Nevirapine hypersensitivity.

Treatment of HIV-1 infections with nevirapine is associated with skin and liver toxicity. These two organ toxicities range from mild to severe, in rare cases resulting in life-threatening liver failure or toxic epidermal necrolysis. The study of the mechanistic steps leading to nevirapine-induced skin rash has been facilitated by the discovery of an animal model in which nevirapine causes a skin rash in rats that closely mimics the rash reported in patients. The similarity in characteristics of the rash between humans and rats strongly suggests that the basic mechanism is the same in both. The rash is clearly immune-mediated in rats, and partial depletion of CD4(+) T cells, but not CD8(+) T cells, is protective. We have demonstrated that the rash is related to the 12-hydroxylation of nevirapine rather than to the parent drug. This is presumably because the 12-hydroxy metabolite can be converted to a reactive quinone methide in skin, but that remains to be demonstrated. Although the rash is clearly related to the 12-hydroxy metabolite rather than the parent drug, cells from rechallenged animals respond ex vivo to the parent drug by producing cytokines such as interferon-gamma with little response to the 12-hydroxy metabolite, even when the rash was induced by treatment with the metabolite rather than the parent drug. This indicates that the response of T cells in vitro cannot be used to determine what caused an immune response. We are now studying the detailed steps by which the 12-hydroxy metabolite induces an immune response and skin rash. This animal model provides a unique tool to study the mechanistic details of an idiosyncratic drug reaction; however, it is likely that there are significant differences in the mechanisms of different idiosyncratic drug reactions, and therefore the results of these studies cannot safely be generalized to all idiosyncratic drug reactions.

Oxidative stress and thiol depletion in plasma and peripheral blood lymphocytes from HIV-infected patients: toxicological and pathological implications.

OBJECTIVES: To determine, first, whether the plasma and lymphocytes of HIV-positive individuals and AIDS patients have alterations in the major thiols glutathione and cysteine, and/or their oxidative disulphide and mixed disulphide products; and, secondly, whether thiol/disulphide status differs in patients with sulphonamide drug hypersensitivity reactions. DESIGN: Thiols provide critical cellular defence against toxic drug reactive intermediates and endogenous oxidative stress, and may modulate HIV replication. Glutathione is reported to be low in HIV-positive individuals and AIDS patients, but this is controversial and the mechanism responsible is unknown. Also unknown is whether altered thiol/disulphide status determines the predisposition of HIV-positive and AIDS patients to drug reactions. METHODS: Thiols and disulphides were measured by high-performance liquid chromatography. RESULTS: Both plasma thiols were decreased by approximately 58% in HIV-positive individuals and AIDS patients compared with uninfected controls (P < 0.05), with increases of up to threefold in oxidized products (P < 0.05). Similarly, in lymphocytes, thiols were decreased by 30-35% (P < 0.05), with apparent increases in oxidized products. For both glutathione and cysteine, the thiol/disulphide ratios also were decreased (P < 0.05). The plasma and lymphocyte glutathione thiol/disulphide ratios were highly correlated (r = 0.7661; P = 0.0001) among all subjects. No parameters differed in patients with drug reactions, or with antiretroviral therapy. CONCLUSIONS: The enhanced thiol oxidation in HIV-positive individuals and AIDS patients indicates oxidative stress, which also contributes to thiol depletion, and may enhance damage to macromolecular targets. These mechanisms may contribute to enhanced viral replication and other pathological outcomes. HIV-positive individuals' and AIDS patients' predisposition to drug hypersensitivity reactions appears to be unrelated to thiol/disulphide status.

Metabolism of 35S- and 14C-labeled propylthiouracil in a model in vitro system containing thyroid peroxidase.

