Tri-iodo-l-thyronine (T3) is essential for maintaining normal cardiac contractile function by regulating transcription of numerous T3-responsive genes. Both hormone availability and relative amounts of nuclear thyroid hormone receptor isoforms (TRα1, TRβ1) determine T3 effectiveness. Cultured neonatal rat ventricular myocytes grown in T3-depleted medium expressed predominantly TRα1 protein, but within 4 h of T3 treatment, TRβ1 protein increased significantly, whereas TRα1 was decreased by 46 ± 5%. Using replication-defective adenoviruses to overexpress TRα1 in cardiomyocytes, we studied the mechanisms by which T3 mediated the decrease in TRα1 protein. Inhibitors of the proteosome pathway resulted in an accumulation of ubiquitylated TRα1 in the nucleus and prevented T3-induced degradation of ubiquitylated TRα1, suggesting that T3 induced proteosome-mediated degradation of TRα1; however, TR ubiquitylation was T3 independent. TRα1 transcriptional activity, measured using transient transfection of a thyroid hormone-responsive element (TRE) reporter plasmid, was T3 dose dependent and inversely proportional to nuclear TRα1 content, with 10 nM T3 having maximum effect. Quantitative RT-PCR showed that both endogenous and adenovirus-expressed TRα1 mRNAs were significantly decreased to 54 ± 11 and 25 ± 5%, respectively, within 4 h of T3 treatment. Measurements of TRα1 mRNA half-life in actinomycin D-treated cardiomyocytes showed that T3 treatment significantly decreased TRα1 mRNA half-life from 4 h to less than 2 h, whereas it had no effect of TRβ1 mRNA half-life. These data support a role for both the proteosome degradation pathway and altered mRNA stability in T3-induced decrease of nuclear TRα1 in the cardiomyocyte and provide novel cellular targets for therapeutic development.
- neonatal rat ventricular myocyte
the importance of thyroid hormone in maintaining homeostasis in the cardiovascular system is underscored by its effects on cardiac contractile function and phenotype that include the regulation of expression of numerous ion channels, contractile proteins, and calcium regulators (8, 16, 17, 19). Heart failure and conditions resulting in cardiac hypertrophy produce a phenotype that closely resembles the hypothyroid cardiac phenotype with impaired diastolic and systolic contractile properties (2, 14, 18, 22, 37, 43). Therefore, the role of thyroid hormones in this clinical setting has received considerable attention (23, 26, 29). Tri-iodo-l-thyronine (T3) actions are primarily mediated by nuclear thyroid hormone receptors (TRs) that regulate transcription of target genes by either repressing or activating T3-responsive promoters, depending on hormone status and the nature of the DNA binding site (3, 44). Therefore, both hormone and receptor play an important role in this process. Recent studies indicate that the half-lives of many transcription factors, including nuclear receptors, are controlled by ubiquitin-dependent proteolysis and that agonist binding attenuates receptor-mediated transcription by targeting the receptors for degradation. Ligand-mediated receptor degradation has been reported for estrogen and progesterone receptors (27), thyroid hormone receptor (6), retinoid X receptor (RXR) (30), and peroxisome proliferator-activated receptor (PPARγ) (11). This mechanism of receptor downregulation is largely conserved among the nuclear receptor superfamily and may be an important mechanism by which receptor signaling can be regulated.
Although two primary TR isoforms TRα1 and TRβ1 are expressed in the heart, studies of transgenic animal models suggest that TRα1 is the predominant isoform regulating cardiac phenotype and function, with TRβ1 potentially playing a redundant role (9, 38). The present study confirms the presence of TRα1 and TRβ1 isoforms in the cardiomyocyte and shows that T3 regulation of the two TR isoforms occurs by distinct mechanisms that may provide a means for expression of genes that are unique to each isoform. Studies using intact heart tissue have failed to document any change in TR protein content in response to thyroid status or cardiac disease, although changes in TR mRNA content have been reported (7, 10, 14, 15, 36). Because cardiomyocytes represent a minority of cells (∼20%) within the heart tissue, and although the cardiac myocytes are primarily binucleated (∼85%), the nuclear receptor proteins from the myocytes may be under represented when whole tissue samples are analyzed (13, 21, 33). To circumvent this shortcoming, we used pure cultures of rat ventricular myocytes to investigate the effects of T3 on TRs and were able to define distinct effects on TRα1 and TRβ1 isoforms not previously reported.
Early studies (32) published before the cloning and identification of various TR isoforms suggested that T3 reduced the amount of its nuclear receptor protein in pituitary (GH1) cells by decreasing its half-life and reducing its rate of synthesis, implying that T3 had effects at the level of mRNA (or rate of mRNA translation) and protein turnover. These data showed that thyroid hormone receptor half-life in GH1 cells was reduced from 4.7 to 3.3 h by T3 treatment. This insightful report is supported by the results of the current study in which T3 was shown to effect TRα1 mRNA half-life and proteosome-mediated TRα1 protein degradation, thus providing multiple mechanisms by which thyroid hormone can regulate cardiomyocyte function.
MATERIALS AND METHODS
Cell culture reagents including Hanks' balanced salt solution (HBSS), DMEM-F12, fetal bovine serum (FBS), and antibiotic solution were obtained from GIBCO-BRL (Grand Island, NY). N-Acetyl-Leu-Leu-Nle-CHO (ALLN), N-Acetyl-Leu-Leu-Met-CHO (ALLM), carbobenzoxy-l-leucyl-l-leucyl-l-leucinal (MG-132), cathepsin L inhibitor IV, and cathespsin B inhibitor III were from Calbiochem (San Diego, CA). Cyclohexamide (CHX), actinomycin D, T3, and reverse T3 (rT3) were from Sigma (St. Louis, MO). Antibody to TRα1 (PAI-211A) was from Affinity BioReagents (Golden, CO), and TRα1/β1-antibody (C1) and TFIID (TATA box binding protein, TBP) were from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-ubiquitin antibody was from Calbiochem. Horseradish peroxidase-conjugated goat anti-mouse and anti-rabbit secondary antibodies were purchased from Bio-Rad Laboratories (Hercules, CA). Radioactive chemicals and chemiluminescent reagents for Western blot analysis were from New England Nuclear Life Science Products (Boston, MA). Real-time quantitative PCR reagents were from Eurogentec (Belgium). All other reagents, including protease and phosphatase inhibitors, were of the highest available quality from Sigma, Calbiochem, or Pierce (Rockford, IL).
Animals used in these experiments were treated in accordance with National Institutes of Health Guide for the Care and Use of Laboratory Animals (DHHS Publication No. 85-23), and study protocols were approved by the Institutional Animal Care and Use Committee. Twelve Sprague-Dawley rats (160–180 g) (Charles River Laboratories, Wilmington, MA) were made hypothyroid by surgical thyroidectomy 3 wk before initiation of the study protocol. Six animals received T3 (5 μg/100 g body wt) as two separate bolus injections given subcutaneously over a 24-h period. Twelve hours after the second injection, the rats were anesthesized, the hearts were removed, and the left ventricles were immediately frozen in liquid nitrogen and stored at −80°C until analyzed for RNA. Hearts were similarly obtained from six thyroidectomized animals that were not treated. Euthyroid Sprague-Dawley rats (200 g body wt) were also used for the preparation of nuclear extracts from livers and hearts (left ventricles) as we have previously published (40).
Isolation and culture of neonatal rat ventricular myocytes.
Neonatal rat ventricular myocytes (NRVM) were isolated from hearts of 2-day-old rats by collagenase digestion as previously described (31). Myocytes were plated at ∼1.5 × 104/cm2 on collagen-coated six-well plates or 60-mm dishes and cultured for the first 20 h in DMEM/Ham's F-12 (GIBCO-BRL) containing l-glutamine, 10% FBS, Ara-C, and antibiotics. Cell cultures were subsequently maintained in serum-free medium consisting of DMEM/F-12 (GIBCO-BRL) containing transferrin (5 mg/l), selenium (5 μg/l), insulin (120 IU/l), l-glutamine, and antibiotics in an incubator equilibrated at 5% CO2, 37°C. Media were changed daily.
Replication-defective adenoviruses containing the full-length coding sequence of rat thyroid hormone receptor TRα1 was constructed by first subcloning the cDNA (kindly provided by Dr. Ron Evans, The Salk Institute, San Diego, CA) into the pShuttle vector with 5′ sequence modification to optimize translation and then subcloning into the adenoviral genome according to the manufacturer (Clontech Laboratories, Palo Alto, CA). Recombinant adenoviruses containing the TR cDNA sequences were amplified by sequential infection of HEK-293 cells and purified by CsCl gradient ultracentrifugation. The viral titer of each preparation of replication-defective adenovirus was determined by dilution assay in HEK-293 cells or by Clontech's Adeno-X Rapid Titer kit. Replication-defective adenovirus encoding nuclear localized β-galactosidase Ad-βgal (kindly provided by Dr. Allen Samarel, Loyola University, Maywood, IL) was used to control for nonspecific effects of adenoviral transduction of cultured cells.
Viral transduction of cultured cardiomyocytes.
After the first 20 h in serum-containing medium, the NRVM cultures were washed with HBSS and exposed to adenovirus in DMEM/F-12 medium for 1 h in the cell culture incubator (5% CO2, 37°C). Subsequently, the cells were washed with HBSS and maintained in serum-free medium for 24 or 40 h before experimentation as indicated in results and figures. Most experimental protocols involved the addition of reagents for less than 8 h and therefore were conducted 40 h after viral transduction when expression of TRα1 had reached a maximum level. Stock solutions of T3 and reverse T3 were freshly prepared in 0.05 M NaOH, and dilutions were made to final concentrations as indicated. Final concentrations in the cell culture medium were as follows for these reagents: 100 μM ALLN, 100 μM ALLM, 10 μM MG-132, 10 nM cathespsin-B and -L inhibitors, 50 μM CHX, and 5 μg/ml actinomycin D.
For protein analysis, cells were harvested, and cytosolic and nuclear fractions were prepared by Dounce homogenization in ice-cold buffer containing 0.3 M sucrose, 10 mM HEPES (pH 7.9), 10 mM KCl, 1.5 mM MgCl2, 0.1 mM EGTA, 0.5 mM DTT, 0.5 mM PMSF, 0.4% Nonidet-40, and protease inhibitor cocktail, including leupeptin, aprotinin, and antipain (each at 5 μg/ml). After centrifugation at 10,000 g for 1 min at 4°C, the supernatants were collected as cytoplasmic fractions, and nuclear extracts were prepared by resuspension of the pellet in high salt buffer containing 420 mM NaCl at 4°C for 30 min. After centrifugation at 10,000 g for 10 min at 4°C, the supernatants were collected and used for nuclear protein analysis.
Transfection of cardiomyocytes.
After 20 h in culture, NRVM cultured in 12-well plates were cotransfected with a plasmid (0.75 μg) containing two thyroid-responsive element (TRE) sequences (DR+4/PAL) inserted upstream of sarcoplasmic reticulum calcium ATPase-1 (SERCA1) minimal promoter (−141 bp) in phRL (Renilla luciferase reporter) (kindly provided by Dr. W. Simonides, VU University Medical Center, Amsterdam, The Netherlands) and pGL3-basic (firefly luciferase reporter) (Promega, Madison, WI). After the 6-h period of transfection (Lipofectin, GIBCO-BRL), the cells were washed twice with HBSS, replenished with fresh medium, and then transduced for 1 h at 37°C with Ad-TRα1 [2 multiplicity of infection (MOI)] or Ad-βgal (5 MOI). After adenovirus exposure, the cells were washed with HBSS and cultured in serum-free medium during the experiment of 48 h. Treatment protocols with T3 for derivation of time- and dose-response curves, as well as treatment protocols with reverse T3, are detailed in the figure legends. Cells (5 × 105) were harvested in 150 μl of lysis buffer according to the manufacturer's protocol, and 20 μl of cell lysate were analyzed for Renilla and firefly luciferase activities by using the dual luciferase reporter assay system (Promega). Results are expressed as normalized luciferase activity, which is derived from Renilla luminescence units normalized for firefly luciferase activity in the same volume of cell lysate. Cell lysates were also resolved by SDS-PAGE and analyzed for expression of TRα1 by Western blot analysis as detailed below.
Real-time quantitative PCR.
Individual RNA samples were prepared from ∼106 cells plated in six-well plates using a commercially available kit (RNeasy Mini Kit; Qiagen, Valencia, CA). RNA from tissue (liver and LV) samples were isolated by the acid phenol-chloroform method as we have previously published (29) followed by treatment with deoxyribonuclease and purification using the RNeasy Qiagen kit. The integrity of the RNA was assessed using the Bioanalyzer (Agilent Technologies, Foster City, CA). All RNA samples were DNase treated before analysis by real-time Q-PCR using TaqMan chemistry and an ABI Prism 7700 sequence detector (PerkinElmer-Applied Biosystems, Foster City, CA). Forward (tccgaagcctgcctgc) and reverse (gatgaggtgtgaggatgtttgc) primers and TaqMan probe (catgtggagaggcgtggtcattg) for TRβ1 annealed between 346 and 426 bp of the cloned cDNA (Genbank J03819). Forward (aggaagtctaagcctcaggcg) and reverse (cagctttgtcccttctctcca) primers and TaqMan probe (cagagggtgtgcggagctggtg) for TRα1 annealed between 1547 and 1622 bp of the cloned cDNA (Genbank M18028). Q-PCR primer sequences for TRα2 were selected between 881 bp and 944 bp (Genbank M31174). Primers and TaqMan probes labeled with TET-TAMRA were synthesized using Applied Biosystems 394 DNA Synthesizer in the Institutional Core Facility. Expression amount of rat GAPDH mRNA was used for normalization of each sample, and analysis of each specific mRNA was conducted in duplicate. Relative expression of the mRNAs was calculated by the “delta delta Ct” method, and results are expressed as fold change with respect to the corresponding experimental control.
Western blot analysis.
Protein concentrations in the nuclear fractions were determined by Micro BCA assay (Pierce). After normalization of sample protein content, nuclear proteins were resolved by electrophoresis on 5% stacking/10% polyacrylamide-SDS gels, and the resolved proteins were transferred to nitrocellulose membrane (Bio-Rad) at 200 mA for 4 h using Towbin's transfer buffer [96 mM glycine, 12.5 mM Tris (pH 8.3), and 10% methanol]. The membrane was then blocked by incubation in 5% nonfat milk-TBST (10 mM Tris·HCl, pH 8, 150 mM NaCl, and 0.05% Tween-20) 30 min at room temperature. Blots were incubated at 37°C for 2 h with antibodies diluted as follows: polyclonal antibodies TRα1 (PAI-211A) and ubiquitin at 1:1,000 dilution in 5% milk-TBST and monoclonal antibodies against TBP (diluted 1:40) and TRα1-TRβ1 C1 (diluted 1:100). After the membrane in TBST was extensively washed, secondary antibodies (either goat anti-rabbit or goat anti-mouse IgG conjugated with horseradish peroxidase at 1:4,000 dilution in 5% milk-TBST) were incubated with the membrane for 1 h at room temperature. After extensive washes, the signal was developed using a chemiluminescence reagent according to the manufacturer’s instructions (Western Lightning Reagent Plus; PerkinElmer) and detected by exposure to X-ray film. The protein bands on the X-ray film were quantified by laser-scanning densitometry (GS-800 Calibrated Densitometer; Bio-Rad), and the band density was calculated by Quantity One 4.2.2 software. Results are expressed as arbitrary densitometric units (du).
Electrophoretic mobility shift assay.
Double-strand DNA probes used in the electrophoretic mobility shift assays (EMSA) included the TREs derived from the cardiac α-myosin heavy chain (MHC) gene (5′-ttggctctggAGGTGAcaggAGGACAgc-3′) and SERCA2 gene (5′-cggAGGCAAgccaAGGACAc-3′) as well as F2 chick ovalbumin gene (5′-ttgACCCCAgctgAGGTCAagt-3′). The sense DNA strand was 5′-end labeled with T4 polynucleotide kinase and [γ-32P]ATP to ∼2 × 106 dpm/pmol, purified over Sephadex G-25 spin columns and then annealed to 50-fold excess of the unlabeled antisense strand. TRα1 and RXRγ (subcloned into pCI vectors) were synthesized in the in vitro rabbit reticulocyte lysate (RL) TNT transcription/translation system (Promega) and were analyzed by EMSA concomitantly with nuclear extracts derived from NRVM transduced with adenovirus-expressed TRα1 (Ad-TRα1) to determine whether the TR bound to TREs as hetero- or homodimers. Binding reactions contained 3 to 5 μg nuclear extract protein or 3 μl RL, 1 μg poly(dI-dC), 12 mM HEPES (pH 7.9), 10 mM KCl, 4 mM Tris·HCl, 0.6 mM DTT, 1 mM MgCl2, 12% glycerol, and 0.6 mM PMSF. Reactions were preincubated for 15 min at room temperature before incubation for 30 min with labeled double-stranded (ds)DNA (20–40 fmoles at 40,000 dpm/reaction) as we have previously published (41). Reaction products were resolved on 6% native PAGE, vacuum dried, and exposed to X-ray film.
All data are presented as means ± SE derived from three to six distinct samples per experiment derived from a minimum of two separate cell preparations. Unpaired Student's t-test was used for statistical analysis between groups, and significance was determined at P < 0.05.
Effect of thyroid hormone on TR isoforms in vivo and in cultured cardiomyocytes.
Immunoblot analysis using an antibody that recognizes both TRα1 and TRβ1 proteins showed that these two TR protein isoforms were expressed in relatively equal amounts in both normal adult rat liver and heart tissue, with perhaps slightly higher amounts of TRα1 than TRβ1 protein in the adult left ventricle (Fig. 1). These data corroborate observations from transgenic models of TR isoform-specific null mice in which TRα1 has a predominant physiological role in the heart (9, 38). However, culture conditions had a marked effect on the differential expression of the TR isoforms in NRVM. As shown in Fig. 1, expression of the TRβ1 protein in NRVM was low when cultured in the absence of thyroid hormone but was increased significantly with T3 treatment. Unliganded TRα1 has recently been shown to repress TRβ gene expression in fetal hearts when circulating thyroid hormone levels are low (24). In contrast, the TRα1 protein content in cardiomyocytes, which was higher than TRβ1 in T3-depleted conditions, was decreased by 54 ± 5% with exposure to T3 (Fig. 1). The observed increase in TRβ1 protein is likely to be the result of T3-induced transcription at previously identified TREs located within the promoter of this gene (35), whereas the mechanism of T3-mediated decrease of TRα1 may be proteosome-mediated degradation of the protein as has been described for numerous nuclear hormone receptors including thyroid receptors (6, 11, 27, 30). To address this possibility, we developed an adenovirus approach to overexpress TRα1 in cultured NRVM to facilitate the study of this phenomenon.
Effect of T3 on adenovirus-expressed TRα1 protein.
The Ad-TRα1 localized primarily to the nucleus in ∼3:1 ratio of nuclear to cytoplasmic compartments. We routinely used a low MOI (2 MOI) so that expression levels of TRα1 would remain low to minimize any potential nonspecific effects of the overexpressed TR in the cell. Most importantly, the Ad-TRα1 behaved similarly to that observed for the endogenous TRα1 in the cultured NRVM. T3 (10−8 M) treatment resulted in a rapid decrease of nuclear TRα1 that could be observed within 4 h of treatment (Fig. 2A). After 8 h of exposure to T3, the amount of TRα1 was ∼20–30% of pretreatment values and remained unchanged at that level up to 24 h. Nuclear TBP content did not change in response to T3, and it was used for normalization of each sample on the immunoblots (Fig. 2). To determine that this effect of T3 was specific to the Ad-TRα1 protein, we transduced myocytes with Ad-βgal in which the bacterial LacZ gene product β-galactosidase (β-gal) localized to the nucleus. As shown in Fig. 2B, T3 had no effect on the nuclear content of β-gal protein at 4 and 8 h and at 24 h (not shown). Furthermore, to rule out any potential effect of T3 on the cytomegalovirus (CMV) promoter that drives the expression of TRα1 in the adenovirus vector, we transiently transfected cultured NRVM with a CMV promoter-luciferase reporter plasmid and treated the cells with T3 for up to 24 h. No effect of T3 on the CMV promoter was observed (data not shown).
Time and dose dependence of T3-mediated decrease of TRα1 and effects on transcription.
We used EMSA to show that Ad-TRα1 in NRVM exhibited TRE-binding activity. As shown in Fig. 3, TRα1 and RXRγ, which were synthesized in programmed reticulocyte lysates, combined to form heterodimers on three different TREs (Fig. 3, lanes 2, 5, and 8), and that TRα1 also formed homodimers on the chick ovalbumin F2 TRE (lane 1). Using these results as a comparison, we determined that the Ad-TRα1 in NRVM preferentially formed heterodimers on the TREs of the cardiac α-MHC and SERCA2 genes and homodimers on the F2 TRE (Fig. 3, lanes 6, 9, and 3, respectively). Furthermore, T3 could not displace the TR heterodimer complex from the TRE, as has been previously published by others (39) (data not shown). These data support the presence of functional Ad-TRα1 in cardiomyocytes that bind to cardiac-specific TREs as heterodimers with endogenous RXR proteins.
To determine whether the virally expressed TRα1 was transcriptionally active, cultured cardiomyocytes were first transiently cotransfected with a plasmid containing a minimal promoter sequence with a direct repeat (DR+4) TRE plus a palindromic TRE driving a luciferase reporter gene and a control promoterless reporter plasmid, and then the cells were transduced with Ad-TRα1. T3-mediated activation of this minimal TRE promoter required coexpression of the Ad-TRα1 as shown by a lack of luciferase activation when Ad-βgal was expressed (Fig. 4B). As shown in Fig. 4A, Western blot analysis showed that TRα1 protein was significantly reduced following 8 h of T3 treatment; however, luciferase activity was not induced at that time, possibly due to the inherent lag between transcription and translation and insufficient accumulation of luciferase protein for analysis. However, by 24 h, luciferase activity was increased significantly from both baseline and the 8-h time point (Fig. 4A). Furthermore, activation of the TRE promoter construct showed a T3 dose dependency with a maximum activation at 10−8 M concentration (Fig. 4B). Transcriptional activation correlated inversely with TRα1 protein content indicating that the T3-induced decrease of TRα1 protein was dose dependent. The 10−11 M T3 was essentially ineffective in inducing the TRE promoter and similarly had no effect on the content of TRα1 protein (Fig. 4B). The specificity of this effect to T3 was shown by treating the cardiomyocytes with rT3, which had no effect on either TRα1 protein content or induction of the TRE promoter (Fig. 4C).
T3-mediated decrease of TRα1 occurs at the protein level.
To determine whether the decrease in TRα1 protein in response to T3 occurred at the posttranslational level, we treated the cultured cardiomyocytes with CHX, an inhibitor of protein translation. As seen in Fig. 5, treatment of the cultured cardiomyocytes with CHX alone for 4.5 h resulted in a 46 ± 7% (P < 0.05) reduction in nuclear TRα1 content as would be expected based on the relatively short half-life of this protein (32). Furthermore, exposure of the CHX-treated myocytes to T3 (10−8 M) for 4 h resulted in a further decrease in TRα1 content to 37 ± 5% of control (P < 0.01), suggesting that the T3-mediated decrease of TRα1 occurred at the protein level (Fig. 5).
To delineate the mechanisms by which TRα1 protein was decreased in response to T3, we focused on the proteosome pathway that has been shown to regulate the stability of many transcription factors, including nuclear receptors. As shown in Fig. 6A, treatment of the cardiomyocytes with the proteosome inhibitor ALLN for 4 h resulted in increased TRα1 protein with evidence of accumulation of higher molecular weight TRα1 proteins, suggesting that the unliganded receptor may be degraded by the ubiquitin-proteosome pathway. Treatment with ALLN for 18 h showed even greater increases in higher molecular weight immunoreactive TRα1 proteins (data not shown). Furthermore, ALLN prevented the T3-mediated decrease of TRα1 protein supporting a role of the proteosome complex in the T3-induced degradation of TRα1 (Fig. 6A). We observed immunoreactivity of the TRα1 protein band when the immunoblots were reprobed with anti-ubiquitin antibody, suggesting that the TRα1 protein was ubiquitylated (Fig. 6B). Because ALLN has been reported to also have calpain and cathepsin activities, we treated the cultured myocytes with ALLM, a calpain inhibitor, and with inhibitors of cathepsin L and B (data not shown). Because none of these agents inhibited T3-induced TRα1 protein degradation, we can attribute the ALLN result solely to its effect on the proteosome. Studies using another proteosome inhibitor MG132 showed similar effectiveness in inhibiting the T3-mediated degradation of TRα1 (data not shown).
T3-mediated decrease of TRα1 protein is regulated at the mRNA level.
We undertook experiments to measure TR isoform mRNA concentrations because previous studies, including our own, had shown that thyroid hormone could regulate TRα1 mRNA expression level (1, 12). Using quantitative real-time PCR, we determined that TRα1 mRNA in cardiomyocytes transduced with either Ad-βgal or with Ad-TRα1 was significantly decreased within 4 h of treatment with T3 to 0.54 ± 0.11 and 0.25 ± 0.05, respectively, of non-T3-treated values (Fig. 7). T3 had no effect on the β-galactosidase mRNA expressed from Ad-βgal, thus supporting the hypothesis that the effect of T3 was specific for TRα1 mRNA and not a result of adenovirus expression of the TR (Fig. 7). To further support the specificity of the effect of T3 on TRα1, we measured endogenous TRα2 and TRβ1 mRNAs. As shown in Fig. 8, T3 had no effect on TRα2 mRNA content but significantly increased the TRβ1 mRNA (2.7 ± 0.4-fold) within 4 h of treatment, which was sustained at that level when measured after 24 h of treatment. These data obtained in cultured NRVM were supported by observations in animal studies in which thyroidectomized rats were treated with T3 (5 μg/100 g body wt), and the RNA was isolated from the left ventricles 24 h after T3 injection. As shown in Fig. 8, TRα1 mRNA was significantly decreased by 30%, whereas TRβ1 mRNA was increased 1.9-fold in the left ventricle of T3-treated animals. The amount of TRα2 mRNA remained unchanged.
We measured the effect of T3 on the half-life of TRα1 mRNA by treating the cultured cardiomyocytes with actinomycin D, an inhibitor of DNA-dependent RNA polymerase II. As clearly shown in Fig. 9A, T3 markedly decreased the half-life of TRα1 mRNA from an estimated 4 h to less than 2 h. In contrast, T3 had no effect on the half-life of TRβ1 mRNA, when measured in the same RNA samples (Fig. 9B).
A variety of experimental and clinical cardiac disease conditions that result in left ventricular hypertrophy and remodeling produce a hypertrophic cardiac phenotype that include changes in expression of many T3-responsive genes (14, 15, 17, 23). Several reports have documented the level of expression of TR mRNA isoforms in human ischemic and dilated cardiomyopathy, with one study (7) showing an increase in both TRα1 and TRα2 mRNAs, while other studies (14, 36) have shown significant decreases in TRα1 mRNA. Overall, these studies have not answered the question of the role of T3 and TRs in determining the heart failure phenotype. Kinugawa et al. (15) published that the cardiac phenotype in experimental pathological cardiac hypertrophy was similar to hypothyroidism and showed that mRNA levels of all three isoforms TRα1, TRα2, and TRβ1 were downregulated in this model. However, these data are discordant with expression of the TR isoforms in the hypothyroid heart in which expression of the TRα mRNA isoforms is increased, along with increased nuclear T3 binding capacity (10). In a recent publication, Mai et al. (24) describe the repressive function of thyroid hormone aporeceptors (unliganded TRs) in a nonpatholoigcal hypothyroid condition such as exists during the fetal period. They showed that in the fetal heart, TRα aporeceptors repressed expression of TRβ as well as several genes encoding ion channels involved in cardiac contractile activity and heart rate. Despite our current understanding of the mechanisms of action of T3 receptors and the dominant negative effects of unliganded receptors, we have yet to ascertain the role of the TRs in determining the hypertrophic cardiac phenotype.
In the present study, we used primary cultures of NRVM that retain the physical and phenotypic characteristics of cardiac myocytes and that represent a pure population of cardiac myocytes. Although numerous studies have documented changes in the expression of TR mRNA isoforms in response to thyroid status and cardiac hypertrophy, fewer studies have reported TR protein content in the heart, and these data have been largely inconsistent. Thus in the present study using cultures of neonatal rat cardiac myocytes, several key findings have emerged. Significantly, the study provides the first identification of the TRα1 protein as the predominant isoform expressed in cardiomyocytes under T3-depleted conditions and raises the possibility that unliganded TRα1, and not TRβ1, determines the hypothyroid cardiac phenotype. Second, the present study provides the first observation of rapid changes in nuclear content of both TRα1 and TRβ1 proteins in response to T3 treatment. Within several hours of T3 exposure, TRα1 protein was decreased significantly by ∼50% and expression of TRβ1 protein was induced. Whereas the T3-induced increase in TRβ1 is likely to be mediated at the transcriptional level (35), the rapid decrease in TRα1 protein in response to T3 treatment had not been previously reported in cardiac myocytes. This study provides the first identification of two potential mechanisms responsible for the observed rapid decrease of TRα1 in response to T3, the proteosome protein degradation pathway, and a T3-induced decrease in TRα1 mRNA half-life.
To study the mechanism of T3-mediated TR degradation, we used replication-defective adenovirus to overexpress the TRα1 protein in the cultured myocytes because of the low abundance of the endogenous protein and the low transfection efficiency of this cell type. We ascertained that 1) the virally expressed TRα1 protein bound to TREs, 2) it was transcriptionally active in response to T3, and 3) it recapitulated the T3-mediated TRα1 decrease observed with the endogenous protein. T3 had no effect on CMV promoter activity nor did T3 alter expression of bacterial β-gal that was expressed from a similarly constructed replication-deficient adenovirus vector. Together, these data supported the utility of the adenovirus expression system to study TRα1 in cardiomyocytes.
The rapidity of the T3-mediated decrease of TRα1 was similar to that reported for agonist-stimulated degradation of RXRs, which were also shown to be rapidly degraded in response to ligand binding when RXR served as a heterodimeric partner for RAR and TR (30). The decrease in TRα1 protein could be seen within 1 h of T3 treatment, with a new steady state reached at ∼4 to 6 h, with no significant change noted up to 24 h after treatment. Furthermore, the decrease of TRα1 protein in response to T3 was dose dependent with 10−8 M having a maximum effect. Experiments in which transcription was measured using transient transfection of a reporter plasmid showed that the transcriptional activity was dependent on both the amount of TRα1 protein and the concentration of T3. Clearly, higher levels of unliganded TR caused greater repression of the TRE-containing promoter, and with increasing concentrations of T3 with its greater effect on TRα1 degradation, transcriptional activity was also increased. Not surprising then, as reported by Dace et al. (6) using CV1 and GC cells, that inhibition of TR degradation by lactacystin, a proteosome inhibitor, resulted in repression of a T3-inducible promoter.
The experiments using CHX to inhibit protein translation showed convincingly that the T3-mediated decrease of TRα1 occurred posttranslationally. Ligand-induced receptor degradation by the proteosome pathway has been reported for several other receptors of the steroid hormone receptor family, including PPAR (11), estrogen receptor (27), and T3 receptors (6). Although not previously shown in cardiomyocytes, we have shown using various proteosome inhibitors that T3 induced the proteolysis of TR by this pathway and that the TR was ubiquitylated. As reported by Dace et al. (6), we also found that T3 did not induce TR ubiquitylation but rather it appeared to stimulate degradation of ubiquitylated TR. The molecular processes by which T3 binding to the ubiquitylated TR results in its degradation remain unknown. Our studies suggest that ubiquitylated TR was present in the nucleus; however, it is not clear whether the T3-mediated degradation of the ubiquitylated protein occurred in the nuclear or cytoplasmic compartment. TRs have been shown to shuttle rapidly between nuclear and cytoplasmic compartments and that interactions with corepressors and RXR appear to maintain the TR in the nucleus (4, 25). Some studies have shown that T3 does not influence the mobility of the TRβ between compartments (25), whereas others have shown that T3 enhances the retention of TRα in the nucleus (4). It is unclear whether these disparate results represent mechanistic differences that are specific to the TR isoform. Thus the precise molecular mechanism by which T3 stimulates TRα degradation by the proteosome, potentially in the nucleus, requires further study.
We tested the hypothesis that changes in TRα1 mRNA turnover may also be involved in the regulation of TR expression, because studies have shown that T3 reduces the half-life of other mRNAs, including thyrotropin β-subunit mRNA (19, 34). Using quantitative RT-PCR, we found that endogenous TRα1 mRNA as well as Ad-TRα1 mRNA decreased 50% to 70% within 4 h of T3 treatment. The specificity of this effect was ascertained by the lack of T3 effect on adenovirus-expressed β-galactosidase mRNA. To determine whether the mechanism responsible for this effect was transcriptional or posttranscriptional, we measured mRNA half-life by inhibiting transcription with actinomycin D. The results showed for the first time that T3 decreased TRα1 mRNA half-life from 4 h to <2 h. Because the half-lives of both TRα1 protein and mRNA can be measured in hours, it is possible that the observed rapid decrease in TRα1 protein following T3 treatment can be explained by an initial induction of TR protein degradation followed by reduced TR protein synthesis due to decreased mRNA. The relative contributions of protein versus mRNA degradation in response to T3 cannot be ascertained by these data and require further study. Insightful studies published by Raaka and Samuels in 1981 (32) showed that nuclear thyroid receptors in pituitary cells were reduced when treated with T3 by a decrease in its rate of synthesis and secondly by a decrease in receptor half-life, suggesting that T3 induced changes in both TR mRNA and protein content.
The lack of effect of T3 on TRβ1 mRNA half-life suggests that the observed increases in TRβ1 mRNA and protein in response to T3 are mediated at the transcriptional level and that TRβ1 expression is regulated by completely different mechanisms than TRα1. These divergent sites of action of the ligand may provide distinctive targets for the development of therapeutic agents directed toward receptor-specific functions.
In summary, our data support a role for both the proteosome complex and mRNA stability in the observed rapid T3-induced decrease of TRα1 protein in cardiomyocytes. The mechanism by which T3 regulates its receptor mRNA half-life is unknown, although it might be interesting to speculate based on recent data published by Xu and Koenig (42), showing that TRs are RNA binding proteins. Alternatively, T3 may alter the expression of proteins that modulate mRNA stability, as has been recently reported for a neuron-specific protein HuD that binds to specific mRNA sequences (5). Our data underscore the importance of cellular mechanisms regulating RNA turnover and stability in determining phenotype and function of the cardiomyocyte, and therefore, provide an avenue for discovery of novel therapeutics directed at these new cellular targets.
This work was supported by National Heart, Lung, and Blood Institute Grants HL-71623 and HL-03775.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 by the American Physiological Society