Abstract
Ziprasidone (Geodone), a novel atypical antipsychotic agent, is recently approved for the treatment of schizophrenia. It undergoes extensive metabolism in preclinical species and humans after oral administration, and only a very small amount of administered dose is excreted as unchanged drug. In vitro studies using human liver microsomes have shown that the oxidative metabolism of ziprasidone is mediated primarily by CYP3A4. However, coadministration of ziprasidone with ketoconazole, a CYP3A4 inhibitor, showed only a modest increase in its exposure. Therefore, in vitro metabolism of ziprasidone was investigated in hepatic cytosolic fractions to further understand its clearance mechanisms in preclinical species and humans. The major metabolite from incubation of ziprasidone in cytosolic fractions of rat, dog, and human was characterized by liquid chromatography-tandem mass spectrometry and found to be the product of reductive cleavage. Derivatization and hydrogen/deuterium exchange were used to deduce that the addition of two hydrogen atoms had occurred at the benzisothiazole moiety. Further studies to determine the enzyme involved in the formation of this metabolite are currently in progress. The identification of this novel metabolite in cytosol has clarified the clearance mechanism of ziprasidone in humans and preclinical species.
Ziprasidone (ZIP), a substituted benzisothiazolyl-piperazine (Fig. 1), is a recently approved novel antipsychotic agent effective in the treatment of schizophrenia. It exhibits potent and highly selective dopamine D2 and serotonin 5-HT2 receptor antagonistic activities. It also has a high affinity for the 5-HT1A, 5-HT1D, and 5-HT2C receptor subtypes that could contribute to the overall therapeutic effect (Seeger et al., 1995). The metabolism of ZIP has been studied both in vitro and in vivo in preclinical species and humans (Prakash et al., 1997a,c, 2000). In vitro studies using human liver microsomes suggested the formation of four oxidative metabolites, ZIP-sulfoxide, ZIP-sulfone, BITP, and Ox-AA (Fig. 1). Further in vitro studies using P450 isoform-selective chemical inhibitors, correlation studies, and metabolism by recombinant human P450 isoforms suggested that the formation of these oxidative metabolites is mediated mainly by human liver CYP3A4 (Prakash et al., 2000). However, phase II clinical studies showed only modest increases in both the area under the curve and Cmax of ZIP (<40% increase with single and multiple dose evaluations) with concomitant administration of ketoconazole, a potent inhibitor of CYP3A4 (Miceli et al., 2000b). It is known that substrates that are primarily dependent on CYP3A4 for metabolism demonstrate intersubject variability in exposure on the order of 50-fold and have significant increases in exposure with concomitant administration of ketoconazole (Lin and Lu, 1998; Thummel and Wilkinson, 1998). On the other hand, coadministration with carbamazepine (200 mg b.i.d.), an inducer of CYP3A4, resulted in only a small decrease (<40%) in exposure of ZIP (Miceli et al., 2000a). In addition, concomitant administration of cimetidine (400 mg/day), a general inhibitor of P450 isoforms, showed only modest increase in both the area under the curve and Cmax of ZIP (<40% increase with single and multiple dose; Wilner et al., 2000). Therefore, it could be hypothesized that the major clearance of ZIP in preclinical species and humans is due to metabolism by non-P450 enzymes.
In vivo metabolism studies suggested that ZIP is extensively metabolized both in animals and humans, and only a small amount (<1%) of drug is excreted unchanged in urine (Prakash et al., 1997a,c). In addition to metabolites found in vitro, a major metabolite of ZIP, S-methyl-dihydro-ZIP, was identified in rats, dogs, and humans. This metabolite accounted for >60% of the administered dose in feces of humans following a single 20-mg oral dose of radiolabeled ZIP (Prakash et al., 1997c). Therefore, it could be envisioned that enzymes other than P450 isoforms form the S-methyl-dihydro-ZIP metabolite. The formation of the S-methyl-dihydro-ZIP metabolite was speculated to involve an initial reductive cleavage of the benzisothiazole moiety to give an intermediate that can be further converted to a methyl thioether by methylation with S-thiomethyl transferases (Prakash et al., 1997b). However, the intermediate metabolite was not detected either in vivo or in vitro, possibly due to its instability or its rapid metabolism to S-methyl metabolite. The present work reports the in vitro metabolism of ZIP in hepatic cytosolic fractions from preclinical species and humans to further understand the clearance mechanisms of ZIP in these species. The major metabolite of ZIP in these incubations was characterized by LC-MS/MS, hydrogen/deuterium (H/D) exchange, and chemical derivatization with N-dansylaziridine and found to be the product of reductive cleavage of the benzisothiazole ring. Taking all these data together, the new metabolite was identified as dihydroziprasidone (6-chloro-5-(2-{4-[imino-(2-mercapto-phenyl)-methyl]-piperazin-1-yl}-ethyl)-1,3-dihydro-indol-2-one). The identification of this novel metabolite in cytosol has provided the understanding of the clearance mechanism of ziprasidone in humans and preclinical species.
