Ratio of Metalloproteinase 2 to Tissue Inhibitor of Metalloproteinase 2 in Medullary Thyroid Carcinoma

Objectives  To develop an index for the ratio of metalloproteinase 2 (MMP-2) to its tissue inhibitor (TIMP-2) in immunostained medullary thyroid carcinoma specimens and to correlate it with clinical and pathologic prognostic factors. Metalloproteinases, enzymes related to the degradation of the extracellular matrix, take part in carcinogenesis and have been associated with the prognosis of neoplasias. Nevertheless, medullary carcinoma is rarely considered in research analysis. Researchers tend to favor the ratio of enzymes to their inhibitors over the absolute concentrations of these enzymes.

Design  Retrospective study of surgical samples.

Setting  Head and Neck Surgery and Endocrinology Departments, Universidade de São Paulo Medical School Hospital.

Patients  Surgical specimens from 33 patients who had been observed for a mean of 76.8 months (range, 4-201 months) were immunohistochemically stained for MMP-2 and TIMP-2. Only patients whose clinical and pathologic data were complete and whose specimens were preserved were included in the study.

Main Outcome Measures  The ratio between the expressions of MMP-2 and TIMP-2 was based on a staining index (immunostaining extent and intensity) of each of the markers.

Results  Proportionally large expressions of TIMP-2 over MMP-2 correlated with low occurrences of positive findings on initial cervical examination for the presence of thyroid nodules and/or lymphadenopathy (P = .02) and cervical lymph node metastases (P < .001), conditions correlated with prognosis. A correlation with cure at the end of follow-up (P = .01) was also observed. (P< .05 was considered statistically significant.)

Conclusion  The ratio of MMP-2 to TIMP-2 expression is an additional and novel prognostic predictor of the outcome of medullary carcinoma treated surgically.

Metalloproteinases (MMPs) are enzymes with a wide range of physiologic functions, and their role, free of regulatory mechanisms, in the spread of cancer is of great interest.¹ Their actions can be direct (eg, when they promote local and metastatic dissemination through the degradation of the extracellular matrix and basal membrane) or indirect (eg, when they induce angiogenesis). The most important role of these enzymes in neoplastic invasion, however, is not restricted to the creation of pathways through the extracellular matrix but in the rearrangement of matrix components to better accommodate cellular migration.² Different MMPs interact among themselves and participate in all stages of carcinogenesis. Their ability to degrade allows each tumor cell to interact appropriately with its surroundings.¹

Metalloproteinases are inactivated by tissue inhibitors of metalloproteinases (TIMPs) through the formation of complex inhibitors,³ which are important for establishing a balance between the synthesis and degradation of the extracellular matrix. While TIMPs were originally thought to work against the progression of neoplasia, this is not always the case. They perform multiple biological functions, including activities that promote cell growth.³ Metalloproteinase 2 (MMP-2) is the main and possibly most important MMP active in neoplasms. It is a marker of the malignant phenotype,^4-6 and TIMP-2 is its specific tissue inhibitor. At certain concentrations, TIMP-2 also inhibits membrane type 1 MMP, an enzyme directly involved in the activation of MMP-2; TIMP-2 affects the function of MMP-2 by preventing its activation in addition to inhibiting it directly.¹ Researchers have favored relative quantities of MMPs and TIMPs in addition to absolute quantities in interpreting data on malignant neoplasias.^7-9

Research has shown the presence of MMP-2 and TIMP-2 in thyroid tissue and has demonstrated an association with the prognosis of patients with thyroid neoplasia.^3,6,10-17 The greater emphasis, however, has been placed on epithelial tissue tumors. In a preliminary study,¹⁸ our research group observed the presence of MMP-2 and TIMP-2 in medullary thyroid carcinoma (MTC) specimens and validated the novel use of immunohistochemical staining for MMP-2 as a prognosis marker. The data obtained in this preliminary analysis were used to create an index that correlated the proportion of expressions of these 2 tumor markers with the prognostic and clinical evolution indexes for patients who underwent surgical treatment for MTC.

Studies conducted on MTC prognostic factors are varied, and their conclusions are diverse. The outcomes of patients, even if appropriately treated, are difficult to predict regardless of whether persistent or relapsing diseases are present.^19,20 These aspects drive the search for additional prognostic factors. We intend to gain a better understanding of the molecular biology of MTC by using immunohistochemical expressions of the selected proteases in tumor specimens in an effort to identify new prognosis predictors.

Thirty-three surgically treated patients with MTC were selected for this retrospective study. They were observed in the Head and Neck Surgery and Endocrinology Departments of the Universidade de São Paulo Medical School Hospital in Brazil. Only patients whose clinical and pathologic data were found in their medical records and whose specimens were available and reasonably well preserved were included in the study. The mean follow-up time was 76.8 months (range, 4-201 months; median, 61 months) between June 1982 and February 2006. The study was approved by the ethics committee for human studies.

The patients were aged between 9 and 65 years, with a mean (SD) age of 33.58 (14.39) (median age, 33 years). There were 22 female patients (67%). Nineteen patients had sporadic disease (58%), and 14 had familial disease. At the time of surgery, 19 patients had cervical lymph node metastasis, and 1 had distant metastasis. Eight patients had stage I disease; 4 had stage II disease; 3 had stage III; 16 had stage IVA; and 1 had stage IVC. Two cases were not classified owing to lack of data.

Based on previously published criteria,²¹ patients were considered to be biochemically cured after surgery if their serum calcitonin levels fell to the upper limit of the laboratory reference range or below. They were considered to have a relapse if their serum calcitonin levels became elevated 6 months or longer after having satisfied the criteria for a biochemical cure. They were classified as having persistent disease if their serum calcitonin level was elevated or if their levels were normal for less than 6 months. Patients who were considered cured had carcinoembryonic antigen (CEA) levels lower than laboratory reference limits, although 6 patients who were not cured also had normal CEA levels. At the time of the first postsurgical ambulatory visit, patients were classified as cured or as having persistent disease. At the time of the last ambulatory visit, the patients' clinical status was classified as alive without disease, alive with disease, or dead from MTC. No patient in this study died from non-MTC causes.

Tissue specimens were fixed in 10% formalin, set in paraffin blocks, and cut into 3-μm sections. Parallel and sequential histologic sections were immunohistochemically stained with MMP-2 and TIMP-2 reagents using the streptavidin-biotin-peroxidase complex amplification method (DakoCytomation, Carpinteria, California). Expression of MMP-2 was detected with a 1:700 dilution of a mouse monoclonal antibody (MS-806; Labvision, Fremont, California), which reacted with both the latent and active forms of the enzyme. Expression of TIMP-2 was detected with a 1:40 dilution of a mouse monoclonal antibody (MS-1485; Labvision). Immunostaining for MMP-2 was observed in the tumor stroma (Figure 1), while TIMP-2 immunostaining was observed in tumor cells (Figure 2). The extent and intensity of immunostaining were used to formulate the immunostaining indexes detailed in Table 1. Immunostaining for TIMP-2 was semiquantitatively estimated as the percentage of neoplastic cells stained in each specimen (extent index). The staining intensity (intensity index) was observed and a composite index calculated (extent index cells plus intensity index). The extent index for MMP-2 was evaluated using an automatic process. The intensity index and composite index for MMP-2 were estimated in a manner similar to that of TIMP-2 (Table 1).

Colon adenocarcinoma specimens were used as positive controls, and MTC vascular wall specimens, usually staining intensely for both reagents, were used as internal positive controls. As a negative control, the primary antibody was replaced with phosphate-buffered saline. Slides stained for MMP-2 were studied under a light microscope (Carl Zeiss Inc, Göttingen, Germany) with original magnification ×10, and the stained area was quantified with a Kontron Electronic 300 image analysis system (Carl Zeiss Inc). The TIMP-2–stained material was analyzed at original magnification ×40. Slides were analyzed by a single observer (C.R.J.), a pathologist who did not know the clinical features of the patients.

Assuming that TIMP-2 should behave like the MMP-2 inhibitor, we divided the MMP-2 composite indexes for the primary tumor by the TIMP-2 composite indexes in each case. Two possible results were obtained: 1.0 for those patients whose specimens received scores with similar indexes for both markers and 0.7 for those patients whose MMP-2 composite index was 2 and whose TIMP-2 composite index was 3. This index was called the relative index. Thirteen patients had a relative index of 0.7, and 20 had a relative index of 1.0. The values obtained were correlated with the developments observed shortly after surgery (initial evaluation), evidence of relapse, clinical and/or imaging evidence of local and/or distant disease in the absence of cure, and final clinical conditions. Clinical and pathologic aspects previously associated with the prognosis of MTC (presence of thyroid nodules and/or lymphadenopathy on admission,²² presence of cervical lymph node metastases,^20,22,23 and TNM stage^19,20,22) were also correlated with the relative index. These aspects presented a statistically significant association with the final clinical conditions of this group of patients (P = .002, P < .001 and P < .001, respectively).

