Journal Code Article ID Dispatch: 22.09.16 CE: B M C 3 8 2 8 No. of Pages: 8 ME: 1 58 Received: 28 June 2016 Revised: 18 August 2016 Accepted: 24 August 2016 2 59 DOI 10.1002/bmc.3828 3 60 4 61 5 RESEARCH ARTICLE 62 6 63 7 64 8 A UPLC–MS/MS approach for simultaneous determination of 65 9 66 10 eight in rat plasma, and its application to 67 11 ‐ ‐ ‐ 68 12 pharmacokinetic studies of Fu Zhu Jiang Tang tablet in rats 69 13 70 † † 14Q1 Yi Tao1,2* | Xi Chen1,2 | Yanhui Jiang1 | Baochang Cai1,2 71 15 72 16 73 1 School of Pharmacy, Nanjing University of 17 74 Chinese Medicine, Nanjing, China Abstract

18 2 – 75 Jiangsu Key Laboratory of Chinese Medicine The purpose of this study is to establish and validate a UPLC MS/MS approach to determine 19 Processing, Nanjing University of Chinese eight flavonoids in biological samples and apply the method to pharmacokinetic study of 76 Medicine, Nanjing, China 20 Fu‐Zhu‐Jiang‐Tang tablet. A Waters BEH C18 UPLC column was employed with methanol/ 77 21 Correspondence 0.1% formic acid–water as mobile phases. The mass analysis was carried out in a triple quadru- 78 Yi Tao, School of Pharmacy, Nanjing University 22 pole mass spectrometer using multiple reaction monitoring with negative scan mode. A one‐step 79 of Chinese Medicine, Nanjing 210023, China. 23 Email: [email protected] protein precipitation by methanol was used to extract the analytes from blood. Eight major 80 24 flavonoids were selected as markers. Our results showed that calibration curves for 81 25 3′‐hydroxypuerarin, mirificin, , 3′‐methoxypuerarin, , rutin, astragalin and 82 26 displayed good linear regression (r2 > 0.9986). The intra‐day and inter‐day precisions 83 27 (RSD) of the eight flavonoids at high, medium and low levels were <8.03% and the bias of the 84 28 accuracies ranged from −5.20 to 6.75%.The extraction recoveries of the eight flavonoids were 85 29 from 91.4 to 100.5% and the matrix effects ranged from 89.8 to 103.8%. The validated approach 86 30 was successfully applied to a pharmacokinetic study in Sprague–Dawley rats after oral adminis- 87 31 tration of FZJT tablet. Double peaks were emerged in curves of mean plasma concentration for 88 32 3′‐methoxypuerarin, which was reported for the first time. 89 33 90 34 KEYWORDS 91

35 Fu‐Zhu‐Jiang‐Tang tablet, pharmacokinetics, UPLC–MS/MS 92 36 93 37 94 38 95 39 1 | INTRODUCTION and 2‐caffeoylquinic acid isomers, which exhibit anti‐diabetic 96 40 effects through inhibition of α‐glucosidase (Zhang, Han, He, & Duan, 97 41 Fu‐Zhu‐Jiang‐Tang (FZJT) tablet, which consists of six well‐known 2008). , such as puerarin and daidzin, are the major constit- 98 42 traditional Chinese medicines – Pueraria lobata (PL), Morus alba (MA), uents of PL, and exert hypoglycemic and hypolipemic roles by elevating 99 43 Panax notoginseng (PN), Astragalus membranaceus (AM), Momordica insulin expression and maintaining metabolic homoeostasis (Wu et al., 100 44 charantia (MC) and Cortex lycii (CL) – has been proven to show benefi- 2013a). Ginsenosides, which belong to PN, can improve lipid profiles, 101 45 cial effects on patients with type II diabetes. A number of articles have inhibit peroxidation and increase the activity of antioxidant enzymes 102 46 been published on the traditional use of the six individual herbs for the (Xia, Sun, Zhao, & Hypolipidemic, 2011). Astragalosides, which are 103 47 treatment of diabetes (Habicht, Ludwig, Yang, & Krawinkel, 2014; Li, the major constituents of AM, show protective effects on kidney of 104 48 Wang, Xue, Gu, & Lin, 2011; Ulbricht et al., 2015; Uzayisenga, Ayeka, diabetic rats (Motomura et al., 2009). The extract of MC displays insulin 105 49 & Wang, 2014; Zhang, Pugliese, Pugliese, & Passantino, 2015). For secretagog and insulinomimetic activity, which may be ascribed to ste- 106 50 instance, the nonanthocyanin phenolics of MA known to date are rutin, roidal compounds and polypeptides (Raman and Lau, 1996). Our previ- 107 51 ous study characterized and determined the major constituents of 108 52 FZJT tablet using UPLC‐Q‐TOF/MS and HPLC‐UV (Tao et al., 2016). 109 †These two authors contributed equally. 53 Several analytical approaches have been established for pharma- 110 Abbreviations used: AM, Astragalus membranaceus; CL, Cortex lycii; FZJT, Fu‐ cokinetic determination of the above‐mentioned anti‐diabetic 54 Zhu‐Jiang‐Tang; MA, Morus alba; MC, Momordica charantia; MRM, multiple 111 55 reaction monitoring; PL, Pueraria lobata; PN, Panax notoginseng. constituents in vivo. For example, He et al. developed a sensitive 112 56 113 57 Biomedical Chromatography 2016; 1–8 wileyonlinelibrary.com/journal/bmc Copyright © 2016 John Wiley & Sons, Ltd. 1 114 1 2 TAO ET AL. 58 2 59 3 LC–MS/MS method to simultaneously determine rutin and astragalin maintained in an air‐conditioned room under the constant room 60 4 in rat plasma after oral administration of total flavonoids from mulberry temperature (23 ± 2°C), relative humidity level (50 ± 15%) and lighting 61 5 leaves (He et al., 2013). Xu, Li, and Zhang, (2015) quantified puerarin (12 h light/dark cycle) with free access to food and water for 7 days 62 6 and daidzin in rat plasma using LC‐MS/MS and applied them to a phar- before the experiment. Before the herb extract administration, the rats 63 7 macokinetic study of Gegenqinlian decoction. Furthermore, Yan et al. were fasted overnight but had access to water ad libitum. The experi- 64 8 (2014) established an LC‐MS/MS method to simultaneously quantify mental protocols were reviewed and approved by the Animal Ethics 65 9 puerarin, daidzin and daidzein in female rat plasma following oral Committee of Nanjing University of Chinese Medicine. 66 10 administration of Ge‐Gen decoction. These studies just focus on one 67 11 or two herbals. There is no report on the method for the pharmacoki- 68 2.2 | UPLC‐MS/MS conditions 12 netic study of the six‐herb preparation – FZJT tablet. The pharmacoki- 69 13 netic behavior of FZJT tablet after oral administration remains unclear. Liquid chromatography was carried out on a Shimadzu ultra perfor- 70

