Introduction!

Lancea tibetica Hook. F. et Thoms (Family: Scro-phulariaceae) is a Tibetan medicinal plant usedfor the prevention and treatment of leukemia, in-testinal angina, heart disease, and cough in tradi-tional Chinese medicine [1]. Recently, some pat-ents included this plant for treatment of pneu-monic influenza and cardiovascular and cerebro-vascular diseases [2–4]. Phytochemical investiga-tions have identified several different types ofcompounds from L. tibetica including phenylpro-panoids, lignans, flavonoids, terpenoids, and ste-roids [5–8]. The pharmacological and antioxidanteffects of L. tibetica were mainly attributed to thepresence of phenylpropanoid glycosides and lig-nans [9,10]. In addition to anti-inflammatory,anti-viral, and antioxidant properties, the hydro-philic glycosylated phenylpropanoids verbasco-side and isoverbascoside were reported to displayantitumor activities [6,9]. These compoundscould also induce the proliferation of murinebone marrow-derived dendritic cells and showsynergistic effects in coordination with cytokine[11]. In this plant, eleven furofurano lignans,

which contain a core skeleton 2,6-diaryl-3,7-diox-abicyclo[3.3.0]octane, have been isolated. Oneunique tetrahydrofuran type lignan glucoside,viz., tibeticoside A, has been isolated as well. Fu-rofurano lignan compounds are lantibetin, lanti-beside B, lantibeside C, lantibeside D, sylvatesmin(phyllygenol), phillyrin, simplexoside, lantibeside,sesamin, and sesaminol 2′-O-β-D-glucoside. Soket al. [12] summarized that furfuran type lignansfrom edible plants possessed anticancer, antioxi-dant, cardiovasculoprotective, neuroprotective,and anti-inflammatory activities. In an in vitroantitumor assay, sylvatesmin showed strong cyto-toxicity against human hepatoma cells (SMMC-7721), human uterine cervix carcinoma cells (He-La), and mouse melanoma cells (B16) [9,10].The chemical fingerprint analysis of medicalplants requires standard compounds (markers)having specific chemical composition and/or bio-logical activity. The analytical method should beable to identify corresponding plants unambigu-ously. In the plant L. tibetica, lignan componentsincluding tibeticoside A, lantibeside, lantibesideB, and lantibeside C are unique. All these com-pounds have only been reported from L. tibetica.

Abstract!

An HPLC method was developed for simultaneousdetermination of one phenylpropanoid glycoside,verbascoside (1), and nine lignans, including lan-tibeside (2), phillyrin (3), lantibeside B (4), lanti-beside C (5), tibeticoside A (6), styraxjaponosideC (7), sylvatesmin (8), (+)-piperitol (9), and hors-fieldin (10), from the Tibetanmedicinal plant Lan-cea tibetica Hook. F. et Thoms. The analysis wasperformed within 45min. The extraction methodwas optimized with different solvent systems.The HPLC method was validated for linearity, re-peatability, accuracy, limits of detection, and lim-its of quantification. The limits of detection and

limits of quantification of 10 analytes were foundto be less than 0.1 and 0.5 µg/mL, respectively. TheRSD for intra- and inter-day analyses was lessthan 4.2%, and the recovery efficiency was 90–105%. The method was used to analyze differentpopulations of L. tibetica collected in China. HPLCprofiles showed that the concentrations of ana-lytes were different in samples collected from dif-ferent areas of China. Verbascoside was the domi-nant component in three out of five plant sam-ples; compounds 2, 3, 6, and 8 accounted for over62% yields in total lignan contents. The method isuseful for identification, quality assurance, andquality control of L. tibetica and its related prod-ucts.

