INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

PHYSIOLOGICAL CONCENTRATIONS of thyroid hormone potentiate the antiviral state induced by homologous interferon (IFN)- (21, 22). Antiviral action of IFN- requires tyrosine phosphorylation of the signal transducer and activator of transcription STAT1 (7). Maximal antiviral activity of IFN- is obtained when Ser-727 of STAT1 is phosphorylated (17), perhaps by mitogen-activated protein kinase (MAPK) (8) or another serine kinase (45). The 44- and 42-kDa MAPK isoforms extracellular signal-regulated kinase 1 and 2 (ERK1 and ERK2, respectively) are ubiquitously expressed serine/threonine kinases that are activated by dual-specificity MAPK kinases (MEK1 and MEK2) in response to diverse agonists (28). A number of receptor tyrosine kinases, cytokine receptors, and heterotrimeric G proteins have been shown to activate MEK and MAPK (3, 12).

We have recently shown that thyroid hormone promotes tyrosine phosphorylation and nuclear uptake of STAT1 (20) and have speculated that the hormone may activate the MAPK pathway to obtain maximal activation of STAT1 and potentiation of the biological activity of IFN-. We report here that thyroid hormone [L-thyroxine (T₄) or 3,5,3'-triiodo-L-thyronine (T₃)] indeed activates the MAPK cascade in HeLa and CV-1 cells, in both the absence and presence of IFN-. Components of the mechanism by which thyroid hormone activates the MAPK pathway are also described. Because neither HeLa cells (33) nor CV-1 cells (23) contain functional nuclear thyroid hormone receptor (TR), the actions of thyroid hormone on kinase activities in these cell lines are not mediated by TR.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Materials. T₄, T₃, 3,3',5'-triiodothyronine (rT₃), 3,3',5,5'-tetraiodothyroacetic acid (tetrac), T₄-agarose, protein A-agarose, and guanosine 5'-O-(3-thiotriphosphate) (GTPS) were obtained from Sigma Chemical (St. Louis, MO), and recombinant human IFN- was from BioSource International (Camarillo, CA). Pertussis toxin was obtained from Calbiochem (San Diego, CA). CGP-41251 was a gift from Novartis Pharma (Basel, Switzerland), and Lipofectin was obtained from GIBCO BRL (Grand Island, NY). Genistein was obtained from ICN Biochemicals (Costa Mesa, CA), geldanamycin came from the Drug Synthesis and Chemistry Branch, National Cancer Institute (Bethesda, MD), PD-98059 was from Calbiochem (La Jolla, CA), and U-73122 and U-73343 were obtained from Dr. Robert Smallridge (Mayo Clinic, Jacksonville, FL). HeLa cells were obtained from the American Type Culture Collection (Manassas, VA), CV-1 cells were from Dr. Paul M. Yen (National Institutes of Health), and 293T cells were from Dr. Kevin Pumiglia (Albany Medical College, Albany, NY). U3A, STAT1_wt, and STAT1_A727 cells, derived from human fibroblasts (26, 42), were obtained with the permission of Dr. George Stark (Cleveland Clinic Foundation, Cleveland, OH) from the laboratories of Drs. James E. Darnell, Jr., (Rockefeller University, New York, NY) and Ke Shuai (University of California School of Medicine, Los Angeles, CA). [³²P]NAD was obtained from DuPont-NEN (Boston, MA).

Cell culture and preparation of nuclear fractions. Confluent HeLa and CV-1 cells grown in 100-mm culture dishes were treated with 0.25% hormone-depleted fetal bovine serum-containing medium (22) for 48 h. The U3A cell series were grown in the same serum-supplemented medium with 400 µg/ml G418 added. The total and free T₄ concentrations in this serum-supplemented medium were 2.3 × 10¹¹ M and 10¹⁴ M, respectively, and the free T₃ concentration was below detectable levels (23). Hormone, hormone analogs, IFN-, and/or inhibitors were then added for different time periods as indicated. Stock solutions (10⁴ M) of hormone and analogs were prepared in 0.04 N KOH-4% propylene glycol, and dilutions were made to final concentrations as indicated. In all experiments in which T₄ was added to cultures, the total and free T₄ concentrations were 10⁷ M and 10¹⁰ M, respectively, and total and free T₃ levels were below the limits of measurement. The hormone solvent had no effect on signal transduction studies. T₄-agarose was provided as a suspension in 0.5 M NaCl containing ~6 mM T₄ and was diluted in culture medium to a final T₄ concentration of 10⁷ M.

