Introduction

Abstract Introduction Results Discussion References

FIN¹ (Proscar; fig. 1) is an orally active 4-azasteroid synthesized by MRL that has been demonstrated to be effective in the clinical management of BPH (1-3), one of the most common diseases exhibited by middle-aged and older men (4). Abnormalities of prostate growth are common only to humans and male dogs. Although there are differences between human and canine BPH, the condition in the dog has many features in common with the human disease, and this model is widely accepted as the best animal model for studying BPH (5, 6).

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Proposed metabolic pathway for FIN in dogs. FIN used in this study was radiolabeled with 14C (*) and 3H ().

FIN is a potent, mechanism-based inhibitor of human type 2 5-reductase (7), the enzyme expressed in prostate, liver, genital skin, and male accessory sex glands, whereas type 1 5-reductase is expressed in the sebaceous glands of human skin and is poorly inhibited by FIN (8-10). In animal studies in both rats (11) and dogs (12-14), treatment with FIN resulted in a significant reduction in the size of the prostate gland. These findings provide further support for the use of the dog as a model for the human disease.

The metabolism and excretion of FIN in rats and human subjects and its in vitro metabolism have been reported (15-18). FIN undergoes extensive hepatic metabolism through oxidative pathways giving metabolites that are eliminated primarily through bile. After an oral dose of [¹⁴C]FIN (38.1 mg) to humans, 39% of the dose was excreted renally. The major radiolabeled component recovered from the urine was the carboxylic acid metabolite (17). Recently, CYP3A has been shown to be the major enzyme involved in the biotransformation of FIN (19). However, only limited data have been published on the in vivo metabolism of FIN in the dog (20). The aim of the present study was to describe the metabolic fate and pharmacokinetics of FIN in the dog after an intravenous dose and oral doses of 10 and 80 mg/kg. Metabolic patterns are of interest because they demonstrate the dog resembles humans in major metabolic pathways. This is important because the dog is the relevant animal model for the study of BPH, and in addition served as one of the primary species used for investigation of the pharmacology and toxicology of FIN.

Materials and Methods

Chemicals. FIN (molecular weight: 372), the IS (the 4-N-methyl analog of FIN), and the oxidative metabolites (-OH-FIN, -carboxy-FIN, FIN--al, and Me FIN--oate) were prepared at MRL (Rahway, NJ). [1,2-³H]FIN and [t-butyl-stem-¹⁴C]FIN (radiopurities >98% at specific activities of 18.5 and 9.5 µCi/mg, respectively) were used for preparation of the intravenous dose and the 10 mg/kg oral dose. For the 80 mg/kg oral dose, [¹⁴C]FIN was diluted to a specific activity of 1 µCi/mg. Both radiolabeled compounds were prepared at MRL. All organic solvents (EM Science, Gibbstown, NJ) were HPLC grade. Water was purified in a Millipore Milli-Q system (Bedford, MA). Carbosorb, Permafluor, and Monophase S (Packard, Downers Grove, IL) were used for tissue combustions. All other chemicals and reagents were used as described previously (21).

Dosing and Sample Collection. Five male beagle dogs (body weight 10-16 kg, 2-7 years old) were obtained from either White Eagle Farms (Doylestown, PA) or Marshall Farms (North Rose, NY). Animals were housed and maintained according to MRL IACUC-approved guidelines. Dogs were fasted (water ad libitum) for ~16 hr before dosing. Three dogs were dosed with [¹⁴C]FIN (10 mg/kg po): one dog in the group received [³H]/[¹⁴C]FIN. In a separate study, two of the three dogs received [¹⁴C]FIN at 80 mg/kg po. The drug was suspended in 0.5% methylcellulose and given by gavage. Two dogs were dosed intravenously with [¹⁴C]FIN at 5 mg/kg (one animal received the doubly labeled drug). For administration of the intravenous dose, the dog was restrained in a sling, and a catheter was implanted in the brachiocephalic vein. The intravenous dose was prepared as a solution in propylene glycol/ethanol [70:30 (v/v)] and filtered through a 0.2-µm syringe filter. The solution was infused at a constant rate (~0.3 ml/min) over 30 min with the aid of a Harvard Apparatus syringe pump. After dosing, all animals were housed in metabolism cages, and food was returned 4 hr later. Serial blood samples (heparinized) were obtained from the jugular vein during the infusion, at 5 min postinfusion, then at timed intervals up to 72 hr postdose; urine and feces were collected daily at the same time, except that after the 80 mg/kg dose excreta were collected for each 24-hr period up to 8 days postdose. Plasma was separated. All samples were stored at 20°C until they were assayed.

Analysis of Radioactivity. Plasma and urine samples were assayed for total radioactivity by direct LSC (Packard Tri-Carb Scintillation Spectrometer model 1900 TR) using a liquid scintillant (Insta-Gel, Packard). Before analysis, a separate set of plasma samples, blood and fecal homogenates (1:4 dilutions of feces), were air-dried, combusted in a tissue oxidizer (Packard model B-306), and the resultant ¹⁴CO₂ and/or tritiated water was trapped and then mixed with scintillant. Concentrations of radioactivity were expressed as drug equivalents/milliliter (or gram) of sample. Difference in the total radioactivity obtained by the two methods (direct counting vs. combustion) was taken as evidence for the presence of a volatile tritiated species in plasma. Alternately, the presence of volatile tritium in a sample was determined by the difference of radioactivity in the sample determined before and after evaporation under a stream of nitrogen.

