| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 280, Issue 30, 27533-27543, July 29, 2005
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
||
From the
GlaxoSmithKline, 25 Avenue du Québec, 91951 Les Ulis cedex, France, ¶AMC Liver Center, Meibergdreef 69-71, 1105 BK Amsterdam, The Netherlands, and the Dyslipoproteinemia and Atherosclerosis Research Unit (U551), National Institute for Health and Medical Research, Hôpital de la Pitié, Paris Cedex 13, France
Received for publication, March 22, 2005 , and in revised form, May 31, 2005.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
(TR) ligand increase APOA5 mRNA and protein levels in hepatocytes. Our data revealed that T3-activated TR directly regulates APOA5 promoter through a functional direct repeat separated by four nucleotides (DR4). Interestingly, we show that upstream stimulatory factor 1, a transcription factor associated with familial combined hyperlipidemia and elevated triglyceride levels in humans, and upstream stimulatory factor 2 cooperate with TR, resulting in a synergistic activation of APOA5 promoter in a ligand-dependent manner via an adjacent E-box motif. In rats, we observed that apoAV levels declines with thyroid hormone depletion but returned to normal levels upon T3 administration. In addition, treatments with a TR-selective agonist increased apoAV and diminished triglyceride levels. The identification of APOA5 as a T3 target gene provides a new potential mechanism whereby thyroid hormones can influence triglyceride homeostasis. Additionally, these data suggest that TR may be a potential pharmacological target for the treatment of hypertriglyceridemia. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
30% of the United States adult population exhibits hypertriglyceridemia (1), an independent risk factor for coronary artery disease (2) and a key feature of the highly prevalent metabolic syndrome (1, 3). Therefore, understanding the factors that regulate triglyceride (TG)1 levels is of major interest and may provide new opportunities for therapeutic intervention in atherogenic dyslipidemia.
Elevation of plasma TG is associated with hypothyroidism (4–8). Indeed, hypertriglyceridemia is clearly associated with hypothyroidism in obese patients, who are characterized by attenuated rates of clearance of very low density lipoprotein (VLDL)-TG, relative to those in obese euthyroid subjects (6). Such elevation in TG levels has been attributed to low lipoprotein lipase (LPL) (9) or low hepatic triglyceride lipase activities (6, 10, 11). In contrast, hyperthyroid patients exhibit elevated rates of clearance of VLDL and normal or decreased circulating TG levels (6), whereas treatment with thyroid hormones (TH) is associated with elevation in both LPL and hepatic triglyceride lipase activities (9, 10, 11) and concomitantly with a tendency to TG lowering (6, 9, 11). At present, the molecular mechanisms by which TH may regulate lipase activities and circulating TG levels in humans remain to be defined.
3,5,3'-Triiodo-L-thyronine (T3) exerts its biological actions through binding to specific nuclear receptors that modulate gene expression (12). Typically, the T3 receptor (TR) complexes with the retinoid X receptor (RXR) to form a heterodimer that binds to specific DNA sequence elements known as TREs, composed of two half-core PuGGTCA motifs with specific nucleotide spacing and orientation (13). There are several TRs, which are encoded by two distinct genes, TR
(NR1A1) and TR (NR1A2). The TR gene gives rise to the ligand-binding protein TR1 and splice variants that do not bind T3. Several amino-terminal protein variants are produced from the TR gene: TR1; TR2, which is largely restricted to anterior pituitary and hypothalamus (12); and the two recently identified, low level expressed, TR3 and TR
3 (14).
The widely expressed isoforms TR1 and TR1 have divergent N-terminal regions but display remarkably sequence homology throughout the rest, especially in their DNA binding domains (12, 15). As a consequence, both isoforms can bind T3 with similar affinities, and heterodimers of RXR and either TR isoform recognize the same motifs on DNA (12). Despite their structural similarities, distinct patterns of expression of TRs may account for isoform-specific phenotypic functional differences observed in in vivo investigations (12). In the liver, TR1 is the predominant receptor isoform, representing 80% of T3-binding activity (16, 17). T3 influences lipid metabolism through the hepatic regulation of some key TRE-bearing genes, including the lipogenic fatty acid synthase and malic enzyme (18, 19), the mitochondrial fatty acid oxidation rate-controlling enzyme carnitine palmitoyltransferase-I (CPT-I) (20, 21), the rate-limiting enzyme in bile acid synthesis CYP7A1 (22), and the sterol regulatory element-binding protein-2, which in turn activates the low density lipoprotein receptor (LDLr) and other genes directly involved in cholesterol homeostasis (23). The administration of thyroid hormones lowers plasma cholesterol in hypothyroid patients. Unfortunately, the natural TH cannot be used therapeutically to treat hypercholesterolemia in euthyroid individuals because they have undesirable effects in the heart, where TR1 is the predominant isoform (12). These differences in TR isoform expression, together with the fact that TR1 and TR1 isoforms can bind TH analogs with subtle differences in affinity, have spawned attempts to develop thyromimetics with higher selectivity toward TR1 versus TR1 that may have cholesterol-lowering effects but minimal cardiac toxicity (12).
