Evaluation of the cytotoxicity and intestinal absorption of a self-emulsifying drug delivery system containing sodium taurocholate

European Journal of Pharmaceutical Sciences 106 (2017) 212–219

Contents lists available at ScienceDirect

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

Evaluation of the cytotoxicity and intestinal absorption of a self-emulsifying drug delivery system containing sodium taurocholate

MARK

Hang Gaoa, Miao Wangb, Dandan Sunb, Shilin Sunb, Cheng Sunb, Jianguo Liua, Qingxiang Guanb,⁎ a b

The First Hosptial of Jilin University, No. 71, Xinmin Street, Changchun 130021, PR China School of Pharmacy, Jilin University, No. 1266, Fujin Road, Changchun 130021, PR China

A R T I C L E I N F O

A B S T R A C T

Chemical compounds studied in this article: Puerarin (PubChem CID: 5281807) Maisine™35-1 (PubChem CID: 5367328) Propane-1,2-diol (PubChem CID: 1030) Sodium taurocholate (PubChem CID: 23666345) D-Glucose (PubChem CID: 5793) Verapamil (PubChem CID: 2520) 2,4-dinitrophenol (PubChem CID: 1493) Monosodium phosphate (PubChem CID: 23672064) Trypsin (PubChem CID: 72699210) Methylthiazolyldiphenyl - tetrazolium bromide (PubChem CID: 64966)

Currently, many surfactants used in self-emulsifying drug delivery systems (SMEDDS) can cause gastrointestinal mucosal irritation and systemic toxicity. In the present study, SMEDDS were loaded with pueraria flavones, using sodium taurocholate to replace polyoxyl 40 dydrogenated castor oil (Cremophor® RH 40) as the surfactant (PFSMEDDSNR) to reduce the toxicity of SMEDDS using Cremophor® RH 40 as the surfactant (PF-SMEDDSR). The absorption rate constants (Ka) and intestinal permeability coefficients (Peff) were measured. The effects of Pglycoprotein inhibitor (verapamil), adenosine triphosphate (ATP) inhibitor (2,4-dinitrophenol), and carrier inhibitor on Ka and Peff values in the ileum were determined. Biological safety was also evaluated. The Ka and Peff values increased for PF-solution concentrations of 200 μg/ml > 100 μg/ml > 400 μg/ml in individual segments of the intestines. The results indicated that Peff values of PF-SMEDDSNR were distinctly higher than those of SMEDDS loaded with pueraria flavones using Cremophor®RH 40 as the surfactant (PF-SMEDDSR) and PFsolution in four intestinal segments. However, the Ka values of PF-SMEDDSNR were higher only in the jejunum and ileum segments compared with those of PF-SMEDDSR and PF-solution. The Ka and Peff values without verapamil were significantly lower than those with verapamil. 2,4-Dinitrophenol had no effect on Ka and Peff values. The Ka and Peff values of PF-SMEDDSNR significantly decreased after perfusing B-SMEDDSNR for 1 h prior to the study. The cell viabilities after exposure to SMEDDSNR were higher than those of SMEDDSR in the range of 81–324 μg/ml. Lactate dehydrogenase release from cells treated with PF-SMEDDSNR or B-SMEDDSNR was significantly lower than that from cells treated with PF-SMEDDSR or B-SMEDDSR at surfactant concentrations of 243 and 324 μg/ml. However, there were no differences with SMEDDS treatment at surfactant concentrations of 0–162 μg/ml. Hence, we conclude that SMEDDS using sodium taurocholate as the surfactant can reduce the toxicity of SMEDDS, meanwhile, maintain the characteristics of SMEDDS, and enhance intestinal absorption.

Keywords: Self-emulsifying drug delivery system Toxicity Bile salts Intestinal absorption in situ single-pass intestinal perfusion

1. Introduction Self-emulsifying drug delivery systems (SMEDDS) are colloidal systems, comprising oils, surfactants, and cosurfactants, which form oilin-water microemulsions in aqueous media with gentle agitation (Agrawal et al. 2015; Wang et al. 2015; Yeom et al. 2016). Recently, much attention has been focused on the development of SMEDDS for their potential to enhance solubility, membrane permeability, and bioavailability of water-insoluble drugs (Agrawal et al. 2015; McConville and Friend 2013; Yeom et al. 2016). SMEDDS have been widely used in the pharmaceutical industry, benefiting from the success of cyclosporin A softgel formulations (e.g., Sandimmune® and Neoral®) (Agrawal et al. 2015; Gao and Morozowich 2006; Yeom et al. 2016). Unfortunately, SMEDDS require a number of surfactants during their preparation. Many surfactants used in SMEDDS can cause

⁎

gastrointestinal mucosa irritation and systemic toxicity (Chen et al. 2011; Huang et al. 2014). Some strategies have been explored to minimise the toxicity of surfactants. Tian et al. prepared a supersaturated SMEDDS by adding polyvinylpyrrolidoneK30 into SMEDDS composed of biphenyl dimethyl dicarboxylate. The in vivo pharmacokinetic data confirmed that these supersaturated SMEDDS could enhance oral drug availability and reduce toxicity (Tian and Quan 2011). Huang et al. developed a novel puerarin self-microemulsion drug delivery system by replacing polysorbate 80 with natural emulsifiers to reduce surfactant-derived toxicity in SMEDDS (Huang et al. 2012). In this study, pueraria flavones (PF) were selected as the model drug. PF are major active ingredients extracted from Radix Puerariae, traditional Chinese herb medicines that are effective in treating diseases, such as hypertension, hypercholesterolemia, coronary heart disease, and angiocardiopathy (Fan et al. 2014; Guo et al. 2009). However,

Corresponding author at: Department of Pharmaceutics, School of Pharmacy, Jilin University, Room 102, No. 1266, Fujin Road, Changchun 130021, PR China. E-mail address: [email protected] (Q. Guan).

http://dx.doi.org/10.1016/j.ejps.2017.06.005 Received 28 February 2017; Received in revised form 26 May 2017; Accepted 3 June 2017 Available online 04 June 2017 0928-0987/ © 2017 Elsevier B.V. All rights reserved.

European Journal of Pharmaceutical Sciences 106 (2017) 212–219

H. Gao et al.

O

HO

OH

2.3. Characterisation of SMEDDS The particle size, polydispersity index (PDI), and zeta potential of SMEDDS were measured on a dynamic light scattering particle size analyser (Zetasizer Nano ZS, Malvern Instruments, UK) at 25 °C. All samples were diluted 100-fold with distilled water, stirred at 100 rpm for 1 min and measured. Each experiment comprised 10 to 120 runs, which depended on the acquisition of a stable reading. In order to estimate the physical stability of SMEDDS after using sodium taurocholate to Cremophor® RH 40 as surfactants, particle size, PDI and zeta potential of SMEDDS were measured at different time points and compared under room temperature. Moreover, transparency was observed and absorbance diluted by 100-fold distilled water before or after centrifuge was measured at 500 nm by Utlraviolet spectrophotometer. Centrifugal stability coefficient (Ke) values were calculated according to eq. (1). All experiments were performed in triplicate. All data were calculated as mean ± S.D.