Quantitative Determination of 10Phenylpropanoid and Lignan Compoundsin Lancea tibetica by High-Performance LiquidChromatography with UV Detection

Authors Zong-Hua Song1,2, Yan-Hong Wang1, Zhong-Zhi Qian2, Troy J. Smillie1, Ikhlas A. Khan1,3

Affiliations 1 National Center for Natural Products Research, School of Pharmacy, University of Mississippi, University, Mississippi, USA2 Chinese Pharmacopeia Commission, Beijing, P.R. China3 Department of Pharmacognosy, University of Mississippi, University, Mississippi, USA

Key wordsl" Lancea tibetical" Scrophulariaceael" HPLC‑UVl" verbascosidel" lignan

received Sept. 28, 2010revised January 24, 2011accepted February 2, 2011

BibliographyDOI http://dx.doi.org/10.1055/s-0030-1270833Published online February 23,2011Planta Med 2011; 77:1562–1566 © Georg ThiemeVerlag KG Stuttgart · New York ·ISSN 0032‑0943

CorrespondenceProf. Dr. Ikhlas A. KhanNational Centerfor Natural Products ResearchSchool of PharmacyUniversity of MississippiUniversity, MS 38677USAPhone: + 16629157821Fax: + 16629157989ikhan@olemiss.edu

1562

Song Z-H et al. Quantitative Determination of… Planta Med 2011; 77: 1562–1566

Original Papers

One HPLC method has been developed to quantify ursolic acidand oleanolic acid in this plant [13]; however, no analyticalmethods using lignans as markers for L. tibetica characterizationhave been reported. As part of a program to address the issues ofauthenticity of material source, safety, quality assurance, andquality control of botanical and dietary supplements products,nine furofurano lignans along with verbascoside, isoverbasco-side, β-sitosterol, and oleanolic acid were identified fromL. tibetica. In the present work, an HPLC method was developedfor quantitative determination of ten compounds, verbascoside(1), lantibeside (2), phillyrin (3), lantibeside B (4), lantibeside C(5), tibeticoside A (6), styraxjaponoside C (7), sylvatesmin (8),(+)-piperitol (9), and horsfieldin (10), in L. tibetica. The chemicalstructures are given inl" Fig. 1. The HPLCmethod is validated andapplied for analysis of different collections of L. tibetica from dif-ferent areas in China. The content of analytes and their HPLCchromatogram profile were also discussed in this paper.

Materials and Methods!

Instrumentation and chromatographic conditionsAll analyses were performed on aWaters Alliance 2695 HPLC sys-tem (Waters Corp.) equipped with a 2996 photodiode array de-tector and a computerized data station, operated with WatersEmpower-2 software. Separation was achieved on a Gemini C18column (250 × 4.6mm id; 5-µm particle size; Phenomenex, Inc.)operated at 25°C. The column was equipped with a 2-cm LC-18guard column (Phenomenex, Inc.). The mobile phase consistedof water (A) and acetonitrile (B), both containing 0.1% acetic acid.The following gradient elution was used: 0min, 20% B; in next3min to 22% B; in next 17min to 27% B; continually increasingsolvent B to 32% B at 30min and to 68% B at 45min, respectively.In the following 2min the eluent composition was increased to100% B and the column was flushed for 8 minutes. Each run wasfollowed by a 10-min wash with 100% acetonitrile and an equili-bration period of 15min. The flow rate was 1.0mL/min, and theinjection volume was 10 µL. The identified compounds were de-tected at 280 nm.

Chemicals, standards, and plant samplesAcetonitrile, water, and acetic acid were HPLC grade and pur-chased from Fisher Scientific. Standard compounds (1–10) wereisolated at the National Center for Natural Products Research(NCNPR). The identity and purity were confirmed by chromato-graphic methods, spectroscopic data (1D-, 2D‑NMR and HR-ESI-MS) and by comparison with published spectra data [5–8,14].The purity of these standard compounds was calculated to be97.6%, 98.6%, 99.2%, 96.5%, 99.1%, 98.1%, 96.7%, 99.5%, 98.2%,and 99.5% for compounds 1–10, respectively. Stock solutions ofthe standards were stable for at least three months at 4°C.Plant samples, whole plants of L. tibetica coded as NCNPR # 8084,7782, 7783, 7784, and 7785, were collected from different re-gions of China and identified by Dr. Yueguo Zhong from Chong-qing Research Institute of TCMs. Specimens of the sample are de-posited at the NCNPR, University of Mississippi, Mississippi, USA.

