• E-mail
• Print
• Comment
• Font Size
• Digg
• del.icio.us
• Discuss article

Toxicological Significance of Mechanism-Based Inactivation of Cytochrome P450 Enzymes By Drugs

Posted on: Tuesday, 12 June 2007, 03:00 CDT

By Masubuchi, Yasuhiro Horie, Toshiharu

Cytochrome P450 (P450) enzymes oxidize xenobiotics into chemically reactive metabolites or intermediates as well as into stable metabolites. If the reactivity of the product is very high, it binds to a catalytic site or sites of the enzyme itself and inactivates it. This phenomenon is referred to as mechanism-based inactivation. Many clinically important drugs are mechanism-based inactivators that include macrolide antibiotics, calcium channel blockers, and selective serotonin uptake inhibitors, but are not always structurally and pharmacologically related. The inactivation of P450s during drug therapy results in serious drug interactions, since irreversibility of the binding allows enzyme inhibition to be prolonged after elimination of the causal drug. The inhibition of the metabolism of drugs with narrow therapeutic indexes, such as terfenadine and astemizole, leads to toxicities. On the other hand, the fate of P450s after the inactivation and the toxicological consequences remains to be elucidated, while it has been suggested that P450s modified and degraded are involved in some forms of tissue toxicity. Porphyrinogenic drugs, such as griseofulvin, cause mechanism-based heme inactivation, leading to formation of ferrochelatase-inhibitory N-alkylated protoporphyrins and resulting in porphyria. Involvement of P450-derived free heme in halothane- induced hepatotoxicity and catalytic iron in cisplatin-induced nephrotoxicity has also been suggested. Autoantibodies against P450s have been found in hepatitis following administration of tienilic acid and dihydralazine. Tienilic acid is activated by and covalently bound to CYP2C9, and the neoantigens thus formed activate immune systems, resulting in the formation of an autoantibody directed against CYP2C9, named anti-liver/kidney microsomal autoantibody type 2, whereas the pathological role of the autoantibodies in drug- induced hepatitis remains largely unknown. Keywords Autoantibody, Cytochrome P450, Drug Interaction, Free Heme, Hepatotoxicity, Mechanism-Based Inactivation, N-Alkylprotoporphyrin, Reactive Metabolite

INTRODUCTION

Cytochrome P450 (P450) enzymes are hemoproteins that oxidize xenobiotics to generate hydrophilic drug metabolites. The P450- dependent oxidative metabolisms result in generation of chemically reactive metabolites or intermediates as well as stable metabolites. The reactive species thus formed covalently bind to hepatic macromolecules, which has been implicated in liver toxicities. The P450 enzymes that generate these reactive species are also targets of this covalent binding of them. Xenobiotics converted by the P450 enzymes to highly reactive intermediates that bind covalently to a catalytic site or sites of the enzyme itself leads to irreversible inhibition and inactivation of the enzymes. These compounds are commonly referred to mechanism-based inactivators or inhibitors. Although the rigorous criteria for mechanism-based inactivators have been established (Rando, 1984; Silverman, 1988), a broad definition of mechanism-based inactivators, which have been implicated as "catalysis-dependent" inactivators, has also been accepted. The commonly accepted mechanism-based inactivators of P450 enzymes include: (1) compounds that bind covalently to the apoprotein; (2) compounds that alkylate the porphyrin framework of the heme; (3) compounds that destroy the prosthetic heme group, leading to heme- derived fragments that covalently modify the apoprotein; and (4) compounds that quasi-irreversibly coordinate to the prosthetic heme iron atom, which is referred to metabolic intermediate (MI) complex formation (Correia and Oritz de Montellano, 2005; Halpert, 1995; Kent et al., 2001; Murray, 1987, 1997; Murray and Reidy, 1990; Oritz de Montellano and Correia, 1983; Zhou et al., 2004, 2005).

Mechanism-based inactivators of P450 enzymes are likely to be more specific than reversible inhibitors, because the former must be a substrate of a P450 and requires the metabolic activation by the enzyme for its inhibition. Because of the high specificity of the mechanism-based inactivators, they are used as therapeutic agents that exert their effects by selectively inhibiting target enzyme. For example, aromatase inhibitors, such as 4hydroxyandrostendione and exemestane, are mechanism-based inactivators of aromatase P450 and are used for the treatment of estrogen-dependent breast cancer to inhibit estrogen production in cancer tissues (Brueggemeier et al., 2005). Another application of the mechanism-based inactivators is for determining which P450 enzymes may be responsible for catalyzing a particular reaction in microsomal preparations. A typical and well-established mechanism-based inactivator used as a selective P450 inhibitor is furafylline for CYP1A2 (Kunze and Trager, 1993; Racha et al., 1998; Tassaneeyakul et al., 1994). Troleandomycin for C YP3 A4, diethyldithiocarbamate for C YP2E1, and orphenadorine for CYP2B6 are also widely used, although the specificity of the latter pairs has not been completely validated (Bourrie et al., 1996; Chang et al., 1994; Eagling et al., 1998; Ekins et al., 1997; Newton et al., 1995; Ono et al., 1996; Sai et al., 2000).

Because drug-induced mechanism-based inactivation of P450 is initiated by conversion of the drug into reactive species, this group of drugs may be involved in metabolism-dependent drug toxicities. On the other hand, because the reactive species limits its own generation and the resultant toxicity, the inactivation could be categorized as a defense system against the adverse drug reactions. A large number of therapeutic agents have been shown to act as a mechanism-based inactivator of P450 (Correia and Oritz de Montellano, 2005; Halpert, 1995; Kent et al., 2001; Murray, 1987, 1997; Murray and Reidy, 1990; Oritz de Montellano and Correia, 1983; Zhou et al., 2004,2005), which is independent of their pharmacological activities. The irreversible nature of the binding of the reactive intermediates to targeted P450 allows enzyme inhibition to be prolonged after elimination of the causal drug. Therefore, a typical outcome of the inactivation of P450 during drug therapy is inhibition of metabolism of the causal drug itself and coadministered drugs, resulting in serious drug interactions. Occasionally, the covalent binding of a reactive intermediate to P450 prosthetic heme or apo-P450 protein directly leads to liver dysfunction and other forms of toxicity, through the toxicity of the heme fragment derived from the destroyed heme group, or activation of immune systems by modified P450 protein (neoantigen). This article reviews the toxicological consequences of drug-induced inactivation of P450 enzymes.

THERAPEUTIC AGENTS THAT INACTIVATE HUMAN CYTOCHROME P450S

Many clinically important drugs are mechanism-based inactivators of P450 enzymes that include macrolide antibiotics, calcium channel blockers, and selective serotonin uptake inhibitors, which are not always pharmacologically related. These are structurally unrelated, although it is known that drugs containing several common moieties, such as an acetylene function, a tertiary amine function, and a furan ring, are metabolized by P450 enzymes into reactive metabolites or intermediates that bind to heme or protein moiety of the P450 molecule. In this section, therapeutic agents that behave as mechanism-based inactivators of P450 are categorized by their possible causal structures.

Acetylenes

A number of compounds containing the acetylenic functional group are mechanism-based inactivators of P450 enzymes. The reactive intermediates derived from the terminal acetylene have been known to alkylate the prosthetic heme group as well as to bind covalently to the protein of the P450 enzymes. The oxygenation of the internal carbon of acetylene would generate a reactive intermediate, alpha- ketocarbene, which would lead to N-alkylation of the prosthetic heme, whereas the oxygenation of the terminal carbon of the acetylene would lead to formation of a ketene intermediate, which can be hydrolyzed to a carboxylic acid product, or can acylate the active site of P450 (Figure 1) (Chan et al., 1993; Ortiz de Montellano and Komives, 1985).

