EJPB 11699

No. of Pages 12, Model 5G

8 September 2014 European Journal of Pharmaceutics and Biopharmaceutics xxx (2014) xxx–xxx 1

Contents lists available at ScienceDirect

European Journal of Pharmaceutics and Biopharmaceutics journal homepage: www.elsevier.com/locate/ejpb

2

Research paper

6 4 7

Prediction of oral absorption of cinnarizine – A highly supersaturating poorly soluble weak base with borderline permeability

5 8

Q1

a

9 10

b

12 11 13 1 3 5 0 16 17 18 19 20 21 22 23 24 25 26 27 28 29

Mark Berlin a, Karl-Heinz Przyklenk b, Annette Richtberg b, Wolfgang Baumann b, Jennifer B. Dressman a,⇑ Institute of Pharmaceutical Technology, Goethe University, Frankfurt am Main, Germany Hennig Arzneimittel GmbH & Co KG, Flörsheim am Main, Germany

a r t i c l e Q2

i n f o

Article history: Received 6 June 2014 Accepted in revised form 20 August 2014 Available online xxxx Keywords: Dissolution Supersaturation Precipitation Cinnarizine Permeability Absorption Physiologically based pharmacokinetic model

a b s t r a c t Two important driving forces for oral absorption of active pharmaceutical ingredients are drug dissolution and permeability in the gastrointestinal tract. Poorly soluble weak bases typically exhibit high solubility under fasted gastric conditions. However, the solubility of such drugs usually decreases drastically in the fasted small intestine, constraining drug absorption. Since there is a discrepancy in solubility between the fasted state stomach and intestine, it is crucial to examine the influence of dissolution, supersaturation and precipitation on the oral absorption of poorly soluble weak bases during and after fasted state gastric emptying. Cinnarizine is a poorly soluble weak base with borderline permeability, exhibiting supersaturation and precipitation under simulated fasted state gastric emptying conditions. Interestingly, supersaturation and precipitation of cinnarizine under fed state conditions is not expected to occur, since the drug shows good solubility in fed state biorelevant media and exhibits a positive food effect in pharmacokinetic studies. The present work is aimed at investigating the dissolution, supersaturation and precipitation behavior of marketed cinnarizine tablets under fasted and fed state conditions using biorelevant dissolution and transfer methods. In order to predict the in vivo performance of these cinnarizine formulations, the in vitro results were then coupled with different physiologically based pharmacokinetic (PBPK) models, which considered either only dissolution or a combination of dissolution, supersaturation and precipitation kinetics. The results of the in silico predictions were then compared with in vivo observations. The study revealed that under fasting conditions, plasma profiles could be accurately predicted only when supersaturation and precipitation as well as dissolution were taken into account. It was concluded that for poorly soluble weak bases with moderate permeability supersaturation and precipitation during fasted state gastric emptying may have an essential influence on oral drug absorption and thus on in vivo drug performance. Ó 2014 Elsevier B.V. All rights reserved.

31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57

Abbreviations: API, active pharmaceutical ingredient; AUC, area under the curve; BCS, biopharmaceutics classification system; Cmax, maximal concentration; FaSSGF, Fasted State Simulated Gastric Fluid; FaSSIF, Fasted State Simulated Intestinal Fluid; FaSSIF-V2, Fasted State Simulated Intestinal Fluid – version 2; FaSSIF-V2(PO4), Fasted State Simulated Intestinal Fluid – version 2 with phosphate buffer; fd, fraction of drug dissolved; FeSSGF, Fed State Simulated Gastric Fluid; FeSSIF-V2, Fed State Simulated Intestinal Fluid – version 2; GER, gastric emptying rate; GI, gastrointestinal; HPLC, high pressure liquid chromatography; kp, precipitation constant; Papp, apparent permeability; Peff, effective permeability; PBPK, physiologically based pharmacokinetics; PK, pharmacokinetics; PSA, polar surface area; PTFE, polytetrafluoroethylene; rpm, rotations per minute; SAP, surface activity profiling; tmax, time of maximal concentration; Vd, volume of distribution; Vd/F, apparent volume of distribution. ⇑ Corresponding author. Institute of Pharmaceutical Technology, Goethe University Frankfurt, Max-von-Laue-Strasse 9, 60438 Frankfurt am Main, Germany. Tel.: +49 (0) 69 798 29680. E-mail address: [email protected] (J.B. Dressman).

1. Introduction

58

The prediction of oral drug absorption of active pharmaceutical ingredients (APIs) is a topic which is gaining importance in pharmaceutical research and development. Early investigations have shown that oral absorption of APIs is generally limited by two key factors, gastrointestinal (GI) solubility and intestinal permeability [1]. The introduction of biorelevant media, which mimic the fluids of the GI tract, has facilitated the assessment of the in vivo behavior of poorly soluble drugs [2–4]. Biorelevant dissolution testing has become a significant tool for qualitative in vitro–in vivo correlation of drugs, especially of BCS class II compounds [5–7]. Further, coupling of biorelevant dissolution with physiologically based pharmacokinetic (PBPK) models opens the

59

http://dx.doi.org/10.1016/j.ejpb.2014.08.011 0939-6411/Ó 2014 Elsevier B.V. All rights reserved.

Please cite this article in press as: M. Berlin et al., Prediction of oral absorption of cinnarizine – A highly supersaturating poorly soluble weak base with borderline permeability, Eur. J. Pharm. Biopharm. (2014), http://dx.doi.org/10.1016/j.ejpb.2014.08.011

60 61 62 63 64 65 66 67 68 69 70

EJPB 11699

No. of Pages 12, Model 5G

8 September 2014 2 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136

M. Berlin et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2014) xxx–xxx

possibility of quantitatively assessing the in vivo performance of such APIs [7–13]. Dissolution testing under simulated intestinal conditions is generally a very useful tool to assess the in vivo behavior of poorly soluble drugs, but it may have limited predictive power in the case of weak bases. While weakly basic drugs tend to dissolve well in the stomach, their solubility in the upper small intestine is significantly lower due to the considerably higher pH values in this environment. In order to evaluate the behavior of weak bases during gastric emptying, Kostewicz et al. introduced the transfer method [14]. The study revealed that certain weakly basic drugs tend to supersaturate and then precipitate when moving from the stomach into the intestine. On the other hand, creation of a supersaturated solution of the drug may lead to an enhanced driving force for uptake from the small intestine [15,16]. Thus, it is important to understand both the degree of supersaturation that can be achieved and the tendency to precipitate in the small intestine during and after gastric emptying in order to quantitatively describe the in vivo performance of weak bases with PBPK models. Several commercial PBPK simulation software, such as SimCypÒ, PKSimÒ and GastroPlusÒ, permit the incorporation of supersaturation and precipitation kinetics into their models [17]. Shono and co-workers developed a STELLAÒ model which considered possible supersaturation and precipitation of the weak base, nelfinavir [18]. Yet the model used only a statistical estimation of precipitation of the drug; no in vitro supersaturation and precipitation investigations were carried out. The STELLAÒ software has also successfully been used to quantitatively predict the in vivo performance of a BCS class IV base which showed minor supersaturation in the fasted state transfer model [19]. Likewise in a recent study on dantrolene precipitation from a salt formulation in the fasted GI tract, Kambayashi et al. were able to predict the plasma profile of this compound by using dissolution and precipitation data obtained from fasted state biorelevant dissolution testing [20]. However, since dantrolene is a poorly soluble weak acid, its gastrointestinal behavior is expected to be quite different than that of weak bases such as cinnarizine. Cinnarizine is a lipophilic weak base (pKa1 = 1.95 and pKa2 = 7.47 (both basic)) with a molecular weight of 368.514 g/ mol and an octanol–water distribution coefficient (log P) of 5.6 [21,22]. It exhibits a polar surface area (PSA) of 2.4–11.8 Å and low solubility at pH values greater than 4.5 [21,23,24]. A study on weak bases has shown cinnarizine to supersaturate and precipitate when being transferred from a highly acidic compartment (0.1 M HCl) into Fasted State Simulated Intestinal Fluid (FaSSIF) adjusted to a pH value of 5.5 [25]. Interestingly, no supersaturation was observed when the drug was moved from a compartment containing the McIlvaine buffer (pH 5.0) into FaSSIF (pH 6.5). This observation could partially explain the gastric acidity dependent bioavailability of cinnarizine, which has been observed in humans and dogs [26,27]. A permeability investigation of cinnarizine in the Caco-2 cell culture model has shown a moderate Papp value of 4.2  106 cm/s in comparison with mannitol (1.1  106 cm/s), a poorly permeable reference compound [28]. Pharmacokinetic investigation of cinnarizine in beagle dogs revealed an oral bioavailability of 55–60%. The authors concluded that reduced oral bioavailability values are attributed to presystemic metabolism rather than to permeability constraints [21]. Cinnarizine demonstrated a positive food effect in pharmacokinetic studies on ArlevertÒ (20 mg cinnarizine tablets) in healthy human volunteers. The administration of ArlevertÒ after a high fat meal resulted in a moderate increase of Cmax and AUC values by 8% and 44%, respectively. One possible explanation for the food effect is that there is precipitation of cinnarizine in the small intestine in the fasted state, but not in the fed state. Another factor for the food effect could be cinnarizine instability, which was recently reported in

ex vivo studies by Koumandrakis et al. [29]. However, it is not known to what extent degradation of the compound takes place in vivo. The aim of this study was to analyze the pre-absorptive behavior of three different marketed formulations of cinnarizine in order to predict their in vivo performance. The roles of drug dissolution, supersaturation and precipitation were examined using biorelevant dissolution and transfer methods. An advanced STELLAÒ model, enabling the introduction of supersaturation and precipitation kinetics obtained from in vitro investigations, was developed for cinnarizine. The model included permeability restrictions, permitting incorporation of Caco-2 permeability data for the prediction of the drug flux through the human intestine. The in vitro results were then entered into the model, which included disposition parameters obtained from available in vivo data. Finally, the simulation results were compared with observed in vivo plasma profiles.

