Abstract
We investigated cytochrome P450 (P450)-catalyzed metabolism of the important cancer drugs paclitaxel and docetaxel in rat, pig, minipig, and human liver microsomes and cDNA-expressed P450 enzymes. In rat microsomes, paclitaxel was metabolized mainly to C3′-hydroxypaclitaxel (C3′-OHP) and to a lesser extent to C2-hydroxypaclitaxel (C2-OHP), di-hydroxypaclitaxel (di-OHP), and another unknown monohydroxylated paclitaxel. In pig and minipig microsomes, this unknown hydroxypaclitaxel was the main metabolite, whereas C3′-OHP was a minor product. In minipigs, C2-OHP was the next minor product. In human liver microsomes, 6α-hydroxypaclitaxel (6α-OHP) was the main metabolite, followed by C3′-OHP and C2-OHP. Among different cDNA-expressed human P450 enzymes (CYP1A2, 1B1, 2A6, 2C9, 2E1, and 3A4), only CYP3A4 enzyme formed C3′-OHP and C2-OHP. Docetaxel was metabolized in pig, minipig, rat, and human liver microsomes mainly to hydroxydocetaxel (OHDTX), whereas CYP3A-induced rat microsomes produced primarily diastereomeric hydroxyoxazolidinones. Human liver microsomes from 10 different individuals formed OHDTX at different rates correlated with CYP3A4 content. Troleandomycin as a selective inhibitor of CYP3A inhibited the formation of C3′-OHP, C2-OHP, and di-OHP, as well as the unknown OHP produced in rat, minipig, and pig microsomes. In human liver microsomes, troleandomycin inhibited C3′-OHP and C2-OHP formation, and a suitable inhibitor of human CYP2C8, fisetin, strongly inhibited the formation of 6α-OHP, known to be catalyzed by human CYP2C8. In conclusion, the metabolism of docetaxel is the same in all four species, but metabolism of paclitaxel is different, and 6α-OHP remains a uniquely human metabolite. Pigs and minipigs compared with each other formed the same metabolites of paclitaxel.
Paclitaxel (Taxol) and docetaxel (Taxotere) are mitotic poisons stimulating polymerization of microtubuli and inhibiting their depolymerization (Spencer and Faulds, 1994). These drugs cause the formation of deformed tubulin structures, e.g., asterisks and stars (Schiff et al., 1979), and induce apoptosis (Fan, 1999; Aoudjit and Vuori, 2001). They have been successfully used in breast, ovarian, and lung cancer therapy (Rowinsky et al., 1992; Huizing et al., 1995; Klener, 1995).
Paclitaxel is metabolized primarily in liver, and its metabolites are excreted in the bile (Monsarrat et al., 1990, 1993). Four metabolites have been identified in rat liver microsomes and hepatocytes, mainly paclitaxel hydroxylated at the phenyl C3′ at the side chain of C13 (C3′-OHP1) and, to a lesser extent, C2-OHP (Walle et al., 1993; Anderson et al., 1995) (Fig. 1). Three other metabolites were found in dexamethasone-treated rats, two of which were identified as deacetylpaclitaxel and deacetylated C2-OHP (Anderson et al., 1995). Nine metabolites were detected in vivo in rat bile, and C3′-OHP, C2-OHP, and baccatin III were identified (Monsarrat et al., 1990). In rats, the hydroxylated metabolites are produced by CYP3A1/2 (Walle et al., 1993; Anderson et al., 1995). In human liver microsomes, three metabolites were identified in vitro; 6α-OHP (with a OH group on the phenyl at C6 of the taxane ring) was the main metabolite, whereas C3′-OHP was formed to a lesser extent (Cresteil et al., 1994; Kumar et al., 1994; Rahman et al., 1994). A minor di-OHP metabolite, formed from 6α-OHP, was found in human liver slices and microsomes (Harris et al., 1994b). In an in vivo study, these and seven other metabolites were detected in the bile of patients, and 10-deacetylpaclitaxel, 10-deacetylbaccatin III, and baccatin III were also identified (Monsarrat et al., 1998). Six metabolites of paclitaxel were identified in urine and 6α-OHP, C3′-OHP, di-OHP, and 10-deacetylpaclitaxel were also identified in plasma (Monsarrat et al., 1998). 6α-OHP is formed by human liver CYP2C8 (Rahman et al., 1994). CYP2C8 also oxidizes other clinically important drugs such as the antidiabetic troglitazone, the antiarrhythmic amiodarone, the 3-hydroxy-3-methylglutaryl-CoA reductase inhibitor cerivastatin, retinoids, and arachidonic acid (Nakajima et al., 2003). C3′-OHP is produced by CYP3A4 (Harris et al., 1994b; Rahman et al., 1994). The formation of di-OHP results from stepwise catalysis by CYP2C8 and CYP3A4 (Harris et al., 1994b). 6α-OHP was about 30-fold less active than paclitaxel in tumor cell lines in vitro (Harris et al., 1994a), and C3′-OHP, C2-OHP, and baccatin III were also less cytotoxic (Monsarrat et al., 1990). Human CYP1A2, 2E1, 3A4, and 3A5 expressed in Escherichia coli, HepG2, or human lymphoblastoid cells catalyzed only C3′-OHP formation (Harris et al., 1994b).
