Article pubs.acs.org/molecularpharmaceutics
Luminal Lipid Phases after Administration of a Triglyceride Solution of Danazol in the Fed State and Their Contribution to the Flux of Danazol Across Caco-2 Cell Monolayers Maria Vertzoni,† Constantinos Markopoulos,†,§ Moira Symillides,† Constantinos Goumas,‡ Georgios Imanidis,§ and Christos Reppas*,† †
Laboratory of Biopharmaceutics and Pharmacokinetics, National & Kapodistrian University of Athens, Zografou, Greece Department of Gastroenterology, Red Cross Hospital of Athens, Athens, Greece § Institute of Pharma Technology, University of Applied Sciences Northwestern Switzerland, Muttenz, Switzerland ‡
ABSTRACT: The first aim of this study was to characterize the luminal contents and their micellar phase after the administration of a heterogeneous liquid meal to healthy adults. The second aim was to evaluate the impact of micellar lipids and coarse lipid particles on danazol flux through intestinal monolayers. A third aim was to compare the micellar composition in the upper small intestine with the composition of fed state simulating intestinal fluid (FeSSIF-V2), a medium that has been proposed for investigating dissolution of poorly soluble drugs in the fed state. Danazol (150 mg), predissolved in the olive oil portion of the meal, was administered via the gastric port of a two-lumen tube to the antrum of eight adults. Aspirates from the ligament of Treitz [collected up to 4 h postdosing (∼15 mL every 30 min)] were characterized physicochemically. Comparison of these characteristics with FeSSIF-V2 indicates that FeSSIFV2 is an appropriate medium for evaluating drug solubilization in the luminal micellar phase in the fed state. Individual aspirates and their corresponding micellar phases were also diluted with aqueous transport medium and subjected to Caco-2 cell permeation experiments. Permeability coefficients for danazol in the diluted aspirates were smaller than those for the diluted micellar phases, which in turn were similar to those for aqueous transport medium. The high danazol concentrations overcompensated the reduced permeability coefficient values in the diluted aspirates in terms of total drug flux. We conclude that drug dissolved in the coarse lipid particles formed after administration of a triglyceride solution can directly contribute to the flux of lipophilic drugs across the intestinal mucosa. KEYWORDS: human aspirates, micellar composition, fed state, danazol, Caco-2 permeability coefficients, transport rates
■
The first objective of this study was to characterize the luminal contents and their micellar phases in the fed state. The second objective was to evaluate the impact of nonmicellar and micellar lipid components on the flux of danazol (nonionizable, log P 4.24) across Caco-2 cell monolayers. For these purposes, a pharmacologically relevant dose of danazol (150 mg) was predissolved in the olive oil portion of a high fat, heterogeneous liquid meal and administered to healthy volunteers. Aspirates were then collected from the ligament of Treitz. After appropriate dilution of the aspirates and their micellar phases with aqueous transport medium, the permeability coefficients of danazol across Caco-2 cell monolayers were measured, and the contributions of various phases to transport were estimated. A third objective was to compare the micellar composition in the upper small intestine in the fed state with FeSSIF-V2,
INTRODUCTION In the fasted state, the drug can reach the intestinal mucosa via the aqueous phase and the mixed bile salt micelles. Mixed bile salt micelles enhance the solubilization of lipophilic compounds, but primarily due to their size, their diffusion toward the intestinal epithelium is slower than that of the single lipophilic drug molecules. Depending on their capacity, solubilization in mixed bile salt micelles can enhance mass transport of lipophilic compounds.1 After administration of a lipid dosage form and when dosage forms are administered in the fed state, other colloidal structures are also available in the intestinal lumen for transferring the drug toward the intestinal epithelium. Although it is generally believed that, even in these situations, lipophilic drugs are transported primarily via the mixed bile salt micelles,2 there is little relevant information in the literature. For example, it is known that lipids (in coadministered food and/or in the dosage form) enhance the bioavailability of danazol by increasing the extent of luminal solubilization3 but whether danazol reaches the intestinal mucosa primarily via the bile salt micelles or via other structures has not been directly investigated. © 2012 American Chemical Society
