1604 Biol. Pharm. Bull. 39, 1604–1610 (2016) Vol. 39, No. 10 Regular Article

Acyl-glucuronide as a Possible Cause of Trovafloxacin-Induced Liver Toxicity: Induction of Chemokine (C-X-C Motif) Ligand 2 by Trovafloxacin Acyl-glucuronide Ryo Mitsugi,a Kyohei Sumida,a Yoshiko Fujie, a Robert H. Tukey,b Tomoo Itoh, a and Ryoichi Fujiwara*,a a Department of Pharmaceutics, School of Pharmacy, Kitasato University; 5–9–1 Shirokane, Minato-ku, Tokyo 108–8641, Japan: and b Laboratory of Environmental Toxicology, Departments of Chemistry & Biochemistry and Pharmacology, University of California at San Diego; La Jolla, CA 92023, U.S.A. Received February 27, 2016; accepted July 11, 2016

Trovafloxacin is an antibiotic that was withdrawn from the market relatively soon after its release due to the risk of hepatotoxicity. Trovafloxacin is mainly metabolized to its acyl-glucuronide by uridine 5 -diphos- phate (UDP)–glucuronosyltransferase (UGT) 1A1. In this study, we examined whether the acyl-glucuronide is involved in the development of hepatotoxicity. A UGT1A1-induced cell model was developed and the toxicity of trovafloxacin acyl-glucuronide was evaluated. The UGT1A1-induced cell model was developed by treating HepG2 cells with chrysin for 48 h. Chemokine (C-X-C motif) ligand 2, a cytokine involved in drug-induced liver injury, was uniquely induced by trovafloxacin in the UGT1A1-induced HepG2 cells. Induction of UGT1A1 resulted in a decrease in cell viability. An in vivo animal study further demonstrated the impor- tance of UGT1A1 in the trovafloxacin-induced liver toxicity. Although the complete mechanism of trovaflox- acin-induced liver injury is still unknown, trovafloxacin acyl-glucuronide can be involved in the development of toxic reactions in vitro and in vivo. Key words trovafloxacin; drug-induced liver injury; acyl-glucuronide; chrysin; HepG2 cell

Uridine 5′-diphosphate (UDP)–glucuronosyltransferases cyto- and genotoxicity.12) Microarray expression analysis is a (UGTs; EC 2.4.1.17) are a family of membrane-bound en- promising tool to identify genes associated with a drug treat- zymes that catalyze glucuronidation of endogenous and exog- ment. To identify genes that were associated with the trova- enous compounds by transferring the glucuronic acid moiety floxacin-induced liver toxicity, several research groups carried of UDP-glucuronic acid to the substrates.1) Human UGTs are out the microarray expression analysis in human hepatocytes, mainly divided into two distinct families, UGT1 and UGT2, mice, and rats.13–15) The group of genes that were specifically on the basis of evolutionary divergence and homology.2) The induced by the trovafloxacin treatment included topoisomerase UGT1 gene is located on chromosome 2q37 and produces nine I (TOP1), B-cell leukemia/lymphoma 2 (BCL-2)-associated functional enzymes, UGT1A1, UGT1A3, UGT1A4, UGT1A5, transcription factor 1 (BCLAF), Mitofusin1 (MFN1), Metallo- UGT1A6, UGT1A7, UGT1A8, UGT1A9, and UGT1A10, by thionein (MT) 2A, MT1H, and MT1X.13) Although these genes exon sharing.3) The unique first exons encode N-terminal do- might have been induced by the acyl-glucuronide in the hepa- main and the common exons 2 to 5 encode C-terminal domain tocytes, there still was a possibility that the parent compound of UGT1A proteins. Since all of UGTs recognize and utilize itself was involved in the induction of the genes. This was UDP-glucuronic acid as a co-substrate, it has been suggested because a certain amount of trovafloxacin still remained in the that the C-terminal domain is responsible for the co-substrate body even 24 h after the oral and intravenous administration binding. In contrast, UGT1A proteins exhibit overlapping but of trovafloxacin.16–18) distinct substrate specificities, suggesting that the N-terminal Previously, we determined UGT1A1 as the main UGT iso- domain is responsible for the substrate binding. While the form responsible for trovafloxacin acyl-glucuronidation.19) To liver is the most contributing tissue to the metabolism,4) recent investigate whether trovafloxacin acyl-glucuronide is involved findings suggest that extrahepatic tissues such as small intes- in trovafloxacin-induced liver injury, in the present study, a tine play an important role in glucuronidation of endogenous UGT1A1-induced cell model was developed and the toxicity and exogenous