Werner & Ingbar's The Thyroid: A Fundamental & Clinical Text, 9th Edition

10C.Biological Actions of Thyrotropin

Stephen W. Spaulding

Thyrotropin (TSH) is the major regulator of thyroid hormone synthesis and secretion. TSH is best known for actions that are mediated by cyclic 3′, 5′-adenosine monophosphate (cyclic AMP), but TSH regulates many other signaling pathways. When stimulated, TSH receptors usually couple with large G-protein αβγ-heterotrimers that activate adenylyl cyclase, but TSH receptors can also associate with αβγ-heterotrimers that stimulate small guanosine triphosphatases, phospholipid metabolism, phosphatidylinositide 3 (PI3)-kinase, and other signaling pathways. After activation, various pathways can be down-regulated, permitting tight control of hormone production as well as of growth and death of thyroid follicular cells (thyrocytes). Many of the signaling pathways in thyrocytes are also regulated by other hormones, growth factors, and cytokines, in addition to receiving signals from the colloid, from adjacent cells, and from the extracellular matrix. Studies of normal rodents have provided much understanding of the actions of TSH, and studies of gene-knockout mice are providing new insights, but certain aspects of signaling, cell regeneration, and death differ between rodents and humans (1,2). Some responses to TSH vary between species and even between strains of animal (3). In vitro models can provide information not otherwise obtainable, but paracrine factors and direct connections with neighboring cells and structures are missing or distorted, and there is substantial variation in TSH responses between in vitro models. It is important to emphasize, however, that pathways found to be regulated by TSH in such experimental models have proven to be important clinically as causes of thyroid autonomy, and of benign and malignant tumors.


Although TSH is a tonic regulator of growth and function of the adult thyroid, it is not required for early thyroid development (see Chapter 2). Before any TSH or TSH receptors are expressed in the embryo, thyroid precursor cells express a combination of three transcription factors found uniquely in thyrocytes: TTF-1 (thyroid transcription factor-1 or Titf1), TTF-2 (or Foxe1), and Pax-8 (paired box gene 8). TTF-1 and Pax-8 are phosphoproteins required for the development of thyrocyte precursor cells, while TTF-2 is needed for correct migration of the cells. During normal thyroid development, TSH and its receptor first appear at about the 16th embryonic day in rats and mice, after which small follicular structures appear and thyroid hormone production begins. In mutant mice in which the TSH receptor has been “knocked out,” full maturation of the thyroid is impaired, although some follicle formation occurs in late embryonic life. These mice survive only if they are treated with thyroid hormone (4). The thyroid vestigium in these animals contains very little sodium-iodide symporter protein or thyroid peroxidase, but noniodinated thyroglobulin is present. If forskolin is added to normal thyroid precursor tissue from 15-day rodent embryos in order to directly activate adenylyl cyclase and increase cyclic AMP production, follicular structures form and iodine-concentrating activity is induced prematurely (5). Similarly, if the thyroid vestigium from TSH-receptor knockout mice is incubated with forskolin, some organified iodine can be detected, indicating that signaling mechanisms required for iodide uptake and organification have been preserved, even in the absence of a TSH receptor, and confirming the importance of cyclic AMP for the normal function of the thyroid (6).

In addition to playing essential roles in thyroid development, TTF-1 and Pax-8 are involved in regulating the expression of several thyroid-specific genes. In FRTL-5 (Fisher rat thyrocyte cell line-5) cells, TSH decreases TTF-1 expression and down-regulates expression of the TSH receptor gene—whose promoter contains TTF-1 binding sites but no Pax-8 binding site—but only if serum and insulin are present in the culture medium (7). In contrast, TSH increases Pax-8 expression and up-regulates the expression of the thyroglobulin gene—whose promoter contains an overlapping binding site for both Pax-8 and TTF-1—regardless of whether serum and insulin are present (7). TTF-1 and Pax-8 can exist as a complex and act synergistically to promote transcription from the rat thyroglobulin gene promoter in rat PCCl3 thyrocytes (8).

Patients with a mutation in one TTF-1 allele have high serum TSH concentrations, and some are overtly hypothyroid. Similarly, mice with one defective TTF-1 allele also have high serum TSH concentrations (9). Despite the high levels of TSH, their serum thyroxine (T4) levels remain somewhat low, but administering exogenous TSH can normalize the T4 (9). Their thyroid glands contain diminished levels of TSH receptor and thyroglobulin messenger RNA (mRNA), whereas expression of other thyroid-specific genes is not significantly decreased. Thus biallelic expression of TTF-1 is necessary for the thyroid to be fully responsive to TSH (9). Although both TTF-1 and Pax-8 are phosphoproteins, and cyclic AMP and protein kinase A (PKA) are involved in their regulation (10), it appears that the effect of PKA may not be mediated by directly phosphorylating them (11).

Pax-8 and TTF-1 activities are influenced by their redox state (12,13). In FRTL-5 cells, TSH increases the DNA-binding activity of endogenous Pax-8 and TTF-1 by reducing their oxidized forms, accompanied by increased expression of thioredoxin (13). Thioredoxin is a major intrathyroidal antioxidant that restores binding activity to these and other redox-regulated transcription factors (13). Ref-1 (redox factor-1) repairs DNA damaged by oxidative stress, and also regulates the redox state and DNA-binding activity of transcription factors like AP-1, NF-κB, and p53, in addition to protein kinase A TTF-1 and Pax-8. Ref-1 associates with thioredoxin in the nucleus, reducing oxidized Pax-8 and increasing its binding to the promoter regions of the sodium-iodide symporter and thyroglobulin genes (12,14). TSH increases Ref-1 levels in FRTL-5 cells by a cyclic AMP-mediated mechanism (15,16), and also causes translocation of Ref-1 from the cytoplasm to the nucleus of FRTL-5 cells (16) Clinically, histochemical studies indicate that Ref-1 levels are high in hyperfunctioning thyroid adenomas, particularly in the nuclear fraction (16). Some of the effects of TSH on AP-1 target genes may be mediated by actions of Ref-1 on the redox state of the different components that can comprise AP-1 dimers (15,17). Ref-1 may also guard cells from apoptosis by protecting them from oxidative stressors like hydrogen peroxide (H2O2), which may be another response regulated by this mediator of TSH action (see later in the chapter).


When TSH binds to its transmembrane receptor, it induces a change in the receptor that activates αβγ heterotrimers on the inner surface of the basolateral plasma membrane of thyrocytes (Fig. 10C.1). Activation causes a guanosine triphosphate (GTP) molecule to replace the guanosine diphosphate (GDP) molecule that is bound to the Gα-subunit associated with a Gβγ-dimer in the basal state. The activated GTP-containing Gα-subunit then dissociates from the Gβγ-dimer, permitting the free Gα-subunit and membrane-bound Gβγ-dimer to interact with various downstream effectors. The specific pathway that is activated by a given TSH molecule is determined by the particular α-, β-, and γ-subunits that comprise the heterotrimer to which the TSH receptor has coupled. The duration of the signals produced by a given Gα-subunit could simply depend on when its intrinsic GTPase hydrolyzes the GTP back to GDP, which inactivates the Gα-subunit and permits it to re-associate with whatever Gβγ-dimer is available. However, other proteins can regulate both the rate at which GTP initially displaces the GDP from an inactive Gα-subunit, as well as the rate at which the GTP is hydrolyzed. The dose and duration of the TSH stimulus, the extent of glycosylation of the TSH, and the structure of the receptor also affect the responses. If the duration of TSH stimulation is protracted, the thyroid may react by decreasing the expression of TSH receptors and heterotrimer components. When G-coupled receptors are occupied by ligand, they can be phosphorylated by G-protein receptor kinases, which can down-regulate the response to subsequent stimuli by permitting the binding of arrestins that promote internalization and degradation of the receptors. Uncoupled G-protein heterotrimers are also able to interact with other receptors in thyroid plasma membranes, such as those for catecholamines, lysophosphatidic acid, prostaglandins, and cholinergic and purinergic agents (18).

FIGURE 10C.1. Regulation of the thyrotropin (TSH) receptor and G-protein heterotrimers. Top diagram: In the basal state, the TSH receptor does not contain a ligand, while the inactive Gα-subunit in the heterotrimer contains a guanosine diphosphate (GDP) molecule. Bottom diagram: TSH binds to and activates the receptor, and causes guanosine triphosphate (GTP) to replace the GDP on the Gα-subunit. The activated Gα-subunit dissociates from the Gβγ-subunit dimer, which then interacts with various downstream effectors. The heterotrimer shown in the diagram contains a GαS-subunit, which activates adenylyl cyclases. However, other heterotrimers may contain Gαi/0-subunits that can inhibit adenylyl cyclases, Gαq/11-subunits that can activate phospholipases, or GαS12/13-subunits that can activate phospholipases or guanosine exchange factors that activate small GTPases. An activated receptor is susceptible to phosphorylation by receptor kinases, which then permits arrestin proteins to bind that promote receptor internalization and degradation.


Under basal conditions, a stimulatory Gα protein (GαS) is the predominant Gα isoform in the thyroid. GαS stimulates adenylyl cyclases and increases production of cyclic AMP, which not only activates cyclic AMP-dependent PKAs, but also independently activates guanine nucleotide exchange proteins activated by cyclic AMP (EPAC) proteins that regulate certain small GTPases. Cyclic nucleotides are degraded by phosphodiesterases. One action of TSH is to increase the activity of cyclic AMP-specific phosphodiesterases, which creates a negative feedback loop that limits the cyclic AMP response to TSH as well as to other factors that stimulate cyclic AMP production in the thyroid. Clinically, patients who have autonomous thyroid adenomas that contain an activating mutation—either in the TSH receptor gene or in the Gαs gene—commonly express a higher level of cyclic AMP phosphodiesterase activity in the tumor than is present in the surrounding normal thyroid tissue (19), which suggests that some cells in the adenomas retain the homeostatic tendency to limit cyclic AMP-dependent responses. In FRTL-5 cells, TSH rapidly increases phosphodiesterase activity via cyclic AMP-dependent phosphorylation of the enzyme (20), while longer exposure to TSH increases the transcription of phosphodiesterases as well (21). Transfecting these cells with a constitutively active mutant GαS causes phosphodiesterase-4 activities to increase, as determined by the selective inhibitor rolipram (22). Rolipram has no effect on the proliferation of wild-type cells, but in cells expressing the constitutively active mutant GαS, rolipram causes TSH-independent proliferation, indicating that cyclic AMP plays a key role in the proliferation of FRTL-5 cells (22). It is important to note, however, that “immortalized” thyrocyte cell lines may respond to TSH differently than thyrocytes grown in primary culture, to say nothing of thyroid tissue in vivo. Thus, for example, FRTL-5 cells are known to dedifferentiate with increasing number of cell passages (23), may be tetraploid (24), and can harbor mutations in important genes (23,25).

Cyclic AMP activates cyclic AMP-dependent PKAs by binding to the regulatory subunit dimer in the holoenzyme (Fig. 10C.2). The dimerization region on R-subunits also permits them to bind to AKAPs (A-kinase anchoring proteins), which are scaffolding proteins that tether different PKA isoforms in specific locations within the cell. When the four cyclic AMP, binding sites on the R-subunit dimer are occupied, catalytic C-subunits are released, which then incorporate phosphate groups on certain serine or threonine residues in neighboring substrates, modulating their interactions with other proteins. The CREB, CREM, and ATF-1 family of transcription factors are important nuclear substrates of PKA, although they are phosphorylated by other kinases as well. Phosphorylation by PKA permits CREB dimers to bind to cyclic AMP-response elements (CREs) in various genes, and to recruit components of the transcription complex, thereby activating cyclic AMP-responsive genes. Inducible cyclic AMP early repressor (ICER) is a cyclic AMP-inducible antagonistic transcription factor that is produced from an alternative promoter in the CREM gene, and which down-regulates its own expression. In hypophysectomized rats, injection of TSH initially raises the level of TSH receptor mRNA, but it falls as ICER mRNA levels peak (26). TSH also transiently increases ICER levels in primary dog thyrocytes, which correlates with down-regulation of TSH receptors (27).

FIGURE 10C.2. Regulation of cyclic and cyclic AMP-dependent protein kinases (PKA). Top diagram: In the basal state, PKA heterotrimers consist of a regulatory subunit dimer and two catalytic (Cat) subunits, and may be tethered in specific subcellular locations by binding to A-kinase anchoring proteins (AKAP). Bottom diagram: Thyrotropin (TSH) activates adenylyl cyclase and converts adenosine triphosphate to cyclic AMP. When all the cyclic AMP binding sites on PKA are occupied, catalytic subunits are released so they can phosphorylate nearby substrates. Subsequently, TSH activates cyclic AMP-specific phosphodiesterases that degrade the cyclic AMP, allowing catalytic subunits to reassociate with regulatory subunit dimers.

Clinically, some thyroid adenomas bear somatic mutations in the TSH-receptor gene that act in a dominant fashion to activate the receptor in the absence of ligand (see Chapter 25). However, mutations that inactivate the receptor act in a recessive fashion, so both alleles of the gene must be affected before clinical TSH resistance or hypothyroidism occur (see Chapter 48). Mutations in the TSH-receptor gene may alter the basal level of activity of the receptor or the ability of the receptor to couple with certain G α-subunits, selectively affecting, for example, the ability of TSH to generate inositol phosphate signals (28,29).

When a dominant negatively acting isoform of CREB is experimentally targeted to the mouse thyroid, the mice are born hypothyroid (10). The thyroid glands of these mice are poorly developed despite the high serum TSH levels, and they express only low levels of the mRNAs encoding the TSH receptor, thyroglobulin, thyroid peroxidase, and the transcription factors Pax-8, TTF-1, and TTF-2, indicating the importance of CREB in both the growth and the function of the thyroid (10). The state of CREB phosphorylation also mediates thyrocyte responsiveness to TSH stimulation. When TSH-starved FRTL-5 cells are given TSH, there is an initial burst in cyclic AMP-dependent transcription that then decreases gradually over a period of hours, reflecting a progressive dephosphorylation of CREB by protein phosphatase 1 (30). More prolonged exposure to TSH causes decreased translation of PKA catalytic subunits, further desensitizing the response to cyclic AMP (30). Finally, chronically high levels of TSH decrease the density of TSH receptors in the thyrocytes of rats in vivo, as well as in FRTL-5 cells in vitro (31).

Still another way that a change in the cyclic AMP responsiveness of thyrocytes will occur is if TSH activates a receptor that couples with a G-protein heterotrimer containing a Gαi/o-subunit, which inhibits adenylyl cyclases (32).


In addition to regulating cyclic AMP pathways, TSH receptors can couple with G-protein heterotrimers containing Gαq/11-subunits, which activate phospholipases that generate inositol phosphates and diacylglycerol, which in turn activate Ca++/phospholipid-dependent kinases (PKC). Higher concentrations of TSH are needed to activate these Gα-isoforms than are needed to activate GαS. If a receptor couples with a heterotrimer containing a Gα12-subunit, it can affect small GTPases by interacting with proteins that activate guanosine exchange factors, as well as other pathways (33). It is important to note that when TSH stimulates the release of any of the Gα-isoforms, it simultaneously releases Gβγ-dimers that then interact with other proteins. The Gβγ-dimers, which are prenylated and remain membrane-bound, can activate certain phospholipases, phosphatidylinositol-3-kinases (PI3-kinases) and small GTPases.


Small guanosine triphosphatases (GTPase) are molecular switches that regulate membrane trafficking of receptors, transporters, and ion channels, as well as regulating the activity of many kinases (Fig. 10C.3). Clinical and experimental data show that several small GTPases play important roles in thyroid differentiation, proliferation, apoptosis and tumorogenesis. Inactive small GTPases contain GDP, but if the GDP is replaced by GTP, the proteins become active. The switch to the active state is regulated by guanosine exchange factors that promote the displacement of GDP by GTP (34), while inactivation is enhanced by GTPase-activating proteins. Several hundred of these regulatory molecules have effects on different small GTPases. Their activities in turn can be regulated by cyclic AMP, Ca++, diacylglycerol, and redox conditions, as well as by other small GTPase regulators. Small GTPases often undergo prenylation, which promotes their association with membranes.

FIGURE 10C.3. Regulation of small guanosine triphosphatase (GTPase) proteins. Small GTPases are commonly membrane-associated due to their covalent modification by various kinds of prenylation (indicated with a trident). In the basal state, a small GTPase contains a guanosine diphosphate (GDP) molecule. Activation occurs when GTP displaces the GDP, which can be promoted by various guanosine exchange factors. Activated small GTPases can recruit factors to membranes, and can activate kinases and transporters. Subsequently, the protein's intrinsic GTPase activity hydrolyzes the GTP back to GDP, which can be accelerated by various GTPase activating factors.

Ras small GTPases are protooncogenes located predominantly in membranes, and are of clinical importance, since mutations in Ki-Ras, N-Ras, and Ha-Ras are found in multinodular goiters, microfollicular adenomas, and papillary and follicular carcinomas (35,36). When activated, Ras proteins regulate Raf kinase as well as PI3-kinase–signaling cascades. Posttranslational modifications of Ras proteins affect their association with other proteins, with intracellular membranes, and the signaling pathways they regulate. Experimentally, when an oncogenic Ha-Ras mutant is expressed in normal human thyrocytes, it causes proliferation but does not affect differentiated function (37). In rat thyroid cell lines, however, experimental expression of a constitutively active Ras causes apoptosis (38) and dedifferentiation (37,39). In FRTL-5 cells, TSH induces proliferation via PKA-mediated phosphorylation of Ras, which stimulates its interaction with PI3-kinase (40). Cyclic AMP also stimulates proliferation of Wistar rat thyrocytes. Active GTP-Ras must be present before TSH can stimulate proliferation of these cells, and this response is independent of PKA and tyrosine kinase activities (41,42). TSH and cyclic AMP also stimulate proliferation of cultured dog thyrocytes; however, both agents actually reduce the level of GTP-Ras-GTP (43). These varied effects of TSH on Ras may reflect species and strain differences reported in Ras and in the proteins that interact with Ras.

The Rap proteins (Rap-1a, -1b, and Rap-2) are small GTPases (Fig. 10C.4) that affect tyrosine kinases and integrin-mediated cell adhesion. The Rap phosphoproteins are homologous to Ras, and they bind to many of the proteins to which Ras binds, but counteract some of the transforming effects of Ras. TSH activates Rap activity in a number of thyroid cell lines. Expressing a constitutively active Rap-1a mutant in Wistar rat thyrocytes causes differentiation, while cells expressing a dominant negative mutant Rap-1 no longer proliferate in response to TSH (41). The activation of Rap-1 is mimicked by cyclic AMP but does not require PKA, although PKA does shorten the duration of Rap-1 activation, whereas blocking PKA prolongs the activation of Rap-1 by TSH (41). PKA stabilizes Rap-1GAP, a GTPase activator of Rap that is a tumor suppressor, and if TSH is removed from Wistar rat cells, the level of Rap-1GAP falls (41,42). In PCCl3 rat thyrocytes that over-express Rap-1GAP, PKA no longer can inhibit Akt (42a). In FRTL-5 cells, TSH also increases Rap-GTP levels by a cyclic AMP-mediated effect that does not involve PKA (43). TSH and forskolin activate Rap in primary cultures of dog thyrocytes, however carbachol, phorbol esters, insulin, and epidermal growth factor also have the same effect, and these agents do not activate PKA (43). Furthermore, microinjecting Rap-1b alone or in combination with a catalytic subunit of PKA into dog thyrocytes does not affect their proliferation or differentiation (43). A synthetic transgene that can be switched from a constitutively active form of Rap1b to a dominant negative form has been targeted to the mouse thyroid (43a). The constitutively active transgene perchlovate for 6 months causes nodules to form in the hyperplastic goiters that normally occur. These nodules express the constitutively active protein. If the goitrogens are removed, those findings reversed within 2 months, but if the goitrogens are continued for a year, frank invasion of the capsule, blood vessels and surrounding tissue occurs. Switching the transgene to the dominant negative form reduces the size of goiters produced by the goitrogens and the amount of cell proliferation, indicating Rap1b can have an oncogenic effect when linked to the mitogenic action of chronically elevated cyclic AMP levels (43a). While TSH and cyclic AMP do influence Rap activity, it is affected by other pathways as well. The non-PKA mediated effects of cyclic AMP on Rap may be mediated by EPAC proteins, which bind cyclic AMP directly, and which promote the displacement of GDP by GTP and thus activate Rap. EPAC-1 mRNA is known to be expressed at high levels in the thyroid (34), while EPAC-2 can be expressed in several isoforms due to alternative mRNA splicing, so the relative importance of different EPAC-mediated effects could underlie some of the differences in the actions of cyclic AMP that are not mediated by PKA.

FIGURE 10C.4. Regulation of the small guanosine triphosphatase (GTPase) Rap-1. Activation of the small GTPase Rap-1 causes membrane recruitment and activation of various kinases and other small GTPases, and alters membrane signaling. GTP-Rap-1 levels can be increased via guanosine exchange factors, and can be decreased by GTPase activating proteins such as Rap-1 glyceraldehyde-3-phosphate (GAP). TSH and cyclic adenosine monophosphate (cyclic AMP) activate Rap-1 in Wistar rat thyrocytes, presumably via guanine nucleotide exchange proteins activated by cyclic AMP and not via protein kinases (PKA), although PKA can phosphorylate Rap-1. TSH stabilizes Rap-1 GAP, and this effect is mediated by PKA, possibly via its action on phosphoprotein phosphatase. If TSH is removed from the medium, Rap-1 GAP levels decrease due to phosphorylation by glycogen synthase kinase 3β, which causes proteasomal degradation of Rap-1 GAP.


The Raf protein kinases (e.g., B-Raf and Raf-1) are phosphoproteins that become recruited to membranes by binding to activated small GTPases like GTP-Ras (Fig. 10C.5). The phosphorylation and dephosphorylation of Raf regulates its affinity for GTP-Ras, although phosphoprotein-binding proteins can regulate the access of GTP-Ras to these phosphorylated sites. Once Raf has been recruited to the membrane, its activity is regulated by membrane receptors that signal through kinases such as PI3-kinase and Akt. Activated Raf is a mitogen-activated protein (MAP) kinase kinase kinase (MAPKKK) that phosphorylates and activates MEK (MAPK/ERK kinase or MAPKK), which in turn activates MAP kinases. Many papillary thyroid carcinomas contain a specific mutation in B-Raf, indicating obvious clinical importance. In FRTL-5 cells, forskolin activates and recruits B-Raf to membranes, where it co-localizes with Rap-1 and subsequently activates ERK (44). Interestingly, these effects of forskolin are not mediated by PKA (44).

FIGURE 10C.5. Regulation of Raf kinases. The Raf protein kinase family can be recruited to the plasma membrane by activated small guanosinetriphosphatases, like Rap and Ras. The ability of Raf to be recruited depends on the state of certain phosphorylated sites on Raf, and access to them can be blocked if the sites are occupied by phosphoprotein-binding proteins. Once Rafs have been recruited to membranes, they are activated by various kinases, and in turn activate nitrogen-activated protein (MAP) kinase cascades.


PI3-kinases are heterodimers that can contain one of several possible noncatalytic subunits that suppress the activity of a catalytic subunit (Fig. 10C.6). One type of noncatalytic subunit contains SH-2 or SH-3 domains that bind to phosphotyrosine residues produced when membrane tyrosine kinase receptors are activated, while another type of noncatalytic subunit of PI3-kinase binds to the Gβγ-dimers that become available on the plasma membrane when G-coupled receptors are activated. Since PI3-kinases can be activated both by G-protein–coupled receptors (like TSH receptors) and receptor tyrosine kinases (like insulin and IGF-1 receptors), these two kinds of signals intersect, and differences in signaling responses can arise (45). Once PI3-kinase holoenzymes are bound at the membrane, their conformation changes and the activated catalytic subunit then phosphorylates proteins and lipids in the plasma membrane. The lipid kinase activity increases the membrane levels of several phosphatidylinositol 3′phosphates, which have differing affinity and selectivity for proteins with pleckstrin homology domains. Proteins that contain these domains include Akt, phosphoinositide-dependent kinase-1, and phospholipase-C isoforms, which can also be activated once they become membrane associated. The major lipid phosphatase responsible for down-regulating the signaling of phosphatidylinositol phosphates is PTEN, which is a tumor suppressor as well as a regulator of cell migration and cell size. Clinically, PTEN is silenced in several types of thyroid tumor.

FIGURE 10C.6. Regulation of phosphatidylinositide-3-kinases (PI3-kinases). In the basal state, PI3-kinases are heterodimers containing a catalytic subunit (Cat) that is suppressed by a noncatalytic subunit. When TSH stimulates a receptor, it releases Gβγ dimers that can recruit one type of noncatalytic subunit to the plasma membrane and activate its associated catalytic subunit. Membrane receptor tyrosine kinases can recruit another type of noncatalytic subunit to the plasma membrane and also activate its associated catalytic subunit. The activated PI3-kinase catalytic subunits then phosphorylate proteins as well as lipids in the plasma membrane. By increasing the levels of phosphatidylinositol-3-phosphate molecules in membranes, PI3-kinases attract proteins with pleckstrin homology domains (shown as three tridents), permitting those proteins to be activated as well. Subsequently, the lipid phosphatase PTEN turns off this signal.

