Metabolic profile of Fructus Gardeniae in human plasma and urine using ultra high-performance liquid chromatography coupled with high-resolution LTQ-orbitrap mass spectrometry.

Abstract 1. In China, Fructus Gardeniae was used as a traditional Chinese medicine (TCM) with a wide array of biological activities. The bioactive components identified in Fructus Gardeniae mainly included iridoids, flavonids, pigments, and so on. Among them, iridoids were regarded as important compounds in Fructus Gardeniae. Though analyses of the constituents in biological samples after oral administration of Fructus Gardeniae effective fraction or its active compounds have been reported, few efforts have been made to investigate the metabolic profile of Fructus Gardeniae in humans. In this study, the constituents and metabolites of Fructus Gardeniae in human blood and urine after oral administration of Fructus Gardeniae were investigated using ultra high-performance liquid chromatography (UHPLC) coupled with high-resolution LTQ-Orbitrap mass spectrometery. 2. Totally, 14 constituents (two parent compounds and 12 metabolites) of Fructus Gardeniae were identified in human plasma and urine either by comparing the retention time and mass spectrometry data with that of reference compounds or by the accurate high-resolution MS/MS data of the chemicals. The compounds identified were mainly iridoid glycosides such as geniposide and the derivatives of genipin-O-glucuronide. Among them, 11 metabolites were detected in human plasma and urine while the other three metabolites including geniposidic acid (M1), demethylation derivative of genipin-O-glucuronide (M2), and dehydration product of mono-hydroxylated genipin-O-glucuronide (M9) were only discovered in human urine. Further, the possible metabolic pathways of Fructus Gardeniae in vivo were proposed and the peak area–time curve of the most abundant metabolite genipin-O-glucuronide (M13) in human plasma after oral administration of Fructus Gardeniae was depicted. The results suggested that a metabolic difference existed between rats and humans. 3. The results obtained in the present research would provide basic information to understand the metabolic profile of Fructus Gardeniae in humans and explore the chemicals responsible for the hepatotoxicity of Fructus Gardeniae in vivo. Moreover, it would be beneficial for us to further study the pharmacokinetic behavior of Fructus Gardeniae in humans systematically.

The bioactive components identified in Fructus Gardeniae mainly included iridoids (scandoside methyl ester, gardenoside, genipin, geniposide, and acetylgeniposide) (Wang et al., 2012;Zhou et al., 2005Zhou et al., , 2007, flavonids (Cai et al., 2011), pigments (corin, crocetin, and neocrocin A), organic acids (chlorogenic acids, vanillic acid, quinic acid, and their derivatives) (Clifford et al., 2010;Isacchi et al., 2009;Kim et al., 2006), and other lipophilic compounds (Cai et al., 2015). Previous studies mainly focused on the identification or simultaneous determinations of bioactive compounds in Fructus Gardeniae as quality control markers (Bergonzi et al., 2012;Coran et al., 2014;Du et al., 2008;Han et al., 2015;He et al., 2006;Lee et al., 2014;Ouyang et al., 2011;Yang et al., 2011). Among them, eight bioactive constituents including geniposidic acid, chlorogenic acid, genipin-1-b-gentiobioside, geniposide, genipin, rutin, crocin-1, and crocin-2 were suggested as quality control and producing areas differentiation markers of Fructus Gardeniae (Wu et al., 2014;Yin et al., 2015). Nowadays, medicinal herbs are increasingly drawing attention as alternative treatment approaches and the most important consideration involving medicinal plants is to identify the active compounds responsible for the pharmacological activities. Numerous researches have demonstrated that the major active constituents in Fructus Gardeniae were responsible for the majority of medical effects of this fruit. Plasma pharmacochemistry demonstrated that only the constituents absorbed into the blood have the chance to exhibit the effects . Recently, analyses of the constituents in rat plasma after oral administration of Fructus Gardeniae effective fraction or active compounds (genipin and geniposide) have been reported (Hou et al., 2008;Wang et al., 2013c;Yang et al., 2012;Zhou et al., 2010). Besides, the metabolic profiles of geniposide in rat urine (Han et al., 2011) and genipin in biological samples (rat urine, plasma, feces, and bile) have been clarified (Ding et al., 2013a). However, to our knowledge, few efforts have been made to investigate the metabolic profile of Fructus Gardeniae in human plasma and urine in the literature up to now.
With the development of advanced instrument, highresolution mass spectrometry (HR-MS) is being used extensively in metabolic analyses owing to its accurate measurements of the mass-to-charge ratio (m/z) of molecular, fragments, and retention time (Ding et al., 2013a;Dunn et al., 2013;Han et al., 2011;Ren et al., 2014;Zuo et al., 2015). Considering the importance of drug metabolism in the body, the aim of our study was to investigate the constituents of Fructus Gardeniae in human blood and urine after oral administration of Fructus Gardeniae using ultra highperformance liquid chromatography (UHPLC) coupled with high-resolution LTQ-Orbitrap mass spectrometery.

Plant materials
Fructus Gardeniae were purchased from Hebei jingcao Pharmaceutical Co., Ltd. (Hebei, China). The batch number was 20150614. The collection of Fructus Gardeniae was permitted by Affiliated Hospital of Inner Mongolia University for the Nationalities (Tongliao, China). The Fructus Gardeniae were authenticated as Gardenia jasminoides Ellis by professor Buhebateer (College of Mongolian Medicine, Inner Mongolia University for the Nationalities).

Chemicals and reagents
Seven standards including geniposidic acid, chlorogenic acid, shanzhiside methyl ester, genipin-1-b-gentiobiosid, geniposide, rutin, and ursolic acid were purchased from the National Institute for the Control of Pharmaceutical and Biological Products (Beijing, China); MS-grade formic acid was purchased from Thermo Fisher Biochemical Products Co., Ltd. HPLC-grade methanol, acetonitrile (Fisher, Fair Lawn, NJ) and purified mili-Q water were used for UHPLC-MS analysis.

Quality control of Fructus Gardeniae
Before oral administration, the quality of Fructus Gardeniae was evaluated by UHPLC-HR-MS. Certain amounts of geniposidic acid, chlorogenic acid, shanzhiside methyl ester, genipin-1-b-gentiobiosid, geniposide, rutin, and ursolic acid were dissolved with methanol individually (100 mg/mL for each compound) and stored at 4 C. According to the literature (Wu et al., 2014), an accurately weighted 1.0 g pulverized Fructus Gardeniae was extracted in an ultrasonic bath with 50 mL 50% methanol for 60 min to get the Fructus Gardeniae sample. The supernatants of the extract were centrifuged at 12 000 rpm for 10 min at 4 C prior to UHPLC-MS analysis.

Collection and preparation of biosamples
The pulverized Fructus Gardeniae was orally given to human volunteers (n = 6) at a clinical dose (5 g/60 kg). Blood specimens (2 mL) were obtained before dosing and subsequently at 30 min, 1, 2, 3, 4, 4.5, 5, 6, 8, 10, and 12 h after oral administration of Fructus Gardeniae. Samples were collected in heparinized eppendorf tubes and centrifuged at 3000 rpm for 15 min. The plasma was stored at À80 C until assay. Urine was also collected before dosing and subsequently during 0-12 h. This study was ethically approved by the Medical Ethics Committee of Affiliated Hospital of Inner Mongolia University for the Nationalities (Tongliao, China). The ethical approval number was NM-LL-2009-07-01. All the people voluntarily joined this study with informed consents. The baseline characteristics of the volunteers are shown in Table 1.
The human plasma and human urine samples were prepared according to the procedure described by Zuo et al. with slight modifications (Zuo et al., 2015;Zuo et al., 2014). At each time point, 100 mL aliquot of the plasma of each volunteer (n = 6) was mixed. Further, 30 mL of 1/100 ascorbic acid (1 g of ascorbic acid dissolved in 100 mL of saline) and 1800 mL of methanol were added to the mixed plasma, followed by vortexing for 5 min and centrifuging at 10 000 rpm for 15 min to remove the precipitation. The supernatant was transferred to a clean tube and evaporated to dryness under nitrogen blow at 35 C. The residue was redissolved in 100 mL of 60% methanol, vortexed for 3 min, and centrifuged at 10 000 rpm for 10 min. The supernatant was analyzed by UHPLC-HR-MS. The preparation of urine samples was conducted the same as the above procedure with plasma samples, except that 1 mL of urine (n = 6) was mixed with 50 mL of ascorbic acid (1/100) and 3 mL of methanol and finally the residue of urine samples was redissolved in 200 mL of 60% methanol. The supernatant was analyzed by UHPLC-HR-MS. Plasma and urine samples before administration were regarded as controls.

Instrumentation and chromatographic conditions
The assay was performed by a Dionex UltiMate 3000 hyperbaric liquid chromatography system coupled to an LTQ Orbitrap mass spectrometer via an ESI interface from Thermo Fisher Scientific (Bremen, Germany). The liquid chromatography system consisted of a diode array detector, an auto-sampler, a column compartment, and two pumps. Xcalibur 3.0, Metworks 1.3, and Mass Frontier 7.0 software packages (Thermo Fisher Scientific Inc., Fair Lawn, NJ) were used for data collection and data analysis.
Liquid chromatographic separations of Fructus Gardeniae and its metabolites were carried out using a Thermo Hypersil BDS C 18 column (Thermo Fisher Scientific Inc., Fair Lawn, NJ) (150 mm Â 2.1 mm, 2.4 mm). The LC separation was optimized according to the reported methods with slight modification (Fu et al., 2014b;Wu et al., 2014). The mobile phase consisted of 0.1% formic acid in water (solvent A) and methanol (solvent B). The optimized gradient elution was as follows: 5-8% B at 0-5 min; 8% B at 5-15 min; 8-30% B at 15-20 min; 30-95% B at 20-25 min; 95% B at 25-32 min; 5% B for equilibration of the column at 32-38 min. The injection volume was 3.0 mL and the flow rate was 0.3 mL/min. The temperature controlled column oven was set at 30 C and the sampler was set at 4 C.
The ESI source parameters were set as follows: the heater temperature and the capillary temperature were 350 C, the spray voltage was set at 3.5 kV in the positive mode and À3.5 kV in the negative ion mode, the normalized collision energy was 35 eV, sheath gas (N 2 ) flow was 35 psi, and the aux gas flow was 10 psi. The ESI source was operated in the negative mode for Fructus Gardeniae extract. To guarantee accurate identifications of metabolites, both positive and negative ion modes were used for the analysis of human plasma and urine. In the Fourier Transform (FT) cell, full MS scans were acquired in the range of m/z 50-1500 with a mass resolution of 30 000. In the MS/MS analysis, data requirements were set as data-dependent scans.

Qualitative analysis of Fructus Gardeniae
The quality of Fructus Gardeniae was evaluated by UHPLC-MS before oral administration. Totally, seven main bioactive constituents, most of which were reported as quality control and producing areas differentiation markers of Fructus Gardeniae, were confirmed in the Fructus Gardeniae extract with standard references ( Figure S1). The results showed that genipin-1-b-gentiobiosid and geniposide were the most abundant bioactive components, indicating that the Fructus Gardeniae used in our study achieved an excellent quality (Wu et al., 2014).

The proposed fragmentation pathways of four iridoid glycosides
In the present study, the possible ESI-HR-MS/MS fragmentation pathways of four iridoid glycosides (geniposide, geniposidic acid, shanzhiside methyl ester, and genipin-1b-gentiobioside) were deduced to assist the identification of metabolites in human plasma and urine. In the negative ion  (Fu et al., 2014a,b;Wu et al., 2010;Zuo et al., 2015), their fragmentation pathways were proposed (shown in Figure S2). The neutral losses of CH 3 OH, H 2 O, and C 6 H 10 O 5 were the main fragmentation patterns for iridoid glycosides in the negative ionization mode. In addition, the characteristic ions at m/z 225.0767 (C 11 H 13 O 5 ) and 123.0452 (C 7 H 7 O 2 ) were found in the MS/MS spectra of the four iridoid glycosides, which played an important role in the metabolite identification (Table 2).

Metabolite identification in human plasma and urine
First, a wide range of mass values from 150 to 1000 one by one in order were extracted without mass defect filter (MDF) by comparing the UPLC-MS chromatograms of the administrated versus control samples to discover the potential metabolites. Based on the obtained m/z of the potential metabolites, extractions of their HR-MS data with MDF were performed by comparing the UPLC-MS chromatograms of the administrated versus control samples to further identify the metabolites. The key parameters were carefully modulated as follows: the maximum tolerance of the MDF was set at 5 ppm; elemental compositions for expected and unexpected metabolite peaks were generated based on the extensively possible formula of the compounds. The structural elucidation of the constituents was further clarified by comparing their retention time and MS/MS fragmentations with those of standards or data reported in the literatures. False positive results were not reported. ) less than those of geniposide, suggesting that M1 was a demethylation product of geniposide, namely, geniposidic acid (Zhou et al., 2010;Zuo et al., 2015). And further, M1 was confirmed as geniposidic acid with standard reference (Figure 1) (Figure 2), a demethylation product of geniposide. Thus, M2 was preliminarily inferred as a demethylation derivative of genipin-O-glucuronide, which was previously reported as a metabolite of genipin in rat bile (Ding et al., 2013a) (Table 3)       literature (Ding et al., 2013a;Han et al., 2011 mass measurement showed that their chemical formula was 2 Da (2H) more than that of M13, suggesting two dihydrogen derivatives of M13. Their product ions at m/z 385.1134, 371.0974, 327.1069, and 227.0925 were 2 Da more than those of M13 as well. Their fragment at m/z 175.0248 indicated that a glucuronic acid was attached to the aglycone moiety. According to the literature, the fragments of M10 and M11 were the same as those of the metabolites reported (Ding et al., 2013a;Zuo et al., 2015) and they were deduced as ringopened derivatives of genipin-O-glucuronide.
M12 was a prototype in the plasma and urine. It diluted at   (Figure 4) assured its structure as genipin-O-glucuronide (Ding et al., 2013a;Zuo et al., 2015). For steric reasons, the glucuronation site of M13 at C-1 was more likely.  Figure 5). Other fragments at m/z 193. 0355, 175.0248, and 113.0244 were the same as those of M7 and M8. According to the above evidence, M14 was identified as a reduced product of M7 or M8, namely, the ring opened and reduced derivative of genipin-O-glucuronide (Han et al., 2011).
In general, 14 constituents (two parent compounds and twelve metabolites) of Fructus Gardeniae were identified in human plasma and urine by UHPLC-MS/MS data (Table S1). Among them, 11 metabolites were detected in human plasma and urine while the other three metabolites (M1, M2, and M9) were only discovered in human urine. Based on the metabolites detected in human plasma and urine, the possible metabolic pathways of Fructus Gardeniae in vivo were proposed ( Figure 6).

Discussion
Fructus Gardeniae, a commonly used TCM called Zhizi in Chinese, was first recorded in the book named ''Shen Nong's Herbal Classic'' in China. In Chinese medical theory, with a kind of cool and bitter character, it could ease the mind, reduce pathogenic fire, eliminate damp-heat, and remove heat-toxicity from blood. Nowadays, the exploitations of gardenia plants had been involved in food additives, dyestuffs, cultivation of ornamental plant, antiseptic, and new medicines Park et al., 2013). Zhang et al. (2014) reviewed that there are nearly 18 healthcare functions of the health food containing Fructus Gardeniae, and 23 formulas containing Fructus Gardeniae have assisted function to protect chemical liver injury and modulate the immune system. Liu et al. (2013) reviewed that crocins and iridoid glycosides exhibited considerable biological activities. At present, consumers are becoming more and more interested in naturally occurring colorants with bioactives, which could offer health benefits. Natural colorants derived from Fructus Gardeniae and its processed products are particularly of significance in the food industry (Wrolstad & Culver, 2012). In Japan, gardenia blue iridoid pigment extracted from Gardenia jasminoides is approved for food use. In Asian countries, geniposide is used as a functional food and traditional medicine, and is also applied as a food coloring (Cai et al., 2015). However, Fructus Gardeniae extracts were reported to exhibit hepatotoxicity (Ding et Wang et al., 2013a;Yang et al., 2006). Thus, being a functional food and TCM with multi-pharmacological activities and hepatotoxicity, an in-depth study of its constituents absorbed into the blood and its metabolites in vivo was necessary and significant.
In previous studies, Yang et al. (2012) detected 13 compounds in rat blood after oral administration of Fructus Gardeniae extract. Zhou et al. (2010) reported that 7 major iridoid glycosides (parent compounds) were characterized in rat plasma after intravenous administration of Gardenia jasminoides Ellis. In their studies, no conjugated metabolite was found in rat serum. Further, Wang et al. (2013c) identified four iridoid glucosides (parent compounds) and one of their metabolite in rat serum after administration of Fructus Gardeniae target fraction. By comparison, in our study, 11 constituents (one parent compound and 10 metabolites) in human plasma and 14 compounds (two parent compounds and 12 metabolites) in human urine were tentatively characterized after oral administration of Fructus Gardeniae (shown in Figure 6). Although glucuronidated derivatives of iridoids and iridoid glycosides were the main compounds detected in humans and mainly parent compounds were discovered in rats after oral administration of Fructus Gardeniae, the metabolic pathway of Fructus Gardeniae in humans agreed with those of geniposide and genipin in rats (Han et al., 2011;Ding et al., 2013a,b), except that hydroxylation was another metabolic pathway of Fructus Gardeniae in humans. The results suggested that a metabolic difference existed between rats and humans. It was reported that the noxious property of Gardeniae extracts was related to the hepatotoxicity and genetoxic of genipin  and geniposide (Ding et al., 2013b;Wang et al., 2013a;Yang et al., 2006). Our results provided a collaborative evidence that iridoid glycosides, especially M13 (the most abundant metabolite in human plasma and urine), a genipin-Oglucuronide metabolite, might be the chemicals responsible for the hepatotoxicity of Fructus Gardeniae in vivo. Further, the peak area-time curve of M13 in human plasma after oral administration of Fructus Gardeniae at a clinical dose was depicted based on the chromatographic peak area (Figure 7). As shown in Figure 7, the time of M13 to reach the maximum plasma concentration (T max ) was at 4.5 h. Zheng et al. (2012) reported that the T max of geniposidic acid in rat was 1.0 h, demonstrating an obvious pharmacokinetic difference of iridoid glycosides between humans and animals. In addition, by comparison with one blood collection time (20 min) in rats (Yang et al., 2012), our study collected human plasma at different time points (30 min, 1, 2, 3, 4, 4.5, 5, 6, 8, 10, and 12 h) to get a more scientific, reasonable, and comprehensive metabolite identification result. However, the structures of the compounds in TCM were so complex and many metabolites contain similar moieties that it would be hard to unambiguously confirm all the constituents solely by mass spectrometry (Zuo et al., 2015). Thus, the metabolite identification in the present study is preliminary and speculative, but it might be helpful to understand the metabolic profile of Fructus Gardeniae in vivo and discover the active constituents of Fructus Gardeniae responsible for its hepatotoxicity and pharmacological activities.

Conclusions
In the present study, a simple and specific UHPLC-HR-MS method for the quality control of Fructus Gardeniae and its metabolic profile in human plasma and urine was developed. Totally, two parent compounds and 12 metabolites of Fructus Gardeniae were identified in human plasma and urine based on the characteristic fragmentation behaviors of four iridoids and the accurate high-resolution MS/MS data of the chemicals. Further, the possible metabolic pathways of Fructus Gardeniae in vivo were proposed and the peak area-time curve of the most abundant metabolite M13 in human plasma after oral administration of Fructus Gardeniae was depicted based on the chromatographic peak area. The results obtained in the present research would provide basic information to understand the metabolic process of Fructus Gardeniae in humans. Moreover, it would be beneficial for us to further study the pharmacokinetic behavior of Fructus Gardeniae in humans systematically.