Thyroid nodules are extremely common with prevalence rates approaching 50-60% of adults under 60 years old. Because only approximately 5% of thyroid nodules are malignant, accurate pre-operative characterization of thyroid nodules is critical in selecting patients appropriate for surgical thyroidectomy. Fine needle aspiration (FNA) is the single most important diagnostic procedure in the evaluation of thyroid nodules.
Table 1. Molecular markers for thyroid nodules and lymph nodes:
Potential diagnostic markers for thyroid nodules and lymph nodes
Telomerase Galectin-3 Thyroid peroxidase Thyroglobulin Oncofetal fibronectin ret/PTC? oncogenes PaxH/PPARy oncogene B-Raf mutations Nm23
High mobility Group I (Y) Protein
GLUT-1 CA 19-9 CD 15 HUME-1 CD 30 CD 57 CD 97 Lcu-7
Epithelial Membrane Antigen Cyclooxygenase 2 Cytokcratin 19 Oytokeratin 20
For small, solid nodules, experienced cytopathologists can accurately distinguish most benign nodules and papillary cancers. However, cytological features do not distinguish benign from malignant follicular neoplasms, and cystic papillary thyroid cancers are a common cause of false negative results. Importantly, only 15% of cytological follicular neoplasm will ultimately be follicular carcinomas; therefore, 85% of individuals that undergo surgery for these nodules will have done so unnecessarily. Finally, by its nature, cytopathologic interpretation of FNA samples is subjective. For these reasons, the application of molecular analysis to better characterize thyroid nodule cytologic samples has been an area of intense interest.
With the advent of methods, such as reverse transcriptase-polymerase chain reaction (RT-PCR), in which tiny amounts of samples are suitable for analysis, and increases in the number of antibodies suitable for immunocytochemistry, the possibility of improving FNA-based characterization of thyroid nodules is now possible. In the initial section of this review, several of the most carefully studied molecular markers (Table 1) for thyroid FNA will be discussed.
Telomeres are chromosomal end structures, consisting of tandem repeats of TTAGGG that play a critical role in the protection of chromosomes during cell division and are important in chromosome positioning during replication (1). Chromosomes typically lose about 50 to 200 nucleotides of telomeric sequence from chromosomal ends per cell division because DNA polymerase is unable to replicate the ends of linear DNA. The resultant progressive shortening of chromosomes as cells divide has been described a cellular "biological clock"; once the chromosomes are shortened to a critical length through telomeric loss, cell growth stops and apoptosis is induced. Therefore, preservation of chromosomal end length during division would be expected to retard this natural "aging" of cells and result in continuous cell growth.
Telomerase is an enzyme that extends telomeres, thereby preserving chromosomal length. Structurally, telomerase is a ribonucleic acid-protein complex containing a catalytic component, the human telomerase reverse transcriptase (hTERT) (2). Telom-erase expression and activity have been identified in immortalized human cell lines and cells and in bone marrow cells that normally divide, but not in normal human adult epithelial cells. However, expression of telomerase and demonstration of telomerase activity using a PCR-based assay (TRAP; Telomeric Repeat Amplification Protocol) has been described in a variety of malignancies and other dividing cells, such as germinal cells of the ovary and testis (1). Based on these results, detection of telom-erase activity and expression of hTERT have been explored as potential distinguishing markers for the presence of malignant thyroid cells.
Saji et al. (3) evaluated surgical pathology samples from thirty papillary thyroid cancers, three benign nodules and ten normal thyroid specimens for telomerase activity by TRAP and found that 67% of the malignant tissues had telomerase activity compared to 0% of the benign nodules. In this study, 64% of the papillary thyroid cancers that had a non-diagnostic preoperative FNA were positive for telomerase activity. Haugen et al. (4) and Onoda, et al. (5) obtained similar results when they investigated surgical thyroid specimens.
The ability of telomerase help differentiate follicular adenoma from carcinoma was reported by Umbricht et al. (6) who studied frozen tissue samples from patients undergoing thyroidectomy for follicular neoplasm on FNA. TRAP assays were performed on 44 follicular thyroid tissue specimens and 22 normal thyroid tissue samples. The authors reported a sensitivity of 100% and a specificity of 76% for detecting follicular carcinoma. The false positive samples occurred mostly in tumors that also had lym-phocytic infiltration, as lymphocytes are known to have detectable telomerase activity. However, it is possible that the presence of telomerase activity in histologically benign specimens may represent an early step in the development of an invasive tumor. This issue of potentially creating an assay that is more sensitive than the clinical gold standard is a problem for many RT-PCR-based assays.
