Rodent-specific induction of cholestasis and cholangioma through the reduction of bile salt export pump by bardoxolone methyl

Abstract

Bardoxolone methyl, also known as CDDO-Me or RTA 402, was a new investigational drug that improves the estimated glomerular filtration rate by activating the Keap1-Nrf2 pathway. Bardoxolone methyl is a synthetic triterpenoid compound derived from oleanolic acid (OA). OA-mediated cholestasis has not been reported in the clinical treatment of treat liver disorders but has been reported in mice. Cholestasis can be transient (e.g., during pregnancy) but can also be chronic and is a risk factor for hepatobiliary carcinoma. Therefore, it is important to evaluate and understand the risk of drug-induced cholestasis and its mechanisms, including species differences. We evaluated the effects of bardoxolone methyl on the in vivo hepatobiliary systems in rats and monkeys, as well as sandwich-cultured hepatocytes. Bardoxolone methyl was administered daily to rats and monkeys for 26 or 52 weeks, respectively. As a result, bardoxolone methyl was associated with the development of cholestasis and cholangioma in rats but not in monkeys. In an in vitro evaluation using sandwich-cultured hepatocytes treated with bardoxolone methyl, cholestasis was observed in rat hepatocytes, but not in monkey or human hepatocytes. A gene expression analysis showed that rat-specific cholestasis was caused by the reduction of the bile salt export pump gene expression after treatment with bardoxolone methyl. These results strongly suggest that the effects of bardoxolone methyl on the hepatobiliary system differ among animal species, especially between rodents and non-rodents. In conclusion, the risk of cholestasis and cholestasis-derived carcinogenicity associated with bardoxolone methyl are expected to be quite low in humans.

INTRODUCTION

Bardoxolone methyl, also known as CDDO-Me or RTA 402, was a new investigational drug that activates the Keap1-Nrf2 pathway (Rojas-Rivera et al., 2012). Bardoxolone methyl promotes the nuclear translocation of Nrf2 and transcription of its downstream genes in mice (Shelton et al., 2015), rats (Kadıoğlu et al., 2021), monkeys (Reisman et al., 2012), and humans (Pergola et al., 2011b; 2011a; Chin et al., 2018; Nangaku et al., 2020) by interacting with Keap1, the Nrf2 inhibitory molecule. The Nrf2 pathway promotes antioxidative and cytoprotective responses, which result in protection from the oxidative stress, pro-inflammatory signaling, and mitochondrial dysfunction (Itoh et al., 1999; Dinkova-Kostova et al., 2005). Because of Nrf2-mediated attractive biological responses, Nrf2-activating drugs such as oleanolic acid (OA) have been developed for liver disorders, including cholestasis, in China, and have been used with dimethyl fumarate for the treatment of relapsing multiple sclerosis (Dinkova-Kostova et al., 2005; Yates et al., 2007; Lin et al., 2016; Lipton and Satoh, 2017). In particular, bardoxolone methyl has been reported to improve the glomerular filtration rate (GFR) and estimated GFR (eGFR); therefore, bardoxolone methyl was expected to improve the renal function in patients with chronic kidney diseases due to type 2 diabetes mellitus, Alport syndrome, and autosomal dominant polycystic kidney disease (Pergola et al., 2011a; 2011b; Chin et al., 2018; Nangaku et al., 2020).

Bardoxolone methyl is a synthetic triterpenoid compound derived from OA, which is associated with the side effect of cholestasis. Species differences have been reported in this side effect. In the clinical use of OA in China, no clinical signs (e.g., jaundice) and biochemical evidence (elevation of alkaline phosphatase [ALP], gamma-glutamyltransferase [γGTP] and total bilirubin [T-Bil]) were reported, while these findings have been observed in mice after repeated administration (Dinkova-Kostova et al., 2005; Heathcote, 2007; Yates et al., 2007; Lin et al., 2016). However, the mechanisms underlying the species differences are still unknown. Species differences in toxicological findings are often observed because non-clinical toxicity studies using at least two animal species, rodents and non-rodents, are generally required for drug development to evaluate the drug’s toxicological profile. A poor understanding of toxicological profiles, including species differences and their mechanisms, makes it difficult to evaluate whether each toxicity is likely to occur in humans. Thus, for properly assessing the risks of toxicity in humans, it is important to understand the mechanisms of toxicities, including inter-species differences such as OA-mediated cholestasis.

Cholestasis is the accumulation of bile acids, which can result from suppressing the function of just a single specific gene, and it is a risk factor for neoplastic changes. Bile acids are synthesized from cholesterol in the hepatocytes of the liver and are transported to the bile duct through bile acid transporters. Bile acids are toxic themselves, so their levels in the liver are regulated strictly to prevent toxicity to hepatobiliary tissues. Some nuclear receptors are known to be activated to regulate bile acid homeostasis by directly binding to bile acids. Among them, farnesoid X receptor (FXR) is the most important nuclear receptor because it promotes bile acid clearance by controlling the expression of bile acid transporters. Fxr knockout mice show elevated serum bile acid concentrations (Sinal et al., 2000) and hepatocarcinoma induced by cholic acid (Kong et al., 2016), and FXR mutations cause progressive familial intrahepatic cholestasis in humans (Gomez-Ospina et al., 2016). Bile salt export pump (BSEP) is an important bile acid transporter that exports bile acid from the hepatocytes to the bile duct. In addition to cholestasis, cholestasis-induced carcinoma is observed in both humans with BSEP mutations and Bsep knockout mice (Zamek-Gliszczynski et al., 2003; Okada et al., 2008, 2022). Therefore, cholestasis could be caused by the suppression of the function of only one specific gene and may be a risk factor for cholestasis-induced hepatobiliary carcinoma.

