RTS and GA (G10105, 97%) were purchased from Sigma-Aldrich Chemical Co. (St. Louis, MO, United States). Cocktail substrate and metabolite standards, NADPH tetrasodium salt, glucose-6-phosphate dehydrogenase (G-6-PD), glutathione (GSH), and all other chemicals, unless indicated, were purchased from Sigma-Aldrich Chemical Co. (St. Louis, MO, United States). Pyrrole-GSH conjugate, 7,9-diGSH-(±)-6,7-dihydro-7-hydroxy-1-hydroxymethyl-5H-pyrrolizine (7,9-diGS-DHP), was prepared as previously reported (Lin et al., 2000; Ma et al., 2015). HPLC-grade ethanol, acetone, formic acid, and acetonitrile were purchased from Merck (Darmstadt, Germany). Anti-CYP3A1 antibody was purchased from Santa Cruz Biotechnology (sc-53246, Santa Cruz, CA, United States). Horseradish peroxidase-conjugated goat anti-rabbit and goat anti-mouse were purchased from Cell Signaling Technology Inc. (Danvers, MA, United States). Anti-RECA-1 antibody (ab197727) and Alexa Fluor 488-conjugated goat anti-mouse IgG (ab150117) were supplied by Abcam (Cambridge, MA, United States).

Liquorice aqueous extract prepared from the root of Glycyrrhizae glabra L. (No. LC151018) was collected from Kazakhstan in Oct. 2015 and dried at 55°C for 12 h, and then authenticated by the co-author Prof. Yang YE’s group. The dried root of liquorice was extracted three times by refluxing with water (1:5 w/v) for 2 h per time. The water extract was then combined and lyophilized to provide liquorice extract (EX). EX was accurately weighed, completely dissolved in distilled water, and analyzed by HPLC-UV. The HPLC-UV analysis was performed on an Agilent 1100 liquid chromatograph system (Agilent Technologies, Inc., Palo Alto, United States). The chromatographic separation was performed on an Agilent C18 (4.6 × 250 mm, Shodex, Tokyo, Japan) column. The mobile phase was a mixture of water containing 1% formic acid (A) and acetonitrile (B), with a gradient elution as follows: 0–15 min, 75% A; 15–25 min, 25% A; 25–35 min, 75% A. The flow rate was set at 1 ml/min. The detection wavelength was 254 nm, and the injection volume was 20 µL. A representative HPLC-UV chromatogram of EX is shown in Supplementary Figure S1. The obtained EX was stored at -20°C before use.

Male Sprague Dawley rats (about 200–220 g) were supplied by the Laboratory Animal Service Centre at the Chinese University of Hong Kong. The animal room was maintained at 25 ± 1°C with a 12-h light-dark cycle and 55 ± 5% humidity. The rats had free access to standard rodent chow and water. All procedures were approved by the Animal Experimental Ethics Committee, the Chinese University of Hong Kong under the regulations of Hong Kong SAR government.

Rats were randomly divided into the following twelve groups: 1 and 2) Single- or Multiple-dose of distilled water groups (Vehicle), 3 and 4) Single-dose of RTS groups (RTS), 5 and 6) Single- or Multiple-dose of EX groups (SD-EX and MD-EX), 7 and 8) Single- or Multiple-dose of GA groups (SD-GA and MD-GA), 9) Single-dose of EX + RTS group (RTS-SD-EX), 10) Multiple-dose of EX + RTS group (RTS-MD-EX), 11) Single-dose of GA + RTS group (RTS-SD-GA), and 12) Multiple-dose of GA + RTS group (RTS-MD-GA). EX and GA were dissolved in distilled water at 37°C for oral administration with the dosage of EX and GA as 500 mg/kg and 50 mg/kg. RTS was dissolved in distilled water at room temperature to make the final concentration of 40 mg/kg. RTS was given following the last dose of distilled water/GA/EX. Detailed dosage regimens are described in Supplementary Figure S2. All rats were sacrificed at 48 h after the last dosing. Serum and liver samples were collected and stored at −80°C until use. The dose of RTS chosen for this study was based on our previous findings in that 40 mg/kg of RTS caused significant acute liver injury with moderate to severe SECs damage in male SD rats (Chen et al., 2021). According to the intake of the Compound Glycyrrhizin Tablets (containing 25 mg GL/per tablet), the suggested daily intake of GL is 225 mg/day for humans, which is approximately 20 mg/kg/day GA for rats. In our experiments, we used 50 mg/kg/day GA and 500 mg/kg/day EX (suggested daily intake of 5–20 g of liquorice root) for 5 consecutive days, which is close to the normal dose and has also been reported in many publications (Lin et al., 1999; Zhang et al., 2019; Yan et al., 2021).

For the toxicokinetic study, jugular vein cannulation was performed on rats on the day before drug administration. The right jugular vein was cannulated with a polyethylene tube (0.4 mm i.d. × 0.8 mm o.d., SIMS Portex, United Kingdom) for blood sampling. The rats were maintained in individual metabolic cages and fasted overnight with free access to water. Rats were orally dosed with GA (50 mg/kg) or distilled water (<5 ml/kg) 2 hours prior to RTS (30 mg/kg). After RTS dosing, blood samples (about 0.2 ml/sample) were collected at 5, 10, 20, 40, 60, 120, 240, 360, 480, 720, and 1440 min and placed in heparinized tubes. Saline (0.2 ml) containing 25 units of heparin/mL was injected after each blood sampling for compensation of blood withdrawal. The blood samples were then centrifuged at 3,000 x g for 30 min, and plasma samples were harvested and stored at −20°C until use.

The largest lobe of the liver was sliced, fixed in 10% buffered-neutral formalin for 24 h, and embedded in wax. Sections of 5 µm in thickness were subjected to hematoxylin and eosin staining before being examined. Liver injury (necrosis, endothelial cell damage, and hemorrhage) was evaluated with a LEICA DM5000B Microscope (Leica, Heidelberg, Germany).

The serum alanine transaminase (ALT) level was measured by a kit obtained from Sigma-Aldrich Chemical Co. (St. Louis, MO, United States). The total bilirubin level was measured by the QuantiChromTM Bilirubin Assay kit according to the manufacturer’s instructions (BioAssay Systems, Hayward, CA, United States). Hepatic malondialdehyde (MDA) level was measured using a kit obtained from Nanjing Jiancheng Bioengineering Institute (Nanjing, China). Total GSH content in the liver was measured following a standard spectrophotometric method using Ellaman’s reagent.

The largest lobe of the rat liver was sliced, cold-embedded in Tissue-Tek OCT compound, and processed for immunohistochemistry according to the Abcam IHC frozen sections staining protocol. The tissue slices were incubated with rat endothelial cell antigen (RECA)-1 antibody and Alexa Fluor 488-conjugated goat anti-mouse IgG, and then stained with DAPI and observed under Olympus FluoViewTM FV1000 confocal microscope.

The pyrrole-protein adducts were measured according to our previously developed method (Lin et al., 2011; Ruan et al., 2014). Briefly, 50 mg liver or 100 µL serum/plasma was mixed with acetone and centrifuged at 900 x g for 10 min. The precipitated protein was treated with silver nitrate (20 mg/ml) in ethanol containing 5% trifluoroacetic acid to release pyrrole moiety from the pyrrole-protein adducts. After centrifugation, an aliquot of the supernatant was incubated with 4-dimethylaminobenzaldehyde (20 mg/ml) at 56°C for 10 min. The mixture was filtered and subjected to LC-MS/MS analysis. 7,9-diGS-DHP was used as a standard with a concentration ranging from 0.2 to 2000 nM and underwent the same preparation process to construct a calibration curve. To measure the concentration of RTS and 7,9-diGS-DHP, liver, serum, plasma samples, or the incubation mixture were mixed with five volumes of acetone (for liver samples) or three volumes of acetonitrile (for serum and plasma samples) to precipitate the proteins. After centrifuging at 20,000 x g for 30 min, the supernatants were collected and analyzed by LC-MS/MS according to our previously developed methods (Ma et al., 2019; He et al., 2021a; Ma et al., 2021).

To determine the effect of GA on the formation of pyrrole-GSH conjugate (7,9-diGS-DHP), GA (0, 2.5, 5, 10, 30, 50, and 80 μM, prepared in DMSO) or ketoconazole (KCZ, 25 μM, prepared in DMSO) was incubated with 200 µM RTS in the incubation mixture containing 1 mg/ml rat liver microsomes, 1 mM NADP⁺, 10 mM G-6-P, 5 mM MgCl₂, 2 mM GSH, 1 U/ml G-6-PD and potassium phosphate buffer (0.1 M, pH 7.4). The reaction was initiated by adding RTS and stopped with the addition of ice-cold acetonitrile after 120 min incubation at 37°C. The mixtures were centrifuged, and supernatants were analyzed by LC-MS/MS.

A reported cocktail assay (Otten et al., 2011) was used for evaluating the effect of GA on different CYP activities by LC-MS/MS with high sensitivity and selectivity. The concentrations of probe substrates were listed in Supplementary Table S1. Various concentrations of GA (0, 2.5, 5, 10, 30, 50, and 80 µM) were added to the reaction mixture containing mixed CYPs probe substrates, NAPDH-regenerating system (5 mM MgCl₂, 1 mM NADP⁺, 1 mM G-6-PD), and 1 mg/ml rat liver microsomes.

The kinetic study was performed to determine the inhibitory mechanism of GA on rat CYP3A1. Briefly, GA (0, 10, 25 and 50 µM) was incubated in 200 µL incubation system containing 1 mg/ml rat liver microsomes, 5 mM MgCl₂, 1 mM NADP⁺, 1 mM G-6-PD, and nifedipine (0.1, 0.5, 1, 2, 5, 10, 20, 50, 100, and 200 µM). The mixture was pre-incubated at 37°C for 15 min. The reaction was initiated by adding nifedipine for 45 min incubation at 37°C, and then quenched with ice-cold acetonitrile and centrifuged at 15,000 x g for 20 min. The supernatant was filtered through a 0.22 µm membrane filter before the LC-MS/MS analysis.

The LC-MS/MS analysis was performed on an Agilent 6460 Triple Quadrupole LC/MS System using a Waters Acquity BEH C18 column (2.1 × 100 mm, 1.7 mm). For the detection of pyrrole-protein adducts, the mobile phase of water containing 0.1% formic acid (A) and acetonitrile containing 0.1% formic acid (B) was used with a gradient elution as follows: 0–5 min, 35–95% B; 5.5–6 min, 95–35% B. The flow rate was 0.3 ml/min. The injection volume was 3 μL. For the detection of pyrrole-GSH conjugates and RTS, a gradient elution was used as follows: 0–1 min, 2% B; 1–7 min, 2–25% B, 7–7.5 min, 25–95% B. 7.5–10 min, 95% B. The flow rate was 0.3 ml/min. The injection volume was 5 µL. The mass spectrometer was operated in multiple reactions monitoring for data acquisition. The multiple reaction monitoring (MRM) transitions, fragmentor, and collision energy for 7,9-diGS- RTS, RTS, and pyrrole-protein adducts are listed in Supplementary Table S1.

For the cocktail assay, a gradient elution was used as follows: 20% A (0–1 min), 20–95% A (1–4 min), 95% A (4–6 min). The MRM transitions and linearity information for each substrate are indicated in Supplementary Table S2. The calibration curves of individual CYP substrates and their corresponding metabolites were constructed by plotting the peak area versus the spiked concentration.

Proteins were separated by gel electrophoresis and electrophoretically transferred to nitrocellulose paper. The nitrocellulose paper was then incubated with rat CYP3A1 antibody (1:1000) overnight and then reacted with horseradish peroxidase-conjugated secondary antibody. Bands were developed using an ECL chemiluminescence detection kit. Equal loading of proteins was verified by GAPDH. The semi-quantification of each band was measured by ImageJ software.

Data are expressed as mean ± SD. The Student’s t test was used for the comparison between two groups. The one-way ANOVA followed with a post hoc Bonferroni’s multiple comparison test was used for the comparison among multiple groups. Toxicokinetic parameters were calculated by noncompartmental methods using WinNonlin version 4.0 (Pharsight, Mountain View, CA, United States). The enzyme kinetic parameters were estimated from the best fit line using least-squares linear regression of the inverse substrate concentration versus the inverse velocity (Lineweaver-Burk plots), and the mean values were used to calculate the K_m and V_max. The inhibition constant (K _i ) value of GA was determined by the secondary Lineweaver-Burk plots. The statistical significance was set as ^* p < 0.05, ^** p < 0.01 or ^*** p < 0.001.

Compared with the vehicle groups, serum ALT and hepatic MDA levels were unchanged after both single and multiple administrations of EX and GA but elevated significantly at 48 h after RTS treatment. While a single dose of GA significantly reduced RTS-induced ALT and MDA elevations (Figures 2A,C), the multiple doses of both EX and GA also significantly attenuated these RTS-induced elevations (Figures 2B,D). In addition, hepatic GSH level increased significantly (2.82-fold, p < 0.001) at 48 h after RTS dosing as a feedback response to the activation of the antioxidant system, while this elevation was attenuated by the treatment of both EX and GA, especially their multiple dosage regimens (Figures 2E,F).

Protective effects of EX and GA on RTS-induced hepatotoxicity were also evaluated by histopathological examination. No obvious liver damage was observed in rats treated with EX or GA alone (Figures 3A,B). A single dose of RTS caused severe hepatotoxicity in rats, as demonstrated by massive hemorrhage and hepatocytes necrosis observed in the liver sections, whereas both single and multiple doses of EX or GA significantly ameliorated RTS-induced hemorrhage and necrosis (Figure 3A). The damage of sinusoidal endothelial cells was further measured by the immunofluorescence of RECA-1. Sinusoids, tightly distributed around central veins, were markedly damaged by RTS, and the injured sinusoids merged to form a large space with wide gaps among individual sinusoids, while this damage was significantly less in EX or GA treated groups (Figure 3B). The degrees of hemorrhage, sinusoidal dilation, and necrosis were assessed and scored (Table 2) according to a modified DeLeve’s system (Deleve et al., 1999) as summarized in Supplementary Table S3. All these biochemical and histopathological results clearly demonstrated that both EX and GA produced similar hepato-protective effects against PA-induced liver toxicity. Considering the individual hepato-protective effect of EX or GA multiple treatments exhibited a better protective effect than the single treatment scenario.

The formation of pyrrole-protein adducts is a critical initiating event of RTS-induced hepatotoxicity. As shown in Figure 4, the formation of pyrrole-protein adducts in the liver was significantly reduced by both single (Figure 4A) and multiple (Figure 4B) treatments with GA and EX to similar levels. Similarly, the serum pyrrole-protein adducts levels were also significantly reduced with the treatment of EX and GA in both single and multiple treatments to the same extent (Figures 4C,D). The following studies to delineate the mechanism underlying protective effect on RTS-induced hepatotoxicity were carried out with GA that is reported as the primary bioactive form in vivo (Öztürk et al., 2017).

For the study of the effect of GA on the toxicokinetic profile of RTS, rats treated with a single dose of GA and RTS exhibited significantly higher plasma C _max (4806.31 ± 133.74 ng/ml vs. 3188.07 ± 144.65 ng/ml, p < 0.001) and AUC_0–1440min (341.52 ± 7.57 min×µg/mL vs. 234.49 ± 34.07 min×µg/mL, p < 0.05) of RTS compared to the group treated with RTS alone (Figure 5A). In addition, the kinetic profile of pyrrole-protein adducts revealed an inhibitory effect of GA on the formation of pyrrole-protein adducts (Supplementary Table S4). In RTS + GA group, GA significantly reduced AUC_0–1440min of pyrrole-protein adducts to 856.49 ± 11.05 min × µM (16.2% reduction) compared to that (1022.49 ± 11.05 min × µM) in RTS group (Figure 5B). The results indicated that GA significantly inhibited the metabolic activation of RTS, leading to the reduced formation of pyrrole-protein adducts in rats.

Furthermore, the in vitro incubation with rat liver microsomes was also performed to confirm the inhibitory effect of GA on the metabolic activation of RTS. The dehydro-RTS, produced from metabolic activation of RTS by rat liver microsomes, was trapped by GSH to form pyrrole-GSH conjugate (7,9-diGS-DHP), since our previous study demonstrated that pyrrole mono-GSH adducts could not be detected in the blood of rats dosed with RTS (Lian W. and Lin G. 2017). The formation of 7,9-diGS-DHP after co-incubation with GA or ketoconazole (KCZ, CYP3A1 inhibitor) was then detected. GA significantly inhibited the formation of 7,9-diGS-DHP in a concentration-dependent manner with the estimated IC₅₀ value of 5.07 µM, while in parallel, the concentration of intact RTS increased along with the concentration of GA (Figure 5C). In addition, the formation of 7,9-diGS-DHP was significantly reduced by 62% with the co-incubation of KCZ (25 µM) (Figure 5C). All the results confirmed that GA significantly inhibited metabolic activation of RTS, and thus reduced the production of toxic metabolite dehydro-RTS evidenced by the decreased formation of 7,9-diGS-DHP.

It is well-known that metabolic activation of PAs is mediated by hepatic CYPs, thus the effects of GA on the activity and protein expression of CYPs were determined by the cocktail assay. The results showed that GA inhibited the activities of rat CYPs, in particular CYP3A1, CYP2A2, and CYP2B2 (Figure 6). CYP3A1 (homologue to human CYP3A4), the predominant CYP isoform responsible for the metabolic activation of RTS, was inhibited by GA in a concentration-dependent manner with 68.5% inhibition at the highest concentration (80 µM) of GA tested (Figure 6A). For other tested CYPs which play minor roles in the metabolism of RTS (Ruan et al., 2014), the activities of CYP2A2 and CYP2B2 were also reduced by GA (Figures 6B,C), while the activities of CYP2C11 and CYP2D1 were not altered by GA (Figures 6D,E). Furthermore, the effect of GA on protein expression of CYP3A1 was tested in rats, and the results revealed that multiple oral administration of GA once a day for 5 consecutive days did not affect the protein expression level of rat hepatic CYP3A1 (Figure 6F), further confirming that only the activity but not expression of CYP3A1 could be inhibited by GA.

To further explore the inhibitory mechanism of GA towards CYP3A1, an enzyme kinetic study was performed using nifedipine as the probe substrate for rat CYP3A1. The results from the Michaelis-Menten curve (Figure 7A) demonstrated that GA inhibited CYP3A1-mediated nifedipine metabolism with K _m value increased along with the increase in GA concentrations (Figure 7C). Furthermore, the Lineweaver-Burk plot (Figure 7B) indicated that GA concentration-dependently altered the slope but did not affect the Y-intercept, and the slope of each curve correlated well with the concentration of GA (R ² = 0.9952) with K _i value of 8.52 µM (Figure 7B). All the results demonstrated that GA competitively inhibited the enzyme activity of CYP3A1.