Abstract
The purpose of the study was to elucidate human intestinal cytochrome P450 isoform(s) involved in the metabolism of an antihistamine, ebastine, having two major pathways of hydroxylation andN-dealkylation. The ebastine dealkylase in human intestinal microsomes was CYP3A4, based on the inhibition studies with antibodies against CYP1A, CYP2A, CYP2C, CYP2D, CYP2E, and CYP3A isoforms and their selective inhibitors. However, ebastine hydroxylase could not be identified. We then examined the inhibitory effects of anti-CYP4F antibody and 17-octadecynoic acid, an inhibitor of the CYP4 family, on ebastine hydroxylation in intestinal microsomes, since CYP4F was recently found to be the predominant ebastine hydroxylase in monkey intestine; and a novel CYP4F isoform (CYP4F12), also capable of hydroxylating ebastine, was found to exist in human intestine. However, the inhibitory effects were only partial (about 20%) and thus it was thought that, although human CYP4F was involved in ebastine hydroxylation, another predominant enzyme exists. Further screening showed that the hydroxylation was inhibited by arachidonic acid. CYP2J2 was selected as a candidate expressed in the intestine and closely related to arachidonic acid metabolism. The catalytic activity of recombinant CYP2J2 was much higher than that of CYP4F12. Anti-CYP2J antibody inhibited the hydroxylation to about 70% in human intestinal microsomes. These results demonstrate that CYP2J2 is the predominant ebastine hydroxylase in human intestinal microsomes. Thus, the present paper for the first time indicates that, in human intestinal microsomes, both CYP2J and CYP4F subfamilies not only metabolize endogenous substrates but also are involved in the drug metabolism.
Ebastine is a potent nonsedative H1-receptor antagonist (Fig.1), and after oral administration to experimental animals and humans, the agent is almost completely metabolized to the pharmacologically active principle, the carboxylated metabolite (carebastine), and other inactive metabolites (Fujii et al., 1994; Matsuda et al., 1994; Yamaguchi et al., 1994). Carebastine alone was the major metabolite detectable in the blood. Our previous in situ studies using rats indicated that the small intestine extensively converted the orally given ebastine to carebastine via hydroxylated ebastine and the dealkylated metabolite (Fujii et al., 1997). Therefore, it seemed that small intestine plays an important role in the first-pass metabolism of this drug, and the enzymes responsible for its metabolism exist there.
Metabolic pathways of ebastine.
We reported that ebastine was primarily metabolized by human liver microsomes to two metabolites, hydroxy- and desalkyl-ebastine (Hashizume et al., 1998). N-Dealkylation to desalkyl-ebastine was mediated by CYP3A4, whereas hydroxylation to hydroxy-ebastine, the most important intermediate metabolite yielding carebastine, was mediated by unidentified P450(s) other than CYP3A4. Our recent studies revealed that two novel CYP4F isoforms (P450 MI-2 and CYP4F12) obtained from monkey and human small intestine, respectively, were involved in the ebastine hydroxylation (Hashizume et al., 2001a,b). Results obtained in an inhibition study using anti-CYP4F antibody indicated the involvement of CYP4F isoform (P450 MI-2) in the drug metabolism in monkey intestinal microsomes, although this subfamily had been recognized to be connected with the endobiotic metabolism.
Based on these findings, we attempted to elucidate P450 isoform(s) involved in the metabolism of ebastine in human small intestine. We found that CYP2J and CYP4F were involved in the metabolism of ebastine. The present finding suggests that CYP2J and CYP4F play an important role in the overall first-pass metabolism of drugs in humans.
Experimental Procedures
Materials.
