Treatment of pulmonary arterial hypertension with the endothelin receptor antagonist bosentan has been associated with transient increases in liver transaminases. Mechanistically, bosentan inhibits the bile salt export pump (BSEP) leading to an intrahepatic accumulation of cytotoxic bile salts, which eventually results in hepatocellular damage. BSEP inhibition by bosentan is amplified by its accumulation in the liver as bosentan is a substrate of organic anion-transporting polypeptide (OATP) transport proteins. The novel endothelin receptor antagonist macitentan shows a superior liver safety profile. Introduction of the less acidic sulfamide moiety and increased lipophilicity yield a hepatic disposition profile different from other endothelin receptor antagonists. Passive diffusion rather than OATP-mediated uptake is the driving force for macitentan uptake into the liver. Interaction with the sodium taurocholate cotransporting polypeptide and BSEP transport proteins involved in hepatic bile salt homeostasis is therefore limited due to the low intrahepatic drug concentrations. Evidence for this conclusion is provided by in vitro experiments in drug transporter-expressing cell lines, acute and long-term studies in rats and dogs, absence of plasma bile salt changes in healthy human volunteers after multiple dosing, and finally the liver safety profile of macitentan in the completed phase III morbidity/mortality SERAPHIN (Study with an Endothelin Receptor Antagonist in Pulmonary Arterial Hypertension to Improve Clinical Outcome) trial.
About a decade ago, endothelin receptor antagonists were introduced as a therapeutic concept for the treatment of pulmonary arterial hypertension, a debilitating and finally fatal disease for which no oral treatment option was available before. The dual endothelin receptor antagonist bosentan (Tracleer; Actelion Pharmaceuticals, Allschwil, Switzerland) was approved in 2001 as the first member of this new class, followed by sitaxentan (Thelin; Encysive Pharmaceuticals, Houston, TX) in 2006 and ambrisentan (Letairis/Volibris; Gilead Sciences, Foster City, CA) in 2007. Macitentan (Opsumit, N-[5-(4-bromophenyl)-6-[2-[(5-bromo-2-pyrimidinyl)oxy]ethoxy]-4-pyrimidinyl-N′-propylsulfamide; Actelion Pharmaceuticals) has been developed as a new generation endothelin receptor antagonist with sustained receptor binding and improved receptor potency, pharmacokinetic properties, and liver safety profile (Iglarz et al., 2008; Raja, 2010). Most of these improvements result from a modified tissue distribution as macitentan can freely diffuse into tissues rather than being dependent on active transport.
Bosentan was approved at doses of 62.5 and 125 mg twice a day, but was initially studied at higher doses for the treatment of hypertension (Krum et al., 1998) and chronic heart failure (Sütsch et al., 1998). Chronic heart failure patients treated with 500 mg of bosentan twice a day had an 18% incidence of elevated alanine aminotransferase levels versus 4% on placebo. In a subset of patients concomitantly treated with the antidiabetic glyburide, 29% had elevated alanine aminotransferase versus 4% and 0% on either placebo alone or placebo and glyburide, respectively. Changes in liver transaminases were accompanied by dose-dependent increases in plasma bile salts and alkaline phosphatase.
Inhibition of the bile salt export pump (BSEP) by bosentan and its metabolites has been identified as the likely mechanism underlying the observed changes in plasma transaminases (Fattinger et al., 2001). BSEP is an ATP-dependent transport protein located at the hepatocanalicular membrane and mediates the rate-limiting step in bile salt secretion from blood into bile (Gerloff et al., 1998; Stieger et al., 2000). Bosentan and its metabolites inhibited taurocholate transport in vitro in canalicular rat liver membrane vesicles and in Spodoptera frugiperda (Sf9) cell vesicles overexpressing rat bsep. In rats, plasma bile salts increased in a dose-dependent manner after intravenous dosing of bosentan (Stieger et al., 2000; Kis et al., 2009).
These initial findings in rats were later confirmed with human BSEP (Mano et al., 2007) and led to the hypothesis that bosentan treatment initially triggers a disruption of bile salt homeostasis through dose-dependent blockade of BSEP-mediated bile salt excretion into bile, eventually resulting in their accumulation in liver cells. As bile salts are cytotoxic at high concentrations, the observed liver transaminase elevations in man are believed to result from the secondary bile salt toxicity in hepatocytes.
The hepatic disposition of bosentan is mediated by organic anion-transporting polypeptide (OATP) transport (Treiber et al., 2007) followed by extensive metabolism through CYP3A4 and CYP2C9 and finally excretion of the metabolites into bile (Weber et al., 1999). As a consequence, bosentan pharmacokinetics are sensitive to concomitant CYP3A4 and/or OATP inhibitors. Although the potent CYP3A4 inhibitor ketoconazole increased bosentan in plasma by only about 2-fold (van Giersbergen et al., 2002), more pronounced elevations were observed with the OATP inhibitor rifampicin (van Giersbergen et al., 2007), the human immunodeficiency virus protease inhibitor ritonavir/lopinavir (Kaletra; Abbott Laboratories, Abbott Park, IL) (Dingemanse et al., 2010), and cyclosporin A (Binet et al., 2000), the latter two being combined CYP3A4/OATP inhibitors.
Conceptually, there are several options for designing drugs with an improved side-effect profile. On the one hand, improving receptor affinity and pharmacokinetic properties might yield drugs that are effective at lower doses. The alternative approach is to avoid interactions with targets critically involved in toxicity. Both approaches were combined in the discovery of macitentan. The present report summarizes the experimental evidence demonstrating that macitentan does not interact with hepatic transport proteins critically involved in bile salt trafficking and drug accumulation in the liver.
Materials and Methods
Chemicals and Reagents
Macitentan was obtained from Lonza AG (Visp, Switzerland) with a purity of 99.8%. [14C]Radiolabeled macitentan with a specific activity of 55 mCi/mmol was purchased from GE Healthcare (Little Chalfont, UK). Metabolite ACT-132577 ([5-(4-bromophenyl)-6-(2-(5-bromopyrimidin-2-yloxy)ethoxy)-pyrimidin-4-yl]-sulfamide) was obtained either from the chemistry department of Actelion Pharmaceuticals Ltd, or from SynphaBase (Pratteln, Switzerland), with purity in excess of 97%. [14C]ACT-132577 with a specific activity of 56 mCi/mmol was obtained from Quotient Bioresearch (Rushden, Northamptonshire, UK). Both radiolabeled compounds were supplied as acetonitrile solutions with radiochemical purities in excess of 97%.
