ARCHIVES

OF BIOCHEMISTRY

AND

BIOPHYSICS

Vol. 286, No. 1, April, pp. 30-37, 1991

Reductive Metabolism of Nitroprusside Hepatocytes and Human Erythrocytes D. N. Ramakrishna Faculty of Pharmacy,

Rao, Sahar Elguindi, University

and Peter J. O’Brien’

of Toronto, 19, Russell Street, Toronto. Canada M5S 1Al

Received August 21, 1990, and in revised form November

15, 1990

The metabolism of nitroprusside by hepatocytes or subcellular fractions involves a one-electron reduction of nitroprusside to the corresponding metal-nitroxyl radical. Thiol compounds also reduced nitroprusside to the metal-nitroxyl radical apparently via a thiol adduct. The nitroprusside reduction by microsomes was shown to be due to cytochrome P450 reductase as an antibody to cytochrome P450 reductase inhibits the microsomal reduction of nitroprusside, and the inhibitors of cytochrome P450 such as carbon monoxide or metyrapone had no effect, The reduction of nitroprusside by mitochondria in the presence of NADH or NADPH also produced the metal-nitroxyl radical. In hepatocytes, both mitochondria and the cytochrome P450 reductase are involved in the reduction of nitroprusside. The reductive metabolism of nitroprusside was found to produce toxic by-products, namely, free cyanide anion and hydrogen peroxide. We have also detected thiyl radicals formed in the thiol compound reduction of NP. We propose that cyanide and hydrogen peroxide are important toxic species formed in the metabolism of nitroprusside. The rate of reductive metabolism of nitroprusside by rat hepatocytes was much higher than with human erythrocytes. Therefore the major site of nitroprusside metabolism in uivo may be liver and not blood as originally proposed. o lssl Academic

Press,

in Rat

Inc.

Nitroprusside [NP;2 pentacyanonitrosylferrate (II)] is a potent hypotensive agent widely used for controlled hypertension during surgery, in hypertensive emergencies, I To whom correspondence should be addressed. ’ Abbreviations used: BCNU, 1,3-bis (2-chloroethyl)-l-nitrosourea; DCNB, 1,2 dichloro-knitrobenzene; DMPO, 5,5’-dimethyl-l-pyrrolineN-oxide; DTPA, diethylenetriamine pentaacetic acid; DTT, dithiothreitol; ESR, electron spin resonance; GSH, glutathione; GSSG, glutathione disulfide; MNO, metal-nitroxyl; NP, nitroprusside; SOD, superoxide dismutase.

and to improve heart function after myocardial infarction (l-6). The proposed mechanism for the hypotensive action of this drug involves the formation of a nitrosyl hemoglobin by the interaction of the drug with hemoglobin (7,8). The nitrosylated hemoglobin stimulates guanylate cyclase which results in the hydrolysis of GTP to cyclicGMP (9, 10); this activates protein kinases that leads to protein phophorylation/dephosphorylation, subsequently producing the relaxation in vascular smooth muscle (ll13). The nitric oxide moiety of NP released by other pathways also reacts with hemoglobin to produce the nitrosyl hemoglobin (7, 14). The use of NP in the treatment of high blood pressure and heart attack is limited, because of the fact that it caused deaths in some cases following its administration (15, 16). This has been suggested to be due to cyanide, apparently released from NP per se in human body (17, 18), but this has been disputed (14, 19). However, it is known that the reductive metabolites of NP can liberate free cyanide anion (14). Although the chemistry of this compound has been studied, some of its reaction mechanisms are not very well understood. We have recently shown that NP cytotoxicity in isolated hepatocytes can be attributed to cyanide formation and oxygen activation (20). The product of the one-electron reduction is a paramagnetic species which undergoes redox cycling with oxygen. This paramagnetic species can also undergo further reactions to produce various diamagnetic species. Although there is enough evidence to show that NP does not liberate cyanide per se, the reductive metabolites of NP do liberate cyanide. The photoproduct of NP, namely, [Fe (CN), H2012- can produce cyanide, but physiologically this pathway is not significant, as [Fe (CN), H2012- is not formed in uiuo. (14). The metabolism of NP produces various metabolites in addition to cyanide and hydrogen peroxide as by-products; these metabolic products are shown in Scheme 1 (14, 21-32). Liver is a major organ where the reductive metabolism of NP is expected to occur. Therefore we unooo3-9861/91

30

Copyright 0 All

rights

1991 of reproduction

$3.00

by Academic Press, Inc. in any form reserved.

