Human and simulated intestinal fluids as solvent systems to explore food effects on intestinal solubility and permeability

European Journal of Pharmaceutical Sciences 63 (2014) 178–186

Contents lists available at ScienceDirect

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

Human and simulated intestinal fluids as solvent systems to explore food effects on intestinal solubility and permeability Jef Stappaerts a, Benjamin Wuyts a, Jan Tack b, Pieter Annaert a, Patrick Augustijns a,⇑ a b

Drug Delivery and Disposition, KU Leuven Department of Pharmaceutical and Pharmacological Sciences, Leuven, Belgium Department of Gastroenterology, University Hospitals Leuven, Belgium

a r t i c l e

i n f o

Article history: Received 16 May 2014 Received in revised form 26 June 2014 Accepted 14 July 2014 Available online 22 July 2014 Keywords: Food effect Solubility Permeability In situ intestinal perfusion b-Blockers

a b s t r a c t The mixed micelles and vesicles present in the intraluminal environment of the postprandial state exhibit suitable solubilizing capacities for lipophilic drugs. This increase in solubility, however, is accompanied by a decrease in the free fraction caused by micellar entrapment of these lipophilic compounds. In this study, both simulated and aspirated human intestinal fluids of fasted and fed state conditions were used to evaluate the influence of food on the intestinal disposition of a series of structurally related b-blockers, with varying log P values. Using the in situ intestinal perfusion technique with mesenteric blood sampling in rats, it was demonstrated that fed state conditions significantly decreased the absorptive flux of the more lipophilic compounds metoprolol, propranolol and carvedilol, whereas the influence on the flux of the hydrophilic b-blocker atenolol was limited. The solubility of BCS class II compound carvedilol was found to increase significantly in simulated and aspirated media of the fed state. Intestinal perfusions using intestinal media saturated with carvedilol, revealed a higher flux in the fasted state compared to the fed state, despite the higher solubility in the fed state. This study underscores the importance of addressing the complex nature of the behavior of compounds in the intraluminal environment in fasted and fed state conditions. Moreover, our data point out the value of studying the effect of food on both solubility and permeability using biorelevant experimental conditions. Ó 2014 Elsevier B.V. All rights reserved.

1. Introduction Unraveling the mechanisms by which the prandial state affects the absorption of drugs is of great importance to understand and predict the effect of food on the pharmacokinetics of orally administered drugs. The intake of drugs together with a meal has been shown to influence the rate and the extent of absorption of numerous drugs (Singh, 2005). Several mechanisms may be at the origin of these food effects; best known are a delayed gastric emptying time and an increase in the gastric pH in the postprandial state, which can significantly alter the solubility of drugs and, consequently, the amount of drug that will be presented to the small intestine for absorption. In general, the intestinal solubility of poorly soluble, lipophilic drugs will increase in the fed state. The ⇑ Corresponding author. Address: Drug Delivery and Disposition, KU Leuven Department of Pharmaceutical and Pharmacological Sciences, Gasthuisberg O&N 2, Herestraat 49 box 921, 3000 Leuven, Belgium. Tel.: +32 16 330301; fax: +32 16 330305. E-mail address: [email protected] (P. Augustijns). http://dx.doi.org/10.1016/j.ejps.2014.07.009 0928-0987/Ó 2014 Elsevier B.V. All rights reserved.

vesicles and mixed micelles that are formed upon mixing of secreted bile with food constituents exhibit excellent solubilizing capacities for these lipophilic compounds (Clarysse et al., 2009a). Therefore, the resulting intraluminal concentrations are often much higher than the thermodynamic solubility of these compounds in simple aqueous buffer solutions. These findings have prompted researchers to utilize simulated intestinal fluids of fasted and fed state conditions (FaSSIF, FeSSIF) in studies on the effect of food on drug absorption. These media have been optimized and adapted in accordance with the expanding body of literature on the composition of aspirated human intestinal fluids in the fed and fasted state (Kleberg et al., 2010). Despite these efforts to approximate the composition of intestinal fluids as closely as possible, still considerable differences in experimental results are observed when comparing simulated to aspirated intestinal fluids. For instance, in a study of Holmstock et al., the solubility of indinavir in FeSSIF was found to be more than fourfold higher than its solubility in human aspirated fluids of the fed state. Moreover, the indinavir flux towards the mesenteric blood was found to be significantly lower upon perfusion with aspirated fluids of the fed state

179

J. Stappaerts et al. / European Journal of Pharmaceutical Sciences 63 (2014) 178–186

compared to perfusion with simulated fluids (Holmstock et al., 2013). Several studies have highlighted the importance of studying the effect of food on both intestinal solubility and permeability. The increase in solubility in the fed state is often accompanied by a decrease in free fraction and, consequently, a less strong driving force of compounds from the intraluminal environment towards the blood, as a result of entrapment in micelles. Therefore, it may be misleading to predict the effect of food on oral bioavailability based solely on solubility profiling. Several authors have observed this solubility–permeability trade-off in micellar solvent systems (Fischer et al., 2011a; Katneni et al., 2006; Miller et al., 2011; Yano et al., 2010). Simulated intestinal fluids of the fasted state are commonly used as solvent systems in Caco-2 experiments and the effects on solubility and permeability exerted by the micelles present in these biorelevant media can be studied using this in vitro model (Frank et al., 2012b; Ingels et al., 2002). However, due to the fact that simulated intestinal fluids of the fed state and human aspirated intestinal fluids are mostly incompatible with the Caco-2 model, determining the permeability for compounds in biorelevant media remains an important challenge in the evaluation of food effects. Work is ongoing to generate media that exhibit a composition that is relevant for the fed state and are compatible with Caco-2 cells (Markopoulos et al., 2013). It remains challenging, however, to render media compatible with Caco-2 cells, while maintaining substances such as bile salts within biorelevant concentration ranges. More robust experimental absorption models such as the in situ perfusion technique generate the possibility to study permeability for compounds in presence of biorelevant simulated intestinal fluids and aspirated human intestinal fluids (Holmstock et al., 2013). In this study, the impact of coadministration of food on the intestinal disposition was evaluated by using simulated and aspirated human intestinal fluids of fasted and fed state conditions in the in situ intestinal perfusion technique with mesenteric blood sampling in rat. A series of b-blockers which are structurally related but vary significantly in lipophilicity were selected as model compounds (Table 1). Atenolol is a BCS class III compound which is hydrophilic and exhibits mostly paracellular transport across the small intestinal barrier. The more lipophilic compounds metoprolol and propranolol are highly soluble and retain favorable permeability, rendering them typical BCS class I compounds. Being a BCS class II compound, carvedilol exhibits poor solubility and dissolution characteristics. The effect of food on the intestinal solubility–permeability interplay for the four b-blockers was assessed in simulated and aspirated human intestinal fluids of the fasted and fed state. 2. Materials and methods 2.1. Chemicals Atenolol, metoprolol tartrate, propranolol and orlistat were purchased from Sigma–Aldrich (St. Louis, MO). Talinolol was obtained