PI3-kinases mediate various actions of TSH in FRTL-5 and Wistar rat thyrocyte cell lines (40,45,46,47,48,49,50). Since PI3-kinases can phosphorylate proteins like insulin regulatory substrate-1, TSH can activate pathways that are also regulated by insulin and insulin-like growth factor 1 (IGF-1) (45,46). TSH also induces IGF-1 production in FRTL-5 cells, as well as potentiating IGF-1–mediated proliferation via PI3-kinase, tyrosine kinase, and MAPK pathways (45,46). Anti-TSH receptor–immunocomplexes prepared from FRTL-5 cells contain PI3-kinase activity, and treating the cells with TSH increases the amount of PI3-kinase precipitated, as well as the amount of phosphotyrosine present. Additional evidence for the important role played by PI3-kinase in mediating the action of cAMP in these cells was provided by the use of the PI3-kinase inhibitor wortmannin, which blocked cyclic AMP-mediated phosphotyrosine production (46,47). PKA stimulates formation of a complex between PI3-kinase and Ras (40), and increases the phosphorylation and amount of PI3-kinase regulatory subunit that precipitates with anti-TSH receptor antibody in FRTL-5 cells (47). IGF-1 and TSH are synergistic in causing FRTL-5 cell proliferation, but in contrast, IGF-1 blocks the induction of sodium-iodine symporter mRNA by TSH and cyclic AMP (48). The inhibitory effect of IGF-1 on the symporter can be eliminated by blocking PI3-kinase, but is not affected by inhibiting PKA, PKC, or MEK (48). In Wistar rat thyrocytes, PI3-kinase is required for cyclic AMP to stimulate mitogenesis, and this involves both PKA-dependent and -independent pathways (49,50). Although TSH amplifies insulin and IGF-1 signaling in primary cultures of dog thyrocytes, it does not change PI3-kinase activity (51).

One of the kinases with a pleckstrin domain that is attracted by increased phosphatidylinositol phosphate levels in membranes is 3-phosphoinositide-dependent protein kinase-1. Once it is located in the membrane, it can be activated by various kinases: when phosphorylated by PI3-kinase, it can undergo subcellular translocation. TSH alters the subcellular distribution of this kinase in FRTL-5 cells via both PI3-kinase and PKA mediated pathways (47). 3-phosphoinositide-dependent protein kinase-1 then acts by phosphorylating the activation loop of members of the PKA, PKC, and cyclic GMP-dependent kinase families.

S6K is another pleckstrin domain–containing kinase that is phosphorylated by many kinase pathways: SK6 in turn phosphorylates the 40S ribosomal protein S6 and the initiation factor 4E-binding protein-1 required for cells to traverse from G1 to S phase. If FRTL-5 cells are maintained in medium without any TSH, insulin, or serum, the addition of TSH stimulates p70-S6K activity, but this response can be blocked either by inhibitors of PKA or of PI3-kinase (47). In Wistar rat thyrocytes, TSH and cyclic AMP stimulate S6K both by Rap-1A–and cyclic AMP-dependent pathways (49,50). In primary dog thyrocytes, TSH rapidly activates p70-S6K but does not affect p90-S6K (51).

The mammalian target of rapamycin (mTOR) is yet another pleckstrin domain–containing kinase whose phosphorylation is regulated by small GTPases. The mammalian target of rapamycin senses nutrient levels, receives mitogenic signals, and transmits signals to S6K to regulate cell proliferation. Rapamycin, a specific inhibitor of mTOR, blocks TSH/cAMP-mediated phosphorylation of S6K in FRTL-5 cells, and causes cell cycle arrest, implicating mTOR in this action of TSH (48).

Protein kinase B (also called PKB) kinases are phosphoproteins located in cytosol in a relatively inactive form under basal conditions. Increasing the membrane level of phosphatidylinositol phosphates causes Akts to bind to the membrane, which alters the structure of the Akts and permits their activation (52,53). In FRTL-5 cells and human thyroid cancer cells, TSH does not affect Akt phosphorylation, although insulin does (55,56). Akt is involved in FRTL-5 cell proliferation: expressing a constitutively active Akt-l mutant in these cells increases basal DNA synthesis, and serum or insulin/IGF-1 are no longer needed for TSH to induce DNA synthesis (55). In Wistar rat thyrocytes (grown with insulin but without serum), TSH causes phosphorylation of Akt. Expressing a constitutively active Rap increases TSH-mediated phosphorylation of Akt in these cells (53). In contrast, in rat PCCl3 cells (grown with insulin but without serum), TSH inhibits Akt activity via PKA, presumably via PKA-mediated phosphorylation and activation of Rap-1b (54). When a constitutively active form of Rap is expressed in PCCl3 cells it decreases Akt phosphorylation, and potentiates the inhibitory effect of forskolin on Akt phosphorylation (54). These three different effects of TSH on Akt in three rat thyrocyte models could reflect the different ways that these immortalized cells were obtained: Wistar cells were cloned from fibroblasts of thyroids from immature rats, PCCl3 cells were cloned from 18-month-old rats, and FRTL-5 cells were obtained by culturing young rat thyroid cells in low levels of calf serum. If vascular cells are treated with H2O2, Akt becomes oxidized, and its phosphorylation is affected (this action of H2O2on Akt appears to be mediated by PI3-kinase, since it is blocked by wortmannin) (52).


TSH regulates thyroid peroxidase expression via cyclic AMP pathways, whereas H2O2 production involves Ca++/ phosphoinositol pathways in all species. In dog thyrocytes, TSH and cyclic AMP agonists stimulate the production of H2O2 (57), and high doses of TSH can increase H2O2 production in FRTL-5 cells as well (13). It is noteworthy that exogenous H2O2 can induce apoptosis in FRTL-5 cells (58), and a variety of factors are involved in controlling the toxic effects of H2O2, lipid hydroperoxides, and free radicals. TSH increases the expression of Ref-1 (15), the transcription factor that repairs DNA damaged by oxidative stress, and Ref-1 also protects cells from oxidative apoptotic stimuli. Changes in the redox state of a cell alter the activity of CHOP, a transcription factor that monitors endoplasmic reticulum stress and which also can regulate apoptosis. CHOP is expressed in rat thyroids, as well as in FRTL-5 and PCCl3 cells. If TSH is removed from the culture medium of FRTL-5 cells, CHOP levels fall unless forskolin is present (59). This protective effect of forskolin on CHOP levels involves PKA, mTOR, and the level of reactive oxygen species (59). In FRTL-5 cells, TSH increases the expression of antioxidants like peroxiredoxin-1 and thioredoxin (13,58). Thioredoxin reductase is induced by Ca++/phosphoinositol pathways, but not by cyclic AMP in human thyrocytes (60). Duox-2 is a flavoprotein involved in H2O2 production, and which is regulated by TSH and cyclic AMP in pig (61), human, and dog thyrocytes (62). Insulin also increases Duox-2 mRNA expression in FRTL-5 cells grown in TSH-free low-serum medium, but the addition of insulin switches the stimulatory action of forskolin on Duox-2 to an inhibitory one (63). This illustrates how a hormone that does not directly act via cyclic AMP can dramatically alter a thyrocyte's response to cyclic AMP.


Many kinds of stress can activate MAPKKK/MAPKK/ MAP kinase cascades, and thus regulate cell growth, transformation, and apoptosis. The downstream cascades can be divided into three subfamilies: p38 MAPKs and the JNKs (c-jun N-terminal kinases) (both associated with regulation in response to cellular stresses), and the ERKs (extracellular signal–regulated kinases), which are associated with regulation of cell proliferation and differentiation, although the actions of the subfamilies can overlap.

When pig thyrocytes are cultured with TSH, they form follicle-like structures. If the TSH is removed, this stress causes the cells to migrate and become confluent monolayers. The loss of higher organization and the migration can be prevented if TSH or cyclic AMP agonists are added to the culture medium (64). The loss of organization involves activation-specific phosphorylation of ERK, which accumulates in the thyrocyte nuclei. ERK inhibitors block the migratory responses, and PI3-kinase inhibitors block the initiation of cell movement, but not the activation of ERK, suggesting that two pathways are involved in the cell migration that is tonically inhibited by the presence of TSH or cyclic AMP (64).

TSH and cyclic AMP agonists also cause a rapid but transient phosphorylation of ERK in FRTL-5 cells (45). The importance of this response was confirmed by inhibiting MEK (MAPKK), which blocks MEK-dependent phosphorylation of ERK as well as cyclic AMP-mediated DNA synthesis (45). The cyclic AMP-mediated DNA synthesis and ERK reponses were not blocked by the PKA inhibitor H89, whereas H89 does inhibit p38 MAP kinase phosphorylation in response to TSH (65). This cyclic AMP-dependent activation of p38 MAP kinase appears to involve reactive oxygen species (65). In dog and human thyrocytes, insulin must be present in order for TSH or cyclic AMP to be mitogenic, and its action is also inhibited when MEK is inhibited (66).

When cytokine receptors in the plasma membrane are activated, they attract the cytoplasmic tyrosine kinase JAK (Janus kinase), creating docking sites for SH-2–containing proteins like the STATs (signal transducer and activator of transcription. Several cytokines are known to alter thyrocyte growth and function, and TSH and cyclic AMP influence these actions of cytokines. As has been observed for other heterotrimeric G-protein–coupled receptors, activation of the TSH receptor can alter the tyrosine phosphorylation and activation of STATs. In FRTL-5 cells, TSH causes tyrosine phosphorylation of STAT-3, induces the formation of a STAT-3/DNA complex (67), and induces the expression of the feedback inhibitor SOCS-1 (suppressor of cytokine signaling-1), a target gene of STAT-3 (68). Expression of a dominant negative STAT-3 inhibited the action of TSH on SOCS-1 expression, confirming that TSH is acting via JAK/STAT signaling (67). The action of TSH and forskolin on STAT-3 phosphorylation can be blocked by inhibitors of PKC, but not by inhibitors of PKA (69).


TSH activates Gαq-containing heterotrimers to stimulate isoforms of phospholipase A, C, and D, and provides diacylglycerol (DAG) and inositol 1,4,5-trisphosphate, which can mobilize intracellular pools of Ca2+ to activate PKC. Cross regulation and indirect effects have made it difficult to establish which are the primary actions of TSH. For example, the activity and subcellular location of phospholipase D is regulated by multiple pathways, including phospholipids, small GTPases, PKC, as well as via indirect effects through tyrosine kinases. The phosphatidic acid that phospholipase D produces acts as a cofactor in a number of signaling cascades, in addition to being converted to DAG by phosphatidic acid phosphatases. Phospholipases also release arachidonic and linoleic acids that can be metabolized by lipoxygenase to form free radicals and lipid peroxides, which provide another inflammatory stress that can induce apoptosis. Arachidonate also may be converted to prostanoids. The stimulation of phospholipase D activity by TSH appears to involve both cyclic AMP and PKC, since inhibitors of PKC and PKA have separate inhibitory actions on TSH-mediated phospholipase D activity (70).


TSH stimulates blood flow in the thyroid by increasing nitric oxide synthase and nitric oxide production (71). Chronic TSH stimulation increases the vascularity of the thyroid gland, TSH increases the expression of FGF-1 (fibroblast growth factor 1) and in primary cultures of human thyrocytes (72). If a dominant negative FGF-1 mutation is expressed in FRTL-5 cells, it inhibits the actions of TSH on cell growth and iodide uptake, and if it is expressed in mice, it reduces the goiter and the vascular growth produced by a low iodine diet plus perchlorate, indicating that FGF receptor–signaling affects several thyroid responses (73). Adding FGF-2 to human thyrocytes increases their growth but inhibits their iodide uptake (72). Vascular endothelial growth factor (VEGF) isoforms have different effects in different models, but in primary cultures of human thyrocytes, VEGF stimulates nitric oxide release (74). Histochemical studies show that thyroid follicles with a cuboidal epithelium are metabolically active, and express TSH receptor, sodium-iodine symporter, pendrin, thyroid peroxidase, and thyroglobulin that contains T4, whereas follicles with a flat epithelium are metabolically inactive and express little of these substances. The capillary networks that surround active follicles contain nitric oxide synthase III and endothelin, but the capillaries around flat follicles do not, suggesting there is a paracrine connection that communicates the activity of epithelial cells to the neighboring vascular cells (75). Thyrocytes are known to produce angiogenic factors, including IGF-I and angiopoietins, but the primary regulators of thyroid vascularity remain to be established. Interestingly, if rats are given a low iodine diet plus goitrogens to raise the circulating level of TSH, the vascular endothelium begins to proliferate before the thyrocytes (76). Chronic TSH stimulation of rats causes thrombospondin, an inhibitor of angiogenesis, to disappear progressively from thyroid endothelial cells and stroma (77).


Thyroid hormones can act on the thyroid gland directly, and on the other hand, TSH and cyclic AMP increase nuclear level of triiodothyronine (T3) receptors as well as increasing the activity of deiodinases in human thyrocytes, responses that can be inhibited by activating PKB (78). The nuclear T3 receptor is involved in the growth of the thyroid, and if one gives T3 to hypophysectomized rats, the density of TSH receptors expressed on the thyroid increases, representing another way that thyroid hormone can affect the function of the thyroid directly (31). Studies in vivo (79) and in FRTL-5 cells (80) indicate that epidermal growth factor acts as an autocrine factor produced within the thyroid in response to changes in thyroid hormone levels.

Somatostatin is commonly added to culture media to promote thyrocyte differentiation in vitro, but the somatostatin gene is expressed in FRTL-5 cells and is regulated by TSH, indicating it is an autocrine inhibitory factor (81). Adding exogenous somatostatin to FRTL-5 cells inhibits the proliferative responses both to TSH and to insulin, while it increases the expression of p27kip1, the cyclin-dependent kinase cell cycle inhibitor (81). Somatostatin also lessens TSH-mediated activation of adenylyl cyclase, and inhibits PI3-kinase in a PKA-independent fashion in these cells (82). In PCCl3 cells, somatostatin inhibits the proliferation by activating a membrane phosphotyrosine phosphatase, which prevents the activation of MAP kinase and stabilizes p27kip1 (83,84).


When a goiter undergoes involution, organized cell death is involved (85). Apoptosis occurs in both stromal as well as vascular cells if a normal level of iodine is added to the diet of rats that have been fed a low iodine diet plus methimazole (86). In rats with methimazole-induced goiter, withdrawal of the methimazole initially inhibits cell proliferation, while apoptosis becomes evident about a week later (87).

TSH and cyclic AMP agonists protect FRTL-5 cells from apoptosis that normally would be induced by actinomycin D or H2O2 (88). The extracellular matrix is an important regulator of cell development and survival. If serum is removed from the incubation medium of FRTL-5 cells, TSH must be present in order to prevent apoptosis. This action involves increased cell adhesion and is mediated in part by PKA (89). If serum is removed from the incubation medium, human thyrocytes degrade extracellular matrix, lose integrin-fibronectin interactions, and apoptosis ensues. Adding an inhibitor of PI3-kinase promotes apoptosis in this model (90). TSH protects Wistar rat thyrocytes from nitroprusside-induced apoptosis, and this effect involves PKA and S6K (91). Rap-1 also plays a role, since in cells expressing a constitutively active Rap-1, the apoptotic effect of nitroprusside is accentuated, whereas cells expressing a dominant negative Rap-1 resist the apoptotic effect of nitroprusside (91).


TSH affects various members of the complex network of cell cycle activators and inhibitors, but no single common pathway applies for all models. It is commonly held that cyclin-dependent kinases (CDKs) need to be activated before cells can enter the S phase of the cell cycle, and that the level of the cyclin inhibitor protein p27kip1must decrease to allow CDK activation. In FRTL-5 cells, TSH does cause the expected decrease in p27kip1 and increases cyclin D1 levels, whereas phosphodiesterase inhibitors curtail the cyclin D1 response to TSH and IGF-1 (21). Cyclic AMP is involved in the proliferation of synchronized FRTL-5 cells: the levels of RI and RII isoforms of PKA both change with the cell cycle, and experimentally altering these regulatory subunit responses affects cell cycle progression (92). The predominant cyclin D isoform in many thyrocytes is cyclin D3, and this isoform appears to be important clinically because it is overexpressed in most follicular adenomas (93). In rats given propylthiouracil, cyclin D3 levels increase within 5 days, while in PCCl3 cells grown in serum alone, adding TSH increases cyclin D3 synthesis within 6 hours (93). Cyclin D3 also plays a role in TSH-mediated proliferation of primary cultures of dog thyrocytes, since proliferation is suppressed when cyclin D3 is neutralized by microinjecting antibody to cyclin D3 (94). If TSH, serum or insulin are added to starved WRT or PCCI3 cells, there is no increase in cell proliferation (94a). However if serum is combined with either TSH or forskolin, the cells proliferate rapidly. Adding the full complement of growth factors causes CDK2 levels to increase at 3 hours, followed by an increase in cyclin D1 and then in cyclin D3. Although total p27kip1 levels decreased only slightly at 15 hours, nuclear staining of p27kip1 in WRT cells decreased wthin 6 hours. TSH or forskolin did not deplete nuclear p27kip1, but they accelerated and potentiated this effect of serum. The disappearance of nuclear p27kip1 staining correlated with progression of cells from G1 to S (94a).


Functional TSH receptors are expressed in several nonthyroid tissues in which the actions of TSH are not mediated by cyclic AMP. For example, the action of TSH on neurotransmitter-induced ion currents in submucosal glands from human trachea is blocked by tyrosine kinase inhibitors, and is not mimicked by cyclic AMP (95). In abdominal preadipocytes and orbital fibroblasts, TSH increases S6K activity, and this response can be blocked by wortmannin, the inhibitor of PI3-kinase (96). Studies in TSH receptor knockout mice suggest that TSH may have a direct effect on osteoclast formation that is mediated by JNK phosphorylation, and may affect osteoblast differentiation and type 1 collagen formation via Wnt and VEGF signaling pathways 97. If 3T3-L1 preadipocytes are grown without serum, TSH acts as a survival factor (98). TSH has no effect on cyclic AMP levels in these cells, but it increases phosphotyrosine and phosphatidylinositol phosphate levels—presumably via PI3-kinase, since TSH increases the level of activated S6K and Akt, and since these responses can be blocked by PI3-kinase inhibitors (98).


Many nonthyroid cells have been transfected to express TSH receptors, and in some cells TSH activates pathways that are not prominent in thyroid models, probably because in the spectrum of proteins expressed is different from the spectrum expressed in thyrocytes. Furthermore, only a low level of TSH receptor is normally expressed in thyrocytes, whereas experimental expression of a high level of TSH receptor may drive interactions that would not occur normally. This is also true when PKA subunits are overexpressed in cells (99). Nonetheless, the results obtained by such studies can be interesting. For example, cyclic AMP normally suppresses the growth of wild-type mouse NIH-3T3 fibroblasts. These fibroblasts express a relatively low level of the RIIβ isoform of A-kinase, and if RIIβ is overexpressed, the cells undergo apoptosis, which can be reversed by adding 8-Br cyclic AMP. If wild-type NIH-3T3 fibroblasts are transfected with a human TSH receptor gene under control of a tetracycline repressor gene and then tetracycline is removed, TSH causes cyclic AMP levels to rise and cell growth is inhibited, as would be expected. However, if both TSH receptors and RIIβ are expressed in these cells, TSH actually stimulates growth, which appears to involve CREB phosphorylation, Rap-1, Akt, and ERK activity (100).


As different mechanisms of action of TSH are uncovered in different models, we are obtaining a deeper understanding of the way different signaling pathways come together to control thyroid function, growth, and apoptosis. The clinical insights we have obtained from such studies have provided a deeper understanding of several thyroid diseases and of diseases in other organs as well.


1. Hamad NM, Elconin JH, Karnoub AE, et al. Distinct requirements for Ras oncogenesis in human versus mouse cells. Genes Dev 2002;16:2045–57.

2. Smogorzewska A, de Lange T. Different telomere damage signaling pathways in human and mouse cells. EMBO J 2002;21: 4338–48.

3. Pohlenz J, Maqueem A, Cua K, et al. Improved radioimmunoassay for measurement of mouse thyrotropin in serum: strain differences in thyrotropin concentration and thyrotroph sensitivity to thyroid hormone. Thyroid 1999;9:1265–71.

4. Postiglione MP, Parlato R, Rodriguez-Mallon A, et al. Role of the thyroid-stimulating hormone receptor signaling in development and differentiation of the thyroid gland. Proc Natl Acad Sci U S A 2002;99:15462–7.

5. Pic P, Michel-Bechet M, el Atiq F, et al. Forskolin stimulates cAMP production and the onset of the functional differentiation in the fetal rat thyroid in vitro. Biol Cell 1986;57: 231–7.

6. Marians RC, Ng L, Blair HC, et al. Defining thyrotropin-dependent and -independent steps of thyroid hormone synthesis by using thyrotropin receptor-null mice. Proc Natl Acad Sci USA 2002;99:15776–81.

7. Medina DL, Suzuki K, Pietrarelli M, et al. Role of insulin and serum on thyrotropin regulation of thyroid transcription factor-1 and PAX8 genes expression in FRTL-5 thyroid cells. Thyroid 2000;10:295–303.

8. Di Palma T, Nitsch R, Mascia A, et al. The paired domain-containing factor Pax8 and the homeodomain-containing factor TTF1 directly interact and synergistically activate transcription. J Biol Chem 2003;278:3395–3402.

9. Moeller LC, Kimura S, Kusakabe T, et al. Hypothyroidism in TTF1 haploinsufficiency is caused by reduced expression of the TSH receptor. Mol Endocrinol 2003;17:2295–2302.

10. Nguyen LQ, Kopp P, Martinson F, et al. A dominant negative CREB (cAMP response element-binding protein) isoform inhibits thyrocyte growth, thyroid-specific gene expression, differentiation, and function. Mol Endocrinol 2000;14:1448–1461.

11. Feliciello A, Allevato G, Musti AM, et al. Thyroid transcription factor 1 phosphorylation is not required for PKA-dependent transcription of the thyroglobulin promoter. Cell Growth Differ 2000;11:649.

12. Puppin C, Arturi F, Ferretti E, et al. Transcriptional regulation of human NIS gene: a role for Redox factor-1. Endocrinology 2004;145:1290.

13. Kambe F, Nomura Y, Okamoto T, et al. Redox regulation of thyroid-transcription factors, PAX8 and TTF1, is involved in their increased DNA-binding activities by thyrotropin in rat thyroid FRTL-5 cells. Mol Endocrinol 1996;10:801–812.

14. Cao X, Kambe F, Ohmori S, et al. Oxidoreductive modification of two cysteine residues in paired domain by Ref-1 regulates DNA-binding activity of PAX8. Biochem Biophys Res Commun 2002;297:288–293.

15. Asai T, Kambe F, Kikumori T, et al. Increase in Ref-1 mRNA and protein by thyrotropin in rat thyroid FRTL-5 cells. Biochem Biophys Res Commun 1997;236:71–74.

16. Russo D, Celano M, Bulotta S, et al. APE/Ref-1 is increased in nuclear fractions of human thyroid hyperfunctioning nodules. Mol Cell Endocrinol 2002;194:71–76.

17. Kambe F, Miyazaki T, Seo H. Differential induction of fos and jun family genes by thyrotropin in rat thyroid FRTL-5 cells. Thyroid 1996;6:123–128.

18. Iacovelli L, Franchetti R, Grisolia D, et al. Selective regulation of G protein-coupled receptor-mediated signaling by G protein-coupled receptor kinase 2 in FRTL-5 cells: analysis of thyrotropin, alpha(1β)-adrenergic, and A (1) adenosine receptor-mediated responses. Mol Pharmacol 1999;56:316–324.

19. Persani L, Lania A, Alberti L, et al. Induction of specific phosphodiesterase isoforms by constitutive activation of the cAMP pathway in autonomous thyroid adenomas. J Clin Endocrinol Metab 2000;85:2872–2878.

20. Oki N, Takahashi SI, Hidaka H, et al. Short term feedback regulation of cAMP in FRTL-5 thyroid cells. Role of PDE4D3 phosphodiesterase activation. J Biol Chem 2000;275:10831–10837.

21. Takahashi SI, Nedachi T, Fukushima T, et al. Long-term hormonal regulation of the cAMP-specific phosphodiesterases in cultured FRTL-5 thyroid cells. Biochim Biophys Acta 2001; 1540:68–81.

22. Nemoz G, Sette C, Hess M, et al. Activation of cyclic nucleotide phosphodiesterases in FRTL-5 thyroid cells expressing a constitutively active Gsα. Mol Endocrinol 1995;9:1279–1287.

23. Green LM, Murray DK, Tran DT, et al. A spontaneously arising mutation in connexin32 with repeated passage of FRTL-5 cells coincides with increased growth rate and reduced thyroxine release. J Mol Endocrinol 2001;27:145–163.

24. Tasevski V, Benn D, Peters G, et al. The Fischer rat thyroid cell line FRTL-5 exhibits a nondiploid karyotype. Thyroid 1998; 8:623–626.

25. Derwahl M, Seto P, Rapoport B. An abnormal splice donor site in one allele of the thyroid peroxidase gene in FRTL5 rat thyroid cells introduces a premature stop codon: association with the absence of functional enzymatic activity. Mol Endocrinol 1990;4:793–799.

26. Lalli E, Sassone-Corsi P. Thyroid-stimulating hormone (TSH)-directed induction of the CREM gene in the thyroid gland participates in the long-term desensitization of the TSH receptor. Proc Natl Acad Sci USA 1995;92:9633–9637.

27. Uyttersprot N, Costagliola S, Dumont JE, et al. Requirement for cAMP-response element (CRE) binding protein/CRE modulator transcription factors in thyrotropin-induced proliferation of dog thyroid cells in primary culture. Eur J Biochem 1999;259:370–378.

28. Arseven OK, Wilkes WP, Jameson JL, et al. Substitutions of tyrosine 601 in the human thyrotropin receptor result in increase or loss of basal activation of the cyclic adenosine monophosphate pathway and disrupt coupling to Gq/11. Thyroid 2000;10:3–10.

29. Wonerow P, Chey S, Fuhrer D, et al. Functional characterization of five constitutively activating thyrotrophin receptor mutations. Clin Endocrinol (Oxf) 2000;53:461–468.

30. Armstrong R, Wen W, Meinkoth J, et al. A refractory phase in cyclic AMP-responsive transcription requires down regulation of PKA. Mol Cell Biol 1995;15:1826–1832.

31. Denereaz N, Lemarchand-Beraud T. Severe but not mild alterations of thyroid function modulate the density of thyroid-stimulating hormone receptors in the rat thyroid gland. Endocrinology 1995;136:1694–1700.

32. Holzapfel HP, Bergner B, Wonerow P, et al. Expression of G(alpha)(s) proteins and TSH receptor signalling in hyperfunctioning thyroid nodules with TSH receptor mutations. Eur J Endocrinol 2002;147:109–116.

33. Shi CS, Lee SB, Sinnarajah S, et al. Regulator of G-protein signaling 3 (RGS3) inhibits Gβγ 2-induced inositol phosphate production, mitogen-activated PKA activation, and Akt activation. J Biol Chem 2001;276:24293–24300.

34. Kawasaki H, Springett GM, Mochizuki N, et al. A family of cAMP-binding proteins that directly activate Rap1. Science 1998;282:2275–2279.

35. Namba H, Rubin SA, Fagin JA. Point mutations of Ras oncogenes are an early event in thyroid tumorigenesis. Mol Endocrinol 1990;4:1474–1479.

36. Vasko V, Ferrand M, Di Cristofaro J, et al. Specific pattern of Ras oncogene mutations in follicular thyroid tumors. J Clin Endocrinol Metab 2003;88:2745–2752.

37. Gire V, Wynford-Thomas D. Ras oncogene activation induces proliferation in normal human thyroid epithelial cells without loss of differentiation. Oncogene 2000;19:737–744.

38. Cheng G, Lewis AE, Meinkoth JL. Ras stimulates aberrant cell cycle progression and apoptosis in rat thyroid cells. Mol Endocrinol 2003;17:450–459.

39. Francis-Lang H, Zannini M, De Felice M, et al. Multiple mechanisms of interference between transformation and differentiation in thyroid cells. Mol Cell Biol 1992;12:5793–5800.

40. Ciullo I, Diez-Roux G, Di Domenico M, et al. cAMP signaling selectively influences Ras effectors pathways. Oncogene 2001; 20:1186–1192.

41. Tsygankova OM, Kupperman E, Wen W, et al. Cyclic AMP activates Ras. Oncogene 2000;19:3609–3615.

42. Tsygankova OM, Feshchenko E, Klein P, et al. TSH/cAMP and GSK3 beta elicit opposing effects on Rap1GAP stability. J Biol Chem 2004;279:5501–7.

42a. Lou L, Urbani J, Ribeiro-Neto F, et al. cAMP inhibition of Akt is mediated by activated and phosphorylated Rap1b. J Biol Chem 2002;277:32799–32806.

43. Dremier S, Vandeput F, Zwartkruis FJ, et al. Activation of the small G protein Rap1 in dog thyroid cells by both cAMP-dependent and -independent pathways. Biochem Biophys Res Commun 2000;267:7–11.

43a. Ribiero-Neto F, Leon L, Urbani-Brocard, et al. cAMP-dependent oncogenic action of Rap1b in the thyroid gland. J Biol Chem 2004 e-pub Aug 26.

44. Iacovelli L, Capobianco L, Salvatore L, et al. Thyrotropin activates mitogen-activated protein kinase pathway in FRTL-5 by a PKA-independent mechanism. Mol Pharmacol 2001;60: 924–933.

45. Ariga M, Nedachi T, Akahori M, et al. Signalling pathways of insulin-like growth factor-I that are augmented by cAMP in FRTL-5 cells. Biochem J 2000;348[Pt 2]:409–416.

46. Nedachi T, Akahori M, Ariga M, et al. Tyrosine kinase and phosphatidylinositol 3-kinase activation are required for cyclic adenosine 3′,5′-monophosphate-dependent potentiation of deoxyribonucleic acid synthesis induced by insulin-like growth factor-I in FRTL-5 cells. Endocrinology 2000;141:2429–2438.

47. Suh JM, Song JH, Kim DW, et al. Regulation of the phosphatidylinositol 3-kinase, Akt/protein kinase B, FRAP/mammalian target of rapamycin, and ribosomal S6 kinase 1 signaling pathways by thyroid-stimulating hormone (TSH) and stimulating type TSH receptor antibodies in the thyroid gland. J Biol Chem 2003;278:21960–21971.

48. Garcia B, Santisteban P. PI3K is involved in the IGF-I inhibition of TSH-induced sodium-iodide symporter gene expression. Mol Endocrinol 2002;16:342–352.

49. Cass LA, Summers SA, Prendergast GV, et al. PKA-dependent and -independent signaling pathways contribute to cyclic AMP-stimulated proliferation. Mol Cell Biol 1999;19:5882–5891.

50. Cass LA, Meinkoth JL. Ras signaling through PI3K confers hormone-independent proliferation that is compatible with differentiation. Oncogene 2000;19:924–932.

51. Coulonval K, Vandeput F, Stein RC, et al. Phosphatidylinositol 3-kinase, protein kinase B and ribosomal S6 kinases in the stimulation of thyroid epithelial cell proliferation by cAMP and growth factors in the presence of insulin. Biochem J 2000;348[Pt 2]:351–358.

52. Ushio-Fukai M, Alexander RW, Akers M, et al. Reactive oxygen species mediate the activation of Akt/protein kinase B by angiotensin II in vascular smooth muscle cells. J Biol Chem 1999;274:22699–22704.

53. Mei FC, Qiao J, Tsygankova OM, et al. Differential signaling of cyclic AMP: opposing effects of exchange protein directly activated by cyclic AMP and cAMP-dependent protein kinase on protein kinase B activation. J Biol Chem 2002;277:11497–11504.

55. Saito J, Kohn AD, Roth RA, et al. Regulation of FRTL-5 thyroid cell growth by phosphatidylinositol (OH) 3 kinase-dependent Akt-mediated signaling. Thyroid 2001;11:339–351.

56. Ringel MD, Hayre N, Saito J, et al. Overexpression and overactivation of Akt in thyroid carcinoma. Cancer Res 2001;61: 6105–6111.

57. Corvilain B, Collyn L, van Sande J, et al. Stimulation by iodide of H2O2 generation in thyroid slices from several species. Am J Physiol Endocrinol Metab 2000;278:E692–E699.

58. Kim H, Lee TH, Park ES, et al. Role of peroxiredoxins in regulating intracellular hydrogen peroxide and hydrogen peroxide-induced apoptosis in thyroid cells. J Biol Chem 2000;275: 18266–18270.

59. Pomerance M, Carapau D, Chantoux F, et al. CCAAT/enhancer-binding protein-homologous protein (CHOP) expression and transcriptional activity are regulated by cyclic AMP in thyroid cells. Mol Endocrinol 2003;17:2283–22894.

60. Howie AF, Arthur JR, Nicol F, et al. Identification of a 57-kilodalton selenoprotein in human thyrocytes as thioredoxin reductase and evidence that its expression is regulated through the calcium-phosphatidylinositol signaling pathway. J Clin Endocrinol Metab 1998;83:2052–2058.

61. Morand S, Chaaraoui M, Kaniewski J, et al. Effect of iodide on nicotinamide adenine dinucleotide phosphate oxidase activity and Duox2 protein expression in isolated porcine thyroid follicles. Endocrinology 2003;144:1241–1248.

62. De Deken X, Wang D, Many MC, et al. Cloning of two human thyroid cDNAs encoding new members of the NADPH oxidase family. J Biol Chem 2000;28;23227–23233.

63. Morand S, Dos Santos OF, Ohayon R, et al. Identification of a truncated dual oxidase 2 mRNA in two rat thyroid cell lines. Insulin and forskolin regulation of DUOX2 mRNA levels in FRTL-5 cells and porcine thyrocytes. Endocrinology 2003;144: 567–574.

64. Yap AS. Initiation of cell locomotility is a morphogenetic checkpoint in thyroid epithelial cells regulated by ERK and PI3-kinase signals. Cell Motil Cytoskeleton 2001;49:93–103.

65. Pomerance M, Abdullah HB, Kamerji S Thyroid-stimulating hormone and cyclic AMP activate p38 mitogen-activated protein kinase cascade. Involvement of protein kinase A, rac1, and reactive oxygen species. J Biol Chem 2000;275:40539–40546.

66. Vandeput F, Perpete S, Coulonval K, et al. Role of the different mitogen-activated protein kinase subfamilies in the stimulation of dog and human thyroid epithelial cell proliferation by cyclic adenosine 5′-monophosphate and growth factors. Endocrinology 2003;144:1341–1349.

67. Park ES, Kim H, Suh JM, et al. Involvement of JAK/STAT (Janus kinase/signal transducer and activator of transcription) in the thyrotropin signaling pathway. Mol Endocrinol 2000;14: 662–670.

68. Park ES, Kim H, Suh JM, et al. Thyrotropin induces SOCS-1 (suppressor of cytokine signaling-1) and SOCS-3 in FRTL-5 thyroid cells. Mol Endocrinol 2000;14:440–448.

69. Park YJ, Park ES, Kim MS, et al. Involvement of the protein kinase C pathway in thyrotropin-induced STAT3 activation in FRTL-5 thyroid cells. Mol Cell Endocrinol 2002;194:77–84.

70. Ginsberg J, Gupta S, Matowe WC, et al. Activation of phospholipase D in FRTL-5 thyroid cells by forskolin and dibutyryl cyclic AMP. Endocrinology 1997;138:3645–3651.

71. Colin IM, Nava E, Toussaint D, et al. Expression of nitric oxide synthase isoforms in the thyroid gland: evidence for a role of nitric oxide in vascular control during goiter formation. Endocrinology 1995;136:5283–5290.

72. Cocks HC, Thompson S, Turner FE, et al. Role and regulation of the fibroblast growth factor axis in human thyroid follicular cells. Am J Physiol Endocrinol Metab 2003;285:E460–E469.

73. Davies EL, Ramsden JD, Cocks H, et al. Adenovirus-mediated expression of dominant negative fibroblast growth factor (FGF) receptor 1 in thyroid cells blocks FGF effects and reduces goitrogenesis in mice. J Clin Endocrinol Metab 2003;88: 4472–4480.

74. Patel VA, Logan A, Watkinson JC, et al. Isolation and characterization of human thyroid endothelial cells. Am J Physiol Endocrinol Metab. 2003;284:E168–E176.

75. Gerard AC, Many MC, Daumerie C, et al. Structural changes in the angiofollicular units between active and hypofunctioning follicles align with differences in the epithelial expression of newly discovered proteins involved in iodine transport and organification. J Clin Endocrinol Metab 2002;87:1291–1299.

76. Smeds S, Wollman SH. 3H-thymidine labeling of endothelial cells in thyroid arteries, veins, and lymphatics during thyroid stimulation. Lab Invest 1983;48:285–291.

77. Patel VA, Hill DJ, Eggo MC, et al. Changes in the immunohistochemical localisation of fibroblast growth factor-2, transforming growth factor-beta 1 and thrombospondin-1 are associated with early angiogenic events in the hyperplastic rat thyroid. J Endocrinol 1996;148:485–499.

78. Imai Y, Toyoda N, Maeda A, et al. Type 2 iodothyronine deiodinase expression is upregulated by the PKA-dependent pathway and is downregulated by the protein kinase C-dependent pathway in cultured human thyroid cells. Thyroid 2001; 11: 899–907.

79. Ozawa S, Sheflin LG, Spaulding SW. Thyroxine increases epidermal growth factor levels in the mouse thyroid in vivo. Endocrinology 1991;128:1396–1403.

80. Di Fulvio M, Coleoni AH, Pellizas CG, et al. Tri-iodothyronine induces proliferation in cultured bovine thyroid cells: evidence for the involvement of epidermal growth factor-associated tyrosine kinase activity. J Endocrinol 200l;166:173–182.

81. Medina DL, Santisteban P. Somatostatin is expressed in FRTL-5 thyroid cells and prevents thyrotropin-mediated down-regulation of the cyclin-dependent kinase inhibitor p27kip1. Endocrinology 1999;140:87–95.

82. Medina DL, Toro MJ, Santisteban P. Somatostatin interferes with thyrotropin-induced G1-S transition mediated by cAMP-dependent protein kinase and phosphatidylinositol 3-kinase. Involvement of RhoA and cyclin E x cyclin-dependent kinase 2 complexes. J Biol Chem 2000;275:15549–15556.

83. Florio T, Arena S, Thellung S, et al. The activation of the phosphotyrosine phosphatase eta is responsible for the somatostatin inhibition of PCCl3 thyroid cell proliferation. Mol Endocrinol 2001;15:1838–1852.

84. Trapasso F, Iuliano R, Boccia A, et al. A rat protein tyrosine phosphatase eta suppresses the neoplastic phenotype of retrovirally transformed thyroid cells through the stabilization of p27(Kip1). Mol Cell Biol 2000;20:9236–9246.

85. Tamura M, Kimura H, Koji T, et al. Role of apoptosis of thyrocytes in a rat model of goiter: a possible involvement of Fas system. Endocrinology 1998;139:3646–53.

86. Patel VA, Hill DJ, Sheppard MC, et al. Apoptosis during goitre involution—the role of Bcl-2. J Endocrinol 2000; 164:323–330.

87. Riesco JM, Juanes JA, Carretero J, et al. Cell proliferation and apoptosis of thyroid follicular cells are involved in the involution of experimental nontumoral hyperplastic goiter. Anat Embryol (Berl) 1998;198:439–450.

88. Sato H, Abe Y, Noguchi M, Kurokawa K, et al. Inhibitory effect of TSH on apoptosis induced by actinomycin D in a functioning rat thyroid cell line. Endocr J 1999;46:309–315.

89. Li X, Lu S, Miyagi E, et al. Thyrotropin prevents apoptosis by promoting cell adhesion and cell cycle progression in FRTL-5 cells. Endocrinology 1999;140:5962–5970.

90. Illario M, Amideo V, Casamassima A, et al. Integrin-dependent cell growth and survival are mediated by different signals in thyroid cells. J Clin Endocrinol Metab 2003;88:260–269.

91. Saavedra AP, Tsygankova OM, Prendergast GV, et al. Role of cAMP, PKA and Rap1A in thyroid follicular cell survival. Oncogene 2002;21:778–788.

92. Feliciello A, Gallo A, Mele E, et al. The localization and activity of cAMP-dependent protein kinase affect cell cycle progression in thyroid cells. J Biol Chem 2000;275:303–311.

93. Motti ML, Boccia A, Belletti B, et al. Critical role of cyclin D3 in TSH-dependent growth of thyrocytes and in hyperproliferative diseases of the thyroid gland. Oncogene 2003;22:7576–7586.

94. Van Keymeulen A, Bartek J, Dumont JE, et al. Cyclin D3 accumulation and activity integrate and rank the comitogenic pathways of thyrotropin and insulin in thyrocytes in primary culture. Oncogene 1999;18:7351–7359.

94A. Lewis AE, Fikaris AJ, Prendergast GV et al. Thyrotropin and serum regulate thyroid cell proliferation through differential effects on p27 expression and localization. Mol Endocrinol 2004; 18:2321–32.

95. Tamada T, Sasaki T, Saitoh H, et al. A novel function of thyrotropin as a potentiator of electrolyte secretion from the tracheal gland. Am J Respir Cell Mol Biol 2000;22:566–573.

96. Bell A, Gagnon A, Grunder L, et al. Functional TSH receptor in human abdominal preadipocytes and orbital fibroblasts. Am J Physiol Cell Physiol 2000;279:C335–C340.

97. Abe E, Marians RC, Yu W, et al. TSH is a negative regulator of skeletal remodeling. Cell 2003; 115:151–162.

98. Bell A, Gagnon A, Dods P, et al. TSH signaling and cell survival in 3T3-L1 preadipocytes. Am J Physiol Cell Physiol 2002; 283:C1056–C10564.

99. Spaulding SW. The ways in which hormones change cyclic adenosine 3′,5′-monophosphate-dependent protein kinase subunits, and how such changes affect cell behavior. Endocr Rev 1993;14:632–650.

100. Porcellini A, Messina S, De Gregorio G, et al. The expression of the thyroid-stimulating hormone (TSH) receptor and the cAMP-dependent protein kinase RII beta regulatory subunit confers TSH-cAMP-dependent growth to mouse fibroblasts. J Biol Chem 2003;278:40621–40630.