Main

Several tumor types, differing in biological and clinical behavior, originate from the thyroid epithelial follicular cells. They include well-differentiated, indolent papillary thyroid and follicular thyroid carcinomas (PTC and FTC), as well as extremely aggressive anaplastic carcinoma.1 Studies performed in several laboratories, including ours, have demonstrated that distinct molecular events are associated with specific tumor types.2 FTC is characterized by the PAX8/PPARγ rearrangement and activating mutations of RAS genes.1 PTC is associated with rearrangements involving the RET and NTRK1 tyrosine kinase receptors,2 and the V600E BRAF-activating mutation.3, 4, 5 More recently, microarray studies have identified several genes that might be important in the molecular pathogenesis and the malignant progression of thyroid cancer and could be used as diagnostic or prognostic molecular markers. These candidate genes are involved in several different processes, such as cell adhesion, cell cycle progression, mitogenic control and tumorigenesis.6, 7, 8, 9, 10 In spite of that, the molecular pathogenesis of thyroid cancer is still incomplete; in particular, a role of tumor suppressor genes has not yet been assessed. Metallothioneins (MTs) are low-molecular weight proteins of 6–7 kDa, with high content of cysteine (30%) and complete absence of aromatic amino acids and histidine11 capable of binding heavy metals with high affinity. Human MTs are encoded by a family of genes clustered on chromosome 16 and containing at least 10 functional members, and are organized in four major isoforms: MT1 (seven functional genes), MT2, MT3 and MT4.12 MT1 and MT2 are ubiquitously expressed, and MT2 accounts for 80% of MT expression, at least in some tissues.13 Expression of MT3 and MT4 is restricted to brain and reproductive organs, and to stratified squamous epithelia of skin and tongue, respectively. While MT3 and MT4 are constitutively expressed, MT1 and MT2 are coordinately regulated by a variety of developmental and environmental signals, such as metals, oxidative stress, cytokines and glucocorticoid hormones.14 The primary functions of MTs have not been defined with certainty. They regulate the availability of heavy metals, especially zinc, which associates with hundreds of proteins regulating gene expression, metabolism and cell growth. MTs also participate in metal detoxification and protection against oxidative stress.12, 14

The role and the expression of MTs in human tumors have been extensively investigated, but the results are often contradictory among different cancer types; therefore, the real contribution of these important molecules to malignant transformation has not yet been clarified.15 MT overexpression has been observed in several tumors such as melanoma, breast and lung cancer, and in some cases has been associated with a poor prognosis16, 17 and with resistance to treatments.18 Conversely, other tumor types, such as colorectal cancer and hepatocellular carcinoma, show a downregulation of MT expression during cancer progression.19, 20 Recently, the molecular mechanism for the downregulation of MT genes in liver tumors has been unveiled; it involves the PI3K cascade.21 With respect to thyroid tumors, by microarray analysis Huang et al6 identified MTs as genes downregulated in PTC with respect to normal thyroid tissue. Subsequently, they showed that in PTC the expression of MT1G gene is repressed by promoter methylation, suggesting an oncosuppressor role.22 In agreement with this finding, immunohistochemistry analysis showed reduced MT expression in PTC.23 To further define the role of MTs in thyroid carcinogenesis, we performed expression and biological studies. The expression of several MT genes in a collection of normal and tumor tissue samples was investigated by analyzing gene expression profile data and by performing quantitative real-time PCR (QRT-PCR). In addition, by immunohistochemistry, MT expression was also investigated at the protein level. Both approaches documented downregulation of several MT genes in PTC and, to a lesser extent, in FTC. The functional consequences of MT downregulation were addressed in K1 cells, in which the expression of MT1G is silenced by promoter methylation. Restoration of MT1G expression by cDNA transfection reduced in vitro growth rate and tumorigenicity of K1 cells. Overall, our data suggest that MT genes act as oncosuppressors in thyroid epithelial carcinogenesis.