Apolipoproteins (apo) play a determinant role in lipid homeostasis. Alterations in the levels of these functionally specialized proteins dramatically influence plasma lipid concentrations. Recently, the gene coding for a new apolipoprotein family member, APOA5, was identified 30 kb proximal to the APOA1/C3/A4 gene cluster and shown to be a major determinant of plasma TG levels (24, 25). Mice expressing a human APOA5 transgene (24) or injected with adenoviral vectors overexpressing mouse APOA5 (26) exhibit reduction in plasma TG concentrations to one-third the levels of control mice. Conversely, in knock-out mice lacking APOA5, plasma TG levels were 4-fold elevated compared with their wild-type littermates (24). In humans, common polymorphisms across the APOA5 locus have been associated with elevated plasma TG concentrations (24, 27), familial combined hypertriglyceridemia (28), and increased risk of cardiovascular disease (29, 30). Furthermore, inherited apoAV deficiency is associated with severe hypertriglyceridemia in humans (31). ApoAV is a highly hydrophobic protein mainly expressed in liver that circulates at low concentrations associated with high density lipoprotein (24, 25); in addition, apoAV appears to reduce plasma TG by inhibiting hepatic VLDL-TG production and stimulating LPL-mediated VLDL-TG hydrolysis (32).
Inasmuch as apoAV concentration is a key determinant of plasma TG levels and since the apolipoprotein gene family is highly regulated at the transcriptional level (33), recent research has been oriented to the identification of factors that control APOA5 expression. Little is known regarding the transcriptional regulation of this newly discovered gene, although it has been shown to date that APOA5 is regulated by peroxisome proliferator-activated receptor- (NR1C1), farnesoid X-activated receptor (NR1H4), and sterol regulatory element-binding protein-1c, all of which are transcription factors directly implicated in triglyceride metabolism (34–36).
Since T3 influences VLDL-TG metabolism, we investigated whether T3 might regulate APOA5 expression. In this study, we provide evidence that T3 induces APOA5 expression. Furthermore, our findings reveal that TR directly regulates APOA5 in a ligand-dependent manner via a functional TRE within the promoter. In addition, in rats in vivo, apoAV content correlated with thyroid status, and moreover a TR ligand increased apoAV and simultaneously diminished TG levels. Therefore, our findings provide evidence for molecular crosstalk between thyroid status and intravascular TG metabolism.
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
T/–108A T/–101G A/–100T A and to nt –80A C/–77T C of the human APOA5 promoter. The vector pGL3-TK contains a fragment corresponding to nt –109 to +20 of the thymidine kinase (TK) gene promoter of herpes simplex virus (36) subcloned into the BglII/HindIII sites of pGL3-basic vector. The reporter plasmids p(AVDR4)n-TK (n = 1–5) were generated by insertion of 1–5 copies of a double-stranded oligonucleotide containing wild type (5'-GAT CCT GGG AGG CAG CTG AGG TCA ACT TA-3') or mutant (5'-GAT CCT GGG AAG TTG CTG AGA ACA ACT TA-3') sequences spanning nt –117 to –94 of human apoAV promoter into the BglII site of pGL3-TK. Plasmids expressing human cDNAs for TR1 and TR1 were provided by J. A. Holt (GlaxoSmithKline, Research Triangle Park, NC). Plasmid DNA was prepared using the Qiagen endotoxin-free maxipreparation method and quantified spectrophotometrically. The integrities of all plasmids were verified by DNA sequencing.
Cell Transfection and Reporter Assays—Human hepatoblastoma HepG2 cells were cultured in Eagle's basal medium supplemented with nonessential amino acids, 2 mM L-glutamine, 100 units/ml penicillin, 100 µg/ml streptomycin sulfate (medium A), and 10% (v/v) fetal calf serum. On day 0, cells were seeded on 24-well plates at a density of 3.5 x 105 cells/well. On day 1, cells were transfected with FuGENE 6 reagent (Roche Applied Science) according to the manufacturer's instructions. Typically, each well of a 24-well plate received 200 ng of firefly luciferase reporter plasmid and 100 ng of a plasmid expressing human TR1, TR1, and/or USF. Effector plasmid dosage was kept constant by the addition of appropriate amounts of the empty expression vector pSG5. 100 ng/well of a sea pansy (Renilla reniformis) luciferase plasmid pRL-null (Promega) was included in all transfections as an internal control for transfection efficiency. On day 2, cells were switched to Dulbecco's modified Eagle's medium/F-12 medium (Invitrogen) supplemented with nonessential amino acids, 2 mM L-glutamine, 100 units/ml penicillin, and 100 µg/ml streptomycin sulfate (medium B) and 1% (v/v) delipidated calf serum (Sigma) and containing, when indicated, T3 (Sigma), N-[3,5-dimethyl-4-(4'-hydroxy-3'-isopropylphenoxy)-phenyl]-oxamic acid (CGS-23425) (Novartis), or vehicle (water or Me2SO, respectively). On day 3, cell lysates were prepared by shaking the cells in 200 µl of 1x Promega lysis buffer for 15 min at room temperature. Firefly and Renilla luciferase activities were measured using a Dual-Luciferase® reporter assay system (Promega) and a Lumistar luminometer (BMG Lab Technologies). Firefly luciferase activity values were divided by Renilla luciferase activity values to obtain normalized luciferase activities. To facilitate comparisons within a given experiment, activity data were presented either as relative luciferase activities or as -fold induction over the normalized activity of the reporter plasmid in the absence of nuclear receptor cotransfection and agonist supplementation. The data are expressed as the means ± S.D. Statistic significance analyses were done with Student's t test.
