Abstract
Objective. Chinese Medicinal Properties (CMP) play a vital role in theoretical research and clinical practice. However, the traditional CMP system is subjective, qualitative, fixed, inconsistent, and obscured. Nowadays, quantifying CMP research achieved a notable progress. This study aims to review and reflect the relevance between qualitative CMP and quantitative material components. Methods. A raw literature search was performed firstly in CNKI and Pubmed database to get a rough idea on the general advances in measuring CMP. Then, a strict literature search and data extraction from two dependent research studies were performed to analyze the relevance and discrimination between CMP and material components. Results. The quantitative CMP research mainly focused on the microelements and chemical compositions. The largest microelements research listed 747 Chinese Materia Medica (CMM) (6780 flavors) and 120,000 element data. The measurement of chemical composition of CMM has risen rapidly in the 1990s and continues till the present. Thirty-seven articles were finally identified for the relevance analysis of CMP and material components. Of these, 18 and 19 articles correspondingly focused on the chemical compositions and microelements, and 26 and 11 articles correspondingly focused on their correlation and discrimination relationship. The most commonly used method for correlation analysis is intuitive analysis. The support vector machine maybe highly efficient and would act as the preferred method in discriminant analysis. Twelve (67%) and 5 (26%) articles’ data came from the literature search in chemical compositions and microelement research studies. Four studies indicated that the research objects are the basic substances and material basis of CMP, 15 articles claimed that the chemical compositions were significantly related to CMP, 12 research studies concluded that the regularity and causality were identified between the research objects and CMP, and 9 research studies successfully established discriminant models for CMP basing on the detected substances. Conclusions. The relevance research between qualitative CMP and quantitative material components achieved a positive progress, though it is weak and defective. Standardizing the qualitative CMP system, establishing series comprehensive databases for the material components, innovating statistical and data mining methods, and integrating doctors’ experiences are important and feasible for future research.
1. Introduction
The medicinal property (or nature, categorization) is the basis of prescription and efficacy in Traditional Chinese Medicine (TCM), which plays a vital role in theoretical research and clinical practice. It was established and enriched by many highly skilled and respectable ancient Chinese physicians and pharmacists in the last thousands of years. Nowadays, the Chinese Medicinal Properties (CMP) system contains four Qi (or four natures, si qi), five flavors (wu wei), ascending, descending, floating and sinking (sheng jiang fu chen), channel entering (or meridians, gui jing), and toxicity [1]. This system has made great contribution to classify Chinese Materia Medica (CMM), activate the efficacy of drugs and formulas, heal body, and even evaluate and diagnose diseases.
However, the current CMP system has obvious shortcomings. It is (a) subjective, that is, artificially classified by ancient physicians and pharmacists just based on their own feelings or experience in clinical practice; (b) qualitative, that is, described only with text rather than data; (3) fixed, that is, rarely changed after proposed by the original person; (4) inconsistent, that is, the medicinal properties for some herbs were described differently in different medical books; and (5) the latent characteristics of medicinal properties are unknown due to outdated natural techniques in ancient times. These limitations obviously hinder the communication of TCM and CMM, both internally and externally to Western medicine, as well as global doctors and researchers and, secondly, decrease the technical capabilities of clinicians, especially for beginners and integrative doctors who need to devote more energies to realize and think about these vague drug characteristics.
It is, however, gratifying that these limitations may not be insurmountable. Quantifying the properties of CMM is one important and feasible method. CMM quantification is a scientific, standardized, specific, micro, and quantifiable modern research category, which contains dose thresholds from a scientific view and syndrome differentiation and dose from philosophical and artistic views [2]. Of those, quantifying CMP is the core set which intends to measure, standardize, apply, and visualize the latent natures and representative characterizations of CMM. It transforms the traditional qualitative CMP system into the modern quantitative one. This study aims to analyze the advances and limitations of the current quantitative CMP research and provide advice for future research.
2. The General Advances in Measuring CMP
The material components of CMM are mainly divided into inorganic, organic, and physical compositions. In the <Chemistry of Chinese Materia Medica>, it was divided into more detailed categories: sugars, glycosides, terpenoids, phenylpropanoids, flavonoids, steroids and volatile oils, alkaloids, steroids, triterpenoids, and tannins, as well as fatty acids, organic phosphorus compounds, amino acids, cyclic peptides, proteins, enzymes, and minerals [3]. Currently, the quantitative CMP research mainly focused on the compositions measurement of CMM, which has made significant strides with the rapidly developing microscopic measurement technologies, especially in the last decades.