METABOLISM

[Fe (CN); NO?1 \ ,

RS-

“go; L ly

+ RSNO-

RS‘+ NO

4CN-

NO

Nitmprusside

? [Fe (CN)$-

OF NITROPRUSSIDE

[Fe (NO)z (WI

Fe’+ + [Fe (CN),14Femcyanide

OH- + ‘OH + Fe’+

SCHEME

1. Metabolism

of nitroprusside.

dertook this investigation to compare the metabolism NP in rat hepatocytes and human erythrocytes. MATERIALS

AND

of

METHODS

Hepatocytes were isolated from male Sprague-Dawley rats, by liver perfusion as described by Moldeus et al. (33). The cell viability was measured by the Trypan blue exclusion method, and the hepatocytes used in these studies were at least 80% viable. Rat liver microsomes were isolated in 150 mM KCl-50 mM Tris buffer, pH 7.4, as previously described (34). Rat liver mitochondria were prepared in 150 mM KCl50 mM Tris buffer, pH 7.4, by the method of Bellomo et al. (35). The supernatant of the first 100,OOOgcentrifugation, after removing mitochondria, was used as the cytosol fraction. Heat-treated hepatocytes and beat-treated microsomes and mitochondria were prepared by heating the respective samples (2 ml) under nitrogen in a boiling water bath for 10 min. The inhibition of microsomal cytochrome P450 with carbon monoxide was performed by bubbling the sample with carbon monoxide for about 10 min. The inhibition with metyrapone was carried out by treating microsomes with metyrapone (1-5 mMj for 30 min. The inhibition of cytochrome P450 reductase with the antibody (produced against pig liver cytochrome P450 reductasej was performed by incubating the rat liver microsomes for about 30 min prior to the addition of NP. The control sample contained preimmune sera instead of the antibody. Inhibitors such as allopurinol, dicumarol, cyanide, and 4-methyl pyrazole were added to hepatocytes and incubated for 5-10 min prior to the addition of NP. Human erythrocytes were prepared from fresh blood obtained from Canadian Red Cross. The blood was first centrifuged at 1OOOgfor 5 min, and the sediment was washed three times with isotonic saline at 4OC. The cells were suspended in 144 mM NaCl and 10 mM phosphate buffer, pH 7.4. The supernatant of the first centrifugation was used as plasma.

31

Diluted erythrocyte cytosol was prepared by diluting the cells with 4 vol of water and centrifuging at 20,OOOgfor 20 min. The supernatant was used as erythrocyte cytosol. In experiments using DCNB (1 mMj, erythrocytes were treated with DCNB for 1 h before they were treated with NP. The protease treatment of erythrocytes or diluted erythrocyte cytosol were performed by incubating the respective samples with protease (25 mg/ml) for 2 h at room temperature. The protein content was determined by the method of Lowry et al. (36). Cyanide concentrations were measured spectrophotometrically by using the reaction of cyanide with methemoglobin to form a cyanohemoglobin complex (37). The GSH and GSSG contents in hepatocytes were measured by the method of Reed et al, using HPLC (38). The determination of total GSH and GSSG in hepatocytes and the medium were carried out on the supernatant after precipitation of the protein with metaphophoric acid (final concentration 5%). The hepatocyte suspension was centrifuged at 200g for 1 min to remove the intact cells. Oxygen consumption by various samples was measured by using a Clark oxygen electrode (Yellow Springs Instruments, Model 5300) at 37’C. Samples for ESR measurements were prepared by the addition of NP (usually 6 mM) to microsomes containing externally added NADPH (1 mM), or erythrocytes, erythrocyte cytosol, hypoxanthine (0.081 mM)/xanthine oxidase (0.09 units/ml), or a thiol, previously bubbled with nitrogen for 10 min. The sample was aspirated into a flat cell as described by Mason (39). Hepatocyte samples were prepared by treating the cells (usually 6 X 10s cells/ml) with NP; the sample became anoxic by the time ESR measurement was started. The ESR spectra were recorded immediately after the addition of NP and the spectra were scanned several times (usually 8-min scans). Spin-trapping experiments were done by the addition of DMPO to a system containing NP and a reductant at pH 7.4. ESR measurements were performed using a TE cavity, in dark, in an aqueous flat-cell (10 mm wide), at room temperature using a Varian E-6 ESR spectrometer. Sodium nitroprusside was obtained from Fisher Scientific, and collagenase (Clostridium hitolyticumj, NADPH, and catalase were obtained from Boehringer-Mannheim Biochemicals, and 1,3-bis (2chloroethyl)l-nitrosourea was a gift from Bristol Lab, Canada. The antibody for the pig liver cytochrome P450 reductase and the preimmune sera were gifts from Dr. B. S. Masters, Wisconsin Medical College (Milwaukee, WI). Superoxide dismutase, xanthine oxidase, protease (from Streptomyces griesusj type XIV, 5,5’-dimethyl-l-pyrroline N-oxide, paraquat, and diethylenetriamine pentaacetic acid, allopurinol, and dicumarol were obtained from Sigma. 4-Methyl pyrazole was obtained from Aldrich, and HPLC grade solvents were obtained from Caledon, Ontario. DMPO was distilled under vacuum before use, and it was stored at -20°C under nitrogen. All other reagents used in these studies were of good quality commercially available.