Sample preparationDry plant material was grounded using a coffee bean grinder(SmartGrind™; Black and Decker) and passed through a No. 40sieve. About 200mg of plant samples were weighed and soni-cated in 2.0mL of 80% methanol for 30min followed by centrifu-gation for 10min at 4000 rpm. The supernatant was transferredto a 10-mL volumetric flask. The procedure was repeated fourtimes and the respective supernatants combined. The final vol-umewas adjusted to 10mL. Prior to application, an adequate vol-ume (ca. 2mL) was passed through a 0.45-µm polyethersulfonemembrane (Whatman). The first 1.0mL was discarded, and theremaining volume was collected in a LC sample vial. Each samplesolution was injected in triplicate.

Preparation of standard solutionsStock solutions of standard compounds were prepared at a con-centration of 1.0mg/mL inmethanol. The calibration curves wereprepared at seven different concentration levels. The range of theconcentration was 0.5–100 µg/mL.

Validation procedureThe HPLC method was validated in terms of precision, accuracy,and linearity according to ICH guidelines [15]. The accuracy ofthe assay method was evaluated in triplicate using two concen-tration levels of 10 and 20 µg/mL. The limit of detection (LOD)

Fig. 1 Structure of 10 standard compounds:verbascoside (1); lantibeside (2); phillyrin (3);lantibeside B (4); lantibeside C (5); tibeticoside A(6); styraxjaponoside C (7); sylvatesmin (8);(+)-piperitol (9), and horsfieldin (10).

1563

Song Z-H et al. Quantitative Determination of… Planta Med 2011; 77: 1562–1566

Original Papers

and limit of quantification (LOQ) were determined by injectingdilute solutions of the standards with known concentrations.The LOD and LOQ were defined as the signal-to-noise ratio equalto 3 and 10, respectively.

Results and Discussion!

In a preliminary phase, different reverse-phase HPLC columnswere tested in order to optimize the condition of separation. Thetested columns were Phenomenex Synergi Fusion-RP(150 × 4.6mm, 4 µm), Synergi Hydro-POLAR (150 × 4.6mm, 4µm), Gemini C18 (150 × 4.6mm, 5 µm), and Gemini C18 (250 ×4.6mm, 4 µm). The best separation efficiency and peak shapewere achieved by a 250 × 4.6mm Phenomenex Gemini C18 col-umn. Optimal chromatographic conditions were obtained by us-ing acetonitrile and water, both containing 0.1% acetic acid. Ace-tonitrile was preferred over alcohol type solvents because it en-hanced the separation as well as a relatively low back pressure.Acetic acidwas used as an additive because it improved the peaksshape and separation. Various column temperatures between 25and 40°Cwere tested, and the temperature finally usedwas 25°C.Selecting 280 nm as the detection wavelength resulted in an ac-ceptable response and enabled the detection of all 10 compoundsused in this study.Optimization of the extraction solvents and procedures was es-sential in the development of analytical methods because differ-ent solvents and procedures will influence the extraction effi-ciency of different compounds, as well as the overall sensitivityand selectivity of the method. Various extraction solvent systemswere evaluated for the maximum recoveries of all ten com-pounds. The plant material (LT 8084) was extracted with differ-ent solvent systems (1) 100% methanol, (2) 80% methanol, (3)70% methanol, (4) 60% methanol, and (5) 100% acetonitrile. Therecoveries of the major compounds were found to be higher(> 92%) with 80%methanol thanwith other solvents. The contentof verbascoside, a dominant plant component in LT 8084, was0.92% and 0.77%, respectively, in the fourth and fifth extracts.Therefore, the extraction procedure of the plant material was op-timized by using 80% MeOH, and it was extracted five times us-ing the sonication method.The HPLC method was validated for precision, accuracy, and line-arity. The specificity was determined by injecting individual sam-