Some estrogens and progestogens that have terminal acetylene moiety are known as mechanism-based inactivators of P450 (Figure 2). 17alpha-Ethinylestradiol, the major constituent of many oral contraceptives, is a mechanism-based inactivator of CYP3A4 (Guengerich, 1990; Lin et al., 2002). The inactivation of CYP3A4 by 17alpha-ethinylestradiol involves destruction of the heme and covalently binding to the apoprotein. 17alpha-Ethinylestradiol can also inactivate CYP2B6, whereas it only modifies the apoprotein in CYP2B6 (Kent et al., 2002), suggesting that the metabolic activation of this ethinyl compound can result in different reactivities toward heme versus apoprotein with different P450s. Other 17alpha- acetylenic steroids, gestodene, desogestrel, levonorgestrel, and norethisterone have also been shown to inactivate CYP3A4 (Guengerich, 1990). Among them, gestodene (Figure 2) is the most potent mechanism-based inactivator in terms of the pseudo-first- order constant of inactivation (k^sub inact^).

Mifepristone (RU-486), an antiprogestin agent, has an internal acetylene with a methyl group (Figure 2) and has also been characterized as a mechanism-based inactivator of CYP3A4. Mifepristone irreversibly modifies the CYP3A4 apoprotein, as shown by gel electrophoresis, rather than heme adduction and fragmentation, which were ruled out by a lack of change in the spectroscopic properties of CYP3A4 (He et al., 1999). It has been proposed that mifepristone is oxidized and converted via a 1,2- methyl shift to a reactive ketene intermediate (Figure IB) that reacts with a nucleophilic residue at the enzyme active site (He et al., 1999). The other major adult human liver CYP3A enzyme, CYP3A5, is not subject to mechanism-based inactivation by mifepristone, suggesting that CYP3A5 is unable to produce a putative ketene (Khan, He, Correia et al., 2002). Because mifepristone also has a tertiary amine structure (Figure 2) and is subject to W-demethylation by CYP3A4, it has also been reported that inactivation of CYP3A4 by mifepristone is associated with the formation of a nitrosoalkane- metabolic intermediate (MI) complex (see next section, Figure 3A) through the oxidation of the dimethylamino moiety (Jang and Benet, 1998). This was supported by the fact that lilopristone and onapristone, analogues of mifepristone without an acetylene moiety, can also inactivate CYP3A4 (Jang and Benet, 1998). However, mifepristone N-demethylation also occurs with CYP3A5, which is not inactivated by mifepristone (Khan, He, Correia et al., 2002). Therefore, it seems that mifepristone inactivates CYP3A4 mainly via the formation of the reactive ketene, while lilopristone and onapristone inactivate it via nitrosoalkane-MI complex formation. FIG. 1. The oxidation of terminal acetylenes by P450s leading to their heme or protein modification. The oxygenation of the internal carbon of the terminal acetylene would generate a reactive intermediate, N-ketocarbene, which would lead to N-alkylation of the prosthetic heme (A), whereas the oxygenation of the terminal carbon of acetylene would lead to formation of a ketene intermediate, which can be hydrolyzed to carboxylic acid product, or can acylate the active site of P450 (B).

FIG. 2. Steroids as mechanism-based inactivators of CYP3A4. An estrogen, 17a-ethinylestradiol, and a progestogen, gestodene, have a terminal acetylene moiety, while an antiprogestin, mifepristone has an internal acetylene moiety. Both of these functions may be involved in the inactivation of CYP3A4.

Amines and Related Drugs

Alkyl and aromatic amines, including a number of clinically useful antibiotics such as erythromycin, belong to a large class of agents that form quasi-irreversible MI complexes. These amines are oxidized to intermediates that coordinate tightly with the ferrous heme and give rise to a spectrum with an absorbance maximum at 445- 55 nm. Complex formation requires a primary amine but secondary and tertiary amines, which can give P450 complexes if the primary amine function is unmasked by N-demethylation reactions as in the case of troleandomycin. Then the primary amines thus formed are further oxidized by P450 enzymes involving oxidation of the nitrogen to nitroso species that coordinate to the iron (Figure 3A).

Macrolide antibiotics such as troleandomycin and erythromycin potently inhibit CYP3A4-mediated drug metabolism, resulting in many drug interactions (see next section ), and the former has been withdrawn from the market. These macrolides, which have a tertiary amine function, can be metabolized by CYP3A4 via N-demethylation to form reactive nitrosoalkanes that interact with P450 giving a nitrosoalkane-MI complex. A 14-membered ring has been regarded as an essential structure to form an MI complex, since similar drugs with a 16-membered rings, such as josamycin and miocamycin, generally do not form MI complexes (Periti et al., 1992; von Rosensteil and Adam, 1995). However, among the 14-membered macrolides, clarithromycin moderately interacts and dirithromycin poorly interacts with CYP3A4 (Lindstrom et al., 1993; Periti et al., 1992; von Rosensteil and Adam, 1995). The ability to form an MI complex may be accounted for by various factors, including lipophilicity, steric hindrance, and pK^sub a^ value, and it is thus a good in vitro marker to estimate clinical drug interactions. However, there are some discrepancies, such as azithromycin and dirithromycin, which poorly interfere with CYP3A4 in vitro but exhibit a potential for drug interactions in clinical settings (Westphal, 2000).

FIG. 3. Mechanisms for quasi-irreversible coordination of drugs to prosthetic heme iron atom, which is referred to as metabolic intermediate (MI) complex formation. (A) Tertiary and secondary amines can give P450 complexes if the primary amine function is unmasked by N-demethylation reactions, which are further oxidized to nitroso species that coordinate with the iron. (B) Methylenedioxy compounds are subject to oxidation of the methylene bridge to a species, probably carbene that forms a tight but reversible complex with a heme iron atom.

Recently some calcium-channel antagonists (Figure 4) have been identified as CYP3A4 inactivators, and these include diltiazem, verapamil, and nicardipine (Jones et al., 1999; Ma et al., 2000; Wang et al., 2004; Yeo and Yeo, 2001). These compounds contain an amine functional group and undergo /V-dealkylation, resulting in Mi- complex formation (Jones et al., 1999; Ma et al., 2000; Wang et al., 2004; Yeo and Yeo, 2001 ), while other mechanisms have also been suggested, such as potent reversible inhibition by stable metabolites (Ma et al., 2000). Furthermore, CYP3A5 is less susceptible to the inactivation commonly induced by these drugs (McConn et al., 2004; Wang et al., 2005), probably because CYP3A5 has lower ability to generate the putative nitrosoalkane intermediates from these drugs. A relatively new calcium channel antagonist, mibefradil (Figure 4), which was withdrawn from the market due to toxic drug interactions, has been demonstrated to be a potent mechanism-based inactivator of CYP3A4 (Prueksaritanont et al., 1999). Unlike other calcium-channel antagonists, catalysis- dependent inhibition of CYP3A4 by mibefradil is irreversible (Lim et al., 2005; Ma et al., 2000) and it may not be metabolized into primary amines (Welker et al., 1998). Thus, an MI complex may not be involved in the CYP3A4 inactivation by mibefradil, while no alternative mechanism has been provided. Amiodarone, an antiarrhythmic agent, inactivates CYP3A4, which has a tertiary amine moiety, but it is not known if a nitrosoalkane MI complex is responsible for the inactivation. The W-deethylated metabolite inactivates not only CYP3A4 but also other P450s, such as CYP2D6 (Ohyama et al., 2000).

FIG. 4. Calcium-channel antagonists as mechanism-based inactivators of CYP3A4. These calcium-channel antagonists are structurally unrelated but have tertiary amine functions, which may be N-dealkylated to primary amines (except mibefradil), followed by oxidation to nitroso species to form MI complexes.

FIG. 5. Selective serotonin uptake inhibitors as mechanismbased inactivators of P450s. Fluoxetine, a secondary amine, inactivates CYP3A4 probably via formation of a nitrosoalkane Mi-complex (Figure 3A). Paroxetine, a methylenedioxy compound, inactivates CYP2D6 via oxidation of the methylene bridge to a species, probably carbene, which forms an MI complex (Figure 3B).