Immunoprecipitation and immunoblotting. After normalization of sample protein content, immunoprecipitation was performed using polyclonal anti-phosphotyrosine antibody (Transduction Laboratories, Lexington, KY). After overnight incubation at 4°C with rocking, protein A-agarose was added and samples were rocked for 1 h at 4°C. After two washes with hypotonic buffer containing 0.2% NP-40, immunoprecipitates were eluted with 2× sample buffer, and proteins were separated by discontinuous SDS-PAGE (7.5-9.0%). Proteins were transferred to Immobilon membranes (Millipore, Bedford, MA) by electroblotting. After being blocked with 5% milk in Tris-buffered saline containing 0.1% Tween, membranes were incubated with 1:1,000 monoclonal anti-MAPK antibody (ERK2) (Transduction Laboratories) or with 1:1,000 monoclonal anti-STAT1 antibody (Transduction Laboratories) overnight. For selected studies, 1:1,000 polyclonal anti-tyrosine/threonine-phosphorylated MAPK antibody (New England BioLabs, Beverly, MA) was used for immunoblots. In some experiments (see Fig. 5), nuclear extracts were immunoprecipitated with monoclonal anti-STAT1 or anti-MAPK antibody (Transduction Laboratories), and the immunoprecipitates were then separated by PAGE and immunoblotted with antibody to MAPK or STAT1, respectively. Polyclonal antibodies to Ser-727-phosphorylated STAT1 and to amino acids 73-93 of TR1 were generously provided by Dr. David Frank (Dana-Farber Cancer Institute, Boston, MA) and Dr. William Chin (Brigham and Women's Hospital, Boston, MA), respectively. The secondary antibodies were rabbit anti-mouse IgG or goat anti-rabbit IgG (1:1,000, DAKO, Carpenteria, CA). Immunoblots were visualized by enhanced chemiluminescence (ECL; Amersham Life Science, Arlington Heights, IL) and quantitated by digital imaging (BioImage, Millipore). Immunoblots shown are representative of two or more experiments.

Cell treatments. Cells were treated with thyroid hormone or analogs and/or IFN- in the concentrations indicated. Different concentrations of the protein kinase C (PKC) inhibitor CGP-41251 (5-100 nM), genistein, a protein tyrosine kinase (PTK) inhibitor (1-100 µg/ml), or pertussis toxin (20-1,000 ng/ml) were added to cultures for 70 min, and T₄ (10⁷ M) was added for the last 30 min. U-73122, a phospholipase C (PLC) inhibitor (1-10 µM), its inactive analog U-73343, or GTPS (10⁸ to 10⁵ M) was added for 60 min, and T₄ was added for the last 30 min. Geldanamycin (1-10 µM) or PD-98059 (30 µM) was applied to cells for 16 h, and 10⁷ M T₄ was added for the last 30 min. Cells were harvested, and nuclear proteins were prepared as described. DMSO (0.1%) was the solvent for all inhibitors and had no effect itself on immunoprecipitation, immunoblotting, or antiviral studies.

Oligonucleotide transfection. HeLa cells were treated, as described by Glennon et al. (15), with 2.5 µg/ml Lipofectin for 6 h, and sense or antisense oligonucleotides (Operon Technologies, Alameda, CA) were applied for 48 h in a concentration of 10 µM. Selected cells from each treatment group were then exposed to 10⁷ M T₄ for 40 min and subsequently harvested for preparation of nuclear extracts, immunoprecipitation, and immunoblotting as described above.

Antiviral studies. Cells were exposed to IFN- (1.0 IU/ml) in the presence or absence of T₄ (10⁷ M) for the last 24 h of MAPK antisense oligonucleotide transfection. In studies with MAPK pathway inhibitors, cells were treated with T₄, with or without geldanamycin or PD-98059, for 24 h; the medium was then replaced, and cells were treated with IFN- (1.0 IU/ml) for 24 h. STAT1_wt and STAT1_A727 cells were treated with IFN-, with or without T₄, for 24 h. After these various treatments, the cells were infected with vesicular stomatitis virus, and an antiviral plaque assay was performed as described previously (22), with results expressed in plaque-forming units (pfu) per milliliter. In our antiviral studies, a maximal effect of IFN- is seen at a cytokine concentration of 1,000 IU/ml, but T₄ potentiation of IFN- action is greatest at submaximal IFN- concentrations [1-10 IU/ml (22)]. The plaque assay method does not allow accurate measurement of virus yield below 10⁴ pfu/ml. One-way ANOVA was used to determine statistical significance.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Thyroxine and analogs cause tyrosine phosphorylation and nuclear translocation of MAPK. Initial studies were performed with T₄ alone, in the absence of IFN-. Confluent HeLa cells were treated with 0.25% thyroid hormone-depleted serum-supplemented medium for 2 days, and T₄ (10⁷ M total, 10¹⁰ M free) was then added to cultures for 10-60 min. This concentration of T₄ is physiological. Cells were harvested, and nuclear extracts were immunoprecipitated with anti-phosphotyrosine antibody. Immunoprecipitated proteins were eluted from protein A-agarose, separated by electrophoresis, and immunoblotted with anti-MAPK antibody. Increased nuclear content of tyrosine-phosphorylated MAPK was detected at 10 min after exposure to T₄ (Fig. 1A). This transient effect was maximal at 30-40 min and significantly reduced or absent at 60 min. In four experiments, the increase in band density with T₄ addition for 30 min, compared with a control sample without hormone, was 25 ± 8-fold (means ± SE). Nuclear tyrosine-phosphorylated MAPK was also detected by immunoblotting of nuclear samples with anti-tyrosine/threonine-phosphorylated MAPK antibody, as shown in Fig. 1B. In the study shown, accumulation of nuclear activated MAPK was maximal at 30 min, and there was a 34-fold increase in combined band intensities of ERK1 and ERK2.

