β-Catenin and Cyclin D1 in Mucoepidermoid Carcinoma of Variable Histologic Grades

Objective  To analyze the expression of β-catenin and cyclin D1 in mucoepidermoid carcinoma (MEC) of variable histologic grades to establish a correlation between the expression of these proteins and the different histopathologic grades of this neoplasia.

Design  Immunohistochemical analysis of MEC.

Setting  Pathological Anatomy Service, Discipline of Oral Pathology, Department of Dentistry, Federal University of Rio Grande do Norte, Natal, Brazil.

Patients  Fifteen cases of MEC, graded and categorized according to criteria reported in the literature into 5 tumors with a low grade of malignancy, 4 with an intermediate grade, and 6 with a high grade.

Main Outcome Measures  Labeling patterns of β-catenin and cyclin D1.

Results  No significant difference in β-catenin labeling patterns was observed between low- and intermediate-grade tumors or between low- and high-grade tumors (P = .60 and P = .06, respectively; Fisher exact test), despite a strong tendency toward a difference in the latter. In contrast, a significant difference was noted between intermediate- and high-grade tumors (P = .03). For cyclin D1, no labeling was observed in any high-grade cases, and only 3 cases showed overexpression of this protein. Comparison of the labeling patterns among the different histologic grades revealed no significant difference.

Conclusions  The reduced expression of β-catenin observed in all high-grade MECs is probably due to the loss of its adhesion function, which confers a greater invasive potential to these tumors. The overexpression of cyclin D1 observed in only 3 MEC cases suggests that this protein does not participate in the etiopathogenesis of these tumors, which implies that other genes are likely responsible.

Mucoepidermoid carcinoma (MEC) is the most common malignant epithelial tumor of the salivary glands, affecting mainly the parotid and minor salivary glands.^1-4 It accounts for 15.5% of all salivary gland tumors and 29% of all malignant tumors in both the major and minor salivary glands.⁵ Histologically, MEC consists of a variable proportion of neoplastic cells, including mucosal, intermediate, and epidermoid cells, showing a cystic or solid growth pattern.³ Various studies have demonstrated that some clinical characteristics, such as tumor location and stage, and histopathologic characteristics of MEC, such as cell proliferation rate and local invasiveness of adjacent tissues, are correlated with the clinical behavior and prognosis of the tumor.

Cyclin D1 is an essential protein for cell cycle progression from the G₁ to the S phase and is frequently overexpressed or genetically altered in breast, prostate, and colon cancers.⁶ These alterations in the cyclin D1 protein and gene have also been identified in squamous cell carcinomas of the head and neck.^7-10

β-Catenin plays an important role in E-cadherin–mediated cell adhesion and cell differentiation and tumor progression^3,11 and is also involved in the Wnt-1 signaling pathway. Mitotic signals released through the Wnt-1 pathway activate some proteins, a process that results in the transcription of genes that regulate cell growth and differentiation.¹² β-Catenin participates in this pathway by forming a heterodimer with the T-cell transcription factor (TCF), and in the absence of a Wnt signal, cytosolic β-catenin is degraded by an adenomatous polyposis coli–dependent mechanism. However, in the presence of stimulation of the Wnt pathway, cytosolic β-catenin is stabilized, accumulates in the nucleus, and binds to transcription factors of the TCF/lymphoid enhancer factor family, initiating the transcription of genes such as c-myc, cyclin D1, MMP-7, and WISP-1.^11,13 The Wnt pathway is highly regulated in normal cells, but its regulation is generally altered during tumorigenesis.¹³

Few studies have investigated the expression of β-catenin and cyclin D1 in MEC, and the simultaneous expression of these 2 proteins in these tumors seems to have been poorly explored. Literature findings demonstrate that cyclin is frequently overexpressed or genetically altered in some types of cancer and that the expression of β-catenin is altered in some neoplasms, including its reduction in the cell membrane and accumulation in the nucleus. Therefore, our initial hypothesis was that cyclin D1 is overexpressed in MECs, especially in high-grade tumors, and that the expression of β-catenin is reduced in the cell membrane mainly in high-grade tumors in addition to its nuclear accumulation in these cases. The objective of the present study was to determine the immunohistochemical expression of β-catenin and cyclin D1 in MEC of variable histologic grades to establish a correlation between the expression of these proteins and different histopathologic grades of this neoplasia.