Pearson χ² and Fisher exact tests were used for statistical analysis. The error margin used in the statistical test decision was 5.0% (P < .05). The analysis was performed with SAS software, version 8.0 (SAS Institute Inc, Cary, North Carolina), and SPSS, version 11 (SPSS Inc, Chicago, Illinois).

After initial surgical treatment, 18 patients were considered cured (55%), while 15 had persistent neoplasia (46%). Among those initially cured, 5 experienced relapses (28% of those initially cured; 15% of the entire group). Among the patients who were not cured, 8 eight had biochemical disease (24%) (elevated levels of serum calcitonin with no other clinical or image evidence of cervical or distant metastatic disease); 5 had cervical disease (15%); 1 had distant metastasis (3%); and 6 had locoregional and distant metastatic diseases (18%). At the end of clinical follow-up, 13 patients were cured (39%); 16 had active disease (49%); and 4 had died of the disease (12%), all with clinical and/or imaging evidence of cervical and/or distant neoplasia.

Statistical significance was obtained by correlating the indexes with the positive initial physical examination findings for the presence of thyroid nodules and/or lymphadenopathy (P = .02), presence of cervical lymph node metastases (P < .001), and final clinical conditions (alive without disease, alive with disease, or dead of disease) (P = .01) (Table 2).

To our knowledge, only 2 studies have specifically addressed MMPs in MTC.^10,14 In our earlier analysis of records from a group of patients similar to those in the present study,¹⁸ we observed the markings for MMP-2 in the tumor stroma and always at maximum intensity. A significant correlation was observed between the MMP-2 composite index (Table 1) and final clinical conditions (P < .001) as well as initial findings after surgery (P = .02): poorer conditions were found in those patients with higher indexes.

Immunomarkings for TIMP-2 were observed in the neoplasm cells. However, none of the proposed indexes for this marker provided a statistically significant association with clinical evaluations.

Tomita¹⁴ selected 3 cases of C-cell hyperplasia and 10 specimens of MTC that stained with antibodies for MMP-2 and -9 and TIMP-1 and -2. These enzymes were included in the list of markers of neuroendocrine cells, which include C thyroid cells. It was observed that nonpathologic C cells tended to have a stronger staining for MMPs than the neoplastic ones, while neoplastic cells also presented weaker staining for TIMPs. This is where the importance of the ratio of MMP to its specific inhibitor emerges. In half of the neoplasms studied, the staining for MMP was just as intense as for nonpathologic C cells.

Cvejić et al¹⁰ studied 22 sporadic cases selected for MMP-2 immunostaining. As in our study, all of the specimens were stained by the marker, and follicular cells did not present markings. However, no apparent association was noted between MMP-2 expression and neoplasm stage because patients who already had lymph node metastasis presented immunostaining intensities similar to those with restricted disease.

Campo et al¹⁵ observed the immunohistochemical distribution of MMP-2 in various thyroid tissues, including 3 MTC specimens. They observed stronger immunostaining for more aggressive neoplasms, although the MTC specimens exhibited moderate staining.

Korem et al¹⁶ observed weak to moderate immunostaining for MMP-2 in 4 MTC specimens. Of the controls, benign conditions, and follicular and medullary carcinomas, only the papillary thyroid carcinoma specimens (n = 12) showed intense immunostaining and greater enzymatic activity, although without correlation to the presence of lymph node metastasis. It has been suggested that follicular and medullary carcinomas, like more aggressive neoplasms, must rely on other MMPs for their growth processes. The quantifications of TIMP-2 were similar between healthy and pathologic specimens.

Komorowski et al⁵ attributed great importance to MMP-2 and TIMP-2 in the pathogenesis of thyroid cancer. Serum concentrations of MMP-2 in malignant thyroid neoplasms, including in 3 MTC specimens, were substantially greater than in controls and accompanied by slightly higher concentrations of TIMP-2, with an emphasis on the balance between the 2 enzymes.

The presence of high levels of MMP-2 and TIMP-2 messenger RNA (mRNA) has already been identified in epithelial thyroid cell cultures, cells derived from thyroid carcinomas, and fibroblasts derived from thyrocytes.¹⁷ In epithelial thyroid carcinomas, intense expressions of MMP-2 mRNA were identified in fibroblasts adjacent to the neoplastic cells, especially along tumor invasion lines.¹³ Significantly elevated levels and activity of MMP-2 were also identified in papillary carcinoma specimens compared with follicular adenomas and controls and with other MMPs. Expression of TIMP-2 was less relevant.¹² Researchers detected MMP-1 mRNA in malignant thyroid cells and found none in benign cells (cell cultures).²⁴ They also concluded that the positive MMP-1:TIMP-1 ratio would favor the degradation of the extracellular matrix by malignant thyrocytes.

Significantly higher MMP-2 activation rates have been observed in advanced-stage papillary carcinoma with lymph node metastasis, and the protease was seen as the primary cause of metastasis.¹² It was also observed that the MMP:TIMP ratios were significantly higher in carcinoma samples and that TIMP-2 expressions were not higher in the carcinoma specimens than in adenomas and controls.¹² Significantly higher MMP-2 activity has been observed in papillary carcinomas than in nonneoplastic tissues.¹¹ A significant correlation has been noted between MMP-2 immunostaining and the clinical and pathologic aspects of tumor size, lymph node metastasis, tumor stage, and intrathyroid and vascular invasions. Correlations were also observed between TIMP-2 immunostaining and tumor size, stage, and intrathyroid and vascular invasions, with a less promising outlook for patients with the highest TIMP-2 staining indexes.¹¹

After identifying a correlation between MMP-2 immunostaining and clinical evolution by means of the composite index applied to primary MTC specimens, and despite finding no evidence of the participation of TIMP-2 in the progression of or protection from MTC, we tried to attribute the inhibitory function of MMP-2 to TIMP-2. Dividing the MMP-2 composite index values by the TIMP-2 composite index values gives us the relative index for each case. Patients with relative indexes of 1.0 presented equivalent markings for the 2 proteins, and patients with indexes of 0.7 presented predominance of markings, and perhaps of activity, of TIMP-2 over MMP-2, which possibly indicated less aggressive neoplasms. This was one hypothesis put forward and confirmed through statistical analysis that grouped the patients under the 2 possible results and compared them with clinical and pathologic variables. The observations favored the patients whose primary tumors showed higher TIMP-2 than MMP-2 immunostaining indexes, likely indicating an inhibitory effect of TIMP-2 on the carcinogenic effects of MMP-2. Paradoxically, however, TIMPs may not offer protection against malignant neoplasms. They have been shown to participate in cell proliferation processes,¹⁷ and their expressions have already been correlated with a compromised prognosis.^3,11

Metalloproteinase 2 and TIMP-2 are correlated enzymes, and the consequences of their activities depend more on their proportions than on their absolute concentrations. Certainly, other factors including other MMPs are involved in this complex carcinogenesis mechanism.

Our patients whose MTC specimens showed higher TIMP-2 than MMP-2 immunohistochemical expression had better prognoses, while patients whose specimens presented proportional immunohistochemical expression for the 2 markers showed evidence of a poorer prognosis.

Correspondence: Beatriz Godoi Cavalheiro, MD, PhD, Rua Periquito 160-153, 04514-050 São Paulo, Brazil (bgcavalheiro@yahoo.com.br).

Accepted for Publication: November 14, 2008.

Author Contributions: Drs Cavalheiro and Brandão had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Cavalheiro, Junqueira, and Brandão. Acquisition of data: Cavalheiro and Junqueira. Analysis and interpretation of data: Cavalheiro and Junqueira. Drafting of the manuscript: Cavalheiro, Junqueira, and Brandão. Critical revision of the manuscript for important intellectual content: Cavalheiro and Brandão. Statistical analysis: Cavalheiro and Junqueira. Administrative, technical, and material support: Cavalheiro and Junqueira. Study supervision: Cavalheiro and Brandão.

Financial Disclosure: None reported.

Previous Presentation: This study was presented in part at the Seventh International Conference on Head and Neck Cancer; July 19-23, 2008; San Francisco, California.

Additional Contributions: The Immunohistochemistry Laboratory of the Pathology Department of the Universidade de São Paulo Medical School assisted in the preparation of the study materials.