Fluoxetine, a selective serotonin uptake inhibitor, has a secondary amine function (Figure 5), which inactivates CYP3A4, probably via formation of a nitrosoalkane MI complex (Figure 3A) with CYP3A4 (Mayhew et al., 2000). On the other hand, paroxetine, another selective serotonin uptake inhibitor, potently inactivates CYP2D6 (Bertelsen et al., 2003). Paroxetine is a unique example of a drug that is a quasi-irreversible P450 inactivator with methylenedioxy moiety. The oxidation of the methylene bridge to a species, probably carbene, forms a tight but reversible complex with a heme iron atom (Figure 3B, Figure 5).

Furans

8-Methoxypsoralen (methoxsalen) is one of the psoralens (Figure 6), a family of furanocoumarin derivatives, and has been used to treat diseases like psoriasis, vitiligo, and cutaneous T-cell lymphoma. Methoxsalen has been shown to inactivate CYP2A6 (Koenigs et al., 1997; Koenigs and Trager, 1998; Tinel et al., 1987). The mechanism of inactivation by methoxsalen appears to be an initial oxidation to generate an epoxide that reacts with a nucleophilic amino acid at the active site of CYP2A6 (Koenigs and Trager, 1998). On the other hand, 5-methoxypsoralen (bergapten), an isomer of methoxsalen (Figure 6), which is a constituent of Seville orange juice, was found to inactivate CYP3A4 (Malhotra et al., 2001 ). Furanocoumarins in grapefruit juice, bergamottin (Figure 6), 6',7'- bergamottin, and furanocoumarin dimers, inactivate CYP3A4 (He et al., 1998; Schmiedlin-Ren et al., 1997; Tassaneeyakul et al., 2000). The inactivation by bergamottin may involve apoprotein modification. A recent report suggests that bergamottin also inactivates CYP3A5 and CYP2B6 through heme destruction in addition to covalent binding to protein (Lin et al., 2005).

The furanopyridine L-754,392 was a potential HIV protease inhibitor but its further development was discontinued because of its potent inhibition of CYP3A4 (Chiba et al., 1995; Lightning et al., 2000; Sahali-Sahly et al., 1996). The inactivation of CYP3A4 by L-754,392 involves the intact furan ring and proceeds via initial epoxidation of the furanopyridine, followed by epoxide ring opening catalyzed by an active-site nucleophile (Chiba et al., 1995; Lightning et al., 2000; Sahali-Sahly et al., 1996). Furafylline, which was a potential replacement for theophylline in the treatment of asthma, had its clinical development terminated and it is now used experimentally as a specific inhibitor of CYPl A2(Kunze and Trager, 1993;Rachaetal., 1998; Tassaneeyakul et al., 1994). Furafylline inactivates CYPl A2 via apoprotein modification, while it is suggested that the xanthine, but not the furan moiety of furafylline, is directly involved in the inactivation (Racha et al., 1998). FIG. 6. Psoralens as mechanism-based inactivators of P450s. 8- Methoxypsoralen (methoxsalen) appears to undergo an initial oxidation to generate an epoxide that reacts with the apoprotein of CYP2A6. 5-Methoxypsoralen (bergapten) and bergamottin may also be converted into similar epoxides, and mainly inactivate CYP3A4.

Sulfur Containing Compounds

Tienilic acid, a substituted thiophene, is a diuretic drug (Figure 7) and was withdrawn from the market because of its hepato- and renal toxicity. Tienilic acid is oxidized by CYP2C9 to form 5- hydroxytienilic acid and a product covalently binds to the protein at the active site, leading to enzyme inactivation. Evidence suggests that formation of an electrophilic thiophene sulfoxide and/ or a thiophene epoxide that reacts with the CYP2C9 protein nucleophile inactivates the enzyme (Jean et al., 1996; Koenigs et al., 1999; Lopez-Garcia et al., 1994; Lopez Garcia et al., 1993). A recent study with rat liver microsomes identified glutathione (GSH) adducts of two reactive intermediates, a thiophene S-oxide and a thiophene epoxide, after the oxidation of 2-phenylthiophene, a model thiophene compound (Dansette et al., 2005). The nonsteroidal anti- inflammatory agent suprofen is also a mechanism-based inactivator of CYP2C9 (O'Donnell et al., 2003). Similar to tienilic acid, suprofen is a substituted thiophene (Figure 7) and may be subject to epoxidation of thiophene ring leading to covalent modification of apoprotein. Ticlopidine, another thiophene (Figure 7), which is clinically used as an anti-platelet aggregation agent, is the first selective mechanism-based inactivator of CYP2C19 (Ha-Duong et al., 2001). It is suggested that a similar electrophilic intermediate to tienilic acid, possibly thiophene S-oxide, is involved in the inactivation of CYP2C19. A more recent study suggests that CYP2B6 is also effectively inactivated by ticlopidine, as well as by clopidogrel, a related thienopyridine anti-platelet agent (Richter et al., 2004).

FIG. 7. Thiophenes as mechanism-based inactivators of P450s. Tienilic acid is oxidized to an electrophilic thiophene sulfoxide and/or a thiophene epoxide that reacts with CYP2C9 protein nucleophile to inactivate the enzyme. Suprofen also inactivates CYP2C9. Ticlopidine is a mechanism-based inactivator of CYP2C19 via conversion of a similar electrophilic intermediate to tienilic acid, possibly thiophene S-oxide.

Ritonavir, an HIV-I protease inhibitor, which has thiazolyl groups, inactivates CYP3 A4, implying that this functional group plays an important role in the enzyme inactivation (Koudriakova et al., 1998). On the other hand, other HIV-I protease inhibitors without the thiazolyl moiety but with an amine, such as amprenavir and nelfinavir, inactivate CYP3A4 via formation of an MI complex (Ernest et al., 2005; Lillibridge et al., 1998). Another sulfur- containing drug, troglitazone, is an antidiabetic thiazolidinedione. It was withdrawn from the market because of idiosyncratic hepatotoxicity. A recent study demonstrated that troglitazone was a mechanism-based inactivator of CYP3A4, and the other thiazolidinediones, rosiglitazone and pioglitazone, also inactivated CYP3A4, but to lesser extent than troglitazone (Lim et al., 2005). It is possible that reactive metabolites formed from oxidative cleavage of the common thiazolidinedione ring and/or a quinone methide formed from the troglitazone-specific chromane moiety, both of which have been identified as their GSH adducts (Kassahun et al., 2001), are involved in the inactivation of CYP3A4. Disulfiram and its primary metabolite, diethyldithiocarbamate, are mechanism-based inactivators of CYP2E1 (Guengerich et al., 1991). Diethyldithiocarbamate has also been shown to inhibit other P450s such as CYP2A6 and CYP2C19(Changetal., 1994;Onoetal., 1996; Sai et al., 2000). No metabolite of disulfiram directly involved in inactivation of the P450s has been identified.