Analysis of Unchanged FIN and Metabolites in Plasma, Urine, and Fecal Samples. Concentrations of FIN in plasma and urine were determined by an HPLC method with UV and/or radiometric detection as previously reported (21), then modified to include the analysis of metabolites (17). The N-methyl analog of FIN was used as the IS to monitor the procedural recovery and reproducibility of the assay. UV response of the IS in each analyte was compared with that of an equivalent amount injected on-column. Typically, extraction efficiencies averaged 84%, with <5% variation among the samples. The LOQ of [³H]/[¹⁴C]FIN (1 ng-eq/ml) was based on radioactivity data.

Plasma. One- to 2-ml aliquots of plasma were prepared for analysis by dilution to 15 ml with water and addition of 4.1 µg of the IS. The sample was passed through a Sep-Pak C₁₈ cartridge (Waters Associates, Milford, MA), followed by elution with methanol/H₂O) [70:30 (v/v)]. The eluate was taken to dryness with nitrogen, diluted with H₂O, and passed through a Sep-Pak CN (cyanpropyl) cartridge. Typically, 1-4% of the total radioactivity was not retained on the cyano cartridge; thus, it was characterized as acidic based on the selectivity of the phase. Unretained fractions were stored frozen until analysis. Retained radioactivity was eluted from the cyano cartridge with methylene chloride. The solvent was evaporated and the analyte was redissolved in methanol for analysis by HPLC. To recover the radioactivity quantitatively, samples were applied to C₁₈ cartridges. Retained radioactivity was desorbed with 100% methanol. The eluate was concentrated by evaporation, redissolved in methanol, and analyzed by HPLC using method 3.

Urine. A 5-ml aliquot of urine (0-24 and 24-48 hr) was mixed with 5 ml 0.2 M NaK₂PO₄ (pH 7.4) buffer and 4.1 µg of the IS. These samples were then extracted sequentially at pH 7.4 and pH 2 with methylene chloride and ethyl acetate; solvent fractions and the aqueous layer were assayed for total radioactivity. In a separate study, incremental specimens collected from 0 to 72 hr after 80 mg/kg FIN were pooled on a proportional basis before fractionation as described. Before HPLC analysis, each sample of neutral and acidic urinary radioactivity was purified further by use of serial adsorptions-desorptions from C₁₈ cartridges. The pH of the aqueous layers was adjusted to 5, and the sample was stored frozen until used in enzymatic deconjugation experiments. Incubation with -glucuronidase. Before incubation, the aqueous fraction from 0 to 72 hr dog urine (50 ml, containing on average 18% of the total urinary radioactivity) was extracted with ethyl acetate to remove products resulting from nonenzymatic hydrolysis. Each of the resulting aqueous samples was applied to eight C₁₈ Sep-Paks: radioactivity was desorbed quantitatively with 100% methanol. The effluent was evaporated and the residue dissolved in 0.2 M sodium acetate buffer (pH 5.0). Aliquots of samples (50 µg equivalents of radioactivity) were incubated overnight at 37°C with and without -glucuronidase (2000 units, Helix aspersa, Type HA-4; Sigma Chemical Co., St. Louis, MO). In a parallel incubation, D-saccharo-1,4-lactone (0.5 mM) was added. After incubation, samples were extracted as described previously for urine. The resulting extracts, containing neutral products, were analyzed by HPLC using method 1 (described herein); acidic extracts, containing little radioactivity, were not assayed.

Feces. Ten-milliliter-aliquots of feces were diluted with pH 7.4 buffer and extracted at neutrality and at pH 2 in the same way as described previously for urine, except that the aqueous fractions of feces contained little radioactivity and were not saved.