[14C]Ebastine [4′-tert-butyl-4-[4-([ring-U-14C]diphenylmethoxy)piperidino]butyrophenone] was synthesized by using the method described previously (Fujii et al., 1994), with a specific activity of 1.08 MBq/mg and radiochemical purity of 99%. Authentic metabolites [hydroxy-ebastine, 4′-(2-hydroxy-1,1-dimethylethyl)-4-[4-(diphenylmethoxy)piperidino]butyrophenone; desalkyl-ebastine, 4-(diphenylmethoxy)piperidine; and carebastine, 4′-(2-carboxy-1,1-dimethylethyl)-4-[4-(diphenylmethoxy)piperidino]butyrophenone] were supplied by Almirall-Prodesfarma S. A. (Barcelona, Spain). NADPH and dilauroyl-l-3-phosphatidylcholine were purchased from Wako Pure Chemical Industries (Osaka, Japan). 17-Octadecynoic acid (17-ODYA), leukotriene B4, and arachidonic acid were from Cayman Chemical (Ann Arbor, MI). Furafylline, sulfaphenazole, and ketoconazole were products of Salford Ultrafine Chemical and Research (Manchester, UK). SKF-525A, coumarin, orphenadrine, tranylcypromine, quinidine, diethyldithiocarbamate, and lauric acid were purchased from Sigma Chemical (St. Louis, MO). Human small intestine cDNA library, restriction enzymes, T4 DNA ligase, Pyrobest DNA polymerase and human NADPH-P450 reductase were obtained from Takara Shuzo (Shiga, Japan). Other chemicals used were of the highest grade commercially available. Recombinant CYP1A1, CYP1A2, CYP1B1, CYP2A6, CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP2E1, CYP3A4, and CYP4A11 were purchased from Gentest (Woburn, MA) or Sumika Chemical Analysis Service (Osaka, Japan).
Microsomes.
Human small intestinal microsomes (lot 9, 10, 11, 12, 13, 49, 50, 52, 53, 54, and 58) were obtained from the International Institute for the Advancement of Medicine (Exton, PA). Human pooled liver microsomes (lot 610) were obtained from XenoTech (Kansas City, KS). All microsomes were stored at −80°C until use.
Cloning and Expression of CYP2J2.
The entire coding region of CYP2J2 was isolated from the human small intestine cDNA library with the polymerase chain reaction. Polymerase chain reaction was carried out with Pyrobest DNA polymerase as follows: denaturation at 96°C for 30 s, annealing at 55°C for 30 s, and extension at 72°C for 90 s, followed by 35 cycles. The upstream and downstream parts of the coding region were amplified based on the CYP2J2 cDNA sequence (GenBank accession no. U37143) (Wu et al., 1996) by using primer sets as follows: the upstream primers, 2J2-hind (5′- AAA AAG CTT AAA AAA ATG CTC GCG GCG ATG GGC TCT −3′) and 2J2–750a (5′- TTC CAG TTG CTG AAG AGA GT −3′); the downstream primers, 2J2–560 (5′- TTC AAG ATC AAC AAT GCA GT −3′) and 2J2-Eco (5′- GGG GAA TTC CAA TAT TAC ACC TGA GGA AC −3′). Each fragment was cut withHindIII and NcoI or NcoI andEcoRI. After a pBluescript vector (Stratagene, La Jolla, CA) was digested by HindIII and EcoRI, the vector and two digested fragments were ligated at the same time: the upstream and downstream fragments were joined at the NcoI site and inserted between the HindIII and EcoRI sites of a vector. The resulting cDNA was subcloned. The nucleotide sequence of the isolated clone was confirmed by sequence.
The pBluescript vector containing the entire coding region of CYP2J2 was digested with EcoRI, blunt-ended by T4 DNA polymerase, and ligated with aHindIII linker. After being subcloned, the cDNA was cut out of the plasmid with HindIII. The fragment was introduced into the HindIII site of the yeast expression vector pGYR1 with the glyceraldehyde 3-phosphate dehydrogenase promoter and the yeast NADPH-P450 reductase genes (Sakaki et al., 1992) for expression. pGYR1 was a kind gift of Dr. Y. Yabusaki of Sumitomo Chemical Co. (Hyogo, Japan). The resulting plasmid was used for transformation of the Saccharomyces cerevisiae AH22 strain (Murakami et al., 1990). Yeast was cultured as described (Murakami et al., 1986; Sakaki et al., 1991) and the microsome was prepared as described previously (Oeda et al., 1985).