De Deken et al. (7) demonstrated a decrease of telomere length as well as increased variability in the telomere size in benign nodules without measurable telomerase activity, as compared to normal thyroid tissues. These data suggest that the benign nodular cells may have progressed through more mitotic divisions than the adjacent normal tissue and therefore may be closer to their limit for further growth, consistent with benign tumor growth.
The cloning of the human telomerase reverse transcriptase cDNA allowed for RT-PCR analysis of expression of its mRNA in clinical samples. This created an opportunity to design a more user-friendly telomerase assay that could be applied to FNA samples. Saji et al. (8) studied 19 malignant and 18 benign thyroid surgical samples for evidence of hTERT gene expression by RT-PCR. HTERT mRNA was detected in 15 (79%) of the malignant and 5 (28%) of the benign tumors; all 5 benign lesions with demonstrable hTERT gene expression had lymphocytic infiltration on final pathology. HTERT mRNA was not detected in any of the normal thyroid specimens. The results correlated with TRAP assay results from the same samples. Similar results were reported by the same authors using thyroid FNA samples (9).
Fine needle aspiration specimens have also been investigated for telomerase activity using TRAP assay. Sebesta et al. (10) failed to show any additional usefulness of measuring telomerase activity in a small study of FNA samples. It is likely that RT-PCR of hTERT is more sensitive than TRAP assay due to the logarithmic amplification inherent in RT-PCR, thus accounting for a greater sensitivity when used to analyze FNA samples.
The frequency of telomerase-positive results, either by RT-PCR or TRAP assay for papillary thyroid cancer varies between studies. For example, some report that as many as 67% of papillary thyroid cancers are positive (3), while others indicate a much lower percentage of about 20% (11). The small number of cases in many of these studies makes interpretation quite difficult. Similarly, results regarding the association between telomerase activity and tumor aggressiveness also vary; some studies demonstrate a correlation between telomerase activity and tumor progression (5, 12) while others do not (3). It is clear that larger, more extensive studies are needed before telomerase can be considered an effective diagnostic tool.
Galectin-3 is a member of the lectin family that regulates the functions of its protein targets by interacting with attached galactose-containing glycoprotein side chains. As a group, galectins regulate cell growth and differentiation, intercellular recognition and adhesion, as well as malignant transformation. Galectin-3 levels have been directly correlated with metastatic potential in fibrosarcoma and melanoma cell lines. In vitro studies have indicated that expression of galectin proteins is elevated in thyroid cancer cell lines and microarray analysis has demonstrated increased levels of galectin-3 in papillary thyroid cancers (13) This led to their investigation as molecular markers used to distinguish between benign and malignant thyroid tissues.
Xu, et al. (14) evaluated protein derived from 41 surgical thyroid specimens for galectin-1 and galectin-3 expression by Western blot. Elevated levels of both proteins were demonstrated in thyroid cancer compared to normal thyroid tissue. Similarly, normal thyroid tissue did not express galectin 1 or 3 by immunohistochemical analysis, but high levels of both proteins were detected in both papillary and follicular thyroid cancers, and in regional nodal metastases. A second group evaluated 41 malignant and 35 benign thyroid tissue specimens (15) for both galectin-3 protein and RNA levels. Galectin-3 expression was identified in 18 of 18 of papillary cancer samples, 4 of 8 follicular cancers, 2 of 3 poorly differentiated cancers, 5 of 5 anaplastic cancers, 3 of 6 medullary, and 1 of 1 Hurthle cell cancers. By contrast, none of the normal or benign nodular tissues expressed galectin-3, other than those with lymphocytic infiltration. Levels of galectin-3 mRNA appeared to correlate with protein levels in papillary cancers and normal tissue (15).
To determine if galectin-3 immunocytochemistry could be applied to FNA samples, Orlandi, et al. (16) evaluated FNA and surgical pathology specimens from 64 patients who had undergone thyroidectomy whose preoperative diagnosis was malignant (n = 15), indeterminate (n = 37), and benign (n = 12). The final histo-logic diagnosis included 18 papillary and 17 follicular cancers, as well as 29 follicular adenomas. All papillary thyroid cancers expressed galectin-3 in both FNA and surgical specimens. For the follicular cancers, immunoactive galectin-3 was detected in all surgical specimens in a heterogeneous pattern, and in all but 3 FNA samples. By contrast, only 3 of 29 benign follicular adenomas expressed galectin-3.