Cell Treatments—On day 0, human hepatoblastoma HepG2 cells or rat hepatoma McArdle-RH7777 (ATCC, Manassas, VA) were plated on 24-well plates at 5 x 105 or 105 cells/well, respectively, in medium A and 10% (v/v) fetal calf serum or in Dulbecco's modified Eagle's medium supplemented with 4.5 g/liter glucose, 2 mM L-glutamine, 100 units/ml penicillin, 100 µg/ml streptomycin sulfate, 10% (v/v) fetal bovine serum, and 10% (v/v) horse serum, respectively. On day 2, cells were refed with medium B supplemented with 1% (v/v) delipidated calf serum (Sigma), and T3 (Sigma), CGS-23425 (Novartis) or vehicle (water or Me2SO, respectively). On day 3, cells were washed twice with PBS and harvested for isolation of RNA or Western analysis. Human primary hepatocytes in 24-well plates fed with Williams E medium supplemented with 100 nM dexamethasone, 100 units/ml penicillin, 100 µg/ml streptomycin sulfate, 4 µg/ml insulin (medium C), and 1% (v/v) fetal calf serum were provided on day 0 by Biopredict (batch Hep220069 MW24). On day 1, cells were refed with medium C supplemented with 1% (v/v) delipidated calf serum and containing 50 nM T3, 10 nM CGS-23425, or vehicle (water or Me2SO, respectively). After 24 h, the medium was replaced by 500 µl/well fresh medium C containing 50 nM T3, 10 nM CGS-23425, or vehicle (water or Me2SO, respectively). After 6 h, media were collected, and the cells were washed twice with PBS and harvested for isolation of RNA or Western analysis.
Western Blot Analysis—For determination of secreted apoAV, media from triplicate wells were pooled, and proteins were precipitated by the addition of 180 µl of trichloroacetic acid (Sigma) and agitation overnight at 4 °C. After centrifugation at 2 x 104 g at 4 °C for 15 min, protein pellets were washed twice in 500 µl of cold acetone, dried at room temperature, and resuspended in 100 µl of 1x NuPage LDS sample buffer (Invitrogen). For determination of cellular apoAV, whole cell lysates were prepared by shaking the cells on 24-well plates in 100 µl/well lysis buffer A (PBS, 1% Triton, 50 mM NaF, 5 mM sodium pyrophosphate, 10 µl/ml protease inhibitor mixture from Sigma) for 30 min at 4 °C. The lysates were clarified by centrifugation at 104 x g at 4 °C for 5 min.
60-µl aliquots of resuspended samples from the incubation media and 40 µg of proteins from whole cell lysates were boiled at 100 °C for 5 min, electrophoresed on 10% polyacrylamide MOPS NuPAGE® Novex gels (Invitrogen), and transferred onto nitrocellulose membranes in NuPAGE® transfer buffer (Invitrogen). Membranes were preincubated for 1 h at room temperature in blocking buffer, 5% nonfat dry milk in PBST (10 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.1% (v/v) Tween 20). Subsequently, blots were incubated overnight at 4 °C with rabbit anti-human apoAV, rabbit anti-rat apoAV (25), or mouse anti--actin (A5441; Sigma) in blocking buffer. After washing five times in PBST for 5 min, blots were incubated with IRDyeTM infrared dye 38 (800 nm; Rockland Immunochemicals) and Alexa Fluor® 680 (700 nm; Molecular Probes, Inc., Eugene, OR) labeled goat anti-mouse and anti-rabbit secondary antibodies diluted at 1:5000 in PBST for 1 h at room temperature. Signals were detected by using OdysseyTM Infrared Imaging System (LI-COR).
