Oral granular cell tumors: A clinicopathologic and immunocytochemical study

Oral granular cell tumors: A clinicopathologic and immunocytochemical study

Oral granular cell tumors: A clinicopathologic and immunocytochemical study Carol M. Stewart, D.D.S., M.S.,” Ronald E. Watson, D.D.S., M.A.E.,b L. R. ...

5MB Sizes 0 Downloads 11 Views

Oral granular cell tumors: A clinicopathologic and immunocytochemical study Carol M. Stewart, D.D.S., M.S.,” Ronald E. Watson, D.D.S., M.A.E.,b L. R. Eversole, D.D.S., M.S.D., M.A.,’ Werner Fischlschweiger, Ph.D.,d and Alan S. Leider, D.D.S., M.A.,’ Gainesville, Fla., and San Francisco, Calif: UNIVERSITY DENTISTRY

OF FLORIDA

COLLEGE

OF DENTISTRY

AND

UNIVERSITY

OF THE PACIFIC

SCHOOL

OF

To investigate the histogenesis of the granular cell, a large series of granular cell tumors was studied for clinical and histopathologic features with emphasis on immunocytochemical markers. The nongingival granular cell tumors (NGGCT) were found to be more prevalent among females than males by a ratio of 2: 1 and arose on the tongue (67%), the buccal mucosa (13%) the lips (8%) the soft palate @P/o), and other sites (6%). With the use of the avidin-biotin-peroxidase method, polyclonal rabbit antisera were employed. The antisera were directed to the following antigens: S-100 protein, myoglobin. myosin, actin, desmin, alpha-1-antitrypsin, and muramidase. Results indicated that granular cell tumors are not homogenous for immunocytochemical markers. Nongingival granular cell tumors were universally positive for S- 100 protein and failed to exhibit immunoreactivity for myogenous or histiocytic markers. Alternatively, the gingival granular cell tumor of infancy was negative for all markers, whereas rhabdomyoma was reactive with myogenous markers and a subpopulation of tumor cells displayed S-100 protein immunoreactivity. The granular cell ameloblastoma was reactive only with antiserum to alpha- 1-antitrypsin. Ultrastructurally, granular cells from one of two NGGCT showed a direct evolution from skeletal muscle fibers. It is concluded that the oral NGGCT is a tumor positive for S- 100 protein that may arise from muscle or nerve sheath. (ORAL SURC ORAL MED ORAL PATHOL 1988;65:427-35)

S

ince the original description of the granular cell myoblastoma by Abrikossoff in 1926,’ the histogenesis of the granular cell tumor (GCT) has remained enigmatic. Many cell types have been implicated in the histogenesis, including muscle cells (myoblasts), 112histiocytes, 3,4fibroblasts s,6 neural sheath cells,7-gand undifferentiated mesenchymal cells.10-13 Immunocytochemical studies have been conducted

“Assistant Professor, Division of Oral Medicine, Department of Oral Diagnostic Sciences, University of Florida College of Dentistry. bAssistant Professor, Division of Oral Medicine, Department of Oral Diagnostic Sciences, University of Florida College of Dentistry. ‘Professor and Chairman, Department of Oral Diagnostic Sciences, University of Florida College of Dentistry. dProfessor, Department of Oral Biology, University of Florida College of Dentistry. ‘Professor and Head, Division of Pathology, Department of Diagnostic Sciences, University of the Pacific School of Dentistry.

to elucidate the histogenesis of GCTs. With the use of putative myogenous markers, the theory of a myogenic origin has not been supported. Mukai14 reported on a series of 18 GCTs with which antistriated muscle protein antisera was used. The granular cells in all tumors reacted negatively to antimyoglobin and antimyosin antiserum. Thompson’5 reported that none of seven GCTs of the tongue appeared to contain myoglobin. Ingram and coworkerP reported that myoglobin was not localized in GCT of the breast. Matthews and Mason” reported on 15 casesof oral GCT. The granular cells showed a weak reaction for actin in contrast to the strong cytoplasmic staining found in smooth and striated muscle. Slootweg and coworkers’* reported desmin to be absent in six oral granular cell tumors; three were granular cell myoblastomas (GCM) of the tongue, two were congenital gingival granular cell tumors (CGGT), and one was a granular cell ameloblastoma. Some investigators have indicated that S-100 427

428

Stewart

et al.

Fig. 1. Bar graph depicting age distribution of nongingival granular cell tumors according to gender.

