Estrogen receptor ligands ameliorate fatty liver through a nonclassical estrogen receptor/Liver X receptor pathway in mice

Authors

  • Song-iee Han,

    1. Graduate School of Life and Environmental Sciences, University of Tsukuba, Tsukuba, Ibaraki, Japan
    Search for more papers by this author
  • Yoko Komatsu,

    1. Graduate School of Life and Environmental Sciences, University of Tsukuba, Tsukuba, Ibaraki, Japan
    Search for more papers by this author
  • Akiko Murayama,

    1. Graduate School of Life and Environmental Sciences, University of Tsukuba, Tsukuba, Ibaraki, Japan
    2. Center for Tsukuba Advanced Research Alliance (TARA), Graduate School of Comprehensive Human Sciences, University of Tsukuba, Tsukuba, Ibaraki, Japan
    Search for more papers by this author
  • Knut R. Steffensen,

    1. Division of Clinical Chemistry, Department of Laboratory Medicine, Karolinska Institutet, Karolinska University Hospital Huddinge, Stockholm, Sweden
    Search for more papers by this author
  • Yoshimi Nakagawa,

    1. Department of Internal Medicine (Endocrinology and Metabolism), Faculty of Medicine, University of Tsukuba, Ibaraki, Japan
    Search for more papers by this author
  • Yuka Nakajima,

    1. Graduate School of Life and Environmental Sciences, University of Tsukuba, Tsukuba, Ibaraki, Japan
    Search for more papers by this author
  • Michiko Suzuki,

    1. Graduate School of Life and Environmental Sciences, University of Tsukuba, Tsukuba, Ibaraki, Japan
    Search for more papers by this author
  • Shohei Oie,

    1. Graduate School of Life and Environmental Sciences, University of Tsukuba, Tsukuba, Ibaraki, Japan
    Search for more papers by this author
  • Paolo Parini,

    1. Division of Clinical Chemistry, Department of Laboratory Medicine, Karolinska Institutet, Karolinska University Hospital Huddinge, Stockholm, Sweden
    Search for more papers by this author
  • Lise-Lotte Vedin,

    1. Center for Biosciences, Department of Biosciences and Nutrition, Karolinska Institutet, Stockholm, Sweden
    Search for more papers by this author
  • Hiroyuki Kishimoto,

    1. Center for Tsukuba Advanced Research Alliance (TARA), Graduate School of Comprehensive Human Sciences, University of Tsukuba, Tsukuba, Ibaraki, Japan
    Search for more papers by this author
  • Hitoshi Shimano,

    1. Center for Tsukuba Advanced Research Alliance (TARA), Graduate School of Comprehensive Human Sciences, University of Tsukuba, Tsukuba, Ibaraki, Japan
    2. Department of Internal Medicine (Endocrinology and Metabolism), Faculty of Medicine, University of Tsukuba, Ibaraki, Japan
    3. International Institute for Integrative Sleep Medicine (WPI-IIIS), University of Tsukuba, Ibaraki, Japan
    Search for more papers by this author
  • Jan-Åke Gustafsson,

    1. Center for Biosciences, Department of Biosciences and Nutrition, Karolinska Institutet, Stockholm, Sweden
    2. Center for Nuclear Receptors and Cell Signaling, Department of Biology and Biochemistry, University of Houston, Houston, TX
    Search for more papers by this author
  • Junn Yanagisawa

    Corresponding author
    1. Graduate School of Life and Environmental Sciences, University of Tsukuba, Tsukuba, Ibaraki, Japan
    2. Center for Tsukuba Advanced Research Alliance (TARA), Graduate School of Comprehensive Human Sciences, University of Tsukuba, Tsukuba, Ibaraki, Japan
    • Address reprint requests to: Junn Yanagisawa, Ph.D., Graduate School of Life and Environmental Sciences, and Center for Tsukuba Advanced Research Alliance (TARA), Graduate School of Comprehensive Human Sciences, University of Tsukuba, Tsukuba, Ibaraki 305-8577, Japan. E-mail: junny@agbi.tsukuba.ac.jp; fax: +81-29-853-7322.

    Search for more papers by this author

  • Potential conflict of interest: Nothing to report.

  • This work was supported by the Research Fellowship (to S.H. and Y.K.) from the Japan Society for the Promotion of Science. J-A.G. was supported by the Swedish Cancer Society, the Texas Emerging Technology Fund (under Agreement 18, no. 300-9-1958), and the Robert A. Welch Foundation (E-0004).

Abstract

Liver X receptor (LXR) activation stimulates triglyceride (TG) accumulation in the liver. Several lines of evidence indicate that estradiol-17β (E2) reduces TG levels in the liver; however, the molecular mechanism underlying the E2 effect remains unclear. Here, we show that administration of E2 attenuated sterol regulatory element-binding protein (SREBP)-1 expression and TG accumulation induced by LXR activation in mouse liver. In estrogen receptor alpha (ERα) knockout (KO) and liver-specific ERα KO mice, E2 did not affect SREBP-1 expression or TG levels. Molecular analysis revealed that ERα is recruited to the SREBP-1c promoter through direct binding to LXR and inhibits coactivator recruitment to LXR in an E2-dependent manner. Our findings demonstrate the existence of a novel liver-dependent mechanism controlling TG accumulation through the nonclassical ER/LXR pathway. To confirm that a nonclassical ER/LXR pathway regulates ERα-dependent inhibition of LXR activation, we screened ERα ligands that were able to repress LXR activation without enhancing ERα transcriptional activity, and, as a result, we identified the phytoestrogen, phloretin. In mice, phloretin showed no estrogenic activity; however, it did reduce SREBP-1 expression and TG levels in liver of mice fed a high-fat diet to an extent similar to that of E2. Conclusion: We propose that ER ligands reduce TG levels in the liver by inhibiting LXR activation through a nonclassical pathway. Our results also indicate that the effects of ER on TG accumulation can be distinguished from its estrogenic effects by a specific ER ligand. (Hepatology 2014;59:1791–1802)

