Development and evaluation of self–nanoemulsifying drug delivery system of rhubarb anthraquinones

Accepted Manuscript Development and evaluation of self–nanoemulsifying drug delivery system of rhubarb anthraquinones Jincheng Li, Yanbin Shi, Yuan Ren, Zhaotong Cong, Guotai Wu, Nana Chen, Xiaoning Zhao, Li Li PII:

S1773-2247(17)30046-1

DOI:

10.1016/j.jddst.2017.04.002

Reference:

JDDST 338

To appear in:

Journal of Drug Delivery Science and Technology

Received Date: 17 January 2017 Revised Date:

16 February 2017

Accepted Date: 2 April 2017

Please cite this article as: J. Li, Y. Shi, Y. Ren, Z. Cong, G. Wu, N. Chen, X. Zhao, L. Li, Development and evaluation of self–nanoemulsifying drug delivery system of rhubarb anthraquinones, Journal of Drug Delivery Science and Technology (2017), doi: 10.1016/j.jddst.2017.04.002. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Graphical abstract The optimized rhubarb anthraquinones loaded nanoemulsion can obviously enhance solubilization of rhubarb anthraquinones in distilled water. Pharmacokinetics in rats showed that the Cmax and AUC of rhubarb anthraquinones loaded nanoemulsion were enhanced

Morphology of RhA–NE

AC C

EP

TE D

Pseudo–ternary phase diagram

M AN U

SC

RI PT

comparing to rhubarb anthraquinones suspension.

In vivo concentration–time curve

In vitro drug release

ACCEPTED MANUSCRIPT

Development and evaluation of self–nanoemulsifying drug delivery

1

system of rhubarb anthraquinones

3

Jincheng Li a, Yanbin Shi

a,*

, Yuan Ren b, Zhaotong Cong a, Guotai Wu b, Nana Chen a,

4

Xiaoning Zhao a, Li Li a

5

a

6

China

7

b

8

University of Chinese Medicine, Lanzhou 730000, China

RI PT

2

School of Pharmacy, Lanzhou University, 199 Donggang West Road, Lanzhou 730000,

M AN U

9

SC

Key Laboratory of TCM Pharmacology and Toxicology of Gansu Province, Gansu

10

11

TE D

12

13

EP

14

AC C

15

16

17

18 *

Address correspondence to Yanbin Shi, School of Pharmacy, Lanzhou University, 199 Donggang

West Road, Lanzhou 730000, P. R. China. Tel. 86-931-8915685; Fax. 86-931-8915686; E-mail: [email protected]

1

ACCEPTED MANUSCRIPT

Abstract

20

Rhubarb anthraquinones (RhA) as a group were formulated in nanoemulsion based system

21

with the aim of improving its solubility and oral bioavailability. RhA loaded nanoemulsion

22

(RhA–NE) was prepared using spontaneous nanoemulsification method. Solubility of RhA in

23

oils, surfactants and co–surfactants was determined to select nanoemulsion components.

24

Surfactants and co–surfactants were screened for their ability to emulsify selected oily phase.

25

Pseudo–ternary phase diagrams were constructed to identify area of nanoemulsification. A

26

four–factor–five–level central composite design was carried out to attain optimal formulation.

27

The optimized formulations of RhA–NE consisted of capryol 90/ethyl oleate containing RhA

28

as oil phase, cremophor RH 40 as surfactant, transcutol HP as co–surfactant, oleic acid as

29

stablizer and distilled water as water phase. The RhA–NE was characterized by dynamic light

30

scanning, transmission electron microscope, solubilizing capacity, encapsulation efficiency,

31

drug loading and stability. The optimized RhA–NE can obviously enhance solubilization of

32

RhA in distilled water. The appearance and RhA contents were basically unchanged after 60

33

days storage at room temperature in brown bottle. More than 50% RhA released from

34

nanoemulsion within 240 h in vitro. Pharmacokinetics in rats showed that the Cmax and AUC

35

of RhA–NE were enhanced compared to that of RhA suspension.

36

Keywords: Rhubarb anthraquinones; Spontaneous nanoemulsification; Central composite

37

design; Dissolution profile; Pharmacokineitics

38

Chemical compounds studied in this article:

39

aloe–emodin (PubChem CID: 10207); rhein (PubChem CID: 10168); emodin (PubChem CID:

40

3220); chrysophanol (PubChem CID: 10208); physcion (PubChem CID: 10639)

41

1. Introduction

AC C

EP

TE D

M AN U

SC

RI PT

19

42

Rhubarb, officially recorded in Chinese, European and Japanese Pharmacopoeia, has been

43

used for thousands of years in China [1]. Among the extracts obtained from the dried rhizome

44

and root of Chinese official rhubarbs, anthraquinone derivatives including emodin,

45

aloe–emodin, rhein, physcion, chrysophanol and their glucosides are commonly accepted as

46

the main active components [2]. Many studies showed that these anthraquinones had many 2

ACCEPTED MANUSCRIPT 47

biological and pharmacological properties, for example, anti–bacterial [3,4], anti–fungal [5],

48

anti–viral [6], anti–oxidant [7–9], neuroprotective [10], anti–cancer [11,12], laxative [13–15],

49

hepatoprotective [16], anti–angiogenic [17], anti–inflammatory [18–20] and so on. Pharmacokinetic analysis indicated that the anthraquinones mainly presented as

51

glucuronides/sulfates in serum, intestine and liver, whereas free forms of most anthraquinones

52

were predominant in kidney and liver. In brain, neither free forms nor conjugated metabolites

53

have been detected [21–23]. Glucuronidation generally results in poor gastrointestinal

54

absorption. Apart from the problem of low bioavailability of free anthraquinones, poor water

55

solubility for rhubarb anthraquinones remains a major obstacle to their development and

56

clinical application.

SC

RI PT

50

Recently, various nanonization strategies have been made and developed to improve drug

58

solubility and bioavailability, such as, drug nanocrystals, nanoemulsions, polymeric micelles,

59

cyclodextrins, melt extrusion and liposomes [24–26]. Nanoemulsions have uniform and

60

extremely small droplet sizes, typically in the range of 10 – 100 nm. In addition, high kinetic

61

stability, low viscosity and optical transparency make them very attractive for many industrial

62

applications [27]. Nanosized drug particles can maximize absorption and hence

63

bioavailability of poorly water soluble drug candidates [24, 28]. In self–nanoemulsifying drug

64

delivery systems (SNEDDs), nanoemulsion can be spontaneously formed following oily

65

phase being mixed with surfactant/co–surfactant mixture and water at an optimal proportion.

66

Pseudo–ternary is useful for the screen of surfactant, co–surfactant and Smix ratio.

67

Nanoemulsion composition could affect the physicochemical characteristics of the

68

nanoemulsion, such as mean droplet size (MDS), zeta potential (ZP) and polydispersity index

69

(PDI). Central composite design–response surface methodology (CCD–RSM) is a useful tool

70

to optimize the parameters of a rhubarb anthraquinone loaded nanoemulsion and contributes

71

to understand the relationship between independent variables and response variables.

AC C

EP

TE D

M AN U

57

72

The main objectives of this work were: (i) to screen the components including oil,

73

surfactant, co–surfactant and stabilizer for rhubarb anthraquinones loaded nanoemulsion

74

(RhA–NE) formulation; (ii) to study the effect of processing variables such as drug content,

75

oil, surfactant and stabilizer on MDS, PDI and ZP; (iii) to analyze in vitro release properties

76

and pharmacokinetics in rats to validate advantages of the optimized RhA–NE. 3

ACCEPTED MANUSCRIPT 77

2. Materials and methods

78

2.1. Materials The dried radix et rhizoma of Rheum palmatum L., produced in Lixian county in Gansu

80

province of China, was purchased from the Lixian Pharmaceutical Company and identified by

81

professor Yongjian Yang, School of Pharmacy, Lanzhou University. A voucher specimen (No.

82

856002) was deposited in the Institute of Pharmacognosy. Chemical reference substances of

83

aloe–emodin, rhein, emodin, chrysophanol and physcion were purchased from the National

84

Institute for the Control of Pharmaceutical and Biological Products (Beijing, China), and their

85

chemical structures were shown in Fig. 1. Capryol 90 and transcutol HP were kindly donated

86

by Gattefossé (Shanghai, China). Cremophor RH 40 and solutol HS 15 were purchased from

87

Beijing Fengli Jingqiu Pharmaceutical Co., Ltd. (Beijing, China). Methanol of HPLC grade

88

was purchased from Shandong Yuwang Industry Company (Jinan, China). All other chemicals

89

and reagents of analytical grade were bought from local commercial company (Lanzhou,

90

China).