compounds.5,6) of trovafloxacin acyl-glucuronide was evaluated. We further Trovafloxacin is an antibiotic that was released on the mar- employed Ugt1-knockout mice to examine the importance of ket in 1998.7) This promising agent was withdrawn from the trovafloxacin acyl-glucuronide in the trovafloxacin-induced market relatively soon after its release due to the risk of hepa- liver injury in vivo. totoxicity including acute liver failure. Trovafloxacin is mainly metabolized by UGTs to its acyl-glucuronide in humans.8) MATERIALS AND METHODS While glucuronides are usually pharmacologically inactive, certain types of glucuronide, especially acyl-glucuronide, Chemicals and Reagents UDP-glucuronic acid can exhibit an increased reactivity compared to the parent (UDPGA), alamethicin, chrysin, estradiol, and estradiol 3-O- compounds.9) Acyl-glucuronide-associated toxicity has been glucuronide were purchased from Sigma-Aldrich (St. Louis, reported in vivo and in vitro10,11); however, there are also re- MO, U.S.A.). Trovafloxacin was purchased from Wako Pure ports showing that acyl-glucuronidation did not induce the Chemical Industries, Ltd. (Osaka, Japan). Recombinant human

* To whom correspondence should be addressed. e-mail: [email protected] © 2016 The Pharmaceutical Society of Japan Vol. 39, No. 10 (2016) Biol. Pharm. Bull. 1605 tumor necrosis factor α (TNF-α) was purchased from Roche for 24 h. Forty-eight hours after changing the culture medium (Mannheim, Germany). Primers were commercially synthe- to a normal or a chrysin-containing DMEM medium, the cells sized at Life Technologies (Carlsbad, CA, U.S.A.). All other were co-treated with trovafloxacin (50 µM) and TNF-α (4 ng/ chemicals and solvents were of analytical grade or the highest mL) for 24 h. Cell viability was measured using a 3-(4,5-di- grade commercially available. methylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) Cell Culture and Chemical Treatments The human Cell Counting Kit (Nacalai Tesque, Kyoto, Japan) according to hepatoma HepG2 cells were obtained from DS Pharma Bio- the manufacturer’s protocol. medical Co., Ltd. (Osaka, Japan). HepG2 cells were grown Animals Heterozygous Ugt1 (Ugt1+/−) mice23) and in Dulbecco’s modified Eagle’s medium (DMEM) containing C57BL/6NCrSlc mice were used to obtain wild type (Ugt1+/+), 100 U/mL penicillin, 100 µg/mL streptomycin, and 10% fetal Ugt1+/−, and UGT1 knockout (Ugt1−/−) mice. Genomic DNA bovine serum (FBS) and were maintained at 37°C in a humid- was isolated from tail biopsies and was used as a template for 24) ified atmosphere containing 5% of CO2. Before the treatment, genotyping PCR. All animals received food and water ad li- HepG2 cells were seeded into six-well plates at 5×105 cells/ bitum, and mouse handling and experimental procedures were well. After 24 h, the culture medium was changed to a normal conducted in accordance with our animal care protocol, which or a chrysin-containing DMEM medium and subsequently was previously approved by Kitasato University. cells were maintained for 48 h until harvesting. RNA was iso- Two-day-old mice were subcutaneously treated with tro- lated from the cells and was used for the quantitative-reverse- vafloxacin (150 mg/kg) or canola oil. Three hours after the transcription PCR (Q-PCR) analysis. The microsomal fraction injections, mice were subcutaneously treated with lipopolysac- was also obtained from the cells. Control and the UGT1A1- charide (LPS) (5 mg/kg) or saline. Nine hours after the second induced HepG2 cells were further treated with trovafloxacin treatment, blood was obtained from the submandibular vein (50 µM) for 24 h. and serum was prepared. Serum alanine aminotransferase Q-PCR Analysis cDNA was synthesized from total RNA (ALT) levels were determined using a Transaminase CII-test using ReverTra Ace qPCR RT Master Mix (Toyobo, Tokyo, Wako kit (Wako Pure Chemical Industries, Ltd.) according to Japan) according to the manufacturer’s protocol. Q-PCR was the manufacturer’s protocol. performed with THUNDERBIRD SYBR qPCR Mix (Toyobo), Statistical Analysis Significant differences of UGT1A1 and the reactions were run in a CFX96 Real-Time PCR Detec- expression in HepG2 cells were analyzed by Dunnett’s test. tion System (Bio-Rad, Hercules, CA, U.S.A.). Expression of The Tukey–Kramer test was used to determine the signifi- glyceraldehyde 3-phosphate dehydrogenase (GAPDH) mRNA cance in the gene expression studies. The unpaired t-test was was used as an internal control for the cDNA quantity and used to determine the significance in the enzyme and cell vi- quality. Primer pairs that can detect UGT1A1 and GAPDH ability assays. p<0.05 was considered as significant. were reported previously.20,21) Other primers used were newly established with a program, Primer Blast (National Institutes RESULTS of Health). We confirmed that all of the primer sets produced specific bands and that bands were not detected when the PCR Development of the UGT1A1-Induced Cell Model reaction was conducted without cDNA. After an initial dena- Chrysin, phenytoin, carbamazepine, and phenobarbital can in- turation at 95°C for 30 s, the amplification was performed by duce UGT1A1 in vitro and in vivo.25–27) To determine the most denaturation at 95°C for 5 s, annealing at an appropriate tem- potent UGT1A1 inducer in HepG2 cells, we preliminarily perature for 30 s, and extension at 72°C for 30 s for 45 cycles. conducted a cell-based induction assay.28) HepG2 cells were Enzyme Assay in Vitro and in Cells Microsomes were treated with a lower (10 µM) and a higher (50 µM) concentra- prepared as described before.22) Estradiol 3-O- and trova- tion of chrysin, phenytoin, carbamazepine, and phenobarbital. floxacin acyl-glucuronidations were determined according to Forty-eight hours after the treatment, RNA was isolated and previously reported methods with slight modifications.19,22) a Q-PCR analysis was carried out to determine the UGT1A1 Briefly, a typical incubation mixture (200 µL of total vol- expression level in the cells. We observed that phenytoin, car- ume) contained 50 mM Tris–HCl (pH 7.4), 4 mM MgCl2, 2 mM bamazepine, and phenobarbital moderately induced UGT1A1 UDPGA, 50 µg/mL alamethicin, 1.5 mg/mL microsomes, and in HepG2 cells. In contrast, it was demonstrated that chry- 10 µM estradiol or 50 µM trovafloxacin. The reaction was initi- sin significantly induced UGT1A1. In a subsequently per- ated by the addition of UDPGA after a 3-min preincubation at formed study, chrysin concentration-dependently induced the 37°C. After incubation at 37°C for 90 min for estradiol gluc- UGT1A1 mRNA expression in HepG2 cells (Fig. 1A). The uronidation and 60 min for trovafloxacin glucuronidation, the highest expression of UGT1A1 was observed when the cells reaction was terminated by adding 200 µL of cold methanol. were treated with 50-µM chrysin. When HepG2 cells were After removal of the protein by centrifugation at 12000×g treated with even higher concentrations of chrysin such as for 5 min, supernatant was subjected to HPLC to quantitate 80 and 100 µM, slight and moderate cytotoxicity was induced, estradiol 3-O-glucuronide and trovafloxacin acyl-glucuronide. respectively. It was further demonstrated that the microsomes 19,22) The conditions of the HPLC analysis were reported before. prepared from the HepG2 cells treated with 50-µM chrysin HepG2 cells were treated with 100 µM estradiol. Seven exhibited an 8-fold higher estradiol 3-O-glucuronidation activ- hours after incubation at 37°C, a portion of the cell-culturing ity (Fig. 1B). The estradiol 3-O-glucuronidations in the cells media was collected and mixed with the same amount of were also determined in the absence of additional UDPGA. acetonitrile. After removal of proteins by centrifugation at Estradiol was added into the cell-culturing media and was in- 12000×g for 5 min, supernatant was subjected to HPLC. cubated for 7 h. Estradiol 3-O-glucuronide was detected in the Cell Viability Assay HepG2 cells were seeded into 96- cell-culturing medium of HepG2 cells. The amount of gluc- well plates at 1×104 cells/well and were maintained at 37°C uronide was three-fold greater in the medium of the UGT1A1- 1606 Biol. Pharm. Bull. Vol. 39, No. 10 (2016)

Fig. 1. Induction of UGT1A1 mRNA and Activity in HepG2 Cells