Abbreviations
9-cRA

9-cis-retinoic acid

Abs

antibodies

ChIP

chromatin immunoprecipitation

E2

estradiol-17β

EMSA

electrophoretic mobility shift assay

ERα

estrogen receptor alpha

ERE

estrogen response element

GST

glutathione S-transferase

HA

hemagglutinin

HFD

high-fat diet

IgG, ; IP

intraperitoneally

L-ERα

liver-specific ERα

KO

knockout

LXR

liver X receptor

LXRE

LXR response element

mRNA

messenger RNA

MUT LXRE

mutated LXRE

MetS

metabolic syndrome

NCoR

nuclear receptor corepressor

OVX

ovariectomized

PBS

phosphate-buffered saline

PCR

polymerase chain reaction

RXR

retinoid X receptor

SCD

stearoyl CoA desaturase

SEM

standard error of the mean

SRC-1

steroid receptor coactivator 1

SREBP

sterol regulatory element-binding protein

T0

T0901317

TG

triglyceride

WT

wild type

Fatty liver, caused by triglyceride (TG) accumulation in the liver, has been associated with metabolic syndrome (MetS) and is known to trigger type 2 diabetes, atherosclerosis, and other metabolic diseases.[1, 2] It is desirable to elucidate the mechanism of fatty liver pathogenesis.

Liver X receptors (LXRs; LXRα and LXRβ) are sterol sensors and bind oxysterols to regulate genes critical to cholesterol, lipid, and glucose metabolism.[3, 4] In the liver, LXRs activate lipogenesis by increasing expression of sterol regulatory element-binding protein (SREBP)-1c, which controls expression of key genes involved in fatty acid biosynthesis.[5] LXRα is expressed in tissues with high metabolic activity, including liver, adipose tissue, and macrophages, whereas LXRβ is ubiquitously expressed. LXRs form an obligate heterodimer with the retinoid X receptor (RXR) and bind to a specific DNA sequence called the LXR response element (LXRE).

In mammals, estradiol-17β (E2) and its receptors, estrogen receptor (ER)α and ERβ, play a major role in reproductive functions as well as in the regulation of food intake and lipid and glucose homeostasis.[6-8] Recent studies have shown that E2 reduces TG accumulation in high-fat diet (HFD)-fed and ob/ob mice[9, 10]; however, the underlying mechanism in the liver remains to be completely elucidated.

In this study, we demonstrate the existence of a novel liver-dependent mechanism controlling TG accumulation through the ER/LXR pathway. Elevation of SREBP-1 expression and TG accumulation was attenuated in livers of wild-type (WT) mice by administration of E2. In contrast, E2 did not affect SREBP-1 expression and TG levels in ERα knockout (KO) and liver-specific ERα KO (L-ERα KO) mouse liver. These results suggest that ERα regulates SREBP-1 expression and TG accumulation in the liver. We showed that ERα was recruited to the SREBP-1c promoter by directly binding to LXR. Furthermore, we identified the phytoestrogen, phloretin, which repressed LXR transcriptional activity and TG accumulation through ERα without enhancement of ERα transcriptional activity in mouse liver. Our results demonstrate that ER ligands reduce TG levels in the mouse liver by inhibiting LXR transcriptional activity through a nonclassical pathway.

Materials and Methods

Mice and Cells

Five-week-old female C57BL/6 WT mice were purchased from the CLEA Japan, Inc. (Kawasaki, Japan), and mice were ovariectomized (OVX) for 2 weeks. Mice were treated with T0901317 (T0; Cayman Chemical Company, Ann Arbor, MI) and/or E2 (Sigma-Aldrich, St. Louis, MO) or phloretin (Sigma-Aldrich) by intraperitoneal (IP) administration for 12 hours, 2 days, or 8 days. For the HFD experiment, mice were put on normal chow or HFD (21% fat, 0.2% cholesterol, TD88137; Harlan Teklad, Indianapolis, IN) and were administered with E2 or phloretin by Alzet mini osmotic pump (0.25 µl/hour) for 4 weeks. ERα conditional (ERαflox/flox) and null alleles with C57BL/6J background have been previously described.[11] L-ERα KO mice were generated by breeding ERαflox/flox mice with B6.Cg-Tg (Alb-cre)21Mgn/J (The Jackson Laboratory, Bar Harbor, ME), in which the promoter of the albumin gene drives expression of cre-recombinase protein specifically in the liver, as described previously.[12] LXRα and LXRα/β KO mice have been previously described.[13] All animal husbandry and animal experiments were consistent with the University of Tsukuba's Regulation of Animal Experiment Committee (Ibaraki, Japan). 293E cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum.

Oil Red O Staining

Liver tissues were harvested in cold phosphate-buffered saline (PBS), fixed overnight at 4°C in 4% paraformaldehyde in PBS, cryoprotected in 30% sucrose in PBS, embedded in optimal cutting temperature compound (Sakura Finetek Europe B.V., Alphen aan den Rijn, The Netherlands), and frozen. Frozen tissues were cut into 5-µm-thick cryosections and stained with Oil Red O (Sigma-Aldrich).

Measurement of Liver and Plasma TG Content

Liver tissues were homogenized in 5% Triton X-100 in water (1:10, w/v). Samples were slowly heated to 80°C for 5 minutes. Insoluble materials were removed by centrifugation. TG secretion into plasma was determined after tail vein injection of tyloxapol (Sigma-Aldrich) after 6 hours of fasting and blood was collected by the tail vein. TG concentration in the supernatant was determined by the FUJI DRI-CHEM analyzer 3000 (Fujifilm, Tokyo, Japan).

Real-Time Reverse-Transcription Polymerase Chain Reaction

Mouse liver tissues were homogenized in Sepazol, and total RNA was extracted according to the manufacturer's instructions (Nacalai Tesque, Inc., Kyoto, Japan). Complementary DNA was synthesized from total RNA using RevatraAce reverse transcriptase (Toyobo, Tokyo, Japan) and oligo dT primer. Real-time polymerase chain reactions (PCRs) were performed to amplify fragments representing for the indicated messenger RNA (mRNA) expression using Thermal Cycle Dice Real Time System (TP800; Takara Bio, Otsu, Japan) and SYBR Premix Ex Taq (Takara Bio). Primer sequences are provided in Supporting Table 1A.

Antibodies, Coimmunoprecipitation, Glutathione S-Transferase Pull-Down Assay, and Immunoblotting

Rabbit anti-RXRα (sc-774; Santa Cruz Biotechnology, Santa Cruz, CA), anti-p300 (sc-585; Santa Cruz Biotechnology), anti-nuclear receptor corepressor (NCoR; sc-1609; Santa Cruz Biotechnology), anti-steroid receptor coactivator 1 (SRC-1; sc-8995; Santa Cruz Biotechnology), anti-RIP140 (ab3425; Abcam, Cambridge, MA), anti-TRRAP (ab73546; Abcam), anti-ASC-2 (ab18193; Abcam), anti-BRG-1 (sc-10768; Santa Cruz Biotechnology) polyclonal antibodies (Abs), rabbit anti-LXRα/β (gift from Dr. J.-Å. Gustafsson), rabbit anti-ERα (constructed by MBL International, Nagoya, Japan), mouse anti-ERα (diagenode, AC-066-100), anti-FLAG-M2 (F1804; Sigma-Aldrich), and anti-hemagglutinin (HA; 11867423001; Roche Diagnostics, Indianapolis, In) Abs were used according to the manufacturer's instructions. Nuclear extracts of mouse liver were prepared as described previously.[14] Extracted liver proteins were immunoprecipitated with rabbit immunoglobulin G (IgG; sc-2027; Santa Cruz Biotechnology) or anti-LXRα/β Ab. Glutathione S-transferase (GST) pull-down assay was performed as described previously.[15] Bound proteins were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis, transferred to polyvinylidine difluoride membranes (Millipore, Billerica, MA), and detected with appropriate primary Abs and horseradish-peroxidase–conjugated secondary Abs. Specific proteins were visualized using an enhanced chemiluminescence western blotting detection system (Amersham Biosciences, Amersham, UK).

Expression Vectors

DNA fragments encoding full-length, deletion mutants of human ERα and full-length mouse RXRα were amplified by PCR and cloned into the pcDNA3 vector containing sequences coding for FLAG. Similarly, the human LXRα and β full-length and deleted DNA were cloned into the pcDNA3 vector containing both FLAG and HA. We prepared GST fusion protein constructs of full-length LXRα/β, ERα, and RXRα by subcloning the protein-coding regions into pGEX vector.

Chromatin Immunoprecipitation Assay

Liver tissues were fixed in 1% formaldehyde, and chromatin immunoprecipitation (ChIP) assays were performed essentially as described previously.[16] For re-ChIP assay, 293E cells were cotransfected with pcDNA3 LXRα and ERα expression plasmid by PerFectin Transfection Reagent (Gene Therapy Systems, Inc., San Diego, CA). DNA was amplified by real-time PCR, as described above. Primers for real-time PCR of the SREBP-1c gene upstream region are provided in Supporting Table 1B.

Electrophoretic Mobility Shift Assay

Electrophoretic mobility shift assay (EMSA) was performed as described previously.[17] ERα, LXRα, and RXRα proteins were synthesized in vitro using the TNT T7 quick-coupled transcription/translation system (Promega, Madison, WI). Oligonucleotide sequences are provided in Supporting Table 1C.

Avidin-Biotin Complex DNA-Binding Assay

Avidin resin (Promega) was incubated with biotin-conjugated consensus WT LXRE or mutated LXRE (MUT LXRE) oligonucleotides, followed by incubation with liver extracts in binding buffer (20 mM of HEPES, 100 mM of KCl, 0.5 mM of ethylenediaminetetraacetic acid, 0.1% Triton X-100, and 1 mM of dithiothreitol) for 30 minutes. The subsequent LXRE-protein complexes trapped on the resin were then eluted and western blotted. Oligonucleotide sequences are provided in Supporting Table 1D.

