- Email: [email protected]
Chemical pattern-aided classification to simplify the intricacy of morphological taxonomy of Epimedium species using chromatographic fingerprinting
Journal of Pharmaceutical and Biomedical Analysis 52 (2010) 452–460
Contents lists available at ScienceDirect
Journal of Pharmaceutical and Biomedical Analysis journal homepage: www.elsevier.com/locate/jpba
Chemical pattern-aided classification to simplify the intricacy of morphological taxonomy of Epimedium species using chromatographic fingerprinting Pei-Shan Xie a,∗ , Yu-Zhen Yan a , Bao-Lin Guo b , C.W.K. Lam a , S.H. Chui a , Qiong-Xi Yu a a b
Macau Institute for Applied Research in Medicine and Health, Macau University of Science and Technology, Macau Institute of Medicinal Plant Development, Chinese Academy of Medical Science and Peiking Union Medical College, Beijing, China
a r t i c l e
i n f o
Article history: Received 12 August 2009 Received in revised form 7 January 2010 Accepted 9 January 2010 Available online 20 January 2010 Keywords: Epimedium herb Prenylated flavonoids Bioactive-fraction-aided classification HPLC fingerprinting ABCI fingerprint region
a b s t r a c t Epimedium herb (Yinyanghuo), one of the popular Chinese materia medica, is a multiple species colony of Epimedium genus belonging to Berberidaceae. There are five species of Epimedium that have been officially adopted in Chinese Pharmacopoeia under the same crude drug name ‘Yinyanghuo’ comprising Epimedium brevicornu, E. koreanum, E. sagittatum, E. pubescens, and E. wushanense. In addition, non-official species like E. acuminatum, E. miryanthum and E. leptorrhizum are also mix-used. Frequently, the morphological taxonomical identification is very difficult during on-site inspection for species authentication in the market. Researchers are often bewildered by the multiple species ambiguity when putting this crude drug in use. Referring to the bioactive constituents that are vital for therapeutic efficacy, the key to clarifying the multiple species confusion should rely on analysis of the bioactive composition. It is well known that medicinal Epimedium herbs contain special C-8 prenylated flavonol glycosides which contribute to various bioactivities and the major four, epimedin A (A), epimedin B (B), epimedin C (C) and icariin (I), are unanimously used as bioactive markers for quality control. In this study, HPLC-DAD fingerprinting was performed for investigating the molecular spectrum of various Epimedium species. It was found that the four major flavonoids constitute the middle part of the chromatographic profiles to form a specific region (named as ‘ABCI fingerprint region’) being dominant in the HPLC profiles of all medicinal Epimedium species, and the five official species express five different ‘ABCI’ patterns (different peak: peak ratios). Our study found that the convergent tendency of the ‘ABCI region’ among multiple species of Epimedium could facilitate differentiation of complex commercial samples based on similar bioactive composition should confer similar bioactivities. Merging the different species that possess the same ‘ABCI region’ pattern into the same group can create a simpler bioactive-fraction-aided classification array by clustering the commercial samples into three bioactive ingredients-based fingerprint patterns – ‘E.b. pattern’, ‘E.k. pattern’ and ‘extensive E.w. pattern’. This approach offers the feasibility of characterizing and qualitycontrolling complex samples in the same genus designated under a single herbal drug entity on the premise of possessing the same bioactive ingredients pattern and supported by long-term traditional usage. © 2010 Elsevier B.V. All rights reserved.
1. Introduction The Epimedium herb (Yinyanghuo), one of the popular Chinese materia medica (CMM), is a multiple species colony of genus Epimedium belonging to the Berberidaceae widely distributed in China. The aerial parts of five species have officially been adopted in Chinese Pharmacopoeia under the same crude drug name Yinyanghuo comprising the Short-horned epimedium herb1
∗ Corresponding author. Tel.: +853 28836521. E-mail addresses: [email protected], [email protected] (P.-S. Xie). 1 The English names of all the species despite ‘Tianpingshan mountain epimedium herb’ were cited from Xie Zong-wan “Chinese-Latin-English dictionary of official 0731-7085/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jpba.2010.01.025
(Epimedium brevicornu Maxim.), Korean epimedium herb (E. koreanum Nakai), Sagittate epimedium herb (E. sagittatum (Sieb. & Zucc.) Maxim.), Pubescent epimedium herb (E. pubescens Maxim.) and Wushan mountain epimedium herb (E. wushanense T.S. Ying) [1]. In addition to the five official species, the aerial parts of Acuminate epimedium herb (E. acuminatum) and Tianpingshan mountain epimedium herb (E. miryanthum Stearn) have also been mix-used in the market. From the plant taxonomy point of view, Epimedium is an intricate genus for species classification [2,3]. Morphological
names of Chinese materia medica” 2004, Beijing, Beijing Science and Technology Publication House, 7-1205–1218. According to ‘China Flora’ Vol. 29, pp. 262–330, Short-horned epimedium herb should be ‘E. brevicornu’, not ‘E. brevicornum’.
P.-S. Xie et al. / Journal of Pharmaceutical and Biomedical Analysis 52 (2010) 452–460
453
2.2. Chemical reference substances (CRS) Epimedin A, epimedin B, epimedin C and icariin were provided by Professor. B. L. Guo or purchased from the National Institute for Control of Pharmaceutical and Biological Products, Beijing, China. 2.3. Chemicals Acetonitrile and methanol (chromatographic grade) were purchased from Merck Co. Ltd., Germany, and ethanol and glacial acetic acid (analytical grade) from Guangzhou Chemicals, Guangdong, China. Deionized water was prepared by using a Milli-Q water purification system (Millipore Corp., Bradford, MA, USA). 2.4. Preparation of sample solution
Fig. 1. Chemical structure of major prenylated flavonoids. Rha, rhamnose; Glc, glucose; Xyl, xylose.
variation of leaves within species and resemblance among species under the genus Epimedium makes it difficult to identify species properly; some subtle differentiation even baffled the taxonomist [4]. In certain cases the morphological taxonomical identification between species is too difficult during on-site inspection for species authentication among commercial samples in the market. Chemical and pharmacological experts and clinical practitioners are often bewildered by such species confusion when putting this crude drug in use. With such species uncertainty and mix-use, research studies or clinical observation cannot generate reliable results. There is a consensus that a single herbal drug name should exclusively designate only one species. However, a number of Chinese CMMs are historically composed of multiple species involved under a single entity. Given that the bioactive constituents are vital for therapeutic efficacy, the key to clarifying the multiple species confusion could be analysis of the bioactive composition. It is well known that the medicinal ‘Epimedium herb’ contain special C-8 prenylated flavonol glycosides which contribute to the immunomodulatory effect, osteoblastic proliferative activity and alleged sex hormone functions. Four major compounds epimedin A (A), epimedin B (B), epimedin C (C) and icariin (I) have been used as bioactive markers for quality control [5–11] (Fig. 1). Therefore selecting these major prenylated flavonoids as specific ingredients in Epimedium herb for further study is rational. In this study, high performance liquid chromatography coupled with diode array detector (HPLC-DAD) fingerprinting analysis was performed for studying the molecular spectrum of various Epimedium species. The aim was to investigate the possibility of applying such technique for clarifying the perplex problem of multiple species of Epimedium co-existing in the market.