O OH

O HO

OH OH Fig. 1. Chemical structure of puerarin.

the effectiveness of PF as cardiovascular disease remedies is restricted by their poor solubility (Guo et al. 1998). In the present study, we prepared PF-SMEDDS using either Cremophor® RH 40 (PF-SMEDDSR) or sodium taurocholate and Cremophor® RH 40 (PF-SMEDDSNR) as surfactants. We then evaluated these drug carrier systems as follows. (i)The intestinal absorption of PF-SMEDDSNR in four individual intestinal segments (i.e., duodenum, jejunum, ileum, and colon) was determined via in situ single-pass intestinal perfusion (SPIP) and compared to that of PF-SMEDDSR and PF-solution. (ii)The intestinal absorption mechanism of PF-SMEDDSNR was investigated preliminarily with three absorptive inhibitors (i.e., P-glycoprotein (P-gp) inhibitor, carrier inhibitor, and ATP inhibitor), and (iii) the cellular toxicity of the SMEDDS was studied by MTT method and LDH release (LDH) assay.

Ke = [(A0 − A) A0 ] × 100%

(1)

where A0 represents absorbance before centrifuge, A represents absorbance after centrifuge. 2.4. Release study in vitro Briefly, in vitro release studies were carried out using a dialysis method in 45 mLof pH 6.8 phosphate buffer saline (PBS) on a horizontal stirring instrument. The temperature was maintained at (37 ± 0.5) °C and the stirring speed was set to 100 rpm. PF-SMEEDSR and PF-SMEEDSNR were diluted with pH 6.8 PBS respectively with a final concentration up to 500 μg/ml respectively. Then 5 mL of PFSMEEDSR or PF-SMEEDSNR solution were transferred into a cellophane membrane dialysis bag (8–12 kDa) respectively. At definite time intervals (0, 0.5, 1, 2, 4, 8, 12, 22 and 24 h), An aliquot of 5 ml sample was withdrawn and the same temperature and equivalent volume of fresh release medium was immediately compensated. All samples were filtrated through 0.45 μm membrane and concentration of puerarin was analysed using HPLC method.

2. Materials and methods 2.1. Materials Puerarin (Fig. 1, purity > 99.0%) was purchased from the National Institute for the Control of Pharmaceutical and Biological Products (No. 110752-200912, Beijing, P.R. China).PF were purchased from Xi'an SaiBang Pharmaceuticals and Technology Co., Ltd. (Xi'an, P.R. China, purity: 66.7%). Maisine™ 35-1 was obtained from Gattefossé (SaintPriest, France). Propane-1, 2-diol was provided by Tianjin Guangfu Fine Chemical Research Institute (Tianjin, P.R. China). Polyoxyl 40 dydrogenated castor oil (Cremophor® RH 40) was provided by BASF (Ludwigshafen, Germany). Sodium taurocholate was purchased from Shanghai Ryon Biological Technology Co., Ltd. (Shanghai, P.R. China). Dulbecco's modified Eagle medium (DMEM), foetal bovine serum (FBS), trypsin, methylthiazolyldiphenyl-tetrazolium bromide (MTT), and HPLC-grade methanol were all purchased from Thermo Fisher Scientific Inc. (Madison, USA). F-12 K medium was provided by Wisent Bio products (Montreal, Canada).Endothelial cell growth supplement was purchased from ScienCell (USA). Lactate dehydrogenase (LDH) kit was obtained from Nanjing Jiancheng Bioengineering Institute. All the other chemicals used in the study were of analytical grade and obtained commercially.

2.5. Single-pass intestinal perfusion studies 2.5.1. Preparation of solutions The Krebs-Ringer solution was composed of NaCl (133 mM), KCl (4.7 mM), NaH2PO4 (2.7 mM), NaHCO3 (16.3 mM), MgCl2 (0.2 mM), Dglucose (7.7 mM), and CaCl2 (3.3 mM). Perfusate solutions of PF (100, 200, and 400 μg/ml) were obtained by dissolving PF in Krebs-Ringer solution under ultrasound bath. Perfusate solutions of PF-SMEDDSNR and PF-SMEDDSR were prepared by dissolving PF-SMEDDSNR or PFSMEDDSR, respectively, in Krebs-Ringer solution with a final PF concentration of 200 μg/ml. All sample solutions were prepared at 37 ± 0.5 °C. 2.5.2. Perfusion experiments All experiments were conducted according to the Guidelines for Care and Use of Laboratory Animals of Jilin University. Male Wistar rats (230–270 g) were supplied by the Central Animal Laboratory of Jilin University, P.R. China, and were randomly divided into six groups. The rats were housed and handled at 23 ± 1 °C and a humidity of 65–75% with a 12 h/12 h light/dark cycle. Rats were fasted for 24 h with free access to water prior to experiments. The procedure for the in situ SPIP was performed as previously reported (Dahan et al. 2009; Ho et al. 2008; Mora et al. 2015; Reis et al. 2013). Rats were anaesthetised and affixed supine on a heated surface (37 °C) under suitable lighting. The abdomen was opened through a midline incision of 3–4 cm. The duodenum (4 cm distance from pylorus), jejunum (15 cm distance from pylorus), ileum (20 cm upward from ileocecal valve), and colon (3 cm downward from ileocecal valve)

2.2. Preparation of self-microemulsion formulation Blank self-microemulsion drug delivery systems consisted of Maisine™ 35-1, Cremophor® RH 40, and propane-1,2-diol at mass ratios of 1:3:6 (B-SMEDDSR), and Maisine™ 35-1, sodium taurocholate, Cremophor® RH 40, and propane-1,2-diol at mass ratios of 1:1.5:1.5:6 (B-SMEDDSNR). The mixtures were thoroughly equilibrated at 37 ± 0.5 °C with 100 rpm magnetic stirring. PF-SMEDDSR and PFSMEDDSNR were obtained by adding PF into B-SMEDDSR and BSMEDDSNR at a mass ratio of 1:4.5, respectively, and mixing with a water bath shaker at 37 ± 0.5 °C for 72 h. 213

