In vitro and in vivo performance of novel supersaturated self-nanoemulsifying drug delivery systems (super-SNEDDS)

Journal of Controlled Release 160 (2012) 25–32

Contents lists available at SciVerse ScienceDirect

Journal of Controlled Release journal homepage: www.elsevier.com/locate/jconrel

In vitro and in vivo performance of novel supersaturated self-nanoemulsifying drug delivery systems (super-SNEDDS) N. Thomas a, R. Holm b, A. Müllertz c, d,⁎, T. Rades a, c a

School of Pharmacy, University of Otago, 9054 Dunedin, New Zealand Preformulation, H. Lundbeck A/S, Ottiliavej 9, DK-2500 Valby, Denmark Department of Pharmacy, Faculty of Health and Medical Sciences, University of Copenhagen, DK-2100 Copenhagen, Denmark d Bioneer:FARMA, Department of Pharmacy, Faculty of Health and Medical Sciences, University of Copenhagen, DK-2100 Copenhagen, Denmark b c

a r t i c l e

i n f o

Article history: Received 25 December 2011 Accepted 26 February 2012 Available online 3 March 2012 Keywords: Supersaturation SNEDDS Precipitation Bioavailability Amorphous

a b s t r a c t Novel supersaturated self-nanoemulsifying drug delivery systems (super-SNEDDS) containing the poorly water-soluble drug halofantrine above equilibrium solubility (150% Seq) were compared in vitro and in vivo with conventional SNEDDS containing the drug below equilibrium solubility (75% Seq). Pre-concentrates comprising of either medium chain lipids (Captex 300/Capmul MCM) or long chain lipids (soybean oil/ Maisine), Cremophor RH40 and ethanol were formulated maintaining the lipid-to-surfactant-to-cosolvent ratio constant (55:35:10, w/w %). The ability of super-SNEDDS to increase the absorption of halofantrine in dogs, as well as the predictivity of the dynamic in vitro lipolysis model was studied. In vitro lipolysis of SNEDDS and super-SNEDDS showed rapid drug precipitation from all formulations while the same drug concentrations in the digestion medium were found during digestion of equal amounts of SNEDDS and superSNEDDS. Elevated halofantrine solubilisation during in vitro lipolysis was observed only when multiple capsules of conventional SNEDDS were subjected to in vitro digestion. After lipolysis the isolated super-SNEDDS pellets were characterised by XRPD revealing no crystalline halofantrine from any of the investigated formulations. Subsequent dissolution studies of the super-SNEDDS pellet in the lipolysis medium demonstrated enhanced dissolution of halofantrine suggesting that halofantrine in the pellet was amorphous. The enhanced dissolution of the amorphous halofantrine was also reflected in vivo since two capsules of conventional SNEDDS were needed to achieve similar AUC and Cmax as obtained after dosing of a single capsule of super-SNEDDS. The study demonstrated that the absorption of halofantrine was not hampered by drug precipitation. Super-SNEDDS lead to precipitation of halofantrine in an amorphous form, which can be the driving force for enhanced absorption. Since super-SNEDDS were also physically stable for at least 6 months they represent a potential novel oral lipid-based drug delivery system for low aqueous soluble compounds. © 2012 Elsevier B.V. All rights reserved.

1. Introduction Modern drug discovery programs have produced a wealth of new chemical entities including an increasing number of lipophilic and poorly water-soluble compounds [1,2]. The physicochemical properties of these compounds are ultimately reflected in poor and erratic absorption and varying bioavailability. In order to overcome poor water solubility and limited dissolution rate, advanced drug delivery systems are needed. One popular approach is the utilisation of lipid and surfactant based drug delivery systems, in particular selfnanoemulsifying drug delivery systems (SNEDDS) [3–8]. SNEDDS preconcentrates are isotropic mixtures of oil, surfactant, cosolvent

⁎ Corresponding author at: Department of Pharmaceutics and Analytical Chemistry, Faculty of Pharmaceutical Sciences, University of Copenhagen, DK-2100 Copenhagen, Denmark. Tel.: + 45 35 33 64 40; fax: + 45 35 33 60 30. E-mail address: [email protected] (A. Müllertz). 0168-3659/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.jconrel.2012.02.027

and drug that readily disperse in aqueous environments upon mild agitation generating ultrafine nanoemulsions. The rationale to use SNEDDS for the delivery of poorly water soluble drugs is that the drug in such preconcentrates is presented in solution, hence the dissolution step, required for solid, crystalline compounds, is avoided. Additionally, upon dispersion and subsequent digestion of SNEDDS a variety of colloidal species is formed facilitating drug absorption [3,9,10]. Traditionally, the dose of a drug that can be administered in SNEDDS is limited by the solubility of the drug in the preconcentrate. Thus, in order to avoid the administration of multiple units of a dosage form (usually soft gelatine or liquid filled hard gelatine capsules), a drug should exhibit adequate solubility in the SNEDDS preconcentrates to allow sufficient and convenient dosing for a desired therapeutic intervention. However, the solubility of drugs in the preconcentrates is often restricted, which may limit the usefulness of lipid based delivery systems [11]. Strategies for the development of SNEDDS usually include studies to ensure that the drug does not

26

N. Thomas et al. / Journal of Controlled Release 160 (2012) 25–32

precipitate upon dispersion or digestion, since re-dissolution of the precipitate is believed to decrease absorption [8,12]. This has resulted in the development of SNEDDS preconcentrates containing the drug below its equilibrium solubility (Seq), typically between 50 and 90% of Seq, restricting the access of many compounds to this promising technology [13,14]. In order to inhibit drug precipitation after dispersion of SNEDDS preconcentrates in the GI tract (GIT), the use of supersaturatable self-emulsifying drug delivery systems (S-SEDDS) has been suggested [15,16]. S-SEDDS are thermodynamically stable SEDDS containing a polymer (such as PVP or HPMC) that should inhibit excessive drug precipitation, temporarily maintaining a supersaturated solution of the drug in the GIT after dispersion, thereby possibly increasing absorption. However, this approach does not yield higher drug loads in the preconcentrates. The current paradigm aiming at the inhibition of drug precipitation might need to be put into perspective however, since recent findings obtained from in vitro lipolysis have revealed that some drugs precipitate in an amorphous form [17]. Sassene et al. [17] found considerably faster dissolution rates of cinnarizine precipitated during in vitro lipolysis compared to control experiments containing crystalline drug. The implications of this in vitro result remain yet to be elucidated in vivo. However, if precipitation is of less concern, at least for some drugs, this might enable a broader application of SNEDDS, particularly those containing increased amounts of hydrophilic excipients and greater drug loads that are more likely to exhibit drug precipitation [18]. Supersaturated delivery systems are a possible way to increase the thermodynamic activity of a drug and have been thoroughly investigated for topical drug delivery [19–21], however, to our knowledge, they have never been investigated for orally dosed SNEDDS preconcentrates. The aim of the present study therefore is to evaluate the feasibility of super-SNEDDS preconcentrates, both with regard to in vitro lipolysis and in vivo performance. Using the poorly soluble antimalarial drug halofantrine as a model drug, super-SNEDDS preconcentrates are subjected to in vitro lipolysis to assess the distribution of drug between a solubilised and a precipitated state during digestion. In vivo performance is evaluated in a bioavailability study in beagle dogs.