Other Miscellaneous Compounds

Some drugs that do not belong to the classification just described are mechanism-based inactivators of P450, particularly CYP3A4. Midazolam, a sedative drug, inactivates CYP3A4, probably via midazolam A-hydroxylation (Khan, He, Domanski et al., 2002). It has also been suggested that protein modification rather than heme destruction leads to CYP3A4 inactivation (Khan, He, Domanski et al., 2002). The nonsteroidal anti-inflammatory agent diclofenac, at high concentrations, inactivates CYP3A4 (Masubuchi et al., 2002). It has been suggested that a 5-hydroxylation step of diclofenac but not a p- benzoquinone imine derivative of 5-hydroxydiclofenac, a proposed toxic metabolite, is critical for the inactivation of CYP3A4 (Masubuchi et al., 2002). Tamoxifen, a nonsteroidal antiestrogen, is used for the treatment of hormone-dependent breast cancer. Inactivation of CYP3A4 by tamoxifen and Ndesmethyltamoxifen was found and was proposed to result from formation of MI complex (Zhao et al., 2002). On the other hand, another study observed a mechanism- based inactivation of CYP2B6 and also, partially, CYP2D6 and CYP2C9, but not CYP3A4 (Sridar et al., 2002). It is proposed that further metabolism of 4-hydroxytamoxifen leads to the formation of a catechol that could then react with the CYP2B6 apoprotein (Sridar et al., 2002). Raloxifene is a selective estrogen modulator and is effective in the treatment of osteoporosis. It inactivates CYP3A4, and analysis of the GSH-adduct suggests CYP3A4-mediated formation of raloxifene arene oxides or a reactive quinone intermediate, which modifies the CYP3A4 apoprotein (Chen et al., 2002). Delavirdine, an HIVreverse-transcriptase inhibitor, also inactivates CYP3A4, and this may occur via apoprotein modification (Voorman et al., 1998). Involvement of one of the CYP3A4-dependent metabolites, 6'- hydroxydelavirdine, or another unidentified CYP3A4-dependent metabolite has been implicated in the inactivation (Voorman et al., 1998). Nefazodone, a nontricyclic antidepressant, induces hepatotoxicity as well as metabolism-mediated inactivation of CYP3A4 (Kalgutkaret al., 2005). An electrophilic metabolite, 2-chloro-1,4- benzoquinone, was proposed to be involved in the hepatotoxicity, but its role in the CYP3A4 inactivation is unknown (Kalgutkaret al., 2005). Recently, several drugs that are known as mechanism-based inactivators of CYP3A4, such as verapamil, amiodarone, and fluoxetine, were also found to inactivate CYP2C8 (Polasek et al., 2004).

DRUG INTERACTIONS CAUSED BY INACTIVATION OF CYTOCHROME P450 ENZYMES

Pharmacokinetic drug interactions often occur as a change in hepatic drug metabolism. As well as reversible inhibition, the inactivation of P450s during drug therapy results in a reduced rate of drug metabolism. Unlike reversible inhibition, the enzyme inhibition by mechanism-based inactivators is prolonged even after elimination of the causal drug, because enzyme activities can only be restored by de novo protein synthesis. Thus, the mechanism-based inactivation may cause severe drug toxicity due to prolonged elevations in the plasma concentrations of coadministrated drugs. However, not all drug interactions are necessarily of clinical importance and there are several contributory factors. Drugs causing interactions and drugs that are interacted with by other agents were recently referred to as "perpetrators" and "victims," respectively (Venkatakrishnan et al., 2003). In general, adverse drug reactions are regarded as side effects of "victims," the plasma concentrations of which are significantly elevated by "perpetrators." Thus, the potential to develop an adverse drug reaction is determined by the nature of both perpetrators and victims as follows. For the perpetrator, the relative potency as a P450 inhibitor (potency of inactivation, in this case, i.e., low K^sub i^ and large k^sub inact^) in the clinically relevant concentration range of each drug must be a determinant to cause adverse drug reaction. For the victim, P450-mediated metabolism must represent its major elimination pathway, and then the impaired metabolism must produce a significant alteration in its plasma concentration profile relative to the concentration-response relationship. Moreover, drug interactions generally lead to serious toxicities only if the victims have a narrow therapeutic index. We focused on drug interactions induced by mechanism-based inactivator-type perpetrators, which resulted in various adverse reactions of victim drugs. In this section, victims are categorized by their adverse reactions. In some cases, such drug interactions have been responsible for the recent withdrawal of several drugs from the market, including both perpetrators and victims.

Nonsedating Antihistamines and Other Drugs That Induce Torsades de Pointes

Torsades de pointes is a life-threatening ventricular arrhythmia that occurs in the setting of electrocardiographic QT interval prolongation. It has been shown to occur with the nonsedating antihistamines terfenadine and astemizole (Paakkari, 2002; Woosley, 1996), which were recently withdrawn from the market, because they led to the QT interval prolongation. These drugs appear to act in a concentration-dependent manner to block delayed rectifier potassium current in cardiac conduction pathways, which is the basis for QT interval prolongation (Paakkari, 2002; Woosley, 1996). The extensive presystemic elimination of these drugs results in their low circulating plasma concentrations. Conditions in which plasma concentrations are markedly elevated are associated with increased risk of developing this arrhythmia. Because the elimination of terfenadine and astemizole is predominantly determined by CYP3A4 metabolism, marked elevation in their plasma concentrations, which reach a level sufficient to cause torsades de pointes, is caused by coadministration of CYP3A4 inhibitors including mechanism-based inactivators such as erythromycin, clarithromycin, diltiazem, ritonavir, and grapefruit juice components (Dresser et al., 2000; Kivisto et al., 1994; Paris et al., 1994; Thummel and Wilkinson, 1998). Terfenadine and astemizole are prodrugs that are metabolized primarily by CYP3A4 to their therapeutically active metabolites, fexofenadine and norastemizole, respectively. The parent drugs, but not the active metabolites, are responsible for the cardiotoxic effect. Therefore, fexofenadine is now marketed instead of terfenadine (Handley et al., 1998). Members of other classes of drugs-a gastrointestinal promotility agent, cisapride (Bedford and Rowbotham, 1996; Michalets and Williams, 2000), and an antipsychotic drug, pimozide (Desta et al., 1999)-also cause torsades de pointes when their metabolism is inhibited, thereby reaching concentrations high enough to cause the cardiotoxic effect. Cisapride was withdrawn from the market because of heart rhythm abnormalities similar to those caused by terfenadine. These drugs are also metabolized extensively by CYP3A4, and inactivation of CYP3A4 by macrolide antibiotics and grapefruit juice components led to drug interactions and cardiotoxic effects.

Theophylline

Theophylline is subject to drug interactions because of its narrow therapeutic index and extensive metabolism by P450s. Decreases in the rate of metabolism can lead to serious toxicities, including convulsions, and heart arrhythmias that can be serious enough to cause death. It has been reported that fluvoxamine enhances these serious theophylline toxicities (DeVane et al., 1997; van den Brekel and Harrington, 1994). Theophylline is mainly oxidized to inactive metabolites by CYP1A2, and its inhibition by fluvoxamine, a mechanism-based inactivator of CYPl A2, was demonstrated (Rasmussen et al., 1995). Furthermore, the inhibition potency by fluvoxamine is much greater in vitro than in vivo (Yao et al., 2001). It has also been reported that in vivo theophylline metabolism is impaired by macrolide antibiotics and some calcium channel blockers, while it is not known whether these amine drugs act as mechanism-based inactivators of CYP3A4 or other P450s (Upton, 1991).

HMG-CoA Reductase Inhibitors

The HMG-coenzyme A (CoA) reductase inhibitors (statins) are an important class of cholesterol-lowering medications. Toxic adverse reactions induced by the statins are myopathy and rhabdomyolysis, which occur under conditions in which the plasma concentrations of parent drug and metabolite are markedly elevated by perturbations including metabolic drug interactions. Lovastatin, simvastatin, atorvastatin, and cerivastatin are metabolized mainly by CYP3A4, whereas fluvastatin is metabolized by CYP2C9, and pravastatin by non- P450-dependent pathways (Corsini et al., 1999). Thus, mechanism- based inactivators, such as macrolide antibiotics, HIV protease inhibitors, and grapefruit juice components, caused marked increases in the plasma concentrations of lovastatin, simvastatin, and cerivastatin, leading to their toxicities, myopathy and rhabdomyolysis (Christians et al., 1998; Corsini et al., 1999; Dresser et al., 2000; Herman, 1999; Williams and Feely, 2002). Such interactions as well as those with other drugs like gemfibrozil resulted in the withdrawal of cerivastatin from the market because of deaths from rhabdomyolysis.

