Verification of the formulation and efficacy of Danggui Buxue Tang (a decoction of Radix Astragali and Radix Angelicae Sinensis): an exemplifying systematic approach to revealing the complexity of Chinese herbal medicine formulae

This article exemplifies a systematic approach to revealing the complexity of Chinese herbal medicine formulae through three levels of scientific research: standardization of herbs, verification of ancient formulae and mechanism studies. We use Danggui Buxue Tang (DBT) as an example for this approach. Among thousands of traditional Chinese medicine herbal formulae, almost all of which consist of multiple herbs, DBT is one of the simplest. Containing only two herbs, namely Radix Astragali (RA) and Radix Angelicae Sinensis (RAS), DBT is traditionally used to treat ailments in women. The weight ratio of RA to RAS in DBT was prescribed to be 5:1 as early as in 1247 AD. In addition to advanced chemical analysis of herbal constituents, DNA genotyping techniques have been developed for reliable standardization of RA and RAS. Chemical evaluation shows that main active constituents in DBT, including astragaloside IV, calycosin, formononetin and ferulic acid, were most abundant after extraction at the RA to RAS ratio of 5:1, whereas other tested RA to RAS ratios only gave sub-optimal levels of the active constituents. Biological evaluation indicates that bioactivities of DBT, e.g. immuno-modulatory, oesteotropic and estrogenic effects are also best exerted at the RA to RAS ratio of 5:1. Correlation analysis demonstrates statistically significant relationship between the tested chemical constituents and tested bioactivities. Up- and down-regulation of expression of some genes as potential biomarkers has been detected by using gene chip technology. This systematic approach on the basis of herbal standardization, chemical and biological verification and mechanism studies, as exemplified in this article, will be useful to reveal the complexity of not only DBT but also other Chinese medicine herbal formulae.


Background
Traditional Chinese medicine (TCM) has been used to improve the well-being of the Chinese people for thousands of years. TCM products, many of which were raw materials, made up only 3% of the 16 billion USD international herbal medicine market in 2004 [1,2]. Since the market opening-up of China, international pharmaceutical companies have been gaining a market share in both conventional and herbal medicine products in China. In the 21 st century, TCM products should meet stringent international quality and safety standards through modernization; otherwise they will lose their competitiveness.
Standardization as the basis of modernization and internationalization of TCM is the key to ensure the safety and efficacy of TCM products. At present, lack of standardization in TCM products impedes the development of TCM. For instance, it is common that different herbs have the same name or a single herb has different names in the market. Some herbs cultivated in different regions or harvested in different seasons may vary considerably in their chemical and biological properties. Most of the TCM products do not have specific biomarkers. TCM is traditionally administered in the form of a decoction with a combination of different herbs. The complexity of biological effects of the interactions among different compounds within a decoction complicates experimental studies to reveal the action mechanisms.
Among thousands of TCM formulae, Danggui Buxue Tang (DBT) is one of the simplest. The formula consists of only two herbs: Radix Astragali (RA, Huangqi) and Radix Angelicae Sinensis (RAS, Danggui) in a weight ratio of 5:1. According to a traditional method, the herbs are boiled together in two bowls of water at moderate heat until the final volume has been reduced to one bowl [2]. In a book entitled Neiwaishang Bianhuo Lun in 1247 AD, DBT was first described by Li Dongyuan, one of the four wellknown TCM physicians during the Jin and Yuan Dynasties in China.
In this review, we summarize recent findings of DBT to exemplify a systematic approach to revealing the complexity of Chinese herbal medicine formulae through three levels of scientific research: standardization of raw materials, verification of ancient formulae and mechanism studies.