2. Materials and methods 2.1. Materials Halofantrine hydrochloride (HAL HCl) was purchased from APAC Pharmaceuticals LLC (Hangzhou, China). Halofantrine base (HAL) was subsequently prepared from HAL HCl as reported previously [22]. The internal standard (IS) 2,4-dichloro-6-trifluromethyl-9-(1[2-(dibutylamino)ethyl]) phenathrenemethanol hydrocloride was a kind gift of GlaxoSmithKline (West Sussex, UK). Porcine pancreatic lipase and bile extract, soybean oil (LC triglycerides), 4bromobenzeneboronic acid (BBBA), trizma maleate, calcium chloride, and sodium hydroxide were obtained from Sigma-Aldrich (St. Louis, MO, USA). Maisine 35-1 (LC mono-, di-, and triglycerides, LC mixed glycerides) was a gift from Gattefossé (St. Priest, France). Epikuron 200 (phosphatidylcholine) was supplied by Cargill (Hamburg, Germany). Cremophor RH40 (PEG 40 hydrogenated castor oil) was donated by BASF (Ludwigshafen, Germany). Captex 300 (MC triglycerides) and Capmul MCM (MC mono-, di-, and triglycerides, MC mixed glycerides) were a kind gift from Abitec (Columbus, OH, USA). HPLC grade acetonitrile, methanol, tert. butylmethylether (TBME), and sodium chloride were obtained from Merck (Darmstadt, Germany). Purified water was obtained from a Millipore Milli-Q Ultrapure Water purification system (Billerica, MA, USA). All other chemicals were of analytical grade.

2.2. Determination of saturation solubility The equilibrium solubility of HAL in preconcentrates was determined at 25 °C. Approximately 3 g SNEDDS preconcentrate was added to amber Teflon-sealed glass vials containing 0.3 g drug. The suspensions were stirred on a magnetic stirrer at 25 °C for up to 7 days with the help of a magnetic stirrer bar (10 mm). After 6 days the suspensions were centrifuged at 13,400 rpm (12.1 × 103 g at rmax) for 15 min at room temperature (Eppendorf, Wesseling, Germany) followed by filtration of the supernatant through a 0.2 μm PTFE filter (Millipore, Billerica, MA, USA). After appropriate dilution with methanol drug concentration in the filtrates were quantified by HPLC. After 7 days the procedure was repeated and equilibrium solubility was confirmed by drug concentrations of two consecutive measurements not varying by more than 5% [23]. Similarly, HAL equilibrium solubility was determined in the aqueous supernatants obtained after ultracentrifugation from digested blank (drug free) preconcentrates. Quantification of HAL by HPLC was carried out on an Agilent 1200 system (Agilent Technologies, Santa Clara, CA, USA) with autosampler employing a validated method described previously [24]. 2.3. Preparation of preconcentrates For the current study previously developed and characterised SNEDDS preconcentrates were used, based either on medium (MC) or long chain (LC) lipids, containing the same ratios of lipid, surfactant and cosolvent [25]. The lipid phase was prepared by blending the triglycerides with the molten mixed glycerides. The lipid mixture was then cooled to ambient temperature and accurately weighted into dust-free glass vials. The surfactant and ethanol were added to the lipid mixture and the vials were sealed with Teflon-lined screw-caps followed by intimate mixing, generating isotropic, liquid SNEDDS preconcentrates. Drug loaded preconcentrates were prepared by weighing the required amount of drug into dust-free glass vials and addition of preconcentrates. Drug loads are expressed as percentage of the equilibrium solubility of HAL in preconcentrates at 25 °C and are summarised in Table 1. Prior to sealing of the glass vials the preconcentrates were purged with nitrogen to avoid oxidation of the lipids. Preconcentrates containing drug loads below Seq were ultrasonicated for 5 min to aid the drug dissolution process after which the vials were placed in a waterbath (TW20, Julabo, Seelbach, Germany) for equilibration at 37 °C. For the preparation of supersaturated preconcentrates (super-SNEDDS, drug above Seq) sonication was extended to 30 min. The preconcentrates were further subjected to a heating phase (5 h at 60 °C), followed by a cooling phase (overnight at 37 °C). After the drug had dissolved macroscopically the preconcentrates were investigated by polarising light microscopy, PLM, (Motic BA 300 POL, Richmond, BC, Canada) for the presence of undissolved drug. Before SNEDDS and super-SNEDDS were subjected to in vitro lipolysis and before administration to the study animals 0.8 g of the preconcentrate were filled into hard gelatine capsules (size 000; Capsugel, North Peapack, NJ, USA). 2.4. Stability of SNEDDS and super-SNEDDS preconcentrates In order to assess the chemical and physical stability of SNEDDS and super-SNEDDS sealed glass vials containing approximately 5 g Table 1 Drug loads of MC-SNEDDS, MC-super-SNEDDS, LC-SNEDDS, and LC-superSNEDDS. MC-SNEDDS (75% HAL Seq) MC-super-SNEDDS (150% HAL Seq) LC-SNEDDS (75% HAL Seq) LC-super-SNEDDS (150% HAL Seq)

35.2 mg/g 70.4 mg/g 48.0 mg/g 96.0 mg/g

N. Thomas et al. / Journal of Controlled Release 160 (2012) 25–32

super-SNEDDS preconcentrates were stored at 25 °C for 6 months after which they were analysed for HAL content by HPLC. Furthermore super-SNEDDS were investigated both visually and by PLM on a regular basis for the presence of crystalline drug precipitating during storage.

2.5. Particle size characterisation of dispersed preconcentrates 0.8 g SNEDDS and super-SNEDDS preconcentrates were filled in hard gelatin capsules and were immersed in 300 ml Milli-Q water (37 °C) with stainless steel sinkers. A type 2 USP dissolution apparatus (DT 600, Erweka, Heusenstamm, Germany), equipped with standard vessels and paddle speed set at 100 rpm was employed for stirring. After 30 min the particle size of the resulting dispersions was measured without further dilution of the sample by dynamic light scattering (DLS) using a nanosizer-ZS (Malvern Instruments, Worcestershire, UK) at 37 °C.

27

2.8. Dissolution studies of halofantrine precipitated during lipolysis Prior to dissolution studies performed in a standard USP dissolution 2 apparatus (DT 600, Erweka, Germany) the solubility of HAL in lipolysis medium was determined (pH 6.5, 37 °C) after 72 h to estimate sink conditions. After ultracentrifugation three pellets obtained from three centrifugation tubes were combined and dispersed in 5 ml Milli-Q water. The dispersion was then added to 445 ml lipolysis medium maintained at pH 6.5 and 37 °C. The pH was verified before and after dissolution and solubility experiments. The paddle speed was set at 75 rpm and dissolution was carried out over 180 min. During the dissolution study 3 ml samples were withdrawn at designated time points (1, 2, 5, 10, 20, 40, 60, 120, and 180 min) and the volume was substituted by fresh lipolysis medium of the same temperature and pH. Samples were filtered (0.45 μm, PTFE, Millipore, MA, USA) and the first milliliter of the filtrate was discarded. Following appropriate dilution drug concentration in the filtrate was analysed by HPLC. 2.9. In vivo study

2.6. In vitro lipolysis Dynamic in vitro lipolysis was carried out as described previously [26,27] with some modifications. Briefly, one or two capsules containing SNEDDS (0.8 g/capsule) or one capsule super-SNEDDS preconcentrates (0.8 g) was immersed into 250 ml lipolysis medium (containing 5 mM bile salt, 1.25 ml phosphatidyl choline, 2 mM trizma maleate buffer, 150 mM sodium chloride) with the aid of a stainless steel sinker. After 7 min the pH was automatically adjusted to 6.5 by a computer-controlled pH-stat device (Metrohm Titrino 744, Tiamo Version 1.3, Switzerland) before in vitro lipolysis was initiated by the addition of 50 ml freshly prepared pancreatic lipase solution (pH 6.5, 37 °C). The lipase activity in the final 300 ml lipolysis medium was 350 U/ml and the rate of lipolysis was controlled by the constant addition of CaCl2 (0.5 M, 0.045 mmol/min) throughout the duration of lipolysis (60 min). The fatty acids released during in vitro digestion were automatically titrated with 1 M NaOH to maintain the pH at 6.5. 9 ml digestion medium were withdrawn in 15 min intervals up to 60 min in vitro lipolysis. In each sample lipase activity was immediately inhibited by the addition of 45 μl 4-bromobenzeneboronic acid (1 M in methanol). This was followed by ultracentrifugation of 8.0 g samples in 13.5 ml polycarbonate tubes at 50,000 rpm (2.3 × 10 5 g at rmax) for 50 min at 37 °C in a Beckman L-80 Ultracentrifuge using a 70.1 Ti rotor (Beckman Coulter, USA). After appropriate dilution the resulting clear supernatant and the pellet were subsequently analysed by HPLC for their HAL content.

2.7. X-ray powder diffraction The isolated pellets obtained after 60 min in vitro lipolysis of super-SNEDDS were analysed within 2 h after sampling by X-ray powder diffraction (XRPD) in order to elucidate the solid state of HAL present in the pellet. The X-ray diffractometer consisted of a PANalytical X'Pert Pro MPD PW 3040/60 (PANalytical B.V., Almelo, The Netherlands) with CuKα as radiation source (1.542 Å). Control experiments were conducted using blank pellets obtained from lipolysis of drug-free preconcentrates. The pellets were spiked with the corresponding amount of crystalline drug present in pellets obtained from drug loaded preconcentrates. The pellets were applied to flat aluminium holders and scanned from 5 to 35° (2θ) using a scanning speed of 0.1285°/min and a step size of 0.0084° with an applied voltage and current of 40 kV and 30 mA, respectively. The diffractograms were prepared by X'Pert High Score version 2.2.4 (PANalytical B.V., Almelo, The Netherlands).

All animal care and experimental studies were approved by the Animal Welfare Committee, appointed by the Danish Ministry of Justice, and were carried out in compliance with EC Directive 2010/ 63/EU, the Danish law regulating experiments on animals and NIH Guidelines for the Care and Use of Laboratory Animals. The SNEDDS and super-SNEDDS were tested in 6 male beagle dogs (11.8– 13.8 kg) in a non-randomised cross-over study, with a washout period of 7 days between each treatment. The dogs were fasted for 20–24 h before the initiation of the experiment and fed again 8 h after administration. The dose was given either as an intravenous injection of an emulsion (1.5 ml/kg), prepared as described previously [28], and orally as one or two capsules containing freshly prepared SNEDDS or super-SNEDDS formulations (0.80 g/capsule). Blood samples (0.5 ml) were taken from the cephalic vein by individual venepuncture and collected into EDTA coated tubes. Samples were collected before administration and after 30, 60, 90 min, and 2, 3, 4, 6, 8, 24 and 28 h after drug administration and additionally after 5 and 15 min for the i.v. group. Plasma was harvested immediately by centrifugation at 3200 g (15 min, 4–8 °C) and stored at −20 °C until further analysis. 2.10. Quantitative analysis of plasma samples Halofantrine plasma samples were processed using the method described by Humberstone et al. [24] with some modifications. Briefly, plasma (0.2 ml) was spiked with IS (100 μl, 2 μg/ml in acetonitrile), proteins were precipitated with 1 ml acetonitrile, followed by liquid–liquid-extraction with 6 ml TBME. After centrifugation for 10 min at 4000 rpm and 8 °C (6–16 K, Sigma, Osterode am Harz, Germany) the supernatant was transferred to plastic vials containing 100 μl 0.005 M HCl in acetonitrile. The ether phase was evaporated under a stream of nitrogen for 50 min at 40 °C (TurboVap LV, Zymark, MA, USA). Following reconstitution of the residue in 0.1 ml methanol, 25 μl of the sample was injected in a LaChrom HPLC system (Merck, Darmstadt, Germany) and analysed as previously described [29]. 2.11. Pharmacokinetic analyses The pharmacokinetic parameters following oral administration of HAL were obtained by a non-compartmental analysis using WinNonlin Professional version 5.2 (Pharsight Corporation, Mountain View, CA, USA). The area under the curve (AUC) of the plasma concentration versus time curves was determined by the linear trapezoidal method from time t = 0 h to the last plasma concentration measured at t = 28 h after dosing. The maximum HAL plasma concentration

