Abstract
An increased appreciation of the importance of transporter and enzyme interplay in drug clearance and a desire to delineate these mechanisms necessitates the utilization of models that contain a full complement of enzymes and transporters at physiologically relevant activities. Additionally, the development of drugs with longer half-lives requires in vitro systems with extended incubation times that allow characterization of metabolic pathways for low-clearance drugs. A recently developed coculture hepatocyte model, HepatoPac, has been applied to meet these challenges. Faldaprevir is a drug in late-stage development for the treatment of hepatitis C. Faldaprevir is a low-clearance drug with the somewhat unique characteristic of being slowly metabolized, producing two abundant hydroxylated metabolites (M2a and M2b) in feces (∼40% of the dose) without exhibiting significant levels of circulating metabolites in humans. The human HepatoPac model was investigated to characterize the metabolism and transport of faldaprevir. In human HepatoPac cultures, M2a and M2b were the predominant metabolites formed, with extents of formation comparable to in vivo. Direct glucuronidation of faldaprevir was shown to be a minor metabolic pathway. HepatoPac studies also demonstrated that faldaprevir is concentrated in liver with active uptake by multiple transporters (including OATP1B1 and Na+-dependent transporters). Overall, human HepatoPac cultures provided valuable insights into the metabolism and disposition of faldaprevir in humans and demonstrated the importance of enzyme and transporter interplay in the clearance of the drug.
Introduction
Faldaprevir (BI 201335; FDV) is a novel, selective inhibitor of the hepatitis C virus (HCV) NS3/NS4 protease that is currently showing high efficacy in phase 3 clinical trials in HCV genotype 1–infected patients (Sulkowski et al., 2013a,b). Recent drug-drug interaction (DDI) guidances from regulators indicates that, where possible, sponsors should attempt to define relevant clearance pathways and characterize the enzymes involved in metabolic clearance (Food and Drug Administration, 2012; European Medicines Agency, 2012). During preclinical evaluations, significant enrichment of faldaprevir into the target organ, liver, was observed in rats dosed with faldaprevir (White et al., 2010), which suggests a possible role of uptake transporters. Transporters are a key aspect of intracellular concentrations in the liver (Chu et al., 2013). A number of in vitro cell-based systems can be used to define the interaction of a drug with transporters (Ghibellini et al., 2006; Brouwer et al., 2013). It is much more challenging to identify an appropriate in vitro cell system when attempting to integrate transport with metabolism. For example, there are known limitations of suspended hepatocytes and sandwich-cultured hepatocytes in the expression of drug transporters, drug-metabolizing enzymes (DMEs), or both (Bow et al., 2008; Swift et al., 2010).
Faldaprevir is a low-clearance drug in humans (Duan et al., 2012) that predominantly undergoes oxidation to form two abundant hydroxylated metabolites (M2a and M2b) detected in feces at ∼40% of the administered dose without exhibiting appreciable levels of circulating metabolites (Chen et al., 2014). To effectively investigate the interplay of faldaprevir with transporters and DMEs, it was important to select a cellular system in which sufficient levels of metabolites would be generated for this slowly metabolized drug. Addition of metabolites to cell systems in vitro does not guarantee achieving clinically relevant intracellular concentrations. An additional challenge for faldaprevir is that metabolism in humans (mainly through oxidation) is different than in the toxicology species (mice, rats, and monkeys) because animals have significantly lower levels of M2a and M2b excreted and glucuronidation is the major pathway in the rat (Ramsden et al., 2014). So animals are not good surrogates to study the clearance of faldaprevir. In addition, although no glucuronide was observed in the human [14C] absorption, distribution, metabolism, and excretion (ADME) study, glucuronidation of faldaprevir cannot be fully excluded because levels of parent drug excreted in feces can be a function of nonabsorbed drug, direct excretion of parent in bile, and/or excretion of a glucuronide metabolite of parent that is then subsequently cleaved by gut bacteria (Sousa et al., 2008). Therefore, the extent of faldaprevir glucuronidation needed to be evaluated using in vitro models.
HepatoPac is a coculture of hepatocytes with mouse fibroblasts that has been shown to maintain hepatocyte function (metabolism and canalicular efflux transport) for extended periods of time, for example, up to 6 weeks for albumin and urea secretion (Khetani and Bhatia, 2008). In our hands, high DME activities for cytochromes P450 have been effectively maintained routinely for 2.5 weeks (data on file, Boehringer Ingelheim Pharmaceuticals, Inc.). Additionally, HepatoPac can accurately replicate in vivo metabolite profiles (Wang et al., 2010) and has the capability of determining the intrinsic clearance of low-clearance drugs by allowing incubation for up to 7 days without a change in medium (Chan et al., 2013). Our studies with rat HepatoPac have also demonstrated an in vitro–to–in vivo correlation in the rat for faldaprevir liver uptake and metabolism, especially glucuronidation (Ramsden et al., 2014). Based on these criteria, human HepatoPac was considered an appropriate in vitro model and was investigated as an integrated system in an attempt to characterize the metabolism and transport of faldaprevir, with a view to explain clinical findings as well as attempt to develop models for predicting clinical DDIs.
