Experimental Procedures

Chemicals.

Almokalant [(RS)-4-[3-ethyl[3-propylsulphinyl)-propyl]amino]- 2-hydroxypropoxybenzonitrile] and its metabolites were synthesized at AstraZeneca R&D, Mölndal, Sweden (Astra Hässle R&D). The CYP2C19-specific substrate S-mephenytoin was a kind gift from Sandoz Pharma (Basel, Switzerland). Hydroxytolbutamide, 4′-hydroxymephenytoin, dextrorphan, and 6-hydroxychlorzoxazone were purchased from Ultrafine Chemicals (Manchester, UK). [4-¹⁴C]Testosterone (specific activity, 59 mCi/mmol) was purchased from Amersham Pharmacia Biotech (Bucks, UK). Reference testosterone metabolites (6β-, 7α-, and 16α-hydroxytestosterone) were from the Steroid Reference Collection (Prof. D. N. Kirk, London, UK). Roche Molecular Biochemicals (Welwyn Garden City, UK) kindly provided debrisoquine and 4-hydroxydebresoquine sulfate. Other chemicals were obtained mainly from Sigma Chemical Company (St. Louis, MO) and Boehringer Ingelheim (Ingelheim, Germany) and were of the highest purity available.

Tissue.

The respective local ethical committees approved the collection and use of all human tissues in this study.

Analysis of Almokalant and Its Metabolites.

Two different reversed-phase HPLC systems combined with UV detection or mass spectrometry were used for measuring unchanged almokalant and its metabolites formed in vitro. Method 1 was used for analysis of microsomal incubations fortified with NADPH. Method 2 was used for measuring microsomes fortified with UDPGA, liver slices, and primary human hepatocytes. Metabolites formed by cDNA-expressed enzymes were quantified by liquid chromatography-mass spectrometry. The identity of metabolites produced in the various in vitro systems was verified by liquid chromatography-mass spectrometry in comparison with synthetic references.

HPLC analysis.

The HPLC system used for UV measurements consisted of a Pharmacia LKB HPLC Pump 2248 (Bromma, Sweden), a Spectra Focus UV detector (Spectra Physics, San José, CA) operated at 240 nm, and an automatic sample injection system (ASPEC XL, Gilson, Villiers-le-Bel, France) operated in a partial loop mode. At a flow rate of 1 ml/min, the solvent gradient program was formed by a LKB LC Controller 2252 equipped with a low-pressure mixer. The analytical columns were a 5-μm Zorbax SB C₁₈ column (15 cm × 4.6-mm i.d., Eka Nobel, Surte, Sweden, method 1) and a 5-μm HiChrom C₁₈ column (15 cm × 4.6-mm i.d., Scantec Lab, Partille, Sweden, method 2). Both columns were protected by a cyano Brownlee guard column (7 μm, 15 × 3.2 mm) from Applied Biosystems (San José, CA). The chromatograms were stored and reprocessed by means of peak area calculation using a computerized data integration program (Spectra Physics, Spectra system software, PC 1000).

Method 1 (quantifying almokalant and phase I metabolites in microsomal incubations fortified with NADPH).

The mobile phase A consisted of 10% methanol and 10% acetonitrile in a formic acid buffer (12 mM, pH 3.5). The corresponding composition of mobile phase B was 10% methanol and 60% acetonitrile in formic acid. The gradient elution profile was 0 to 50% B for 0 to 10 min and 50 to 0% B for 10 to 11 min. Thereafter the system was equilibrated for 14 min before injection of the next sample.

Method 2 (quantifying almokalant and metabolites in incubations using tissue slices, hepatocytes, and microsomes fortified with UDPGA).

A 10 mM ammonium acetate buffer at pH 7 was used as phase A with 50% acetonitrile as phase B. The gradient program was 27% B for 0 to 12 min, 27 to 100% B for 12 to 17 min, and 100% B for 17 to 20 min. The gradient was reversed over 1 min and equilibrated for 19 min before injection of the next sample. The biological samples were centrifuged for 10 min at 10,000g, and 50 μl of the clear supernatant was injected in duplicate. Standard solutions of almokalant and the metabolites were prepared and diluted in mobile phase A to a final concentration of 100, 50, 10, 5, 1, 0.5, 0.25, and 0.1 μM, respectively. Each standard solution was analyzed in triplicate by methods 1 and 2, and the calibration curves were obtained by plotting the mean peak area against the known concentration of the analytes. Ther² values of the resulting standard curves were 0.98, and the limit of quantification was set to 100 nM.

Mass spectrometry analysis (verifying metabolites in the various in vitro systems and quantifying products formed in incubations using cDNA-expressed P450 enzymes).

For the analysis by mass spectrometry, the liquid chromatographic system (Series 1050, Hewlett Packard, Agilent Technologies, Waldbronn, Germany) consisted of an autosampler and two HPLC pumps operating at a flow rate of 1 ml/min. About 30% of the eluent was diverted to the triple stage mass spectrometer (Finnigan TSQ7000, ThermoFinnigan, San José, CA) equipped with an electrospray interface. The spray voltage in the electrospray interface was kept at 4.5 kV, resulting in an ion spray of about 2.6 μamps. The temperature of the capillary was 200°C, and the voltage of the capillary and of the tube lenses was maintained at 11.7 and 55.2 V, respectively. Positive ions were recorded in a scanning mode of the first quadrupole at unit mass resolution. The mass range was betweenm/z 200 to 600, and the scan speed was 2 s/spectrum. The mass spectra were stored in the computer system, which controlled both the HPLC and the TSQ7000. The analytical column and mobile phases were the same as given above for method 1. The gradient elution profile was 0 to 70% B during 0 to 10 min. The gradient was reversed over 1 min, and the system was equilibrated for at least 5 min before injection of the next sample. The computer-reconstructed ion current of the protonated molecular ions of almokalant (m/z 353), M3 (m/z 221), M13 (m/z 325), M16 (m/z369), and M18a/b (m/z 529) was used to determine the peak area automatically, according to preset parameters. The single point calibration method with external standardization was applied where the target concentration of almokalant in the sample at time 0 was taken. These standard samples were analyzed in triplicate, and the mean value was taken to calculate the concentrations of the compounds.

Modeling.

Molecular modeling of mammalian microsomal P450s from the CYP102 bacterial crystal structure template has been carried out for all of the major P450 families associated with the phase I metabolism of drugs and other foreign compounds (Lewis, 1996; Lewis et al., 1999). The current program has focused on human P450s from families CYP1, CYP2, and CYP3, which represent the major catalysts of oxidative drug metabolism in humans. Three-dimensional models of 10 human P450s have been constructed from the CYP102 hemoprotein domain, for which the X-ray coordinates are known both in the substrate-bound and substrate-free states. The methodologies used in homology modeling of P450 from CYP102 are described in detail elsewhere, including a general reference to the rationale for using the CYP102 structure as a preferred modeling template.

A number of potential probe substrates (25 total) for various human P450 enzymes have been identified and tested in models for CYPlA1, CYP1A2, CYP2A6, CYP2C9, CYP2C19, CYP2D6, CYP2E1, and CYP3A4. After this preliminary work, almokalant was tested in the models.

cDNA-Expressed P450 Enzymes.

The following human P450 enzymes were produced in yeast: CYP1A1, CYP1A2, CYP2A6, CYP2B6, CYP2C9, CYP2C19, CYP2D6, CYP2E1, and CYP3A4. The expression levels were relatively high, 30 to 300 pmol/mg of microsomal protein. All microsomes produced good difference spectra showing a homogenous peak at about 450 nm in the reduced carbon monoxide-bound state of the hemoprotein with no signs of P420. To achieve good expression of the CYP2C19 and CYP2D6 enzymes, the plasmid cDNAs were modified to include a triplet of adenine bases just before the initiating ATG codon (Krynetski et al., 1995).

Expression was carried out with the Saccharomyces cerevisiaestrain W(R), which has been genetically modified to also overexpress the yeast reductase (Truan et al., 1993). A galactose-inducible promoter in the plasmid and in the yeast genome, respectively, was used to control the expression. After transformation of the plasmid into the W(R) yeast strain, using lithium acetate selection of clones was achieved by growing the yeast in adenine- and uracil-deficient media. To achieve higher expression levels, the yeast were first grown to a high density with glucose as the main energy source; thereafter the expression of the plasmid was initiated by the addition of galactose. Preparation of microsomes was as described in Oscarson et al. (1997), with the exception that the microsomes were collected in the last step by ultracentrifugation. P450 and P450 reductase were measured as described (Oscarson et al., 1997). Control microsomes were prepared from yeast transformed with the V60 plasmid without insert. The catalytic properties of the yeast microsomes were evaluated using probe substrates and were found to be active with kinetics for the marker substrates (K_m values are expressed in micromolar, V_max in nanomoles per nanomole per minute) as follows: ethoxyresorufin-O-deethylation, CYP1A1, K_m = 1.4,V_max = 16; CYP1A2,K_m = 1.5, V_max= 1.5; coumarin 7-hydroxylation, CYP2A6, K_m= 0.5, V_max = 16; ethoxycoumarin-O-deethylation, CYP2B6,K_m = 40, V_max = 40; diclofenac-4-hydroxylation, CYP2C9, K_m= 6, V_max = 30; bufuralol-1-hydroxylation,K_m = 3, V_max = 30; chlorzoxazone-6-hydroxylation, CYP2E1,K_m = 210, V_max= 1.9; carbamazepine 10,11-epoxidation, CYP3A4,K_m = 250, V_max= 4.7.

Incubations of yeast microsomes with 100 μM almokalant were performed using 400 μg of microsomal protein corresponding to 10 to 100 pmol of P450 depending on the P450 form in question. The reaction was carried out in 0.1 M potassium phosphate buffer, pH 7.4, and was started after 3 min of preincubation at 37°C by adding 10 μl of NADPH, resulting in a final concentration of 1 mM. After 20 min, the reaction was terminated by rapid cooling in liquid nitrogen. Thereafter the samples were stored at −70°C until analysis. Almokalant incubations with CYP3A4 were also carried out with additional human cytochromeb₅. In this case the microsomes were mixed at a 1:1 ratio of P450 to cytochrome b₅ and put on ice for 10 min before the incubation procedure was initiated as above.

NADPH-Dependent Liver Microsomal Metabolism.

Human liver samples were obtained from kidney transplantation donors. Macroscopically normal tissue was homogenized in 4 volumes of ice-cold phosphate buffer, pH 7.4, and microsomes were separated by differential centrifugation. The final microsomal pellet was suspended in 0.1 M phosphate buffer. The liver preparations were thoroughly characterized with respect to the presence and properties of the major P450 enzymes used for the pharmaceutical screening.

P450 Inhibition and Correlation Studies.

Correlation studies were performed using a panel of human hepatic microsomal samples that had previously been characterized for their content and activity of all of the major drug-metabolizing forms of P450 (CYP1A2, CY1B1, CYP2A6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP2E1, CYP3A4, CYP3A5, and CYP4A11), using form-specific antibodies and Western immunoblotting and appropriate probe substrates. The characterization of these samples has been described in detail previously (Edwards et al., 1998). The formation rates of the oxidative metabolites of almokalant (M3, M13, and M16) by this panel of hepatic samples were determined by incubating microsomal protein (1.0 mg) with almokalant (20 μM) and NADPH (1.2 mM) in phosphate buffer, pH 7.4, for 60 min at 37°C. The conditions of the assay were previously shown to be linear with respect to time of incubation and protein concentration.

In P450 inhibition studies, the following enzyme assays, which display at least some enzyme specificity, were used: ethoxyresorufinO-deethylation (CYP1A1/2) (Burke et al., 1977); phenacetinO-deethylation (CYP1A2) (Murray and Boobis, 1986), adapted for HPLC using [¹⁴C]phenacetin as substrate; coumarin 7-hydroxylation (CYP2A6) (Aitio, 1978), with slight modifications as described by Raunio et al. (1988, 1990); tolbutamide (methyl) hydroxylation (CYP2C9) (modified from Knodell et al., 1987 andSullivan-Klose et al., 1996); mephenytoin 4′-hydroxylation (CYP2C19) (Wrighton et al., 1993); dextromethorphanO-demethylation (CYP2D6) (modified from Park et al., 1984and Kronbach et al., 1987); debrisoquine 4-hydroxylation (CYP2D6) (Khan et al., 1982); chlorzoxazone 6-hydroxylation (CYP2E1) (Peter et al., 1990); and testosterone 6β-hydroxylation (CYP3A4/5) (Waxman et al., 1991). The incubation conditions were similar for the different reactions. If not otherwise stated in the next section, the analytical method was applied according to the references cited.

Determination of the Metabolites in P450-Specific Assays.

Metabolites of ethoxyresorufin O-deethylase and coumarin 7-hydroxylase assays were measured fluorometrically. For the HPLC method of the metabolites of chlorzoxazone, dextromethorphan, and mephenytoin a Superspher 100 RP-18 column (125-4, Merck, Darmstadt, Germany) with a guard column of the same matrix was used. For the HPLC method for paracetamol (from phenacetin), off-line radiometric detection following peak location by UV absorbance at 254 nm, with a spherisorb C₁₈ column (Waters Associates, Milford, MA) was used. For the determination of hydroxytolbutamide, a LiChrospher 100 RP-18 column with guard column (125-4, Merck) was used. For debrisoquine, gas chromatography-mass spectrometry was used as reported by Kahn et al. (1982).

In Vitro Inhibition of P450-Selective Activities by Almokalant and Reference Inhibitors.

To determine IC₅₀ values, almokalant was added at a wide range of concentrations (final concentrations: 0.1, 1, 10, 100, and 1000 μM) into the incubation mixture in a small volume of an appropriate solvent, and activity was compared with that of control incubations into which only solvent was added. The reference inhibitors used were furafylline (preincubated for 10 min) for CYP1A2, methoxsalen for CYP2A6, sulfaphenazole for CYP2C9, omeprazole for CYP2C19, quinidine for CYP2D6, pyridine for CYP2E1, and ketoconazole for CYP3A4. The specificity of these inhibitors with respect to the P450-associated activities used here has been partially characterized (see Pelkonen et al., 1998). Inhibition was determined independently in two different laboratories.

In Vitro Inhibition of Almokalant Metabolism by P450-Selective Inhibitors.

The P450-selective inhibitors furafylline (1, 2, and 5 μM), coumarin (10, 20, and 30 μM), quercetin (3, 10, and 30 μM), sulfaphenazole (3, 10, and 30 μM), S-mephenytoin (100, 200, and 300 μM), quinidine (1, 5, and 20 μM), pyridine (1, 5, and 20 μM), and ketoconazole (1, 5, and 20 μM) were incubated with almokalant (20 μM) for 60 min at 37°C. The activity was compared with that of control incubations into which only solvent was added.

Hepatic Microsomal UDPGA-Dependent Metabolism.

Liver microsomes (pooled microsomes from five subjects) were prepared according to Ernster et al. (1962) from tissue (excess material removed during surgery) frozen in liquid nitrogen in small cubes (1 cm³) and stored at −70°C until preparation. Microsomal protein concentration was measured according to the method of Markwell et al. (1978). Microsomes were activated by preincubation with Brij (0.17 mg/mg of microsomal protein on ice for 20 min). Samples were then incubated with almokalant for 45 min at 37°C. The incubation mixture (total volume of 300 μl) contained activated microsomes (1 mg/ml), 0.1 mM Tris-maleate buffer (pH 7.4), 3 mM UDPGA, 5 mM MgCl₂, and almokalant (10–10,000 μM). The reaction was terminated by adding 200 μl of methanol. The incubation conditions were optimized to give linear formation of conjugates with time and protein concentration.

Studies Using Human Hepatocytes from Liver Biopsies.

Surgical liver biopsies (weighting 1–3 g) were obtained from patients undergoing cholecystectomy after informed consent was obtained. Patients had no known liver pathology, nor did they receive medication during the weeks before surgery. None of the patients were habitual consumers of alcohol or other drugs. A total of six liver biopsies (two males and four females) were used. Patients' ages ranged from 35 to 77 years. Hepatocytes were isolated using a two-step tissue microperfusion technique as described elsewhere (Gómez-Lechón et al., 1997). Cellular viability, estimated by the dye exclusion test with 0.4% trypan blue in saline, was higher than 90%. Hepatocytes were seeded on fibronectin-coated plastic dishes (3.5 μg/cm²) at a density of 8 × 10⁴ viable cells/cm² and cultured in Ham's F-12/Lebovitz L-15 (1:1) medium supplemented with 2% newborn calf serum, 10 mM glucose, 50 mU/ml penicillin, 50 μg of streptomycin/ml, 0.2% bovine serum albumin, and 10 nM insulin. One hour later the medium was changed, and after 24 h, cells were shifted to serum-free, hormone-supplemented medium (10 nM insulin and 10 nM dexamethasone). The medium was changed daily. Under these culture conditions, cells are metabolically competent (Donato et al., 1995;Gómez-Lechón et al., 1997; Gómez-Lechón and Castell, 1998). For metabolic studies, hepatocytes were incubated with 250 μM almokalant. Treatment was started after 24 h of culture, and incubation medium and cells were subsequently frozen after 24 and 48 h of continuous incubation with the compound, to be further analyzed for the metabolic profile. Parallel plates without cells, incubated with medium containing the compound, were collected and frozen after the same incubation periods.

Studies Using Human Hepatocytes Isolated from Liver Resections.

Adult human hepatocytes were obtained from liver resections from primary or secondary tumors. Hepatocytes were seeded in Williams' medium supplemented with 10% fetal calf serum at a density of 10⁶ cells in 2 ml of medium in a 2.5-cm diameter Petri dish. The medium supplemented with 5 × 10⁻⁵ M hydrocortisone but lacking serum was renewed daily (Gugen-Guillouzo and Guillouzo, 1986; Abdel Razzak et al., 1993).

Almokalant was added to 24-h cultures at 100 μM. Incubations lasted 10, 24, and 48 h before media and cells were harvested. The compound was renewed at 24 h with the medium. Media and cells were stored at −80°C until analyses.

Preparation and Incubation of Precision-Cut Liver Slices.

The sources of the tissue culture materials were as described previously (Beamand et al., 1993; Lake et al., 1998). Samples of human liver (surplus to transplant requirements) were collected and transported to TNO BIBRA on ice. Tissue cylinders from liver samples were prepared using a 10-mm diameter motor-driven tissue-coring tool. From the cylinders, liver slices (200–300 μm) were prepared in oxygenated (95% O₂/5% CO₂) Earle's balanced salt solution containing 25 mM d-glucose, 50 Tg/ml gentamicin, and 2.5 Tg/ml Fungizone using a Krumdieck tissue slicer (Alabama Research and Development Corporation, Munford, AL). The liver slices were floated onto Vitron type C titanium roller inserts (two slices per insert) and cultured in glass vials containing 1.7 ml of culture medium in a Vitron dynamic organ incubator (Vitron Inc., Tucson, AZ). The culture medium consisted of RPMI 1640 containing 5% (v/v) fetal calf serum, 0.5 mMl-methionine, 1 TM insulin, 0.1 mM hydrocortisone-21-hemisuccinate, 50 μg/ml gentamicin, and 2.5 μg/ml Fungizone. Liver slice cultures were maintained at 37°C in an atmosphere of 95% O₂/5% CO₂. After 60 min, the medium was changed to fresh serum-free RPMI 1640 medium containing 0.5 mMl-methionine, 1 TM insulin, 0.1 mM hydrocortisone-21-hemisuccinate, and 0 to 500 μM A1 (dissolved directly in the culture medium). Liver slices were incubated for periods of 30 to 360 min and the incubations terminated by removing the vials from the incubator and plunging them into ice. Appropriate blank incubations (i.e., liver slices in medium without any A1 and A1 in medium without any liver slices) were also performed. To achieve a good recovery of parent compound and metabolites (Worboys et al., 1995), the liver slices were removed from the mesh of the roller inserts and homogenized in the culture medium by sonication (Beamand et al., 1993). Liver slice/medium homogenates were assayed for total protein content by the method of Lowry et al. (1951) using bovine serum albumin as standard. Total protein content was determined to allow for any differences in liver slice thickness between vials and to provide a scaling factor for intrinsic clearance calculations. The liver slice/medium homogenates stored at −80°C before dispatch to the laboratory undertaking the analysis of A1 metabolites.

