Clinical Significance of Cellular Distribution of Moesin in Patients with Oral Squamous Cell Carcinoma

572 Vol. 10, 572–580, January 15, 2004 Clinical Cancer Research

Hiroichi Kobayashi,1 Junji Sagara,2 those with a better prognosis and might improve the defini- Hiroshi Kurita,1 Masayo Morifuji,3 tion of suitable therapy for each. Masamichi Ohishi,3 Kenji Kurashina,1 and Shun’ichiro Taniguchi2 INTRODUCTION Departments of 1Dentistry and Oral Surgery and 2Molecular Oral cancer, 1 of the 10 most common cancers in the world, Oncology and Angiology, Aging and Adaptation, Shinshu University remains a morbid and often fatal disease. Despite marked advances School of Medicine, Matsumoto, Japan and 3First Department of Oral of management and diagnosis of oral squamous cell carcinoma and Maxillofacial Surgery, Faculty of Dentistry, Kyushu University, (OSCC), the overall survival ratio has showed only a modest Fukuoka, Japan increase in recent years. Therefore, the development of molecular markers is needed to improve the diagnosis and assessment of ABSTRACT tumor progression and metastasis in OSCC patients. Moesin is a member of the ERM (ezrin/radixin/moesin) Purpose: Moesin is a linking protein of the submembra- family, which shares ϳ78% amino acid sequence identity with neous cytoskeleton and plays a key role in the control of cell each other. ERM proteins, part of the band 4.1 superfamily, act morphology, adhesion, and motility. The aim of the present as a membrane-cytoskeleton linker in actin-enriched specialized study was to elucidate the clinical significance of expression plasma membrane structures, especially microvilli, ruffling patterns of moesin in patients with oral squamous cell car- membranes, and cleavage furrows and thus play a key role in the cinoma (OSCC). control of cell morphology, adhesion, and motility (1–6). The Experimental Design: Immunohistochemistry for moe- integral membrane proteins such as CD44, CD43, intercellular sin monoclonal antibody was performed on 103 paraffin- adhesion molecules 1 and 2, and actin are identified as ligands embedded specimens from patients with primary OSCC, for ERM proteins (7–9). Merlin, encoded by the neurofibroma- including 30 patients with locoregional lymph node metas- tosis type 2, is classified as a tumor suppressor protein (10). tasis, and in the sections from nude mice transplanted with Because moesin shares high homology with Merlin and colo- two cell lines derived from a single human tongue cancer calizes with it beneath the plasma membrane, it has been spec- (SQUU-A and SQUU-B). ulated that moesin may also be a tumor suppressor (11, 12). Results: Expression patterns of moesin in OSCCs were However, recent studies have indicated that ERM proteins are divided into three groups: membranous pattern; mixed pat- up-regulated in various kinds of tumors (13–17). Whether moe- tern; and cytoplasmic pattern. These expression patterns sin is functional as a tumor suppressor in carcinogenesis and correlated with tumor size, lymph node metastasis, mode of tumor development remains unclear. invasion, differentiation, and lymphocytic infiltration. In Because little is known about the role of moesin in the oral about two-thirds of the patients with metastatic lymph node, normal mucosa and oral lesions, including leukoplakia, verru- homogeneous cytoplasmic expression was detected in the cous carcinoma, and small cell carcinoma, we initiated a series metastatic lymph nodes. In addition, SQUU-B with high of studies aimed at characterizing expression of moesin (18). In metastatic potential showed more reduced levels of mem- this particular study, we report that OSCC cells constitutively brane-bound moesin than SQUU-A with low metastatic po- display several expression patterns of moesin, thereby providing tential. A multivariate analysis demonstrated that expres- a new biomolecular marker for use in prediction of metastasis sion patterns of moesin can be an independent prognostic and poor prognosis. factor. Conclusions: Our results suggest that moesin expres- MATERIALS AND METHODS sion contributed to discriminating between patients with the Patients and Tumor Sample. The study group consisted potentiality for locoregional lymph node metastasis and of 103 patients with OSCC who were diagnosed at the Department of Dentistry and Oral Surgery, Shinshu University School of Med- icine. Tissues of primary (n ϭ 103) and metastatic (n ϭ 30) lesions of OSCCs were collected during biopsy or operation after patients signed the informed consent form approved by the Institution Received 11/1/02; revised 10/14/03; accepted 10/20/03. The costs of publication of this article were defrayed in part by the Review Committee. Permission to perform this study was obtained payment of page charges. This article must therefore be hereby marked from the committee. Thirty-one patients (T1–T2) without lymph advertisement in accordance with 18 U.S.C. Section 1734 solely to node metastasis underwent radiotherapy alone while 72 patients indicate this fact. underwent surgery. The grade of tumor differentiation was deter- Requests for reprints: Hiroichi Kobayashi, Department of Dentistry and Oral Surgery, Shinshu University School of Medicine, Asahi 3-1-1, mined according to the criteria proposed by the WHO (19). Mode Matsumoto 390-8621, Japan. Phone: 81-263-37-2677; Fax: 81-263-37- of invasion was classified according to Jakobsson’s classification 2676; E-mail: [email protected]. (20). Median follow-up time was 32.0 months (range, 2–115