Mibefradil, a benzimidazole-containing calcium channel blocker (Figure 4), behaves as a mechanism-based inactivator of CYP3A4 (discussed earlier), leading to potent inhibition of the several drugs and resultant toxicities, including lifethreatening rhabdomyolysis in patients on lovastatin and simvastatin (Corsini et al., 1999; Krayenbuhl et al., 1999). Mibefradil was finally withdrawn after a study that showed a trend to higher mortality in patients on statins and given mibefradil. The toxicities and increase in mortality were the side effects of the victim drugs, statins in this case, exacerbated by the multiple interactions caused by the perpetrator, mibefradil (see next subsection).

Calcium-Channel Antagonists and Other Drugs That Induce Symptomatic Hypotension

Hypotension is a dose-dependent adverse effect of many antihypertensive drugs. It occurs by dihydropyridine calcium antagonists when mechanism-based inactivators of CYP3A4 are combined (Anderson and Nawarskas, 2001). Pharmacokinetic drug interactions between grapefruit juice and dihydropyridine calcium antagonists are widely known, while the changes in blood concentration by the interaction differ markedly between victim drugs (Dresser et al., 2000). Significant drug interactions caused by mibefradil occur with dihydropyridines and beta-blockers. It has been reported that nifedipine dosing 24 h after the last dose of mibefradil causes hypotension and bradycardia, resulting in death of the patient (Mullins et al., 1998).

Sildenafil (Viagra), a selective inhibitor of cyclic guanosine monophosphate (cGMP)-specific phosphodiesterase type 5 (PDE5), is available for the treatment of patients with erectile dysfunction. It inhibits the degradation of cGMP from the nitric oxide (NO) pathway, producing selective vasodilation in the corpus cavernosal. However, under conditions of excessive sildenafil exposure by the inactivation of CYP3A4, a major enzyme metabolizing sildenafil, particularly with nitrate therapy, systemic vasodilation is observed, resulting in fatal hypotension and myocardial infarction (Cheitlin et al., 1999; Simonsen, 2002).

Other Victims

Carbamazepine is a relatively safe antiepileptic drug that is subject to dose-dependent neurologic toxicities. Since CYP3A4 is a major enzyme mediating carbamazepine metabolism, its inhibition by mechanism-based inactivators of CYP3A4, such as macrolide antibiotics and calcium-channel antagonists has been shown to cause ataxia (Patsalos et al., 2002; Spina et al., 1996). Psychopharmacological agents are also susceptible to CYP3A4- mediated drug interactions. The plasma levels of benzodiazepines, such as midazolam, triazolam, and alprazolam, are enhanced markedly by some macrolides, selective serotonin reuptake inhibitors, and calcium-channel antagonists such as verapamil and diltiazem, leading to potentially hazardous excessive sedation (Tanaka, 1999; Yuan et al., 1999). Toxicities associated with tricyclic antidepressants as a result of inhibition of P450, particularly CYP2D6, include bradycardia, seizures, and delirium. Paroxetine, an inactivator of CYP2D6, has been shown to significantly increase the plasma concentrations of tricyclic antidepressants such as imipramine and desipramine, with signs of toxicity in some cases. These and other clinically important drug interactions due to mechanism-based inactivators of P450s are included in comprehensive reviews of drug interactions (Levy et al., 2000; Rodrigues, 2002).

TOXICITIES DERIVED FROM THE HEME OF CYTOCHROME P450 MODIFIED BY REACTIVE METABOLITES

The fate of the P450 molecule after inactivation by chemical agents is unclear. However, several lines of evidence suggest that modified or destroyed P450 enzymes are sometimes involved in liver and other tissue toxicities. In particular, P450 hemereactive metabolite adduct formation results in modified heme, free heme, and free iron, which should lead to the toxicities. In this section, we focus on basic studies of toxicological outcomes resulting from modified heme of P450s, which are produced via their mechanism- based inactivation, by chemicals including therapeutic agents.

Xenobiotic-lnduced Disruption of Hepatic Heme Homeostasis and Porphyrinogenic Reactions

Some of mechanism-based inactivators of P450s are known to deplete hepatic heme and trigger acute attacks of hepatic porphyria in genetically predisposed individuals (Halpert et al., 1994; Smith and De Matteis, 1980). Heme synthesis is regulated by its own negative feedback mechanism at the level of its rate-limiting enzyme 5-aminolevulinic acid synthase (ALAS) (Figure 8). The major portion of the heme synthesized in hepatocytes is consumed by the synthesis of P450 enzymes. Destruction of P450 prosthetic heme by chemical agents results in acute hepatic heme depletion, which leads to a stimulating effect on ALAS by increased transcription, translation, and the translocation of ALAS from the cytosol to the mitochondria. As a result of the elevated level of 5-aminolevulinic acid (ALA), porphyrin biosynthesis is accelerated. Therefore, destruction of P450 enzymes leads to accumulation of porphyrins but to only a modest extent (De Matteis, 1988; Oritz de Montellano and Correia, 1983). There is an additional mechanism for depleting free heme and accumulation of porphyrins in the liver as follows.

Ferrochelatase is a mitochondrial enzyme that introduces iron into protoporphyrin, the last step in heme biosynthesis (Figure 8). Lower activities of this enzyme lead to defective heme biosynthesis and accumulation of protoporphyrin in some tissues such as liver and red blood cells. Porphyrinogenic reactions by the mechanism-based inactivators of P450s that alky late the porphyrin framework of the heme have been proposed to proceed via following several steps (De Matteis, 1988; De Matteis et al., 1987; De Matteis and Marks, 1996; Marks et al., 1988): (1) During biotransformation of these xenobiotics by certain P450 enzymes, they are converted within the active site of the P450 molecule into highly reactive alkyl radicals; (2) the reactive species can alkylate one of the four pyrrole nitrogens of the heme moiety at the active site of the P450, resulting in inactivation of the P450 enzyme; (3) the modified heme of the P450 dissociates from the apoprotein P450 and its iron is released, yielding N-alkylprotoporphyrin IX (N-alkylPP, Figure 9), which is generally composed of 4 isomers; (4) some of the N- alkylPPs (such as N-methylPP) potently inhibit ferrochelatase; and (5) hepatic ferrochelatase inhibition by N-alkylPPs leads not only to reduced heme production but also accumulation of a large amount of porphyrins and other heme precursors and eventually porphyria. This scheme evolved from extensive studies with potent porphyrinogenic agents, such as 4-alkyldihydropyridines, as described in the next section. FIG. 8. Postulated mechanisms of chemical-induced protoporphyria. Heme synthesis is regulated by its own negative feedback at the rate-limiting enzyme 5-aminolevulinic acid synthase (ALAS). Acute hepatic heme depletion caused by destruction of P450 prosthetic heme leads to stimulating effect on ALAS and an elevated level of 5-aminolevulinic acid (ALA), followed by accelerated porphyrin biosynthesis. On the other hand, alkylation of the heme moiety P450 by porphyrinogenic agents results in dissociation of the modified heme from the apoprotein P450 with the release of its iron, giving N-alkylprotoporphyrin IX (N-alkylPP, Figure 9). Some of the N-alkylPPs, such as N-methylPP, potently inhibit ferrochelatase, the enzyme that introduces iron into protoporphyrin as the last step in the heme biosynthesis. Ferrochelatase inhibition by N-alkylPPs leads not only to a decrease in heme production but also to accumulation of a large amount of porphyrins.

FIG. 9. Structure of N-alkylprotoporphyrin IX. One of the four possible isomers is shown, where the pyrrole nitrogen on ring A is alkylated.

The porphyrinogenic effects of several xenobiotics have been shown to depend on their ability to alkylate prosthetic heme and to yield corresponding N-alkylPP, and the ability of resultant N- alkylPP to inhibit ferrochelatase. Therefore, structural features are critical determinants of the induction of protoporphyria. Historically, drug-induced protoporphyria has been studied with experimental animals, such as rodents and chicks, and subsequently also with human liver microsomes. Recent studies found that N- methylPP was a more potent inhibitor of human ferrochelatase than those of experimental animals (Gamble et al., 2000), suggesting the clinical significance of inhibition of ferrochelatase by N-alkylPPs in the pathogenesis of drug-induced porphyria.

