Development, in vitro and in vivo evaluation of a self-emulsifying drug delivery system (SEDDS) for oral enoxaparin administration

Accepted Manuscript Development, in vitro and in vivo evaluation of a self-emulsifying drug delivery system (SEDDS) for oral enoxaparin administration Ožbej Zupančič, Julia Anita Grießinger, Julia Rohrer, Irene Pereira de Sousa, Lukas Danninger, Alexandra Partenhauser, Nadine Flavia Elli Sündermann Laffleur, Andreas Bernkop-Schnürch PII: DOI: Reference:

S0939-6411(16)30600-2 http://dx.doi.org/10.1016/j.ejpb.2016.09.013 EJPB 12297

To appear in:

European Journal of Pharmaceutics and Biopharmaceutics

Received Date: Revised Date: Accepted Date:

26 November 2015 13 September 2016 23 September 2016

Please cite this article as: O. Zupančič, J. Anita Grießinger, J. Rohrer, I. Pereira de Sousa, L. Danninger, A. Partenhauser, N.F. Elli Sündermann Laffleur, A. Bernkop-Schnürch, Development, in vitro and in vivo evaluation of a self-emulsifying drug delivery system (SEDDS) for oral enoxaparin administration, European Journal of Pharmaceutics and Biopharmaceutics (2016), doi: http://dx.doi.org/10.1016/j.ejpb.2016.09.013

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Development, in vitro and in vivo evaluation of a self-emulsifying drug delivery system (SEDDS) for oral enoxaparin administration

Ožbej Zupančiča, Julia Anita Grießingerb, Julia Rohrera, Irene Pereira de Sousaa, Lukas Danningera, Alexandra Partenhausera, Nadine Elli Sündermannc Flavia Laffleura, Andreas Bernkop-Schnürcha*

a

Department of Pharmaceutical Technology, Institute of Pharmacy, Leopold-Franzens-

University Innsbruck, Innrain 80/82, 6020 Innsbruck, Austria / Europe b

ThioMatrix GmbH, Research Center Innsbruck, Trientlgasse 65, 6020 Innsbruck, Austria /

Europe c

Division of Developmental Immunology, Center for Biomodels and Experimental Medicine

(CBEM), Innsbruck Medical University, Innrain 80/82, 6020 Innsbruck, Austria *

Corresponding author:

Department of Pharmaceutical Technology Institute of Pharmacy Leopold-Franzens-University Innsbruck Innrain 80/82 6020 Innsbruck, Austria / Europe Tel.: +43-512- 507-58601 Fax: +43-512-507- 58699 E-Mail: [email protected] Key words: hydrophobic ion pairing, oral LMWH delivery, self-emulsifying drug delivery systems

1

Abstract Aim: It was the aim of this study to develop SEDDS for oral enoxaparin administration and evaluate it in vitro and in vivo. Methods: The emulsifying properties of SEDDS composed of long chain lipids (LCSEDDS), medium chain lipids (MC-SEDDS), short chain lipids (SC-SEDDS) and no lipids (NL-SEDDS) were evaluated. Thereafter, enoxaparin was incorporated via hydrophobic ion pairing in the chosen SEDDS, which were evaluated regarding their mucus permeating properties, stability towards pancreatic lipase, drug release profile and cytotoxicity. Finally, in vivo performance of SEDDS was evaluated. Results: The average droplet size of chosen LC-SEDDS, MC-SEDDS and NL-SEDDS ranged between 30 - 40 nm. MC-SEEDS containing 30% Captex 8000, 30% Capmul MCM, 30% Cremophor EL and 10% propylene glycol and NL-SEDDS containing 31.5% Labrafil 1944, 22.5% Capmul PG-8, 9% propylene glycol, 27% Cremophor EL and 10% DMSO exhibited 2-fold higher mucus diffusion than LC-SEDDS and were therefore chosen for further studies. The enoxaparin-dodecylamine complex (ENOX/DOA) was incorporated in a payload of 2% (w/v) into MC-SEDDS and NL-SEDDS. After 90 min 97% of MC-SEDDS and 5% of NL-SEDDS were degraded by pancreatic lipase. Both MC-SEDDS and NLSEDDS showed sustained in vitro enoxaparin release. Furthermore, orally administrated MCSEDDS and NL-SEDDS yielded an absolute enoxaparin bioavailability of 2.02% and 2.25%, respectively. Conclusion: According to the above mentioned findings, SEDDS could be considered as a potential oral LMWH delivery system.

2

1. Introduction Low molecular weight heparins (LMWHs) are among the most potent anticoagulants being represented by 55% of the US anticoagulant market [1]. They are administered intravenously or subcutaneously in prophylaxis of deep vein thrombosis and pulmonary embolism [1–4]. LMWHs are polyanionic glycosaminoglycans with a mean molecular weight of about 5 Da [5]. Due to their relatively high molecular weight, dense anionic net charge and hydrophilicity intestinal membrane permeability is poor [4]. Additionally, LMWHs are unstable in acidic conditions of the stomach [3,4]. Hence, no commercial oral LMWH formulation is available yet [4]. Development of oral LMWH formulations would be beneficial in offering higher patient compliance and reduced expenses associated with prolonged hospital stay and patient monitoring [3,4,6]. In order to sufficiently increase LMWH oral bioavailability, various formulations such as liposomes [7], microparticles [8], enteric coated delivery systems [9] and emulsions [10] have been evaluated. Among all approaches the use of absorption enhancers seems most promising [11]. Those include lipids and surfactants such as sodium caprate [3], bile acids [3], Labrasol [3],

Gelucire

44/14

[3],

fatty

acids

[11],

glyrthenic

acid

[4],

N-[8-(2-

hydroxybenzoyl)amino]caprylate [4], sodium N-[10-(2-hydroxybenzoyl)amino]decanoate [12], combination of phosphatidylcholine with monoacylglycerol [13] and deoxycholic acid [2,6]. Furthermore, self-emulsifying drug delivery systems (SEDDS) composed of such lipids, surfactants and co-solvents [14] might be a promising approach to enhance oral LMWH bioavailability. After contact with gastrointestinal fluids SEDDS spontaneously form transparent and kinetically stable emulsions under gentle agitation, which is provided in vivo by the motility of gastrointestinal tract [15]. Some SEDDS formulations also exhibit mucus permeating [16,17] and absorption enhancing properties [18]. However, SEDDS for oral LMWH delivery were to the best of our knowledge not developed yet.

3

Therefore, it was the aim of this study to incorporate a model LMWH enoxaparin into SEDDS and characterize the developed formulation in vitro and in vivo. Firstly, emulsification properties of different SEDDS formulations composed of long chain lipids (LC-SEDDS), medium chain lipids (MC-SEDDS), short chain lipids (SC-SEDDS) and SEDDS containing no lipids (NL-SEDDS) were evaluated. Afterwards, the chosen SEDDS containing enoxaparin were further characterized regarding mucus permeation, drug release profile, stability towards intestinal lipase degradation and cytotoxicity. Finally, the oral efficiency of SEDDS was evaluated in vivo.