Received: Revised: Accepted: Published: 1189
September 15, 2011 April 3, 2012 April 7, 2012 April 7, 2012 dx.doi.org/10.1021/mp200479f | Mol. Pharmaceutics 2012, 9, 1189−1198
Molecular Pharmaceutics
Article
Greece) had been predissolved, 1.25 eggs, 25 g of sucrose, 2.7 g of NaCl, vanilla flavor, and tap water to reach a final volume of 500 mL. Prior to administration, the meal was vigorously mixed by mechanical blending for 10 s to generate a coarse emulsion.10 Samples of the duodenal contents (∼15 mL) were aspirated from a location near the end of the duodenum every half hour up to 4 h after the administration of the meal. For each sample, volume, pH, and buffer capacity were measured immediately upon collection. After adding a cocktail of lipase/ protease inhibitors (2% v:v) consisting of 50 mM diisopropylfluorophosphate, 50 mM diethyl(p-nitrophenyl)phosphate, 50 mM acetophenone, and 250 mM phenylboronic acid,8 the sample was divided in two subsamples (Figure 1). The first subsample was further divided in 5 portions that were kept at −70 °C until analysis of danazol, osmolality, total protein content, bile acids, and lipids. The second subsample was ultracentrifuged (410174g, 37 °C, 2 h) so that four phases were obtained (Figure 1), that is, triglyceride (TG) phase, interphase, primarily micellar phase, and pellet. It should be acknowledged that although collection of pure micellar phase may not be possible,11 for the sake of simplicity in this manuscript, the more precise description “primarily micellar phase” will be referred to as the “micellar phase”. Likewise, lipids that are present in dispersed form in the aspirate are referred to as “coarse lipid particles”. The TG and pellet phases were analyzed for their danazol content. Part of the micellar phase was carefully separated and divided into 6 portions (Figure 1) that were stored at −70 °C until measuring the following parameters: danazol content, danazol solubility, osmolality, total protein content, bile salts, and lipids. At the end of the collection period and before the tube was removed, the final position of the tube was confirmed fluoroscopically, the tube was removed, and after a brief examination, the subject was offered a complementary lunch, and then, he/she was discharged from the clinic. Danazol Solubility Measurements. The solubility of danazol was measured in samples of the micellar phases obtained from the individual aspirates. Because of limited volumes of these phases, the solubility in each sample could be measured only once. For each measurement, 1 mL of the micellar phase and 1 mg of danazol were transferred into 2 mL polypropylene vials. Vials were sealed with polypropylene caps and put in an oscillating water bath (37 °C, 75 oscillations/min). After they reached equilibrium (4 h after initiation of shaking), samples were centrifuged for 10 min at 11400g at 37 °C. The supernatant was appropriately diluted with acetonitrile, centrifuged again (10 min, 11400g, 10 °C), and assayed for drug concentration by HPLC. The equilibration time has been estimated beforehand in fed state simulating intestinal fluid (FeSSIF).12 The adequacy of centrifugation vs filtration (0.45 μm regenerated cellulose filters) was confirmed by measuring solubility in FeSSIF in triplicate. Caco-2 Cell Monolayer Experiments with Aspirates and Their Micellar Phases. The origin of the cell lines and the materials used for culture and transport experiments as well as standard culture conditions for Caco-2 cells have been described previously.13 The aqueous transport medium consisted of DMEM base (8.26 g/L) without L-glutamine, glucose, phenol red, sodium pyruvate, sodium bicarbonate, or FCS (foetal calvine serum), which was supplemented with D-glucose (4.5 g/L), HEPES (4.76 g/L), NaCl (1.987 g/L), and L-glutamine (0.876 g/L) (all from Sigma, St. Louis, MO) and adjusted to pH 7.4. Six-well Transwell plates (Costar, Corning Corporation, NY) with a surface area of 4.71 cm2 and a 0.4 μm pore size polycarbonate
