Introduction

Top Abstract Introduction Results Discussion References

Dimethyl-4,4'-dimethoxy-5,6,5',6'-dimethylene dioxybiphenyl-2,2'-dicarboxylate (DDB¹) is a synthetic compound derived from Schizandrin C, a component of Fructus schizandrae. DDB protects liver against carbon tetrachloride-, D-galactosamine-, thioacetamide-, and prednisolone-induced hepatic injury in experimental animals, although the exact mechanism is not well characterized (Liu et al., 1979, 1982). This compound has also been reported to be effective in improving liver functions of patients with chronic hepatitis (Lee et al., 1991). Currently, DDB is the most widely used remedy for patients with chronic viral hepatitis B in Asia. Plasma concentration of DDB after oral administration was relatively low (Dr. Y. Lee, Chunman University, Kwang-Ju, Korea, personal communication), and the half-life of this compound was 2 to 3 h, suggesting that metabolism of DDB plays a role in clearing the drug from the body. DDB was reported to modulate cytochrome P450 (P450) activities and glutathione S-transferase. Several investigators reported that DDB alone or in combination with garlic oil induced CYP2B1/2 levels in rats (Liu et al., 1981; Kim et al., 1995). However, few studies have dealt with the metabolism and pharmacokinetics of DDB. The information on the ability of DDB to modulate drug-metabolizing enzymes led us to hypothesize that P450 enzymes in the liver could metabolize DDB.

In this study, we investigated the in vitro metabolism of DDB in the presence of human liver microsomes. Five metabolites were identified using mass and NMR spectral analysis. We then identified the P450 isoforms mainly involved in the metabolism of DDB using several approaches: 1) the effect of selective chemical inhibitors on DDB metabolism, 2) determination of DDB metabolism by microsomes from cells expressing recombinant P450 isoforms, and 3) study of the correlation of DDB metabolism with marker activities of each P450 in different human liver microsomes.

Experimental Procedures

Materials. DDB (Fig. 1) was synthesized at the Kyung-Dong Pharmaceutical Co. (Kyungkido, Korea). Glucose 6-phosphate, -NADP⁺, glucose-6-phosphate dehydrogenase, troleandomycin, diethyldithiocarbamate, and tolbutamide were purchased from Sigma Chemical Co. (St. Louis, MO). Sulfaphenazole and ketoconazole were obtained from RBI/Sigma (Natick, MA). Methyl iodide, d₃-methyl iodide, and N,O-bis(trimethylsilyl)-trifluoroacetamide were purchased from Aldrich Chemical Co. (Milwaukee, WI). All other chemicals were of the highest commercially available grade.

View larger version (16K): [in this window] [in a new window]   Fig. 1.   HPLC chromatograms from analysis of an incubation containing DDB, human liver microsomes, and an NADPH-generating system.

Microsomes. Frozen human liver samples were kindly donated by Dr. Guengerich (Vanderbilt University, Nashville, TN). The microsomal fraction was prepared according to the method described elsewhere (Guengerich et al., 1986). Human B-lymphoblastoid-derived P450 microsomes were purchased from GENTEST Corp. (Woburn, MA). For the correlation experiment, 10 different human liver microsomes were purchased from Human Biologics Inc. (Phoenix, AZ).

Microsomal Incubations. Individual incubations (final volume = 0.5 ml) consisted of 1.0 mg/ml microsomal protein in 100 mM phosphate buffer (pH 7.4) with final concentrations of 5 mM glucose 6-phosphate, 1 mM -NADP⁺, and 1 U/ml glucose-6-phosphate dehydrogenase. DDB, buffer, and microsomes were mixed and kept at 37°C for 3 min, and the reaction was started by adding an NADPH-generating system. Incubations were conducted at 37°C and stopped by adding 0.1 ml 1 N HCl and 1 ml chloroform/isopropanol (85:15, v/v). Organic phase was taken after vortexing for 2 min and dried under nitrogen stream. Residue was reconstituted in a high-performance liquid chromatography (HPLC) mobile phase and injected into a HPLC column.