Bosentan was obtained from the chemistry department of Actelion Pharmaceuticals Ltd. Sodium taurocholate was obtained from Sigma-Aldrich (Buchs, Switzerland), and [3H]taurocholic acid with a specific activity of 4.6–5.0 Ci/mmol was purchased from PerkinElmer (Boston, MA) as a solution in methanol:ethanol (1:3) at a concentration of 1 mCi/ml. Estrone-3-sulfate and atorvastatin calcium trihydrate were from Sigma-Aldrich. [3H]Estrone-3-sulfate and [3H]atorvastatin calcium with specific activities of 50 Ci/mmol and 10 Ci/mmol, respectively, were purchased from American Radiolabeled Chemicals (St. Louis, MO) as solutions in ethanol or ethanol:water (1:1). Cyclosporin A was purchased from Fluka (Buchs, Switzerland) and rifampicin from Sigma-Aldrich. The liquid scintillation cocktails Filter-Count and IRGA Safe Plus were purchased from PerkinElmer (Zürich, Switzerland). Baculovirus-infected Sf9 cell membrane vesicles overexpressing human BSEP were obtained from SOLVO Biotechnology (Budapest, Hungary). All media and supplements for Chinese hamster ovary (CHO) and CHO Flp InTM cells were obtained from Invitrogen AG (Basel, Switzerland).
Preparation of Stock Solutions.
For BSEP and sodium taurocholate cotransporting polypeptide (NTCP) inhibition experiments, macitentan and ACT-132577 stock solutions were initially prepared in dimethyl sulfoxide (DMSO) in concentration ranges of 1 μM to 100 mM BSEP and 1 μM to 100 mM NTCP, and then diluted with the buffer used in the transport experiments described herein. For cellular transport experiments, macitentan and ACT-132577 DMSO stock solutions were prepared in a range from 0.1 µM to 100 mM and again diluted with transport buffer. Stock solutions of cyclosporin A and rifampicin were prepared in DMSO in a concentration range from 1–50 mM. DMSO was also used to prepare the 10 and 100 mM stock solutions of atorvastatin, taurocholic acid, and estrone-3-sulfate.
CHO Flp In cells overexpressing human NTCP were cultured at passage numbers 5 to 19 on tissue culture dishes of 55-cm2 growth area (Sarstedt, Newton, NC) at 37°C in a humidified atmosphere containing 5% carbon dioxide. Cells were maintained in Ham’s F-12 medium supplemented with 10% fetal calf serum, penicillin/streptomycin (100 IU/ml), l-glutamine (1 mM), and hygromycin B (500 μg/ml). For transport experiments, cells from a maximally 90% confluent 58-cm2 tissue culture dish were detached with trypsin- EDTA, uniformly resuspended in Ham’s F-12 medium, and seeded on tissue culture dishes (Corning, Tewksbury, MA). The cells were used for transport experiments 72 to 96 hours later, when they were 80–90% confluent. At 24 hours before starting the transport experiments, cells were additionally induced by adding 5 mM sodium butyrate (Sigma-Aldrich) to the medium.
CHO cells overexpressing human OATP1B1, OATP1B3, and OATP2B1 and wild-type CHO cells were cultured at passage numbers 9 to 60 on tissue culture dishes (Corning) at 37°C in a humidified atmosphere containing 5% carbon dioxide. All cell lines were maintained in Dulbecco’s modified Eagle’s medium containing 1 g/l glucose and supplemented with 10% fetal calf serum, penicillin/streptomycin (100 IU/ml) and l-proline (0.05 mg/ml). The culture medium for the OATP-expressing CHO cells additionally contained geneticin (500 μg/ml). For transport experiments, cells from a confluent 55-cm2 tissue culture dish were detached with trypsin-EDTA, uniformly resuspended in Dulbecco’s modified Eagle’s medium, and seeded on tissue culture dishes of 8-cm2 growth area. Cells were used for transport experiments 72 to 96 hours later, when they were 90–100% confluent. At 24 hours before starting the transport experiment, the cells were additionally induced by adding 5 mM sodium butyrate to the medium.
Cryopreserved human hepatocytes (lot SSR; Bioreclamation IVT, Brussels, Belgium) were seeded on collagen-coated 24-well plates (Nunc, Thermo Scientific, Wohlen, Switzerland) at a density of 0.2 × 106 viable cells per well. Cells were allowed to attach for about 4 hours in William’s medium E (Life Technologies Europe B.V., Zug, Switzerland) supplemented with 10% fetal calf serum, 10 mg/ml insulin, and 10 mg/ml penicillin/streptomycin before use in uptake experiments.
Transport Experiments with Overexpressing CHO Cells.
Transport experiments with CHO Flp In cells expressing NTCP were run using three 8-cm2 tissue culture dishes for each concentration investigated. After washing the cells three times with 2 ml of prewarmed (37°C) sodium or choline buffer, we initiated the uptake experiment by adding 1 ml of buffer containing either [14C]-labeled macitentan at various concentrations or 5 μM [3H]-labeled taurocholic acid (appropriately diluted with nonlabeled material). The sodium-containing buffer was composed of 20 mM HEPES (pH 7.4), 116.4 mM NaCl, 1 mM NaH2PO4, 5.3 mM KCl, 0.8 mM MgSO4, and 5.5 mM d-glucose. The choline-containing buffer had the overall same composition but sodium chloride and NaH2PO4 were replaced with 116.4 mM choline chloride and 1 mM KH2PO4, respectively. After incubation at 37°C for 40 seconds, cellular uptake was stopped by addition of 2 times 2 ml of ice-cold choline buffer containing 0.5% bovine serum albumin (Sigma-Aldrich). Bovine serum albumin was included in the washing buffer to minimize unspecific binding. Cells were washed four times with approximately 2 ml of ice-cold choline buffer and then solubilized by addition of 1 ml of 1% (w/v) Triton X-100. After incubation for at least 20 minutes, 0.5 ml of the cell lysate was mixed with 3.5 ml of scintillation cocktail IRGA Safe Plus, and the total radioactivity was determined using a Tri-Carb 2300 TR liquid scintillation analyzer (Packard Bioscience, Zürich, Switzerland). Twenty-five microliter–aliquots of the cell lysates were used to determine the protein content of each sample. Inhibition experiments were performed by simultaneous addition of 5 μM [3H]taurocholic acid and predefined concentrations of macitentan or ACT-132577. Incubations and sample work-up were done as outlined above. NTCP-mediated transport rates were calculated as the difference between sodium and choline buffer.
Transport experiments with OATP-expressing and wild-type CHO cells were run using three 8-cm2 tissue culture dishes for each concentration. After washing the cells three times with 2 ml of prewarmed (37°C) transport buffer, the uptake experiment was initiated by adding 1 ml of buffer containing macitentan or ACT-132577 at concentrations of 0.01–100 μM and 0.01–300 μM, respectively. Cellular uptake was determined at 37°C and stopped after 40 seconds by addition of two times 2 ml of ice-cold transport buffer containing 0.5% bovine serum albumin. The latter was included in the washing buffer to minimize nonspecific binding of radioactive compounds. Cells were then rapidly washed four times with each 2 ml of ice-cold transport buffer and solubilized by addition of 1 ml of 1% (w/v) Triton X-100. After incubation for at least 20 minutes, 0.5 ml of the cell lysate was mixed with 5 ml of scintillation cocktail IRGA Safe Plus and total radioactivity determined using a Tri-Carb 2300 TR liquid scintillation analyzer. Twenty-five microliter–aliquots of the cell lysates were used to determine total protein content. Before each transport experiment, the time dependence of cellular uptake was individually determined to optimize experimental conditions.
