DMSO, MTT, trypan blue, and other chemicals were purchased from Sigma. Paclitaxel (Phyxol^®), purity >99%, was obtained from Sinphar Pharmaceutical (Yilan County, Taiwan). Dulbecco’s modified minimal essential medium (DMEM), streptomycin, penicillin, and fetal bovine serum (FBS) were obtained from Gibco BRL. Trifluoroacetic acid, methanol, and acetonitrile were used as LC/MS/MS grade reagents (Merck, Darmstadt, Germany). Anhydrous ethyl ether (AEE) was used for sample preparation (Merck, Darmstadt, Germany).

The XSLJZT prescription was based on the unified formula announced by the Committee on Chinese Medicine and Pharmacy of the Department of Health (Taipei, Taiwan). It consists of Ginseng radix (Panax ginseng C. A. Mey. (2.5 g), Ren She), Atractylodis macrocephalae rhizoma (Atractylodes macrocephala Koidz. (5 g), Bai zhu); Hoelen (Wolfiporia cocos (Schw.) Wolf (5 g); Fu ling), and Glycyrrhizae Radix et Rhizoma. (Glycyrrhiza uralensis Fisch. ex DC. (2 g) Gan cao), Citri Reticulatae Pericarpium (Citrus reticulata Blanco, (2 g), Chen pi), Pinelliae Rhizoma (Pinellia ternata (Thunb.) Makino, (2.5 g), Ban xia), Amomi Fructus (Amomum villosum Lour. (2 g), Sha ren), and Aucklandiae Radix (Aucklandia costus Falc. (2 g), Mu xiang) (Liu et al., 2017). All materials were purchased from Sun Ten Pharmaceutical (New Taipei City, Taiwan) and authenticated by a non-profit organization, Brion Research Institute of Taiwan (New Taipei City, Taiwan). Voucher specimens (No. GR-20180001 for Ginseng radix, No. AMR-20180001 for Atractylodis macrocephalae rhizoma, No. H-20180001 for Hoelen, No. GLR-20180001 for Glycyrrhizae Radix et Rhizoma, No. CR-20180001 for Citri reticulatae pericarpium, No. PR-20180001 for Pinellia rhizoma, No. AF-20180001 for Amomi Fructus and No. AR-20180001 for Aucklandiae Radix was deposited at the College of Pharmacy, Taipei Medical University. XSLJZT was prepared using the following method: total drug weight of 112 g was placed in an extractor, distilled water (10-fold) was added and boiled for 30 min until half the volume was left. This procedure was repeated once. The two extracts were combined and filtered through gauze layers. The residue was discarded, and the filtrate was lyophilized at −20°C to yield a drug powder, which was stored at 4°C. The yield of XSLJZT was 25.6%.

The XSLJZT sample (0.5 g) was extracted using 20 ml of 70% methanol through ultrasonic oscillation at 25°C for 15 min and then vortexed at 160 rpm at 40°C for 40 min. The sample was then filtered through a 0.45 μm syringe filter, and a 20 μl sample was directly injected into the Waters HPLC system (Milford, MA, United States), which comprised a Waters 600 pump system, Waters 2996 Photodiode array detector, Waters 717 plus autosampler, and Sugai U-620 column oven (Wakayama City, Japan). A Cosmosil 5C18-MS-II reversed-phase column (5 μm, 4.6 mm × 250 mm, Nacalai Tesque, Japan) equipped with a Lichrospher RP-18 end-capped guard column (5 μm, 4.0 mm × 10 mm, Merck, Germany) was used as the stationary phase. The gradient elution was performed using the eluents A and B (A: acetonitrile; B: 0.1 % H₃PO₄) according to the following profile: 0–25 min, 19%–20% A and 81%–80% B; 25–60 min, 20%–40% A and 80%–60% B; 60–90 min, 40%–55% A and 60%–45% B; 90–100 min, 55%–60% A and 45%–40% B; 100–125 min, 19% A and 81% B. The flow rate was 1 ml/min, and the column temperature was maintained at 35°C. The following marker substances in XSLJZT were chosen according to the regulations of the Taiwan Herbal Pharmacopoeia and Pharmacopoeia of the People’s Republic of China: glycyrrhizin and liquiritin for Glycyrrhiza uralensis, costunolide and dehydrocostus lactone for Aucklandia lappa; and hesperidin for Citrus reticulata. We used an ultraviolet detection wavelength of 225 nm for costunolide and dehydrocostus lactone, 250 nm for glycyrrhizin, and 280 nm for liquiritin and hesperidin.

An LC/MS/MS system was used for analytical separation. It consisted of a Quattro Ultima mass spectrometer (Micromass, Manchester, United Kingdom), a Waters Alliance 2795 pump and autosampler (Waters, MA, United States), Biosil ODS 4.6 × 250 mm, 5-µm column (Biotic Chemical, Taipei, Taiwan). Data acquisition was performed using MassLynx vers. 4.0, (Micromass). The mobile phase consisted of 70% CH₃CN and 0.1% formic acid. The flow rate was 1.00 ml/min with a post column split of 1/10 to tandem MS. The ionization mode was electrospray/positive ionization, and mass scanning was conducted in the multiple reaction monitor (MRM) mode. The paclitaxel precursor ion was detected at m/z 854.36, and the product ion at m/z 286.04. The capillary voltage was 3.2 kV, cone voltage was 35 eV, source temperature was 80°C, desolation temperature was 400°C, and collision voltage was 25 eV. A primary standard stock solution of paclitaxel (1,000 μg/ml) was prepared by dissolving 1 mg of pure paclitaxel in 1 ml methanol to prepare a standard stock solution of 1 mg/ml. A working solution was prepared by diluting the stock solution in 50% (v/v) methanol to obtain the following concentrations: three standard solutions of 10, 100, and 1,000 ng/ml, representing low-, intermediate-, and high-strength stock solutions prepared using serial dilution.

All method validation experiments were performed according to the Food and Drug Administration (FDA) guidelines (Health, U.D.o. and H. Services, 2001). The matrix effect, recovery rate, accuracy, and precision were evaluated. The liquid–liquid extraction method was used for the preparation of all biological samples.

Male SD rats, provided by the Laboratory Animal Center of Taipei Medical University, were used to determine the effects of XSLJZT on the pharmacokinetics of paclitaxel. All animal experiments were done in accordance with regulations of the animal experimentation committee of Taipei Medical University (IACUC, approval no. LAC-2015-0105). The animals had a 12 h light/dark cycle and had free access to food and water. Each group consisted of six rats, and paclitaxel (2 mg/kg, i. v.) was administered to the control group, while other animals were divided into four groups according to the different treatments they received. The first treatment group (which represented co-treatment) received XSLJZT (250 mg/kg, p. o.) 2 h prior to paclitaxel (2 mg/kg, i. v.) administration, while in the second, third, and fourth groups (pretreatment), XSLJZT (250 mg/kg, p. o.) was administered for 3, 5, and 7 days prior to paclitaxel administration, respectively. Blood samples (220 μl) were collected from the jugular vein at the following time intervals: 1, 5, 10, 15, 30, 60, 120, 180, 240, 300, and 360 min following paclitaxel administration, and centrifuged at 1800 rpm for 10 min to collect the plasma. Samples were prepared using a liquid-liquid extraction method, and LC/MS/MS was used to detect paclitaxel.

To determine the influence of XSLJZT on the metabolism of paclitaxel, human liver microsomes were used (HLMs). The HLMs (0.5 mg/ml) were added with phosphate buffered saline (0.1 M, pH = 7.4) (PBS) and XSLJZT (0.5, 5, and 10 mg/ml) and paclitaxel (1, 4, 8, and 16 µM) and). The reaction was initiated by the addition of an NADPH generating system (1 mM NADP, 10 mM glucose 6-phosphate, 2 IU/ml glucose 6-phosphate dehydrogenase, and 5 mM MgCl2), and the samples were incubated for 30 min at 37°C in a shaking water bath. The reaction was terminated by the addition of acetonitrile (200 µL) and centrifugation at 18,000 × g for 10 min. The supernatant was extracted into new Eppendorf tubes and evaporated to dryness. The residues were then reconstituted for the analysis. Time dependency was evaluated by pre-incubating the HLMs with XSLJZT (0.5, 5, and 10 mg) for different pre-incubation times (0, 10, and 30 min). After each pre-incubation period, paclitaxel (16 µM) was added, and the samples were incubated for 30 min (Wattanachai et al., 2011; Cheng et al., 2017).

WinNonlin standard edition version 1.1 software was used to analyze the PK parameters. Each dataset was analyzed using a two-compartment model. The PK parameters observed included the drug concentration at zero time (C₀), the area under the concentration versus time curve (AUC), half-life (T_1/2), clearance (CL), and mean residence time (MRT). For the statistical analysis, two groups were compared with an unpaired Student’s t-test using Sigma Plot 10.0 software. Statistical significance was set to p < 0.05. Data were summarized as the mean ± standard deviation.

The retention times for liquiritin, hesperidin, glycyrrhizin, costunolide, and dehydrocostus lactone were 10.04, 18.27, 67.17, 89.25, and 91.92 min, respectively (Figure 1). Each Gram of XSLJZT contained liquiritin, hesperidin, glycyrrhizin, costunolide, and dehydrocostus lactone at 3.13, 16.86, 3.84, 0.04, and 0.15 mg, respectively.

The linearity of the calibration curve was established by plotting the peak ratio of paclitaxel. Good linearity was achieved over the concentration range of 5–1,000 ng/ml with a coefficient of estimation (r ²) of 0.9998. Chromatograms of the standard and paclitaxel-spiked rat plasma are shown in Figures 2A,B. The LOD of paclitaxel was 2 ng/ml with an S/N ratio of 5.3, and the LOQ was 5 ng/ml with an S/N ratio of 33.9. The values of both precision and accuracy were within 15% (Table 1), which is an acceptable criterion. Both the extraction efficiency and matrix effect were within the acceptable limits of >80% (Table 2). These results proved that the LC/MS/MS method was reliable and reproducible; thus, it was used for sample analysis.

The effects of XSLJZT on paclitaxel plasma concentrations were evaluated by comparing the changes in PK parameters between the treatment groups and the control group. The time-concentration curve is displayed in Figure 3, and the PK parameters are summarized in Table 3. The results indicated that XSLJZT significantly increased the AUC in the pretreated groups. The highest increase was observed in the 7-days-pretreated group with an AUC of 786.44 ± 193.14 min μg/ml as compared to the control group which had an AUC of 186.19 ± 64.42 min μg/ml. No significant change in the AUC (212.36 ± 138.92 min μg/ml) was observed in the co-treatment group. XSLJZT significantly prolonged the half-life (T_1/2) from 17.11 ± 2.48 min to 22.77 ± 2.08, 30.75 ± 8.81, 34.74 ± 7.97, and 37.56 ± 4.75 in the co-treatment groups and in the pretreatment groups of 3, 5, and 7 days, respectively. However, the MRT was significantly prolonged only in the pretreatment groups of 5 and 7 days.

SD rat liver tissues and Hep-G2 cells were used to evaluate the influence of XSLJZT on CYP 3A4 enzymes. XSLJZT inhibited the enzyme expression in both the rat liver tissues and Hep-G2 cells. In the rat liver tissue, XSLJZT inhibited CYP3A1/2 expression starting from 5 days pretreatment. However, the highest effect was observed in the 7 days pretreatment group (Figure 4). In Hep G2 cells, XSLJZT inhibited the expression of CYP3A4 in a dose-and time-dependent manner. The highest inhibition was observed at a concentration of 800 μg/ml XSLJZT (Figures 5, 6).

HLMs were incubated with paclitaxel (16, 8, 4, and 1 µM) with or without XSLJZT (0.5, 5, and 10 mg/ml) for 30 min. The residual paclitaxel concentration significantly increased in the groups where HLMs were incubated with both paclitaxel and XSLJZT as compared to the control group, which was incubated with paclitaxel only. This indicated that XSLJZT inhibited paclitaxel metabolism, which in turn increased its concentration (Figure 7). To determine the influence of XSLJT on enzyme activity at different incubation times, the HLMs were pre-incubated with different concentrations of XSLJZT (0.5, 5, and 10 mg/ml) for 0, 10, and 30 min. No significant difference was observed with the different pre-incubation times, indicating that XSLJZT did not significantly inhibit XSLJZT in a time-dependent manner (Figure 8). To understand the inhibition kinetics, different concentrations of XSLJZT (5 and 10 mg/ml) were incubated with different concentrations of paclitaxel (4, 8, and 16 µM). A Lineweaver-Burk plot and a secondary plot indicated that XSLJZT inhibited the enzyme activity in HLMs by competitive inhibition with a Ki value of 6.4 mg/ml (Figures 9A,B).