Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Activation of the renin-angiotensin-aldosterone system is linked to high blood pressure, cardiac hypertrophy, cardiac and vascular fibrosis, renal injury, magnesium loss, baroreceptor sensitivity, ventricular arrhythmias, and increased mortality in patients with heart failure (Struthers, 1996; Sun et al., 1997; Pitt et al., 1999). Traditionally, clinicians have tended to focus on the role of angiotensin II in these negative outcomes, but increasingly strong evidence indicates that aldosterone, the central mineralocorticoid hormone in salt and blood volume homeostasis, also plays a significant and important role. As an example, the Randomized Aldactone Evaluation Study showed that, in patients with heart failure, the addition of aldosterone antagonists to standard therapies, including angiotensin-converting enzyme (ACE¹) inhibitors and diuretics, decreased mortality rates by 30% (Pitt et al., 1999).

In part, the prior emphasis on angiotensin II has arisen from the erroneous assumption that ACE inhibitors and angiotensin II receptor blockers concomitantly inhibit aldosterone production. Although this assumption is true after initial treatment with these pharmacologic agents, several studies have shown that aldosterone levels rise to pretreatment levels or higher in the majority of patients who are treated chronically with ACE inhibitors or angiotensin II receptor blockers. This phenomenon has been dubbed "aldosterone escape" (MacFadyen et al., 1977; Staessen et al., 1981; Pitt et al., 1999). Aldosterone receptor blockers are therefore expected to play increasingly important roles in therapies for hypertension and heart failure in the future.

Eplerenone (EP) is the first highly selective aldosterone receptor blocker designed to effectively block aldosterone at receptor sites in tissues (de Gasparo et al., 1987; McMahon, 2001). The stable 9,11-epoxide group in the steroid molecule of EP (Fig. 1) reduces the progestational and antiandrogenic actions associated with its nonselective predecessor, spironolactone, while maintaining its beneficial aldosterone-blocking properties (de Gasparo et al., 1987). Because of its selectivity, EP is expected to provide important clinical benefits not previously available with spironolactone. EP has been approved recently in the United States for the treatment of hypertension.

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Chemical structures of [14C]EP and its proposed primary oxidative metabolism to 6-hydroxy EP and 21-hydroxy EP. The asterisk indicates the position of the labeled carbon atom.

The present study was conducted to further define how EP is metabolized in humans and dogs, one of the animal species used in toxicity models. Specifically, the studies were designed to determine which P450 isozyme(s) are responsible for the metabolism of EP, to identify the major in vitro metabolites formed by P450 isozymes and to determine whether EP could inhibit the major P450 isoforms in vitro.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Materials. [¹⁴C]EP (lot numbers, GDS-8181-53A, GDS-8181-51A, GDS-8181-80B, and GDS-8181-28A, specific activity 16.5 or 125 µCi/mg, purity >95%), nonradiolabeled EP (lot numbers B90081 and 99K008-F3A) and 6-hydroxyeplerenone (lot number GDS-6346-096A) were obtained from Pharmacia Corporation (Skokie, IL). Male and female dog liver microsomes were obtained from In Vitro Technologies (Baltimore, MD) and pooled equally for the appropriate experiments. Pooled human liver microsomes and the Reaction Phenotyping kit were obtained from XenoTech LLC (Kansas City, KS). Microsomes prepared from cDNA-transfected human B-lymphoblastoid cell lines that stably express CYP1A2, CYP2A6, CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP2E1, CYP3A4, or CYP3A5, and several of their respective monoclonal or polyclonal antibodies, were obtained from BD Gentest Corporation (Woburn, MA). The specific CYP3A inhibitor, 6',7'-dihydroxybergamottin was synthesized by Pharmacia Corporation using the method reported by Bellevue et al. (1997). Dog CYP2D15 and CYP3A12 were cloned by Pharmacia Corporation using methods reported previously (Paulson et al., 1999). The recombinant dog P450 enzymes were expressed in S9 insect cells, commercially available from Invitrogen (Carlsbad, CA), using the bacmid methodology described by Luckow et al. (1993). [¹⁴C]-(S)-mephenytoin was obtained from Amersham Biosciences Inc. (Piscataway, NJ). All other chemicals and reagents were obtained from either Sigam-Aldrich (St. Louis, MO), ICN Biomedicals Inc. (Costa Mesa, CA), or Salford Ultrafine Chemicals (Manchester, UK), and were of the highest quality available.