RESULTS

Figure 1A shows the formation of a relatively broad three-line ESR spectrum in the presence of NP, rat liver microsomes, and NADPH. In the absence of NADPH (Fig. lB), microsomes (Fig. lC), or NP (Fig. 1D) in the assay system, the three-line spectrum was not detected. Heat-treated microsomes (Fig. 1E) did not generate the three-line spectrum. The large line width, the hyperfine coupling constant of about 14.25 G, and the g-value of 2.028 are consistent with the reported value for the metal nitroxide radical (21). Metabolic studies, by monitoring the stability of the MNO’ radical showed that the signal amplitude of the ESR spectrum in the presence of microsomes and NADPH did not diminish significantly 60 min after the addition of NP. The effect of ascorbate on the microsomal reduction of NP is shown in Fig. 2. In the complete system containing

32

RAO, ELGUINDI,

AND

O’BRIEN

D

E F FIG. 1. A, the ESR spectrum of the MN0 * radical generated in a nitrogen-bubbled system of NP (6 mM), microsomes (3.4 mg/ml), and NADPH (1 mM) in 150 mM KCl-50 mM Tris buffer, pH 7.4. B, same as in A, but without NADPH. C, same as in A, but without microsomes. D, same as in A, but without NP. E, same as in A, but with heat-treated microsomes. The instrumental conditions were: 20-mW microwave power, 12.5 G/min scan rate, for A, 0.3-s time constant, 1-G modulation amplitude, and 4 X lo3 receiver gain. For B-E, l-s time constant, 5-G modulation amplitude, and 2 X ld receiver gain. uN (MNO’) = 14.25 G. A blip in the spectrum E is due to instrumental aberration.

NP, NADPH, and microsomes and the buffer containing DTPA (Fig. 2A), or the complete system containing ascorbate, the MN0 . radical was formed as expected, but no ascorbyl radical was detected (Fig. 2B). Ascorbate also reduced NP to generate the MNO. radical and the ascorbyl radical (aH = 1.8 G, g = 2.004; Fig. 20. In the presence of microsomes without NADPH, ascorbate also reduced NP to produce the MN0 . and ascorbyl radicals (Fig. 2D). Figure 2E shows that NADPH by itself did not reduce NP. Similarly NADH also did not reduce NP (data not shown). Ascorbate did not autoxidize to the ascorbyl radical in the presence of a chelating agent DTPA (Fig. 2F), thereby showing that the ascorbyl radical detected in the presence of microsomes, NP, and ascorbate is due to the oxidation of ascorbate by NP. Rat liver microsomes treated with carbon monoxide or metyrapone (1 or 5 mM) to inhibit cytochrome P450 also reduced NP to the MNO. radical in the presence of NADPH. The antibody to cytochrome P450 reductase (5 mg/ml) incubated with microsomes for 30 min inhibited the formation of the MN0 * radical; the amplitude of the ESR spectrum of the MN0 * radical was about 50% that of the control sample. Tables I and II summarize the formation and detection of the MN0 * radical under various conditions, by cell organelles and cells, respectively.

FIG. 2. A, the ESR spectrum of the MNO. radical generated in a nitrogen-bubbled system of NP (6 mM), microsomes (3.4 mg/ml), NADPH (1 mM), and DTPA (1 mM) in 150 mM KCl-50 mM Tris buffer, pH 7.4. B, same as in A, but with ascorbate (0.5 mM). C, same as in B, but without microsomes and NADPH. D, same as in B, but without NADPH. E, same as in A, but without microsomes. F, ascorbate (0.5 mM) and DTPA (1 mM) in the above buffer. The instrumental conditions were: For A, same as for Fig. 1A; for B-F, 20-mW microwave power; for B, 25 G/min scan rate; for C-F, 12.5 G/min scan rate; for B, 0.1-s time constant; for C-F, l-s time constant; for B, 1-G modulation amplitude; for C-F, 0.5-G modulation amplitude, and for B, 2 X lo3 receiver gain; for C-F, 2 X 10’ receiver gain.

NP was also reduced by rat liver mitochondria (data not shown). Addition of NADH (1 mM), or respiratory substrates such as succinate enhanced the amplitude of the ESR spectrum of the MN0 * radical five times. Cy-

TABLE

I

The Generation of the MN0 * Radical by Cell Organelles and the Cell-Free Medium Cell organelle

+ NP (6 mM)

MN0

. radical formed

Microsomes/NADPH (1 mM) Microsomes/antibody (5 mg) Microsomes/carbon monoxide Microsomes/metyrapone (5 mM) Mitochondria Mitochondria + succinate Cytosol Cytosol + NADH (1 mM) Cytosol + NADPH (1 mM) Note. Microsomal protein, 3.4 mg/ml; mitochondrial ml; cytosolic protein, 1 mg/ml.

protein,

3 mg/

METABOLISM TABLE

OF NITROPRUSSIDE

II

The Generation of the MN0 * Radical by Rat Hepatocytes and Human Erythrocytes Cells + NP (6 mM)