The effect of the OATP inhibitors cyclosporin A and rifampicin on the uptake of 1 μM macitentan was investigated for all three OATP transporters. The inhibition experiment was started by addition of 1 ml of prewarmed transport buffer containing radiolabeled macitentan and the inhibitor in a concentration range from 0.05–100 μM. After incubation at 37°C for 20 seconds, cellular uptake was terminated by addition of 2 times 2 ml of ice-cold transport buffer containing 0.5% bovine serum albumin. The sample work-up in these inhibition experiments was performed as previously outlined. The final content of organic solvent in the transport experiments never exceeded 1%. [3H]Estrone-3-sulfate was used as a positive control.
Transport Experiments with Membrane Vesicles.
For transport experiments, membrane vesicles expressing human BSEP (50 μg total protein) were incubated in the presence and absence of 5 mM ATP. Incubations were performed at 37°C for 1 minute or 3 minutes in transport buffer containing 10 mM HEPES (pH 7.4), 50 mM sucrose, 100 mM KNO3, 10 mM Mg(NO3)2, and 5 μM [3H]taurocholic acid. Taurocholate uptake was stopped by the addition of ice-cold washing buffer containing 10 mM Tris-HCl (pH 7.4), 50 mM sucrose, and 100 mM KCl, followed by collection of membrane vesicles on a cellulose nitrate membrane filter (pore size 0.45 µm) using a rapid filtration system (Millipore, Zug, Switzerland). Before the experiments, filters were saturated with 1 mM nonlabeled taurocholic acid to minimize nonspecific binding of radiolabeled compound. Retained membranes were then washed twice with ice-cold buffer and transferred into scintillation vials. After addition of 3.5 ml of scintillation cocktail, the total radioactivity was determined on a Tri-Carb 2300 TR liquid scintillation analyzer. Inhibition experiments were performed by incubating membrane vesicles simultaneously with 5 μM [3H]taurocholic acid and various concentrations of macitentan ACT-132577. BSEP-mediated transport rates were calculated as the difference of results obtained in the presence or absence of ATP.
Macitentan Partitioning in Human Hepatocytes.
Cellular uptake of macitentan was determined in triplicate with plated, cryopreserved human hepatocytes. After we had removed the William’s medium E, the cells were washed twice with 0.5 ml of prewarmed Hanks’ balanced salt solution (HBSS). Medium and washing solutions were pooled, and the number of unattached cells was counted using a Vi-CELL counter (Beckman Coulter, Nyon, Switzerland) to estimate the number of plated hepatocytes in the well. The hepatic uptake experiment was started by the addition of 200 µl of prewarmed (37°C) incubation solution containing macitentan in HBSS with 1% DMSO at a final concentration of approximately 100 nM. After 10 minutes’ incubation at 37°C on an orbital shaker at 300 rpm, uptake was terminated by removal of the supernatant followed by washing of the cells with 2 × 0.5 ml of ice-cold phosphate-buffered saline (pH 7.4). Supernatants were fortified with one volume equivalent of acetonitrile containing tetradeuterated macitentan as analytic standard. Hepatocytes were lysed by the addition of 200 µl of a 2:3 mixture of HBSS and acetonitrile containing tetradeuterated macitentan, and incubation at room temperature for 15 minutes. Calibration samples were prepared and worked up in parallel in a concentration range from 2 to 1000 nM by diluting the macitentan stock solution in DMSO with a 1:1 (v/v) mixture of either acetonitrile and HBSS, or hepatocyte lysate. All samples were placed in 96-well plates pending analysis by liquid chromatography–tandem mass spectrometry (LC-MS/MS).
Macitentan Binding in Human Hepatocytes.
Macitentan binding to human hepatocyte homogenate was determined by the use of rapid equilibrium dialysis and a membrane with a molecular mass cutoff of 8 kDa (Thermo Fisher Scientific, Reinach, Switzerland). Before equilibrium dialysis, human hepatocytes (1 × 106 cells/ml) were metabolically inactivated by an initial incubation at 37°C and 800 rpm on a thermomixer for 48 hours, followed by three freeze-thaw cycles at room temperature and −20°C, and finally sonication for 10 seconds (Vibracell 75043; Bioblock Scientific, Illkirch, France). Macitentan at a final concentration of 0.5 μM was added to the hepatocyte homogenate as a 0.5-mM stock solution in DMSO. Two hundred microliter–aliquots of this mixture were transferred into the donor compartment of the rapid equilibrium device and dialyzed against 350 µl HBSS at 37°C for 4 hours on an orbital shaker in an atmosphere containing 5% CO2. At the end of dialysis, 50-μl aliquots of the donor compartment were diluted with HBSS, while 50 µl of the receiver compartment were diluted with 50 µl of blank hepatocyte homogenate to generate samples with the same analytical matrix. Three independent experiments were performed with three replicates each.
Sample work-up for LC-MS/MS analysis consisted of protein precipitation with three volume equivalents of methanol containing tetradeuterated macitentan as analytic standard. After centrifugation at 3220g and 4°C for 20 minutes, 5-µl aliquots were transferred in a 96-well plate pending analysis. Calibration samples were prepared and worked up in parallel in a concentration range from 0.5–1000 nM by diluting the macitentan stock solution in DMSO with a 1:1 (v/v) mixture of hepatocyte homogenate and HBSS.
Quantification of Macitentan by LC-MS/MS.
The analytic equipment consisted of a Shimadzu HPLC System (Shimadzu, Reinach, Switzerland) connected to an API5000 (AB SCIEX, Concord, ON, Canada). Data acquisition was performed with the Analyst software package (version 1.5.1; AB SCIEX). The chromatographic analysis was achieved on a Phenomenex Luna C8 column (5 µm, 2.0 × 20 mm i.d.) at room temperature with a flow rate of 0.6 ml/min (Phenomenex, Torrance, CA). Mobile phases consisted of 0.1% aqueous formic acid and acetonitrile. The mass transitions used for macitentan and its tetradeuterated internal standard were 589 to 201 and 593 to 205, respectively, both with a scan time of 50 milliseconds.
Determination of Total Protein Content.
Total protein content was determined using the Pierce bicinchoninic acid assay (Pierce Science, Lausanne, Switzerland) with quantification at a wavelength of 590 nm on a SpectraCount spectrophotometer (Packard Bioscience) according to the supplier’s protocol. Bovine serum albumin was used as a standard. Raw data were analyzed using the PlateReader software I-Smart (version 3.0 for Windows; Packard Bioscience).
Data from the inhibition experiments were evaluated by plotting the inhibitor concentration (logarithmic scale) against the BSEP- or NTCP-mediated transport of taurocholic acid. IC50 values were then determined from the plot by nonlinear regression using Eq. 1 with a constraint Bottom > 0 (Giacomini et al., 2010):(1)in which y is the transport expressed as the percentage of inhibition relative to control, x is the inhibitor concentration (μM), s is the slope at the point of inversion, and Top and Bottom are the maximum and minimum transport rates. For all graphical data evaluations, the GraphPad Prism software package (version 5.0; GraphPad Software Inc., La Jolla, CA) was used. The fitted parameters are presented as best-fit parameter and standard error.
Cellular uptakes were normalized for total protein content and are either expressed as pmol/mg protein or are further normalized for incubation time and expressed as pmol/(mg∙min). The OATP-mediated net uptake rates are calculated as the difference of OATP-expressing and wild-type CHO cells for each individual concentration and are presented as mean and S.D. Uptake ratios were calculated from the OATP-expressing and wild-type cells.
The partitioning ratio of macitentan between human hepatocytes and in the incubation medium (Kp) was calculated using Eq. 2:(2)where Ahepatocyte is the amount of macitentan in hepatocyte lysate, Vhepatocyte is the hepatocyte volume, and cincubation solution is the macitentan concentration in the incubation medium at the end of the experiment. The hepatocyte volume was estimated from cell diameters measured before plating (23 µm). This value is in good agreement with previously published data, which are 16.2 µm for human hepatocytes (Mateus et al., 2013) and 24 µm for rat hepatocytes (Treijtel et al., 2005).
The unbound fraction in the hepatocyte homogenate (fu,homogenate) was calculated using Eq. 3:(3)where chomogenate and cbuffer are the macitentan concentrations in the donor and receiver compartments at the end of dialysis. The unbound fraction in human hepatocytes (fu, hepatocyte) was derived from Eq. 4:(4)where D is the homogenate dilution factor. The volume of human hepatocytes was again determined from cell diameters (20 µm) before homogenization. The ratio of unbound macitentan concentrations between cells and medium (Kp,uu) as a measure for drug accumulation was calculated using Eq. 5:(5)where fu, incubation solution is the unbound macitentan concentration of the incubation medium. In this equation, fu, incubation solution is assumed to be 1 as the medium does not contain proteins.
Bile Salt Measurements in Animals and Man
Quantification of Bile Salts in Plasma and Serum.
Bile salt concentrations in plasma and serum were determined using an enzymatic assay based on the reduction of NAD to NADH, which is subsequently used to reduce nitrotetrazolium blue to formazan, followed by colorimetric quantification of the latter at a wavelength of 530 nm. For samples from the intravenous rat model, bile salts in plasma were quantified using a kit from Sigma Diagnostics (St. Louis, MO) and a set of calibration samples ranging from 0–100 μM that was run on the same 96-well plate as the unknown samples. The commercial kit is designed to quantify bile salts in serum but can also be used for plasma (validation data not shown). Bile salts in sera from the externally performed rat and dog toxicity studies and the multiple ascending–dose study with macitentan in human healthy subjects (Sidharta et al., 2013) were analyzed in the respective preclinical and clinical research laboratories.
For the acute cholestasis model, male Wistar rats, 8–12 weeks of age, were delivered from RCC Ltd., Biotechnology and Animal Breeding Division, Füllinsdorf, Switzerland, and used after an acclimatization period of at least 7 days. Body weights were between 221–345 g at the day of the experiment. All animals were housed under climate-controlled conditions with a 12-hour light/dark cycle in accordance with the guidelines of the Basel Cantonal Veterinary Office (license no. 169). All animals were maintained under identical conditions and had free access to drinking water and food (batch 3418; Provimi Kliba, Kaiseraugst, Switzerland).
The multiple-dose toxicity studies in Wistar or Sprague-Dawley rats and Beagle dogs were conducted in certified contract research organizations in compliance with principles of Good Laboratory Practice. All animal experiments adhered to the Principles of Laboratory Animal Care (National Institutes of Health, 8th edition; http://grants.nih.gov/grants/olaw/Guide-for-the-care-and-use-of-laboratory-animals.pdf).
Bosentan, Macitentan, and ACT-132577 in the Acute Rat Model.
Macitentan, its metabolite ACT-132577, and bosentan were formulated as microsuspensions in 7.5% gelatin and intravenously administered via the tail vein at a dose of 25 mg/kg (n = 6) and a dosing volume of 1 ml/kg. All formulations were prepared freshly on the day of experiment and stirred well before administration. About 0.5 ml of blood were collected into EDTA-containing vials from the sublingual vein before dosing and at 10, 45, and 120 minutes after dosing. The effect of the gelatin vehicle was investigated in a control experiment. Plasma was prepared by centrifugation at approximately 4000g and stored frozen at −20°C pending analysis.
Bile Salt Measurements in the Rat and Dog Toxicity Studies.
In the multiple-dose oral toxicity studies, male and female animals were treated once daily with macitentan either by gavage (rats) or with capsules (dogs). A control group receiving the methylcellulose vehicle or empty capsules was an integral part of all study designs. Doses were selected based on previous dose-range finding studies and preceding studies of shorter duration. Treatment duration was 4, 13, and 26 weeks in the rat, and 4, 13, and 39 weeks in the dog. Bile salts in plasma were determined at the end of the toxicity study as part of the regular clinical chemistry program.
Bile Salt Measurements in the Multiple Ascending Dose Study in Man.
The multiple ascending dose study with macitentan in healthy human subjects was designed as a double-blind, placebo-controlled, randomized study to investigate the tolerability, safety, pharmacokinetics, and pharmacodynamics of macitentan (Sidharta et al., 2013). The study followed the principles of the Declaration of Helsinki and Good Clinical Practice. The protocol and informed consent form was approved by an independent ethics committee (ethics committee of the Landesärztekammer Baden-Württemberg, Stuttgart, Germany). A total of 32 men received doses of 1, 3, 10, and 30 mg of macitentan or placebo for 10 days once daily in fasted state. Each dose was administered sequentially to a group of eight subjects (six on macitentan, two on placebo). The safety evaluation comprised the collection of adverse event data, including assessments of seriousness, severity, relationship to study drug, and outcome. The safety assessment comprised laboratory variables, vital signs, 12-lead electrocardiogram, physical condition, and body weight. Pharmacokinetic parameters of macitentan and its metabolite ACT-132577 were assessed for all doses. Bile salts in serum were determined before the dose on day 1 and day 10 of treatment.
Physiologically-Based Pharmacokinetic Modeling
The study used Simcyp Population-Based ADME Simulator (version 12; Simcyp Limited, Sheffield, UK), a physiologically-based pharmacokinetic (PBPK) computer model combined with genetic, physiologic, and demographic variables using Monte Carlo methods and equations derived from population databases obtained from literature sources.
