QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Perkins, E. J. Articles by Cassidy, K. C. Search for Related Content PubMed PubMed Citation Articles by Perkins, E. J. Articles by Cassidy, K. C.

PRECLINICAL CHARACTERIZATION OF 2-[3-[3-[(5-ETHYL-4'-FLUORO-2-HYDROXY[1,1'-BIPHENYL]-4-YL)OXY]PROPOXY]-2-PROPYLPHENOXY]BENZOIC ACID METABOLISM: IN VITRO SPECIES COMPARISON AND IN VIVO DISPOSITION IN RATS

Lilly Research Laboratories, Indianapolis, Indiana

(Received March 13, 2003; Accepted August 11, 2003)

	Abstract

Top Abstract Materials and Methods Results Discussion References

 
Assessment of the pharmacokinetics of [¹⁴C]2-[3-[3-[(5-ethyl-4'-fluoro-2-hydroxy[1,1'-biphenyl]-4-yl)oxy]propoxy]-2-propylphenoxy-]benzoic acid ([¹⁴C]LY293111), an experimental anti-cancer agent, suggested long-lived circulating metabolites in rats. In vivo metabolites of LY293111 were examined in plasma, bile, urine, and feces of Fischer 344 (F344) rats after oral administration of [¹⁴C]LY293111. Metabolites were profiled by high-performance liquid chromatography-radiochromatography, and identified by liquid chromatography (LC)/mass spectrometry and LC/NMR. The major in vivo metabolites of LY293111 identified in rats were phenolic (ether), acyl, and bisglucuronides of LY293111. Measurement of radioactivity in rat plasma confirmed that a fraction of LY293111-derived material was irreversibly bound to plasma protein and that this bound fraction increased over time. This was consistent with the observed disparity in half-lives between LY293111 and total radioactivity in rats and monkeys, and is likely due to covalent modification of proteins by the acyl glucuronide. In vitro metabolism of [¹⁴C]LY293111 in liver slices from CD-1 mice, F344 rats, rhesus and cynomolgus monkeys, and humans indicates that glucuronidation was the primary metabolic pathway in all species. The acyl glucuronide was the most prevalent radioactive peak (16% of total ¹⁴C) produced by F344 rat slices, whereas the ether glucuronide was the major metabolite in all other species (26-36% of total ¹⁴C). Several minor hydroxylated metabolites were detected in F344 rat slice extracts but were not observed in other species. The data presented suggest that covalent modification of proteins by LY293111 acyl glucuronide is possible in multiple species, although the relative reactivity of this metabolite appears to be low compared with those known to cause adverse drug reactions.

Initial preclinical excretion and pharmacokinetic studies using radiolabeled LY293111 indicated that its oral bioavailability is limited partially by its poor absorption in rats and monkeys, with about half of an oral dose excreted in feces of biliary cannulated rats (Cramer et al., 1995). Furthermore, plasma concentrations of total radioequivalents did not correspond directly to LY293111 concentrations in rats and monkeys dosed with [¹⁴C]LY293111, suggesting the occurrence of significant metabolism in vivo. Therefore, it was hypothesized that first-pass metabolism may also significantly limit LY293111 bioavailability in laboratory animals. The metabolic fate of LY293111 was not examined in these early studies.

Glucuronidation of carboxylic acid-containing drugs is a frequent occurrence in vivo, resulting in the formation of ß-1-O-acyl glucuronides (Bolze et al., 2002). These reactive metabolites have been widely demonstrated to undergo chemical hydrolysis and intramolecular rearrangement (acyl migration), which can lead to irreversible binding to endogenous protein nucleophiles through transacylation and glycation mechanisms (Bailey and Dickinson, 2003). Because of the relatively large molecular weight of LY293111 (mol. wt. = 544.62) and its free phenolic and carboxyl groups, direct phase II conjugation and biliary excretion were predicted to account for at least a portion of its metabolic fate.

In the current studies, the pharmacokinetics, excretion, and in vivo metabolism of LY293111 was investigated in rats. In vitro comparison of metabolites was also conducted in rat, mouse, monkey, and human liver slices. Because of the difficult nature of structural elucidation of glucuronide metabolites, liquid chromatography/nuclear magnetic resonance (LC/NMR) was utilized in addition to the more standard analytical techniques.

	Materials and Methods

Top Abstract Materials and Methods Results Discussion References

 
Reagents. Methanol and acetonitrile (HPLC reagent grade) were purchased from Mallinckrodt (St. Louis, MO). Ultima-Flow M and Ready Protein⁺ liquid scintillation cocktails were purchased from PerkinElmer Life Sciences (Boston, MA) and Beckman Coulter (Fullerton, CA), respectively. Waymouth's 752/1 medium without phenol red or glutamine was purchased from Cell and Molecular Technologies, Inc. (Phillipsburg, NJ). Fetal bovine serum was purchased from Harlan (Indianapolis, IN). Unlabeled LY293111 sodium was synthesized at Lilly Research Laboratories using previously published methods (Sawyer et al., 1995; Sawyer, 1996).

Synthesis of [¹⁴C]LY293111. ¹⁴C-labeled LY293111 ([¹⁴C]LY293111) was synthesized in a five-step convergent procedure utilizing CuCN, K¹⁴CN, and 1-fluoro-2-iodobenzene to produce the initial ¹⁴C-containing reactant, ¹⁴C-1-fluorobenzonitrile. Coupling of ¹⁴C-1-fluorobenzonitrile with 2-propylbenzene-1,3-diol formed the desired ¹⁴C-labeled 2-(3-hydroxy-2-propylphenoxy)-benzonitrile, which was reacted further with 2-benzyloxy-4-(2-chloro-ethoxy)-5-ethyl-4'-fluorobiphenyl to afford ¹⁴C-labeled 2-[3-[3-(2-benzyloxy-5-ethyl-4'-fluorobiphenyl-4-yloxy)-2-propyloxy]-2-propylphenoxy]benzonitrile. Debenzylation and subsequent hydrolysis of the [¹⁴C]nitrile functionality produced [¹⁴C]LY293111 sodium. [¹⁴C]LY293111 sodium coeluted with authentic LY293111 by HPLC using a Zorbax RX C8 column at 30°C, with a linear gradient of 82:18 methanol/0.01 N H₂SO₄ to 90:10 methanol/0.01 N H₂SO₄ over 35 min (flow rate = 1 ml/min). Radiochemical purity was determined to be greater than 97% by HPLC with radiochemical detection.

Rat Pharmacokinetic Studies. [¹⁴C]LY293111 was administered to male Fischer 344 (hereafter called F344) rats (190-250 g; Harlan) as either a single 10 mg/kg (47 µCi/kg) i.v. dose or a 30 mg/kg (45 µCi/kg) oral gavage (p.o.) dose. Dose solutions (4.8 mg/ml i.v.; 7.5 mg/ml p.o.) were prepared in purified water using the sodium salts of the labeled and unlabeled compounds. Blood samples were collected into heparinized tubes via cardiac puncture from three to four rats per time point. The time points included were 0.08, 0.17, 0.25, 0.5, 1, 2, 4, 8, 12, 18, 24, 36, 48, 72, and 96 h in the intravenous dose group and 0.5, 1, 2, 3, 4, 8, 12, 18, 24, 36, 48, 72, and 96 h in the oral dose group. Plasma was separated by centrifugation and frozen at -70°C until assayed.

Quantitation of total ¹⁴C in dosing solutions and plasma was conducted by liquid scintillation counting (LSC) with external standardization to correct for quenching. The plasma concentration of LY293111 was determined by reversed phase HPLC using a Zorbax RX C8 column (250 x 4.6 mm) with UV detection at 255 nm. Samples were prepared by acetonitrile precipitation and centrifuged to remove particulate matter. The supernatant was transferred to a clean tube, evaporated to dryness under nitrogen, and reconstituted in mobile phase (acetonitrile/methanol/0.125 M ammonium phosphate buffer, pH 2.35; 65:10:25, v/v). Chromatography was carried out under isocratic conditions with a flow rate of 0.8 ml/min. The limit of quantitation was 0.005 µg/ml. Pharmacokinetic parameters were calculated by noncompartmental analysis, using the trapezoidal rule for area under the curve calculation.

In Vivo Rat Metabolism Studies. Bile duct-cannulated male F344 rats were obtained from Taconic Farms (Germantown, NY) and housed in Nalgene metabolism cages (Nalge Nunc International, Naperville, IL). Rats were administered a single p.o. dose of [¹⁴C]LY293111 at 30 mg/kg (77 µCi/kg) in aqueous solution. Bile and urine were collected into refrigerated conical tubes containing 2 M acetic acid for 48 h after dosing. Urine samples were combined into 0- to 24-h and 24- to 48-h pools for each animal. Feces and cage wash were collected at 24 and 48 h. Rats were euthanized for carcass analysis at the end of the collection period.

Rat bile samples were diluted with an equal volume of water and centrifuged before analysis. Fecal homogenates from the rat (approximately 1.5-g aliquots of the 0- to 24-h and 24- to 48-h samples) were extracted by mixing with methanol (10 ml) for approximately 5 min, followed by centrifugation. The supernatants were removed and dried under nitrogen at 50°C until approximately 100 µl remained. An additional 400 µl of methanol/water (1:1, v/v) was added to each extract. The samples were vortex-mixed for 1 min followed by centrifugation for 5 min.

To obtain sufficient plasma samples for metabolite profiling and assessment of covalent plasma protein binding, a separate group of male F344 rats was administered a 30 mg/kg (100 µCi) oral dose of [¹⁴C]LY293111 sodium. Blood was collected into heparinized tubes from three animals per time point at 2, 8, 12, and 24 h. Plasma was isolated by centrifugation and stored at -70°C before analysis. Plasma pools (300 µl) for metabolic profiling were prepared by combining equal aliquots of each time point. Plasma proteins were precipitated with 600 µl of acetonitrile, and samples were centrifuged to remove particulate matter.

Assessment of Covalent Binding to Plasma Protein. Rat plasma pools were prepared for assessment of irreversible protein binding by combining 250-µl aliquots from each sample at each time point (n = 3). Total radioactivity was determined by LSC from a 100-µl aliquot of the pooled plasma. Aliquots of the plasma (500 µl) were subsequently extracted with acetonitrile (1 ml) and centrifuged for 5 min. The residual radioactivity in the supernatant was determined by LSC. The remaining protein pellet was washed twice (with vortexing) with 1 ml of acetonitrile/water (2:1, v/v) and centrifuged. Each wash supernatant fraction was assessed by LSC. To the remaining pellet, 0.75 ml of water was added and vigorously vortexed. The sample was heated in a 60°C water bath for 1 h to dissolve the pellet and quantitatively transferred to a scintillation vial. Ready Protein⁺ (PerkinElmer Life Sciences) scintillation fluid (15 ml) was added to each scintillation vial prior to LSC.

In Vitro Metabolism Studies. Precision-cut liver slices were prepared from livers of human (n = 2), rat (pooled, n = 3), cynomolgus monkey (n = 1), rhesus monkey (n = 1), and mouse (pooled, n = 9). Livers were collected from anesthetized male F344 rats (17-19 weeks old; Harlan) and CD-1 mice (6 weeks old; Harlan), and were immediately placed in ice-cold saline before preparation of slices. Livers from adult male cynomolgus and rhesus monkeys were collected after phenobarbital overdose and were chilled in ice-cold modified Sack's preservation solution (Monden and Fortner, 1982). Two human liver samples were obtained from Tissue Transformation Technologies (Edison, NJ). Human donor 045-22445 was a 64-year old Asian female smoker. Human donor 045-23079 was a 60-year old white male smoker. The cause of death for both human donors was listed as intracranial hemorrhage. The donor livers were obtained for possible transplantation and therefore were cold-perfused with University of Wisconsin (UW) organ preservation solution to preserve tissue viability. Liver samples were shipped on ice in UW solution, and slices were prepared immediately upon receipt.

Cylindrical tissue cores were prepared by passing a sharpened stainless steel tube through the liver tissue with a motorized coring press. Slices were prepared with a Brendel/Vitron tissue slicer (Vitron, Tuscon, AZ) to a thickness of approximately 200 to 250 µm. Two slices per species were incubated with 50 µM [¹⁴C]LY293111 in a dynamic roller culture incubator at 37°C under an atmosphere of 95% O₂/5% CO_2. Samples were collected at 4 or 24 h, homogenized by sonication, quick-frozen on dry ice, and stored at -70°C before analysis. A portion of the 24-h incubation homogenates was divided into aliquots (0.25 ml) that were further incubated at 37°C overnight (approximately 16 h) in the absence or presence of Sigma Type HP-2 ß-glucuronidase (25 µl, 2,755 units) to assess the stability of glucuronidated metabolites.

Homogenized in vitro samples were prepared for analysis by addition of 2.5 ml of 90:10 acetonitrile/isopropanol to 0.25 ml of liver slice sample. The samples were vortexed and centrifuged to remove particulate matter. The supernatants were dried under nitrogen and reconstituted in 0.25 ml of acetonitrile/10 mM ammonium acetate (20:80, v/v).

Metabolism of 7-ethoxycoumarin was used as a positive control to demonstrate the capacity of the liver slice preparations to retain integrated phase I and phase II metabolism (Barr et al., 1991). Briefly, liver slices were incubated with 50 µM 7-ethoxycoumarin for 3 h at 37°C, and incubations were terminated by homogenization. Samples were quick-frozen on dry ice and stored at -70°C before analysis. A previously described fluorometric method (Steemsa et al., 1994) was used to determine concentrations of 7-hydroxycoumarin (7-HC), 7-HC-O-sulfate, and 7-HC-O-glucuronide (Table 1).

HPLC/¹⁴C Analysis. The HPLC/¹⁴C analysis of in vitro and in vivo metabolism samples was conducted using an LC-10A binary gradient HPLC system equipped with an SIL-10A autosampler (Shimadzu Scientific Instruments Inc., Columbia, MD) and either a PerkinElmer Life Sciences 500TR or a Berthold LB509 (Berthold, Nashua, NH) radiochemical detector equipped with a 500-µl liquid cell. Separation of parent compound and metabolites was achieved using a 10 mM ammonium acetate (A)/methanol (B) gradient on a Supelco Discovery C18 column (4.6 x 250 mm, 5-µm particle size). A typical gradient profile consisted of 55% B for 3.5 min, then using a linear gradient to 80% B at 43.5 min, and ramping to 90% B at 60 min. The flow rate was 1 ml/min at room temperature. Ultima-Flow M liquid scintillation cocktail was added to the effluent at 3 ml/min through an in-line mixing tee before entering the radiochemical detector.

Because of low radioactivity in plasma extracts, metabolic profiling of plasma was obtained by HPLC with microplate solid scintillation counting. Plasma extracts were separated by HPLC, and the column effluent was collected at 20-s intervals into 96-well LumaPlates (PerkinElmer Life Sciences). The plates were counted off-line by microplate solid scintillation counting (PerkinElmer Life Sciences TopCount) and the results were plotted to obtain radiochromatograms.

LC/MS-MS Analysis. LC/MS-MS analysis was conducted using the HPLC conditions described, with the HPLC flow split between the radiochemical detector and mass spectrometer. The analysis was conducted on a Finnigan LCQ DECA mass spectrometer (ThermoQuest, San Jose, CA) using positive-ion ESI and positive-ion APCI. For ESI analysis, the capillary heater was set at 225°C and the spray voltage at 5 kV. For APCI analysis, the APCI probe was set at 350°C and the spray current at 5 µA. MS-MS was carried out with collision energy 50% and isolation width of 4.0.

Purification of Metabolites from Rat Bile. A Waters (Milford, MA) HPLC system consisting of a model 600 pump and controller, model 2700 autosampler, model 2487 UV-visible detector, and a fraction collector was utilized for metabolite purification. Aliquots (1 ml) of bile (diluted with an equal volume of water) were separated on a Supelco Discovery C-18 column (5-µm, 10 x 250 mm) using a gradient of 10 mM ammonium acetate (A) and methanol (B) at a flow rate of 4 ml/min. The initial solvent composition of 45% A and 55% B was maintained for 3.5 min, followed by an increase of solvent B to 80% over 40 min and a further increase to 90% B at 42.5 min. The mobile phase was maintained at 90% B for an additional 13 min. Metabolites M2, M7, and M9 (Fig. 2) were isolated by collecting the column effluent in 30-s intervals from 20 to 70 min. Replicate injections were made until the entire sample of diluted bile (12 ml) was fractionated. Fractions were evaluated for purity using a Finnigan LCQ mass spectrometer equipped with a photodiode array detector. The fractions containing the metabolites of interest were combined for each metabolite and dried to approximately 4 ml. The isolated metabolites were repurified by HPLC using similar chromatographic conditions except that solvent A was 0.2% formic acid. Metabolites M2, M7, and M9 were isolated by collecting the column effluent in 20-s-interval fractions from 45 to 50 min. Fractions totaling 4 ml were combined and further repurified for each metabolite. Final samples containing M2, M7, or M9 were lyophilized to dryness. Structural confirmation of isolated metabolites was determined by LC/MS-MS and LC/NMR analyses.

View larger version (23K): [in this window] [in a new window]   FIG. 2. Proposed metabolism pathway of LY293111.

LC/NMR Analysis. Fractions containing M2, M7, and M9 were reconstituted separately in 50:50 MeOH/H₂O. The ¹H NMR for each metabolite was obtained using a "column trapping" technique in conjunction with LC/NMR. The column trapping technique required concentrating the analyte of interest on a Keystone Aquasil 30 x 3 mm C-18 column using D₂O (2 ml/min; Varian 9012 HPLC pump and 9050 variable wavelength detector; Varian, Inc., Palo Alto, CA). After removing protonated solvents, the HPLC flow was reversed, and the analyte was eluted from the column and into the LC/NMR flow cell with 90% deuterated acetonitrile and 10% deuterium oxide (1 ml/min). NMR analyses were carried out using standard pulse sequences on a Varian Unity Inova 600 MHz spectrometer equipped with a Varian triple resonance flow probe and a 60-µl flow cell.