Metabolism of EP Using Human and Dog Microsomes. [¹⁴C]EP, prepared by combining appropriate amounts of radiolabeled and nonradiolabeled compound in an aqueous solution of acetone (1% v/v final concentration), was incubated with human and dog liver microsomes and cDNA expressed P450 microsomes (human CYP1A2, CYP2A6, CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP2E1, CYP3A4, and CYP3A5, and dog CYP2D15 and CYP3A12). The incubation mixtures consisted of microsomal protein, 0.1 mM EDTA, 0.1 M sodium phosphate buffer (pH 7.4), 5 mM MgCl₂ · 6H₂O, 4 mM D-glucose 6-phosphate, 1.25 mM NADP+, and 2 U/ml D-glucose-6-phosphate dehydrogenase (Sigma-Aldrich, type VII from baker's yeast) in 1.5-ml capacity disposable polypropylene microcentrifuge tubes. The concentrations of human and dog microsomal proteins used ranged from 0.1 to 4.0 mg/ml, depending on the experiment. The final substrate concentration and incubation time also depended on the experiment, however, under the conditions used for kinetic experiments, metabolic rate of EP was linear with respect to substrate concentration and incubation time, as well as protein concentration. Also, preliminary experiments indicated the EP was stable in 37°C phosphate buffer for at least 2 h. Reaction mixtures were preincubated for 5 min in a 37°C shaking water bath. The reaction was initiated with the addition of [¹⁴C]EP and terminated with the addition of 200 µl of MeOH. Samples were centrifuged; the supernatants were transferred to borosilicate glass test tubes and evaporated to dryness under a stream of nitrogen (N-EVAP; Organomation Associates, Inc., Berlin, MA). Samples were reconstituted in 200 µl of an aqueous solution of acetonitrile (10%) and adjusted to a pH of 3.0 with 10 µl of HCl. All samples were analyzed by an appropriate analytical procedure.

Identification of 6- and 21-Hydroxy EP. Large-scale reactions of [¹⁴C]EP in cDNA-expressed human CYP3A4 were performed to isolate the major metabolites of EP. These samples were prepared as scaled-up versions of the microsomal incubations described above to obtain enough metabolites for identification by an LC-MS procedure. Metabolites eluting from 21.5 to 22.5 min and from 22.5 to 23.5 min on an HPLRC system were collected by fraction collection. The identity of metabolites was determined by comparing their MS fragmentation pattern with reference 6-hydroxy EP and 21-hydroxy EP that were identified previously by LC-MS and NMR (C. S. Cook, L. M. Berry, A. Karim, J. Qian, M. Lonien, J. Hribar, R. Bible, unpublished data).

Kinetics of EP Metabolism. An in vitro kinetic study of [¹⁴C]EP metabolism in human liver microsomes was performed as described previously. The 30-min reactions contained human liver microsomes (1.5 mg/ml) and EP at concentrations of 3.62, 6.03, 36.2, 60.3, 121, 181, 241, 302 µM. Extracts were analyzed for production of 6-hydroxy EP, 21-hydroxy EP, and disappearance of EP (total metabolism) using an HPLRC procedure. Mean metabolic rates were fit to an equation describing Michaelis-Menten kinetics using WinNonlin (Pharsight Corp., Mountain View, CA), and V_max and K_m values were obtained. Intrinsic clearance (CL_int) was calculated by the following equation, assuming 45 mg of microsomal protein/g liver (Houston and Carlile, 1997; Naritomi et al., 2001) and 26 g of liver per kg body weight (Davies and Morris, 1993).

