Anti-Atherosclerotic Effects of A Novel Synthetic Tissue-Selective Steroidal LXR Agonist in LDLR−/− Mice

Abstract

Liver X receptor (LXR) agonists have the potential to treat atherosclerosis based on their ability to enhance reverse cholesterol transport. However, their side effects, such as induction of liver lipogenesis and triglyceridemia, may limit their pharmaceutical development. In contrast to the non-steroidal LXR agonist T0901317, 3α, 6α, 24-trihydroxy-24, 24-di(trifluoromethyl)-5β-cholane (ATI-829), a novel potent synthetic steroidal LXR agonist, was a poor inducer of SREBP-1c expression in hepatoma HepG2 cells, while both compounds increased ABCA1 expression in macrophage THP-1 cells. In male LDLR−/− mice, ATI-829 selectively activated LXR target gene expression in mouse intestine and macrophages, but not in liver. A significant increase in liver triglyceride and plasma triglyceride-rich small VLDL was observed in T0901317 but not ATI-829-treated mice. Compared to vehicle-treated mice, atherosclerosis development was significantly inhibited in the innominate artery after treatment with either compound. However, in the aortic root, inhibition of atherosclerosis was only observed in the right (RC) but not left coronary related sinus (LC) of mice treated with either compound. Lesions in the innominate artery were less complex after treatment with either compound and contained mostly macrophage foam cells. In contrast, LC lesions were more complex and had a large collagen-positive fibrous cap and less macrophage foam cell area after treatment with either compound. The T0901317-induced hypertriglyceridemia was accompanied by an increase in small triglyceride-rich VLDL that may influence LXR agonist-mediated anti-atherosclerotic effects at certain vascular sites. ATI-829 by selectively activating LXR in certain tissues without inducing hypertriglyceridemia is a good candidate for drug development.

Introduction

Dyslipidemia is a key risk factor for cardiovascular disease. The statins, a potent class of plasma LDL cholesterol lowering drugs, significantly reduce cardiovascular mortality in hypercholesterolemic patients. However, statins do not completely prevent the progression of atherosclerosis in many susceptible individuals. Indeed, even among individuals with the same cholesterol levels there is great disparity in the expression of clinical vascular disease and plasma total cholesterol levels are within recommended levels in almost half of all coronary heart disease patients (The Bezafibrate Infarction Prevention Study Group, 1992). Statins block cholesterol and lipid biosynthesis and reduce plasma apoB-containing lipoproteins, which in turn decreases the influx of cholesterol into plaques. However, statins do not directly enhance cholesterol removal from arterial walls, which is also critical for prevention and treatment of atherosclerosis. Thus, the addition of strategies for enhancing reverse cholesterol transport may be useful for prevention or treatment of atherosclerosis especially in those patients with normal plasma cholesterol levels or patients that do not respond adequately to statin treatment.

Liver X receptors (LXR) are members of the nuclear receptor superfamily and were discovered in our laboratory (Song et al., 1994) and by others (Peet et al., 1998). They have recently been identified as oxysterol receptors that among other properties mediate cholesterol efflux from foam cells to nascent and mature HDL by activating target genes, such as ATP-binding cassette transporter A1 and G1 (ABCA1, ABCG1) and apolipoprotein E (apoE) (Janowski et al., 1996; Repa et al., 2000; Wang et al., 2006; Yancey et al., 2007). The discovery of LXR regulation of reverse cholesterol transport has opened a door to the possible development of LXR agonists as drugs for treatment of atherosclerosis. However, LXR activation in liver may lead to the accumulation of triglycerides due to up-regulation of sterol regulatory element binding protein 1c (SREBP-1c), a sterol-responsive transcription factor, which induces some of the key enzymes involved in fatty acid synthesis, such as fatty acid synthase (FAS) and acetyl-coenzyme A carboxylase (ACC) (Schultz et al., 2000; Grefhorst et al., 2002). Therefore, activation of LXR in liver induces liver de novo lipogenesis and elevation of plasma triglyceride levels. This significant side effect from LXR activation presents a challenge for pharmaceutical development of ideal LXR agonists.

A few possible strategies have been proposed to minimize the side effects of LXR agonist treatment. LXRβ-selective agonists may not induce hepatic lipogenesis (Lund et al., 2006; Quinet et al., 2006). However, the crystal structures of LXR alpha and beta are very similar in the ligand-binding domains and it may be a challenge to develop subtype-selective ligands (Williams et al., 2003; Svensson et al., 2003). Another approach is the identification of target gene-selective or tissue-selective LXR agonists. We previously observed that cholestenoic acid was an LXR agonist that had hypolipidemic effects in hypercholesterolemic rats, mice and hamsters and did not increase plasma triglyceride levels (Song and Liao, 2001). In line with these latter observations, several natural or synthetic steroidal LXR agonists were recently reported to selectively activate gene expression with minimal SREBP-1c gene activation in liver and strong induction of ABCA1 expression in macrophages (Adams et al., 2004; Quinet et al., 2004; Beyea et al., 2007). However, the use of steroidal LXR agonists as anti-atherosclerotic agents has not been examined perhaps because of their relatively low potency and low in vivo bioavailability. In this work, a new potent steroidal LXR agonist, ATI-829, was designed and synthesized. The agonist was formulated in a microemulsion for effective in vivo delivery. Activation of LXR target gene expression was characterized in both cultured human cell lines and different mouse organs. The impact of LXR agonist on plasma lipid levels and lipoprotein profile and on the development and progression of atherosclerosis in two artery regions were examined in male LDL receptor deficient (LDLR−/−) mice.

Methods

Chemicals

T0901317 was purchased from Cayman Chemical Company (Ann Arbor, MI). 22(R)-HC, 24(S)-HC, and 24(S), 25-EC were purchased from Steraloids Inc (Newport, RI). ATI-829 was synthesized from hyodeoxycholic acid and structure and purity (>99%) were confirmed by ¹H , ¹³C , ¹⁹F nuclear magnetic resonance spectroscopy and high resolution mass spectrometry. Details on compounds are available in the supplemental information (Supplemental Fig.1 and 2).

Cell Culture

Human embryonic kidney (HEK) 293, hepatoma HepG2, and macrophage THP-1 cell lines were obtained from the American Type Culture Collection (Manassas, VA). HEK293 and HepG2 cells were cultured in Dulbecco's modified Eagle's media supplemented with 10% FBS, 2 mM L-glutamine, and 100 U/ml penicillin and 100 μg/ml streptomycin (Invitrogen Life Technologies, Carlsbad, CA). THP-1 cells were maintained in Roswell Park Memorial Institute 1640 media containing 10% fetal bovine sera, 2mM L-glutamine, and 55 μM β-mercaptoethanol (Invitrogen Life Technologies). Confluent THP-1 cells were treated with 50−100 ng/ml phorbol 12, 13-dibutyrate (Sigma-Aldrich, St. Louis, MO) for 3 days to induce differentiation into adherent macrophages. HEK293 cells were plated on 48-well plates 24 h before ligand or vehicle addition. LXR ligands were dissolved in dimethylsulfoxide and added to cells for 24 to 48 h. The activity of LXR ligands was analyzed by a trans-activation assay as described previously (Song and Liao, 2001).

Animals and Diets

All work with mice followed National Institutes of Health guidelines for care and use of animals in experimentation. Animal work was reviewed and approved by the University of Chicago Institutional Animal Care and Use Committee. Male LDLR−/− mice on a C57BL/6 background were obtained from The Jackson Laboratory and fed a Western diet (TD88137, Harlan TEKLAD, Madison, WI) for 12 weeks starting at the age of 8 weeks for atherosclerosis studies. Mice were administered the LXR agonists T0901317 at a dose of 2 mg/kg, or ATI-829 at doses of 5 and 10 mg/kg by gavage daily in a 20% microemulsion (Gao et al., 1998) starting with initiation of the Western diet. The control group (vehicle) received microemlusion only. At 20 weeks of age, the mice were fasted for 4 hours, anesthetized with ketamine (100 mg/kg) and xylazine (10 mg/kg), and a blood sample removed from the retro-orbital plexus. The mice were then perfused transcardially with PBS followed by paraformaldehyde and the heart and upper vasculature removed and prepared for histology as described (Reardon et al., 2001).

For gene expression studies, fifteen week old male LDLR−/− mice were fed a Western diet and treated with or without LXR agonist as described above for 2 weeks to analyze LXR target gene expression in tissues and lipid levels in liver. Resident peritoneal macrophages were obtained by flushing the peritoneal cavity with ice-cold PBS and pelleted by centrifugation at 1500 rpm for 10 min. Cell pellets were resuspended in RNAlater (Ambion, Austin, TX) for RNA isolation. Intestine and liver samples were also stored in RNAlater for RNA isolation. Liver samples for lipid analysis were frozen at −70°C. Blood was collected from the retro-orbital plexus for lipid and lipoprotein analysis.

Determination of Plasma Concentration of T0901317 and ATI-829

Male C57BL/6 mice were gavaged with different doses of T0901317 and ATI-829 for pharmacokinetic studies. After gavage of compound, blood from the retro-orbital plexus was collected at different time points and no more than three bleedings were performed on each mouse. The serum was frozen at −70°C until analysis. An equal volume of 5% ammonium hydroxide was added to thawed serum followed by 10 volumes of methyl t-butyl ether. After mixing, the sample was centrifuged at 3000 rpm for 10 minutes and the methyl t-butyl ether layer transferred into a glass tube and dried in a TurboVap (Caliper Life Sciences, Hopkinton, MA) at 35°C. Dried extract was reconstituted in acetonitrile: water (50:50, v/v) and analyzed by high performance liquid chromatography with electrospray tandem mass spectrometry detection (LC/MS/MS) (Sciex, API-3000) with a SB-C18 column (Zorbax, 5μm) to determine the concentration of the compound in the extracts. All chemicals were purchased from Sigma-Aldrich (St. Louis, MO).

