Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Compound I¹ is a selective and potent inhibitor of farnesyl-protein transferase (IC₅₀ of 2 nM) that has been evaluated clinically for the treatment of cancer. FPTase is believed to be an appropriate target in cancer therapy because several mammalian proteins that contribute to the regulation of cell replication require post-translation modification by FPTase to achieve biological activity (Gibbs et al., 1997; Gibbs and Oliff, 1997; Kohl et al., 1998; Lobell and Kohl, 1998). Farnesyl transferase inhibitors as anticancer agents were developed based on the assumption that cellular transformation by Ras oncoproteins requires posttranslational modification of farnesylation catalyzed by FPTase (Gibbs, 2000).

This report describes the pharmacokinetics and metabolic fate of ¹⁴C-labeled compound I in bile duct-cannulated rats and dogs, the two species used in the toxicological evaluation of this compound. In addition, data from animal and human liver microsomal intrinsic clearance values were scaled to predict hepatic clearance and half-life in humans and compared with the in vivo data.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Chemicals. All chemicals and solvents were of HPLC or analytical grade. Water was distilled and purified through a Milli-Q reagent system (Millipore Corp., Bedford, MA). Ultima Gold (liquid scintillation counting), Carbo-sorb E, and Permaflour E+ (for combustion of samples) were obtained from Packard Instrument Co. (Meriden, CT). -Glucuronidase (from Helix pomatia, type H-1 with sulfatase activity) was obtained from Sigma Chemical Co. (St. Louis, MO). ¹⁴C-labeled I and some of its metabolites (M4, M8, M10, M11, and M14) were prepared at Merck Research Labs. The ¹⁴C label was incorporated at the cyano-moiety (Fig. 1), having a specific activity of 22.4 µCi/mmol and a radiochemical purity of >98% as determined by radio-HPLC. All other chemicals were purchased from commercial sources and were of the highest purity available.

View larger version (8K): [in this window] [in a new window]   Fig. 1.   Chemical structure of compound I. *, represents the site of 14C label; a to i and 2 to 6 represent proton labels for NMR analysis.

Animal Studies.

Rats Male Sprague-Dawley rats (n = 3), weighing approximately 300 g, had an indwelling cannula implanted in the right jugular vein for drug administration and blood sampling. The surgery was performed under light pentobarbital anesthesia (40 mg/kg i.p.) 1 day before the experiment. During the study, all animals were housed in individual metabolism cages and were unrestrained throughout the experiment. Following an overnight fast, a 10-mg/kg dose was given by oral gavage and a 5-mg/kg dose by bolus intravenous injection. Oral bioavailability was determined using a solution of free base in 0.05 M citric acid, and the intravenous dose was administered as a saline solution. Blood samples were obtained serially at selected time points after drug administration via the jugular vein cannula. After centrifugation, the resultant plasma was stored frozen at 20°C.

Dogs. Oral bioavailability of a 10-mg/kg oral dose was assessed in fasted beagle dogs (n = 3) using a solution formulation in 0.05 M citric acid. Dose proportionality studies were conducted by administering compound I i.v. at doses of 2-, 5-, and 10-mg/kg doses of compound I as a solution in saline. For assessment of excretion and metabolic pathways, bile and urine were collected from fasted, male, bile duct-cannulated beagle dogs (n = 3). An i.v. dose was administered at 10 mg/kg in saline containing compound I ¹⁴C radiotracer at a final specific activity of 0.89 µCi/mg. Bile and urine were collected up to 72 h postdose and were kept at 20°C until analysis.

Pharmacokinetic Calculations. Area under plasma concentration-time curve (AUC) was determined by linear-log trapezoidal interpolation. The apparent half-life was estimated from the log-linear portion of the slope of time-concentration profiles. The apparent volume of distribution (V_dss) of the drug was calculated by the following equation:
where AUMC is the total area under the first moment of the drug concentration curve from zero to infinity. Clearance (CL) was calculated as:

Bioanalytical.

Quantitation of compound I in plasma by LC-MS/MS Concentrations of compound I in plasma were determined by a LC-MS/MS method following a protein precipitation procedure. An aliquot of the extract was analyzed by LC-MS/MS in the positive ion mode using an SCIEX API III triple quadrupole mass spectrometer (PE SCIEX, Concord, ON, Canada) equipped with a heated nebulizer interface using a discharge current of 3 µA and an orifice potential of 35 V. The heated nebulizer was maintained at 450°C. The instrument was operated in the selected reaction monitoring mode. The precursor and product ions of compound I were 406 to 195. The linear range for the assay was 10 to 2500 ng/ml.

