Abstract
Dabigatran etexilate, a double prodrug of dabigatran, is a reversible, competitive, direct thrombin inhibitor that has been approved for use in many countries. A recent guideline from the European Medicines Agency on drug-drug interactions proposed dabigatran etexilate as a sensitive in vivo and in vitro probe substrate for intestinal P-glycoprotein (P-gp) inhibition. We therefore performed a series of in vitro studies to determine the best experimental conditions for evaluation of P-gp involvement on the transport process of dabigatran etexilate across colorectal adenocarcinoma Caco-2 cell monolayers. Experiments using expressed carboxylesterase 1 (CES1) and CES2 bactosomes revealed that dabigatran etexilate was hydrolyzed into BIBR 1087 by CES1 expressed in our Caco-2 cells. The impact of CES1-mediated BIBR 1087 formation during transcellular transport experiments was assessed by comparing several combinations of three experimental approaches: radioactivity detection using [14C]dabigatran etexilate as substrate, liquid chromatography-tandem mass spectrometry (LC-MS/MS) quantification of dabigatran etexilate, and in the presence and absence of a CES inhibitor bis(p-nitrophenyl) phosphate (BNPP). The experimental approach that was based on the use of nonlabeled dabigatran etexilate together with LC-MS/MS quantification and the addition of BNPP was selected as the most favorable condition in which to correctly evaluate the permeability coefficient (Papp) of dabigatran etexilate and its transcellular transport by P-gp. The in vitro Caco-2 study at the selected condition revealed that dabigatran etexilate is a P-gp substrate with an efflux ratio of 13.8 and an intrinsic Papp, which is the Papp under the condition of complete blockage of P-gp by P-gp inhibitor, of 29 × 10−6 cm/s.
Introduction
Dabigatran etexilate (Pradaxa, Pradax, Prazaxa) is a novel oral, reversible direct thrombin inhibitor approved for the prevention of thromboembolism after hip and knee surgery and for secondary prevention of stroke in patients with nonvalvular atrial fibrillation (Eriksson et al., 2005). After oral administration, the inactive double prodrug dabigatran etexilate is rapidly hydrolyzed to active dabigatran via two short-lived intermediates, BIBR 1087 and BIBR 951 (Fig. 1) (Starling et al., 1997; Blech et al., 2008; Stangier, 2008) to yield the pharmacologically active drug, dabigatran. About 20% of circulating dabigatran is conjugated to glucuronic acid by UDP -glycuronosyltransferase 1 (UGT) 1A9, 2B7, and 2B15 to yield a pharmacologically equipotent glucuronide (Ebner et al., 2010). Dabigatran and its glucuronide metabolites represent >90% of the systemically available dose and are predominantly (>80%) renally excreted (Blech et al., 2008).
In recent years, research on carrier-mediated drug transport processes has helped to elucidate their central role in drug disposition (Dresser et al., 2001; Litman et al., 2001; Kusuhara and Sugiyama, 2002; Ohtsuki and Terasaki, 2007; Nakanishi and Tamai, 2012). This knowledge has provided a better understanding of factors that affect drug bioavailability, distribution, and elimination and, thereby, determination of the correct clinically effective and safe drug dose (Ayrton and Morgan, 2001; Fricker and Miller, 2002; Goh et al., 2002; Mizuno et al., 2003; Giacomini et al., 2010). The two major classes of carrier-mediated drug transporters comprise solute carrier transporters, which generally facilitate drug entry into cells, and ATP-binding cassette transporters that use energy from ATP-hydrolysis to efflux drugs from the cytoplasm into the extracelluar space (Mizuno et al., 2003). Among the latter, P-glycoprotein (P-gp) first appeared in the regulatory guidelines for investigations of drug-drug interaction (DDI) (FDA, 2012). Therefore, numerous researchers have investigated the role of P-gp on DDIs, and a plethora of in vitro data has been accumulated in transporter community.
Transcellular transport experiments across the colorectal adenocarcinoma Caco-2 cell monolayer are the most common experimental system, not only in pharmaceutical companies but also in academic institutes, to evaluate P-gp. Recent concerns about large interlaboratory differences in experimental data from transcellular experiments using Caco-2 cells have gradually spread to the transporter community (Bentz et al., 2013). Several reasons for the relatively large interlaboratory differences have been proposed and are under discussion (e.g., different source of Caco-2 cells, equations for estimating inhibition potency, pH gradient in apical-to-basal (AtoB) assay).