MN0

Hepatocytes Hepatocytes + cyanide (10 mM) Hepatocytes + I-methyl pyrazole (2 mM) Hepatocytes + dicumarol (0.2 mM) Hepatocytes + allopurinol (2 mM) Hepatocytes + diethylmaleate (5 mM) Heat-treated hepatocytes Blood plasma Erythrocytes Erythrocyte cytosol Erythrocytes + protease (25 mg/ml) Erythrocyte cytosol + protease (25 mg/ml) Erythrocytes + carbon monoxide Erythrocytes t cyanide (1 mM) Erythrocytes + 1,2-dichloro 4-nitrobenzene (1 mM) Note. Hepatocytes, 6 X 10s cells/ml; erythrocytes ml; erythrocyte cytosol (protein), 50 mg/ml.

* radical formed + +

+ + +

+ + + + +

(protein),

340 mg/

tosol, or cytosol treated with NADH or NADPH, or heattreated mitochondria did not produce the MN0 - radical. Isolated hepatocytes also reduced NP to the MN0 * radical (Fig. 3A), but the MN0 * was not detected in the absence of NP (Fig. 3B) or hepatocytes (Fig. 3C) or in the presence of heat-treated hepatocytes (Fig. 3D). The ascorbyl radical was not detected in the hepatocytes and NP incubation mixture (Fig. 3). The signal amplitude of the ESR spectrum of the MN0 * radical did not significantly diminish over an extended period of time, e.g., in hepatocytes (6 X lo6 cells/ml), NP (6 mM) was reduced to the MNO. radical, and the signal amplitude did not diminish significantly even 90 min after the addition of NP. However, in the ascorbate (0.5 mM) reduction of NP (6 mM) in 150 mM KCl-50 mM Tris buffer, pH 7.4, the MN0 * radical signal decayed significantly after about 60 min. Hepatocytes treated with diethyl maleate (5 mM) to deplete glutathione also reduced NP to the MN0 . radical with the same signal amplitude as that of the control sample. Hepatocytes (6 X lo6 cells/ml) treated for 5-10 min with allopurinol (2 mM) to inhibit xanthine oxidase, dicumarol (0.2 mM) to inhibit DT-diaphorase, 4-methyl pyrazole (2 mM) to inhibit alcohol dehydrogenase, or KCN (10 mM) to inhibit the mitochondrial electron transport did not inhibit the generation of the MNO. radical (Table II). The reduction of NP by xanthine oxidase and hypoxanthine under anoxic conditions also lead to the formation of the MNO. radical. In the presence of oxygen this radical was not detected (data not shown). These results are in accordance with the report of Misra who also showed that NP can be reduced by xanthine oxidase (40).

33

Human erythrocytes treated with NP produced the MNO. radical; however, plasma did not reduce NP. Erythrocytes or diluted erythrocyte cytosol treated with protease for 2 h to hydrolyze the NP reducing activity of erythrocytes also did not produce the MN0 . radical. Prior treatment of erythrocytes for 1 h with DCNB, the GSHtransferase substrate to deplete GSH did not prevent the reduction of NP to the MN0 . radical. Erythrocytes bubbled with carbon monoxide to complex oxyhemoglobin also reduced NP to the MN0 * radical, but the amplitude of the ESR spectrum of the MNO- radical was about 50% of the amplitude of the MN0 * radical. Treatment of erythrocytes with cyanide to complex methemoglobin prevented the formation of the MN0 * radical. Therefore we believe that the MN0 . radical formed by the reduction of NP by erythrocytes involves hemoglobin and methemoglobin reductase, and not GSH. A qualitative comparison of the ability of rat hepatocytes and human erythrocytes to reduce NP was studied to understand which one of these two tissues is important in NP metabolism. NP reduced by hepatocytes (protein, 10 mg/ml) to the MN0 * radical had an ESR signal amplitude four times higher than that formed by a much higher concentration of erythrocytes (protein, 344 mg/ ml). This shows that the reduction of NP by human erythrocytes occurs much more slowly than in rat hepatocytes. Figure 4 shows the relative effect of different thiols on the reduction of NP. DTT, cysteamine, cysteine, and

FIG. 3. A, the ESR spectrum of the MNO. radical generated in a system of NP (6 mM), DTPA (1 mM), and hepatocytes (6 X lo6 cells/ ml) in Krebs-Hensleit buffer, pH 7.4. B, same as in A, but without NP. C, same as in A, but without hepatocytes. D, same as in A, but with heat-treated hepatocytes. The instrumental conditions were: 20-mW microwave power, 12.5 G/min scan rate, l-s time constant, 1-G modulation amplitude; and for A, 10 X 10s receiver gain; and for B-D, 2 X lo4 receiver gam.