In previous communications we described an in vitro model system containing highly purified thyroid peroxidase (TPO) for studying the mechanism of inhibition of thyroid hormone biosynthesis by the antithyroid drugs, 6-propylthiouracil (PTU) and 1-methyl-2-mercaptoimidazole (MMI). We showed that inhibition of iodination of thyroglobulin in this system may be reversible or irreversible depending on the relative concentrations of iodide and drug and the TPO concentration. Metabolism of the drugs occurred under both conditions, but was more limited under irreversible conditions of inhibition. It was of interest to examine the nature of the drug metabolites associated with reversible and irreversible conditions of inhibition. For this purpose we have employed the 35S- and 14C-labeled drugs and a recently developed reverse phase HPLC procedure. Results of a similar study with MMI were reported in an earlier communication. In the present study we report our findings with PTU. Under conditions of reversible inhibition, PTU was readily metabolized and by 15 min was reduced to a few percent of the starting value. The earliest detectable metabolite with both [35S]- and [14C]PTU was the disulfide, which reached a peak in about 15 min and then slowly declined. Coincident with the decline in the disulfide was the appearance of more polar metabolites. In the case of [35S]PTU, these corresponded to sulfate/sulfite, PTU sulfonate, and a product tentatively identified as PTU sulfinate. The latter two were also observed as 14C-labeled metabolites produced from [14C]PTU. Two nonpolar desulfurated 14C-labeled metabolites were also observed. Surprisingly, these did not correspond to either propyluracil or propyldeoxyuracil, the anticipated most likely products of PTU desulfuration. The identity of these desulfurated metabolites of PTU in the TPO model system remains to be determined. Under conditions of irreversible inhibition of iodination, a relatively small fraction of PTU was metabolized. PTU disulfide was, again, the earliest detectable metabolite, and it declined with time. However, only small amounts of other metabolites were observed, in contrast to the results obtained under conditions of reversible inhibition of iodination. As in the case of MMI, the difference in metabolic pattern between reversible and irreversible conditions is primarily related to the rapid inactivation of TPO that occurs under irreversible conditions. In general, the metabolism of PTU by the TPO model system resembled that previously observed with MMI. With both drugs, the disulfide was the earliest detectable metabolite, and under conditions of reversible inhibition of iodination, an appreciable fraction of the sulfur was oxidized as far as sulfate/sulfite.(ABSTRACT TRUNCATED AT 400 WORDS)

The contribution of N-hydroxylation and acetylation to dapsone pharmacokinetics in normal subjects.

The relative importance of N-hydroxylation and acetylation of dapsone to the oral clearance of dapsone (100 mg) was investigated in seven healthy volunteers. Plasma dapsone and monoacetyldapsone concentrations rose rapidly with subsequent similar monoexponential elimination. The oral clearance of dapsone was low (33 +/- 14 ml/min), with a threefold variability. Four subjects were identified as fast acetylators; however, differences in acetylation did not explain the variability in oral clearance. The cumulative urinary recoveries of dapsone and its hydroxylamine were approximately 20% of the dose. The formation clearance of hydroxylamine, which exhibited a tenfold intersubject variability, was closely associated with the oral clearance of dapsone (r = 0.96). Thus, the formation of the hydroxylamine is more important than acetylation in determining dapsone's intersubject variability in oral clearance. Variation in N-hydroxylation may have clinical consequences, because the hydroxylamine is considered to be important in dapsone-mediated toxicity.

Drug-induced hypothyroidism: the thyroid as a target organ in hypersensitivity reactions to anticonvulsants and sulfonamides.

Inherited defects in detoxification of reactive metabolites of drugs predispose patients to "hypersensitivity" reactions. Covalent interaction of metabolites with cell macromolecules leads to cytotoxic and immunologic outcomes, manifested clinically by multisystem syndromes with variable organ involvement. Hypothyroidism developed in 5 of 202 patients (age range, 1 to 81 years) we investigated for hypersensitivity reactions to anticonvulsants or sulfonamides shortly after their reaction. None had previous personal or family histories of autoimmune disease. All had low thyroxine levels, elevated levels of thyroid stimulating hormone, and autoantibodies including antimicrosomal antibodies. Patients were 2 to 18 years of age at presentation, and two were male. All returned to a euthyroid state within a year of presentation, and all remain well. The demographics, clinical presentation, and course of the patients is atypical of idiopathic lymphocytic thyroiditis. We investigated the pathogenesis of thyroid toxicity using the hydroxylamine metabolite of sulfamethoxazole as a model. The hydroxyalmine was toxic to thyroid cells in vitro, which did or did not express thyroid peroxidase activity, whereas the parent sulfonamide was toxic only to cells with active thyroid peroxidase. The purified enzyme converted sulfamethoxazole to the hydroxylamine. Formation of reactive drug metabolites by thyroid peroxidase in a host who is genetically unable to detoxify the metabolites may lead directly to cytotoxicity. Covalent binding to macromolecules, including thyroid peroxidase, also may lead to expression of neoantigens and formation of autoantibodies. Patients who have sustained hypersensitivity reactions to drugs should be investigated for possible involvement of the thyroid.

Is it possible to more accurately predict which drug candidates will cause idiosyncratic drug reactions?