HPLC. All analyses were performed using a Waters M600 liquid chromatography system (Waters Corp., Milford, MA) consisting of M 600 quaternary pump, M717 autosampler, and M486 UV detector operated at 280 nm. Analyses were done at ambient temperature on a Capcell-pak C₈ column (4.6 × 250 mm, 5 µm, Shiseido, Japan). The flow rate was 1.0 ml/min. Separations were conducted using a 25-min gradient: 45% B for 1 min, followed by a linear increase to 70% B over 12 min, then maintained for 10 more min. Solvent A consisted of 10 mM ammonium phosphate adjusted with 1N HCl solution to pH 3.0; solvent B consisted of methanol/water (95:5, v/v).

Mass Spectroscopy (MS) and Electrospray MS (ESI-MS). The mass spectrometry system was based on a HP 5989A MS Engine mass spectrometer with a HP 59987A electrospray MS interface (Hewlett Packard, Palo Alto, CA). Samples dissolved in methanol containing 0.1% formic acid were introduced into the ESI interface by the syringe pump through a heated nebulizer probe (150°C) using nitrogen as a nebulizing gas. CapEx voltage was 100 V.

Electron Impact Ionization MS (EI-MS). A Hewlett Packard gas chromatography/mass selective detector (5890/5972) was used in EI mass analysis. A cross-linked Ultra-1 capillary column (18-m × 0.2-mm i.d., 0.33-µm film thickness; Hewlett Packard) was directly connected to the ion source. Helium at a flow rate of 0.9 ml/min was used as carrier gas. Initial oven temperature was 150°C; it was held for 1 min and subsequently increased by 20°C/min to 300°C and held there for 5 min. Before analysis, isolated metabolites were either methylated by treatment with 200 µl of CH₃I/acetone (10:90, v/v) containing 50 mg of K₂CO₃ at 60°C for 2 h or trimethylsilylated by treatment with 50 µl of N-methyl-N-trimethylsilyl-trifluoroacetamide/CH₃CN (30:70, v/v) at 80°C for 30 min. The reaction mixtures were directly injected into the column in a split mode.

NMR Spectroscopy. ¹H NMR experiments for structural elucidation of metabolites were carried out on a Bruker AMX 300 spectrometer in 5-mm tubes (Bruker Analytik GmbH, Rheinstetten, Germany). The metabolites were dissolved in 0.5 ml of CDCl₃ and ¹H NMR spectra were recorded. Chemical shifts for the ¹H NMR spectra are reported in parts per million relative to trimethylsilyl using the residual solvent (CDCl₃) signal at 7.2 ppm.

Metabolism by B-Lymphoblastoid Microsomes. Incubations with B-lymphoblastoid microsomes were conducted as described above with 50 µM DDB and microsomes containing 200 pmol of P450/ml, with an incubation time of 2 h.

Correlation Analysis. Correlation of the formation of DDB metabolites with P450 isoform-specific activities in the microsomes prepared from 10 different human livers (HepatoScreen Test Kit, Human Biologics) was studied at a concentration of 10 µM DDB. The analysis was done in duplicate, and the experiments were repeated for the conformation.

Chemical Inhibition. The chemical inhibitors used in this study were as follows: furafylline for CYP1A2 (Kunze and Trager, 1993); sulfaphenazole for CYP2C9 (Baldwin et al., 1995); quinidine for CYP2D6 (Harris et al., 1994); diethyldithiocarbamate for CYP2E1 (Guengerich et al., 1991); ketoconazole for CYP3A4; and troleandomycin for CYP3A4 (Rodrigues, 1994). Concentration of DDB in all chemical inhibition experiments was 10 µM. The reaction mixture was preincubated for 10 min at 37°C in the presence of inhibitor and an NADPH-generating system before the addition of DDB in the case of mechanism-based inhibitors. Other inhibitors were added directly to the incubation mixture at the start of each experiment. The experiments were repeated with duplicate determination.

