218
Chem. Res. Toxicol. 1991,4, 218-222
N-Chlorination and Oxidation of Procainamide by Myeloperoxidase: Toxicological I mplicationst Jack P. Uetrecht* and Nasir Zahid Faculties of Pharmacy and Medicine, University of Toronto and Sunnybrook Medical Centre, and The University of Toronto Centre for Drug Safety Research, Toronto, Canada M5S 2S2 Received July 23, 1990 In previous studies we had shown that procainamide is metabolized to reactive metabolites by activated leukocytes, and evidence pointed to involvement of myeloperoxidase (MPO). In this study we examine the metabolism of procainamide by MPO/H202, in the presence and absence of chloride ion. In the absence of chloride ion, the metabolism was very similar to that seen with activated leukocytes. The major metabolite was formed by oxidation of the arylamine group to a hydroxylamine. In the presence of chloride ion, a much greater degree of metabolism occurred, and the major product (40% of the starting procainamide) was a reactive species that could not be isolated. This metabolite spontaneously rearranged to 3-chloroprocainamide, and from its mass spectrum and chemical reactions, we deduce its structure to be N-chloroprocainamide. The N-chloroprocainamide metabolite reacted very rapidly with reducing agents, such as ascorbate, and also reacted with protein such as albumin, the major product in both cases being procainamide. This metabolite also chlorinated phenylbutazone. When radiolabeled procainamide was oxidized by MPO/H202 in the presence of albumin, covalent binding of the radiolabel to albumin occurred, and binding was greater under conditions in which N-chloroprocainamide was formed. It is probable that the failure to observe N-chloroprocainamide, when procainamide is oxidized by activated leukocytes, is due to its rapid reaction with the cells. We propose that modification of neutrophils (or neutrophil precursors in the bone marrow) by these reactive metabolites is responsible for procainamide-induced agranulocytosis. In a similar manner, procainamide-induced lupus could be due to modification of monocytes by monocyte-generated reactive metabolites.
Procainamide is an effective antiarrhythmic drug, but its chronic use is limited by a high incidence of drug-induced lupus and agranulocytosis. It was expected that procainamide would be replaced by more effective agents such as flecainide and encainide; however, these two agents have recently been shown to be associated with an increase in mortality (1). The use of a sustained-release form of procainamide has made its use more convenient, but it has also been reported to be associated with an increased incidence of agranulocytosis (2). Thus, although cardiac arrhythmias are a major cause of death, the choice of therapy is severely limited. The incidence of drug-induced lupus associated with the use of procainamide is higher than with any other drug. With chronic therapy most patients develop antinuclear antibodies (ANA) and about 20% develop symptomatic lupus (3, 4). The mechanism of drug-induced lupus is unknown, and possible mechanisms have been reviewed elsewhere ( 5 6 ) . The observation that N-acetylprocainamide does not induce lupus, even in patients that have developed procainamide-induced lupus, as well as in the general association of lupus with drugs that have a primary arylamine functional group, suggests that this group plays an important role in the mechanism of procainamide-induced lupus. We have demonstrated that the arylamine of procainamide is oxidized to a hydroxylamine by rat and human hepatic microsomes (7). This hydroxylamine is 'Portions of this work were presented at the 39th Annual Meeting of the American Society of Pharmacology and Experimental Therapeutics and appeared in the literature (20). *Address correspondence to this author at the Faculty of Pharmacy, University of Toronto, 19 Russell St., Toronto, Canada M5S 2s2.
