Chinese Journal of Natural Medicines  2018, Vol. 16Issue (11): 829-837  DOI: 10.3724/SP.J.1009.2018.00829
0

Cite this article as: 

YAO Qing-Qing, LI Li, XU Ming-Cheng, HU Hai-Hong, ZHOU Hui, YU Lu-Shan, ZENG Su. The metabolism and hepatotoxicity of ginkgolic acid (17:1) in vitro[J]. Chinese Journal of Natural Medicines, 2018, 16(11): 829-837.
[Copy]

Research funding

This study was supported by the National Key Project of China (No. 2017YFC0908600), the National Natural Science Foundation of China (No. 81173120) and the National Natural Science Foundation of Zhejiang Province (No. LQ15H310003)

Corresponding author

E-mail:yuls@zju.edu.cn (YU Lu-Shan)
Tel/Fax:86-571-88208407, E-mail:zengsu@zju.edu.cn (ZENG Su)

Article history

Received on: 28-Apr-2018
Available online: 20 November, 2018
The metabolism and hepatotoxicity of ginkgolic acid (17:1) in vitro
YAO Qing-Qing1 , LI Li2 , XU Ming-Cheng1 , HU Hai-Hong1 , ZHOU Hui1 , YU Lu-Shan1 , ZENG Su1     
1 Institute of Drug Metabolism and Pharmaceutical Analysis, Zhejiang Province Key Laboratory of Anti-Cancer Drug Research, College of Pharmaceutical Sciences, Hangzhou 310058, China;
2 Department of Pharmacy & Geriatrics Institute of Zhejiang Province, Zhejiang Hospital, Hangzhou 310013, China
[Abstract]: Pharmacological activities and adverse side effects of ginkgolic acids (GAs), major components in extracts from the leaves and seed coats of Ginkgo biloba L, have been intensively studied. However, there are few reports on their hepatotoxicity. In the present study, the metabolism and hepatotoxicity of GA (17:1), one of the most abundant components of GAs, were investigated. Kinetic analysis indicated that human and rat liver microsomes shared similar metabolic characteristics of GA (17:1) in phase Ⅰ and Ⅱ metabolisms. The drug-metabolizing enzymes involved in GA (17:1) metabolism were human CYP1A2, CYP3A4, UGT1A6, UGT1A9, and UGT2B15, which were confirmed with an inhibition study of human liver microsomes and recombinant enzymes. The MTT assays indicated that the cytotoxicity of GA (17:1) in HepG2 cells occurred in a time-and dose-dependent manner. Further investigation showed that GA (17:1) had less cytotoxicity in primary rat hepatocytes than in HepG2 cells and that the toxicity was enhanced through CYP1A-and CYP3A-mediated metabolism.
[Key words]: Ginkgolic acid (17:1)     Cytotoxicity     Liver microsomes     Recombinant enzyme     Hepatotoxicity    
Introduction

Ginkgolic acids (GAs), a mixture of structurally related n-alkyl phenolic acid compounds, are abundant in the leaves and seed coat of Ginkgo biloba L., a traditional Chinese medicine [1]. Previous studies have reported that GAs or standardized leaf extracts of G. biloba (EGb 761) elicit anti-cancer, anti-bacterial, and anti-HIV activities and have been used to treat cerebrovascular disorders and cognitive decline [2-7]. The molecular mechanisms underlying the anti-cancer and anti-bacterial activities of GAs have been further investigated. Baek et al. have found that GA (17 : 1), 6-[(10Z)-heptadecenyl] salicylic acid, one of the most abundant components of GAs, has anti-cancer effects against multiple myeloma cells by inhibiting the signal transducer and activator of transcription 3 (STAT3) signaling pathway [8]. GAs can partially inhibit the growth of pancreatic cancer by targeting the pathway driving lipogenesis [9]. In addition, GAs also demonstrate considerable anti-invasion and anti-metastatic activity in cancer cells, which may be explained by the inhibition of NF-κB essential modulator (NEMO) sumoylation, thus leading to the reduction of NF-κB activity and downregulation of metastasis-related genes in breast cancer cells, including uPA, PAI-1, CXCR4, and MMP-9, or by the inactivation of the PI3K/Akt/mTOR signaling pathway in lung cancer cells [10-11].

GAs have bioactive properties and are also known to have cytotoxic and allergenic potential [12-14]. Yao et al. have reported that ginkgolic acid (15 : 1) could induce MDCK cell necrosis via mitochondrial and lysosome damages and cell cycle arrest [15]. Berg et al. have examined the cytotoxicity and mutagenicity of three major GAs (13 : 0, 15 : 1, 17 : 1) in different cell lines, including male Chinese hamster lung fibroblasts (V79 cells) and salmonella typhimurium strains (TA97a, TA98, TA100 and TA102), showing that these tested GAs have cytotoxic potency but no mutagenic action [16]. Similarly, GAs are also known to be cytotoxic against human breast adenocarcinoma MCF-7 cells, human lung adenocarcinoma A549 cells and human leukemia HL-60 cells [17].

Liver, which contains an abundance of phase Ⅰ and phase Ⅱ metabolizing enzymes, is the major site for drug metabolism and biotransformation of most endogenous or exogenous substrates. Notably, liver is also the target site of toxicity caused by xenobiotics, which is one of the key factors that limit drug development and application. Jiang et al. have characterized GA (15 : 1)-induced liver damage in mice through a nuclear magnetic resonance (NMR)-based metabolomics approach in combination with biochemistry assays [18]. In our previous study, GA (15 : 1) showed cytotoxic activity in primary rat hepatocytes and HepG2 cells, and CYP1A-and CYP3A-mediated metabolites could enhance the cytotoxicity [19]. To date, there have been no reports about the hepatotoxicity of GA (17 : 1). Thus, the present study aimed to investigate the metabolic characteristics and metabolic pathways of GA (17 : 1) and to evaluate the hepatotoxicity of GA (17 : 1) in vitro.