The physicochemical properties and blood binding of macitentan—molecular weight 588 g/mol, logD 2.9, pKa 6.2, plasma protein binding 99.6%, and blood/plasma ratio 0.55—were entered into Simcyp. The corresponding values for ACT-132577 were molecular weight 546 g/mol, logD 1.5, pKa 6.1, plasma protein binding 99.5%, and blood/plasma ratio 0.55. Published data were used as source for the pharmacokinetic, metabolism, and excretion data used for the development of the PBPK model (Bruderer et al., 2012b; Atsmon et al., 2013). Only 4% of the oral dose was excreted in feces as unchanged macitentan after oral dosing of [14C]-labeled macitentan to healthy volunteers, suggesting almost complete absorption of the dose from the gut. Unchanged macitentan was not detected in urine.
Oral absorption was modeled using a simple first-order model, with the fraction of the dose absorbed from the gut (fa) set to 1 without variation. The rate constant of absorption (ka), absorption lag time, and Qgut, a hybrid term including both villous blood flow and permeability through the enterocyte membrane (Yang et al., 2007; Pang and Chow, 2012), were optimized to fit the observed plasma concentration profile of macitentan. Optimized values were: ka 0.3 h−1, lag time 1 hour, and Qgut 9.5 l/h. The volume of distribution was calculated based on a full PBPK model using the tissue partitioning equations of Rodgers et al. (Rodgers et al., 2005; Rodgers and Rowland, 2006, 2007). The predicted volume of distribution at steady state using tissue volumes for a healthy volunteer population was 0.36 l/kg for macitentan and 0.22 l/kg for ACT-132577. Distribution was assumed to be perfusion-limited for all organs. The liver to plasma partitioning coefficients for macitentan and ACT-132577 were predicted as 0.18 and 0.09, based on physiologic liver volume, intracellular and extracellular water content, neutral and acidic (phospho)lipid content, binding to albumin, and (predicted) binding to lipoproteins.
For the purpose of modeling, it was assumed that macitentan excretion into feces was the result of biliary excretion rather than incomplete absorption. Consequently, 4% of the clearance was set to occur via unchanged excretion into the bile, and the remaining 96% was set to be cleared by hepatic metabolism. Renal clearance was set to zero as no unchanged macitentan was detected in urine. Macitentan blood clearance in man is unknown. However, oral clearance can be calculated using Eq. 6:(6)The mean plasma area under the concentration versus time curve (AUC) of macitentan was 5759 ng·h/ml at a dose of 10 mg (Atsmon et al., 2013). The oral clearance (CLpo) was calculated as 1.8 l/h. The metabolic oral clearance (96% of the total oral clearance) of macitentan was scaled to microliter per minute per milligram protein and microliter per minute per million hepatocytes with the assumption, based on in vitro and clinical interaction data, that 62% of macitentan is metabolized to ACT-132577 (Actelion, data on file) using the scaling factors in Simcyp. The intrinsic clearance (CLint) was derived from CLpo using the well-stirred liver model and Eq. 7:(7)This resulted in a liver microsomal clearance CLint of 0.48 μl/min/pmol CYP3A4 for the formation of ACT-132577. The intrinsic clearance of 1.5 μl/min/million hepatocytes was used to calculate biliary clearance in the PBPK model. The remainder of the metabolic clearance of macitentan was not assigned to a specific enzyme.
ACT-132577 has not been dosed intravenously to humans, so the human clearance is unknown. However, the clearance of a metabolite can be calculated using Eq. 8 (Rowland and Tozer, 1989):(8)Application of Eq. 8 using the observed AUC values for macitentan and ACT-132577 (corrected for the difference in molecular weight) after a single 10-mg macitentan dose (Atsmon et al., 2013) and the 62% fraction of macitentan metabolized to ACT-132577 resulted in a metabolite clearance of 0.32 l/kg.
A coefficient of variation of 30% was assumed for the input parameters single compartment absorption rate constant, lag time, Qgut, and intrinsic clearance. For all other parameters, variation during the simulations was based on (physiologic) variation of the population database within Simcyp. The population selected for the trial design was a healthy volunteer population, male subjects, aged 18–45 years, in fed state. Ten virtual trials of 10 subjects each were run (total size: 100) for a single dose of 10 mg and for steady-state simulations with a 30-mg macitentan loading dose followed by 11 daily doses of 10 mg.
Inhibition of taurocholate uptake by macitentan and metabolite ACT-132577 was investigated using membrane vesicles expressing human BSEP at concentrations up to 100 and 300 μM, respectively (Fig. 1). IC50 values of macitentan and ACT-132577 derived from an analysis of pooled data were 18 ± 5 μM (n = 6) and 60 ± 14 μM (n = 4), respectively. Taurocholate was also the substrate in the inhibition experiments with human NTCP overexpressed in CHO cells (Fig. 2). Macitentan inhibited taurocholate uptake with a mean IC50 value of 18 ± 2 μM (n = 3), whereas ACT-132577 showed a mean IC50 of 14 ± 2 μM (n = 2).
We have previously reported on the cellular uptake of macitentan in OATP1B1- and OATP1B3-overexpressing cells in the context of the clinical drug-drug interaction studies with cyclosporin A and rifampicin (Bruderer et al., 2012a). Table 1 summarizes these results of the uptake experiments with macitentan and ACT-132577 together with new data on OATP2B1. Estrone-3-sulfate was used as a positive control. Uptake ratios between OATP-overexpressing and wild-type cells were calculated. Based on these data, neither macitentan nor ACT-132577 are considered substrates for OATP1B1 or OATP2B1. Cellular uptake rates of both compounds into CHO wild-type cells exceeded that of the OATP substrate estrone-3-sulfate by at least 150-fold (at 5 μM), indicating that their cellular uptake is mostly driven by high passive diffusion. Figure 3 displays the net uptake rates for macitentan in OATP1B3-overexpressing CHO cells. Net uptake rates consistently differed from those in wild-type CHO cells, and the linear increase up to the highest concentration of 100 μM suggests that macitentan is likely a substrate of OATP1B3. No saturation in macitentan uptake was observed in this concentration range, indicating that the affinity of macitentan for OATP1B3 transport is rather low. Passive permeation was again the major contributor to overall cellular uptake as uptake ratios never exceeded 1.2. The role of OATP in the overall cellular uptake is therefore considered of little clinical relevance. Based on the results in Table 1, metabolite ACT-132577 is not a substrate for OATP1B3.
To verify the above conclusion, macitentan uptake into OATP-overexpressing cells was additionally determined in the presence of the known OATP inhibitors cyclosporin A (Shitara et al., 2003) and rifampicin (Vavricka et al., 2002; Hirano et al., 2006). The results are summarized in Table 2. No consistent effect of cyclosporin A was observed over the concentration range up to 100 μM, that is, at concentrations largely exceeding its Ki value of 0.2 μM (Shitara et al., 2003). Cyclosporin A inhibition was not evident in OATP1B3 cells, nor was there any consistent effect of rifampicin in either cell line. Control experiments with both inhibitors using [3H]atorvastatin as an OATP substrate yielded concentration-dependent decreases in net uptake rates for both compounds (data not shown). Overall, these OATP inhibition experiments support the conclusion that macitentan cellular uptake is mostly dependent on passive diffusion with only a small component of OATP1B3-mediated uptake.
