Spa

MMS

a

ARR1AA

KDUFLRP

1

ffC[idaihSa

mC

h0

Journal of Pharmaceutical and Biomedical Analysis 117 (2016) 372–379

Contents lists available at ScienceDirect

Journal of Pharmaceutical and Biomedical Analysis

journa l homepage: www.e lsev ier .com/ locate / jpba

imultaneous determination of four furostanol glycosides in ratlasma by UPLC–MS/MS and its application to PK study after oraldministration of Dioscorea nipponica extracts

in Liao, Cong Dai, Mengping Liu, Jiefeng Chen, Zuanguang Chen, Zhiyong Xie,eicun Yao ∗

chool of Pharmaceutical Sciences, Sun Yat-sen University, Guangzhou 510006, PR China

r t i c l e i n f o

rticle history:eceived 26 July 2015eceived in revised form5 September 2015ccepted 16 September 2015vailable online 21 September 2015

eywords:ioscorea nipponicaPLC–MS/MSurostanol glycosides

a b s t r a c t

A novel, sensitive and rapid ultra-performance liquid chromatography–tandem mass spectrometric(UPLC–MS/MS) method for simultaneous quantification of four furostanol glycosides in rat plasma wasestablished and validated. Ginsenoside Rb1 was used as an internal standard. Plasma samples were pre-treated by liquid–liquid extraction with n-butanol and chromatographed on a C18 column (2.1 × 50 mmi.d., 2.6 �m) using a gradient elution program consisting of acetonitrile and water (containing 0.03%formic acid and 0.1 mM lithium acetate) at a flow rate of 0.4 mL/min. Lithium adduct ions were employedto enhance the response of the analytes in electrospray positive ionization mode and multiple reac-tion monitoring transitions were performed for detection. All calibration curves exhibited good linearity(r > 0.999) over the range of 10–20,000 ng/mL for protodioscin and 2–4000 ng/mL for protogracillin, pseu-doprotodioscin and pseudoprotogracillin. The recoveries of the whole analytes were more than 80.3% and

ithium adductat plasmaharmacokinetics

exhibited no severe matrix effect. Meanwhile, the intra- and inter-day precisions were all less than 10.7%and accuracies were within the range of −8.1–12.9%. The four saponins showed rapid excretion and rel-ative high plasma concentrations when the validated method was applied to the PK study of Dioscoreanipponica extracts by intragastric administration at low, medium and high dose to rats. Moreover, the

mpou

T1/2 and AUC0–t of each coat different dose levels.

. Introduction

Dioscorea nipponica (DN), a folk medicine widely used in Chinaor thousands of years [1], as the major active ingredient of someamous Chinese patent medicine, Di’ao Xinxuekang capsule [2],huan–Shan–Long Injection® [3] and Guge Fengtong preparation4], is an important medicinal plant with high content of saponinsn its rhizome. Up to now, numerous bioactivities of DN have beenocumented. For instance, DN showed therapeutic effect on goutyrthritis based on the SDF-1/CXCR 4 and p38 MAPK pathway both

n vitro and in vivo [5]. Besides, it also exhibited cytotoxicity againstuman oral cancer HSC-3 cells [6], human HepG2 cell [7], humanH-SY5Y neuroblastoma cells [8], murine melanoma B16F10 cellsnd human melanoma A2058 cells [9]. In addition, in a large amount

∗ Corresponding author at: Laboratory of Pharmaceutical Analysis, School of Phar-aceutical Sciences, Sun Yat-sen University, Guangzhou Higher Education Meg

enter, Guangzhou 510006, PR China. Fax: +86 20 39943040.E-mail address: yaomeicun@gmail.com (M. Yao).

ttp://dx.doi.org/10.1016/j.jpba.2015.09.021731-7085/© 2015 Elsevier B.V. All rights reserved.

nd turned out to behave in a dose-dependent pattern by comparing them

© 2015 Elsevier B.V. All rights reserved.

of animal experiment, DN presented anti-obesity [10], anti-diabetic[11], anti-thrombotic [12], anti-arthritic [13], liver-protection [14]bioactivities.

Meanwhile, furostane- and spirostane-type skeleton saponinsare two major active ingredients in DN [15]. However, spirostane-type, such as dioscin [16] and gracillin [17], showed an extremelylow bioavailability in rats. Therefore, in this study, the relativehigher bioavailability components, furostane-type saponins [3,18],were the main ingredients we paid attention to. Among thesefurostanol glycosides [19–21], protodioscin (PD), protogracillin(PG), pseudoprotodioscin (PPD), pseudoprotogracillin (PPG) werethe relative higher contents in DN and the chemical structuresof them are shown in Fig. 1. What is more, PD and PPD showedanti-cancer [22,23] and anti-hyperlipidemic [24] effects as well. Inconsideration of multiple bioactivities of DN above and mission ofthe age to promote modernization of TCM, it is pressingly necessary

for us to figure out its pharmacokinetics (PK) behaviors.

To date, analytical methods for the quantification of the activechemicals of DN included high performance liquid chromatog-raphy with ultraviolet (UV) [20] or evaporative light scattering

M. Liao et al. / Journal of Pharmaceutical and Biomedical Analysis 117 (2016) 372–379 373

f the f

dLdppoosiocb

efsf

2

2

yCpoidm

2

(b

Fig. 1. The chemical structures o

etection (ELSD) [25]. Recently, an increasing number ofC-ESI–MS/MS methods were applied to characterization andetermination of DN [4,19,21]. Meanwhile, several researches hadreviously investigated the metabolism and PK of DN and someure steroidal saponins in DN [3,18,19,26]. However, these meth-ds were mainly focused on the quantification of total saponinsr one of them in plasma, and few of them were reported for theimultaneous determination of multiple saponins in biological flu-ds. Accordingly, as an explicit and effective herb medicine, it isf great significance to establish a sensitive and specific analyti-al method for simultaneous quantification of multiple saponins iniological samples and evaluate its PK behaviors.

In our study, a newly developed UPLC–MS/MS method wasstablished and validated for the simultaneous determination ofour furostanol glycosides in rat plasma. The rapid, sensitive andpecific method was successfully applied to evaluate the PKs ofour saponins after oral administration of DN extracts to rats.

. Experimental

.1. Plant materials

D. nipponica samples, cultivated with a growth period of 2ears and dried in the sun, were collected from Jilin Province,hina. The moisture content of DN sample was about 8.15%. Thelant materials were authenticated by Prof. Depo Yang, Schoolf Pharmaceutical Sciences, Sun Yat-sen University. Correspond-

ng voucher specimens (voucher number: DN-JL-20150604) wereeposited in Lab of Pharmaceutical Analysis and Quality Assess-ent, Sun Yat-sen University.

.2. Chemicals and reagents

The reference standards of PD, PG, PPD, PPG, ginsenoside Rb1internal standards, IS), with purity higher than 98% (checkedy NMR and HPLC-ELSD), were achieved from Beijing Xinrong

our furostanol glycosides and IS.

Technologies Co., Ltd. (Beijing, China). HPLC grade methanol andacetonitrile were purchased from Honeywell Burdick & Jackson(Muskegon, Michigan, USA). Formic acid was bought from ChengduKelong Chemical Reagent Factory (Sichuan, China). Lithium acetatewas from Aladdin Industrial Corporation (Shanghai, China). Deion-ized water up to a resistivity of 18.2 M� was prepared with aMilli Q-purification system (Barnstead International Inc., Dubuque,Iowa, USA).

2.3. Instrument and UPLC–MS/MS conditions

Sample analysis was performed on an UPLC–MS/MS system(Waters Corporation, Milford, Connecticut, USA) composed of anautosampler, a column oven (set at 30 ◦C) and a binary solventdelivery manager. A tandem mass spectrometry was connectedto the LC system through electrospray ionization (ESI) inter-face. LC separations were obtained on an Kinetex XB-C18 column(50 mm × 2.1 mm, 2.6 �m; Phenomenex, Torrance, California, USA)preceded by a SecurityGuard ULTRA Cartridges UHPLC precolumn.The gradient elution consisted of acetonitrile (mobile phase A) andultrapure water contained 0.03% formic acid and 0.1 mM lithiumacetate (mobile phase B) was set as follows: starting at 20% A–80%B, increasing A to 50% in 1.5 min, then decreasing A to 20% in 0.1 min,keeping constant for 0.5 min. The flow rate was 0.40 mL/min and thesample injection volume was 10 �L at room temperature. [M + Li]+

ions were used as the precursor for quantification in multiple-reaction monitoring (MRM) mode. The optimized cone voltage andcollision energy for four saponins and IS are shown in Table 1. Otherparameters of MS analysis were as follows: capillary voltage, 3 kV;

ultra-high pure nitrogen and argon were used as desolvation gas(900 L/h) and collision gas (0.21 mL/min), respectively. Cone gasflow rate, 50 L/h; source temperature and desolvation tempera-ture was 150 ◦C and 400 ◦C. MasslynxTM 4.1 software was used foracquiring and processing data.

374 M. Liao et al. / Journal of Pharmaceutical and Biomedical Analysis 117 (2016) 372–379

Table 1MS/MS parameters for detection of four saponins and IS.

Compound Parent ion (m/z) Daughter ion (m/z) Cone voltage (v) Collision energy [eV]

PD 1056 1038a 60 65892b

PG 1072 1054a 60 70908b

PPD 1038 892a 60 65566b

PPG 1054 908a 60 65349b

IS 1116 349a 95 55

2

tAi2ehltUo

2

t3ncipaweImgoDtva

2

pmofi4faPsos

774b

a Transition used for quantification.b Transition used for qualitatively analysis.

.4. Animals

Fifteen SPF-grade healthy male Wistar rats (225 ± 25 g) (Cer-ificate no. SCXK2011-0029) were obtained from the Laboratorynimal Center in Sun Yat-sen University. Animals were kept

n an environmentally controlled breeding room (temperature:3–26 ◦C, relative humidity: 45–60%) for 3 days before starting thexperiments and free accessible to food and water. Diet was pro-ibited for 12 h before the experiment, while water was taken ad

ibitum. The studies were approved by the Animal Ethics Commit-ee of Sun Yat-sen University and followed the Guide for Care andse of Laboratory Animals published by the US National Institutef Health [27].

.5. Preparation of DN extracts

DN (50 g) was chopped into 60 mesh pieces and extracted threeimes by ultrasonic (120 w, 40 kHz) with methanol (1:10, w/v) for0 min each time. The combined filtrate was concentrated to dry-ess and then suspended with water to gain DN extracts (DNE) at aoncentration equivalent to 1 g/mL of the raw DN material beforet is used. The stability data of extract DN in water is listed in Sup-lementary Table 2. To figure out the dosage of administration, themount of four saponins in the extract solution were quantifiedith UPLC–MS/MS. The quantities of PD, PG, PPD and PPG in the

xtract were 11.849, 0.991, 1.195 and 0.160 mg/mL, respectively.n order to picture the global chemical profile of DN, an HPLC-UV

ethod was established. The method was as follows: chromato-raphic separation was performed at 35 ◦C and detected at 203 nmn an Apollo C18 column (250 × 4.60 mm; 5 �m; Alltech, Illinois,eerfield, USA); the mobile phase consisted of acetonitrile–water

hrough gradient elution at a flow rate of 0.8 mL/min; the injectionolume was 10 �l. According to area normalization methods, themounts of the four analytes came to more than 50% (Fig. 4).

.6. Preparation of standard and quality control (QC) samples

The stock solutions of PD, PG, PPD, PPG and IS were pre-ared with methanol at concentration of 2 mg/mL, respectively. Theixed standard solutions of PD, PG, PPD and PPG for the preparation

f calibration curves were serially diluted with methanol to givenal concentrations of 20, 50, 200, 800, 4000, 20,000, 30,000 and0,000 ng/mL for PD and 4, 10, 40, 160, 800, 4000, 6000, 8000 ng/mLor the others, respectively. The working solutions at low, medium

nd high concentration were prepared in 50, 4000, 30,000 ng/mL forD and 10, 800, 6000 ng/mL for the rest by diluting the initial stockolution with methanol. A 200 ng/mL solution of the IS was alsobtained by dilution of the IS stock solution with methanol. Bothtock and working solutions were stored at 4 ◦C and the stability of

stock solutions were validated and related data was appended inSupplementary Table 1.

2.7. Samples preparation

Rat plasma samples were stored at −20 ◦C and thawed toroom temperature when used. To every 100 �L plasma sample,50 �L IS working solution (200 ng/mL), 50 �L methanol and 500 �Ln-butanol were added. After vortex for 3 min, the mixture was cen-trifuged at 16,000 × g for 5 min at 4 ◦C, and then the organic layerwas transferred to another blank Eppendorf tube and evaporatedto dryness at 45 ◦C. The residue was reconstituted with 100 �Lacetonitrile–water (20:80, v/v), a 10 �L aliquot was injected intothe UPLC–MS/MS for analysis after centrifugation at 16,000 × g for5 min at 4 ◦C.

2.8. Method validation

The establishment of validation method was performed inaccordance with the USA Food and Drug Administration Bioana-lytical Method Validation Guidance [28].

Calibration standards were prepared in triplicate and analyzedin duplicate in three independent runs. Linear regression method atleast squares principle was applied to determine the slope, inter-cept and square regression coefficient (r) of the linear regressionequation. The lower limit of quantitation (LLOQ) was determinedas the lowest concentration with values for precision and accuracywithin ±20% and the analyte response at the LLOQ was more thanfive times the response compared to blank samples at the sameretention time. To evaluate specificity, chromatograms of the ana-lytes at the LLOQ level were compared to those of blank plasmasamples in quintuplicate.

Intra- and inter-day precisions and accuracies of the methodwere determined by testing five replicates of each of the QC sam-ples at low, middle and high concentration (25 ng/mL, 2000 ng/mL,15,000 ng/mL for PD and 5 ng/mL, 400 ng/mL, 3000 ng/mL for theother three) in three separate analytical runs. Precision is expressedas the relative standard deviation (RSD) and accuracy as the relativeerror (RE), respectively.

To assess potential interferences by matrix, matrix effect wascarried out by comparing the responses achieved from post-extraction blank rat plasma and mobile phase spiked with low,middle and high concentrations of the analyte, separately.

The absolute recovery of the analytes, after liquid–liquid extrac-tion, was determined by comparing the responses of the analytesfrom QC samples at low, middle and high concentrations withanalytes spiked in post-extracted blank rat plasma at equivalentconcentrations.

M. Liao et al. / Journal of Pharmaceutical and Biomedical Analysis 117 (2016) 372–379 375

Fig. 2. Product ion mass spectra, proposed collision-induced fragmentations of precursor ions and monitored transitions of the four analytes and IS. (A) PG, (B) PD, (C) PPD,(

atTac

apTsla

2

mAst1(tn5e

D) PPG, (E) IS.

Carryover was determined by measuring the peak area for thenalytes by injecting two processed blank matrix samples sequen-ially after injecting an upper limit of quantitation (ULOQ) sample.he response in the first blank matrix at the retention times of eachnalyte should be less than 20% of the response of its LLOQ sample,orrespondingly.

The stabilities of the target chemicals in rat plasma were evalu-ted by analyzing five aliquots of low and high QC samples tested forre-treatment, post-treatment and three cycles of freeze–thaws.he results were compared with those for freshly prepared QCamples and the percentage concentration deviation was calcu-ated. Accuracy should be within 85–115% of the nominal valuesnd precision within ±15% RSD.

.9. Pharmacokinetic study

Fifteen rats were randomly divided into low (5 g/kg DNE),edium (10 g/kg DNE), high (20 g/kg DNE) dose group, equally.

ll groups were treated with intragastric administration. Bloodamples (200 �L) were acquired with heparinized tubes by punc-ure of the retro-orbital sinus at 0 (predose), 5, 10, 30, 60, 90,50, 210, 270, 360, 480, 600 min after dosing. Plasma samples

100 �L) were gained by centrifuging at 16,000 × g for 1 min andhen samples were stored at −20 ◦C until analysis. Pharmacoki-etic parameters for the analytes were calculated with Winnonlin.0.1 (Pharsight, USA) by non-compartmental analysis. Data wasxpressed as mean ± SD.

3. Results and discussion

3.1. Optimization for UPLC–MS/MS conditions

For purpose of exploring more sensitive ionization mode, bothpositive and negative modes were compared using the abundanceof four reference compounds and IS as index. As a result, theresponse of the analytes in positive ionization mode was muchhigher than in negative. Ultimately, [M + Li]+ was chosen as the pre-cursor ion for both target compounds and IS in positive ionizationmode. The advantages of [M + Li]+ over [M + Na]+ and [M + H]+ forsuch kind of chemicals were clarified in our previous study [18]. Theoptimized both precursor ions and daughter ions for the analytesare listed in Table 1. Meanwhile, the proposed collision-inducedfragmentations of precursor ions of the analytes are shown in Fig. 2.

The chromatographic conditions were optimized to achievesymmetrical peak shapes, enhance the signal response and obtainshorter chromatographic cycle times for simultaneous analysis offour components and IS. In the first place, compared with WatersUPLC BEH C18 column (50 mm × 2.1 mm, 1.7 �m) and PhenomenexKinetex XB-C18 column (50 mm × 2.1 mm, 2.6 �m), the latter wasour final choice for better peek shapes and a relative shorter reten-

tion time (detailed in Supplementary Fig. 1). Meanwhile, methanolwas without our consideration as mobile phase for hydroxyl offurostanols at position C-22 could be transformed into methylethers in methanol under the high temperature of ion source[29]. Furthermore, water phase with 0.03% (v/v) FA could suppress

376 M. Liao et al. / Journal of Pharmaceutical and Biomedical Analysis 117 (2016) 372–379

Table 2The regression equations, linear range and LLOQs of the four saponins.

Analyte Range (ng/mL) Calibration curves Correlation coefficient (r) LLOQ (ng/mL)

824

055.896

.879

slE(wIde

3

imopoTueho

3

3

opcpro

Fsl

PG 2–4000 Y = 0.0055X − 0.0PD 10–20000 Y = 0.0045X − 0.3PPD 2–4000 Y = 1.1321X − 12PPG 2–4000 Y = 1.1207X − 22

ignal noise and increase the response of analytes. Besides, 0.1 mMithium acetate was applied to assist the formation of [M + Li]+.ventually, the mobile phases composed of (A) acetonitrile andB) aqueous formic acid (0.03%, v/v) and 0.1 mM lithium acetateith a gradient elution at a flow rate of 0.4 mL/min were used.

n this study, ginsenoside Rb1 was selected as the internal stan-ard because of its similarity of chemical structure, retention andxtraction behavior with the four saponins.

.2. Sample pretreatment

On account of the differences among the analytes in the phys-ochemical properties, it is important to explore an appropriate

ethod to prepare the samples with a high recovery and with-ut endogenous interferences for all these analytes. To begin with,rotein precipitation (PPT) was evaluated. However, PPT was leftut of account because of the severe matrix effect for PD and PG.hen, a conventional liquid–liquid extraction (LLE) approach wassed and ethyl acetate, n-butanol and acetone were all tested asxtraction solvent. Finally, n-butanol was adopted because of itsigher extraction efficiency (>80%) for all these saponins withoutbvious endogenous interference.

.3. Validation results

.3.1. Specificity and selectivityAs shown in Fig. 3, no severe endogenous interferences were

bserved in the retention times for four saponins and IS by com-

aring chromatograms of blank plasma of five different rats withhromatograms of plasma spiked with IS and the target com-ound standards at LLOQ level. Representative chromatograms ofat plasma are also presented at 1 h after intragastric administrationf DNE at low, medium and high dose in Fig. 3.

ig. 3. Extracted ion current chromatograms of the analytes and the IS (100 ng/mL) acqupiked with the analytes at LLOQ and IS, i.e., 10 ng/mL for PD and 2 ng/mL for the rest; (C)ow (5 g/kg), medium (10 g/kg), high (20 g/kg) dose of DNE to rats, respectively. 1: IS, 2: P

0.9998 20.9994 100.9999 20.9996 2

3.3.2. Linearity of calibration curves and LLOQAs depicted in Table 2, all calibration curves were found to be

adequate for the subsequent analysis of rat plasma samples andexhibited good linearity over the concentration of 10–20,000 ng/mLfor PD and 2–4000 ng/mL for the others with correlation coefficient(r) within the range of 0.9994–0.9999. The LLOQs were all suitablefor the determination of analytes in the pharmacokinetic studies.Moreover, according to Table 3, the accuracies and precisions of theanalytes at LLOQ levels were all below 10% which were acceptable.

3.3.3. Accuracy and precisionThe UPLC–MS/MS method provided high within-run and

between-run precision as well as accuracy of quantification offour saponins in rat plasma. As described in Table 3, the intra-day and inter-day accuracies of the four analytes were −8.1–12.0%and −7.1–12.9%, respectively. Meanwhile, precisions representedas RSD no more than 10.7% for both within-run and between-run. Ina word, the results above suggested that the method was accurate,reliable and reproducible.

3.3.4. CarryoverConspicuous carryover was observed after injection of the first

blank plasma samples, it was approximately 30% of the LLOQ forPD and no more than 5% for the other three and IS. Nevertheless,carryovers of the whole analytes were less than 5% of the LLOQ afterinjection of the second blank plasma samples which were withinthe acceptable range. Therefore, in our succeeding sample analysis,all samples were tested from low to high concentration. In addition,a blank plasma sample would be injected if the former sample was

close to ULOQ level.

3.3.5. Matrix effect and recoveryThe average extraction recoveries and matrix effects of the QC

samples are summarized in Table 4. Recoveries of the four saponins

ired during the analysis of plasma samples: (A) blank rat sample; (B) blank plasma, (D), (E) incurred rat plasma obtained at 30 min after intragastric administration ofG, 3: PD, 4: PPD, 5: PPG.

M. Liao et al. / Journal of Pharmaceutical and Biomedical Analysis 117 (2016) 372–379 377

Table 3Intra-day and inter-day precisions and accuracies for the determination of the four saponins from the assay samples (mean ± SD).

Analytes Nominal conc. (ng/mL) Intra-day (n = 5) Inter-day (n = 15)

Observedconcentration(ng/mL)

Precision (RSD, %) Accuracy (RE, %) Observedconcentration(ng/mL)

Precision (RSD, %) Accuracy (RE, %)

PG 2.0 2.1 ± 0.1 5.3 4.0 2.1 ± 0.1 5.6 3.35.0 4.6 ± 0.3 7.1 −7.2 4.7 ± 0.2 5.3 −6.6

400.0 417.7 ± 11.6 2.8 4.4 422.0 ± 9.5 2.3 5.53000.0 2963.7 ± 71.6 2.4 −1.2 2965.6 ± 77.4 2.6 −1.1

PD 10.0 10.6 ± 0.4 4.2 6.2 10.7 ± 0.6 5.3 6.7

25.0 25.4 ± 0.8 3.3 1.6 26.4 ± 1.3 4.9 5.52000.0 2240.3 ± 22.5 1.0 12.0 2257.6 ± 30.3 1.3 12.9

15000.0 14359.6 ± 308.0 2.1 −4.3 14639.8 ± 428.7 2.9 −2.4

PPD 2.0 2.2 ± 0.2 7.7 8.0 2.2 ± 0.2 10.7 8.35.0 4.8 ± 0.2 3.7 −4.4 4.8 ± 0.3 5.9 −4.2

400.0 387.1 ± 6.2 1.6 −3.2 386.7 ± 11.0 2.8 −3.33000.0 2775.1 ± 138.8 5.0 −7.5 2893.6 ± 163.6 5.7 −3.5

PPG 2.0 2.0 ± 0.2 10.6 −2.0 2.0 ± 0.2 7.8 −1.75.0 4.8 ± 0.3 5.7 −4.0 4.8 ± 0.2 4.9 −3.6

400.0 367.7 ± 25.8 7.0 −8.1 371.7 ± 19.4 5.2 −7.1−

awmarst

FPla

3000.0 2938.3 ± 190.3 6.5

t all levels were higher than 80.3% and IS was about 81.5%, whichas adequate and acceptable for the PK studies. Meanwhile, theatrix effects of IS and QC samples of the four target compounds

t three concentrations were observed to be within 97.3% and the

ange of 86.9–107.4%, respectively. Data above suggested that ionuppression or enhancement from plasma matrix is negligible forhis method.

ig. 4. (A) HPLC-UV (203 nm) profile of DNE; 1: PD, 2: PG, 3: PPD, 4: PPG, 5: dioscin and

G, PPD and PPG after oral administration of DNE at low (×), medium (�) and high (�) dow dose level; the plasma concentration of PPG at low dose level was under LLOQ. Vertict each does level (mean ± SD, n = 5). *Values are significantly different from each other (P

2.1 3072.9 ± 208.8 6.8 2.4

3.3.6. Stability testsThe results of stability evaluation under different conditions

were illustrated in Table 5. The deviations of the measured con-centrations from the standard ones for the whole analytes in

the stability tests were within the 15% assay variability limit.Namely, the four saponins were stable enough in our anticipantconditions.

gracillin. (A1), (A2), (A3), (A4) Mean rat plasma concentration–time profiles of PD,ose to rats. The figures in top right corner of (A1)–(A3) were the amplified ones ofal bars represent SD. The comparison of AUC0–t (B) and T1/2 (C) of the four saponins

< 0.05), **Values are significantly different from each other (P < 0.01).

378 M. Liao et al. / Journal of Pharmaceutical and Biomedical Analysis 117 (2016) 372–379

Table 4The absolute recovery and matrix effect of four saponins and IS in rat plasma (n = 5).

Analytes Spiked concentration Recovery (%) RSD (%) Matrix effect (%) RSD (%)(ng/mL) (mean ± SD) (mean ± SD)

PG 5.0 103.3 ± 6.0 5.8 107.4 ± 6.7 6.2400.0 89.8 ± 0.7 0.7 91.5 ± 1.9 2.1

3000.0 86.0 ± 5.2 6.0 101.8 ± 6.4 6.3

PD 25.0 96.8 ± 1.5 1.5 106.0 ± 2.9 2.72000.0 91.2 ± 9.6 10.5 86.9 ± 4.7 5.4

15000.0 80.4 ± 6.6 8.2 88.8 ± 1.3 1.4

PPD 5.0 84.0 ± 9.0 10.7 100.6 ± 8.1 8.0400.0 90.0 ± 6.1 6.8 89.6 ± 5.2 5.8

3000.0 83.3 ± 10.8 13.0 91.9 ± 4.5 4.9

PPG 5.0 92.6 ± 8.3 8.9 91.1 ± 12.4 13.6400.0 91.3 ± 9.4 10.3 91.5 ± 6.6 7.2

3000.0 80.3 ± 4.6 5.7 93.6 ± 2.8 2.9

IS 100.0 81.5 ± 7.8 9.6 97.3 ± 12.6 13.0

Table 5The stability of the four saponins in rat plasma under different storage conditions (n = 5).

Analytes Nominal conc. (ng/mL) 25 ◦C for 4 h Frozen for 15 days Three-freeze–thaw cycles 4 ◦C for 24 h

Precision(RSD, %)

Accuracy(RE, %)

Precision(RSD, %)

Accuracy(RE, %)

Precision(RSD, %)

Accuracy(RE, %)

Precision(RSD, %)

Accuracy(RE, %)

PG 5.0 5.4 −4.4 5.6 −2.8 4.7 2.4 10.2 −5.23000.0 3.8 2.7 7.4 −4.9 1.7 4.3 2.6 5.5

PD 25.0 7.4 −8.5 4.7 −4.0 2.9 1.9 3.2 7.415000.0 1.1 0.1 0.8 −9.5 1.2 0.9 2.2 1.3

PPD 5.0 3.4 −6.0 7.2 −6.0 11.2 −8.0 11.6 −11.63000.0 6.6 10.3 7.2 −8.2 3.4 9.5 4.5 −4.3

PPG 5.0 7.9 −6.4 11.4 −7.6 12.1 −3.2 8.0 −6.03000.0 3.4 5.2 4.0 0.5 1.3 −1.5 8.6 −5.4

Table 6Pharmacokinetic parameters of four saponins after intragastric administration of DNE at doses of 5 g/kg (n = 5), 10 g/kg (n = 5), and 20 g/kg (n = 5) to rats.

Analytes Dose T1/2 Tmax Cmax AUC0−t AUC0−∞ MRT0−∞mg/kg min min ng/mL ng min/mL ng min/mL min

PG 4.96 195.9 ± 26.2 52.5 ± 15.0 37.5 ± 12.6 10231.9 ± 3162.7 11597.1 ± 3270.5 287.6 ± 39.59.91 150.7 ± 26.3 52.5 ± 15.0 486.3 ± 303.7 59249.1 ± 31,887.2 61963.0 ± 31,488.8 165.7 ± 41.5

19.82 116.4 ± 34.6 67.5 ± 15.0 2011.4 ± 426.2 238463.5 ± 61210.8 242221.1 ± 61,137.3 129.1 ± 21.2

PD 59.25 194.8 ± 27.0 35.0 ± 28.9 293.3 ± 120.5 71135.7 ± 23397.3 79507.9 ± 23,160.5 271.6 ± 49.4118.49 142.9 ± 32.2 37.5 ± 15.0 4557.7 ± 2638.5 482572.4 ± 246825.0 503093.0 ± 245815.1 151.8 ± 33.9236.98 95.9 ± 13.2 82.5 ± 15.0 15575.1 ± 3115.0 1869125.1 ± 412759.7 1890521.7 ± 4064858.0 120.8 ± 25.4

PPD 5.98 120.1 ± 38.6 90.0 ± 42.4 25.5 ± 7.9 5927.4 ± 2667.4 6254.5 ± 2793.4 211.5 ± 47.111.95 103.7 ± 37.4 52.5 ± 15.0 436.1 ± 282.0 50868.0 ± 28654.3 51547.3 ± 28429.8 126.2 ± 27.223.90 87.0 ± 15.4 82.5 ± 15.0 1903.0 ± 626.9 259101.6 ± 63653.1 261328.4 ± 64499.0 131.5 ± 21.5

.0

.2

3

aamAimtT

itc

PPG 1.60 165.5 ± 15.1 50.0 ± 17.3 37.9 ± 123.20 153.9 ± 39.9 67.5 ± 15.0 86.7 ± 52

.4. Application of the assay method

The established method depicted above was successfullypplied to the PK study of four furostanol glycosides in rat plasmafter oral administration of DNE at low, medium and high dose. Theean plasma concentration–time profiles and the comparison of

UC0–600 min and T1/2 of the different dose groups for four saponinsn rats are described in Fig. 4. Furthermore, non-compartmental

ode was applied to calculate all of the pharmacokinetic parame-ers by using Winnonlin 5.0.1 software and the results are listed in

able 6.

As illustrated in Table 6, the pharmacokinetic parametersncluding half-time (T1/2), time to reach the maximum concentra-ions (Tmax), maximum plasma concentration (Cmax), area underoncentration–time curve (AUC0–t and AUC0–∞), mean residence

4154.1 ± 1088.3 4411.9 ± 1019.0 167.7 ± 42.612887.3 ± 7547.3 13823.7 ± 7950.6 206.2 ± 88.0

time (MRT0–∞) are listed as mean ± SD. Significant differences ofT1/2 were observed between low and high dose groups for PGand PD. What is more, AUC0–600 min between low and high dosegroups, medium and high dose groups was statistical significancefor PG, PD and PPD as well. Generally, four furostanol glycosidesexhibited rapid excretion and relative high plasma concentra-tions in rats. In our preliminary studies and reported in literatures[20,21], spirostanol saponins, such as dioscin and gracillin, alsoaccounted a high amount in DN. However, it is difficult to simul-taneously quantify them with furostanol glycosides in a PK study

for the extremely low plasma concentrations and relative longerhalf-time [17]. Meanwhile, as a well-known effective componenttreating with cardiovascular disease [2], it may be reasonablefor us to monitor its ingredients working quickly and effectivelyin vivo.

and B

asvmAea

4

mnrtpafpcc

A

nRR

A

t

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

animals.pdf,2011.[28] US Food and Drug Administration, Center for Drug Evaluation and Research,

http://www.fda.gov/downloads/Drugs/Guidance/ComplianceRegulatoryInformation/Guidances/UCM070107.pdf, 2001.

M. Liao et al. / Journal of Pharmaceutical

In addition, it is interesting to find that Cmax and AUC of PPDfter intragastric administration of DN were much higher than aingle oral administration of PPD in rats compared with our pre-ious study [18]. It is supposed that the other ingredients in DNay enhance the absorption of PPD and improve its bioavailability.

s a consequence, this ubiquitous phenomenon in TCM may wellxplain the reason why most of TCMs work effectively despite theppearance of possessing simply trace chemical constituents.

. Conclusion

In summary, a specific, sensitive and rapid UPLC–MS/MSethod was the first time established and validated to simulta-

eously quantify the concentration of four furostanol glycosides inat plasma with a LLOQ value of 10 ng/mL for PD and 2 ng/mL forhe others. In addition, the method was successfully applied to theharmacokinetic study of the four saponins following intragastricdministration at low, medium and high dose of DNE to rats. Theour furostanol glycosides showed rapid excretion and relative highlasma concentrations in rats. Moreover, the T1/2 and AUC0–t of eachompound turned out to behave in a dose-dependent manner byomparing them at three dose levels.

cknowledgments

This work was funded by the Guangzhou Science and Tech-ology Program (No. 2014J4100171), the National Key Technology&D Program during the Twelfth Five-Year Plan Period of People’sepublic of China (No. 2013BAD10B04-2).

ppendix A. Supplementary data

Supplementary data associated with this article can be found, inhe online version, at http://dx.doi.org/10.1016/j.jpba.2015.09.021.

eferences

[1] Chinese Pharmacopoeia Commission, Pharmacopoeia of the People’s Republicof China, People’s Medical Publishing House, Beijing, 2010, pp. 250–251.

[2] Y.N. Yu, S.Y. Hu, G.X. Li, J. Xue, Z.M. Li, X.L. Liu, X.Y. Yang, B. Dong, D.H. Wang,X.F. Wang, S.R. Liu, J. Liu, B.W. Chen, L.Y. Wang, S.S. Liu, Q.G. Chen, C.T. Shen, Z.Wang, Y.Y. Wang, Comparative effectiveness of Di’ao Xin Xue Kang capsuleand compound danshen tablet in patients with symptomatic chronic stableangina, Sci. Rep. 4 (2014) 7058.

[3] T.J. Wang, Z.B. Liu, J. Li, M. Zhong, J.P. Li, X.H. Chen, K.S. Bi, Determination ofprotodioscin in rat plasma by liquid chromatography–tandem massspectrometry, J. Chromatogr. B 848 (2007) 363.

[4] X.L. Cheng, L.W. Qi, Q. Wang, X.G. Liu, B. Boubertakh, J.Y. Wan, E.H. Liu, P. Li,Highly efficient sample preparation and quantification of constituents fromtraditional Chinese herbal medicines using matrix solid-phase dispersionextraction and UPLC–MS/MS, Analyst 138 (2013) 2279.

[5] L. F. Lu, D.H. Liu, X.Z. Yu, Q. Li, S.M. Liu Zhou, Therapeutic effect of RhizomaDioscoreae Nipponicae on gouty arthritis based on the SDF-1/CXCR 4 and p38MAPK pathway: an in vivo and in vitro study, Phytother. Res. 28 (2014)280–288.

[6] M.H. Chien, Y.H. Ying, Y.S. Hsieh, Y.C. Chang, C.M. Yeh, J.L. Ko, W.S. Lee, J.H.Chang, S.F. Yang, Dioscorea nipponica Makino inhibits migration and invasion

of human oral cancer HSC-3 cells by transcriptional inhibition of matrixmetalloproteinase-2 through modulation of CREB and AP-1 activity, FoodChem. Toxicol. 50 (2012) 558–566.

[7] L. Zhao, T.J. Wang, L. Cheng, J. Wang, K.S. Bi, X.H. Chen, HPLC fingerprintanalysis of Dioscorea nipponica Mak. and the correlation with its anticanceractivity on human HepG2 cell, J. Shenyang Pharm. Univ. 30 (2013) 366–373.

[

iomedical Analysis 117 (2016) 372–379 379

[8] M.A. Rahman, H.J. Yang, N.H. Kim, S.O. Huh, Induction of apoptosis byDioscorea nipponica Makino extracts in human SH-SY5Y neuroblastoma cellsvia mitochondria-mediated pathway, Animal Cells Syst. 18 (2014) 41–51.

[9] M.L. Ho, Y.S. Hsieh, J.Y. Chen, K.S. Chen, J.J. Chen, W.H. Kuo, S.J. Lin, P.N. Chen,Antimetastatic potentials of Dioscorea nipponica on melanoma in vitro and invivo, Evid. Based Complement. Altern. Med. (2015) 1–13, http://dx.doi.org/10.1155/2011/507920(2011).

10] C.S. Kwon, H.Y. Sohn, S.H. Kim, J.H. Kim, K.H. Son, J.S. Lee, J.K. Lim, J.S. Kim,Anti-obesity effect of Dioscorea nipponica Makino with lipase-inhibitoryactivity in rodents, Biosci. Biotech. Biochem. 67 (2003) 1451–1456.

11] H. Yu, L.L. Zheng, L. Xu, L.H. Yin, Y. Lin, H. Li, K.X. Liu, J.Y. Peng, Potent effects ofthe total saponins from Dioscorea nipponica Makino againststreptozotocin-induced type 2 diabetes mellitus in rats, Phytother. Res. 29(2015) 228–240.

12] N.X. Zhao, Y.M. Han, S.J. Zhang, Antithrombotic components in Dioscoreanipponica, Chin. Tradit. Herb. Drugs 42 (2011) 652–655.

13] X.J. Liang, T.Y. Sun, W.J. Dong, Y.Z. Shang, H.R. Song, Effects of total saponinfrom Rhizoma Dioscorea nipponica on inflammatory cytokines incollagen-induced arthritis rats, Lishizhen Med. Mater. Med. Res. 24 (2013)522–524.

14] H. Yu, L.L. Zheng, L.H. Yin, L. Xu, Y. Qi, X. Han, Y.W. Xu, K. Liu, J.Y. Peng,Protective effects of the total saponins from Dioscorea nipponica Makinoagainst carbon tetrachloride–induced liver injury in mice through suppressionof apoptosis and inflammation, Int. Immunopharmacol. 19 (2014) 233–244.

15] M. Sautour, A.C. Mitaine-Offer, M.A. Lacaille-Dubois, The Dioscorea genus: areview of bioactive steroid saponins, J. Nat. Med. 61 (2007) 91.

16] K. Li, Y. Tang, J.P. Fawcett, J. Gu, D. Zhong, Characterization of thepharmacokinetics of dioscin in rat, Steroids 70 (2005) 525–530.

17] B.W. Wang, S.G. Ji, H. Zhang, L. Zhao, L. Lv, Y.Y. Li, G.C. Zhou, G.Q. Zhang, Liquidchromatography tandem mass spectrometry in study of the pharmacokineticsof six steroidal saponins in rats, Steroids 78 (2013) 1164–1170.

18] M. Liao, X. Chen, J.F. Chen, M.P. Liu, J.Y. Wang, Z.G. Chen, Z.Y. Xie, M.C. Yao,Determination of pseudoprotodioscin in rat plasma by UPLC–MS/MS: assaydevelopment and application to pharmacokinetic study, J. Chromatogr. B(2015), http://dx.doi.org/10.1016/j.jchromb.2015.05.004.

19] T.N. Tang, T. Yi, H.M. Chen, Z.Z. Zhao, Z.T. Liang, H.B. Chen, Quantitativecomparison of multiple components in Dioscorea nipponica and D. panthaicaby ultra-high performance liquid chromatography coupled with quadrupoletime-of-flight mass spectrometry, Phytochem. Anal. 24 (2013) 413–422.

20] C.Z. Liu, H.Y. Zhou, Q. Yan, Fingerprint analysis of Dioscorea nipponica byhigh-performance liquid chromatography with evaporative light scatteringdetection, Anal. Chim. Acta 582 (2007) 61–68.

21] S.H. Lin, D.M. Wang, D.P. Yang, J.H. Yao, Y. Tong, J.P. Chen, Characterization ofsteroidal saponins in crude extract from Dioscorea nipponica Makino by liquidchromatography tandem multi-stage mass spectrometry, Anal. Chim. Acta599 (2007) 98.

22] E. Abdel-Sattar, M.M. Shabana, S. El-Mekkawy, Protodioscin andpseudoprotodioscin from Solanum intrusum, Res. J. Phytochem. 2 (2008)100–105.

23] K. Hu, X.S. Yao, Protodioscin (NSC-698 796): its spectrum of cytotoxicityagainst sixty human cancer cell lines in an anticancer drug screen panel,Planta Med. 68 (2002) 297–301.

24] T.J. Wang, R.C.Y. Choi, J. Li, J. Li, C.W.C. Bi, L.Q. Zang, Z.B. Liu, T.T.X. Dong, K.S.Bi, K.W.K. Tsim, Antihyperlipidemic effect of protodioscin, an activeingredient isolated from the Rhizomes of Dioscorea nipponica, Planta Med. 76(2010) 1642–1646.

25] H.F. Pan, X. Hong, Y.R. Li, Determination of diosgenin in Dioscorea nipponica byHPLC-ELSD, Chin. J. Exp. Tradit. Med. Form. 08 (2011) 119–121.

26] H. Ren, J.P. Chen, B.Y. Tan, D.M. Wang, S.H. Lin, Y.P. Zhang, D.P. Yang, T.Y.W.Loo, H.S.B. Yeung, M. Jin, X. Li, The disposition and pharmacokinetics ofDioscorea nipponica Makino extract in rats, Afr. J. Biotechnol. 7 (2008)4015–4018.

27] US NIH, Guide for the Care and Use of Laboratory Animals, Eighth ed., http://grants.nih.gov/grants/olaw/Guide-for-the-care-and-use-of-laboratory-

29] L. Dinan, J. Harmatha, R. Lafont, Chromatographic procedures for the isolationof plant steroids, J. Chromatogr. A 935 (2001) 105–123.