A comparable, chemical and pharmacological analysis of the traditional Chinese medicinal herbs Oldenlandia diffusa and O. corymbosa and a new valuation of their biological potential.

Introduction

The herb of Oldenlandia diffusa (Willd.) Roxb. (Syn. Hedyotis diffusa Willd., Family Rubiaceae) is a well-known folk medicine in Southern China and has been used for the treatment of hepatitis, tonsillitis, sore throat, appendicitis, urethral infection and malignant tumors of the liver, lung and stomach (Xu et al., 1997). A recent systematic survey on confusable Chinese herbal medicines has revealed that O. corymbosa (L.) Lam is also indiscriminately available as O. diffusa in the wholesale and food markets (Zhao and Li, 2004). According to the Chinese Pharmacopoeia, O. diffusa has been recorded as an ingredient of chinese patent medicine (Pharmacopoeia of China, 2005). The popular use of the two herbs has attracted a great deal of attention to their differences of pharmacological activity.

Previous pharmacological studies of O. diffusa have demonstrated that it has antitumor, immunomodulatory (Yoshida et al., 1997; Shan et al., 2001; Chung et al., 2002), antimutagenic (Wong et al., 1992a, b, 1993a, b), anti-inflammatory, hepatoprotective (Lin et al., 2002), anti-oxidative (Lu et al., 2000; Kim et al., 2005) and neuroprotective (Kim et al., 2001) activities. It has been suggested that the aqueous extract of O. diffusa has immunomodulating activity and it may stimulate the immune system to eliminate tumor cells (Shan et al.,2001). The aqueous extract has also been proven effective in inhibiting the growth of eight different cancer cell lines and inducing apoptosis (Gupta et al., 2004). On the other hand, ursolic acid, a main component of methanol extracts not found in aqueous extracts of O. diffusa, exhibited significant antitumor effects (Kim et al., 1998). Oral administration of O. diffusa to mice significantly inhibited the growth of murine renal carcinoma cells and O. diffusa extracts enhanced macrophage function in vitro (Wong et al.,1996). Polysaccharides isolated from O. diffusa also showed antitumor activities against transplanted Sarcoma-180 cells in mice (Li et al., 2002). Therefore, antitumor activity has been identified as the main pharmacological effect of O. diffusa,

However, previous phytochemical studies of O. diffusa have demonstrated that it contains iridoid glucosides, triterpenoids, flavonoids and polysaccharides (Nishihama ct al., 1981; Wu et al., 1991, 1992, 2005; Lu and He, 1996; Lu et al., 2000; Kim et al., 2001; Ren et al., 2005). Among them, iridoid glucosides were the main components of O. diffusa, but there is limited data relating iridoid glucosides to the antitumor activity.

Here, the antiproliferative effect of different fractions from O. diffusa and O. corymbosa on human colon carcinoma CaCo2 cells and hepatoma HepG2 cells was appraised. The high-performance liquid chromatography (HPLC) method was used for the first time to compare the chemical profiles of the two herbs containing their different fractions. Five compounds, including ursolic acid (1), oleanolic acid (2), asperuloside (3), E-6-O-p-coumaroyl scandoside methyl ester (4)and E-6-O-p-coumaroyl scandoside methyl ester- 10-0-methyl ether (5), isolated from O. diffusa in our laboratory were used as chemical markers (Fig. I). Among the five compounds, compounds 3, 4 and 5 belong to the iridoid glucosides. As the main components of O. diffusa, the anti-angiogenic activity of the above five compounds was also tested for the first time using the zebrafish angiogenic model, as anti-angiogenic agents could inhibit tumor angiogenesis by blocking nutrient and oxygen supply.

[FIGURE 1 OMITTED]

Materials and methods

Materials

After comparison of the preliminary HPLC finger-prints of many batches of the two herbs from different herbal markets, the experimental samples were selected as the representative samples for comparison. The herbal sample of O. diffusa (Willd.) Roxb. was purchased from Luoding County, Guangdong Province in August 2004, while the O. corymbosa (L.) Lam sample was purchased from Hong Kong in October 2004. These materials were authenticated by Professor Zhao Zhongzhen of the School of Chinese Medicine, Hong Kong Baptist University, and voucher specimens were deposited at the HKBU BOC (HK) Chinese Medicines Centre.