Preparation of Antibodies.
Polyclonal antibodies against CYP1A, CYP2C, CYP2D, and CYP3A were purchased from Daiichi Pure Chemicals (Tokyo, Japan). Polyclonal anti-CYP2E1 antibody and monoclonal anti-CYP2A6 antibody were obtained from Gentest. Anti-CYP2J and CYP4F antibodies were prepared as described previously (Imaoka et al., 1990; Hashizume et al., 2001b). The anti-CYP2J antibody could cross-react with recombinant human CYP2J2 but not with the following recombinant P450 isoforms: CYP1A1, CYP1A2, CYP1B1, CYP2A6, CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP2E1, CYP3A4, and CYP4A11 (data not shown). Anti-CYP2J and CYP4F IgG and control IgG were purified from rabbit sera using protein A Sepharose CL-4B (Amersham Biosciences UK, Ltd., Little Chalfont, Buckinghamshire, UK).
Incubations.
The reaction mixture for the microsomal metabolism consisted of [14C]ebastine (kinetic study: 1.25, 5, 10, 20, and 50 μM; others 5 or 20 μM), microsomes (human intestine, 0.04–0.2 mg; human liver, 0.08 mg; yeast cells expressing CYP4F12, 95 pmol; CYP4F2, 107 pmol; and CYP2J2, 0.07–1.0 pmol), and 0.8 mM NADPH in a final volume of 0.5 ml of 50 mM potassium phosphate buffer (pH 7.4). For the reconstituted system, 22 pmol of native CYP2J3 (originally named P450 3b) purified from rat liver previously (Imaoka et al., 1990), human NADPH-P450 reductase (0.5 unit), and dilauroyl-l-3-phosphatidylcholine (5 μg) were used in place of the above microsomes. Assay conditions were such that metabolite formation and parent drug consumption were linear with respect to incubation time and protein concentration. For the kinetic study, protein concentrations of individual microsomes were adjusted to avoid further metabolism of hydroxy-ebastine, and the initial velocities of ebastine hydroxylation and N-dealkylation were measured. The reaction was started by the addition of NADPH and stopped after incubation at 37°C for 30 min by the addition of acetonitrile (2 ml). After centrifugation at 800g for 10 min, an aliquot of the supernatant was evaporated to dryness with a centrifugal concentrator. The residue was dissolved in 100 μl of methanol and 30 μl was injected onto a 250 × 4.6 mm Inertsil ODS-3 reverse-phase column (GL Science, Tokyo, Japan) maintained at 40°C (Hashizume et al., 1998). The mobile phase consisted of 12 mM ammonium acetate buffer (pH 4.5) and acetonitrile, at a flow rate of 1.0 ml/min. The proportion of acetonitrile was maintained at 35% from 0 to 3 min and then increased to reach 85% at 25 min. The elution profile of the radioactive compounds was monitored by a FLO-ONE β radioactivity flow detector (Packard Instrument Co., Meriden, CT).
Chemical Inhibition.
The following P450 isoform-selective inhibitors were used: 50 μM furafylline (an inhibitor for CYP1A2), 500 μM coumarin (for CYP2A6), 50 μM orphenadrine (for CYP2B6), 10 μM sulfaphenazole (for CYP2C), 20 μM tranylcypromine (for CYP2C19), 10 μM quinidine (for CYP2D6), 50 μM diethyldithiocarbamate (for CYP2E1), 1 μM ketoconazole (for CYP3A), and 100 μM lauric acid (for CYP4A). The concentrations of inhibitors described above were decided based on their published IC50,K i, orK m values for P450 isoform-specific reactions. 17-ODYA (a mechanism-based inhibitor for the CYP4 family; 5, 25, and 100 μM), leukotriene B4 (a CYP4F substrate; 10, 50, and 100 μM), and arachidonic acid (a CYP4F substrate; 10, 50, and 100 μM) were also used (Shak et al., 1985;Kikuta et al., 1994; Zou et al., 1994; Powell et al., 1998). Inhibitors except furafylline and 17-ODYA were incubated at 37°C for 30 min in a final volume of 0.5 ml of 50 mM potassium phosphate buffer (pH 7.4) containing 0.2 mg of pooled human intestinal microsomes (n = 3), 0.8 mM NADPH, and 5 μM [14C]ebastine. The substrate concentration selected was 2 times the K m value for ebastine hydroxylation in human intestinal microsomes. Furafylline and 17-ODYA were preincubated with human intestinal microsomes and NADPH for 20 min in the absence of [14C]ebastine. The assay was carried out as described under Incubations.