In a more recent study (17), different antibodies against human galectin-3 were used in an immunohistochemical study of thyroid surgical specimens (13 benign and 62 malignant). Immunoactive galectin-3 was most prevalent in the papillary thyroid cancers (33 of 45), but some benign lesions were 3 of 8 benign adenomas demonstrated immunoactive galectin-3.
Finally, Bernet, et al. (18) applied quantitative RT-PCR to galectin-3 analysis to determine if a particular "cut-point" of galectin-3 gene expression correlated best with malignancy. In this study, markedly elevated levels of galectin-3 mRNA were identified in papillary cancers compared with normal tissue. There was no difference between the galectin-3 mRNA levels in follicular adenomas and carcinomas.
Based on the above results, galectin-3 immunocytochemistry seems to be a promising new marker of thyroid cancer that could be applied to FNA analysis. It appears that classic molecular approaches, such as quantitative RT-PCR may not be helpful for the conundrum of follicular neoplasm FNA results. However, additional studies are still required.
Thyroid peroxidase (TPO) is a thyroid-specific enzyme that catalyzes iodide oxidation, thyroglobulin iodination, and iodothyronine coupling. Reduced expression of TPO impairs thyroid follicular cell function correlates with a loss of differentiated thyroid function and has been well described in thyroid cancer cell lines and tumor samples. Thus, immunohistochemical staining for TPO expression and molecular analysis of the TPO gene have been studied for use as diagnostic tools for thyroid cancer.
DeMicco, et al. reported a retrospective study of 150 FNA samples including 125 benign tissues (19), and demonstrated that 113 of 125 benign lesions were characterized by immunoactive TPO in more than 80% of cells while <80% of the cells expressed TPO in all 25 malignant lesions. Thus, using this level of TPO-expressing cells as a positive, they reported a sensitivity of 100% and a specificity of 90%.
Christensen, et al. reported their prospective experience using this method in 124 consecutive FNAs using the same anti-TPO primary antibody (20). In their hands, TPO immunohistochemistry (>80% cut-off) correctly identified all cases of cancer. Only one benign follicular adenoma was identified as malignant by this immunohisto-chemical criterion. These investigators concluded that TPO immunohistochemistry of FNA samples using the 80% cut-off values has a sensitivity of 100% and a specificity of 99%. These results are obviously subjective and may be antibody dependent.
Because germline mutations of the TPO gene that cause functional loss of TPO activity cause of congenital hypothyroidism, loss of heterozygocity (LOH) at the TPO gene locus has been implicated as a cause of the organification defect typical of benign and malignant thyroid tumors. However, in a study of 40 hypoactive thyroid nodules (21), LOH of the TPO gene was noted in only 6, making this an unlikely method for evaluating thyroid nodules preoperatively.
Thus, it appears that immunostaining for thyroid peroxidase may be a valuable addition to the analysis of FNA samples. Studies with additional available antibodies may be useful from a practical standpoint.
Fibronectins are high-molecular-weight glycoproteins found in the extracellular matrix. Oncofetal fibronectin is characterized by the presence of the oncofetal domain (IIICS domain), which is absent in normal fibronectin. Overexpression of this variant of fibronectin has been demonstrated in many epithelial cancers and it has been studied as a molecular marker of malignancy. Several investigators have evaluated the utility of the oncofetal fibronectin mRNA as a marker of thyroid malignancy.
Higashiyama, et al. (22) evaluated 19 malignant and 33 benign surgical thyroid specimens by competitive RT-PCR and demonstrated elevated levels in papillary and anaplastic cancers versus benign tissues. Levels were variable in follicular carcinomas and were not clearly different from follicular adenomas. The same group also reported detection of oncofetal fibronectin mRNA on surgical samples using in situ hybridization and reported similar results (23).
Takano et al. (24) examined 72 FNA samples (23 normal, 14 adenomatous goiters, 13 follicular adenomas, 3 follicular carcinomas, 18 papillary carcinomas and 1 anaplastic cancer) for expression of oncofetal fibronectin mRNA using RT-PCR. 95% of the papillary or anaplastic carcinomas by cytology also expressed oncofetal fibronectin mRNA compared to only 4% (n = 109) of benign specimens. In contrast, none of the 6 follicular tumors expressed oncofetal fibronectin. Fifty of these patients underwent surgery, based on the results of the surgical histology, oncofetal fibronectin RT-PCR was 97% sensitive and 100% specific. These results are similar to Higashiyama, et al. as all but one cancer sample included in this study was papillary. These results suggested that oncofetal fibronectin mRNA amplification was an accurate marker of papillary, but not follicular carcinoma. A potential cause of false positive results is the expression of oncofetal fibronectin in thyroid fibroblasts (25). Despite the fibroblast data, the results of the immunhistochemical and molecular studies suggest that measurement of oncofetal fibronectin expression may be useful as an adjunctive test for identifying papillary thyroid carcinoma.
Ret/PTC oncogenes are genomic rearrangments that couple the tyrosine kinase domain of the Ret receptor to different regions leading to aberrant expression and activation of Ret. To date, there are 8 Ret/PTC proteins, however, the prevalence is greatest for Ret/PTC 1, 2, and 3. Translocations involving Ret are particularly prevalent in papillary carcinomas that develop following exposure to radiation. Because these rearrangements are largely limited to thyroid carcinomas, the expression of PTC oncogenes has been studied as molecular markers for thyroid malignancy.
In a study of 73 thyroid specimens from which both FNA and surgically obtained tissue was available, Cheung, et al. (26) evaluated the presence of PTC1-5 by RT-PCR. Only Ret/PTC 1, 2 or 3 were detected in the samples; Ret/PTC translocations were not detected on FNA and surgical samples from 39 benign tissue samples, including 11 follicular adenomas, 25 nodular hyperplasia's and 3 Hashimoto's thyroiditis cases. In contrast, Ret/PTC 1, 2, or 3 expression was detected in 17 FNA samples and 21 surgical specimens derived from 33 malignant thyroid tumors. Of importance, this molecular method was more accurate than routine cytopathology in these samples.
Conflicting results were reported by Elisei, et al. (27) who studied 154 patients referred to surgery for FNA-characterized benign nodules (n = 65) or papillary thyroid cancer (n = 89). Expression of Ret/PTC-1 and Ret/PTC-3, the most common Ret/PTC oncogenes, was identified in both benign and malignant nodules. RET protein expression has been evaluated by immunohistochemistry in papillary thyroid cancers (28). Overall, expression of Ret was heterogenous and was demonstrated in regions of cellular atypia in both malignant and benign lesions. Thus, based on these data, it appears that Ret/PTC may not be helpful in pre-operative diagnosis due to a relatively low prevalence in many populations with papillary thyroid cancer and potential issues with specificity. However, more studies are needed to clarify a role for Ret/PTC rearrangement or Ret overexpression in the diagnosis of thyroid nodules.
Kroll, et al. (29) identified a chromosomal translocation t(2;3)(q13;p25) causing a fusion gene between Pax8 and the peroxisome proliferator activated receptor gamma in follicular thyroid carcinomas. Specifically, 5 of 8 follicular cancers expressed the fusion gene, while all of the 20 follicular adenomas, 10 papillary thyroid carcinomas and 10 other benign nodules did not express the rearranged gene, suggesting that detection of fusion gene expression might accurately identify follicular carcinomas preoperatively.
The specificity of the Pax 8-PPARy may not be complete, as other groups (30, 31) have reported expression of in benign follicular adenomas, albeit at a lower frequency than follicular carcinomas. The importance of expression of Pax8-in follicular adenomas on malignant transformation is uncertain. It has been speculated that overexpression of alone, even in the absence of a defined chromosomal rearrangement, may be a marker of malignant transformation. Detection of overexpression by immunohistochemistry appears to be more sensitive, but also, less specific for detection of follicular carcinoma (31).
Mutations in the serine-threonine kinase, B-Raf have been described in 35-70% of papillary thyroid carcinomas, with almost no overlap with other known oncogenes or other benign or malignant thyroid lesions (32-34). Because this mutation appears quite specific for papillary thyroid cancer, and it is limited to two specific mutations, detection of the mutations has been proposed as an adjunctive test for FNA analysis (32). This method would likely be useful only for papillary thyroid cancer detection, however.