137

2. Materials and methods

154

2.1. Chemicals and reagents

155

StugeronÒ 25 mg tablets (lot ACL1800) were purchased commercially from Janssen-Cilag AG, Baar, Switzerland. ArlevertÒ 20 mg/40 mg Cinnarizine/Dimenhydrinate tablets (lot 324011), 30 mg/60 mg Cinnarizine/Dimenhydrinate tablets (Cinnarizine 30 mg) (lot 11/228) and cinnarizine drug substance (lot ZR001575PUE521-MIC) were kindly donated by Hennig Arzneimittel GmbH, Flörsheim, Germany. Long-life heat-treated and homogenized milk, containing 3.5% fat (UTH) was purchased from AF Deutschland GmbH, Düsseldorf, Germany. Glyceryl monooleate (GMO, Rylo M19 Pharma, 99.5% monoglyceride, lot 173403401989490) was kindly donated by Danisco Specialities, Brabrand, Denmark. Egg phosphatidylcholine (Lipoid E PC, 99.1% pure, lot 108015-1/42) was kindly donated by Lipoid GmbH, Ludwigshafen, Germany. Sodium taurocholate (lot 2011040152) was obtained from Prodotti Chimici E Alimentari S.P.A., Basaluzzo, Italy. Sodium oleate powder (82.7% pure, lot 51110) was purchased from Riedelde Haën, Seelze, Germany. Pepsin (0.51 U/mg, lot 1241256) was purchased from Fluka Chemie AG, Buchs, Switzerland. Maleic acid (lot BCBC3155) was obtained from Sigma–Aldrich, Steinheim, Germany. Ammonium acetate (lot A0028716849), sodium acetate trihydrate (lot A792767747) and monobasic sodium phosphate monohydrate (lot A02140949112) were of analytical grade and purchased from Merck KGaA, Darmstadt, Germany. Acetic acid (lot 12B220509), acetonitrile (lot I691330327), hydrochloric acid (lot 10L060526), methanol (lot I630607212), sodium chloride (lot 10J110040) and sodium hydroxide (lot 09G300017) were of analytical grade and acquired from VWR International, Leuven, Belgium.

156

2.2. Biorelevant media composition

184

Biorelevant media were used for solubility studies as well as for dissolution tests and transfer experiments. Media characterizing the pre- and postprandial conditions of the stomach were the Fasted State Simulated Gastric Fluid, FaSSGF, and the Fed State Simulated Gastric Fluid, FeSSGF [3,4]. Conditions in the pre- and postprandial intestine were simulated by the Fasted State Simulated Intestinal Fluid, FaSSIF-V2, and Fed State Simulated Intestinal Fluid, FeSSIF-V2 [4]. In the fed state the bile salt secretion is known to be higher, resulting in higher bile salt and lecithin concentrations in FeSSIF-V2 than in FaSSIF-V2. Moreover FeSSIF-V2 contains glycerol monooleate and sodium oleate which represent products of lipolysis of dietary fats.

185

Please cite this article in press as: M. Berlin et al., Prediction of oral absorption of cinnarizine – A highly supersaturating poorly soluble weak base with borderline permeability, Eur. J. Pharm. Biopharm. (2014), http://dx.doi.org/10.1016/j.ejpb.2014.08.011

138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153

157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183

186 187 188 189 190 191 192 193 194 195 196

EJPB 11699

No. of Pages 12, Model 5G

8 September 2014 3

M. Berlin et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2014) xxx–xxx

205

When FaSSGF is emptied into FaSSIF-V2 during transfer experiments a pH change in the acceptor phase is inevitable. To avoid drastic pH shifts during transfer experiments in the preprandial intestinal compartment, the pH of FaSSGF was adjusted to a value of pH 2 [30]. To further comply with the transfer experiment requirements, the buffer capacity of FaSSIF-V2 was increased by exchanging the maleate buffer with a phosphate buffer (FaSSIFV2(PO4) [14]. The compositions of simulated stomach and intestinal media used for this study are summarized in Table 1.

206

2.3. Quantitative analysis of cinnarizine samples

207

222

Cinnarizine samples obtained from solubility, dissolution and transfer experiments were quantitatively analyzed via high performance liquid chromatography (HPLC). The HPLC setup consisted of a Merck Hitachi D-7000 interface, a L-4250 UV–VIS detector, a L-7200 autosampler, a L-7100 pump (Merck Hitachi, Darmstadt, Germany) and a Merck LiChroCART 125-4 RP-18 (5 lm) column (Merck KGaA, Darmstadt, Germany). EZchome Elite v2.8 software (Hitachi, Tokyo, Japan) was used for the evaluation of chromatograms. The mobile phase comprised a 1% ammonium acetate/0.2% acetic acid aqueous buffer and acetonitrile in a ratio of 1:4. The flow rate was set at 1.5 ml/min while the injection volume was adjusted to 20 ll. The measurements were performed at a detection wavelength of 254 nm. Using this method the retention time was approximately 3.5 min and the cinnarizine limit of quantification was 50 ng/ml.

223

2.4. Solubility measurements

224

Solubility of cinnarizine in FaSSGF, FaSSGF pH 2, FaSSIF-V2, FaSSIF-V2(PO4) and FeSSIF-V2 was measured utilizing the UniprepÒ (Whatman Inc., Sanford, ME, USA) method. A combination of FaSSGF pH 2 and FaSSIF-V2(PO4) in a ratio of 1:2 was used to additionally evaluate the solubility of cinnarizine in the acceptor medium after completion of transfer experiments. To assess the bile salt dependent solubility of cinnarizine, measurements also were performed in blank buffers of FaSSIF-V2 and FeSSIF-V2. Briefly, solubilities were determined by introducing an excess amount of cinnarizine into a UniprepÒ vial, filled with 3 ml of biorelevant medium, closed with the UniprepÒ cap, slightly shaken to remove bubbles and stored on an orbital shaker at 37 °C for 24 h. Subsequently, the dispersions were filtered through the integrated

197 198 199 200 201 202 203 204

208 209 210 211 212 213 214 215 216 217 218 219 220 221

225 226 227 228 229 230 231 232 233 234 235 236 237

UniprepÒ 0.45 lm PTFE filter. 500 ll of the filtrate were diluted with 500 ll of methanol and analyzed by HPLC. For the determination of the solubility of cinnarizine in FeSSGF the standard shake flask method was utilized. An excess amount of drug was introduced into a 20 ml scintillation glass vial, 15 ml of the medium was added, mixed and the vial was incubated in an orbital shaker at 37 °C for 24 h. Due to insoluble milk proteins in the medium a filtration through a 0.45 lm filter was not possible. Thus, the dispersions were filtered through Whatman 2.7 lm glass microfiber syringe filters (25 mm GD/X, GE Healthcare UK Ltd., Buckinghamshire, UK). Subsequently, soluble proteins were precipitated by adding 2 ml of acetonitrile to 1 ml of filtrate. The mixtures were centrifuged at 7500 rpm for 3 min and the supernatant was analyzed by HPLC. For each medium the measurements were performed in triplicate.

238

2.5. Dissolution tests

253

To mimic the fasted state gastric dissolution conditions the mini paddle assembly (Erweka DT 80, Heusenstamm, Germany), a scaled down USP 2 apparatus, was utilized. The volume in the fasted stomach the volume of FaSSGF was adjusted to 250 ml. The conditions in the fed state stomach and the fasted and fed intestine were mimicked using the USP 2 dissolution test apparatus with 500 ml of each medium (FeSSGF, FaSSIF-V2 and FeSSIF-V2). All dissolution tests were carried out at 37 ± 0.5 °C and a revolution speed of 75 rpm. Sampling for all dissolution tests was performed at 0, 5, 10, 15, 20, 30, 45, 60, 90, 120, 180 and 240 min. The samples were taken manually via a 5 ml glass syringe, which was connected to a stainless steel sampling cannula and a 10 lm polyethylene cannula filter (Erweka GmbH, Heusenstamm, Germany). At each sampling point 5 ml of biorelevant medium was aspirated into the glass syringe and filtered through a 0.45 lm PTFE filter (Rezist 30, GE Healthcare UK Ltd., Buckinghamshire, UK) whereby 4 ml was returned to the dissolution vessel. For each medium, samples were prepared for analysis according to procedures described in the solubility measurement section. All dissolution tests were performed in triplicate.