Immunoinhibition Study. Rabbit polyclonal anti-human P450 antibodies were prepared with cDNA-expressed P450 proteins, and specificity of antibodies were confirmed by inhibiting P450 isoform-specific activities. Microsomes (1 mg of protein/ml), IgG, and phosphate buffer were combined and kept for 3 min at 37°C, followed by 15 min at room temperature. The drug (final concentration = 10 µM) was then added, and the reaction mixture was kept at for 5 min at 37°C. The reaction was started by adding an NADPH-generating system, and incubation was performed for 1 h at 37°C. The metabolism of DDB was determined by using the protocol described under Microsomal Incubations.

	Results

Top Abstract Introduction Results Discussion References

Identification and Characterization of Metabolites. DDB was converted to several metabolites in an NADPH-dependent manner in human and rat liver microsomes. The rate of metabolism of DDB ranged from 3.30 to 8.31 nmol of substrate metabolized/mg of microsomal protein/min. Five metabolites with retention times of 6.4, 9.5, 10.7, 12.7, and 15.2 min were chromatographically resolved (Fig. 1), and UV spectra of all metabolites obtained from the diode array detector were similar to that of parent DDB (data not shown). Structures of each metabolite were characterized on the basis of mass and ¹H NMR spectral analysis.

DDB. The EI spectrum of DDB showed a molecular ion at m/z 418 and fragment ions at m/z 387 [M  (-OCH₃)], 359 [M  (-COOCH₃)], 344 [359  (CH₃)], and 328 [344  (O)]. The ¹H NMR spectrum showed proton signals at 3.67 (s, 6H, -COOCH₃), 3.98 (s, 6H, -OCH₃), 5.99 (s, 4H, methylene dioxy), and 7.38 ppm (s, 2H, aromatic protons).

Metabolites M1, M2, and M5. The protonated molecular ion of M1 was observed at m/z 393 in positive ESI spectrum (Fig. 2). The EI spectrum of M1 after derivatization with methyl iodide gave rise to a molecular ion at m/z 434 and fragment ions at m/z 403, 375, and 345. Mass spectrum obtained after the derivatization with d₃-methyliodide showed a molecular ion at m/z 443 and fragment ions at m/z 409, 384, and 348, suggesting that three free hydroxyl- and/or carboxyl-functional groups exist in M1 (Fig. 3). ¹H NMR spectrum showed that one carboxymethyl and one methylene proton disappeared (Fig. 4). Therefore, M1 was identified as 2-methyl-4,4'-dimethoxy-5',6'-methylene-5,6-dihydroxyl dioxybiphenyl-2,2'-dicarboxylate. Positive ESI spectrum of M2 showed a protonated molecular ion [M + H]⁺ at m/z 407 (Fig. 2), and negative ESI spectrum demonstrated a strong molecular ion [M  H] at m/z 405 (data not shown). A molecular weight of 406 amu was assigned to M2. In the case of M5, a protonated molecular ion at m/z 375 in the positive ESI spectrum was noted (Fig. 2). In the negative ESI spectrum, a [M  H] ion at m/z 373 was obtained along with a very abundant ion at m/z 405 (data not shown). Therefore a molecular weight of 374 amu was assigned to M5. The ¹H NMR spectrum of M1 was identical to that of M5 (data not shown). EI-mass analysis of M2 and M5 after derivatization with methyl iodide showed a molecular ion at m/z 434 and fragment ions at m/z 409, 384, and 348, which was identical with the spectrum of M1. In addition, EI-mass analysis of M2 after derivatization with d₃-methyl iodide showed a molecular ion at m/z 440 and fragment ions at m/z 406 and 381, suggesting that two hydroxyl groups exist in this metabolite (Fig. 3). M5 produced the same spectra as with that of M1. Based on the molecular weight, identical EI-mass, and ¹H NMR spectra of M1 and M5, the following pathway could be deduced. Initial cleavage of the methylenedioxy moiety of DDB by P450s generated dihydrodiol metabolite (M2). Further intramolecular lactonization between the free hydroxyl group at the C6 and carboxymethyl group at the C2' resulted in the generation of M5. The opening of the lactone ring of M5 could proceed due to the ring strain factor, resulting in the generation of M1. The relative percentage of these three metabolites was dependent on the pH of reconstitution solvents. As shown in Fig. 5, M1 was marginally detected, while the intensity of M2 was similar to that of M5 in 50% methanol solution. M2 completely disappeared at pH 8.0, whereas the amount of M1 gradually increased with increasing pH.