	Results

Top Abstract Materials and Methods Results Discussion References

 
Rat Pharmacokinetics. A large difference between mean elimination half-lives (t_1/2) of LY293111 and total ¹⁴C was observed following administration of [¹⁴C]LY293111 sodium to male F344 rats (Fig. 1; Table 2). The t_1/2 of total radioactivity was 50 to 68 h, with a corresponding LY293111 t_1/2 of 2.6 to 4.4 h following oral (30 mg/kg) and i.v. (10 mg/kg) dosing, respectively. Because of the prolonged residence time of total radioactivity compared with LY293111, the mean area under the plasma concentration versus time curve for total ¹⁴C was approximately twice that of LY293111. The oral bioavailability of total radiocarbon (25%) was similar to that of LY293111 (22.9%), suggesting that absorption and/or first-pass biliary excretion may be limiting factors for oral availability of LY293111 in rats. In the orally dosed group, maximal concentrations (C_max) of LY293111 and total ¹⁴C were reached at 1 h after the oral dose. However, a second minor plasma peak was apparent at 4 h, consistent with possible enterohepatic recirculation (Fig. 1).

View larger version (16K): [in this window] [in a new window]   FIG. 1. Plasma profile of LY293111 and total LY293111-related radioequivalents in rat plasma following a single intravenous (10 mg/kg) (A) or oral (30 mg/kg) (B) dose of [14C]LY293111 sodium.

Excretion and in Vivo Metabolism in Rat. The excretion of LY293111 and its metabolites was evaluated over 48 h in bile duct-cannulated F344 rats (n = 5) given a single oral 30 mg/kg dose of [¹⁴C]LY293111 sodium (77 µCi/kg). Recovery of total radioactivity was determined by LSC of bile, feces, urine, cage wash, and the dissolved carcass. Total mean recovery of radioactivity was 93 ± 2%, with the majority of radiocarbon (85 ± 2% of the dose) excreted within the first 24 h after dosing (Table 3). One-half of the administered dose (52 ± 5%) was excreted in the bile, whereas the excretion of radioactivity in urine was minimal (0.2% of the dose), indicating that the radioactivity excreted in bile represents the entire fraction of the dose absorbed. This is in contrast to the observed radioactive bioavailability of 25% and suggests that a large fraction of the absorbed dose is eliminated via first-pass excretion in the bile.

Bile and fecal metabolites. Parent drug and six metabolite peaks (M2, M4, M7, M9, M10, and M12) were identified in rat bile (Table 4). In 0- to 6-h bile, which accounted for 31% of the dose, the majority of the radioactivity was due to three metabolites (34% M2, 32% M7, and 32% M9). These metabolites were subsequently identified as acyl, ether, and bis-glucuronides, respectively (Fig. 2, described below). The remaining radioactivity in 0- to 6-h bile consisted of the parent drug and several other minor metabolites (less than 1% of radioactivity each). Approximately 39% of the dose, assumed to represent the unabsorbed fraction, was excreted in feces of bile-cannulated rats, with no metabolites identified in fecal extracts (Table 4). The extraction efficiency of radioactivity for feces was 98%, and the total column recovery was calculated to be 107%.

Plasma metabolites. Due to the possible instability of acyl glucuronides, plasma extracts were not dried and were assayed shortly after preparation. The majority of circulating radioactivity in plasma was comprised of parent drug, the acyl glucuronide of LY293111 (M2), and the phenolic glucuronide (M7), which represented 44%, 21%, and 5% of the plasma radioactivity, respectively (Fig. 3). Minor amounts of the bis-glucuronide of LY293111 (M9) and a proposed rearrangement product of M2 (M1) were also detected by LC/MS.

View larger version (15K): [in this window] [in a new window]   FIG. 3. Representative radiochromatographic (microplate scintillation counting) profile of rat plasma (2 h postdose; pooled from three rats).

Assessment of covalent protein binding in plasma. Preliminary metabolism studies in rats suggested that extraction efficiency of [¹⁴C]LY293111-related radioactivity was incomplete and that it decreased with increasing time after dosing in rat plasma. It was hypothesized that the incomplete recovery of radioactivity was related to covalent binding of the drug-related material to plasma protein. To further investigate this hypothesis, the protein pellet was re-extracted twice and the radioactive mass balance between the plasma extract and protein pellet was examined (Table 5). After acetonitrile precipitation, recovery of radioactivity in the supernatant for 2-, 8-, 12-, and 24-h plasma decreased for each later time point; from 86% for the 2-h plasma to 63% for the 24-h plasma. Subsequent washing of the protein pellet only accounted for an additional 4 to 8% of the total radioactivity. The remaining radioactivity was detected in the protein pellet, representing 7% (2-h) to 32% (24-h) of the total radioactivity, suggesting that one or more of the LY293111 metabolites undergo covalent binding to plasma protein.

Structural Identification of Metabolites. Table 6 summarizes the HPLC retention times and LC/MS-MS results for LY293111 and metabolites observed. Characteristic product ions are listed for all of the metabolites observed.

M2 (LY293111 acyl glucuronide). The LC/MS spectrum for peak M2 showed m/z 738 as the base peak, corresponding to the addition of NH₄⁺ and glucuronic acid to LY293111. MS-MS of m/z 738 gave m/z 545 (the aglycone), confirming the addition of glucuronic acid to LY293111. The base peak ion, m/z 527, corresponds to the loss of water from the aglycone ion. Product ions characteristic for LY293111 were observed at m/z 313, 295, 273, 255, 245, and 217. The position of conjugation of M2 was determined by proton NMR (Fig. 4). The key feature observed is the anomeric proton of the glucuronide moiety, a doublet in a region of the spectrum where few proton resonances are found. The chemical shift for the anomeric proton for M2 is observed at 5.7 ppm, consistent with literature values for ester-linked glucuronides (Sidelmann et al., 1997).

