Quantitative Determination of 7 Saikosaponins in Xiaochaihu Granules Using High-Performance Liquid Chromatography with Charged Aerosol Detection

Bupleuri Radix (Chaihu, in Chinese) is the principal drug in Xiaochaihu granules (XGs) that is a famous Chinese medicine preparation in China. Since previous analytical methods have not focused on the multiactive saikosaponins of Chaihu, it is difficult to effectively control the quality of XG on the market. In this manuscript, the simultaneous determination of 7 saikosaponins (saikosaponins C, I, H, A, B2, G, and B1) in XG by HPLC with charged aerosol detection (CAD) and confirmation by LC-Q-Orbitrap HRMS were described. The saikosaponins were purified on an SPE cartridge and determined on a Waters CORTECTS C18 column (4.6 mm × 150 mm, 2.7 μm) by gradient elution using 0.01% acetic acid aqueous solution and acetonitrile. The results showed good linearity with the r2 values higher than 0.998 for all analytes. The average recoveries at three different concentration levels ranged from 80% to 109% and the intraday and interday precision (relative standard deviations, RSD%) were in the range of 1.0%∼1.9% and 1.4%∼2.1%, respectively. The established HPLC-CAD method was subsequently applied to 15 batches of XG to investigate the batch-to-batch consistency and controllability. The proposed method could potentially be used for the quality control of XG and also be helpful in the quality evaluation of Chaihu and its related preparations.


Introduction
Xiaochaihu granules (XGs) are derived from a traditional Chinese medicine (TCM) classical formula Xiaochaihu-tang (named Sho-saiko-to in Japanese). It is a modern Chinese medicine preparation composed of 7 herbs including Bupleuri Radix (Chaihu, in Chinese), Scutellaria radix (Huangqin), Codonopsis radix (Dangshen), Glycyrrhiza radix (Ganchao), Pinellia tuber (Banxia), Jujube fruit (Dazhao), and Ginger rhizome (Shengjiang). It has the function of inducing sweat to dispel heat, soothing liver, and harmonizing stomach [1]. Nowadays, XG produced by dozens of manufactures has been sold as an over-thecounter medicine and widely used for the treatment of influenza, bronchitis, pneumonitis, hepatic fibrosis, and gastrointestinal diseases in clinical practice [2,3]. Due to its good efficacy, few side effects, and convenient administration, XG has been one of the best-selling anticold medicines and used to treat roughly millions of patients annually in China.
Chaihu is the principal drug in XG that plays a major therapeutic role in the treatment of the main disease and syndrome. Literature studies reported that saikosaponins have many pharmacology activities, such as antiviral, antibacterial, antihepatitis, and immunomodulating effects [4][5][6], and the major active constituents are saikosaponin A (SSa), saikosaponin C (SSc), and saikosaponin D (SSd) in Chaihu [7,8]. Since the unstable allyl oxide bonds could be broken under mildly acidic condition or by heating, they might be partially or completely converted into saikosaponin B 1 (SSa ⟶ SSb 1 ), saikosaponin I or H (SSc ⟶ SSi or SSh), and saikosaponin B 2 (SSd ⟶ SSb 2 ) accordingly. Because of the lacking of UV chromophores, HPLC-UV [9] is imperfect for saikosaponins analysis. e evaporative light scattering detector (ELSD) [10] and charged aerosol detector (CAD) [10] are usually recommended to analyze those compounds with weak or no UV absorption. Eom [10] compared CAD and ELSD methods in the simultaneous analysis of 10 saikosaponins in Chaihu. However, both methods were shown to be not suitable for saikosaponins analysis in XG due to the low content and multicomponent mutual interference. In addition, mass spectrometry (MS) technology [8,11] has been mainly focused on the in vivo metabolism study. Bao et al. [8] established an LC-MS/MS method to determine several saikosaponin derivatives present in Chaihu and its phytopharmaceuticals of Xiaochaihu-tang, which could be applied to the DMPK evaluation and study. LC-MS technique has some limitations for large-scale routine analysis of herbal medicine as a result of its expensive price, high maintenance cost, and low penetration. Furthermore, most methods aimed at Chaihu were very difficult to be transferred for XG analysis because of the structural transformation between active components [7,8]. e sample processing procedure is also a key factor in the analysis. Several novel extraction methods, such as microwave assisted extraction [12], ultrasound assisted extraction [12], ultrasoundand slat-assisted liquid-liquid extraction [13,14], matrix solidphase dispersion [15,16], and vortex-assisted liquid-liquid microextraction [17], have been developed for natural product analysis. Considering the practicability and operability of the quantitative method used for QC laboratories, routine methods including liquid-liquid extraction and solid-phase extraction were compared and evaluated in this paper. e current Chinese Pharmacopeia (ChP) method [1] ignores the assay of saikosaponins and the Japanese Pharmacopoeia (JP) method [18] only focuses on single ingredient SSb 2 . It is widely acknowledged that the synergistic effects of multiple-active components are responsible for the therapeutic effects of TCM preparation. erefore, the reported methods and current standards are insufficient to effectively ensure the quality of XG.
In this work, 7 saikosaponins (saikosaponins C, I, H, A, B 2 , G, and B 1 ) present in XG were unambiguously identified by comparing with reference standards using LC-Q-Orbitrap HRMS. All analytes were purified and concentrated on a C18 SPE cartridge. e established HPLC-CAD method was proved to be specific, reliable, and practical for the quantitative determination of 7 saikosaponins and successfully applied to the quality evaluation of 15 batched of XG thereof.

Sample Preparation
, and SSb 1 (0.5 mg·mL −1 ) was prepared with methanol. A series of mixed standard solutions were obtained by appropriate dilution of the stock solution for plotting calibration curve. All solutions were stored in the refrigerator at 4°C until analysis.

Sample Solution
Preparation. An aliquot of 3 g XG dissolved in 6 mL water was loaded onto a C18 SPE cartridge that was preconditioned with 6 mL methanol and 10 mL water sequentially.
e cartridge was washed with 10 mL 10% methanol aqueous solution containing 5% concentrated ammonia and followed by 30% aqueous methanol solution. After that, the cartridge was eluted using 10 mL methanol to desorb the analytes. e eluate was evaporated to dryness using a vacuum concentrator. e residue was dissolved with methanol and then transferred into a 2 mL volumetric flask which was brought up to its volume with methanol. All solutions were filtered through a 0.22-μm filter membrane before LC analysis.

HPLC-CAD Chromatographic Condition.
A Dionex Ultimate 3000 LC system ( ermo Fisher Scientific, USA) consisting of vacuum degasser, quaternary pump, autosampler, thermostatted column compartment, and CAD, connected with Chromeleon software, was employed for quantitative analysis. e column, mobile phase, and column temperature were the same as that of LC-MS condition.

Validation of the Method.
e quantitative HPLC-CAD method was fully validated in terms of specificity, linearity, limit of detection (LOD), limit of quantitation (LOQ), intra-and interday precision, repeatability, and accuracy in accordance with ICH guideline Q2(R1) [19]. Specificity was assessed by comparing the chromatograms obtained from blank solution, mix standard solution, test sample, and test sample without Chaihu. A series of mix standard solutions at six different concentration levels were prepared and plotted the calibration curve, where y and x represent the logarithm of peak area and logarithm of concentration (μg·mL −1 ). LOD and LOQ were determined by diluting the reference standard solution with methanol when the signal-to-noise rations (S/N) were about 3 : 1 and 10 : 1, respectively. For intra-and interday precision, the mix standard solution was determined in the same day (n � 6) and in three consecutive days, respectively. Six replicates as per the test method representing a single batch were prepared and analyzed to demonstrate the repeatability. e precision and repeatability were evaluated by calculating the relative standard deviations (RSDs) of each peak area. e recovery was used to express the accuracy of the method, which was performed by spiking known amounts of the sample to three concentration levels (50%, 100%, and 150%) of the mix standard solution.