Real-time PCR Quantification of mRNAs—Total RNA was prepared from human primary hepatocytes, HepG2, and rat McArdle cells with the RNeasyTM Mini kit, the QIAshredderTM homogenizers, and the RNase-free DNase set (Qiagen) according to the manufacturer's instructions. Human normal liver total RNA was purchased from BD Biosciences (reference number 64099-1, batch 4120140; Caucasian 51-year-old man, sudden death). A 1-µg aliquot was used as a template for cDNA synthesis employing the TaqManTM Reverse Transcription Reagent kit (Applied Biosystems). Primers were designed with Primer Express Software (PerkinElmer Life Sciences). The sequences of forward and reverse primers used for the amplifications are as follows: 18 S, 5'-GGG AGC CTG AGA AAC GGC-3' and 5'-GGG TCG GGA GTG GGT AAT TT-3'; hTR1, 5'-CAG CCG CTT CCT CCA CAT-3' and 5'-CCG CCT GAG GCT TTA GAC TT-3'; hTR1, 5'-CTG CAC ATG AAG GTG GAA TG-3' and 5'-TCG AAC ACT TCC AGG AAC AA-3'; cyclophilin, 5'-CAT CTG CAC TGC CAA GAC TGA-3' and 5'-CCA CAA TAT TCA TGC CTT CTT TCA-3'; hAPOC3, 5'-CTT CTC AGC TTC ATG CAG GGT TA-3' and 5'-ACG CTG CTC AGT GCA TCC TT-3'; hLDLr, 5'-GTT GCT GGC AGA GGA AAT GAG AAG-3' and 5'-CAA AGG AAG ACG AGG AGC ACG AT-3'; hAPOA5, 5'-AGC TGG TGG GCT GGA ATT T-3' and 5'-GGC CAC CTG CTC CAT CAG-3'; and rAPOA5, 5'-ACA CGG TCG AGC TGA TGG-3' and 5'-GGC CTT GGT GCC TTT TCC-3', respectively. The specificity and efficiency of the primers were validated as previously described (35). The reactions contained 4 µl of diluted (1:10) cDNA, a 300 nM concentration of the forward and reverse primers, and 2x SYBRTM Green PCR Master Mix (Applied Biosystems) in a final volume of 20 µl. Real-time PCRs were carried out in 384-well plates by using the ABI PRISMTM 7900 sequence detection system (Applied Biosystems). Levels of TR1 and TR1 were normalized by 18 S to compensate for variations in input RNA amounts. Levels of APOA5, APOC3, and LDLr were normalized to cyclophilin to compensate for variations in input RNA amounts (cyclophilin levels were unaffected by the treatments). The amounts of mRNAs were calculated using the comparative CT method as described in Ref. 38. All assays were performed in triplicate during two independent experiments, and the reverse transcriptase (RT)-PCRs were carried out in duplicate for each sample.
In Vitro Transcription/Translation and EMSAs—TR, RXR, USF1, and USF2 proteins were synthesized in vitro from the expression plasmid by using TNT® Quick Coupled transcription/translation system (Promega) according to the instructions of the manufacturer. In order to obtain an unprogrammed lysate as a negative control for EMSA, a reaction was performed with the empty vector pSG5. Double-stranded oligonucleotides corresponding to the sequence spanning nt –117 to –94 and nt –88 to –69 of human APOA5 promoter (AVDR4 and AVEbox, respectively) or modified versions harboring mutations in the DR4 hexamers and E-box described under "Plasmids" (mutAVDR4 and mutAVE, respectively) were radiolabeled by fill in with the Klenow fragment of DNA polymerase I and used as probes. The control probes contain the DR4 sequence of the human TR proximal TR response element (39) and the E-box of the rat CPT-I first intron (40), respectively. Protein-DNA binding assays and electrophoreses of samples were performed as described (41). Gels were dried and analyzed using a PhosphorImager STORM 860 and ImageQuant software (Amersham Biosciences).
Hypothyroid Rats—All experimental protocols were performed in accordance with the policies of the institutional Animal Care and Use Committee. 12-week-old male Wistar rats (Charles River France) were housed under controlled conditions (22 °C, 12 h/12 h dark/light cycle) with food and water ad libitum. Twelve rats were divided into a control and an experimental group. The experimental group was rendered hypothyroid by gavage with a daily single dose of 10 mg/kg 6-n-propyl-2-thiouracil (PTU) (Sigma) aqueous solution for 3 weeks. Seventeen days after the initiation of the PTU treatment, the experimental group received an intraperitoneal injection of saline (PTU, n = 4) or 300 µg/kg T3 (PTU + T3, n = 4) daily for 4 days. Control rats (n = 4) received daily gavage and intraperitoneal injection of only the solvents for the same periods of time. Six hours after the final administration, the animals were killed, and livers were excised. Liver sections of 300 mg were placed in 4 ml of lysis buffer A and homogenized for 20–40 s using a conventional rotor-stator. The lysates were clarified by centrifugation at 104 x g at 4 °C for 5 min. 40 µg of protein were used for Western analysis.
