3.1.2 Biosynthesis, transport, metabolism, and actions of thyroid hormones

In healthy humans with a normal iodine intake, the thyroid follicular cells produce predominantly the prohormone thyroxine (3,3′,5,5′-tetraiodothyronine; T₄), which is converted in peripheral tissues to the bioactive hormone 3,3′,5-triiodothyronine (T₃) or to the inactive metabolite 3,3′,5′-triiodothyronine (reverse T₃). The bioavailability of thyroid hormone in target tissues depends to a large extent on the supply of plasma T₄ and T₃, the activity of transporters mediating the cellular uptake and/or efflux of these hormones, as well as the activity of deiodinases and possibly other enzymes catalyzing their activation or inactivation. Thyroid function is regulated most importantly by the hypophyseal glycoprotein thyroid-stimulating hormone (TSH), also called thyrotropin. In turn, TSH secretion from the anterior pituitary is stimulated by the hypothalamic factor thyrotropin-releasing hormone (TRH). TSH secretion is down-regulated by negative feedback action of thyroid hormone on the hypothalamus and the pituitary. The contribution of locally produced T₃ versus uptake of plasma T₃ is much greater for some tissues such as the brain and the pituitary than for most other tissues. Plasma TSH is an important parameter for the diagnosis of thyroid dysfunction but is not representative for the thyroid state of all tissues. In this chapter various aspects will be discussed of: (a) the neuroendocrine regulation of thyroid function, (b) the biosynthesis of thyroid hormone (i.e. the prohormone T₄), (c) the activation and inactivation of thyroid hormone in peripheral tissues, and (d) the mechanism by which T₃ exerts it biological activity. A schematic overview of the hypothalamus– pituitary–thyroid–periphery axis is presented in Fig. 3.1.2.1.

TRH is a tripeptide with the structure pyroglutamyl-histidyl-proline amide (pGlu-His-Pro-NH₂) in which the C-terminal carboxyl group is blocked by amidation and the N-terminal α-amino group is blocked by cyclization. Beside stimulating TSH secretion, TRH also stimulates prolactin secretion and, under certain pathological conditions, growth hormone secretion from the anterior pituitary. TRH is not only produced in the hypothalamus but is widely distributed through the central nervous system where it functions as a neurotransmitter or neuromodulator. Centrally mediated actions of TRH include neurobehavioural, haemodynamic, and gastrointestinal effects. TRH is also detected in the posterior pituitary and in different peripheral tissues, such as the pancreas, the heart, the testis, the adrenal, and the placenta. Little is known about the function of TRH in these tissues.

Hypophysiotropic TRH is produced in neurons, the cell bodies of which are located in the paraventricular nucleus of the hypothalamus (1). The biosynthesis of TRH involves the production of a large precursor protein (proTRH) which, in humans, consists of a sequence of 242 amino acids. This proTRH contains six copies of the TRH progenitor sequence Gln-His-Pro-Gly, flanked at both sides by pairs of the basic amino acids Arg and/or Lys (Fig. 3.1.2.2). Cleavage of proTRH at the basic amino acids by prohormone convertases (e.g. PC1 and PC2) and further removal of remaining basic residues by carboxypeptidases results in the liberation of the progenitor sequences. A specific glutaminyl cyclase catalyses the formation of the pGlu ring at the N-terminus and a so-called peptidylglycine α-amidating mono-oxygenase converts Pro-Gly to ProNH₂ at the C-terminus (2). The processing of proTRH takes place in vesicles that transport mature TRH and intervening peptides along the axons of the TRH neurons to the median eminence, where they are released into the portal vessels of the hypophyseal stalk.

TRH is transported over a short distance through the hypophyseal stalk to the anterior lobe of the pituitary, where it stimulates the production and secretion of TSH (and prolactin). These actions of TRH are initiated by its binding to the type 1 TRH receptor (TRHR1), which is expressed on both the thyrotroph (TSH-producing cell) and the lactotroph (prolactin-producing cell) (3). This receptor belongs to the family of G-protein-coupled receptors, characteristically containing seven transmembrane domains. Human TRHR1 is a protein consisting of 398 amino acids, and binding of TRH induces a change in its interaction with the trimeric G-protein, resulting in the stimulation of phospholipase C activity. The activated phospholipase C catalyses the hydrolysis of phosphatidylinositol-4,5-diphosphate to the second messengers inositol-1,4,5-triphosphate and diacylglycerol, which initiate a cascade of reactions, including an increase in cellular Ca²⁺ levels and protein kinase C activity, that ultimately stimulates the release as well as the synthesis of TSH (and prolactin) (3). TRH stimulation of TSHβ gene expression is also dependent on the pituitary-specific transcription factor 1.

In addition to the TRHR1 expressed in the anterior pituitary, a second TRH receptor (TRHR2) has been cloned and characterized in rat and mouse brain which probably mediates most central actions of TRH (3). In humans, only one type of TRH receptor exists, namely TRHR1.

TRH is subject to rapid degradation in the blood as well as in different tissues. Although multiple enzymes are involved, a very important role is played by the TRH-degrading ectoenzyme TRHDE, which catalyses the cleavage of the pGlu-His bond (4):

This enzyme has been characterized as a zinc-containing metalloproteinase, which in humans consists of 1024 amino acids. It has a single transmembrane domain and is inserted in the plasma membrane such that most of the protein is exposed on the cell surface (ectopeptidase), in particular in brain, pituitary, liver, and lung. Enzymatic cleavage of the protein close to the cell membrane releases most of the protein in a soluble and enzymatically active form into the circulation, representing the origin of plasma TRHDE. Plasma TRHDE appears to be derived mostly from the liver. In the brain and the pituitary, where the enzyme is probably located in close vicinity of the TRH receptor, TRHDE supposedly plays an important role in the local regulation of TRH bioavailability. Interestingly, TRHDE activity in the pituitary and in plasma is increased in hyperthyroidism and decreased in hypothyroidism, which may contribute to the negative feedback control of TSH secretion by thyroid hormone (4).

TSH is a glycoprotein produced by the thyrotropic cells of the anterior pituitary. Like the other hypophyseal hormones, luteinizing hormone and follicle-stimulating hormone, it is composed of two subunits. The α-subunit is identical and the β-subunit is homologous among the three hormones (5). Although hormone specificity is conveyed by the β-subunit, dimerization with the α-subunit is required for biological activity. Human TSH consists of 205 amino acids; 92 in the α-subunit and 113 in the β-subunit. It has a molecular weight of 28 kDa, 20% of which is contributed by three complex carbohydrate groups: two on the α-subunit and one on the β-subunit. The structure of these carbohydrate groups is important for the biological activity of TSH and is dependent on the stimulation of the thyrotroph by TRH (5).

In addition to the stimulation by TRH and negative feedback by thyroid hormone, TSH production and secretion is also subject to negative regulation by hypothalamic somatostatin and dopamine, and by cortisol (6). The inhibitory effect of cortisol is exerted to an important extent at the hypothalamic level.

TSH binds to a specific TSH receptor located in the plasma membrane of the follicular cell. Like the TRH receptor, this is also a G-protein-coupled receptor which, in humans, is a protein consisting of 764 amino acids with an exceptionally long extracellular N-terminal domain (7). The TSH receptor is preferentially coupled to a G_sα-subunit of the trimeric G-protein. Binding of TSH to its receptor induces the dissociation of the G-protein subunits, resulting in the activation of the membrane-bound adenylate cyclase and, thus, in the stimulation of cAMP formation as second messenger. The increased cAMP levels induce a series of events, including the activation of protein kinase A activity, that ultimately results in the stimulation of the biosynthesis and secretion of thyroid hormone (8). In particular, the expression of genes coding for key proteins for hormone production (e.g. the iodide transporter, thyroglobulin, and thyroid peroxidase) is increased through mechanisms that also involve different thyroid-specific transcription factors such as TTF1, TTF2, and PAX8. At high TSH concentrations, the TSH receptor also couples to the G_qα-subunit, resulting in the activation of the phosphoinositide pathway, which is also involved in the regulation of thyroid function and growth (8).

As discussed elsewhere in this section, hyperthyroidism is often caused by an autoimmune process in which TSH receptor stimulating antibodies play an important role. Hyperthyroidism may also be caused by a hyperfunctioning adenoma. In most patients with a toxic adenoma, somatic mutations have been identified in the TSH receptor, which result in the constitutive activation of this receptor (9). In other patients, somatic mutations have been found in the G_sα-subunit that result in the constitutive activation of the G-protein in the absence of TSH. Together, mutations in the TSH receptor and the G_sα-subunit account for the majority of toxic thyroid adenomas. Also, germline, gain-of-function mutations have been identified in patients with congenital, nonautoimmune hyperthyroidism. Conversely, germline, loss-of-function mutations have been described in patients with TSH resistance (9). Such a loss-of-function mutation has also been identified as the cause of the hypothyroidism in the hyt/hyt mouse. However, patients with TSH resistance may be clinically euthyroid because the partial defect in TSH receptor function is compensated by increased plasma TSH levels (9).

The functional unit of the thyroid gland is the follicle, composed of a single layer of epithelial cells surrounding a colloidal lumen in which thyroid hormone is stored as an integral part of its precursor protein thyroglobulin. The biosynthesis of thyroid hormone comprises the following steps, which are depicted schematically in Fig. 3.1.2.3 (8, 10):

Iodine is an essential trace element required for the synthesis of thyroid hormone. It is not surprising, therefore, that the basolateral membrane of the follicular cell contains an active transporter that mediates uptake of I⁻ together with Na⁺. This sodium-iodide symporter (NIS) has been characterized as a protein consisting, in humans, of 618 amino acids and 13 transmembrane domains (12). Supposedly, these domains form a channel through which I⁻ and Na⁺ are transported in a stoichiometry of 1:2. The surplus of positive charge indicates that I⁻ transport is electrogenic and further driven by the Na⁺ gradient. TSH stimulates the expression of the NIS gene to such an extent that the intracellular iodide concentration may be up to 500 times higher than its extracellular level. The NIS is not completely specific for iodide but also binds other anions, some of which are even transported (12).

An important example is perchlorate (ClO₄⁻) which potently inhibits iodide uptake by the NIS, an effect utilized in the perchlorate discharge test used for the diagnosis of an organification defect, i.e. impaired incorporation of iodine in thyroglobulin. Perchlorate inhibits the uptake but not the release of iodide from the thyroid. Therefore, if perchlorate is administered after a dose of radioactive iodide, it will provoke a marked release of radioactivity from the thyroid in case of an organification defect but not from a normal thyroid gland. Pertechnetate (TcO₄⁻) is another anion transported by the NIS, and this observation is utilized in the scanning of the thyroid gland using radioactive ^99mTcO₄⁻. Of course, the latter is not incorporated in thyroglobulin and, thus, cannot be used to test the hormone production capacity of the thyroid.

It is not sufficient that iodide is transported across the plasma membrane. Since the iodination of thyroglobulin takes place at the luminal surface of the apical membrane, iodide also has to pass this membrane. A transporter putatively involved in this process has been identified and termed pendrin, since the gene coding for this protein is mutated in patients with Pendred’s syndrome (13). This is a congenital condition characterized by deafness due to a cochlear defect and hypothyroidism due to an organification defect as indicated by a positive perchlorate discharge test. Pendrin is capable of transporting bicarbonate, chloride, and iodide (13), and is expressed only in the thyroid and the cochlea. The exact function of pendrin in the transport of iodide across the apical membrane is subject to debate. Most likely, it is not the only protein capable of releasing iodide into the follicular lumen. Efflux of iodide from thyroid follicular cells is acutely stimulated by TSH, which may involve recruitment and/or activation of an iodide exporter such as pendrin. The function of pendrin in the cochlea probably lies in the secretion of bicarbonate into the endolymph.

Thyroglobulin is an exceptionally large glycoprotein consisting of two identical subunits. Each mature subunit in human thyroglobulin contains 2748 amino acids and has a molecular weight of approximately 330 kDa (14). The TG gene is located on human chromosome 8q24.2-q24.3; it covers about 300 kb of genomic DNA and consists of 48 exons.

DUOX2 is a large and complex glycoprotein embedded in the apical membrane of the thyrocyte. Mature human DUOX2 contains 1527 amino acids and has seven putative transmembrane domains, an NADPH-binding domain, an FAD-binding domain, a haem-binding domain, two calcium-binding EF hands, and a peroxidase domain (15). It catalyses the oxidation of NADPH from the cytoplasm and delivers its product H₂O₂ to the luminal surface of the membrane. The haem group appears to be the site of H₂O₂ generation and its location within transmembrane domains fits with the vectorial (enzyme/transport) function of DUOX2. Functional expression of DUOX2 requires the presence of the maturation factor DUOXA2, a protein consisting of 320 amino acids and five putative transmembrane domains (16). The DUOX2 and DUOXA2 genes are clustered together with the homologous DUOX1 and DUOXA1 genes on human chromosome 15q15.

TPO is a glycoprotein consisting of 933 amino acids and featuring a single transmembrane domain. A short C-terminal domain is located in the cytoplasm but most of the protein is exposed on the luminal surface of the apical membrane, which also contains a haem-binding domain, the active centre of the enzyme (17). Functional TPO may exist as a homodimeric structure linked through a disulfide bond. The human TPO gene covers about 150 kb on chromosome 2p25, distributed over 17 exons. In addition to full-length TPO1, the TPO2 splice variant is generated by the skipping of exon 10, resulting in the loss of 57 amino acids in the middle of the protein (17). TPO2 has no enzyme activity and its function is unknown.

Thyroid hormone synthesis takes place at the luminal surface of the apical membrane in the scaffold of the thyroglobulin molecule and consists of two important reactions that are both catalysed by TPO, i.e. the iodination of Tyr residues and the subsequent coupling of iodotyrosine to iodothyronine residues (17). The structures of these compounds are illustrated in Fig. 3.1.2.4. The prosthetic haem group of TPO undergoes a two-electron oxidation by H₂O₂ (supplied by DUOX2) to the intermediate compound 1 (Cpd1). Cpd1 may carry out either a one-electron oxidation reaction, by which it is converted to the intermediate Cpd2, or a two-electron oxidation by which native TPO is regenerated. TPO-catalysed iodination probably involves a two-electron oxidation of I⁻ to I⁺ with subsequent electrophilic substitution of Tyr residues in thyroglobulin, producing 3-iodotyrosine (monoiodotyrosine, MIT). Substitution of MIT residues with a second iodine produces 3,5-diiodotyrosine (DIT).

Coupling of two suitably positioned iodotyrosine residues results in the formation of an iodothyronine residue at the site of the acceptor iodotyrosine, leaving a dehydroalanine residue at the site of the donor iodotyrosine (17). It is generally believed that coupling involves the one-electron oxidation of each donor and acceptor iodotyrosine residue, generating radicals that rapidly combine to produce an iodothyronine residue. Coupling of the diiodophenol moiety of one DIT residue to the phenolic oxygen of a second DIT residue results in the formation of T₄, while coupling of the iodophenol moiety of MIT to a DIT residue yields T₃. Coupling of DIT and MIT to generate reverse T₃ is apparently a rare event, since thyroidal secretion of reverse T₃ is negligible. This probably also holds for formation of 3,3′-T₂ by coupling of two MIT residues.

Although Tyr is the building block of thyroid hormone, the Tyr content of thyroglobulin is not greater than that of most other proteins. Of the 67 Tyr residues per thyroglobulin subunit, about 20–25 are available for iodination, but the capacity for iodothyronine formation is limited (17). Each thyroglobulin subunit has only four hormonogenic sites, Tyr residues that can ultimately be transformed into iodothyronines. At three sites (positions 5, 1290, and 2553 in the mature protein) T₄ can be formed, while at the fourth site (position 2746) T₃ is preferentially produced. However, at normal levels of iodination the average yield is 1–1.5 molecules of T₄ and approximately 0.1 molecule of T₃ per thyroglobulin subunit. At this stage the iodothyronines are still in peptide linkage with the thyroglobulin backbone and remain stored as such in the lumen until their secretion is required.

In response to TSH stimulation, thyroglobulin is resorbed from the lumen largely by both macro- and micropinocytosis (8, 10). The former type of endocytosis is associated with the formation of large pseudopodia that engulf colloid and the thyroglobulin contained therein, resulting in the formation of large cytoplasmic vesicles also known as colloid droplets. The second process concerns the receptor-mediated endocytosis of thyroglobulin, involving the binding of thyroglobulin to apical membrane proteins. Megalin, a very large (c.600 kDa) cargo protein located in the apical membrane of different cell types, including thyrocytes, may be involved although it appears to function primarily in the transcellular transport of poorly iodinated thyroglobulin (18).

Both types of endosomes fuse with lysosomes, generating so-called phagolysosomes. In these vesicles thyroglobulin is hydrolysed by lysosomal proteases, i.e. cathepsins (19), resulting in the liberation of T₄, a small amount of T₃, as well as excess MIT and DIT molecules. MIT and DIT are probably exported from the vesicles via a specific transporter (20). Thus, they have access to the iodotyrosine dehalogenase (DEHAL1 or IYD), located in the endoplasmic reticulum, which catalyses their deiodination by NADH (21, 22). The iodide thereby released is reutilized for iodination of thyroglobulin.

Human DEHAL1 is a homodimer of a 289-amino acid protein containing an N-terminal membrane anchor and a conserved nitroreductase domain with an FMN-binding site (21, 22). The DEHAL1 gene is located on chromosome 6q24-q25 and consists of five exons. Since DEHAL1 lacks an NADH-binding sequence, iodotyrosine deiodinase activity requires the involvement of a reductase, which has not yet been identified. DEHAL1 is also expressed in the liver and kidneys.

Surprisingly little is still known about the exact mechanism of thyroid hormone secretion. One option involves the transcellular transport of the thyroglobulin-liberated iodothyronines in vesicles, which fuse with the basolateral membrane and release their content in the extracellular compartment. Alternatively, iodothyronines may be released via transporters from the vesicles into the cytoplasm, and subsequently secreted through transporters located in the basolateral membrane. In the latter route, some T₄ may be converted before secretion to T₃ by iodothyronine deiodinases present in the thyrocyte (see below). Recent findings suggest that the transporter MCT8 (see below) plays an important role in thyroid hormone secretion.

In an average human subject, T₄ and T₃ are secreted in a ratio of about 15:1, i.e. about 100 μg (130 nmol) T₄ and 6 μg (9 nmol) T₃ per day. The latter represents approximately 20% of daily total T₃ production (23). Hence, most T₃ is produced by deiodination of T₄ in peripheral tissues.

Administration of a large amount of iodide usually results in an acute but transient decrease in thyroid hormone secretion (8, 10). The mechanism of this inhibition of thyroid hormone secretion by excess iodide is unknown. Excess iodide will also induce an inhibition of the synthesis of thyroid hormone; this phenomenon is known as the Wolff–Chaikoff effect (8, 10). The mechanism appears to involve, among other things, the formation of an iodinated lipid (iodolactone) that inhibits several steps in thyroid hormone synthesis. This includes the inhibition of iodide uptake by the NIS, which results in a decrease in the intracellular iodide concentration and, thus, a decrease in iodolactone formation. This relieves the inhibited hormone synthesis, known as the escape from the Wolff–Chaikoff effect, that occurs despite the continued administration of excess iodide.

Thiourea derivatives have been known since the pioneering work of Astwood in the 1940s as potent inhibitors of thyroid hormone synthesis (24). Two of these, methimazole and 6-propyl-2-thiouracil are widely used in the medical treatment of patients with hyperthyroidism (Fig. 3.1.2.5). Their antithyroid activity is based on the potent inhibition of TPO, the mechanism of which depends on the available iodide concentration (17). In the presence of iodide, the thiourea inhibitors compete with the Tyr residues in thyroglobulin for the TPO–I⁺ iodination complex, preventing the formation of thyroid hormone. The thiourea inhibitors are thus converted to the sulfenyl iodide derivatives which undergo further oxidation of the sulfur ultimately to sulfate.

Methimazole is a more potent inhibitor of TPO than propylthiouracil (17), and lower doses of methimazole (or the prodrug carbimazole) are required for the treatment of hyperthyroidism compared with propylthiouracil. Besides inhibiting thyroid hormone (i.e. T₄) synthesis by TPO, propylthiouracil also inhibits conversion of T₄ to T₃ by the type 1 iodothyronine deiodinase located not only in the thyroid but also in liver and kidney (see below). In contrast, methimazole does not affect D1 activity.

In plasma, thyroid hormone is bound to three proteins, thyroxine-binding globulin (TBG), transthyretin (TTR, previously known as thyroxine-binding prealbumin (TBPA)), and albumin (Table 3.1.2.1) (25). Human TBG is a 54-kDa glycoprotein produced in the liver and consists of 395 amino acids and four carbohydrate residues. The TBG gene is located on the human chromosome Xq22.2, spans about 5.5 kb, and contains five exons (26). Among the different thyroid hormone transport proteins it shows by far the highest affinity for T₄, with an equilibrium dissociation constant (K_d) of approximately 0.1 nM, but also the lowest plasma concentration (c.15 mg/l) (25).

TTR is composed of four identical subunits, each consisting of 127 amino acids. The TTR gene is located on human chromosome 18q11.2-q12.1, covers about 7 kb, and contains four exons (27). TTR has a cigar-shaped structure with two identical binding channels, each formed by two symmetrically positioned subunits, with ligand entry sites at opposite ends of the TTR molecule. Binding of a T₄ molecule in one site hinders the binding of another T₄ molecule in the second site. Binding of the first T₄ molecule to TTR is characterized by a K_d value of approximately 10 nM, and the plasma concentration of TTR amounts to approximately 250 mg/l (25). Plasma TTR is produced in the liver, but the protein is also expressed in the choroid plexus where it is probably involved in T₄ transfer from plasma to the cerebrospinal fluid. Furthermore, TTR is expressed in trophoblasts where it may participate in the transplacental transfer of maternal T₄ to the fetus. TTR also binds retinol-binding protein and thus also plays an important role in vitamin A transport (27).

Albumin has multiple low-affinity binding sites for thyroid hormone, with K_d values for T₄ of 1–10 µM, but it has by far the highest plasma concentration (c.40 g/l) (25). Iodothyronines also bind to lipoproteins, in particular high-density lipoprotein. Although the proportion of plasma T₄ and T₃ bound to lipoproteins is low compared with the other plasma transport proteins, it may be important to target thyroid hormone specifically to lipoprotein receptor-expressing tissues (25).

The resultant of the concentrations and affinities of the different thyroid hormone-binding proteins is that in normal human subjects approximately 75% of plasma T₄ is bound to TBG, approximately 15% is bound to albumin, and approximately 10% is bound to TTR (25). The total binding capacity of these proteins is so high that only approximately 0.02% of plasma T₄ is free (non-protein-bound). The affinity of T₃ for the different proteins is roughly 10% of that of T₄. Therefore, plasma T₃ shows a similar distribution to T₄ over the different proteins, and the free T₃ fraction in normal plasma amounts to approximately 0.2%. Thus, while the mean normal plasma total T₄ (c.100 nmol/l) and T₃ (c.2 nmol/l) levels differ about 50-fold, the difference in the mean normal free T₄ (c.20 pmol/l) and free T₃ (c.5 pmol/l) is only about fourfold. Reverse T₃ binds with intermediate affinity to the plasma proteins (25).

Since it is the plasma free T₄ and free T₃ concentrations that determine the tissue availability of thyroid hormone, they are more important parameters than the plasma total T₄ and T₃ concentrations in the assessment of thyroid status. Both concentration and thyroid hormone-binding affinity of the different plasma proteins are influenced by a variety of (patho)physiological factors (25). Since it binds most thyroid hormone in plasma, variations in TBG concentration are more important than variations in TTR or albumin concentrations. Inherited TBG excess is a rare phenomenon caused by TBG gene duplication. Inherited TBG deficiency is often caused by a single base mutation in the TBG gene, resulting in a decreased T₄ affinity or a decreased protein stability. More severe TBG gene defects are responsible for a complete lack of serum TBG in affected hemizygous males (26). Beside genetic variation, TBG levels are also influenced by various endogenous and exogenous factors. Notably, plasma TBG levels are increased by oestrogens, whereas they are decreased by androgens. In addition, different endogenous factors, such as free fatty acids, and drugs, such as salicylates, competitively inhibit T₄ binding to TBG (25).

A large number of mutations have also been identified in the TTR gene, some of which are associated with a decrease in T₄ binding affinity, whereas others (e.g. Ala109Thr and Thr119Met) result in an increased affinity for T₄ (28). More importantly, however, TTR mutations often cause neuropathic or cardiomyopathic amyloidosis, resulting from the deposition of insoluble TTR fibrils in nerves or the heart (28). Finally, binding of thyroid hormone to albumin is subject to genetic variation. In particular, a specific increase in the binding of T₄ to albumin is frequently observed in otherwise healthy subjects, which may lead to the false diagnosis of hyperthyroidism if inadequate methods for analysis of plasma free T₄ are used (25). This phenomenon of familial dysalbuminaemic hyperthyroxinaemia has been attributed to mutations in the albumin gene, resulting in a marked increase in T₄ affinity (29).

Perturbation of plasma iodothyronine binding provokes an adaptation of the hypothalamus–pituitary–thyroid axis until normal free T₄ and free T₃ concentrations are again obtained. Therefore, measurement of plasma free T₄ rather than total T₄ levels is, together with analysis of plasma TSH, the cornerstone of the diagnosis of thyroid disorders.

Because iodothyronines are lipophilic compounds, it has been generally assumed that they readily pass the plasma membrane by simple diffusion. However, the polar nature of the alanine side chain (‘zwitterion’) is a serious obstacle for passage through the lipid bilayer of the cell membrane. However, studies in recent years have established that tissue uptake of thyroid hormone does not take place by diffusion but is mediated by specific plasma membrane transporters (30). Most studies have been carried out in isolated rat hepatocytes, but carrier-mediated uptake of iodothyronines has been demonstrated in a variety of cells, including neuronal cells, astrocytes, erythrocytes, thymocytes, choriocarcinoma cells, fibroblasts, (cardio)myocytes, and anterior pituitary (tumour) cells (30).

The kinetics of T₄ and T₃ uptake by isolated rat hepatocytes suggest the involvement of multiple mechanisms with different affinities (30). The high-affinity components are characterized by K_m values in the nanomolar range and most likely represent cellular uptake of the iodothyronines by specific transporters. The low-affinity components are characterized by K_m values in the micromolar range and may represent uptake of the iodothyronines by nonspecific transporters or binding to the cell surface. Although T₃ is capable of inhibiting T₄ uptake, and vice versa, the large difference between the K_m and K_i values for each iodothyronine suggests the involvement of different transporters (30).

High-affinity transport of thyroid hormone into rat liver cells is an active process, dependent on the ATP content of the cells. A decrease in cellular ATP has a greater effect on T₄ and reverse T₃ uptake than on T₃ uptake, supporting the involvement of different transporters. In vitro studies in isolated human hepatocytes as well as in vivo studies in human subjects also suggest energy- dependent thyroid hormone uptake into the human liver. Uptake of iodothyronines by hepatocytes is inhibited by the Na⁺,K⁺-ATPase inhibitor ouabain, suggesting Na⁺ dependence of the transporters involved (30). Thyroid hormone uptake in liver is inhibited by different iodinated compounds such as the antiarrhythmic drug amiodarone and the radiographic contrast agents iopanoic acid and ipodate (see below).

The mechanisms of thyroid hormone uptake appear to differ between tissues as studies in rat pituitary cells suggest a common transporter for T₄ and T₃, whereas neonatal rat cardiomyocytes show preferential uptake of T₃ over T₄ (30). In view of the iodothyronine structure, it is not surprising that thyroid hormone uptake by different cell types is mediated, at least in part, by amino acid transporters showing partial (L-type) or complete (T-type) preference for aromatic amino acids (31, 32).

Since 2000, a number of thyroid hormone transporters have been identified at the molecular level (Fig. 3.1.2.6). These include the Na-taurocholate cotransporting polypeptide (NTCP), different members of the organic anion transporting polypeptide (OATP) family, the L-type amino acid transporters (LATs), and members of the monocarboxylate transporter family (33, 34).

Of these transporters, only NTCP (SLC10A1) transports its ligands in a Na⁺-dependent manner (35). It is exclusively expressed in liver and transports primarily bile acids. Human NTCP consists of 349 amino acids and has seven transmembrane domains. The NTCP gene is located on chromosome 14q24.1 and has five exons. There are no other thyroid hormone transporters in the SLC10 family. NTCP shows a preference for sulfated over nonsulfated iodothyronines and probably is not the major Na⁺-dependent thyroid hormone transporter in liver, which therefore remains elusive.

The human OATP family contains of 11 members, most of which have been shown to transport iodothyronine derivatives (36). In general they are multispecific, transporting a variety of ligands, not only anionic but also neutral and even cationic compounds. OATPs are glycoproteins containing around 700 amino acids and 12 transmembrane domains. The human OATP1 subfamily contains four members (OATP1A2, 1B1, 1B3, 1C1) with quite interesting properties. They are encoded by a gene cluster on chromosome 12p12 containing 14–15 exons. OATP1B1 and 1B3 are expressed only in liver and show preference for sulfated over nonsulfated iodothyronines as ligands (37). The latter also holds for OATP1A2, which is expressed in different tissues. OATP1C1 is by far the most interesting transporter in this subfamily, showing a high preference for T₄ as the ligand and almost exclusive expression in the brain, especially in choroid plexus and capillaries. It thus appears very important for T₄ transport into the brain (36).

T₄ and T₃ are also transported by two members of the heterodimeric amino acid transporters LAT1 and LAT2 (38). These transporters are glycoproteins consisting of two subunits, a heavy chain and a light chain. In humans, there are two possible heavy chains (SLC3A1,2) and least 13 possible light chains (SLC7A1–11,13,14). The heavy chains contain a single transmembrane domain, and the light chains contain 12–14 transmembrane domains. LAT1 is composed of the SLC3A2 (4F2hc or CD98hc) heavy chain and the SLC7A5 light chain, and LAT2 is composed of the same heavy chain and the SLC7A8 light chain. These transporters are expressed in various tissues, and stimulated in activated immune cells and in tumours. Both LAT1 and LAT2 facilitate the bidirectional transport of a variety of aliphatic and aromatic amino acids as well as iodothyronines over the plasma membrane (38).

Two important thyroid hormone transporters come from an unexpected family, the monocarboxylate transporter (MCT) family, named such because MCT1–4 facilitate transport of monocarboxylates such as lactate and pyruvate (39, 40). Functional expression of MCT1–4 requires their interaction with the ancillary proteins basigin (CD147) or embigin. The MCT family contains 14 members, but the function of most of these transporters is as yet unknown. However, two members from this family, MCT8 and MCT10, have recently been identified as important thyroid hormone transporters. Of these, MCT10 also transports the aromatic amino acids Trp, Tyr, and Phe, but so far only iodothyronines have been identified as ligands for MCT8.

Human MCT8 consists of 613 or 539 amino acids, depending on which of the two possible translation start sites is used, and MCT10 has 515 amino acids. They are homologous proteins with about 50% amino acid identity between ‘short’ MCT8 and MCT10 (Fig. 3.1.2.7). Like the other MCTs, both MCT8 and MCT10 have 12 transmembrane domains. However, they are not glycosylated and they also do not appear to require ancillary proteins for functional expression. They have identical gene structures; the MCT8 gene is located on human chromosome Xq13.2, and the MCT10 gene is located on chromosome 6q21-q22. Both consist of six exons and five introns, with a large approximately 100 kb first intron. MCT8 and MCT10 show wide but different tissue distributions.

MCT8 and MCT10 are the most active and specific thyroid hormone transporters known today (34, 41, 42). MCT8 is importantly expressed in brain, where it is localized in choroid plexus, capillaries, and neurons in different brain areas. MCT8 appears to be essential for T₃ uptake in central neurons and, thus, for the crucial action of thyroid hormone during brain development. Mutations in MCT8 have recently been identified as the cause of the Allan–Herndon–Dudley syndrome that occurs in male patients and is characterized by severe psychomotor retardation in combination with highly elevated serum T₃ levels (34, 41, 42) (see Chapter 3.4.8).

Thyroid hormone metabolism takes place intracellularly (see next section) and requires cellular uptake of iodothyronines over the plasma membrane. Thus, T₄ uptake in T₃-producing tissues is one of the factors determining peripheral T₃ production (30). A diminished liver T₄ uptake may therefore contribute to the decreased T₃ production underlying the low T₃ syndrome induced by nonthyroidal illness and fasting. This may be due in part to inhibition of T₄ transporters by plasma factors such as bilirubin and free fatty acids, which are increased in illness. Radiographic agents and other iodinated compounds such as the antiarrhythmic drug amiodarone also inhibit liver uptake of T₄ which may contribute to the decrease in serum T₃ induced by their administration. Since T₃ exerts most of its effects by binding to intracellular (nuclear) receptors, thyroid hormone bioactivity also depends on the activity of T₃ transporters in different tissues.

The thyroid gland of a healthy human adult with an adequate iodine intake produces predominantly the prohormone T₄ and only a small amount of the bioactive hormone T₃. It is generally accepted that, in humans, approximately 80% of circulating T₃ is produced by enzymatic outer ring deiodination (ORD) of T₄ in peripheral tissues (23). Alternatively, inner ring deiodination (IRD) of T₄ produces the inactive metabolite reverse T₃, thyroidal secretion of which is negligible. Deiodination is also an important pathway by which T₃ and reverse T₃ are further metabolized. T₃ largely undergoes IRD to the inactive compound T₂, which is also the main metabolite produced from reverse T₃ by ORD (Fig. 3.1.2.8). Thus, the bioactivity of thyroid hormone is determined to an important extent by the enzyme activities responsible for the ORD (activation) or IRD (inactivation) of iodothyronines.

Three iodothyronine deiodinases (D1–3) are involved in the reductive deiodination of thyroid hormone (Fig. 3.1.2.9) (43). They are homologous proteins consisting of 249–278 amino acids, with a single transmembrane domain located at the N-terminus. The deiodinases are inserted in cellular membranes such that the major part of the protein is exposed on the cytoplasmic surface. This is consistent with the reductive nature of the cytoplasmic compartment required for the deiodination process. Probably all three deiodinases are functionally expressed as homodimers (43).

The most remarkable feature of all three deiodinases is the presence of a selenocysteine (Sec) residue in the centre of the amino acid sequence. As in other selenoproteins, this Sec residue is encoded by a UGA triplet, which in mRNAs for nonselenoproteins functions as a translation stop codon. The translation of the UGA codon into Sec requires the presence of a particular stemloop structure in the 3′-untranslated region of the mRNA, termed Sec-insertion sequence (SECIS) element, Sec-tRNA, and a number of cellular proteins, including SECIS-binding protein (SBP2). A bona fide SECIS element has been identified in the mRNA of all deiodinases (43).

D1 is a membrane-bound enzyme expressed predominantly in liver, kidneys, and thyroid (43). It catalyses the ORD and/or IRD of a variety of iodothyronine derivatives, although it is most effective in the ORD of reverse T₃. In the presence of dithiothreitol (DTT) as the cofactor, D1 displays high K_m and V_max values. Hepatic D1 is probably a major site for the production of plasma T₃ and clearance of plasma reverse T₃. D1 activity in liver and kidney is increased in hyperthyroidism and decreased in hypothyroidism, representing the regulation of D1 activity by T₃ at the transcriptional level.

Hepatic and renal D1 activities are strongly reduced in rats fed a selenium-deficient diet, resulting in a decrease in serum T₃ and an increase in serum T₄. The Sec residue is essential for the function of D1 since substitution with Cys reduces enzyme activity to 1%, while substitution with Leu yields a completely inactive protein. Rapid inactivation of D1 by iodoacetate is probably due to modification of the highly reactive Sec residue. Moreover, D1 activity is extremely sensitive to inhibition by very low concentrations of gold thioglucose by formation of a stable complex with the Sec residue. Thus, Sec is the catalytic centre of D1 (Fig. 3.1.2.10).

The different deiodinases require thiols as cofactor. Although reduced glutathione is the most abundant intracellular thiol, its activity is very low compared with the unnatural thiol DTT, which is often used in in vitro studies. Alternative endogenous cofactors include dihydrolipoamide, glutaredoxin, and thioredoxin. D1 shows ping-pong-type kinetics in catalysing the deiodination of iodothyronines by DTT. D1 activity is potently inhibited by propylthiouracil, and this inhibition is uncompetitive with substrate and competitive with cofactor. Together, these findings suggest that the catalytic mechanism of D1 involves the transfer of an iodinium ion (I⁺) from the substrate to the selenolate (Se⁻) group of the enzyme, generating a selenenyl iodide intermediate which is reduced back to native enzyme by thiols such as DTT or converted into a dead-end complex by propylthiouracil (Fig. 3.1.2.10).

D2 is expressed primarily in brain, anterior pituitary, brown adipose tissue, thyroid, and to some extent also in skeletal muscle (43, 44). D2 mRNA is also expressed in human heart, but it is unknown to what extent this is translated into functional deiodinase. In brain, D2 mRNA has been localized in astrocytes, in particular also in tanycytes lining the third ventricle in the arcuate nucleus–median eminence region. D2 is a low-K_m, low-capacity enzyme possessing only ORD activity, with a preference for T₄ over reverse T₃ as the substrate. The amount of T₃ in brain, pituitary, and brown adipose tissue is derived to a large extent from local conversion of T₄ by D2 and to a minor extent from plasma T₃ (23, 43). The enzyme located in the anterior pituitary and the arcuate nucleus of the hypothalamus appears very important for the negative feedback regulation of TSH and TRH secretion (1).

In general, D2 activity is increased in hypothyroidism and decreased in hyperthyroidism. This is explained in part by substrate-induced inactivation of the enzyme by T₄ and reverse T₃ involving the ubiquitin-proteasome system (43). However, inhibition of D2 activity and mRNA levels by T₃ has also been demonstrated in the brain and pituitary. The substrate (T₄, reverse T₃) and product (T₃)-dependent down-regulation of D2 activity is important to maintain brain T₃ levels in the face of changing plasma thyroid hormone levels.

In mammals, D2 mRNA contains a second UGA codon just upstream of a UAA stop codon (43). It remains to be determined to what extent this second TGA codon specifies the incorporation of a second Sec residue or acts as a translation stop codon. The amino acid sequence downstream of this second Sec is not required for enzyme activity.

D3 activity has been detected in different human tissues, brain, skin, liver, and intestine, where activities are much higher in the fetal stage than in the adult stage (43). D3 is also abundantly expressed in placenta and pregnant uterus. D3 has only IRD activity, catalysing the inactivation of T₄ and T₃ with intermediate K_m and V_max values. D3 in tissues such as the brain is thought to play a role in the regulation of intracellular T₃ levels, while its presence in placenta, pregnant uterus, and fetal tissues may serve to protect developing organs against undue exposure to active thyroid hormone. Indeed, fetal plasma contains low T₃ (and high reverse T₃) concentrations. However, local D2-mediated T₃ production from T₄ is crucial for brain development. Also in adult subjects, D3 appears to be an important site for clearance of plasma T₃ and production of plasma reverse T₃. In brain, but not in placenta, D3 activity is increased in hyperthyroidism and decreased in hypothyroidism, which at least in brain is associated with parallel changes in D3 mRNA levels (23, 43).

In contrast to the marked decrease in hepatic and renal (but not thyroidal) D1 activities, there are only minor effects of selenium deficiency on tissue D2 and D3 activities (45). This may be explained by findings that the selenium state of different tissues varies greatly in selenium-deficient animals. In addition, the efficiency of the SECIS element to facilitate read-through of the UGA codon may differ among selenoproteins, which could result in the preferred incorporation of Sec into D2 or D3 over other selenoproteins.

The presence of Sec in a strongly conserved region of the proteins suggests the same catalytic mechanism for the different deiodinases. However, D2 and D3 are much less susceptible than D1 to the mechanism-based inhibitors propylthiouracil, iodoacetate, and gold thioglucose (43). This could be explained if the reactivity of the selenol group in D2 and D3 is much lower than that in D1. Indeed, substitution of Sec in D1 with the much less reactive Cys is associated with a dramatic decrease in its sensitivity to inhibition by gold thioglucose and propylthiouracil. Interestingly, the amino acid two positions downstream of the catalytic Sec residue (Ser in D1, Pro in D2 and D3) plays an important role in determining the reactivity of the catalytic Sec residue (43).

Intriguing metabolites are generated by side-chain metabolism of iodothyronines (Fig. 3.1.2.11). Presumably by action of aromatic l-amino acid decarboxylase (AADC) iodothyronines are converted into iodothyronamines. In particular two of these, 3-iodothyronamine (T₁AM) and thyronamine (T₀AM), have high affinity for the trace amine receptor TAR1, and exert acute and dramatic effects on heart rate, body temperature, and physical activity, inducing a torpor-like state (46). Thus, these thyroid hormone metabolites appear to have neurotransmitter-like properties, adding a novel dimension to the already diverse effects of the conventional thyroid hormone structures.

Presumably by further conversion of iodothyronamines by the monoamine oxidases MAO-A or MAO-B, the iodothyroacetic acid metabolites 3,3′,5,5′-tetraiodothyroacetic acid (Tetrac) and 3,3′,5-triiodothyroacetic acid (Triac) are generated from T₄ and T₃, respectively (Fig. 3.1.2.11) (47). Although, in general, oxidative deamination is an inactivating pathway for monoamines, Triac has significant thyromimetic activity and its affinity for the T₃ receptor TRα1 is equal to that of T₃ and for the TRβ receptor it is even higher that of T₃ (see next section). There may be multiple pathways leading from T₄ and T₃ to T₁AM and T₀AM with different orders for the successive decarboxylation and deiodination steps. Iodothyronamines are deiodinated by the different deiodinases (48), but it is unknown which iodothyronines are substrates for AADC or which iodothyronamines are converted by MAO-A or MAO-B. Also, the exact biological functions of the iodothyronamine and iodothyroacetic acid metabolites remain to be established.

In addition to deiodination, iodothyronines are metabolized by conjugation of the phenolic hydroxyl group with sulfate or glucuronic acid (Fig. 3.1.2.11). Sulfation and glucuronidation are so-called phase II detoxification reactions, which increase the water solubility of substrates and, thus, facilitate their biliary and/or urinary clearance. However, iodothyronine sulfate levels are normally very low in plasma, bile, and urine, as these conjugates are rapidly degraded by D1, suggesting that sulfate conjugation is a primary step leading to the irreversible inactivation of thyroid hormone (49, 50). Thus, the IRD of T₄ sulfate to reverse T₃ sulfate and of T₃ sulfate to T₂ sulfate is orders of magnitude faster than the IRD of nonsulfated T₄ and T₃, whereas the ORD of T₄ sulfate to T₃ sulfate is completely blocked. Plasma levels (and biliary excretion) of iodothyronine sulfates are increased if D1 activity is inhibited by drugs such as propylthiouracil, and during fetal development, nonthyroidal illness, and fasting. Under these conditions, T₃ sulfate may function as a reservoir of inactive hormone from which active T₃ may be recovered by action of tissue sulfatases and bacterial sulfatases in the intestine.

Sulfotransferases represent a family of enzymes with a monomer molecular weight of approximately 34 kDa, located in the cytoplasm of different tissues, in particular liver, kidney, intestine, and brain. They catalyse the transfer of sulfate from 3′-phosphoadenosine-5′-phosphosulfate to usually a hydroxyl group of the substrate. Different phenol sulfotransferases have been identified with significant activity towards iodothyronines. These include human SULT1A1, 1A2, 1A3, 1B1, and 1C2 (49). They have a large substrate preference for T₂, which is sulfated orders of magnitude faster than T₃ or reverse T₃, whereas sulfation of T₄ is hardly detectable.

Surprisingly, human oestrogen sulfotransferase (SULT1E1) is an important isoenzyme for sulfation of thyroid hormone. Although human SULT1E1 shows much greater affinity for oestrogens (K_mc.nM) than for iodothyronines (K_mc.μM), it sulfates T₂ and T₃ as efficiently as other SULTs, and is much more efficient in sulfating reverse T₃ and T₄ (49). Human tissues expressing SULT1E1 include liver, uterus, and mammary gland (51). In particular, the enzyme expressed in the endometrium may be a significant source of the high levels of iodothyronine sulfates in human fetal plasma. Different human SULTs have also been shown to catalyse the sulfation of iodothyronamines (52).

In contrast to the sulfates, iodothyronine glucuronides are rapidly excreted in the bile. However, this is not an irreversible pathway of hormone disposal since, after hydrolysis of the glucuronides by bacterial β-glucuronidases in the intestine, part of the liberated iodothyronines is reabsorbed, constituting an enterohepatic cycle (50, 53). Nevertheless, about 20% of daily T₄ production appears in the faeces, probably through biliary excretion of glucuronide conjugates. Glucuronidation is catalysed by UDP-glucuronyltransferases (UGTs) that utilize UDP-glucuronic acid as cofactor. UGTs are localized in the endoplasmic reticulum of predominantly liver, kidney, and intestine. Most UGTs are members of the UGT1A and UGT2B families (54).

Glucuronidation of T₄ and T₃ is catalysed by different members of the UGT1A family, 1A1, 1A3, and 1A7–10. Usually, this involves the glucuronidation of the hydroxyl group (Fig. 3.1.2.11), but human UGT1A3 also catalyses the glucuronidation of the side-chain carboxyl group, with formation of so-called acyl glucuronides (55). Interestingly, Tetrac and Triac are much more rapidly glucuronidated in human liver than T₄ and T₃, and this occurs predominantly by acyl glucuronidation (56).

In rodents, metabolism of thyroid hormone is accelerated through induction of T₄-glucuronidating UGTs by different classes of compounds, including barbiturates, fibrates, and polychlorinated biphenyls (57, 58). This may result in a hypothyroid state as the thyroid gland is not capable of compensating for the increased hormone loss. In humans, thyroid function may be affected by induction of T₄ glucuronidation by antiepileptics, but overt hypothyroidism is rare (59). Administration of such drugs to T₄-replaced hypothyroid patients may necessitate an increase in the T₄ substitution dose.

Thyroid hormone is critical for the development of different tissues, in particular the brain, but it is also essential for an optimal function of most tissues in adult life (60). It is probably the most important factor regulating thermogenesis, as reflected by the increase in the basal metabolic rate in hyperthyroid subjects and the decrease observed in hypothyroid individuals (61–63). The positive effect of thyroid hormone on the resting metabolic rate appears to be largely mediated by the stimulation of so-called futile cycles. This concerns the cycling of substrates of the intermediary metabolism as well as that of cations such as Na⁺, K⁺, and Ca²⁺ across cellular membranes. Such cycles result in the net hydrolysis of ATP, the energy of which is dissipated as heat. Thyroid hormone increases the synthesis as well the degradation of proteins, lipids, and carbohydrates, predominantly by stimulating the expression of key enzymes involved in these processes. Examples of these are the lipogenic enzymes, malic enzyme, fatty acid synthase, and glucose-6-phosphate dehydrogenase, and the gluconeogenic enzyme phosphoenolpyruvate carboxykinase.

Special forms of substrate cycling take place between the cytoplasm and the mitochondrion, such as the glycerol-3-phosphate/dihydroxyacetone phosphate shuttle in which cytoplasmic and mitochondrial α-glycerophosphate dehydrogenase (αGPD) isoenzymes participate (61, 62). This represents one way to enable oxidation of cytoplasmic NADH in the mitochondrion, which is impermeable to this cofactor. Thyroid hormone stimulates the expression of mitochondrial αGPD, and the increased electron flow via this enzyme is associated with an increased heat production relative to ATP synthesis.

Thyroid hormone also increases the activity of Na⁺,K⁺-ATPase, an enzyme located in the plasma membrane of all tissues, in particular kidney, heart, and skeletal muscle, which is responsible for the maintenance of the Na⁺ and K⁺ gradients across this membrane. This increased Na⁺,K⁺-ATPase activity is only functional if associated with—and perhaps triggered by—the activation of processes that tend to dissipate these gradients (61, 62). Tissue uptake of glucose, amino acids, fatty acids, and other nutrients predominantly occurs by cotransport with Na⁺ via specific plasma membrane transporters. Stimulated cycling of these substrates by thyroid hormone, which may also involve increased expression of the transporters, is thus accompanied by a significant cellular Na⁺ influx. In addition, thyroid hormone may promote the permeability of the cell membrane for Na⁺ and K⁺ by activation of channels for these ions. In myocytes, the increased Na⁺,K⁺-ATPase activity accelerates the repolarization of the sarcolemma following a depolarization stimulus that contributes to the tachycardia induced by thyroid hormone. T₃ stimulates the expression of both (α and β) subunits of Na⁺,K⁺-ATPase by increasing the transcription of the genes as well as by stabilization of the mRNAs (61).

Another important target for thyroid hormone action is the Ca²⁺-ATPase located in the sarcoplasmic reticulum of muscle cells (63). Innervation of the myocyte triggers the release of large amounts of Ca²⁺ from the sarcoplasmic reticulum into the cytoplasm, where it binds to the actomyosin complex that initiates contraction. Relaxation of the muscle requires the reuptake of the Ca²⁺ into the sarcoplasmic reticulum by Ca²⁺-ATPase at the expense of ATP. There are two Ca²⁺-ATPase isoenzymes, SERCA1 that is characteristic for fast-type skeletal muscle and SERCA2 that is characteristic for slow-type skeletal muscle and heart. T₃ increases Ca²⁺-ATPase activity by stimulating the transcription of both SERCA1 and SERCA2 genes, which explains the increased relaxation rate of the muscle induced by T₃ (63).

It is difficult to estimate how much the increased Ca²⁺-ATPase activity accounts for the T₃-induced increase in resting energy expenditure of muscle, since the extent of futile Ca²⁺ cycling is unknown in resting muscle. This depends not only on the activity of the Ca²⁺-ATPase but also on the rate of Ca²⁺ leak from the sarcoplasmic reticulum. However, it has been estimated that excess Ca²⁺ cycling in contracting muscle may account for up to 50% of the T₃-dependent energy expenditure during work or shivering (63). The remainder of the T₃-induced energy turnover in contracting muscle is largely accounted for by the change in the expression of two forms of the myosin heavy chain which are characterized by high (MHCα) and low (MHCβ) ATPase activities and contraction rates. T₃ stimulates the expression of the MHCα gene, whereas it inhibits the expression of the MHCb gene (63). A similar T₃-induced shift in MHC expression is also observed in the heart (64).

In addition, T₃ increases the expression of the uncoupling protein UCP1 in brown adipose tissue (BAT) (61, 62). This is an important mechanism by which T₃ stimulates nonshivering cold-induced thermogenesis. UCP1 is an ion transporter located in the inner mitochondrial membrane which dissipates the proton gradient over this membrane generated by the respiratory chain, producing heat instead of ATP. Significant amounts of BAT were thought to be present only in small mammals and the human infant. Recently, however, significant BAT depots have also been demonstrated in the neck and shoulder region of normal adults, especially in cold-adapted subjects and more so in younger females than in older males (65). Cold exposure leads to a dramatic stimulation of D2 expression in BAT, and the resultant induction of local T₃ production plays an important role in the stimulation of BAT activity. This includes increased mobilization and burning of lipids as well as stimulated UCP1 expression, together resulting in a major increase in heat production (61, 62).

UCP1 is expressed exclusively in BAT. Other members of the UCP family are expressed in other human tissues, including UCP2 in a variety of tissues including heart and skeletal muscle, UCP3 in skeletal muscle, and UCP4 and UCP5 in brain. The expression of UCP2 and UCP3 is also under positive control of thyroid hormone, but their role in T₃-induced thermogenesis has not been established (66).

The regulation of the mitochondrial proteins UCP1 and αGPD by thyroid hormone is mediated predominantly by interaction of the nuclear T₃ receptor with the promoters of these genes (61, 66). However, there is also evidence for direct effects of thyroid hormone on the mitochondria, the mechanism of which is incompletely understood but may involve interaction of T₃ and other iodothyronines such as 3,3'-T₂ and 3,5-T₂ with cytochrome c oxidase (67). Many studies have reported effects of thyroid hormone on cellular processes that are not mediated by the nuclear T₃ receptor, including stimulation of transport of glucose, amino acids, and ions over the cell membrane, stimulation of actin polymerization in neurons, and stimulation of mitogen-activated protein kinase activity. The last is mediated by the binding of iodothyronines to integrin, a plasma membrane receptor. The interested reader is referred to a recent extensive review of these extranuclear actions of thyroid hormone (68).

Specific thyroid hormone-binding sites have also been detected in the cytoplasm in different tissues. A notable example is the NADPH-dependent cytoplasmic thyroid hormone-binding protein present in rat liver, which appears to be important for the trafficking of thyroid hormone to the nucleus or mitochondria (69).

Most biological actions of T₃ are initiated by its binding to nuclear T₃ receptors (70–72). These proteins are members of the superfamily of ligand-dependent transcription factors, which also includes the receptors for steroids (e.g. cortisol, oestradiol, and testosterone), 1,25-dihydroxyvitamin D₃, retinoic acid, and 9-cis-retinoic acid. The last, so-called retinoid X receptor (RXR) is an important member of this gene family, because it forms functional heterodimers with a number of other nuclear receptors, including T₃ receptors. Two T₃ receptor genes have been identified; the α gene is located on human chromosome 17 and the β gene on human chromosome 3. By alternative exon utilization of both genes, four major receptor isoforms, TRα1, TRα2, TRβ1, and TRβ2, are generated, which consist of 410–514 amino acids (Fig. 3.1.2.12). Although the β gene (150 kb) is much larger than the α gene (c.30 kb), they have similar genomic structures, comprising 10 (β) or 11 (α) exons, and their coding sequences show a high degree of homology (70–72).

As in the other members of the nuclear receptor family, functional key domains have been recognized in the T₃ receptors, in particular the DNA-binding domain (DBD), which is approximately 100 amino acids long, and the ligand-binding domain (LBD), which is approximately 250 amino acids in length (70–72). The amino acid sequences of the TRα and TRβ subtypes are most homologous in their DBD and LBD and least homologous at their N-terminus. The latter contains the ligand-independent AF1 transactivation domain, while an AF2 domain necessary for homo- and heterodimerization and ligand-dependent activation is located at the C-terminus. The short sequence between the DBD and the LBD is usually referred to as the hinge region.

The structural difference between TRα1 and TRα2 is located at the C-terminus of the proteins, where the sequences of the last 40 amino acids in TRα1 and 122 amino acids in TRα2 differ completely due to alternative splicing. The alteration in the LBD of TRα2 is associated with a complete loss of T₃ binding. Therefore, this splice variant is not a bona fide T₃ receptor, but for convenience it will still be referred to here as TRα2. TRα2 has a weak negative effect on the action of T₃ through the other T₃ receptors. The N-terminal domains of TRβ1 (106 amino acids) and TRβ2 (159 amino acids) differ almost completely due to utilization of alternative transcription start sites. Apparently, this domain provides TRβ2 with specific properties required for T₃-induced down-regulation of TRH and TSH genes (70–72).

The high homology between the LBDs of TRα1 and TRβ explains their very similar ligand specificity, with affinities decreasing in the order T₃ more than T₄ more than reverse T₃. However, the metabolite Triac also binds to the T₃ receptors with an affinity equal to (TRα1) or even greater than (TRβ1) that of T₃ (73). Nevertheless, T₃ is the major endogenous iodothyronine occupying the nuclear thyroid hormone receptors, which are thus true T₃ receptors. Recently, several TRβ-specific agonists have been developed with pharmacologically interesting and selective effects on the liver, resulting in lowering of body weight, lipid, and cholesterol without detrimental effects on the heart (72, 73). Most likely, the tissue-specific effects of these compounds is not only determined by their affinity for the T₃ receptor isoforms but also by the diverse ligand-preference of thyroid hormone transporters in different tissues. Interestingly, nonselective T₃ receptor antagonists have been developed as well (72, 73).

The different T₃ receptor isoforms show distinct tissue distributions (70–72). The TRα1 is the predominant T₃ receptor expressed in brain, heart, and bone, whereas TRβ1 is the major receptor in other tissues, including liver, skeletal muscle, kidney, and fat. TRβ2 is preferentially expressed in the anterior pituitary and the hypothalamic area of the brain. These locations suggest the particular involvement of TRβ2 in the feedback inhibition of TSH and TRH secretion by thyroid hormone. Exon utilization specifying TRβ2 expression in the anterior pituitary is under the control of pituitary-specific transcription factor 1, response elements for which are located in the TRβ gene promoter (74). Regulation of the expression of T₃-responsive genes involves the binding of the T₃ receptors to so-called T₃ response elements (TREs) in the promoter region of these genes (70–72). TREs usually consist of two half-sites arranged as repeats or palindromes. The most prevalent TRE half-site sequence is AGGTCA, and the direct repeat of this half-site spaced by four nucleotides (DR4) is a particularly powerful TRE. However, some TREs show marked deviation from this ‘consensus’ half-site sequence, which, moreover, is also recognized by other receptors such as RXR and the retinoic acid receptor. This may be the basis for ‘cross-talk’ between different nuclear receptors and their target genes. Although T₃ receptors may bind as homodimers to the TREs, T₃ effects on gene expression are usually mediated by T₃ receptor/RXR heterodimers.

Binding of the T₃ receptor/RXR heterodimer to TRE does not require T₃ or 9-cis-retinoic acid, the ligand for RXR. The DBDs of these (and other) nuclear receptors contain two ‘zinc fingers’ (peptide loops that chelate a zinc atom) that fit in the grooves of the DNA and are, thus, very important for the specificity of the receptor-promoter interaction (70–72). In the absence of T₃ and irrespective of the presence of 9-cis-retinoic acid, binding of the T₃ receptor/RXR heterodimer to the TRE results in suppressed gene transcription mediated by the binding of corepressor proteins such as NCoR (nuclear corepressor) or SMRT (silencing mediator of retinoid and thyroid hormone receptors) to a specific region (CoR box) of the unliganded T₃ receptor (Fig. 3.1.2.13). These corepressors directly or indirectly inhibit the activity of the basal transcription machinery.

Binding of T₃ induces a conformational change in the T₃ receptor, which results in the release of the corepressors and the recruitment of coactivator proteins such as SRC1 (steroid receptor coactivator-1) and CBP (cAMP response element-binding protein (CREB)-binding protein) (70–72). The AF2 domain, a highly conserved 9-amino acid sequence located at the C-terminus of the different nuclear receptors, plays an important role in the binding of the coactivators. The latter directly or indirectly stimulate the activity of the basal transcription machinery. One mechanism by which transcription is stimulated involves the histone acetyltransferase activity of the coactivators or of other proteins with which they interact. Acetylation of histones loosens the chromatin structure and thus facilitates interaction of the transcription machinery with the DNA. Conversely, corepressors may recruit proteins with deacetylase activity.

The above discussion of the mechanism of action of T₃ concerns the expression of genes which are under positive control of thyroid hormone. However, a roughly equal number of genes are negatively regulated by T₃, in particular those involved in the negative feedback regulation of the hypothalamus–pituitary–thyroid axis, i.e. the TSHb and the TRH genes. In the promoter regions of these genes negative TREs have been identified that often consist of only one half-site. In the TSHb gene such a negative TRE has been found in close proximity to the AP-1 site which mediates the stimulation of TSHb gene transcription by TRH. As mentioned above, there appears to be a specific role for TRβ2 in the regulation of the negative TREs in the TSHb and TRH genes (70–72). In contrast to gene regulation through positive TREs, binding of TRβ2 to negative TREs in the absence of T₃ probably results in the activation of gene transcription. In the presence of T₃, transcription is inhibited. The exact mechanism of this negative regulation of gene expression by T₃ and any T₃ receptor is still unclear.

TSHb gene transcription is also strongly inhibited by 9-cis-retinoic acid, and this effect is mediated by the pituitary-specific RXRγ1 subtype, and involves both TRE-dependent and TRE-independent interactions with the TSHb gene promoter. The clinical relevance of this effect is underscored by a recent study showing that treatment of patients with T-cell lymphoma with bexarotene, another RXR-selective ligand, induces central hypothyroidism (75). It is also interesting to mention that the TRH gene promoter contains a glucocorticoid response element. Hypothalamic TRH-producing cells also express the glucocorticoid receptor, and the interaction of this receptor with its response element appears to mediate the inhibition of TRH synthesis by glucocorticoids (76).

In addition to the regulation of TSHβ and α-subunit gene expression, T₃ also acutely inhibits TSH secretion, the exact mechanism of which is still unresolved. Although T₃ is the active hormone exerting the inhibition of TSH production and secretion, serum T₄ appears to be a major player in the negative feedback regulation of the hypothalamus–pituitary–thyroid axis by acting as a precursor for local D2-mediated generation of T₃ at these central sites (43, 77).

Recent research in two particular areas has led to important advances in our understanding of the mechanism of action of T₃. One type of study has utilized T₃ receptor knockout and mutant mice in which one or more of the different T₃ receptor isoforms is deleted or mutated (72). These studies reveal which organ functions critically depend on the type of T₃ receptors they express. Much knowledge regarding the molecular mechanisms of T₃ receptor/T₃ action has also been gained from studies in patients with thyroid hormone resistance associated with mutations in the THRβ gene. For a thorough discussion of this subject, the reader is referred to Chapter 3.4.8.