Gastrointestinal transfer: In vivo evaluation and implementation in in vitro and in silico predictive tools

Gastrointestinal transfer: In vivo evaluation and implementation in in vitro and in silico predictive tools

European Journal of Pharmaceutical Sciences 63 (2014) 233–242 Contents lists available at ScienceDirect European Journal of Pharmaceutical Sciences ...

1MB Sizes 1 Downloads 100 Views

European Journal of Pharmaceutical Sciences 63 (2014) 233–242

Contents lists available at ScienceDirect

European Journal of Pharmaceutical Sciences journal homepage: www.elsevier.com/locate/ejps

Gastrointestinal transfer: In vivo evaluation and implementation in in vitro and in silico predictive tools Bart Hens a,1, Joachim Brouwers a,1, Bart Anneveld b, Maura Corsetti c, Mira Symillides d, Maria Vertzoni d, Christos Reppas d, David B. Turner e, Patrick Augustijns a,⇑ a

Drug Delivery & Disposition, KU Leuven, Gasthuisberg O&N 2, Herestraat 49, Box 921, 3000 Leuven, Belgium TNO, Zeist, The Netherlands Translational Research Center for Gastrointestinal Disorders (TARGID), KU Leuven, Belgium d Laboratory of Biopharmaceutics and Pharmacokinetics, National and Kapodistrian University of Athens, Greece e Simcyp (a Certara Company) Limited, Sheffield S2 4SU, UK b c

a r t i c l e

i n f o

Article history: Received 25 May 2014 Received in revised form 4 July 2014 Accepted 15 July 2014 Available online 23 July 2014 Keywords: Gastrointestinal transfer Paromomycin TIM-1 Three compartmental in vitro model Simcyp Gastric emptying

a b s t r a c t Introduction: The purpose of this study was to explore the transfer of drug solutions from stomach to small intestine and its impact on intraluminal drug concentrations in humans. The collected intraluminal data were used as reference to evaluate simulations of gastrointestinal transfer currently implemented in different in vitro and in silico absorption models. Methods: Gastric and duodenal concentrations of the highly soluble and non-absorbable compound paromomycin were determined following oral administration to 5 healthy volunteers under the following conditions: fasted state, fed state and fed state in the presence of a transit-stimulating (domperidone) or transit-inhibiting (loperamide) agent. Based on the obtained intraluminal concentration–time profiles, gastrointestinal transfer (expressed as the half-life of gastric emptying) was analyzed using physiologically-based parameter estimation in SimcypÒ. Subsequently, the observed transfer profiles were used to judge the implementation of gastrointestinal transfer in 2 in vitro simulation tools (the TNO Intestinal Model TIM-1 and a three-compartmental in vitro model) and the SimcypÒ population-based PBPK modeling platform. Results: The observed duodenal concentration–time profile of paromomycin under fasting conditions, with a high average Cmax obtained after 15 min, clearly indicated a fast transfer of drug solutions from stomach to duodenum (estimated gastric half-life between 4 and 13 min). The three-compartmental in vitro model adequately reflected the in vivo fasted state gastrointestinal transfer of paromomycin. For both TIM-1 and SimcypÒ, modifications in gastric emptying and dilutions were required to improve the simulation of the transfer of drug solutions. As expected, transfer from stomach to duodenum was delayed in the fed state, resulting in lower duodenal paromomycin concentrations and an estimated gastric half-life between 21 and 40 min. Administration of domperidone or loperamide as transit-stimulating and transitinhibiting agent, respectively, did not affect the fed state gastric half-life of emptying. Conclusion: For the first time, the impact of gastrointestinal transfer of solutions on intraluminal drug concentrations was directly assessed in humans. In vitro and in silico simulation tools have been validated and optimized using the in vivo data as reference. Ó 2014 Elsevier B.V. All rights reserved.

1. Introduction After oral intake, the absorption of a drug depends on several intraluminal processes and physiological variables. Following ⇑ Corresponding author. Address: Drug Delivery & Disposition, Campus Gasthuisberg O&N 2, Box 921, Herestraat 49, 3000 Leuven, Belgium. Tel.: +32 16 330301; fax: +32 16 330305. E-mail address: [email protected] (P. Augustijns). 1 These authors should both be considered as first author. http://dx.doi.org/10.1016/j.ejps.2014.07.008 0928-0987/Ó 2014 Elsevier B.V. All rights reserved.

release from its dosage form, a drug has to dissolve in the gastrointestinal fluids and, subsequently, permeate the intestinal mucosa. These intraluminal processes are influenced by different physiological variables including gastrointestinal hydrodynamics, secretions of gastric and intestinal fluids and gastrointestinal transit. An adequate implementation of these physiologic factors in in vitro and in silico simulation tools is of crucial importance to accurately predict the behavior of orally administered drugs. However, this implementation often remains an issue, hampering the predictive value of multiple simulation models (Kostewicz et al., 2014).

234

B. Hens et al. / European Journal of Pharmaceutical Sciences 63 (2014) 233–242

For low solubility compounds, the rate and extent of absorption will mainly be determined by the intestinal concentrations that can be reached upon oral intake. Predicting the in vivo performance of absorption-enabling formulation strategies for this type of compounds therefore relies on the accurate simulation of intraluminal concentrations. This requires, inter alia, the adequate implementation of drug transfer from stomach to duodenum and subsequent dilution with mucosal secretions, bile and pancreatic fluid. The importance of gastrointestinal transfer is exemplified in the performance prediction of enabling strategies that rely on the creation of supersaturation, i.e. intraluminal concentrations exceeding thermodynamic solubility. Supersaturation can be created through formulation strategies, but, for basic compounds, also upon transit from stomach to the intestine (Brouwers et al., 2009). The degree of intestinal supersaturation and, consequently, the tendency for precipitation, can be significantly affected by the rate of gastrointestinal transfer and the extent of duodenal dilution. While the biorelevant evaluation of intraluminal supersaturation has recently been progressing (Bevernage et al., 2010, 2011), the adequate integration of gastrointestinal transfer remains an issue. Knowledge of gastrointestinal transfer is presently based on indirect methodologies such as the use of plasma concentration– time profiles of high solubility, high permeability compounds, e.g. paracetamol (Heading et al., 1973). Another widely reported technique is scintigraphy (Madsen, 2013) which translates the radiation profile of a specific marker compound into the transfer of the marker from stomach to duodenum. To avoid exposure to ionizing radiation, magnetic resonance imaging (MRI) is gaining more attention (Schiller et al., 2005). The use of water-sensitive magnetic resonance imaging can be used to monitor the transit of dosage forms, along with the movement of gastrointestinal fluids. This technique has recently demonstrated very short gastric residence times of a magnetically-marked capsule: below 3 min upon swallowing for five volunteers with the fastest emptying of the capsule out of the stomach being 15 s (Weitschies et al., 2005). Similar techniques have been used to study gastric emptying of liquid solutions. By use of MRI, the gastric emptying of a 300 ml aqueous solution was investigated: the observed gastric half-life of emptying (t1/2,G) amounted to 15.8 min (Steingoetter et al., 2006). In addition, Oberle et al. designed a study to correlate interdigestive motility and gastric emptying for 200 and 50 ml aqueous solutions of phenol red. The authors observed that a 200 ml aqueous solution leaves the stomach with a half-life of 11.8 min, being less dependent on gastric motility than a 50 ml aqueous solution (Oberle et al., 1990). Compared to gastric emptying of solid dosage forms, both studies demonstrated a trend for faster emptying of an aqueous solution (Culen et al., 2013). Complementary to these techniques, the purpose of this study was to provide a more direct approach to evaluate the impact of gastrointestinal transfer of drugs in solution on their intraluminal concentrations. To exclude the confounding effects of absorption, dissolution and/or precipitation, we selected the non-absorbable, highly soluble paromomycin as a model drug. Following oral intake of a solution, gastric and duodenal paromomycin concentrations were simultaneously monitored in healthy volunteers. The observed concentration–time profiles directly reflect the gastrointestinal transfer and simultaneous dilution by gastrointestinal secretions. Gastrointestinal transfer was explored in the fasted and fed states. In addition, the impact of the transit-stimulating agent MotiliumÒ (domperidone) and the transit-inhibiting agent ImodiumÒ (loperamide HCl) was assessed. In a next step, the generated human data set was used as reference to judge the implementation of gastrointestinal transfer in two existing in vitro tools (TNO Intestinal Model-1 (Blanquet et al., 2004) and a simpler three-compartment in vitro model (Psachoulias et al., 2012)) and an in silico PBPK model (SimcypÒ) (Jamei et al., 2009; Rostami-Hodjegan, 2012).

