Chem. Res. Toxicol. 1988, I , 148-151
148
Art i c 1es N-Chlorination of Phenytoin by Myeloperoxidase to a Reactive Metabolitet Jack Uetrecht*J and Nasir Zahids Faculties of Pharmacy and Medicine, University of Toronto and Sunnybrook Medical Centre, Toronto, Ontario, M 5 S 1 A l Canada Received January 18, 1988
Several types of phenytoin toxicity appear t o involve leukocytes. We h a d previously demonstrated t h a t other drugs were metabolized t o reactive metabolites by activated neutrophils and monocytes or t h e combination of myeloperoxidase (MPO) a n d hydrogen peroxide. In this study we found that phenytoin was chlorinated by MPO/H202/C1- t o N,N’-dichlorophenytoin which is chemically reactive. Failure t o demonstrate t h a t activated neutrophils also formed N,N’-dichlorophenytoin appeared t o be due t o t h e rapid reaction of N,N’-dichlorophenytoin with neutrophils. W e were able t o demonstrate t h a t phenytoin covalently bound t o albumin in the presence of MPO/H202/C1- and t o neutrophils, b u t only if the cells were activated. Such activation leads t o t h e release of MPO a n d t h e generation of H2O2. We, therefore, speculate that t h e toxicity of phenytoin may be due t o the formation of N,”-dichlorophenytoin by activated neutrophils or monocytes.
Use of the anticonvulsant, phenytoin, is associated with several types of adverse reactions including birth defects ( 1 , 2 ) ,pseudolymphoma ( 3 , 4 ) ,drug-induced lupus ( 5 , 6 ) , generalized hypersensitivity reactions (7-9), and neutropenia (10-12). It has been proposed that a reactive arene oxide metabolite is responsible for the toxicity of phenytoin (13). Mephenytoin has a very similar structure and profile of toxicity as phenytoin (14). Unlike phenytoin, mephenytoin has a chiral carbon and only the S enantiomer is metabolized to a phenol (presumably through an arene oxide intermediate). Therefore, it was proposed that the S enantiomer is responsible for the toxicity of mephenytoin. However, when this hypothesis was tested, it was found that the R enantiomer was more toxic than the S enantiomer (15). Moreover, the R enantiomer undergoes extensive N-demethylation and the N-demethylated derivative was more toxic than the parent drug. This suggested that the nitrogen might be involved in the mechanism of toxicity. Further evidence against the arene oxide hypothesis is that there are several other anticonvulsants, such as ethosuximide and trimethadione, that have heterocyclic structures similar to phenytoin and very similar profiles of toxicity (161,but they do not have a phenyl ring which could be metabolized to an arene oxide. Several of the toxicities mentioned above involve leukocytes. Certainly neutropenia and pseudolymphoma inPortions of this work were presented a t an annual meeting of the American Society for Pharmacology and Experimental Therapeutics and appeared in the Pharmacologist 29, 187 (1987). Faculties of Pharmacy and Medicine. Faculty of Pharmacy. +
*
0893-228x/88/2701-0148$01.50/0
volve neutrophils and lymphocytes, respectively. Presumably, idiosyncratic reactions and lupus are mediated through the immune system and also involve leukocytes. We have previously demonstrated that other drugs such as procainamide, dapsone, and sulfadiazine, which are associated with drug-induced lupus, agranulocytosis, and idiosyncratic reactions, are metabolized to reactive hydroxylamine metabolites by leukocytes (17-19). Two of the leukocyte cell types, neutrophils and monocytes, contain myeloperoxidase ( M P 0 ) l and a hydrogen peroxide generating system. It appears as if the M P O / H 2 0 2combination is responsible for the formation of the hydroxylamine metabolites by neutrophils and monocytes. Evidence for the role of M P O / H 2 0 2in the formation of hydroxylamines by leukocytes includes: formation of the same hydroxylamines by purified MPO/H,O,, inhibition by azide (which inhibits MPO) and catalase (which destroys H202),and lack of inhibition by prostaglandin synthetase inhibitors. Formation of hydroxylamine metabolites by leukocytes only occurred when the cells were activated by phorbol myristate acetate (PMA) or opsonized zymosan. This activates the hydrogen peroxide generating system and causes the release of MPO. This process is known as the respiratory burst. In addition to oxidation, in the presence of chloride ion, this system can also chlorinate drugs (20). We, therefore, set out to determine if phenytoin could also be metabolized to a reactive metabolite by neutrophils or MPO/H,Oz. Metabolites which involved the hydantoin nitrogen would be of special interest. . Abbreviations: MPO, myeloperoxidase; PhlA phorbol myristate acetate
0 1988 American Chemical Society
N - Chlorination of Phenytoin by Myeloperoxidase