from Arzneimittelwerk Dresden (Radebeul, Germany). Carvedilol was purchased from Sequoia Research Products (Pangbourne, UK). Sodium acetate trihydrate and methanol were purchased from VWR International (Leuven, Belgium). Acetonitrile and dimethylsulfoxide (DMSO) were purchased from Acros Organics (Geel, Belgium). Dichloromethane was obtained from Fisher Scientific (Leuven, Belgium). Phosphate buffered saline (PBS) and Hanks’ balanced salt solution (HBSS) were provided by Lonza (Basel, Switzerland). Simulated intestinal fluid (SIF) powder was purchased from Biorelevant (Surrey, UK). Ketamine (Anesketin) and xylazin (Xyl-M 2%) were obtained from Eurovet (Heusden, Belgium) and VMD (Arendonk, Belgium), respectively. Stock solutions were prepared in DMSO. Water was purified with a Maxima system (Elga Ltd., High Wycombe Bucks, UK). Ensure Plus (Abbott Laboratories B.V., Zwolle, The Netherlands) was used to simulate a standard meal. One portion of 200 mL has an energy content of 1.263 kJ of which lipids, carbohydrates and proteins constitute 29%, 54% and 17% on energy basis, respectively; the osmolality amounts to 670 mOsm/kg; the pH is 6.6. 2.2. Media Transport medium (TM) consisted of HBSS buffered with HEPES (10 mM) to pH 7.4. Fasted and fed state simulated intestinal fluid (FaSSIF and FeSSIF) were made according to the manufacturer’s protocol. Shortly, FaSSIF and FeSSIF were prepared by dissolving SIF powder in a blank FaSSIF phosphate buffer (2.24 mg/ml) and a blank FeSSIF acetate buffer, respectively (11.2 mg/ml). Modified FeSSIF (pH 6.5) was prepared by dissolving SIF powder in the blank FaSSIF phosphate buffer (11.2 mg/ml). Human intestinal fluids (HIF) from the duodenum of healthy volunteers were collected in two different nutritional states (11 volunteers for FaHIF and 9 volunteers for FeHIF) according to the method described by Bevernage et al. (2011). The human intestinal fluids were collected every 15 min for up to 120 min from the duodenum (D2–D3) after the intake of 200 mL of water (fasted state) or a liquid meal (Ensure Plus 400 mL) + 200 mL of water (fed state). Lipase activity was inhibited immediately upon aspiration by addition of orlistat (final concentration of 1 lM) to the test tubes. For each nutritional state, one pooled sample was made by combining the aspirates from all volunteers. The pooled HIF were stored at 30 °C until further use. An approval for the experiments with humans was granted by the University Hospital Medical Ethics Commission of the KU Leuven. 2.3. Solubility measurements The apparent solubility of carvedilol was determined in TM, FaSSIF, FeSSIF and in HIF of the fasted and fed state using the standard shake flask method. Upon saturation of the human intestinal fluids, the pH of FaHIF and FeHIF amounted to approximately 7.5 and 5.8, respectively. All solubility experiments were performed in triplicate. Approximately 1 mg of carvedilol was added to

Table 1 Key physicochemical and biopharmaceutical characteristics of atenolol, metoprolol, propranolol and carvedilol.

Atenolol Metoprolol Propranolol Carvedilol a b c

MW

AUC ratio (fed/ fasted)a

BCS

exp Log Pc

Log DpH5.0

Log DpH6.5

Log DpH7.4

pKa

% Uncharged pH 5.0

% Uncharged pH 6.5

% Uncharged pH 7.4

fa b

266.3 267.4 259.3 406.5

0.8 1.389 1.521 1.033

III I I II

0.16 1.88 3.48 4.19

2.8 1.47 0.64 0.3

2.48 1.14 0.32 1.22

1.80 0.47 0.36 2.07

9.6 9.6 9.5 7.8

0 0 0 0.02

0.07 0.07 0.07 0.58

0.54 0.54 0.54 4.36

50 95 90 43

AUC ratios (fed/fasted) adopted from Singh (2005). fa: Fraction absorbed values adopted from Haslam et al. (2011). Experimental Log P values were adopted from following references: Poulin and Theil (2002), Ruell et al. (2003), Taylor et al. (1981), Wang et al. (1991).

