Multi-Component Comparative Pharmacokinetics in Rats After Oral Administration of Fructus aurantii Extract, Naringin, Neohesperidin, and Naringin-Neohesperidin

Citrus × aurantium L., Chinese name: Fructus Aurantii (FA) has been largely used as Qi-invigorating herb in China for centuries. The main components (meranzin hydrate, naringin, neohesperidin, meranzin, nobiletin) have good physiological activity with relatively high abundance in FA. Few multi-component comparative pharmacokinetics are simultaneously accessible for the flavone glycosides, polymethoxy flavones, and coumarins in FA. In this work, a reliable and rapid ultra-high performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS) method was established and validated to determine the five ingredients in the SD rat plasma, and further applied to the pharmacokinetic studies after oral administration of monomer, drugs in compatibility, and FA extract. After hydrolysis with β-glucuronidase and sulfatase, the concentration of naringin and neohesperidin in rat plasma were expressed respectively by the total concentration of naringenin and hesperitin which was determined by UPLC-MS/MS. Double-peak phenomenon was observed for naringin and neohesperidin, which may be due to the enterohepatic circulation or multiple site absorption of the two flavone glycosides. Meranzin hydrate and meranzin (coumarins) were absorbed rapidly (Tmax, about 1.0 h) but eliminated slowly (t1/2z exceeds 6.5 h). Nobiletin, a typical polymethoxy flavone, was also rapidly absorbed according to Tmax and AUC(0-t). DAS 3.1 software suggests the pharmacokinetic profiles of the five components in rats be depicted as a two-compartment pharmacokinetic model. There were significant differences in pharmacokinetic parameters for naringenin and hesperetin between the compatibility, FA extract group vs monomer group: ① remarkable increases in the values of AUC(0–∞), AUC(0–t) and Cmax; ② obvious decrease of CLZ/F; and ③ longer tmax and t1/2z. The results suggest that compatibility can promote mutual absorption and affect the elimination behaviors.

Our previous study indicated that the components (MH, meranzin, and nobiletin) have good physiological activity with relatively high abundance in FA besides naringin and neohesperidin. We also found that the drug quality is closely related to the content proportion of naringin vs neohesperidin in FA, which can partly elucidate the geo-authentic origin of the medicinal plant . So, it is interesting to investigate the pharmacokinetic differences of naringin and neohesperidin between the different forms (monomer, compatibility, and extract). Some literatures indicated that naringin and neohesperidin transiently exist in the plasma, and the glucuronides/sulfates of their aglycones (naringenin and hesperetin) were the main circulating metabolites (Felgines et al., 2000;Fang et al., 2006;Xu et al., 2019). Free naringenin and hesperetin can be easily obtained from the naringenin-glucronaide and hesperetinglucronaide in plasma by hydrolysis with b-glucuronidase and sulfatase. In addition, naringenin and hesperetin are relatively low content in FA. So, the concentration of the absorbed naringin and neohesperidin can be detected in the corresponding aglycone form (naringenin and hesperetin).
Hence, a rapid and sensitive ultra-high performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS) method was developed and validated for the simultaneous determination of the main components (naringin, neohesperidin, MH, meranzin, and nobiletin) in rat plasma, and used for the pharmacokinetic study of the compounds in different forms: monomer, compatibility, and extract.

Mass Spectrometry
The mass spectrometer was operated in negative mode for quercetin (IS), naringenin, and hesperetin, and positive mode for meranzin hydrate, meranzin, nobiletin. Quantification was obtained using multiple reaction monitoring (MRM) mode in the following positive/negative MS/MS scan segments. Segment I: 1.5-2.3min in positive mode, meranzin hydrate; Segment II: 2.2-4.5min in negative mode quercetin (IS), naringenin, and hesperetin; Segment III: 4.4-7.5min in positive mode for meranzin and nobiletin. The optimal MS parameters were as follows: Interface ESI, Nebulizing Gas Flow of 3.0 L/min; Heating Gas Flow of 10 L/min; Interface Temperature of 300°C; DL Temperature of 250°C; Heat Block Temperature of 400°C; Drying Gas Flow of 10 L/min; Nebulizing Gas of 3.0 L/min, Drying Gas F1 of 10.0 L/min, Conversion Dy: 10 kV, Detector Volt: 1.92 kV. MS/MS operating conditions were optimized by infusion of the standard solution (200 ng/ml) of each analyte and I.S. into the ESI source via a syringe pump. The optimal MRM parameters for these compounds are shown in Table 1.

Animals
Male Sprague-Dawley (SD) rats [Certificate No. SCXK(Xiang) 2016-0002] weighing 200 ± 20 g were purchased from Hunan Silaike Laboratory Animal Ltd (Changsha, China). The animals were kept in a controlled breeding room with the following conditions: a temperature of 22 ± 2°C, a relative humidity of (65 ± 5)%, and a 12 h light-dark cycle (Yuan et al., 2017). The Experimental Animal Ethic Committee of JXUTCM approved all animal protocols. The animal experiments were carried out according to the European Community guidelines for the use of experimental animals.

Preparation of FA Extract
The dried drug was extracted twice by refluxing with boiling Methanol (1:10, w/v) for 2 h, then the two filtrates were merged and evaporated to dryness under vacuum (Tong et al., 2012). The extract (14.56 g) was weighed and the extract yield was 29.14%. The obtained extract was stored at 4°C until use. To count the dosage, the contents of the main compounds in the FA extract were quantitatively measured by HPLC (Zhang et al., 2018;Chen et al., 2019), and the contents of naringin, neohesperidin, meranzin hydrate, meranzin and nobiletin in the extract were 171.6, 167.5, 1.98, 0.94, and 1.70 mg/g raw herb, respectively (see supplementary material: Figure S1).

Biological Sample Preparation
The 20 ml of b-glucuronidase and 15 ml of sulfatase were added into 50 ml of plasma sample. After vortexed for 30 s, the mixture was incubated at 37°C for 2 h in a water bath, then placed in ice water for 5 min to terminate the reaction (Xu et al., 2019). IS solution (20 µl, 5 mg/ml) and acetonitrile (1.0 ml) were added into the mixture and vortexed for 2 min. The mixture was centrifuged at 4,000 rpm at 4°C for 10 min, and the supernatant (850 µl) was transferred into a new eppendorf tube and evaporated to dryness. The residue was redissolved in 100 ml of 50% methanol, and vortexed for 3 min, then centrifuged at 14,000 rpm at 4°C for 10 min, finally the supernatant was used for UPLC-MS/MS analysis.

Method Validation
This method was validated according to the current US FDA Bioanalytical Method Validation Guidance (Guidance for Industry: Bioanalytical Method Validation, 2001). The following parameters were determined: selectivity, linearity, LLOQ, LOD, precision, accuracy, extraction recovery, matrix effect, and stability.

Linearity Parameters
To assess the linearity ranges, a series of the mixed standard solutions (six concentration levels) were prepared in triplicate. Each calibration curve (y = ax + b) was established by plotting the peak area ratio of analyte to IS (y) against the concentrations (x) of the calibration solution with a least square linear regression. The limit of detection (LOD) and limit of quantification (LOQ) were determined when the peak height was three times and ten times the background noise, respectively (Yuan et al., 2017).

Specificity and Selectivity
Blank plasma samples or the samples spiked with the IS were detected for endogenous or IS interference (Zhang et al., 2018). Specificity and selectivity were investigated by comparing the chromatograms of five individual batches of blank plasma samples, the samples spiked with the 5 analytes and IS, and plasma samples at 0.75 h after oral administration of FA extracts.

Accuracy and Precision
The intra-and inter-day precisions and accuracies were calculated by an analysis of variance based on the replicate analysis of QC samples, and the work was accompanied by a standard calibration curve on each analytical run (Yuan et al., 2017). Five samples were measured for each concentration level on the optimal conditions five times within the same day for intra-day variance, and the three different days for inter-day variance. The analytic precision was denoted as the RSD (relative standard deviation), and the accuracy was considered the RE (relative error) of the measured average deviated from the nominal value. The precision and accuracy values (RSD%) were required to be within ±20% for LLOQ (lower limit of quantification) and within ±15% for other concentrations.

Extraction Recovery and Matrix Effect
The extraction recoveries of the compounds were assessed by comparing the mean peak areas from five plasma samples in the pre-extraction spiked with the analytes and the post-extraction at three QC levels. The matrix effects were evaluated by comparing the peak areas obtained from the spiked plasma matrix with the pure standard solutions at the same concentrations. There was no matrix effect if the ratio was between 85-115%.

Stability
The QC samples were assayed under several different conditions to evaluate the stability of the compounds in rat plasma. For the short-term stability, fresh plasma samples were kept at room temperature for 24 h before the sample preparation. After keeping the samples at room temperature for 24 h, post-preparation stability was tested. Freeze-thaw cycles (-20°C/room temperature) stability were not treated at -20°C for freezing-room temperature melting three times (repeated freezing and thawing three times). In each freeze-thaw cycle, the samples were frozen and stored at -20°C for 24 h, and subsequently thawed at room. To evaluate the long-term stability of QC plasma samples, following storage at -20°C for one month. The stability data were acceptable when the bias was within ±15% of the actual value (Yuan et al., 2017).

Pharmacokinetics
Rats were randomly divided into the following four groups ( The rats were intragastric administrated with 1.5 ml of water at the time points of 0.25, 0.5, 2, 4, and 6 h, respectively. The rats had free access to water during the experiment. The blood samples were then immediately centrifuged at 4,000 rpm for 10 min. The plasma was frozen and stored at −80°C until analysis. The plasma concentration of the studied compounds was determined 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 the last time point), AUC (0−∞) (the area under the plasma concentration-time curve from 0 to time infinity), MRT (0-t) (average residence time from 0 to the last time point), and t 1/2z (the elimination half-life) were fitted with the proprietary DAS 3.1 pharmacokinetic program (Chinese Pharmacological Society).

Statistical Analysis
All statistical analyses were performed using the SPSS 21.0 software package from SPSS, Inc. (Chicago, IL, USA). *P values < 0.05, or **P values < 0.01 was considered statistically significant difference.

UPLC-MS/MS Conditions
According to the direct full-scan ESI mass spectra, the ion intensities of naringenin, hesperetin, and quercetin (IS) was higher in negative mode, and the other three compounds (meranzin hydrate, meranzin, and nobiletin) got greater sensitivity in positive mode. So, the switch scan method (ESI +/-) was used for the corresponding compounds, and the product ion mass spectra of the six compounds are shown in Figure 1.
Based on previous work, chromatographic conditions such as the constituents and the gradient profiles of the mobile phase were further improved to adapt the separation and detection of six components in rat plasma. In addition, the optimum mobile phase was water (containing 0.1% formic acid) and acetonitrile.
Various sample preparation methods including protein precipitation with methanol and acetonitrile were investigated to obtain an efficient cleanup of the biological samples. In this work, acetonitrile was chosen as the protein precipitation agent with simplicity and efficiency.

Specificity
Typical chromatograms of blank, spiked plasma and plasma samples are shown in Figure 2. No interference was found between endogenous compounds or xenobiotics. The developed method results in six single sharp peaks at the retention time of meranzin hydrate, quercetin (IS), naringenin, hesperetin, meranzin, and nobiletin were 2.03, 2.46, 3.29, 3.63, 5.09, and 5.84 min, respectively.

Linearity Parameters
The linearity parameters of the five components in rat plasma using weighed (1/x 2 ) least-squares linear regression analysis are summarized in Table 2. Within the investigated linear range, good linearity data was obtained with correlation coefficients >0.9941. The results indicate the proposed UPLC-MS/MS method has a wide linear range and high sensitivity.

Precision and Accuracy
The precision and accuracy results are shown in Table 3, intraday precision was less than 7.9%, intra-day accuracy was between -8.66 to 9.68%, and inter-day precision was less than 10% for the five compounds. Inter-day accuracies were in the range of -9.34-9.94% expressed by relative errors (RE%). The relative standard deviation and relative error values did not exceed ± 15% (Table  3), this indicate the accuracy and precision are acceptable.

Extraction Recovery and Matrix Effect
The extraction recovery and matrix effect were summarized in Table 4. The recoveries were between 82.33 to 89.92%, and the matrix effects of the five compounds were in the range of 94.36-101.65%. These data indicate that the UPLC-MS/MS method provides suitable, reproducible, and reliable data and can be used as an analytical tool in rat plasma.

Stability
The stability data under different experimental conditions are summarized in Table 5. The results showed all RSD% were less than 8.8% and RE% were in the range from -13.50 to 14.81%, which suggested that rat plasma samples containing 5 analytes were stable under routine laboratory conditions and no additional procedures were necessary to stabilize the sample for pharmacokinetic studies.

Pharmacokinetics
The proposed UPLC-MS/MS method was applied to a multicomponent pharmacokinetic study of meranzin hydrate, naringenin, hesperetin, meranzin, and nobiletin in SD rats after oral administration of the extract, NA, NHE, and NA-NHE. The plasma concentration-time profiles of the five components were shown in Figure 3, and the corresponding pharmacokinetic parameters were shown in Tables 6 and 7. The physiological disposition conformed to a two-compartment model of the 5 analytes in rats fitted by the DAS 3.1 software.
The results suggested that there were significant differences in pharmacokinetic parameters of naringenin and hesperetin between different drug forms, including AUC (0-t) , AUC (0-∞) , C max , t max , t 1/2z , and CL Z/F . The main differences including   remarkable increases in the values of AUC (0-∞) , AUC (0-t) and C max , obvious decrease of CL Z/F , and longer t max and t 1/2z . The comparisons of C max and AUC (0-t) of different groups were shown in Figure 4.

DISCUSSION
UPLC-MS/MS is a powerful analytical tool for the trace compounds in biosamples with high selectivity and sensitivity (Yuan et al., 2017). Various influencing factors were investigated carefully including separation and detection conditions in this work, and the optimal analytical parameters were obtained for the studied analytes. In addition, the mass spectrometer can be switched freely between ESI (Electro-spray Ionization) +/detection modes in a single injection, and it ensures that each analyte can get the best response. Therefore, the developed method achieved higher sensitivity than the literatures did with lower LODs (see Table 2).
A study  has reported that flavone glycosides (naringin, hesperidin, and neohesperidin) could not been detected in the rat plasma even after orally administering 30 g/kg FA decoction, which might be associated with the hydrolysis caused by microbial bacterial from the gastrointestinal tract (Hsiu et al., 2002). Naringin can be quickly absorbed, and metabolized into its aglycone and naringenin glucuronide. Hesperidin was also quickly transferred to hesperetin due to the similar structure Wang et al., 2018). In general, naringin and hesperidin are hydrolyzed in the gastrointestinal tract by the enzymes of intestinal bacteria followed by absorption and conjugation of their aglycones, and the main present forms are naringenin and hesperitin glucuronide/sulfatase conjugations and a small amount of the free aglycones in plasma (Xu et al., 2019). Therefore, the concentration of naringin and neohesperidin in mouse plasma were expressed respectively by the total (free + conjugated) concentration of naringenin and hesperitin which was determined by UPLC-MS/MS after hydrolysis with bglucuronidase and sulfatase in this work.   As shown in Figure 3, the change trend in concentration-time profiles of naringenin and hesperetin are basically consistent with previous reports (Xu et al., 2010;Tong et al., 2012;Xu et al., 2019): both had double peaks. The double-peak phenomenon may be due to the enterohepatic circulation and drug reabsorption of naringin and neohesperidin in rats, which was also reported for other flavonoid glycosides from the extract of TCM Lu et al., 2007;Zhang et al., 2012;Okada et al., 2017). In addition, these components can be detected in rat plasma within 5 min after dosing with a short T max (less than 1 h), and which indicates that they be absorbed easily and rapidly after oral administration. The values of C max and AUC of naringenin and hesperitin are relatively big, which indicates their bioavailability is not as low as reported in the literatures . The metabolic behaviors of the flavonoid glycosides might also be explained by the enterohepatic circulation or multiple site absorption . Meanwhile, the double-peak phenomenon and the so-called low bioavailability might also suggest that the reabsorption of the metabolites secreted in bile for FA should be more obvious.
Meranzin hydrate and meranzin, as the typical coumarins, were absorbed rapidly (T max , about 1 h) but eliminated slowly (t 1/2z exceeds 6.5 h). Nobiletin, a polymethoxy flavone, was also rapidly absorbed with T max value of 1.17 ± 0.26 h and AUC (0-t) value of 472.64 ± 118.70, which was consistent with the previous report (Onoue et al., 2013;Shen et al., 2019). The average t 1/2z of nobiletin was 2.10 ± 0.37 h in rats, suggested that nobiletin should be rapidly distributed and eliminated (Shen et al., 2019), which was partly different from the previous report, probably owing to the interactions among components in FA extract (Qiao et al., 2012;Zhang et al., 2018).
Statistical moment analysis (corresponding to noncompartmental model) and compartmental model analysis can be simultaneously performed in a run of batch analysis by DAS 3.1. In the current in vivo data analysis, the non-compartmental model has become the mainstream processing method and has been recommended by the drug review authorities in various countries (Yuan et al., 2017;Xu et al., 2019). In particular, the statistical moment analysis is more reliable for the drugs with double-peak phenomena. In this work, naringenin and hesperetin are found to have exactly these pharmacokinetic properties. So, the statistical moment parameters are shown in Tables 6 and 7. In the compartmental model analysis, various models and different statistical weights were compared in this work, and two-compartment parameters were found to be the optimal fitting values with the minimum AIC (Akaike's Information Criterion) and the best R (correlation coefficient) values (Yuan et al., 2017;Yang et al., 2019). The compartmental model parameters are shown in Tables S1 and S2.
As mentioned above, the pharmacokinetic characteristics of naringenin and hespeietin are closely related with enterohepatic circulation or multiple site absorption/reabsorption, which indicates their pharmacokinetics may be affected by the forms in which they exist. Therefore, we design the following four groups to investigate the effects: pure monomer (naringenin and hesperitin), compatibility (NA-NHE), and FA extract group. The results demonstrated that there were significant differences in pharmacokinetic parameters between different drug forms, including AUC (0-t) , AUC (0-∞) , C max , t max , t 1/2z , and CL Z/F . The main differences were observed between the compatibility group and the other two groups: ① remarkable increases in the values of AUC (0-∞) , AUC (0-t) and C max ; ② obvious decrease of CL Z/F ; and ③ longer t max and t 1/2z . The comparisons of C max and AUC (0-t) between different groups were shown in Figure 4. The results suggest that compatibility can promote mutual absorption and affect the metabolic behaviors. Drug-drug interactions often occur, and the detailed effect of compatibility on the pharmacokinetics needs further study.

CONCLUSION
The proposed UPLC-MS/MS method was suitable for the accurate determination of meranzin hydrate, naringenin, hesperetin, meranzin, and nobiletin in rat plasma with high selectivity and sensitivity. The analytical method has been successfully used for the pharmacokinetic studies following oral administration of monomer, drugs in compatibility, and FA extract. The five components in FA can rapidly be absorbed. Drug compatibility can promote mutual absorption and affect the metabolic behaviors.

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the article/ Supplementary Material.

ETHICS STATEMENT
The animal study was reviewed and approved by The Experimental Animal Ethic Committee of Jiangxi University of Traditional Chinese Medicine.

AUTHOR CONTRIBUTIONS
JY, FW, and WY conceived and designed the experiments. FW, XL, RQ, MZha, MZho, and HC performed the experiments. FW, JY, and WY analyzed the data. JY and FW wrote the paper. All authors contributed to the article and approved the submitted version.