Glycycoumarin (purity>98%, Cat# DSTDG010801) was used as a reference compound and was purchased from Lemeitian Medical Technology Co., Ltd. (Chengdu, China). Psoralidin (Cat# 180526), used as an internal standard (IS), was obtained from Ronghe Pharmaceutical Sciences (Shanghai, China). β-Glucuronidase (Type H-1, Cat# SLCH4420) was obtained from Sigma (St. Louis, MO, United States). Solutol HS-15 (Cat# 159186) was obtained from MedChemExpress LLC (Shanghai, China). Anhydrous sodium acetate (Cat# C12095781) was purchased from Macklin Biochemical Technology Co., Ltd. (Shanghai, China). HPLC-grade acetonitrile and methanol were purchased from Fisher Scientific (Emerson, IA, United States). MCI GEL was purchased from Mitsubishi Chemical (Tokyo, Japan). Silica gel was obtained from Qingdao Haiyang Chemical Co., Ltd. (Qingdao, China). Analytical-grade ethyl acetate, methanol, and dichloromethane were obtained from Fuyu Fine Chemical Co., Ltd. (Tianjin, China).

The roots and rhizomes of licorice (G. uralensis Fisch.) were obtained from the Institute of Medicinal Plant Development, Chinese Academy of Medical Sciences, and Peking Union Medical College in November 2018 in Beijing. It was identified by Associate Professor Cha Qin from the Institute of Chinese Materia Medica of Bijie City. A voucher specimen (No. 2018-1101) was deposited at the Key Laboratory of Pharmacy of Bijie City First People’s Hospital, Guizhou, China.

Licorice (21 kg) was powdered and extracted thrice with 90% (v/v) ethanol for 2 h under reflux. The extracts were pooled and filtered, concentrated in vacuo to remove ethanol, and free-dried to obtain a dry licorice alcohol extract (3.6 kg). The licorice extract was dissolved in distilled water and extracted using ethyl acetate. The resulting ethyl acetate extract (860 g) was applied to a silica gel column and successively eluted with 0, 5, 10, 30, 50, and 100% (v/v) methanol in dichloromethane. The eluents from the 5% methanol application were collected and subjected to MCI GEL column chromatography for decolorization. The extracted compounds were then separated by repeated column chromatography and preparative high-performance liquid chromatography (HPLC) to isolate GCM. The chemical structure of the isolated GCM was identified by comparing the mass and fragmentation profiles obtained by electrospray ionization mass spectrometry (ESI-MS), as well as nuclear magnetic resonance (¹H- and ¹³C-NMR) spectra (Bruker AV-600, Fällanden, Switzerland) with those available in the literature. We also compared the GCM HPLC retention times and online UV spectra with those of the reference compound. The purity of GCM was assessed using HPLC-UV analysis.

Male Sprague-Dawley rats (200 ± 20 g) were obtained from SJA Laboratory Animal Co. Ltd. (Hunan, China). All animals were kept under standard conditions of light, humidity, and temperature, and were allowed free access to a standard diet and water. All animal experiments were approved by the Animal Ethics Committee of Bijie City First People’s Hospital (BY[2019]-02-198). On the day before GCM administration, rats were subjected to a minor surgical procedure. Briefly, a polyurethane catheter (0.59 mm ID, 0.94 mm OD; Skillsmodel Limited, Beijing, China) was cannulated into the right jugular vein while under anesthesia induced using an intraperitoneal dose of chloral hydrate at 350 mg/kg. After surgery, the rats were individually placed in metabolism cages to allow recovery for at least 24 h. The rats were fasted overnight with free access to water before GCM administration.

GCM (2.0 mg) and IS (2.0 mg) were accurately weighed and then dissolved in methanol to prepare respective stock solutions with a concentration of 2 mg/mL each. After dilution with methanol, the stock solution was prepared as a series of standard working solutions. The working solutions were used to spike the plasma, bile, urine, or tissue homogenates to obtain final GCM concentrations of 1, 2, 5, 10, 20, 50, 100, 200, 500, 1000, and 2000 ng/mL for the standard curve of all samples. The final concentration of IS used in the plasma, bile, urine, and tissue homogenates was 100 ng/mL. Four concentrations (5, 15, 150, and 1500 ng/mL) of quality control (QC) samples were freshly prepared for method validation.

Free (parent drug) and/or total (including parent and glucuronidated metabolites) GCM concentrations in rat plasma, tissues, urine, and bile were quantitatively assayed using LC-MS/MS before and after hydrolytic treatment with β-glucuronidase.

The stability of GCM in the gastrointestinal tract, including chemical stability and enzymatic stability, was assessed using an in vitro method as described previously (Kong et al., 2013). Briefly, GCM was spiked into different pH (1.2, 6.8, and 7.4) buffers and gastrointestinal content solutions containing the GCM at a final concentration of 0.8 μg/mL. The solutions were then mixed and incubated at 37°C. Aliquots of 200 μL samples were collected at 0.5, 1, 2, and 4 h, and 200 μL ice-cold methanol containing 200 ng/mL IS was added to stop the incubation. After centrifugation, 10 μL of the supernatant was used for analysis. Each experiment was performed in triplicate.

An LC-MS/MS system comprising an Agilent 1260 HPLC system (Palo Alto, CA, United States) and an Applied Biosystems 4500 Q-Trap mass spectrometer equipped with an electrospray ionization source (Foster City, CA, United States) was used. Chromatographic separation was performed on a C18 column (100 × 4.6 mm, 3 μm, Thermo Fisher Scientific, Waltham, MA, United States) maintained at 30°C. The mobile phases comprised 0.1% (v/v) formic acid in water (A) and acetonitrile (B) at a flow rate of 0.4 mL/min. Phase B was linearly increased from 40% to 90% over a period of 0.1 min, maintained at 90% for 2 min, and then decreased to 40% for re-equilibration. The total run time was 4 min. For mass spectrometry detection, the positive ion mode was used and the ion spray voltage was set at 5,500 V. The operating conditions were as follows: ion source temperature, 450°C; curtain gas, 20 psi; collision gas, medium. Ion source gases 1 and 2 were both at 60 psi. The collision energies were 37 and 42 V, and the declustering potentials for GCM and IS were 110 and 136 V, respectively. Quantification was performed by multiple reaction monitoring (MRM) of molecular ions and product ions at m/z 369.3 and 284.7 for GCM and 337.0 and 280.9 for IS (Supplementary Figure S1). Analyst^® software (Version 1.6.3) was used for data acquisition and peak integration.

Pharmacokinetic parameters were calculated based on changes in the plasma drug concentration versus the time profile of each rat using a non-compartmental model using WinNonlin software (Pharsight Corporation, Mountain View, CA, United States, Version 6.1). Plasma maximum concentration (Cmax) and the time of Cmax (Tmax); initial plasma concentration (C₀) for i. v. administration; terminal elimination half-life (t_{1/2, λz}); area under the plasma concentration versus time curve from time zero to the last sample collection time point (AUC_0–t) and to infinity (AUC_0–∞); total body clearance (CL); and volume of distribution (V_{d, λz}). Absolute bioavailability (F) was calculated based on the AUC_{0–∞, oral}/AUC_{0–∞, i. v.} at equivalent doses.

After several purification steps (silica gel column, MCI GEL™ column, and reverse-phase HPLC), pure GCM was isolated. Its chemical structure was characterized based on the high-resolution ESI-MS (HRESI-MS), ¹H-NMR, and ¹³C-NMR spectra (Supplementary Figures S2–S5, Supplementary Material) in accordance with the literature (Ji et al., 2016). Furthermore, the HPLC retention time and online UV spectrum of the isolated GCM were the same as those of the reference standard. The purity of the isolated GCM was determined to be 97.4% using HPLC-UV analysis.

Before the pharmacokinetic studies, a LC-MS/MS method was established and validated for the quantitative analysis of free and total GCM concentrations in rat plasma, urine, bile, and tissues.

To better understand the oral absorption and disposal process of GCM in the body, free and total plasma GCM concentrations (the parent form and glucuronidated metabolites) were determined. The plasma concentration versus time profiles of free and/or total GCM in rats after receiving an i. v. dose (10 mg/kg) or oral administration (20 mg/kg) are shown in Figure 3. The pharmacokinetic parameters of the free and/or total GCM are presented in Table 1. After i. v. dosing, only a minor quantity of parent GCM was detected in plasma samples after the first 6 h post-administration; it was almost undetectable in plasma collected after 6 h. It might be transformed into its glucuronidated metabolites because relatively high GCM concentrations were detected in the plasma after enzymatic treatment with β-glucuronidase. The results showed that the AUC_0-∞ of conjugated GCM was approximately four times higher than that of free GCM. This implies that glucuronidated metabolites are the major form of GCM in rats. The glucuronidated metabolites were eliminated slowly from the body, with average t_{1/2, λz} values of 2.46 h after i. v. administration, more than 3 folds those of free GCM.

After oral administration of 20 mg/kg, free GCM was almost undetectable in the plasma after 2 h. However, after treatment with β-glucuronidase, GCM was detected in the plasma over 24 h after dosing. Total GCM reached a C_max of 232.18 ng/mL at a T_max of 0.79 h, and was eliminated from the body with a t_{1/2, λz} of 4.75 h; the AUC_0-∞ was 1,017.85 ng h/mL.The absolute bioavailability of GCM was not determined because the concentration of free GCM after oral administration was too low to be detected at most time points. The bioavailability of total GCM was estimated to be 9.22%, which was calculated by an AUC_0-∞ of total GCM. This result suggested that a certain amount of GCM could be absorbed into the body and then quickly metabolized into glucuronidated metabolites via intestinal and/or hepatic first-pass effects.

GCM is reported to be rapidly absorbed into circulation and is involved in Phase I and Phase II biotransformation; further, GCM glucuronidation and hydroxylation are the main metabolic pathways involved (Wang et al., 2014). The isoprenyl methyl group is the major metabolic participant in hydroxylation reactions. The glucuronidation products include the parent GCM and its hydroxyl metabolites conjugated to glucuronic acid. However, in this study, our developed method could only monitor GCM but not detect other metabolites. Cytochrome P450 enzymes are also reported to be involved in GCM metabolism because many GCM metabolites have been demonstrated to be catalyzed in rat liver microsomes (Wang et al., 2014). Thus, the most likely reason for the low oral bioavailability of GCM might be related to its extensive first-pass effect when transiting the liver and/or intestine, which deserves further study.

In addition to hepatic metabolism, chemical stability, and drug metabolism in the intestine are considered important factors for low oral bioavailability (Crozier et al., 2009; Stalmach et al., 2012). To further explore whether the low bioavailability of GCM was related to intestinal metabolism, GCM was investigated in vitro using different pH buffer solutions and in gastrointestinal content samples. The results indicated that GCM was relatively stable in these tested solutions (pH 1.2, 6.8, and 7.4) and in stomach content solutions after 4 h of incubation at 37°C, but was unstable in the small intestine and colon content solutions (Figure 4). This result is consistent with reports on polyphones, which undergo rapid and extensive first-pass effects via conjugation in the intestinal tract and are poorly bioavailable (Stalmach et al., 2012; Ma et al., 2014; Wang et al., 2014; Wang et al., 2015). Thus, these results indicate the intestinal tract and hepatic first-pass metabolism are the main reasons for low GCM bioavailability.

The tissue distribution profile of drugs or compounds in vivo is closely related to their pharmacological effects, which determine the strength and duration of these effects. Further, understanding the characteristics of tissue distribution of drugs or compounds is helpful to further understand the sites and extent of their accumulation in vivo (Yuan et al., 2015). This will be beneficial in predicting adverse drug reactions, which are significant for providing guidance, especially for certain drugs or compounds metabolized in the body.

Our previous studies have shown that intraperitoneal injection of GCM exerts strong hepatoprotective and anti-liver cancer activity (Yan et al., 2018; Zhang et al., 2018). We thus questioned whether GCM is well-distributed in the liver tissue. Combining previous modes of drug administration (Song et al., 2015; Song et al., 2016; Yan et al., 2018), reducing the influence of individual absorption differences in animals, and directly observing the distribution of GCM in rats, intraperitoneal administration was used in the experiments reported here. After intraperitoneal administration of 20 mg/kg, GCM was widely distributed in all tissues, except the brain, as shown in Figure 5. Despite being a lipophilic polyhydroxy compound, GCM was not detected in brain samples. These results indicate that GCM cannot cross the blood-brain barrier. The highest GCM concentration was observed in the liver, followed by that in the spleen, lung, heart, and kidney; the concentration in the liver increased during 0.5–1 h, and then decreased. However, the GCM levels significantly decreased to undetectable levels by 6 h after dosing in most tissues. The highest concentration detected in the liver implied that the liver may be the main binding site for GCM, and that hepatic clearance may represent the primary route for GCM elimination. This may also help explain the hepatic first-pass effect of GCM. The concentration of GCM in plasma decreased rapidly after 0.5 h, which is in agreement with the results of the pharmacokinetic studies. Compared with the plasma concentrations, GCM can be quantified in various tissue samples, meaning that it has a relatively high degree of tissue distribution. Overall, these results provide an important material basis for the pharmacological effects of GCM.

The total amount of a drug in the circulatory system can be calculated after i. v. administration, which can yield the precise excretion rate of a parent drug in bile and urine. However, oral administration cannot result in this because of individual differences in drug absorption. Therefore, we performed i. v. administration to explore the excretion rate of GCM. The urinary and biliary recoveries of free and total GCM after i. v. administration are displayed in Table 1. The results indicated that a small amount of GCM was excreted as glucuronidated metabolites from bile with approximately 1.97%, 1.71%, 1.51%, and 1.33% of the dose administered over the periods of 0–2, 2–4, 4–6 and 6–8 h post-dosing, respectively. Furthermore, very low amounts of GCM in the parent form were excreted from bile: <1% of the delivered dose over 0–8 h. The urinary recovery of GCM over 24 h after dosing was 0.39% and 0.97% of the administered dose for free and total GCM, respectively. In the study of another polyphenol compound, a large number of sulfonated metabolites can also be excreted through bile and urine in addition to glucuronidated metabolites (Chang et al., 2012). This suggests that biliary and urinary excretion represent minor pathways for glucuronidated GCM metabolites.

The liver is rich in a large number of phase I and II metabolic enzymes, including CYP450, UDP-glucuronosyltransferases (UGTs), and sulfotransferases, which are the main sites of drug metabolism (Zhang et al., 2020). Our results show that a large amount of GCM is distributed in the liver and further metabolized in the liver. In addition to phase I metabolism, parent GCM and its phase I metabolites can undergo glucuronidation reactions by UGT enzymes because there is an abundance of UGT enzymes including UGT1A1, UGT1A3, UGT1A4, UGT1A6, UGT1A9, UGT1A19, and UGT2B7 in the liver (Wang et al., 2014). More recently, GCM was found to strongly inhibit P450 family members, including CYP1A2, CYP2B6, CYP2C8, CYP2C9, and CYP2C19, as well as UGT1A19 in human liver microsomes (Li et al., 2017; Kuang et al., 2021). Furthermore, GCM also significantly inhibits CYP2D6 activity, which mediates the metabolism of approximately 30% of the drugs on the market (Qiao et al., 2014a). Glycyrol, with a similar structure to that of GCM, is another major bioactive coumarin in licorice, as well as being a metabolite of GCM, and is not absorbed into the blood after oral administration (Wang et al., 2015). However, glycyrol also displays competitive inhibitory effects on CYP1A1 and CYP2C9 in human liver microsomes(Kim et al., 2016). In summary, the above results suggest that GCM may cause potential drug interactions when co-administered with agents metabolized by such enzymes. Therefore, as a lead compound with developmental value, the mechanisms of GCM absorption and its inhibitory effects on P450 should be investigated further.