Standardization of Radix Astragali and Radix Angelicae Sinensis
A reliably reproducible chemical composition of DBT is a prerequisite in delineating the biological effects of this Chinese medicine preparation. The quality of RA and RAS may be considerably influenced by weather, geographic location, soil conditions, and the methods of cultivation and processing. Some Chinese medicinal materials with excellent quality are only produced in certain regions of China which are often referred to as 'the best growth region' or 'Didao'. Therefore, how to authenticate and choose the best RA and RAS plays a critical role in ensuring the quality of DBT (Figure 1).
With neighbor-joining and maximum parsimony analyses, phylogenetic trees were mapped according to their sequence diversity. A. membranaceus and A. membranaceus var.mongholicus have the highest sequence homology. The common substitute of RA in some parts of China is the roots of Hedysarum polybotrys which has very different The authentic sources of RA and RAS  HPLC and spectrophotometry were used to determine the levels of isoflavonoids, astragalosides, polysaccharides, amino acids and trace elements, which are the main active constituents in different Astragalus species and RA collected in different seasons and of various ages. The results indicated that RA of three years of age from Shanxi, China ( Figure 2a) contained the highest amounts of isoflavonoids, saponins and polysaccharides [6,10].
Determination of the active constituents in RA and RAS and Angelica gigas Nakai, mainly found in Japan and Korea respectively, are also sold as RAS in the markets of South East Asia [11][12][13][14] (Figure 1). Studies have shown that the three commonly used Angelica roots vary in their chemical composition, pharmacological properties and efficacy [9,11]. The 5S-rRNA spacer domains of the three species of Angelica were amplified and their nucleotide sequences were determined. The sequence of A. sinensis is 72.87% and 73.58% identical to those of A. acutiloba and A. gigas respectively, while A. acutiloba and A. gigas are 93.57% identical in their sequences [9]. The phylogenetic tree clearly reveals that the three Angelica species are divided into two clusters: A. sinensis is in one cluster and A. acutiloba and A. gigas are in another.
The main chemical constituents of Angelica roots are ferulic acid, Z-ligustilide, angelicide, brefeldin A, butylidenephthalide, butyphthalide, succinic acid, nicotinic acid, uracil and adenine [9,[15][16][17]. The levels of ferulic acid and Z-ligustilide are often used as chemical markers for the quality control of Angelica roots [16]. In A. sinensis roots from Gansu, China, the levels of ferulic acid and Zligustilide are about ten-fold higher than those of the roots of A. acutiloba (from Japan) and A. gigas (from Korea) [9,17] (Figure 2b). Su Jing (659 AD) in Tang Bencao and Li Shizhen (1596 AD) in Bencao Gangmu recorded that Angelica roots of two years of age produced in Gansu were the authentic source. RAS from Gansu contains about two-fold higher amounts of Z-ligustilide and ferulic acid than those RAS from Yunnan, Shanxi or Sichuan, China [9] (Figure 2b). To ensure the best quality of DBT decoction, we suggest that standardized RA from Shanxi and standardized RAS from Gansu should be used in all DBT preparations ( Figure 2c).

Verification of the DBT formula Chemical evaluation
Li Dongyuan (1247 AD) documented that RA and RAS combined at a ratio of 5:1 demonstrated the best efficacy. In a previous study [18], DBT was prepared by boiling the herbal mixture under various conditions and the results indicated that the 5:1 ratio indeed provided the maximum levels of active constituents of DBT. Furthermore, the levels of active constituents and biological activities of DBT extracts were investigated with preparations of RA and RAS at ratios of 1:1, 2:1, 3:1, 4:1, 5:1, 7:1 and 10:1.
Used as chemical markers, the main active constituents in DBT include RA-derived astragaloside IV, calycosin and formononetin, RAS-derived ferulic acid and ligustilide, and total saponins, total flavonoids and total polysaccharides [19]. The detected levels of the chemical markers varied significantly among the seven preparations ( Figure 3). The level of astragaloside IV of the 5:1 ratio preparation was the highest, 2-fold higher than the 10:1 ratio prepara-tion which recorded the lowest level [19]. The 5:1 ratio preparation also contained the highest level of calycosin, formononetin, and ferulic acid. As regards the levels of total saponins, total flavonoids and total polysaccharides, the 5:1 ratio DBT preparation recorded the highest levels ( Figure 3) [19].
There are several possibilities for higher levels of active chemical constituents in DBT preparations. Firstly, compounds such as saponins (over 2% in total dry weight) [10,17] in RA may help increase the solubility of other compounds extracted from RAS. For example, astragaloside increases the solubility of RAS-derived ferulic acid and ligustilide. Secondly, ferulic acid and ligustilide are readily oxidized under heat, which means they can be degraded when boiled [9]. However, when RAS is boiled together with RA, compounds derived from RA may prevent this oxidization process, thereby producing a higher yield of ferulic acid and ligustilide in DBT preparations. Thirdly, the stability of those active constituents may be improved by having a cocktail of different chemicals. Further research is required for better understanding of this complexity.

Biological evaluation
According to TCM theories, DBT replenishes qi and nourishes xue (the blood). DBT is therefore used for treating menopausal symptoms [2]. Due to a deficiency of ovarian hormones, especially estrogen, women in menopause often suffer from hot flashes, sweating, anxiety, mood swings and an increased risk for other health problems, such as reduction of bone mineral density and cardiovascular diseases [20]. Apart from a lack of estrogen, the immune system is also involved in the menopausal symptoms. Steroid hormones may modulate the immune response [21] and immune reactions may also regulate the ovarian function [22]. Various bioactivities related to menopausal symptoms, such as osteotropic effect, estrogenic effect, anti-platelet aggregation effect and immunomodulatory effect have been used to evaluate the functional roles of DBT.
DBT extract was applied to a cultured human MG-63 osteosarcoma cell. Bone cell proliferation and differentiation were measured by 3-(4, 5-dimethylthioazol-2-yl)-2, 5diphenyltetrazolium bromide (MTT) assay and alkaline phosphatase (ALP) assay. DBT induced both the proliferation and differentiation of osteoblast MG-63 cells in a dose-dependent manner. In both assays, DBT showed stronger effects than RA or RAS alone, In the MTT assay, the 5:1 ratio DBT extract stimulated MG-63 cell proliferation, which was 10-20% higher than the extracts of other ratios ( Figure 4). For bone cell differentiation, the 5:1 ratio DBT preparation induced ALP activity to the highest level among all ratios and showed the strongest osteotropic effect [19].
The estrogenic effects of DBT were tested by a cellular reporter system of transcriptional activation of estrogen receptor/promoter. A promoter/reporter construct (pERE-Luc) corresponding to the responsive elements of estrogen receptor was stably transfected into MCF-7 cells. The DBT extracts of various ratios were applied onto the cultures for 2 days. Two parameters, namely cell number and promoter activity (luciferase activity), were determined. While DBT was not able to alter the proliferation of MCF-7 cells, the estrogen-driven promoter activity was markedly induced by DBT ( Figure 4); the 5:1 ratio DBT showed the strongest effect in inducing the promoter activity than RA, RAS alone or the extracts of other ratios [19].
In anti-platelet aggregation assay, the activity of DBT in preventing ADP-induced platelet aggregation was determined. The ratios 5:1 and 7:1 DBT extracts demonstrated higher levels of activity in preventing platelet aggregation than either RA, RAS alone or the extracts of other ratios ( Figure 4) [19].
In a study of immuno-modulatory effects, DBT preparations of various ratios were applied to cultured T-lymphocytes and macrophages. In cultured T-lymphocytes, DBT induced markedly cell proliferation, interleukin-2 secretion and the phosphorylation of extracellular signalregulated kinase (ERK1/2). In addition, the phagocytosis of cultured macrophages was elevated by DBT treatment. The immuno-modulatory effects of 5:1 ratio DBT were the strongest [23] (Table 1).
In addition to the in vitro assays, the 5:1 ratio of RA and RAS in DBT was further tested and verified by animal studies. In DBT-administrated mice, the 5:1 ratio preparation was the most effective decoction in triggering immune responses [24,25].
The pharmacological studies in animals also suggest that DBT has the ability to promote hematopoiesis, to stimulate blood circulation, to prevent osteoporosis and to counter oxidative stress [19,26,27]. Moreover, DBT is known to enhance myocardial mitochondria and glutathione status in red blood cells, thereby increasing their resistance to injury induced by oxidative stress [28]. In rats, DBT protected against myocardial ischemia-reperfusion injury in a dose-dependent manner [28]. A more potent cardio-protection was demonstrated in DBTtreated rats than in rats treated with either extracts of RA, RAS alone, or a mixture of RA and RAS (not boiled together). When the mice were administered orally with DBT, the serum collected from abdominal aorta was added to an in vitro cultivating system of mouse hematopoietic progenitor cells. The decoction-contained serum showed promoting actions to CFU-GM and CFU-E. Once again, the action of the 5:1 ratio DBT was 97.81% stronger than that of the 1:1 ratio extract [29,30] (Table 2).

Mechanism studies Correlation between chemical fingerprints and bioactivities of DBT
Fifty-four chemical peaks were detected in DBT extracts by an HPLC analysis ( Figure 5a) and a total of over 100 DBT extracts from various preparations were analyzed [27]. Among these 54 peaks, the markers for RA-derived astragaloside IV, calycosin and formononetin, and for RASderived ferulic acid and ligustilide were identified. In analysis of correlation, the identified 54 peak areas together with the contents of total saponins, total flavonoids and total polysaccharides were considered as independent variables. The results of the four bioactivities, namely proliferation and differentiation of MG-63 cells, estrogenic property in MCF-7 cells and anti-platelet aggregation activity, were considered as dependent variables. By analyzing the correlation of these two kinds of variables, coefficients of correlation between the HPLC data of the 57 chemicals and the bioassay data of the DBT extracts were obtained. The values of the coefficients indicate possible relationship of these chemical peaks with bioactivities, where positive values suggest positive effects of chemicals DBT induced cell proliferation, interleukin-2 secretion and the phosphorylation of extracellular signal-regulated kinase (ERK1/2) in cultured T-lymphocytes. The 5:1 ratio DBT showed the strongest immuno-modulatory effects.

Cultured T-lymphocytes and macrophages
DBT of various ratios of RA and RAS, compared with PHA, PMA, Zymosan A and negative control Gao et al. [23] Biological activities of RA, RAS and DBT of various RA to RAS ratios on bioactivities and negative values suggest negative effects.
In the assay of MG-63 cell proliferation, astragaloside IV, formononetin, total saponins and total flavonoids are correlated with the bioactivities (Figure 5b). In the assay of MG-63 cell differentiation, formononetin, total saponins and total flavonoids are correlated with the bioactivities. In the analysis of estrogen promoter in MCF-7 cells, ferulic acid are correlated with the bioactivities. Calycosin and total polysaccharides were two very important factors in the assay of anti-platelet aggregation. On the other hand, the amount of ligustilide showed negative effects in all bioassays (Figure 5b). Other components of DBT, such as those corresponding to peaks 5 to 15, have high correlation coefficients with the bioactivities, but are yet to be identified.

Specific estrogenic and immuno-modulatory effects of DBT
The estrogenic effects of DBT were investigated by determining the levels of phosphorylation of estrogen receptor α (ERα) and extracellular signal-regulated kinase 1/2 (ERK1/2) in cultured MCF-7 cells. In contrast to estrogen, DBT triggered the phosphorylation of ERα and ERK1/2 at both S118 and S167 in a time-dependent manner.
Although the activity of the estrogen-responsive element in pERE-Luc stably expressing MCF-7 cells was activated by extracts of either RA or RAS alone, or by a mixture of RA and RAS, the phosphorylation of ERα at S167 and of ERK1/2 were only found in DBT-treated cultures. Interestingly, the specific estrogenic effects of DBT were not only shown in the MCF-7 cells [31].
In cultured T-lymphocytes, the phosphorylation of the ERK 1 (about 42 kDa) and ERK 2 (about 44 kDa) was increased by DBT [30]. The induction was transient. An approximately eight-fold increase of ERK phosphorylation was detected 20 minutes after DBT was applied, whereas the phosphorylation was undetectable in the cultures treated with extracts of either RA or RAS alone [31]. Moreover, the phosphorylation of ERK in T-lymphocytes could not be activated by a simple mixture of extracts of RA and RAS. This result suggests that boiling RA and RAS together is essential for DBT to exert estrogenic effects.

Genomics
For decades, scientists mainly isolated pure chemicals from herbal extracts and then screened for biological activities and possible targets. This strategy does not garantee to isolate and/or identify active chemicals from well-known medicinal plants because a single chemical compound may not fully account for the overall effects of herbal extracts. Recent advances in genomics and proteomics have enriched our tool sets to reveal the complex nature of TCM decoctions. An experiment on DBT-regulated genes was carried out in our laboratory using gene Li YK et al. [25] DBT alleviated cardiac injury in ischemia reperfusion.
Wu DZ et al. [26] DBT increased the levels of RBC, WBC, and BMNC. Some DBT promoted the proliferation of BMNC and increased the level of CFU-Mix. DBT inhibited growth of GM-CFU, while the decoction-containing serum promoted growth of GM-CFU.
Kunming mice, GM-CFU DBT was administered orally; serum collected from abdominal aorta was added to an in vitro cultivating system of mouse hematopoietic progenitor cells.
Kunming mice, CFU-E DBT was administered orally; serum collected from abdominal aorta was added to an in vitro cultivating system of mouse hematopoietic progenitor cells.
Zhang YH et al. [30] Correlation coefficients between the data of 57 chemicals and the four bioassays  hours. The isolated mRNAs were analyzed using microarray, which is a quantitative method to investigate the change of mRNA expression profiles between the control and treatment groups. A total of 8064 genes were screened. Significant changes in gene expression were found after the treatment of DBT, RA or RAS (Table 3). A total of 883 genes were either up or down regulated by DBT treatment of which 403 genes were DBT-specific; 660 genes were regulated by RA treatment of which 172 genes were RA-specific; 1,062 genes were regulated by RAS treatment of which 473 genes were RAS-specific. In addition, 279 genes were commonly regulated by the extracts of DBT, RA or RAS. The genomic analysis demonstrated not only the activation effect of DBT in stimulating the proliferation and differentiation of the cultured osteoblasts but also a set of candidates of biomarkers that are specifically activated by DBT. These DBT-specific changes in gene expression may be useful in developing biomarkers for quality control of DBT. After identification of these DBTspecific genes and their roles, it will be easier to elucidate the action mechanism of DBT.

Conclusion
In verification studies of DBT decoction, the quality of herbal materials is ensured by authentication analysis. The ancient formula of RA to RAS ratio at 5:1 has been confirmed in both chemical composition and biological responses both in vivo and in vitro. Mechanism studies have also revealed some therapeutic effects of DBT. It is hoped that a systematic research and development approach, as exemplified in this article, will provide an effective method to develop Chinese herbal medicine products. *Significant changes of gene expression are defined as regulation, which can be either up-regulation when fluorescent signal in the sample was 200% greater than that of control, or down-regulation when the signal was 50% less than that of control.