In Vitro and in Vivo Regulation of Thyrotropin Receptor mRNA Levels in Dog and Human Thyroid Cells*

Regulation of thyrotropin (TSH) receptor (TSHr) mRNA accumulation as compared with two other thyroid differentiation markers (thyroglobulin and thy-roperoxidase (TPO)) has been investigated by North- ern blot. In dogs in uiuo, chronic stimulation of the thyroid by treatment with antithyroid drugs did not modify TSHr mRNA although it increased the levels of thyroglobulin and TPO mRNA. In dogs treated with thyroxin, the quiescent thyroids expressed normal levels of TSHr and TPO mRNA but depressed levels of thyroglobulin mRNA. In primary cultures of dog thyrocytes, dedifferentiation of the cells by treatment with epidermal growth factor or 12-0-tetradecanoyl-phorbol- 13-acetate led to decreased TSHr mRNA lev- els and nearly abolished thyroglobulin and TPO gene expression. However, TSHr mRNA was always pres- ent, compatible with the fact that these cells, when treated by TSH, reexpress differentiation. Treatment of the cells with TSH or forskolin transiently increased the TSHr mRNA level after 20 h, an effect inhibited by cycloheximide. This up-regulation was confirmed at the protein level: forskolin-treated cells showed an enhanced CAMP response to TSH and an increased binding thyroglobulin. Probes were stripped from the nylon membrane by washing for 1 h at 65 "C in 50% formamide, 10 mM sodium phosphate, pH 6.5 and for 15 min at room temperature in 2 X SSC, 0.1% SDS. CAMP Determinations-CAMP was determined using a radio- immunoassay method, as described (34). Binding Studies-For binding experiments, cells were harvested, and a membrane fraction was prepared (35). Binding of l*sI-labeled TSH (Trak, Republic of was performed at 4 "C for 120 min (23). Bound radioactivity was separated by centrifugation and measured. Nonspecific binding was determined in the presence of 30 milliunits/ml unlabeled TSH. All experiments were carried out at least twice, with, in each experiment, duplicates of cell Petri dishes (for culture experiments) and samples for each treatment.

Thyrotropin (TSH)l is the main agent regulating the thyroid gland (1-3). Stimulation of thyroid follicular cells by T S H promotes cell proliferation and stimulates the synthesis and secretion of the thyroid hormones, requiring an iodinated glycoprotein precursor, thyroglobulin, and the enzyme thyroperoxidase (TPO), two important markers of the differentiated state of the thyrocyte (4). TSH exerts most of its effects via a receptor positively coupled to adenylate cyclase (1-3).
* This work was supported by the Ministbre de la Politique Scientifique, Fonds National de la Recherche Scientifique, Association Belge contre le Cancer, Tklivie, and La Loterie Nationale. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
2-555- 46-55. Previous studies have demonstrated that TSH, via CAMP, increases the mRNA and the transcription of the thyroglobulin and TPO genes (5-9). Although receptors serve as regulators of cellular activities, they are themselves subject to regulation. Stimulation of a receptor is often followed by a desensitization involving negative regulation of the protein level or activity; the former control in general also involves decreased expression of the receptor gene (10-13). Nevertheless, up-regulation of receptors or receptor mRNA after agonist stimulation has also been demonstrated (14-16). In the case of the thyroid, the continuous stimulation of the TSH receptor by autoantibodies in Graves' disease or by excessive secretion of TSH causes hyperthyroidism and goiter, suggesting a very incomplete or nonexistent desensitization (17,18). The TSH receptor has been shown to be up-regulated in vitro (19) and in vivo (20) whereas other studies on FRTL, and human thyroid cells reported down-regulation (21,22). The recent cloning of the TSH receptor provided us with the opportunity to explore the control of thyrotropin receptor gene expression (23)(24)(25). TSHr mRNA levels were investigated in vivo in dogs and in vitro in primary cultures of dog and human thyroid cells subjected to a variety of stimuli and were compared with the amounts of thyroglobulin and TPO mRNA. Our results show that there is little modulation of TSHr mRNA by TSH, in agreement with the clinical observations; TSH through cyclic AMP transiently up-regulates and chronically down-regulates to a small extent TSHr mRNA. On the other hand dedifferentiation of thyroid cells by treatment with epidermal growth factor and phorbol esters greatly decreases TSH receptor gene expression.

MATERIALS AND METHODS
Animal studies were conducted in accordance with the highest standards of care. Dogs were treated for brain studies by oral administration for 4 weeks of methimazole ("1) (2 X 60 mg/day Strumazol) and propylthiouracil (PTU) (2 X 150 mg/day Propylthiurit) to increase the circulating TSH level, or by thyroxine (T,) (2 X 600 &day Elthyrone) to decrease it. Triiodothyronine (T3) and T, concentrations in the serum were followed by radioimmunoassay to ensure that the treatments were effective (T3, Amerlex; T4, Gammacoat, Clinical assays, Travenol Genentech Diagnostics, South San Francisco, CA. For low values, 20 pl instead of 10 pl was used in the assay).
On the day of the experiment, 1 h prior to thyroid resection, the animals received 50 mg/kg bromodeoxyuridine (BrdU) by intravenous injection. Dogs were anesthetized and the thyroid lobes resected. They were dissected free of connective tissue and frozen in small pieces in liquid nitrogen or cut in slices for histological or BrdU labeling analysis. For histology, classical Bouin-fixed, paraffin-embedded thyroid tissues were sectioned at 7 pm and stained by hematoxylin-eosin, green Masson trichrome, or periodic acid-Schiff methods. BrdU labeling analyses were performed as described (26). Two independent series of dogs were handled in this way.
Human thyroid cells, obtained from normal tissue adjacent to cold thyroid adenomas, were prepared in the same way as the dog cells and cultured in the same control medium containing in addition 1.25 pg/rnl transferin (Sigma) (29).
RNA Puri/ication-To purify the RNA, thyroid tissues were ground into a fine powder under liquid nitrogen and homogenized in 4 M guanidinium thiocyanate; thyroid cells were scraped from the culture dishes in the same solution. Lysed cells were layered over a cushion of 5.7 M cesium chloride in an SW56 tube and centrifuged at 36,000 rpm at 15 "C for 17 h. Total RNA was extracted as described by Chirgwin et al. (30). After treatment with 60 pg/ml proteinase K in 0.3 M, NaCl 0.1% SDS and phenol/chloroform extractions, the RNA was precipitated and resuspended in water for spectrophotometric quantification.
Northern Blot Analysis-After denaturation using glyoxal according to the procedure of McMaster and Carmichael (31) equal aliquots (10-15 pg) of total RNA were fractionated by electrophoresis on a 1% agarose gel in 10 mM phosphate buffer, pH 7. Acridine orange staining of independent lanes ascertained that the amounts of RNA were equal in all samples. Denaturated RNAs were transferred by diffusion blotting to a nylon membrane (Pall Biodyne A) using SSC X 20 (SSC X 1 = 0.15 M NaCI, 0.015 sodium citrate) as described (32). Prehybridization (4 h at 42 'C) and hybridization (overnight at 42 "C) were carried out in 50% formamide, 5 X Denhardt's (0.1% Ficoll, 0.1% polyvinylpyrrolidone), 5 X SSPE (20 X SSPE = 3.6 M NaCI, 0.2 M sodium phosphate, pH 8.3, 20 mM EDTA), 0.3% SDS, 250 pg/ml denatured salmon testes DNA, 200 pg/ml bovine serum albumin. The hybridization solution contained in addition 10% dextran sulfate (w/v) and the heat-denatured probe. cDNA probes (dog and human TSHr cDNA, 2.4-kb BamHI-XhoI insert of a pSVL construct; dog and human TPO cDNA, 2-kb BamHI and 1.9-kb EcoRI inserts of pBS constructs respectively, dog thyroglobulin cDNA, 5-kb EcoRI insert of a pBS construct) were a-'*P labeled by random priming extension to a specific activity of approximately loe cpm/pg (33). Filters were washed four times for 10 min in 2 X SSC, 0.1% SDS at room temperature and four times for 30 min in 0.1 X SSC, 0.1% SDS at 55 or 65 "C for heterologous or homologous hybridization, respectively. They were then autoradiographed at -70 "C using Hyperfilm MP (Amersham Corp.) and Siemens intensifying screens. The same filters were probed successively for TSHr, TPO, and thyroglobulin. Probes were stripped from the nylon membrane by washing for 1 h at 65 "C in 50% formamide, 10 mM sodium phosphate, pH 6.5 and for 15 min at room temperature in 2 X SSC, 0.1% SDS.
CAMP Determinations-CAMP was determined using a radioimmunoassay method, as described (34).
Binding Studies-For binding experiments, cells were harvested, and a membrane fraction was prepared (35). Binding of l*sI-labeled TSH (Trak, Henning GmbH, Berlin, Federal Republic of Germany) was performed at 4 "C for 120 min (23). Bound radioactivity was separated by centrifugation and measured. Nonspecific binding was determined in the presence of 30 milliunits/ml unlabeled TSH.
All experiments were carried out at least twice, with, in each experiment, duplicates of cell Petri dishes (for culture experiments) and samples for each treatment.

TSHr mRNA Accumulation in Viuo
Dogs treated with antithyroid drugs and 1-thyroxine were biologically hypo-and hyperthyroid at the time of thyroid resection. Serum TI levels were 3.3 and 6.8 pg/dl in T4-treated dogs, 0.8 and 0.5 pg/dl in MMI/PTU treated dogs, and 1.1, 2.6, 1.0, 1.2 pg/dl in control dogs for the first and second series, respectively. Serum Ts levels of the second series were < 30 ng/dl for MMI/PTU, 105 ng/dl for T,-treated dogs, and 31 and 42 ng/dl for control animals. Histological examination showed very hyperplasic thyroids in the antithyroid drugtreated animals with almost total disappearance of the colloid, and normal follicles full of colloid in the other thyroids. RrdU labeling analysis of the second series of dogs demonstrated nuclear labeling in 8.1% of the cells in the dog treated with antithyroid drugs and 0.09% and 0.24% of the cells in the control dogs. The T,-treated animals showed virtually absent labeling, less than 0.03% of the cells. There was therefore no doubt that the animals treated with antithyroid drugs were hypothyroid and that their thyroids were chronically hyperstimulated as a consequence. Conversely, thyroxin-treated dogs had higher thyroid hormone levels, and their thyroids were therefore quiescent.
Total RNA was extracted from the dog thyroids and subjected to Northern blot analysis. Fig. 1 shows the Northern blot corresponding to the second series of animals. As described previously (23), a 2.4-kb cDNA probe corresponding to the coding region of the dog TSHr hybridized to a 4.9-kb mRNA transcript. The levels of these transcripts did not differ greatly between normal and MMI/PTUor T,-treated dogs. The in-curved shape of the band is caused by the presence of the 28 S RNA, which has nearly the same size and so displaces the TSHr mRNA. The same blots were then hybridized with a dog TPO cDNA probe which revealed a major transcript around 4.1 k b a n d a minor one around 3.5 kb. These transcripts have already been reported (36,37). In the same way, a dog thyroglobulin cDNA probe hybridized to an approximately 8.5-kb mRNA, as described (38,39). In contrast to the small variation in TSHr mRNA levels the thyroglobulin mRNA level was clearly increased in MMI/ PTU and strongly decreased in T4-treated dogs, respectively, reflecting the treatment given. TPO mRNA level was increased in the MMI/PTU-treated dogs, but in T4-treated dogs the situation was not clear: in one experiment it was increased ( Fig. l ) , and in a second experiment (not shown) it was similar to the control levels.
Thus, modifying in uiuo the circulating concentrations of TSH mainly affected thyroglobulin mRNA, to a lesser extent TPO mRNA, but TSHr mRNA was only affected to a small extent. Regulation of Thyrotropin Receptor mRNA FCS and then maintained for 5 days in control medium without serum or supplemented with T S H (1 milliunit/ml), forskolin M), or EGF (25 ng/ml) + 10% FCS. Blot hybridization analyses were then performed to determine the levels of TSHr, thyroglobulin, and TPO mRNA after these incubations. Chronic stimulation by TSH or forskolin, leading t o a highly differentiated state (28), resulted in a slight downregulation of TSHr mRNA and in a high expression of TPO and thyroglobulin mRNA (Fig. 2). On the other hand, dedifferentiating the cells by EGF and FCS (28) reduced TSHr mRNA levels and in a more dramatic way TPO and thyroglobulin mRNA, whose levels became undetectable. The high basal TSHr mRNA level in the control cells is noteworthy by comparison with the very low thyroglobulin or TPO mRNA basal levels, the last one being undetectable.

TSHr mRNA Accumulation in Vitro Chronic Stimulation in Dog and
To investigate TSHr mRNA expression in human cells, human thyrocytes were seeded in 1% FCS-containing medium and then incubated (without serum) 2 days in the presence of forskolin M ) or in control medium followed by a 6-h incubation with forskolin M) or TSH (250 microunits/ ml). The 2-day forskolin incubation did not result in TSHr mRNA down-regulation, as the human TSHr mRNA levels were identical in control and forskolin-treated cells. As in dog thyrocytes and in other systems (36,37,40,41), thyroglobulin and TPO mRNA levels were strongly increased following the 2-day forskolin stimulation (Fig. 3). The 6-h treatment with forskolin or TSH did not modify TSHr and thyroglobulin mRNA levels but increased TPO mRNA levels, as expected (5,7). Contrary to the dog TPO transcripts, the major human T P O mRNA appeared around 3.5 kb. This transcript has been described as generating the protein (42). The longer transcript could be a precursor.
Thus, except for the slight down-regulation of TSHr mRNA, the data obtained in human and dog thyrocytes submitted to chronic activation of the cyclic AMP cascade show little variation in TSHr mRNA levels as opposed to high variations in TPO and thyroglobulin mRNA.
Accumulation of TSHr mRNA in Response to Short Time Exposures to Various Agents-Dog thyrocytes were seeded (day 0) and maintained for 1 day in control medium supplemented with 1% FCS. T o get enough material, cell prolifera- tion was then stimulated by EGF (25 ng/ml) and 10% FCS (day 1). After 3 days of such treatment, confluence was achieved while cells lost most of their differentiated characteristics (28, 40) (day 4). Expression of the differentiated functions is restored by washing out the EGF and serum for 2 days followed by addition of an adenylate cyclase activator (at day 6) (28, 40). Fig. 4A shows that TSHr mRNA levels were low but still detectable after a %day treatment with EGF and FCS; they were increased again after a further 2 days in control medium. All the experiments described were performed a t day 6.
A 20-h incubation with TSH (1 milliunit/rnl) or forskolin (lo-" M ) led to an increase in TSHr, thyroglobulin, and TPO mRNA levels whereas treatment with EGF (25 ng/ml) or TPA (10 ng/ml) markedly reduced thyroglobulin and TSHr mRNA, more so for TPA than EGF for the TSHr mRNA (Fig. 4A). The kinetics of TSHr mRNA induction in the presence of TSH or forskolin showed that the increase culminated at 20 h (Fig. 4R). This up-regulation was short lived; indeed, Fig. 4C shows that the mRNA levels had already decreased after 24 h. Longer exposure times resulted in TSHr mRNA levels equal to or lower than the control levels ( Fig. 4, B and C). Similar results have been obtained by in situ hybridization.' By contrast and as expected (5, 7), TPO mRNA became apparent after 6 h, and thyroglobulin mRNA content was strongly enhanced after 15 h. A similar time course for TSHr mRNA increase was observed when the cells were seeded and maintained for 4 days in the control medium, without strong proliferative and dedifferentiating pretreatment (not shown).
The increase in TSHr mRNA reflects also an increase in TSH receptors themselves. Indeed, in the same culture conditions as above the capacity of thyroid cells to respond to a TSH (1 milliunit/ml) stimulation by a cAMP elevation became higher when these cells were first pretreated with forskolin. Table I shows that a 15-or 60-min TSH (1 rnilliunit/ ml) incubation of control cells led to a 9-or %fold increase in cAMP intracellular concentrations, respectively, but that these increases became 45-or 68-fold, respectively, when the cells were first pretreated 2 days with forskolin (10 or 5 p M ) (from day 6). An increase in TSH receptors is also suggested * V. Pohl. submitted for puhlication.  by binding experiments performed on cells submitted to the same protocol with the last 2 days in control medium or in forskolin M)-containing medium; a 2.6-and 2.9-fold increase in binding was observed after the fonkolin treatment ( Table I). Sensitivity to Cycloheximide-To determine whether TSH and forskolin exert their actions through newly synthesized protein(s), cells (at day 6) were incubated for 20 h in the presence of cycloheximide (10 pg/ml), alone or with TSH. At this concentration, cycloheximide inhibited more than 90% ["S]methionine incorporation in total proteins (not shown).
Cycloheximide alone markedly reduced TSHr mRNA basal levels, and simultaneous addition of cycloheximide and TSH resulted in an even more pronounced decrease (Fig. 5). This suggests that protein synthesis is involved not only in the TSH stimulation of TSHr but also in the maintainance of the basal levels. As expected (5, 7 ) , the thyroglobulin mRNA increase was also blocked by cycloheximide whereas TPO mRNA remained unaffected under these conditions (not shown). Washing of the cells after cycloheximide treatment is followed by recovery of thyroglobulin mRNA expression (5).
Role of Insulin-Insulin, acting presumably through insulin-like growth factor receptors (43), is an important regulatory factor for dog thyroid cells in culture. A role of insulin has clearly been demonstrated in thyroglobulin gene expression but not in TPO gene expression (Ti, 7 , 9, [44][45][46]  investigate the role of this hormone on TSHr mRNA, cells were incubated in the presence or absence of insulin (5 pg/ ml) during the 2 days following EGF and FCS treatment and during the TSH stimulation. Fig. 6 shows that the basal and TSH-enhanced levels of TSHr mRNA were not significantly different in the absence or in the presence of insulin. Only the basal level might be slightly lowered in absence of this agent. Consequently, the effects of this hormone on TSHr mRNA appeared to be relatively weak or nonexistent by comparison with its effects on thyroglobulin mRNA since hybridization of the same blots with the thyroglobulin probe showed a clear and expected effect of insulin.
Stability of TSHr rnRNA-To estimate the stability of TSHr mRNA, cells were treated with actinomycin D (5 pg/ ml), an inhibitor of transcription. As the TSHr mRNA basal levels were easily detectable, actinomycin D was added to cells in control medium (at day 6) for several periods of time. The addition of this agent for periods up to 8 h did not lead to a marked decrease in TSHr mRNA basal levels (Fig. 7), but after a 16-h treatment a strong decrease could be observed, and after 20 h the mRNA levels were no longer detectable (barely detectable after overexposure; not shown). This suggests that the TSHr mRNA is relatively stable since its halflife can be estimated to be of the order of 10-15 h.

DISCUSSION
In the response of cells to extracellular signals, two elements are involved in the primary event: the signal and the receptor. The cell can modulate its response to any factor by controlling the number or the activity of receptors available to this factor. When this results in a decreased response the process is called desensitization. One of the mechanisms involved is the downregulation, i.e. the decrease in the number of receptors which may result from the regulation of any step involved in receptor gene expression or turnover. Changes in the steady-state levels of the receptor mRNA have been shown to be induced by the ligand. Treatment of GHa cells by thyrotropin-releasing hormone (13), of MCF-7 cells by estradiol (12), of rabbit aortic smooth muscle cells by norepinephrine (ll), and of hamster smooth muscle DDTIMF-2 cells with @-adrenergic agonists (10) all result in a decrease in both the receptors mRNA levels and, when measured, in the number of receptors themselves. The physiological meaning of such down-regulation is clear: it represents a feedback mechanism by which a cell decreases and shortens its response to persistent stimulation.
However, down-regulation is not a general rule. In DDTIMF-2 cells, Collins et al. (47) observed an increase in &-adrenergic receptor mRNA after a short time exposure to epinephrine as well as a down-regulation after prolonged exposure. Similar kinetics have been obtained for the luteinizing hormone receptor in 8-bromo-CAMP-treated MA-10 Leydig cells (48). An upregulation has also been observed for the mRNAs of EGF receptor (15) and growth hormone receptor (16).
Northern blot analysis were performed to study the regulation of TSHr mRNA in in uiuo treated dogs and in oiho primary cultures of dog and human thyroid cells. Validation of our results was made by comparison of the TSHr mRNA regulation with the previously studied TPO and thyroglobulin mRNA regulations. As we have shown previously, TSHr mRNA hybridization revealed a 4.9-kb transcript and two transcripts of 4.6 and 4.4 kb in dog and human thyroid cells, respectively. TPO mRNA hybridization revealed several transcripts (approximately 4.1 and 3.5 kb) in dog and human cells, as described previously (36,37). The smaller transcripts (2.2 and 1.7 kb), described in human cells (36) and cloned by Nagayama et al. (49), were not observed because the cDNA probe used in our experiments is a 3' probe whereas the 2.2and 1.7-kb mRNA represent the 5' end of TPO mRNA. Thyroglobulin mRNA transcripts appeared around 8.5 kb in dog and human cells.
Modulation of the TSH circulating concentrations by treat-ing dogs with MMI/PTU or T4 resulted in little variation in TSHr mRNA. On the other hand, thyroglobulin and TPO mRNA levels were increased in the MMI/PTU-treated dogs, but only thyroglobulin mRNA levels showed a considerable decrease in the T4-treated dogs. Lack of material precluded the measurement of TSH binding, which still leaves open the possibility of a down-regulation of receptors. However, in guinea pigs no such down-regulation was observed (20). The increase in thyroglobulin mRNA observed in the MMI/PTUtreated dogs was not reported in PTU-treated rats (50,51) where it was shown that the thyroglobulin gene was maximally transcribed under normal TSH levels. It is probable that this reflects species difference such that, as expected, thyroglobulin gene expression can be up-regulated by TSH, but this phenomenon is not observed in the normally highly active rat thyroid.
In primary cultures of dog and human thyroid cells, TSH through cAMP promotes function, proliferation, and differentiation (3, 28, 40) whereas EGF, TPA, or serum induces mitogenesis but represses differentiation expression (28,40,52). Chronic stimulation of dog thyrocytes for 5 days by TSH or forskolin led to a slight down-regulation of TSHr mRNA. The TSH-promoted down-regulation is homologous, and the forskolin-promoted down-regulation is heterologous since the TSH receptor itself is not involved in this process. The latter effect suggests that cyclic AMP is involved in TSH receptor mRNA down-regulation. TPO and thyroglobulin mRNA were on the contrary highly expressed in these cells, which is compatible with a highly differentiated state as also demonstrated by high levels of iodide transport (28). Human thyrocytes cultured 2 days in the presence of forskolin showed similar results except that no down-regulation was observed for the human TSHr mRNA. Thus, in both species, little variation in TSHr mRNA levels was observed by comparison with the high variations in TPO and thyroglobulin mRNA, already described in dog (5,40), calf (53), and human (36,41) thyroid cells and in FRTL,cells (7,54,55). Chronic exposure to EGF and FCS induced a striking fibroblast-like morphology and decreased the differentiation characteristics (iodide trapping and organification, thyroglobulin gene expression) (28,40). TSHr mRNA levels were strongly repressed, and TPO and thyroglobulin mRNA became undetectable. Thus, despite the fact that the cells became highly dedifferentiated, they still retained detectable amounts of TSHr mRNA. The changes induced by EGF were reversible after its removal; and in the presence of TSH, normal epithelial morphology was then restored. Thyroid cells, even dedifferentiated, never turn the TSHr gene completely off, they always keep expressing the gene which will allow them to return to differentiation upon stimulation by TSH. This is compatible with data obtained on human thyroid tumors,3 demonstrating that the TSH receptor is the last differentiated characteristic to be lost in increasingly dedifferentiated thyroid neoplasia. A relationship between TSHr mRNA expression and the degree of dedifferentiation was also reported by Ohta et al. (56) in neoplastic human thyroid tissue. This persistence of TSHr mRNA under all these conditions explains why these cells, even when apparently undifferentiated, are still able to respond to TSH in terms of differentiation, proliferation, and function.
Short term incubations of dog thyrocytes with TSH or forskolin resulted in an up-regulation of TSHr mRNA; its levels were maximally increased until 20 h and decreased thereafter to levels equal or lower than the control levels. In the same cells, TPO and thyroglobulin mRNA levels were G. Brabant, submitted for publication. enhanced after 6 and 10-15 h, respectively, in accordance with other studies on thyroid cells (5, 7). Treatment (20 h) with EGF or TPA markedly reduced TSHr, TPO, and thyroglobulin mRNA. The decrease observed was stronger in the presence of TPA, which could be correlated with the stronger inhibitory effect of this agent on the expression of differentiation in these cells (9).
The observed transient up-regulation of TSHr mRNA contrasts with previous report on the regulatory effects of TSH on its own mRNA; in FRTLs cells, Akamizu et al. (21) reported a down-regulation within 8 h of TSHr mRNA in response to agents increasing cAMP levels. This represents another of the multiple differences in regulation between this cell line and thyroid cells in vivo or in primary culture (57,58). In transfected Chinese hamster ovary cells, down-regulation of TSHr mRNA was not observed (22). In dog and human thyroid cells, desensitization, i.e. a moderate decrease in cAMP response to TSH, is observed after 2 h until 16 h (59). This phenomenon, impaired by cooling the cells and resistant to protein synthesis inhibition, presumably takes place at the level of the receptor itself (60). In DDTIMF-2 cells it has been demonstrated that cAMP elevation after receptor activation stimulated the &adrenergic receptor gene and also led to receptor phosphorylation (on specific protein kinase A phosphorylation sites) involving receptor desensitization (61). Similarly, a rise in cyclic AMP induces both an increased catabolism of the R and C subunits of cyclic AMPdependent protein kinase and thus a decrease in enzyme level and an increase in the level of the corresponding mRNA (62). It is tempting to consider a similar autoregulation by cAMP of the TSH receptor. As suggested by Collins et al. (61) this would make physiological sense: the transient increase in TSHr mRNA might allow the cell to maintain receptor number, by compensating an eventual loss of receptors.
Nevertheless, although in DDTIMF-2 cells the number of &-adrenergic receptors did not increase with its mRNA level (47). In MA-10 cells, on the contrary, luteinizing hormone/ chorionic gonadotropin receptors are increased (48), and in dog thyrocytes the increase in TSHr mRNA is also accompanied by an increase in TSH receptors themselves. Cells treated with forskolin showed an enhanced cAMP response to a TSH stimulation by comparison with untreated cells, and binding experiments revealed higher levels of TSH receptor sites in forskolin-treated cells than in control cells. This suggests that like ACTH, which positively regulates its own receptors and cAMP response in cultured bovine adrenal cells (14), the TSH-induced cyclic AMP accumulation, as mimicked by forskolin, has a positive effect on both TSH receptor number and response. These receptors appear also to be very stable and little desensitized since this response is maintained even after a 2-day forskolin treatment while TSHr mRNA had already declined. Increased content of G,, the CY subunit of the G protein which activates adenylate cyclase, might contribute to the higher cyclic AMP response of TSH-and forskolin-treated cells (63). The long latency time observed before forskolin or TSH could increase TSHr mRNA levels suggested that prior protein synthesis might be required as in the case of thyroglobulin gene induction (5, 7). To test this hypothesis, experiments were performed in the presence of cycloheximide. Cycloheximide lowered the basal TSHr mRNA levels and completely abolished induction by TSH, suggesting that newly synthesized protein(s) are involved not only in the TSH enhancement of TSHr mRNA level but also in maintaining the basal levels.
In our model of primary culture of dog thyroid cells, insulin markedly potentiated the action of forskolin and TSH on the rate of DNA synthesis (64). Insulin, acting presumably through insulin-like growth factor I receptors (43), has been shown to be involved in the control of thyroglobulin but not of TPO gene expression (5,7,9,(44)(45)(46). However, the TSHr mRNA levels were not significantly different with or without insulin. This is in agreement with data4 demonstrating that the cAMP response to TSH is identical in the presence or absence of insulin. By comparison, rehybridization of the same blots with a thyroglobulin probe clearly demonstrated that the basal and a part of the stimulated levels of thyroglobulin gene expression depended on the presence of insulin in the culture medium, in agreement with the concept of a multihormonal regulation of the thyroglobulin gene (9, 45, 46).
Experiments performed in the presence of actinomycin D showed that the TSHr mRNA in control medium was very stable, since about a 16-h incubation in the presence of this drug was necessary to observe a decline in TSHr mRNA. Half-lives of the same order of magnitude have been observed for the &adrenergic receptor mRNA (12 h) (65). Moreover, microinjection of recombinant TSHr mRNA (23) in Y1 cells conferred a TSH-responsive phenotype on them, which was maintained for at least 20 h, suggesting also that receptor and/or mRNA were stable. Taken together with the transient TSH-promoted stimulation of TSHr mRNA (observed at 15 and 20 h and ended by 24 h), these data suggest that an active mechanism is involved in the rapid decline of the raised mRNA level.
In summary, we presented here data on TSHr mRNA regulation which we compared with two other well known thyroid differentiation markers: TPO and thyroglobulin. Regulation of TPO and thyroglobulin mRNA, used to validate our results on TSHr mRNA, is in accordance with previous results obtained in our laboratory and in others. TSHr mRNA expression is clearly linked to differentiation expression although it is never abolished with dedifferentiating treatments. A physiological meaning of this persistence under dedifferentiation conditions could be that contrary to the genes it controls, the TSH receptor represents the identity and address of this cell, i.e. its ability to respond to its specific signal and to function in the physiological regulatory network. TSHr mRNA is relatively little modulated by TSH in vivo as in vitro although a short up-regulation has been demonstrated. As for thyroglobulin mRNA, the responses to TSH are cyclic AMP mediated, slow, and cycloheximide sensitive. According to the classification of the CAMP-regulated genes by Roesler et al. (66) (one group characterized by a rapid response to CAMP, insensitive to cycloheximide, and a second group characterized by a slow response, sensitive to cycloheximide), the TSHr gene would belong to the second group, which includes the thyroglobulin gene. Our data on TSH receptors demonstrate that gene expression accounts at least in part for regulation of receptor expression. The absence of important down-regulation is compatible with clinical observations on hyperthyroidism. The mechanisms of changes in the TSHr mRNA (transcriptional activation or post-transcriptional stabilization) and the regulation of the protein itself remain open questions.