Drug Safety 7 (Suppl. I): 51-56, 1992 0114-5916/92/0001-0051/$3.00/0 © Adis International Limited. All rights reserved. DSSUP3329
Metabolism of Clozapine t by Neutrophils Possible Implications for Clozapine-Induced Agranulocytosis Jack P. Uetrecht Faculties of Pharmacy and Medicine, University of Toronto and Sunnybrook Medical Centre, Toronto, Canada
Summary
Many types of adverse drug reactions appear to involve reactive metabolites which, by their very nature, usually have short biological half-lives. Therefore, reactive metabolites formed by neutrophils, or neutrophil precursors in the bone marrow, would seem more likely to be responsible for drug-induced agranulocytosis than metabolites formed in the liver. We have found that several drugs associated with a relatively high incidence of drug-induced agranulocytosis are metabolised by activated neutrophils to chemically reactive metabolites. In preliminary experiments with clozapine, we found that clozapine was metabolised by neutrophils. It also reacted with hypochlorous acid, the principal oxidant generated by neutrophils, to form a reactive intermediate. This intermediate has a half-life of I minute in buffer, but reacts very rapidly with glutathione. We believe that this intermediate is a nitrenium ion. Such a metabolite could be responsible for clozapine-induced agranulocytosis, either by direct toxicity or through an immunemediated mechanism.
1. Idiosyncratic Drug Reactions Idiosyncratic drug reactions represent a significant medical problem because they are virtually impossible to prevent and can be life-threatening (Mathews 1984). They also pose difficulty for the development of new drugs because they are not detected by animal testing and are often not apparent until the drug has been released on the market (Bakke et al. 1984). Very little is known about the mechanism of these reactions. The characteristics of idiosyncratic reactions suggest that they are not due to direct cytotoxicity. Specifically, the reaction does not occur in most tClozaril®/Leponex@
people or in animals, no matter how high the dose of the drug (Park et al. 1987; Pohl et al. 1988; Uetrecht 1990). In addition, there is usually a delay of more than a week between starting the drug and the onset of the adverse reaction; however, rechallenge of a patient who has experienced an idiosyncratic drug reaction usually results in an immediate reaction. One common type of idosyncratic reaction is agranulocytosis. There is strong evidence that agranulocytosis due to amino phenazone (aminopyrine) is mediated by an antibody (Barrett et al. 1976; Goudemand et al. 1976; Moeschlin & Wagner 1952). This evidence includes the observation that serum from patients with acute aminophenazone-induced agranulocytosis caused com ple-
52
ment-dependent agglutination of normal neutrophils. One investigator went so far as to infuse blood from a patient into himself, and this caused an immediate drop in the recipient's neutrophil count. Even though the observed antibodies are directed against mature neutrophils, the bone marrow is affected in almost all cases of aminophenazoneinduced agranulocytosis as well as in agranulocytosis caused by other drugs. Antibodies against neutrophils have also been found in many other cases of drug-induced agranulocytosis; however, in most cases there is no proof that these antibodies are pathogenic (Fibbe et al. 1986; Petz & Fudenberg 1975; Pisciotta 1978; Weitzman & Stossel 1978). In many other studies of drug-induced agranulocytosis, no evidence of antineutrophil antibodies was found, and most negative studies were probably never published. In contrast to aminophenazone-induced agranulocytosis, the agranulocytosis associated with the use of chlorpromazine has more characteristics of direct cytotoxicity (Pisciotta 1969; Pisciotta et al. 1958). Specifically, the onset of agranulocytosis is more gradual, and the time-course is not accelerated on re-exposure. Pisciotta has shown that chlorpromazine is toxic to bone marrow cells and appears to be more toxic to bone marrow cells from patients who have had chlorpromazine-induced agranulocytosis (Pisciotta 1971). During the last several decades, accumulated evidence has shown that many types of toxicity are due to reactive intermediates (Nelson 1982; Nelson & Pearson 1990). A few drugs such as penicillin and several of the anticancer agents are chemically reactive, but most reactive species arise from metabolism of the drug. Reactive intermediates can cause either direct cytotoxicity or immune-mediated reactions by acting as haptens, as in penicillin-induced hypersensitivity (Parker 1982) or halothane-induced hepatic necrosis (Kenna et al. 1988). Most drug metabolism occurs in the liver; however, by their very nature, most reactive metabolites have very short biological half-lives and would not be expected to reach the bone marrow if formed elsewhere. It would seem that reactive intermediates formed by neutrophils or other cells
Drug Safety 7 (Suppl. 1) 1992
in the bone marrow would be more likely to lead to agranulocytosis than reactive metabolites formed outside the bone marrow.