Drug-induced cholestasis (DIC) is a common hepatotoxicity caused by multiple factors, and a useful evaluation has been developed to evaluate the risk of DIC. In Sweden, cholestatic alone or mixed cholestatic liver injury are observed in almost half of 784 liver injury cases between 1970 and 2004 (Björnsson and Olsson, 2005). In the United States, 20% of the jaundice cases among the elderly population are attributed to DIC (Lewis, 2000). DIC results from the dysfunction of bile acid homeostasis, which can be attributed to a wide variety of causes, including direct inhibition of bile acid transporters, altered nuclear receptor regulation, inflammation, and other factors (Yang et al., 2013). To assess the risk of DIC, it is useful to evaluate effects of drug candidates on hepatobiliary parameters/systems in in vivo toxicity studies, which are usually conducted during drug development (Heathcote, 2007; Padda et al., 2011). In addition, in vitro matrigel/collagen sandwich-cultured hepatocyte is useful for evaluating the risk of DIC and elucidating its mechanisms and species differences, because cholestasis can be evaluated using fluorescence microscopy in animal and human hepatocytes with bile canaliculi expressing transporters (Swift et al., 2010). These evaluations of the risk of DIC are especially important for drugs such as bardoxolone methyl, which are expected to be used over a lifetime (Lesage et al., 2001; Padda et al., 2011; Razumilava and Gores, 2014) because DIC is not only a tentative side effect but also a potential cause of hepatobiliary carcinoma, as mentioned above.

In this study, we aimed to evaluate the hepatobiliary toxicity of bardoxolone methyl to facilitate its safe use in humans. To achieve this purpose, we performed toxicity studies in rats and monkeys for 26 and 52 weeks, respectively. In addition, we performed in vitro evaluations of cholestasis using matrigel/collagen sandwich-cultured hepatocytes derived from multiple animal species and a gene expression analysis to elucidate the mechanisms of bardoxolone methyl-induced cholestasis and its species differences, as well as for the extrapolation of the findings to humans.

MATERIALS AND METHODS

Test articles and reagents

Bardoxolone methyl and RTA 401 (a bardoxolone methyl carboxylate metabolite at C17-acid specific to rodents used for evaluating metabolite-mediated cholestasis) were supplied by Reata Pharmaceuticals. OA and carboxy dichlorofluorescein diacetate (CDFDA) were purchased from Fujifilm Wako Pure Chemical Corporation (Osaka, Japan). MK571 (an MRP2 inhibitor used as the positive control for the cholestasis evaluation) was purchased from Cayman Chemical (Ann Arbor, MI, USA).

Toxicity study in animals

Bardoxolone methyl (amorphous micronized) was solved/suspended in sesame oil and orally administered to male and female Crl:CD(SD) rats (Charles River Laboratories, Portage, MI, USA; n=35, each) from the age of 8 weeks for 26 weeks, with an interim necropsy (n=10, each) after 13 weeks of study, followed by a 4-week recovery period. It was also orally administered to male and female cynomolgus monkeys (Covance Research Products, Inc., Denver, PA, USA; country of origin, Vietnam; n=9, each) for 52 weeks, followed by a 4-week recovery period. At the initiation of study, the dosages were 0 (sesame oil), 1, 3 and 10 mg/kg/day for male rats, 0 (sesame oil), 0.1, 1.5 and 4.5 mg/kg/day for female rats, and 0 (sesame oil), 30, 100 and 300 mg/kg/day for male and female cynomolgus monkeys. However, the dosages for male rats in each group were decreased to 0.5, 1.5 and 4.5 mg/kg/day (represented as 0.5/1, 1.5/3 and 4.5/10 mg/kg/day in this article) due to unexpected mortality observed on Days 35, 35 and 41, respectively. The dosages for the monkeys in the low-dose and middle-dose groups were also decreased to 5 and 30 mg/kg/day (represented as 5/30 and 30/100 mg/kg/day), respectively, as no significant differences were observed across the three doses.

Blood samples for blood chemistry were collected into tubes without anticoagulant via the vena cava after carbon dioxide inhalation following overnight fasting (rats) or via the femoral vein following overnight fasting and water deprivation (monkeys). On the day of necropsy, rats were euthanized by carbon dioxide inhalation followed by exsanguination via the abdominal vena cava, while monkeys were euthanized by sedation with ketamine, followed by an intravenous overdose of sodium pentobarbital solution via the saphenous or cephalic vein and exsanguination by severing the femoral vessels.

The following items were evaluated: observations of morbidity, mortality, clinical signs; measurement of body weight; examinations of clinical pathology (blood), necropsy, organ weights, and histopathology at 13 weeks for rats and 52 weeks for monkeys. All animal experiments were approved by the institutional animal ethics committee of the testing facility (MPI Research, Inc., Mattawan, MI, USA) and were performed in accordance with its guidance.

Toxicokinetic analysis of bardoxolone methyl in whole blood concentrations

Blood samples for the toxicokinetic (TK) analysis were collected from groups of animals via the orbital sinus after anesthesia with carbon dioxide/oxygen inhalation on Day 223 (rats) or from all surviving animals via the femoral vein on Day 350 (monkeys), at 0.5 (rat only), 1, 2, 4, 6, 8, 12, and 24 hr after dosing without fasting. Samples were placed in tubes containing K₃EDTA anticoagulant. The bardoxolone methyl concentrations in the whole blood were determined using an LC-MS/MS system (Agilent 1200 Series HPLC System). The TK parameters were calculated for bardoxolone methyl from mean concentration-time data in the test species using a non-compartmental module of WinNonlin® Enterprise, which was validated by MPI Research.

Cell culture

Cryopreserved rat and human hepatocytes were purchased from Biopredic International (Rennes, France), and monkey hepatocytes were purchased from Biopredic International and GIBCO (USA). Cells were seeded at 1-1.2×10⁵ cells/well in collagen I-coated 48-well plates and were cultured according to the protocols of Biopredic International. Three days (rat hepatocytes), 4 days (monkey hepatocytes) or 8 days (human hepatocytes) after cell seeding, bile canaliculi-formed cells were used for the following evaluation.

Cholestasis evaluation

Cells were treated with test articles at the final concentration (bardoxolone methyl, RTA 401 and OA, 1 μmol/L; MK571, 100 μmol/L) of 0.1% dimethyl sulfoxide (DMSO) in transport buffer (125 mmol/L NaCl, 4.8 mmol/L KCl, 5.6 mmol/L d-glucose, 1.2 mmol/L CaCl₂, 1.2 mmol/L KH₂PO₄, 1.2 mmol/L MgSO₄, 25 mmol/L HEPES, adjusted to pH 7.4 with NaOH) containing 10 μmol/L CDFDA. Thirty minutes after incubation at 37°C under 5% CO₂, cells were washed with Dulbecco’s phosphate-buffered saline containing CaCl₂ and MgCl₂ (D-PBS (+), GIBCO), then fluorescence microscopy was performed to evaluate the fluorescence intensity as the bile acid excretion ability.