Calculations.

Clearance intrinsic (CL_int) is defined as the proportionality constant between initial elimination rate (V₀) and the drug concentration (C). According to the Michaelis-Menten equation,V₀/C =V_max/K_m +C = CL_int. At substrate concentrations considerably lower than K_m, CL_int =V_max/K_m.

At low substrate concentrations, CL_int can also be calculated from initial rates (CL_int = initial rate/C) (Ito et al., 1998).

When estimating CL_int in vivo from CL_int in vitro the following constants were used: 200 mg of protein/g of liver (used for tissue slices and hepatocytes) and 45 mg of microsomal protein/g of liver (Bayliss et al., 1990;Houston, 1994).

To calculate the hepatic clearance (CL_H), the well stirred model was applied excluding protein binding: CL_H = Q_H · CL_int/Q_H + CL_int.

Hepatic blood flow, Q_H = 1450 ml/min (Davies and Morris, 1993).

Results

Modeling.

Almokalant was able to fit the putative active site of CYP3A4 (Lewis et al., 1999) in a variety of ways which would lead toN-dealkylations (three possible routes) andS-oxidation. In the latter respect, there is some analogy with the known CYP3A4-mediated metabolism of omeprazole. However, despite the presence of a basic nitrogen atom (protonatable at physiological pH), almokalant is not likely to act as a CYP2D6 substrate due to the presence of the cyano group at the preferred site of oxidation. Moreover, almokalant did not display the known characteristics of substrates of other human hepatic P450s with the possible exception of CYP2C19 where the disposition of hydrogen bond acceptors indicates that an interaction may occur.

cDNA-Expressed Enzymes.

The turnover of almokalant by the expressed enzymes was low. Almokalant does not appear to be a specific substrate for any of the major human drug-metabolizing P450 forms but was metabolized to a small extent by several different forms.

The dealkylated metabolite, M3, was produced mainly by CYP1A1 with a possible small contribution from P450s 2C19 and 2D6. The deethylated metabolite, M13, was formed in incubations containing P450s 1A1, 1A2, 2B6, 2C9, 2C19, and 2D6. The sulfone, M16, was produced mainly by P450s 1A1, 2C9, 219, and 2D6 (Table 2). Addition of cytochrome b₅ to the incubation with CYP3A4 did not influence the activity.

Human Liver Microsomal Fraction.

Human liver microsomal fraction incubated with almokalant produced three NADPH-dependent metabolites (M3 via N-dealkylation, M13 via N-deethylation, and M16 via sulfoxidation). The formation rates were low, and initial saturation curves indicated thatK_m values were in the millimolar range. It was therefore decided that metabolite formation at 20 μM would be well below K_m and should therefore be on the part of the saturation curve where initial rates correlate with CL_int.

The formation of M3, M13, and M16 by human liver microsomes incubated with 20 μM almokalant over a period of 120 min was conducted. The metabolite formation rates were linear for 60 min. Over that period the formation rate was as follows: M3, 4.5 pmol/mg/min; M13, 3.8 pmol/mg/min; and M16, 1.8 pmol/mg/min (Table3).

Two conjugated metabolites, M18a and M18b, were formed in human liver microsomes incubated with UDPGA. To estimate kinetic parameters for the formation of these two metabolites the Michaelis-Menten equation was fitted to the data. V_max andK_m values for the formation of M18 a were 658 pmol/mg/min and 936 μM, respectively. TheV_max and K_mvalues of M18b were 6222 pmol/mg/min and 1444 μM, respectively (Fig.2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Saturation curve for the formation of the glucuronidated metabolite M18a (A) and M18b (B) of almokalant in human hepatic microsomal fraction.

Correlation Analysis.