254

2.6. Transfer experiments

274

Previous work by Kostewicz et al. [14] investigated supersaturation and precipitation of weak bases during gastric emptying using the transfer model. Those studies were conducted by pre-dissolving the drugs in a donor compartment with

275

Table 1 Composition of biorelevant gastric and intestinal media used in the cinnarizine experiments. FaSSGF

FaSSGF pH 2.0

FeSSGF

FaSSIF-V2

FaSSIF-V2(PO4)

FeSSIF-V2

Composition Sodium taurocholate (mM) Lecithin (mM) Glycerol monooleate (mM) Pepsin (mg/ml) Sodium oleate (mM) Sodium chloride (mM) Hydochloric acid (mM) Maleic acid (mM) Monobasic sodium phosphate (mM) Acetic acid (mM) Sodium acetate (mM) Sodium hydroxide (mM) Milk/buffer ratio

0.08 0.02 – 0.1 – 34.2 25.1 – – – – – –

0.08 0.02 – 0.1 – 34.2 12.6 – – – – – –

– – – – – 237.0 – – – 17.1 29.8 – 1/1

3 0.2 – – – 68.62 – 19.12 – – – 34.8 –

3 0.2 – – – 62.6 – – 28.6 – – 8.7 –

10 2 5 – 0.8 125.5 – 55.02 – – – 81.65 –

Characteristic parameters pH Osmolality (mOsm/kg) Buffer capacity (mM/DpH)

1.6 120 ± 2.5 –

2.0 120 ± 2.5 –

5.0 400 ± 10 25

6.5 180 ± 10 10

6.5 180 ± 10 12

5.8 390 ± 10 25

Please cite this article in press as: M. Berlin et al., Prediction of oral absorption of cinnarizine – A highly supersaturating poorly soluble weak base with borderline permeability, Eur. J. Pharm. Biopharm. (2014), http://dx.doi.org/10.1016/j.ejpb.2014.08.011

239 240 241 242 243 244 245 246 247 248 249 250 251 252

255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273

276 277 278

EJPB 11699

No. of Pages 12, Model 5G

8 September 2014 4

M. Berlin et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2014) xxx–xxx

314

Simulated Gastric Fluid (SGF) and transferring the solution into an acceptor compartment filled with FaSSIF or FeSSIF. In the current work the setup of the original transfer model was adapted to better reflect conditions in a PK study. Usually, when a tablet is taken with a glass of water it starts to disintegrate and to release the drug in the fasted stomach while being simultaneously emptied into the intestine. To represent this physiological behavior under fasted conditions, the Mini Paddle apparatus with 250 ml FaSSGF pH 2 was used as the donor compartment and the USP 2 paddle apparatus with 500 ml FaSSIF-V2(PO4) was used as the acceptor compartment. The paddle rotation speed was set at 75 rpm in each compartment and the temperature was adjusted to 37 ± 0.5 °C. The tablets were introduced into the donor compartment and the content simultaneously allowed to dissolve and to transfer into the acceptor compartment using a peristaltic pump (Ismatec Type IPC-8 V3.01, Glattbrugg-Zürich, Switzerland). Gastric emptying was mimicked through implementation of different transfer rates. Kostewicz et al. introduced zero order transfer rates of 9, 4, 2 and 0.5 ml/min. Population studies have shown that the gastric emptying proceeds on average according to first order kinetics with a rate constant of 2.8 h1 in the fasted state and according to zero order kinetics at a rate of approximately 4 kcal/min in the fed state [11]. To comply with average human physiology in the fasted state a first order transfer with a rate constant of 3 h1 was implemented and results were compared with the three zero order rates of 9, 4, and 0.5 ml/min, representing faster than usual gastric emptying, approximately average, and slower than usual gastric emptying rates, respectively. The sampling times were adapted to each transfer rate. Samples were removed from the acceptor compartment and prepared for analysis using the procedure described in Section 2.4. All transfer experiments were performed in triplicate. Due to the high solubilities of cinnarizine in both gastric and small intestinal fed state media, supersaturation and precipitation was not expected to occur. So, in the case of cinnarizine, transfer experiments simulating the fed state were deemed irrelevant and will not further be addressed.

315

2.7. Pharmacokinetic data of cinnarizine tablets

316

One fed state and two fasted state PK studies were considered for this investigation. The fed state study consisted of two different treatments. 28 healthy males were enrolled in a dose proportionality study in which ArlevertÒ 20 mg cinnarizine tablets were compared with Cinnarizine 30 mg tablets. In each treatment the subjects received a cinnarizine tablet (ArlevertÒ or Cinnarizine 30 mg) after a standardized breakfast (high fat meal). Plasma samples were collected at 0, 0.25, 0.5, 0.75, 1, 1.25, 1.5, 1.75, 2.00, 2.5, 3, 4.00, 5, 6, 8, 12, 16 and 24 h [31]. The first fasted state study was conducted in 16 healthy subjects who were treated with one ArlevertÒ tablet after an overnight fast. Blood samples were taken at 1.5 h prior and 0.5, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 12, 14, 16, 18, 20 and 24 h after oral administration [31]. In the second fasted state study 24 healthy male volunteers received one StugeronÒ 25 mg tablet orally. Plasma concentrations were determined for the time points of 0, 0.5, 1, 1.5, 2, 2.5, 3, 3.5, 4, 6, 8, 12, 16, 24, 36, 48, and 72 h [32]. Generally, intravenous plasma profiles are considered necessary to investigate the post absorptive disposition kinetics of APIs. In the case of cinnarizine, only oral data were available, so the distribution and elimination kinetics were evaluated using observed oral plasma profiles of cinnarizine formulations in the fasted and fed state with the help of a commercial pharmacokinetic software (WinNonlinÒ Professional v. 4.1, Pharsight Corporation, Mountain View, CA, USA). Pharmacokinetic studies of ArlevertÒ demonstrated a positive food effect for cinnarizine. Therefore it could be assumed that drug exposure was maximal under fed state

279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313

317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342

conditions. The apparent volume of distribution Vd/F (i.e. Vd adjusted for bioavailability factor F) for the fed state was estimated to be 225.6 ± 35.6 l. For the fasted state studies the Vd/F was estimated to be higher, 336.6 ± 74.7 l for ArlevertÒ and 352 ± 199 l for StugeronÒ. However the standard error for the fasted state estimations is high and seems less reliable. As the F value in the fed state was estimated to be closer to 1, the apparent volume of distribution Vd/F should be closer to the ‘‘real’’ volume of distribution, Vd. Consequently, the fed state Vd/F was implemented into the simulation software for the prediction of both preprandial and postprandial plasma profiles. The distribution parameters k12, k21 and the elimination constant k10 used for the fasted and the fed state were implemented specific to the conditions of administration.

343

2.8. In silico dissolution, supersaturation and precipitation models

357

For the simulation of plasma profiles of cinnarizine the STELLAÒ simulation software (v. 9.1.3, Isee Systems, Inc., Lebanon, NH, USA) was utilized. The dissolution-only PBPK model (Fig. 1), which has been previously described in several publications [9,10,18,19], was employed to predict the fasted and fed state profiles of all cinnarizine formulations (ArlevertÒ, Cinnarizine 30 mg tablets and StugeronÒ). This model utilizes dissolution kinetics and solubility of poorly soluble APIs in simulated gastric and intestinal media without considering drug supersaturation and precipitation. For the fed state it assumes a gastric volume of 750 ml and a zero order gastric emptying rate of 4 kcal/min. Simulations of the fasted state consider a gastric volume of 250 ml and a first order gastric emptying rate of 2.8 h1 [11]. Dissolution kinetics are described by the Noyes–Whitney dissolution model, which is presented in Eq. (1)

358

344 345 346 347 348 349 350 351 352 353 354 355 356

359 360 361 362 363 364 365 366 367 368 369 370 371

372

1=3

dW t DCN ¼ W 2=3 ðX s  W t Þ ¼ zW 2=3 ðC s  C t Þ dt Vdq2=3

ð1Þ

where Wt is the mass of drug dissolved in time t, D is the diffusion coefficient, C is the shape factor, N is the number of particles to be dissolved, W is the mass of drug remaining to be dissolved, Xs is the mass of drug that saturates the volume V, d is the thickness of

Fig. 1. PBPK model for in vivo cinnarizine formulation behavior, in which only dissolution is concerned (forthwith ‘‘dissolution-only model’’).

Please cite this article in press as: M. Berlin et al., Prediction of oral absorption of cinnarizine – A highly supersaturating poorly soluble weak base with borderline permeability, Eur. J. Pharm. Biopharm. (2014), http://dx.doi.org/10.1016/j.ejpb.2014.08.011

374 375 376 377 378

EJPB 11699

No. of Pages 12, Model 5G

8 September 2014 M. Berlin et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2014) xxx–xxx

380 381 382 383 384 385 386 387 388 389 390 391 392 393

394 396

397 398 399 400 401

402 404

405 406 407 408 409 410 411 412 413

414 416

the diffusion layer, q is the particle density, Cs is the solubility of the drug and Ct is the concentration of the drug at time t. The expression DCN/Vdq2/3 can be merged into the initial dissolution rate constant z, which was obtained from dissolution test results in simulated gastric and intestinal media for each of the cinnarizine formulations. The model based on the Noyes–Whitney theory was demonstrated to be reliable for the investigation of dissolution kinetics of drug compounds in several studies [7,9–12,18,19]. For the fasted state an alternative model was established, which considers supersaturation and precipitation during and after gastric emptying. An illustration of the model is presented in Fig. 2. The initial gastric volume was assumed to be 250 ml and the gastric emptying rate to follow first order kinetics (2.8 h1) [11]. The total drug emptied from the stomach into the intestinal compartment at a given time is described in the following equation

dM t ¼ ke M dt

ð2Þ

Mt is the total mass of drug emptied into the intestine at time t, ke is the gastric emptying constant and M is the maximum amount of drug which can be emptied (M = dose). During transfer the amount of dissolved drug in the intestine at a given time point can be described by the kinetics shown in Eq. (3)