34

RAO, ELGUINDI,

F ---~----~G FIG. 4. A, the ESR spectrum of the MN0 * radical generated in a nitrogen bubbled system of NP (6 mM), DTPA (1 mM), and DTT (2 mM) in 150 mM KCl-50 mM Tris buffer, pH 7.4. B, same as in A, but with cysteamine (2 mM) and without DTT. C, same as in A, but with cysteine (2 mM) and without DTT. D, same as in A, but with penicillamine (2 mM) and without DTT. E, same as in A, but with N-acetyl cysteine (2 mM) and without DTT. F, same as in A, but with diethyldithiocarbamate (2 mM) and without DTT. G, same as in A, but with bovine serum albumin (25 mg/ml) and without DTT. The instrumental conditions were: 20-mW microwave power, 12.5 G/min scan rate, l-s time constant; for A-C, 1-G modulation amplitude; for D-G, 5-G modulation amplitude; for A, 4 X lo3 receiver gain; for B, C and E-G, 10 X lo3 receiver gain; and for D, 2 X 10’ receiver gain.

penicillamine reduced NP to the MN0 * radical; DTT is a relatively stronger reducing agent, and reduced NP more rapidly than other thiols (Fig. 4A). The amplitude of the MN0 . radical in the presence of DTT was twice as large as those detected in the presence of cysteamine and cysteine (Figs. 4B and 4C). The amplitude in the presence of penicillamine was very much lower (Fig. 4D). IV-Acetyl cysteine, diethyldithiocarbamate, and bovine serum albumin (protein thiol) did not generate detectable levels of the MN0 * radical (Figs. 4E, 4F and 4G). The reduction of NP by glutathione with or without DTPA first resulted in the formation of a broad singlet (g = 2.032; Fig. 5A) which then slowly disappeared resulting in the formation of the MN0 . radical (Figs. 5B-5D). The rate of formation of the MN0 * radical was dependent on the buffer component. For example, in the presence of phosphate (0.1 M) buffer, the MNO. radical was produced much more slowly than in Hepes or Tris buffer (data not shown). A similar singlet species was formed in the cysteine reduction of NP, the singlet first formed was slowly replaced by the MN0 * radical. This singlet species was also formed

AND

O’BRIEN

in the cysteamine reduction of NP. The singlet species could not be characterized due to the lack of structure in the ESR spectrum (Fig. 5A), although it is tempting to find a relationship with the formation of the MN0 . radical. The hyperfine coupling constants obtained for the MNO. formed from the reduction of NP with various thiol compounds ranged from 14.2 G to 14.8 G. Dopamine reduction of NP to the MN0 + radical resulted in the oxidation of dopamine to the corresponding o-semiquinone radical. Catechol, norepinephrine, and phydroquinone also reduced NP, resulting in the formation of the MN0 . radical and the corresponding semiquinone radicals (data not shown). A significant oxygen consumption by the solutions containing NP and catechol, NP and norepinephrine, and NP and dopamine was also observed; this oxygen consumption was inhibited by the addition of catalase. The reduction of NP by sodium borohydride or dithionite also produced the MN0 * radical (data not shown), and no singlet species was formed in any of these systems. The reductive metabolism of NP is known to produce thiyl, superoxide, and hydroxyl radicals (Scheme l), and we therefore attempted to trap these radicals using DMPO. Spin-trapping in the presence of NP and thiol compounds such as cysteine and GSH resulted in the production of a DMPO-thiyl adduct. Addition of catalase did not diminish the ESR spectrum nor did the addition

-WE--

FIG. 5. A, the ESR spectrum of the MNO. radical generated in a nitrogen-bubbled system of NP (6 mM) and GSH (6 mM) in 100 mM Hepes buffer, pH 7.4. B, same as in A, but second scan. C, same as in A, but third scan. D, same as in A, but fourth scan, after a delay of about 3 min after the third scan. The instrumental conditions were: 20mW microwave power, 12.5 G/min scan rate, l-s time constant; for AC, 5-G modulation amplitude; for D, 1-G modulation amplitude; and for A-C, 2 X 10’ receiver gain; for D, 2.5 X lo3 receiver gain.

METABOLISM

35

OF NITROPRUSSIDE

of dimethyl sulfoxide produce a DMPO carbon-centered radical. Nitroprusside by itself did not catalyze the Fenton reaction in the presence of hydrogen peroxide, as we did not detect DMPO/OH adduct. In an earlier report (20), we showed that NP induced cell death in hepatocytes. The concentration of NP required to induce more than 50% cell death in 2 h was 10 mM, and total cell death occurred when the concentration of NP was 15 mM. Substantial oxidation of hepatocytic GSH to GSSG occurred at 5 min (results not shown). The total amount of cyanide formed by the reduction of NP by various thiols was measured, and the results are shown in Table III. Thiols which reduce NP liberate cyanide from NP, and thiols such as N-acetyl cysteine and bovine serum albumin (protein thiol) which did not reduce NP did not generate cyanide (Table III). DISCUSSION