further converted to a nitroso derivative, and it appears to be the nitroso derivative that is responsible for covalent binding to microsomal protein (8). We have demonstrated that activated leukocytes (both neutrophils and monocytes) can also oxidize procainamide to the hydroxylamine and nitroso metabolites (9). This appears to be due to oxidation by the combination of myeloperoxidase (MPO)' and hydrogen peroxide. Leukocytes are the major cells responsible for the induction of an immune response; therefore, it would seem that the formation of reactive metabolites by these cells would be more likely to lead to an immunological reaction than if the metabolites were formed in the liver. Formation of reactive metabolites by neutrophils or neutrophil precursors in the bone marrow is also likely to be responsible for procainamide-induced agranulocytosis. Such reactive metabolites could cause drug-induced lupus and agranulocytosis by direct toxicity, and we have demonstrated that the hydroxylamine of procainamide is toxic to lymphocytes at micromolar concentrations while the parent drug is not (10). More appealing is a mechanism in which the reactive metabolites act as haptens to generate an immunogen, and antineutrophil antibodies have been observed in procainamideinduced agranulocytosis (11, 12). In similar studies of the metabolism of phenytoin, we found that N,N'-dichlorophenytoin was a product when phenytoin was incubated with MPO/H202/C1-, but not when phenytoin was incubated with activated neutrophils (13). When synthetic N,N'-dichlorophenytoin was incubated with neutrophils, it disappeared with a half-life of less than 1 min. It is therefore likely that the reason that Abbreviation: MPO,myeloperoxidase.
Q893-228~/91/ 27Q4-Q218$Q2.5Q/Q 0 1991 American Chemical Society
N- Chlorination of Procainamide N,”-dichlorophenytoin was not observed when phenytoin was incubated with activated neutrophils is that it reacted with the neutrophils. Likewise, procainamide might undergo N-chlorination by the MPO/H202/C1- system of neutrophils and yet not be detected because it reacts rapidly with the neutrophils. Chlorination of aliphatic amines by activated neutrophils is a well-known reaction (14). Arylamines are also known to be chlorinated by hypochlorite (15),which is probably the active chlorinating species generated by myeloperoxidase (16). N-Chloroprocainamide would be expected to be chemically reactive and could contribute to the toxicity of procainamide. Therefore, the objective of this study was to compare the metabolism of procainamide by MPO/H2O2/C1-with that previously observed in activated neutrophils with an emphasis on finding evidence of N-chlorination.
Materials and Methods Materials. MPO was obtained from Alpha Therapeutic Corp., Los Angeles, CA. Albumin (bovine, fraction V) and procainamide were obtained from Sigma Chemical Co., St. Louis, MO. The hydroxylamine of procainamide was synthesized by the reduction of the nitro derivative with hydrogen in the presence of a triethyl phosphite poisoned platinum catalyst as described previously (7). Desethylprocainamide was a gift from Dr. Tsuen Ruo and Dr. Arthur Atkinson at the Northwestern University Medical School, Chicago, IL. [“CIProcainamide (carbonyl label, 2.8 mCi/mM, >98% purity) was obtained from a custom synthesis by Amersham (Arlington, IL). (A) 3-Chloroprocainamide. The hydroxylamine of procainamide was heated in 5 mL of concentrated hydrochloric acid at 74 OC for 5 h. The product was neutralized with sodium bicarbonate and made alkaline with sodium hydroxide. This mixture was extracted with ethyl acetate, the ethyl acetate layer dried with potassium carbonate, and the solvent evaporated. It was placed on a silica gel column and eluted with ethyl acetate/methanol(41 v/v). The product was a light yellow oil which was 98% pure by HPLC and contained a small amount of procainamide. (B) 2-Chloroprocainamide. 