Fig. 1 Staining of normal oral epithelium with monoclonal antibody against moesin. A, expression level was gradually reduced from the parabasal layer toward the top layer. B, reactivity for moesin monoclonal antibody was prominent in the cell membrane of basal layer and parabasal layer cells. Bar, 100 ␮minA;10␮minB.

months). The study population consisted of 59 men and 44 women Cell Lines and Culture. Human oral cancer cell lines, averaging 65.0 years of age (range, 27–88 years). For controls, SQUU-A and SQUU-B, were established as reported previously normal oral mucosa were obtained from consenting patients during (21). These cell lines were cultured in Eagle’s MEM (Nissui, removal of a lower wisdom tooth. Tokyo, Japan) supplemented with 10% fetal bovine serum (Life

Fig. 2 Three staining patterns in oral squamous cell carcinomas with monoclonal antibody against moesin. A, membranous expression pattern, predominant expression in the cell membrane. B, mixed expression pattern, reactivity in the cell membrane about equal to that in the cytoplasm. C, cytoplasmic expression pattern, predominant cytoplasmic labeling. Bar,10␮m.

Fig. 3 Immunohistochemical localization of moesin in primary tissue and metastatic lymph nodes of the same oral squamous cell carcinoma patient. A and B, membranous expression of moesin in primary tissue. C and D, cytoplasmic expression in front of an invasive margin of primary tumor cells. E and F, cytoplasmic expression homogeneously observed. Bar, 100 ␮minA, C, and E;10␮minB, D, and F.

Technologies, Inc., Grand Island, NY), penicillin (100 IU/ml), zuoka, Japan) and were housed under conventional conditions streptomycin (100 mg/ml), and fungizone (1 mg/ml) at 37°Cin with free access to animal chow and water. Under general ϳ ϫ 5 an atmosphere of 5% CO2. anesthesia with diethylether, 3.5 10 /0.035 ml viable cells Animals and Experimental Treatment. Female were injected in the s.c. tissue of the right side of the tongue. BALB/c mice (6 weeks old) were purchased from SLC (Shi- The mice were sacrificed 5 weeks after the infection, and

crowaved (500 w, 25 min) for antigen retrieval in 0.01 M citric acid (pH 6.0). After being washed with distilled water and 0.05 M Tris-buffered saline (TBS; pH 7.6), the specimens were treated with mouse anti-moesin antibody diluted by TBS containing 1% BSA at 4°C overnight, washed three times with TBS, and then incubated with goat antimouse immunoglobulin polyclonal antibody diluted by TBS containing 1% BSA for 60 min at room temperature. After being washed three times with TBS, the sections were developed in 0.05 M Tris-buffer (pH 7.6) containing 25 mg/125 ml 3,3Ј- diaminobenzidine and 0.0015% hydrogen peroxide for 7 min. The sections were then washed in water, counterstained with Mayer’s hematoxylin, dehydrated, cleaned, and coverslipped. Negative controls were treated by replacing the primary antibody with TBS 1% BSA. Scoring of Results. Sections were examined by two independent researchers (H. Ko., H. Ku.) in an effort to provide a consensus of staining pattern. Moesin or ␤-actin expression of neoplastic cell in primary lesions was classified as follows: membranous pattern—membranous expression of moesin or ␤-actin was more dominant than cytoplasmic expression; mixed pattern—membranous expression of moesin or ␤-actin was ap- Fig. 4 Comparison of percentage of cytoplasmic expression of moesin in primary tissues and metastatic lymph nodes in the same oral squa- proximately equal to cytoplasmic expression; and cytoplasmic mous cell carcinoma patient with cervical lymph node metastasis. In all pattern—membranous expression of moesin or ␤-actin was primary tissues, cytoplasmic expression type was heterogeneously ob- weaker than cytoplasmic expression. We used the expression served (range, 10–92%), whereas in about two-thirds of the metastatic pattern of moesin used as the predominant pattern on the whole lymph nodes, cytoplasmic expression pattern was homogeneously dis- played. histological section of the tumor.

tongues were resected. The care and use of these experimental Table 1 Expression pattern of moesin in oral squamous cell animals were in accordance with institutional guidelines. carcinoma according to clinicopathological features of patients Antibodies and Immunohistochemical Staining. Expression pattern of moesin Mouse monoclonal antibody (CR-22) was kindly provided by Total Drs. Shoichiro Tsukita and Sachiko Tsukita, Kyoto University Features no. Membranous Mixed Cytoplasmic P (Kyoto, Japan); this antibody has a higher affinity for moesin Total no. of patients 103 28 38 37 (22). The specificity of antibody against moesin was confirmed Age (yrs) 0.3277 Ͻ65 44 8 20 16 with Western blotting and immunoprecipitation of the cell ly- Ͼ65 59 20 18 21 sates from human peripheral leukocyte (23) and human malig- Gender 0.8481 nant melanoma cells (11). In addition, the samples were re- Male 59 15 23 21 ported to show similar staining in frozen sections as well as Female 44 13 15 16 T classification 0.0012 formalin-fixed, paraffin-embedded section (11, 23). Further- ϭ T1 12 65 24 24 17 ϭ more, our previous study also validated the specificity of this T2 53 ϭ antigen by Western blot analysis from OSCC tissues and im- T3 14 3841420 ϭ munohistochemical staining in frozen OSCC tissues (18). T4 24 Ͻ Horseradish peroxidase-conjugated goat antimouse polyclonal N classification 0.0001 N0 69 26 27 16 antibody (Dako, Copenhagen, Denmark) and anti-actin (␤) ϩ N1 N2 3421121 monoclonal antibody (Abcam, Cambridge, United Kingdom) Mode of invasion 0.0017 were purchased, respectively. 1 ϭ 6 82 27 31 24 ϭ All samples were fixed in 10% formalin and embedded 2 38 3 ϭ 38 in paraffin to prepare serial sections. Expression of moesin 4 ϭ 21 21 1 7 13 was examined by the indirect peroxidase technique as de- Differentiation 0.0002 scribed previously (23). Tissues embedded in paraffin were Well ϭ 65 65 25 24 16 cut into 3-␮m sections and mounted onto silane-coated glass Moderate ϭ 28 3831421 ϭ slides. The slides were dewaxed in xylene, dehydrated in Poor 10 Lymphocytic 0.0034 descending dilutions of ethanol, and preincubated with a infiltration solution of 1% hydrogen peroxide in methanol to suppress Low ϭ 29 29 4 8 17 the endogenous peroxidase for 30 min at room temperature. Moderate ϭ 53 74 24 30 20 ϭ After being rinsed in distilled water, the sections were mi- Marked 21

calculated by univariate and multivariate analysis using Cox regression models. The correlation between expression patterns of moesin and ␤-actin was estimated by Spearman’s rank correlation.

RESULTS Moesin Expression in Normal Oral Epithelia. In basal layer cells as well as restricted parabasal layer cells and spinous layer cells, the membranous expression pattern of moesin is dominant. Weak immunoreactivity for moesin was seen in the cytoplasm of basal layer cells, and no apparent staining with anti-moesin antibody was observed in cornified layer cells (Fig. 1, A and B). Moesin Expression in Primary and Metastatic Tumors. Fig. 5 Kaplan-Meier survival curves of patients with oral squamous cell carcinoma according to expression pattern of moesin. The survival In a previous study, we showed that moesin expression de- curves were analyzed by the log-rank test. creased in the membrane and increased in the cytoplasm in accordance with the degree of transformation and malignancy of oral lesions, including OSCC, verrucous carcinoma, and dys- Thirty cases with locoregional lymph node metastasis were plastic lesion. In this study, we focused on moesin expression in evaluated for heterogeneity of tumor cells in the primary lesions primary tumors and metastatic lymph nodes. as well as in the metastatic lesions by examination in 10 ran- The cellular distribution pattern of moesin differed sub- domized fields of sections at a magnification of ϫ400. Expres- stantially in primary tumors and metastatic lymph nodes. We sion percentages of cytoplasmic expression pattern of moesin divided moesin distribution patterns into three types: membra- was calculated from these. nous (Fig. 2A); mixed (Fig. 2B); and cytoplasmic (Fig. 2C) Statistics. The relationships between expression of moe- patterns. In OSCC patients with locoregional lymph node me- sin and clinicopathological indices such as age, gender, tumor tastasis, primary tumors showed various distribution patterns of size, lymph node metastasis, differentiation, mode of invasion, moesin as in Fig. 2, but most metastatic tumors in lymph nodes and lymphocytic infiltration were evaluated by Mann-Whitney’s showed the cytoplasmic distribution pattern. A case of OSCC U test. Kaplan-Meier survival curves were constructed and patient with cervical lymph node metastasis is presented in Fig. log-rank tests performed to assess whether the expression pat- 3. Membranous or cytoplasmic patterns are seen in the primary tern of moesin in neoplastic cells had any effect on overall tumor of the patient (Fig. 3, A–D), but all of the metastatic survival of patients with oral cancer. Relative risk of death was tumors in the lymph nodes display the cytoplasmic pattern (Fig.

Table 2 Univariate and multivariate analysis of clinicopathological data and expression pattern of moesin in 103 cases of oral cancer Variable HRa 95% CIb P Univariate analysis Age (Ͼ65/Ͻ65 years) 1.837 0.729–4.631 0.1973 Gender (male/female) 1.128 0.461–2.761 0.4927 ϩ ϩ T classification (T1 T2/T3 T4) 0.325 0.132–0.797 0.0141 ϩ N classification (N0/N1 N2) 0.262 0.107–0.643 0.0035 Mode of invasion (1 ϩ 2 ϩ 3/4) 0.297 0.121–0.729 0.0081 Differentiation (well/moderate ϩ poor) 0.209 0.080–0.554 0.0013 Lymphocytic infiltration (marked ϩ moderate/low) 0.274 0.114–0.662 0.0040 Moesin 0.0014 Cytoplasmic 1.000 Mixed 0.203 0.067–0.614 Membranous 0.07 0.009–0.532 Multivariate analysis ϩ ϩ T classification (T1 T2/T3 T4) 0.765 0.228–2.567 0.6639 ϩ N classification (N0/N1 N2) 0.691 0.202–2.359 0.5550 Mode of invasion (1 ϩ 2 ϩ 3/4) 0.712 0.265–1.915 0.5008 Differentiation (well/moderate ϩ poor) 0.408 0.141–1.183 0.0988 Lymphocytic infiltration (marked ϩ moderate/low) 0.493 0.190–1.282 0.1470 Moesin 0.0470 Cytoplasmic 1.000 Mixed 0.305 0.098–0.947 Membranous 0.163 0.020–1.361 a Hazard ratio (HR) estimated from Cox proportional hazard regression model. b Confidence interval (CI) of the estimated HR.

Fig. 6 Moesin localization in sections from nude mice transplanted with two cell lines. Tumor cell, at least in part, showed membranous expression of moesin in SQUU-A with low metastatic ability (A and B), but not in the SQUU-B with high metastatic ability (C and D). Bar,50␮minA and C;10␮minB and D.

3, E and F). Comparison in moesin expression between primary Univariate regression analysis showed that overall and metastatic tumors in the same patients demonstrated that survival correlated with tumor size, cervical lymph node metastatic cells predominantly showed cytoplasmic pattern, metastasis, mode of invasion, differentiation, lymphocytic whereas primary tumors showed heterogeneous expression pat- infiltration, and expression pattern of moesin. Furthermore, a terns (Fig. 4). multivariate analysis using the Cox proportional hazards Relationships between Expression Pattern of Moesin model also showed that the expression pattern of moesin and Clinicopathological Parameters and Prognosis in OSCC (P ϭ 0.0470) is significantly associated with overall survival Patients. We compared moesin expression pattern with clin- (Table 2). icopathological parameters in 103 OSCC patients (Table 1). Moesin Level and Cellular Localization in Oral Can- There was an association of expression pattern of moesin with cers. In the next experiments, we investigated the expression tumor size, cervical lymph node involvement, mode of invasion, of moesin in established cell lines of OSCC. Mixed or predom- differentiation, and lymphocytic infiltration. However, the ex- inantly cytoplasmic expression patterns of moesin were detected pression pattern of moesin did not differ significantly with in a large number of SQUU-A cells with low metastatic poten- respect to age and gender. tial. These cells showed predominantly membranous expression By the Kaplan-Meier curves, there was significant differ- of moesin in some part (Fig. 6, A and B), whereas the whole of ence among three groups, membranous pattern, mixed pattern, SQUU-B cells with high metastatic activity exhibited a down- and cytoplasmic pattern (␹2 ϭ 18.841, P Ͻ 0.0001 by log-rank regulation of membranous expression and an increase in cyto- test; Fig. 5). plasmic expression (Fig. 6, C and D).

Fig. 7 Distribution of ␤-actin in normal oral epithelia and primary oral squamous cell carcinoma tissues. A, normal oral epithelia. B, well- differentiated carcinoma tissue without lymph node metastasis. C, poorly differentiated carcinoma tissue with lymph node metastasis. Bar,25␮m.

Correlation between Expression Patterns of Moesin mode of invasion, differentiation, and lymphocytic infiltration. and ␤-Actin in Primary Tumors. In normal oral epithelia, Furthermore, our results demonstrated that tumor cells with basal layer and parabasal layer cells showed homogeneous cytoplasmic expression of moesin showed higher incidence of cytoplasmic staining against anti-actin (␤) monoclonal antibody lymph node metastasis than tumor cells with membranous ex- and spinous layer cells showed juxtamembranous staining to- pression of moesin. This is consistent with the observation that ward the top layer (Fig. 7A). Moreover, in the cytoplasm of all using a murine model in which almost all the whole cells in a of the tumor cells, ␤-actin was stained. Especially, in well highly metastatic cell line (SQUU-B) showed cytoplasmic ex- differentiated tumor cells without lymph node metastasis, ␤- pression of moesin, whereas a small number of cells in a less actin was strongly stained beneath the cell membrane (Fig. 7B), metastatic cell line (SQUU-A) showed membranous expression. whereas ␤-actin was homogeneously observed in poorly differ- Although the biological significance of cellular translocation of entiated tumor cells with lymph node metastasis (Fig. 7C). moesin is unclear, there are several explanations for these find- Twenty-six (92.8%) of 28 patients whose tumors showed mem- ings. Firstly, conformational and functional change of moesin branous expression pattern of moesin showed membranous ex- results in redistribution of this molecule in tumor cells. Inactive ␤ pression of -actin, and 31 (83.8%) of 37 patients whose tumors moesin is self-associated between the COOH-terminal domain showed cytoplasmic expression pattern of moesin showed and NH -terminal domain existing in the cytoplasm (24). Upon ␤ 2 mixed expression pattern of -actin. However, no patients receipt of appropriate activators, phosphatidylinositol 4,5- ␤ showed cytoplasmic expression pattern of -actin. Significant bisphosphate (25) or phosphorylation of Thr558 (26), moesin ␤ correlation between expression patterns of moesin and -actin translocates from the cytoplasm to the juxtamembrane by dis- Ͻ was observed (P 0.0001; Table 3). ruption of its intramolecular binding. Thus, the change of bal- ance of activator and/or inactivator for moesin may bring about DISCUSSION cellular translocation of the molecule. It has been suggested that In this study, we noted that expression pattern of moesin phosphatidylinositol 4,5-bisphosphate production is activated correlated with tumor size, cervical lymph node metastasis, by oncogenic Ras through phosphatidylinositol 3-kinase and G protein Rac-induced malignant transformation (27). Secondly, CD44, a cell surface receptor involved in cell adhesion, tumor Table 3 Correlation between expression patterns of moesin and invasion, and metastasis, has been cleaved by membrane-type 1 ␤ a -actin in primary oral squamous cell carcinoma matrix metalloproteinase in carcinoma cells at a membrane- Expression pattern of ␤-actin proximal domain, thereby suggesting that functional moesin Membranous Mixed Cytoplasmic Total migrates with CD44 degraded from the cell surface to the cytoplasm (28). Thirdly, because significant correlation between Expression pattern of moesin ␤ Membranous 26 2 0 28 expressions pattern of moesin and -actin was observed in Mixed 24 14 0 38 primary carcinoma tissues, change of moesin distribution in Cytoplasmic 6 31 0 37 tumor cells may reflect organization of the actin cytoskeleton Total 56 47 0 103 and an altered cellular environment. Fourthly, according to a P Ͻ 0.0001 by Spearman’s rank correlation. carcinogenesis, it is possible that mutant of moesin cannot

cross-link between plasma membrane and actin filament, which REFERENCES shows an increase in the cytoplasm of neoplastic cells. 1. Algrain, M., Turunen, O., Vaheri, A., Louvard, D., and Arpin, M. Of interest, our previous studies found that OSCC dis- Ezrin contains cytoskeleton and membrane-binding domains accounting played higher rate of cytoplasmic expression pattern when com- for its proposed role as a membrane-cytoskeletal linker. J. Cell Biol., pared with the pattern in normal oral epithelium, oral epithelial 120: 129–139, 1993. dysplasia, and verrucous carcinoma (18). Furthermore, it has 2. Vaheri, A., Carpen, O., Heiska, L., Halender, T. S., Jaaskelainen, J., Majander-Nordenswan, P., Sainio, M., Timonen, T., and Turunen, O. been reported that squamous cell carcinoma in the skin showed The ezrin protein family: membrane-cytoskeleton interactions and dis- cytoplasmic expression of moesin (29). However, another study ease association. Curr. Opin. Cell Biol., 9: 659–666, 1997. demonstrated that membranous labeling of ezrin was highly 3. Crepaldi, T., Gautreau, A., Comglio, P., Louvard, D., and Arpin, M. observed in metastatic tissues as compared with the primary Ezrin is an effector of hepatocyte growth factor-mediated migration and tissues or hyperplastic specimens (30). A recent study indicated morphogenesis in epithelial cells. J. Cell Biol., 138: 423–434, 1997. that ezrin and moesin might be regulated differently because the 4. Bretscher, A. Regulation of cortical structure by the ezrin-radixin- moesin protein family. Curr. Opin. Cell Biol., 11: 109–116, 1999. binding activity of ezrin and moesin to L-selectin differed when 5. Tsukita, S., and Yonemura, S. ERM (ezrin/radixin/moesin) family: phorbol myristate acetate stimulation and protein kinase C in- From cytoskeleton to signal transduction. Curr. Opin. Cell Biol., 9: hibitor were used (31). Thus, the discrepancy between the study 70–75, 1997. and our series may suggest different regulation among ERM 6. Tsukita, S., and Yonemura, S. Cortical actin organization: Lessons proteins and a distinct role for moesin, depending on the type of from ERM (ezrin/radixin/moesin) proteins. J. Biol. Chem., 274: 34507– cancer. 34510, 1999. Interestingly, all specimens in the primary tumors revealed 7. Tsukita, Sa., Oishi, K., Sato, N., Sagara, J., Kawai, A., and Tsukita, Sh. ERM family members as molecular linkers between the cell surface heterogeneous expression pattern of moesin, but most speci- glycoprotein CD44 and actin-based cytoskeletons. J. Cell Biol., 126: mens in the metastatic lymph nodes homogeneously showed 391–401, 1994. cytoplasmic expression of moesin. We detected cytoplasmic 8. Yonemura, S., Hirao, M., Doi, Y., Takahashi, N., Kondo, T., Tsukita, expression of moesin in the marginal carcinoma cells of the Sa., and Tsukita, Sh. Ezrin/radixin/moesin (ERM) proteins bind to a nests, which seem to constitute the advancing front of cancer positively charged amino acid cluster in the juxta-membrane cytoplasmic domain of CD44, CD43, and ICAM-2. J. Cell Biol., 140: 885–895, invasion. We speculate that because carcinoma cells at the 1998. invasive front in primary tumors show cytoplasmic expression 9. Heiska, L., Alfthan, K., Gronholm, M., Vilja, P., Vaheri, A., and of moesin, which greatly degrade the extracellular matrix as Carpen, O. Association of ezrin with intercellular adhesion molecule-1 well as CD44, these cells probably are involved in pathological and -2 (ICAM-1 and ICAM-2). J. Biol. Chem., 273: 21893–21900, processes of tumor invasion and metastasis. 1998. The tumor size and regional lymph node involvement 10. Martuza, R. L., and Eldridge, R. Neurofibromatosis 2 (bilateral acoustic neurofibromatosis). N. Engl. J. Med., 318: 684–688, 1988. are known as indicators for tumor aggressiveness and poor 11. Ichikawa, T., Masumoto, J., Kaneko, M., Saida, T., Sagara, J., and outcome in OSCC patients. In this study, we observed the Taniguchi, S. Moesin and CD44 expression in cutaneous melanocytic same association using univariate analysis. Furthermore, our tumours. Br. J. Dermatol., 138: 763–768, 1998. data indicated that expression pattern of moesin has a prog- 12. Tokunou, M., Niki, T., Saitoh, Y., Imamura, H., Sakamoto, M., and nostic value: patients whose tumors showed cytoplasmic Hirohashi, S. Altered expression of the ERM proteins in lung adeno- expression had a poorer overall survival compared with pa- carcinoma. Lab. Investig., 80: 1643–1650, 2000. tients whose tumors showed cell membranous expression. In 13. Bohling, T., Turunen, O., Jaaskelainen, J., Carpen, O., Sainio, M., Wahlstrom, T., Vaheri, A., and Haltia, M. Ezrin expression in stromal this study, multivariate analysis demonstrated dramatically cells of capillary hemangioblastoma. An immunohistochemical survey that the expression pattern of moesin is the only independent of brain tumors. Am. J. Pathol., 148: 367–373, 1996. prognostic factor in patients with OSCC. In view of its 14. Fazioli, F., Wong, W. T., Ullrich, S. J., Sakaguchi, K., Appella, E., clinical usefulness, early detection methods clearly predict and Di Fiore, P. P. The ezrin-like family of tyrosine kinase substrates: locoregional lymph node metastasis as well as poor clinical receptor-specific pattern of tyrosine phosphorylation and relationship to malignant transformation. Oncogene, 8: 1335–1345, 1993. outcome may improve strategic planning. We therefore rec- 15. Kaul, S. C., Mitsui, Y., Komatsu, Y., Reddel, R. R., and Wadhwa, ommend assessing the expression pattern of moesin using R. A highly expressed 81-kDa protein in immortalized mouse fibroblast: pretreatment specimens in OSCC clinical trials in an effort to its proliferative function and identity with ezrin. Oncogene, 13: 1231– validate the most appropriate treatments. 1237, 1996. This is the first study to state that the expression pattern of 16. Ohtani, K., Sakamoto, H., Rutherford, T., Chen, Z., Satoh, K., and Naftolin, F. Ezrin, a membrane-cytoskeletal linking protein, is involved moesin is an independent prognostic indicator for patients with in the process of invasion of endometrial cancer cells. Cancer Lett., 147: OSCC. The expression pattern of moesin may be a clinically 31–38, 1999. useful marker for selection of patients for specific treatments; 17. Akisawa, N., Nishimori, I., Iwamura, T., Onishi, S., and Holling- moreover, modulation of the expression pattern of moesin is a sworth, M. A. High levels of ezrin expressed by human pancreatic potential therapeutic strategy for improving clinical outcome in adenocarcinoma cell lines with high metastatic potential. Biochem. Biophys. Res. Commun., 258: 395–400, 1999. OSCC patients. 18. Kobayashi, H., Sagara, J., Masumoto, J., Kurita, H., Kurashira, K., and Taniguchi, S. Shift in cellular localization of moesin in normal oral epithelium, oral epithelial dysplasia, verrucous carcinoma and oral sqau- ACKNOWLEDGMENTS mous cell carcinoma. J. Oral Pathol. Med., 32: 344–349, 2003. We thank Drs. Shoichiro Tsukita and Sachiko Tsukita for their kind 19. Pindborg, J. J., Reichert, P. A., Smith, C. J., and van der Waal, I. gift of anti-moesin monoclonal antibody (CR-22). World Health Organization International Histological Classification of

Tumours: Histological Typing of Cancer and Precancer of the Oral teins and regulates their head-tail association. J. Cell Biol., 140: 647– Mucosa. Berlin: Springer Publ., 1971. 657, 1998. 20. Jakobsson, P. A., Eneroth, C. M., Killander, D., Moberger, G., and 26. Matsui, T., Yonemura, S., Tsukita, S., and Tsukita, S. Activation Martensson, B. Histologic classification and grading of malignancy in of ERM proteins in vivo by Rho involves phosphatidyl-inositol carcinoma of the larynx. Acta Radiol. Ther. Phys. Biol., 12: 1–8, 1973. 4-phosphate 5-kinase and not ROCK kinases. Curr. Biol., 9: 1259– 21. Morifuji, M., Taniguchi, S., Sakai, H., Nakabeppu, Y., and Ohishi, 1262, 1999. M. Differential expression of cytokeratin after orthotopic implantation 27. Takenawa, T., Itoh, T., and Fukami, K. Regulation of phosphatidy- of newly established human tongue cancer cell lines of defined meta- linositol 4,5-biphosphate levels and its roles in cytoskeletal re-organi- static ability. Am. J. Pathol., 156: 1317–1325, 2000. zation and malignant transformation. Chem. Phys. Lipids, 98: 13–22, 22. Sato, N., Yonemura, S., Obinata, T., Tsukita, Sa., and Tsukita, Sh. 1999. Radixin, a barbed end-capping actin-modulating protein, is concentrated 28. Kajita, M., Itoh, Y., Chiba, T., Mori, H., Okada, A., Kinoh, H., and at the cleavage furrow during cytokinesis. J. Cell Biol., 113: 321–330, Seiki, M. Membrane-type 1 matrix metalloproteinase cleaves CD44 and 1991. promotes cell migration. J. Cell Biol., 153: 893–904, 2001. 23. Masumoto, J., Sagara, J., Hayama, M., Hidaka, E., Katsuyama, T., 29. Ichikawa, T., Masumoto, J., Kanoko, M., Saida, T., Sagara, J., and and Taniguchi, S. Differential expression of moesin in cells of hema- Taniguchi, J. Expression of moesin and its associated molecule CD44 in topoietic lineage and lymphatic systems. Histochem. Cell Biol., 110: epithelial skin tumors. J. Cutan. Pathol., 25: 237–243, 1998. 33–41, 1998. 30. Ohtani, K., Sakamoto, H., Rutherford, T., Chen, Z., Kikuchi, A., 24. Gray, R., and Bretscher, A. Ezrin self-association involves bind- Yamamoto, T., Satoh, K., and Naftolin, F. Ezrin, a membrane-cytoskel- ing of an N-terminal domain to a normally masked C-terminal etal linking protein, is highly expressed in atypical endometrial hyper- domain that includes the F-actin binding site. Mol. Biol. Cell, 6: plasia and uterine endometrioid adenocarcinoma. Cancer Lett., 179: 1061–1075, 1995. 79–86, 2002. 25. Matsui, T., Maeda, M., Doi, Y., Yonemura, S., Amano, M., Kai- 31. Ivetic, A., Deka, J., Ridley, A., and Ager, A. The cytoplasmic tail of buchi, K., Tsukita, Tsukita, Sa., and Tsukita, Sh. Rho-kinase phospho- L-selectin interacts with members of the ezrin-radixin-moesin (ERM) rylates COOH-terminal threonines of ezrin/radixin/moesin (ERM) pro- family of proteins. J. Biol. Chem., 277: 2321–2329, 2002.

Downloaded from clincancerres.aacrjournals.org on October 3, 2021. © 2004 American Association for Cancer Research. Clinical Significance of Cellular Distribution of Moesin in Patients with Oral Squamous Cell Carcinoma

Hiroichi Kobayashi, Junji Sagara, Hiroshi Kurita, et al.

Clin Cancer Res 2004;10:572-580.

Updated version Access the most recent version of this article at: http://clincancerres.aacrjournals.org/content/10/2/572

Cited articles This article cites 30 articles, 11 of which you can access for free at: http://clincancerres.aacrjournals.org/content/10/2/572.full#ref-list-1

Citing articles This article has been cited by 4 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/10/2/572.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/10/2/572. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.