Macitentan uptake into CHO Flp In cells overexpressing the sodium-dependent taurocholate cotransporting polypeptide NTCP was investigated in the same concentration range as used for OATP transporters. Uptake ratios were determined from experiments in the presence or absence of sodium using taurocholic acid at 5 μM as a positive control. The results are summarized in Table 3. Uptake ratios were around unity over the entire concentration range, indicating that macitentan is not a NTCP substrate.
The potential for intracellular accumulation of macitentan has been determined in human hepatocytes using a previously published method with small modifications (Mateus et al., 2013). Two parallel experiments were performed, in which, first, the partitioning ratio (Kp) of macitentan between hepatocytes and culture medium was determined. Mean Kp was 724 ± 96 indicating a significant partitioning of macitentan into human hepatocytes. Macitentan binding in hepatocytes (fu,hepatocytes) was derived from binding in hepatocyte homogenates (fu,homogenate) as determined by rapid equilibrium dialysis. The corresponding values for fu,homogenate and fu,hepatocytes were 0.189 ± 0.009 and 0.00099 ± 0.00004, respectively. These data suggest that macitentan is highly bound in human hepatocytes and that the free fraction therein is only around 0.1%. The ratio of unbound macitentan concentrations in hepatocytes and medium (Kp,uu) as a measure for hepatocellular accumulation was calculated as 0.7 ± 0.1. Macitentan thus does not accumulate in human liver cells, as the unbound concentrations in liver cells and incubation medium are similar. The significant partitioning of macitentan into liver cells is likely the consequence of its elevated lipophilicity and is compensated for by the high binding to cellular components.
The acute effects of macitentan and ACT-132577 on bile salt homeostasis were tested upon intravenous dosing to the rat. This model was developed by Fattinger et al. (2001) to mechanistically rationalize the increased plasma bile salts in clinical trials with bosentan. Macitentan and ACT-132577 were individually tested in this model at an intravenous dose of 25 mg/kg, with plasma samples taken before and at 10, 45, and 120 minutes after dosing. As individual plasma bile salt concentrations varied significantly between animals before drug administration, the results are also expressed as individual differences from predose values. Bosentan was included as a positive control. Results for all three endothelin receptor antagonists are summarized in Table 4. Bosentan increased plasma bile salts by 18 ± 13 μM at 10 minutes after the dose, which then returned to predose values within 45 minutes after dosing. Neither macitentan nor ACT-132577 elicited such an increase in plasma bile salts. After 10 minutes, the mean increases were 3.7 ± 4.4 μM for macitentan and 2.5 ± 6.7 μM for ACT-132577, and, thus, were not different from vehicle (0.4 ± 3.4 μM).
Bile salts in serum were systematically determined in the oral toxicology program of macitentan in the rat and dog as part of the clinical chemistry program. Table 5 summarizes the data collected for all dose groups at the end of the respective study period. In the 4-week rat study, there was no increase in mean bile salts in the male animals up to the highest dose of 1500 mg/kg and in female rats up to 450 mg/kg. Similar to the observations in the intravenous rat model, significant interindividual variability was evident in these rat studies, most likely resulting from differences in food consumption as animals had free access to food over the entire study. No difference between dose groups was noted in the bile salt concentrations in the 26-week toxicity study, in which rats received macitentan doses up to 250 mg/kg.
In the 4-week dog study, macitentan doses up to 250 mg/kg were given. Interanimal variability was significantly lower compared with the rat. There was no difference in mean serum bile salts between dose groups in male or female animals. A similar picture was obtained for doses up to 100 mg/kg in the 13-week study, during which bile salt data were collected after 4 weeks of treatment and at study end. No change in bile salts was observed across dose groups for the entire study duration. Doses in the 39-week dog study were 5, 30, and 100 mg/kg at the start of the study. After 20 weeks, the high dose had to be reduced to 75 mg/kg. Serum bile salt data were collected at the end of the study and confirmed the observations from the studies of shorter duration.
Changes in serum bile salts were also monitored in the multiple-ascending dose study with macitentan in which healthy volunteers received macitentan doses of 1, 3, 10, and 30 mg for a period of 10 days (Sidharta et al., 2013). Each dose group consisted of six individuals on active treatment and two on placebo. The placebo data were pooled from the four active dose groups. Bile salts were collected on the first day before macitentan dosing and on day 10 at the end of study. Results are shown in Table 6. Bile salt concentrations in serum were in a narrow range from 7–19 μM, and there was no discernible trend toward increased serum levels at any dose. Inspection of the individual data revealed a maximum difference of 5 μM between measurements on days 1 and 10, which was observed in a subject receiving placebo. These data confirm the above animal data and show that macitentan treatment in man is not associated with changes in serum bile salts.
A PBPK model of macitentan was developed to allow comparison of the in vitro transport data to concentrations of macitentan and ACT-132577 in the portal vein, systemic circulation, and liver. Table 7 summarizes the observed and predicted human pharmacokinetic parameters of macitentan and ACT-132577 after single and repeat dosing as derived from the PBPK model. The derived plasma concentration versus time profiles for both compounds after a single oral dose are depicted in Fig. 4, A and B. Projected plasma exposure (AUC) and Cmax data were close to the mean observed data. The model predicts a mean Tmax of 7 hours—slightly earlier than the observed mean of 9 hours but still well within the range of individual values. Similarly, the Tmax for ACT-132577 was predicted at 26 hours versus observed 48 hours. Overall, the PBPK model adequately describes the observed plasma concentration profiles of macitentan and ACT-132577. It was therefore considered suitable to predict portal vein and liver concentrations as well.
Figure 4C shows the time course of mean macitentan and ACT-132577 concentrations in the portal vein, plasma, and liver. Peak concentrations of macitentan in the portal vein and liver after a single 10-mg dose to healthy male subjects were estimated as 354 nM (208 ng/ml) and 62 nM (36 ng/ml), respectively. The corresponding values for ACT-132577 were 420 nM (223 ng/ml) and 39 nM (21 ng/ml). As ACT-132577 is generated by hepatic metabolism, the plasma and portal vein concentrations are identical. The concentration versus time profiles for both compounds at steady state are illustrated in Fig. 4D. Predicted total peak concentrations for macitentan at steady state were 400 nM (235 ng/ml) in plasma, 419 nM (247 ng/ml) in the portal vein, and 72 nM (43 ng/ml) in the liver. The corresponding values for ACT-132577 were 1722 nM (915 ng/ml) for plasma/portal vein and 159 nM (85 ng/ml) for the liver.
