In vitro solubility, dissolution and permeability studies combined with semi-mechanistic modeling to investigate the intestinal absorption of desvenlafaxine from an immediate- and extended release formulation

PHASCI 3296

No. of Pages 11, Model 5G

16 June 2015 European Journal of Pharmaceutical Sciences xxx (2015) xxx–xxx 1

Contents lists available at ScienceDirect

European Journal of Pharmaceutical Sciences journal homepage: www.elsevier.com/locate/ejps 6 7

5

In vitro solubility, dissolution and permeability studies combined with semi-mechanistic modeling to investigate the intestinal absorption of desvenlafaxine from an immediate- and extended release formulation

8

F. Franek a,⇑, A. Jarlfors a,b, F. Larsen c, P. Holm d, B. Steffansen a

3 4

9 10 11 12 13 14 1 2 6 3 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

a

Faculty of Health and Medical Sciences, University of Copenhagen, Universitetsparken 2, Copenhagen, Denmark Department of Pharmacy, Uppsala University, Biomedical Centre, P.O. Box 580, SE-751 23 Uppsala, Sweden Clinical Pharmacology, H. Lundbeck A/S, Ottiliavej 9, DK-2500 Valby, Denmark d Biologics and Pharmaceutical Science, H. Lundbeck A/S, Ottiliavej 9, DK-2500 Valby, Denmark b c

a r t i c l e

i n f o

Article history: Received 12 November 2014 Received in revised form 31 May 2015 Accepted 15 June 2015 Available online xxxx Chemical compounds studied in this article: Desvenlafaxine succinate (PubChem CID: 6918664) Keywords: SimCyp pH-dependent permeability HPMC Biorelevant

a b s t r a c t Desvenlafaxine is a biopharmaceutics classification system (BCS) class 1 (high solubility, high permeability) and biopharmaceutical drug disposition classification system (BDDCS) class 3, (high solubility, poor metabolism; implying low permeability) compound. Thus the rate-limiting step for desvenlafaxine absorption (i.e. intestinal dissolution or permeation) is not fully clarified. The aim of this study was to investigate whether dissolution and/or intestinal permeability rate-limit desvenlafaxine absorption from an immediate-release formulation (IRF) and PristiqÒ, an extended release formulation (ERF). Semi-mechanistic models of desvenlafaxine were built (using SimCypÒ) by combining in vitro data on dissolution and permeation (mechanistic part of model) with clinical data (obtained from literature) on distribution and clearance (non-mechanistic part of model). The model predictions of desvenlafaxine pharmacokinetics after IRF and ERF administration were compared with published clinical data from 14 trials. Desvenlafaxine in vivo dissolution from the IRF and ERF was predicted from in vitro solubility studies and biorelevant dissolution studies (using the USP3 dissolution apparatus), respectively. Desvenlafaxine apparent permeability (Papp) at varying apical pH was investigated using the Caco-2 cell line and extrapolated to effective intestinal permeability (Peff) in human duodenum, jejunum, ileum and colon. Desvenlafaxine pKa-values and octanol–water partition coefficients (Do:w) were determined experimentally. Due to predicted rapid dissolution after IRF administration, desvenlafaxine was predicted to be available for permeation in the duodenum. Desvenlafaxine Do:w and Papp increased approximately 13-fold when increasing apical pH from 5.5 to 7.4. Desvenlafaxine Peff thus increased with pH down the small intestine. Consequently, desvenlafaxine absorption from an IRF appears rate-limited by low Peff in the upper small intestine, which ‘‘delays’’ the predicted time to the maximal plasma concentration (tmax), consistent with clinical data. Conversely, desvenlafaxine absorption from the ERF appears rate-limited by dissolution due to the formulation, which tends to negate the influence of pH-dependent permeability on absorption. We suggest that desvenlafaxine Peff is mainly driven by transcellular diffusion of the unionized form. In the case of desvenlafaxine, poor metabolism does not imply low intestinal permeability, as indicated by the BDDCS, merely low duodenal/jejunal permeability. Ó 2015 Published by Elsevier B.V.

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 58 59 60 61 62 63

65 64

1. Introduction

66

Desvenlafaxine is a serotonin–norepinephrine reuptake inhibitor (SNRI) used to treat depression (Hadfield et al., 2004; Kornstein et al., 2014; Pae, 2011).

67 68

⇑ Corresponding author.

Desvenlafaxine is described as a high solubility compound (FDA Approval Package: NDA 21-992, 2008; Hadfield et al., 2007) and is accordingly classified as such in the biopharmaceutics classification system (BCS) (Amidon et al., 1995; EMA, 2010; FDA, 2000) and the biopharmaceutics drug disposition classification system (BDDCS) (Benet et al., 2011; Wu and Benet, 2005). After oral- and intravenous administration of desvenlafaxine, approximately 50% of the administered dose is excreted unchanged

http://dx.doi.org/10.1016/j.ejps.2015.06.012 0928-0987/Ó 2015 Published by Elsevier B.V.

Please cite this article in press as: Franek, F., et al. In vitro solubility, dissolution and permeability studies combined with semi-mechanistic modeling to investigate the intestinal absorption of desvenlafaxine from an immediate- and extended release formulation. Eur. J. Pharm. Sci. (2015), http://dx.doi.org/ 10.1016/j.ejps.2015.06.012

69 70 71 72 73 74 75 76

PHASCI 3296

No. of Pages 11, Model 5G

16 June 2015 2

F. Franek et al. / European Journal of Pharmaceutical Sciences xxx (2015) xxx–xxx

120

via the urine (Nichols et al., 2012). This classifies desvenlafaxine as a poorly metabolized compound according to the BDDCS since <90% of the administered dose is metabolized (Benet et al., 2008). Consequently, desvenlafaxine permeability is implied to be low by the BDDCS because it is classified as a poorly metabolized compound (Benet et al., 2011; Wu and Benet, 2005). However, after oral administration of desvenlafaxine extended-release formulation (ERF), marketed as PristiqÒ, the absolute average bioavailability is 80%, ranging from 40% to 100% (Nichols et al., 2012). Due to rapid dissolution, the absolute bioavailability of desvenlafaxine from an immediate-release formulation (IRF) may be higher than from the ERF (although this has not, to the authors knowledge, been investigated). Desvenlafaxine might thus be classified as ‘‘highly permeable’’ according to the BCS if the %fraction absorbed is above 90% (Amidon et al., 1995; EMA, 2010; FDA, 2000). Consequently, desvenlafaxine given as an IRF may on one hand be classified as BCS class 1 (indicating high permeability), but on the other hand as BDDCS class 3 (indicating low permeability), which may appear unusual (Dahan et al., 2010; Liu et al., 2012). In vitro- and clinically-derived data describing absorption and disposition can be used as input into pharmacokinetic models to predict plasma concentration–time profiles (Jones et al., 2006; Patel et al., 2014). Fully mechanistic models include data from in vitro studies only, since in vitro studies often investigate and propose mechanisms influencing pharmacokinetics (Tsamandouras et al., 2015). Semi-mechanistic models however, combine in vitro and clinical data to predict plasma concentration–time profiles. In vitro solubility-, dissolution- and permeability studies can investigate and propose absorption mechanisms (Dahan et al., 2010; Neuhoff et al., 2003). Combined with clinically-derived data describing disposition, the in vitro-derived data on absorption may be used as input into a semi-mechanistic model in order to predict plasma concentration–time profiles (Jones et al., 2006; Patel et al., 2014). Subsequently, predicted and reported plasma concentration–time profiles may be compared in order to investigate whether the proposed absorption-mechanisms are supported by clinical studies (Jones et al., 2012; Neuhoff et al., 2013). The aim of the present study was to investigate whether intestinal permeability and/or dissolution is rate-limiting for intestinal absorption of desvenlafaxine from an IRF and ERF. This was done by conducting in vitro studies that investigate desvenlafaxine absorption and use the results as the ‘‘mechanistic part’’ to semi-mechanistically model and predict desvenlafaxine pharmacokinetics, which were finally compared to results from clinical studies.

121

2. Materials and methods

122

2.1. Materials

123

2.1.1. Formulation and chemicals 76 mg desvenlafaxine succinate, equal to 50 mg of desvenlafaxine, PristiqÒ hypromellose (HPMC)-based ERF (NDC 0008-1211-50), were obtained from Pfizer, United States. Desvenlafaxine succinate was obtained from Teva API, Israel. Potassium phosphate monohydrate (Sigma–Aldrich, United States), Sodium tetraborate decahydrate (Merck, Germany), Tris(h ydroxymethyl)-aminomethane (Sigma–Aldrich, United States) were of analytical grade and used to make buffers for solubility studies. Hank’s Balanced Salt Solution (HBSS) and sodium bicarbonate were purchased from Gibco, USA. DL-[4-3H]-propranolol hydrochloride (81.5 ci/mmol) was purchased from Amersham Life Science, UK. All other reagents were of standard grade or better.

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

124 125 126 127 128 129 130 131 132 133 134 135

136 137 138

2.1.2. Cell culture Human colon adenocarcinoma cell line from ‘‘Deutsche Sammlung von Mikroorganismen und Zellkulturen’’ (DSMZ)

(Braunschweig, Germany), were obtained at passages 3 and 4 (N = 2). Cells were seeded at 8.93  104 cells/cm2 at a seeding density of 1.7  105 cells/mL onto TranswellÒ polycarbonate membrane inserts (pore size 0.4 lm, 1.12 cm2 growth area) (Grandvuinet et al., 2013). Cells were grown in Dulbecco’s Modified Eagle’s medium (glucose 4.5 mg/mL) supplemented with

139

L-glutamine

(788 g/mL), penicillin (90 U/mL), streptomycin (90 g/mL), 10% fetal bovine serum, and 1% nonessential amino acids (alanine, asparagine, aspartic acid, glutamine acid, glycine, proline and serine). Experiments were carried out on the 18th day after seeding.

145

2.2. Semi-mechanistic modeling overview and input requirements

150

SimCypÒ (V13 R2, SimCypÒ, a Certara company, Sheffield, UK) was used to build five semi-mechanistic models, as described in Section 2.7. All models include a disposition sub-model (non-mechanistic model) and – in cases when the pharmacokinetics after oral administration are predicted – also an absorption sub-model (mechanistic model), as illustrated in Fig. 1 (Jamei et al., 2009a,b; Rowland Yeo et al., 2010; Yu et al., 2000). Model disposition (i.e. distribution and clearance) input were adopted from an intravenous clinical study reported in literature (Nichols et al., 2012) and is further described in Section 2.5. For models that predict desvenlafaxine pharmacokinetics after oral administration, in vitro investigated dissolution and permeability (described in Section 2.3) was extrapolated to in vivo (described in Section 2.4). In vitro results were thus used for the prediction of desvenlafaxine absorption from the IRF and ERF. Reported pharmacokinetic data that was used as comparison to model predictions of desvenlafaxine pharmacokinetics is further described in Section 2.6. In vitro and pharmacokinetic (PK) input data for desvenlafaxine semi-mechanistic models are given in Table 1, and were either collated from literature, experimentally determined or assumed. Assumed input data values are SimCypÒ defaults. These parameter values were subjected to sensitivity analyses to investigate their potential impact on the predicted desvenlafaxine pharmacokinetics.

151

2.3. In vitro studies to support modeling of desvenlafaxine absorption

176

2.3.1. Desvenlafaxine pH-dependent solubility Desvenlafaxine succinate pH-dependent solubility was investigated in triplicates by the shake flask method. Desvenlafaxine succinate was added with excess to phosphate-, tris- and borate-buffers with different pH (6.8–10) and incubated at 37 °C under agitation for at least 24 h. The pH of the suspension was adjusted until stable, filtered (0.22 lm Millex-GV Hydrophilic (PVDF)) at 37 °C, diluted 10 times with phosphate buffer (pH 6.8) and analyzed by HPLC (Section 2.3.5).

177

2.3.2. Biorelevant in vitro dissolution studies using the USP3 dissolution apparatus A biorelevant in vitro dissolution method was used to predict desvenlafaxine fasted-state in vivo dissolution from the ERF (Franek et al., 2014; Klein, 2005). The biorelevant dissolution method utilized the USP3 dissolution apparatus (Agilent Bio-Dis Reciprocating Cylinder apparatus, USA) with 6 vessels containing 250 mL biorelevant media, kept at a temperature of 37 ± 0.5 °C. The reciprocating cylinder with a top and bottom stainless steel screen size of 840 lm was used as tablet holder. It moves from one vessel to the next at pre-determined time-intervals and reciprocates in the biorelevant media at 10 dips per minute (dpm). The fasted-state biorelevant media (and time spent reciprocating in

186

Please cite this article in press as: Franek, F., et al. In vitro solubility, dissolution and permeability studies combined with semi-mechanistic modeling to investigate the intestinal absorption of desvenlafaxine from an immediate- and extended release formulation. Eur. J. Pharm. Sci. (2015), http://dx.doi.org/ 10.1016/j.ejps.2015.06.012

140 141 142 143 144

146 147 148 149

152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175

178 179 180 181 182 183 184 185

187 188 189 190 191 192 193 194 195 196 197 198

PHASCI 3296

No. of Pages 11, Model 5G

16 June 2015 F. Franek et al. / European Journal of Pharmaceutical Sciences xxx (2015) xxx–xxx

3

Fig. 1. A schematic illustration of desvenlafaxine semi-mechanistic models (left) and their data input requirements (right). All five of the semi-mechanistic models contain a disposition sub-model, which is divided into portal vein-, liver- and systemic-compartments interconnected by blood-flow (arrows between compartments). Elimination occurs via hepatic- and renal clearance from the liver- and systemic compartment, respectively. The data for the disposition sub-model was taken from a clinical study. Four of five semi-mechanistic models also contain an absorption sub-model. It divides the gastro intestinal (GI) tract into nine segments (stomach, duodenum, jejunum I & II, ileum I–IV and colon) where desvenlafaxine intestinal dissolution (in all segments) and subsequent permeation (only in intestinal segments) from an immediate release formulation (IRF) or extended release formulation (ERF) are simulated based on data from in vitro studies.

Table 1 Input data values for desvenlafaxine semi-mechanistic models. Input

Value

Reference/comments

Disposition Vss (L/kg)-observed CLiv CLR

3.42 21.9 (CV 32%) 12.12

Nichols et al. (2012) Nichols et al. (2012) Nichols et al. (2012)

10 5 10 From 134 (pH 1.3) to 0.1 (pH 10) 1.2

SimCyp default Predicted from Avdeef et al. (2004) SimCyp default Experimentally determined and from FDA Approval Package: NDA 21-992 (2008) SimCyp default

See Fig. 2B

Experimentally determined

Absorption Dissolution from IRF r (lm) D (104 cm2/min) h (lm) Cs (mg/mL) Particle density (g/mL) Dissolution from the ERF Biorelevant in vitro dissolution profile Permeability Peff (104 cm/s) duodenum Peff (104 cm/s) jejunum (I–II) Peff (104 cm/s) ileum (I–IV) Peff (104 cm/s) colon Colonic transit time (h) Other MW (g/mol) log Po:w – experimental Compound type pKa1 (acid) – experimental pKa2 (base) – experimental CB/CP fu – experimental

a

b

1 or 6.2 1a or 6.2b 6.2ab 1.8a or 6.2b 34 (CV 78%)

Extrapolated from Papp Extrapolated from Papp Extrapolated from Papp Extrapolated from Papp Meier et al. (1995)

263.38 1.8 Ampholyte 10.45 9.18 0.55 0.7

Experimentally determined Experimentally determined Experimentally determined Experimentally determined Default Sproule et al. (2008)

Vss: volume of distribution at steady state, CLiv: total intravenous clearance, CLR: renal clearance, IRF: immediate release formulation, r: particle size radius, D: diffusion coefficient, h: height of the diffusion layer surrounding the particle, Cs: saturation solubility of desvenlafaxine (at the particle surface), ERF: extended release formulation, Peff: effective intestinal permeability, CV: coefficient of variation, MW: molecular weight, Po:w: octanol–water partition coefficient at isoelectric point, fu: fraction unbound in plasma, CB/CP blood-to-plasma partition ratio. a For semi-mechanistic models that account for pH dependent permeability (i.e. M3 and M5). b For semi-mechanistic models that do not account for pH dependent permeability (i.e. M2 and M4).