4-Amino-2-chlorobenzoic acid (34 mg, Aldrich Chemical Co., Milwaukee, WI) was dissolved in 5 mL of acetonitrile, and N,N-diethylethylenediamine(34 pL, Aldrich), 1 N hydrochloric acid (200 pL), and I-(ethoxycarbonyl)-2-ethoxy-l,2-dihydroquinoline(50 mg, Aldrich Gold Label) were added. The mixture was stirred for 48 h. The product was purified on a preparative silica gel TLC plate eluted with ethyl acetate/methanol (8:2 v/v). The Rr of the product was 0.2. It was 97% pure by HPLC. (C) N-Chloroprocainamide. Samples of N-chloroprocainamide were made by the action of sodium hypochlorite on procainamide (final concentrations 0.05% and 0.1 mM, respectively) allowed to react for only 1min, and then 0.1 mL was injected onto a HPLC column. N-Chloroprocainamide was obtained by collection of the appropriate peak from HPLC, but it could not be isolated in solid form. (D) Chlorophenylbutazone. Phenylbutazone (0.25 mL of a 4 mM solution in ethanol) was added to 0.75 mL of 0.1 mM pH 6 phosphate buffer. Sodium hypochlorite (10 pL of a 5% solution) was added, and the solution immediately became turbid. After 2 min, the mixture was extracted with ethyl acetate, the ethyl acetate layer was washed with water and dried with sodium sulfate, and the solvent was removed with a stream of nitrogen. The residue was 95% pure by HPLC. Incubations. The baseline conditions for the incubations were as follows: MPO concentration, 1.25 units/mL; procainamide concentration, 0.1 mM; hydrogen peroxide concentration, 0.2 mM; incubation medium, 0.2 mL of phosphate (0.1 M phosphate, pH 7.5); temperature, 25 OC; and time, 5 min. Some of the incubations also contained ascorbate (up to 0.1 mM) and/or chloride ion (150 mM). These conditions were varied to determine their influence on oxidation. Analytical. The incubation mixture was injected onto a HPLC column without further preparation. The HPLC system consisted of a Beckman llOB pump (Berkley, CAI, a 4.6 X 150 mm column
Chem. Res. Toxicol., Vol. 4, No. 2, 1991 219 packed with a 5-pm Spherisorb ODS I1 (Jones Chromatography, Mid Glamorgan, U.K.), and a Beckman 160 detector a t a wavelength of 254 nm. The solvent consisted of water/acetonitrile/acetic acid/triethylamine (80201:0.05 v/v) at a flow rate of 1 mL/min. Under these conditions the retention times for procainamide and its hydroxylamine, desethyl, 3-chloro, and N-chloro metabolites were 3.7, 2.9, 2.2, 8.6, and 10.1 min, respectively, and the standard curves for the hydroxylamine, desethyl, and 3-chloro metabolites were linear with correlation coefficients >0.99. 2-Chloroprocainamide had a retention time of 4.9 min under these conditions, but it was not observed as a metabolite. The solvent for chlorophenylbutazone consisted of water/ acetonitrile/acetic acid/triethylamine (50:501:0.05 v/v), and the retention time was 26.2 min. (A) N-Chloroprocainamide Reactions. N-Chloroprocainamide rearranged spontaneously to 3-chloroprocainamide. In order to study this reaction, the peak with a retention time of 10.1 min was collected from the HPLC and added to 0.1 M phosphate buffer at pH 3.6. An Ultracarb column was used in these experiments (see HPLC/MS method) because it does not require triethylamine and triethylamine reacts slowly with the N-chloroprocainamide. Although this reaction did not appear to interfere significantly with other experiments, it does introduce significant error in these experiments because the rate of rearrangement is relatively slow. These solutions were then reinjected into the HPLC, and the appearance of 3-chloroprocainamide was followed with time. This experiment was done three times. Under these conditions the major product was 3-chloroprocainamide, but a trace of procainamide and a small amount of an unknown product were also produced. N-Chloroprocainamide was prepared as before and purified by HPLC using the solvent listed above. The peak containing N-chloroprocainamide (0.1 mL) was mixed with phenylbutazone or monochlorodimedon (10 pL of a 0.2 mM solution) and the product analyzed by HPLC. N-Chloroprocainamide in 25 pL of HPLC solvent was added directly to an albumin solution (100 pL of pH 6 phosphate buffer, albumin concentration 1 mg/mL). The rate of N-chloroprocainamide disappearance was monitored by HPLC. (B) Quantification of N-Chloroprocainamide. The Nchloro metabolite could not be isolated. Therefore, in order to calculate its concentration, the chromatograms of 6 samples a t two different concentrations were run, and then these samples were treated with solid ascorbate to convert the N-chloroprocainamide to procainamide. The chromatograms were run again, and the peak areas before and after ascorbate were compared. The peak areas of procainamide after ascorbate treatment were 0.692 that for the N-chloro metabolite before treatment. Therefore, the standard curve for the procainamide was used for the N-chloro metabolite after