91

2.2. Preparation and quantification of the rhubarb extract

TE D

M AN U

SC

RI PT

79

92

In our previous study, a combined procedure for extraction and purification of free

93

anthraquinones from rhubarb has been reported [29]. Briefly, the mixture of rhubarb powder

94

and 20% sulfuric acid solution were heated in water bath at 70

95

distilled water until the penetrating water approximately displayed a central pH. Hydrolyzed

96

rhubarb powder was dried and ultrasonically extracted with 90% ethanol. The extraction

97

solution was rotary evaporated to recycle ethanol, and the aqueous solution was extracted

98

triply with chloroform. The chloroform layer was undergone alkali–solution and

99

acid–isolation. The sediment was dissolved in methanol, and the soluble part was dried to

100

AC C

EP

, filtered and washed with

obtain brown rhubarb anthraquinones.

101

The content of total anthraquinones together with five compounds were analyzed by high

102

performance liquid chromatography (HPLC) method. The HPLC system used was Agilent

103

technology

104

automatic–sampler. Chromatographic separation was achieved on a Diamonsil column (250 ×

1260

series

(Agilent

technologies,

4

INC.,

USA)

equipped

with

an

ACCEPTED MANUSCRIPT 105

4.6 mm, 5 µm, C18, Dikma) at 40

106

phosphoric acid (88:12, v/v) with an isocratic elution. The flow rate was adjusted to 1.0

107

mL/min. The raw data were acquired at 254 nm and processed with an OpenLAB Software.

108

Calibration curves were generated from a series of concentrations of mixed standard solutions

109

containing five reference substances of aloe–emodin, rhein, emodin, chrysophanol and

110

physcion.

111

2.3. Screening of components

112

2.3.1. Selection of oily phase

SC

RI PT

. The mobile phase consisted of methanol–0.1%

The saturation solubility of RhA in oils (isopropyl myristate, isopropyl palmitate, capric

114

triglyceride, capryol 90, castor oil, ethyl oleate and mixed oils) was determined by using

115

shake flask method. RhA′s saturation solubility in 15% (w/w) surfactant (labrasol, cremophor

116

RH 40, solutol HS 15, tween 20, tween 80, gelucire 44/14 and poloxamer 188) solution and

117

co-surfactants (transcutol HP, 1, 2–propylene glycol, glycerol, n–butanol, PEG 400, PEG 600

118

and ethanol) were also determined for further reference. Briefly, each kind of oils, aqueous

119

surfactant solutions and co–surfactants of equal mass were transferred to glass tubes and then

120

excess amount of RhA was added, respectively. The tubes were vortexed at 25 ℃ for 72 h to

121

achieve equilibrium. After attaining equilibrium each tube was centrifuged at 14000 rpm for

122

20 min. The supernatant was suitably diluted with methanol and anthraquinones′ solubility

123

was determinied by a validated HPLC method.

124

2.3.2. Screening of surfactants

AC C

EP

TE D

M AN U

113

125

Emulsification ability of various surfactants was screened by two kinds of methods. One

126

is turbidimetric method. In brief, equal amounts of surfactant and the selected oily phase were

127

gently vortexed and heated around 45 – 60 ℃. 50 mg of isotropic mixture was subsequently

128

diluted with distilled water to 50 mL to obtain emulsion. The resulting emulsions were

129

visually observed for the relative turbidity, and their transmittance was assessed at 549.2 nm

130

on UV–visible spectrophotometer (Shimadzu, Japan) using double distilled water as blank.

131

Another is oil titration method. Briefly, the selected oily phase was intermittently added to

132

1.0 mL of 15% (w/w) surfactant solution under vigorous vortex until a monophasic clear 5

ACCEPTED MANUSCRIPT 133

solution became opaque and recorded of the added oils was recorded. The monophasic clear

134

solution with the maximum amount of oil added was tested for their cloud point in water bath.

135

2.3.3. Selection of co–surfactants and the surfactant and co–surfactant (Smix) ratio Co–surfactants were separately combined with the screened surfactant at a fixed mass

137

ratio of 1:1. For the construction of each phase diagram, sixteen combinations of oil and the

138

above–mentioned mixture were to define the boundaries of phase precisely formed in the

139

phase diagrams. Pseudo–ternary phase diagrams of oil, Smix and aqueous phase were

140

developed using aqueous titration method. The area of oil–in–water (o/w) nanoemulsion

141

region was adopted as the assessment criteria for the evaluation of co-surfactants for

142

emulsifying ability, and the maximal area of o/w nanoemulsion region indicated the best

143

choice.

M AN U

SC

RI PT

136

The screened Smix were combined with various mass ratios of 3:1 to 1:3 (w/w), and then

145

the selected oil and Smix of different ratios were mixed with mass ratios of 1:9 to 9:1 (w/w).

146

Aqueous titration method was carried out for construction of pseudo-ternary phase diagrams

147

for choosing the optimal Smix ratio.

148

2.4. Thermodynamic stability and dispersibility

TE D

144

In order to avoid inclusion of metastable systems, thermodynamic stability studies like

150

centrifugation, heating cooling cycle and freeze–haw cycle were performed. (a)

151

Centrifugation: nanoemulsion formulations were centrifuged at 12000 rpm for 30 min. Those

152

formulations that did not show any phase separation were taken for the heating cooling cycle

153

test. (b) Heating cooling cycle: Six cycles between refrigerator temperature (4

154

with storage at each temperature for not less than 48 h were investaigated for their stability.

155

The stable formulations were subjected to freeze–thaw cycle test. (c) Freeze–thaw cycle:

156

Three freeze–thaw cycles between -20

157

AC C

EP

149

and +25

) and 45

were finally done.

The efficiency of dispersibility was assessed using a dissolution apparatus. 1 mL of each

158

nanoemulsion was added to 500 mL of water at 37

159

paddle rotating at 50 rpm provided gentle agitation. The in vitro performance of the

160

formulations was visually assessed using the grading system reported by Shafiq [30]. 6

. A standard stainless steel dissolution

ACCEPTED MANUSCRIPT 161

2.5. Effect of stabilizer and pH on nanoemulsion Sodium oleate, sodium citrate, cholic acid, oleic acid, sorbitan monolaurate (span 20),

163

sorbitan monopalmitate (span 40) and pH modifiers were conventionally employed as a

164

stabilizer to improve ZP values. Stabilizers content was preliminarily set as 1% (w/w).

165

Sodium hydroxide solution (0.1 mol/L) and hydrochloric acid solution (0.1 mol/L) were used

166

to adjust pH of nanoemulsion formulation to 5.0, 7.0, 9.0 measured by a basic pH meter

167

(PB-10, Sartorius, Germany). The MDS, PDI and ZP were analyzed by Malvern Zetasizer

168

Nano ZS (Model ZEN3600, Malvern Instruments Ltd, UK), and droplet size and PDI were

169

analyzed at a wavelength of 633 nm and a scattering angle of 173° at 25

170

2.6. Nanoemulsion preparation

M AN U

SC

RI PT

162

.

171

RhA–NE was formulated using capryol 90/ethyl oleate (3:1, w/w) served as dispersed

172

phase, deionized water as the continuous phase, cremophor RH 40 as a surfactant, transcutol

173

HP as a co–surfactant, and oleic acid as a stabilizer. Nanoemulsion was prepared by mixing

174

cremophor RH 40/transcutol HP (1:1, w/w) and oleic acid at 45

175

ingredients completely, and then homogenized with oil phase containing rhubarb

176

anthraquinones at room temperature, followed by dropwise addition of water under

177

continuous stirring to form transparent and easily flowable o/w nanoemulsion.

178

2.7. Optimization design

EP

TE D

to dissolve all the

The optimization of nanoemulsion was conducted with application of central composite

180

design (CCD). Preliminary experiments indicated that the variables of drug content (4.5 – 6.5

181

mg/mL in oil, A), oil content (15 – 25%, w/w, B), mixed emulsifier content (35 – 45%, w/w,

182

C) and stabilizer content (0.75 – 1.75%, w/w, D) played vital roles during optimization of

183

RhA–NE. Thus, A four–factor CCD was employed to systemically investigate the main

184

effects, interaction effects and quadratic effects of the variables on three responses: ZP, MDS,

185

and PDI of RhA–NE. A total of 29 experimental runs involving factorial points, axial points

186

and replicates of center points were generated by using Design-Expert version 8.0 software

187

(Stat–Ease, Minneapolis, USA) at five levels for each variable, and the experiments were

188

conducted in random to minimize the influence of extraneous factors.