2. Drug Metabolism by Neutrophils A major function of neutrophils and monocytes is the destruction of pathogenic organisms (Kleban off & Clark 1978). When neutrophils encounter such organisms, they are activated, and there is a large increase in their oxygen uptake. This is referred to as a respiratory burst. The oxygen is converted to superoxide by the enzyme NADPH oxidase. The superoxide is in turn converted to hydrogen peroxide. Simultaneously, granules in the neutrophils release myeloperoxidase and several other agents. Hydrogen peroxide converts myeloperoxidase to its oxidised form called compound I (Harrison et al. 1980). Compound I is a strong oxidant, and it appears that it can metabolise drugs. Chloride, the major substrate of myeloperoxidase oxidised by compound I, is converted to hypochlorous acid (Weiss 1989) which is a strong oxidant and an active bactericidal agent used to kill bacteria in municipal water supplies. Most of the superoxide formed by neutrophils is converted to hypochlorous acid. These systems are summarised in figure 1. In addition to mature neutrophils, some neutrophil precursors in the bone marrow are also capable of a significant respiratory burst and release myeloperoxidase on activation (Bainton et al. 1971; Zakhireh & Root 1979). We have demonstrated that many drugs associated with drug-induced agranulocytosis are oxidised by activated neutrophils to reactive metabolites by the combination of myeloperoxidase, hydrogen peroxide and chloride ion or simply by hypochlorous acid (Uetrecht 1990). Specifically, primary arylamines such as procainamide (Uetrecht et al. 1988a; Uetrecht & Zahid 1991), dapsone (Uetrecht et al. 1988b) and sulphonamides (Cribb et al. 1990) are oxidised to hydroxylamines, nitroso metabolites and chloramines. The chloramines are not observed in the neutrophil incubations because they react very rapidly with the cells. In addition, propylthiouracil is metabolised by a series of reactions to a sulphonic acid, which is chemically reactive,
Metabolism of Clozapine by Neutrophils
53
MPO
Compound 1
+ HOCI
+
(200 nmol/10 6 cells)
Fig. 1. Summary of oxidants produced by neutrophils during the respiratory burst. Myeloperoxidase (MPO) is converted by hydrogen peroxide (H202) to its oxidised form, compound I, which in turn oxidises chloride to hypochlorous acid (HOCl).
and several of the intermediates are probably also chemically reactive (Waldhauser & Uetrecht 1991). Phenytoin is chlorinated to N,N'-dichlorophenytoin, and this is associated with a small degree of covalent binding of the drug to neutrophils (Uetrecht & Zahid 1988). Hydralazine is oxidised to an unidentified reactive intermediate which results in covalent binding of the drug to neutrophils (Hofstra et al. 1991). Benzene, which is associated with agranulocytosis and aplastic anaemia, is probably not oxidised by the myeloperoxidase system; however, phenol, a hepatic metabolite of benzene, is metabolised by the myeloperoxidase system (Eastmond et al. 1986). Activated neutrophils are metabolically very reactive and each cell can metabolise a large number of molecules of drug. However, the number of neutrophils activated at anyone time is likely to be low, and once a neutrophil is maximally activated it probably has a rather short life span. Therefore, neutrophils are unlikely to make a significant contribution to the elimination of drugs. In contrast to metabolism by the cytochrome P-450 system, metabolism by neutrophils appears to have a low degree of substrate specificity, and drugs that will
be metabolised can probably be predicted on the basis of whether they have functional groups that are easily oxidised. This is not surprising if the major oxidant is hypochlorous acid. Unlike the cytochrome P-450 system, hypochlorous acid cannot oxidise most carbon-hydrogen bonds, although there are exceptions such as phenylbutazone (Ichihara et al. 1986). Oxidation by hypochlorous acid depends on the presence of a heteroatom, such as nitrogen or sulphur, that is easily oxidised. Metabolism of drug by neutrophils occurs only when the cells are activated (Uetrecht 1990). Therefore, an infection or other inflammatory condition in which the cells are activated could be one of the risk factors for drug-induced agranulocytosis. It appears that the incidence of agranulocytosis associated with vesnarinone, a new drug used for the treatment of heart failure, was increased in patients who received an influenza vaccine during their therapy. We found that influenza vaccine, which had been opsonised by incubation with serum, activated neutrophils so that they metabolised vesnarinone to a metabolite that covalently bound to the neutrophils (Uetrecht et al. 1989).
54
Drug Safety 7 (Suppl. 1) 1992
3. Metabolism of Clozapine by Neutrophils The structure of clozapine shows several nitrogens that could be oxidised. Chlorination of the distal nitrogen of the piperazine ring would likely lead to the formation of an iminium ion and could lead to opening of the ring or further oxidation to a lactam. The product of chlorination of the nitrogen in the 5 position would presumably lose chloride ion to give a relatively stable nitrenium ion. Formation of the nitrenium ion would result in an aromatic 4N + 2 11" electron system, and the positive charge would be extensively delocalised into all 3 ring systems. We have studied the reaction of clozapine with hypochlorous acid in a diode array spectrophotometer. When the 2 components were mixed there was immediate formation of a red colour with a maximum wavelength of about 460nm. This colour disappeared with a half-life of about 1 minute (fig. 2). The final products of the reaction appear to be phenols; however, we have not conclusively identified them at present. These final metabolites are the same whether formed with the myeloperoxidase system or hypochlorous acid, but
we cannot follow the reaction by absorption spectroscopy in the presence of myeloperoxidase. We propose that the intermediate with the red colour is the nitrenium ion (fig. 3) and the final metabolites are phenols formed by the reaction of this nitrenium ion with water. Although the half-life of this metabolite is about 1 minute in phosphate buffer, it reacts very rapidly with glutathione and would be expected to have a very short half-life in vivo. In a biological system such an intermediate would be expected to react with protein sulfhydryl groups and lead to covalent binding of the drug to biological macromolecules.
4. Possible Implications of Clozapine Metabolism by Neutrophils If metabolism of clozapine by activated neutrophils leads to the formation of a reactive nitrenium ion, this could be the initial step in the mechanism of clozapine-induced agranulocytosis. Certainly, it would be expected that the formation of a reactive metabolite of a drug on the surface of neutrophils, or neutrophil precursors in the bone marrow, would be more likely to result in agranulocytosis than if
0.640 ~
c
0.480
OJ
.0
0
«'"
.0
0.320
0.160
0.00 300
400
500
600
Wavelength (nm)
Fig. 2. Absorption spectra of the reaction mixture of c10zapine and hypochlorous acid. Sodium hypochlorite (final concentration 40 ILmol/L) was added to a cuvette containing c\ozapine (40 ILmol/L in phosphate-buffered saline at pH6). The cuvette was inverted to mix the reactants and immediately placed in a diode array spectrophotometer. The first spectrum is the top one and the interval between spectra is 24 seconds. The spectra of c\ozapine and hypochlorous acid have no significant absorption beyond 400nm at this concentration.
ss
Metabolism of Clozapine by Neutrophils
I
Nilrenium ion dusky red colour.
Clozapine
I". '" 1 min
Fig. 3. Proposed structure of the product formed from the reaction of clozapine and hypochlorous acid.
those same metabolites were formed in the liver. Such a reactive metabolite could cause either direct toxicity or lead to an immune-mediated reaction. Clozapine-induced agranulocytosis has characteristics that fall somewhere between those of aminophenazone and chlorpromazine. It appears to have a more idiosyncratic nature than the agranulocytosis associated with chlorpromazine; however, one case was presented at this conference in which a patient who developed c1ozapine-induced agranulocytosis was rechallenged, and it required essentially the same length of time to onset of agranulocytosis on rechallenge as it had on initial exposure. It is conceivable that part of this delay is due to a requirement for a viral infection to activate neutrophils. This must not be the major risk factor for the initial development of agranulocytosis because most patients who develop agranulocytosis do so in the first 3 months of therapy. Therefore, if a viral infection does represent a risk factor, most patients must be exposed to such an infection within a 3-month period. In general, most investigations have sought antineutrophil antibodies as evidence for immune-mediated agranulocytosis; however, the presence of antineutrophil antibodies does not prove that they were responsible for induction of agranulocytosis. Although a rapid onset of agranulocytosis could be expected on re-exposure to a drug that has caused agranulocytosis if the reaction were antibody-mediated, it is less clear that this would be the case if the reaction were mediated by cytotoxic T cells,
and the possibility that drug-induced agranulocytosis involves a cell-mediated immune reaction does not appear to have received much attention. Another possibility is that c1ozapine-induced agranulocytosis involves either an antibody or a T cell that recognises reactive metabolites bound to leucocytes rather than unmodified cells. Current methods of detection would fail because, even if the drug is added to the incubation, the cells would not metabolise drug to the reactive metabolite unless they were activated. When cells are activated they clump and are not suitable for detection of antibody. Identification and synthesis of the reactive metabolite of c1ozapine, as well as the reactive metabolites of other drugs, may permit their use for the detection of antibodies or T cells responsible for agranulocytosis.
Acknowledgement We would like to thank Sandoz Pharma Ltd, Basle, for the donation of clozapine and a small donation to the research. The research is supported by a grant from the Medical Research Council of Canada (MA 10036).
References Bainton OF, Ullyot JL, Farquhar MG. The development of neutrophilic polymorphonuclear leukocytes in human bone marrow. Journal of Experimental Medicine 134: 907-934, 1971 Bakke OM, Wardell WM, Lasagna L. Drug discontinuations in the United Kingdom and the United States, 1964-1983; Issues of safety. Clinical Pharmacology and Therapeutics 35: 559-567, 1984 Barrett AJ, Weller E, Rozengurt N, Longhurst P, Humble JG.
56
Aminopyrine agranulocytosis: drug inhibition of granulocyte colonies in the presence of patient's serum. British Medical Journal 2: 850-851, 1976 Cribb AE, Miller M, Tesoro A, Spielberg SP. Peroxidase-dependent oxidation of sulfonamides by monocytes and neutrophils from humans and dogs. Molecular Pharmacology 38: 744-751, 1990 Eastmond DA, Smith MT, Ruzo LO, Ross D. Metabolic activation of phenol by human myeloperoxidase and horseradish peroxidase. Molecular Pharmacology 30: 674-679, 1986 Fibbe WE, Claas FHJ, Van der Star-Dijlstra W, Schaafsma MR, Meyboom RHB, et al. Agranulocytosis induced by propylthiouracil: evidence of a drug dependent antibody reacting with granulocytes, monocytes and haematopoietic progenitor cells. British Journal of Haematology 64: 363-373, 1986 Goudemand J, Plouvier J, Bauters F, Goudemand M. Les agranulocytosis aigues induites par Ie pyramidon ou les phenothiazines. Semaine des Hopitaux 52: 1513-1520, 1976 Harrison J, Araiso T, Palcic M, Dunford H. Compound I ofmyeloperoxidase. Biochemical and Biophysical Research Communications 94: 34-40, 1980 Hofstra AH, Matassa LC, Uetrecht JP. Metabolism of hydralazine by activated neutrophils: implications for hydralazine-induced lupus. Journal of Rheumatology, in press, 1991 Ichihara S, Tomisawa H, Fukazawa H, Tateishi M, Joly R, et al. Involvement ofleukocytes in the oxygenation and chlorination reaction of phenylbutazone. Biochemical Pharmacology 35: 3935-3939,.1986 Kenna JG, Satoh H, Christ DO, Poh1 LR. Metabolic basis for a drug hypersensitivity: antibodies in sera from patients with halothane hepatitis recognize liver neoantigens that contain the trifluoroacetyl group derived from halothane. Journal of Pharmacology and Experimental Therapeutics 245: 1103-1109, 1988 KlebanotT SJ, Clark RA. The neutrophil: function and clinical disorders. Elsevier/North-Holland Inc., Amsterdam, 1978 Mathews KP. Clinical. spectrum of allergic and pseudoallergic drug reactions. Journal of Allergy and Clinical Immunology 74: 558556, 1984 Moeschlin S, Wagner K. Agranulocytosis due to the occurrence of leukocyte-agglutinins. Acta Haematologica 8: 29-41, 1952 Nelson SO. Metabolic activation and drug toxicity. Journal of Medicinal Chemistry 25: 753-765, 1982 Nelson SO, Pearson PG. Covalent and noncovalent interactions in acute lethal cell injury caused by chemicals. Annual Review of Pharmacology and Toxicology 30: 169-195, 1990 Park BK, Tingle MD, Grabowski PS, Coleman JW, Kitteringham NR. Disposition and immunogenicity of dinitrofluorobenzene, a model compound for the investigation of drugs as haptens. Biochemical Pharmacology 36: 591-599, 1987
Drug Safety 7 (Suppl. 1) 1992
Parker CWo Allergic reactions in man. Pharmacological Reviews 34: 85-104, 1982 Petz LD, Fudenberg HH. Immunologic mechanisms in drug-induced cytopenias. Progress in Hematology 9: 185-206, 1975 Pisciotta AV. Agranulocytosis induced by certain phenothiazine derivatives. Journal of the American Medical Association 208: 1862-1868, 1969 Pisciotta AV. Studies on agranulocytosis IX. A biochemical defect in chlorpromazine-sensitive marrow cells. Journal of Laboratory and Clinical Medicine 78: 435-448, 1971 Pisciotta V. Drug-induced agranulocytosis. Drugs 15: 132-143, 1978 Pisciotta AV, Ebbe S, Lennon EJ, Metzger GO, Madison FW. Agranulocytosis following administration of phenothiazine derivatives. American Journal of Medicine 25: 210-223, 1958 Pohl LR, Satoh H, Christ DO, Kenna JG. The immunologic and metabolic basis of drug hypersensitivities. Annual Review of Pharmacology 28: 367-287, 1988 Uetrecht JP. Drug metabolism by leukocytes, its role in druginduced lupus and other idiosyncratic drug reactions. Critical Reviews in Toxicology 20: 213-235, 1990 Uetrecht J, Zahid N. N-chlorination of phenytoin by myeloperoxidase to a reactive metabolite. Chemical Research in Toxicology 1: 148-151, 1988 Uetrecht JP, Zahid N. N-chlorination and oxidation of procainamide by myeloperoxidase: toxicological implications. Chemical Research in Toxicology 4: 218-222, 1991 Uetrecht J, Zahid N, Rubin R. Metabolism of procainamide to a hydroxylamine.by human neutrophils and mononuclear leukocytes. Chemical Research in Toxicology I: 74-78, 1988a Uetrecht J, Zahid N, Shear NH, Biggar WD. Metabolism of dapsone to a hydroxylamine by human neutrophils and mononuclear cells. Journal of Pharmacology and Experimental Therapeutics 245: 274-279, 1988b Uetrecht JP, Zahid N, Spielberg SP. Oxidation of OPC-8212 to a reactive intermediate by influenza vaccine-activated neutrophils: possible relationship to agranulocytosis. European Journal of Clinical Pharmacology 36 (Suppl.): A53, 1989 Waldhauser L, Uetrecht J. Oxidation of propylthiouracil to reactive metabolites by activated neutrophils: implications for agranulocytosis. Drug Metabolism and Disposition 19: 354359, 1991 Weiss SJ. Tissue destruction by neutrophils. New England Journal of Medicine 320: 365-376, 1989 Weitzman SA, Stossel TP. Drug-induced immunological neutropenia. Lancet 1: 1068-1072, 1978 Zakhireh B, Root RK. Development of oxidase activity by human bone marrow granulocytes. Blood 54: 429-439, 1979 Correspondence and reprints: Dr Jack Uetrecht, Faculties of Pharmacy and Medicine, 19 Russell Street, University of Toronto, Toronto, Canada M5S 2S2.
Metabolism of Clozapine t by Neutrophils Possible Implications for Clozapine-Induced Agranulocytosis Jack P. Uetrecht Faculties of Pharmacy and Medicine, University of Toronto and Sunnybrook Medical Centre, Toronto, Canada
Summary
Many types of adverse drug reactions appear to involve reactive metabolites which, by their very nature, usually have short biological half-lives. Therefore, reactive metabolites formed by neutrophils, or neutrophil precursors in the bone marrow, would seem more likely to be responsible for drug-induced agranulocytosis than metabolites formed in the liver. We have found that several drugs associated with a relatively high incidence of drug-induced agranulocytosis are metabolised by activated neutrophils to chemically reactive metabolites. In preliminary experiments with clozapine, we found that clozapine was metabolised by neutrophils. It also reacted with hypochlorous acid, the principal oxidant generated by neutrophils, to form a reactive intermediate. This intermediate has a half-life of I minute in buffer, but reacts very rapidly with glutathione. We believe that this intermediate is a nitrenium ion. Such a metabolite could be responsible for clozapine-induced agranulocytosis, either by direct toxicity or through an immunemediated mechanism.
1. Idiosyncratic Drug Reactions Idiosyncratic drug reactions represent a significant medical problem because they are virtually impossible to prevent and can be life-threatening (Mathews 1984). They also pose difficulty for the development of new drugs because they are not detected by animal testing and are often not apparent until the drug has been released on the market (Bakke et al. 1984). Very little is known about the mechanism of these reactions. The characteristics of idiosyncratic reactions suggest that they are not due to direct cytotoxicity. Specifically, the reaction does not occur in most tClozaril®/Leponex@
people or in animals, no matter how high the dose of the drug (Park et al. 1987; Pohl et al. 1988; Uetrecht 1990). In addition, there is usually a delay of more than a week between starting the drug and the onset of the adverse reaction; however, rechallenge of a patient who has experienced an idiosyncratic drug reaction usually results in an immediate reaction. One common type of idosyncratic reaction is agranulocytosis. There is strong evidence that agranulocytosis due to amino phenazone (aminopyrine) is mediated by an antibody (Barrett et al. 1976; Goudemand et al. 1976; Moeschlin & Wagner 1952). This evidence includes the observation that serum from patients with acute aminophenazone-induced agranulocytosis caused com ple-
52
ment-dependent agglutination of normal neutrophils. One investigator went so far as to infuse blood from a patient into himself, and this caused an immediate drop in the recipient's neutrophil count. Even though the observed antibodies are directed against mature neutrophils, the bone marrow is affected in almost all cases of aminophenazoneinduced agranulocytosis as well as in agranulocytosis caused by other drugs. Antibodies against neutrophils have also been found in many other cases of drug-induced agranulocytosis; however, in most cases there is no proof that these antibodies are pathogenic (Fibbe et al. 1986; Petz & Fudenberg 1975; Pisciotta 1978; Weitzman & Stossel 1978). In many other studies of drug-induced agranulocytosis, no evidence of antineutrophil antibodies was found, and most negative studies were probably never published. In contrast to aminophenazone-induced agranulocytosis, the agranulocytosis associated with the use of chlorpromazine has more characteristics of direct cytotoxicity (Pisciotta 1969; Pisciotta et al. 1958). Specifically, the onset of agranulocytosis is more gradual, and the time-course is not accelerated on re-exposure. Pisciotta has shown that chlorpromazine is toxic to bone marrow cells and appears to be more toxic to bone marrow cells from patients who have had chlorpromazine-induced agranulocytosis (Pisciotta 1971). During the last several decades, accumulated evidence has shown that many types of toxicity are due to reactive intermediates (Nelson 1982; Nelson & Pearson 1990). A few drugs such as penicillin and several of the anticancer agents are chemically reactive, but most reactive species arise from metabolism of the drug. Reactive intermediates can cause either direct cytotoxicity or immune-mediated reactions by acting as haptens, as in penicillin-induced hypersensitivity (Parker 1982) or halothane-induced hepatic necrosis (Kenna et al. 1988). Most drug metabolism occurs in the liver; however, by their very nature, most reactive metabolites have very short biological half-lives and would not be expected to reach the bone marrow if formed elsewhere. It would seem that reactive intermediates formed by neutrophils or other cells
Drug Safety 7 (Suppl. 1) 1992
in the bone marrow would be more likely to lead to agranulocytosis than reactive metabolites formed outside the bone marrow.
2. Drug Metabolism by Neutrophils A major function of neutrophils and monocytes is the destruction of pathogenic organisms (Kleban off & Clark 1978). When neutrophils encounter such organisms, they are activated, and there is a large increase in their oxygen uptake. This is referred to as a respiratory burst. The oxygen is converted to superoxide by the enzyme NADPH oxidase. The superoxide is in turn converted to hydrogen peroxide. Simultaneously, granules in the neutrophils release myeloperoxidase and several other agents. Hydrogen peroxide converts myeloperoxidase to its oxidised form called compound I (Harrison et al. 1980). Compound I is a strong oxidant, and it appears that it can metabolise drugs. Chloride, the major substrate of myeloperoxidase oxidised by compound I, is converted to hypochlorous acid (Weiss 1989) which is a strong oxidant and an active bactericidal agent used to kill bacteria in municipal water supplies. Most of the superoxide formed by neutrophils is converted to hypochlorous acid. These systems are summarised in figure 1. In addition to mature neutrophils, some neutrophil precursors in the bone marrow are also capable of a significant respiratory burst and release myeloperoxidase on activation (Bainton et al. 1971; Zakhireh & Root 1979). We have demonstrated that many drugs associated with drug-induced agranulocytosis are oxidised by activated neutrophils to reactive metabolites by the combination of myeloperoxidase, hydrogen peroxide and chloride ion or simply by hypochlorous acid (Uetrecht 1990). Specifically, primary arylamines such as procainamide (Uetrecht et al. 1988a; Uetrecht & Zahid 1991), dapsone (Uetrecht et al. 1988b) and sulphonamides (Cribb et al. 1990) are oxidised to hydroxylamines, nitroso metabolites and chloramines. The chloramines are not observed in the neutrophil incubations because they react very rapidly with the cells. In addition, propylthiouracil is metabolised by a series of reactions to a sulphonic acid, which is chemically reactive,
Metabolism of Clozapine by Neutrophils
53
MPO
Compound 1
+ HOCI
+
(200 nmol/10 6 cells)
Fig. 1. Summary of oxidants produced by neutrophils during the respiratory burst. Myeloperoxidase (MPO) is converted by hydrogen peroxide (H202) to its oxidised form, compound I, which in turn oxidises chloride to hypochlorous acid (HOCl).
and several of the intermediates are probably also chemically reactive (Waldhauser & Uetrecht 1991). Phenytoin is chlorinated to N,N'-dichlorophenytoin, and this is associated with a small degree of covalent binding of the drug to neutrophils (Uetrecht & Zahid 1988). Hydralazine is oxidised to an unidentified reactive intermediate which results in covalent binding of the drug to neutrophils (Hofstra et al. 1991). Benzene, which is associated with agranulocytosis and aplastic anaemia, is probably not oxidised by the myeloperoxidase system; however, phenol, a hepatic metabolite of benzene, is metabolised by the myeloperoxidase system (Eastmond et al. 1986). Activated neutrophils are metabolically very reactive and each cell can metabolise a large number of molecules of drug. However, the number of neutrophils activated at anyone time is likely to be low, and once a neutrophil is maximally activated it probably has a rather short life span. Therefore, neutrophils are unlikely to make a significant contribution to the elimination of drugs. In contrast to metabolism by the cytochrome P-450 system, metabolism by neutrophils appears to have a low degree of substrate specificity, and drugs that will
be metabolised can probably be predicted on the basis of whether they have functional groups that are easily oxidised. This is not surprising if the major oxidant is hypochlorous acid. Unlike the cytochrome P-450 system, hypochlorous acid cannot oxidise most carbon-hydrogen bonds, although there are exceptions such as phenylbutazone (Ichihara et al. 1986). Oxidation by hypochlorous acid depends on the presence of a heteroatom, such as nitrogen or sulphur, that is easily oxidised. Metabolism of drug by neutrophils occurs only when the cells are activated (Uetrecht 1990). Therefore, an infection or other inflammatory condition in which the cells are activated could be one of the risk factors for drug-induced agranulocytosis. It appears that the incidence of agranulocytosis associated with vesnarinone, a new drug used for the treatment of heart failure, was increased in patients who received an influenza vaccine during their therapy. We found that influenza vaccine, which had been opsonised by incubation with serum, activated neutrophils so that they metabolised vesnarinone to a metabolite that covalently bound to the neutrophils (Uetrecht et al. 1989).
54
Drug Safety 7 (Suppl. 1) 1992
3. Metabolism of Clozapine by Neutrophils The structure of clozapine shows several nitrogens that could be oxidised. Chlorination of the distal nitrogen of the piperazine ring would likely lead to the formation of an iminium ion and could lead to opening of the ring or further oxidation to a lactam. The product of chlorination of the nitrogen in the 5 position would presumably lose chloride ion to give a relatively stable nitrenium ion. Formation of the nitrenium ion would result in an aromatic 4N + 2 11" electron system, and the positive charge would be extensively delocalised into all 3 ring systems. We have studied the reaction of clozapine with hypochlorous acid in a diode array spectrophotometer. When the 2 components were mixed there was immediate formation of a red colour with a maximum wavelength of about 460nm. This colour disappeared with a half-life of about 1 minute (fig. 2). The final products of the reaction appear to be phenols; however, we have not conclusively identified them at present. These final metabolites are the same whether formed with the myeloperoxidase system or hypochlorous acid, but
we cannot follow the reaction by absorption spectroscopy in the presence of myeloperoxidase. We propose that the intermediate with the red colour is the nitrenium ion (fig. 3) and the final metabolites are phenols formed by the reaction of this nitrenium ion with water. Although the half-life of this metabolite is about 1 minute in phosphate buffer, it reacts very rapidly with glutathione and would be expected to have a very short half-life in vivo. In a biological system such an intermediate would be expected to react with protein sulfhydryl groups and lead to covalent binding of the drug to biological macromolecules.
4. Possible Implications of Clozapine Metabolism by Neutrophils If metabolism of clozapine by activated neutrophils leads to the formation of a reactive nitrenium ion, this could be the initial step in the mechanism of clozapine-induced agranulocytosis. Certainly, it would be expected that the formation of a reactive metabolite of a drug on the surface of neutrophils, or neutrophil precursors in the bone marrow, would be more likely to result in agranulocytosis than if
0.640 ~
c
0.480
OJ
.0
0
«'"
.0
0.320
0.160
0.00 300
400
500
600
Wavelength (nm)
Fig. 2. Absorption spectra of the reaction mixture of c10zapine and hypochlorous acid. Sodium hypochlorite (final concentration 40 ILmol/L) was added to a cuvette containing c\ozapine (40 ILmol/L in phosphate-buffered saline at pH6). The cuvette was inverted to mix the reactants and immediately placed in a diode array spectrophotometer. The first spectrum is the top one and the interval between spectra is 24 seconds. The spectra of c\ozapine and hypochlorous acid have no significant absorption beyond 400nm at this concentration.
ss
Metabolism of Clozapine by Neutrophils
I
Nilrenium ion dusky red colour.
Clozapine
I". '" 1 min
Fig. 3. Proposed structure of the product formed from the reaction of clozapine and hypochlorous acid.
those same metabolites were formed in the liver. Such a reactive metabolite could cause either direct toxicity or lead to an immune-mediated reaction. Clozapine-induced agranulocytosis has characteristics that fall somewhere between those of aminophenazone and chlorpromazine. It appears to have a more idiosyncratic nature than the agranulocytosis associated with chlorpromazine; however, one case was presented at this conference in which a patient who developed c1ozapine-induced agranulocytosis was rechallenged, and it required essentially the same length of time to onset of agranulocytosis on rechallenge as it had on initial exposure. It is conceivable that part of this delay is due to a requirement for a viral infection to activate neutrophils. This must not be the major risk factor for the initial development of agranulocytosis because most patients who develop agranulocytosis do so in the first 3 months of therapy. Therefore, if a viral infection does represent a risk factor, most patients must be exposed to such an infection within a 3-month period. In general, most investigations have sought antineutrophil antibodies as evidence for immune-mediated agranulocytosis; however, the presence of antineutrophil antibodies does not prove that they were responsible for induction of agranulocytosis. Although a rapid onset of agranulocytosis could be expected on re-exposure to a drug that has caused agranulocytosis if the reaction were antibody-mediated, it is less clear that this would be the case if the reaction were mediated by cytotoxic T cells,
and the possibility that drug-induced agranulocytosis involves a cell-mediated immune reaction does not appear to have received much attention. Another possibility is that c1ozapine-induced agranulocytosis involves either an antibody or a T cell that recognises reactive metabolites bound to leucocytes rather than unmodified cells. Current methods of detection would fail because, even if the drug is added to the incubation, the cells would not metabolise drug to the reactive metabolite unless they were activated. When cells are activated they clump and are not suitable for detection of antibody. Identification and synthesis of the reactive metabolite of c1ozapine, as well as the reactive metabolites of other drugs, may permit their use for the detection of antibodies or T cells responsible for agranulocytosis.
Acknowledgement We would like to thank Sandoz Pharma Ltd, Basle, for the donation of clozapine and a small donation to the research. The research is supported by a grant from the Medical Research Council of Canada (MA 10036).
References Bainton OF, Ullyot JL, Farquhar MG. The development of neutrophilic polymorphonuclear leukocytes in human bone marrow. Journal of Experimental Medicine 134: 907-934, 1971 Bakke OM, Wardell WM, Lasagna L. Drug discontinuations in the United Kingdom and the United States, 1964-1983; Issues of safety. Clinical Pharmacology and Therapeutics 35: 559-567, 1984 Barrett AJ, Weller E, Rozengurt N, Longhurst P, Humble JG.
56
Aminopyrine agranulocytosis: drug inhibition of granulocyte colonies in the presence of patient's serum. British Medical Journal 2: 850-851, 1976 Cribb AE, Miller M, Tesoro A, Spielberg SP. Peroxidase-dependent oxidation of sulfonamides by monocytes and neutrophils from humans and dogs. Molecular Pharmacology 38: 744-751, 1990 Eastmond DA, Smith MT, Ruzo LO, Ross D. Metabolic activation of phenol by human myeloperoxidase and horseradish peroxidase. Molecular Pharmacology 30: 674-679, 1986 Fibbe WE, Claas FHJ, Van der Star-Dijlstra W, Schaafsma MR, Meyboom RHB, et al. Agranulocytosis induced by propylthiouracil: evidence of a drug dependent antibody reacting with granulocytes, monocytes and haematopoietic progenitor cells. British Journal of Haematology 64: 363-373, 1986 Goudemand J, Plouvier J, Bauters F, Goudemand M. Les agranulocytosis aigues induites par Ie pyramidon ou les phenothiazines. Semaine des Hopitaux 52: 1513-1520, 1976 Harrison J, Araiso T, Palcic M, Dunford H. Compound I ofmyeloperoxidase. Biochemical and Biophysical Research Communications 94: 34-40, 1980 Hofstra AH, Matassa LC, Uetrecht JP. Metabolism of hydralazine by activated neutrophils: implications for hydralazine-induced lupus. Journal of Rheumatology, in press, 1991 Ichihara S, Tomisawa H, Fukazawa H, Tateishi M, Joly R, et al. Involvement ofleukocytes in the oxygenation and chlorination reaction of phenylbutazone. Biochemical Pharmacology 35: 3935-3939,.1986 Kenna JG, Satoh H, Christ DO, Poh1 LR. Metabolic basis for a drug hypersensitivity: antibodies in sera from patients with halothane hepatitis recognize liver neoantigens that contain the trifluoroacetyl group derived from halothane. Journal of Pharmacology and Experimental Therapeutics 245: 1103-1109, 1988 KlebanotT SJ, Clark RA. The neutrophil: function and clinical disorders. Elsevier/North-Holland Inc., Amsterdam, 1978 Mathews KP. Clinical. spectrum of allergic and pseudoallergic drug reactions. Journal of Allergy and Clinical Immunology 74: 558556, 1984 Moeschlin S, Wagner K. Agranulocytosis due to the occurrence of leukocyte-agglutinins. Acta Haematologica 8: 29-41, 1952 Nelson SO. Metabolic activation and drug toxicity. Journal of Medicinal Chemistry 25: 753-765, 1982 Nelson SO, Pearson PG. Covalent and noncovalent interactions in acute lethal cell injury caused by chemicals. Annual Review of Pharmacology and Toxicology 30: 169-195, 1990 Park BK, Tingle MD, Grabowski PS, Coleman JW, Kitteringham NR. Disposition and immunogenicity of dinitrofluorobenzene, a model compound for the investigation of drugs as haptens. Biochemical Pharmacology 36: 591-599, 1987
Drug Safety 7 (Suppl. 1) 1992
Parker CWo Allergic reactions in man. Pharmacological Reviews 34: 85-104, 1982 Petz LD, Fudenberg HH. Immunologic mechanisms in drug-induced cytopenias. Progress in Hematology 9: 185-206, 1975 Pisciotta AV. Agranulocytosis induced by certain phenothiazine derivatives. Journal of the American Medical Association 208: 1862-1868, 1969 Pisciotta AV. Studies on agranulocytosis IX. A biochemical defect in chlorpromazine-sensitive marrow cells. Journal of Laboratory and Clinical Medicine 78: 435-448, 1971 Pisciotta V. Drug-induced agranulocytosis. Drugs 15: 132-143, 1978 Pisciotta AV, Ebbe S, Lennon EJ, Metzger GO, Madison FW. Agranulocytosis following administration of phenothiazine derivatives. American Journal of Medicine 25: 210-223, 1958 Pohl LR, Satoh H, Christ DO, Kenna JG. The immunologic and metabolic basis of drug hypersensitivities. Annual Review of Pharmacology 28: 367-287, 1988 Uetrecht JP. Drug metabolism by leukocytes, its role in druginduced lupus and other idiosyncratic drug reactions. Critical Reviews in Toxicology 20: 213-235, 1990 Uetrecht J, Zahid N. N-chlorination of phenytoin by myeloperoxidase to a reactive metabolite. Chemical Research in Toxicology 1: 148-151, 1988 Uetrecht JP, Zahid N. N-chlorination and oxidation of procainamide by myeloperoxidase: toxicological implications. Chemical Research in Toxicology 4: 218-222, 1991 Uetrecht J, Zahid N, Rubin R. Metabolism of procainamide to a hydroxylamine.by human neutrophils and mononuclear leukocytes. Chemical Research in Toxicology I: 74-78, 1988a Uetrecht J, Zahid N, Shear NH, Biggar WD. Metabolism of dapsone to a hydroxylamine by human neutrophils and mononuclear cells. Journal of Pharmacology and Experimental Therapeutics 245: 274-279, 1988b Uetrecht JP, Zahid N, Spielberg SP. Oxidation of OPC-8212 to a reactive intermediate by influenza vaccine-activated neutrophils: possible relationship to agranulocytosis. European Journal of Clinical Pharmacology 36 (Suppl.): A53, 1989 Waldhauser L, Uetrecht J. Oxidation of propylthiouracil to reactive metabolites by activated neutrophils: implications for agranulocytosis. Drug Metabolism and Disposition 19: 354359, 1991 Weiss SJ. Tissue destruction by neutrophils. New England Journal of Medicine 320: 365-376, 1989 Weitzman SA, Stossel TP. Drug-induced immunological neutropenia. Lancet 1: 1068-1072, 1978 Zakhireh B, Root RK. Development of oxidase activity by human bone marrow granulocytes. Blood 54: 429-439, 1979 Correspondence and reprints: Dr Jack Uetrecht, Faculties of Pharmacy and Medicine, 19 Russell Street, University of Toronto, Toronto, Canada M5S 2S2.
