The Expression of the Thyroid-stimulating Hormone (TSH) Receptor and the cAMP-dependent Protein Kinase RII β Regulatory Subunit Confers TSH-cAMP-dependent Growth to Mouse Fibroblasts*

TSH activates its specific receptor in thyroid cells and induces cAMP, a robust stimulator of thyroid cell proliferation. Conversely, cAMP is a potent inhibitor of growth in mouse fibroblasts. To dissect the signals mediating cAMP-dependent growth, we have expressed in mouse fibroblasts the human thyrotropin receptor (TSHR) or a constitutively active mutant, under the control of the tetracyclin promoter. Both TSHR and cAMP levels were modulated by tetracyclin. In the presence of serum, activation of TSHR by TSH induced growth arrest. In the absence of serum, cells expressing TSHR stimulated with TSH, replicated their DNA, but underwent apoptosis. Co-expression of cAMP-dependent protein kinase (PKA) regulatory subunit type II (RIIβ) inhibited apoptosis and stimulated the growth of cells only in the presence of TSH. Expression of RIIβ-PKA, in the absence of TSHR, induced apoptosis, which was reversed by cAMP. Growth, stimulated by TSHR-RIIβ-PKA in mouse fibroblasts, was also dependent on Rap1 activity, indicating cAMP-dependent growth in thyroid cells. As for the molecular mechanism underlying these effects, we found that in normal fibroblasts, TSH induced AKT and ERK1/2 only in cells expressing TSHR and RII. Similarly, activation of TSHR increased cAMP levels greatly, but was unable to stimulate CREB phosphorylation and transcription of cAMP-induced genes in the absence of RII. These data provide a simple explanation for the anti-proliferative and proliferative effects of cAMP in different cell types and indicate that RII-PKAII complements TSHR action by stably propagating robust cAMP signals in cell compartments.

cAMP stimulates or inhibits the growth of many cell types (1)(2)(3)(4) Thyroid cells are exquisitely dependent on TSH 1 for growth and differentiation (5,6). cAMP mediates most TSH effects (7). Conversely, cAMP arrests mouse 3T3 fibroblasts in G 1 (8). In both cases, the downstream effects of cAMP are mediated by PKA (9) and by cAMP-binding proteins, GEF, that stimulate GTP binding effectors, such as Rap1 (10). Rap1 has been shown to mediate both positive and negative effects of cAMP on cell proliferation (11)(12)(13). Paradoxically, the same downstream targets appear to mediate negative or positive cAMP effects. The cdk2 inhibitor p27 has been shown to be induced by cAMP-PKA both in thyroid cells (14) and in other cell lines arrested in G 1 by cAMP (15,16).
To dissect the signals involved in cAMP stimulation of growth, we have expressed the TSH receptor (TSHR) in mouse fibroblasts under the control of an inducible promoter. Note that a previous study reported the permanent expression of the receptor in mouse fibroblasts, stably transfected with wild-type TSHR and constitutively active mutants. Under these conditions, a strong selection was applied to the cells, and the phenotypes of the clones were not easily interpretable in terms of cAMP-dependent growth (17).
We report here that under conditional expression of TSHR, only the co-expression of RII␤ subunit of PKA and TSHR resulted in TSH-dependent growth. Both TSHR or RII␤ PKA expressed alone induced apoptosis or growth arrest. Their growth-stimulating activity was dependent on the ability not only to generate but to propagate cAMP signals to downstream nuclear substrates.

EXPERIMENTAL PROCEDURES
Materials and Reagents-Unless otherwise specified, drugs and chemicals were obtained from Sigma, and cell culture supplies were purchased from standard suppliers, e.g. Falcon, Invitrogen, Hyclone.
Cell Lines-Mouse fibroblasts NIH 3T3 were grown in Dulbecco's modified Eagle's medium containing 10% fetal calf serum and 2 mM glutamine (standard medium). The tTA-repressor-expressing clones were grown in standard medium containing puromycin (2.5 g/ml), tetracyclin (1.0 g/ml), and geneticin (G418) (200 g/ml). The selection and tetracyclin were removed 48 h before starting the experimental procedures described below.
For each NTC line, we have generated permanent clones, which expressed the regulative subunit of the PKA type II ␤ (RII␤) by transfecting a plasmid encoding the human RII␤ cDNA (3.6-kb EcoRI fragment) under the control of the CMV promoter, carrying the geneticin resistance gene (G-418) (20).
