CANCER GENOMICS & PROTEOMICS 4: 403-410 (2007)
Characterization of Microarray Gene Expression
Profiles of Early Stage Thyroid Tumours
KATALIN GOMBOS1, ESZTER SZELE1, ISTVÁN KISS1, TÍMEA VARJAS1,
LÁSZLÓ PUSKÁS2, LÁSZLÓ KOZMA3, FERENC JUHÁSZ4, ERIK KOVÁCS5,
ISTVÁN SZANYI6 and ISTVÁN EMBER1
1Institute of Public Health, Faculty of Medicine, 5Heart Institute and First Department
of Internal Medicine and 6Department of Otolaryngology Head and Neck Surgery,
Faculty of Medicine, University of Pecs, Pecs; 2Functional Genomical Laboratory,
Biological Research Center of Hungarian Academy of Sciences, Szeged;
3Regional Blood Transfusion Center of Debrecen, Debrecen;
4First Department of Surgery, University of Debrecen,
Faculty of Medicine, Debrecen, Hungary
Abstract. Background: Microarray analysis offers the opportunity of screening transcriptional expression profile of neoplastic cells on a genomic level. Defining consistent changes in gene expression pattern of tumours enables the detection of genes essential for tumorigenesis and might provide biomarkers to early recognition of malignant behaviour and new therapeutical targets. Patients and Methods: A high-density oligonucleotide array with 20,000 human gene-specific oligonucleotide was used to analyze benign and early-stage malignant thyroid tumours of epithelial origin: follicular adenoma, follicular carcinoma and papillary carcinoma, compared to normal thyroid tissue. Results: Significant expression differences of 279 genes - underexpression of 252 and overexpression of 27 genes - were found. The overlapping genes of the different histological types were examined extensively. Among these genes a limited set acting on the same transcriptional pathway, through NF-κB, were found. Conclusion: The role of overlapping genes in histologically different tumours has not been clarified, but might represent early or pivotal steps of carcinogenesis. All investigated histiotypes of tumours contained significantly modulated genes acting on the NF-κB regulatory pathway. Our findings suggest that modulation of NF-κB signalling plays a crucial role in early thyroid carcinogenesis.
Correspondence to: Katalin Gombos, Institute of Public Health, Faculty of Medicine, University of Pecs, Szigeti ut 12. H-7624 Pecs, Hungary. Tel: +36 72 536 394, Fax: +36 72 536 395, e-mail: firstname.lastname@example.org
Key Words: Microarray, gene expression profiling, thyroid tumour, follicular adenoma, follicular carcinoma, papillary carcinoma.
The most frequent endocrine neoplasms are the tumours of the thyroid gland, with an incidence of 9 per 100,000 cases per year in industrialized countries (1). Benign thyroid nodules are far more common. An estimated 5-10% of the population develops a clinically significant thyroid nodule during their lifetime (2). Therefore distinguishing benign from malignant nodules is a frequent diagnostic dilemma in the clinical practice. Importance of early recognition of thyroid malignancies lies in the fact that in the nodular form of the disease surgical treatment could provide complete recovery if the diagnosis is established before the development of local invasion and metastasis. The evaluation of thyroid nodules by fine-needle aspiration (FNA) has been a standard for almost 30 years. However, the quality and quantity of the cytological samples taken by FNA are frequently not suitable for cytological evaluation. Invasiveness, metastatic features and many other molecular features of tumours are not discernible on cytology and only after surgical resection and histological evaluation can they be recognised (3).
High throughput technologies, such as microarray, offer the opportunity of screening the entire transcriptional expression profile of neoplastic cells on a genomic level. Defining consistent changes in the gene expression pattern of tumours might lead to detection of genes essential for tumour development and might provide new biomarkers to classify tumours based on molecular features, early molecular identification of malignant behaviour, as well as new targets for therapy (4).
Patients and Methods
Tissue samples. Tumour tissue samples were obtained from patients with benign and clinically early-stage malignant thyroid tumours (T1, N0, M0) undergoing surgery at the First Department of Surgery in Debrecen University. All tumour samples were obtained with the permission of institutional review boards and informed consent of the patients. Dissected samples were frozen immediately after surgery at stored at -80oC. Histopathological classification was confirmed by an endocrine pathologist. Tumour samples of follicular adenoma (n=8), follicular carcinoma (n=7), papillary carcinoma (n=10) and normal thyroid tissue (from patients who underwent total thyroidectomy n=20) were selected for our investigations.
RNA isolation. Total RNA was isolated from 40 mg of each tissue sample with NucleoSpin RNA purification kit (Macherey-Nagel, Durren, Germany) according to the manufacturer's instructions. The quality and quantity of all RNA preparations were assessed by agarose gel electrophoresis and spectrophotometry (NanoDrop, Rockland, DE, USA). Total RNA was used both for microarray analysis and quantitative real-time PCR.
Microarray design. A total of 20,000 human gene-specific, amino-modified oligonucleotides were purchased (Sigma-Operon, Budapest, Hungary) and arrayed onto PXM oligonucleotide slides (Full Moon Biosystems, Sunnyvale, CA, USA) using a MicroGrid Total Array System (BioRobotics, Cambridge, UK) spotter. DNA elements were deposited in duplicate in distinct areas of the array. After printing, slides were incubated in a humid chamber for 14 hours at 42o.C.
Microarray preparation and hybridization. Two micrograms of total RNA were reverse transcribed using a poly-dT primed Genisphere Expression Array 900 Detection system (Genisphere, Hatfield, PA, USA) in 20 Il total volume using 20 U RNAsin (Fermentas, Vilnius, Lithuania), 1x first strand buffer and U of RNAse H(-) point mutant M-MLV reverse transcriptase (Fermentas). All the other probe preparation steps were carried out according to the manufacturer's instructions (Genisphere). Both the first step cDNA hybridization and the second step capture reagent (2.5 µl of each Cy5 and Cy3) hybridization were carried out in a Ventana hybridization station (Ventana Discovery, Tucson, AZ, USA) using the antibody protocol. First hybridization was performed at 40oC for 6 hours in FGL2 hybridization buffer (10x Denhart solution, 0.25M sodium phosphate buffer pH 7.0, 1 mM ethylene-diamine-tetraacetic acid, 1x salt sodium citrate, 0.5% sodium dodecyl sulphate) then 2.5 µl of each Cy5 and Cy3 capture reagents were added to the slides in 200 Il "Chiphyb" hybridization buffer (Ventana) and they were incubated at 42oC for 2 hours. After hybridization, the slides were washed in 0.2x salt sodium citrate twice at room temperature for 10 minutes, then dried and scanned.
Scanning and data analysis. Each array was scanned under a green (543 nm for Cy3 labeling) or a red laser (633 nm for Cy5 labeling) using a ScanArray Lite (GSI Lumonics, Billerica, MA, USA) scanning confocal fluorescence scanner. Scanned output files were analyzed using the GenePix Pro 5.0 software (Axon Instruments Inc., Foster City, CA, USA). For each channel, the median values of feature and local background pixel intensities were determined. The background corrected expression data was filtered for flagged spots and weak signal. Technical replicates on the same array were averaged. Data was excluded in case where technical replicates were significantly different. Normalization was performed using the print-tip LOWESS method. The one-sample t-test was used in order to determine the genes to be regarded as regulated. Expression ratios were log-transformed to fulfil the t-test's requirement for a normal distribution. Genes for which the mean of log-ratios across the biological replicates was equal to zero at a significance level α=0.05 were considered to have an unchanged expression. On the other hand, genes having a p-value smaller than · and an average fold-change (increase or decrease) of the four data points of at least 2.0-fold were considered as modulated genes.
From among the group of significantly modulated genes, attention was focused on those showing higher than two-fold expression change constitutively in all of the tumour samples in a histopathological group.
Gene expression analysis. In the case of 5 genes, gene expression changes observed on microarray were confirmed by quantitative real-time (QRT) PCR. Total RNA from each sample (2 µg) was reverse transcribed in the presence of poly (dT) sequences and was used as template in the QRT-PCR. Primers were designed using the PrimerExpress software (Applied Biosystem, Foster City, CA, USA). Reactions were performed in a total volume of 20 µl containing 0.2 mM of dNTP, 1xPCR reaction buffer (ABGene, Epsom, UK), 5 pmol of each primer, 4 mM of MgCl2, 1x SybrGreen I (Molecular Probes, Eugene, Oregon, USA) and 0.5 units of thermostat Taq DNA polymerase (ABGene). Quantitative PCR was performed using a RotorGene 3000 instrument (Corbett Research, Sidney, Australia) according to the SybrGreen protocol. Curves were analyzed by RotoGene software (Corbett Research) using dynamic tube and slope correction methods, ignoring data from cycles close to baseline. Relative expression ratios were normalized to the geometric mean of two housekeeping genes, GAPDH and hypoxanthin phosphorybosyltransferase. Expression ratios were calculated using the Pfaffl method (5). All the PCRs were performed four times in separate runs. Regulation was confirmed in the correct direction for each of the 5 genes.
Significant expression differences of 279 genes, underexpression of 252 and overexpression of 27 genes, were found in the thyroid tumour samples of follicular adenoma, follicular carcinoma and papillary carcinoma compared to the normal thyroid tissue. Underexpressions were more expressed in all tumour samples than overexpressions and showed an increase in number in the malignant histiotypes (Figure 1). While the overexpressions were specific for each tumour histiotype (Table I), the underexpressed genes exhibited many overlaps between the tumour groups (Tables II and III). Moreover 25 out of 49 underexpressed genes found in follicular adenoma exhibited overlaps with the malignant histological types.
Analysing these overlapping genes between the benign and malignant histiotypes, the underexpression of Bruton's tyrosine kinase (BTK), G protein coupled receptor 30 (GPR 30), apelin, E-cadherin (CDH9) and CDP-diacylglycerol (CDP-DAG) were identified in follicular adenoma and follicular carcinoma, while the underexpressed genes in follicular adenoma and papillary carcinoma included interleukin 12 receptor β-2 (IL12R,-2) (Table II).
Overlaps of underexpressions in the malignant histiotypes (follicular carcinoma and papillary carcinoma) involved phospholipase D1 (PLD1) and mitogen-activated protein kinase 10 (MAPK10). We also found some novel genes down-regulated in follicular and papillary thyroid carcinoma, which had not been reported yet in thyroid malignancies, such as SOX21 and NIL-2-A. However the consequent significant underexpression of three particular genes in all investigated tumour groups was remarkable: Peroxisome proliferator-activated receptor-gamma (PPARγ), eosinophil-derived neurotoxin and urotensin II (Table III).
We found small number of constitutively up-regulated genes in the investigated tumour histiotypes and this number showed further decrease in the malignant groups, whereas these up-regulated genes were characteristic to the different tumour histiotypes. The most considerable was the EDA1 overexpression identified in follicular thyroid adenoma and the cyclin D1 (CYLD1) overexpression represented in papillary carcinoma. Among the overexpressed genes just one was found to be significantly regulated in all follicular tumour types: NF-κB.
Microarray analysis enables us to take a genome-wide cross-sectional snapshot of tumours. Analysis and comparison of these molecular snapshots makes it possible to identify characteristic steps of deregulation and dedifferentiation in tumorigenesis. Both the advances and the limitations of this genome-wide screening arise from the nature of the methodology. It can reflect a momentary picture of a global network of all transcriptional events, but is unable to represent the dynamically changing molecular process of tumour development. For that reason in this present investigation we compared the microarray pattern of early-stage thyroid tumours looking not particularly for significantly modulated candidate genes, but for a set of genes acting on similar antiapoptotic and signalling pathways, and for those showing overlaps between the early-stage epithelial thyroid tumour types. Although the role of the overlapping genes between the histologically different tumour types with a common epithelial origin has not been clarified, they might represent early or pivotal steps of deregulation in tumorigenesis.
In this investigation follicular tumour types exhibited the down-regulation of BTK, CDP diacylglycerol, GPR30 and Ecadherin. BTK and CDP diacylglycerol act on the same inositol-phospholipid signal transducing pathway, responsible for activating protein kinase C. BTK was previously demonstrated to be a mediator of oxidative stress induced apoptosis. BTK deficient B cells showed high resistance to oxidative stress-induced apoptosis (6). We found GPR 30 underexpressed both in follicular adenoma and follicular carcinoma and expression-ratios were higher in adenoma tissues. GPR 30 binds estrogen with high affinity and resides in the endoplasmatic reticulum, where it activates multiple intracellular signalling (7). GPR 30 and MAPK pathway were previously found to mediate the growth response to estrogens (17 beta esrtadiol, genistein and 4 hydroxytamoxifen) that could induce the progression of thyroid carcinogenesis (8). We recognized down-regulation of E-cadherin which was characteristic for both follicular tumour types (follicular adenoma and follicular carcinoma). E-cadherin is a cellular adhesion molecule mediating Ca2+-dependent homophilic cell-cell adhesion thus playing important role in the maintenance of normal tissue architecture (9) and its down-regulation among thyroid carcinomas has already been demonstrated (10). In a real-time quantitative analysis of Ecadherin expression in thyroid neoplasms carried out by Smyth et al. follicular thyroid carcinomas consistently had significantly decreased E-cadherin expression compared with papillary thyroid carcinomas (11).
The most remarkable overlapping gene between follicular adenoma and papillary carcinoma was the interleukin 12 receptor β-2 (IL12Rβ-2). The IL12Rβ-2 gene has previously found to be silenced in tumour cells from different human B-cell malignancies. It drives T helper responses, enhances T and NK cell cytotoxicity, and induces interferon-γ production by T and NK cells. In addition, IL-12 exerts antitumour activity through interferon-γ-dependent mechanisms (12).
We found the underexpressions phospholipase D1 (PLD1) and mitogen-activated protein kinase 10 (MAPK10) overlapping in the malignant histiotypes. These genes have previously been described and investigated in thyroid neoplastic models. MAPK10 was identified in human thyroid cells and tissue and in PCR-validated homozygous deletions it found to be a candidate tumour suppressor gene as a target specific transcriptional factor inducing apoptosis and generating tissue specific responses (13, 14). The underexpression of PLD1 seems to be contradictory at first sight as it is a downstream target of the RAS/RalA GTPase cascade implicated in mitogenic and oncogenic signalling contributing to the progression to a malignant phenotype in cells. Its expression is regulated by PKC, ADP ribosylation factors and Rho family proteins affecting signal transduction, vesicular trafficking and cytoskeletal reorganization, implicated in the metastatic potential of malignancies (15, 16). We identified PLD1 down-regulation both in follicular carcinoma and papillary carcinoma. A possible reason for this result might be the early developmental stage of the tumour samples. On clinical and histopathological evaluation, neither of the tumour sample types showed metastatic features either locally or systemically.
There were some novel genes among the consequently down-regulated ones that we identified in the groups of follicular and papillary thyroid carcinoma, which had not been reported yet in thyroid malignancies, such as SOX21 and NIL-2-A. Members of the SOX family are transcription factors required for development and differentiation of multiple cell lineages (17). NIL-2-A encodes a zinc finger protein inhibiting T lymphocyte specific interleukin-2 gene expression (18). The role of these genes in thyroid neoplasms has not been investigated yet.