Metabolism of [¹⁴C]EP Using a Reaction Phenotyping Kit. The Reaction Phenotyping kit (XenoTech, LLC) contained 16 separate preparations of human liver microsomes from individual donors. Each microsomal preparation was characterized by the manufacturer for P450 enzyme activities using specific marker reactions including the O-dealkylation of 7-ethoxyresorufin (CYP1A2), the 7-hydroxylation of coumarin (CYP2A6), the deethylation of 7-ethoxy-4-trifluoro-methylcoumarin (CYP2B6), the 6-hydroxylation of Taxol (CYP2C8), the methylhydroxylation of tolbutamide (CYP2C9), the 4'-hydroxylation of (S)-mephenytoin (CYP2C19), the O-demethylation of dextromethorphan (CYP2D6), the 6-hydroxylation of chlorzoxazone (CYP2E1), the N-demethylation of dextromethorphan (CYP3A), the 6-hydroxylation of testosterone (CYP3A), and the 12-hydroxylation of lauric acid (CYP4A9/11). The samples were prepared for incubation as described earlier (final [¹⁴C]EP concentration of 10 µM, final protein concentration of 0.5 mg/ml, 15-min incubation). To facilitate the possible metabolism of EP by P450 isozymes other than CYP3A, the experiment was repeated in the presence of the CYP3A inhibitor ketoconazole in two different ways. The first was the same as outlined above, except ketoconazole was included at a final concentration of 10 µM. The second included ketoconazole at a final concentration of 10 µM, [¹⁴C]EP at a final concentration of 50 µM, and microsomal protein at a final concentration of 1.0 mg/ml in a 30-min incubation. In these cases, the samples were preincubated for 5 min with inhibitor before adding EP, to ensure maximal inhibition of the CYP3A isoforms. Concentrations of [¹⁴C]EP and metabolites were determined using an HPLRC procedure. Correlation coefficients (r), comparing the metabolic rates of EP in each microsomal preparation to the precharacterized P450 activities (data provided by XenoTech, LLC), were determined using Microsoft Excel 2000 (Microsoft Corporation, Seattle, WA).

Inhibition of EP Metabolism by Chemical Inhibitors. The effects of P450-selective chemical inhibitors on [¹⁴C]EP metabolism with liver microsomes were evaluated by determining the substrate metabolism rate in the absence (control) or presence of the following known isoform-selective inhibitors: furafylline (CYP1A2), tranylcypromine (CYP2A6/2C19), orphenadrine (CYP2B6), sulfaphenazole (CYP2C9), quinidine (CYP2D6), dihydrocapsaicin (CYP2E1), and ketoconazole, 6',7'-dihydroxybergamottin, and troleandomycin (CYP3A). The selective chemical inhibitors used to inhibit human P450 isoforms in human liver microsomes were also used to inhibit the same subfamily of dog P450 isoforms in dog liver microsomes. Chemical inhibitors were prepared in an aqueous solution of acetone. Control incubations contained the same final concentration of acetone (1%). Pooled human and dog liver microsomal samples were prepared and incubated (final [¹⁴C]EP concentration of 50 µM) as described previously. However, inhibitors were also included for a 10-min preincubation in the presence of the NADPH regenerating system prior to adding [¹⁴C]EP to allow for the maximal inhibition of P450 isozymes. The uniform inhibitor concentration (10 µM) was chosen to be at least 3 times higher than the K_i values found in the literature. Those inhibitors that were evaluated for selectivity at various concentrations were considered adequately so at 10 µM (Gannett et al., 1990; Newton et al., 1995; Bourrie et al., 1996; Bellevue et al., 1997; Sai et al., 2000; Parkinson, 2001; Taavitsainen et al., 2001). Following incubation, concentrations of [¹⁴C]EP and metabolites were determined using an HPLRC procedure.

Inhibition of EP Metabolism by Human P450 Antibodies. Four P450-specific monoclonal antibodies (anti-CYP2A6, anti-CYP2D6, anti-CYP2E1, and anti-CYP3A4) and two polyclonal antibodies (anti-CYP1A1/2 and anti-CYP2C) were used to inhibit P450 isoforms in pooled human liver microsomes. Incubates were prepared and incubated (final [¹⁴C]EP concentration of 50 µM) as described previously, except that human liver microsomes and antibodies (1 mg of protein/ml) were preincubated for 15 min at 37°C before adding the substrate. Alternatively, control incubates contained an equivalent volume of 25 mM Tris buffer (pH of 7.4). Concentrations of [¹⁴C]EP and metabolites were determined using an HPLRC procedure.