Fig. 2. Pie chart illustrating gival granular cell Cumors.

protein is a reliable immunocytochemical marker for the Schwann cell component of peripheral nerve neoplasms.19-2’Support for a Schwann cell origin came from the demonstration of S- 100 protein in the granular cells by Slootweg and coworkers,‘* Stefansson and Wollmann,22Nakazato and associates,23and Armin and associates,24Bedetti and co11eagues,25 Mukai,14 and Ingram and colleagues.‘6 It is now recognized that S-100 protein is not restricted to Schwann cellsz6-29 and has been associated with other types of neoplasms.20~ 30,3’ Pennys and co11eagues32, 33suggestedthat a monoclonal antibody to myelin basic protein (MBP) was a specific marker for neoplastic Schwann cells. Mukai14 reported positive staining in granular cells of 18 granular cell lesions for S-100 protein, P2 protein, and PO protein. Clark and associates34 included granular cell tumors in a study of Schwann cell neoplasms.He reported that the granular cells of the granular cell tumors all showed positive staining for S-100 protein, but no neoplastic cells in the Schwann cell neoplasms showed staining for either myelin basic protein or P2 protein. Glial fibrillary acid protein (GFAP), an intermediate-sized filament (IF) protein, had been reported to be a specific marker for astrocytes and tumors derived from them.35 It had been reported that schwannomas appeared not to contain the protein.)” Slootweg and associates’* observed no reactivity to GFAP in the oral granular cell tumors. Vimentin, an IF protein, had been demonstrated in schwannomas.37Slootweg and associates’sreported vimentin reactivity in CGGT, reactivity to a lesser extent in GCM, and no reactivity in granular cell ameloblastoma.

Leu-7 exhibits binding to neural structures3” and to preparations of myelin-associated glycoprotein.39 Smolle and colleagues40reported positive reactions (5% to 30%) with the use of Leu-7 (HNK-1) monoclonal antibody in 10 of 13 granular cell tumors. One of the negative reactions was from a GCT of the tongue and two were from tumors of the trunk. On the basis of these studies, which used neural sheath immunocytochemical markers, a definitive neurogenic origin for all GCTs could not be substantiated. Matthews and Mason’? stained 15 oral GCMs for histiocytic products. None of the tumors were positive for lysozyme, alpha- 1-antitrypsin, or cathepsin B. Twelve lesions gave a granular reaction for cathepsin D, which would support the view that the granules represent secondary lysozymes. Slootweg and associates’8reported that muramidase was negative in all six oral granular cell lesions; however, peanut lectin binding was positive in all cases. Because all histiocytes do not contain muramidase, these reports do not confirm or negate a histiocytic origin. Carcinoembryonic antigen (CEA) staining has been reported in a series of these tumors.16,‘7.4’A study of gingival granular cell tumors of the newborn did not duplicate these findings.42 Ingram and colleaguesi reported CEA localization in tumor cells in all cases. Matthews and Mason17reported positive CEA staining with the use of rabbit antiserum. We have elected to evaluate a large series of oral GCTs to elucidate clinical and histopathologic features with emphasis on immunocytochemical markers aimed at providing additional data in regard to tissue site of origin. The results were compared with

site distribution

for nongin-

Oral granular cell tumors

Volume 65 Number 4

429

Fig. 3. Nongingival granular cell tumors. A, Granular cells invested by skeletal muscle fibers. (Hematoxylin and eosin stain. Magnification, x480.) B, Granular cells associated with neural fibers. (Hematoxylin and eosin stain. Magnification, X200.)

respect to the following GCTs: gingival granular cell tumor, rhabdomyoma, granular cell ameloblastoma, and granular cell ameloblastic fibroma. METHODS

Forty-eight instances of oral nongingival granular cell tumor (NGGCT) were retrieved from the oral pathology laboratory files at the University of Florida and the University of the Pacific. Clinical data with respect to age, gender, and site distribution were obtained. Twenty cases were randomly selected for detailed histopathologic and immunocytochemical studies. For comparative purposes, four additional lesions exhibiting populations of granular cells were included: (1) gingival granular cell tumor (courtesy of Dr. R. Corio); (2) oral rhabdomyoma (courtesy of Dr. R. Corio); (3) granular cell ameloblastoma (courtesy of Dr. C. Tomich); and (4) granular cell ameloblastic fibroma. Formalin-fixed, paraffin-embedded tissues were employed for both routine light microscopic evaluation and immunocytochemical marker studies. Eight sections of 6 pm were obtained from each block and adhered to glass slides cleaned with acid and coated with 0.1% polylysine. Sections were subsequently baked for 30 minutes at 56” C, dewaxed in xylene, and hydrated through graded ethanols. Endogenous peroxidase was quenched by application of 3% hydrogen peroxide for 5 minutes; this was followed by a rinse for 2 minutes in phosphate-buffered saline (PBS), pH 7.5. Nonspecific protein binding was

Table I. Histologic characteristics of 20 nongingival granular cell tumors Histologic

characteristics

PronouncedPEH* Mild/moderate PEH Absence of PEH Muscular investment of granular cell Neural investment of granular cell Both muscular and neural investment *pseudoepitheliomatous

% 25 20 55 70 5 3

hyperplasia

attenuated by incubation for 20 minutes with 0.1% bovine albumin/O.l% bovine gamma globulins. Sections were blotted without rinsing and various primary antisera were applied to each section for 30 minutes at room temperature with three subsequent rinses in PBS for 3 minutes. After incubation, the avidin-biotin-peroxidase polyvalent detection system (Immunon, Detroit, Michigan) was used. Secondary biotinylated antibody was reacted for 30 minutes, and three PBS washeswere undertaken. The sections were then incubated with peroxidase-reactive 3amino-9-ethylcarbazole chromogen in the presence of 0.1% hydrogen peroxide substrate.43s44 The tissue sections were finally rinsed in distilled water, counterstained lightly with hematoxylin, and mounted in a glycerin-water medium. The primary antisera used in the aforementioned avidin-biotin-peroxidase method were all polyclonal

430

Stewart et al.

Oral Surg April 1988

Table

II. Immunocytochemical markers in granular cell tumors Tumor Marker Myogenous Myosin Actin Myoglobin Desmin Neurogenous s-100 Histiocytic Muramidase alpha- 1-antitrypsin

GCT

GGCT

----r

Rhabdomyoma

GCA

-

+ ++ ++ +

-

* * * *

+++

-

t

-

*

-

-

-

+

* *

-

j ,

GC.46 -..-_.--

*Granular cell ameloblastic fibroma exhibited excessive endogenousperoxidase and alkaline phosphatase,yielding noninterpretable results.

rabbit antibodies directed to (1) S-100 protein (Dako, Santa Barbara, California), (2) myoglobin (Dako), (3) myosin (Immunon), (4) actin (Immunon), (5) desmin (Dako), (6) alpha- 1-antitrypsin (Immunon), and (7) muramidase (Immunon). All primary antisera were applied undiluted. Negative control sections were incubated in PBS, antiserum free buffer. Positive controls for each of the aforementioned listed antisera consisted of tissues containing the following cellular elements: (1) S-100 protein-nerve fibers, intraepithelial Langerhans’ cells; (2) myoglobin, myosin, actin, and desminvascular smooth muscle, skeletal muscle; and (3) alpha- 1-antitrypsin and muramidase-inflammatory lesions with histiocytic infiltrates. For ultrastructural study, two tissue blocks were deparaffinized in xylene and then hydrated through ethanols into water. They were then refixed in 3% glutaraldehyde; this was followed by postfixation in 2% osmium tetroxide in 0.1 M sodium cacodylate buffer. After dehydration through graded ethanols, tissue blocks were embedded in EPOX 8 12 (Ernest Fullam, Inc., Latham, New York). Sections were stained with uranyl acetate and lead citrate and studied with a Zeiss (Oberkochen, West Germany) 10A transmission electron microscope. RESULTS Clinical findings

Clinical findings were based on 48 oral GCTs, 19 accessedfrom the files of the University of Florida and 29 from the University of the Pacific School of Dentistry from 1977 to 1986, inclusive. Oral GCTs had a predilection toward females of greater than 2:l with 34 female patients and 14 male patients involved. Most (85%) of the tumors were nearly evenly distributed among the third, fourth, and fifth decades (Fig. 1). The age ranged from 7 to

