Antidyslipidemic potential of a novel farnesoid X receptor antagonist in a hamster model of dyslipidemia: Comparative studies of other nonstatin agents

Abstract We attempted to clarify the therapeutic capability of antagonists of the farnesoid X receptor (FXR), a nuclear receptor that regulates lipid and bile acid metabolism. Herein, we report the antidyslipidemic effects of a novel synthesized FXR antagonist, compound‐T1, utilizing a dyslipidemic hamster model. Compound‐T1 selectively inhibited chenodeoxycholic acid‐induced FXR activation (IC 50, 2.1 nmol·L−1). A hamster model of diet‐induced hyperlipidemia was prepared to investigate the antidyslipidemic effects of compound‐T1 through comparative studies of the nonstatin lipid‐modulating agents ezetimibe, cholestyramine, and torcetrapib. In the hamster model, compound‐T1 (6 mg·kg−1·day−1, p.o.) increased the level of plasma high‐density lipoprotein (HDL)‐cholesterol (+22.2%) and decreased the levels of plasma non‐HDL‐cholesterol (−43.6%) and triglycerides (−31.1%). Compound‐T1 also increased hepatic cholesterol 7α‐hydroxylase expression and fecal bile acid excretion, and decreased hepatic cholesterol content. Moreover, the hamster model could reflect clinical results of other nonstatin agents. Torcetrapib especially increased large HDL particles compared with compound‐T1. Additionally, in the human hepatoma Huh‐7 cells, compound‐T1 enhanced apolipoprotein A‐I secretion at a concentration close to its IC 50 value for FXR. Our results indicated the usefulness of the hamster model in evaluating FXR antagonists and nonstatin agents. Notably, compound‐T1 exhibited beneficial effects on both blood non‐HDL‐cholesterol and HDL‐cholesterol, which are thought to involve enhancement of cholesterol catabolism and apolipoprotein A‐I production. These findings aid the understanding of the antidyslipidemic potential of FXR antagonists with a unique lipid and bile acid modulation.


| INTRODUCTION
High blood cholesterol is a risk factor for cardiovascular events. 1,2 3-Hydroxymethyl 3-glutaryl-CoA reductase inhibitors, also known as statins, markedly reduce both blood cholesterol and cardiovascular risk in hyperlipidemia patients. [3][4][5] However, for patients who cannot achieve their cholesterol-lowering targets with statins, owing to insufficient efficacy or adverse effects, nonstatin agents are also effective as an alternate or add-on therapy to statins. 6 One such nonstatin agent is cholestyramine, a bile acid sequestrant, which forms a complex with bile acids in the intestine and is then eliminated into feces. Through the augmentation of bile acid synthesis and cholesterol catabolism, cholestyramine demonstrated a significant reduction in blood cholesterol. 7,8 A recent meta-analysis and large-scale epidemiologic analysis suggested that the cholesterol-lowering effects of bile acid sequestrants, including cholestyramine, may contribute to the prevention of cardiovascular diseases. 9 Another nonstatin medication is the cholesterol absorption inhibitor ezetimibe, which targets intestinal uptake of dietary and biliary cholesterols; good efficacy and tolerability of combination therapy with statin and ezetimibe have been reported. 10 Another approach aimed at the reduction in cardiovascular risk is the increase in blood high-density lipoprotein (HDL)-cholesterol levels 11 by the inhibition of cholesteryl ester-transfer protein (CETP), which mediates transfer of cholesteryl ester from HDLs to atherogenic lipoproteins. Clinical trials involving torcetrapib, the first CETP inhibitor, were terminated early owing to off-target toxicity, 12,13 and despite remarkable increases in HDL-cholesterol, torcetrapib did not reduce the progression of atherosclerosis. 14,15 The definitive clinical efficacies of CETP inhibition are under investigation in the ongoing REALIZE study of anacetrapib, a more recent CETP inhibitor. 16 As a possible concept for another novel nonstatin agent, we focused on the inhibition of farnesoid X receptor (FXR), a member of the nuclear receptor superfamily. FXR is expressed at high levels in tissues involved in bile acid metabolism such as the liver, intestine, and kidney. 17,18 In the liver, FXR indirectly suppressed expression of cholesterol 7a-hydroxylase (CYP7A1), a rate-limiting enzyme in the cholesterol catabolic pathway, 19 and also directly downregulated the expression of apolipoprotein A-I, a major protein constituent of HDL. 20 We therefore hypothesized that repression of FXR would improve dyslipidemia via both non-HDL-cholesterol reduction and HDL-cholesterol elevation, and we successfully confirmed that a synthesized FXR antagonist, compound-T3, displayed these blood changes in cynomolgus monkeys. 21,22 In this study, to clarify the therapeutic advantage of an FXR antagonist over the nonstatin agents ezetimibe, cholestyramine, and torcetrapib, we investigated  Figure 1) and conducted comparative studies in a high-fat diet-induced dyslipidemic hamster model.

