Supplemental material to this article can be found at:

Transcription

1 Supplemental material to this article can be found at: X/42/11/ $ DRUG METABOLISM AND DISPOSITION Drug Metab Dispos 42: , November 2014 Copyright ª 2014 by The American Society for Pharmacology and Experimental Therapeutics In Vitro Assay of Six UDP-Glucuronosyltransferase Isoforms in Human Liver Microsomes, Using Cocktails of Probe Substrates and Liquid Chromatography Tandem Mass Spectrometry s Kyung-Ah Seo, Hyo-Ji Kim, Eun Sook Jeong, Nagi Abdalla, Chang-Soo Choi, Dong-Hyun Kim, and Jae-Gook Shin Department of Pharmacology and PharmacoGenomics Research Center, Inje University College of Medicine, Busan, Korea (K.-A.S., H.-J.K., E.S.J., N.A., D.-H.K., J.-G.S.); and Department of General Surgery, Inje University Busan Paik Hospital, Busan, Korea (C.-S.C.) Received May 8, 2014; accepted August 13, 2014 ABSTRACT UDP-glucuronosyltransferase (UGT) mediated drug drug interactions are commonly evaluated during drug development. We present a validated method for the simultaneous evaluation of drugmediated inhibition of six major UGT isoforms, developed in human liver microsomes through the use of pooled specific UGT probe substrates (cocktail assay) and rapid liquid chromatography tandem mass spectrometry (LC-MS/MS) analysis. The six probe substrates used in this assay were estradiol (UGT1A1), chenodeoxycholic acid (UGT1A3), trifluoperazine (UGT1A4), 4-hydroxyindole (UGT1A6), propofol (UGT1A9), and naloxone (UGT2B7). In a cocktail incubation, UGT1A1, UGT1A9, and UGT2B7 activities were substantially inhibited by other substrates. This interference could be eliminated Introduction Most pharmacokinetic drug drug interactions occur at the metabolic level and usually involve changes in the activity of the major drug-metabolizing enzymes. Identification of these enzymes allows us to predict potential drug drug interactions, which is critical for new drug development. Although cytochrome P450 (P450) enzymes are mainly responsible for the initial oxidative metabolism of xenobiotic compounds, a considerable number of drugs (approximately 15% of approved drugs on the market) are known to be metabolized by UDPglucuronosyltransferases (UGTs), either directly or after initial oxidative metabolism (Williams et al., 2004). Therefore, rapid and sensitive tools for in vitro evaluation of compound-mediated inhibition of UGT isoform activities, along with those for P450s, are required for studies of drug drug interactions in drug discovery. Several in vitro P450 cocktail methods have been developed, in which a mixture of several P450-selective substrates is included in a single human microsomal incubation, and the metabolism of the substrates is determined by liquid chromatography tandem mass spectrometry (LC-MS/MS) (Dixit et al., 2007; Pillai et al., 2013). Selective substrates, antibodies, or inhibitors of UGT isoforms can be This research was supported by the National Research Foundation of Korea [Grant R funded by the Korean Government]. dx.doi.org/ /dmd s This article has supplemental material available at dmd.aspetjournals.org. by dividing substrates into two incubations: one containing estradiol, trifluoperazine, and 4-hydroxyindole, and the other containing chenodeoxycholic acid, propofol, and naloxone. Incubation mixtures were pooled for the simultaneous analysis of glucuronyl conjugates in a single LC-MS/MS run. The optimized cocktail method was further validated against single-probe substrate assays using compounds known to inhibit UGTs. The degree of inhibition of UGT isoform activities by such known inhibitors in this cocktail assay was not substantially different from that in single-probe assays. This six-isoform cocktail assay may be very useful in assessing the UGT-based drug-interaction potential of candidates in a drug-discovery setting. employed in metabolism studies with human liver microsomes, and have been extremely useful in estimating the contribution of each UGT isoform to metabolism of the compound of interest (e.g., a new chemical entity) (Manevski et al., 2010). However most individual UGTs exhibit distinct but overlapping substrate selectivity and differ in their regulation of expression, their genetic polymorphism, and in other factors known to influence the activity of drug-metabolizing enzymes in humans (Lépine et al., 2004; Court, 2005; Itäaho et al., 2008). As a result, few selective substrates and inhibitors useful for phenotyping UGTs have been identified to date (Donato et al., 2010). Some UGT substrates have been used as probe drugs without proper validation, which can lead to biased study results (Hanioka et al., 2001). A cocktail method using multiple UGT substrates was recently developed for determining UGT activity in vitro (e.g., in human liver microsomes) (Gagez et al., 2012), but no validated method for measuring the inhibitory potential of a given compound on the major UGT enzymes has yet been reported. The purpose of this study was to develop a new cocktail method for simultaneous evaluation of the activities of six major human liver microsomal UGT isoforms (UGT1A1, UGT1A3, UGT1A4, UGT1A6, UGT1A9, and UGT2B7). We evaluated the specificity and sensitivity of each probe substrate and validated those substrates with specific UGT inhibitors. We explored the optimal experimental conditions to avoid potential interactions among the cocktail drugs, and developed an analytical method for cocktail experiments using LC-MS/MS. Downloaded from dmd.aspetjournals.org at ASPET Journals on March 8, 2019 ABBREVIATIONS: LC-MS/MS, liquid chromatography tandem mass spectrometry; MRM, multiple reaction monitoring; P450, cytochrome P450; SN-38, 7-ethyl-10-hydroxy-camptothecin; UDPGA, uridine 59-diphosphoglucuronic acid; UGT, UDP-glucuronosyltransferase. 1803

2 1804 Seo et al. Materials and Methods Chemicals and Reagents. Alamethicin (from Trichoderma viride), uridine 59-diphosphoglucuronic acid (UDPGA), 1-napthol, b-estradiol, bilirubin, chenodeoxycholic acid, fluconazole, hecogenin, lithocholic acid, naloxone, niflumic acid, propofol, trifluoperazine, troglitazone, and b-estradiol-3-b-dglucuronide were obtained from Sigma-Aldrich (St. Louis, MO). 4-Hydroxyindole and propofol glucuronide were obtained from Toronto Research Chemicals (North York, ON, Canada). Recombinant human UGT isoforms (UGT1A1, UGT1A3, UGT1A4, UGT1A6, UGT1A9, UGT2B4, UGT2B7, UGT2B15, and UGT2B17) and pooled human liver microsomes were purchased from BD Gentest Co. (Woburn, MA). High-performance liquid chromatography grade acetonitrile and methanol were purchased from J. T. Baker (Phillipsburg, NJ). All other chemicals were the highest analytical grade commercially available. Microsomal Incubations. The incubation mixtures consisted of 0.25 mg/ml of pooled human liver microsomes, 25 mg/ml alamethicin, 50 mm Tris-HCl (ph 7.5), 10 mm MgCl 2, and substrates (various UGT enzyme-specific substrates or a substrate cocktail set) in a total volume of 125 ml. After preincubation on ice for 15 minutes, reactions were initiated by the addition of 5 mm UDPGA and were incubated for 1 hour at 37 C in a shaking water bath. The reactions were terminated by the addition of 125 ml acetonitrile containing estrone glucuronide (2 mm, internal standard) and centrifuged at 10,000g for 5 minutes at 4 C. An aliquot of the supernatant was injected into LC-MS/MS for the determination of glucuronide conjugates. Selectivity Screening of Known UGT Isoform Substrates. Incubation mixtures containing 0.25 mg recombinant human UGTs, substrates, and 25 mg/ml alamethicin were reconstituted in 50 mm Tris-HCl (ph 7.5) and preincubated on ice for 15 minutes. The selective substrates were estradiol (10 mm) for UGT1A1, chenodeoxycholic acid (5 mm) for UGT1A3, trifluoperazine (10 mm) for UGT1A4, 4-hydroxyindole (10 mm) for UGT1A6, propofol (50 mm) for UGT1A9, and naloxone (250 mm) for UGT2B7. The concentration of each probe substrate was initially chosen near its K m value reported elsewhere (Supplemental Table 1). Under these conditions, drug interactions among substrates were observed in the cocktail incubation and their concentrations were reduced to one-half to one-quarter of their K m values to avoid such interactions. The final volume of the organic solvents in each incubation mixture was 1% (v/v). Reactions were initiated by adding 5 mm UDPGA, and were incubated for 1 hour at 37 C in a shaking water bath. Reactions were terminated by the addition of 125 ml acetonitrile containing estrone glucuronide (2 mm, internal standard) and centrifuged at 10,000g for 5 minutes at 4 C. Aliquots of the supernatants were analyzed by LC-MS/MS for the identification of the glucuronide metabolites. LC-MS/MS Analysis of Glucuronide Metabolites of Selective Substrates. LC-MS/MS analysis was performed on an API 4000 LC-MS/MS system (Applied Biosystems, Foster City, CA), coupled with an Agilent 1100 series highperformance liquid chromatography system (Agilent Technologies, Wilmington, DE). The separation was performed on a Synergi RP 80A column (2 150 mm, 4 mm; Phenomenex, Torrance, CA) using a mobile phase of 0.1% formic acid and acetonitrile (60:40, v/v). The flow rate was 0.2 ml/min. Electrospray ionization was performed in positive and negative ion modes with nitrogen as the nebulizing, turbo, and curtain gases with the optimum values set at 50, 50, and 30 (arbitrary units). The turbo ion spray interface was operated at 4500 V in the positive ion mode and at V in the negative ion mode. Multiple reaction monitoring (MRM) mode, using specific precursor/product ion transition, was employed for quantification. Detection of the positive ions was performed by monitoring the transitions of m/z for trifluoperazine glucuronide, for 4-hydroxyindole glucuronide, and for naloxone-3-glucuronide. Detection of the negative ions was performed by monitoring the transitions of m/z for estradiol- 3-glucuronide, for chenodeoxycholic acid glucuronide, for propofol glucuronide, and for the internal standard estrone glucuronide. Peak areas for all compounds were automatically integrated using Analyst software (version 1.4; Applied Biosystems). Concentrations of glucuronides that lacked reference compounds were estimated as molar equivalents, with respect to the calibration curve of the respective parent probe. Chemical Inhibition. The inhibitory effects of known UGT isoformselective inhibitors on the formation of probe-drug glucuronides were evaluated to identify the feasibility of the cocktail method for screening the inhibitory effects of test compounds. Inhibitors used in this study were as follows: bilirubin (50 mm) for UGT1A1, lithocholic acid (10 mm) for UGT1A3, hecogenin (5 mm) for UGT1A4, troglitazone (100 mm) for UGT1A6, niflumic acid (5 mm) for UGT1A9, and fluconazole (2.5 mm) for UGT2B4 and UGT2B7. The formation rates of probe-drug glucuronides were determined from reaction mixtures incubated in the presence or absence of inhibitors. With the exception of the addition of UGT isoform-specific inhibitors, all other incubation conditions were as described above. Data Analysis. In microsomal incubation studies, the apparent kinetic parameters of biotransformation (K m and V max ) were determined by fitting a one-enzyme Michaelis Menten: a substrate inhibition: or a Hill equation: V ¼ V max½sš K m þ½sš V max ½SŠ V ¼ 1 þ½sš K m þ½sš K si V ¼ V max½sš n ½SŠ n 50 þ½sšn The calculated parameters included the maximum rate of formation (V max ), substrate concentration at half-maximal rate (apparent K m or S 50 ), and the intrinsic clearance (CL int = V max /apparent K m or S 50 ). UGT-mediated activities in the presence of inhibitors were expressed as a percentage of the corresponding control values. A sigmoid curve was fitted to the data, and the enzyme inhibition parameter (IC 50 ) was calculated using a nonlinear least-squares regression analysis of the plot of percent control activity versus concentration of the test inhibitor. Calculations were performed using WinNonlin software (Pharsight, Mountain View, CA). The percentages of inhibition were calculated by the ratio of the amounts of metabolites formed, with and without the specific inhibitor. Results Glucuronidation of UGT Isoform-Selective Substrates. Our initial efforts were focused on the selection of six UGT isoform-specific substrates suitable for cocktail incubations. It is generally known that UGT isoforms show broad substrate specificity. The following probe substrates for each UGT isoform used in the cocktail assay were selected on the basis of previous reports and on our preliminary screening results: b-estradiol for UGT1A1, chenodeoxycholic acid for UGT1A3, trifluoperazine for UGT1A4, 4-hydroxyindole for UGT1A6, propofol for UGT1A9, and naloxone for UGT2B4/UGT2B7. A simultaneous analytical method using LC-MS/MS for six UGT isoformspecific probe metabolites and an internal standard was developed for the cocktail assay of UGT activity in human liver microsomes. The MRM transitions and optimized collision-induced dissociation conditions are described in Table 1. The specificity of the tandem mass spectrometer allowed a fast liquid chromatography gradient to be employed. The representative chromatograms for six probe metabolites in microsomal incubation mixtures are presented in Fig. 1. There was no interference from other substrates or metabolites at any of the retention times of interest for any metabolite MRM channel. In the case of b-estradiol, two glucuronides were observed in the microsomal incubation: one at a retention time of 3.23 was b-estradiol-3-glucuronide and the other at 3.82 was b-estradiol-17-glucuronide. The formation of b-estradiol-3-glucuronide is mediated by UGT1A1, whereas the formation of b-estradiol-17-glucuronide is mainly catalyzed by UGT2B7 (Alkharfy and Frye, 2002). The selectivity of each UGT substrate was evaluated using cdnaexpressed human UGT isoforms (Fig. 2). The concentration of each Downloaded from dmd.aspetjournals.org at ASPET Journals on March 8, 2019

3 In Vitro Cocktail Analysis for the Inhibition of Six UGTs 1805 TABLE 1 MRM parameters for the major metabolites of six UGT probe substrates UGT Isoform Metabolite Transition Polarity CE m/z ev UGT1A1 b-estradiol-3-glucuronide ES 2 50 UGT1A3 Chenodeoxycholic acid glucuronide ES 2 50 UGT1A4 Trifluoperazine glucuronide ES + 35 UGT1A6 4-Hydroxyindole glucuronide ES + 20 UGT1A9 Propofol glucuronide ES 2 35 UGT2B7 Naloxone-3-glucuronide ES + 30 Internal standard Estrone glucuronide ES 2 38 CE, collision energy; ES 2, electrospray negative ion mode; ES 1, electrospray positive ion mode. substrate was optimized to avoid interactions among probe substrates. The formation rate of b-estradiol-3-glucuronide by UGT1A1 was 11-fold greater than that by UGT1A3. Conversely, the formation of chenodeoxycholic acid glucuronide by UGT1A3 was 14-fold greater than that by UGT1A1. UGT2B7 showed minimal activity on chenodeoxycholic acid. 4-Hydroxyindole was glucuronidated mainly by UGT1A6, with minor activity by UGT1A9. Kinetic analysis also demonstrated that UGT1A1, UGT1A3, and UGT1A6 could play major roles in the glucuronidation of estradiol, chenodeoxycholic acid, and 4-hydroxyindole, respectively (Supplemental Figs. 1 3; Supplemental Table 2). The glucuronidation of trifluoperazine, propofol, and naloxone was almost exclusively catalyzed by UGT1A4, UGT1A9, and UGT2B7, respectively. Our results indicate that the UGT isoformselective targets used in this experiment are appropriate substrates, representing the corresponding UGT isoform activities, when incubated in a cocktail. Fig. 1. MRM chromatograms from the analysis of the major metabolites of UGT substrates. (A) Estradiol-3-glucuronide. (B) Chenodeoxycholic acid glucuronide. (C) Trifluoperazine glucuronide. (D) 4-Hydroxyindole glucuronide. (E) Propofol glucuronide. (F) Naloxone-3-glucuronide. cps, counts per second. Downloaded from dmd.aspetjournals.org at ASPET Journals on March 8, 2019

4 1806 Seo et al. Fig. 2. Representative plots of the formation of estradiol-3-glucuronide from estradiol (A), chenodeoxycholic acid glucuronide from chenodeoxycholic acid (B), trifluoperazine glucuronide from trifluoperazine (C), 4-hydroxyindole glucuronide from 4-hydroxyindole (D), propofol glucuronide from propofol (E), and naloxone-3- glucuronide from naloxone (F), by cdna-expressed human UGT isoforms. Activities shown are means of duplicate determinations from a single experiment. CDCA, chenodeoxycholic acid; TFP, trifluoperazine. Comparison of UGT Isoform Activities between Individual and Cocktail Incubations. Potential interactions among UGT substrates were evaluated during simultaneous incubations with human liver microsomes. The simultaneous incubation of six substrates with human liver microsomes showed glucuronidation activities different from those obtained with single individual incubations (Fig. 3A). The formation of estradiol-3-glucuronide, chenodeoxycholic acids, propofol glucuronide, and naloxone-3-glucuronide was inhibited by greater than 30% when six substrates were coincubated with microsomes. When pairs of substrates were incubated, an interaction between estradiol and propofol was observed. In the presence of estradiol, propofol glucuronidation catalyzed by UGT1A9 was reduced to approximately 50% of basal activity. Inhibition of UGT1A9 activity was independent of the concentration of propofol. When estradiol was replaced with the UGT1A1-selective substrate SN-38 (7-ethyl-10-hydroxy-camptothecin; Hanioka et al., 2001), SN-38 glucuronidation was inhibited by both trifluoperazine and naloxone. Naloxone glucuronidation was also inhibited by other UGT isoform-selective substrates. Replacement of naloxone with the UGT2B7-selective substrate efavirenz (Bae et al., 2011) inhibited the glucuronidation of trifluoperazine and propofol. When zidovudine was added as an UGT2B7-selective substrate (Barbier et al., 2000) to cocktail incubations, the glucuronidation of the drug was inhibited by estradiol (Supplemental Fig. 4). These results collectively indicated that simultaneous incubation of all six UGT isoform-selective substrates with human liver microsomes caused interactions among substrates that resulted in the inhibition of at least one or two UGT isoforms. Therefore, two cocktails of substrates were prepared for the microsomal incubation step. Cocktail A included estradiol, trifluoperazine, and 4-hydroxyindole, and cocktail B contained chenodeoxycholic acid, propofol, and naloxone. These mixtures were pooled after incubation and analyzed together by LC-MS/MS to reduce total assay time. As shown in Fig. 3B, each UGT isoform s activity was not substantially inhibited by other substrates within the cocktail sets except UGT1A3 (percent inhibition,20%). UGT1A3 activity was enhanced 1.3-fold over single-substrate incubations. Assay Validation Using UGT Isoform-Selective Inhibitors. The utility of this cocktail incubation as a screening tool for UGT inhibition was evaluated using known UGT inhibitors. The IC 50 value of each UGT isoform-selective inhibitor was determined in both individual and cocktail incubations. As shown in Fig. 4, the inhibition profile of each inhibitor was not substantially different between the two incubation methods, with the exception of that of lithocholic acid, a UGT1A3 inhibitor. The IC 50 values measured by the different approaches are summarized in Table 2. The IC 50 value of lithocholic acid for the formation of chenodeoxycholic acid glucuronide in single incubations was 2.6-fold lower than in cocktail incubations (Table 2). The effects of isoform-selective inhibitors on other UGT isoforms were also evaluated in cocktail incubations (Fig. 5). Bilirubin, a UGT1A1-selective inhibitor (Williams et al., 2002), resulted in greater inhibition of UGT1A1 activity compared with those of the activities of UGT1A4 or UGT1A6. Hecogenin, lithocholic acid, and niflumic acid demonstrated selective inhibition of UGT1A4, UGT1A3, and UGT1A9 activities, respectively, without affecting other isoform activities measured in cocktail incubations. Fluconazole inhibited UGT2B7 activity in a concentrationdependent manner up to 10 mm, although UGT1A3 and UGT1A9 activities were also inhibited by 40 and 21%, respectively, at 10 mm fluconazole. Troglitazone is reported to be a UGT1A6 inhibitor (Ito et al., 2001). However, this compound inhibited the activity of UGT1A1 and UGT1A4 to a greater extent than UGT1A6 in cocktail incubations. This was also observed in individual incubations with estradiol and trifluoperazine, suggesting that the inhibition observed in cocktail incubations was not due to substrate interactions. Troglitazone caused Downloaded from dmd.aspetjournals.org at ASPET Journals on March 8, 2019

5 In Vitro Cocktail Analysis for the Inhibition of Six UGTs 1807 Fig. 3. Effects of cocktail incubation on UGT isoform activities in human liver microsomes. Six substrates were incubated together (A) and six substrates were divided into two groups prior to incubation (B). Each bar represents the relative percentage of the activity assessed by individual incubation with estradiol for UGT1A1 (10 mm), chenodeoxycholic acid for UGT1A3 (5 mm), trifluoperazine for UGT1A4 (10 mm), 4-hydroxyindole for UGT1A6 (10 mm), propofol for UGT1A9 (50 mm), and naloxone for UGT2B7 (250 mm). Each activity shown is the mean of triplicate experiments. Each bar represents the mean 6 S.D. of triplicate determinations from a single experiment. greater inhibition of UGT1A6 activities when incubated with recombinant UGT1A6 instead of microsomes (Supplemental Fig. 5). Discussion During the early stages of drug development, knowledge of the metabolic characteristics of new drug candidates is very important in selecting lead compounds and in minimizing failures during clinical studies due to major kinetic problems, such as drug drug interactions. For this reason, several in vitro methods have been developed and are being used to study drug metabolism and metabolic interactions in the early phases of drug discovery and development (Pelkonen et al., 2005). The aim of this study was to develop a simple and rapid cocktail assay to simultaneously monitor the activity of hepatic UGT isoforms in human liver microsomes. The probe substrates for six human hepatic UGT isoforms were selected from the literature and from our own preliminary screening of their specificity for each isoform. The specificity of each substrate was evaluated using cdna-expressed UGTs. It is well known that UGTs exhibit partially distinct but frequently overlapping substrate specificities, which make it difficult to identify a selective substrate for each UGT isoform (Lépine et al., 2004). In addition, substrates selective for one UGT isoform often modulate the activities of other isoforms. Therefore, considerable efforts have been made to choose probe Downloaded from dmd.aspetjournals.org at ASPET Journals on March 8, 2019 Fig. 4. Effects of various inhibitors on UGT isoform activity in human liver microsomes in individual and cocktail incubations. UGT isoform-selective inhibitors used were bilirubin (A; 50 mm), lithocholic acid (B; 10 mm), hecogenin (C; 5 mm), troglitazone (D; 100 mm), niflumic acid (E; 5 mm), and fluconazole (F; 2.5 mm). Each bar represents the mean 6 S.D of triplicate determinations from a single experiment.

6 1808 Seo et al. TABLE 2 IC 50 values obtained in this study using individual substrate and substrate cocktails, and published IC 50 values for six UGT inhibitors UGT Isoform Substrate Inhibitor Individual Substrate IC 50 Reported IC 50 Reference Cocktail Substrate mm UGT1A1 b-estradiol Bilirubin and 30 Williams et al., 2002; Tachibana et al., 2005 UGT1A3 Chenodeoxycholic acid Lithocholic acid a UGT1A4 Trifluoperazine Hecogenin and 15 Uchaipichat et al., 2006; Edavana et al., 2013 UGT1A6 4-Hydroxyindole Troglitazone Ito et al., 2001 UGT1A9 Propofol Niflumic acid and 0.4 Mano et al., 2006; Miners et al., 2011 UGT2B7 Naloxone Fluconazole and 2500 Mano et al., 2007; Donato et al., 2010 a Data not available. substrates relatively specific for single UGT isoforms that do not interfere with other isoform activities. We found that trifluoperazine, propofol, and naloxone were almost exclusively glucuronidated by UGT1A4, UGT1A9, and UGT2B7, respectively, and these results are consistent with data reported elsewhere (Court, 2005; Di Marco et al., 2005; Uchaipichat et al., 2006). The formation of estradiol-3- glucuronide is mediated mainly by UGT1A1, whereas estradiol-17- glucuronide is generated by UGT2B7 (Alkharfy and Frye, 2002). Chenodeoxycholic acid is reported to be glucuronidated by UGT1A3 (Trottier et al., 2006). We also found that UGT1A1 was the major enzyme catalyzing the formation of estradiol-3-glucuronide. Although UGT1A3 also contributed to the glucuronidation of estradiol, its rate of formation was approximately 9% of that seen with UGT1A1. On the other hand, chenodeoxycholic acid was mainly glucuronidated by UGT1A3, with UGT1A1 catalyzing glucuronide formation at only 7% of the UGT1A3 rate. Fallon et al. (2013b) recently reported that the average protein level of UGT1A1 is 4.5-fold higher than that of UGT1A3 in human liver microsomes (36.2 versus 8.0 pmol/mg protein). As reported in the same study, BD supersomes expressed 2.6-fold more recombinant UGT1A1 than UGT1A3 (Fallon et al., 2013a). When intrinsic clearance values obtained from kinetic analyses (Supplemental Fig. 1; Supplemental Table 1) and relative ratios of expression are considered, the contributions of UGT1A3 to estradiol-3- glucuronide formation and UGT1A1 to chenodeoxycholic acid glucuronide formation in human liver microsomes were estimated to be 8.2% of UGT1A1 and 7.9% of UGT1A3, respectively. The relative contributions of UGT1A6 and UGT1A9 to the formation of 4-hydroxylindole were estimated to be 81.3 and 18.7%, respectively, Downloaded from dmd.aspetjournals.org at ASPET Journals on March 8, 2019 Fig. 5. The effects of UGT isoform-selective inhibitors on other UGT isoform activities in human liver microsomes in substrate cocktails. Bilirubin (A), hecogenin (B), troglitazone (C), lithocholic acid (D), niflumic acid (E), and fluconazole (F). Cocktail A contained estradiol, trifluoperazine, and 4-hydroxyindole, and cocktail B consisted of chenodeoxycholic acid, propofol, and naloxone. Data represent the mean 6 S.D. of triplicate determinations from a single experiment.

7 In Vitro Cocktail Analysis for the Inhibition of Six UGTs 1809 when calculated in the same way. These results collectively indicate that the substrates selected for this study were suitable as probe substrates for each UGT isoform. When relatively selective substrates for six hepatic UGTs (UGT1A1, UGT1A3, UGT1A4, UGT1A6, UGT1A9, and UGT2B7) were incubated in a cocktail assay, UGT1A1, UGT1A9, and UGT2B7 activities were substantially inhibited relative to those seen in individual incubations (Fig. 3A). Substrates affecting the activity of other UGT isoforms were initially identified by measuring activity in pairwise incubations. Identified substrates were then replaced by others reported to be isoform selective, as described in Results. However, crossinteractions among substrates could not be avoided, even after the replacement of estradiol with the UGT1A1-selective substrates SN-38 or etoposide, or exchanging naloxone for the UGT2B7 substrates efavirenz or zidovudine. Therefore, we employed a single LC-MS/MS analysis of incubation mixtures pooled from two separate microsomal incubations with substrates. Substrates were divided into two groups: cocktail A included estradiol, trifluoperazine, and 4-hydroxyindole, and cocktail B contained chenodeoxycholic acid, propofol, and naloxone. For five UGT isoforms (but not for UGT1A3), activities in these sets were similar to those observed in individual incubations. In group B incubations, propofol glucuronidation by UGT1A9 was increased 30% over individual incubations (Fig. 3B). This may be due to catalytic activation. However, this activation did not change the IC 50 value of niflumic acid, a known UGT1A9 inhibitor (Table 2). Although the availability of selective UGT inhibitors is currently limited, they represent the most powerful tool available for reaction phenotyping. The best known UGT inhibitors are hecogenin for UGT1A4 (Uchaipichat et al., 2006), niflumic acid for UGT1A9 (Mano et al., 2006), and fluconazole for UGT2B7 (Donato et al., 2010; Miners et al., 2010). Bilirubin and lithocholic acid are known to be substrates for UGT1A1 and UGT1A3, respectively. These compounds are also used for inhibition studies for UGT1A1 (Alkharfy and Frye, 2002; Soars et al., 2003) and UGT1A3 (Matern et al., 1984; Verreault et al., 2006). Our results demonstrated that hecogenin and niflumic acid resulted in strong and selective inhibition of UGT1A4 and UGT1A9, respectively, as expected (Fig. 4). Fluconazole was a moderately selective inhibitor; we found that it inhibited both UGT1A1 and UGT2B7. Bilirubin inhibited UGT1A1 activity, but also weakly inhibited UGT1A4 activity. Troglitazone was chosen as a UGT1A6 inhibitor based on a report that it inhibited recombinant UGT1A6-mediated 1-naphthol glucuronidation with an IC 50 of 28 mm (Hanioka et al., 2001). However, troglitazone did not inhibit 1-naphthol glucuronidation under our experimental conditions, and had no effect on the glucuronidation of 4-hydroxyindole, another reaction mediated by UGT1A6 in microsomal incubations (data not shown). Unexpectedly, UGT1A1 and UGT1A4 were inhibited by troglitazone, with IC 50 values less than 10 mm. This discrepancy may be due to use of enzymes from different sources. With our cdna-expressed human UGT1A6, troglitazone inhibited the glucuronidation of 4-hydroxyindole and 1-naphtol, consistent with the results of Hanioka et al. (2001). These results suggest that recombinant UGTs may not be suitable for evaluating the inhibition potential of chemicals, particularly in the case of UGT1A6. We found that all inhibitors tested showed similar inhibition profiles with both individual substrates and substrate cocktails (Fig. 4). The IC 50 values of the selective UGT inhibitors determined using the substrate cocktails were in good agreement with those determined using individual substrates, and were comparable to those reported by other groups (Table 2). This suggests that the inhibitory potential of test compounds can be accurately determined using our cocktail assay, rather than individual substrate incubations. In conclusion, a method was developed for high-throughput inhibition screening of the major human hepatic UGT enzymes (UGT1A1, UGT1A3, UGT1A4, UGT1A6, UGT1A9, and UGT2B7) using in vitro substrate cocktails and LC-MS/MS analysis. Probe substrates were selected after evaluation of isoform selectivity to minimize possible interference by other UGT isoforms. Six substrates divided into two cocktails for incubation and pooled for analysis in a single run allowed us to evaluate the activity of six UGT isoforms without cross-interference. With known UGT isoform-selective inhibitors, this cocktail assay produced similar inhibition profiles to those obtained from singlesubstrate incubations, suggesting that this assay can be a useful tool for rapid screening of UGT inhibition and for the prediction of clinical drug interactions. Authorship Contributions Participated in research design: D.-H. Kim, Shin. Conducted experiments: Seo, H.-J. Kim, Jeong, Abdalla. Performed data analysis: Seo, H.-J. Kim, Jeong, Choi, D.-H. Kim, Shin. Wrote or contributed to the writing of the manuscript: Seo, D.-H. Kim, Shin. References Alkharfy KM and Frye RF (2002) Sensitive liquid chromatographic method using fluorescence detection for the determination of estradiol 3- and 17-glucuronides in rat and human liver microsomal incubations: formation kinetics. J Chromatogr B Analyt Technol Biomed Life Sci 774: Bae SK, Jeong YJ, Lee C, and Liu KH (2011) Identification of human UGT isoforms responsible for glucuronidation of efavirenz and its three hydroxy metabolites. Xenobiotica 41: Barbier O, Turgeon D, Girard C, Green MD, Tephly TR, Hum DW, and Bélanger A (2000) 39-azido-39-deoxythimidine (AZT) is glucuronidated by human UDP-glucuronosyltransferase 2B7 (UGT2B7). Drug Metab Dispos 28: Court MH (2005) Isoform-selective probe substrates for in vitro studies of human UDPglucuronosyltransferases. Methods Enzymol 400: Di Marco A, D Antoni M, Attaccalite S, Carotenuto P, and Laufer R (2005) Determination of drug glucuronidation and UDP-glucuronosyltransferase selectivity using a 96-well radiometric assay. Drug Metab Dispos 33: Dixit V, Hariparsad N, Desai P, and Unadkat JD (2007) In vitro LC-MS cocktail assays to simultaneously determine human cytochrome P450 activities. Biopharm Drug Dispos 28: Donato MT, Montero S, Castell JV, Gómez-Lechón MJ, and Lahoz A (2010) Validated assay for studying activity profiles of human liver UGTs after drug exposure: inhibition and induction studies. Anal Bioanal Chem 396: Edavana VK, Dhakal IB, Williams S, Penney R, Boysen G, Yao-Borengasser A, and Kadlubar S (2013) Potential role of UGT1A4 promoter SNPs in anastrozole pharmacogenomics. Drug Metab Dispos 41: Fallon JK, Neubert H, Goosen TC, and Smith PC (2013a) Targeted precise quantification of 12 human recombinant uridine-diphosphate glucuronosyl transferase 1A and 2B isoforms using nano-ultra-high-performance liquid chromatography/tandem mass spectrometry with selected reaction monitoring. Drug Metab Dispos 41: Fallon JK, Neubert H, Hyland R, Goosen TC, and Smith PC (2013b) Targeted quantitative proteomics for the analysis of 14 UGT1As and -2Bs in human liver using NanoUPLC-MS/MS with selected reaction monitoring. J Proteome Res 12: Gagez AL, Rouguieg-Malki K, Sauvage FL, Marquet P, and Picard N (2012) Simultaneous evaluation of six human glucuronidation activities in liver microsomes using liquid chromatography-tandem mass spectrometry. Anal Biochem 427: Hanioka N, Ozawa S, Jinno H, Ando M, Saito Y, and Sawada J (2001) Human liver UDPglucuronosyltransferase isoforms involved in the glucuronidation of 7-ethyl-10-hydroxycamptothecin. Xenobiotica 31: Itäaho K, Mackenzie PI, Ikushiro S, Miners JO, and Finel M (2008) The configuration of the 17-hydroxy group variably influences the glucuronidation of beta-estradiol and epiestradiol by human UDP-glucuronosyltransferases. Drug Metab Dispos 36: Ito M, Yamamoto K, Sato H, Fujiyama Y, and Bamba T (2001) Inhibitory effect of troglitazone on glucuronidation catalyzed by human uridine diphosphate-glucuronosyltransferase 1A6. Eur J Clin Pharmacol 56: Lépine J, Bernard O, Plante M, Têtu B, Pelletier G, Labrie F, Bélanger A, and Guillemette C (2004) Specificity and regioselectivity of the conjugation of estradiol, estrone, and their catecholestrogen and methoxyestrogen metabolites by human uridine diphospho-glucuronosyltransferases expressed in endometrium. J Clin Endocrinol Metab 89: Manevski N, Kurkela M, Höglund C, Mauriala T, Court MH, Yli-Kauhaluoma J, and Finel M (2010) Glucuronidation of psilocin and 4-hydroxyindole by the human UDP-glucuronosyltransferases. Drug Metab Dispos 38: Mano Y, Usui T, and Kamimura H (2006) In vitro inhibitory effects of non-steroidal antiinflammatory drugs on 4-methylumbelliferone glucuronidation in recombinant human UDPglucuronosyltransferase 1A9 potent inhibition by niflumic acid. Biopharm Drug Dispos 27: 1 6. Mano Y, Usui T, and Kamimura H (2007) Comparison of inhibition potentials of drugs against zidovudine glucuronidation in rat hepatocytes and liver microsomes. Drug Metab Dispos 35: Matern S, Matern H, Farthmann EH, and Gerok W (1984) Hepatic and extrahepatic glucuronidation of bile acids in man. Characterization of bile acid uridine 59-diphosphateglucuronosyltransferase in hepatic, renal, and intestinal microsomes. J Clin Invest 74: Downloaded from dmd.aspetjournals.org at ASPET Journals on March 8, 2019

8 1810 Seo et al. Miners JO, Bowalgaha K, Elliot DJ, Baranczewski P, and Knights KM (2011) Characterization of niflumic acid as a selective inhibitor of human liver microsomal UDP-glucuronosyltransferase 1A9: application to the reaction phenotyping of acetaminophen glucuronidation. Drug Metab Dispos 39: Miners JO, Mackenzie PI, and Knights KM (2010) The prediction of drug-glucuronidation parameters in humans: UDP-glucuronosyltransferase enzyme-selective substrate and inhibitor probes for reaction phenotyping and in vitro-in vivo extrapolation of drug clearance and drugdrug interaction potential. Drug Metab Rev 42: Pelkonen O, Turpeinen M, Uusitalo J, Rautio A, and Raunio H (2005) Prediction of drug metabolism and interactions on the basis of in vitro investigations. Basic Clin Pharmacol Toxicol 96: Pillai VC, Strom SC, Caritis SN, and Venkataramanan R (2013) A sensitive and specific CYP cocktail assay for the simultaneous assessment of human cytochrome P450 activities in primary cultures of human hepatocytes using LC-MS/MS. J Pharm Biomed Anal 74: Soars MG, Ring BJ, and Wrighton SA (2003) The effect of incubation conditions on the enzyme kinetics of udp-glucuronosyltransferases. Drug Metab Dispos 31: Tachibana M, Tanaka M, Masubuchi Y, and Horie T (2005) Acyl glucuronidation of fluoroquinolone antibiotics by the UDP-glucuronosyltransferase 1A subfamily in human liver microsomes. Drug Metab Dispos 33: Trottier J, Verreault M, Grepper S, Monté D, Bélanger J, Kaeding J, Caron P, Inaba TT, and Barbier O (2006) Human UDP-glucuronosyltransferase (UGT)1A3 enzyme conjugates chenodeoxycholic acid in the liver. Hepatology 44: Uchaipichat V, Mackenzie PI, Elliot DJ, and Miners JO (2006) Selectivity of substrate (trifluoperazine) and inhibitor (amitriptyline, androsterone, canrenoic acid, hecogenin, phenylbutazone, quinidine, quinine, and sulfinpyrazone) probes for human udp-glucuronosyltransferases. Drug Metab Dispos 34: Verreault M, Senekeo-Effenberger K, Trottier J, Bonzo JA, Bélanger J, Kaeding J, Staels B, Caron P, Tukey RH, and Barbier O (2006) The liver X-receptor alpha controls hepatic expression of the human bile acid-glucuronidating UGT1A3 enzyme in human cells and transgenic mice. Hepatology 44: Williams JA, Hyland R, Jones BC, Smith DA, Hurst S, Goosen TC, Peterkin V, Koup JR, and Ball SE (2004) Drug-drug interactions for UDP-glucuronosyltransferase substrates: a pharmacokinetic explanation for typically observed low exposure (AUCi/AUC) ratios. Drug Metab Dispos 32: Williams JA, Ring BJ, Cantrell VE, Campanale K, Jones DR, Hall SD, and Wrighton SA (2002) Differential modulation of UDP-glucuronosyltransferase 1A1 (UGT1A1)-catalyzed estradiol-3- glucuronidation by the addition of UGT1A1 substrates and other compounds to human liver microsomes. Drug Metab Dispos 30: Address correspondence to: Dr. Jae-Gook Shin or Dr. Dong-Hyun Kim, Department of Pharmacology and PharmacoGenomics Research Center, Inje University College of Medicine, Gaegum-Dong, Busanjin-Gu, Busan , Korea. phshinjg@inje.ac.kr or dhkim@inje.ac.kr Downloaded from dmd.aspetjournals.org at ASPET Journals on March 8, 2019

9 Supplemental data for Drug Metabolism and Disposition In vitro assay of six UGT isoforms in human liver microsomes, using cocktails of probe substrates and LC-MS/MS Kyung-Ah Seo, Hyo-Ji Kim, Eun Sook Jeong, Nagi Abdalla, Chang-Soo Choi, Dong Hyun Kim, and Jae-Gook Shin Department of Pharmacology and PharmacoGenomics Research Center, Inje University College of Medicine, Busan, Korea (K.-A.S., H.-J.K., E.S.J., N.A., D.H.K, and J.-G.S) Department of General Surgery, Inje University Busan Paik Hospital, Busan, Korea (C.-S. Choi)

10 Supplemental Table 1. Published K m values for six UGT substrates. UGT isoform Substrates Recombinant Reported K m (µm) HLMs References UGT1A1 β-estradiol (Soars et al., 2003, Alkharfy and Frye, 2002) UGT1A3 Chenodeoxycholic acid (Trottier et al., 2006, Gagez et al., 2012, Matern.S. et al., 1984) UGT1A4 Trifluoperazine (Uchaipichat et al., 2006, Gagez et al., 2012) UGT1A6 4-Hydroxyindole (Manevski et al., 2010), in house UGT1A9 Propofol UGT2B7 Naloxone (Liang et al., 2011, Fujiwara et al., 2007, Soars et al., 2001, Soars et al., 2003) (Coffman et al., 1998, Donato et al., 2010, Soars et al., 2001)

11 Supplemental Table 2. Kinetic parameters for the glucuronide conjugates mediated by human cdna-expressed UGT isoforms. Kinetic Parameters Substrate UGT isoforms K m V max CL int (V max /K m ) (µm) (nmol/min/mg protein) (µl/min/mg protein) Hill coefficient (n) Estradiol Chenodeoxycholic acid 4-Hydroxyindole UGT1A UGT1A UGT1A UGT1A UGT1A UGT1A

12 (A) (B) Supplemental Figure 1. Kinetics for the formation of estradiol 3-glucuronide from estradiol in human cdna-expressed UGT1A1 and 1A3 (A) and the respective corresponding Eadie- Hofstee plots (B). An increasing concentration of estradiol was incubated with each recombinant UGT and UDPGA at 37 o C for 30 min. The kinetic data were fitted by Michaelis- Menten equation for UGT1A3 and a Hill equation for UGT1A1. Each data point represents the mean ± SD of triplicate determinations.

13 (A) (B) Supplemental Figure 2. Kinetics for the formation of chenodeoxycholic acid glucuronide from chenodeoxycholic acid in human cdna-expressed UGT1A1 and 1A3 (A) and the respective corresponding Eadie-Hofstee plots (B). Increasing concentrations of chenodeoxycholic acid was incubated with each recombinant UGT and UDPGA at 37 o C for 30 min. The kinetic data were fitted by Michaelis-Menten equation for UGT1A3 and a Hill equation for UGT1A1. Each data point represents the mean ± SD of triplicate determinations.

14 (A) (B) Supplemental Figure 3. Kinetics for the formation of 4-hydroxyindole glucuronide from 4- hydroxyindole in human cdna-expressed UGT1A6 and 1A9 (A) and the respective corresponding Eadie-Hofstee plots (B). Increasing concentrations of 4-hydroxyindole was incubated with each recombinant UGT and UDPGA at 37 o C for 30 min. The kinetic data were fitted by Michaelis-Menten equation for UGT1A6 and a Hill equation for UGT1A9. Each data point represents the mean ± SD of triplicate determinations.

15 Supplemental Figure 4. Inhibition of UGT isoform selective activities by various substrates in human liver microsomes: propofol glucuronidation by estradiol (A), SN-38 glucuronidation by trifluoperazine and naloxone (B), trifluoperazine and propofol glucuronidation by efavirenz (C), and zidovudine (AZT) glucuronidation by estradiol and trifluoperazine (D). Each data point represents the mean of duplicate determinations.

16 120 Relative activity (%) Hydroxyindole glucuronidation 1-Napthol glucuronidation Hydroxyindole (10 + Troglitazone (100 M) 1-Napthol (20 + Troglitazone (100 M) Supplemental Figure 5. Inhibition of UGT1A6-mediated 4-hydroxyindole and 1-napthol glucuronidation by troglitazone in human liver microsomes (black) and recombinant UGT1A6 (grey). Each data point represents the mean of duplicate determinations.

17 References Alkharfy KM and Frye RF. (2002) Sensitive liquid chromatographic method using fluorescence detection for the determination of estradiol 3- and 17-glucuronides in rat and human liver microsomal incubations: formation kinetics. J Chromatogr B Analyt Technol Biomed Life Sci 774: Coffman BL, King CD, Rios GR, and Tephly TR. (1998) The glucuronidation of opioids, other xenobiotics, and androgens by human UGT2B7Y(268) and UGT2B7H(268). Drug Metab Dispos 26: Donato MT, Montero S, Castell JV, Gomez-Lechon MJ, and Lahoz A. (2010) Validated assay for studying activity profiles of human liver UGTs after drug exposure: inhibition and induction studies. Anal Bioanal Chem 396: Fujiwara R, Nakajima M, Yamanaka H, Nakamura A, Katoh M, Ikushiro S, Sakaki T, and Yokoi T. (2007) Effects of coexpression of UGT1A9 on enzymatic activities of human UGT1A isoforms. Drug Metab Dispos 35: Gagez AL, Rouguieg-Malki K, Sauvage FL, Marquet P, and Picard N. (2012) Simultaneous evaluation of six human glucuronidation activities in liver microsomes using liquid chromatography-tandem mass spectrometry. Anal Biochem 427: Liang SC, Ge GB, Liu HX, Shang HT, Wei H, Fang ZZ, Zhu LL, Mao YX, and Yang L. (2011) Determination of propofol UDP-glucuronosyltransferase (UGT) activities in hepatic microsomes from different species by UFLC-ESI-MS. J Pharm Biomed Anal 54: Manevski N, Kurkela M, Hoglund C, Mauriala T, Court MH, Yli-Kauhaluoma J, and Finel M. (2010) Glucuronidation of psilocin and 4-hydroxyindole by the human UDPglucuronosyltransferases. Drug Metab Dispos 38: Matern.S., Matern.H., Farthmann.E.H., and Gerok.W. (1984) Hepatic and extrahepatic glucuronidation of bile acids in man. Characterization of bile acid uridine 5'-diphosphateglucuronosyltransferase in hepatic, renal, and intestinal microsomes. J Clin Invest 74: Soars MG, Ring BJ, and Wrighton SA. (2003) The effect of incubation conditions on the enzyme kinetics of udp-glucuronosyltransferases. Drug Metab Dispos 31: Soars MG, Riley RJ, Findlay KA, Coffey MJ, and Burchell B. (2001) Evidence for significant differences in microsomal drug glucuronidation by canine and human liver and kidney. Drug Metab Dispos 29:

18 Trottier J, Verreault M, Grepper S, Monte D, Belanger J, Kaeding J, Caron P, Inaba TT, and Barbier O. (2006) Human UDP-glucuronosyltransferase (UGT)1A3 enzyme conjugates chenodeoxycholic acid in the liver. Hepatology 44: Uchaipichat V, Mackenzie PI, Elliot DJ, and Miners JO. (2006) Selectivity of substrate (trifluoperazine) and inhibitor (amitriptyline, androsterone, canrenoic acid, hecogenin, phenylbutazone, quinidine, quinine, and sulfinpyrazone) "probes" for human udpglucuronosyltransferases. Drug Metab Dispos 34: