes. Red cell suspensions were incubated in the presence (h) and absence (j) of cysteamine (20 mM) for 15 min at 377C prior to addition of various concentrations of DDS–NOH. After 30 min of incubation at 377C, aliquots of the red cells were removed and analyzed for protein– SSG by HPLC-EC. Values are means of duplicate determinations.
tion or was present at a steady-state level, Na2S was added to the red cell incubates 0, 5, 10, and 20 min after the addition of DDS–NOH (100 mM). Incubation was continued for a further 1, 5, or 20 min, after which samples were removed for spectral analysis. The peak height at 620 nm was proportional to the time of incubation post Na2S and independent of the duration of incubation prior to Na2S addition. Calculation of DO.D./min for the various Na2S accumulation times indicated a steady-state production of ferrylhemoglobin in these DDS–NOH-treated rat red cells equivalent to a DO.D. of ca. 0.025–0.03/min. The data indicate that ferrylhemoglobin does not accumulate to measurable levels under these conditions.
1189
between glutathione and cellular protein, and by the formation of disulfide-linked adducts between hemoglobin and certain membrane skeletal proteins.10 We have hypothesized that the latter reaction induces a change on the internal surface of the red cell that generates an antigenic site on the external cell surface, which in turn, provokes the immune system to remove the damaged red cell from the circulation by macrophages in the spleen.6 We have also reported that glutathione and hemoglobin thiyl radicals are generated in rat red cells incubated with hemolytic concentrations of phenylhydroxylamine.15 This observation suggested that mixed disulfides arise via thiyl radical attack on free sulfhydryl groups of cellular proteins; however, the origin of the thiyl radicals in phenylhydroxylamine-treated red cells has remained unclear. One possibility is that oxygen-centered free radicals are responsible for the generation of thiyl radicals. This viewpoint is supported by the extensive studies of Kiese and colleagues,7 which showed that arylhydroxylamines enter red cells and undergo a coupled oxidation reaction with oxyhemoglobin (Fe2/), forming the nitrosoarene and methemoglobin (Fe3/), respectively. The nitrosoarene is reduced by an NADPH-dependent reductase to the arylhydroxylamine, which is available to oxidize another equivalent of oxyhemoglobin. In this fashion, many equivalents of methemoglobin are produced per equivalent of arylhydroxylamine. Although the precise chemistry underlying this enzymatic cycle remains to be fully elucidated, molecular oxygen is thought to act as the
DISCUSSION
In previous studies on the mechanism underlying the hemolytic anemia induced by aniline and related arylamine drugs such as dapsone and primiquine, we have utilized an in vitro incubation/in vivo response assay that has allowed us to identify biochemical events within red cells that correlate with the toxic response.6 Using this experimental approach we have shown that rat red cells exposed in vitro to phenylhydroxylamine and DDS–NOH rapidly lose GSH in a concentrationand time-dependent manner. This loss of GSH is accompanied by the formation of disulfide-linked adducts
Fig. 6. Effect of pretreatment with cysteamine on DDS–NOH-induced mixed disulfide formation in rat erythrocytes. Red cell suspensions were incubated in the presence of cysteamine (20 mM) for 15 min at 377C prior to addition of various concentrations of DDS– NOH. After 30 min of incubation at 377C, aliquots of the red cells were removed and analyzed for protein–SSG (m) and protein-cysteamine mixed disulfides (Prot-S-S-CysNH2, j) by HPLC-EC. Values are means of duplicate determinations.
/ 2b29 2433 Mp 1189 Thursday Mar 13 10:16 PM EL–FRB (vol.27#6 ’97) 2433
1190
T. P. BRADSHAW et al.
Fig. 7. Spectrophotometric detection of ferrylhemoglobin in rat erythrocytes exposed to DDS–NOH. Erythrocytes (40% suspension) were exposed to the indicated concentrations of DDS–NOH (0–200 mM) for 10 min at 377C. Sodium sulfide (2 mM) was added prior to DDS– NOH to convert the ferryl species to the sulfheme derivative (lmax 620 nm). Inset: Absorbance was measured at 620 nm 20 min after addition of various concentrations of DDS–NOH to red cell incubates containing 2 mM sodium sulfide.
acceptor of electrons for the oxidation steps of this reaction, resulting in its reduction to superoxide anion radical followed by dismutation to hydrogen peroxide.27–29 Under normal conditions, superoxide and hydrogen peroxide are removed by superoxide dismutase and catalase/glutathione peroxidase, respectively. However, these protective mechanisms are likely to be overwhelmed by the continuous and excessive formation of active oxygen species produced by arylhydroxylamine redox cycling. Superoxide and hydrogen peroxide would then be available to interact with iron (either free or heme iron) to generate highly reactive intermediates, including hydroxyl radical and ferryl heme species, which in turn, could react with cellular sulfhydryl groups to generate thiyl radicals.12,30,31 The present studies report that incubation of rat red cells with hemotoxic concentrations of DDS–NOH resulted in the detection of hydroxyl radical adducts of DMPO in rat red cells. The DMPO–OH radical adduct signal was dependent on the presence of DDS–NOH, red cells, and the spin trap (Fig. 1), and it had EPR spin constants that were identical to those of a photolytically generated DMPO–OH radical adduct standard (Table 1). That free hydroxyl radicals exist in the red cell under hemolytic conditions was further supported by the observation that salicylic acid was converted to 2,3-
and 2,5-dihydroxybenzoic acid under these incubation conditions (Fig. 2). Together, these data provide strong support for the postulate that hydroxyl radicals are formed in the red cell and, hence, could be responsible for the development of hemolytic injury. If hydroxyl radicals were responsible for the production of DMPO–OH, then a Fenton-type reaction would have to be invoked to explain this observation. However, a variety of studies using purified hemoglobin preparations and hydrogen peroxide have suggested that hemoglobin (or heme) per se is unable to catalyze a Fenton reaction.32,33 In these cell-free systems, other reactions are considered to predominate, in particular, the reaction of hydrogen peroxide with hemoglobin and methemoglobin to form ferryl heme species.34,35 On the other hand, the nature of redox reactions within the intact red cell remain poorly understood, and it seems inappropriate at this time to discard the possibility that a Fenton-type reaction can occur within the red cell. Of importance, it has long been known that oxidative damage to hemoglobin results in the release of heme from methemoglobin,32,36 as well as the release of iron in a diffusable and redox active form.37–39 Although the amount of free iron produced in DDS–NOH-treated red cells is unknown and warrants further investigation, the extent of hemoglobin oxidation under hemolytic conditions is considerable (ú40% methemoglobin), and, hence, it is plausible that significant amounts of iron in a form capable of catalyzing a Fenton reaction are available under hemolytic conditions. A second possibility is that DMPO–OH arose from active oxygen species other than hydroxyl radical. For example, superoxide anion radical is known to react with DMPO to form the superoxide radical adduct, DMPO– OOH, which subsequently could have been reduced to DMPO–OH by cellular reducing agents.40 Because DMPO–OOH was not detected by EPR in these studies, this possibility could not be confirmed. However, superoxide is known to react extremely slowly with nitroxide spin traps (10 M01s01) compared to its spontaneous and SOD-catalyzed dismutation.41 In contrast, the reaction between hydroxyl radicals and DMPO is extremely rapid (2 1 109 M01s01).42 In the present studies, the DMPO– OH radical adduct signal could be observed immediately after addition of DDS–NOH-treated cells to the EPR flat cell (4-min scan time). That some part of the observed DMPO–OH radical adduct arose via this mechanism cannot be eliminated by the present data, though in view of the production of 2,3- and 2,5-dihydroxybenzoate from salicylate under these experimental conditions (Fig. 2), it seems unlikely that this mechanism is a major contributor. A third possibility is that ferryl heme species were responsible for generating DMPO–OH, and could the-