The unexpected occurrence of idiosyncratic drug reactions during late clinical trials or after a drug has been released can lead to a severe restriction in its use or failure to release/withdrawal. This leads to considerable uncertainty in drug development and has led to attempts to try to predict a drug's potential to cause such reactions. It appears that most idiosyncratic drug reactions are due to reactive metabolites; however, many drugs that form reactive metabolites are associated with a very low incidence of idiosyncratic drug reactions. Therefore. screening drug for their ability to generate reactive metabolites is likely to cause the rejection of many good drug candidates. There is evidence to suggest that an idiosyncratic drug reaction is more likely if there is some "danger signal'. Thus drugs that cause some degree of cell stress or damage may be more likely to lead to a high incidence of idiosyncratic drug reactions. The exact nature of the putative danger signals is unknown. However, a screen of the effects of drugs known to be associated with a high incidence of idiosyncatic reactions using expression genomics and proteomics may reveal a pattern or patterns of mRNA and protein expression that predict which drugs will cause a high incidence of idiosyncratic drug reactions. Although idiosyncratic drug reactions are not usually detected in animal tests because they are just as idiosyncratic in animals as they are in humans, it is likely that drug reactive metabolites would also cause similar cell stress in animals. It is more likely that in most cases it is differences in the immune response to the reactive metabolites that determine which individuals will develop an idiosyncratic reaction. If the expression of certain proteins in animals treated with a drug candidate could be used as a screening method to predict a drug's potential to cause a high incidence of idiosyncratic drug reactions, it would greatly facilitate the development of safer drugs.

Mechanism of idiosyncratic drug reactions: reactive metabolite formation, protein binding and the regulation of the immune system.

Drug-induced adverse reactions, especially type B reactions, represent a major clinical problem. They also impart a significant degree of uncertainty into drug development because they are often not detected until the drug has been released onto the market. Type B reactions are also termed idiosyncratic drug reactions by many investigators due to their unpredictable nature and our lack of understanding of the mechanisms involved. It is currently believed that the majority of these reactions are immune-mediated and are caused by immunogenic conjugates formed from the reaction of a reactive metabolite of a drug with cellular proteins. It has been shown that most drugs associated with idiosyncratic reactions form reactive metabolites to some degree. Covalent binding of reactive metabolites to cellular proteins has also been shown in many cases. However, studies to reveal the role of reactive metabolites and their protein-adducts in the mechanism of drug-induced idiosyncratic reactions are lacking. This review will focus on our current understanding and speculative views on how a reactive metabolite of a drug might ultimately lead to immune-mediated toxicity.

Myeloperoxidase as a generator of drug free radicals.

Reactive metabolites are believed to be responsible for many types of toxicity, including idiosyncratic drug reactions. Bone marrow is a frequent target of idiosyncratic reactions, and, since these reactions have characteristics that suggest involvement of the immune system, the formation of reactive metabolites by leucocytes could also play a role in the aetiology of idiosyncratic drug reactions. The major oxidation system in neutrophils and monocytes is a combination of NADPH oxidase and myeloperoxidase. This system oxidizes primary arylamines, such as sulphonamides, to reactive metabolites and these drugs are also associated with a high incidence of agranulocytosis, generalized idiosyncratic reactions and/ or drug-induced lupus. Clozapine is oxidized by this system to a relatively stable nitrenium ion; clozapine is also associated with a high incidence of agranulocytosis. Arylamines that have an oxygen or nitrogen in the para position, such as amodiaquine, vesnarinone and 5-aminosalicylic acid, are oxidized to quinone-like reactive intermediates. Aminopyrine is oxidized to a very reactive dication. Such reactive metabolites could also inhibit neutrophil function and mediate some of the therapeutic effects of these drugs: for example, the use of dapsone for dermatitis herpetiformis and the use of 5-aminosalicylic acid for inflammatory bowel disease.

Acetylator phenotype and lupus erythematosus.

There are several known therapeutic implications of acetylator phenotype; among them, the association of a higher incidence of procainamide- and hydralazine-induced lupus in slow acetylators. Presumably, this is because acetylation of the aromatic amine or hydrazine functional group leads to a non-toxic product. Several other drugs which have been implicated in drug-induced lupus also contain an aromatic amine or hydrazine group. The clinical and laboratory characteristics of drug-induced and idiopathic lupus are similar but the degree to which the pathophysiological mechanisms are related, if at all, is unknown. There is also evidence reported for an association between the slow acetylator phenotype and idiopathic lupus. If true, this relationship should provoke some new experimental approaches to investigation into the mechanism of idiopathic lupus.