Reagent grade acetonitrile and methanol were purchased from E. Merck (Darmstadt, Germany); Dulbec-co's Modified Eagle Medium (DMEM), RPMI 1640 medium, 100 x antibiotic-antimycotic and fetal bovine serum from [Gibco sup.TM] Invitrogen Corporation (USA); phosphate buffer saline (PBS) from Oxoid Ltd. (UK); trypsin-EDTA, Sulforhodamine B (SRB) stain and Trisbase from Sigma (USA); trichloroacetic acid (TCA) from Aldrich Chemical. Water was purified using a Milli-Q (Millipore, Bedford, MA, USA) water purification system. A spectrum MAX340 microplate reader was used to read the absorbance. The reference compounds of ursolic acid (1), oleanolic acid (2), asperuloside (3), E-6-O-p-coumaroyl scandoside methyl ester (4) and E-6-0-p-coumaroyl scandoside methyl ester- 10-O-P-coumaroy seandoside mthly ester (4) and E-6-O-P-coumaroyl scandoside methly ester-10-O methyl ether (5) were isolated and purified from the dried berb of O. diffusa in our laboratory (Liang et al., 2006). Their structures were determined on the basis of HR-MS, 'H-NMR and (sup.13)C-NMR spectral data.

HPLC conditions

HPLC fingerprinting analysis was performed on an Agilent 1100 series LC system consisting of a G1311A Quart pumps, a G1322A degasser, a G1315A photo-diode array detector (DAD) and a G1313A ALS. Separation was performed at room temperature on an Alltima C(sub.18) analytical column (250mm x 4.6 mm, 5um, Alltech Associates, Inc. USA) coupled with a C(sub.18) guard column (7.5 mm x 4.6 mm, 5 um, Alltech Associates, Inc. USA) that was eluted with acetonitrile/water at a flow rate of 1 ml/min by a discontinuous gradient in which acetonitrile was adjusted to 0%, 7%, 15%, 20%, 35%, 75% and 90%, at 0, 10, 25, 45, 55, 65 and 75min, respectively. Detections were performed at 238 and 210 nm.

Preparation of plant extracts and fractions for antiproliferative analysis

Fifteen gram powdered whole plant was extracted using 200 ml methanol for 4 h on a soxhlet apparatus. Extracts were dried under reduced pressure to obtain 1941 mg of O. diffusa residue and 1752mg of O. corymbosa residue, respectively. Each solid residue was dissolved in water and extracted with petroleum ether (for disposing lipid and pigments), chloroform (fraction 1, 113.8mg of O.diffusa and 329 mg of O. corymbosa), ethyl acetate (fraction 2, 128.9mg of O. diffusa and 69 mg of O. corymbosa) and n-butanol (fraction 3, 128.6 mg of O. diffusa and 119mg of O. corymbosa). These plant extracts were dissolved in DMSO-ethanol (1:1) to producing stock solutions.

Preparation of plant extracts and fractions for HPLC analysis

One gram of powdered whole plant was extracted using 50 ml methanol for 4h on a soxhlet apparatus. Two extracts of each sample were prepared and dried under reduced pressure. One of the extracts of each sample was dissolved in 25 ml methanol for analysis as methanol crude extract. Another extract was dissolved in water and extracted sequentially with petroleum ether, chloroform, ethyl acetate and n-butanol. The last three fractions were dried under reduced pressure and dissolved in 10 ml methanol. These preparations were filtered through 0.45 pm membranes and 10 pi samples were analyzed by HPLC.

Human cell lines

Human colon carcinoma CaCo2 cells and hepatoma HepG2 cells were maintained in DMEM medium and RPMI 1640 medium, respectively, supplemented with 1% antibiotic-antimycotic and 10% fetal bovine serum.

Antiproliferative assay

Cells grown to approximately 80-90% confluent were detached by 10% trypsin-EDTA treatment, washed with PBS and collected by centrifugation. Cells were resuspended in 5ml of medium and viable cells were identified by trypan blue exclusion and counted using a hemocytometer. Cells were diluted with growth medium and seeded into 96-well microtiter plates at the pre-determined optimal plating densities of 1.5 x [10.sup.3] cells/100[micro]l/well for CaCo2 cells and 5x [10.sup.3] cells/100[micro]l/well for HepG2 cells. Cells were incubated at 37[degrees] for 24h before experimentation.

Stock solutions of plant extracts in 1:1 DMSO-ethanol were diluted by appropriate growth media to the desired concentrations before each experiment and were added at 100[micro]l/well (Table 1). Cells were incubated at 37[degrees c] for 72h and cell number was estimated by SRB protein assay. Each treatment was performed in triplicates. The antiproliferative effect of oleanolic acid was also determined by the same method.

Table 1. The antiproliferative effect of the extracts from Oldenlandia
diffusa and O. corymbosa

Extraction        [Ic.sub.50] [micro]g/ml [+ or -] S.D. (n = 3)

                     HepG2 cell line     CaCo2 cell line

OD-M              468.3 [+ or -] 179.6   953.6 [+ or -] 109.9
OD-1              112.0 [+ or -] 6.9     272.8 [+ or -] 71.8
OD-2             1387.7 [+ or -] 130.6   >1863
OD-3             1322.0 [+ or -] 246.8  1704.3 [+ or -] 147.8
OC-M              563.3 [+ or -] 137.0  1185.7 [+ or -] 55.3
OC-1              116.7 [+ or -] 14.4    508.6 [+ or -] 106.2
OC-2              544.0 [+ or -] 78.6    744.0 [+ or -] 35.4
OC-3              980.7 [+ or -] 125.6  1474.9 [+ or -] 144.0
Oleanolic acid     10.4 [+ or -] 0.3      37.5 [+ or -] 7.6
acid

OD-M The methanol extract of O.diffusa; OC-M The methanol
extract of O. corymbosa; 1,2,3 mean fraction 1,2,and 3.

SRB assay

The SRB assay is a colorimetric assay assessing growth inhibition by estimating cell numbers indirectly by staining total cellular protein with the dye SRB (Skehan et al., 1990). Cells were fixed by adding 100[micro]l/well of 15% cold TCA and incubated at 4[degrees]C for 1 h. Wells were washed five times with cold water and dried in air. in all, 50[micro]l of SRB solution (0.4g in 100 ml 1% acetic acid) was added to each well. After 10 min wells were rinsed with 1% acetic acid four times to remove all loosely bound dye. The plates were air dried and 100 [mu]l of 10[mu]M Tris base, pH 10.5, was added to each well to solubilize the dye. The plates were shaken gently for 20 s before the absorbance was read on a microplate reader at 515 nm. Cell survival was estimated as the percentage absorbance compared to the control non-treated cells. [IC(sub.50) values were obtained by applying the Anilisa program and by plotting the percentage of survival versus extract concentration, interpolated by cubic spine.

Antiangiogenic zebrafish assay

Zebrafish embryos were generated by natural pair-wise mating and maintained in embryo water (0.2 g/1 of Instant Ocean Salt in distilled water) at 28.5 [degrees]C. At 24 h post fertilization, the embryos were dechorionated and distributed into 96-well microplates, one embryo per well. Then test compounds at various concentrations were added to the well. At least 20 embryos were used for each concentration. At 72 h post fertilization. the embryos were performed by a quantitative endogenous alkaline phosphatase assay as described (Parng et al., 2004).

Results and discussion

Antiproliferative effects of the extracts from O. diffusa and O. corymbosa

The [IC .sub.50] values of the methanol extract of O. diffusa were 468.3 [+ or -]179.6 [mu]g/ml on Hep[G.sub.2] cells and 953.6[+ or -]10.9.9 [micro]g/ml on Ca[Co.sup.2] cells, respectively. As for O. corymbosa, the [1C.sup.50] values of the methanol extract were 563.3 [+ or -] 179.6 [micro]g/ml on Hep[G.sup.2] cells and 1185.7[+ or -]55.3 [micro]g/ml on Ca[Co.sub 2] cells, respectively. Such results showed that there are almost no antiproliferative effects of the methanol extracts from the two herbs (Table 1). On the two cancer cell lines, the IC30 values of chloroform fractions from the two herbs were the lowest while the ethyl acetate and n-butanol fractions of O. diffusa had no noticeable antiproliferative effect with [IC.sub.50] values of more than 1000 ug/ml. From the antiproliferative assay, it was clear that the chloroform fractions of both herbs were almost the same as active against HepG2 cells but O. diffusa was more active on CaCo2 cells. The results also showed that HepG2 cells were more sensitive to those fractions than CaCo2 cells. For example, the [IC.sub.50 ] values on Hep[G.sub.2] cells of the chloroform fraction from O. diffusa was 112.0 [+ or -] 6.9 [mu]g/ml compared to 272.8 [+ or -]71.8 [micro]g/ml for Ca[Co.sup.2] cells.

HPLC analysis

The examination of the chemical profiles of the methanol extracts from O. diffusa (Fig. 2A) and O. corymbosa (Fig. 2B) by HPLC showed that they contained quite different compounds apart from compounds 1, 2 and 3. The content of compound 4 was noticeably lower in O. corymbosa than in O. diffusa. In addition, compound 5 was absent from O. corymbosa but markedly present in O. diffusa. Comparison of these peaks was made on the basis of their ultraviolet absorption spectra and retention time. Fig. 2 also indicated that the compounds 4 and 5 were the main components of O. diffusa.

[FIGURE 2 OMITTED]

Because the chloroform fractions of the two herbs contributed mostly to the slightly antiproliferative effect, they were further analyzed by HPLC. As shown in Figs. 3 and 4, the main different peaks in the chromatograms of the chloroform fractions were peak I for O. diffusa as well as peaks 4 and 5 for O. corymbosa. Compared with the chemical marker, peak I was identified as E-6-O-p-coumaroyl scandoside methyl ester-10-O-methyl ether. The peaks 2 and 3 were identified as oleanolic acid and ursolic based on standard chemical markers (Fig. 4). The common peaks, named as Group A, were found in both chromatograms of the chloroform fractions of the two herbs. However, the intensity of the Group A peaks in the chromatograms of the chloroform fractions was obviously higher than those of the ethyl acetate and n-butanol fractions in the two herbs. Compared with the chromatograms of the three fractions of O. diffusa, the main components of compounds 3, 4 and 5 were not attributable to the antiproliferative effect. It was deduced that the compounds of Group A peaks were the main antiproliferative contstituents.

[FIGURE 3 OMITTED]

The antiproliferative effect of oleanolic acid was also studied here. The [IC.sup.50] values of oleanolic acid on Ca[Co.sup.2] cell lines and HepG2 cell lines were 37.5 and 10.4 [micro]g/ml, respectively, which indicated that oleanolic acid has antiproliferative effect. Ursolic acid has been shown to cause a significant inhibition of the proliferation of a number of culture tumor cells, such as A549 (human lung), SK-OV-3 (ovary) and HCT--15 (colon) (Kim et al., 1998). Therefore, the antiproliferative effect of O. diffusa and O. corymbosa could be attributed mainly to ursolic acid and olenaolic acid.

Antiangiogenic zebrafish assay

As the main differences between O. diffusa and O. corymbosa were the contents of compounds 4 and 5, as well as compounds 1 and 3, which were the main components of the two herbs, another screening model of the zebrafish angiogenic model was used to appraise the antianiogenic activity of compounds 1, 3, 4 and 5.The results showed that the compounds 1, 4 and 5 had almost no effect on zebrafish angiogenic vessel formation at the concentrations tested. Compound 3 inhibited zebrafish angiogneic vessel formation by 30% at 40 [mu]M and was a very weak inhibitor compared to 2-methoxyestradiol, a putative antiangiogenic agent currently under clinic trial. The latter can inhibit angiogenic vessel formation by 34% at 10 [mu]M (Fig. 5). Such results suggested that the mechanism of antitumor activity of O. diffusa may be not through antiangiogenic activity.

[FIGURE 5 OMITTED]

Conclusions

The HPLC analysis of the methanol extracts of both Oldenlandia species have revealed that they differ in their chemical composition. Additionally, the pharmacological investigation using human hepatoma HepG2 cells, an angiogenic model and colon carcinoma Ca[Co.sup.2] cells has revealed that the methanol extracts of both herbs showed only a very low antiproliferative and almost no antiangiogenic effect. In the light of these results, it is not very likely that constituents 1-5 are responsible for the antitumoral activity of the Old enlandia herbs. Therefore, the use of Oldenlandia diffusa in TCM might be due to the immunomodulating effect of the polysaccharides that are present only in the water extract (decoction) used in TCM medical practices. Our investigation underlines once more that "modernization of Chinese medicine" has to take into consideration a thorough quality proof of TCM herbs and a correlation of their chemical composition with the traditionally intended medication.

[FIGURE 6 OMITTED]

Acknowledgments

The authors would like to acknowledge their thanks to Ms. Shen Xiaoling (Department of Biology and Chemistry, City University of Hong Kong) for comments. The project was supported by the Faculty Research Grant of Hong Kong Baptist University (FRG/04-05/II-55).

References

Chung, H.S., Jeong, H.J., Hong, S.H., Kim, M.S., Kim, S.J., Song, B.K., Jeong, I.S., Lee, E.J., Ahn, J.W., Baek, S.H., Kim, H.M., 2002. Induction of nitric oxide synthase by Oldenlandia diffusa in mouse peritoneal macrophages. Biol. Pharm. Bull. 25, 1142-1146.

Gupta, S., Zhang, D., Yi, J., Shao, J., 2004. Anticancer activities of Oldenlandia diffusa. J. Herb. Pharmacother. 4, 21-33.

Kim,, S.H., Ahn, B.Z., Ryu, S.Y., 1998. Antitumor effects of ursolic acid isolated from Oldenlandia diffusa. Phytother. Res. 12, 553-556.

Kim, Y., Park, E.J., Kim, J., Kim, Y., Kim, S.R., Kim. Y.Y., 2001. Neuroprotective constituents from Hedyotis diffusa. J. Nat. Prod. 64, 75--78.

Kim, D.H., Lee, H.J., Oh, Y.J., Kim, M.J., Kim, S.H., Jeong, T.S., Baek, N.I., 2005. Iridoid glycosides isolated from Oldenlandia diffusa inhibit LDL-oxidation. Arch. Pharm. Res. 28, 1156-1160.

Li, R., Zhao, H.R., Lin, Y.N., 2002. Anti-tumor effect and protective effect on chemotherapeutic damage of water soluble extracts from Hedyotis diffusa. J. Chin. Pharm. Sci 11, 54-58.

Liang, Z.T., Jiang, Z.H., Leung, K.S.Y., Zhao, Z.Z., 2006. Determination of iridoid glucosides for quality assessment of Herba Oldenlandiae by high-performance liquid chromatography. Chem. Pharm. Bull. 54, 1131-1137.

Lin, C.C., Ng, L.T., Yang, J.J., Hsu, Y.F., 2002. Antiinflammatory and hepatoprotective activity of pehhue-juwa-chi-cao in male rats. Am. J. Chin. Med. 30, 225-234.

Lu, H.C., He, J., 1996. A study on chemical constituents of Oldenlandia diffusa (Willd) Roxb. Nat. Prod. Res. Dev. 8, 34-37.

Lu, C.,M., Yang, J.J., Wang, P.Y., Lin, C.C., 2000. A new acylated flavonol glycoside and antioxidant effects of Hedyotis diffusa. Planta Med. 66, 374-377.

Nishihama, Y., Masuda, K., Yamaki, M., Takagi, S., Sakina, K., 1981. Three new iridoid glucosides from Hedyotis diffusa. Planta Med. 43, 28-33.

Parng, C.,, Seng, W.L., Semino, C., Megrath, P., 2004. Zebrafish: a preclinical model for drug screening. Assay Drug Dev. Technol. 1, 43-48.

Pharmacopoeia of China, 2005. Chemical Industry Publishing House, Beijing, p. 22.

Ren, F.Z., Liu, G.S., Zhang, L., Niu, G.Y., 2005. Studies on chemical constituents of Hedyotis diffusa Willd. Chin. Pharm. J. 40, 502-504.

Shan, B.N., Zhang, J.Y., Du, X.N., Li, Q.X., 2001. Immunomodulatory activity and anti-tumor activity of Oldenlandia diffusa in vitro. Zhongguo Zhong Xi Yi Jie He Za Zhi 21, 370-374.

Skehan, P., Storeng, R., Scudiero, D., Monks, A., McMahon, J., Vistica, D., Warren, J.T., Bokesch, H., Kenney, S., Boyd, M.R., 1990. New colorimetric cytotoxicity assay for anticancer-drug screening. J. Natl. Cancer Inst. 82, 1107-1112.

Wong, B.Y., Lau, B.H., Teel, R.W., 1992a. Chinese medicinal herbs modulate mutagenesis, DNA binding and metabolism of benzo[a]pyrene 7, 8-dihydrodioI and benzofa] pyrene 7, 8-dihydrodiol-9, 10-epoxide. Cancer Lett. 62, 123-131.

Wong, B.Y., Lau, B.H., Tadi, P.P., Teel, R.W., 1992b. Chinese medicinal herbs modulate mutagenesis, DNA binding and metabolism of aflatoxin B1. Mutat. Res. 279, 209-216.

Wong, B.Y., Lau, B.H., Yamasaki, T., Teel, R.W., 1993a. Modulation of cytochrome P-450IA1-mediated mutagenicity, DNA binding and metabolism of benzo[a]pyrene by Chinese medicinal herbs. Cancer Lett. 68, 75-82.

Wong, B.Y., Lau, B.H., Yamasaki, T., Teel, R.W., 1993b. Inhibition of dexamethasone-induced cytochrome P450-mediated mutagenicity and metabolism of aflatoxin B1 by Chinese medicinal herbs. Eur. J. Cancer Prev. 2, 351-356.

Wong, B.Y., Lau, B.H., Jia, T.Y., Wan, C.P., 1996. Oldenlandia diffusa and Scutellaria barbata augment macrophage oxidative burst and inhibit tumor growth. Cancer Biother. Radiopharm. 11, 51-56.

Wu, H., Tao, X., Chen, Q., Lao, X., 1991. Iridoids from Hedyotis diffusa. J. Nat. Prod. 54, 254-256.

Wu, H.M., Huang, S.Y., Lao, X.F., Ma, K., Zhou, H.Q., 1992. A study on the structure of the polysaccharide from Hedyotis diffusa Willd. Youji Huaxue 12, 428-431.

Wu, K.S., Zhang, K., Tan, G.S., Zeng, G.Y., Zhou, Y.J., 2005. Studies on constituents of Oldenlandia diffusa. Chin. Pharm. J. 40, 817-819.

Xu, G.J., Xu, L.S., Wang, Z.T., 1997. In: Species System-atization and Quality Evaluation of Commonly Used Chinese Traditional Drugs (South-China Edition), vol. IV. Fujian Science and Technology Press, Fuzhou, p. 658.

Yoshida, Y., Wang, M.Q., Liu, J.N., Shan, B.E., Yamashita, U., 1997. Immunomodulating activity of Chinese medicinal herbs and Oldenlandia diffusa in particular. Int. J. Immunopharmacol. 19, 359-370.

Zhao, Z.Z., Li, T.Y.S., 2004. Authentication of Easily Confused Chinese Materia Medica in Hong Kong. The Chinese Medicine Merchants Association Ltd., Hong Kong, pp. 144-145.

Zhitao Liang (a), Mingfang He (b), Wangfun Fong (a),Zhihong Jiang (a),Zhongzhen Zhao (a) *

(a) School of Chinese Medicine, Hong Kong Baptist University, Kowloon, Hong Kong Special Administrative Region, PR China (b)Department of Biology, The Chinese University of Hong Kong, Shatin, NT, Hong Kong, PR China

* Correspondiong author. Tel: [phone]; fax: [phone]. E-mail Address [e-mail]. (Z.Zhao).