Introduction

Leflunomide (LEF) is an immunosuppressive agent that mainly inhibits dihydroorotate dehydrogenase, the rate-limiting enzyme in the biosynthesis of pyrimidines. LEF was marketed as a disease-modifying antirheumatic drug in 19981. However, many reports were released to describe its side effects during its clinical application, which included fatal hepatitis and liver failure. Due to these clinical cases of hepatotoxicity, the Food and Drug Administration (FDA) labeled LEF with a black box warning in 2011 (http://www.fda.gov/Safety/MedWatch/SafetyInformation/ucm228392.htm). Several cytochrome P450 enzymes (CYPs), such as CYP1A2, CYP2C19 and CYP3A4, are involved in the biotransformation of LEF into its major metabolite teriflunomide (TER)2. One clinical report showed that a patient co-administered with LEF and the CYP3A inhibitor itraconazole developed fatal hepatitis3. However, the contribution of CYPs to LEF-induced liver toxicity has not yet been elucidated completely, and the limited available in vitro results are controversial. For example, it was suggested that the metabolites of LEF might be more toxic to the liver, as LEF-induced cytotoxicity was attenuated by the nonspecific CYPs inhibitor ABT in immortalized human hepatocytes4. In contrast, another study indicated that LEF cytotoxicity was enhanced by several CYPs inhibitors in primary rat hepatocytes5. As far as we know, there is no direct in vivo studies investigating the role of CYPs in LEF induced hepatotoxicity. Besides, although TER treatment significantly increased aminotransferase level, which leads to discontinue therapy in clinical studies6,7,8, there is no investigation about the liver toxicity of the metabolite of LEF, TER. Therefore, it's worthwhile to investigate the mechanism of TER induced liver toxicity and its contribution to the hepatotoxicity of LEF. At the same time, both LEF and TER were reported to be high affinity substrates of efflux transporter breast cancer resistance protein (BCRP)9. Since there are multiple anti-rheumatic drugs are reported to be the substrates of BCRP10,11,12, and the change of BCRP function may lead to toxicity13, the potential drug-drug interaction risk mediated by BCRP or other transporters should be investigated. Therefore, we want to explore whether transporters are involved in the liver toxicity of LEF and TER.

In this study, we investigated the role of hepatic metabolism and transport in LEF-induced hepatotoxicity. We checked whether LEF toxicity was modulated by CYPs using nonspecific CYPs inhibitors in primary hepatocytes. Then, hepatic cytochrome P450 reductase null (HRN) mice were employed to verify the contribution of CYPs to the plasma concentration of LEF and its liver toxicity. Given that neither LEF nor TER significantly changed the mRNA expression of BCRP in either rat or human hepatocytes; and TER could significantly reduce sodium/bile acid cotransporter (NTCP) expression in human hepatocytes. Rat hepatocytes and NTCP-transfected HEK293 cells were used to investigate the contribution of selected transporters to the hepatotoxicity of LEF and TER. Then, SD rats were orally administered with TER for 4 weeks to further certify in vitro findings about potential mechanism of TER toxicity.

Materials and methods

Chemicals

LEF (99.5%; batch No 130603) and TER (>99.9%; batch No 121123) were kindly provided by Cinkate Pharmaceutical Intermediates Co, Ltd (Shanghai, China). For in vitro assays, LEF and TER were dissolved in Dimethyl sulfoxide (DMSO). All reagents used for cell culture were purchased from GIBCO unless otherwise specified. Dimethyl sulfoxide (DMSO), proadifen (SKF), aminobenzotriazole (ABT), collagenase (type IV), phenacetin (Phe), tetramethylrhodamine ethyl ester (TMRE), troglitazone (Tro) and 3-methylcholanthrene (3-MC) were purchased from Sigma-Aldrich (St Louis, MO, USA). BD MatrigelTM Basement Membrane Matrix and rat tail collagen (type I) were obtained from BD Biosciences (Palo Alto, CA, USA). 3-(4, 5-Dimethylthiazollthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT) was purchased from Sangon Biotech (Shanghai) Co, Ltd. BSA protein assay kit was obtained from Pierce Chemical (Rockford, IL, USA).