Fifty millilitres of 50% ethanol was added onto 0.1 g of pulverized leaves and refluxed on a 100 ◦ C water bath for 1 h. The filtrate was evaporated to dryness under reduced pressure. The residue was dissolved in 30% methanol, adjusted to 5 ml and, after standing overnight at about 4 ◦ C in a refrigerator, filtered through a 0.45 m membrane. The final filtrate constituted the sample solution for analysis. 2.5. Preparation of CRS solution Five milligrams of each of epimedin A, epimedin B, epimedin C and icariin was dissolved in 50 ml of methanol. 2.6. HPLC equipment and conditions The Agilent 1100 HPLC system with DAD and Agilent Chemstation software (Agilent, Palo Alto, CA, USA) and Chromatographic Fingerprinting Evaluation software (Chromafinger Solution 2005 software developed by Chromap Institute of Herbal Medicine Research, Zhuhai, China) were used in this study. The HPLC conditions were as follows: column, Agilent Zorbax Ecllipse Plus-C18 (particle size 5 m, diameter 4.6 mm, length 250 mm); mobile phase, methanol–acetonitrile–0.5% acetic acid. The linear gradient elution was as in Table 1 with flow rate of 1.0 ml/min. Column temperature: 20 ◦ C. Detection wavelength: 270 nm. Sample injection: 20 l. 2.7. Data analysis At first, the four major C-8 prenylated flavonoids in the HPLC profiles were attributed by comparison with the CRS – epimedin A, epimedin B, epimedin C and icariin. The fingerprint common patterns were generated by inputting the original data suits of all authenticated samples acquired from the HPLC workstation to the Fingerprint Solution software, the followed data processing simulated all the HPLC profiles one by one, and exhibited on the
2. Materials and methods 2.1. Crude drugs Eighty-one batches of samples of ‘Epimedium herb’ (Yin Yang Huo) were collected from Guizhou, Henan, Sichuan, Shaanxi, Anhui, Guangxi, Jilin and Liaoning provinces over western, eastern and north-eastern China. Some of them were commercial samples. Species identification was conducted by Professor B.L. Guo of the Institute of Medicinal Plant Development, Beijing, China. The vouchers of the authenticated samples were deposited in Professor Guo’s laboratory of the Institute of Medicinal Plant Development.
Table 1 Elution gradient program. Time (min)
Methanol (%)
Acetonitrile (%)
0.5% acetic acid (%)
0 30 45 45.01 60 65 85 90 100
0 0 0 11 11 4 4 0 0
12 25 25 23.5 23.5 35 35 50 50
88 75 75 65.5 65.5 61 61 50 50
454
P.-S. Xie et al. / Journal of Pharmaceutical and Biomedical Analysis 52 (2010) 452–460
computer’s screen, aligning manually the major peaks by clicking the peaks apex in order to ensure correct recognition, the computer-simulated profile by averaging all the profiles data produced serves as the common pattern of the species. The similarity can be calculated and expressed by correlative coefficient, as well as principal component analysis (PCA) was also carried out. 2.8. Methodology validation Validation of the HPLC fingerprint method was carried out under the regulation of Chinese Pharmacopoeia Commission [12]. Accuracy as represented by the relative standard deviation (RSD) of peak retention time of icariin in 5 runs of same sample solution was 2.2%. Reproducibility values as calculated from the RSD of peak area and peak retention time of icariin in 5-sample solution prepared from the same sample was 1.6% and 0.19%, respectively. Stability as assessed by the RSD value of peak area of icariin in 0, 1, 2, 4, 8 and 24 h averaged 1.8%. All of these results were in accordance with the requirements of the Chinese Pharmacopoeia Commission.
3. Results 3.1. Establishment of the common pattern of the five official species in Chinese Pharmacopoeia and designation of ‘ABCI fingerprint’ region 3.1.1. The common pattern of the official species of Epimedium The common patterns of the HPLC profiles of the five official Epimedium species leaves, namely E. koreanum (n = 5), E. brevicornu (n = 6), E. pubescence (n = 9), E. wushanense (n = 13) and E. sagittatum (n = 8) showed the dominant specific region between the retention times of about 37–50 min (under the experimental condition in this study). It consisted of 5–7 peaks involving the four essential C-8 prenylated flavonoids – epimedin A (A) (peak 2), epimedin B (B) (peak 3), epimedin C (C) (peak 4) and icariin (I) (peak 6) as well as 1–3 minor unknown flavonoids peaks in between. It was named as ‘ABCI fingerprint region’ (‘ABCI region’ for short) [3]. This region in the chromatographic profile should be overwhelming in the qualitative identification and bioactivity-oriented QA/QC assessment of Epimedium (Fig. 2).
Fig. 2. The HPLC profiles’ common patterns of the leaves of five official Epimedium spp. The dominant specific ‘ABCI fingerprint regions indicated by red frame, cf. Fig. 3, 4.
P.-S. Xie et al. / Journal of Pharmaceutical and Biomedical Analysis 52 (2010) 452–460
455
Fig. 3. ABCI region’ (TR 37 min – 50 min) of HPLC profiles of the official five Epimedium species (cf. Fig. 8). Peaks (1), (5), (7): unknown flavonoids; (2) epimedin A; (3) epimedin B; (4) epimedin C; (6) icariin.
3.1.2. Designation of specific ‘ABCI region’ fingerprints in the HPLC profiles of Epimedium species Observing the ‘ABCI fingerprint region’ in the HPLC profiles of all the samples, a composition fluctuation was found among the five official species, particularly simultaneous regular elevation of one peak and decline of the other that occurred among peaks of epimedin C, icariin and epimedin B, constituting different characteristic fingerprints. Thus reading the absorbance abundance of the peaks alongside the y-axis, five species contributed five patterns and constituted the independent characters of the five official species. Briefly, the feature of E. koreanum was icariin (peak 6) being the strongest peak, named as ‘E.k. pattern’; epimedin B (peak 3) tied with icariin (peak 6) constituted the distinct feature of E. brevicornu, named as ‘E.b. pattern’; the sole epimedin C peak (peak 4) dominated the ‘ABCI region’ in E. wushanense, named as ‘E.w. pattern’; the prevailing peaks of epimedin C (peak 4) followed by icariin (peak 6) of ‘ABCI region’ in E. sagittatum was named as ‘E.s. pattern’ and E. pubescens, named as ‘E.p. pattern’. However, peak 4 was generally higher than peak 6 in ‘E.s. pattern’ and vice versa in ‘E.p. pattern’. In terms of expression of the discrepancy by formulae, using small letter symbolizes weak peak, capital letter represents medium–stronger peak and capital bold letter indicates the strongest peak, the formula for ‘E.k. pattern’: ‘a b c I’; for ‘E.b.