Docetaxel is metabolized to the same products in humans and rats in vitro and in mice, rabbits, and dogs in vivo. The main metabolite formed in vivo is hydroxylated at C13 of the tert-butyl side chain. It is further oxidized and cyclized via an unstable aldehyde to two stereomeric hydroxyoxazolidinones. Hydroxydocetaxel can also be oxidized, via an unstable carboxylic acid, to oxazolidinedione (Marre et al., 1996; Royer et al., 1996; Monsarrat et al., 1997). CYP3A4/5 dominates human metabolism of docetaxel (Shou et al., 1998), indicating that their different individual expression might result in different metabolism of docetaxel among individuals.
We investigated in vitro metabolism of paclitaxel and docetaxel in pig and minipig microsomes in comparison with humans and rats. Metabolism of taxanes has not yet been investigated in pigs or minipigs. The aim was to determine whether pigs metabolize paclitaxel and docetaxel in a similar way and whether their metabolism resembles that in humans, who had been considered the only species to form 6α-OHP. Minipigs and pigs might be suitable experimental animals to predict metabolic pathways in humans, because the most important human P450 isoform (CYP3A) has ortholog isoform present in minipigs at comparable levels and activities, and there is no need to induce P450 enzyme levels (Anzenbacher et al., 1998). Moreover, CYP3A29 has been identified in the pig according to European Molecular Biology Laboratory/GenBank/DNA Data Bank of Japan databank submissions (Jurima-Romet et al., 2000) as an ortholog of human CYP3A4. Furthermore, metabolism in pigs supposedly closely resembles that of humans, and pig livers have even been considered future candidates for human transplantation.
Materials and Methods
Chemicals. Paclitaxel was obtained from Fluka (Buchs, Switzerland); 6α-OHP was purchased from BD Gentest (Woburn, MA). C3′-OHP, pregnenolon-16α-carbonitrile (PCN), and fisetin were from Sigma-Aldrich Chemie (Steinheim, Germany). Docetaxel, hydroxydocetaxel, two stereomeric hydroxyoxazolidinones, oxazolidinedione, and 7-epidocetaxel were generous gifts of Aventis (Strasbourg, France).