Materials and Methods
Chemicals and Reagents. Commercially obtained chemicals and solvents were of HPLC or analytical grade. [14C]ZIP (Fig. 1), specific activity 8.2 mCi/mmol, and dihydro-ZIP were synthesized at Pfizer Global Research and Development (Groton, CT). Dansylaziridine, menadione, allopurinol, CD3 COOD, and D2O were purchased from Sigma-Aldrich (St. Louis, MO).
Incubations. The cytosolic fractions of rat, dog, and human liver were prepared using standard centrifugal methods (Prakash et al., 2000). The incubations of ZIP with cytosolic fractions of rat, dog, and human liver were conducted using standard procedures. A typical incubation mixture (1 ml) contained [14C]ZIP (10 μM in 5 μl of methanol), 10 mg/ml cytosolic protein, 2-hydroxypyrimidine (40 μM), and 100 mM potassium phosphate buffer, pH 7.4. The reaction was initiated by the addition of the cytosolic fraction, and the incubation was carried out for 20 min. The reaction was stopped by the addition of methanol (5 ml). Control incubations using boiled cytosols were also carried out under the same conditions. The reaction mixtures were centrifuged (1900g) and the supernatants were transferred to clean tubes to be analyzed by LC-MS/MS. Incubations were also conducted with allopurinol (a mechanism-based inhibitor of xanthine oxidase) and menadione (a specific inhibitor of AO) at a concentration of 100 μM each as described above.
HPLC-MS. The mixture of metabolites was subjected to chromatography on an HPLC system that consisted of an HP-1050 solvent delivery system, an HP-1050 auto-injector, a radioactivity monitor (β-RAM; IN/US Systems, Inc., Tampa, FL) and an SP 4200 computing integrator (Spectra Physics, San Jose, CA). Chromatography was conducted on a YMC basic HPLC column (4.6 × 150 mm, 3 μm) with a binary mixture of 10 mM ammonium acetate (pH 5.0; solvent A) and methanol (solvent B), and the flow rate was established at 1 ml/min. Analysis of metabolites was performed on a PerkinElmerSciex API III HPLC-MS/MS system (PerkinElmerSciex Instruments, Boston, MA) using ion spray. The effluent from the HPLC column was split, and about 50 μl/min was introduced into the atmospheric ionization source via an ion spray interface. The remaining effluent was directed into the flow cell of a β-RAM. The β-RAM response was recorded in real time by the mass spectrometer that provided simultaneous detection of radioactivity and mass spectrometry data. The ion spray interface was operated at 5000 V, and the mass spectrometer was operated in the positive ion mode. CID studies were performed using argon gas at collision energy of 25 to 30 eV and a collision gas thickness of 2 × 1015 molecules/cm2.
Derivatization with Dansylaziridine. The supernatant from the incubation was dried under nitrogen in a TurboVap LV evaporator (Caliper Life Sciences, Hopkinton, MA). The dried incubation mixture was dissolved in 2 M KOH (100 μl), 500 μl of 3 mM N-dansylaziridine in methanol was added, and the reaction mixture was stirred at room temperature for 1 h (Orford et al., 1989). Following evaporation of the solvents, the residue was dissolved in mobile phase and an aliquot was injected into the HPLC-MS/MS system without further purification
Results and Discussion
Metabolite Profiles of ZIP in Cytosolic Fractions. Representative HPLC-radiochromatograms of metabolites from incubations of ZIP with cytosolic fractions of rat, dog, and human liver are shown in Fig. 2. A major metabolite designated M11, along with a few additional minor metabolites (M4, M4A, M7, and M8) were detected in the chromatograms. These metabolites were not detected with boiled liver cytosols, suggesting the formation of these metabolites by enzymatic reactions. We have earlier reported the structural characterization of metabolites M4, M4A, M7, and M8 in humans, but M11 was not detected in vivo in humans (Prakash et al., 1997c). Full-scan mass spectrum of metabolite M11 revealed a protonated molecular ion [M + H]+ at m/z 415, 2 Da higher than the drug. The product ion mass spectrum of m/z 415 produced fragment ions at m/z 280, 263, 237, 194, and 159 (Fig. 3a). The ion at m/z 280 corresponds to a charge-initiated fragmentation of the piprazinyl nitrogen-benzisothiazole carbon bond with the expulsion of the benzisothiazole as a neutral molecule (Prakash et al., 1997a). The ions at m/z 263 resulted by the loss of ammonia from fragment ion at m/z 280. The presence of other characteristic fragment ions at m/z 194 and 263 in its CID spectrum further suggested that the addition of 2 Da had occurred remote from the oxindole part of the molecule.