A parameter sensitivity analysis of the PBPK model was performed to investigate the effect of various degrees of active hepatic uptake on the pharmacokinetic parameters of macitentan, with the liver as a permeability-limited instead of a perfusion-limited organ. Passive uptake clearance (CLint, passive) into human hepatocytes was estimated from the macitentan logD of 2.9 as 5.3 μl/min/106 cells (Ménochet et al., 2012). Four different values for active uptake clearance (CLint, active)—that is, 0, 10, 25, and 50 μl/min/106 cells—were then arbitrarily assigned to OATP1B1 as a prototypical hepatic uptake transporter. Ten virtual trials of 10 subjects each were again run (total size: 100) for a single dose of 10 mg without any other changes to the PBPK model. A comparison between the predicted and observed pharmacokinetic parameters is shown in Table 8. In the permeability-limited PBPK liver model, the best fit between modeled and observed macitentan AUC0–24 h is achieved with a CLint, active of 25 μl/min/106 cells, for which a ratio of unbound liver and plasma concentrations of 1.05 is predicted. The outcome is then closest to the perfusion-limited PBPK model, in which the ratio of unbound concentrations is per definition set to unity.
“Dosis sola facit venenum” is the scientific legacy of the Swiss physician, alchemist, astrologer, and philosopher Paracelsus (1493–1541). About 500 years later, its longer version “all things are poison, and there is nothing without poisonous qualities . . . it is only the dose which make a thing poison” is still a major paradigm in modern pharmacology in its attempt to find drugs with adequate risk-benefit profiles. Nowadays, the interplay of target affinity and local drug concentrations has replaced the historical dose term as a measure for drug quantity. Whether a drug reaches its target in sufficient concentrations is mostly dependent on its physicochemical properties, as drug distribution to the majority of organs and tissues is driven by passive diffusion. Although total drug concentrations can vary significantly among tissues as a consequence of differing binding properties, free drug concentrations strive to equilibrate. This general concept needs to be expanded for organs expressing transport proteins that are capable of maintaining nonequilibrium conditions and that, consequently, depending on their transport direction, result in lower or higher free drug concentrations on one side of a membrane. Liver, kidney, brain, and placenta are typical examples of such organs, and the discovery and characterization of transporters like the ATP-dependent efflux pump P-glycoprotein or the organic anion-transporting polypeptide family OATP has fundamentally changed our understanding of drug disposition.
Endothelin receptor antagonists have been established as a therapeutic concept for the treatment of pulmonary arterial hypertension. From a chemical perspective, they require a negative charge (Wu, 2000; Boss et al., 2002) for their interaction with Arg326 of the endothelin receptors (Breu et al., 1995). In the case of sitaxsentan and ambrisentan, this negative charge is provided by a carboxylic acid whereas bosentan contains an aromatic sulfonamide. As a consequence, the majority of drug molecules are negatively charged at physiologic pH, which–together with high binding to plasma proteins—limits distribution into tissues. On the other hand, the negative charge makes them candidate substrates for anion transporters, and interactions with OATP transport proteins have indeed been described (Katz et al., 2006; Treiber et al., 2007; Spence et al., 2010). Hepatic side effects—mostly manifesting as transient reversible elevations of liver transaminases in plasma—have also been reported for some marketed endothelin receptor antagonists (Humbert et al., 2007; Galie et al., 2011). Active drug uptake by OATP transporters has the potential to increase hepatic drug concentrations and thus to contribute to liver injury. The combination of BSEP inhibition and OATP-mediated accumulation in liver cells is the likely mechanism for the cholestatic effect of bosentan (Fattinger et al., 2001; Treiber et al., 2007). In line with this hypothesis is the observation in pulmonary arterial hypertension patients concomitantly receiving the human immunodeficiency virus protease inhibitor lopinavir/ritonavir, a potent inhibitor of both CYP3A4 metabolism and OATP transport (Hull et al., 2009; Annaert et al., 2010): bosentan trough plasma concentrations were increased by 48-fold in this patient population but were not associated with a higher frequency of liver injury (Dingemanse et al., 2010), most likely because the hepatic drug burden was reduced as a consequence of blocked OATP uptake.
These points illustrate the necessity to discover novel endothelin receptor antagonists with an improved safety profile. The medicinal chemistry program leading to the discovery of macitentan and its pharmacologic profile in animal models were published previously (Iglarz et al., 2008; Bolli et al., 2012). Key structural changes in macitentan versus bosentan constitute the replacement of the sulfonamide by a sulfamide moiety and the increase in overall compound lipophilicity. The sulfamide function in macitentan is less acidic than the sulfonamide in bosentan, as evidenced by the difference in pKa values of 6.2 and 5.1 (Iglarz et al., 2008), respectively, thus increasing the proportion of molecules present in nonionized state at physiologic pH. The octanol/water partition coefficient logD7.4, a measure of lipophilicity, is 2.9 for macitentan and 1.3 for bosentan. Both factors facilitate membrane permeability and penetration into tissues. These changes were achieved without compromising target affinity. Macitentan antagonizes the specific binding of endothelin (ET)-1 to recombinant ETA and ETB receptors with IC50 values of 0.5 and 391 nM, respectively (Iglarz et al., 2008). In ex vivo models using rat aorta and trachea preparations, macitentan behaved as a more dual endothelin receptor with pA2 values of 7.6 and 5.9 for ETA and ETB receptors, respectively (Iglarz et al., 2008). Metabolic stability was optimized in vitro and favorably translated in animals (Bolli et al., 2012) and humans (Sidharta et al., 2011, 2013). As a consequence, macitentan was used at daily doses of 3 mg and 10 mg in patients in the completed phase III Study with an Endothelin Receptor Antagonist in Pulmonary Arterial Hypertension to Improve Clinical Outcome (SERAPHIN) trial (Pulido et al., 2013). Macitentan has an active metabolite, ACT-132577, that forms in a cytochrome P450–catalyzed reaction, circulates in human plasma, and likely contributes to the overall efficacy.
Beyond improved receptor-binding affinities and pharmacokinetic properties over bosentan, the structural changes in macitentan have altered the hepatic disposition and interaction profile with hepatic transporters. As illustrated in Figs. 1 and 2, macitentan and ACT-132577 are inhibitors of the two major transport proteins responsible for hepatic bile salt trafficking: NTCP and BSEP. NTCP (SLC10A1) is the major transporter responsible for moving bile salts from blood into liver cells (Doring et al., 2012) whereas BSEP mediates the rate-limiting step in bile salt secretion from blood into bile (Gerloff et al., 1998; Stieger et al., 2000) (Fig. 5). NTCP was equally inhibited by macitentan and ACT-132577 with IC50 values in the range of 14–18 μM, whereas BSEP was inhibited with IC50 ranging from 18–60 μM. These in vitro data are thus not different from those of bosentan with reported IC50 values on BSEP and NTCP of 25–77 μM (Mano et al., 2007; Dawson et al., 2012; Warner et al., 2012) and 24–30 μM (Leslie et al., 2007), respectively.