Since the 1960s, a large number of researchers have studied the active ingredients and basic substances of CMM. In 1973, Nozdryukhina discovered the biological activity efficacy of microelements in Chinese herbs [4]. This finding has greatly influenced the microresearch of CMM during the later decade, and plenty evidence were accumulated. In 2013, Li et al. [5] published series articles listed 747 CMM (6780 flavors) and 120,000 element data which came from 1036 literatures and 34 kinds of analytical technologies. Since the 1990s, with the development of chemical compositions detection technologies in China, the measurement of chemical compositions of CMM has risen rapidly and continues till the present. This research category has achieved greater progresses and accumulated more evidences than microelements research (Figure 1). However, as for the published articles in Pubmed, the difference is not so obvious, but the total number (no more than 25 per year) is significantly smaller than the articles in CNKI. Besides many professional papers, journals, and books, many chemical databases and TCM databases provided detailed information on the chemical compositions of CMM.
3. The Advances in the Relevance and Discrimination between CMP and Chemical Components
3.1. Search Strategy
The searches were performed in the CNKI and Pubmed database with predefined search strategies: “SU=(“si qi”+“wu wei”+“gui jing”+“sheng jiang fu chen”+“du xing”+“yao xing”+“gong xiao”+“xin”+“ ku”+“gan”+“xian”+“ping”+“han”+“ re”+“wen”+“liang”+ “suan”) (“liang hua”+“ding liang”+“pan bie”+“guan xi”+“guan lian”+“xiang guan”) AND AB=(“yuan su”+“wei liang yuan su”+“you ji hua xue”+“you xiao cheng fen”+“hua xue cheng fen”+“huo xing wu zhi”+“gan lei”+“tang lei”+“kun lei”+“huang tong lei”+“hui fa you”+“tie lei”+“sheng wu jian”+“zai ti lei”+“san tie lei”+“rou lei”)” limited in “Medicine” discipline category in CNKI database; “(((Quantitative[Title] OR Discrimination[Title] OR Relationship[Title])) AND (Elements[Title/Abstract] OR Microelement[Title/Abstract] OR Trace elements[Title/Abstract] OR Organic Chemistry[Title/Abstract] OR Active Ingredients[Title/Abstract] OR Chemical Ingredients[Title/Abstract] OR Active Substances[Title/Abstract])) AND (Four Qi[Title/Abstract] OR Five Flavors[Title/Abstract] OR channel entering[Title/Abstract] OR properties[Title/Abstract] OR nature[Title/Abstract])” in Pubmed database. The last retrieval time is Dec. 24, 2019.
3.2. Data Management and Extraction
The literature was included if it met all the following criteria: (a) analyzed the relevance or discrimination between CMP and material components; (b) adopted clear statistical methods or tools.
The literature was excluded if it met any of the following criteria: (a) only detected chemical components of CMM without relevance or discrimination analysis; (b) qualitative interpretations, hypotheses, speculations, or reviews of CMP without quantitative evidences; (c) duplicated publications; and (d) full text was not available.
Two researchers (HU Wen, HUANG Zhi-bang) independently searched and traced relevant references. After obtaining the preliminary information, they screened the literature according to the predefined inclusion and exclusion criteria. If the results are inconsistent, they presented them to the coordinators (HOU Zheng-kun, LIU Feng-bin) for further evaluation until an agreement was reached. Then, HU and HUANG independently extracted information from the literature with a uniform data form. HOU checked the final data and retracked and re-extracted the information if a difference or bias was perceived.
3.3. Main Results
3.3.1. Search Results
A total of 1580 and 272 initial records were obtained from CNKI and Pubmed database, respectively. Following the data extraction and management procedure, 35 Chinese documents and 2 English documents were finally included for analysis [6–42] (Figure 2). Of these, 18 [6–23] and 19 [23–42] articles respectively focused on the chemical compositions and microelements to analyze the correlation or discrimination with CMP (Tables 1 and 2).