H/D exchange followed by mass spectrometry has long been recognized as a valuable means to study the mechanism of ion formation and metabolic pathways of xenobiotics, and to differentiate the isomeric structures of metabolites (Kamel et al., 2002; Kamel and Munson, 2004). H/D exchange techniques are useful for determination of the presence, number, and position of H/D-exchangeable functional groups on the metabolite structures and serve as an aid for structure elucidation of metabolites (Nassar, 2003). Solution phase H/D exchange of M11 using D2O showed a shift of the protonated molecular ion from m/z 415 to m/z 419 for the full exchanged species [M(d3) + D]+. On the other hand, the full-scan MS of ZIP after D2O treatment showed the deuterated molecular ion (M(d) + D)+ at m/z 415 (2 mass units higher than the corresponding [M + H]+). The increase of 2 mass units for the deuterated molecular ion of M11 compared to the deuterated molecular ion of the parent is in agreement with the presence of two exchangeable hydrogen atoms. The product ion MS spectrum of m/z 419 showed fragment ions at m/z 283, 265, 195, and 160 (Fig. 3b). These data indicated that the addition of two hydrogen atoms had occurred at the benzisothiazole moiety by its reductive cleavage. Based on these data, the structure of the major metabolite M11 was proposed as dihydroziprasidone.
The proposed structure of M11 was further corroborated by its derivatization with dansylaziridine, a specific derivatizing reagent for the thiol group (Orford et al., 1989). Treatment of M11 with dansylaziridine formed a new product at a HPLC retention time of 42.2 min. The full-scan MS of derivatized product revealed a protonated molecular ion at m/z 691, 176 Da higher than the (M + H)+ of M11, suggesting the addition of a dansylaziridine moiety. The product ion MS spectrum of m/z 691 showed fragment ions at m/z 455, 412, 309, 280, 263, 194, and 170, consistent with the proposed structure (Fig. 4a). The assignment of these ions was confirmed by parallel CID spectrum of m/z 693 (MH+, 37Cl), which gave the fragment ions at m/z 455, 412, 309, 282, 265, 196, and 194 (not shown). The structure of M11 was confirmed unambiguously by comparison of its retention time and MS spectral data with a synthetic standard.
This is the first report on the identification of such an intermediate metabolite of ZIP. There are several marketed drugs, including the anticonvulsant zonisamide and the antipsychotic risperidone, and iloperidone, which contain the 1,2-isooxazole ring structure. It has been reported that these drugs undergo reductive metabolism resulting in the cleavage of N-O bond to form the intermediate imines (Stiff et al., 1992; Mannens et al., 1993; Mutlib et al., 1995; Dalvie et al., 2002). Although the intermediate imines themselves have never been isolated, due to their rapid hydrolysis in aqueous media to ketones, Sugihara et al. (1996) provided the evidence for this mechanism when they demonstrated stoichiometric production of ammonia and 2-sulfamoyalacetylphenol from zonisamide. We report here that the 1,2-benzisothiazole ring structure of ZIP also undergoes similar reductive metabolism, resulting in the cleavage of N-S bond to form an intermediate amidine. The structure of this intermediate amidine was characterized by LC-MS/MS, H/D exchange, and chemical derivatization with a thiol-specific derivatization reagent (Orford et al., 1989). The intermediate amidine was not detected in vivo in preclinical species due either to its instability or to its rapid metabolism to a methyl thioether by methylation with S-thiomethyl transferases (Prakash et al., 1997b). Hillenweck et al. (1997) have shown the formation of methylthio metabolites of chlorothanil, a fungicide used in agriculture, in digestive contents of rat, dog, and human.
It is well established that the reductive metabolism of zonisamide is catalyzed by several enzymes, liver microsomal cytochrome P450, especially CYP3A4 (Nakasa et al., 1993), and AO (Sugihara et al., 1996), as well as mammalian intestinal bacteria (Kitamura and Tatsumi, 1984; Kitamura et al., 1997). However, the reduction of the isoxazole ring in risperidone was attributed mainly to gut microflora (Mannens et al., 1993). Recently, Tschirret-Guth and Wood (2003) have reported that 3-(indole-1-yl)-1,2-benzisoxazoles are reductively N-dearylated by rat liver microsomes under anaerobic conditions. One of the proposed mechanisms for the reductive N-dearylation involved the cleavage of the N-O bond followed by hydrolysis.
It is speculated that the cleavage of the benzisothiazole ring of ZIP could also be mediated by any of the above enzyme systems. M11 or its S-methyl metabolite were not detected in the incubation of ZIP with human fecal samples under anaerobic conditions, suggesting that the gut microflora may not be playing a role in the formation of these metabolites (data not shown). Furthermore, the ring cleaved S-methyl metabolites of ZIP were earlier found in both urine and serum of humans and preclinical species, suggesting that the liver enzymes mediate the opening of the benzisothiazole ring. Our previous in vitro studies ruled out the possibility of involvement of microsomal enzyme in the formation of M11 (Prakash et al., 2000). However, the present study has demonstrated that the cytosolic fractions from rat, dog, and human can catalyze the reduction of the benzisothiazole ring. Our preliminary studies in human liver cytosolic fractions showed that the formation of M11 was inhibited slightly by a xanthine oxidase inhibitor, allopurinol (22%). However, 100 μM menadione (a specific AO inhibitor) largely abolished the formation of dihydroziprasidone (>70% inhibition), suggesting that the formation of dihydroziprasidone metabolite may be mainly mediated by AO (A. Kamel, unpublished work). These results indicated that the opening of the benzisothiazole ring might be mediated by AO. As reported previously, AO, a liver cytosolic metalloflavoprotein, in the presence of its electron donor, catalyzes the reduction of a variety of compounds such as N-oxides, hydroxamic acids, and oximes (Kitamura and Tatsumi, 1984; Tatsumi and Ishigai, 1987). The reduction of oximes formed the intermediate ketimines, which subsequently can undergo nonenzymatic hydrolysis to ketones (Tatsumi and Ishigai, 1987). The benzisothiazole derivatives may also undergo reductive metabolism by AO to the corresponding amidine. The details of the mechanism for the formation of M11 are currently under investigation in this laboratory.
In summary, ZIP is eliminated by two distinct pathways in humans and preclinical species. In humans, cytosolic enzyme(s) mediates approximately two-thirds of the ZIP metabolism and, therefore, reduces the potential of cytochrome 450-based drug-drug interaction with coadministered drugs. Finally, the identification of this novel metabolite of ZIP allows us to understand the clearance mechanism of ZIP and the results from clinical studies.
Acknowledgments
We thank Drs. Keith McCarthy and Stan Walinsky for synthesizing radiolabeled ziprasidone and synthetic standard, Larry Cohen for providing cytosol, and Drs. Hassan Fouda, Scott Obach, Robert Ronfeld, and Larry Tremaine for helpful suggestions and advice.
Footnotes
-
This work was presented in part at the 4th American Society of Mass Spectrometry and Allied Topics, Chicago, Illinois, May 27–31, 2001.
-
Article, publication date, and citation information can be found at http://dmd.aspetjournals.org.
-
doi:10.1124/dmd.105.004036.
-
ABBREVIATIONS: ZIP, ziprasidone (5-[2-{4-(1,2-benzisothiazol-3-yl)piperazin-1-yl}ethyl]-6-chloro-1,3-dihydro-indol-2-one) hydrochloride hydrate); 5-HT, 5-hydroxytryptamine; BITP, 3-(piperazin-1-yl)-1,2-benzisothiazole; Ox-AA, 6-chloro-2-oxo-2,3-dihydro-1H-indol-5-yl)acetic acid; dihydro-ZIP, 6-chloro-5-(2-{4-[imino-(2-mercapto-phenyl)-methyl]-piperazin-1-yl}-ethyl)-1,3-dihydro-indol-2-one; S-methyl-dihydro-ZIP, (6-chloro-5-(2-{4-[imino-(2-methylsulfanyl-phenyl)-methyl]-piperazin-1-yl}-ethyl)-1,3-dihydro-indol-2-one; P450, cytochrome P450; LC-MS/MS, liquid chromatography-tandem mass spectrometry; H/D, hydrogen/deuterium; HPLC, high-performance liquid chromatography; CID, collision-induced dissociation; MS, mass spectrometry; AO, aldehyde oxidase.
- Received February 28, 2005.
- Accepted April 6, 2005.
- The American Society for Pharmacology and Experimental Therapeutics