European Journal of Pharmaceutical Sciences 106 (2017) 212–219

H. Gao et al.

standard solution into blank intestinal perfusate. The HPLC methods were validated for various parameters, including linearity (R2 > 0.9995) and relative standard deviation of inter-day and intraday precision (< 1.0% and 1.5%, respectively). Specificity was verified based on the absence of an interference peak at the retention time of puerarin with blank intestinal perfusate.

were carefully exposed, cannulated at the proximal and distal ends with two flexible polyethylene cannulae and then replaced into the abdomen to minimise curling and kinking. The incision was covered with cotton soaked in 37 °C isotonic saline solution to keep the area moist. All perfusate solutions were placed in a 37 °C water bath before perfusion through the intestinal segments. At the beginning of the study, individual intestinal segments were rinsed with 37 °C isotonic saline for 10 min in order to clear out any residual debris. Segments were then perfused with Krebs-Ringer solution (i.e., blank perfusate solution) at a constant flow rate of 0.2 ml/min for 30 min to ensure steady state conditions had been reached. Pefusate samples were collected in pre-weighed glass tubes at 15-min intervals for 75 min. All samples were weighed and stored at − 20 °C until analysis. Perfusate samples were extracted using methanol with an intestinal perfusate: methanol ratio of 1:3 (v/v), vortexed for 2 min, and then centrifuged at a speed of 10,000 × g for 10 min. The supernatants were filtered through a 0.45 μm membrane filter. Drug concentrations were determined via high-performance liquid chromatography (HPLC). Following completion of the experiment, the length (l) of the perfused intestinal segment was measured. Finally, all animals were euthanized. The net water flux, resulting from water absorption and efflux in the intestinal perfused segment, was measured by gravimetric method (Escribano et al. 2012). The intestinal absorption rate constant (Ka) and intestinal permeability coefficient (Peff) of the drug were calculated according to the following formulae:

K a = (1 − Cout(corr) Cin)⋅Q πr 2l

(2)

Peff = −ln(Cout (corr ) Cin )⋅Q 2πrl

(3)

2.5.6. Data analysis To evaluate the differences among the experimental groups of the in situ SPIP experiment, all results were calculated as mean ± S.D. For multiple-group comparisons, the data were analysed using a mixed model analysis of variance (ANOVA). For a two-group comparison, a two-tailed nonpaired Students t-test was used. P < 0.05 and P < 0.01 were considered statistically significant. 2.6. Biological safety assessment 2.6.1. Cell culture and treatment Human umbilical vein endothelial cells (HUVECs) were maintained in 6-well culture plates supplemented with F-12 K medium, containing 0.4% heparin sodium, 10% FBS, and 1% endothelial cell growth supplement at 37 °C in a humidified, 5% CO2 incubator. The medium was changed every other day. Cells were subcultured via trypsinisation when the cell density reached 80%. For experiments, cells were seeded in a 96-well plate at an initial density of 5 × 104 cells per well and incubated for 24 h at 37 °C. Cells were exposed to SMEDDS with different concentrations of surfactants (0, 81, 162, 243, and 324 μg/ml with a corresponding PF concentration of 0, 50, 100, 150, and 200 μg/ml, respectively).

Cin (μg/ml) represents the drug concentration of the flux intestinal perfusate, Cout (corr) (μg/ml) represents the outlet drug concentration of perfusate corrected for water flux, Q denotes the flow rate (0.2 ml/ min), r is the radius of the intestinal segment (cm), and l is the length of the intestinal segment (cm). The radius (r) of the intestinal segment was assumed to be 0.18 cm (Mora et al. 2015).

2.6.2. MTT test Cell viability was assessed after 48 h exposure to SMEDDS. Twenty microliters of MTT solution (5 mg/ml) was added into each well and incubated for 4 h at 37 °C. The supernatant was removed and 150 μl/ well of DMSO was added to dissolve the formazan crystals. The absorbance was measured at 492 nm with a microplate reader. The cell proliferation ratio was defined as cell viability and calculated according to the following formula (Beg et al. 2015):

2.5.3. Studies of intestinal absorption mechanism The effects of P-gp inhibitor (verapamil) on the Ka and Peff values of PF-SMEDDSNR, PF-SMEDDSR, and PF-solution were evaluated. The effects of ATP inhibitor (i.e., 2, 4-dinitrophenol, DNP, 100 μg/ml) and carrier (i.e., B-SMEDDSNR) on the Ka and Peff values of PF-SMEDDSNR were also determined. The effect of carrier on intestinal absorption was also investigated by perfusing PF-SMEDDSNR after initially perfusing BSMEDDSNR for 1 h.

cell viability (%) = (ODsample − ODblank ) (ODcontrol − ODblank ) × 100% (4) where OD refers to optical density (absorbance). All data are expressed as mean ± S.D. 2.6.3. Assay of lactate dehydrogenase release LDH release to the medium is an indicator of cellular injury. After 48 h exposure to SMEDDS, cell media was collected and centrifuged at 624 ×g for 5 min. The supernatant was assayed, using the LDH kit, at 440 nm, according to manufacturer's protocols.

2.5.4. Studies of physical absorption of intestinal gut The study was designed to investigate whether the loss of PF from intestinal perfusate was caused solely by intestinal absorption. In brief, after terminating in situ SPIP experiments, the intestinal mucus membranes were everted with a glass rod, washed with isotonic saline, and then placed into 10 ml perfusate solution in a 37 ± 0.5 °C water bath. After 2 h incubation, an aliquot of 500 μl perfusate was sampled and analysed as described in section 2.4.2. Data were presented as mean ± S.D.

3. Results 3.1. Characterisation of SMEDDS The average particle size, PDI, and zeta potential of SMEDDS were measured and summarised in Table 1. The order of the mean particle diameters was PF-SMEDDSR > PF-SMEDDSNR > B-SMEDDSR > BSMEDDSNR. The results demonstrated that the particle size and PDI were greatly increased after adding PF into B-SMEDDSNR and BSMEDDSR (P < 0.01). The zeta potential of all samples offered negative surface charge. B-SMEDDSNR and PF-SMEDDSNR carried more negative charge due to NATC. Remarkable differences on average particle size as well as zeta potential of PF-SMEDDSNR versus B-SMEDDSNR, PFSMEDDSR versus B-SMEDDSR, B-SMEDDSNR versus B-SMEDDSR, PFSMEDDSR versus PF-SMEDDSNR (P < 0.01) were found. No distinct differences were observed between the zeta potential of PF-SMEDDSNR

2.5.5. HPLC method validation Puerarin was always analysed as the representative component of PF in the drug absorption study. An aliquot of 20 μl of the supernatant was analysed using HPLC. The LC-20AT HPLC (Shimadzu, Japan) system comprised a LC-20AT pump and SPD-20A UV detector, controlled by Lab-solution software. The analysis was performed using a Diamonsil C18 column (4.6 × 250 mm, 5 μm, Dikma, China) with a refillable pre-column (C18, 4.6 × 10 mm, Dikma, China) at a column temperature of 30 °C. The mobile phase was 25% (v/v) methanol. The flow rate was adjusted to 1.0 ml/min, and spectrophotometric detection was at 250 nm. Calibration curves were prepared by spiking puerarin 214

European Journal of Pharmaceutical Sciences 106 (2017) 212–219

H. Gao et al.

Table 1 The results of particle size, PDI and zeta potential (Mean ± S.D., n = 3). Formulations

Size (nm)

B-SMEDDSR B-SMEDDSNR PF-SMEDDSR PF-SMEDDSNR

20.68 10.65 37.70 34.32

± ± ± ±

PDI 1.87 0.54⁎⁎ 0.24a 1.24ΨΨ,b

0.22 0.23 0.28 0.30

Zeta potential (mV) ± ± ± ±

0.01 0.03 0.01c 0.01d

− 16.60 − 27.73 − 19.33 − 26.50

± ± ± ±

1.39 1.75⁎⁎ 1.23e 1.30ΨΨ

⁎⁎

P < 0.01 versus B-SMEDDSR. P < 0.01 versus PF-SMEDDSR. P < 0.01 versus B-SMEDDSR. P < 0.01 versus B-SMEDDSNR. P < 0.01 versus B-SMEDDSR. P < 0.01 versus B-SMEDDSNR. P < 0.05 versus B-SMEDDSR.

ΨΨ a b c d e

Fig. 3. The PDI of B-SMEDDSR, B-SMEDDSNR, PF-SMEDDSR and PF-SMEDDSNR at 0, 8, 24, 48 and 72 h. *P < 0.05 and **P < 0.01 indicate significant differences within SMEDDSNR and SMEDDSR. #P < 0.05 and ##P < 0.01 indicate significant differences between B-SMEDDS and PF-SMEDDS. Data are expressed as mean ± S.D. (n = 3).

and B-SMEDDSNR whereas a distinct differences was found between zeta potential of PF-SMEDDSR and B-SMEDDSR (P < 0.05). PDI values of all four SMEDDS formulations were very similar and all below 0.30. The PDI results showed that all SMEDDS formulations possessed good homogeneity. The particle size, PDI and zeta potential of four SMEDDSNR formulations were presented in Fig. 2–4. The results showed that the particle size, PDI and zeta potential of B-SMEDDSR, B-SMEDDSNR, PFSMEDDSR and PF-SMEDDSNR under room temperature offered no change in 3 days. Ke value of B-SMEDDSNR (2.3 ± 0.3) at a mass ratio of 5:5 between Cremophor®RH 40 and NATC was greatly smaller than that of B-SMEDDSNR (1.5 ± 0.3) at a mass ratio of 10:0 between Cremophor® RH 40 and NATC(P < 0.01), which demonstrated that BSMEDDSNR at a mass ratio of 5:5 between Cremophor®RH 40 and NATC presented better physical stability.

3.3. In situ intestinal absorption studies In the in situ SPIP experiments, three concentrations (100, 200, and 400 μg/ml) of PF-solution were used to evaluate the effect of concentration on Ka and Peff values. The results were shown in Fig.6. The order of Ka and Peff values in the four intestinal segments were duodenum > jejunum > ileum ≥ colon. The Ka and Peff values of PFsolution at 100 and 200 μg/ml concentration in duodenum presented significant differences in comparison to jejunum, ileum and colon (*P < 0.05 and **P < 0.01) The Ka and Peff values of PF-solution at 200 and 400 μg/ml concentration in duodenum and in jejunum presented significant differences (*P < 0.05). The order of Ka and Peff values were 200 μg/ml > 100 μg/ml > 400 μg/ml in all four intestinal segments. The maximum values of Ka and Peff values were (2.4 ± 0.93) × 10− 2 min− 1 and (0.66 ± 0.23) × 10− 3 cm·min− 1, respectively, in the duodenum at a PF concentration of 200 μg/ml, which revealed that the intestinal segment with the greatest PF-solution absorption was the duodenum. These results were in agreement with those reported previously (Lai et al. 2009). Based on the above results, PF-solution, PF-SMEDDSNR, and PFSMEDDSR at 200 μg/ml concentration were selected for additional study and the results are presented in Fig. 7. Overall, the order of Ka and Peff values was PF-SMEDDSNR > PF-SMEDDSR > PF-solution. The order of Ka and Peff values of PF-SMEDDSNR and PF-SMEDDSR in the intestinal segments was ileum > jejunum > duodenum > colon, which was different from those of PF-solution. The Ka values of PFSMEDDSNR and PF-SMEDDSR were both higher than those of PF-solution in jejunum and ileum (**P < 0.01 or *P < 0.05). The Ka value of PF-SMEDDSNR was higher than that of PF-solution in colon (**P < 0.01). There were differences between the Ka values of PFSMEDDSNR and PF-SMEDDSR in jejunum, ileum, and colon (**P < 0.01 or *P < 0.05). However, the Peff values of PF-SMEDDSNR were higher than those of PF-SMEDDSR only in the duodenum and

3.2. Release study in vitro The release profiles of peurarin from PF-SMEDDSR and PFSMEDDSNR were shown in Fig.5. The amount of released peurarin increased as time prolonged and presented a slow release trend until 12 h. The percentage of peurarin release was (75.3 ± 2.4) % and (93.2 ± 2.7) % at 12 h respectively and then a stable release plateau of peurarin was found after 12 h. The release rate of peurarin from PFSMEDDSNR was faster and higher than that of PF-SMEDDSR during the whole release procedure (P < 0.01). The faster release of peurarin was probably attributed to smaller particle size of SMEDDS nanoparticles having bigger surface area and better contact to release medium, which could be advantageous to drug release. Based on Noyes-Whitney equation, the increase of drug solubility and decrease of particle size would result in an increase in dissolution rate (Böhm and Müller 1999; Hintz and Johnson 1989).

Fig. 2. The particle size of B-SMEDDSR, B-SMEDDSNR, PFSMEDDSR and PF-SMEDDSNR at 0, 8, 24, 48 and 72 h. *P < 0.05 and **P < 0.01 indicate significant differences within SMEDDSNR and SMEDDSR. #P < 0.05 and ## P < 0.01 indicate significant differences between BSMEDDS and PF-SMEDDS. Data are expressed as mean ± S.D. (n = 3).

215

European Journal of Pharmaceutical Sciences 106 (2017) 212–219

H. Gao et al.

Fig. 4. The Zeta potential of B-SMEDDSR, B-SMEDDSNR, PFSMEDDSR and PF-SMEDDSNR at 0, 8, 24, 48 and 72 h. *P < 0.05 and **P < 0.01 indicate significant differences within SMEDDSNR and SMEDDSR. #P < 0.05 and ##P < 0.01 indicate significant differences between B-SMEDDS and PF-SMEDDS. Data are expressed as mean ± S.D. (n = 3).

respectively. The Peff values of PF-solution, PF-SMEDDSNR, and PFSMEDDSR with verapamil versus without verapamil were (0.59 ± 0.06) × 10− 3 cm·min− 1 vs. (0.38 ± 0.15) × 10− 3 cm·min− 1, (1.35 ± 0.17) × 10− 3 cm·min− 1 vs. (1.20 ± 0.15) × 10− 3 cm·min− 1 and (1.20 ± 0.09) × 10− 3 cm·min− 1 vs. (1.19 ± 0.21) × 10− 3 cm·min− 1 respectively. Carrier, B-SMEDDSNR, also affected PF intestinal absorption in SMEDDSNR (Fig. 9). The Ka and Peff values of PFSMEDDSNR were (3.33 ± 0.85) × 10− 2 min− 1 and (1.20 ± 0.15) × 10− 3 cm·min− 1, but after perfusing with B-SMEDDSNR for 1 h, the values were significantly reduced to (2.51 ± 0.41) × 10− 2 min− 1 and (0.69 ± 0.21) × 10− 3 cm·min− 1, respectively (*P < 0.05 or ** P < 0.01). No difference was found between the Ka and Peff values of PFSMEDDSNR with or without DNP. Fig. 5. Release profile of Puerarin from PF-SMEDDSNR and PF-SMEDDSNR formulation determined by a dialysis method. *P < 0.05 and **P < 0.01 indicate significant differences within PF-SMEDDSNR and PF-SMEDDSR. Data are expressed as mean ± S.D. (n = 3).

3.5. Studies of the physical absorption of intestinal gut The PF remaining of incubated fluids from PF-solution, PFSMEDDSNR, and PF-SMEDDSR were shown in Table 2. PF remaining were found to be > 96.5%, which revealed that the loss of PF from intestinal perfusate could be attributed to intestinal absorption during the experiment.

jejunum (**P < 0.01), while no differences were found in the ileum and colon. The Peff values of PF-SMEDDSNR and PF-SMEDDSR were remarkably higher than those of PF-solution in all four intestinal segments (**P < 0.01 or *P < 0.05). The results suggested there were different absorption mechanisms between PF-SMEDDS and PF-solution, but the absorption mechanisms between PF-SMEDDSNR and PFSMEDDSR were similar.

3.6. Biological safety assessment 3.6.1. MTT test The effects of B-SMEDDSNR, B-SMEDDSR, PF-SMEDDSNR, and PFSMEDDSR on HUVECs were shown in Fig. 10. The MTT assay showed that HUVEC viability decreased in a dose-dependent manner with all formulations. The cell viabilities after exposure to B-SMEDDSNR and PFSMEDDSNR were higher than those after exposure to PF-SMEDDSR and B-SMEDDSR in the range of 81–324 μg/ml (*P < 0.05). The differences in cell viabilities among groups were more distinct when surfactant concentrations reached 324 μg/ml. At this concentration, 24.52 ± 0.55 and 22.62 ± 1.04% of cells died after exposure to PFSMEDDSNR and B-SMEDDSNR, respectively, while almost 50% of HUVECs exposed to PF-SMEDDSR and B-SMEDDSR were killed (45.10 ± 0.47 and 46.89 ± 1.14%, respectively). The cell viabilities were significantly lower at 324 μg/ml surfactant concentrations compared to controls without surfactants (#P < 0.05). However, when the

3.4. Intestinal absorption mechanism studies The effects of verapamil, carrier (B-SMEDDSNR) and DNP on Ka and Peff values were shown in Figs. 8 and 9. The Ka values of PF-solution, PFSMEDDSNR, and PF-SMEDDSR without verapamil were obviously lower than those of the three formulations with verapamil (*P < 0.05 or **P < 0.01, Fig. 8). The Ka values of PF-solution, PF-SMEDDSNR, and PFSMEDDSR with verapamil and without verapamil were (3.20 ± 0.15) × 10− 2 min− 1 vs. (1.65 ± 0.48) × 10− 2 min− 1, (3.75 ± 0.26) × 10− 2 min− 1 vs. (3.33 ± 0.85) × 10− 2 min− 1 and (4.22 ± 0.12) × 10− 2 min− 1 vs. (2.88 ± 0.37) × 10− 2 min− 1,

Fig. 6. The profile of Ka values (A) and Peff values (B) of PFsolution in four intestinal segments. *P < 0.05 and **P < 0.01 indicate significant differences within concentration groups. #P < 0.05 and ##P < 0.01 indicate significant differences between concentration groups. Data are expressed as mean ± S.D. (n = 6).

216

European Journal of Pharmaceutical Sciences 106 (2017) 212–219

H. Gao et al.

Fig. 7. The profiles of Ka values (A) and Peff values (B) obtained with a PF concentration of 200 μg/ml in PFSMEDDSNR, PF-SMEDDSR, and PF-solution. Data are expressed as mean ± S.D. (n = 6). *P < 0.05 and **P < 0.01 indicate significant differences within individual intestinal segments.

concentration of surfactants was 81 μg/ml, cell viabilities were > 95%. The viabilities of the cells exposed to B-SMEDDSNR and PF-SMEDDSNR were not significantly different, and neither were those exposed to PFSMEDDSR and B-SMEDDSR.

parameters of SMEDDS, which affect the rate of drug release, drug stability, aggregation, and dispersion. The reason that the particle sizes of PF-SMEDDSNR were lower than those of PF-SMEDDSR might be the special structure of sodium taurocholate, which contains hydrophilic and hydrophobic groups and has supramolecular self-assembly capability. The zeta potential of PF-SMEDDSNR was larger than that of BSMEDDSR and PF-SMEDDSR, possibly the result of negative charges from bile salts. The in situ SPIP model mimics the situation in fasting humans, and offers a valuable technique to evaluate absorption ability (Dezani et al., 2016; Sun et al. 2014). Characteristics of this method include low flow rate (0.2 ml/min), short perfusion time (≤ 2 h), and high tissue activity. The phenol red method and gravimetric method are usually used to correct perfusate volume to avoid experimental error from the small intestine absorbing water. However, it was reported that phenol red was absorbed in the small intestine to some extent, which influenced the drug's intestinal absorption, especially for drugs with low solubility (Nie et al. 2005). Hence, we used the gravimetric method to evaluate drug absorption during the experiment, which provided a possibility to absorb drugs without first-pass metabolism through the liver. In our preliminary experiments, the solubility of PF in the KrebsRinger solution at 37 °C was 83.2 mg/ml, which indicated that PF could be thoroughly dissolved in the perfusate solution. Therefore, the absorption parameters calculated by this method were able to reflect the true characteristics of drug membrane permeability in each intestinal segment. The Ka and Peff values were calculated using the decreased weight method. We had already verified all validation parameters to ensure that this method was reliable and useful. The Ka and Peff values first increased and then decreased as the PF concentration increased, which demonstrated that drug absorption saturated. We speculated that PF might use a carrier-mediated transport mechanism rather than simple passive transport. The Ka and Peff values of PF-SMEDDSNR were higher than those of PF-SMEDDSR in each intestinal segment, possibly attributable to bile salts or sodium taurocholate. Phospholipids are main components of intestinal membranes. Bile salts and phospholipids can form mixed micelles, which have a good affinity for intestinal epithelial cell membranes. Bile salt/phospholipid mixed micelles could carry insoluble drugs in their hydrophobic core to increase drug dissolution, promote

3.6.2. Assay of lactate dehydrogenase (LDH) release The results of LDH release are presented in Fig. 11. The LDH release from H2O2-treated cells as a positive control into the medium greatly increased compared to that from the control and other test groups (##P < 0.01). LDH release from cells was not significantly different at surfactant concentrations from 0 to 162 μg/ml. LDH release from cells treated with PF-SMEDDSNR or B-SMEDDSNR was significantly lower than that from cells treated with PF-SMEDDSR or B-SMEDDSR at surfactant concentrations of 243 and 324 μg/ml (*P < 0.05). LDH release from cells increased at surfactant concentrations > 162 μg/ml. At a surfactant concentration of 324 μg/ml, LDH release from cells exposed to PF-SMEDDSNR, B-SMEDDSNR, PF-SMEDDSR, and B-SMEDDSR was 28.01 ± 1.50, 27.66 ± 1.05, 53.61 ± 1.07, and 52.36 ± 0.94%, respectively. There were also significant differences (#P < 0.05) in LDH released from cells treated with PF-SMEDDSNR and B-SMEDDSNR, or PF-SMEDDSR and B-SMEDDSR as surfactant concentrations increased from 162 to 324 μg/ml. 4. Discussion Bile salts are anionic, steroidal biosurfactants, generated from endogenous bile, which are able to form stable micelles with fatty acids and phospholipids (Dongowski et al. 2005). Sodium deoxycholate, sodium cholate, and sodium taurocholate are the main components of human bile salts. In our study, we investigated the in vitro stability of SMEDDS when surfactants (Cremophor® RH 40) were replaced with sodium deoxycholate, sodium cholate, and sodium taurocholate. The results indicated that the centrifugal stability and light transmittance of SMEDDS were optimal when the Cremophor® RH 40 to sodium taurocholate mass ratio was 5:5 (data not shown). Hence, we selected SMEDDS comprising Maisine 35-1:Cremophor® RH 40:sodium taurocholate:propane-1, 2-diolat mass ratios of 1:1.5:1.5:6 and a 21.99% drug load for further study. Particle size, PDI, and zeta potential are important evaluation

Fig. 8. The profiles of Ka values (A) and Peff values (B) obtained from 200 μg/ml concentrations of PF-solution, PFSMEDDSNR, and PF-SMEDDSR with or without verapamil. *P < 0.05 and **P < 0.01 compared to the corresponding control group. Data are expressed as mean ± S.D. (n = 6).

217

European Journal of Pharmaceutical Sciences 106 (2017) 212–219

H. Gao et al.

Fig. 9. The profiles of Ka values (A) and Peff values (B) of PF-SMEDDSNR in the rat ileum. Values are expressed as mean ± S.D. (n = 6). Control bars represent 200 μg/ml concentration of PF-SMEDDSNR without inhibitors. *P < 0.05 compared to control group.

Table 2 The mean percentage of PF remaining after 2 h in perfusate solution at 37 °C in four intestinal segments (Mean ± S.D., n = 6). Formulations

PF-solution PF-SMEDDSNR PF-SMEDDSR ⁎ #

PF remaining (%) Duodenum

Jejunum

Ileum

Colon

97.0 ± 1.6 100.0 ± 3.0# 96.2 ± 2.8

99.2 ± 2.9 96.5 ± 2.5 97.3 ± 1.4

101.3 ± 2.6 97.6 ± 3.4⁎ 99.4 ± 1.7

97.7 ± 3.0 98.7 ± 3.6 98.0 ± 3.8

P < 0.05 versus PF-solution. P < 0.05 versus PF-SMEDDSR.

Fig. 11. The profiles of LDH release at different surfactant concentrations. Human umbilical vein endothelial cell (HUVEC) viability was evaluated using the LDH assay (n = 5), following PF-SMEDDSNR, B-SMEDDSNR, PF-SMEDDSR, or B-SMEDDSR treatment. Data are presented as mean ± S.D. (n = 5). *P < 0.05 within surfactant concentration groups. #P < 0.05 between surfactant concentration groups. ##P < 0.01 between the positive and negative controls.

mucosa of the gastrointestinal tract to increase drug transmembrane transport. The optimal absorption of PF-solution occurred in the duodenum, while the optimal absorption of PF-SMEDDSNR and PFSMEDDSR was in the ileum. There are two probable explanations for these results. First, it was reported that the apical membranes of rat intestines, especially the ileum, express P-gp abundantly (Sun et al. 2011). When drug concentrations exceed the limits of P-gp, the surplus drug could be transported out of the cells, which would cause decreased drug absorption. Second, a special, lymphatic transport pathway may exist that plays an important role in transporting SMEDDS. Liu et al. (Liu et al. 2015) found that a self-microemulsifying delivery system could promote puerarin intestinal lymphatic transport and blood absorption. Peyer's patches (PPs) are formed by groups of lymphoid follicles, which are covered with specialised epithelial cells called microfold cells (M cells). When nanoparticles reached M cells of the luminal surface, the nanoparticles were immediately engulfed by pinocytosis, transported in intracellular vesicles, and then released beneath the M cell. PF could be entrapped in the integrated microemulsion and then pinocytosedby PPs on intestinal epithelial cells. According to a published report (Wang et al. 2010), the distribution of PPs was abundant in the ileum, which might be one reason for greater absorption of SMEDDS in the ileum. The reasons for PF's lower absorption in the colon might be attributed to fewer villi in the colon, which produces a smaller effective absorption surface area than other intestinal segments do. Some studies have shown that puerarin is a substrate for P-gp (Cui et al. 2007, 2008). At the intestinal level, P-gp allows transport of structurally unrelated drugs to the gut lumen. P-gp inhibitions considered a major cause of drug-drug interactions. Verapamil is a classic P-gp inhibitor and is often used to investigate whether drug absorption is affected by P-gp-mediated efflux. Our results showed that verapamil greatly influenced PF intestinal absorption and intestinal permeability.

Fig. 10. Cell viability-concentration profiles with the MTT test. Human umbilical vein endothelial cells (HUVECs) were exposed to B-SMEDDSNR, PF-SMEDDSNR, B-SMEDDSR, or PF-SMEDDSR prior to viability testing. Data are presented as mean ± S.D. (n = 5). *P < 0.05 shows significant differences of B-SMEDDSNR and PF-SMEDDSNR versus BSMEDDSR and PF-SMEDDSR. #P < 0.05 shows a significant difference between untreated cells and those treated with 324 μg/ml surfactant.

drug transmembrane transport, and improve intestinal absorption (Duan et al. 2014). Furthermore, sodium deoxycholate and sodium caprate were advantageous to promote drug absorption by inhibiting the function of P-gp. P-gp is an ATP-dependent multidrug efflux pump and represents an important membrane transporter pump. This energydependent membrane transporter functions to transfer drugs from the epithelial basement membrane to the apical membrane (Lo and Huang 2000). Therefore, we speculated that sodium taurocholate also could inhibit the function of P-gp to increase drug absorption. Another reason might be due to much smaller particle size and faster drug release of PFSMEDDSNR in comparison to PF-SMEDDSR. The Peff values of PF-SMEDDSNR and PF-SMEDDSR were higher than those of PF-solution in the four intestinal segments, while the Ka values of PF-SMEDDSNR and PF-SMEDDSR were higher than those of PF-solution in two intestinal segments (i.e., jejunum and ileum).This could be explained by the small particle size of the self-microemulsion, which could avoid gastrointestinal tract degradation to overcome the enzyme and membrane barriers. Meanwhile, the lower microemulsion surface tension was advantageous to increase droplet contact time with the 218

European Journal of Pharmaceutical Sciences 106 (2017) 212–219

H. Gao et al.

Chen, Y., Chen, C., Zheng, J., Chen, Z., Shi, Q., Liu, H., 2011. Development of a solid supersaturatable self-emulsifying drug delivery system of docetaxel with improved dissolution and bioavailability. Biol. Pharm. Bull. 34, 278–286. Cui, S., Zhao, C., He, Z., 2007. Uptaking characteristics of puerarin in Caco-2 model system. Chinese Traditional and Herbal Drugs 38, 836–839. Cui, S., Zhao, C., He, Z., 2008. Study on absorption mechanism of puerarin using rat everted gut sac. Lishizhen medicine and materia medica research 19, 1715–1716. Dahan, A., West, B.T., Amidon, G.L., 2009. Segmental-dependent membrane permeability along the intestine following oral drug administration: evaluation of a triple singlepass intestinal perfusion (TSPIP) approach in the rat. Eur. J. Pharm. Biopharm. 36, 320–329. Dezani, T.M., Dezani, A.B., Junior, J.B.D.S., Serra, C.H.D.R., 2016. Single-pass intestinal perfusion (SPIP) and prediction of fraction absorbed and permeability in humans: a study with antiretroviral drugs. Eur. J. Pharm. Biopharm. 104, 131–139. Dongowski, G., Fritzsch, B., Giessler, J., Härtl, A., Kuhlmann, O., Neubert, R.H.H., 2005. The influence of bile salts and mixed micelles on the pharmacokinetics of quinine in rabbits. Eur. J. Pharm. Biopharm. 60, 147–151. Duan, Y., Xi, Y., Yang, X., Du, H., Zhai, G., 2014. Preparation of curcumin-loaded bile salt /phospholipid mixed micelles and its physicochemical properties. Journal of Shangdong University (health sciences) 52, 63–67. Escribano, E., Sala, X.G., Salamanca, J., Navarro, C.R., Regué, J.Q., 2012. Single-pass intestinal perfusion to establish the intestinal permeability of model drugs in mouse. Int. J. Pharm. 436, 472–477. Fan, H., Wang, Y., Liu, X., Sun, S., Zhang, G., Guan, Q., 2014. Study on preparation and evaluation in vitro of self-microemulsifying dropping pills of puerarin flavonoids. Special Wild Economic Animal and Plant Research 3, 44–47. Gao, P., Morozowich, W., 2006. Development of supersaturatable selfemulsifying drug delivery system formulations for improving the oral absorption of poorly soluble drugs. Expert Opinion on Drug Delivery 3, 97–110. Guo, J., Sun, Q., Zhou, Q., Yao, K., 1998. Studies on isoflavones from radix puerariae drug carrier systems. Chin. Pharm. J. 33, 598–601. Guo, Y., Meng, S., Xu, H., Liu, J., 2009. A brief review on herbal pharmacokinetics study for puerarin and flavone of radix puerariae lobatae. Chinese Journal of Experimental Traditional Medical Formulae 15, 82–85. Hintz, R.J., Johnson, K.C., 1989. The effect of particle size distribution on dissolution rate and oral absorption. Int. J. Pharm. 51, 7–17. Ho, Y., Lai, M., Yu, H., Huang, D., Hsueh, W., Tsai, T., Lin, C., 2008. Application of rat In situ single-pass intestinal perfusion in the evaluation of presystemic extraction of indinavir under different perfusion rates. J. Formos. Med. Assoc. 107, 37–45. Huang, S., Long, X., Yuan, F., Chen, L., Zhang, M., Cai, B., Liu, C., 2012. Preparation and evaluation of berbenine hydrochloride self-microemulsion containing natural emulsifier in vitro and in vivo. Chinese Journal of New Drugs 21, 2794–2799. Huang, S., Long, X., Yuan, F., Cai, B., Ding, R., 2014. Comparative study on properties of natural emulsifier self-microemulsion and nonionic surfactants self-microemulsion of puerarin in vitro and in vivo. Chin Pharm J 49, 1530–1534. Lai, J., Liao, Z., Liang, X., Zhao, G., 2009. Research of absorption enhancers on intestinal absorption of puerarin. Journal of Jiangxi University of TCM 21, 30–32. Liu, C., Long, X., Chen, L., Chen, Y., Luo, Q., Liang, H., 2015. Study of self-microemulsifying drug delivery system in promoting lymphatic and hematological transport of puerarin in rats after oral administration. Traditional Chinese Drug Research & Clinical Pharmacology 26, 809–813. Lo, Y., Huang, J., 2000. Effects of sodium deoxycholate and sodium caprate on the transport of epirubicin in human intestinal epithelial caco-2 cell layers and everted gut sacs of rats. Biochem. Pharmacol. 59, 665–672. McConville, C., Friend, D., 2013. Development and characterisation of a self-microemulsifying drug delivery systems (SMEDDSs) for the vaginal administration of the antiretroviral UC-781. Eur. J. Pharm. Biopharm. 83, 322–329. Mora, M.J., Petiti, J.P., Longhi, M.R., Torres, A.I., Granero, G.E., 2015. Intestinal uptake and toxicity evaluation of acetazolamide and its multicomponent complexes with hidroxypropyl-β-cyclodextrin in rats. Int. J. Pharm. 478, 258–267. Nie, S., Pan, W., Yang, X., Liu, H., Liu, Z., 2005. Evaluation of gravimetry in the rat singlepass intestinal perfusion technique. Chinese Journal of New Drugs 14, 1176–1179. Reis, J.M., Dezani, A.B., Pereira, T.M., Avdeef, A., Serra, C.H.R., 2013. Lamivudine permeability study: a comparison between PAMPA, ex vivo and in situ single-pass intestinal perfusion (SPIP) in rat jejunum. Eur. J. Pharm. Sci. 48, 781–789. Sun, M., Zhai, X., Xue, K., Hu, L., Yang, X., Li, G., Si, L., 2011. Intestinal absorption and intestinal lymphatic transport of sirolimus from self-microemulsifying drug delivery systems assessed using the single-pass intestinal perfusion (SPIP) technique and a chylomicron flow blocking approach: linear correlation with oral bioavailabilities in rats. Eur. J. Pharm. Sci. 43, 132–140. Sun, Y., He, X., Yang, X., Dong, C., Zhang, C., Song, Z., Lu, M., Yang, Z., Li, P., 2014. Absorption characteristics of the total alkaloids from Mahonia bealei in an in situ single-pass intestinal perfusion assay. Chin. J. Nat. Med. 12, 554–560. Tian, X., Quan, D., 2011. Assessment of a supersaturated self-microemulsifying formulation of biphenyl dimethyl dicarboxylate in vitro and in vivo. Chin. Pharm. J. 46, 600–604. Wang, W., Zhou, G., Leng, M., Zhou, H., Chen, Y., 2010. The changes of Peyer's patches lymphocytes in experimental terminal ileitis and its significance. Chinese Journal of Clinical Gastroenterology 22. Wang, Z., Gao, Q., Ma, C., Li, Y., 2011. Experimental study on the application of MTT colorimetric method in counting live Escherichia coli. Acta Sci. Circumst. 31, 3642–3650. Wang, X., Jiang, S., Wang, X., Liao, J., Yin, Z., 2015. Preparation and evaluation of nattokinase-l oaded self-double-emulsifying drug delivery system. Asian Journal of Pharmaceutical Sciences 10, 386–395. Yeom, D.W., Son, H.Y., Kim, J.H., Kim, S.R., Lee, S.G., Song, S.H., Chae, B.R., Choi, Y.W., 2016. Development of a solidified self-microemulsifying drug delivery system (SSMEDDS) for atorvastatin calcium with improved dissolution and bioavailability. Int. J. Pharm. 506, 302–311.

Verapamil inhibited P-gp and reduced efflux of puerarin, thus increasing PF absorption. B-SMEDDSNR was used as a carrier inhibitor to assess the effect of carriers on intestinal absorption. DNP is a mitochondrial uncoupler that inhibits the formation of ATP. Therefore, DNP is utilised as an ATP inhibitor. The Ka and Peff values of PFSMEDDSNR were not significantly different with or without DNP (P > 0.05), which suggests that the absorption of PF-SMEDDSNR is not energy-dependant. Collectively, P-gp and carriers influenced the intestinal absorption of PF-SMEDDSNR, and no energy was consumed during membrane transport. Drug membrane transport includes passive transport, carrier-mediated transport, and membrane mobile transport. Facilitated diffusion is a type of carrier-mediated transport, which requires carrier, but not ATP, and possesses saturation kinetics. Hence, we hypothesised that the membrane transport of PF-SMEDDSNR was facilitated diffusion, rather than simple passive transport. The MTT method requires mitochondrial dehydrogenases to reduce MTT to form purple formazan, which positively correlates with viable cell numbers (Wang et al. 2011). In order to ensure that PF was not toxic to cells, a series of PF concentrations was tested in preliminary experiments. We found that cell viability remained at > 95% when PF concentration was 200 μg/ml. The cell viability significantly decreased with PF concentrations > 200 μg/ml. Hence, we selected 200 μg/ml PF-SMEDDS, with a corresponding surfactant concentration of 324 μg/ml, for further study. The results of MTT and LDH assays indicated that the toxicity from surfactants in SMEDDS was diminished by partly replacing Cremophor® RH 40 with sodium taurocholate. 5. Conclusion In this study, SMEDDSNR with nano-sized droplet diameters were successfully prepared. The Ka and Peff values of PF-SMEDDSNR were higher than those of PF-SMEDDSR and PF-solution. The region of optimal absorption in the intestine and the mechanism of absorption of PF-SMEDDSNR and PF-SMEDDSR were different from those of PF-solution. The greatest absorption of PF-SMEDDSNR and PF-solution occurred in the ileum and duodenum, respectively. PF-SMEDDSNR transport was P-gp-and carrier-dependent, but energy-independent. Carriermediated transport and lymphatic transport were speculated to coexist during the absorption of PF-SMEDDSNR across the intestinal membrane. Alteration of PF-SMEDDSNR by partially replacing Cremophor®RH 40 with sodium taurocholate enabled to lessen cellular toxicity of the surfactant in SMEDDS, meanwhile, maintain the original performance of SMEDDS and enhance intestinal absorption of PF. The absorption mechanism studies of PF-SMEDDSNR are still in progress. Conflict of interest The authors declare no conflict of interest. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ejps.2017.06.005. Acknowledgements All authors are grateful to Lichun Zhao professor who provides us with human umbilical vein endothelial cells. The research is supported by Health Department of Jilin University (2016225). References Agrawal, A.G., Kumar, A., Gide, P.S., 2015. Self emulsifying drug delivery system for enhanced solubility and dissolution of glipizide. Colloids Surf. B: Biointerfaces 126, 553–560. Beg, S., Sharma, G., Thanki, K., Jain, S., Katare, O.P., Singh, B., 2015. Positively charged self-nanoemulsifying oily formulations of olmesartan medoxomil: systematic development, in vitro, ex vivo and in vivo evaluation. Int. J. Pharm. 493, 466–482. Böhm, B.H.L., Müller, R.H., 1999. Lab-scale production unit design for nanosuspensions of sparingly soluble cytotoxic drugs. Pharm. Sci. Technol 2, 336–339.

219