Immunoinhibition.
The anti-P450 antibodies, preimmune sera (as control for antibodies against CYP1A, CYP2C, CYP2D, CYP2E, and CYP3A), 25 mM Tris-HCl buffer (pH 7.4) (as control for anti-CYP2A6 antibody), or preimmune IgG (as control for antibodies against CYP4F and CYP2J) were preincubated with intestinal microsomes (0.2 mg), liver microsomes (0.08 mg), and yeast microsomes expressing CYP2J2 (0.2 mg) at room temperature for 30 min. The reaction was started by the addition of [14C]ebastine (final concentration, 5 or 20 μM) and was stopped with 2 ml of acetonitrile. The assay was carried out as described under Incubations.
Data Analysis.
The values represent the average of duplicate or triplicate determinations. Kinetic parameters (V
max andK
m) were obtained by fitting data to the following Michaelis-Menten equation using the curve-fitting software GraFit version 3.0 (Erithacus Software, Staines, UK):
Results
Metabolism of Ebastine by Human Intestinal Microsomes.
Figure2 shows a typical reverse-phase HPLC radiochromatogram obtained after incubation of [14C]ebastine with human intestinal microsomes. In the presence of NADPH, human intestinal microsomes metabolized ebastine to two major metabolites, hydroxy- and desalkyl-ebastine. Carebastine, the active principle, was also partly formed from hydroxy-ebastine. Hydroxylation and N-dealkylation of ebastine followed simple Michaelis-Menten kinetics in human intestinal microsomes under the condition in which carebastine was not formed (Fig. 3). The kinetic parameters of five individual microsomes are summarized in Table1. Data demonstrated large interindividual variability in the formation of the two major metabolites, and the K m values were significantly different (p < 0.03) between the hydroxylation (2.6 ± 2.3 μM) and the N-dealkylation (39 ± 24 μM). There are no significant differences in theV max andV max/K mvalues between the hydroxylation and the N-dealkylation (p > 0.05). Both reactions were inhibited almost completely by SKF-525A and anti-NADPH-P450 reductase antibody, suggesting that these reactions were mediated by P450 (data not shown).
Reverse-phase HPLC radiochromatogram of metabolites formed in incubations of [14C]ebastine with human intestinal microsomes. Human intestinal microsomes (lot 12, 0.2 mg) were incubated at 37°C for 30 min with 20 μM [14C]ebastine and 0.8 mM NADPH in a final volume of 0.5 ml. Metabolites were resolved by reverse-phase HPLC and the eluent radioactivity was monitored with an online radioactive flow detector. Carebastine, an acid metabolite, was formed under this condition.
Velocity versus substrate concentration curve for ebastine hydroxylation and N-dealkylation by the human intestinal microsomes. In the presence of NADPH, human intestinal microsomes (lot 12, 0.06 mg) were incubated at 37°C for 30 min with [14C]ebastine (1.25, 5, 10, 20, and 50 μM) in a final volume of 0.5 ml, and the metabolites were analyzed by HPLC. Open and closed circles indicate the catalytic activities of hydroxylation andN-dealkylation of ebastine, respectively. Each point represents mean (±S.D.) of triplicate determinations. Lines represent functions determined by nonlinear least-squares regression analysis using the Michaelis-Menten equation.
Kinetic parameters for hydroxylation and N-dealkylation of ebastine by human intestinal microsomes
First Screening for Ebastine Hydroxylase in Human Intestinal Microsomes.
