oral pathology Editor: LEWIS R. EVERSOLE, DDS, MSD, MA
Oral Diagnosis, Medicine & Pathology School of Dentistry 53-058 UCLA Health Sciences Center Los Angeles, California 90024
Immunoprofile of mucoepidermoid carcinomas of minor salivary glands Joseph A. Regezi, DDS, MS,” Richard J. Zarbo. MD, DMD,b and John G. Batsakis. IUD.~ Ann Arbor and Detroit, Mich., and Houston, Texas UNIVERSITY ANDERSON
OF MICHIGAN, HENRY FORD HOSPITAL/WAYNE CANCER CENTER, UNIVERSITY OF TEXAS
STATE UNIVERSITY,
AND M. D.
Because the data on the antigenic phenotype of mucoepidermoid carcinoma (MEC) are incomplete and for low- and high-molecular-weight somewhat disparate, 45 MECs were evaluated immunohistochemically keratins, vimentin. g&al fibrillary acidic protein, smooth muscle actin, and S- 100 protein. Tumors stained uniformly for keratins and, on occasion, focally for vimentin. Tumors were nonreactive with antibodies to glial fibrillary acidic protein and, with few exceptions, to muscle-specific actins and S-100 protein. Clear cell and papillary histologic variants were seen as potential diagnostic pitfalls. If used with hematoxylin-and-eosin-stained sections, limited potential is seen for this antibody panel in surgical pathology. Myoepithelial cell-associated antigens are expressed to a very limited extent in MECs. (ORAL SURG ORAL MED ORAL PATHOL
1991;71:189-92)
A
lthough unique cellular antigens have not been defined in any salivary gland neoplasms, panels of multiple, less specific immunomarkers may be of use in surgical pathology. Antigenic profiles, as determined through immunohistochemical studies, may provide enough selectivity to serve as adjuncts to diagnosis and classification. Also, with the assumption that neoplastic cells express antigens that are found on their normal cell counterparts, knowledge of immunophenotypes may be helpful in elucidating the level of myoepithelial participation/differentiation in these neoplasms. Some antigenic markers on mucoepidermoid carcinoma (MEC) cells have been confirmed. A wide Qepartment of Oral Medicine/Pathology/Surgery, School of Dentistry, University of Michigan, Ann Arbor, bDepartment of Pathology, Henry Ford Hospital/Wayne State University, Detroit. %hairman, Department of Pathology, M. D. Anderson Cancer Center, University of Texas, Houston. 7/14/23051
spectrum of keratins has been identified for MECs. With some notable exceptions,‘-4 coexpression of other intermediate filaments, vimentin, desmin, and glial fibrillary acidic protein (GFAP) generally has not been a feature of MECS.~-‘OPositive immunoreactivity for actin filaments has been reported by one group of investigators,4 and myosin filament stains were reported as negative by another.6 Only S-100 protein-positive dendritic and phagocytic cells have been found in otherwise negative MECs by most investigators,5v‘2 1l* I2 whereas S- lOOpositive tumor cells have been reported by others.4*I3 Immunoreactivity for macrophage-associated antigens antichymotrypsin and Leu-M 13 (but not lysozymei4) has also been noted. MECs are thought to arise from an undifferentiated cell(s) associated with salivary gland excretory ducts. ’ 5 Because putative myoepithelial cells have beenreported recently in the baseof theseducts,16and becausesome myoepithelial cell-associated antigens (vimentin, GFAP, and actin) have beenfound in some MECs, it has been suggested that the myoepithelial 189
190
Regezi, Zarbo, and Batsakis
ORAL SURG ORAL
MED ORAL
PATHOL
February
Table
I. Antibodies used
Antigen
( Dilution
LMK
1:5000
HMK Vimentin GFAP Muscle actins
1:lOOO 1:25 1:3000 1:75,000
S- 100 protein
I:5000
/
Source Enzo Biochemical, New York, N.Y. Enzo Biochemical Dako, Santa Barbara, Calif. Enzo Biochemical Dr. A. M. Gown, University of Washington, Seattle HSC Research Development Corp., Toronto, Canada
cell may have a significant role in the development of MECs. Nonetheless, many questions remain because of variable morphologic*7-‘9 and immunologic findings (previously cited). Because the data on MEC antigenic phenotype are incomplete and somewhat disparate, additional studies are warranted. The purpose of this investigation, therefore, is to evaluate immunohistochemically several casesto define further antigenic determinants of MECs. This information may help define a tumor immunoprofile that could be useful in diagnosis and classification, and it may help answer some of the questions relating to the role of myoepithelium in the genesisof MECs. MATERIAL
AND METHODS
Diagnosis was confirmed microscopically on 42 MECs of minor salivary glands (University of Michigan oral pathology files) with accepted histopathologic criteria.20-22Clinical data were compiled, and histologic variations were recorded. Sections for immunohistochemistry were cut from formalin-fixed tissue in paraffin blocks and mounted on gelatin-coated slides. Antibodies to low-molecularweight keratin (LMK; clone 35BH11, cytokeratin No. 8 of Moll catalog23), and high-molecular-weight keratins (HMK, clone 34BE12, cytokeratin Nos. 1,5, 10,14), vimentin (clone V9), GFAP (clone GFP-8A), muscle-specific actins (clone HHF-35), and S-100 protein (polyclonal) were evaluated on dewaxed tissue sections by an avidin-biotin-peroxidase method (Table I). All sectionswere digested with 0.4% pepsin (Sigma Chemical Co., St. Louis, MO.) for 30 minutes and incubated overnight at 4” C with primary antibody. Subsequent incubations were with biotinylated IgG followed by avidin-biotin-peroxidase complex (Vector Laboratories, Burlingame, Calif.). Sections were developedin aminoethylcarbazole and counterstained with Mayer’s hematoxylin. Sections from a multitissue normal organ sausage block were used for antibody titration and assessment
199 I
of specificity and sensitivity of antigen detection in each run.24 Two substitution-negative controls of nonimmune rabbit serum or immunoglobulin of the same species, isotype, and protein concentration as the primary antibody were used. Only casesin which internal positive controls were reactive were included in the study. Immunoreactivity was scored on a scale of 0 to 4+, with 0 as negative or nonreactive, 1t representing scattered spotty staining, 2t indicating up to 25% of tumor cells positive, 3+ indicating 25% to 50% of tumor cells positive, and 4t indicating more than 50% of tumor cells positive. Immunoreactivity was also evaluated relative to the various cell types found in these tumors (mucous cells, intermediate cells, duct cells, clear cells, and dendritic cells). RESULTS
The patient population was composedof 19 males and 23 females, with an average age of 46 years and an age range of 12 to 82 years. The regional distribution of the tumors was as follows: hard palate, 12; buccal mucosa, 10; retromolar pad, 7; floor of mouth, 5; soft palate, 3; lower lip, 2; upper lip, 1; gingiva, 1; and mandible (intraosseous), 1. Histopathology
MECs typically exhibited a lobular and infiltrative growth pattern. Microcystic spaceswere frequently seen in the well-differentiated lesions, and in nine cases only a single cystic space with intraluminal and/or intramural proliferation was found. Papillary patterns were observed focally in three of the welldifferentiated lesions (Fig. 1). Foci of clear cells were noted in five cases. Stroma between lobules of neoplastic cells frequently showed fibroblastic proliferation and occasional scarring and hyalinization. Free mucin was infrequently seen in these neoplasms. A patchy lymphocytic infiltrate was occasionally seen. Twenty-five lesions were classified as low-grade tumors, typified by well-differentiated mucous cells and cystic spaces.Thirteen were intermediate-grade tumors, which characteristically contained fewer mucous cells and exhibited a more solid and lobular growth pattern. In the four tumors that were diagnosed as high-grade MECs, few mucous cells were seen.These lesions were poorly differentiated, showing nuclear pleomorphism and increased nuclear cytoplasmic ratios. lmmunohistochemistry
All tumors were positive (3+ or 4+) for LMK and HMK (Table II). A distinctive staining pattern did not emerge for either antibody. With both keratin antibodies, ductal cells and intermediate cells were consistently positive, clear cells were occasionally
Volume 7 I Number
Fig.
Immunohistochemistry
of mucoepidermoid
191
carcinomas
2
1. Papillary
and cystic focus in low-grade MEC.
positive, and mucous cells were usually negative, although basal staining was occasionally noted. There was also no difference in staining intensity for the various MEC subtypes. Staining of tiny nests of tumor cells at the periphery of these lesions was helpful in locating invasive neoplastic cells at surgical margins. In normal minor salivary gland, ductal cells were positive for LMK, but mucous cells were usually nonreactive. Normal ductal cells and acinar cells were frequently positive for HMK. Occasional basal cells or acini and excretory ducts stained positive for HMK, and some of the ductal cells exhibited dendritic processes. Positive vimentin staining (l+ or 2+) was seen, usually in small foci, 16 in MECs (all three subtypes) and was usually found in the intermediate cells (Fig. 2). Vimentin reactivity was also seenin dendritic cells in three cases.In normal glands vimentin staining was frequently found in myoepithelial cells. All tumors were nonreactive for GFAP. In normal gland, myoepithelial cells were occasionally positive. Generally, MECs were nonreactive for smooth muscle actin. However, two intermediate-grade tumors exhibited a 2+ reaction in intermediate tumor cells. In normal salivary gland, myoepithelium was positive for smooth muscle actin and ducts were negative. One low-grade lesion showed positive nuclear and cytoplasmic staining (2+) of neoplastic (nonmucin-producing) ductal cells; all other MECs were nonreactive for S- 100 protein. Dendritic cells reactive for S-100 protein were noted in four lesions. In normal gland, myoepithelial cells and intercalated duct cells were positive for S-100 protein. DISCUSSION
Most MECs of minor salivary gland were welldifferentiated lesions falling into the low- or intermediate-grade class. The various subtypes (low, intermediate, high grades) did not exhibit any significant immunohistochemical staining differences, indicating
Fig. 2. Positivevimentin stainof low-grade MEC. Darker cells in center (arrows) are immunoreactive. Note staining within connective tissue partition (top right).
Table
II. Immunoreactivity s-100 LMK
HMK
Vimentin
MEC
+
+
16142
Normal gland
+
+
Myo+
+.
Positive
immunostaining;
-,
negative:
GFAP
Actin
protein
-
-*
-t
Myo+ Myo+ Myo/duct+ duct, intercalcated
ducts;
myo,
myoepithelium. *Two
cases showed
tone
case showed
2+ reactivity. 2+ reactivity.
that, relative to the antigens studied, immunohistochemistry was of no value in the subclassification of these lesions. Of histologic interest was the appearanceof occasionalclear cell or papillary foci in MECs, a possible point of confusion with other salivary gland lesions.25-28 It is well known that each of the five antigenically distinct intermediate filaments (keratins, vimentin, desmin, GFAP, neurofilaments) that are limited individually to various normal cells, also generally characterize their neoplastic counterparts. Therefore immunostaining neoplasmsfor intermediate filament neoplasmsshould provide information relative to their general origin or classification. Because salivary glands are of epithelial origin, keratin intermediate filaments are found in parenchymal cells and their neoplastic counterparts. In the MECs studied, both LMK and HMK were found, but distinct histologic cell types were not identified. Although specific keratin subtypes have not been identified individually in MECs, the presenceof LMK of simple epithelium would be expected to dominate over HMK of keratinized epithelium and myoepithelium. Because normal myoepithelial cells express antigenie determinants for keratins, vimentin, GFAP,
192 Regezi, Zarbo, and Batsakis smooth muscle actin, and S- 100 protein, any of these may be expressedby various salivary gland tumors. Coexpression of intermediate filaments in salivary gland tumors has been reported, and as many as four types (keratin, vimentin, GFAP, desmin) have been identified in sometumo& the reason for this is likely related to the myoepithelial cell. In this study coexpression of keratin and vimentin was seen. However, the former was consistently present and the latter infrequent and in small foci. The other antigenic determinants associated with myoepithelial cells, GFAP, actin, and S-100 protein were minimally expressed.The relatively low reactivity for these antigens may be of somevalue in tumor identification, especially when compared with other salivary gland tumors and taken in the context of hematoxylin-and-eosin-stained sections. MECs appear to be composed predominantly of pure epithelial cells. Myoepithelial differentiation appears to be minimal as evidenced by the very limited expression of antigenic determinants that normally characterize these cells. REFERENCES I. Gustafsson H, Virtanen I, Thornell L. Expression of cytokeratins and vimentin in salivary gland carcinomas as revealed with monoclonal antibodies. Virchows Arch ]A] 1988;412: 515-24. 2. Gustafsson H, Virtanen 1, Thornell L. Glial fibrillary acidic protein and desmin in salivary neoplasms.Virchows Arch [B] 1989;57:303-13. 3. Hamper K, Schmitz-Watjen W, Mausch H, Caselitz J, Seifert G. Multiple expression of tissue markers in mucoepidermoid carcinomas and acinic carcinomas of the salivary glands. Virchows Arch IA1 1989:414:407-l3. 4. Hassanin M, Ghosh L, Das A, Waterhouse J. Immunohistochemical and fluorescentmicroscopic study of histogenesisof salivary mucoepidermoid carcinoma. J Oral Pathol Med 1989;18:291-8. 5. Nakazato Y, lshida Y, Takahashi K, Suzuki K. Immunohistochemical distribution of S-100 protein and glial fibrillary acidic protein in normal and neoplastic salivary glands. Virchows Arch [A] 1985;405:299-310. 6. Matsushima R. Nakavama I. Shim& M. Immunohistochemical localization of keiatin, vimentin, and myosin in salivary gland tumors. Acta Path01Jpn 1988;38:445-54. I. Regezi J, Lloyd R, Zarbo J, McClatchey K. Minor salivary gland tumors: a histologic and immunohistochemical study. Cancer 1985;55:108-15. 8. Seifert G. The importanceof tumor markers in oral pathology. Pathol Res Pratt 1985;179:625-8. 9. Seifert G, Caselitz J. Markers of oral and salivary gland tumors: immunocytochemical investigations. Cancer Detect Prev 1985;8:23-4. 10. Zarbo R, Regezi J, Hatfield J, et al. Immunoreactive glial fibrillary acidic protein in normal and neoplastic salivary glands: a combined immunohistochemical and immunoblot study. Surg Path01 1988;1:55-63.
ORAL S~JRG ORAL
MED ORAL PATHOL February 199I
11 Cracker J, Jenkins R, Campbell J, Fuggle W, Shah V. Immu-
nohistochemical demonstration of S-100 protein in salivary gland neoplasms. J Path01 1985;146:115-21. 12. Zarbo R, Regezi J, Batsakis J. S-100 protein in salivary gland tumors: an immunohistochemical study of 129 cases. Head Neck Surg 1986;8:268-75. 13. Kahn H, Baumal R, Marks A, Dardick I, Van Nostrand P. Myoepithelial cells in salivary gland tumors. Arch Path01Lab Med 1985;109:190-5. 14. Saito I, Teratani K, Inoue M, Fanatsu K, Moro I. Immunohistochemical characterization of functional markers in human minor salivary gland tumors. J Oral Path01 1984;13: 52.5-34. 15. Regezi J, Batsakis J. Histogenesisof salivary gland neoplasms. Otolarvnnol Clin North Am 1977:10:297-307. 16. Dardick y, Rippstein P, Skimming L, Boivin M, Parks W, Dairkee S. Immunohistochemistry and ultrastructure of myoepithelium and modified myoepithelium of the ducts of human major salivary glands: histogenetic implications for salivary gland tumors. ORAL SURC ORAL. MED ORAL PATHOL 1987; 64:703-15. 17. Dardick 1, Daya D, Hardie J, van Nostrand A. Mucoepidermoid carcinoma: ultrastructural and histogenetic aspects. J Oral Path01 1984;13:342-58. 18. Nikai H, El-Bardaie M, Takata T, Ogawa I, Ijuhin N. Histologic evaluation of myoepithelial participation in salivary gland tumors. Int J Oral Maxillofac Surg-1986;15:597-6051 19. Nicolatou 0. Harwick R. Putone - P. Leifer C. Ultrastructural characterization of intermediate ceils of mucoepidermoid carcinoma of the parotid. Oral Surg 1979;48:324-36. 20. Evans H. Mucoepidermoid carcinoma of salivary glands: a study of 69 caseswith special attention to histologic grading. Am J Clin Path01 1984~81:696-701. 21. Hamper K, Schimmelpenning H, Caselitz J, et al. Mucoepidermoid tumors of the salivary glands: correlation of cytophotometrical data and prognosis. Cancer 1989163:707-l7. 22. Nascimento A, Amaral A, Prado L, Kligerman J, Silveira T. Mucoepidermoid carcinoma of salivary glands: a clinicopathologic study of 46 cases. Head Neck Surg 1986;8:409-17. 23. Moll R, Franke W, Schiller D, Greiger B, Krepler R. The catalog of human cytokeratins: patterns of expression in normal enithelia. tumors. and cultured cells. Cell 1982;31:1l-24. 24. Battifora H. Methods in laboratory investigation: the multitumor (sausage) tissue block. Lab Invest 1986;55:244-8. 25. Fliss D, Zirkin H, Puterman M, Tovi F. Low-grade papillary adenocarcinoma of buccal mucosasalivary gland origin. Head Neck Surg 1989;11:237-41. 26. Luna MA, Ordonez NG, Mackay B, Batsakis JG, Guillamondegui 0. Salivary epithelial-myoepithelial carcinomas of intercalated ducts: a clinical, electron microscopic, and immunocytochemical study. ORAL SURG ORAL MED ORAL PATHOL 1985;59:482-90.
21. Mills S, Garland T, Allen M. Low-grade papillary adenocarcinema of palatal salivary glands. Am J Surg Pathol 1984; 81367-74. 28. Palmer R. Epithelial-myoepithelial carcinoma: an immunocytochemical study. ORAL SURG ORAL MED ORAL PATHOL 1985;59:51l-5. Reprint requests to:
Joseph A. Regezi, DDS, MS Department of Oral Pathology School of Dentistry University of Michigan Ann Arbor. MI 48104-1078
Oral Diagnosis, Medicine & Pathology School of Dentistry 53-058 UCLA Health Sciences Center Los Angeles, California 90024
Immunoprofile of mucoepidermoid carcinomas of minor salivary glands Joseph A. Regezi, DDS, MS,” Richard J. Zarbo. MD, DMD,b and John G. Batsakis. IUD.~ Ann Arbor and Detroit, Mich., and Houston, Texas UNIVERSITY ANDERSON
OF MICHIGAN, HENRY FORD HOSPITAL/WAYNE CANCER CENTER, UNIVERSITY OF TEXAS
STATE UNIVERSITY,
AND M. D.
Because the data on the antigenic phenotype of mucoepidermoid carcinoma (MEC) are incomplete and for low- and high-molecular-weight somewhat disparate, 45 MECs were evaluated immunohistochemically keratins, vimentin. g&al fibrillary acidic protein, smooth muscle actin, and S- 100 protein. Tumors stained uniformly for keratins and, on occasion, focally for vimentin. Tumors were nonreactive with antibodies to glial fibrillary acidic protein and, with few exceptions, to muscle-specific actins and S-100 protein. Clear cell and papillary histologic variants were seen as potential diagnostic pitfalls. If used with hematoxylin-and-eosin-stained sections, limited potential is seen for this antibody panel in surgical pathology. Myoepithelial cell-associated antigens are expressed to a very limited extent in MECs. (ORAL SURG ORAL MED ORAL PATHOL
1991;71:189-92)
A
lthough unique cellular antigens have not been defined in any salivary gland neoplasms, panels of multiple, less specific immunomarkers may be of use in surgical pathology. Antigenic profiles, as determined through immunohistochemical studies, may provide enough selectivity to serve as adjuncts to diagnosis and classification. Also, with the assumption that neoplastic cells express antigens that are found on their normal cell counterparts, knowledge of immunophenotypes may be helpful in elucidating the level of myoepithelial participation/differentiation in these neoplasms. Some antigenic markers on mucoepidermoid carcinoma (MEC) cells have been confirmed. A wide Qepartment of Oral Medicine/Pathology/Surgery, School of Dentistry, University of Michigan, Ann Arbor, bDepartment of Pathology, Henry Ford Hospital/Wayne State University, Detroit. %hairman, Department of Pathology, M. D. Anderson Cancer Center, University of Texas, Houston. 7/14/23051
spectrum of keratins has been identified for MECs. With some notable exceptions,‘-4 coexpression of other intermediate filaments, vimentin, desmin, and glial fibrillary acidic protein (GFAP) generally has not been a feature of MECS.~-‘OPositive immunoreactivity for actin filaments has been reported by one group of investigators,4 and myosin filament stains were reported as negative by another.6 Only S-100 protein-positive dendritic and phagocytic cells have been found in otherwise negative MECs by most investigators,5v‘2 1l* I2 whereas S- lOOpositive tumor cells have been reported by others.4*I3 Immunoreactivity for macrophage-associated antigens antichymotrypsin and Leu-M 13 (but not lysozymei4) has also been noted. MECs are thought to arise from an undifferentiated cell(s) associated with salivary gland excretory ducts. ’ 5 Because putative myoepithelial cells have beenreported recently in the baseof theseducts,16and becausesome myoepithelial cell-associated antigens (vimentin, GFAP, and actin) have beenfound in some MECs, it has been suggested that the myoepithelial 189
190
Regezi, Zarbo, and Batsakis
ORAL SURG ORAL
MED ORAL
PATHOL
February
Table
I. Antibodies used
Antigen
( Dilution
LMK
1:5000
HMK Vimentin GFAP Muscle actins
1:lOOO 1:25 1:3000 1:75,000
S- 100 protein
I:5000
/
Source Enzo Biochemical, New York, N.Y. Enzo Biochemical Dako, Santa Barbara, Calif. Enzo Biochemical Dr. A. M. Gown, University of Washington, Seattle HSC Research Development Corp., Toronto, Canada
cell may have a significant role in the development of MECs. Nonetheless, many questions remain because of variable morphologic*7-‘9 and immunologic findings (previously cited). Because the data on MEC antigenic phenotype are incomplete and somewhat disparate, additional studies are warranted. The purpose of this investigation, therefore, is to evaluate immunohistochemically several casesto define further antigenic determinants of MECs. This information may help define a tumor immunoprofile that could be useful in diagnosis and classification, and it may help answer some of the questions relating to the role of myoepithelium in the genesisof MECs. MATERIAL
AND METHODS
Diagnosis was confirmed microscopically on 42 MECs of minor salivary glands (University of Michigan oral pathology files) with accepted histopathologic criteria.20-22Clinical data were compiled, and histologic variations were recorded. Sections for immunohistochemistry were cut from formalin-fixed tissue in paraffin blocks and mounted on gelatin-coated slides. Antibodies to low-molecularweight keratin (LMK; clone 35BH11, cytokeratin No. 8 of Moll catalog23), and high-molecular-weight keratins (HMK, clone 34BE12, cytokeratin Nos. 1,5, 10,14), vimentin (clone V9), GFAP (clone GFP-8A), muscle-specific actins (clone HHF-35), and S-100 protein (polyclonal) were evaluated on dewaxed tissue sections by an avidin-biotin-peroxidase method (Table I). All sectionswere digested with 0.4% pepsin (Sigma Chemical Co., St. Louis, MO.) for 30 minutes and incubated overnight at 4” C with primary antibody. Subsequent incubations were with biotinylated IgG followed by avidin-biotin-peroxidase complex (Vector Laboratories, Burlingame, Calif.). Sections were developedin aminoethylcarbazole and counterstained with Mayer’s hematoxylin. Sections from a multitissue normal organ sausage block were used for antibody titration and assessment
199 I
of specificity and sensitivity of antigen detection in each run.24 Two substitution-negative controls of nonimmune rabbit serum or immunoglobulin of the same species, isotype, and protein concentration as the primary antibody were used. Only casesin which internal positive controls were reactive were included in the study. Immunoreactivity was scored on a scale of 0 to 4+, with 0 as negative or nonreactive, 1t representing scattered spotty staining, 2t indicating up to 25% of tumor cells positive, 3+ indicating 25% to 50% of tumor cells positive, and 4t indicating more than 50% of tumor cells positive. Immunoreactivity was also evaluated relative to the various cell types found in these tumors (mucous cells, intermediate cells, duct cells, clear cells, and dendritic cells). RESULTS
The patient population was composedof 19 males and 23 females, with an average age of 46 years and an age range of 12 to 82 years. The regional distribution of the tumors was as follows: hard palate, 12; buccal mucosa, 10; retromolar pad, 7; floor of mouth, 5; soft palate, 3; lower lip, 2; upper lip, 1; gingiva, 1; and mandible (intraosseous), 1. Histopathology
MECs typically exhibited a lobular and infiltrative growth pattern. Microcystic spaceswere frequently seen in the well-differentiated lesions, and in nine cases only a single cystic space with intraluminal and/or intramural proliferation was found. Papillary patterns were observed focally in three of the welldifferentiated lesions (Fig. 1). Foci of clear cells were noted in five cases. Stroma between lobules of neoplastic cells frequently showed fibroblastic proliferation and occasional scarring and hyalinization. Free mucin was infrequently seen in these neoplasms. A patchy lymphocytic infiltrate was occasionally seen. Twenty-five lesions were classified as low-grade tumors, typified by well-differentiated mucous cells and cystic spaces.Thirteen were intermediate-grade tumors, which characteristically contained fewer mucous cells and exhibited a more solid and lobular growth pattern. In the four tumors that were diagnosed as high-grade MECs, few mucous cells were seen.These lesions were poorly differentiated, showing nuclear pleomorphism and increased nuclear cytoplasmic ratios. lmmunohistochemistry
All tumors were positive (3+ or 4+) for LMK and HMK (Table II). A distinctive staining pattern did not emerge for either antibody. With both keratin antibodies, ductal cells and intermediate cells were consistently positive, clear cells were occasionally
Volume 7 I Number
Fig.
Immunohistochemistry
of mucoepidermoid
191
carcinomas
2
1. Papillary
and cystic focus in low-grade MEC.
positive, and mucous cells were usually negative, although basal staining was occasionally noted. There was also no difference in staining intensity for the various MEC subtypes. Staining of tiny nests of tumor cells at the periphery of these lesions was helpful in locating invasive neoplastic cells at surgical margins. In normal minor salivary gland, ductal cells were positive for LMK, but mucous cells were usually nonreactive. Normal ductal cells and acinar cells were frequently positive for HMK. Occasional basal cells or acini and excretory ducts stained positive for HMK, and some of the ductal cells exhibited dendritic processes. Positive vimentin staining (l+ or 2+) was seen, usually in small foci, 16 in MECs (all three subtypes) and was usually found in the intermediate cells (Fig. 2). Vimentin reactivity was also seenin dendritic cells in three cases.In normal glands vimentin staining was frequently found in myoepithelial cells. All tumors were nonreactive for GFAP. In normal gland, myoepithelial cells were occasionally positive. Generally, MECs were nonreactive for smooth muscle actin. However, two intermediate-grade tumors exhibited a 2+ reaction in intermediate tumor cells. In normal salivary gland, myoepithelium was positive for smooth muscle actin and ducts were negative. One low-grade lesion showed positive nuclear and cytoplasmic staining (2+) of neoplastic (nonmucin-producing) ductal cells; all other MECs were nonreactive for S- 100 protein. Dendritic cells reactive for S-100 protein were noted in four lesions. In normal gland, myoepithelial cells and intercalated duct cells were positive for S-100 protein. DISCUSSION
Most MECs of minor salivary gland were welldifferentiated lesions falling into the low- or intermediate-grade class. The various subtypes (low, intermediate, high grades) did not exhibit any significant immunohistochemical staining differences, indicating
Fig. 2. Positivevimentin stainof low-grade MEC. Darker cells in center (arrows) are immunoreactive. Note staining within connective tissue partition (top right).
Table
II. Immunoreactivity s-100 LMK
HMK
Vimentin
MEC
+
+
16142
Normal gland
+
+
Myo+
+.
Positive
immunostaining;
-,
negative:
GFAP
Actin
protein
-
-*
-t
Myo+ Myo+ Myo/duct+ duct, intercalcated
ducts;
myo,
myoepithelium. *Two
cases showed
tone
case showed
2+ reactivity. 2+ reactivity.
that, relative to the antigens studied, immunohistochemistry was of no value in the subclassification of these lesions. Of histologic interest was the appearanceof occasionalclear cell or papillary foci in MECs, a possible point of confusion with other salivary gland lesions.25-28 It is well known that each of the five antigenically distinct intermediate filaments (keratins, vimentin, desmin, GFAP, neurofilaments) that are limited individually to various normal cells, also generally characterize their neoplastic counterparts. Therefore immunostaining neoplasmsfor intermediate filament neoplasmsshould provide information relative to their general origin or classification. Because salivary glands are of epithelial origin, keratin intermediate filaments are found in parenchymal cells and their neoplastic counterparts. In the MECs studied, both LMK and HMK were found, but distinct histologic cell types were not identified. Although specific keratin subtypes have not been identified individually in MECs, the presenceof LMK of simple epithelium would be expected to dominate over HMK of keratinized epithelium and myoepithelium. Because normal myoepithelial cells express antigenie determinants for keratins, vimentin, GFAP,
192 Regezi, Zarbo, and Batsakis smooth muscle actin, and S- 100 protein, any of these may be expressedby various salivary gland tumors. Coexpression of intermediate filaments in salivary gland tumors has been reported, and as many as four types (keratin, vimentin, GFAP, desmin) have been identified in sometumo& the reason for this is likely related to the myoepithelial cell. In this study coexpression of keratin and vimentin was seen. However, the former was consistently present and the latter infrequent and in small foci. The other antigenic determinants associated with myoepithelial cells, GFAP, actin, and S-100 protein were minimally expressed.The relatively low reactivity for these antigens may be of somevalue in tumor identification, especially when compared with other salivary gland tumors and taken in the context of hematoxylin-and-eosin-stained sections. MECs appear to be composed predominantly of pure epithelial cells. Myoepithelial differentiation appears to be minimal as evidenced by the very limited expression of antigenic determinants that normally characterize these cells. REFERENCES I. Gustafsson H, Virtanen I, Thornell L. Expression of cytokeratins and vimentin in salivary gland carcinomas as revealed with monoclonal antibodies. Virchows Arch ]A] 1988;412: 515-24. 2. Gustafsson H, Virtanen 1, Thornell L. Glial fibrillary acidic protein and desmin in salivary neoplasms.Virchows Arch [B] 1989;57:303-13. 3. Hamper K, Schmitz-Watjen W, Mausch H, Caselitz J, Seifert G. Multiple expression of tissue markers in mucoepidermoid carcinomas and acinic carcinomas of the salivary glands. Virchows Arch IA1 1989:414:407-l3. 4. Hassanin M, Ghosh L, Das A, Waterhouse J. Immunohistochemical and fluorescentmicroscopic study of histogenesisof salivary mucoepidermoid carcinoma. J Oral Pathol Med 1989;18:291-8. 5. Nakazato Y, lshida Y, Takahashi K, Suzuki K. Immunohistochemical distribution of S-100 protein and glial fibrillary acidic protein in normal and neoplastic salivary glands. Virchows Arch [A] 1985;405:299-310. 6. Matsushima R. Nakavama I. Shim& M. Immunohistochemical localization of keiatin, vimentin, and myosin in salivary gland tumors. Acta Path01Jpn 1988;38:445-54. I. Regezi J, Lloyd R, Zarbo J, McClatchey K. Minor salivary gland tumors: a histologic and immunohistochemical study. Cancer 1985;55:108-15. 8. Seifert G. The importanceof tumor markers in oral pathology. Pathol Res Pratt 1985;179:625-8. 9. Seifert G, Caselitz J. Markers of oral and salivary gland tumors: immunocytochemical investigations. Cancer Detect Prev 1985;8:23-4. 10. Zarbo R, Regezi J, Hatfield J, et al. Immunoreactive glial fibrillary acidic protein in normal and neoplastic salivary glands: a combined immunohistochemical and immunoblot study. Surg Path01 1988;1:55-63.
ORAL S~JRG ORAL
MED ORAL PATHOL February 199I
11 Cracker J, Jenkins R, Campbell J, Fuggle W, Shah V. Immu-
nohistochemical demonstration of S-100 protein in salivary gland neoplasms. J Path01 1985;146:115-21. 12. Zarbo R, Regezi J, Batsakis J. S-100 protein in salivary gland tumors: an immunohistochemical study of 129 cases. Head Neck Surg 1986;8:268-75. 13. Kahn H, Baumal R, Marks A, Dardick I, Van Nostrand P. Myoepithelial cells in salivary gland tumors. Arch Path01Lab Med 1985;109:190-5. 14. Saito I, Teratani K, Inoue M, Fanatsu K, Moro I. Immunohistochemical characterization of functional markers in human minor salivary gland tumors. J Oral Path01 1984;13: 52.5-34. 15. Regezi J, Batsakis J. Histogenesisof salivary gland neoplasms. Otolarvnnol Clin North Am 1977:10:297-307. 16. Dardick y, Rippstein P, Skimming L, Boivin M, Parks W, Dairkee S. Immunohistochemistry and ultrastructure of myoepithelium and modified myoepithelium of the ducts of human major salivary glands: histogenetic implications for salivary gland tumors. ORAL SURC ORAL. MED ORAL PATHOL 1987; 64:703-15. 17. Dardick 1, Daya D, Hardie J, van Nostrand A. Mucoepidermoid carcinoma: ultrastructural and histogenetic aspects. J Oral Path01 1984;13:342-58. 18. Nikai H, El-Bardaie M, Takata T, Ogawa I, Ijuhin N. Histologic evaluation of myoepithelial participation in salivary gland tumors. Int J Oral Maxillofac Surg-1986;15:597-6051 19. Nicolatou 0. Harwick R. Putone - P. Leifer C. Ultrastructural characterization of intermediate ceils of mucoepidermoid carcinoma of the parotid. Oral Surg 1979;48:324-36. 20. Evans H. Mucoepidermoid carcinoma of salivary glands: a study of 69 caseswith special attention to histologic grading. Am J Clin Path01 1984~81:696-701. 21. Hamper K, Schimmelpenning H, Caselitz J, et al. Mucoepidermoid tumors of the salivary glands: correlation of cytophotometrical data and prognosis. Cancer 1989163:707-l7. 22. Nascimento A, Amaral A, Prado L, Kligerman J, Silveira T. Mucoepidermoid carcinoma of salivary glands: a clinicopathologic study of 46 cases. Head Neck Surg 1986;8:409-17. 23. Moll R, Franke W, Schiller D, Greiger B, Krepler R. The catalog of human cytokeratins: patterns of expression in normal enithelia. tumors. and cultured cells. Cell 1982;31:1l-24. 24. Battifora H. Methods in laboratory investigation: the multitumor (sausage) tissue block. Lab Invest 1986;55:244-8. 25. Fliss D, Zirkin H, Puterman M, Tovi F. Low-grade papillary adenocarcinoma of buccal mucosasalivary gland origin. Head Neck Surg 1989;11:237-41. 26. Luna MA, Ordonez NG, Mackay B, Batsakis JG, Guillamondegui 0. Salivary epithelial-myoepithelial carcinomas of intercalated ducts: a clinical, electron microscopic, and immunocytochemical study. ORAL SURG ORAL MED ORAL PATHOL 1985;59:482-90.
21. Mills S, Garland T, Allen M. Low-grade papillary adenocarcinema of palatal salivary glands. Am J Surg Pathol 1984; 81367-74. 28. Palmer R. Epithelial-myoepithelial carcinoma: an immunocytochemical study. ORAL SURG ORAL MED ORAL PATHOL 1985;59:51l-5. Reprint requests to:
Joseph A. Regezi, DDS, MS Department of Oral Pathology School of Dentistry University of Michigan Ann Arbor. MI 48104-1078
