Tetrahydrofurofuranoid Lignans, Eudesmin, Fargesin, Epimagnolin A

pharmaceutics

Article Tetrahydrofurofuranoid Lignans, Eudesmin, Fargesin, Epimagnolin A, Magnolin, and Yangambin Inhibit UDP-Glucuronosyltransferase 1A1 and 1A3 Activities in Human Liver Microsomes

Ria Park 1,†, Eun Jeong Park 1,†, Yong-Yeon Cho 1, Joo Young Lee 1 , Han Chang Kang 1 , Im-Sook Song 2 and Hye Suk Lee 1,*

1 College of Pharmacy, The Catholic University of Korea, Bucheon 14662, Korea; [email protected] (R.P.); [email protected] (E.J.P.); [email protected] (Y.-Y.C.); [email protected] (J.Y.L.); [email protected] (H.C.K.) 2 College of Pharmacy and Research Institute of Pharmaceutical Sciences, Kyungpook National University, Daegu 41566, Korea; [email protected] * Correspondence: [email protected]; Tel.: +82-2-2164-4061; Fax: +82-32-342-2013 † These authors contributed equally to this work.

Abstract: Eudesmin, fargesin, epimagnolin A, magnolin, and yangambin are tetrahydrofurofuranoid lignans with various pharmacological activities found in Magnoliae Flos. The inhibition potencies 0 of eudesmin, fargesin, epimagnolin A, magnolin, and yangambin on six major human uridine 5 - diphospho-glucuronosyltransferase (UGT) activities in human liver microsomes were evaluated Citation: Park, R.; Park, E.J.; Cho, using liquid chromatography–tandem mass spectrometry and cocktail substrates. Eudesmin, fargesin, Y.-Y.; Lee, J.Y.; Kang, H.C.; Song, I.-S.; epimagnolin A, magnolin, and yangambin inhibited UGT1A1 and UGT1A3 activities, but showed Lee, H.S. Tetrahydrofurofuranoid negligible inhibition of UGT1A4, UGT16, UGT1A9, and UGT2B7 activities at 200 µM in pooled Lignans, Eudesmin, Fargesin, human liver microsomes. Moreover, eudesmin, fargesin, epimagnolin A, magnolin, and yangambin Epimagnolin A, Magnolin, and Yangambin Inhibit noncompetitively inhibited UGT1A1-catalyzed SN38 glucuronidation with Ki values of 25.7, 25.3, 3.6, UDP-Glucuronosyltransferase 1A1 26.0, and 17.1 µM, respectively, based on kinetic analysis of UGT1A1 inhibition in pooled human and 1A3 Activities in Human Liver liver microsomes. Conversely, the aforementioned tetrahydrofurofuranoid lignans competitively Microsomes. Pharmaceutics 2021, 13, inhibited UGT1A3-catalyzed chenodeoxycholic acid 24-acyl-glucuronidation with 39.8, 24.3, 15.1, 187. https://doi.org/10.3390/ 37.6, and 66.8 µM, respectively in pooled human liver microsomes. These in vitro results suggest pharmaceutics13020187 the necessity of evaluating whether the ﬁve tetrahydrofurofuranoid lignans can cause drug–drug interactions with UGT1A1 and UGT1A3 substrates in vivo. Academic Editors: Charles M. Heard and Maria Rosaria Lauro Keywords: eudesmin; fargesin; epimagnolin A; magnolin; yangambin; human liver microsomes; Received: 6 December 2020 UDP-glucuronosyltransferase Accepted: 28 January 2021 Published: 1 February 2021

Publisher’s Note: MDPI stays neutral 1. Introduction with regard to jurisdictional claims in published maps and institutional affil- Eudesmin, fargesin, epimagnolin A, magnolin, and yangambin (Figure1) are the iations. pharmacologically active tetrahydrofurofuranoid lignans found in Magnolia denudata Desrousseaux, M. fargesii Cheng (family: Magnoliaceae), and Zanthoxylum armatum DC (family: Rutaceae) [1–5]. These lignans possess anti-inflammatory activity through the inhibitory effects on 5-lipoxygenase and nitric oxide synthase [5–9] and antitumor activity [10–12]. In addition to the anti-inflammatory and antitumor activity of tetrahydrofurofu- Copyright: © 2021 by the authors. Licensee MDPI, Basel, Switzerland. ranoid lignans, eudesmin showed bactericidal effect on Helicobacter pylori reference strain This article is an open access article 26,695 and suppressed inflammatory and immune responses in Helicobacter pylori-infected distributed under the terms and mice [13]. Fargesin exhibits additional biological activities, including cardioprotective ef- conditions of the Creative Commons fect in rats with ischemia/reperfusion injury [14] and lipid lowering and hypoglycemic Attribution (CC BY) license (https:// effects in high-fat diet-induced obese mice [15], and antihypertensive effects in 2K1C hy- creativecommons.org/licenses/by/ pertensive rats [16]. Magnolin also ameliorates contrast-induced nephropathy in rats via 4.0/). anti-oxidation and anti-apoptosis [17].

Pharmaceutics 2021, 13, 187. https://doi.org/10.3390/pharmaceutics13020187 https://www.mdpi.com/journal/pharmaceutics Pharmaceutics 2021, 13, 2 of 14

tective effect in rats with ischemia/reperfusion injury [14] and lipid lowering and hypoglycemic effects in high-fat diet-induced obese mice [15], and antihypertensive effects in 2K1C hypertensive rats [16]. Magnolin also ameliorates contrast-induced nephropathy in Pharmaceutics 2021, 13, 187 2 of 13 rats via anti-oxidation and anti-apoptosis [17].

FigureFigure 1. 1. ChemicalChemical structures structures of of ( (AA)) eudesmin, eudesmin, ( (BB)) fargesin, fargesin, ( (CC)) epimagnolin epimagnolin A, A, ( (D)) magnolin, magnolin, and ( E) yangambin.

TheThe ethanol ethanol extract extract of of the the dried dried flower flower buds buds of M. of fargesiiM. fargesii (encoded(encoded as NDC-052 as NDC-052 con- tainedcontained eudesmin, eudesmin, fargesin, fargesin, epimagnolin epimagnolin A, magnol A, magnolin,in, and yangambin and yangambin as 4.1, as 3.4, 4.1, 11.9, 3.4, 21.5, 11.9, and21.5, 9.1%, and respectively, 9.1%, respectively, as determined as determined by the LC-APCI-MS/MS by the LC-APCI-MS/MS method [18]) method has [been18]) hasde- been developed as an effective alternative or complement to standard asthma therapy veloped as an effective alternative or complement to standard asthma therapy based on based on their biological activities [19]. Consequently, taking NDC-052 (600 mg/day for their biological activities [19]. Consequently, taking NDC-052 (600 mg/day for 8 weeks per 8 weeks per oral) in adult asthmatic patients was safe and tolerated [19]. The add-on oral) in adult asthmatic patients was safe and tolerated [19]. The add-on therapy of NDC- therapy of NDC-052 (600 mg/day for 8 weeks per oral) with inhaled corticosteroids in 052 (600 mg/day for 8 weeks per oral) with inhaled corticosteroids in asthmatic patients asthmatic patients had a beneficial effect on asthma control [19]. The pharmacokinet- had a beneficial effect on asthma control [19]. The pharmacokinetics of herbal constituents ics of herbal constituents in rats have been investigated in addition to their biological in rats have been investigated in addition to their biological activities. Eudesmin, yangam- activities. Eudesmin, yangambin, epimagnolin A, fargesin, and magnolin were identi- bin, epimagnolin A, fargesin, and magnolin were identified in rat plasma following oral fied in rat plasma following oral administration (5.5–22 mg/kg) of NDC-052 to rats [18]. administration (5.5–22 mg/kg) of NDC-052 to rats [18]. The area under the plasma concen- The area under the plasma concentration curve (AUC) and maximum plasma concentration curve (AUC) and maximum plasma concentration (Cmax) values of eudesmin, tration (Cmax) values of eudesmin, yangambin, and epimagnolin linearly increased with yangambin, and epimagnolin linearly increased with dose increase [18]. However, AUC dose increase [18]. However, AUC and Cmax values of fargesin and magnolin were not andincreased Cmax values with dose of fargesin proportionality, and magnolin suggesting were thenot nonlinearincreased pharmacokineticwith dose proportionality, properties suggesting the nonlinear pharmacokinetic properties of fargesin and magnolin [18]. Ad- of fargesin and magnolin [18]. Additionally, the AUC and Cmax values of magnolin in ditionally,rats were 7568 the AUC± 1085 and ng C·h/mLmax valuesand 2493of magnolin± 513 ng/mL, in ratsrespectively, were 7568 ± when1085 ng·h/mL administered and 2493NDC-052 ± 513 (22.2 ng/mL, mg/kg respectively, containing when 4 mg/kg admi ofnistered magnolin) NDC-052 [20] and (22.2 had mg/kg similar containing results with 4 mg/kg of magnolin) [20] and had similar results with the previous report [18]. However, the previous report [18]. However, the AUC and Cmax values of magnolin in rats were the3630 AUC± 581 and ng C·h/mLmax valuesand 1340of magnolin± 113 ng/mL, in ratsrespectively, were 3630 ± when581 ng·h/mL administered and 1340 magnolin ± 113 ng/mL,alone (4 respectively, mg/kg) [21 ].when These administered results suggested magnol thein alone possibility (4 mg/kg) of drug [21]. interactions These results among sug- gestedthe constituents the possibility of herbal of drug drugs. interactions among the constituents of herbal drugs. TheThe herb-drugherb-drug interactionsinteractions among among the the herbal herbal drugs drugs (e.g., (e.g.,Hypericum Hypericum perforatum perforatum, Ginkgo, Ginkgobiloba, Camelliabiloba, Camellia sinensis sinensis, Glycyrrhiza, Glycyrrhiza glabra ,glabraAllium, Allium sativum sativum, Coptis, Coptis chinensis chinensis, and, Silybumand Si- lybummarianum marianum) and their) and constituents their constituents via the inhibitionvia the inhibition or induction or induction of major of drug-metabolizing major drug-me- tabolizingenzymes, cytochromeenzymes, cytochrome P450 (CYP) andP450 uridine (CYP) 50 -diphosphoand uridine (UDP)-glucuronosyltransferase 5′-diphospho (UDP)-glucuronosyltransferase(UGT) have been reported(UGT) have [22– been27]. Thus,reported evaluating [22–27]. theThus,in evaluating vitro inhibitory the in effects vitro inhib- of the itoryactive effects components of the active of herbal components drugs on of human herbal CYP drugs and on UGT human activities CYP and is necessary UGT activities for the isassessment necessary offor herb-drug the assessment interactions. of herb-drug interactions. Aschantin,Aschantin, a bioactive tetrahydrofurofuranoidtetrahydrofurofuranoid lignan lignan found found in inM. M. biondii biondiiand andHernan- Her- nandiadia nymphaeifolia nymphaeifolia, exhibited, exhibited the the time-dependent time-dependent inhibition inhibition of of CYP2C8 CYP2C8 ( K(Ki:i: 10.210.2 μµM and −1 −1 −1 −1 kkinactinact: 0.056: 0.056 min min), CYP2C9), CYP2C9 (Ki (: K3.7i: 3.7μMµ Mand and kinactkinact: 0.044: 0.044 min min), CYP2C19), CYP2C19 (Ki: 5.8 (K i:μ 5.8M andµM −1 −1 and kinact: 0.048 min ), and CYP3A4 (Ki: 12.6 µM and kinact: 0.062 min ) in human liver microsomes [28]. The inactivation efficiency (kinact/Ki) of aschantin on CYP2C8, CYP2C9, µ CYP2C19, and CYP3A4 activities was calculated as 5.49, 11.9, 8.28, 4.92 mL/ mol/min, respectively, which were in the range of the inactivation efficiency of clarithromycin (2.8 mL/µmol/min) and verapamil (11.2 mL/µmol/min) on CYP3A4, representative examples of clinical CYP3A4-mediated drug interactions [29]. Aschantin weakly inhibited Pharmaceutics 2021, 13, 187 3 of 13