Formation of N-Alkylprotoporphyrin IX After Interaction of Liver Microsomes with 4-Alkyldihydropyridines

Extensive studies have been performed for formation of N-alkylPP in liver microsomes with porphyrinogenic 4-alkyldihydropyridines (4- alkyl analogous to 3,5-diethoxycarbonyl-1,4-dihydrocollidine, DDC, Figure 10). A green pigment with inhibitory effects on ferrochelatase was isolated from the livers of mice and rats treated with DDC, but not with the nonporphyrinogenic analogue 3,5- diethoxycarbonylcollidine (DC) (Tephly et al., 1979). Then N- methylPP was identified as a potent inhibitor of ferrochelatase in the liver of mice, rats, and chick embryos (Cole et al., 1981; De Matteis et al., 1981; Ortiz de Montellano et al., 1981; Tephly et al., 1981). Involvement of P450 was further demonstrated. 4- Alkyldihydropyridines with specific 4-alkyl functional groups are converted into the corresponding radical cation, leading to elimination of the 4-alkyl group as an alkyl radical (Figure 10), and these reactions are mediated by P450 (Augusto et al., 1982). Other studies with similar 4-alkyldihydropyridines revealed the loss of P450 and corresponding production of N-alkylPPs (de Matteis et al., 1982) with ferrochelatasereducing activity (Marks et al., 1985). Therefore, a source of N-alkylPP was identified as the prosthetic group of P450 heme that was alkylated through mechanism- based inactivation of P450 by DDC and its analogues. In addition, for a N-alkylPP to accumulate, the alkyl group needs to be of a special size and shape (methyl, ethyl, and propyl), because bulky residues would be sterically hindered in their access to the pyrrole nitrogen atoms and would, instead, attack more accessible sites in the molecule of heme (De Matteis et al., 1987; de Matteis et al., 1982).

FIG. 10. Postulated mechanisms of oxidation of 4-alkyl- and 4- aryldihydropyridines. 4-Alkyldihydropyridines with specific functional groups, such as 4-methyl- (3,5-diethoxycarbonyl-1,4- dihydrocollidine, DDC) and 4-ethyl- (3,5-diethoxycarbonyl-2,6- dimethyl-4-ethyl-1,4-dihydropyridine, DDEP), are converted into the corresponding radical cations, leading to elimination of the 4- alkyl group as an alkyl radical, which alkylates the prosthetic group of P450 heme and forms N-alkylPP (closed arrow). 4Aryldihydropyridines, such as nifedipine, may also be converted into radical cations, which are metabolized to the corresponding pyridine derivatives retaining the 4-aryl group (open arrow).

The 4-ethyl analogue of DDC, 3,5-diethoxycarbonyl-2,6dimethyl-4- ethyl- 1,4-dihydropyridine (DDEP), has been used for further identification of P450 subject to the inactivation. DDEP inactivates rat CYPlA, CYP2C6, CYP2C11, and CYP3A1/2 (Correia et al., 1987; Kimmett et al., 1994; Riddick et al., 1989; Tephly et al., 1986). Each P450 follows a different fate after inactivation, namely, CYP2C6 and CYP2C11 are destroyed by N-ethylation of hemes, and DDEP- inactivated CYP2C11 is more labile and more readily proteolysed than CYP2C6 (Correia et al., 1987; Correia, Yao et al., 1992). On the other hand, the prosthetic heme destruction products of CYP3A1/2 irreversibly bind to the P450 apoprotein. This results in rapid degeneration by the cytosolic ubiquitin-dependent proteolytic system (Correia, Davoll et al., 1992; Correia et al., 1987; Correia, Yao et al., 1992). Thus, CYP2C6 and CYP2C11, rather than CYP3A1/2, have been identified as major sources of N-ethylPP in rat livers. Gender differences in rats (male > female) that were attributable to CYP2C11-mediated metabolism have been observed for DDEP-induced N- ethylPP formation (Wong et al., 1998). Besides conversion of the heme moiety of P450 to N-alkylPPs, it leads to formation of heme- derived protein adducts. However, the weak porphyrinogenicity of bulky DDC analogues, such as a 4-isopropyl derivative, cannot be explained by utilization of heme for the adducts instead of N- alkylPP formation (Riddick and Marks, 1990).

DDEP inactivates CYP3A4 in human liver microsomes, but the DDEP radical is stable enough to inactivate other P450s. Thus, it may not be a strictly mechanism-based inactivator (Bocker and Guengerich, 1986). This was supported by the fact that an ethyl radical can be trapped with a nitrone spin-trap reagent (Augusto et al., 1982). Studies with cDNA-expressed human P450s in human lymphoblastoid cell lines revealed the mechanism-based inactivation of CYP IAl, CYP1A2, C YP2C9, and CYP3A4 by DDEP (McNamee et al., 1997). On the other hand, DDEP-induced N-alkylPP formation with CYP2C9 was detected with this system (Gamble et al., 2003), and that with CYP1A2 and CYP2C9 was also detected with baculovirus-infected insect cells, but not in human liver microsomes (Lavigne et al., 2002). Thus, DDEP-induced inactivation of other P450s such as CYP3A4 may not lead to N- alkylPP formation.

Calcium Antagonists as Porphyrinogenic Agents

Dihydropyridine calcium antagonists such as nifedipine and nitrendipine possess similar structures to 4-alkyldihydropyridines, but their 4-positions are aryl substituents. These 4- aryldihydropyridines are metabolized to corresponding pyridine derivatives containing the 4-aryl group, whereas 4- alkyldihydropyridines formed those in which a hydrogen atom is present at the 4-position and the alkyl group is lost in human liver microsomes (Figure 10). Therefore, only the latter compounds inactivate P450 via heme alkylation by alkyl radicals (Bocker and Guengerich, 1986). Although nifedipine was found neither to be a mechanism-based inactivator of P450 nor to exhibit ferrochelatase- lowering activity in chick embryo liver cells, it possesses a potent porphyrinogenic activity (Marks et al., 1986). Because the major porphyrin that accumulates in response to nifedipine is uroporphyrin, nifedipine might cause porphyria not by ferrochelatase inhibition, but by inhibition of uroporphyrinogen decarboxylase (Marks et al., 1986). It has been suggested that oxidative stress by free radicals could block the latter enzyme (De Matteis, 1988; Marks et al., 1986). Recent studies indicated an important role of P450, particularly CYP1A2, in chemical-induced uroporphyria (German et al., 2002; Uno et al., 2004). Accumulation of porphyrins other than protoporphyrin was also observed with some lipophilic 4- alkyldihydropyridines that do not inhibit ferrochelatase; for example, 4-isopropyldihydropyridine accumulated coproporphyrin, and 4-benzyl analogous accumulated uro- and heptacarboxylic acid porphyrin (Marks et al., 1985).