14 Pharmacokinetics is a branch of pharmacology dedicated to deter- mance liquid chromatography (UPLC) unit with an Acquity BEH C18 71 15 mining the fate of substances administered externally to a living organ- column (2.1 × 50 mm, 1.7 μm particle size) and inline 0.2 μm stainless 72 16 ism, which is of great importance in drug development and can explain steel frit filter. The elution solvents were optimized using methanol– 73 17 a variety of events related to the efficacy of herbal medicines (Katz, water or acetonitrile–water (with or without 0.1% formic acid) as the 74 18 Murray, Bhathena, & Sahelijo, 2008). The composition of herbal prep- mobile phase system. Finally, a gradient program was employed with 75 19 arations is complex and different constituents may have interactions. the mobile phase of solvent A (0.1% formic acid in water) and solvent 76 20 The pharmacokinetics of multi‐herbal preparations may provide valu- B (methanol) as follows: 5–30% B (0–2 min), 30% B (2.0–2.5 min), 77 21 able information for their clinical usage. In this study, our aim was to 30–100% B (2.5–6.0 min), 100% B (6.0–7.0 min). A subsequent 78 22 develop, optimize and validate a simple, sensitive and specific UPLC‐ re‐equilibration time (1 min) should be performed before next injec- 79 23 MS/MS method for the simultaneous determination of the bioactive tion. The flow rate was 0.30 mL/min and the injection volume was 80 24 compounds in rat plasma after oral administration of the FZJT total 5 μL. The column temperature was maintained at 35°C. 81 25 fraction. An AB sciex QTRAP 5500 triple quadruple mass spectrometer was 82 26 Our previous study showed that steroidal compounds and poly- employed for mass spectrometric detection. The detection was 83 27 peptides were not detected using UPLC‐Q‐TOF/MS. Therefore, only operated in the multiple reaction monitoring (MRM) mode under unit 84 28 compounds which can be detected were used for pharmacokinetic mass resolution in the mass analyzers. The mass parameters for each 85 29 analysis. Meanwhile, these eight flavonoids showed anti‐diabetic compound were optimized, including cone voltage, collision energy 86 30 effects. For instance, rutin and astragalin exhibited anti‐diabetic and so on. The MRM analysis was conducted by monitoring the 87 31 effects through inhibition of α‐glucosidase (Tao, Zhang, Cheng, & precursor ion to product ion transitions of m/z 431.0 → 311.0 for 88 32 Wang, 2013). Puerarin, daidzin and daidzein have been found to signif- 3′‐hydroxypuerarin, 415.0 → 267.0 for puerarin, 547.0 → 295.0 for 89 33 icantly decrease the baseline fasting plasma glucose and total choles- mirificin, 445.1 → 325.0 for 3′‐methoxypuerarin, 415.2 → 251.9 for 90 34 terol and increase insulin levels in several in vivo models (Liu et al., daidzin, 609.2 → 299.9 for rutin, 447.1 → 283.9 for astragalin, 91 35 2015). 3′‐Hydroxypuerarin, 3′‐methoxypuerarin, mirificin and daidzein 253.1 → 224.0 for daidzein, 405.2 → 243.0 for 2,3,5,4′‐tetrahydroxy‐ 92 36 exhibited anti‐diabetic effects by reducing oxidative stress to which stilbene‐2‐O‐β‐D‐glucoside (IS), respectively. Analyst (version 1.5.2, 93 37 the pancreatic cells were exposed (Bebrevska et al., 2010). In addition, AB Sciex) and PeakView software (version 1.2, AB Sciex) was used for 94 38 standards for the eight flavonoids were available. Thus, the eight flavo- data acquisition and instrument control. 95 39 noids were used as the determining index. 96 40 97 2.3 | Stock solutions 41 2 | EXPERIMENTAL 98 42 Reference standards of 3′‐hydroxypuerarin, puerarin, mirificin, daidzin, 99 43 rutin, astragalin, 3′‐methoxypuerarin, daidzein and IS were separately 100 2.1 | Chemicals and reagents 44 weighed and dissolved in methanol to prepare stock solutions. A series 101 45 3′‐Hydroxypuerarin, puerarin, 3′‐methoxypuerarin, mirificin, daidzin of standard solutions were obtained by further dilution of the stock 102 46 and daidzein were obtained from Chengdu Purechem standard Co. solution with methanol. Calibration samples were prepared by adding 103 47 Astragalin and 2,3,5,4′‐tetrahydroxy‐stilbene‐2‐O‐β‐D‐glucoside (IS) the series standard solutions (10 μL) to blank rat plasma (190 μL) to 104 48 were obtained from Sichuan Weikeqi Biotechnology Co. Rutin was obtain concentrations of 1.2–156.5 ng/mL for 3′‐hydroxypuerarin, 105 49 purchased from Chengdu Must Bio‐technology Co. The purity of each 2.0–1251 ng/mL for mirificin, 1.1–188.4 ng/mL for puerarin, 106 50 standard substance was >98%. FZJT tablet (lot no. 150320) and FZJT 1.0–1250 ng/mL for 3′‐methoxypuerarin, 0.8–250 ng/mL for daidzin, 107 51 total fraction were obtained from By‐Health Co. Methanol and aceto- 1.8–250.8 ng/mL for rutin, 1.5–125.2 ng/mL for astragalin and 108 52 nitrile were obtained from Merck (Darmstadt, Germany). Ultra‐pure 1.3–500.4 ng/mL for daidzein. 109 53 water used in the experiment was generated by a Milli‐Q water purifi- Quality control (QC) samples were prepared independently at 110 54 cation system (Millipore Corporation, Billerica, MA, USA). All other concentrations of 4.5, 50 and 120 ng/mL for 3′‐hydroxypuerarin, 111 55 chemicals were of analytical grade. 10.5, 90 and 1050 ng/mL for mirificin, 3.5, 40 and 165 ng/mL for 112 56 Sprague–Dawley rats (280 ± 20 g) were obtained from Slaccas puerarin, 3.2, 45 and 150 ng/mL for 3′‐methoxypuerarin, 3, 35 and 113 57 Experiment Animal Company (Shanghai, China). The rats were 195 ng/mL for daidzin, 8.5, 75 and 195 ng/mL for rutin, 5.2, 65 and 114 1 TAO ET AL. 3 58 2 59 3 105 ng/mL for astragalin and 5.5, 55 and 350 ng/mL for daidzein using 2.5.4 | Matrix effect and extraction recovery 60 4 the same method. The extraction recoveries of the analytes at the three QC levels were 61 5 evaluated by comparing the peak area ratios of the analytes with the 62 6 IS in the post‐extraction spiked samples with those acquired from 63 2.4 | Sample preparation 7 the pre‐extraction spiked samples. The matrix effects were measured 64 8 Protein precipitation and liquid–liquid extraction methods were by comparing the peak areas of the analytes in the post‐extraction 65 9 attempted and compared during sample preparation. For protein spiked samples with those of the standard solutions. The extraction 66 10 precipitation, the effects of methanol and acetonitrile on extraction recovery and matrix effect of the analytes at the three QC levels were 67 11 rate were investigated. For liquid–liquid extraction, ethylacetate, repeated for six replicates. 