the catalytic activities of UGT1A1 (IC50, 131.7 µM), UGT1A6 (IC50, 144.1 µM), and UGT1A9 (IC50, 71.0 µM) in human liver microsomes [28]. The inhibitory effects of yangambin, epimagnolin A, eudesmin, fargesin, and magnolin on CYP activities in human liver microsomes were previously investigated [30]. Fargesin inhibited 0 competitively CYP2C9-catalyzed diclofenac 4 -hydroxylation (Ki, 16.3 µM) [30]. It also exhib- −1 ited the time-dependent inhibition of CYP2C8 (Ki: 10.7 µM and kinact: 0.082 min ), CYP2C19 −1 −1 (Ki: 3.7 µM and kinact: 0.102 min ), and CYP3A4 (Ki: 23.0 µM and kinact: 0.050 min ) in human liver microsomes. However, fargesin negligibly inhibited the catalytic activities of CYP1A2, CYP2A6, CYP2B6, and CYP2D6. In addition, eudesmin, epimagnolin A, magnolin, and yangambin slightly inhibited eight major CYP activities in human liver microsomes, indicating that they may not cause CYP-mediated drug interactions [30]. However, no reports exist on in vitro and in vivo inhibitory effects of eudesmin, fargesin, epimagnolin A, magnolin, and yangambin on human UGT enzymes. Therefore, this study aimed to investigate the in vitro inhibitory potentials and inhibition kinetics of eudesmin, fargesin, epimagnolin A, magnolin, and yangambin, that are major tetrahydrofurofuranoid lignans contained in NDC-052 and also found in rat plasma samples [18], on UGT1A1, UGT1A3, UGT1A4, UGT1A6, UGT1A9, and UGT2B7 activities using pooled human liver microsomes to provide the underlying drug interaction potentials of eudesmin, fargesin, epimagnolin A, magnolin, and yangambin.

2. Materials and Methods 2.1. Materials and Reagents Fargesin (purity, 99.2%) and magnolin (purity, 98.9%) were obtained from PhytoLab GmbH & Co. (Vestenbergsgreuth, Germany). Eudesmin (purity, 99%) was purchased from Selleck Chemicals LLC. (Houston, TX, USA). Epimagnolin A (purity, 98%) and yangambin (purity, 98%) were obtained from ChemFaces Biochemical Co. Ltd. (Wuhan, China). Ul- trapooled human liver microsomes (catalog number 452161) were obtained from Corning Life Sciences (Woburn, MA, USA). Alamethicin, UDP-glucuronic acid (UDPGA), Trizma base, N-acetylserotonin, chenodeoxycholic acid, ketoconazole, naloxone, naloxone 3-β- D-glucuronide, mycophenolic acid, and triﬂuoperazine dihydrochloride were provided by Sigma-Aldrich (St. Louis, MO, USA). Chenodeoxycholic acid 24-acyl-β-glucuronide, SN-38 glucuronide, and mycophenolic acid β-D-glucuronide were provided by Toronto Research Chemicals (Toronto, ON, Canada). SN-38 was provided by Santa Cruz Biotech- nology (Dallas, TX, USA). Water, methanol, and acetonitrile of high-performance liquid chromatography grade were provided by Fischer Scientiﬁc (Fair Lawn, NJ, USA).

2.2. Inhibitory Potentials of Eudesmin, Fargesin, Epimagnolin A, Magnolin, and Yangambin on Six Major UGT Activities in Ultrapooled Human Liver Microsomes The inhibitory potentials of eudesmin, fargesin, epimagnolin A, magnolin, and yangambin on six UGT activities were measured as described previously [31]. Briefly, reaction mixture (100 µL) containing 50 mM Tris buffer (pH 7.4), 10 mM MgCl2, 25 µg/mL alamethicin, 5 mM UDPGA, ultrapooled human liver microsomes (0.2 mg/mL), various concentrations of eudesmin, fargesin, epimagnolin A, magnolin, or yangambin in acetonitrile (final concentrations of 0.1, 1, 10, 20, 50, 100, and 200 µM each, acetonitrile less than 1% (v/v), and the UGT probe cocktails of set A (0.5 µM SN-38 for UGT1A1, 2 µM chenodeoxycholic acid for UGT1A3, and 0.5 µM trifluoperazine for UGT1A4) or set B (1 µM N-acetylserotonin for UGT1A6, 0.2 µM mycophenolic acid for UGT1A9, and 1 µM naloxone for UGT2B7) were incubated at 37 ◦C for 60 min. Fifty microliters of meloxicam and propofol glucuronide (internal standards, ISs) in acetonitrile was added to the reaction tube, and the mixtures were centrifuged at 13,000× g for 4 min at 4 ◦C. Fifty microliters of the supernatants from set A and set B reaction tubes were mixed, and 3 µL-aliquot of the mixture was analyzed by LC-MS/MS. Pharmaceutics 2021, 13, 187 4 of 13

2.3. Kinetic Analysis The incubation mixtures (100 µL) including 50 mM Tris buffer (pH 7.4), human liver microsomes (0.1 mg/mL), various concentrations of SN-38 (0.2–2 µM) for UGT1A1 or chenodeoxycholic acid (1–10 µM) for UGT1A3, 5 mM UDPGA, 25 µg/mL alamethicin, 10 mM MgCl2, and various concentrations of eudesmin (2–40 µM), fargesin (2–30 µM), epimagnolin A (0.2–5 µM for UGT1A1; 2–30 µM for UGT1A3), magnolin (2–40 µM), or yangambin (2–40 µM for UGT1A1; 5–80 µM for UGT1A3) were incubated for 30 min to determine the Ki values and inhibition mode for inhibition of eudesmin, fargesin, epimagnolin A, magnolin, and yangambin on UGT1A1 and UGT1A3 activities. The inhibitory effect of NDC-052 on UGT1A1 and UGT1A3 activities was also measured by incubating the incubation mixtures (100 µL) including 50 mM Tris buffer (pH 7.4), human liver microsomes (0.1 mg/mL), SN-38 (0.5 µM) for UGT1A1 or chenodeoxycholic acid (2 µM) for UGT1A3, 5 mM UDPGA, 25 µg/mL alamethicin, 10 mM MgCl2, and various concentrations of ND-052 (0.1–100 µg/mL) for 30 min. The reactions were stopped by adding 100 µL of 500 ng/mL meloxicam (IS for SN-38 glucuronide) or propofol glucuronide (IS for chenodeoxycholic acid 24-acyl-β-glucuronide) in ice-cold acetonitrile. The incubation mixtures were then centrifuged at 13,000 × g for 5 min, and 30 µL of the supernatant was diluted with 70 µL of water. Aliquots (3 µL) were analyzed by LC-MS/MS.

2.4. LC-MS/MS Analysis UGT metabolites were simultaneously quantified using Agilent 6495 triple quadrupole mass spectrometer equipped with Agilent 1290 Infinity UPLC (Agilent Technologies, Wilm- ington, DE, USA) as described previously [31]. Analytes were eluted from an Atlantis dC18 column (3 µm, 2.1 mm internal diameter × 100 mm, Waters Co., Milford, MA, USA) by the gradient elution of mobile phase A (5% acetonitrile in 0.1% formic acid) and mobile phase B (95% acetonitrile in 0.1% formic acid) at flow rate of 0.3 mL/min. The electrospray ionization source settings in both positive and negative ion modes were drying gas flow, 14 L/min; nebulizer pressure, 40 psi; sheath gas flow, 11 L/min; drying gas temperature, 200 ◦C; sheath gas temperature, 380 ◦C; nozzle voltage, 500 V; capillary voltage, 4500 V; and fragmentor voltage, 380 V. Selected reaction monitoring (SRM) transitions and collision energy in the positive ion mode were m/z 503.9→309.9 and 32 eV for naloxone 3-β-D- glucuronide, m/z 394.9→219.0 and 10 eV for N-acetylserotonin β-D-glucuronide, m/z 583.9→407.9 and 26 eV for trifluoperazine glucuronide, m/z 568.9→392.9 and 30 eV for SN-38 glucuronide, and m/z 351.9→115.0 and 20 eV for meloxicam (IS). SRM transitions and collision energy in the negative ion mode: m/z 567.1→391.2 and 34 eV for chenodeoxycholic acid 24-acyl-β-glucuronide, m/z 495.0→319.0 and 20 eV for mycophenolic acid glucuronide, and m/z 353.0→177.0 and 28 eV for propofol glucuronide (IS). The linear concentrations for trifluoperazine glucuronide were 4–1200 pmol, but those of five other UGT metabolites were 1–300 pmol. The relative error and coefficient of variation values for quality control samples of six metabolites were −2.5–3.0% and 5.9–9.8%.

2.5. Data Analysis

IC50 values (i.e., the inhibitor concentrations required for 50% inhibition of the control activity) of eudesmin, fargesin, epimagnolin A, magnolin, and yangambin on UGT activities were estimated using SigmaPlot (version 12.0; Systat Software, San Jose, CA, USA). The apparent kinetic inhibition constant (Ki) values of eudesmin, fargesin, epimagnolin A, magnolin, and yangambin and the mode of inhibition were calculated from Dixon plot transformation [31].

3. Results 3.1. Inhibitory Effect of Eudesmin on Human Uridine 50-diphospho-glucuronosyltransferase (UGT) Isoforms The inhibitory effects of ﬁve tetrahydrofurofuranoid lignans on six major human UGT isoforms were measured in ultrapooled human liver microsomes and the IC50 values calcu- Pharmaceutics 2021, 13, 187 5 of 13

Pharmaceutics 2021, 13, 5 of 14 lated from the concentration dependent inhibition curves of UGT activities are summarized in Table1.

Table 1. Inhibitory potentials of eudesmin, fargesin, epimagnolin A, magnolin, and yangambin on six major UGT enzyme magnolin, and yangambin and the mode of inhibition were calculated from Dixon plot activities in ultrapooled human liver microsomes. transformation [31].

IC50 (µM) UGT Enzyme Activity 3. Results Eudesmin Fargesin Epimagnolin A Magnolin Yangambin 1A1 SN-38 glucuronidation3.1. Inhibitory Effect of Eudesmin 24.3 on Human 24.7 Uridine 5′-diphospho-glucuronosyltransferase 7.5 21.3 29.7 1A3 chenodeoxycholic acid 24-acyl-glucuronidation(UGT) Isoforms 26.6 21.5 26.6 22.9 56.5 1A4 triﬂuoperazine N-glucuronidation >200 182.7 >200 >200 >200 1A6 N-acetylserotonin glucuronidationThe inhibitory effects of 195.6 five tetrahydro 193.9furofuranoid >200 lignans >200on six major >200 human 1A9 mycophenolic acid glucuronidationUGT isoforms were measured 173.2 in ultrapooled 110.9 human liver >200 microsomes 145.7 and the IC >20050 values 2B7 naloxone 3-β-D-glucuronidationcalculated from the concentration >200 dependent 94.7 inhibition >200 curves of UGT >200 activities >200are summarized in Table 1. TheThe UGT1A1UGT1A1 andand UGT1A3UGT1A3 activitiesactivities werewere inhibited inhibitedby by the the presence presence of of eudesmin eudesmin (0.1–200(0.1–200µ μM)M) in in a a concentration concentration dependent dependent manner manner with with IC IC5050values valuesof of24.3 24.3and and26.6 26.6µ μM,M, respectively,respectively, when when measured measured by by the the formation format ofion SN-38 of SN-38 glucuronide glucuronide and chenodeoxycholic and chenodeoxy- acidcholic 24-acyl-glucuronide. acid 24-acyl-glucuronide. The UGT1A6 The UGT1A6 and UGT1A9 and UGT1A9 activities activities were were slightly slightly inhibited inhib- byited the by presence the presence of eudesmin of eudesmin with IC with50 values IC50 values of 195.6 of and195.6 173.2 andµ 173.2M, respectively, μM, respectively, when measuredwhen measured by the by formation the formation of N-acetylserotonin of N-acetylserotonin glucuronide glucuronide and mycophenolic and mycophenolic acid glucuronide.acid glucuronide. However, However, UGT1A4 UGT1A4 and UGT2B7 and UGT2 activitiesB7 activities were were not inhibited not inhibited by eudesmin by eudes- upmin to up 200 toµ M200 tested μM tested (Figure (Figure2, Table 2,1 Table). 1).

FigureFigure 2. 2. InhibitoryInhibitory effects of of eudesmin eudesmin (0.1–200 (0.1–200 μµM)M) on on the the UGT1A1-, UGT1A1-, UGT1A3-, UGT1A3-, UGT1A4-, UGT1A4-, UGT1A6-, UGT1A6-, UGT1A9-, UGT1A9-, and andUGT2B7-mediated UGT2B7-mediated glucuronidation glucuronidation of 0.5 of μ 0.5M µSN-38,M SN-38, 2 μM 2 chenodeoxycholicµM chenodeoxycholic acid, acid,0.5 μM 0.5 trifluoperazine,µM triﬂuoperazine, 1 μM 1Nµ-ace-M Ntylserotonin,-acetylserotonin, 0.2 μM 0.2 mycophenolicµM mycophenolic acid, and acid, 1 μ andM naloxone, 1 µM naloxone, a probe a substrate probe substrate for UGT1A1, for UGT1A1, UGT1A3, UGT1A3, UGT1A4, UGT1A4, UGT1A6, UGT1A9, and UGT2B7, respectively, in human liver microsomes. The data are expressed as means ± standard deviations UGT1A6, UGT1A9, and UGT2B7, respectively, in human liver microsomes. The data are expressed as means ± standard (n = 3). deviations (n = 3). 3.2. Inhibitory Effect of Fargesin on Human UGT Isoforms 3.2. InhibitoryThe UGT1A1 Effect ofand Fargesin UGT1A3 on Humanactivities UGT were Isoforms inhibited by the presence of fargesin (0.1– 200 TheμM) UGT1A1in a concentration and UGT1A3 dependent activities manner were with inhibited IC50 values by the of presence24.5 and 21.5 of fargesin μM, re- (0.1–200spectively,µM) whenin a concentrationmeasured by the dependent formation manner of SN-38 with glucuronide IC50 values and of 24.5chenodeoxycholic and 21.5 µM, respectively,acid 24-acyl-glucuronide, when measured respectively. by the formation The ofmetabolic SN-38 glucuronide activities andof UGT1A4, chenodeoxycholic UGT1A6, acidUGT1A9, 24-acyl-glucuronide, and UGT2B7 were respectively. slightly inhibit Theed metabolic by the presence activities of fargesin of UGT1A4, with IC UGT1A6,50 values of 182.7, 193.9, 110.9, and 94.7 μM, respectively (Figure 3, Table 1).

Pharmaceutics 2021, 13, 187 6 of 13

Pharmaceutics 2021, 13, 6 of 14 UGT1A9, and UGT2B7 were slightly inhibited by the presence of fargesin with IC50 values of 182.7, 193.9, 110.9, and 94.7 µM, respectively (Figure3, Table1).

FigureFigure 3. 3.Inhibitory Inhibitory effects effects of of fargesin fargesin (0.1–200 (0.1–200µ μM)M) on on UGT1A1-, UGT1A1-, UGT1A3-, UGT1A3-, UGT1A4-, UGT1A4-, UGT1A6-, UGT1A6-, UGT1A9-, UGT1A9-, and and UGT2B7- UGT2B7- mediatedmediated glucuronidation glucuronidation of of 0.5 0.5µ μMM SN-38, SN-38, 2 2µ μMM chenodeoxycholic chenodeoxycholic acid, acid, 0.5 0.5µ μMM triﬂuoperazine, trifluoperazine, 1 1µ μMMN N-acetylserotonin,-acetylserotonin, 0.20.2µ μMM mycophenolic mycophenolic acid, acid, and and 1 1µ μMM naloxone, naloxone, a a probe probe substrate substrate for for UGT1A1, UGT1A1, UGT1A3, UGT1A3, UGT1A4, UGT1A4, UGT1A6, UGT1A6, UGT1A9, UGT1A9, andand UGT2B7, UGT2B7, respectively, respectively, in in human human liver liver microsomes. microsomes. The The data data are are expressed expressed as as means means± ±standard standard deviations deviations ( n(n= = 3). 3).

Table 1. Inhibitory potentials of eudesmin, fargesin, epimagnolin A, magnolin, and yangambin on six major UGT enzyme 3.3. Inhibitory Effect of Epimagnolin A on Human UGT Isoforms activities in ultrapooled human liver microsomes. The UGT1A1 and UGT1A3 activities were inhibited by the presence of epimagnolin A IC50 (µM) UGT Enzyme(0.1–200 ActivityµM) in a concentration dependent manner with IC50 values of 7.5 and 26.6 µM, respectively, when measuredEudesmin by the formationFargesin ofEpimagno SN-38 glucuronidelin A Magnolin and chenodeoxy- Yangambin 1A1 SN-38 glucuronidationcholic acid 24-acyl-glucuronide. 24.3 The metabolic 24.7 activities 7.5 of UGT1A4, UGT1A6, 21.3 UGT1A9, 29.7 1A3 chenodeoxycholic acidand 24-acyl-glucuronidation UGT2B7 were negligibly inhibited 26.6 by 21.5 epimagnolin 26.6 A up to 200 µM 22.9 tested (Figure 56.5 4, 1A4 trifluoperazine N-glucuronidationTable1). >200 182.7 >200 >200 >200 1A6 N-acetylserotonin glucuronidation 195.6 193.9 >200 >200 >200 3.4. Inhibitory Effect of Magnolin on Human UGT Isoforms 1A9 mycophenolic acid glucuronidation 173.2 110.9 >200 145.7 >200 2B7 naloxone 3-β-D-glucuronidationThe UGT1A1 and UGT1A3>200 activities 94.7 were inhibited>200 by the presence>200 of magnolin>200 (0.1–200 µM) in a concentration dependent manner with IC50 values of 21.3 and 22.9 µM, respectively, when measured by the formation of SN-38 glucuronide and chenodeoxycholic 3.3. Inhibitory Effect of Epimagnolin A on Human UGT Isoforms acid 24-acyl-glucuronide. The UGT1A9 activity was slightly inhibited by the presence of The UGT1A1 and UGT1A3 activities were inhibited by the presence of epimagnolin magnolin with IC50 values of 145.7 µM, when measured by the formation of mycophenolic acidA (0.1–200 glucuronide. μM) in a However, concentration the UGT1A4,dependent UGT1A6, manner with and IC UGT2B750 values activities of 7.5 and were 26.6 notμM, inhibitedrespectively, by the when presence measured of magnolin by the upformat to 200ionµ Mof testedSN-38 (Figureglucuronide5, Table and1). chenodeoxycholic acid 24-acyl-glucuronide. The metabolic activities of UGT1A4, UGT1A6, UGT1A9, 3.5.and Inhibitory UGT2B7 Effectwere ofnegligibly Yangambin inhibited on Human by UGTepimagnolin Isoforms A up to 200 μM tested (Figure 4, TableThe 1). UGT1A1 and UGT1A3 activities were inhibited by the presence of yangambin (0.1–200 µM) in a concentration dependent manner with IC50 values of 29.7 and 56.5 µM, respectively, when measured by the formation of SN-38 glucuronide and chenodeoxycholic acid 24-acyl-glucuronide. The metabolic activities of UGT1A4, UGT1A6, UGT1A9, and UGT2B7 were negligibly inhibited by yangambin up to 200 µM tested (Figure6, Table1).

Pharmaceutics 2021, 13, 7 of 14

Pharmaceutics 2021, 13, 187 7 of 13 Pharmaceutics 2021, 13, 7 of 14

Figure 4. Inhibitory effects of epimagnolin A (0.1–200 μM) on UGT1A1-, UGT1A3-, UGT1A4-, UGT1A6-, UGT1A9-, and UGT2B7-mediated glucuronidation of 0.5 μM SN-38, 2 μM chenodeoxycholic acid, 0.5 μM trifluoperazine, 1 μM N-acetylserotonin, 0.2 μM mycophenolic acid, and 1 μM naloxone, a probe substrate for UGT1A1, UGT1A3, UGT1A4, UGT1A6, UGT1A9, and UGT2B7, respectively, in human liver microsomes. The data are expressed as means ± standard deviations (n = 3).

3.4. Inhibitory Effect of Magnolin on Human UGT Isoforms The UGT1A1 and UGT1A3 activities were inhibited by the presence of magnolin (0.1–200 μM) in a concentration dependent manner with IC50 values of 21.3 and 22.9 μM, FigureFigure 4. Inhibitory 4. Inhibitory effects effects of epimagnolinof epimagnolin A A (0.1–200(0.1–200 μµM)M) on on UGT1A1-, UGT1A1-, UGT1A3-, UGT1A3-, UGT1A4-, UGT1A4-, UGT1A6-, UGT1A6-, UGT1A9-, UGT1A9-, and respectively, when measured by the formation of SN-38 glucuronide and chenodeoxy- and UGT2B7-mediatedUGT2B7-mediated glucuronidationglucuronidation of of0.5 0.5 μMµ SN-38,M SN-38, 2 μM 2 chenodeoxycholicµM chenodeoxycholic acid, 0.5 acid, μM trifluoperazine, 0.5 µM triﬂuoperazine, 1 μM N-ace- 1 µM tylserotonin, 0.2 μM mycophenoliccholic acid acid, 24-acyl-glucuronide. and 1 μM naloxone, a probeThe UGT1A9 substrate foractivity UGT1A1, was UGT1A3, slightly inhibitedUGT1A4, UGT1A6, by the pres- N-acetylserotonin, 0.2 µM mycophenolic acid, and 1 µM naloxone, a probe substrate for UGT1A1, UGT1A3, UGT1A4, UGT1A9, and UGT2B7, respectively,ence of magnolin in human with liver ICmicrosomes50 values .of The 145.7 data μ areM, expressed when measured as means by ± standard the formation deviations of my- UGT1A6,(n = UGT1A9, 3). and UGT2B7,cophenolic respectively, acid inglucuronide. human liver However, microsomes. the TheUGT1A4, data are UGT1A6, expressed and as UGT2B7 means ± activitiesstandard deviations (n = 3). were not inhibited by the presence of magnolin up to 200 μM tested (Figure 5, Table 1). 3.4. Inhibitory Effect of Magnolin on Human UGT Isoforms The UGT1A1 and UGT1A3 activities were inhibited by the presence of magnolin (0.1–200 μM) in a concentration dependent manner with IC50 values of 21.3 and 22.9 μM, respectively, when measured by the formation of SN-38 glucuronide and chenodeoxycholic acid 24-acyl-glucuronide. The UGT1A9 activity was slightly inhibited by the presence of magnolin with IC50 values of 145.7 μM, when measured by the formation of mycophenolic acid glucuronide. However, the UGT1A4, UGT1A6, and UGT2B7 activities were not inhibited by the presence of magnolin up to 200 μM tested (Figure 5, Table 1).

FigureFigure 5. Inhibitory 5. Inhibitory effects effects of magnolin of magnolin (0.1–200 (0.1–200µM) onμM) UGT1A1-, on UGT1A1-, UGT1A3-, UGT1A3-, UGT1A4-, UGT1A4-, UGT1A6-, UGT1A6-, UGT1A9-, UGT1A9-, and and UGT2B7- mediatedUGT2B7-mediated glucuronidation glucuronidation of 0.5 µM SN-38, of 0.5 2μµMM SN-38, chenodeoxycholic 2 μM chenodeoxycholic acid, 0.5 µ acid,M triﬂuoperazine, 0.5 μM trifluoperazine, 1 µM N 1-acetylserotonin, μM N-acetylserotonin, 0.2 μM mycophenolic acid, and 1 μM naloxone, a probe substrate for UGT1A1, UGT1A3, UGT1A4, UGT1A6, 0.2 µMUGT1A9, mycophenolic and UGT2B7, acid, andrespectively, 1 µM naloxone, in human aliver probe microsomes substrate. The for data UGT1A1, are expressed UGT1A3, as means UGT1A4, ± standard UGT1A6, deviations UGT1A9, and UGT2B7,(n = 3). respectively, in human liver microsomes. The data are expressed as means ± standard deviations (n = 3).

3.6. Kinetic Analysis of Eudesmin, Fargesin, Epimagnolin A, Magnolin, and Yangambin on UGT1A1 and UGT1A3 Activities in Human Liver Microsomes Figure 5. Inhibitory effects of magnolin (0.1–200 μM) on UGT1A1-, UGT1A3-, UGT1A4-, UGT1A6-, UGT1A9-, and K UGT2B7-mediated glucuronidationThe Dixon of 0.5 μ plotsM SN-38, were 2 μ usedM chenodeoxycholic to calculate acid, the 0.5i μvaluesM trifluoperazine, and inhibition 1 μM N modes-ace- of eu- tylserotonin, 0.2 μM mycophenolicdesmin, fargesin, acid, and epimagnolin1 μM naloxone, A,a probe magnolin, substrate and for UGT1A1, yangambin UGT1A3, for theUGT1A4, inhibition UGT1A6, of catabolic UGT1A9, and UGT2B7,activities respectively, of in UGT1A1 human liver and microsomes UGT1A3. The enzymes data are expressed (Figures as7 andmeans8, ± Table standard2). deviations In addition, eu- (n = 3). desmin, fargesin, epimagnolin A, magnolin, and yangambin noncompetitively inhibited SN-38 glucuronidation by UGT1A1 with Ki values of 25.7, 25.3, 3.6, 26.0, and 17.1 µM, respectively, in pooled human liver microsomes (Figure7, Table2). Conversely, eudesmin, fargesin, epimagnolin A, magnolin, and yangambin showed the competitive inhibition of Pharmaceutics 2021, 13, 8 of 14

Pharmaceutics 2021, 13, 187 8 of 13 3.5. Inhibitory Effect of Yangambin on Human UGT Isoforms The UGT1A1 and UGT1A3 activities were inhibited by the presence of yangambin (0.1–200 μM) in a concentration dependent manner with IC50 values of 29.7 and 56.5 μM, chenodeoxycholicrespectively, when acid measured 24-acyl-glucuronidation by the formation of SN-38 by UGT1A3 glucuronide with andKi valueschenodeoxy- of 39.8, 24.5, 15.1,cholic 37.6, acid and 24-acyl-glucuronide. 66.8 µM, respectively The metabolic (Figure8 ,activities Table2). of UGT1A4, UGT1A6, UGT1A9, and UGT2B7 were negligibly inhibited by yangambin up to 200 μM tested (Figure 6, Table 1).

Pharmaceutics 2021, 13, 9 of 14

Table 2. Kinetic parameters for the inhibition of eudesmin, fargesin, epimagnolin A, magnolin, and yangambin on UGT1A1 and UGT1A3 enzyme activities in pooled human liver microsomes.

Ki (µM, Inhibition Mode) Figure 6. Inhibitory effects of yangambinSN-38 (0.1–200 µM) on UGT1A1-,Chenodeoxycholic UGT1A3-, Acid 24-acyl-Glucuronidation UGT1A4-, UGT1A6-, UGT1A9-, Figure 6. Inhibitory effects of yangambin (0.1–200 μM) on UGT1A1-, UGT1A3-, UGT1A4-, UGT1A6-, UGT1A9-, and Glucuronidation (UGT1A1) (UGT1A3) and UGT2B7-mediatedUGT2B7-mediated glucuronidationglucuronidation of of0.5 0.5μMµ SN-38,M SN-38, 2 μM 2chenodeoxycholicµM chenodeoxycholic acid, 0.5 μ acid,M trifluoperazine, 0.5 µM triﬂuoperazine, 1 μM N-ace- 1 µM Eudesmin 25.7 (noncompetitive) 39.8 (competitive) N-acetylserotonin,tylserotonin, 0.2 μµMM mycophenolic mycophenolic acid, acid, and 1 andμM naloxone, 1 µM naloxone, a probe substrate a probe for substrate UGT1A1, forUGT1A3, UGT1A1, UGT1A4, UGT1A3, UGT1A6, UGT1A4, Fargesin 25.3 (noncompetitive) 24.5 (competitive) UGT1A9, and UGT2B7, respectively, in human liver microsomes. The data are expressed as means ± standard deviations UGT1A6, UGT1A9,Epimagnolin and UGT2B7, A respectively, 3.6 (noncompet in humanitive) liver microsomes. The15.1 data (competitive) are expressed as means ± standard (n = 3). deviations (n = 3).Magnolin 26.0 (noncompetitive) 37.6 (competitive) Yangambin 17.1 (noncompetitive) 66.8 (competitive) 3.6. Kinetic Analysis of Eudesmin, Fargesin, Epimagnolin A, Magnolin, and Yangambin on UGT1A1 and UGT1A3 Activities in Human Liver Microsomes

The Dixon plots were used to calculate the Ki values and inhibition modes of eudesmin, fargesin, epimagnolin A, magnolin, and yangambin for the inhibition of catabolic activities of UGT1A1 and UGT1A3 enzymes (Figures 7 and 8, Table 2). In addition, eudesmin, fargesin, epimagnolin A, magnolin, and yangambin noncompetitively inhibited SN- 38 glucuronidation by UGT1A1 with Ki values of 25.7, 25.3, 3.6, 26.0, and 17.1 μM, respectively, in pooled human liver microsomes (Figure 7, Table 2). Conversely, eudesmin, fargesin, epimagnolin A, magnolin, and yangambin showed the competitive inhibition of chenodeoxycholic acid 24-acyl-glucuronidation by UGT1A3 with Ki values of 39.8, 24.5, 15.1, 37.6, and 66.8 μM, respectively (Figure 8, Table 2). Next, we measured the IC50 values of NDC-052 that contained eudesmin, fargesin, epimagnolin A, magnolin, and yangambin at a content of 4.1, 3.4, 11.9, 21.5, and 9.1%, respectively [18] on UGT1A1- and UGT1A3-mediated glucuronidation. The UGT1A1 and UGT1A3 activities were inhibited by the presence of NDC-052 (0.1–100 μg/mL) in a concentration dependent manner and the IC50 values of NDC-052 for UGT1A1 and UGT1A3 were calculated as 38.1 and 65.0 μg/mL, respectively.

Figure 7. Dixon plots for the inhibitory effects of (a) eudesmin, (b) fargesin, (c) epimagnolin A, (d) magnolin, and (e) Figure 7. Dixonyangambin plots for on the UGT1A1-catalyzed inhibitory effects SN-38 glucuronidation of (a) eudesmin, in pooled (b human) fargesin, liver microsomes. (c) epimagnolin Each symbol A, represents (d) magnolin, the and (e) yangambin on UGT1A1-catalyzedconcentration of SN-38: ● SN-38, 0.2 μM; glucuronidation ○, 0.5 μM; ▼, 1 μM; in▽, pooled2 μM. The human data are means liver microsomes.± SDs (n = 3). Each symbol represents the concentration of SN-38: •, 0.2 µM; , 0.5 µM; H, 1 µM; 5, 2 µM. The data are means ± SDs (n = 3). #

PharmaceuticsPharmaceutics2021 2021, 13, 13, 187, 109 of of 13 14

FigureFigure 8.8. DixonDixon plotsplots forfor thethe inhibitoryinhibitory effectseffects ofof (a(a)) eudesmin,eudesmin, ((bb)) fargesin,fargesin, ((cc)) epimagnolinepimagnolin A,A, ((dd)) magnolin,magnolin, andand ((ee)) yangambinyangambin onon UGT1A3-catalyzedUGT1A3-catalyzed chenodeoxycholicchenodeoxycholic acidacid 24-acyl24-acyl glucuronidationglucuronidation inin pooledpooled humanhuman liverliver microsomes.microsomes. Each symbol represents the concentration of chenodeoxycholic acid: ●, 1 μM; ○ , 2 μM; ▼, 5 μM; ▽, 10 μM. The data are Each symbol represents the concentration of chenodeoxycholic acid: •, 1 µM; , 2 µM; H, 5 µM; 5, 10 µM. The data are means ± SDs (n = 3). means ± SDs (n = 3). # 4. Discussion Table 2. Kinetic parameters for the inhibition of eudesmin, fargesin, epimagnolin A, magnolin, and yangambin on UGT1A1 and UGT1A3 enzyme activities inEudesmin, pooled human fargesin, liver microsomes. epimagnolin A, magnolin, and yangambin were selected as major tetrahydrofurofuranoid lignans that exist in the ethanol extract of flower bud of M. fargesii (NDC-052) and are also presentKi (µM, in Inhibition the rat plasma Mode) samples following oral administration [18]. Therefore, the inhibitory potentials (ICChenodeoxycholic50 values) of eudesmin, Acid fargesin, SN-38 Glucuronidation (UGT1A1) epimagnolin A, magnolin, and yangambin on 24-acyl-GlucuronidationUGT1A1, UGT1A3, UGT1A4, (UGT1A3) UGT1A6, EudesminUGT1A9, and 25.7 UGT2B7 (noncompetitive) enzyme activities were evaluated 39.8 in (competitive) pooled human liver micro- Fargesinsomes in the 25.3 concentration (noncompetitive) ranges of 0.1–200 μM of individual 24.5 (competitive) tetrahydrofurofuranoid Epimagnolin Alignan (Figures 3.6 (noncompetitive) 2–6; Table 1). The concentration range of lignans 15.1 (competitive) tested in this study could Magnolincover the plasma 26.0 (noncompetitive) concentration of lignans in rats after an 37.6 oral (competitive) administration of NDC-052 Yangambin 17.1 (noncompetitive) 66.8 (competitive) (22 mg/kg) and the expected gastric concentrations of five lignans. For example, the Cmax values of eudesmin, fargesin, epimagnolin A, magnolin, and yangambin were 2.2 ± 0.41 μM, Next,0.38 ± we0.18 measured μM, 3.07 the± 0.72 IC 50μM,values 5.95 ± of 1.12 NDC-052 μM, and that 0.52 contained ± 0.30 μM, eudesmin, respectively, fargesin, after epimagnolinsingle oral administration A, magnolin, of and NDC-052 yangambin (22 mg/kg) at a content [18]. Herbal of 4.1, extracts 3.4, 11.9, are 21.5, diluted and by 9.1%, gas- respectivelytrointestinal [fluid18] on after UGT1A1- oral administration and UGT1A3-mediated by a factor of glucuronidation. oral dilution volume The (500–2200 UGT1A1 andmL UGT1A3in humans) activities [32,33]. were Therefore, inhibited the by ex thepected presence gastric of NDC-052 concentrations (0.1–100 ofµ eudesmin,g/mL) in afargesin, concentration epimagnolin dependent A, magnolin, manner and and yangambin the IC50 values could ofbe NDC-052calculated foras 29–127 UGT1A1 μM, and 25– UGT1A3110 μM, were78–342 calculated μM, 140–618 as 38.1 μ andM, and 65.0 µ55–244g/mL, μ respectively.M by considering the content of five lignans (eudesmin, fargesin, epimagnolin A, magnolin, and yangambin as 4.1, 3.4, 11.9,

Pharmaceutics 2021, 13, 187 10 of 13

4. Discussion Eudesmin, fargesin, epimagnolin A, magnolin, and yangambin were selected as major tetrahydrofurofuranoid lignans that exist in the ethanol extract of flower bud of M. fargesii (NDC-052) and are also present in the rat plasma samples following oral administration [18]. Therefore, the inhibitory potentials (IC50 values) of eudesmin, fargesin, epimagnolin A, magnolin, and yangambin on UGT1A1, UGT1A3, UGT1A4, UGT1A6, UGT1A9, and UGT2B7 enzyme activities were evaluated in pooled human liver microsomes in the concentration ranges of 0.1–200 µM of individual tetrahydrofurofuranoid lignan (Figures2–6; Table1). The concentration range of lignans tested in this study could cover the plasma concentration of lignans in rats after an oral administration of NDC-052 (22 mg/kg) and the expected gastric concentrations of five lignans. For example, the Cmax values of eudesmin, fargesin, epimagnolin A, magnolin, and yangambin were 2.2 ± 0.41 µM, 0.38 ± 0.18 µM, 3.07 ± 0.72 µM, 5.95 ± 1.12 µM, and 0.52 ± 0.30 µM, respectively, after single oral administration of NDC-052 (22 mg/kg) [18]. Herbal extracts are diluted by gastrointestinal fluid after oral administration by a factor of oral dilution volume (500–2200 mL in humans) [32,33]. Therefore, the expected gastric concentrations of eudesmin, fargesin, epimagnolin A, magnolin, and yangambin could be calculated as 29–127 µM, 25–110 µM, 78–342 µM, 140–618 µM, and 55–244 µM by considering the content of five lignans (eudesmin, fargesin, epimagnolin A, magnolin, and yangambin as 4.1, 3.4, 11.9, 21.5, and 9.1%, respectively) [18] and oral dose of NDC-052 in adult asthmatic patients (600 mg/day for 8 weeks) [19]. Eudesmin, fargesin, epimagnolin A, magnolin, and yangambin reversibly and noncompetitively inhibited UGT1A1-catalyzed SN-38 glucuronidation with Ki values of 25.7, 25.3, 3.6, 26.0, and 17.1 µM, respectively, indicating the possible interactions with UGT1A1 substrates (e.g., bilirubin, irinotecan, trovafloxacin, belinostat, dolutegravir, pterostilbene, and resveratrol) [34–38]. In addition, eudesmin, fargesin, epimagnolin A, magnolin, and yangambin were competitive inhibitors of UGT1A3 with Ki values of 39.8, 24.5, 15.1, 37.6, and 66.8 µM, respectively, indicating that they should be used carefully with UGT1A3 substrates (e.g., chenodeoxycholic acid, fimasartan, losartan, candesartan, and zolarsartan) [39–41] to avoid possible drug interactions. However, they showed weak or negligible inhibition toward the other four UGT isoforms such as UGT1A4, UGT1A6, UGT1A9, and UGT2B7 tested in this study (IC50 values of about 100 µM or higher). The five tetrahydrofurofuranoid lignans possess a similar chemical structure, which made reversible inhibition of five lignans to UGT1A1 and UGT1A3 but not to UGT1A4, UGT1A6, UGT1A9, and UGT2B7 isozymes. Moreover, these five tetrahydrofurofuranoid lignans were all found in NDC-052 with Cmax values of 2.2 ± 0.41 µM (eudesmin), 0.38 ± 0.18 µM (fargesin), 3.07 ± 0.72 µM (epimagnolin A), 5.95 ± 1.12 µM (magnolin), and 0.52 ± 0.30 µM (yangambin), respectively, in rats [18]. Therefore, these five lignans may potentiate coordinately the likelihood of UGT1A1- and UGT1A3-mediated drug interaction with co-administration of victim drugs. As expected, NDC-052 inhibited UGT1A1- and UGT1A3-mediated glucuronidation in a concentration dependent manner with IC50 values of 38.1 and 65.0 µg/mL, respectively. When the IC50 values were calculated as the concentration of individual lignans by using the content of five lignans in NDC-052, 38.1 and 65.0 µg/mL of NDC-052 contained 4.0 and 6.9 µM of eudesmin, 3.5 and 6.0 µM of fargesin, 10.9 and 18.6 µM of epimagnolin A, 19.7 and 33.6 µM of magnolin, and 7.8 and 13.2 µM of yangambin, respectively. These concentrations were lower than the IC50 values of individual lignans in Table1, which suggested that the inhibitory effect of NDC-052 on the metabolic activities of UGT1A1 and UGT1A3 may be caused by the coexistence of lignans rather than the inhibition by single lignan. However, we also should note that the total content of five lignans accounts for 50% of NDC-052 and unveiled components such as terpenoids, alkaloids, and flavonoids may serve as inhibitors of UGT enzymes [42]. The efficacy of NDC-052 has been evaluated for the treatment of asthma in guinea pig chronic asthma model at a repeated oral dose (50 mg/kg/day for 8 weeks) as well as in Pharmaceutics 2021, 13, 187 11 of 13

adult asthmatic patients (600 mg/day for 8 weeks) [19]. Although the plasma and gastric concentrations of individual eudesmin, fargesin, epimagnolin A, magnolin, and yangambin in asthmatic patients were not reported, the sum of the plasma and gastric concentrations for these five lignans may reach or exceed Ki values to UGT1A1 and UGT1A3 in the above efficacy model considering the Cmax of the previous study [18] and the expected gastric concentration of five lignans after repeated oral administration. Therefore, these five lignans may potentiate coordinately the likelihood of UGT1A1- and UGT1A3-mediated drug interaction with co-administration of victim drugs. In addition, among the five tetrahydrofurofuranoid lignans (e.g., eudesmin, epimagnolin A, fargesin, magnolin, and yangambin), only fargesin inhibited competitively CYP2C9 −1 (Ki, 16.3 µM) and irreversibly inhibited CYP2C8 (Ki: 10.7 µM and kinact: 0.082 min ), −1 CYP2C19 (Ki: 3.7 µM and kinact: 0.102 min ), and CYP3A4 (Ki: 23.0 µM and kinact: 0.050 min−1) in human liver microsomes [22]. Consequently, it indicated that CYP2C8, CYP2C9, CYP2C19, and CYP3A4 activities might be inhibited by the presence of fargesin that may cause reversible inhibition on CYP2C9 and the formation of intermediate metabolites of fargesin which may cause irreversible inhibition ofCYP2C8, CYP2C19, and CYP3A4. This could cause CYP2C8, CYP2C9, CYP2C19, and CYP3A4-mediated drug interaction with co-administered victim drugs or co-existing components such as magnolin, epimagnolin A, eudesmin, and yangambin. In a previous study, the AUC of magnolin was linearly increased with an increase in the intravenous dose (0.5–2 mg/kg) and in the oral dose (1–4 mg/kg) [21]. In addition, magnolin was mainly eliminated by metabolism [21]. Three major metabolites (i.e., O- desmethylmagnolin, i.e., M1 and M2, and hydroxymagnolin, M4) were formed by CYP2C8, CYP2C9, CYP2C19, and CYP3A4 in the in vitro studies of magnolin metabolism [43]. However, the AUC of magnolin was not proportionally increased with increasing oral dose of NDC-052 (5.5–22 mg/kg containing 1–4 mg/kg magnolin) and about two-fold higher than that following oral administration of magnolin alone [18,20]. Taken together, fargesin may be involved in the two-fold increase of magnolin AUC by inhibiting the CYP2C8, CYP2C9, CYP2C19, and CYP3A4-mediated magnolin metabolism when orally administered as NDC-052 to the same dose of single magnolin. For the elucidation of the coordinated drug interaction among the similar structure possessing components such as these tetrahydrofurofuranoid lignans in the NDC-052, the CYPs and UGTs-mediated metabolism of these tetrahydrofurofuranoid lignans awaits further investigation. In this regard, this study would provide the molecular mechanism for understanding the drug interactions among the structurally related herbal components and the concomitantly administered victim drugs.

5. Conclusions Eudesmin, fargesin, epimagnolin A, magnolin, and yangambin noncompetitively inhibited UGT1A1-catalyzed SN-38 glucuronidation with Ki values of 25.7, 25.3, 3.6, 26.0, and 17.1 µM, respectively. Conversely, the aforementioned tetrahydrofurofuranoid lignans competitively inhibited UGT1A3-catalyzed chenodeoxycholic acid 24-acyl-glucuronidation with Ki values of 39.8, 24.5, 15.1, 37.6, and 66.8 µM, respectively. However, these ﬁve tetrahydrofurofuranoid lignans weakly or negligibly inhibited the glucuronidation activities of UGT1A4, UGT1A6, UGT1A9, and UGT2B7. Consequently, these in vitro results suggest that eudesmin, fargesin, epimagnolin A, magnolin, and yangambin alone and the herbal preparations containing these tetrahydrofurofuranoid lignans should be examined in terms of potential in vivo pharmacokinetic drug interactions caused by inhibition of UGT1A1 and UGT1A3 activities.

Author Contributions: Conceptualization, R.P., E.J.P., and H.S.L.; methodology, R.P., E.J.P., Y.-Y.C., and H.S.L.; investigation, R.P., E.J.P., and I.-S.S.; writing—original draft preparation, R.P. and E.J.P.; writing—review and editing, J.Y.L., H.C.K., Y.-Y.C., I.-S.S., and H.S.L.; supervision, H.S.L.; funding acquisition, H.S.L. All authors have read and agreed to the published version of the manuscript. Pharmaceutics 2021, 13, 187 12 of 13

Funding: This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (NRF-2020R1A2C2008461 and NRF2020R1A4A2002894) and The Catholic University of Korea in 2020. Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: The data presented in this study are available upon request. Conﬂicts of Interest: The authors declare no conﬂict of interest.

References 1. Zhou, X.; Chen, C.; Ye, X.; Song, F.; Fan, G.; Song, F. Analysis of lignans in Magnoliae Flos by turbulent flow chromatography with online solid-phase extraction and high-performance liquid chromatography with tandem mass spectrometry. J. Sep. Sci. 2016, 39, 1266–1272. [CrossRef][PubMed] 2. Bhatt, V.; Sharma, S.; Kumar, N.; Sharma, U.; Singh, B. Simultaneous quantification and identification of flavonoids, lignans, coumarin and amides in leaves of Zanthoxylum armatum using UPLC-DAD-ESI-QTOF–MS/MS. J. Pharm. Biomed. Anal. 2017, 132, 46–55. [CrossRef][PubMed] 3. Kumar, V.; Kumar, S.; Singh, B.; Kumar, N. Quantitative and structural analysis of amides and lignans in Zanthoxylum armatum by UPLC-DAD-ESI-QTOF–MS/MS. J. Pharm. Biomed. Anal. 2014, 94, 23–29. [CrossRef][PubMed] 4. Guo, T.; Su, D.; Huang, Y.; Wang, Y.; Li, Y. Ultrasound-Assisted Aqueous Two-Phase System for Extraction and Enrichment of Zanthoxylum armatum Lignans. Molecules 2015, 20, 15273–15286. [CrossRef] 5. Guo, T.; Deng, Y.-X.; Xie, H.; Yao, C.-Y.; Cai, C.-C.; Pan, S.-L.; Wang, Y.-L. Antinociceptive and anti-inflammatory activities of ethyl acetate fraction from Zanthoxylum armatum in mice. Fitoterapia 2011, 82, 347–351. [CrossRef] 6. Lim, H.; Son, K.H.; Bae, K.H.; Hung, T.M.; Kim, Y.S.; Kim, H.P. 5-Lipoxygenase-inhibitory constituents from Schizandra fructus and Magnolia flos. Phytother. Res. 2009, 23, 1489–1492. [CrossRef][PubMed] 7. Baek, J.A.; Lee, Y.D.; Lee, C.B.; Go, H.K.; Kim, J.P.; Seo, J.J.; Rhee, Y.K.; Kim, A.M.; Na, D.J. Extracts of Magnoliae flos inhibit inducible nitric oxide synthase via ERK in human respiratory epithelial cells. Nitric Oxide 2009, 20, 122–128. [CrossRef] 8. Kim, J.S.; Kim, J.Y.; Lee, H.J.; Lim, H.J.; Lee, D.Y.; Kim, D.H.; Ryu, J.-H. Suppression of inducible nitric oxide synthase expression by furfuran lignans from flower buds of Magnolia fargesii in BV-2 microglial cells. Phytother. Res. 2009, 24, 748–753. [CrossRef] 9. Yue, B.; Ren, Y.J.; Zhang, J.J.; Luo, X.P.; Yu, Z.L.; Ren, G.Y.; Sun, A.N.; Deng, C.; Wang, Z.T.; Dou, W. Anti-Inflammatory Effects of Fargesin on Chemically Induced Inflammatory Bowel Disease in Mice. Molecules 2018, 23, 1380. [CrossRef] 10. Yoo, S.M.; Lee, C.J.; Kang, H.C.; Lee, H.S.; Lee, J.Y.; Kim, K.D.; Kim, D.J.; An, H.J.; Cho, Y.Y. Epimagnolin targeting on an active pocket of mammalian target of rapamycin suppressed cell transformation and colony growth of lung cancer cells. Mol. Carcinog. 2019, 58, 1221–1233. [CrossRef] 11. Lee, C.-J.; Lee, M.-H.; Yoo, S.-M.; Choi, K.-I.; Song, J.H.; Jang, J.-H.; Oh, S.-R.; Ryu, H.W.; Lee, H.S.; Dong, Z.; et al. Magnolin inhibits cell migration and invasion by targeting the ERKs/RSK2 signaling pathway. BMC Cancer 2015, 15, 1–12. [CrossRef] [PubMed] 12. Jun, A.Y.; Kim, H.-J.; Park, K.-K.; Son, K.H.; Lee, D.H.; Woo, M.-H.; Chung, W.-Y. Tetrahydrofurofuran-type lignans inhibit breast cancer-mediated bone destruction by blocking the vicious cycle between cancer cells, osteoblasts and osteoclasts. Investig. New Drugs 2014, 32, 1–13. [CrossRef][PubMed] 13. Yang, J.S.; Wang, C.M.; Su, C.H.; Ho, H.C.; Chang, C.H.; Chou, C.H.; Hsu, Y.M. Eudesmin attenuates Helicobacter pylori-induced epithelial autophagy and apoptosis and leads to eradication of H. pylori infection. Exp. Ther. Med. 2018, 15, 2388–2396. [PubMed] 14. Wang, X.; Cheng, Y.; Xue, H.; Yue, Y.; Zhang, W.; Li, X. Fargesin as a potential β1 adrenergic receptor antagonist protects the hearts against ischemia/reperfusion injury in rats via attenuating oxidative stress and apoptosis. Fitoterapia 2015, 105, 16–25. [CrossRef][PubMed] 15. Lee, Y.-S.; Cha, B.-Y.; Choi, S.-S.; Harada, Y.; Choi, B.-K.; Yonezawa, T.; Teruya, T.; Nagai, K.; Woo, J.-T. Fargesin improves lipid and glucose metabolism in 3T3-L1 adipocytes and high-fat diet-induced obese mice. BioFactors 2012, 38, 300–308. [CrossRef] 16. Sha, S.; Xu, D.; Wang, Y.; Zhao, W.; Li, X. Antihypertensive effects of fargesin in vitro and in vivo via attenuating oxidative stress and promoting nitric oxide release. Can. J. Physiol. Pharmacol. 2016, 94, 900–906. [CrossRef] 17. Wang, F.; Zhang, G.; Zhou, Y.; Gui, D.; Li, J.; Xing, T.; Wang, N. Magnolin Protects against Contrast-Induced Nephropathy in Rats via Antioxidation and Antiapoptosis. Oxidative Med. Cell. Longev. 2014, 2014, 1–7. [CrossRef] 18. Jeong, J.H.; Kim, D.K.; Ji, H.Y.; Oh, S.-R.; Lee, H.-K.; Lee, H.S. Liquid chromatography-atmospheric pressure chemical ionization tandem mass spectrometry for the simultaneous determination of dimethoxyaschantin, dimethylliroresinol, dimethylpinoresinol, epimagnolin A, fargesin and magnolin in rat plasma. Biomed. Chromatogr. 2010, 25, 879–889. [CrossRef] 19. Park, C.S.; Kim, T.-B.; Lee, J.-Y.; Park, J.Y.; Lee, Y.C.; Jeong, S.S.; Lee, Y.D.; Cho, Y.S.; Moon, H.-B. Effects of Add-On Therapy with NDC-052, an Extract fromMagnoliae Flos, in Adult Asthmatic Patients Receiving Inhaled Corticosteroids. Korean J. Intern. Med. 2012, 27, 84–90. [CrossRef] Pharmaceutics 2021, 13, 187 13 of 13

20. Kim, N.J.; Oh, S.R.; Lee, H.K.; Lee, H.S. Simultaneous determination of magnolin and epimagnolin A in rat plasma by liquid chromatography with tandem mass spectrometry: Application to pharmacokinetic study of a purified extract of the dried flower buds of Magnolia fargesii, NDC-052 in rats. J. Pharm. Biomed. Anal. 2009, 50, 53–57. [CrossRef] 21. Kim, N.J.; Song, W.Y.; Yoo, S.D.; Oh, S.-R.; Lee, H.-K.; Lee, H.S. Pharmacokinetics of magnolin in rats. Arch. Pharmacal Res. 2010, 33, 933–938. [CrossRef][PubMed] 22. Na, D.H.; Ji, H.Y.; Park, E.J.; Kim, M.S.; Liu, K.-H.; Lee, H.S. Evaluation of metabolism-mediated herb-drug interactions. Arch. Pharmacal Res. 2011, 34, 1829–1842. [CrossRef][PubMed] 23. Ma, B.-L.; Ma, Y. Pharmacokinetic herb–drug interactions with traditional Chinese medicine: Progress, causes of conflicting results and suggestions for future research. Drug Metab. Rev. 2016, 48, 1–26. [CrossRef][PubMed] 24. Meng, Q.; Liu, K. Pharmacokinetic interactions between herbal medicines and prescribed drugs: Focus on drug metabolic enzymes and transporters. Curr. Drug Metab. 2015, 15, 791–807. [CrossRef][PubMed] 25. Brantley, S.J.; Argikar, A.A.; Lin, Y.S.; Nagar, S.; Paine, M.F. Herb–Drug Interactions: Challenges and Opportunities for Improved Predictions. Drug Metab. Dispos. 2014, 42, 301–317. [CrossRef] 26. Roe, A.L.; Paine, M.F.; Gurley, B.J.; Brouwer, K.R.; Jordan, S.; Griffiths, J.C. Assessing Natural Product–Drug Interactions: An End-to-End Safety Framework. Regul. Toxicol. Pharmacol. 2016, 76, 1–6. [CrossRef] 27. Zuo, Z.; Huang, M.; Kanfer, I.; Chow, M.S.S.; Cho, W.C.S. Herb-Drug Interactions: Systematic Review, Mechanisms, and Therapies. Evid. Based Complement. Alternat. Med. 2015, 2015, 239150. [CrossRef] 28. Kwon, S.-S.; Kim, J.-H.; Jeong, H.-U.; Cho, Y.-Y.; Oh, S.-R.; Lee, H.S. Inhibitory Effects of Aschantin on Cytochrome P450 and Uridine 50-diphospho-glucuronosyltransferase Enzyme Activities in Human Liver Microsomes. Molecules 2016, 21, 554. [CrossRef] 29. Albaugh, D.R.; Fullenwider, C.L.; Fisher, M.B.; Hutzler, J.M. Time-Dependent Inhibition and Estimation of CYP3A Clinical Pharmacokinetic Drug-Drug Interactions Using Plated Human Cell Systems. Drug Metab. Dispos. 2012, 40, 1336–1344. [CrossRef] 30. Kim, J.-H.; Kwon, S.-S.; Jeong, H.-U.; Lee, H.S. Inhibitory Effects of Dimethyllirioresinol, Epimagnolin A, Eudesmin, Fargesin, and Magnolin on Cytochrome P450 Enzyme Activities in Human Liver Microsomes. Int. J. Mol. Sci. 2017, 18, 952. [CrossRef] 31. Park, E.J.; Park, R.; Jeon, J.-H.; Cho, Y.-Y.; Lee, J.Y.; Kang, H.C.; Song, I.-S.; Lee, H.S. Inhibitory Effect of AB-PINACA, Indazole Carboxamide Synthetic Cannabinoid, on Human Major Drug-Metabolizing Enzymes and Transporters. Pharmaceutics 2020, 12, 1036. [CrossRef][PubMed] 32. Kim, S.-B.; Kim, K.-S.; Kim, D.-D.; Yoon, I.-S. Metabolic interactions of rosmarinic acid with human cytochrome P450 monooxyge- nases and uridine diphosphate glucuronosyltransferases. Biomed. Pharmacother. 2019, 110, 111–117. [CrossRef][PubMed] 33. Amidon, G.L.; Lennernäs, H.; Shah, V.P.; Crison, J.R. A Theoretical Basis for a Biopharmaceutic Drug Classification: The Correlation of in Vitro Drug Product Dissolution and in Vivo Bioavailability. Pharm. Res. 1995, 12, 413–420. [CrossRef][PubMed] 34. Goon, C.P.; Wang, L.Z.; Wong, F.C.; Thuya, W.L.; Ho, P.C.-L.; Goh, B.C. UGT1A1 Mediated Drug Interactions and its Clinical Relevance. Curr. Drug Metab. 2016, 17, 100–106. [CrossRef][PubMed] 35. Fujiwara, R.; Sumida, K.; Kutsuno, Y.; Sakamoto, M.; Itoh, T. UDP-glucuronosyltransferase (UGT) 1A1 mainly contributes to the glucuronidation of trovafloxacin. Drug Metab. Pharmacokinet. 2015, 30, 82–88. [CrossRef] 36. Dellinger, R.W.; Garcia, A.M.G.; Meyskens, F.L. Differences in the Glucuronidation of Resveratrol and Pterostilbene: Altered Enzyme Specificity and Potential Gender Differences. Drug Metab. Pharmacokinet. 2014, 29, 112–119. [CrossRef] 37. Wang, L.-Z.; Ramírez, J.; Yeo, W.; Chan, M.-Y.M.; Thuya, W.-L.; Lau, J.-Y.A.; Wan, S.-C.; Wong, A.L.-A.; Zee, Y.-K.; Lim, R.; et al. Glucuronidation by UGT1A1 Is the Dominant Pathway of the Metabolic Disposition of Belinostat in Liver Cancer Patients. PLoS ONE 2013, 8, e54522. [CrossRef] 38. Reese, M.J.; Savina, P.M.; Generaux, G.T.; Tracey, H.; Humphreys, J.E.; Kanaoka, E.; Webster, L.O.; Harmon, K.A.; Clarke, J.D.; Polli, J.W. In Vitro Investigations into the Roles of Drug Transporters and Metabolizing Enzymes in the Disposition and Drug Interactions of Dolutegravir, a HIV Integrase Inhibitor. Drug Metab. Dispos. 2012, 41, 353–361. [CrossRef] 39. Alonen, A.; Finel, M.; Kostiainen, R. The human UDP-glucuronosyltransferase UGT1A3 is highly selective towards N2 in the tetrazole ring of losartan, candesartan, and zolarsartan. Biochem. Pharmacol. 2008, 76, 763–772. [CrossRef] 40. Erichsen, T.J.; Aehlen, A.; Ehmer, U.; Kalthoff, S.; Manns, M.P.; Strassburg, C.P. Regulation of the human bile acid UDP- glucuronosyltransferase 1A3 by the farnesoid X receptor and bile acids. J. Hepatol. 2010, 52, 570–578. [CrossRef] 41. Jeong, E.-S.; Kim, Y.-W.; Kim, H.-J.; Shin, H.-J.; Shin, J.-G.; Kim, K.H.; Chi, Y.H.; Paik, S.H.; Kim, D.-H. Glucuronidation of fimasartan, a new angiotensin receptor antagonist, is mainly mediated by UGT1A3. Xenobiotica 2014, 45, 10–18. [CrossRef] [PubMed] 42. Shen, Y.; Li, C.G.; Zhou, S.; Pang, E.; Story, D.; Xue, C.C. Chemistry and Bioactivity of Flos Magnoliae, A Chinese Herb for Rhinitis and Sinusitis. Curr. Med. Chem. 2008, 15, 1616–1627. [CrossRef][PubMed] 43. Kim, D.K.; Liu, K.H.; Jeong, J.H.; Ji, H.Y.; Oh, S.R.; Lee, H.K.; Lee, H.S. In vitro metabolism of magnolin and characterization of cytochrome P450 enzymes responsible for its metabolism in human liver microsomes. Xenobiotica 2011, 41, 358–371. [CrossRef] [PubMed]