Inhibition of Human P450 Isozyme Activities by EP. Inhibition of P450 enzymes was measured with model substrates for cDNA-derived enzymes (BD Gentest Corporation) in microsomes prepared from human lymphoblastoid cell lines (CYP1A2, CYP2C9, CYP2D6, and CYP3A4) or from baculovirus-infected insect cells (CYP2C19). A single concentration (approximately 2 times apparent K_m) of each model substrate was used, and their metabolic rates were determined by the measurement of an appropriate metabolite. Each enzyme assay included multiple EP concentrations separated by approximately 1/2 log (0, 0.01, 0.03, 0.1, 0.3, 1, 3, 10, 30, 100, and 300 µM) and two concentrations of a positive control inhibitor specific to the enzyme in question. Acetonitrile (final concentration 1%) was used as the substrate and inhibitor cosolvent. For this experiment, each enzyme assay required distinctive incubation methods. These conditions were provided by BD Gentest Corporation and used with minor modifications. Briefly, the incubation conditions used for each enzyme assay were as follows:

HPLRC. An aliquot (200 µl) of sample was injected into an HPLRC system comprised of a Hewlett Packard series 1050 autosampler and pump (Hewlett Packard Analytical Direct, Wilmington, DE), equipped with an Alltech Altima C₁₈ guard column (5 µ, 7.5 × 4.6 mm) and an Alltech Alltima C₁₈ HPLC column (5 µ, 150 × 4.6 mm) (Alltech Associates, Deerfield, IL). A linear gradient system was used consisting of acetonitrile and water, from 10:90 by volume to 50:50 over 40 min, at a flow rate of 1.0 ml/min. The column was re-equilibrated with mobile phase A for 15 min between injections. A Radiomatic Flo-One/eta series A-500 radiochemical detector (PerkinElmer Life Sciences, Boston, MA), pumping Packard Flo-Scint III at 3 ml/min, detected the radiochromatographic profiles. Assay sensitivity of the system allowed for the detection of any radioactive metabolic peak areas greater than 1% of the total sample radioactivity.

LC-MS. Identification of the hydroxy metabolites of EP was performed on a Finnigan TSQ-700 triple quadrupole LC-MS system (Thermo Finnigan, San Jose, CA) equipped with an atmospheric pressure chemical ionization unit. The corona current was set to 5 µA, the vaporizer temperature to 500°C, the capillary temperature to 200°C, the sheath gas to 80 psi, and the auxiliary gas to 10 ml/min. Samples were scanned at a mass range of 70 to 750 amu every 2 s. Samples were injected onto a 3.5 µ, 3 × 150 mm Zorbax SB-C₁₈ LC column (MAC-MOD Analytical Inc., Chadds Ford, PA) at 50°C. The mobile phase gradient consisted of a 40-min ramp from 100% mobile phase A (95% H₂O, 5% MeOH, 12.5 mM ammonium acetate, pH 4.5) to 100% mobile phase B (5% H₂O, 95% MeOH, 12.5 mM ammonium acetate, pH 4.5).

Determination of Metabolites of P450 Marker Substrates. Methodologies for determination of P450 marker substrate metabolites were provided by BD Gentest Corporation and used with minor modifications. Breifly, sample aliquots were injected into an HPLC system using a Waters model 510 pump and a Waters model 717 plus autosampler (Waters, Milford, MA). Separation occurred at 45°C and 1 ml/min on a Supelco 4.6 × 250 mm Nucleosil C₁₈, 5 µ HPLC column (Bellefonte, PA), with linear mobile phase gradients. Differing mobile phases, gradients of water and methanol, or water, acetonitrile, and perchlorate were used, depending on the analyte. A Waters model 486 tunable absorbance detector monitored for the metabolites acetaminophen (absorbance at 244 nm), 4'-hydroxydichlofenac (absorbance at 280 nm), and 6-hydroxytestosterone (absorbance at 254 nm). A Packard model 150TR Radiomatic Flow-One flow scintillation analyzer (PerkinElmer Life Sciences) was used to detect [¹⁴C]4'-hydroxymephenytoin. A Foci System 3 scanning spectrofluorometer (Optical Technology Devices, Inc., Elmsford, NY) monitored for 1'-hydroxybufuralol (excitation at 252 nm and emission at 302 nm). When using absorbance or fluorescence methods, analytes were quantified by comparison to an appropriate standard curve.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Metabolism of EP In Liver Microsomes and cDNA-Expressed P450 Isozymes. Metabolic profiles of [¹⁴C]EP after incubation with human and dog liver microsomes are shown in Fig. 2, A and B. The metabolite peaks eluting at the HPLC retention time of 22 to 22.5 min and 23 to 23.5 min were identified as 6-hydroxy EP and 21-hydroxy EP, respectively, as described below. The major metabolite of [¹⁴C]EP produced with human microsomes was 6-hydroxy EP, and a relatively small amount of 21-hydroxy EP was also found. With dog microsomes, the major metabolites were also 6-hydroxy and 21-hydroxy EP.

