Introduction

Top Abstract Introduction Results Discussion References

Haloperidol (HAL¹; 4-(4-chlorophenyl)-1-[4-(4-fluorophenyl)-4-oxobutyl]-4-piperidinol) is one of the most widely used antipsychotic drugs (Wysowski and Baum, 1989; Shader, 1994; Lohse et al., 1996). HAL is extensively metabolized with only about 1% of the administered dose excreted unchanged in urine (Forsman and Larsson, 1978). The metabolism of HAL can be summarized as shown in Fig. 1 and consists of the following pathways: 1) N-dealkylation, which leads to the formation of 4-(4-chlorophenyl)-4-hydroxypiperidine (CPHP) (Soudijn et al., 1967; Fang and Gorrod, 1993; Gorrod and Fang, 1993; Fang et al., 1996a); 2) reduction of the ketone group, which leads to the formation of reduced HAL (RH) (Forsman and Larsson, 1978); 3) the reverse oxidation of RH back to HAL (Forsman and Larsson, 1978; Midha et al., 1988; Someya et al., 1992); 4) N-dealkylation of RH leading to the formation of CPHP (Fang et al., 1996a); 5) a pyridinium analog of HAL (HP⁺) (Subramanyam et al., 1990; Fang and Gorrod, 1991; 1993; Gorrod and Fang, 1993); 6) the dehydration product of HAL leading to the formation of haloperidol 1,2,3,6-tetrahydropyridine (HTP), which was proposed to be an intermediate product for the formation of HP⁺ from HAL (Fang and Gorrod, 1991; 1993; Gorrod and Fang, 1993); and 7) the pyridinium analog of RH (RHP⁺) (Eyles et al., 1994; Van der Schyf et al., 1994). RHP⁺ is reported to be produced either by reduction of HP⁺ or by oxidation from RH (Eyles et al., 1996). Both HP⁺ and RHP⁺ have been shown to be present in plasma and postmortem brain of patients administered HAL (Eyles et al., 1994; 1997); 8) oxidation of RHP⁺ back to HP⁺; 9) a few unknown metabolites were also identified (Fang and Gorrod, 1991; 1993; Gorrod and Fang, 1993). The major phase II metabolite of HAL is reported to be the glucuronidation product of HAL.

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Enzymes responsible for the different metabolic pathways of haloperidol. Dotted arrow indicates a minor pathway.

Dose individualization is required for the use of HAL because of a narrow therapeutic window (Ulrich et al., 1998) and a large interindividual variability with the plasma levels of HAL. The interindividual variability may be related to enzyme polymorphism (Llerena et al., 1992a,b; Lane et al., 1997), interethnic differences (Potkin et al., 1984; Chang et al., 1987; Jann et al., 1989; Someya et al., 1990), and smoking status (Perry et al., 1993; Shimoda et al., 1999). HAL is reported to interact with a number of other drugs through inhibition or induction of cytochrome P450 (P450) enzymes (Kudo and Ishizaki, 1999). A good understanding of the enzymes responsible for the metabolism of HAL would assist in rationalizing the interindividual variations and drug-drug interactions involving HAL.

There has not been a comprehensive study on the enzyme(s) responsible for the metabolism of HAL covering most of the metabolic pathways and known isoenzymes. In this study, a thorough in vitro study was carried out to study the P450 isoenzymes responsible for the metabolism of HAL and the secondary metabolism of its metabolites.

Experimental Procedures

Materials. Haloperidol was purchased from Sigma (St. Louis, MO). CPHP and CPTP were purchased from Aldrich Chemical Company (Milwaukee, WI). HTP, HP⁺, and RH were synthesized as previously described (Gorrod and Fang, 1993). RHP⁺ was a kind gift from Dr. Neal Castagnoli (Virginia Tech, Blacksburg, VA). All other chemicals were of analytical grades.

Methods. HPLC analysis. CPHP, CPTP, RH, HAL, HTP, HP⁺, and RHP⁺ were analyzed using a slight modification of an HPLC method described previously (Fang and Gorrod, 1993). Briefly, the system comprised a Waters model 510 solvent delivery system, a Waters WISP 710B Autoinjector, and a Waters 2487 dual  absorbance detector (Waters, Milford, MA) set at 220 nm and 245 nm. Signals from the UV detector (220 nm for CPHP and RH; 245 nm for CPTP, HAL, HTP, HP⁺, and RHP⁺) were collected and processed at the same time by a Waters Millennium chromatography manager system. A Hypersil CN, 5-µm column (4.6 × 250 mm) (Phenomenex, Torrance, CA) coupled with a SecurityGuard guard cartridge system (Phenomenex) was used. The mobile phase consisted of acetonitrile/ammonium acetate buffer (1 M)/water (67:1:32 by volume). The mixture was adjusted to pH 5.4 with acetic acid, and the solvent was delivered at a flow rate of 1 ml/min.

Enzymatic studies. Incubation procedures were as follows; reaction mixtures (100 µl) consisted of microsomal preparation (10 µl), a cofactor-generating system consisting of -nicotinamide adenine dinucleotide phosphate (1.3 mM), glucose 6-phosphate (3.3 mM), glucose-6-phosphate dehydrogenase (0.4 U/ml) and MgCl₂ (3.3 mM), and appropriate concentrations of substrates in phosphate buffer (0.1 M, pH 7.4). Control incubates contained heat-inactivated microsomes or control microsomes transfected with a control vector. The reaction mixtures were incubated at 37°C, and biological reactions were terminated by the addition of acetonitrile (50 µl). The denatured proteins were removed by centrifugation, and the clear supernatant was subjected to HPLC analysis.

Correlation studies. For correlation analysis, HAL, RH, HTP, or RHP⁺ (20 µM) was incubated with a bank of human liver microsomal preparations from 12 donors (10 µl/incubate) for 20 min. Sample preparation and HPLC analysis of the metabolites were carried out as described under Enzymatic Studies. The quantities of HAL metabolites increased linearly with increasing incubation time (data not shown). The rates of formation of the HAL metabolites were correlated to the catalytic activities of CYP1A2, CYP2C8, CYP2C9, CYP2C19, CYP2D, CYP2E, and CYP3A in the microsomes (data supplied by GENTEST). The rates of formation of the HAL metabolites were also correlated to the immunoactivities of CYP1A2, CYP2C9, CYP2D6, CYP2E1, CYP3A4, and CYP3A5 in the microsomes (data supplied by GENTEST). Correlations of metabolite formation velocities with enzyme activities were evaluated by simple linear regression (GraphPad Prism software; GraphPad Software, San Diego, CA). Whether the slope of the regression line is significantly different from zero is determined by an F-test.

cDNA-Expressed Enzymes. HAL, RH, and RHP⁺ (100 µM for all substrates) were incubated with recombinant CYP1A1, CYP1A2, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP2E1, CYP3A4, and CYP3A5. Ten microliters of the microsomal preparations were used for each incubation, and the incubation was carried out for 60 min. Sample preparation and HPLC analysis of the metabolites were carried out as described under Enzymatic Studies.

	Results

Top Abstract Introduction Results Discussion References

HAL, RH, and RHP⁺ were investigated as substrates of recombinant CYP1A1, CYP1A2, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP2E1, CYP3A4, and CYP3A5. Among these isoenzymes, CYP3A4, CYP3A5, CYP1A1, CYP1A2, CYP2C8, CYP2C9, CYP2C19, and likely CYP2D6 showed some catalytic activity toward these substrates (Table 1). HAL was shown to be metabolized to CPHP by CYP3A4, CYP3A5, CYP1A1, CYP2C19, CYP2C8, and CYP2C9. CYP3A4, CYP1A1, and CYP3A5 were also responsible for the formation of HP⁺ from HAL. CYP3A4 catalyzed the dehydration of HAL to HTP. RH was shown to be oxidized to HAL by CYP3A4, CYP1A1, and CYP3A5. RH was metabolized to CPHP by CYP3A4. RH was converted to RHP⁺ by CYP3A4 and CYP3A5. The oxidation of RHP⁺ to HP⁺ was also investigated and was shown to be catalyzed by CYP1A1, CYP1A2, and CYP3A4.

In a second series of experiments, the above-mentioned metabolic pathways were investigated using a panel of 12 human liver microsomal preparations (Table 2). The metabolic activities toward different pathways were correlated with the isoenzyme-specific activities of the microsomes (using selective substrates of the isoenzymes; Table 3) and the immuno-quantified amount of each isoenzyme in the microsomal preparations (Table 4).

	Discussion

Top Abstract Introduction Results Discussion References

The current study demonstrated that CYP3A4 is the most important P450 isoenzyme responsible for most of the metabolic pathways of HAL. These pathways include the dealkylation of HAL and RH, reverse oxidation of RH to HAL, aromatization of the pyridine ring of both HAL and RH, and metabolism of HAL to HTP. Recombinant CYP3A4 is shown to be capable of catalyzing the above-mentioned metabolic pathways investigated and had higher activities than other isoenzymes. The "panel study" showed that metabolic activities of all the pathways investigated correlate with the selective catalytic (Table 3) and immuno (Table 4) activities of CYP3A4. Recombinant CYP1A2, CYP2C8, CYP2C9, CYP2C19, CYP3A5, and CYP2D6 were shown to catalyze one or more metabolic pathways; with human liver microsomes, however, these metabolic activities did not correlate with the isoenzyme-selective substrate activities and isoenzyme immuno-reactivities. Since CYP3A4 is the most abundant P450 isoenzyme expressed in human liver, CYP3A4 should play an important role in the in vivo metabolism of HAL as well.

The oxidation of RHP⁺ to HP⁺ can be catalyzed by recombinant CYP1A1, CYP1A2, and CYP3A4. It is interesting to note that recombinant CYP1A1 showed the highest activity for this reaction, which was about 50 times more active than CYP1A2 and 220 times more active than CYP3A4. In the panel study, the rate of oxidation of RHP⁺ to HP⁺ was shown to be correlated with CYP3A4 levels (Table 3 and Table 4). Correlation with CYP1A1 activities was not possible because there is currently a lack of a selective substrate for CYP1A1. In addition, CYP1A1 is not consistently detected in human liver, and it was proposed that CYP1A1 is not normally present in human liver unless induced. It is therefore not surprising that CYP3A4 was found to be responsible for the oxidation of RHP⁺ in liver microsomal preparations. RHP⁺ is among the best selective substrates for CYP1A1 discovered so far. No good probe substrate is currently available for determination of CYP1A1 activity. Ethoxyresorufin O-deethylation was found to be catalyzed by CYP1A1 and CYP1A2; the intrinsic clearance of recombinant CYP1A1 is 10 times higher than that of CYP1A2 (Penman et al., 1994). Verlukast epoxidation was another metabolic pathway reported to be selectively catalyzed by CYP1A1 in mice, rats, rhesus monkeys, and humans (Grossman et al., 1993). It was shown that lymphoblasts-expressed human CYP1A1 had high catalytic activity, whereas CYP1A2 exhibited activity similar to that of the uninfected lymphoblasts, which constitutively express low levels of P450s. However, the role of other P450 isoenzymes in the metabolism of verlukast is not known. A more thorough study is needed for verlukast to be accepted as a selective substrate of CYP1A1. Since CYP1A1 is not normally present in human liver, studies with other tissues, such as placenta or lung, is needed to establish whether RHP⁺ can be used as a selective probe substrate of CYP1A1. These studies are currently being conducted in this laboratory.

The findings from the current study correlate well with the evidence reported for individual metabolic pathways. The N-dealkylation of HAL was shown to be catalyzed by CYP3A4 by a number of experimental techniques, such as the use recombinant P450 isoenzymes (Fang et al., 1997; Pan et al., 1997; Kudo and Odomi, 1998), a correlation study with human liver microsomes prepared from nine donors, and inhibition by inhibitors and antibodies (Pan et al., 1998). Back-oxidation of RH to HAL was reported to be catalyzed by recombinant CYP3A4 (Fang et al., 1997; Kudo and Odomi, 1998). The current study further demonstrated that the activities for the oxidation of RH to HAL in the panel of human liver microsomes correlate with the catalytic (Table 3) and immuno (Table 4) activities of CYP3A4. This provided further evidence that CYP3A4 is primarily responsible for the oxidation of RH to HAL. The formation of HP⁺ from HAL was investigated using recombinant P450 isoenzymes (Fang et al., 1997), and it was demonstrated that recombinant CYP3A4 can catalyze this reaction. Using human liver microsomes, Usuki et al. (1996) demonstrated that CYP3A4 plays a major role in the formation of HP⁺ from HAL by a correlation study in a panel of human liver microsomes and an inhibition study using selective chemical and immunochemical inhibitors.

The formation of RH from HAL was found to be enantio-selective and S()-RH is produced in higher quantities than its R(+)-enantiomer (Eyles and Pond, 1992; Eyles et al., 1998). This ketone reductase belongs to a group of enzymes called carbonyl reductases, which are found ubiquitously in mammalian tissue (Inaba and Kovacs, 1989). RH used in the current study is a racemic mixture, and further study using S()-RH and R(+)-RH is needed to establish whether the oxidation of RH back to HAL is stereoselective.

In a previous report, a small quantity of CPHP was detected in the microsomal incubation mixture of HAL with recombinant CYP2D6 (Fang et al., 1997) because a sensitive electron-capture gas chromatography method was used to measure the levels of CPHP. In the current study, a less sensitive HPLC-UV (220 nm) method was used to detect CPHP, and the level of CPHP was below the sensitivity limit in the incubation mixture of HAL with recombinant CYP2D6. CYP2D6 is generally considered as a high-affinity low-capacity enzyme, whereas CYP3A4 is a low-affinity high-capacity enzyme. Therefore, the contribution of CYP2D6 in vivo may be more significant than suggested by these in vitro studies in which a high substrate concentration was used. In fact, the contribution of CYP2D6 in the metabolism of HAL has been indicated by the observation that plasma concentrations of HAL correlate to the polymorphism of CYP2D6 (Llerena et al., 1992a,b). CYP2D6 is also considered to be a factor causing the interethnic differences in the pharmacokinetics of HAL (Kudo and Ishizaki, 1999). Thus, a more detailed enzyme kinetic study is needed to clarify the role of CYP2D6 in the metabolism of HAL.

The mechanism of the dehydration of HAL to HTP is not clear. HTP was shown to be produced from HAL enzymatically in in vitro studies by several research groups (Gorrod and Fang, 1993; Tomlinson et al., 1993; Igarashi et al., 1995). There must be an intermediate oxidative step involved in this reaction because of the involvement of P450 enzymes. Castagnoli and coworkers (Usuki et al., 1998) proposed the possibility that the first step of the aromatization of the piperidine ring is the -carbon hydroxylation, which leads to the 2,3-dihydropiperidine analog of HAL (HDP⁺). HDP⁺ then chemically oxidized to HP⁺. This mechanism did not include the formation of HTP, which has been consistently detected in the incubation mixture of HAL. It is therefore proposed that HDP⁺ can be autooxidized by disproportionation to HTP and HP⁺ (Fig. 2). HTP so formed can then be oxidized back to HDP⁺ by CYP3A4 and CYP2D6 (Fang et al., 1997). A similar disproportionation reaction has been proposed for the transformation of N-methyl-4-phenyl-2,3-dihydropyridinium to N-methyl-4-phenylpyridinium (MPP⁺) (Chiba et al., 1985).

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Proposed mechanism for the metabolism of haloperidol to HP+ and HTP.

Human liver microsomes had the lowest activities for the RH  RHP⁺ pathway (Table 2). Since relatively high plasma levels of RHP⁺ were detected in patients taking HAL (Eyles et al., 1994), RHP⁺ is unlikely to be produced from the oxidation of RH by liver microsomes. RHP⁺ is more likely produced from the reduction of HP⁺ by carbonyl reductase (Eyles et al., 1996).

Knowledge of the isoenzymes responsible for the metabolism of HAL is useful in rationalizing some of the known drug-drug interactions observed clinically. It has been reported that plasma concentrations of HAL can be elevated by coadministration of inhibitors of P450 enzymes (Kudo and Ishizaki, 1999), such as fluvoxamine and fluoxetine (Goff et al., 1991; Daniel et al., 1994). Inhibition of CYP3A4 seems to be the mechanism of these drug-drug interactions.

The clearance of HAL was shown to be enhanced by carbamazepine (Jann et al., 1985; Kidron et al., 1985), phenytoin and phenobarbital (Linnoila et al., 1980), and rifampicin (Takeda et al., 1986). These interactions were proposed to be due to the induction of P450 enzymes responsible for the metabolism of HAL. The present study seems to indicate that induction of CYP3A4 is the mechanism of these drug interactions.

Smoking is reported to enhance the metabolism of HAL, particularly at low doses (Perry et al., 1993; Shimoda et al., 1999). Smoking is known to induce CYP1A1 and 1A2 (Zevin and Benowitz, 1999). The present study found that recombinant CYP1A1 could catalyze the HAL  CPHP, HAL  HP⁺, RH  HAL, and RHP⁺  HP⁺ pathways. Whether CYP1A1 plays an important role in these reactions is not known because no data are available for this isoenzyme to carry out the correlation analysis. It is interesting that HAL plasma levels were decreased more significantly at low doses. Since CYP3A4 is known to be a low-affinity high-capacity enzyme, it is possible that CYP1A1 plays a more important role in the metabolism of HAL at low concentrations because it has higher affinity for HAL than CYP3A4. CYP1A1 is largely an extrahepatic enzyme that is present in lung, intestine, placenta, etc. (McKinnon et al., 1991; Hakkola et al., 1996). Extrahepatic metabolism may play an important role in the metabolism of HAL among smokers.

HP⁺ is a structural analog of the toxic metabolite of N-methyl-4-phenyl-1,2,3,6-tetrahydropyridine, MPP⁺. MPP⁺ was shown to be ultimately responsible for Parkinsonian symptoms induced by N-methyl-4-phenyl-1,2,3,6-tetrahydropyridine. The pyridinium metabolite of HAL was therefore proposed to be involved in some of the movement disorders induced by HAL. This was supported by the observations that HP⁺ was neurotoxic both in vitro (Fang et al., 1995, 1996b) and in vivo (Rollema et al., 1994).

In summary, the present study systematically investigated the cytochrome P450 isoenzymes responsible for the metabolism of HAL. CYP3A4 was demonstrated to be responsible for catalyzing all the metabolic pathways investigated, those include the dealkylation of HAL and RH, reverse oxidation of RH to HAL, aromatization of the pyridine ring of both HAL and RH, and dehydration of HAL to HTP. The oxidation of RHP⁺ to HP⁺ can be catalyzed by CYP1A1, CYP1A2, and CYP3A4. The recombinant CYP1A1 showed much higher activities than CYP1A2 and CYP3A4 in catalyzing this reaction. Further studies are currently being carried out to evaluate RHP⁺ as a selective substrate of CYP1A1.