Establishment and Characterization of Amitrole-Induced Mouse Thyroid Adenomatous Nodule-Derived Cell Lines

Abstract

Background:

Thyroid cancer cell lines have been of great value for the study of thyroid cancer. However, the availability of benign thyroid adenoma cell lines is limited.

Methods:

Cell lines were established from thyroid adenomatous nodules that developed in mice treated with the goitrogen amitrole. Expression of epithelial, mesenchymal, and thyroid markers of these established cell lines was determined, and the effect of lentivirus-transduced overexpression of NKX2-1, a master regulator of thyroid development, on the thyroid marker expression was examined. Signal transduction and cell proliferation were evaluated after treatment with insulin-like growth factor-I (IGF-I) and the selective IGF-I receptor (IGF-IR) inhibitor NVP-ADW742. Xenograft studies were performed to examine tumorigenicity of the cells in mice. Whole-genome sequencing (WGS) was used to comprehensively determine the genetic mutations in the established two cell lines.

Results:

Five mouse thyroid adenomatous nodules-derived cell lines named CAT (cells from amitrole-treated thyroids) were established. Among these, two cell lines, CAT458/458s (CAT458s: a subline of CAT458) and CAT459, were found to be positive for epithelial markers and negative for a mesenchymal marker. NKX2-1-positive CAT459 cells showed higher messenger RNA (mRNA) expression of some thyroid differentiation markers than NKX2-1-negative CAT458s cells, and NKX2-1 overexpression increased and/or induced their expression. IGF-I signaling was transduced in thyrotropin receptor (Tshr)-negative CAT458s and 459 cells, and NVP-ADW742 suppressed their proliferation. No tumors developed in mice after subcutaneous injection of CAT458s or 459 cells. The WGS analysis revealed the presence of missense mutations in the tumor suppressor genes such as Polk (encoding DNA polymerase kappa) and Tgfb1 (encoding transforming growth factor beta 1), while no mutations were found in the prominent thyroid cancer-related genes Braf, Trp53 (encoding p53), and Tert (encoding telomerase reverse transcriptase).

Conclusions:

Two mouse thyroid adenomatous nodule-derived cell lines with different thyroid differentiation marker expression were established. NKX2-1 induced partial differentiation of these cell lines. They lacked tumorigenicity and prominent gene mutations involved in thyroid cancer development, while missense mutations were found in some tumor suppressors as revealed by WGS. The CAT458s and 459 provide a new tool to further clarify the process of thyroid multistep carcinogenesis and differentiation.

Introduction

Many gene mutations have been identified in thyroid carcinogenesis. Among them, the Braf mutation is the most common and is thought to directly transform normal thyroid follicular cells to papillary thyroid cancer (PTC) without an adenoma intermediate step.¹ On the contrary, some types of thyroid cancers are known to arise from thyroid adenomas, leading to malignant transition from the benign adenoma, as proposed in the multistep carcinogenesis model.¹

Human thyroid cancer cell lines, including those derived from PTC, follicular thyroid cancer (FTC), and anaplastic thyroid cancer (ATC), have been extensively used to study thyroid cancer.² To investigate normal thyroid functions, the spontaneously immortalized normal rat thyroid cell lines, FRTL-5^3–5 and PCCL3,^5,6 and normal human thyroid cell lines immortalized with SV40, Nthy-ori 3-1⁷ and N-thy-Ts.1,⁸ have been used. However, thyroid adenoma cell lines have not been established and characterized. Thyroid adenoma cell lines as model benign thyroid epithelial cells would help unravel the mechanisms of thyroid carcinogenesis and thyroid differentiation.

Amitrole, 3-amino-1, 2, 4-triazole, widely used herbicide for nonfood cropland, was reported to inhibit thyroid peroxidase (TPO) and iodine uptake^9
–11 and induce tumors in rat thyroids.^12,13 In addition, amitrole induces thyroid adenomas, but not thyroid cancer in Nkx2-1 (fl/fl) mice.¹⁴

In the present study, cell lines were established from thyroid adenomatous nodules developed in mice fed amitrole-containing diet. Among the five cell lines established, two cell lines showed expression of epithelial and thyroid markers, suggesting that these cell lines originated from thyroid follicular epithelial cells, while the other three cell lines were considered as mesenchymal cells. The differentiation, proliferation, and tumorigenicity capacity of the two thyroid epithelial cell lines were characterized, and whole-genome sequencing (WGS) was carried out to identify gene mutations in these cell lines. The results suggest that they are consistent with benign thyroid cells with different degrees of thyroid differentiation. These cell lines could provide a novel tool to obtain a more complete understanding of thyroid carcinogenesis and differentiation.

Materials and Methods

Generation of thyroid adenomatous nodule-derived cell lines and cell cultures

Method for the generation of CAT (cells from amitrole-treated thyroids) cell lines and culture conditions for CAT and MLg cells purchased from ATCC are described in the Supplementary Methods.

Cultured cell analyses

The lentiviral vector expression system was used to establish CAT458s and 459 cells stably expressing tetracycline (Tet)-inducible rat NKX2-1 protein, as previously described.¹⁵ Methods for production of recombinant lentivirus, selection of infected cells, induction of NKX2-1, cell proliferation assay, CCK-8 assay, and the reagents used in the study are described in the Supplementary Methods.

Biochemical and molecular biological methods

Total RNA extraction, complementary DNA (cDNA) synthesis, and quantitative reverse transcription–polymerase chain reaction (qRT-PCR), and protein extraction and Western blotting are described in the Supplementary Methods. Primers and antibodies used in this study are listed in Supplementary Tables S1 and S2, respectively.

Animal procedures and cytological/histological analysis

All animal studies were performed in accordance with the Using Animals in Intramural Research Guidelines (National Institutes of Health Animal Research Advisory Committee, National Institutes of Health [NIH]) with approval of the Institutional Animal Care and Use Committee (No. 01-094, LM-077). Subcutaneous xenografting of CAT458s and 459 cells in immunocompromised nude mice and preparation of cells and mouse tissues for cytological and histological analyses were performed as described in the Supplementary Methods.