Figure 1 shows that the NADPH-dependent microsomal reductase reduced NP to the MN0 * radical. A gvalue of 2.028, and the hyperfine coupling constant, uN = 14.25 G, suggests that this paramagnetic species is the MN0 * (metal-nitroxyl) radical. This species may be characterized as [Fe (CN)SNO]3- or [Fe (CN)4NO]2-, but Glidewell and Johnson suggested that it is the latter species (22). Microsomes with NADPH in the presence of oxygen, or in the presence of heat-treated microsomes with NADPH did not generate the MN0 * radical, but in the absence of oxygen, microsomes with NADPH produced the MN0 * radical, suggesting that a reductase is responsible for the reduction of NP and not superoxide. Ascorbate also reduced NP to produce the MN0 * and the ascorbyl radicals. However, when ascorbate, microsomes, and NADPH were together only MN0 * radical was formed (Fig. 2B). This suggests that ascorbate is less effective than cytochrome P450 reductase (see later) in reducing NP. In the presence of microsomes, the reductase is the major reducing factor in the reduction of NP. The lack of detection of the ascorbyl radical in the above system is not due to microsomes, because the ascorbyl radical was detected in the presence of NP, microsomes, DTPA, and without NADPH (Fig. 2D). The MNO. radical was not detected in the presence of NP and NADPH (Fig. 2E). The inhibition of microsomal cytochrome P450 by carbon monoxide or metyrapone did not inhibit the formation of the MNO. radical, but the antibody to cytochrome P450 reductase inhibited the formation of the MN0 * radical. This suggested that the microsomal NADPH-cytochrome P450 reductase, and not cytochrome P450 is involved in the reduction of NP to the MNO. radical. The mitochondrial reduction of NP may be due to respiratory reductases, because addition of substrates such as succinate enhanced the rate of the formation of the MN0 * radical. Cofactors such as NADH also enhanced

TABLE

III

Liberation of Cyanide from Nitroprusside by Thiol Compounds and Microsomes/NADPH Thiols Dithiothreitol Cysteine Cysteamine Penicillamine Glutathione N-Acetyl-cysteine Bovine serum albumin Microsomes/NADPH

Total CN-liberated

(PM)

113f 3 128 f 7 142 t 8 113f 9 63+ 8 17+ 6 17k 8 107 +_ 13

Note. NP, 100 HIM; thiol, 300 pM; bovine serum albumin, microsomes, 850 pg and NADPH (1 mM)

10 mg/ml;

the amplitude of the ESR spectrum of the MN0 . radical, implying that the membrane bound NADH: b5 reductase is involved in the reduction of NP. Other nitro compounds such as nitrofurantoin and nifurtimox are known to be reduced by this mitochondrial enzyme (23). This result shows that mitochondrial reduction of NP is important in the toxic effect of NP, since the production of cyanide by the reductive metabolism of NP can block cellular respiration. In hepatocytes, the reductases of the endoplasmic reticula and mitochondria seem to be important reducing enzymes in the reduction of NP (Fig. 3A), because heatdenatured hepatocytes did not reduce NP to the MN0 . radical (Fig. 3D). Hepatocytes contain GSH (about 50 nM/l@ cells), but this concentration is not sufficient to compete with the membrane bound reductases for the reduction of NP. We also point out that the NP reduction by GSH initially results in an ESR singlet (g = 2.032) which is slowly replaced by the MN0 * radical (Fig. 5). This was not observed when NP was reduced by hepatocytes (Fig. 3). Dicumarol (an inhibitor of DT-diaphorase), or allopurinol (an inhibitor of xanthine oxidase), or 4-methyl pyrazole (an inhibitor of alcohol dehydrogenase) did not inhibit the NP reduction, suggesting that the soluble enzymes such as DT-diaphorase, xanthine oxidase, or alcohol dehydrogenase were not effective in hepatocytes, although isolated xanthine oxidase can reduce NP. Blocking the mitochondrial electron transport chain by cyanide also had no effect on NP reduction by hepatocytes. Human erythrocytes also reduce NP to the MNO. radical, but rat hepatocytes had a significantly higher concentration of this radical than erythrocytes at comparable NP to protein concentrations (see results). This suggested a high steady-state level of this radical in hepatocytes, probably due to a higher rate of formation. As GSH depletion did not affect the steady-state level of the MNO’ radical, and no ascorbyl radical was detected, suggests that NADH, methemoglobin reductase and hemoglobin are involved in NP reduction.