199 200 201 202 203 204 205

each media) were as follows: Vessel 1 contained fasted state simulated gastric fluid (FASSGF), pH 1.8 (60 min), vessel 2 and 3 contained fasted state simulated intestinal fluid (FASSIF) at pH 6.5 (15 min) and 6.8 (15 min), respectively. Vessel 4 contained FASSIF with half of the amount of bile-components compared to regular FASSIF at pH 7.2 (30 min). Vessel 5 contained blank FASSIF at pH 7.5 (120 min) and vessel 6 contained simulated

colonic fluid (SCoF) at pH 5.8 (720 min). Biorelevant dissolution studies were conducted in triplicates. 5 mL samples were withdrawn at appropriate time-points, filtered (AcrodiscÒ PSF GxF/GHP 0.45 lm) and analyzed by HPLC (Section 2.3.5). To ensure that mass balance was maintained, the amount of desvenlafaxine remaining in each tablet residue at the end of the dissolution studies was determined. This was done by collecting and disintegrating

Please cite this article in press as: Franek, F., et al. In vitro solubility, dissolution and permeability studies combined with semi-mechanistic modeling to investigate the intestinal absorption of desvenlafaxine from an immediate- and extended release formulation. Eur. J. Pharm. Sci. (2015), http://dx.doi.org/ 10.1016/j.ejps.2015.06.012

206 207 208 209 210 211 212

PHASCI 3296

No. of Pages 11, Model 5G

16 June 2015 4 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274

F. Franek et al. / European Journal of Pharmaceutical Sciences xxx (2015) xxx–xxx

tablet residues under stirring in 250 mL phosphate buffer (pH 6.8), from which desvenlafaxine concentration was determined by HPLC analysis. 2.3.3. In vitro permeability studies using the Caco-2 cell-line Experiments were performed at two separate occasions in triplicates (N = 2, n = 3), on the 18th day post-seeding. Apical (0.5 mL) to basolateral (1 mL) flux of 1 mM (apical) desvenlafaxine was measured in buffers supplemented with HBSS and sodium bicarbonate. The basolateral compartment was kept at pH 7.4 (using 10 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) buffer). Desvenlafaxine flux was measured at three apical pH levels, i.e. pH 5.5 (10 mM 2-(N-morpholino)ethanesulfonic acid (MES) buffer), 6.8 (10 mM MES buffer) and 7.4 (10 mM HEPES buffer). Furthermore, DL-[4-3H]-propranolol hydrochloride flux was measured at apical pH 7.4 (10 mM HEPES buffer). Samples were taken from the basolateral compartment after 15, 30, 60, 90, 120 and 150 min and the slope of the last four measurements was used to calculate the flux value, which was divided by the donor concentration to calculate Papp. Samples including desvenlafaxine were analyzed by HPLC (Section 2.3.5). Samples including DL-[4-3H]-propranolol, were analyzed by liquid scintillation counting (Packard Tri-Carb 21000TR) after the addition of 2 mL of Ultima Gold Scintillation fluid (PerkinElmer, Waltham, Massachusetts). A one way analysis of variances (ANOVA) using GraphPad Prism 4 (GraphPad Software, Inc.) was employed for the statistical evaluation of desvenlafaxine Papp at the different (apical) pH-levels. 2.3.4. Determination of desvenlafaxine pKa, Do:w and Po:w Stock solutions were created by dissolving 1–2 mg desvenlafaxine succinate in 0.163 ± 0.053 (SD) M KCl. The pKa values were determined by UV-metric titration using a Sirius T3 autotitrator and diode-array detector (Sirius Analytical Instruments Ltd., East Sussex, UK). The stock solutions were pre-acidified to pH 2 with 0.500 M HCl and titrated with 0.500 M KOH from pH 2 to 12 at 25 °C thrice. The autotitrator used a diode-array-detector, which recorded the UV–Vis spectrum of the solution in-line. Thus the titration was mapped via spectra of protonated and unprotonated species, and the pKa values calculated using Sirius T3 v1.1 software. Desvenlafaxine logarithm of the octanol–water partition coefficient Log(Do:w) profile was determined by titration. N-octanol was added in increasing amounts to desvenlafaxine stock solutions and consequently titrated as described above. A difference curve was created from each of these titrations by blank subtraction and apparent pKa values were calculated. From the change in apparent pKa values with the n-octanol:water ratio combined with real pKa values, the (Log)Do:w profile, the octanol–water partition coefficient at the isoelectric point (Po:w) and desvenlafaxine fraction neutral (fn) were calculated. 2.3.5. Quantitative analysis of desvenlafaxine by HPLC A previously published HPLC method was modified and used to analyze desvenlafaxine (Shah et al., 2011). Desvenlafaxine was analyzed by HPLC (Merck Hitachi L-7000, VWR International, Tokyo, Japan), and was thus injected onto a Waters XBridge C18 column (3.5 lm, 150 mm  4.6 mm) held at 45 °C and eluted with a mixture of phosphate buffer (pH 6.8):methanol (60:40, v/v) at a flow rate of 1.0 mL/min. Quantitation was based on peak area measurement at k = 226 nm. The retention time for desvenlafaxine was approximately 4.4 min. 2.4. In vitro–in vivo extrapolation of desvenlafaxine dissolution and permeability The absorption sub-model, which was used to extrapolate desvenlafaxine dissolution and permeability from in vitro studies

into oral absorption, is available in SimCypÒ and named the ‘‘advanced dissolution, absorption and metabolism (ADAM)-model’’ (Jamei et al., 2009b). It divides the gastrointestinal-tract into nine different segments; one gastric, one duodenal-, two jejunal-, four ileal and one colonic segment (illustrated in Fig. 1). The ADAM-model can also be used to predict in vivo intestinal metabolism from in vitro studies (Jamei et al., 2009b). However, no intestinal metabolism was assumed because desvenlafaxine metabolism via CYP3A4, the predominant CYP-enzyme in the intestine (Paine et al., 1997), is low (DeMaio et al., 2011).

275

2.4.1. Desvenlafaxine dissolution from the IRF and ERF Desvenlafaxine in vivo dissolution rate (DR) from the IRF is in a gastro-intestinal (GI) segment at time (t) extrapolated by the Wang and Flanagan diffusion-layer equation (Jamei et al., 2009b; Wang and Flanagan, 1999), available in the absorption sub-model:

285

 1 1 C s  C lumen ðtÞ þ DRðtÞ ¼ 4pr 2 ðtÞD rðtÞ h

276 277 278 279 280 281 282 283 284

286 287 288 289

290



ð1Þ

292

where D is the diffusion coefficient, h is the thickness of the diffusion layer surrounding the particle, r is the particle size radius, Cs is the saturation solubility of desvenlafaxine at the particle surface and Clumen is the concentration of desvenlafaxine in the intestinal fluid. This equation may thus be used to predict in vivo dissolution time of the full IRF dose. In order to predict desvenlafaxine in vivo dissolution from the ERF, biorelevant in vitro dissolution studies were conducted (Section 2.3.2). Subsequently, the predicted in vivo dissolution profile was used as input to the absorption sub-model to determine desvenlafaxine in vivo dissolution rate from the ERF (Jamei et al., 2009b).

293

2.4.2. Desvenlafaxine pH-dependent- and independent intestinal permeability Desvenlafaxine effective intestinal permeability (Peff) across an intestinal segment was extrapolated from the in vitro apparent permeability of desvenlafaxine across filter grown Caco-2 cells at apical pH 7.4 (Papp, pH=7.4) (Darwich et al., 2010; Sun et al., 2002; Tchaparian et al., 2008):

305

294 295 296 297 298 299 300 301 302 303 304

306 307 308 309 310 311

312

Peff ¼ 100:939logðPapp;

pH¼7:4 Þ0:8787

 Sc

ð2Þ

314

In the absorption sub-model, the extrapolation also incorporates a scalar (Sc) to adjust for inter-laboratory differences between assays. The scalar was calculated from the scalar-compound (in this study: propranolol) Papp, pH=7.4, propranolol, determined from same conditions as desvenlafaxine Papp, pH=7.4 (i.e. apical pH 7.4):

315 316 317 318 319 320

321

Sc ¼

43  106 Papp;

ð3Þ

pH¼7:4; propranolol

The absorption sub-model considers the Peff extrapolated using Eq. (2) as the same for each intestinal segment. Thus, extrapolating Peff using Eq. (2) only, does not consider the effect of pH on permeability. In order to account for pH-dependent permeability, an empiric equation was used (Eq. (4)). To determine effective permeability at pH n (Peff, pH=n) of an intestinal segment:

323 324 325 326 327 328 329 330

331

Peff;

pH¼n

¼ Peff;

pH¼7:4 

Papp; pH¼n Papp; pH¼7:4

ð4Þ

where Peff, pH=7.4 is equal to Peff from Eq. (2), and is multiplied by the difference between apparent permeability at apical pH n (Papp, pH=n) and Papp, pH=7.4 in order to determine Peff, pH=n.