pattern’: ‘a B c I’; for ‘E.p pattern’: ‘a b C I’ or ‘a b C I’; for ‘E.s. pattern’: ‘a b C I’ and for ‘E.w. pattern’: ‘a b C i’ (Figs. 3 and 4). 3.2. Principal component analysis (PCA) of 46 samples of Epimedium species Taking HPLC profile of E. koreanum as reference, five patterns were categorized clearly. The first principal component (PC1) was epimedin C (score 0.68), and the second (PC2) was epimedin B and icariin (scores 0.70 and 0.69, respectively). These meant that the most influential factors for different Epimedium species were located in “ABCI region”. With regard to the “shape” of ABCI region, it seems that the hydrophilic component of epimedin C which contains two moieties of rhamnose at C-3 position is the dominant in the ‘E.w. pattern’ (a b C i), ‘E.s. pattern’ (a b C I) and ‘E.p. pattern’ (a b C I), icariin (one rhamnose at C-3 position and almost insoluble in water) dominated in ‘E.k. pattern’ (a b c I). And the dominant peak in ‘E.b. pattern’ is epimedin B (a B c I) (Fig. 5). 3.3. Survey of non-official Epimedium species in the market Plant taxonomical studies have reported that there are 47 species under the genus Epimedium distributed in China [3]. How-
456
P.-S. Xie et al. / Journal of Pharmaceutical and Biomedical Analysis 52 (2010) 452–460
Fig. 4. Histogram of ‘ABCI fingerprint region’ of HPLC profiles of the five official Epimedium species based on the contents determined by HPLC quantitative analysis. (E.b.) E.brevicornum; (E.s.) E.sagittatu ; (E.w.) E. wushanense; (E.k.) E. koreanum; (E.p.): E. pubescens.
ever, most species do not form commercial resources for medicinal use. Besides, several rich resources of non-official species have also been sold in some local crude drug markets in east China, north-west China and south China, blending contingently with the official species spread over several herbal drug markets and manufacturer’s storehouses [2]. In this study, the leaves of eight non-official species: E. acuminatum Franch (n = 4), E. myrianthum Stearn (n = 5), E. elongatum Komarov (n = 2), E. dolichostemon Stearn (n = 1), E. mikinorii Stearn (n = 1), E. leptorrhizum Stearn (n = 3), E. davidii Franch (n = 2) and E. membranaceum Franch (n = 2) were analyzed. A comparative study on HPLC fingerprinting among those species with the five official species showed the following: (1) Convergence of ‘ABCI region’ pattern among species of Epimedium. It was found that different species possibly shared the same ‘ABCI region’ pattern. Comparative observation showed that E. acuminatum, E. myrianthum and E. mikinorii samples displayed the same feature of ‘E.s. pattern’ (a b C I), while E. dolichostemon
and E. myrianthum and its var. jianheanse showed ‘E.w. pattern’ composition (a b C i), and E. elongatum presented ‘E.p. pattern’ (a b C I) (Fig. 6). By all account, the existence of different species sharing the same bioactivity related to ‘ABCI pattern’ indicated the possibility of such species having similar biological activities and therapeutic efficacy. (2) Dissimilation of ‘ABCI region’ pattern within species of Epimedium. In contrast, the chromatographic spectra with some individual Epimedium species were seen to deviate out of the original pattern possibly due to genetic or environmental influence. This is illustrated by E. koreanum and E. sagittatum. Compared with the normal samples, some samples collected from different habitats or in different seasons were found to exhibit almost blank “ABCI region” in their HPLC profiles (Fig. 7). This dissimilation suggests that these herbs cannot contribute the same efficacy as the normal ones. On the other hand, several samples of E. brevicornu exhibited ‘E.p. pattern’ and ‘E.k. pattern’. Some samples of E. davidii possessed ‘E.k. pattern’ and two samples of E. membranaceum showed ‘E.p. pattern’. (3) Lack of ‘ABCI region’ in E. leptorrhizum. The aerial part of E. leptorrhizum has also been used as medicinal ‘Epimedium herb’ in the local market. However, the HPLC profile obtained in our study showed lack of ‘ABCI region’ or only a trace amount. This seemed not an occasional occurrence and similar observation on this species has also been published [3,5,13]. Considering this species is widely distributed and all the samples contained no or trace amount of ABCI, genetic alteration has probably occurred. Accordingly, E. leptorrhizum should be deleted from the list of medicinal Epimedium herb. 4. Discussion Species authentication of herbal drugs is a prerequisite for ensuring their quality for clinical administration, pharmacological study, and quality control. However, not infrequently, multiple species are encountered in a single crude drug in CMM. This is a perplex historical problem because China has a vast territory and diverse plant varieties and regional custom of herbal crude drug usage. Designating any of the single species alone and depriving
Fig. 5. The plot of Principle Component Analysis (PC1 – PC2) of 46 samples of Epimedium spp. Accumulated mean square: PC1 = 70.5%; PC2 = 86.9%.
P.-S. Xie et al. / Journal of Pharmaceutical and Biomedical Analysis 52 (2010) 452–460
the others for the sake of ‘purifying’ the resource of such crude drug is not rational. Yinyanghuo (Epimedium herb) is a typical example. Research studies of Epimedium herbs have shown the convergence in composition of active ingredients among species
457
(Fig. 6A) facilitating their re-classification. On the other hand, dissimilation of active components within species (Fig. 7) results in some abnormal patterns. Some examples have broken down the boundaries of taxonomical species exposing the composition of
Fig. 6. Convergence tendency of ‘ABCI region’ in HPLC profiles of different species. E. acuminatum, E. myrianthum, E. mikinorii have similar pattern of E. sagittatum (a b C I); E. longatum has E. pubescens pattern (E.p. pattern: a b C I).
458
P.-S. Xie et al. / Journal of Pharmaceutical and Biomedical Analysis 52 (2010) 452–460
Fig. 6. (Continued ).
ABCI region out of its original pattern (leaves of E. brevicornu) merged to other patterns. Both findings provide an alternative solution of simplifying the complex morphological taxonomy during on-sight inspection and for quality assurance in the manufacturing process. According to the composition of the “ABCI fingerprint region”, all of the samples can be categorized into five patterns. Further converging the five patterns can be reduced to three patterns – ‘extensive E.w. pattern’ (including ‘E.w.’, ‘E.p.’, and ‘E.s. patterns’), ‘E.b. pattern’ and ‘E.k. pattern’ (Fig. 8) expressed by
Cluster Analysis and Classification Index (see below). Therefore it paved a practical way to alleviate the confusion of co-existence of multiple taxonomical species. For instance, if E. wushanense was selected for usage originally in a prescription or product, then E. pubescens, E. sagittatum, and E. acuminatum can be suggested as alternative entities based on sharing the same ‘ABCI region’ pattern. Classification Index of leaves of Epimedium herb based on bioactive ingredients composition (ABCI fingerprint region).