Fifteen cases of MEC were selected from the files of the Pathological Anatomy Service, Discipline of Oral Pathology, Department of Dentistry, Federal University of Rio Grande do Norte. The experiment was approved by the university’s ethics review committee. Hematoxylin-eosin–stained specimens were examined under a light microscope by 2 pathologists (M.C.C.M. and M.C.O.), and tumors were graded according to the criteria reported by Hicks et al¹⁴ into low, intermediate, and high grade. Five cases were classified as low grade, 4 as intermediate grade, and 6 as high grade. Paraffin-embedded specimens were cut into 3-μm-thick sections and examined by immunohistochemical analysis using the streptavidin-biotin method. Antigen retrieval for antibodies was performed with citrate, pH 6.0, in a steamer for 30 minutes. Polyclonal anti–β-catenin antibody (Neomarkers, Fremont, Calif), diluted in a ratio of 1:500 and incubated for 30 minutes, and monoclonal anti–cyclin D1 antibody (clone P2D11F11; Novocastra, Newcastle, England), diluted in a ratio of 1:25 and incubated for 120 minutes, were used as primary antibodies. The sections were exposed to a second antibody, and the complex used for development was diluted in a ratio of 1:100. Endogenous peroxidase activity was blocked by incubation with hydrogen peroxide (10 volumes). The reaction was developed with 0.03% diaminobenzidine as a chromogen, and the material was counterstained with Mayer hematoxylin.

Immunohistochemical labeling for β-catenin was based on the criteria described by Shieh et al³ using a semiquantitative scale on which the results were classified into 4 categories according to the percentage of immunolabeled cells. Areas of nonneoplastic squamous epithelium and salivary gland epithelium present in the specimens were used as a positive internal control. When more than 90% of tumor cells were stained uniformly at the same intensity as the normal epithelium, the result was defined as positive. If the labeling of neoplastic cells was weaker than that of the normal epithelium and/or heterogeneous (10%-90%), the result was defined as reduced expression. When no labeling was observed or less than 10% of cells were labeled, the result was classified as negative. If the labeling was strong in more than 50% of tumor cells compared with the normal epithelium, the pattern was defined as reverse positive. For statistical analysis, cases that showed a positive labeling pattern were considered to be normal, whereas the other 3 patterns (negative, reduced, and reverse positive) were classified as altered.

Cyclin D1 was analyzed using the criterion proposed by Haas et al,¹⁰ with some modifications, whereby an approximate fraction of tumor cells that showed a positive nuclear reaction was classified as follows: 0% to 5% of positive cells were scored as negative, 5% to 20% were scored as reduced expression, 20% to 50% were scored as positive, and more than 50% were scored as strongly positive. Overexpression was defined as labeling of more than 20% of cells.

The immunohistochemical results were analyzed by the Fisher exact test for comparison among the different histologic grades and each labeling category. The level of significance was defined as P<.05.

In the present study, 10 patients were female and 5 were male. With respect to anatomical location, 10 tumors were located in the minor salivary glands and 5 in the major salivary glands. Patients ranged in age from 9 to 85 years, with a mean of 51.6 years.

Positivity for β-catenin labeling was detected in the squamous epithelium and normal salivary gland epithelium present in areas adjacent to the tumor (Figure 1). Expression was mainly observed in the intercellular junctions of the basal and suprabasal layers but also in the cytoplasm of the epithelial lining. In the normal salivary gland epithelium, cytoplasmic labeling was observed in the excretory ducts.

In MEC, the labeling pattern varied from case to case and within the same case, and labeling was mainly observed in the cell membrane and cytoplasm, with a predominance in the latter. In some cases, nuclear labeling was also detected (Figure 2, Figure 3, and Figure 4).

The labeling results for β-catenin according to the different histopathologic grades are given in Table 1. In all of the evaluated cases considered abnormal, reduced expression of β-catenin was identified. For each marker, pairwise comparisons of 2 histologic grades at a time were performed with the Fisher exact test. No significant difference in β-catenin labeling patterns was observed between low- and intermediate-grade MEC (P = .60) or between low- and high-grade tumors (P = .06), whereas a significant difference was detected between intermediate- and high-grade cases (P = .03).

For cyclin D1, nuclear labeling was observed in focal areas of the basal and suprabasal layers of the normal squamous epithelium present in areas adjacent to the tumor (Figure 5), in addition to cytoplasmic labeling in the excretory ducts of the normal salivary gland epithelium present in the specimens studied. No labeling of MEC areas was observed in any high-grade cases, and only 3 cases showed overexpression of cyclin D1, including 2 low-grade tumors and 1 intermediate-grade tumor (Figure 6 and Table 2).

No significant difference in cyclin D1 labeling was observed between low- and intermediate-grade MEC (P = .60), low- and high-grade cases (P = .18), or intermediate- and high-grade cases (P = .40). The association between cyclin D1 and β-catenin labeling is given in Table 3.

Because of its histopathologic diversity, various grading systems have been used to classify MEC in an attempt to identify a correlation between the histopathologic aspects and the prognostic and overall survival presented by this tumor. On the basis of these aspects, we decided to use the 3-tiered grading system revised by Hicks et al,¹⁴ which verified that this system correlates with clinical and pathologic factors that influence the prognostic and overall survival in affected individuals. Moreover, this grading system was more suitable for the analysis of incisional biopsy specimens, which were used in the sample evaluated.

β-Catenin is a 92-kDa protein that, together with α- and γ-catenin, was first associated with the cytoplasmic chain of E-cadherin. β-Catenin binds to E-cadherin, α-catenin, and consequently the microfilament network of the cytoskeleton. An additional function of intracellular signaling has also been attributed to β-catenin, with this protein involved in the Wnt signaling pathway. These 2 distinct biological functions of β-catenin determine its cell localization; as a protein involved in cell-cell adhesion, it is located in the cytoplasm membrane and maintains the epithelial architecture, whereas the free cytoplasmic protein participates in signaling by binding to TCF and activating its transcriptional activity in the nucleus, which results in the expression of target genes such as cyclin D1, c-myc, and MMP-7. Therefore, despite the lack of nuclear localization, the translocation of β-catenin to the nucleus is mediated by binding to TCF after an increase in the free cytoplasmic pool through the Wnt signaling pathway.¹⁵ The increased levels of β-catenin might favor neoplastic transformation by inducing the expression of these target genes and consequently causing uncontrolled cell proliferation.¹⁶

The present results obtained for the normal epithelium revealed β-catenin labeling in the cell membrane, mainly in the basal and suprabasal layers, in agreement with the findings of Shieh et al³ and Lopez-Gonzalez et al.¹² In the excretory ducts, labeling was observed in the cell membrane, as also detected by Shieh et al,³ in addition to cytoplasmic labeling.

In MEC, β-catenin labeling was observed in the membrane, cytoplasm, and nucleus. In general, there was a predominance of labeling in the cytoplasm compared with the cell membrane, a fact also observed by Seidler et al¹⁷ for colorectal adenocarcinomas. Several investigators have reported nuclear or cytoplasmic detection of β-catenin, and this redistribution has been associated with its transactivation potential mediated by the Wnt signaling pathway.¹²

In the present study, few cases exhibited nuclear labeling, including 3 low-grade tumors and only 1 high-grade tumor. This finding agrees with Wu et al,¹⁸ who detected nuclear β-catenin labeling in ovarian endometrioid adenocarcinomas, especially in well-differentiated tumors, whereas Lee et al¹⁹ observed a correlation between the nuclear localization of β-catenin and high-grade serous ovarian carcinomas. These findings are probably due to the variable β-catenin labeling in different types of tumors.

Comparison of β-catenin labeling between the different histopathologic grades revealed no significant difference between low- and intermediate-grade tumors. In contrast, a significant difference was observed between intermediate- and high-grade tumors, and low- and high-grade tumors tended to be different, although this was not statistically significant.

Abnormal β-catenin labeling was observed in all high-grade MEC cases, showing reduced immunoreactivity. This finding confirms other studies that demonstrate a decrease in β-catenin expression in high-grade tumors.^3,12,20,21 This reduced expression is probably due to a loss of function of β-catenin as an adhesion molecule, since in high-grade invasive or metastatic tumors, cohesiveness between cells and between cells and the substrate tends to diminish because of rupture of the E-cadherin–β-catenin complex, thus increasing cell mobility, a fact that confers a greater invasive potential.^22,23

Shieh et al³ observed an altered expression of β-catenin in patients with low- or high-grade MEC and in an advanced clinical stage. In the same study, multivariate analysis of β-catenin expression and histopathologic grade demonstrated a significant correlation between the abnormal expression of this protein and a reduced survival rate, and the authors suggested the combined use of histopathologic grade and immunohistochemical labeling for β-catenin for the determination of the clinical prognosis of these patients.

Cyclins are molecules that regulate the activity of cyclin-dependent kinases. The programmed synthesis and individual degradation of cyclins ensure the maintenance of the order of cell cycle events.²⁴ The D-type cyclins are considered to be key regulators in the progression from the G₁ to the S phase.²⁵

Cyclin D1 binds to cdk4 or cdk6, forming a complex that binds to the pRb protein, phosphorylating and inactivating it and leading to the release of transcription factor E2F, thus initiating cell proliferation through the G₁ phase.¹⁰ The cyclin-cdk complex controls cycle progression by means of ordered activation and inactivation. Any disturbance in this mechanism may play an important role in the pathogenesis of various malignant tumors. Some studies indicate that overexpression of cyclin D1 occurs at the beginning of tumor development and might therefore be considered an early marker of cell proliferation, whereas others have demonstrated late overexpression of this protein during the development of malignant tumors.²⁶

In the present study, nuclear immunoreactivity to cyclin D1 was observed in focal areas of the basal and suprabasal layers of the normal lining epithelium present in areas adjacent to some tumors, a fact also reported by Rousseau et al,⁸ whereas Schoelch et al⁷ and Haas et al¹⁰ did not detect such labeling. This difference is probably due to the different clones and types of antigen retrieval used. Cytoplasmic labeling of cyclin D1 was detected in the excretory ducts of normal salivary glands. However, Etges et al²⁷ did not observe the expression of this protein in 6 normal salivary gland cases studied.

Perez-Ordonez et al²⁸ observed weak expression of cyclin D1 in polymorphous low-grade adenocarcinomas. Shintani et al²⁹ detected the expression of cyclin D1 in adenoid cystic carcinoma and observed overexpression in 4 of 22 cases, 3 of them showing the solid pattern, which has the worst prognosis. Overexpression of cyclin D1 was also correlated with the cell proliferation rate, suggesting that this overexpression is related to the high proliferative activity of adenoid cystic carcinoma.

In the present study, overexpression of cyclin D1 was observed only in 3 MEC cases, including 2 low-grade tumors and 1 intermediate-grade tumor. This finding contradicts the observation of some investigators who indicated a higher expression of this protein in high-grade and poorly differentiated tumors,^7,8 tumors with a poor prognosis,^30,31 and also nodal metastases.⁹ In addition, amplification of the cyclin D1 gene has been associated with a poor prognosis.^32-34

In 3 cases of MEC, Etges et al²⁷ observed positivity for cyclin D1, but the histopathologic grade of these tumors was not reported. In addition, no immunolabeling was detected in 6 normal salivary gland cases used as a comparative parameter.

In human hepatocellular and colorectal carcinoma, evidence indicates that cyclin D1 is the main transcriptional target of the Wnt–β-catenin signaling pathway.³⁵ This association does not seem to exist in the MEC cases studied herein, because although most low- and intermediate-grade tumors showed β-catenin expression mainly in the cytoplasm, overexpression of cyclin D1 was detected in only 3 cases. Additional studies that use techniques that demonstrate gene expression of these proteins are necessary to provide more convincing evidence.

Furthermore, in the study by Haas et al,¹⁰ overexpression of cyclin D1 was rare in tonsillar carcinomas but frequent in carcinomas of the tongue, implying a relationship between cyclin D1 overexpression and distinct tumor sites. These findings led the authors to conclude that distinct molecular alterations occur at different tumor sites in head and neck cancer. On the basis of these considerations, we suggest that cyclin D1 does not seem to play a role in the etiopathogenesis of the MEC cases studied. Another important factor is that according to Si and Liu,²⁶ cyclin does not seem to be sufficient for malignant transformation in most cells, requiring the synergistic effect of other proto-oncogenes. Therefore, cyclin does not represent an exclusive factor, and consequently, other nuclear transcription factors might be involved in the pathogenesis of MEC.

Reduced cell membrane expression of β-catenin was mainly demonstrated in high-grade tumors, suggesting a probable disturbance in the complex intercellular adhesion system, a fact that might contribute to the poor prognosis shown by this type of tumor. However, cytoplasmic labeling was generally observed in the same tumors, and nuclear labeling was detected in 1 case. Nevertheless, no cyclin D1 expression was observed in these cases even though nuclear expression of cyclin D1, one of the target genes of β-catenin, was expected with the evident cytoplasmic labeling of the latter. This suggests that other genes are involved in the pathogenesis of MEC.

Correspondence: Lélia Batista de Souza, PhD, Faculdade de Odontologia, Laboratório de Patologia Oral, Universidade Federal do Rio Grande do Norte, Av Senador Salgado Filho, 1787 CEP-59056-000 Natal RN, Brazil (leliasouza@dod.ufrn.br).

Submitted for Publication: October 8, 2004; final revision received April 1, 2005; accepted April 1, 2005.