which, based on luminal data after administration of a commercially available stable emulsion (Ensure plus), has been proposed as a medium to simulate the micellar solubilization of drugs in the fed small intestine.5 It would be interesting to know if the nutrient composition of a meal with similar caloric content and its digestion in the upper gastrointestinal tract have any impact on luminal environment. A high fat meal (750 kcal, 13% carbohydrate, 73% fat, and 14% protein) was used in the present study in accordance with current guidelines for the composition of meal to be used in oral bioavailability studies.6,7
■
EXPERIMENTAL SECTION Collection of Human Aspirates. Human aspirates were collected at the Red Cross Hospital of Athens after receiving approval by the (Greek) National Organization for Medicines (KΔ-IS-98-01-05). Eight healthy adults (seven males and one female) participated in the study. Subjects were between 22 and 34 years old and deviated by not more than 15% from their ideal body weight. Inclusion Criteria. Willingness of the subject to participate was indicated by his/her signed informed consent, age 18− 60 years, weight within 15% of ideal body weight as determined by Metropolitan Life Tables, verification of suitability by a general physical examination, and ability to abstain from cigarette smoking, alcohol, and over-the-counter and prescription medication(s) for 3 days prior to and throughout the experimental day. In addition, a blood sample was taken to assess electrolyte balance, kidney and liver function, blood morphologic characteristics, and lipid levels, and the subject had to be found healthy in all of these examinations to qualify. Exclusion Criteria. The existence of a major health problem (cardiovascular, pancreatic, hepatic, thyroid, etc.) and/or existence of any condition requiring prescription drug therapy, recent history of gastrointestinal symptoms regardless of the severity (e.g., heartburn, constipation, hemorrhoids, etc.), women who were pregnant, lactating, or had been on birth control pills for less than 3 months, receipt of an investigational agent (new or generic) within 30 days prior to the initiation of study, the presence of antibodies indicating active acute or chronic HIV, HBV, or HCV infection, use of medication that may affect GI function (including antibiotics) within 30 days of the study, and irregular bowel habits were exclusion criteria. Study Protocol. Each subject reported to the clinic in the morning after fasting for at least 12 h. If female, the subject was tested for pregnancy. The subject was then intubated nasally using a sterile two-lumen duodenal tube (Freka Trelumina Ch/Fr 16/9, 150 cm, ref no. 7550911). The two-lumen tube was 150 cm long with an external diameter of 5.3 cm. A series of holes (55−65 cm proximal to the end of the tube) were used to access the antrum of the stomach. A further series of handmade holes (13.5−23.5 cm proximal to the end of the tube) were used to aspirate samples from the duodenum. Insertion of the tube was assisted by a guiding wire, and its position was monitored fluoroscopically. After its final position was reached and the wire was removed, a heterogeneous liquid meal containing the drug was administered to the antrum using 60 mL (capacity) syringes. The composition of the administered meal was based on a previous meal used for studying luminal lipids during duodenal fat digestion,8 after adjusting the caloric content so that it becomes closer to that of Ensure plus9 and to the high-fat requirement by the regulatory agencies.6,7 Specifically, the meal consisted of 62.5 g of olive oil in which 150 mg of danazol (NLT 98.5% pure, Sanofi-Synthelabo A.E., Peania, 1190
dx.doi.org/10.1021/mp200479f | Mol. Pharmaceutics 2012, 9, 1189−1198
Molecular Pharmaceutics
Article
Figure 1. Diagram showing how each aspirated sample was treated and divided to be physicochemically characterized. A double-lined box implies a specific vial that was stored at −70 °C until the relevant measurement was made.