2. Materials and methods 2.1 Materials Enoxaparin (Lovenox®, average MW 4500 Da, anti-Xa 10.000 IU/mL) was purchased from Sanofi-Aventis GmbH (Austria), Float-a-lyser® (MWCO = 1000 kDa) was purchased from Spectrumlabs (Netherlands), Capmul MCM EP (caprylic/capric mono- & diglycerides, HLB = 5.5), Capmul PG-8 EP/NF (propylene glycol monocaprylate, HLB = 6.7) and Captex 8000 (tricaprylin) were a gift from Abitec (USA), Peceol (glycerol monooleate, HLB = 3), Labrafil M 1944 CS (oleoyl macrogol-6 glycerides, HLB = 9), Labrasol (caprylocaproyl macrogol-8 glycerides, HLB = 12), Maisine 35-1 (glycerol monolinoleate, HLB = 1) and Transcutol HP (diethylene glycol monoethyl ether) were a gift from Gattefosse (France), Mygliol 840 (propylene glycol dicaprylate/dicaprate) was purchased from Sasol (Germany), Cremophor EL (polyethoxylated-35 castor oil, HLB = 13), propylene glycol, olive oil, sesame oil, triacetin, dodecylamine hydrochloride, cetrimonium bromide, benzalkonium chloride, Azure A hydrochloride, fluorescein diacetate (FDA), bile salts (1:1 mixture of sodium cholate and sodium deoxycholate), lipase from porcine pancreas (Type II, 54 units/mg using triacetin) were purchased from Sigma Aldrich (Germany). All other chemicals, reagents and solvents were of analytical grade and received from commercial sources. 4

2.2 Size exclusion HPLC (SEC-HPLC) for enoxaparin determination Enoxaparin was quantified by a slightly modified SEC-HPLC method as described previously [19]. A Hitachi EliteLaChrom HPLC-system equipped with L-2130 pump, L-2200 autosampler and L-2450 photodiode array UV detector was used. A BioSep-SEC-S 2000 column (300 x 7.80 mm, 5 µm) was used as stationary phase. The mobile phase consisted of 1 mg/mL L-arginine solution (pH was adjusted to 6.5 with HCl) at room temperature and flow rate of 1 mL/min. Injection volume was 90 µL and the detection wavelength 215 nm. 2.3 Preparation and characterisation of SEDDS formulations Lipids, surfactants and co-solvents (1 g) were homogenized by a vortex mixer followed by ultrasonication for 30 min. The semisolid components were melted before use. The resulting SEDDS preconcentrates (1% w/v) were emulsified in 5 mL of 50 mM phosphate buffer pH 6.8. Mean droplet size and polydispersity index of emulsions were determined by dynamic light scattering using a PSS Nicomp 380 DLS (Santa Barbara, CA, United States). Measurements were performed at 25 0C or 37 0C. The appearance of the emulsion was visually assessed by using the grading system [20]. In brief, `Grade A´ corresponds to a rapidly forming emulsion in less than 60 s, having a clear or slightly bluish appearance. `Grade B´ refers a rapidly forming emulsion appearing in a bluish color. `Grade C´ demonstrates a dark bluish to bright white (milky) emulsion that is formed within 120 s. `Grade D´ is a dull, greyish white emulsion that is slow to emulsify (longer than 180 s) and `Grade E´ demonstrates a formulation with poor or minimal emulsification exhibition with large oil droplets. After storing SEDDS formulations at room temperature for 24 h they were analyzed visually with respect to turbidity and phase separation [21]. Transmittance of 1% (w/v) SEDDS was determined spectrophotometrically (Uvi Light XT2, Secomam, Ales Cedex, France) at 560 5

nm using distilled water as reference [21,22]. Stability of anhydrous SEDDS preconcentrates and emulsions was evaluated as already described previously [17]. In brief, SEDDS preconcentrates were centrifuged (10 min, 15000 rpm). The preconcentrates which showed no phase separation were emulsified (1% w/v) in 50 mM phosphate buffer pH 6.8 and examined for kinetic stability via centrifugation (10 min, 6000 rpm). Additionally, their cloud point was determined as follows: 1% (w/v) SEDDS were emulsified in deionized water and placed on a water bath at 25 0C. The water bath temperature was slowly increased (2 0C/min) until a sudden appearance of turbidity in SEDDS was observed. Moreover, thermal stability of SEDDS was also measured by incubating them for 120 min at 37 °C under gentle agitation (300 rpm). Mean droplet size was measured at the beginning of the incubation and after 120 min.

2.4 In vitro digestion study For SEDDS in vitro digestion studies a slightly modified method as described previously was used [23]. The composition of digestion medium pH 6.5 was 2 mM Trizma maleate, 150 mM NaCl, 5 mM CaCl2 and 5 mM bile salts. SEEDS (1% w/v) were emulsified in the digestion medium in a thermostatically controlled vessel at 37 0C and pH was adjusted to 6.5 with either 0.1 M HCl or 0.1 M NaOH. Freshly prepared pancreatic lipase suspension (81 U/ mL) was added to the medium to initiate lipolysis. Throughout the study pH was kept constant at 6.5 by adding 0.1 M NaOH. The rate of SEDDS digestion was calculated from NaOH consumpiton.

2.5 SEDDS mucus diffusion study The diffusion of the SEDDS through freshly prepared mucus was investigated using the rotating tube technique as described previously [24]. In brief, defined sized silicon tubes were filled with freshly obtained porcine intestinal mucus and closed at one end with a silicone cap. 6

Fluorescein diacetate (FDA) labelled SEDDS were prepared as follows: 0.1 mg of FDA was dissolved in 100 mg SEDDS overnight under stirring at room temperature. The formulation was then centrifuged (10 min, 13.400 rpm) to remove any non-dissolved FDA. Additionally, mucus diffusion of pure FDA suspension (0.1 mg in 100 µL of 50 mM phosphate buffer pH 6.8) was also evaluated to confirm that the measured fluorescence represents SEDDS diffusion and not the released FDA molecules. Then, at the open end 50 µl of FDA labelled SEDDS (1% w/v) in 50 mM phosphate buffer pH 6.8 were added and the silicone tube was closed securely with a second cap. During the experiment all tubes were kept under horizontal rotating condition (50 rpm) in an incubator at 37 °C. After 240 min the tubes were frozen. To determine the depth of diffusion into the mucus layer, the frozen tubes were cut into slices of 2 mm in length, starting cutting at the end where the SEDDS were applied. The pieces were incubated in 400 µl of 5 M NaOH for 30 min under ultrasonication and another 30 min at 37 °C under gentle agitation (500 rpm) in order to quantitatively hydrolyze FDA to sodium fluorescein. The 100% value was prepared by directly adding 50 µl of FDA labelled SEDDS into 400 µl of 5 M NaOH. The sample fluorescence was measured in 96-well plates at an excitation wavelength of 480 nm and an emission wavelength of 520 nm on a micro-plate reader (Fluostar Galaxy, Austria). 2.6 Enoxaparin complex preparation and its incorporation in SEDDS To 100 µL of enoxaparin solution (2 mg/mL) 400 µL of benzalkonium chloride (0.25 – 5.0 mg/mL), of cetyltriammonium bromide (0.25 – 5.0 mg/mL) and of dodecylamine hydrochloride solution (0.25 – 5.0 mg/mL) was added dropwise at room temperature. The precipitated enoxaparin-surfactant complex was isolated by centrifugation (15 min, 18.000 rpm). The precipitate was washed twice with deionized water, lyophilized (Christ Gamma 116 LSC Freeze dryer) and stored at room temperature for further use. The amount of precipitated enoxaparin was determined by measuring the remaining amount of enoxaparin in 7

the supernatant by SEC-HPLC as described above. To measure if enoxaparin was quantitatively adsorbed on tube material, 500 µL of enoxaparin standard solutions (0.2 mg/mL, 0.1 mg/mL and 0.05 mg/mL) were added in the reaction tubes. The samples were centrifuged (15 min, 18.000 rpm) and the enoxaparin recovery rate (method accuracy) was determined by SEC-HPLC according to ICH Q2 (R1) recommendations. Each standard solution was injected in triplicate. ENOX/DOA was dissolved in formulation MC10 as follows: 20 mg of ENOX/DOA was dissolved in a mixture of 300 mg of Capmul MCM and 100 mg of propylene glycol, protected from light at 25 0C. After 72 h ENOX/DOA was completely dissolved and 300 mg of Captex 8000 and 300 mg of Cremophor EL were added. The formulation was homogenized by vortexing and stored at room temperature, protected from light. ENOX/DOA was dissolved in formulation NL9 as follows: 20 mg of ENOX/DOA was dissolved in 100 µL of DMSO overnight and added to 900 mg of formulation NL9. The formulation was homogenized by vortexing and stored at room temperature, protected from light.

2.7 In vitro release studies Release of enoxaparin from SEDDS was evaluated in 50 mM phosphate buffer pH 6.8 at 37 0

C by using a dialysis tube method as described previously [25]. Briefly, enoxaparin stock

solution (ENOX, 1 mg/mL) was prepared in 50 mM phosphate buffer pH 6.8. Dialysis tube (Float-a-lyser®, MWCO = 1000 kDa) was filled with 1 mL of stock solution, 1 mL of enoxaparin SEDDS emulsion (ENOX-SEDDS, prepared by emulsifying 20 mg of SEDDS in 1 mL of stock solution) and 1 mL of SEDDS emulsion containing ENOX/DOA complex (ENOX/DOA-SEDDS, prepared by emulsifing 20 mg of SEDDS containing a ENOX/DOA payload of 2% w/v in 1 mL of phosphate buffer). All samples were dialyzed against 15 mL of phosphate buffer at 370 C under constant stirring (300 rpm). At predetermined time points 8

aliquots of 50 µL were withdrawn from the medium and replaced with phosphate buffer. The amount of released heparin was determined by a slightly modified azure A assay [26]. To a 50 µL of sample 150 µL of aqueaous azure A solution (0.05 mg/mL) was added. The sample absorbance was measured after background subtraction at 530 nm. The total amount of enoxaparin in SEDDS samples (100% value) was determined by SEC-HPLC as described above.

2.8 In vivo studies In vivo studies were performed according to the Principles of Laboratory Animal Care and were approved by the Ethical Comitee of Austria (Vienna). Male Sprague-Dawley rats with a mean body weight of 250 - 300 g were obtained from Janvier Labs (Saint Berthevin, France). Rats were randomly divided in 6 groups (n = 6), where the first group served as positive control and recieved 100 µL of i.v. enoxaparin injection (0.2 mg/kg). All other groups were treated by oral administration via oral gavage and dosing volume of 5 mL/kg. The second group recieved aqueous enoxaparin solution (10 mg/kg), the third and forth group recieced placebo SEDDS fromulations (1g SEDDS/kg). Finally, the fifth and sixth group recieved SEDDS containing 2% (w/w) ENOX/DOA (10 mg enoxaparin/kg, 1g SEDDS/kg). The animals were fasted 12 h prior to oral administration and had free acess to water during the experiment. Blood samples of 200 µL were withdrawn from the tail vein and immediatly mixed with 25 µL of 3.8% sodium citrate to prevent clotting. Blood samples were centrifuged (20 min, 3000 rpm) and plasma was collected and stored at -20 0C until analysis. The amount of enoxaparin in samples was determined by Biophen® Heparin Anti-Xa kit using manual method. Area under curve (AUC0–360) was determined by the linear trapezoidal rule. Absolute bioavailability was calculated by comparing AUC0–360 for orally administered enoxaparin with that of intravenously administered enoxaparin using the following equation (1): 9

(1)

2.9 Resazurin assay The potential cytotoxic effect of 0.25% (w/v) and 0.5% (w/v) SEDDS emulsions in minimum essential medium (MEM) was determined by a slightly modified method as described previously [27]. Approximately 2.5 x 104 Caco-2 cells per well were seeded to a 24-well plate. The cells were incubated for 21 days in minimum essential medium supplemented with 10% (v/v) heat inactivated fetal calf serum (FCS) and penicillin/streptomycin solution (100 units/ 0.1 mg/L) at 95% humidity and 37 0C in an 5% CO2 atmosphere. Pure MEM and 1% (w/v) Triton X 100 served as negative and positive controls, respectively. After incubating for 1 h and 4 h, samples (500 µL) were removed from the cells and washed with isotonic phosphate buffered saline (250 µL). Subsequently, 250 µL of resazurin solution (44 µM) in FCS- and penicillin/streptomycin- free MEM was added to each well and incubated for 3 h. The fluorescence of the supernatant was measured after background subtraction using an excitation wavelength of 540 nm and an emission wavelength of 590 nm. Cell viability rates of the samples were calculated according to Equation (2): (2)

where As is the fluorescence of samples and Ac is the fluorescence measured after treatment of cells with MEM

2.10 Statistical data analysis Statistical data analysis was performed on IBM SPSS Statistics 21 with p < 0.05 as the minimal level of significance. The analysis of variance (ANOVA) was used to compare different groups. All values are expressed as means ± SD. 10

3. Results and discussion 3.1 Preparation and characterisation of SEDDS formulations The selection criteria of excipients for SEDDS development was based on the excipients classification on non-polar lipids, polar class I lipids, polar class II lipids, hydrophobic surfactants (HLB < 12), hydrophilic surfactants (HLB > 12) and co-solvents [28]. To evaluate the impact of carbon chain length on SEDDS characteristics, four different types of SEDDS were developed: SEDDS containing long chain lipids (LC-SEDDS, composed of lipids with > C10 carbon chain length), SEDDS containing medium chain lipids (MC-SEDDS, composed of lipids with C8 - C10 carbon chain length), SEDDS containing short chain lipids (SC-SEDDS, composed of lipids with < C8 carbon chain length) and SEDDS containing no lipids (NLSEDDS, containing only surfactants and co-solvents) (Tables 1 - 4). LC-SEDDS containing long chain lipids, namely Maisine 35-1 tend to be instable in combination with Cremophor EL (Table 1). Additionally, a phase separation of LC2, LC4 and LC9 preconcentrates was observed after centrifugation. This indicates poor miscibility of Maisine 35-1 with Peceol and Labrafil 1944 which are both derivatives of oleic acid. Also, corn oil seems not to be compatible with Capmul MCM. LC-SEDDS composed of more than 20% (w/w) long chain lipids formed dense milky and unstable emulsions (LC1-LC9). However, when Peceol was used in amounts below 20%, stable emulsions (LC10-LC18) were formed. Moreover, SCSEDDS failed to form droplet sizes below 165 nm and were therefore excluded from further studies (Table 2). In contrast, MC-SEDDS and NL-SEDDS were shown to possess more favourable emulsification characteristics (Tables 3 and 4). In most cases, the emulsions formed were clear. A combination of Cremophor EL with Capmul MCM [15,25] and Labrafil 1944 [16] has already been shown to form SEDDS with average droplet size below 50 nm. Additionally, Cremophor EL is also known to be biocompatible and is therefore very often 11

used in SEDDS formulations [20]. Interestingly, a monodisperse system with polydispersity index (PDI) of 0.04 was observed only by formulation MC10 containing triglyceride Captex 8000. Additionally, MC7 containing Mygliol 840 also had a relatively narrow size distribution with a PDI value of 0.22. All other formulations had PDI values between 0.31 – 0.52, which is, according to the literature, not considered optimal in respect of emulsion stability and delivery system efficiency. The suggested acceptable PDI values are up to the range 0.250-0.300 [29–33]. The high PDI of MC-SEDDS seems to correlate with higher amount (> 60%) of relatively hydrophilic components like Capmul MCM, Cremophor EL and Labrafil 1944 and lower (< 30%) amount of highly lipophilic oils with HLB values < 1 like Captex 8000 and Mygliol 840. Furthermore, the chosen formulations from each group with similar droplet sizes – LC10, MC10 and NL9 were emulsified in phosphate buffer and their cloud point, which is defined as the temperature at which there is a sudden appearance of cloudiness, was measured both visually and by determining the emulsions UV transmittance [25,34]. The measured cloud points for LC10, MC10 and NL9 were 47 0C, 52 0C and 48 0C, respectively. These findings indicate the formation of stable emulsions in vivo without the risk of phase separation and uncontrolled drug release from SEDDS [34]. Moreover, after incubating for 2 h at 37 0C the droplet sizes of LC10, MC10 and NL9 did not significantly change (p < 0.05) from initial 58.7 ± 37.7, 38.7 ± 6.7 and 41.5 ± 9.9 nm to 60.2 ± 37.7, 38.2 ± 6.2 and 44.7 ± 12.6 nm, respectively. Moreover, NL9 could be considered as a Type IV lipid formulation system, which contains no oils and is composed of surfactant and co-solvent mixtures. Although such compositions do not meet classical emulsion definition, were they added to Lipid Formulation Classification System [28,35]. Accordingly, the water insoluble surfactants in NL9 like Labrafil 1944 and Capmul PG-8 were unable to dissolve in water or form micelles and could therefore be considered as “oils”. Henceforth, they were emulsified in the aqueous phase by Cremophor EL, presumably forming similar colloidal structures as

12

MC10 on the basis of their similar appearance, droplet sizes, PDI and emulsification properties. 3.2 Mucus diffusion study Mucus layer is a hydrogel composed of glycoproteins posing a net negative charge, mucin being its main component. In order to reach the epithelium, SEDDS containing the drug must overcome the mucus barrier to ensure the sufficient drug levels in the systemic circulation [36]. Due to SEDDS lipophilic nature, their interaction with mucus layer is presumably low. Moreover, many SEDDS components like oleic acid, monoglycerides of caprylic acid, propylene glycol esters of caprylic acid, Tween 80, Labrasol, PEG 400 and Transcutol [36] as well as triacetin and Cremophor RH40 [17] were shown to readily diffuse through mucus. Because particle size influences mucus diffusion speed [37] only SEDDS with similar size distribution from each group were compared. Mucus permeation of the chosen FDA labeled SEDDS formulations LC10, MC10 and NC9 is presented in Figure 1. In the 0-2 mm segment similar amount diffused SEDDS was measured by formulations LC10 and NL9, whereas 20% of FDA was diffused by MC10. Formulation LC10 barely diffused to segment 6-8 mm, whereas an almost 2- fold increase in diffusion of formulations MC10 and NL9 was measured. With each next segment, an about 2-fold decrease of diffused FDA was observed by all formulations. However, a significant difference (p < 0.05) in diffusion of LC10 and both MC10 and NL9 formulations indicated that carbon chain length might have an impact on SEDDS mucus diffusion. Furthermore, addition of 10% DMSO to formulation NL9 had no significant impact (p < 0.05) on NL9 in vitro mucus permeability (p < 0.05) (data not shown). Noteworthy, in a preliminary study it was shown that pure FDA suspension did not diffuse through mucus. In addition, FDA incorporation in SEDDS hat no negative impact on their droplet size and stability. Hence, it could be assumed that the lipophilic FDA remained incorporated in SEDDS throughout the experiment [17]. 13

3.3 Enoxaparin complex preparation and its incorporation in SEDDS Enoxaparin is a hydrophilic molecule with a dense anionic charge. In order to dissolve LMWH in lipids, its lipophilicity had to be altered by masking the charge. In previous studies, a LMWH derivate using covalent attachment of deoxycholic acid [6] and noncovalent ion pairing of LMWH with deoxycholyl ethylenediamine [12] have shown to increase intestinal LMWH absorption. In this study, a straightforward ion pairing method using commercially available cationic surfactants resulted in precipitation of enoxaparin-surfactant complex (Figure 2). Firstly, the recovery of enoxaparin standard solutions was 99.23%, indicating that enoxaparin was not quantitatively adsorbed on the tube material. Dodecylamine hydrochloride most efficiently precipitated enoxaparin form aqueous solution, yielding a flakey-like precipitate. Precipitation of enoxaparin at surfactant concentrations above 1.0 mg/mL was practically quantitative. Increased concentration up to 5.0 mg/mL did not reduce the overall enoxaparin precipitation yield, which remained around 100%. Dodecylamine hydrochloride concentration of 1 mg/mL was the lowest concentration to fully precipitate enoxaparin and was chosen for further studies. Higher concentrations of surfactant may reduce the amount of enoxaparin in the complex. In this study, SEC-HPLC showed that ENOX/DOA complex contained 56.3 ± 2.4% of enoxaparin. In addition, the droplet sizes of MC10 and NL9 did not significantly change after ENOX/DOA incorporation (p < 0.05).

3.4 In vitro release studies Hydrophilic drugs incorporated in SEDDS have the tendency to rapidly migrate in the aqueous phase after emulsification [25]. In order to evaluate if enoxaparin remained in the oily droplets after emulsification of SEDDS with phosphate buffer a release study using a 14

dialysis membrane was performed. Figure 3 presents the release profiles ENOX, ENOXSEDDS and ENOX/DOA-SEDDS. ENOX served as control to measure the impact of the membrane on drug release. No significant difference (p < 0.05) in the release profiles of ENOX and ENOX-SEDDS was observed indicating that by only emulsifying placebo SEDDS in aqueous enoxaparin solution does not result in its incorporation in the oily phase. In contrast, significant difference (p < 0.05) in release profiles between aforementioned samples and ENOX/DOA-SEDDS was observed. Both formulations MC10 and NL9 containing 2% w/v ENOX/DOA showed sustained drug release. After 6 h, 45% of enoxaparin was released by formulation NL9 and 33% by formulation MC10. The absence of burst release is beneficial to ensure contact of enoxaparin with lipid and amphiphilic SEDDS components. Sustained release also indicates that hydrophobic ion pairing led to efficient incorporation of enoxaparin into SEDDS. Furthermore, enoxaparin is a hydrophilic molecule with a strong anionic net charge density. Additionally, the molecule has numerous anionic centers. Although hydrophobic ion pairing is a reversible process, both of those factors might contribute to stronger enoxaparin-surfactant ionic interactions and thus prolonged lipophilic character of ENOX/DOA, sustaining its release from the lipophilic SEDDS core. 3.5 In vitro digestion study When administered orally, SEDDS are prone to digestion in duodenum by pancreatic lipase. It hydrolyses triglycerides in positions 1 and 3, resulting in two free fatty acids and one 2monoacylglyceride [28]. The rate of SEDDS digestion by pancreatic lipase can be estimated by measuring the consumption of NaOH needed to neutralize the released fatty acids during lipolysis. Figure 4 illustrates the consumption of 0.1 M NaOH during lipolysis of 1% (w/v) formulations NL9 and MC10. In 90 min practically no NaOH was consumed by formulation NL9 indicating its lipolytic stability. The main oil components of NL9 are Labrafil 1944 and Capmul PG-8, both lacking classical triglyceride structure. Hence, they are apparently poor 15

lipase substrates. In contrast, MC10 was almost completely digested within 90 min, where 4.3 mL of NaOH was consumed. MC10 is composed of tricaprylate Captex 8000 and a mixture of mono-, di- and triglycerides Capmul MCM, which have already been shown to be lipase substrates [25]. In addition, although the normal small intestinal passage time is considered to be about 4 h was the lipolysis rate measured only for 90 min. Pancreatic lipase follows Michaelis–Menten kinetics and the vast majority of SEDDS would be already be digested in this timeframe. Besides, according to mucus diffusion studies it could be assumed that within 90 min a fair amount of SEDDS would already diffuse into mucus layer, where pancreatic lipase has no effect. 3.6 In vivo studies The objective of in vivo study was demonstrate that orally administered enoxaparin dissolved in the oily phase of SEDDS was able to reach the systemic circulation. Two SEDDS formulations containing 2% (w/v) of enoxaparin-dodecylamine complex, namely NL9 and MC10, were compared to aqueous enoxaparin solution. Blank NL9 and MC10 served as negative controls to show that its components had no impact on plasma anti-Xa activity. In contrast, intravenous enoxaparin served as positive control. The results in Table 5 show the main pharmacokinetic parameters of intravenous enoxaparin, NL9 and MC10. Moreover, no anti-Xa activity was detected by aqueous enoxaparin solution, which is in good accordance with current knowledge. In contrast, as illustrated in Figure 5 both SEDDS containing enoxaparin showed an about 2-fold increase in anti-Xa activity at least 6 h after oral administration compared to aqueous enoxaparin solution, resulting in absolute bioavailability of 2.25% and 2.02%, respectively. This indicates that sufficient amount of enoxaparin in SEDDS was able to cross the mucus and the absorption barrier in gastrointestinal tract. Interestingly, there was no significant difference in pharmacokinetic profiles between MC10 and NL9, assuming that both SEDDS formulations sufficiently solubilized enoxaparin. 16

3.7 Resazurin assay To investigate the potential cytotoxic effects of chosen SEDDS, cell viability of a Caco-2 monolayer was determined. Caco-2 cells were incubated with SEDDS formulations MC10, NL9, MC10 containing 2% (w/w) ENOX/DOA and NL9 containing 2% (w/w) ENOX/DOA. Table 6 illustrates the toxicities of the tested formulations after 4h of incubation and concentrations up to 0.5% (w/v). It was shown that both NL9 and MC10 placebo SEDDS had no negative impact on Caco-2 cells viability. However, there was a significant reduction in cell viability in both SEDDS containing ENOX/DOA. Indeed, this may be due to known toxic effects of dodecylamine and related cationic surfactants [38]. However, it should be noticed that the Caco-2 cytotoxicity model may not completely mimic the actual in vivo conditions due to several factors. To begin with, Caco-2 cells generally show higher viability rate if the initial Caco-2 seeding density is higher and the culture duration of 21 days [39]. In addition, according to the literature data, the volume of liquid in small intestine ranges from 30 - 420 mL in fasted state and from 18 - 660 mL in the fed state [40]. Hence, the maximum SEDDS concentration of 0.5% (w/v) was chosen as the maximum value as it is assumed that under in vivo conditions SEDDS would be diluted to tolerable concentrations. Moreover, Caco-2 model also lacks a protective mucus layer and the ability of the monolayer to recover from trauma over time, which is present under in vivo conditions [4].