Radiochemical analysis. Quantitation of radioactivity was achieved by liquid scintillation counting of samples in a Beckman LS5601 liquid scintillation counter (Beckman Coulter, Inc., Fullerton, CA) with external quench correction. Aliquots of urine and bile were assayed by direct addition to 15 ml of Ready-Safe (Beckman) scintillation cocktail.

LC-MS-radiometric analysis. Metabolite profiles in urine and bile were determined using a Finnigan LCQ mass spectrometer (Thermo Finnigan, San Jose, CA) coupled to an inline radioactivity flow detector that allowed simultaneous mass spectral and radioactive detection. The instrument was operated in the electrospray ionization mode with a 5-kV potential, and the transfer capillary temperature was maintained at approximately 200°C. The chromatographic column was a BDS Hypersil C₁₈ column (4.6 × 150 mm, 5-µm particle size; Keystone Scientific, Bellefonte, PA). A linear gradient commenced with an initial concentration of 95% solvent A (5 mM ammonium acetate, pH 4.5) and 5% solvent B (acetonitrile) and increased to 10% A over a period of 18 min at a flow rate of 1 ml/min. The samples were centrifuged (4000 rpm × 10 min), filtered through a 0.22-µm filter and injected directly on the HPLC system. The column effluent was split such that approximately 0.1 ml/min was introduced into the ion trap mass spectrometer equipped with an electrospray ionization source, and the remainder (0.9 ml/min) was mixed with 3 flow volumes of the Ultima M scintillation cocktail before detection by a Packard Radiometric Detector (515 TR; Packard Instruments Co.).

LC-NMR. A linear gradient consisted of 95% solvent A (D₂O), 5% solvent B (acetonitrile), and increased to 25% A in over a period of 25 min and then to 5% A from 25 to 35 min at a flow rate of 1 ml/min using a BDS Hypersil C₁₈ column (4.6 × 150 mm). The eluent was monitored by UV at 235 nm with a Varian 9065 photodiode array UV detector (Varian, Palo Alto, CA).

In Vitro Metabolism. To determine the intrinsic clearance required for the prediction of hepatic clearance, varying concentrations of compound I were incubated with liver microsomes (2 mg/ml protein) from rats, dogs, monkeys, and humans at 37°C for up to 60 min. Reactions were stopped by the addition of 2 volumes of acetonitrile, and the disappearance of parent drug was determined by LC-MS/MS. Intrinsic clearance was calculated from the ratio of V_max/K_m (well stirred hepatic model), scaled up to whole liver, and this value was applied, in conjunction with the experimentally determined unbound fraction in blood, to calculate hepatic clearance using the equations below (Wilkinson and Shand, 1975; Lin et al., 1996; Lin, 1997).
V_max and K_m are experimentally determined parameters using hepatic microsomes;
where Q_H is hepatic blood flow and f_u is fraction of compound unbound in plasma;
predicted half-life calculated assuming monoexponential pharmacokinetics and that distribution volume in humans is in the same range as that in rats, dogs, and monkeys.

Plasma protein binding and blood to plasma distribution ratio. The binding of ¹⁴C-labeled compound I to human, dog, rat, and monkey plasma protein was determined by an ultrafiltration method. Compound I was added to plasma to yield a final concentration of 40 µg/ml. The drug solution was added in a volume equal to or less than 1% of the plasma volume. Following incubation of the plasma samples at 37°C for 15 min, an aliquot of plasma (1 ml) was immediately transferred to a centrifuge tube (Amicon Co., Danvers, MA). The tube was then centrifuged at 1500g for 30 min at 37°C; the ultrafiltrates were removed and analyzed by liquid scintillation counting. The free drug fraction was determined by dividing the amount (dpm) of drug in the ultrafiltrate by the amount in the original plasma sample. The percentage of protein bound values were not corrected for nonspecific binding (<5%). The equilibrium blood to plasma distribution ratio of ¹⁴C-labeled compound I (2 µM) was determined at 37°C in freshly drawn rat, dog, monkey, and human blood after incubation for 15 min. Total radioactivity in whole blood and plasma was determined by combustion of the respective samples using a Packard 307 Oxidizer (Packard Instruments Co.).

Combined -glucuronidase/sulfatase enzyme hydrolysis. Urine and bile samples (0.1 ml) were added to 0.9 ml of sodium acetate buffer (pH 5, 0.1 M) and treated with 2000 units of -glucuronidase/sulfatase, and the mixture was incubated at 37°C overnight (Shimada et al., 1997). The reaction was stopped by adding 2 ml of acetonitrile and centrifuged at 4000 rpm. The supernatant was evaporated under nitrogen and reconstituted in 0.2 ml of the HPLC mobile phase (95% ammonium acetate and 5% acetonitrile). Incubation of urine and bile samples without the enzymes served as controls.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Absorption and Pharmacokinetics. Following i.v. administration (5 mg/kg), compound I was cleared rapidly in rats in a polyphasic manner with a terminal half-life of 41 min. The plasma clearance (CL_p) and V_dss were 41.2 ml/min/kg and 1.2 l/kg, respectively (Table 1). Less than 2% of the dose was excreted in rat bile and urine as unchanged drug over a period of 24 h, indicating that biotransformation was the major route of elimination of compound I. When the drug was administered orally as a solution in 0.05 M citric acid at a dose of 10 mg/kg, the absorption was rapid; the plasma concentration of compound I (C_max, 7.36 µM) was reached in about 16 min, and the bioavailability was 45% (Table 1).

Excretion. Following i.v. administration of the radiolabeled I, most of the radioactivity was recovered within 72 h postdose in rats and dogs (Table 3). In rats, about two-thirds of the dose was excreted in bile and about a third of the dose in urine, whereas in the case of dogs, about 60% of the dose was excreted in urine and 29% was recovered in bile. In both species, the parent drug accounted for less than 2% of the recovered dose.

Plasma Protein Binding and Blood to Plasma Partition Ratio. The fraction unbound was 33.6, 32.5, and 30.9% in the rat, dog, and human plasma, respectively, indicating that I was not bound extensively to plasma proteins. The mean equilibrium blood-to-plasma concentration ratio in rats, dogs, and humans at the therapeutically relevant concentration of 2 µM was 0.83, 0.53, and 0.68. This modest extent of distribution into red blood cells indicated that plasma clearance underestimates slightly the blood clearance.

Prediction of Human Hepatic Clearance. Using the well stirred model (Wilkinson and Shand, 1975; Lin et al., 1996; Lin, 1997), hepatic clearance was calculated for laboratory animals and humans from in vitro metabolic data and the experimentally determined unbound fraction in plasma. The predicted clearance values based on the in vitro data were in good agreement with the observed in vivo plasma clearance in dogs and monkeys, whereas it was lower in case of rats (Table 2). Using the same approach with three different pools of human liver microsomes, the clearance of compound I in humans was predicted to be low, with a mean value of 3 ml/min/kg. In humans, provided that compound I is cleared primarily by hepatic metabolism exhibiting monoexponential pharmacokinetic behavior, and assuming that the distribution volume in humans to be within the range estimated in rats, dogs, and monkeys (0.6-1.2 l/kg), the corresponding half-life was predicted to be in the range of 2.3 to 4.6 h.

Metabolism. Radiolabeled components of urine and bile were analyzed for unchanged compound I and its metabolites using tandem LC-MS-radiometric detection. Representative radiochromatograms from rat and dog urine and bile are shown in Fig. 2. The radiochromatograms indicate presence of minor amounts of unchanged I (<2%) and about 19 metabolites in urine and bile (Table 3).

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Representative radiochromatographic profile of a 0- to 24-h sample after intravenous infusion of 14C-labeled compound I. A, dog urine; B, dog bile; C, rat urine; D, rat bile.

View larger version (13K): [in this window] [in a new window]   Fig. 3.   MS/MS of 14C-labeled compound I (M + H+ = 406 Da) and proposed fragmentation pattern.

View this table: [in this window] [in a new window]   TABLE 4 LC-1H NMR data of compound I and its major metabolites excreted in dog and rat urine and bile See Fig. 1 for proton assignments. Chemical shifts () are expressed in ppm. 1H-1H spin-spin coupling constants (J) are expressed in Hz.

