Development of Salvianolic acid B–Tanshinone II A–Glycyrrhetinic acid compound liposomes: Formulation optimization and its effects on proliferation of hepatic stellate cells

International Journal of Pharmaceutics 462 (2014) 11–18

Contents lists available at ScienceDirect

International Journal of Pharmaceutics journal homepage: www.elsevier.com/locate/ijpharm

Pharmaceutical Nanotechnology

Development of Salvianolic acid B–Tanshinone II A–Glycyrrhetinic acid compound liposomes: Formulation optimization and its effects on proliferation of hepatic stellate cells Jiahao Lin 1 , Xiuli Wang 1 , Qing Wu, Jundong Dai, Huida Guan, Weiyi Cao, Liangying He, Yurong Wang ∗ Department of Chinese Pharmaceutics, School of Chinese Materia Medica, Eastern Campus, Beijing University of Chinese Medicine, Beijing 100102, PR China

a r t i c l e

i n f o

Article history: Received 16 September 2013 Received in revised form 15 December 2013 Accepted 19 December 2013 Available online 27 December 2013 Keywords: Co-delivery Salvianolic acid B Tanshinone II A Glycyrrhetinic acid Liposomes Hepatic stellate cells Chemical compounds studied in this article: Salvianolic acid B (PubChem CID: 13991587) Tanshinone II A (PubChem CID: 164676) Glycyrrhetinic acid (PubChem CID: 10114)

a b s t r a c t The aim of this study was to systematically optimize and characterize the co-encapsulation process of Salvianolic acid B (Sal B), Tanshinone II A (TSN) and Glycyrrhetinic acid (GA) into liposomes. The liposomes (GTS-lip) were prepared using film hydration method combined with probe sonication to encapsulate two hydrophobic components (TSN and GA), and using pH gradient method to load hydrophilic component Sal B. The concentration of encapsulated drugs was measured by a reversed phase high performance liquid chromatography (RP-HPLC) method. Systematic optimization of encapsulation process was performed using single factor test, orthogonal test in combination with Box–Behnken Design. Optimum conditions are as follows: ratio of GA to lipid (w/w) = 0.08, ratio of Sal B to lipid (w/w) = 0.12 and pH of buffer = 3.3. Based on the conditions mentioned above, encapsulation efficiency of Sal B, TSN and GA reached target levels: (96.03 ± 0.28)%, (80.63 ± 0.91)% and (88.56 ± 0.17)%, respectively. The GTS-lip had a unimodal sizedistribution and a mean diameter of 191.3 ± 6.31 nm. Morphology determination of the GTS-lip indicated that the liposomes were spherical, and there was no free drug crystal in the visual field of transmission electron microscopy. Also, the  potential of GTS-lip was detected to be −11.6 ± 0.35 mV. In vitro release investigation of GTS-lip suggested that the release rate of GTS-lip significantly decreased compared to drug solution. The accumulative release percentage of TSN, GA and Sal B were 10% in 36 h, 4% in 36 h and 77% in 24 h. Meanwhile, GTS-lip exhibited definite activity on proliferative inhibition of hepatic stellate cells (HSC). GTS-lip decreased the viability of the HSC to higher than 75% at two high drug concentration groups in 24 h. At the same time, GTS-lip of two low drug concentration groups increased the inhibition rates by 2.3 folds and 1.9 folds separately at 48 h compared to 24 h. By contrast, inhibition activity of G-T-S solution group showed less change between 48 h and 24 h. The prolonged and enhanced activity in 48 h which GTS-lip group manifested might contribute to its sustained release effect. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Salvianolic acid B (Sal B), Tanshinone II A (TSN) and Glycyrrhetinic acid (GA) are three main active ingredients of anti-hepatic fibrosis couplet herbs Salvia miltiorrhiza – Glycyrrhiza glabra, which have been proved to be more effective against hepatic fibrosis than each single herb by therapeutic experiments (Lin et al.,

∗ Corresponding author at: Department of Chinese Pharmaceutics, School of Chinese Materia Medica, Eastern Campus, Beijing University of Chinese Medicine, No. 6, Zhong Huan Nan Lu, Wang Jing, Chaoyang District, Beijing 100102, PR China. Tel.: +86 10 84738616; fax: +86 10 84738611. E-mail addresses: [email protected], [email protected] (Y. Wang). 1 These authors contributed equally to this work and should be considered co-first authors. 0378-5173/$ – see front matter © 2014 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.ijpharm.2013.12.040

2008; Liu et al., 2005) and clinical data (Hou, 2005; Li, 2007). Previous studies found that TSN which inhibited hepatic stellate cells (HSC) growth by inducing S phase cell cycle arrest in activate drat HSCs via altering cyclins E, A and cdk2 might be anti-fibrotic candidate (Che et al., 2010). Sal B is reported to reverse liver fibrosis in rats by preventing liver cell injury, inhibiting proliferation of HSC and collagen production in vitro (Liu et al., 2002; Wang et al., 2010). GA can prevent the deterioration of hepatic fibrosis and other chronic liver diseases by inhibiting the proliferation and collagen production of HSCs in culture, and down-regulating the mRNA expression of type III and I procollagen (Luk et al., 2007; Wang et al., 2001). Recent reports also show that carriers modified with GA have higher accumulation in the hepatocytes because of abundant GA receptors on hepatocyte membranes (Sheng-Jun et al., 2007; Zhang et al., 2012). So GA performs not only as activity compound, but also as targeting leader. New combination therapy

12

J. Lin et al. / International Journal of Pharmaceutics 462 (2014) 11–18

Fig. 1. Schematic representation of liposomes, which co-encapsulate Salvianolic acid B, Tanshinone II A and Glycyrrhetinic acid.

strategy of three compounds aforementioned for hepatic fibrosis, taking the advantages of co-delivery more than one therapeutic agent in one delivery system, has recently been shown to be more effective than single agent through potential synergistic effects of different treatment mechanisms (Deng et al., 2007; Wang et al., 2007). However, all the three ingredients exhibit poor absolute bioavailability and therefore restricted therapeutic potential when they are administered clinically via the oral route such as capsules or tablets. The dissolutions of TSN and GA in biological liquids are both practically poor due to their poor water solubility. Thus, the oral bioavailability of TSN and GA is quite low. In addition, Sal B is a BCS Class III drug with high solubility and low permeability (Li et al., 2012). Previous studies consistently suggested a very poor membrane permeation of Sal B (Zhou et al., 2009) as well as the low oral bioavailability in vivo (Hyun Kim et al., 2005; Wu et al., 2006). Hence, it is significant to construct a highly efficient delivery system with multifunction in order to acquire cooperation and prolonged biological effect of the formulation consisted of TSN, GA and Sal B. It is known that liposomes for delivery of various drugs have been studied extensively, and some of them have been marketed. Therapeutic agents can be encapsulated in the liposomes, or associated externally to the lipid bilayer. The co-delivery of several agents in liposomes may serve many functions, including solubilizing hydrophobic ingredients, maintaining drugs in a monomeric state for systemic administration, enhancing the co-delivery of the agents in absolute terms by increased circulation times (in some cases, improving uptake), and increasing tissue selectivity (Wang et al., 2006; Yu et al., 2013). Liposomes, moreover, could be administered intravenous route, which appears as the most ideal approach to administer Sal B, TSN and GA in vivo due to their low oral bioavailability and instantaneous pharmacodynamic effects. At present, there are rare reports about co-delivery liposomes, in particular liposomes co-encapsulating three components with different polarity. Low entrapment of therapeutic active agents is one of the main obstacles of liposome quality (Xu et al., 2011). The objective of this investigation is to explore the feasibility of

co-encapsulating TSN, GA and Sal B into liposomes with high encapsulation efficiency (>80%) and small particle size (100–200 nm) through establishing an optimized preparation method (Fig. 1). The co-delivery system was characterized in terms of encapsulation efficiency, particle size, morphology, release property and so on. Meanwhile, in vitro evaluation was described through its effect on proliferation of hepatic stellate cells. 2. Materials and methods 2.1. Chemicals and reagents Sal B, TSN and GA used as raw materials of liposome were purchased from Baoji GuoKang Biotechnology Co., Ltd. (Shanxi, China) with the purity >98%, while what used as quantitative analytes were obtained from National Institutes for Food and Drug Control (Beijing, China). Cholesterol was purchased from Amresco, Inc. (OH, USA). Soybean phospholipid was purchased from Lipoid (Germany). Methanol was of HPLC grade (Thermo Fisher Scientific, USA) and all other reagents were of analytical grade. Ultracel YM100 centrifugal devices (100 kDa) were purchased from Millipore (Billerica, MA, USA). 2.2. Cell line The human hepatic stellate cell line was provided by Wangjing Hospital (Beijing, China) and cultured using Dulbecco’s Modified Eagle’s Medium (DMEM) with low glucose and fetal bovine serum (FBS) purchased from Hyclone (Logan, UT, USA) in a 5% CO2 air incubator at 37 ◦ C. Trypsin solution and trypsin neutralizer solution were purchased from Sciencell (Carlsbad, CA, USA). 2.3. Preparation of GTS-lip Two hydrophobic ingredients, namely TSN and GA were embedded into phospholipid bilayers to form Tanshinone II A – Glycyrrhetinic acid compound liposomes (GT-lip) firstly. Preparation

J. Lin et al. / International Journal of Pharmaceutics 462 (2014) 11–18



Table 1 Levels of factors used in BBD.

OD =

Factors

Code

Ratio of GA to lipid (w/w) (A) Ratio of Sal B to lipid (w/w) (B) pH of buffer (C)

X1 X2 X3

Range and levels

dt

(2)

−1

0

1

where, n = number of responses.

0.026 0.115 3.3

0.051 0.128 5.4

0.077 0.141 6.9

2.5. Encapsulation efficiency determination

2.4. Experiment methods Encapsulation efficiency of three drugs and particle size of GTSlip were selected as evaluation indexes of the whole screening progress which was divided into two primary steps to obtain optimal conditions with less and reasonable experiments. The liposome preparation process was the same as Section 2.3. 2.4.1. Step one: encapsulation of two hydrophobic components Three mainly effective factors (ratio of lipid to chloroform (w/w) (A), ratio of lipid to TSN (n/n) (B) and ratio of lipid to GA (n/n) (C)) among six key single ones (type of phospholipid, ratio of lipid to chloroform, ratio of lipid to TSN, ratio of lipid to GA, water bath temperature, power of ultrasonic probe) associated with this process got further optimized using orthogonal L9 (34 ) test design. The three factors were tested at three different levels choosing encapsulation efficiency of TSN (EETSN ) as evaluation index. 2.4.2. Step two: encapsulation of one hydrophilic component Ratio of GA and Sal B to lipid, pH of buffer got further optimized and were subjected to response surface methodology (RSM) with a three factor – three coded level Box–Behnken Design (BBD) (Table 1) as they were considered to be key ones based on single factor test, which was used to screen various new factors involved in this step (incubation time and temperature, ratio of GA and Sal B to lipid, pH of buffer). Encapsulation efficiency of Sal B (EESal B , Y1 ) and TSN (EETSN , Y2 ) were determined as two response variables. In addition, the overall desirability (OD, Y3 ) with value between 0 and 1 was involved and defined by aggregating the individual desirability values in this study which contained correlated multiple responses. To evaluate OD, each response Yi was converted into an individual desirability function di that may vary over the range 0 ≤ di ≤ 1. If Yi meets the target value, then di = 1, and if it falls beyond the acceptable limit, then di = 0. Otherwise, di was calculated using formula (1). OD was expressed using a geometric mean (formula (2)). According to the preliminary study results, the ideal value range of Y1 and Y2 were determined 78–98% and 65–95%, respectively. The relationship between each factor and index was fitted using data processing software Design – Expert trial version 8.05b (Stat-Ease Inc., Minneapolis). Yi − Ymin Ymax − Ymin

1/n

t=1

of GT-lip was based upon film-ultrasound method. The required amount of TSN, GA, soybean phospholipid and cholesterol were codissolved by 10 ml ethanol. The organic solvent was then removed by vacuum evaporation process at 30 ◦ C for about 30 min. Lipid film was later hydrated with small amount of deionized water. Then ultrasonic probe was applied to adjust the diameter of GTlip (Scientz-II D, Xinzhi, China) with ultrasonic power of 380 W and treatment time of 5 min. The obtained suspension was used as stock liposome to encapsulate hydrophilic ingredient Sal B secondly. A specified amount of Sal B was dissolved by 3 ml glycine – HCl buffer liquid with corresponding pH. Subsequently, a water bath was performed after the Sal B solution was dropped into GT-lip with adequate mixing to obtain GTS-lip finally.

d1 =

n 

13

(1)

One ml GTS-lip was centrifuged at 224 × g at the room temperature for 5 min. A fixed amount of supernatant was diluted into a suitable concentration with methanol and the content of TSN and GA involved in supernatant was called the concentration of encapsulated drug (Ce) (Lin et al., 2009). The residual supernatant was centrifuged continuously at 12,600 × g for 30 min. Then 400 ␮l of finally obtained supernatant was put into an Ultracel YM-100 centrifugal device and centrifuged at 2016 × g for 30 min. The filtrate was collected to determine the free-drug concentration of Sal B (Cf ). To determine the concentration of total drug (Ct ), 100 ␮l liposomes were dissolved by 700 ␮l methanol. The concentration of three drugs was measured by RP-HPLC analysis simultaneously (SHIMADZU LC-20AT pump liquid chromatograph; Supelco Discovery® C18 column, 250 mm × 4.6 mm, 5 ␮m; SupelguardTM Discovery® C18 guard cartridge, 20 mm × 4.0 mm, 5 ␮m). A gradient mobile phase system consisting of MeOH (A)–H2 O (B) with 0.5% formic acid was pumped at a flow rate of 0.8 ml/min with 30 ◦ C column oven temperature, and the column eluents were monitored by dual-wavelength (253 nm for TSN and GA, 289 nm for Sal B). The gradient elution method was as follows: 0–8 min: 45% A → 88% A; 8–20 min: 88% A; 20–25 min: 88% A → 45% A. The encapsulation efficiency (EE) was calculated as follows: EE(%) =

Ce × 100 Ct

 EE(%) =

1−

Cf Ct

(3)

 × 100

(4)

where formula (3) was for EE calculation of TSN and GA, while (4) was for Sal B. 2.6. Particle size and zeta potential Particle size and zeta potential of the liposomes were measured using a Zetasizer Nano ZS analyser (Malvern Instruments Co., Worcestershire, UK). 15 ␮l of prepared liposomes were diluted 100fold with ultra-pure water and shaken up prior to its measurement at 25 ◦ C. The dynamic light scattering data was collected using a helium laser as the light source and mean results were provided by photon correlation spectroscopy (PCS). 2.7. Visualization by transmission electron microscopy (TEM) The morphology of GTS-lip was observed using transmission electron microscopy apparatus (JEM-1230, JEOL, Tokyo, Japan). A drop of sample placed on a carbon-coated copper grid was negatively stained with 2% phosphotungstic acid, then viewed and photographed. 2.8. In vitro drug release study The release of three drugs from GTS-lip was investigated through two different dialysis methods with opposite dialysis directions (Levy and Benita, 1990), and was compared with drugs in physically mixed solution. For TSN and GA, 10 ml release medium (ultra-pure water with 0.5% SDS to solubilize hydrophobic drugs) was placed in a dialysis pocket (MWCO 8000–10,000, Sigma). The pocket was then immersed in 100 ml release medium, which

14

J. Lin et al. / International Journal of Pharmaceutics 462 (2014) 11–18

Table 2 Experimental runs and results of responses for Box–Behnken design. Run

Factor 1 X1

Factor 2 X2

Factor 3 X3

Response 1 Y1

Response 2 Y2

Response 3 Y3

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

−1 −1 1 0 1 0 −1 0 0 0 0 0 1 −1 1

−1 0 1 −1 −1 0 1 1 −1 0 0 1 0 0 0

0 −1 0 1 0 0 0 1 −1 0 0 −1 −1 1 1

72.96 93.52 79.53 88.04 90.26 74.96 80.00 86.50 95.07 75.00 77.06 92.88 95.68 89.25 85.51

71.84 72.20 79.76 85.88 73.36 61.32 79.79 80.15 85.89 65.37 60.21 77.85 82.79 83.74 60.99

0.000 0.431 0.194 0.591 0.413 0.000 0.223 0.463 0.771 0.000 0.000 0.565 0.724 0.593 0.000

contained equivalent liposome or mixed drug solution. While stirring the release medium using the magnetic stirrer at 37 ◦ C, samples (2 ml) were taken at predetermined time intervals from the pocket over 36 h, and isovolumetric fresh medium was replenished at the same time. For Sal B, the inverse process was performed, which meant 4 ml liposome or mixed drug solution was put in the pocket, which was immersed in 100 ml physiological saline totally to maintain sinking condition. Except samples were obtained from the release medium outside of the pocket, the remaining operations were the same as the inverse method mentioned earlier. Concentration of three drugs was determined by HPLC. 2.9. In vitro effects on proliferation of hepatic stellate cells To evaluate the GTS-lip’ activity on proliferative inhibition of hepatic stellate cells (HSC), an in vitro cell culture system using human hepatic stellate cell (HHSC) line was developed. Cells within logarithmic growth phase were suspended in DEME with 10% FBS after 0.25% trypsin digestion. 1 × 105 cells/200 ␮l was plated in 96-well plates and left to adhere and grow for 24 h. Then the cellculture medium was substituted by a series of concentrations of GTS-lip, G-T-S mixed solution with corresponding concentrations, blank-lip, or blank solvent (200 ␮l) and maintained at 37 ◦ C in the humidified atmosphere containing 5% CO2 in air for 24 h and 48 h. The proliferative inhibition rate of the HSCs was determined by MTT assay (Lian et al., 2005) choosing 490 nm as detection wavelength and calculated with the formula (5) below. The conformation of HSCs was monitored by phase-contrast microscopy (Nikon, ECLIPSE TE2000-S, Japan). Values were presented as mean ± SD. Comparison between groups was carried out by Student’s unpaired t-test at the 0.05 significance level. PI(%) =

A0 − Ai × 100 A0

(5)

PI was proliferative inhibition rate of the HSCs, while Ai and A0 were absorbance of several administration groups and blank solvent group respectively. 3. Results 3.1. The optimization of step one: encapsulation of two hydrophobic components The influence to encapsulation efficiency of TSN decreased in the order: A, C, B according to the R values of orthogonal L9 (34 ) test. The maximum EETSN was obtained when ratio of lipid to chloroform, ratio of lipid to TSN, ratio of lipid to GA were 6:1,

30:1, 24:1, respectively (A1 , B3 , C3 ). In addition, the results of several confirmatory experiments according to the optimal composition showed that EETSN , EEGA , particle size of GT-lip were (81.50 ± 0.76)%, (98.63 ± 0.90)%, (120.5 ± 1.62) nm, respectively. 3.2. The optimization of step two: encapsulation of one hydrophilic component A total of fifteen experimental runs were carried out in random order to optimize the three individual parameters in the BBD. Table 2 outlined the experimental condition of each run and their results. The relationship between the three independent variables (X1 , X2 and X3 ) and responses (Y1 , Y2 or Y3 ) was modeled using the following nonlinear polynomial regression equation: Y = b0 + b1 X1 + b2 X2 + b3 X3 + b4 X1 2 + b5 X2 2 + b6 X3 2 + b7 X1 X2 + b8 X1 X3 + b9 X2 X3 where Y represented the response variable and b0 means constant term, b1 –b9 were regression coefficients. In order to simplify the regression model and make accurate prediction for future formulations, insignificant terms (P > 0.2) were rejected, and the model underwent secondary fitting checked by F-test and P-value. Final equations in terms of actual factors were obtained as follows: Y1 = 75.67 + 1.91X1 − 3.48X3 *** − 4.44X1 X2 ** + 2.69X1 2 + 2.32X2 2 + 12.63X3 2*** (R2 = 0.9299, P = 0.0003). Y2 = 62.30 − 8.33X1 X3 *** + 3.19X1 2 + 10.70X2 2*** + 9.44X3 2** 2 (R = 0.8454, P = 0.0005) Y3 = 0.014 − 0.11X3 ** − 0.11X1 X2 − 0.22X1 X3 *** + 0.18X2 2*** + 0.41X3 2*** (R2 = 0.9126, P = 0.0002).where *** and ** indicated the significance at P < 0.01 and P < 0.05, respectively. It can be seen from the models that both Y1 and Y3 have negative correlations with X3 , which means lower pH of buffer might lead to high values of EESal B and OD to meet our requirement. In addition, interaction between X1 and X3 performed for Y2 and Y3 could not be neglected. The determination coefficient R2 and P-value derived from the analysis of variance (ANOVA) demonstrated that the three models were highly significant and reliable for prediction within the range of experimental variables. In order to visualize the predictive models, three-dimensional (3D) surface graphs were mapped by plotting each response versus two significant factors while the third was fixed at its central point. The response surface diagrams of three responses were shown in Fig. 2. Fig. 2 illustrates that three response variables ascended at different degree with an increase of ratio of GA to lipid regardless of the other two factors. The similar trend occurred to Y1 when the ratio of Sal B to lipid increased (Fig. 2A). This phenomenon could be explained that when the amount of lipoid was fixed, limited space was established between the phospholipid bilayers

J. Lin et al. / International Journal of Pharmaceutics 462 (2014) 11–18

15

and (80.63 ± 0.91)% fit well with the maximum predicted value of 99.05% and 86.07%, which demonstrated the accuracy of the predictive model. Meanwhile, EEGA value of GTS-lip was determined as (88.56 ± 0.17)%. The particle size of GTS-lip was 191.3 ± 6.31 nm with even distribution. Value of  potential was detected to be −11.6 ± 0.35 mV. 3.3. Morphology of GTS-lip The morphology of GTS-lip was observed by transmission electron microscopy and the observation results were displayed in Fig. 3. The TEM image shows that most liposomes were spherical particles with approximate size and uniform dispersion. 3.4. In vitro release study Fig. 4 showed the different release properties of Sal B, TSN and GA from GTS-lip prepared according to optimal conditions and from mixed solution consisting of the same drug formulation with liposomes. It was noticed that the release rate of GTS-lip significantly decreased compared to drug solution. For hydrophilic component Sal B, it released over 90% in the initial 5 h from drug solution, while GTS-lip released 30%. Then 47% of the entrapped Sal B was further released during the subsequent 19 h incubation. The Sal B release profile was prominently prolonged by the liposomal encapsulation. For hydrophobic component, though both TSN and GA exhibited obvious effect of slow release whether in liposomes or in mixed solution, there existed a certain degree difference all the time. TSN and GA released 18% and 13% from drug solution respectively while 10% and 4% from liposome within 36 h. Based on our previous study, solubility of TSN and GA in solubilization medium (water with 0.5% SDS) was still very low (TSN 0.0089 mg/ml, GA 0.3369 mg/ml). In this case, there may not be big enough concentration gradient for drug to pass through the dialysis membrane. 3.5. In vitro effects on proliferation of hepatic stellate cells

Fig. 2. Response surface graphs.

and inside the inner water phase. Along with the increasing quantity of drug within certain range, more drug could be encapsulated through interaction between drug and lipid, or even drug and drug. According to the Design – Expert software, the optimum conditions for preparation of GTS-lip were obtained as follows: ratio of GA to lipid (w/w) = 0.08, ratio of Sal B to lipid (w/w) = 0.12 and pH of buffer = 3.3. Several confirmatory experiments were done according to the optimal conditions to test the suitability of the model equations. The mean experimental EESal B and EETSN value of (96.03 ± 0.28)%

The activation of HSC is one of the central pathophysiological mechanisms of hepatic fibrogenesis. Once they are activated, HSCs exhibited a high proliferation rate. This is a mechanism that plays a key role in the progression of fibrosis in chronic liver disease (Bataller and Brenner, 2005; Friedman, 2008). Thus, we investigated the effects of GTS-lip and mixed solution of Sal B, TAN, GA on the proliferation of human HSC. The results indicated that neither blank liposome nor blank solvent had inhibition effect against HSC, nevertheless, GTS-lip and G-T-S solution groups had. The effects varied at different degree with different concentrations and time. Fig. 5(A) shows the cell inhibition rate of HSC exposed to GTS-lip and G-T-S solution for 24 h. High levels of inhibition rate were found out with the drug concentration increased, and C3 and C4 decreased the viability of the HSCs to higher than 75% in two drug forms. Compared to G-T-S solution group, the inhibition rate of GTS-lip group manifested slower increasing trend within 24 h, however, its inhibition activity enhanced in the next 24 h, which was exhibited in Fig. 5(B). The results showed that GTS-lip of C1 and C2 concentration groups increased the inhibition rates by 2.3-folds and 1.9-folds separately at 48 h compared to 24 h. By contrast, inhibition activity of G-T-S solution group showed less change between 48 h and 24 h. This may be attributed to sustained release effect of GTSlip. Microscopical analysis revealed visual signs of cell bubbling of HSC stimulated with G-T-S solution group and GTS-lip group at C2 concentration for 24 h (Fig. 6A and B). Both groups induced cell death in different extent, which showed out a detachment of HSCs while morphology of blank liposome group and blank solvent group (Fig. 6C and D) did not get affected.

16

J. Lin et al. / International Journal of Pharmaceutics 462 (2014) 11–18

Fig. 3. TEM photographs of GTS-lip magnified to 40,000 times (A) and 100,000 times (B).

4. Discussion It’s crucial to choose proper quality attributes as target profile at the beginning for process optimization of GTS-lip. Since the encapsulation efficiency affects the quality and clinical effects of the liposomes to a great extent (Chen et al., 2010), the evaluation of EE is a regulatory requirement (FDA, 2002, 2010; ICH, 2003). On the other hand, the particle size should be controlled considering the principle of injection administration. Accordingly, particle size (range of 100–200 nm) and drug encapsulation efficiency (>80%) were defined and further delineated to identify all factors. EE was closely related to structure and character of drug and interaction between drug and phospholipid. Hydrophilic compound is usually encapsulated into inner water phase while lipophilic compound is generally entrapped between lipid bilayers. In this study, TSN and GA are lipophilic ingredients while Sal B is the hydrophilic one. In fact, the total loading space is limited once the amount of phospholipid is fixed. Meanwhile, requirement for small unilamellar vesicles in this study leads to reduction of loading space and decrease of EE. Accordingly, keeping EE of TSN, GA and Sal B all over 80% appears to be a big challenge. During the optimum process, TSN performed harder to be entrapped compared to GA. The strong liposolubility, acicular crystal structure and small molecular weight of TSN might induce its leakage from lipid bilayers during the hydration link (Yu et al., 2002). Compared to hydrophobic drugs, the high water solubility makes it difficult to achieve a high degree of entrapment for hydrophilic drugs. As the hydrophilic compound

Fig. 4. In vitro release of drugs (Sal B, TAN and GA) from GTS-lip and mixed solution.

in this study, Sal B presented high EE by using pH gradient method and this method has not been reported to entrap Sal B in other literatures. Based on our early research, Sal B performed some lipophilic character at strong acid environment, which might improve affinity between Sal B and phospholipid, and this is beneficial to guarantee the EE of Sal B. ´ Researchers (Curi c´ et al., 2013; Xu et al., 2011) provide better understanding for the process of preparation of liposomes and the variables occurring in it by the Ishikawa diagram. The Ishikawa diagram contains the common causes separated into five groups: materials, men, environment, equipment and method. These groups include many controllable and uncontrollable variables, and some of which may have potential interactions. Specially,

Fig. 5. In vitro inhibition of GTS-lip and mixed solution against HSC at 24 h (A) and 48 h (B).

J. Lin et al. / International Journal of Pharmaceutics 462 (2014) 11–18

17

Fig. 6. Phase-contrast images of cultured HSC after 24 h incubation with C2 concentration of G-T-S mixed solution (A), GTS-Lip (B), blank-lip (C) and blank-solvent (D).

in this study, there are four quality characters to control (EE of three drugs and particle size of liposomes) throughout the whole preparation process of liposomes encapsulating compounds with different polarities. Thus, in order to detect both the main effects and the interactions between the factors providing the most information with minimal number of runs and without sacrificing the quality of the results (Singh et al., 2011), the optimization process in this study is decomposed into two steps (encapsulation of two hydrophobic components firstly and one hydrophilic component subsequently). During each step, OFAT (analyzing one factor at time) was used first to find out main effective factors (Kleppmann, 2008) DOE (design of experiments) was used additionally to understand the factor coherences. Based on the preliminary on factor test results, we found that the particle size could reach the standard after the probe ultrasonic process, and EEGA could keep higher than 80% all along. Therefore, maximal EE of TAN and Sal B were determined to be two critical objectives in the subsequent DOE process, and this decreased unnecessary work of the complex system. Response surface methodology (RSM) is a collection of mathematical and statistical technique which quantifies the functional relationship between a number of measured response variables and several explanatory factors, hereby to acquire an optimal response by using a sequence of tests (Yang et al., 2012). Most of the work in response surface methodology (RSM) has been focused on the case where there is only one response of interest (Ma et al., 2013; Zhao et al., 2012). In this study, determination of optimum conditions on the input variables would require simultaneous consideration of all responses of interest, which is called a multiresponse problem (Myers and Anderson-Cook, 2009). Multicriteria methodology was applied as an approach when various responses have to be considered at the same time and it is necessary to find optimal compromises between the total numbers of responses taken into account (Bezerra et al., 2008). In this method, the simultaneous optimization process is reduced to find the levels of factors

that demonstrate the maximum overall desirability (OD), which is defined as the weighted geometric average of the individual desirability. In this study, application of OD avoids wasting of reagents and materials, and brings objectivity in the optimization of multiple response procedures. Release profiles of GTS-lip indicated that Sal B was completely entrapped in the liposome with little free in solution or absorbed on the liposome surface, and therefore there was no burst release at the beginning. In this case, the release rate depends on the membrane permeability, which is affected by the fluidity of lipid bilayer (Volodkin et al., 2007). For TSN and GA, their poor solubility in aqueous medium leads to low concentration gradient between two sides of dialysis membrane, and this slows down the passive diffusion rate of drug. Thus, drug release rate is significantly low both from solution and from GTS-lip. Indeed, all three drugs embody significant sustained release effect from GTS-lip in this research. This was also demonstrated by their in vitro effects on proliferation of hepatic stellate cells. However, although reverse dialysis method was used to simulate the condition of drug in blood after intravenous injection, there is still inevitable difference between vivo and vitro environment. Thus, further research such as pharmacokinetics and bio-distribution of GTS-lip should be taken in our undergoing studies. 5. Conclusion After systematical screening and optimization including single factor test, orthogonal test and BBD, the particle size of liposomes could be controlled as 191.3 ± 6.31 nm and EE of three drugs could all be kept over 80% with little variation. The present study suggested that GTS-lip could inhibit the proliferation of human HSCs effectively. Meanwhile, all three drugs displayed sustained release trend to some extent in liposomes. We successfully obtained liposomes containing TSN, GA and Sal B simultaneously.

18

J. Lin et al. / International Journal of Pharmaceutics 462 (2014) 11–18

Acknowledgement This work was mainly supported by National Natural Science Foundation of China (81202928); Beijing Natural Science Foundation (7123118); Chinese pharmaceutical innovative team of compound prescription in Beijing University of Chinese Medicine (no. 2011-CXTD-13). References Bataller, R., Brenner, D.A., 2005. Liver fibrosis. J. Clin. Invest. 115, 209–218. Bezerra, M.A., Santelli, R.E., Oliveira, E.P., Villar, L.S., Escaleira, L.A., 2008. Response surface methodology (RSM) as a tool for optimization in analytical chemistry. Talanta 76, 965–977. Che, X.H., Park, E.J., Zhao, Y.Z., Kim, W.H., Sohn, D.H., 2010. Tanshinone II A induces apoptosis and s phase cell cycle arrest in activated rat hepatic stellate cells. Basic Clin. Pharmacol. Toxicol. 106, 30–37. Chen, C., Han, D., Cai, C., Tang, X., 2010. An overview of liposome lyophilization and its future potential. J. Control. Release 142, 299–311. ´ ´ A., Reul, R., Möschwitzer, J., Fricker, G., 2013. Formulation optimization of Curi c, itraconazole loaded PEGylated liposomes for parenteral administration by using design of experiments. Int. J. Pharm. 448, 189–197. Deng, X., Wang, Q., Zhang, X., Liu, Y., Zeng, Y., Jia, L., Li, P., Cai, D., 2007. Effective mechanism of the glycyrrhetinic acid with tanshinones on immunological liver fibrosis in rat. China Pharm. 10, 741–744. FDA, 2002. Liposome drug products, chemistry, manufacturing, and controls; human pharmacokinetics and bioavailability; and labeling documentation (draft guidance). Guid. Ind. 67, 54220. FDA, 2010. Draft Guidance on Doxorubicin Hydrochloride. FDA. Friedman, S.L., 2008. Mechanisms of hepatic fibrogenesis. Gastroenterology 134, 1655–1669. Hou, C., 2005. Observation of the therapeutic effects of compound glycyrrhizin combined with radix salviae miltiorrhizae on hepatic fibrosis. China Pharm. 16, 53. Hyun Kim, H., Kim, J., Young Ji, H., Chul Kim, Y., Hwan Sohn, D., Mu Lee, B., Suk Lee, H., 2005. Pharmacokinetics of lithospermic acid B isolated from Salvia miltiorrhiza in rats. J. Toxicol. Environ. Health A 68, 2239–2247. ICH, 2003. Stability Testing of New Drug Substances and Products Q1A (R2). ICH. Kleppmann, W., 2008. Taschenbuch Versuchsplanung: Produkte und Prozesse optimieren. Hanser Verlag. Levy, M., Benita, S., 1990. Drug release from submicronized o/w emulsion: a new in vitro kinetic evaluation model. Int. J. Pharm. 66, 29–37. Li, J., Liu, P., Liu, J.-P., Yang, J.-K., Zhang, W.-L., Fan, Y.-Q., Kan, S.-L., Cui, Y., Zhang, W.J., 2012. Bioavailability and foam cells permeability enhancement of salvianolic acid B pellets based on drug-phospholipids complex technique. Eur. J. Pharm. Biopharm. 83, 76–86. Li, Y., 2007. The therapeutic effects of compound glycyrrhizin combined with radix salviae miltiorrhizae on hepatic fibrosis. Proc. Clin. Med. J., 16. Lian, L.H., Park, E.J., Piao, H.S., Zhao, Y.Z., Sohn, D.H., 2005. Aloe emodin-induced apoptosis in t-HSC/Cl-6 cells involves a mitochondria-mediated pathway. Basic Clin. Pharmacol. Toxicol. 96, 495–502. Lin, L., Wang, W., Niu, Y., 2009. Determination of entrapment efficiency for fluorescence red GG liposomes by low-speed centrifugation. J. Zhejiang Univ. Technol. 37, 535–537. Lin, Y.L., Hsu, Y.C., Chiu, Y.T., Huang, Y.T., 2008. Antifibrotic effects of a herbal combination regimen on hepatic fibrotic rats. Phytother. Res. 22, 69–76.

Liu, P., Hu, Y.-Y., Liu, C., Zhu, D.-Y., Xue, H.-M., Xu, Z.-Q., Xu, L.-M., Liu, C.-H., Gu, H.-T., Zhang, Z.-Q., 2002. Clinical observation of salvianolic acid B in treatment of liver fibrosis in chronic hepatitis B. World J. Gastroenterol. 8, 679–685. Liu, S., Peng, D., Li, P., Xiang, R., 2005. Study on components and compatibility of medicines of youhuaganshen compound in resisting hepatic fibrosis. China J. Tradit. Chin. Med. Pharm. 20, 373–375. Luk, J.M., Zhang, Q.S., Lee, N.P., Wo, J.Y., Leung, P.P., Liu, L.X., Hu, M.Y., Cheung, K.F., Hui, C.K., Lau, G.K., 2007. Hepatic stellate cell-targeted delivery of M6P-HSAglycyrrhetinic acid attenuates hepatic fibrogenesis in a bile duct ligation rat model. Liver Int. 27, 548–557. Ma, W.-C., Zhang, Q., Li, H., Larregieu, C.A., Zhang, N., Chu, T., Jin, H., Mao, S.-J., 2013. Development of intravenous lipid emulsion of ␣-asarone with significantly improved safety and enhanced efficacy. Int. J. Pharm. 450, 21–30. Myers, R.H., Anderson-Cook, C.M., 2009. Response Surface Methodology: Process and Product Optimization Using Designed Experiments. Wiley, Wiley.com. Sheng-Jun, M., Yue-Qi, B., Hui, J., Da-Peng, W., Ru, H., Shi-Xiang, H., 2007. Preparation, characterization and uptake by primary cultured rat hepatocytes of liposomes surface-modified with glycyrrhetinic acid. Pharmazie 62, 614–619. Singh, B., Kapil, R., Nandi, M., Ahuja, N., 2011. Developing oral drug delivery systems using formulation by design: vital precepts, retrospect and prospects. Expert Opin. Drug Deliv. 8, 1341–1360. Volodkin, D., Mohwald, H., Voegel, J.-C., Ball, V., 2007. Coating of negatively charged liposomes by polylysine: drug release study. J. Control. Release 117, 111–120. Wang, J.Y., Zhang, Q.S., Guo, J.S., Hu, M.Y., 2001. Effects of glycyrrhetinic acid on collagen metabolism of hepatic stellate cells at different stages of liver fibrosis in rats. World J. Gastroenterol. 7, 115–119. Wang, Q.-L., Tao, Y.-Y., Yuan, J.-L., Shen, L., Liu, C.-H., 2010. Salvianolic acid B prevents epithelial-to-mesenchymal transition through the TGF-␤1 signal transduction pathway in vivo and in vitro. BMC Cell Biol. 11, 31. Wang, T., Deng, Y., Geng, Y., Gao, Z., Zou, J., Wang, Z., 2006. Preparation of submicron unilamellar liposomes by freeze-drying double emulsions. Biochim. Biophys. Acta 1758, 222–231. Wang, X., Zhang, Z.R., Fu, H., Liu, J., Chen, Q., Nie, Y., Deng, L., Gong, T., 2007. Simultaneous determination and pharmacokinetic study of water-soluble and lipid-soluble components of danshen in rat plasma using HPLC-UV method. Biomed. Chromatogr. 21, 1180–1185. Wu, Y.-T., Chen, Y.-F., Hsieh, Y.-J., Jaw, I., Shiao, M.-S., Tsai, T.-H., 2006. Bioavailability of salvianolic acid B in conscious and freely moving rats. Int. J. Pharm. 326, 25–31. Xu, X., Khan, M.A., Burgess, D.J., 2011. A quality by design (QbD) case study on liposomes containing hydrophilic API: I. Formulation, processing design and risk assessment. Int. J. Pharm. 419, 52–59. Yang, S., Chen, J., Zhao, D., Han, D., Chen, X., 2012. Comparative study on preparative methods of DC-Chol/DOPE liposomes and formulation optimization by determining encapsulation efficiency. Int. J. Pharm. 434, 155–160. Yu, B., Zhang, Z., Liu, W., 2002. Method to improve entrapment efficiency of liposomes. Chin. J. Pharm. 33, 564–568. Yu, H., Teng, L., Meng, Q., Li, Y., Sun, X., Lu, J., Lee, J., Teng, R.L., 2013. Development of liposomal Ginsenoside Rg3: formulation optimization and evaluation of its anticancer effects. Int. J. Pharm. 450, 250–258. Zhang, C., Wang, W., Liu, T., Wu, Y., Guo, H., Wang, P., Tian, Q., Wang, Y., Yuan, Z., 2012. Doxorubicin-loaded glycyrrhetinic acid-modified alginate nanoparticles for liver tumor chemotherapy. Biomaterials 33, 2187–2196. Zhao, X., Liu, J., Hu, Y., Fan, Y., Wang, D., Yuan, J., Xu, L., Cui, L., Jing, Z., 2012. Optimization on condition of glycyrrhetinic acid liposome by RSM and the research of its immunological activity. Int. J. Biol. Macromol. 51, 299–304. Zhou, L., Chow, M.S., Zuo, Z., 2009. Effect of sodium caprate on the oral absorptions of danshensu and salvianolic acid B. Int. J. Pharm. 379, 109–118.