View larger version (24K): [in this window] [in a new window]   Fig. 2.   Representative HPLC profile of [14C]EP (48 µM, 1 h incubation) metabolism by human liver microsomes (A), dog liver microsomes (B), human cDNA-expressed CYP3A4 (C), and dog cDNA-expressed CYP3A12 (D). M1, 6-hydroxy EP; M2, 21-hydroxy EP.

View larger version (23K): [in this window] [in a new window]   Fig. 3.   MS of 6-hydroxy EP (A) and 21-hydroxy EP (B) isolated from cDNA-expressed CYP3A4 incubates.

Kinetics of EP Metabolism. The V_max and K_m (±S.E.) measured from 6-hydroxylation and 21-hydroxylation were 0.973 ± 0.132 nmol/min/mg and 217 ± 56.0 µM, and 0.143 ± 0.035 nmol/min/mg and 211 ± 101 µM, respectively (Fig. 4). The corresponding values, as measured from the disappearance of EP (total metabolism), were 1.76 ± 0.24 nmol/min/mg and 303 ± 70.0 µM. Even though enough curvature was present in the data to obtain estimated V_max and K_m values using WinNonlin, solubility of EP was the major limitation in determining metabolic rates at concentrations higher than 300 µM, thereby causing a relatively large standard error in their estimation. The CL_int for 6-hydroxylation, 21-hydroxylation, and total metabolism of EP were 5.24, 0.79, and 6.78 ml/min/kg, respectively, when calculated from the V_max and K_m values. Using CL_int and hepatic blood flow, the estimated CL_H of EP due to 6-hydroxylation, 21-hydroxylation, and total metabolism were 1.84, 0.31, and 2.30 ml/min/kg. In this respect, CL_H measured from 6-hydroxy and 21-hydroxy metabolites alone accounted for 93.4% of total EP CL_H as estimated using human liver microsomes.

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Rate of 6-hydroxylation, 21-hydroxylation, and total metabolism in a pool of human liver microsomes. Error bars indicate standard error of the mean for an n of 3.

Correlation with P450 Isozyme Activities. The metabolic rate of [¹⁴C]EP was correlated with specific P450 isozyme activities from human liver microsomes in the Reaction Phenotyping kit (Table 1). The total rate of [¹⁴C]EP metabolism was best correlated with the dextromethorphan N-demethylation (CYP3A) and testosterone 6-hydroxylation (CYP3A) rates compared with rates of other marker metabolic reactions. When metabolism of EP to the two major metabolites was compared with P450 isozyme activities, the formation rates of both 6-hydroxy and 21-hydroxy EP were best correlated with CYP3A activity. In this experiment, [¹⁴C]EP also seemed to correlate moderately (an r value of approximately 0.5) with coumarin 7-hydroxylation (CYP2A6), 7-ethoxy-4-trifluoromethylcoumarin O-dealkylation (CYP2B6) and paclitaxel 6-hydroxylation (CYP2C8). To verify whether these isoforms play a minor role in EP metabolism and whether the correlation improves by inhibition of CYP3A activities, the phenotyping experiment was repeated in the presence of ketoconazole. With ketoconazole, 10 µM EP metabolism was completely inhibited, and metabolites were not measurable in any of the microsomal preparations. Therefore, correlation values could not be determined. Using the modified conditions described earlier (i.e., using 50 µM EP, 1 mg protein/ml and 30-min incubation), moderate metabolism was observed in the presence of ketoconazole. In this case, the metabolic rates tended to be reduced by 90% when normalized for initial substrate concentration, and both 6-hydroxy and 21-hydroxy metabolites were observed. However, the correlation values calculated for 6-hydroxylation, 21-hydroxylation and total metabolism remained the same as in the experiment without ketoconazole. That is, correlations with CYP3A activity remained greater than 0.9, whereas correlations with CYP2A6, CYP2B6, and CYP2C8 activities remained approximately 0.5. 