HPLC Analysis. The HPLC system consisted of two pumps (Spectroflow 400), a UV detector and gradient controller (Spectroflow 783) from Kratos/Applied Biosystems (San Jose, CA), and Rheodyne 7125 injector (Cotati, CA). Data acquisition and integration were performed by use of a PE Nelson Analytical Model 2600 Data System, 760 series interface (Cupertino, CA). All analyses were conducted on a Zorbax C₈ analytical column (4.6 × 250 mm, Mac-Mod Analytical, Inc., Chadd's Ford, PA) with a LC-8 packed guard column (Supelco, Bellefonte, PA). Flow rate was 1 ml/min. Column effluent was monitored by UV absorbance at 210 nm. HPLC separations were conducted at ambient temperature. Four HPLC methods were used as follows: methods 1 and 2 for separation of the major and minor neutral metabolites, respectively; method 3 for separation of acidic metabolites; and method 4 for purification of those metabolites partially resolved by method 1. In method 1, the column was eluted isocratically with a mobile phase consisting of methanol/acetonitrile/H₂O [39:26:35 (v/v/v)]. Typical retention times for FIN, its metabolites, and the IS with this system are as follows (in minutes): FIN, 13.5; 6-OH-FIN, 7.5; -OH-FIN, 6.1; ,6-(OH)₂-FIN, 4.5; and IS, 21. In method 2, the column was eluted with a gradient from 0% to 100% B in 25 min with solvent A as methanol/acetonitrile/H₂O [21:14:65 (v/v/v)] and solvent B as methanol/acetonitrile/H₂O [45:30:25 (v/v/v)]. The retention times of FIN--al and Me FIN--oate were 21 and 22 min, respectively. In method 3, the mobile phase consisted of solvent A [acetonitrile/H₂O, 10:90 (v/v) containing 0.2% H₃PO₄] and solvent B [acetonitrile/H₂O, 90:10 (v/v) containing 0.2% H₃PO₄], and the column was eluted with a linear gradient from 0 to 100% B in 60 min. Typical retention times were -carboxy-FIN, 26.8 min (25-28) and FIN, 36-37 min. Method 4 used two C₈ columns connected in series and eluted at a flow rate of 0.7 ml/min with a mobile phase of methanol/acetonitrile/H₂O [33:22:45 (v/v/v)]. Retention times were noted for -OH-FIN (30.3-32.3 min) and the di-OH-FIN metabolites (13-18 min). For off-line radioactivity detection and metabolite isolation, fractions of column effluent were collected at 0.5- or 1-min intervals (ISCO Retriever II). Radioactivity content of each fraction was determined by LSC, and the results were expressed as a percentage of the total radioactivity that was recovered from the column. For each extract of plasma, urine, and feces, the percentage of the total radioactivity contained in each fraction was plotted to establish an elution profile of radioactivity. Concentrations of FIN and metabolites were calculated from the elution profiles as described previously (17). Percentage-based values obtained by HPLC were converted to radioactivity and subsequently converted to amounts by use of the appropriate specific activity; thereafter, amounts were reported as nanograms (or micrograms) of drug and metabolites. In performing this conversion, the small difference in the molecular weights of the various compounds was not taken into account. A comparison of concentrations measured by a specific HPLC assay and the percentage-based total radioactivity method revealed a very close correlation (17, 22).

Metabolite Isolation and Identification. Larger amounts of plasma, urine, and feces were processed to obtain metabolites, using essentially the same methods employed to obtain the respective metabolic profiles. Identification of metabolites was determined by cochromatography with authentic metabolite standards, NMR spectroscopy, and MS.

Spectral Studies. MS analysis of metabolites was performed by several techniques. DCI mass spectra (98 eV) were acquired on a Nermag R 1010C quadrupole mass spectrometer (Houston, TX) operated in the positive-ion mode with experimental conditions as follows: source temperature of 120°-130°C, reagent gas either methane or ammonia, and source pressure 0.2 torr. Mass spectra were acquired from m/z 250-500 at 1 scan/sec. Data reduction and analysis were conducted using SIDAR software. Samples were introduced into the ionization chamber by a direct insertion probe on a platinum filament, which was heated rapidly (current gradient of 50-550 mA at 20 mA/sec) to desorb the sample. Low-resolution EI mass spectra were obtained using a LKB 9000 mass spectrometer (Rockville, MD) operated as follows: electron energy, 70 eV; source temperature, 270°C; accelerating potential, 3.5 kV; and trap current, 60 µA. Molecular mass measurements and collision-induced dissociation to produce product ion spectra were performed on a SCIEX API III tandem mass spectrometer (Toronto, Canada) using the heated nebulizer interface and positive-ion detection. Metabolites were introduced into the instrument by flow injection with a mobile phase of acetonitrile/3 mM ammonium acetate (1:1, v/v) at 40 µl/min. Mass spectra were acquired from m/z 150-600 at a scan rate of 1.25 sec/scan, with the orifice potential set at 60 V. NMR spectra were recorded at either 400 MHz on a Varian XL-400 spectrometer or at 500 MHz on a Varian VXR500S instrument (Palo Alto, CA) using CDCl₃ as solvent. Data were collected using a 45° flip angle and a 1-sec acquisition time. Chemical shifts are given in ppm relative to IS tetramethylsilane at 0 and the residual CHCl₃ signal set a 7.26 ppm.

Plasma Protein Binding of [¹⁴C]FIN. Fresh heparinized dog plasma samples were mixed with [¹⁴C]FIN (specific activity: 24.7 µCi/mg) at concentrations of 0.02, 0.2, 0.5, and 2.0 µg/ml, and incubated at 37°C for 30 min. Immediately after incubation, 1 ml of sample was pipetted into an ultrafiltration device (Amicon Centrifree micropartition system; Amicon, Danver, MA) and then centrifuged at 2000g for 1 hr. An aliquot of the filtrate (plasma water) and an aliquot of the initial plasma to which the drug had been added were assayed for radioactivity. fu was calculated from the ratio of drug concentration in the ultrafiltrate to the initial drug concentration in the plasma; the ratio was expressed as a percentage (fu%). Nonspecific binding of [¹⁴C]FIN was determined using the same method in the absence of plasma.