We found a consequent down-regulation of three particular genes in all investigated tumour histiotypes: Peroxisome proliferator-activated receptor-gamma (PPARA), eosinophil-derived neurotoxin and urotensin II. PPARγ is a member of a nuclear hormone receptor subfamily of transcription factors and found to be a promoter-specific repressor of nuclear factor of kappa light polypeptide gene enhancer in B-cells (NF-κB) (19). In a mouse model for thyroid carcinogenesis, Kato Y et al. found that reduced PPARγ protein abundance led to the activation of NF-κB signalling pathway and resulted in repression of critical genes involved in apoptosis and promotion of follicular thyroid carcinogenesis (20). There is also evidence showing that the PPARγ transcriptional pathway can include terminal differentiation of malignant breast epithelial cells and prostate adenocarcinoma cells (21). Eosinophil-derived neurotoxin is a cytosolic ribonuclease exhibiting tumour cell toxicities and telomerase degradative activity in Kaposi's sarcoma and Hodgkin's lymphoma and found to be useful for targeted therapy in both cancer types (22-23). The role of urotensin II in carcinogenesis is unclear. In a study UII-receptor activation was functionally coupled to phospholipase C-mediated Ca2+ mobilization (24).
Among the few constitutively up-regulated genes found in this investigation, some characteristic ones such as cyclin D1 (CYLD1), were represented in papillary carcinoma, and EDA1 in follicular adenoma. The CYLD1 gene has previously been identified in thyroid carcinogenesis in both papillary (25) and follicular thyroid carcinoma (26) with excellent specificity, furthermore its regulatory role on NF-κB signalling has already been recognised and currently researched. EDA1 was found to be a member of the TNF-related ligand family regulating epithelial development (27). It activates NF-κBpromoted transcription and has a regulatory role in JNK (including MAPK10) pathways (28). Among the overexpressed genes we found just one to be significantly regulated in all follicular tumour types: NF-κB. NF-κB belongs to the category of rapidly acting primary transcriptional factors, acting as a first responder to a wide variety of cell surface receptor stimuli regulating genes that control cell proliferation and cell survival. In most cell types, NF-κB is retained in the cytoplasm in its inactive form. A wide variety of external stimuli (including TNF, IL-1, PMA, H2O2, endotoxin and hypoxia) can induce its activation. Most of these agents induce NF-κB to translocate to the nucleus, where it activates transcription of target genes partaking in the suppression of apoptosis. Its activity was constitutively elevated in primary human thyroid carcinomas and correlated malignant phenotype (29-31).
Activated NF-κB was also detected in an in vitro model of human thyroid cancer that resembles the in vivo differentiated and undifferentiated thyroid tumours. In these cell lines persistent NF-κB activity was progressively detected in papillary thyroid carcinoma cells to follicular carcinoma cells until reaching the highest levels in anaplastic carcinoma cells reflecting the sustained activation of NF-κB confers an advantage for clonal selectivity (32).
All the histiotypes of tumours investigated here contained significantly modulated genes acting on the NF-κB regulatory pathway (Figure 2). In follicular histiotypes (follicular adenoma and follicular carcinoma) NF-κB was found to be up-regulated. Follicular thyroid adenomas exhibited the overexpression of EDA1, wich has a regulatory role in NF-κB-promoted transcription and JNK signalling. Malignant phenotypes (follicular thyroid carcinoma and papillary thyroid carcinoma) additionally represented constitutive down-regulation of genes (PPARγ and MAPK10/JNK) partaking in NF-κB inhibition. Furthermore in papillary thyroid carcinoma another regulatory gene (CYLD1) was consequently up-regulated, which is closely connected to NF-κB signalling. Our findings support the previous assumptions that modulation of NF-κB signal transduction plays a crucial role in early thyroid carcinogenesis and further investigations could possibly open the door to diagnostic, therapeutic or preventative perspectives.
1 Harrison's Principles of Internal Medicine 15th edition. Braunwald E, Fauci A, Kasper DL, Hauser SL, Longo DL and Jameson JL, McGraw-Hill Company Medical Publishing Division, New York, pp. 2079, 2001.
2 Barden CB, Shister KW, Zhu B, Guiter G, Greenblatt DY, Zeiger MA and Fahey III TJ: Classification of follicular thyroid tumors by molecular signature: results of gene profiling. Clin Cancer Res 9: 1792-1800, 2003.
3 Finley DJ, Arora N, Zhu B, Gallagher L and Fahey III TJ:
Molecular profiling distinguishes papillary carcinoma from benign
thyroid nodules. J Clin Endocrinol Metab 89: 3214-3223, 2004.
4 Gebauer G, Krones-Herzig A, Glinskii AB and Glinsky GV: Identification of genes associated with tumor progression using microarrays. Anticancer Res 25: 1477-1482, 2005.
5 Pfaffl MW, Horgan GW and Dempfle L: Relative expression software tool (REST©) for group-wise comparison and statistical analysis of relative expression results in real-time PCR. Nucleic Acids Res 30(9): e36, 2002.
6 Uckun F, Ozer Z and Vassilev A: Bruton's tyrosine kinase prevents activation of the anti-apoptotic transcription factor STAT3 and promotes apoptosis in neoplastic B-cells and B-cell precursors exposed to oxidative stress. Br J Haematol 136: 574589, 2007.
7 Bologa CG, Revankar CM, Young SM, Edwards BS, Arterburn JB, Kiselyov AS, Parker MA, Tkachenko SE, Savchuck NP, Sklar LA, Oprea TI and Prossnitz ER: Virtual and biomolecular screening converge on a selective agonist for GPR30. Nat Chem Biol 2: 207-212, 2006.
8 Vivacqua A, Bonofiglio D, Albanito L, Madeo A, Rago V, Carpino A, Musti AM, Picard D, Ando S and Maggiolini M: 17 ,-Estradiol, genistein, and 4-hydroxytamoxifen induce the proliferation of thyroid cancer cells through the G protein-coupled receptor GPR30. Mol Pharmacol 70: 1414-1423, 2006.
9 Sato H, Ino Y, Miura A, Abe Y, Sakai H, Ito K and Hirohashi
S: Dysadherin: expression and clinical significance in thyroid carcinoma. J Clin Endocrinol Metab 88: 4407-4412, 2003.
10 Mitselou A, Ioachim E, Peschos D, Charalabopoulos K, Michael M, Agnantis NJ and Vougiouklakis T: E-cadherin adhesion molecule and syndecan-1 expression in various thyroid pathologies. Exp Oncol 29: 54-60, 2007.
11 Smyth P, Sheils O, Finn S, Martin C, O'Leary J and Sweeney EC: Real-time quantitative analysis of E-cadherin expression in ret/PTC-1-activated thyroid neoplasms. Int J Surg Pathol 9: 265272, 2001.
12 Airoldi I, Di Carlo E, Cocco C, Sorrentino C, Fais F, Cilli M, D'Antuono T, Colombo MP and Pistoia V: Lack of Il2rb2 signaling predisposes to spontaneous autoimmunity and malignancy. Blood 106: 3846-3853, 2005.
13 Lapouge G, Millon R, Muller D, Abecassis J, Eber M, Bergerat JP and Klein-Soyer C: Cisplatin-induced genes as potential markers for thyroid cancer. Cell Mol Life Sci 62: 53-64, 2005.
14 Kim YH, Girard L, Giacomini CP, Wang P, Hernandez-Boussard T, Tibshirani R, Minna JD and Pollack JR: Combined microarray analysis of small cell lung cancer reveals altered apoptotic balance and distinct expression signatures of MYC family gene amplification. Oncogene 25: 130-138, 2006.
15 Joseph T, Wooden R, Bryant A, Zhong M, Lu Z and Foster DA: Transformation of cells overexpressiong a tyrosine kinase by phospholipase D1 and D2. Biochem Biophys Res Commun
289: 1019-1024, 2001.
16 Riebeling C, Muller C and Geilen CC: Expression and regulation of phospholipase D isoenzymes in human melanoma cells and primary melanocytes. Melanoma Res 13: 555-562, 2003.
17 Wang H, McKnight NC, Zhang T, Lu ML, Balk SP and Yuan
X: SOX9 is expressed in normal prostate basal cells and regulates androgen receptor expression in prostate cancer cells. Cancer Res 67: 528-536, 2007.
18 Williams TM, Montoya G, Wu Y, Eddy RL, Byers MG and Shows TB: The TCF8 gene ancoding a zinc finger protein (Nil2-a) resides on human chromosome 10p11. 2. Genomics 14: 194-196, 1992.
19 Pascual G, Fong AL, Ogawa S, Gamliel A, Li AC, Perissi V, Rose DW, Willson TM, Rosenfeld MG and Glass CK: A SUMOylation-dependent pathway mediates transrepression of inflammatory response genes by PPAR-A. Nature 437: 759763, 2005.
20 Kato Y, Ying H, Zhao L, Furuya F, Araki O, Willingham MC and Cheng SY: PPARA insufficiency promotes follicular thyroid carcinogenesis via activation of the nuclear factor-kappaB signaling pathway. Oncogene 25: 2736-2747, 2006.
21 Mueller E, Sarraf P, Tontonoz P, Evans RM, Martin KJ, Zhang M, Fletcher C, Singer S and Spiegelman BM: Terminal differentiation of human breast cancer through PPARA. Mol Cell 3: 465-470, 1998
22 Norrback KF, Enblad G, Erlanson M, Sundstrom C and Roos
G: Telomerase activity in Hodgkin's disease. Blood 92: 567573, 1998.
23 Suzuki M, Saxena SK, Boix E, Prill RJ, Vasandani VM, Ladner JE, Sung C and Youle RJ: Engineering receptor-mediated cytotoxicity into human ribonucleases by steric blockade of inhibitor interaction. Nat Biotechnol 17: 265-270, 1999.
24 Pati S, Lee Y and Samaniego F: Urinary proteins with proapoptotic and antitumor activity. Apoptosis 5: 21-28, 2000.
25 Meirmanov S, Nakashima M, Kondo H, Matsufuji R, Takamura N, Ishigaki K, Ito M, Prouglo Y, Yamashita S and Sekine I: Correlation of cytoplasmic beta-catenin and cyclin D1 overexpression during thyroid carcinogenesis around Semipalatinsk nuclear test site. Thyroid 13: 537-545, 2003.
26 Moyano L, Franco C, Carreno L, Robinson P and Sanchez G: HBME-1 and cyclin D1 as diagnostic markers for follicular thyroid carcinoma. Rev Med Chil 132: 279-284, 2004.
27 Ezer S, Bayes M, Elomaa O, Schlessinger D and Kere J: Ectodysplasin is a collagenous trimeric type II membrane protein with a tumor necrosis factor-like domain and colocalizes with cytoskeletal structures at lateral and apical surfaces of cells. Hum Mol Genet 8: 2079-2086, 1999.
28 Cui CY, Durmowicz M, Tanaka TS, Hartung AJ, Tezuka T, Hashimoto K, Ko MS, Srivastava AK and Schlessinger D: EDA targets revealed by skin gene expression profiles of wild-type, Tabby and Tabby EDA-A1 transgenic mice. Hum Mol Genet
11: 1763-1773, 2002.
29 Ichikawa H, Takada Y, Murakami A and Aggarwal BB: Identification of a novel blocker of IIBalpha kinase that enhances cellular apoptosis and inhibits cellular invasion through suppression of NF-IB regulated gene products. J Immun 174: 7383-7392, 2005.
30 Aggarwal BB: Nuclear factor IB: the enemy within. Cancer Cell
6: 203-208, 2004.
31 Palona I, Namba H, Mitsutake N, Starenki D, Podtcheko A, Sedliarou I, Ohtsuru A, Saenko V, Nagayama Y, Umezawa K, Yamashita S: BRAFV600E promotes invasiveness of thyroid cancer cells through nuclear factor IB activation. Endocrinology
147: 5699-5707, 2006.
32 Pacifico F, Mauro C, Barone C, Crescenzi E, Mellone S, Monaco M, Chiappetta G, Terrazzano G, Liguoro D, Vito P, Consiglio E, Formisano S and Leonardi A: Oncogenic and antiapoptotic activity of NF-IB in human thyroid carcinomas. J Biol Chem 52: 54610-54619, 2004.
Received August 24, 2007
Revised November 14, 2007
Accepted Novembert 20, 2007