Effects of P450 isoform-selective inhibitors on ebastine hydroxylation and N-dealkylation by human intestinal microsomes are shown in Fig.4. Ebastine N-dealkylation was almost completely inhibited by ketoconazole (selective for CYP3A), whereas the hydroxylation was not affected by all inhibitors examined: furafylline (for CYP1A2), coumarin (for CYP2A6), orphenadrine (for CYP2B6), sulfaphenazole (for CYP2C), tranylcypromine (for CYP2C19), quinidine (for CYP2D6), diethyldithiocarbamate (for CYP2E1), ketoconazole (for CYP3A), and lauric acid (for CYP4A11). Figure5 shows the effects of polyclonal antibodies against CYP1A, CYP2A, CYP2C, CYP2D, CYP2E, and CYP3A on ebastine hydroxylation and N-dealkylation by human intestinal microsomes. Similar to chemical inhibitors, none of the anti-P450 antibodies inhibited ebastine hydroxylation and only anti-CYP3A antibody inhibited the dealkylation completely. These results indicate that CYP3A was the major ebastine dealkylase, whereas P450(s) other than the ordinary drug-metabolizing P450 was involved in hydroxylation.
Effects of P450 isoform-selective inhibitors on hydroxylation and N-dealkylation of ebastine by human intestinal microsomes. Human intestinal microsomes (final: 0.4 mg/ml) were incubated with 5 μM [14C] ebastine and 0.8 mM NADPH at 37°C for 30 min in the presence of various P450 inhibitors. Values in parentheses indicate the micromolar concentrations of inhibitors. Hatched and closed bars indicate the catalytic activities of hydroxylation and N-dealkylation of ebastine, respectively. Data are the average of duplicate determinations and expressed as percentage of control activity.
Effects of polyclonal antibodies against CYP1A, CYP2A, CYP2C, CYP2D, CYP2E, and CYP3A on hydroxylation andN-dealkylation of ebastine by human intestinal microsomes. Human intestinal microsomes (0.2 mg) were preincubated with various amounts of antisera or preimmune serum for 20 min at room temperature, and then incubated at 37°C for 30 min with 5 μM [14C]ebastine and 0.8 mM NADPH in a final volume of 0.5 ml. Open and closed circles indicate the catalytic activities of hydroxylation and N-dealkylation of ebastine, respectively. Data are the average of duplicate determinations and expressed as percentage of control activity.
Role of CYP4F in Ebastine Hydroxylation.
We determined the kinetic parameters for ebastine hydroxylation by the recombinant human CYP4F12 and CYP4F2, which were recently shown to be expressed in human small intestine and to participate in ebastine hydroxylation (Hashizume et al., 2001a). The K m andV max values of CYP4F12 were 3.0 μM and 0.354 nmol/min/nmol P450, respectively. Kinetic parameters of CYP4F2 could not be determined because catalytic activity (the activity at 20 μM ebastine, 0.005 nmol/min/nmol P450) was too low. In addition, CYP4F12 was found to produce the carboxylated metabolite, carebastine, via hydroxy-ebastine by the prolonged incubation (over 60 min).
To assess the contribution of CYP4F12 to ebastine hydroxylation in human intestinal microsomes, the effects of polyclonal anti-CYP4F antibody on the hydroxylation were examined. As shown in Fig.6, the inhibitory effect of anti-CYP4F antibody on ebastine hydroxylation by human intestinal microsomes was found to be only partial (about 20%); in contrast, the antibody completely inhibited ebastine hydroxylation by the recombinant CYP4F12 and leukotriene B4 ω-hydroxylation by human intestinal microsomes (data not shown). In addition, 17-ODYA, a mechanism-based inhibitor for the CYP4 family, also inhibited ebastine hydroxylation by human intestinal microsomes to an extent similar to that of the antibody (Fig. 7). These results indicate that although human intestinal CYP4F isoform was certainly involved in ebastine hydroxylation, its contribution in human intestinal microsomes was only partial, which was quite different from that observed in monkey intestinal microsomes (about 70% contribution;Hashizume et al., 2001b). This suggests that another predominant ebastine hydroxylase exists in human intestinal microsomes.
Effects of polyclonal anti-CYP4F antibody on ebastine hydroxylation by human intestinal microsomes. Human intestinal microsomes (0.2 mg) were preincubated with various amounts of anti-CYP4F antibody (40 mg IgG/ml) for 20 min at room temperature, and then incubated at 37°C for 30 min with 5 μM [14C]ebastine and 0.8 mM NADPH in a final volume of 0.5 ml. Closed circles indicate hydroxylation activity of ebastine. Data are the average of triplicate determinations and expressed as percentage of control activity.
Effects of 17-ODYA on ebastine hydroxylation by human intestinal microsomes. Human intestinal microsomes (0.2 mg) were preincubated with various amounts of 17-ODYA, as an inhibitor for the CYP4 family, at 37°C for 30 min and then incubated with 5 μM [14C]ebastine and 0.8 mM NADPH in a final volume of 0.5 ml. Closed circles indicate hydroxylation activity of ebastine. Data are the average of duplicate determinations and expressed as percentage of control activity.
Identification of Major Ebastine Hydroxylase in Human Intestinal Microsomes.
Results obtained above suggest that the ebastine hydroxylase belongs to P450 isoform(s) that have not been recognized to be ordinary drug-metabolizing P450s and, instead, seemingly possesses the substrate specificity to endogenous compounds. Thereby, to identify the predominant ebastine hydroxylase(s) in human intestinal microsomes, leukotriene B4 and arachidonic acid (both are CYP4F substrates) were used as the inhibitory probes. Leukotriene B4 inhibited ebastine hydroxylation by human intestinal microsomes to 85% of control (Fig.8A) and arachidonic acid, to 2% of control (Fig. 8B), suggesting that arachidonic acid more strongly interacts with the predominant ebastine hydroxylase.
Effects of leukotriene B4 (A) and arachidonic acid (B) on ebastine hydroxylation by human intestinal microsomes. Human intestinal microsomes (0.2 mg) were incubated at 37°C for 30 min with 5 μM [14C]ebastine and 0.8 mM NADPH in the presence of leukotriene B4 (A) or arachidonic acid (B). Closed circles indicate hydroxylation activity of ebastine. Data are the average of duplicate determinations and expressed as percentage of control activity.
Based on the results, CYP2J2 was selected out of P450 isoforms as a candidate expressed in the intestine (Zeldin et al., 1997) and closely related to arachidonic acid metabolism (Rifkind et al., 1995). We cloned CYP2J2 cDNA from the human small intestine cDNA library and expressed this P450 in yeast cells as described, and the kinetics for ebastine hydroxylation by the recombinant CYP2J2 were examined (Table2). TheV max value of ebastine hydroxylation by CYP2J2 was much higher than that by CYP4F12 (40.6 versus 0.354 nmol/min/nmol P450, respectively). TheK m value of CYP2J2 was half that of CYP4F12. TheV max/K mvalue of CYP2J2 was 250-fold higher than that of CYP4F12. To assess the involvement of the CYP2J subfamily in ebastine hydroxylation, we measured ebastine hydroxylase activity of purified native CYP2J3 (Imaoka et al., 1990). The reconstituted CYP2J3 protein also catalyzed ebastine hydroxylation (6.4 nmol/min/nmol P450, at 10 μM ebastine). At higher concentrations of CYP2J2 and CYP2J3, ebastine was rapidly oxidized to the carboxylic acid metabolite, carebastine, via hydroxy-ebastine and then disappeared completely.
Summary of kinetic parameters for ebastine hydroxylation by recombinant human CYP2J2, CYP4F12, and CYP4F2
To confirm the role of CYP2J2 as the main catalyst of ebastine hydroxylation in human intestinal microsomes, immunoinhibition experiments were performed using polyclonal anti-CYP2J antibody. As shown in Fig. 9, anti-CYP2J antibody inhibited ebastine hydroxylation by recombinant CYP2J2 (Fig. 9A) and human intestinal microsomes (Fig. 9B). In addition, anti-CYP2J antibody had the inhibitory effect on ebastine hydroxylase activity in human liver microsomes (25% inhibition at 10 mg IgG/mg microsomal protein). This is consistent with reports that CYP2J2 is also expressed in human liver (Wu et al., 1996; Gu et al., 2000). These results indicate that CYP2J2 is the predominant ebastine hydroxylase in human intestinal microsomes.