180

J. Stappaerts et al. / European Journal of Pharmaceutical Sciences 63 (2014) 178–186

microcentrifuge tubes containing 0.5 mL of the above mentioned media and placed in a prewarmed shaking incubator [37 °C at 175 rpm (KS 4000 i control incubator from Ika (Staufen, Germany))] for 30 h. The solid phase was separated from the dissolved part using centrifugation (15 min, 20.817g at 37 °C). The top layer was carefully removed by aspiration. The supernatant of the samples was diluted 1/100 or 1/1000 with methanol:water (20:80 v/v) and quantified using an HPLC system with fluorescence detection. 2.4. Animals Purpose-bred, male Sprague-Dawley rats (Janvier, Le Genest Saint-Isle, France) were used for in situ intestinal perfusion experiments. The rats were housed according to the Belgian and European laws, guidelines and policies for animal experiments, housing and care in the Central Animal Facilities of the university. Approval for this project was granted by the Institutional Ethical Committee for Animal Experimentation. 2.5. In situ intestinal perfusion Rats of approximately 350 g were anaesthetized using a mixture of ketamine (87.5 mg/kg) and xylazin (0.875 mg/kg). The right jugular vein was cannulated with a heparinized (50 IU/ml) polyethylene cannula (o.d. 1.02 mm; Portex, Kent, UK) for blood supply from donor rats during the perfusion experiment. A laparotomy was performed and a distal segment of the small intestine was isolated by inserting two glass cannulas (o.d. 4 mm, i.d. 3 mm) at the proximal and distal end of the segment. A mean intestinal segment radius of 0.2 cm was used in further calculations. Polyethylene tubing was connected to the inlet cannula and a perfusion pump (Minipuls3, Gilson, Middleton, USA) was placed between the perfusate reservoir and the inlet cannula. After removal of the intestinal content, the intestinal segment was preincubated with TM, FaSSIF, or FeSSIF. FaSSIF and FeSSIF were also used as preincubation medium for the experiments with FaHIF and FeHIF. The mesenteric vein draining the isolated part of the intestinal segment was cannulated using the top end (1 cm) of a catheter (InsyteWÒ 0.7  19 mm; Beckton Dickinson, Salt Lake City, Utah). The cannula was secured with a knot and connected to a piece of silastic tubing (o.d. 1.19 mm, i.d. 0.64 mm; Helix Medical, USA) for blood collection. After cannulation of the mesenteric vein, the experiment was initiated by replacing the preincubation medium with the desired perfusion medium (t = 0). Perfusion experiments using media with a predefined concentration of atenolol (50 lM), metoprolol (20 lM), propranolol (20 lM) and carvedilol (20 lM) in TM, FaSSIF and FeSSIF were performed using an open loop set-up, whereas experiments with aspirated human intestinal fluids and experiments with media saturated with carvedilol were performed in the closed loop setup; this was done to avoid excessive use of aspirated intestinal fluids and carvedilol. Similarly, atenolol, metoprolol, propranolol and carvedilol were combined with the same goal of minimizing the use of human intestinal fluids. For the perfusion experiments with saturated media, FaSSIF, FeSSIF, FaHIF or FeHIF were saturated overnight with carvedilol, and the resulting suspensions were used as perfusion media, allowing further dissolution upon absorption of dissolved carvedilol. The perfusion flow rate for all in situ experiments amounted to 1 ml/min. The blood flow from the mesenteric vein was continuously collected over 5-min intervals. Donor blood was supplied via the jugular vein at a rate of 0.3 ml/min using a syringe pump (Pilot A2, Fresenius Vial, Grenoble, France) to maintain hemodynamic conditions. Perfusate samples were taken to verify the donor concentration (Cdonor). Perfusate samples from the saturated media

were taken every 5 min and centrifuged directly after sampling. The supernatant was diluted 1/100 or 1/1000 with methanol:water (20:80 v/v) and quantified using an HPLC system with fluorescence detection. All samples were stored at 20 °C prior to analysis. 2.6. Dialysis We compared the bioaccessible fraction of atenolol, metoprolol, propranolol and carvedilol in every intestinal medium by performing dialysis experiments using the HTD 96b from HTDialysis, LLC (Gales Ferry, CT, US). The donor and acceptor compartments were separated by cellulose membrane strips with a molecular weight cutoff of 12–14 kDa (±3 nm), which are expected to be impermeable for micelles, since they were shown to be about 7 nm and 50 nm in diameter in FeSSIF and FaSSIF, respectively (Kloefer et al., 2010). The membranes were hydrated according to the manufacturer’s instructions. The acceptor compartment was filled with 150 lL of a phosphate buffer for which the osmolality and pH were equalized to this of the donor solution. This was done by preparing a concentrated phosphate buffer, which was subsequently diluted with water to correct for osmotic value; the pH was adjusted using diluted NaOH or HCl. The donor compartment was loaded with an equal volume of intestinal medium containing atenolol (50 lM), metoprolol (20 lM), propranolol (20 lM) and carvedilol (20 lM). Samples of the donor and acceptor compartment were taken after 1, 2 and 6 h. Although this time frame was too short to reach equilibrium between donor and acceptor compartment, as is the case in several other recently published articles, this technique allows us to distinguish between the bioaccessible fractions in each medium (Fischer et al., 2011a; Frank et al., 2012a,b; Holmstock et al., 2013). Samples from the acceptor compartment were directly injected into the HPLC system. 2.7. Analysis Samples were analyzed using an HPLC method with fluorescence detection. Samples obtained from the aforementioned solubility and dialysis experiments and perfusate samples obtained from the in situ experiments were directly injected into the HPLC system. Before quantification of the b-blockers in blood samples by HPLC, atenolol, metoprolol, propranolol and carvedilol were extracted from the blood. After diluting 150 ll of blood in 1250 ll of a mixture of HCl (0.2 M) and MeOH (80:20), 100 ll of internal standard solution (talinolol, 2 lM) and 500 ll of NaOH were added. After extraction with 10 ml of dichloromethane and centrifugation (2880 g, 10 min), the organic layer was transferred to a clean test tube and evaporated to dryness under a gentle stream of air. The test tubes were rinsed with 1 ml MeOH, which was again evaporated to dryness. The residue was dissolved in 150 ll of a solution of water and DMSO (70:30 v/v), of which 100 ll was injected in the HPLC system. The b-blockers and the internal standard were detected using a fluorescence program monitored by a Jasco fluorescence detector (FP-1520); excitation 271 nm, emission 302 nm for atenolol and metoprolol; excitation 249 nm, emission 333 nm for talinolol and propranolol; excitation 249 nm, emission 400 nm for carvedilol. The HPLC system consisted of a Waters 600 series separations module and a Novapak C18 column under radial compression (Waters, Milford, MA). The retention times of atenolol, metoprolol, talinolol (internal standard), propranolol and carvedilol amounted to 3.7, 8.1, 14.1, 15.0 and 17.2 min, respectively. The observed peaks were integrated using Empower Pro (Empower 2) software. The calibration curve was linear over the concentration range of 9.76 nM–5 lM. The assessment of intraday repeatability, determined at 1.25 lM,

J. Stappaerts et al. / European Journal of Pharmaceutical Sciences 63 (2014) 178–186

resulted in a relative S.D. (n = 5) and deviation from the theoretical concentration below 5% for all b-blockers. 2.8. Calculations The intestinal flux of the b-blockers was determined from the cumulative amount of compound appearing in the mesenteric blood between 30 and 60 min according to the following equation:

