Trovafloxacin acyl-glucuronide induces chemokine (C-X-C motif) ligand 2 in HepG2 cells

Summary

UDP-glucuronosyltransferases (UGTs) catalyze glucuronidation of endogenous and exogenous compounds by transferring the glucuronic acid moiety of UDP-glucuronic acid to the substrates. While glucuronides are usually pharmacologically inactive, certain types of glucuronides, especially acyl-glucuronide, can exhibit the increased reactivity compared to the parent compounds. Trovafloxacin is an antibiotic that was withdrawn from the market relatively soon after its release due to the risk of hepatotoxicity. Trovafloxacin was mainly metabolized to its acyl-glucuronide; therefore, it was hypothesized that the acyl-glucuronide was involved in the development of hepatotoxicity. To investigate whether trovafloxacin acyl-glucuronide can be involved in the trovafloxacin-induced liver injury, in the present study, a cell-based UGT1A1-induced 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 hours. We demonstrated for the first time that 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. While there are reports showing the less toxicity of acyl-glucuronides, certain kinds of acyl-glucuronide, including trovafloxacin acyl-glucuronide, can be associated with the development of toxic reactions in vitro and in vivo.

Introduction

UDP-glucuronosyltransferases (UGTs; EC 2.4.1.17) are a family of membrane-bound enzymes that catalyze glucuronidation of endogenous and exogenous compounds by transferring the glucuronic acid moiety of UDP-glucuronic acid to the substrates.¹⁾ Human UGTs are mainly divided into two distinct families, UGT1 and UGT2, on the basis of evolutionary divergence and homology.²⁾ The UGT1 gene is located on chromosome 2q37 and produces nine functional enzymes, UGT1A1, UGT1A3, UGT1A4, UGT1A5, UGT1A6, UGT1A7, UGT1A8, UGT1A9, and UGT1A10, by exon sharing. ³⁾ The unique first exons encode N-terminal domain and the common exons 2 to 5 encode C-terminal domain of UGT1A proteins. Since all of UGTs recognize and utilize UDP-glucuronic acid as a co-substrate, it has been suggested that the C-terminal domain is responsible for the co-substrate binding. In contrast, UGT1A proteins exhibit overlapping but distinct substrate specificities, suggesting that the N-terminal domain is responsible for the substrate binding. While the liver is the most contributing tissue to the metabolism, ⁴⁾ recent findings suggest that extrahepatic tissues such as small intestine play an important role in glucuronidation of endogenous and exogenous compounds. ^5,6)

Trovafloxacin is an antibiotic that was released on the market in 1998. ⁷⁾ This promising agent was withdrawn from the market relatively soon after its release due to the risk of hepatotoxicity including acute liver failure. Trovafloxacin is mainly metabolized by UGTs to its acyl-glucuronide in humans. ⁸⁾ While glucuronides are usually pharmacologically inactive, certain types of glucuronides, especially acyl-glucuronide, can exhibit the increased reactivity compared to the parent compounds. ⁹⁾ It was therefore hypothesized that the trovafloxacin acyl-glucuronide was involved in the development of toxicity in the liver. Acyl-glucuronide-associated toxicity has been reported in vivo and in vitro; ^10,11) however, there are also reports showing that acyl-glucuronidation did not induce the cyto- and genotoxicity. ¹²⁾ Microarray expression analysis is a promising tool to identify genes associated with a drug treatment. To identify genes that were associated with the trovafloxacin-induced liver toxicity, several research groups carried out the microarray expression analysis in human hepatocytes, mice, and rats. ^13–15) The group of genes that were specifically induced by the trovafloxacin treatment included topoisomerase I (TOP1), B-cell leukemia/lymphoma 2 (BCL-2)-associated transcription factor 1 (BCLAF), Mitofusin1 (MFN1), Metallothionein (MT) 2A, MT1H, and MT1X. ¹³⁾ Although these genes might have been induced by the acyl-glucuronide in the hepatocytes, there still was a possibility that the parent compound itself was involved in the induction of the genes. This was because a certain amount of trovafloxacin still remained in the body even 24 hours after the oral and intravenous administration of trovafloxacin. ^16–18)

Previously, we determined UGT1A1 as the main UGT isoform responsible for trovafloxacin acyl-glucuronidation. ¹⁹⁾ To investigate whether the trovafloxacin acyl-glucuronide can be involved in the trovafloxacin-induced liver injury, in the present study, a cell-based UGT1A1-induced model was developed and the toxicity of trovafloxacin acyl-glucuronide was evaluated.

Materials and Methods

Chemicals and Reagents

UDP-glucuronic acid, alamethicin, chrysin, estradiol, and estradiol 3-O-glucuronide were purchased from Sigma–Aldrich (St Louis, MO). Trovafloxacin was purchased from Wako Pure Chemical (Osaka, Japan). Recombinant human tumor necrosis factor α (TNF-α) was purchased from Roche (Mannheim, Germany). Primers were commercially synthesized at Life Technologies (Carlsbad, CA). All other chemicals and solvents were of analytical grade or the highest grade commercially available.

Cell culture and chemical treatments

The human hepatoma HepG2 cells were obtained from DS Pharma Biomedical Co Ltd (Osaka, Japan). HepG2 cells were grown in Dulbecco’s modified Eagle’s medium containing 100 U/mL penicillin, 100 μg/mL streptomycin, and 10% fetal bovine serum and were maintained at 37°C in a humidified atmosphere containing 5% of CO². Before the treatment, HepG2 cells were seeded into six-well plates at 5 × 10⁵ cells/well. After 24 hours, the culture medium was changed to a normal or a chrysin-containing DMEM medium and subsequently cells were maintained for 48 hours until harvesting. RNA was isolated from the cells and was used for the quantitative-reverse-transcription PCR (Q-PCR) analysis. The microsomal fraction was also obtained from the cells. Control and the UGT1A1-induced HepG2 cells were further treated with trovafloxacin (50 μM) for 24 hours.

Q-PCR analysis

Complementary DNA (cDNA) was synthesized from the total RNA using ReverTra Ace qPCR RT Master Mix (Toyobo, Tokyo, Japan) according to the manufacturer’s protocol. Quantitative RT-PCR was performed with THUNDERBIRD SYBR qPCR Mix (Toyobo), and the reactions were run in a CFX96 Real-Time PCR Detection System (Bio-Rad, Hercules, CA). Expression of glyceraldehyde 3-phosphate dehydrogenase (GAPDH) mRNA was used as an internal control for the cDNA quantity and quality. Primer pairs that can detect UGT1A1 and GAPDH were reported previously. ^20,21) Other primers used were newly established with a program, Primer Blast (National Institutes of Health). We confirmed that all of the primer sets produced specific bands and that bands were not detected when the PCR reaction was conducted without cDNA. After an initial denaturation at 95°C for 30 seconds, the amplification was performed by denaturation at 95°C for 5 seconds, annealing at an appropriate temperature for 30 seconds, and extension at 72°C for 30 seconds for 45 cycles.

Enzyme assay in vitro and in cells

Microsomes were prepared as described before. ²²⁾ Estradiol 3-O- and trovafloxacin acyl-glucuronidations were determined according to previously reported methods. ^19,22) and Fujiwara ewith slight modifications. Briefly, a typical incubation mixture (200 μl of total volume) contained 50 mM Tris-HCl (pH 7.4), 4 mM MgCl₂, 2 mM UDPGA, 50 μg/ml alamethicin, 1.5 mg/ml microsomes, and 10 μM estradiol or 50 μM trovafloxacin. The reaction was initiated by the addition of UDPGA after a 3-min preincubation at 37°C. After incubation at 37°C for 90 min for estradiol glucuronidation and 60 min for trovafloxacin glucuronidation, the reaction was terminated by addition of 200 μl of cold methanol. After removal of the protein by centrifugation at 12,000 g for 5 min, supernatant was subjected to high-performance liquid chromatography (HPLC) to quantitate the formation of estradiol 3-O-glucuronide and trovafloxacin acyl-glucuronide. The conditions of the HPLC analysis were reported before. ^19,22)

HepG2 cells were treated with 100 μM estradiol. Seven hours after incubation at 37°C, a portion of the cell-culturing media was collected and mixed with the same amount of acetonitrile. After removal of proteins by centrifugation at 12,000 g for 5 min, supernatant was subjected to HPLC.

Cell viability assay

HepG2 cells were seeded into 96-well plates at 1 × 10⁴ cells/well and were maintained at 37°C for 24 hours. Twenty-four hours after changing the culture medium to a normal or a chrysin-containing DMEM medium, the cells were co-treated with trovafloxacin (50 μM) and TNF-α (4ng/mL) for 24 hours. Cell viability was measured using a MTT Cell Counting Kit (Nacalai Tesque) according to the manufacture’s protocol.

Animals

Heterozygous Ugt1 (Ugt1^+/−) mice²³⁾ were crossed to C57BL/6NCrSlc mice to obtain wild type (Ugt1^+/+), Ugt1^+/−, and UGT1 knockout (Ugt1^−/−) mice. Genomic DNA was isolated from tail biopsies and was used as a template for genotyping PCR.²⁴⁾ All animals received food and water ad libitum, and mouse handling and experimental procedures were conducted in accordance with our animal care protocol, which was previously approved by Kitasato University.

Two-day-old mice were subcutaneously treated with trovafloxacin (150 mg/kg) or canola oil. Three hours after the injections, mice were subcutaneously treated with lipopolysaccharide (LPS) (5 mg/kg) or serum. Nine hours after the second treatment, blood was obtained from the submandibular vein and serum was prepared. Serum alanine aminotransferase (ALT) levels were determined using a Transaminase CII-test Wako kit (Wako Pure Chemical) according to the manufacture’s protocol.

Statistical analysis

All data are shown as mean ± SD. Significant differences of UGT1A1 expression in HepG2 cells were analyzed by using Dunnett’s test. Enzyme assay and cell viability were assessed for statistical significance using the unpaired t-test. For the gene expression study (Fig. 2A, 2D), Tukey-Kramer test was used to calculate statistical significance. The criterion of significance was p < 0.05.

Figure 2. Expression levels of UGT1A1 (A) and toxicity-associated genes (B) in the control and UGT1A1-induced HepG2 cells.

UGT1A1-induced HepG2 cells were developed by treating the cells with Chrysin (50 μM) for 48 hours. Control and the UGT1A1-induced HepG2 cells were further treated with trovafloxacin (50 μM) for 24 hours. Total RNA was isolated from the cells and the mRNA expressions of UGT1A1, MFN1, MT2A, MT1H, MT1X, TOP1, and BCLAF were determined by Q-PCR. Data are mean of duplicate determinations. TFX, Trovafloxacin.

Results

Development of the UGT1A1-induced cell model

Chrysin, phenytoin, carbamazepine, and phenobarbital have an ability to induce UGT1A1 in vitro and in vivo. ^25–27) To determine the most potent UGT1A1 inducer in HepG2 cells, we preliminarily conducted a cell-based induction assay. ²⁸⁾ HepG2 cells were treated with a lower (10 μM) and a higher (50 μM) concentration of chrysin, phenytoin, carbamazepine, and phenobarbital. Forty-eight hours after the treatment, RNA was isolated and a Q-PCR analysis was carried out to determine the UGT1A1 expression level in the cells. We observed that phenytoin, carbamazepine, and phenobarbital moderately induced UGT1A1 in HepG2 cells. In contrast, it was demonstrated that chrysin significantly induced UGT1A1. In the subsequently performed study, chrysin concentration-dependently induced the UGT1A1 mRNA expression in the HepG2 cells (Fig. 1A). The highest expression of UGT1A1 was observed when the cells were treated with 50-μM chrysin. When HepG2 cells were treated with even higher concentrations of chrysin such as 80 μM and 100 μM, slight and moderate cytotoxicity was induced, respectively. It was further demonstrated that the microsomes prepared from the HepG2 cells treated with 50-μM chrysin exhibited the 8-fold higher estradiol 3-O-glucuronidation activity (Fig. 1B). The estradiol 3-O-glucuronidations in the cells were also determined in the absence of additional UDPGA. Estradiol was added into the cell-culturing media and was incubated for 7 hours. Estradiol 3-O-glucuronide was detected in the cell-culturing medium of HepG2 cells. The amount of glucuronide was three-fold greater in the medium of the UGT1A1-induced HepG2 cells (Fig. 1C). Trovafloxacin acyl-glucuronide was not detected in the reaction mixture including the microsomes prepared from the control HepG2 cells. Meanwhile, a slight but detectable amount of trovafloxacin acyl-glucuronide was observed in the reaction mixture including the microsomes prepared from the UGT1A1-induced HepG2 cells with 1.1 pmol/min/mg (Fig. 1D). Estradiol is a selective substrate of UGT1A1; ²⁹⁾ therefore, it was demonstrated that the HepG2 cells treated with 50-μM chrysin was the UGT1A1-induced cell model.

Figure 1. Induction of UGT1A1 mRNA and activity in HepG2 cells.

A, HepG2 cells were treated with Chrysin (20–100 μM) for 48 hours. 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 hours. Microsomes were prepared and the UGT1A1 activity was determined using the UGT1A1 specific substrate, estradiol. Data are mean ± sd (n = 5).

Expression of trovafloxacin-induced genes in HepG2 cells

It has been shown that TOP1, BCLAF MFN1, MT2A, MT1H, MT1X were significantly induced in the liver when hepatocytes and animals were treated with trovafloxacin. ¹³⁾ To investigate whether these trovafloxacin-associated genes could be induced by trovafloxacin acyl-glucuronide or not, the mRNA expression of the genes was quantified in the HepG2 cells and the cells treated with trovafloxacin in the presence or absence of chrysin, the UGT1A1 inducer. It was shown that trovafloxacin itself did not induce UGT1A1 in the HepG2 cells (Fig. 2A). MT2A, TOP1, and BCLAF was induced 2- to 3-fold by the treatment of trovafloxacin (Fig. 2B), which was in agreement with the previous reports. ¹³⁾ In the UGT1A1-induced cells, however, the expression level of MT2A, TOP1, and BCLAF was still 2- to 3-fold higher in the presence of trovafloxacin compared to the level in the control cells. Increased amount of UGT1A1 in the HepG2 cells did not affect the expression of MT2A, TOP1, and BCLAF mRNA, suggesting that the trovafloxacin acyl-glucuronide was not involved in the induction of these genes. The expression levels of MFN1, MT1H, and MT1X were also examined in the control and UGT1A1-induced HepG2 cells; however, none of these genes were specifically induced by trovafloxacin in the UGT1A1-induced HepG2 cells (Fig. 2B).

Expression of toxicity-associated genes in HepG2 cells

It has been demonstrated that NATCH, LRR, and pyrin domain-containing protein 3 (NALP3), receptor for advanced glycation endproducts (RAGE), interleukin (IL)-6, and IL-23p19 were highly induced by drugs that were associated with DILI in the cell-based assay. ³⁰⁾ In HepG2 cells, IL-6 was induced 2-fold when the cells were treated with trovafloxacin (Fig. 2C). Meanwhile, trovafloxacin did not induce NALP3, RAGE, or IL-23p19 in the HepG2 cells. In the UGT1A1-induced model, such induction pattern was still the same as observed in the regular HepG2 cells.

CXCL-2, S100A9, and IL-1β are also genes associated with DILI. ^30,31) To further determine whether these genes could be specifically responsive to trovafloxacin acyl-glucuronide, the expression levels of CXCL-2, S100A9, and IL-1β were examined in the control and UGT1A1-induced HepG2 cells. It was confirmed that chrysin highly induced UGT1A1 in the HepG2 cells (Fig. 2A). CXCL-2 was induced 4-fold by the treatment of trovafloxacin in the HepG2 cells (Fig. 2D). In the UGT1A1-induced cells, trovafloxacin induced CXCL-2 more than 10-fold compared to the level in the control cells. S100A9 was not induced by trovafloxacin; however, this gene was highly induced by chrysin (Fig. 2D). While IL-1β was induced by trovafloxacin in HepG2 cells, the gene was similarly induced by trovafloxacin in the UGT1A1-induced model (Fig. 2D). CXCL-2 was uniquely induced by trovafloxacin in the UGT1A1-induced HepG2 cells, indicating that CXCL-2 was specifically responsive to trovafloxacin acyl-glucuronide.

Importance of UGT1A1 in the trovafloxacin-induced liver toxicity in vitro and in vivo

The effect of the UGT1A1 induction on cell viability was examined by MTT assay. Co-treatment of the control HepG2 cells with trovafloxacin and TNF-α decreased the cell viability by 50% (Fig. 3A-left column). When the UGT1A1-induced HepG2 cells were used, the co-treatment decreased the cell viability by 65% (Fig. 3A-right column). To further investigate the importance of UGT1A1 in the trovafloxacin-induced liver toxicity in vivo, we utilized Ugt1-knockout mice. It was previously shown that Ugt1 knockout mice display no UGT1 activities, while wild type and heterozygous Ugt1 mice similarly have higher glucuronidation activities.²³⁾ When wild type and heterozygous Ugt1 mice were treated with trovafloxacin as well as LPS, serum ALT levels were increased 5-fold. In contrast, such increase was not observed in the Ugt1 knockout mice (Fig. 3B). These data indicate that UGT1A1 is highly involved in the trovafloxacin-induced hepatotoxicity in vitro and in vivo.

Figure 3. Expression levels of toxicity-associated genes in the control and UGT1A1-induced HepG2 cells.

Control and the UGT1A1-induced HepG2 cells were treated with trovafloxacin (50 μM) for 24 hours. Total RNA was isolated from the cells and the mRNA expressions of NALP3 (A), RAGE (B), IL-6 (C), and IL-23p19 (D) were determined by Q-PCR. Data are mean of duplicate determinations. TFX, Trovafloxacin.

Discussion

DILI is one of the leading causes of acute liver failure in the U.S. While trovafloxacin was withdrawn from the market due to the high risk of developing the severe liver damage, the molecular mechanism underlying the trovafloxacin-induced hepatotoxicity remains to be cleared. Importantly, whether the reactive metabolite – trovafloxacin acyl-glucuronide – can be involved in the toxic reaction or not has been inconclusive. Various cell-based assays have been developed to evaluate the cyto- and genotoxicity of compounds. ¹²⁾ An addition of liver microsomes or drug-metabolizing enzyme-expressing systems, as well as the substrates, into the cell-culturing medium is a convenient method to generate reactive metabolites in the cell-culturing medium. In fact, this method can be used to determine the hepatotoxic potential of compounds in preclinical drug development. ³⁰⁾ Since most of drug-metabolizing enzymes are localized in the endoplasmic reticulum membrane, extremely hydrophilic metabolites such as glucuronides are usually produced inside the cells. Meanwhile, such metabolites would be generated outside of the plasma membrane if liver microsomes and enzyme-expressing systems were experimentally added into the medium. To properly evaluate the effect of trovafloxacin acyl-glucuronide on the hepatic cells, in the present study, we developed a UGT1A1-induced HepG2 cells.

Due to the detection limit of the instruments used, we were not able to quantify the amount of trovafloxacin acyl-glucuronide in the cells. However, the microsomes prepared from the UGT1A1-induced HepG2 cells exhibited higher estradiol 3-O- and trovafloxacin acyl-glucuronidation activities (Fig. 1B and D), indicating that a higher amount of trovafloxacin acyl-glucuronide should have been produced in the UGT1A1-induced cells than the amount in the control HepG2 cells. A previous study reported that trovafloxacin was mainly metabolized to its glucuronide in humans, while it is partially metabolized to N-acetyltrovafloxacin and sulfate conjugate. ⁸⁾ Therefore, it was assumed that the effect of other trovafloxacin-metabolizing enzymes and their metabolites on the trovafloxacin-induced cytotoxicity was slight. Thus, it is assumed that the 10-fold induction of CXCL-2 (Fig. 2D) was specifically caused by trovafloxacin acyl-glucuronide. In vitro MTT assays and in vivo studies with Ugt1 knockout mice also indicated that UGT1A1 was highly involved in the trovafloxacin-induced hepatotoxicity. CXCR2 is a receptor of CXCL2. It was previously demonstrated that ischemia-reperfusion caused significant liver injury in wild type mice, but not in CXCR2-deficient mice. ³²⁾ Moreover, it was demonstrated that a treatment of primary hepatocytes with recombinant CXCL2 induced LDH release in the cells. ³²⁾ Induction of CXCL2 was also observed in mice treated with hepatotoxic α-naphthylisothiocyanate, carbon tetrachloride, and acetaminophen. ^33–35) Previously, it was indicated that certain types of cells and cytokines, such as Th17 cells and TNF-α, are commonly involved in various drug-induced hepatotoxicity. ³⁶⁾ Therefore, CXCL-2 can also be specifically induced by certain toxicants and the induction might play a significant role in the development of liver injury. The detailed mechanism underlying the induction of CXCL-2 by trovafloxacin acyl-glucuronide needs to be elucidated in the future. Since Toll like receptor 2 (TLR2) is tightly associated with the development of liver injury as well as the gene expression of hepatic CXCL-2^{34, 37)}, TLR2 might be the key factor in trovafloxacin acyl-glucuronide-associated liver injury.

A large inter-individual variability in the expression and the function of UGT1A1 has been reported. UGT1A1*6 and *28, the two major genetic polymorphisms of UGT1A1, can significantly reduce the enzymatic activity of UGT1A1 in the body. SN-38, an active metabolite of an anti-cancer agent irinotecan, is selectively glucuronidated by UGT1A1. ³⁸⁾ It has been known that the carriers of UGT1A1*6 and *28 have a higher risk of developing adverse reactions when normal doses of irinotecan are administered. Meanwhile, such genetic polymorphism carriers might have a lower risk of developing the acyl-glucuronide-induced hepatotoxicity due to their low enzyme activity of UGT1A1. Not only the UGT1A1 protein itself, but also other enzymes localized in the endoplasmic reticulum membrane where UGTs are expressed can regulate the function of UGTs.

Hetero-oligomerization of different UGT isoforms can result in an alteration of the enzyme activities. ^39,40) Ishii et al. recently showed that cytochrome P450 (CYP) 3A4 functionally activated UGTs. ^41,42) Furthermore, we previously demonstrated that a large number of microsomal proteins interact each other to possibly affect their function. ^43,44) As only 150 people developed liver toxicity including 14 cases of acute liver failure among more than 2 million people who received trovafloxacin, there is a significant inter-individual variability in the risk of developing trovafloxacin-induced liver injury. A complete characterization of the activities of drug-metabolizing enzymes in vivo prior to the administration of drugs might prevent the development of the reactive metabolite-associated DILI. Inter-individual variability in the expression and function of transporters that can transport drugs and metabolites across the plasma membrane⁴⁵⁾ might also be involved in the wide inter-individual variability in the development of the reactive metabolite-associated DILI.

In conclusions, we treated HepG2 cells with chrysin, a known UGT inducer, ^46,47) to induce UGT1A1. We demonstrated for the first time that CXCL-2, a cytokine involved in DILI, was uniquely induced by trovafloxacin in the UGT1A1-induced HepG2 cells. While there are reports showing the less toxicity of acyl-glucuronides, certain kinds of acyl-glucuronide, including trovafloxacin acyl-glucuronide, can be associated with the development of toxic reactions in vitro and in vivo. To avoid the development of such reactive metabolite-associated DILI, a complete characterization of the activities of drug-metabolizing enzymes in vivo prior to the administration of drugs might be required.