Intrinsic left ventricular contractility in normal subjects.

The influence of autonomic tone on left ventricular (LV) contractility, along with the range of normal values and the effects of exercise on contractile state, were studied in 12 normal volunteers. Serial reproducibility was examined in a subgroup of 6. LV contractility was estimated by the LV peak-systolic pressure to end-systolic volume relation (pressure-volume relation), and the ratio of peak-systolic pressure to end-systolic volume (pressure/volume ratio). The cuff blood pressure and radionuclide ventriculogram were recorded at rest, during exercise and during pharmacologic pressure-afterloading with phenylephrine, before and after vagal and beta-adrenergic "blockade." Both the pressure/volume ratio and ejection fraction increased during the stimulus of exercise (both p less than or equal to 0.008). After blockade, the pressure-volume relations were highly linear (r = 0.95 +/- 0.05 [standard deviation], n = 12), and there was no systematic difference in their slopes induced by blockade. The serial studies of pressure-volume relations showed no significant differences. The results demonstrated that vagal and sympathetic tone were not important in the support of LV contractility in normal subjects at rest, and that the pressure-volume relation and pressure/volume ratio are reproducible between studies. Also, the findings confirmed that both the pressure/volume ratio and the ejection fraction were sensitive to exercise-induced changes in contractility. This demonstration of intrinsic LV contractility in normal subjects, plus the reproducibility of the measurements, supports the feasibility of serial study of LV contractility.

Carbamazepine metabolism to a reactive intermediate by the myeloperoxidase system of activated neutrophils.

Carbamazepine is an anticonvulsant which is associated with a significant incidence of hypersensitivity reactions including agranulocytosis. We have postulated that many drug hypersensitivity reactions, especially agranulocytosis and lupus, are due to reactive metabolites generated by the myeloperoxidase (MPO) (EC 1.11.1.7) system of neutrophils and monocytes. This led to a study of the metabolism and covalent binding of carbamazepine with MPO/H2O2/Cl- and neutrophils. Metabolism and covalent binding were observed in both systems and the same pathway appeared to be involved; however, the metabolism observed with the MPO system was approximately 500-fold greater than that observed with neutrophils. The metabolites identified were an intermediate aldehyde, 9-acridine carboxaldehyde, acridine, acridone, choloroacridone, and dichloroacridone. We postulate that the first intermediate in the metabolism of carbamazepine is a carbonium ion formed by reaction of hypochlorous acid (HOCl) with the 10,11 double bond. Although we have no direct proof for the proposed carbonium ion, it provides the most likely mechanism for the observed ring contraction. Iminostilbene, a known metabolite of carbamazepine, was also metabolized by a similar pathway leading to ring contraction; however, the rate was much faster and the first step may involve N-chlorination and a nitrenium ion intermediate. These data confirm that carbamazepine is metabolized to reactive intermediates by activated leukocytes. Such metabolites could be responsible for some of the adverse reactions associated with carbamazepine, especially reactions such as agranulocytosis and lupus which involve leukocytes.

Metabolism of clozapine by neutrophils. Possible implications for clozapine-induced agranulocytosis.

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 1 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 immune-mediated mechanism.

Current trends in drug-induced autoimmunity.

Idiosyncratic adverse drug reactions have characteristics that suggest involvement of the immune system. In particular, drug-induced lupus which is an autoimmune syndrome, must be immune-mediated. A major working hypothesis for the first step in the mechanism of drug-induced autoimmunity is that the drug, or more commonly a reactive metabolite of the drug, must irreversibly bind to some structure. In view of the reactive nature of these metabolites, in most cases it is likely that the metabolite must be formed in the organ where toxicity occurs. The liver is the major site of drug metabolism and it is a common target for idiosyncratic drug reactions. In the case of immune reactions directly involving leukocytes, the enzyme system most likely responsible for the formation of reactive metabolites is the NADPH oxidase/myeloperoxidase system found in neutrophils and monocytes. In some cases, the reactive metabolite results in the production of antibodies or T-cells directed against the altered structure. However, in many other cases, the mechanism appears to be more complex than this. In some cases, true auto-antibodies are produced that do not require the presence of the drug, and furthermore, the antibodies produced often are the same as those induced by other stimuli, such as viruses. This suggests either molecular mimicry or a common alteration in the processing and presentation of antigens such that cryptic antigens are presented. Another possibility is that the reactive metabolite directly alters the class II MHC molecule leading to a graft-vs-host reaction.

Myeloperoxidase-mediated activation of xenobiotics by human leukocytes.

Peripheral blood leukocytes contain a variety of enzymes that are capable of metabolising xenobiotics. The enzyme myeloperoxidase (MPO) appears to be the most important for drug metabolism. MPO is a peroxidase/oxidase and generates the powerful oxidant hypochlorous acid. MPO- or MPO-generated oxidants are capable of oxidizing a wide variety of compounds and a broad range of functional groups, especially those that contain nitrogen and sulfur. Leukocytes have a role in immune response; therefore, reactive intermediates generated by leukocyte metabolism of xenobiotics may have a role in idiosyncratic drug reactions, particularly those that are immune-mediated such as drug-induced lupus or agranulocytosis.

Factors that modify penicillamine-induced autoimmunity in Brown Norway rats: failure of the Th1/Th2 paradigm.

Idiosyncratic drug reactions appear to be immune-mediated. Immune responses are driven by helper T cells (Th); Th1 responses promote cell-mediated immunity, whereas Th2 responses drive antibody-mediated reactions. Th1 cytokines inhibit Th2 responses and Th2 cytokines inhibit Th1 responses; therefore, it may be possible to prevent idiosyncratic drug reactions by changing the Th1/Th2 cytokine balance. We tested this hypothesis in an animal model in which penicillamine causes an autoimmune syndrome in Brown Norway rats. This syndrome has the hallmarks of a Th2-mediated response and we tried to inhibit it with a polymer of inosine and cytosine (poly I:C), a Th1 cytokine-inducer. However, we found that a single dose of poly I:C, given at the onset of penicillamine treatment, significantly increased both the incidence (100 vs. 60%) and accelerated the onset (30+/-4 vs. 39+/-5 days) of penicillamine-induced autoimmunity when compared with controls. To rule out other effects of poly I:C that might overshadow the induction of Th1 cytokines, we directly tested the effects of the prototypic Th1 cytokine, interferon-gamma. Although not as dramatic, interferon-gamma-pretreatment also appeared to make the syndrome worse. Conversely, when we used misoprostol, a prostaglandin-E analog that inhibits Th1 cytokines, it completely protected the animals. Just one dose of misoprostol prior to initiation of penicillamine treatment was sufficient to provide this protection. The syndrome was also completely inhibited by aminoguanidine, an inhibitor of iNOS. These results, although dramatic, suggest that the effects of these agents were not mediated by their effects on Th1/Th2 balance, but rather by some other mechanism.

Reactive intermediates in the oxidation of hydralazine by HOCl: the major oxidant generated by neutrophils.

The use of the antihypertensive hydralazine is associated with an autoimmune syndrome resembling systemic lupus erythematosus. Adverse drug reactions, such as drug-induced lupus, often involve reactive intermediates. Oxidation of hydralazine by liver microsomes or activated leukocytes leads to reactive intermediates that covalently bind to protein and may be involved in hydralazine-induced lupus. Oxidation of hydralazine to a reactive intermediate by cells involved in immune response, such as leukocytes, would be more likely to lead to an autoimmune reaction, such as drug-induced lupus, than would oxidation by cells in the liver. Leukocytes possess a defense system that generates HOCl in response to invading microorganisms. Hydralazine was oxidized to a reactive intermediate by HOCl generated by activated leukocytes. The reactive intermediate was trapped with N-acetylcysteine and the adduct was identified as 1-phthalazylmercapturic acid. The reactive intermediate is likely the diazonium salt of hydralazine. Two stable products were formed in the reaction, phthalazine and phthalazinone. Although phthalazine is oxidized to phthalazinone by HOCl, the rate of the reaction is much too slow to explain the rapid production of phthalazinone. It is more likely that most of the phthalazinone is formed by reaction of the putative diazonium salt with water. We propose that this reactive metabolite is responsible for hydralazine-induced lupus.

Predicting drug-induced agranulocytosis: characterizing neutrophil-generated metabolites of a model compound, DMP 406, and assessing the relevance of an in vitro apoptosis assay for identifying drugs that may cause agranulocytosis.