17

4. Conclusion In this study the proof of concept that LMWH could be orally delivered via SEDDS was demonstrated. SEDDS composed of medium chain lipids (MC-SEDDS) and no lipids (NLSEDDS) with the average droplet size of about 30 nm exhibited enhanced mucus permeating properties. The lipophilicity of enoxaparin was sufficiently increased by hydrophobic ion pairing with dodecylamine hydrochloride. The incorporated ENOX/DOA complex (2% w/w) in both SEDDS formulations showed sustained enoxaparin release. Moreover, NL-SEDDS were stable towards pancreatic lipase degradation, whereas MC-SEDDS were degraded within 90 min. In conclusion, the in vivo study showed 2-fold increased enoxaparin bioavailability when delivered via SEDDS, indicating that the full potential of SEDDS has not been reached yet. 5. References [1] Dickinson DM, Liu J, Linhardt RJ. Chemoenzymatic Synthesis of Heparins. In: Endo T, editor. Glycoscience: Biology and Medicine. Tokyo: Springer; 2015. p. 419–426. [2] Park JW, Jeon OC, Kim SK, Al-Hilal TA, Moon HT, Kim CY, Byun Y. Anticoagulant efficacy of solid oral formulations containing a new heparin derivative. Molecular Pharmaceutics 2010;7:836– 43. [3] Mori S, Matsuura A, Rama Prasad, Yarasani Venkata, Takada K. Studies on the Intestinal Absorption of Low Molecular Weight Heparin Using Saturated Fatty Acids and Their Derivatives as an Absorption Enhancer in Rats. Biol. Pharm. Bull. 2004;27:418–21. [4] Motlekar NA, Srivenugopal KS, Wachtel MS, Youan BC. Evaluation of the Oral Bioavailability of Low Molecular Weight Heparin Formulated With Glycyrrhetinic Acid as Permeation Enhancer. Drug Development Research 2006;67:166–74. [5] Radivojša Matanović M, Grabnar I, Gosenca M, Grabnar PA. Prolonged subcutaneous delivery of low molecular weight heparin based on thermoresponsive hydrogels with chitosan nanocomplexes: Design, in vitro evaluation, and cytotoxicity studies. International Journal of Pharmaceutics 2015;488:127–35.

18

[6] Kim SK, Lee DY, Lee E, Lee Y, Kim CY, Moon HT, Byun Y. Absorption study of deoxycholic acid-heparin conjugate as a new form of oral anti-coagulant. Journal of Controlled Release Official Journal of the Controlled Release Society 2007;120:4–10. [7] Ueno M, Nakasaki T, Horikoshi I, Sakuragawa N. Oral administration of liposomally-entrapped heparin to beagle dogs. Chem. Pharm. Bull. 1982;30:2245–7. [8] Thanou M, Verhoef JC, Nihot M, Verheijden JHM, Junginger HE. Enhancement of the intestinal absorption of low molecular weight heparin (LMWH) in rats and pigs using Carbopol® 934P. Pharmaceutical Research 2001;18:1638–41. [9] Jiao YY, Ubrich N, Hoffart V, Marchand-Arvier M, Vigneron C, Hoffman M, Maincent P. Preparation and characterization of heparin-loaded polymeric microparticles. Drug Development and Industrial Pharmacy 2002;28:1033–41. [10] Engel RH, Riggi SJ. Intestinal absorption of heparin: A study of the interactions of components of oil‐ in‐ water emulsions. Journal of Pharmaceutical Sciences 1969;58:1372–5. [11] Ross BP, Toth I. Gastrointestinal absorption of heparin by lipidization or coadministration with penetration enhancers. Current Drug Delivery 2005;2:277–87. [12] Lee DY, Lee J, Lee S, Kim SK, Byun Y. Liphophilic complexation of heparin based on bile acid for oral delivery. Journal of Controlled Release Official Journal of the Controlled Release Society 2007;123:39–45. [13] Lohikangas L, Wilen M, Einarsson M, Artursson P. Effects of a new lipid-based drug delivery system on the absorption of low molecular weight heparin (Fragmin) through monolayers of human intestinal epithelial Caco-2 cells and after rectal administration to rabbits. European Journal of Pharmaceutical Sciences 1994;1:297–305. [14] Pouton CW, Porter, Christopher J H. Formulation of lipid-based delivery systems for oral administration: materials, methods and strategies. Advanced Drug Delivery Reviews 2008;60:625– 37. [15] Hintzen F, Perera G, Hauptstein S, Müller C, Laffleur F, Bernkop-Schnürch A. In vivo evaluation of an oral self-microemulsifying drug delivery system (SMEDDS) for leuprorelin. International Journal of Pharmaceutics 2014;472:20–6. [16] Karamanidou T, Karidi K, Bourganis V, Kontonikola K, Kammona O, Kiparissides C. Effective incorporation of insulin in mucus permeating self-nanoemulsifying drug delivery systems. European Journal of Pharmaceutics and Biopharmaceutics 97 (2015): 223-229. [17] Friedl H, Dünnhaupt S, Hintzen F, Waldner C, Parikh S, Pearson JP, Wilcox MD et al. Development and evaluation of a novel mucus diffusion test system approved by selfnanoemulsifying drug delivery systems. Journal of Pharmaceutical Sciences 2013;102:4406–13. [18] Hintzen F, Laffleur F, Sarti F, Müller C, Bernkop-Schnürch A. In vitro and ex vivo evaluation of an intestinal permeation enhancing self-microemulsifying drug delivery system (SMEDDS). Journal of Drug Delivery Science and Technology 2013;23:261–7. [19] Matanović MR, Grabnar I, Grabnar PA, Roškar R. Development and validation of a simple and sensitive size-exclusion chromatography method for quantitative determination of heparin in pharmaceuticals. Acta Pharmaceutica (Zagreb, Croatia) 2015;65:43–52.

19

[20] Shen H, Zhong M. Preparation and evaluation of self-microemulsifying drug delivery systems (SMEDDS) containing atorvastatin. The Journal of Pharmacy and Pharmacology 2006;58:1183– 91. [21] Shahnaz G, Hartl M, Barthelmes J, Leithner K, Sarti F, Hintzen F, Rahmat D et al. Uptake of phenothiazines by the harvested chylomicrons ex vivo model: Influence of self-nanoemulsifying formulation design. European Journal of Pharmaceutics and Biopharmaceutics 2011;79:171–80. [22] Singh AK, Chaurasiya A, Awasthi A, Mishra G, Asati D, Khar RK, Mukherjee R. Oral bioavailability enhancement of exemestane from self-microemulsifying drug delivery system (SMEDDS). AAPS PharmSciTech 2009;10:906–16. [23] Thomas N, Müllertz A, Graf A, Rades T. Influence of lipid composition and drug load on the In Vitro performance of self-nanoemulsifying drug delivery systems. Journal of Pharmaceutical Sciences 2012;101:1721–31. [24] Grießinger J, Dünnhaupt S, Cattoz B, Griffiths P, Oh S, Gómez, Salvador Borrós I, Wilcox M et al. Methods to determine the interactions of micro- and nanoparticles with mucus. European Journal of Pharmaceutics and Biopharmaceutics 96 (2015): 464-476 [25] Zupančič O, Partenhauser A, Lam HT, Rohrer J, Bernkop-Schnürch A. Development and in vitro characterisation of an oral self-emulsifying delivery system for daptomycin. European Journal of Pharmaceutical Sciences 2016;81:129–36. [26] Radivojša M, Grabnar I, Ahlin Grabnar P. Thermoreversible in situ gelling poloxamer-based systems with chitosan nanocomplexes for prolonged subcutaneous delivery of heparin: design and in vitro evaluation. European Journal of Pharmaceutical Sciences 2013;50:93–101. [27] Partenhauser A, Laffleur F, Rohrer J, Bernkop-Schnürch A. Thiolated silicone oil: synthesis, gelling and mucoadhesive properties. Acta Biomaterialia 2015;16:169–77. [28] Müllertz A, Ogbonna A, Ren S, Rades T. New perspectives on lipid and surfactant based drug delivery systems for oral delivery of poorly soluble drugs. Journal of Pharmacy and Pharmacology 2010;62:1622–36. [29] Morales D, Gutiérrez JM, Garcia-Celma MJ, Solans YC. A study of the relation between bicontinuous microemulsions and oil/water nano-emulsion formation. Langmuir 2003;19:7196– 200. [30] Bernardi DS, Pereira TA, Maciel NR, Bortoloto J, Viera GS, Oliveira GC, Rocha-Filho PA. Formation and stability of oil-in-water nanoemulsions containing rice bran oil: in vitro and in vivo assessments. Journal of Nanobiotechnology 2011;9:1. [31] Kumar M, Misra A, Babbar AK, Mishra AK, Mishra P, Pathak K. Intranasal nanoemulsion based brain targeting drug delivery system of risperidone. International Journal of Pharmaceutics 2008;358:285–91. [32] Müller RH, Schmidt S, Buttle I, Akkar A, Schmitt J, Brömer S. SolEmuls®—novel technology for the formulation of iv emulsions with poorly soluble drugs. International Journal of Pharmaceutics 2004;269:293–302. [33] Tadros T, Izquierdo P, Esquena J, Solans C. Formation and stability of nano-emulsions. Advances in colloid and Interface Science 2004;108:303–18. [34] Gupta S, Chavhan S, Sawant KK. Self-nanoemulsifying drug delivery system for adefovir dipivoxil: design, characterization, in vitro and ex vivo evaluation. Colloids and Surfaces A: Physicochemical and Engineering Aspects 2011;392:145–55. 20