Materials and Methods
Human HepatoPac cultures, with proprietary maintenance medium and serum-free probing medium, were acquired from Hepregen Corporation (Medford, MA). The cultures were prepared from cryoplateable human hepatocytes purchased from Invitrogen Life Technologies (Grand Island, NY). Rifamycin SV and sodium butyrate was purchased from Sigma-Aldrich Inc. (St. Louis, MO). Radiolabeled rosuvastatin and taurocholic acid (TCA) were purchased from American Radiochemicals (St Louis, MO) and PerkinElmer (Waltham, MA), respectively. Faldaprevir, [14C]faldaprevir, d7-faldaprevir, IN79158 (M2a), IN79157 (M2b), and BI 203279 (faldaprevir glucuronide) (see structures of faldaprevir and its metabolites in Table 1) were synthesized at Boehringer Ingelheim Pharmaceuticals, Inc. (Ridgefield, CT). All other reagents and solvents were of analytical grade or higher purity and were obtained from commercial suppliers.
Cell Culture
Human HepatoPac cultures were prepared at Hepregen Corporation. A female Caucasian donor was selected from a panel of seven available donors for preparation of HepatoPac cultures. This donor was screened against other donors and determined to be representative of average CYP3A and glucuronidation activities and was also genotyped as OATP1B1 wild type *1a/*1a and CYP3A5*3*3. Cryopreserved human hepatocytes were seeded with 32,000 (in a 24-well plate) or 5000 (in a 96-well plate) hepatocytes per well and cultured for 9 days post-seeding before initiating experiments. The maintenance medium was supplied by Hepregen and contained 10% bovine serum. For incubations, equal volumes of proprietary serum-free probing medium and maintenance medium, both supplied by Hepregen, were mixed, providing human HepatoPac incubation medium with 5% bovine serum (hHIM). The plates were cultured in an incubator with 10% CO2 and 99% relative humidity at 4°C or 37°C.
Protein Binding in hHIM
Teflon dialysis cells (Spectrum, Rancho Dominguez, CA) and dialysis membranes (Spectra/Por) with 12,000–14,000 mol. wt. cutoff were used for equilibrium dialysis to determine binding of [14C]faldaprevir (0.3–21 μM; n = 5 replicates per concentration) in hHIM. Dialysis cells were incubated at 37°C for 4 hours. After that, the contents of each side were transferred to scintillation vials and processed by liquid scintillation counting.
Metabolite Profiling
[14C]Faldaprevir was incubated with human HepatoPac cultured in a 24-well plate format, with and without pretreatment of rifampicin (25 μM), for 48 hours. Reactions were initiated by adding faldaprevir (final concentration of 5 μM in hHIM) to human HepatoPac plates or control fibroblast-only plates (plates containing only fibroblast feeder cells). At 24, 48, 72, and 96 hours, the entire medium (400 μl) was removed from each well and transferred to 2-ml Fisher low-bind vials (Fisher Scientific, Pittsburgh, PA) containing 600 μl of reaction termination solution (99.9% acetonitrile and 0.1% acetic acid). These samples were designated as the medium samples. The wells were quickly washed twice with ice-cold medium to remove residual compound. After the washes were complete, an equivalent volume of medium (400 μl) was added to each well and the cells were scraped from the bottom of the well using the sharp end of a pipette tip to aid in detachment. These samples, designated as the lysate samples, were then transferred and treated in the same manner as the medium samples. Removal of cell monolayers was confirmed microscopically. After centrifugation of both medium and lysate samples, the supernatants were evaporated under a stream of nitrogen gas using a rotary evaporator at 37°C. Post-evaporation samples were reconstituted in 100 μl of the reaction termination solution and 50 μl of water. Both medium and lysate samples were analyzed by liquid chromatography–radiochromatography-tandem mass spectrometry (LC-radiochromatography-MS/MS).
Determination of Metabolite Formation Rate and Liver Enrichment
Faldaprevir was incubated with human HepatoPac in 96-well plates at final incubation concentrations of 0.3, 1, 6, and 21 μM in hHIM (65 μl). The concentrations were chosen to cover the free and total concentrations observed at clinically relevant doses of faldaprevir. At various time points up to 96 hours, the entire aliquot of medium was removed and added to 130 μl of reaction termination solution consisting of acetonitrile/water [60:40 (v/v), containing 0.1% acetic acid and 0.1 μM d7-faldaprevir, an internal standard]. These samples were designated as medium samples. The cells were washed twice with ice-cold blank medium and resuspended in blank medium (65 μl). The lysate samples were then prepared in the same manner as the medium samples. All treatments were also repeated on control fibroblast-only plates. Samples were filtered through a membrane via centrifugation at 4°C for 15 minutes, and the supernatant was analyzed by LC-MS/MS.
Faldaprevir Glucuronide Stability Study
To evaluate its stability, faldaprevir glucuronide, at final concentrations of 0.1 and 1 μM, was incubated with blank plates containing hHIM only, fibroblast-only plates, or human HepatoPac plates using 96-well plates. Medium and lysate samples were removed at designated time points (0, 0.5, 1, 2, 4, 8, 12, 16, 22, 24, 48, and 72 hours) and analyzed by LC-MS/MS. Half-lives (t1/2) for glucuronide degradation were determined.