AC C

179

7

ACCEPTED MANUSCRIPT 189

2.8. Characterization of the optimized RhA–NE

190

2.8.1. General characteristics of RhA–NE Prior to MDS, PDI and ZP measurement, nanoemulsion samples were necessarily to be

192

diluted in 1:100 (v/v) ratios with deionized water until the appropriate concentrations of

193

droplets were obtained in order to avoid multi–scattering effect. All the measurements were

194

repeated three times. A digital display viscometer (NDJ–8S (N), Shanghai, China) equipped

195

with different size of rotors was used to automatically measured the viscosity. UV

196

spectrophotometer (UV–2550, Shimadzu, Japan) was applied to analyze the T% of RhA–NE,

197

and the λmax was set at 700 nm. Morphology of the nanoemulsion was observed using

198

transmission electron microscopy (Tecnai G2 TF20, FEI, USA).

199

2.8.2. Solubilizing capacity

M AN U

SC

RI PT

191

Solubilization of RhA in nanoemulsion was compared with RhA in sodium dodecyl

201

sulfate (SDS) aqueous solution (0.01 mol/L) and distilled water, respectively. Briefly, excess

202

amount of RhA was individually added to 2.0 g blank nanoemulsion, 0.01 mol/L SDS

203

aqueous solution and distilled water, followed by continually magnetic stirring for 24 h at

204

room temperature. The obtained samples were centrifuged at 14000 rpm for 20 min, and the

205

supernatant was diluted 100 times with methanol and analyzed by HPLC method.

206

2.8.3. Encapsulation efficiency and drug loading

EP

TE D

200

1 mL of RhA–NE in 3500 Da dialysis tubing (Solarbio, China) against 100 mL distilled

208

water at room temperature for 72 h was dialyzed. The dialyzed formulation was removed and

209

free anthraquinones were determined by HPLC equipped with DAD detector. Encapsulation

210

efficiency (EE) and drug loading (DL) were respectively calculated by EE = (Wt – Wf)/ Wt ×

211

100% and DL = (Wt – Wf)/ Wc × 100%, therein, Wt refers to the total amount of RhA in

212

RhA–NE, Wf was the amount of free RhA, and Wc was that of total ingredients in RhA–NE.

213

2.8.4. Stability of RhA–NE

AC C

207

214

Stability was preliminarily investigated by appearance observation with coalescence,

215

phase separation and demulsification after high–speed centrifugation at 12,000 rpm for 30 8

ACCEPTED MANUSCRIPT min. Three batches of RhA–NE were stored in the sealed colorless and/or brown bottles, and

217

then carried out on high–temperature test (45

), high–light exposure test (4500 ± 500 lx) and

218

long term storage at the temperature of 25

in drug stability tester (Type WD–A, Tianjin

219

pharmacopoeia standard instrument Co., LTD, China). For temperature and light stress tests,

220

the MDS, PDI and free anthraquinones content were assessed on the 0, 5th and 10th days,

221

respectively. While for long–term test, they were separately assessed on the 10th, 30th and 60th

222

days.

223

2.8.5. In vitro drug release

The in vitro release of free anthraquinones from RhA–NE was performed by a dialysis

225

method with the mixture of phosphate buffer (PBS, pH 7.0) and 2.0% (w/v) SDS as

226

dissolution medium. The dialysis bags (Solarbio, China) with a molecular weight cut-off of

227

3500 Da were soaked in the boiling water for 30 min before use. 1 mL of freshly prepared

228

RhA-NE was transferred into the dialysis bags and tightly sealed. The bags were then placed

229

in conical flasks with 100 mL of release medium and horizontally shaken at 37

230

100 rpm (IS–RDD3 thermostatic oscillator, Crystal Technology & Industries, Inc., USA). At

231

predetermined intervals, 1 mL of sample was withdrawn for HPLC analysis, and the same

232

volume of blank dissolution medium was supplemented. Meanwhile, the release of RhA

233

suspension made by dispersing 1.0 mg RhA into 1.0 mL mixed solution of 0.5% CMC–Na

234

and 0.2% tween 80 mixed solution was performed as control. Release profiles were analyzed

235

by applying four different mathematical models, which were depicted as follows:

236

Zero order kinetic equation: F(t) = k0t

(1)

237

First order kinetic equation: F(t) = 1－exp(－k1t)

(2)

238

Higuchi equation: F(t) = kHt1/2

(3)

239

Ritger–Peppas equation: F(t) = kRtn

(4)

240

therein, k0, k1, kH and kR are all the dissolution constant of the corresponding model, F(t) is

241

the cumulative percentage of drug released at time t [31].

242

2.9. Pharmacokinetics in rats

at a rate of

EP

TE D

M AN U

224

AC C

SC

RI PT

216

9

ACCEPTED MANUSCRIPT 243

2.9.1. Animal experiment Wistar rats were randomly and equally divided into RhA–NE group and RhA suspension

245

group. RhA–NE and RhA suspension were orally administrated by gavage with a single dose

246

of 15 mL/kg (equivalent to 15 mg/kg). Every three rats were punctured for collecting blood

247

samples into heparinized EP tubes at predetermined intervals, followed by centrifuging whole

248

blood samples at 12000 rpm. The resulting plasma layer was collected and stored in EP tube

249

at –20 ℃ until analysis.

250

2.9.2. Sample extraction

SC

RI PT

244

100 µL of 1,8–dihydroxy anthraquinone solution (0.128 µg/mL) as internal standard (IS)

252

and 200 µL of methanol as well as 200 µL 2 mol/L hydrochloric acid were added into 1.0 mL

253

plasma sample. The mixture was thoroughly vortex-mixed, followed by adding 4.0 mL of

254

ethyl acetate, mixed and centrifuged. The supernatant was collected and dried at 40 ℃ with a

255

rotary evaporator under reduced pressure. The residue was redissolved with 200 µL methanol

256

for HPLC analysis.

257

2.9.3. HPLC–FLD Analysis

TE D

M AN U

251

An Agilent technology 1260 series HPLC system equipped with YMC–Pack–ODS–C18

259

column (150 × 4.6 mm, 5 µm) was applied to separate all samples. The mobile phase was

260

selected as methanol–0.1% phosphoric acid (75:25, v/v) with isocratic elution, and the flow

261

rate was adjusted to 0.8 mL/min. Fluorescence detection (FLD) was employed with an

262

excitation wavelength of 435 nm and an emission wavelength of 515 nm. The column

263

temperature was maintained at 40 ℃. The injection volume of all samples was 20 µL.

264

2.10. Statistical analysis

AC C

EP

258

265

The software of Microsoft Excel version 2007 (Microsoft Co., USA), Design–Expert

266

version 8.0 (Stat–Ease, Minneapolis, USA) and Origin Pro version 8.0 (Origin Lab Co., USA)

267

were used for statistical analysis. The pharmacokinetic parameters and the compartment

268

models were analyzed by pharmacokinetic program DAS 3.2.7 edited by Drug and Statistics,

269

Mathematical Pharmacology Professional Committee of China. Numerical data are expressed 10

ACCEPTED MANUSCRIPT as means ± SD. The significance of differences was examined using the ANOVA method.

271

Results with p < 0.05 were considered to be significant.

272

3. Results and discussion

273

3.1. Quantitation of rhubarb anthraquinones

RI PT

270

The weight percentage (w/w, %) of the RhA including aloe emodin, rhein, emodin,

275

chrysophanol and physcion in rhubarb extract was separately listed in Table 1. The total peak

276

area of the RhA was up to 98.0% of that of all the peaks.

277

3.2. Formulation of nanoemulsion

278

3.2.1. Selection of excipients

M AN U

SC

274

In order to attain the maximum solubility of the multi-components (RhA), oil from

280

different classes like long chain fatty acid ester (isopropyl myristate, isopropyl palmitate),

281

long chain triglyceride (castor oil), long chain unsaturated fatty acid ester (ethyl oleate),

282

medium chain triglyceride (capric triglyceride), synthetic medium fatty acid ester (capryol 90),

283

and the mixed oils were chosen for the solubility tests. Higher is the solubility of the drug in

284

the oil phase, lower is the volume of oil required to dissolve a single dose of drug. This, in

285

turn, would lead to lower consumption of surfactants and co-surfactants for the preparation of

286

nanoemulsion. RhA showed a better solubility from high to low level in capryol 90/ethyl

287

oleate (3:1, w/w), capryol 90/capric triglyceride (GTCC) (2:1, w/w) and capryol 90. The

288

solubility of RhA in capryol 90/ethyl oleate (3:1, w/w) and capryol 90/GTCC (2:1, w/w) was

289

5.816 ± 0.233 mg and 5.912 ± 0.090 mg, respectively. Ethyl oleate and GTCC were both

290

reported to be stable and have low toxicity, but for the concern of economic, capryol 90/ethyl

291

oleate (3:1, w/w) was preferably selected for loading RhA.