28

N. Thomas et al. / Journal of Controlled Release 160 (2012) 25–32

(Cmax) and its time necessary to occur (tmax) were determined from the individual plasma concentration versus time curves and the absolute bioavailability (Fa) was calculated as: Fa ¼

   AUCPO DoseIV ⋅ AUCIV DosePO

where PO denotes per oral and IV intravenous administration of the formulations. The plasma concentration–time profiles of halofantrine after intravenous dosing were fitted to a two compartment model. 2.12. Statistical analysis GraphPad Prism Version 5.01 (GraphPad Software, San Diego, CA, USA) was used for statistical analyses of the data. Student's t-test and analyses of variance (ANOVA) followed by Tukey's post-test were employed to analyse the mean of two groups and more than two groups, respectively (α = 0.05).

of 60 nm (72%) and 4.8 μm (28%), the latter representing the upper limit for particle size measurements by DLS. The DLS results were macroscopically reflected as the drug-loaded dispersions were slightly milky emulsions compared to the clear-opalescent appearance of drug-free dispersions. Upon increasing the HAL concentration to 150% Seq in MC-super-SNEDDS, dispersions were even milkier in appearance not feasible for reliable size determination by DLS, even after further dilution of the samples. Upon dispersion of LC-SNEDDS (75% Seq HAL drug load), the particle size changed from 45 nm (blank LC-SNEDDS) to approximately 80 nm (80% of the population) and 1.1 μm (20%) while the PDI increased from 0.07 to 0.42. Dispersions of LC-super-SNEDDS containing 150% HAL load were heterogeneous (as seen with MC-super-SNEDDS) and were therefore not suitable for DLS measurements. The increase in particle size, PDI and the concomitant milkiness are most likely due to the precipitation of HAL from the dispersed preconcentrates as a result of the reduced solubilisation capacity caused by the dilution of the cosolvent in the bulk phase. This was verified using PLM where aggregates of crystalline drug were visible in dispersed super-SNEDDS (data not shown).

3. Results and discussion 3.4. In vitro lipolysis of preconcentrates 3.1. Solubility of compounds in preconcentrates

The content of HAL in SNEDDS and super-SNEDDS preconcentrates stored at 25 °C in sealed glass vials was determined 6 months after preparation. The HAL content in MC- and LC-SNEDDS corresponded to the declared amount. Approximately 97% and 92% of the declared amount of HAL were found in MC-super-SNEDDS and LCsuper-SNEDDS, respectively, indicating reasonable chemical stability. Since the supersaturated state is not thermodynamically stable, MC- and LC-super-SNEDDS were inspected visually and by polarising light microscopy (PLM) for the presence of precipitated drug in preconcentrates during storage in glass vials after their preparation on a regular basis. No crystalline drug was observed indicating that formulations were physically stable for 6 months (upon submission of the manuscript). 3.3. Droplet size of dispersed SNEDDS and super-SNEDDS Hard gelatin capsules filled with SNEDDS and super-SNEDDS typically disintegrated and fully dispersed their content within 6 min. Since particle size has been commonly used for SNEDDS characterisation the particle size was measured by DLS after 30 min dispersion time in Milli-Q water. Blank SNEDDS, containing no drug, generated uniform dispersions (polydispersity index (PDI) b 0.1) and particle sizes between 33 nm and 45 nm (z-average) for dispersed MC and LC-SNEDDS, respectively. When HAL was incorporated into MC-SNEDDS at 75% Seq the PDI increased to 0.7 indicating a heterogeneous dispersion. In the DLS measurement two distinctive particle populations (estimated from the intensity signal) were observed, with an approximate diameter

a) Halofantrine concentration (μg/ml)

3.2. Stability of SNEDDS and super-SNEDDS

The concentration of HAL in solution during in vitro lipolysis of MC-SNEDDS and MC-super-SNEDDS preconcentrates is depicted in Fig. 1, panel a). 200

150

100

50

0 0

15

30

45

60

45

60

Time (min)

b) Halofantrine concentration (μg/ml)

HAL (cLogP 8.5) is a very lipophilic drug and is practically insoluble in water [30,31]. Due to the higher solubility in lipids, HAL free base was used in this study rather than the commercially available halofantrine hydrochloride. The solubility of HAL in MC-SNEDDS preconcentrates was 47.0 ± 2.9 mg/g but could be increased to 64.0 ± 3.3 mg/g by substitution of the lipid component in MC-SNEDDS (medium chain lipids derived from coconut oil) by long chain lipids (soybean oil, Maisine) present in LC-SNEDDS. In order to maintain the relative composition of the preconcentrates and thus to allow a better comparison between the systems, the same lipid/surfactant/ cosolvent ratio (55/35/10, % w/w) was used for both MC- and LCSNEDDS preconcentrates.

1000

750

200

100

0 0

15

30

Time (min) Fig. 1. HAL in solution during lipolysis of SNEDDS and super-SNEDDS preconcentrates based on MC-lipids (panel a) and LC-lipids (panel b). Formulations included: (○) 1 capsule SNEDDS, (□) 2 capsules SNEDDS, and (◊) 1 capsule super-SNEDDS. The drug loads were 75% of HAL equilibrium solubility (Seq) in SNEDDS and 150% HAL Seq in super-SNEDDS. The Seq of HAL in the lipolysis medium containing digestion products of one capsule drug-free preconcentrates at corresponding time points is indicated by *. Data represent mean ± SD, n = 3.