MATERIALS AND METHODS

Sample Collection and Microarray Analysis

The thyroid samples were collected at the Department of Pathology at the ‘Istituto Nazionale dei Tumori’ (Milan, Italy) from 1984 to date. All the patients gave their written informed consent. The specimens were snap frozen. A frozen section of each cryopreserved sample was stained with hematoxylin and eosin, and samples eligible for microarray analysis, ie containing exclusively tumor or normal tissue, were selected. The cDNA microarray used in this study contains 4451 unique clones selected from the human sequence-verified IMAGE clone collection (Research Genetics/Invitrogen, Carlsbad, CA, USA) along with plant genes, printing controls (water/DMSO) and spike genes (Amersham Bioscience, Amersham, UK) added as internal controls. Each cDNA fragment was PCR-amplified and spotted in triplicate on type 7 star slides (Amersham Bioscience, UK). The procedures of RNA isolation, probe labeling, sample hybridization and scanning were performed as previously described.24 Supervised analysis (class comparison and hierarchical clustering) was performed using BRB ArrayTool v.3.3_Beta1 (http://linus.nci.nih.gov/BRB-ArrayTools.html).

The expression levels of the three MT genes in all the tissue samples hybridized on the array were measured as log ratio between the expression level of the specimens and that of the reference. Genes showing a positive log ratio value (ie more expressed in the thyroid sample than in the reference) are called upregulated; those with a negative log ratio value are called downregulated.

QRT-PCR

Total RNA from thyroid tissues was reverse transcribed using the High Capacity cDNA Archive kit (Applied Biosystems, Foster City, CA, USA). Primers specific for the different MT isoforms, located in the divergent 5′ and 3′ untranslated regions, were used.25 PCR products were subjected to direct nucleotide sequencing, and their correct identity was confirmed using BLAST software (http://www.ncbi.nlm.nih.gov/blast). The housekeeping gene 18S was amplified using the following primers: Fw 5′-CGCCGCTAGAGGTGAAATTC-3′, Rev 5′-CTTTCGCTCTGGTCCGTCTT-3′.26 QRT-PCR reactions were carried out in triplicate using 2 × SYBR GREEN PCR Master Mix (Applied Biosystems) on the ABI PRISM 7700 Sequence Detector System (Applied Biosystems) and the amplification conditions reported by Tai et al.27 Relative quantification of expression levels was calculated using the comparative Ct method;28 for each MT isoform, the normal sample showing the lowest expression was selected as a calibrator for the calculation of the ΔΔCt value; samples with an expression level lower than the calibrator were called downregulated.

Immunohistochemistry

Formalin-fixed and paraffin-embedded 2-μm tumor sections were treated with 3% H2O2 to inhibit endogenous peroxidase, and then subjected to antigen retrieval using 5 mM citrate buffer (pH 6) in autoclave at 95°C for 15 min. Sections were incubated with Ultra V Block (Lab Vision, Fremont, CA, USA) for 10 min, to limit nonspecific reactivity, and then incubated with MT mouse monoclonal antibody (clone E9; Dako, Carpinteira, CA, USA) diluted 1:150 in ChemMate™ antibody diluent (Dako) for 1 h. All stains were carried out by primary antibody enhancer and polymer following provided protocol (UltraVision LP large-volume detection system HRP polymer; Lab Vision) and developed with liquid DAB+ substrate chromogen system (Dako). Breast was used as positive control. A semiquantitative assessment of the MT immunostaining was determined using the multiplicative ‘quick-score’ method, which accounts for both the intensity and the extent of cell staining.29 To obtain the relative expression, the score values of each sample were normalized for that of normal tissue.

Constructs

The MT1G cDNA was isolated by RT-PCR using the primers described by Tai et al27 and cloned into the pTargeT expression vector. The insert was extracted by XhoI/EcoRI digestion and transferred into the XhoI/EcoRI-cut pcDNA3 expression vector. The resulting construct carries the MT1G cDNA in the sense orientation with respect to the CMV promoter present in the vector. This plasmid was used as a template for the production of a myc-tagged cDNA. The MT1G cDNA was amplified with primers T7 (AATACGACTCACTATAGGG) and GCGGATCC GGCGCAGCAGCTGCACTTCTC, containing nucleotides 166–183 of MT1G mRNA (italic), and the BamHI restriction site (bold). The PCR product was digested with EcoRI and BamHI endonucleases and inserted into the pcDNA3.1/myc-His(−)A carrying compatible ends. The resulting construct encodes the MT1G-myc fusion protein.

Cell Culture and Transfection

The K1 cell line, derived from a human PTC and obtained from ECACC, was cultured at 37°C, 5% CO2 humidified atmosphere, in DMEM:HAM’S F12:MCDB 105 (2:1:1) containing 2 mM glutamine and 10% FCS. Transient transfection was performed with Fugene6 (Roche, Indianapolis, IN, USA) following manufacturer's instructions. Stable transfection was performed with Nucleofector™ II (Amaxa Biosystem, Gathesburg, MD, USA). Plasmid DNA (1 μg) (pcDNA3.1-MT1G-myc or pcDNA3.1-myc, as control) was added to a mix containing 90 μl of human keratinocyte nucleofector solution and 20 μl of Supplement 1 Human Keratinocyte Nucleofector Kit (Amaxa Biosystem). K1 cells (1 × 106) were resuspended in the above solution and were transfected by applying the T003 program using the Nucleofector™ II (Amaxa Biosystem). Transfected cells were seeded onto two 10 cm Petri dishes, and G418 selection (400 μg/ml) was applied 24 h later. G418-resistant colonies were isolated after 2 weeks of selection and propagated in G418-containing medium.

Growth Curves

To determine the growth rate of K1-MT1G-myc clones, the alamarBlue® Assay (Biosource) was used, following the supplier information. Fifty cells per well were plated onto 96-well multiwell (Corning). After adhesion of cells, the alamarBlue (Biosource) was added to the medium at 20% v/v. The fluorescence at λ=590 and 535 nm was detected every 24 h, from day 1 to day 8, using a microplate reader (Tecan).

Western Blot Analysis

Proteins were extracted in RIPA buffer, quantified by Bradford's assay with the BIO-RAD Protein Assay (Bio-Rad, Munchen, Germany). Total protein extracts (70–100 μg) were separated on NuPAGE 4–12% Bis–Tris precasted gel (Invitrogen) and blotted on Hybond™-C supermembrane (Amersham Bioscience, Uppsala, Sweden). Hybridization was carried out with anti-MT E9 antibody (Dako, Carpinteira, CA, USA) or with the anti-myc 9E10 antibody hybridoma (Roche) and with the anti-β actin antibody (Sigma, St Louis, MO, USA). The immunoreactive bands were visualized using horseradish peroxidase-conjugated secondary antibodies and enhanced chemiluminescence (Amersham Bioscience, Sweden).

In Vivo Studies

Animal studies were reviewed and approved by the Ethics Committee for Animal Experimentation of the Fondazione IRCCS ‘Istituto Nazionale dei Tumori’ and are in accordance with the guidelines of the UK Coordinating Committee for Cancer Research.30. Female CD-1 nu/nu mice (8- to 9-week old) (Charles River, Calco, Italy) were injected subcutaneously into the left flank with 5 × 106 cells in 0.2 ml PBS volume. Tumor growth was assessed by evaluating tumor latency, ie days to reach 0.1 g, and by monitoring tumor weight (TW) twice a week. TW was estimated by the formula TW(g)=d2 × D/2 where d and D are the shortest and the longest diameters of the tumor, respectively, measured in cm.

RESULTS

Analysis of MT Gene Expression in Thyroid Tumors

We explored the expression level of MT genes by analyzing cDNA microarray data produced in our laboratory using RNA from a collection of 59 thyroid samples, including 9 normal thyroid tissues (N), 8 follicular adenomas (FA), 11 FTCs and 31 PTCs. RNAs were hybridized to a microarray spotted with 4451 unique cDNA clones; a pool of 10 human cell lines was used as a reference.24 Gene expression levels of the tumor histotype were compared to that of normal thyroid using class comparison analysis (http://linus.nci.nih.gov/BRB-ArrayTools.html). The MT1G, MT1H and MT1X genes were identified within a group of genes with low expression levels and strong downregulation in PTC samples compared to normal tissue (data not shown). Analysis of the expression levels of the three MT genes in all the tissue samples hybridized on the array are reported in Figure 1 and showed that the different MT isoforms are downregulated in at least 19 out of 31 (60%) of PTC and 3 out 11 (27%) of FTC. Our findings confirm the previous data on MT RNA downregulation in PTC6 and, in addition, they also suggest that MTs may be downregulated in FTC.

Figure 1
figure 1

cDNA array expression of MT genes. Graphic representation of the expression levels of the three MT clones spotted on the cDNA array in normal thyroid samples (yellow), follicular adenomas (green), follicular carcinomas (red) and papillary carcinomas (blue). The expression levels (relative expression, RE) of the three MT genes in all the tissue samples hybridized on the array were measured as log ratio between the expression level of the specimens and that of the reference. Genes showing a positive log ratio value (ie more expressed in the thyroid sample than in the reference) are called upregulated; those with a negative log ratio value are called downregulated.

Full size image

To corroborate the microarray data, we performed a QRT-PCR experiment with primers located on the divergent 5′ and 3′ untranslated regions of each MT gene; these primers do not cross react with other members of the family and thus are able to discriminate among the different MT isoforms.25 In addition to the three MT1 genes present on the array, we also selected primers for the analysis of the expression of MT1A, MT1B, MT1E, MT1F and MT2A, thus including all the MT genes ubiquitously expressed in human tissues. Informative results (based on expression level and/or reproducibility) were obtained for MT1E, MT1G, MT1X and MT2A. The ΔΔCt calculation was used to determine the expression level of the samples analyzed by QRT-PCR; the normal sample showing the lowest expression level was used as a calibrator. Samples with an expression level lower than the calibrator were called downregulated, and their frequencies are reported in Table 1. MT1E is downregulated in 80.9% of PTC and in 50% of FTC; MT1G is downregulated in 60% of PTC and 9% of FTC; MT1X is downregulated only in PTC (76%); MT2A shows downregulation in PTC (65%) and FTC (40%). On the whole these data indicate that (a) MT downregulation occurs frequently in differentiated epithelial thyroid tumors, being detectable in 38–71% of the cases analyzed, (b) although involving both differentiated follicular thyroid cancer types, MT downregulation is prevalent in PTC and (c) MT downregulation is infrequent in FA, since only a single specimen showing MT2A downregulation was detected (data not shown).

Table 1 MT downregulation in differentiated thyroid tumors
Full size table

Our results are in keeping with a previous observation reporting MT downregulation in PTC6 and, interestingly, they provide also important information about other MT isoforms and other thyroid tumor histotypes.

Evaluation of MT Protein Expression by Immunohistochemistry

With respect to expression analysis at protein level, two contradictory studies show either overexpression or downregulation of MT proteins in PTC as compared to normal thyroid.31. We investigated the MT protein expression by immunohistochemistry in the same samples analyzed by microarray. Corresponding archival paraffin-embedded sections, including both tumor and normal tissue were selected, and immunohistochemistry was performed using the E9 mouse monoclonal antibody, which reacts with both MT1 and MT2 isoforms. Normal thyroid follicular cells showed an intense staining, whereas in the tumor counterpart a heterogeneous staining pattern was observed. A semiquantitative assessment of the MT immunostaining was determined using the multiplicative ‘quick-score’ method,29 and the MT expression level of each sample with respect to normal thyroid cells was calculated. The results are shown in Figure 2a. All the tumor samples showed reduced MT expression compared to normal controls. In the FTC series, the reduction was 22–61%; in one sample only, it was greater than 60%. In the PTC series, reduction was in the range 28–75%, and it was greater than 60% in 9 out 19 samples. In Figure 2b, two PTC samples with different extent (53 and 83%, respectively) of MT protein downregulation are shown. On the whole, the immunohistochemistry analysis demonstrates that the overall expression of MT proteins is frequently decreased in thyroid tumors, particularly in PTC. Our results are in agreement with a recent report22 in which the same antibody was used, and demonstrate that MT gene expression downregulation correlate with reduced protein level.

Figure 2
figure 2

Detection of MT expression by immunohistochemistry. (a) Immunohistochemistry was performed on 2-μm cut formalin-fixed and paraffin-embedded sections as described in Materials and methods. A semiquantitative assessment of the MT immunostaining was determined using the multiplicative ‘quick-score’ method, which accounts for both the intensity and the extent of cell staining. The relative expression was obtained by normalizing the score values of each sample for that of normal tissue. (b) Section of two PTC samples (13 and 16) with different extent of MT downregulation. The normal (N) and tumor (T) components are indicated.

Full size image

Restoration of MT1G Gene Expression in K1 Thyroid Carcinoma Cell Line

The K1 cell line, derived from a PTC, does not express the MT1G gene as a consequence of promoter methylation, as demonstrated by its re-expression following azacytidine treatment22 (our unpublished results). Other MT isoforms are expressed, as detected by western blot analysis with E9 antibody (Figure 3). To assess the role of MT downregulation in PTC, we restored the expression of MT1G in K1 cells. We constructed the MT1G-myc expression plasmid carrying the MT1G cDNA fused in frame with the myc epitope and producing a 9 kDa MT1G-myc protein detected by both anti-myc and anti-MT (E9) antibodies and, in the latter case, distinguishable from the endogenous MT proteins (6 kDa) (Figure 3). The MT1G-myc expression plasmid was transfected into K1 cells, and transfected clones were isolated by selection in geneticin. Western blot analysis identified several independent clones expressing the exogenous MT1G-myc protein at variable levels (Figure 3, right). To investigate the effect of MT1G restoration in K1 cells, the growth rate of K1/MT1G-myc clones (namely K1/1G-A, -S, -W and -X) was determined. As a control, the K1/1G-D clone (not expressing MT1G-myc) and the K1/vec clone (transfected with the empty vector) were used. As shown in Figure 4a, clones expressing the MT1G-myc protein display a growth rate reduced with respect to control cells. As a consequence, when plated at low density K1/1G cells formed colonies smaller than those of control cells; an example is shown in Figure 4b.

Figure 3
figure 3

Expression of MT1G-myc protein in K1 cells. K1 cells transiently transfected with MT1G-myc construct (left) and K1 clones stably transfected with MT1G-myc construct (right) were analyzed by western blot for the expression of MT1G-myc protein using the E9 anti-MT antibody. As a control, untransfected (−) or mock-transfected K1 cells were used. Endogenous MTs and the MT1G-myc proteins are indicated by arrows.

Full size image
Figure 4
figure 4

Effect of MT1G-myc protein expression in K1 cells. (a) Growth curve of K1/1G and K1/vec clones were determined by the alamarBlue method as described in the Materials and methods section. Fluorescence was determined at the indicated time points. (b) Colonies formed by the K1/1G-W and by the K1/vec control clone are shown. In total, 600 cells were plated onto 10 cm Petri dishes; colonies were fixed and stained after 2 weeks. (c) In vivo tumorigenicity assay. K1/1G-X, K1/vec clones and parental K1 cells were inoculated subcutaneously into the left flank of athymic nude mice as described in Materials and methods. Tumor growth was monitored twice a week from day 10 to day 30. (d) The expression of MT1G-myc protein in K1/1G-X cells, before (1) and after (2) implant in mice was determined by western blot with anti-myc antibodies. As a control, parental K1 and K1/vec cells are shown.

Full size image

We next investigated the impact of MT1G-induced growth reduction on the in vivo tumorigenicity of K1 cells by inoculating K1/1G-X cells in nude mice. As a control, both the parental K1 cell line and the K1/vec clone were used. For each cell line, 5 × 106 cells were injected subcutaneously. TW was monitored from day 10 to day 30 after implant; the results are reported in Figure 4c. The control cells induced tumor formation in 4/4 mice with tumor latency of 10 days, and none of the tumors regressed. The K1/1G-X cell line induced tumor formation in 4/5 mice; among the tumor-bearing mice, one underwent complete remission. The growth rate of tumors induced by K1/1G-X cells was reduced with respect to that of tumors induced by parental K1 and K1/vec cell lines. At day 30 mice were killed; tumors were explanted and used for cell culture establishment. Western blot analysis showed that tumor-derived cells express the MT1G-myc protein (Figure 4d). These results demonstrate that restoration of MT1G expression reduces the tumorigenicity of K1 thyroid carcinoma cells.

DISCUSSION

By expression and functional studies, we have provided evidence that MT genes may act as tumor suppressor in thyroid tumorigenesis.

Expression data (array and QRT-PCR) showed that downregulation of MT gene expression (namely MT1E, MT1G, MT1X and MT2A) is associated with both thyroid differentiated tumor types, although it appears prevalent in PTC. Our results are in keeping with a previous observation reporting MT downregulation in PTC6 and, interestingly, they also provide important information about other MT isoforms and other thyroid tumor histotypes. Indeed, they demonstrate that expression of MT isoforms can be modulated to different extents in different thyroid tumor histotypes. While this paper was in preparation, a paper reported downregulation of 13 different MT genes in PTC,32 in agreement with our study and the previous report by Huang et al.6

At variance with the other two studies, we also investigated the MT expression at the protein level, by immunohistochemistry. In this respect, two contradictory previous studies are available. An early report31 showed that MTs are higher expressed in PTC than in normal thyroid tissues, whereas a more recent study documented the absence of MT expression in a consistent fraction (65%) of PTC.23 Our data are in keeping with the second report: we detected a considerable reduction of MT expression in PTC, in agreement also with gene expression studies. The discrepancy with the early study is most likely related to the different antibodies used.

The mechanism(s) leading to MT mRNA downregulation in tumors may involve MT promoter methylation as well as transcriptional repression. With respect to PTC, Huang et al22 have shown that repression of MT1G is associated with promoter downregulation.

The molecular consequences of MT downregulation in some cancers, including thyroid carcinoma, are not completely understood. Considering the antioxidant role of MTs, their downregulation should have a detrimental effect for the cell. However, this could be overcome by the activation of compensatory mechanisms. On the other hand, the lack of MTs could result in greater availability of zinc for a variety of zinc-finger transcription factor, whose increased activity could account for the rapid proliferation of cancer cells. Moreover, MT downregulation may act together with deregulation of other genetic elements in promoting tumor cell growth.

To explore the contribution of MT downregulation to thyroid tumor phenotype, we restored the MT1G gene in the K1 cell line, in which the endogenous gene is silenced by promoter methylation,22 whereas other MT isoforms are expressed. With respect to the parental K1 cells, K1 clones expressing exogenous MT1G showed reduction of both in vitro proliferation rate and in vivo tumorigenicity, thus demonstrating that MT1G counteracts the growth properties of transformed cells. It is worth noting that K1 cells harbor the V600E BRAF mutation, which has been recently shown to be important to maintain proliferation, transformation and tumorigenicity of BRAF-mutant PTC cell lines.33 Our data demonstrate that even in the presence of the V600E BRAF mutation and of expression of several other MT isoforms, restoration of MT1G expression is capable of modulating the transformed properties of K1 cells.

Taken together, our results demonstrate that low MT gene expression is associated with PTC, and that MTs could play an oncosuppressor role in thyroid carcinogenesis. However, further studies are needed to elucidate the precise contribution of MTs to thyroid carcinogenesis. These involve the dissection of the role of each MT isoform, and the determination of possible cross-talk between MT downregulation and intracellular pathways triggered by the various PTC-associated oncogenes.