The reference standards for naringenin, nobiletin, narirutin, naringin, hesperidin, neohesperidin, sulfamethoxazole (internal standard, IS1), and honokiol (internal standard, IS2) were purchased from the Shanghai Yuanye Bio-Technology Co., Ltd.(Shanghai, China). Meranzin hydrate (MH) was obtained from the Huaxi Medical University Medicine Factory (Chengdu, China). Formic acid, methanol and acetonitrile (HPLC grade) were purchased from Sigma (United States). Water was purified by a Milli-Q water purification system (Millipore, Burlington, MA, United States). ZQ was obtained from the pharmacy in Xiangya Hospital, Central South University and authenticated by Professor S. Y. Hu.

Chromatographic analysis was carried out on an Agilent Technologies 1200 series LC system (Agilent Corporation, Santa Clara, CA, United States), equipped with a G1311A quaternary pump, a G1329A autosampler, a G1322A on-line degasser, and a G1316A thermostat. The chromatographic separation was achieved on an 300SB-C18 column (4.6 mm × 250 mm, i.d., 5 μm, Agilent, United States) and an equivalent guard column (5 μm, Phenomenex, United States). Gradient elution with 0.1% formic acid (A) and acetonitrile (B) was applied for chromatographic separation at a flow rate of 1.0 mL/min. The mobile phase was used as follows: 23% B from 0 to 6.0 min; 23–65% B from 6.0 to 6.5 min; 65–65% B from 6.5 to 11.5 min; 65–23% B from 11.5–12.0 min; 23% B from 12.0 to 18.0 min. The total analysis time was 18.0 min, and the injection volume was 10 μL.

An ABI 3200Q-TRAP quadrupole linear ion trap mass spectrometer with an electrospray ion source (ESI) was used in this study. The ESI source was operated in the positive ionization mode for naringenin, nobiletin, MH and sulfamethoxazole and in the negative ion mode for narirutin, naringin, hesperidin, neohesperidin and honokiol. The precursor-to-product ion transitions in multiple reaction monitoring mode (MRM) were used for mass analysis and quantification. The precursor-to-production pairs and optimized declustering potential, entrance potential, collision energy and collision cell exit potential for each analyte and IS are showed in Table 1. Other MS/MS setting conditions were as follows: curtain gas pressure at 20 psi, ion source gas 1 at 60 psi, ion source gas 2 at 55 psi, ion spray voltage at 5500 V (in positive ion source) or -4500 V (in negative ion source), and temperature at 600°C. High purity nitrogen was employed as both the nebulizing and drying gasses. The process was performed by Analyst 1.4.1 data acquisition and processing software (applied Biosystems/MDS Sciex).

The crude ZQ was extracted according to our previously described method (Zhang et al., 2017a). Briefly, the ZQ was infiltrated in distilled water (1:12, w/v) for 30 min and boiled twice for 30 min under 100°C. The two decoctions were merged, evaporated under pressure at 60°C, and lyophilized, outputting the ZQ lyophilized powder with a yield of 19.7%. To count the dosage, the contents of the main compounds in the ZQ extract were quantitatively measured using the same analysis conditions as described previously. In additionally, the lyophilized powder was evenly dispersed in distilled water at dosage of ZQ (20 g/kg) before the experiment.

All animal experimental procedures were performed in line with the guidelines for the institutional guidelines of the Animal Care and Use Committee of Central South University (Changsha, China). The protocol was approved by the animal experimental committee of Central South University. All efforts were used to minimize the pain and suffering of the rats. SPF male Sprague–Dawley rats (220–250 g) were supplied by Shanghai SLAC Laboratory Animal Co. Ltd. (Shanghai, China), and housed under ambient temperature of 23 ± 2°C with a 12 h light/dark cycle and 50% relative humidity in the laboratory, and free access to a normal standard chow diet and water. All the rats were fasted for 12 h with free access to tap water prior to the experiment.