DMP 406 is a clozapine analogue developed by Dupont-Pharma for the treatment of schizophrenia. Unfortunately it caused agranulocytosis in dogs during preclinical studies. Clozapine also causes agranulocytosis and this is believed to be due to a reactive nitrenium ion metabolite produced by neutrophils. We studied the oxidation of DMP 406 by activated neutrophils and found that the major reactive species that is produced is not a nitrenium ion but rather an imine. This metabolite is similar to the reactive metabolite that has been proposed to be responsible for mianserin-induced agranulocytosis. Therefore we also studied the oxidation of mianserin by activated neutrophils and found that, although the major species is an iminium ion, it also bears a lactam moiety in the piperazine ring resulting from further oxidation. We usually find that HOCl is a good model system for the production of reactive metabolites of drugs that are formed by activated neutrophils, but in the case of both DMP 406 and mianserin, the products produced were significantly different than those formed by activated neutrophils. In contrast, the combination of horseradish peroxidase and hydrogen peroxide (HRP/H(2)O(2)) formed a very similar pattern of products, and this system was used to produce sufficient quantities of metabolites to allow for identification. The reactive metabolites of both DMP 406 and mianserin reacted with a range of nucleophiles, but in many cases the reaction was reversible. The best nucleophile for trapping these reactive metabolites was cyanide. It has been demonstrated that the products of clozapine oxidation by HRP/H(2)O(2), presumably the nitrenium ion, induced apoptosis in neutrophils at therapeutic concentrations of clozapine. It has been suggested that this process is involved in the mechanism of clozapine-induced agranulocytosis. We tested DMP 406 and mianserin in this system to see if the ability of a reactive metabolite of a drug to cause apoptosis could predict the ability of that drug to cause agranulocytosis. We used clozapine as a positive control and we also tested olanzapine, a drug that forms a reactive metabolite similar to that of clozapine but is given at a lower dose and does not cause agranulocytosis. We found that DMP 406 did not increase apoptosis at concentrations below 50 microM, and although mianserin did increase apoptosis at 10 microM this is above the therapeutic concentration. Olanzapine caused an increase in apoptosis at the same concentration as clozapine (1 microM), but because its therapeutic concentration is lower, this concentration was above the pharmacological range. There was no increase in apoptosis with any drug in the absence of HRP/H(2)O(2). These results indicate that this assay is unable to reliably predict the ability of different types of drugs to cause agranulocytosis. This is not a surprising result given that different drugs may induce agranulocytosis by different mechanisms.

Metabolism of drugs by leukocytes.

Neutrophils and monocytes can metabolize drugs to reactive metabolites, especially those drugs that have nitrogen or sulfur in a low oxidation state. The major system involved in this oxidation is the combination of NADPH oxidase and myeloperoxidase which generates HOCl. Although this system is unlikely to be quantitatively important, i.e. it is unlikely to have a significant effect on the pharmacokinetics of a drug, the reactive metabolites produced appear to have significant biological effects. Reactive metabolites, by their very nature, have short half-lives, and most of their effects will be exerted on the cells that formed them. Therefore, they are likely to be important for adverse reactions that involve leukocytes, such as agranulocytosis and immune-mediated reactions. However, the mechanism of these reactions is unknown and evidence for the association of leukocyte-derived reactive metabolites with such reactions is circumstantial at present. There is also circumstantial evidence to link the formation of such reactive metabolites to the antiinflammatory effects of some drugs. Possible mechanisms include the scavenging of other reactive species or inhibition of cells, especially neutrophils and macrophages, involved in inflammation. The oxidation of drugs by leukocytes requires activation of the cells; therefore, infection or other inflammatory conditions that activate leukocytes may represent one of the risk factors for idiosyncratic drug reactions.

Reactive metabolites and agranulocytosis.

Central to most hypotheses of the mechanism of idiosyncratic drug-induced blood dyscrasias is the involvement of reactive metabolites. In view of the reactive nature of the majority of such metabolites, it is likely that they are formed by, or in close proximity to the blood cells affected. The major oxidative system of neutrophils generates hypochlorous acid. We have demonstrated that the drugs associated with the highest incidence of agranulocytosis are oxidized to reactive metabolites by hypochlorous acid and/or activated neutrophils. There are many mechanisms by which such reactive metabolites could induce agranulocytosis. In the case of aminopyrine-induced agranulocytosis, most cases appear to involve drug-dependent anti-neutrophil antibodies, and these are likely to be induced by cell membrane antigens modified by the reactive metabolite of aminopyrine. The target of agranulocytosis associated with many other drugs is usually neutrophil precursors and may involve cytotoxicity or a cell-mediated immune reaction induced by a reactive metabolite. In the case of aplastic anaemia, there is evidence in some cases for involvement of cytotoxic T cells, which could either be induced by metabolites generated by neutrophils, or more likely, by reactive metabolites generated by stem cells.