/ 2b29 2433 Mp 1190 Thursday Mar 13 10:16 PM EL–FRB (vol.27#6 ’97) 2433
Dapsone hydroxylamine in rat red cells
oretically have been responsible for the hydroxylation of salicylate. Indeed, ferrylhemoglobin is known to be a strong oxidant analogous to other peroxidases and catalase complex I,12,43 and the present data clearly document that ferrylhemoglobin is present in rat red cells under DDS–NOH hemolytic conditions (Fig. 7). However, the characteristic radical adduct signal for 5,5dimethylpyrrolidone-2(2)-oxyl-(1), which would have indicated a reaction between ferryl heme species and the spin trap,44 was not observed. Furthermore, ferrylhemoglobin could not be detected spectrophotometrically and was recognized only by trapping it as the stable sulfheme derivative. This observation, along with studies on the time dependence of ferrylhemoglobin formation, suggested a low steady-state level of this species. The low steady-state level of ferrylhemoglobin in DDS–NOH-treated red cells is consistent with earlier studies,20 which showed that when oxyhemoglobin concentration is in excess, ferrylhemoglobin is reduced rapidly by oxyhemoglobin to methemoglobin. To what extent ferrylhemoglobin reduction contributes to DDS– NOH-induced methemoglobin formation remains to be determined; however, the concentration-response curves for DDS–NOH-induced methemoglobin formation6 and ferryl heme formation (Fig. 7, inset) are similar, raising the possibility that ferrylhemoglobin may contribute to methemoglobin levels at hemolytic concentrations of DDS–NOH. Whether or not ferrylhemoglobin has a role in the DDS–NOH-dependent hemolytic process is also unclear. While steric hinderance considerations suggest that a direct role in the generation of skeletal-protein thiyl radicals is unlikely, the possibility of an indirect role cannot be excluded at this time. The presence of ferrylhemoglobin does, however, clearly indicate that hydrogen peroxide and, by inference, superoxide anion radical, are generated in DDS–NOH-treated red cells. Collectively, the data in the present studies demonstrate the complexity of the arylhydroxylamine redox cycle in the intact red cell, and indicate that further studies will be necessary to determine the individual contributions of these reactive intermediates to the hemolytic response. Although the present studies clearly support a role for active oxygen species in dapsone-induced hemolytic anemia, the link between active oxygen species generation and thiyl radical formation remains to be firmly established. On the one hand, a DMPO–SG radical adduct signal was not detected in DDS–NOHtreated red cells. The reason for this is not clear, but may be a reflection of the rapid loss of GSH (ca. 85– 95%) that occurs upon addition of hemolytic concentrations of DDS–NOH to rat red cells.10,45 Because DMPO–SG is unstable and must be continuously gen-
1191
erated to be detected by EPR,17 the lack of a strong DMPO–SG signal that could compete with that of DMPO–OH presumably is a consequence of the extensive depletion of red cell GSH by DDS–NOH. Because glutathione thiyl and hydroxyl radical adducts of DMPO give rise to similar four-line (1:2:2:1) EPR signals,21 a minor DMPO–SG spectrum could have been obscured by the DMPO–OH signal. On the other hand, addition of cysteamine to replenish the level of nonprotein sulfhydryl groups in the red cell did result in conversion of the EPR spectrum from a hydroxyl radical adduct signal to that of a cysteamine thiyl radical adduct signal (Fig. 3). Furthermore, cysteamine suppressed protein-SSG formation (Fig. 5) and, like intracellular GSH, became bound to cellular protein via disulfide linkages in a concentration-dependent manner (Fig. 6). In contrast, when GSH, which cannot readily penetrate the cell membrane, was added to the media in place of cysteamine, the DMPO–OH signal was not quenched, indicating that the attacking radical species is formed within the red cell and does not diffuse across the cell membrane. A fourth possibility is that glutathione, hemoglobin, and cysteamine thiyl radicals arise via reaction of their sulfhydryl groups with compound-centered (i.e., dapsone hydronitroxide) free radicals. In support of this viewpoint, Maples et al.13 have shown by EPR that phenylhydroxylamine can undergo a one-electron oxidation to a phenylhydronitroxide radical, and that this radical is capable of generating glutathione thiyl radicals upon addition of GSH and DMPO to the system. However, the toxicological relevance of this data is uncertain because it was obtained in phosphate buffer; i.e., in the absence of hemoglobin or of intact red cells, and at a concentration of phenylhydroxylamine (1 mM), which is about an order of magnitude greater than the Cmax for methemoglobin formation in the intact rat red cell.46 When examined at hemotoxic concentrations (50–100 mM), DDS–NOH did not give rise to an EPR spectrum indicative of the presence of dapsone hydronitroxide in the rat red cell. It is not yet known whether the failure to detect the hydronitroxide radical reflects a relative insensitivity of the EPR spectroscopy or whether it indicates an absence of significant levels of this species in the cell. On chemical grounds, it is likely that the hydronitroxide radical is an intermediate during the two-electron oxidation of arylhydroxylamines to nitrosoarenes in the red cell, and thus could be involved in the redox cycling of arylhydroxylamines thought to occur within or near the heme pocket.7 The degree to which such radicals, if formed, might escape the redox cycle into the cytosol to react with critical intracellular targets, and hence, participate in the hemolytic process, is un-
/ 2b29 2433 Mp 1191 Thursday Mar 13 10:16 PM EL–FRB (vol.27#6 ’97) 2433
1192
T. P. BRADSHAW et al.
known. It is noteworthy that while exogenous cysteamine quenched the DMPO–OH signal, exogenous GSH did not. Because dapsone hydronitroxide radicals may be expected to have sufficient lipophilicity and stability to penetrate the cell membrane and leave the cell, the inability of extracellular GSH to suppress the DMPO–OH signal implies that the arylhydronitroxide is not the species that attacks thiol groups and generates thiyl radicals. Based on the current data, we have no evidence that a compound-centered free radical species, such as dapsone hydronitroxide, is generated in rat red cells under hemotoxic conditions or that it plays a direct role in DDS–NOH hemotoxicity. On the other hand, in view of its likely formation during redox cycling, its probable high turnover, and possible lack of sensitivity to EPR detection, it is premature to rule out its participation in the hemolytic process. In summary, the present studies indicate that a number of potentially toxic active oxygen species are generated in rat red cells exposed to hemolytic concentrations of DDS–NOH. Their formation under hemolytic conditions is compatible with one or more of these species having a causal role in the hemolytic process. While the precise roles(s) of active oxygen species in red cell injury is not yet understood and warrants further investigation, the present data support the hypothesis that red cell damage arises from the interaction of cellular thiols with oxygen radicals generated during the redox cycling of arylhydroxylamines with oxyhemoglobin, and that subsequent binding of hemoglobin with skeletal protein via disulfide linkages could provide the trigger for premature sequestration of these oxidatively damaged cells by the spleen.6 Acknowledgements — This study was supported by NIH Grant HL30038. The authors wish to thank Jennifer Schulte and Leslie Edwards for their excellent technical assistance in the preparation of this manuscript.
REFERENCES 1. Degowin, R. L.; Eppes, R. B.; Powell, R. D.; Carson, P. E. The haemolytic effects of diaphenylsulfone (DDS) in normal subjects and in those with glucose-6-phosphate-dehydrogenase deficiency. Bull. World Health Organ. 35:165–179; 1966. 2. Mayer, K.; Ley, A. B. Hemolysis of red cells due to sulfone. Ann. Intern. Med. 72:711–714; 1970. 3. Beutler, E. Hemolytic anemia in disorders of red cell metabolism. New York: Plenum Medical; 1978. 4. Grossman, S. J.; Jollow, D. J. Role of dapsone hydroxylamine in dapsone-induced hemolytic anemia. J. Pharmacol. Exp. Ther. 244:118–125; 1988. 5. Kosower, N. S.; Zipser, Y.; Faltin, Z. Membrane thiol-disulfide status in glucose-6-phosphate dehydrogenase deficient red cells. Relationship to cellular glutathione. Biochim. Biophys. Acta 691:345–352; 1982.
6. Jollow, D. J.; Bradshaw, T. P.; McMillan, D. C. Dapsone-induced hemolytic anemia. Drug Metab. Rev. 27:107–124; 1995. 7. Kiese, M. Methemoglobinemia: A comprehensive treatise. Cleveland, OH: CRC Press; 1974. 8. Rostorfer, H. H.; Cormier, M. J. Formation of ‘‘hydrogen peroxide’’ in the reaction of oxyhemoglobin with methemoglobin forming agents. Arch. Biochem. Biophys. 71:235–249; 1957. 9. Ellederova´, D.; Wagner, J.; Ka´cl, K. H2O2-Bildung wa¨hrend ha¨moglobinoxidation durch phenylhydroxylamin in erythrozyten. Arch. Exp. Pathol. Pharmacol. 260:109–110; 1968. 10. Grossman, S. J.; Simson, J.; Jollow, D. J. Dapsone-induced hemolytic anemia: Effect of N-hydroxy dapsone on the sulfhydryl status and membrane proteins of rat erythrocytes. Toxicol. Appl. Pharmacol. 117:208–217; 1992. 11. Ross, D.; Norbeck, K.; Moldeus, P. The generation and subsequent fate of glutathionyl radicals in biological systems. J. Biol. Chem. 260:15028–15032; 1985. 12. Cadenas, E. Biochemistry of oxygen toxicity. Annu. Rev. Biochem. 58:79–110; 1989. 13. Maples, K. R.; Eyer, P.; Mason, R. P. Aniline-, phenylhydroxylamine-, nitrosobenzene-, and nitrobenzene-induced hemoglobin thiyl free radical formation in vivo and in vitro. Mol. Pharmacol. 37:311–318; 1990. 14. Winterbourn, C. C. Superoxide as an intracellular radical sink. Free Radic. Biol. Med. 14:85–90; 1993. 15. Bradshaw, T. P.; McMillan, D. C.; Crouch, R. K.; Jollow, D. J. Identification of free radicals produced in rat erythrocytes exposed to hemolytic concentrations of phenylhydroxylamine. Free Radic. Biol. Med. 18:279–285; 1995. 16. Floyd, R. A.; Soong, L. M.; Stuart, M. A.; Reigh, D. L. Spin trapping of free radicals produced from nitrosoamine carcinogens. Photochem. Photobiol. 28:857–862; 1978. 17. Davies, M. J.; Forni, L. G.; Shuter, S. L. Electron spin resonance and pulse radiolysis studies on the spin trapping of sulphur-centered radicals. Chem. Biol. Interact. 61:177–188; 1987. 18. Richmond, R.; Halliwell, B.; Chauhan, J.; Darbre, A. Superoxide-dependent formation of hydroxyl radicals: Detection of hydroxyl radicals by the hydroxylation of aromatic compounds. Anal. Biochem. 118:328–335; 1981. 19. Grootveld, M.; Halliwell, B. Aromatic hydroxylation as a potential measure of hydroxyl-radical formation in vivo. Identification of hydroxylated derivatives of salicylate in human body fluids. Biochem. J. 237:499–504; 1986. 20. Giulivi, C.; Davies, K. J. A novel antioxidant role for hemoglobin. The comproportionation of ferrylhemoglobin with oxyhemoglobin. J. Biol. Chem. 265:19453–19460; 1990. 21. Buettner, G. R. Spin trapping: ESR parameters of spin adducts. Free Radic. Biol. Med. 3:259–303; 1987. 22. Pou, S.; Hassett, D. J.; Britigan, B. E.; Cohen, M. S.; Rosen, G. M. Problems associated with spin trapping oxygen-centered free radicals in biological systems. Anal. Biochem. 177:1–6; 1989. 23. Walters, F. P.; Kennedy, F. G.; Jones, D. P. Oxidation of myoglobin in isolated adult rat cardiac myocytes by 15-hydroperoxy5,8,11,13-eicosatetraenoic acid. FEBS Lett. 163:292–296; 1983. 24. Galaris, D.; Sevanian, A.; Cadenas, E.; Hochstein, P. Ferrylmyoglobin-catalyzed linoleic acid peroxidation. Arch. Biochem. Biophys. 281:163–169; 1990. 25. Evelyn, K. A.; Malloy, H. T. Microdetermination of oxyhemoglobin, methemoglobin, and sulfhemoglobin in a single sample of blood. J. Biol. Chem. 126:655–662; 1938. 26. Berzofsky, J. A.; Peisach, J.; Blumberg, W. E. Sulfheme proteins. I. Optical and magnetic properties of sulfmyoglobin and its derivatives. J. Biol. Chem. 246:3367–3377; 1971. 27. Wallace, W. J.; Maxwell, J. C.; Caughey, W. S. The mechanisms of hemoglobin autoxidation. Evidence for proton-assisted nucleophilic displacement of superoxide by anions. Biochem. Biophys. Res. Commun. 57:1104–1110; 1974. 28. Misra, H. P.; Fridovich, I. The generation of superoxide radical during the autoxidation of hemoglobin. J. Biol. Chem. 247:6960– 6962; 1972. 29. Fridovich, I. Oxygen radicals, hydrogen peroxide, and oxygen
/ 2b29 2433 Mp 1192 Thursday Mar 13 10:16 PM EL–FRB (vol.27#6 ’97) 2433
Dapsone hydroxylamine in rat red cells
30. 31. 32. 33. 34.
35.
36. 37.
38.
toxicity. In: Pryor, W. A., ed. Free radicals in biology. New York: Academic Press; 1976:239–271. McCord, J. M.; Day, E. D. Superoxide-dependent production of hydroxyl radical catalyzed by iron-EDTA complex. FEBS Lett. 86:139–142; 1978. Halliwell, B.; Gutteridge, J. M. Biologically relevant metal iondependent hydroxyl radical generation. An update. FEBS Lett. 307:108–112; 1992. Gutteridge, J. M. Iron promoters of the Fenton reaction and lipid peroxidation can be released from haemoglobin by peroxides. FEBS Lett. 201:291–295; 1986. Puppo, A.; Halliwell, B. Formation of hydroxyl radicals from hydrogen peroxide in the presence of iron. Is haemoglobin a biological Fenton reagent? Biochem. J. 249:185–190; 1988. Galaris, D.; Mira, D.; Sevanian, A.; Cadenas, E.; Hochstein, P. Co-oxidation of salicylate and cholesterol during the oxidation of metmyoglobin by H2O2. Arch. Biochem. Biophys. 262:221– 231; 1988. Galaris, D.; Cadenas, E.; Hochstein, P. Glutathione-dependent reduction of peroxides during ferryl- and met-myoglobin interconversion: A potential protective mechanism in muscle. Free Radic. Biol. Med. 6:473–478; 1989. Itano, H. A. Phenyldiimide, hemoglobin, and Heinz bodies. Proc. Natl. Acad. Sci. USA 67:485–492; 1970. Ferrali, M.; Signorini, C.; Ciccoli, L.; Comporti, M. Iron release and membrane damage in erythrocytes exposed to oxidizing agents, phenylhydrazine, divicine and isouramil. Biochem. J. 285:295–301; 1992. Ferrali, M.; Signorini, C.; Ciccoli, L.; Comporti, M. Iron released
39.
40. 41. 42. 43. 44.
45.
46.
1193
from an erythrocyte lysate by oxidative stress is diffusible and in redox active form. FEBS Lett. 319:40–44; 1993. Ciccoli, L.; Signorini, C.; Alessandrini, C.; Ferrali, M.; Comporti, M. Iron release, lipid peroxidation, and morphological alterations of erythrocytes exposed to acrolein and phenylhydrazine. Exp. Mol. Pathol. 60:108–118; 1994. Rosen, G. M.; Freeman, B. A. Detection of superoxide generated by endothelial cells. Proc. Natl. Acad. Sci. USA 81:7269–7273; 1984. Rotilio, G.; Bray, R. C.; Fielden, E. M. A pulse radiolysis study of superoxide dismutase. Biochim. Biophys. Acta 268:605–609; 1972. Finkelstein, E.; Rosen, G. M.; Rauckman, E. J. Spin trapping of superoxide and hydroxyl radical: Practical aspects. Arch. Biochem. Biophys. 200:1–16; 1980. Osawa, Y.; Korzekwa, K. Oxidative modification by low levels of HOOH can transform myoglobin to an oxidase. Proc. Natl. Acad. Sci. USA 88:7081–7085; 1991. Mao, G. D.; Thomas, P. D.; Poznansky, M. J. Oxidation of spin trap 5,5-dimethyl-1-pyrroline-1-oxide in an electron paramagnetic resonance study of the reaction of methemoglobin with hydrogen peroxide. Free Radic. Biol. Med. 16:493–500; 1994. McMillan, D. C.; Simson, J. V.; Budinsky, R. A.; Jollow, D. J. Dapsone-induced hemolytic anemia: Effect of dapsone hydroxylamine on sulfhydryl status, membrane skeletal proteins and morphology of human and rat erythrocytes. J. Pharmacol. Exp. Ther. 274:540–547; 1995. Harrison, J. H., Jr.; Jollow, D. J. Contribution of aniline metabolites to aniline-induced methemoglobinemia. Mol. Pharmacol. 32:423–431; 1987.
/ 2b29 2433 Mp 1193 Thursday Mar 13 10:16 PM EL–FRB (vol.27#6 ’97) 2433
PII S0891-5849(96)00542-4
Original Contribution FORMATION OF FREE RADICALS AND PROTEIN MIXED DISULFIDES IN RAT RED CELLS EXPOSED TO DAPSONE HYDROXYLAMINE Timothy P. Bradshaw,*1 David C. McMillan,* Rosalie K. Crouch,† and David J. Jollow* Departments of *Cell and Molecular Pharmacology and Experimental Therapeutics, and †Ophthalmology, Medical University of South Carolina, Charleston, SC 29425, USA (Received 9 August 1995; Revised 19 June 1996; Re-revised 23 September 1996; Accepted 30 September 1996)
Abstract—The hemolytic activity of dapsone is well known to reside in its N-hydroxylamine metabolites. Addition of dapsone hydroxylamine (DDS–NOH) to red cell suspensions causes damage such that when reintroduced into the circulation of isologous rats, the injured cells are rapidly removed by the spleen. Hemolytic activity is associated with the extensive formation of disulfide-linked hemoglobin adducts on red cell membrane skeletal proteins. To determine if free radicals could be involved in this process, rat red cells were incubated with DDS–NOH in the presence of the spin trap, 5,5*-dimethyl-1-pyrroline-N-oxide (DMPO) and subjected to EPR analysis. Addition of DDS–NOH (25–50 mM) to a red cell suspension gave rise to a four-line (1:2:2:1) EPR spectrum with coupling constants identical to those of a DMPO-hydroxyl radical adduct (DMPO–OH) standard. No other radicals were detected; however, preincubation of red cells with cysteamine caused the DDS–NOH-generated DMPO–OH signal to be replaced by a cysteamine thiyl radical adduct signal. DDS–NOH-treated red cells were also found to contain ferrylhemoglobin, indicating the presence of hydrogen peroxide. Furthermore, DDS–NOH was found to stimulate salicylate hydroxylation in red cell suspensions, confirming the presence of oxygen radicals. These data support the hypothesis that oxygen radicals are involved in the mechanism underlying dapsone-induced hemolytic anemia. q 1997 Elsevier Science Inc. Keywords—Dapsone, Dapsone hydroxylamine, Erythrocytes, Rat, Hemolytic anemia, Hydroxyl radical, Cysteamine, Ferrylhemoglobin
sone administration in amounts sufficient to account for the toxicity of the parent drug.4 The hemolytic activity of arylamine drugs in humans has long been associated with a drug-induced oxidative stress within the red cell as evidenced by oxidation of cellular reduced glutathione (GSH) to oxidized glutathione (GSSG) and glutathione-protein mixed disulfides (protein-SSG), and by the enhanced sensitivity of individuals deficient in erythrocytic glucose-6-phosphate dehydrogenase activity.3,5,6 Oxidative stress within red cells is thought to result from a cyclic oxidation-reduction reaction that occurs between the arylhydroxylamine metabolite and oxyhemoglobin, yielding the nitrosoarene and methemoglobin, respectively.7 This reaction has been shown to produce hydrogen peroxide,8,9 and thus, is widely considered to generate other active oxygen species (i.e., superoxide anion
INTRODUCTION
Hemolytic anemia and methemoglobinemia are wellknown dose-limiting side effects in the therapeutic use of the arylamine drug dapsone.1–3 Because dapsone is not hemotoxic when incubated with red cells in vitro, it has long been recognized that toxicity is mediated by metabolites rather than the parent compound. We have previously shown that the hemolytic metabolites of dapsone are its N-hydroxy derivatives, dapsone hydroxylamine (DDS–NOH) and monoacetyl dapsone hydroxylamine. These metabolites are equipotent and direct-acting hemotoxicants, and are formed after dapAddress correspondence to: David J. Jollow, Ph.D., Department of Pharmacology, Medical University of South Carolina, 171 Ashley Avenue, Charleston, SC 29425-2251. 1 Present address of Timothy Bradshaw: Oligomer Development Department, Glaxo-Wellcome Research Institute, 5 Moore Drive, Research Triangle Park, NC 27709. 1183
/ 2b29 2433 Mp 1183 Thursday Mar 13 10:16 PM EL–FRB (vol.27#6 ’97) 2433
1184
T. P. BRADSHAW et al.
and hydroxyl radicals), which are capable of causing cellular injury. In studies on the effects of DDS–NOH in rat red cells, we have observed a rapid loss of GSH with concomitant formation of protein-SSG, disulfide-linked hemoglobin polymers, and disulfide-linked hemoglobin adducts on certain membrane skeletal proteins.10 We have hypothesized that oxygen-centered free radicals induce formation of both glutathione and hemoglobin thiyl radicals, and that the latter react with skeletal proteins of the cell membrane to form hemoglobin-skeletal protein adducts. This ‘‘damage’’ to critical membrane skeletal proteins is considered to initiate premature removal of the injured red cells by the spleen.6 An alternate hypothesis is that ‘‘compoundcentered’’ (e.g., arylhydronitroxide) free radicals are responsible for thiyl radical generation and subsequent hemoglobin-skeletal protein adduct formation.11–14 The present studies were undertaken to determine whether oxygen and/or compound-centered free radicals could be detected in rat red cells exposed to hemolytic concentrations of DDS–NOH and, thus, have a role in the mechanism underlying DDS–NOH-induced hemolytic injury. DDS–NOH was added to red cell suspensions in the presence and absence of a spin trap, and the incubates analyzed by EPR spectroscopy. The data indicate that in vitro treatment of rat red cells with concentrations of DDS–NOH at and below those necessary to induce a hemolytic response results in the generation of active oxygen species, including superoxide, hydrogen peroxide, hydroxyl radical, and ferryl heme species. Moreover, free radicals generated within red cells were able to transform the exogenous thiolcontaining compound, cysteamine, to a thiyl radical adduct. Addition of cysteamine to red cells also suppressed protein–SSG formation with concomitant production of cysteamine-protein mixed disulfides (protein–SSCysNH2). No evidence was observed that indicated the presence of a dapsone hydronitroxide free radical under hemolytic conditions. These data support the hypothesis that oxygen radicals, formed in the red cell as a consequence of DDS–NOH redox cycling, generate hemoglobin thiyl radicals, which in turn, attack membrane skeletal proteins.
MATERIALS AND METHODS
Chemicals Diethylenetriaminepentaacetic acid (DTPA), GSH, cysteamine, sodium salicylate, 2,3-dihydroxybenzoic acid, and 3,5-dihydroxybenzoic acid were obtained from Sigma Chemical Co. (St. Louis, MO). DMPO was purchased from Aldrich Chemical Co. (Milwaukee,
WI). DDS–NOH was synthesized as described previously.4 All other chemicals and reagents were of the best commercially available grade. Animals Male Sprague–Dawley rats (130–150 g) were purchased from Camm Research, Inc. (Wayne, NJ), and maintained on food and water ad lib. Animals were acclimated for 1 week to a 12-h light-dark cycle prior to their use. Red cells were collected from the descending aorta of ether-anesthetized rats into heparinized tubes, washed three times in isotonic phosphate-buffered saline (pH 7.4) supplemented with 10 mM D-glucose (PBSG), and resuspended in PBSG prior to their use. EPR studies Incubation mixtures (1 ml) contained 100 mM DMPO, 0.1 mM DTPA and red cells (5–20% suspension) in PBSG at 47C under aerobic conditions. Following a 5-min preincubation on ice, reactions were initiated by the addition of DDS–NOH dissolved in acetone (5 ml). EPR spectra were then recorded at ambient temperature on a Varian E-4 spectrometer operating at 10 mW power with a microwave frequency of 9.45 GHz and a field set of 3380 G as described previously.15 Additional spectra were recorded under the conditions described above following a 15-min preincubation at 377C in the presence and absence of 20 mM cysteamine or GSH. DMPO–adduct standards for hydroxyl radical and GSH thiyl radical were generated photolytically as described previously.16,17 HPLC analysis of salicylate hydroxylation in red cells Hydroxyl radical formation was detected with the use of salicylate as a trapping agent.18 Rat red cells (1.0 ml, 40% suspension) were preincubated with sodium salicylate (2.5 mM) for 5 min prior to the addition of DDS–NOH (180 mM). After a 30-min incubation at 377C, a 200 ml aliquot of the cell suspension was hemolyzed by a single freeze-thaw cycle. The hemolysate was then centrifuged (13,000 1 g for 3 min), and an aliquot of the supernatent (20 ml) was injected onto the HPLC. Products of hydroxyl radical attack on salicylate (2,3- and 2,5-dihydroxybenzoic acid) were separated on a Waters reverse-phase C18 column (Resolve) cartridge. The dihydroxybenzoic acids were eluted with an aqueous mobile phase consisting of monochloroacetic acid (75 mM), octanesulfonic acid (1.5 mM), EDTA (0.7 mM), and acetonitrile (5%) at a flow rate of 1.3
/ 2b29 2433 Mp 1184 Thursday Mar 13 10:16 PM EL–FRB (vol.27#6 ’97) 2433
Dapsone hydroxylamine in rat red cells
1185
ml/min.19 The dihydroxybenzoic acids were detected using a BAS LC-4 amperometric detector (Bioanalytical Systems, West Lafayette, IN) controlled at 5 nA full scale with a filter of 0.1 Hz in the reduction mode. A dual glassy-carbon electrode was operated in series; the upstream (generator) electrode was set at /750 mV, and the downstream electrode (detector) was set at /200 mV vs. Ag/AgCl. HPLC analysis of red cell sulfhydryls Following a 15-min preincubation in the presence and absence of cysteamine, DDS–NOH dissolved in acetone (10 ml) was added slowly to red cell suspension (2 ml, 40% suspension) and allowed to incubate in PBSG for 30 min at 377C. At the end of the incubation period, aliquots of the suspension (200 ml) were lysed and analyzed for protein–SSG and cysteamine-protein mixed disulfide (protein–SCysNH2) content by HPLCEC as described previously.10 The concentration of GSH or cysteamine released from protein following sodium borohydride-catalyzed reduction of disulfide bonds was determined from standard curves generated using known amounts of GSH and cysteamine treated identically to the experimental samples. Spectrophotometric detection of ferrylhemoglobin Rat red cells (1 ml, 40% suspension) were exposed to various concentrations of DDS–NOH. Aliquots (75 ml) were removed after designated intervals, lysed in 5 ml of hemolysis buffer (5 mM Na2HPO4, 0.5% Triton X-100, pH 7.5), and analyzed in a Shimadzu UV-160A double-beam UV-visible recording spectrophotometer. The amount of ferrylhemoglobin was measured in red cell incubations treated with sodium sulfide (2 mM), which converts ferrylhemoglobin to sulfhemoglobin (l max 620 nm).20 RESULTS
Formation of DMPO radical adducts in red cells Previous studies in our laboratory have shown that when rat 51Cr-labeled red cells are incubated in PBSG containing DDS–NOH for 2 h at 377C and then administered intravenously to isologous rats, the damaged red cells are removed rapidly from the circulation by the spleen. The hemolytic response is concentration dependent, with an EC50 of approximately 150 mM.4 To determine whether free radicals could be detected under these experimental conditions, rat red cells (5% suspension) were preincubated with 100 mM DMPO and then exposed to DDS–NOH. As shown in Fig. 1A,
Fig. 1. Detection of DMPO-radical adduct signals in rat erythrocytes exposed to DDS–NOH. (A) EPR spectrum recorded 5 min after addition of DDS–NOH (25 mM) to a 5% red cell suspension in PBSG containing 100 mM DMPO and 0.1 mM DTPA. (B) As in A, except DMPO was omitted. (C) As in A, except DDS–NOH was omitted. (D) As in A, except red cells were omitted. Receiver gain, 32,000; modulation amplitude, 2.5 G; time constant, 3.0.
addition of DDS–NOH (25 mM) to red cells generated a four-line (1:2:2:1) EPR signal. The appearance of the radical adduct signal was dependent on the presence of DMPO (Fig. 1B) and DDS–NOH (Fig. 1C). When red cells were excluded from the incubation mixture, a weak four-line signal was detected (Fig. 1D). At concentrations of DDS–NOH greater than 75 mM, the radical adduct signal deteriorated markedly and was not replaced by any new signal, even at concentrations of DDS–NOH as high as 200 mM (data not shown). In addition, no radical adduct signals were detected in lysed red cell incubates, which indicated that generation of this radical adduct is dependent on the the intact red cell. Based on the superficial structure of the signal, the EPR spectrum obtained from DDS–NOH-treated red cells (Fig. 1A) could be assigned as either a hydroxyl or thiyl free radical adduct signal.21 To discriminate between these two possibilities, authentic hydroxyl radical–DMPO (DMPO–OH) and glutathione thiyl radical-DMPO (DMPO–SG) adducts were prepared and
/ 2b29 2433 Mp 1185 Thursday Mar 13 10:16 PM EL–FRB (vol.27#6 ’97) 2433
1186
T. P. BRADSHAW et al.
their hyperfine splitting constants compared against those of the experimental sample. As shown in Table 1, the nitrogen (aN) and b-hydrogen (aH) coupling constants of the experimental sample and the DMPO-OH adduct standard were identical.
Hydroxylation of salicylate in red cells Although the formation of DMPO–OH has been cited frequently as evidence for the existence of hydroxyl radicals in a variety of in vitro systems, this radical adduct could have been formed as consequence of superoxide radical attack on the spin trap followed by reduction of the spin adduct to DMPO–OH.22 Thus, we utilized an alternative method for hydroxyl radical detection that exploits the ability of hydroxyl radicals to attack salicylate with subsequent formation of hydroxylated products that can be measured by HPLCEC.19 As shown in Fig. 2, addition of DDS–NOH (180 mM) to red cells containing salicylate (2.5 mM) resulted in the the formation of 2,3- and 2,5-dihydroxybenzoic acid (Fig. 2D). No peaks corresponding to 2,3- and 2,5dihydroxybenzoic acid were observed in red cells that contained DDS–NOH (Fig. 2B) or salicylate (Fig. 2C) alone.
Effect of cysteamine on radical adduct formation Rat red cells have been shown to lose GSH rapidly and extensively upon exposure to hemolytic concentrations of DDS–NOH, and this loss of GSH is associated with the formation of mixed disulfides between glutathione and the soluble protein of the red cell.10 Although glutathione thiyl radicals were not detected by EPR in DDS–NOH-treated red cells, their role as intermediates in the oxidation of GSH to GSSG and mixed disulfides has been well established.12 Given that GSH is lost rapidly from the red cell upon exposure to DDS–NOH, we examined the possibility that a thiyl radical signal could be observed in red cell suspensions
Table 1. Hyperfine Splitting Constants of DMPO Adducts Spin Adduct DMPO–OH standarda DMPO–SG standardb Red cell DMPO adductb
aN (G)
aH (G)
Reference
14.9 15.4 14.9
14.9 15.8 14.9
21 17
a Radical adduct standards formed by UV photolysis of hydrogen peroxide (DMPO–OH) and GSSG (DMPO–SG). The splitting constants for the adduct standards were determined experimentally. b Splitting constants determined from the EPR spectrum shown in Fig. 1A. Precise measurements were made from the original spectrum.
Fig. 2. Detection of salicylate hydroxylation in rat erythrocytes exposed to DDS–NOH. HPLC-EC elution profile of (A) 2,5- and 2,3-dihydroxybenzoic acid (DHBA) standards; (B) an erythrocyte suspension containing DDS–NOH (180 mM); (C) an erythrocyte suspension containing salicylate (2.5 mM); and (D) an erythrocyte suspension containing both DDS–NOH and salicylate. Erythrocytes (40% suspension) were preincubated for 5 min with salicylate prior to addition of DDS–NOH. The cells were then incubated for 30 min at 377C, and analyzed by HPLC-EC as described in Materials and Methods.
containing a high concentration of an exogenous thiolcontaining compound. Rat red cell suspensions were preincubated in the presence of the lipophilic thiol, cysteamine (20 mM), for 15 min prior to the addition of 50 mM DDS–NOH.
/ 2b29 2433 Mp 1186 Thursday Mar 13 10:16 PM EL–FRB (vol.27#6 ’97) 2433
Dapsone hydroxylamine in rat red cells
As shown in Fig. 3B, the presence of cysteamine resulted in replacement of the DMPO–OH control signal (Fig. 3A) with a stronger signal with different structure and coupling constants (Table 2). This new spectrum was generally similar in appearance to that of the photolytically generated cysteamine–DMPO radical adduct (DMPO–SCysNH2; Fig. 3C). However, the observed nitrogen coupling constants (aN) were not identical (Table 2). This reason for this difference is not clear; however, a mixture of DMPO radical adducts (i.e., DMPO–OH plus DMPO–SCysNH2) in the experimental sample could account for this difference. The possibility that the free radical species respon-
1187
Table 2. Hyperfine Splitting Constants of DMPO Adducts Spin Adduct DMPO–SG standarda DMPO–SCysNH2 standardb Red cell DMPO adductc
aN (G)
aH (G)
Reference
15.4 15.3 15.1
15.8 16.3 16.3
17
a Radical adduct standard formed by UV photolysis of GSSG (DMPO–SG). The splitting constants for the adduct standard were determined experimentally. b Radical adduct standard formed by UV photolysis of cysteamine (DMPO–SCysNH2). Splitting constants determined from the EPR spectrum shown in Fig. 2C. Precise measurements were made from the original spectrum. c Splitting constants of red cell DMPO adduct determined from the EPR spectrum shown in Fig. 2B. Precise measurements were made from the original spectrum.
sible for generating DMPO–SCysNH2 was capable of diffusing out of the red cell was examined by preincubating red cells with GSH, which cannot penetrate the plasma membrane, before addition of DDS–NOH. When GSH (20 mM) was substituted for cysteamine in DDS–NOH-treated red cell suspensions, no change in the DMPO–OH EPR spectrum was observed (data not shown), which indicated that the attacking free radical species was formed and remained within the intact red cell.
Effect of cysteamine on mixed disulfide formation
Fig. 3. Effect of addition of cysteamine to rat erythrocytes on the EPR signal induced by DDS–NOH. (A) EPR spectrum recorded 8 min after addition of DDS–NOH (50 mM ) to a 5% red cell suspension in PBSG containing 100 mM DMPO and 0.1 mM DTPA. (B) As in A, except cysteamine (20 mM) was added 15 min before DDS– NOH. (C) Photolytically generated DMPO–cysteamine adduct standard. Receiver gain, 32,000 (A), 3200 (B and C); modulation amplitude, 2.5 G; time constant, 3.0 (A), 1.0 (B and C).
To determine if the cysteamine-thiyl radicals generated on addition of cysteamine to DDS–NOH-treated rat red cells could suppress the formation of glutathione-protein mixed disulfides, red cells were pretreated with cysteamine (20 mM) 30 min prior to addition of various concentrations of DDS–NOH. After a 30-min incubation in the presence of DDS–NOH, an aliquot of the incubation mixture was removed for quantitation of protein–SSG and protein–SCysNH2 formation. Our laboratory has previously developed a highly sensitive HPLC-EC assay for determination of GSH, GSSG, and protein–SSG content in red cell suspensions.10 To demonstrate that this assay could also be used to quantitate protein–SCysNH2 formation, cysteamine, and GSH were added to red cell lysates, and the samples were then processed and analyzed by HPLC-EC. Figure 4A shows a typical HPLC elution profile of a sample from a red cell lysate treated with cysteamine and GSH. The data indicated that both thiols could be measured using the same electrochemical detector conditions, and were sufficiently well separated to permit quantitation of both thiols in the same experimental samples. A standard curve generated using known amounts of GSH and cysteamine (Fig. 4B) was used to quantitate mixed disulfide content in ex-
/ 2b29 2433 Mp 1187 Thursday Mar 13 10:16 PM EL–FRB (vol.27#6 ’97) 2433
1188
T. P. BRADSHAW et al.
Fig. 4. (A) HPLC-EC elution profile of GSH and cysteamine (CysNH2) standards. (B) Standard curve generated by addition of known amounts of GSH (j) and CysNH2 (l) to PBSG. Standards were treated in the same manner as the red cell samples.
perimental samples. Recovery of both GSH and cysteamine from incubation mixtures was ú95%. In agreement with previous studies,10 addition of DDS–NOH to rat red cells induced a concentrationdependent increase in protein–SSG formation (Fig. 5). Inclusion of cysteamine (20 mM) in the incubation mixture resulted in greater than 80% suppression of protein–SSG formation. This response was accompanied by a concentration-dependent formation of protein–SCysNH2 (Fig. 6). Collectively, these data indicate that cysteamine can substitute for GSH as a nucleophilic reactant in red cells, and support the hypothesis that thiyl radicals originate via oxygen-centered radical attack on cellular thiol groups.