2. Materials and methods 2.1. Chemicals Paromomycin sulfate, glycine, orlistat and 9-fluorenyl methoxycarbonyl chloride (FMOC-CL) were obtained from Sigma Aldrich (St. Louis, MO, USA). Acetonitrile and NaCl were purchased from Fisher Scientific (Leicestershire, UK), while NaH2PO4H2O was received from Acros Organics (Geel, Belgium). Merck (Overijse, Belgium) supplied boric acid. BDH Laboratory Supplies (Poole, UK) provided HCl and NaOH. KOH was obtained from Riedel-De Haën (Seelze–Hannover, Germany). Sodium acetate and acetic acid were supplied by VWR (Leuven, Belgium). Simulated intestinal fluid (SIF) powder was purchased from Biorelevant (Croydon, UK). Lonza (Verviers, Belgium) was the supplier of Hanks’ balanced salt solution (HBSS), Dulbecco’s modified Eagle’s medium, penicillin–streptomycin (10,000 IU/ml), nonessential amino acid medium (100), trypsin–EDTA solution, fetal bovine serum, and HEPES. Purified water was obtained by using a Maxima system (Elga Ltd., High Wycombe Bucks, UK). For the Caco-2 experiments, the cell culture medium consisted of Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum, 1% nonessential amino acid, and 100 IU/ml penicillin–streptomycin. Transport medium consisted of HBSS containing 25 mM glucose and 10 mM HEPES (pH 7.4). Fasted state simulated intestinal fluid (FaSSIF), including sodium taurocholate (3 mM) and lecithin (0.75 mM) was made by dissolving 2.24 mg SIF powder per milliliter FaSSIF buffer (pH 6.5), which contains NaOH (10.5 mM), Na2HPO4 (28 mM) and NaCl (106 mM). 2.2. Paromomycin: luminal solubility & intestinal permeability The solubility of paromomycin was determined by incubating an excess of powder (5 mg paromomycin sulfate) to 0.5 ml of human gastric and intestinal fluids for 28 h at 37 °C in a prewarmed shaking incubator (175 rpm) (KS4000i incubator, Ika, Staufen, Germany). Samples were centrifuged for 10 min at 20.817 g and the supernatant was analyzed for paromomycin concentration (see Section 2.6). To confirm the non-absorbable properties of paromomycin, a Caco-2 experiment was set up. Caco-2 cells were cultured as described previously (Bevernage et al., 2012). On the day of the experiment the cell culture medium was refreshed one hour before the experiment. After rinsing the cells twice, the transepithelial electrical resistance (TEER) was measured after 25 min. Only monolayers with TEER values higher than 300 X  cm2 were used. After aspirating the apical and basolateral media, 1.5 ml of transport medium (pH 7.4) containing 0.2% D-atocopheryl polyethylene glycol 1000 succinate (to create sink conditions) was added to the basolateral compartment; the apical compartment consisted of 0.5 ml of transport medium (pH 7.4) or FaSSIF (pH 6.5) containing paromomycin at different concentrations (1400 lM in transport medium, 200 and 50 lM in FaSSIF). Samples were taken from the donor and acceptor compartment after 60 min and analyzed for paromomycin. After completion of the experiment, TEER values were measured to check the final integrity of the monolayer. 2.3. In vivo study To evaluate the gastrointestinal transfer in vivo, gastric and duodenal concentrations of paromomycin were monitored in five healthy volunteers (two men, three women; aged between 22 and 24 years). Exclusion criteria included gastrointestinal disorders, HIV, hepatitis B or C infection, use of medication, pregnancy and frequent X-ray exposure. The procedure followed the tenets

B. Hens et al. / European Journal of Pharmaceutical Sciences 63 (2014) 233–242

of the Declaration of Helsinki and was approved by the Committee of Medical Ethics of the University Hospitals Leuven, Belgium (ML9149) and the Federal Agency for Medicines and Health Products (530610). All volunteers provided written informed consent to participate in this study. After a fasting period of 12 h, two double-lumen polyvinyl catheters [Salem Sump Tube 14 Ch (external diameter 4.7 mm), Covidien, Dublin, Ireland] were introduced via the mouth or nose and positioned into (i) the duodenum (D2/D3) and (ii) the stomach (antrum; in front of the pylorus); positioning was checked by fluoroscopy. Gastrointestinal transfer was evaluated in four different conditions: (1) fasted state, (2) fed state, (3) fed state in the presence of the transit-stimulating agent domperidone and (4) fed state in the presence of the transit-inhibiting agent loperamide. In all conditions, an immediate release tablet of 250 mg paromomycin (GabbroralÒ, Pfizer S.r.l., Ascoli Piceno, Italy) was dissolved in 250 ml of tap water (resulting in a concentration of 1624 lM paromomycin) and given orally to the volunteers. To simulate the fed state, 400 ml of Ensure PlusÒ nutrient shake (Abbott Laboratories B.V., Zwolle, The Netherlands) was administered twenty minutes prior to the paromomycin solution. The energy content of this nutrient drink is 2526 kJ/600 kcal (lipids 29%, carbohydrates 54% and proteins 17%), the pH is 6.6 and the osmolality is 670 mOsm/kg. Finally, the influence of transit-modifying agents was investigated by administering 2 tablets of loperamide (2 mg/tablet, ImodiumÒ, Janssen-Cilag SpA, Latina LT, Italy) or 2 tablets of domperidone (10 mg/tablet, Motilium Ò, Janssen-Cilag S.A., Val De Reuil, France) twenty minutes prior to the liquid meal (i.e., forty minutes prior to paromomycin intake). Due to their formulation design (orodispersible tablets), MotiliumÒ and ImodiumÒ show a maximal effect in less than one hour, according to their summary of product characteristics (SmPC). During the experiment, volunteers were sitting in an upright position. Samples were taken every ten minutes in the first hour, and every fifteen minutes up to two hours (fasted state) or four hours (fed state). The sampling volume was kept as small as possible (<3 ml). The pH of the gastrointestinal fluids was measured immediately after aspiration (Hamilton Knick PortamessÒ, Bonaduz, Switzerland). Aspirated fluids were stored on ice prior to analysis on the day of the experiment. Collected fed state duodenal fluids were treated with a solution of orlistat in ethanol, to inhibit lipase activity and prevent lipolysis (final concentration of 1 lM orlistat in the test tube). The total amount of paromomycin that was removed by aspiration per experiment ranged from 0.95% to 9.33% of the administered dose.

235

the GET and the initial stomach volume in the fed state, all other physiological parameters of the gastrointestinal tract, were kept at the standard HV values in Simcyp. This means that the default settings for duodenal fluid transit time, secretion rates (pancreatic juice, etc.) and water absorption rate constants were used and assumed to be correct. Determination of the best fit was based on the method of least squares. Table 3 gives an overview of the various physiological variables, the trial design and the physicochemical properties of paromomycin applied for the simulations in the fasted and fed states in SimcypÒ. Data are presented as mean ± standard error of the mean (S.E.M.) for five subjects. Pharmacokinetic parameters (Cmax, Tmax and AUC) were compared using an ANOVA test combined with a Bonferroni’s multiple comparison test; differences were considered statistically significant at p < 0.05. 2.5. Evaluation of gastrointestinal transfer and ongoing secretions as implemented in in vitro and in silico tools The simulation of fasted state gastrointestinal transfer, as implemented in in vitro tools that have been described in the literature, and in a PBPK modeling platform (SimcypÒ Simulator), was validated based on the intraluminal reference data obtained from the in vivo study. 2.5.1. TNO intestinal model (TIM-1) The gastric-small intestinal model TIM-1 (TNO, Zeist, The Netherlands (Blanquet et al., 2004)) was tested to evaluate the gastrointestinal transfer of paromomycin in fasting conditions after addition of the paromomycin solution (1 tablet in 250 ml of water) to the stomach compartment (Fig. 1). Two fasted state experimental setups characterized by different gastric emptying times and modified duodenal secretions were explored (Table 1). The first experiment corresponded to the standard TIM-1 process parameters as described by Brouwers et al. (2011). Because of the fast gastrointestinal transfer derived from the in vivo study, an additional experiment was performed to accommodate this situation. In experiment 2, the gastric emptying

2.4. Treatment and statistical analysis of in vivo data To estimate the gastric emptying time observed for each volunteer in the in vivo study, parameter estimation was performed for the mean gastric emptying time (GET) in the SimcypÒ Simulator; emptying from the gastric compartment is treated as a first order process. Based on this first order process of gastric emptying, the gastric half-life (time needed to remove half of stomach content; t1/2,G) can be calculated by the following equation:

t 1=2;G ¼ GET  ln 2

ð1Þ

GET was varied (from 6 to 24 min in the fasted state and from 7.5 to 60 min in the fed state) for a ‘‘representative individual of a healthy volunteer (HV) population’’, to find the best fit of the simulated duodenal concentration–time profiles to the observed profiles for all five volunteers in the four different conditions. For simulations in the fed conditions, the initial stomach volume was reduced from 1000 ml to 650 ml (based on the administered volume of nutrient drink and water in the in vivo study). Apart from

Fig. 1. TIM-1 system. A. stomach compartment; B. pyloric sphincter; C. duodenum compartment; D. peristaltic valve; E. jejunum compartment; F. peristaltic valve; G. ileum compartment; H. ileo-caecal sphincter; I. stomach electrolytes, HCl and enzyme secretion; J. bile, pancreatic juice and bicarbonate secretion; K. jejunum/ ileum electrolyte secretion; L. pre-filter; M. hollow fiber membrane; N. filtrate pump; P. pH electrodes; Q. level sensor; R. temperature sensor; S. pressure sensor. In the present experiments, a hollow fiber membrane, to simulate absorption, was excluded. Only compartments A, B and C were of interest in the present study.

236

B. Hens et al. / European Journal of Pharmaceutical Sciences 63 (2014) 233–242

Table 1 Process parameters of TIM-1 for simulating the behavior of an aqueous solution of 250 mg paromomycin in the upper gastrointestinal tract in fasting conditions. Experiment 1 refers to experimental conditions described by Brouwers et al. (2011). Experiment 2 refers to the adapted experimental parameters applied for better fitting. Experiment 1: fasted state standard conditions

Experiment 2: fasted state adapted conditions

Gastric compartment Dosing Water GabbroralÒ Start residue HPMC + bile Amylase Flux fluid secretions

230 ml 1 tablet 10 ml 10 ml 2 mg 1 ml/min

230 ml 1 tablet 10 ml 10 ml 2 mg 1 ml/min

pH profile Time (min) 0 10 30 60 240

pH 3 2.2 1.8 1.7 1.7

pH 3 2.2 1.8 1.7 1.7

20 min

10 min

Gastric emptyinga Gastric half-life of emptying to duodenum Gastric emptying time

0.47 h

0.23 h

Duodenal compartmentb Volume Flux fluid secretions

55 ml 1 ml/min

55 ml 3 ml/min

pH profile Time (min) 0 10 30 60 240

pH 6.3 6.3 6.3 6.3 6.3

pH 6.3 6.3 6.3 6.3 6.3

Fig. 2. Schematic representation of the three-compartmental in vitro model described by Psachoulias et al. (2012). Gastric content flows with a rate F1 from the gastric to the duodenal compartment. Contents of duodenal compartment are removed with a flow rate F (ml/min). The volume in the duodenal compartment is kept constant by a continuous infusion of concentrated intestinal fluid from the reservoir compartment with a rate F2 (F1 + F2 = F) (adapted from Psachoulias et al. (2012)).

emptying. Recently, Symillides et al. (2013) optimized the setup to a continuous model. Since the design of the clinical study (administration of a solution) and the solubility of paromomycin ensured that no dissolution or precipitation will take place in the GI tract, intraluminal paromomycin concentrations could be predicted by simulation using the mathematical equations (see below) underlying the fluid transfer behavior of the stepwise and the continuous model.

a Transfer from content in the stomach to the duodenum was regulated by opening or closing peristaltic valves that connect the two compartments. This  b



t

t

implemented transfer can be described by the formula: f ¼ 1  2 1=2 where f is the fraction of content delivered to the duodenum, t is the time of delivery, t1/2 is the half-life of delivery to the duodenum and b is the factor that describes the shape of the curve (Elashoff et al., 1982; Minekus et al., 1999). b Duodenal volume and secretion remain constant during the experiments.

time and duodenal secretions were modified to improve the duodenal concentration–time profile that was observed in the in vivo study presented in this manuscript. In both experiments, only the gastric and duodenal compartments were investigated (Table 1). Luminal samples of the gastric and duodenal compartment were taken every ten minutes for the first hour and every fifteen minutes for the next hour. Samples were stored at 26 °C prior to analysis. Before the start of each experiment, the entire system was washed with detergent and rinsed with water. Experiments were done in triplicate and results are presented as mean ± S.E.M. 2.5.2. Three-compartment in vitro model Psachoulias et al. (2012) developed a three-compartment in vitro model to simulate fasted state gastrointestinal transfer and secretions in the upper small intestine (Fig. 2). As a function of time, gastric contents are transported to the duodenal compartment with a rate F1, while duodenal contents are removed to the waste with a rate F. The fluid volume in the duodenal compartment is kept constant by simultaneous infusion with concentrated simulated intestinal fluid from a reservoir compartment (rate F2 = F  F1). This in vitro tool has previously been used as a stepwise (=non-continuous) model consisting of 5 individual steps with different flow rates to simulate first-order gastric

2.5.2.1. Stepwise approach (Psachoulias et al., 2012). The stepwise methodology consists of 5 individual steps with a duration of 15 min and a decreasing flow rate F, F1 and F2 for each experiment (Table 2). An overall duodenal concentration–time profile is obtained by combining the mid-time concentrations (7.5 min) of the 5 individual experiments. Starting from a gastric volume of 250 ml, this setup simulates a first-order gastric emptying with a gastric half-life of approximately 15 min. For each experiment, duodenal paromomycin concentrations were simulated at the mid-point (7.5 min) using the following equation:

CðtÞ ¼

  5F t 4 Q0  1 1  e 4V I 5 V G0

ð2Þ

where C stands for the drug concentration (lM) in the intestinal compartment as a function of time (t); Q0 is the dose of drug added to the gastric compartment in the beginning of the experiment (406 lmol = 250 mg of paromomycin); F1 stands for the flow rate (ml/min) from gastric to duodenal compartment for each individual experiment; VG0 is the initial fluid volume in the gastric compartment (250 ml) and VI is the constant volume in the intestinal compartment (60 ml). 2.5.2.2. Continuous approach (Symillides et al., 2013). Recently, the stepwise model has been converted into a continuous model. From a practical point of view, this setup carries out an easier way of handling and is less labor-intensive. During the course of the experiment, the flow rate from gastric to duodenal compartment, F1, reduces as follows:

F 1 ¼ FekG t

ð3Þ

where kG is the first order gastric emptying rate constant, fixed at 0.046 min1 which corresponds to a half-life of gastric emptying (t1/2,G) of 15 min. The flow rate from duodenum to waste, F, is maintained constant during the course of the experiment and can be calculated by:

237

B. Hens et al. / European Journal of Pharmaceutical Sciences 63 (2014) 233–242 Table 2 Flow rates of the three-compartment in vitro model (stepwise approach) (Psachoulias et al., 2012). Stepwise approach: individual steps

Time interval after initiation of emptying (min)

In vitro Gastric to duodenal compartment (F1 ml/min)

Reservoir to duodenal compartment (F2 ml/min)

Duodenal compartment to waste (F ml/min)

Step Step Step Step Step

0–15 15–30 30–45 45–60 60–75

8 4 2 1 0.5

2 1 0.5 0.25 0.125

10 5 2.5 1.25 0.625

1 2 3 4 5

F ¼ V G0 kG ¼ V G0

lnð2Þ t1=2;G

ð4Þ

where VG0 stands for the initial gastric volume of 250 ml. The volume in the duodenal compartment is maintained at 40 ml for the continuous approach (instead of 60 ml in the stepwise approach). Duodenal paromomycin concentrations were simulated using the following underlying equation:

1 kG CðtÞ ¼ Q 0 V I kG  F=V I



e

  F VI

t

! kG t

e

ð5Þ

where C is the concentration of paromomycin that appears in the duodenal compartment as a function of time (t), Q0 stands for the dose of drug added to the gastric compartment in the beginning of the experiment (406 lmol = 250 mg of paromomycin), VI is the duodenal volume (40 ml), kG is the first order gastric emptying rate constant and F is the flow rate from duodenal compartment to the waste, expressed in ml/min. The stepwise and the continuous approach were evaluated by comparing the corresponding simulated concentration profiles in the duodenal compartment (Eqs. (2) and (5), respectively) with the actual concentrations of paromomycin measured in the duodenum, achieved from the in vivo study. 2.5.3. PBPK modeling platform SimcypÒ Simulator The predictive value of the fasted state gastrointestinal transfer as implemented in the commercially available PBPK modeling platform SimcypÒ (SimcypÒ Simulator version 13 (Release 1), Sheffield, UK) was judged based on the observed in vivo data for paromomycin (Fig. 3). The Simcyp transfer model considers baseline fluid volumes in the stomach, each of seven small intestinal compartments and a single large intestinal compartment. It also considers fluid secretion rates into these compartments viz. saliva gastric juices, pancreatic, Brunners gland, bile and mucosal secretions in the

relevant regions and the reabsorption rates of fluid from these compartments. Fluid taken with dose is added to the stomach compartment at dosing time which then transits and/or is reabsorbed in a time-dependent manner until the regional basal steady state volumes are re-attained (unless of course simulation times are very short). In SimcypÒ, the clinical study was simulated by administering paromomycin (a solution of 250 mg in 250 ml water) to a virtual population of 5 healthy volunteers (HVs) (two men and three women, all aged between 22 and 24 years) using standard fasting state conditions (Table 3). Subsequently, the simulated duodenal concentration–time profiles were compared to the in vivo data. 2.6. Analysis of paromomycin in simulated and human gastrointestinal fluids The determination of paromomycin in human (in vivo study) and simulated (TIM-1 study) gastrointestinal fluids was based on a

Table 3 Physiological parameters, compound input variables and trial design for simulating the fasted and fed state behavior of paromomycin in SimcypÒ. Fasted state

Fed state

Gastrointestinal attributes Mean gastric emptying time (h) CV gastric emptying time (%) Initial volume of stomach fluid (ml) Initial volume of stomach fluid CV (%)

0.4 38 50 30

1 38 650 30

Physicochemical characteristics Molecular weight (g/mol) Log Po:w Compound type pKa 1

615.63 8.31 Monoprotic base 12.9

Absorption Absorption model

Gut wall permeability (from Caco-2 experiments) (106 cm/s) Type of formulation Trial design Populationa No. of trialsb No. of subjects in trialb Minimum age (years)b Maximum age (years)b Proportion of femalesb Single dose (mg) Fluid intake with dose (ml) Fig. 3. The Simcyp fluid dynamics model – detail for the stomach and duodenal compartments only where Vj, Rsec,j, KRe-abs,j, and Kt,j respectively refer to the luminal fluid volume, fluid secretion rate (zero order), fluid re-absorption rate constant and transit rate constant for the gut segment. Note that (i) the fluid volume in the intestine is generally much smaller than the cylindrical volume, (ii) the fluid volume taken with dose (default 250 ml) can be adjusted according to the trial design and (iii) all these values have associated inter-individual variability within the SimcypÒ simulator.

Advanced dissolution, Absorption & Metabolism (ADAM) 0 Solution Healthy volunteer 1 5 22 24 0.6 250 250*

a Virtual population was applied for validation of the standard, implemented gastric emptying time in SimcypÒ. Population representative was used for determining the gastric emptying time for each individual volunteer in a fasted state, fed state, fed state in the presence of MotiliumÒ and fed state in the presence of ImodiumÒ. b Only applicable for simulations with a virtual population. * Fixed for all subjects.