The Mouse Model of Griseofulvin-lnduced Porphyria

The antifungal drug griseofulvin develops hepatic porphyria and liver injury in mice (Knasmuller et al., 1997). Griseofulvin generated green pigment, and the corresponding ferrochelatase- inhibitory porphyrin is spectrally similar to the N-alkylPP obtained from the DDC-treated liver (De Matteis and Gibbs, 1980; De Matteis et al., 1987). Subsequently, two N-alkylPPs were isolated from griseofulvintreated mouse liver, the major component of which was N- griseofulvinPP, which is unable to inhibit ferrochelatase (Bellingham et al., 1995; Holley et al., 1991). The minor one is N- methylPP, which may arise as a secondary product from the major one (Holley et al., 1991). Because incubation of NgriseofulvinPP gained inhibitory potency to ferrochelatase, it was concluded that N- griseofulvinPP elicited the inhibitory effects through conversion to N-methylPP (Figure 11 ) (De Matteis and Marks, 1996). Although limited studies have been available for destructive effects of griseofulvin on P450s (Denk et al., 1977; Williams and Simonet, 1986), involvement of P450 in N-methylPP generation was suggested (De Matteis et al., 1991). On the other hand, griseofulvin is an inducer of P450, particularly CYP2A5 in mice (Salonpaa et al., 1997; Salonpaa et al., 1995). Because induction of P450 accelerates heme consumption, it may be a promoting factor for porphyrinogenic chemical-induced hepatic heme deficiency. In fact, a phenobarbital/ DDEP model has been established as acute heme deficiency and hepatic porphyria (Halpert et al., 1994). Recently, griseofulvin-induced mouse models of cholestatic liver injury and erythropoietic protoporphyria have been proposed, and their gene expression profiling and other analysis have demonstrated upregulation of genes of heme synthesis and catabolism, P450 genes, and biliary salt, phospholipids, and cholesterol secretion rates (Davies et al., 2005; Gant et al., 2003; Plosch et al., 2002). FIG. 11. Postulated mechanisms of generation of ferrochelatase-inhibitory N- alkylprotoporphyrin from griseofulvin. Griseofulvin alkylates the prosthetic group of P450 heme to generate N-griseofulvinPP, which is unable to inhibit ferrochelatase. N-GriseofulvinPP is converted into N-methylPP, which is a potent inhibitor of ferrochelatase.

Other Porphyrinogenic Agents

3-[(Arylthio)ethyl]sydnone (TTMS) was developed as a potential antiarthritic compound, but research into its use was abandoned after it was found to cause hepatic protoporphyrin IX accumulation in dogs and rodents. TTMS causes a depletion of hepatic P450 with concurrent N-vinylPP formation and an inhibition of ferrochelatase activity, indicating that TTMS acts in a similar manner to 4- alkyldihydropyridines in eliciting its porphyrinogenic effects (Ortiz de Montellano and Grab, 1986; Sutherland et al., 1986). CYP3A2 is responsible for NvinylPP formation in male rat liver after TTMS administration (McNamee and Marks, 1996; Wong et al., 1998). In humans, involvement of CYP3A4 in TTMS-induced N-vinylPP formation was also shown by in vitro studies with human liver microsomes and cDNA-expressed human P450s (Gamble et al., 2003; Lavigne et al., 2002).

Allylisopropylacetamide (AIA) is an analogue of the hypnotics aprobarbital and allylisopropylacetylurea, which have a common terminal olefin (Figure 12). AIA is the prototype of a group of compounds that lower hepatic free heme levels by destruction of the heme moiety of P450. A green pigment was obtained from the rats given AIA, which was from the modified heme moiety of P450 (Oritz de Montellano and Correia, 1983). Although AIA bound covalently to protoporphyrin IX as a 1:1 porphyrin-AIA adduct (Ortiz de Montellano and Mico, 1980), which was further characterized as an N-AIA protoporphyrin adduct (Wong et al., 1999), the adduct did not inhibit ferrochelatase (De Matteis et al., 1980). It has been proposed that porphyria can be induced, even without inhibition of ferrochelatase, through severe and/or continuous heme depletion, where repeated mechanism-based inactivation dissociates the modified heme moiety from intact and heme-reconstitutable apoprotein of P450 (Bornheim et al., 1987; De Matteis and Marks, 1996; Marks et al., 1988). Porphyria via a similar mechanism may occur by other compounds including the already described AIA-analogous drugs and secobarbital (He et al., 1996) (Figure 12), and other drugs with unsaturated carboncarbon bonds such as ethinylestradiol (discussed earlier), all of which cause alkylation of P450 heme and formation of a green pigment (Oritz de Montellano and Correia, 1983).

FIG. 12. Structures of possible porphyrinogenic olefins. Allylisopropylacetamide (AIA) lowers hepatic free heme levels by the destruction of the heme moiety of P450, leading to porphyria via the postulated mechanisms described in the text. The AIA-analogous hypnotics, allylisopropylacetylurea, aprobarbital, secobarbital, and allobarbital, have a common terminal olefin, and also cause alkylation of P450 with subsequent porphyria by similar mechanisms.

FIG. 13. Metabolism of halothane via two major P450-dependent pathways leading to modification of P450. Halothane is oxidatively metabolized to a trifluoroacetyl chloride, which reacts with water to form trifluoroacetic acid and covalently modifies hepatic proteins, leading to immune-mediated liver injury (see later discussion). Halothane is metabolized via a reductive pathway to yield 1,1,1-trifluoro-2-chloro-ethyl free radical, which is converted into chlorotrifluoroethane (CTE) and chlorodifluoroethylene (CDE). The reductive metabolism also leads to the destruction of P450, probably via modification of prosthetic heme.

Possible Involvement of P450 Destruction in Halothane Hepatotoxicity

Halothane anesthesia causes liver injury and destruction of hepatic P450. Halothane is metabolized via two major P450-dependent pathways (Figure 13), both of which are involved in halothane- induced liver injury. At normal oxygen concentrations, halothane is oxidatively metabolized to trifluoroacetyl chloride (Kharasch et al., 1996), which covalently modifies hepatic proteins and leads to immune-mediated liver injury (see later discussion). In contrast, under hypoxic conditions, halothane is metabolized via a reductive pathway to yield 1,1,1 -trifluoro-2-chloro-ethyl free radical, which reacts with cellular proteins and lipids (Awad et al., 1996; de Groot and Noll, 1983). CYP2A6 is involved in reductive radical formation and resultant lipid peroxidation in human liver microsomes (Minoda and Kharasch, 2001; Spracklin et al., 1996). The reductive metabolism of halothane also leads to destruction of P450 in rat and human liver microsomes, probably via modification of prosthetic heme (Baker et al., 1991; Krieter and van Dyke, 1983; Manno et al., 1991; Manno et al., 1992). In contrast to porphyrinogenic drugs, halothane- induced heme modification does not lead to porphyrin accumulation, but to a decrease in protoporphyrin IX as well as microsomal heme, probably via destruction of its tetrapyrrole ring (Manno et al., 1991). On the other hand, exposure of rats to halothane results in a decrease in hepatic P450, followed by a rapid increase in free heme concentration, and resultant decrease in ALAS expression and induction of heme oxygenase-1 (HO-1) (Odaka et al., 2000). Free heme acts as a potent pro-oxidant, which may contribute to oxidative tissue injury (Kumar and Bandyopadhyay, 2005). Because HO-1 catalyzes heme decomposition (Wagener et al., 2003), its induction may be a protective response to the heme toxicity. This is supported by the observation that induction of HO-1 by hemin pretreatment resulted in abrogation of halothane-induced hepatotoxicity (Odaka et al., 2000). Accordingly, it could be that that free heme derived from the destroyed P450 as well as other heme proteins are involved in halothane-induced liver injury (Figure 14).

FIG. 14. Postulated mechanisms of hepatotoxicity mediated by P450- derived heme in the reductive metabolism of halothane. Free heme, probably derived from P450 enzymes destroyed during the reductive halothane metabolism, may act as a potent prooxidant and contribute to oxidative tissue injury. Heme oxygenase-1 (HO-1) catalyzes heme decomposition to form carbon monoxide (CO), ferritin, and bilirubin, which may be a protective response to the heme toxicity.

The effects of other compounds analogous to halothane have been also examined. Isoflurane, a less hepatotoxic volatile anesthetic than halothane, is also less effective on P450, ALAS, and HO-1 (Buzaleh et al., 2000; Yamasaki et al., 2001). Hydrofluorocarbons such as HCFC-123 (2,2-dichloro-1,1,1-trifluoroethane) are replacements for ozone-depleting chlorofluorocarbons. HCFC-123 is reductively metabolized into free radicals and/or carbene species, which inactivate CYP2E1 and CYP2B via modification of their heme groups (Ferrara et al., 1997). Decreased P450 activities in olfactory microsomes were observed after treatment of microsomes with HCFC-123 as well as halothane under anaerobic conditions (Marini et al., 2001). However, the implications for hepatotoxicity and olfactory mucosal toxicity are unknown. A haloalkane, carbon tetrachloride (CCI4), causes liver injury and is widely used as an animal model of hepatotoxicity (Weber et al., 2003). During anaerobic metabolism, CCI4 is reductively metabolized into the trichloromethyl free radical, which binds covalently to the prosthetic heme of P450, leading to its destruction (Manno et al., 1988). As well as halothane, the CCl4-induced destruction of P450 accompanies an increase in free heme, which may contribute to free radical generation and induction of HO-1 (Nakahira et al., 2003). Furthermore, inhibition of HO-1 activity by tin-mesoporphyrin results in aggravation of the hepatotoxicity of CCl4, suggesting that HO-1 acts as a protective component against the hepatotoxicity mediated by free heme, which may be derived from the destroyed P450 (Manno et al., 1988; Nakahira et al., 2003).

Role of P450 Heme in Cisplatin-Induced Renal Toxicity

Cisplatin is an effective anticancer drug. The most common adverse effect limiting its use is the nephrotoxicity that develops primarily in the proximal tubules (Safirstein et al., 1986; Screnci and McKeage, 1999), and it has been used as an experimental model of acute renal failure. The underlying mechanism of this toxicity is not well understood, but reactive oxygen species (ROS) may be involved in the pathogenesis (Baliga et al., 1998a). Treatment of rats with cisplatin causes a decrease in renal P450 content, accompanied by an increase in bleomycin-detectable iron content (Baliga et al., 1998b). The mechanism for the P450 destruction has not been elucidated. It was suggested that iron, probably derived from P450, plays an important role in mediating renal injury through generation of hydroxy radicals (Baliga et al., 1998a, 1998b). Induction of renal HO-I is accompanied by nephrotoxicity (Agarwal et al., 1995). Induction of HO-I by hemin, prior to cisplatin exposure, reduces cisplatin-induced ROS generation and toxicity and, in contrast, inhibition of HO-1 by tin-protoporphyrin (SnPP) results in aggravation of the effects of cisplatin (Agarwal et al., 1995; Schaaf et al., 2002; Shiraishi et al., 2000), suggesting that heme is involved in cisplatininduced toxicity. The protective role of HO- 1 is also supported by the fact that toxicity is exacerbated in HO- I null mice (Agarwal and Nick, 2000). Further studies of the chemical inhibition of CYP2E1 and CYP2E1-null mice suggest a role of CYP2E1 in cisplatin-induced nephrotoxicity as a site for the generation of ROS and/or a source of catalytic iron (Liu et al., 2002; Liu and Baliga, 2003). Thus, it may be that catalytic iron derived from destroyed CYP2E1 participates in cisplatin-induced renal injury. ROLE OF COVALENT BINDING OF DRUG REACTIVE METABOLITE TO CYTOCHROME P450 IN IMMUNE-MEDIATED HEPATITIS

Immunological mechanisms are involved in many adverse drug reactions. In certain forms of drug-induced hepatitis, serum from patients has been found to express specific autoantibodies against P450 enzymes. The autoantibody specificity is commonly restricted to the particular P450s responsible for bioactivation of the drug. In some cases, mechanism-based inactivation of the corresponding P450s by these drugs and/or covalent binding of their reactive metabolites to apoprotein of the P450 have been demonstrated through in vitro experiments. Thus, it has been proposed that the appearance of autoantibodies against the specific P450s in the serum of patients with drug-induced hepatitis is attributed to drug-modified P450 that behaves as a neoantigen and activates immune system, resulting in formation of autoantibodies directed against the specific P450s (Beaune et al., 1994; Beaune and Lecoeur, 1997; Manns and ObermayerStraub, 1997; Obermayer-Straub and Manns, 1996; Robin et al., 1997). However, the role of the autoantibodies in the pathogenesis of drug-induced hepatitis remains largely unknown.

Hapten Hypothesis and Danger Hypothesis

Immune-mediated mechanisms for drug toxicity, particularly hepatotoxicity, which is mediated by drug-protein adducts, are based on the "hapten hypothesis" (Griem et al., 1998; Park et al., 1998; Pohl et al., 1988). The drug undergoes bioactivation and covalent binding to protein. The haptenized protein, including modified P450, is internalized by phagocytes, such as Kupffer cells and dendritic cells, and is presented by major histocompatibility complex class II (MHCII) on the antigen-presenting cells (APC) to helper T cells, leading to their activation. The cytotoxic T cells are targeted against hepatocytes that express haptenized protein, and B cells producing autoantibodies or antibodies against haptenized protein mediate antibody-dependent toxicity. On the other hand, the recently proposed "danger hypothesis," which proposes that the immune system only responds to so-called danger signals (Matzinger, 1994), can be adapted to drug toxicity (Gruchalla, 2001; Kaplowitz, 2005; Park et al., 1998; Pirmohamed et al., 2002; Seguin and Uetrecht, 2003; Uetrecht, 1999). The danger hypothesis is additive to the hapten hypothesis and explains the idiosyncratic nature of drug toxicity. Haptenization alone might be insufficient to trigger an immune reaction and results in tolerance. Development of a full immune response to hapten requires a second co-stimulatory trigger, the danger signal. Danger signals include cytokines derived from damaged hepatocytes and host-dependent factors, such as a viral or bacterial infection and other inflammatory diseases (Figure 15).

Tienilic Acid-Induced Hepatitis and Autoantibody to CYP2C9

Tienilic acid, a diuretic drug, was used in the treatment of hypertension, but was withdrawn from the market because of rare, but severe, cases of hepatotoxicity. In 60% of patients suffering from severe hepatitis after administration of tienilic acid, a specific antibody directed against unmodified liver and kidney microsomal proteins was detected, which was termed antiliver/kidney microsome type-2 (LKM-2) (Homberg et al., 1984). This LKM-2 autoantibody specifically recognized CYP2C9 in human liver (Beaune et al., 1987). A conformational epitope was then identified on CYP2C9, which was recognized by this antibody (Lecoeur et al., 1996). The study with rats indicated that LKM-2 autoantibodies recognized both native epitopes on the P450 and epitopes that were modified by tienilic acid (Pons et al., 1991; Robin et al., 1996). As shown in earlier discussion in this article, tienilic acid is oxidized by CYP2C9 to form electrophilic thiophene sulfoxide and/or a thiophene epoxide that reacts with CYP2C9 protein at the active site (Jean et al., 1996; Koenigs et al., 1999; Lopez-Garcia et al., 1994; Lopez Garcia et al., 1993). Therefore, large quantities of one particular modified protein, CYP2C9 in this case, are obtained, which may become a neoantigen and trigger an immune response (Lecoeur et al., 1994). However, it is not known whether LKM-2 autoantibody is simply a by-product of the underlying pathogenetic process or whether it is directly involved in the pathogenesis of the disease. One possibility for a pathogenetic role of the LKM-2 autoantibody would be liver cell injury initiated by the antibody binding to the cell surface. The rat model demonstrated that CYP2C11 modified with tienilic acid were visible on the outer surface of hepatocytes, indicating that the LKM-2 autoantibody may play a role in the disease process via antibody-dependent cytotoxicity and complement (Robin et al., 1996).

FIG. 15. Role of P450 in hapten hypothesis and danger hypothesis. The drug is converted into a reactive metabolite and binds covalently to P450. The haptenized P450 is possessed by the antigen- presenting cells (APC) and presented by major histocompatibility complex class II (MHCII) on APC to helper T cells, leading to their activation. The cytotoxic T cells are targeted against hepatocytes that express haptenized protein, and B cells producing autoantibodies or antibodies against hapten

Source: Critical Reviews in Toxicology

More News in this Category