Materials and Methods Chemicals and reagents

The GA (17 : 1) was prepared and assayed in our laboratory (purity > 98.8%) [20]. Dulbecco's modified Eagle's medium (DMEM) and non-essential amino acids were purchased from Invitrogen (Carlsbad, CA, USA). Fetal bovine serum (FBS) was purchased from Sijiqing Biological Engineering Materials Co., Ltd. (Hangzhou, China). Trypsin, α-naphthoflavone, sulfaphenazole, quinine, cimetidine, 4-methylpyrazole, ketoconazole, β-naphthoflavone, rifampin, phenacetin, midazolam, DL-isocitric acid trisodium salt hydrate, isocitrate dehydrogenase β-nicotinamide adenine dinucleotide phosphate sodium salt (β-NADP), and β-nicotinamide adenine dinucleotide phosphate reduced tetrasodium salt (β-NADPH) were obtained from Sigma Co. (St. Louis, MO, USA). Human liver microsomes and CYP3A4, CYP1A2, UGT2B15, UGT1A9, and UGT1A6 recombinant enzymes were obtained from BD Gentest (Woburn, MA, USA). The human hepatoma HepG2 cells were purchased from the Shanghai Cellular Research Institute (Shanghai, China). Methanol and acetonitrile were chromatography grade. All other chemicals used in the present study were of the highest grade that was commercially available.

Preparation of rat liver microsomes

Sprague-Dawley (SD) rats (male, weighing 180-210 g) were housed in a room with temperature at 25 ℃, a 12-h dark-light cycle, humidity of 50% ± 10%, and had access to water and a standard laboratory rodent diet ad libitum. After 12-h fasting, rat liver microsomes were prepared according to previously described ultracentrifugation methods [21]. Briefly, the rat liver tissues were minced thoroughly, homogenized in homogenization buffer (sucrose solution, 0.25 mol·L-1), and then centrifuged at 9000 g for 20 min at 4 ℃ to precipitate erythrocytes, nuclei and cytoderm. Then, the supernatants were centrifuged at 19 000 g for 20 min at 4 ℃ to precipitate mitochondria, and the supernatants obtained from this procedure were centrifuged again at 100 000 g for 60 min at 4 ℃. Finally, the supernatants were discarded, and the pellets were re-suspended with a sucrose solution containing 20 mmol·L-1 Tris (pH 7.4). The total protein concentrations in the rat liver microsomes were quantified with the bicinchoninic acid (BCA) protein assay. The microsomes were stored at -80 ℃ until use.

Isolation and culture of primary rat hepatocytes

Primary rat hepatocytes were isolated through the two-step collagenase perfusion method [19]. The cells were maintained in DMEM supplemented with 10% FBS, 1% non-essential amino acids, 100 U·mL-1 of streptomycin, 100 U·mL-1 of penicillin and 0.1 μmol·L-1 of insulin, in a humidified atmosphere containing 5% CO2 at 37 ℃.

HepG2 cell culture

HepG2 cells were cultured in DMEM supplemented with 10% FBS, 100 U·mL-1 of streptomycin and 100 U·mL-1 of penicillin in a humidified incubator at 37 ℃ with 5% CO2. The cells were detached from the flasks with 0.25% trypsin-EDTA when the cells reached 90% confluency. The cell passage ratio was set at 1 : 3.

Incubation conditions

Phase Ⅰ metabolism reactions of human liver microsomes and recombinant enzymes were performed under optimized incubation times and protein concentrations (0.5 mg·mL-1). The incubation system (100 μL) was consisted of microsomal protein or recombinant enzymes, 0.1 mol·L-1 Tris-HCl buffer (pH 7.4), 15 mmol·L-1 of MgCl2, 12 mmol·L-1 of DL-isocitric acid trisodium salt hydrate, 0.38 units of isocitrate dehydrogenase, and GA (17 : 1) with or without inhibitors or inducers. After pre-incubation at 37 ℃ for 5 min, the reaction was initiated by the addition of β-NADP and β-NADPH (0.26 mmol·L-1 and 0.12 mmol·L-1), respectively.

Phase Ⅱ metabolism reactions were performed in a phosphate-buffered system including 112.5 μg·mL-1 Triton X-100, 10 mmol·L-1 of MgCl2, 50 nmol·L-1 Tris-HCl (pH 7.4), liver microsomes or recombinant enzymes, and substrates. After pre-incubation for 5 min, 4 μL of UDPGA (77.4 mmol·L-1, diluted in 0.1 mol·L-1 K2HPO4, pH 7.5) was added to initiate the glucuronidation reaction.

At the end of the incubation period (30 or 60 min), 100 μL of ice-cold acetonitrile containing the internal standard (600 μmol·L-1 GA [15 : 1]) was added to terminate the reaction. After vortexing and centrifugation at 13 000 r·min-1 for 10 min, the supernatants were analyzed by HPLC. The organic solvent in the incubation mixture was controlled within 0.5% (V/V). All the experiments were performed in triplicate.

Instrument and analytical conditions

Agilent 1100 HPLC system (Agilent, Santa Clara, CA USA), equipped with a binary solvent delivery pump, column oven, DAD detector, and autosampler, was used for the analysis. Chromatographic separation was performed on Agilent Zorbax Eclipse C18 (4.6 mm × 250 mm, 5 μm, Santa Clara, CA, USA). The mobile phase was consisted of methanol and water containing 0.1% H3PO4 in ratio of 95 : 5 (V/V). Substrates were analyzed by ultraviolet detection at 245 nm with a flow rate of 1 mL·min-1. The injection volume was 30 μL. This method has been validated in our lab without endogenous interference from microsomes and incubation system in the chromatograms. The calibration curve was linear over the concentration range of 10-800 μmol·L-1 (r2 = 0.9938) for GA (17 : 1). Intra-and inter-day precisions and accuracy, determined by QC samples at three concentration levels (20, 100 and 640 μmol·L-1), the relative standard deviation (RSD) and relative error (RE) were all within 10%. Extraction recovery ranged from 93.2% to 94.1%. GA (17 : 1) was stable at room temperature for 12 h.

Chemical inhibition studies in rat and human liver microsomes

To determine which cytochrome P450 or UDP-glucuronosyltransferases isoforms are involved in GA (17 : 1) metabolism, specific CYP and UGT chemical inhibitors, such as α-naphthoflavone for CYP1A1/2, sulfaphenazole for CYP2C6, quinine for CYP2D1, cimetidine for CYP2C11, 4-methylpyrazole for CYP2E1, ketoconazole for CYP3A2, propofol for UGT1A8/9, quercetin for UGT1A2, diclofenac potassium for UGT2B1, phenylbutazone for UGT1A7 and trifluoperazine for UGT1A4, were introduced to the rat liver microsomes for chemical inhibition studies. The concentration ranges of inhibitors were 0-100 and 0-200 μmol·L-1 for the CYP and UGT chemical inhibitors, respectively. To further confirm the results, chemical inhibition studies of human liver microsomes were conducted.

Chemical induction studies of rat liver microsomes

The SD rats were randomly divided into three groups. One group was used as the control, and the other two groups were administered with a cytochrome P450-specific chemical inducer, β-naphthoflavone (inducer of CYP1A, dissolved in tea oil, i.p., 80 mg·kg-1·d-1) or dexamethasone (inducer of CYP3A, dissolved in saline, i.g., 100 mg·kg-1·d-1). After continuous administration for 3 days, the rat liver microsomes were prepared. Phenacetin and midazolam were used as probe substrates to verify enzyme activity. The effects of the chemical inducers on GA (17 : 1) metabolism in rat liver microsomes were then compared.

Determination of effect of GA (17 : 1) on CYP activity

The co-incubation system included rat liver microsomes, GA (17 : 1) and classic CYP substrates, such as midazolam (CYP3A1/2), phenacetin (CYP1A2), dextromethorphan (CYP2D1), chlorzoxazone (CYP2E1) and diclofenac (CYP2C6). The Ki values were calculated to evaluate whether GA (17 : 1) affected CYP activity. The concentrations of these specific substrates were set at 20, 50, and 100 μmol·L-1, and the concentration of GA (17 : 1) ranged from 0 to 400 μmol·L-1. After a 30-min incubation, the samples were treated using the same aforementioned method. Further verification experiments were performed with human liver microsomes under the same assay conditions.

MTT assay

The cytotoxicity of GA (17 : 1) in HepG2 cells and primary rat hepatocytes was determined using the MTT assay, according to the published method with some modifications [22]. Briefly, the cells were seeded at a density of 1 × 104 cells/mL in a 96-well plate. After 24 h, the medium was removed and replaced with 100 μL of medium containing GA (17 : 1) at various concentrations (0-80 μmol·L-1). At the same time, medium containing 0.5% DMSO was used as the negative control. After culturing for the designed times, the medium containing GA (17 : 1) was discarded and replaced with 100 μL of serum-free medium containing 0.05% MTT. The plates were then incubated at 37 ℃ for additional 4 h. The culture medium was removed, and 100 μL of DMSO was added to each well and the plates were shaken gently for 10 min to dissolve the formazan. Quantitation was performed with a SpectraMax M2 microplate reader at 570 nm. The absorbance of the treated cells was compared to the absorbance of the control cells that were exposed only to the vehicle, which was considered to represent 100% viability.

Cytotoxicity studies in different cell lines

After 24 h of culture, the medium of the HepG2 cells was changed to a serum-free medium containing various concentrations of GA (17 : 1) (final concentrations: 0, 5, 10, 20, 40, 50, and 80 μmol·L-1) or vehicle (0.5% DMSO) as a control. The cells were treated for additional 4, 12 or 24 h before MTT assays were performed. In addition, the effects of the CYP inducers and inhibitors on GA (17 : 1) cytotoxicity were determined.

Induction studies were performed using HepG2 and primary rat hepatocytes. After 24 h of culture, the cell culture medium was replaced with an induction medium containing 25 μmol·L-1 of β-naphthoflavone (inducer of CYP1A2) or 10 μmol·L-1 of rifampin (inducer of CYP2C9, CYP3A4 and UGT1A). The inducer-containing medium was replaced every 24 h. After 72 h of pretreatment, the cells were exposed to GA (17 : 1) for additional 24 h. The cytotoxicity was then determined by MTT assay.

For the inhibition studies, primary rat hepatocytes were co-incubated with different concentrations of GA (17 : 1) and an inhibitor (10 μmol·L-1 of α-naphthoflavone for CYP1A or ketoconazole for CYP3A) followed by MTT assay.

Each concentration for both induction and inhibition assays was tested in triplicate.

Statistical analysis

The results were expressed as means ± standard deviation (SD). The data obtained from the cytotoxicity studies were statistically analyzed with the unpaired Student's t-test using GraphPad Prism 5.0 (GraphPad, San Diego, CA, USA).

Results Kinetic assay

Human or rat liver microsomes were incubated with GA (17 : 1) at 20, 40, 80, 100, 200, 400, and 600 μmol·L-1. Enzyme kinetic curves are shown in Figs. 1 and 2. The enzyme kinetic parameters were calculated with the Michaelis-Mentent equation, according to the substrate reduction (Table 1). The Vmax, Km and CLint values of phase Ⅰ metabolism were 3.27 ± 0.27 μmol·min-1·mg-1 protein, 125.7 ± 39.9 μmol·L-1 and 0.026 ± 0.0068 L·min-1·mg-1 protein, respectively, for the rat liver microsomes, and 2.53 ± 0.24 μmol·min-1·mg-1 protein, 174.5 ± 41.2 μmol·L-1 and 0.015 ± 0.0044 L·min-1·mg-1 protein, respectively, for the human liver microsomes. For the phase Ⅱ enzyme kinetic studies, the Vmax, Km and CLint values were 3.64 ± 0.42 μmol·min-1·mg-1 protein, 206.6 ± 56.1 μmol·L-1 and 0.017 ± 0.0037 L·min-1·mg-1 protein, respectively, for the rat liver microsomes, and 2.97 ± 0.36 μmol·min-1·mg-1 protein, 209.4 ± 59.8 μmol·L-1 and 0.014 ± 0.0012 L·min-1·mg-1 protein, respectively, for the human liver microsomes. The results indicated that the characteristics of the GA (17 : 1) metabolic kinetics were similar in the human and rat liver microsomes.

Figure 1 Oxidation kinetics of GA (17 : 1) in rat (A) or human (B) liver microsomes (mean ± SD, n = 3)
Figure 2 Glucuronidation kinetics of GA (17 : 1) in rat (A) or human (B) liver microsomes (mean ± SD, n = 3)
Table 1 Kinetic parameters for the phase Ⅰ and phase Ⅱ metabolism of GA (17:1) in human and rat liver microsomes (mean ± SD, n = 3)
Metabolic pathways of GA (17 : 1) in vitro

Chemical inhibition studies in rat liver microsomes were a convenient method used in the present study to determine which cytochrome P450 isoforms were involved in GA (17 : 1) metabolism. As shown in Fig. 3A, the results indicated that α-naphthoflavone and ketoconazole could potently inhibit the CYP-mediated metabolism of GA (17 : 1). At a concentration of 40 μmol·L-1, α-naphthoflavone and ketoconazole inhibited microsome enzyme activity by approximately 40%. At a concentration of 100 μmol·L-1, α-naphthoflavone and ketoconazole were able to strongly enhance the inhibitory effect to 60% and 75%, respectively. In contrast, no inhibitory effects from sulfaphenazole, quinine, 4-methylpyrazole and cimetidine were observed. Therefore, GA (17 : 1) might be primarily metabolized by CYP1A1/2 and CYP3A2 isoforms in rat liver microsomes.

Figure 3 Effects of inhibitors and inducers on the metabolism of GA (17 : 1) as determined by using specific CYPs inhibitors or inducers in liver microsomes. (A) Effect of CYP inhibitors on the metabolism of GA (17 : 1) in rat liver microsomes. (B) Induction of CYP inducers increased the metabolism of GA (17 : 1) in rat liver microsomes. *P < 0.05, vs the negative control (mean ± SD, n = 3). (C) Effect of α-naphthoflavone and ketoconazole on the metabolism of GA (17 : 1) in HLMs. (D) Chromatograms of the metabolism of GA (17 : 1) in HLMs and rhCYPs. A: rhCYP3A4, B: HLMs, C: HLMs without NADP/NADPH, D: rhCYP1A2, and E: blank HLMs

To further confirm whether CYP1A and CYP3A isoforms were involved in GA (17 : 1) metabolism, chemical inducer studies were performed with β-naphthoflavone and dexamethasone. Phenacetin and midazolam were used as classical substrates to test enzyme activity (data not shown). After incubation with the induced rat liver microsomes, the metabolic rate of GA (17 : 1) was significantly increased by 1.9-fold (β-naphthoflavone-treated group) and 2.3-fold (dexamethasone-treated group), respectively, compared with the controls (Fig. 3B). Combined with the results of the inhibition studies, our findings suggested that CYP1A1/2 and CYP3A2 were involved in GA (17 : 1) metabolism in rats.

To confirm whether human CYP1A2 and CYP3A4 participate in GA (17 : 1) metabolism, specific inhibitors of CYP1A2 (α-naphthoflavone) and CYP3A4 (ketoconazole) were tested in human liver microsome studies, and the metabolism profile using human liver microsomes and recombinant human CYP (rhCYP) enzymes were determined. The results shown in Fig. 3C indicated that CYP1A2 and CYP3A4 were two enzymes that mediated GA (17 : 1) metabolism, which were consistent with the aforementioned results of the rat liver microsome studies. The metabolites from recombinant CYP1A2 and CYP3A4 were slightly different compared with human liver microsomes (Fig. 3D). Human liver microsomes could catalyze GA (17 : 1) to form metabolites M1, M2, and M3. M1 might be primarily produced by CYP3A4, whereas M3 was produced by CYP1A2 and CYP3A4.

Selective UGT inhibitor studies showed that propofol and phenylbutazone exhibited notable inhibitory effects on UGT activity on GA (17 : 1) in rat liver microsomes (Fig. 4). At a concentration of 100 μmol·L-1, propofol and phenylbutazone inhibited UGT enzyme activity by approximately 50%. In addition, the inhibitory effect became more significant as the concentration of propofol and phenylbutazone increased to 200 μmol·L-1. In contrast, the inhibitory effects mediated by high concentrations of quercetin, trifluoperazine and diclofenac potassium were lower than 20%. Thus, it could be preliminarily inferred that rat UGT1A7 and UGT1A9 participate in GA (17 : 1) glucuronidation.

Figure 4 Effects of UGTs inhibitors on GA (17:1) UGT activity in rat liver microsomes (mean ± SD, n = 3)

Rat UGT1A7 and UGT1A9 correspond to UGT1A6 and UGT1A9 in humans. An investigation using human liver microsomes and UGT recombinant enzymes further verified the important role of UGT1A6 and UGT1A9 in GA (17 : 1) glucuronidation. After incubation with 100 μmol·L-1 of GA (17 : 1) for 50 min, the metabolites produced by human liver microsomes or recombinant UGT1A6 and UGT1A9 showed no differences (Fig. 5). Moreover, recombinant UGT2B15 also formed the same metabolite, suggesting that species differences exist between humans and rats.

Figure 5 Chromatograms of the metabolism of GA (17 : 1) in HLMs and rhUGTs. A: UGT2B15, B: HLMs, C: UGT1A9, D: UGT1A6, E: blank HLMs, and F: HLMs without UDPGA
Inhibitory effects of GA (17 : 1) on CYPs

To explore whether GA (17:1) affected CYP activity, GA (17:1) was incubated with CYP-specific substrates. The Ki values of the CYPs were calculated using Prism based on the changes in substrate concentration (Table 2). The results showed that GA (17 : 1) exerted no notable effects on CYP3A2, CYP1A2, CYP2D1, and CYP2E1 (Ki > 100 μmol·L-1), except for CYP2C6 in rats. The Ki of CYP2C6 reached 5.26 ± 1.07 μmol·L-1, indicating a moderate inhibitory effect.

Table 2 Inhibitory effect of GA (17 : 1) on CYPs (mean ± SD, n = 3)

Rat CYP2C6 corresponds to CYP2C9 in humans. Therefore, tolbutamide and diclofenac, two substrates of CYP2C9, were used to verify the effect on GA (17 : 1) metabolism. Table 3 shows that GA (17 : 1) had a moderate inhibitory effect on CYP2C9 activity of both human liver microsomes and recombinant enzymes.

Table 3 The IC50 (μmol·L-1) values of GA (17 : 1) for CYP2C9 (mean ± SD, n = 3)
Cytotoxicity of GA (17 : 1) in HepG2 cells

After exposure to various concentrations of GA (17 : 1) (final concentrations 0, 5, 10, 20, 40, 50, and 80 μmol·L-1) for 4, 12, and 24 h, the cytotoxic effects of GA (17 : 1) on HepG2 cells were evaluated with the MTT assay.

As shown in Fig. 6, the cytotoxicity of GA (17 : 1) in HepG2 cells showed time-dependent and dose-dependent cytotoxicity. The IC50 values were calculated with GraphPad Prism 5 with the log (inhibitor) vs normalized response-variable slope equation. The estimated IC50 values for the 4-, 12-, and 24-h cultures were > 80, 67.0 ± 4.91, and 43.1 ± 4.85 μmol·L-1, respectively.

Figure 6 Cytotoxic effects of GA (17 : 1) on HepG2 cells as determined by the MTT assay. *P < 0.05, vs control (mean ± SD, n = 3)
Effects of the CYP inducers on GA (17 : 1) hepatotoxicity

To confirm the effects of certain cytochrome P450 isoforms on GA (17 : 1) cytotoxicity, studies of the effects of CYP inducers on GA (17 : 1) cytotoxicity were conducted using HepG2 cells and primary rat hepatocytes.

The cytotoxicity of GA (17 : 1) in β-naphthoflavone-or rifampin-pre-treated HepG2 cells is shown in Fig. 7A. The viability of HepG2 cells during the 12-h incubation period was closely related to the concentrations of GA (17 : 1). The IC50 values of the control group (0.1% DMSO), rifampin-pre-treated group and β-naphthoflavone-pre-treated group were 64.6 ± 5.17, 39.0 ± 5.02, and 36.0 ± 5.53 μmol·L-1, respectively, which indicated that the inducers could enhance the cytotoxicity of GA (17 : 1) in HepG2 cells.

Figure 7 Effects of cytochrome P450-specific chemical inducers on GA (17 : 1) cytotoxicity in HepG2 cells (A) and primary rat hepatocytes (B) as determined by the MTT assay. *P < 0.05 vs control (mean ± SD, n = 3)

Due to the short survival time of primary rat hepatocytes in vitro after isolation, the cells were cultured only for 24 h for the cytotoxicity studies. Fig. 7B shows that, at concentrations ranging from 40 to 80 μmol·L-1 of GA (17 : 1), cell viability was notably decreased after β-naphthoflavone or rifampin induction, compared with the control group (untreated hepatocytes). The IC50 values of the control group, rifampin-pre-treated group and β-naphthoflavone-pre-treated groups were > 80.0, 60.4 ± 7.96, and 49.0 ± 7.68 μmol·L-1, respectively.

These results indicated that CYP1A and CYP3A were involved in GA (17 : 1) biotransformation and that the hepatotoxicity of the metabolites was more potent than GA (17 : 1).

Effects of CYP inhibitors on GA (17 : 1) hepatotoxicity

To confirm the contributions of CYP3A and CYP1A in GA (17 : 1) metabolism, the cytotoxicity of different concentrations of GA (17 : 1) was determined after co-incubation with CYP inhibitors. As CYP1A2 and CYP3A4 activity was not detected in HepG2 cells [23], we used only primary hepatocytes as the cell model in the chemical inhibition study. Although primary hepatocytes had prominent advantages, including normal physiological concentrations of enzymes and cofactors and maintenance of the morphological integrity of primary hepatocytes, which reflect the metabolic environment in vivo, the short survival time and rapid degradation of cytochrome P450 enzymes limited the application of primary hepatocytes [24-25]. Thus, rifampicin or β-naphthoflavone was added to increase the expression levels of CYP3A and CYP1A when the primary hepatocytes had adhered to the dishes for 24 h. The effects of α-naphthoflavone and ketoconazole on the cytotoxicity of GA (17 : 1) in primary rat hepatocytes are shown in Fig. 8. These two inhibitors markedly increased cell viability (IC50 values were > 80 μmol·L-1 for all groups). Even after co-incubation with 80 μmol·L-1 GA (17 : 1), the loss of cell viability was only approximately 30%. Thus, these results indicated that the chemical inhibitors could alleviate the cytotoxicity of GA (17 : 1), which is primarily catalyzed by CYP1A and CYP3A, in hepatocytes.

Figure 8 The effect of cytochrome P450-specific chemical inhibitors on GA (17 : 1) cytotoxicity in primary rat hepatocytes as determined by the MTT assay. *P < 0.05 vs control (mean ± SD, n = 3)
Discussion

Several reports have investigated the pharmacokinetics, tissue distribution, excretion, and bioavailability of GAs in rats [22, 26]. A pharmacokinetic study has demonstrated that GA (17 : 1) is rapidly absorbed, with an absolute bioavailability of approximately 19.5% [26]. GA (17 : 1) has been shown to be a P-gp and BCRP substrate in cell and rat models, indicating that efflux function could be an important factor limiting the bioavailability of GA (17 : 1) [22]. Oral bioavailability could also be impacted by the first-pass metabolism, particularly the small intestine and liver, which contain abundant CYP enzymes [27-28]. Before GA (17 : 1) enters the circulatory system, a portion of GA (17 : 1) may be metabolized by CYP3A4 in the small intestine or liver, thus leading to low oral bioavailability.

In addition, enzyme kinetics can reflect the affinity of drugs to the metabolizing enzymes and the ability of enzyme catalysis. Thus, enzyme kinetic parameters (Km, Vmax and CLint) of GA (17 : 1) metabolism by human and rat liver microsomes are used to evaluate and compare enzymatic functions. Compared with GA (15 : 1) [19], the Vmax and CLint values of GA (17 : 1) were significantly different (P < 0.05), suggesting that GA (17 : 1) would be metabolized slower in the liver than GA (15 : 1). The affinity of GA (17 : 1) and GA (15 : 1) were similar because there were no significant differences in the Km values (P > 0.05). GA (17 : 1) and GA (15 : 1) share a common structure, comprising a benzene ring but a different side alkyl chain. As a result, CYP1A1/2 and CYP3A2 are the primary enzymes that mediate GA (17 : 1) and GA (15 : 1) metabolism in rat liver microsomes. Metabolites from the hydroxylation and glucuronidation of GA (17 : 1) and GA (15 : 1) have been identified in vitro and in vivo [29-31]. In addition, GA (17 : 1) exerts a moderate inhibitory effect on the activity of CYP2C9 in vitro. However, according to Chinese Pharmacopiea (2015 Edition), the maximum administration dose of GA (17 : 1) in human per day cannot reach the inhibitory concentration of GA (17 : 1). Thus, EGb inhibition of CYP2C9 related drug-drug interactions could be ignored.

The metabolism of xenobiotics is not only a detoxification process but also a procedure by which toxic metabolites are generated. The results in the present study indicated that GA (17 : 1) showed less cytotoxicity in primary rat hepatocytes than in HepG2 cells, possibly due to species differences and the different expression levels of phase Ⅱ enzymes. Phase Ⅱ metabolism is generally considered to be a detoxification process. Our results demonstrated that GA (17 : 1) was a substrate of rat UGT1A7 and UGT1A9 and human UGT1A6, UGT1A9 and UGT2B15. UGT1A6 and UGT1A9 were not detected in HepG2 cells but were abundant in human hepatocytes [23]. UGT1A7 was highly expressed in primary rat hepatocytes [32-33]. Although GA (17 : 1) and its CYP-mediated metabolites are more toxic to hepatocytes, we speculated that glucuronidation might ease the cytotoxicity in primary rat hepatocytes. GA (17 : 1) showed less toxicity in HepG2 cells but greater toxicity in primary rat hepatocytes than GA (15 : 1) [19]. The viability of HepG2 cells was less than 40% after 40 μmol·L-1 GA (15 : 1) treatment for 24h, while the cell viability was greater than 50% after 40 μmol·L-1 GA (17 : 1) treatment for same time. As there was little CYP1A and CYP3A expression in HepG2 cells, the toxicity might be directly caused by the parent compound. Thus, GA (17 : 1) seemed to be less toxic than GA (15 : 1).

Both GA (17 : 1) and GA (15 : 1) showed cytotoxicity in HepG2 cells, and CYP1A-and CYP3A-mediated metabolism could activate GA parent compounds to become more toxic. Compared with the β-naphthoflavone and rifampicin groups, the cytotoxicity of GA (17 : 1) in HepG2 cells showed no statistically significant differences, which differed from our previous research demonstrating that GA (15 : 1) showed more toxicity in HepG2 cells after treatment with β-naphthoflavone than with rifampicin. The results indicated that CYP1A contributed more to GA (15 : 1) metabolism, while CYP1A and CYP3A contributed equally to GA (17 : 1) metabolism [19]. Yang et al. have found that, compared with GA (13 : 0) and GA (15 : 1), GA (17 : 1) showed the strongest inhibitory dose-and time-dependent effect in SMMC-7721 cells, another human hepatoma cell model [34]. A previous study in our lab showed that GA (17 : 1) was predominantly accumulated in the liver and kidneys in rats [26]. Combined with the results from the present study, it is more likely that GA (17 : 1) would cause severe hepatotoxicity. In addition, the mechanism of GA cytotoxicity might be attributed to non-specific sirtuin (SIRT) inhibition, cell cycle arrest, and fragmentation of chromosomal DNA [2, 35].

In conclusion, the present study elucidated the phase Ⅰ and phase Ⅱ metabolism of GA (17 : 1) and evaluated the hepatotoxicity of GA (17 : 1). The results showed that human CYP1A2, CYP3A4, UGT1A6, UGT1A9, and UGT2B15 primarily mediated GA (17 : 1) metabolism. More importantly, the metabolites generated by CYPs were more toxic to primary hepatocytes and human hepatoma cells than parent compound. Thus, hepatotoxicity caused by GA (17 : 1) and its toxic metabolites should be investigated further.

Acknowledgement

We thank Ms. HU Hai-Hong for managing the instruments for drug analysis.

References
[1]
Jaggy H, Koch E. Chemistry and biology of alkylphenols from Ginkgo biloba L.[J]. Pharmazie, 1997, 52(10): 735-738.
[2]
Zhou C, Li X, Du W, et al. Antitumor effects of ginkgolic acid in human cancer cell occur via cell cycle arrest and decrease the Bcl-2/Bax ratio to induce apoptosis[J]. Chemotherapy, 2010, 56(5): 393-402. DOI:10.1159/000317750
[3]
Itokawa H, Totsuka N, Nakahara K, et al. A quantitative structure-activity relationship for antitumor activity of long-chain phenols from Ginkgo biloba L.[J]. Chem Pharm Bull, 1989, 37(6): 1619-1621. DOI:10.1248/cpb.37.1619
[4]
Yang X, Chen J, Qian Z, et al. Study on the antibacterial activity of ginkgolic acids[J]. J Chin Mater Med, 2002, 25(9): 651-653.
[5]
Lu JM, Yan S, Jamaluddin S, et al. Ginkgolic acid inhibits HIV protease activity and HIV infection in vitro[J]. Med Sci Monit, 2012, 18(8): BR293-298.
[6]
Mazza M, Capuano A, Bria P, et al. Ginkgo biloba and donepezil: a comparison in the treatment of Alzheimer's dementia in a randomized placebo-controlled double-blind study[J]. Eur J Neurol, 2006, 13(9): 981-985. DOI:10.1111/ene.2006.13.issue-9
[7]
Montes P, Ruiz-Sanchez E, Rojas C, et al. Ginkgo biloba extract 761: A review of basic studies and potential clinical use in psychiatric disorders[J]. CNS Neurol Disord Drug Targets, 2015, 14(1): 132-149. DOI:10.2174/1871527314666150202151440
[8]
Baek SH, Lee JH, Kim C, et al. Ginkgolic acid C 17 : 1, derived from Ginkgo biloba leaves, suppresses constitutive and inducible STAT3 activation through induction of PTEN and SHP-1 tyrosine phosphatase[J]. Molecules, 2017, 22(2): 276. DOI:10.3390/molecules22020276
[9]
Ma J, Duan W, Han S, et al. Ginkgolic acid suppresses the development of pancreatic cancer by inhibiting pathways driving lipogenesis[J]. Oncotarget, 2015, 6(25): 20993-21003.
[10]
Hamdoun S, Efferth T. Ginkgolic acids inhibit migration in breast cancer cells by inhibition of NEMO sumoylation and NF-kappaB activity[J]. Oncotarget, 2017, 8(21): 35103-35115.
[11]
Baek SH, Ko JH, Lee JH, et al. Ginkgolic acid inhibits invasion and migration and TGF-beta-Induced EMT of lung cancer cells through PI3K/Akt/mTOR inactivation[J]. J Cell Physiol, 2017, 232(2): 346-354. DOI:10.1002/jcp.v232.2
[12]
Hecker H, Johannisson R, Koch E, et al. In vitro evaluation of the cytotoxic potential of alkylphenols from Ginkgo biloba L.[J]. Toxicology, 2002, 177(2-3): 167-177. DOI:10.1016/S0300-483X(02)00189-0
[13]
Benezra C. Molecular recognition in allergic contact dermatitis. The concept of double-headed haptens[J]. Dermatol Clin, 1990, 8(1): 13-16. DOI:10.1016/S0733-8635(18)30513-8
[14]
Al-Yahya AA, Al-Majed AA, Al-Bekairi AM, et al. Studies on the reproductive, cytological and biochemical toxicity of Ginkgo biloba in Swiss albino mice[J]. J Ethnopharmacol, 2006, 107(2): 222-228. DOI:10.1016/j.jep.2006.03.014
[15]
Yao QQ, Liu ZH, Xu MC, et al. Mechanism for ginkgolic acid (15 : 1)-induced MDCK cell necrosis: Mitochondria and lysosomes damages and cell cycle arrest[J]. Chin J Nat Med, 2017, 15(5): 375-383.
[16]
Berg K, Braun C, Krug I, et al. Evaluation of the cytotoxic and mutagenic potential of three ginkgolic acids[J]. Toxicology, 2015, 327: 47-52. DOI:10.1016/j.tox.2014.10.001
[17]
Oh J, Hwang IH, Hong CE, et al. Inhibition of fatty acid synthase by ginkgolic acids from the leaves of Ginkgo biloba and their cytotoxic activity[J]. J Enzyme Inhib Med Chem, 2013, 28(3): 565-568. DOI:10.3109/14756366.2012.658786
[18]
Jiang L, Si ZH, Li MH, et al. 1H NMR-based metabolomics study of liver damage induced by ginkgolic acid (15 : 1) in mice[J]. J Pharm Biomed Anal, 2017, 136: 44-54. DOI:10.1016/j.jpba.2016.12.033
[19]
Liu ZH, Zeng S. Cytotoxicity of ginkgolic acid in HepG2 cells and primary rat hepatocytes[J]. Toxicol Lett, 2009, 187(3): 131-136. DOI:10.1016/j.toxlet.2009.02.012
[20]
Xia H, Wang X, Li L, et al. Development of high performance liquid chromatography/electrospray ionization mass spectrometry for assay of ginkgolic acid (15 : 1) in rat plasma and its application to pharmacokinetics study[J]. J Chromatogr B Analyt Technol Biomed Life Sci, 2010, 878(28): 2701-2706. DOI:10.1016/j.jchromb.2010.08.009
[21]
Zhou Q, Yao TW, Zeng S. Chiral metabolism of propafenone in rat hepatic microsomes treated with two inducers[J]. World J Gastroenterol, 2001, 7(6): 830-835. DOI:10.3748/wjg.v7.i6.830
[22]
Li L, Yao QQ, Xu SY, et al. Cyclosporin A affects the bioavailability of ginkgolic acids via inhibition of P-gp and BCRP[J]. Eur J Pharm Biopharm, 2014, 88(3): 759-767. DOI:10.1016/j.ejpb.2014.06.012
[23]
Guo L, Dial S, Shi L, et al. Similarities and differences in the expression of drug-metabolizing enzymes between human hepatic cell lines and primary human hepatocytes[J]. Drug Metab Dispos, 2011, 39(3): 528-538. DOI:10.1124/dmd.110.035873
[24]
Wilkening S, Stahl F, Bader A. Comparison of primary human hepatocytes and hepatoma cell line Hepg2 with regard to their biotransformation properties[J]. Drug Metab Dispos, 2003, 31(8): 1035-1042. DOI:10.1124/dmd.31.8.1035
[25]
Hewitt NJ, Lechon MJ, Houston JB, et al. Primary hepatocytes: current understanding of the regulation of metabolic enzymes and transporter proteins, and pharmaceutical practice for the use of hepatocytes in metabolism, enzyme induction, transporter, clearance, and hepatotoxicity studies[J]. Drug Metab Rev, 2007, 39(1): 159-234. DOI:10.1080/03602530601093489
[26]
Li L, Guo C, Hui Z, et al. Pharmacokinetics, tissue distribution and excretion studies of ginkgolic acid (17 : 1) in Rats[J]. Lat Am J Pharm, 2014, 33(2): 278-287.
[27]
Kawauchi S, Nakamura T, Yasui H, et al. Intestinal and hepatic expression of cytochrome P450s and mdr1a in rats with indomethacin-induced small intestinal ulcers[J]. Int J Med Sci, 2014, 11(12): 1208-1217. DOI:10.7150/ijms.9866
[28]
Bruyere A, Decleves X, Bouzom F, et al. Effect of variations in the amounts of P-glycoprotein (ABCB1), BCRP (ABCG2) and CYP3A4 along the human small intestine on PBPK models for predicting intestinal first pass[J]. Mol Pharm, 2010, 7(5): 1596-1607. DOI:10.1021/mp100015x
[29]
Xia H, Liu Z, Hu H, et al. Identification of ginkgolic acid (15 : 1) metabolites in rats following oral administration by high-performance liquid chromatography coupled to tandem mass spectrometry[J]. Xenobiotica, 2013, 43(5): 454-460. DOI:10.3109/00498254.2012.725141
[30]
Qian Y, Zhu Z, Duan JA, et al. Simultaneous quantification and semi-quantification of ginkgolic acids and their metabolites in rat plasma by UHPLC-LTQ-Orbitrap-MS and its application to pharmacokinetics study[J]. J Chromatogr B Analyt Technol Biomed Life Sci, 2017, 1041-1042: 85-93. DOI:10.1016/j.jchromb.2016.11.038
[31]
Liu ZH, Chen J, Yu LS, et al. Structural elucidation of metabolites of ginkgolic acid in rat liver microsomes by ultra-performance liquid chromatography/electrospray ionization tandem mass spectrometry and hydrogen/deuterium exchange[J]. Rapid Commun Mass Spectrom, 2009, 23(13): 1899-1906. DOI:10.1002/rcm.v23:13
[32]
Metz RP, Auyeung DJ, Kessler FK, et al. Involvement of hepatocyte nuclear factor 1 in the regulation of the UDP-glucuronosyltransferase 1A7 (UGT1A7) gene in rat hepatocytes[J]. Mol Pharmacol, 2000, 58(2): 319-327. DOI:10.1124/mol.58.2.319
[33]
Richert L, Binda D, Hamilton G, et al. Evaluation of the effect of culture configuration on morphology, survival time, antioxidant status and metabolic capacities of cultured rat hepatocytes[J]. Toxicol In Vitro, 2002, 16(1): 89-99. DOI:10.1016/S0887-2333(01)00099-6
[34]
Yang XM, Wang YF, Li YY, et al. Thermal stability of ginkgolic acids from Ginkgo biloba and the effects of ginkgol C 17 : 1 on the apoptosis and migration of SMMC7721 cells[J]. Fitoterapia, 2014, 98: 66-76. DOI:10.1016/j.fitote.2014.07.003
[35]
Ryckewaert L, Sacconnay L, Carrupt PA, et al. Non-specific SIRT inhibition as a mechanism for the cytotoxicity of ginkgolic acids and urushiols[J]. Toxicol Lett, 2014, 229(2): 374-380. DOI:10.1016/j.toxlet.2014.07.002