In June 2012, the European Medicines Agency (EMA) released guidelines on the investigation of drug interactions. These guidelines propose the use of dabigatran etexilate as one of the most sensitive probe substrates for intestinal P-gp inhibition (EMA, 2012). To obtain good in vitro-in vivo correlation, adequate in vitro experimental conditions should be set for P-gp evaluation of dabigatran etexilate. Here we describe a series of in vitro studies performed to determine the adequate experimental conditions for evaluation of P-gp involvement in the transporter process of dabigatran etexilate across the Caco-2 cell monolayer.
Materials and Methods
Chemicals
Cyclosporin A (CsA), bis(p-nitrophenyl) phosphate (BNPP), propranolol, and digoxin were obtained from Sigma-Aldrich (St. Louis, MO). [3H(G)]digoxin was obtained from PerkinElmer (Waltham, MA). d-[1-14C]Mannitol and DL-[4-3H]propranolol hydrochloride were purchased from GE Healthcare (Little Chalfort, UK). Dabigatran etexilate, [14C]dabigatran etexilate, [13C]dabigatran etexilate, and BIBR 1087 SE were synthesized by Boehringer Ingelheim Pharma GmbH and Co. KG, (Ingelheim, Germany). All other chemicals were of the highest reagent grade available from commercial sources.
Biologic Materials
Caco-2 cells were purchased from the Leibniz Institute DSMZ-German Collection of Microorganisms and Cell Cultures (Braunschweig, Germany). Expressed human carboxylesterase 1 (CES1) and CES2 bactosomes were from Cypex Limited (Dundee, Scotland).
Cell Cultures
Caco-2 cells were maintained at 37°C, 8% CO2, and 90% relative humidity in 75-cm2 cell culture flasks with 15–20 ml of Dulbecco’s modified Eagle’s medium (DMEM) culture medium (500 ml; BioChrom AG, Berlin, Germany) supplemented with 3.7 g/l NaHCO3 (Wako, Tokyo, Japan), 50 ml of fetal bovine serum (Invitrogen, Carlsbad, CA), 5 ml of MEM-nonessential amino acids, 10 mM) (Invitrogen), 5 ml of penicillin-streptomycin (Invitrogen), and 5 ml of l-glutamine (200 mM) (Invitrogen). Cells were passaged weekly using 0.25% trypsin/0.2% EDTA solution after washing with phosphate-buffered saline (−) and 0.02% EDTA solution. About 1 × 106 cells were seeded per flask. The culture medium was changed three times a week. Transwell filter inserts with a polycarbonate membrane (Corning, NY) were coated with 0.2 ml/well of collagen R solution (Serva, (Heidelberg, Germany)); 1–2 days after coating, the filter inserts were washed with phosphate-buffered saline (+). Caco-2 cells were seeded at a density of 1.5 × 105 cells/cm2 on collagen-coated transwell filter inserts. The filter inserts were placed on open reservoir plates, and the apical compartments were supplied with 0.25 ml for 24 wells, the outer wells of 24-well plate with 32 ml/plate of DMEM culture medium. The cells were cultured on 24 wells at 37°C, 8% CO2, and 90% relative humidity in DMEM culture medium for 15 to 16 days; P-gp functionality and tightness of the monolayer were confirmed as comparable to 12 wells cultured for 21 days.
Transcellular Transport Assay
Transport experiments were performed as triplicate incubations using different filter inserts for both the AtoB and basal-to-apical (BtoA) direction. Cells were equilibrated in transport buffer (pH 7.2; 128 mM sodium chloride, 5.4 mM potassium chloride, 1 mM magnesium sulfate hexahydrate, 1.8 mM calcium chloride dihydrate, 1.2 mM disodium hydrogen phosphate 12-water, 0.41 mM sodium dihydrogen phosphate dihydrate, 15 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid, 4.2 mM sodium bicarbonate, and 20 mM glucose) for 30 minutes. The apical (basal) donor side of the cell monolayer was filled with transport buffer containing the substrate with or without P-gp inhibitor (CsA) or the CES inhibitor (BNPP). The basal (apical) receiver chamber was filled with transport buffer containing 2% bovine serum albumin (BSA) supplemented with or without P-gp inhibitor (CsA) or the CES inhibitor. Assays were started after a preincubation period of 30 minutes in the presence of compounds (including substrates and P-gp or CES inhibitors). Samples (50 μl) were taken at 0 and 90 minutes from the donor compartment, representing the actual start and end concentrations in the donor compartment; those subjected to liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis were mixed with 50 μl of transport buffer containing 4% BSA in the wells of a shallow 96-well plate to prevent adsorption on the polypropylene surface. Samples from the receiver compartment were also taken at 0, 30, 60, and 90 minutes and processed in the same way as described for measuring test compound passed through the monolayer. Initial volumes in the receiver compartment were replenished with fresh transport buffer containing 2% BSA after each sampling point. Samples analyzed by LC-MS/MS were stored at −20°C or below until analysis.
Metabolism Incubation
Hydrolytic activity toward [14C]dabigatran etexilate was assessed by incubating [14C]dabigatran etexilate with expressed human CES1/2 bactosomes and measuring the formation of BIBR 1087, BIBR 951, and BIBR 953, respectively, via enzymatic hydrolysis. Incubations with bactosomes were carried out in 0.05 M phosphate buffer, pH 7.4, containing 5 mM magnesium chloride in a total volume of 200 μl. Incubations containing [14C]dabigatran etexilate were initiated by the addition of substrate and terminated by adding 200 μl of 0.2 M hydrochloric acid followed by vortex mixing. After centrifugation (10,000g for 10 minutes) at 4°C, the supernatants were directly injected into the high-performance liquid chromatography (HPLC) system.
Radio-HPLC Analysis
Samples containing [14C]dabigatran etexilate–related radioactivity (25 or 100 μl) were injected into an HPLC system composed of an autosampler HTC PAL (Chromtech GmbH, Idstein, Germany), HPLC pump PU-980 (Jasco, Gross-Umstadt, Germany), ternary gradient unit LC-980-02 (Jasco), three-line degaser DG-980-50 (Jasco), ultraviolet/visible spectroscopic detector UV-975 (Jasco), and flow scintillation analyzer TR 525 (PerkinElmer, Hamburg, Germany). Chromatography was performed for a total run time of 33 minutes on a LiChroCART Purospher RP 18-e 5-μm 125-2 analytical column (Merck, Darmstadt, Germany) with a LiChroCART Purospher RP 18-e 5-μm 10-2 guard column. As mobile phases, we used 0.05 M formic acid adjusted to pH 4.0 with ammonia solution (A) and acetonitrile (B), applying the following gradient (0 minutes: 10% (B); 21 minutes: 60% (B); 23 minutes: 80% (B); 24 minutes: 80% (B); 25 minutes: 10% (B) at a flow rate of 0.4 ml/min. In addition to flow scintillation analytes were quantified on a Sciex API 3000 triple quadrupole mass spectrometer (Applied Biosystems, Foster City, CA) operated in positive electrospray mode; 100 μl of solutions containing 497–292,000 dpm was found in a linear calibration range. The quality-control samples (second highest and second lowest concentration of the linearity assessment) were prepared at concentrations of 800 dpm/ml and 100,000 dpm/100 μl. Accuracy and precision of samples of quality control were assessed before and after each sample batch. Measurements of the quality-control samples showed assay imprecision of 14.7% CV (800 dpm/100 μl), and 4.4% CV (100,000 dpm/100-μl sample size N = 31).
Radioactivity Measurement
Radioactivity was determined in a liquid scintillation counter (LSC; TRI-CARB 3100TR and 3110TR, PerkinElmer) or a TopCount NXT (PerkinElmer). For measurement by LSC, samples transferred into the vials were mixed with a scintillation cocktail (4 ml). Radioactivity was determined by 3-minute measurement. For measurement by TopCount NXT, samples transferred into the 96-well sample collection plate were dried up at 50°C. Radioactivity (cpm) was determined by 10-minute measurement and then converted to dpm using the built-in cpm-quenching calibration curve.
LC-MS/MS Analysis for Nonlabeled Dabigatran Etexilate and BIBR 1087.
Sample preparation.
Frozen samples in 96-well shallow plates were thawed at room temperature, vortexed, and spun down at room temperature. Aliquots (10 μl) of donor solution of dabigatran etexilate taken at the start (0 minutes) and at 90 minutes were diluted 10-fold with 90 μl of transport buffer containing 2% BSA. Aliquots of donor solution of BIBR 1087 were diluted 100-fold with transport buffer containing 2% BSA to 100 μl. Diluted donor solutions and receiver solutions (100 μλ) were mixed with 100 μl of internal standard and vortexed well. After incubation at room temperature for 30 minutes, the plates were centrifuged for 5 minutes at 10,000g at 15°C. Aliquots of 130 μl of the resulting supernatants were transferred into 96-well deep plates.
Chromatographic and MS conditions.
Analytes were extracted by online solid-phase extraction (column-switching) on an Oasis HLB (25 μm, 2.1 mm × 20 mm) enrichment column (Waters Corporation, Milford, MA). Chromatography was performed on a Purospher STAR RP-18 analytical column (3 μm, 55 mm × 2 mm) (Merck). Mobile phases were 0.01 M ammonium formate buffer, pH 4.5 (A), and acetonitrile (B) using a programmed gradient [0 minutes: 8% (B), 2.1 minutes: 50% (B), 2.5 minutes: 80% (B), 3 minutes: 92% (B), 9.9 minutes: 92% (B), 10 minutes: 8% (B)] at a flow rate of 0.3 ml/min. Analytes were quantified on a Sciex API 3000 triple quadrupole MS (Applied Biosystems) operated in positive electrospray mode. Transitions from m/z = 628.4 to m/z = 289.4, from m/z = 600.3 to m/z = 289.4 and from m/z = 634.4 to m/z = 295.1 were recorded for dabigatran etexilate, BIBR 1087, and [13C]dabigatran etexilate as an internal standard for both. The obtained calibration curves were linear over the concentration range from 1.00 to 1000 nM for dabigatran etexilate and BIBR 1087. Accuracy and precision of samples of calibration samples were assessed in the analytical phase. For dabigatran etexilate, inaccuracy ranged from −15.2% to 10.6%, and precision was within 10.4%. For BIBR 1087, inaccuracy ranged from −11.8% to 10.0%, and precision was within 10.4%.
Calculations.
Net permeability coefficient (Papp).
The net permeability coefficient (Papp) was calculated from the initial radiolabeled ligand concentration in the donor compartment and the transport rate into the receiver compartment using the equations given as follows:
Papp: Permeability coefficient (cm/s)
Ct0: Radioactivity in the donor compartment at time 0 (dpm/ml or nmol/ml)
A: Area of the filter (cm2)
VR: Volume in the receiver compartment (ml)
ΔCR/Δt: Change in substance concentration over time in the receiver compartment (dpm/ml ⋅ s or nmol/ml ⋅ s)
The transport rate (VR ☓ΔCR/Δt) was calculated from the linear part of the drug concentration versus the time curve in the receiver compartment.
Efflux ratio.
The ratio of BtoA to AtoB transport was calculated using the equation given below:
Papp BtoA BtoA transport
Papp AtoB AtoB transport
Results
Hydrolysis of Dabigatran Etexilate.
Caco-2 cells express CES; however, the expression profiles of CES isoforms differ from those in human small intestine and colon tissue (Ohura et al., 2010). The involvement of CES1 and CES2 on the formation from dabigatran etexilate to the mono-prodrugs BIBR 1087, BIBR 951, as well as dabigatran was investigated under several experimental conditions (Fig. 2). Formation of BIBR 1087 from dabigatran etexilate occurred by both nonenzymatic hydrolysis and CES1-catalyzed hydrolysis, whereas formation of BIBR 951 was catalyzed by CES2 only. Caco-2 cells are reported to express CES1 rather than CES2 (Imai et al., 2005), and the level of mRNA expression of CES1 and CES2 in Caco2 cells cultured on transwell filters for 15 to 16 days was investigated. As shown in Fig. 3, Caco-2 cells used for dabigatran etexilate evaluation showed higher CES1 expression compared with CES2, comparable to previously reported results (Imai et al., 2005). Therefore, it is reasonable to hypothesize that CES1-catalyzed BIBR 1087 formation occurs during the course of the transcellular transport experiment.
Transcellular Transport of Nonlabeled Dabigatran Etexilate Using LC-MS/MS Quantification.