(A) HepG2 cells were treated with chrysin (20–100 µM) for 48 h. Total RNA was isolated from the cells and the UGT1A1 mRNA expression was determined by Q-PCR. (B) HepG2 cells were treated with chrysin (50 µM) for 48 h. Microsomes were prepared and the UGT1A1 activity was determined using the UGT1A1 specific substrate, estradiol. (C) Estradiol was added to the cell-culturing media and the amount of estradiol 3-O-glucuronide was determined. (D) Trovafloxacin acyl-glucuronidation was determined in microsomes prepared from HepG2 cells. Data are the mean±S.D. (n=5). N.D., not detected. * p<0.05 compared to control. induced HepG2 cells (Fig. 1C). Trovafloxacin acyl-glucuronide trol cells. Increased amount of UGT1A1 in the HepG2 cells was not detected in the reaction mixture including the micro- did not affect the expression of MT2A, TOP1, and BCLAF somes prepared from the control HepG2 cells. Meanwhile, a mRNA, suggesting that trovafloxacin acyl-glucuronide was slight but detectable amount of trovafloxacin acyl-glucuronide not involved in the induction of these genes. The expression was observed in the reaction mixture including the micro- levels of MFN1, MT1H, and MT1X were also examined in the somes prepared from the UGT1A1-induced HepG2 cells with control and UGT1A1-induced HepG2 cells; however, none of 1.1 pmol/min/mg (Fig. 1D). Estradiol is a selective substrate these genes were specifically induced by trovafloxacin in the of UGT1A129); therefore, it was demonstrated that the HepG2 UGT1A1-induced HepG2 cells (Fig. 2B). cells treated with 50-µM chrysin was the UGT1A1-induced Expression of Toxicity-Associated Genes in HepG2 Cells cell model. It has been demonstrated that neuronal apoptosis inhibitory Expression of Trovafloxacin-Induced Genes in HepG2 protein, MHC class II, HET-E, telomerase protein 1 domain-, Cells It has been shown that TOP1, BCLAF MFN1, MT2A, leucine-rich repeat-, and pyrin domain-containing protein MT1H, and MT1X were significantly induced in the liver 3 (NALP3), receptor for advanced glycation endproducts when hepatocytes and animals were treated with trovafloxa- (RAGE), interleukin (IL)-6, and IL-23p19 were highly induced cin.13) To investigate whether these trovafloxacin-associated by drugs that were associated with drug-induced liver injury genes are induced by trovafloxacin acyl-glucuronide or not, (DILI) in the cell-based assay.30) In HepG2 cells, IL-6 was the mRNA expression of the genes was quantified in the induced 2-fold when HepG2 cells were treated with trova- HepG2 cells and the cells treated with trovafloxacin in the floxacin (Fig. 2C). Meanwhile, trovafloxacin did not induce presence or absence of chrysin. It was confirmed that chrysin NALP3, RAGE, or IL-23p19 in HepG2 cells. In the UGT1A1- highly induced UGT1A1 in the HepG2 cells, while trovafloxa- induced model, such induction pattern was still the same as cin itself did not induce UGT1A1 (Fig. 2A). MT2A, TOP1, observed in the control HepG2 cells. and BCLAF were induced 2- to 3-fold by trovafloxacin (Fig. Chemokine (C-X-C motif) ligand 2 (CXCL-2), S100A9, 2B), which was in agreement with the previous reports.13) In and IL-1β are also genes associated with DILI.30,31) To further the UGT1A1-induced cells, however, the expression level of determine whether these genes are specifically responsive MT2A, TOP1, and BCLAF was still 2- to 3-fold higher in the to trovafloxacin acyl-glucuronide, the expression levels of presence of trovafloxacin compared to the level in the con- CXCL-2, S100A9, and IL-1β were examined in the control and Vol. 39, No. 10 (2016) Biol. Pharm. Bull. 1607

Fig. 2. Expression Levels of UGT1A1 and Toxicity-Associated Genes in the Control and UGT1A1-Induced HepG2 Cells

UGT1A1-induced HepG2 cells were developed by treating the cells with chrysin (50 µM) for 48 h. Control and the UGT1A1-induced HepG2 cells were further treated with trovafloxacin (50 µM) for 24 h. Total RNA was isolated from the cells and the mRNA expression of UGT1A1 was determined (A). The mRNA expressions of MFN1, MT2A, MT1H, MT1X, TOP1, and BCLAF, which are reported trovafloxacin-induced toxicity-associated gene, were determined by Q-PCR (B). The mRNA expressions of NALP3, RAGE, IL-6, and IL-23p19, which are reported DILI-associated gene, were determined by Q-PCR (C). The mRNA expressions of NALP3, RAGE, IL-6, and IL-23p19, which are reported DILI-associated gene, were determined by Q-PCR (C). The mRNA expressions of CXCL-2, S100A9, and IL-1β, which are immunotoxicity- associated genes, were determined by Q-PCR (D). Data are the mean of two independent experiments (B and C) or the mean±S.D. of three independent experiments (A and D). TFX, Trovafloxacin. * p<0.05 compared to the level in HepG2 cells treated with TFX. † p<0.05 compared to the level in HepG2 cells treated with chrysin.