Luciferase Reporter Assay

Mouse SREBP-1c promoter region (−572 to 0) and three tandem estrogen response element (ERE) repeats were cloned into a pGL3-basic reporter plasmid. For luciferase assays, 293E cells were cotransfected with reporter plasmid and pcDNA3 LXRα, RXRα, and ERα WT or ΔC expression plasmid. phRG-TK (Promega) is a reference plasmid used to normalize transfection efficiency. Luciferase assays were performed on cell extracts according to the manufacturer's instruction (Promega).

Statistical Analysis

Data represent the mean ± standard error of the mean (SEM). Statistical significance was evaluated by unpaired two-tailed Student t test. P < 0.05 was considered a significant difference.

Results

E2 Inhibits TG Accumulation and SREBP-1 Expression Through the ER/LXR Pathway in the Liver

LXRs are known to lead to TG accumulation in the liver because of increased expression of lipogenic genes.[3, 4] To investigate the effect of E2 on LXR-dependent lipogenesis, we examined TG accumulation in the liver. The LXR ligand, T0 and/or E2, were injected IP into OVX female mice and TG levels in the liver were determined by Oil Red O staining and TG extraction assays. T0 increased TG levels in the liver and E2 attenuated T0-dependent TG accumulation (Fig. 1A,B). We measured liver TG secretion by injecting mice with tyloxapol to inhibit TG catabolism. Accumulation of plasma TG after tyloxapol injection was significantly less in T0 + E2-treated mice than T0-treated mice (Fig. 1C). These results suggest that E2 reduces LXR-dependent lipogenesis in the liver.

Figure 1.

E2 inhibits T0-dependent TG accumulation and SREBP-1 expression in the liver. OVX mice were IP administered T0 (50 mg/kg) and/or E2 (10 mg/kg). (A and B) TG levels in the liver were determined by Oil Red O staining (A) and TG extraction assays (B) 8 days after ligand administration. (C) Mice were IP administered T0 and/or E2 for 2 days. Plasma TG at 0 and 120 minutes after tail vein injection of tyloxapol (500 mg/kg) was measured. (D) SREBP-1, LXRα, and LXRβ mRNA levels in the liver were determined by real-time PCR 12 hours after ligand administration. (E) mRNA levels of SCD-1 were determined by real-time PCR 8 days after ligand administration. Data represent the mean ± SEM (n = 3-12). *P < 0.05; **P < 0.01; n.s., not significant.

Next, we examined expression levels of SREBP-1, which is a master regulator of TG accumulation in the liver. Consistent with previous reports,[18] T0 increased expression of SREBP-1 (Fig. 1D). Elevation of SREBP-1 expression by T0 was reduced by E2 treatment (Fig. 1D). LXR mRNA levels were not altered by E2 (Fig. 1D). The stearoyl CoA desaturase (SCD)-1 gene, a target gene of SREBP-1 known to be involved in TG synthesis,[5] was up-regulated by T0. T0-dependent SCD-1 expression was also suppressed by E2 treatment (Fig. 1E). Note that the TG and SREBP-1 mRNA levels in non-OVX female mouse liver were lower than those in OVX female mouse liver by T0 treatment without changing protein levels of LXR and ER (Supporting Fig. 1A-C). These results suggest that E2 suppresses LXR-dependent expression of SREBP-1 and its downstream targets in mouse liver.

E2 binds to ERα and β, which then regulate the transcription of various target genes.[19, 20] Because the expression level of ERα is much higher than that of ERβ in the liver,[21] we investigated whether ERα is necessary for the effect of E2 on SREBP-1 expression and TG accumulation using ERα KO and L-ERα KO mice. In the presence of T0, TG levels and SREBP-1 expression in livers of ERα KO or L-ERα KO mice were comparable to those in WT mice (Fig. 2A-D). However, elevation of TG accumulation and SREBP-1 expression induced by T0 were not affected by E2 treatment (Fig. 2A-D), indicating that E2 reduction of TG accumulation and SREBP-1 expression in the liver requires ERα.

Figure 2.

Effects of E2 on TG accumulation and SREBP-1 expression are mediated through the ER/LXR pathway in the liver. OVX WT, ERα KO, L-ERα KO, and LXRα KO mice were IP administered T0 (50 mg/kg) and/or E2 (10 mg/kg). (A, C, and E) TG levels in the liver were determined by TG extraction assays. (B, D, and F) SREBP-1 mRNA levels in the liver were determined by real-time PCR. Data represent mean ± SEM (n = 3-10). *P < 0.05; n.s., not significant.

To confirm that the effects of E2 were mediated through LXR, we examined TG accumulation and SREBP-1 expression in livers of LXRα KO mice. In LXRα KO mouse livers, levels of TG and SREBP-1 expression were not increased by T0 treatment, and E2-dependent reduction of TG and SREBP-1 expression levels was also not observed (Fig. 2E,F). These results indicate that the effects of E2 on TG accumulation and SREBP-1 expression are mediated through the ER/LXR pathway in the liver.