Ferrylhemoglobin formation in red cells Hydrogen peroxide generated within red cells can undergo a variety of reactions, one of which is with oxyhemoglobin to form ferryl heme species ([HbFeIV – OH]30).20 Because ferryl heme species are potent oxidants capable of contributing to intracellular damage,12,23,24 it was of interest to determine whether ferrylhemoglobin is formed in rat red cells under hemolytic conditions. Addition of DDS–NOH to rat red cells induced a concentration-dependent increase in the absorbance at 630 nm, consistent with the formation of methemoglobin (HbFe3/). Addition of sodium cya-
nide to the cuvette to form cyanomethemoglobin abolished the peak at 630 nm, confirming its identity as methemoglobin.25 However, removal of the methemoglobin peak did not reveal a spectrum indicative of the presence of ferrylheme species (i.e., increase in absorbance at 545 and 580 nm).20 To examine the possibility ferrylhemoglobin was formed as an intermediate in DDS–NOH-treated rat red cells, sodium sulfide (2 mM) was added to the incubates before addition of DDS–NOH. Sodium sulfide has been used to trap the ferryl oxidation state of myoglobin23 and hemoglobin20 in cellular systems by converting it irreversibly to sulfhemoglobin.26 Addition of hemolytic concentrations of DDS–NOH to red cell incubates containing Na2S resulted in the appearance of a new peak at 620 nm, consistent with the formation of sulfhemoglobin.26 Addition of cyanide to the cuvette had no effect on the appearance of the spectrum. In Na2S-pretreated red cells ferrylhemoglobin was observed to accumulate with increasing incubation time, and this accumulation was concentration dependent (Fig. 7 inset), with an EC50 of about 30 mM DDS– NOH. Under the experimental conditions described above, the magnitude of the peak height at 620 nm represented the amount of ferrylhemoglobin that had accumulated (as sulfhemoglobin) in the incubation mixtures as a function of incubation time. To determine whether ferryl hemoglobin was accumulating during the incuba-
/ 2b29 2433 Mp 1188 Thursday Mar 13 10:16 PM EL–FRB (vol.27#6 ’97) 2433
Dapsone hydroxylamine in rat red cells
Fig. 5. Effect of pretreatment with cysteamine on DDS–NOH-induced protein–SSG formation in rat erythrocytes. Red cell suspensions were incubated in the presence (h) and absence (j) of cysteamine (20 mM) for 15 min at 377C prior to addition of various concentrations of DDS–NOH. After 30 min of incubation at 377C, aliquots of the red cells were removed and analyzed for protein– SSG by HPLC-EC. Values are means of duplicate determinations.
tion or was present at a steady-state level, Na2S was added to the red cell incubates 0, 5, 10, and 20 min after the addition of DDS–NOH (100 mM). Incubation was continued for a further 1, 5, or 20 min, after which samples were removed for spectral analysis. The peak height at 620 nm was proportional to the time of incubation post Na2S and independent of the duration of incubation prior to Na2S addition. Calculation of DO.D./min for the various Na2S accumulation times indicated a steady-state production of ferrylhemoglobin in these DDS–NOH-treated rat red cells equivalent to a DO.D. of ca. 0.025–0.03/min. The data indicate that ferrylhemoglobin does not accumulate to measurable levels under these conditions.
1189
between glutathione and cellular protein, and by the formation of disulfide-linked adducts between hemoglobin and certain membrane skeletal proteins.10 We have hypothesized that the latter reaction induces a change on the internal surface of the red cell that generates an antigenic site on the external cell surface, which in turn, provokes the immune system to remove the damaged red cell from the circulation by macrophages in the spleen.6 We have also reported that glutathione and hemoglobin thiyl radicals are generated in rat red cells incubated with hemolytic concentrations of phenylhydroxylamine.15 This observation suggested that mixed disulfides arise via thiyl radical attack on free sulfhydryl groups of cellular proteins; however, the origin of the thiyl radicals in phenylhydroxylamine-treated red cells has remained unclear. One possibility is that oxygen-centered free radicals are responsible for the generation of thiyl radicals. This viewpoint is supported by the extensive studies of Kiese and colleagues,7 which showed that arylhydroxylamines enter red cells and undergo a coupled oxidation reaction with oxyhemoglobin (Fe2/), forming the nitrosoarene and methemoglobin (Fe3/), respectively. The nitrosoarene is reduced by an NADPH-dependent reductase to the arylhydroxylamine, which is available to oxidize another equivalent of oxyhemoglobin. In this fashion, many equivalents of methemoglobin are produced per equivalent of arylhydroxylamine. Although the precise chemistry underlying this enzymatic cycle remains to be fully elucidated, molecular oxygen is thought to act as the
DISCUSSION
In previous studies on the mechanism underlying the hemolytic anemia induced by aniline and related arylamine drugs such as dapsone and primiquine, we have utilized an in vitro incubation/in vivo response assay that has allowed us to identify biochemical events within red cells that correlate with the toxic response.6 Using this experimental approach we have shown that rat red cells exposed in vitro to phenylhydroxylamine and DDS–NOH rapidly lose GSH in a concentrationand time-dependent manner. This loss of GSH is accompanied by the formation of disulfide-linked adducts
Fig. 6. Effect of pretreatment with cysteamine on DDS–NOH-induced mixed disulfide formation in rat erythrocytes. Red cell suspensions were incubated in the presence of cysteamine (20 mM) for 15 min at 377C prior to addition of various concentrations of DDS– NOH. After 30 min of incubation at 377C, aliquots of the red cells were removed and analyzed for protein–SSG (m) and protein-cysteamine mixed disulfides (Prot-S-S-CysNH2, j) by HPLC-EC. Values are means of duplicate determinations.
/ 2b29 2433 Mp 1189 Thursday Mar 13 10:16 PM EL–FRB (vol.27#6 ’97) 2433
1190
T. P. BRADSHAW et al.
Fig. 7. Spectrophotometric detection of ferrylhemoglobin in rat erythrocytes exposed to DDS–NOH. Erythrocytes (40% suspension) were exposed to the indicated concentrations of DDS–NOH (0–200 mM) for 10 min at 377C. Sodium sulfide (2 mM) was added prior to DDS– NOH to convert the ferryl species to the sulfheme derivative (lmax 620 nm). Inset: Absorbance was measured at 620 nm 20 min after addition of various concentrations of DDS–NOH to red cell incubates containing 2 mM sodium sulfide.