View larger version (25K): [in this window] [in a new window]   Fig. 2.   ESI-MS spectra of DDB its and metabolites.

View larger version (30K): [in this window] [in a new window]   Fig. 3.   Gas chromatography/MS spectra of DDB, M1, M2, M3, and M4 (deuterium methyl derivatives).

View larger version (17K): [in this window] [in a new window]   Fig. 4.   Proton NMR spectra of DDB, M1, and M4.

View larger version (12K): [in this window] [in a new window]   Fig. 5.   HPLC analysis of solvent extracts of in vitro incubation mixtures after dissolution in 50% methanol (A) and Tris buffer (pH 8.0) (B). The reaction mixtures were incubated for 10 min and extracted with chloroform/isopropanol (85:15). The residues were dissolved in the above-mentioned solvents and then injected into HPLC.

Metabolite M3. The protonated molecular ion [M + H]⁺ was observed at m/z 361 in positive ESI spectrum (Fig. 2), and the [M  H] ion was obtained at m/z 359 in negative ESI spectrum, suggesting that the molecular weight of this metabolite was 360. EI mass spectrum of M3 after derivatization with methyl iodide gave rise to a molecular ion at m/z 434 and fragment ions at m/z 403, 375, and 345. Mass spectrum obtained after the derivatization with d₃-methyliodide showed a molecular ion at m/z 446 and fragment ions at m/z 429, 412, and 387, suggesting that four free hydroxyl- and/or carboxyl- functional groups exist in M3 (Fig. 3). Thus this metabolite was generated by O-demethylation of M5.

Metabolite M4. The molecular weight of M4 was 404. This was deduced from the positive ESI spectrum, which showed an abundant [M + Na]⁺ ion at m/z 427, appearing with the [M + H]⁺ ion at m/z 405 (Fig. 2). In the negative ESI spectrum a very intense [M  H] at m/z 403 was obtained. The EI spectrum of M4 after the derivatization with methyl iodide was identical to that of parent DDB. The spectrum obtained a spectra molecular ion at m/z 421 and fragment ions at m/z 390, 362, and 331, suggesting that one free hydroxyl group exists in the M4 (Fig. 3). ¹H NMR spectrum demonstrated that all signals corresponding to methylenedioxy and carboxymethyl groups were still present and the signal corresponding to methoxy protons was reduced by 50% (Fig. 4). These results collectively suggest that M4 was O-demethylated DDB.

Correlation Studies. The formation rates of individual metabolites of DDB were compared with marker activities of each P450 in a series of microsomal preparations containing varying levels of enzyme (Table 1). Strong correlation was observed between the rates of formation of M1, M2, and M5 and the testosterone 6-hydroxylation activity of each sample. The rates of formation of these metabolites were also correlated with tolbutamide hydroxylation activity. However, the formation rate of M4 was well correlated with the caffeine N3-demethylation activity. In contrast, mild correlation was noted between M3 formation and testosterone 6-hydroxylation/caffeine N3-demethylation activities.

