Thyroid hormone induces cardiac myocyte hypertrophy in a thyroid hormone receptor alpha1-specific manner that requires TAK1 and p38 mitogen-activated protein kinase

Abstract

Alterations in TR [thyroid hormone (TH) receptor]1 isoform expression have been reported in models of both physiologic and pathologic cardiac hypertrophy as well as in patients with heart failure. In this report, we demonstrate that TH induces hypertrophy as a direct result of binding to the TRalpha1 isoform and, moreover, that overexpression of TRalpha1 alone is also associated with a hypertrophic phenotype, even in the absence of ligand. The mechanism of TH and TRalpha1-specific hypertrophy is novel for a nuclear hormone receptor and involves the transforming growth factor beta-activated kinase (TAK1) and p38. Mitigating TRalpha1 effects, both TRalpha2 and TRbeta1 attenuate TRalpha1-induced myocardial growth and gene expression by diminishing TAK1 and p38 activities, respectively. These findings refine our previous observations on TR expression in the hypertrophied and failing heart and suggest that manipulation of thyroid hormone signaling in an isoform-specific manner may be a relevant therapeutic target for altering the pathologic myocardial program.

Figure 1. Cardiac myocyte expression of human TRs.

a. Immunostaining. Neonatal rat cardiac myocytes (MCs) were exposed to adenovirus at 100 MOI for 72 h. Upper panels are immunofluoresence pictures of cells infected with the indicated AdTRs incubated with the C1 antibody that only recognizes human TR isoforms (α and β). Bottom panels were the same cells co-incubated with antibody to sarcomeric α-actin which identifies cardiac myocytes. Less than 5% of cells were sarcomeric actin-negative (nonmyocytes, NMCs). A myocyte without expression of human TRβ₁ (↑↑) is identified. Note the restriction of hTRβ₁ expression to the nucleus of these cells while both hTRα₁ and hTRα₂ appear to be distributed in both nuclear and cytoplasmic compartments. b. Electrophoretic mobility shift assay for the DR4 (direct repeat-4) TRE. Cells were exposed to adenovirus at 50 MOI for 48 h. For supershift assays, the same human TR-specific antibodies used in Figure 1c were used, and are denoted as “+” isoform-specific Ab. B1 and B2 consists of heterodimers of retinoid X receptor (RXRα, β, or γ) and TR (1 molecule of each), and homodimers of TRs (2 TR molecules), respectively. No monomer binding was observed. Competitor lanes were with unlabeled oligonucleotide. The 200MOI lane for hTRβ was included since this was the only condition where cytosolic hTRβ was found c. Quantification and sub-cellular location of human TR over-expression in neonatal rat cardiac myocytes. Myocytes were infected with the individual AdTRs at the indicated MOIs for 48 hours. Fractionated cell extracts were prepared and subjected to Western blotting with human-specific TR antibodies in the upper panels (hence no rat TR is detected in un-infected lanes). In the binding experiments, cell extracts from equal numbers of cells were subjected to [¹²⁵I]T₃-binding assay as described previously (44). Notably, expression of TRα₁ was readily found in both nuclear and cytoplasmic fractions while AdTRβ₁ expression was generally limited to the nucleus.

Figure 2. Over-expression of TRα₁ induces myocyte hypertrophy

a. MOI-dependent effects on protein synthesis by AdTRs. Cells were infected with AdβGal, AdTRα₁, AdTRα₂, or AdTRβ₁ for 48h at the designated MOI. Radiolabeled protein content (RLP) was normalized to AdβGal at identical MOIs and at the 0.3MOI level for the subsequent increases in AdβGal itself. For comparison, the RLP seen with 100nM T₃ alone is shown. b. Effects of T₃ or GC-1 on protein synthesis. Cells were infected with AdβGal, AdTRα₁, AdTRα₂, or AdTRβ₁ for 48h at 10MOI with various concentrations of T₃ or GC-1. Values were normalized to vehicle + AdβGal at 10MOI. c. Effects of AdTRα₂ or AdTRβ₁ on AdTRα₁-induced hypertrophy. Cells were treated with AdTRα₁ at 10MOI (□) with the addition of AdTRα₂ or AdTRβ₁ at the indicated MOIs for 48h.

Figure 3. TH and TRα₁ hypertrophy is p38-dependent.

a. Dose-dependent effects of SB202190 on T₃ and AdTRα₁-induced myocyte growth. Cells were pretreated with the indicated dose of SB202190 (“SB”) or null SB202474 (“Null”) for 30 min, followed by the addition of AdβGal (50MOI, not shown), AdβGal+T₃ (100 nM) or AdTRα₁ for 48h. Values were normalized to that of AdβGal + vehicle. b. Cells were pretreated with vehicle (DMSO) or U0126 (1 μM) for 30 min or AdJNK1DN, AdMKK3DN, or Adp38αDN for 24 h. Cells were subsequently infected with AdβGal (50MOI, not shown), and treated with T₃ (100 nM, AdβGal+ T₃), or AdTRα₁ for 48 h. Values were normalized to that of AdβGal + vehicle. c. T₃/TRα stimulation of p38MAPK. Cells were treated with T₃ (100 nM) or AdTRα₁ (50MOI) for the designated times (left and middle panels) or infected with AdβGal or AdTRs (50MOI) for 24h and T₃ (100nM) added for an additional 15 minutes (right panel). Phospho-p38 was then determined by Western blotting. Both T₃ and AdTRα₁ activate p38 (bottom panel). AdβGal cells were treated with T₃ (15 min) or AdTRα₁ (24h) at indicated doses. In vitro p38 activity was measured by immune complex kinase assay with GST-ATF2.

Figure 4. TH and TRα₁ activate MKK3/6 and TAK1, but not ERK or JNK.

a. Activated (phosphorylated) MKK3 (upper band) and MKK6 (lower band) expression increase in T₃ and AdTRα₁ treated cells. b. In vitro TAK1 activity was measured by immune complex kinase assay using MalMKK3 in cells treated with T₃ alone (15 min) or in the presence of the indicated AdTRs (48h infection). c. Phosphorylated and total ERK1/2 and JNK1/2 expression were also examined in similarly treated T₃ and TR infected cells. As a positive control, cells were treated with 20% of fetal bovine serum (FBS) for 30 min.

Figure 5. Cytosolic TRα₁ interacts with TAK1.

a. TRα₁ and TRα₂, (but not TRβ₁) interact with TAK1. Lanes 1–3: Human-specific TR antibody was validated for Western blotting with control human TRs synthesized in rabbit reticulocyte lysate [TRα₁ (~48kDa), TRα₂ (~58kDa), and TRβ₁ (~52kDa)]. Doublets represent lysate-specific in vitro processing and are not seen in AdTR-infected cells. Lanes 4–6: Myocytes were infected with AdTRs at 50 MOI for 24h followed by immunoprecipitation of endogenous TAK1. This was subjected to Western for TR. Lanes 7–9: Expression of human TRs in each sample was confirmed using the same antibody. b. Whole cell extract from un-infected cells was immunoprecipitated with rabbit IgG or rat-specific TRα₁ antibody, and subjected to Western blotting for TAK1. c. Western blotting and immunofluorescence microscopy for endogenous cardiac myocyte TAK1 expression. d and e. TRβ₁ (but not TRα₁ and TRα₂) interacts with p38 and diminishes its kinase activity. Cells were infected with AdTRs and Adp38αWT for 24h. Total p38 was immunoprecipitated, and subjected to Western for humanTR (C1). e. In vitro synthesized human TRβ₁ or control rabbit reticulocyte lysate was mixed with active MKK6 or active p38α (~68kDa), and their activities measured on unactive recombinant GST-p38α (~64kDa) or GST-ATF2 (~40kDa), respectively. SB202190 was used at 10nM.

Figure 6. TR isoform-specific changes in the cardiac myocyte gene program.

Cells were treated with AdβGal at 50 MOI with or without T₃ (100nM) for 72h and compared with cells infected with AdTRα₁ or AdTRα₂, or AdTRβ₁ at 50 MOI. Values of the corresponding Adβ Gal group were set at 100%, and data is presented as % change from 100%, n=3–4. As such, a value of 0% equals no change from AdβGal infected cells and 100% represents a doubling of signal. All signals were corrected for RNA loading using an internal GAPDH signal.

Figure 7. Proposed schema of T₃/TR isoform-specific action on cardiac myocyte MAPK signaling and gene program.

See text for details.