Data Analysis. The principal pharmacokinetic parameters were estimated by model-independent methods from plasma concentration-time data. C_max and t_max were the observed values. k_el was determined by least squares linear regression of the terminal concentration-time data points (7-24 or 48 hr); t_1/2 was calculated as 0.693/k_el. AUC from time of dosing to the 24- or 48-hr sampling time was determined by the trapezoidal rule and was extrapolated to infinite time by addition of the term C/k_el, where C is the concentration at the last quantifiable sampling time. After correcting for dose, the oral bioavailability of FIN was calculated as AUC_po/AUC_iv. AUMC was calculated from the plasma data using the trapezoidal rule and extrapolated to time infinity (23). MRT was determined from the intravenous data by the expression MRT_iv = AUMC_0-/AUC_0-, CL_p = Dose/AUC, and Vd_ss = CL_p · MRT_iv. CL_b of unchanged FIN was estimated from the fecal data by the expression CL_b = FIN_0-t/Plasma AUC, wherein FIN_0-t is the amount (µg) of FIN excreted in feces over 24 or 72 hr.

	Results

Abstract Introduction Results Discussion References

Concentrations of Total Radioactivity, FIN, and Metabolites in Plasma: Pharmacokinetic Studies. At all time points and for all dose groups, HPLC analysis (method 1) showed FIN was a major component in plasma, and -OH-FIN was the major metabolite. HPLC analysis (method 3) of AFR showed -carboxy-FIN was a major metabolite; the other (more polar) metabolite observed has not been identified. Composite plasma profiles to total radioactivity, concentrations of FIN, and -OH-FIN are shown in fig. 2. Five minutes after completion of the 30-min 5 mg/kg iv infusion, the concentration of FIN was 3.1 µg/ml and accounted for most (84%) of the plasma radioactivity. Plasma concentrations of FIN exhibited a multiexponential pattern of elimination, with an initial half-life of 1.6 hr and a terminal half-life of 3.4 hr observed over the 7- to 24-hr interval. CL_p was 4.8 ml/min/kg, Vd_ss was 1.1 liter/kg, MRT was 4.2 hr, and AUC was 20.4 µg · hr/ml. As drug levels fell and contributions of FIN to plasma radioactivity declined with time, concentrations of -OH-FIN increased in a time-dependent manner, accounting for nearly 50% of the radioactivity at 7 hr.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Plasma concentration-time profiles of total radioactivity equivalents (squares), FIN (circles), and the metabolite -OH-FIN (diamonds) in dogs after administration of [14C]FIN. Each concentration on the curve represents the mean ± SD of three dogs dosed orally at 10 mg/kg (top) and the average values of two dogs dosed intravenously at 5 mg/kg (bottom).

Characterization of Plasma Radioactivity. After either dosing route and for up to 7 hr postdose, the ³H/¹⁴C-labeled radioactivity in plasma exhibited similar characteristics. Procedural recoveries and quantification of FIN and its metabolites based on either radioisotope were essentially the same. A time-dependent increase in the amount of volatile tritium present in plasma has been observed after administration of [³H]FIN. In the present study, when plasma samples from the two dogs given 10 mg/kg [³H]/[¹⁴C]FIN either orally or intravenously were passed through a C₁₈ Sep-Pak, the fraction of radioactivity not adsorbed increased from <3% over the 5-min to 7-hr interval and to >15% at 24 hr and >70% at 48 hr, with a less than corresponding increase in ¹⁴C-labeled material. Tritium radioactivity declining slowly in plasma was not due to polar metabolites, but rather tritiated water; the mechanism for its formation from [³H]FIN is unknown. The small but increasing percentage of ¹⁴C-label not retained on the cartridge suggested the time-dependent formation of very polar metabolites.

Plasma Protein Binding. The in vitro protein binding of [¹⁴C]FIN in dog plasma seems to be independent of concentration over the range (0.02-2.0 µg/ml) investigated. Mean fu was 17.8 ± 0.8%. Nonspecific binding of FIN to the ultrafiltration device could not be determined due to the hydrophobicity of the compound. In the absence of plasma, relevant concentrations of FIN could not be maintained in an aqueous solution.

Excretion of [¹⁴C]FIN and Its Metabolites. The recoveries of total radioactivity in urine and feces are presented in table 2. For all dose groups, most of the administered radioactivity was excreted in the feces over a 72-hr period, with the majority excreted during the first 48 hr. Less than 5% of the oral doses and <10% of the intravenous dose were excreted renally. Radioactivity excreted in urine and feces was characterized by its solvent-partitioning properties. Results indicated that the major urinary metabolite excreted after the 10 mg/kg oral dose were divided between neutral (44%) and polar (i.e. not extractable) compounds (39%). After the 80 mg/kg oral dose, renal excretion of neutral material increased to >70%, with a concomitant decrease in the polar fraction of radioactivity (17%). At both dose levels, only 12% of the urinary radioactivity was characterized as acidic, which is consistent with the small amount of acidic fraction radioactivity found in dog plasma. After intravenous dosing, dogs exhibited a similar pattern of urinary excretion, but inverse amounts of neutral and acidic metabolites. The major fraction (94%) of fecal radioactivity excreted over the 0- to 72-hr interval was characterized as neutral, whereas 4.5% was acidic and only 1.5% was nonextractable. Partially fractionated radioactivity in urinary and fecal extracts was also characterized according to its chromatographic properties. Representative urinary (0-72 hr) and fecal (24-48 hr) profiles of neutral metabolites are shown in fig. 3, and quantitative information on the cumulative excretion of FIN and its major neutral and acidic metabolites is presented in table 3 (urine) and table 4 (feces).

