|Home | About | Journals | Submit | Contact Us | Français|
NKX2-1 is a homeodomain transcription factor that is critical for genesis of the thyroid and transcription of the thyroid-specific genes. Nkx2-1-thyroid-conditional hypomorphic mice were previously developed in which Nkx2-1 gene expression is lost in 50% of the thyroid cells. Using this mouse line as compared with wild-type and Nkx2-1 heterozygous mice, a thyroid carcinogenesis study was carried out using the genotoxic carcinogen N-bis(2-hydroxypropyl)-nitrosamine (DHPN), followed by sulfadimethoxine (SDM) or the non-genotoxic carcinogen amitrole (3-amino-1,2,4-triazole). A significantly higher incidence of adenomas was obtained in Nkx2-1-thyroid-conditional hypomorphic mice as compared with the other two groups of mice only when they were treated with DHPN + SDM, but not amitrole. A bromodeoxyuridine incorporation study revealed that thyroids of the Nkx2-1-thyroid-conditional hypomorphic mice had >2-fold higher constitutive cell proliferation rate than the other two groups of mice, suggesting that this may be at least partially responsible for the increased incidence of adenoma in this mouse line after genotoxic carcinogen exposure. Thus, NKX2-1 may function to control the proliferation of thyroid follicular cells following damage by a genotoxic carcinogen.
Chemically induced thyroid follicular cell tumors in rodents are thought to arise from two different modes of action, mutagenic and anti-thyroid or a combination of the two (1). N-bis(2-hydroxypropyl)-nitrosamine (DHPN) is a genotoxic mutagen that initiates carcinogenesis in target organs such as lung, thyroid, kidneys and liver (2–4). Many studies were carried out to examine the effect of anti-thyroidal agents on DHPN-induced thyroid carcinogenesis, including sulfadimethoxine (SDM) (5–7), xylazine hydrochloride (8), propylthiouracil (7) and phenobarbital (9). The first three chemicals exhibit anti-thyroidal activity by inhibiting thyroid hormone synthesis through inhibition of iodide uptake and/or thyroid peroxidase (TPO) activity. This leads to lower serum thyroid hormone levels and the ensuing elevation of serum thyrotropin or thyroid-stimulating hormone (TSH), which in turn stimulates follicular cells causing an enlargement of the thyroid, which itself can lead to thyroid tumors (10,11). A triazole herbicide, amitrole (3-amino-1,2,4-triazole), also inhibits TPO activity and iodide uptake (1). When this non-genotoxic carcinogen was administered to mice, an increased incidence of thyroid and pituitary tumors were found, whereas thyroid and liver tumors were induced in rats (1,12).
Phenobarbital exerts anti-thyroid activity through an indirect mechanism involving the liver (1,13). It induces hepatic microsomal enzymes that stimulate thyroid hormone metabolism, including upregulation of uridine diphosphate-glucuronosyltransferase activity toward T4 and/or increase biliary excretion of T4. This results in increased serum T4 clearance (1,13). Most herbicides and some other chemicals induce thyroid tumors through this indirect mechanism.
NKX2-1 (formerly called NKX2.1, TTF1, TITF1 or T/EBP) is a homeodomain transcription factor that is critical for the genesis of the thyroid, lung and ventral forebrain (14–16). In the thyroid, NKX2-1 regulates the expression of genes specifically expressed in this organ such as those encoding TPO (17,18), thyroglobulin (19), TSH receptor (20,21) and Na/I symporter (22), the genes that are responsible for thyroid hormone synthesis. An additional role for NKX2-1 in the maintenance of ordered architecture and function of the differentiated thyroid was demonstrated using Nkx2-1(fl/fl);TPO-Cre (Nkx2-1-thyroid-conditional hypomorphic) mice (23), which express Cre recombinase under control of the TPO gene promoter (24). In the Nkx2-1-thyroid-conditional hypomorphic mice, thyroid-specific Nkx2-1 disruption was found by 1 month of age, but in only about one-half of the follicular cells. This is distinct from the Nkx2-1-heterozygous mice in which each cell expresses one-half the amount of NKX2-1. Thyroid follicular cells that had lost NKX2-1 expression appeared to undergo degeneration (23). Interestingly, ~20% of Nkx2-1-thyroid-conditional hypomorphic mice exhibited atrophic/degenerative thyroids with extremely high TSH levels, whereas the remaining ~80% of the mice developed thyroids with extraordinary dilated follicles as they age (23). Although the reason for the dual phenotypes is not known, a possible role for NKX2-1 in the pathogenesis of thyroid diseases in humans was suggested because gradual and/or sporadic ablation of NKX2-1 gene could take place under natural circumstances.
In order to determine a possible role for NKX2-1 in thyroid carcinogenesis, we carried out a thyroid chemical carcinogenesis study using DHPN as a mutagenic carcinogen followed by SDM as a promoter or amitrole as an anti-thyroid agent administered to Nkx2-1-thyroid-conditional hypomorphic, Nkx2-1-heterozygous and Nkx2-1-floxed (wild-type) mice. A significantly higher incidence of adenoma was found in Nkx2-1-thyroid-conditional hypomorphic mice as compared with the other two groups only when they were treated with DHPN + SDM, but not amitrole. The results suggest that NKX2-1 may play a role in controlling cell proliferation in thyroid follicular cells when they are damaged by a genotoxic carcinogen.
Animals used for this study were Nkx2-1(fl/fl) (Nkx2-1-floxed), Nkx2-1(fl/fl);TPO-Cre (Nkx2-1-thyroid-conditional hypomorphic) and Nkx2-(fl/ko) (Nkx2-1-heterozygous) mice, hereafter called Flox, Flox-Cre and Het, respectively. They were a mixed background of several strains, including C57BL/6, 129Sv, FVB and Black Swiss. Characterization of these mice was described previously (23). Flox (Nkx2-1-floxed) mice are considered as wild-type based on their health status, fertility, histological and molecular biological characteristics of their tissues.
DHPN, SDM and amitrole were purchased from Sigma–Aldrich (St Louis, MO). Twelve mice (six males and six females) per genotype were injected with DHPN subcutaneously (5600 mg/kg) at the age of 4 weeks. Treatment of SDM through drinking water (2000 p.p.m.) started a week later and continued for 180 days until the day of euthanasia. For the amitrole carcinogenesis study, mice were fed a ground meal diet containing amitrole (0.1 g/100 g diet) at the age of 5 weeks and feeding continued for 180 days. On day 180 after the initiation of SDM or amitrole treatment, all mice were euthanized, and the thyroid and pituitary were fixed in 10% neutral buffered formalin, embedded in paraffin, sectioned at 4–5 μm and stained with hematoxylin and eosin; the age of mice at the time of necropsy was ~30 weeks. Pathological evaluation was the summary of those independently carried out by three pathologists. All animal studies were carried out after approval by the National Cancer Institute Animal Care and Use Committee.
Several paraffin sections of 4–5 μm thickness were combined and used for DNA extraction using EX-WAX paraffin-embedded DNA Extraction Kit (Millipore, Billerica, MA). A part of exons 1 and 2 of H-, K and N-Ras genes and exon 18 of Braf gene were amplified and the amplified fragments were sequenced. Primers used for polymerase chain reaction (PCR) amplification were shown in Table I. PCR condition used was 95°C for 5 min, 40 times of 94°C for 1 min, 55°C for 1 min and 72°C for 1 min, followed by 72°C for 7 min. Amplified fragments were run on agarose gels and each band was subjected to DNA purification using QIAquick PCR Purification Kit (Qiagen, Valencia, CA), followed by DNA sequencing analysis (Beckman coulter, model CEQ-200XL, Fullerton, CA).
Bromodeoxyuridine (BrdU) was injected intraperitoneally at 50 μg/g body wt 2 h before euthanasia. Thyroid sections were prepared on sialinated slides as described above, which were pretreated in a microwave oven in Tris–ethylenediaminetetraacetic acid buffer (10 mM Tris and 1 mM ethylenediaminetetraacetic acid-Na) at pH 9.0 for 30 min at 95°C. The slides were cooled to room temperature for at least 30 min and then incubated for 30 min in 3% H2O2 (vol/vol) in methanol to block endogenous peroxidase activity. They were further treated with 10% normal goat serum, 0.05% Tween 20 and 3% skim milk in phosphate-buffered saline to block non-specific binding sites. For the assessment of proliferative activity of thyroid follicular cells, a double immunostaining for BrdU and calcitonin was applied. Positive cells for both BrdU and calcitonin were excluded from the analysis due to C cells that express calcitonin (25). Sections were incubated with rat polyclonal anti-BrdU antibody (1:100 dilution, MCA2060; AbD Serotec, Raleigh, NC) (26,27) and rabbit polyclonal anti-calcitonin antibody (1:1000 dilution; ICN, Irvine, CA) overnight at 4°C in a humidified chamber. After rinsing three times in phosphate-buffered saline for 15 min, the tissue sections were incubated with alkaline phosphatase anti-Rabbit IgG (H + L) (1:200 dilution, AP-1000; Vector Laboratories) for 30 min. Following rinsing in phosphate-buffered saline for 15 min, fuchsin substrate (K0624; DAKO, Carpinteria, CA) was used to visualize specific immunoreactive staining for calcitonin. The sections were then incubated with biotinylated goat anti-rat IgG (H + L) (1:200 dilution, BA-9400; Vector Laboratories), followed by processing with the ABC method using commercially available kit (Vector Laboratories). Immunocomplexes were visualized with 3,3′-diaminobenzidine tetrahydrochloride (DAKO). The slides were rinsed in tap water, counterstained with Mayer's hematoxylin and mounted on crystal mount (Biomedia, Foster City, CA).
The BrdU-positive follicular cells were evaluated for each animal using an equal number of areas from both thyroid lobes at a ×40 objective magnification. At least 2000 intrafollicular cells (at least 1000 cells per lobe) were counted in the fields randomly chosen that equally include the center and the periphery of each lobe. Among BrdU positives, only those from true thyroid follicular cells, but not C cells, were counted based on the established counting rules (25). The BrdU-positive cell numbers were expressed as the mean per 1000 follicular cells. The difference between each group was analyzed by t-test using GraphPad Prism 5 software (La Jolla, CA); P<0.05 was considered statistically significant.
In order to examine a possible role for NKX2-1 in thyroid carcinogenesis, Flox (Nkx2-1-floxed), Flox-Cre (Nkx 2-1-thyroid-conditional hypomorphic) and Het (Nkx2-1-heterozygous) mice were subjected to a thyroid carcinogenesis bioassay using two protocols; administration of DHPN as a mutagen followed by SDM as a promoter and amitrole as an anti-thyroid agent. Their thyroids were examined at ~30 weeks of age.
In both DHPN + SDM- and amitrole-treated groups, Flox and Het mouse thyroids had a statistically significantly higher incidence of goiter (diffuse follicular hyperplasia) (Figure 1C) as compared with thyroids from untreated mice within the same genotype (Table II). The incidence of goiter in carcinogen-treated Flox-Cre mice was lower than those in Flox or Het mice and was similar to untreated groups. In contrast, a higher incidence of focal hyperplasia (Figure 1A and D) was found in Flox-Cre mice treated with DHPN + SDM or amitrole as compared with untreated mice, or Flox or Het mice within the same carcinogen treatment, most of them with statistical significance (Table II). A statistically significantly higher incidence of adenomas was found in DHPN + SDM-treated Flox-Cre mice as compared with untreated or other genotypes of mice. The adenoma incidence in amitrole-treated mice was similar among three genotypes, all of which are statistically significantly higher than those of untreated group of mice. Carcinoma by itself or within the adenoma were found in a few thyroids from Flox and Flox-Cre mice treated with DHPN + SDM and one Flox-Cre mouse treated with amitrole (Figure 1A and B; Table II). Some adenomas were oncocytic, which were found in one Flox and two Flox-Cre mouse thyroids (Figure 1E–G).
The degree of pituitary hyperplasia was also assessed; in DHPN + SDM treatment, severe hyperplasia (Figure 1H) was found at a higher incidence in most Flox-Cre mice as compared with the other two groups, in which most pituitary showed mild hyperplasia (Table II). There seems to be in agreement with the higher incidence of severe pituitary hyperplasia and adenomas in Flox-Cre thyroids. It is probable that pituitary hyperplasia is from TSH cells that stimulate the thyroid, and/or the mutagenic carcinogen causes damage in thyroid follicular cells that results in pituitary TSH cell hyperplasia. In amitrole-treated mice, the degree of pituitary hyperplasia was similar among three groups, which appeared to agree with no significant changes found in the incidence of thyroid adenomas among these three groups (Table II). These results demonstrate that Flox-Cre thyroid-conditional hypomorphic mice are more susceptible than the other two groups of mice to DHPN + SDM carcinogenesis, but not amitrole carcinogenesis.
Since DHPN is known to cause Ras gene mutations in rodents (28,29), mutations for codons 12 and 13 in exon 1 and codon 61 in exon 2 of the mouse H-, K- and N-Ras genes were analyzed by PCR amplification, followed by DNA sequencing of the purified DNA fragments. BrafV600E mutation was also examined since its overexpression in thyroid follicular cells in transgenic mice resulted in papillary cancers (30). Our analysis demonstrated no mutations in any genomic sequences analyzed in any groups of mouse thyroids.
In order to gain insight into the relationship between incidence of adenoma and genotypes of DHPN + SDM-treated mice, the cell proliferation rate was examined by BrdU labeling (Figure 2). BrdU was administered to various ages of three genotypes of mice, and their thyroids were examined for BrdU incorporation 2 h later (Figure 2A). Two groups of mice were studied including those <40 days of age and those between 45 and 126 days of age (Figure 2B). In both age groups, the number of BrdU-positive follicular cells was >2-fold higher in Flox-Cre thyroids than those of Flox and Het thyroids. Among the latter thyroids, there were no statistically significant differences. The cell proliferation rate in Flox-Cre thyroids tended to be higher in younger mice as compared to older mice, although not statistically significant. These results suggest that the higher cell proliferation rate in Flox-Cre thyroids may be at least partially responsible for the increased incidence of neoplastic thyroid lesions in these mice administered the mutagenic carcinogen DHPN.
In this study, we demonstrated that Flox-Cre thyroid-conditional hypomorphic mice developed adenoma in their thyroids at statistically significantly higher incidences as compared with Flox and Het mice only when they were treated with DHPN, followed by SDM, but not with amitrole. We also found that the Flox-Cre thyroids have >2-fold higher cell proliferation rate as compared with Flox or Het mice when examined by BrdU incorporation. Since DHPN is a mutagen that causes DNA damage, the higher cell proliferation rate found in Flox-Cre thyroids is probably to result in higher susceptibility to the effect of a mutagen, leading to enhanced incidences of adenomas. In contrast, thyroid carcinogenesis by the non-genotoxic carcinogen amitrole involves thyroid hormone imbalance that does not act at the genomic level. The higher cell proliferation rate in Flox-Cre thyroids seems not to play a role in amitrole carcinogenesis.
Our previous study demonstrated that Flox-Cre thyroids had only approximately one-half of Nkx2-1 alleles disrupted and exhibited about half the level of NKX2-1 expression as compared with Flox thyroids when measured by quantitative PCR (23). The level of NKX2-1 expression in Flox-Cre was similar to that of Het thyroids. However, immunohistochemical analysis revealed that in Flox-Cre thyroids, one-half the level of NKX2-1 expression was the result of a mixture of cells that either express or do not express NKX2-1, whereas in Het thyroids, all thyroid follicular cells expressed NKX2-1 presumably at one-half the level of Flox thyroids. Note that immunohistochemical analysis did not allow us to determine differences in the level of NKX2-1 expression among cells. The reason why the Nkx2-1 gene was not disrupted in all follicular cells of Flox-Cre thyroids is not known (23). It could be because TPO is not expressed in all follicular cells (24) or TPO expression is positively regulated by NKX2-1 itself (17,18). The latter is most likely because Pten(fl/fl);TPO-Cre mouse thyroid, in which the Pten gene is not under control of NKX2-1, exhibited a recombination efficiency reaching almost 100% (31). Approximately, half of the follicular cells in Flox-Cre thyroid-conditional hypomorphic thyroids lost NKX2-1 expression by 1 month of age as judged by immunohistochemical analysis, and those cells that had lost NKX2-1 expression underwent degeneration (23). Despite this, ~80% of Flox-Cre mouse thyroids became extraordinary dilated as they aged with almost all follicular cells expressing NKX2-1 (23). In order to explain the apparent dimorphic phenotypes of Flox-Cre thyroids, we hypothesized that stem/progenitor cells may be present in the thyroid (32), which actively participate in the regeneration of Flox-Cre mouse thyroids (23). Accordingly, a higher rate of cell proliferation may be seen in the latter mouse thyroids. The current BrdU incorporation results provide the evidence that Flox-Cre mouse thyroids indeed have a higher cell proliferation rate. These results suggest a role for NKX2-1 in the pathogenesis of thyroid diseases, particularly cancer in humans; the somatic mutation of NKX2-1 gene could occur due to exposure to genotoxic mutagens or radiation, which render NKX2-1 inactive, leading to higher cell proliferation in thyroid, which in turn augments the damage in DNA and/or chromosomes caused by genotoxic mutagens and/or irradiation. Mutagens and gamma radiation are among the risk factors for thyroid cancers, which induce genetic mutations and/or genomic instability (33). The role for NKX2-1 in thyroid carcinogenesis has recently been demonstrated by a genome-wide association study (34). Common variants on 9q22.33 and 14q13.3 were shown to be associated with thyroid cancer; the gene nearest to the 9q22.33 locus was FOXE1 (TTF2), another gene critical for thyroid organogenesis and function (15,16), whereas NKX2-1 was among the genes located at the 14q13.3 locus. Both variants contributed to an increased risk of both papillary and follicular thyroid cancer (34).
Extensive studies have been carried out to understand the role of genetic background involved in pathogenetic mechanisms of thyroid carcinogenesis, particularly in humans (33,35). Among the genes involved in thyroid carcinogenesis that have been extensively studied include RAS, RET, BRAF, PTEN, TP53, CTNNB1 (gene encoding β-catenin), GNAS1 (gene encoding guanosine triphosphate-binding protein Gs α subunit), TSHR and PPARG genes. The mutation and/or rearrangement of these genes are thought to be associated with increased risk of thyroid cancers (33,35–38). For instance, RET arrangement accounts for 50–90% of post-Chernobyl childhood thyroid cancer, whereas the RAS and the TP53 mutations are found in up to 60 and 88% of undifferentiated thyroid carcinomas, respectively (33). The role of genes in thyroid carcinogenesis that are involved in cell cycle regulation and/or cell adhesion molecules and extracellular matrix has also been described (33,35). Contrary to extensive studies done on human thyroid cancers, very few studies have been carried out in rodents, particularly in mice.
Frequent mutation of K-Ras codon 12 was found in DHPN-induced tumors of thyroid, kidney and lung in Wistar rats (28) and thyroid carcinoma lines derived from tumors in rat induced by DHPN (29), suggesting that K-Ras mutation is involved in carcinogenesis with DHPN. N-RAS mutation at codon 61 was prevalent in thyroid follicular tumors in humans (39). Our analysis did not reveal mutations in any of the H-, K- and N-Ras genes at codon 12 nor other known hot spots, codon 13, 61 and 63 in either DHPN- or amitrole-treated thyroids. Although the exact reason for this discrepancy is not known, a species difference could be considered; in a DHPN-induced carcinogenesis study, Kitahori et al. (28,29) used rat, whereas we used mice. It is known that not all species react the same to the same carcinogens. Further, it is possible that DHPN induced mutations within the Ras gene other than those we analyzed. Alternatively, H-, K- and N-Ras genes mutations, if produced, could be involved in later events during tumor formation in the thyroid. In this regard, the previous studies reporting Ras gene mutations used thyroid carcinomas (28,29). We had a few incidences where carcinomas were found in DHPN + SDM- and amitrole-treated mice but did not find any Ras gene mutations in the thyroids with carcinoma. As mentioned above, the involvement of genes other than RAS has been described in the pathogenetic mechanisms of thyroid follicular cell neoplasia in humans (33,35–38). It is quite possible that many of these genes share similar roles in thyroid carcinogenesis in mice. Due to the limitation of genomic DNAs available and that no RNAs could be successfully retrieved from paraffin sections to perform gene rearrangement studies, only the H-, K- and N-Ras and BrafV600E mutations were analyzed in the current studies. The fact that we did not find any mutations in these genes does not exclude the presence of mutations/rearrangements in other genes. The pathogenetic mechanism for the precise molecular pathway for the DHPN + SDM and amitrole carcinogenesis remains to be determined.
Intramural Research Program of the National Cancer Institute, Center for Cancer Research (Project no. Z01 BC 005522).
We would like to thank Stephanie Oldham and Jennifer Harris for the carcinogenesis study and Frank J.Gonzalez for critical review of the manuscript.
Conflict of Interest Statement: None declared.