Flux ¼

dQ dt  A

where Q is the cumulative amount of drug appearing in the mesenteric blood and A is the surface area of the perfused cylindrical intestinal segment. 2.9. Statistics Statistical analysis was performed using an unpaired, twotailed t-test. P < 0.05 was used as a criterion for statistical significance. 3. Results and discussion The major aim of this study was to explore the effect of food on intestinal drug disposition through the use of human intestinal fluids collected in fasted state versus fed state conditions. The b-blockers atenolol, metoprolol, propranolol and carvedilol were selected as model compounds. These b-blockers are structurally related but span a considerable range in log P values. Table 1 gives an overview of the key physicochemical and biopharmaceutical properties of these compounds. None of the selected compounds are known to be substrates of drug transporters present in the small intestine, although carvedilol has been demonstrated to be an inhibitor of the efflux transporter P-glycoprotein (Jonsson et al., 1999). The in situ intestinal perfusion model with mesenteric blood sampling in rat was selected as the most appropriate tool to obtain reliable data in a biorelevant setting. Compared to cell models, the in situ technique exhibits high robustness, which renders it extremely useful to evaluate intestinal absorption from complex media, such as human intestinal fluids. A second part of this study focuses on the complex solubility–permeability interplay existing in the intraluminal environment of the fasted and fed state; for this study, carvedilol was selected as model compound, because of its poor solubility and dissolution characteristics (BCS II). In addition to the human intestinal fluids, several other commonly used solvent systems such as TM (transport medium; HEPES-buffered HBSS, pH 7.4) and simulated intestinal fluids of the fasted and fed state (FaSSIF and FeSSIF) were selected and their effect on intestinal transport and solubility were evaluated in comparison with the human intestinal fluids. 3.1. Intestinal transport of b-blockers (predetermined concentrations) A first set of experiments was designed to evaluate the influence of food on the absorptive flux of the b-blockers at a predetermined concentration: 50 lM for atenolol and 20 lM for metoprolol, propranolol and carvedilol. This allowed us to explore the effect of the different solvent systems on intestinal transport and gauge the effect of micellar interactions and pH effects. Upon analysis of the mesenteric blood samples, notable differences were observed in the absorptive flux; a distinct intercompound variability was observed, as well as a compoundspecific effect of the perfusion media used (Fig. 1). For atenolol, the most hydrophilic compound of the b-blockers tested, the different perfusion media had almost no effect on the measured flux,

181

with the exception of FeHIF, which significantly reduced the flux of atenolol. This drop in flux in FeHIF is most likely caused by the interaction of atenolol with constituents that are present in human intestinal fluids upon food processing; this interaction may give an explanation for the 20% decrease in AUC of atenolol when it was taken together with food, as was observed in a clinical trial (Melander et al., 1979). For the more lipophilic compounds metoprolol, propranolol and carvedilol, fed state conditions notably affected transport of b-blockers towards the mesenteric blood, resulting in significant decreases in flux values in the fed state compared to fasted state conditions. This decrease was observed for simulated (FeSSIF versus FaSSIF) as well as for aspirated intestinal fluids (FeHIF versus FaHIF). Phenomena that may be at the origin of this observation are entrapment of the b-blockers in the mixed micelles present in the simulated and aspirated intestinal fluids and a pH effect as the pH in fed state conditions is typically lower than in fasted state conditions. The micellar inclusion of lipophilic compounds indeed seems to decrease the free concentration of the lipophilic b-blockers and therefore the driving force from the intraluminal environment towards the blood, resulting in a decrease in transport across the small intestinal barrier. The extent of micellar entrapment increases with increasing lipophilicity: as compared to the values observed in FaHIF, the intestinal transport of carvedilol, propranolol and metoprolol in FeHIF amounted to 4%, 8% and 26%, respectively. The simulated media were not able to fully reproduce the absolute flux values that were observed for the human intestinal fluids. This could be attributed to the presence of lipid degradation products and other constituents that are particular to human intestinal fluids, which may also interact with the b-blockers, thereby potentially influencing their transport across the small intestinal barrier. These constituents are generally not included in simulated intestinal fluids. Moreover, in vivo, secreted bile consists of a myriad of different bile salts whereas in the simulated intestinal fluids, only taurocholate is included. Table 2 gives an overview of the characteristics of the simulated and aspirated intestinal fluids. Despite the obvious dissimilarity in composition of the media, it appears from the in situ data that the rank order and relative differences in flux values in aspirated fluids of the fasted and fed state are maintained for the b-blockers in the simulated fluids. The flux of carvedilol, propranolol and metoprolol in FeSSIF amounted to 8%, 15% and 26%, respectively, of the flux values observed in FaSSIF. Since atenolol, metoprolol, propranolol and carvedilol are all weakly basic drugs, differences in pH of the perfusion media may also affect intestinal transport due to a different degree of ionization. Indeed, the pH in the fed state lays around 5.0–5.5, whereas in the fasted state, a slightly higher pH is commonly observed (6.0–7.0) (Clarysse et al., 2009b; Sjögren et al., 2014). It is conceivable that the lower percentage of uncharged compound in the fed state may also result in a decreased flux across the small intestinal barrier compared to the fasted state, especially for carvedilol (pKa = 7.8). To assess the relative impact of pH and micellar interactions on the flux of the b-blockers, a series of perfusions were performed in FeSSIF, of which the pH was modified to 6.5, the pH of FaSSIF. Carvedilol, the most lipophilic compound of the b-blockers tested, exhibited a more than 4-fold higher flux in FaSSIF than in FeSSIF of which the pH was adjusted to 6.5 (further referred to as modified FeSSIF), suggesting strong interactions between carvedilol and the mixed micelles, present in FeSSIF. When comparing FeSSIF (pH 5.0) with modified FeSSIF (pH 6.5) it is clear that, in addition to the effects of micellar entrapment, pH has an important effect on transport over the pH range tested. Nevertheless, the main cause for the observed drop in carvedilol flux in the fed state seems to be the micellar entrapment. Propranolol and metoprolol

182

J. Stappaerts et al. / European Journal of Pharmaceutical Sciences 63 (2014) 178–186

atenolol

metoprolol

4.5

fasted state

fed state

3.5 3.0 2.5

fasted state *

20 flux x 104 (nmol/(s x cm2)

flux x 104 (nmol/(s x cm2)

4.0

*

2.0 1.5 1.0

fed state

15

*

10

n.s. 5

0.5 0.0

0

TM

FaSSIF

FaHIF

FeSSIF

FeSSIF (mod)

FeHIF

TM

FaSSIF

FaHIF

FeSSIF

FeSSIF (mod)

carvedilol

propranolol 20

30

fasted state *

25

fed state

20

*

15 10

n.s.

5

fed state

fasted state

18 flux x 104 (nmol/(s x cm2)

flux x 104 (nmol/(s x cm2)

FeHIF

16 14 12 10 8 6

*

*

4

*

2 0

0 TM

FaSSIF

FaHIF

FeSSIF

FeSSIF (mod)

TM

FeHIF

FaSSIF

FaHIF

FeSSIF

FeSSIF (mod)

FeHIF

Fig. 1. Flux values of atenolol, metoprolol, propranolol and carvedilol towards the mesenteric vein in in situ intestinal perfusion experiments in Sprague–Dawley rats. Donor concentrations were 50 lM for atenolol and 20 lM for metoprolol, propranolol and carvedilol. Perfusion media used were TM, FaSSIF, FeSSIF, modified FeSSIF (pH 6.5), FaHIF and FeHIF. Bars represent the mean + SD (n = 3).

Table 2 Individual bile salt and lecithin concentrations in FaHIF, FeHIF, FaSSIF and FeSSIF. Medium

FaHIF FeHIF FaSSIF FeSSIF

Bile salt concentration (lM)

Lecithin concentration (lM)

TUDC

TCDC

TDC

GUDC

GCDC

GDC

GC

TC

Total

16 54

461 978

242 463

53 234

1047 2442

616 1263

1275 2896

514 887 3000 15000

4224 9217 3000 15000

110 3000 750 3750

TUDC: tauroursodeoxycholate, TCDC: taurochenodeoxycholate, TDC: taurodeoxycholate, GUDC: glycoursodeoxycholate, GCDC: glycochenodeoxycholate, GDC: glycodeoxycholate, GC: glycocholate, TC: taurocholate.

both demonstrated a significantly lower flux in FeSSIF compared to FaSSIF. Since similar fluxes were observed in FeSSIF and modified FeSSIF (pH 6.5), it can be concluded that the effect of food on the flux of metoprolol and propranolol is largely due to entrapment of these compounds in mixed micelles. Despite the fact that, given a pKa of 9.6 the extent of uncharged compound in FaSSIF is more than 30-fold higher than in FeSSIF, this seems to be of minor significance for the overall intestinal transport of these BCS class I compounds. This lack of pH effect on intestinal permeability was also seen in a study reported by Incecayir et al. who demonstrated similar effective permeability values for metoprolol at pH 6.5 and pH 7.5 in rat intestinal perfusion experiments (Incecayir et al., 2013). On the other hand, an earlier study by Dahan et al. reported an increase in the intestinal permeability for metoprolol in rat jejunum and ileum (Dahan et al., 2010). Although speculative at this point of time, dissimilarities among these studies may originate

from differences in composition of perfusion media: MES and HEPES buffer (Incecayir et al., 2013) and phosphate buffers (Dahan et al., 2010) are plain aqueous buffers whereas the behavior of compounds in micellar media may be different. Moreover, the presence of an acidic microclimate at the unstirred water layer of the small intestinal barrier may offset the effect of the pH of the perfusion media on the overall absorption (Lucas, 1983; Yeap et al., 2013a). For atenolol, the fluxes in FaSSIF and FeSSIF were similar, indicating a lack of interaction between this hydrophilic compound and the mixed micelles. 3.2. Free fraction dialysis of b-blockers (predetermined concentrations) In view of the fact that micellar entrapment plays a pivotal role in the transport of the b-blockers, we further explored the affinity

183

J. Stappaerts et al. / European Journal of Pharmaceutical Sciences 63 (2014) 178–186

for the micelles by free fraction dialysis experiments at predefined concentrations. This free, bioaccessible fraction is generally assumed to decrease for lipophilic drugs when the concentration of micelle forming constituents increases above the critical micellar concentration (Frank et al., 2014; Miller et al., 2011). Dialysis experiments were performed over a time span of 6 h and generally support the in situ findings. Again, marked differences were observed between the hydrophilic atenolol and the more lipophilic compounds metoprolol, propranolol and carvedilol (Fig. 2). For atenolol, the appearance at the acceptor side was very similar for all the media tested, which is in good agreement with the observed flux in the in situ experiments. In contrast, for propranolol and carvedilol, large differences were observed in transport profiles across the dialysis membranes. Moreover, the free concentration in FeHIF was extremely low for these compounds. These findings correspond to the low flux observed in FeHIF for propranolol and carvedilol in situ. It is clear from the results obtained in situ and in vitro that food has a substantial effect on the transport of lipophilic compounds from the intraluminal environment towards the blood and that this effect is at least partly caused by increased micellar entrapment in the postprandial state. Despite the various points of agreement in the free fraction dialysis profiles and the in situ data, the in vitro set-up seems to be less discriminative between the different media with reference to permeation. More specifically, when FeSSIF was used as donor medium, the flux across the dialysis membranes appears to be relatively high compared to the in situ findings. For example, similar acceptor concentrations were measured in FaSSIF and FeSSIF for metoprolol and carvedilol during the dialysis experiments, whereas in situ, metoprolol and carvedilol flux values were significantly different for these perfusion media. Consequently, it appears that the use of dialysis membranes does not fully capture the complexity of the intestinal absorption of compounds. Discrepancies between determination of free fraction and permeability values have also been observed by Fischer et al. (2011b). However, in general, determination of free fraction by dialysis gives a relatively good

3.3. Solubility of carvedilol The experiments so far were performed using a predefined concentration of the b-blockers, allowing the determination of the intrinsic absorption capacity. For low solubility compounds, the extent of absorption is also strongly affected by the dissolution and solubility in the intraluminal environment. In view of the fact that atenolol, metoprolol and propranolol are freely soluble, solubility experiments were focused on carvedilol. Carvedilol is a BCS class II compound and exhibits poor solubility and dissolution characteristics. Several attempts have been made to overcome the problematic solubility of carvedilol, including the use of carvedilol in lipid based formulations or solid dispersions (Stillhart et al., 2014; Yuvaraja and Khanam, 2014). It is known, however, that for poorly soluble, lipophilic compounds such as carvedilol, the intestinal solubility may increase significantly in the fed state through solubilization by the mixed micelles. Using simulated intestinal fluids, Fagerberg et al. demonstrated that endogenous compounds present in the intraluminal environment, such as bile salts and phospholipids, already generate a significant improvement in the dissolution rate and solubility of carvedilol (Fagerberg et al., 2010). In our study, the apparent solubility of carvedilol was determined using the standard shake flask method in TM, FaSSIF, FeSSIF and in HIF collected in the fasted and fed state. Consistent with the findings of Fagerberg, the presence of bile salts and phospholipids in the simulated and aspirated intestinal fluids significantly improved the solubility of carvedilol compared to its solubility in transport medium (Fig. 3). The solubility in FeHIF was almost 3-fold higher than in FaHIF. Moreover, the solubility in FeSSIF was found to be approximately 4-fold higher than in FaSSIF. Remarkably, there are considerable differences in absolute solubility values between simulated and aspirated intestinal fluids. The

atenolol

metoprolol

15

Concentration (µM)

30

Concentration (µM)

indication of the interaction between compounds and the micellar fraction of the medium and, consequently, the bioaccessible fraction of a compound.

20

10

10

5

0

0 0

2

4

6

8

0

2

Time (h) propranolol

6

8

6

8

carvedilol

8

Concentration (µM)

Concentration (µM)

15

4

Time (h)

10

5

6 4 2 0

0 0

2

4

Time (h)

6

8

0

2

4

Time (h)

Fig. 2. Concentrations of the b-blockers measured in the acceptor compartment as a function of time in the dialysis setup. Donor concentrations were 50 lM for atenolol and 20 lM for metoprolol, propranolol and carvedilol in TM (d), FaSSIF (j), FaHIF (.), FeSSIF (N) and FeHIF (). The acceptor medium consisted of buffer for which the osmolality and pH was equalized to this of the donor solution. Values are shown as the mean ± SD (n = 3).

184

J. Stappaerts et al. / European Journal of Pharmaceutical Sciences 63 (2014) 178–186 1600

Flux (nmol/(s x cm2))

Concentraon (μM)

*

1.2E-02

1400 1200 1000 800 600 400 200

1.0E-02

*

8.0E-03 6.0E-03 4.0E-03 2.0E-03 0.0E+00

0 FaSSIF

FeSSIF

FaHIF

FeHIF

Fig. 3. Solubility of carvedilol in TM, FaSSIF, FeSSIF, FaHIF and FeHIF, measured after 24 h. Solubility was determined using the shake flask method at 37 °C. Bars represent the mean + SD (n = 3).

solubility of carvedilol in FaSSIF and FeSSIF is 3 and 4 times higher than its solubility in FaHIF and FeHIF, respectively, indicating that solubility measurements in simulated intestinal fluids do not always correctly represent the actual intraluminal concentrations that can be expected in vivo (Holmstock et al., 2013; Wuyts et al., 2013). It is important to note, however, that the buffer capacity of the aspirated fluids is much lower than for the simulated fluids (Perez de la Cruz Moreno et al., 2006). The observed pH in FaHIF and FeHIF was therefore slightly higher than in the simulated fluids, resulting in a lower percentage of ionized compound, leading to a lower solubility. Solubility values observed in FaSSIF and FeSSIF in this study are more than twofold higher than the results published by Fagerberg et al. (2010). These inconsistencies may be due to different experimental techniques used to measure solubilities: in the study of Fagerberg et al. in situ UV probes are applied to evaluate dissolution and solubility. Moreover, the preparation protocol to prepare the simulated intestinal fluids differs from the one used in this study. 3.4. Intestinal transport of carvedilol (used as suspension) To be able to integrate the positive effect of food on the solubility of carvedilol into transport studies, simulated and aspirated intestinal fluids were saturated with carvedilol and the resulting suspensions were used to assess the intestinal transport in the in situ intestinal perfusion with mesenteric blood sampling in rat. The use of saturated suspensions allowed further dissolution upon absorption of dissolved carvedilol. Both in simulated and in aspirated intestinal fluids, equilibrium solubilities of carvedilol were much higher than the concentration of 20 lM that was applied in the previous perfusion experiments (Fig. 3). As a result, the absorptive flux in the saturated perfusion media increased accordingly (Fig. 4). In agreement with the data from the solubility studies, this relative increase in intestinal absorption was stronger in fed state conditions. Nevertheless, it appears that, despite increased solubility, the higher extent of micellar interactions with carvedilol in the fed state still results in flux values from the saturated perfusion media of the fed state that are lower than values in fasted state conditions; the observed intestinal transport of carvedilol in FeSSIF and FeHIF amounted to 26% and 16% of the values observed in FaSSIF and FaHIF, respectively. This drop in flux in fed state conditions is substantial, but less significant than was observed using media with predefined concentrations (Fig. 1). Therefore, it is pertinent to take the effects of food on both solubility and permeability into account. In a pharmacokinetic study reported by Louis et al., food was shown to have no effect on the oral bioavailability of carvedilol

FaSSIF

FaHIF

FeSSIF

FeHIF

Fig. 4. Flux values of carvedilol towards the mesenteric vein in in situ intestinal perfusion experiments in Sprague–Dawley rats. Perfusion media used were TM, FaSSIF, FeSSIF, FaHIF and FeHIF, saturated with carvedilol. Bars represent the mean + SD (n = 3).

(Louis et al., 1987). This finding could not be fully captured by the observations from the present study as a negative effect of food on the flux of carvedilol from a saturated suspension was observed (Fig. 4). Although the clinical data reported by Louis et al. were obtained in elderly, they emphasize the importance to consider the myriad of factors that may be important when studying the effect of concomitant intake of food together with an oral dosage form. It is important to note that the human intestinal fluids used in this study, were aspirated from the proximal site in the small intestine (D2 and D3 sections of the duodenum) and that lipase activity was inhibited directly upon aspiration of the fluids. Therefore, the low free fraction observed for the b-blockers is a representation of the situation that occurs at proximal sites of the small intestine at a relatively early phase in the process of dispersion and digestion of the intraluminal contents. A factor which may be at the origin of the discrepancy between the in vivo and in situ results may be the creation of supersaturation in vivo, induced either by lipid processing or gastrointestinal transfer. The highly dynamic process of digestion, dispersion and absorption may induce supersaturation and precipitation of compounds. Several authors have proposed absorption of fatty acids at the acidic microclimate of the unstirred water layer, digestion or dispersion of lipids or even dilution by the significant volume of digestive secretions to be at the origin of supersaturation phenomena (Anby et al., 2012; Yeap et al., 2013a,b). Stillhart et al. demonstrated that the processing of lipid based formulations containing carvedilol may result in supersaturation followed by precipitation of carvedilol in the amorphous state. This amorphous solid may rapidly redissolve and still be available for absorption (Stillhart et al., 2014). Moreover, several recently published articles report promising results on the use of amorphous solid dispersions as a formulation strategy to increase apparent solubility without depression of intestinal permeability (Frank et al., 2012a, 2014; Miller et al., 2012). A second trigger that may induce supersaturation is the transfer from the stomach to the more basic intraluminal environment. Carvedilol is a weak base with a pKa of approximately 7.8, which ensures the ionized state in the acidic environment of the stomach. Reported solubility of carvedilol in simulated gastric fluids exceeds the solubility that was observed in this study both in FaHIF and FeHIF (Murthy and Raju, 2011). As a result, gastrointestinal transfer of carvedilol may generate an initial period of supersaturation in the small intestine, followed by precipitation of carvedilol. Moreover, studies performed by Fagerberg et al. demonstrated favorable dissolution characteristics of carvedilol in FeSSIF, which may potentially lead to intestinal carvedilol concentrations near the equilibrium concentration in the fed state (Fagerberg et al., 2010).