3.3.2. Correlation, Discrimination, Quantitative Research, and Methods
Firstly, there are 26 articles focused on the correlation relationships between CMP and chemical compositions [6–9, 11, 14, 16–20, 22, 23, 37] or microelements [24–29, 31, 33–36, 39]. Among them, the most commonly used method is intuitive analysis [6, 7, 9, 17, 33, 34, 37], that is, the researcher investigates the macroscopic drug properties from a microscopic angle by synthesizing from the numeric data to find the correlations between microscopic substances and CMP. The other analysis methods include factor analysis, cluster analysis, logarithm, standard error, t-test, variance, and correlation analysis, which also aim to find the correlations between the chemical components and CMP based on the composition and content differences of CMM with different medicinal properties.
Secondly, there are 11 articles in which the discrimination relationships between CMP and chemical compositions [10, 12, 13, 15, 21] or microelements [30, 32, 38, 40–42] are studied. The research methods used include cluster analysis, principal component analysis, discriminant function, Fisher analysis, Bayes discriminant analysis, support vector machine, and so on. It is notable that Yang [10] established a logistic regression equation for four levels of organic components and identified the relationship between inorganic elements and cold-heat of herbs. The results confirmed that the high Fe and low Mn content are the character of cold CMM and low Fe and high Mn content are the character of hot CMM. In the prediction and discrimination research on the 80 CMM properties, basing on the types and contents of inorganic elements, the accuracy rates of Fisher, Bayes, and support vector machine methods were 60%, 78.75%, and 95%, correspondingly. Qin et al. [15] established a support vector machine discriminant model based on the volatile components of 28 antiepidemic drugs, and the accuracy rate was 93.6%. The limited evidence showed that the support vector machine maybe highly efficient and would act as the preferred method in discriminant analysis. In addition, Hu et al. [28] analyzed the relationship between four characteristics and 32 kinds of trace elements in 115 Chinese herbs with linear discriminant analysis and established a model with 70.34% correct classification for cool, warm, and neutral. Liang [37] proposed that the four Qi and five flavors should be quantified with general acid-base theory, the principle of soft-hard and acid-base, Frontline orbit theory, and the relationship between electronegativity difference and ionic properties of the chemical bond.
3.3.3. Data Source of Chemical Composition of CMP
Of the 18 articles which analyzed chemically active ingredients, 12 (67%) research studies’ data came from literature search [6–10, 13, 14, 16–18, 20, 21], and only 6 (33%) research studies’s data came from self-test experiments [11, 12, 15, 19, 22, 23]. Of the 19 articles which analyzed microelements, the number is 5 (26%) [24–26, 34, 40] and 14 (74%), correspondingly, but the tests mainly concentrated in the 1990s and early 2000s.
3.3.4. Main Findings and Conclusions of Research Studies
The abovementioned quantitative research explores the CMP from multiple angles, such as chemical active ingredients, microelements, protein content, molecular charge, molecular weight, and so on. Four literatures indicated that the research objects are the basic substances and material basis of nonrepresentational CMP [12, 28, 35, 39], 15 articles claimed that the chemical compositions of CMM were significantly related to the CMP [6, 8–11, 14, 15, 20, 22, 23, 26, 27, 30, 32, 37], and 12 researches concluded that the regularity and causality existed between the research objects and CMP [7, 16–19, 24, 25, 29, 31, 33, 34, 36]. For example, Liu et al. [16] found that the CMM with diuresis efficacy which attributed in upper Jiao (上焦) major contain flavonoids and followed with terpenoids, in middle Jiao (中焦) major contain terpenoids and followed with steroidal, and in lower Jiao (下焦) major contain terpenoids and followed with flavonoids. Jia et al. [18] pointed out that composite Chinese Medicinals mainly owned the nature of being cold and cool and mainly contain volatile oils, flavonoids, terpenes, alkaloids, and acids, which are closely associated with their efficacy. Gong and Zhang proposed that the contents of Fe, Zn, and Cu are higher in CMM with cold property and salty flavor [24] and Zn, Fe, Cu, and Mn contents are higher in CMM which enter the liver channel [25]. Also, 9 research studies established discriminant function models for CMP, which can successfully discriminate CMP based on the detected substances [10, 13, 15, 21, 30, 32, 38, 40, 41].
4. Discussion
As mentioned above, a number of research studies have been performed on the measuring of material components of CMM, conversely, less focused on the relationships between the material components and CMP. In terms of data sources, many studies abandoned self-test but adopt literature reviews. Even so, many literatures did not strictly define the data sources and inclusion and exclusion criteria. As for the analysis methods, the earlier studies usually performed analysis with basic statistical methods and indicators, such as frequency, logarithm, mean, standard error, standard deviation, and so on; now, more and more studies use complex methods, such as factor analysis, cluster analysis, support vector machine, and Fisher analysis, that is, some progress in methodologies has been achieved. But, more advanced methods, and maybe more suitable for the nonlinear and black-box characters of CMP, such as support vector, neural network analysis, decision tree, and random forest, are still rarely being used. Focusing on the research results, it is clear that the chemical component of CMM is the material basis of CMP, and the regularity and causality do existed between them which can be quantified by discriminant functions.
Another important feature is that the existing quantitative research studies mainly focused on the microelements and chemical compositions of the four Qi (cold, heat) and five flavors (Xin) in CMP, but little on the other microdata and CMP characters, such as channel entering (gui jing) and toxicity, especially, no research was interested in the ascending, descending, and floating and sinking (sheng jiang fu chen). For example, as for the channel entering, Wang et al. [20] found that the CMM in the heart channel contains plenty of terpenoids, flavonoids, and volatile oils, in the liver channel contains plenty of flavonoids, organic acids, and terpenoids, and in the lung channel contains plenty of flavonoids, volatile oils, glycosides, and terpenoids and concluded that there are explicit interrelations of channel tropism of CMM in the chemical constituent. Wang [43] considered the biological basis of different Chinese meridian herbs with network pharmacology and molecular docking methods and found special proteins related to 12 channels. As for the toxicity, Li et al. [44] established a quantitative structure-toxicity model for aconitine compounds basing on partial least squares methods; using the same methods, Xiao et al. [45] established a similar quantitative structure-toxicity model for norditerpenoid alkaloids, and Xin et al. [46] proposed an integrative model for pharmacodynamic research of the aconiti kusnezoffii radix basing on TK-TD from the perspectives of structure-activity and dose effect relationship. As for the processing methods of CMM, Zhong et al. [47] found that different ginger juices will lead to the different changes of the effective material group, for example, jatrorrhizine hydrochloride, coptisine hydrochloride, palmatine hydrochloride, berberine hydrochloride, and epiberberine in Coptidis Rhizoma, resulting in the pharmacological effect and property difference between Coptidis Rhizoma processed with fresh ginger juice and Zingiberis Rhizoma juice.
In addition, some new values and directions had been found by the active researchers. Jing et al. [48] studied the nature-effect relationship of CMM with the main active components, targets, and protein interaction network and found the common characteristics of the bitter-liver combination and the specific characteristics of cold or warm medicinal properties at the molecular network level. Similar research studies were performed to find the nature-effect relationship of Salviae Miltiorrhizae Radix et Rhizomaand and Carthami Flos [49] and Curcumae Longae Rhizoma, Curcumae Radix, and Curcumae Rhizoma [50]. Wang [51] synthesized and analyzed the information on “chemical composition-biological activity-targets” of Acai (Euterpe Oleraceae Mart.) and Maca (Lepidium meyenii Walp.), two new foreign resources which did not exist in the classical CMM contents, and successful predicted and validated their four Qi, five flavors, ascending, descending, floating and sinking, channel entering, and efficacy. Guohui et al. [52] used the Mahalanobis distance learned by distance metric learning algorithm to measure the similarity of ultraviolet spectrum data of the existing 61 herbs and constructed their prediction and recognition model of the cold and hot nature with better predictive stability and extrapolation compared with the existing classical models. Besides, there is one study which explained the substantial base of the nature of Radix Scutellariae through investigating the impacts of the herb and its splitting components on the relevant enzyme expressions of the substance and energy metabolism in rats with cold and heat syndrome and found that aglycones and glycosides are the material base of the cold nature of Radix Scutellariae [53].
However, the current research studies have obvious limitations: firstly, the comprehensiveness and accuracy in the tremendous amount of chemical components data collection procedure directly influence the accuracy and reliability of the experimentation results. For example, the information selection and analysis and import of CMM sources, component categories, and compositions content could produce errors or selective bias and lead to inevitable deviations between the experimental results and true values. More seriously, many studies obtained the data from the secondary literature, 46% from our findings, and subjectively selected some chemical components to perform the analysis without regulation and interpretation. This undetected and unexpressed heterogeneity obviously harms the research credibility. Secondly, the qualities of the research methods used to analyze the relationships between the material components and CMP are various. For example, the subjective analysis method used in many studies is too dependent on the personal experience of experts and lacks credible standards. Using the difference analysis methods, such as standard deviation and t-test, it is difficult to solve the complex nonlinearity and diversity problems in the discrimination, prediction, and quantification of CMP. Even for the advanced statistical methods, such as factor analysis, cluster analysis, support vector machine, and Fisher analysis, their working efficacies are also not completely satisfied. Thirdly, there are great gaps between the laboratory researches and clinical practices, that is, all the abovementioned quantitative research studies of CMP are limited in the laboratories and the literature and rarely involved in the clinical services. Unfortunately, the microdata on CMP are not completely consistent with the macrosubjective judgments in real world. The busy physicians still highly rely on their own subjective experiences in clinical practice but not microcomplex objective data. This embarrassing situation makes the quantitative CMP research lack breakthrough and application.
The reasons for the abovementioned limitations are complexly. Firstly, the latent inner properties of CMM are too abstract and indelible, and the chemical components are obviously affected by the region, geology, climate, weather, material extraction time, and processing methods of CMM. Secondly, it is still impossible to accurately measure all the components of CMM with current detection methods, and the results are very susceptible to the technologies and tools. Thirdly, the lack of systematic human, material, and financial investments and managements results in the fragmentation of quantitative CMP research. Fourthly, the lack of multidisciplinary talents has limited the scope of related research, and it is difficult to break into new research fields. However, in any case, these problems are likely to be overcome in the near future.
5. Conclusions
In the long course of the history of traditional Chinese medicine, the properties system of Chinese Materia Medica has been established, intermingled, and enriched with the continuous contributions from many ancient doctors and pharmacists, which provides a solid theoretical basis for clinical practice. However, it cannot be evaded that the classical qualitative Chinese medicinal properties system has several deficiencies, which limit the recognition, application, and communication of the traditional Chinese medicine and Chinese Materia Medica. Therefore, measuring, standardizing, applying, and visualizing the Chinese medicinal properties are emphasized and advocated by more and more researchers. Currently, there are many research studies which measured the material components of Chinese Materia Medica and obtained huge information about the microelements and chemical compositions. In addition, there are limited studies which analyzed the relationships between the material components and Chinese medicinal properties. However, there are still many shortcomings in the current research studies, which are mainly related to the complex attributes of Chinese Materia Medica and imperfect measurement techniques, but they are still expected to be resolved. In future research, it is very important and also feasible to standardize the qualitative Chinese medicinal properties system, establish series comprehensive databases for the material components of Chinese Materia Medica, innovate the statistical methods for mining the relationships between the material components and Chinese medicinal properties, and develop a mixed system of Chinese medicinal properties for clinical applications.
Conflicts of Interest
The authors declare no conflicts of interest.
Authors’ Contributions
Prof. Hou ZK is the guarantor of integrity of the entire study and was responsible for the conception and design of the study and acquisition, analysis, and interpretation of data. Prof. LIU FB and Prof. LIANG YY are the advisers for the study. HU W, Ph.D., HUANG ZY, M.D., and HUANG ZB, Ph.D. are responsible for the data collection and analysis. All authors read and approved the final manuscript. The corresponding author attests that all listed authors meet authorship criteria and that no others meeting the criteria have been omitted.
Acknowledgments
This work was supported by the National Natural Science Foundation of China (No. 81774450), Science and Technology Planning Project of Guangdong Province, China (No. 2017A020215107), Pearl River S&T Nova Program of Guangzhou, China (No. 201710010077), and Key Science Technology Research of Health Qigong Management Center of General Administration of Sports of China (QG2017037). The sponsors had no role in study design, data collection, data analysis, data interpretation, or the writing of the report.