Lipid and Lipoprotein Analysis

Two hundred μl of plasma was fractionated on tandem Superose 6 fast protein liquid chromatography (FPLC) columns (Reardon et al., 2001). Cholesterol and triglyceride in the even-numbered FPLC fractions and in plasma samples were measured by using commercial kits (Stanbio Laboratory, Boerne, TX). Liver samples were homogenized, and the lipids were extracted as described (Bligh and Dyer, 1959) and measured as described above.

Electron microscopy (EM) of Lipoproteins

VLDL particles from pooled FPLC fractions were placed on a carbon-coated EM grid and negatively stained with 1% uranyl acetate. Particles were examined with a FEI Tecnai F30 electron microscope. OpenLab software version 3.1.5 was used in the measurement of particle size.

RNA Isolation and Gene Expression Analysis

Total RNA from cultured cells, mouse macrophages, intestine and liver was isolated by using the RNeasy Mini System (QIAGEN, Valencia, CA). First strand cDNA was synthesized by utilizing SuperScript III System (Invitrogen Life Technologies, Carlsbad, CA). Sequences of gene-specific primers and probe are shown in supplemental information (Supplemental Sequences of Gene-Specific Primers and Probe Information) and were purchased from IDT (Coralville, IA). Quantitative real time PCR was performed by using “TaqMan Universal PCR Master Mix” from Applied Biosystems. The results were normalized to 36B4 mRNA.

Histology and Lesion Analysis

Lesions in the innominate artery were quantitated from 4 digitally captured oil red O-stained 10-μm sections, each separated by 100 μm, and located between 150 and 450 μm distal to the branch point of the innominate artery from the aortic arch. Aortic sinus lesions were evaluated from 3 sections, each separated by 100 μm, beginning at the site of appearance of the coronary artery. Adjacent sections were stained with trichrome stain (Masson) (Sigma-Aldrich, St. Louis, MO), or immunostained for ABCA1 or macrophages using rabbit polyclonal antibody to ABCA1 (Novus Biologicals Inc, Littleton, CO) or MOMA-2 antibody (Serotec Inc, Raleigh, NC), respectively. OpenLab software version 3.1.5 was used in the quantification.

Statistical Analysis

Values are presented as means ± standard error of the mean (SEM) and were analyzed by using one-way ANOVA and Fisher's PLSD post hoc test or unpaired Student t-test. A statistically significant difference is set at p<0.05.

Results

In vitro characterization of ATI-829

Previously we have shown that an analog of hyodeoxycholic (3α, 6α-dihydroxy-5β-cholanic acid) is an agonist of LXR (Song and Liao, 2001). We modified the side chain of hyodeoxycholic acid by introducing the acidic bistrifluoromethyl carbinol present in T0901317 to create ATI-829 (Fig. 1A). ATI-829 is an LXR-specific agonist and failed to enhance the activity of various other known nuclear receptors (Fig. 1B), which were activated by their own agonists (Supplemental Fig.3). T0901317, ATI-829, 24(S), 25-epoxycholesterol, 24(S)-hydroxycholesterol, and 22(R)-hydroxycholesterol activated a DR4-luciferase reporter gene in HEK293 cells expressing LXRα or LXRβ. None of the compounds activated a DR4-luciferase reporter gene in HEK293 cells without co-transfection of LXRα or LXRβ expression vectors (Supplemental Fig.4). T0901317 and ATI-829 were the most potent agonists with LXRα, while the natural oxysterols were much less potent (Fig. 1C). With LXRβ, T0901317 was again the most potent agonist, but ATI-829 and the natural oxysterols were much weaker LXRβ agonists (Fig. 1D). Because several natural and synthetic LXR steroidal agonists selectively activate certain LXR target genes in different cell types (Beyea et al., 2007; Quinet et al., 2004), the selectivity of ATI-829 for activating gene expression was examined in human hepatoma HepG2 cells and macrophage THP-1 cells. In comparison with T0901317, ATI-829 only slightly induced SREBP-1c expression in HepG2 cells (Fig. 1E). However, a robust induction of ABCA1 expression in THP-1 cells was observed with both compounds (Fig. 1F). In addition, a 10-fold higher concentration of ATI-829 was required to activate the LXR target gene SREBP-1c in HepG2 cells compared to the concentration of ATI-829 needed to activate the LXR target gene ABCA1 in THP-1 cells (Fig. 1E and 1F). These data indicate that ATI-829 might have tissue-selectivity in activating LXR target genes.

Fig. 1.

Chemical structures of T0901317 and ATI-829 (A), nuclear receptor-specific transactivation by ATI-829 (B) and effect of T0901317, ATI-829, 22(R)-HC, 24(S)-HC, and 24(S), 25-EC on luciferase reporter gene expression in HEK293 cells expressing LXRα (C) or LXRβ (D) and co-transfected with a DR-4 luciferase reporter. The ability of ATI-829 to act as an agonist for other nuclear receptors was examined in HEK293 cells co-transfected with various nuclear receptors and their cognate luciferase reporter genes. Firefly luciferase expression was normalized using a co-transfected sea pansy luciferase expression vector and normalized firefly luciferase expression is presented relative to luciferase expression in transfections without added agonist. n = 9 in T0901317 and ATI-829 experiments and n = 6 in 22(R)-HC, 24(S)-HC, and 24(S), 25-EC experiments. Effect of T0901317 and ATI-829 on the induction of SREBP-1c mRNA expression in HepG2 cells (E) (n = 8) and ABCA1 expression in THP-1 cells (F) (n = 6). Specific mRNA levels were measured by quantitative RT-PCR, normalized with 36B4, and presented relative to vehicle control.

Pharmacokinetics of ATI-829

We formulated ATI-829 and T0901317 in a microemulsion for oral delivery (Gao et al., 1998) and compared their pharmacokinetics. The maximum concentration (Cmax) of T0901317 attained in mouse plasma was about 1400 ng/ml for a dose of 10 mg/kg and 210 ng/ml for a dose of 2 mg/kg (Fig. 2). The Cmax for ATI-829 in mouse plasma was approximately 200 ng/ml for a dose of 10 mg/kg (Fig. 2). The half-life of these compounds in blood was around 5 hours for T0901317 and approximately 7 hours for ATI-829 (Fig. 2).

Fig. 2.

Pharmacokinetics of LXR agonists. C57 BL/6 mice were orally administered with either 10 mg/kg ATI-829 (n=5), or 2 mg/kg (n=6) and 10 mg/kg T0901317 (n=3), and bled at different time points (from 10 minutes to 24 hours). Liquid chromatography with electrospray positive tandem mass spectrometry detection (LC/MS/MS) was used to determine the concentration of compounds in serum after extraction.

ATI-829 selectively activated LXR target genes in vivo

ATI-829 at a dose of 10 mg/kg/day and T0901317 at a dose of 2 mg/kg/day were orally administered for 2 weeks to male LDLR−/− mice fed a Western diet. Based on the pharmacokinetics, at these doses the two compounds should have a similar blood concentration and half-life in mice. LXR target gene expression in various tissues was determined by using quantitative real-time PCR. In mouse small intestine, both compounds induced ABCA1 expression more than 4 fold and significantly increased ABCG5 and ABCG8 mRNA up to 2 to 3 fold (Fig. 3A). Furthermore, a slight but significant increase in ABCA1 mRNA was found in peritoneal macrophages from mice treated with either ATI-829 or T0901317 (Fig. 3B). T0901317 treatment also significantly increased ABCG1 and apoE mRNA in peritoneal macrophages, while only a slight increase in ABCG1 and apoE mRNA was observed in peritoneal macrophages from ATI-829-treated mice (Fig. 3B). In T0901317-treated mice, hepatic expression of the lipid transporter ABCA1 was significantly induced more than 2.5 fold (Fig. 3C). The mRNA of genes involved in liver lipogenesis, such as SREBP-1c, ACC and FAS were also clearly elevated 2.5 to 3.5 fold after T0901317 treatment (Fig. 3C). In addition, a marked induction of genes involved in lipoprotein metabolism and clearance was observed in liver, which included a 2.5-fold and 2.7-fold increase of lipoprotein lipase and phospholipid transfer protein mRNA respectively (Fig. 3C). In contrast to these results with T0901317-treated mice, there was no significant change in LXR target gene expression in the livers of ATI-829-treated mice (Fig. 3C). Consistent with the effect of ATI-829 on LXR target genes in liver, hepatic triglyceride levels were significantly increased only in T0901317-treated mice (Fig. 3D).

Fig. 3.

Effect of LXR agonists on gene expression in mouse intestine (A), peritoneal macrophages (B), and liver (C) and mouse hepatic triglycerides (D). Male LDLR−/− mice on a Western diet were gavaged daily for 2 weeks with vehicle (control, n=8), 10 mg/kg ATI-829 (n=8), or 2 mg/kg T0901317 (n=7). Specific mRNA levels were measured by quantitative RT-PCR, normalized with 36B4, and presented relative to controls. Significant differences between control and treatment groups are indicated as * P<0.05.