Inhibition of ebastine hydroxylase activity of recombinant CYP2J2 (A) and human intestinal microsomes (B) by anti-CYP2J antibody. Microsomes (0.2 mg) from yeast cells expressing CYP2J2 and human intestine were preincubated with various amounts of anti-CYP2J antibody (40 mg IgG/ml) for 30 min at room temperature, and then incubated at 37°C for 30 min with [14C]ebastine (for CYP2J2, 20 μM; for intestinal microsomes, 5 μM) and 0.8 mM NADPH in a final volume of 0.5 ml. Closed circles indicate hydroxylation activity of ebastine. Data are the average of triplicate determinations and expressed as percentage of control activity.
Discussion
Recently, clinical investigations have demonstrated that the small intestine participates in the “first-pass” metabolism of orally administered drugs like cyclosporine, midazolam, tacrolimus, and verapamil (Wu et al., 1995; Thummel et al., 1996; Lin et al., 1999). All of these drugs were substrates of CYP3A4; therefore, most studies of intestinal “first-pass” metabolism have focused on CYP3A and little is known about the involvement of other intestinal P450 isoforms. Therefore, the drug that is metabolized by P450(s) other than CYP3A in human small intestine appears to provide new findings for understanding intestinal drug metabolism in humans.
In the present study, we attempted to elucidate P450 isoform(s) involved in the small intestinal metabolism of an antihistamine, ebastine, metabolism of which has been characterized as follows. 1) Rat small intestine in situ almost completely converted the orally given ebastine to carebastine (via hydroxyl-ebastine) and desalkyl-ebastine (Fujii et al., 1997). 2) Ebastine was primarily metabolized by human liver microsomes to desalkyl- and hydroxy-metabolites (Hashizume et al., 1998). The former was shown to be produced by CYP3A4, while the latter, the intermediate metabolite leading to carebastine, was mediated by unidentified P450(s). 3) Very recently, CYP4F was found to be the predominant ebastine hydroxylase in monkey intestinal microsomes (Hashizume et al., 2001b), and a novel human CYP4F isoform (named CYP4F12), being capable of hydroxylating ebastine, was also found to exist in the small intestine (Hashizume et al., 2001a).
Human intestinal microsomes, similar to human liver microsomes, metabolized ebastine to desalkyl-ebastine and hydroxy-ebastine. Inhibition experiments using nine P450 isoform-selective inhibitors and six anti-P450 antibodies demonstrated that CYP3A4 was the main ebastine dealkylase, whereas unidentified P450(s) was involved in the hydroxylation.
We recently purified an ebastine hydroxylase belonging to the CYP4F subfamily from monkey intestinal microsomes (Hashizume et al., 2001b). The reconstituted P450 showed ebastine hydroxylase activity (V max, 37.0 nmol/min/nmol P450) and anti-CYP4F antibody strongly inhibited the hydroxylation by monkey intestinal microsomes, suggesting that CYP4F was the predominant ebastine hydroxylase in monkey intestinal microsomes. Our study also indicated that both CYP4F12 and CYP4F2 were expressed in human small intestine, and they had the ability of ebastine hydroxylation (Hashizume et al., 2001a). The kinetic analysis in the present study revealed that recombinant CYP4F12 was a quite efficient catalyst of ebastine hydroxylation. Although the results of inhibition experiments using anti-CYP4F antibody and 17-ODYA demonstrated the involvement of CYP4F in ebastine hydroxylation in human intestinal microsomes, the inhibitory effect was unexpectedly small (about 20%). This strongly suggested the presence of another predominant hydroxylase in human intestinal microsomes.
After further screening, ebastine hydroxylation by human intestinal microsomes was found to be strongly inhibited by arachidonic acid. Although arachidonic acid inhibited microsomal monooxygenation mediated by CYP1A2, CYP2C8, and CYP2C19 (Yamazaki and Shimada, 1999), these P450 isoforms did not have ebastine hydroxylase activity. Among P450 enzymes concerned with arachidonic acid metabolism, we could select CYP2J2 because this isoform was reported to be expressed in the small intestine (Zeldin et al., 1997). TheV max value of CYP2J2 for hydroxylation (40.6 nmol/min/nmol P450) was found to be much higher than those of CYP4F12 (0.354 nmol/min/nmol P450) and CYP3A4 (0.023 nmol/min/nmol P450; Hashizume et al., 1998). The value was as high as theV max of the CYP4F isoform purified from monkey intestinal microsomes (37 nmol/min/nmol P450; Hashizume et al., 2001b). In addition, anti-CYP2J antibody inhibited ebastine hydroxylation to about 70% of control by human intestinal microsomes. Results obtained in the kinetic analysis and immunoinhibition experiments indicated that CYP2J2 was the predominant ebastine hydroxylase in human intestinal microsomes. The immunoblotting data reported by Zeldin et al. (1997) suggest that CYP2J2 content in human intestinal microsomes was not high. However, the present studies demonstrated that ebastine hydroxylase activity in human intestinal microsomes was attributable to CYP2J2 for the most part. This finding is supported by our preliminary unpublished immunoquantification result (T. Hashizume and S. Imaoke), using antibody raised against the recombinant CYP2J2, that the high correlation (r = 0.924, p < 0.003) existed between the relative CYP2J2 content and ebastine hydroxylase activities in seven human intestinal microsomes.
It would be interesting to consider the enzymes responsible for metabolism of another antihistamine, terfenadine, which possesses certain chemical structural similarities to ebastine, as follows: the presence of a terminal tertiary butyl group, two aromatic rings, and a protonatable nitrogen. Terfenadine also has two similar major metabolic pathways of hydroxylation and N-dealkylation, and the carboxylated metabolite formed from the former is the active principle. However, unlike ebastine, both hydroxylation andN-dealkylation of terfenadine are reported to be catalyzed predominantly by CYP3A4 (Yun et al., 1993; Jurima-Romet et al., 1994;Ling et al., 1995; Rodrigues et al., 1995). Recently, a complete computer-assisted conformational and electronic characterization study demonstrated that the preferred three-dimensional spatial orientations were different in the molecular locations of the highest occupied and lowest unoccupied molecular orbitals of the two agents (Segarra et al., 1999). Furthermore, it was reported that, for terfenadine, additional points of interaction with macromolecules as a hydrogen bond donor were found, whereas noncardiotoxic antihistamines including ebastine were lacking. Therefore, the three-dimensional structural and electronic features of these two compounds seem to be related to the difference in the contribution of CYP3A to their microsomal metabolism.
In conclusion, the present paper for the first time indicates that, in human intestinal microsomes, both CYP2J and CYP4F subfamilies not only metabolize endogenous substrates such as arachidonic acid and leukotriene B4 but also are involved in the drug metabolism. To date, actual participation in the intestinal first-pass metabolism in humans has been elucidated for CYP3A4 alone, and in contrast to hepatic metabolism, only a limited number of P450 isoforms of small intestine have been characterized. The results here clearly revealed a potential role of CYP2J and CYP4F subfamilies in the xenobiotic metabolism in small intestinal microsomes and thus will serve for further understanding of the intestinal first-pass metabolism.
Acknowledgments
We thank Prof. Dr. Tadanobu Inaba (University of Toronto), Dr. Satoshi Matsumoto (Dainippon Pharmaceutical Co., Ltd.) and Mayuko Osada (Osaka City University) for helpful discussions. We are also grateful to Masashi Nakao (Dainippon Pharmaceutical Co., Ltd.) for providing [14C]ebastine.
Abbreviations
- P450
- cytochrome P450
- 17-ODYA
- 17-octadecynoic acid
- HPLC
- high-performance liquid chromatography
- Received July 13, 2001.
- Accepted October 5, 2001.
- The American Society for Pharmacology and Experimental Therapeutics