View larger version (27K): [in this window] [in a new window]   Fig. 3.   Typical radiochromatograms from HPLC separation (isocratic method 1) of samples containing the neutral fraction of radioactivity extracted from (top) 0- to 72-hr urine and (bottom) from 24- to 48-hr feces obtained from dogs after oral administration of [14C]FIN at 80 mg/kg. Parent drug and its metabolites are shown on the chromatograms as FIN and I, and II, and IV, respectively. Asterisk shows the retention time (fractions), wherein a mixture of the di-OH metabolites (V, VI, and VII) elute.

Urine. Unlike plasma, urine contained only a small amount of parent drug (0.02-0.15% of the dose) after the 10 and 80 mg/kg oral doses. Radiochromatographic profiles demonstrated extensive metabolism and were qualitatively similar after oral and intravenous dosing. Use of one column (method 1) resolved the monohydroxy metabolites (the -OH and 6-OH derivatives of FIN), but only partially resolved the neutral, dihydroxylated metabolites eluting between 4-5 min. When radioactivity in these column fractions was purified further by use of serial columns (method 4), the metabolite profiles were relatively complex, with three dihydroxylated metabolites identified (6-OH, 6-OH and 15-OH derivatives of -OH-FIN) and a number of trace metabolites unidentified. -Carboxy-FIN was the major metabolite present in the acidic urinary extracts. Treatment of the polar urinary fraction with -glucuronidase resulted in a substantial increase in the amount of neutral radioactivity. The metabolic profile of the latter was essentially the same as that obtained from the original free fraction of neutral metabolites. Most of the radioactivity obtained as consequence of the enzymatic hydrolysis eluted at the retention time of -OH-FIN, with a minor fraction corresponding to the diols. Hydrolysis was partially inhibited by D-saccharo-1,4-lactone, a specific inhibitor the -glucuronidase. Spontaneous hydrolysis of urinary radioactivity resulted in an increase in the quantity of diol metabolites. Results indicated the presence of stable and labile conjugates in urine.

Feces. The majority of administered radioactivity was excreted in the feces. Unchanged FIN was a prominent component in all fecal samples, accounting for 5% of the 10 mg/kg oral dose and increasing to 53% (dog 1) and 77% (dog 2) when the dose was increased to 80 mg/kg. Consistent with these data and also reflecting the increase in dose, CL_b of unchanged FIN in dog 1 increased from 0.4 to 2.5 ml/min/kg, whereas after administration of the high dose to dog 2, the CL_b value was 7 ml/min/kg. HPLC results showed that the neutral and acidic metabolites present in fecal extracts corresponded to those observed in urine.

Identification of Metabolites. In dogs, FIN was transformed to at least six metabolites that were detected in various amounts by radiochromatographic analysis of plasma, urine, and feces samples. These metabolites were isolated and purified to allow comparison with authentic standards for structural identification (see Materials and Methods). Structures of metabolites of FIN and the proposed pathways are shown in fig. 1. Where possible, LC/MS/MS fragmentation further supported the structures of metabolites after the pattern of product ions was compared with that produced by reference compounds or previously identified metabolites. As summarized in table 5, MS results show ions consistent with the metabolite structures. Spectral features of FIN and its metabolites are described herein.

FIN. The most highly retained radiochromatographic component in plasma, urine, and feces was determined to be FIN. The EI/MS spectrum of FIN showed the [M]^+· ion at m/z 372 and an ion m/z 357 consistent with the loss of the methyl group [M-CH₃]⁺ and ions at m/z 317 [M-C₄H₇]⁺ and m/z 300 [M-NHC(CH₃)₃]⁺ resulting from the stepwise fragmentation of the side chain. Cleavage of the entire side chain resulted in a diagnostic fragment ion at m/z 272 corresponding to an unaltered tetracyclic nucleus. The positive-ion DCI/MS and LC/MS spectra exhibited the [M+H]⁺ ion at m/z 373. No fragment ions were produced during the LC/MS ionization process, whereas the MS/MS spectrum showed an intense product ion at m/z 317. Loss of ring A (C1-C3 and the NH group) from the molecular ion yielded the ions observed at m/z 305 and m/z 69. The chemical shifts obtained in the NMR spectrum were: 6.79 d, 10.0, 1H, H1; 5.82 dd, 10.0, 1H, H2; 5.15 br s, 1H, NH; 5.07 br s, 1H, NH; 3.34 m, 1H, H5, 1.36 s, 9H, (CH₃)₃; 0.98 s, 3H, 19-CH₃; 0.70 s, 3H, 18-CH₃.