membrane were used. Soybean lecithin (S 100) was donated by Lipoid GmbH (Ludwigshafen, Germany). Twelve aspirated samples (each representing a specific individual and aspiration time) and their corresponding micellar phases were used in the Caco-2 cell line experiments. Aspirated samples and their corresponding micellar phases were selected on a volume availability basis. It has been previously reported that human aspirates may have to be diluted with aqueous transport medium to make Caco-2 permeation experiments possible, even in the fasted state.14 In this study, the degree of dilution of aspirates and their micellar phases required to ensure cell monolayer integrity was determined in the Transwell as follows: The culture medium was removed, and the cells were washed with Dulbecco's PBS containing Ca2+ and Mg2+. Cells were incubated for at least 1 h at 37 °C in 8% CO2 with 0.8 and 2.4 mL of the aqueous transport medium on the apical and the basolateral side, respectively, and the transepithelial electrical resistance (TEER) was measured using an EVOM epithelial voltohmmeter (World Precision Instruments, Sarasota, FL). TEER values were typically between 500 and 650 Ω × cm2. Occasionally, values as low as 250 Ω × cm2 were measured, and these values were tolerated on the grounds that permeability of danazol was not expected to depend on the paracellular pathway. The transport medium on the apical side was replaced by aspirates or dilutions thereof with transport medium, the plates were agitated at 37 °C in a water-saturated atmosphere under an incubator hood (TH15, Edmund Bühler GmbH, Tübingen & Hechingen, Germany) at a stirring rate of 75 rpm on an orbital shaker (KS15, Edmund Bühler GmbH, Tübingen & Hechingen, Germany), and TEER was measured at 15 min and, if the value was maintained, at 30, 60, 90, 120, and 180 min. Successive dilutions of aspirates were tested until at least 75% of the TEER value was maintained for up to 180 min. A 1:16 dilution of the aspirates and a 1:8 dilution of
their micellar phase were found to be necessary for all aspirated samples to fulfill this monolayer integrity criterion. For transport experiments, a 1:16 dilution was applied to both the aspirates and their micellar phase to facilitate comparison of results. Drug transport was measured in the apical-to-basolateral direction under experimental conditions identical to those of the integrity test described above, with the exception that the aqueous transport medium in the acceptor compartment was replaced by a liposome dispersion prepared with Lecithin S100 in the transport medium.13 This was necessary to improve the solubility of danazol in the acceptor solution (exact match of solubility in the receiving compartment may not be possible given the changing composition in the receiving compartment due to the potential transport of lipid components in addition to danazol whereas use of identical blank aspirated sample is impossible to have). The addition of the aspirates to the apical (donor) chamber marked the beginning of the permeation experiment. Samples of 100 μL were withdrawn at 10, 20, 30, 45, 60, 90, 120, and 180 min from the acceptor solution and at 180 min from the apical solution. Withdrawn volumes were not replaced. TEER was measured at the end of each experiment. All experiments were performed in triplicate. To assess mass balance, the danazol content of the cellular compartment at the end of the permeation experiment was determined as follows: The donor and the acceptor solutions were removed, and the cells were washed with Dulbecco's PBS not containing Ca2+ or Mg2+. An amount of 0.3 mL of trypsinEDTA solution was added to the insert, which was then transferred in a Petri dish, and after incubation at 37 °C for 15 min and the addition of 1 mL of cell culture medium, the cell monolayer was mechanically scraped off the membrane (BD Falcon Cell Scraper, BD Biosciences Discovery Labware, Bedford, United States). The cell suspension was collected in 1191
dx.doi.org/10.1021/mp200479f | Mol. Pharmaceutics 2012, 9, 1189−1198
Molecular Pharmaceutics
Article
