Development and Evaluation of Optimized Sucrose Ester Stabilized Oleanolic Acid Nanosuspensions Prepared by Wet Ball Milling with Design of Experiments

Abstract

The aim of this study was to develop optimized sucrose ester (SE) stabilized oleanolic acid (OA) nanosuspensions (NS) for enhanced delivery via wet ball milling by design of experiments (DOE). In this study, SEOA NS batches were prepared by wet ball milling method. Mean particle sizes and polydispersity indices were determined using a nanosizer. The percent encapsulation efficiency, saturation solubility and in vitro dissolution rate were obtained with analyses using HPLC. Preparation methods were optimized by DOE using the Minitab software. The in vitro bioefficacy was obtained by methyl thiazolyl tetrazolium (MTT) measurements in A549 human non small cell lung cancer cell line. The in vivo pharmacokinetics profile was determined using LC-electrospray ionization (ESI)-MS/MS. The study produced spherical SEOA NS particles (ca. 100 nm in diameter) which were found to be able to increase OA saturation solubility considerably. Optimized SEOA-GBD NS (milled at 600 rpm for 3 h, sucrose monolaurate (SEL) : sucrose monopalmitate (SEP) at 9 : 1, w/w; SE : OA at 1 : 1, w/w) was found to be physically stable over 14 d at 4°C. The NS showed much higher dissolution rate, cytotoxicity and bioavailability when compared with the free drug. Thus, the prepared OA as SE stabilized NS particles by wet ball milling enhanced the saturation solubility, in vitro dissolution rate, bioefficacy and in vivo bioavailability of OA. The use of sugar esters may also be potentially applied to other hydrophobic drugs.

In the search for lead drug compounds, nature-derived compounds remain one of the important sources. Many of them have problems associated with low aqueous solubility, low dissolution rate, and consequently, poor bioavailability which limit their potential therapeutic usefulness.^1,2)

Oleanolic acid (OA) is a poorly water-soluble plant-derived triterpenoid which exists widely in natural plants in the form of free acid or aglycones for triterpenoid saponins.³⁾ It has shown to possess hepatoprotective effects, used in both acute liver injury and chronic liver fibrosis and cirrhosis.⁴⁾ In addition, OA had been reported to possess anti-cancer effects including inhibition of tumor initiation and promotion, and induction of tumor cell differentiation and apoptosis.^5,6) Other reported pharmacological activities include anti-inflammatory,⁷⁾ antimicrobial activity,⁸⁾ hypolipidemic and anti-diabetic effects.⁹⁾ However, the low bioavailability of OA limits its application.

The low dissolution rate and poor bioavailability of poorly soluble drugs may be mitigated by altering drugs’ particle sizes.^10,11) The formation of nanosuspension (NS) increased the surface area of a drug due to reduced particle size can improve drug dissolution rate and is often accepted as one of the most expeditious and cost-effective method to improve the dissolution properties of poorly water-soluble drugs. Wet ball milling is a common top-down method to prepare NS. Unlike dry ball milling, the wet method can reduce the amount of heat generated, thereby minimizing the risk of chemical degradation due to heat.

As a group of non-ionic surfactants synthesized by esterification of sucrose with fatty acids, and with safe and biodegradable properties, sucrose esters (SEs) have been widely used for preparing various formulations.^12–14) Although the SEs have many favorable advantages in product formulations, they have not been used much as stabilizers in preparation of NS till in recent years.^15–20) Our previous reported research is the first to employ SEs as sole stabilizers for preparing NS.²¹⁾

Design of experiments (DOE) was first exploited in 1958 by Fisher^22,23) and has since been extensively used for formulation development. DOE is a methodology for studying any response that varies as a function of one or more independent variables. By observing the response under a planned matrix of settings, a statistically valid mathematical model for the responses within the designed space can be determined.²⁴⁾ The use of DOE is very useful for the screening of experimental parameters with much reduced number of experiments.²⁵⁾

Copious publications have discussed the preparation of NS and their stability and oral bioavailability. However, the optimization of NS preparation condition has drawn less attention. Even less publications discuss optimizing NS preparation by DOE.²⁶⁾

Hence, the aim of this study was to produce optimized SEOA NS by tuning the operational parameters for wet ball milling through the application of DOE using Minitab (version 15, Minitab Inc., State College, Pennsylvania, U.S.A.) and to study the in vitro and in vivo properties of the optimized products produced.

MATERIALS AND METHODS

Materials

Sucrose monolaurate (SEL) (batch M07A001, >90% purity) and sucrose monopalmitate (SEP) (batch M07C003, >90% purity) were obtained from Compass Foods Pte. Ltd. (Singapore). OA was purchased from Nanjing Qinze Pharmaceutics Co., Ltd. (Nanjing, China). A549 human non small cell lung cancer cell line (NSCLC®) was purchased from American Type Culture Collection (Rockville, MD, U.S.A.). Methyl thiazolyl tetrazolium (MTT) and F12 Ham Kaighn’s modification (F12K®) medium were purchased from Sigma-Aldrich Co., Ltd. (St. Louis, MO, U.S.A.). Fetal bovine serum (FBS) was purchased from Hyclone Laboratories Co., Ltd. (Logan, UT, U.S.A.). 4-(2-Hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES®) was purchased from Applichem Co., Ltd. (Darmstadt, Germany). MilliQ water (18.2 MΩ cm at 25°C) was obtained from a Millipore Direct-Q ultra-pure water system (Billerica, MA, U.S.A.) and used throughout the study.

Methods. Preparation of SEOA NS by Wet Ball Milling

SEOA NS formulations were prepared by the wet ball milling (S1®, Retsch, Haan, Germany) with 10 mm diameter stainless steel balls. Procedurally, SEP and SEL were mixed with OA in a predetermined ratio and MilliQ water was added to a final volume of 100 mL. The mixture was next transferred into the stainless steel milling chamber containing stainless steel balls. Ball milling was carried out for 25 min followed by a 5 min break for cooling then repeating the cycle until completion of the proposed total milling time. After intervals, aliquots of milled suspension were extracted and transferred into clean BD Falcon® 50 mL conical tubes (BD Bioscience, Mississauga, Canada), centrifuged at 13000 rpm for 10 min and filtered through 0.22 µm to produce the clear NS samples.

NS preparation was optimized with respect to milling time, rotational milling speed, ratio of SEP to SEL, and ratio of SE (SEL and SEP) to OA, by 4 factors and 2 levels factorial design. Milling time was set as 1 h (low level) or 3 h (high level); rotational milling speed was 300 rpm (low level) or 600 rpm (high level); SEL : SEP was 1 : 1 (low level) or 9 : 1 (high level) and SE : OA was 1 : 1 (low level) or 10 : 1 (high level). Details of variables for SEOA NS preparation are shown in Table 1.