Chemical Inhibition. Inhibition of metabolism of [¹⁴C]EP in human liver microsomes was examined in the presence of various chemical inhibitors at a concentration of 10 µM. Total metabolism was markedly reduced with CYP3A inhibitors ketoconazole (95%), troleandomycin (83%), and 6',7'-dihydroxybergamottin (86%). When formation of 6-hydroxy EP and 21-hydroxy EP was measured individually, metabolism of [¹⁴C]EP to these metabolites was significantly inhibited by ketoconazole, troleandomycin, and 6',7'-dihydroxybergamottin (Fig. 5, A and B). Formation of 6-hydroxy EP was also significantly reduced by dihydrocapsaicin (31%) compared with control.

View larger version (40K): [in this window] [in a new window]   Fig. 5.   Inhibition of 6-hydroxy EP (A) and 21-hydroxy EP (B) formation in human liver microsomes and 6-hydroxy EP (C) and 21-hydroxy EP (D) formation in dog liver microsomes by P450 chemical inhibitors at 10 µM. Data are expressed as a percent of control values. CON, control; FUR, furafylline; TRA, tranylcypromine; ORP, orphenadrine; SUL, sulfaphenazole; QUI, quinidine; DIH, dihydrocapsaicin; KET, ketoconazole; DHB, 6',7'-dihydroxybergamottin; TAO, troleandomycin. Error bars indicate standard error of the mean for an n of 3. , p < 0.05; , p < 0.01; , p < 0.001.

Immunoinhibition of 6- and 21-Hydroxy EP Formation. Inhibition of [¹⁴C]EP metabolism in human microsomes was examined with human P450 antibodies. The total metabolism of [¹⁴C]EP was inhibited by 84% with a monoclonal anti-CYP3A4 antibody. Inhibition of total metabolism of EP was minimal when human liver microsomes were incubated with other antibodies. When formation of 6-hydroxy and 21-hydroxy EP was measured individually, metabolism of [¹⁴C]EP to these two metabolites was also significantly inhibited only by anti-CYP3A4 (Fig. 6).

View larger version (48K): [in this window] [in a new window]   Fig. 6.   Inhibition of EP metabolism to 6-hydroxy EP and 21-hydroxy EP in human liver microsomes by P450 antibodies. Data are expressed as a percent of control values. Error bars indicate standard error of the mean for an n of 3. , p < 0.05; , p < 0.01; , p < 0.001.

Inhibition of P450 Isozyme Activities by EP. As indicated by metabolism of the marker substrates, EP at concentrations up to 300 µM did not remarkably reduce (<20% reduction) the activities of CYP1A2, CYP2C9, CYP2C19, or CYP2D6 (Fig. 7). Positive control inhibitors for these enzymes, however, did markedly reduce metabolism of marker substrates, indicating proper function of the assay. On the other hand, EP inhibited CYP3A4 catalytic activity by approximately 34% at 100 µM and approximately 45% at 300 µM. The IC₅₀ for this interaction was not calculated because all inhibition values were below 50%. CYP3A activity was inhibited more than 75% by 0.1 µM and nearly 95% by 1.0 µM ketoconazole, the positive control inhibitor.

View larger version (30K): [in this window] [in a new window]   Fig. 7.   Inhibition of metabolism of marker substrates for CYP1A2 (PHE, phenacetin), CYP2C9 (DIC, diclofenec), CYP2C19 (MEP, (S)-mephenytoin), CYP2D6 (BUF, bufuralol), and CYP3A4 (TES, testosterone) by EP at concentrations of 30, 100, and 300 µM (A) and by isozyme-specific positive control inhibitors (B) at the low and high concentrations. BEN, 7,8-benzoflavone for CYP1A2; SUL, sulfaphenazole for CYP2C9; TRA, tranylcypromine for CYP2C19; QUI, quinidine for CYP2D6; KET, ketoconazole for CYP3A4. The low and high concentrations are 0.3 and 3.0 µM for BEN, SUL and TRA, respectively, and 0.1 and 1.0 µM for QUI and KET, respectively. Values are means of n = 2.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The present study demonstrated that 6- and 21-hydroxy EP are the major in vitro metabolites in both human and dog and that the metabolism to these hydroxy compounds was primarily mediated by CYP3A4 in humans and CYP3A12 in dogs. The involvement of CYP3A4 in the metabolism of EP in humans is well supported by the following findings: 1) metabolism of [¹⁴C]EP to 6- and 21-hydroxy EP was best correlated with dextromethorphan N-demethylation and testosterone 6-hydroxylation activities, which are typical markers for CYP3A activity; 2) correlation with other P450 marker reactions did not improve in the presence of CYP3A-selective inhibitor ketoconazole; 3) metabolism to both 6- and 21-hydroxy EP was significantly inhibited by only CYP3A-selective inhibitors ketoconazole, troleandomycin, and 6',7'-dihydroxybergamottin; 4) metabolic activity was extensively inhibited by human anti-CYP3A4 antibody, but not by other antibodies; 5) EP was extensively metabolized to 6- and 21-hydroxy [¹⁴C]EP by human cDNA-expressed CYP3A4 and only minimally by CYP3A5, but not by other CYP isozymes. With dog liver microsomes, metabolism of [¹⁴C]EP to both 6- and 21-hydroxy EP, was inhibited by CYP3A specific inhibitors, ketoconazole, troleandomycin, and 6',7'-dihydroxybergamottin, whereas the metabolism was not inhibited by the inhibitors for other P450 isozymes. Furthermore, metabolism of [¹⁴C]EP was observed with dog cDNA-expressed CYP3A12, and there was no notable metabolism with dog CYP2D15. These results indicate that formation of 6- and 21-hydroxy [¹⁴C]EP in the dog was also primarily mediated by CYP3A.

The chemical inhibitors used in this study were chosen based on a review of literature that ascertained their potency and selectivity. Nevertheless, in P450-selective chemical inhibition experiments with human liver microsomes, dihydrocapsaicin, a selective CYP2E1 inhibitor, reduced metabolism of [¹⁴C]EP significantly at a concentration of 10 µM (Fig. 5). However, there was no apparent metabolism with cDNA-expressed CYP2E1, EP metabolism (alone or in the presence of ketoconazole) was not well correlated with CYP2E1 activities (Table 1), nor was EP metabolism inhibited by anti-CYP2E1 (Fig. 3). Therefore, the minor reduction of EP metabolism by dihydrocapsaicin may be nonselective at the concentrations studied.

Examining EP metabolism using the Reaction Phenotyping kit of human microsomes revealed some moderate correlations with CYP2A6, CYP2B6, and CYP2C8 activities (r values of approximately 0.5), in addition to higher correlations with CYP3A activities. These correlations were not improved in the presence of ketoconazole. Thus, these P450 isozymes were not involved in EP metabolism in vitro. Instead, these correlations appear to be artifacts of the microsomal samples chosen by XenoTech for this particular version of the Reaction Phenotyping kit. Using the marker reaction data alone, XenoTech determined that a selective CYP3A4/5 substrate reaction such as testosterone 6-hydroxylation would have an apparent correlation of 0.46, 0.55 to 0.58, and 0.57 to marker reactions for CYP2A6, CYP2B6, and CYP2C8 activities, respectively (data provided by Xenotech). If EP were a substrate of one of these isoforms, a correlation value of significantly higher than 0.5 to 0.6 is expected, particularly in the presence of ketoconazole. This demonstrates the importance of evaluating the correlations between the marker metabolic reactions prior to final interpretation of the results from this type of phenotyping experiment.

