Development of the quantified relationship between chemical compositions and antioxidant activity of Radix et Rhizoma Salviae Miltiorrhizae, Radix Puerariae Lobatae and the herbal formula of Danshengegen Tang (DGT) using liquid chromatography and chemometric techniques

Abstract

Herbal medicine (HM) is widely used for preventing and healing diseases in these years. It is necessary to ascertain its consistency, safety and efficacy through getting related information and evidences from scientific investigation. Correlation of the chemical composition with traditionally intended medication is one of the important targets. It is difficult to achieve this owing to the complexity of HM and the holistic concept of the working principle involved in disease treatment. Currently, the authentication of HM is based on "compound-oriented approach" and "pattern-oriented approach". The former approach mainly monitors the contents of several selected chemical markers for the measurement and it is widely accepted by authorized agencies in the world. As for the latter approach, it considered all the detectable components appeared in the chromatographic fingerprint for assessment. Still, the efficacy of HM cannot be reflected from these two approaches. The major reason is that the bioactive constituents are not the key concern. Thus, it is interested to explore the bioactive components in HM as they are responsible for the pharmacologic activity. This could aid in improving the quality control measures, getting better understanding on the therapeutic effect of HM and finding drug lead candidates from the natural sources. Usually, only a few so called active ingredients are considered even if the bioactivity aspect is under investigation. How to include more active ingredients and significant ones is still a great challenge to scientists. In this study, the quantified relationship between the chromatographic fingerprint and bioactivity of HM through the Quantitative Pattern-Activity Relationship (QPAR) approach was proposed and chemometric techniques were developed and applied to these two types of experimental data. The QPAR model thus established by us provides two pieces of crucial information about HM. They are (1) a model for predicting total bioactivity from the chromatographic fingerprint and (2) the features in the chromatographic profile responsible for the bioactivity. In this work, the Chinese herbal medicine Radix et Rhizoma Salviae Miltiorrhizae (DS), Radix Puerariae Lobatae (YG) as well as the herbal formula of DanshenGegen Tang (DGT) and a synthetic mixture system (MIX) were studied by both "compound-oriented approach" and "pattern-oriented approach". Their chemical fingerprints through HPLC-DAD-MS and RRLC-DAD instruments and antioxidant activities from FRAP assay were measured for data processing. In addition, the QPAR approach was applied to YG and MIX also to evaluate the performance of our QPAR methods developed and information mined. The bioactivity of HM is affected by the inherent variations of HM, the manufacturing process and other factors. Definitely, one of them is sample extraction. The extraction parameters affect the class of components and their quantities to be extracted. This leads to variation in the bioactivity level of the extract obtained when different sample preparation procedures are used. In this work, the optimal extraction condition of Radix et Rhizoma Salviae Miltiorrhizae (DS) in response to the antioxidant activity was explored and determined by response surface methodology (RSM). Through which, we investigated the relationship of bio-response and extraction parameters. In this way, the optimal antioxidant activity of DS was achieved by using 80% methanol with the volume of 90ml for 45minutes under ultrasound assisted extraction. Twenty DS samples collected from different provinces of P. R. China were investigated under the optimal extraction conditions. It was found that the antioxidant activity of DS is related to the amounts of phenolic acids instead of diterpenoids. The antioxidant activity of another HM, Radix Puerariae Lobatae (YG), was also studied. Both the bioactivity and chemical analysis were carried out on seventy eight YG samples. Through these two types of experimental data acquired, the quantified relationship between the chromatographic fingerprint and antioxidant activity of YG was established using the chemometric methods, PLS and UVE-PLS. Through which, on top of the chemical composition information, the antioxidant activity level of any YG sample including the unknown one can be reflected from its chromatographic fingerprint. The performance of determining the antioxidant activity level from the established relationship is good with the root mean square error for training set of 76.88 (6.37%) and the root mean square error of prediction for validation set of 79.59 (6.60%). The correlation coefficients of the predicted and experimental antioxidant capacities of the samples in the training and validation sets are 0.9318 and 0.8926 respectively. Exploration of bioactive components in YG was further attempted through establishing the QPAR relationship by another chemometrics algorithm, Target Projection coupled with Selectivity Ratio (TP/SR). The potential bioactive candidates were revealed at three regions. Their respective retention times are (G1) 12.5min, (G2) 13.3min and (G3) 15.6min. This prediction was proven by our experimental fractionation of YG samples. Furthermore, the discovery of bioactive components was done on another system which we called it as synthetic mixture system (MIX). They were made up of twelve different classes of reference compounds commonly found in HM. The most active components in this system were picked up by this algorithm. They were gallic acid (GA) and quercetin (QT). These findings show that our quantified relationship of chemical fingerprint and bioactivity could aid in studying the pharmacological activity of HM in more detail, drug discovery from the natural source, and upgrading the quality assessment HM with both chemical composition and bioactivity considered at the same time. In another study, the optimal ratio of two component herbs, DS and YG, in the DGT preparation was determined with the application of the chromatographic technique and in-vitro antioxidant assay. Rapid resolution liquid chromatography (RRLC) was utilized in this investigation because of its rapid analysis, improved resolution and sensitivity. It helped markedly in analyzing complex systems like DGT with many components involved. Concerning both chemical composition and antioxidant activity, the best combination ratio of DS and YG was determined to be 6:4. The outcomes and methods proposed could provide a useful tool in finding out the optimal combination of the component herbs in other herbal formulae.