However, the mechanism of hepatic uptake differs between bosentan and macitentan. Bosentan uptake into liver cells is largely dependent on OATP transport in animals and humans (Treiber et al., 2004, 2007). In contrast, its increased lipophilicity allows macitentan to enter cells by passive diffusion, as shown in the uptake experiments with OATP-expressing and wild-type cells (Table 1). The same conclusion was drawn in a recent study on macitentan interactions with other hepatic transport proteins (Weiss et al., 2013). Macitentan is a weak substrate for OATP1B3 (Fig. 3) but not for OATP1B1, OATP2B1, or NTCP. Active transport contributes less than 20% to overall cellular uptake. Consequently, macitentan uptake is not vulnerable to OATP inhibition by cyclosporin A or rifampicin (Table 3). These conclusions were confirmed in a clinical drug-drug interaction study with cyclosporin A, an established potent inhibitor of OATP-mediated (Shitara et al., 2003) and NTCP-mediated transport (Mita et al., 2006). Concomitant cyclosporin A increased macitentan exposure in humans by only 1.1-fold (Bruderer et al., 2012a), compared with a 30-fold increase observed with the proven OATP substrate bosentan (Treiber et al., 2007) in the presence of cyclosporin A (Binet et al., 2000).
Bosentan elicited a dose-dependent increase in plasma bile salts in the rat after a single intravenous administration (Fattinger et al., 2001). Macitentan and ACT-132577 were both tested in this model alongside with bosentan, all at a dose of 25 mg/kg. Results with bosentan reproduced literature data, while no effect was observed with macitentan or ACT-132577 (Table 4). These data provide evidence in a more physiologic setting that neither compound interferes with bile salt homeostasis under conditions that led to an acute cholestatic effect with bosentan. Supportive data upon chronic dosing was provided by the bile salt data from the rat and dog toxicity studies. After 26 or 39 weeks of treatment with macitentan at doses up to 250 mg/kg (rat) and 100/75 mg/kg (dog), there was no pattern of elevated plasma bile salts in any toxicity study nor were there any histologic findings pointing to a cholestatic potential of macitentan (Actelion, data on file). In contrast, bosentan-treated animals showed elevated bile salt and alanine aminotransferase levels, which were considered to be consequences of a functional cholestasis (Actelion, data on file).
A PBPK model was developed for macitentan and ACT-132577 with the aim of estimating blood and liver concentrations. Total mean macitentan concentrations projected from PBPK modeling at steady state in plasma, portal vein, and liver are 400, 419, and 72 nM, respectively. The corresponding values for ACT-132577 were 1722 nM for plasma/portal vein and 159 nM for the liver. Correction for plasma protein binding leads to predicted unbound peak macitentan concentrations of 1.60 nM in plasma, 1.68 nM in the portal vein, and 0.29 nM in the liver. Free concentrations for ACT-132577 in plasma/portal vein and liver were 8.61 and 0.80 nM, respectively. These concentrations are significantly below those observed in the bile salt transport experiments with NTCP- and BSEP-expressing cells. The key assumption of free partitioning of macitentan between blood and liver and equal unbound drug concentrations in both compartments is supported by the experimentally determined Kp,uu of 0.7 in human hepatocytes and the sensitivity analysis of the PBPK model probing for the impact of various degrees of active hepatic uptake on macitentan pharmacokinetics. In this permeability-limited PBPK model, the best fit between observed and predicted macitentan pharmacokinetics is achieved when the ratio of unbound plasma and liver concentrations approaches unity (Table 8).
Serum bile salts were measured in the multiple ascending dose study with macitentan in which daily doses up to 30 mg were given to healthy volunteers for 10 days (Sidharta et al., 2013). As shown in Table 6, there was no difference in bile salts when the dose groups were compared with placebo. All individual values were in a narrow range from 7.7–19 μM. Liver transaminases but not bile salts were measured in the long-term phase III SERAPHIN trial with macitentan (Pulido et al., 2013). The incidence of liver enzyme elevations more than 3 times the upper limit of normal was not different between placebo and the 3-mg and 10-mg dose groups. The incidences of alanine or aspartate aminotransferase elevations more than 3 times above the upper limit of normal were 4 and 3% in the 3 mg and 10 mg dose groups, respectively, compared with 4% on placebo (Pulido et al., 2013). Similarly, the incidence of plasma bilirubin elevations was equally distributed over all dose groups and placebo.
In conclusion, there is cumulative evidence demonstrating the superior liver safety profile of the next generation endothelin receptor antagonist macitentan. Unlike bosentan, macitentan is devoid of a pattern of functional cholestasis in long-term clinical trials. On a molecular basis, this difference is the result of discrete changes in chemical structure. While macitentan maintains equal inhibitory potency to bosentan on NTCP and BSEP transport in vitro, replacement of the sulfonamide by a less acidic sulfamide moiety and an increased lipophilicity result in a complete change in the hepatic disposition profile. Bosentan uptake into the liver is an active process, mediated by OATP transport proteins, likely leading to accumulation in liver cells. In contrast, macitentan partitions into the liver mostly by passive diffusion. Local drug concentrations are thus limited by the extensive binding of macitentan to plasma proteins to levels that are unlikely to exert an inhibitory potential on hepatic bile salts. These conclusions from biochemical and drug disposition data were confirmed in an acute rat model and long-term toxicity studies in the rat and dog. In clinical trials, no change in plasma bile salts was observed in healthy volunteers upon multiple-dose treatment up to 30 mg per day, nor was the incidence of liver transaminase elevations different in macitentan- versus placebo-treated patients in the long-term morbidity/mortality SERAPHIN trial. Macitentan at a daily dose of 10 mg is therefore not expected to interfere with hepatic bile salt transport in clinical practice.
The authors thank Aude Weigel, Eric Soubieux, Julia Friedrich, and Stephanie Bernhard for dedication and experimental contributions, and Susan Flores and Charlotte Gonzales for help in preparing the manuscript.
Participated in research design: Treiber, Äänismaa, Delahaye, Treher, Hess, Sidharta.
Conducted experiments: Treiber, Äänismaa, Delahaye, Treher, Hess, Sidharta.
Performed data analysis: Treiber, Äänismaa, Treher, de Kanter, Delahaye, Hess, Sidharta.
Wrote or contributed to the writing of the manuscript: Treiber, Äänismaa, Treher, de Kanter, Hess, Sidharta.
- Received February 19, 2014.
- Accepted April 23, 2014.
- area under the plasma concentration versus time curve
- bile salt export pump
- Chinese hamster ovary
- intrinsic clearance
- oral clearance
- dimethyl sulfoxide
- Hanks’ balanced salt solution
- liquid chromatography–tandem mass spectrometry
- sodium taurocholate cotransporting polypeptide
- organic anion-transporting polypeptide
- physiologically-based pharmacokinetic (modeling)
- Study with an Endothelin Receptor Antagonist in Pulmonary Arterial Hypertension to Improve Clinical Outcome
- Copyright © 2014 by The American Society for Pharmacology and Experimental Therapeutics