W t ¼ Mt f d

ð3Þ

where Wt is the amount of dissolved drug at time t, Mt is the total mass that has been transferred at time t and fd is the fraction of drug in the intestine which is dissolved at transfer time t. Analysis of the transfer experiment results revealed that fd was constant over time during the transfer process for cinnarizine. After the drug is emptied into the intestine it is assumed to be in a metastable supersaturated state. Precipitation kinetics of the supersaturated drug were previously described by Kambayashi et al. [20] and are presented in the following equation

dPt ¼ kp ðW t  X s Þ; dt

t P t0

ð4Þ

where Pt is the amount of precipitated drug at a given time after the transfer from the stomach to the small intestine is complete (t0 ), kp is the first order precipitation constant and Xs is the amount of the free base that saturates the volume V in the intestine (based on the API solubility in the intestinal fluids). Fig. 3 illustrates the evaluation of supersaturation and precipitation kinetics with the aid of ‘‘theoretical’’ and ‘‘observed’’ concentration curves in the intestine during and after gastric emptying. Dissolution of solid and re-dissolution of precipitated weak base in the intestine are described by the Noyes–Whitney based dissolution kinetics presented in Eq. (1). Human oral bioavailability or mass balance studies are preferred for evaluation of the permeability characteristics of drugs [33]. Since such data were not available for cinnarizine, an alternative approach was required to estimate the permeability of cinnarizine. A study of oral and intravenous b-cyclodextrin cinnarizine formulations in beagle dogs revealed cinnarizine to have an oral bioavailability of at least 60% [21]. Surface activity profiling (SAP) has proven to be a practical tool to evaluate the oral absorption and blood brain barrier penetration of drugs in humans [34,35]. Internal investigation of cinnarizine using SAP predicted an oral fraction absorbed of 65% (data not shown). Studies in humans and dogs have suggested that cinnarizine bioavailability is quite dependent on gastric acidity, and thus on solubility rather than permeability restrictions of the drug [26,27]. A recent Caco-2 study demonstrated moderate permeability values of cinnarizine [22,28]. Taking all these results into consideration, it was assumed for modeling purposes that cinnarizine exhibits moderate to good permeability and therefore can be classified as a borderline BCS class II/IV drug. The drug flux from the intestinal lumen through the gut wall into the venous blood stream was calculated using Eq. (5) [36]

ð5Þ

where dYt/dt is the rate of uptake of dissolved drug through the intestinal wall at time t, Papp is the permeability coefficient which was evaluated from Caco-2 data (4.2  106 cm/s) [28], Aeff is the effective surface area in the intestine and C is the drug concentration in the intestine. The intestinal volumes and effective surface area were obtained from literature. The effective surface area was assumed to be 3.6 m2, while the intestinal volume was set at 320 ml in the fed state and at 212 ml in the fasted state [37,38]. To enable plasma profile predictions of cinnarizine formulations in the fed and fasted state, the disposition kinetics (as described in Section 2.7) and GI condition parameters (Table 2) along with

100

0

Fig. 2. PBPK model for cinnarizine formulation behavior, in which dissolution, supersaturation and subsequent precipitation are accounted for (forthwith ‘‘supersaturation and precipitation PBPK model’’).

417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447

448

dY t ¼ Papp Aeff C dt

% dissolved/transferred

379

5

Fig. 3. Evaluation of supersaturation (fd) and precipitation kinetics after gastric emptying (kp). Total mass of drug emptied into intestine at time t is shown as a theoretical concentration curve (- -). The observed percentage (mass) of drug dissolved in the intestine at time t and drug precipitation at times after gastric emptying is complete, t0 , are presented in the continuous curve (-). The vertical line (- -) represents the time at which gastric emptying is complete, t0 .

Please cite this article in press as: M. Berlin et al., Prediction of oral absorption of cinnarizine – A highly supersaturating poorly soluble weak base with borderline permeability, Eur. J. Pharm. Biopharm. (2014), http://dx.doi.org/10.1016/j.ejpb.2014.08.011

450 451 452 453 454 455 456 457 458 459 460 461

EJPB 11699

No. of Pages 12, Model 5G

8 September 2014 M. Berlin et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2014) xxx–xxx

Table 2 Gastrointestinal and pharmacokinetic disposition parameters for the PBPK model.

462 463 464 465 466 467 468 469 470 471 472 473 474 475

476 478

Parameter

Fed state

Fasted state

Gastric volume Intestinal volume [37,38] Gastric emptying rate [11] Permeability [28] Aeff [37] Vd/F k12 k21 k10

750 ml 320 ml 4 kcal/min 4.2  106 cm/s 3.6 m2 225.6 l 0.218 0.020 0.116

250 ml 212 ml 2.8 h1

0.145 0.056 0.148

in vitro solubility, dissolution, supersaturation and precipitation results were integrated according to the described PBPK models. 2.9. Statistical comparison of predicted plasma profiles using different transfer experiment approaches To evaluate the influence of the experimental setup of the transfer model on the fasted state plasma profile predictions, results of different zero order transfer rates (9, 4 and 0.5 ml/min) were implemented into the supersaturation and precipitation PBPK model and compared with data obtained from first order transfer results. All predictions were compared with in vivo data using point estimates ratios of the bioequivalence test parameters AUC and Cmax and the difference factor ( f1), which is a model-independent comparator of plasma concentration time profiles. The calculation of the f1 value was performed using Eq. (6):

Pn  t¼1 jRt  T t j Pn  100 f1 ¼ t¼1 Rt

ð6Þ

solubility of cinnarizine[μg/ml]

6

10000 1000 100 10 1 1077

540

0.1

0.16

2.82

2.51

6.45

43.03

1.13

73.42

3.49

k ) .0 4) .8 nk lan 1:2 H5 H5 bla (PO bl p p 4) ( .8 b Fp 2p 2p -V2 5.0 6.5 SGF SGF H5 (PO SIF SSG IF-V IF-V pH p pH 2 S e S S a V F F FaS FaS 2 2 S S F -V -V Fa Fe SG SIF SIF SIF FeS FaS FeS FaS .0 : H2 p SGF FaS .6 H1

.0 H2

ank

.5 H6

Fig. 4. Solubility of cinnarizine in biorelevant media (logarithmic scaling). Solubility values in lg/ml along with the standard errors are shown on the bar for each medium.

with the corresponding blank buffer, a 17-fold increase in solubility is observed for the biorelevant medium. A previous solubility study of cinnarizine in the older version of FaSSIF (V1) confirms this observation [40]. FaSSIF contains a 3.75 times higher concentration of lecithin than FaSSIF-V2 and exhibits a 3.4 times higher solubilization of cinnarizine. Since FaSSIF-V2 and FaSSIF-V2(PO4) have similar compositions, apart from the buffer, differences in cinnarizine solubility between these media are not quite statistically significant (p > 0.05). Conditions of the fed state stomach are represented by FeSSGF pH 5.0, which contains fats and proteins due to the presence of 50% milk in the composition of the medium. Since cinnarizine is very lipophilic in its unionized form (log P  5.6) [22,41] it shows interactions with lipids leading to a significant increase of cinnarizine solubility [26,28,41]. Likewise, due to interactions of cinnarizine with fats and proteins in FeSSGF the solubility is considerably higher (43 lg/ml) than in the corresponding aqueous buffer (6.45 lg/ml). Cinnarizine exhibits relatively high solubility in the medium simulating the fed state intestinal milieu. Comparing to FaSSIF-V2 the higher solubility in FeSSIF-V2 can be explained by three factors. First, the pH in FeSSIF-V2 is considerably lower than in FaSSIF-V2 resulting in a higher extent of cinnarizine ionization. Second, the concentration of bile salts and lecithin is significantly higher in FeSSIF-V2. And third, the fed state medium contains monoglycerides and fatty acids, which may additionally contribute to cinnarizine solubilization. The solubility of cinnarizine in the medium representing the fasted state intestinal conditions after ‘‘gastric emptying’’ (mixture of FaSSGF pH 2.0 and FaSSIF-V2(PO4) in the ratio 1 to 2) is higher than in pure FaSSIF-V2(PO4). Since bile salt and lecithin concentrations are lower in the mixture, one would expect a lower solubility of cinnarizine. However, the combination of FaSSGF pH 2.0 and FaSSIF-V2(PO4) results in a lower pH (5.95) than that of FaSSIFV2(PO4) (pH = 6.5) leading to a higher degree of drug ionization and thus to a higher solubility.

507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522

484

where n is the number of time points, Rt is the plasma concentration of the in vivo profile at time point t, Tt is the plasma concentration of the simulated plasma profile at time point t [39]. The comparison was performed for plasma concentrations at time points obtained from the in vivo studies (up to 24 h). The f1 cut-off limit of 20% was used in this study.

485

2.10. Sensitivity and model validation analysis

486

495

A sensitivity analysis was performed to identify the pre-systemic parameters which are most important for the simulation of cinnarizine plasma profiles. For this purpose the fasted and fed state gastric emptying rate (GER) and effective permeability (Peff) values were varied in the range of 0.25–5 times GER and Peff. Fraction dissolved ( fd) values were varied between 0.25 and 1, and the precipitation constant (kp) values were varied 0.25–5-fold. To investigate the impact of dissolution on plasma profile predictions in the fed state, initial dissolution rates (z values) in FeSSGF and FeSSIF-V2 were varied 0.25–5-fold.