CGS-23425 Treatment of Fat-fed Rats—Six-week-old male Sprague-Dawley rats (Charles River France) were maintained on a chow diet supplemented with 1.5% cholesterol and 0.5% cholic acid (fat-fed) for 14 days before experiments. Fat-fed rats were treated orally by gavage with a daily single dose of 100 µg/kg/day CGS-23425 (Novartis) or vehicle (0.5% hydroxypropyl methylcellulose, 1% Tween 80) for 1 week. Four hours after the last dose, the animals were killed, and blood and livers were collected. Lysates from liver sections (300 mg) were prepared as described above. Total plasma triglyceride levels were measured enzymatically using the TG-PAP150 kit (Biomerieux, France).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
1 and TR1 isoforms are expressed in HepG2 and human primary hepatocytes, suggesting that both isoforms may be readily available for T3-dependent regulation studies. More specifically, TR1 mRNA levels in HepG2 are 4-fold higher than in human primary hepatocytes and 5-fold higher than in human liver (2–Ct x 104 values were 7.00 ± 0.53, 1.84 ± 0.35, and 1.37 ± 0.06, respectively), whereas TR1 levels are similar in the three cases (2.33 ± 0.12 for HepG2, 2.08 ± 0.36 for human primary hepatocytes, and 2.91 ± 0.05 for human liver). To determine whether T3 can modulate APOA5 gene expression, we incubated HepG2 cells in the presence or absence of T3. Treatment with T3 significantly increased APOA5 mRNA levels at 6 h, and a 2-fold induction was achieved after 24 h of T3 addition (Fig. 1A). Furthermore, increasing concentrations of T3 resulted in a dose-dependent induction of APOA5 expression (Fig. 1B). In addition, we aimed to know whether APOA5 expression might be increased by TH analogs with potential therapeutic interest (12). CGS-23425 is a synthetic thyromimetic with cholesterol-lowering effects but minimal cardiac toxicity in rats that has been reported to be more selective toward TR1 versus TR1 and to show a higher binding affinity to intact hepatic nucleic than T3 (42). As shown in Fig. 1B, CGS-23425 increased APOA5 mRNA levels, attaining a 2.2-fold increase at 5 nM.
Similarly, human primary hepatocytes and rat hepatoma McArdle cells were incubated for 24 h in medium containing T3, CGS-23425, or vehicle. As shown in Fig. 1, C–F, treatment with T3 or CGS-23425 increased APOA5 mRNA levels. These effects were specific, since the expressions of LDLr (Fig. 1, C and D) and CPT-I (Fig. 1, E and F) were also increased, in accordance with previous studies (40, 43), whereas cyclophilin, used as internal control, remained unaffected by T3 or CGS-23425 treatments, and in addition, no significant effect on APOC3 mRNA levels was observed (Fig. 1, C and D).
T3 and the Thyromimetic CGS-23425 Increase ApoAV Protein Levels in Human Hepatocytes—Western blot analyses performed on whole cell lysates from HepG2 (data not shown) and human primary hepatocytes (Fig. 2) incubated for 24 h with T3, CGS-23425, or vehicle revealed that the quantity of cellular apoAV protein was markedly increased by both TR ligands (Fig. 2). Moreover, treatments with T3 or the thyromimetic CGS-23425 led to a significant increase in levels of apoAV protein secreted into the medium by human primary hepatocytes (Fig. 2).
T3 and CGS-23425 Increase Human APOA5 Expression at the Transcriptional Level via the Nuclear Receptor TRs—To determine whether APOA5 was directly responsive to T3-activated TRs, we performed functional analysis of the human APOA5 promoter. In transient transfection assays in HepG2 cells, treatment with 50 nM T3 increased the activity of the firefly luciferase reporter gene driven by the –617/+18 sequence of the human APOA5 promoter (Fig. 3A). The effect of T3 was promoter-dependent, because it was not observed with the promoterless pGL3-basic vector. Cotransfection of human TR1 or TR1 expression plasmids robustly enhanced T3-induced promoter activity. In contrast, APOA5 promoter activity was not affected significantly by cotransfection of TR1 or TR1 in the absence of T3 (Fig. 3A). No statistically significant differences on T3 induction of APOA5 promoter were observed between TR1 and TR1 transactivations. In order to confirm that the lipid-lowering thyromimetic CGS-23425 induces the APOA5 gene at the transcriptional level, similar transient transfection assays were performed in HepG2 cells with the human APOA5 promoter along with a human TR1 expression plasmid. As shown in Fig. 3B, increasing concentrations of CGS-23425 resulted in a dose-dependent induction of the luciferase activity. The promoterless pGL3-basic vector was unaffected by CGS-23425 (data not shown).