Active Uptake of Faldaprevir in the Absence and Presence of Rifamycin SV
Reactions were initiated for each experiment by adding designated concentrations of faldaprevir (0.01–100 μM) in hHIM (65 μl) to HepatoPac and fibroblast-only plates in the absence and presence of rifamycin SV (100 μM). The plates were cultured at 4°C or 37°C. At each time point (0, 2.5, 5, 15, 60, 120, 240, 480, 720, and 1440 minutes), medium and lysate samples were removed and analyzed by LC-MS/MS. Uptake of radiolabeled rosuvastatin (0.1 μM) was used as a control for OATP-mediated uptake.
Evaluation of Faldaprevir as an OATP Substrate
A transient transfection system was used to express OATP transporters. Prior to transfection, cDNAs were amplified and characterized. Uptake assays, using a prototypical radiolabeled substrate (cholecystokinin-8 for OATP1B3, estradiol 17-β-d-glucuronide for OATP1B1, and estrone 3-sulfate for OATP2B1), were used to confirm functional expression. HEK293 cells were seeded into 12-well plates coated with poly-d-lysine at 400,000 cells/ml. At 16–24 hours post-seeding, cells were transfected with cDNA containing a transporter of interest or vector control using the FuGENE-6 transfection reagent, per the FuGENE-6 manufacturer’s protocol (Fugent LLC and Promega, Madison, WI). To increase expression, at 24 hours post-transfection, 5 mM sodium butyrate was added. At 48 hours post-transfection, uptake of 1 μM [14C]faldaprevir at 37°C in OptiMEM (GIBCO Life Technologies, Grand Island, NY) with 4% bovine serum albumin (BSA) (pH 7.4) was determined in the absence and presence of rifamycin SV (100 μM). At designated time points the reaction was stopped by aspirating medium and washing each well three times with cold OptiMEM. After washing, the cells were lysed with 1% SDS and analyzed by scintillation counting.
Sodium-Dependent Uptake of Faldaprevir
Sodium-dependent uptake of faldaprevir was evaluated using osmotically balanced transport buffers with sodium (Na+) or without sodium (−Na+). Transport buffers (pH 7.4) were prepared as follows: 20 mM HEPES, 15 mM glucose, 1.25 mM MgSO4, 1.2 mM KH2PO4, 4.7 mM KCl, 1.2 mM CaCl2, and either 110 mM NaCl and 25 mM NaHCO3 (for Na+) or 110 mM C5H14ClNO and 25 mM KHCO3 (for –Na+). The −Na+ buffer was osmotically balanced using potassium and choline. Additional Na+ and −Na+ buffers were also prepared containing 5% of physiologic BSA content (2 mg/ml) by adding 200 mg BSA to 100 ml of each transport buffer. Reactions were initiated for each experiment by adding faldaprevir (0.1, 0.3, 3, and 30 μM) in fresh hHIM (65 μl) or transport buffers (± Na+ in the absence and presence of BSA) to HepatoPac and fibroblast-only plates. Incubations containing the inhibitor rifamycin SV (100 μM) were also included. At each time point (0, 5, 10, 15, 30, 60, 120, and 240 minutes), sampling was performed as previously described and samples were analyzed by LC-MS/MS. Uptake of [3H]TCA (0.1 μM) was used as a control for sodium-dependent uptake [sodium taurocholate cotransporting peptide (NTCP)].
Equipment and Chromatographic Conditions
Samples of [14C]faldaprevir were analyzed by LC-LTQ Orbitrap (Thermo Scientific, San Jose, CA) coupled with a flow scintillation radiomatic detector (PerkinElmer). Metabolite separation was achieved using a Phenomenex Gemini C18, 3 μm, 150 mm × 4.6 mm column (Phenomenex, Torrance, CA). The aqueous and organic mobile phase consisted of 95:5 (v/v) water/acetonitrile (A) and 95:5 (v/v) acetonitrile/water (B), respectively. Both mobile phases contained 0.1% acetic acid. The gradient conditions were 100% A for 5 minutes, changed to 15% B over 15 minutes, changed to 75% B over 40 minutes and held at 100% B for 5 minutes. The LC flow rate was 0.7 ml/min. A sample splitter was used to separate the LC flow with 1 part going to the mass spectrometer and 19 parts to the radiomatic detector.
The metabolite formation and uptake samples were analyzed on a 4000 Qtrap (AB Sciex, Thornhill, Ontario, Canada) attached to an Acquity ultra performance LC system (Waters Corp., Milford, MA). The aqueous and organic mobile phase consisted of 95:5 (v/v) water/acetonitrile (A) and 95:5 (v/v) acetonitrile/water (B), respectively. Both mobile phases contained 0.1% acetic acid. Samples were eluted through an Acquity ultra performance LC column BEH C18 1.7 μm (2.1 mm × 50 mm) following a 10-minute gradient with 5% B to 38% B over 1 minute at a flow rate of 0.5 ml/min, to 51.5% B over 9 minutes at a 0.7 ml/min flow rate. Multiple reaction monitoring analysis was performed in positive ionization mode to quantify levels of faldaprevir (869.3 → 422.1), M2a and M2b (885.3 → 422.1), faldaprevir glucuronide (1047.2 → 871.1), and internal standard d7-faldaprevir (876.3 → 429.1).
Data Analysis
Correction of the Faldaprevir Glucuronide Formation Rate with Degradation Rate.
The t1/2 of faldaprevir glucuronide under incubation conditions was determined. The rate of formation of faldaprevir glucuronide was corrected for the rate of degradation using eq. 1 (Spahn-Langguth and Benet, 1993):(1)where G = amount of glucuronide degraded at each time point; k0 = zero-order initial degradation rate; kd = degradation rate, determined as 0.693/t1/2; and t = time.
Scaling Rates of Formation of M2a, M2b, and Faldaprevir Glucuronide.
In vivo formation rates of M2a and M2b in humans were calculated using eq. 2:
(2)In vitro formation rates of M2a, M2b, and faldaprevir glucuronide were scaled to total liver using eq. 3 assuming 1.2 × 108 hepatocytes per g of liver and 1800 g of liver per human (Davies and Morris, 1993; Houston and Carlile, 1997). Faldaprevir glucuronide (pmol) formed was first corrected based on eq. 1.
(3)Deriving Km and Vmax Values.
The formation rates were calculated for each metabolite. Faldaprevir glucuronide formed was first corrected based on eq. 1. Free substrate concentrations were used after correction for protein binding, determined by equilibrium dialysis. The Km and Vmax parameters were determined by nonlinear regression analysis using GraphPad Prism 5 (GraphPad, Inc., La Jolla, CA) and fitted to the Michaelis-Menten equation (eq. 4):(4)where [S] is free substrate concentration, v is reaction velocity in pmol/min/total liver, Vmax is the normalized maximum reaction velocity in pmol/min/total liver, and Km is the Michaelis-Menten constant.
Estimation of Clearance.
Intrinsic clearance (CLint) for each metabolic pathway was derived from eq. 5 using therapeutically appropriate free substrate concentrations [S] based on steady-state Cmax observed during clinical trials, corrected for plasma protein binding.
(5)The Fm values for cytochrome P450– and UDP glucuronosyltransferase (UGT)-mediated metabolism were determined by comparing the respective individual CLint to the total CLint at each substrate concentration.
Calculation of Intracellular Concentrations and Enrichment.
The intracellular concentrations of faldaprevir and metabolites were corrected for fibroblast binding. Fibroblast control plates were treated in the same manner as the HepatoPac plates. Because each well in HepatoPac plates has 75% surface area as fibroblasts and 25% surface area as hepatocytes (Khetani and Bhatia, 2008; Khetani et al., 2013), the total pmol/ml values determined from the fibroblast lysate were multiplied by 0.75 and then subtracted from the values determined in HepatoPac plates, to exclude the amount of analytes nonspecifically binding to fibroblasts. The hepatocyte-specific concentration (pmol/ml) was calculated using eq. 6:
(6)This concentration was corrected with the hepatocyte intracellular volume (0.0000324 ml for 5000 hepatocytes) to calculate the intracellular concentration (pmol/ml) (eq. 7). A hepatocyte volume of 6.48 pl/hepatocyte was used (personal communication with Dr. Kenneth Brouwer from Qualyst Transporter Solutions, Durham, NC, 2010). This value is in line with other literature values (Swift et al., 2010).
(7)The enrichment value was calculated using eq. 8:
(8)To estimate the expected liver enrichment at therapeutically relevant concentrations, a one-site binding model was used to determine the kinetics of enrichment. The reciprocal values of enrichment (Y) were plotted on the y-axis and free incubation concentrations were plotted on the x-axis to derive the Kd (equilibrium binding constant) and Bmax (1/maximum binding) using eq. 9:
(9)Determination of Kinetic Parameters for Uptake of Faldaprevir.
The uptake rates of faldaprevir (pmol/min/million cells) were calculated for total uptake (active uptake and passive permeability), active uptake inhibited by rifamycin SV, and active uptake not inhibited by rifamycin SV (calculated as total active uptake minus active uptake inhibited by rifamycin SV) after incubation with HepatoPac for 240 minutes. The uptake rates were corrected for nonspecific binding to the cocultured fibroblasts as described above (eq. 6). The Km and Jmax parameters were determined by nonlinear regression analysis using GraphPad Prism 5 and fit to the Michaelis-Menten equation using free substrate concentrations. CLint for total uptake, uptake inhibited by rifamycin SV, and uptake not inhibited by rifamycin SV was derived from eq. 10:
(10)Passive CLint (CLint,passive) was calculated after subtracting active uptake from the total uptake using eq. 11.
(11)Estimation of Total Hepatic Clearance.
The total hepatic blood clearance (CLh,B) in humans was estimated with three models, including the standard well stirred model (eq. 12) (Yang et al., 2007) and two modified well stirred models incorporating hepatic uptake (eq. 13) and the binding difference in liver and plasma (eq. 14) (Webborn et al., 2007; Poulin et al., 2012), with details described previously (Li et al., 2014).
(12)(13)(14)where Qh refers to hepatic blood flow (1617 ml/min), fu,p is the free fraction of faldaprevir in plasma (0.002), Cb/Cp is the blood-to-plasma ratio (0.613), and fu,liver is the free fraction of drug in liver (estimated as 0.018) (Li et al., 2014). CLint,met is the sum of intrinsic clearance values for the formation of M2a, M2b, and faldaprevir glucuronide in the liver. The scaling factor of hepatic microsomal recovery was 40 mg/g of liver (Barter et al., 2007), and the total liver weight was 1800 g per 70 kg body weight (Davies and Morris, 1993).
Results
Overall Metabolic Profile for [14C]Faldaprevir Incubation with Human HepatoPac
Metabolite profiling and identification were performed after incubation of [14C]faldaprevir at 5 μM with human HepatoPac for up to 6 days. The structures of the metabolites were proposed based on LC-MS/MS analysis and are shown in Table 1. The metabolites were designated according to their original names in the [14C] ADME study in humans (Chen et al., 2014), and the same LC method was used. Quantitative results for the extents of formation of metabolites at 96 hours in the HepatoPac incubation medium were compared with those from the [14C] ADME study in humans (Chen et al., 2014) and are shown in Table 2. M2a and M2b were observed starting from 4 hours, and at 96 hours M2a and M2b were the primary metabolites formed, accounting for 21% and 45%, respectively, of total metabolites. Faldaprevir glucuronide was only detected after 96 hours and represented 9.4% of the total metabolites. The radioactivities of lysate samples were too low to be analyzed using LC-radiochromatography. Downstream metabolites of these metabolites, such as hydroxylation of faldaprevir glucuronide, were not detected based on extraction of molecular ions or multiple reaction monitoring transitions.
The effect of enzyme induction on faldaprevir metabolism was evaluated by preincubating HepatoPac with a known pregnane X receptor agonist (rifampicin 25 μM) for 48 hours followed by a 96-hour incubation with 5 μM faldaprevir. Upon induction by rifampicin, faldaprevir levels decreased by 40%, while the levels of M2a, M2b, and glucuronide increased by 3.5-, 1.6-, and 1.6-fold, respectively.
Determination of Kinetic Parameters for Faldaprevir Metabolism with Human HepatoPac
Faldaprevir at total concentrations of 0.3, 1, 6, and 21 μM was incubated with human HepatoPac. All three metabolites were detected in the medium samples, while only M2a and M2b were detected in the lysate samples. Representative profiles for metabolite formation over time at 1 μM of faldaprevir are shown in Fig. 1.
The formation rates of M2a and M2b, calculated by adding the formation rates from the medium and lysate samples at each time point, were shown to be linear up to 96 hours (Fig. 1). The t1/2 of faldaprevir glucuronide was consistent across concentrations (0.1 and 1 μM) and matrices (HepatoPac plates, fibroblast-only plates, and hHIM only) and averaged 30.9 ± 9.4 hours. Because the t1/2 of faldaprevir glucuronide did not decrease between cells and medium-only incubations, depletion of faldaprevir glucuronide was not mediated by metabolic processes. Therefore, formation rates of faldaprevir glucuronide were calculated after correction of its degradation rate (eq. 1) and were also shown to be linear up to 96 hours (Fig. 1).
To generate the free Km, the free fractions of faldaprevir in hHIM (fu,in vitro) were measured by equilibrium dialysis at faldaprevir concentrations of 0.3, 1, 6, and 21 μM and were used to calculate the free substrate concentrations (Table 3). The formation rates and the free substrate concentrations were then fitted to the Michaelis-Menten equation to derive the free Km and Vmax (pmol/ml/total liver) values for formation of M2a, M2b, and faldaprevir glucuronide (Fig. 2; Table 4). Using these data, the CLint values for each pathway were defined at therapeutically relevant concentrations of faldaprevir and the fraction of metabolism undergoing glucuronidation was estimated to be 11%–12% (Table 4).
Comparison of In Vitro and In Vivo Formation Rates of M2a and M2b
At total concentrations of 0.3 and 1 μM faldaprevir, with corresponding free concentrations (Table 3) close to free Cmax (0.0183 μM) in the human [14C] ADME study (Chen et al., 2014), the rates of formation of M2a and M2b with human HepatoPac were comparable to the in vivo formation rates estimated from the ADME study (Table 4). Faldaprevir glucuronide was not observed in the human [14C] ADME study and thus was not included in the comparison.
Liver Enrichment in HepatoPac.
Enrichment factors were calculated at total faldaprevir concentrations of 0.3, 1, 6, and 21 μM based on eq. 8. There was a time dependence observed for the extent of enrichment, whereby at earlier time points the enrichment values were lower than at later time points. Therefore, the liver enrichment values were calculated after equilibrium was reached, excluding the early time points (0, 0.5, and 1 hour). In addition, there was a trend of lower enrichment values at higher substrate concentrations, suggesting saturation of liver enrichment at higher substrate concentrations. Using a one-site binding model (Fig. 3), free Kd and Bmax (i.e., 1/maximal binding) values were determined to be 0.0049 μM and 0.0502 μM (corresponding to 20-fold maximal binding), respectively, which were used to calculate the extent of enrichment at therapeutically relevant faldaprevir concentrations (Table 5). The enrichment values were calculated as 21.9- to 31.5-fold for 240-mg and 120-mg daily (QD) doses, respectively.
Uptake into HepatoPac and Evaluation of Transporter Interactions
Active OATP-mediated uptake was confirmed in the HepatoPac model using rosuvastatin, a prototypical OATP substrate, as indicated by greater accumulation of rosuvastatin into hepatocytes at 37°C compared with 4°C (Fig. 4A). Accumulation of rosuvastatin at 37°C was also completely inhibited in the presence of an OATP inhibitor, rifamycin SV.
Significant passive permeability (temperature independent) and active uptake (temperature dependent) into cells were observed in the incubations of faldaprevir from 0 to 60 minutes (Fig. 4B). Coincubation of rifimycin SV largely inhibited the active uptake component at these early time points. The total uptake of faldaprevir increased at later time points (4 to 24 hours) (Fig. 4C). Coincubation of rifamycin SV, however, only partially inhibited the active uptake component, suggesting the presence of other uptake mechanism(s) that cannot be inhibited by rifamycin SV. Using transfected HEK293 cells expressing OATP1B1, an ∼3-fold higher uptake of faldaprevir compared with vector control cells was observed and the uptake was completely inhibited by rifamycin SV, confirming that faldaprevir is a substrate for OATP1B1 (Fig. 5). Uptake of faldaprevir was limited (<2-fold compared with vector control cells) in HEK293 cells expressing OATP1B3 and OATP2B1, suggesting that faldaprevir is not a substrate of those transporters.
TCA uptake is mediated by NTCP and was used as a positive control substrate for sodium-dependent uptake. The uptake of TCA over time was evaluated in the transport buffers (± sodium and ± 5% BSA) and hHIM, and was linear up to 30 minutes (Fig. 6A). In the absence of sodium, the total uptake of TCA decreased by an average of 86%, confirming that NTCP was functional in the human HepatoPac model (Fig. 6B). The presence of BSA had no impact on the uptake of TCA. The uptake in hHIM was similar to the uptake in the transport buffer containing sodium (Fig. 6A), indicating that the maximal sodium-dependent uptake was observed in incubations using hHIM. Rifamycin SV inhibited sodium-dependent TCA uptake by 76% (Fig. 6B). Uptake of rosuvastatin was also assessed in this study. Its uptake was decreased by 32% at 30 minutes in the absence of sodium and was significantly reduced in the presence of rifamycin SV (Fig. 7).
In the absence of sodium, the average uptake of faldaprevir, across all time points and concentrations, was reduced by 55.3% ± 8.6% compared with the uptake in the presence of sodium (Fig. 8). Rifamycin SV completely inhibited the sodium-dependent uptake of 0.3 μM faldaprevir (Fig. 8A) but only slightly inhibited the total sodium-dependent uptake (∼20%) at 3 μM faldaprevir (Fig. 8B).
Determination of Km and Jmax Values for Uptake of Faldaprevir.
Uptake rates of faldaprevir were determined at total substrate concentrations of 0.01, 0.03, 0.1, 0.3, 1, 3, 30, and 100 μM at different time points. To calculate free Km, the total substrate concentrations were corrected with free fractions determined based on a linear regression model of the fu,in vitro data obtained using 0.3, 1, 6, and 21 μM faldaprevir (Table 3). As mentioned before, faldaprevir uptake into hepatocytes in the HepatoPac model reached equilibrium after 4 hours. Therefore, the kinetic profiles of faldaprevir were generated for total uptake (active uptake + passive permeability), active uptake inhibited by rifamycin SV, and active uptake not inhibited by rifamycin SV using 240-minute data points, which were fitted to the Michaelis-Menten model (Fig. 9).
Kinetic parameters for the different transport processes, including Km, Jmax, and CLint, were calculated (Table 6). The passive permeability component was described by subtracting the active process CLint from the total (active + passive) CLint. The total uptake inhibited by rifamycin SV accounted for 35%–45% of the total uptake clearance at therapeutically relevant concentrations. Passive permeability appeared to comprise 24%–30% of the total uptake. The uptake that was not inhibited by rifamycin SV accounted for 31%–35% of the total uptake at therapeutically relevant concentrations (Table 6).
Calculating the Hepatic Clearance of Faldaprevir.
The hepatic clearance of faldaprevir was determined using the well stirred liver model, a modified well stirred liver model incorporating uptake kinetics (Webborn et al., 2007), and a modified well stirred liver model incorporating the binding difference in liver and plasma (Poulin et al., 2012). The hepatic clearances estimated from all three methods are shown in Table 7.
Discussion
Faldaprevir is a low-clearance drug that is metabolically very stable in vitro (Duan et al., 2012). However, the human ADME study revealed two abundant monohydroxylated metabolites, M2a and M2b, excreted in feces, accounting for ∼40% of the dose, but not found in human plasma (Chen et al., 2014). Individual in vitro enzyme and transporter systems were used to elucidate the fate of M2a and M2b (Li et al., 2014). An integrated system, HepatoPac, with both active transport and DMEs, was used to provide an overall picture of faldaprevir disposition in humans.
The metabolic profile generated in vitro with [14C]faldaprevir and human HepatoPac was compared with results from the human [14C] ADME study (Tables 1 and 2). Five of seven in vivo metabolites were identified in vitro. Quantitative radiometric analysis demonstrated that M2a and M2b were the predominant metabolites in vitro, providing a combined level of 66% of total metabolites, compared with 88% of total fecal metabolites. Faldaprevir glucuronide represented ∼9% of total metabolites in vitro. However, the in vivo ratio for M2a and M2b (88%) may be overestimated as the glucuronide metabolite could be hydrolyzed back to the parent by gut bacteria in vivo. Differences in the relative levels of M2a and M2b were interesting. M2a and M2b were excreted in feces at equivalent levels. In vitro, M2b was formed at ∼1.6-fold-higher levels than M2a. These minor discrepancies may reflect CYP3A5 polymorphism of the HepatoPac donor (*3*3) used in this study compared with the eight healthy volunteers in the ADME study. This is supported by the results where M2b formation was ∼1.5-fold higher than M2a formation in subjects with CYP3A5*3*3 (77% of population) based on studies using rCYP3A4 and rCYP3A5 (Li et al., 2014). No further biotransformation of these oxidative metabolites, e.g., glucuronidation, was detected in vitro. Induction by rifampicin pretreatment was included to evaluate the impact of CYP3A4 induction on faldaprevir metabolism. These studies indicated that rifampicin (a known potent CYP3A4 inducer) moderately increased formation of M2a, M2b, and faldaprevir glucuronide, consistent with the clinical observation of increased faldaprevir clearance by another inducer, efavirenz (Kiser et al., 2013).
In vitro formation rates of M2a and M2b, when scaled to whole liver, were comparable to in vivo formation rates (Table 4) assuming that formation and excretion of M2a and M2b into feces were linear over the collection period (168 hours). This assumption is reasonable considering the low clearance (Duan et al., 2012) and long t1/2 (20–30 hours) (Manns et al., 2011) of faldaprevir. Interestingly, faldaprevir inactivates CYP3A4 (data on file, Boehringer Ingelheim Pharmaceuticals, Inc.). Because CYP3A4 is predominantly responsible for formation of M2a and M2b (Li et al., 2014), it was expected that the rate of formation would decrease over time in hepatocytes due to inactivation of CYP3A4 by faldaprevir. However, the formation rates of M2a and M2b were constant up to 96 hours (Fig. 1), suggesting that the rate of CYP3A4 synthesis was equal to the rate of degradation and inactivation.
Human HepatoPac is a useful model for gauging metabolic profiles in humans, particularly in situations where circulating and excreted metabolites may not accurately reflect the metabolism occurring in liver. The in vitro–to–in vivo correlation observed in metabolic profile between rat HepatoPac and rat in vivo, particularly glucuronidation (Ramsden et al., 2014), supports this contention. Using human HepatoPac, M2a and M2b were the major metabolites (88%–89%) with glucuronidation representing 11%–12% of the total metabolism (sum of the three metabolites) calculated based on the ratio of CLint (glucuronide versus M2a + M2b) at unbound therapeutic concentrations of faldaprevir (Table 4), indicating that glucuronidation is a minor metabolic pathway in humans. Interaction with UGTs is an emerging regulatory concern (Food and Drug Administration, 2012). These data suggest that the DDI potential of faldaprevir with UGT inhibitors is low.
As liver is the primary site of HCV infection, accumulation of drug in liver can be considered advantageous. In rats, faldaprevir demonstrated a 42-fold enrichment in liver compared with plasma (Duan et al., 2012), which was accurately predicted by rat HepatoPac (Ramsden et al., 2014). Therefore, in this study, human HepatoPac was used to predict in vivo liver enrichment in humans. In human HepatoPac there was a time dependence to establish equilibrium between medium and intracellular concentrations, similar to rat HepatoPac, possibly due to the interplay of uptake and efflux and also a saturation of liver enrichment at higher incubation concentrations (>0.12 μM as free concentration) (Fig. 3). Liver partitioning was estimated to be 31.5- and 21.9-fold for 120-mg QD and 240-mg QD doses, respectively. As an alternative approach to estimate liver enrichment, the free Km values for the formation of M2a and M2b, generated from human HepatoPac, were compared with the free Km values generated with human liver microsomes (HLM) (Li et al., 2014). Differences in these values may represent the enrichment value, since free Km parameters should reflect concentrations of drug at the site of metabolism. The ratios of Km values for hepatocyte/HLM were 19.7 for M2a and 23.4 for M2b (Table 4), comparable to the enrichment values above (21.9 and 31.5).
Liver enrichment can result from active uptake (Chu et al., 2013), saturation of efflux (Swift et al., 2010), higher tissue binding (Nagar and Korzekwa, 2012), and/or sequestration into subcellular compartments such as liposomes or lysosomes (Nadanaciva et al., 2011). Expression of efflux transporters has been demonstrated in HepatoPac (Khetani and Bhatia, 2008). Using the probe substrates rosuvastatin and TCA, both OATPs and NTCP, respectively, were shown to be functional (Fig. 4A; Fig. 6). Interestingly, rifamycin SV, generally used as a selective OATP inhibitor, also inhibited sodium-dependent uptake of TCA to hepatocytes by ∼76%, indicating that rifamycin SV can also inhibit uptake mediated by NTCP (Fig. 6B), which is consistent with a finding that rifamycin SV inhibited NTCP in polarized LLC-PK1 cells (Mita et al., 2006). In addition, rosuvastatin uptake decreased by 32% in the absence of sodium (Fig. 7), in line with observations that rosuvastatin is also a substrate of NTCP (Ho et al., 2006). Therefore, it is likely that complete inhibition of rosuvastatin uptake by rifamycin SV combines inhibition of both OATPs and NTCP.
In human HepatoPac, the active uptake of faldaprevir was partially inhibited by rifamycin SV (Fig. 4), consistent with faldaprevir being a substrate of OATP1B1, a finding confirmed using HEK293 cells expressing OATP1B1 (Fig. 5). The partial inhibition of the sodium-dependent uptake of faldaprevir by rifamycin SV at 3 μM faldaprevir and complete inhibition at 0.3 μM faldaprevir (Fig. 8) suggest that one or more sodium-dependent transporters may be involved. Uptake kinetic parameters for faldaprevir were generated (Fig. 9; Table 6). Three processes contributed approximately equally to the total uptake of faldaprevir, namely, active uptake inhibited by rifamycin SV, active uptake not inhibited by rifamycin SV, and passive permeability (Table 6).
Using these data, the hepatic clearance of faldaprevir was determined by applying the well stirred liver model, a modified well stirred liver model incorporating uptake kinetics (Webborn et al., 2007), and a modified well stirred liver model considering different binding in liver and plasma (Poulin et al., 2012). The Poulin method includes ionization properties of faldaprevir and the albumin content differences between plasma and liver, which is important because clearance of highly protein-bound drugs, especially those with high albumin binding, tend to be underpredicted using the standard well stirred model (Ring et al., 2011). This model is limited as it does not include active uptake. The uptake of faldaprevir into human HepatoPac involves multiple transporters. Therefore, uptake-mediated CLint and passive CLint were incorporated into the estimation of hepatic clearance using the Webborn method. Hepatic clearance was underestimated using the well stirred liver model when compared with oral clearance (CL/F) determined in the clinic (Table 7). The absolute bioavailability (F) of faldaprevir is unknown. However, because ∼50% of the dose undergoes metabolism in humans, based on ADME data (Chen et al., 2014), and first-pass metabolism is estimated to be minimal (Li et al., 2014), it is expected that F > 0.5. Both the well stirred liver model incorporating uptake and the Poulin method were more predictive of the observed CL/F values, suggesting that active uptake or different binding in liver and plasma, coupled with metabolic clearance, plays a predominant role in the overall clearance of faldaprevir. The activities of both M2a and M2b metabolites were evaluated in a separate study. Both metabolites are pharmacologically active and 2- to 4-fold more potent than the parent (Li et al., 2014). However, considering their expected low exposure in liver due to slow formation and efficient efflux into bile, the contribution of M2a and M2b to clinical efficacy is expected to be limited (Li et al., 2014).
The use of human HepatoPac provided valuable information in elucidating the metabolism and disposition of faldaprevir in humans, confirming the importance of the interplay between DMEs and transporters in the clearance of the drug. These studies suggest that faldaprevir is taken up into liver by passive permeability and active uptake by multiple transporters (including OATP1B1 and Na+-dependent transporters), metabolized by CYP3A, and excreted through passive permeability and/or biliary efflux transporters, such as P-glycoprotein (Li et al., 2014), explaining why faldaprevir is almost completely excreted via bile.
Acknowledgments
The authors thank Patrick Baum for performing genotyping studies, Arti Mathur for characterizing UGT activity, Kirsten Mease for performing the HEK293 transporter studies, and Monica Keith-Luzzi for conducting protein binding studies in HepatoPac medium. We thank Dr. Timothy S. Tracy for scientific advice.
Authorship Contributions
Participated in research design: Ramsden, Tweedie, Taub, Chan, Li.
Conducted experiments: Ramsden.
Performed data analysis: Ramsden, Li.
Wrote or contributed to the writing of the manuscript: Ramsden, Tweedie, Li.
Footnotes
- Received November 8, 2013.
- Accepted December 23, 2013.
This research was funded by Boehringer Ingelheim Pharmaceuticals, Inc.
Abbreviations
- ADME
- absorption, distribution, metabolism, and excretion
- BSA
- bovine serum albumin
- CLh,B
- hepatic blood clearance
- CLint
- intrinsic clearance
- CL/F
- oral clearance
- DDI
- drug-drug interaction
- DME
- drug-metabolizing enzyme
- FDV
- faldaprevir
- fu,in vitro
- free fraction of drug in hHIM
- fu,liver
- free fraction of drug in liver
- fu,p
- free fraction of drug in plasma
- HCV
- hepatitis C virus
- hHIM
- Human HepatoPac incubation medium
- HLM
- human liver microsomes
- Kd
- equilibrium binding constant
- Km
- Michaelis-Menten constant
- LC-MS/MS
- liquid chromatography–tandem mass spectrometry
- NTCP
- sodium taurocholate cotransporting peptide
- Qh
- hepatic blood flow in humans
- QD
- daily
- t1/2
- half-life
- TCA
- taurocholic acid
- UGT
- UDP glucuronosyltransferase
- Copyright © 2014 by The American Society for Pharmacology and Experimental Therapeutics