acceptor of electrons for the oxidation steps of this reaction, resulting in its reduction to superoxide anion radical followed by dismutation to hydrogen peroxide.27–29 Under normal conditions, superoxide and hydrogen peroxide are removed by superoxide dismutase and catalase/glutathione peroxidase, respectively. However, these protective mechanisms are likely to be overwhelmed by the continuous and excessive formation of active oxygen species produced by arylhydroxylamine redox cycling. Superoxide and hydrogen peroxide would then be available to interact with iron (either free or heme iron) to generate highly reactive intermediates, including hydroxyl radical and ferryl heme species, which in turn, could react with cellular sulfhydryl groups to generate thiyl radicals.12,30,31 The present studies report that incubation of rat red cells with hemotoxic concentrations of DDS–NOH resulted in the detection of hydroxyl radical adducts of DMPO in rat red cells. The DMPO–OH radical adduct signal was dependent on the presence of DDS–NOH, red cells, and the spin trap (Fig. 1), and it had EPR spin constants that were identical to those of a photolytically generated DMPO–OH radical adduct standard (Table 1). That free hydroxyl radicals exist in the red cell under hemolytic conditions was further supported by the observation that salicylic acid was converted to 2,3-
and 2,5-dihydroxybenzoic acid under these incubation conditions (Fig. 2). Together, these data provide strong support for the postulate that hydroxyl radicals are formed in the red cell and, hence, could be responsible for the development of hemolytic injury. If hydroxyl radicals were responsible for the production of DMPO–OH, then a Fenton-type reaction would have to be invoked to explain this observation. However, a variety of studies using purified hemoglobin preparations and hydrogen peroxide have suggested that hemoglobin (or heme) per se is unable to catalyze a Fenton reaction.32,33 In these cell-free systems, other reactions are considered to predominate, in particular, the reaction of hydrogen peroxide with hemoglobin and methemoglobin to form ferryl heme species.34,35 On the other hand, the nature of redox reactions within the intact red cell remain poorly understood, and it seems inappropriate at this time to discard the possibility that a Fenton-type reaction can occur within the red cell. Of importance, it has long been known that oxidative damage to hemoglobin results in the release of heme from methemoglobin,32,36 as well as the release of iron in a diffusable and redox active form.37–39 Although the amount of free iron produced in DDS–NOH-treated red cells is unknown and warrants further investigation, the extent of hemoglobin oxidation under hemolytic conditions is considerable (ú40% methemoglobin), and, hence, it is plausible that significant amounts of iron in a form capable of catalyzing a Fenton reaction are available under hemolytic conditions. A second possibility is that DMPO–OH arose from active oxygen species other than hydroxyl radical. For example, superoxide anion radical is known to react with DMPO to form the superoxide radical adduct, DMPO– OOH, which subsequently could have been reduced to DMPO–OH by cellular reducing agents.40 Because DMPO–OOH was not detected by EPR in these studies, this possibility could not be confirmed. However, superoxide is known to react extremely slowly with nitroxide spin traps (10 M01s01) compared to its spontaneous and SOD-catalyzed dismutation.41 In contrast, the reaction between hydroxyl radicals and DMPO is extremely rapid (2 1 109 M01s01).42 In the present studies, the DMPO– OH radical adduct signal could be observed immediately after addition of DDS–NOH-treated cells to the EPR flat cell (4-min scan time). That some part of the observed DMPO–OH radical adduct arose via this mechanism cannot be eliminated by the present data, though in view of the production of 2,3- and 2,5-dihydroxybenzoate from salicylate under these experimental conditions (Fig. 2), it seems unlikely that this mechanism is a major contributor. A third possibility is that ferryl heme species were responsible for generating DMPO–OH, and could the-
/ 2b29 2433 Mp 1190 Thursday Mar 13 10:16 PM EL–FRB (vol.27#6 ’97) 2433
Dapsone hydroxylamine in rat red cells
oretically have been responsible for the hydroxylation of salicylate. Indeed, ferrylhemoglobin is known to be a strong oxidant analogous to other peroxidases and catalase complex I,12,43 and the present data clearly document that ferrylhemoglobin is present in rat red cells under DDS–NOH hemolytic conditions (Fig. 7). However, the characteristic radical adduct signal for 5,5dimethylpyrrolidone-2(2)-oxyl-(1), which would have indicated a reaction between ferryl heme species and the spin trap,44 was not observed. Furthermore, ferrylhemoglobin could not be detected spectrophotometrically and was recognized only by trapping it as the stable sulfheme derivative. This observation, along with studies on the time dependence of ferrylhemoglobin formation, suggested a low steady-state level of this species. The low steady-state level of ferrylhemoglobin in DDS–NOH-treated red cells is consistent with earlier studies,20 which showed that when oxyhemoglobin concentration is in excess, ferrylhemoglobin is reduced rapidly by oxyhemoglobin to methemoglobin. To what extent ferrylhemoglobin reduction contributes to DDS– NOH-induced methemoglobin formation remains to be determined; however, the concentration-response curves for DDS–NOH-induced methemoglobin formation6 and ferryl heme formation (Fig. 7, inset) are similar, raising the possibility that ferrylhemoglobin may contribute to methemoglobin levels at hemolytic concentrations of DDS–NOH. Whether or not ferrylhemoglobin has a role in the DDS–NOH-dependent hemolytic process is also unclear. While steric hinderance considerations suggest that a direct role in the generation of skeletal-protein thiyl radicals is unlikely, the possibility of an indirect role cannot be excluded at this time. The presence of ferrylhemoglobin does, however, clearly indicate that hydrogen peroxide and, by inference, superoxide anion radical, are generated in DDS–NOH-treated red cells. Collectively, the data in the present studies demonstrate the complexity of the arylhydroxylamine redox cycle in the intact red cell, and indicate that further studies will be necessary to determine the individual contributions of these reactive intermediates to the hemolytic response. Although the present studies clearly support a role for active oxygen species in dapsone-induced hemolytic anemia, the link between active oxygen species generation and thiyl radical formation remains to be firmly established. On the one hand, a DMPO–SG radical adduct signal was not detected in DDS–NOHtreated red cells. The reason for this is not clear, but may be a reflection of the rapid loss of GSH (ca. 85– 95%) that occurs upon addition of hemolytic concentrations of DDS–NOH to rat red cells.10,45 Because DMPO–SG is unstable and must be continuously gen-
1191
erated to be detected by EPR,17 the lack of a strong DMPO–SG signal that could compete with that of DMPO–OH presumably is a consequence of the extensive depletion of red cell GSH by DDS–NOH. Because glutathione thiyl and hydroxyl radical adducts of DMPO give rise to similar four-line (1:2:2:1) EPR signals,21 a minor DMPO–SG spectrum could have been obscured by the DMPO–OH signal. On the other hand, addition of cysteamine to replenish the level of nonprotein sulfhydryl groups in the red cell did result in conversion of the EPR spectrum from a hydroxyl radical adduct signal to that of a cysteamine thiyl radical adduct signal (Fig. 3). Furthermore, cysteamine suppressed protein-SSG formation (Fig. 5) and, like intracellular GSH, became bound to cellular protein via disulfide linkages in a concentration-dependent manner (Fig. 6). In contrast, when GSH, which cannot readily penetrate the cell membrane, was added to the media in place of cysteamine, the DMPO–OH signal was not quenched, indicating that the attacking radical species is formed within the red cell and does not diffuse across the cell membrane. A fourth possibility is that glutathione, hemoglobin, and cysteamine thiyl radicals arise via reaction of their sulfhydryl groups with compound-centered (i.e., dapsone hydronitroxide) free radicals. In support of this viewpoint, Maples et al.13 have shown by EPR that phenylhydroxylamine can undergo a one-electron oxidation to a phenylhydronitroxide radical, and that this radical is capable of generating glutathione thiyl radicals upon addition of GSH and DMPO to the system. However, the toxicological relevance of this data is uncertain because it was obtained in phosphate buffer; i.e., in the absence of hemoglobin or of intact red cells, and at a concentration of phenylhydroxylamine (1 mM), which is about an order of magnitude greater than the Cmax for methemoglobin formation in the intact rat red cell.46 When examined at hemotoxic concentrations (50–100 mM), DDS–NOH did not give rise to an EPR spectrum indicative of the presence of dapsone hydronitroxide in the rat red cell. It is not yet known whether the failure to detect the hydronitroxide radical reflects a relative insensitivity of the EPR spectroscopy or whether it indicates an absence of significant levels of this species in the cell. On chemical grounds, it is likely that the hydronitroxide radical is an intermediate during the two-electron oxidation of arylhydroxylamines to nitrosoarenes in the red cell, and thus could be involved in the redox cycling of arylhydroxylamines thought to occur within or near the heme pocket.7 The degree to which such radicals, if formed, might escape the redox cycle into the cytosol to react with critical intracellular targets, and hence, participate in the hemolytic process, is un-
/ 2b29 2433 Mp 1191 Thursday Mar 13 10:16 PM EL–FRB (vol.27#6 ’97) 2433
1192
T. P. BRADSHAW et al.
known. It is noteworthy that while exogenous cysteamine quenched the DMPO–OH signal, exogenous GSH did not. Because dapsone hydronitroxide radicals may be expected to have sufficient lipophilicity and stability to penetrate the cell membrane and leave the cell, the inability of extracellular GSH to suppress the DMPO–OH signal implies that the arylhydronitroxide is not the species that attacks thiol groups and generates thiyl radicals. Based on the current data, we have no evidence that a compound-centered free radical species, such as dapsone hydronitroxide, is generated in rat red cells under hemotoxic conditions or that it plays a direct role in DDS–NOH hemotoxicity. On the other hand, in view of its likely formation during redox cycling, its probable high turnover, and possible lack of sensitivity to EPR detection, it is premature to rule out its participation in the hemolytic process. In summary, the present studies indicate that a number of potentially toxic active oxygen species are generated in rat red cells exposed to hemolytic concentrations of DDS–NOH. Their formation under hemolytic conditions is compatible with one or more of these species having a causal role in the hemolytic process. While the precise roles(s) of active oxygen species in red cell injury is not yet understood and warrants further investigation, the present data support the hypothesis that red cell damage arises from the interaction of cellular thiols with oxygen radicals generated during the redox cycling of arylhydroxylamines with oxyhemoglobin, and that subsequent binding of hemoglobin with skeletal protein via disulfide linkages could provide the trigger for premature sequestration of these oxidatively damaged cells by the spleen.6 Acknowledgements — This study was supported by NIH Grant HL30038. The authors wish to thank Jennifer Schulte and Leslie Edwards for their excellent technical assistance in the preparation of this manuscript.
REFERENCES 1. Degowin, R. L.; Eppes, R. B.; Powell, R. D.; Carson, P. E. The haemolytic effects of diaphenylsulfone (DDS) in normal subjects and in those with glucose-6-phosphate-dehydrogenase deficiency. Bull. World Health Organ. 35:165–179; 1966. 2. Mayer, K.; Ley, A. B. Hemolysis of red cells due to sulfone. Ann. Intern. Med. 72:711–714; 1970. 3. Beutler, E. Hemolytic anemia in disorders of red cell metabolism. New York: Plenum Medical; 1978. 4. Grossman, S. J.; Jollow, D. J. Role of dapsone hydroxylamine in dapsone-induced hemolytic anemia. J. Pharmacol. Exp. Ther. 244:118–125; 1988. 5. Kosower, N. S.; Zipser, Y.; Faltin, Z. Membrane thiol-disulfide status in glucose-6-phosphate dehydrogenase deficient red cells. Relationship to cellular glutathione. Biochim. Biophys. Acta 691:345–352; 1982.