Preparation and Characterization of Microsomes. Male Wistar SPF rats, purchased from VELAZ (Prague, Czech Republic) (weight 280–300 g), were kept on a pellet diet and tap water ad libitum 1 week before experiments. For induction of CYP3A1/2, PCN was given in five intragastric intubations, 25 mg/kg in olive oil, in 12-h intervals during 3 days. Animals were killed 24 h after the last PCN dose, giving high CYP3A1/2 induction as described before (Nedelcheva et al., 1998). They were euthanized by decapitation, and livers were rapidly excised and chilled in ice-cold physiological saline and homogenized in an ice-cold solution of 150 mM KCl/50 mM Tris-HCl, pH 7.4. Human liver samples were obtained from transplantation donors in accordance with Czech law. Livers were collected from 10 male organ donors with unknown drug history who had died from head injury (gun shot, traffic accidents) at the age of 16 to 42 years. During the heart or kidney transplantation procedure, the liver samples were excised and immediately frozen and then maintained in liquid nitrogen. The time period between the beginning of blood exchange and freezing the liver in liquid nitrogen was 9 to 46 min; this period, as well as age, did not correlate with various P450 activities and immunochemical levels (Nedelcheva et al., 1999). These liver samples were homogenized in the same way as the rat livers. Untreated castrated male minipigs (Brno White variety of Goettingen minipig; Research Institute of Veterinary Medicine, Brno, Czech Republic) weighing 22 to 31 kg and 6 months old, were kept in a breeding facility of The Palacky University (Olomouc, Czech Republic). Untreated castrated male pigs and female pigs (Czech Markedly Fleshy breed) were obtained from Slaughter House Cesky Brod (Cesky Brod, Czech Republic) and the Research Institute of Animal Production (Prague-Uhrineves, Czech Republic). All pigs were fed regular commercial pig food. The livers from pigs and minipigs were obtained 10 min after electrocution, followed by exsanguination. Microsomes were prepared by differential centrifugation as described previously (Gut et al., 1993). P450 concentration was estimated according to the method of Omura and Sato (1964). The concentration of microsomal protein was determined according to the method of Lowry et al. (1951), with bovine serum albumin as standard. Specific induction of CYP3A1/2 in rats by pregnenolone 16α-carbonitrile pretreatment (inducing higher levels than dexamethasone) was immunochemically characterized as described previously (Nedelcheva et al., 1998; Gut et al., 2000). Specific P450 contents are shown in Table 1.
Incubations. The incubation mixture included 150 mM KCl/50 mM Tris-HCl buffer (pH 7.4), 1 mg/ml microsomal protein, and 10 μM paclitaxel or docetaxel, added as 10 mM solutions in methanol (up to 0.1% v/v). Even a 1% concentration of methanol did not inhibit metabolism of paclitaxel or docetaxel in our experiments (data not shown). Selection of a 10 μM concentration was related to Km values reported and observed by us, and these levels are close to Cmax levels during infusion. Samples with microsomes from different species were incubated at the same time in several experiments to exclude interexperiment variability. The reaction was initiated by adding an NADPH-generating system (final concentrations 1 mM NADP, 10 mM glucose 6-phosphate, 10 mM MgCl2, 0.5 U/ml glucose-6-phosphate dehydrogenase. The samples were incubated at 37°C for 30 min in a shaking water bath. The incubation was stopped by chilling the samples and immediate extraction with ethyl acetate. Control samples were incubated in the absence of the NADPH-generating system.
Metabolite Formation by cDNA-Expressed Human P450 Enzymes.E. coli membranes with simultaneously expressed CYP3A4 or several other P450s and NPR [NADP(H)-cytochrome P450 reductase] were used for analysis of paclitaxel metabolism. Plasmids based on the bicistronic system were obtained from F. P. Guengerich (preparation described by Parikh at al., 1997). Samples of 1.0 ml total volume contained CYP3A4-NPR (150 pmol of P450/incubation and NPR at similar levels) with 10 μM paclitaxel (added as 1 μl of 10 mM solution in methanol) and were incubated as described above.
Sample Preparation and HPLC Analysis. After the reactions with microsomes and cDNA-expressed CYP3A4, paclitaxel or docetaxel and their metabolites were extracted into distilled ethyl acetate (two times at 3.5 ml). These ethyl acetate extracts were evaporated to dryness under a nitrogen stream. The dry extracts were dissolved in 200 μl of the mobile phase methanol/water, 65:35 (v/v) for paclitaxel (Cresteil et al., 1994). The mobile phase for docetaxel (acetonitrile/methanol/tetrahydrofuran/water, 21:35:2:42) gave better resolution of metabolites than the previously used mixture of acetonitrile/methanol/water (26:32:42) or the previously mentioned mixture of methanol/water, or a mixture of ammonium acetate/acetonitrile (43:57, v/v) (Garg and Ackland, 2000). The HPLC conditions were as follows: HPLC Agilent Series 1100 system with autosampler, 20-μl sample loop; Macherey-Nagel column (4 × 250 mm with Nucleosil 10-5 C18), flow rate 1.2 ml/min, detection wavelength 232 nm (paclitaxel) and 230 nm (docetaxel). The data were analyzed by CSW software (Chromatography Station for Windows, version 1.7; DataApex, Prague, Czech Republic). Only products formed in the presence of paclitaxel or docetaxel and NADPH and not present in the absence of either of them were considered as drug metabolites. 6α-OHP was identified by MS analysis and by comparison with an authentic standard (BD Gentest). C3′-OHP was identified by means of MS and by comparison with authentic standard (Sigma-Aldrich, St. Louis, MO). Both these metabolites were quantified according to calibration curves, which were linear in the range of 0.1 to 100 μM concentrations. Commercial standards are not available for C2-OHP and for the unknown metabolite of paclitaxel; therefore, paclitaxel was used as a standard, assuming an identical molar extinction coefficient, as described in Vaclavikova et al. (2003). Docetaxel metabolites were detected and quantified according to the signal of authentic standard products generously provided by Aventis.