-OH-FIN (I). -OH-FIN was the principal metabolite in plasma and a major metabolite in urine and feces. Its EI/MS spectrum showed a mean [M]^+· at m/z 388, 16 mass units greater than that of FIN and contained ions at m/z 370 [M-H₂O]^+·, m/z 357 [M-CH₂OH]⁺, m/z 317 [M-C₄H₇O]⁺, m/z 300 [M-NHC(CH₃)₂CH₂OH]⁺, and m/z 272the ion indicating no transformation had taken place on the ring system. The positive-ion DCI/MS and LC/MS spectra exhibited the [M+H]⁺ ion at m/z 389. MS/MS of the parent ion gave a product ion spectrum containing a base ion at m/z 272, as well as ions at m/z 317, m/z 73, and m/z 321the latter resulting from cleavage through ring A. MS fragmentation patterns indicated that oxidation had occurred on one of the methyl groups of the t-butyl side chain. The NMR spectrum is distinguished from that of FIN by a CH₂OH and two methyl peaks at 1.30 ppm and 1.28 ppm. These findings are consistent with hydroxylation of one of the t-butyl methyl groups. In other resects, the spectrum closely resembles that of FIN. NMR spectral signals for the metabolite were observed at 6.78 d, 10.0, 1H, H1; 5.82 dd, 10.0, 2.0, 1H, H2; 5.32 br s, 1H, NH; 5.16 br s, 1H, NH; 3.60 dd, ~10.0, 6.0, 1H: 3.57 dd, ~10.0, 6.0, 1H, -CH₂OH; 1.30 s, 3H, -CH₃; 1.28 s, 3H, -CH₃; 0.98 s, 3H, 19-CH₃; 0.70 s, 3H, 18-CH₃. Spectral data and the HPLC retention time (using method 1) of this metabolite matched that of an authentic standard of -OH-FIN.

6-OH-FIN (II). This was a minor metabolite. Its EI/MS spectrums showed an intense ion at m/z 388 [M]^+· and contained fragment ions at m/z 373 [M-CH₃]⁺, m/z 357 [M-CH₃-H₂O]⁺, m/z 333 [M-C₄H₇]⁺, and m/z 316 [M-NHC(CH₃)₃]⁺. The ion observed at m/z 288 indicated the presence of an oxidized ring system, whereas the ion at m/z 270 was consistent with the loss of the elements of water from m/z 288. The exact position and stereochemistry of the hydroxyl group was established by NMR spectroscopy. Analysis was conducted using a specimen of the urinary metabolite, because authentic material was not available. Key NMR features include the appearance of a new signal at 3.75 ppm consistent with a CHOH and the loss of the smaller of the two coupling constants that characterize a normal H5 resonance. These two findings provided strong evidence for replacement of the C6 equatorial proton by OH. NMR spectral signals for the metabolite were observed at 6.74 d, 10.0, 1H, H1; 6.05 br s, 1H, NH; 5.84 dd, 10.0, ~2, H2; 5.10 br s, 1H, NH; 3.75 m, 1H, H6; 3.15 d, 10.0, 1H, H5; 1.35 s, 9H, (CH₃)₃; 0.99 s, 3H, 19-CH₃; 0.70 s, 3H, 18-CH₃. The HPLC retention time and MS fragmentation patterns of a radiolabeled component isolated from plasma were essentially identical to those of the urinary metabolite, and thus it was postulated that 6-OH-FIN was present as a minor metabolite in plasma. Subsequently, the same metabolite was identified in vitro by Ishii et al. (18).

-Carboxy-FIN (I). Formation of this carboxylic acid metabolite presumably involves further oxidation of the hydroxymethyl group of -OH-FIN. The presence of -carboxy-FIN in the acidic fraction of plasma radioactivity and in urinary and fecal extracts (obtained by extraction of samples with methylene chloride at pH 2) was inferred by HPLC retention time comparison (using method 3) with the synthetic standard. The EI/MS of the metabolite showed an [M]^+· at m/z 402, 30 amu greater than that of FIN, and contained ions m/z [M-COOH]⁺ and at m/z 272 indicating the lack of metabolic change on the tetracycline ring system. The positive-ion DCI/MS and LC/MS spectra exhibited [M+H]⁺ ions at m/z 403. MS/MS of the parent ion gave a product ion spectrum containing an ion at m/z 272 and many of the ions produced from FIN, as well as the ion at m/z 335, diagnostic for cleavage of this molecule through ring A. The NMR spectrum was characterized by the presence of two methyl peaks at 1.58 ppm and 1.57 ppm and by the absence of the t-butyl signal. In other respects, the spectrum closely resembled that of the parent drug. Compared with the spectrum of the -CH₂OH metabolite, further downfield displacement of the remaining two methyls of the original t-butyl group would be expected by a carboxy group. NMR signals were observed at 6.73 d, 10.0, 1H, H1; 5.81 dd, 10.0, ~2.0, 1H, H2; 5.65 br s, 1H, NH; 3.33 m, 1H, H5; 1.58 s, 3H, -CH₃; 1.57 s, 3H, -CH₃; 0.97 s, 3H, 19-CH₃; 0.70 s, 3H, 18-CH₃.

Dihydroxylated Metabolites (V-VII). These compounds were found as unconjugated metabolites obtained by extraction of urine and feces with methylene chloride at pH 7, and were liberated following spontaneous and/or enzymatic hydrolysis of the polar urinary radioactivity. Taken together, they accounted for a substantial amount of radioactivity. Dihydroxy metabolites also were observed in dog plasma (fig. 4); however, they were isolated from the acidic fraction of radioactivity that suggests they, too, may have been present initially as labile (possibly sulfate) conjugates. Urinary HPLC (method 1) isolate containing metabolites V, VI, and VII, when examined under the second HPLC system (method 4) using the dual-column technique, showed that the original peak was a mixture of at least three components in one dog, and primarily one metabolite [,6-(OH)₂-FIN] in the other. Major components were examined, and three proposed structures are described in this section. The EI/MS of ,6-(OH)₂-FIN showed an intense [M]^+· at m/z 404, 16 amu greater than that of -OH-FIN, indicating that a second oxidation had taken place. MS fragmentation patterns gave ions that were characteristic of oxidation on the t-butyl side chain and as well as on the ring system. The DCI/MS and LC/MS spectra exhibited the [M+H]⁺ ion at m/z 405. MS/MS of the parent ion gave a product ion spectrum containing the base ion at m/z 270 (m/z 288-H₂O) and m/z 73, the latter corresponding to a product ion of -OH-FIN. Like ,6-(OH)₂-FIN, metabolites VI and VII showed an [M+H]⁺ ion at m/z 405 and gave similar product ion spectra, indicating that the three metabolites are isomers differing only in the position and/or stereochemistry of the hydroxyl group on the tetracyclic ring system.