| Nuclear receptor panel assays
The agonistic and antagonistic activities of human nuclear receptors, FXR and PPARs PPARa, PPARd, and PPARc were measured as previously described. 23,24 The agonistic and antagonistic activities for the human retinoid X receptor-a (RXRa) and the human liver X receptors (LXRs) LXRa and LXRb were also measured. Briefly, the reporter construct for the RXRa reporter gene assay, pGL3-DR1 9 4-tk-luc, was created by the insertion of four copies of direct repeat 1 element (AGGTCA-N-AGGTCA) upstream of the herpes virus thymidine kinase promoter and the luciferase reporter gene. To construct the LXRa and LXRb reporter gene assays, pGL3-DR4 9 4-tk-luc was generated by the insertion of four copies of direct repeat 4 (AGGTCA-NNNN-AGGTCA) at the same position as pGL3-DR1 9 4tk-luc.

| Multiple-dose study
Hamsters were fed with a high-fat diet (a chow diet containing 10.3% nonsalt butter, Oriental Yeast Co., Ltd) and housed individually in cages after the age of 6 weeks. After high-fat diet feeding for 1 week, drug treatments were commenced at the age of 7 weeks.
Food consumption was determined for each animal every day. Compound-T1, ezetimibe, and torcetrapib were suspended in 0.5% immediately into storage at À30°C for lipid level measurements and at À80°C for gene expression measurements.

| Measurement of plasma C4 levels
Plasma C4 levels were measured following the method described in our previous report. 23 Briefly, 25 lL plasma was mixed vigorously with 10 lL of 7b-hydroxy-4-cholesten-3-one (50 lgÁmL À1 in dimethylsulfoxide; Calbiochem, USA), as an internal standard, and 500 lL acetonitrile, and then the mixture was sonicated for 5 minutes. After centrifugation, 100 lL supernatant was processed using a HPLC system (ClassVP; Shimadzu, Japan) fitted with a Nova-Pak C18 steel column (Waters Corporation, USA), and analyzed with 95% acetonitrile (flow rate: 1 mLÁmin À1 ) as the mobile phase. The absorbance of C4 was detected at 241 nm; absorbance areas of C4 and the internal standard were calculated, and the concentrations of C4 were quantified against an internal standard.

| Measurement of hepatic lipid levels
The frozen liver samples (approximately 1.0 g) were homogenized by the addition of 3.35% sodium sulfate solution (9 mL) and shaken for 10 minutes after the addition of a 3:2 solution of hexane-isopropanol (3:2) solution. Subsequently, 100 lL of the solution was evaporated to dryness and solidified at 50°C in a nitrogen atmosphere to form a dried residue. The residue was dissolved in a 100 lL of diox-   Table S1). Expectedly, these three agents  Table S2). Torcetrapib significantly elevated HDLcholesterol level but did not lower non-HDL-cholesterol and TG levels, which was in contrast to compound-T1. Although no changes in body weight were observed in all groups, a significant reduction in food intake was observed in the 10 mgÁkg À1 Áday À1 of compound-T1-treated group (À22.5%, data not shown).
To perform a detailed comparison between the HDL-cholesterolelevating effects of compound-T1 and torcetrapib, we conducted lipoprotein profiling by HPLC gel filtration system ( Figure 5). The selectivity of compound-T1 for human nuclear receptors was measured as described in the Methods section.  change in the peak value of particle size. At doses of 3 mgÁkg À1 Áday À1 and 10 mgÁkg À1 Áday À1 , compound-T1 also increased HDL levels with no change in the peak value of particle size, but an increase larger HDL particles (fraction number 16) was observed, especially at the higher treatment concentration. In contrast, torcetrapib produced a more marked increase in HDLs with a larger particle diameter; in particular, a dose of 100 mgÁkg À1 Áday À1 of torcetrapib resulted in an increase in much larger particles (fraction number [13][14][15], which were hardly observed in the control and compound-T1-treated groups. Torcetrapib did not produce any changes in the VLDL and LDL fractions.

| Comparative studies on fecal and hepatic lipid profiles in a dyslipidemic hamster model
The changes in fecal lipid content after treatment with compound-T1, ezetimibe, and cholestyramine are summarized in Figure 6A.
Both compound-T1 and cholestyramine significantly increased fecal total bile acid content, while ezetimibe increased both fecal total cholesterol and TG content. No other significant changes were observed in all groups.
Next, the changes in fecal lipid content were compared between compound-T1 and torcetrapib, and are summarized in Figure 6B.
Compound-T1 dose-dependently increased fecal total bile acid content, while torcetrapib resulted in a significant decrease. A significant reduction in fecal total cholesterol was observed in both groups receiving the highest dosage of compound-T1 (10 mgÁkg À1 Áday À1 ) and torcetrapib (100 mgÁkg À1 Áday À1 ). There was no significant change in fecal TG in all groups.
The effects of compound-T1, ezetimibe, and cholestyramine on hepatic lipid content are summarized in Figure 7. Compound-T1 dose-dependently lowered hepatic total cholesterol and cholesteryl ester. Compound-T1 also decreased hepatic free cholesterol at a dose of 10 mgÁkg À1 Áday À1 . Torcetrapib did not show any significant change in hepatic total cholesterol, cholesteryl ester, and free cholesterol content. In addition, hepatic TG content was decreased in compound-T1-treated group, but increased in the 100 mgÁkg À1 Áday À1 torcetrapib-treatment group.
To confirm the FXR antagonistic activities of compound-T1 after multiple doses, we measured the hepatic expression levels of CYP7A1. Hepatic CYP7A1 mRNA levels were dose-dependently increased in the compound-T1-treated groups and in the cholestyramine-treated group. There were no significant changes in the ezetimibe-treated groups (Figure 8).

| Effects of compound-T1 on apolipoprotein A-I production in Huh-7 cells
To elucidate the mechanism of the HDL-cholesterol-elevating effects of compound-T1, we investigated its effect on apolipoprotein A-I production using the human hepatoma Huh-7 cell line under physiological CDCA conditions. Apolipoprotein A-I secretion was significantly reduced by the addition of 30 lmolÁL À1 CDCA, and this reduction was dose-dependently rescued by compound-T1 with an ED 50 value of 1.8 nmolÁL À1 (Figure 9).