68 12 chloroform and n‐hexane were used to extract the compounds from 69 13 plasma samples. 70 2.5.5 | Stability 14 As a result, 100 μL of collected plasma sample was transferred to a 71 Six QC samples of each concentration at low, medium and high levels 15 1.5 mL centrifuge tube. An aliquot of 300 μL of the IS working solution 72 were prepared in the blank plasma and stored at −20°C for the freeze– 16 (30 ng/mL in methanol) was added to the plasma sample. The tubes 73 thaw experiments. Three freeze–thaw cycles were carried out within 17 were vortexed for 3 min and then centrifuged at 13,000 g for 5 min. 74 36 h to evaluate the freeze–thaw stability. Short‐term stability was 18 The supernatant was injected into the UPLC‐MS/MS system for 75 conducted by storing the QC samples at room temperature for 4 h. 19 analysis. 76 Long‐term stability was assessed on the same QC samples which had 20 77 been stored at −20°C for 30 days. Autosampler stability was 21 78 2.5 | Method validation performed by setting pretreated QC samples in autosampler (the tem- 22 79 perature was 4°C) for 24 h before analysis. 23 2.5.1 | Selectivity and specificity 80 24 The selectivity of the method was evaluated by comparing the 81 25 chromatograms of blank plasma with the corresponding plasma 2.6 | Pharmacokinetic study of FZJT tablet 82 26 samples spiked with the eight flavonoids, as well as real samples 83 Six Sprague–Dawley rats weighing 280 ± 20 g were fasted for 12 h 27 collected from treated rats. 84 before the experiment, with the exception of free access to water. 28 Specificity toward the endogenous plasma matrix constituents 85 The experiment was approved by the Animal Ethics Committee of 29 was assessed by comparing the chromatograms of blank rat plasma 86 Nanjing University of Chinese Medicine. According to the manufac- 30 from six sources, blank plasma spiked with the analytes at the concen- 87 turer, 1 g FZJT tablet consisted of 0.5 g FZJT total fraction and 0.5 g 31 trations of the lower limit of quantification (LLOQ) and the plasma 88 excipients. TIC chromatograms in negative and positive mode of FZJT Q2 32 samples from the rats after oral administration of FZJT total fraction. 89 total fraction are shown in Figure (in the Supporting Information). The 33 90 routine dose for human is 6 g FZJT tablet/day. Thus, six rats received 34 91 2.5.2 | Linearity and sensitivity an intragastric administration of 2.7 g/kg FZJT total fraction (equiva- 35 92 Calibration curves were constructed by plotting the peak‐area ratio of lent to 8.30 mg/kg of 3′‐hydroxypuerarin, 2.69 mg/kg of mirificin, 36 93 the eight flavonoids to the IS vs the concentrations of the flavonoids. 66.90 mg/kg of puerarin, 53.35 mg/kg of 3′‐methoxypuerarin, 37 94 The regression equations were achieved using a weighted least‐ 9.40 mg/kg of daidzin, 7.35 mg/kg of rutin, 5.48 mg/kg of astragalin 38 95 squares linear regression (the weighting factor, 1/×2). The lower limit and 0.65 mg/kg of daidzein). The FZJT total fraction was suspended 39 96 of quantification (LLOQ) was determined based on at least 10 times in 0.5% carboxymethyl cellulose sodium (w/v). Blood samples (about 40 97 of signal‐to‐noise ratio at which the precision (expressed by relative 500 μL) were collected in heparinized tubes via the postorbital venous 41 98 standard deviation, RSD) and accuracy (calculated by relative error, plexus veins from each rat before the dose and at 0.083, 0.176, 0.333, 42 99 RE) were lower than ±20%. 0.67, 1.0, 2.0, 4.0, 6.0, 8.0, 12.0 and 24.0 h after administration, and 43 100 were immediately centrifuged. Plasma was transferred into clean tubes 44 101 and stored at −80°C until analysis. The pharmacokinetic parameters 45 2.5.3 | Precision and accuracy 102 were calculated by the noncompartmental analysis of plasma concen- 46 Three validation batches, each containing six replicates of QC samples 103 tration vs time data using the DAS 2.0 software package. 47 at low, medium and high concentration levels (4.5, 50 and 120 ng/mL 104 48 for 3′‐hydroxypuerarin, 10.5, 90 and 1050 ng/mL for mirificin, 3.5, 105 49 40 and 165 ng/mL for puerarin, 3.2, 45 and 150 ng/mL for 3 | RESULTS AND DISCUSSION 106 50 3′‐methoxypuerarin, 3, 35 and 195 ng/mL for daidzin, 8.5, 75 and 107 51 195 ng/mL for rutin, 5.2, 65 and 105 ng/mL for astragalin and 5.5, 55 108 3.1 | Optimization of mass spectrometry conditions 52 and 350 ng/mL for daidzein) were assayed to assess the precision 109 53 and accuracy of the method on three different days. The precision is The chemical structures of the eight flavonoids and IS are displayed in 110 54 expressed by relative standard deviation (RSD) between the replicate Figure 1. The mass responses of the eight flavonoids in both positive F1 111 55 measurements. Accuracy is defined as relative error (RE), which is and negative modes were investigated, with better response obtained 112 56 calculated using the formula RE = [(measured value − theoretical in the negative ionization mode (see Figure ). This phenomenon may be 113 57 value)/theoretical value] × 100. explained by the loss of a proton to yield [M − H]− being easier than 114 1 4 TAO ET AL. 58 2 59 3 60 4 61 5 62 6 63 7 64 8 65 9 66 10 67 11 68 12 69 13 70 14 71 15 72 16 73 17 74 18 75 19 76 20 FIGURE 1 Chemical structures of eight 77 compounds and 2,3,5,4′‐tetrahydroxy‐ 21 78 stilbene‐ 2‐O‐β‐D‐glucoside(IS) 22 79 23 adding an proton to produce [M + H]+ for the eight flavonoids under were investigated by adding 0.1% formic acid to the water phase. 80 24 soft ionization mode. Throughout the development of the method, The addition of 0.1% formic acid significantly increased the signal 81 25 the selectivity, reproducibility and robustness of the MS method were intensity of the eight compounds, especially for the methanol– 82 26 monitored and adjusted to establish methods that could be validated water system. Meanwhile, the addition of 0.1% formic acid was 83 27 and applied to a pharmacokinetic study. The following MRM modes beneficial to the peak shape of the compounds. Thus, the metha- 84 28 were found to be specific and intense for the analysis of nol–water (0.1% formic acid) system was selected as the elution 85 29 3′‐hydroxypuerarin (m/z 431.0–311.0), puerarin (m/z 415.0–267.0), solvent. 86 30 mirificin (m/z 547.0–295.0), 3′‐methoxypuerarin (m/z 431.0–311.0), 87 31 daidzin (m/z 445.1–325.0), rutin (m/z 609.2–299.9), astragalin (m/z 88 32 447.1–283.9), daidzein (m/z 253.1–224.0) and IS (m/z 405.2–243.0). 89 3.3 | Optimization of extraction procedure 33 T1 The details are presented in Table 1. 90 34 Protein precipitation and liquid–liquid extraction methods were 91 35 attempted and compared during sample preparation. As shown in 92 3.2 | Optimization of liquid chromatography 36 Figure , the result of the liquid–liquid extraction method was unsatis- 93 conditions 37 factory, which may be due to the compounds having a high chemical 94 38 First, gradient elution was carried out using methanol–water or polarity and not being able to be extracted by ethyl acetate, chloro- 95 39 acetonitrile–water as the mobile phase system and the results are form and n‐hexane. Thus, protein precipitation was chosen to prepare 96 40 shown in Figure . For compounds 1, 3, 4 and 5, the methanol– plasma on account of its high recovery. The extraction rates of vari- 97 41 water mobile phase system yielded better signal intensity than that ous organic solvents (methanol, acetonitrile) were compared. Finally, 98 42 of acetonitrile–water. For compounds 6–8, the acetonitrile–water methanol was chosen as a reliable solvent for protein precipitation 99 43 mobile phase system yielded better signal intensity than that of owing to its excellent precipitation, preferable recovery and small 100 44 methanol–water. Moreover, the signal intensities of the compounds matrix effect. 101 45 102 46 TABLE 1 Multiple reaction monitoring transitions and parameters for the detection of the analytes and IS 103 47 104 Analytes Precursor ion (m/z) Product ion (m/z) Cone voltage (V) Collision energy (eV) 48 105 3′‐Hydroxypuerarin 431.0 311.0 ‐150 ‐32 49 106 Puerarin 415.0 267.0 ‐175 ‐42 50 107 Mirificin 547.0 295.0 ‐195 ‐38 51 108 3′‐Methoxypuerarin 445.1 325.0 ‐175 ‐32 52 109 Daidzin 415.2 251.9 ‐200 ‐36 53 110 Rutin 609.2 299.9 ‐185 ‐48 54 111 Astragalin 447.1 283.9 ‐120 ‐38 55 112 Daidzein 253.1 224.0 ‐25 ‐34 56 113 IS 405.2 243.0 ‐90 ‐26 57 114 1 TAO ET AL. 5 58 2 59 3 3.4 | Validation of the method 3.4.4 | Extraction recovery and matrix effect 60 4 61 3.4.1 | Selectivity and specificity The mean extraction recoveries of eight analytes were determined 5 using six replicates of QC samples of each level. The results are shown 62 At the retention times of 3′‐hydroxypuerarin (2.18 min), mirificin 6 in Table S3. The mean extraction recoveries of the eight flavonoids 63 (2.38 min), puerarin (2.41 min), 3′‐methoxypuerarin(2.46 min), daidzin 7 were from 91.4 to 100.5% and the matrix effects ranged from 89.8 to 64 (2.67 min), rutin (2.86 min), astragalin (3.19 min), daidzein (3.75 min) 8 103.8%. As seen from Table S3, the values of matrix effect >100% were 65 and IS (2.98 min), no obvious endogenous interferences were observed 9 only detected at the lower QC concentration of 3′‐hydroxypuerarin, 66 from blank plasma. Representative plasma chromatographs, such as 10 puerarin and rutin. This phenomenon may be explained by the system- 67 blank plasma, plasma at the LLOQ level and plasma samples obtained 11 atic error of the UPLC‐MS/MS instrument or the sampling error. The 68 from intragastric administration of FZJT total fraction in rats are shown 12 results indicated that the plasma matrix peaks did not affect the 69 F2 in Figure 2. 13 quantification of the eight flavonoids in the experimental conditions. 70 14 71 15 3.4.5 | Stability 72 3.4.2 | Linearity and LLOQs 16 73 The stability of the flavonoids during the sample processing proce- 17 The calibration curves were constructed by plotting the ratio of peak 74 dures and storage was evaluated by analysis of three levels of QC 18 areas of the eight flavonoids to that of IS vs the nominal concentration 75 samples. As shown in Table S4, the results indicated that the eight fla- 19 of calibration standards and fitted to linear regression (y = ax + b). The 76 vonoids in rat plasma were all stable for 4 h at room temperature, for 20 concentrations of analytes in test samples were calculated by the 77 30 days' storage at −80°C, after three freeze–thaw cycles and at 4°C 21 regression parameters from the calibration curves. The lowest limit of 78 for 24 h, with accuracy in the range of −2.31–7.08, −7.81–9.36, 22 quantification (LLOQ) was defined as the lowest concentration of 79 −2.94–10.55 and −6.22–9.16%, respectively. The results suggest that ‐ ‐ 23 analyte in spiked plasma resulting in a signal to noise ratio of 10:1. 80 the eight flavonoids are stable under these conditions. 24 The results in Table (Supporting Information) show that all of the 81 2 25 calibration curves displayed good linear regression (r > 0.9986). The 82 3.5 | Pharmacokinetic analysis 26 LLOQs were determined as 1.2, 2.0, 1.1, 1.0, 0.8, 1.8, 1.5 and 1.3 ng/mL 83 ′‐ ′‐ 27 for 3 hydroxypuerarin, mirificin, puerarin, 3 methoxypuerarin, This developed approach was applied to determine the plasma concen- 84 28 daidzin, rutin, astragalin and daidzein, respectively. The acceptable trations of the eight flavonoids after intragastric administration of 85 29 precision variations were <20% with an accuracy error within ±20%. FZJT tablet. The mean plasma concentration–time profiles of 86 30 3′‐hydroxypuerarin, mirificin, puerarin, 3′‐methoxy‐ puerarin, daidzin, 87 31 rutin, astragalin and daidzein are illustrated in Figure 3. In addition, F3 88 3.4.3 | Precision and accuracy 32 the derived pharmacokinetic parameters are summarized in Table 2. T2 89 33 The intra‐day and inter‐day precisions and accuracy of the eight flavo- The assay was sensitive for the simultaneous determination of 90 34 noids were investigated by analyzing three levels of QC samples. All 3′‐hydroxypuerarin, mirificin, puerarin, 3′‐methoxypuerarin, daidzin, 91 35 the data are shown in Table S2. All the results of the tested samples rutin, astragalin and daidzein in rat plasma after intragastric administra- 92 36 were within the acceptance criterion of ±15%. tion of FZJT total fraction. 93 37 94 38 95 39 96 40 97 41 98 42 99 43 100 44 101 45 102 46 103 47 104 48 105 49 106 50 107 51 108 52 109 53 110 54 111 55 112 56 FIGURE 2 Representative multiple reaction monitoring chromatograms: (A) blank plasma; (B) blank plasma spiked with the analytes at LLOQs; (C) 113 57 blank plasma spiked with the analytes; and (D) plasma samples after oral administration of Fu‐Zhu‐Jiang‐Tang (FZJT) total fraction 114 1 6 TAO ET AL. 58 2 59 3 60 4 61 5 62 6 63 7 64 8 65 9 66 10 67 11 68 12 69 13 70 14 71 15 72 16 73 17 74 18 75 19 76 20 77 21 78 22 79 23 80 24 81 25 82 26 83 27 84 28 85 29 86 30 87 31 88 32 89 33 90 34 91 35 92

36 FIGURE 3 Mean plasma concentration–time 93 37 profiles of eight compounds after oral admin- 94 38 istration of FZJT tablet 95 39 96 40 The results in Table 2 indicate that flavonoids had homogeneous 3′‐methoxypuerarin can be transformed into 3′‐methoxyl‐6′′‐O‐ 97

41 absorption rates, and the rank order of elimination half‐life (t1/2) was acetylpuerarin by Caco‐2 cells (Wu, Xu, & Yang, 2013b). The double 98 42 daidzein >3′‐methoxypuerarin > mirificin > rutin > puerarin PK peak phenomenon of 3‐methoxypuerarin might not be attributed Q3 99 43 >3′‐hydroxypuerarin > daidzin > astragalin. The mean plasma concen- to bacteria‐mediated biotransformation, but can be explained by distri- 100 44 tration–time curve profiles of puerarin, 3′‐methoxypuerarin, bution, reabsorption and enterohepatic circulation. 101 45 3′‐hydroxypuerarin and mirificin exhibited a consistent trend in vivo As shown in Figure 3(E and F), daidzin and daidzein achieved the 102 46 owing to their parallel chemical structures. As shown in Figure 3 maximum plasma concentration at 0.56 and 0.67 h, respectively. The 103

47 (A–D), puerarin, 3′‐hydroxypuerarin, mirificin and 3′‐methoxypuerarin values of t1/2 of daidzin and daidzein were 0.36 and 6.44 h, respec- 104 48 were quickly absorbed into the body and could be detected 5 min after tively. Meanwhile, the area under concentration–time curve (AUC) of 105 49 oral administration. The four compounds achieved the maximum daidzin was much higher than that of daidzein. Double peaks emerged 106 50 plasma concentration between 0.28 and 0.67 h. Meanwhile, the in curves of mean plasma concentration for daidzein. This phenome- 107 51 plasma concentrations of the four compounds were eliminated rapidly, non was also been reported in the literature (Yan et al., 2014). The 108 52 within 4 h of intragastric administration. However, the pharmacoki- double PK peak phenomenon of daidzein might be attributed to the 109 53 netic parameters of 3′‐methoxypuerarin were not entirely in biotransformation of most daidzin into daidzein. It was reported that 110 54 accordance with the data reported in previous literature (Zhao, Zhao, daidzin, ‐O‐, was easily hydrolyzed by glucosidases 111 55 Liu, Han, & Simultaneous, 2012). Double peaks emerged in curves of in small intestine (Miao et al., 2013). Meanwhile, puerarin and mirificin 112 56 mean plasma concentration for 3′‐methoxypuerarin (see Figure 3D), were reported to be transformed into the same aglycone, daidzein, by 113 57 which is reported for the first time. A previous report showed that Caco‐2 cells (Wu, Wu, Wang, & Yang, 2015). 114 1 TAO ET AL. 7 58 2 59 3 Astragalin and rutin are . As shown in Figure 3 60

4 , area (G and H), the times to reach the maximum plasma concentration (Tmax) 61 −∞ 5 0 of astragalin and rutin were 0.39 ± 0.14 and 0.44 ± 0.18 h, respectively. 62

6 Meanwhile, their t1/2 ranged from 0.31 to 0.76 h. It was demonstrated 63 7 that two flavonoid glycosides were rapidly absorbed and eliminated in 64 point; AUC 8 ‐ rat plasma after oral administration of FZJT total fraction. 65 9 Compared with several previous studies, the established method 66 10 had three advantages (see Table ). First, LLOQs of 3′‐hydroxypuerarin, 67 11 puerarin, mirificin, 3′‐methoxypuerarin, daidzin and daidzein in this 68 12 study were lower than those of previous studies. Second, a one‐step 69 13 protein precipitation method instead of liquid– liquid extraction 70 14 method was employed in this study. The recoveries of 71 15 3′‐hydroxypuerarin, puerarin, mirificin, 3′‐methoxypuerarin, rutin and 72 16 astragalin were much higher than those of previous studies. Lastly, 73 17 the validated method can be applied to simultaneously determine the 74 18 concentrations of eight flavonoids in plasma. The information 75 =6)

19 n described in the present study should be informative for future appli- 76

20 time curve from zero to the last quantifiable time cation of FZJT tablets in clinical therapy. 77 – 21 78 22 79 4 | CONCLUSIONS 23 80 24 81 A multi‐component pharmacokinetic study was conducted for the 25 82 simultaneous determination of 3′‐hydroxypuerarin, mirificin, puerarin, 26 83 3′‐methoxypuerarin, daidzin, rutin, astragalin and daidzein in rat 27 84 plasma in negative ionization mode after oral administration of FZJT 28 85 tablet by UPLC‐MS/MS. The excellent selectivity, sensitivity, precision, 29 86 accuracy and extraction recovery proved that the method is suitable Methoxypuerarin Daidzin Daidzein Astragalin Rutin , area under the plasma concentration t ′‐ 30 – 87 0 for pharmacokinetic study. The analysis was fast and easy owing to 31 88 the relative short chromatographic running time and straightforward 32 89 sample pretreatment. The rapid absorption of bioactive constituents life; AUC 33 ‐ 90 gave a reasonable explanation for the rapid effects of FZJT tablet in 34 91 the treatment of type II diabetes. The co‐existing compounds in FZJT 35 92 tablet could alter the pharmacokinetic behaviors of these eight constit-

36 Dawley rats after oral administration of FZJT tablet (mean ± SD, 93

– uents compared with other prescriptions of traditional Chinese medi- , elimination half

37 , mean residence time. 94 cine. The results will be helpful for the further investigation of the 1/2 −∞ t 38 0 95 mechanism of action of FZJT tablet and may provide a useful tool for 39 96 its clinical application. 40 97 41 98 ACKNOWLEDGMENTS 42 99 This study was supported by the Nutritional Science Foundation of 43 100 Hydroxypuerarin Mirificin 3 ‐

′‐ By Health Co. (no. TY0141103), the Natural Science Foundation for 44 101 the Youth of Jiangsu Province (no. BK20140963) and the Priority 45 102 Academic Program Development of Jiangsu Higher Education 46 103 , time to maximum concentration; Institution.

47 max 104 T 48 105 REFERENCES time curve from time zero to infinity; MRT

49 – 106

Puerarin 3 Bebrevska, L., Foubert, K., Hermans, N., Chatterjee, S., Van Marck, E., 50 107 De Meyer, G., … Apers, S. (2010). In vivo antioxidative activity of a 51 quantified Pueraria lobata root extract. Journal of Ethnopharmacology, 108 52 127(1), 112–117. 109 53 Habicht, S. D., Ludwig, C., Yang, R. Y., & Krawinkel, M. B. (2014). Momordica 110 charantia and type 2 diabetes: from in vitro to human studies. Current 54 111 Pharmacokinetic parameters of the eight analytes in male Sprague Diabetes Reviews, 10(1), 48–60. (ngh/L) 110.82 ± 13.52 21.53 ± 3.45 107.81 ± 13.40 131.76 ± 11.49 77.68 ± 9.25 53.08 ± 8.19 17.93 ± 0.96 28.52 ± 4.78 (h) 2.54 ± 0.36 7.39 ± 1.20 9.90 ± 1.30 4.71 ± 0.14 3.63 ± 0.47 8.87 ± 1.94 4.18 ± 0.21 5.10 ± 1.28 (ngh/L) 109.87 ± 13.16 16.14 ± 1.05 99.88 ± 8.70 131.71 ± 11.56 71.79 ± 4.20 52.20 ± 7.87 16.52 ± 0.39 27.64 ± 4.44 t 55 t ∞ 112 – – –

(h) 0.61 ± 0.25 0.50 ± 0.19 0.28 ± 0.08 0.67 ± 0.21 0.56 ± 0.18 0.67 ± 0.30… 0.39 ± 0.14 0.44 ± 0.18 0 0 0 He, J., Feng, Y., Ouyang, H. Z., Yu, B., Chang, Y. X., Pan, G. X., Gao, X. M. A. (ng/mL) 88.98 ± 10.86 9.97 ± 0.97 25.34 ± 2.28 37.85 ± 4.13 57.91 ± 3.07 8.73 ± 0.88 14.50 ± 1.78 12.54 ± 1.23 (h) 0.58 ± 0.29 0.38 ± 0.16 1.53 ± 0.66 3.55 ± 0.80 0.36 ± 0.09 6.44 ± 2.48 0.31 ± 0.04 0.76 ± 0.20 56 , maximum plasma concentration; (2013). sensitive LC‐MS/MS method for simultaneous determination 113 max max 1/2 max AUC T t C AUC MRT under the plasma concentration TABLE 2 57 C of six flavonoids in rat plasma: application to a pharmacokinetic study 114 1 8 TAO ET AL. 58 2 59 3 of total flavonoids from mulberry leaves. Journal of Pharmaceutical and Wu, S., Xu, W., & Yang, X. W. (2013b). Studies on biotransformation of 60 – 4 Biomedical Analysis, 84, 189 195. chemical constituents of tongmai formula by human intestinal flora. 61 Chinese Journal of Chinese Materia Medica, 38(20), 3510–3519. 5 Katz, D. A., Murray, B., Bhathena, A., & Sahelijo, L. (2008). Defining drug 62 disposition determinants: a pharmacogenetic–pharmacokinetic Wu, S., Wu, W., Wang, F. R., & Yang, X. W. (2015). Study of the biotrans- 6 63 strategy. Nature Reviews Drug Discovery, 7(4), 293–305. formation of tongmai formula by human intestinal flora and its 7 intestinal permeability across the Caco‐2 Cell Monolayer. Molecules, 64 Li, M., Wang, W., Xue, J., Gu, Y., & Lin, S. (2011). Meta‐analysis of the 20(10), 18704–18716. 8 clinical value of Astragalus membranaceus in diabetic nephropathy. 65 9 Journal of Ethnopharmacology, 133(2), 412–419. Xia, W., Sun, C., Zhao, Y., & Hypolipidemic, W. L. (2011). antioxidant activ- 66 ities of sanchi (radix notoginseng) in rats fed with a high fat diet. 10 67 Liu, B., Kongstad, K. T., Qinglei, S., Nyberg, N. T., Jager, A. K., & Staerk, D. Phytomedicine, 18(6), 516–520. 11 (2015). Dual high‐resolution alpha‐glucosidase and radical scavenging 68 profiling combined with HPLC‐HRMS‐SPE‐NMR for identification of Xu, B., Li, P., & Zhang, G. (2015). Comparative pharmacokinetics of 12 69 minor and major constituents directly from the crude extract of Pueraria puerarin, daidzin, , glycyrrhizic acid, liquiritin, berberine, – 13 lobata. Journal of Natural Products, 78(2), 294–300. palmatine and jateorhizine by liquid chromatography mass spectrome- 70 try after oral administration of Gegenqinlian decoction and active 14 71 Miao, W. J., Wang, Q., Bo, T., Ye, M., Qiao, X., Yang, W. Z., … Guo, D. A. components alignment (ACA) to rats. Journal of Chromatography B, 15 (2013). Rapid characterization of chemical constituents and rats metab- 988,33–44. 72 olites of the traditional Chinese patent medicine Gegen‐Qinlian‐Wan 16 Yan, Y., Chai, C. Z., Wang, D. W., Wu, J., Xiao, H. H., Huo, L. X., … Simulta- 73 by UHPLC/DAD/qTOF‐MS. Journal of Pharmaceutical and Biomedical 17 neous, Y. B. Y. (2014). determination of puerarin, daidzin, daidzein, 74 Analysis, 72,99–108. 18 paeoniflorin, albiflorin, liquiritin and in rat plasma and its 75 Motomura, K., Fujiwara, Y., Kiyota, N., Tsurushima, K., Takeya, M., application to a pharmacokinetic study of Ge‐Gen Decoction by a liquid 19 Nohara, T., … Ikeda, T. (2009). Astragalosides isolated from the root chromatography–electrospray ionization–tandem mass spectrometry. 76 20 of astragalus radix inhibit the formation of advanced glycation end Journal of Pharmaceutical and Biomedical Analysis, 95,76–84. 77 products. Journal of Agricultural and Food Chemistry, 57(17), 21 Zhang, W., Han, F., He, J., & Duan, C. (2008). HPLC‐DAD‐ESI‐MS/MS anal- 78 7666–7672. 22 ysis and antioxidant activities of nonanthocyanin phenolics in mulberry 79 Raman, A., & Lau, C. (1996). Anti‐diabetic properties and phytochemistry of (Morus alba L.). Journal of Food Science, 73(6), 512–518. 23 – 80 Momordica charantia L. (Cucurbitaceae. Phytomedicine, 2(4), 349 362. Zhang, K., Pugliese, M., Pugliese, A., & Passantino, A. (2015). Biological 24 81 Tao, Y., Zhang, Y., Cheng, Y., & Wang, Y. (2013). Rapid screening and iden- active ingredients of traditional Chinese herb Astragalus membranaceus 25 tification of alpha‐glucosidase inhibitors from mulberry leaves using on treatment of diabetes: a systematic review. Mini‐Reviews in Medicinal 82 26 enzyme‐immobilized magnetic beads coupled with HPLC/MS and Chemistry, 15(4), 315–329. 83 – 27 NMR. Biomedical Chromatography, 27(2), 148 155. Zhao, X., Zhao, Y. L., Liu, X. M., Han, W., & Simultaneous, Y. Z. G. (2012). 84 determination of six in rat plasma after administration of 28 Tao, Y., Jiang, Y., Li, W., & Cai, B. (2016). Rapid characterization and deter- 85 ‐ ‐ mination of isoflavones and triterpenoid saponins in Fu‐Zhu‐Jiang‐Tang total flavonoid from Gegen by ultra HPLC MS/MS. Journal of Separa- 29 – 86 tablets using UHPLC‐Q‐TOF/MS and HPLC‐UV. Analytical Methods, tion Science, 35(8), 984 993. 30 8(21), 4211–4219. 87 31 Ulbricht, C., Costa, D., Dam, C., D'Auria, D., Giese, N., Isaac, R., … Windsor, SUPPORTING INFORMATION 88 32 R. C. (2015). An evidence‐based systematic review of (Pueraria 89 Additional Supporting Information may be found online in the 33 lobata) by the Natural Standard Research Collaboration. Journal of 90 Dietary Supplements, 12(1), 36–104. supporting information tab for this article. 34 91 Uzayisenga, R., Ayeka, P. A., & Wang, Y. (2014). Anti‐diabetic potential of 35 92 Panax notoginseng saponins (PNS): a review. Phytotherapy Research, 36 28(4), 510–516. How to cite this article: Tao, Y., Chen, X., Jiang, Y., and Cai, B. 93 37 Wu, K., Liang, T., Duan, X., Xu, L., Zhang, K., & Li, R. (2013a). Anti‐diabetic (2016), A UPLC–MS/MS approach for simultaneous determi- 94 38 effects of puerarin, isolated from Pueraria lobata (Willd.), on nation of eight flavonoids in rat plasma, and its application to 95 ‐ 39 streptozotocin diabetogenic mice through promoting insulin expression pharmacokinetic studies of Fu‐Zhu‐Jiang‐Tang tablet in rats. 96 and ameliorating metabolic function. Food and Chemical Toxicology, 60, 40 97 341–347. Biomed. Chromatogr., doi: 10.1002/bmc.3828 41 98 42 99 43 100 44 101 45 102 46 103 47 104 48 105 49 106 50 107 51 108 52 109 53 110 54 111 55 112 56 113 57 114 Author Query Form

Journal: Biomedical Chromatography Article: bmc_3828

Dear Author,

During the copyediting of your paper, the following queries arose. Please respond to these by annotating your proofs with the necessary changes/additions. • If you intend to annotate your proof electronically, please refer to the E-annotation guidelines. • If you intend to annotate your proof by means of hard-copy mark-up, please use the standard proofing marks. If manually writing corrections on your proof and returning it by fax, do not write too close to the edge of the paper. Please remember that illegible mark-ups may delay publication. Whether you opt for hard-copy or electronic annotation of your proofs, we recommend that you provide additional clarification of answers to queries by entering your answers on the query sheet, in addition to the text mark-up.

Query No. Query Remark

Q1 AUTHOR: Please confirm that given names (red) and surnames/family names (green) have been identified correctly. Q2 AUTHOR: Please define TIC Q3 AUTHOR: Please define PK

USING e-ANNOTATION TOOLS FOR ELECTRONIC PROOF CORRECTION

Required software to e-Annotate PDFs: Adobe Acrobat Professional or Adobe Reader (version 7.0 or above). (Note that this document uses screenshots from Adobe Reader X) The latest version of Acrobat Reader can be downloaded for free at: http://get.adobe.com/uk/reader/

Once you have Acrobat Reader open on your computer, click on the Comment tab at the right of the toolbar:

This will open up a panel down the right side of the document. The majority of tools you will use for annotating your proof will be in the Annotations section, pictured opposite. We’ve picked out some of these tools below:

1. Replace (Ins) Tool – for replacing text. 2. Strikethrough (Del) Tool – for deleting text.

Strikes a line through text and opens up a text Strikes a red line through text that is to be

box where replacement text can be entered. deleted.

How to use it How to use it  Highlight a word or sentence.  Highlight a word or sentence.  Click on the Replace (Ins) icon in the Annotations  Click on the Strikethrough (Del) icon in the section. Annotations section.  Type the replacement text into the blue box that appears.

3. Add note to text Tool – for highlighting a section 4. Add sticky note Tool – for making notes at to be changed to bold or italic. specific points in the text.

Highlights text in yellow and opens up a text Marks a point in the proof where a comment box where comments can be entered. needs to be highlighted.

How to use it How to use it

 Highlight the relevant section of text.  Click on the Add sticky note icon in the

 Click on the Add note to text icon in the Annotations section.

Annotations section.  Click at the point in the proof where the comment

 Type instruction on what should be changed should be inserted. regarding the text into the yellow box that  Type the comment into the yellow box that appears. appears.

USING e-ANNOTATION TOOLS FOR ELECTRONIC PROOF CORRECTION

5. Attach File Tool – for inserting large amounts of 6. Add stamp Tool – for approving a proof if no

text or replacement figures. corrections are required.

Inserts an icon linking to the attached file in the Inserts a selected stamp onto an appropriate

appropriate pace in the text. place in the proof.

How to use it How to use it  Click on the Attach File icon in the Annotations  Click on the Add stamp icon in the Annotations section. section.  Click on the proof to where you’d like the attached  Select the stamp you want to use. (The Approved file to be linked. stamp is usually available directly in the menu that appears).  Select the file to be attached from your computer or network.  Click on the proof where you’d like the stamp to

 Select the colour and type of icon that will appear appear. (Where a proof is to be approved as it is,

in the proof. Click OK. this would normally be on the first page).

7. Drawing Markups Tools – for drawing shapes, lines and freeform annotations on proofs and commenting on these marks. Allows shapes, lines and freeform annotations to be drawn on proofs and for comment to be made on these marks..

How to use it

 Click on one of the shapes in the Drawing Markups section.  Click on the proof at the relevant point and draw the selected shape with the cursor.  To add a comment to the drawn shape, move the cursor over the shape until an arrowhead appears.

 Double click on the shape and type any text in the red box that appears.

For further information on how to annotate proofs, click on the Help menu to reveal a list of further options: