TSH overcomes Braf^V600E-induced senescence to promote tumor progression via downregulation of p53 expression in papillary thyroid cancer

Abstract

The BRAF^V600E mutation is found in approximately 40% of papillary thyroid cancers (PTC). Mice with thyroid-specific expression of Braf^V600E (TPO–Braf^V600E) develop PTC rapidly with high levels of serum thyroid-stimulating hormone (TSH). It is unclear to what extent the elevated TSH contributes to tumor progression. To investigate the progression of Braf^V600E-induced PTC (BVE–PTC) under normal TSH, we transplanted BVE–PTC tumors subcutaneously into nude and TPO–Braf^WT mice. Regression of the transplanted tumors was observed in both nude and TPO–Braf^WT mice. They were surrounded by heavy lymphocyte infiltration and oncogene-induced senescence (OIS) was demonstrated by strong β-gal staining and absence of Ki-67 expression. In contrast, BVE–PTC transplants continued to grow when transplanted into TPO–Braf^V600E mice. The expression of Trp53 was increased in tumor transplants undergoing OIS. Trp53 inactivation reversed OIS and enabled tumor transplants to grow in nude mice with characteristic cell morphology of anaplastic thyroid cancer (ATC). PTC-to-ATC transformation was also observed in primary BVE–PTC tumors. ATC cells derived from Trp53 knockout tumors had increased PI3K/AKT signaling and became resistant to Braf^V600E inhibitor PLX4720, which could be overcome by combined treatment of PI3K inhibitor LY294002 and PLX4720. In conclusion, BVE–PTC progression could be contained via p53-dependent OIS and TSH is a major disruptor of this balance. Simultaneous targeting of both MAPK and PI3K/AKT pathways offer a better therapeutic outcome against ATC. The current study reinforces the importance of rigorous control of serum TSH in PTC patients.

INTRODUCTION

Thyroid cancer is the most common type of endocrine malignancy and its incidence has risen rapidly in recent years, especially among women.¹ Histologically, it can be classified into papillary thyroid cancer (PTC), follicular thyroid cancer and anaplastic thyroid cancer (ATC). PTC accounts for more than 80% of thyroid cancer cases.² Surgery combined with radioactive iodine therapy is still the treatment of choice for both PTC and follicular thyroid cancer.³ However, the 10-year recurrence rate is about 20–30% among patients who are older than 45 years and have large invasive tumors or extensive lymph-node metastases.^4,5 Currently, there is no effective treatment for radioiodine-resistant metastatic disease, with a 10-year survival rate of less than 15%.⁶ A better understanding of thyroid cancer biology is necessary to develop new treatment strategies.

The RAS–RAF–MEK–ERK MAP kinase signaling pathway (MAPK) has an important role in the initiation and progression of PTC. Among genetic alterations detected in PTC, BRAF^V600E is the most common mutation (44%) and has been associated with poorer prognosis and more aggressive clinical outcome.⁷ BRAF^V600E can downregulate the expression of genes (NIS, TG, TPO) involved in thyroid hormone synthesis and promote dedifferentiation processes in PTC.^8,9 In humans, serum thyroid-stimulating hormone (TSH) increases with age and mean TSH is significantly higher in thyroid cancer patients.¹⁰ Even though serum TSH is within normal range from 0.4 to 4.0 (mIU/l) in most PTC patients, the frequency of malignancy increases when TSH is greater than 0.9 mU/l.^11–13 In Braf^V600E-induced PTC mouse models, however, serum TSH is more than 100-fold higher owing to severe hypothyroidism.^14–16 The difference may be due to the fact that, in humans, PTC develops from minority loci of thyrocytes carrying BRAF^V600E and the surrounding majority of normal thyrocytes can still maintain normal thyroid function, whereas, in mice, all the thyrocytes carry mutant Braf^V600E and normal thyroid function cannot be maintained, resulting in hypothyroidism with elevated TSH. In mice with normal serum TSH¹⁶ or blocked TSH signaling,¹⁵ tumorigenesis may occur but is significantly delayed, resulting in small and localized tumors. Shimamura et al.¹⁷ demonstrated recently that Braf^V600E cannot induce a tumor when it is expressed postnatally in thyrocytes without TSH stimulation. These studies indicate that development of aggressive BVE–PTC requires constant TSH stimulation.

Braf^V600E-induced senescence occurs in human and mouse naevi, protecting melanocytes against Braf^V600E-driven proliferation and progression into melanoma.^18,19 This process has not been demonstrated in mouse models of Braf^V600E-induced thyroid cancer and elevated TSH may be involved in preventing oncogene-induced senescence (OIS). It is known that TSH stimulates the growth and development of thyroid cancer and elevated serum TSH is associated with higher thyroid cancer incidence and recurrence.^20–22 The study by Franco et al.¹⁵ demonstrated that, once PTC is established, suppression of TSH does not change the phenotype. However, it is not known whether tumors will continue to progress or OIS will occur to suppress tumor progression under normal serum TSH. In the present study, we investigated BVE–PTC progression under normal serum TSH and explored the molecular mechanisms regulating the process.

RESULTS

Regression of BVE–PTC transplants in nude and TPO–Braf^WT mice To investigate whether tumors can continue to grow under normal TSH, we collected 16 BVE–PTC tumors from 4- to 6-month-old TPO–Braf^V600E mice and transplanted them subcutaneously into eight nude and eight TPO–Braf^WT mice, respectively. These tumor transplants were monitored for up to 7 months. As shown in Figure 1A, serum TSH levels from nude and TPO–Braf^WT mice were more than 100-fold lower than TPO–Braf^V600E mice. The serum TSH levels from five TPO–Braf^V600E mice (5 months old) were all above the detection limit of 50 000 pg/ml. The average TSH levels from five TPO–Braf^WT and five nude mice of the same age were 439.6 ± 39.8 and 426.4 ± 9.6 pg/ml, respectively. The mice were genotyped and a representative result is shown in Figure 1B. Tumor transplants were not able to grow and often regressed by more than 50% over the 7-month period. Large cysts were formed in 2 of the 16 tumor transplants. Histology of tumor transplants showed heavy lymphocyte infiltration surrounding the tumor tissue (Figure 1C). Macrophages were often present in the empty spaces of the tumor tissue, which were more frequently seen in the TPO–Braf^WT mice (Figure 1C, c and d) than in the nude mice (Figure 1C, a and b), indicating tumor clearance by the macrophages. We next investigated whether BVE–PTC tumor transplants could grow subcutaneously in TPO–Braf^V600E mice under high serum TSH. As shown in Figure 1D, tumor transplants could continue to grow in TPO–Braf^V600E mice. The average weight of 4-month-old tumor transplants from TPO–Braf^V600E mice was 1410 ± 1067 mg as compared with 36.3 ± 11.4 mg from TPO–Braf^WT mice (P < 0.01). The average weight of tumors before transplantation was 76.0 ± 12.6 mg. The histology of the tumor transplants from TPO–Braf^V600E mice showed significant tumor growth and less lymphocyte infiltration, although macrophages could still be seen in the empty spaces of tumor tissue. Furthermore, the BVE–PTC tumor transplants could continue to grow in TPO–Braf^WT mice after they developed hypothyroidism following treatment of anti-thyroid drug propylthiouracil (0.1% in drinking water).

Figure 1.

Regression of BVE–PTC transplants in nude and TPO–Braf^WT mice. (A) Serum TSH level from TPO–Braf^V600E, TPO–Braf^WT and nude mice. The serum TSH levels from five TPO–Braf^V600E mice (5 months old) are all above the detection limit of 50 000 pg/ml. The average TSH levels from five TPO–Braf^WT and five nude mice of same age are 439.6 ± 39.8 and 426.4 ± 9.6 pg/ml, respectively. (B) Genotyping of TPO–Braf^V600E (a) and TPO–Braf^V600E–Trp53^−/− mice (b). The PCR product was run on a 1.5% agarose gel. The LSL–Braf^V600E was recombined only in the thyroid (activation of Braf^V600E) as a result of Cre-mediated deletion of a floxed transcriptional stop sequence. The LSL-Braf^V600E allele was not detected in the BVE-PTC cell line, indicating complete Cre-mediated recombination in the thyroid tumor cell line (a, lower panel), and no fibroblast contamination. Wild-type and mutant Trp53 alleles from BVE–PTC tumors are shown by 680-bp and 320-bp fragments, respectively. (C) Histology of BVE–PTC transplants. Tumors from 4- to 6-month-old TPO–Braf^V600E mice were transplanted subcutaneously into nude and TPO–Braf^WT mice. They were removed for histology after 4–7 months. The tumor transplants are surrounded by heavy lymphocyte infiltration (indicated by dark arrows) from both nude mice (a and b) and TPO–Braf^WT mice (c and d). Macrophages are often present in the empty spaces of tumor tissue (indicated by light arrows). (D) Tumor growth in TPO–Braf^V600E mice. TPO–Braf^V600E mice are about the half size of TPO–Braf^WT mice due to severe hypothyroidism and tumor transplants are indicated by arrows (a). The size of tumor transplants after 4-month implantation in TPO–Braf^V600E and TPO–Braf^WT mice (b). Histology of tumor transplants from TPO–Braf^V600E mice (c and d). There is more tumor tissue and less lymphocyte infiltration. Macrophages in the empty spaces of tumor tissue are indicated by light arrows.

Induction of senescence by Braf^V600E in BVE–PTC transplants via upregulation of Trp53

To investigate the mechanisms leading to tumor regression in tumor transplants under normal serum TSH, we studied p53 expression by western blot analysis of primary BVE–PTC tumors and tumor transplants. As shown in Figure 2A, more than fivefold increase in p53 expression was observed in a BVE–PTC tumor transplant when compared with that from a 2-month-old primary tumor. There was also a gradual increase in p53 expression from primary tumors: lower expression in the early stage of the tumor (2 months old) and higher expression in the late stage of the tumor (4–6 months old). This may represent a compensatory mechanism of p53 against Braf^V600E-driven tumor progression. No significant change was observed in Cdkn2a(p16^INK4A) and Cdkn1a(p21^CIP1/WAF1) expression (data not shown). The p53 expression was associated with strong senescence-associated beta-galactosidase (SA-beta-Gal) activity (strong SA-beta-Gal staining) and the absence of cell proliferation (absence of Ki-67 immunostaining) in the tumor transplants, whereas no SA-beta-Gal staining and strong Ki-67 immunostaining were demonstrated in primary BVE–PTC tumors (Figure 2B). These data suggest that Braf^V600E-induced senescence has a major role in tumor regression and is mediated by p53. Heavy lymphocyte and macrophage infiltration in the tumor transplants indicate that innate immune response may be triggered by OIS for tumor clearance. To investigate what cytokines and chemokines were involved, we analyzed 32 cytokines and chemokines in the fluid of cysts from two tumor transplants and compared with those in the 48-h conditioned culture medium from the BVE–PTC cell line which we established from a 6-month-old primary BVE–PTC tumor. As shown in Figure 2C, the production of inflammatory cytokines and chemokines was significantly increased in the tumor fluid such as MIP-1α (macrophage inflammatory protein-1α, which is involved in the acute inflammatory response for the recruitment and activation of polymorphonuclear leukocytes), MIP-1β (a chemoattractant for natural killer cells and monocytes), interleukin (IL)-1β (produced by activated macrophages), and TNF (tumor necrosis factor, also called TNF-α), which is also involved in the acute inflammatory response and produced mainly by activated macrophages. The significance of elevated IL-10 (an anti-inflammatory cytokine) and eotaxin in the fluid is unclear, but they may have a role in the acute inflammatory response. In contrast, many tumor-promoting cytokines and chemokines were significantly reduced in the tumor fluid, such as VEGF (vascular endothelial growth factor), G-CSF (granulocyte colony-stimulating factor),²³ LIF (leukemia inhibitory factor),²⁴ IL-13,²⁵ KC (keratinocyte chemoattractant),²⁶ and MIP-2 (macrophage inflammatory protein 2), which has a role in carcinogenesis by stimulating engraftment,²⁷ and MCP-1 (monocyte chemoattractant protein-1) (Figure 2D). The production of MCP-1 by tumors has been shown to be responsible for the recruitment of immunosuppressive macrophages, and neutralization of MCP-1 reduces the growth of prostate, breast and lung cancer in mice.²⁸ These data suggest that OIS can reduce autocrine production of many tumor-promoting cytokines/chemokines by tumor cells, and enhance autocrine/paracrine production of anti-tumor cytokines/chemokines by both tumor cells and infiltrating lymphocytes.

Figure 2.

Braf^V600E-induced senescence in BVE–PTC transplants. (A) Western blot analysis of p53 expression from a 6-month-old BVE–PTC tumor transplant and primary BVE–PTC tumors of different ages. The expression of p53 is significantly higher in the tumor transplants. There is also a gradual increase in p53 expression from primary BVE–PTC tumors: lower expression in early tumors and higher expression in late tumors. Quantification of the western blot was performed using ImageJ software. (B) Strong SA-beta-Gal staining (blue color, a) and no Ki-67 immunostaining (c) in the BVE–PTC tumor transplants. No SA-beta-Gal staining (b) and strong Ki-67 immunostaining (d) in a primary BVE–PTC tumor. (C) Increased release of cytokines and chemokines in the fluid of BVE–PTC transplants. Cytokine and chemokine levels were assayed in the fluid of cysts from two BVE–PTC transplants and compared with those in the 48-h culture medium from a BVE–PTC cell line. The value from culture medium alone was subtracted and data are expressed as mean ± s.e.m. of triplicate wells. Only cytokines/chemokines with significant difference are shown (P < 0.05). (D) Decreased level of cytokines and chemokines in the fluid of cysts from BVE–PTC transplants. Only cytokines/chemokines with significant difference are shown.

Inhibition of Braf^V600E-induced senescence by TSH in BVE–PTC cells via downregulation of p53 expression

To investigate whether TSH can downregulate p53 expression and inhibit Braf^V600E-induced senescence, we analyzed p53 expression in primary BVE–PTC cell cultures. In the absence of TSH, primary BVE–PTC cells could not replicate and displayed characteristic features of cellular senescence such as enlarged and flattened cell morphology with multiple or enlarged nuclei, a vacuolated cytoplasm and increased SA-beta-Gal staining (Figure 3A). When TSH was added into the culture, primary BVE–PTC cells gradually replicated into an immortalized cell line with reduced p53 expression, loss of cellular senescence and decreased SA-beta-Gal staining (Figure 3A). During 5 months of continued culture (about 40 passages), the immortalized cell line eventually became TSH-independent and highly tumorigenic when injected into nude or TPO–Braf^WT mice: all five nude and TPO–Braf^WT mice grew tumors within 6 weeks following subcutaneous injection of 10⁶ cells. In a BVE–PTC cell line with 18 passages, p53 expression was downregulated by 1.6-fold upon 72-h incubation with 10mU/ml TSH (Figure 3B). p-AKT expression was increased by 1.7-fold. The reduction of E-Cadherin expression, a known tumor-suppressor gene involved in cancer progression, was even more significant and became undetectable (Figure 3B). The expression of p53 was further reduced in the immortalized BVE–PTC cell line (about 40 passages) (Figure 4D). It has been reported that increased AKT signaling can negatively regulate p53 activity via AKT-mediated phosphorylation and activation of MDM2²⁹ and inhibit E-Cadherin expression.³⁰ The current data show that TSH can overcome Braf^V600E-induced senescence by downregulating p53 and E-Cadherin expression through upregulation of PI3K/AKT pathway, resulting in tumor progression.

Figure 3.

Inhibition of Braf^V600E-induced senescence by TSH in BVE–PTC cells. (A) Primary BVE–PTC cells were cultured in the presence of 2 mU/ml of bovine TSH for 2 months and stained for SA-beta-Gal activity (a) and p53 expression (c). Without bovine TSH in the culture, primary BVE–PTC cells could not replicate and was under senescence with strong SA-beta-Gal staining (b) and p53 immunostaining (d). (B) Western blot analysis of p53, p-AKT and E-Cadherin expression in a BVE–PTC cell line cultured for 72 h in the presence or absence of 10 mU/ml of bovine TSH. TSH can downregulate both p53 and E-Cadherin expression and increase p-AKT expression. (C) Quantification of p53 expression from the western blot was performed using ImageJ software. (D) Quantification of p-AKT expression from the western blot using ImageJ software.

Figure 4.

Trp53 knockout leads to ATC transformation and loss of Braf^V600E-induced senescence. (A) Transformation of PTC to ATC from a 5-month-old TPO–Braf^V600E–Trp53^+/− mouse (heterozygous Trp53 knockout, a, b, c) and a 3-month-old TPO–Braf^V600E–Trp53^−/− mouse (homozygous Trp53 knockout, d, e, f). ATC tumors (indicated by arrows) can be seen together with PTC. ATC tumors are larger from TPO–Braf^V600E–Trp53^−/− mice (d, e, f ) than from TPO–Braf^V600E–Trp53^+/− mice (a, b, c), and composed of highly pleomorphic giant cells and spindle cells with large, bizarre nuclei containing prominent nucleoli and mitotic figures (c and f). The papillary architecture is replaced by undifferentiated architecture. (B) Loss of tumor regression as a result of Trp53 knockout. Thyroid tumors were removed from 4-month-old TPO–Braf^V600E (with wild-type Trp53) mice and TPO–Braf^V600E–Trp53^−/− (with Trp53 knockout) mice. They were transplanted to nude mice and observed for 4 months. Regression of tumor transplant from a TPO–Braf^V600E mouse (left) and continued tumor growth from a TPO–Braf^V600E–Trp53^−/− mouse (right) are shown. (C) Histology of BVE–PTC tumor transplants. PTC architecture from tumor transplants with Trp53^−/− is completely replaced by undifferentiated cellular structure with many giant and spindle cells characteristic of ATC. Lymphocyte infiltration is not present (a and b). BVE–PTC tumor transplants with Trp53^+/+ are surrounded by heavy lymphocyte infiltration as indicated by arrows and are localized (c and d). (D) Western blot analysis of p53, p-ERK and p-AKT expression among normal thyroid, BVE–PTC, BVE–ATC-c1 and BVE–ATC-c2 cell lines. p-AKT expression is significantly increased in BVE–ATC-c1 and BVE–ATC-c2 cell lines as a result of Trp53 knockout.

Trp53 inactivation leads to ATC transformation and loss of Braf^V600E-induced senescence

To further confirm the role of p53 in Braf^V600E-induced senescence and tumor regression, we created two mouse strains: TPO–Braf^V600E–Trp53^−/− (homozygous Trp53 knockout) and TPO–Braf^V600E–Trp53^+/− (heterozygous Trp53 knockout) by crossing LSL–Braf^V600E and TPO–Cre mice with TSG–p53. AtC development was observed from both TPO–Braf^V600E–Trp53^−/− and TPO–Braf^V600E–Trp53^+/−, although there was a 2–3-month delay and tumors grow slower from TPO–Braf^V600E–Trp53^+/− mice (the loss of WT p53 allele in the tumors was not observed). ATC transformation could be detected as early as 12 weeks (Figure 4A). Histologically, ATC were co-existed with PTC, and ATC tumors were larger from TPO–Braf^V600E–Trp53^−/− mice than from TPO–Braf^V600E–Trp53^+/− mice (Figure 4A). Characteristic features of ATC were present, such as undifferentiated cellular structure, pleomorphic giant cells and spindle cells. In contrast to BVE–PTC tumor transplants with wild-type Trp53, BVE–PTC tumor transplants with either homozygous or heterozygous Trp53 knockout were able to continue to grow in nude mice without elevated serum TSH (Figure 4B). These data indicate that one allelic inactivation of Trp53 is sufficient for ATC transformation and TSH-independent tumor growth. Histology of tumor transplants with Trp53 knockout showed that PTC architecture was completely replaced by ATC architecture and lymphocyte infiltration was absent (Figure 4C, a and b) whereas, in primary tumors, both PTC and ATC components were present (Figure 4A). This may be due to the fact that only the ATC component from primary tumors could continue to grow in nude mice and the PTC component before transformation to ATC was unable to grow under normal serum TSH even though Trp53 is inactivated, resulting in the enrichment of the ATC component.

The BVE–ATC cells from TPO–Braf^V600E–Trp53^−/− tumor transplants could replicate in cell cultures without TSH and two cell lines were established: BVE–ATC-c1 and BVE–ATC-c2. As compared with the BVE–PTC cell line, p-AKT expression was significantly increased in both cell lines as a result of Trp53 knockout (Figure 4D). The increased p-AKT expression in the normal thyroid cells is probably due to the presence of 2 mU/ml bovine TSH in the primary thyroid cell culture. p-ERK levels were also elevated in ATC cells lines when compared with normal, but the level of increase was similar to that in BVE-PTC cell line (Figure 4D). These data indicate that PI3K/AKT pathway activation has an essential role in PTC-to-ATC transformation.

Effectiveness of simultaneous inhibition of both MAPK and PI3K/AKT pathways against ATC cells

Because both MAPK and PI3K/AKT pathways are activated in BVE–ATC cells, targeting both pathways may offer a better therapeutic outcome for ATC. To confirm the hypothesis, BVE–PTC and BVE–ATC-c1 cell lines were treated with PI3K inhibitor LY294002 or BRAF^V600E inhibitor PLX4720 alone or in combination for short term (up to 72 h) and long term (14 days) effects on cell proliferation. The short-term effects on cell proliferation were determined by a non-radioactive MTS assay, and the long-term effects of PLX4720 and LY294002 inhibitors were measured by a colony formation assay. BVE–ATC cells became resistant to PLX4720 treatment (Figures 5A and B, P < 0.01). There were more than 60% of viable BVE–ATC cells after10 μm PLX4720 treatment for 72 h (Figure 5A) and 75% of viable cells after 4 μm PLX4720 treatment for 14 days (Figure 5B). By contrast, less than 30% and 50% of BVE–PTC cells were viable following 72 h treatment (10 μm) (Figure 5A) and 14-day treatment (4 μm) (Figure 5B), respectively. The resistance to PLX4720, however, could be reversed by combined treatment of PLX4720 and LY294002 inhibitors (Figures 5A and B). The viable BVE–ATC cells were reduced to 35% after combined PLX4720 and LY294002 treatment (10 μm each) for 72 h. The effects of combined treatment were even more obvious during long-term treatment using a lower concentration (4 μm) of each inhibitor: 35% from a combined treatment versus 74% from a single PLX4720 treatment (Figure 5B). These data demonstrate the synergistic effects of simultaneous targeting of both MAPK and PI3K/AKT pathways against ATC cells.

Figure 5.

Effect of simultaneous inhibition of both MAPK and PI3K/AKT pathways against ATC cells. (A) Cell proliferation of BVE–PTC and BVE–ATC-c1 cell lines. The cells were plated into 96-well plates (1 × 10³ cells/well) and treated with PI3K inhibitor LY294002 or Braf^V600E inhibitor PLX4720 alone or in combination for up to 72 h. Cell proliferation was determined by a non-radioactive MTS assay kit. Cells treated with 0.1% dimethyl sulfoxide (DMSO) only were served as a vehicle control. Data are expressed as percentage of the vehicle control (100%) in mean ± s.e.m. of triplicate wells. (B) Colony formation assay. Cells were plated into 12-well plates (5 × 10² cells/well) and cultured in the presence of different concentrations of PLX4720, LY294002 or both for 14 days to determine long-term effects of PLX4720 and LY294002. Cells were then stained with 0.5% crystal violet dye (in methanol:de-ionized water, 1:5) for 10 min. After three washes with de-ionized water, the crystal violet dye was released from cells by incubation with 1% sodium dodecyl sulfate (SDS) for 2 h before optical density (OD)_{570 nm} measurement. Data are expressed as percentage of the vehicle control.

DISCUSSION

In the present study, we have demonstrated that OIS is initiated in the BVE–PTC tumors when serum TSH is normal and this process is dependent on normal p53 function. In addition, TSH can reduce p53 expression through upregulation of the PI3K/AKT pathway and inhibit OIS, resulting in tumor progression. Either homozygous or heterozygous inactivation of Trp53 leads to the loss of OIS and promotes transformation of PTC into ATC with subsequently increased activation of the PI3K/AKT pathway. Furthermore, simultaneous targeting of both MAPK and PI3K/AKT pathways may offer a better therapeutic outcome against ATC. These data demonstrate the significant oncogenic role of TSH in thyroid tumor progression: with long-term TSH stimulation, thyroid cancer cells can overcome OIS, gradually become TSH-independent and progress into poorly differentiated cancer.

OIS is a tumor-suppressive mechanism against oncogenic events.³¹ Two major pathways are involved in the process: p16^INK4A–Rb and ARF–p53.³² Unlike replicative senescence, OIS cannot be bypassed by the expression of hTERT,³³ but can be bypassed by inactivation of the p16^INK4A–Rb or ARF–p53 pathway.³² Previous studies have demonstrated that mutations in oncogenes (Ras and Braf and tumor-suppressor genes (Pten and NF1) can trigger OIS in vivo in human tumors and in mouse tumor models,^31,34–36 and is dose-dependent.³⁷ For example, low levels of Ras activation stimulate cellular proliferation and mammary epithelial hyperplasias, whereas high levels of Ras activation induce cellular senescence.³⁷ Heterozygous Pten inactivation promotes tumor initiation and proliferation,³⁸ whereas homozygous loss of Pten inhibits prostate tumorigenesis by triggering a powerful p53-dependent cellular senescence.³⁵ Although senescence occurs mainly in benign tumors to prevent the progression of benign lesions into malignant tumors,^18,36,39 evidence of cellular senescence in early-stage human prostate cancer has also been shown and is p53-dependent.³⁵ Vizioli etal.⁴⁰ have reported evidence of OIS in human primary thyrocytes and thyroid cancers. In thyroid cancer, the expression of OIS markers such as p16^INK4A, p21^CIP1/WAF1 and IGFBP7 is upregulated at early stages and lost during thyroid tumor progression. Although p53 expression is not investigated in thyroid cancer specimens, its expression is induced following Braf^V600E transfection in primary thyrocytes.⁴⁰ The current study has shown that OIS can even occur in vivo in advanced PTC and can be triggered by decreased serum TSH and activation of p53. To our knowledge, this is the first report demonstrating the induction of OIS in advanced tumors. Conventional chemotherapy and radiotherapy also induce cellular senescence in tumors but with significant side effects of cytotoxicity to normal tissues. Treatments aimed at the induction of OIS such as controlling serum TSH, blocking the TSH signaling pathway or stabilizing p53 function may be the cost-effective approach in thyroid cancer therapy, at least for well-differentiated thyroid cancers without p53 mutation. However, these interventions become ineffective when normal p53 function is disrupted, for example, by somatic mutations. Indeed, heterozygous p53 mutation is frequently found in ATC patients who are often non-responsive to TSH-suppressive therapy,⁴¹ and less commonly in well-differentiated thyroid cancers.⁴²

Braf^V600E-induced senescence in benign human naevi occurs through the activation of the p16^INK4A–Rb pathway,¹⁸ whereas, in benign lung tumors, the ARF–p53 pathway is activated.⁴³ Recently, it has been shown that p16^INK4A–Rb pathway is involved in the RAS-induced senescence in human thyrocytes.⁴⁴ Utilization of the cellular senescence pathway is likely cell-type-specific and probably oncogene-specific. It has been reported that NF1, Pten, Raf or Ras mutations can induce a negative feedback response and promote senescence by inhibiting the PI3K/AKT pathway activated by these mutations.^36,45 In mouse prostate cancer, the growth arrest induced by complete Pten inactivation is through the p53-dependent pathway.³⁵ Although Trp53 inactivation fails to produce a tumor phenotype, complete Pten inactivation triggers OIS, resulting in non-lethal invasive prostate cancer after long latency. When both Pten and Trp53 are inactivated, OIS is lost and lethal invasive prostate cancer develops at a young age.³⁵ In mouse thyroid, Pten inactivation causes only benign goiter and follicular adenoma,⁴⁶ whereas combined inactivation of Trp53 and Pten results in follicular thyroid cancer and ATC, the most lethal forms of thyroid cancer.⁴⁷ Although OIS is not explored in those studies, the p53-dependent cellular senescence pathway is likely involved in preventing the progression of benign goiter and follicular adenoma into follicular thyroid cancer and ATC. McFadden et al.⁴⁸ have reported recently that homozygous deletion of Trp53 leads to accelerated progression of PTC to ATC and TSH suppression does not affect tumor progression to ATC in a similar model. Their data are in agreement with our finding that tumor growth will become TSH-independent and likely to progress to ATC when p53 is inactivated. The current study further demonstrates that heterozygous deletion of Trp53 in the context of Braf^V600Ecould also result in ATC transformation, and is sufficient to reverse OIS, indicating haplotype insufficiency of Trp53 to counter ATC transformation. This is consistent with clinical studies that heterozygous, not homozygous, TP53 mutations are often found in ATC patients and cell lines.^41,49 The large lymphocyte infiltration and increased secretion of inflammatory cytokines and chemokines from the senescent tumor cells suggests that OIS can trigger an innate immune response to enhance tumor regression and clearance, and this process is dependent on functional Trp53. Indeed, upregulation of inflammatory cytokines and an innate immune response has been shown in p53-induced cellular senescence in mouse liver cancer.⁵⁰

Activation of the PI3K/AKT pathway not only has a key role in tumor progression, but it also triggers OIS to maintain the balance of growth. In ATC, BRAF^V600E is frequently found together with either gain-of-function mutations in the p110 catalytic subunit of PI3K (PIK3CA) or loss-of-function mutations in TP53. Charles et al.⁵¹ have shown that mice carrying an activated PIK3CA^H1047R are unable to develop tumor. However, concomitant mutations of both PIK3CA^H1047R and BRAF^V600E lead to ATC, suggesting that single activation mutation in either MAPK or PI3K pathway may trigger OIS, whereas concomitant mutations in both pathways would likely neutralize OIS, leading to ATC development. We have identified in the current study that elevated serum TSH or Trp53 mutation could override OIS, resulting in tumor progression. PTEN is a key negative regulator of PI3K/AKT signaling and can increase p53 expression.^52,53 Induction of cellular senescence as a novel therapeutic approach against prostate cancer has been demonstrated.⁴⁵ In the presence of functional p53, the PTEN inhibitor VO-OHpic⁵⁴ induces senescence, whereas the Mdm2 inhibitor Nutlin-3, a p53-stabilizing drug,⁵⁵ enhances senescence by super-activation of p53.⁴⁵ These may represent a viable approach for thyroid cancer patients who are not responsive to TSH-suppressive therapy, but have functional p53 or PTEN. In the absence of functional p53, simultaneous targeting of both MAPK and PI3K/AKT pathways may offer a better therapeutic outcome.

MATERIALS AND METHODS

Mice

LSL–Braf^V600E, TPO–Cre and TSG–p53 strains have been described previously.^56–58 LSL–Braf^V600E mice carry a latent mutant allele of Braf and are kept as heterozygotes. LSL–Braf^V600E was crossed with TPO–Cre mice to generate the TPO–Braf^V600E strain where Braf^V600E is exclusively expressed in thyroid follicular cells under its native promoter through Cre-recombinase-mediated deletion of a floxed transcriptional stop sequence. The resulting TPO–Braf^V600E strain expressed mutant Braf^V600E transcripts at a physiological level. TSG–p53 (Trp53 knockout, B6.129–Trp53^tm1Brd N12) mice were obtained from Taconic (Hudson, NY, USA). To knockout Trp53 in TPO–Braf^V600E mice, TSG–p53 mice were first crossed with LSL–Braf^V600E or TPO–Cre mice to generate Trp53^+/−; Braf^V600E strain or TPO–Cre; Trp53^+/− strain. Trp53^+/−; Braf^V600E mice and TPO–Cre; Trp53^+/− mice were then bred together to create TPO–Braf^V600E–Trp53^−/− (homozygous Trp53 knockout) or TPO–Braf^V600E–Trp53^+/− (heterozygous Trp53 knockout) mice. Female athymic BALB/c-nu/nu mice (4–8 weeks of age) were acquired from the Jackson Laboratory (Bar Harbor, ME, USA). Mice were provided with autoclaved food and water ad libitum. The number of animals was determined by the formula: n = log β/log p; where β is the probability of committing a Type II error (0.05) and p represents the proportion of the animals in the group that have no tumor regression. If tumor regression expected to occur in 30% with 95% chance of detecting the regression, the number of animals needed: n = log 0.05/log 0.7 = 8.4. Thus, nine animals need to be used. We used 16 animals in the study. The animals were randomly chosen and divided into control and experimental groups. No blinding was carried out. The study was approved by the Animal Care and Usage Committee of the institution and conducted in compliance with the Public Health Service Guidelines for the Care and Use of Animals in Research.

Genotyping of transgenic mice

Genotyping of Cre-mediated recombination of the LSL–Braf^V600E targeted allele has been described previously.⁵⁷ Briefly, the following primers were used to detect LSL–Braf^V600E recombination in the mouse tissue: primer A, 5′-AGTCAATCATCCACAGAGACCT-3′, primer B: 5′-GCTTGGCTGGACGTAAACTC-3′, and primer C, 5′-GCCCAGGCTCTTTATGAGAA-3′. Primers A+C detect the wild-type allele (466 bp) and Cre-recombined allele (Lox-Braf^V600E) yielding a 518-bp fragment. Primers B+C detect the LSL–Braf^V600E allele (140 bp). For genotyping of wild-type and mutant Trp53, the following primers were used: p53-X5, 5′-TAAGTCAGAAGCCGGGAGATGG-3′, p53-X3, 5′-AGCCTGAGCATGGAAGTAAGAC-3′, and NEO-19, 5′-CTATCAGGACATAGCGTTGG-3′. p53-X5 and p53-X3 detect wild-type Trp53 yielding a 680-bp fragment. p53-X5 and NEO-19 detect mutant Trp53 yielding a 320-bp fragment. PCR conditions are 95 °C for 5 min followed by 35 cycles of amplification (95 °C for 1 min, 60 °C for 1 min, 72 °C for 1 min with a final extension at 72 °C for 10 min).

Normal thyroid and tumor cell culture

Normal thyroids and thyroid tumors were collected aseptically from donor mice (TPO–Braf^WT, TPO–Braf^7600E and TPO–Braf^V600E–Trp53^−/−) using blunt dissection, mechanically dissociated by mincing and passing through a 40-μM mesh sterile screen, and suspended in DMEM/F12 growth medium (10% fetal bovine serum, 100 units/ml penicillin, 100 μg/ml streptomycin). Cells were further dissociated by incubation in the growth medium containing 100 U/ml type I collagenase (Sigma-Aldrich, St Louis, MO, USA) and 1.0 U/ml dispase I (Roche Diagnostics, Indianapolis, IN, USA) at 37 °C in a rocking water bath for 60 min. The cell suspension was washed twice and re-suspended in a 10-mm culture dish with DMEM/F12 growth medium containing 2 mU/ml bovine TSH (Sigma-Aldrich) to establish a primary normal thyroid culture or a BVE–PTC cell line. Two BVE–ATC cell lines (BVE–ATC-c1 and BVE–ATC-c2) were established from TPO–Braf^V600EμTrp53^−/− tumor transplants by limited dilution and subcloning, and cultured in DMEM/F12 growth medium without bovine TSH. Their genetic background was confirmed by genotyping.

Tumor transplants

To evaluate tumor progression under normal and high serum TSH, thyroid tumors were removed aseptically from TPO–Braf^V600E mice at 4–6 months of age and were implanted subcutaneously on the flank of nude mice (n = 8), TPO–Braf^V600E (n = 8) or their wild-type littermates (n = 8, TPO–Braf^WT) under general anesthesia using a ketamine/xylazine combination (80 mg/kg:10 mg/kg, intraperitoneally). The incision was closed with sutures and tumor growth was monitored weekly. All animals were maintained in a sterile environment on a daily 12-h light/12-h dark cycle. To evaluate tumor progression with Trp53 knockout, thyroid tumors were removed from TPO–Braf^V600E–Trp53^−/− mice or TPO–Braf^V600E–Trp53^+/− at 3 months of age and were implanted subcutaneously on the flank of eight nude mice in each group under the same procedure as above. The genetic background of tumor transplants was confirmed by genotyping.

TSH and cytokine/chemokine measurements

Blood was collected by cardiac puncture. Serum TSH was measured by Luminex’s xMAP using a MILLIPLEX MAP Mouse Pituitary Magnetic Bead Panel following the manufacturer’s instructions (MPTMAG-49K, EMD Millipore Corporation, Billerica, MA, USA). Cytokines and chemokines in the culture medium and tumor fluid were measured by Luminex’s xMAP using a MILLIPLEX MAP Mouse Cytokine/Chemokine Magnetic Bead Panel—Premixed 32 Plex (MCYTMAG-70K-PX32). The following 32 cytokine and chemokine levels were determined: VEGF, Eotaxin, G-CSF, GM-CSF, IFN-γ, IL-1α, IL-1β, IL-2, IL-3, IL-4, IL-5, IL-6, IL-7, IL-9, IL-10, IL-12 (p40), IL-12 (p70), IL-13, IL-15, IL-17, IP-10, KC-like, LIF, LIX, MCP-1, M-CSF, MIG, MIP-1α, MIP-1β, MIP-2, RANTES and TNF-α.

Histology and immunohistochemistry

Histology and immunohistochemical staining were described previously.⁵⁹ Briefly, 4-μm-thick formalin-fixed paraffin-embedded tissue sections were prepared and stained with hematoxylin and eosin, Ki-67 (1:400 dilution) or p53 (1:100 dilution) using a DAKO LSAB+kit (DAKO, Carpinteria, CA, USA). Ki-67 and p53 antibodies were obtained from Cell Signaling Technology, Danvers, MA, USA.

Senescence-associated expression of beta-galactosidase (SA-beta-Gal) activity

SA-beta-Gal activity was detected using an Abcam’s Senescence Detection Kit following the manufacturer’s procedure (Abcam, Cambridge, MA, USA).

Western blot analysis

Cell lysates were obtained by extraction in RIPA buffer (20 mm Tris-HCl, pH7.4, 150 mm NaCl, 5 mm EDTA, 1% NP-40) containing Pierce’s Halt Protease Inhibitor Cocktail (Thermo Scientific, Rockford, IL, USA). Protein concentration was determined by Bradford’s assay using a Bio-Rad protein assay kit (Bio-Rad, Hercules, CA, USA). Proteins (40 μg) were loaded onto a 12% sodium dodecyl sulfate–polyacrylamide gel. Proteins were transferred to a polyvinyl difluoride membrane and subjected to western blot analysis using the following antibodies from either Cell Signaling Technology (Boston, MA, USA): p53 (cat# 2524), phospho-ERK1/2 (cat# 4370), E-Cadherin (cat# 3195), phospho-AKT (cat# 4060), total-ERK1/2 (cat# 9102), total-AKT (cat# 9272) and β-actin (cat# 4970) or Santa Cruz Biotechnology (Dallas, TX, USA): p16^INK4A (cat# sc-1207) and p21^CIP1/WAF1 (cat# sc-397). All antibodies were used at 1:1000 dilutions except for total-ERK1/2, total-AKT and β-actin where 1:5000 dilutions were used. Quantification of western blots was performed using ImageJ software (http://rsb.info.nih.gov/ij/).

Cell proliferation assay

Cell proliferation was measured by a non-radioactive MTS assay kit according to the manufacturer’s procedure (Promega Corp, Madison, WI, USA). Briefly, cells were plated in triplicate into 96-well plates (1 × 10³ cells/well) containing 10% serum and different concentrations of PLX4720 (selective Braf^V600E inhibitor), LY294002 (PI3K inhibitor) or both for up to 72 h (Selleck Chemicals, Houston, TX, USA). For the final 4 h of incubation, 20 μl of CellTiter 96 AQ_ueous One Solution reagent were added into each well for measurement of cell viability.

Colony formation assay

Cells were plated into 12-well plates (×10² cells/well) and cultured for 14 days in the absence or presence of different concentrations of PLX4720, LY294002 or both. Cells were then fixed with methanol for 10 min and stained with 0.5% crystal violet dye (in methanol:de-ionized water, 1:5) for 10 min. After three washes with de-ionized water to remove excess crystal violet dye, the crystal violet dye was released from cells by incubation with 1% sodium dodecyl sulfate for 2 h before optical density (OD)_{570 nm} measurement.

Statistical analysis

Unpaired Student’s t-test (two-tailed) was used. A P value of 0.05 or less was considered significant.