B. Hens et al. / European Journal of Pharmaceutical Sciences 63 (2014) 233–242

3. Results and discussion 3.1. Paromomycin: luminal solubility & intestinal permeability characteristics To ensure that the intraluminal concentration–time profiles of paromomycin in healthy volunteers only reflect gastrointestinal transfer and dilution, in vitro experiments were performed to confirm adequate solubility and permeability characteristics. The solubility of paromomycin in human gastric and intestinal fluids exceeded 14 mM, indicating no risk for precipitation from the administered oral solution (1624 lM). Furthermore, the summary of product characteristics (SmPC) indicates that paromomycin will not be absorbed after oral intake of the immediate release tablet GabbroralÒ. This was confirmed in a dual chamber setup, where no transport of paromomycin across Caco-2 monolayers was observed within a one-hour incubation (basal concentrations below limit of detection, transport <1%; data not shown). Paromomycin was completely recovered in the apical compartment. 3.2. In vivo studies 3.2.1. Fasted state Fig. 4 depicts the mean concentration–time profiles of paromomycin in the antrum and duodenum of five volunteers in fasting conditions. Concentrations were monitored for two hours after oral intake of a 250 ml solution of paromomycin (1624 lM). For the duodenal profiles, a Tmax of less than 30 min was observed for each volunteer. Even after 5 min, very high concentrations of paromomycin were detected in the duodenum (mean concentration 708 lM). Compared to the paromomycin concentration in the administered aqueous solution (1624 lM), maximal duode-

1400 1200

pH

derivatization of paromomycin with 9-fluorenyl methoxycarbonyl chloride (FMOC-CL) (Khan and Kumar, 2011), followed by HPLCFLUO analysis. After centrifugation (20,817 g, 15 min, 4 °C) of the aspirated gastrointestinal fluids, 20 lL of the supernatant was added to 430 lL of borate buffer (0.4 M, adjusted to pH 8 with 50% w/v potassium hydroxide solution). Upon addition of 500 lL of a FMOC-CL solution in acetonitrile (4 mM), the mixture was incubated in the dark and repeatedly vortexed to allow for paromomycin derivatization. After 10 min, the reaction was stopped by adding 50 lL of glycine solution in borate buffer (0.1 M). Potential protein in the samples was separated by centrifuging the samples (14,000 rpm, 5 min, 37 °C). Subsequently, 10 lL of supernatant was injected into the HPLC system consisting of an Alliance 2695 separations module and a Novapak C-18 column under radial compression (Waters, Milford, MA, USA); the derivatization product of paromomycin was detected by fluorescence (Waters 2475 Multiwavelength Fluorescence Detector) at an excitation wavelength of 260 nm and an emission wavelength of 315 nm. An isocratic run with acetonitrile:water (87:13 v/v) was performed with a flow rate of 1 ml/min to generate a retention time of 7.8 min. After 10 min, the column was rinsed during 1 min with methanol: 25 mM acetic acid buffer pH 3.5 (75:25 v/v), followed by 2 min with water: 25 mM acetic acid buffer pH 3.5 (75:25 v/v) and subsequently re-equilibrated with the mobile phase for 2 min. Calibration curves were made based on a stock solution of paromomycin in water (1400 lM). Linearity was observed between 1300 lM and 10 lM. The accuracy and precision errors were less than 9% and 6%, respectively, for a concentration of 50 lM in human intestinal fluids collected in the fasted state (FaHIF). Quality control samples of 700 lM paromomycin, which were analyzed together with the samples of the in vivo study, resulted in a relative standard deviation of less than 5%.

paromomycin (µM)

238

1000 800

9 8 7 6 5 4 3 2 1 0

0

0.5

1

Time (h)

1.5

2

600 400 200 0

0

0.5

1

1.5

2

Time (h) Fig. 4. Mean paromomycin concentration–time profiles and pH profiles in the antrum (N) and duodenum (j) of the five volunteers in the fasted state (mean ± S.E.M.).

nal concentrations suggest a dilution of less than two-fold. These observations indicate a very fast transfer of the administered solution from the stomach to the upper small intestine. Based on the duodenal concentration–time profiles, the gastric half-life of emptying, estimated using PBPK modeling (SimcypÒ Simulator), ranged from 4 to 13 min, with a mean gastric half-life of 5.2 min. It should be noted that paromomycin concentrations observed in the gastrointestinal fluids are the result of different simultaneously ongoing physiological processes including gastrointestinal secretions, water absorption and small intestinal transit. By making use of the PBPK modeling program SimcypÒ Simulator to estimate the gastric emptying time, these processes are taken into account. The concentration–time profiles for paromomycin in the antrum were highly variable among the five volunteers (Fig. 4 and Table 4). Remarkably, for four out of five volunteers, the area under the curve (AUC) of the concentration–time profile of the antrum was smaller than the AUC of the concentration–time profile of the duodenum. The observation of gastric concentrations lower than expected based on the administered solution (1624 lM) might be attributed to the fast transfer of the solution from stomach to small intestine, in combination with the limited mixing and, consequently, inhomogeneity of gastric contents in the fasted state.

3.2.2. Fed state In a next step, the influence of food on gastrointestinal transfer was investigated. Paromomycin concentrations in the antrum and duodenum were monitored for four hours. The mean profiles of the five volunteers are shown in Fig. 5. Compared to the fasted state, lower concentrations of paromomycin were observed in the duodenum. Based on the duodenal concentration–time profiles, the half-life of gastric emptying was significantly increased from 5.2 min in the fasted state to 33 min in the fed state (p < 0.05). Half-lifes of gastric emptying for each individual varied between 21 min and 40 min. Also the mean Tmax differed between the fasted (15 min) and fed (39 min) states (Table 5). These observations are in agreement with the digestive response to the intake of a liquid meal resulting in a slower transpyloric flow of the drug solution. The intake of food stimulates the enteric, central and autonomic nervous systems (Camilleri, 2006; Hellström et al., 2006). The resulting antral contraction waves lead to (i) prolonged residence time and (ii) better mixing and thus a more homogenous distribution of the gastric contents. In the present study this is reflected in (i) the increased gastric half-life for paromomycin and (ii) higher

239

B. Hens et al. / European Journal of Pharmaceutical Sciences 63 (2014) 233–242

Table 4 Descriptive parameters of the concentration–time profiles of paromomycin in the antrum and duodenum for the five healthy volunteers (HV) in the fasted state. The last column describes the parameters of the mean concentration–time profile for the five volunteers in the fasted state. Fasted state

HV 1

HV 2

HV 3

HV 4

HV 5

Mean concentration–time profile

Gastric half-life (min) AUCAntrum (lmol h/l) AUCDuodenum (lmol h/l) Cmax Antrum (lM) Tmax Antrum (min) Cmax Duodenum (lM) Tmax Duodenum (min)

5.2 452 623 549 25 739 5

4.2 337 222 493 25 224 25

12.5 784 868 910 15 1064 15

4.2 180 496 131 75 676 15

6.2 322 522 176 45 587 15

5.2 433 585 585 25 1047 15

1000

8 7 6 5

pH

800

4

paromomycin (µM)

3 2 1

600

0

0

0.5

1

1.5

2

Time (h)

2.5

3

3.5

4

400

200

0

0

05 0.5

1

15 1.5

2

2 2.5 5

3

3 3.5 5

4

Time (h) Fig. 5. Mean paromomycin concentration–time profiles and pH profiles in the antrum (N) and duodenum (j) of five volunteers in the fed state (mean ± S.E.M.).

(and less variable) antral paromomycin concentrations. In the individual duodenal concentration–time profiles (data not shown) several separated peaks can be observed. This phenomenon might be due to an additional flow of solution by the ‘‘Magenstrasse’’ (stomach road) described by Pal et al. (2007) as an alternative route for liquids to travel directly from the fundus to the upper small intestine, independent of antral contraction waves (ACW). In 2002, this phenomenon was studied by Hausken et al., by giving 300 ml of a low-calorie, liquid meat soup to eight healthy volunteers and, subsequently, monitoring the transpyloric flow from stomach to duodenum. In the absence of peristaltic antral activity, a considerable transpyloric flow was observed during gastric emptying (Hausken et al., 2002). 3.2.3. Fed state: co-administration with MotiliumÒ To evaluate the influence of a transit-stimulating agent on the gastrointestinal transfer of dissolved drugs, paromomycin was co-administered with the dopamine D2-receptor antagonist domperidone (MotiliumÒ) in postprandial conditions. In all five volunteers, domperidone co-administration resulted in an increased duodenal Cmax for paromomycin (average of 291 lM in the fed state versus an average of 319 lM in the fed state co-administered with 20 mg domperidone). This may be the

result of enhanced antral-duodenal contractions and more pronounced peristalsis across the pylorus that are induced by domperidone (Reddymasu et al., 2007). Despite the enhanced duodenal Cmax, domperidone did not significantly affect the mean gastric emptying time in postprandial conditions (estimated mean gastric half-life of 33 min in both conditions). The lack of effect on gastric emptying might be attributed to the calorific value of the test meal that was administered (600 kcal). Indeed, several studies investigated the gastrointestinal prokinetic effect of the dopamine D2-receptor antagonist at different doses with administration of liquid or solid meals (Broekaert, 1979; De Schepper et al., 1978; Gounaris et al., 2010; Markey and Shafat, 2012; Nagahata et al., 1995; Reddymasu et al., 2007; Tatsuta et al., 1989; Valenzuela and Dooley, 1984). While the gastric emptying of paracetamol, co-administered with a liquid meal of 200 kcal, increased significantly in the presence of domperidone (Tatsuta et al., 1989), more recent studies could not prove the pro-gastric effect of domperidone for high-fat test meals (Parker and Chapman, 2004). 3.2.4. Fed state: co-administration with ImodiumÒ In a final condition, the influence of the transit-inhibiting agent loperamide HCl (ImodiumÒ) on the fed state gastrointestinal transfer of paromomycin was evaluated. The rationale for using loperamide was based on its inhibiting effect on gastrointestinal motility due to interaction with opioid receptors along the intestinal tract. However, no significant effects of loperamide were observed on the parameters describing the concentration–time profiles in the antrum and duodenum (data not shown). Literature review confirms that administering loperamide HCl at different doses to healthy volunteers does not decrease gastric emptying time (Kirby et al., 1989; van Wyk et al., 1992). The pharmacodynamic activity for loperamide HCl is mainly located at the level of the colon where it acts as an opiate agonist for the l-receptors (Ooms et al., 1984) to reduce the peristaltic movements of the colon. Since the site of action is mainly the colon, the influence on the transition from stomach to small intestine will be minimal. 3.3. Evaluation of gastrointestinal transfer and ongoing secretions as implemented in in vitro and in silico tools In addition to the in vivo assessment in healthy volunteers, the present study aimed to judge the predictive value of gastrointesti-

Table 5 Descriptive parameters of the concentration–time profiles of paromomycin in the antrum and duodenum for the five healthy volunteers (HV) in the fed state. The last column describes the parameters of the mean concentration–time profile for the five volunteers in the fed state. Fed state

HV 1

HV 2

HV 3

HV 4

HV 5

Mean concentration–time profile

Gastric half-life (min) AUCAntrum (lmol h/l) AUCDuodenum (lmol h/l) Cmax Antrum (lM) Tmax Antrum (min) Cmax Duodenum (lM) Tmax Duodenum (min)

39.5 1084 312 928 15 185 35

29 1193 381 691 35 319 55

22.9 397 436 318 25 315 25

25 1226 339 670 15 254 55

20.8 532 261 437 15 291 25

33.2 1127 461 743 25 291 55

240

B. Hens et al. / European Journal of Pharmaceutical Sciences 63 (2014) 233–242

nal transfer and ongoing dilution, as implemented in current in vitro tools. An adequate implementation of gastrointestinal transfer is crucial to simulate biorelevant intraluminal concentrations and strengthen the predictive power of these tools in drug and formulation development. For this purpose, two in vitro models were selected based on their ability to simulate duodenal concentration–time profiles. This involves simulating (i) gastrointestinal transfer, (ii) drug dilution by duodenal secretions, and (iii) drug removal from the duodenum by either transit or absorption. Also, a commercially available in silico tool was tested. 3.3.1. TNO intestinal model (TIM-1) The setup of the in vivo study in the fasted state was simulated in the gastric-small intestinal model (TIM-1). Two experiments were performed, both in triplicate. The first experiment implemented the standard TIM-1 conditions in fasting conditions (described by Brouwers et al. (2011)). A 250 ml aqueous, non-caloric solution of 1 tablet of paromomycin was introduced into the stomach compartment and the run started for two hours (Fig. 6A). The emptying halftime from the stomach was programmed at 20 min. The obtained duodenal concentration–time curve for TIM-1, resulting from a programmed emptying halftime of 20 min, approaches the SimcypÒ predicted duodenal concentration–time curve with a corresponding gastric halftime of 22 min. Compared to the mean gastric half-life of emptying observed in vivo (5.2 min) this indicates that TIM-1 underestimates the gastric emptying rate of drugs in solution. Taking these observations into account, a novel TIM-1 simulation was performed, employing adjusted parameters (Table 1). Decreasing the gastric emptying halftime (from 20 min to 10 min) while increasing duodenal secretion (from 1 ml/min to 3 ml/min), resulted in a reasonably good fit to the duodenal concentration–time profile as seen in humans (Fig. 6B). As the current standard configuration deploys a certain fixed inner volume in this compartment, the increased rate 1800 1600 1400 1200 1000 800 600 400 200 0

0

0.5 05

1

1.5 15

2

Time (h)

paromomycin (µM)

B

3.3.2. Three-compartment in vitro model The three-compartmental in vitro model, initially described by Psachoulias et al. (2012) to explore the intestinal drug behavior of highly permeable, lipophilic weak bases, was evaluated for the prediction of duodenal paromomycin concentrations in the fasted state (starting from a 250 ml solution of 1 tablet GabbroralÒ). Duodenal concentration–time profiles were simulated using both the original (stepwise) and the optimized (continuous) model. In Fig. 7, the simulated duodenal concentration–time profiles are compared with the mean duodenal concentration–time profile of the five volunteers in fasting conditions. Both simulations fit reasonably well with the in vivo profile, both in terms of Cmax and the subsequent decline in duodenal concentrations. This suggests that the model adequately reflects the gastrointestinal transfer and subsequent dilution of paromomycin. It should be noted that data from the three-compartmental model have been successfully related to the in vivo data of 4 lipophilic (highly permeable) weak bases (Psachoulias et al., 2012). Based on data from this study (Fig. 7), it seems that the continuous approach can also be useful for the study of low permeability compounds. 3.3.3. PBPK modeling platform Simcyp simulator Systems-based PBPK modeling is a useful approach to predict both the mean and population variability of the in vivo behavior of a drug; this is done by integrating physicochemical and in vitro data on drug and formulation into a physiological model of the human body (Jamei et al., 2009; Rostami-Hodjegan, 2012). The physiological information is adapted where known according to the ethnic, disease, pediatric etc. population of interest. Implementation of correct physiological parameters is critical for the predictive power of PBPK modeling. Therefore, the in vivo data obtained in the present study was used to judge the simulation of gastrointestinal transfer of drugs in solution in fasting conditions, based upon parameters for a North European Caucasian Healthy Volunteer population as implemented in the SimcypÒ Simulator. In a virtual population of 5 healthy volunteers, the intraluminal behavior of paromomycin following oral intake was simulated using standard fasting state parameters in SimcypÒ (Table 3).

1600 1400

1400

1200

1200

1000 800 600 400 200 0

0

0.5

1

1.5

2

Time (h) Fig. 6. A: TIM-1 fasted state experiment 1: Mean paromomycin concentration– time profiles in the stomach (N) and duodenum (d) compartment of TIM-1 compared with the mean in vivo concentration–time profile in the duodenum (j) (n = 3; mean ± S.E.M.). The arrow indicates the difference between the gastric halflife of emptying for the TIM-1 (20 min) and the in vivo data (5.2 min). B: TIM-1 fasted state experiment 2: Mean paromomycin concentration–time profiles of the stomach (N) and duodenum (d) compartment of TIM-1 compared with the mean in vivo concentration–time profile in the duodenum (j) (n = 3; mean ± S.E.M.).

Paromomycin (µM)

paromomycin (µM)

A

of duodenal secretions (applied in experiment 2) was required to reflect the in vivo duodenal residence time. These experiments demonstrate how human reference studies may guide the optimization of TIM-1 parameters to improve the biorelevance and predictive power of the system.

1000 800 600 400 200 0

0

0.5

1

1.5

2

Time (h) Fig. 7. Simulated duodenal concentration–time profiles of paromomycin, using the three-compartmental model of Psachoulias et al. (2012), in both the stepwise approach (Eq. (2)) (j) and the continuous approach (Eq. (5)) (- - -). Reference data are given by the mean (±S.E.M.) in vivo duodenal concentration–time profile (N).

paromomycin (µM)

B. Hens et al. / European Journal of Pharmaceutical Sciences 63 (2014) 233–242

1400

Mean In Vivo

1200

GET 0.125 h

References

GET 0.4 h

1000 800 600 400 200 0

0

0.5 05

1

1.5 15

241

2

Time (h) Fig. 8. Simulated duodenal concentration–time profiles of paromomycin in a virtual population using SimcypÒ Simulator: comparison between the implemented gastric emptying time (GET) of 0.4 h (  ) versus the modified GET of 0.125 h (- - -), corresponding to the in vivo situation. For reference purposes, the mean in vivo duodenal concentration–time profile in the fasted state is also shown (mean ± S.E.M.) (j).

Comparing predicted and human data, it was clear that duodenal paromomycin concentrations cannot be adequately simulated employing the default mean GET in SimcypÒ (24 min; corresponding to a gastric half-life of emptying of 16.5 min) (Fig. 8). Reducing the mean GET of the virtual population to 7.5 min (corresponding to a gastric half-life of emptying of 5.2 min) significantly improved the prediction of duodenal paromomycin concentrations. Implementing a faster fluid gastric emptying rate in SimcypÒ may thus optimize simulations that involve the transfer of an aqueous, non-caloric drug solution. Obviously, in the case of immediate release tablets or capsules gastric residence time may be altered due to disintegration time or otherwise differ for a non-disintegrating monolithic controlled release or gastro-retentive formulation. In addition, pharmacodynamic effects of the drug itself may also alter GET. 4. Conclusion Complementary to imaging data, the concentration–time profiles and transfer rates for paromomycin provide unique information on the transfer of drugs in solution from the stomach to the duodenum, in both fasted and fed state: a fast transfer (estimated gastric half-life of emptying between 4 and 13 min) resulted in a dilution of less than 1:2 in fasting conditions. These results served as intraluminal reference data to validate, and possibly optimize, simulation of gastrointestinal transfer in existing in vitro models and in physiologically-based pharmacokinetic models. It was demonstrated that current in vitro and in silico tools tend to overestimate the gastric half-life for drugs in solution, possibly resulting in a poor estimation of in vivo intraluminal drug concentrations. Nevertheless, straightforward modification of model parameters allowed for better simulation of the in vivo situation. Acknowledgments This work has received support from (1) the Institute for the Promotion of Innovation through Science and Technology in Flanders (IWT) and (2) the Innovative Medicines Initiative Joint Undertaking (http://www.imi.europa.eu) under Grant Agreement No. 115369, resources of which are composed of financial contribution from the European Union’s Seventh Framework Program (FP7/ 2007–2013) and EFPIA companies’ in kind contribution. We also would like to thank Nancy Haesendonck, Marleen Hellemans and Kathy Van Herck (Gastroenterology, University Hospitals Leuven, Belgium) for their assistance during the in vivo studies.