Identification of Other Paclitaxel Metabolites by MS/MS. HPLC elution fractions, representing metabolites of paclitaxel, were collected as defined peaks during HPLC analysis and dried under a nitrogen stream. For analysis by MS, the samples were dissolved in 125 μl of 10 mM ammonium formate/methanol (1:1), transferred into autosampler vials, and measured using an HPLC-electrospray ionization MS method, with a Luna 3-μm C18 column, 1 × 150 mm, mobile phase methanol/10 mM ammonium formate (7:3 v/v). The structures of taxanes and their metabolites are shown in Fig. 1. The characteristics of the respective products of paclitaxel are listed in Table 2.
Enzyme Kinetic Assays. For kinetic assays, 2.5 to 25 μM paclitaxel or docetaxel concentrations were used. The apparent Vmax and Km values were calculated from equations fitting the data in double-reciprocal plots in the program Microsoft Excel (Microsoft, Redmond, WA). The H6 human liver microsomal sample, exhibiting the highest enzyme activity, provided the best precision and enabled analysis of all four detectable human metabolites. Microsomes from control and PCN-pretreated rats were used for the same purpose, because of high activity.
Inhibition Studies. Inhibition studies of paclitaxel metabolism were performed in the presence of 20, 30, 60, or 100 μM troleandomycin (inhibitor of CYP3A) and 30 μM fisetin (inhibitor of CYP2C8) (Vaclavikova et al., 2003) in the incubation mixtures. Troleandomycin was preincubated for 5 min at 37°C (Nedelcheva et al., 1998) with the microsomes or cDNA-expressed CYP3A4 in the presence of the NADPH-generating system before the addition of paclitaxel (10 μM). Fisetin (1 mM stock solution in methanol used to make 30 μM concentration in the incubation mixture) was added to the tubes, the methanol was evaporated to dryness under a nitrogen stream, and fisetin was redissolved in 150 mM KCl/50 mM Tris-HCl buffer (pH 7.4) by sonication at 80°C for 20 min. After the 20-min sonication, fisetin had not lost its activity. The other components of incubation were added subsequently, and the incubations and HPLC analysis were performed as described above. Inhibition studies were not done with docetaxel, because CYP3A enzymes apparently formed all the metabolites studied.
Statistics. Data are presented as means ± S.D. of the experimental values. Analysis of statistical significance (*p < 0.05 or **p < 0.01), where appropriate, was performed using a one-sided Student's t test. The p values corresponded to at least two separate determinations.
Results
Microsomal Metabolism of Paclitaxel. The metabolites isolated from incubations with rat liver microsomes were characterized: tR = 3.16 min, MH + 886.3, di-OHP; tR = 3.92 min, MH + 870.3, OHP; tR = 4.73 min, MH + 870.3, OHP; tR = 5.56 min, MH + 870.3, OHP (which was different from 6α-OHP and was concluded to be a precursor of di-OHP with the OH position not yet characterized) (Table 2). Rats formed, primarily, C3′-OHP; less, C2-OHP and a minor, previously not reported di-OHP, whose exact positions of hydroxyls could not be identified from MS/MS spectra. C3′-OHP was identified by MS as OHP (tR = 3.92 min) and by comparison with an authentic standard. C2-OHP was identified as OHP (tR = 4.73 min) by MS and specifically as C2-OHP, because it was the second most abundant product next to C3′-OHP found in rat liver microsomes (Table 3), in accordance with Walle et al. (1993) and Anderson et al. (1995). HPLC conditions used by Anderson and Walle (1993) are described in the legend for Table 2. Moreover, rat microsomes metabolized paclitaxel to a previously not reported OHP (tR = 5.56 min), which was characterized by HPLC-MS. This unknown OHP was quantitatively the third most abundant rat metabolite and was observed in CYP3A1/2-induced as well as in uninduced rat liver microsomes (Table 3). This metabolite has a close, but reproducibly different tR (80.40% ± 0.15% of paclitaxel) from that of 6α-OHP found in human microsomes (tR = 79.55 ± 0.15% of paclitaxel), and the percentage tR proved to give higher reproducibility than absolute tR values in minutes.