[35] Pouton CW, Porter CJH. Formulation of lipid-based delivery systems for oral administration: materials, methods and strategies. Advanced Drug Delivery Reviews 2008;60:625–37. [36] Laffleur F, Bernkop-Schnürch A. Strategies for improving mucosal drug delivery. Nanomedicine (London, England) 2013;8:2061–75. [37] Cone RA. Barrier properties of mucus. Advanced drug delivery reviews 2009;61:75–85. [38] Nałęcz-Jawecki G, Grabińska-Sota E, Narkiewicz P. The toxicity of cationic surfactants in four bioassays. Ecotoxicology and Environmental safety 2003;54:87–91. [39] Bu P, Narayanan S, Dalrymple D, Cheng X, Serajuddin ATM. Cytotoxicity assessment of lipid-based self-emulsifying drug delivery system with Caco-2 cell model: Cremophor EL as the surfactant. European Journal of Pharmaceutical Sciences 2016;91:162–71. [40] Mudie DM, Amidon GL, Amidon GE. Physiological parameters for oral delivery and in vitro testing. Molecular Pharmaceutics 2010;7:1388–405.

21

Figure legend: Figure 1

Figure 2

Figure 3

Figure 4 Figure 5

Mucus diffusion of indicated SEDDS formulations: LC10 (gray), MC10 (black) and NL9 (white) via the rotating tube method. Data are shown as mean ± SD (n =3). Precipitation of enoxaparin from aqaeous solution (2 mg/ml) with dodecylamine chloride [○], cetyltetramethylammonium bromide [□] and benzalkonium chloride [◊]. The precipitated enoxaprain-surfactant complex was isolated by centrifugation and the remaining enoxaparin anaysed by SEC-HPLC. Data are shown as mean ± SD (n =3). In vitro release of enoxaparin in 50 mM phosphate buffer pH 6.8 from enoxaparin aqueous solution [■], enoxaparin aqueous solution containing 2% (w/V) formulation NL9 [♦], enoxaparin aqueous solution containing 2% (w/V) formulation MC10 [◊], 2% (w/V) formulation NL9 containing 2% (w/w) ENOXDOA complex [●] and 2% (w/V) formulation MC10 containing 2% (w/w) ENOX-DOA complex [○] via dialysis bag method. Data are shown as mean ± SD (n =3). Consumption of 0.1 M NaOH by in vitro pancreatic lipase digestion of SEDDS formulations MC10 [■] and NL9 [●].Data are shown as mean ± SD (n =3). Oral absorbtion profiles of SEDDS formulations NL9 (□) and MC10 (■) containing 2% (w/w) ENOX/DOA. Data are shown as mean ± SD (n = 6).

Table legend: Table 1 Table 2 Table 3 Table 4 Table 5 Table 6

Development and characterization of SEDDS formulations containing long chain lipids (LC-SEDDS). Data are shown as mean ± SD (n =3). Development and characterization of SEDDS formulations containing short chain lipids (SC-SEDDS). Data are shown as mean ± SD (n=3). Development and characterization of SEDDS formulations containing medium chain lipids (MC-SEDDS). Data are shown as mean ± SD (n = 3). Development and characterization of SEDDS formulations containing no lipids (NL-SEDDS). Data are shown as mean ± SD (n = 3). Main in vivo study pharmacokinetic parameters (n = 6). Cell viability rates of Caco-2 cells after incubation for 1 h and 4 h with 0.25% (w/v) and 0.5% (w/v) placebo SEDDS and SEDDS containing 2% (w/v) ENOX/DOA. Data are shown as mean ± SD (n = 3).

22

diffused FDA [%]

20 15 10 5 0 2

4

6

8

10

12

14

distance [mm]

Figure 1: Mucus diffusion of indicated SEDDS formulations: LC10 (gray), MC10 (black) and NL9 (white) via the rotating tube method. Data are shown as mean ± SD (n =3).

precipitated enoxaparin [%]

100 80 60 40 20 0 0.0

0.5

1.0

1.5

2.0

surfactant concentration [mg/mL]

Figure 2: Precipitation of enoxaparin from aqaeous solution (2 mg/ml) with dodecylamine chloride [○], cetyltetramethylammonium bromide [□] and benzalkonium chloride [◊]. The precipitated enoxaprain-surfactant complex was isolated by centrifugation and the remaining enoxaparin anaysed by SEC-HPLC. Data are shown as mean ± SD (n =3).

23

released enoxaparin [%]

100

80

60

40

20

0 0

1

2

3

4

5

6

time [h] Figure 3: In vitro release of enoxaparin in 50 mM phosphate buffer pH 6.8 from enoxaparin aqueous solution [■], enoxaparin aqueous solution containing 2% (w/V) formulation NL9 [♦], enoxaparin aqueous solution containing 2% (w/V) formulation MC10 [◊], 2% (w/V) formulation NL9 containing 2% (w/w) ENOX-DOA complex [●] and 2% (w/V) formulation MC10 containing 2% (w/w) ENOX-DOA complex [○] via dialysis bag method. Data are shown as mean ± SD (n =3).

consumption NaOH [mL]

5 4 3 2 1 0 0

30

60

90

time [min]

24

enoxaparin concentration [IU/mL]

Figure 4: Consumption of 0.1 M NaOH by in vitro pancreatic lipase digestion of SEDDS formulations MC10 [■] and NL9 [●].Data are shown as mean ± SD (n =3).