70 years with a mean age of 34 years. The average age at time of occurrence was 30 years for female patients and 38 years for male patients. Thirty-four (7 1%) of the tumors occurred in white persons (twenty six female and eight male) and eleven (23%) occurred in black persons (six female and five male). One female patient was of Hispanic origin, and in two instances race was not specified. The site distribution of the 48 granular cell tumors rounded to the nearest whole percent is reported in Fig. 2. The tongue was the predominant site of involvement (67%); in 32 of the 48 cases. Of those tongue lesions, 48% occurred on the dorsum, 15% occurred on the lateral border, and 4% occurred on the ventral surface. Of the remaining tumors, 13% involved the buccal mucosa, 4% the upper lip, 4% the lower lip, 6% the palate, 2% the floor of the mouth, 2% the gingiva, and 2% the retromolar area. Histologic

findings

As seen with light microscopy, all tumors displayed histopathologic features consistent with the diagnosis of granular cell tumor. The histologic characteristics of the 20 cases, examined in detail, are listed in Table I. In 45% of the tumors, the overlying epithelium displayed varying degrees of pseudoepitheliomatous hyperplasia (PEH). One fourth of the cass displayed pronounced PEH, and one fifth displayed mild-to-moderate PEH. A close association between tumor cells and striated muscle was evident. A transition was noted from normal muscle fibers to degenerating muscle fibers, progressing to the development of granular cells enveloped by a common sarcolemmal membrane. The muscle membrane investment of tumor cells was apparent in 70% of the cases (Fig. 3, A). A similar continuum from normal fascicles of nerve fibers to their gradual replacement by tumor cells was

Oral granular cell tumors 431

Volume65 Number 4

Fig. 4. Immunocytochemical markers for nongingival granular cell tumor (NGGCT) and rhabdomyoma. A, NGGCT, tumor cells exhibit positivity for S-100 protein. (Magnification, X40.) B, NGGCT, tumor cells are negative for and adjacent muscle fibers are positive for actin. (Magnification, x40.) C, Rhabdomyoma; actin positivity. (Magnification, x350.) D, Rhabdomyoma; some cells display S-100 protein reactivity. (Avidin-Biotin-Peroxidase. Magnification, x400.)

observed in 5% of the cases (Fig. 3, B). Both muscular and neural investment of tumor cells was observed in 3% of the cases. lmmunocytochemical

findings

The results obtained with the cytochemical markers employed in this study appear in Table II. The common oral GCT was found to be negative for all markers except S- 100 protein; all 20 specimens assayedwere strongly positive (Fig. 4, A, and B). In comparison, rhabdomyoma showed relatively uniform immunoreactivity for the myogenous markers

actin and myoglobin with a subpopulation of tumor cells exhibiting positivity for S-100 protein (Fig. 4, C and D). The gingival granular cell tumor was found to be universally negative for all markers used in this study. Most granular cells occupying the stellate reticulum zone in granular cell ameloblastoma were immunoreactive for alpha- 1-antitrypsin but were negative for all other markers. Granular cell ameloblastic fibroma was universally positive for all markers; however, the negative control specimen was found to be positive, as well indicating a high, nonquenchable, level of endogenous peroxidase. For

432

Stewart

et al.

Oral Surg April 1988

Fig. 5. Ultrastructure of nongingival granular cell tumors. A, Muscle-associated tumor showing granular cells adjacent to skeletal muscle fiber (upper righfl. Tumor cells show smooth plasma membranes with cytoplasmic vacuoles, densebodies, and autophagic granules. (Magnification, x3300.) B, Evolving granular cell with dense bodies arising from muscle fiber terminus. (Magnification, x7,700). C, Granular cell with angulate bodies in cytoplasmic process(arrowheads]. (Magnification, x2.800.) D, Neural associatedtumor showing granular cells (right) and cells containing cytoplasmic filaments ileft). (Magnification, x2,900.)

this reason, all reactions were repeated with the use of a streptavidin alkaline phosphataseconjugate with similar results (i.e., all markers as well as negative control stained positively), thereby mitigating any meaningful aquisition of data in this tumor. Furthermore, it should be noted that the specimen had been decalcified subsequent to fixation. With the exception of granular cell ameloblastic fibroma, all tumors exhibited only mild background staining comparable to that seenin adjacent collagen when primary antibody was replaced by buffer (i.e., negative control specimens). Positive controls were consistently reactive, as expected. Peripheral nerve fibers were positive for S-100 protein, although

skeletal muscle fibers were also slightly reactive in most instances. Actin, myosin, and myoglobin were positive in both vascular smooth muscle and skeletal muscle, whereas desmin labeled skeletal muscle and yet failed to stain smooth muscle. Alpha-l-antitrypsin and muramidase labeled inflammatory histiocytes positively. Ultrastructural

findings

Two nongingival granular cell tumors were examined ultrastructurally; one tumor exhibited light microscopic features indicative of muscle origin, whereas the secondtumor showed evidence of neural derivation. Both specimenswere artifactually altered