496

3. Results and discussion

497

3.1. Solubility of cinnarizine in biorelevant media

3.2. Dissolution results of cinnarizine tablets in biorelevant media

543

498

Solubility results of cinnarizine in biorelevant media are presented in Fig. 4. Earlier studies have shown that the solubility of cinnarizine is highly pH dependent [26], which is confirmed by this study. The solubility in FaSSGF pH 1.6 and FaSSGF pH 2.0, which represent the fasted stomach, is high (1077 lg/ml and 540 lg/ml respectively). In less acidic media such as FaSSIF-V2 pH 6.5, which characterizes the fasted state upper intestinal environment, cinnarizine solubility is much lower. The compound exhibits high bile salt and lecithin dependent solubility. Comparing the FaSSIF-V2

Dissolution profiles of ArlevertÒ (20 mg) and StugeronÒ (25 mg) tablets in fasted state gastric and intestinal media, FaSSGF pH 1.6 and FaSSIF-V2, are presented in Fig. 5. For clarity the theoretical percentage dissolved for 20 mg cinnarizine in FaSSIF-V2, based on the API solubility in this medium, is included. The results are consistent with solubility data in Section 3.1. As expected, both formulations exhibited a rapid drug release and complete dissolution in FaSSGF pH 1.6 (ArlevertÒ tablets within 15 min and StugeronÒ tablets within 30 min). In FaSSIF-V2 the dissolution was slow,

544

479 480 481 482 483

487 488 489 490 491 492 493 494

499 500 501 502 503 504 505 506

Please cite this article in press as: M. Berlin et al., Prediction of oral absorption of cinnarizine – A highly supersaturating poorly soluble weak base with borderline permeability, Eur. J. Pharm. Biopharm. (2014), http://dx.doi.org/10.1016/j.ejpb.2014.08.011

523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542

545 546 547 548 549 550 551 552

EJPB 11699

No. of Pages 12, Model 5G

8 September 2014 7

100

100

80

80

% dissolved

% dissolved

M. Berlin et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2014) xxx–xxx

60 40 20

0 0

10

20

30

40

50

60

me [minutes]

554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585

40 20

0

553

60

0

30

60

90

120

150

180

210

240

me [minutes]

Fig. 5. Dissolution of ArlevertÒ and StugeronÒ in fasted state gastric and intestinal media (some symbols are bigger than the standard deviation bars): ArlevertÒ (d) and StugeronÒ (j) in FaSSGF pH 1.6 and ArlevertÒ (s) and Cinnarizine 30 mg tablets (h) in FaSSIF-V2 (curves are overlapping) along with the theoretical % dissolved for 20 mg cinnarizine in FaSSIF-V2, based on its solubility (- -).

Fig. 6. Dissolution of cinnarizine tablets in the fed gastric and intestinal state biorelevant media (some symbols are bigger than the standard deviation bars): ArlevertÒ (d) and Cinnarizine 30 mg tablets (N) in FeSSGF and ArlevertÒ (s) and Cinnarizine 30 mg tablets (4) in FeSSIF-V2 along with the theoretical % dissolved for 30 mg cinnarizine in FeSSGF, based on its solubility (- -).

releasing only 5.6% cinnarizine from both products and still climbing toward the solubility values after 240 min. The different dissolution behaviors of the two formulations in FaSSGF pH 1.6 and FaSSIF-V2 are reflected by the z values reported in Table 3. Interestingly, ArlevertÒ (20 mg) exhibits comparable plasma profiles to StugeronÒ (25 mg) in vivo despite the dose difference. This observation is qualitatively in accordance with the higher z values of ArlevertÒ in FaSSGF pH 1.6 and FaSSIF-V2. Fig. 6 shows dissolution results of ArlevertÒ 20 mg and Cinnarizine 30 mg tablets in the fed state media, FeSSGF and FeSSIF-V2. In FeSSGF, the fed gastric medium, 47% of the ArlevertÒ 20 mg dose was released after 240 min. With a solubility of 43 lg/ml of cinnarizine in FeSSGF (see Fig. 4), 100% dissolution of this dose would be possible in 500 ml FeSSGF if sufficient time was available. The dissolution rate of Cinnarizine 30 mg tablets was significantly slower, releasing only 27% of the dose in 240 min. This might be partially attributed to the higher dose (30 mg), which precludes more than 71% dissolution in 500 ml of FeSSGF. In FeSSIF-V2, the fed intestinal medium, both formulations exhibited faster and more extensive dissolution than in FeSSGF. The higher dosed Cinnarizine 30 mg tablets exhibited lower percentage dissolution values than ArlevertÒ (55% vs. 74%).

AUCfasted = 1.44). Indeed, if only intestinal data were to be considered in the in silico model, predictions would underestimate the in vivo performance of ArlevertÒ in the fasted state (see also Section 3.4). As previous observations from in vitro transfer experiments indicated that smaller than expected food effects may be attributable to supersaturation of cinnarizine in the intestinal compartment in the fasted state [25], the transfer model results were applied in the simulations. It is therefore advantageous to investigate the sequential dissolution, supersaturation and precipitation behavior of cinnarizine formulations under fasted gastric and intestinal conditions by implementing the transfer model. Fig. 7 shows cinnarizine concentrations in the acceptor phase during transfer experiments at 3 h1 (A), 9 ml/min (B), 4 ml/min (C) and 0.5 ml/min (D) transfer rates. In each diagram, the theoretical cumulative concentration curve of cinnarizine in the acceptor compartment, representing the case in which all cinnarizine dissolves and stays in solution during transfer is also shown. For all transfer rates high supersaturation values were observed in vitro, reaching up to six times the solubility values. The values of supersaturation for zero order transfer rates seem to follow a pattern: the higher the transfer rate the higher the maximum concentration achieved in solution. During transfer the discrepancy between the theoretical and observed concentrations is partly due to continuing dissolution (since part of the drug will be transferred as solid material) and partly due to precipitation of drug from solution. However, after transfer is completed, it can be assumed that any decrease in concentration is due solely to precipitation. Thus, the kinetics of this part of the curve were fitted to describe the precipitation. The ability of ArlevertÒ over StugeronÒ to achieve higher concentrations in solution (Section 3.2) is supported by most of the transfer experiment results. The supersaturation ratio is generally higher for ArlevertÒ than for StugeronÒ. However, ArlevertÒ exhibits faster precipitation than StugeronÒ after transfer is completed (i.e. after time t0 ) at most transfer rates. Supersaturation and precipitation kinetics for each formulation and transfer rate are presented as fraction dissolved ( fd) and first order precipitation (kp) constants in Table 4.

586

3.4. PBPK simulation results of cinnarizine formulations

624

3.4.1. Investigation of permeability restrictions of cinnarizine Due to the good solubility of cinnarizine in both FeSSGF and FeSSIF-V2 and a slow gastric emptying rate in the fed state, complete drug release of cinnarizine tablets (ArlevertÒ and Cinnarizine 30 mg) in the upper GI tract is expected. Precipitation is not likely to occur in the fed state since solubility in FeSSIF-V2 is higher than the solubility in FeSSGF. For these reasons, the dissolution-only

625

3.3. Investigation of supersaturation and precipitation of cinnarizine during transfer experiments Comparison of the dissolution behavior of ArlevertÒ in FaSSIFV2 and FeSSIF-V2 reveals disparate maximum concentrations in these two media. Even though the results are in agreement with the positive food effect which is observed for ArlevertÒ, the in vitro behavior only qualitatively reflects the in vivo observations. The solubility ratio of cinnarizine in FeSSIF-V2/FaSSIF-V2 (26.04) and the quotient of maximum concentrations measured during dissolution in FeSSIF-V2 and FaSSIF-V2 (5.05) would both suggest a much higher food effect than is observed in vivo (AUCfed/ Table 3 Initial dissolution rates (z values) for the various biorelevant media and cinnarizine formulations. Dissolution medium

Formulation

z Value ± SE (ml/mg2/3/h)

FeSSGF FeSSIF-V2 FaSSGF pH 1.6 FaSSIF-V2 FeSSGF FeSSIF-V2 FaSSGF pH 1.6 FaSSIF-V2

ArlevertÒ ArlevertÒ ArlevertÒ ArlevertÒ Cinnarizine 30 mg Cinnarizine 30 mg StugeronÒ StugeronÒ

0.110 ± 0.01 0.173 ± 0.004 5.16 ± 0.536 0.200 ± 0.016 0.062 ± 0.003 0.081 ± 0.001 2.28 ± 0.324 0.136 ± 0.003

Please cite this article in press as: M. Berlin et al., Prediction of oral absorption of cinnarizine – A highly supersaturating poorly soluble weak base with borderline permeability, Eur. J. Pharm. Biopharm. (2014), http://dx.doi.org/10.1016/j.ejpb.2014.08.011

587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623

626 627 628 629 630 631

EJPB 11699

No. of Pages 12, Model 5G

8 September 2014 8

M. Berlin et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2014) xxx–xxx 100

100

A

80

% dissolved

80

% dissolved

B

60 40 20

60 40 20 0

0 0

30

60

90

120

150

180

0

30

60

100

100

C

80

120

150

180

D

80

% dissolved

% dissolved

90

me [minutes]

me [minutes]

60 40 20

60 40 20

0

0 0

30

60

90

120

150

180

0

60

120

180

me [minutes]

240

300

360

420

480

me [minutes]

Fig. 7. Supersaturation and precipitation of ArlevertÒ (s) and StugeronÒ (d) cinnarizine tablets at 3 h1 (A), 9 ml/min (B), 4 ml/min (C) and 0.5 ml/min (D) transfer rates. The theoretical transfer curve (- -) is displayed along with the solubility (– –) of 25 mg cinnarizine in the combination of FaSSGF pH 2.0 and FaSSIF-V2(PO4) in the ratio 1:2.

Table 4 Fraction dissolved ( fd) and precipitation (kp) constants for ArlevertÒ and StugeronÒ for different transfer rates. Transfer rate First order (3 h1) First order (3 h1) 9 ml/min 9 ml/min 4 ml/min 4 ml/min 0.5 ml/min 0.5 ml/min

Formulation

fd

kp (h1)

Decrease in concentration at t (h)

0.7685

0.4997

1

StugeronÒ

0.5186

0.1980

1.17

ArlevertÒ StugeronÒ ArlevertÒ StugeronÒ ArlevertÒ StugeronÒ

0.7479 0.5202 0.5824 0.6101 0.7880 0.7460

0.6742 0.1886 0.4643 0.4732 0.4153 0.3537

0.5 0.5 1.167 1.167 4 5

Arlevert

Ò

PBPK model was used to evaluate the influence of drug permeability on the plasma profile predictions in the fed state. Initial dissolution rates (z values) in FeSSGF and FeSSIF-V2 as well as the corresponding cinnarizine solubility values in both media were implemented into one version of the model, in which permeability limitations were disregarded. In a second version the permeability was restricted. The implementation of permeability limitations into the dissolution-only PBPK model exhibited good fits with

632 633 634 635 636 637 638 639

Q4

respect to AUC, Cmax and tmax as shown in Table 5 (profiles are shown in Supplementary materials). By contrast, a lack of permeability restriction results in substantial overestimation of the Cmax and an earlier appearance of tmax at both dosages of cinnarizine. The permeability of BCS class II drugs in STELLAÒ models is usually assumed not to be limited [11]. However, Markopoulos et al. have recently suggested that drug association into micelles in the biorelevant media effectively reduce the permeability [44]. A STELLAÒ model of aprepitant, which is at the best a BCS class II compound (Caco-2 permeability of 7.8  106 cm/s) required permeability restrictions to accurately predict the oral absorption of the drug [10,42,43]. Further, Wagner et al. [19] were able to predict the oral absorption of Compound A, a BCS class IV drug with a Caco-2 permeability of 0.8  106 cm/s, using permeability restrictions in the STELLAÒ model. Cinnarizine, which is assumed to be a borderline BCS class II/IV drug, shows a Caco-2 permeability of 4.2  106 cm/s. As for Compound A, effective human permeability was calculated by directly introducing the Caco-2 data (conducted in buffer solution) into Eq. (5) [19]. Although the Caco-2 methodology may have varied among these compounds, it is likely that the permeability of cinnarizine lies in between the values of aprepitant and Compound A.

Table 5 Comparison of in vivo data with in silico predictions of different PBPK models for ArlevertÒ, Cinnarizine 30 mg tablets and StugeronÒ in the fed and fasted state. Fed state

Value

Fasted state

AUC0–24 (ng  h/ml) Cmax (ng/ml) ArlevertÒ In vivo 315.22 Predicted (restricted permeability) 315.38 Predicted (unrestricted permeability) 324.3

35.69 35.83 50.18

Cinnarizine 30 mg In vivo 466.50 Predicted (restricted permeability) 477.46 Predicted (unrestricted permeability) 485.40

55.68 55.42 77.61

Value

tmax (h)

AUC0–24 (ng  h/ml) Cmax (ng/ml)

tmax (h)

5.000 In vivo 4.630 Predicted (dissolution-only) 4.000 Predicted (supersaturation and precipitation)

218.72 332.24 238.06

33.00 24.69 33.63

2.000 2.125 1.750

232.81 331.05 214.71

29.87 18.54 29.18

2.000 2.250 1.875

StugeronÒ 5.000 In vivo 5.000 Predicted (dissolution-only) 4.125 Predicted (supersaturation and precipitation)

Please cite this article in press as: M. Berlin et al., Prediction of oral absorption of cinnarizine – A highly supersaturating poorly soluble weak base with borderline permeability, Eur. J. Pharm. Biopharm. (2014), http://dx.doi.org/10.1016/j.ejpb.2014.08.011

640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662

EJPB 11699

No. of Pages 12, Model 5G

8 September 2014 9

M. Berlin et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2014) xxx–xxx

demonstrated with the slowest transfer rate (0.5 ml/min). Overall, the first order transfer rate (3 h1) revealed best predictions of oral absorption of immediate release (IR) cinnarizine formulations in the fasted state.

727

3.6. Sensitivity analysis

731

The sensitivity analysis is shown for the fasted and fed state profiles of ArlevertÒ in Figs. 10 and 11. For the fasted state (Fig. 10) the plasma profile predictions are sensitive to gastric emptying rate (GER) shown in Panel A. Panel B reveals that the effective permeability (Peff) is also an influential factor for predictions of cinnarizine plasma profiles. Additionally, the fraction dissolved constant ( fd) influences the plasma profile forecasts (Panel C). Variations of the constant fd represent changes in the degree of drug supersaturation during gastric emptying, and thus in the concentration of dissolved API available for absorption from the intestinal lumen into the blood stream. By contrast, the

732

708

3.4.2. ‘‘Dissolution-only PBPK model’’ vs. ‘‘supersaturation and precipitation PBPK model’’ for modeling fasted state behavior Initially the dissolution-only PBPK model was used in order to simulate the fasted state plasma profiles of ArlevertÒ and StugeronÒ. In this model, the permeability was assumed to be restricted, since permeability constraints were necessary to accurately simulate the pharmacokinetics in the fed state (see Section 3.4.1). Analogously to the fed state simulations the model uses dissolution and solubility data, but does not consider supersaturation and precipitation of cinnarizine in the intestine. In a recent study the STELLAÒ software was used to evaluate the effects of possible supersaturation and precipitation on oral absorption of a nelfinavir, a weak base [18]. The investigators implemented their assumptions into the model and could demonstrate qualitatively that a consideration of supersaturation and precipitation is important to accurately predict oral drug absorption of nelfinavir in the fasted state. Thus, simulations of the fasted state were additionally conducted using the supersaturation and precipitation PBPK model. This model introduces fraction dissolved (fd) and first order precipitation (kp) constants to predict the supersaturation and precipitation of cinnarizine in the intestine. Using the example of ArlevertÒ, predictions of supersaturation and precipitation of cinnarizine utilizing fd (0.7685) and kp (0.4997 h1), obtained from first order transfer experiments, and the gastric emptying rate of 2.8 h1 were conducted (shown in Supplementary materials). Dissolution and re-dissolution of nondissolved and precipitated cinnarizine was simulated using solubility data (2.82 lg/ml) as well as the z value of ArlevertÒ (0.2 mg/ml2/3/h) in FaSSIF-V2. Fig. 8 presents the predictions of fasted state plasma profiles of ArlevertÒ and StugeronÒ comparing the two models. It demonstrates the significance of supersaturation and precipitation during and following gastric emptying (transfer rate = 3 h1) in the fasted state. The dissolution-only PBPK model underestimates the Cmax and overestimates the AUC of both formulations, while the predictions of supersaturation and precipitation model result in a good fit (Table 5). The simulation outcome demonstrates that dissolution experiments and solubility values deliver useful data for quantitative predictions of the plasma profiles of cinnarizine when supersaturation and precipitation of the drug is not expected i.e. in the fed state. In contrast, when supersaturation and precipitation is expected to occur i.e. in the fasted state, the use of transfer experiment data is essential to quantitatively assess oral absorption of this weak base.

709

3.5. Statistical comparison analysis of different transfer rates

710

Table 6 shows the comparison of fasted state in vivo data with in silico predictions using the results of various transfer experiments. Plasma predictions of ArlevertÒ and StugeronÒ at each transfer rate are presented in Fig. 9. Results of the first order transfer rate lead to good plasma profile predictions for both ArlevertÒ and StugeronÒ. Interestingly, plasma profiles of ArlevertÒ matched well with the predictions using the supersaturation and precipitation values from the transfer rate of 4 ml/min. A comparison of the first order and 4 ml/min transfer profiles of ArlevertÒ reveals several matching factors which may lead to similar predictions (i.e. the transfer time for both transfer rates lasted for about 1 h, after 1 h approximately the same amount of cinnarizine was released (60%) and the precipitation constants were similar for both transfer rates). However, in the case of StugeronÒ the use of the transfer rate of 4 ml/min lead to poor in vivo predictions. The transfer rate of 9 ml/min overestimated the AUC and Cmax of both formulations. Very poor predictions of the in vivo performance were

Fig. 8. Fasted state plasma profile predictions of ArlevertÒ (A) and StugeronÒ (B) tablets using the dissolution – only and supersaturation and precipitation (obtained from transfer experiments at a rate of 3 h1) PBPK models.

665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707

711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726

in in vivo vivo- -Arlevert® Arlevert® in silico silico- -supersaturaon supersaturaon andand precipitaon precipitaon PBPK PBPK model model in silico silico- -dissoluon-only dissoluon-only PBPK PBPK model model

A 35

concentraon [ng/ml]

664

30 25 20 15 10 5 0 0

2

4

6

8

10

12

14

16

18

20

22

24

22

24

me [h] in vivo vivo- -Stugeron® Stugeron® in-silico- -supersaturaon in-silico supersaturaon andand precipitaon precipitaon PBPK model PBPK mod in silico silico- -dissoluon-only dissoluon-only PBPK PBPK model model

B 35

concentraon [ng/ml]

663

30 25 20 15 10 5 0 0

2

4

6

8

10

12

14

16

18

20

me [h]

Table 6 Statistical comparison of in silico predictions to in vivo observations using the difference factor (f1) and point estimate ratios of bioequivalence test parameters for different IR formulations of cinnarizine and transfer experiments (best fits in boldface). Fasted state

ArlevertÒ (20 mg, 3 h1) ArlevertÒ (20 mg, 9 ml/min) ArlevertÒ (20 mg, 4 ml/min) ArlevertÒ (20 mg, 0.5 ml/min) StugeronÒ (25 mg, 3 h1) StugeronÒ (25 mg, 9 ml/min) StugeronÒ (25 mg, 4 ml/min) StugeronÒ (25 mg, 0.5 ml/min)

Value Ratio AUC0–24

Ratio Cmax

f1

1.095 1.207 1.019 1.493 0.922 1.095 1.221 1.783

1.019 1.179 0.953 0.757 0.977 1.170 1.327 0.971

13.84 21.80 12.14 46.00 10.31 21.78 30.30 60.27

Please cite this article in press as: M. Berlin et al., Prediction of oral absorption of cinnarizine – A highly supersaturating poorly soluble weak base with borderline permeability, Eur. J. Pharm. Biopharm. (2014), http://dx.doi.org/10.1016/j.ejpb.2014.08.011

728 729 730

733 734 735 736 737 738 739 740 741 742

EJPB 11699

No. of Pages 12, Model 5G

8 September 2014 10

M. Berlin et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2014) xxx–xxx 45

A

40

40

35

35

concentraon [ng/ml]

concentraon [ng/ml]

45

30 25 20 15 10

B

30 25 20 15 10

5

5 0

0 2

0

4

6

8

10

12

14

16

18

20

22

24

0

2

4

6

8

10

me [h] 45

C

40

40

35

35

concentraon [ng/ml]

concentraon [ng/ml]

45

12

14

16

18

20

22

24

14

16

18

20

22

24

me [h]

30 25 20 15 10

D

30 25 20 15 10

5

5

0

0 0

4

2

6

8

10

12

14

16

18

20

22

24

0

2

4

6

8

10

me [h]

12

me [h]

Fig. 9. Fasted state plasma profile predictions of ArlevertÒ ( ) and StugeronÒ ( ) tablets using the supersaturation and precipitation PBPK model after implementing results of different transfer rates. The in vivo plasma profiles of ArlevertÒ (s) and StugeronÒ (d) are displayed in against profile predictions for each transfer rate, i.e. 3 h1 (A), 9 ml/ min (B), 4 ml/min (C) and 0.5 ml/min (D). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

reference reference

(0.25 -- 5) (0.25 5)xxGER GER

B

Arlevert® 20 Arlevert® 20mg mgininvivo vivo

50 45 40 35 30 25 20 15 10 5 0

A

0

2

4

6

8

10

12

14

16

18

20

22

concentraon [ng/ml]

concentraon [ng/ml]

A

reference reference

(0.25 -- 5) (0.25 5)xxPeff Peff

B

0

24

2

4

6

8

10

12

me [h] reference reference

fd (0.25 (0.25 --1) 1)

D

Arlevert® 20 Arlevert® 20mg mgininvivo vivo

50 45 40 35 30 25 20 15 10 5 0

C

0

2

4

6

8

10

12

14

16

18

20

22

24

me [h]

14

16

18

20

22

24

me [h]

concentraon [ng/ml]

concentraon [ng/ml]

C

Arlevert® 20 Arlevert® 20mg mgininvivo vivo

50 45 40 35 30 25 20 15 10 5 0

reference reference

(0.25 -- 5) (0.25 5)xxkp kp

Arlevert® 20 Arlevert® 20mg mgininvivo vivo

50 45 40 35 30 25 20 15 10 5 0

D

0

2

4

6

8

10

12

14

16

18

20

22

24

me [h]

Fig. 10. Sensitivity analysis of the gastric emptying rate (A), the effective surface area (B), the fraction dissolved constant (C) and precipitation constant (D) on cinnarizine plasma profiles in the fasted state. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

743 744 745 746 747 748 749

precipitation constant represents the reduction rate of dissolved API in the intestine. Although changes in the precipitation constant influence the plasma predictions (Panel D), it has less influence than variations in fd on the prediction of the cinnarizine plasma profile. In the fed state variations in GER and Peff show consistency with the fasted state observations (Panels A and B). Variation in initial

dissolution rates (z values) in FeSSGF and FeSSIF-V2 had very little to no impact on pharmacokinetic predictions (Panel C and D). The negligible role of the z values on the in vivo pharmacokinetic predictions of cinnarizine is explained by the high solubility of the drug in the fed state media. The sensitivity analysis further demonstrates that the plasma predictions of cinnarizine highly depend on permeability, in

Please cite this article in press as: M. Berlin et al., Prediction of oral absorption of cinnarizine – A highly supersaturating poorly soluble weak base with borderline permeability, Eur. J. Pharm. Biopharm. (2014), http://dx.doi.org/10.1016/j.ejpb.2014.08.011

750 751 752 753 754 755 756

EJPB 11699

No. of Pages 12, Model 5G

8 September 2014 11

M. Berlin et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2014) xxx–xxx

A

reference reference

(0.25 -- 5) (0.25 5)xxGER GER

B

in vivo vivo

A

50

concentraon [ng/ml]

concentraon [ng/ml]

reference reference

(0.25 -- 5) (0.25 5)xxPeff Peff

40 30 20 10

B

50 40 30 20 10 0

0 0

2

4

6

8

10

12

14

16

18

20

22

0

24

2

4

6

8

me [h] D reference reference

10

12

14

16

18

20

22

24

me [h]

C (0.25 -- 5) (0.25 5)xxzzFeSSGF FeSSGF

reference reference

in vivo vivo

(0.25 -- 5) (0.25 5)xxzzFeSSIF-V2 FeSSIF-V2

in vivo vivo

60

60

C

50

concentraon [ng/ml]

concentraon [ng/ml]

in vivo vivo

60

60

40 30 20 10

D

50 40 30 20 10 0

0 0

2

4

6

8

10

12

14

16

18

20

22

24

0

2

me [h]

4

6

8

10

12

14

16

18

20

22

24

me [h]

Fig. 11. Sensitivity analysis of the gastric emptying rate (A), the effective surface area (B), the z value in FeSSGF (C) and the z value in FeSSIF-V2 (D) on cinnarizine plasma profiles in the fed state. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

765

accordance with its borderline BCS class II/IV drug status. It also confirms that supersaturation and precipitation play a crucial role for the bioavailability of cinnarizine the fasted state. Generally speaking, the sensitivity analysis reveals that the variability in plasma profiles under fasted conditions is expected to be higher than in the fed state. This is also in accordance with the in vivo observations. While the coefficient of variation for AUC in the fasted state was high (50%), for the fed state it was significantly lower (27%) [31].

766

4. Conclusions

767

786

Implementation of dissolution test results in a physiologically based pharmacokinetic (PBPK) model resulted in an accurate prediction of the fed state plasma profiles of ArlevertÒ and Cinnarizine 30 mg tablets. A precise in vivo plasma forecast of ArlevertÒ and StugeronÒ in the fasted state could only be achieved by introducing transfer experiment results into an advanced PBPK model, i.e. it was necessary to consider both supersaturation and precipitation in the small intestine as well as the dissolution characteristics. Moreover, integration of permeability restrictions was essential to accurately predict the in vivo performance of all dosage forms of cinnarizine in both the fed and the fasted state. The PBPK model developed in this work makes it possible to simulate plasma profiles of poorly soluble weak bases which have a tendency to supersaturate and then precipitate during gastric emptying. The model enables the evaluation of formulation differences and food effects on the in vivo performance during formulation development. It also enables an implementation of permeability restrictions based on Caco-2 data in the PBPK model, which is essential for forecasting in vivo performance of borderline BCS class II/IV and BCS class IV drugs.

787

Appendix A. Supplementary material

788

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ejpb.2014.08.011.

757 758 759 760 761 762 763 764

768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785

789

References

790

[1] G.L. Amidon, H. Lennernas, V.P. Shah, J.R. Crison, A theoretical basis for a biopharmaceutic drug classification: the correlation of in vitro drug product dissolution and in vivo bioavailability, Pharm. Res. 12 (1995) 413–420. [2] E. Galia, E. Nicolaides, D. Horter, R. Lobenberg, C. Reppas, J.B. Dressman, Evaluation of various dissolution media for predicting in vivo performance of class I and II drugs, Pharm. Res. 15 (1998) 698–705. [3] M. Vertzoni, J. Dressman, J. Butler, J. Hempenstall, C. Reppas, Simulation of fasting gastric conditions and its importance for the in vivo dissolution of lipophilic compounds, Eur. J. Pharm. Biopharm. 60 (2005) 413–417. [4] E. Jantratid, N. Janssen, C. Reppas, J.B. Dressman, Dissolution media simulating conditions in the proximal human gastrointestinal tract: an update, Pharm. Res. 25 (2008) 1663–1676. [5] C. Reppas, M. Vertzoni, Biorelevant in-vitro performance testing of orally administered dosage forms, J. Pharm. Pharmacol. 64 (2012) 919–930. [6] K. Kleberg, J. Jacobsen, A. Mullertz, 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 (2010) 1656–1668. [7] J.B. Dressman, C. Reppas, In vitro–in vivo correlations for lipophilic, poorly water-soluble drugs, Eur. J. Pharm. Sci. 11 (Suppl. 2) (2000) S73–S80. [8] K. Thelen, E. Jantratid, J.B. Dressman, J. Lippert, S. Willmann, Analysis of nifedipine absorption from soft gelatin capsules using PBPK modeling and biorelevant dissolution testing, J. Pharm. Sci. 99 (2010) 2899–2904. [9] Y. Shono, E. Jantratid, N. Janssen, F. Kesisoglou, Y. Mao, M. Vertzoni, C. Reppas, J.B. Dressman, Prediction of food effects on the absorption of celecoxib based on biorelevant dissolution testing coupled with physiologically based pharmacokinetic modeling, Eur. J. Pharm. Biopharm. 73 (2009) 107–114. [10] Y. Shono, E. Jantratid, F. Kesisoglou, C. Reppas, J.B. Dressman, Forecasting in vivo oral absorption and food effect of micronized and nanosized aprepitant formulations in humans, Eur. J. Pharm. Biopharm. 76 (2010) 95–104. [11] E. Nicolaides, M. Symillides, J.B. Dressman, C. Reppas, Biorelevant dissolution testing to predict the plasma profile of lipophilic drugs after oral administration, Pharm. Res. 18 (2001) 380–388. [12] D. Juenemann, E. Jantratid, C. Wagner, C. Reppas, M. Vertzoni, J.B. Dressman, Biorelevant in vitro dissolution testing of products containing micronized or nanosized fenofibrate with a view to predicting plasma profiles, Eur. J. Pharm. Biopharm. 77 (2011) 257–264. [13] R. Takano, K. Sugano, A. Higashida, Y. Hayashi, M. Machida, Y. Aso, S. Yamashita, Oral absorption of poorly water-soluble drugs: computer simulation of fraction absorbed in humans from a miniscale dissolution test, Pharm. Res. 23 (2006) 1144–1156. [14] E.S. Kostewicz, M. Wunderlich, U. Brauns, R. Becker, T. Bock, J.B. Dressman, Predicting the precipitation of poorly soluble weak bases upon entry in the small intestine, J. Pharm. Pharmacol. 56 (2004) 43–51. [15] Y.Y. Yeap, N.L. Trevaskis, C.J. Porter, The potential for drug supersaturation during intestinal processing of lipid-based formulations may be enhanced for basic drugs, Mol. Pharm. (2013).