36

RAO. ELGUINDI,

Scheme 1 shows the generation of cyanide, when [Fe (CN), N013- is converted to [Fe (CN), NO]‘-, and second, when the latter is converted to [Fe (NO), (SR,)]-. The superoxide radical is also formed at two different sites, during the reoxidation of both [Fe (CN), N0’J3- and [Fe (CN), NO]‘-. The superoxide radical is generated in a third step when the disulfide radical anion is oxidized to disulfide by oxygen (41,42). This step is also important, because the thiyl radical formed at two different steps in the metabolism of NP can lead to the formation of disulfide radical anion (Scheme 1). Although the involvement of thiol compounds in the reduction of NP in liver is less important, thiol compounds may be important in the metabolism of the MN0 * radical (Scheme 1). Other reducing agents such as ascorbate found in the cell are less likely to be involved in the reduction of NP when mitochondria or cytochrome P450 reductase is present. A relative comparison of NP reduction by various thiol compounds suggests that DTT reduced NP more readily than cysteine and cysteamine, but penicillamine was a poor reducing agent (Fig. 4), N-Acetyl cysteine, diethyldithiocarbamate, and bovine serum albumin did not reduce NP. Determination of cyanide liberated from the reduced metabolites of NP also corresponded well with those thiol compounds which reduced NP (Table III). Thiols that do not reduce NP liberated very little cyanide. Thus thiols are important in the subsequent reactions of the metabolites of NP, after the initial reduction of NP by the cytochrome P450 reductase and mitochondria. The g = 2.032 singlet species appears to be a direct result of NP reduction by thiols, but it is more stable in cysteine or GSH reduction of NP (Fig. 5). We hoped to trap the thiyl, the superoxide, and the hydroxyl radical in the reductive metabolism of NP, but we were unsuccessful in trapping any of these radicals in microsomes. In the presence of NP and thiol compounds we were able to trap only thiyl radical. The lack of sensitivity of this spin adduct to the addition of catalase and dimethylsulfoxide suggests that this spin adduct is a DMPO-thiyl radical. It is known that quinols such as p-hydroquinone reduce NP (21). Therefore we studied the reduction of NP by catechols and catecholamines. All three catechols, namely, 1,2-dihydroxybenzene, norepinephrine, and dopamine reduce NP to the MN0 * radical. Oxygen consumption measurements showed the formation of the superoxide and hydrogen peroxide, suggesting the reduction of NP (data not shown). This result may be important in the kidney where there is a high concentration of catecholamines in chromaffin cells. Toxicity studies using rat hepatocytes showed significant NP-induced cell death (20). This is to be expected as NP is reduced to the MN0 . radical by hepatocytes, and this undergoes redox cycling to generate the superoxide radical, which can disproportionate to produce

AND

O’BRIEN

H202. The [Fe (CN), N013- formed after the reduction produces one mole-ion of CN- and [Fe (CN), N012- (14). The metal complexes are routinely employed to linebroaden the ESR signals due to the radicals present outside the cell on the assumption that the highly charged metal complexes do not cross the cell membrane (43). In this regard, it is interesting to note that Kreye and Reske have shown that NP passes through the erythrocyte membrane very slowly (44). In isolated hepatocytes, NP may also be passing through the cell membrane slowly, probably through an anion transport channel on the membrane, but NP reduction also occurs outside the cell, due to the reductases from some broken cells. In conclusion, endoplastic reticular NADPH:cytochrome P450 reductase and mitochondrial dehydrogenases in hepatocytes may be the major pathway for reducing NP to the MN0 * radical. A relative comparison shows that the MN0 . radical exists at a much higher concentration in liver than in erythrocytes. The toxicity of NP occurs in hepatocytes rather than in erythrocytes. This is further supported by the toxicity data of NP to hepatocytes (20). In human erythrocytes, NP is probably reduced to the MN0 . radical by hemoglobin (22) and methemoglobin reductase. REFERENCES 1. Johnson, C. C. (1928) Proc. Sot. Exp. Biol. Med. 26, 102. 2. Page, I. H. (1951) J. Amer. Med. Assoc. 147, 1311. 3. Currey, C. L., and Hosten, A. 0. (1975) J. Amer. Med. Assoc. 232, 1367-1369. 4. Schiffmann, H., and Fuchs, P. (1966) Actu Anaesthesiol. Scan&. 23, 704-709. 5. Siegel, P., Maraca, P. P., and Green, J. R. (1971) Brit. J. Anoesth. 43, 790-795. 6. Franciosa, J. A., Guiha, N. H., Limas, C. J., Rodriguera, E., and Cohn, J. N. (1972) Lancet 1.650-654. I. Ignarro, L. J., Adams, J. B., Horwitz, P. M., and Wood, K. S. (1986) J. Biol. Chem. 261,4997-5001. 8. Butler, A. R., Glidewell, C., Johnson, I. L., and McIntosh, A. S. (1987) Znorg. Chim. Acta. 138, 159-162. 9. Ignarro, L. J., Degnan, J. N., Baricos, W. H., Kadowitz, P. J., and Wolin, M. S. (1982) Biochim. Biophys. Acta 718, 49-59. 10. Craven, P. A., and De Rubertis, F. R. (1983) B&him. Biophys. Actu 746,310-319. 11. Rapoport, R. M., Draznin, M. B., and Murad, F. (1982) Proc. Nutl. Acad. Sci. USA 79,6470-6474. 12. Rapoport, R. M., Draznin, M. B., and Murad, F. (1983) Nature (London) 306,174-176. 13. Fiscus, R. R., Rapoport, R. M., and Murad, F. (1984) J. Cyclic Nucleotide Protein Phospho. Res. 9, 415-425. 14. Butler, A. R., and Glidewell, C. (1987) C&m. Sot. Reu. 16, 361380. 15. Davies, D. W., Kadar, D., Steward, D. J., and Munro, I. R. (1975) Canad. Anaesth. Sot. J. 22, 547-552. 16. Merifield, A. J., and Blundell, M. D. (1974) Br. J. Anaesth. 46,324. 17. Vesey, C. J., Cole, P. V., Linnell, J. C., and Wilson. (1974) Brit. Med. J. 2, 140-142. 18. Vesey, C. J., Cole, P. V., and Simpson, P. J. (1976) Brit. J. Anuesth. 48,651-660.

METABOLISM

37

OF NITROPRUSSIDE

19. Bisset, W. 1. K., Butler, A. R., Glidewell, C., and Reglinski, J. R. (1981) &it. J. Anoesth. 53, 1015-1018. 20. Elguindi, S., and O’Brien, P. J. (1990) in Biological Reactive Intermediates III (Snyder, M., Ed.), in press. 21. Mulvey, D., and Waters, W. A. (1975) J. Chem. Sot. Dalton Trans. 951-959. 22. Glidewell, C., and Johnson, I. L. (1987) Znorg. Chim. Acta 132,145147. 23. Moreno, S. N. J., Mason, R. P., and Docampo, R. (1984) J. Biol. Chem. 259,6298-6305. 24. Bisset, W. I. K., Burdon, M. G., Butler, A. R., Glidewell, C., and Reglinski, J. R. (1981) J. Chem. Res. (S) 299, (M) 3501. 25. Wolfe, S. K., and Swinehart, J. H. (1973) Znorg. Chen. 14,10491053. 26. Morando, P. J., Borghi, E. B., de Schteingart, L. M., and Blesa, M. A. (1981) J. Chem. Sot. Dalton Trans. 435-440. 27. Johnson, M. D.. and Wilkins, R. G. (1984) Znorg. Chem. 23, 231235. 28. Ignarro, L. J., Lipton, H., Edwards, J. C., Baricos, W. H., Hyman, P. J., Kadowitz, P. J., and Grietter, C. A. (1981) J. Pharm. Exp. Ther. 218, 739-749. 29. Giniyatullin, N. G., and Kargin, Y. M. (1976) Zzu. Vysch. Uchben. Zaued. Khim. Tekhnol. 19, 1668-1671. 30. Gawron, O., and Fernando, J. (1961) J. Amer. Chem. Sot. 83,29062908. 31. Antal, K., Banyal, I., and Beck, M. T. (1985) J. Chem. Sot. Dalton Trans. 1191-1193.

32. Andrade, 650.

C., and Swinehart,

J. H. (1972) Znorg. Chem. 11, 648-

33. Moldeus, P., Hogherg, J., and Orrenius, S. (1978) in Methods in Enzymology (Hoffee, P. A., and Jones, M. E., Eds.), Vol. 51, pp. 6071, Academic Press, New York. 34. Perez-Reyes, E., Kalyanaraman, Pharmacol. 17, 239-244.

B., and Mason, R. P. (1980) Mol.

35. Bellomo, G., Martino, A., Richelmi, P., Moore, G., Jewell, S., and Orrenius, S. (1984) Eur. J. Biochem. 140, l-6. 36. Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, (1951) J. Biol. Chem. 193,265-275. 37. Kiese, M. (1974) in Methemoglobinemia: pp. 3-7, CRC Press, Cleveland, OH.

A Comprehensive

R. J.

Treatise,

38. Reed, D. J., Babson, J. R., Beatty, P. W., Brodie, A. E., Ellis, W. W., and Potter, D. W. (1980) Anal. Biochem. 106, 55-62. 39. Mason, R. P. (1984) in Methods in Enzymology (Packer, L., Ed.), Vol. 105, pp. 416-422, Academic Press, Orlando, FL. 40. Misra, H. P. (1984) J. Biol. Chem. 259, 12,678-12,684. 41. Ross, D., Norbeck, 15,028-15,032.

K., and Moldeus, P. (1985) J. Biol. Chem. 260,

42. Rao, D. N. R., Fischer, V., and Mason, R. P. (1990) J. Biol. Chem. 265,844-847. 43. Berg, S. P., and Nesbitt, D. M. (1979) Biochim. Biophys. Acta. 548, 608-615. 44. Kreye, V. A. W., and Reske, S. N. (1982) Naunyn-Schmeidberg’s Arch. Pharmacol. 320, 260-265.

Reductive metabolism of nitroprusside in rat hepatocytes and human erythrocytes.

The metabolism of nitroprusside by hepatocytes or subcellular fractions involves a one-electron reduction of nitroprusside to the corresponding metal-...
896KB Sizes 0 Downloads 0 Views