Quantitative determination, principal component analysis and discriminant analysis of eight marker compounds in crude and sweated Dipsaci Radix by HPLC-DAD

Abstract Context: Dipsaci Radix is derived from the dry root of Dipsacus asper Wall.ex Henry (Dipsacaceae). It has attracted increasing attention as one of the most popular and precious herbal medicines in clinical use. Objective: To develop a HPLC-DAD method for quantitative analysis and quality control of eight active components in crude and sweated Dipsaci Radix. Materials and methods: The eight components in Dipsaci Radix were analyzed by HPLC-DAD on an Agilent Eclipse XDB-C18 column within a gradient elution of acetonitrile and 0.05% formic acid aqueous solution. ESI-MS spectra were acquired on a triple quadrupole mass spectrometer. Validation was performed in order to demonstrate linearity, precision, repeatability, stability, and accuracy of the method. The results were processed with principal component analysis (PCA) and discriminant analysis (DA). Results: The eight components showed good linearity (R2 > 0.9991) in the ranges of 60.40–1208.00, 151.00–3020.00, 3.06–61.20, 30.76–615.20, 5.13–102.60, 10.17–203.40, 10.20–204.00, and 151.60–3032.00 mg/mL, respectively. The overall recoveries were in the range of 99.03–102.38%, with RSDs ranging from 1.89% to 4.05%. Through PCA, the degree of importance of the eight components in sequence was CA > AVI > IA > LA > LN > IC > IB > CaA. The crude and sweated Dipsaci Radix were distinguished obviously by DA. Discussion and conclusion: The method, using HPLC-DAD analysis in combination with PCA and DA, could provide a more comprehensive and quantitative chemical pattern recognition and quality evaluation to crude and sweated Dipsaci Radix.


Introduction
Dipsaci Radix, derived from the dry roots of Dipsacus asper Wall. ex Henry (Dipsacaceae) (Chinese Pharmacopoeia Commission 2015), is a traditional herbal medicine with a long history for the treatment of bone fractures and low back pain in China (Wong et al. 2007;Peng et al. 2010;Niu et al. 2012Niu et al. , 2015. Sweating, which used to be a traditional method of processing fresh Dipsaci Radix in the production area for drying (Jin et al. 2010), is no longer commonly adopted. So the effect of the sweating to Dipsaci Radix should be studied based on the perspective of the chemical compositions.
In the last decades, the phytochemistry of Dipsaci Radix has been extensively investigated, and the results indicate that loganic acid, chlorogenic acid, caffeic acid, loganin, isochlorogenic acid A, isochlorogenic acid B, isochlorogenic acid C, and asperosaponin VI are the main active components (Wei et al. 2011;Liu et al. 2012;Du et al. 2014;Ling et al. 2014). Pharmacological studies on the components showed that they all had various biological activities. Loganic acid has anti-inflammatory activity due to COX inhibition (Ram ırez-Cisneros et al. 2015), and could be used to reduce intraocular pressure (Szumny et al. 2015). Loganin and chlorogenic acid also have anti-inflammatory activity (Lou et al. 2015;Kim et al. 2015a). Caffeic acid inhibits lipid peroxidation (Kim et al. 2015b). Isochlorogenic A, B, C have anti-inflammatory and antimicrobial effects . Asperosaponin VI has antioxidant activity (Song et al. 2014), and could promote bone cell proliferation (Niu et al. 2011).
To our knowledge, previously reported analytical methods were employed to rapidly separate and identify lots of components in Dipsaci Radix (Wei et al. 2011;Liu et al. 2012;Du et al. 2014;Ling et al. 2014), but to quantify only several components in crude Dipsaci Radix Liu et al. 2011;Fan et al. 2013;Du et al. 2013;Zhao et al. 2013;Fan et al. 2015;Wang et al. 2015;Zhang et al. 2015;Du et al. 2016). In this study, a high performance liquid chromatography with diode array detector (HPLC-DAD) method was developed to quantify eight major bioactive components simultaneously in both the crude and its processed products. The method is simple, quick, and cheap with good reproducibility. It offers a new method that could be employed for the quality control of not only crude Dipsaci Radix, but also its processed products.

Chemicals and reagents
Acetonitrile of HPLC grade and methanol for analysis were provided by Tedia Co. (Fairfield, OH). Formic acid of reagent grade was purchased from Zhejiang SanYing Chemical Co., Ltd.

Plant material
Dipsaci Radix used in this study were collected from different origins (Table 1), the crude and sweated samples all were processed from the same batch of fresh herbs in the production area. The crude samples were produced after drying the fresh herbs directly by oven, and the sweated ones were produced after drying the fresh herbs pre-sweated to internal turning green. These herbal samples were authenticated by Professor Pingfan Lai (Zhejiang Chinese Medical University, Hangzhou, China). These specimens were preserved in the Research Center of TCM Processing Technology, Zhejiang Chinese Medical University.

Preparation of standard solutions
Primary stock standard solutions of the eight compounds were prepared in methanol with a concentration of 1208.0 lg/mL for LA, 3020.0 lg/mL for CA, 61.2 lg/mL for CaA, 615.2 lg/mL for LN, 102.6 lg/mL for IB, 203.4 lg/mL for IA, 204.0 lg/mL for IC and 3032.0 lg/mL for AVI. The standard stock solution was further diluted with methanol to make10 different concentrations including 1, 3/4, 1/2, 2/5, 3/10, 1/4, 1/5, 3/20, 1/10 and 1/20 of the original concentration. The solutions were filtered through a polyvinylidene difluoride filter of 0.45 lm and stored at 4 C.

Preparation of sample solutions
The dried powder of Dipsaci Radix (0.500 g, 80 mesh) was accurately weighed and added into dark brown calibrated flasks (100 mL). Methanol (25 mL) was added, the weight was accurately measured, and the sample was sonicated for 30 min. The solution was weighed again, and the loss in weight was made up with methanol. The supernatants were filtered through a 0.45 lm membrane prior to injection.

Calibration curves and limits of detection and quantification
The calibration curves were performed with 10 different concentrations in triplicate. The regression equations were established by plotting the peak area (y) versus concentration (x) of each analyte. The linearity was measured by correlation coefficient (R 2 ) values. Limit of detection (LOD) and quantification (LOQ) were determined by injecting a series of standard solutions until the signal-to-noise ratio (S/N) for each compound was 3 for LOD and 10 for LOQ, respectively. The results were given in Table 2.

Precision, repeatability and stability
The intra-and inter-day precision was determined by analyzing calibration samples during a single day and on six different days, respectively. The intra-day variation was determined by analyzing the six replicates on the same day and inter-day variation was determined on six consecutive days. Overall intra-and inter-day variations were less than 1.77%.
To further evaluate the repeatability of the developed assay, Dipsaci Radix was analyzed in six replicates as described above. The contents of eight compounds were calculated from the corresponding calibration curves. The relative standard deviations (RSDs) were taken as measurements of repeatability. Stability was tested with Dipsaci Radix at room temperature and analyzed at 0, 2, 4, 8, 12, 24, and 48 h within 2 days, As a result, the RSDs of repeatability test and stability test were both less than 3.45%. The results were given in Tables 3 and 4.

Accuracy
Accuracy was determined by the recovery test. An appropriate amount of Dipsaci Radix powder was weighed and spiked with 80%, 100% and 120% of known amount of each standard compound. They were then treated and analyzed as described above.
Each sample was analyzed in six replicates. The total amount of each analyte was calculated from the corresponding calibration curve. Mean recoveries of eight compounds were 99.03-102.38%. The results were given in Table 5.
Results and discussion

Optimization of chromatographic conditions
High performance liquid chromatography conditions including column type, column temperature, flow rate and mobile phases were assessed to accomplish the simultaneous separation of the eight analytes. The theoretical plate, symmetry factor and resolution were evaluated. To evaluate the suitability, three different columns, Zorbax Extend-C18, Eclipse XDB-C18 and Inertsil ODS-SP were compared with regard to the three analytical factors. As a result, the Eclipse XDB-C18 was the best for separation. Furthermore, other chromatographic variables were also optimized on the Eclipse XDB-C18 column, including mobile phases (water-methanol, water-acetonitrile and aqueous formic acid-acetonitrile), the column temperatures (20, 25, and 30 C) and the flow rates (0.8 and 1.0 mL/min). Eventually, the optimal separation was achieved on an Agilent Eclipse XDB-C18 (250 mm Â 4.6 mm, 5.0 lm) at a column temperature of 30 C with a flow rate of 1.0 mL/min. The HPLC chromatograms were shown in Figure 1.  Figure (Kouno et al. 1990).

Principal component analysis
Principal component analysis (PCA) is a well-known approach to give an interpretable overview of the main information in a multivariate dataset. It could generate fewer principal components (PCs) which are independent of the original variables but show linear combinations of them, and simultaneously explain most features of the aboriginal data ). The PCA was performed by using the contents of the eight compounds as the variances and the first 3 PCs were extracted with a cumulative contribution rate of 80.054%. The multiple regression models of each PC were obtained as follows: PC 1 ¼ 0.506X 1 þ 0.945X 2 À 0.672X 3 þ 0.398X 4 À 0.156X 5 þ 0.594X 6 À 0.076X 7 þ 0.759X 8 ; PC 2 ¼ 0.743X 1 þ 0.034X 2 À 0.140X 3 þ 0.857X 4 þ 0.706X 5 À 0.656X 6 À 0.096X 7 À 0.462X 8 ; PC 3 ¼ 0.174X 1 þ 0.193X 2 À 0.397X 3 þ 0.050X 4 À 0.126X 5 þ 0.273X 6 þ 0.946X 7 À 0.124X 8 , where Xi was the standardized area of common peak i. From the point of variance contribution rate, when eigenvalue k 1 ¼ 2.718, PC 1 contribution rate was 33.972%, which was the largest, and contained the most information. In three main components (Figure 3), the first principal component (PC 1 ) coefficient in sequence was X 2 > X 8 > X 6 > X 1 > X 4 > X 7 > X 5 > X 3 , the coefficient represent the

Conclusions
The method established in this paper was specific, accurate, and sensitive for simultaneous quantification of eight compounds in Dipsaci Radix. The method, using HPLC-DAD analysis in combination with PCA and DA, could provide a more comprehensive and quantitative chemical pattern recognition and quality evaluation to Dipsaci Radix samples. In the meantime, it provided a scientific basis for clarifying the mechanism of Dipsaci Radix sweating in the production area.   "À"decreased; "þ" increased.