Volume 65 Number 4

as a result of prior formalin fixation; however, the morphologic features were found to be reasonably preserved. The lesion intimately associated with muscle was represented by aggregates of irregularly shaped polygonal cells with centrally placed oval or spherical nuclei. These granular cells had a plasma membrane encompassing a vacuolated cytosol containing numerous oval dense bodies and autophagic granules. Externally, compacted collagen fibers were observed in the stroma. Whereas most tumor cells were separated from intact muscle fibers, each with their own plasma membranes (Fig. 5, A), numerous sections disclosed the presence of evolutionary or transitional cells. Evolving granular cells were seen to develop at muscle fiber terminals without an intervening plasma membrane (Fig. 5, B). In yet other regions, granular cells with cytoplasmic dense bodies contained angulate structures resembling myofilaments, although no cross banding sarcomeres could be discerned among these remnants (Fig. 5, C). Pinocytic vesicles were not observed, nor was there any evidence of active phagocytosis. The lesion that was associated with nerve exhibited similar features, although more individual cells contained cytoplasmic filaments and myelin figures (Fig. 5, D). DISCUSSION

Since the original description of the nongingival granular cell tumor by Abrikossoff,’ the controversy over histogenesishas received considerable attention. Light microscopic features have suggested origin in skeletal muscle fibers for many of these tumors, whereasother appear to be intimately associatedwith peripheral nerve.‘.2*7-9Extensive studies employing cytochemical markers have led most authors to conclude that the nongingival granular cell tumor is derived from nerve sheath cells rather than from muscle by virtue of S-100 protein immunoreactivity when contrasted with myogenic protein negativity.‘4-25 Alternatively, others have suggested a histiocytic origin on the basis of both cytochemical and ultrastructural studies, particularly in regard to the pronounced granularity being attributed to the presence of lysosomal bodies.3,4Ultrastructural studies have also failed to support a muscular origin for this tumor on the basisof a lack of myofilaments in granular cells and the failure of researchersto identify morphologic evidenceof a transition from skeletal muscle to tumor cells with a common plasma membrance (i.e., sarcolemma1 investment of granular cell). Despite the cytochemical evidence for neurogenous origin, many nongingival granular cell tumors at

Oral granular cell tumors

433

the light microscopic level appear to be intimately associated with skeletal muscle fibers and indeed, appear to be in continuity with sarcolemmal membranes from contiguous skeletal muscle fibers cut in longitudinal section. In this study, we found muscle investment of tumor cells to be more commonly encountered (70%) than neural investment (5%). In accordance with other investigations, the immunocytochemical marker studies employed here disclosed consistent positivity for S-100 protein with a lack of immunoreactivity with the use of myogenous markers (myosin, actin, myoglobin, desmin). Importantly, normal skeletal muscle was slightly reactive with S- 100 antisera, and in comparative studies, with a benign neoplasm of undisputed muscular origin (rhabdomyoma) nearly half the tumor cell population was found to be positive for S-100 protein. Rhabdomyoma also exhibited immunoreactivity with actin and myoglobin. Alternatively, the cytochemical markers employed in this study were universally negative when applied to sections of gingival granular cell tumor of the newborn. It should be stressed that the demonstration of cellular proteins by immunocytochemical methods is a qualitative measure of gene expression and cytodifferentiation rather than a de facto indiator of cellular origin. During a pathologic process, cells may not necessarily emulate their progenitors. Thus, the presence of S-100 protein should not be considered a marker indicative of neurogenous origin; indeed, this protein has been demonstrated in numerous cellular phenotypes unrelated to nerve or nerve sheath cells.3’ The identification of S- 100 protein in normal skeletal muscle and in the cytoplasm of rhabdomyoma tumor cells further mitigates the reliability of this marker as an indicator of histogenesis. Sometimes routine light microscopy can be as revealing as more sophisticated analyses. The ultrastructural findings reported here also support a myogenic origin for some of these tumors, and it is not unreasonable to assume that the granular cells may evolve from degenerating muscle cells that, in turn may lose cellular proteins and organelles that ordinarily would serve as footprints of histogenesis. Alternatively, on light microscopic grounds, some nongingival granular cell tumors in this group are intimately associated with nerve and nerve sheath cells. It is our contention that the nongingival granular cell tumor is a reactive proliferation composed of lysosome-rich cells with positivity for S-100 protein that may evolve from degenerating muscle or Schwann cells and that the resultant phenotype, regardless of cellular origin, is homogenous histologically, ultrastructurally, and cytochemically. Fur-