Macherey-Nagel, Switzerland) was used. The mobile phase consisted of acetonitrile/water/acetic acid 55:45:0.5 v:v:v, and the flow rate was 0.25 mL/min. The detection wavelength in UV was 286 nm. The MS detector employed atmospheric pressure electrospray ionization source and was operated in the scan mode at positive polarity with a capillary voltage of 4000 V, fragmentor of 100 V, drying gas flow of 10 L/min, drying gas temperature of 350 °C, and nebulizer pressure of 30 psig. Danazol was detected in MS at m/z 338. UV was used for quantifying the danazol concentration in donor and receiving compartments according to previously described method.17 Because of an interfering peak, MS was used for quantifying the danazol concentration in extracts from the walls of wells. The lowest concentration quantified in those samples was 1.2 ng/mL. All chemicals used for the optimization and application of analytical methods were of analytical grade and purchased from Sigma Aldrich Chemie GmbH or E. Merck (Germany), except for egg lecithin (E PC), which was kindly donated by Lipoid GmbH. Data Treatment. The percentage of danazol in the nontriglyceride phases (interphase, micellar phase, and pellet) was estimated from the amount of danazol in the aspirate (prior to ultracentrifugation) and the amount of danazol in the olive oil of the aspirate (after ultracentrifugation). Permeability coefficients were estimated by fitting a five compartment kinetic model to the data using EASY-FIT Model Design software (http://www.ai7.uni-bayreuth.de/easy_fit.htm, Dr. K. Schittkowski, University of Bayreuth, Germany). The apical, the cellular, and the basolateral compartment and the apical and the basolateral Transwell wall comprised the five compartments. Drug transport between the apical, the cellular, and the basolateral compartment was assumed to occur by (passive) diffusion and binding to the Transwell wall by a physicochemical adsorption process. The change of drug concentration or mass in each of the five compartments as a function of time was described by the following equations:
15 mL centrifuge tubes (BD Falcon blue max 15 mL, BD Biosciences Discovery Labware) and centrifuged for 5 min at 200g (5810R, Eppendorf AG, Hamburg, Germany), the supernatant was discarded, and the pellet was resuspended in 750 μL of deionized water at 37 °C and transferred to an Eppendorf microtube. The suspension was vortexed, stored at −80 °C for at least 15 min, thawed in a thermomixer (Thermomixer comfort, Eppendorf, Hechingen, Germany) at 37 °C and 1400 oscillations per minute (opm) supplemented with 750 μL of acetonitrile, and placed in an ice−water bath for 20 min. The sample was thawed in the thermomixer for 10 min at 37 °C and 1400 opm and centrifuged for 3 min at 16000g (5415R, Eppendorf AG), and the water−acetonitrile supernatant was stored in a microtube at 4 °C. A volume of 750 μL of acetonitrile was added to the remaining pellet, and the suspension was subjected to six pulses with an ultrasonic disintegrator (Branson Sonifier 250, model 100-132-868, Branson Ultrasonics Corporation, Danbury, CT; output control: 2/duty cycle: 30%). The suspension was agitated in the thermomixer at 37 °C for 5 min and centrifuged at 16000g for 10 min. The supernatant (acetonitrile) was transferred to another microtube, the pellet was subjected to a second ultrasonication−extraction cycle, and the supernatants were evaporated under nitrogen stream. The residue was taken up in the stored water−acetonitrile solvent of the first extraction, and the sample was agitated for 3 min at 37 °C and centrifuged at 16000g for 25 min before HPLC analysis. Danazol was extracted from the plastic parts of the Transwell plate upon completion of permeation experiment by adding 0.8 mL of acetonitrile to the apical and 2.4 mL to the basolateral compartment. The plate was sealed with parafilm to avoid evaporation and was agitated as described in the cell integrity test for 45 min. Samples of 500 μL were withdrawn from each compartment, supplemented with 500 μL of water, and analyzed by HPLC. Analytical Methods. Osmolality, pH, and total protein content were measured as described previously.9 Depending on the sample, the buffer capacity was measured by dropwise addition of 0.1 or 0.01 M HCl solution and calculated as described previously.9 Lipids, that is, TGs, diglycerides (DGs), monoglycerides (MGs), fatty acids (FAs), phosphatidylcholine (PC), lyso-phosphatidylcholine (Lyso-PC), cholesterol (CHO), and cholesterol ester (CE) were assayed as described previously.15 Glycerol trioleate was