0.3

0.2

0.1

0 0

2

4

6

time [h]

Figure 5: Oral absorbtion profiles of SEDDS formulations NL9 (□) and MC10 (■) containing 2% (w/w) ENOX/DOA. Data are shown as mean ± SD (n = 6).

25

Table 1: development and characterisation of SEDDS formulations containing long chain lipids (LC-SEDDS).Data are shown as mean ± SD (n =3). surfactants HLB < 12 Labrafil Capmul 1944 PG 8

surfactants co-solvent HLB > 12 emulsification transmittance time [s] [%] Sesam Corn Capmul Cremophor Peceol Maisine 35-1 Labrasol PG Transcutol oil oil MCM EL LC1 20 40 30 10 LC2 10 50 30 10 LC3 20 40 30 10 LC4 40 20 30 10 LC5 20 40 30 10 LC6 20 30 40 10 LC7 40 30 20 10 LC8 30 30 30 10 LC9 30 30 30 10 LC10 10 40 40 10 87 95.7 LC11 10 40 40 10 211 17.3 LC12 20 20 50 10 190 2.7 LC13 10 40 40 10 64 18.4 LC14 20 30 40 10 41 79.5 LC15 12.5 37.5 37.5 12.5 29 4.3 LC16 10 30 50 10 64 40.0 LC17 10 20 60 10 87 50.3 LC18 10 40 40 10 114 87.8 X – the formulation could not be characterized due to phase separation or no emulsification X* – phase separation of SEDDS preconcentrates after centrifugation lipid phase

No

appearance X X* X X* X X X X X* clear bluish dark bluish dark bluish bluish dark bluish bluish bluish clear

mean droplet size [nm]

polydispersity index [P.I.]

30.5 ± 0.3 216.5 ± 4.0 184.3 ± 8.2 182.6 ± 2.0 97.0 ± 1.2 242.5 ± 10.6 203.0 ± 4.7 213.1 ± 4.2 32.5 ± 0.5

0.13 0.37 0.35 0.28 0.36 0.14 0.29 0.27 0.10

Table 2: development and characterisation of SEDDS formulations containing short chain lipids (SC-SEDDS).Data are shown as mean ± SD (n=3). lipid phase No Triacetin SC1 SC2 SC3 SC4 SC5

30 20 20 40 30

Capmul MCM

Maisine 35-1 20

40 30 20 30

surfactants HLB > 12 Cremophor EL 40 30 40 30 30

co-solvent PG 10 10 10 10 10

emulsification time [s]

transmittance [%]

appearance

mean droplet size [nm]

polydispersity Index

44 21 105 119 23

41.0 26.1 58.9 22.2 42.6

bluish dark bluish slightly bluish dark bluish dark bluish

165.5 ± 6.4 265.3 ± 12.0 236.6 ± 19.7 173.6 ± 5.1 242 ± 9.1

0.38 0.31 0.53 0.34 0.36

27

Table 3: development and characterisation of SEDDS formulations containing medium chain lipids (MC-SEDDS).Data are shown as mean ± SD (n = 3).

No NL1 NL2 NL3 NL4 NL5 NL6 NL7 NL8 NL9

surfactants HLB < 12 Labrafil Capmul PG 8 1944 30 40 30 40 25 25 35 25 40 20 45 25 35 25

surfactants HLB > 12 Cremophor Labrasol EL 40 20 40 10 40 20 30 20 40 30 30 20 30

co-solvent PG

Transcutol

emulsification time [s]

transmittance [%]

appearance

mean droplet size [nm]

Polydispersity Index [P.I.]

45 125 22 21 31 20 22 19 28

99.4 98.8 99.3 98.8 95.7 93.7 93.5 70.6 87.7

clear clear clear clear clear slightly bluish slightly bluish slightly bluish slightly bluish

12.6 ± 1.3 25.0 ± 1.0 25.2 ± 0.9 25.8 ± 0.3 37.3 ± 0.1 39.2 ± 0.2 39.8 ± 0.6 68.2 ± 0.2 39.6 ± 0.6

1.03 0.07 0.05 0.06 0.31 0.05 0.03 0.08 0.03

10 10 10 10 10 10 10 10 10

Table 4: development and characterisation of SEDDS formulations containing no lipids (NL-SEDDS).Data are shown as mean ± SD (n = 3). lipid phase No. MC1 MC2 MC3 MC4 MC5 MC6 MC7 MC8 MC9 MC10

Captex 8000 20 10

Miglyol 840

20 10 20 20 30

30

Capmul MCM 40 50 40 50 30 30 30 20 30 30

surfactant HLB < 12 Labrafil 1944

30 30

surfactants HLB > 12 Cremophor EL 30 30 30 30 40 40 30 40 30 30

co-solvent PG 10 10 10 10 10 10 10 10 10 10

emulsification time [s]

transmittance [%]

appearance

mean droplet size [nm]

Polydispersity Index [P.I.]

24 25 28 20 47 110 78 268 234 20

83.7 17.6 66.0 1.0 92.6 64.9 85.5 88.7 74.4 98.6

clear dark bluish slightly bluish dark bluish clear clear slightly bluish clear* clear* slightly bluish

66.1 ± 2.1 194.9 ± 9.9 142.5 ± 1.4 229.0 ± 9.4 36.1 ± 0.4 121.0 ± 7.6 43.7 ± 0.6 73.5 ± 1.4 129.3 ± 1.4 33.3 ± 0.8

0.45 0.38 0.30 0.39 0.31 0.53 0.22 0.52 0.34 0.04

28

Table 5: Main in vivo study pharmacokinetic parameters (n = 6). formulation enoxaparin solution intravenous NL9 containing 2% ENOX/DOA MC10 containing 2% ENOX/DOA

Cmax [IU/mL] 0.3133 0.2412 0.2714

tmax [h] 1.5 1.5

AUC0-6 [IU h/mL] 0.9898 0.9983 1.1122

absolute bioavailability [%] 100 2.02 2.25

Table 6: Cell viability rates (in %) of Caco-2 cells after incubation for 1 h and 4 h with 0.25% (w/v) and 0.5% (w/v) placebo SEDDS and SEDDS containing 2% (w/v) ENOX/DOA. Data are shown as mean ± SD (n = 3). Formulation

0.25 % (w/v)

0.5 % (w/v)

1h

4h

1h

4h

MC10

95.2 ± 0.4

97.2 ± 1.2

98.2 ± 0.3

96.6 ± 1.1

MC10 ± 2% ENOX/DOA

87.7 ± 1.6

67.5 ± 2.1

80.8 ± 2.3

59.3 ± 3.2

NL9

96.1 ± 0.3

98.4 ± 0.4

97.4 ± 2.5

94.2 ± 1.7

NL9 ± 2% ENOX/DOA

84.8 ± 2.7

65.2 ± 0.77

77.9 ± 3.1

55.8 ± 1.3

29

Graphical abstract

30