multiplying peak areas by 0.692. Although the conversion of N-chloroprocainamide to procainamide may not have been quantitative, procainamide was the only significant product observed. [14C]Procainamide was used to check the method of N chloroprocainamide quantification. To do this, [14C]procainamide (0.05 pCi) was reacted with sodium hypochlorite (final concentration 0.00125%) for 5 min in 0.2 mL of phosphate-buffered saline (pH 6), and then a 100-pL aliquot was injected into the HPLC. The N-chloroprocainamide peak was collected and separated into two equal aliquots. One of these aliquots was counted in a scintillation counter, and the amount of N-chloroprocainamide was calculated. The other aliquot was converted to procainamide with ascorbate, and the amount of N-chloroprocainamide was calculated by assuming complete conversion to procainamide. This was done a total of four times. The concentration of N chloroprocainamide determined by converting N-chloroprocainamide to procainamide was 1.2 times that determined by radioactivity. This indicates that the method, although it may contain a small degree of error, is valid. (C) Mass Spectra. Mass spectra of desethylprocainamide, 3-chloroprocainamide, and 2-chloroprocainamide were performed on a VG-Analytical ZAB-SE mass spectrometer using methane chemical ionization. The desethylprocainamide metabolite, isolated by HPLC, gave an M + 1 ion at m / z 208 (base peak) and fragment ions a t m / z 163 (74%) and 120 (98%). The mass
220 Chem. Res. Toricol., Vol. 4, No. 2, 1991 spectrum of 3-chloroprocainamide consisted of the molecular ion at m / z 270 (base peak) with a chlorine isotope peak a t m / z 272 (33%) and fragment ions at m/z 143 (7%), 171 (6%), 197 (lo%), and 199 (4%). The mass spectrum of synthetic 2-chloroprocainamide was similar, with a molecular ion at m / z 270 (base peak), isotope peak at m / z 272 (32%),and fragment ions at m/z 143 (4%), 197 (9%), and 199 (3%). The mass spectrum of chlorophenylbutazone done in the electron impact mode produced a molecular ion at m / z 342 (27%) with a chlorine isotope peak at m / z 344 and fragment ions at m/z 183 (loo%), 77 (83%),and 286 (10%);the last of these has a chlorine isotope peak at m / z
Uetrecht a n d Zahid
0
10
20
30
0
Time (mln)
20 40 SO 80 Pfocalnamlde (ItM)
100
05
10
288.
The HPLC/MS of a mixture of 3-chloroprocainamide and N-chloroprocainamidewas performed on a Hewlett Packard 5989A mass spectrometer with a Thermospray LC/MS interface in the positive ion mode. The material for this mass spectrum was produced by adding sodium hypochlorite (final concentration 0.012%) to 8 mM procainamide dissolved in the HPLC solvent described below. After 1 min of mixing, the reaction mixture was injected into the HPLC/MS. The HPLC conditions consisted of a 4.6 X 100 mm column packed with 5-wm Ultracarb 20 (Phenomenex, Torrance, CA) and a mobile phase of water/ methanol/acetic acid ( 8 5 1 5 1 v/v) at a flow rate of 0.7 mL/min. Under these conditions 3-chloroprocainamide and N-chloroprocainamide had retention times of 7 and 8.6 min, respectively. The base peak of the mass spectrum of 3-chloroprocainamidewas the molecular ion at m/z 270, with a chlorine isotope peak at m/z 272 (32%). There were no significant fragment ions. The base peak of the N-chloroprocainamide mass spectrum was also the molecular ion at m / z 270, with a chlorine isotope peak at m / z 272 (40%), and in addition, there was a fragment at m / z 236 (31%) due to loss of chlorine. Covalent Binding. [14C]Procainamide(0.2 WCi), MPO (0.25 unit), and albumin (250 gg) were dissolved in 0.2 mL of either phosphate buffer, 0.1 M phosphate, pH 7.5 (the conditions that produced the maximum amount of the hydroxylamine), or phosphate-buffered saline, 0.1 M phosphate and 150 mM NaCl, pH 6 (the conditions that produced the maximum amount of N-chloroprocainamide). The reaction was initiated by adding hydrogen peroxide to a final concentration of 0.1 mM, and the mixture was incubated for varying periods of time. At the end of the incubation, the albumin was precipitated by addition of 3 volumes of acetone. The albumin was washed repeatedly with acetone, and then it was redissolved in 100 pL of water. The protein concentration of the solution was determined by the method of Bradford and an aliquot was added to scintillation fluid (ACS, Amersham, Oakville, Ontario) and counted in a scintillation counter.