Fig. 7. HPLC profiles of E. sagittatum (leaves) from different habitats (top: grown in Guizhou province, southwest China; lower: three samples were grown in Anhui province, east China).
P.-S. Xie et al. / Journal of Pharmaceutical and Biomedical Analysis 52 (2010) 452–460
459
Fig. 8. Dendrogram of HPLC fingerprint cluster analysis. Converging all samples into three categories based on the composition of “ABCI region”. E.k. pattern: dominant peak is icariin (peak 6); extensive E.w pattern: epimedin C (peak 4); E.b pattern: epimedin B (peak 3).
1. “ABCI region” distinct 1.1. Epimedin C (peak 4) dominant (extensive E.w. pattern) 1.1.1. Epimedin C (peak 4) dominant exclusively (E.w. pattern) (a b C 5 i) E. wushanense E. dolichostemon E. myrianthum E. myrianthum var. jianheanse 1.1.2. Epimedin C (peak 4)>/
Consistent content of the major bioactive C-8 prenylated flavonoids is another supporting factor for breaking down the boundary of the taxonomical species to merge the commodities of medicinal Epimedium herb into the same group based on the resembling ‘ABCI region’. Comparing the integrated peak areas, the ratio of ‘ABCI’ amount, and total flavonoids amount (ABCI:total) of the samples with the same ‘pattern’, it is possible to estimate the relative amount of the main bioactive ingredients for quality evaluation. A set of unpublished quantitative data on ‘ABCI’ acquired by quantitative HPLC analysis in our lab demonstrated the concordance between precisely determined content and the estimation by integrated HPLC peaks (Fig. 9). Some papers have been published on HPLC fingerprinting analysis [3,5,14,15] of Epimedium species. All of them focus on the “ABCI region” as the characteristics of HPLC profiles of Epimedium herbs for phytotaxonomy and quality analysis. As a typical example, B.L. Guo et al. attempted to delineate the subtle variation of the “ABCI peak group” of 35 species of Epimedium for differentiating the species more strictly. All the samples tested were divided into four main types and nine subtypes [5] to compile with the complex morphological taxonomic system devised by W.T. Stearn [16,17]. Plant taxonomy needs divergence strategy to distinguish the visual variance of the appearance in a subtle way to define the
Fig. 9. Positive correlation between the quantitative data (above) and the integrated peaks area (lower) of total ‘ABCI’. It showed the reconcilability of precise determined contents and relative quantities from the integrated peaks area of HPLC profiles. The latter is rough but effective and easily available.
uniqueness of species. The complex classification verified the diversity of the botanical kingdom. However, for solving the practical difficulty caused by multiple species herbal medicine commodities coexisted in the Chinese herbal drug market, divergence approach is obviously not the right choice. We need reverse thinking. Converging the different species that possess the same HPLC fingerprint pattern towards one group will facilitate the efficacious use of multiple species of Chinese herbal medicines like the case of Epimedium herb. Convergence or divergence depends on the final purposes. Acknowledgement We thank the Science and Technology Development Fund of Macao Government for financial support (Grant Number: 045/2005/A). References [1] Chinese Pharmacopoeia Commission, Chinese Pharmacopoeia I, Chemical Industry Press, Beijing, 2005, p. 229. [2] B.L. Guo, P.G. Xiao, Comment on main species of Herba Epimedii, Chin. J. Chin. Mater. Med. 28 (2003) 303–307. [3] Li Pei-kuan, B.L. Guo, W.H. Huang, Studies on fingerprinting and identification of main species of Herba Epimedii, Chin. J. Chin. Mater. Med. 33 (2008) 1662–1668. [4] B.L. Guo, P.G. Xiao, The flavonoids in Epimedium L. and their taxonomic significance, Acta Phytotaxon. 37 (1999) 228–243.
460
P.-S. Xie et al. / Journal of Pharmaceutical and Biomedical Analysis 52 (2010) 452–460
[5] B.L. Guo, L.K. Pei, P.G. Xiao, Further research on taxonomic significance of flavonoids in Epimedium (Berberidaceae), J. Sys. Evol. 46 (2008) 874–885. [6] S.P. Wong, P. Shen, L. Lee, J. Li, E.L. Yong, Pharmacokinetics of prenylflavonoids and correlations with the dynamics of estrogen action in sera following ingestion of a standardized Epimedium extract, JPBA 50 (2009) 216–223. [7] Z.X. Shou, L. Shen, Y.P. Yang, J. Xie, P.Q. Zhou, L. Gao, Effects of epimedium total flavonoids on bone mineral density and bone metabolism-related indices in primary osteoporosis, J. Clin. Rehabil. Tissue Eng. Res. 13 (2009) 2191–2195. [8] H. Wu, E.J. Lien, L.L. Lien, Chemical and pharmacological investigations of Epimedium species: a survey, Prog. Drug Res. 60 (2003) 1–57. [9] M.K. Lee, Y.J. Choi, S.H. Sung, D.I. Shin, J.W. Kim, Y.C. Kim, Antihepatotoxic activity of icariin, a major constituent of Epimedium koreanum, Planta Med. 61 (1995) 523–526. [10] S.P. Yap, P. Shen, M.S. Butler, Y. Gong, C.J. Loy, E.L. Yong, New estrogenic prenylflavone from Epimedium brevicornum inhibits the growth of breast cancer cells, Planta Med. 71 (2005) 114–119.
[11] K.M. Chen, B.F. Ge, H.P. Ma,.Y. Liu,.H. Bai, Y. Wang, Icariin, a flavonoid from the herb Epimedium enhances the osteogenic differentiation of rat primary bone marrow stromal cells, Pharmazie 60 (2005) 939–942. [12] Chinese Pharmacopoeia Commission, Guidance on chromatographic fingerprint experimental research of TCM Injection (in Chinese) (2002). [13] L.K. Pei, W.H. Huang, T.G. He, B.L. Guo, Systematic studies on quality of main species of ‘herba epimedii’, Chin. J. Chin. Mater. Med. 32 (2007) 2217–2222. [14] C.Y. Huang, L.H. Zhao, L.H. Mei, Chin. Studies on the HPLC fingerprint of Epimedium brevicornum crude drug, J. Nat. Med. 1 (2003) 146–148. [15] L.X. Wang, C.Z. Wang, X.D. Geng, Studies on the RPLC fingerprints and quality evaluation of Epimedium L, Crude drugs. Acta Chim. Sin. 64 (2006) 551–555. [16] W.T. Stearn, Epimedium and Vancouveria (Berberidaceae), a monograph, J. Linn. Soc. 51 (1938) 409–535. [17] W.T. Stearn, The Genus Epimedium and other Herbaceous Berberidaceae Including the Genus Podophyllium, Timber Press, Oregon, 2002.
Contents lists available at ScienceDirect
Journal of Pharmaceutical and Biomedical Analysis journal homepage: www.elsevier.com/locate/jpba
Chemical pattern-aided classification to simplify the intricacy of morphological taxonomy of Epimedium species using chromatographic fingerprinting Pei-Shan Xie a,∗ , Yu-Zhen Yan a , Bao-Lin Guo b , C.W.K. Lam a , S.H. Chui a , Qiong-Xi Yu a a b
Macau Institute for Applied Research in Medicine and Health, Macau University of Science and Technology, Macau Institute of Medicinal Plant Development, Chinese Academy of Medical Science and Peiking Union Medical College, Beijing, China
a r t i c l e
i n f o
Article history: Received 12 August 2009 Received in revised form 7 January 2010 Accepted 9 January 2010 Available online 20 January 2010 Keywords: Epimedium herb Prenylated flavonoids Bioactive-fraction-aided classification HPLC fingerprinting ABCI fingerprint region
a b s t r a c t Epimedium herb (Yinyanghuo), one of the popular Chinese materia medica, is a multiple species colony of Epimedium genus belonging to Berberidaceae. There are five species of Epimedium that have been officially adopted in Chinese Pharmacopoeia under the same crude drug name ‘Yinyanghuo’ comprising Epimedium brevicornu, E. koreanum, E. sagittatum, E. pubescens, and E. wushanense. In addition, non-official species like E. acuminatum, E. miryanthum and E. leptorrhizum are also mix-used. Frequently, the morphological taxonomical identification is very difficult during on-site inspection for species authentication in the market. Researchers are often bewildered by the multiple species ambiguity when putting this crude drug in use. Referring to the bioactive constituents that are vital for therapeutic efficacy, the key to clarifying the multiple species confusion should rely on analysis of the bioactive composition. It is well known that medicinal Epimedium herbs contain special C-8 prenylated flavonol glycosides which contribute to various bioactivities and the major four, epimedin A (A), epimedin B (B), epimedin C (C) and icariin (I), are unanimously used as bioactive markers for quality control. In this study, HPLC-DAD fingerprinting was performed for investigating the molecular spectrum of various Epimedium species. It was found that the four major flavonoids constitute the middle part of the chromatographic profiles to form a specific region (named as ‘ABCI fingerprint region’) being dominant in the HPLC profiles of all medicinal Epimedium species, and the five official species express five different ‘ABCI’ patterns (different peak: peak ratios). Our study found that the convergent tendency of the ‘ABCI region’ among multiple species of Epimedium could facilitate differentiation of complex commercial samples based on similar bioactive composition should confer similar bioactivities. Merging the different species that possess the same ‘ABCI region’ pattern into the same group can create a simpler bioactive-fraction-aided classification array by clustering the commercial samples into three bioactive ingredients-based fingerprint patterns – ‘E.b. pattern’, ‘E.k. pattern’ and ‘extensive E.w. pattern’. This approach offers the feasibility of characterizing and qualitycontrolling complex samples in the same genus designated under a single herbal drug entity on the premise of possessing the same bioactive ingredients pattern and supported by long-term traditional usage. © 2010 Elsevier B.V. All rights reserved.
1. Introduction The Epimedium herb (Yinyanghuo), one of the popular Chinese materia medica (CMM), is a multiple species colony of genus Epimedium belonging to the Berberidaceae widely distributed in China. The aerial parts of five species have officially been adopted in Chinese Pharmacopoeia under the same crude drug name Yinyanghuo comprising the Short-horned epimedium herb1
∗ Corresponding author. Tel.: +853 28836521. E-mail addresses: [email protected], [email protected] (P.-S. Xie). 1 The English names of all the species despite ‘Tianpingshan mountain epimedium herb’ were cited from Xie Zong-wan “Chinese-Latin-English dictionary of official 0731-7085/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jpba.2010.01.025
(Epimedium brevicornu Maxim.), Korean epimedium herb (E. koreanum Nakai), Sagittate epimedium herb (E. sagittatum (Sieb. & Zucc.) Maxim.), Pubescent epimedium herb (E. pubescens Maxim.) and Wushan mountain epimedium herb (E. wushanense T.S. Ying) [1]. In addition to the five official species, the aerial parts of Acuminate epimedium herb (E. acuminatum) and Tianpingshan mountain epimedium herb (E. miryanthum Stearn) have also been mix-used in the market. From the plant taxonomy point of view, Epimedium is an intricate genus for species classification [2,3]. Morphological
names of Chinese materia medica” 2004, Beijing, Beijing Science and Technology Publication House, 7-1205–1218. According to ‘China Flora’ Vol. 29, pp. 262–330, Short-horned epimedium herb should be ‘E. brevicornu’, not ‘E. brevicornum’.
P.-S. Xie et al. / Journal of Pharmaceutical and Biomedical Analysis 52 (2010) 452–460
453
2.2. Chemical reference substances (CRS) Epimedin A, epimedin B, epimedin C and icariin were provided by Professor. B. L. Guo or purchased from the National Institute for Control of Pharmaceutical and Biological Products, Beijing, China. 2.3. Chemicals Acetonitrile and methanol (chromatographic grade) were purchased from Merck Co. Ltd., Germany, and ethanol and glacial acetic acid (analytical grade) from Guangzhou Chemicals, Guangdong, China. Deionized water was prepared by using a Milli-Q water purification system (Millipore Corp., Bradford, MA, USA). 2.4. Preparation of sample solution
Fig. 1. Chemical structure of major prenylated flavonoids. Rha, rhamnose; Glc, glucose; Xyl, xylose.
variation of leaves within species and resemblance among species under the genus Epimedium makes it difficult to identify species properly; some subtle differentiation even baffled the taxonomist [4]. In certain cases the morphological taxonomical identification between species is too difficult during on-site inspection for species authentication among commercial samples in the market. Chemical and pharmacological experts and clinical practitioners are often bewildered by such species confusion when putting this crude drug in use. With such species uncertainty and mix-use, research studies or clinical observation cannot generate reliable results. There is a consensus that a single herbal drug name should exclusively designate only one species. However, a number of Chinese CMMs are historically composed of multiple species involved under a single entity. Given that the bioactive constituents are vital for therapeutic efficacy, the key to clarifying the multiple species confusion could be analysis of the bioactive composition. It is well known that the medicinal ‘Epimedium herb’ contain special C-8 prenylated flavonol glycosides which contribute to the immunomodulatory effect, osteoblastic proliferative activity and alleged sex hormone functions. Four major compounds epimedin A (A), epimedin B (B), epimedin C (C) and icariin (I) have been used as bioactive markers for quality control [5–11] (Fig. 1). Therefore selecting these major prenylated flavonoids as specific ingredients in Epimedium herb for further study is rational. In this study, high performance liquid chromatography coupled with diode array detector (HPLC-DAD) fingerprinting analysis was performed for studying the molecular spectrum of various Epimedium species. The aim was to investigate the possibility of applying such technique for clarifying the perplex problem of multiple species of Epimedium co-existing in the market.
Fifty millilitres of 50% ethanol was added onto 0.1 g of pulverized leaves and refluxed on a 100 ◦ C water bath for 1 h. The filtrate was evaporated to dryness under reduced pressure. The residue was dissolved in 30% methanol, adjusted to 5 ml and, after standing overnight at about 4 ◦ C in a refrigerator, filtered through a 0.45 m membrane. The final filtrate constituted the sample solution for analysis. 2.5. Preparation of CRS solution Five milligrams of each of epimedin A, epimedin B, epimedin C and icariin was dissolved in 50 ml of methanol. 2.6. HPLC equipment and conditions The Agilent 1100 HPLC system with DAD and Agilent Chemstation software (Agilent, Palo Alto, CA, USA) and Chromatographic Fingerprinting Evaluation software (Chromafinger Solution 2005 software developed by Chromap Institute of Herbal Medicine Research, Zhuhai, China) were used in this study. The HPLC conditions were as follows: column, Agilent Zorbax Ecllipse Plus-C18 (particle size 5 m, diameter 4.6 mm, length 250 mm); mobile phase, methanol–acetonitrile–0.5% acetic acid. The linear gradient elution was as in Table 1 with flow rate of 1.0 ml/min. Column temperature: 20 ◦ C. Detection wavelength: 270 nm. Sample injection: 20 l. 2.7. Data analysis At first, the four major C-8 prenylated flavonoids in the HPLC profiles were attributed by comparison with the CRS – epimedin A, epimedin B, epimedin C and icariin. The fingerprint common patterns were generated by inputting the original data suits of all authenticated samples acquired from the HPLC workstation to the Fingerprint Solution software, the followed data processing simulated all the HPLC profiles one by one, and exhibited on the
2. Materials and methods 2.1. Crude drugs Eighty-one batches of samples of ‘Epimedium herb’ (Yin Yang Huo) were collected from Guizhou, Henan, Sichuan, Shaanxi, Anhui, Guangxi, Jilin and Liaoning provinces over western, eastern and north-eastern China. Some of them were commercial samples. Species identification was conducted by Professor B.L. Guo of the Institute of Medicinal Plant Development, Beijing, China. The vouchers of the authenticated samples were deposited in Professor Guo’s laboratory of the Institute of Medicinal Plant Development.
Table 1 Elution gradient program. Time (min)
Methanol (%)
Acetonitrile (%)
0.5% acetic acid (%)
0 30 45 45.01 60 65 85 90 100
0 0 0 11 11 4 4 0 0
12 25 25 23.5 23.5 35 35 50 50
88 75 75 65.5 65.5 61 61 50 50
454
P.-S. Xie et al. / Journal of Pharmaceutical and Biomedical Analysis 52 (2010) 452–460
computer’s screen, aligning manually the major peaks by clicking the peaks apex in order to ensure correct recognition, the computer-simulated profile by averaging all the profiles data produced serves as the common pattern of the species. The similarity can be calculated and expressed by correlative coefficient, as well as principal component analysis (PCA) was also carried out. 2.8. Methodology validation Validation of the HPLC fingerprint method was carried out under the regulation of Chinese Pharmacopoeia Commission [12]. Accuracy as represented by the relative standard deviation (RSD) of peak retention time of icariin in 5 runs of same sample solution was 2.2%. Reproducibility values as calculated from the RSD of peak area and peak retention time of icariin in 5-sample solution prepared from the same sample was 1.6% and 0.19%, respectively. Stability as assessed by the RSD value of peak area of icariin in 0, 1, 2, 4, 8 and 24 h averaged 1.8%. All of these results were in accordance with the requirements of the Chinese Pharmacopoeia Commission.
3. Results 3.1. Establishment of the common pattern of the five official species in Chinese Pharmacopoeia and designation of ‘ABCI fingerprint’ region 3.1.1. The common pattern of the official species of Epimedium The common patterns of the HPLC profiles of the five official Epimedium species leaves, namely E. koreanum (n = 5), E. brevicornu (n = 6), E. pubescence (n = 9), E. wushanense (n = 13) and E. sagittatum (n = 8) showed the dominant specific region between the retention times of about 37–50 min (under the experimental condition in this study). It consisted of 5–7 peaks involving the four essential C-8 prenylated flavonoids – epimedin A (A) (peak 2), epimedin B (B) (peak 3), epimedin C (C) (peak 4) and icariin (I) (peak 6) as well as 1–3 minor unknown flavonoids peaks in between. It was named as ‘ABCI fingerprint region’ (‘ABCI region’ for short) [3]. This region in the chromatographic profile should be overwhelming in the qualitative identification and bioactivity-oriented QA/QC assessment of Epimedium (Fig. 2).
Fig. 2. The HPLC profiles’ common patterns of the leaves of five official Epimedium spp. The dominant specific ‘ABCI fingerprint regions indicated by red frame, cf. Fig. 3, 4.
P.-S. Xie et al. / Journal of Pharmaceutical and Biomedical Analysis 52 (2010) 452–460
455
Fig. 3. ABCI region’ (TR 37 min – 50 min) of HPLC profiles of the official five Epimedium species (cf. Fig. 8). Peaks (1), (5), (7): unknown flavonoids; (2) epimedin A; (3) epimedin B; (4) epimedin C; (6) icariin.
3.1.2. Designation of specific ‘ABCI region’ fingerprints in the HPLC profiles of Epimedium species Observing the ‘ABCI fingerprint region’ in the HPLC profiles of all the samples, a composition fluctuation was found among the five official species, particularly simultaneous regular elevation of one peak and decline of the other that occurred among peaks of epimedin C, icariin and epimedin B, constituting different characteristic fingerprints. Thus reading the absorbance abundance of the peaks alongside the y-axis, five species contributed five patterns and constituted the independent characters of the five official species. Briefly, the feature of E. koreanum was icariin (peak 6) being the strongest peak, named as ‘E.k. pattern’; epimedin B (peak 3) tied with icariin (peak 6) constituted the distinct feature of E. brevicornu, named as ‘E.b. pattern’; the sole epimedin C peak (peak 4) dominated the ‘ABCI region’ in E. wushanense, named as ‘E.w. pattern’; the prevailing peaks of epimedin C (peak 4) followed by icariin (peak 6) of ‘ABCI region’ in E. sagittatum was named as ‘E.s. pattern’ and E. pubescens, named as ‘E.p. pattern’. However, peak 4 was generally higher than peak 6 in ‘E.s. pattern’ and vice versa in ‘E.p. pattern’. In terms of expression of the discrepancy by formulae, using small letter symbolizes weak peak, capital letter represents medium–stronger peak and capital bold letter indicates the strongest peak, the formula for ‘E.k. pattern’: ‘a b c I’; for ‘E.b.
pattern’: ‘a B c I’; for ‘E.p pattern’: ‘a b C I’ or ‘a b C I’; for ‘E.s. pattern’: ‘a b C I’ and for ‘E.w. pattern’: ‘a b C i’ (Figs. 3 and 4). 3.2. Principal component analysis (PCA) of 46 samples of Epimedium species Taking HPLC profile of E. koreanum as reference, five patterns were categorized clearly. The first principal component (PC1) was epimedin C (score 0.68), and the second (PC2) was epimedin B and icariin (scores 0.70 and 0.69, respectively). These meant that the most influential factors for different Epimedium species were located in “ABCI region”. With regard to the “shape” of ABCI region, it seems that the hydrophilic component of epimedin C which contains two moieties of rhamnose at C-3 position is the dominant in the ‘E.w. pattern’ (a b C i), ‘E.s. pattern’ (a b C I) and ‘E.p. pattern’ (a b C I), icariin (one rhamnose at C-3 position and almost insoluble in water) dominated in ‘E.k. pattern’ (a b c I). And the dominant peak in ‘E.b. pattern’ is epimedin B (a B c I) (Fig. 5). 3.3. Survey of non-official Epimedium species in the market Plant taxonomical studies have reported that there are 47 species under the genus Epimedium distributed in China [3]. How-
456
P.-S. Xie et al. / Journal of Pharmaceutical and Biomedical Analysis 52 (2010) 452–460
Fig. 4. Histogram of ‘ABCI fingerprint region’ of HPLC profiles of the five official Epimedium species based on the contents determined by HPLC quantitative analysis. (E.b.) E.brevicornum; (E.s.) E.sagittatu ; (E.w.) E. wushanense; (E.k.) E. koreanum; (E.p.): E. pubescens.
ever, most species do not form commercial resources for medicinal use. Besides, several rich resources of non-official species have also been sold in some local crude drug markets in east China, north-west China and south China, blending contingently with the official species spread over several herbal drug markets and manufacturer’s storehouses [2]. In this study, the leaves of eight non-official species: E. acuminatum Franch (n = 4), E. myrianthum Stearn (n = 5), E. elongatum Komarov (n = 2), E. dolichostemon Stearn (n = 1), E. mikinorii Stearn (n = 1), E. leptorrhizum Stearn (n = 3), E. davidii Franch (n = 2) and E. membranaceum Franch (n = 2) were analyzed. A comparative study on HPLC fingerprinting among those species with the five official species showed the following: (1) Convergence of ‘ABCI region’ pattern among species of Epimedium. It was found that different species possibly shared the same ‘ABCI region’ pattern. Comparative observation showed that E. acuminatum, E. myrianthum and E. mikinorii samples displayed the same feature of ‘E.s. pattern’ (a b C I), while E. dolichostemon
and E. myrianthum and its var. jianheanse showed ‘E.w. pattern’ composition (a b C i), and E. elongatum presented ‘E.p. pattern’ (a b C I) (Fig. 6). By all account, the existence of different species sharing the same bioactivity related to ‘ABCI pattern’ indicated the possibility of such species having similar biological activities and therapeutic efficacy. (2) Dissimilation of ‘ABCI region’ pattern within species of Epimedium. In contrast, the chromatographic spectra with some individual Epimedium species were seen to deviate out of the original pattern possibly due to genetic or environmental influence. This is illustrated by E. koreanum and E. sagittatum. Compared with the normal samples, some samples collected from different habitats or in different seasons were found to exhibit almost blank “ABCI region” in their HPLC profiles (Fig. 7). This dissimilation suggests that these herbs cannot contribute the same efficacy as the normal ones. On the other hand, several samples of E. brevicornu exhibited ‘E.p. pattern’ and ‘E.k. pattern’. Some samples of E. davidii possessed ‘E.k. pattern’ and two samples of E. membranaceum showed ‘E.p. pattern’. (3) Lack of ‘ABCI region’ in E. leptorrhizum. The aerial part of E. leptorrhizum has also been used as medicinal ‘Epimedium herb’ in the local market. However, the HPLC profile obtained in our study showed lack of ‘ABCI region’ or only a trace amount. This seemed not an occasional occurrence and similar observation on this species has also been published [3,5,13]. Considering this species is widely distributed and all the samples contained no or trace amount of ABCI, genetic alteration has probably occurred. Accordingly, E. leptorrhizum should be deleted from the list of medicinal Epimedium herb. 4. Discussion Species authentication of herbal drugs is a prerequisite for ensuring their quality for clinical administration, pharmacological study, and quality control. However, not infrequently, multiple species are encountered in a single crude drug in CMM. This is a perplex historical problem because China has a vast territory and diverse plant varieties and regional custom of herbal crude drug usage. Designating any of the single species alone and depriving
Fig. 5. The plot of Principle Component Analysis (PC1 – PC2) of 46 samples of Epimedium spp. Accumulated mean square: PC1 = 70.5%; PC2 = 86.9%.
P.-S. Xie et al. / Journal of Pharmaceutical and Biomedical Analysis 52 (2010) 452–460
the others for the sake of ‘purifying’ the resource of such crude drug is not rational. Yinyanghuo (Epimedium herb) is a typical example. Research studies of Epimedium herbs have shown the convergence in composition of active ingredients among species
457
(Fig. 6A) facilitating their re-classification. On the other hand, dissimilation of active components within species (Fig. 7) results in some abnormal patterns. Some examples have broken down the boundaries of taxonomical species exposing the composition of
Fig. 6. Convergence tendency of ‘ABCI region’ in HPLC profiles of different species. E. acuminatum, E. myrianthum, E. mikinorii have similar pattern of E. sagittatum (a b C I); E. longatum has E. pubescens pattern (E.p. pattern: a b C I).
458
P.-S. Xie et al. / Journal of Pharmaceutical and Biomedical Analysis 52 (2010) 452–460
Fig. 6. (Continued ).
ABCI region out of its original pattern (leaves of E. brevicornu) merged to other patterns. Both findings provide an alternative solution of simplifying the complex morphological taxonomy during on-sight inspection and for quality assurance in the manufacturing process. According to the composition of the “ABCI fingerprint region”, all of the samples can be categorized into five patterns. Further converging the five patterns can be reduced to three patterns – ‘extensive E.w. pattern’ (including ‘E.w.’, ‘E.p.’, and ‘E.s. patterns’), ‘E.b. pattern’ and ‘E.k. pattern’ (Fig. 8) expressed by
Cluster Analysis and Classification Index (see below). Therefore it paved a practical way to alleviate the confusion of co-existence of multiple taxonomical species. For instance, if E. wushanense was selected for usage originally in a prescription or product, then E. pubescens, E. sagittatum, and E. acuminatum can be suggested as alternative entities based on sharing the same ‘ABCI region’ pattern. Classification Index of leaves of Epimedium herb based on bioactive ingredients composition (ABCI fingerprint region).
Fig. 7. HPLC profiles of E. sagittatum (leaves) from different habitats (top: grown in Guizhou province, southwest China; lower: three samples were grown in Anhui province, east China).
P.-S. Xie et al. / Journal of Pharmaceutical and Biomedical Analysis 52 (2010) 452–460
459
Fig. 8. Dendrogram of HPLC fingerprint cluster analysis. Converging all samples into three categories based on the composition of “ABCI region”. E.k. pattern: dominant peak is icariin (peak 6); extensive E.w pattern: epimedin C (peak 4); E.b pattern: epimedin B (peak 3).
1. “ABCI region” distinct 1.1. Epimedin C (peak 4) dominant (extensive E.w. pattern) 1.1.1. Epimedin C (peak 4) dominant exclusively (E.w. pattern) (a b C 5 i) E. wushanense E. dolichostemon E. myrianthum E. myrianthum var. jianheanse 1.1.2. Epimedin C (peak 4)>/
Consistent content of the major bioactive C-8 prenylated flavonoids is another supporting factor for breaking down the boundary of the taxonomical species to merge the commodities of medicinal Epimedium herb into the same group based on the resembling ‘ABCI region’. Comparing the integrated peak areas, the ratio of ‘ABCI’ amount, and total flavonoids amount (ABCI:total) of the samples with the same ‘pattern’, it is possible to estimate the relative amount of the main bioactive ingredients for quality evaluation. A set of unpublished quantitative data on ‘ABCI’ acquired by quantitative HPLC analysis in our lab demonstrated the concordance between precisely determined content and the estimation by integrated HPLC peaks (Fig. 9). Some papers have been published on HPLC fingerprinting analysis [3,5,14,15] of Epimedium species. All of them focus on the “ABCI region” as the characteristics of HPLC profiles of Epimedium herbs for phytotaxonomy and quality analysis. As a typical example, B.L. Guo et al. attempted to delineate the subtle variation of the “ABCI peak group” of 35 species of Epimedium for differentiating the species more strictly. All the samples tested were divided into four main types and nine subtypes [5] to compile with the complex morphological taxonomic system devised by W.T. Stearn [16,17]. Plant taxonomy needs divergence strategy to distinguish the visual variance of the appearance in a subtle way to define the
Fig. 9. Positive correlation between the quantitative data (above) and the integrated peaks area (lower) of total ‘ABCI’. It showed the reconcilability of precise determined contents and relative quantities from the integrated peaks area of HPLC profiles. The latter is rough but effective and easily available.
uniqueness of species. The complex classification verified the diversity of the botanical kingdom. However, for solving the practical difficulty caused by multiple species herbal medicine commodities coexisted in the Chinese herbal drug market, divergence approach is obviously not the right choice. We need reverse thinking. Converging the different species that possess the same HPLC fingerprint pattern towards one group will facilitate the efficacious use of multiple species of Chinese herbal medicines like the case of Epimedium herb. Convergence or divergence depends on the final purposes. Acknowledgement We thank the Science and Technology Development Fund of Macao Government for financial support (Grant Number: 045/2005/A). References [1] Chinese Pharmacopoeia Commission, Chinese Pharmacopoeia I, Chemical Industry Press, Beijing, 2005, p. 229. [2] B.L. Guo, P.G. Xiao, Comment on main species of Herba Epimedii, Chin. J. Chin. Mater. Med. 28 (2003) 303–307. [3] Li Pei-kuan, B.L. Guo, W.H. Huang, Studies on fingerprinting and identification of main species of Herba Epimedii, Chin. J. Chin. Mater. Med. 33 (2008) 1662–1668. [4] B.L. Guo, P.G. Xiao, The flavonoids in Epimedium L. and their taxonomic significance, Acta Phytotaxon. 37 (1999) 228–243.
460
P.-S. Xie et al. / Journal of Pharmaceutical and Biomedical Analysis 52 (2010) 452–460
[5] B.L. Guo, L.K. Pei, P.G. Xiao, Further research on taxonomic significance of flavonoids in Epimedium (Berberidaceae), J. Sys. Evol. 46 (2008) 874–885. [6] S.P. Wong, P. Shen, L. Lee, J. Li, E.L. Yong, Pharmacokinetics of prenylflavonoids and correlations with the dynamics of estrogen action in sera following ingestion of a standardized Epimedium extract, JPBA 50 (2009) 216–223. [7] Z.X. Shou, L. Shen, Y.P. Yang, J. Xie, P.Q. Zhou, L. Gao, Effects of epimedium total flavonoids on bone mineral density and bone metabolism-related indices in primary osteoporosis, J. Clin. Rehabil. Tissue Eng. Res. 13 (2009) 2191–2195. [8] H. Wu, E.J. Lien, L.L. Lien, Chemical and pharmacological investigations of Epimedium species: a survey, Prog. Drug Res. 60 (2003) 1–57. [9] M.K. Lee, Y.J. Choi, S.H. Sung, D.I. Shin, J.W. Kim, Y.C. Kim, Antihepatotoxic activity of icariin, a major constituent of Epimedium koreanum, Planta Med. 61 (1995) 523–526. [10] S.P. Yap, P. Shen, M.S. Butler, Y. Gong, C.J. Loy, E.L. Yong, New estrogenic prenylflavone from Epimedium brevicornum inhibits the growth of breast cancer cells, Planta Med. 71 (2005) 114–119.
[11] K.M. Chen, B.F. Ge, H.P. Ma,.Y. Liu,.H. Bai, Y. Wang, Icariin, a flavonoid from the herb Epimedium enhances the osteogenic differentiation of rat primary bone marrow stromal cells, Pharmazie 60 (2005) 939–942. [12] Chinese Pharmacopoeia Commission, Guidance on chromatographic fingerprint experimental research of TCM Injection (in Chinese) (2002). [13] L.K. Pei, W.H. Huang, T.G. He, B.L. Guo, Systematic studies on quality of main species of ‘herba epimedii’, Chin. J. Chin. Mater. Med. 32 (2007) 2217–2222. [14] C.Y. Huang, L.H. Zhao, L.H. Mei, Chin. Studies on the HPLC fingerprint of Epimedium brevicornum crude drug, J. Nat. Med. 1 (2003) 146–148. [15] L.X. Wang, C.Z. Wang, X.D. Geng, Studies on the RPLC fingerprints and quality evaluation of Epimedium L, Crude drugs. Acta Chim. Sin. 64 (2006) 551–555. [16] W.T. Stearn, Epimedium and Vancouveria (Berberidaceae), a monograph, J. Linn. Soc. 51 (1938) 409–535. [17] W.T. Stearn, The Genus Epimedium and other Herbaceous Berberidaceae Including the Genus Podophyllium, Timber Press, Oregon, 2002.