ZQ lyophilized powder or fluoxetine (Eli Lilly, United States) was dissolved with saline. To evaluate dose-dependent effects, ZQ (5, 10, 20 g/kg) or fluoxetine (20 mg/kg) was orally administered as a single dosage 0.5 h before the test. To study the precise relationship between antidepressant effects and pharmacokinetics, ZQ (20 g/kg) was orally administered as a single dose 0.5, 1, 2, or 4 h before the test. Similarly, vehicle rats were orally administered an equal volume of saline using the identical procedure.

The swimming tests were conducted according to a previously described method (Flores-Burgess et al., 2017) with minor modifications. In brief, all experiments were operated between 12:00 h and 15:00 h in the light cycle. Each rat was individually put into a vertical glass cylinder with a diameter of 20 cm, water depth of 30 cm and water temperature of 25°C. A 15-min forced swimming test was conducted 24 h before the 5-min test. Each rat was deemed to be immobile when it stopped scrabbling, remained floating stilly in the water and only made necessary struggling to maintain its head over water. The total duration of immobility was recorded during the 5-min test.

Standard stock solutions of naringenin (2420 ng/mL), nobiletin (1750 ng/mL), MH (10480 ng/mL), narirutin (5720 ng/mL), naringin (5190 ng/mL), and hesperidin (2375 ng/mL) were prepared in methanol, and neohesperidin (3640 ng/mL) was dissolved in hot water. Then, the calibration standard stock solutions of gradient concentrations were obtained by serial dilution with 90% methanol/water. The internal standard (IS) solutions of sulfamethoxazole and honokiol were prepared to the concentrations of 754 ng/mL and 274 ng/mL in methanol, respectively. All working solutions were enclosed and stockpiled at 4°C, and then they were returned to room temperature before test.

The calibration standards were prepared by spiking 20 μL of appropriate standard working solution into 180 μL of blank plasma to obtain the concentrations ranging from 242.0 to 3.78 ng/mL for naringenin, 175.0–2.73 ng/mL for nobiletin, 1048.0–16.38 ng/mL for MH, 572.0–8.94 ng/mL for narirutin, 519.0–8.11 ng/mL for naringin, 237.5–3.71 ng/mL for hesperidin, and 364.0–5.69 ng/mL for neohesperidin.

Quality control (QC) samples at low, middle and high concentrations (3.78, 30.25, and 242.0 ng/mL for naringenin; 2.73, 21.88, and 175.0 ng/mL for nobiletin; 16.38, 131.0, and 1048.0 ng/mL for MH; 8.94, 71.5, and 572.0 ng/mL for narirutin; 8.11, 64.88, and 519.0 ng/mL for naringin; 3.71, 29.69, and 237.5 ng/mL for hesperidin; and 5.69, 45.5, and 364.0 ng/mL for neohesperidin) were also prepared by the same procedures. The QC samples in plasma were extracted in the same way as the standards samples on each analysis day.

To identify the main active compounds of ZQ in target tissue, six rats subjected to FST were sacrificed under anesthesia 0.5 h after oral administration of ZQ extract (20 g/kg), and then the hippocampal and cortex tissue was separated rapidly from the brain on an ice tray and stored at -80°C. All tissues (∼50.0 mg) were homogenized with 1.5 mL of ice-cold 0.9% saline.

Sample (200 μL) and 10 μL of mixed IS solution containing sulfamethoxazole (754 ng/mL) and honokiol (274 ng/mL) was added to Eppendorf tubes and extracted with 1 mL acetonitrile. The mixed solution was vortexed for 3 min and centrifuged at 10000 rpm for 10 min at 4°C to acquire the supernatant. The supernatant (800 μL) was moved to a clean Eppendorf tube, and then desiccated under a flow of nitrogen at 40°C. Finally, the residue was redissolved in 100 μL acetonitrile/water (90:10, v/v) with vortex-mixing for 3 min and centrifuged at 13000 rpm for 10 min. A 10 μL aliquot of the sample was injected into the LC/MS/MS system for analysis.

The analytical method was validated for its selectivity, linearity, sensitivity, accuracy, precision, matrix effect, recovery and stability according to US FDA bioanalytical method validation guidance (Zhang et al., 2017b).

Blank rat plasma samples or samples spiked with the IS were detected for endogenous or IS interference. Specificity and selectivity were investigated by comparing the chromatograms of six individual batches of blank rat plasma samples, plasma samples added with the 7 analytes and IS, and plasma samples 30 min after oral administration of ZQ extracts.

The calibration curve was obtained by plotting the peak area ratio (y) of the analytes to IS vs. the plasma concentration (x) of the analytes with weighted (1/x) least square linear regression. The lower limits of quantification (LLOQ) was conformed to the lowest concentration on the standard curve with an acceptable accuracy (relative error, RE ≤± 20%) and a precision (relative standard deviation, RSD ≤ 20%).

The intra- and inter-day precisions were assessed by RSD, and developed by analyzing three concentrations levels of QC samples. Six samples were measured for each concentration level on the same experimental day and the three successive days, respectively. The analytic precision was denoted as the RSD, and the accuracy was considered the RE of the measured average deviated from the nominal value. The acceptable norm for the RSD was required to be less than 15%, and the RE was required to be within ±15%.

The extraction recoveries of the analytes were calculated by comparing the peak areas gained from six plasma samples in the pre-extraction spiked with the analytes and the post-extraction at three QC levels. The matrix effects were evaluated by contrast the peak areas obtained from the plasma matrix added with working standard solutions with the pure standard solutions at the same concentrations. The extraction recovery and matrix effect of the IS were also measured by the same approach at identical concentration.

The stabilities of the 7 analytes in rat plasma was evaluated by assaying six replicates of QC samples at low, middle, and high concentrations in practical experimental conditions: samples post-preparation stored in auto-sampler condition for 12 h, and storage at -80°C for 1 month. The freeze- thaw stability was determined after the experience of three freeze-thaw cycles (from -80°C to room temperature). Samples were identified as stable if the accuracy bias was within 15.0% of the nominal concentration.

All of the rats were subjected to FST for 24 h before oral administration of the ZQ extract (20 g/kg). Plasma samples (about 0.5 mL) were gathered in a heparinized 1.5 mL Eppendorf tube from the venous plexus of the eye socket at 0.083, 0.25, 0.5, 0.75, 1.0, 1.5, 2.0, 3.0, 4.0, 6.0, 8.0, 12.0, and 24.0 h under anesthesia. Each sample was immediately centrifuged at 10000 rpm for 15 min to acquire the plasma. At last, the plasma was immediately placed at -80°C until measure.

All data were expressed as the mean ± SD. The plasma concentration of the 7 compounds during the experiment was calculated according to the daily calibration curve. The pharmacokinetic parameters, including C_max (the peak drug plasma concentration), T_max (the time to C_max), AUC_0-t (the area under the plasma concentration-time curve from 0 to time), AUC_0-∞ (the area under the plasma concentration-time curve from 0 to time infinity), and t_1/2 (the elimination half-life) were calculated by performing a non-compartment model using WinNonlin Professional Edition Version 2.1 (Pharsight Corporation, Sunnyvale, CA, United States). The immobile time in the FST test was evaluated by one-way ANOVA. Statistical significance was denoted at p < 0.05. Statistical analysis was operated by GraphPad Prism 5 software.

The analytical results indicated that the contents of naringenin, nobiletin, Meranzin hydrate (MH), narirutin, naringin, hesperidin and neohesperidin in ZQ extract were 0.075, 0.067, 0.011, 0.220, 1.233, 0.122, and 1.471 mg/g, respectively. Quantitative data also showed that naringin and neohesperidin were the more predominant flavonoids in ZQ extract. Furthermore, it is reported that the contents of the ingredients in ZQ were significantly affected by different geographical origins (Li et al., 2016).

ZQ (20 g/kg) and Fluoxetine (20 mg/kg) significantly decreased the immobility time of rats (111.39 ± 33.26, 116.07 ± 28.65 s) compared with the vehicle (173.18 ± 40.36 s, p < 0.05) in the FST (Figure 1). However, there were no statistically significant discrepancy in the ZQ extract (5 or 10 g/kg) (128.89 ± 44.71, 154.43 ± 43.94 s, p > 0.05). In line with previous studies (Zhang Y.J. et al., 2012), our results further confirmed the antidepressant effect of ZQ against FST by decreasing immobility time. The FST is used as a valid behavioral test to assess the efficacy of antidepressant drugs (Flores-Burgess et al., 2017). That is, the FST is also applied to survey the behavioral despair as well as passive, coping and negative behaviors which are regarded as a main symptoms of depression (Slattery and Cryan, 2012; Haj-Mirzaian et al., 2016). The behavioral despair is an act of “giving up hope of escaping” after stress, which can be reflected by the immobile behavior in animal modes of depression (Cryan et al., 2005). As such, remarkable reduction in immobility time in this test after giving drug is considered as a main indication of an antidepressant-like action.

In additional, the relationship between the antidepressant effect and pharmacokinetics was precisely investigated. As shown in Figure 1B, the immobility time was significantly reduced following oral administration of ZQ (20 g/kg) at 0.5 and 1 h (117.79 ± 26.55, 125.27 ± 27.54 s) before the test compared with the vehicle (175.24 ± 42.57 s, p < 0.05), whereas no significant decrease was observed at 2 and 4 h (146.53 ± 26.39, 166.81 ± 34.51 s) before the test. Many phytochemicals have been revealed in ZQ extract (Duan et al., 2014), however, absorbed bioactive compounds (ABCs) in ZQ in vivo remains unknown.

We optimized the mobile phase to obtain good chromatographic conditions. Compared to methanol, the mass response and resolution of the analytes and IS were greatly enhanced by acetonitrile instead of methanol after several trial tests. Additionally, peak symmetry and sensitivity of naringenin, narirutin and hesperidin were significantly perfected by the addition of 0.1% formic acid. Isomers of narirutin vs. naringin and hesperidin vs. neohesperidin have almost close polarity and identical precursor-to-product ion pairs in MS. Therefore, two pairs of isomers were separated completely with different retention times by optimization of the gradient elution (Figure 3). Similarly, the run time was also shortened in our study.

The standard solutions of the 7 analytes and IS were directly injected along with the mobile phase into the mass spectrometer with an ESI source. To seek more strong response values of all analytes and IS, both the positive and negative ionization modes were compared. The response of naringenin, nobiletin, MH and sulfamethoxazole surveyed in the positive [M-H]⁺ ionization mode was stronger compared with the negative ionization mode in the experiment. However, the response in the negative [M-H]^- ionization mode became higher than that in the positive ionization mode for narirutin, naringin, hesperidin, neohesperidin and honokiol. Figure 2 explains the product ion scan spectra of the analytes and IS. The MS/MS transitions and energy parameters of the 7 compounds and IS are listed in Table 1.

Due to their different pKa values, dissolution properties and concentrations as well as stabilities in the serous matrix, the sample extraction method for the various analytes in plasma was optimized, and the extraction procedure should be simple, time-saving, repeatable, minimize matrix effects and be as convenient as possible. To achieve desired recovery of the multiple compounds, liquid-liquid extraction (LLE) was applied because this method could concentrate the analytes. Therefore, acetonitrile, instead of methanol, was eventually chosen to obtain higher extraction recovery in our experiment.

To find what active components were absorbed in the blood and brain, which played latently acute antidepressant effect, we qualitatively assayed the samples by LC/MS-MS. Compared with the retention of the standards, we found that naringenin, nobiletin, MH, narirutin, naringin, hesperidin and neohesperidin were absorbed into the hippocampus and cortex of acute depressive rats after oral administration of ZQ extract (Figure 3). Interestingly, it has been reported that naringenin, nobiletin, MH, naringin and hesperidin play a protective role against acute depression (Pathak et al., 2013; Yi et al., 2014; Choi et al., 2017). Therefore, these findings support the possibility that these are the main active compounds of ZQ, which are representative of its antidepressant effect. However, the precise relationship between the antidepressant effects of ZQ and the pharmacokinetics remains unknown. This issue is a challenge to understanding pharmacokinetic significance and the use of ZQ.