DNA Microarray-Based Screening and Characterization of Traditional Chinese Medicine

The application of DNA microarray assay (DMA) has entered a new era owing to recent innovations in omics technologies. This review summarizes recent applications of DMA-based gene expression profiling by focusing on the screening and characterization of traditional Chinese medicine. First, herbs, mushrooms, and dietary plants analyzed by DMA along with their effective components and their biological/physiological effects are summarized and discussed by examining their comprehensive list and a list of representative effective chemicals. Second, the mechanisms of action of traditional Chinese medicine are summarized by examining the genes and pathways responsible for the action, the cell functions involved in the action, and the activities found by DMA (silent estrogens). Third, applications of DMA for traditional Chinese medicine are discussed by examining reported examples and new protocols for its use in quality control. Further innovations in the signaling pathway-based evaluation of beneficial effects and the assessment of potential risks of traditional Chinese medicine are expected, just as are observed in other closely related fields, such as the therapeutic, environmental, nutritional, and pharmacological fields.


Introduction
Herbal medicine is an important part of the medical practices in traditional Chinese medicine (TCM), and consists of a variety of plant species [1,2]. While the effectiveness of herbal medicine has sometimes been considered doubtful [3,4], this could be due in part to the difficulty of controlling the quality of herbs and the amounts of their effective components. Thus, various methods have been used to confirm their efficacy and identify the effective components [5]. Modern technologies, such as high-performance liquid chromatography (HPLC) and quantitative reverse-transcription polymerase chain reaction (qRT-PCR), have been continually developed and utilized to replace conventional technologies for the comprehensive and cost-effective quality control of herbal medicine [6]. Although recent progress in the human and other genome projects and the development of a variety of omics technologies-such as genomics and transcriptomics-have contributed to the identification and utilization of effective components and quality control/authentication of medicinal plants [7][8][9][10][11], we are still searching for the best strategy for the maximum utilization of herbal medicine. This review focuses on DNA microarray-based gene expression profiling, a key technology in transcriptomics, and shows how researchers used this technology for the screening and characterization of TCM. In particular, cross-examination of the data on TCMs, and their constituent herbs, mushrooms, and dietary plants, has become quite important to evaluate the knowledge accumulated and to assess the advantages/disadvantages of TCM and its effective applications.
TCM includes practices such as acupuncture, moxibustion, Chinese herbal medicine, tui na, dietary therapy, tai chi, and qi gong, which is rooted in the ancient philosophy of Taoism and is more than 2500 years old [12]. The original TCM has been further developed and modified to adapt it to people of various nationalities and genetic backgrounds (giving rise to different types of health problems, dietary and nutritional customs/practices, and ways of thinking and beliefs about medicine), based on variations in the types of herbs and their ingredients in countries such as Japan and Korea, where Kampo and traditional Korean medicine (TKM), respectively, were developed. Therefore, herbal medicine includes a number of plant species and ways of processing them. Furthermore, herbal medicine is often used with other constituents such as mushrooms and dietary plants, and thus their extracts and effective chemicals are also discussed here.
DNA microarrays are a type of biotechnological device used to detect alterations in genomic DNA and mRNA and to monitor genes and their expressions associated with various functions; thus, they have been widely used in basic research and industrial research/development (reviewed by Kiyama & Zhu [13]). DNA microarray assay (DMA) has been used to screen and characterize useful materials among mixtures of chemicals and extracts of natural resources including plants. DMA has advantages and disadvantages compared with other technologies. It has been used as a diagnostic device, such as for the genotyping of drug-metabolizing genes and predicting the metastatic risk of breast cancer, which is attributable to its unique characteristic of providing sufficient complexity to differentiate the variations needed for diagnosis and the reliability needed to predict genotypes/gene expression profiles accurately.    Extracts from herbs have been analyzed by DMA, including the following: extracts of the whole, or parts such as flower and leaves, of alkanet root, American ginseng, barbed skullcap, beach vitex, beth root, black cohosh, cancer bush, Chinese figwort, boneset, dong quai, field horsetail, greater celandine, kava, leigongteng, moutan, mum, orchid, pink lapacho, purple coneflower, salai, St. John's wort, turmeric, and wild yam. Meanwhile, the extracts of root, radix or rhizome were made from Chinese goldthread, Chinese peony, cistanche, danshen, goldthread, huang-qi, kudzu, and sheng-di-huang; the extracts of seeds were from lotus; or the extracts enriched in essential oil were from lovage and turmeric.