The hepatic microsomal formation rates of M3, M13, and M16 were highly correlated with each other (r ≥ 0.8, P≤ 0.005). The formation rate of all three of these metabolites correlated with CYP3A4-dependent monooxygenase activity (midazolam 1′-hydroxylase) (M3: r = 0.90, P < 0.001; M13: r = 0.66, P = 0.05; M16:r = 0.88, P < 0.002). The formation rate of M3 and M16 also correlated with microsomal CYP3A4 apoprotein content (M3: r = 0.76, P < 0.02; M16:r = 0.78, P < 0.02) (Table4).

Inhibition Analysis.

The effect of almokalant on P450 model activities was independently investigated in two laboratories. Almokalant was a weak inhibitor of CYP2C9-dependent activity (tolbutamide hydroxylation) with an IC₅₀ value of 100 to 200 μM. The IC₅₀ values for all of the other activities tested were consistently ≥1 mM (phenacetin O-deethylation, coumarin 7-hydroxylation, dextromethorphan O-demethylation, chlorzoxazone 6-hydroxylation, and testosterone 6β-hydroxylation) (data not shown).

Using P450-selective inhibitors, the formation of M3 and M13 from almokalant (20 μM) in human liver microsomes was inhibited 90% by 10 μM coumarin (CYP2A6 inhibitor) and 80% by 1 μM ketoconazole (CYP3A4 inhibitor), respectively. The other inhibitors tested (furafylline, quercetin, sulfaphenazole, S-mephenytoin, quinidine, and pyridine) did not show specific inhibition of metabolite formation (data not shown). The formation of M16 was too low to be investigated.

Human Hepatocytes.

Two different protocols for isolation and incubation of human hepatocytes were used. In cells isolated from surgical biopsies (n = 4), human hepatocytes were incubated with 250 μM almokalant. Samples of culture media were withdrawn at indicated time points up to 48 h. Two metabolites were identified, the glucuronides M18a and M18b. Their formation was linear for 24 h. The formation rate was 78 ± 36 pmol/mg/min for M18a and 228 ± 64 pmol/mg/min (n = 4) for M18b (Table 3).

Human hepatocytes isolated from two different liver resections were incubated with almokalant at 100 μM, over a period of 48 h (Table 3).

The two conjugated metabolites, M18a and M18b, were identified. Their formation was linear for 24 h. The formation rate of M18a was 13.1 and 76 pmol/mg/min, respectively, for the two preparations. The formation rates of M18b were 56.9 and 196 pmol/mg/min, respectively, in the two experiments.

Precision-Cut Liver Slices.

Liver slices were incubated with 20, 50, 200, and 500 μM almokalant for 90 min. The formation rates for the conjugated (M18b), dealkylated (M3), N-deethylated (M13), and sulfone (M16) metabolites did not plateau over the concentration range used, indicating that theK_m values for these metabolites were greater than the highest concentration almokalant tested (500 μM), consistent with the data obtained with hepatic microsomal fractions.

Liver slices incubated with 50 μM almokalant produced metabolites M13, M16, and M18b linearly throughout the 90-min experimental period, whereas the concentration of M3 decreased during the experimental period. It is possible that M3 was further metabolized to products not detected in the system used. The formation rates for M13, M16, and M18b were 1.3, 0.7, and 9.5 pmol/mg/mg of slice protein, respectively. The formation rate of M3 (estimated over a 30-min incubation) was 4.3 pmol/min/mg of tissue slice protein. The concentration of glucuronide M18a was below the limit of detection (Table 3).

Liver slices from a second subject were incubated with 50 μM almokalant for 6 h. Three metabolites were identified, the two conjugates (M18a and -b) and the N-deethylated product (M13). The two conjugates were formed linearly for 6 h, whereas formation of M13 leveled off after 3 h. The formation rates for M13, M18a, and M18b were 3.2, 3.9, and 10.8 pmol/mg/mg of slice protein, respectively (Table 3).

Clearance Estimates.

The clearance estimates calculated from the various in vitro systems can be seen in Table 5. NADPH-dependent metabolite formation by human liver microsomal fraction with 20 μM almokalant was well below K_m values for the formation of any of these metabolites; hence, the clearance could be calculated from the initial rates of metabolite formation over a 60-min incubation period. The rates were 4.5, 3.8, and 1.8 pmol/min/mg of microsomal protein for the formation of M3, M13, and M16, respectively. Thus, the CL_int by oxidative metabolism in a 70-kg man was calculated as 31.6 ml/min. The hepatic clearance was then calculated using the well stirred model as 31 ml/min, i.e., hepatic extraction by oxidative metabolism would be minimal.

The rate of almokalant glucuronide formation was considerably greater than that for NADPH-dependent metabolism, and full saturation curves were obtained. The CL_int(V_max/K_m) was 0.70 and 4.3 μl/min/mg microsomal protein for the formation of M18a and M18b, respectively. The CL_int for metabolism by glucuronidation in a 70-kg man was calculated as 359 ml/min. The well stirred hepatic model predicts a hepatic clearance of 287 ml/min, i.e., hepatic extraction by glucuronidation is predicted as 20%.