Please cite this article in press as: Franek, F., et al. In vitro solubility, dissolution and permeability studies combined with semi-mechanistic modeling to investigate the intestinal absorption of desvenlafaxine from an immediate- and extended release formulation. Eur. J. Pharm. Sci. (2015), http://dx.doi.org/ 10.1016/j.ejps.2015.06.012

333 334 335 336

PHASCI 3296

No. of Pages 11, Model 5G

16 June 2015 5

F. Franek et al. / European Journal of Pharmaceutical Sciences xxx (2015) xxx–xxx Table 2 Pharmacokinetic (PK) studies of healthy volunteers used as semi-mechanistic model input or comparison. PK study no a

1 2b 3b 4b 5b 6b 7b 8b 9b 10b 11b 12b 13b 14b 15b

Dosing regime

Age range (Years)

n

Gender

Reference

50 mg SD i.v. 75 mg SD o.IRF 150 mg SD o.IRF 75 mg SD o.ERF 100 mg SD o.ERF 100 mg SD o.ERF 100 mg SD o.ERF 2  50 mg SD o.ERF 150 mg SD o.ERF 2  75 mg SD o.ERF 200 mg SD o.ERF 200 mg SD o.ERF 200 mg SD o.ERF 100  3 mg SD o.ERF 100  6 mg SD o.ERF

18–45 18–45 18–45 18–45 20–50c 18–45 36–62 18–45 18–45 18–45 18–45 18–65 18–45 20–50c 20–50c

14 35 35 35 21 14 12 20 35 20 33 26 20 20 19

M M + Fd M + Fd M + Fd M + Fd M 66% M M M + Fd M 52% M M + Fd M M + Fd M + Fd

174-US (2008), Nichols et al. (2012) CSR 4509 (2008) CSR 4509 (2008) CSR 4509 (2008) Behrle et al. (2005), CSR 49864 (2008) 174-US (2008), Nichols et al. (2012) CSR 58916 (2008) CSR 48798 (2008) CSR 4509 (2008) CSR 48798 (2008) 181-US (2008) CSR 54268 (2008) CSR 48798 (2008) Behrle et al. (2005), CSR 49864 (2008) Behrle et al. (2005), CSR 49864 (2008)

SD: single dose, i.v.: intravenous, o.: oral, IRF: immediate release formulation, ERF: extended release formulation, n: amount of study subjects, M: male, F: female, HV: healthy volunteers. a PK data used as semi-mechanistic model input. b PK data used as comparison to semi-mechanistic model PK predictions. c Age range unknown, 20–50 assumed. d Gender distribution for PK study unknown, 50% male and female assumed.

337

2.5. Desvenlafaxine distribution and clearance

338

348

Desvenlafaxine volume of distribution at steady state (Vss), total- and renal-clearance were adopted from an intravenous clinical study reported in literature (Nichols et al., 2012) and used as input to a simple disposition model described previously (Rowland Yeo et al., 2010). In SimCypÒ this is the so-called ‘‘the minimal PBPK model’’. This disposition model is schematically illustrated in Fig. 1, and consists of a portal vein-, liver- and systemic-compartments interconnected by blood-flows (arrows between the compartments). Elimination occurs by hepatic clearance from the liver-compartment (i.e. the total- minus the renal clearance) and by renal clearance from the systemic compartment.

349

2.6. Clinical pharmacokinetic studies of desvenlafaxine

350

Plasma concentration–time profiles and PK parameter values, i.e. systemic exposure (AUC), maximum concentration (Cmax) and time to Cmax (tmax), were adopted from publically available clinical PK study results (FDA Approval Package: NDA 21-992, 2008). The PK studies are listed in Table 2. Data from PK-study 1 was used as input for distribution and clearance and data from PK studies 2–15 were used for comparison to semi-mechanistic model PK predictions.

339 340 341 342 343 344 345 346 347

351 352 353 354 355 356 357

358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373

2.7. Semi-mechanistic models used for prediction of desvenlafaxine pharmacokinetics The PK predictions were conducted so that dosage regimen, age-range and gender of the virtual subjects matched real subjects. The PK studies including real subjects are specified in Table 2. Further, to ensure reasonable confidence in the PK predictions, each virtual PK study had twenty times more virtual subjects than the actual PK study. The match was considered to be ‘‘good’’ when predicted PK parameters (AUC, Cmax, tmax) were within 0.75–1.25-fold from observed. A predicted/observed PK parameter ratio between 0.5–0.75 and 1.25–2 was considered a poor match. Having too many predicted/observed PK parameter ratios that are poor or worse would indicate insufficient model performance. A total of five semi-mechanistic models were built; one intravenous – (M1), two IRF (M2 and M3) – and two ERF – (M4 and M5) models.

Table 3 Overview of semi-mechanistic model 2–5 similarities and differences. Dissolution

Permeability

pH independent pH dependent

IRF

ERF

M2 M3

M4 M5

IRF: Immediate release formulation. ERF: Extended release formulation.

M1 considers desvenlafaxine disposition only, as described in Section 2.5. M1 thus only simulates desvenlafaxine pharmacokinetics after intravenous administration. M2–5 use the same sub-model for disposition as M1, but also include the absorption sub-model that considers dissolution and permeation. An overview of M2–5 is given in Table 3. For M2, desvenlafaxine in vivo dissolution is extrapolated from the in vitro determined pH-solubility-profile by using Eq. (1). M2 does not account for pH-dependent permeability and thus uses Eq. (2) only to extrapolate Peff (described in Section 2.4.2). Permeability is thus the same in each intestinal segment. M3 extrapolates desvenlafaxine in vivo dissolution the same as M2. However, M3 accounts for pH-dependent intestinal permeability by using Eqs. (2) and (4) to extrapolate Peff. Consequently, the Peff of the intestinal segments is adjusted to match the mid-range pH of the intestinal fluid in duodenum/jejunum, ileum and colon pH at 6.1, 7.4 and 6.5, respectively (Bown et al., 1974). In M4 and M5, desvenlafaxine in vivo dissolution is predicted from the biorelevant in vitro dissolution studies described in Section 2.3.2. However, M4 does not account for pH-dependent permeability whereas M5 does.

374

3. Results

395

3.1. Desvenlafaxine predicted in vivo dissolution and permeability

396

Desvenlafaxine in vivo dissolution from the IRF (75–150 mg) was extrapolated using the pH solubility profile in Fig. 2A. The absorption sub-model (using Eq. (1)) predicted that the total desvenlafaxine dose from the IRF dissolves no later than 3 min after administration, for any virtual individual. The sensitivity

397

Please cite this article in press as: Franek, F., et al. In vitro solubility, dissolution and permeability studies combined with semi-mechanistic modeling to investigate the intestinal absorption of desvenlafaxine from an immediate- and extended release formulation. Eur. J. Pharm. Sci. (2015), http://dx.doi.org/ 10.1016/j.ejps.2015.06.012

375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394

398 399 400 401

PHASCI 3296

No. of Pages 11, Model 5G

16 June 2015 F. Franek et al. / European Journal of Pharmaceutical Sciences xxx (2015) xxx–xxx

1000

100

A % desvenlafaxine dissolved

Desvenlafaxine (succinate) solubility (mg/mL)

6

100

10

1

0.1

0

5

10

75

50

25

0

15

B

0

4

pH

8

12

16

Time (h)

Fig. 2. (A) Desvenlafaxine (succinate) solubility-pH profile, where open circles represent values from publically available data (FDA Approval Package: NDA 21-992, 2008), full circles represent mean values from in house determinations with error-bars representing standard deviation (n = 3). The error-bars may be so small that they are not visible. (B) Desvenlafaxine predicted in vivo dissolution profile from the ERF by biorelevant in vitro dissolution. The error-bars represent the standard deviation (n = 3).

405 406 407 408 409 410 411 412 413 414 415 416 417

418 419 420 421

3.2. Predicted desvenlafaxine pharmacokinetics after intravenous infusion and oral administration of IRF The observed and simulated plasma concentration–time profiles after 1 h intravenous infusion of desvenlafaxine were similar

16

7

A

14

6

Ileum

12

5

10

4

8 3

6 4

2

Colon

1

Duodenum & Jejunum

2 0

5

(Fig. 4A). This was also indicated by AUC, Cmax and tmax predicted/observed ratio, which were within 0.75–1.25 (Fig. 5). Accounting for pH-dependent permeability after IRF administration of desvenlafaxine ‘‘delays’’ predicted tmax from approximately 2 to 3.5 h, which improved tmax and plasma concentration–time profile prediction (Fig. 4B). The improved prediction of tmax was indicated by the better predicted/observed tmax ratio, which increased from <0.65 to 1 when accounting for pH-dependent permeability (Fig. 5). When administered as an IRF, desvenlafaxine was predicted to be almost completely absorbed in the small intestine since the combined duodenum, jejunum and ileum %fraction absorbed (%fa) was >97%, no matter if pH-dependent permeability was accounted for or not (Fig. 6A). Consequently, colonic absorption is predicted to be minimal, i.e. <3% fa, because desvenlafaxine was predicted to be absorbed before reaching the colon. When not accounting for pH-dependent permeability, desvenlafaxine administered as an IRF was predicted to be mainly absorbed in the upper small intestine (duodenum/jejunum 90% fa) compared to the lower small intestine (ileum 10% fa) (Fig. 6A). The opposite was observed when accounting for pH-dependent permeability; only 40% of the dose was predicted

5.5

6

6.5

apical pH

7

7.5

8

0

100

B

10

Do:w

404

analyses indicated that varying parameters used to predict desvenlafaxine in vivo dissolution from the IRF had negligible influence on desvenlafaxine absorption and pharmacokinetics. In contrast to the IRF, the predicted in vivo dissolution rate of desvenlafaxine from the ERF was lower; about 75% of the desvenlafaxine dose was predicted to be dissolved 16 h after administration (Fig. 2B). 100% (±5%) of the stated desvenlafaxine dose from the ERF was recovered. Desvenlafaxine Papp decreased approximately 13-, 5- and 2-fold from apical pH 7.4 to 5.5 (p < 0.001), 6.8 to 5.5 (p < 0.001) and 7.4 to 6.8 (p < 0.001), respectively (Fig. 3A). Consequently, when pH-dependent permeability was accounted for, the extrapolated Peff decreased with apical pH. Desvenlafaxine Do:w decreased 13.8-, 4.4- and 3.1-fold (Fig. 3B) and fn decreased 78-, 20- and 4-fold from pH 7.4 to 5.5, 6.8 to 5.5 and 7.4 to 6.8, respectively.