ERα is Recruited to the SREBP-1c Promoter Through Direct Binding to LXR

We then investigated the molecular mechanisms of regulation of SREBP-1 expression by ERα. First, we determined ERα occupancy on the SREBP-1c promoter. ChIP experiments using mouse liver extracts showed that ERα was recruited to the LXRE in the SREBP-1c promoter with LXRα/β and RXRα (Fig. 3B-D). Recruitment of ERα, LXRα/β, and RXRα to the promoter region was not affected by treatment with E2 and/or T0 (Fig. 3B-D). Accordingly, we tested whether coactivator p300 and corepressor NCoR occupancy in the SREBP-1c promoter was affected by E2 treatment. In agreement with previous reports,[22, 23] T0 treatment induced the dissociation of corepressor NCoR from the SREBP-1c promoter and increased recruitment of p300 to the SREBP-1c promoter (Fig. 3E,F). E2 treatment did not affect T0-dependent NCoR dissociation, indicating that E2-mediated repression of LXR activity does not occur because of the recruitment of NCoR (Fig. 3E). In contrast, recruitment of p300 to the SREBP-1c promoter by T0 treatment was abrogated by E2 in WT mouse liver (Fig. 3F). To test whether p300 occupancy of the SREBP-1c promoter was regulated by ERα, we performed ChIP assays using ERα KO mouse liver extracts. In ERα KO mouse liver, recruitment of p300 to the promoter region induced by T0 was not affected by E2 treatment (Fig. 3G). Thus, our results indicate that T0 enhances the association of the coactivator complex comprising p300 and LXR and that recruitment of the coactivator complex to LXR is abrogated by E2 treatment.

Figure 3.

ERα is recruited to the SREBP-1c promoter and inhibits T0-dependent p300 recruitment in E2-dependent manner. (A) A schematic view of primer locations at the SREBP-1c promoter is shown. (B-G) Livers from WT (B-F) or ERα KO (G) mice treated with the indicated ligands were analyzed by ChIP assays using the indicated Abs. Immunoprecipitated DNA was assessed by real-time PCR using primers specific for LXRE or the distal region of the SREBP-1c promoter. Samples were normalized to the input DNA. Data represent mean ± SEM (n = 3-9). *P < 0.05; **P < 0.01; n.s., not significant.

Our results indicate that ERα occupied the SREBP-1c promoter in complex with LXR and RXR. However, the SREBP-1c promoter does not contain the ERE, a DNA element necessary for ERα binding. Accordingly, we hypothesized that ERα is recruited to the LXRE in the SREBP-1c promoter through LXR. To test this hypothesis, we first examined interaction between ERα and LXRα/β in mouse liver. ERα coimmunoprecipitated with LXRα/β, indicating that ERα interacts with LXRα/β (Fig. 4A). We then investigated the direct binding of ER, LXR, and RXR to the LXRE of the SREBP-1c promoter using EMSA. LXRα and RXRα did not bind to the LXRE individually, but when they were mixed together, an LXRα/RXRα–LXRE complex was formed (Supporting Fig. 2A). ERα did not directly bind to the LXRE, whereas it bound to the ERE (Supporting Fig. 2A,B). Coincubation of ERα with LXRα/RXRα resulted in an increased intensity of retarded band corresponding to the LXRα/RXRα–LXRE complex (Supporting Fig. 2A,C), and this band decreased with the addition of unlabeled WT ERE oligonucleotides, suggesting that ERα enhances a complex of LXRα/RXRα with the LXRE (Supporting Fig. 2C). Next, we examined whether ERα is recruited to the LXRE in combination with LXRα/β using Avidin-biotin complex DNA-binding assay. Nuclear extracts prepared from mouse liver were incubated with the LXRE (WT LXRE) oligonucleotide conjugated to beads. Western blotting analysis revealed that the proteins that were recruited to the LXRE oligonucleotide included ERα, p300, and SRC-1 with LXRα/β and RXRα (Fig. 4B). Subsequently, we generated a MUT LXRE oligonucleotide, which did not bind with LXRα/β and incubated the MUT LXRE with mouse liver nuclear extracts. Immunoblotting analysis confirmed that LXRα/β and RXRα did not interact with the MUT LXRE (Fig. 4B). Consistent with these results, ERα, p300, and SRC-1 were not recruited to the MUT LXRE (Fig. 4B). Finally, we performed a ChIP assay using extracts derived from WT or LXRα/β KO mouse liver. Recruitment of ERα to the SREBP-1c promoter was not observed in LXRα/β KO mouse liver (Fig. 4C). These results indicate that ERα forms a complex with LXRα/β and RXRα and that this complex interacts with the LXRE.

Figure 4.

ERα is recruited to the LXRE through direct binding to LXR. (A) Nuclear extracts of mouse liver were immunoprecipitated (IP) with rabbit IgG or anti-LXRα/β Ab. Bound proteins were immunoblotted with anti-ERα and anti-LXRα/β Abs. (B) Nuclear extracts of mouse liver were incubated with WT or MUT LXRE oligonucleotides, which were conjugated to beads. In the no-LXRE lane, nuclear extracts were conjugated to beads without LXRE oligonucleotides. Proteins that interacted with the oligo-beads were detected by immunoblotting using the indicated Abs. (C) Livers from WT or LXRα/β KO mice were analyzed by ChIP assays using the anti-ERα Ab. Immunoprecipitated DNA was assessed by real-time PCR using primers specific for LXRE of the SREBP-1c promoter. Samples were normalized to the input DNA. (D) Full-length and truncated Flag-ERα (indicated schematically at the left panel) were synthesized in vitro and used in GST pull-down assays with glutathione-Sepharose–immobilized GST, GST-LXRα, or GST-LXRβ. Precipitates were analyzed by immunoblotting using anti-FLAG-M2 Ab. (E) 293E cells were cotransfected with the indicated expression plasmids. Re-ChIP assays were performed using the indicated Abs. Immunoprecipitated DNA was assessed by real-time PCR using primers specific for LXRE of the SREBP-1c promoter. Samples were normalized to the input DNA. (F) 293E cells were cotransfected with a luciferase reporter plasmid containing the SREBP-1c promoter (−572 to 0) and the indicated expression plasmids. Cells were treated with T0 (10−7 M) and 9-cRA (10−6 M) and/or E2 (10−8 M) for 24 hours. Cell extracts were examined using luciferase assays. Data represent mean ± SEM (n = 3-9). *P < 0.05; n.s., not significant.