Herbal Medicine, Effective Chemicals, and Their Effects
The extracts from mushrooms were also analyzed after extraction of the whole or the fruiting body, such as that from buna-shimeji, caterpillar fungus, common mushroom, himematsutake, hiratake, lingzhi, lumpy bracket, maitake, and turkey tail; after extraction of the mycelium, such as that from caterpillar fungus, himematsutake, lingzhi, lumpy bracket, maitake, shiitake, and turkey tail; or after extraction of polysaccharides from lingzhi or triterpenes from hiratake, hoelen, and lingzhi.
On the other hand, mixtures of chemicals analyzed by DMA are: grape antioxidant dietary fiber (roughage/dietary fiber), oil palm phenolics, phytosterol mixtures (mixtures of steroid compounds), plant phospholipid/lipid conjugates, polysaccharides, and polyunsaturated fatty acids (PUFAs).

Mechanisms of Action by Traditional Chinese Medicine
DNA microarrays for gene expression analysis can be categorized into two types, global and focused DNA microarrays, based on their application [13,182]. Global DNA microarrays contain thousands to hundreds of thousands of probes representing some or all of the cDNA, expressed sequence tags (ESTs), and various types of expression marker, such as those for the estimation of mRNA copy numbers within cells. Meanwhile, focused DNA microarrays contain a few dozen to thousands of probes designed for specific purposes, such as the study of tissue/cell-type specificity, functional specificity, and expression profiling. Focused DNA microarrays are sometimes more appropriate for the study of the mechanisms of action when the action is known, such as in the case of comparative risk assessment of chemicals and the prediction of cancer metastatic risks.

Genes and Pathways Responsible for the Action
The signaling pathways analyzed by DMA are as follows (see Kiyama & Zhu [13]; Kiyama et al. [183]): MAPK (such as G protein-coupled receptor (GPCR)/MAPK, MAPK/c-Jun N-terminal kinase (JNK), and NF-κB/MAPK/ERK) and other (such as angiogenesis, ErbB/human epidermal growth factor receptor (HER), nuclear receptor, and ubiquitin/proteasome) signaling pathways, or apoptosis pathways (such as those for death receptor, infectious response, and p53-dependent apoptosis), autophagy pathways (such as those for phosphatidylinositol-3-kinase (PI3K)/Akt/mTOR signaling and starvation stress response), cell cycle/DNA damage pathways (such as G1/S checkpoint and G2/M DNA damage checkpoint signaling pathways), cellular metabolism pathways (such as AMP-activated protein kinase (AMPK) and insulin receptor signaling pathways), chromatin/epigenetic regulation pathways (such as those for DNA methylation, heterochromatin, and histone modification), cytoskeletal regulation and adhesion pathways (such as those related to actin, adherens junction, and microtubule dynamics), development and differentiation pathways (such as hedgehog, Notch, TGF-β, and Wnt/β-catenin signaling pathways), immunology and inflammation pathways (such as those for B-cell receptor signaling, cytokine receptor signaling, inflammatory response, rheumatoid arthritis, T-cell activation, and TLR-induced immune response), neuroscience pathways (such as Alzheimer's disease-and Parkinson's disease-related signaling pathways) and translational control pathways (such as eIF2, eIF4/P70S6K, and mTOR signaling pathways).
Since genes and pathways responsible for the action of TCM are related to various cell functions, it is almost impossible to understand the mechanisms of action just by studying the mixture of chemicals.

Cell Functions Involved in the Action
The major cell functions analyzed by DMA for TCM include: adipogenesis, anti-atherosclerosis, anti-carcinogenesis, anti-inflammation, apoptosis, carcinogenesis, chemoprevention, hypoxia, and life-span extension (Table 2; Figure 1).
Adipogenesis is a cellular differentiation process in whichpreadipocytes are transformed into differentiated adipocyte cells, and involves features such as morphological change, growth arrest, lipogenic gene expression, and the production of hormones and growth factors (such as leptin and TNF-α). Among the components found in the extract of Toddalia asiatica, aculeatin was found to promote the differentiation of mouse 3T3-L1 preadipocytes into adipocytes [132]. DMA revealed the involvement of PPAR-γ target genes in the process of activation by aculeatin, which is not a ligand of PPAR-γ, suggesting the presence of additional signaling mechanisms.
Atherosclerosis is a chronic inflammatory response of white blood cells in arterial blood vessels, which is promoted by low-density lipoproteins (LDLs), carriers of cholesterol, and triglycerides, and results in the formation of atherosclerotic plaques that are rich in macrophages and foam cells. Estrogenic activity was detected by DMA-based gene expression profiling in the extract of Agaricus blazei, which was attributable to brefeldin A [138]. The extract has no estrogen receptor-dependent cell proliferation activity, while showing activation of estrogen signaling (such as activation of ERK, Akt and P70S6K) and beneficial effects for patients with high levels of oxidized LDLs (see Section 3.3). Carcinogenesis, alternatively referred to as oncogenesis or tumorigenesis, is a process by which normal cells are transformed into cancer cells characterized by uncontrolled cell division; it involves Carcinogenesis, alternatively referred to as oncogenesis or tumorigenesis, is a process by which normal cells are transformed into cancer cells characterized by uncontrolled cell division; it involves a progression of changes at the cellular, genetic, and epigenetic levels. Several chemicals exhibiting anti-carcinogenic effects were isolated or identified from natural products, such as 3,3 -diindolylmethane from cruciferous vegetables [145], grifolin from Albatrellus confluens [153], paeoniflorin from Paeonia lactiflora [161], and quercetin from various dietary plants [161], and further analyzed by DMA. 3,3 -Diindolylmethane is estrogenic and shows gene expression profiles favoring tumor promotion [145]. Grifolin acts negatively against the cell cycle and cell growth through inhibiting ERK and Rb pathways, and downregulates the expression of cyclin D1, cyclin E, and CDK4 (a gene for a cyclin-dependent kinase), and upregulates the expression of CKI (a CDK inhibitor gene) [153]. Paeoniflorin enhances the expression of HSP70, which helps to protect cells from stress, and modulates the expression of CDC2, FOSL1, and EGR1, regulators of cell growth and proliferation [161]. Quercetin, on the other hand, induces p53-independent apoptosis by enhancing the expression of death-receptor or TNFR signaling genes, such as the genes for caspase-10, DFF45, FAS, IκBα, IL1R (Interleukin-1 receptor), TNFR1, and TRAILR [171].
Inflammation is a protective response to cell injury, and involves the local vascular system, the immune system, and various cells within the injured tissue. Ergosterol peroxide produced by Sarcodon aspratus suppresses inflammatory response in macrophages by inhibiting TNF-α secretion and down-regulating the expression of interleukin1α/β (IL-1α/β) through pathways such as C/EBPβ, ERK, JNK, MAPK, and NF-κB [147].
Apoptosis is the process of programmed cell death that may occur in multicellular organisms in response to various stresses, such as heat, hypoxia, increased intracellular calcium concentration, nutrient deprivation, receptor-ligand binding, radiation, and viral infection. Several chemicals are related to the promotion of apoptosis and thus have been used as effective components in herbal medicine. Emodin extracted from the rhizomes of Rheum palmatum showed testicular toxicity, including the induction of apoptosis, most likely through pathways such as IGF-1, TGF/Wnt, and TNFR1 signaling [146]. β-Hydroxyisovalerylshikonin extracted from Lithospermum erythrorhizon is an inhibitor of protein-tyrosine kinases and induces apoptosis by suppressing TRAP1, a TNF-associated protein and a member of the HSPs, as well as the production of ROS [155]. Tanshinone IIA found in the root of Salvia miltiorrhiza induces peroxisome proliferator-activated receptor (PXR)/NF-κB/CCL2-mediated apoptosis in leukemia cells [180]. 2,4,3 ,5 -Tetramethoxystilbene extracted from fruit, berries, and grapes is a derivative of resveratrol and a strong inducer of apoptosis by increasing the expression of tubulin, stress response, and pro-apoptotic genes [178].
Chemoprevention refers to the administration of a medication, such as drugs and vitamins, for the purpose of preventing disease or infection, and various chemicals have been developed especially for cancer chemoprevention. Myricetin [158] and sulforaphane [177] isolated from dietary plants show chemopreventive activity against cancer through activating Nrf2-mediated antioxidant response or PI3K/Akt signaling pathways, respectively.
Hypoxia is a condition in which a cell is deprived of adequate oxygen supply and has been shown to stimulate various biological and physiological responses. Paeonol isolated from Paeonia suffruticosa induces the expression of hypoxia-inducible genes, including hypoxia-inducible factor 1 (HIF-1)-target genes, through suppressing the NF-κB signaling pathway and inhibiting amyloid precursor protein (APP) activity [162].
Life extension has been studied in terms of slowing down or reversing the processes of aging in order to extend both the maximum and the average lifespan, and the effects of anti-aging products, nutrition, physical fitness, skin care, hormone replacements, vitamins, supplements, and herbs have been examined. Diallyl trisulfide isolated from garlic increases the longevity of nematodes through activation of the pro-longevity transcription factor gene skn-1 and the products of its target genes [144].
Conditions such as chronic (arthritis, asthma, cancer, diabetes, and viral diseases) and neurodegenerative (Parkinson's and Alzheimer's diseases) diseases have been treated with TCM [1], among which some were investigated by DMA and explored by animal tests and/or clinical studies to eventually achieve clinical applications. Other than the cell functions discussed above, the diseases with extensive impacts were also investigated. For example, antidepressant, anti-diabetic, anti-obesity, neuromodulation, and neuroprotection effects, and the treatments of neurological, Parkinson's, and Alzheimer's diseases associated with TMC and/or constituent herbs/mushrooms/dietary plants were studied by means of DMA (Table 1), or their effective components, such as ginsenosides (for diabetes), (−)-hydroxycitric acid (for obesity), obovatol (for neuroinflammation), and salvianolic acid B (for neuroprotection), were studied by means of DMA (Table 2).

Activities Found by DNA Microarray Assays (Silent Estrogens)
Activities found by DMA are often detected as cell signals in specific pathways, such as angiogenesis, ErbB/HER, MAPK, nuclear receptor, and ubiquitin/proteasome signaling pathways, and/or in cell functions, such as apoptosis, autophagy, cell cycle/DNA damage/cytoskeletal formation, cellular metabolism, chromatin/epigenesis regulation, development/differentiation, immunology/inflammation response, neurological diseases, and translational control [183]. While most of these cell signaling pathways and cell functions can be detected by other technologies, there might be some activities that can be detected exclusively by DMA. One such activity is by a group of estrogens, silent estrogens, which show estrogenic gene expression profiles without showing positive effects on cell proliferation [13].
Estrogen is a female hormone that is responsible for various biological and physiological activities, including receptor-mediated stimulation of the proliferation of cells in tissues such as the breast and ovary. Several chemicals and mixtures of chemicals, such as brefeldin A [138], licorice extracts [150], and oil degradation products [184], were found to show gene expression profiles similar to that for estrogen, although they did not stimulate the proliferation of estrogen receptor-positive breast cancer MCF-7 cells. Although the signaling pathway for cell proliferation could theoretically be separated from those for other cell functions, this separation has not been possible because most of the cells examined for estrogenic activity contain estrogen receptors and the technologies used were not suited to such a purpose. Recent findings about more complicated signaling pathways/networks, such as autocrine/paracrine/homeostatic networks and crosstalk/bypassing of cell signals, include pathways not necessarily involving cell proliferation or the cells containing estrogen receptors [185,186]. Thus, estrogenic activity can be detected even for silent estrogens because DMA can separate various signaling pathways, and the similarity of chemicals can be analyzed at the levels of gene expression and cell signaling.
Chinese medicinal plants were analyzed by DNA-based technologies, such as those detecting amplified fragment length polymorphism (AFLP), cleaved amplified polymorphic sequence (CAPS), inter-simple sequence repeat (ISSR), random amplified polymorphic DNA (RAPD), restriction fragment length polymorphism (RFLP), and simple sequence repeat (SSR), and those by amplification refractory mutation system (ARMS), DNA amplification fingerprinting (DAF), hybridization, microarray assay and sequencing [8], or by assays at the cell/tissue/animal levels, such as gene knockout and cell membrane chromatography and transgenics [187].
The literature surveyed for transcriptomics using DMA in the study of 297 frequently used medicinal herbs in China showed that most of the studies focused on finding their biological effects [10]. However, there are cases where DMA was applied to screen effective components and for quality control of TCM. In this section, we discuss how DMA has been used for quality control of TCM.
Additional applications of effective chemicals in TCM analyzed by DMA include patent filing, drug discovery, and clinical trials. An effective chemical identified and characterized by these applications can be patented as a new chemical, or otherwise as a known chemical with a new application.

DNA Microarray Assays for Quality Control of Traditional Chinese Medicine
DMA has been used to detect and evaluate various activities of pure chemicals and mixtures of chemicals [13,188]. When DMA was applied for the study of TCM, gene sets specific to herbs were selected. For example, a set of 55 genes were screened by DMA in order to understand the effect of Qingfei Xiaoyan Wan formula on asthma by the regulation of gene/protein networks [80]. A set of 92 genes was found to be differentially regulated by Toki-shakuyaku-san, a formula effective for circulation problems [89]. A set of nine marker genes was initially screened by DMA and subsequently confirmed by RT-PCR to assess the batch-to-batch consistency of the biological effects of ISF-1, a formula used for the management of post-stroke disorders [6].
Once gene sets have been selected, they can be used for screening and/or quality control of the herbs. For example, DMA was used to screen TCM species with inhibitory effects on Cytochrome P450 (CYP450) intended to treat HIV infection [189]. High-throughput DMA were applied to screen for anti-mitotic effects (independent of toxicity) on the proliferation of MDA-MB-231 cells from 897 aqueous extracts of commonly used natural products, and less than 1.34% of the extracts tested showed growth inhibitory properties at a concentration of less than 0.0183 mg/mL [27]. The DMA based on the yeast transcriptome was used for quality control of the extracts of Equisetum arvense [30].
Specific activity was examined to evaluate the quality of materials in food and supplements. For example, estrogenic activity was examined by DMA using a customized DNA microarray containing 172 estrogen-responsive genes in order to evaluate food materials, such as phytoestrogens [127] and ginsenosides [148], and for the extracts of plants and mushrooms, such as soybeans [127], Glycyrrhiza glabra [150], and Agaricus blazei [138]. The estrogen-responsive genes were further classified into six functional groups (enzymes, signaling, proliferation, transcription, transport, and others) and some showed preferences for specific groups [183].

Protocols of DNA Microarray Assays for Quality Control
Schemes of new DMA-based protocols for quality control of herbal extracts are summarized in Figure 2. A simplified protocol is based on gene expression profiles of different sources of herbal extracts (A1 to A5), which are compared with that of a standard (S) by correlation coefficients (R-values) based on linear regression (Figure 2A). Deviations (such as in A3, Figure 2A) can be detected by comparing R-values. While gene sets for expression profiling can be used without selection (Figure 2A), the genes having specific cell functions could be used to improve the level of quality control ( Figure 2B,C), where the gene sets (G1 to G3) can be selected arbitrarily, such as those showing stable (G1), less stable (G2) or unstable (G3) reproducibility when various lots (A1 and A2) of the herbal extracts are compared with a standard ( Figure 2B). Alternatively, gene sets can be selected according to gene functions ( Figure 2C), where profiling can be performed with functionally grouped genes (F1 to F3) to give a protocol of efficacy-based quality control of herbal extracts. The advantages for using these protocols are that: (1) cell function-based gene expression profiling is good for efficacy-based quality control; (2) gene sets can be selected depending on the purpose; and (3) various activities can be monitored by different gene sets. to F3) to give a protocol of efficacy-based quality control of herbal extracts. The advantages for using these protocols are that: (1) cell function-based gene expression profiling is good for efficacy-based quality control; (2) gene sets can be selected depending on the purpose; and (3) various activities can be monitored by different gene sets.

Conclusions and Perspectives
Owing to the progress of new biotechnological tools, new approaches have been developed, including systems biology, signal transduction study, and RNA-sequencing-based whole genome/exome analysis, which provide us with massive amounts of omics data. However, we still do not know how to use these tools effectively. For example, we now know a number of chemicals showing estrogenic activity and the number is steadily increasing because they affect cells, tissues, and organs through complex and novel pathways of cell signaling, such as intracellular signal transduction, signal crosstalk/bypassing, and intercellular networks of autocrine/paracrine signaling [185,186]. This complex status would be the same as in the study and the application of TCM because, since estrogenic activity is one of the most important activities of effective components in herbs and has been shown to affect the human body, the findings about estrogen would be true for more complex statuses, which include other hormones and growth factors. Chemicals having other hormonal and growth-factor activities could show such complexities, and thus new approaches are needed to understand the effects of chemicals. Other than the pharmacological use of chemicals, DMA-based gene expression profiling and pathway-based testing of chemicals have been developed for the diagnosis of diseases and risk assessment of endocrine disruptors, where the development of new types of diagnostic tool is in progress, such as the in vitro diagnostic multivariate index assay (IVDMIA), and toxicity pathway-based risk management protocols, such as those proposed by the U.S. National Research Council [190]. Thus, a pathway-based evaluation of beneficial effects and the assessment of potential risks by means of omics technologies are needed for the study and development of TCM.