6α-Hydroxypaclitaxel (6α-OHP), formed as major metabolite in human microsomes, was not formed by microsomes from pigs, minipigs, or rats. The human microsomes from different individuals also formed C3′-OHP and C2-OHP (Table 3).
The pig and minipig microsomes produced predominantly the above-mentioned unknown OHP, which created 95% of all detected products of paclitaxel, whereas it was a minor product in rats. C3′-OHP was a minor metabolite in minipigs and both sexes of pigs; C2-OHP was a minor product, which was detected only in minipigs. The overall metabolic rates in minipigs were higher than in pigs and untreated rats (Table 3).
Kinetic Characterization of Paclitaxel Metabolism. We compared the kinetic parameters of paclitaxel oxidation not previously reported in pig and minipig microsomes with those of human and rat microsomes. The formation of various paclitaxel metabolites in rat, pig, minipig, and human microsomes all showed hyperbolic kinetics. The sample curve of estimation of kinetic characteristics is shown in Fig. 4A. The Km and Vmax values for major paclitaxel oxidation, calculated from Lineweaver-Burk plots, are presented in Table 4.
Chemical Inhibition Studies with Paclitaxel. Inhibition of the formation of paclitaxel metabolites by selective CYP3A inhibitor TAO is shown in Fig. 2, A to C. In the CYP3A1/2-induced rat microsomes from rats treated with PCN, TAO significantly inhibited the conversion of paclitaxel to all four detected metabolites (Fig. 2A). In minipig microsomes, TAO inhibited the formation of C3′-OHP, C2-OHP, and the unknown OHP (tR at 80.4% of paclitaxel) (Fig. 2B). In the male and female pig microsomes, the formation of C3′-OHP and of the unknown OHP (80.4% tR) was inhibited by TAO (Fig. 2C); inhibition of this unknown OHP in minipigs was stronger than in pigs. In human microsomes, TAO strongly inhibited the production of C3′-OHP and C2-OHP. As expected, TAO did not inhibit the formation of the major human metabolite 6α-OHP, which is catalyzed by CYP2C8 (Rahman et al., 1994). Fisetin significantly inhibited the formation of 6α-OHP in human liver microsomes and exerted only a mild but significant effect on the formation of other paclitaxel metabolites (Fig. 3).
Metabolism of Paclitaxel by cDNA-Expressed P450s. Human CYP3A4 expressed in E. coli metabolized paclitaxel to C3′-OHP and C2-OHP (Table 3), indicating that CYP3A4 formed both these products. Moreover, TAO inhibited the formation of both metabolites by ≥90% (Fig. 3). In contrast, fisetin, which significantly inhibited the formation of the CYP2C8-produced 6α-OHP, inhibited the formation of C3′-OHP and C2-OHP markedly less than did TAO, in a manner similar to that of human microsomes.
Microsomal Metabolism of Docetaxel. In contrast to paclitaxel, the rat, human, pig, and minipig microsomes formed the same metabolites of docetaxel, with hydroxydocetaxel being the main product. In contrast, CYP3A-induced microsomes from rats treated with PCN produced a markedly different pattern of docetaxel metabolites when analyzed under the same conditions (10 μM docetaxel, 30-min incubation, 1 mg/ml microsomal protein), in that virtually all docetaxel was consumed and both diastereomeric hydroxyoxazolidinones were predominant metabolites. The difference was related to high CYP3A expression in PCN microsomes. Identical data were obtained when the incubation lasted only 10 min (Table 5). However, reduction of the microsomal protein concentration to 0.1 mg/ml resulted in a pattern of docetaxel metabolites similar to that of untreated rats and the other species used (data not shown).
Kinetic Characteristics of Docetaxel Metabolism. Docetaxel metabolism in pig, minipig, human, and untreated rat microsomes revealed differences in Km, Vmax, and enzyme efficiency (Vmax/Km) (Table 6). The sample curve of estimation of kinetic characteristics is shown in Fig. 4B. Possible differences in metabolic rates and consequent inactivation of docetaxel may also occur among the studied species in vivo.
Discussion
This study compared metabolic profiles and rates of paclitaxel and docetaxel in different species, especially in pigs and minipigs, both of which have not been studied before in this context. Minipig liver microsomes expressed activities characteristic of human CYP3A4 (nifedipine oxidation), 2A6 (coumarin 7-hydroxylation), 2D6 (bufuralol 1′-hydroxylation), 2C9 (tolbutamide hydroxylation), and 2E1 (p-nitrophenol hydroxylation) (Anzenbacher et al., 1998). The testosterone 6β-hydroxylation activity of minipig microsomes also corresponded to values for human liver microsomal activity (Anzenbacher et al., 1998), and this enzyme activity was higher in minipigs than in pigs (Skaanild and Trios, 1997). In pigs, we paid particular attention to CYP3A and CYP2C8, which are important in the metabolism of paclitaxel and docetaxel, and we have found great similarity between the metabolic profiles of testosterone in human and pig livers in comparison with rats and dogs. This correlated with the fact that CYP3A4 is the major P450 isozyme in the human liver and suggested that a CYP3A isozyme must also be the major hepatic P450 in the pig. The same testosterone metabolic profile was found between pig and human hepatocytes (Donato et al., 1999). The cross-reactivity of human CYP3A4 antibodies with pig liver microsomal protein suggests at least partial homology between the human CYP3A4 and its equivalent isoform in pig liver. Furthermore, human cDNA probe for CYP3A4 hybridized to pig liver mRNA indicated that the pig CYP3A29 gene has considerable homology to the human gene (Jurima-Romet et al., 2000). Minipigs also contained proteins similar to human CYP3A4 and/or 3A5 as shown by immunoblotting (Anzenbacher et al., 1998). Moreover, the N-terminal amino acid sequences of the minipig liver microsomal CYP3A and pig CYP3A29 were identical and showed 60% identity with human CYP3A4 (Soucek et al., 2001). Despite all of these similarities between minipigs, pigs, and humans, we found marked species differences in paclitaxel metabolism. The 40% difference between the amino acid sequence of human CYP3A4 and CYP3A29 ortholog in pigs may result in large differences in substrate specificity. Therefore, it is not surprising that we have found differences in paclitaxel metabolism. Minipig and regular pig microsomes oxidized paclitaxel mainly to an OHP different from all fully characterized OHP metabolites. It was a minor metabolite in rats and absent in humans (Vaclavikova et al., 2003), and may be 4-hydroxypaclitaxel or 10-hydroxypaclitaxel (Slatter et al., 1997). The results show that structurally and metabolically similar pig/minipig and human P450 enzymes may form different products of paclitaxel and suggest significant differences in their catalytic sites. 6α-OHP was the main metabolite in human microsomes and C3′-OHP was the main product in rats. Thus, the metabolic profile of paclitaxel in pigs and minipigs differed completely from that in humans and also in rats. Considering the fact that small differences between amino acid sequence of P450 orthologs may result in significantly different metabolism of xenobiotics, identical metabolic profiles of paclitaxel and docetaxel found here suggest that pig and minipig CYP3A enzymes are closely related. The overall metabolic rates in untreated pigs, minipigs, and rats were similar but significantly lower than those in most human microsomes due to the fact that humans metabolize paclitaxel mainly using CYP2C8. The 10-fold higher metabolic rate in microsomes of rats treated with PCN than in untreated rats corresponded to CYP3A1/2 induction. The Vmax/Km for 6α-OHP (6.5) was significantly higher than that for any other metabolite in these species except PCN-treated rats, which represent an artificial case with severalfold higher CY3A1/2 levels than in uninduced rats.
TAO is a known inhibitor of CYP3A activity in human (Harris et al., 1994b), rat (Nedelcheva et al., 1998), and minipig (Anzenbacher et al., 1998) microsomes. In this study we investigated the effect of TAO on paclitaxel metabolism in pigs and minipigs in comparison with its effect in rats and humans. Although human P450s and their established markers and inhibitors are not always appropriate for extrapolation to other animal P450s, TAO was found to be a good inhibitor of human CYP3A4 ortholog form CYP3A29 in minipigs. Harris et al. (1994b) observed that the formation of C3′-OHP was inhibited by 20 μM TAO to 15% of control activity. In this work with PCN-treated rat liver microsomes, TAO (30, 60, or 100 μM) caused a concentration-related inhibition of C3′-OHP to 15% and that of C2-OHP to 5% of control activity (Fig. 2A). The inhibition of C3′-OHP and C2-OHP formation in minipig microsomes by 30 μM TAO was virtually complete as well as that of C3′-OHP in male and female pig microsomes. The formation of the unknown OHP was inhibited in a dose-related manner by 30, 60, and 100 μM TAO to 10% of control activity in minipigs and less in female (to 30%) and male (to 45%) pigs; the inhibition in microsomes isolated from PCN-treated rats reached 30% of control activity without evidence of a concentration effect. In human liver microsomes and CYP3A4-containing membranes, TAO significantly inhibited the production of C3′-OHP and C2-OHP but did not affect 6α-OHP formation, whereas the opposite was true for the effect of fisetin. The role of human CYP3A4 in C2-OHP formation was clearly demonstrated by its production with cDNA-expressed CYP3A4 as well as its inhibition by TAO. The formation of C3′-OHP, C2-OHP in all four species, and unknown OHP in minipigs, pigs, and rats was strongly inhibited by TAO. Moreover, our data indicate that human CYP2C8 was the only P450 involved in the formation of 6α-OHP. These inhibition studies indicated that the minor paclitaxel metabolite in human microsomes (44% tR of paclitaxel) was apparently not the di-OHP reported by Harris et al. (1994a), since it was not sufficiently inhibited by fisetin, although CYP2C8 is involved in di-OHP formation and fisetin was shown to inhibit CYP2C8.
The sensitivity of particular P450 enzymes to TAO was high in pigs, minipigs, and humans, but significantly less in rats and also varied for the different metabolites. In a previous study, TAO exerted differential inhibition of nifedipine oxidation catalyzed by CYP3A in minipig and human microsomes (Anzenbacher et al., 1998). It may be due to structural differences in the CYP3A active sites of minipigs and humans. Different sensitivity to inhibition by TAO also supports the idea that catalytic sites of CYP3A enzymes in these species differ. Fisetin seemed to be a rather selective inhibitor of CYP2C8 in human microsomes. Fisetin (30 μM) inhibited CYP2C8-catalyzed formation of 6α-OHP by more than 80%. This result was particularly interesting since this 6α-hydroxylation of paclitaxel is used as a marker substrate activity of human CYP2C8 (Projean et al., 2003). However, fisetin is not a specific inhibitor of CYP2C8, since it also inhibited CYP3A-catalyzed formation of C3′-OHP in human microsomes by up to 40% in this study and by less than 20% in another study with human and rat microsomes (Vaclavikova et al., 2003). Moreover, it was a good inhibitor (IC50 = 9 μM) of CYP1A1/1A2-dependent 7-ethoxycoumarin O-deethylase activity in rat liver microsomes (Moon et al., 1998). However, CYP1A1/1A2 do not participate in taxane metabolism.
In untreated rats, pigs, minipigs, and humans, docetaxel was metabolized primarily to hydroxydocetaxel, and two diastereomeric hydroxyoxazolidinones were minor products, whereas CYP3A1/2-induced rat microsomes metabolized docetaxel mainly to both hydroxyoxazolidinones, and hydroxydocetaxel was a minor product. Shorter incubations and less microsomal protein resulted in a metabolic pattern similar to that of untreated rats or the other species. The Km values and the metabolic rates of hydroxydocetaxel formation were similar in all the tested species. However, some human livers metabolized docetaxel at 10-fold lower rates than did the other species, apparently due to great variability in CYP3A4/5 expression. A new mobile phase (acetonitrile/methanol/tetrahydrofuran/water, 21: 35:2:42 v/v) enabled isocratic separation of hydroxydocetaxel and both diastereomeric hydroxyoxazolidinones within a reasonable time of 33 min for complete analysis.
It seems interesting that humans metabolized paclitaxel at significantly higher rates than the other species, but the opposite was true for docetaxel. This phenomenon is partly due to the fact that paclitaxel and docetaxel are metabolized by CYP3A orthologs, whereas paclitaxel is also oxidized by the uniquely human CYP2C8.
In conclusion, both taxanes are oxidized to products that are less antineoplastic than the parent drug. The antineoplastic effects of taxanes observed in vivo are apparently related to metabolic rates rather than to metabolic profiles. Humans (in vitro) metabolize paclitaxel at higher rates than do the other untreated species, and the anticancer effects may, therefore, be less than might be observed in the other species. CYP3A orthologs formed similar products of docetaxel. In the case of paclitaxel, humans, pigs, minipigs, and rats, respectively, formed different major metabolites, apparently due to different catalytic sites of particular P450s and due to participation of specific human CYP2C8, which is responsible for the human 6α-hydroxylation of paclitaxel. Although inhibition studies with TAO indicate participation of CYP3A ortholog forms in paclitaxel metabolism in different species, further studies are needed to fully characterize pig, minipig, and human P450 orthologs. Knowledge of which P450 enzymes form the respective metabolites enables extrapolation of P450 expression in vivo to metabolic inactivation, and consequent species and possibly human individual differences in the antineoplastic effects of taxanes, which can help to predict pharmacological interactions.
Acknowledgments
We gratefully acknowledge the generous gift of the authentic samples of docetaxel and its five metabolites by Agnes Gonthier (Aventis, Strasbourg, France). Pig livers were kindly provided by Dr. B. Hlavackova, Slaughter House Cesky Brod, and by Dr. M. Rezek, Research Institute of Animal Production.
Footnotes
-
↵1 Abbreviations used are: C3′-OHP, C3′-hydroxypaclitaxel; P450, cytochrome P450; C2-OHP, C2-hydroxypaclitaxel; di-OHP, di-hydroxypaclitaxel; 6α-OHP, 6α-hydroxypaclitaxel; TAO, troleandomycin; PCN, pregnenolon-16α-carbonitrile; MS, mass spectrometry; MS/MS, tandem MS; tR, retention time; HPLC, high performance liquid chromatography; NPR, NADP(H)-cytochrome P450 reductase;
-
This work was financed by Internal Grant Agency of the Ministry of Health of the Czech Republic Grants NL/6715-3 and NL/7567-3 (I.G., R.V., S.H., P.S.), U.S. Public Health Service Grants R01 CA90426 and P30 ES00267 (F.P.G.), and Grant Agency of the Czech Republic Grant 203/02/1152 (P.S.).
-
Citation of meeting abstracts where the work was presented: Václavíková R, Horský S, and Gut I (2003) Interspecies variability of paclitaxel metabolism in humans, rats, minipigs and regular pigs, in XXII. Xenobiochemical Symposium, 2003 Jun 9–11, Smolenice, Slovak. Abstract in memorial volume Posters, Section 4, p. 51–52; Václavíková R, Horský S, and Gut I (2003) New in vitro metabolites of paclitaxel in humans, rats, minipigs and regular pigs and P450 involved in their formation, at the 13th International Conference on Cytochromes P450, 2003 Jun 29–Jul 3, Prague, Czech Republic. Abstract in Chemické Listy 97(6), S188, WP54; Václavíková R, Horský S, and Gut I, New in vitro metabolites of paclitaxel in humans, rats, minipigs and regular pigs. Proceedings: Cytochromes P450, Biochemistry, Biophysics and Drug Metabolism, Monduzzi Editore, International Proceedings division, Work Code D629C0043; Václavíková R, Svobodová S, Horský S, Šimek P, and Gut I (2003) Poster; Taxanes: antitumor effects and interspecies differences in metabolism. Active presentation of poster in EURO-TOX 2003, Sep 28–Oct 1, Florence, Italy.
- Received November 3, 2003.
- Accepted February 25, 2004.
- The American Society for Pharmacology and Experimental Therapeutics