Animals

Male Sprague-Dawley (SD) rats (8 weeks old) and male C57 BL6 mice (6 weeks old) housed in the SPF class experimental animal room were purchased from Shanghai SLAC Laboratory Animal Co, Ltd (Shanghai, China). HRN mice, without metabolic activity, were generated as previously reported14. Male SD rats, male wild-type (WT) and HRN mice (7 weeks old) were housed under standard laboratory conditions (temperature 25±1 °C, humidity 50%±10% and 12 h light/12 h dark cycle) in the institutional animal facility with free access to food and water.

All animal experiments were conducted in compliance with the Guidance for Ethical Treatment of Laboratory Animals, and the experimental protocols were approved by the Institutional Animal Care and Use Committee at the Shanghai Institute of Materia Medica (Shanghai, China). For all animal experiments, LEF or TER was suspended in a 0.5% carboxymethylcellulose sodium solution (CMC-Na+) and administered by intragastric gavage in a volume of 10 mL/kg body weight.

Experimental design

Initially, in in vitro assays, we used rat and human hepatocytes to investigate the interaction between CYPs and LEF, TER. To verify the observed phenomena in in vitro assays, we investigated the contribution of CYPs to LEF-induced hepatotoxicity in HRN mice, and obtained consistent results with these different models, suggesting that the species differences were minimal in this context. Then, we observed TER significantly decreased NTCP expression in human hepatocytes. Therefore, we used rat hepatocytes and NTCP-transfected HEK293 cells to study the effect of TER on NTCP function and got consistent results. Eventually, we conducted in vivo assay in SD rats to confirm the results of in vitro assays. The details of this study design is shown in Figure 1.

Figure 1
figure 1

Flow chart of the study design. LEF, leflunomide; TER, teriflunomide.

PowerPoint slide

Full size image

Isolation and culture of primary rat and human hepatocytes

Primary rat hepatocytes were obtained from SD rats using a two-step collagenase digestion method with some modifications15,16. Cell viability, determined via trypan blue exclusion, was greater than 85%. Viable hepatocytes were plated in culture plates coated with type I rat tail collagen and incubated at 37 °C under 5% CO2. The medium was changed after attachment. For the biliary excretion assay, after 24 h, cells were overlaid with 0.25 mg/mL Matrigel in ice-cold medium without fetal bovine serum (FBS) to form a sandwich configuration during the subsequent culture period, as previously described17. Cryopreserved primary human hepatocytes were purchased from GIBCO. After attachment for 6 h, the plating medium was changed to culture medium without FBS for the subsequent culture period.

Quantitative determination of gene expression via RT-PCR

Primary rat and human hepatocytes were treated with 10 μmol/L LEF or TER for 48 h. Total RNA was isolated with the TRIzol reagent (Life Technologies, CA, USA), and cDNA synthesis was performed using the PrimeScript RT Reagent Kit (Takara, Shiga, Japan). Quantitative analysis of the gene expression of several CYPs and transporters was conducted via real-time PCR using a Qiagen Rotor Gene Q instrument (Qiagen, Germany). β-Actin was used for internal normalization. All the primers (Table 1) were synthesized at Sangon Biotech (Shanghai) Co, Ltd.

Table 1 Sequences of the primers used for real-time reverse transcription (RT-PCR) analyses.
Full size table

Determination of CYP1A2 enzyme activity

The activity of the CYP1A2 enzyme was measured using its specific substrate phenacetin (Phe 2 μmol/L) as a probe18. After human hepatocytes were treated with LEF (10 μmol/L) and TER (10 μmol/L) for 48 h, 2 μmol/L Phe dissolved in medium was added to the cells, followed by an additional 2 h incubation after the hepatocytes were rinsed with PBS. The concentrations of Phe in medium at 0, 0.5, 1, and 2 h were analyzed via LC–MS/MS (LCMS-8030; Shimadzu, Kyoto, Japan). 3-MC (2 μmol/L) was used as a positive control. Enzymatic activity was normalized according to the protein quantification results.

Cell viability assays

Primary rat hepatocytes were seeded into 96-well plates. The hepatocytes were then treated with various concentrations of LEF and TER (10-1000 μmol/L). Then, the assay was conducted following the instruction as previously described19. For the inhibition assay, hepatocytes were pre-incubated with the nonspecific CYPs inhibitor ABT (1 mmol/L) for 1 h.