AC C

EP

TE D

279

292

Non-ionic surfactants were employed since an important concern with respect to the

293

selection of surfactants was their toxicity. They have been reported to be the least toxicity, less

294

affected by pH and ionic strength changes compared to those ionic surfactants. Generally, a

295

surfactant should have HLB > 10 to form an o/w emulsion [28]. Therefore, tween80, tween 20,

296

labrasol, gelucire 44/14, poloxamer 188, cremophor RH 40 and solutol HS 15 having HLB 11

ACCEPTED MANUSCRIPT value more than 10 were respectively applied to select suitable surfactant. According to the

298

result of the solubility test and turbidimetric method, the solution added with cremophor RH

299

40 and solutol HS 15 had better ability to dissolve RhA (graph presented in Supplementary

300

material), and the resulting solution reached the goal of obtaining higher light transmittance

301

(data presented in Supplementary material). Based on oil titration method, tween 20 could

302

emulsify the largest amount of oil and the obtained solution had high cloud point, hinting

303

good thermodynamic stability. After comprehensive consideration, cremophor RH 40, solutol

304

HS 15 and tween 20 were chosen as the surfactants for further study.

RI PT

297

Co–surfactant can increase interfacial fluidity by penetrating into the surfactant film,

306

creating void space among surfactant molecules, reducing the interfacial tension, changing the

307

curvature of the oil–water interface, which resulting in an easily formation of nanoemulsion

308

[32]. It was found that changing the surfactant type in systems did affect nanoemulsion region

309

(graph presented in Supplmentary material). The o/w nanoemulsion area obtained with

310

cremophor RH 40 was much larger than that constructed with solutol HS 15 and tween 20

311

with the same co–surfactant used except for PEG 600. Meanwhile, cremophor RH 40 had the

312

highest solubility of five anthraquinones. Therefore, cremophor RH 40 was finally selected as

313

surfactant.

TE D

M AN U

SC

305

Cremophor RH 40 and different co–surfactants were submitted to investigate their kinetic

315

stability, thermodynamic stability and dispersibility (data presented in Supplementary

316

material). As we found that transcutol HP performed the most stable effect on the

317

nanoemulsion system. The dependence of kinetic and thermodynamic stability on

318

co–surfactants type may be attributed to differences in their physicochemical properties, such

319

as melt point, HLB and viscosity. As the viscosity of PEG, propylene glycol and glycerol was

320

higher than that of transcutol HP, n–butanol and ethanol, the diffusivity of PEG, propylene

321

glycol and glycerol reduced, and the effect of Smix on lowering surface tension between

322

oil/water interface decreased [33, 34]. In addition, the higher the viscosity of Smix was, the

323

easier it is affected by temperature changes. When ethanol or n–butanol was applied as

324

co–surfactants, the nanoemulsion formulation exhibited thermal instability under H/C cycle,

325

which may be due to lower boiling point of ethanol and the long hydrocarbon chain structure

AC C

EP

314

12

ACCEPTED MANUSCRIPT 326

of n–butanol. Besides, transcutol HP showed the highest solubility of five anthraquinones

327

(graph presented in Supplementary material). After overall consideration, transcutol HP was

328

preferably chosen as co–surfactant. The Smix ratio between cremophor RH 40 and transcutol HP was vital to the ability to

330

form o/w nanoemulsion. It can be seen in Fig. 2, based on percentage of nanoemulsion area

331

accounted for the whole triangle area, a descending order of nanoemulsion area was Smix 1:1

332

(25.44%) > Smix 1:2 (24.15%) > Smix 2:1 (20.99%) > Smix 1:3 (20.17%) > Smix 3:1

333

(19.19%) > Smix 4:1 (15.16%) > Smix 1:0 (9.37%). When surfactant was used alone (Smix

334

ratio 1:0), large microemulsion gel area (blue region) was obtained while small o/w

335

nanoemulsion region was found towards aqueous rich apex and Smix rich apex. In this case,

336

the maximum concentration of oil emulsified was up to 30% (w/w) by using 20% (w/w) of

337

Smix. As the ratio of Smix was changed to 1:1, the microemulsion gel area was inverted to

338

o/w nanoemulsion area, and the maximum oil emulsified was 25% (w/w) using 38% (w/w) of

339

Smix. This may be attributed to the fact that the addition of co–surfactant may lead to greater

340

penetration of the oil phase in the hydrophobic region of the surfactant monomers thereby

341

further decreasing the interfacial tension, leading to increase in the fluidity of the interface,

342

thus facilitating nanoemulsion formation. However, if Smix decreased to 1:3 ratio,

343

nanoemulsion region reduced, maximally making 20% (w/w) oil soluble at 47% (w/w) Smix.

344

For reach the goal of obtaining kinetic, thermodynamic stable and dispersible nanoemulsion

345

system, further study was continued.

EP

TE D

M AN U

SC

RI PT

329

Hundreds of formulations can be prepared from nanoemulsion region through

347

pseudo–ternary phase diagrams. Generally, for certain concentration of oil selected, minimum

348

concentration of Smix was adopted from the phase diagram for minimal toxicity. There was

349

no significant difference in the phase behavior and nanoemulsion area of phase diagrams as

350

RhA was encapsulated in the formulations. When surfactant was used alone (Smix ratio 1:0),

351

H/C and Freeze–thaw unstable phenomenon was significantly noticed. As the ratio of Smix

352

gradually decreased to 1:1, the nanoemulsion system changed to be easily dispersed and

353

thermodynamic stable (data presented in Supplementary material). This may be explained that

354

the addition of transcutol HP (HLB = 4.2) increased in the fluidity of the interface, and also

AC C

346

13

ACCEPTED MANUSCRIPT transformed the HLB of the Smix to a desirable value, thus increasing the entropy of the

356

system. When Smix further decreased to 1:3, thermodynamic unstable phenomenon was

357

noticeably found at 45 ℃. Taken the nanoemulsion area and the stability into consideration,

358

the Smix ratio of 1:1 was selected for the RhA–NE system. At this stage, the preferable mass

359

percentage of oil, Smix and water for nanoemulsion formulation was set as 20%, 40%, 40%,

360

respectively.

361

3.2.2. Effect of stabilizer and pH on nanoemulsion

RI PT

355

Stabilizers are needed to stabilize the nanoparticles against inter-particle forces and

363

prevent them from aggregating. Generally, absolute values higher than 30 mV indicate good

364

stability, while values above 60 mV indicate excellent long–term stability [35, 36]. In Fig. 3A,

365

all stabilizers except for cholic acid can enhance absolute value of ZP with a descending order

366

of sodium oleate > sodium citrate > oleic acid > span 40 > span 20. Sodium oleate not only

367

reduced the MDS lower than 20 nm, but also resulted in the ZP value lower than –30 mV.

368

Whereas, the PDI increased to 0.261 ± 0.025 and pH significantly rose up to 9.89 ± 0.08. For

369

sodium citrate, MDS and ZP values were 24.12 ± 0.93 nm and –17.45 ± 1.30 mV, and PDI

370

and pH increased to 0.135 ± 0.008 and 7.16 ± 0.12, respectively. Thus, it can be inferred that

371

ZP value of the system was mainly related to pH change. Subsequently, the influence of pH

372

on nanoemulsion system was further explored.

EP

TE D

M AN U

SC

362

In Fig. 3B, significant difference in the ZP of the nanoemulsion with change in pH from

374

5.0 to 9.0 was found. As pH value increased, ZP value became more and more negative. This

375

descending trend was in close agreement with several studies that non–ionic elmusifiers

376

reduced the absolute magnitude of ZP of suspension in a wide range of pH [37]. Relatively

377

stable nanoemulsions were formed at pH 6.0 and 7.0, while extensive particle

378

growth/aggregation occurred at lower or higher pH values, which was attributed to either

379

chemical (hydrolysis) or physical (electrical charge) effects [38]. In the process of stability

380

test, high pH and light performed double effect on free anthraquinones, aggravating the

381

decompose reaction. Based on comprehensive consideration, 1.0% oleic acid was preferably

382

selected as stabilizer and no pH adjustment was done. The pH, MDS, PDI and ZP of the

AC C

373

14

ACCEPTED MANUSCRIPT 383

nanoemulsion system added with stabilizer were 6.09 ± 0.02, 26.08 ± 0.36 nm, 0.115 ± 0.010,

384

–16.0 ± 0.9 mV, respectively.