View larger version (36K): [in this window] [in a new window]   FIG. 4. Expanded proton NMR spectra for LY293111, M2, M7, and M9

M7 (LY293111 ether glucuronide). The LC/MS spectrum for peak M7 gave m/z 738 as the base peak, indicating the addition of NH₄⁺ and glucuronic acid to LY293111. MS-MS of m/z 738 gave the aglycone (m/z 545) and the aglycone -H₂O (m/z 527). Product ions characteristic for LY293111 were observed at m/z 313, 295, 273, 255, and 245. Peak M7 is identified as the direct glucuronic acid conjugate of LY293111. The position of conjugation of M7 was determined by proton NMR (Fig. 4). Similar to M2, the key feature observed is the anomeric proton of the glucuronide moiety. The chemical shift of the anomeric proton for M2 is observed at 5.0 ppm, consistent with literature values for aromatic ether glucuronides (Sidelmann et al., 1997).

M8 (LY293111 bis-glucuronide). The LC/MS spectrum for peak M8 gave m/z 914 as the base peak, indicating the addition of NH₄⁺ and two molecules of glucuronic acid to LY293111. MS-MS of m/z 914 gave m/z 879, corresponding to the loss of NH₃ and H₂O. The base peak m/z 703 is generated by the loss of glucuronic acid, NH₃, and H₂O. A very weak aglycone ion was observed at m/z 545 along with product ions at m/z 527 and 295. This metabolite is identified as a bis-glucuronic acid conjugate of LY293111. The regio- and/or stereochemistry of acyl glucuronide conjugation could not be definitively determined using mass spectrometry. M8 was observed only in the 24-h mouse and primate liver slice samples and not appreciably in the 4-h samples from the other species. Since M8 and M9 are both bis-glucuronides of LY293111 and M9 was observed in both the 4-h and 24-h samples, it is therefore proposed that M8 results from the acyl migration reaction of the acyl glucuronic acid moiety of M9.

M9 (LY293111 bis-glucuronide). The LC/MS spectrum for peak M9 gave m/z 914 as the base peak, the addition of NH₄⁺ and two molecules of glucuronic acid to LY293111. MS-MS of m/z 914 gave m/z 738, corresponding to the loss of glucuronic acid. The base peak m/z 527 is formed by the loss of two glucuronic acid molecules, NH₃ and H₂O. The aglycone ion (LY293111) was observed at m/z 545. Product ions m/z 295, 273, and 255 confirm that the aglycone is LY293111. M9 was isolated from rat bile for NMR analysis (Fig. 4). The key features observed are the anomeric protons of the glucuronide moieties, doublets in a region of the spectrum where few proton resonances are found. The chemical shift of the anomeric protons for M9 are observed at 5.7 and 5.1 ppm, consistent with literature values for ester- and ether-linked glucuronides (Sidelmann et al., 1997).

Minor metabolites. Metabolites M1, M3, M4, M5, M6, M10, M11 and M12 are relatively minor (< 3% of total radioactivity); therefore, extensive characterizations were not performed for these metabolites. However, ESI-LC/MS analysis showed that M1 is a direct glucuronide of LY293111, m/z 738, and proposed to be an acyl-migrated ester-glucuronide of M2. Although oxidation of LY293111 was not a major biotransformation pathway, several minor monohydroxylated metabolites were observed. The ESI-LC/MS spectrum for M3 showed the ammonium adduct at 578 m/z, whereas APCI-LC/MS spectra for M5 and M6 showed protonated molecular ions at 561 m/z. Conjugation of M3 with glucuronic acid gave metabolites M4 (M3 monoglucuronide, m/z 754) and M10 (M3 bis-glucuronide, m/z 930). M11 (m/z 273) is formed by oxidation followed by O-dealkylation. M12 was putatively identified by ESI-LC/MS as a glucuronide conjugate of M11.

In Vitro Metabolism. The in vitro metabolite profiles of [¹⁴C]LY293111 were examined in liver slices from male Fischer 344 rat, rhesus and cynomolgus monkey, male CD-1 mouse, and humans. These data were collected to assess the in vivo predictability of the in vitro data for rats and to compare the metabolite profiles between preclinical species used for toxicology and efficacy studies to that in humans. In these studies, the metabolism of LY293111 was qualitatively similar among species, although several unique minor metabolites (M3, M4, M5, M6, M10, and M11) were observed in rat slices (Table 7). The major metabolites M2, M7, and M9 (Fig. 2) were observed in all species evaluated and were confirmed by LC/MS and LC/NMR as the acyl (ester) glucuronide, ether glucuronide, and bis-glucuronide, respectively. In all species, except the rat, the ether glucuronide was the predominant metabolite at 24 h. The acyl glucuronide ranged from 4% (monkey) to 16% (rat) of the total radioactivity in 24-h samples. Although both rhesus and cynomolgus monkey liver samples (24 h) contained relatively low levels of the acyl monoglucuronide, the amount of bis-glucuronide was high (19% 25%), with a significant amount (6%) of the proposed acyl migrated bis-glucuronide (M8) as well. Metabolite M8 was observed in all species at 24 h except the rat, while M1, a proposed rearrangement product of M2, was only observed in the 24 h rat liver slice. This suggests that acyl migration is possible with both M2 and M9. Overall recovery of radioactivity in each liver slice sample was greater than 95%.

The stability of LY293111 metabolites was assessed in 24-h rat liver slice samples by overnight incubation at 37°C with or without ß-glucuronidase (Table 8). In the absence of ß-glucuronidase, the radioactive peak area for the acyl glucuronide (M2) decreased 46%, with a corresponding increase in its putative rearrangement product (M1). Furthermore, the bis-glucuronide (M9) was no longer quantifiable, whereas the percentage area for the ether glucuronide (M7) increased in proportion to the loss of the bis-glucuronide. The amount of LY293111 was unchanged after overnight incubation. In the samples treated with ß-glucuronidase, the metabolites M1, M2, M4, M9, and M10 were no longer quantifiable, whereas the percentage area for M7 was unchanged. An increase was observed in the minor oxidative metabolites M3 and M5 after ß-glucuronidase treatment, corresponding to the loss of the hydroxylated glucuronide metabolites M4 and M9. Surprisingly, the O-dealkylation product M11 was increased 6.3 fold by incubation with ß-glucuronidase, which may explain the relatively small increase in parent drug after cleavage of the acyl glucuronide. This dealkylation was not observed in the absence of ß-glucuronidase, suggesting that this reaction is related to the enzyme preparation utilized and not due to chemical instability of the parent molecule.

These results further support the identification of these metabolites as glucuronide conjugates and indicate that the acyl glucuronides undergo degradation and rearrangement under simulated physiological conditions.

	Discussion

Top Abstract Materials and Methods Results Discussion References

 
The pharmacokinetic profiles of LY293111 and related radioequivalents in rats given oral or intravenous doses of [¹⁴C]LY293111 sodium suggested that one or more metabolites have a longer residence time than has the parent molecule. The direct conjugation of glucuronic acid to compounds containing carboxyl or phenolic functional groups is a common metabolic occurrence, with bis-glucuronides occasionally identified in compounds with both of these functional groups (Hyneck et al., 1988; Sidelmann et al., 1997; Jaggi et al., 2002). Because LY293111 contains both a free phenolic and a carboxyl group, it was assumed that direct glucuronidation at these positions was the likely pathway of biotransformation.

The excretion of [¹⁴C]LY293111-derived radioequivalents occurred almost exclusively in the bile, with about one-half of an oral dose absorbed. However, the absolute bioavailability of LY293111 was <23%, suggesting significant first-pass elimination. Furthermore, the oral bioavailability of total radioequivalents was only 25%, which suggests that this first-pass effect does not result in higher concentrations of circulating metabolites compared with an i.v. dose. Therefore, it appears that as much as 50% of the absorbed dose in rats is conjugated and excreted into the bile before reaching the systemic circulation.

Detailed examination of the in vivo metabolite profile of LY293111 revealed that glucuronidation is indeed the major pathway of biotransformation, with little oxidative metabolism observed. The acyl glucuronide was the major metabolite observed in rat plasma, whereas the relative amounts of the acyl, ether, and bis-glucuronides were similar in bile. Very little parent drug was detected in the bile, indicating that the first-pass excretion of LY293111 radioequivalents occurs subsequent to glucuronidation. This is also consistent with the observation of a secondary plasma peak of LY293111 at 4 h, due to enterohepatic recirculation.

A species comparison of in vitro metabolism of LY293111 demonstrated that glucuronidation was the primary metabolic pathway in all species examined (rat, mouse, monkey, and human). The acyl, ether, and bis-glucuronides (M7, M2, and M9/M8, respectively) were the only metabolites detected after 24-h incubations with liver slices from mice, monkeys, and humans, whereas several minor oxidative metabolites were also observed in rat samples. The qualitative pattern of in vitro metabolites from rat liver slices corresponded closely to that seen in vivo, suggesting that the liver slice data are predictive of the in vivo metabolites likely to be observed in other species.

The primary method of structural identification used in these studies was LC/MS-MS, which is clearly a technique of choice for high-throughput and rapid structural identification of metabolites. However, positional isomers of phase II metabolites (e.g., glucuronides and sulfates) can be difficult to determine by LC/MS-MS alone. The main difficulty encountered is that for phase II metabolites the major product ions most often result from the loss of a glucuronic acid moiety (176 Da) or sulfate (80 Da). As described for metabolites M2 and M7, the MS-MS spectra did not provide adequate information to definitively establish the site of glucuronic acid conjugation. The base peak in the MS-MS spectra for both metabolites was the aglycone ion (m/z 545). In this case, the definitive determination of metabolite structure was performed using ¹H NMR with column trapping (Ehlhardt et al., 1998; Mutlib et al., 2002). Due to the propensity of ester glucuronides to undergo acyl migration, the described column trapping technique was useful to minimize sample handling. After the semipreparative isolation of M2, M7, and M9, this technique negated the typical solvent removal step required for NMR analysis. Instead, the HPLC-effluent collected using semipreparative chromatography was concentrated in situ on the head of an HPLC column, then back flushed into the NMR flow probe using D₂O and deuterated acetonitrile solvent. The ¹H NMR spectra for M2 and M7 clearly show the chemical shifts for the glucuronic acid anomeric protons at approximately 5.0 and 5.7 ppm, indicative of ether- and ester-linked glucuronides (Sidelmann et al., 1997). Likewise, the ¹H NMR spectrum for M9, the bis-glucuronide, shows the chemical shifts for the two anomeric protons at approximately 5.0 and 5.7 ppm. Therefore, LC/NMR methodology utilizing this unique column trapping technique proved useful for definitive structural determination of the acyl and ether glucuronide metabolites.

Although glucuronidation is typically a mechanism of detoxification and elimination, it is well documented that formation of unstable ß-1-O-acyl glucuronides can lead to covalent modification of protein nucleophiles (Smith et al., 1986; Dickinson and King, 1991; Dahms et al., 1997; Wang and Dickinson, 1998; Akira et al., 2002). The capacity of an acyl glucuronide to bind covalently to proteins in vitro appears to be directly related to its degree of instability (Bolze et al., 2002; Benet et al., 1993). Recent data also implicate reactive acyl-CoA derivatives of fatty acid-like drugs in protein modification through transacylation (Sallustio et al., 2000; Li et al., 2002). The covalent modification of macromolecules via acyl glucuronide intermediates has been linked to idiosyncratic adverse drug reactions and led to the withdrawal of several NSAIDs, including zomepirac, aclofenac, and benoxaprofen, from the marketplace. However, there are many currently marketed, and generally accepted as safe, carboxylate drugs with known acyl glucuronide metabolites (Bailey and Dickinson, 2003). Therefore, it cannot be assumed that evidence of covalent protein binding will lead to detrimental biological consequences.

In the current studies, degradation and rearrangement of the acyl and bis-glucuronides of LY293111 were observed following overnight incubation (37°C) of samples containing metabolites generated from rat liver slices. Although a definitive degradation rate constant was not determined in these studies, 56% of the acyl glucuronide was still present after 16 h at 37°C, suggesting that LY293111 acyl glucuronide is more stable than many published examples of reactive acyl glucuronides (Benet et al., 1993; Bolze et al., 2002). Acyl glucuronides of benzoic acid derivatives, such as furosemide, have been shown to be resistant to degradation, and consequently, these metabolites tend to be less reactive with proteins (Ebner et al., 1999; Bolze et al., 2002). However, there are examples of drugs with relatively stable acyl glucuronides, such as mefenamic acid and etodolac, which cause significant covalent modification of albumin in vitro (Smith et al., 1992; McGurk et al., 1996). Therefore, it is not unexpected that acyl glucuronidation of LY293111 leads to in vivo protein modification, despite its relatively slow rate of degradation. Whereas the ether glucuronide was resistant to ß-glucuronidase, the acyl glucuronide and its rearrangement product (M1) were susceptible to enzymatic hydrolysis. This is noteworthy because although acyl glucuronides are often more susceptible to hydrolysis than phenolic glucuronides, their acyl-migrated isomers are generally resistant to ß-glucuronidase (Bailey and Dickinson, 2003).

An assessment of rat plasma samples confirmed that a fraction of LY293111-derived radioactivity was irreversibly bound to plasma protein and that the degree of binding increased over time. This finding is consistent with the differential in half-lives between LY293111 and total radioactivity in rats, and suggests that the long radioactive t_1/2 is due to covalent modification of plasma proteins. This hypothesis is further supported by the fact that the apparent terminal t_1/2 of non-LY293111 radioactivity (50-68 h) is similar to the reported half-life of rat serum albumin (Jeffay, 1960; Reed et al., 1988). Although the possible role of acyl-CoA thioester intermediates cannot be conclusively ruled out from the current experiments, there has been no evidence to date of acyl-CoA conjugates or related amino acid conjugates of LY293111 in any species examined. The exact nature of the protein-bound radioactivity has not been identified, but the involvement of the acyl glucuronides in such binding is well established.

In conclusion, LY293111 is metabolized almost exclusively by direct glucuronidation, resulting in acyl, ether, and bis-glucuronides. The acyl and bis-glucuronides appear to undergo rearrangement in vitro and may be responsible for covalent binding of plasma proteins in vivo, resulting in a prolonged t_1/2 of circulating radioactivity in rats dosed with [¹⁴C]LY293111. Additional studies are needed to further elucidate this phenomenon and to conclusively identify the specific proteins involved. The profile of in vitro metabolites was qualitatively similar among rats, mice, monkeys, and humans; however, the relative quantities of the glucuronide metabolites differed among species. In addition to the prevalent glucuronides, there were several minor hydroxylated metabolites present in rat liver slice incubations. However, these metabolites were not observed in other species and would not be predicted to be significant in vivo metabolites in these species. These data suggest that covalent modification of proteins via LY293111 acyl glucuronide is possible in humans; however, this metabolite appears to be less reactive than those known to cause adverse drug reactions, such as zomepirac and tolmetin. Therefore, although the details of LY293111 acyl glucuronide-related protein binding warrant further investigation, the current data do not pose significant concerns in an oncology patient population.

	Acknowledgments

 
We acknowledge and thank Dr. William Ehlhardt for his thoughtful comments and valuable review of the manuscript.

	Footnotes

 
¹ Abbreviations used are: LY293111, 2-[3-[3-[(5-ethyl-4'-fluoro-2-hydroxy[1,1'-biphenyl]-4-yl)oxy]propoxy]-2-propylphenoxy]benzoic acid; HPLC, high-performance liquid chromatography; 7-HC, 7-hydroxycoumarin; ESI, electrospray ionization; APCI, atmospheric pressure chemical ionization; LC/NMR, liquid chromatography/nuclear magnetic resonance.

Address correspondence to: Dr. Everett J. Perkins, Eli Lilly and Company, Lilly Corporate Center, Indianapolis, IN 46285. E-mail: eperkins{at}lilly.com

	References

Top Abstract Materials and Methods Results Discussion References

 


Akira K, Uchijima T, and Hashimoto T (2002) Rapid internal acyl migration and protein binding of synthetic probenecid glucuronides. Chem Res Toxicol 15: 765-772.[CrossRef][Medline]

Bailey MJ and Dickinson RG (2003) Acyl glucuronide reactivity in perspective: biological consequences. Chem Biol Interact 145: 117-137.[CrossRef][Medline]

Barr J, Weir AJ, Brendel K, and Sipes IG (1991) Liver slices in dynamic organ culture. I. An alternative in vitro technique for the study of rat hepatic drug metabolism. Xenobiotica 21: 331-339.[Medline]

Bolze S, Bromet N, Gay-Feutry C, Massiere F, Boulieu R, and Hulot T (2002) Development of an in vitro screening model for the biosynthesis of acyl glucuronide metabolites and the assessment of their reactivity toward human serum albumin. Drug Metab Dispos 30: 404-413.[Abstract/Free Full Text]

Cramer JW, Farid NA, Richardson LA, and Rinkema LE (1995) Pharmacokinetics and excretion of [¹⁴C]LY293111, a new LTB₄ receptor antagonist, in rat and monkey. Pharm Res (NY) 12 (9 Suppl): S-343/Abstr PPDM 8066.

Dahms M, Lotz R, Lang W, Renner U, Bayer E, and Spahn-Langguth H (1997) Elucidation of phase I and phase II metabolic pathways of Rhein: species differences and their potential relevance. Drug Metab Dispos 25: 442-252.[Abstract/Free Full Text]

Dickinson RG and King AR (1991) Studies on the reactivity of acyl glucuronides—II. Interaction of diflunisal acyl glucuronide and its isomers with human serum albumin in vitro. Biochem Pharmacol 42: 2301-2306.[CrossRef][Medline]

Ebner T, Heinzel G, Prox A, Beschke K, and Wachsmuth H (1999) Disposition and chemical stability of telmisartan 1-O-aclyglucuronide. Drug Metab Dispos 27: 1143-1149.[Abstract/Free Full Text]

Ehlhardt WJ, Woodland JM, Baughman TM, Vandenbranden M, Wrighton SA, Kroin JS, Norman BH, and Maple SR (1998) Liquid chromatography/nuclear magnetic resonance spectroscopy and liquid chromatography/mass spectrometry identification of novel metabolites of the multidrug resistance modulator LY335979 in rat bile and human liver microsomal incubations. Drug Metab Dispos 26: 42-51.[Abstract/Free Full Text]

Hyneck ML, Smith PC, Munafo A, McDonagh F, and Benet LZ (1988) Disposition and irreversible protein binding of tolmetin in humans. Clin Pharmacol Ther 44: 107-114.[Medline]

Jaggi R, Addison RS, King AR, Suthers BD, and Dickinson RG (2002) Conjugation of desmethyl napoxen in the rat—a novel acyl glucuronide-sulfate diconjugate as a major biliary metabolite. Drug Metab Dispos 30: 161-166.[Abstract/Free Full Text]

Jeffay H (1960) The metabolism of serum proteins. III. Kinetics of serum protein metabolism during growth. J Biol Chem 235: 2352-2356.[Free Full Text]

Li C, Benet LZ, and Grillo MP (2002) Studies on the chemical reactivity of 2-phenylpropionic acid 1-O-acyl glucuronide and S-acyl-CoA thioester metabolites. Chem Res Toxicol 15: 1309-1317.[CrossRef][Medline]

Marder P, Spaethe SM, Froelich LL, Cerimele BJ, Petersen BH, Tanner T, and Lucas RA (1996) Inhibition of ex vivo neutrophil activation by oral LY293111, a novel leukotriene B4 receptor antagonist. Br J Clin Pharmacol 42: 457-464.[CrossRef][Medline]

Marshall M, Diaz B, and Brozinick JL (2002a) LY293111 inhibits tumor cell growth in vitro through an apparent PPAR gamma agonist activity. American Association for Cancer Research 93rd Annual Meeting, San Francisco, CA, April 6-10, 2002. 43: 963.

Marshall M, Diaz B, Froelich L, Barnard D, Hocket R, Weitzman A, Sawyer S, Law K, Starling J, Iverson P, and Fleisch JH (2002b) The LTB4 receptor antagonist, LY293111, inhibits tumor cell growth in vitro independently of the LTB4 receptor. American Association for Cancer Research 93rd Annual Meeting, San Francisco, CA, April 6-10, 2002. 43: 957.

McGurk KA, Remmel RP, Hosagrahara VP, Tosh D, and Burchell B (1996) Reactivity of mefenamic acid 1-O-acyl glucuronide with proteins in vitro and ex vivo. Drug Metab Dispos 24: 824-849.

Monden M and Fortner JG (1982) Twenty-four and 48 h canine liver preservation by simple hypothermia with prostacyclin. Ann Surg 196: 38-42.[Medline]

Mutlib AE, Shockcor J, Chen S, Espina R, Pinto DJ, Orwat MJ, Prakash SR, and Gan L (2002) Disposition of 1-[3-(aminomethyl)phenyl]1H-pyrazole-5-5-carboxamide (DPC 423) by novel metabolic pathways. Characterization of unusual metabolites by liquid chromatography/mass spectrometry and NMR. Chem Res Toxicol 15: 48-62.[CrossRef][Medline]

Reed RG, Davidson LK, Burrington CM, and Peters T (1988) Non-resolving jaundice: bilirubin covalently attached to serum albumin circulates with the same metabolic half-life as albumin. Clin Chem 34: 1992-1994.[Abstract/Free Full Text]

Sallustio BC, Nunthasomboon S, Drogemuller CJ, and Knights KM (2002) In vitro covalent binding of nafenopin-CoA to human liver proteins. Toxicol Appl Pharmocol 163: 176-182.

Sawyer JS (1996) LY293111 sodium. Drugs Future 21: 610-614.

Sawyer JS, Bach NJ, Baker SR, Baldwin RF, Borromeo PS, Cockerham SL, Fleisch JH, Floreancig P, Froelich LL, Jackson WT, et al. (1995) Synthetic and structure/activity studies on acid-substituted 2-arylphenols: discovery of 2-[2-propyl-3-[3-[2-ethyl-4-(4-fluorophenyl)-5-hydroxyphenoxy]-propoxy]phenoxy]benzoic acid, a high-affinity leukotriene B4 receptor antagonist. J Med Chem 38: 4411-4432.[CrossRef][Medline]

Sidelmann U, Christiansen E, Krogh L, Cornett C, Tjornelund J, and Hansen S (1997) Purification and 1H NMR spectroscopic characterization of phase II metabolites of tolfenamic acid. Drug Metab Dispos 25: 725-731.[Abstract/Free Full Text]

Smith PC, McDonagh AF, and Benet LZ (1986) Irreversible binding of zomepirac to plasma protein in vitro and in vivo. J Clin Investig 77: 934-939.

Smith PC, Song WQ, and Rodriguez RJ (1992) Covalent binding of etodolac acyl glucuronide to albumin in vitro. Drug Metab Dispos 20: 962-965.[Medline]

Steemsa A, Beamand JA, Walter DG, Price RJ, and Lake BG (1994) Metabolism of coumarin and 7-ethoxycoumarin by rat, mouse, guinea pig, cynomolgus monkey and human precision-cut liver slices. Xenobiotica 24: 893-907.[Medline]

Tong W, Ding X, Henning R, Witt RC, Standop J, Pour PM, and Adrian TE (2002) Leukotriene B4 receptor antagonist LY293111 inhibits proliferation and induces apoptosis in human pancreatic cancer cells. Clin Cancer Res 8: 3232-3242.[Abstract/Free Full Text]

Wang M and Dickinson RG (1998) Disposition and covalent binding of diflunisal and diflunisal acyl glucuronide in the isolated perfused rat liver. Drug Metab Dispos 26: 98-104.[Abstract/Free Full Text]

This Article Abstract Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Perkins, E. J. Articles by Cassidy, K. C. Search for Related Content PubMed PubMed Citation Articles by Perkins, E. J. Articles by Cassidy, K. C.