Minireview: Thyrotropin Receptor Signaling in Development and Differentiation of the Thyroid Gland: Insights from Mouse Models and Human Diseases

IN 1927, UHLENHUTH (1) demonstrated that the anterior lobe of the pituitary gland of salamanders produces a factor capable of controlling the function of the thyroid. This factor was subsequently called TSH. Forty years later, Pastan and co-workers (2) demonstrated that TSH exerts its effects through a protease-sensitive structure on the thyroid cell plasma membrane, thus postulating the existence of a specific receptor (TSH-R; the current official name is Tshr for mouse genetic locus and TSHR for the human locus). The relevance of the interaction between TSH and its receptor in the physiology and pathology of the thyroid became immediately clear, and indeed in 1968 Stanbury (3) suggested that congenital hypothyroidism in the absence of goiter could be due to an impaired response to TSH. The mechanisms triggered by Tshr upon interaction with the ligand, which regulate both proliferation and function of thyroid cells, have been exhaustively studied mainly in cell culture models. Several comprehensive reviews summarize these aspects of TSH signaling (4–8). This review will focus on a different aspect of TSH signaling, its role in the development and differentiation of the thyroid, as determined by the study of alterations in this pathway in both mice and humans. Indeed, these studies have provided important insights about the physiological roles of TSH/Tshr signaling, in some cases underscoring the pivotal role that this system plays in regulating the size and function of the thyroid gland; in others demoting it to, at best, a marginal modulator. Furthermore, the animal models have been useful in discovering new important functions of such signaling systems in extrathyroidal tissues (9), which are beyond the scope of this review. Several mutant mouse lines that lack either a functional Tshr or TSH have been reported and described. Tshrhyt/Tshrhyt (formerly hyt/hyt) mice (10) are spontaneous mutant mice characterized by hypothyroidism and failure to respond to TSH. These mice have a point mutation in the coding region of the Tshr gene (11) that causes the replacement of a highly conserved proline (Pro556) in transmembrane domain IV with a leucine. This mutation leads to a defective binding with TSH (12). More recently, a mouse model in which the Tshr gene has been inactivated by homologous recombination in embryonic stem cells has been generated (13). As expected, both of these mouse types display hypothyroidism with thyroid hypoplasia. Another model of impaired TSH signaling is the pitdw/pitdw (formerly Snell dwarf or dw/dw) mouse (14). These hypothyroid mice carry a loss of function mutation in the sequence encoding the POU domain of the transcription factor Pit1. Mice homozygous for this mutation do not express TSH, GH, or prolactin (15). Furthermore, a mouse was generated in which the gene encoding the α-glycoprotein hormone subunit (α-GSU) had been disrupted (16). The α-subunit is common to the pituitary hormones TSH, LH, and FSH. As a consequence, α-GSU-null mice are hypothyroid and hypogonadal. It is important to stress that mice with mutations resulting in loss of the TSH receptor have a thyroid phenotype comparable with that of mice with mutations knocking out TSH. Small differences observed in the phenotype could be due to the diverse genetic background of the mice. This observation suggests that the constitutive activity of the receptor alone (17) is not sufficient to maintain the level of a physiological response in vivo. In addition, the thyroid phenotype of mutants in the TSH/Tshr pathway is not worsened by the absence of additional hormones as in the case of the pitdw /pitdw or α-GSU-null mice. This observation allows us to conclude that other pituitary hormones have no supplementary effect, either direct or indirect, on thyroid function. The expression of Tshr during thyroid morphogenesis in mammals has been mainly analyzed in rodents. Tshr mRNA is barely detected in the embryonic thyroid of mice and rats on embryonic d 14 (E14) (our unpublished observations; Fig. 1) and E15, respectively (18, 19). The level of expression increases on subsequent days of fetal life. The Pit1-dependent expression of TSHβ in the caudomedial cells of the pituitary begins on E15.5 (20), whereas α-GSU mRNA is expressed in the mouse pituitary from E12.5 (21). Information about expression during human development is more limited. The onset of TSHR expression in humans has not yet been reported. TSH is detectable in the pituitary (22) and serum (23) at approximately 12 wk gestation, and its levels increase during the second trimester of gestation (24). Onset of functional differentiation in mouse developing thyroid. A, Serial sagittal sections on E14 (A–D) and E15 (E–H) were hybridized with the indicated antisense probe. B, Serial sagittal sections on E14 (A and B) and E16 (C and D) were stained with the indicated antibodies. Tshr mRNA is first barely detected in the embryo thyroid on E14 and coincides with that of TPO. NIS expression begins on E16. [Reprinted from Postiglione et al. (35).] Although partial, these data are sufficient to determine that the early steps of thyroid organogenesis are independent of the TSH signaling pathway. Organogenesis of the thyroid gland is accomplished through a complex process. At an early stage of embryonic life (E9 in the mouse embryo), a group of endodermal cells, identified by the coexpression of four transcription factors, Titf1, Pax8 (18), Foxe1 (formerly called TTF-2) (25), and Hhex (26), detaches from the wall of the primitive pharynx and migrates toward the base of the neck. After completion of the migration of the primordium, by E14.5 the functional differentiation of the thyroid begins, and it is completed through both synthesis of a series of proteins that are essential for thyroid hormone biosynthesis and formation of the follicles. Activation of the TSH/Tshr pathway in the developing thyroid is coincident with the onset of thyroid hormone (TH) biosynthesis, and, therefore, it could be an important trigger for the expression of genes such as thyroglobulin (Tg), thyroid peroxidase (TPO), and sodium-iodide symporter (NIS) that encode proteins belonging to TH biosynthetic machinery. Because the expression of Tg, TPO, and NIS genes depends (at least in part) on the presence of the transcription factors Titf1, Pax8, and Foxe1 (27), it is plausible that the TSH/Tshr pathway acts through these transcription factors. In support of this view, several reports have shown that in vitro the TSH/Tshr pathway can regulate the expression of Titf1, Pax8, and Foxe1. Rat thyroid cells cultured in the absence of TSH express the transcription factor Titf1. This expression is down-regulated by the addition of either TSH or forskolin (28). This TSH-dependent down-regulation of Titf1 is due to a decrease in the transcription rate from the Titf1 promoter (29). However, in primary cultures of dog thyrocytes, there is no significant variation in the levels of Titf1 mRNA and protein in response to cAMP agonists (30). This discrepancy can be attributed to the differences between the two systems used (primary cultures vs. continuous cell line) or to species-specific differences in Titf1 gene regulation. Unlike Titf1, the steady state levels of Pax8 mRNA and protein are clearly increased after the addition of forskolin in both rat thyroid cells (29) and primary cultured dog thyrocytes (31). Foxe1 was identified as a thyroid-specific nuclear factor induced by TSH and forskolin (32). The transcription of Foxe1 strictly depends on the presence of either TSH or IGF-I or high doses of insulin in thyroid cells in culture (25, 33). The hormonal induction of Foxe1 mRNA levels required ongoing protein synthesis (33). However, in primary cultures of dog thyrocytes, the DNA-binding activity of Foxe1 is clearly detectable even in the absence of TSH stimulation (34). A detailed analysis of the gene expression pattern in the developing thyroid of the mutant mice described above has recently been reported (35) and has shown a rather different picture. Indeed, using this in vivo approach, it has been demonstrated that the absence of a functional Tshr or its ligand does not affect the expression of Titf1, Pax8, and Foxe1. Even in pitdw/pitdw mice, which lack TSH, GH and, as consequence of this, IGF-I, the expression of thyroid-enriched transcription factors is comparable to that in wild-type mice. These data strongly suggest that, at least during fetal life, the expression of Titf1, Pax8, and Foxe1 is not strictly dependent on TSH or IGF-I. However, it is still conceivable that these hormones regulate the expression of these transcription factors in vivo, but in their absence, another hormone or growth factor could trigger their transcription. In the developing thyroid, after the induction of Tshr, the expression of NIS begins (35) and the levels of both Tg and TPO increase (19). These observations are in agreement with several reports demonstrating, in different systems, that TSH/Tshr signaling positively regulates the synthesis of genes essential for thyroid physiology, such as NIS, TPO, and Tg. TSH stimulates iodide accumulation in rat thyroid cells through a cAMP-dependent mechanism (36). The TSH/Tshr pathway is involved in regulating the expression and activity of NIS at different levels: TSH up-regulates NIS protein in the rat thyroid (37) and the mRNA in rat thyroid cells (38). The TSH-induced stimulation of NIS expression requires protein synthesis (38). In rats, an enhancer that recapitulates the most relevant aspects of NIS regulation has been identified (39). A strong increase in the expression of NIS mRNA has been reported in dog thyroid and human thyroid primary cultures after goitrogenic treatment (40–42). In addition to the regulation of NIS transcription and biosynthesis, TSH/Tshr signaling is also required to target NIS to the plasma membrane (43). The phosphorylation of specific residues dependent on the presence of TSH might be involved in regulation of the subcellular distribution of NIS (43). The regulation of TPO by TSH is mostly at the mRNA level. In dog thyroid primary cultures, TSH or cAMP agonist induces TPO mRNA accumulation by an increase in the gene transcription rate. This effect is rapid and does not require protein synthesis, suggesting that the TPO promoter can be directly controlled via cAMP regulatory elements (44, 45). On the contrary, in rat thyroid cells in culture, TSH has been shown to increase the steady state TPO mRNA level (46, 47) by indirect mechanisms. Regulatory sequences responsive to TSH and cAMP signals have been identified at the 5′-flanking region in functional studies of both human (48) and rat (32) TPO promoters. Animal models deficient in TSH/Tshr signaling are an excellent tool to confirm the importance of TSH in the control of NIS and TPO expression in the whole organism. In Tshrhyt/Tshrhyt, Tshr-null, and pitdw/pitdw mice, the expression of some thyroid-specific genes has been carefully analyzed at the end of thyroid organogenesis (35) and in adult life (13). NIS and TPO are almost undetectable in mutant thyroids on E17 (Fig. 2). The same result was observed in the thyroids of adult Tshr-null mice, which fail to express the NIS protein. TSH signals control the expression of NIS and TPO in mouse developing thyroid gland. Wild-type (A–C), Tshrhyt/Tshrhyt (D–F), pit1dw/pit1dw (G–I), and TSHR-KO (J–L) E17.5 embryos were stained with anti-Tg (A, D, G, and J), and anti-NIS (B, E, H, and K), or hybridized with TPO antisense probe (C, F, I, and L). In the absence of functioning TSH/Tshr signaling, the expression of both NIS and TPO was almost abolished, whereas Tg appears only slightly decreased. [Reprinted from Postiglione et al. (35).] Notably, several patients affected by congenital hypothyroidism due to loss of function mutations in the TSHR gene have been described. In these subjects, 99Tc scintigraphy fails to demonstrate the presence of thyroid tissue, whereas ultrasonography reveals hypoplastic thyroid tissue (49–51). The phenotype is consistent with down-regulation of the NIS gene. Along the same lines is the report that in Graves’ disease thyroid tissues and in toxic adenomas, both NIS and TPO mRNA are increased (52, 53). These findings strongly suggest that in humans also, TSH signals control the transcription of NIS and TPO genes. The influence of TSH on Tg has been known for many years (54). The role of TSH signals in Tg expression has been extensively investigated in different systems, and there is a general consensus that TSH increases the level of Tg mRNA and that cAMP is the physiological mediator of this effect (55, 56). Interestingly, the extent of this regulation is variable in these different reports. In rat (57, 58) and porcine (59) thyroid cells, TSH approximately doubles the rate of transcription of Tg. In hypophysectomized rats (60), TSH control seems to be tighter. However, recent experiments in mice carrying loss of function mutations in Tshr suggest that Tsh/Tshr signaling has a moderate effect on Tg expression in vivo (see below). The increase in Tg mRNA transcription in response to TSH or a cAMP agonist is slow. Furthermore, cycloheximide inhibits the increase in Tg expression (44), indicating that TSH induction requires newly synthesized proteins. These data suggest that TSH uses different mechanisms to control TPO and Tg synthesis (44, 46, 61). The effect of TSH on the Tg promoter is not direct, as observed for the TPO promoter, but is mediated by other factors that are the primary targets of cAMP induction (44, 61). Consistent with these data is the fact that no canonical cAMP response elements have been identified in the 5′ region of the gene. The mediators of the TSH signal on the Tg promoter are still unknown. Pax8 has been hypothesized to be a link between the TSH pathway and Tg transcription. However, it has been shown that Pax8 alone is not sufficient to reactivate the expression of Tg in the absence of TSH stimulation (62). The expression of Tg has been studied in the animal models described above at the end of thyroid organogenesis (35) and in adult mice (13). Unlike both NIS and TPO, which are absent in E17 mutant thyroid, Tg appears to be only slightly decreased in mutant mice compared with wild-type animals. Also in the adult mutant, the expression of Tg is only barely affected. These studies in animal models might be consistent with data obtained from patients with inactive TSHR that present relatively normal circulating Tg levels. In addition, no increased Tg mRNA has been reported in either Graves’ disease or toxic adenomas (52, 53). In conclusion, the studies that we have summarized allow us to state two important concepts. The first is that there is not a tight control of TSH over Tg expression, and thus TSH is not the major regulator of Tg transcription. The second is that the known role of Tshr signaling in Tg iodination is based, besides the well established activation of H2O2 generation (63, 64), on the coordinated and tight control of NIS and TPO expression. It should be of great interest to discover whether other players of this pathway are also under TSH control. The developing thyroid begins to expand significantly in size only after the end of migration from the pharyngeal cavity, at the same stage at which follicular cells begin to synthesize TH. By E15–16 in mice and E70 in humans, the gland exhibits its definitive shape, two lobes connected by a narrow isthmus. It is worth noting that the onset of this expansion coincides with the appearance of TSH in serum. The growth of the thyroid continues after birth. In mice, the weight of the thyroid increases during the first month of life (65). In humans, the growth of the thyroid roughly parallels body growth during childhood (66). At the adult stage, the estimated turnover time for follicular cells, using in vitro labeling experiments, is approximately 8.5 yr (67). TSH, mainly using cAMP as a second messenger, induces cell proliferation in rat thyroid cell lines and primary cultures of thyrocytes from different species. However, studies of different models demonstrated that in the thyroid other distinct mitogenic cascades are present and promote cell proliferation. The most relevant ones are the pathways stimulated by growth factors such as IGF-I or epidermal growth factor. Furthermore, a third cascade, the phospholipase C cascade, induced by tumor-promoting phorbol esters or α1-adrenergic agonist, has been shown to enhance the proliferation and dedifferentiation of thyroid cells (4). In all systems, TSH is the main regulator of thyroid growth, whereas IGF-I (or insulin at supraphysiological concentrations) is required for the mitogenic action of TSH (4). Indeed, IGF-I/insulin alone does stimulate the proliferation of either dog or human thyroid cells in primary culture, whereas rat thyroid cell lines can weakly proliferate in response to IGF-I/insulin alone (4). These data have been confirmed in vivo by the comparison of the thyroid phenotype of two transgenic mice, one overexpressing in thyroid the A2 adenosine receptor (68), which causes constitutive activation of adenylyl cyclase, and the other characterized by simultaneous overexpression in the gland of IGF-I and its receptor (69). In adult A2 adenosine receptor transgenic mice, the dramatic enlargement of the gland, whose weight increases more than 100 times, is due to strong hyperplasia of the thyroid. Indeed, bromodeoxyuridine incorporation reveals that the proliferation of thyroid cells continues throughout life. On the contrary, transgenic mice overexpressing IGF-I/IGF-I receptor show mild enlargement of the thyroid as a consequence of hypertrophy of the gland without an increase in the cell population. The gland displays an increased follicular lumen area and a slight decrease in the number of cells per square millimeter compared with the wild-type thyroid. In humans, IGF-I has little mitogenic effect; actually patients with acromegaly display enlarged thyroids (70). However, IGF-I signaling in humans in vivo seems to be required for the mitogenic action of TSH, as demonstrated by the low endemic goiter prevalence among pygmies (71). In support of the fact that TSH/Tshr signaling per se has a relevant role in thyroid growth in adults is the thyroid phenotype displayed by all animal models carrying natural or induced mutations in Tshr or its cognate ligand (13, 14, 16, 72). All mutants are hypothyroid and show a hypoplastic adult thyroid characterized by small sparse follicles, areas not organized into follicles, and a reduced number of cells. Although these data confirm the mitogenic effect of the TSH-induced cAMP pathway in the adult thyroid, the growth control of the fetal thyroid is still puzzling. Pioneering works in reported that the thyroid in or in hypophysectomized More recently, it has been reported that the size of the thyroid in both Tshrhyt/Tshrhyt and pitdw/pitdw mice on E17 is to that in wild-type mice Furthermore, the number of cells in the thyroid is comparable in mutants and wild-type In addition, the of mutant thyroids is not affected by the absence of a functional Tshr, indicating that is Although these data do not a role for TSH in controlling the growth of the embryonic thyroid, suggest that in thyroid gland development the proliferation of thyrocytes could be controlled by a Consistent with this is the observation that in A2 adenosine receptor transgenic mice (see the of the thyroid at is comparable to that in wild-type mice. These data the of might regulate the expansion of thyroid cells in the developing thyroid. At the we can only suggest some growth through its has been demonstrated to promote the proliferation of rat and dog thyroid cells in This factor and its receptor are both expressed in the thyroid, thus suggesting that a regulation could be involved in controlling the growth of the thyroid during fetal life in vivo. Another regulator could be a of the growth factor has been reported to synthesis in thyroid cells in of the has been detected in the developing thyroid This suggests that thyroid cells are to respond to present in the which has mitogenic activity in other such as the developing pituitary and In in vitro only signals are to trigger proliferation and differentiation whereas the other pathways proliferation and This observation could signals are in a functional thyroid in whereas can be by other signals during In thyroid organogenesis is by but the gland continues to In addition, the functions of the are accomplished in humans at but only after in rodents. This suggests that the data from the mouse models be directly to the The mechanisms controlling the growth of fetal thyroid could be different in the different species. In conclusion, recent data with the that signaling is essential for thyroid function. However, as as gene expression is only a of the genes involved in TH biosynthesis is controlled by TSH, thus indicating that there is no control of the thyroid differentiation Furthermore, thyroid gland size seems to be controlled during life an and in the adult these are in mice and humans to be for on the This was in by with Animal and per of and for in and growth α-glycoprotein hormone sodium-iodide thyroid thyroid TSH