When cultured human hepatocytes isolated from biopsies were incubated with 250 μM almokalant, a concentration well below theK_m values was obtained for the formation of the two conjugates, and the combined formation of M18a and M18b was 306 pmol/mg/min. Using this rate, the CL_int value calculated was 246 ml/min in a 70-kg man. Hepatic clearance estimated from the well stirred model was 212 ml/min, giving a predicted hepatic extraction ratio of 0.15.

When the cells isolated from liver resections were incubated with 100 μM almokalant, the combined formation rates for M18a and M18b were 70 pmol/mg/min, and in the second experiment 273 pmol/mg/min. The hepatic CL_int calculated from the formation rates was 917 and 235 ml/min for the first and second experiment, respectively. Using the well stirred model, the hepatic clearance was calculated as 561 and 202 ml/min, respectively.

For precision-cut liver slices clearance estimates were obtained using data from both of the two experiments. The concentration of almokalant was 50 μM and thus well below K_m values. CL_int was therefore calculated using initial rates for the formation of metabolites. In experiment A, four metabolites, M3, M13, M16, and M18b, were formed. The formation rates of all metabolites were linear for 90 min except for M3. The formation rate of M3 (estimated over a 30-min incubation) was 4.3 pmol/min/mg of tissue slice protein. The formation rates of M13, M16, and M18b (90-min incubations) were 1.1, 0.52, and 6.4 pmol/min/mg of liver protein. The incubation time was probably too short to yield detectable amounts of M18a, which would result in an underestimation of the clearance values. The CL_int calculated from data for the four metabolites detected was 71 ml/min in a 70-kg man. The hepatic clearance was estimated as 55 ml/min when protein binding was taken into account and 68 ml/min when protein binding was not included in the model. In experiment B, three metabolites, M13, M18a, and M18b, were detected. The formation of M13 was linear for 3 h and the rate of formation was 3.2 pmol/min/mg of tissue slice protein. The formation of M18a and M18b was linear for 6 h and the rates were 3.9 and 10.8 pmol/min/mg of liver protein. The CL_int in a 70-kg man estimated from the formation of the three metabolites was 101 ml/min. Using the well stirred model hepatic clearance was estimated to 94 ml/min.

Formation clearance to the two glucuronidated metabolites M18a and M18b in the different in vitro systems can be seen in Table6. In all systems, the formation of M18b dominates over the formation of M18a. The ratio M18b/M18a is highest with hepatic microsomal fraction (7.1-fold), whereas the ratio in hepatocytes and liver slices varies between 2.8- and 5.3-fold.

Discussion

During its development the pharmacokinetics of almokalant were well characterized. Despite almost complete absorption, almokalant has an oral bioavailability of only 55%, presumably due to presystemic metabolism in the liver and/or intestinal tract. Total body plasma clearance is 12 ml/min/kg of body weight, and the renal clearance is 2 ml/min/kg of body weight. Less than 1% of the dose is recovered in the feces. The plasma protein binding is 20%. The volume of distribution at steady state is 2.6 l/kg. After oral administration (25.6-μmol dose), almokalant is rapidly absorbed with a mean peak plasma concentration of 59 μmol/l within 50 min. The mean terminal half-life is 3.6 h. The metabolism of almokalant in vivo in humans has been studied extensively (Fig. 1). Following administration of tritium-labeled almokalant to healthy volunteers the given radioactivity could be almost completely recovered in urine within 24 h. About 10% of the dose was excreted unchanged.

In the urine, glucuronides (M18a and M18b), which are stereoisomers and the result of direct conjugation of the hydroxyl group, are the dominant metabolites (69% of an oral dose). These were also the major metabolites in hepatocyte and precision-cut liver slice incubations. Oxidative metabolites, most likely the result of P450-dependent activity, constitute less than 10% of the dose in human urine and were also only minor products in whole cell in vitro systems. Metabolites detected in urine samples in vivo not found in the in vitro systems were the aliphatic hydroxylated product (M6) and the acidic M17. M6 is a minor metabolite, and the formation rate may be too slow to produce detectable amounts in vitro. M17 is the product of several metabolic steps, which may not be seen, in the in vitro systems. The results in vivo thus indicate that P450 enzymes are of little importance in the clearance of almokalant. The results from the present in vitro studies also clearly suggest that the P450 enzymes play little role in the clearance of the drug.

The P450 protein homology models predicted the main P450-dependent metabolites (N-dealkylation and sulfoxidation products) of almokalant. Hepatic microsomal fraction and liver slices also formed these metabolites. Molecular modeling suggested that CYP3A4 was the enzyme responsible for these metabolic pathways, whereas almokalant was excluded as a substrate for CYP2D6. In contrast, recombinant expressed CYP3A4 did not metabolize almokalant, even in the presence of exogenous cytochrome b₅, whereas CYP2D6 produced all three metabolites. However, the turnover in these experiments using cDNA-expressed enzymes was low. Furthermore, almokalant was metabolized by several of the expressed P450 enzymes. On the other hand, different conclusions were reached using P450-selective inhibitors and hepatic microsomal fraction. The formation of M3 was inhibited by coumarin, a CYP2A6 substrate, and the formation of M13 was inhibited by ketoconazole, a CYP3A4 inhibitor. However, almokalant did not inhibit, or was only a very poor inhibitor, of the respective P450-selective reaction. Furthermore, the formation rate of M3 and M16 correlated with apoprotein content and CYP3A4-depdendent activities in a panel of human liver microsomes. M13 correlated with CYP2C19- and CYP3A4-dependent activities. These ambiguous in vitro results probably indicate a problem to identify responsible enzymes for minor metabolic routes by in vitro methods. Almokalant may also be a relatively unselective and poor substrate for P450 enzymes, further complicating the attempt to identify the enzymes catalyzing the oxidative metabolic routes.

All of the in vitro systems used identified conjugation of almokalant with glucuronic acid as the major metabolic clearance route. The conjugated metabolite M18b dominated over M18a in all of the in vitro systems, which also reflects the ratio between the two conjugates found in urine. In urine, the concentration of M18b is 1.4-fold higher than that of M18a. In the in vitro systems, the clearance to M18b was 2.8- to 7.1-fold greater than to M18a. The largest difference between clearance to M18b and M18a was that obtained using human hepatic microsomal fraction. However, the artificial conditions using saturating concentrations of UDPGA and Brij treatment of microsomal membranes may have affected the specificity of the UDP glucuronosyltransferases. Furthermore, in vivo almokalant might also be conjugated by extrahepatic enzymes, which could alter the metabolite ratio. When comparing the clearance values obtained with the different in vitro systems, the formation of M18b varied over 5-fold whereas formation of M18a varied less than 3-fold. Since the formation of the two glucuronides is the result of stereoselective metabolism at the hydroxy group, presumably the in vitro systems differentially favor the formation of one of the almokalant enantiomers, either because of stereoselectivity of the enzyme involved or because the enantiomers of almokalant are conjugated by different enzymes, which are differentially retained or accessible in the various systems. Nevertheless, it was of interest that the relative formation of oxidative and glucuronidated metabolites found with hepatic microsomal fraction was very similar to that observed in vitro.

Although the in vitro systems identified conjugation of almokalant to be the major metabolic clearance route, the clearance estimated from the in vitro data underpredicted the clearance found in vivo. The total metabolic clearance of almokalant is 700 ml/min in a 70-kg man. The predicted clearance from studies with hepatic microsomal fraction using the sum of UDPGA- and NADPH-dependent metabolism amounts to half the clearance found in vivo. The rate of metabolism in isolated human hepatocytes showed large variation. There was a 4-fold difference between the preparations from liver resections, which might reflect the variation in the metabolic capacity of the livers dependent on age, diseases, drug treatment, etc. In contrast to the in vitro results, the clearance of almokalant in vivo exhibited a small variation, possibly because this study was done on a homogenous group of young healthy male subjects.

The predicted hepatic clearance amounts to 24 to 80% of the in vivo clearance. The clearance estimated from studies with precision-cut liver slices was only 10 to 15% of the clearance in vivo. All in vitro systems therefore consistently underpredict the in vivo metabolic clearance of almokalant. It should be noted, however, that for drugs dependent on oxidative metabolism for their elimination, in vitro systems have been shown to consistently underpredict their clearance (Obach, 1999; Carlile et al., 1999). No reasonable explanation for the consistent underpredictions of clearance by the different in vitro systems has been suggested.

Compared with the numerous predictions for drugs metabolized by P450 enzymes, there are few such studies on conjugated drugs wherein the in vivo clearance has been predicted from in vitro intrinsic clearance data. It has been speculated that poor in vitro-in vivo scaling for conjugated drugs might be due to differences in enzyme access to UDPGA, the latency of the enzyme in the endoplasmic reticulum, or β-glucuronidase activity in the in vitro system (Ito et al., 1998). Prediction of the in vivo metabolic clearance of troglitazone by sulfation and glucuronidation in different mouse strains using cytosol and Brij 58 activated microsomes was, however, reasonably good (Izumi et al., 1997). In a study by Furlan et al. (1999), four drugs that are mainly conjugated were used to compare the intrinsic clearance predicted from data obtained with microsomal fraction obtained from healthy and cirrhotic human livers. The decrease in clearance in patients with cirrhosis was reasonably well predicted. However, no quantitative predictions of the in vivo clearance were reported.

In summary, liver microsomes fortified with UDPGA or NADPH, primary hepatocytes in culture, and precision-cut liver slices all suggest that almokalant is cleared metabolically by conjugative routes, with a minor contribution from P450-dependent metabolism. However, the intrinsic clearance predicted by these in vitro systems was low compared with that observed in vivo, as has been reported relatively consistently for a number of other drugs. There may be a systematic reason for this.