Article pubs.acs.org/molecularpharmaceutics
Combined in Vitro−in Vivo Approach To Assess the Hepatobiliary Disposition of a Novel Oral Thrombin Inhibitor Elin M. Matsson,† Ulf G. Eriksson,§ Johan E. Palm,§ Per Artursson,†,‡ Maria Karlgren,†,‡ Lucia Lazorova,†,‡ Marie Bran̈ nström,§ Anja Ekdahl,§ Kristina Dunér,§ Lars Knutson,∥ Susanne Johansson,§ Kajs-Marie Schützer,§ and Hans Lennernas̈ *,† †
Department of Pharmacy, Uppsala University, Box 580, SE-751 23 Uppsala, Sweden Uppsala University Drug Optimization and Pharmaceutical Profiling Platform, Department of Pharmacy, Uppsala University, Box 580, SE-751 23 Uppsala, Sweden § AstraZeneca R&D Mölndal, SE-431 83 Mölndal, Sweden ∥ Department of Surgical Sciences, Uppsala University Hospital, SE-751 85 Uppsala, Sweden ‡
S Supporting Information *
ABSTRACT: Two clinical trials and a large set of in vitro transporter experiments were performed to investigate if the hepatobiliary disposition of the direct thrombin inhibitor prodrug AZD0837 is the mechanism for the drug−drug interaction with ketoconazole observed in a previous clinical study. In Study 1, [3H]AZD0837 was administered to healthy male volunteers (n = 8) to quantify and identify the metabolites excreted in bile. Bile was sampled directly from the jejunum by duodenal aspiration via an oro-enteric tube. In Study 2, the effect of ketoconazole on the plasma and bile pharmacokinetics of AZD0837, the intermediate metabolite (ARH069927), and the active form (AR-H067637) was investigated (n = 17). Co-administration with ketoconazole elevated the plasma exposure to AZD0837 and the active form approximately 2-fold compared to placebo, which may be explained by inhibited CYP3A4 metabolism and reduced biliary clearance, respectively. High concentrations of the active form was measured in bile with a bile-to-plasma AUC ratio of approximately 75, indicating involvement of transporter-mediated excretion of the compound. AZD0837 and its metabolites were further investigated as substrates of hepatic uptake and efflux transporters in vitro. Studies in MDCK-MDR1 cell monolayers and P-glycoprotein (P-gp) expressing membrane vesicles identified AZD0837, the intermediate, and the active form as substrates of P-gp. The active form was also identified as a substrate of the multidrug and toxin extrusion 1 (MATE1) transporter and the organic cation transporter 1 (OCT1), in HEK cells transfected with the respective transporter. Ketoconazole was shown to inhibit all of these three transporters; in particular, inhibition of P-gp and MATE1 occurred in a clinically relevant concentration range. In conclusion, the hepatobiliary transport pathways of AZD0837 and its metabolites were identified in vitro and in vivo. Inhibition of the canalicular transporters P-gp and MATE1 may lead to enhanced plasma exposure to the active form, which could, at least in part, explain the clinical interaction with ketoconazole. KEYWORDS: direct thrombin inhibitor, biliary clearance, transport proteins, liver transport, drug−drug interactions, ketoconazole, AZD0837
■
INTRODUCTION AZD0837 is an oral anticoagulant prodrug intended for the prevention and treatment of thromboembolic diseases, developed by AstraZeneca.1−3 The compound has been studied in four phase II clinical trials in patients with atrial fibrillation.4 Following administration, AZD0837 is bioconverted to the active form, AR-H067637, a selective and reversible direct thrombin inhibitor.5 The biotransformation of AZD0837 to the active form occurs via an intermediate (AR-H069927) and includes two metabolic reactions (Figure 1). In vitro results have shown that CYP3A4, 2C9, 2C19, and 2J2 may catalyze the demethylation, and that the subsequent reduction is mediated by the N-hydroxylamine reductase enzyme system (AstraZeneca data on file).6 The influence of CYP3A4 inhibition on the pharmacokinetics of AZD0837 and its two metabolites in vivo, © XXXX American Chemical Society
has been investigated in interaction studies performed with grapefruit juice and ketoconazole.7 No effects were observed from the concomitant administration of grapefruit juice (200 mL), an inhibitor that is thought to reduce CYP3A4 activity primarily in the intestine.8,9 In contrast, the strong CYP3A4 inhibitor ketoconazole (400 mg, single dose) increased the area under the plasma concentration−time curve (AUC) and the maximum concentration (Cmax) of both AZD0837 and the active form: 2-fold and 1.5-fold for the prodrug, respectively, and 2.3-fold and 1.9-fold for the active form.7 Assuming that Received: June 11, 2013 Revised: September 4, 2013 Accepted: September 30, 2013
A
dx.doi.org/10.1021/mp400341t | Mol. Pharmaceutics XXXX, XXX, XXX−XXX
Molecular Pharmaceutics
Article
Figure 1. Direct thrombin inhibitor prodrug AZD0837 metabolic conversion into the pharmacologically active compound, AR-H067637, via an intermediate, AR-H069927. The biotransformation involves two metabolic reactions: (1) oxidation mediated by CYP3A4, CYP2C9, CYP2C19, and CYP2J2 and (2) reduction catalyzed by the N-hydroxylamine reductase enzyme system.