791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837

Please cite this article in press as: M. Berlin et al., Prediction of oral absorption of cinnarizine – A highly supersaturating poorly soluble weak base with borderline permeability, Eur. J. Pharm. Biopharm. (2014), http://dx.doi.org/10.1016/j.ejpb.2014.08.011

EJPB 11699

No. of Pages 12, Model 5G

8 September 2014 12 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884

Q3

M. Berlin et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2014) xxx–xxx

[16] Y.L. Hsieh, G.A. Ilevbare, B. Van Eerdenbrugh, K.J. Box, M.V. Sanchez-Felix, L.S. Taylor, PH-induced precipitation behavior of weakly basic compounds: determination of extent and duration of supersaturation using potentiometric titration and correlation to solid state properties, Pharm. Res. 29 (2012) 2738–2753. [17] E.S. Kostewicz, L. Aarons, M. Bergstrand, M.B. Bolger, A. Galetin, O. Hatley, M. Jamei, R. Lloyd, X. Pepin, A. Rostami-Hodjegan, E. Sjogren, C. Tannergren, D.B. Turner, C. Wagner, W. Weitschies, J. Dressman, PBPK models for the prediction of in vivo performance of oral dosage forms, Eur. J. Pharm. Sci. (2013). [18] Y. Shono, E. Jantratid, J.B. Dressman, Precipitation in the small intestine may play a more important role in the in vivo performance of poorly soluble weak bases in the fasted state: case example nelfinavir, Eur. J. Pharm. Biopharm. 79 (2011) 349–356. [19] C. Wagner, E. Jantratid, F. Kesisoglou, M. Vertzoni, C. Reppas, J.B. Dressman, Predicting the oral absorption of a poorly soluble, poorly permeable weak base using biorelevant dissolution and transfer model tests coupled with a physiologically based pharmacokinetic model, Eur. J. Pharm. Biopharm. 82 (2012) 127–138. [20] A. Kambayashi, J.B. Dressman, An in vitro–in silico–in vivo approach to predicting the oral pharmacokinetic profile of salts of weak acids: case example dantrolene, Eur. J. Pharm. Biopharm. 84 (2013) 200–207. [21] T. Jarvinen, K. Jarvinen, N. Schwarting, V.J. Stella, Beta-cyclodextrin derivatives, SBE4-beta-CD and HP-beta-CD, increase the oral bioavailability of cinnarizine in beagle dogs, J. Pharm. Sci. 84 (1995) 295–299. [22] K. Peeters, R. De Maesschalck, H. Bohets, K. Vanhoutte, L. Nagels, In situ dissolution testing using potentiometric sensors, Eur. J. Pharm. Sci. 34 (2008) 243–249. [23] S. Branchu, P.G. Rogueda, A.P. Plumb, W.G. Cook, A decision-support tool for the formulation of orally active, poorly soluble compounds, Eur. J. Pharm. Sci. 32 (2007) 128–139. [24] C.A. Bergstrom, C.M. Wassvik, K. Johansson, I. Hubatsch, Poorly soluble marketed drugs display solvation limited solubility, J. Med. Chem. 50 (2007) 5858–5862. [25] C.H. Gu, D. Rao, R.B. Gandhi, J. Hilden, K. Raghavan, Using a novel multicompartment dissolution system to predict the effect of gastric pH on the oral absorption of weak bases with poor intrinsic solubility, J. Pharm. Sci. 94 (2005) 199–208. [26] H. Ogata, N. Aoyagi, N. Kaniwa, A. Ejima, N. Sekine, K. Kitamura, M. Inoue, Gastric acidity dependent bioavailability of cinnarizine from two commercial capsules in healthy volunteers, Int. J. Pharm. 29 (1986). [27] I. Yamada, T. Goda, M. Kawata, K. Ogawa, Application of gastric aciditycontrolled beagle dog to bioavailability study of cinnarizine, Yakugaku Zasshi 110 (1990) 280–285. [28] A.T. Larsen, P. Akesson, A. Jureus, L. Saaby, R. Abu-Rmaileh, B. Abrahamsson, J. Ostergaard, A. Mullertz, Bioavailability of cinnarizine in dogs: effect of SNEDDS loading level and correlation with cinnarizine solubilization during in vitro lipolysis, Pharm. Res. (2013).

[29] N. Koumandrakis, M. Vertzoni, C. Reppas, Increasing the biorelevance of simulated intestinal fluids for better predictions of drug equilibrium solubility in the fasted upper small intestine, in: 9th World Meeting on Pharmaceutics, Biopharmaceutics and Pharmaceutical Technology, Lisbon, Portugal, 31 March–3 April, 2014. [30] Y. Fei, E.S. Kostewicz, M.T. Sheu, J.B. Dressman, Analysis of the enhanced oral bioavailability of fenofibrate lipid formulations in fasted humans using an in vitro–in silico–in vivo approach, Eur. J. Pharm. Biopharm. (2013). [31] Hennig Arzneimittel – Data on file. [32] H. Nowacka-Krukowska, M. Rakowska, K. Neubart, M. Kobylinska, Highperformance liquid chromatographic assay for cinnarizine in human plasma, Acta Pol. Pharm. 64 (2007) 407–411. [33] Guideline on the Investigation of Bioequivalence, European Medicines Agency, 2010. [34] K. Kiehm, Development of a Novel Screening Tool for the Prediction of Oral Drug Absorption based on Surface Activity Profiling, Shaker Verlag, 2009. [35] A.C. Petereit, K. Swinney, J. Mensch, C. Mackie, S. Stokbroekx, M. Brewster, J.B. Dressman, Prediction of blood–brain barrier penetration of poorly soluble drug candidates using surface activity profiling, Eur. J. Pharm. Biopharm. 75 (2010) 405–410. [36] G.M. Grass, Simulation models to predict oral drug absorption from in vitro data, Adv. Drug Deliv. Rev. 23 (1997) 199–219. [37] N. Fotaki, M. Symillides, C. Reppas, Canine versus in vitro data for predicting input profiles of L-sulpiride after oral administration, Eur. J. Pharm. Sci. 26 (2005) 324–333. [38] E.L. McConnell, H.M. Fadda, A.W. Basit, Gut instincts: explorations in intestinal physiology and drug delivery, Int. J. Pharm. 364 (2008) 213–226. [39] P. Costa, J.M. Sousa Lobo, Modeling and comparison of dissolution profiles, Eur. J. Pharm. Sci. 13 (2001) 123–133. [40] M. Lindenberg, A Biopharmaceutics Classification Scheme for the Development of New Drugs, Shaker Verlag, 2008. p. 126. [41] G.A. Kossena, W.N. Charman, B.J. Boyd, C.J. Porter, Influence of the intermediate digestion phases of common formulation lipids on the absorption of a poorly water-soluble drug, J. Pharm. Sci. 94 (2005) 481–492. [42] R. Angi, T. Solymosi, Z. Otvos, B. Ordasi, H. Glavinas, G. Filipcsei, G. Heltovics, F. Darvas, Novel continuous flow technology for the development of a nanostructured Aprepitant formulation with improved pharmacokinetic properties, Eur. J. Pharm. Biopharm. (2013). [43] Y. Wu, A. Loper, E. Landis, L. Hettrick, L. Novak, K. Lynn, C. Chen, K. Thompson, R. Higgins, U. Batra, S. Shelukar, G. Kwei, D. Storey, The role of biopharmaceutics in the development of a clinical nanoparticle formulation of MK-0869: a Beagle dog model predicts improved bioavailability and diminished food effect on absorption in human, Int. J. Pharm. 285 (2004) 135–146. [44] C. Markopoulos, F. Thoenen, D. Preisig, M. Symillides, M. Vertzoni, N. Parrott, C. Reppas, G. Imanidis, Biorelevant media for transport experiments in the Caco2 model to evaluate drug absorption in the fasted and the fed state and their usefulness, Eur. J. Pharm. Biopharm. 86 (2014) 438–448.

885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932

Please cite this article in press as: M. Berlin et al., Prediction of oral absorption of cinnarizine – A highly supersaturating poorly soluble weak base with borderline permeability, Eur. J. Pharm. Biopharm. (2014), http://dx.doi.org/10.1016/j.ejpb.2014.08.011