used for quantifying TGs, 1,2 dioleoyl-rac-glycerol and dipalmitin were used for quantifying DGs, and 1-monooleyl-rac-glycerol was used for quantifying MGs. Oleic acid, palmitic acid, linoleic acid, and stearic acid were used for quantifying FAs. Cholesterol oleate was used for quantifying CEs. Bile acids, that is, taurocholic acid (TC), glycocholic acid (GC), taurochenodeoxycholic acid (TCDC), glycochenodeoxycholic acid (GCDC), and glycodeoxycholic acid (GDC), were quantified as described previously.16 The determination of danazol in the aspirates, in the TG, micellar, and pellet phases of the aspirates, and, in solubility measurements was performed using a previously described assay methodology.17 The analytical column was a Hypersil BDS C18 (250 mm × 4.6 mm, 5 μm). The mobile phase consisted of water:acetonitrile (45:55 v:v), and the flow rate was 1 mL/min. The detection wavelength was 286 nm. The injection volume was 50 μL, and the retention time was about 12 min. In the Caco-2 cell experiments, danazol was assayed with HPLC-UV-MS (Agilent series 1200). A C18 reversed phase column (CC 125 mm × 2 mm, Nucleodur 100-5, C18 ec,
dCa P(Ca − KCc)S k C (W − Wa) k W =− − on a tot + off a dt Va Va Va (1)
dC b P(KCc − uC b)S k C (5Wtot − Wb) k W = − on b + off b dt Vb Vb Vb (2)
dCc P(Ca − KCc)S P(KCc − uC b)S = − dt Vc Vc
(3)
dWa = konCa(Wtot − Wa) − koff Wa dt
(4)
dWb = konC b(5Wtot − Wb) − koff Wb dt
(5)
where a, b, and c denote the apical, basal, and cellular compartment, respectively, C is the danazol concentration, K is the drug partition coefficient between the apical and the cellular compartment, kon is the second order rate constant of drug binding to the surface of the Transwell wall, koff is the first order rate constant of drug desorption off the surface of the Transwell wall, P is the permeability coefficient, u is the drug partition coefficient between the apical and the basolateral compartment (to account for a possible asymmetry in terms of danazol 1192
dx.doi.org/10.1021/mp200479f | Mol. Pharmaceutics 2012, 9, 1189−1198
Molecular Pharmaceutics
Article
Figure 2. Danazol concentration vs time plots and danazol mass vs time plots (experimental data and best fitted lines) during the Caco-2 transport experiment performed after diluting (1:16 with aqueous transport medium) the sample aspirated 90 min after meal administration to volunteer #6 (a and b) and the micellar phase of the aspirate (c and d). The total lipid concentrations in the aspirate and in its micellar phase (prior to dilutions) were 88.98 and 24.26 mM, respectively. In a and c, the dotted line and associated white circles refer to the danazol concentration in the apical compartment, whereas the continuous line and associated black circles refer to the danazol concentration in the basolateral compartment. In b and d, the dotted line and associated white circles refer to danazol mass in the cells, whereas the continuous line and associated black circles refer to danazol mass in the Transwell walls. r2 values for the fits to the aspirate data (a and b) and to the micellar phase data (c and d) were 0.996 and 0.9995, respectively.
solubility between the apical and the basolateral compartment), V is the volume of solution in each compartment, W is danazol mass bound on the walls, Wtot is the total binding capacity of the apical Transwell wall, and S is the surface area of the monolayer. Va was constant during the experiment (0.8 mL), and Vc was 0.0094 mL based on a monolayer thickness of 20 μm. The reduction of the volume of solution in the basal compartment (Vb) due to sampling as a function of time was accounted for by the following relation that was determined empirically using regression analysis.
these parameters occurred. The drug partition coefficient between the apical and the cellular compartment, K, was estimated to be 0.00805 ± 0.00491 and 0.00194 ± 0.00079 for the aspirates and for their micellar phases, respectively. These low values indicate a high affinity of the drug for the interior of the cell, consistent with the high lipophilicity of danazol. The value for the aspirates is significantly larger than the value for the micellar phase (p < 0.001) in agreement with the higher lipid concentration in the aspirates. The drug partition coefficient between the donor and the acceptor compartments, u, was