WGS and validation of gene mutations

Genomic DNA extraction, WGS, and validation were performed as described in the Supplementary Methods. Primers used for validation were designed based on the Ensembl 102 GRCm38 database.¹⁶ The sequences are listed in Supplementary Table S1. Forward primers used for validation were used as sequencing primers unless otherwise mentioned.

In silico amino acid mutation analysis

Predicted mutation impacts on stability were estimated using protein structure-^17
–19 and sequence^20

–23-based approaches. Various methods^24

–30 were employed to predict the association with a disease. The details are described in the Supplementary Methods.

Database analysis

The mutated gene lists by WGS were analyzed using Database for Annotation, Visualization and Integrated Discovery (DAVID) knowledgebase.^31,32 Protein alignment comparison between mouse and human was performed using Clustal Omega,^33,34 and the mutations in cancer patients were analyzed using the database Catalogue Of Somatic Mutations In Cancer (COSMIC).³⁵ The details are described in the Supplementary Methods and Supplementary Tables S6–S9.

Statistical analysis

Two-tailed Student's t-tests or Welch's t-tests were used for comparison of two samples. Results were considered significant at p < 0.05.

Results

Establishment of five cell lines from thyroid adenomatous nodules induced by amitrole in mice

Nkx2–1(fl/fl) mice were fed an amitrole-containing diet starting at 5 weeks of age to induce thyroid adenomas, as previously described.¹⁴ Mice developed thyroid tumors within 6–12 months post-initiation of the diet (Fig. 1A, B). The thyroid adenomatous nodules, initially characterized based on gross pathology, were resected, whole tumor tissue was dissociated, and the cells were subjected to continuous culture. The nodules were similar to adenomas found in the thyroids of amitrole-treated Nkx2-1(fl/fl);Tpo-Cre mice¹⁴ (Supplementary Fig. S1). The primary cells that continuously grew after five passages were considered as permanent cell lines, and five CAT lines, CAT411, 413, 427, 458, and 459, were successfully established from five individual tumors in individual mice (Fig. 1A).

FIG. 1.

CAT458/458s and 459 cells show epithelial morphology and epithelial marker expression. (A) Schematic representation of establishment of cell lines from amitrole-induced mouse thyroid adenomatous nodules. (B) Representative picture of developed thyroid adenomatous nodules. Yellow dashed lines: thyroid lobes; black dashed lines with arrows: thyroid adenomatous nodules. (C) Representative pictures of cell morphology of the established cell lines, CAT411, 413, 427, 458, 458s (a subline of CAT458), and 459. Scale bar, 200 μm. (D) Lysates (8 μg) of CATs, MLg, and mouse thyroid were subjected to Western blotting with indicated antibodies. CAT, cells from amitrole-treated thyroids.

CAT411, 413, and 427 cells were spindle-shaped cells (Fig. 1C), as was MLg, a mouse lung fibroblast cell line (Supplementary Fig. S2). On the contrary, CAT458, 458s, a subline of 458, and 459 showed cuboidal cell morphology (Fig. 1C), suggesting that they are epithelial cells. In fact, the expression of the epithelial markers, E-cadherin and Claudin-1, was found in CAT458, 458s, 459, and mouse thyroid, but not in 411, 413, 427, and MLg cells (Fig. 1D). In contrast, Vimentin, a mesenchymal marker, was not found in CAT458, 458s, and 459, but expressed in the other CATs, MLg, and in thyroid (Fig. 1D), consistent with cellular morphologies. These results suggest that CAT458/458s and 459 cells originated from epithelial cells, while the other CATs were from mesenchymal cells, presumably residing in the thyroid.

CAT458/458s and 459 cells originate from thyroid epithelial cells

To confirm that CAT458/458s and 459 epithelial cells originated from thyroids, the expression of thyroid transcription factors critical for thyroid development and function^36,37 was first examined. Nkx2-1, Pax8, and Foxe1 messenger RNAs (mRNAs) were found in CAT458/458s (Nkx2-1 not detected in CAT458 cells) and CAT459 cells, while Nkx2-1 mRNA was not detected in all other CATs and MLg cells (Fig. 2A). It was noted that NKX2-1 protein, a master regulator of thyroid development,^36,38 was expressed in CAT459 cells at comparable levels as found in mouse thyroid (Fig. 2B). On the contrary, PAX8 protein was detected in only mouse thyroid (Fig. 2B).

FIG. 2.

CAT458/458s and 459 cells express thyroid transcription factors and thyroid markers. (A, C) qRT-PCR analysis of the indicated mouse thyroid gene mRNAs using cDNAs from CATs, MLg, and mouse thyroid. For each gene, the CAT459 value was set as 1 except Tshr and Tpo in which the value of thyroid was set as 1. Data are presented as mean ± SD from two to four independent samples. Dot indicates a value obtained in triplicate from each sample (see the Supplementary Methods). *p < 0.05, **p < 0.01, ***p < 0.005 for comparison of CAT458s with CAT459 cell samples by two-tailed Student's t-tests or Welch's t-tests. (B) Lysates (8 μg) of CATs, MLg, and mouse thyroid were subjected to Western blotting with the indicated antibodies. The samples were the same as those used for Figure 1D. See GAPDH and β-actin in Figure 1D as loading controls. (D) H&E staining of CAT458s and 459 cells. Representative pictures are shown. Scale bar, 10 μm. cDNA, complementary DNA; H&E, Hematoxylin & Eosin; mRNA, messenger RNA; n.d., not detected; NS, not significant; qRT-PCR, quantitative reverse transcription–polymerase chain reaction; SD, standard deviation; Tpo, thyroid peroxidase; Tshr, thyrotropin receptor.

Next, mRNAs encoding thyroid markers³⁷ were quantified. Both Duox1 and Duox2, NADPH oxidases involved in thyroid hormone synthesis, were expressed in CAT458/458s and 459 cells (Fig. 2C). Notably, Duox1 mRNA was much higher in CAT458/458s and 459 cells than in the mouse thyroid (Fig. 2C) due to an unknown mechanism. However, Tg encoding thyroglobulin, a main precursor of thyroid hormones, and Slc5a5, encoding sodium iodide symporter (NIS), were barely expressed in CAT458/458s (Tg not detected in CAT458s cells) and CAT459 cells (Fig. 2C). Furthermore, no expression of thyrotropin receptor (Tshr) and Tpo mRNAs was detected in either CAT458/458s or 459 cells (Fig. 2C), as reported in human thyroid cancer cell lines.³⁹ Collectively, these results suggest that CAT458/458s and 459 cells originated from thyroid follicular epithelial cells.