A CL_H value of 2.3 ml/min/kg was obtained from the total rate of EP metabolism using the well stirred model. The plasma clearance of EP after oral administration (CL/F) to humans was 2.7 ml/min/kg (C. S. Cook, L. M. Berry, A. Karim, J. Qian, M. Lonien, J. Hribar, R. Bible, unpublished data). Since F (fraction bioavailable) is expected to be greater than 0.7, and EP is extensively metabolized in humans, the in vivo hepatic clearance will be between 1.9 and 2.7 ml/min/kg. Thus, despite the solubility limitations in determining V_max and K_m, the CL_H predicted by the in vitro values is in close agreement with in vivo data.

Steroid molecules such as testosterone and progesterone are metabolized to 6-hydroxy metabolites by CYP3A (Yamazaki and Shimada, 1997), and it is not surprising to find that 6-hydroxylation of EP is also catalyzed by CYP3A4. However, 6-hydroxylation of testosterone and progesterone was also mediated by CYP2C9 and 2C19 although the rates of metabolism by these isozymes were much lower than those by CYP3A. Unlike these steroid molecules, there was no evidence that 6-hydroxy EP was formed by these isozymes at the assay sensitivity level of 0.08 pmol/min/mg of protein. Moreover, EP was not substantially metabolized in the presence of cDNA-expressed CYP3A5, demonstrating that EP metabolism is highly selective to CYP3A4. Some steroid molecules, such as testosterone and estradiol, are good substrates of both CYP3A5 and CYP3A4, although the rates of metabolism of these compounds tended to be lower with CYP3A5 (Williams et al., 2002). In addition to 6-hydroxylation, EP was also metabolized to 21-hydroxy EP, which has a hydroxyl group on the -lactone ring. To our knowledge, EP is the first example that demonstrates involvement of CYP3A4 in the hydroxylation of a -lactone ring. Hydroxylation of a -lactone ring has been shown for canrenone (Karim, 1978) and the naphthalenic lignan lactone 5-lipoxygenase inhibitor, L-702539 (Chauret et. al., 1995). However, the P450 isozyme responsible for this type of metabolism has not been reported.

Even though the in vitro metabolites, 6-hydroxy EP and 21-hydroxy EP, were formed by CYP3A subfamily enzymes in dogs, as observed in humans, the proportion of these two metabolites in dogs was substantially different from that in human. The major in vitro metabolite of EP formed in human liver microsomes was 6-hydroxy EP, whereas only a small amount of 21-hydroxy was formed. In dog liver microsomes, on the other hand, there was relatively more 21-hydroxy EP formed than 6-hydroxy EP. These results are in agreement with in vivo findings in which 6-hydroxy EP is the major metabolite in humans (C. S. Cook, L. M. Berry, A. Karim, J. Qian, M. Lonien, J. Hribar, R. Bible, unpublished data), whereas in dogs, 21-hydroxy EP is the major metabolite (Cook et al., 2000). CYP3A4 and CYP3A12 share 79% amino acid sequence similarity. However, the active sites of these two isozymes appeared to be substantially different. This was well demonstrated by the finding that the human anti-CYP3A4 antibody did not inhibit CYP3A12 activity, as measured by N-dealkylation of disopyramide (Zhang et al., 2001)

Although EP is selectively metabolized by CYP3A4, it is highly unlikely that EP will significantly reduce the metabolism of other drugs catalyzed by CYP3A4, or by other P450 isozymes, due to its high IC₅₀ values (>300 µM) against these P450 isozymes. However, potent CYP3A inhibitors like ketoconazole, 6',7'-dihydroxybergamottin, and troleandomycin significantly reduced EP metabolism. Therefore, it is anticipated that other potent CYP3A4 inhibitors (e.g., erythromycin, fluconazole, cyclosporin, saquinavir) will also reduce the metabolism of EP, thereby potentially increasing in vivo exposure to EP.

In summary, the results of the present study indicate that EP was primarily metabolized to 6- and 21-hydroxy metabolites by CYP3A4 in humans and by CYP3A12 in dogs. Additionally, it is highly unlikely that EP will reduce the metabolism of other drugs, whereas only very potent CYP3A inhibitors are expected to substantially reduce the metabolism of EP in vivo.