Bevernage, J., Brouwers, J., Clarysse, S., Vertzoni, M., Tack, J., Annaert, P., Augustijns, P., 2010. Drug supersaturation in simulated and human intestinal fluids representing different nutritional states. J. Pharm. Sci. 99, 4525–4534. Bevernage, J., Forier, T., Brouwers, J., Tack, J., Annaert, P., Augustijns, P., 2011. Excipient-mediated supersaturation stabilization in human intestinal fluids. Mol. Pharm. 8, 564–570. Bevernage, J., Brouwers, J., Annaert, P., Augustijns, P., 2012. Drug precipitationpermeation interplay: supersaturation in an absorptive environment. Eur. J. Pharm. Biopharm. 82, 424–428. Blanquet, S., Zeijdner, E., Beyssac, E., Meunier, J.-P., Denis, S., Havenaar, R., Alric, M., 2004. A dynamic artificial gastrointestinal system for studying the behavior of orally administered drug dosage forms under various physiological conditions. Pharm. Res. 21, 585–591. Broekaert, A., 1979. Effect of domperidone on gastric emptying and secretion. Postgrad. Med. J. 55 (Suppl. 1), 11–14. Brouwers, J., Brewster, M.E., Augustijns, P., 2009. Supersaturating drug delivery systems: the answer to solubility-limited oral bioavailability? J. Pharm. Sci. 98, 2549–2572. Brouwers, J., Anneveld, B., Goudappel, G.-J., Duchateau, G., Annaert, P., Augustijns, P., Zeijdner, E., 2011. Food-dependent disintegration of immediate release fosamprenavir tablets: in vitro evaluation using magnetic resonance imaging and a dynamic gastrointestinal system. Eur. J. Pharm. Biopharm. 77, 313–319. Camilleri, M., 2006. Integrated upper gastrointestinal response to food intake. Gastroenterology 131, 640–658. Culen, M., Rezacova, A., Jampilek, J., Dohnal, J., 2013. Designing a dynamic dissolution method: a review of instrumental options and corresponding physiology of stomach and small intestine. J. Pharm. Sci. 102, 2995–3017. De Schepper, A., Wollaert, F., Reyntjens, A., 1978. Effects of oral domperidone on gastric emptying and motility. A double-blind comparison with placebo and metoclopramide. Arzneimittelforschung 28, 1196–1199. Elashoff, J.D., Reedy, T.J., Meyer, J.H., 1982. Analysis of gastric emptying data. Gastroenterology 83, 1306–1312. Gounaris, A., Costalos, C., Varchalama, E., Kokori, F., Grivea, I.N., Konstantinidi, K., Syrogiannopoulos, G.A., 2010. Gastric emptying of preterm neonates receiving domperidone. Neonatology 97, 56–60. Hausken, T., Mundt, M., Samsom, M., 2002. Low antroduodenal pressure gradients are responsible for gastric emptying of a low-caloric liquid meal in humans. Neurogastroenterol. Motil. 14, 97–105. Heading, R.C., Nimmo, J., Prescott, L.F., Tothill, P., 1973. The dependence of paracetamol absorption on the rate of gastric emptying. Br. J. Pharmacol. 47, 415–421. Hellström, P.M., Grybäck, P., Jacobsson, H., 2006. The physiology of gastric emptying. Best Pract. Res. Clin. Anaesthesiol. 20, 397–407. Jamei, M., Turner, D., Yang, J., Neuhoff, S., Polak, S., Rostami-Hodjegan, A., Tucker, G., 2009. Population-based mechanistic prediction of oral drug absorption. AAPS J. 11, 225–237. Khan, W., Kumar, N., 2011. Characterization, thermal stability studies, and analytical method development of paromomycin for formulation development. Drug Test. Anal. 3, 363–372. Kirby, M.G., Dukes, G.E., Heizer, W.D., Bryson, J.C., Powell, J.R., 1989. Effect of metoclopramide, bethanechol, and loperamide on gastric residence time, gastric emptying, and mouth-to-cecum transit time. Pharmacotherapy 9, 226–231. Kostewicz, E.S., Abrahamsson, B., Brewster, M., Brouwers, J., Butler, J., Carlert, S., Dickinson, P.A., Dressman, J., Holm, R., Klein, S., Mann, J., McAllister, M., Minekus, M., Muenster, U., Müllertz, A., Verwei, M., Vertzoni, M., Weitschies, W., Augustijns, P., 2014. In vitro models for the prediction of in vivo performance of oral dosage forms. Eur. J. Pharm. Sci. 57, 342–366. Madsen, J.L., 2013. Scintigraphic assessment of gastrointestinal motility: a brief review of techniques and data interpretation. Clin. Physiol. Funct. Imaging 34, 243–253. Markey, O., Shafat, A., 2012. Does domperidone, a D2-antagonist alter gastric emptying rates and appetite sensations in healthy adults with high-fat meal? A block-randomised, single-blind placebo-controlled study. Ir. J. Med. Sci. 181, 215–219. Minekus, M., Smeets-Peeters, M., Bernalier, A., Marol-Bonnin, S., Havenaar, R., Marteau, P., Alric, M., Fonty, G., Huis in’t Veld, J.H., 1999. A computer-controlled system to simulate conditions of the large intestine with peristaltic mixing, water absorption and absorption of fermentation products. Appl. Microbiol. Biotechnol. 53, 108–114. Nagahata, Y., Azumi, Y., Kawakita, N., Wada, T., Saitoh, Y., 1995. Inhibitory effect of dopamine on gastric motility in rats. Scand. J. Gastroenterol. 30, 880–885. Oberle, R.L., Chen, T.S., Lloyd, C., Barnett, J.L., Owyang, C., Meyer, J., Amidon, G.L., 1990. The influence of the interdigestive migrating myoelectric complex on the gastric emptying of liquids. Gastroenterology 99, 1275–1282. Ooms, L.A., Degryse, A.D., Janssen, P.A., 1984. Mechanisms of action of loperamide. Scand. J. Gastroenterol. Suppl. 96, 145–155. Pal, A., Brasseur, J.G., Abrahamsson, B., 2007. A stomach road or ‘‘Magenstrasse’’ for gastric emptying. J. Biomech. 40, 1202–1210. Parker, B.A., Chapman, I.M., 2004. Food intake and ageing – the role of the gut. Mech. Ageing Dev. 125, 859–866.

242

B. Hens et al. / European Journal of Pharmaceutical Sciences 63 (2014) 233–242

Psachoulias, D., Vertzoni, M., Butler, J., Busby, D., Symillides, M., Dressman, J., Reppas, C., 2012. An in vitro methodology for forecasting luminal concentrations and precipitation of highly permeable lipophilic weak bases in the fasted upper small intestine. Pharm. Res. 29, 3486–3498. Reddymasu, S.C., Soykan, I., McCallum, R.W., 2007. Domperidone: review of pharmacology and clinical applications in gastroenterology. Am. J. Gastroenterol. 102, 2036–2045. Rostami-Hodjegan, A., 2012. Physiologically based pharmacokinetics joined with in vitro-in vivo extrapolation of ADME: a marriage under the arch of systems pharmacology. Clin. Pharmacol. Ther. 92, 50–61. Schiller, C., Fröhlich, C.-P., Giessmann, T., Siegmund, W., Mönnikes, H., Hosten, N., Weitschies, W., 2005. Intestinal fluid volumes and transit of dosage forms as assessed by magnetic resonance imaging. Aliment. Pharmacol. Ther. 22, 971– 979. Steingoetter, A., Fox, M., Treier, R., Weishaupt, D., Marincek, B., Boesiger, P., Fried, M., Schwizer, W., 2006. Effects of posture on the physiology of gastric emptying: a magnetic resonance imaging study. Scand. J. Gastroenterol. 41, 1155–1164.

Symillides, M., Vertzoni, M., Reppas, C. Optimizing the in vitro testing methodology for the evaluation of drug transfer in the upper small intestine based on intraluminal data (poster). In: BBBB International Conference, Athens, 26–28 September 2013. Tatsuta, M., Iishi, H., Okuda, S., 1989. Gastric emptying and gastrointestinal symptoms in patients with atrophic gastritis and the effects of domperidone. Scand. J. Gastroenterol. 24, 251–256. Valenzuela, J.E., Dooley, C.P., 1984. Dopamine antagonists in the upper gastrointestinal tract. Scand. J. Gastroenterol. Suppl. 96, 127–136. Van Wyk, M., Sommers, D.K., Moncrieff, J., 1992. Influence of cisapride, metoclopramide and loperamide on gastric emptying of normal volunteers as measured by means of the area under the curve of the cumulative fraction absorbed-time profiles of paracetamol. Methods Find. Exp. Clin. Pharmacol. 14, 379–382. Weitschies, W., Kosch, O., Mönnikes, H., Trahms, L., 2005. Magnetic marker monitoring: an application of biomagnetic measurement instrumentation and principles for the determination of the gastrointestinal behavior of magnetically marked solid dosage forms. Adv. Drug Deliv. Rev. 57, 1210–1222.