Plasma lipid and lipoprotein analysis

Male LDLR−/− mice fed a Western diet were treated with vehicle, T0901317 or ATI-829 for two weeks. Analysis of plasma lipids and lipoproteins revealed a marked increase in total and VLDL cholesterol in T0901317-treated mice but not in ATI-289 treated mice (Fig. 4A and 4C). HDL cholesterol was slightly increased in ATI-829 treated mice (Fig. 4A and 4C). Only T0901317 dramatically increased plasma triglyceride levels, resulting in an increase in VLDL and LDL triglyceride (Fig. 4B and 4D). Treatment with T0901317 significantly raised the ratio of triglyceride to cholesterol in VLDL particles from 1.1 ± 0.11 (n=8) in the control group to 2.55 ± 0.16 (n=7, p<0.05) and in LDL particles from 0.22 ± 0.02 (n=8) in control group to 0.79 ± 0.06 (n=7, p<0.05). The triglyceride to cholesterol ratio in VLDL and LDL particles was similar in both ATI-829 and vehicle-treated mice. Most interestingly, the VLDL peak in T0901317-treated mice was broader extending into higher numbered FPLC fractions (Fig. 4A and 4B), indicating the possible accumulation of smaller-sized plasma VLDL particles in T0901317-treated mice but not in ATI-829-treated mice.

Fig. 4.

Effect of LXR agonist on lipoprotein profiles (A and B) and plasma lipid levels (C and D) in male LDLR−/− mice after 2 weeks of treatment with vehicle (control, n=8), 10 mg/kg/day of ATI-829 (n=8), or 2 mg/kg/day (n=7) of T0901317. Lipoprotein profiles were determined by FPLC. Significant differences between vehicle and treatment groups are indicated as * P<0.05.

To further examine the influence of ATI-829 and T0901317 on the size of plasma VLDL particles, pooled VLDL particles (FPLC fractions 7 to 14) were analyzed by electron microscopy (EM). Large VLDL particles (>100 nm in diameter) were observed in male LDLR−/− mice treated with T0901317 for one week (Fig. 5A) and in male LDLR−/− mice treated with ATI-829 or vehicle for up to 2 months (Fig. 5D to 5F). However, very few large VLDL particles (>100 nm in diameter) were observed after 2 weeks (Figure 5B) or 2 months (Figure 5C) of T0901317 treatment. The average diameter of the VLDL particles in the plasma of male LDLR−/− mice treated for 2 months with T0901317 was significantly reduced compared to vehicle-treated male LDLR−/− mice (Fig. 5G). ATI-829 treatment did not influence the average diameter of the VLDL particles. The percent of total VLDL particles with diameters less than 40 nm was dramatically increased (Fig. 5H), while the number of larger (> 50 nm in diameter) VLDL particles was significantly decreased in T0901317-treated male LDLR−/− mice (Figure 5H). In contrast, ATI-829 treatment only significantly altered the number of VLDL particles in the 50 to 60 nm and 40 to 50 nm ranges (Fig. 5H). Based on this data we conclude that ATI-829 differs from T0901317 in in vivo organ selectivity for activation of LXR target gene expression, which in turn causes a differential biological response in lipoprotein production and/or metabolism.

Fig. 5.

Effect of LXR agonist on plasma VLDL size. Male LDLR−/− mice were fed a Western diet and gavaged with 2 mg/kg/day of T0901317 for 1 week (A), 2 weeks (B), 2 months (C), or with 10 mg/kg/day of ATI-829 for 2 weeks (D), 2 months (E), or with vehicle for 2 months (F). FPLC fractions 7 to 14 were pooled for EM analysis. Size bar is 100 nm. The average size (G) and size distributions (H) of VLDL particles from mice treated with vehicle (n=3), 10 mg/kg/day of ATI-829 (n=3), or 2mg/kg/day of T0901317 (n=3) were determined after 2 months of treatment. Significant differences between vehicle and treatment groups are indicated as * P<0.05.

Anti-atherosclerotic effects of ATI-829

The impact of ATI-829 and T0901317 on the development of atherosclerosis was examined in male LDLR−/− mice fed a Western diet. After 12 weeks of treatment, there was no change in plasma total cholesterol levels in the vehicle and different treatment groups (Fig. 6A). However, T0901317 treatment significantly increased the cholesterol levels in VLDL particles and decreased the cholesterol levels in LDL and HDL particles (Fig. 6A). A significant increase in total, VLDL and LDL triglyceride was observed only in mice treated with T0901317 (Fig. 6B). In addition, hepatic triglyceride levels were significantly increased only in T0901317-treated but not ATI-829 treated mice (Fig. 6C). Therefore, based on these observations and the EM studies, T0901317 but not ATI-829 persistently induced an enhanced hypertriglyceridemia in male LDLR−/− mice with an increase of triglyceride-rich small VLDL particles.

Fig. 6.

Effect of LXR agonist on plasma lipid levels (A and B) and hepatic triglyceride levels (C) in male LDLR−/− mice after 12 weeks of treatment with vehicle (control, n=12), 10 mg/kg/day of ATI-829 (n=14), or 2 mg/kg/day (n=11) of T0901317. Significant differences between vehicle and treatment groups are indicated as * P<0.05.