View larger version (27K): [in this window] [in a new window]   Fig. 4.   Typical radiochromatograms from HPLC separation (gradient method 3) of samples containing the AFR extracted from pooled plasma samples of (top) dogs and (bottom) humans after oral administration of [14C]FIN at 10 mg/kg and 38.1 mg, respectively. Monocarboxylic acid metabolite shown as IV on the chromatograms accounted for ~1% (dog) and ~21% (human) of the total radioactivity in plasma. Human metabolic data were taken from a previous study (17).

Me FIN--oate (III). Trace amounts of this derivative of -carboxy-FIN were detected in plasma. The HPLC (method 1) isolate corresponding to the region of compound III, when examined under a second HPLC system (method 2), showed the major radiolabeled component was -OH-FIN. Negligible amounts of radioactivity coeluted at the retention time (21.2 min) of the side chain aldehyde derivative of FIN, indicating the aldehyde metabolite was not present in plasma. About 4% of the partially purified radioactivity eluted at 22 min, the retention time of the authentic methyl ester derivative of -carboxy-FIN. The EI/MS of this compound showed an [M]^+· ion at m/z of 416 (44 amu and 14 amu greater than that of FIN and -carboxy-FIN, respectively), and key features of FIN and its identified metabolites, along with fragment ions corresponding to losses of OCH₃ and COOCH₃. Its DCI/MS spectrum exhibited the [M+H]⁺ ion at m/z 417. Authentic Me FIN--oate gave the same mass spectrum as radiolabeled component III. The mechanism for its formation is unknown, and possibly it is formed ex vivo as an artifact of sample handling. Compound III was not formed by reaction of the aldehyde or carboxy derivatives of FIN in methanol. Authentic compounds were stable in methanol; however, when the carboxylic acid was treated with diazomethane in ethereal methanol, it gave Me FIN--oate.

	Discussion

Abstract Introduction Results Discussion References

FIN was labeled with carbon-14 at the tertiary carbon of the t-butyl side chain, because previous studies in animals (data not shown) and results obtained subsequently in humans confirmed that the label position was stable metabolically (17). HPLC-based assays allowed the monitoring of concentrations of drug and metabolites in plasma, urine, and feces for a minimum of 24 hr after dosing.

Absorption estimates based on plasma radioactivity (equivalents) indicated that FIN at 10 mg/kg was well absorbed. In contrast, the 80 mg/kg dose when given to the same two dogs was absorbed slowly; peak levels of radioactivity and parent drug were reached in 4 hr in one dog, but not until 30 hr in the other. It is possible that, when a large amount of FIN is administered, its low aqueous solubility may increase its dissolution time, resulting in a reduction in the rate of absorption. Thus, the 8-fold increase in dose resulted in less than proportional increases in C_max values for both radioactivity and FIN. On the other hand, it seems the higher dose was reasonably well absorbed in both animals, as indicated by the nearly proportionate increase in AUC values of total radioactivity and FIN.

FIN was eliminated rapidly in all dose groups, with the majority of the administered radioactivity excreted in feces during the first 48 hr, providing indirect evidence of excretion via the bile. Only a small fraction of the oral and intravenous doses, 5 and 10%, respectively, was excreted renally. Elimination of FIN was governed entirely by metabolism, as negligible amounts of parent drug were detected in urine after either dosing route or in feces of intravenous-dosed dogs. After oral dosing, 5% of the 10 mg/kg dose was excreted in feces as intact FIN, confirming the lower dose was well absorbed. On the other hand, a large amount (53-77%) of the 80 mg/kg dose was excreted unchanged in contrast to that observed at the 10 mg/kg dose, which was eliminated mainly as metabolites. These results suggested that, at the high dose, a saturation of metabolic pathways may have occurred and was compensated for by increased biliary excretion of intact FIN. Investigation of the underlying mechanism for the difference in patterns between the 10 and 80 mg/kg doses was beyond the scope of the present study.

In dog plasma, ~82% of [¹⁴C]FIN was bound to protein with no indication of concentration dependency over the range 0.02-2.0 µg/ml; whereas at that range in human plasma, protein binding was higher (90%) and seemed to be slightly concentration-dependent (17). In both species, most of the carbon-14 measured in blood was associated with the plasma. The blood-to-plasma ratio of radioactivity remained constant in samples up to 24 hr postdose, indicating that the biotransformation of FIN did not give rise to metabolites that changed the selective distribution of radioactivity to either the plasma or red blood cells.