Extrapolated Peff (10-4 cm/s)

403

Desvenlafaxine Papp (10-6 cm/s)

402

1

0.1

0.01

5

6

7

8

9

pH

Fig. 3. (A) Desvenlafaxine apparent permeability (Papp, left y-axis), experimentally determined at apical pH 5.5, 6.8 and 7.4. Full circles denote mean Papp and error-bars denote the in vitro experimental standard-deviation (N = 2, n = 3). The error-bars may be so small that they are not visible. The right y-axis indicates effective intestinal permeability (Peff), extrapolated from Papp. When not accounting for pH-dependent permeability (i.e. M2 and M4), the Peff extrapolated from Papp at apical pH 7.4 was used for all intestinal segments. However, when accounting for pH-dependent permeability (i.e. M3 and M5), the extrapolation of Peff is influenced by intestinal segment apical pH. The beginning of the arrow indicates the intestinal segments apical pH that was used, and the end of the arrow indicates that segments extrapolated Peff. (B) Desvenlafaxine octanol–water partition coefficients (Do:w) over pH.

Please cite this article in press as: Franek, F., et al. In vitro solubility, dissolution and permeability studies combined with semi-mechanistic modeling to investigate the intestinal absorption of desvenlafaxine from an immediate- and extended release formulation. Eur. J. Pharm. Sci. (2015), http://dx.doi.org/ 10.1016/j.ejps.2015.06.012

422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443

PHASCI 3296

No. of Pages 11, Model 5G

16 June 2015 7

A

1000

100

10

1

0

6

12

18

24

30

36

42

48

Desvenlafaxine concentration (ng/mL)

Desvenlafaxine concentration (ng/mL)

F. Franek et al. / European Journal of Pharmaceutical Sciences xxx (2015) xxx–xxx

B 300

200

100

0

0

4

8

C

250

200

150

100 50

0

0

8

16

24

32

40

12

16

20

24

Time (h)

48

56

64

72

Desvenlafaxine concentration (ng/mL)

Desvenlafaxine concentration (ng/mL)

Time (h)

D

1800

1500 1200 900

600 300 0

0

8

16

24

32

40

48

56

64

72

Time (h)

Time (h)

Fig. 4. Predicted- (lines) and observed-mean (empty circles) plasma concentration–time profiles after (A) 50 mg 1 h-intravenous infusion (PK-study 1), (B) 75 mg IRF oral (PK study 2), (C) 100 mg ERF oral (PK study 7), and (D) 6  100 mg ERF oral (PK-study 15)-administration of desvenlafaxine. In B–D, full- and dashed-black lines denote predictions from semi-mechanistic models that account (M3 or M5) and do not account (M2 or M4) for pH-dependent permeability, respectively. Gray lines indicate the 5-th and 95-th percentile predictions (in C and D: for M5). C and D denote best and worst predictions of PK-studies 1–15, respectively. PK-study information is given in Table 2.

444 445 446 447 448

449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467

to be absorbed in the duodenum/jejunum vs. 60% in the ileum. The absorption in the lower small intestine thus appears greater and occurs later than in the upper small intestine when accounting for pH-dependent permeability (Fig. 6A), which explains the ‘‘delay’’ in predicted tmax (Figs. 4B and 5).

3.3. Predicted desvenlafaxine pharmacokinetics after oral administration of the ERF Accounting for pH-dependent permeability had minimal influence on predictions of desvenlafaxine plasma concentration–time profiles when administered as an ERF (Fig. 4C and D). The small difference between predictions was also indicated by the small differences between predicted/observed PK parameter ratios; accounting for pH-dependent permeability decreases, at most, mean predicted AUC, Cmax and tmax by 4%, 7% and 3%, respectively (Fig. 5, PK studies 4–15). The predicted/observed PK parameter ratios were within 0.8–1.2 for tmax and within 0.65–1.2 for AUC and Cmax (excluding PK-study 15, with AUCpredicted/observed of 0.62), when accounting for – or not accounting for – pH-dependent permeability. Approximately 80% of the administered dose from the ERF was predicted to be absorbed, no matter if accounting for pH-dependent permeability or not. This is indicated in Fig. 6B by adding duodenum/jejunum, ileum and colon %fa. Approximately half of the absorbed dose was predicted to be absorbed in the small

intestine (i.e. duodenum/jejunum and ileum), the other half in the colon.

468

4. Discussion

470

4.1. Desvenlafaxine absorption from the IRF is rate-limited by pH-dependent intestinal permeability

471

Desvenlafaxine disposition was predicted adequately (Figs. 4A and 5). This was to be expected since desvenlafaxine distribution and clearance were not mechanistically modeled, i.e. extrapolated from in vitro studies or physicochemical properties, but adopted from an intravenous clinical pharmacokinetic study (Nichols et al., 2012). The mechanistic part of semi-mechanistic modeling refers to the investigation of desvenlafaxine absorption, which is the focus of this study. Desvenlafaxine dissolution from the IRF is predicted to occur rapidly due to its high solubility at gastrointestinal pH. This is underlined by desvenlafaxine’s low (<250 mL) dose/solubility ratio (FDA, 2000) for the highest strength of the IRF (150 mg); 1 and 11 mL at pH 1.2 and 7.3, respectively (Fig. 2 A). The entire desvenlafaxine dose from the IRF is thus available for permeation in the duodenum. Desvenlafaxine permeability is pH-dependent and increases with apical pH (Fig. 3A). In order to account for the in vitro observed pH-dependent permeability when predicting in vivo permeability, Eq. (4) was used in addition to Eq. (2) to extrapolate Papp

473

Please cite this article in press as: Franek, F., et al. In vitro solubility, dissolution and permeability studies combined with semi-mechanistic modeling to investigate the intestinal absorption of desvenlafaxine from an immediate- and extended release formulation. Eur. J. Pharm. Sci. (2015), http://dx.doi.org/ 10.1016/j.ejps.2015.06.012

469

472

474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491

PHASCI 3296

No. of Pages 11, Model 5G

16 June 2015 8

F. Franek et al. / European Journal of Pharmaceutical Sciences xxx (2015) xxx–xxx

AUCpredicted/ AUCobserved

1.5 1.25

1 0.75

0.5 PK study no0

1 i.v.

2

Formulation

--IRF--

3

4

------------------------------------------ERF----------------------------------------

5

6

7

8

9

10

11

12

13

14

15

Dose (mg)

50

75

150

75

----------100----------

---150---

-------200-------

300 600

2

3

4

5

9

11

14

Cmax, predicted/ Cmax, observed

1.5 1.25 1 0.75

0.5 PK study no0 Formulation

i.v.

--IRF--

-------------------------------------------ERF---------------------------------------

Dose (mg)

50

75

150

75

----------100----------

---150---

-------200-------

300 600

3

4

5

9

11

14

1

6

7

8

10

12

13

15

tmax, predicted/ tmax, observed

1.5 1.25 1 0.75

0.5 0 PK study no Formulation

1

2

i.v.

--IRF--

Dose (mg)

50

75

150

6

7

8

10

12

13

15

-------------------------------------------ERF-------------------------------------75

----------100----------

---150---

-------200-------

300 600

100

A

Predicted %fa desvenlafaxine (ERF)

Predicted %fa desvenlafaxine (IRF)

Fig. 5. The predicted vs. observed ratio of the pharmacokinetic (PK) parameters systemic exposure (AUC), maximum concentration (Cmax) and time until Cmax (tmax). The xaxis denotes corresponding PK studies, their administration-route/formulation and dose (see Table 2 for more information). Empty and full circles denote predictions from semi-mechanistic models that account (M3 or M5) and do not account (M2 or M4) for pH-dependent permeability, respectively. Triangles denote predicted (using M1) vs. observed PK parameter ratio after intravenous (i.v.) 1-h infusion.

75

50

25

0 Duodenum & Jejunum

Ileum

Colon

100

B

75

50

25

0 Duodenum & Jejunum

Ileum

Colon

Fig. 6. Predicted % fraction absorbed (%fa) desvenlafaxine in three different regions of the intestinal tract when administered as (A) IRF and (B) ERF. White and black columns denote %fa predictions from semi-mechanistic models that account (M3 or M5) and do not account (M2 or M4) for pH-dependent permeability, respectively.

492 493 494 495 496 497 498 499

to Peff. The equation assumes that Papp changes in the same extent as Peff when apical pH is changed. The drawback of using this approach to account for pH-dependent permeability is that Peff is not directly linked to pH within the absorption sub-model. The Peff is firstly extrapolated using Eq. (4), which is not part of the absorption sub-model, and secondly used as model input. Every virtual individual thus has the same Peff throughout the intestinal tract, independent of their

individual (simulated) intestinal pH. Consequently, the (oral) semi-mechanistic models cannot predict the inter-individual variability of desvenlafaxine pharmacokinetics due to pH-dependent permeability. We suggest that desvenlafaxine in vivo permeability is pH-dependent and consequently the majority of an IRF dose is absorbed in the ileum, which ‘‘delays’’ desvenlafaxine tmax. The importance of accounting for pH-dependent permeability to an

Please cite this article in press as: Franek, F., et al. In vitro solubility, dissolution and permeability studies combined with semi-mechanistic modeling to investigate the intestinal absorption of desvenlafaxine from an immediate- and extended release formulation. Eur. J. Pharm. Sci. (2015), http://dx.doi.org/ 10.1016/j.ejps.2015.06.012

500 501 502 503 504 505 506 507

PHASCI 3296

No. of Pages 11, Model 5G

16 June 2015 F. Franek et al. / European Journal of Pharmaceutical Sciences xxx (2015) xxx–xxx 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573

adequate prediction of desvenlafaxine tmax and consequently the pharmacokinetic profile and parameters is indicated in Figs. 4B and 5. Due to increasing pH along the small intestine (Bown et al., 1974), desvenlafaxine permeability is predicted to be lower in the duodenum/jejunum than in the ileum (Fig. 3A). Consequently, the majority of the desvenlafaxine dose from the IRF is predicted to be absorbed after its passage through the duodenum/jejunum, i.e. in the ileum (Fig. 6A), which is shown as a ‘‘delayed’’ tmax (Figs. 4B and 5). A ‘‘delayed’’ tmax, due to low duodenal/jejunal and high ileal permeability has previously been observed after oral administration of sotalol (Dahan et al., 2012, 2010; Liu et al., 2012). It may be argued that only two reported PK studies were used as comparison to IRF semi-mechanistic model PK predictions (vs. 11 for the ERF) (Fig. 5). However, both IRF PK studies contain relatively high amount of subjects (n = 35) compared to ERF PK studies (n often 20) (Table 2). When building mechanistic models, it is useful and important to conduct sensitivity analyses to investigate which parameters under what conditions impact pharmacokinetics and thus need to be characterized thoroughly (Tsamandouras et al., 2015). Sensitivity analyses outline parameters that may (or likely will not) influence mean pharmacokinetics in studies with few individuals. Both reported IRF PK studies indicate that the absorption-sub model and thus the semi-mechanistic IRF model underperforms when pH-dependent permeability is not accounted for since desvenlafaxine tmax is predicted ‘‘poorly’’ (predicted/observed ratio between 0.5 and 0.75). The sensitivity analyses of parameters used to predict in vivo dissolution from the IRF indicate that the poor prediction of absorption is probably not due to poor prediction of in vivo dissolution. Rather, it is likely due to not accounting for the underlying mechanism of pH-dependent permeability. In order to investigate if desvenlafaxine permeability is pH-dependent in vivo, the IntellicapÒ, an electronic drug delivery device, may be used (Becker et al., 2014). Desvenlafaxine release from the IntellicapÒ may be triggered by remote in a desired intestinal region (at a desired pH). By coordinated blood-sampling, desvenlafaxine in vivo absorption may be measured in the upper small intestine and lower small intestine and thus compared. Desvenlafaxine intestinal permeability is probably driven by diffusion-driven transcellular permeability (Ptrans) of its unionized form. Ptrans is largely influenced by compound lipophilicity and ionization, and thus increases with Do:w and fn (Reynolds et al., 2009;Winiwarter et al., 1998). Since desvenlafaxine Do:w increases with pH to the same extent as Papp (Fig. 3B), it is likely that Ptrans of the unionized form to large extent drives intestinal permeability. The mechanism of desvenlafaxine intestinal permeation is similar to sotalol, for which the pH-dependent intestinal permeability is also driven by Ptrans of its unionized form (Dahan et al., 2010). Sotalol is thus good candidate for further qualifying the use of Eq. (4) to extrapolate pH-dependent in vitro Papp to in vivo Peff. Desvenlafaxine is a BDDCS class 3 compound and its permeability may thus be influenced by carriers (Benet et al., 2011; Wu and Benet, 2005). However, few studies have investigated desvenlafaxine carrier-mediated permeability. Efflux studies using Caco-2 monolayers have shown that desvenlafaxine is not a substrate for P-glycoprotein (Oganesian et al., 2008). Furthermore, desvenlafaxine oral absorption-rate is linear with increasing dose (Behrle et al., 2005), which indicates absence of saturable carriers (Sugano et al., 2010). Consequently, we currently consider carriers to play a minor role in desvenlafaxine intestinal permeation. Paracellular pore radii and thus paracellular permeability (Ppara) decreases down the intestinal tract (Fordtran et al., 1965;Billich and Levitan, 1969). Hence, compounds with permeability mainly driven by Ppara, have decreased permeability down the (small)

9

intestinal tract (Artursson et al., 1993). Desvenlafaxine permeability however, increases down the small intestinal tract, as indicated by the ‘‘delayed’’ tmax at 3.5 h (Fig. 4B). Consequently, we consider Ppara to play a minor role in desvenlafaxine intestinal permeability.

574

4.2. Desvenlafaxine absorption from the ERF is rate-limited by dissolution

578

Desvenlafaxine predicted in vivo dissolution from the ERF appears rate-limited by the formulation (Fig. 2B). Increasing attrition forces by increasing USP3 dissolution apparatus reciprocation rate from 10 to 50 dpm or by adding beads into the reciprocating cylinder does not influence desvenlafaxine dissolution from the ERF (data not shown). This is likely because the ERF contains high amount of long-chain HPMC-polymer (Hadfield et al., 2007) that forms a strong gel when in contact with GI fluid, which decreases the influence of attrition-forces on dissolution rate (Franek et al., 2014; Sung et al., 1996). In contrast to the IRF, the predicted desvenlafaxine tmax after oral administration of the ERF is minimally influenced by pH-dependent permeability (Figs. 4B–D and 5). The ERF controls the rate of desvenlafaxine dissolution and thus the amount available for intestinal permeation (Franek et al., 2014; Turner et al., 2004). Consequently, the desvenlafaxine dose is predicted to be absorbed to lesser extent in the small intestine and to greater extent in the colon when administered as an ERF compared to the IRF (Fig. 6). This diminishes the effect of pH-dependent permeability on tmax.

580

5. Conclusions

600

Based on in vitro studies and semi-mechanistic modeling predictions compared to clinical data, we suggest that desvenlafaxine absorption from the IRF is rate-limited by permeability, which increases with pH. Consequently, desvenlafaxine permeability is probably highest in the ileum, which we suggest is the reason for a clinically observed ‘‘delayed’’ tmax at 3.5 h when administered as an IRF. We suggest that desvenlafaxine intestinal permeability is driven by transcellular diffusion of its unionized form. Desvenlafaxine absorption from the ERF is likely rate-limited by dissolution due to the formulation. Consequently, the influence of pH-dependent permeability on desvenlafaxine absorption from the ERF is predicted to be minimal. We have proposed an absorption mechanism for desvenlafaxine that may explain why it is a BCS class 1 (high solubility, high permeability) and BDDCS class 3 compound (high solubility, poor metabolism). Thus, in the case of desvenlafaxine, poor metabolism does not imply low intestinal permeability, merely low duodenal/jejunal permeability.

601

Conflict of interest

619

The authors report no conflict of interest.

575 576 577

579

581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599

602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618

620

6. Uncited references

621

Bettini et al. (2001), Davis et al. (1986), Fallingborg et al. (1989), and USP (2011).

622

Acknowledgements

624

The authors would like to thank Erling B. Jørgensen, preformulation specialist at Lundbeck A/S for conducting pKa, Do:w and Po:w analysis and practical help with setting up the HPLC. We would like to acknowledge Lundbeck A/S and the Drug Research Academy, University of Copenhagen, for co-financing this project.

625

Please cite this article in press as: Franek, F., et al. In vitro solubility, dissolution and permeability studies combined with semi-mechanistic modeling to investigate the intestinal absorption of desvenlafaxine from an immediate- and extended release formulation. Eur. J. Pharm. Sci. (2015), http://dx.doi.org/ 10.1016/j.ejps.2015.06.012

623

626 627 628 629

PHASCI 3296

No. of Pages 11, Model 5G

16 June 2015 10

F. Franek et al. / European Journal of Pharmaceutical Sciences xxx (2015) xxx–xxx

630

References

631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712

174-US, 2008. Protocol 3151A1-174-US: An open-label, randomized, single-dose, dose-proportionality study of desvenlafaxine SR in healthy subjects. From: FDA Approval Package (NDA: 21-992) – Clinical Pharmacology and Biopharmaceutics Review Part 4, pp. 227–239. (accessed 11.11.14). 181-US, 2008. Protocol 3151A1-181-US: an open-label, randomized, single-dose, 4period crossover study to determine the effect of a low-fat, medium-fat, and high fat meal in relationship to fasting on the relative bioavailability, pharmacokinetic profile, safety and tolerability of DVS-233 SR in healthy adults. From: FDA Approval Package (NDA: 21-992) – Clinical Pharmacology and Biopharmaceutics Review Parts 5 and 6, pp. 315–372. (accessed 11.11.14). Amidon, G.L., Lennernäs, H., Shah, V.P., Crison, J.R., 1995. A theoretical basis for a biopharmaceutic drug classification: the correlation of in vitro drug product dissolution and in vivo bioavailability. Pharm. Res. 12, 413–420. Artursson, P., Ungell, A.L., Löfroth, J.E., 1993. Selective paracellular permeability in two models of intestinal absorption: cultured monolayers of human intestinal epithelial cells and rat intestinal segments. Pharm. Res. 10, 1123–1129. Avdeef, A., Nielsen, P.E., Tsinman, O., 2004. PAMPA-a drug absorption in vitro model: 11. Matching the in vivo unstirred water layer thickness by individualwell stirring in microtitre plates. Eur. J. Pharm. Sci. 22, 365–374. Becker, D., Zhang, J., Heimbach, T., Penland, R.C., Wanke, C., Shimizu, J., Kulmatycki, K., 2014. Novel orally swallowable IntelliCapÒ device to quantify regional drug absorption in human GI tract using diltiazem as model drug. AAPS PharmSciTech 15, 1490–1497. Behrle, J.A., Nichols, A.I., McGrory, S.B., Raible, D., 2005. An open-label, randomized, single-dose, dose-proportionality study of oral doses of a sustained-release formulation of desvenlafaxine succinate in healthy subjects. Clin. Pharmacol. Ther. 77, 82. Benet, L.Z., Amidon, G.L., Barends, D.M., Lennernäs, H., Polli, J.E., Shah, V.P., Stavchansky, S.A., Lawrence, X.Y., 2008. The use of BDDCS in classifying the permeability of marketed drugs. Pharm. Res. 25, 483–488. Benet, L.Z., Broccatelli, F., Oprea, T.I., 2011. BDDCS applied to over 900 drugs. AAPS J. 13, 519–547. Bettini, R., Catellani, P.L., Santi, P., Massimo, G., Peppas, N.A., Colombo, P., 2001. Translocation of drug particles in HPMC matrix gel layer: effect of drug solubility and influence on release rate. J. Control. Release 70, 383–391. Billich, C.O., Levitan, R., 1969. Effects of sodium concentration and osmolality on water and electrolyte absorption from the intact human colon. J. Clin. Invest. 48, 1336–1347. Bown, R.L., Gibson, J.A., Sladen, G.E., Hicks, B., Dawson, A.M., 1974. Effects of lactulose and other laxatives on ileal and colonic pH as measured by a radiotelemetry device. Gut 15, 999–1004. CSR 4509, 2008. Protocol 0600D3-167-FR: an open-label, randomized, three-period crossover study of the relative bioavailability and tolerability of two single doses of two formulations of desvenlafaxine in healthy adults. From: FDA Approval Package (NDA: 21-992) – Clinical Pharmacology and Biopharmaceutics Review Part 3, pp. 147–168. (accessed 11.11.14). CSR 48798, 2008. Protocol 0600D3-186-US: an open-label, 4-period, randomized crossover study of the pharmacokinetic profile and tolerability of 3 new dosage strengths Of DVS-233 SR in healthy adults. From: FDA Approval Package (NDA: 21-992) – Clinical Pharmacology and Biopharmaceutics Review Part 6, pp. 392– 405. (accessed 11.11.14). CSR 49864, 2008. Protocol 0600D3-172-US: an open-label, randomized, single-dose, dose-proportionality study of desvenlafaxine SR in healthy subjects. From: FDA Approval Package (NDA: 21-992) – Clinical Pharmacology and Biopharmaceutics Review Part 6, pp. 392–405. (accessed 11.11.14). CSR 54268, 2008. Protocol 1351A1-177-US: the pharmacokinetic profile, safety, and tolerability of three different 200-Mg tablet formulations of DVS-233 SR, in healthy adults from two different manufacturing sites. From: FDA Approval Package (NDA: 21-992) – Clinical Pharmacology and Biopharmaceutics Review Parts 4 and 5, pp. 259–275. (accessed 11.11.14). CSR 58916, 2008. Protocol 3151A1-169-EU: an open-label, singe 100-mg dose, parallel-group study of the pharmacokinetics and safety of DCS-233 SR in subjects with chronic hepatic impairment and in matched healthy adults. From: FDA Approval Package (NDA: 21-992) – Clinical Pharmacology and Biopharmaceutics Review Part 5, pp. 276–314. (accessed 11.11.14). 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. Dahan, A., Lennernäs, H., Amidon, G.L., 2012. The fraction dose absorbed, in humans, and high jejunal human permeability relationship. Mol. Pharm. 9, 1847–1851.

Darwich, S., Neuhoff, S., Jamei, M., Rostami-Hodjegan, A., 2010. Interplay of metabolism and transport in determining oral drug absorption and gut wall metabolism: a simulation assessment using the ‘‘Advanced Dissolution, Absorption, Metabolism (ADAM)’’ model. Curr. Drug Metab. 11, 716–729. Davis, S.S., Hardy, J.G., Fara, J.W., 1986. Transit of pharmaceutical dosage forms through the small intestine. Gut 27, 886–892. DeMaio, W., Kane, C.P., Nichols, A.I., Jordan, R., 2011. Metabolism studies of desvenlafaxine. J. Bioequiv. Bioavail. 3, 151–160. EMA, 2010. Committee for Medicinal Products for Human Use: Guideline on the Investigation of Bioequivalence. European Medicines Agency (EMA), London. Fallingborg, J., Christensen, L.A., Ingeman-Nielsen, M., Jacobsen, B.A., Abildgaard, K., Rasmussen, H.H., 1989. PH-profile and regional transit times of the normal gut measured by a radiotelemetry device. Aliment. Pharmacol. Ther. 3, 605–613. FDA Approval Package: NDA 21-992, 2008. Center for Drug Evaluation and Research: FDA Approval Package (NDA: 21-992) – Clinical Pharmacology and Biopharmaceutics Review. (accessed 11.11.14). FDA, 2000. Food and Drug Administration: Guidance for Industry: Waiver of In Vivo Bioavailability and Bioequivalence Studies for Immediate-release Solid Oral Dosage Forms Based on a Biopharmaceutics Classification System. Food and Drug Administration, Rockville, MD. Fordtran, J.S., Rector Jr., F.C., Ewton, M.F., Soter, N., Kinney, J., 1965. Permeability characteristics of the human small intestine. J. Clin. Invest. 44, 1935–1944. Franek, F., Holm, P., Larsen, F., Steffansen, B., 2014. Interaction between fed gastric media (Ensure PlusÒ) and different hypromellose based caffeine controlled release tablets: comparison and mechanistic study of caffeine release in fed and fasted media versus water using the USP dissolution apparatus 3. Int. J. Pharm. 461, 419–426. Grandvuinet, A.S., Gustavsson, L., Steffansen, B., 2013. New insights into the carriermediated transport of estrone-3-sulfate in the Caco-2 cell model. Mol. Pharm. 10, 3285–3295. Hadfield, A.F., Shah, S.M., Winkley, M.W., Sutherland, K.W., Provost, J.A., Park, A., Shipplett, R.A., Russell, B.W., Weber, B.T., 2004. Succinate salt of O-desmethylvenlafaxine. Patent, US 10/073,743 US 6673838 B2. Hadfield, A.F., Shah, S.M., Provost, J.A., 2007. Solubility, permeability, and bioavailability; oral dosage. Patent, US 11/334,223 US 7291347 B2. Jamei, M., Marciniak, S., Feng, K., Barnett, A., Tucker, G., Rostami-Hodjegan, A., 2009a. The Simcyp population-based ADME simulator. Expert Opin. Drug Metab. Toxicol. 5, 211. Jamei, M., Turner, D., Yang, J., Neuhoff, S., Polak, S., Rostami-Hodjegan, A., Tucker, G., 2009b. Population-based mechanistic prediction of oral drug absorption. AAPS J. 11, 225–237. Jones, H.M., Parrott, N., Ohlenbusch, G., Lavé, T., 2006. Predicting pharmacokinetic food effects using biorelevant solubility media and physiologically based modelling. Clin. Pharmacokinet. 45, 1213–1226. Jones, H.M., Dickins, M., Youdim, K., Gosset, J.R., Attkins, N.J., Hay, T.L., Gurrell, I.K., Logan, Y.R., Bungay, P.J., Jones, B.C., 2012. Application of PBPK modelling in drug discovery and development at Pfizer. Xenobiotica 42, 94–106. Klein, S., 2005. Biorelevant dissolution test methods for modified release dosage forms. PhD-Thesis. Johann Wolfgang Goethe-Universität, Germany. Kornstein, S.G., McIntyre, R.S., Thase, M.E., Boucher, M., 2014. Desvenlafaxine for the treatment of major depressive disorder. Expert Opin. Pharmacother. 15, 1449– 1463. Liu, W., Okochi, H., Benet, L.Z., Zhai, S.D., 2012. Sotalol permeability in cultured-cell, rat intestine, and PAMPA system. Pharm. Res. 29, 1768–1774. Meier, R., Beglinger, C., Dederding, J.P., Meyer-Wyss, B., Fumagalli, M., Rowedder, A., Turberg, Y., Brignoli, R., 1995. Influence of age, gender, hormonal status and smoking habits on colonic transit time. Neurogastroenterol. Motil.: Off. J. Eur. Gastroint. Motil. Soc. 7, 235–238. Neuhoff, S., Ungell, A.L., Zamora, I., Artursson, P., 2003. PH-dependent bidirectional transport of weakly basic drugs across Caco-2 monolayers: implications for drug–drug interactions. Pharm. Res. 20, 1141–1148. Neuhoff, S., Yeo, K.R., Barter, Z., Jamei, M., Turner, D.B., Rostami-Hodjegan, A., 2013. Application of permeability-limited physiologically-based pharmacokinetic models: Part I-digoxin pharmacokinetics incorporating P-glycoproteinmediated efflux. J. Pharm. Sci. 102, 3145–3160. Nichols, A.I., Behrle, J.A., Richards, L.S., Parker, V.D., Posener, J.A., Fruncillo, R., Paul, J., 2012. The absolute bioavailability of desvenlafaxine in healthy subjects. J. Bioequiv. Bioavail. 4, 18–23. Oganesian, A., Shilling, A.D., Young-Sciame, R., Tran, J., Watanyar, A., Azam, F., Kao, J., Leung, L., 2008. Desvenlafaxine and venlafaxine exert minimal in vitro inhibition of human cytochrome P450 and P-glycoprotein activities. Psychopharmacol. Bull. 42, 47–63. Pae, C.U., 2011. Desvenlafaxine in the treatment of major depressive disorder. Expert Opin. Pharmacother. 12, 2923–2928. Paine, M.F., Khalighi, M., Fisher, J.M., Shen, D.D., Kunze, K.L., Marsh, C.L., Perkins, J.D., Thummel, K.E., 1997. Characterization of interintestinal and intraintestinal variations in human CYP3A-dependent metabolism. J. Pharmacol. Exp. Ther. 283, 1552–1562. Patel, N., Polak, S., Jamei, M., Rostami-Hodjegan, A., Turner, D.B., 2014. Quantitative prediction of formulation-specific food effects and their population variability from in vitro data with the physiologically-based ADAM model: a case study using the BCS/BDDCS Class II drug nifedipine. Eur. J. Pharm. Sci.: Off. J. Eur. Fed. Pharm. Sci. 57, 240–249.

Please cite this article in press as: Franek, F., et al. In vitro solubility, dissolution and permeability studies combined with semi-mechanistic modeling to investigate the intestinal absorption of desvenlafaxine from an immediate- and extended release formulation. Eur. J. Pharm. Sci. (2015), http://dx.doi.org/ 10.1016/j.ejps.2015.06.012

713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796

PHASCI 3296

No. of Pages 11, Model 5G

16 June 2015 F. Franek et al. / European Journal of Pharmaceutical Sciences xxx (2015) xxx–xxx 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820

Reynolds, D.P., Lanevskij, K., Japertas, P., Didziapetris, R., Petrauskas, A., 2009. Ionization-specific analysis of human intestinal absorption. J. Pharm. Sci. 98, 4039–4054. Rowland Yeo, K., Jamei, M., Yang, J., Tucker, G.T., Rostami-Hodjegan, A., 2010. Physiologically based mechanistic modelling to predict complex drug-drug interactions involving simultaneous competitive and time-dependent enzyme inhibition by parent compound and its metabolite in both liver and gut-the effect of diltiazem on the time-course of exposure to triazolam. Eur. J. Pharm. Sci. 39, 298–309. Shah, D.A., Nakrani, R.S., Baldania, S.L., Chhalotiya, U.K., Bhatt, K.K., 2011. Stability indicating liquid chromatographic method for the estimation of desvenlafaxine in pharmaceutical dosage form. Chem. Ind. Chem. Eng. Quart. 17, 341–348. Sproule, B.A., Hazra, M., Pollock, B.G., 2008. Desvenlafaxine succinate for major depressive disorder. Drugs Today (Barc) 44, 475–487. Sugano, K., Kansy, M., Artursson, P., Avdeef, A., Bendels, S., Di, L., Ecker, G.F., Faller, B., Fischer, H., Gerebtzoff, G.g., 2010. Coexistence of passive and carriermediated processes in drug transport. Nat. Rev. Drug Discovery 9, 597–614. Sun, D., Lennernäs, H., Welage, L.S., Barnett, J.L., Landowski, C.P., Foster, D., Fleisher, D., Lee, K.D., Amidon, G.L., 2002. Comparison of human duodenum and Caco-2 gene expression profiles for 12,000 gene sequences tags and correlation with permeability of 26 drugs. Pharm. Res. 19, 1400–1416. Sung, K.C., Nixon, P.R., Skoug, J.W., Ju, T.R., Gao, P., Topp, E.M., Patel, M.V., 1996. Effect of formulation variables on drug and polymer release from HPMC-based matrix tablets. Int. J. Pharm. 142, 53–60.

11

Tchaparian, E.H., Xu, G., Jin, L., 2008. Cell based experimental models as tools for the prediction of human intestinal absorption: correlation with oral drug absorption and jejunal permeability in humans. In: Poster Presented at ISSX 2008. Tsamandouras, N., Rostami-Hodjegan, A., Aarons, L., 2015. Combining the ‘bottom up’ and ‘top down’ approaches in pharmacokinetic modelling: fitting PBPK models to observed clinical data. Br. J. Clin. Pharmacol. 79, 48–55. Turner, S., Federici, C., Hite, M., Fassihi, R., 2004. Formulation development and human in vitro-in vivo correlation for a novel, monolithic controlled-release matrix system of high load and highly water-soluble drug niacin. Drug Dev. Ind. Pharm. 30, 797–807. USP, 2011. United States Pharmacopoeia 34: General Notices and Requirements. United States Pharmacopoeia, 6. Wang, J., Flanagan, D.R., 1999. General solution for diffusion-controlled dissolution of spherical particles. 1. Theory. J. Pharm. Sci. 88, 731–738. Winiwarter, S., Bonham, N.M., Ax, F., Hallberg, A., Lennernäs, H., Karlén, A., 1998. Correlation of human jejunal permeability (in vivo) of drugs with experimentally and theoretically derived parameters. A multivariate data analysis approach. J. Med. Chem. 41, 4939–4949. Wu, C.Y., Benet, L.Z., 2005. Predicting drug disposition via application of BCS: transport/absorption/elimination interplay and development of a biopharmaceutics drug disposition classification system. Pharm. Res. 22, 11–23. Yu, L.X., Gatlin, L., Amidon, G.L., 2000. Predicting oral drug absorption in humans. In: Amidon, G.L., Lee, P. (Eds.). Marcel Dekker, New York, pp. 377–409.

821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845

Please cite this article in press as: Franek, F., et al. In vitro solubility, dissolution and permeability studies combined with semi-mechanistic modeling to investigate the intestinal absorption of desvenlafaxine from an immediate- and extended release formulation. Eur. J. Pharm. Sci. (2015), http://dx.doi.org/ 10.1016/j.ejps.2015.06.012