Effect of Human P450 Inhibitors on DDB Metabolism. Effects of various P450 inhibitors on the metabolism of DDB were examined (Table 2). Troleandomycin (Pessayre et al., 1983) and ketoconazole (Maurice et al., 1992), two widely used selective inhibitors of CYP3A4, strongly inhibited the formation of M1, M2, and M5 in a concentration-dependent manner. Sulfaphenazole, a selective inhibitor of CYP2C9, also inhibited the formation of these metabolites. The formation of M3 and M4 was also inhibited by these inhibitors to a lesser degree. Furafylline, a selective inhibitor of CYP1A2, preferentially inhibited the formation of M3 and M4. However, diethyldithiocarbamate and quinidine, selective inhibitors of CYP2E1 and CYP2D6, respectively, had little effect on the metabolism of DDB in human liver microsomes.

Immunoinhibition. To determine whether CYP1A2 and CYP3A4 have a major role in the metabolism of DDB, immunoinhibition studies were carried out using the anti-CYP3A4 and anti-CYP1A2 rabbit antibodies raised against human recombinant CYP1A2 and CYP3A4. As shown in Fig. 6, the formation of M4 was inhibited only by anti-CYP1A2 antibody in a concentration-dependent manner. The formation of M3 was also inhibited by anti-CYP1A2 antibody (data not shown). In contrast, anti-CYP3A4 and anti-CYP2C9 antibodies inhibited the formation of M5, but the degree of inhibition was 30 to 50% even at 10 mg of antibody/nmol of P450. The formation of M1, M2, and M3 was also inhibited by anti-CYP3A4 and anti-CYP2C9 antibodies with a similar inhibition pattern (data not shown).

View larger version (12K): [in this window] [in a new window]   Fig. 6.   Inhibition of the formation of DDB metabolites by rabbit anti-human P450 polyclonal antibodies. Human liver microsomes (HL126) were preincubated with antibody for 15 min at room temperature, and DDB (10 µM) was added. The reaction was then initiated by the addition of an NADPH-generating system. The analysis of metabolites was performed as described under Experimental Procedures.

Metabolism of DDB by Microsomes Derived from P450 cDNA-Expressed Lymphoblastoid Cells. CYP2D6 and CYP2E1 B-lymphoblastoid microsomes did not metabolize DDB. CYP1A2 microsomes formed M4, whereas CYP3A4 microsomes generated M1, M2, and M5 (Fig. 7; see Fig. 8 for metabolic pathway of DDB and its metabolites). CYP2C9 also resulted in the formation of M3 and M5, but the rate of formation of these metabolites by CYP2C9 was approximately one-fourth of that by CYP3A4. Equimolar mixed microsomes of CYP1A2 and CYP3A4 generated all the metabolites, including M3.

View larger version (25K): [in this window] [in a new window]   Fig. 7.   Metabolism of DDB by specific P450 cDNA-transfected human B-lymphoblastoid cell microsomes. Incubations were conducted for 2 h, with microsomes containing 200 pmol of P450 (GENTEST Corp.), 50 µM DDB, and an NADPH-generating system.

View larger version (16K): [in this window] [in a new window]   Fig. 8.   Metabolic pathway of DDB in human liver microsomes.