Atherosclerosis was quantified in the innominate artery and aortic root. In the innominate artery, treatment with 10 mg/kg/day of ATI-829 and 2 mg/kg/day of T0901317 significantly decreased atherosclerotic plaque size by 60% and 86%, respectively (Fig. 7A). With 5 mg/kg/day of ATI-829 there was a trend towards reduction (23%) of atherosclerosis at this site (Fig. 7A). In the aortic root, the mean atherosclerotic lesion area was significantly reduced by 27% and 29% in mice treated with 10 mg/kg/day of ATI-829 and 2 mg/kg/day of T0901317, respectively (Fig. 7B). A dose of 5 mg/kg/day of ATI-829 caused only a small reduction (16%) that was not significant (Fig. 7B). Overall, both T0901317 and ATI-829 have greater efficacy in blocking development of atherosclerosis in the innominate artery than in the aortic root after 3 months on a Western diet.

Fig 7.

Effect of LXR agonists on atherosclerosis lesion area in the innominate artery (A), aortic root (B) and individual sinuses of the aortic root (C-F). Male LDLR−/− mice on a Western diet were gavaged daily for 12 weeks with vehicle (control, n=12), 5 (n=12) or 10 mg/kg/day (n=14) of ATI-829, or 2 mg/kg/day (n=11) of T0901317. Atherosclerotic lesions were measured as described in Methods. LC and RC are the left and right coronary associated sinuses, respectively and NC is the sinus not associated with a coronary artery. Significant differences between vehicle and compound treated groups are indicated as * P<0.05.

We also examined the efficacy of these compounds on blocking development of atherosclerosis in different sinuses of the aortic root. Lesions in the aortic root were designated LC and RC for left and right coronary artery-associated sinuses and NC for the remaining sinus that is not associated with a coronary artery (Fig. 7C). Interestingly, all treatments were able to significantly reduce the plaque size in RC. The reduction was 51%, 72% and 68% for treatment with 5 and 10 mg/kg/day of ATI-829 and 2 mg/kg/day of T0901317, respectively (Figure 7D). However, no difference was observed in plaque area in LC or NC in the different treatment groups (Fig. 7E and 7F).

The quality of the lesions was also evaluated. In general the lesions in the innominate artery with both T0901317 and ATI-829 treatments were not only smaller but also much less complex with a predominance of macrophage foam cells compared to the control group (Fig. 8A to 8D and 8M). In contrast, the lesions in the LC sinus from T0901317-treated mice and mice treated with 10mg/kg/day of ATI-829 were more complex with fewer macrophage foam cells and larger fibrous caps compared to vehicle-treated mice (Fig. 8E to 8N). In addition, the larger LC fibrous caps in either T0901317 or 10mg/kg/day of ATI-829 treated mice were collagen-positive areas after trichrome staining (Fig. 8I to 8L). Immunostaining of aortic root sections also revealed that both compounds significantly induced ABCA1 protein expression in both macrophages in atherosclerotic lesions and in the vascular wall of male LDLR−/− mice (Fig. 9).

Fig. 8.

Analysis of atherosclerotic lesion composition. Macrophage foam cells in lesions of the innominate artery (A-D) and aortic root (E-H) from mice treated with vehicle (A, E), 5 (B, F) or 10 (C, G) mg/kg/day of ATI-829, or 2 mg/kg/day of T0901317 (D, H) were immunostained with MOMA-2 antibody to detect macrophages. LC lesions from mice treated with vehicle (I), 5 (J) or 10 (K) mg/kg/day of ATI-829, or 2 mg/kg/day of T0901317 (L) were stained for collagen using trichrome stain (Masson). Percentage of macrophage foam cell area in the innominate and sinus plaques (M) and the average thickness of fibrous cap in LC lesions (N) were also quantitated. Significant differences between vehicle and compound treated groups are indicated as * P<0.05. Arrows indicate the fibrous cap. Original magnification: 100 × (A-D), 40 × (E-H) and 200 × (IL).

Fig. 9.

Expression of ABCA1 protein within aortic sinus plaques from male LDLR−/− mice. Aortic sinus sections were analyzed for ABCA1 and MOMA-2 protein expression by immunostaining. Original magnification: 200 ×.

Discussion

Long-term liver de novo lipogenesis caused by activation of LXR will produce fatty liver (Grefhorst et al., 2002) with the potential for liver cirrhosis and carcinogenesis. In addition, stimulation of lipogenesis leads to elevation of plasma triglycerides, which is an independent risk factor for coronary heart disease (Cullen, 2000). Obviously, side effects from the activation of LXR in liver will limit development of LXR agonists as pharmaceuticals for treatment of atherosclerosis. In this study, the novel synthetic steroidal LXR agonist ATI-829 appears to selectively activate LXR target gene expression in the intestine and macrophages rather than in liver. Enhanced expression of lipogenesis-associated genes such as SREBP-1c, ACC and FAS in liver was not triggered by ATI-829 treatment. Therefore, no fatty liver and plasma triglyceride accumulation was found in ATI-829 treated male LDLR−/− mice. Given its ability to activate gene expression in macrophages and intestine, ATI-829 has the potential for pharmacological development, since it lacks undesirable side effects associated with enhanced synthesis of triglycerides in liver. The mechanisms regarding its tissue-selectivity are still not clear and may depend on tissue-selective expression of LXR coactivators and corepressors (Misti et al., 1998), ligand-selective modulation of LXR interaction with coactivators and corepressors (Albers et al., 2006) or post-translational effects of LXR agonists on SREBP processing (Adams et al., 2004; Beyea et al., 2007).