Fractionation of the urinary and fecal radioactivity from dogs into neutral and acidic metabolites provided a basis for comparing the difference in the excretion patterns and profiles of FIN metabolites exhibited by dogs with those observed in humans (17). Also, preliminary purification of the sample by solvent extraction was an advantage, because it would have been more difficult to achieve chromatographic separation of positional isomers (dihydroxylated metabolites V and VII) in the presence of the acidic metabolite(s).

At all time points and for all dose groups, parent drug was a major component in plasma with -OH-FIN (metabolite I) as the major metabolite. This also was observed in humans after oral administration of FIN. Plasma concentration profiles of unchanged FIN after intravenous dosing showed that the drug declined plasma in a multiexponential fashion. The relatively low value of Vd_ss (1.1 liter/kg) suggests that FIN was not extensively distributed to the tissues. Peak concentrations of -OH-FIN after the oral dose were reached ~4 hr later than those of FIN, suggesting the metabolite was formed subsequent to the first pass. AUC values for FIN and the -OH metabolite were similar after intravenous and oral dosing at 10 mg/kg, indicating a low first-pass effect that is in agreement with the high bioavailability determined from dose-normalized plasma data. The metabolic pathway may have been started with the high dose, as evidenced by the small increase in the amount of plasma radioactivity accounted for as parent drug and the small decrease accounted for as -OH-FIN.

When 38 mg of [¹⁴C]FIN was administered to human subjects in a metabolism study, human plasma contained an AFR, accounting for ~27% of the total radioactivity, with the major acidic component identified as -carboxy-FIN, metabolite IV with an AUC value of 1.4 µg · hr/ml (17); however, canine plasma exhibited a minor AFR (<4%), with the -carboxy metabolite accounting for ~1%. Plasma metabolic profiles from the dog and human are compared in fig. 4. The possible difference in systemic exposure to this acidic material between humans at the therapeutic dose of 5 mg and dogs undergoing toxicity testing at 80 mg/kg prompted the metabolism study in dogs at 80 mg/kg. An 8-fold increase in dose from 10 to 80 mg/kg resulted in less than proportional increases in the AUC values obtained for AFR, but no change was observed in the percentage of plasma radioactivity identified as -carboxy-FIN. It can be concluded from these data that, at 80 mg/kg of FIN, systemic levels of the -carboxy metabolite in dogs were nearly 3-fold higher than those found circulating in the plasma of humans who received 38 mg of FIN in the metabolism study, and it was estimated that the levels of metabolite circulating in the dog may be ~20-fold higher than those present in plasma of clinical subjects receiving 5 mg of drug.

FIN was metabolized extensively through oxidative pathways (fig. 1). The major biotransformations included hydroxylation of the t-butyl side chain to give -OH-FIN with subsequent oxidation of the side chain to -carboxy-FIN or hydroxylation of the tetracyclic ring system to the major dihydroxylated,6-(OH)₂, ,6-(OH)₂, and ,15-(OH)₂derivatives of FIN. Hydroxylation of the t-butyl side chain is not uncommon and, for example, was observed for timolol by Tocco et al. (24). The pathway leading to the formation of -carboxy-FIN involves formation of an aldehyde intermediate, which is then metabolized to the acid. The in vitro formation of the aldehyde from FIN was detected by Ishii et al. (18) using rat liver microsomal preparations and by Huskey et al. (19) using human liver microsomal preparations. By use of human liver microsomes containing recombinant CYP enzymes, Huskey et al. demonstrated that CYP3A catalyzed each step of the biotransformation of FIN to -carboxy-FIN. Although we were able to confirm the formation of the aldehyde intermediate in vitro using rat liver microsomes (unpublished data), we were not able to detect this metabolite in vivo (plasma).

The 6-OH derivative of FIN was identified as minor metabolite in dog plasma, whereas the 6- and 6-hydroxylated metabolites were identified as the additionally hydroxylated derivatives of -OH-FIN. Derivatives of FIN hydroxylated at the 6-position were detected in human samples (15) and in the rat (16). The stereospecificity in hydroxylation of FIN at the 6-position differs from that of testosterone which, in most species, including the dog (25, 26), is usually hydroxylated at the 6-position. In the rat, 6-hydroxylation is conducted by the members of CYP3A family of isozymes (27), whereas the 6-OH metabolite has been found as a minor metabolite of testosterone catalyzed by CYP2A1 (28). Characterization of the isozyme responsible for the hydroxylation of FIN at the 6-position rather than 6-position of the ring system is beyond the scope of the present study, but the explanation may involve the difference in the conformation of the A-B ring of the azasteroid, compared with that of testosterone, resulting in the  side of FIN being more accessible to the enzymes.

In summary, the absorption, plasma distribution, metabolism, and elimination of FIN have been studied in male beagle dogs after oral (10 mg/kg) and intravenous (5 mg/kg) administration, and also after the oral route at the dose level (80 mg/kg) used in the toxicity testing of the compound. The high systemic bioavailability of FIN indicates that the drug is well absorbed and the first-pass metabolism is low. Despite quantitative differences in the amounts of neutral vs. acidic metabolites formed, we have shown the drug is metabolized by dog and humans via similar pathways. The physiological disposition of FIN in the dog seems to be a reasonable paradigm for humans.