NKX2-1-positive CAT459 cells showed higher expression of thyroid marker mRNAs including Foxe1, Duox1, Duox2, and Tg than the NKX2-1-negative CAT458s cells, except for a slightly lower expression of Slc5a5 mRNA (Fig. 2A, C). Consistent with this finding, Hematoxylin & Eosin staining revealed that CAT458s cells showed a higher degree of cellular atypia than CAT459 cells, and CAT459 cells had more cytoplasm than CAT458s cells (Fig. 2D). These results suggest that NKX2-1-positive CAT459 cells are more differentiated.

NKX2-1 induces partial thyroid differentiation in CAT458s and 459 cells

Next, to examine the effect of NKX2-1 on the expression of thyroid differentiation markers, CAT458s and 459 cells stably expressing Tet-inducible rat NKX2-1, named 458s- and 459-TNKX cells, respectively, were established (Fig. 3A). Overexpression of NKX2-1 under doxycycline (DOX) was confirmed in both cell lines (Fig. 3B, C). Enlargement of cells was found 3 days after DOX treatment in both cell lines (Fig. 3D), suggesting that NKX2-1 may induce thyroid differentiation. Furthermore, the expression of Pax8 and thyroid markers including Duox1, Duox2, Tg, and Tshr mRNAs was higher in both 458s- and 459-TNKX cells with NKX2-1 overexpression (Fig. 3E). On the contrary, Foxe1, Slc5a5, and Tpo mRNAs were not induced by NKX2-1 (Supplementary Fig. S3). These results suggest that NKX2-1 can induce and further enhance partial thyroid differentiation in CAT458s and 459 cells.

FIG. 3.

NKX2-1 induces thyroid marker expression in CAT458s and 459 cells. (A) Schematic representation of establishment of 458s- and 459-TNKX cells stably expressing Tet-inducible rat NKX2-1. (B, E) qRT-PCR for Nkx2-1 and other thyroid-specific gene mRNAs using cDNAs from 458s- and 459-TNKX cells treated with DMSO or DOX for 3 days. For Pax8 and Duox1, cell line names are as those listed for Tg and Tshr; 458s-TNKX (left), 459-TNKX (right). For each gene, the value of 459-TNKX cells treated with DMSO was set as 1 except Tshr in which the value of 458s-TNKX cells treated with DOX was set as 1. Data are presented as mean ± SD from three independent samples. Dot indicates a value obtained in triplicate from each sample (see the Supplementary Methods). *p < 0.05, **p < 0.01, ***p < 0.005 for comparison of DMSO-treated samples with DOX-treated samples by two-tailed Student's t-tests or Welch's t-tests. (C) Lysates of 458s- and 459-TNKX cells treated with DMSO or DOX for 3 days were subjected to Western blotting with indicated antibodies. (D) Representative pictures of cell morphology of 458s- and 459-TNKX cells treated with DMSO or DOX for 3 days. Scale bar, 100 μm. DMSO, dimethyl sulfoxide; DOX, doxycycline; Tet, tetracycline.

Insulin-like growth factor-I signaling is partly responsible for the growth of Tshr-negative CAT458s and 459 cells

It is known that thyrotropin (TSH) signals through the TSHR and is essential for growth and differentiation of thyroid epithelium.⁴⁰ However, Tshr expression is deficient in mouse thyroid adenomatous nodule-derived CAT458s and 459 cells (Fig. 2C), as reported for human thyroid cancer cells.³⁹ Studies using FRTL-5 cells demonstrated that insulin-like growth factor-I (IGF-I) works synergistically with TSH in thyroid follicular epithelium.^41,42 Hence, studies were conducted to determine whether IGF-I signaling is transduced independent of TSHR and is required for the growth of Tshr-negative CAT458s and 459 cells.

Igf1r (encoding IGF-I receptor) mRNA and IGF-I receptor β (IGF-IRβ) protein expression was found in CAT458s and 459 cells (Supplementary Fig. S4A and Fig. 4A), while Igf1 mRNA levels were barely detected (Supplementary Fig. S4B). Induction of phosphorylation of downstream targets of IGF-I signal, AKT and S6 ribosomal protein,⁴⁰ was confirmed in both CAT458s and 459 cells after IGF-I treatment (Fig. 4A). Proliferation of CAT459 cells was promoted by IGF-I; however, IGF-I treatment failed to promote the growth of CAT458s cells (Fig. 4B and Supplementary Fig. S5A). Whether endogenous IGF-IR-mediated signaling affects the growth of these cells was next examined using NVP-ADW742 (ADW742), a selective IGF-IR inhibitor.⁴³ ADW742 inhibited the proliferation of both cell lines (Fig. 4C and Supplementary Fig. S5B). These results suggest that IGF-I signaling is transduced and at least partly responsible for the growth of Tshr-negative CAT458s and 459 thyroid cells.

FIG. 4.

IGF-I signaling is partly responsible for the proliferation of Tshr-negative CAT458s and 459 cells. (A) Lysates of CAT458s and 459 cells treated with PBS or IGF-I (100 ng/mL) for 1 hour under 2% FBS were subjected to Western blotting with indicated antibodies. (B) Representative pictures of CAT458s and 459 cells treated with PBS or IGF-I (100 ng/mL) for 3 days under 2% FBS. Scale bar, 200 μm (upper). Cell proliferation assay results (lower). Data are presented as mean ± SD from three biological replicates. **p < 0.01 by two-tailed Student's t-tests. (C) Representative pictures of CAT458s and 459 cells treated with DMSO or ADW742 for 3 days under 2% FBS. Scale bar, 200 μm (upper). Cell proliferation assay results (lower). Data are presented as mean ± SD from three biological replicates. ***p < 0.005 by two-tailed Student's t-tests. FBS, fetal bovine serum; IGF-I, insulin-like growth factor-I; PBS, phosphate-buffered saline.

CAT458s and 459 cells are not tumorigenic in vivo

Since whole tumor tissues were used to establish CAT458s and 459 cell lines, there was a possibility that a small number of cancer cells that resided in the tumor might have been selected and formed the resultant cell lines. Whether CAT458s and 459 cells are tumorigenic in vivo was therefore examined by subcutaneous injection of these cells into immunocompromised nude mice. Tumors did not develop in mice when examined 4–8 weeks after injection (Table 1). Indeed, viable injected cells were not observed in subcutaneous tissues 6 days after injection (Supplementary Fig. S6). These results suggest that CAT458s and 459 cells are consistent with benign thyroid cells, different from thyroid cancer cells.

Table 1.

Tumorigenicity of CAT458s and 459 Cells in Nude Mice

	CAT458s	CAT459
3 × 10⁶ Cells	0/5	0^a/5
1 × 10⁶ Cells	0/6	0/6

1 or 3 × 10⁶ Cells of CAT458s or 459 cells were subcutaneously implanted into nude mice. Incidence of developed growing subcutaneous tumors 4–8 weeks after implantation was evaluated by gross examination (shown as developed growing tumor/mouse).

A tiny fibrous nodule (containing possible epithelial cells, examined by H&E staining, figure not shown) remained in one mouse even 8 weeks after injection of 3 × 10⁶ of CAT459 cells.

CAT, cells from amitrole-treated thyroids; H&E, Hematoxylin & Eosin.

WGS reveals missense gene mutations in tumor suppressor genes in CAT458s and 459 cells

Finally, WGS was carried out to identify somatic gene mutations present in CAT458s and 459 cells. Analysis for structural variations (SV) was first carried out, which showed that both cells have similar relative prevalence among the variations, although CAT459 cells have more SV than 458s cells (Supplementary Fig. S7). Next, missense gene mutations of CAT458s and 459 cells were examined. Since CAT cell lines were derived from Nkx2-1(fl/fl) mice on a mixed genetic background,¹⁴ all the normal variants reported in Ensembl 102 and mgp v5 dbSNP142 were excluded from the mutation lists of CAT458s and 459 cells (Fig. 5A). This resulted in the identification of 65 CAT458s-specifc and 53 CAT459-specifc mutated genes (Fig. 5B and Supplementary Tables S3 and S4).

FIG. 5.

WGS reveals missense mutations in tumor suppressor genes in CAT458s and 459 cells. (A) Schematic representation of extraction of missense gene mutations in CAT458s and 459 cells demonstrated by WGS. (B) Venn diagram of missense gene numbers in CAT458s and 459 cells. (C) Venn diagram indicating missense mutated genes in “Cancer-related genes” and “Thyroid/Thyroid hormone-related genes” (see the Supplementary Methods) in CAT458s and 459 cells. Known or possible tumor suppressor genes are shown in red. (D) Chromatograms of genome sequences around the identified missense mutation in Atm, Polk, and Tgfb1, Thrb, and Med12 in CAT458s and 459 cells. WGS, whole-genome sequencing.

From each 65 and 53 pools of genes for the CAT458 and CAT459 cells, respectively, “Cancer-related genes” and “Thyroid/thyroid hormone-related genes” were extracted (see the Supplementary Methods). Among “Cancer-related genes,” prominent gene mutations known to be involved in thyroid cancer, such as Braf, Trp53 (encoding p53), and Tert (encoding telomerase reverse transcriptase),^1,44,45 were not identified, suggesting that CAT458s and 459 cells are consistent with benign thyroid adenoma cells. Indeed, most of human cancer cell lines possess missense gene mutations in these genes.^46,47 Instead, missense mutations in tumor suppressor genes Atm (encoding ataxia telangiectasia mutated)⁴⁸ and Polk (encoding DNA polymerase kappa)⁴⁹ in CAT458s, and Tgfb1 (encoding transforming growth factor beta 1)⁵⁰ in CAT459 were found (Fig. 5C).

Possible tumor suppressors Thrb (encoding thyroid hormone receptor β)^51,52 and Med12 (encoding mediator complex subunit 12)^53
–55 were additionally found in the CAT458s- and 459-specific mutated gene lists, respectively, in “Thyroid/thyroid hormone-related genes” (Fig. 5C). Therefore, these five genes were selected, and their mutations were validated by PCR amplification followed by sequencing (Fig. 5D and see the mutations details in Supplementary Tables S3 and S4). It was also confirmed that these mutations are absent in CAT411, 413, and 427 cells (Supplementary Fig. S8).

In silico analysis of these mutations in tumor suppressors was carried out to predict protein stability and association with a disease. Predicted stability change by the mutations showed a tendency toward a slight protein instability, suggesting that the mutations might lead to loss of function (Table 2). Meta-Single Nucleotide Polymorphism results with various frequently used methods^24

–28 showed the possibility that the mutations except for Atm are associated with a disease, especially Thrb with a higher likelihood (Table 2). Predictions with PolyPhen-2²⁹ also suggested that these mutations may cause disease (Table 2). In addition, predictions of the corresponding mutations in human with MutationTaster2021³⁰ also suggested that the mutations in THRB and TGFB1 are deleterious (Supplementary Table S5). These predictions are consistent with the possibility that the mutations found in CAT458s and 459 cells contribute to thyroid adenoma development.

Table 2.

In Silico Analysis of Amino Acid Mutations in Tumor Suppressor Genes in CAT458s and 459 Cells

Discussion

Five cell lines were established from amitrole-induced mouse thyroid adenomatous nodules, among which the CAT458/458s and 459 cell lines possess thyroid epithelial properties. CAT459 cells positive for thyroid master transcription factor, NKX2-1, showed higher expression of thyroid markers than CAT458s cells. In addition, NKX2-1 overexpression induced the expression of some thyroid markers in both cell lines, suggesting that NKX2-1 can induce partial differentiation in these cells, as was reported in mouse embryonic stem (ES) cells.³⁶ It should be noted that Foxe1 expression is much higher in CAT459 cells than in CAT458s cells. FOXE1 is another essential transcription factor for human thyroid development,^37,56 and FOXE1 expression is induced by NKX2-1 and PAX8 in mouse ES cells.³⁶ However, in the current study, NKX2-1 overexpression did not induce Foxe1 expression in either cells, suggesting that Foxe1 expression may be regulated independent of NKX2-1. Further studies are necessary to establish the function and regulation of FOXE1 in adult mouse thyroid cell differentiation.

High expression of TSHR, TPO, and NIS is required for thyroid hormone synthesis through TSHR signaling in thyroids. The rat normal thyroid cell lines FRTL-5 and PCCL-3 maintain thyroglobulin production and DNA synthesis in response to TSH under two-dimensional culltures,^3,5 and thus, they are commonly used for normal thyroid function research.

On the contrary, CAT458s and 459 cells barely express or lack Tg, Scl5a5, Tpo, and Tshr mRNA expression. Similar findings were reported in studies using various types of human thyroid cancer cell lines and transformed rat thyroid cell lines.^5,39,47 These findings suggest that CAT458s and 459 cells possess characteristics close to cancer cells that can grow in the absence of TSH. Indeed, bovine TSH (bTSH) treatment did not induce levels of Tg, Slc5a5, and Tpo mRNAs (Supplementary Fig. S9), known target genes associated with TSH signaling.⁵⁷ Instead, IGF-I signaling is partly responsible for the growth of Tshr-negative CAT458s and 459 cells. Igf1 mRNA expression is abundant in CAT411, 413, and 427 mesenchymal cell lines, suggesting the possibility that thyroid adenoma-associated fibroblasts could be a source of IGF-I.

Mouse and human thyroid cancer cell lines are known to develop tumor after subcutaneous, intravenous, and orthotopic injection into immunocompromised mice.^47,58,59 When CAT458s and 459 cells were subcutaneously injected into nude mice, small subcutaneous nodules were observed in some mice 6 days after injection; however, only fibrosis and/or inflammatory cells with or without nonviable injected cells were found histologically. Consistent with this, subcutaneous nodules disappeared by 4–8 weeks after injection, except for one tiny fibrous nodule without any growing tumors, suggesting that CAT458s and 459 cells are most likely different from cancer cells. Since CAT458s and 459 cells were derived from thyroid adenomas induced by amitrole treatment, which is known to inhibit iodide uptake as well as TPO, these cells may be useful for the study of iodine deficiency-induced thyroid adenomatous nodules in humans.⁶⁰ However, amitrole may have other unknown effects on the thyroid through inhibition of TPO activity or unknown mechanisms, which might limit the use of these cell lines to understand the nature and/or mechanism of thyroid carcinogenesis.

SV are generally thought to cause malignant transformation resulting in cancer, rather than benign adenoma. Thus, missense gene mutations of CAT458s and 459 cells were assessed to obtain insights into the mutations that occur in the early stages of multistep thyroid carcinogenesis. Besides 65 CAT458s-specific and 53 CAT459-specific mutated genes, 123 commonly mutated genes were obtained. Among them, Actl6b (actin-like 6b), Ptprj (protein tyrosine phosphatase receptor type J), Rxrb (retinoid X receptor beta), and Tnxb (tenascin XB) were categorized as “Cancer-related genes” and Atf6b (activating transcription factor 6 beta) and Rxrb as “Thyroid/thyroid hormone-related genes.” However, it is unlikely that the two independent cell lines share mutated genes at the same locus. It is most likely that they are variants derived from the mouse genetic background that were not listed in the public databases used. Indeed, the presence of heterozygous and/or homozygous mutations in Rxrb ⁶¹ and tumor suppressor genes Actl6b ⁶² and Ptprj ⁶³ was confirmed in all five CAT cell lines (data not shown). Therefore, only CAT458s- or 459-spcific missense mutations were studied.

Tshr and Gnas1 (encoding Gsα) point mutations were reported in autonomously functioning thyroid nodules.^1,60,64,65 However, these gene mutations were not found in CAT458s and 459 cells by WGS. It is likely that CAT458s and 459 cells were derived from follicular thyroid adenomas, which could be precancerous. The mutated amino acids in coding proteins due to missense mutations of tumor suppressor genes such as Polk and Thrb in CAT458s and Tgfb1 and Med12 in CAT459 cells are conserved between mice and humans (Supplementary Fig. S10A).

COSMIC database³⁵ analysis for the presence of missense mutations of these four genes in human cancer patients identified the same G845E mutation in POLK in one glioma patient,⁶⁶ G845V in POLK in one lung squamous cell cancer patient,⁶⁷ and S83I in TGFB1 in one enodmetrioid carcinoma patient in The Cancer Genome Atlas (TCGA) study (Supplementary Fig. S10B). These findings suggest that these mutations might be involved in early carcinogenesis. On the contrary, the corresponding mutations in THRB and MED12 were not found in the COSMIC database. The Thrb mutation found in CAT458s cells is located in the DNA-binding domain, different from that in the ligand-binding domain observed in human thyroid retinal hormone resistance syndrome (RTH) patients, which caused spontaneous FTC when introduced in mice.^51,52,68

The D878N mutation in MED12 is located in exon 19 (out of 45, middle region). On the contrary, the G44 located in exon 2 of MED12 is one of the hot spot mutation sites observed in some types of malignant tumors such as fatal forms of non-ATC.^53
–55 The G44 is involved in binding to Cyclin C/Cyclin-dependent kinase (CDK) 8, and the mutations at this site were reported to abrogate the activity of CDK8, and the dysregulation of CDK8 possibly contributes to malignancy.^53
–55 These findings suggest that the mutations in Thrb and Med12 we identified may not be potent enough to transform benign thyroid cells. Further studies would be required for clarification of thyroid multistep carcinogenesis.

In conclusion, two mouse thyroid adenomatous nodule-derived cell lines with different thyroid differentiation marker expression were established that possess missense mutations in tumor suppressor genes. However, they lack tumorigenicity in immunocompromised mice, suggesting that they are consistent with benign adenoma cells. CAT458s and 459 cells would provide a tool to further understand the nature and/or mechanism of thyroid carcinogenesis and differentiation.

Footnotes

Acknowledgments

Mice under amitrole diet were maintained at the animal facility run by the National Cancer Institute (NCI) Laboratory Animal Sciences Program. WGS and Sanger sequencing were conducted at Sequencing Facility and the Center for Cancer Research (CCR) Genomics Core, respectively, at the NCI, NIH.

Authors' Contributions

Y.-T.S.: Formal analysis, investigation, writing—original draft, writing—review and editing. N.H.: Investigation, writing—review and editing. J.M.W.: Formal analysis and writing—review and editing. H.L.: Formal analysis and writing—review and editing. R.E.C.: Formal analysis and writing—review and editing. M.P.L.: Formal analysis and writing—review and editing. S.K.: Conceptualization, supervision, writing—review and editing.

Author Disclosure Statement

The authors declare that they have no conflict of interest.

Funding Information

This work was supported by an Intramural Research Program of the National Cancer Institute, Center for Cancer Research (ZIA BC 005522).

Supplementary Material

Supplementary Methods

Supplementary Figure S1

Supplementary Figure S2

Supplementary Figure S3

Supplementary Figure S4

Supplementary Figure S5

Supplementary Figure S6

Supplementary Figure S7

Supplementary Figure S8

Supplementary Figure S9

Supplementary Figure S10

Supplementary Table S1

Supplementary Table S2

Supplementary Table S3

Supplementary Table S4

Supplementary Table S5

Supplementary Table S6

Supplementary Table S7

Supplementary Table S8

Supplementary Table S9

References

Kondo

, Ezzat

, Asa

. Pathogenetic mechanisms in thyroid follicular-cell neoplasia. Nat Rev Cancer, 2006; 6(4):292–306; doi: 10.1038/nrc1836

Schweppe

, Klopper

, Korch

, et al. Deoxyribonucleic acid profiling analysis of 40 human thyroid cancer cell lines reveals cross-contamination resulting in cell line redundancy and misidentification. J Clin Endocrinol Metab, 2008; 93(11):4331–4341; doi: 10.1210/jc.2008-1102

Ambesi-Impiombato

, Parks

, Coon

. Culture of hormone-dependent functional epithelial cells from rat thyroids. Proc Natl Acad Sci U S A, 1980; 77(6):3455–3459; doi: 10.1073/pnas.77.6.3455

Bidey

, Lambert

, Robertson

. Thyroid cell growth, differentiation and function in the FRTL-5cell line: A survey. J Endocrinol, 1988; 119(3):365–376; doi: 10.1677/joe.0.1190365

Berlingieri

, Akamizu

, Fusco

, et al. Thyrotropin receptor gene expression in oncogene-transfected rat thyroid cells: Correlation between transformation, loss of thyrotropin-dependent growth, and loss of thyrotropin receptor gene expression. Biochem Biophys Res Commun, 1990; 173(1):172–178; doi: 10.1016/s0006-291x(05)81037-x

Fusco

, Berlingieri

, Di Fiore

, et al. One- and two-step transformations of rat thyroid epithelial cells by retroviral oncogenes. Mol Cell Biol, 1987; 7(9):3365–3370; doi: 10.1128/mcb.7.9.3365-3370.1987

Gregory Powell

, Wang

, Allard

, et al. The PAX8/PPARgamma fusion oncoprotein transforms immortalized human thyrocytes through a mechanism probably involving wild-type PPARgamma inhibition. Oncogene, 2004; 23(20):3634–3641; doi: 10.1038/sj.onc.1207399

Wynford-Thomas

, Bond

, Wyllie

, et al. Conditional immortalization of human thyroid epithelial cells: A tool for analysis of oncogene action. Mol Cell Biol, 1990; 10(10):5365–5377; doi: 10.1128/mcb.10.10.5365-5377.1990

Pan

, Sun

, Zhang

. The effects of amitrole on thyroglobulin and iodide uptake in FRTL-5 cells. Toxicol Ind Health, 2011; 27(2):187–192; doi: 10.1177/0748233710386405

10.

Crofton

. Thyroid disrupting chemicals: Mechanisms and mixtures. Int J Androl, 2008; 31(2):209–223; doi: 10.1111/j.1365-2605.2007.00857.x

11.

Alexander

. Antithyroid action of 3-amino-1,2,4-triazole. J Biol Chem, 1959; 234(1):148–150; doi: 10.1016/s0021-9258(18)70353-0

12.

Jukes

, Shaffer

. Antithyroid effects of aminotriazole. Science, 1960; 132(3422):296–297; doi: 10.1126/science.132.3422.296

13.

Tsuda

, Hananouchi

, Tatematsu

, et al. Tumorigenic effect of 3-amino-1H-1,2,4-triazole on rat thyroid. J Natl Cancer Inst, 1976; 57(4):861–864; doi: 10.1093/jnci/57.4.861

14.

Hoshi

, Hoshi

, Okamoto

, et al. Role of NKX2-1 in N-bis(2-hydroxypropyl)-nitrosamine-induced thyroid adenoma in mice. Carcinogenesis, 2009; 30(9):1614–1619; doi: 10.1093/carcin/bgp167

15.

Iwadate

, Takizawa

, Shirai

, et al. An in vivo model for thyroid regeneration and folliculogenesis. Lab Invest, 2018; 98(9):1126–1132; doi: 10.1038/s41374-018-0068-x

16.

Yates

, Achuthan

, Akanni

, et al. Ensembl 2020. Nucleic Acids Res, 2020; 48(D1):D682–D688; doi: 10.1093/nar/gkz966

17.

Laimer

, Hofer

, Fritz

, et al. MAESTRO—Multi agent stability prediction upon point mutations. BMC Bioinformatics, 2015; 16:116; doi: 10.1186/s12859-015-0548-6

18.

Montanucci

, Capriotti

, Birolo

, et al. DDGun: An untrained predictor of protein stability changes upon amino acid variants. Nucleic Acids Res, 2022; 50(W1):W222–W227; doi: 10.1093/nar/gkac325

19.

Chen

, Lu

, Zhang

, et al. PremPS: Predicting the impact of missense mutations on protein stability. PLoS Comput Biol, 2020; 16(12):e1008543; doi: 10.1371/journal.pcbi.1008543

20.

Zhang

, DeLuca

, Keskin

, et al. MULTIPRED2: A computational system for large-scale identification of peptides predicted to bind to HLA supertypes and alleles. J Immunol Methods, 2011; 374(1–2):53–61; doi: 10.1016/j.jim.2010.11.009

21.

Cheng

, Randall

, Baldi

. Prediction of protein stability changes for single-site mutations using support vector machines. Proteins, 2006; 62(4):1125–1132; doi: 10.1002/prot.20810

22.

Capriotti

, Fariselli

, Casadio

. I-Mutant2.0: Predicting stability changes upon mutation from the protein sequence or structure. Nucleic Acids Res, 2005; 33(Web Server issue):W306–W310; doi: 10.1093/nar/gki375

23.

, Panday

, Alexov

. SAAFEC-SEQ: A sequence-based method for predicting the effect of single point mutations on protein thermodynamic stability. Int J Mol Sci, 2021; 22(2):606; doi: 10.3390/ijms22020606

24.

Capriotti

, Altman

, Bromberg

. Collective judgment predicts disease-associated single nucleotide variants. BMC Genomics, 2013; 14(Suppl 3):S2; doi: 10.1186/1471-2164-14-S3-S2

25.

, Ebert

, Muruganujan

, et al. PANTHER version 16: A revised family classification, tree-based classification tool, enhancer regions and extensive API. Nucleic Acids Res, 2021; 49(D1):D394–D403; doi: 10.1093/nar/gkaa1106

26.

Capriotti

, Fariselli

. PhD-SNPg: Updating a webserver and lightweight tool for scoring nucleotide variants. Nucleic Acids Res, 2023; 51(W1):W451–W458; doi: 10.1093/nar/gkad455

27.

Vaser

, Adusumalli

, Leng

, et al. SIFT missense predictions for genomes. Nat Protoc, 2016; 11(1):1–9; doi: 10.1038/nprot.2015.123

28.

Bromberg

, Yachdav

, Rost

. SNAP predicts effect of mutations on protein function. Bioinformatics, 2008; 24(20):2397–2398; doi: 10.1093/bioinformatics/btn435

29.

Adzhubei

, Schmidt

, Peshkin

, et al. A method and server for predicting damaging missense mutations. Nat Methods, 2010; 7(4):248–249; doi: 10.1038/nmeth0410-248

30.

Steinhaus

, Proft

, Schuelke

, et al. MutationTaster2021. Nucleic Acids Res, 2021; 49(W1):W446–W451; doi: 10.1093/nar/gkab266

31.

Huang da

, Sherman

, Zheng

, et al. Extracting biological meaning from large gene lists with DAVID. Curr Protoc Bioinformatics, 2009;Chapter 13:Unit 13.11; doi: 10.1002/0471250953.bi1311s27

32.

Sherman

, Hao

, Qiu

, et al. DAVID: A web server for functional enrichment analysis and functional annotation of gene lists (2021 update). Nucleic Acids Res, 2022; 50(W1):W216–W221; doi: 10.1093/nar/gkac194

33.

Sievers

, Higgins

. The Clustal Omega multiple alignment package. Methods Mol Biol, 2021; 2231:3–16; doi: 10.1007/978-1-0716-1036-7_1

34.

Madeira

, Pearce

, Tivey

ARN

, et al. Search and sequence analysis tools services from EMBL-EBI in 2022. Nucleic Acids Res, 2022; 50(W1):W276–W279; doi: 10.1093/nar/gkac240

35.

Tate

, Bamford

, Jubb

, et al. COSMIC: The Catalogue of Somatic Mutations in Cancer. Nucleic Acids Res, 2019; 47(D1):D941–D947; doi: 10.1093/nar/gky1015

36.

Antonica

, Kasprzyk

, Opitz

, et al. Generation of functional thyroid from embryonic stem cells. Nature, 2012; 491(7422):66–71; doi: 10.1038/nature11525

37.

Fernandez

, Lopez-Marquez

, Santisteban

. Thyroid transcription factors in development, differentiation and disease. Nat Rev Endocrinol, 2015; 11(1):29–42; doi: 10.1038/nrendo.2014.186

38.

Kimura

, Hara

, Pineau

, et al. The T/ebp null mouse: Thyroid-specific enhancer-binding protein is essential for the organogenesis of the thyroid, lung, ventral forebrain, and pituitary. Genes Dev, 1996; 10(1):60–69; doi: 10.1101/gad.10.1.60

39.

van Staveren

, Solis

, Delys

, et al. Human thyroid tumor cell lines derived from different tumor types present a common dedifferentiated phenotype. Cancer Res, 2007; 67(17):8113–8120; doi: 10.1158/0008-5472.CAN-06-4026

40.

Davies

, Andersen

, Latif

, et al. Graves' disease. Nat Rev Dis Primers, 2020; 6(1):52; doi: 10.1038/s41572-020-0184-y

41.

Fukushima

, Yoshihara

, Furuta

, et al. Nedd4-induced monoubiquitination of IRS-2 enhances IGF signalling and mitogenic activity. Nat Commun, 2015; 6:6780; doi: 10.1038/ncomms7780

42.

Hakuno

, Takahashi

. IGF1 receptor signaling pathways. J Mol Endocrinol, 2018; 61(1):T69–T86; doi: 10.1530/JME-17-0311

43.

Mitsiades

, Mitsiades

, McMullan

, et al. Inhibition of the insulin-like growth factor receptor-1 tyrosine kinase activity as a therapeutic strategy for multiple myeloma, other hematologic malignancies, and solid tumors. Cancer Cell, 2004; 5(3):221–230; doi: 10.1016/s1535-6108(04)00050-9

44.

Xing

. Molecular pathogenesis and mechanisms of thyroid cancer. Nat Rev Cancer, 2013; 13(3):184–199; doi: 10.1038/nrc3431

45.

Pozdeyev

, Gay

, Sokol

, et al. Genetic analysis of 779 advanced differentiated and anaplastic thyroid cancers. Clin Cancer Res, 2018; 24(13):3059–3068; doi: 10.1158/1078-0432.CCR-18-0373

46.

Saiselet

, Floor

, Tarabichi

, et al. Thyroid cancer cell lines: An overview. Front Endocrinol (Lausanne), 2012; 3:133; doi: 10.3389/fendo.2012.00133

47.

Henderson

, Ahn

, Ryu

, et al. Development and characterization of six new human papillary thyroid carcinoma cell lines. J Clin Endocrinol Metab, 2015; 100(2):E243–E252; doi: 10.1210/jc.2014-2624

48.

Renwick

, Thompson

, Seal

, et al. ATM mutations that cause ataxia-telangiectasia are breast cancer susceptibility alleles. Nat Genet, 2006; 38(8):873–875; doi: 10.1038/ng1837

49.

Yadav

, Mukhopadhyay

, Anbalagan

, et al. Somatic mutations in catalytic core of POLK reported in prostate cancer alter translesion DNA synthesis. Hum Mutat, 2015; 36(9):873–880; doi: 10.1002/humu.22820

50.

Colak

, Ten Dijke

. Targeting TGF-beta signaling in cancer. Trends Cancer, 2017; 3(1):56–71; doi: 10.1016/j.trecan.2016.11.008

51.

Kato

, Ying

, Willingham

, et al. A tumor suppressor role for thyroid hormone beta receptor in a mouse model of thyroid carcinogenesis. Endocrinology, 2004; 145(10):4430–4438; doi: 10.1210/en.2004-0612

52.

Kaneshige

, Kaneshige

, Zhu

, et al. Mice with a targeted mutation in the thyroid hormone beta receptor gene exhibit impaired growth and resistance to thyroid hormone. Proc Natl Acad Sci U S A, 2000; 97(24):13209–13214; doi: 10.1073/pnas.230285997

53.

Ibrahimpasic

, Xu

, Landa

, et al. Genomic alterations in fatal forms of non-anaplastic thyroid cancer: Identification of MED12 and RBM10 as novel thyroid cancer genes associated with tumor virulence. Clin Cancer Res, 2017; 23(19):5970–5980; doi: 10.1158/1078-0432.CCR-17-1183

54.

Gonzalez

, Akula

, Burleson

. The role of mediator subunit 12 in tumorigenesis and cancer therapeutics. Oncol Lett, 2022; 23(3):74; doi: 10.3892/ol.2022.13194

55.

Zhang

, O'Regan

, Xu

. The emerging role of mediator complex subunit 12 in tumorigenesis and response to chemotherapeutics. Cancer, 2020; 126(5):939–948; doi: 10.1002/cncr.32672

56.

Clifton-Bligh

, Wentworth

, Heinz

, et al. Mutation of the gene encoding human TTF-2 associated with thyroid agenesis, cleft palate and choanal atresia. Nat Genet, 1998; 19(4):399–401; doi: 10.1038/1294

57.

Morillo-Bernal

, Fernandez-Santos

, De Miguel

, et al. Ghrelin potentiates TSH-induced expression of the thyroid tissue-specific genes thyroglobulin, thyroperoxidase and sodium-iodine symporter, in rat PC-Cl3 Cells. Peptides, 2011; 32(11):2333–2339; doi: 10.1016/j.peptides.2011.09.013

58.

Vanden Borre

, McFadden

, Gunda

, et al. The next generation of orthotopic thyroid cancer models: Immunocompetent orthotopic mouse models of BRAF V600E-positive papillary and anaplastic thyroid carcinoma. Thyroid, 2014; 24(4):705–714; doi: 10.1089/thy.2013.0483

59.

Zhang

, Gaskins

, Yu

, et al. An in vivo mouse model of metastatic human thyroid cancer. Thyroid, 2014; 24(4):695–704; doi: 10.1089/thy.2013.0149

60.

Paschke

. Molecular pathogenesis of nodular goiter. Langenbecks Arch Surg, 2011; 396(8):1127–1136; doi: 10.1007/s00423-011-0788-5

61.

Gauchotte

, Lacomme

, Brochin

, et al. Retinoid acid receptor expression is helpful to distinguish between adenoma and well-differentiated carcinoma in the thyroid. Virchows Arch, 2013; 462(6):619–632; doi: 10.1007/s00428-013-1419-z

62.

Revill

, Wang

, Lachenmayer

, et al. Genome-wide methylation analysis and epigenetic unmasking identify tumor suppressor genes in hepatocellular carcinoma. Gastroenterology, 2013; 145(6):1424.e1–e25–1435.e1–e25; doi: 10.1053/j.gastro.2013.08.055

63.

Iuliano

, Le Pera

, Cristofaro

, et al. The tyrosine phosphatase PTPRJ/DEP-1 genotype affects thyroid carcinogenesis. Oncogene, 2004; 23(52):8432–8438; doi: 10.1038/sj.onc.1207766

64.

Parma

, Duprez

, Van Sande

, et al. Somatic mutations in the thyrotropin receptor gene cause hyperfunctioning thyroid adenomas. Nature, 1993; 365(6447):649–651; doi: 10.1038/365649a0

65.

Trulzsch

, Krohn

, Wonerow

, et al. Detection of thyroid-stimulating hormone receptor and Gsalpha mutations: In 75 toxic thyroid nodules by denaturing gradient gel electrophoresis. J Mol Med (Berl), 2001; 78(12):684–691; doi: 10.1007/s001090000170

66.

Lee

, Wang

, Sa

, et al. Spatiotemporal genomic architecture informs precision oncology in glioblastoma. Nat Genet, 2017; 49(4):594–599; doi: 10.1038/ng.3806

67.

Xiong

, Li

, et al. Exome sequencing identifies MXRA5 as a novel cancer gene frequently mutated in non-small cell lung carcinoma from Chinese patients. Carcinogenesis, 2012; 33(9):1797–1805; doi: 10.1093/carcin/bgs210

68.

Parrilla

, Mixson

, McPherson

, et al. Characterization of seven novel mutations of the c-erbA beta gene in unrelated kindreds with generalized thyroid hormone resistance. Evidence for two “hot spot” regions of the ligand binding domain. J Clin Invest, 1991; 88(6):2123–2130; doi: 10.1172/JCI115542