For cAMP-dependent transcription assay each clone was transiently transfected in triplicte samples by the calcium phosphate procedure (21) with 5 g of pRSV-LacZ and 5 g of pCRE-CAT (22). 24 h after transfection, cells were serum-starved for 48 h (0.5% fetal bovine serum). Control cells were stimulated with forskolin (25 M) for 4 h. The cells were harvested and total protein extracted (see below). CAT enzymatic assay was performed as described (21). Transfection efficiency was normalized for ␤-galactosidase assay.
Total RNA Extraction and Northen Blot Analysis-Total RNA was extracted from 10 7 cells grown in a 150-mm dish by the guanidium isothiocyanate procedure; 20 g of each RNA sample was electrophoresed on a 1% agarose gel containing formaldehyde and transferred to nylon membranes (Amersham Biosciences) using standard capillary techniques. The filter was then hybridized with specific radioactive probes overnight at 65°C in Church Buffer (0.5 M Na 2 PO 4 , pH 7.2, 7% SDS, 1 mM EDTA). Radioactivity was revealed by autoradiography using preflashed films.
The following probes were used: a 0.9-kb DNA fragment corresponding to hTSHR cDNA region coding for codons 58 -381; and the fulllength rat glyceraldehyde-3-phosphate dehydrogenase cDNA. All probes were labeled with [ 32 P]dGTP by nick translation.
Quantitative analysis of the hTSHR mRNA amount was performed by densitometry using the NIH Image software for Apple Macintosh.
cAMP Assays-For each experimental point 2 ϫ 10 5 cells of each clone were seeded in a 6-well tissue culture tray. After 24 h the cells were serum-starved for 18 h; then serum-free medium with 0.5 mM IBMX and different TSH concentrations (0, 0.1, 1, and 10 mU/ml) were added for 40 min. Cells were lysed with cold ethanol containing 1% 1 M HCl and incubated overnight at ϩ4°C. The medium was collected, neutralized, lyophilized, and resuspended in 0.2 ml of 10 mM Tris-HCl pH 7.8, 1 mM EDTA and used for cAMP assay (Amersham Biosciences, cyclic AMP 3 H assay system). Each experiment was repeated three times in triplicate samples.
Growth Analysis-Proliferation was analyzed under different conditions: normal medium or low serum (0.5%) with and without TSH 10 mU/ml. For each experiment 10 4 cells were seeded in a 12-well tissue culture tray; cells were collected and counted after 24 h or 3, 6, 9, or 15 days.
The MTT colorimetric assay for proliferation was performed in a 96-well flat-bottomed tissue culture tray for each time point: 2500 cells for each clone were seeded in eight replicas for each experimental point; 6 h after plating the standard medium was removed, and 100 l of culture medium without serum were applied. The determinations were carried out every 24 h as follow: 0.01 ml of MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide) stock (5 mg/ml) was added to each well. After 4 h of incubation at 37°C the medium was removed and 0.1 ml of isopropyl alcohol/0.04 N HCl was added. The absorbance was measured after 1 h on an ELISA plate reader. We have used a Dynatech MR580 reader with a test wavelength of 570 nm and a reference wavelength of 630 nm.
[ 3 H]thymidine incorporation assays were carried out as follows. 5000 cells were seeded in a 96-well flat-bottomed tissue culture tray. After 72 h of growth in the absence of tetracyclin serum was removed for 18 h (0.5% fetal bovine serum). The following day the low serum medium was replaced by standard medium, and [ 3 H]thymidine (0.5 Ci) and TSH 10 mU/ml (as required) were added. Cells were collected 6, 12, and 18 h after stimulation. DNA was extracted with the Cell Harvester; radioactivity was determined by liquid scintillation. All experiments were repeated three times in triplicate.
Cell cycle analysis was carried out by FACS. 10 6 cells were plated in 100-mm dishes; 8 h after seeding, cells were subject to different treatments. After 48 h, 2.5-3 ϫ 10 6 cells were collected, washed, and resuspended in 1 ml of PBS. Ethanol fixation was carried out by adding 10 ml of ice-cold 70% ethanol. After 3 h, fixed cells were washed and stained for 30 min at room temperature in 0.1% Triton X-100, 0.2 mg/ml DNase-free RNaseA, 20 g/ml propidium iodide. Fluorescence was determined by FACS and analyzed by Cell Fit Cell-Cycle Analysis Version 2.
Protein Extraction and Western Blot Analysis-10 7 cells were lysed in TBS, 1% Triton X-100 lysis buffer (150 mM NaCl, 50 mM Tris-HCl, pH 8, 5 mM EDTA, 1 mM NaF, 1 mM Na 4 P 2 O 7 , 1.5 mM KH 2 PO 4 , 0.4 mM Na 3 VO 4 ) and incubated on ice for 20 min. After centrifugation at 12,000 ϫ g, the protein concentrations were determined using the BioRad protein assay reagent (Bio-Rad, Hercules, CA). 100 g of protein were added to an equal volume of 2ϫ sample buffer (125 mM Tris-HCl, pH 6.8, 4% SDS, 10% glycerol, 0.006% bromphenol blue, 2% mercaptoethanol), boiled for 5 min, separated on 7% SDS-PAGE, and electrophoretically transferred onto nitrocellulose membranes (Schleicher & Schuell, Germany). After transferring, Ponceau staining was used to confirm equal protein loading. The membranes were incubated for 1 h at room temperature with PBS containing 5% nonfat dry milk (Bio-Rad) and then incubated overnight at 4°C in PBS/2.5% nonfat dry milk containing 1 g/ml of specific antibody. The membranes were washed twice with PBS and reincubated for 1 h with horseradish peroxidase-conjugated anti-rabbit or anti-mouse secondary antibody (Amersham Biosciences). Signals were detected using enhanced chemiluminescence according to the manufacturer's instruction (ECL detection system; Amersham Biosciences). Anti-rat PCREB antibodies (23) were purchased from Upstate Biotechnology, Lake Placid, NY. Specific anti-RII␤ antibodies were generated by immunizing rabbits with a synthetic RII␤ peptide (peptide 31-57 from the AUG of the rat sequence), cross-linked to soybean trypsin inhibitor (24). Anti-Rap-1b was purchased from Santa Cruz Biotechnology, Santa Cruz, CA.
FIG. 1. Conditional expression of human TSH receptor under the control of the tet promoter. Northern blot analysis of two clones stable transfected, respectively, with the wild-type (TSHR) and the T632I mutant variant of the human TSH receptor. Cells were grown into 150-mm dishes for 48 h in standard medium without or with tetracyclin at the indicated concentration. Maximal inhibition was observed at 1.0 g/ml drug. Left panel, TSHR mRNA levels in the same clones after 15 day of culture in the absence of tetracyclin. All samples were normalized for loading and for glyceraldehyde-3-phosphate dehydrogenase mRNA expression (lower panel).
TUNEL Assay-10 4 cells were grown on slides (ordinary glass slides washed in ethanol and soaked in 0.01% poly-L-lysine). After treatment, slides were fixed in 4% paraformaldehyde/1ϫ PBS, 10 min, at room temperature and washed one time in PBS ϩ 50 mM glycine, 10 min at room temperature, and three times for 5 min in PBS. Cells were permeabilized with 0.5% Triton X-100, 1ϫ PBS for 10 min, washed 3 ϫ 5 min in PBS, and equilibrated with 200 l of 1ϫ TdT buffer ϩ 1 mM cobalt chloride (under a 60 ϫ 24 mm coverslip), for 5 min. The TUNEL reaction was carried out using the In Situ Cell Death Detection Kit (Roche Applied Science). Apoptotic rate was determined by FACS analysis on propidium iodide-stained cells.
For FACS analysis, after 48 of tet starvation, 10 6 cells for each clone were plated in 100-mm dishes and starved from serum for 18 h. Terminal transferase reaction (Roche Applied Science) and staining with antiBRDU-FITC (BD Pharmingen) was performed on paraformaldheyde/ethanol-fixed cells, according to the manufacturer's instructions. After TUNEL reaction, cells were washed and stained for 30 min at room temperature in 0.1% Triton X-100, 0.2 mg/ml DNase-free RNaseA, 20 g/ml propidium iodide. Fluorescence was determined by FACS and analyzed by Cell Quest software.