6. Jollow, D. J.; Bradshaw, T. P.; McMillan, D. C. Dapsone-induced hemolytic anemia. Drug Metab. Rev. 27:107–124; 1995. 7. Kiese, M. Methemoglobinemia: A comprehensive treatise. Cleveland, OH: CRC Press; 1974. 8. Rostorfer, H. H.; Cormier, M. J. Formation of ‘‘hydrogen peroxide’’ in the reaction of oxyhemoglobin with methemoglobin forming agents. Arch. Biochem. Biophys. 71:235–249; 1957. 9. Ellederova´, D.; Wagner, J.; Ka´cl, K. H2O2-Bildung wa¨hrend ha¨moglobinoxidation durch phenylhydroxylamin in erythrozyten. Arch. Exp. Pathol. Pharmacol. 260:109–110; 1968. 10. Grossman, S. J.; Simson, J.; Jollow, D. J. Dapsone-induced hemolytic anemia: Effect of N-hydroxy dapsone on the sulfhydryl status and membrane proteins of rat erythrocytes. Toxicol. Appl. Pharmacol. 117:208–217; 1992. 11. Ross, D.; Norbeck, K.; Moldeus, P. The generation and subsequent fate of glutathionyl radicals in biological systems. J. Biol. Chem. 260:15028–15032; 1985. 12. Cadenas, E. Biochemistry of oxygen toxicity. Annu. Rev. Biochem. 58:79–110; 1989. 13. Maples, K. R.; Eyer, P.; Mason, R. P. Aniline-, phenylhydroxylamine-, nitrosobenzene-, and nitrobenzene-induced hemoglobin thiyl free radical formation in vivo and in vitro. Mol. Pharmacol. 37:311–318; 1990. 14. Winterbourn, C. C. Superoxide as an intracellular radical sink. Free Radic. Biol. Med. 14:85–90; 1993. 15. Bradshaw, T. P.; McMillan, D. C.; Crouch, R. K.; Jollow, D. J. Identification of free radicals produced in rat erythrocytes exposed to hemolytic concentrations of phenylhydroxylamine. Free Radic. Biol. Med. 18:279–285; 1995. 16. Floyd, R. A.; Soong, L. M.; Stuart, M. A.; Reigh, D. L. Spin trapping of free radicals produced from nitrosoamine carcinogens. Photochem. Photobiol. 28:857–862; 1978. 17. Davies, M. J.; Forni, L. G.; Shuter, S. L. Electron spin resonance and pulse radiolysis studies on the spin trapping of sulphur-centered radicals. Chem. Biol. Interact. 61:177–188; 1987. 18. Richmond, R.; Halliwell, B.; Chauhan, J.; Darbre, A. Superoxide-dependent formation of hydroxyl radicals: Detection of hydroxyl radicals by the hydroxylation of aromatic compounds. Anal. Biochem. 118:328–335; 1981. 19. Grootveld, M.; Halliwell, B. Aromatic hydroxylation as a potential measure of hydroxyl-radical formation in vivo. Identification of hydroxylated derivatives of salicylate in human body fluids. Biochem. J. 237:499–504; 1986. 20. Giulivi, C.; Davies, K. J. A novel antioxidant role for hemoglobin. The comproportionation of ferrylhemoglobin with oxyhemoglobin. J. Biol. Chem. 265:19453–19460; 1990. 21. Buettner, G. R. Spin trapping: ESR parameters of spin adducts. Free Radic. Biol. Med. 3:259–303; 1987. 22. Pou, S.; Hassett, D. J.; Britigan, B. E.; Cohen, M. S.; Rosen, G. M. Problems associated with spin trapping oxygen-centered free radicals in biological systems. Anal. Biochem. 177:1–6; 1989. 23. Walters, F. P.; Kennedy, F. G.; Jones, D. P. Oxidation of myoglobin in isolated adult rat cardiac myocytes by 15-hydroperoxy5,8,11,13-eicosatetraenoic acid. FEBS Lett. 163:292–296; 1983. 24. Galaris, D.; Sevanian, A.; Cadenas, E.; Hochstein, P. Ferrylmyoglobin-catalyzed linoleic acid peroxidation. Arch. Biochem. Biophys. 281:163–169; 1990. 25. Evelyn, K. A.; Malloy, H. T. Microdetermination of oxyhemoglobin, methemoglobin, and sulfhemoglobin in a single sample of blood. J. Biol. Chem. 126:655–662; 1938. 26. Berzofsky, J. A.; Peisach, J.; Blumberg, W. E. Sulfheme proteins. I. Optical and magnetic properties of sulfmyoglobin and its derivatives. J. Biol. Chem. 246:3367–3377; 1971. 27. Wallace, W. J.; Maxwell, J. C.; Caughey, W. S. The mechanisms of hemoglobin autoxidation. Evidence for proton-assisted nucleophilic displacement of superoxide by anions. Biochem. Biophys. Res. Commun. 57:1104–1110; 1974. 28. Misra, H. P.; Fridovich, I. The generation of superoxide radical during the autoxidation of hemoglobin. J. Biol. Chem. 247:6960– 6962; 1972. 29. Fridovich, I. Oxygen radicals, hydrogen peroxide, and oxygen
/ 2b29 2433 Mp 1192 Thursday Mar 13 10:16 PM EL–FRB (vol.27#6 ’97) 2433
Dapsone hydroxylamine in rat red cells
30. 31. 32. 33. 34.
35.
36. 37.
38.
toxicity. In: Pryor, W. A., ed. Free radicals in biology. New York: Academic Press; 1976:239–271. McCord, J. M.; Day, E. D. Superoxide-dependent production of hydroxyl radical catalyzed by iron-EDTA complex. FEBS Lett. 86:139–142; 1978. Halliwell, B.; Gutteridge, J. M. Biologically relevant metal iondependent hydroxyl radical generation. An update. FEBS Lett. 307:108–112; 1992. Gutteridge, J. M. Iron promoters of the Fenton reaction and lipid peroxidation can be released from haemoglobin by peroxides. FEBS Lett. 201:291–295; 1986. Puppo, A.; Halliwell, B. Formation of hydroxyl radicals from hydrogen peroxide in the presence of iron. Is haemoglobin a biological Fenton reagent? Biochem. J. 249:185–190; 1988. Galaris, D.; Mira, D.; Sevanian, A.; Cadenas, E.; Hochstein, P. Co-oxidation of salicylate and cholesterol during the oxidation of metmyoglobin by H2O2. Arch. Biochem. Biophys. 262:221– 231; 1988. Galaris, D.; Cadenas, E.; Hochstein, P. Glutathione-dependent reduction of peroxides during ferryl- and met-myoglobin interconversion: A potential protective mechanism in muscle. Free Radic. Biol. Med. 6:473–478; 1989. Itano, H. A. Phenyldiimide, hemoglobin, and Heinz bodies. Proc. Natl. Acad. Sci. USA 67:485–492; 1970. Ferrali, M.; Signorini, C.; Ciccoli, L.; Comporti, M. Iron release and membrane damage in erythrocytes exposed to oxidizing agents, phenylhydrazine, divicine and isouramil. Biochem. J. 285:295–301; 1992. Ferrali, M.; Signorini, C.; Ciccoli, L.; Comporti, M. Iron released
39.
40. 41. 42. 43. 44.
45.
46.
1193
from an erythrocyte lysate by oxidative stress is diffusible and in redox active form. FEBS Lett. 319:40–44; 1993. Ciccoli, L.; Signorini, C.; Alessandrini, C.; Ferrali, M.; Comporti, M. Iron release, lipid peroxidation, and morphological alterations of erythrocytes exposed to acrolein and phenylhydrazine. Exp. Mol. Pathol. 60:108–118; 1994. Rosen, G. M.; Freeman, B. A. Detection of superoxide generated by endothelial cells. Proc. Natl. Acad. Sci. USA 81:7269–7273; 1984. Rotilio, G.; Bray, R. C.; Fielden, E. M. A pulse radiolysis study of superoxide dismutase. Biochim. Biophys. Acta 268:605–609; 1972. Finkelstein, E.; Rosen, G. M.; Rauckman, E. J. Spin trapping of superoxide and hydroxyl radical: Practical aspects. Arch. Biochem. Biophys. 200:1–16; 1980. Osawa, Y.; Korzekwa, K. Oxidative modification by low levels of HOOH can transform myoglobin to an oxidase. Proc. Natl. Acad. Sci. USA 88:7081–7085; 1991. Mao, G. D.; Thomas, P. D.; Poznansky, M. J. Oxidation of spin trap 5,5-dimethyl-1-pyrroline-1-oxide in an electron paramagnetic resonance study of the reaction of methemoglobin with hydrogen peroxide. Free Radic. Biol. Med. 16:493–500; 1994. McMillan, D. C.; Simson, J. V.; Budinsky, R. A.; Jollow, D. J. Dapsone-induced hemolytic anemia: Effect of dapsone hydroxylamine on sulfhydryl status, membrane skeletal proteins and morphology of human and rat erythrocytes. J. Pharmacol. Exp. Ther. 274:540–547; 1995. Harrison, J. H., Jr.; Jollow, D. J. Contribution of aniline metabolites to aniline-induced methemoglobinemia. Mol. Pharmacol. 32:423–431; 1987.
/ 2b29 2433 Mp 1193 Thursday Mar 13 10:16 PM EL–FRB (vol.27#6 ’97) 2433