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Metabolites of 14C-labeled compound I in dog and rat urine and bile.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The pharmacokinetics and metabolism of compound I displayed species differences in rats compared with dogs. Compound I had a lower plasma clearance in dogs (10.6 ml/min/kg) compared with that in rats (41 ml/min/kg), but the half-life was similar in both the species (~40 min). The similar half-life, yet different clearance values, could partly be explained by the difference in V_dss (about half in dog compared to rat; Table 1). Moreover, the mean equilibrium blood-to-plasma concentration ratio in rats and dogs was 0.83 and 0.53, respectively, indicating that the plasma CL_b > CL_p in both the species. Oral bioavailability was complete in dogs but somewhat less in rats. Since the rat CL_b (~49 ml/min/kg) is moderate relative to hepatic blood flow, the incomplete bioavailability may be due in part to hepatic first-pass metabolism.

The use of in vitro systems affords the opportunity to produce quantitative data on human clearance before initiation of pharmacokinetics studies in vivo (Houston, 1994; Ito et al., 1998; Lin, 1998; Obach, 1999). Hence, these in vitro-in vivo correlations have become very useful in helping select drug candidates with desirable pharmacokinetic properties in humans. In the present study, using the well stirred model (Wilkinson and Shand, 1975; Lin et al., 1996; Lin, 1997), hepatic clearance of compound I was estimated from in vitro CL_int values using liver microsomes from rats, dogs, monkeys, and humans. With the exception of rats, the predicted clearance values in dogs and monkeys were in good agreement to those observed in vivo (Table 2). The predicted clearance values in rats were about 10-fold lower than the observed values, and one possible explanation is a large extrahepatic metabolism component of clearance. Using the predicted human clearance value of 3 ml/min/kg, at initial phase I dose of 35 mg/m²/day (administered by continuous infusion), the projected steady-state concentration was calculated as 0.5 µM.

The predicted pharmacokinetic values were in good agreement with results obtained in subsequent clinical studies in patients with solid tumors. Following administration of a dose of 35 mg/m²/day, the mean steady state concentration was 0.8 µM (range, 0.3-1.3; n = 3), the plasma clearance range was 2.3 to 4.3 ml/min/kg, and the half-lives for - and -phases were 0.6 and 3.3 h, respectively (Soignet et al., 1999). As expected, based on the preclinical studies, metabolism was the major route of clearance with less than 1% of the parent drug detected in the urine.

Metabolism of compound I was extensive with only a small percentage of the unchanged drug (<2% of the dose) excreted into the urine and bile of dogs and rats (Table 3). Biliary excretion was the major route of elimination in rats, whereas the reverse was true in dogs. N-Dealkylation at two different positions of the molecule was the most important biotransformation pathway, resulting in formation of about a dozen products, including subsequent phase II conjugation products. Metabolism of compound I was extensive in rat, dog, monkey, and human liver microsomes, and qualitatively the metabolic pathways were similar in all species (unpublished results, Merck Research Labs). The dealkylated metabolites that contain the imidazole portion of the molecule, but lack the radiotracer (M5, M12 and M14), were not quantified. Metabolite M12 was formed via either oxidation of M14 or N-dealkylation of M17 or M19. Based on LC-MS/MS data, the site of oxidation was identified to be on the chlorophenyl ring, but the regio-chemistry was not determined. Another major biotransformation observed in rats and dogs is the hydroxylation on the C-4 atom of the chlorophenyl ring (M17). The corresponding phase II metabolism was also different between rats and dogs. In case of rats, M17 underwent mostly sulfation (M15), whereas glucuronidation was the preferred pathway in dogs (M9).

Both dog and rat showed extensive dealkylation and oxidation of compound I, but there are a few differences in metabolite profiles. The process of dealkylation between the imidazole and piperazinone rings most likely goes through an aldehyde intermediate, which either undergoes reduction to M8 or oxidation to M4 (Testa and Caldwell, 1995). There was some species difference between these two pathways in rats and dogs. In the case of rats, the reductive pathway to M8 and its glucuronide accounted for about 27% of the dose, whereas about 15% was accounted for in the dog. Conversely, the oxidative pathway leading to M4 and its glucuronide M3 accounted for about 30% of the dose in dogs and about 11% in rats.

In conclusion, compound I exhibited some differences in pharmacokinetics and metabolic behavior of the laboratory animals. It had a relatively short half-life and was cleared primarily by phase I and II metabolism in rats and dogs. In this study, reasonably accurate human clearance and half-life was predicted using human and animal in vitro metabolism data. Such predicted clearance data could be useful in selecting the best candidate for progression into development and also to assist in determination of the optimum first dose in man.