Materlals and Methods Synthesis of N,N‘-Dichlorophenytoin. Phenytoin (2.52 g, 0.01 mol; Sigma Chemical Company, St. Louis, MO) was added to sodium hypochlorite (30 mL of a 5% solution, Aldrich Chemical Company, Milwaukee, WI) over a period of about 5 min. After 10 min the turbid solution was extracted with three 20-mL portions of ethyl acetate. The extract was dried over MgS04 and the solvent removed with a rotary evaporator. The residue was recrystallized from ethyl acetate/hexane (51,v/v). The yield was 65%. The product was a white crystallinesolid with a melting point of 150-154 OC (with sublimation). It contained 5% phenytoin by HPLC. Anal. Calcd: C, 56.1; H, 3.1; N, 8.7; 0, 10.0; C1, 22.1. Found: C, 56.8; H, 3.4; N, 9.4; 0, 7.4; C1, 23.5. IR: 1750,1795 cm-’; the N-H stretch between 3200 and 3600 cm-’ found in phenytoin is absent. MS: m / z [relative intensity] 251 (M - 2C1+ 1) [loo]; 287 (M - C1 + 1) [lo]; 289 (C1 isotope of 287) [3.5]; 321 (M + 1) [1.9]; 323 (C1 isotope of 321) [1.2]; 325 (C1 isotope of 321) [0.2];349 (M + 29) [0.2]. This is very similar to the relative peak abundance of the dechlorinated ions produced under the same conditions from N,”-dichloro-5,5-dimethylhydantoin (Aldrich Chemical Co., Milwaukee, WI). N-Nf-Dichlorophenytoin was reduced to phenytoin by ascorbic acid, N-acetylcysteine, or sodium borohydride. Analytical Details. HPLC was performed by using a Beckman llOB pump and a Shimadzu SPD-6AV absorbance detector at a wavelength of 240 nm (Shimadzu Corp, Kyoto, Japan). The column was a 15-cm Jones column packed with 5-pm Spherisorb ODS-2 (Jones Chromatography Limited, Llanbradach, U.K.). The solvent was water, acetonitrile, and acetic acid (59:40:1, v/v) at a flow rate of 1 mL/min. Under these conditions the retention time of phenytoin was 4.2 min, and that of N&’-dichlorophenytoin was 12.3 min. The retention time of other known metabolites of phenytoin under these conditions were as follows: p hydroxyphenytoin (Sigma Chemical Co.), 2.05 min; m-hydroxyphenytoin (Sigma Chemical Co.), 2.14 min. Reaction of N,N’-Dichlorophenytoin with Neutrophils, Tryptophan, and DNA. N,N’-Dichlorophenytoin (0.04 mM) was incubated with neutrophils (0.75 X lo6 cells in 0.5 mL of Hanks buffer), tryptophan (0.04 mM), or DNA (1.2 mg/mL, type I, Sigma Chemical Co.) in pH 7 phosphate buffer at 37 “C and the rate of disappearence of N&’-dichlorophenytoin determined by analyzing the solution by HPLC as a function of time. Enzymatic Chlorination of Phenytoin. MPO (0.25 unit in 5 pL of phosphate buffer, Alpha Therapeutic Corporation, Los Angeles, CA) was added to 185 pL of pH 6 phosphate buffer containing NaCl(l50 mM). Phenytoin (5 pL of a 4 mM solution in ethanol) was added to the enzyme solution and vortexed. The reaction was started by adding hydrogen peroxide (5 pL to produce a final concentration of 0.4 mM). After incubation, the product was assayed by injecting 15 pL into the HPLC. The identity of the product was determined by its retention time on HPLC (a mixture of metabolite and synthetic N,N’-dichlorophenytoin produced only one peak) and its reaction with ascorbate and N-acetyl cysteine to give phenytoin. The requirement of C1- for the formation of the metabolite is also consistent with the proposed N,N’-dichlorophenytoin structure. Covalent Binding. [3H]Phenytoin (1.0 mCi in 10 pL of ethanol, 99% pure by TLC, labeled on the para position of the phenyl ring, New England Nuclear, Boston, MA), MPO (0.25unit), and H202( 5 pL of 16 mM to give a final concentration of 0.4 mM) were added to 200 pL of phosphate buffered saline (0.1 M phosphate, pH 6.0) and incubated at 37 “C for 10 min. Albumin (0.2 mg, Sigma, Bovine fraction V) was added and the solution incubated for a further 20 min. Acetone (5 mL) was added and the mixture placed on ice for 10 min to allow precipitation. The mixture was centrifuged, and, after the liquid was removed, the protein was redissolved in 100 pL of water. The protein was again precipitated with acetone and redissolved in water three times. The radioactivity in the final wash was equal to background. Parallel control incubations were performed in which various factors were omitted. The protein was dried with nitrogen and again dissolved in water (200 pL). Half of the protein solution was dissolved in Aquasol (Beckman) and counted in a scintillation counter. The protein in the other half of the protein solution was
Chem. Res. Toxicol., Vol. 1, No. 3, 1988 149 80
1
60
1
40
-
20
-
4-
” . 5
7
6
a
PH
Figure 1. PH dependence of phenytoin chlorination by MPO/H2O2/C1-. Conditions: MPO, 2.5 units/mL; H202,0.4 m M chloride, 150 mM; phenytoin, 0.1 mM; incubation time, 4 min. The data are the mean f SE from four experiments. quantified by using absorbance at 280 nm and comparison with an albumin standard curve. Neutrophils were isolated from human blood by use of differential centrifugation in lymphocyte separation medium (Litton Bionetics, Charleston, SC) as described previously (19). The cells had a viability of greater than 95% as determined by Trypan-Blue exclusion. [3H]Phenytoin (1.0 mCi in 10 pL of ethanol) was added to a suspension of neutrophils in Hanks buffer (0.5 mL, 1.5 X lo6 cells/mL). The cells were activated by adding phorbol myristate acetate (20 ng in 10 pL of DMSO, Sigma Chemical Co.) and incubated at 37 OC for 45 min. The cells were then collected on a Millipore filter (type GS, Millipore, Bedford, MA) and washed with absolute ethanol until the washes contained only background radioactivity. The filter was placed in Aquasol and placed in the dark for 24 h and radioactivity determined in a scintillation counter.