	Discussion

Top Abstract Introduction Results Discussion References

Although DDB has been widely used in Asian countries, its metabolic profile and the enzymes responsible for its metabolism have not been fully elucidated. The present study has demonstrated the characterization of the in vitro metabolites of DDB and the identification of P450s involved in the metabolism in human liver microsomes. This characterization was invaluable in understanding the in vivo disposition of DDB and the prediction of possible interactions with coadministered drugs. The results have shown that the major in vitro metabolites are formed via two primary metabolic pathways: 1) O-demethylation of the methoxy group (M4) and 2) demethylenation of the methylenedioxyphenyl group, which results in the catechol derivative (M1). Intracellular lactonization of M1 results in the lactone derivative (M5), which is either hydrolyzed to M1 or further demethylated to the formation of M3. Catechol formation of the methylenedioxyphenyl moiety has been well documented. P450 enzymes mediate the hydroxylation of the methylene carbon, and resulting hydroxylated intermediate can undergo hydrolysis to yield the catechol metabolite. Several protease inhibitors possessing methylenedioxyphenyl moiety, methylenedioxybenzene, and methylenedioxyamphetamine have been demonstrated to be converted to corresponding catechol (Anders and Sunram, 1984; Kumagai et al., 1992; Chiba et al., 1998). Once catechol metabolite is generated, the hydroxyl group at C2 can attack the carboxymethyl group of the other ring, subsequently forming the lactone derivative in the incubation mixture. The hydrolysis product of the lactone derivative was also detected in the chromatogram. The interconversion among M1, M2, and M5 occurs depending on the environment condition. The lactone metabolite (M5) seems to be mostly derived from M3 rather than from M1. The intramolecular lactonization between the free adjacent hydroxyl group and the carboxyl group occurs more easily when the carboxyl group is methylated. It is not clear whether O-demethylation of the carboxymethyl moiety in the formation of M1 is a P450-mediated reaction. M1 can result from either the hydrolysis of M5 or from P450-mediated demethylation of M2. Rechromatography of the isolated M2 and M5 after dissolution in the buffer (pH 6-8) showed that the relative intensity of the M1 peak increased at higher pH, suggesting that M1 is mostly generated due to the hydrolysis of M5.

The P450 isoforms involved in the formation of the human liver microsomal metabolites of DDB have been identified using several complementary techniques because methylenedioxyphenyl compounds can form a metabolite complex with P450 enzymes. The initial studies with a panel of human liver microsomes showed that the formation of M1, M2, and M5 correlated with CYP3A-specific activities and with a CYP2C9-specific activity to a lesser degree. Ketoconazole and troleandomycin were found to effectively inhibit M5 formation with fairly low IC₅₀ values, further indicating the major role of CYP3A4 in this pathway. Sulfaphenazole, an inhibitor of CYP2C9, also inhibited the formation of these metabolites. But the IC₅₀ value for the inhibition of M5 by sulfaphenazole (~80 µM) is much higher than the values reported for the inhibition of the metabolism of CYP2C9-specific substrates (Newton et al., 1995; Kumar et al., 1997), suggesting that CYP2C9 probably contributes minimally to the formation of M5 and related metabolites. The immunoinhibition and cDNA-expressed cell microsome results further support the notion that CYP3A contributes substantially to the formation of M5. The major role of CYP3A in the formation of M1 and M2 is also supported by the results that M1 and M5 are nonenzymatically generated from M2.

The multidirectional approach clearly indicates that CYP1A2 is major contributor to the formation of M4. The formation of M4 in a panel of human liver microsomes correlated only with CYP1A2-specific activities. Furafylline, an inhibitor of CYP1A2, effectively inhibited the formation of M4 with a K_i value of 4.3 µM, and anti-CYP1A2 antibody almost completely attenuated this reaction. CYP1A2 and CYP3A seem to be required to form M3 in human liver microsomes. The formation of M3 was inhibited by CYP1A2- and CYP3A4-specific inhibitors and correlated with both CYP1A2- and CYP3A4- specific reactions. In the case of cDNA-expressed cell microsomes, M3 was detected only in the incubation with equimolar mixed microsomes of CYP1A2 and CYP3A4, further demonstrating that both CYP1A2 and CYP3A4 are essential in the formation of M3.

In summary, DDB is rapidly metabolized by human liver microsomes via two primary metabolic pathways, including demethylenation and O-demethylation, and by subsequently generated secondary metabolites. Demethylenation of the methylenedioxy moiety is catalyzed by CYP3A with a minor contribution of CYP2C9, whereas O-demethylation of the carboxymethyl moiety is exclusively mediated by CYP1A2.