J. Stappaerts et al. / European Journal of Pharmaceutical Sciences 63 (2014) 178–186

4. Conclusion The dynamic interplay between compounds and the highly complex intraluminal environment and its effect on intestinal absorption was demonstrated in this study by using human intestinal fluids as solvent system. Fluids were collected from the upper small intestine, thus reflecting a state in which digestion and dispersion are still in a relatively early phase. Effects of food were especially observed for the more lipophilic drugs. The hydrophilic compound atenolol exhibited limited interactions with mixed micelles, whereas it was observed that the micellar entrapment became more prominent with increasing lipophilicity for the b-blockers metoprolol, propranolol and carvedilol. This micellar inclusion was shown to increase the solubility of carvedilol in fed state conditions. Furthermore, carvedilol flux was negatively influenced by the lower pH in the fed state, resulting in a lower extent of unionized compound. Intestinal transport experiments using media saturated with carvedilol revealed that, despite a higher solubility in the postprandial intraluminal environment, the high level of micellar entrapment in the fed state resulted in a negative effect of food on the flux of carvedilol from saturated intestinal media. Therefore, it would be of great interest for follow up studies to integrate pancreatic enzymes into the in situ environment and evaluate the intraluminal behavior of carvedilol in human intestinal fluids containing digestive enzymes. In conclusion, this study demonstrates the necessity to evaluate both intestinal solubility and permeability in biorelevant conditions to capture the effect of food on intestinal absorption. Acknowledgements This research was funded by grants from: (1) Fund for Scientific Research in Flanders (FWO), and (2) ‘Onderzoeksfonds’ of the KU Leuven in Belgium. References Anby, M.U., Williams, H.D., McIntosh, M., Benameur, H., Edwards, G.A., Pouton, C.W., Porter, C.J.H., 2012. Lipid digestion as a trigger for supersaturation: evaluation of the impact of supersaturation stabilization on the in vitro and in vivo performance of self-emulsifying drug delivery systems. Mol. Pharm. 9, 2063– 2079. 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. Clarysse, S., Psachoulias, D., Brouwers, J., Tack, J., Annaert, P., Duchateau, G., Reppas, C., Augustijns, P., 2009a. Postprandial changes in solubilizing capacity of human intestinal fluids for BCS class II drugs. 1456–1466. Clarysse, S., Tack, J., Lammert, F., Duchateau, G., Reppas, C., Augustijns, P., 2009b. Postprandial evolution in composition and characteristics of human duodenal fluids in different nutritional states. J. Pharm. Sci. 98, 1177–1192. Dahan, A., Miller, J.M., Hilfinger, J.M., Yamashita, S., Yu, L.X., Lennernäs, H., Amidon, G.L., 2010. High-permeability criterion for BCS classification: segmental/pH dependent permeability considerations. Mol. Pharm. 7, 1827–1834. Fagerberg, J.H., Tsinman, O., Sun, N., Tsinman, K., Avdeef, A., Bergström, C.A.S., 2010. Dissolution rate and apparent solubility of poorly soluble drugs in biorelevant dissolution media. Mol. Pharm. 7, 1419–1430. Fischer, S.M., Brandl, M., Fricker, G., 2011a. Effect of the non-ionic surfactant Poloxamer 188 on passive permeability of poorly soluble drugs across Caco-2 cell monolayers. Eur. J. Pharm. Biopharm. Off. J. Arbeitsgemeinschaft Für Pharm. Verfahrenstechnik EV 79, 416–422. Fischer, S.M., Flaten, G.E., Hagesæther, E., Fricker, G., Brandl, M., 2011b. In-vitro permeability of poorly water soluble drugs in the phospholipid vesicle-based permeation assay: the influence of nonionic surfactants. J. Pharm. Pharmacol. 63, 1022–1030. Frank, K.J., Rosenblatt, K.M., Westedt, U., Hölig, P., Rosenberg, J., Mägerlein, M., Fricker, G., Brandl, M., 2012a. Amorphous solid dispersion enhances permeation of poorly soluble ABT-102: true supersaturation vs. apparent solubility enhancement. Int. J. Pharm. 437, 288–293. Frank, K.J., Westedt, U., Rosenblatt, K.M., Hölig, P., Rosenberg, J., Mägerlein, M., Brandl, M., Fricker, G., 2012b. Impact of FaSSIF on the solubility and dissolution-/permeation rate of a poorly water-soluble compound. Eur. J. Pharm. Sci. Off. J. Eur. Fed. Pharm. Sci. 47, 16–20. Frank, K.J., Westedt, U., Rosenblatt, K.M., Hölig, P., Rosenberg, J., Mägerlein, M., Fricker, G., Brandl, M., 2014. What is the mechanism behind increased