ples, and no interference was observed for any of the compo-nents. The chromatograms were checked for the appearance ofany extra peaks. Linear calibration plots were obtained over sev-en concentration levels (0.5–100 µg/mL). The results showed alinear correlation between the peak area and concentration(l" Table 1). The precision study used the plant sample LT 8084and followed the validated procedure for sample preparation.The precision of the assay method was evaluated by carrying outthree independent assays on three consecutive days. The % RSD ofall ten analytes was determined to be within the acceptable limitof 5.0%.Multiple injections illustrated that the resultswere highlyreproducible and had a low standard deviation. The RSD of assayresults obtained in inter-day and intra-day study (l" Table 2) waswithin 4.5% and showed a maximum of 4.05% for tibeticoside Ain intra-day, and 4.2% for styraxjaponoside C in inter-day studies,confirming a good precision of the developed method. The accu-racy of the method was determined for the related substance byspiking samples with a known amount of all 10 standard com-pounds. The accuracy of the assay method was evaluated in trip-licate at two concentration levels, 10 and 20 µg/mL, of 10 analytesin the samples (l" Table 1). The percentage recovery ranged from90 to 105%.Five plant samples of L. tibetica, designated as LT 8084, LT 7782,LT 7783, LT 7784, and LT 7785, were analyzed by the LC‑UVmeth-od. l" Fig. 2 shows LC‑UV chromatograms of five samples at280 nm. The chromatograms showed the presence or absence of10 selected components in tested samples. The concentration ofcompounds 1–10 are shown in l" Table 3. The compounds associ-ated with each chromatographic peak in the plant samples wereidentified by spiking the samples with standard compounds andby comparison of UV spectra and retention times with those ofthe standards (l" Fig. 2).Upon comparing chromatograms of plant material, it was deter-mined that compound 1 was the dominant component in sam-ples LT 8084, LT 7782, and LT 7784, and a major component inLT 7785. However, in sample LT 7783, the amount of compound1was much less than in the rest of the samples. Among those lig-nan compounds, the total concentrations of compounds 2, 3, 6,and 8 accounted for over 62% of the total lignan contents in thisplant. It is worth noting that compounds 2, 3, and 8 share thesame aglycone sylvatesmin, a cytotoxic compound in vitro. Asshown in l" Fig. 2 and Table 3, the major compounds 1–3, 6, and8 were detected in all analyzed samples. For plant samples, the

Table 1 Calibration data for standard compounds 1–10 including regression equation, correlation coefficient, recovery rate, limit of detection, and limit of quan-tification.

Compounds Equation R2 Recovery LOD

(µg/mL)

LOQ

(µg/mL)

10 µg/mL 20 µg/mL

Verbascoside Y = 1.22e + 004 X − 1.12e + 004 0.9999 105.9 101.2 0.1 0.5

Lantibeside Y = 7.42e + 003 X + 4.91e + 003 0.9998 105.7 100.9 0.1 0.5

Phillyrin Y = 6.99e + 003 X + 1.32e + 003 0.9998 92.2 98.3 0.1 0.5

Lantibeside B Y = 6.23e + 003 X − 2.97e + 003 0.9997 104.9 103.7 0.1 0.5

Lantibeside C Y = 7.37e + 003 X − 9.02e + 002 0.9998 91.6 92.5 0.1 0.5

Tibeticoside A Y = 1.21e + 004 X − 5.19e + 003 0.9998 95.4 90.5 0.1 0.5

Styraxjaponoside C Y = 5.42e + 003 X − 4.57e + 002 0.9999 91.7 97.4 0.1 0.5

Sylvatesmin Y = 1.39e + 004 X − 4.99e + 003 0.9998 100.8 96.5 0.1 0.5

(+)-Piperitol Y = 1.10e + 004 X + 3.53e + 003 0.9998 93.2 99.1 0.1 0.5

Horsfieldin Y = 1.31e + 004 X − 3.05e + 003 0.9998 102.3 97.5 0.1 0.5

X = concentration of analytes, Y = peak area

1564

Song Z-H et al. Quantitative Determination of… Planta Med 2011; 77: 1562–1566

Original Papers

content of compound 1was 0.421%, 0.409%, 0.405%, 0.014%, and0.056% for LT 8084, LT 7782, LT 7784, LT 7783, and LT 7785, re-spectively. Interestingly, two samples, LT 7783 and LT 7785, con-tained little compound 1, but it was found that sylvatesmin (8)

concentrations in LT 7783 and LT 7785 were higher than in sam-ples LT 8084, LT 7782, and LT 7784. For instance, in LT 7783 and LT7785, compound 8 was present with concentrations of 0.067%and 0.103%, respectively, but in samples LT 8084, LT 7782, and

Fig. 2 HPLC chromatograms of the mixture ofstandard compounds (A) and of 80% methanolicextracts of Lancea tibetica plant materials [LT 8084(B), LT 7782 (C), LT 7783 (D), LT 7784 (E), and LT7785 (F)] at UV 280 nm: verbascoside (1); lantibe-side (2); phillyrin (3); lantibeside B (4); lantibeside C(5); tibeticoside A (6); styraxjaponoside C (7); sylva-tesmin (8); (+)-piperitol (9); and horsfieldin (10).

Table 2 Intra- and inter-day precision of plant sample LT 8084 assayed under optimized conditions for compounds 1–10 by using the HPLC‑UV method.

Compounds Intra-day (n = 3) Inter-Day (n = 9)

Day 1 Day 2 Day 3

Verbascoside 0.413 (1.53) 0.416 (2.68) 0.433 (0.37) 0.421 (2.77)

Lantibeside 0.138 (1.42) 0.137 (2.40) 0.140 (0.31) 0.138 (1.66)

Phillyrin 0.022 (2.32) 0.022 (2.99) 0.023 (0.38) 0.022 (2.23)

Lantibeside B 0.029 (3.52) 0.027 (3.23) 0.028 (1.44) 0.028 (3.74)

Lantibeside C 0.042 (2.23) 0.043 (2.46) 0.043 (0.84) 0.043 (3.30)

Tibeticoside A 0.020 (4.05) 0.020 (0.30) 0.020 (1.51) 0.020 (2.91)

Styraxjaponoside C 0.003 (3.95) 0.003 (3.58) 0.003 (3.39) 0.003 (4.20)

Sylvatesmin 0.010 (1.31) 0.010 (1.62) 0.010 (2.24) 0.010 (2.58)

(+)-Piperitol 0.003 (1.29) 0.003 (3.40) 0.003 (1.20) 0.003 (2.32)

Horsfieldin 0.003 (2.47) 0.003 (3.89) 0.003 (1.69) 0.003 (2.85)

Values in mg/100mg of plant sample; relative standard deviation (% CV) are given in parentheses

1565

Song Z-H et al. Quantitative Determination of… Planta Med 2011; 77: 1562–1566

Original Papers

LT 7784, its concentrations corresponded to 0.010%, 0.013%, and0.009%, respectively. Compared with compounds 1–3 and 8 inwhich hydroxy groups were protected by a methyl group, the li-gnan composition of compounds 4, 5, 7, 9, and 10, in which twoo-hydroxys connected on phenyl moiety were etherized bymethylene, appeared as only minor components in all samplesexcept for LT 7785 (compound 10, 0.047%) (l" Figs. 1 and 2). Com-pounds 7 and 10 were not detected in samples LT 7782 and LT7783.The HPLC method for chemical analysis of verbascoside and ninelignans was found to be capable of identifying L. tibetica. Thismethod exhibited an excellent performance in terms of sensitiv-ity and baseline separation of all marker compounds without in-terference. The HPLC method was validated and the extractionprocedure was optimized. This method was successfully appliedto quantitative determination of ten compounds in plant samplescollected from different regions in China. Variation of verbasco-side and lignan contents was found in plant samples from differ-ent areas of China, and these contents related to chemical struc-tures were discussed. This method could be useful for identifica-tion of L. tibetica and for quality assurance and quality control ofits related products.

Acknowledgements!

This research is supported in part by “Science Based Authentica-tion of Dietary Supplements” and “Botanical Dietary SupplementResearch” funded by the Food and Drug Administration grantnumbers 5U01FD002071-09 and 1U01FD003871-01, and theUnited States Department of Agriculture, Agricultural ResearchService, Specific Cooperative Agreement No. 58-6408-2-0009.The authors are grateful to Dr. Yue-Guo Zhong from the Chong-qing Research Institute of TCMs for the assistance in collectingand identifying Lancea tibetica samples. We thank Dr. Jon Parcherfor comments on the manuscript.

References1 Commission of Chinese Herbs. Zhonghua Bencao (Volume of TibetanMedicine) (in Chinese), 1st edition. Shanghai: Shanghai Science andTechnology Press; 2002: 192

2 Chen D. Traditional Chinese medicine preparation (internal adminis-tration) for treating pneumonic influenza. Faming Zhuanli ShenqingGongkai Shuomingshu CN 101411790 A 20090422; 2009

3 Baima J. A Tibetan medicinal composition for treating cardiovascularand cerebrovascular diseases, and its preparation method. FamingZhuanli Shenqing Gongkai Shuomingshu CN 1313115 A 20010919;2001

4 Baima J. A Chinese medicinal composition for treating cardiovascularand cerebrovascular diseases, and its preparation method. FamingZhuanli Shenqing Gongkai Shuomingshu CN 1304770 A 20010725;2001

5 Li T, Hao XJ, Gu QQ, ZhuWM.Minor furofurano lignans from the Tibetanherb, Lancea tibetica. Planta Med 2008; 74: 1391–1396

6 Su BN, Zhu QX, Gao K, Yuan CS, Jia ZJ. Lignan and phenylpropanoid gly-cosides from Lancea tibetica and their antitumor activity. Planta Med1999; 65: 558–561

7 Zhang H, Wang M, Zhen R, Zhang S, Chen Y. Chemical constituents ofLancea tibetica Hook. F. et. Thoms. J Lanzhou Univ Nat Sci 1987; 23:156–158

8 Zheng R, Wang M, Zhang H. Study on the chemical constituents of Lan-cea tibetica Hook F. et Thoms. Acta Bot Sin 1985; 27: 402–406

9 Li J, Zheng Y, Zhou H, Su BN, Zheng RL. Differentiation of human gastricadenocarcinoma cell line MGc80-3 induced by verbascoside. PlantaMed 1997; 63: 499–502

10 Zhao C, Qiu R, Zheng R. In vitro antitumor activities of furan lignans.J Lanzhou Univ Nat Sci 2000; 36: 66–68

11 Tang Y, Huang D, Xie M, Xie X, Wan Y, Dai D. Effects of verbascoside andisoverbascoside on proliferation of dendritic cells. Chin Pharm J 2008;43: 1785–1787

12 Sok DE, Cui HS, Kim MR. Isolation and bioactivities of furfuran type lig-nan compounds from edible plants. Recent Pat Food Nutr Agric 2009;1: 87–95

13 Lu YC, Ma HX. Simultaneous determination of ursolic acid and olenolicacid in Lancea tibetica Hook f. et Thoms by HPLC. West China J PharmSci 2009; 24: 302–303

14 Banerji A, Pal S. Constituents of Piper sylvaticum: structure of sylvates-min. J Nat Prod 1982; 45: 672–675

15 ICH harmonised tripartite guideline, validation of analytical proce-dures: text and methodology Q2 (R1), available at http://private.ich.org/cache/compo/363-272-1.html#Q2A. Accessed February 21, 2011

Table 3 Contents of compounds 1–10 in different plant samples (mg/100mg).

Compounds LT 8084 LT 7782 LT 7783 LT 7784 LT 7785

Verbascoside 0.421 (2.77) 0.409 (2.46) 0.014 (3.28) 0.405 (1.60) 0.056 (1.19)

Lantibeside 0.138 (1.66) 0.123 (1.41) 0.021 (2.39) 0.123 (1.51) 0.098 (1.55)

Phillyrin 0.022 (2.23) 0.052 (1.15) 0.004 (1.43) 0.020 (1.77) 0.023 (2.06)

Lantibeside B 0.028 (3.74) 0.007 (3.02) 0.004 (3.81) 0.019 (2.31) 0.006 (3.17)

Lantibeside C 0.043 (3.30) DUL DUL 0.075 (4.60) 0.045 (3.49)

Tibeticoside A 0.020 (2.91) 0.010 (1.05) 0.005 (3.43) 0.018 (2.69) 0.018 (2.91)

Styraxjaponoside C 0.003 (4.20) ND ND 0.004 (4.23) 0.003 (4.18)

Sylvatesmin 0.010 (2.58) 0.013 (1.23) 0.067 (1.17) 0.009 (3.55) 0.103 (2.18)

(+)-Piperitol 0.003 (2.32) DUL 0.011 (2.60) DUL 0.007 (3.73)

Horsfieldin 0.003 (2.85) ND ND 0.003 (3.91) 0.047 (2.24)

ND = not detected; DUL = detected under limits of quantification

1566

Song Z-H et al. Quantitative Determination of… Planta Med 2011; 77: 1562–1566

Original Papers