To identify the regions in ERα responsible for binding to LXRα/β, we generated truncated forms of ERα and performed GST pull-down assays. Binding between ERα and LXRα/β was observed in the in vitro GST pull-down assay, demonstrating that ERα bound directly to LXRα/β. ERα WT and ERα ABC bound LXRα/β; however, ERα DEF and ΔC did not, indicating that the C (DNA-binding) region in ERα is necessary for interaction between ERα and LXRα/β (Fig. 4D). We next generated truncated forms of LXRα and LXRβ and performed GST pull-down assays (Supporting Fig. 3). The results revealed that the ABC domain of LXR was necessary for interaction between LXR and ERα. Furthermore, we generated GST-RXRα and GST-ERα and subsequently performed the GST pull-down assay with ERα and RXRα, respectively. The results showed that ERα does not interact directly with RXRα (Supporting Fig. 4). Accordingly, we hypothesized that ERα interacts with LXR of the LXR/RXR heterodimer. We speculate that ERα is recruited to the LXRE through interaction with its C domain and ABC domain of LXR.

To confirm whether ERα was recruited to the LXRE through LXR, we performed re-ChIP assays. We performed the first ChIP assay using anti-LXRα/β Ab and the second using anti-ERα Ab. We observed SREBP-1c promoter-specific enrichment in the second ChIP fraction with the ERα Ab (Fig. 4E). Consistent with the GST pull-down results, ERα ΔC was not recruited to the SREBP-1c promoter (Fig. 4E). Also, EMSA showed that unlike ERα WT, the addition of ERα ΔC with LXRα/RXRα did not increase the intensity of the LXRα/RXRα–LXRE band (Supporting Fig.2D). Next, we explored the effect of ERα on LXR transcriptional activity by performing reporter assays with a luciferase reporter plasmid containing the SREBP-1c promoter. In this assay, LXR transcriptional activity was induced by T0 and RXR ligand, 9-cis retinoic acid (9-cRA) treatment. When ERα WT was coexpressed with LXR and RXR, potentiation of LXR transcriptional activity by T0 and 9-cRA was repressed by E2 treatment (Fig. 4F). However, potentiation of LXR transcriptional activity induced by T0 and 9-cRA was not affected by E2 treatment in ERα ΔC-expressing cells (Fig. 4F). These results strengthened the evidence that ERα is recruited to the LXRE through LXR.

Phloretin Inhibits T0-Dependent SREBP-1 Expression and TG Accumulation Without Enhancing ER Transactivation

In the classical pathway, ERs bind to a specific ERE and regulate the transcription of target genes.[24] However, it has also been shown that ERs regulate transcription at alternative response elements to which they are recruited by protein-protein interactions; this is defined as the nonclassical pathway.[19, 20] The findings presented above demonstrate that E2 can suppress LXR-dependent transcription through ERα-LXR interactions (nonclassical pathway).

To test whether the ER-LXR nonclassical pathway is essential for E2-dependent suppression of SREBP-1c gene expression and TG accumulation in mice, we attempted to identify ERα ligands that inhibit LXR-mediated transcription without inducing ERα transcriptional activation. We obtained 511 compounds that were expected to bind ERα from 100,000 compounds using an in silico screen and evaluated the effects of the selected compounds on ERα- and LXR-dependent transcription by luciferase assays in 293E cells. As a result, we identified the phytoestrogen, phloretin (Supporting Fig. 5A and Fig. 5A), which has previously been reported to bind ERα.[25] Phloretin did not activate the ERE reporter and repressed the SREBP-1c promoter. To confirm whether phloretin binds to ERα, we performed in vitro ligand-binding assays (Supporting Fig. 5B). The ligand-binding assay revealed that phloretin bound to ERα, and the affinity of phloretin for ERα was much lower than that of E2 (half maximal effective concentration/E2, 0.6 nM; phloretin, 8.03 µM). Our results demonstrate that phloretin is a ligand for ERα and that it suppresses LXR-dependent transcription without enhancing ERα-dependent transcription in vitro.

Figure 5.

Phloretin inhibits T0-dependent SREBP-1 expression and TG accumulation without enhancing ER transactivation. (A) Structure of phloretin. (B) OVX mice were treated with E2 (10 mg/kg) or phloretin (100 mg/kg) for 8 days, and quantitative weight of uteri was determined. (C) OVX mice were administered with the indicated ligands. After 12 hours, mRNA levels of STAT5A and LRH-1 in the liver were determined by real-time PCR. (D) WT and ERα KO mice were IP administered T0 (50 mg/kg) and/or phloretin (100 mg/kg). SREBP-1, LXRα, and LXRβ mRNA levels in the liver were determined by real-time PCR 12 hours after administration. (E and F) WT and ERα KO mice received daily IP administrations of the indicated ligands. After 8 days, TG levels in the liver were determined by Oil Red O staining (E) and TG extraction assays (F). Data represent mean ± SEM (n = 3-7). *P < 0.05; n.s., not significant.