Gene expression analysis

Matrigel/collagen I sandwich-cultured rat and human hepatocytes were treated with bardoxolone methyl, RTA 401 or OA at the concentrations of 0.1 and 1 μmol/L (final concentration of 0.1% DMSO) in sandwich maintenance medium (Biopredic International) for 24 hr (n=5). Subsequently, RNA was extracted using an RNeasy Mini Kit (QIAGEN), and cDNA was prepared using a PrimeScript RT reagent kit (TaKaRa) according to the manufacturers’ protocols. For quantitative real-time PCR (qPCR), TaqPath™ qPCR Master Mix CG and TaqMan™ Gene Expression Assays (Applied Biosystems) were used according to the manufacturer’s protocol. Each sample was evaluated in duplicate. qPCR was performed with the 7500 Fast RT-PCR system (Applied Biosystems): Step 1, 50°C for 2 min; Step 2, 95°C for 20 sec; Step 3, 95°C for 3 sec; Step 4, 60°C for 30 sec (Steps 3 and 4 were repeated for 40 cycles). The evaluated genes and primer IDs are shown in Table 1. The expression levels of genes in each sample were calculated by the ΔΔCt method.

Table 1. The evaluated genes and primers.

RESULTS

Evaluation of hepatobiliary effects in animals

To evaluate the effects of bardoxolone methyl on the hepatobiliary system, the compound was orally administered to rats and monkeys for 26 and 52 weeks, respectively. From 13 weeks of dosing, significantly increased mortality (Supplemental Table 1) accompanied by abnormal clinical signs (Supplemental Table 2) and decreased body weight gain (Supplemental Fig. 1) were observed in rats in the high-dose group; all remaining animals in the high-dose group were necropsied at the end of 13-week administration. In the following sections, the results of males are representatively shown because most of these findings, including mortality, were commonly present in both sexes (data not shown).

Regarding blood chemistry (Fig. 1), the 13-week administration of bardoxolone methyl significantly increased aspartate aminotransferase (AST), alanine aminotransferase (ALT), γGTP, T-Bil, and cholesterol (Cho) in male rats in the high-dose group in comparison to the control group. ALT in the middle-dose group, and γGTP and Cho in the low- and middle-dose groups were also significantly increased by bardoxolone methyl administration in rats. No significant changes were observed in ALP in rats in any of the bardoxolone methyl groups. In monkeys, bardoxolone methyl administration significantly reduced ALP, T-Bil, and Cho in the middle- and high-dose groups. T-Bil was also significantly decreased by bardoxolone methyl administration in the monkeys in the low-dose group. No significant changes were observed in AST, ALT, or γGTP in monkeys in any of the bardoxolone methyl groups.

Fig. 1

Blood chemistry results in rats and monkeys. Bardoxolone methyl was orally administered to male Crl:CD(SD) rats for 13 weeks and male cynomolgus monkeys for 52 weeks. After a dosing period of 13 weeks (rats, n=10) and 52 weeks (monkeys, n=6), blood levels of AST, ALT, ALP, γGTP, T-Bil, and Cho were analyzed. Plots show the values of individual animals. Bars indicate the mean value of each group. *, p<0.05; **, p<0.01 (a Dunnett’s test or Welch’s t-test)

The absolute and relative liver weights were increased by bardoxolone methyl administration in all dose groups in rats and in the middle- and high-dose groups in monkeys (Fig. 2). At necropsy, cyst, discoloration (red and tan), and/or enlargement of the liver were observed only in rats in middle- and high-dose groups (Table 2). There were no abnormal findings in the liver or gallbladder in monkeys. Histopathology (Fig. 3 and Table 3) revealed benign cholangioma (Fig. 3B) and bile duct hyperplasia (Fig. 3C) in the liver in rats in the bardoxolone methyl groups. In particular, bile duct hyperplasia was noted in all of the bardoxolone methyl groups. In monkeys, bile duct hyperplasia and inflammation (Fig. 3E) in the liver were noted in the high-dose group and hypertrophy/hyperplasia were noted in the gallbladder mucosa (Fig. 3G) in all bardoxolone methyl dose groups. Hypertrophy in the gallbladder mucosa was characterized by an increased height of mucosal epithelial cells, which was associated with an increased cell count (hyperplasia) and an increased number and height of mucosal projections (folds) into the lumen (Fig. 3G).

Fig. 2

Absolute and relative liver weights in rats and monkeys. For the animals shown in Fig. 1, organ weights were measured at necropsy. Plots show the values of individual animals. Bars indicate the mean value of each group. **, p<0.01 (a Dunnett’s test or Welch’s t-test)

Table 2. Necropsy results of liver in rats and monkeys.

Fig. 3

Histopathology results in rats and monkeys. For the animals shown in Fig. 1, microscopic histopathology was conducted with hematoxylin/eosin staining. A representative image in each test condition is shown for the liver from rats treated with vehicle (sesame oil, A) and with bardoxolone methyl (1.5/4.5 mg/kg; B, C), for the liver from monkeys treated with vehicle (sesame oil, D) and with 300 mg/kg bardoxolone methyl (E), and the gallbladder from monkeys treated with vehicle (F) and bardoxolone methyl (G).

Table 3. Histopathological changes in rats and monkeys.

At the end of a 4-week recovery period after the 26- and 52-week administration periods in rats and monkeys, respectively, increased liver weight, macroscopic hepatic cysts, microscopically benign cholangiomas, biliary cysts, and bile duct hyperplasia were noted in male rats in the middle-dose group, and bile duct hyperplasia was noted in male monkeys in the middle- and high-dose groups (Supplemental Table 3). The other changes observed in the administration period recovered after the recovery period.

In the TK evaluation of blood concentrations of bardoxolone methyl, the AUC_0-24 in rats in the middle-dose group was almost same as that in monkeys in the low-dose group, and was 2-3 times lower than that in monkeys in the middle- and high-dose groups and that observed in clinical trials in humans (Table 4).

Table 4. The results of TK analysis.

Collectively, bardoxolone methyl was associated with stronger hepatotoxic changes in rats than in monkeys, while the exposure to bardoxolone methyl was lower in rats than in monkeys. The highest dose rat group was not evaluated due to a deterioration of their general condition.

In vitro cholestasis evaluation using cultured hepatocytes

In the bardoxolone methyl toxicity studies in animals, cholestatic changes were observed only in rats. Next, to evaluate the effects of bardoxolone methyl on the bile acid excretion and its species differences, we conducted in vitro cholestasis assays using matrigel/collagen sandwich-cultured hepatocytes from multiple animal species. In this assay, CDFDA passively diffuses into hepatocytes and is hydrolyzed to carboxydichlorofluorescein (CDF), a fluorescent substance mimicking bile acid (Zamek-Gliszczynski et al., 2003). Intracellular CDF is excreted into bile canaliculi through bile acid transporters. Thus, cholestasis can be assessed in vitro using CDF as an indicator of bile acid excretion. To confirm that the cholestasis assay was functioning properly, MK-571, an MRP2 inhibitor, was used as a positive control. To address whether the cholestatic changes were due to metabolite- and/or structure-dependent toxicity, RTA 401 (a bardoxolone methyl carboxylate metabolite specific to rodents) and OA were subjected to the assay together with bardoxolone methyl.

Matrigel/collagen sandwich-cultured rat, monkey, and human primary hepatocytes were treated with test articles in the CDFDA-containing buffer (Fig. 4). All the hepatocytes treated with vehicle showed the excretion of CDF into the bile canalicular lumen. MK571, an MRP2 inhibitor as the positive control for the assay, caused intracellular accumulation of CDF, indicating the inhibition of bile acid excretion. Three triterpenoid compounds, bardoxolone methyl, RTA 401, and OA, also showed the intracellular accumulation of CDF in rat hepatocytes. On the other hand, in monkey and human hepatocytes, these compounds showed the excretion of CDF into the bile canaliculi, indicating no inhibition of bile acid excretion. These results suggest that the inhibitory effects on bile acid excretion of the triterpenoid compounds are specific to rat hepatocytes.

Fig. 4

In vitro evaluation of cholestasis using cultured hepatocytes. Rat, monkey, and human hepatocytes were cultured with matrigel/collagen for 3 d (rat), 4 d (monkey), or 8 d (human). The cells were treated with test articles (bardoxolone methyl, RTA 401 and OA, 1 μmol/L; MK571, 100 μmol/L), or vehicle (0.1% DMSO) in transport buffer containing 10 μmol/L CDFDA and were incubated at 37°C and 5% CO₂ for 30 min. The fluorescence intensity was evaluated by fluorescence microscopy as the bile acid excretion ability. The representative image in each test condition is shown.

Gene expression analyses using hepatocytes

Next, we performed gene expression analyses of Nrf2- and bile acid excretion-related genes to investigate the effects of bardoxolone methyl on bile acid homeostasis. The genes evaluated included those encoding Nrf2 and its target NAD(P)H quinone dehydrogenase 1 (NQO1), as an Nrf2-related gene; FXR and its targets: CYP7A1, BSEP, sodium taurocholate co-transporting polypeptide (NTCP), solute carrier organic anion transporter family member (OATP) 1B3 (OATP1B3), OATO1B2, and multidrug resistance protein 4 (MRP4), as bile acid uptake/efflux transporters; (Table 1) (Yang et al., 2013).

First, we conducted a preliminary study to determine an evaluation time point, and 24 hr was chosen as the treatment condition because increases in the NQO1/Nqo1 expression were observed in both rat and human hepatocytes after 24 hr of treatment (Supplemental Fig. 2). Then, rat and human hepatocytes were treated with bardoxolone methyl, RTA 401, or OA at 0.1 and 1 μmol/L for 24 hr, and mRNA levels were determined (Fig. 5). Increased NQO1/Nqo1 mRNA levels and decreased NRF2/Nrf2 and FXR/Fxr mRNA levels were observed in both rat and human hepatocytes treated with bardoxolone methyl and RTA 401. Bsep mRNA levels were decreased with bardoxolone methyl and OA, but not with RTA 401, in rat hepatocytes (Fig. 5a). On the other hand, BSEP mRNA levels were unchanged by bardoxolone methyl and OA, and were increased by RTA 401 in human hepatocytes (Fig. 5b). Increases in Mrp4 mRNA levels with all three test articles and in Nrf2, Nqo1 and Fxr mRNA levels with OA were also observed in rat hepatocytes, while no remarkable changes in these genes were observed in human hepatocytes. Decreased NTCP mRNA levels were only observed with bardoxolone methyl in human hepatocytes. No other obvious differences were observed between human and rat hepatocytes.

Fig. 5

Gene expression analysis using cultured rat and human hepatocytes. Rat (A) and human (B) hepatocytes were cultured with matrigel/collagen for 3 d (rat) or 8 d (human). The cells were treated with vehicle (0.1% DMSO), bardoxolone methyl, RTA 401, or OA at concentrations of 0.1 and 1 μmol/L in culture medium for 24 hr (n = 5, each). RNA extraction and cDNA preparation were performed, followed by qPCR to determine mRNA levels. Each sample was evaluated in duplicate. Bars indicate mean values, and error bars indicate standard deviation.

DISCUSSION

In this study, we evaluated the effect of bardoxolone methyl on the hepatobiliary system. First, we conducted a long-term nonclinical toxicity study using rats and monkeys. In rats, neoplastic lesions (e.g., cholangioma) with elevated hepatic parameters (e.g., γGTP) were observed in the liver. Elevated γGTP is known to be an indicator of chronic drug-induced cholestasis and a biomarker for intrahepatic cholangiocarcinoma (Yin et al., 2013). It has been reported that cholestasis induces cholestasis-related hepatocellular carcinoma and cholangiocarcinoma in Fxr and Bsep knockout mice due to altered bile acid composition-related damage to hepatocytes and cholangiocytes (Sinal et al., 2000; Okada et al., 2008, 2022; Kong et al., 2016). Therefore, the present study suggests that neoplastic lesions such as cholangioma observed in rats were caused by cholestasis.

In contrast to rats, neither elevated hepatic parameters nor neoplastic lesions were observed in monkeys, suggesting a species difference in bardoxolone methyl-induced toxicity between rats and monkeys. In monkeys, bile duct hyperplasia in the liver and hypertrophy/hyperplasia in the gallbladder mucosa were observed. However, these changes showed a recovery trend after drug withdrawal. In addition, no hepatobiliary parameters were elevated in monkeys, suggesting that no occurrence of cholestasis was observed in monkeys. Therefore, given that neoplastic changes are generally irreversible, the hepatobiliary histopathological changes in monkeys are different from those in rats and are not considered to be neoplastic lesions, suggesting species differences in bardoxolone methyl-induced toxicity.

The TK analysis of bardoxolone methyl supported the hepatobiliary toxicity observed in rats as being rat-specific. The AUC_0-24 of bardoxolone methyl was clearly higher in the high-dose group of monkeys than in any group of rats. These results confirmed that bardoxolone methyl showed substantial toxicity in rats despite low exposure and that monkeys tolerated bardoxolone methyl well, even under high exposure. Therefore, bardoxolone methyl-induced cholestasis and related neoplastic lesions were possibly rat-specific changes and were unlikely to occur in monkeys.

The species differences between rats and monkeys in bardoxolone methyl-mediated toxicity could be due to the species differences in the mechanism of action of bardoxolone methyl. Nrf2 activation counteracts cholestatic liver injury via stimulation of hepatic defense systems (Okada et al., 2009), and cholestasis without hepatic injury has been observed in Nrf2-deficient mice (Reisman et al., 2009). These reports indicate that Nrf2 plays an important role in preventing cholestasis and associated hepatic injury. Therefore, the cholestasis-related neoplastic changes in rats were not caused by Nrf2 activation, which is the pharmacological activity of bardoxolone methyl, but rather by off-target toxicity of bardoxolone methyl. In addition, it has been reported that Nrf2 activation increases bile flow in mice with the low expression of Keap1 (Reisman et al., 2009), and no neoplastic lesions were observed in these mice in a 2-year longevity study (Taguchi et al., 2010). Inchinkoto, a herbal medicine and Nrf2 activator, also increased bile flow in rats via Nrf2 activation (Okada et al., 2022). Furthermore, ursodeoxycholic acid, another Nrf2 activator, showed similar hepatic histopathological findings in monkeys to those observed with bardoxolone methyl. These reports suggest that Nrf2 activation increases bile flow, which may result in an increase in the number and size of bile ducts and peribiliary vessels as an adaptive change, not a neoplastic change. Therefore, the pathological findings observed in the hepatobiliary system of monkeys could be attributed not to hepatobiliary toxicity, but to changes resulting from increased bile flow due to Nrf2 activation as the pharmacological action of bardoxolone methyl. Collectively, the species differences in bardoxolone methyl-related changes between rats and monkeys may have resulted from the differences in the mechanisms of action: Nrf2-dependent changes in monkeys and Nrf2-independent toxicity in rats. In other words, bardoxolone methyl-mediated cholestasis in rats could be caused by Nrf2-independent mechanisms, such as by rodents-specific metabolites and triterpenoid structures. In addition, bardoxolone methyl-mediated and Nrf2-independent cholestasis was considered unlikely to occur in humans because no cholestatic findings were observed in the clinical use of Nrf2 activators with triterpenoid structures, bardoxolone methyl or OA (Lin et al., 2016; Lewis et al., 2021).

In vitro evaluation using sandwich-cultured hepatocytes corroborated the idea that cholestasis observed in rats was rat-specific and caused by compounds with a triterpenoid structure, including bardoxolone methyl or its rodents-specific metabolite, suggesting that the toxicity observed in rats has low clinical extrapolability to humans. All of the tested triterpenoid compounds induced cholestasis in rat hepatocytes, but not in monkey or human hepatocytes. The results obtained in vitro were consistent with those obtained in in vivo studies. These results suggest that triterpenoid compounds might inhibit bile acid excretion and induce cholestasis as a rat-specific triterpenoid class effect rather than Nrf2 activation and metabolite-mediated toxicity. Under the conditions in this study, no cholestasis was observed in human hepatocytes, suggesting that bardoxolone methyl-induced cholestasis is unlikely to occur in humans. In fact, these results are consistent with a report showing that no hepatic parameters indicative of cholestasis were observed in human clinical trials (Lewis et al., 2021). Therefore, the clinical extrapolability of bardoxolone methyl-induced cholestasis and neoplastic lesions observed in rats to humans is considered low.

The gene expression analysis revealed that the cholestasis observed in rats may have been caused by the decreased expression of BSEP, which is one of the most important bile acid transporters, and its inhibition, mutations or deficiency are known to cause cholestasis and cholangioma (Knisely et al., 2006; Scheimann et al., 2007; Stieger, 2010). The Bsep expression was decreased only in rat hepatocytes when treated with bardoxolone methyl, which is consistent with the absence of cholestasis in human cells. The expression of Mrp4, encoding a bile acid transporter on the hepatocyte basolateral membrane (Pérez-Pineda et al., 2021), was increased in rat hepatocytes treated with all three triterpenoid compounds; however, it was probably insufficient to prevent cholestasis considering that cholestasis was observed in bile acid-fed Fxr-null mice and progressive familial intrahepatic cholestasis 2 (PFIC2), where the expression of Mrp4 and Bsep was induced and decreased, respectively (Keitel et al., 2005; Zollner et al., 2006). Therefore, bardoxolone methyl-induced cholestasis may be mainly caused by decreased BSEP levels, which was specific to rats. These results further support that the risk of cholestasis and cholestasis-induced neoplastic lesions is low in humans.

It remains to be elucidated why the expression of BSEP was reduced only in rats treated with bardoxolone methyl, but it is possibly due to a rodent-specific class effect of the triterpenoid structure. Similar to bardoxolone methyl, cholestasis was observed in mice treated with OA, but no changes suggestive of cholestasis have been observed in humans treated with OA (Dinkova-Kostova et al., 2005; Yates et al., 2007; Lin et al., 2016; Lipton and Satoh, 2017). One hypothesis regarding these species differences is that the triterpenoid structure affects bile acid homeostasis. Terpenoid compounds bind to FXR, and OA is known to inhibit the functions of FXR (Liu and Wong, 2010). Upon binding to bile acids, FXR activates the expression of its target genes, including BSEP (Jiang et al., 2021). It is also known that Nrf2 activation upregulates the BSEP gene expression in humans, but not in mice or rats (Okada et al., 2008, 2009; Weerachayaphorn et al., 2009). These findings suggest that triterpenoid compounds directly bind FXR to inhibit its function and decrease the expression of BSEP.

In the meantime, the expression of BSEP is compensated by FXR-independent Nrf2 activation in humans (Weerachayaphorn et al., 2009). Due to this mechanism, cholestasis caused by triterpenoid compounds was observed in mice and rats, but not in humans and monkeys. This is consistent with our results and previous reports: bardoxolone methyl only induced cholestasis in rats and no changes in the BSEP expression were noted in human hepatocytes, OA induced cholestasis in mice (Lu et al., 2013), and no cholestatic findings were observed in association with the clinical use of bardoxolone methyl or OA (Lin et al., 2016; Lewis et al., 2021). Collectively, triterpenoid compounds may cause rodent-specific cholestasis by reducing the expression of BSEP, and the clinical extrapolability to humans of triterpenoid compound-induced cholestasis in rodents is low.

This study has revealed important results regarding the safety of bardoxolone methyl. Generally, non-clinical toxicity studies using at least two animal species, rodents and non-rodents, are required for drug development. A poor understanding of the drug’s toxicity, including species differences and their mechanisms, makes it difficult to evaluate whether the toxicity observed in these studies is likely to occur in humans. Therefore, it is important to elucidate the mechanisms of toxicity and to understand the toxicity profile of drug candidates for properly assessing toxicity risks in humans. The present results have revealed that bardoxolone methyl showed rodent-specific hepatobiliary toxicity, which supports the safety of bardoxolone methyl in humans.

In conclusion, the long-term administration of bardoxolone methyl induced rodent-specific cholestasis and cholestasis-induced cholangioma. Based on the results using hepatocytes, the cholestatic changes observed in rats were probably caused by the decreased expression of BSEP in the liver, and they are unlikely to occur in non-rodent species such as monkeys and humans. These results strongly suggest that the hepatobiliary toxicity profiles of bardoxolone methyl are different among animal species, especially between rodents and non-rodents. Therefore, we believe that the risk of cholestasis and cholestasis-derived carcinogenicity associated with bardoxolone methyl is quite low in humans.

ACKNOWLEDGMENTS

The authors deeply appreciate Ayako Miwa, Teppei Shioji, and Dr. Takeshi Uemori for their technical assistance, and Dr. Yui Suzuki for her review on histopathological examination.

Conflict of interest

H.O., T. A., A. T., K. I., M. H. and K. N. are employees of Kyowa Kirin, Co., Ltd. Kyowa Kirin and Reata Pharmaceuticals were clinically developing bardoxolone methyl.

REFERENCES

•  Björnsson,  E. and  Olsson,  R. (2005): Outcome and prognostic markers in severe drug-induced liver disease. Hepatology, 42, 481-489.
•  Chin,  M.P.,  Bakris,  G.L.,  Block,  G.A.,  Chertow,  G.M.,  Goldsberry,  A.,  Inker,  L.A.,  Heerspink,  H.J.,  O’Grady,  M.,  Pergola,  P.E.,  Wanner,  C.,  Warnock,  D.G. and  Meyer,  C.J. (2018): Bardoxolone Methyl Improves Kidney Function in Patients with Chronic Kidney Disease Stage 4 and Type 2 Diabetes: Post-Hoc Analyses from Bardoxolone Methyl Evaluation in Patients with Chronic Kidney Disease and Type 2 Diabetes Study. Am. J. Nephrol., 47, 40-47.
•  Dinkova-Kostova,  A.T.,  Liby,  K.T.,  Stephenson,  K.K.,  Holtzclaw,  W.D.,  Gao,  X.,  Suh,  N.,  Williams,  C.,  Risingsong,  R.,  Honda,  T.,  Gribble,  G.W.,  Sporn,  M.B. and  Talalay,  P. (2005): Extremely potent triterpenoid inducers of the phase 2 response: correlations of protection against oxidant and inflammatory stress. Proc. Natl. Acad. Sci. USA, 102, 4584-4589.
•  Gomez-Ospina,  N.,  Potter,  C.J.,  Xiao,  R.,  Manickam,  K.,  Kim,  M.S.,  Kim,  K.H.,  Shneider,  B.L.,  Picarsic,  J.L.,  Jacobson,  T.A.,  Zhang,  J.,  He,  W.,  Liu,  P.,  Knisely,  A.S.,  Finegold,  M.J.,  Muzny,  D.M.,  Boerwinkle,  E.,  Lupski,  J.R.,  Plon,  S.E.,  Gibbs,  R.A.,  Eng,  C.M.,  Yang,  Y.,  Washington,  G.C.,  Porteus,  M.H.,  Berquist,  W.E.,  Kambham,  N.,  Singh,  R.J.,  Xia,  F.,  Enns,  G.M. and  Moore,  D.D. (2016): Mutations in the nuclear bile acid receptor FXR cause progressive familial intrahepatic cholestasis. Nat. Commun., 7, 10713. Nature Publishing Group.
•  Heathcote,  E.J. (2007): Diagnosis and management of cholestatic liver disease. Clin. Gastroenterol. Hepatol., 5, 776-782.
•  Itoh,  K.,  Wakabayashi,  N.,  Katoh,  Y.,  Ishii,  T.,  Igarashi,  K.,  Engel,  J.D. and  Yamamoto,  M. (1999): Keap1 represses nuclear activation of antioxidant responsive elements by Nrf2 through binding to the amino-terminal Neh2 domain. Genes Dev., 13, 76-86.
•  Jiang,  L.,  Zhang,  H.,  Xiao,  D.,  Wei,  H. and  Chen,  Y. (2021): Farnesoid X receptor (FXR): structures and ligands. Comput. Struct. Biotechnol. J., 19, 2148-2159. The Author(s)
•  Kadıoğlu,  E.,  Tekşen,  Y.,  Koçak,  C. and  Koçak,  F.E. (2021): Beneficial effects of bardoxolone methyl, an Nrf2 activator, on crush-related acute kidney injury in rats. Eur. J. Trauma Emerg. Surg., 47, 241-250. Springer Berlin Heidelberg.
•  Keitel,  V.,  Burdelski,  M.,  Warskulat,  U.,  Kühlkamp,  T.,  Keppler,  D.,  Häussinger,  D. and  Kubitz,  R. (2005): Expression and localization of hepatobiliary transport proteins in progressive familial intrahepatic cholestasis. Hepatology, 41, 1160-1172.
•  Knisely,  A.S.,  Strautnieks,  S.S.,  Meier,  Y.,  Stieger,  B.,  Byrne,  J.A.,  Portmann,  B.C.,  Bull,  L.N.,  Pawlikowska,  L.,  Bilezikçi,  B.,  Ozçay,  F.,  László,  A.,  Tiszlavicz,  L.,  Moore,  L.,  Raftos,  J.,  Arnell,  H.,  Fischler,  B.,  Németh,  A.,  Papadogiannakis,  N.,  Cielecka-Kuszyk,  J.,  Jankowska,  I.,  Pawłowska,  J.,  Melín-Aldana,  H.,  Emerick,  K.M.,  Whitington,  P.F.,  Mieli-Vergani,  G. and  Thompson,  R.J. (2006): Hepatocellular carcinoma in ten children under five years of age with bile salt export pump deficiency. Hepatology, 44, 478-486.
•  Kong,  B.,  Zhu,  Y.,  Li,  G.,  Williams,  J.A.,  Buckley,  K.,  Tawfik,  O.,  Luyendyk,  J.P. and  Guo,  G.L. (2016): Mice with hepatocyte-specific FXR deficiency are resistant to spontaneous but susceptible to cholic acid-induced hepatocarcinogenesis. Am. J. Physiol. Gastrointest. Liver Physiol., 310, G295-G302.
• Lesage, G., Glaser, S., Regulation, A.G., Lesage, G., Glaser, S. and Alpini, G. (2001): Regulation of cholangiocyte proliferation. 73-80.
•  Lewis,  J.H. (2000): DRUG-INDUCED LIVER DISEASE. Med. Clin. North Am., 84, 1275-1311.
•  Lewis,  J.H.,  Jadoul,  M.,  Block,  G.A.,  Chin,  M.P.,  Ferguson,  D.A.,  Goldsberry,  A.,  Meyer,  C.J.,  O’Grady,  M.,  Pergola,  P.E.,  Reisman,  S.A.,  Wigley,  W.C. and  Chertow,  G.M. (2021): Effects of Bardoxolone Methyl on Hepatic Enzymes in Patients with Type 2 Diabetes Mellitus and Stage 4 CKD. Clin. Transl. Sci., 14, 299-309.
•  Lin,  C.,  Wen,  X. and  Sun,  H. (2016): Oleanolic acid derivatives for pharmaceutical use: a patent review. Expert Opin. Ther. Pat., 26, 643-655. Taylor & Francis.
•  Lipton,  S. and  Satoh,  T. (2017): Recent advances in understanding NRF2 as a druggable target: development of pro-electrophilic and non-covalent NRF2 activators to overcome systemic side effects of electrophilic drugs like dimethyl fumarate. F1000 Res., 6, 1-10.
•  Liu,  W. and  Wong,  C. (2010): Oleanolic acid is a selective farnesoid X receptor modulator. Phytother. Res., 24, 369-373.
•  Lu,  Y.F.,  Wan,  X.L.,  Xu,  Y. and  Liu,  J. (2013): Repeated oral administration of oleanolic acid produces cholestatic liver injury in mice. Molecules, 18, 3060-3071.
•  Nangaku,  M.,  Kanda,  H.,  Takama,  H.,  Ichikawa,  T.,  Hase,  H. and  Akizawa,  T. (2020): Randomized Clinical Trial on the Effect of Bardoxolone Methyl on GFR in Diabetic Kidney Disease Patients (TSUBAKI Study). Kidney Int. Rep., 5, 879-890. Elsevier Inc.
•  Okada,  K.,  Shoda,  J.,  Taguchi,  K.,  Maher,  J.M.,  Ishizaki,  K.,  Inoue,  Y.,  Ohtsuki,  M.,  Goto,  N.,  Takeda,  K.,  Utsunomiya,  H.,  Oda,  K.,  Warabi,  E.,  Ishii,  T.,  Osaka,  K.,  Hyodo,  I. and  Yamamoto,  M. (2008): Ursodeoxycholic acid stimulates Nrf2-mediated hepatocellular transport, detoxification, and antioxidative stress systems in mice. Am. J. Physiol. Gastrointest. Liver Physiol., 295, G735-G747.
•  Okada,  K.,  Shoda,  J.,  Taguchi,  K.,  Maher,  J.M.,  Ishizaki,  K.,  Inoue,  Y.,  Ohtsuki,  M.,  Goto,  N.,  Sugimoto,  H.,  Utsunomiya,  H.,  Oda,  K.,  Warabi,  E.,  Ishii,  T. and  Yamamoto,  M. (2009): Nrf2 counteracts cholestatic liver injury via stimulation of hepatic defense systems. Biochem. Biophys. Res. Commun., 389, 431-436. Elsevier Inc.
• Okada, K., Shoda, J., Kano, M., Suzuki, S., Ohtake, N., Yamamoto, M., Takahashi, H., Utsunomiya, H., Oda, K., Sato, K., Watanabe, A., Ishii, T., Itoh, K., Yamamoto, M., Yokoi, T., Yoshizato, K., Sugiyama, Y. and Suzuki, H. (2022): Inchinkoto, a herbal medicine, and its ingredients dually exert Mrp2 / MRP2- mediated choleresis and Nrf2-mediated antioxidative action in rat livers. 1450–1463.
•  Padda,  M.S.,  Sanchez,  M.,  Akhtar,  A.J. and  Boyer,  J.L. (2011): Drug-induced cholestasis. Hepatology, 53, 1377-1387.
•  Pérez-Pineda,  S.I.,  Baylón-Pacheco,  L.,  Espíritu-Gordillo,  P.,  Tsutsumi,  V. and  Rosales-Encina,  J.L. (2021): Effect of bile acids on the expression of MRP3 and MRP4: an In vitro study in HepG2 cell line. Ann. Hepatol., 24, 100325.
•  Pergola,  P.E.,  Raskin,  P.,  Toto,  R.D.,  Meyer,  C.J.,  Huff,  J.W.,  Grossman,  E.B.,  Krauth,  M.,  Ruiz,  S.,  Audhya,  P.,  Christ-Schmidt,  H.,  Wittes,  J. and  Warnock,  D.G.; BEAM Study Investigators. (2011a): Bardoxolone methyl and kidney function in CKD with type 2 diabetes. N. Engl. J. Med., 365, 327-336.
•  Pergola,  P.E.,  Krauth,  M.,  Huff,  J.W.,  Ferguson,  D.A.,  Ruiz,  S.,  Meyer,  C.J. and  Warnock,  D.G. (2011b): Effect of bardoxolone methyl on kidney function in patients with T2D and Stage 3b-4 CKD. Am. J. Nephrol., 33, 469-476.
•  Razumilava,  N. and  Gores,  G.J. (2014): Cholangiocarcinoma. Lancet, 383, 2168-2179.
•  Reisman,  S.A.,  Csanaky,  I.L.,  Yeager,  R.L. and  Klaassen,  C.D. (2009): Nrf2 activation enhances biliary excretion of sulfobromophthalein by inducing glutathione-S-transferase activity. Toxicol. Sci., 109, 24-30.
•  Reisman,  S.A.,  Chertow,  G.M.,  Hebbar,  S.,  Vaziri,  N.D.,  Ward,  K.W. and  Meyer,  C.J. (2012): Bardoxolone methyl decreases megalin and activates nrf2 in the kidney. J. Am. Soc. Nephrol., 23, 1663-1673.
•  Rojas-Rivera,  J.,  Ortiz,  A. and  Egido,  J. (2012): Antioxidants in kidney diseases: the impact of bardoxolone methyl. Int. J. Nephrol., 2012, 321714.
•  Scheimann,  A.O.,  Strautnieks,  S.S.,  Knisely,  A.S.,  Byrne,  J.A.,  Thompson,  R.J. and  Finegold,  M.J. (2007): Mutations in bile salt export pump (ABCB11) in two children with progressive familial intrahepatic cholestasis and cholangiocarcinoma. J. Pediatr., 150, 556-559.
•  Shelton,  L.M.,  Lister,  A.,  Walsh,  J.,  Jenkins,  R.E.,  Wong,  M.H.,  Rowe,  C.,  Ricci,  E.,  Ressel,  L.,  Fang,  Y.,  Demougin,  P.,  Vukojevic,  V.,  O’Neill,  P.M.,  Goldring,  C.E.,  Kitteringham,  N.R.,  Park,  B.K.,  Odermatt,  A. and  Copple,  I.M. (2015): Integrated transcriptomic and proteomic analyses uncover regulatory roles of Nrf2 in the kidney. Kidney Int., 88, 1261-1273.
•  Sinal,  C.J.,  Tohkin,  M.,  Miyata,  M.,  Ward,  J.M.,  Lambert,  G. and  Gonzalez,  F.J. (2000): Targeted disruption of the nuclear receptor FXR/BAR impairs bile acid and lipid homeostasis. Cell, 102, 731-744.
•  Stieger,  B. (2010): Role of the bile salt export pump, BSEP, in acquired forms of cholestasis. Drug Metab. Rev., 42, 437-445.
•  Swift,  B.,  Pfeifer,  N.D. and  Brouwer,  K.L. (2010): Sandwich-cultured hepatocytes: an in vitro model to evaluate hepatobiliary transporter-based drug interactions and hepatotoxicity. Drug Metab. Rev., 42, 446-471.
•  Taguchi,  K.,  Maher,  J.M.,  Suzuki,  T.,  Kawatani,  Y.,  Motohashi,  H. and  Yamamoto,  M. (2010): Genetic analysis of cytoprotective functions supported by graded expression of Keap1. Mol. Cell. Biol., 30, 3016-3026.
•  Weerachayaphorn,  J.,  Cai,  S.Y.,  Soroka,  C.J. and  Boyer,  J.L. (2009): Nuclear factor erythroid 2-related factor 2 is a positive regulator of human bile salt export pump expression. Hepatology, 50, 1588-1596.
•  Yang,  K.,  Köck,  K.,  Sedykh,  A.,  Tropsha,  A. and  Brouwer,  K.L. (2013): An updated review on drug-induced cholestasis: mechanisms and investigation of physicochemical properties and pharmacokinetic parameters. J. Pharm. Sci., 102, 3037-3057.
•  Yates,  M.S.,  Tauchi,  M.,  Katsuoka,  F.,  Flanders,  K.C.,  Liby,  K.T.,  Honda,  T.,  Gribble,  G.W.,  Johnson,  D.A.,  Johnson,  J.A.,  Burton,  N.C.,  Guilarte,  T.R.,  Yamamoto,  M.,  Sporn,  M.B. and  Kensler,  T.W. (2007): Pharmacodynamic characterization of chemopreventive triterpenoids as exceptionally potent inducers of Nrf2-regulated genes. Mol. Cancer Ther., 6, 154-162.
•  Yin,  X.,  Zheng,  S.S.,  Zhang,  B.H.,  Zhou,  Y.,  Chen,  X.H.,  Ren,  Z.G.,  Qiu,  S.J. and  Fan,  J. (2013): Elevation of serum γ-glutamyltransferase as a predictor of aggressive tumor behaviors and unfavorable prognosis in patients with intrahepatic cholangiocarcinoma: analysis of a large monocenter study. Eur. J. Gastroenterol. Hepatol., 25, 1408-1414.
•  Zamek-Gliszczynski,  M.J.,  Xiong,  H.,  Patel,  N.J.,  Turncliff,  R.Z.,  Pollack,  G.M. and  Brouwer,  K.L. (2003): Pharmacokinetics of 5 (and 6)-carboxy-2′,7′-dichlorofluorescein and its diacetate promoiety in the liver. J. Pharmacol. Exp. Ther., 304, 801-809.
•  Zollner,  G.,  Wagner,  M.,  Moustafa,  T.,  Fickert,  P.,  Silbert,  D.,  Gumhold,  J.,  Fuchsbichler,  A.,  Halilbasic,  E.,  Denk,  H.,  Marschall,  H.U. and  Trauner,  M. (2006): Coordinated induction of bile acid detoxification and alternative elimination in mice: role of FXR-regulated organic solute transporter-α/β in the adaptive response to bile acids. Am. J. Physiol. Gastrointest. Liver Physiol., 290, G923-G932.