The effect of CES1-catalyzed BIBR 1087 formation in Caco-2 cells was assessed by conducting transcellular transport experiments with nonlabeled dabigatran etexilate, followed by LC-MS/MS quantification of both dabigatran etexilate and BIBR 1087 in the absence and presence of the CES inhibitor BNPP (Fig. 4) (Ohura et al., 2010). BNPP at a concentration of up to 200 μM did not affect the tightness of the monolayer and P-gp activity, as confirmed by comparing Papp of mannitol, propranolol, and digoxin in the absence and presence of BNPP (data not shown). Recovery rates in the AtoB direction in the absence and presence of BNPP were 13% and 68%, respectively, and those in the BtoA direction in the absence and presence of BNPP were 103% and 106%, respectively (Table 1), when calculated solely from the amounts of dabigatran extexilate retained in both donor and receiver compartments. When BIBR 1087 was additionally used for recovery calculation, recoveries for both directions exceeded 89%. These results suggest that hydrolysis of dabigatran etexilate to BIBR 951, as well as further hydrolysis from both mono-prodrugs to dabigatran, was negligible in the Caco-2 cell system we used. AtoB Papp of dabigatran etexilate in the absence and presence of BNPP was 1.1 × 10−6 cm/s and 5.2 × 10−6 cm/s, and BtoA Papp was 39 × 10−6 cm/s and 72 × 10−6 cm/s, respectively. In this experiment, a decrease of dabigatran etexilate concentration and formation of BIBR 1087 in both compartments were monitored by LC-MS/MS. For the BtoA transport assay in the absence of BNPP, BIBR1087 was not formed in the donor compartment, and the amount of BIBR 1087 in the receiver compartment was 29% relative to the sum of both dabigatran etexilate and BIBR 1087 (total dabigatran etexilate) at the end of transcellular transport assay (90 minutes). In the AtoB transport assay in the absence of BNPP, the major component was BIBR 1087, in both the donor and receiver compartments. In contrast to these results, dabigatran etexilate was the only species found in both compartments and for both directional transport assays in the presence of the CES inhibitor BNPP. The formation of BIBR 1087 in the receiver and donor compartments of the AtoB transport assay was 39% and 23% of total dabigatran etexilate, which was much smaller than that in the AtoB assay in the absence of BNPP but much larger than the BtoA transport assay in the presence of BNPP. These results clearly suggest that the extent of CES1-catalyzed BIBR 1087 formation is greater in the AtoB assay compared with the BtoA assay, regardless of BNPP in the medium.
Papp and Efflux Ratio of Dabigatran Etexilate in Caco-2 Cells in the Presence of BNPP.
To evaluate the Papp- and P-gp-mediated transport of dabigatran etexilate under complete inhibition of CES-catalyzed formation of BIBR1087, a transcellular transport assay using nonlabeled dabigatran etexilate and subsequent LC-MS/MS analysis was performed in the presence of BNPP (Table 2). Since recovery of total dabigatran etexilate during transcellular transport experiment is one of the important parameters for ensuring the validity of Papp data and P-gp profiling in transcellular transport experiment (FDA, 2012), recovery of dabigatran etexilate during the transcellular transport assay was estimated by quantifying both dabigatran etexilate and BIBR 1087. The recovery of dabigatran etexilate in the presence of BNPP was greater than 89% in both AtoB and BtoA assays, indicating that evaluation of the Papp and P-gp profile of dabigatran etexilate could be performed correctly under these conditions. The Papp of dabigatran extexilate across Caco-2 monolayers measured by LC-MS/MS was 5.2 × 10−6 cm/s (AtoB) and 72 × 10−6 cm/s (BtoA) at 1 µM substrate concentration and inhibition of CES activity, resulting in an efflux ratio of 13.8. AtoB transport increased to 15 × 10−6 cm/s at 10 µM of dabigatran etexilate without a change in BtoA transport. However, complete saturation of the transcellular transport of dabigatran etexilate beyond the maximum concentration of 10 μM used in our assay was not obtained as a result of its limited solubility. In addition, the vectorial transport represented with an efflux ratio of 13.8 was completely inhibited to an efflux ratio of 1.03 by CsA, which is a P-gp inhibitor but not an inhibitor to CES1 (IC50 value on CES1 >100 µM, data not shown) (Table 2). The intrinsic Papp under P-gp inhibition was 29 × 10−6 cm/s. These results, obtained under the most adequate experimental conditions, confirmed that dabigatran etexilate is a substrate of P-gp.