CYP3A4 is the major enzyme contributing to the elimination of AZD0837 in vivo, the effect of ketoconazole on the prodrug plasma exposure was expected. The mechanism explaining the changes in the pharmacokinetics of the active form was, however, unclear and resulted in further evaluation as presented here. About 15% of an intravenous dose of the active form has been shown to be renally excreted, in healthy volunteers.10 This suggests that a major fraction of the administered dose is cleared by metabolism and/or excretion into bile. Contribution of biliary excretion to the elimination of the active form has also been indicated by nonclinical data: in pigs, approximately 60% of an intravenous bolus dose of the active form was excreted into bile.11 From these findings it was hypothesized that the interaction between ketoconazole and the active form could be caused by reduced biliary excretion or inhibited metabolism of the thrombin inhibitor. The main objectives of the present in vivo and in vitro studies were to investigate the biliary excretion of AZD0837 and its metabolites in humans and to identify possible transporters involved in the hepatobiliary disposition of these compounds. In two clinical trials, bile was aspirated via an oroenteric tube positioned in the duodenum of healthy volunteers following dosing with AZD0837. The same multichannel tube was used for administration of the study drugs, in an intestinal segment separated from the bile sampling site (Figure 2). The first clinical trial aimed at identifying all metabolites of AZD0837 excreted in bile, including the intermediate metabolite and the active form of AZD0837 that had been detected in plasma, urine, or feces in earlier investigations.10 The second clinical trial aimed at investigating the effect of ketoconazole on the biliary clearance of AZD0837, the intermediate and active form. As high biliary concentrations of the compounds were demonstrated, in vitro assays were performed to clarify the biliary transport pathways for the compounds and provide a mechanistic explanation for the interaction with ketoconazole.
Figure 2. Intestinal path of study drugs. The oro-enteric tube was guided by a physician to the upper part of the small intestine in the healthy volunteers and, once its position was confirmed with fluoroscopy, a balloon attached to the tube was inflated, which separated the site for drug administration and bile sampling. The inflated balloon was placed in close proximity to the papilla of Vater, i.e., where bile enters the lower parts of the duodenum.
from the evening before the day of the study and were on the study day given standardized meals following the removal of the tube. The subjects were not allowed to leave the study site until the coagulation parameter APTT was below predose value +10 s. Study Design. The investigations were conducted at Uppsala University Hospital, Sweden. The protocols were approved by the Swedish Medical Products Agency and the local independent research ethics committee in Uppsala, and the studies were performed in accordance with good clinical practice guidelines and the Declaration of Helsinki. In Study 1 (radioactive dose), eight subjects were planned to be included, to allow for completion of the study with five subjects. The trial included one study day, when the volunteers received 350 mg [3H]AZD0837 (7.4 MBq) via an oro-enteric tube into the proximal jejunum, at a site separated from the site of bile sampling (Figure 2). The dose was followed by 50 mL of water to rinse the syringe and catheter used for administration. The investigational product, 4 mg/mL [3H]AZD0837 solution, was supplied by AstraZeneca R&D Mölndal to the clinical site,
■
EXPERIMENTAL SECTION Clinical Studies. Subjects. The subjects provided their informed consent to participate. Young, male volunteers ascertained to be healthy from their medical history, a physical examination, and routine laboratory tests were included in studies 1 (D1250C00032) and 2 (D1250C00029). The exclusion criteria included intake of any drug (with the exception of occasional use of acetaminophen), St. John’s wort, and grapefruit juice. The individuals had to be fasting B
dx.doi.org/10.1021/mp400341t | Mol. Pharmaceutics XXXX, XXX, XXX−XXX
Molecular Pharmaceutics
Article
where the dose was dispensed. The [3H]-labeling was in a metabolically stable position on AZD0837. In Study 2 (drug−drug interaction), approximately 20 subjects were planned to be included to have at least 10 subjects completing the study sessions. This was an openlabeled, randomized, crossover trial, with a wash-out period of 7−21 days. The study consisted of two study periods when AZD0837 was given either alone (control phase) or together with repeated dosing of ketoconazole (ketoconazole phase). In the control phase, a single dose of AZD0837 (175 mg) was administered in the proximal jejunum via an oro-enteric tube in the morning of the study day. The ketoconazole phase was a four day treatment with ketoconazole (400 mg) (2 × 200 mg tablets of Fungoral, Janssen-Cilag, Sweden) being given orally with breakfast at the clinical site on three successive days. On day 4, ketoconazole (400 mg) was coadministered with AZD0837 (175 mg), and both compounds were given via the oro-enteric tube in the proximal jejunum. Ketoconazole tablets were dispersed in water (400 mg, 8 mg/mL) and administered 30 min prior to the AZD0837 dose. Following each enteral dose of ketoconazole and AZD0837, the syringe and catheter used for administration were rinsed with 50 mL of water. To make the two treatments as similar as possible, a sham volume of 100 mL of water was given 30 min prior to the dosing of AZD0837, in the control phase. The AZD0837 solution (4 mg/ mL) supplied by AstraZeneca R&D Mölndal was dispensed at the clinical site, where 18.5 kBq of [14C]-labeled polyethylene glycol 4000 (PEG4000) was added to the solution. This nonabsorbable marker was used to control for possible leakage from the administration site to the bile sampling site (Figure 2). However, in the analysis of the bile data, the very low concentration of the nonabsorbable marker [14C]PEG4000 prohibited consistent quench correction and could not be used to conclusively rule out leakage from the site of administration to the site of bile aspiration. In both studies, blood samples were taken just before administration of AZD0837 and 0.25, 0.5, 1, 1.5, 2, 2.5, 3, 4, 6, 8, 10, and 24 h thereafter. Bile was continuously sampled in 20 min fractions from the duodenum during three hours after the dosing of AZD0837. It should be noted that bile collected from the duodenum to some extent is diluted with intestinal fluid and pancreatic secretion. To aspirate bile from the intestine, the Loc-I-Gut perfusion tube (Synectics Medical, Stockholm, Sweden) was utilized, and this method has previously been described in detail.12,13 The multichannel tube was positioned in the proximal jejunum of the healthy volunteers, and once its location was confirmed to be correct by fluoroscopy, a balloon attached to the catheter was inflated (Figure 2). The balloon had a dual function: it isolated the drug administration site from the site of the bile sampling and kept the catheter from moving distally along with the peristaltic contractions in the intestine during the bile sampling time. To ensure efficient aspiration of bile from the intestine, the oro-enteric catheter was connected to a vacuum pump (Ameda suction pump type 23, Ameda AG, Zug, Switzerland). Analysis of AZD0837, the Intermediate and the Active Form in Bile and Plasma. Plasma. The analytes were isolated by solid phase extraction by loading the plasma samples (100 μL) on columns (Isolute C6, 25 mg) together with water (100 μL), isotope-labeled internal standards (D6) for all three analytes (50 μL) and a buffer (200 μL, 0.6 M malonic acid, pH 3). The columns were washed with liquid (1 mL) consisting of 15% methanol and 85% buffer (0.1 M malonic acid, pH 3), and
the analytes were eluted with 80% acetonitrile in 0.0076% formic acid (400 μL). The eluate was partially evaporated and dissolved in 0.038% formic acid (400 μL) before liquid chromatography/mass spectrometry (LC-MS/MS) analysis of 10 μL. The three analytes were separated on a reversed phase C18 column (Hypersil HyPURITY, 50 × 2.1 mm, 5 μm) under isocratic conditions at 40 °C, using a mobile phase consisting of 28% acetonitrile in 0.027% formic acid and flow rate 0.3 mL/ min. The mass spectrometer, an API 3000 triple quadrupole with electrospray interface, was used in selective reaction monitoring mode. The linear concentration range for the three analytes was 10−10 000 nM. The average accuracy of the plasma quality control samples, run with each analytical batch in the present studies, were at the concentrations 3 times the lower limit of quantification (LLOQ) and 0.8 times the upper limit of quantification (ULOQ) within the interval 96.9− 103.8%, and the coefficient of variation (CV) was less than 5% for all three analytes at both concentration levels. No interfering peaks occurred in the areas of interest. Bile. The concentration of AZD0837 and its metabolites, in the bile samples, were determined using mixed mode solid phase extraction and LC-MS/MS as has been described elsewhere.14 The linear concentration range was 1.00−1000 μM for the active form and 0.0200−20.0 μM for the intermediate and AZD0837. The average accuracy of the bile quality control samples, run with each analytical batch in the present studies, were at the concentrations 3 times LLOQ and 0.8 times ULOQ within the interval 97.3−102.7%, and the CV was less than 10% for all three analytes at both concentration levels. No interfering peaks occurred in the areas of interest. Determination of [3H]AZD0837 and Metabolic Profile in Bile. The radioactivity in bile was determined in a liquid scintillation counter (Wallac 1414 WinSpectral v1.40) with automatic quench correction. Prior to analysis the samples were mixed with scintillation liquid (Optiphase Hisafe), shaken, and allowed to heat and light stabilize. The bile fractions containing >1% of the dose were pooled together and analyzed using LC combined with radioactivity monitoring and MS. Radioactivity monitoring was used for quantification and recording of the metabolite profile and MS for structural elucidation of metabolites. The LC system consisted of a CTC HTS PAL autosampler (CTC Analytics AG, Zwingen, Switzerland) and an HPLC pump (Hewlett-Packard 1100, Palo Alto, CA) operating at a flow rate of 1 mL/min. Chromatographic separations were performed on an ACE 5 C18 column (150 × 4.6 mm i.d., 5 μm) protected by a precolumn ACE 5 C18 (10 × 4.6 mm). The mobile phases used were 4 mM ammonium acetate dissolved in water (A) and acetonitrile (B): 20−55% B 0−25 min; 55% B 25−30 min; wash with 80% B for 5 min; equilibration for 1 min. One injection (volume 80 μL) of the pooled bile sample was analyzed. The LC eluent was split and transferred for fraction collection into deep-well microplates (Packard Instrument Co., Meriden, CT) using a fraction collector (FC204, Gilson Inc., Middleton, WI) with a collection time of 0.13 min. The microplates were allowed to dry and placed in a scintillation counter (TopCount, Packard Instrument Co.) with 12 detectors, and each well was counted for 60 min. The remainder of the LC eluent was diverted into a hybrid quadrupole time-of-flight mass spectrometer (Micromass QTof 2) with a LockSpray electrospray interface. Specific mass spectrometric source conditions were: capillary voltage 3.20 kV, C
dx.doi.org/10.1021/mp400341t | Mol. Pharmaceutics XXXX, XXX, XXX−XXX
Molecular Pharmaceutics
Article
reagents were obtained from Invitrogen (Carlsbad, CA), SigmaAldrich (St. Louis, MO), or BD Biosciences (Bedford, MA). Cell Culture Procedures. Human embryonic kidney (HEK293) cells transfected with empty vector (HEK-pTREX) and HEK293 cells expressing human OCT1/SLC22A1 (HEK-OCT1), obtained from AstraZeneca R&D Mölndal,15 were seeded on 24-well poly-D-lysine coated plates at a density of 500 000 cells/well three days prior to the transport experiment. The cells were grown in medium consisting of Dulbecco’s modified Eagle’s medium (DMEM)/F-12 1:1 with glutamax-I supplemented with fetal bovine serum (FBS) (10% v/v), Penicillin-Streptomycin (PEST) (0.5% v/v; 10 000 U/mL penicillin and 10 000 μg/mL streptomycin), and 50 mg/mL Geneticin (1% v/v). HEK293-Flp-In cells (Invitrogen) stably transfected with MATE1 (kindly provided by K. M. Giacomini16), OATP1B1*1a, OATP1B3*2, OATP2B1*1, or with empty pcDNA5/ FTR vector,17 were cultured in DMEM supplemented with 10% FBS, 2 mM L-glutamine, and 60 μg/mL (MATE1) or 75 μg/mL (OATP1B1, OATP1B3, and OATP2B1) hygromycin B. Three days prior to the transport experiment, cells were seeded on CellBind 96-well plates at 100 000 cells/well (MATE1, OATP1B3) or 200 000 cells/well (OATP1B1, OATP2B1) in high glucose DMEM without phenol red, supplemented with 10% FBS and 2 mM L-glutamine. Wild type Madin-Darby Canine Kidney cells (MDCK-WT) and MDCK cells transfected with the human ABCB1/MDR1 gene encoding for P-glycoprotein (MDCK-MDR1) were obtained from The Netherlands Cancer Institute, Amsterdam, The Netherlands. The cells were seeded onto 1.13 cm2 Transwell polycarbonate cell culture inserts at a density of 150 000 cells/insert and cultured in DMEM supplemented with 10% FBS, 3 mM L-glutamine, and 1% nonessential amino acids, three days prior to the transport experiment. HEK293-OCT1, -MATE1, -OATP1B1, -OATP1B3, and -OATP2B1 Cell Assays. Functional activity of the respective transporter was verified using the probe substrates [14C]metformin and [ 14 C]TEA (OCT1), [ 14 C]-metformin (MATE1), [3H]E217βG (OATP1B1 and OATP1B3), and [3H]E3S (OATP2B1). In addition, the apparent IC50 value for inhibition of the transporter-mediated uptake of probe substrates by ketoconazole was determined (with the exception of OCT1 for which cimetidine was used as inhibitor). The transfected HEK293 cells were washed and preincubated with transport medium (HBSS with 25 mM HEPES, pH 7.4 (OCT1); HBSS, pH 7.4 (OATPs); HBSS, pH 8.5 (MATE1)). Radiolabeled test compounds were mixed with the corresponding cold compound to obtain the desired concentrations of AZD0837 (1.7 μM and 100 μM for OATPs/MATE1 assay; 0.8 μM for OCT1 assay), the intermediate (17 μM and 100 μM for OATPs/MATE1 assay; 20 μM for OCT1 assay), and the active form (1 μM and 100 μM for OATPs/MATE1 assay; 0.9 μM for OCT1 assay). The intermediate was tested at higher concentrations due to >100-fold lower specific activity of the [14C]-labeled compound (see Materials). Uptake was initiated by adding prewarmed transport medium containing the test compound and was terminated at designated time points by removing the radioactive solutions and adding ice cold HBSS or PBS buffer. Prior to analysis the cells were lysed using NaOH and subsequently neutralized with HCl. Aliquots were transferred to scintillation vials and intracellular accumulation of radioactivity was measured using TopCount or liquid scintillation counters. Samples were removed from the
cone voltage 30 V, MCP 2000 V, TOF 9.10 kV, source temperature 120 °C, and desolvation temperature 300 °C. The instrument was calibrated with sodium formate (mass 80−1000 Da). Pharmacokinetic Analysis. The pharmacokinetic parameters were determined from the plasma and bile concentration− time profiles by noncompartmental analysis using WinNonlin 5.2 (Pharsight Corp., Mountain View, CA). The AUC0−3h and AUC0−24h were derived by means of the linear/logarithmic trapezoidal rule (for the ascending and descending parts of the curve, respectively) to the last quantifiable concentration. Regression analysis of the last 3−5 data points in the terminal part of the log concentration−time curve yielded the terminal rate constant, ke, which was used to calculate the t1/2. The AUC was extrapolated to infinity (AUC0−∞) by dividing the last concentration by k e . The amount of AZD0837, the intermediate, and the active form recovered in bile during the three hours following drug administration, Aebile, was calculated for each compound and summed to obtain the total. The AZD0837 levels in the first bile fraction following drug administration were set to zero to minimize the risk of including possibly contaminated samples in the evaluation of data. The apparent CLbile/FH was estimated for each compound: CL bile Ae bile = FH AUC0−3h,plasma
(1)
where CLbile is the biliary clearance and FH is the fraction of the compound/metabolite that escapes hepatic elimination during the first-pass extraction. The estimated CLbile/FH-values include sinusoidal uptake and canalicular efflux and, for the intermediate and the active form, intracellular formation within the hepatocyte. These values should not be interpreted as true biliary elimination parameters but are used for comparison purposes in the present studies. The flow of bile expelled into the duodenum was determined for each 20 min sampling fraction of bile collected. Statistical Analysis. The data are expressed as means (±SD) in the text and tables and as mean (±SEM) in the figures. Statistical comparisons of the pharmacokinetic parameters (other than tmax) between the two treatment phases in Study 2 were carried out with the paired t test. The Wilcoxon signed rank test was performed (Minitab release 15, Minitab Inc., PA) for the parameter tmax. AUC, Cmax, t1/2, and CLbile data were logarithmically transformed before the statistical analysis. The level of statistical significance was set at p < 0.05. In Vitro Transporter Studies. Materials. Unlabeled metformin, estrone 3-sulfate, estradiol-17β-D-glucuronide, and ketoconazole were purchased from Sigma-Aldrich (St. Louis, MO). [14C]-metformin (3.52 and 4.14 kBq/nmol) was purchased from Moravek Biochemicals (Brea, CA). [14C]TEA (2.04 kBq/nmol) was purchased from American Radiolabeled Chemicals (Brea, CA). [3H]-Estrone 3-sulfate (E3S) (2120 kBq/nmol) and [3H]-estradiol-17β-D -glucuronide (E217βG) (1735 kBq/nmol) were purchased from PerkinElmer Life and Analytical Sciences (Boston, MA). N-Methylquinidine, AZD0837, AR-H069927 (the intermediate), ARH067637 (the active form), and radiolabeled [3H]-N-methylquinidine (2800 kBq/nmol), [3H]-AZD0837 (227 kBq/nmol), [14C]-AR-H069927 (2.1 kBq/nmol), and [3H]-AR-H067637 (609 kBq/nmol) were obtained from AstraZeneca R&D Mölndal (Sweden). All cell plastic and cell culture media and D
dx.doi.org/10.1021/mp400341t | Mol. Pharmaceutics XXXX, XXX, XXX−XXX
Molecular Pharmaceutics
Article
effect of ketoconazole on the transporter-mediated uptake of each probe substrate was determined, in the respective vesicles assay. The vesicular transport studies were performed using a rapid filtration technique on 96-well filter plates according to previously reported protocols.18 Transport was measured both in the transporter containing and control (mock transfected) vesicles, and the ATP-dependent transport (pmol/min/mg protein) was calculated by subtracting the values obtained in the absence of ATP from those in the presence of ATP. A compound was classified as a substrate of the transporter if it showed a statistically significant ATP-dependent transport in the transfected membrane vesicles that was different from that in the mock vesicles. In the BCRP, MRP2, and MRP3 vesicle assays, the transport was examined at one test compound concentration (0.9 μM [3H]AZD0837, 47 μM [14C]AR-H069927 (intermediate), and 0.9 μM [3H]AR-H067637 (active form)) over a range of incubation times (1, 2, 5, 10, and 20 min). In the P-gp vesicle assay, the ATP-dependent transport of test compounds was measured using a 2 min incubation time and a range of eight test concentrations (0.5−500 μM AZD0837, 46−1000 μM intermediate, 0.2−500 μM active form). Km and Vmax values from the P-gp vesicle transport experiments were calculated using the Michaelis−Menten equation. In addition, the IC50 for the ketoconazole inhibition of P-gp mediated transport of [3H]AR-H067637 (active form; 1 μM) was determined.
remaining volume of cell lysate to determine protein concentration using the BCA protein determination kit (Thermo Scientific, Rockford, IL). The results (dpm or cpm) were normalized for total concentration of test compound and for the measured protein concentration of each the wells in the transporter assay plates. Transporter-mediated uptake rate, expressed as pmol/min·mg, was calculated by subtracting the uptake rate obtained in mock cells from that in HEK293transporter cells. Compounds were classified as transporter substrates if the uptake ratio between transporter transfected cells and mock cells was ≥2 and statistically significant (p ≤ 0.05). If a compound was defined as a substrate of a transporter, the concentration dependency of the transport was measured to determine the kinetic parameters, maximum transport rate (Vmax, pmol/min·mg protein) and substrate concentration at half-maximum transport rate (Km, μM), through nonlinear fitting of the data to the standard Michaelis−Menten equation (GraphPad Prism v. 4.02, San Diego, CA). MDCK-MDR1 Cell Monolayer Assay. The integrity of the cell monolayers was checked by measuring the transepithelial electrical resistance before and after each transport experiment. The probe substrate [3H]digoxin demonstrated adequate Pglycoprotein (P-gp) activity in the MDCK-MDR1 cells. Transport of AZD0837, the intermediate and the active form were assessed at 10 μM in both the apical-to-basolateral (A−B) and the basolateral-to-apical (B−A) directions. Immediately after the start of the experiment a sample was removed from the donor compartment, and subsequent samples were taken from the donor and receiver compartment at 30 and 90 min. Samples containing [ 3 H]-digoxin were analyzed using a liquid scintillation counter (Wallac, Turku, Finland). The concentrations of AZD0837 and its metabolites were determined by LC-MS (LLOQ 0.010 μM for all three analytes). The apparent permeability coefficient (Papp) of the test compound was calculated as: Papp = (dQ /dt ) × [1/(A × Cd)]
■
RESULTS Clinical Studies. Subjects. Eight healthy male volunteers (age 29 ± 4 years, body mass index 24 ± 2 kg/m2) were enrolled in Study 1. One of the subjects was excluded because the administration of the study drug deviated from the protocol. Another individual was excluded because the volume of bile obtained during the experiment was not sufficient to allow proper evaluation (50% of the dose) consisted almost exclusively of the active form, which had a bile-to-plasma AUC ratio of approximately 1500.11 However, neither the biliary excretion nor the biliary clearance of the compound were altered after coadministration with ketoconazole. It was therefore suggested that ketoconazole inhibited possible further metabolism of the active form in pigs, which was supported by ketoconazole inhibiting the depletion of the compound in pig liver microsomes.11 In the present clinical studies, the active form was efficiently transported into bile as indicated by the high bile-to-plasma AUC ratio, and no other metabolites were found in bile. Ketoconazole did not affect the amount of the active form excreted into bile during the 3 h collection time. However, the apparent biliary clearance of the compound was reduced by approximately 50%. Given the large intraindividual variability in the biliary clearance of the active form, it was not possible to prove any statistically significant effect of ketoconazole on this parameter. The variability in the bile data is likely caused by the limited sampling time allowed and variability in all bladder contraction (as discussed previously). The apparent effect of ketoconazole on the biliary clearance could be explained by inhibition of the transport proteins involved in the hepatobiliary disposition of the active form, which may lead to changes in the plasma as well as intrahepatic exposure to the compound. It is possible that the lack of effect on terminal t1/2 of the active form can be explained by that inhibitory concentrations of ketoconazole did not cover the entire elimination phase of the active form (ketoconazole t1/2 ∼ 3− 5 h; active form t1/2 ∼ 7 h). The effect of ketoconazole on the active form rather seemed to occur during absorption and the hepatic first pass extraction. Inhibition of P-gp is currently the most well-characterized interaction between ketoconazole and transport proteins,28,29 but the present study shows that ketoconazole also inhibits several other transporters in vitro. In fact, ketoconazole displayed inhibitory potential against all of the transporters that the active form was found to be a substrate of, that is, P-gp, MATE1, and OCT1. However, based on a comparison between I
dx.doi.org/10.1021/mp400341t | Mol. Pharmaceutics XXXX, XXX, XXX−XXX
Molecular Pharmaceutics
Article
(3) Johansson, S.; Cullberg, M.; Eriksson, U. G.; Elg, M.; Dunér, K.; Jensen, E.; Wollbratt, M.; Wåhlander, K. Single-dose pharmacokinetics, pharmacodynamics and safety of AZD0837, a novel oral direct thrombin inhibitor, in young healthy male subjects. Int. J. Clin. Pharmacol. Ther. 2011, 49 (4), 258−267. (4) Ahrens, I.; Peter, K.; Lip, G. Y.; Bode, C. Development and clinical applications of novel oral anticoagulants. Part II. Drugs under clinical investigation. Discovery Med. 2012, 13 (73), 445−450. (5) Deinum, J.; Mattsson, C.; Inghardt, T.; Elg, M. Biochemical and pharmacological effects of the direct thrombin inhibitor AR-H067637. Thromb. Haemost. 2009, 101 (6), 1051−1059. (6) Neve, E. P.; Nordling, A.; Andersson, T. B.; Hellman, U.; Diczfalusy, U.; Johansson, I.; Ingelman-Sundberg, M. Amidoxime reductase system containing cytochrome b5 type B (CYB5B) and MOSC2 is of importance for lipid synthesis in adipocyte mitochondria. J. Biol. Chem. 2012, 287 (9), 6307−6317. (7) Cullberg, M.; Eriksson, U.; Wernevik, L.; Nyström, P.; Svensson, M.; Schützer, K. M.; Nilsson, L. G.; Wåhlander, K. Effect of ketoconazole and grapefruit juice on the pharmacokinetics of the oral direct thrombin inhibitor AZD0837. Basic Clin. Pharmacol. Toxicol. 2007, 101, 130:P205. (8) Lown, K. S.; Bailey, D. G.; Fontana, R. J.; Janardan, S. K.; Adair, C. H.; Fortlage, L. A.; Brown, M. B.; Guo, W.; Watkins, P. B. Grapefruit juice increases felodipine oral availability in humans by decreasing intestinal CYP3A protein expression. J. Clin. Invest. 1997, 99 (10), 2545−2553. (9) Schmiedlin-Ren, P.; Edwards, D. J.; Fitzsimmons, M. E.; He, K.; Lown, K. S.; Woster, P. M.; Rahman, A.; Thummel, K. E.; Fisher, J. M.; Hollenberg, P. F.; Watkins, P. B. Mechanisms of enhanced oral availability of CYP3A4 substrates by grapefruit constituents. Decreased enterocyte CYP3A4 concentration and mechanism-based inactivation by furanocoumarins. Drug Metab. Dispos. 1997, 25 (11), 1228−1233. (10) Johansson, S.; Cullberg, M.; Ekdahl, A.; Wåhlander, K.; Dunér, K.; Eriksson, U. G. Pharmacokinetics of AZD0837, a novel direct thrombin inhibitor, after oral and intravenous administration to healthy male volunteers. Basic Clin. Pharmacol. Toxicol. 2009, 105, 105:TP79. (11) Matsson, E. M.; Palm, J. E.; Eriksson, U. G.; Bottner, P.; Lundahl, A.; Knutson, L.; Lennernäs, H. Effects of ketoconazole on the in vivo biotransformation and hepatobiliary transport of the thrombin inhibitor AZD0837 in pigs. Drug Metab. Dispos. 2011, 39 (2), 239− 246. (12) Bergman, E.; Forsell, P.; Tevell, A.; Persson, E. M.; Hedeland, M.; Bondesson, U.; Knutson, L.; Lennernäs, H. Biliary secretion of rosuvastatin and bile acids in humans during the absorption phase. Eur. J. Pharm. Sci. 2006, 29 (3−4), 205−214. (13) Matsson, E. M.; Eriksson, U. G.; Knutson, L.; Hoffmann, K. J.; Logren, U.; Fridblom, P.; Petri, N.; Lennernäs, H. Biliary excretion of ximelagatran and its metabolites and the influence of erythromycin following intraintestinal administration to healthy volunteers. J. Clin. Pharmacol. 2011, 51 (5), 770−783. (14) Dunér, K.; Hulthe, J.; Tranberg, M. Determination of the thrombin inhibitor AZD0837 and its metabolites in human bile using mixed mode solid phase extraction and LC-MS/MS. J. Chromatogr., B 2012, 904, 29−34. (15) Lohmann, C.; Gelius, B.; Danielsson, J.; Skoging-Nyberg, U.; Hollnack, E.; Dudley, A.; Wahlberg, J.; Hoogstraate, J.; Gustavsson, L. Scintillation proximity assay for measuring uptake by the human drug transporters hOCT1, hOAT3, and hOATP1B1. Anal. Biochem. 2007, 366 (2), 117−125. (16) Chen, Y.; Zhang, S.; Sorani, M.; Giacomini, K. M. Transport of paraquat by human organic cation transporters and multidrug and toxic compound extrusion family. J. Pharmacol. Exp. Ther. 2007, 322 (2), 695−700. (17) Karlgren, M.; Vildhede, A.; Norinder, U.; Wisniewski, J. R.; Kimoto, E.; Lai, Y.; Haglund, U.; Artursson, P. Classification of inhibitors of hepatic organic anion transporting polypeptides (OATPs): influence of protein expression on drug-drug interactions. J. Med. Chem. 2012, 55 (10), 4740−4763.
the reported ketoconazole Cmax,u following dosing with 400 mg to healthy volunteers (approximately 0.15 μM) and the IC50 value of the inhibitor (Table 5),30 inhibition of OCT1 is unlikely to be of any clinical relevance. Thus, in combination with the efficient biliary transport of the active form, our results indicate that inhibition of the canalicular transporters P-gp and/ or MATE1 could, at least in part, explain the interaction between ketoconazole and the active form of AZD0837. In summary, following intraintestinal administration, AZD0837 was excreted into bile predominantly as the active form. Ketoconazole increased the plasma exposure to AZD0837 and its active form, which may, at least in part, be explained by inhibited CYP3A4 metabolism of AZD0837 and reduced biliary clearance of the active form. The active form was shown to be substrate of P-gp, MATE1, and OCT, and ketoconazole was shown to inhibit both P-gp and MATE1 at clinically relevant IC50 values. Hence, the reduced biliary clearance of the active form may be explained by ketoconazole inhibiting the canalicular transporters P-gp and MATE1.
■
ASSOCIATED CONTENT
* Supporting Information S
Concentration dependent transport of identified substrates of hepatic transporters, and concentration dependent inhibition of hepatic transporters by ketoconazole. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*Department of Pharmacy, Uppsala University, Box 580, SE751 23 Uppsala, Sweden. Phone: +46-18-471 4317. Fax: +4618-471 4223. E-mail:
[email protected]. Notes
The authors declare the following competing financial interest(s): The study was sponsored by AstraZeneca R&D Mölndal, Sweden. Ulf Eriksson, Johan Palm, Anja Ekdahl, Kristina Dunér, Susanne Johansson, Marie Brännström, and Kajs-Marie Schützer are employees of AstraZeneca R&D Mölndal.
■
ACKNOWLEDGMENTS The authors are very grateful to Inger Ohlsson, Lisa Wester, and Claes Ericsson for excellent assistance at the clinical site. We would also like to acknowledge Mattias Tranberg, Lars Renberg, Elisabet Berg (AstraZeneca R&D Mölndal, Sweden) and Alex Attema (PRA International, The Netherlands) for the analysis of biological samples; Karima Ben Tabah, Charlotta Vedin Nilsson, and Maria Ulvestad (AstraZeneca R&D Mölndal, Sweden) for excellent assistance with in vitro transporter experiments.
■
REFERENCES
(1) Lip, G. Y.; Rasmussen, L. H.; Olsson, S. B.; Jensen, E. C.; Persson, A. L.; Eriksson, U.; Wåhlander, K. F. Oral direct thrombin inhibitor AZD0837 for the prevention of stroke and systemic embolism in patients with non-valvular atrial fibrillation: a randomized dose-guiding, safety, and tolerability study of four doses of AZD0837 vs. vitamin K antagonists. Eur. Heart J. 2009, 30 (23), 2897−2907. (2) Olsson, S. B.; Rasmussen, L. H.; Tveit, A.; Jensen, E.; Wessman, P.; Panfilov, S.; Wåhlander, K. Safety and tolerability of an immediaterelease formulation of theoral direct thrombin inhibitor AZD0837 in the prevention of stroke and systemic embolism in patients with atrial fibrillation. Thromb. Haemost. 2010, 103 (3), 604−612. J
dx.doi.org/10.1021/mp400341t | Mol. Pharmaceutics XXXX, XXX, XXX−XXX
Molecular Pharmaceutics
Article
(18) Karlsson, J. E.; Heddle, C.; Rozkov, A.; Rotticci-Mulder, J.; Tuvesson, O.; Hilgendorf, C.; Andersson, T. B. High-activity pglycoprotein, multidrug resistance protein 2, and breast cancer resistance protein membrane vesicles prepared from transiently transfected human embryonic kidney 293-epstein-barr virus nuclear antigen cells. Drug Metab. Dispos. 2010, 38 (4), 705−714. (19) Ghibellini, G.; Leslie, E. M.; Brouwer, K. L. Methods to evaluate biliary excretion of drugs in humans: an updated review. Mol. Pharmaceutics 2006, 3 (3), 198−211. (20) Shaffer, E. A. Review article: control of gall-bladder motor function. Aliment. Pharmacol. Ther 2000, 14 (Suppl 2), 2−8. (21) Stolk, M. F.; Van Erpecum, K. J.; Peeters, T. L.; Samsom, M.; Smout, A. J.; Akkermans, L. M.; Vanberge-Henegouwen, G. P. Interdigestive gallbladder emptying, antroduodenal motility, and motilin release patterns are altered in cholesterol gallstone patients. Dig. Dis. Sci. 2001, 46 (6), 1328−1334. (22) Stolk, M. F.; van Erpecum, K. J.; Smout, A. J.; Akkermans, L. M.; Jansen, J. B.; Lamers, C. B.; Peeters, T. L.; van Berge-Henegouwen, G. P. Motor cycles with phase III in antrum are associated with high motilin levels and prolonged gallbladder emptying. Am. J. Physiol. 1993, 264 (4 Pt 1), G596−600. (23) Giacomini, K. M.; Huang, S. M.; Tweedie, D. J.; Benet, L. Z.; Brouwer, K. L.; Chu, X.; Dahlin, A.; Evers, R.; Fischer, V.; Hillgren, K. M.; Hoffmaster, K. A.; Ishikawa, T.; Keppler, D.; Kim, R. B.; Lee, C. A.; Niemi, M.; Polli, J. W.; Sugiyama, Y.; Swaan, P. W.; Ware, J. A.; Wright, S. H.; Yee, S. W.; Zamek-Gliszczynski, M. J.; Zhang, L. Membrane transporters in drug development. Nat. Rev. Drug Discovery 2010, 9 (3), 215−36. (24) Bjornsson, T. D.; Callaghan, J. T.; Einolf, H. J.; Fischer, V.; Gan, L.; Grimm, S.; Kao, J.; King, S. P.; Miwa, G.; Ni, L.; Kumar, G.; McLeod, J.; Obach, R. S.; Roberts, S.; Roe, A.; Shah, A.; Snikeris, F.; Sullivan, J. T.; Tweedie, D.; Vega, J. M.; Walsh, J.; Wrighton, S. A. The conduct of in vitro and in vivo drug-drug interaction studies: a Pharmaceutical Research and Manufacturers of America (PhRMA) perspective. Drug Metab. Dispos. 2003, 31 (7), 815−832. (25) Olkkola, K. T.; Backman, J. T.; Neuvonen, P. J. Midazolam should be avoided in patients receiving the systemic antimycotics ketoconazole or itraconazole. Clin. Pharmacol. Ther. 1994, 55 (5), 481−485. (26) Chien, J. Y.; Lucksiri, A.; Ernest, C. S., II; Gorski, J. C.; Wrighton, S. A.; Hall, S. D. Stochastic prediction of CYP3A-mediated inhibition of midazolam clearance by ketoconazole. Drug Metab. Dispos. 2006, 34 (7), 1208−1219. (27) Zhao, P.; Ragueneau-Majlessi, I.; Zhang, L.; Strong, J. M.; Reynolds, K. S.; Levy, R. H.; Thummel, K. E.; Huang, S. M. Quantitative evaluation of pharmacokinetic inhibition of CYP3A substrates by ketoconazole: a simulation study. J. Clin. Pharmacol. 2009, 49 (3), 351−359. (28) Wang, E. J.; Lew, K.; Casciano, C. N.; Clement, R. P.; Johnson, W. W. Interaction of common azole antifungals with P glycoprotein. Antimicrob. Agents Chemother. 2002, 46 (1), 160−165. (29) Kim, R. B.; Wandel, C.; Leake, B.; Cvetkovic, M.; Fromm, M. F.; Dempsey, P. J.; Roden, M. M.; Belas, F.; Chaudhary, A. K.; Roden, D. M.; Wood, A. J.; Wilkinson, G. R. Interrelationship between substrates and inhibitors of human CYP3A and P-glycoprotein. Pharm. Res. 1999, 16 (3), 408−414. (30) Marcelin-Jimenez, G.; Hernandez, J.; Angeles, A. P.; Contreras, L.; Hinojosa, M.; Rivera, L.; Martinez-Rossier, L.; Amancio, O.; Fernandez, A. Bioequivalence evaluation of two brands of ketoconazole tablets (Onofin-K and Nizoral) in a healthy female Mexican population. Biopharm. Drug Dispos. 2004, 25 (5), 203−209. (31) Ahlin, G.; Chen, L.; Lazorova, L.; Chen, Y.; Ianculescu, A. G.; Davis, R. L.; Giacomini, K. M.; Artursson, P. Genotype-dependent effects of inhibitors of the organic cation transporter, OCT1: predictions of metformin interactions. Pharmacogenom. J. 2011, 11 (6), 400−411.
K
dx.doi.org/10.1021/mp400341t | Mol. Pharmaceutics XXXX, XXX, XXX−XXX