Results and Discussion
3.1. Method Optimization. ELSD and CAD are both universal detectors and have obvious advantages in detecting those compounds with weak or no UV absorption. e performance comparison between these two detectors was studied by some researchers [10,20]. It was demonstrated that CAD was superior in linearity, sensitivity, reproducibility, and peak sharpness. As a result, CAD is used as a valuable alternative to ELSD for pharmaceutical analysis [21][22][23].
XG is a mixture of multiple herbs and thus has a wide range of complex and diverse chemical components [24][25][26][27]. e determination of saikosaponins is difficult because of the low content and multicomponent mutual interference in XG. e sample processing procedure that is able to purify and concentrate the analytes could solve the problem. e liquid-liquid extraction (LLE) using water-saturated n-butanol solution and C18 SPE cartridge were investigated comparatively. It was shown that LLE was unable to remove the major interfering component baicalin present in Huangqin and other impurities thoroughly, while SPE cartridge had obvious superiority in terms of extraction speed, precision, and reliability and was therefore employed for sample preparation. ree brands of SPE  Journal of Analytical Methods in Chemistry cartridges (Waters, Sepax technologies, and Welchrom) were studied, and it was found that Waters cartridge had high cleanup efficiency and good recovery compared with the others. Saikosaponins were fully absorbed on the C18 cartridge when using low proportional methanol in an alkaline solution as the eluent that could get rid of most interfering components. Nevertheless, they were easily desorbed by eluting with a high proportional methanol solution. e HPLC-CAD chromatographic condition described in the reference literature [10] was attempted initially, but found to be unsuitable for saikosaponins analysis in XG in the aspect of drifting retention times, poor peak shape, low resolution between analytes, and their adjacent peaks. Several gradient programs by using different mobile phase additives (formic acid, acetic acid, and ammonium acetate) were investigated to get a better separation and higher response.  Table 1. e peak 3 showed in Figure 2 was tentatively identified as rotundioside D according to the literature [26]. SSd (peak 9) was not detected in the test sample, probably because it had been entirely transformed into SSb 2 (peak 6) during the production of XG. is result was in accordance with literature reports [8]. Figure 3, no interfering peaks were observed at the retention time of target analytes in the test sample, which indicated the method is specific. e linear regression, correlation coefficient (r 2 ), linearity range, LOD, LOQ, intra-and interday precision, repeatability, and recovery of the 7 analytes are listed in Tables 2 and 3. e results showed good linear relationships and high precision. e variations (RSD%) of repeatability were NMT 6.0% for SSc, SSa, SSb 2 , SSg, and SSb 1 and NMT 8.0% for SSi and SSh. e average recoveries at three different concentration levels were between 80% and 109%.

Method Validation. As shown in
ese obtained values illustrated great repeatability and accuracy.

Sample Analysis.
e validated HPLC-CAD quantitative method was subsequently applied to determine 7 saikosaponins (SSc, SSi, SSh, SSa, SSb 2 , SSg, and SSb 1 ) in 15 batches of XG. e results (mg/package, x ± s) are present in Figure 4. e content of saikosaponins in XG was in the order of SSb 2 > SSb 1 > SSa > SSh ≥ SSg > SSi > SSc. e high standard deviation values of SSa, SSb 2 , and SSb 1 among 15 batches probably resulted from the quality difference of Chaihu. Given the structural transformation during processing, the total content of the target analytes in each batch was calculated for quality evaluation. e average content was 1.37 mg per package with a relative standard deviation (RSD %) of 4.4%, which demonstrated acceptable batch-to-batch consistency and controllability. Bao et al. [8] reported that the major saikosaponins contained in the test Xiaochaihu-tang and Sho-saiko-to samples were SSb 2 (26.9%), SSa (25.8%), SSb 1 (22.4%), SSg (14.3%), SSc (6.9%), SSh (5.4%), and SSi (2.9%). e difference between our results and literature data

Conclusions
In summary, 7 saikosaponins (SSc, SSi, SSh, SSa, SSb 2 , SSg, and SSb 1 ) present in XG were unambiguously identified using LC-Q-Orbitrap HRMS and thus selected as quality control components. A specific, sensitive, and accurate HPLC-CAD method in determining the analytes was described in this manuscript. A C 18 SPE cartridge was used to purify and concentrate the analytes by eluting with a 10% methanol aqueous solution containing 5% concentrated ammonia and a 30% methanol aqueous solution, and then by methanol. e peaks of 7 saikosaponins were well-separated, although most of the analytes are isomeric compounds. is quantitative method could potentially meet the requirement for quality analysis for QC laboratories. Also, the proposed method could to be used for quality evaluation of Chaihu and its related preparations.

HPLC-CAD:
High-performance liquid chromatography with charged aerosol detection LC-Q-Orbitrap HRMS: Liquid chromatography with hybrid quadrupole-orbitrap mass spectrometer XG: Xiaochaihu granule TCM: Traditional Chinese medicine NMT: Not more than x ± s: Mean ± standard deviation.

Data Availability
e data used to support this study were obtained from GenChim Testing (Shanghai) Co., Ltd, Shanghai, China, and are available from the corresponding author upon request.

Conflicts of Interest
e authors declare no conflicts of interest.