|
|
|
1 expression plasmid in the presence or the absence of 50 nM T3. As shown in Fig. 4A, the sequence upstream to position –242 could be removed without preventing strong activation of the reporter gene by T3-activated TR. In contrast, deletion of the fragment between nt –242 and –82 completely abolished the induction of APOA5 promoter activity by T3-activated TR, indicating that this region mediates the effects of T3. Analysis of the sequence in the –242/–82 fragment revealed a direct repeat of the hexanucleotide core motif PuGGTCA with a low degree of degeneration separated by 4 nucleotides between nt –113 and –98 (Fig. 4B), thereby conforming to the DR4 response element for TR (TRE).
|
1 expression vector and an APOA5 promoter-luciferase reporter plasmid in which the DR4 sequence was mutated (Fig. 4B). In contrast to the wild-type promoter construct, T3 and T3-activated TR1 failed to induce the activity of the construct bearing the mutated DR4 (Fig. 4C).
The RXR-TR Heterodimer Binds Specifically to the APOA5 DR4 Element—Direct binding of RXR-TR heterodimers to the APOA5 DR4 element was examined. For this purpose, gel shift assays were performed using in vitro translated human RXR and TR and radiolabeled double-stranded oligonucleotides containing the wild-type or a mutated version of the APOA5 DR4 element. The addition of TR1 resulted in the appearance of a weak protein-DNA complex band (Fig. 5, lane 7). This phenomenon most likely corresponds to the binding of TR1 monomers and is also perceived with a control probe containing the DR4 sequence of the human TR proximal TRE (Fig. 5, lane 3) (34). However, when both RXR and TR1 were present, this faint band disappeared, and a robust and more shifted band emerged (Fig. 5, lane 8). In contrast, a labeled probe that is equivalent to APOA5 DR4 but harbors point mutations in the half-sites could not form the faster mobility band with TR1or the strong RXR-TR binding complex (Fig. 5, lanes 11 and 12). Furthermore, the specificity of the RXR-TR1-APOA5 DR4 interaction was confirmed by competition analysis using an excess of cold double-stranded oligonucleotides corresponding to the control DR4 probe, which inhibited the retarded complex, and the mutated APOA5 DR4, which failed to displace the labeled wild-type element (data not shown). Additional EMSAs showed that no TR1 monomers or RXR-TR1 heterodimers could bind when the downstream half-site was mutated, whereas mutation of the upstream hexamer markedly diminished heterodimer binding but enhanced the binding of TR monomers (data not shown), thereby suggesting that RXR binds to the upstream and TR1 to the downstream hexamer, respectively, of APOA5 DR4, as a classical direct repeat TRE (44).
|
1 expression plasmid in the presence or the absence of T3. As demonstrated in Fig. 6, T3-activated TR1 enhanced the activity of APOA5 DR4-driven promoter constructs, whereas the reporter constructs with the TK promoter alone or driven by several copies of mutated APOA5 DR4 were not stimulated. Indeed, an 80-fold induction was attained with five copies, and the response was manifested in a copy number-dependent manner. In addition, we observed that the APOA5 DR4 site conferred 2-fold more T3-activated TR1 responsiveness to the TK promoter than the CPT-I DR4 TRE (data not shown). Taken together, these results show that this APOA5 DR4 motif is a genuine TRE. The USF Transcription Factors Cooperate with TR in the T3-mediated Induction of APOA5—As shown in Fig. 7A, there is a nearby sequence 5'-CACGTG-3' downstream of this DR4 element that constitutes a canonical E-box motif, a binding site for basic helix-loop-helix/leucine zipper proteins. Recently, it was reported that this E-box might make a minor contribution to the sterol regulatory element-binding protein-1c-mediated down-regulation of APOA5 (36). Inasmuch as USF transcription factors appear to be the predominant basic helix-loop-helix/leucine zipper proteins in liver nuclear extracts (45) and because they can physically interact with TR (40), we set out to investigate whether USF might be involved in the T3 induction of APOA5 through this E-box.
|
C/–77T C point mutations (Fig. 7B, lane 4). Furthermore, competition analysis showed that the retarded complex was inhibited by an excess of an unlabeled control probe (Fig. 7B, lanes 8 –10), containing the E-box of CPT-I (40), but not by the mutated APOA5 E-box probe (data not shown).
Transient transfection assays in HepG2 cells revealed that USF activates at least 2-fold the luciferase reporter gene expression vector driven by the –617/+18 sequence of the human APOA5 promoter (Fig. 7C). In the absence of T3, cotransfection of TR1 produced no effect. However, in the presence of T3, TR and USF synergistically activate APOA5, attaining a 20-fold induction (Fig. 7C). Furthermore, as shown in Fig. 8A, an internal deletion USF mutant lacking the basic region required for DNA binding (UB), which forms dimers and sequesters both endogenous USF1 and USF2 (46), down-regulated APOA5 promoter activity both in the absence and in the presence of T3 and transfected TR1. Likewise, an N-terminal truncated USF mutant (UN), composed only of the basic region and basic helix-loop-helix/leucine zipper domains and which binds DNA and forms dimers but is totally inactive (47), also reduced APOA5 promoter activity. Therefore, both the DNA binding and the transcriptional activation domains of USF are required for full synergistic activation with T3-activated TR1. Nevertheless, it is worth noting that, in the presence of both TR and UN, the magnitude of -fold induction by T3 was similar to that achieved in the presence of TR and the wild-type USF (Fig. 8A, compare data of T3/Cont). These findings suggest that despite the fact that this USF mutant lacking the N-terminal activation domain represses APOA5, probably due to competition with the endogenous active USF (47), the enhanced response to T3 is somehow linked to the ability of USF to bind to DNA.
|
1 and USF, thereby confirming that the binding of USF to DNA is required for the action of USF on T3 induction (Fig. 8B).
ApoAV Content Correlates with Thyroid Status in Rats—In order to extend our analyses to animal models, we chemically induced hypothyroidism in rats by the well described treatment with PTU, which inhibits the 5'-deiodinase enzyme required to convert the 3,5,3',5'-tetraiodo-L-thyronine (T4) form of TH into the more bioactive T3 isoform (48). Under these conditions, plasma T4 levels fall significantly by 75–90%, and there is also an 60% reduction of plasma T3 (49–51), whereas plasma thyroid-stimulating hormone concentrations rise (52). As shown in Fig. 9, apoAV protein levels were dramatically diminished in the hypothyroid rats, whereas actin levels remained unaltered. However, administration of T3 restored apoAV protein abundance (Fig. 9). Together, these data demonstrated that apoAV strongly correlates with thyroid status.
The Thyromimetic CGS-23425 Increases ApoAV Protein Levels in Rats—The TR-selective agonist CGS-23425 has been found to exert lipid lowering actions in fat-fed rats (42). Thus, we examined the effects of this thyromimetic on hepatic apoAV levels in fat-fed rats treated with a daily single dose of the drug for 1 week. As expected, CGS-23425 increased the levels of apoAV protein, whereas actin content was unaffected (Fig. 10A). Furthermore, we observed that this TR-selective agonist induced a dramatic decrease both in VLDL-TG (44% of controls, data not shown) and in total plasma TG concentrations (Fig. 10B).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
1 and TR1 isoforms. As demonstrated by EMSA and mutation analyses, activation by T3 may be attributed to a DR4 element located within the proximal APOA5 promoter. Our data from transfection assays using the isolated APOA5 DR4 showed that this TRE is capable of conferring positive T3 responsiveness to a heterologous promoter, further confirming that this DR4 motif corresponds to a genuine TRE. Taken together, our results demonstrate that APOA5 is a direct target of TH.
|
On the other hand, although USF greatly enhances the T3 stimulation of APOA5, several observations suggest that the dependence on USF is not straightforward. First, the APOA5 TRE alone is sufficient to support a robust response to T3 when linked to a basal heterologous promoter. Second, cotransfection of a mutant USF that lacks the DNA-binding domain diminishes the APOA5 promoter activity, probably due to the sequestration of endogenous USF (46), but it produces no significant decrease in the -fold induction by T3 in the presence of TR (5.8 versus 6.0 in Fig. 8A). Furthermore, mutational analyses of the E-box show that the induction by T3-activated TR is similar in both wild-type and mutant constructs. Hence, this E-box is absolutely required for the ligand-dependent synergism between TR and USF, but it appears not to be necessary for the sustenance of T3 response. A similar situation has been described between sterol regulatory element-binding protein and the ubiquitous factor Sp1 in the FAS promoter (54). Whereas sterol regulatory element-binding protein and Sp1 synergistically activate the FAS promoter, mutational analyses revealed that the Sp1 site is dispensable for sterol regulation (54).
While this manuscript was in preparation, Nowak et al. (55) reported that insulin down-regulates APOA5 expression via USF. Interestingly, in agreement with our study, these authors show by chromatin immunoprecipitation assays that the E-box at –81/–76 may bind USF. Even more interestingly, the combination of their report and our study raises the possibility of a cross-talk between insulin and thyroid signals on the same element in APOA5.