385

3.3. Design expert The independent variables and their coded levels are showed in Table 2. The results of 29

387

experimental runs are listed in Supplementary material. The significance of the fitting of the

388

quadratic model was assessed by one way analysis of variance (ANOVA) and the results are

389

listed in Table 3. The proposed second degree polynomial model was fitted to the

390

experimental data to determine the optimum composition of nanoemulsion with lower ZP,

391

MDS and PDI. The selection of model for analyzing the responses was done based on

392

sequential model sum of squares, lack of fit and model summary statistics. The Prob > F value

393

of P < 0.0001, low standard deviation, high R2 and lower predicted residual error sum of

394

square (PRESS) values suggested to select quadratic model for both responses. The Model F

395

values for response ZP, MDS, and PDI were 28.83, 62.08 and 17.91 respectively, indicating

396

the model was significant. ANOVA identifies the significant factors that affect the responses.

397

A, B, C, D, BC, BD, A2, B2, C2 were significant model terms for ZP. Whereas, A, B, C, D, BC,

398

BD, B2 are significant model terms for MDS, and A, B, D, BC, A2, B2, D2 are significant

399

model terms for PDI. Lack of fit “Prob> F” value for MDS was 0.1715, implying lack of fit

400

was not significant relative to the pure error. Lack of fit “Prob> F” values for ZP and PDI

401

were 0.0236 and 0.0249, indicating that only 2.36% and 2.49% chance that the “Lack of fit”

402

F–value could occur due to noise. The value of predicted R2 value was close to that of

403

adjusted R2 value, implying that the model predicted all the responses well. The low values of

404

coefficient of variation (CV%) and acceptable adequate precision indicated that this model

405

can be applied to navigate the design space. The final polynomial equations in terms of coded

406

factors for ZP, MDS and PDI are presented as follows, respectively.

407

ZP = －16.06 － 1.10 A － 0.86 B － 1.61 C － 0.84 D － 0.056 AB － 0.11 AC ＋

408

0.13 AD ＋ 0.38 BC ＋ 0.37 BD ＋ 0.044 CD ＋ 0.58 A2 ＋ 0.35 B2 ＋ 0.52 C2 ＋

409

0.097 D2

410

MDS = ＋26.17 ＋ 0.71 A ＋ 2.34 B － 0.77 C ＋ 0.63 D － 0.18 AB ＋ 0.034 AC

AC C

EP

TE D

M AN U

SC

RI PT

386

15

ACCEPTED MANUSCRIPT － 0.17 AD － 0.49 BC ＋ 0.33 BD － 0.062 CD ＋ 0.13 A2 ＋ 0.33 B2 － 0.038 C2

412

＋ 0.042 D2

413

PDI = ＋0.14 ＋ 0.016 A － 6.917×10-3 B － 3.167×10-3 C － 5.250×10-3 D －

414

1.875×10-3 AB － 2.500×10-3 AC － 0.013 BC － 8.750×10-4 BD － 1.250×10-3 CD －

415

4.121×10-3 A2 ＋ 0.012 B2 ＋ 3.129×10-3 C2 － 6.996×10-3 D2

RI PT

411

The relationship between variables was further studied using contour plots. Fig. 4A and

417

Fig. 4B show that ZP decreased (more negative values) with increase in oil, Smix and

418

stabilizer content. Since oils and stabilizer are composed of fatty acid ester and fatty acid

419

respectively, the negative charge might be attributed to carboxylic acid groups. Besides, the

420

charge on the nanoemulsion droplets may be influenced by the charge on the cremophor RH

421

40′s ester group and glycerol polyoxyethylene ether group [37]. Glycols of co–surfactant also

422

contribute to the negative charge of nanoemulsion owning to their hydroxyl groups. Thus

423

interactive items BC and BD influencing ZP were ascribed to their synergistic effect. Fig. 4C

424

and Fig. 4D show that MDS increased at high level of factor B but low level of factor C. At

425

higher content of oil the surfactant molecules are not enough to cover the oil droplets and

426

reduce interfacial tension at o/w interface, finally leading to higher globule size. MDS was

427

found to decrease with increase in surfactant strongly supported this point. Mixtures of

428

surfactants form a film around dispersed droplets and maintain droplet stability by

429

strengthening interfacial film, and the addition of co–surfactant has also been found to

430

decrease the rate of Ostwald ripening in o/w emulsion [39]. Fig. 4E shows interactive item

431

BC on response PDI. As factors B decreased and factor C increased, PDI was increasing, and

432

vice–versa, which implied suitable proportion of oil and Smix was needed for low PDI.

AC C

EP

TE D

M AN U

SC

416

433

The result of 29 experimental runs showed that the levels of factors A, B, C and D which

434

gave ZP in range of –17.9 – –9.3 mV, MDS in range of 22.54 – 32.79 nm and PDI in range of

435

0.102 – 0.212. The model predicted values of responses ZP, MDS and PDI were −16.6 mV,

436

25.14 nm and 0.121 with factors A, B and C and D values of 5.1 mg, 19.5%, 42.5% and

437

1.50%, respectively. Three batches of nanoemulsion were prepared using these values of

438

factors. The actual values of ZP, MDS and PDI were separately found to be −17.6 ± 1.0 mV,

439

25.37 ± 0.55 nm and 0.136 ± 0.012, which were in close agreement to the predicted values. 16

ACCEPTED MANUSCRIPT 440

3.4. Characterization of the optimized RhA–NE

441

3.4.1. General Characteristics Three batches of RhA–NE prepared according to the optimized formula were used to

443

evaluate general characteristics. MDS and PDI of the optimized RhA–NE were 25.37 ± 0.55

444

nm and 0.136 ± 0.012, respectively. The mean ZP value was –17.6 ± 1.0 mV. A mean

445

viscosity of 73.3 ± 1.4 cP was obtained with the size of SP1 rotor, rotation speed of 60 rpm at

446

25

447

were spherical in morphology and the diameter observed by TEM was smaller than 50 nm

448

with narrow size distribution which was close to the range obtained using CCD experiments

449

(Fig. 5).

450

3.4.2. Solubilizing capacity, EE and DL

RI PT

442

M AN U

SC

, and a mean T% of 88.1 ± 1.9% was achieved at a λmax of 700 nm. RhA–NE droplets

The total anthraquinones′ solubility in blank nanoemulsion was about 23 to 84 times

452

higher than that in SDS (0.01 mol/L) and distilled water. Solubility of aloe–emodin, rhein,

453

emodin, chrysophanol and physcion in blank nanoemulsion was increased to 0.445 ± 0.021,

454

0.296 ± 0.028, 0.456 ± 0.030, 1.067 ± 0.097 and 0.400 ± 0.032 mg/mL, respectively.

455

Solubilization significantly improved was due to a combined action of internal oil phase,

456

surfactant and co–surfactant which can be further explained by solubilization theory of

457

nanoemulsion. The entrapment efficiency of RhA in RhA–NE was about 98.091 ± 0.002 %

458

and the drug loading was 0.101 ± 0.002 %.

459

3.4.3. Stability of RhA–NE

AC C

EP

TE D

451

460

The appearance of the optimized RhA–NE was homogeneous and transparent after being

461

undergone a high–speed centrifugation. For stress stability test, the droplet sizes were nearly

462

unchangeable within 10 days, which were around 24.66 – 35.08 nm. For those samples stored

463

in brown bottles, the content of aloe–emodin, rhein, emodin, chrysophanol and physcion for

464

illuminate stress testing lost about 3.5%, 5.6%, 4.8%, 2.7% and 1.5% on the 10th day

465

compared to that of the first day. And for temperature stress testing, they separately lost about

466

6.2%, 8.5%, 6.8%, 3.2% and 1.2% on the 10th day. The RhA content change was less than 5.0%

467

within 60 days at room temperature (HPLC graphs presented in Supplementary material), and 17

ACCEPTED MANUSCRIPT no obvious MDS and PDI value change was determined. However, for those samples stored in

469

the colorless bottles, the color of samples under strong light changed from brown yellow to

470

deep red on the 10th day. This may be caused by chemical change and interaction between

471

drug molecule and ethylene glycol structure of co–surfactant under strong light condition.

472

3.4.4. In vitro drug release

RI PT

468

RhA are all lipophilic chemicals having limiting solubility in aqueous phase. So the first

474

task was to determine what could act as dissolution medium and how much volume it was

475

applied so as to guarantee sink condition. Aloe–emodin, rhein, emodin, chrysophanol and

476

physcion solubilized in dissolution media were found to be 97.50 ± 2.05%, 98.65 ± 0.39%,

477

96.19 ± 1.46%, 95.35 ± 2.57%, 90.31 ± 2.82% of that solubilized in methanol. Therefore, it

478

was certain that 100 mL dissolution media can meet sink condition. From in vitro release

479

curve of aloe-emodin, rhein, emodin, chrysophanol, physcion and total free anthraquinone

480

(graph presented in Supplementary materials), the delayed release of drug from nanoemulsion

481

may be attributed to the process of drug diffusion from dispersed phase to continuous phase

482

and then to dissolution solution, while drug diffused from suspension to dissolution medium

483

immediately. And compared slope of drug release curves with suspension within 72 h, drug in

484

dosage of nanoemulsion did possess sustained–release property, especially for aloe–emodin,

485

thein and emodin. More than 50% of drug was released from RhA–NE at the end of 240 h.

486

The dissolution rate was in accordance with the polarity of five anthraquinones, namely, the

487

greater the polarity was, the faster the dissolution was observed. Interestingly, we found that

488

RhA–NE exhibited a slower dissolution rate compared to RhA suspension, which posed a

489

conflict with the Ostwald Freundlich and Noyes–Whitney equations. This may be caused by

490

the mixing uniformity of RhA–NE and dissolution media, and SDS hindered the contact of

491

anthraquinones in O/W nanoemulsion to dissolution media [40].

AC C

EP

TE D

M AN U

SC

473

492

Akaike′s information criterion (AIC) and correlation of coefficient square (r2) were

493

adopted as a measure of goodness of fit of the experimental data to all the models. As shown

494

in Table 4, Ritger–Peppas model was most fitted to the release data of RhA. To the

495

determination of the exponent (n), the portion of the release curve where Mt/M∞ < 0.6 should

496

only be used. For a sphere nanoemulsion vector, n ≤ 0.43 indicates it observes Fick′s 18

ACCEPTED MANUSCRIPT diffusion, and 0.43 < n < 0.85 follows anomalous transport which mainly includes corrosion

498

and diffusion. While for irregular–shape suspension vector, the critical values are replaced

499

with 0.45 – 0.5 and 0.89 – 1.0, respectively [31]. The mechanisms of aloe–emodin, rhein,

500

emodin, chrysophanol and physcion release from RhA–NE were correspondingly anomalous

501

transport, diffusion–controlled release, anomalous transport, first–order kinetic and

502

anomalous transport. While the mechanisms of five anthraquinones release from suspension

503

were mainly diffusion–controlled release except for chrysophanol. The mechanism of total

504

anthraquinones release from RhA–NE and suspension was individually anomalous transport

505

and Fick′s diffusion. These results were basically in line with our expectations.

SC

3.5. In vivo pharmacokinetics

M AN U

506

RI PT

497

Non–compartment model was applied for calculating concentration–time data, and the

508

results are shown in Table 5. HPLC chromatograms of blank plasma sample, blank plasma

509

sample spiked with RhA and IS, and RhA–NE group′s plasma samples clearly illustrated

510

there was no obvious interference peak observed at the retention time of five free

511

anthraquinones detected (graphs presented in Supplementary material). Concentration–time

512

curves of each anthraquinone and total anthraquinones are shown in Fig. 6. Compared to the

513

RhA suspension, the RhA–NE had a higher solubility in digestive juice, and therefore

514

leading to an increase of Cmax and AUC. The Cmax and AUC0-∞ of total anthraquinones in

515

RhA–NE group were 1.20 and 2.58 times of which in RhA suspension group. The CLz/F of

516

total anthraquinones in RhA-NE group was smaller than that of suspension, and MRT0-∞ was

517

extended 1.90 times than RhA suspension group. The Tmax of rhein, emodin, physcion were

518

all prolonged in RhA–NE group and the t1/2z of five free anthraquinones were all increased.

519

These results can be explained by small droplet size and application of surfactant and

520

co–surfactant, together with stronger adhesion to gastrointestinal mucosa of nanoemulsion.

521

The Vz/F of the rhein, emodin and chrysophanol in RhA–NE group was smaller compared to

522

that in the RhA suspension group, especially for the rhein. This may be caused by

523

metabolism that aloe–emodin and physcion could be converted to rhein and emodin in rats

524

[41], which lead to the higher concentration of rhein and emodin in plasma. The peak

525

concentration of rhein and emodin in RhA-NE in Fig. 6B and Fig 6C was further back up

AC C

EP

TE D

507

19

ACCEPTED MANUSCRIPT this viewpoint. Multiple peaks were observed in plasma concentration–time curves of each

527

anthraquinone in the two dosage forms, which could be explained by enterohepatic

528

circulation which has been speculated for anthraquinones or the metabolites by Wu [41] and

529

Zhang [42]. The Tmax was 0.67 h for RhA suspension and 1.5 h for RhA–NE, implying drug

530

initially released from RhA suspension faster than that from nanoemulsion in rats'

531

gastrointestinal tract. This can be evidenced by the in vitro drug release profile which

532

showed RhA suspension had higher cumulative percentage of drug release than RhA–NE

533

within 72 h. The MRT0-∞ for RhA–NE was 1.90 times larger than that for RhA suspension,

534

which may be attributed to the in vitro sustained release of RhA–NE in the earlier stage

535

compared to RhA suspension.

SC

4. Conclusion

M AN U

536

RI PT

526

RhA (5.1 mg/mL in oil), Capryol 90/ethyl oleate (19.5%, w/w), cremophor RH 40

538

(21.25%, w/w), transcutol HP (21.25%, w/w), oleic acid (1.50%, w/w) and distilled water

539

(36.5%, w/w) was defined to achieve the optimized RhA–NE with ZP, MDS and PDI were

540

–17.6 ± 1.0 mV, 25.37 ± 0.55 nm and 0.136 ± 0.012, respectively. The RhA–NE can

541

effectively increase the hydrosolubility and improve the oral bioavailability of RhA compared

542

to RhA suspension.

543

Declaration

EP

545

There is no conflict of interest. Acknowledgments

AC C

544

TE D

537

546

The authors acknowledge financial support from Key Laboratory of TCM Pharmacology

547

and Toxicology of Gansu Province (ZDSYS–KJ–2013–005), TCM Science and Technology

548

Research Project of Gansu (GZK–2015–3), P.R. China.

549

Supplementary material

550

References

551 552

[1] J.B. Wang, W.J. Kong, H.J. Wang, H.P. Zhao, H.Y. Xiao, C.M. Dai, X.H. Xiao, Y.L. Zhao, C. Jin, L. Zhang, F. Fang, R.S. Li, Toxic effects caused by rhubarb (Rheum palmatum L.) are reversed on 20

ACCEPTED MANUSCRIPT

EP

TE D

M AN U

SC

RI PT

immature and aged rats, J. Ethnopharmacol. 134 (2011) 216–220. [2] H.X. Zhang, M.C. Liu, Separation procedures for the pharmacologically active components of rhubarb, J. Chromatogra. B 812 (2004) 175–181. [3] C.X. Lu, H.X. Wang, W.P. Lv, P. Xu, J. Zhu, J. Xie, B. Liu, Z.X. Lou, Antibacterial properties of anthraquinones extracted from rhubarb against Aeromonas hydrophila, Fish Sci. 77 (2011) 375–384. [4] P. Raudsepp, D. Anton, A. Roasto, K. Meremäe, P. Pedastsaar, M. Mäesaar, The antioxidative and antimicrobial properties of the blue honeysuckle (Lonicera caerulea L.), Siberian rhubarb (Rheum rhaponticum L.) and some other plants, compared to ascorbic acid and sodium nitrite, Food Control 31 (2013) 129–135. [5] S.K. Agarwal, S.S. Singh, S. Verma, S. Kumar, Antifungal activity of anthraquinone derivatives from Rheum emodi, J. Ethnopharmacol. 72 (2000) 43–46. [6] H.R. Xiong, J. Luo, W. Hou, H. Xiao, Z.Q. Yang, The effect of emodin, an anthraquinone derivative extracted from the roots of Rheum tanguticum, against herpes simplex virus in vitro and in vivo, J. Ethnopharmacol. 133 (2011) 718–723. [7] M. Öztürk, F. Aydoğmuş-Öztürk, M.E. Duru, G. Topçu. Antioxidant activity of stem and root extracts of Rhubarb (Rheum ribes): An edible medicinal plant, Food Chem. 103 (2007) 623–630. [8] H. Matsuda, T. Morikawa, I. Toguchida, J.Y. Park, S. Harima, M. Yoshikawa, Antioxidant Constituents from Rhubarb: Structural Requirements of Stilbenes for the Activity and Structures of Two New Anthraquinone Glucosides, Bioorgan. Med. Chem. 9 (2001) 41–50. [9] J.P.S. Silveira, L.N. Seito, S. Eberlin, G.C. Dieamant, C. Nogueira, M.C.V. Pereda, L.C.D. Stasi, Photoprotective and antioxidant effects of Rhubarb: inhibitory action on tyrosinase and tyrosine kinase activities and TNF-α, IL-1α and α-MSH production in human melanocytes, BMC Complem. Altern. M. 13 (2013) 49–55. [10] T. Liu, H. Jin, Q.R. Sun, J.H. Xu, H.T. Hu, Neuroprotective effects of emodin in rat cortical neurons against β-amyloid-induced neurotoxicity, Brain Res. 1347 (2010) 149–160. [11] K.H. Tsai, H.H. Hsien, L.M. Chen, W.J. Ting, Y.S. Yang, C.H. Kuo, C.H. Tsai, F.J. Tsai, H.J. Tsai, C.Y. Huang, Rhubarb inhibits hepatocellular carcinoma cell metastasis via GSK-3-β activation to enhance protein degradation and attenuate nuclear translocation of β-catenin, Food Chem. 138 (2013) 278–285. [12] CS Shia, G Suresh, YC Hou, YC Lin, PDL Chao, SH Juang, Suppression on metastasis by rhubarb through modulation on MMP-2 and uPA in human A549 lung adenocarcinoma: An ex vivo approach, J. Ethnopharmacol. 133 (2011) 426–433. [13] Y. Qina, J.B. Wang, W.J. Kong, Y.L. Zhao, H.Y. Yang, C.M. Dai, F. Fang, L. Zhang, B.C. Li, C. Jin, X.H. Xiao, The diarrhoeogenic and antidiarrhoeal bidirectional effects of rhubarb and its potential mechanism, J. Ethnopharmacol. 133 (2011) 1096–1102. [14] T.S. Feng, Z.Y. Yuan, R.Q. Yang, S. Zhao, F. Lei, X.Y. Xiao, D.M. Xing, W.H. Wang, Y. Ding, L.J. Du, Purgative components in rhubarbs: Adrenergic receptor inhibitors linked with glucose carriers, Fitoterapia 91 (2013) 236–246. [15] Y.F. Zheng, C.F. Liu, W.F. Lai, Q. Xiang, Z.F. Li, H. Wang, N. Lin, The laxative effect of emodin is attributable to increased aquaporin 3 expression in the colon of mice and HT-29 cells, Fitoterapia 96 (2014) 25–32. [16] X.Y. Xing, Y.L. Zhao, W.J. Kong, J.B. Wang, L. Jia, P. Zhang, D. Yan, Y.W. Zhong, R.S. Li, X.H. Xiao, Investigation of the ″dose-time-response″ relationships of rhubarb on carbon

AC C

553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596

21

ACCEPTED MANUSCRIPT tetrachloride-induced liver injury in rats, J. Ethnopharmacol. 135 (2011) 575–581. [17] Z.H. He, M.F. He, S.C. Ma, P.P.H. But, Anti-angiogenic effects of rhubarb and its anthraquinone derivatives, J. Ethnopharmacol. 121 (2009) 313–317. [18] B.Y. Hu, H. Zhang, X.L. Meng, F. Wang, P. Wang, Aloe-emodin from rhubarb (Rheum rhabarbarum) inhibits lipopolysaccharide-induced inflammatory responses in RAW264.7 macrophages, J. Ethnopharmacol. 153 (2014) 846–853. [19] Z.A. Yang, J.Z. Zheng, Z.D. Wang, Effect of dexamethasone, anisodamine and rhubarb therapy on rats with acute pancreatitis, Journal of Medical Colleges of PLA 24 (2009) 155–160. [20] M.K. Moon, D.G. Kang, J.K. Lee, J.S. Kim, H.S. Lee, Vasodilatory and anti-inflammatory effects of the aqueous extract of rhubarb via a NO-cGMP pathway, Life Sci. 78 (2006) 1550–1557. [21] W. Liu, L. Tang, L. Ye, Z. Cai, B.J. Xia, J.J. Zhang, M. Hu, Z.Q. Liu, Species and Gender Differences Affect the Metabolism of Emodin via Glucuronidation, The AAPS Journal 12 (2010) 424–436. [22] C.S. Shia, S.Y. Tsai, J.C. Lin, M.L. Li, M.H. Ko, P.D.L. Chao, Y.C. Huang, Y.C. Hou, Steady-state pharmacokinetics and tissue distribution of anthraquinones of Rhei Rhizoma in rats, J. Ethnopharmacol. 137 (2011) 1388–1394. [23] S.P. Lin, P.M. Chu, S.Y. Tsai, M.H. Wu, Y.C. Hou, Pharmacokinetics and tissue distribution of resveratrol, emodin and their metabolites after intake of Polygonum cuspidatum in rats, J. Ethnopharmacol. 144 (2012) 671–676. [24] E. Merisko-Liversidge, G.G. Liversidge, Nanosizing for oral and parenteral drug delivery: A perspective on formulating poorly-water soluble compounds using wet media milling technology, Adv. Drug Deliver. Rev. 63 (2011) 427–440. [25] H.B. Chen, C.C. Khemtong, X.L. Yang, X.L. Chang, J.M. Gao, Nanonization strategies for poorly water-soluble drugs, Drug Discov. Today 16 (2016) 354–360. [26] F. Kesisoglou, S. Panmai, Y.H. Wu, Nanosizing – Oral formulation development and biopharmaceutical evaluation, Adv. Drug Deliver. Rev. 59 (2007) 631–644. [27] L.J. Wang, X.F. Li, G.Y. Zhang, J.F. Dong, J. Eastoe, Oil-in-water nanoemulsions for pesticide formulations, J. Colloid. Interf. Sci. 314 (2007) 230–235. [28] V. Bali, M. Ali, J. Ali, Nanocarrier for the enhanced bioavailability of a cardiovascular agent: In vitro, pharmacodynamic, pharmacokinetic and stability assessment, Int. J. Pharm. 403 (2011) 46–56. [29] H.L. Li, L.D. Du, Y.B. Shi, H. Liu, X.Y. Zhang, Y. Li, J.M. Ni, Y. Ren, A combined procedure for extraction and purification of free anthraquinones from Radix et Rhizoma Rhei, Journal of Lanzhou University (Medical Sciences) 40 (2014) 63–70. [30] S. Shafiq, F. Shakeel, S. Talegaonkar, F.J. Ahmad, R.K. Khar, M. Ali, Development and bioavailability assessment of ramipril nanoemulsion formulation, Eur. J. Pharm. Biopharm. 66 (2007) 227–243. [31] P. Costa, J.M.S. Lobo, Modeling and comparison of dissolution profiles, Eur. J. Pharm. Sci. 13 (2001) 123–133.

636 637 638 639 640

[32] D.B. Mahmouda, M.H. Shukr, E.R. Bendas, In vitro and in vivo evaluation of self-nanoemulsifying drug delivery systems of cilostazol for oral and parenteral administration, Int J. Pharm. 476 (2014) 60–69.

AC C

EP

TE D

M AN U

SC

RI PT

597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635

[33] Y.R. Qiu, H. Matsuyama, G.Y. Gao, Y.W. Ou, C. Miao, Effects of diluent molecular weight on the performance of hydrophilic poly(vinyl butyral)/Pluronic F127 blend hollow fiber membrane via 22

ACCEPTED MANUSCRIPT

RI PT

SC

M AN U

TE D

EP

667 668

thermally induced phase separation, J. Membrane SCI. 338 (2009) 128–134. [34] A.H. Saberi, Y. Fang, D.J. McClements, Fabrication of vitamin E-enriched nanoemulsions by spontaneous emulsification: Effect of propylene glycol and ethanol on formation, stability, and properties, Food Res. Int. 54 (2013) 812–820. [35] V. Klang, N. Matsko, A.M. Zimmermann, E. Vojnikovic, C. Valenta, Enhancement of stability and skin permeation by sucrose stearate and cyclodextrins in progesterone nanoemulsions, Int. J. of Pharm. 393 (2010) 152–160. [36] V. Klang, N. Matsko, K. Raupach, N. El-Hagin, C. Valenta, Development of sucrose stearate-based nanoemulsions and optimisation through c-cyclodextrin, Eur. J. Pharm. Biopharm. 79 (2011) 58–67. [37] S.Y. Tang, S. Manickam, T.K. Wei, B. Nashiru, Formulation development and optimization of a novel Cremophore EL-based nanoemulsion using ultrasound cavitation, Ultrason. Sonochem. 19 (2012) 330–345. [38] T.P. Sari, B. Mann, R. Kumar, R.R.B. Singh, R. Sharma, M. Bhardwaj, S. Athira, Preparation and characterization of nanoemulsion encapsulating curcumin, Food Hydrocolloid. 43 (2015) 540–546. [39] S. Sood, K. Jain, K. Gowthamarajan, Optimization of curcumin nanoemulsion for intranasal delivery usingdesign of experiment and its toxicity assessment, Colloids Surf. B: Biointerfaces 113 (2014) 330–337. [40] Y.N. Yi, L.X. Tu, K.L. Hu, W. Wu, J.F. Feng, The construction of puerarin nanocrystals and its pharmacokinetic and in vivo–in vitro correlation (IVIVC) studies on beagle dog, Colloids Surf. B: Biointerfaces 133 (2015) 164–170. [41] W.J. Wu, R. Yan, M.C. Yao, Y. Zhan, Y.T. Wang, Pharmacokinetics of anthraquinones in rat plasma after oral administration of a rhubarb extract, Biomed. Chromatogra. 28 (2014) 564–572. [42] L. Zhang, J.H. Chang, B.Q. Zhang, X.G. Liu, P. Liu, H.F. Xue, L.Y. Liu, Q. Liu, M. Zhu, C.Z. Liu, The pharmacokinetic study on the mechanism of toxicity attenuation of rhubarb total free anthraquinone oral colon-specific drug delivery system, Fitoterapia 104 (2015) 86–96.

AC C

641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666

23

ACCEPTED MANUSCRIPT 669 670

Tables

671 672 673

Table 1 The content of the five anthraquinones ( n = 6, mean ± SD). Rhubarb anthraquinones Rhein

Emodin

11.263±1.037

3.892±0.406

7.524±0.701

Physcion

60.236±6.446

16.063±1.576

M AN U

SC

674 675 676 677 678 679 680 681 682 683 684

Chrysophanol

RI PT

Weight percentage (w/w, %)

Aloe–emodin

Table 2 Coded and real values of the factors for central composite design.

A B C D

–2 (–α)

–1

4.5 15 35 0.75

5.0 17.5 37.5 1.0

AC C

EP

Drug content (w, mg) Oil amount (%, w/w) Smix amount (%, w/w) Stabilizer amount (%, w/w) 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699

Symbols

TE D

Independent variables

24

Coded levels 0 +1 5.5 20 40 1.25

6.0 22.5 42.5 1.5

+2 (+α) 6.5 25 45 1.75

ACCEPTED MANUSCRIPT 700

Table 3 Regression analysis for responses ZP, MDS and PDI for fitting to quadratic model. Response

Model F–value

Model

Lack

P–value

of fit

R2

Lack

Adjusted Predicted 2

of fit

2

R

R

Std

%

Adequate

Dev

CV

precision

F-value P-value 28.83

< 0.0001

9.12

0.0236

0.967

0.9330

0.8128

MDS

62.08

< 0.0001

2.74

0.1715

0.984

0.9683

0.9172

PDI

17.91

< 0.0001

8.87

0.0249

0.947

0.8942

0.7049

20.419

0.4500 1.70

29.242

0.0085 5.77

18.063

SC

701 702 703 704 705 706

0.6000 4.06

RI PT

ZP

707 708 709

AC C

EP

TE D

M AN U

Table 4 The optimal kinetic equations used for fitting in vitro release data Rhubarb Dosage form Fitted Evaluate index anthraquinones equation r2 AIC Aloe–emodin Nanoemulsion Ritger–Peppas –26.57 0.9910 Suspension Ritger–Peppas –23.54 0.9405 Rhein Nanoemulsion Ritger–Peppas –36.12 0.9798 Suspension Ritger–Peppas –38.00 0.9871 Emodin Nanoemulsion Ritger–Peppas –36.56 0.9809 Suspension Ritger–Peppas –31.56 0.9434 Chrysophanol Nanoemulsion First order –91.65 0.9985 Suspension Ritger–Peppas –77.83 0.9923 Physcion Nanoemulsion Ritger–Peppas –48.03 0.9731 Suspension Higuchi –49.20 0.9563 RhA Nanoemulsion Ritger–Peppas –82.77 0.9958 Suspension Ritger–Peppas –84.65 0.9890

25

Parameters k n 0.0681 0.6079 0.1961 0.4280 0.2613 0.2037 0.3933 0.1560 0.0283 0.6080 0.1107 0.3798 0.0028 – 0.0068 0.7703 0.0063 0.6840 0.0177 – 0.0169 0.6320 0.0435 0.4480

ACCEPTED MANUSCRIPT

10

1

Physcion NE Suspension 0.017 0.015 ±0.006 ±0.007 0.262 0.096 ±0.0936 ±0.039 0.303 0.102 ±0.086 ±0.456 15.491 6.197 ±1.223 ±0.598 23.796 6.920 ±4.911 ±1.285 1.500 1.333 ±0.866 ±0.577 16.267 3.430 ±3.695 ±1.170 193.988 114.859 ±83.880 ±12.758 7.978 25.342 ±2.078 ±9.408

RI PT

Chrysophanol NE Suspension 0.229 0.192 ±0.014 ±0.019 3.290 1.957 ±0.494 ±0.471 3.473 2.006 ±0.497 ±0.460 13.277 9.437 ±0.085 ±0.195 15.974 10.865 ±0.584 ±0.524 1.000 1.000 ±0.866 ±0.866 11.0323 10.691 ±0.0968 ±2.232 42.047 73.94 ±8.491 ±31.486 2.625 4.647 ±0.352 ±1.051

SC

Emodin NE Suspension 0.021 0.013 ±0.005 ±0.002 0.548 0.135 ±0.138 ±0.023 0.741 0.173 ±0.228 ±0.030 18.585 7.434 ±0.748 ±0.236 33.847 12.087 ±8.046 ±1.366 1.833 1.667 ±0.289 ±0.577 21.373 7.354 ±5.979 ±1.168 51.200 74.467 ±9.883 ±14.359 1.754 7.064 ±0.654 ±1.163

M AN U

Rhein NE Suspension 0.147 0.121 ±0.020 ±0.007 3.026 0.911 ±1.207 ±0.371 3.855 1.125 ±1.414 ±0.672 16.619 15.254 ±1.003 ±1.515 31.732 24.484 ±9.066 ±12.828 0.833 0.667 ±0.289 ±0.289 22.965 18.607 ±9.625 ±9.860 5.965 14.780 ±4.265 ±2.995 0.169 0.649 ±0.056 ±0.289

TE D

Aloe–emodin NE Suspension Cmax(mg/L) 0.044 0.030 ±0.010 ±0.002 AUC0-t(mg/L*h) 0.364 0.146 ±0.120 ±0.020 0.156 AUC0-∞(mg/L*h) 0.400 ±0.165 ±0.025 MRT0-t(h) 11.728 3.899 ±2.498 ±0.408 15.810 4.643 MRT0-∞(h) ±6.595 ±0.685 Tmax(h) 0.833 0.833 ±0.289 ±0.289 t1/2z(h) 12.249 2.705 ±5.568 ±0.297 Vz/F(L/Kg) 79.550 45.404 ±23.886 ±2.934 CLz/F(L/h/Kg) 4.959 11.770 ±1.675 ±1.947

EP

Parameters

AC C

11 12 13 14

Table 5 The main pharmacokinetic parameters of aloe-emodin, rhein, emodin, chrysophanol and physcion in rat plasma after oral administration of RhA–NE and RhA suspension (mean ± SD, n = 3). Anthraquinones NE Suspension 0.410 0.342 ±0.043 ±0.023 7.491 3.268 ±2.020 ±0.915 8.585 3.329 ±2.357 ±0.921 15.013 10.673 ±0.592 ±0.802 22.168 11.685 ±0.622 ±1.127 1.500 0.667 ±0.866 ±0.289 16.304 9.029 ±1.231 ±2.864 43.480 62.962 ±12.977 ±32.461 1.835 4.724 ±0.486 ±1.190

ACCEPTED MANUSCRIPT

Compounds

R1

Aloe–emodin CH2OH

R2 H

COOH

H

Emodin

CH3

OH

Chrysophanol

CH3

H

Physcion

CH3

OCH3

RI PT

Rhein

AC C

EP

TE D

M AN U

SC

Fig. 1. The chemical structures of five anthraquinones

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Fig. 2. Pseudo–ternary phase diagrams indicating o/w nanoemulsion region composed of oil: capryol 90: ethyl oleate (3:1, w/w), water, different ratios of Smix: cremophor RH 40 and transcutol HP.

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Fig. 3. Stabilizers effect on pH, ZP, MDS and PDI of nanoemulsion system (A) and pH effect on ZP, MDS, PDI of nanoemulsion system (B).

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Fig. 4. Three dimendional contour plots for factors interaction effect on ZP, MDS and PDI (p < 0.05).

AC C

EP

TE D

M AN U

Fig. 5. Morphology of RhA–NE under TEM.

SC

RI PT

ACCEPTED MANUSCRIPT

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

Fig. 6. In vivo concentration–time curves of A. aloe–emodin, B. rhein, C. emodin, D. chrysophanol, E. physcion and F. total rhubarb anthraquinones (RhA).