Opposing Effects of EGF Receptor Signaling on Proliferation and Differentiation Initiated by EGF or TSH/EGF Receptor Transactivation

Abstract

Regulation of thyroid cells by thyrotropin (TSH) and epidermal growth factor (EGF) has been known but different effects of these regulators on proliferation and differentiation have been reported. We studied these responses in primary cultures of human thyroid cells to determine whether TSH receptor (TSHR) signaling may involve EGF receptor (EGFR) transactivation. We confirm that EGF stimulates proliferation and de-differentiation whereas TSH causes differentiation in the absence of other growth factors. We show that TSH/TSHR transactivates EGFR and characterize it as follows: (1) TSH-induced upregulation of thyroid-specific genes is inhibited by 2 inhibitors of EGFR kinase activity, AG1478 and erlotinib; (2) the mechanism of transactivation is independent of an extracellular EGFR ligand by showing that 2 antibodies, cetuximab and panitumumab, that completely inhibited binding of EGFR ligands to EGFR had no effect on transactivation, and by demonstrating that no EGF was detected in media conditioned by thyrocytes incubated with TSH; (3) TSH/TSHR transactivation of EGFR is different than EGFR activation by EGF by showing that EGF led to rapid phosphorylation of EGFR whereas transactivation occurred in the absence of receptor phosphorylation; (4) EGF caused downregulation of EGFR whereas transactivation had no effect on EGFR level; (5) EGF and TSH stimulation converged on the protein kinase B (AKT) pathway, because TSH, like EGF, stimulated phosphorylation of AKT that was inhibited by EGFR inhibitors; and (6) TSH-induced upregulation of thyroid genes was inhibited by the AKT inhibitor MK2206. Thus, TSH/TSHR causes EGFR transactivation that is independent of extracellular EGFR ligand and in part mediates TSH regulation of thyroid hormone biosynthetic genes.

Regulation of thyroid cells in vitro by thyrotropin (TSH) and epidermal growth factor (EGF) has been known for decades (1). In thyroid cells from several species, TSH has been found to induce thyroid cell differentiation and inhibit proliferation, particularly when incubated in the absence of serum or growth factors (2). Differentiation is usually assessed as an increase in the expression of genes that are involved in thyroid hormone biosynthesis. However, when incubated in the presence of serum or growth factors, TSH usually was found to stimulate proliferation (3). By contrast, EGF has been found to consistently stimulate thyroid cell proliferation and to cause dedifferentiation of the adult thyrocytes, usually measured as decrease in the levels of expression of thyroid hormone synthetic genes. Moreover, in canine thyroid cells, it was shown that when EGF and TSH were added simultaneously, EGF antagonized the differentiating effect of TSH (4). This was of interest because many G protein–coupled receptors (GPCRs), like TSH receptor (TSHR), and receptor tyrosine kinases (RTKs), like EGF receptor (EGFR), have been reported to act in concert to regulate biological responses (5‐7). In many of these cell systems, activation of a GPCR was found to activate a signaling pathway(s) typically used by RTKs. This phenomenon was termed RTK transactivation. In a number of cell systems, GPCR activation led to EGFR transactivation (7). Two distinct pathways have been identified for GPCR-mediated EGFR transactivation: an EGFR ligand–independent pathway and an EGFR ligand–dependent pathway. EGFR ligand–independent GPCR-mediated EGFR transactivation occurs independently of GPCR-stimulated generation of EGFR ligands. The ligand-independent transactivation process, especially in cancer cells, is poorly understood (6). The EGFR ligand–dependent pathway involves GPCR activation that leads to activation of membrane-bound matrix metalloproteases–mediated proteolytic cleavage of EGFR ligand precursors. The active ligands are then released into the extracellular space where they can bind to and activate EGFRs on cell surfaces.

Although the majority of studies of thyroid cell biology in vitro have been performed in non-human cells, several studies comparing TSH and EGF actions have been reported using human cells yielding conflicting results. Saunier et al (8) found that TSH activated the mitogen-activated protein kinase pathway in human thyroid follicles. In contrast, Buch et al (9) reported that TSH did not activate this pathway in primary monolayer cultures of human thyroid cells. Perhaps most interestingly, Kraiem et al (10) reported mutually antagonistic interactions between the TSH- and EGF-mediated pathways in cell proliferation and differentiation of cultured human thyroid follicles. We were intrigued by these inconsistent sets of data and decided to study these pathways in primary cultures of human thyroid cells.

We found that TSH and EGF caused opposing effects when added to human thyroid cells in primary culture under the conditions of our experiments. However, activation of TSHR by TSH caused transactivation of EGFR that was part of the signaling pathway leading to enhanced expression of thyroid hormone biosynthetic genes including thyroglobulin (TG), thyroid peroxidase (TPO), deiodinase type 2 (DIO2), sodium/iodide symporter (NIS, SLC5A5), and TSHR.

Material and Methods

Material

Dulbecco’s modified Eagle’s medium (DMEM), penicillin, and streptomycin were purchased from Mediatech Inc (Manassas, VA). Fetal bovine serum (FBS) HyClone was from Cytiva (Marlborough, MA), and bovine serum albumin (BSA) was from MP Biomedicals (Irvine, CA). PBS pH 7.4 was obtained from Quality Biological Inc (Gaithersburg, MD). Trypsin was from Thermo Fisher Scientific (Waltham, MA). RNeasy Mini Kits were from Qiagen (Hilden, Germany), High-Capacity cDNA Archive Kits were from Applied Biosystems (Foster City, CA, USA). Universal Probe Supermix was obtained from Bio-Rad Laboratories (Hercules, CA, USA), primers and probes for human thyroid markers and β-actin were obtained from Taqman, Assay-on-Demand (Applied Biosystems). Human Phospho-EGFR DuoSet IC enzyme-linked immunosorbent assay (ELISA) (Cat # DYC1095B, RRID:AB_2922827), Human Total EGFR DuoSet IC ELISA (Cat # DYC1854, RRID:AB_2922828), EGF Quantikine ELISA (Cat # DEG00, RRID:AB_2922829), Human/Mouse/Rat Phospho-AKT (protein kinase B) (S473), and Pan Specific DuoSet IC ELISA (Cat # DYC887B, RRID:AB_292282) were purchased from R&D Systems (Minneapolis, MN). Bovine TSH was purchased from Sigma-Aldrich (St. Louis, MO) and human TSH Thyrogen® was obtained from Sanofi (Paris, France). EGF was from PeproTech, Inc. (Rocky Hill, NJ). AG1478 and erlotinib were purchased from Tocris Bioscience (Bristol, UK). Cetuximab was from Selleck Chemicals (Houston, TX), and Vectibix® (panitumumab) was from Amgen (Thousand Oaks, CA). NuPAGE precast Bis-Tris gels, and NuPAGE MOPS sodium dodecyl sulfate (SDS) running buffer, Halt^™ protease and phosphatase inhibitor cocktail (100×), RIPA lysis buffer and nitrocellulose membrane filter paper sandwiches were obtained from ThermoFisher Scientific (Waltham, MA). Odyssey blocking buffer and IRDye secondary antibodies were purchased from Li-COR Biotechnology (Lincoln, NE). Anti-TG rabbit monoclonal (Cat # ab156008, RRID:AB_2922822) and anti-β-tubulin mouse monoclonal (Cat # ab231082, RRID:AB_2922824) antibodies were obtained from Abcam (Cambridge, MA). Anti-TPO rabbit polyclonal antibodies (Cat # PA5-81070, RRID:AB_2788319) were purchased from ThermoFisher Scientific (Waltham, MA).

Primary Cultures of Human Thyrocytes

Primary cultures of human thyrocytes were established by isolating cells from normal thyroid tissue samples from patients undergoing surgery for thyroid tumors at the National Institutes of Health Clinical Center as described previously (11). The studies involving human samples were reviewed and approved by the NIDDK Institutional Review Board. Written informed consent was obtained from the patients whose sample were used in the study. Thyrocytes were used only between passages 3 and 5. Cells were cultured in DMEM supplemented with 10% FBS, 10 mM HEPES buffer, pH 7.4, 100 U/mL penicillin, and 100 μg/mL streptomycin (growth media) and incubated in a humidified atmosphere of 5% CO at 37°C. For experiments, thyrocytes were plated in 12-well plates at 1 × 10⁵ cells per well in growth media, that is, DMEM containing 10% FBS. The next day the cells were stepped down overnight in serum-free media comprising DMEM, BSA (0.1%), HEPES (10 mM), penicillin (100 U/mL), and streptomycin (100 μg/mL) (serum-free media). The following day, the cells were treated with the experimental conditions.

Cell Proliferation

Human thyrocytes were seeded at 1 × 10⁵ cells per well in 12-well plates in growth media. Next day cells were treated with 100 ng/mL EGF, 1 mU/mL TSH, or their combination or with 1 mU/mL TSH with the indicated doses of EGF in arresting media. After 5 days, cells were trypsinized and counted on a Vi-Cell counter (Beckman-Coulter, Brea, CA) according to the manufacturer’s instructions. Additionally, the EGF dose response in the presence of 1 mU/mL TSH was in parallel assessed for thyroid differentiation marker expression.

Measurement of mRNA Expression of Thyroid Differentiation Markers

Human thyrocytes were seeded at 1 × 10⁵ cells per well of 12-well plates in growth media, allowed to attach overnight, then arrested overnight in serum-free media. Next day, 5 µM or a dose response of erlotinib and AG1478; 10 µM or a dose response of cetuximab and panitumumab; and 5 µM MK2206 or vehicle (DMSO or PBS) were added for a 20-minute pretreatment, followed by the 48 hours of treatment with either 1 mU/mL TSH or TSH dose response. Levels of human TG, TPO, DIO2, NIS, and TSHR mRNA were measured in total RNA preparations using RNeasy Mini Kits followed by reverse transcription to synthesize first-strand cDNA using a High-Capacity cDNA Archive Kit. Quantitative reverse transcription polymerase chain reaction (RT-PCR) was performed using the prepared cDNA and iTaq™ Universal Probe Supermix and primers and probes were obtained from Taqman, Assay-on-Demand. Quantitative RT-PCR results were normalized to human β-actin.

Western Blotting

Whole-cell RIPA lysates were separated by SDS polyacrylamide gel electrophoresis under reducing conditions on gradient 4% to 12% Precast NuPAGE gels using MOPS SDS running buffer, then transferred to nitrocellulose membranes using a Mini Trans-Blot® electrophoretic transfer cell (Bio-Rad). Membranes were blocked with Odyssey blocking buffer. Antibodies were diluted (TG 1:10 000, and TPO and β-tubulin 1:1000) with Odyssey blocking buffer with 0.1% Tween® 20 and incubated with membranes with gentle shaking overnight at 4°C. Membranes were washed with Tris-buffered saline (TBS) containing 0.1% Tween® 20. IRDye secondary antibodies were diluted 1:10 000 in Odyssey blocking buffer with 0.1% Tween® 20 for 1 hour, washed with TBS containing 0.1% Tween® 20, then TBS, and imaged on an Odyssey CLx using the AutoScan function. Densitometry was performed using Li-COR Image Studio™ Version 2.1. TG and TPO protein expression was normalized to total β-tubulin and normalized to maximal TSH only response.

Measurement of Total and Phospho-EGFR and EGF in Thyrocytes

For phospho-EGFR assessment, human thyrocytes were seeded at 1 × 10⁵ cells per well of 12-well plates in growth media. Next day, they were arrested in serum-free media for 24 hours, then 5 µM or a dose response of erlotinib and AG1478; and 10 µM or a dose response of cetuximab and panitumumab or vehicle (DMSO or PBS) were added for a 20-minute pretreatment followed by a stimulation with EGF (1, 30, or 100 ng/mL), TSH (1, 100 mU/mL, or a dose response) for 5 minutes at 37°C. Following stimulation, cells were washed 2 times with ice-cold PBS, then lysed and assayed by phospho-EGFR ELISA according to the manufacturer’s directions. For total EGFR assessment, thyrocytes were seeded and arrested as described above, then stimulated with EGF (1 and 100 ng/mL) or TSH (1 and 100 mU/mL). After 48 hours, cells were washed twice with ice-cold PBS, then lysed and assayed by total EGFR ELISA according to the manufacturer’s protocol. For EGF secretion, cell culture media conditioned in the presence of TSH dose response for 48 hours was measured in EGF ELISA according to the manufacturer’s instructions.

Phospho-AKT Stimulation in Thyrocytes

Human thyrocytes were seeded at 1 × 10⁵ cells per well of 12-well plates in growth media, allowed to attach overnight, then arrested overnight in serum-free media. Next day 5 µM AG1478 was added for a 20-minute pretreatment followed by a stimulation by TSH (1 or 100 mU/mL), EGF (100 ng/mL), or a combination of TSH and EGF (1 mU/mL and 100 ng/mL, respectively). After 10 minutes, cells were washed 2 times with ice-cold PBS, then the lysates were analyzed for phospho-AKT by ELISA according to the manufacturer’s directions.

Statistics

All analyses were conducted using GraphPad Prism (v8.2.1). Significance was assessed by unpaired Student’s t-test, and P values are reported in the figure legends. Significance testing was conducted using P = .05 as the threshold for statistical significance.

Results

EGF Inhibits TSH-mediated Upregulation of Genes Involved in Thyroid Hormone Biosynthesis and Stimulates Human Thyrocyte Proliferation In Vitro

TSH and EGF have been shown to have an antagonistic effect on TG mRNA levels in dog thyrocytes (4). Here we sought to compare their effects on thyroid differentiation markers—TG, TPO, DIO2, NIS, and TSHR and proliferation in cultures of primary human thyrocytes (Fig. 1). The number of thyrocytes doubled in 5 days in serum-free conditions but the number of cells treated with TSH was not statistically different from the basal (untreated) wells (1.87 vs 2.06 × 10⁵ cells per well, respectively). EGF-treated wells had on average 1.5-fold more cells than TSH-treated wells (2.78 × 10⁵ cells per well). Thyrocytes treated with a combination of TSH and EGF had a higher cell number than the wells treated with TSH alone, but the presence of TSH lowered the proliferative effect that was observed when cells were exposed to EGF alone (2.33 × 10⁵) (Fig. 1A).

Figure 1.

Effects of EGF alone and in combination with TSH on proliferation and differentiation of human thyrocytes in primary culture. (A) Under our culture conditions (in the absence of other growth factors), EGF stimulates proliferation, TSH does not affect proliferation, but inhibits proliferation stimulated by EGF. Human thyrocytes were seeded at 1 × 10⁵ cells per well, allowed to attach overnight, then stimulated in 0.1% BSA-containing DMEM supplemented with antibiotics and HEPES with 100 ng/mL EGF, 1 mU/mL TSH, or their combination. Cells were harvested and counted after 5 days of treatment. The difference between groups was analyzed using unpaired Student’s t-test. The difference between Basal (untreated control) wells and TSH-treated wells was not statistically significant. TSH vs EGF and Control vs EGF were highly significant (P = 0.0002 and 0.008, respectively). P values for TSH vs TSH and EGF and EGF vs TSH and EGF were also statistically significant (.0137 and .0210, respectively). Data are expressed as mean ± SEM, n = 3 patient donors. (B) EGF inhibits differentiation induced by TSH at the level of mRNA. Cells were seeded as described in A and incubated with 1 mU/mL TSH with the indicated doses of EGF for 5 days. Then cells were either harvested and counted or lysed and analyzed for TG, TPO, DIO2, NIS, and TSHR mRNA expression using quantitative RT-PCR. Each experiment contained biological duplicates in a minimum of 3 patient donors. (C and D) EGF inhibits TG and TPO protein expression. Cells were seeded as described in A, allowed to attach overnight, arrested for 24 hours then incubated with 1 mU/mL TSH with the indicated doses of EGF. After 5 days, cells were lysed, TG and TPO expression was measured by Western blot. A representative of 3 Western blots is shown (C). Protein expression was normalized to β-tubulin. Data are shown as mean ± SEM of the percentage of the maximal TSH only response; from 3 patient donors.

The EC₅₀s for EGF-mediated inhibition of the effects of 1 mU/mL TSH on thyroid differentiation markers were 0.43, 0.72, 0.75, 1.3, and 0.64 ng/mL for TG, TPO, DIO2, NIS, and TSHR, respectively (Fig. 1B). TSH-stimulated upregulation of thyroid differentiation markers was markedly inhibited by 100 ng/mL EGF: TG was reduced by 76.8%, TPO by 66.4%, DIO2 by 86.5%, NIS by 87.1%, and TSHR by 65.6% for mRNA expression (Fig. 1B). Fig. 1B demonstrates an inverse relationship between thyrocyte proliferation and differentiation, that is, with increasing concentrations of EGF the number of cells increased, and thyroid differentiation markers were downregulated. Fig. 1C and 1D show that the EGF-mediated inhibition of TG and TPO mRNA correlates with the protein expression of these thyroid markers. TG protein expression was inhibited by 36.7% and TPO was reduced by 54.3%.

These data confirm the antagonistic effect of combination treatment with TSH and EGF on TG reported in Roger et al (4), and expand it to show that antagonism applies to other thyroid-specific genes also including TPO, DIO2, NIS, and TSHR. Under these conditions, TSH did not stimulate proliferation of human thyrocytes, but partially antagonized proliferation stimulated by EGF (Fig. 1).

EGF Receptor Tyrosine Kinase Inhibitors, but not EGFR-blocking Antibodies, Inhibit TSH-mediated Upregulation of Thyroid Genes

Evidence consistent with EGFR transactivation by TSH has been previously reported in human follicular thyroid carcinoma cells via the mitogen-activated protein kinase (MAPK) pathway. In primary thyrocytes, however, MAPK activation occurred independently of EGFR (9). Therefore, we wanted to see if EGFR transactivation was involved in regulation of thyroid-specific gene expression in human thyrocytes.

Because EGFR activation by EGF typically involves autophosphorylation of EGFR (formation of phospho-EGFR) (6), we wanted to determine whether EGFR phosphorylation was involved in EGFR transactivation by TSH. In preliminary experiments (Fig. 2A), we confirmed that 2 EGFR tyrosine kinase inhibitors, AG1478 and erlotinib, act as expected to inhibit phospho-EGFR formation in human thyrocytes.

Figure 2.

EGF receptor tyrosine kinase inhibitor erlotinib inhibits EGF-mediated phosphorylation of EGFR and the TSH-induced increase of thyroid-specific genes in thyrocytes. (A) Erlotinib and AG1478 inhibit phosphorylation of EGFR stimulated by EGF in thyrocytes. Human thyrocytes were seeded at 1 × 10⁵ cells per well, allowed to attach overnight, then arrested in 0.1% BSA-containing DMEM for 24 hours. Erlotinib and AG1478, EGF receptor tyrosine kinase inhibitors were then added at the indicated doses for a 20-minute pretreatment followed by a stimulation by EGF at 30 ng/mL for 5 minutes. Cell lysates were analyzed for phospho-EGFR (pEGFR) by ELISA. Data are expressed as mean ± SD, n = 2 patient donors. (B) EGF receptor tyrosine kinase inhibitor erlotinib inhibits TSH-induced upregulation of TG, TPO, DIO2, NIS, and TSHR mRNAs. Human thyrocytes were seeded at 1 × 10⁵ cells per well, allowed to attach overnight, then arrested in 0.1% BSA-containing DMEM for 24 hours. Erlotinib was then added at 5 µM alone or in combination with 1 mU/mL TSH for 48 hours. Cell lysates were analyzed for TG, TPO, DIO2, NIS, and TSHR mRNAs by RT-qPCR. Each experiment contained biological duplicates. Data are expressed as mean ± SEM, n = 3 patient donors.

We then measured the effect of erlotinib on thyroid-specific gene expression stimulated by TSH (Fig. 2B). Erlotinib (5 µM) decreased upregulation stimulated by TSH (1 mU/mL) of TG mRNA by 76.2%, TPO by 84.0%, DIO2 by 94.6%, and TSHR by 90.0%. Fig. 3 illustrates that AG1478 (5 µM) inhibited the effect of TSH dose response with TG mRNA inhibited by 78.2%, TPO by 88.5%, DIO2 by 61.2%, and TSHR by 93.0% at 1 mU/mL TSH.

Figure 3.

EGF receptor tyrosine kinase inhibitor AG1478 inhibits TSH-induced upregulation of TG, TPO, DIO2, NIS, and TSHR mRNAs. Human thyrocytes were seeded at 1 × 10⁵ cells per well, allowed to attach overnight, then arrested in 0.1% BSA-containing DMEM for 24 hours. AG1478 was then added at 5 µM alone or in combination with TSH at the indicated doses for 48 hours. Cell lysates were analyzed for TG, TPO, DIO2, NIS, and TSHR mRNAs by RT-qPCR. Each experiment contained biological duplicates. Data are expressed as mean ± SEM, n = 3 patient donors.

Two EGFR-blocking antibodies, cetuximab and Vectibix® (panitumumab), that are expected to block ligand binding to EGFR, prevent phospho-EGFR formation stimulated by EGF in human thyrocytes (Fig. 4A). Neither EGFR-blocking antibody had any effect on thyroid gene upregulation mediated by TSH (1 mU/mL) (Fig. 4B). These findings, along with the fact that no detectable EGF was found in the conditioned media in which thyrocytes were cultured in the presence of TSH for 48 hours (data not shown), are consistent with the idea that EGFR transactivation by TSH is ligand-independent in human thyrocytes.

Figure 4.

EGFR-blocking antibodies cetuximab and panitumumab inhibit EGF-stimulated phospho-EGFR but have no effect on TSH regulation of TG, TPO, DIO2, NIS, and TSHR mRNAs (EGFR transactivation). (A) EGFR-blocking antibodies inhibit phospho-EGFR levels stimulated by EGF in thyrocytes. Human thyrocytes were seeded at 1 × 10⁵ cells per well, allowed to attach overnight, then arrested in 0.1% BSA-containing DMEM for 24 hours. Cetuximab and panitumumab were then added at the indicated doses for a 20-minute pretreatment followed by a stimulation by EGF at 30 ng/mL for 5 minutes. Cell lysates were analyzed for pEGFR by ELISA. Data are expressed as mean ± SD, n = 2 patient donors. (B) Cetuximab and panitumumab have no effect on TSH-mediated upregulation of thyroid gene mRNAs. Human thyrocytes were seeded at 1 × 10⁵ cells per well, allowed to attach overnight, then arrested in 0.1% BSA-containing DMEM for 24 hours. Cetuximab or panitumumab were then added at 10 µM alone or in combination with 1 mU/mL TSH for 48 hours. Cell lysates were analyzed for TG, TPO, DIO2, NIS, and TSHR mRNAs by RT-qPCR. Each experiment contained biological duplicates. Data are expressed as mean ± SEM, n = 3 patient donors.

TSH/TSHR Transactivation of EGFR is Different than EGFR Activation by EGF

Once we established that TSH/TSHR transactivation of EGFR was not dependent on secretion of an EGFR ligand, we determined whether TSH, like EGF, caused phosphorylation of EGFR. Fig. 5A demonstrates that EGF treatment caused a rapid phosphorylation of EGFR, but TSH had no effect. Next, we compared the effects of TSH and EGF on EGFR cellular levels because it is known that EGF downregulates EGFR levels (12). After 48 hours of incubation (Fig. 5B). EGFR was markedly downregulated by approximately 61.9% in thyrocytes incubated with EGF whereas TSH did not affect EGFR levels. These observations show that TSH/TSHR transactivation of EGFR, unlike EGFR activation by EGF, occurs in the absence of EGF receptor phosphorylation and downregulation (Fig. 5).

Figure 5.

TSH, unlike EGF, does not stimulate phosphorylation nor downregulation of EGFR. (A) TSH, unlike EGF, does not lead to EGFR phosphorylation. Human thyrocytes were seeded at 1 × 10⁵ cells per well, allowed to attach overnight, then arrested in 0.1% BSA-containing DMEM for 24 hours. Cells were then stimulated by TSH (1 or 100 mU/mL) or EGF (1 or 100 ng/mL) for 5 minutes. Cell lysates were analyzed for phospho-EGFR by ELISA. The difference between groups was analyzed using unpaired Student’s t-test. The differences between basal (untreated) wells and 1 and 100 mU/mL TSH-treated wells were not statistically significant. Basal vs EGF 1 ng/mL and 100 ng/mL were highly significant (P < .0001 for both conditions). Each experiment contained biological duplicates. Data are expressed as mean ± SEM, n = 3 patient donors. (B) TSH, unlike EGF, does not cause downregulation of EGFR. Human thyrocytes were seeded at 1 × 10⁵ cells per well, allowed to attach overnight, then arrested in 0.1% BSA-containing DMEM for 24 hours. Cells were then stimulated by TSH (1 or 100 mU/mL) or EGF (1 or 100 ng/mL) for 48 hours. Cell lysates were analyzed for total EGFR by ELISA. The difference between groups was analyzed using unpaired Student’s t-test. The differences between basal (untreated) wells and 1 or 100 mU/mL TSH-treated wells were not statistically significant. Basal vs EGF 1 ng/mL and 100 ng/mL were significant (P < .0120 and 0.0082, respectively). Each experiment contained biological duplicates. Data are expressed as mean ± SEM, n = 3 patient donors.

EGF and TSH Stimulation Converged on a Pathway that Included Protein Kinase B

It is known that both TSHR and EGFR can activate AKT by stimulating formation of phospho-AKT, which plays roles in many cellular processes such as cell proliferation, transcription and apoptosis (13). First, we wanted to determine whether EGFR receptor kinase inhibitors inhibited TSH stimulation of AKT phosphorylation in human thyrocytes. Fig. 6A shows that TSH stimulates phospho-AKT formation in human thyrocytes and that AG1478 inhibited TSH-mediated activation of AKT. TSH (1 mU/mL) caused a roughly 2-fold increase in phospho-AKT over the basal level that was inhibited by AG1478 by 42.3% and an almost 4-fold increase in phospho-AKT stimulated by 100 mU/mL TSH that was inhibited by 65.4%. Thus, we identified that a step of convergence of the TSHR/EGFR transactivation occurs at AKT (Fig. 6A).

Figure 6.

TSH stimulates phospho-AKT formation in an EGFR-dependent manner, and TSH and EGF synergistically stimulate phospho-AKT formation. (A) EGFR kinase inhibitor AG1478 inhibits TSH-mediated AKT phosphorylation. Human thyrocytes were seeded at 1 × 10⁵ cells per well, allowed to attach overnight, then arrested in 0.1% BSA-containing DMEM for 24 hours. AG1478 (5 µM) was added for a 20-minute pretreatment followed by a stimulation by TSH (1 or 100 mU/mL) for 10 minutes. Cell lysates were analyzed for phospho-AKT by ELISA. The difference between groups was analyzed using unpaired Student’s t-test. The differences between DMSO basal (untreated) wells and 1 and 100 mU/mL TSH-treated wells were highly statistically significant (P < .0001 for both conditions). AG1478-treated Basal vs TSH 1 mU/mL was not significant. For AG1478-treated Basal vs TSH 100 mU/mL P = .0160. The difference between Control DMSO- and AG1478-treated cells was insignificant: P = .6356. Each experiment contained biological duplicates or triplicates. Data are expressed as mean ± SEM, n = 3 patient donors. (B) TSH and EGF synergistically potentiate AKT phosphorylation. Human thyrocytes were seeded at 1 × 10⁵ cells per well, allowed to attach overnight, then arrested overnight in 0.1% BSA-containing DMEM. Cells were stimulated by EGF (100 ng/mL) or TSH (1 mU/mL) or both for 10 minutes. Cell lysates were analyzed for phospho-AKT by ELISA. The difference between groups was analyzed using unpaired Student’s t-test. The differences between basal (untreated) wells and 1 mU/mL TSH-treated wells as well as 100 ng/mL EGF-treated wells were statistically significant (P = .0229 and P < .0001, respectively). The difference between EGF and EGF + TSH was highly significant (P < .0007). Each experiment contained biological duplicates or triplicates. Data are expressed as mean ± SEM, n = 3 patient donors. The dotted line represents the additive level of phospho-AKT independently stimulated by TSH and EGF.

Next, we wanted to determine the effects on AKT phosphorylation stimulated by TSH and EGF separately and in combination (Fig. 6B). EGF response at 100 ng/mL was the most robust (and set at 100% in Fig. 6B). By comparison, the effect of 1 mU/mL TSH stimulation was only 15.8%. However, the combination TSH and EGF treatment was 58% higher than the additive (115.8% dotted line) and, therefore, exhibited synergy.

MK2206, a Selective Allosteric Pan-AKT Inhibitor, Prevents TSH-mediated Upregulation of TG, TPO, DIO2, NIS, and TSHR mRNAs

To support our hypothesis that phospho-AKT was involved in TSHR/EGFR transactivation leading to upregulation of thyroid hormone biosynthetic genes, we studied the effect of a selective allosteric inhibitor of AKT, MK2206, on TSH-stimulated gene expression. Fig. 7 illustrates that MK2206 (5 µM) inhibited stimulation by 1 mU/mL TSH on the mRNA levels of TG by 65.8%, TPO by 70.0%, DIO2 by 69.2%, and TSHR by 70.4% that is consistent with hypothesis of AKT involvement in TSH/TSHR transactivation of EGFR that leads to regulation of thyroid genes.

Figure 7.

Selective allosteric inhibitor of AKT MK2206 inhibits TSH regulation of TG, TPO, DIO2, NIS, and TSHR mRNAs. Human thyrocytes were seeded at 1 × 10⁵ cells per well, allowed to attach overnight, then arrested in 0.1% BSA-containing DMEM for 24 hours. MK2206 was then added at 5 µM alone or in combination with TSH dose response for 48 hours. Cell lysates were analyzed for TG, TPO, DIO2, NIS, and TSHR mRNAs by RT-qPCR. Each experiment contained biological duplicates or triplicates. Data are expressed as mean ± SEM, n = 3 patient donors.

Discussion

GPCR transactivation of RTKs is a well-known mechanism of regulation of multiple cell types (14). Specifically, there have been many examples of GPCR (both classes A and B) transactivation of EGFR, especially well studied in cancer biology and cardiovascular physiology (7). The mechanism of transactivation can significantly diversify signaling networks within cells and has been implicated in promoting both advantageous and disadvantageous physiological and pathophysiological outcomes, making the GPCR/EGFR interactions attractive targets for drug discovery (15). The findings reported herein add to our understanding of the complex roles of EGFR in regulation of thyroid cell function, in particular, of human thyroid cells. They underscore the importance of the mode of activation of EGFR signaling pertaining to the effects on proliferation and differentiation. Under the conditions employed in our experiments (11), where the cells were incubated strictly in arresting media lacking FBS and any hormones or growth factors, we found that EGFR activation initiated by direct binding of EGF to EGFR leads to stimulation of proliferation and dedifferentiation whereas EGFR activation via transactivation by TSH binding to TSHR leads to differentiation without stimulating proliferation. The effect of EGF activation of EGFR to stimulate proliferation was well appreciated previously but was thought to be independent of TSH/TSHR (16). However, a role(s) for TSH/TSHR transactivation of EGFR was not previously studied.

One of the challenges in elucidating the mechanisms of transactivation is caused by the fact that many signaling effectors, such as Src, PI3K, ERK1/2, and MAPK can act as convergence points for multiple signaling pathways, including those that are GPCR and RTK mediated, making it more difficult to tease out signaling events directly attributable to transactivation (15). In in our case, where the Akt pathway is activated by both TSH and EGF receptors, AG1478 and erlotinib have been useful in silencing the EGFR component of transactivation in human thyroid cells. Utilizing these 2 inhibitors of EGFR kinase activity, we demonstrated TSH/TSHR transactivation of EGFR by showing that they inhibited the upregulation of the mRNAs of genes involved in thyroid hormone synthesis—TG, TPO, DIO2, NIS, and TSHR induced by TSH.

The mechanism of transactivation, in general, is most readily divided into 2 types—those mediated by GPCR-stimulated generation of an extracellular RTK ligand which, in an autocrine or paracrine manner, activates an RTK or activation that is independent of an extracellular RTK ligand (5, 6). We demonstrated that TSH/TSHR transactivation of EGFR in human thyroid cells is independent of an extracellular EGFR ligand by showing that 2 antibodies that completely inhibited EGFR activation by EGF had no effect on TSH/TSHR transactivation. Additionally, this observation was supported by the finding that TSH did not cause upregulation of EGF in thyrocytes. Moreover, we showed that TSH/TSHR transactivation of EGFR is different than EGFR activation by EGF by showing that EGF activation led to rapid phosphorylation of EGFR whereas transactivation occurred in the absence of EGFR phosphorylation, and that EGF caused downregulation of the level of EGFR whereas transactivation had no effect on the EGFR level.

Lastly, we demonstrated that both EGF and TSH stimulation of human thyroid cells converged on a pathway that included AKT because (1) TSH, like EGF (17), stimulated phosphorylation of AKT (phospho-AKT) that was inhibited by EGFR kinase inhibitors (18); (2) when both EGF and TSH were added together the effect on phospho-AKT was synergistic; and (3) TSH-induced upregulation of thyroid hormone synthetic gene mRNAs was inhibited by the AKT inhibitor MK2206.

Huang et al (19) demonstrated that activation of TSHR by thyrostimulin, a glycoprotein hormone that like TSH binds to TSHR, promotes proliferation of NIH:OVCAR-3 ovarian cancer cells via trans-regulation of the EGFR pathway by increasing both the protein amount of EGFR present in the cells and the levels of phosphorylated EGFR. The authors concluded that thyrostimulin-induced ovarian cancer cell proliferation was independent of the PKA pathway and seemed to mainly rely on crosstalk with EGFR, which subsequently activates the downstream AKT and ERK signaling systems. Going forward, it would be useful to study whether there are other convergence pathways involved in the mechanism of the ligand-independent EGFR transactivation by TSHR, such as ERK1/2, which is also activated downstream of both receptors, as well as other effectors and compare those pathways in normal and cancerous thyroid tissues.

We reported previously that there was an interaction between TSHR and another RTK, namely the insulin-like growth factor-1 (IGF-1) receptor (IGF1R), in human thyrocytes that we termed TSHR/IGF1R crosstalk (20). TSHR/IGF1R crosstalk, like EGFR transactivation, was independent of an extracellular RTK agonist. However, the molecular details of TSHR transactivation of EGFR are different than those of TSHR/IGF1R crosstalk. For example, TSHR/IGF1R crosstalk is dependent on close proximity of these receptors that is maintained by an apparent scaffolding effect of beta-arrestin 1 (21); we showed that knockdown of beta-arrestin 1 increased the distance between TSHR and IGF1R and prevented crosstalk. In contrast, knockdown of beta-arrestin 1 had no effect on TSHR transactivation of EGFR (data not shown).

It has been shown that EGFR is overexpressed in poorly differentiated and anaplastic thyroid carcinomas (22, 23). In contrast to other human solid cancers that are characterized by the presence of EGFR activating mutations in the tyrosine kinase domain (24), these have been detected only in a minority of thyroid carcinomas (25‐27). Landriscina et al (22) hypothesized that activation of EGF signaling may be driven by the upregulation of its components, rather than by genomic mutation events. Thus, for future directions, it would be beneficial to study how relative levels of expression of TSHR and EGFR in normal thyrocytes and thyroid malignancies may shift thyrocyte signal transduction from EGFR transactivation described herein, which is necessary for thyroid differentiation to EGFR activation by EGF favoring progression toward an angiogenic, dedifferentiated, TSH-independent phenotype.

In conclusion, we have shown that TSH/TSHR causes EGFR transactivation in human thyroid cells in primary culture that is independent of extracellular EGFR ligand and that mediates, in part, TSH regulation of thyroid hormone biosynthetic gene expression. Fully understanding the nature of this transactivation may have clinical relevance for the treatment of thyroid diseases, especially thyroid cancer, but this has not yet been achieved.

Abbreviations

AKT

protein kinase B

BSA

bovine serum albumin

DIO2

deiodinase type 2

DMEM

Dulbecco’s modified Eagle’s medium

EGF

epidermal growth factor

EGFR

epidermal growth factor receptor

ELISA

enzyme-linked immunosorbent assay

FBS

fetal bovine serum

GPCR

G protein–coupled receptor

IGF-1

insulin-like growth factor-1

MAPK

mitogen-activated protein kinase

NIS

sodium/iodide symporter

RTK

receptor tyrosine kinase

RT-PCR

reverse transcription polymerase chain reaction

TBS

Tris-buffered saline

TG

thyroglobulin

TPO

thyroid peroxidase

TSH

thyrotropin

TSHR

thyrotropin receptor

Funding

This work was supported by the Intramural Research Program of the National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health (Z01 DK011006).

Disclosures

The authors declare no competing financial interests. The authors declare no conflict of interest.

Data availability

Original data generated and analyzed during this study are included in this published article or in the data repositories listed in References.