Results N,N’-Dichlorophenytoin is reasonably stable as a solid. There was n o noticeable decomposition when stored for a month. Even in solution at p H 6 or 7, i t had a half-life of 5.5 h. It was rapidly reduced to phenytoin by ascorbic acid, N-acetylcysteine, and NaBH,. It reacted with tryptophan, and the half-life of disappearance of N,N’-dichlorophenytoin was 41 min (data not shown). Although the products of this reaction were not characterized, the products are presumed to be the same as those produced by N-chlorosuccinimide (21). N,”-Dichlorophenytoin also reacted with DNA with a half-life of disappearance of 37 min (data not shown). We were unable to detect the production of any N,N’dichlorophenytoin when phenytoin was incubated with activated neutrophils. However, when synthetic N,Nfdichlorophenytoin was incubated with neutrophils, i t disappeared with a half-life of approximately 30 s (data not shown). In contrast, when phenytoin was incubated with myeloperoxidase, hydrogen peroxide, and chloride it was converted to N,N’-dichlorophenytoin in good yield. The pH optimum for the conversion was about 6.5 (Figure 1). T h e time course of the conversion at p H 6 and 7 is shown in Figure 2. T h e time course was much slower at p H 7, even though a higher MPO concentration was used to make i t easier t o quantify the product accurately. Although N,N’-dichlorophenytoin had a long half-life in buffer, it is clear from Figure 2 that the half-life was much shorter under t h e conditions of t h e incubation with MPO/H202/C1-. T h e optimal hydrogen peroxide concentration was 0.4 m M as shown in Figure 3. Chloride ion was necessary for the reaction, and the yield increased linearly with chloride concentration until well past the physiological range. Above a chloride concentration of 750 mM, the yield N,Nf-dichlorophenytoin began t o decrease
Uetrecht and Zahid
150 Chem. Res. Toxicol., Vol. 1, No. 3, 1988 "" I
Table 11. Covalent Binding of [3H]Phenytoin t o Activated
I
Neutrophils
covalent binding, fmol/106cells'
conditions control" activatedb
11 f 1 177 f ISd
Control conditions: 1.0 mCi of [3H]phenytoinand 0.75 X lo6 neutrophils incubated at 37 "C for 45 min. *Activated by adding 20 ng of PMA. Values are the mean f SE from four experiments. d p < 0.001 using the students t test. 0 4 ' 0 10 I
'
I
20
I
'
.
30
"
,
40
" 50
'
I 60
Scheme I. Proposed Scheme for the Covalent Binding of Phenytoin t o Neutrophils
70
Time (min) CsH,
Figure 2. Rate of phenytoin chlorination by MPO/H202/C1-at pH 6 and 7. Conditions: MPO, 1.25 units/mL at pH 6 and 2.5 units/mL at pH 7; H,02, 0.4 mM; chloride, 150 mM; phenytoin, 0.1 mM. The data are the mean & SE from four experiments.
c
.c6"--;+ C6H5
6
H
~
H>,6 -I: ~
o
f )
0
0
I H
-1
I CI N,N'.Dichlorophenytoin
was only significant if the cells were activated.
Dlscussion
0.0
0.2
0.4
0.6
0.8
1.0
Hydrogen Peroxide (mM)
Figure 3. H202concentration dependence of phenytoin chlorination by MPO/H202/C1-. Conditions: pH 6; MPO, 1.25 units/mL; chloride, 150 mM; phenytoin, 0.1 mM; incubation time, 4 min. The data are the mean f SE from four experiments. Table I. Covalent Binding of [3H]Phenytoin t o Albumin covalent binding, conditions fmolime of albuminb complete system" 437 f 27' -chloride
-MPO [3H]phenytoin and albumin only
61 f 3d 44 f 8 34 f 4
Complete system: MPO/H202/C1-,[3H]phenytoinand albumin. *Values are the mean f SE from four experiments. 'Different than [3H]phenytoin and albumin only with p