(In,
Results The major metabolite of procainamide formed by MPO/H202in the presence of ascorbate was the hydroxylamine. T h e formation of t h e hydroxylamine metabolite a s a function of time, procainamide concentration, MPO concentration, H202concentration, pH, a n d ascorbate concentration is shown in Figure 1. Ascorbate greatly increased t h e yield of the hydroxylamine, presumably by preventing its further oxidation. A trace of desethylprocainamide, a b o u t 0.03 FM, was also observed, but in contrast t o t h e hydroxylamine, its yield was increased by t h e absence of ascorbate t o 8.6 pM (Figure 2). No oxidation of procainamide was detected in t h e absence of
MPO. In t h e absence of ascorbate a n d in t h e presence of chloride ion, a much greater degree of metabolism was observed. T w o procainamide metabolites were observed in high concentration with retention times of 8.6 a n d 10.1 min t h a t were n o t observed in t h e absence of chloride or in neutrophil incubations. T h e peak at 10.1 min disappeared slowly with t i m e or immediately in t h e presence of ascorbic acid or glutathione. T h e reaction with ascorbate or glutathione generated procainamide. T h e peak also reacted rapidly (within 1 min) with monochlorodimedon,
MPO (unltshnl) 1
55
1
65
PH
75
00
Ascorbate (mM)
Figure 1. Formation of the hydroxylamine metabolite as a function of time, procainamide concentration, MPO concentration, H202 concentration, pH, and ascorbate concentration. The base-line conditions for the incubations were as follows: MPO, 1.25 units/mL; procainamide, 0.1 mM; H202,0.2 mM; incubation medium, 0.2 mL of phosphate (0.1 M phosphate, pH 7.5); temperature, 25 "C; ascorbate, 0.1 mM; and time, 5 min. The points represent the mean f SE from 4 experiments.
s .
s
2.
8-
0 O1 0
" ' " ' " ' " ' 10 20 30 40
50
60
Time (min)
Figure 2. Rate of formation of desethylprocainamide in the absence of ascorbate. The incubation conditions were as follows: MPO, 1.25 units/mL; procainamide, 0.1 mM, H202,0.2 mM; incubation medium, 0.2 mL of phosphate (0.1 M phosphate, pH 7.5); and temperature, 25 O C . The points represent the mean f SE from 4 experiments. leading t o procainamide a n d disappearance of the monochlorodimedon. Dichlorodimedon was presumably formed, b u t because of t h e loss of UV absorbance, this could not be readily confirmed. Therefore, we substituted phenylbutazone for t h e montxhlorodimedon and found t h a t this metabolite also rapidly chlorinated phenylbutazone t o chlorophenylbutazone as determined by HPLC with regeneration of procainamide. When this same peak was collected from HPLC, it rearranged t o the metabolite with a retention time of 8.6 min with a half-life of 141 f 8 min. T h e peak with a retention time of 8.6 min was identified as 3-chloroprocainamide by mass spectroscopy, a n d this was confirmed by comparison of its retention t i m e o n HPLC a n d mass spectrum with t h a t of synthetic 3chloroprocainamide. Since 2-chloroprocainamide could give a similar mass spectrum and retention time on HPLC, the 2-isomer was synthesized; however, its retention time
Chem. Res. Toxicol., Vol. 4, No. 2, 1991 221
N- Chlorination of Procainamide
0
10
20
Time (mln)
30
40
0
20
40
60
80
[;
100
Procalnamldo (pM)
0
60
WITH CHLORIDE I ,
120
180
240
Time (min)
40
6.0
5.0
7.0
.