| DISCUSSION
We have previously confirmed that a synthesized FXR antagonist, compound-T3, reduced plasma non-HDL-cholesterol levels and also elevated HDL-cholesterol levels in a primate model. 21,22 However, the development of a rodent model with similar disease conditions to human hyperlipidemia would be still useful for the further comprehension of this new type of agent. However, the rodent animals such as mice, rats, and guinea pigs have different plasma lipid and lipoprotein profiles from humans. 23,25 Actually, we found that another FXR antagonist, compound-T0, exacerbated dyslipidemia in mice due to enhancement of intestinal lipid absorption via acceleration of bile acid excretion. 26 We think that the difference of the effects of FXR antagonist between primate and mouse models would be due to the difference in regulation of key players relating to lipid and bile acids metabolisms, in particular, the hepatic LDL receptor. 27 In addition, CETP is also a key protein involved in plasma cholesterol transport that transfers cholesteryl ester from HDL to LDL and VLDL. 28 Golden Syrian hamsters are known to develop human-like hyperlipidemia following the feeding of a high-fat diet, [29][30][31] and also show both plasma CETP activity 28 and regulation of hepatic LDL receptor. 32 Therefore, we considered that a hyperlipidemic hamster model would be suitable for the purpose described above. To determine the appropriate diet for hamsters, we referenced diet information in nonclinical and clinical studies of orlistat, an antiobesity drug. 33,34 After the hamsters were fed this butter-rich diet for a week, both plasma levels of cholesterol and triglyceride increased to approximately two and three times the normal level respectively F I G U R E 5 Effects of compound-T1 and torcetrapib on plasma lipoprotein distribution in high-fat diet-fed hamsters. Pooled plasma (n = 6 per group) was collected after repeated drug administrations and then was fractionated by gel filtration HPLC. The measurement procedures are described in the Methods section SHINOZAWA ET AL.  (Table S3). Using this hamster model, herein, we sought to clarify the favorable pharmacological profile of a novel FXR antagonist compound-T1, a potential antidyslipidemic agent, through comparative studies with other nonstatin agents ezetimibe, cholestyramine, and torcetrapib.
The first question in our study was to determine whether the clinical results of other nonstatin agents were consistent in our hamster model. Firstly, ezetimibe, a cholesterol absorption inhibitor, exhibited clinically relevant changes with regard to the plasma non-HDL-cholesterol reduction and fecal cholesterol excretion. 35,36 Although a significant increase in fecal TG was also found in ezetimibe-treated hamsters, as reported in a previous study using an obese rodent model, 37 the dramatic effect of ezetimibe on TG excretion has not been reported in a clinical setting. However, an inhibitory effect on the intestinal production of chylomicrons, which consist largely of TG, was reported in patients with hyperlipidemia. 38 Therefore, we assumed that the increased fecal TG in ezetimibe-treated hamsters might result from its intervention in chylomicron TG F I G U R E 9 Effect of compound-T1 on apolipoprotein A-I secretion in CDCA-treated Huh-7 cells. Apolipoprotein A-I concentration in cell supernatant was measured with a sandwich ELISA method, as described in the Methods section. Each value represents the mean AE SEM (n = 3). Statistical analysis was carried out using Student's t-test (**P ≤ .01 vs control) or one-tailed Williams' test ( † P ≤ .025 vs CDCA(+)) observations as described above. A high dosage requirement can sometimes affect patient drug compliance. 46 In this respect, compound-T1 would be therefore superior to cholestyramine. Additionally, in the present study, compound-T1 did not result in a large change in HDL particle size, but torcetrapib did. Future studies are needed to clarify this, but it would be important to investigate whether increased HDL particles can promote the reverse cholesterol transport system to cause cholesterol efflux from atherosclerotic lesions to the liver and feces; in particular, we recommend a focus on apolipoprotein A-I, which has been reported to be a key factor in the effective promotion of beneficial HDL metabolism. 47 Indeed, intravenous infusion of a variant of apolipoprotein A-I-phospholipid complexes resulted in the regression of atherosclerosis in patients with acute coronary syndromes. 48

DISCLOSURE
The authors are employees of Takeda Pharmaceutical Co., Ltd (Osaka, Japan) at which Compound-T1 was synthesized.