Metabolism of LEF and TER in rat hepatocytes

Freshly isolated rat hepatocytes were suspended in 12-well plates. Medium containing LEF or TER (10 μmol/L) was added to the hepatocytes. Subsequently, 100 μL samples were collected at 0, 1, 2 and 4 h respectively and added to 300 μL of ice cold methanol to terminate the reaction. The concentration of LEF and TER in primary rat hepatocytes and medium were determined via liquid chromatography–mass spectrometry tandem mass spectrometry (LC–MS/MS) (LCMS-8030; Shimadzu, Kyoto, Japan) and normalized according to the total protein content. For the inhibition study, primary rat hepatocytes were treated with 30 μmol/L SKF for 30 min before incubated with LEF or TER.

Determination of CLint, scaled of LEF in primary hepatocytes

The scaled intrinsic clearance (CLint, scaled) of LEF was estimated based on the rate of LEF disappearance from the incubation medium, as described previously20. The elimination rate constant, k, for LEF was determined by plotting the natural log of the concentration of LEF according to time in minutes. The elimination rate constant was subsequently used to calculate the half-life (T1/2) of LEF according to T1/2=0.693/k. Then, CLint, in vitro can be derived as follows: CLint, in vitro=(0.693/T1/2)*(V/M), where V/M is equal to the incubation volume per 106 cells. CLint, in vitro was scaled to in vivo CLint, scaled using a hepatocellularity of 120×106 cells/g liver and a human liver weight of 20 g/kg body weight.

Pharmacokinetic and safety study of LEF in WT and HRN mice

In all animal experiments, WT and HRN mice were fasted for 12-14 h prior to the experiment.

Following oral administration of LEF (15 mg/kg), 40 μL of blood (heparin sodium as anti-coagulant) was drawn from the caudal vena cava at designated time points. Plasma was obtained via centrifugation at 4 °C and immediately added to 120 μL of ice-cold acetonitrile containing an internal standard. The plasma concentration of LEF and TER were determined via LC-MS/MS (LCMS-8030; Shimadzu, Kyoto, Japan).

For the survival and toxicity experiments, mice were orally administered with LEF at different doses (0, 25, or 50 mg/kg) once a day, and the number of surviving mice was recorded within 7 d after the treatment. Blood samples were obtained before dissection, and the level of AST and ALT in the serum were analyzed. The animals in the control group only received the vehicle (CMC-Na+).

Accumulation of d8-TCA in freshly isolated primary rat hepatocytes and NTCP-transfected HEK293 cells

Rat hepatocytes and NTCP-transfected HEK293 cells, plated for 4 or 24 h respectively, were washed twice with 300 μL of warm HBSS and then incubated with HBSS for 15 min. Next, 300 μL of d8-TCA (5 μmol/L) dissolved in warm HBSS was added, followed by incubation for an additional 15 min after removing the medium. Then, the solution was removed, and the reaction was terminated by washing with ice cold PBS for three times. The concentration of d8-TCA in the cells was determined via LC–MS/MS (LCMS-8030; Shimadzu, Kyoto, Japan). For the inhibition experiment, cells were incubated with d8-TCA (5 μmol/L) and LEF or TER (10-100 μmol/L) simultaneously. Troglitazone (Tro 20 μmol/L) was used as the positive control. The concentration of d8-TCA in cells was normalized according to total protein content.

Effect of TER on NTCP function in rats

Rats were fasted for 12-14 h prior to the experiment. Then, rats divided into four groups through random allocation were orally administered TER once a day, at multiple doses of 0, 6, and 12 mg/kg. Four weeks later, blood samples were obtained, and the serum level of total bilirubin (TBILI) and direct bilirubin (DBILI) were analyzed. The animals in the control group received only the vehicle.

Blood biochemistry

Serum of rat and mice blood samples were obtained for biochemistry analysis. Alanine aminotransferase (ALT), aspartate aminotransferase (AST), TBILI and DBILI were determined using an Automatic Clinical Analyzer (7080, HITACHI Ltd, Tokyo, Japan).

Statistical analysis

Data are expressed as the mean±standard deviation (SD) unless otherwise stated. Differences between two groups were analyzed using Student's t-test. A one-way ANOVA were used to test statistical significance among groups using GraphPad Prism 5.03. Values of P<0.05 were considered statistically significant.

Results

Effects of LEF and TER on CYPs expression in primary hepatocytes

Following treatment with LEF and TER (10 μmol/L), CYP1A1/2 and CYP7A1 mRNA level were significantly increased by LEF in primary rat hepatocytes (Figure 2A). TER only significantly increased CYP1A2 expression about 6-fold. In primary human hepatocytes, LEF only significantly induced CYP1A1/2 expression, while TER induced both CYP1A2 and CYP7A1 expression (Figure 2B). However, we found that although both LEF and TER significantly increased CYP1A2 mRNA levels, only LEF significantly increased CYP1A1/2 function in human hepatocytes (Figure 2C).

Figure 2
figure 2

Effect of LEF and TER on the gene expression and function of CYPs. Rat (A) and human (B and C) hepatocytes were treated with LEF and TER (10 μmol/L). Gene expression was measured and shown as the fold induction (A and B). The function of CYP1A2 was measured using the specific substrate phenacetin (Phe 2 μmol/L) as a probe (C). 3-MC (2 μmol/L) was employed as the positive control. The data are presented as the mean±SD (n=3). bP<0.05, cP<0.01 compared with the control.

PowerPoint slide

Full size image

Inhibition of hepatic CYPs reduced LEF clearance and increased the toxicity of LEF in rat hepatocytes

ABT (1 mmol/L) significantly increased the toxic effect of LEF in primary rat hepatocytes, with reduction in the IC50 value from 409 μmol/L to 216 μmol/L (Figure 3A) being observed. On the contrary, ABT did not show any effect on the cytotoxicity of TER (Figure 3B). Simultaneously, SKF (30 μmol/L) decreased the hepatic intrinsic clearance of LEF, but not TER, in primary rat hepatocytes: the CLint, scaled of LEF decreased from 136 to 73.9 mL·min−1·kg−1 after SKF treatment. And after 4 h incubation, the accumulation of LEF in primary rat hepatocytes increased 3.68-fold after SKF treatment (Figure 3C and 3D). Consistently, LEF decreased the mitochondrial membrane potential (MMP) of primary rat hepatocytes more apparently than TER (supplementary Figure S1).

Figure 3
figure 3

Effect of the nonspecific CYPs inhibitors ABT (1 mmol/L) and SKF (30 μmol/L) on the cytotoxicity and metabolism of LEF and TER. Cell viability following ABT treatment was determined via the MTT assay (A and B). The concentrations of LEF and TER in rat hepatocytes following SKF treatment were determined via LC–MS/MS (C and D). Data are presented as the mean±SD (n=3 and 6). bP<0.05, cP<0.01 (LEF+SKF treatment vs LEF alone).

PowerPoint slide

Full size image

Pharmacokinetic profiles of LEF in WT and HRN mice

The plasma concentration of LEF increased in HRN mice compared with WT mice (Figure 4A). The area under the time-concentration curve (AUC0-t) for LEF in HRN mice was 483±470 ng·mL−1·h, compared with 158±110 ng·mL−1·h in WT mice, while the generation of TER (AUC0-t) decreased compared with WT mice (73.9±39.8 μg·mL−1·h in HRN mice vs 380±296 μg·mL−1·h in WT mice) (Figure 4A and 4B). However, no significant difference between HRN mice and WT mice was observed because of great individual difference.

Figure 4
figure 4

Plasma concentrations of LEF and TER in WT and HRN mice. WT and HRN mice were orally administered with 15 mg/kg LEF. Then, blood samples were obtained, and the concentrations of LEF (A) and TER (B) in plasma were determined. Data are presented as the mean±SD (n=5).

PowerPoint slide

Full size image

Toxicity evaluation of LEF in WT and HRN mice following continuous oral administration

In WT mice, no test article related death was observed in the 50 mg/kg LEF group (two animals died by accidents in the vehicle control group) (Figure 5A). At 25 mg/kg LEF, there were two HRN mice exhibited moribund signs on the day of dissection. The level of AST and ALT in the serum of HRN mice were significantly increased compared with the control group, whereas these values were not increased in WT mice (Figure 5C). When the dose increased to 50 mg/kg LEF, death of the HRN mice was observed beginning on the 4th d after oral administration, and all of the animals had died 5 d later (Figure 5B).

Figure 5
figure 5

Comparison of LEF-induced hepatotoxicity in WT and HRN mice. After the administration of multiple doses of LEF (05 and 50 mg/kg) for 7 d, the survival rates of WT (A) and HRN (B) mice were recorded daily. (C) The serum levels of AST and ALT were determined before dissection on the 7th d in the 25 mg/kg LEF treatment and control groups. Data are presented as the mean±SD (n=6). bP<0.05 (LEF treatment vs control in HRN mice).

PowerPoint slide

Full size image

Effect of LEF and TER on the expression and function of NTCP

Neither LEF nor TER significantly altered the expression of the selected transporters, ie, P-glycoprotein (P-gp), BCRP, bile salt export pump (BSEP) and NTCP in primary rat hepatocytes (Figure 6A). In primary human hepatocytes, neither LEF nor TER had an effect on the expression of BSEP and BCRP, with the exception of a significant reduction in NTCP expression (Figure 6B). However, it was found that only TER, but not LEF, significantly reduced the uptake of d8-TCA in primary rat hepatocytes and NTCP-transfected HEK293 cells (Figure 6C and 6D). In addition, LEF and TER had no effect on the biliary excretion index (BEI) value of d8-TCA in sandwich-cultured rat hepatocytes (SCRHs) (Supplementary Figure S2).

Figure 6
figure 6

Effect of LEF and TER on the gene expression and function of NTCP. Rat (A) and human (B) hepatocytes were treated with LEF and TER. Gene expression was measured and presented as the fold induction. Rat hepatocytes (C) and NTCP-transfected HEK293 cells (D) were co-treated with d8-TCA and LEF or TER, after which the accumulation of d8-TCA in the cells was measured. Troglitazone (Tro-20 μmol/L) was used as the positive control. Data are presented as the mean±SD (n=3). cP<0.01 compared with the control.

PowerPoint slide

Full size image

Consistently, it was found that TER long term treatment significantly increased the level of TBILI and DBILI in the serum of female rats (Table 2).

Table 2 Levels of TBILI and DBILI in the serum of rats following the oral administration of TER (1–12 mg/kg) for 4 weeks.
Full size table

Discussion

Several studies have demonstrated that LEF is a ligand of the Aryl Hydrocarbon Receptor (AhR)21,22,23, which regulates the expression of CYP1A1/2. Consistently, we determined that LEF significantly induced CYP1A1/2 expression in both primary rat and human hepatocytes (Figure 2). CYP1A1/2 could also be induced by TER in human hepatocytes to a less extent. However, it seems the CYP inhibition instead of CYP induction plays critical roles in the toxicity process of LEF. In this study, it was found that nonspecific CYPs inhibitors SKF and ABT could significantly increase LEF accumulation and cytotoxicity in primary rat hepatocytes (Figure 3). Shi et al claimed that CYPs inhibitors could enhance the liver toxicity of both LEF and TER in primary rat hepatocytes5; however, our data revealed that CYPs only directly contributed to LEF-related, rather than TER-related, detoxification in primary rat hepatocytes. In fact, the mitochondrial membrane potential of primary rat hepatocytes did show more sensitive to LEF than TER (supplementary Figure S1).

Many reports demonstrated that the polymorphism of several CYPs, such as CYP1A2 and CYP2C19, could affect the liver toxicity of LEF in patients with rheumatoid arthritis24,25,26, which could be indirect evidences that CYPs play a critical role in the detoxification of LEF in the body. Our studies with HRN mice are the first direct in vivo evidences that hepatic CYPs are involved in the clearance and toxicity of LEF. The plasma concentration and AUC0-t value of LEF were found to be much higher in HRN mice than in WT mice following single dose of LEF (15 mg/kg), and these mice also showed lower TER exposure in their plasma (Figure 4). Unexpectedly, one of the HRN mice died 10 h after LEF was administered, which may be attributed to the broad variability of LEF pharmacokinetic parameters, as reported previously27. This variability may also explain the great variation in the AUC0-t value of LEF and TER observed in this study. These results suggested that the plasma concentration of LEF may increase when CYPs were knocked out, which could lead to enhanced liver toxicity. In fact, after continuous dosing, at the 25 mg/kg LEF, serum AST and ALT were significantly increased only in the HRN mice (Figure 5C). And the survival rate of HRN mice was significantly lower than that of WT mice when LEF dosage increased to 50 mg/kg for 7 d.

Since LEF and TER are both high affinity substrates of BCRP9, their impacts on liver transporters expression were also explored. It was found they had no effects on the mRNA expression of the selected efflux transporters (BCRP, BSEP), with the exception of the down-regulation of NTCP in human hepatocytes after 48 h of treatment (Figure 6B). Primary rat hepatocytes and NTCP-transfected HEK293 cells were used to further investigate the influences of LEF and TER on the function of NTCP. It was found that TER, but not LEF, could significantly reduce the accumulation of d8-TCA in primary rat hepatocytes, and it inhibited d8-TCA accumulation in NTCP-transfected HEK293 cells in a dose-dependent manner (Figure 6C and 6D). Neither LEF nor TER could directly inhibit the function of BSEP (Supplementary Figure S2). Furthermore, only TER significantly induced the expression of CYP7A1, a rate-limiting enzyme in the biosynthesis of bile acid, by approximately 6-fold in human hepatocytes (Figure 2B). Many drugs disrupt homeostatic mechanisms by directly inhibiting bile acid transporters, such as NTCP or BSEP, leading to bile acid-induced hepatotoxicity28,29,30. Our results implied TER may have similar impacts on the homeostasis of bile acid.

Vrenken et al (2008) reported that TER protects rat hepatocytes from bile acid-induced apoptosis31. Our data may provide an alternative explanation for this observation: the protective role of TER may arise from its inhibition of NTCP function, which leads to less bile acid accumulation in hepatocytes and less apoptosis. However, from another perspective, down-regulation of NTCP and disturbance of bile acid circulation may lead to liver toxicity over the long term32,33,34. In fact, the elimination half-life of TER is approximately 2 weeks, which is thought to result from a combination of extremely low hepatic clearance and enterohepatic recycling27,35. Therefore, the toxicological effect of TER on NTCP expression and function may be sustained over long periods of time. In fact, the level of TBILI and DBILI in female rat serum increased significantly after four weeks treatment of TER (12 mg/kg) (Table 2). This finding implied that TER may have impact on bile acid circulation, which may lead to cholestasis and bile duct injury36. However, since bile acid is the direct biomarker of cholestasis37, bile acid levels in serum will be determined in the future to further verify this hypothesis. In addition, this phenomenon was not observed in male animals, which may be attributed to sex-specific differences in the tissue distribution of TER27.

In summary, CYPs are critical to the detoxification process of LEF-induced liver toxicity. Inactivate hepatic CYPs significantly increase the concentration of LEF in primary hepatocytes and HRN mice, enhancing the hepatotoxicity induced by LEF. Transporters, rather than CYPs, play a unique role in the liver toxicity induced by TER, the major metabolite of LEF. TER significantly decreases the expression and function of NTCP, which may disturb bile acid circulation and cause potential bile acid-related liver issues.

Author contribution

Lei-lei MA, Yang LUAN, Guo-yu PAN participated in research design; Lei-lei MA, Jing WANG, Zhi-tao WU, Le WANG, Chen CHEN, Xuan NI, Yun-fei LIN, Yi-yi CAO conducted experiments; Lei-lei MA, Zhi-tao WU, Xue-feng ZHANG performed data analysis; Lei-lei MA, Yang LUAN, Guo-yu PAN wrote the manuscript.

Abbreviations

ABT, Aminobenzotriazole; ALT, alanine aminotransferase; AST, aspartate aminotransferase; BSEP, bile salt export pump; BEI, biliary excretion index; BCRP, breast cancer resistance protein; CMC-Na+, carboxymethylcellulose sodium; CYPs, cytochrome P450 enzymes; DBILI, direct bilirubin; DDI, drug-drug interaction; HRN, hepatic cytochrome P450 reductase null; LEF, leflunomide; 3-MC, 3-methylcholanthrene; P-gp, P-glycoprotein; Phe, phenacetin; SKF, proadifen; SCRHs, sandwich-cultured rat hepatocytes; CLint, scaled, scaled intrinsic clearance; NTCP, sodium/bile acid cotransporter; TER, teriflunomide; TMRE, tetramethylrhodamine ethyl ester; TBILI, total bilirubin; Tro, troglitazone.