Strikingly, USF1 has been recently identified as the gene on human 1q21–23 that is associated with familial combined hyperlipidemia and especially with high triglycerides in men (56). Therefore, additional studies are warranted to address the relevance of USF-mediated regulation of APOA5 and other lipid-related genes in familial combined hyperlipidemia phenotype.
|
|
ligand CGS-23425 in fat-fed rats (64). Our results show that PTU-induced hypothyroidism reduced apoAV content, whereas T3 treatment restored normal levels, thereby indicating that thyroid status strongly correlates with apoAV levels in rats. Consistent with previous reports, we observed that TG levels were reduced in PTU-treated rats (data not shown). This paradox probably has functional implications, since it suggests that a decrease in apoAV does not necessarily impede an overall reduction of TG levels in hypothyroid rats, which is most likely due to the translational increase in adipose and heart LPL mass reported by others (57). On the other hand, we observed that treatment with CGS-23425 at a heart-sparing dose (42) increased apoAV content and decreased plasma TG levels in fat fed rats. This finding is consistent with the TG-lowering effects exhibited by the hepatoselective thyromimetic SK&F L-94901 in euthyroid rats (65) and by the TR-selective agonist GC-1 in hypothyroid mice (66). Although these data warrant further functional exploration, we and others (65, 66) hypothesize that the hypotriglyceridemic effects of these thyromimetics are most likely due to TR subtype-specific and organ-selective causes. Indeed, SK&F L-94901 does not discriminate TR subtype in terms of binding affinity, but it is more effectively transported to the nucleus in hepatic cells (67); GC-1, which has a 10-fold reduced affinity for TR1, distributes primarily to the liver rather than to the heart or the adipose tissue (66); and the TR-selective ligand CGS-23425 is also markedly hepatoselective (42). Since TR1 is the predominant isoform in adipose tissue (68), it is tempting to speculate that the tissue-selective properties of these thyromimetics contribute to a loss of effect on adipose LPL translation, whereas they induce the expression of hepatic genes such as APOA5. Given the two hypotriglyceridemic mechanisms suggested for apoAV (namely reduction in hepatic VLDL-TG secretion rate (32, 69) and elevation in the efficiency of LPL-mediated TG hydrolysis (32, 70)), the increase in apoAV that we observed in rats is consistent with studies showing lower secretion rates of VLDL in hyperthyroid rats (63), accelerated TG clearance in CGS-23425-treated rats (64), and lower plasma TG concentrations in thyromimetic-treated rodents (65, 66) (this study). On the other hand, GC-1 has been reported to suppress thyroid-stimulating hormone by 40% and to reduce plasma T4 levels by 35% at therapeutic doses, but this was not accompanied by a statistically significant reduction in plasma T3 levels in rats (71). Whereas the administered GC-1 would compensate in terms of TR stimulation, it may not appreciably activate TR1, which could therefore cause a relative TR hypothyroidism. It is not known whether CGS-23425 modifies T3, T4, and thyroid-stimulating hormone plasma levels in rats; future work may address the significance of these potential influences. The molecular mechanisms implicated in the TG-lowering effects of TH in humans are currently unclear. Although the proximal promoter of the human APOC2, which encodes an LPL cofactor, could be transactivated by RXR-TR heterodimers (72), there are no data to indicate stimulation of APOC2 expression by TH. Similarly, it has been shown that RXR-TR heterodimers bind with low affinity to the distal regulatory region of the human APOC3 gene (73), an inhibitor of LPL-mediated lipolysis, although regulation of human APOC3 expression by T3 has not been previously reported. Indeed, our results in human hepatocytes suggest that the effect of T3 on APOC3, if any, should be minimal. On the other hand, TH regulates hepatic apoB mRNA editing in neonatal mice (62) and rats (64), an effect that has been linked to increased hepatic TG secretion (62). However, since apobec-1 is not expressed in human liver, these findings may not be extended to humans.
The current results may provide a potential explanation for the low VLDL-TG clearance rates and low LPL activity observed in hypothyroid patients (6, 9) and for the increase in LPL activity observed in TH therapies (9, 11).
In conclusion, our data reveal that treatment with T3 and pharmacological activation of TR up-regulates APOA5 expression, thereby establishing a novel molecular pathway of T3 action. These results underscore a physiological role of TR in the regulation of genes involved in triglyceride metabolism and, therefore, identify TR as a potential pharmacological target for the treatment of hypertriglyceridemia.
|
FOOTNOTES |
|---|
|| To whom correspondence should be addressed. E-mail: jcrodrig{at}freesurf.fr.
1 The abbreviations used are: TG, triglyceride(s); VLDL, very low density lipoprotein; LPL, lipoprotein lipase; TH, thyroid hormone(s); T3, 3,5,3'-triiodo-L-thyronine; TR, thyroid receptor; RXR, retinoid X receptor; CPT-I, carnitine palmitoyltransferase-I; LDLr, low density lipoprotein receptor; DR, direct repeat; apo, apolipoprotein; nt, nucleotide(s); TK, thymidine kinase; RT, reverse transcriptase; TRE, thyroid receptor response element; PTU, 6-n-propyl-2-thiouracil; USF, upstream transcription factor; T4, 3,5,3',5'-tetraiodo-L-thyronine; CGS-23425, N-[3,5-dimethyl-4-(4'-hydroxy-3'-isopropylphenoxy)-phenyl]-oxamic acid; EMSA, electrophoretic mobility shift assay.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