Discussion
Membrane transporters play a central role in determining drug disposition in the liver, kidney, and intestine (Ayrton and Morgan, 2001; Fricker and Miller, 2002; Goh et al., 2002; Mizuno et al., 2003; Giacomini et al., 2010). Among the efflux transporters, P-gp has gained special attention regarding DDIs, which is reflected in the latest guidance documents published by the U.S. Food and Drug Administration and the EMA. Therein, the potential for P-gp-mediated DDIs is advised to be evaluated during drug development. Caco-2 cells represent a well established in vitro system for such investigations during nonclinical phases to determine efflux ratios. In case a drug candidate is recognized by P-gp as a substrate, as defined by an efflux ratio >2 according to regulatory guidance, or as an inhibitor of P-gp, further clinical studies may be required to assess the in vivo DDI potential. It is obvious that appropriate in vitro and in vivo parameters (e.g., the choice of probe substrates, substrate concentration) are essential for this assessment. The EMA DDI guideline of 2012 proposed dabigatran etexilate, among other in vitro and in vivo test compounds, as a probe substrate for the assessment of intestinal P-gp inhibition (EMA, 2012). Considering its double prodrug nature, a thorough assessment and understanding of in vitro P-gp interaction of dabigatran extexilate are required because the results may trigger extensive clinical DDI assessment. Caco-2 cells express CES1 (Imai et al., 2005), unlike the human small intestine or colon tissue, where CES2 is expressed and dabigatran etexilate requires esterase activity to form pharmacologically active dabigatran. In vitro metabolism studies using CES1/2 bactosomes revealed that the two intermediate prodrugs are formed both by nonenzymatic and CES1-catalyzed hydrolysis (BIBR 1087) and by CES2 only (BIBR 951).
The impact of CES1 expressed in Caco-2 cells on transcellular transport assay was evaluated using nonlabeled dabigatran etexilate and LC-MS/MS quantification of both dabigatran etexilate and BIBR 1087 (Fig. 4; Table 1). In the absence of CES inhibitor BNPP in AtoB transport assay, formation of BIBR 1087 was observed to a large extent in both receiver and donor compartments. To elucidate the possible mechanisms of this finding, we performed Papp and P-gp profiling of BIBR 1087 across the Caco-2 monolayer in the presence of BNPP. The Papp of BIBR 1087 at 10 µM was 0.89 × 10−6 cm/s, and the efflux ratio was 1.00. The substantial formation of BIBR 1087 in the donor compartment of the AtoB assay could be explained by esterases expressed on the outer apical membranes of Caco-2 cells (Ohura et al., 2011), as BIBR 1087 will not pass the apical membranes once formed intracellularly as a result of the low Papp and not being a P-gp substrate. This explanation is supported by the fact that BIBR 1087 formation from dabigatran etexilate in human liver microsomes is not completely inhibited by BNPP, which is a selective CES inhibitor, but by paraoxon, which is a strong inhibitor of both CES and serine esterases (Blech et al., 2008). The appearance of BIBR 1087 in the receiver compartment of the AtoB assay is likely mediated by CES1 activity at the endoplasmatic reticulum of Caco-2 cells. Dabigatran etexilate is thought to enter the cell as a result of its high Papp (29 × 10−6 cm/s) (Table 2) and is then intracellularly hydrolyzed by CES1 to BIBR 1087, which then slowly appears in the receiver compartment. BIBR 1087 was found in both compartments in the AtoB assay in the presence of BNPP, indicating that 200 µM BNPP is not sufficient to completely block esterase activity on the apical membrane of Caco-2 cells (Ohura et al., 2011). Therefore, an experimental condition using [14C]dabigatran etexilate as probe substrate and simple radioactivity detection is not considered adequate for correct evaluation of Papp and P-gp interaction, even in the presence of 200 µM BNPP. On the other hand, the appearance of BIBR 1087 in both the receiver and donor compartments of the BtoA assay was far smaller than in the AtoB assay. Dabigatran etexilate concentrations in the donor compartment of BtoA assays were unchanged under any conditions. Therefore, the BtoA assay will give more reliable results for the Papp and P-gp interaction profile of dabigatran etexilate, regardless of the chosen experimental condition. We conclude that the most favorable experimental condition for the evaluation of Papp and the P-gp interaction of dabigatran etexilate includes using nonlabeled dabigatran etexilate as substrate, LC-MS/MS quantification, and the presence of BNPP in the assay. The Papp and P-gp profile of dabigatran etexilate was evaluated using these conditions, and the correct Papp and P-gp profiles were obtained (Table 2).
Since dabigatran etexilate was proposed as one of the in vitro and in vivo P-gp probe substrates by EMA, more in vitro IC50 determinations of various P-gp inhibitors on P-gp-mediated dabigatran etexilate transcellular transport across Caco-2 cell monolayers are likely to be conducted in the future by different laboratories. To identify the most suitable experimental conditions for correct IC50 values of P-gp inhibitors on P-gp-mediated dabigatran etexilate transcellular transport by using a cost- and time-friendly Caco-2 assay system, the risks of generating overestimations or underestimations at four experimental conditions are compared in Table 3. Regarding AtoB Papps and efflux ratios, only the experimental conditions of nonlabeled dabigatran etexilate/LC-MS/MS quantification/+BNPP returned the correct value. Therefore, the AtoB and efflux ratio at three experimental conditions other than the best one are not considered acceptable for correct IC50 determination. BtoA Papp in all four experimental conditions seems reliable, but a small overestimation of dabigatran etexilate concentration was observed in the receiver compartment of the BtoA assay in experimental conditions using [14C]dabigatran etexilate, radioactivity detection, and without BNPP. Although the small overestimation is critical for P-gp/Papp profiling of dabigatran etexilate because of overestimation of BtoA Papp by recognizing both dabigatran etexilate and BIBR 1087 as dabigatran etexilate, it is acceptable for IC50 determination only in cases in which putative P-gp inhibitors are not inhibitors of CES1.
Taken together, the series of in vitro studies we have conducted clearly suggest that evaluation of Papp and P-gp profiling of dabigatran etexilate can be complex and prone to misjudgment if standard experimental conditions are applied. Herein we present not only adequate experimental conditions for Papp/P-gp profile of dabigatran etexilate but also optimized conditions for IC50 assessment of putative P-gp inhibitors. By sharing this information, we aim to resolve interlaboratory differences of Caco-2 data when dabigatran etexilate is used as in vitro probe substrate for the prospective assessment of DDIs.
Acknowledgments
The authors thank A. Saito, M. Takatani, I. Kitahara, Dr. A. Fukuhara, M. Schuster, P. Diemar-Epple, and S. Stohr for excellent technical assistance in conducting the in vitro experiments. [14C]Dabigatran etexilate and [13C]dabigatran etexilate were kindly provided by Ralf Kiesling, head of the isotope chemistry laboratory, BI PharmaGmbH & Co. KG, Germany.
Authorship Contributions
Participated in research design: Ishiguro, Volz, Ludwig-Schewellinger, Kishimoto, Ebner.
Conducted experiments: Ishiguro, Volz, Ludwig-Schewellinger, Kishimoto.
Performed data analysis: Ishiguro, Volz, Ludwig-Schewellinger, Kishimoto
Wrote or contributed to the writing of the manuscript: Ishiguro, Kishimoto, Volz, Ludwig-Schewellinger, Ebner, Schaefer.
Footnotes
- Received July 18, 2013.
- Accepted November 6, 2013.
This study was supported by Boehringer Ingelheim.
The authors are fully responsible for all content and editorial decisions, were involved at all stages of manuscript development, and have approved the final version.
Abbreviations
- AtoB
- apical-to-basal
- BNPP
- bis(p-nitrophenyl) phosphate
- BSA
- bovine serum albumin
- BtoA
- basal-to-apical
- CES
- carboxylesterase
- CsA
- cyclosporin A
- DDI
- drug-drug interaction
- DMEM
- Dulbecco’s modified Eagle’s medium
- EMA
- European Medicines Agency
- HPLC
- high-performance liquid chromatography
- LC-MS/MS
- liquid chromatography-tandem mass spectrometry
- LSC
- liquid scintillation counter
- Papp
- apparent permeability coefficient
- P-gp
- P-glycoprotein
- Copyright © 2013 by The American Society for Pharmacology and Experimental Therapeutics