Developing an Absorption–Based Quality Control Method for Hu–Gan–Kang–Yuan Capsules by UFLC–QTOF–MS/MS Screening and HPLC–DAD Quantitative Determination

Traditional Chinese Medicine Preparations (TCMPs) contain massive numbers of ingredients responsible for their multiple efficacies. An absorption–based quality control method for complicated TCMPs using Hu–gan–kang–yuan Capsule (HGKYC) as an example was developed. To select proper chemical markers for quality control of HGKYC, an ultra–fast liquid chromatography (UFLC) coupled with electrospray ionization quadrupole time–off light mass spectrometry (UFLC–QTOF–MS/MS) method was used for the rapid separation and structural identification of the constituents in the HGKYC extract and the rat serum after oral administration of HGKYC. As a result, one hundred and seven prototype constituents including flavonoids, organic acid, phenylpropanoids, anthraquinones, saponins, alkaloids, terpenes, phenols and amino acids in HGKYC extract, and 43 compounds found in rat serum after oral administration of HGKYC were unambiguously identified or tentatively characterized by comparing retention times and MS information with those of authentic standards or available literature references. Finally, a simple, low–cost and effective method of simultaneous determination for baicalein, wogonin, paeonol and emodin in HGKYC was developed using high performance liquid chromatography coupled with a diode array detector. In conclusion, an absorption–based quality control pattern was developed and successfully used for evaluating HGKYC.


Introduction
Traditional Chinese Medicine Preparations (TCMPs) are usually composed of multiple Chinese medicinal materials that contain massive numbers of ingredients responsible for their multiple efficacies. In the quality control (QC) of TCMPs, the biggest challenge is the choice of chemical markers due to the complicated nature of the compounds in TCMPs. So far, selecting one compound as the quantitative marker like in Western medicine remains the main QC method used for TCMPs. We investigated and found that among 1493 TCMPs in the Chinese Pharmacopoeia (I volume, 2015 version), 849 TCMPs were evaluated by determining only one marker compound [1]. It is no denying that the Western medicine QC platform is somewhat helpful in evaluating the repeatability of TCMP ingredients among different batches, however, this QC platform may not correlate to the TCMPs' efficacy, because the compounds selected as quantitative markers may not be active or may not be absorbed. Hence, it is inevitable that sometimes the relationship between the "quality" of a TCMP evaluated by some chemical marker and its pharmacological effects can be inconsistent [2].
were purchased from Guangzhou Chemical Reagent factory (Guangzhou, China). Ultrapure water was provided by Southern Medical University (Guangzhou, China).
Three male Sprague-Dawley rats weighing 260˘20 g were obtained from the Animal Center of Southern Medical University. Animals were bred in a breeding room with a temperature of 23˘2˝C humidity of 60%˘5%, and 12 h dark-light cycle. They were given tap water and fed a normal diet and were acclimatized to the facilities for 3 days. The rats were fasted for 12 h before experimentation, while water was taken freely. The animal experiments were carried out in accordance with the Guide for the Care and Use of Laboratory Animals of Southern Medical University.

UFLC-QTOF 5600 + MS/MS
UFLC analysis was performed on a Shimadzu UFLC XR instrument (Shimadzu Corp., Kyoto, Japan), consisting of a binary pump, an auto-sampler, a column oven and a diode-array detector. Samples were separated on a Kinetex C 18 column (100 mmˆ2.1 mm I.D., 2.6 µm, Phenomenex, CA, USA). The mobile phase consisted of acetonitrile (A) and 0.1% aqueous formic acid (v/v) (B). The following gradient elution program was used-linear gradient from 2% to 100% A at 0-30 min), isocratic 100% A at 30-40 min), 100%-2% A at 40-41 min), isocratic 2% A at 41-45 min. The flow rate was kept at 0.3 mL/min. The injected volume was 2 µL and the column temperature was set at 40˝C. The DAD detector scanned from 190 nm to 400 nm. Mass spectrometry was performed on the Triple TOF TM 5600 plus (AB SCIEX, Foster City, CA, USA) a hybrid triple quadrupole time-of-flight mass spectrometer equipped with ESI source. The system was operated with Analyst ® TF 1.6 software (AB SCIEX). The conditions of the MS/MS detector were as follows: first ion source gas 60 psi, second ion source gas 60 psi, curtain gas 35 psi, temperature 550˝C, ion spray voltage floating 5500 V, collision energy 35 V, collision energy spread 15 V, declustering potential 80 V. TOF-MS range was set at m/z 100-1000 and product ions mass range was set at m/z 50-1000. Both positive and negative ion modes were used for compound ionization. Nitrogen was used as nebulizer and auxiliary gas.

Animal Administration
Three SD rats were orally administrated HGKYC powder 10 g/kg. Blood was collected from caudal vein at 4 h after administration and centrifuged at 6000 g for 10 min at 4˝C. Serum was frozen at´20˝C until analysis.

Extract of HGKYC
The contents of HGKYC (0.20 g) were weighed precisely and treated by sonication with 10 mL of methanol for 15 min at 50˝C in an ultrasonic bath (JY92-II, frequency 40 kHz, 250 W, Ningbo Biotechnology Co., Ltd., Ningbo, China). The extract solution was centrifuged at 13,000 rpm for 15 min. Then the supernatant was injected into UFLC and HPLC.

Rat Serum Sample
For identification of multiple constituents in rat, the serum was vortex mixed with 2.5 times amount of acetonitrile (v/v) to precipitate proteins and centrifuged at 13,000 g for 15 min. The supernatant was injected into the UFLC-Q-TOF MS/MS system.
The working solutions were prepared by serial dilution of the stock solutions with methanol. A mixture of all forty-nine reference standards was prepared in methanol and was filtered through a 0.22 µm syringe filter before UFLC-QTOF-MS/MS analysis.

Preparation of Standard Solutions for Quantitative Determination
Standard stock solutions of bacalein, wogonin, paeonol and emodin were precisely weighed and prepared by dissolving them in methanol, respectively. A mixed standard solution was prepared by accurately transferring each-standard stock solution and diluting with methanol. The final concentration of baicalein, wogonin, paeonol and emodin in the mixed standard working solution was 40.25, 46.75, 32.12 and 26.70 µg/mL, respectively. The quality control samples (LQC, MQC and HQC) were 5, 50, and 70 µg/mL for bacalein, 5, 50 and 80 µg/mL for wogonin, 4, 20 and 40 µg/mL for paeonol and emodin, respectively. All solutions were stored at 4˝C until used.

Extract of Negative Control Sample Solution
The negative control samples of HGKYC were prepared by getting rid of one corresponding material medicine from this preparation. The herbs were ground into powders before use, and prepared using the sample preparation protocol (showed in Table 1).

Establishment of Peak Identification
The UFLC-Q-TOF-MS/MS data of samples were analyzed by PeakView software version 1.2 (AB SCIEX), mainly with the XIC manager tool which provided the quasi-molecular weight, mass errors and isotope pattern fit. When a reference standard was available, the compound was identified by comparing the retention time and MS/MS spectra. While the identification of compound without available reference standard was mainly based on MS/MS data and available literature reports. The mass error of molecular ions of all identified compounds was less than˘5 ppm.

Method Validation for Determination of Baicalein, Wogonin, Paeonol and Emodin in HGKYC
A HPLC-DAD method was developed for simultaneous determination of baicalein, wogonin, paeonol and emodin for HGKYC quality evaluation. Specificities were assessed by comparing chromatograms of a standard solution mixture of baicalein, paeonol, wogonin and emodin, extract solution of HGKYC and negative control samples under the same conditions. Linearities were assessed by assaying calibration curves with peak area vs. five different injection amounts of 0.040-0.805 µg, 0.032-0.482 µg, 0.047-0.935 µg and 0.027-0.534 µg for baicalein, paeonol, wogonin and emodin, respectively. The intra-day and inter-day precisions (%, RSD) were established by analyzing QC samples on day 1 and on each of three consecutive days in five replicates. The accuracies were evaluated by means of recovery assays performed by adding known amounts of baicalein, paeonol, wogonin and emodin standards to the sample at the similar concentration in six replicates. The spiked amounts of baicalein, paeonol, wogonin and emodin were 77.10, 96.60, 26.70 and 28.05 µg, respectively. The original amount of baicalein, paeonol, wogonin and emodin in sample solution was 85.23, 104.12, 28.11 and 30.34 µg, respectively. The recoveries were calculated as follows: Recovery p%q " rpdetermined amount´original amountq{amount spikedsˆ100 (1) The stabilities were evaluated by RSD (%) of peak area of baicalein, paeonol, wogonin and emodin by analyzing sample solution in six replicates at 4.0˝C within 1 month.

Optimization of UFLC-QTOF-MS/MS Conditions
The ingredients of HGKYC belonged to different chemical families and showed distinct polarities. Thus, to obtain an effective separation, guarantee high ionization, and minimize ion suppression, the mobile phase system consisting of a mixture of 0.1% formic acid water solution and acetonitrile with gradient elution was used. The MS conditions, such as the parameters of gas pressure, ion spray voltage, capillary temperature and voltage of declustering potential were optimized. The optimized conditions were described in Section 2.2.1. The positive and negative ion modes were detected.

Profiles of Ingredients from HGKYC Extract
The total ion chromatograms (TIC) of HGKYC in positive and negative ion modes were obtained. Figure 1A showed the representative positive signal of samples. To better understand the MS fragmentation patterns of the constituents, forty nine of reference standards, including flavonoids, isoflavonoids, saponins, anthraquinones, organic acid and iridoids were investigated by UFLC-Triple-QTOF-MS/MS firstly.
Under the present conditions, a total of 107 compounds from HGKYC were identified, forty nine of which were unambiguously identified by comparing their retention times, accurate masses and fragment ions with those of the available reference standards, while the others were tentatively identified by elucidating the quasi-molecular ions, fragment ions as well as the available literature reports. The data information of compounds are summarized in Table 2, which includes retention time, molecular formula, measured mass, fragment ions, compound name and related literatures.       [15] "s" Identified with reference compounds; "+": detected and "´"not detected.

Identification of Flavonoids and Isoflavonoids
There are notable amounts of flavonoids and flavonoid derivatives in Scutellariae radix and Epimedii folium. Especially there are amounts of prenylated flavones in Epimedii folium, which have characteristic prenylated group at C-8 position. In MS analysis, the consecutive losses of sugar, H2O, carbonyl and isopentene group (C4H8) were observed as typical fragmentation profiles of these prenylated flavones. A total of 45 flavones or isoflavones were identified in HGKYC, with 18 definitely elucidated and the others tentatively identified.
The loss of an isopentene group was observed as a typical fragmentation profile of prenylated flavones except the consecutive neutral losses of sugar moieties.

Identification of Organic Acids
Because of the presence of COOH groups in organic acid structures, their fragment ions were usually generated by the losses of H 2 O (18 Da), CO 2 (44 Da) and HCOOH (44 Da). Seventeen organic acids have been identified from HGKYC, nine of which were confirmed by reference standards, peaks 6, 10, 11, 22, 24, 31, 32, 35 and 106 were identified as nicotinic acid, p-coumaric acid, gallic acid, chlorogenic acid, caffeic acid, syringic acid, trans-p-coumaric acid, ferulic acid and oleanic acid, respectively. The other organic acids (peaks 13, 14, 20, 29, 45, 97, 105, 107) were tentatively assigned. Peak 14 exhibited an [M´H]´ion at m/z 153.0199 and yielded high intensity fragments ions at m/z 109.0309 and 91.0208 by the successive losses of CO 2 and H 2 O, by comparing with the literature report, peak 14 was identified as protocatechuic acid [13]. Peak 20 gave an [M´H]´ion at m/z 341.0881 (C 15 H 18 O 9 ) and yielded fragment ions at m/z 179.0352 and 135.0454 by the successive losses of one glucose residue and CO 2 , so was inferred that peak 20 was caffeoylglucose by comparison with the available literature [13]. Peak 97 gave an [M + H] + ion at m/z 489.3566 (C 30 H 48 O 5 ) in positive ion mode and yielded major fragment ions at m/z 453.3324 and 407.3254 by the successive losses of 2H 2 O and HCOOH, respectively. By comparison with the available literature, peak 97 was tentatively identified as tormentic acid [15]. Peak 107 had the same [M´H]´ion at m/z 455.3537 (C 30 H 48 O 3 ) with oleanic acid in negative ion mode. Oleanic acid and ursolic acid were reported in Ligustri lucidi fructus, therefore, peak 107 was tentatively characterized as ursolic acid [15] 3.  [31], peak 104 was identified as schizandrin C.

Identification of Anthraquinones
Three anthraquinones have been identified from HKGYC, one of which was identified by comparing with the authentic standard, namely peak 99 identified as emodin [18]. Peaks 27 and 57 were tentatively assigned as mangiferin and polygonimitin B based on the available literature reports [18].

Identification of Alkaloids
Four alkaloids were identified from HGKYC, of which peaks 2, 4 and 15 were definitely identified as choline, betaine and epigoitrin by comparison with their authentic standards [18,19]. Peak  group, because of the unavailability of authentic standards, peak 3 and peak 28 were tentatively identified as proline and riboflavin, respectively, based on the available literature [32,33].

Profiles of Ingredients in Rat Serum after Oral Administration HGKYC
By comparing retention times, UV spectra and MS/MS data with those of compounds analyzed in HGKYC extract sample, forty three prototype compounds were found in rat serum (shown in Table 2 and Figure 1B). As shown in Table 2, the compounds absorbed in blood were far fewer than those in HGKYC extract, which indicated the number of compounds absorbed in blood is limited due to poor absorption or low concentration. On the basis of the availabilities of authentic standards and pharmacological effects reported, baicalein, wogonin, paeonol and emodin were then chosen as the chemical markers for quality control of HGKYC.

Optimization of HPLC Conditions
In this study, different mobile phases, elution modes and detection wavelengths were investigated. Mobile phases of acetonitrile-water and methanol-water with different modifiers including acetic acid, formic acid and phosphoric acid were tested under different gradient elution modes. The detection wavelength was selected according to the maximum absorption wavelengths of baicalein, wogonin, paeonol and emodin shown in UV spectra from DAD. The optimized conditions were described in Section 2.2.2. Finally, excellent separations were achieved and typical chromatograms are shown in Figure 2.

Optimization of Extraction Conditions
Ultrasound-assisted extraction (UAE) has been widely used in sample preparation for the quality control of TCMPs. In the extraction process, extraction solvent, sample-solvent ratio, extraction time and temperature are critical for high extraction effectiveness. Pure and aqueous methanol or ethanol solutions are often used as extraction solvents. In the present study, based on the physicochemical properties of the targeted compounds, different concentrations (50%, 80%, and 100%) of methanol water solutions were examined to extract the four compounds in HGKYC. Considering extract rates of targeted compounds, contents of interfering components and extract time, samples were extracted for 15 min by UAE using 50-time of methanol, 50˝C as extraction temperature.

Sample Determination
The contents of baicalein, paeonol, wogonin and emodin in HGKYC were determined using the validated methods in triplicates. The results are shown in Table 3.

Sample Determination
The contents of baicalein, paeonol, wogonin and emodin in HGKYC were determined using the validated methods in triplicates. The results are shown in Table 3. Table 3. Contents of baicalein, paeonol, wogonin and emodin in HGKYC (µg/g, n = 3, mean˘SD).

Discussion
UFLC-QTOF-MS/MS, a high-efficiency analytical technique, can provide qualitative and quantitative information about samples, but the expensive instrument price and analysis cost limits its applicability in quality control of TCMPs for pharmaceutical companies. HPLC-DAD is a high quality and inexpensive analysis technique, commonly used to analyze quantitatively chemical markers in TCMPs. At present, HPLC-DAD is irreplaceable in QC of TCMPs though it is not as powerful as UFLC-QTOF-MS/MS in functional analysis. Ideally, the more chemical markers selected to evaluate the quality of TCMPs, the better to ensure the consistency between the quality of TCMPs and their pharmacological activities, but that is unpractical because of the limited availability of authentic standards. In this paper, on the basis of the prototype compounds detected in rat serum after oral administration HGKYC, considering the following reasons: (1) the analysis cost; (2) the popularity of analysis equipment; (3) the availabilities of authentic standards; and (4) the reported anti-inflammatory activities and protective effects on hepatocytes of baicalein and wogonin in Scutellariae radix [34][35][36][37], paeonol in Moutan Cortex [38,39] and emodin in Polygoni Cuspidate Rhizome et Radix [40,41], then baicalein, wogonin, emodin and paeonol were selected as chemical markers for evaluating the quality of HGKYC by HPLC-DAD. Hence, the quality control scheme for HGKYC using the absorption-based chemical markers is warranted.

Conclusions
In summary, a reliable and powerful twenty-five minute long analytical method by UFLC-Q-TOF-MS/MS was successfully established for identifying compound profiles in HGKYC extract and in rat serum after oral administration of HGKYC. A total of 107 compounds in HGKYC, 43 compounds of which were also present in rat serum, were identified or tentatively identified based on their retention times, UV spectra, and MS information. The absorption-based quality control scheme for HGKYC provides a valuable demonstration for the quality control of TCMPs.