Next, we evaluated the estrogenic effect of phloretin in vivo. In OVX mice, E2 treatment increased uterine size, whereas phloretin treatment did not (Fig. 5B). To determine whether phloretin induces ERα transcriptional activity in the liver, we determined expression levels of the classical ER target genes, STAT5A and LRH-1.[26, 27] Unlike E2, phloretin did not increase expression levels of these ER target genes in mouse liver (Fig. 5C). These results indicate that phloretin does not exhibit estrogenic activity in mice.

We further examined the effect of phloretin on SREBP-1 expression in mouse liver. Phloretin repressed T0-induced SREBP-1 expression without affecting LXR expression (Fig. 5D). Phloretin-dependent inhibition of SREBP-1 expression induced by T0 was not observed in ERα KO mouse liver (Fig. 5D), indicating that phloretin suppresses T0-dependent SREBP-1 expression through ERα. To test the effect of phloretin on LXR-dependent lipogenesis, we evaluated TG accumulation in the liver. We found that phloretin suppressed T0-dependent TG accumulation in mouse liver, and these effects were not observed in ERα KO mouse liver (Fig. 5E,F). These observations suggest that phloretin reduces T0-dependent LXR transactivation and TG accumulation through ERα without enhancing ERα transcriptional activity.

We then investigated the molecular mechanisms of the effects of phloretin on LXR target gene expression. ChIP assays revealed that phloretin did not affect recruitment of LXRα/β, RXRα, or ERα to the LXRE in the SREBP-1c promoter region (Fig. 6A-C). Similar to the effects of E2, phloretin inhibited T0-dependent recruitment of p300 to the SREBP-1c promoter, but did not affect T0-dependent NCoR dissociation from the SREBP-1c promoter (Fig. 6D). In ERα KO mouse liver, T0-dependent recruitment of p300 to the promoter region was not affected by phloretin (Fig. 6E). These results suggest that binding of phloretin to ERα inhibits recruitment of coactivators to promoters and reduces LXR-dependent transcriptional activity, similar to the effects of E2 binding.

Figure 6.

Phloretin inhibits T0-dependent p300 recruitment to the SREBP-1c promoter. Livers from WT (A-D) or ERα KO (E) mice treated with the indicated ligands were analyzed by ChIP assays using the indicated Abs. Immunoprecipitated DNA was assessed by real-time PCR using primers specific for LXRE of the SREBP-1c promoter. Samples were normalized to the input DNA. Data represent mean ± SEM (n = 3-8). *P < 0.05; **P < 0.01; n.s., not significant.

E2 and Phloretin Inhibit HFD-Dependent LXR Target Gene Expression and TG Accumulation in the Mouse Liver

The HFD-fed mouse is a model of obesity. Previous studies have indicated that an HFD induces LXR transactivation, which causes TG accumulation in the liver.[28, 29] It was known that the oxysterols involved in HFD are endogenous ligands for LXRs.[3, 4] It has also been reported that E2 reduces TG accumulation in HFD-fed mice.[9] Accordingly, we evaluated effects of E2 and phloretin on HFD-dependent LXR target gene expression and TG accumulation in WT and L-ERα KO mice in the liver.

In HFD-fed WT mouse liver, SREBP-1 and SCD-1 expression were elevated, compared to their expression levels in the normal chow-fed mouse liver, as expected (Fig. 7A). We obtained the same results in HFD-fed L-ERα KO mouse livers (Fig. 7A). Real-time PCR demonstrated that both E2 and phloretin inhibited HFD-dependent LXR target gene expression in WT mouse liver, whereas the same effect was not observed in L-ERα KO mouse liver (Fig. 7A). In addition, E2 and phloretin treatment reduced HFD-dependent TG accumulation in WT mouse livers, but those ligands did not reduce TG levels in L-ERα KO mouse liver (Fig. 7B,C). Next, we examined recruitment of LXRα/β, RXRα, ERα, and p300 to the SREBP-1c promoter in HFD-fed mouse livers. The ChIP assay showed that recruitment of LXRα/β, RXRα, and ERα to the SREBP-1c promoter was not changed by HFD or any ER ligand treatments (Supporting Fig. 6). In contrast, recruitment of p300 to the SREBP-1c promoter was increased by HFD, and this effect was abrogated by E2 or phloretin (Fig. 7D). However, in L-ERα KO mice, recruitment of p300 to the SREBP-1c promoter by HFD was not affected by E2 or phloretin treatment (Fig. 7D). These results confirm that increased LXR transactivation and TG accumulation by feeding with HFD were reduced by E2 or phloretin through the ER/LXR pathway.

Figure 7.

E2 and phloretin inhibit HFD-dependent LXR target gene expression and TG accumulation through the ER/LXR pathway in the mouse liver. WT and L-ERα KO mice were fed normal chow or HFD and administered with E2 (10 mg/kg/day) or phloretin (100 mg/kg/day) for 4 weeks. (A) mRNA levels of SREBP-1 and SCD-1 were determined by real-time PCR. (B and C) Liver TG levels were analyzed by TG extraction assay (B) and Oil Red O staining (C). (D) Livers from WT or L-ERα KO mice treated with the indicated ligands were analyzed by ChIP assays using the indicated Abs. Immunoprecipitated DNA was assessed by real-time PCR using primers specific for LXRE of the SREBP-1c promoter. Samples were normalized to the input DNA. Data represent mean ± SEM (n = 3-11). *P < 0.05; **P < 0.01; n.s., not significant. (E) Schematic model for ER dependent regulation of LXR transcription.

Discussion

E2 and ERs are well-known regulators of several aspects of metabolism, including glucose and lipid metabolism; impaired E2 signaling is associated with development of metabolic diseases.[6] E2 is known to regulate adipose tissue development and also affects other metabolic tissues, including liver, pancreatic β cells, and skeletal muscle. Long-term E2 administration has been reported to lower adipokine expression in adipose tissue.[7, 9] Furthermore, E2 reduces expression of some lipogenic genes in adipose tissue, liver, pancreatic β cells, and skeletal muscle.[7, 9, 30, 31] It appears to promote metabolism of free fatty acids toward oxidation and away from TG storage in pancreatic β cells and skeletal muscle.[7, 31] Earlier reports suggested molecular mechanisms for effects of E2 on lipogenesis. E2 treatment decreases expression of LXRα in adipose, liver, and pancreatic β cells,[7, 30, 31] thereby decreasing lipogenic gene expression and TG accumulation. In contrast, it has also been reported that E2 reduces lipogenic gene expression without a decrease in LXRα expression in adipose tissue, liver, and skeletal muscle.[7, 9] Thus, it remains unclear whether a decrease in LXRα expression in adipose tissue and liver is necessary for reduction of lipogenic genes by E2. In pancreatic β cells and skeletal muscle, it was reported that E2 activates adenosine-monophosphate–activated protein kinase phosphorylation, which suppresses SREBP-1c expression and activates β-oxidation.[7, 31]

This study investigated the effects of E2 on LXR-dependent lipogenesis in the liver. LXR leads to TG accumulation in the liver as a result of increased lipogenic gene expression.[3, 4] We showed that E2 reduces LXR-dependent lipogenic gene expression and TG accumulation in the liver. Also, in this study, we investigated the molecular mechanisms underlying the reduction of LXR-dependent lipogenic gene expression in the liver by E2. First, we showed that LXR expression was not decreased by E2 treatment, suggesting that E2 reduces LXR-dependent lipogenic gene expression without changing LXRα expression in the liver. Next, we examined ERα occupancy on the SREBP-1c promoter in mouse liver and showed that ERα is recruited to the LXRE in the SREBP-1c promoter in complex with LXRα/β and RXRα. ERα directly interacts with LXRα/β, and this interaction is necessary for ERα recruitment to the SREBP-1c LXRE. Finally, we demonstrated that an LXR synthetic ligand enhances the association of the coactivator complex comprising p300 and LXR, and that recruitment of the coactivator complex to LXR is abrogated by E2 (Fig. 7E).

ERs regulate transcription of alternative response elements to which they are recruited by protein-protein interactions; this is defined as the nonclassical pathway.[19, 20] To date, several transcription factors, such as Fos, Jun, and Sp1, have been identified as ER-interacting proteins.[17, 32] Our data revealed that E2 can suppress LXR-dependent transcription through ERα-LXR interactions (nonclassical pathway). In the nonclassical pathway, ERs regulate coactivator recruitment to target transcription factor binding sites by E2.[17, 32] Consistent with these reports, our data revealed that p300 recruitment to the SREBP-1c LXRE is regulated by E2.

MetS is increasing worldwide by increased energy intake and reduced physical activity. Fatty liver, caused by TG accumulation in the liver, has been recognized as a condition possibly involved in the pathogenesis of metabolic diseases.[1, 2] Previous studies have indicated that an HFD induces LXR transactivation, which causes accumulation of TG in the liver.[28, 29] We investigated whether ER affects LXR transactivation and TG accumulation induced by HFD in the liver through the molecular mechanism described above. We demonstrated that E2 suppresses expression of HFD-dependent LXR target genes and TG accumulation. Furthermore, we showed that recruitment of p300 to the SREBP-1c promoter was increased by HFD, and this effect was abrogated by E2-bound ERα. Accordingly, we suggest that increased LXR transactivation and TG accumulation by excess energy intake were reduced by E2 through the ER/LXR pathway (Fig. 7E).

The onset of menopause in women markedly increases their risk of developing pathologies associated with MetS, such as obesity, cardiovascular disease, and type 2 diabetes.[33] These risks are reduced by hormone replacement therapy, demonstrating the importance of functional E2 signaling in metabolic tissues. In contrast, E2 leads to side effects, such as an increased risk for breast cancer,[34] through ERα-dependent transcription (classical pathway). In this study, we showed that both E2 and the phytoestrogen, phloretin, inhibit LXR transcriptional activity and reduce TG accumulation, but E2 potentiates ERα-dependent transcription and phloretin does not markedly affect ERα transcriptional activities. These results suggest that regulation of nonclassical LXR transcriptional activity and classical ERα-dependent transcription can be dissociated by some chemical compounds. Such compounds may be useful in effective therapies for MetS with reduced side effects.

Acknowledgment

The authors are grateful to Dr. P. Chambon (IGBMC) for providing ERαflox/flox mice. The authors thank Dr. T. Matsuzaka, Dr. M. Hamada, K. Soma, N. Iwasaki, Dr. H. Hiyoshi (University of Tsukuba), and Y. Imai (University of Tokyo) for their technical support and helpful suggestions. The authors also thank Enago (www.enago.jp) for the English-language review.

Ancillary