N. Thomas et al. / Journal of Controlled Release 160 (2012) 25–32

The pellets obtained after 60 min in vitro lipolysis of MC- and LCsuper-SNEDDS were investigated by XRPD. The typical diffraction patterns of crystalline HAL were visible in blank pellets spiked with the corresponding amount of HAL present in MC-super-SNEDDS (1.2 mg, equivalent to approximately 12% (w/w) of the dry pellet mass) (Fig. 2, panel a) and LC-super-SNEDDS (1.6 mg, 10% w/w) (Fig. 2, panel b). In contrast these characteristic peak patterns were absent both in MC-super-SNEDDS and LC-super-SNEDDS pellets. Since no crystalline HAL patterns could be observed using XRPD, the data suggests that HAL administered in MC-super-SNEDDS and LC-super-SNEDDS precipitated in an amorphous form during in vitro lipolysis. 3.6. Dissolution study The dissolution profiles of HAL from pellets derived from MCsuper-SNEDDS, LC-super-SNEDDS and from blank pellets spiked with the corresponding amount of crystalline drug are shown in Fig. 3. The data indicates that the dissolution rates of HAL from MCsuper-SNEDDS and LC-super-SNEDDS pellets were considerably faster than dissolution of crystalline HAL in the control experiments. As an example, approximately twice as much HAL had dissolved after 30 min from super-SNEDDS pellets, independent of the type of lipid (MC or LC lipids) used for the preconcentrates. The observed

intensity (a.u.)

crystalline HAL 150% MC HAL super-SNEDDS blank MC pellet spiked with crystalline HAL blank MC pellet 10

15

20

25

30

35

Diffraction angle (°2θ)

b) crystalline HAL 150% LC HAL super-SNEDDS blank LC pellet spiked with crystalline HAL blank LC pellet 10

15

20

25

30

35

Diffraction angle (°2θ) Fig. 2. XRPD patterns of crystalline halofantrine (HAL), super-SNEDDS pellets, blank pellets spiked with crystalline HAL, and blank pellets obtained after 60 min in vitro lipolysis and ultracentrifugation. Panels a) and b) represent diffractograms of MC(super)-SNEDDS and LC-(super)-SNEDDS, respectively.

rapid dissolution therefore supports the results obtained from XRPD suggesting the presence of amorphous HAL.

3.7. In vivo study The results obtained from the in vitro lipolysis and dissolution experiments were compared with an in vivo pharmacokinetic study carried out in six fasted beagle dogs. The mean plasma concentration versus time profiles of HAL administered orally as SNEDDS (either one or two capsules, 75% HAL load) and as a single super-SNEDDS capsule (150% HAL load) are presented in Fig. 4.

10

Halofantrine concentration (μg/ml)

3.5. Solid state characterisation of the precipitates

a)

intensity (a.u.)

The lipolysis of two capsules containing MC-SNEDDS (dotted line, panel a)) generated higher concentration of HAL solubilised in the aqueous phase isolated by ultra-centrifugation compared to lipolysis of one capsule MC-SNEDDS and MC-super-SNEDDS. This can most likely be attributed to the generation of more digestion products when adding two capsules as compared to only one capsule of either MC-SNEDDS or MC-super-SNEDDS, facilitating the solubilisation of HAL. Furthermore, the lipolysis of one capsule containing MCSNEDDS and MC-super-SNEDDS yielded comparable concentration versus time profiles. This was also observed during lipolysis of one capsule LC-SNEDDS and LC-super-SNEDDS (Fig. 1, panel b)). However, since one capsule super-SNEDDS contained twice the amount of HAL than one capsule conventional SNEDDS, more HAL precipitated during lipolysis of the super-SNEDDS. The quantification of pellets obtained after 60 min in vitro lipolysis of one capsule LC-superSNEDDS, for example, revealed that approximately 1.6 mg HAL had precipitated whereas only 0.8 mg HAL was found in the pellet originating from one capsule LC-SNEDDS. The equilibrium solubility of HAL in the aqueous phase isolated by ultra-centrifugation during lipolysis of one capsule (drug-free) MCand LC-SNEDDS (shown as stars in Fig. 1) was compared with the HAL concentration found in the aqueous phase during lipolysis of drug-loaded MC-SNEDDS and LC-SNEDDS at corresponding time points. The largest solubilisation of HAL was found at the start of lipolysis (t = 0 min) where the solubilisation capacity of the digestion medium is mainly governed by undigested formulation and bile salts. This was particularly pronounced for dispersed LC-SNEDDS since the equilibrium solubility of HAL was approximately sevenfold greater compared to the digestion medium containing dispersed MC-SNEDDS. However, during the progress of lipolysis and the continuous digestion of the formulation the solubilisation capacity of both MC- and LC-SNEDDS and their digestion products rapidly decreased. Since the concentration of HAL during lipolysis of one capsule MC- and LC-SNEDDS and the corresponding super-SNEDDS never exceeded the corresponding equilibrium solubility of HAL in the digestion products the data suggests that no supersaturated solutions of HAL in the lipolysis medium were generated throughout lipolysis. Instead, rapid precipitation of HAL occurred and thus the solid state properties of the precipitates were evaluated.

29

8

6

4

2

0 0

30

60

90

120

150

180

Time (min) Fig. 3. Dissolution profiles of HAL in MC- (□) and LC-super-SNEDDS (Δ) pellets and corresponding blank pellets spiked with crystalline HAL (closed symbols). Data represent mean ± SD, n = 3.

30

N. Thomas et al. / Journal of Controlled Release 160 (2012) 25–32

plasma conc Hf (ng/ml)

a)

1500

1000

500

0 0

10

20

30

20

30

Time (h)

plasma conc Hf (ng/ml)

b)

4000

3000

2000

1000

0 0

10

Time (h) Fig. 4. Plasma concentration of HAL following oral administration as SNEDDS and super-SNEDDS based on MC-lipids (panel a) and LC-lipids (panel b) to beagle dogs. Treatments included 1 capsule 75% HAL SNEDDS (○), 2 capsules 75% HAL SNEDDS (■), and 1 capsule 150% HAL super-SNEDDS (◊). The drug loads are expressed as % of the equilibrium solubility of HAL in SNEDDS and super-SNEDDS. Data represent mean ± SEM, n = 6.

The pharmacokinetic parameters for MC-SNEDDS and MC-superSNEDDS are provided in Table 2, the corresponding parameters for LC-SNEDDS and LC-super-SNEDDS are provided in Table 3. The dosing of one capsule MC-super-SNEDDS resulted in a significant, approximately 2.5 to 3-fold increase, in the area under the curve (AUC) and in the peak plasma concentration (Cmax) compared to the dosing of one capsule MC-SNEDDS (P b 0.05). The AUC and Cmax increased after dosing of one capsule MC-super-SNEDDS compared to two capsules MC-SNEDDS, but this was not statistically significant (P > 0.05). The AUC and Cmax after dosing of one capsule MC-super-SNEDDS and two capsules MC-SNEDDS were not statistically different from each other (P > 0.05). Similarly, two capsules LC-SNEDDS were needed to achieve the same AUC and Cmax generated by only one capsule LC-super-SNEDDS (P > 0.05). Generally a high variation was observed within the MC treatments. In contrast, LCSNEDDS and LC-super-SNEDDS showed considerably less variability in the pharmacokinetic profiles, except for one dog receiving LCsuper-SNEDDS showing very high plasma concentrations compared to the other five animals. LC-SNEDDS and LC-super-SNEDDS

demonstrated a considerably higher absolute bioavailability (approximately 80 ± 9%) compared to the corresponding MC-SNEDDS and MC-super-SNEDDS. This could be explained by the improved solubilisation for HAL mediated by LC-lipids and their digestion products (see lipolysis graphs), possible lymphatic uptake of HAL and secretion of bile acids stimulated by LC-lipids, further promoting drug solubilisation [32–34]. Compared to the previously reported [24] absolute bioavailability of the marketed halofantrine hydrochloride tablets (8.6 ± 5.3%) the current study demonstrates a 4- to 8-fold increase in the absolute bioavailability of HAL when dosed in a SNEDDS. There were no changes in the time to reach peak plasma concentration (tmax) across all six formulations indicating that the limited quantity of lipid used in this study (approximately 0.4 g per capsule) all had a very similar effect on gastric emptying. In vitro lipolysis has commonly been employed for the in vitro assessment of lipid formulations such as SNEDDS [13,35,36]. The results obtained during this study revealed that more HAL was solubilised when two capsules of SNEDDS were subjected to in vitro lipolysis compared to lipolysis of one capsule filled with super-SNEDDS containing the same amount of HAL. This result was expected since the hydrolysis of more lipids generates more fatty acids and monoglycerides facilitating drug solubilisation by the formation of mixed micelles and other colloidal species. Furthermore, HAL formulated as MC- and LC-super-SNEDDS rapidly precipitated during in vitro digestion of super-SNEDDS leading to a similar HAL concentration in the lipolysis medium as generated during lipolysis of one capsule MC-SNEDDS and LC-SNEDDS, respectively. Based on this observation and the current assumption to avoid precipitation of drugs during in vitro dispersion and digestion, a similar performance of MC- and LCsuper-SNEDDS compared to one capsule MC- and LC-SNEDDS would have been expected in vivo. However, this was not the case. In fact, the in vivo performance of MC- and LC-super-SNEDDS can be considered equivalent to that of two capsules of corresponding SNEDDS (containing the same HAL dose). This finding can be explained by the solid state properties of the precipitated HAL. The current study suggests precipitation of HAL in an amorphous form demonstrating an inherent higher dissolution rate compared to the crystalline form. The precipitation of HAL was intentionally provoked using supersaturated MC- and LC-SNEDDS. The precipitation of HAL in an amorphous form might be explained by the rapid dispersion and digestion of the super-SNEDDS in the digestion medium. Moreover, the bile salts or other compounds present in the digestion medium might stabilise amorphous halofantrine by an interaction similar to the HAL-taurocholate complex suggested by Humberstone et al. [37]. Despite substantial precipitation MC- and LC-super-SNEDDS did not adversely affect the AUC and Cmax. It appears that the precipitated, amorphous HAL originating from MC- and LC-super-SNEDDS was still contributing to absorption. The data suggests that the increased dissolution rate originating from the amorphous HAL precipitate could be the driving force for improved absorption, resulting in similar pharmacokinetics as obtained from multiple units of conventional SNEDDS. Since a drug has to be in solution prior to absorption it might be possible that in the direct vicinity of the precipitating,

Table 2 Pharmacokinetic parameters of HAL following oral administration to dogs in MC-SNEDDS and MC-super-SNEDDS. Data represents mean ± SEM, n = 6. Parameter

MC-SNEDDS 1 cap (28.2 mg HAL)

MC-SNEDDS 2 cap (56.4 mg HAL)

MC-super-SNEDDS 1 cap (56.4 mg HAL)

Cmax (ng/ml) tmax (h) AUC (0–28) (ng∙h/ml) Absolute bioavailability (%)*

388 ± 64.5a 2.7 ± 0.3 4458 ± 639a 41.0 ± 5.0

621 ± 256ab 3.0 ± 0.4 6826 ± 2234ab 31.5 ± 10.4

1167 ± 245b 2.5 ± 0.2 11,263 ± 1719b 53.2 ± 8.6

*Absolute bioavailability was determined as: 100 × (AUCp.o./dosep.o) / (AUCi.v./dosei.v.) after i.v. administration of HAL in an MC-lipid emulsion (1.5 mg/kg); AUC(0–28) was 7489 ± 552 ng∙h/ml. Numbers with different superscripts indicate significant differences.

N. Thomas et al. / Journal of Controlled Release 160 (2012) 25–32

31

Table 3 Pharmacokinetic parameters of HAL following oral administration to dogs in LC-SNEDDS and LC-super-SNEDDS. Data represents mean ± SEM, n = 6. Parameter

LC-SNEDDS 1 cap (38.4 mg HAL)

LC-SNEDDS 2 cap (76.8 mg HAL)

LC-super-SNEDDS 1 cap (76.8 mg HAL)

Cmax (ng/ml) tmax (h) AUC (0–28) (ng∙h/ml) Absolute bioavailability (%)*

1199 ± 94.0a 2.0 ± 0.5 10,768 ± 718a 73.3 ± 4.8

2808 ± 464b 2.2 ± 0.4 22,502 ± 1474b 76.2 ± 2.8

3128 ± 444b 1.7 ± 0.2 26,824 ± 3626b 90.9 ± 11.4

*Absolute bioavailability was determined as: 100 × (AUCp.o./dosep.o) / (AUCi.v./dosei.v.) after i.v. administration of HAL in an MC-lipid emulsion (1.5 mg/kg); AUC(0–28) was 7489 ± 552 ng∙h/ml. Numbers with different superscripts indicate significant differences.

amorphous HAL local supersaturation occurred leading to increased drug absorption. This suggests that precipitation does not necessarily have to be avoided. However, in order to provide for predictable absorption the solubilisation capacity of the local environment should be large enough (as in the case of LC-digestion products) to facilitate reliable absorption from the rapidly re-dissolving amorphous drug. Therefore, formulations based on LC-lipids appeared advantageous promoting greater drug solubilisation and, ultimately, bioavailability of HAL. The present study also implies that the results obtained from in vitro lipolysis have to be interpreted with care. The model has been developed to gain a better understanding of the fate of lipid-based drug delivery systems during digestion and, ultimately, to reveal in vitro/in vivo correlations. While rank-order correlations were possible in some cases, the model could not explain all of the in vivo data in other studies [38,39]. This could be due to the fact that the currently employed digestion models are not designed to consider absorption. As seen in this study rapid dissolution of amorphous precipitates and subsequent absorption could potentially override the previously assumed negative effects of (crystalline) precipitation during in vitro lipolysis, particularly for lipophilic drugs such as HAL, for which absorption is not rate-limiting. Due to their increased drug load super-SNEDDS preconcentrates could be an alternative to conventional SNEDDS preconcentrates. In the current study super-SNEDDS preconcentrates showed rapid precipitation in vitro generating amorphous precipitates. Provided the in vitro lipolysis model reflects the in vivo conditions, the enhanced dissolution of the amorphous precipitate could promote drug absorption while avoiding the inconvenience of multiple dosing. Further studies will be needed to elucidate the importance of excipients on drug precipitation and to predict the solid state properties of precipitates originating during in vitro digestion. The in vitro lipolysis model coupled with an absorption model and solid state characterisation of the pellets could be a valuable tool to increase the knowledge about these promising delivery systems. 4. Conclusion The present study has demonstrated the feasibility of a SNEDDS formulation containing halofantrine at supersaturated levels with reasonable stability over several months. During in vitro lipolysis rapid precipitation of HAL from super-SNEDDS occurred resulting in an amorphous precipitate of HAL that demonstrated enhanced dissolution characteristics. Super-SNEDDS preconcentrates were tested in vivo where they showed similar performance as generated by the same HAL dose administrated in 2 SNEDDS capsules containing HAL at 75% of Seq. This study revealed that drug precipitation during dispersion and digestion cannot necessarily be translated to inferior performance in vivo. Therefore, solid state characterisation of pellets originating during in vitro lipolysis should accompany in vitro digestion experiments in order to understand the physical nature of the precipitate. As this is the first study of its nature more studies will be needed to elucidate which factors govern drug precipitation in vitro and its implication in vivo.

Acknowledgements The personnel in the animal facilities at H. Lundbeck A/S are acknowledged for their skilful handling of the dogs. References [1] C.A. Lipinski, F. Lombardo, B.W. Dominy, P.J. Feeney, Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings, Adv. Drug Deliv. Rev. 23 (1997) 3–25. [2] S. Stegemann, F. Leveiller, D. Franchi, H. de Jong, H. Linden, When poor solubility becomes an issue: from early stage to proof of concept, Eur. J. Pharm. Sci. 31 (2007) 249–261. [3] S. Chakraborty, D. Shukla, B. Mishra, S. Singh, Lipid — an emerging platform for oral delivery of drugs with poor bioavailability, Eur. J. Pharm. Biopharm. 73 (2009) 1–70. [4] P.P. Constantinides, Lipid microemulsions for Improving drug dissolution and oral absorption: physical and biopharmaceutical aspects, Pharm. Res. 12 (1995) 1561–1572. [5] D.H. Hauss, Oral lipid-based formulations, Adv. Drug Deliv. Rev. 59 (2007) 667–676. [6] W.N. Humberstone, Charman, lipid-based vehicles for the oral delivery of poorly water soluble drugs, Adv. Drug Deliv. Rev. 25 (1997) 103–128. [7] C.J.H. Porter, C.W. Pouton, J.F. Cuine, W.N. Charman, Enhancing intenstinal drug solubilisation using lipid-based delivery systems, Adv. Drug Deliv. Rev. 60 (2008) 673–691. [8] C.J.H. Porter, N.L. Trevaskis, W.N. Charman, Lipids and lipid-based formulations: optimizing the oral delivery of lipophilic drugs, Nat. Rev. Drug Discov. 6 (2007) 231–248. [9] D.G. Fatouros, B. Bergenstahl, A. Müllertz, Morphological observations on a lipidbased drug delivery system during in vitro digestion, Eur. J. Pharm. Sci. 31 (2007) 85–94. [10] D.G. Fatouros, I. Walrand, B. Bergenstahl, A. Müllertz, Colloidal structures in media simulating intestinal fed state conditions with and without lipolysis products, Pharm. Res. 26 (2008) 361–374. [11] T.D. Thi, M. Van Speybroeck, V. Barillaro, J. Martens, P. Annaert, P. Augustijns, J. Van Humbeeck, J. Vermant, G. Van den Mooter, Formulate-ability of ten compounds with different physicochemical profiles in SMEDDS, Eur. J. Pharm. Sci. 38 (2009) 479–488. [12] J.F. Cuiné, C.L. Claire, L. McEvoy, W.N. Charman, C.W. Pouton, G.A. Edwards, H. Benameur, C.J.H. Porter, Evaluation of the impact of surfactant digestion on the bioavailability of danazol after oral administration of lipidic self-emulsifying formulations to dogs, J. Pharm. Sci. 97 (2008) 995–1012. [13] J. Christensen, K. Schultz, B. Mollgard, H.G. Kristensen, A. Müllertz, Solubilisation of poorly water-soluble drugs during in vitro lipolysis of medium- and long-chain triacylglycerols, Eur. J. Pharm. Sci. 23 (2004) 287–296. [14] A.M. Kaukonen, B.J. Boyd, C.J.H. Porter, W.N. Charman, Drug solubilization behavior during in vitro digestion of simple triglyceride lipid solution formulations, Pharm. Res. 21 (2004) 245–253. [15] P. Gao, A. Akrami, F. Alvarez, J. Hu, L. Li, C. Ma, S. Surapaneni, Characterization and optimization of AMG 517 supersaturatable self-emulsifying drug delivery system (S-SEDDS) for improved oral absorption, J. Pharm. Sci. 98 (2009) 516–528. [16] P. Gao, B.D. Rush, W.P. Pfund, T. Huang, J.M. Bauer, W. Morozowich, M.-S. Kuo, M.J. Hageman, Development of a supersaturable SEDDS (S-SEDDS) formulation of paclitaxel with improved oral bioavailability, J. Pharm. Sci. 92 (2003) 2386–2398. [17] P.J. Sassene, M.M. Knopp, J.Z. Hesselkilde, V. Koradia, A. Larsen, T. Rades, A. Müllertz, Precipitation of a poorly soluble model drug during in vitro lipolysis: characterization and dissolution of the precipitate, J. Pharm. Sci. 99 (2010) 4982–4991. [18] J.F. Cuiné, W.N. Charman, C.W. Pouton, G.A. Edwards, C.J.H. Porter, Increasing the proportional content of surfactant (Cremophor EL) relative to lipid in selfemulsifying lipid-based formulations of danazol reduces oral bioavailability in beagle dogs, Pharm. Res. 24 (2004) 748–757. [19] K. Moser, K. Kriwet, A. Naik, Y.N. Kalia, R.H. Guy, Passive skin penetration enhancement and its quantification in vitro, Eur. J. Pharm. Biopharm. 52 (2001) 103–112. [20] R.H. Müller, M. Radtke, S.A. Wissing, Solid lipid nanoparticles (SLN) and nanostructured lipid carriers (NLC) in cosmetic and dermatological preparations, Adv. Drug Deliv. Rev. 54 (2002) S131–S155 (Supplement).

32

N. Thomas et al. / Journal of Controlled Release 160 (2012) 25–32

[21] M.A. Pellett, S. Castellano, J. Hadgraft, A.F. Davis, The penetration of supersaturated solutions of piroxicam across silicone membranes and human skin in vitro, J. Control. Release 46 (1997) 205–214. [22] C.J.H. Porter, S.A. Charman, W.N. Charman, Lymphatic transport of halofantrine in the triple-cannulated anesthetized rat model: effect of lipid vehicle dispersion, J. Pharm. Sci. 85 (1996) 351–356. [23] G.A. Kossena, B.J. Boyd, C.J.H. Porter, W.N. Charman, Separation and characterization of the colloidal phases produced on digestion of common formulation lipids and assessment of their impact on the apparent solubility of selected poorly water-soluble drugs, J. Pharm. Sci. 92 (2003) 634–648. [24] A.J. Humberstone, G.J. Currie, C.J.H. Porter, M.J. Scanlon, W.N. Charman, A simplified liquid chromatography assay for the quantitation of halofantrine and desbutylhalofantrine in plasma and identification of a degradation product of desbutylhalofantrine formed under alkaline conditions, J. Pharm. Biomed. Anal. 13 (1995) 265–272. [25] N. Thomas, A. Müllertz, A. Graf, T. Rades, Influence of lipid composition and drug load on the in vitro performance of self-nanoemulsifying drug delivery systems (SNEDDS), J. Pharm. Sci. 101 (2012) 1721–1731, doi:10.1002/jps.23054. [26] N.H. Zangenberg, A. Müllertz, H.G. Kristensen, L. Hovgaard, A dynamic in vitro lipolysis model: I. Controlling the rate of lipolysis by continuous addition of calcium, Eur. J. Pharm. Sci. 14 (2001) 115–122. [27] N.H. Zangenberg, A. Müllertz, H.G. Kristensen, L. Hovgaard, A dynamic in vitro lipolysis model: II: evaluation of the model, Eur. J. Pharm. Sci. 14 (2001) 237–244. [28] M.L. Lind, J. Jacobsen, R. Holm, A. Müllertz, Intestinal lymphatic transport of halofantrine in rats assessed using a chylomicron flow blocking approach: the influence of polysorbate 60 and 80, Eur. J. Pharm. Sci. 35 (2008) 211–218. [29] R. Holm, E.B. Jørgensen, M. Harborg, R. Larsen, P. Holm, A. Müllertz, J. Jacobsen, A novel excipient, 1-perfluorohexyloctane shows limited utility for the oral delivery of poorly water-soluble drugs, Eur. J. Pharm. Sci. 42 (2011) 416–422. [30] R. Holm, C. Porter, A. Müllertz, H. Kristensen, W. Charman, Structured triglyceride vehicles for oral delivery of halofantrine: examination of intestinal lymphatic transport and bioavailability in conscious rats, Pharm. Res. 19 (2002) 1354–1361.

[31] C.J.H. Porter, A.M. Kaukonen, Use of in vitro lipid digestion data to explain the in vivo performance of triglyceride-based oral lipid formulations of poorly watersoluble drugs: studies with halofantrine, J. Pharm. Sci. 93 (2004) 1110–1121. [32] R. Holm, A. Müllertz, G.P. Pedersen, H.G. Kristensen, Comparison of the lymphatic transport of halofantrine administered in disperse systems containing three different unsaturated fatty acids, Pharm. Res. 18 (2001) 1299–1304. [33] S.-M. Khoo, D.M. Shackleford, C.J.H. Porter, G.A. Edwards, Intestinal lymphatic transport of halofantrine occurs after oral administration of a unit-dose lipidbased formulation to fasted dogs, Pharm. Res. 20 (2003) 1460–1465. [34] G. Kossena, W. Charman, C. Wilson, B. O'Mahony, B. Lindsay, J. Hempenstall, C. Davison, P. Crowley, C. Porter, Low dose lipid formulations: effects on gastric emptying and biliary secretion, Pharm. Res. 24 (2007) 2084–2096. [35] A. Dahan, H. Amnon, Rationalizing the selection of oral lipid based drug delivery systems by an in vitro dynamic lipolysis model for improved oral bioavailability of poorly water soluble drugs, J. Control. Release 129 (2008) 1–10. [36] A.T. Larsen, P. Sassene, A. Müllertz, In vitro lipolysis models as a tool for the characterization of oral lipid and surfactant based drug delivery systems, Int. J. Pharm. 417 (2011) 245–255. [37] A.J. Humberstone, C.J.H. Porter, W.N. Charman, A physicochemical basis for the effect of food on the absolute oral bioavailability of halofantrine, J. Pharm. Sci. 85 (1996) 525–529. [38] A. Dahan, A. Hoffmann, Use of a dynamic in vitro lipolysis model to rationalize oral formulation development for poor water soluble drugs: correlation with in vivo data and the relationship to intra-enterocyte processes in rats, Pharm. Res. 23 (2006) 2165–2174. [39] A. Larsen, R. Holm, M. Pedersen, A. Müllertz, Lipid-based formulations for danazol containing a digestible surfactant, Labrafil M2125CS: in vivo bioavailability and dynamic in vitro lipolysis, Pharm. Res. 25 (2008) 2769–2777.