Recombinant Adenovirus Expressing the Dominant-negative RAP-1b Variant-RAP-1b dominant-negative mutant plasmid was previously described (13). The cDNA insert was subcloned into the multiple cloning site of the shuttle plasmid (pAd-CMV-TRK) by standard cloning procedures. The purified shuttle plasmid was digested with the restriction enzymes PmeI to obtain the "rescue" fragment. The fragment was then purified on agarose gel, and 2 g of purified rescue fragment was used for homologous recombination. The adenoviral plasmid pAdEasy-1 (25) was then mixed with the rescue fragment, and the DNA mixture was transformed into the BJ5183 bacterial strain and incubated overnight. Colonies were screened by digesting the DNA with BglII and performing a Southern blot to confirm the presence of the cDNA insert. The DNA with the proper orientation was transformed into the DH5␣ bacterial strain. The recombinant construct, purified using Qiagen Maxi preparation kit, was digested overnight with PacI and transfected into 293 cells using LipofectAMINE (Invitrogen). Adenoviral plaques were twice purified by infecting 293 cells in agar. Virus was purified by CsCl gradient centrifugation, dialyzed and titrated by plaque assay.

Conditional Expression of Wild-type and Mutated TSH Receptor in Mouse
Fibroblasts-TSH receptor was cloned under the control of the inducible tetracyclin promoter. In the configuration we have used, Tet-off (see "Experimental Procedures"), the tetracyclin repressor (tTA), in the presence of the tetracyclin, inhibits the transcription of the downstream gene. Removal of the drug from the medium results in the activation of transcription. We transfected NIH3T3 fibroblasts with the tTA repressor, isolated several clones, and re-transfected them with the Tet-TSHReceptor (TSHR) construct. Several clones were isolated and screened for inhibition of expression of TSHR in the presence of tetracyclin (tet). Fig. 1 shows the levels of TSHR mRNA in two clones, transfected with the wild-type TSHR and the constitutively active mutant T632I, which contains a mutated residue in position 632 (isoleucine mutated to threonine).
TSHR mRNA levels were critically dependent on the time and the concentration of tet added to the medium. Maximal expression was observed 12-24 h after removal of the drug and it remained constant for 15 days of continuous culture in the absence of the drug (Fig. 1). Down-regulation of TSHR mRNA by tet was maximal after 48 -72 h of continuous drug exposure.
To test if TSHR was functionally coupled to G s , we determined the levels of intracellular cAMP. In cells, expressing the TSHR (i.e. in the absence of tet), basal cAMP levels were unchanged, while TSH induced a rapid (30 min) accumulation of cAMP. We screened for clones displaying the same range of cAMP response found in thyroid cells (FRTL-5, see Ref. 26). TSH induction of cAMP was abolished by tet treatment (Fig. 2, panel c). In cells transfected with the constitutively active receptor T632I, basal cAMP levels were increased when the receptor was expressed (i.e. when tet was removed from the medium) (Fig. 2, panel d). In these cells, TSH further increased cAMP levels (Fig. 2, panels c and d) (19).
These data validate our experimental model, since cAMP levels mirror the expression of both the TSHR and T632I receptors. Thus, cAMP basal levels were higher in cells expressing T632I than in cells expressing TSHR in the absence of TSH. This indicates that the wild-type receptor was not expressed at levels that induced its activity in the absence of TSH.
Expression of TSH Receptor Inhibits DNA Synthesis and Growth of Mouse Fibroblasts-It is known that cAMP inhibits the growth of mouse fibroblasts (27)(28)(29). We asked whether the regulated production of cAMP by G s -coupled TSHR could modulate the growth of NIH3T3 fibroblasts. To this end, we have used the wild-type receptor, which activates cAMP production only when exogenous TSH is present, or a mutant version of the receptor, T632I, which constitutively activates adenylyl cyclase in the absence TSH (26,19). Growth of fibroblasts was measured under different conditions (in the absence or presence of tet, serum, or TSH). A general scheme illustrating the temporal order of the various assays is shown in Fig. 3. Under each experimental condition indicated in Fig. 3, we have monitored the expression of TSHR and cAMP levels, following the treatment of the cell with 1 mU TSH for 40 min, to ascertain that the observed effects were not caused by variations in TSHR levels or activity (data not shown). Tetracyclin, tet, at the concentrations used (1.0 g/ml) did not alter the growth of control (cells expressing the tTA-repressor only) and TSHRexpressing fibroblasts (Fig. 4, panels a and b). Control cells displayed a doubling time of ϳ28.4 h. When tet was removed from the medium, the growth of TSHR expressing cells was slightly reduced (doubling time ϭ 29.3 h). Continuous exposure of these cells to TSH (10 mU/ml) inhibited robustly cell growth (Fig. 4, panel b). The doubling time was 42 h, 45% longer than the control. This inhibition was dependent on the expression of TSHR, because it was abolished by tet and was strictly dependent on TSH in cell lines expressing TSHR (Fig. 4, panels a and  b). In cells transfected with T632I, the expression of the active receptor, following removal of tet from the medium, inhibited cell growth in the absence of TSH (Fig. 4, panel c). These data were confirmed by thymidine incorporation curves (Fig. 4, panels d-f) and by MTT assay (data not shown). FACS analysis confirmed that inhibition by TSH was caused by delayed S phase entry (Fig. 5C). The data shown in Figs. 2 and 4 indicate that in our experimental settings, the biological effects on cAMP and DNA synthesis were dependent on TSH. Note that in the absence of TSH, the growth and cAMP levels were similar between TSHR transfected and control cell.
Co-expression of PKA Regulatory Subunit, RII␤, Reverses Growth Inhibition by TSHR-A common feature of thyroid cells and other cell types, stimulated by cAMP, is the robust expression of type II regulatory PKA subunits, in particular, the RII␤ subtype. RII␤ binds to A kinase anchor proteins (AKAPs) with high affinity, is associated to membranes and its affinity for cAMP is lower than that of the RI subunits (14,24,30). Mouse 3T3 fibroblasts contain mostly RI subunits, which form cytosolic PKAI holoenzyme. These cells express low levels of RII␤ subunit (data not shown and Ref. 31). To define the role of RII␤ subunit in TSHR signaling, we have investigated the proliferative response of cell lines expressing the receptor and/or RII␤ subunit. Starting from cell expressing tTA, TSHR, and T632I, we have generated new cell lines stably transfected with RII␤ subunit (Fig. 5A). The expression of RII␤ was not influenced by serum, TSH, or TSHR and it did not modify basal cAMP levels (see legend of Fig. 5 and data not shown). Expression of RII␤ subunit inhibited cell proliferation (Fig. 5B, panel  a). High RII levels shifts the ratio PKAI/PKAII (Refs. 31 and 32 and data not shown). PKAII enzyme, formed by RII-C (C, catalytic subunit), has a lower affinity to cAMP than PKAI (RI-C) (30). The levels of cAMP in growing fibroblasts were not sufficient to activate efficiently the excess of PKAII in the absence of TSHR and TSH (Fig. 5B, panel a) (30). Thus, RII␤induced inhibition of growth or DNA synthesis (Fig. 5B, panel d) was reversed by cAMP (8-Br-cAMP, 100 M) (data not shown). Inefficient activation of PKA prevents entry into cycle of thyroid cells (14) and impairs DNA replication and mitosis in Xenopus oocytes (33) Activation of TSHR by TSH in RII-expressing cells reversed growth inhibition (Fig. 5B, panel b). Thymidine incorporation and cell cycle analysis of the cell lines expressing the various constructs, showed the same type of results (Fig. 5, B and C). Cells expressing RII␤ were mostly in G 0 with few cells in G 2 -M and S phase (2 and 3%, respectively), while TSHR-expressing cells displayed, in the presence of TSH, a longer G 1 phase with a reduction of G 2 -M and S phase from 4.5 to 3 and 7.1 to 5%, respectively ( Fig. 5C and Refs. 14 and 34). Taken together, these data indicate that excess of cAMPinduced G 1 arrest, whereas cAMP depletion by RII expression induced quiescence (14,33,35).
In Low Serum, TSHR Stimulates the Growth of Cells Expressing RII␤ Subunit-The experiments described above were carried out in the presence of 10% serum, which stimulates per se the growth of fibroblasts, because it contains several potent growth factors. To eliminate these effects, we analyzed a possible TSH growth response in low serum (0.5% serum). In these experiments, we have used the constitutively active mutant T632I receptor, which is TSH-independent, to reduce the num-ber of variables. Under these conditions, control cells did not proliferate (see Fig. 4). RII␤ expression reduced the number of the cells (Fig. 6A, panel a). The expression of T632I resulted in a slow growth response (up to 48 h) followed by the reduction of the number of cells (Fig. 6A, panel a). The expression of T632I receptor and RII␤ subunit resulted in a rapid (24 h) growth response followed by a slow increase up to 96 h. The analysis of thymidine incorporation curves indicates the following. (i) DNA synthesis was blocked in RII␤-expressing cells, but was present in T632I lines. (ii) T632I/RII-expressing cells showed a marked and significant DNA synthesis compared with the cell lines expressing separately two genes. The overall effect on growth appeared modest compared with the stimulation of DNA synthesis in T632I/RII cell. This discrepancy was caused by increased G 2 -M length in these cells (data non shown). To test if this reduction was caused by apoptosis, we stained the cells with TUNEL and performed a detailed FACS analysis. The data shown in Fig. 6B indicate that in low serum, the expression of T632I receptor or RII␤ induced apoptosis. Double labeling of cells with propidium iodide and TUNEL indicated that apoptosis induced by T632I receptor occurred in G 2 -M cells (data not shown). Conversely, apoptosis induced by RII␤ occurred early (6 -10 h following serum withdrawal) and was reversed by cAMP. Expression of RII and TSHR completely suppressed apoptosis (Fig. 6B, panels b and c). cAMP efficiently substituted TSHR (Fig. 6B, panels e and f).
It is noteworthy that TSHR and RII␤ efficiently cooperated in all conditions tested, i.e. in the presence or absence of serum. Growth arrest or apoptosis were strictly dependent on the TSHR-dependent Cell Growth Requires a Functional Rap1b-The data shown above indicate that TSHR-RII␤ expression in the absence of serum promote DNA synthesis and the growth of mouse fibroblasts. One downstream target acti-vated by cAMP is Rap1, a small GTP-binding protein that can be activated by GEFs, upon cAMP binding. We used a Rap1b dominant-negative mutant (RapN17) to test whether the inhibition of endogenous Rap1 may affect the TSHR-RII␤-stimulated growth in mouse fibroblasts. Infection of cells with the vector contain significantly higher levels of the protein. All samples were normalized for transfection and for ␤-actin expression (lower panel). The expression of the RII␤ subunit does not affect the cAMP basal levels: 1.81 Ϯ 0.13 pmol/2 ϫ 10 5 cells in NTC versus 1.68 Ϯ 0.19 in NTC/RII␤) B, co-expression of TSHR and PKA, regulatory subunit RII␤, stimulates the growth of mouse fibroblasts. The upper panel shows the growth curve of control fibroblasts (tTA-repressor-expressing clone) and cells expressing the TSHR-WT or T632I and regulatory subunit RII␤. The plots are the mean Ϯ S.E. of three independent experiments in triplicate samples. The lower panel shows the thymidine incorporation profile of the same cell lines indicated above. The cells were processed as described in Fig. 4. C, cell cycle progression of cells expressing TSHR and RII␤. The cell lines indicated were analyzed by FACS. 1 ϫ 10 6 cells were plated in 100-mm dishes; tetracyclin was removed 48 h before seeding, if required, and growth was monitored for 48 h in standard medium in the presence 10 mU/ml of TSH, as indicated. For FACS analysis, cells were harvested at 60% of confluence, ethanol-fixed, and stained with propidium iodide. virus carrying RapN17 efficiently and selectively inhibited TSHR-RII␤-dependent growth (Fig. 7). It is noteworthy that the growth inhibitory effect of cAMP in NIH3T3 cells has been shown to be Rap-dependent (11).
TSHR and RII␤ Activate AKT in a TSH-dependent Manner-To find out the molecular mechanism underlying TSHR/ RII␤ growth stimulatory effect in NIH3T3 fibroblasts, we analyzed the major growth signaling pathways: ERK1/2 and AKT. We determined the phosphorylation of AKT, the serine protein kinase downstream of PI3K, in the cell lines expressing T632I and RII␤ in the presence or absence of TSH. The ratio P-AKT to total AKT, a rather accurate index of AKT activation, was determined in these cell lines (Fig. 8). AKT phosphorylation was markedly inhibited in cells expressing the mutated receptor. 30 min treatment with TSH further inhibited AKT phosphorylation, indicating that the activity of the receptor was stimulated by TSH (see also Fig. 2). RII␤ expression had no effect on the AKT phosphorylation. The expression of the receptor and RII␤ in the presence of TSH significantly stimulated AKT phosphorylation (Fig. 8A). Note that the receptor expressed with RII␤ in the absence of TSH did not inhibit basal AKT phosphorylation (Fig. 8A). To obtain an independent evidence for the regulatory effects of RII␤, we transiently transfected 3T3 NIH fibroblasts with the wild-type receptor and RII␤, and stimulated the cells with TSH for 3 and 30 min. Total cell proteins, normalized for transfection efficiency, were blotted with phospho-ERK1/2 and phospho-AKT antibodies. TSH treatments were carried out in low serum, to rule out interference with other signaling pathways on PI3K-AKT activation. Under these conditions, expression of TSH and TSHR markedly inhibited P-AKT in a time-dependent fashion. Under the same conditions, TSH rapidly stimulated AKT phosphorylation in cells expressing RII␤ and TSHR with a peak at 3 min (Fig.  8B). In the same cell lines, ERK1/2 activation was induced by TSH at 30 min (Fig. 8B). TSH inhibited both ERK1/2 and AKT in cells expressing TSHR. Co-expression of RII␤ and TSHR resulted in a time or dose dependent fashion PI3K-AKT and ERK1/2 activation in response to TSH. AKT activation was essential, because LY294002, a specific PI3K inhibitor, prevented TSH-RII induced DNA synthesis (data not shown). These data indicate that the expression of RII␤ modulates cAMP effects on AKT and ERK1/2 initiated by TSHR.
TSHR and RII Efficiently Propagate cAMP Signals to the Nucleus-Cells expressing RII and TSHR apparently modulate TSH signaling to AKT and ERK1/2. Since PKAII has been associated with efficient transcription of cAMP-induced genes (24,30), we have determined cAMP signaling to the nucleus in the cell lines indicated above, by measuring accumulation of phosphorylated CREB in the nucleus and the transcription of a prototypic cAMP-dependent promoter CRE-CAT (36). The as- says were carried out at 48 h following serum withdrawal. Under these conditions, we have measured the steady-state cAMP-dependent transcriptional activity. Fig. 9 shows that only the cells lines expressing TSHR and RII␤ subunit were able to accumulate phosphorylated CREB and to efficiently transcribe CRE-CAT. The levels of total CREB were not modified by TSHR and RII␤ (Fig. 9A). Acute elevation of cAMP by forskolin (0.1 mM, 30 min) increased CREB phosphorylation in all cell lines, including control cells (data not shown and Refs. 14 and 26). These data indicate that cells expressing TSHR and RII␤ maintain steady elevated and efficient nuclear cAMP signals. DISCUSSION Our data show that the NIH3T3 fibroblasts acquire TSH-cAMP-dependent growth when expressing both TSHR and the PKA-RII␤ regulatory subunit. cAMP induced by TSH was not sufficient to sustain the proliferation of fibroblasts, as it does in thyroid cells. Rise in cAMP levels, as expected, inhibited growth. Growth inhibition by cAMP is a well described phenotype in many cell types (11). cAMP inhibits Raf1 in several cell types, including fibroblasts and thyroid cells (28,29,37) To stimulate cell growth, signaling proteins other than the TSH receptor are needed to convert cAMP signals into a proliferative stimulus. Our results indicate that the RII␤ subunit plays such a role by working as molecular sensor and responding to high cAMP levels generated by TSHR (14). In low serum, FIG. 8. Co-expression of TSHR and RII-PKA modulates AKT and ERK1/2 phosphorylation in a TSH-dependent manner. A, levels of P-AKT in T632I and/or RII␤-expressing NIH3T3 fibroblasts. Western blot analysis for AKT and the P-AKT in stable transfected cells after 48 h of growth in 0.5% serum; TSH was added at concentration of 10 mU/ml for the indicated time. The lower panel shows the P-AKT/AKT ratio after normalization for loading with ␤-actin expression and transfection efficiency. B, levels of P-ERK and P-AKT in WT-TSHR-and/or RII␤expressing NIH3T3 fibroblasts. A representative blot of transiently transfected cells after 48 h of growth in 0.5% serum; TSH was added at concentration of 10 mU/ml for the indicated time. The lower panel shows the P-AKT/AKT and the P-ERK/ERK ratio after normalization for loading with ␤-actin expression and transfection efficiency.
FIG. 9. Co-expression of TSHR and RII-PKA sustains CREB phosphorylation and cAMP-dependent transcription in low serum conditions. A, levels of P-CREB in T632I and/or RII␤-expressing NIH3T3 fibroblasts. Western blot analysis for the P-CREB in stable transfected cells after 48 h of growth in 0.5% serum. The lower panel shows total CREB levels. All samples were normalized for transfection efficiency and ␤-actin expression (data not shown). Acute (30 min) stimulation with forskolin (0.1 mM) induced P-CREB in all cell lines indicated in a comparable fashion. B, CAT activity after 72 h of growth in 0.5% serum. CAT activity of control, TSHR, RII, or TSHR-RII expressing cells, transiently transfected with CRE-CAT and pRSVLacZ, as described under "Experimental Procedures." Extracts, normalized for lacZ activity, were analyzed as described under "Experimental Procedures." The results shown are derived from at least three independent experiments. *, p Ͻ 00.1 (TSHR-RII versus control or TSHR or RII). activation of TSHR in cells expressing RII␤ subunit, generates high cAMP levels and active PKAII that efficiently signals to the nucleus and other cell compartments. These cells proliferate upon TSH receptor activation and cycle through DNA synthesis and mitosis without apparent apoptosis (Fig. 6, A and B). Note that control cells or cells expressing only TSH receptor underwent apoptosis under the same conditions (Fig. 6B, panels b and c). cAMP rise, in the absence of RII␤ subunit, induced DNA synthesis, followed by apoptosis (Fig. 6B, panels e and f). Moreover, cells expressing RII␤ subunit underwent apoptosis, when incubated in 0.5% serum. The simplest interpretation of these data is that cAMP represents an important anti-apoptotic signal in cells deprived of growth factors (22). Inefficient PKAII activation by excess of RII␤, reduces cAMP cytoprotection and amplifies apoptosis following growth factor withdrawal (38). Conversely, excess of cAMP generated by activated TSHR, is able to stimulate the first cycle(s), but it cannot sustain survival in the absence of other growth factors (Fig. 6A) probably because cAMP generated by TSHR in the absence of PKAII inhibits AKT (Fig. 8). Expression of RII and TSHR converts TSH in a powerful stimulator of AKT and ERK (Fig.  8). TSH regulated growth requires PI3 signaling (39). Also, TSH and cAMP stimulate selectively the formation of the complex PI3K-p21 Ras (37). We suggest that PKAII, formed by RII, is an efficient activator of PI3K, which stimulates AKT. The selective activity of PKAII on PI3K may be dependent on membrane anchoring of PKAII, which has been shown to phosphorylate selectively several types of receptors (40). Our results also cast a new light on the mechanisms underlying TSH-or cAMP-dependent cell growth. TSH receptor is a powerful activator of adenylyl cyclase, and the levels of cAMP induced by TSH in thyroid cells as well as in 3T3 cells expressing the exogenous TSHR, are significantly higher than those found in fibroblasts or other cells not expressing specific G s -coupled receptors (26,41). High cAMP levels last longer and are very efficiently propagated to the nucleus when PKAII is present (14). RII␤ subunit, which forms PKAII, has lower affinity for cAMP (as compared with PKAI). RII-containing PKA dissociate at higher cAMP levels (14). Although RII␤ has a lower affinity for cAMP compared with RII␣, under our conditions, RII␣ and RII␤ are equivalent (data not shown and Refs. 14 and 30).
In thyroid cells and in other cells containing a high PKAII/ PKAI holoenzyme ratio (3T3 fibroblasts expressing RII␤), the efficient expression of RII␤ subunit increases the stability and membrane anchoring of PKAII holoenzyme (14,24,42). Membrane-anchoring appears to be an important determinant of the intracellular cAMP gradient, because Ht31, a peptide that inhibits anchoring of PKAII to the membranes, suppressed cAMP gradients in isolated cardiomyocytes (43). Most likely membrane-anchored PKA allows propagation and maintenance of cAMP gradients in different organelles. For example, massive increases in cAMP will not reach the nucleus if the anchoring of PKA to the membranes is dismantled (14,43,44).
In conclusion our data identify RII␤ as an important and specific mediator of TSHR and cAMP signals to downstream membrane and nuclear substrates. Our data also suggest that cAMP-induced transcription is essential for inhibition of apoptosis and stimulation of growth, independent of the cell type analyzed.