185

permeation rate of a poorly soluble drug from aqueous dispersions of an amorphous solid dispersion? J. Pharm. Sci. 103, 1779–1786. Haslam, I.S., O’Reilly, D.A., Sherlock, D.J., Kauser, A., Womack, C., Coleman, T., 2011. Pancreatoduodenectomy as a source of human small intestine for Ussing chamber investigations and comparative studies with rat tissue. Biopharm. Drug Dispos. 32, 210–221. Holmstock, N., Bruyn, T.D., Bevernage, J., Annaert, P., Mols, R., Tack, J., Augustijns, P., 2013. Exploring food effects on indinavir absorption with human intestinal fluids in the mouse intestine. Eur. J. Pharm. Sci. Off. J. Eur. Fed. Pharm. Sci. 49, 27–32. Incecayir, T., Tsume, Y., Amidon, G.L., 2013. Comparison of the permeability of metoprolol and labetalol in rat, mouse, and Caco-2 cells: use as a reference standard for BCS classification. Mol. Pharm. 10, 958–966. Ingels, F., Deferme, S., Destexhe, E., Oth, M., Van den Mooter, G., Augustijns, P., 2002. Simulated intestinal fluid as transport medium in the Caco-2 cell culture model. Int. J. Pharm. 232, 183–192. Jonsson, Ö., Behnam-Motlagh, P., Persson, M., Henriksson, R., Grankvist, K., 1999. Increase in doxorubicin cytotoxicity by carvedilol inhibition of P-glycoprotein activity. Biochem. Pharmacol. 58, 1801–1806. Katneni, K., Charman, S.A., Porter, C.J.H., 2006. Permeability assessment of poorly water-soluble compounds under solubilizing conditions: the reciprocal permeability approach. J. Pharm. Sci. 95, 2170–2185. Kleberg, K., Jacobsen, J., Müllertz, A., 2010. Characterising the behaviour of poorly water soluble drugs in the intestine: application of biorelevant media for solubility, dissolution and transport studies. J. Pharm. Pharmacol. 62, 1656– 1668. Kloefer, B., van Hoogevest, P., Moloney, R., Kuentz, M., Leigh, M.L.S., Dressman, J., 2010. Study of a standardized taurocholate–lecithin powder for preparing the biorelevant media FeSSIF and FaSSIF. diss. Louis, W.J., McNeil, J.J., Workman, B.S., Drummer, O.H., Conway, E.L., 1987. A pharmacokinetic study of carvedilol (BM 14.190) in elderly subjects: preliminary report. J. Cardiovasc. Pharmacol. 10 (Suppl 11), S89–S93. Lucas, M., 1983. Determination of acid surface pH in vivo in rat proximal jejunum. Gut 24, 734–739. Markopoulos, C., Thoenen, F., Preisig, D., Symillides, M., Vertzoni, M., Parrott, N., Reppas, C., Imanidis, G., 2013. Biorelevant media for transport experiments in the Caco-2 model to evaluate drug absorption in the fasted and the fed state and their usefulness. Eur. J. Pharm. Biopharm. Off. J. Arbeitsgemeinschaft Pharm. Verfahrenstechnik EV.. Melander, A., Stenberg, P., Liedholm, H., Scherstén, B., Wåhlin-Boll, E., 1979. Foodinduced reduction in bioavailability of atenolol. Eur. J. Clin. Pharmacol. 16, 327– 330. Miller, J.M., Beig, A., Krieg, B.J., Carr, R.A., Borchardt, T.B., Amidon, G.E., Amidon, G.L., Dahan, A., 2011. The solubility–permeability interplay: mechanistic modeling and predictive application of the impact of micellar solubilization on intestinal permeation. Mol. Pharm. 8, 1848–1856. Miller, J.M., Beig, A., Carr, R.A., Spence, J.K., Dahan, A., 2012. A win-win solution in oral delivery of lipophilic drugs: supersaturation via amorphous solid dispersions increases apparent solubility without sacrifice of intestinal membrane permeability. Mol. Pharm. 9, 2009–2016. Murthy, K.V.R., Raju, V., 2011. Development and validation of new discriminative dissolution method for carvedilol tablets. Indian J. Pharm. Sci. 73, 527. Perez de la Cruz Moreno, M., Oth, M., Deferme, S., Lammert, F., Tack, J., Dressman, J., Augustijns, P., 2006. Characterization of fasted-state human intestinal fluids collected from duodenum and jejunum. J. Pharm. Pharmacol. 58, 1079–1089. Poulin, P., Theil, F.-P., 2002. Prediction of pharmacokinetics prior to in vivo studies. 1. Mechanism-based prediction of volume of distribution. J. Pharm. Sci. 91, 129– 156. Ruell, J.A., Tsinman, K.L., Avdeef, A., 2003. PAMPA—a drug absorption in vitro model: 5. Unstirred water layer in iso-pH mapping assays and pKa flux—optimized design (pOD-PAMPA). Eur. J. Pharm. Sci. 20, 393–402. Singh, B.N., 2005. A quantitative approach to probe the dependence and correlation of food-effect with aqueous solubility, dose/solubility ratio, and partition coefficient (Log P) for orally active drugs administered as immediate-release formulations. Drug Dev. Res. 65, 55–75. Sjögren, E., Abrahamsson, B., Augustijns, P., Becker, D., Bolger, M.B., Brewster, M., Brouwers, J., Flanagan, T., Harwood, M., Heinen, C., Holm, R., Juretschke, H.-P., Kubbinga, M., Lindahl, A., Lukacova, V., Münster, U., Neuhoff, S., Nguyen, M.A., van Peer, A., Reppas, C., Hodjegan, A.R., Tannergren, C., Weitschies, W., Wilson, C., Zane, P., Lennernäs, H., Langguth, P., 2014. In vivo methods for drug absorption – comparative physiologies, model selection, correlations with in vitro methods (IVIVC), and applications for formulation/API/excipient characterization including food effects. Eur. J. Pharm. Sci. 57, 99–151. Stillhart, C., Dürr, D., Kuentz, M., 2014. Toward an improved understanding of the precipitation behavior of weakly basic drugs from oral lipid-based formulations. J. Pharm. Sci. 103, 1194–1203. Taylor, E., Jefferson, D., Carroll, J., Turner, P., 1981. Cerebrospinal fluid concentrations of propranolol, pindolol and atenolol in man: evidence for central actions of beta-adrenoceptor antagonists. Br. J. Clin. Pharmacol. 12, 549– 559. Wang, W., Sasaki, H., Chien, D.-S., Lee, V.H.L., 1991. Lipophilicity influence on conjunctival drug penetration in the pigmented rabbit: a comparison with corneal penetration. Curr. Eye Res. 10, 571–579. Wuyts, B., Brouwers, J., Mols, R., Tack, J., Annaert, P., Augustijns, P., 2013. Solubility profiling of HIV protease inhibitors in human intestinal fluids. J. Pharm. Sci. 102, 3800–3807.

186

J. Stappaerts et al. / European Journal of Pharmaceutical Sciences 63 (2014) 178–186

Yano, K., Masaoka, Y., Kataoka, M., Sakuma, S., Yamashita, S., 2010. Mechanisms of membrane transport of poorly soluble drugs: role of micelles in oral absorption processes. J. Pharm. Sci. 99, 1336–1345. Yeap, Y.Y., Trevaskis, N.L., Porter, C.J.H., 2013a. Lipid absorption triggers drug supersaturation at the intestinal unstirred water layer and promotes drug absorption from mixed micelles. Pharm. Res. 30, 3045–3058.

Yeap, Y.Y., Trevaskis, N.L., Porter, C.J.H., 2013b. The potential for drug supersaturation during intestinal processing of lipid-based formulations may be enhanced for basic drugs. Mol. Pharm. 10, 2601–2615. Yuvaraja, K., Khanam, J., 2014. Enhancement of carvedilol solubility by solid dispersion technique using cyclodextrins, water soluble polymers and hydroxyl acid. J. Pharm. Biomed. Anal. 96, 10–20.