UGT1A1-induced HepG2 cells. CXCL-2 was induced 4-fold more than 10-fold compared to the level in the control cells. by the treatment of trovafloxacin in the HepG2 cells (Fig. 2D). S100A9 was not induced by trovafloxacin; however, this gene In the UGT1A1-induced cells, trovafloxacin induced CXCL-2 was highly induced by chrysin (Fig. 2D). IL-1β was similarly 1608 Biol. Pharm. Bull. Vol. 39, No. 10 (2016)

Fig. 3. Effect of UGT1A1 on Cytotoxicity in Vitro and in Vivo (A) HepG2 cells were co-treated with trovafloxacin (TFX) and TNF-α in the presence or absence of chrysin. (B) Wild type (WT), Ugt1+/− (Hetero), and Ugt1−/− (KO) mice were co-treated with TFX and LPS. Nine hours after the treatment, serum ALT levels were determined. Data are the mean±S.D. (n>3). * p<0.05 compared to con- trol. induced by chrysin in HepG2 cells in the presence or absence extremely hydrophilic metabolites such as glucuronides are of trovafloxacin (Fig. 2D). CXCL-2 was uniquely induced by usually produced inside the cells. Meanwhile, such metabo- trovafloxacin in the UGT1A1-induced HepG2 cells, indicating lites would be generated outside of the plasma membrane if that CXCL-2 might be specifically induced by trovafloxacin liver microsomes and enzyme-expressing systems were ex- acyl-glucuronide. perimentally added into the medium. To properly evaluate the Importance of UGT1A1 in the Trovafloxacin-Induced effect of trovafloxacin acyl-glucuronide on the hepatic cells, Liver Toxicity in Vitro and in Vivo The effect of the in the present study, we developed UGT1A1-induced HepG2 UGT1A1 induction on cell viability was examined by an MTT cells. assay. Co-treatment of control HepG2 cells with trovafloxacin Due to the detection limit of the instruments used, we were and TNF-α decreased the cell viability by 50% (Fig. 3A, left not able to determine the formation of trovafloxacin acyl- column). When UGT1A1-induced HepG2 cells were used, glucuronide in the cells. However, the microsomes prepared the co-treatment decreased the cell viability by 65% (Fig. from the UGT1A1-induced HepG2 cells exhibited higher es- 3A, right column). To further investigate the importance of tradiol 3-O- and trovafloxacin acyl-glucuronidation activities UGT1A1 in the trovafloxacin-induced liver toxicity in vivo, (Figs. 1B, D), indicating that a higher amount of trovafloxacin we utilized Ugt1-knockout mice. It was previously shown that acyl-glucuronide should have been produced in the UGT1A1- Ugt1 knockout mice display no UGT1 activities, while wild induced cells. A previous study reported that trovafloxacin type and heterozygous Ugt1 mice similarly have higher gluc- was mainly metabolized to its glucuronide in humans, while uronidation activities.23) When wild type and heterozygous it is partially metabolized to N-acetyltrovafloxacin and sul- Ugt1 mice were treated with trovafloxacin as well as LPS, fate conjugate.8) Therefore, it was considered that the effect serum ALT levels were increased 5-fold. In contrast, such in- of other trovafloxacin-metabolizing enzymes and their me- crease was not observed in the Ugt1 knockout mice (Fig. 3B). tabolites on the trovafloxacin-induced cytotoxicity was minor. Although it was not statistically significant, these data indicate Thus, it is assumed that the 10-fold induction of CXCL-2 (Fig. that UGT1A1 might be involved in the trovafloxacin-induced 2D) was specifically caused by trovafloxacin acyl-glucuronide. hepatotoxicity in vitro and in vivo. In vitro MTT assays and in vivo studies with Ugt1 knockout mice also indicated that UGT1A1 was highly involved in the DISCUSSION trovafloxacin-induced hepatotoxicity (Fig. 3). C-X-C Motif chemokine receptor 2 (CXCR2) is a receptor of CXCL-2. DILI is one of the leading causes of acute liver failure in the It was previously demonstrated that ischemia–reperfusion U.S.A. While trovafloxacin was withdrawn from the market caused significant liver injury in wild type mice, but not in due to the high risk of developing the severe liver damage, the CXCR2-deficient mice.32) Moreover, it was demonstrated that molecular mechanism underlying the trovafloxacin-induced a treatment of primary hepatocytes with recombinant CXCL-2 hepatotoxicity remains to be cleared. Importantly, whether induced lactate dehydrogenase (LDH) release in the cells.32) the reactive metabolite—trovafloxacin acyl-glucuronide—is Induction of CXCL-2 was also observed in mice treated with involved in the toxic reaction or not has been inconclusive. hepatotoxic α-naphthylisothiocyanate, carbon tetrachloride, Various cell-based assays have been developed to evaluate the and acetaminophen.33–35) Previously, it was indicated that cer- cyto- and genotoxicity of compounds.12) An addition of liver tain types of cells and cytokines, such as Th17 cells and TNF- microsomes or drug-metabolizing enzyme-expressing sys- α, are commonly involved in various drug-induced hepatotox- tems, as well as the substrates, into the cell-culturing medium icity.36) Therefore, CXCL-2 can also be specifically induced by is a convenient method to generate reactive metabolites in the certain toxicants and its induction might play a significant role cell-culturing medium. In fact, this method can be used to in the development of liver injury. The detailed mechanism determine the hepatotoxic potential of compounds in preclini- underlying the induction of CXCL-2 by trovafloxacin acyl- cal drug development.30) Since most of drug-metabolizing en- glucuronide needs to be elucidated in the future. Since Toll zymes are localized in the endoplasmic reticulum membrane, like receptor 2 (TLR2) is tightly associated with the develop- Vol. 39, No. 10 (2016) Biol. Pharm. Bull. 1609 ment of liver injury as well as the gene expression of hepatic primary human hepatocytes to HepG2 cells. Drug Metab. Dispos., CXCL-2,34,37) TLR2 might be the key factor in trovafloxacin 36, 223–233 (2008). acyl-glucuronide-associated liver injury. 14) Liguori MJ, Anderson MG, Bukofzer S, McKim J, Pregenzer JF, Retief J, Spear BB, Waring JF. Microarray analysis in human hepa- In conclusion, we treated HepG2 cells with chrysin, a 38,39) tocytes suggests a mechanism for hepatotoxicity induced by trova- known UGT inducer, to induce UGT1A1. CXCL-2, a floxacin. Hepatology, 41, 177–186 (2005). cytokine involved in DILI, was uniquely induced by trova- 15) Waring JF, Liguori MJ, Luyendyk JP, Maddox JF, Ganey PE, floxacin in the UGT1A1-induced HepG2 cells. In vitro and in Stachlewitz RF, North C, Blomme EA, Roth RA. Microarray analy- vivo studies further demonstrated the importance of UGT1A1 sis of lipopolysaccharide potentiation of trovafloxacin-induced liver in the trovafloxacin-induced liver toxicity. Although the com- injury in rats suggests a role for proinflammatory chemokines and plete mechanism of trovafloxacin-induced liver injury is still neutrophils. J. Pharmacol. Exp. Ther., 316, 1080–1087 (2006). unknown, trovafloxacin acyl-glucuronide can be involved in 16) Lubasch A, Keller I, Borner K, Koeppe P, Lode H. Comparative the development of toxic reactions in vitro and in vivo. In- pharmacokinetics of ciprofloxacin, gatifloxacin, grepafloxacin, le- terindividual variability in the UGT1A1 activity as well as vofloxacin, trovafloxacin, and moxifloxacin after single oral admin- protein–protein interactions involving UGT40,41) might be as- istration in healthy volunteers. Antimicrob. Agents Chemother., 44, 2600–2603 (2000). sociated with the idiosyncrasy of trovafloxacin-induced liver 17) Vincent J, Venitz J, Teng R, Baris BA, Willavize SA, Polzer RJ, injury in humans. Friedman HL. Pharmacokinetics and safety of trovafloxacin in healthy male volunteers following administration of single intra- Conflict of Interest The authors declare no conflict of venous doses of the prodrug, alatrofloxacin. J. Antimicrob. Che- interest. mother., 39 (Suppl. 2), 75–80 (1997). 18) Olsen KM, Rebuck JA, Weidenbach T, Fish DN. Pharmacokinetics REFERENCES of intravenous trovafloxacin in critically ill adults. Pharmacothera- py, 20, 400–404 (2000). 1) Dutton GJ. Acceptor substrates of UDP glucuronosyltransferase and 19) Fujiwara R, Sumida K, Kutsuno Y, Sakamoto M, Itoh T. UDP–gluc- their assay, in Glucuronidation of Drugs and Other Compounds uronosyltransferase (UGT) 1A1 mainly contributes to the glucuroni- (Dutton GJ ed). pp. 69–78, CRC Press, Boca Raton, FL (1980). dation of trovafloxacin. Drug Metab. Pharmacokinet., 30, 82–88 2) Mackenzie PI, Bock KW, Burchell B, Guillemette C, Ikushiro SI, (2015). Iyanagi T, Miners JO, Owens IS, Nebert DW. Nomenclature update 20) Kutsuno Y, Hirashima R, Sakamoto M, Ushikubo H, Michimae for the mammalian UDP glycosyltransferase (UGT) gene superfam- H, Itoh T, Tukey RH, Fujiwara R. Expression of UDP–glucurono- ily. Pharmacogenet. Genomics, 15, 677–685 (2005). syltransferase 1 (UGT1) and glucuronidation activity toward endo- 3) Ritter JK, Chen F, Sheen YY, Tran HM, Kimura S, Yeatman MT, genous substances in humanized UGT1 mouse brain. Drug Metab. Owens IS. A novel complex locus UGT1 encodes human bilirubin, Dispos., 43, 1071–1076 (2015). phenol, and other UDP–glucuronosyltransferase isozymes with 21) Fujiwara R, Takenaka S, Hashimoto M, Narawa T, Itoh T. Expres- identical carboxyl termini. J. Biol. Chem., 267, 3257–3261 (1992). sion of human solute carrier family transporters in skin: possible 4) Tukey RH, Strassburg CP. Human UDP–glucuronosyltransferases: contributor to drug-induced skin disorders. Sci. Rep., 4, 5251 (2014). metabolism, expression, and disease. Annu. Rev. Pharmacol. Toxi- 22) Sumida K, Kawana M, Kouno E, Itoh T, Takano S, Narawa T, col., 40, 581–616 (2000). Tukey RH, Fujiwara R. Importance of UDP–glucuronosyltransfer- 5) Fujiwara R, Maruo Y, Chen S, Tukey RH. Role of extrahepatic ase 1A1 expression in skin and its induction by UVB in neonatal UDP–glucuronosyltransferase 1A1: Advances in understanding hyperbilirubinemia. Mol. Pharmacol., 84, 679–686 (2013). breast milk-induced neonatal hyperbilirubinemia. Toxicol. Appl. 23) Nguyen N, Bonzo JA, Chen S, Chouinard S, Kelner MJ, Hardiman Pharmacol., 289, 124–132 (2015). G, Bélanger A, Tukey RH. Disruption of the ugt1 locus in mice 6) Kaminsky LS, Zhang QY. The small intestine as a xenobiotic- resembles human Crigler–Najjar type I disease. J. Biol. Chem., 283, metabolizing organ. Drug Metab. Dispos., 31, 1520–1525 (2003). 7901–7911 (2008). 7) Gootz TD, McGuirk PR. New quinolones in development. Expert 24) Fujiwara R, Nguyen N, Chen S, Tukey RH. Developmental hyper- Opin. Investig. Drugs, 3, 93–114 (1994). bilirubinemia and CNS toxicity in mice humanized with the UDP 8) Dalvie DK, Khosla N, Vincent J. Excretion and metabolism of tro- glucuronosyltransferase 1 (UGT1) locus. Proc. Natl. Acad. Sci. vafloxacin in humans. Drug Metab. Dispos., 25, 423–427 (1997). U.S.A., 107, 5024–5029 (2010). 9) Pumford NR, Myers TG, Davila JC, Highet RJ, Pohl LR. Immu- 25) Aoshima N, Fujie Y, Itoh T, Tukey RH, Fujiwara R. Glucose in- nochemical detection of liver protein adducts of the nonsteroidal duces intestinal human UDP–glucuronosyltransferase (UGT) 1A1 to antiinflammatory drug diclofenac. Chem. Res. Toxicol., 6, 147–150 prevent neonatal hyperbilirubinemia. Sci. Rep., 4, 6343 (2014). (1993). 26) Soars MG, Petullo DM, Eckstein JA, Kasper SC, Wrighton SA. An 10) Kutsuno Y, Itoh T, Tukey RH, Fujiwara R. Glucuronidation of drugs assessment of UDP–glucuronosyltransferase induction using prima- and druginduced toxicity in humanized UDP–glucuronosyltransfer- ry human hepatocytes. Drug Metab. Dispos., 32, 140–148 (2004). ase 1 mice. Drug Metab. Dispos., 42, 1146–1152 (2014). 27) Ritter JK, Kessler FK, Thompson MT, Grove AD, Auyeung DJ, 11) Miyashita T, Kimura K, Fukami T, Nakajima M, Yokoi T. Evalu- Fisher RA. Expression and inducibility of the human bilirubin ation and mechanistic analysis of the cytotoxicity of the acyl UDP–glucuronosyltransferase UGT1A1 in liver and cultured pri- glucuronide of nonsteroidal anti-inflammatory drugs. Drug Metab. mary hepatocytes: Evidence for both genetic and environmental Dispos., 42, 1–8 (2014). influences. Hepatology, 30, 476–484 (1999). 12) Koga T, Fujiwara R, Nakajima M, Yokoi T. Toxicological evalua- 28) Westerink WM, Schoonen WG. Phase II enzyme levels in HepG2 tion of acyl glucuronides of nonsteroidal anti-inflammatory drugs cells and cryopreserved primary human hepatocytes and their in- using human embryonic kidney 293 cells stably expressing human duction in HepG2 cells. Toxicol. In Vitro, 21, 1592–1602 (2007). UDP–glucuronosyltransferase and human hepatocytes. Drug Metab. 29) Miners JO, Mackenzie PI, Knights KM. The prediction of drug- Dispos., 39, 54–60 (2011). glucuronidation parameters in humans: UDP–glucuronosyltrans- 13) Liguori MJ, Blomme EA, Waring JF. Trovafloxacin-induced gene ferase enzyme-selective substrate and inhibitor probes for reaction expression changes in liver-derived in vitro systems: comparison of phenotyping and in vitro–in vivo extrapolation of drug clearance 1610 Biol. Pharm. Bull. Vol. 39, No. 10 (2016)

and drug–drug interaction potential. Drug Metab. Rev., 42, 196–208 reactive nitrogen and cytokine formation. Toxicol. Sci., 75, 458–467 (2010). (2003). 30) Yano A, Oda S, Fukami T, Nakajima M, Yokoi T. Development of 36) Yokoi T. New prospectives and understanding in drug-induced liver a cell-based assay system considering drug metabolism and im- injury considering drug metabolism and immune- and inflamma- mune- and inflammatory-related factors for the risk assessment of tion-related factors. Folia Pharmacologica Japonica, 144, 22–27 drug-induced liver injury. Toxicol. Lett., 228, 13–24 (2014). (2014). 31) Toyoda Y, Miyashita T, Endo S, Tsuneyama K, Fukami T, Nakajima 37) Shaw PJ, Ganey PE, Roth RA. Trovafloxacin enhances the inflam- M, Yokoi T. Estradiol and progesterone modulate halothane-induced matory response to a Gram-negative or a Gram-positive bacterial liver injury in mice. Toxicol. Lett., 204, 17–24 (2011). stimulus, resulting in neutrophil-dependent liver injury in mice. J. 32) Kuboki S, Shin T, Huber N, Eismann T, Galloway E, Schuster Pharmacol. Exp. Ther., 330, 72–78 (2009). R, Blanchard J, Edwards MJ, Lentsch AB. Hepatocyte signaling 38) Bonzo JA, Bélanger A, Tukey RH. The role of chrysin and the ah through CXC chemokine receptor-2 is detrimental to liver recovery receptor in induction of the human UGT1A1 gene in vitro and in after ischemia/reperfusion in mice. Hepatology, 48, 1213–1223 transgenic UGT1 mice. Hepatology, 45, 349–360 (2007). (2008). 39) Smith CM, Graham RA, Krol WL, Silver IS, Negishi M, Wang H, 33) Xu J, Lee G, Wang H, Vierling JM, Maher JJ. Limited role for CXC Lecluyse EL. Differential UGT1A1 induction by chrysin in primary chemokines in the pathogenesis of alpha-naphthylisothiocyanate- human hepatocytes and HepG2 cells. J. Pharmacol. Exp. Ther., 315, induced liver injury. Am. J. Physiol. Gastrointest. Liver Physiol., 1256–1264 (2005). 287, G734–G741 (2004). 40) Fujiwara R, Nakajima M, Oda S, Yamanaka H, Ikushiro S, Sakaki 34) Moles A, Murphy L, Wilson CL, Chakraborty JB, Fox C, Park EJ, T, Yokoi T. Interactions between human UDP–glucuronosyltransfer- Mann J, Oakley F, Howarth R, Brain J, Masson S, Karin M, Seki E, ase (UGT) 2B7 and UGT1A enzymes. J. Pharm. Sci., 99, 442–454 Mann DAA. A TLR2/S100A9/CXCL-2 signaling network is neces- (2010). sary for neutrophil recruitment in acute and chronic liver injury in 41) Fujiwara R, Itoh T. Extensive protein–protein interactions involving the mouse. J. Hepatol., 60, 782–791 (2014). UDP–glucuronosyltransferase (UGT) 2B7 in human liver micro- 35) James LP, McCullough SS, Lamps LW, Hinson JA. Effect of N- somes. Drug Metab. Pharmacokinet., 29, 259–265 (2014). acetylcysteine on acetaminophen toxicity in mice: relationship to