Another notable finding of the present study is that long-term treatment of mice with T0901317, but not ATI-829, resulted in an accumulation of small triglyceride-rich VLDL in plasma. This finding contrasts with a previous report where wild type mice had an increase in large triglyceride-rich VLDL particles after T0901317 treatment (Grefhorst et al., 2002). However, in this previous study, mice were injected with Triton WR-1339, which protects newly secreted VLDL from metabolism in blood, to produce the large VLDL. In this work, the generation of small VLDL may be the result of LXR-mediated increases in expression of LPL (Zhang et al., 2001) and the apoE/C-I/C-IV/C-II gene cluster (Mak et al., 2002). ApoC-II enhances the catalytic activity of LPL, which hydrolyzes triglycerides in VLDL (Havel et al., 1970). Treatment of mice with ATI-829 did not result in an accumulation of small VLDL in the plasma. Based on the effects of ATI-829 on gene expression in liver, intestine and macrophages, tissue selective activation of gene expression is most likely responsible for differences in plasma lipoprotein metabolism.

In this study, long-term T0901317 treatment reduced LDL and HDL cholesterol, which is consistent with previous work (Levin et al., 2005). These reductions may result from T0901317-induced up-regulation of genes involved in lipoprotein metabolism, such as LPL and PLTP in liver and may also relate to the deficiency of LDL receptor. LDLR−/− mice fed an atherogenic diet have low HDL cholesterol compared to wild type C57BL/6 mice (Reardon et al., 2003; Terasaka et al., 2003; Levin et al., 2005). In general, the plasma lipoprotein profile in LDLR−/− mice is closer to humans than wild type C57BL/6 mice. The majority of plasma lipid in LDLR−/− mice fed an atherogenic diet is present in VLDL. Therefore, VLDL particles may play a critical role in atherogenesis in LDLR−/− mice.

The hypertriglyceridemia induced by the Western diet and the enhanced hypertriglyceridemia induced by T0901317 treatment would be expected to be pro-atherogenic, especially as pro-atherogenic plasma triglyceride-rich small VLDL particles are generated by T0901317 treatment. Small VLDL particles are thought to be atherogenic; whereas, large triglyceride-rich VLDL particles, such as those seen in severe hypertriglyceridemia caused by genetic LPL deficiency, are not associated with increased CHD risk (Cullen, 2000). However, a robust dose-dependent atheroprotection in the innominate artery was observed in mice treated with either compound. In fact, compared to ATI-829 treated mice, a more pronounced response was observed in T0901317-treated mice indicating that accumulation of pro-atherogenic plasma triglyceride-rich small VLDL from T0901317-treated mice did not significantly attenuate its anti-atherosclerotic effects in the innominate artery. While the effects on the innominate artery are clearcut, the responses of lesions in the aortic root are much less notable with no significant change in lesion size observed in the LC and NC sinuses of the aortic root after treatment with either compound. LXR expression levels have been found to differ between the thoracic aorta and the aortic arch (Zhu et al., 2008). LXR expression may also differ between the innominate artery and the aortic sinus and this may be responsible for the observed effects. However, our results are in contrast to a previous report where T0901317 treatment leads to a robust dose-dependent reduction of atherosclerosis in the aortic root of LDLR−/− mice (Terasaka et al., 2003). A major difference between the two studies is a transient hypertriglyceridemia in the previous study compared to a persistent hypertriglyceridemia in this report. In the previous study, sodium cholate was added to the diet. Cholate is an FXR agonist and activation of FXR can down-regulate the expression of short heterodimer partner (SHP) and LXR-associated SREBP-1c expression, which would decrease triglyceride synthesis (Watanabe et al., 2004). In our study, there was no sodium cholate in the diet and this difference may explain the persistent hypertriglyceridemia that we observed. Persistent hypertriglyceridemia may have an impact on the atheroprotective effects of LXR agonists on the aortic sinus.

The vascular response to various manipulations is frequently vascular site selective (VanderLaan et al., 2004). Therefore, the greater effect of a LXR agonist on atherosclerosis in the innominate artery than the aortic root is not totally unexpected. This site selectivity is perhaps driven by the local hemodynamic profile, or is reflective of differences in rate of lesion growth and development at the various vascular sites. A gradient of aortic permeability to LDL in a cephalic to caudal direction has been reported previously (Nielsen et al., 1992). LDL incorporation is greater along the lateral wall with lower shear force and higher susceptibility for atherosclerosis, than along the medial wall (Berceli et al., 1990). The aortic root lesion is generally initiated earlier than is the innominate artery lesion, and therefore the latter may also be less advanced than the LC lesion of the root. This site-specific action of LXR agonists on atherosclerosis in mice with hypertriglyceridemia may be explained at least partly by the influence of regional variation in hemodynamic factors, such as shear stress, arterial blood flow, and their effects on the migration and retention of lipoproteins, especially the plasma small size triglyceride-rich VLDL in this work, into the evolving lesion. Therefore, in comparison with T0901317, the tissue-selective LXR agonist ATI-829 has much less influence on plasma lipid profiles and is a promising drug candidate.

In the face of quite different plasma lipid and lipoprotein levels, the two agonists exert very similar influences on both innominate and aortic sinus atherosclerosis. This suggests that the atheroprotective effects of the LXR agonist may operate through mechanisms at the level of the vessel wall and not circulating lipoproteins. While the LXR agonist increases the level of genes involved in reverse cholesterol transport in the macrophage, which would by itself be expected to be atheroprotective, LXR stimulation also inhibits expression of pro-inflammatory factors, such as inducible nitric oxide synthase, cycloxygenase-1, macrophage chemoattractant protein-1, and macrophage chemoattractant protein-3 perhaps by limiting the activation of NF-κB (Zelcer and Tontonoz, 2006). Indeed it has been shown that bone marrow-derived cells play a major role in the atheroprotective influence of LXR (Tangirala et al., 2002). The monocyte precursor to the foam cell macrophage is likely to be the major actor here. The local anti-inflammatory action of LXR agonists may account for the thickened fibrous cap and more stable plaque observed in the LC region of the aortic root (Glass and Witztum, 2001). However, T and B cells are not essential for lesion development in the presence of extremely high circulating cholesterol levels and anti-inflammatory treatment also may be pro-atherogenic (Li et al., 2002; Szekanecz et al., 2007). The persistent accumulation of small triglyceride-rich VLDL and the inhibition of NF-κB by T0901317 may have adverse consequences and may potentially promote the apoptosis of macrophage and smooth muscle foam cells and the formation of necrotic core in plaques, and which possibly in turn could induce plaque rupture and thrombosis (Glass and Witztum, 2001).

In summary, atheroprotective effects of LXR agonists may be due to their roles in regulating lipid and glucose metabolism and inflammation (Cao et al., 2004). However, given the variety of factors that modulate atherosclerosis, the development of LXR agonists that are tissue, gene or LXRβ-selective, which can raise HDL-cholesterol and stimulate macrophage cholesterol efflux without causing liver and plasma triglyceride accumulation (Lund et al., 2006; Quinet et al., 2006), is critical for the development of LXR-based agonists as therapeutically useful agents. This work strongly supports the possibility of using steroidal-based LXR agonist, such as ATI-829 for development of a new class of drugs for treatment of atherosclerosis.

Abbreviations

FXR

farnesoid X receptor

RAR

retinoic acid receptor

TR

thyroid hormone receptor

VDR

vitamin D receptor

ER

estrogen receptor

AR

androgen receptor

GR

glucocorticoid receptor

LXR

liver X receptor

LDLR−/−

low density lipoprotein receptor deficiency

ATI-829

3α, 6α, 24-trihydroxy-24, 24-di(trifluoromethyl)-5β-cholane

T0901317

N-(2,2,2-trifluoroethyl)-N-[4-[2,2,2-trifluoro-1-hydroxy-1-(trifluoromethyl)ethyl]phenyl]-benzenesulfonamide

22(R)-HC

22(R)-hydroxycholesterol

24(S)-HC

24(S)-hydroxycholesterol

24(S)

25-EC, 24(S), 25-epoxycholesterol

VLDL

very low density lipoprotein

LDL

low density lipoprotein

HDL

high density lipoprotein

apoB

apolipoprotein B

apoE

apolipoprotein E

apoC-II

apolipoprotein CII

SREBP-1c

sterol regulatory element binding protein 1c

ABCA1, ABCG1, ABCG5, ABCG8

ATP-binding cassette transporter A1 or G1 or G5 or G8

FAS

fatty acid synthase

ACC

acetyl-coenzyme A carboxylase

36B4

acidic ribosomal phosphoprotein

LPL

lipoprotein lipase

PLTP

phospholipid transfer protein

CHD

coronary heart disease

DR

direct repeat

PBS

phosphate-buffered saline

FPLC

fast protein liquid chromatography

PCR

polymerase chain reaction

EM

electron microscopy

LC

left coronary artery-associated sinus

RC

right coronary artery-associated sinus

NC

sinus not associated with a coronary artery

NFκB

nuclear factor-κB