View larger version (49K): [in this window] [in a new window]   Fig. 1.   Effect of L-thyroxine (T4) on tyrosine phosphorylation and nuclear translocation of mitogen-activated protein kinase (MAPK). A: nuclear extracts of HeLa cells treated with 107 M T4 for 10-60 min were immunoprecipitated with anti-phosphotyrosine antibody; immunoprecipitated proteins were eluted, separated by gel electrophoresis, and immunoblotted with anti-MAPK antibody. Nuclear accumulation of tyrosine-phosphorylated MAPK is seen as early as 10 min, with a maximal level seen at 30 min of T4 treatment. Band shown is 42-kDa extracellular signal-regulated kinase (ERK) 2. B: aliquots of similar nuclear fractions were immunoblotted with antibody to tyrosine-threonine-phosphorylated MAPK, without immunoprecipitation. Again, accumulation of phosphorylated MAPK is seen within 10 min of T4 treatment, reaches a maximum between 30 and 40 min, and is diminished at 60 min. Bands represent ERK1 (44 kDa) and ERK2 (42 kDa).

7

10

7

View larger version (38K): [in this window] [in a new window]   Fig. 2.   Effect of T4, 3,5,3'-triiodo-L-thyronine (T3), 3,3',5'-triiodothyronine (rT3), 3,3',5,5'-tetraiodothyroacetic acid (tetrac), and T4-agarose on activation of MAPK. A: HeLa cells were treated with T4 (107 M, a physiological concentration), or indicated concentrations of T3 or rT3 for 30 min. Nuclear samples were immunoprecipitated with anti-phosphotyrosine antibody, and precipitates were immunoblotted with anti-MAPK antibody. Neither T3 nor rT3 was as effective as T4 in activating MAPK, although each had some effect at supraphysiological concentrations (107 M). Effects that were seen with T3 and rT3 were maximal at 30 min, as shown. B: HeLa cells were treated with T4, with or without tetrac, for 30 min. T4 stimulated tyrosine phosphorylation and nuclear accumulation of MAPK (lane 2), whereas tetrac had minimal effect (lane 3). However, tetrac reduced T4 action in a dose-dependent manner (lanes 4 and 5, compared with lane 2). C: HeLa cells were treated with 107 M T4, T4-agarose containing 107 M T4 (T4-ag), T4-agarose clarified by 3 washes in PBS (T4-agw), or protein A-agarose (pA-ag) for 30 min, after which nuclear accumulation of activated MAPK was measured. Protein A-agarose (lane 2) had no effect on MAPK activation, whereas T4-agarose (lane 4) was as effective as T4 in activation of MAPK. T4-agarose washed 3 times and reconstituted in same T4 concentration of 107 M (lane 5) was as effective as T4. Lane 1, control (con).

Cell surface action of T₄. The T₄ analog tetrac blocks 1) T₄ potentiation of IFN--induced human leukocyte antigen-DR (HLA-DR) expression (20), 2) T₄ potentiation of IFN--induced antiviral activity (21, 23), and 3) T₄-induced activation of STAT1 (20). Tetrac also partially inhibited the effect of T₄ on tyrosine phosphorylation and nuclear translocation of MAPK, as shown in Fig. 2B (lanes 4 and 5, compared with lane 2). We previously showed that tetrac inhibits binding of T₄ to isolated human erythrocyte membranes (10) and therefore postulated that tetrac blocks T₄ activation of signal-transducing proteins by inhibition of T₄ interaction at the cell surface. To further study this possibility, we treated HeLa cells with T₄-agarose. In Fig. 2C, we show that T₄, T₄-agarose, and T₄-agarose clarified with three PBS washes all stimulated tyrosine phosphorylation and nuclear translocation of MAPK (lanes 3-5), whereas protein A-agarose had no effect on MAPK activation (lane 2). These findings further support the action of T₄ at the cell membrane as an enhancer of signal transduction. We have also conducted these studies in CV-1 cells, which like HeLa cells lack a traditional TR (23), and have obtained similar results (not shown).

Involvement of a G protein-coupled mechanism of hormone action. Della Rocca et al. (11) reported rapid activation of MAPK by -adrenergic agonists via pathways mediated by the G proteins G_i and G_q. To define the possible role of a G protein-coupled mechanism at the cell surface in the action of T₄, GTPS was added to HeLa cells for 70 min, and T₄ was added for the last 30 min of GTPS treatment. Nuclear fractions were prepared and immunoblotted with anti-phosphorylated MAPK antibody, and the results are shown in Fig. 3A. There is T₄-induced phosphorylation and nuclear accumulation of MAPK and dose-dependent reduction of the T₄ effect by GTPS. HeLa cells were also treated with pertussis toxin (20-1,000 ng/ml) for 60 min and with T₄ for 30 min; immunoblots with phosphorylated MAPK antibody showed a dose-dependent reduction of the T₄ effect by pertussis toxin (Fig. 3B). In the absence of T₄, pertussis toxin did not alter MAPK phosphorylation (results not shown). These results suggest a contribution of a pertussis-toxin-inhibitable GTP-binding protein (G_i or G_o) to the T₄ effect and raise the possibility that thyroid hormone binds to a G_i protein-coupled receptor (GPCR) at the cell membrane.

View larger version (43K): [in this window] [in a new window]   Fig. 3.   Effect of guanosine 5'-O-(3-thiotriphosphate) (GTPS) and pertussis toxin on T4-induced activation of MAPK. A: HeLa cells were treated with GTPS (108 to 105 M) for 70 min and T4 (107 M) for 30 min. Nuclear samples were immunoblotted with anti-tyrosine-threonine-phosphorylated MAPK. Activation of MAPK by T4 was reduced by GTPS in a dose-dependent manner. GTPS alone had no effect on MAPK activation. B: HeLa cells were treated with pertussis toxin (PT; 20-1,000 ng/ml) for 60 min and T4 (107 M) for 30 min. Nuclear samples were immunoblotted as in A. Again, there was a dose-responsive reduction in MAPK activation by T4 with addition of pertussis toxin.

Contributions of PKC, PTK, and PLC to the thyroxine effect. We previously reported that T₄ potentiation of IFN--induced HLA-DR expression and antiviral activity is dependent on activities of PKC and PTK (20, 23). Investigation of the effects of kinase inhibitors on T₄ activation of MAPK was therefore undertaken. Genistein (1-100 µg/ml), an inhibitor of PTK activity (30), blocked the action of T₄ on MAPK phosphorylation and nuclear translocation (Fig. 4A). CGP-41251, an inhibitor of calcium-dependent PKC activity [inhibits PKC, -_I, -_II, and - with IC₅₀ values of 24, 17, 32, and 18 nM, respectively (25)], also blocked T₄-induced tyrosine phosphorylation of MAPK in a concentration-dependent manner (Fig. 4B). In Fig. 4C, genistein and CGP-41251 are again shown to inhibit the effect of T₄ on MAPK phosphorylation and nuclear translocation in nuclear samples immunoblotted with antibody to tyrosine/threonine-phosphorylated MAPK. The two bands representing ERK1 (44 kDa) and ERK2 (42 kDa) are similarly affected by T₄ and the inhibitors. The effect of the PLC inhibitor U-73122 (35) and an inactive analog, U-73343, on T₄ activation of MAPK was also examined. Figure 4D shows activation of MAPK by T₄ in HeLa cells and inhibition of this MAPK activation by U-73122 but not by U-73343. Similar results were obtained with 293T cells, which contain TR, as indicated by immunoblotting with antibody to TR (results not shown).

View larger version (36K): [in this window] [in a new window]   Fig. 4.   Role of protein tyrosine kinase, protein kinase C, and phospholipase C (PLC) in activation of MAPK by T4. Detection of phosphorylated MAPK was by anti-phosphotyrosine immunoprecipitation and anti-MAPK immunoblotting in A and B and by anti-phosphorylated MAPK immunoblotting in C and D. In all studies, inhibitor was added for 60 min and T4 (107 M) for last 30 min. A: with addition to HeLa cells of T4 and genistein (Gen; 1-100 µg/ml), there was dose-dependent inhibition of T4 effect on tyrosine phosphorylation and nuclear translocation of MAPK. B: CGP-41251 (CGP; 5-100 nM) also caused dose-dependent inhibition of T4-induced MAPK activation in HeLa cells. C: appearance of two activated MAPK bands (42-kDa ERK2 and 44-kDa ERK1) in nuclear fractions of HeLa cells treated with T4 was also demonstrated with use of antibody to phosphorylated MAPK, without prior immunoprecipitation. Inhibition of T4 effect is again seen with both genistein and CGP-41251, similar to findings in A and B. D: in HeLa cells, U-73122 (1 and 10 µM) but not inactive analog U-73343 reduced or completely inhibited T4-induced nuclear accumulation of phosphorylated MAPK, as shown by immunoblotting with anti-phosphorylated MAPK antibody.

Coimmunoprecipitation of MAPK and STAT1. David et al. (8) described coimmunoprecipitation of MAPK and STAT1 in extracts of cells treated with IFN-. We therefore tested the possibility that a direct interaction between MAPK and STAT1 could be detected in cells treated with T₄, alone or with IFN-. Nuclear extracts were immunoprecipitated with anti-MAPK antibody, and the solubilized immunoprecipitates were subjected to electrophoresis and immunoblotted with anti-STAT1. Nuclear STAT1 appeared in MAPK immunoprecipitates in cells treated with T₄ alone (10⁷ M), as shown in Fig. 5A. In the study shown in Fig. 5B, cells were treated with T₄ with or without IFN- (1-100 IU/ml) for 30 min. The antibody order was reversed, and samples of nuclear extracts were immunoprecipitated with anti-STAT1 antibody and the precipitates were immunoblotted with anti-MAPK antibody. Both T₄ and IFN- separately caused nuclear complexing of STAT1 and MAPK, as reported previously with IFN- (8). At each concentration of IFN-, T₄ enhanced the cytokine effect. Additional studies using an antibody to Ser-727-phosphorylated STAT1 (13) demonstrated increased Ser-727 phosphorylation of STAT1 by T₄ or IFN- and potentiation of the cytokine effect by thyroid hormone (Fig. 5C). T₄ and IFN- (1.0 IU/ml), when applied separately to cells, caused detectable Ser-727 phosphorylation. With addition of hormone together with a low concentration of cytokine (1 IU/ml), there was enhancement of IFN--induced Ser-727 phosphorylation by T₄. With a higher IFN- concentration, there was no further enhancement of the IFN- effect by T₄.

View larger version (44K): [in this window] [in a new window]   Fig. 5.   Induction by T4, with or without interferon- (IFN-), of MAPK and STAT1 nuclear complexes. A: HeLa cells were treated with T4 (107 M) for times indicated. Nuclear extracts were immunoprecipitated with anti-MAPK antibody, and precipitated proteins were separated by PAGE and immunoblotted with anti-STAT1 antibody. Shown is accumulation in up to 30 min of nuclear complexed MAPK and STAT1, with subsequent loss of complex by 120 min. B: HeLa cells were treated with IFN- (1-100 IU/ml) for 30 min in presence or absence of T4 (107 M). Nuclear fractions were immunoprecipitated with anti-STAT1 antibody, and resulting proteins were eluted, separated by PAGE, and immunoblotted with anti-MAPK antibody. Nuclear MAPK coimmunoprecipitated with STAT1 was increased in amount in an IFN- dose-dependent manner (lanes 3, 5, and 7). With each IFN- dose, addition of T4 further enhanced MAPK accumulation and STAT1-MAPK complex formation (lanes 4, 6, and 8). C: HeLa cells were treated with IFN- (1 or 100 IU/ml) and/or T4 (107 M) for 30 min. Nuclear fractions were immunoblotted with antibody to phosphoserine-727-STAT1. Both T4 and a low concentration of IFN- caused some serine phosphorylation of STAT1, and effect of IFN- (1 IU/ml) was enhanced by cotreatment with T4. No T4 enhancement was seen with a higher concentration of IFN-.

Effect of MAPK pathway inhibition. Studies were undertaken to characterize more proximal steps in the MAPK cascade that might contribute to the T₄ effect on activation of MAPK and STAT1 and potentiation of IFN- action. Geldanamycin, a MAPK pathway inhibitor that incompletely depletes cellular content of Raf-1 (32, 36), partially inhibited T₄-stimulated nuclear uptake of tyrosine-phosphorylated MAPK (Fig. 6A, top, lane 3) and tyrosine-phosphorylated STAT1 (Fig. 6A, bottom, lane 3). Similar findings were obtained with an inhibitor of MEK, PD-98059 (27), and are shown in Fig. 6A (lanes 4-6); there was reduction or complete inhibition of T₄-stimulated activation of MAPK and STAT1 by PD-98059 (lane 6, top and bottom).

View larger version (22K): [in this window] [in a new window]   Fig. 6.   Effect of MAPK pathway inhibition on activation and nuclear translocation of MAPK and STAT1 and on T4 potentiation of IFN--induced antiviral activity. A: HeLa cells were treated with T4 (107 M) for 30 min after pretreatment with MAPK pathway inhibitors geldanamycin (Gel) or PD-98059 (PD) for 16 h. Anti-phosphotyrosine immunoprecipitates from nuclear samples were immunoblotted with antibody to MAPK or STAT1. Lanes 1 and 2 and lanes 4 and 5 show T4 activation of MAPK and STAT1 (top and bottom, respectively); inhibition of T4 effect is seen with both geldanamycin (10 µM) and PD-98059 (30 µM) (lanes 3 and 6, respectively). B: antiviral studies were conducted on HeLa cells treated with geldanamycin (1 or 10 µM) or PD-98059 (3 or 30 µM) for 12 h. Medium was then removed, and fresh medium containing IFN- (1 IU/ml) with or without T4 (107 M) was added for 24 h. An antiviral assay was then conducted, with virus yield shown in plaque-forming units (pfu)/ml. Results show means ± SE of 3 studies with each inhibitor. Control virus yield is shown in open bars. IFN- antiviral effect is shown by solid bars, indicating a reduction in virus yield, and significant (P < 0.05 by 1-way ANOVA) potentiation of IFN- effect by T4 is shown by vertical arrows above hatched bars. Both inhibitors blocked T4 potentiation in a dose-dependent manner, but neither had an effect on antiviral action of this IFN- concentration in absence of thyroid hormone, as shown in cross-ruled bars.

7

Effect of MAPK antisense oligonucleotide transfection. To further examine the role of the MAPK pathway in T₄ potentiation of IFN- action, we reduced the HeLa cell content of MAPK by antisense oligonucleotide transfection (15). Absence of activated MAPK in nuclear fractions of cells treated with the antisense oligonucleotide is shown in Fig. 7A (top, lane 7), and there is little tyrosine phosphorylation or nuclear translocation of MAPK in antisense-treated cells exposed to T₄ (top, lane 8), particularly compared with levels in cells treated with Lipofectin alone (lane 4). In the same MAPK-depleted cells, there was a reduction in T₄-stimulated nuclear translocation of tyrosine-phosphorylated STAT1, compared with findings in HeLa cells not exposed to the MAPK antisense oligonucleotide (Fig. 7A, bottom, lane 8 compared with lanes 2 and 4). This finding suggests that the presence of stimulable MAPK is a requirement for optimal activation and nuclear translocation of STAT1. In cells treated with sense oligonucleotide, there was more activated MAPK in nuclei of cells not exposed to T₄ than in cells exposed to the hormone (Fig. 7A, top, lanes 5 and 6). The basal increase in MAPK and reduced activity of MAPK in response to an inducer in these cells have been described by other investigators in a different cell line (29). To test the effect of cellular MAPK depletion on T₄ potentiation of IFN- action, antiviral studies were also conducted on cells transfected with MAPK antisense or sense oligonucleotide or with Lipofectin alone. Significant T₄ potentiation of the antiviral action of IFN- (P < 0.05), shown by the vertical arrows in Fig. 7B, was demonstrated in untreated (control) cells and cells treated with sense oligonucleotide or Lipofectin alone but was diminished in MAPK-depleted cells. In the latter cells, however, the antiviral response to IFN- (1.0 IU/ml) in the absence of T₄ was not diminished. Taken together, the results of experiments with MAPK pathway inhibition and cellular depletion of MAPK provide evidence that T₄-induced activation and nuclear translocation of MAPK and STAT1, and potentiation of the action of IFN-, require activity of Raf-1 and MEK, as well as MAPK. These antiviral experiments with MAPK depletion and MAPK pathway inhibitors were conducted with submaximal concentrations of IFN-.

View larger version (39K): [in this window] [in a new window]   Fig. 7.   Effect of MAPK depletion on activation and nuclear translocation of MAPK and STAT1 and on T4 enhancement of IFN--induced antiviral activity. A: HeLa cells were depleted of MAPK as described in MATERIALS AND METHODS and then exposed to T4 or solvent for 30 min. Nuclear extracts of treated cells were immunoprecipitated with anti-phosphotyrosine antibody, and precipitated proteins were immunoblotted with anti-MAPK or anti-STAT1 antibody. In cells transfected with MAPK antisense oligonucleotide, there was no nuclear tyrosine-phosphorylated MAPK in cells not treated with T4 (lane 7), and a small amount of nuclear activated MAPK in cells that were treated with T4 (107 M; lane 8). Marked T4 activation of MAPK was seen in cells treated with Lipofectin alone (compare lanes 3 and 4). In cells treated with sense oligonucleotide, there was enhancement of nuclear phosphorylated MAPK in control cells (lane 5) and no potentiation in T4-treated cells (lane 6). STAT1 in nuclear fractions was minimally increased by T4 in cells treated with antisense oligonucleotide (bottom, lanes 7 and 8) but was markedly increased in control cells (lanes 1 and 2) and in cells treated with Lipofectin alone (lanes 3 and 4). B: antiviral studies were conducted after exposure of cells to oligonucleotides for 24 h. Cells were treated with IFN- (1.0 IU/ml) with or without T4 (107 M) for 24 h, followed by antiviral assay. Significant T4 potentiation of IFN- action is highlighted by vertical arrows. Data are means ± SE of 3 experiments. In HeLa cells treated with MAPK antisense oligonucleotide, there was loss of T4 potentiation of IFN- action, whereas hormone effect remained intact in cells treated with sense oligonucleotide or with Lipofectin alone. There was, however, no loss of IFN- antiviral effect in cells exposed to oligonucleotide transfection or Lipofectin.

T₄ potentiation of IFN- action requires Ser-727 on STAT1. Maximal antiviral activity of IFN- requires the presence of a serine at position 727 of STAT1 (17). It has been suggested that MAPK may catalyze the Ser-727 phosphorylation of STAT1 (8, 42), although more recent studies suggest that the Ras-MAPK pathway is not involved, at least in IFN--induced serine phosphorylation (45). Because we have demonstrated T₄-induced nuclear accumulation of Ser-727-phosphorylated STAT1 (Fig. 5C), we studied whether T₄ potentiation of IFN- action would be seen in cells lacking serine at position 727. U3A cells lack STAT1 (26, 42) and on exposure to T₄ showed no nuclear accumulation of that protein (Fig. 8A, top). U3A cells with reconstituted wild-type STAT1 (STAT1_wt) showed stimulation of STAT1 tyrosine phosphorylation and nuclear translocation by T₄, whereas U3A cells containing STAT1 with an alanine-for-serine substitution at position 727 (STAT1_A727) showed diminished activation of that protein (Fig. 8A, top). In the same samples, nuclear accumulation of tyrosine-phosphorylated MAPK appeared to be more intense in U3A cells and STAT1_A727 cells than in the STAT1_wt cells (Fig. 8A, bottom), demonstrating that activation of MAPK does not require the presence of STAT1, whereas activation of STAT1 does seem to require the presence of functional MAPK (see Figs. 6 and 7).

View larger version (37K): [in this window] [in a new window]   Fig. 8.   Effect of T4 on activation of signal transduction and on IFN--induced antiviral activity in cells lacking STAT1 or lacking Ser-727 of STAT1. A: U3A cells lacking STAT1, wild-type STAT1 (STAT1wt) cells, and STAT1-deficient cells transfected with STAT1 containing an alanine-for-serine substitution at residue 727 (STAT1A727 cells) were treated with 107 M T4 for 30 min, after which nuclear extracts were immunoprecipitated with anti-phosphotyrosine antibody and precipitates were immunoblotted with anti-STAT1 or anti-MAPK antibody. In U3A cells, no STAT1 is seen, but activated MAPK readily accumulates in T4-treated nuclear fractions. In STAT1wt cells, nuclear accumulation of activated STAT1 and MAPK is seen. In STAT1A727 cells, there is reduced tyrosine-phosphorylated STAT1 in nuclear fraction, but MAPK accumulation is present. B: antiviral studies in HeLa and STAT1wt cells show a similar dose-dependent antiviral response to IFN-, with progressively less T4 potentiation seen as a maximal concentration of IFN- is reached. In STAT1A727 cells, antiviral effect at each concentration of IFN- is reduced compared with effect in other cells, and there is no T4 potentiation of cytokine action at any IFN- concentration.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The present observations support a novel role for thyroid hormone as a modulator of signal transduction. We used a cytokine, IFN-, to demonstrate the biological relevance of thyroid hormone action on two signal transduction pathways. We propose that the enhancement by thyroid hormone of IFN--induced antiviral activity depends on effects of the hormone on the MAPK cascade and, as a result, on the STAT1 pathway that is activated by the cytokine. The consequence of the action of thyroid hormone on MAPK phosphorylation and nuclear translocation is viewed by us to be Ser-727 phosphorylation of the STAT1 dimer that results in increased binding of the STAT complex to one or more IFN--responsive genes. Activated STAT1 is a critical phosphoprotein in the transduction of IFN- and IFN-/ signals (7). Activation of STAT1 by IFN- requires Tyr-701 phosphorylation by Janus kinases (JAK1 and JAK2); maximal transcriptional activity of STAT1 requires, in addition, phosphorylation of Ser-727 (17, 42). MAPK has been implicated in IFN-/ signal transduction, specifically in the Ser-727 phosphorylation of STAT1 (8) and also in Ser-727 phosphorylation of another member of the STAT family, STAT3 (5).

The ability of the two MAPK pathway inhibitors, geldanamycin and PD-98059, to reduce activation of MAPK and STAT1 and to block thyroid hormone potentiation of the antiviral action of IFN- further supports our hypothesis that the mechanism of T₄ in this potentiation involves the MAPK pathway, principally at the levels of MEK and MAPK. Geldanamycin incompletely depletes cellular content of Raf-1 in MCF7 cells (32) and HeLa cells (36), whereas PD-98059 is regarded as a specific inhibitor of MEK activity at the concentrations we used (4, 27). Tyrosine phosphorylation of STAT1 is an integral part of the IFN--stimulated JAK-STAT pathway. In the absence of IFN-, however, the means by which T₄ caused tyrosine phosphorylation of STAT1 required clarification. The results obtained with PD-98059 presented in Fig. 6 suggest that, under the direction of thyroid hormone, MEK tyrosine phosphorylates STAT1, in addition to activating its traditional substrate, MAPK. MEK is a dual-specificity tyrosine-threonine kinase. Phosphorylation of threonines at positions 699 and 704 of STAT1 (44) may, as described by Cobb and Goldsmith (6), provide an environment suitable for phosphorylation of Tyr-701 by MEK. Our results do not exclude the possibility that an unidentified MEK-dependent tyrosine kinase is responsible for T₄-directed tyrosine phosphorylation of STAT1. Depletion of cellular MAPK by antisense oligonucleotide transfection also confirmed that MAPK is required for T₄-induced activation of STAT1 and for T₄ potentiation of the antiviral activity of IFN-.

Fukuda et al. (14) recently reported that the transport of MAPK from cytoplasm to nucleus requires dissociation of the MAPK-MEK complex. It is possible that thyroid hormone may also act at this putative complex to enhance nuclear transfer of MAPK. It should be noted that tyrosine phosphorylation (activation) of MAPK in the present studies was documented by both anti-phosphotyrosine antibody immunoprecipitation with subsequent probing of the immunoprecipitate with anti-MAPK antibody and by the use of antibody to phosphorylated MAPK, which indicated involvement of both ERK1 and ERK2 in the hormone effect. Gorenne et al. (16) recently demonstrated correlation of two measurements of MAPK activation, 1) immunoblotting with antibody to tyrosine/threonine-phosphorylated MAPK and 2) direct measurement of MAPK activity by phosphorylation of myelin basic protein.

The initial step in the mechanism by which thyroid hormone acts on signal transduction is incompletely understood. Several features are clear, however. Because agarose-T₄ is as effective as T₄ in activating MAPK and STAT1, hormone action must begin at the cell surface. Furthermore, the hormone effect is inhibited by tetrac, an iodothyronine analog that, itself, is not a kinase activator but does inhibit the binding of T₄ to human erythrocyte membranes (10). Consistent with this evidence that hormone action requires a putative receptor on the plasma membrane are the observations presented here that T₄ action on signal transduction is pertussis toxin and GTPS sensitive. Thus the initial step in the mechanism is interaction of thyroid hormone with a GPCR. We previously described a G_i protein in erythrocyte membranes that may be involved in hormone action (9, 38), but a GPCR responsive to thyroid hormone has not been previously reported in nucleated cells (39). By nondenaturing PAGE of octylglucoside-solubilized plasma membranes, we recently identified two proteins that bind radiolabeled T₄ and are candidate GPCRs (M. R. Deziel, F. B. Davis, and P. J. Davis, unpublished observations).

PKC has been shown to activate Raf-1 directly (18), a step that is early in the MAPK cascade (Raf-1/MEK/MAPK). The fact that CGP-41251, an inhibitor of traditional PKC isoform activities (PKC, -_I, -_II, and -), prevented the action of thyroid hormone in our model is consistent with an effect of T₄ at the level of Raf-1, although our findings also raise the possibility that activated PKC can mediate the action of T₄ through the phosphorylation of MEK, as described by others (34), thus bypassing Raf-1. With the use of CGP-41251, we have effectively ruled out a contribution of calcium-independent, diacylglycerol-dependent novel PKCs and atypical PKCs, as described by Schönwasser et al. (31), some of which phosphorylate Raf-1 without stimulating its kinase activity. We previously reported that physiological concentrations of T₄ stimulate erythrocyte cytosol PKC activity (19), also supporting a role for PKC stimulation in the T₄ effects we describe in this report. We have elsewhere demonstrated a role for PKC in thyroid hormone potentiation of both antiviral (23) and immunomodulatory (20) actions of IFN-. A role for PLC in iodothyronine action on the MAPK cascade is suggested by results of our studies with the aminosteroid U-73122. That is, hormone activation of PKC (and subsequently of the MAPK pathway) depends on diacylglycerol liberated by PLC.

The proposed sequence of events in the signal transduction pathways acted on by thyroid hormone is shown in Fig. 9. Hormone binding at the cell membrane is followed serially by activation of PLC, PKC, Raf-1, MEK, and MAPK. The possibility that Ras may be involved in the sequence is included in Fig. 9. Although pertussis toxin-sensitive GPCRs have not previously been described to activate Ras through PLC and PKC (39, 40), we have found that thyroid hormone poorly activates MAPK in a dominant negative Ras model (J. Gordinier, F. B. Davis, and P. J. Davis, unpublished observations). Thus Raf-1 may be phosphorylated by PKC in the thyroid hormone-activated cell or PKC may act on Ras (24), thus activating Raf-1.

View larger version (33K): [in this window] [in a new window]   Fig. 9.   Proposed mechanism by which thyroid hormone nongenomically activates MAPK signal transduction cascade. Initial step is postulated to be interaction of hormone with a heterotrimeric Gi protein-coupled receptor (GPCR). The -subunit of GPCR is depicted as a regulator of PLC activity. Evidence for participation of each of upstream steps leading to MAPK activation is reviewed in DISCUSSION. JAK1 and JAK2, Janus kinases 1 and 2; Ras-GTP, complex of activated Ras and GTP; GAS, IFN--activated sequences of IFN-responsive genes; MEK, MAPK kinase.

Apparent from the present studies are not only the promotion by T₄ of the nuclear uptake of activated STAT1 and MAPK but also the recovery of MAPK from nuclear immunoprecipitates made with anti-STAT1 antibody and of STAT1 from anti-MAPK immunoprecipitates. Thus the nuclear fractions in T₄-treated cells contain a STAT1-MAPK complex. Association of MAPK and STAT1 has been described by others in response to IFN-/ (8). We have not detected this complex in cytosol of hormone-treated cells. Ostensibly, the complex reflects the action of MAPK in phosphorylating Ser-727 of STAT1, as suggested by David et al. (8). Documentation of thyroid hormone-potentiated phosphorylation of Ser-727 of STAT1 in the presence of IFN- is shown in Fig. 5C. Thus the complexing of STAT1 and MAPK under the influence of T₄ is not a casual event but is associated with phosphorylation of a specific residue of STAT1. Our studies with the U3A cell series revealed that, in the absence of STAT1, activated MAPK did translocate to the nucleus in the presence of T₄, whereas T₄-induced nuclear translocation of STAT1 was suppressed in the absence of MAPK in antisense oligonucleotide-treated cells.

In the absence of IFN-, thyroid hormone does not induce the antiviral state (22), so that activation of the MAPK pathway by the hormone is insufficient to initiate transcription of antiviral proteins. This is not surprising, since it is the STAT1 pathway that is primarily involved in transducing the IFN signal for the antiviral state and MAPK activity has been shown by others to be facilitative (17, 42). We postulate that there is a biological role for thyroid hormone potentiation in the presence of low levels of IFN-; in fact, we have found that T₄ promotes the antiviral action of IFN- at 0.1 IU/ml, a concentration that in itself is not antiviral (22). Thus, through enhanced activation of MAPK and STAT1, the hormone is able to convert an ineffective level of IFN- into an effective concentration.

It is possible that one or more actions of thyroid hormone on kinase cascades are representative of a newly recognized mechanism of hormone signaling, one that 1,25-dihydroxyvitamin D₃ also utilizes; the vitamin has been reported by others to stimulate activity and translocation of protein kinases via a nongenomic mechanism in acute promyelocytic NB4 cells (2). Other hormones that activate kinases in the MAPK cascade include gonadotropin-releasing hormone (37), norepinephrine (43), and 17-estradiol (41). Inhibition of MEK has been shown to decrease growth hormone-stimulated activation of STAT5 (27) and to decrease insulin stimulation of Ser-727 phosphorylation of STAT3 (4). Thus hormones that have important actions via nuclear receptors, as well as those that act primarily at the cell membrane, can nongenomically alter signal-transducing kinase activities.