.
0.0
0.2
PH
.
.
.
.
0.4
0.6
0.8
1.0
Figure 5. Covalent binding of [14C]procainamideto albumin in the presence of MPO/H202with and without chloride ion. The incubation conditions were MPO, 1.25 units/mL; H202,0.1 mM; [“Clprocainamide, 0.2 pCi (final concentration 0.357 mM); albumin, l mg/mL; and incubation medium, 0.2 mL of phosphate (0.1 M, without NaCl at pH 7.5 or with 0.15 M NaCl at pH 6). The points represent the mean A SE from 3 experiments. H. ”OH
H,O,(mM)
Figure 3. Formation of N-chloroprocainamide and 3-chloroprocainamide as a function of time, procainamide concentration, pH, and H202concentration. The base-line conditions for the incubations were as follows: MPO, 1.25 units/mL; procainamide, 0.1 mM; H202,0.1 mM; incubation medium, 0.2 mL of phosphate-buffered saline (0.1 M phosphate, 0.15 M NaCl, pH 6); temperature, 25 “C; and time, 5 min. The points represent the mean f SE from 4 experiments.
NO
MPOM201/CI’or NaCCl
or pmbabiy activated leukocytes
although it is not dreaiy observed
o~~~‘N-CH,CH,N(C,H~), H
pmF)
O*~-N-CHZCHZN(C~H& H
p3G)
Figure 6. Summary of procainamide reactive metabolite formation by MPO/H202and activated leukocytes. =
o
10
20
30
Time (min)
Figure 4. Rate of reaction between N-chloroprocainamideand albumin. The conditions of the incubation were as follows: N-chloroprocainamide,approximately37 MM,albumin, 1 mg/& incubation medium, a mixture of 25 pL of HPLC solvent and 100 MLof pH 6 phosphate buffer.
on HPLC was much shorter than that of the 3-isomer,thus confirming that the metabolite is the 3-isomer. The reactivity of the metabolite with a retention time of 10.1min suggested that it was N-chloroprocainamide. Procainamide was also converted in good yield to the same product by the action of sodium hypochlorite, and a less than stoichiometric amount of sodium hypochlorite in the same buffer used for the MPO oxidations led to a virtually identical pattern of metabolites as observed with the MPO/H202/C1- system (data not shown). The formation of N-chloroprocainamide and 3-chloroprocainamide as a function of time, procainamide concentration, pH, and Hz02concentration is shown in Figure 3. The concentration of N-chloroprocainamide began to decrease after 20 min while that of 3-chloroprocainamide continued to increase. Under some conditions, approximately 40% of the procainamide is converted to the Nchloro metabolite. Lesser amounts of desethylprocainamide, the hydroxylamine, and other minor metabolites were also observed. As would be expected from its reactivity with ascorbate, N-chloroprocainamide was not observed if ascorbate was added to the incubations containing chloride ion (data not shown). N-Chloroprocainamide reacted with albumin, and the rate of disappearance of N-chloroprocainamide is shown in Figure 4. The major product of this reaction was procainamide, and greater than 50% of the N-chloro-
procainamide disappearance could be accounted for as procainamide as determined by HPLC. When [14C]procainamide was oxidized by MPO/H202, reactive metabolites were generated which covalently bound to albumin, and the conditions that favored Nchlorination (Le., lower pH and the presence of chloride) resulted in increased binding (Figure 5 ) .
Discussion The metabolism of procainamide by MPO/HZO2was similar to that which we had observed previously with activated neutrophils (9). Desethylprocainamide was also observed; however, this is a known metabolite of procainamide, and it is not chemically reactive and is unlikely to be of any toxicological significance. It is likely that the MPO system is responsible for the majority of the metabolism of procainamide observed in activated leukocytes. When chloride was added to the incubation, the major metabolite observed appeared to be N-chloroprocainamide. This metabolite was reasonably stable in aqueous solution, especially a t acidic pH; however, it slowly and spontaneously rearranged to 3-chloroprocainamide and reacted very rapidly with reducing agents or even with albumin, the major product being procainamide. Although it is clear that purified N-chloroprocainamide rearranged to 3chloroprocainamide, some of the 3-chloroprocainamide may have formed by direct chlorination of procainamide. The formation of reactive metabolites of procainamide by MPO or activated leukocytes is summarized in Figure 6. Mass spectrometry indicates that the metabolite with a retention time of 10.1 min is chlorinated procainamide, except that it, unlike 3-~hloroprocainamide,fragments readily with loss of a chlorine. However, there is no evidence from the mass spectrum to indicate the location of
Uetrecht and Zahid
222 Chem. Res, Toxicol., Vol. 4 , No. 2, 1991
the chlorine. The fact that it is an active chlorinating agent and reacts with ketoenolates such as ascorbate, monochlorodimedon, and phenylbutazone indicates that the chlorine is likely to be on a nitrogen. Chlorination of the ring of the 1-position would also lead to a reactive intermediate, and there is precedent for substitution on the position para to an arylamine to form an imine, but we consider it very unlikely in this case (18). In the absence of chloride ion (such as when this metabolite was collected from the HPLC), it is unlikely that 1-chloroprocainamide would rearrange to 3-chloroprocainamide,and we positively identified the 3-chloroprocainamide product by synthesis of both it and the 2-chloro isomer. The product of chlorination of the aliphatic amine or amide would also not be expected to rearrange to 3-chloroprocainamide. The mechanism of rearrangement of N-chloroprocainamide to 3-chloroprocainamide is presumably similar to the mechanism involved in our other synthesis of 3-chloroprocainamide, Le., heating the hydroxylamine in concentrated hydrochloric acid. Thus we are confident in the structural assignment of N-chloroprocainamide shown in Figure 6. Although the major reaction of N-chloroprocainamide with protein\results in its reduction back to procainamide, some appears to covalently bind to albumin since oxidation of [14C]procainamide in the presence of albumin led to covalent binding, and the conditions that favored chlorination increased binding. Such alkylation of protein is likely to be due to the electrophilic nature of the carbon ortho to the chloramine, and such a reaction is unlikely to occur with the aliphatic chloramines that have been found to be produced by neutrophils by other investigators (14). However, the increase in the degree of binding in the presence of chloride ion did not increase in proportion to increases in the degree of metabolism observed on addition of chloride in the metabolic studies. This suggests that the hydroxylamine pathway is somewhat more efficient at leading to covalent binding than the chloramine pathway. We have demonstrated that the metabolism of radiolabeled procainamide by activated neutrophils and monocytes leads to covalent binding of the label to these cells (19). Both the hydroxylamine metabolite (probably after further oxidation to the nitroso derivative) and the N-chloro metabolite are probably involved in this binding; however, since chloride ion is present in vivo, the N-chloro intermediate probably makes the larger contribution, even though we do not directly observe this metabolite in the leukocyte incubations. We propose that modification of monocytes by these reactive intermediates is the initial step in procainamide-induced lupus, and modification of neutrophil precursors in the bone marrow (which also contain MPO) is the initial step in procainamide-induced agranulocytosis. This would only be expected to occur if some stimulus, such as an infection or inflammatory condition, activated the cells. Acknowledgment. We are very grateful to Dr. Henrianna Pang for the mass spectra using the VG instrument and to Dr. Alan Viau at Hewlett Packard Canada for running the HPLC/MS spectra. We are also grateful to
Dr. Ruo and Dr. Atkinson for the donation of desethylprocainamide. This work was supported by the Medical Research Council of Canada (MA 9336 and MT 6499).
References (1) CAST (1989) Preliminary report: Effect of encainide and flecainide on mortality in a randomized trial of arrhythmia suppression after myocardial infarction. N. Engl. J. Med. 321, 406-4 12. (2) Ellrodt, A. G., Murata, G. H., Reidinger, M. S., Stewart, M. E., Mochizuki, C., and Gray, R. (1984) Severe neutropenia associated with sustained-release procainamide. Ann. Intern. Med. 100, 197-201. (3) Woosley, R. L., Drayer, D. E., Reidenberg, M. M., Nies, A. S., Carr, K., and Oates, J. A. (1978) Effect of acetylator phenotype on the rate at which procainamide induces antinuclear antibodies and the lupus syndrome. N. Engl. J. Med. 298, 1157-1159. (4) Henningsen, N. C., Cederberg, A,, Hanson, A., and Johansson, B. W. (1975) Effects of long-term treatment with procaine amide. Acta Med. Scand. 198, 475-482. (5) Uetrecht, J. P. (1988) Mechanism of drug-induced lupus. Chem. Res. Toxicol. 1, 133-143. (6) Uetrecht, J. P. (1990) Drug metabolism by leukocytes, its role
in drug-induced lupus and other idiosyncratic drug reactions. CRC Crit. Reu. Toxicol. 20, 213-235. (7) Uetrecht, J. P., Sweetman, B. J., Woosley, R. L., and Oates, J. A. (1984) Metabolism of procainamide to a hydroxylamine by rat and human hepatic microsomes. Drug Metab. Dispos. 12,77-81. (8) Uetrecht, J. P. (1985) Reactivity and possible significance of hydroxylamine and nitroso metabolites of procainamide. J . Pharmacol. Exp. Ther. 232, 42C-425. (9) Uetrecht, J., Zahid, N., and Rubin, R. (1988) Metabolism of procainamide to a hydroxylamine by human neutrophils and mononuclear leukocytes. Chem. Res. Toxicol. 1, 74-78. (10) Rubin, R. L., Uetrecht, J. P., and Jones, J. E. (1987) Cytotoxicity of oxidative metabolites of procainamide. J. Pharmacol. Exp. Ther. 242, 833-841. (11) Azocar, J. (1984) Anti-myeloid-cell antibody in procainamideinduced agranulocytosis. Lancet 1, 1069-1070. (12) Currie, M. S., Weinberg, J. B., Rustagi, P. K., and Logue, G. L. (1987) Antibodies to granulocyte precursors in selective myeloid hypoplasia and other suspected autoimmune neutropenias: use. of HL-60 cells as targets. Blood 69, 529-536. (13) Uetrecht, J., and M i d , N. (1988) N-Chlorination of phenytoin by myeloperoxidase to a reactive metabolite. Chem. Res. Toxicol. 1, 148-151. (14) Thomas, E. L., Grisham, M. B., and Jefferson, M. M. (1983) Myeloperoxidase-dependent effect of amines on functions of isolated neutrophils. J. Clin. Inuest. 72, 441-454. (15) Gassman, P. G., and Campbell, G . A. (1972) Thermal rear-
rangement of N-chloroanilines. Evidence for the intermediacy of nitrenium ions. J. Am. Chem. SOC.94, 3891-3896. (16) Winterborn, C. (1985) Comparative reactives of various biological compounds with myeloperoxidase-hydrogen peroxidechloride, and similarity of the oxidant to hypochlorite. Biochim. Biophys. Acta 840, 204-210. (17) Bradford, M. M. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72, 248-254. (18) Panda, M., Novak, M., and Magonski, J. (1989) Hydrolysis kinetics of the ultimate hepatocarcinogen N-(su1fonatooxy)-2(acety1amino)fluorene: detection of long-lived hydrolysis intermediates. J. Am. Chem. SOC. 11 1, 4524-4525. (19) Uetrecth, J., and Sokoluk, B. Comparative metabolism and covalent binding of procainamide by neutrophils, monocytes, lymphocytes and platelets. Arthritis. Rheum. (submitted for publication). (20) Uetrecht, J. P., and Zahid, N. (1988) Pharmacologist 30, A98.