434

Stewart et al.

thermore, it may be speculated that granular cell tumors arising from cutaneous sites may arise more often from nerve sheath than their oral counterparts in the tongue. REFERENCES 1. Abrikossoff AI. Uuber myome ausgehend von der quergestreiften willkurlichen muskulatur. Virchows Arch [Pathol Anat] 1926;260:215-33. 2. Murray MR. Cultural characteristics of three granular cell myoblastomas. Cancer 1951;5:857-67. 3. Eversole LR, Sabes WR. Granular sheath cell lesions: report of cases.J Oral Surg 1971;29:867-71. 4. Lucas RB. Pathology of tumours of the oral tissues. 3rd ed. London: Churchill Livingstone, 1976:234-6. 5. Ashlev DJB. Evans’ histoloaical aooearancesof tumours. 3rd ed. London: Churchill Livingston, i978:50-2. 6 Pearse AGE. The histogenesisof granular cell myoblastoma (granular cell perineural fibroblastoma). J Path01 Bacterial 1950;62:351-62. 7 Fisher ER, Wechsler H. Granular cell myoblastoma-a misnomer: electron microscopic and histochemical evidence concerning its Schwann cell derivation and nature (granular celi schwannoma). Cancer 1962;15:936-54. 8. Garancis JC, Komorowski RA, Kuzma JF. Granular cell myoblastoma. Cancer 1970;25:542-50. 9. Weiser G. Granularzelltumor (Granuliires Neurom Feyrter) und SchwannschePhagen. Elektronenoptische Untersuchung von 3 Flllen. Virchows Arch [Pathol Anat] 1978;380:4951. IO. Moscovic EA, Azar HA. Multiple granular cell tumors (myoblastomas). Cancer 1967;20:2032-47. Il. Aparicio SR, Lumsden CE. Light and electron microscope studies on the granular cell myoblastoma of the tongue. J Path01 1969;97:339-55. 12. Sobel HJ, Schwarz R, Marquet E. Light and electron microscope study of the origin of granular cell myoblastoma. J Path01 1973a;109:101-11. 13. Regezi JA, Batsakis JG, Courtney RM. Granular cell tumors of the head and neck. J Oral Surg 1979;37:402-6. 14. Mukai M. Immunohistochemical localization of S-100 protein and peripheral nerve myelin proteins (P2 protein, PO protein) in granular cell tumors. Am J Path01 1983;112:13946. 15. Thompson SH. Myoglobin content of granular cell tumor of the tongue. Oral Surg Oral Med Oral Path01 1984;57:74-6. 16. Ingram DL, Mossier JA, Snowhite J, Leight GS, McCarty KS. Granular cell tumors of the breast. Steroid receptor analysis and localization of carcinoembryonic antigen, myoglobin and S-100 protein. Arch Pathol Lab Med 1984; 108:897-901. 17. Matthews JB, Mason GI. Oral granular cell myoblastoma: an immunohistochemical study. J Oral Pathol 1982;11:343-52. 18. Slootweg P, deWilde P, Vooijs P, Ramaekers F. Oral granular cell lesions.An immunohistochemical study with emphasis on intermediate-sized filaments proteins. Virchows Arch [Pathol Anat] 1983;402:35-45. 19. Clark HB, Hartman BK. S-100 protein as an immunohistochemical marker for neoplasms of glial and Schwann cell origin (abstract). J Neuropathol Exp Neurol 1981;40:335. 20. StefanssonK, Wollmann R, Jerkovic M. S-100 protein in soft tissue tumors derived from Schwann cells and melanocytes. Am J Path01 1982;106:261-8. 21. Santa Cruz DJ, Clark HB, Hartman BK, Moore BW. S-100 protein, an immunohistochemical marker for neural tumors and tumorlike conditions of the skin (abstract). Arch Dermatol 1982;118:951. 22. Stefansson K, Wollmann RL. S-100 protein in granular cell tumors (granular cell myoblastomas). Cancer 198249: 1834-8.

Oral Surg April 1988 23. Nakazato Y, Ishizeki J, Takahashi K, Yamaguchi N. Immunohistochemicai localization of S-100 protein in granular cell myoblastoma. Cancer 1982;49:1624-8. 24. Armin A, Connelly EM, Rowden G. An immunoperoxidase investigation of S-100 protein in granular cell myoblastomas: evidence for Schwann cell derivation. Am J Chn Path01 1983: 79137-44. 2.5 Bedetti CD, Martinez AJ, Beckford NS, May M. Granular cell tumor arising in myelinated peripheral nerves. Light and electron microscopy and immunoperoxidase study. Virchows Arch [Pathol Anat] 1983;402:175-83. 26 Nakajima T, Watanabe S, Sato Y, Kameya ‘I, Hirota T, Shimosato Y. An immunoperoxidase study of S-100 protein distribution in normal and neoplastic tissues. Am J Surg Path01 1982;6:715-27. 2-l Stefansson K, Wollmann RL, Moore BW. Distribution of S-100 protein outside the central nervous system. Brain Res 1982;234:309-17. 28. Stafansson K, Wollmann RL, Moore BW, Arnason BGW. S-100 protein in human chondrocytes. Nature 1982;295: 63-4. 29. Dhillon AP, Rode J. Immunohistochemical studies of S-100 protein and other neural characteristics expressedby granular cell tumor. Diagn Histopathol 1983;6:23-8. 30. Clark HB, Santa Cruz DJ, Hartman BK, Moore BW. S-100 protein, an immunohistochemical marker for malignant melanoma and other melanocytic lesions (abstract). Lab Invest 1982;46:13A. 31. Vanstapel M-J, Gratter KC, de Wolf-Peeters C, Mason DY, Desmet VD. New sites of human S-100 immunoreactivity detected with monoclonal antibodies. Am J Clin Path01 1986; 85:160-8. 32. Pennys NS, Adachi K, Ziegels-Weissman J, Nadji M. Granular cell tumors of the skin contain myelin basic protein. Arch Path01 Lab Med 1983;107:302-3. 33. Pennys NS, Mogollon R, Kowalczyk A, Nadji M, Adachi K. A survey of cutaneous neural lesions for the presence of myelin basic protein: an immunohistochemical study. Arch Dermatol 1984;120:210-3. 34. Clark HB, Minesky JJ, Agrawal D, Agrawal HC. Myelin basic protein and P2 protein are not immunohistochemical markers for Schwann cell neoplasms. A comparative study using antisera to S-100 protein, P2, and myelin basic proteins. Am J Pathol 1985;121:96-101. 35. TascosNA, Parr J, Gonatas NK. Immunocytcchemical study of the glial fibrillary acidic protein in human neoplasmsof the central nervous system. Hum Path01 1982;13:454-8. 36. Ramaekers FCS, Puts JJG, Moesker 0, et al. Antibodies to intermediate filament proteins in the immunohistochemical identification of human tumours: an overview. Histochem J 1983;15:691-713. 31. Altmannsberger M, Osborn M, Trenner J, Holscher A, Weber K, Schauer A. Diagnosis of human childhood rhabdomyosarcoma by antibodies to desmin, the structural protein of muscle specific intermediate fibers. Virchows Arch [Cell Pathol] 1982;39:203-15. 38. Schuller-Petrovic S, Gebhart W, Lassmann H, Rumpold H, Kraft D. A shared antigenic determinant between natural killer cells and nervous tissue. Nature 1983;306:179-81. 39. McGarry RC, Helfand SL, Quarles RH, Roder JC. Recognition of myelin-associated glycoprotein by the monoclonal antibody HNK-1. Nature 1983;306:376-8. 40. Smolle J, Konrad K, Kerl H. Granular cell tumors contain myelin-associated glycoprotein. An immunohistochemical study using Leu 7 monoclonal antibody. Virchows Arch [Pathol Anat] 1985;406:1-5. 41. Shousha S, Lyssiotis T. Granular cell myoblastoma: positive staining for carcinoembryonic antigen. J Clin Path01 1979; 32:219-24. 42. Lack EE, Worsham GF, Callihan MD, Crawford BE, Vawter CF. Gingival granular cell tumors of the newborn (congenital “epulis”). Am J Surg Path01 1981;5:37-46.

Oral granular cell tumors

Volume 65 Number 4 43. