The c-erbAalpha gene encodes two thyroid hormone receptors, TRalpha1 and TRalpha2, that arise from alternative splicing of the TRalpha pre-mRNA. TRalpha2 is not able to bind triiodothyronine (T(3)) and acts as a weak antagonist of TRs. It has been suggested that the balance of TRalpha1 to TRalpha2 is important in maintaining homeostasis. Here, we study the effect of thyroid hormone on the splicing of TRalpha under various conditions in HepG2 cells. First, T(3) was added to HepG2 cells that endogenously express TRalpha. This resulted in a decrease in the TRalpha1:TRalpha2 mRNA ratio after the addition of 10(-)(8 )M or 10(-)(7 )M T(3). Then, HepG2 cells were incubated with sera from hypothyroid or hyperthyroid patients. Sera from hyperthyroid patients (n=6) decreased the TRalpha1:TRalpha2 ratio compared with HepG2 cells incubated with sera from euthyroid patients (n=8). Sera from hypothyroid patients (n=6) had no effect on the TRalpha1:TRalpha2 ratio but supplementation with T(3) caused a decrease in the ratio. Finally, we tested sera from patients with nonthyroidal illness (NTI; n=17) which showed no effect on TRalpha splicing when compared with controls. Free thyroxine levels in sera from hypo-, eu-, and hyperthyroid patients, but not that of NTI patients, were negatively correlated (P<0.01) to the TRalpha1:TRalpha2 ratio. We next studied the expression of the splicing factors hnRNP A1 and ASF/SF2 (SF2) in relation to the splicing of the TRalpha gene. In HepG2 cells incubated with NTI sera a negative relationship was found between the ratio of hnRNP A1:SF2 and the TRalpha1:TRalpha2 ratio. A high hnRNP A1:SF2 ratio is associated with the use of the distal 5'-splice site. The splicing direction should then change towards TRalpha2, which is indeed the case. Rev-ErbA, which is partly complementary to TRalpha2 and could therefore interfere in the splicing process, did not relate to the TRalpha1:TRalpha2 ratio.In conclusion, high T(3) levels induce a low TRalpha1:TRalpha2 ratio which could protect the cell from excessive T(3)-induced gene expression. In vivo, this might be a mechanism to keep tIssues relatively euthyroid during high serum T(3) levels.

Treatment with amiodarone, a potent antiarrhythmic drug, is associated with a dose-dependent increase in plasma cholesterol resulting from a decreased number of liver low-density lipoprotein (LDL) receptors. Similar changes occur in hypothyroidism, and it has been suggested that amiodarone acts via induction of a local 'hypothyroid-like' state in extrathyroidal tissues. The present study was designed to evaluate whether exogenous tri-iodothyronine (T₃) could prevent the effects of amiodarone on LDL cholesterol.

Rats were treated for 14 days with water, amiodarone 10 mg/100 g body weight (BW), or amiodarone and 2·5, 5 or 10 μg T₃/100 g BW respectively. Relative to controls, amiodarone increased plasma LDL cholesterol by 31% and decreased liver LDL receptor mRNA by 56% and protein by 45%; liver T₃ content was reduced by 21%. Addition of T₃ to the treatment with amiodarone dose-dependently reversed all these changes, with a return to control values of plasma cholesterol and the number of liver LDL receptors, although LDL receptor mRNA remained slightly lower. Treatment of rats for 14 days with T₃ alone (5 μg/100 g BW) decreased plasma LDL cholesterol by 19% and increased liver LDL receptor mRNA by 41%.

In conclusion, T₃ prevents the amiodarone-induced changes in plasma LDL cholesterol and liver LDL receptor gene expression. These findings suggest that the inhibitory effect of amiodarone on LDL receptor gene expression is mediated by T₃-dependent pathways.

Journal of Endocrinology (1997) 152, 413–421

Abstract

To evaluate the role of cytokines in the sick euthyroid syndrome, we tried to establish an animal model of non-thyroidal illness in mice by the administration of a sub-lethal dose of bacterial endotoxin (lipopolysaccharide; LPS) which induces a variety of cytokines, including tumour necrosis factor (TNFα), interleukin-1 (IL-1α), interleukin-6 (IL-6) and interferon-γ (IFNγ). When compared with pair-fed controls, a single dose of LPS resulted in (a) systemic illness, (b) induction of TNFα and IL-6 and (c) a decrease of liver 5′-deiodinase mRNA from 4 h onwards followed by a decrease of serum tri-iodothyronine (T₃) and thyroxine (T₄) at 8 h and of serum free T₃ (fT₃) and free T₄ (fT₄) at 24 h; serum TSH remained unchanged.

We then studied whether a single dose or a combination of IL-1α, TNFα, IL-6 or IFNγ could induce the sick euthyroid syndrome in mice, again using pair-fed controls. None of the cytokines except IL-1α caused systemic illness, and IL-1α was the only cytokine that decreased liver 5′-deiodinase mRNA transiently. IL-1α, TNFα or IL-6 did not decrease serum T₃, T₄ and TSH, but administration of IFNγ decreased serum T₄, T₃ and fT₃ in a dose-dependent manner without changes in serum TSH. Administration of all four cytokines together had no synergistic effects; observed changes were of a smaller magnitude than after LPS.

The following conclusions were reached. (1) Administration of LPS in mice is a suitable experimental model for the acute induction of the sick euthyroid syndrome. (2) Acute administration of IL-1α, TNFα or IL-6 in mice does not induce changes in thyroid hormones but IFNγ results in a dose-dependent decrease of serum T₄, T₃ and fT₃ and IL-1α decreases liver 5′-deiodinase mRNA transiently. (3) Combined administration of IL-1α, TNFα, IL-6 and IFNγ had no synergistic effects; observed changes were of a smaller magnitude than after LPS. (4) The LPS-induced sick euthyroid syndrome is currently best explained by a direct thyroidal inhibition due to IFNγ and an extrathyroidal inhibition of liver 5′-deiodinase due to IL-1α, but other still unidentified factors seem to be involved as well.

Journal of Endocrinology (1995) 146, 475–483

The gene expression of thyroid hormone receptors (TR) in ECRF24 immortalized human umbilical vein endothelial cells (HUVECs) was investigated at both the mRNA and the protein level. Endothelin-1 (ET-1) and von Willebrand factor (vWF) production were measured in response to triiodothyronine (T(3)) administration. A real-time PCR technique was used to quantify the presence of mRNAs encoding for the different isoforms of the TR. The binding of T(3) to nuclear TRs was studied in isolated endothelial cell nuclei by Scatchard analysis. Expression of TR at the protein level was investigated by immunocytochemistry and Western blotting using TR-isoform-specific polyclonal rabbit antisera. ET-1 and vWF were measured in cell supernatants with a two-site immunoenzymatic assay. Scatchard analysis yielded a maximum binding capacity of 55 fmol T(3)/mg DNA (+/-200 sites/cell) with a K(d) of 125 pmol/l. Messenger RNAs encoding for the TRalpha1 and the TRalpha2 and the TRbeta1 were observed. The approximate number of mRNA molecules per cell was at least 50 molecules per cell for TRalpha1, five for TRalpha2 and two for TRbeta1. Immunocytochemistry revealed (peri)nuclear staining for TRbeta1, TRalpha1 and TRalpha2. ET-1 and vWF secretion did not increase upon addition of T(3) (10(-10)-10(-6) M). Immortalized ECRF24 HUVECs express TR, but at low levels. The number of TRs per endothelial cell is probably too low to be functional and no change in ET-1 or vWF production was found after addition of T(3). Therefore we conclude that the genomic effects of T(3) are unlikely to occur in these immortalized HUVECs.

Many metabolic processes occur simultaneously in the liver in different locations along the porto-central axis of the liver units. These processes are often regulated by hormones, one of which is thyroid hormone which for its action depends on the presence of the different isoforms of the thyroid hormone receptor (TR). These are encoded by two genes: c-erbA-alpha encoding TRalpha1 and TRalpha2 and their respective Delta isoforms, and c-erbA-beta which encodes TRbeta1, TRbeta2 and TRbeta3. We recently found a zonal (pericentral) expression of and a diurnal variation in the TRbeta1 isoform in rat liver. We were therefore also interested to see whether TRalpha1 and TRalpha2 expression showed similar characteristics. For this reason we raised both polyclonal and monoclonal antibodies against TRalpha1 and TRalpha2 isoforms and characterised these. Antibody specificity was tested using Western blots and immunohistochemistry in liver of TR isoform-specific knockout animals. Using these antibodies we found that the TRalpha1 and TRalpha2 isoforms are zonally expressed around the central vein in rat liver. The experiments show that the portal to central gradient of TRalpha1 is broader than that of TRbeta1. Moreover, the expression of the TRalpha2 protein showed a diurnal variation with a peak in the afternoon when the animals are least active whereas no such variation was found for the TRalpha1 protein.From our data it appears that both the TRalpha1 and TRalpha2 isoforms show a zonal distribution in liver. This finding, together with the observed diurnal rhythm, has major implications for interpreting and timing experiments concerning the TR and its downstream actions in liver.

Nuclear thyroid hormone (T₃) receptors (TR) play a critical role in mediating the effects of T₃ on development, differentiation and normal physiology of many organs. The heart is a major target organ of T₃, and recent studies in knockout mice demonstrated distinct effects of the different TR isoforms on cardiac function, but the specific actions of TR isoforms and their specific localization in the heart remain unclear. We therefore studied the expression of TRα1, TRα2 and TRβ1 isoforms in the mouse heart at different stages of development, using monoclonal antibodies against TRα1, TRα2 and TRβ1. In order to identify distinct components of the embryonic heart, in situ hybridization for cardiac-specific markers was used with the expression pattern of sarcoplasmic reticulum calcium-ATPase 2a as a marker of myocardial structures, while the pattern of expression of connexin40 was used to indicate the developing chamber myocardium and peripheral ventricular conduction system. Here we show that in the ventricles of the adult heart the TRβ1 isoform is confined to the cells that form the peripheral ventricular conduction system. TRα1, on the other hand, is present in working myocardium as well as in the peripheral ventricular conduction system. In the atria and in the proximal conduction system (sinoatrial node, atrio-ventricular node), TRα1 and TRβ1 isoforms are co-expressed. We also found the heterogeneous expression of the TRα1, TRα2 and TRβ1 isoforms in the developing mouse heart, which, in the case of the TRβ1 isoform, gradually revealed a dynamic expression pattern. It was present in all cardiomyocytes at the early stages of cardiogenesis, but from embryonic day 11.5 and into adulthood, TRβ1 demonstrated a gradual confinement to the peripheral ventricular conduction system (PVCS), suggesting a specific role of this isoform in the formation of PVCS. Detailed knowledge of the distribution of TRα1 and TRβ1 in the heart is of importance for understanding not only their mechanism of action in the heart but also the design and (clinical) use of TR isoform-specific agonists and antagonists.