Re-expression of the Nm23 tumor suppressor gene has been demonstrate to reduce the metastatic potential of malignant cells in-vitro and reduced expression of Nm23 occurs in aggressive forms of breast cancer (35) . In thyroid tissues, the interesting finding of increased expression has been demonstrated, primarily in stage IV papillary cancers and anaplastic carcinomas (36). Farley et al. (37) also evaluated 34 thyroid tumors, including 4 follicular adenomas, 19 papillary carcinomas, 6 follicular carcinomas and 5 medullary carcinomas for Nm23 mRNA levels. In this study, overexpression of Nm23 was noted in follicular and medullary cancers, although there was overlap between benign and malignant samples. Similarly, Berthau, et al. (38) reported that immunocytochemical analysis of Nm23 protein expression did not accurately distinguish between benign and malignant lesions. Mechanistically, the finding that overexpression of nm23, rather than reduction of loss of nm23 expression were demonstrated suggests an alternative function for this protein in thyroid cancer (39).
High mobility group I(Y) protein—HMGI(Y)
The high mobility group I (HMGI) proteins are nuclear proteins that regulate chromatin structure and function. HMGI(Y) is particularly highly expressed during embryogenesis, and its reexpression has been described in cancers, but not in normal adult tissues. Chiappetta et al. (40) reported evaluated expression of HMGI(Y) protein by immunohistochemistry on 358 thyroid tissue samples. HMGI(Y) was detected in 18 of 19 follicular carcinomas, 92 of 96 papillary tumors and 11 of 11 anaplastic cancers, but in only 1 of 20 hyperplastic nodules, 44 of 200 benign follicular adenomas and 0 of 12 normal thyroid tissue samples. HMGI(Y) mRNA was detected in 4 of 4 malignant tumors while eight benign FNA samples (6 follicular adenomas and 2 normal thyroid tissue) were negative. Thus, HMGI(Y) may be a potentially useful diagnostic tool for thyroid cancer that warrants further identification.
Because ceruloplasmin, a copper transport protein that shares homology with lacto-ferrin (a molecular marker for several tumor types), it has been investigated as a tumor marker in thyroid cancer. Tuccari et al. (41) evaluated 56 surgical thyroid specimens for ceruloplasmin expression by immunohistochemistry. None of the 15 follicular adenomas expressed ceruloplasmin, while two of two Hurthle cell tumors, all 21 follicular, and all 6 papillary carcinomas were positive. All of the medullary thyroid cancers were negative for ceruloplasmin, as was the normal thyroid tissue surrounding the thyroid cancers. The functional role of ceruloplasmin in thyroid tumors as its potential role as a marker for malignancy require further clarification.
Cytokeratins are structural proteins found in all epithelial cells; several types of keratins have been identified with altered expression patterns in malignancies. In thyroid cancer, immunocytochemical expression for prekeratin was detected in papillary thyroid cancer but not normal thyroid tissues, follicular adenomas and follicular thyroid carcinomas (42). With the development of more specific antibodies that identify cytok-eratin subtypes, a more comprehensive evaluation was able to be performed. Schelfhout et al. (43) used monoclonal antibodies against cytokeratin 8, 18 and 19 to characterize cytokeratin expression in different thyroid histologies. Of these, cytokeratin-19 was overexpressed 12 of 12 papillary cancers, while follicular cancers, follicular adenomas, colloid nodules and normal thyroid tissue were negative or had only weak staining. The authors concluded that staining with antibodies against cytokeratin 19 is a useful diagnostic tool for papillary thyroid cancer. However, these promising results were not able to be confirmed. Sahoo, et al. (44) evaluated 35 surgical thyroid specimens for cytokeratin 19 expression. Although papillary cancers tended to display more intense staining than other tumors, the presence or absence of immunoactive cytokeratin 19 did not distinguish the tumor histologic subtypes. Technical issues could account for the discrepant results and further studies are needed. Cytokeratin 20 has also been evaluated in lymph nodes and peripheral blood of patients with medullary and follicular cell-derived thyroid cancer (see below).
Was this article helpful?
Learning About 10 Ways Fight Off Cancer Can Have Amazing Benefits For Your Life The Best Tips On How To Keep This Killer At Bay Discovering that you or a loved one has cancer can be utterly terrifying. All the same, once you comprehend the causes of cancer and learn how to reverse those causes, you or your loved one may have more than a fighting chance of beating out cancer.