Table 1. Ball Mill Settings and Composition of Components for SEOA NS Preparation

NS batcha)	Milling speed (rpm)	Milling time (h)	SEL to SEP ratio (weight ratio, g)b)	SE to OA ratio (weight ratio, g)b)
FBC	600	1	1 : 1 (5 : 5)	10 : 1 (10 : 1)
HAC	300	1	9 : 1 (9 : 1)	10 : 1 (10 : 1)
EAC	300	1	1 : 1 (0.5 : 0.5)	1 : 1 (1 : 1)
FBD	600	3	1 : 1 (5 : 5)	10 : 1 (10 : 1)
HAD	300	3	9 : 1 (9 : 1)	10 : 1 (10 : 1)
EAD	300	3	1 : 1 (0.5 : 0.5)	1 : 1 (1 : 1)
EBC	600	1	1 : 1 (0.5 : 0.5)	1 : 1 (1 : 1)
EBD	600	3	1 : 1 (0.5 : 0.5)	1 : 1 (1 : 1)
GAC	300	1	9 : 1 (0.9 : 0.1)	1 : 1 (1 : 1)
GAD	300	3	9 : 1 (0.9 : 0.1)	1 : 1 (1 : 1)
GBC	600	1	9 : 1 (0.9 : 0.1)	1 : 1 (1 : 1)
GBD	600	3	9 : 1 (0.9 : 0.1)	1 : 1 (1 : 1)
HBC	600	1	9 : 1 (9 : 1)	10 : 1 (10 : 1)
HBD	600	3	9 : 1 (9 : 1)	10 : 1 (10 : 1)
FAC	300	1	1 : 1 (5 : 5)	10 : 1 (10 : 1)
FAD	300	3	1 : 1 (5 : 5)	10 : 1 (10 : 1)

a) The first alphabet (E, F, G or H) encodes the ratio between SEL to SEP and SE to OA, the second (A or B) encodes the milling speed, and the third (C or D) encodes the milling time. b) In parentheses, actual weight ratio of batch components used, in g.

Determination of Particle Size and Polydispersity Index

Zetasizer Nano-ZS90® (Malvern Instruments, Malvern, U.K.) was used for the determination of particle size and polydispersity index at a wavelength of 532 nm. Within 24 h storage in 4°C fridge, the particle size and polydisperisity index of SEOA NS was measured by being diluted by 4 times volume MilliQ water. The scattering angle was fixed at 90° and the temperature of the sample was maintained at 25°C.

Transmission Electron Microscopy (TEM)

For TEM observation, the copper grids were first coated by 0.25% formvar film and then by carbon in a sequential order. The film faces of copper grids were applied with NS samples and then stained with 5% phosphotungstic acid. Excess applied liquids were carefully blotted off, dried for 10 min under bench lamp and samples were ready for examination. TEM photomicrographs were obtained using a TEM (JEM 2010®, JEOL Co., Ltd., Tokyo, Japan) operated at 200 kV.

Percent Encapsulation Efficiency (EE %) and Saturation Solubility

The method of measurement is followed our previous study.²¹⁾ Briefly, EE % and OA saturation solubility of SEOA NS were determined by the results from HPLC with a C18 column and mobile phase of acetonitrile–MilliQ water (65 : 35, v/v) at 1 mL/min. Column temperature was maintained at 24°C and UV detection wavelength was 210 nm. Standards were prepared by dissolving OA in methanol, in the range of 0.02–0.20 mg/mL. Freshly prepared NS samples were dissolved in methanol for OA concentration determination. All samples were filtered through 0.22 µm membranes before measurements. For calculation of the EE % of OA, the following equation was used,

  

(1)

where OA_NS is the amount of OA in NS and OA_T is the total amount of OA added during preparation.

Lyophilization of SEOA NS and Free OA Solution

SEOA NS and free OA solution were frozen at −80°C overnight and then freeze dried (Labconco Corp., Kansas City, MO, U.S.A.) for 24 h at −70°C and 0.02 mbar.

In Vitro Dissolution Test

Dissolution tests followed the established method.²¹⁾ In brief, a dissolution apparatus (Model 2100c®; Distek, North Brunswick, NJ, U.S.A.) was used according to the USP 29 Apparatus 2 (United States Pharmacopeia Convention, Inc., Rockville, MD, U.S.A., 2006). The dissolution medium was 500 mL phosphate buffer solution (pH 7.4) containing 1% (w/v) sodium dodecyl sulfate (SDS) thermostated at 37±0.5°C with paddles rotated at 100 rpm. SEOA-GBD NS, SEOA-GBD NS lyophilized powder, OA coarse suspension (suspended in N,N-dimethylacetamide (DMAC)–polyethylene glycol (PEG) 400–water in the ratio of 2 : 4 : 1, v/v/v) and SEOA-GBD NS were contained in dialysis bags (MWCO 2000®; Spectrum Medical Industries, Singapore) and suspended in the dissolution media, each bag containing an amount equivalent to about 8 mg OA. Samples (3 mL) were withdrawn at predetermined time intervals and filtered through 0.22 µm filters. After each withdrawal, an equal volume of the dissolution medium was replaced. The OA content in each withdrawn sample was determined by HPLC. All dissolution experiments were triplicated and reported results were the mean values.

Stability Study

The effect of storage time on the stability of SEOA NS at 4°C was investigated. Physical stability was evaluated by the percentage change in mean particle sizes of SEOA NS after storage for 15 and 30 d as well as by the relative concentration, OA concentration as a percentage of original concentration, in the NS after filtration, at the same time intervals. Each measurement was repeated thrice, and all studies were carried out by three independent experiments.

Cytotoxicity of OA and SEOA NS

Followed our established method,²¹⁾ A549 human NSCLC cells were cultured in F12 Ham Kaighn’s modification (F12K®) media, supplemented with 10% FBS, 10 mM HEPES, 100 U/mL penicillin G and 100 µg/mL streptomycin. The cells were maintained at 37°C in a 5% CO₂ humidified incubator. To determine cytotoxicity of OA and SEOA NS, A549 cells were seeded in 96-well plates at a density of 6×10³ cells per well and incubated for 24 h at 37°C in a 5% CO₂ humidified incubator. Culture media were then removed and replaced with 100 µL fresh media (blank), fresh media containing 0.5% dimethyl sulfoxide (DMSO) (control), different concentrations of OA (in media with 0.5% DMSO), SEOA-GBD NS or SE (9 : 1) NS. After 24 and 72 h incubation periods, 10 µL MTT (5 mg/mL in phosphate buffered saline (PBS)) aliquots were added to each well. After incubation at 37°C for another 4 h, the mixtures in the wells were withdrawn, and 110 µL DMSO was added to each well and shaken at 100 rpm for 30 min. Absorbance was measured using a multiplate reader (Molecular Devices, Sunnyvale, CA, U.S.A.) at 590 nm. Proliferation rate (%) was calculated as:

  

(2)

Pharmacokinetics Study. Oral Administration of OA to Rats

The study design and animal handling protocol of the pharmacokinetic study were as previously described²¹⁾ and were approved by the Institutional Animal Care and Use Committee of the National University of Singapore. Adult male Sprague-Dawley rats (250–300 g) were purchased from the Laboratory Animal Center of the National University of Singapore. The rats were housed in controlled environment (22±1°C, relative humidity 60–70%) in the Animal Holding Unit operated at a 12-h light/dark cycle. The rats were allowed free access to food and water. A day before performing the pharmacokinetic study, the rats were anesthetized and a polyethylene tube (i.d. 0.58 mm, o.d. 0.965 mm; Becton Dickinson, Sparks, MD, U.S.A.) was inserted into the right jugular vein for blood sample collection. The rats were randomly divided into three groups, each 5 rats, and given oral doses through gavage. To avoid dietary effects on drug absorption, especially from the inherent bile salt solubilisation capacity in the intestine, the rats were fasted overnight prior to the oral gavage but were allowed free access to water. Rats in groups I and II were administered single doses of SEOA-GBD NS at 10 and 20 mg OA/kg, respectively. As comparison, control rats (group III) were administered coarse OA suspension in N,N-DMAC–PEG400–water (2 : 4 : 1, v/v/v) at 20 mg OA/kg. At pre-determined time intervals (5 min, 0.25, 0.5, 1, 1.5, 2, 3, 4, 6, 8, 10, 12 h), blood samples (200 µL) were collected, cannula flushed and blood withdrawn replaced by an equivalent volume of heparin–saline (20 IU/mL heparin in normal saline). Plasma samples were then collected by after centrifugation (3000×g, 5 min) of the collected blood samples and stored at −80°C until LC-ESI-MS/MS analysis.

Sample Preparation and Calibration

Sample preparation aided by liquid–liquid extraction was adopted from our previous study.²¹⁾ As internal standard (IS), a plasma sample (100 µL) was spiked with a methanol solution (5 µL) of glycyrrhetinic acid (GA) (20 µg/mL) in a 2 mL centrifuge tube. Next, ethyl acetate (300 µL) was added and content was well mixed for 1 min. The tube was then centrifuged (13000×g, 10 min) and the ethyl acetate layer was carefully transferred to another clean tube. The extraction procedure was repeated with two additional aliquots of ethyl acetate and cumulated ethyl acetate layers were collected in the same tube. The cumulated ethyl acetate sample was dried under nitrogen flow at 40°C and the residue was reconstituted with methanol (75 µL) and centrifuged (13000×g, 5 min). The supernatant (10 µL) was injected into the HPLC to determine the OA content. Calibration standards were prepared with 100 µL blank plasma samples, adopting the same preparation procedure. The calibration curve was obtained using blank plasma samples spiked with OA and internal standard. The blank plasma samples were obtained from pooled rat plasma. The calibration curve for OA was linear (r²=0.9907) within the range of 20–2000 ng OA/mL.

Chromatography and Tandem Mass Spectrometry Analysis

The concentrations of OA in plasma were determined as previously described.²¹⁾ Briefly, the HPLC system (Agilent 1100® with G1312A binary pump and G1379A degasser; Agilent, Palo Alto, CA, U.S.A.) with a C18 column (300 mm×2 mm i.d., 3 µm ODS; Hypersil Aquasil, Thermo Scientific, Waltham, MA, U.S.A.) and guard column (Inertsil ODS-3®; GL Sciences, Tokyo, Japan). The flow rate of the mobile phase, consisting of acetonitrile and 10 mM ammonium acetate buffer, pH 6.5 (15 : 85, v/v), was 0.3 mL/min. The mass spectrometer was the Qtrap 3000® with an electrospray ionization (ESI) interface (Applied Biosystems, Toronto, Canada). Negative ion ESI with the collision energy −30 V, curtain gas 10 psi and ion source temperature 200°C were used. Quantification was performed with multiple selected reaction monitoring (MRM) mode. The transition of OA is 455.5/455.5 (m/z) and GA (IS) is 469.5/425.5 (m/z) with a scan time of 100 ms per transition.

Results Analysis

WinNonlin® (Standard, version 5.01, Scientific Consulting Inc., Apex, NC, U.S.A.) was used to analyze the pharmacokinetic parameters and a non-compartmental model was adopted. The area under the plasma concentration (AUC) versus time curve (AUC_0→t) from rats after oral administration (Groups I, II and III) was calculated by the linear trapezoidal rule with the time points from 0 to the last detectable time point. Clearance (Cl) values were calculated using the Eq. 3:

  

(3)

Relative bioavailability (rF %) between oral administration groups was calculated as:

  

(4)

Statistical Analysis

Results were presented as mean±standard deviation (S.D.). Statistical significance of the results was analyzed using two-tail independent sample t-test or one-way ANOVA. Values of p<0.05 were considered statistically significant. Minitab® software (version 15, Minitab Inc., State College, PA, U.S.A.) was used to perform factorial design analysis.

RESULTS

Characteristics of SEOA NS. Particle Size and Polydisperse Index (PDI) of Different SEOA NS Formulations

The particle size and PDI of SE-OA NS were determined by the dynamic light scattering (DLS). Figure 1a shows a typical size distribution of SEOA NS by volume (%) with the measured particle size was below 100 nm. Figure 1b gives the comparison chart of all the 16 formulations prepared and Table 2 contains the detailed size data and statistical analyses by one-way ANOVA. The mean particle sizes of the 16 formulations distributed widely, and the overall averaged size was 93.67±53.53 nm. With the exception of batches FBC (252.5±62.1 nm), FBD (156.9±23.7 nm) and HBD (146.7±14.0 nm), the sizes of all other 13 batches were below 100 nm (Table 2).

Fig. 1. (a) Representative Particle Size Distribution Data Obtained from Zetasizer (Nano-ZS90) Instrument by the Intensity of Volume and (b) Particle Size of Different SEOA NS

Data are presented as mean (nm)±S.D. from three independent experiments repeated triplicate.

Table 2. Comparison of Size and PDI

NS	Size (nm)	PDI
1 FBC	252.5±62.13,4,5,6,7,8,9,10,11,12,13,14,15,16	0.96±0.072,3,7,9,10,11,12,13,14,15,16
2 HBD	146.7±14.01,3,4,5,7,8,9,10,11,13,14,15,16	0.46±0.021,3,4,5,6,7,8,9,15
3 HAC	56.4±22.71,2,6	0.24±0.021,2,4,5,6,8,12,13
4 HAD	76.7±5.91,2,6	0.85±0.042,3,6,7,9,10,11,12,13,14,15,16
5 HBC	80.8±15.61,2,6	0.90±0.112,3,7,9,10,11,12,13,14,15,16
6 FBD	156.9±23.71,3,4,5,7,8,9,10,11,12,13,14,15,16	1.00±0.002,3,4,7,9,10,11,12,13,14,15,16
7 FAC	57.3±18.71,2,6	0.21±0.051,2,4,5,6,8,11,12,13,14
8 FAD	35.8±1.11,2,6,12	0.96±0.042,3,7,9,10,11,12,13,14,15,16
9 EAC	77.1±1.91,2,6	0.31±0.021,2,4,5,6,8
10 EAD	79.8±1.01,2,6	0.33±0.051,4,5,6,8
11 EBC	85.5±2.11,2,6	0.35±0.041,4,5,6,7,8
12 EBD	99.2±1.51,6,8	0.44±0.031,3,4,5,6,7,8,15
13 GAC	87.8±1.41,2,6	0.41±0.021,3,4,5,6,7,8
14 GAD	69.0±1.51,2,6	0.35±0.031,4,5,6,7,8
15 GBC	71.5±1.21,2,6	0.31±0.011,2,4,5,6,8,12
16 GBD	65.7±1.71,2,6	0.34±0.041,4,5,6,8

Data represent 3 independent experiments repeated in triplicate. Values are presented as means±S.D. ¹ Significantly different compared with FBC. ² Significantly different compared with HBD. ³ Significantly different compared with HAC. ⁴ Significantly different compared with HAD. ⁵ Significantly different compared with HBC. ⁶ Significantly different compared with FBD. ⁷ Significantly different compared with FAC. ⁸ Significantly different compared with FAD. ⁹ Significantly different compared with EAC. ¹⁰ Significantly different compared with EAD. ¹¹ Significantly different compared with EBC. ¹² Significantly different compared with EBD. ¹³ Significantly different compared with GAC. ¹⁴ Significantly different compared with GAD. ¹⁵ Significantly different compared with GBC. ¹⁶ Significantly different compared with GBD (p<0.05). Statistics were carried out by one-way ANOVA.

The PDI values of the 16 formulation batches also varied considerably, from 0.21 to 1.0. The overall averaged PDI was 0.53±0.29 but if the five formulations with the largest PDI (FBC, HAD, HBC, FBD, and FAD) values were excluded; the PDI value would drop to 0.34±0.08.

The above observations indicated that some NS formulations had “impaired” the overall property of the 16 SEOA NS formulations prepared by wet ball milling. A better understanding of the impact of the four major preparation parameters on the operational mechanisms in the milling process may be elucidated with the help of the Minitab software. This will be discussed in the discussion section.

Morphology of Milled NS Determined by TEM

The morphology and particle size of SEOA NS prepared were obtained by imaging using the TEM. Figure 2 shows the representative TEM photomicrographs of SEOA NS prepared by wet ball milling. As shown (Figs. 2a, b), NS particles were scatter widely on the cooper grid. The primary NS particles were generally spherical in shape with diameters less than 100 nm. The white outer shells on each separate particle were likely to be constituted of layers of surfactants, SEL and SEP, acting as stabilizers (Fig. 2b). The surfactant shell-like layers around the NS particles were most likely to be important to the integrity of the NS particles as they were believed to act as repulsive barriers or coats and aided the separation of constituent particles in the NS suspension.

Fig. 2. Representative TEM of SEOA NS Produced by Wet Ball Milling Method

(a) Scale bar=200 nm, magnification 15000 times. (b) Scale bar=100 nm, magnification 40000 times.

SEOA NS Encapsulation Efficiency (EE %) and Saturation Solubility

The results of saturation solubility and EE % are shown in Fig. 3 and Table 3. The saturation solubilities of OA in SEOA NS ranged from 2.08 to 5.49 mg/mL, which were much higher than free OA and also higher than the previously reported SEOA NS prepared using the emulsion solvent evaporation method. Of all formulations prepared, HBD (5.49±0.16 mg OA/mL, 54.88±1.64%) and FAD (5.36±0.23 mg OA/mL, 53.63±2.32%) had the best results.

Fig. 3. OA Saturation Solubility in SEOA NS

Data are presented as mean (mg/mL)±S.D. from three independent experiments repeated triplicate.

Table 3. Comparison of OA Saturation Solubility and EE %

NS	Saturation solubility (mg/mL)	EE %
1 FBC	4.10±0.133,4,5,6,7,8,9,10,11,12,13,14,15,16	41.03±1.263,4,5,6,7,8,9,10,11,12,13,14,15,16
2 HAC	4.25±0.143,5,6,7,8,9,10,11,12,13,14,15,16	42.49±1.453,5,6,7,8,9,10,11,12,13,14,15,16
3 EAC	2.08±0.001,2,4,5,6,7,8,10,11,12,13,14,15,16	20.82±0.041,2,4,5,6,7,8,10,11,12,13,14,15,16
4 FBD	4.58±0.041,3,5,6,7,8,9,10,11,12,13,14,15,16	45.79±0.431,3,5,6,7,8,9,10,11,12,13,14,15,16
5 HAD	5.36±0.121,2,3,4,6,7,8,9,10,11,12	53.63±1.171,2,3,4,6,7,8,9,10,11,12
6 EAD	2.77±0.051,2,3,4,5,9,13,14,15,16	27.70±0.541,2,3,4,5,9,13,14,15,16
7 EBC	2.55±0.081,2,3,4,5,,8,12,13,14,15,16	25.51±0.771,2,3,4,5,,8,12,13,14,15,16
8 EBD	3.15±0.071,2,3,4,5,7,9,11,13,14,15,16	31.51±0.711,2,3,4,5,7,9,11,13,14,15,16
9 GAC	2.22±0.291,2,4,5,6,8,10,11,12,13,14,15,16	22.17±2.931,2,4,5,6,8,10,11,12,13,14,15,16
10 GAD	2.81±0.281,2,3,4,5,9,13,14,15,16	28.11±2.791,2,3,4,5,9,13,14,15,16
11 GBC	2.68±0.061,2,3,4,5,8,9,13,14,15,16	26.83±0.601,2,3,4,5,8,9,13,14,15,16
12 GBD	3.11±0.101,2,3,4,5,7,9, 13,14,15,16	31.08±0.961,2,3,4,5,7,9, 13,14,15,16
13 HBC	5.04±0.031,2,3,4,6,7,8,9,10,11,12,14	50.41±0.341,2,3,4,6,7,8,9,10,11,12,14
14 HBD	5.49±0.161,2,3,4,6,7,8,9,10,11,12,13,15	54.88±1.641,2,3,4,6,7,8,9,10,11,12,13,15
15 FAC	5.03±0.111,2,3,4,6,7,8,9,10,11,12,14	50.35±1.081,2,3,4,6,7,8,9,10,11,12,14
16 FAD	5.36±0.231,2,3,4,6,7,8,9,10,11,12	53.63±2.321,2,3,4,6,7,8,9,10,11,12

Data represent 3 independent experiments repeated in triplicate. Values are presented as means±S.D. ¹ Significantly different compared with FBC. ² Significantly different compared with HAC. ³ Significantly different compared with EAC. ⁴ Significantly different compared with FBD. ⁵ Significantly different compared with HAD. ⁶ Significantly different compared with EAD. ⁷ Significantly different compared with EBC. ⁸ Significantly different compared with EBD. ⁹ Significantly different compared with GAC. ¹⁰ Significantly different compared with GAD. ¹¹ Significantly different compared with GBC. ¹² Significantly different compared with GBD. ¹³ Significantly different compared with HBC. ¹⁴ Significantly different compared with HBD. ¹⁵ Significantly different compared with FAC. ¹⁶ Significantly different compared with FAD (p<0.05). Statistics were carried out by one-way ANOVA.

Stability of SEOA NS

Physical stability of SEOA NS can be elucidated from the data shown in Tables 4 and 5. Although the initial particle size was very small upon manufacture (Table 4), the particles in most formulations tended to aggregate with time and grow into much larger entities after prolonged storage. For instance, size changes after storage for 15 d ranged from 15.9% to 3510.6% and for 30 d, 37.4% to 4467.4%. Among the formulations prepared, GBD (23.45±2.04% at 15th day and 37.41±4.06% at 30th day) and GAD (15.87±2.96% at 15th day and 46.98±7.90% at 30th day) were the more stable SEOA NS products.

Table 4. Particle Size Change of SEOA NS

NS	15-d size change (%)	30-d size change (%)
1 FBC	1230.91±774.402,3,7,9.10,11,12,13,14,15,16	1405.94±400.952,3,7,9,10,11,12,13,14,15,16
2 HBD	227.65±36.781,3,4,6,7,8	272.71±212.841,3,4,6,7,8
3 HAC	3510.55±1017.021,2,4,5,6,7,8,9,10,11,12,13,14,15,16	4705.99±1354.431,2,4,5,6,8,9,10,11,12,13,14,15,16
4 HAD	1062.11±181.512,3,7,8,9, 10,11,12,13,14,15,16	1989.83±768.562,3,7,9, 10,11,12,13,14,15,16
5 HBC	532.32±182.683,7,8	1294.42±282.783,7,9, 10,11,12,13,14,15,16
6 FBD	1074.16±520.542,3,7,8,9, 10,11,12,13,14,15,16	1868.55±960.092, 3,7,9, 10,11,12,13,14,15,16
7 FAC	2321.58±861.311,2,3,4,5,6,9,10,11,12,13,14,15,16	4467.42±1481.091,2,4,5,6,8, 9, 10,11,12,13,14,15,16
8 FAD	1877.58±813.642,3,4,5,6, 9,10,11,12,13,14,15,16	1882.04±571.842,3,7, 9, 10,11,12,13,14,15,16
9 EAC	127.13±6.481,3,4,6,7,8	245.30±5.151,3,4,5,6,7,8
10 EAD	97.09±5.971,3,4,6,7,8	101.85±3.101,3,4,5,6,7,8
11 EBC	45.92±1.881,3,4,6,7,8	140.82±8.611,3,4,5,6,7,8
12 EBD	27.58±2.371,3,4,6,7,8	126.64±12.931,3,4,5,6,7,8
13 GAC	37.32±2.751,3,4,6,7,8	79.82±12.821,3,4,5,6,7,8
14 GAD	15.87±2.961,3,4,6,7,8	46.98±7.901,3,4,5,6,7,8
15 GBC	31.16±3.551,3,4,6,7,8	105.81±11.421,3,4,5,6,7,8
16 GBD	23.45±2.041,3,4,6,7,8	37.41±4.061,3,4,5,6,7,8

Data represent 3 independent experiments repeated in triplicate. Original particle sizes of NS were set as 100. Values are presented as means±S.D. ¹ Significantly different compared with FBC. ² Significantly different compared with HBD. ³ Significantly different compared with HAC. ⁴ Significantly different compared with HAD. ⁵ Significantly different compared with HBC. ⁶ Significantly different compared with FBD. ⁷ Significantly different compared with FAC. ⁸ Significantly different compared with FAD. ⁹ Significantly different compared with EAC. ¹⁰ Significantly different compared with EAD. ¹¹ Significantly different compared with EBC. ¹² Significantly different compared with EBD. ¹³ Significantly different compared with GAC. ¹⁴ Significantly different compared with GAD. ¹⁵ Significantly different compared with GBC. ¹⁶ Significantly different compared with GBD (p<0.05). Statistics were carried out by one-way ANOVA.

Table 5. Relative Concentration Change of SEOA NS

NS	15-d relative concentration (%)	30-d relative concentration (%)
1 FBC	99.66±0.744,5,7,8,13	84.78±2.924,5,8,14
2 HAC	87.42±0.473,9,11,12	80.85±5.205,12,14
3 EAC	101.55±4.262,4,5,6,7,8,13	87.23±2.144,5,8,14
4 FBD	83.35±1.681,3,9,10,11,12	74.79±5.291,3,9,10,12,16
5 HAD	83.46±6.831,3,9,10,11,12	68.68±4.961,2,3,6,7,9,10,11,12,13,15,16
6 EAD	87.01±8.103,9,11,12	80.81±6.895,12,14
7 EBC	80.71±0.441,3,9,10,11,12	79.95±9.055,9,12
8 EBD	80.01±1.071,3,9,10,11,12	75.73±5.061,3,9,10,12,16
9 GAC	107.24±11.862,4,5,6,7,8,13,14,15,16	88.98±1.584,5,7,8,13,14
10 GAD	99.24±9.574,5,7,8,13	85.55±0.304,5,8,14
11 GBC	101.87±18.302,4,5,6,7,8,13	82.89±0.345,12,14
12 GBD	103.50±9.742,4,5,6,7,8,13,15	93.68±3.702,4,5,6,7,8,11,13,14,15
13 HBC	84.98±5.371,3,9,10,11,12	78.43±11.695,9,12
14 HBD	91.09±5.439	71.12±0.701,2,3,6,9,10,11,12,15,16
15 FAC	89.36±6.929,12	81.42±6.395,12,14
16 FAD	91.52±6.569	86.23±5.364,5,8,14

Data represent 3 independent experiments repeated in triplicate. Original OA concentrations in NS were set as 100. Values are presented as means±S.D. ¹ Significantly different compared with FBC. ² Significantly different compared with HAC. ³ Significantly different compared with EAC. ⁴ Significantly different compared with FBD. ⁵ Significantly different compared with HAD. ⁶ Significantly different compared with EAD. ⁷ Significantly different compared with EBC. ⁸ Significantly different compared with EBD. ⁹ Significantly different compared with GAC. ¹⁰ Significantly different compared with GAD. ¹¹ Significantly different compared with GBC. ¹² Significantly different compared with GBD. ¹³ Significantly different compared with HBC. ¹⁴ Significantly different compared with HBD. ¹⁵ Significantly different compared with FAC. ¹⁶ Significantly different compared with FAD (p<0.05). Statistics were carried out by one-way ANOVA.

Table 5 contains the relative concentration changes for the SEOA NS preparations after filtration. For the first 15 d, most NS products appeared relatively stable (>80% relative concentration). However, after storage for 30 d, the physical stability values obtained varied from 68.7% to 93.7%. Among the formulations produced, GBD (103.50±9.74% at 15th day and 93.68±3.70% at 30th day) was found to be the most stable one.

In Vitro Dissolution

The influence of SEOA NS on the dissolution rate of OA was investigated by the in vitro dissolution profiles of OA NS with unformulated coarse OA suspension as control (Fig. 4a). The dissolution rate of OA coarse suspension (suspended in N,N-DMAC–PEG400–water at 2 : 4 : 1, v/v/v) was very low, with only about 15% of the drug dissolved after 120 min. On the contrary, the SEOA-GBD NS either in suspension form or as a lyophilized powder both showed a marked increase in the dissolution rate for OA, with 100% of OA dissolution achieved within 20 min.

Fig. 4. Dissolution Profiles of OA Coarse Suspension (▲) (Suspended in N,N-DMAC–PEG400–Water at 2 : 4 : 1, v/v/v), SEOA-GBD NS (◇) and SEOA-GBD NS Lyophilized Powder (◆) (a) and SEOA GBD NS in Dialysis Bag (▲) (b) in pH 7.4 Phosphate Buffer Solution Containing 1% SDS (n=3)

The dissolution rate determined by the dialysis bag method was utilized to examine whether the rapid availability of OA from SEOA-GBD NS was in free molecular form or in NS form. As seen in Fig. 4b, no additional increase OA dissolution was seen within 60 min. The dissolution rate was rather low, with the percent of dissolved OA after 120 min less than 5% and not more than 10% even after 1200 min dissolution time.

Cytotoxicity of SEOA NS

In NS form, the saturation solubility of OA increased from 3.43 µg/mL (free OA) to 3110 µg/mL (SEOA-GBD NS). Owing to the increase in OA saturation solubility, the in vitro cytotoxicity to A549 cell lines measured by MTT assay also increased correspondingly. Formulation of SE OA NS significantly increased the cytotoxicity of OA in both time- and dose-dependent manner (Fig. 5, Table 6). The 72 h IC₅₀ dropped from 120 µM of free OA to 45 µM and 24 h IC₅₀ dropped from 130 µM of free OA to 78 µM. Although OA is not considered a potent anti-lung cancer cell line agent, formulating it as NS form considerably improved its bioefficacy. These data showed that it is possible to increase the efficacy of naturally derived hydrophobic compounds by the transformation of the potential drug candidates into the NS form.

Fig. 5. Dose- and Time-Dependent Growth Inhibition of A549 Cells by Free OA Dissolved in Media Containing 0.05% DMSO, and SEOA-GBD NS

x Axis shows OA concentration (µg/mL) and y shows the percentages of viable A549 cells normalized to that of control (%). * p<0.05 between SEOA-GBD 24 h and free OA 24 h; ^# p<0.05 between SEOA-GBD 72 h and free OA 72 h. Data are presented as mean (µg/mL)±S.D. from three independent experiments repeated in quadruplicate.

Table 6. IC₅₀ Comparison of SEOA-GBD NS and Free OA

Group	24 h (µg/mL)	72 h (µg/mL)
Free OA	59.70±1.01 (130.00 µM)	56.80±1.02 (120.00 µM)
SEOA-GBD NS	36.53±1.06 (78.00 µM)	21.14±1.06 (45.00 µM)
SE (9 : 1) NS	249.10±1.06	212.60±1.00

IC₅₀ values were calculated by nonlinear regression (curve fit) of cytotoxicity data in graphs 6a–d using sigmoidal dose response (variable slope) equation, Graphpad Prism software (Graphpad 4.0, Graphpad Software, Inc., La Jolla, CA, U.S.A.).

Pharmacokinetics Results after Oral Administration

Figure 6 and Table 7 show the plasma pharmacokinetic profiles and the pharmacokinetic parameters of either single oral doses of SEOA NS (10, 20 mg/kg) or coarse OA (20 mg/kg) suspensions. In all cases, OA in NS groups resulted in a significantly (p<0.05) higher C_max than the suspension formulations. There were no significant differences (p>0.05) in T_max and T_1/2 between the two groups. However, the NS groups (Groups I and II) had significantly higher relative bioavailability (rF %) values (13.77 and 13.63 over 1.00) than the coarse OA suspension group (Group III) (p<0.05). Nevertheless, there was no statistical significant differences (p>0.05) between 10 mg/kg and 20 mg/kg NS groups. These findings clearly showed that when compared with coarse OA suspension, preparing OA as NS had markedly enhanced the oral bioavailability of OA.

Fig. 6. Mean Plasma Concentration–Time Profiles Comparison of OA in Rats after Oral Administration of OA NS at 10 (▲, n=5), 20 (■, n=5) mg/kg Doses and Oral Administration of OA Coarse Suspension (▼, n=5, Control) at 20 mg/kg Dose

Vertical bars represent standard deviation.

Table 7. Oral Pharmacokinetics Profiles of SEOA NS and Coarse Suspension

Parameter	Group I	Group II	Group III
Formulation	NS	NS	Suspension
Dose (mg/kg)	10.00	20.00	20.00
AUC (µg·min/mL)	65.57±18.62*,***	129.86±65.98*,**	6.70 ±3.40**,***
Tmax (min)	18.00±6.71	21.00±8.22	13.00±4.50
Cmax (ng/mL)	1101.60±250.84*,***	1896.00±436.38*,**	70.00±42.70**,***
T1/2 (min)	68.00±53.96	65.42±15.25	102.10±16.56
rF %	13.77±3.91	13.63±6.93	1.00**,***

Data are presented as mean±S.D., n=5. * p<0.05 between Groups I and II; ** p<0.05 between Groups II and III; *** p<0.05 between Groups I and III.

DISCUSSION

When designing optimal experimental setting, the experimental method of “changing one separate factor at a time” (COST) or also termed as studying “one variable at a time” (OVAT) is still commonly used.^27–29) However, the aim of optimization may not easily lead to the real optimal set of conditions and may lead to different implications, depending on the starting point investigated.²⁵⁾ The approach also requires many experiments covering a range of possibilities. In contrast, the DOE approach is to use an essential tool for studying complex systems since it offers an organized approach that connects the various experiments in a rational manner, giving rise to more precise information that can be arrived from much fewer experiments.³⁰⁾

When DOE is used, the following issues should be carefully considered²⁵⁾ and they are listed as follows.

a) Factor: Experimental variable, which can be quantitative (time, temperature, etc.) or qualitative (solvent, buffer, etc.)

b) Response: Property of the system that is being measured;

c) Interaction: A state where two or more factors are dependent on each other;

d) Confounding: Effects that cannot be estimated separately.

The aim of this study was to produce optimized SEOA NS by tuning the operational parameters for wet ball milling through the application of DOE and to study the in vitro and in vivo properties of the optimized products produced. Hence the influence of the four main production parameters (milling speed, milling time, SEL : SEP ratio and SE : OA ratio) on the properties of SEOA NS produced were studies by a factorial design and analysis of the results were carried out by the use of the Minitab software.

The three main parameters with significant effects on particle size of NS prepared were compared collectively (Fig. 7a). Among them, RPM (milling speed) had the steepest slope, which meant that it had the largest impact. In general, lowering milling speeds yielded smaller NS particles. This finding may be attributed to the high milling speed which produced more heat-related effects due to impact energy generated during the ball milling preparation process. The NS products were hence more likely to aggregate to form larger entities post-production, to reduce the surface free energy. Higher SEL to SEP ratio led to smaller particle size products and this was probably due to better stabilization effects of surfactant ratio. Lower SE : OA ratio produced lower drug encapsulation by surfactants and, hence, formed smaller particle size products.

Fig. 7. Optimization of Production Factors on SEOA NS

Main effects plot (data means) for particle size (a), PDI (b), saturation solubility (c), 30-d particle size change (d), 30-d relative concentration (e) and response optimization with optimized conditions in brackets (f).

The four main parameters with significant effects on PDI were comparatively evaluated (Fig. 7b). Amongst the parameters, SE : OA ratio was found to produce the largest slope, which meant that it had the strongest influencing effect. Although higher SE : OA ratio showed higher drug encapsulation efficiency and higher saturation solubility, it also resulted in larger particle size products (Fig. 7a) and wider PDI distributions. This implied reduced stability of the production process. Higher energy and heat input associated with faster milling speed and longer milling time resulted in higher encapsulation of the drug but the PDI of the NS produced became broader. It was evident that with higher SEL : SEP ratio, the particle size (Fig. 7a) and PDI were both lower, suggesting that a close relationship of high SEL : SEP ratio value with optimized structure of SEOA NS produced and better product stability.

As the amount of OA used was set the same, 1 g for all formulations studied, the EE % value could be relevant as a determinant parameter to reflect saturation solubility.

Among the parameters with significant effects on saturation solubility, SE to OA ratio showed the steepest slope (Fig. 7c), hence had the strongest influence on saturation solubility. A possible reason for this was that the presence of a higher concentration of surfactant could have brought about better encapsulation, resulting in more drug being encapsulated or the added OA to be converted as NS particles. Prolonging the milling time was also produced an increase in drug encapsulation.

The three production parameters with significant effects on particle size change (after 30 d storage) were evaluated comparatively (Fig. 7d). Among them, SE : OA ratio showed the steepest slope, suggesting that it had the largest influencing effect. In general, increasing the SE : OA ratio yielded higher saturation solubility, EE % and hence, higher concentration of SEOA NS.

According to Lifshitz–Slesov–Wagner (LSW) theory³¹⁾ as shown by Eq. 5:

  

(5)

where ω is the Ostwald ripening rate, r_N is the droplet radius, C(∞) is saturation solubility (i.e., the molecular concentration that is in thermal equilibrium with a macroscopic bulk phase), γ is the interfacial tension, V_m is the molar volume of the dispersed compound, ρ is the density of the dispersed phase, D is the diffusion coefficient in the solvent, R is the gas constant and T is the temperature in K.

Ostwald ripening rate, ω (indicated as change rate of particle size) correlated with the saturation solubility, C(∞), and interfacial tension, γ. More SE could had reduced γ and slowed down the Ostwald ripening process (reducing ripening rate, ω). However, it could have also increased the solubility of OA, which enhanced C(∞) and fastened the Ostwald ripening process. Hence, if too much SE was used, the fast Ostwald ripening process derived from increased saturation solubility may had overrode the slowing down effect and produced conditions detrimental to physical stability of NS generated.

The two production parameters with significant effects on relative OA concentration change were compared (Fig. 7e). Between the two parameters, SE : OA ratio showed a steeper slope, suggesting a more marked influencing effect. Lower SE : OA ratio produced NS possessing a higher relative OA concentration change during storage or a less stable product.

Based on Figs. 7d and e, results generally showed that lowering the SE : OA ratio (1 : 1) led to higher physical stability. In contrast, the augmentation of SE : OA ratio increased the amount of solubilized OA in SEOA NS but in case too much OA was solubilized and beyond the surfactants’ stabilization capability, reduced product stability was encountered.

From the experimental results and discussion above, SE : OA ratio had consistently been found to play an important role in the production of a good SEOA NS product. The optimal SE : OA ratio of 1 : 1 yielded smaller NS particles but of higher physical stability. Although a higher SE : OA ratio of 10 : 1 brought about higher saturation solubility and EE %, the overall NS product stability would be compromised and hence, such products were not be desirable for further studies. Among all the formulations of NS prepared with SE : OA ratio of 1 : 1, SEOA-GBD NS (milling at 600 rpm for 3 h, SEL–SEP at 9 : 1, w/w; SE–OA at 1 : 1, w/w) was found to be the optimal product, given its highest physical stability, relatively high saturation solubility of OA (3.11 mg/mL) and small mean particle size (65.7 nm) with relatively low PDI (0.34).

Further confirmation of the findings was made by viewing the response optimization results derived from the factorial design study (Fig. 7f). Three production parameters were investigated by setting them for minimum outcomes. The upper limits of 30-d particle size change (%), PDI and particle size were set at 100, 0.5, and 100, respectively. Saturation solubility and 30-d relative concentration (%) were sought for maximum outcomes. Their lower limits were set as 1 and 80, respectively. After response optimization, SE-OA-GBD NS was found to be at the optimal formulation.

The in vitro dissolution findings suggested that most of the fast dissolving OA from the NS formulation measured during in vitro dissolution testing was not the dissolved OA drug but nanoparticulates released which is in agreement with our previous report.²¹⁾

The significant A549 tumor cell line growth inhibition by the SEOA-GBD NS indicated the truly effectiveness in formulating the hydrophobic drugs into NS form in enhancing their bioefficacy. In addition, the enhanced anti-cancer effect was most likely due to the increased saturation solubility of OA rather than the surface active effects of the sucrose ester molecules. From Table 6, it is demonstrated that the control empty SE NS (same SEL : SEP ratio but not containing OA) had much larger IC₅₀ values (249.10 µg/mL at 24 h and 212.60 µg/mL at 72 h) than SEOA-GBD NS (6.82 times and 10.01 times, respectively).

When comparing coarse OA suspension with SEOA-GBD NS in vivo, the latter has indicated a highly enhanced oral bioavailability. In future study, the pharmacodynamics study of SEOA-GBD NS on the hepatic lesion rats’ model or tumor rats’ model will be another angle.

CONCLUSION

Investigations on enhancing the delivery of a poorly water soluble compound, oleanolic acid, as SEOA NS prepared by a top-down wet ball milling were carried out and product formulations critically evaluated. The investigations indicated that all the critical product attributes, SE : OA ratio, SEL : SEP ratio, milling time and milling speed exerted certain influences on the characteristics of SEOA NS produced. In this present study, the mean particle sizes of most SEOA NS prepared were less than 100 nm. Except for some variations, the PDI of most formulations were found to be relatively low. The NS particles were generally spherical in shape and observed to be covered by distinct diffuse coating, possibly of the surfactant, SE, on the periphery of particles or their aggregates. Preparation of OA as NS by wet ball milling also increased its saturation solubility considerably, ranging from 2.08 to 5.49 mg/mL. With the help of the statistical software Minitab, critical analysis of the formulation parameters had identified the ideal parameters to prepare the optimized product, SEOA-GBD NS. The optimized product increased the OA dissolution rate markedly. Most of the dissolved OA existed in the NS in solution and not dissolved as the free molecular form. Formulation of OA as NS significantly and substantially increased the cytotoxicity of OA. It reduced the proliferation rate of A549 cell lines to a much greater extent than control OA in a time- and dose-dependent manner. The results indicated this increased activity was attributed to the nanonized drug and not the SE. Further, NS of OA not only increased its saturation solubility and dissolution rate to a great extent but also changed the pharmacokinetic profile of OA after oral administration. Oral bioavailability of OA was enhanced by the NS formulation, which showed much higher C_max and rF % than the coarse suspension group. Dose-independent pharmacokinetics of OA was observed after oral administration at the range of 10 to 20 mg/kg.

Acknowledgment

This project is partially funded from GEA-NUS Grant (N-148-000-008-001) and Singapore Ministry of Health’s National Medical Research Council under IRG NMRC/1187/2008 (R-148-000-114-213). The authors would like to thank Dr. Paul CL Ho for his support in the DLS measurement. We wish to record appreciation to Ms. Loy Gek Luan and Mr. Chong Ping Lee, Department of Biological Sciences, for their support in EM experiments. We would also like to thank Compass Foods Pte. Ltd., Singapore for the generous supply of SE.

REFERENCES

• 1) Rasenack N, Hartenhauer H, Muller BW. Microcrystals for dissolution rate enhancement of poorly water-soluble drugs. Int. J. Pharm., 254, 137–145 (2003).
• 2) Betancor-Fernández A, Perez-Galvez A, Sies H, Stahl W. Screening pharmaceutical preparations containing extracts of turmeric rhizome, artichoke leaf, devil’s claw root and garlic or salmon oil for antioxidant capacity. J. Pharm. Pharmacol., 55, 981–986 (2003).
• 3) Liu J. Pharmacology of oleanolic acid and ursolic acid. J. Ethnopharmacol., 49, 57–68 (1995).
• 4) Liu J, Liu Y, Parkinson A, Klaassen CD. Effect of oleanolic acid on hepatic toxicant-activating and detoxifying systems in mice. J. Pharmacol. Exp. Ther., 275, 768–774 (1995).
• 5) Ovesná Z, Vachalkova A, Horvathova K, Tothova D. Pentacyclic triterpenoic acids: new chemoprotective compounds. Minireview. Neoplasma, 51, 327–333 (2004).
• 6) Konopleva M, Tsao T, Estrov Z, Lee RM, Wang RY, Jackson CE, McQueen T, Monaco G, Munsell M, Belmont J, Kantarjian H, Sporn MB, Andreeff M. The synthetic triterpenoid 2-cyano-3,12-dioxooleana-1,9-dien-28-oic acid induces caspase-dependent and -independent apoptosis in acute myelogenous leukemia. Cancer Res., 64, 7927–7935 (2004).
• 7) Singh GB, Singh S, Bani S, Gupta BD, Banerjee SK. Anti-inflammatory activity of oleanolic acid in rats and mice. J. Pharm. Pharmacol., 44, 456–458 (1992).
• 8) Horiuchi K, Shiota S, Hatano T, Yoshida T, Kuroda T, Tsuchiya T. Antimicrobial activity of oleanolic acid from Salvia officinalis and related compounds on vancomycin-resistant enterococci (VRE). Biol. Pharm. Bull., 30, 1147–1149 (2007).
• 9) Gao D, Li Q, Li Y, Liu Z, Liu Z, Fan Y, Han Z, Li J, Li K. Antidiabetic potential of oleanolic acid from Ligustrum lucidum Ait. Can. J. Physiol. Pharmacol., 85, 1076–1083 (2007).
• 10) Stefaniak AB, Chipera SJ, Day GA, Sabey P, Dickerson RM, Sbarra DC, Duling MG, Lawrence RB, Stanton ML, Scripsick RC. Physicochemical characteristics of aerosol particles generated during the milling of beryllium silicate ores: implications for risk assessment. J. Toxicol. Environ. Health A, 71, 1468–1481 (2008).
• 11) Lanke SSS, Gayakwad SG, Strom JG, D’souza MJ. Oral delivery of low molecular weight heparin microspheres prepared using biodegradable polymer matrix system. J. Microencapsul., 26, 493–500 (2009).
• 12) Paul BK, Mitra RK. Percolation phenomenon in mixed reverse micelles: the effect of additives. J. Colloid Interface Sci., 295, 230–242 (2006).
• 13) Honeywell-Nguyen PL, Frederik PM, Bomans PH, Junginger HE, Bouwstra JA. Transdermal delivery of pergolide from surfactant-based elastic and rigid vesicles: characterization and in vitro transport studies. Pharm. Res., 19, 991–997 (2002).
• 14) Okamoto H, Sakai T, Danjo K. Effect of sucrose fatty acid esters on transdermal permeation of lidocaine and ketoprofen. Biol. Pharm. Bull., 28, 1689–1694 (2005).
• 15) Kuntsche J, Koch MH, Steiniger F, Bunjes H. Influence of stabilizer systems on the properties and phase behavior of supercooled smectic nanoparticles. J. Colloid Interface Sci., 350, 229–239 (2010).
• 16) Klang V, Matsko N, Zimmermann AM, Vojnikovic E, Valenta C. Enhancement of stability and skin permeation by sucrose stearate and cyclodextrins in progesterone nanoemulsions. Int. J. Pharm., 393, 153–160 (2010).
• 17) Takegami S, Kitamura K, Kawada H, Matsumoto Y, Kitade T, Ishida H, Nagata C. Preparation and characterization of a new lipid nano-emulsion containing two cosurfactants, sodium palmitate for droplet size reduction and sucrose palmitate for stability enhancement. Chem. Pharm. Bull., 56, 1097–1102 (2008).
• 18) Ullrich S, Metz H, Mader K. Sucrose ester nanodispersions: microviscosity and viscoelastic properties. Eur. J. Pharm. Biopharm., 70, 550–555 (2008).
• 19) Arıca Yegin B, Benoit JP, Lamprecht A. Paclitaxel-loaded lipid nanoparticles prepared by solvent injection or ultrasound emulsification. Drug Dev. Ind. Pharm., 32, 1089–1094 (2006).
• 20) Lippacher A, Muller RH, Mader K. Semisolid SLN dispersions for topical application: influence of formulation and production parameters on viscoelastic properties. Eur. J. Pharm. Biopharm., 53, 155–160 (2002).
• 21) Li W, Das S, Ng KY, Heng PW. Formulation, biological and pharmacokinetic studies of sucrose ester-stabilized nanosuspensions of oleanolic acid. Pharm. Res., 28, 2020–2033 (2011).
• 22) Keppel G. Design and analysis: A researcher’s handbook. Prentice-Hall, Inc., Englewood Cliffs, NJ (1991).
• 23) Singh B, Kumar R, Ahuja N. Optimizing drug delivery systems using systematic “design of experiments.” Part I: Fundamental aspects. Critical Reviews™ in Therapeutic Drug Carrier Systems, 22, 27–105 (2005).
• 24) Mathews P. Design of Experiments with MINITAB. ASQ Quality Press, Milwaukee, WI (2005).
• 25) Tye H. Application of statistical ‘design of experiments’ methods in drug discovery. Drug Discov. Today, 9, 485–491 (2004).
• 26) Salazar J, Heinzerling O, Muller RH, Moschwitzer JP. Process optimization of a novel production method for nanosuspensions using design of experiments (DoE). Int. J. Pharm., 420, 395–403 (2011).
• 27) Misra S, Raghuwanshi S, Saxena RK. Statistical approach to study the interactive effects of process parameters for enhanced xylitol production by Candida tropicalis and its potential for the synthesis of xylitol monoesters. Food science and technology international=Ciencia y tecnologia de los alimentos internacional, 19, 535–548 (2013).
• 28) Demir T, Gube O, Yucel M, Hames-Kocabas EE. Increased alkalotolerant and thermostable ribonuclease (RNase) production from alkaliphilic Streptomyces sp. M49-1 by optimizing the growth conditions using response surface methodology. World Journal of Microbiology & Biotechnology, 29, 1625–1633 (2013).
• 29) Moradi M, Yamini Y, Rezaei F, Tahmasebi E, Esrafili A. Development of a new and environment friendly hollow fiber-supported liquid phase microextraction using vesicular aggregate-based supramolecular solvent. Analyst, 137, 3549–3557 (2012).
• 30) Arnold SF. Design of experiments with MINITAB. The American Statistician, 60, 205 (2006).
• 31) Lifshitz I, Slyozov V. The kinetics of precipitation from supersaturated solid solutions. J. Phys. Chem. Solids, 19, 35–50 (1961).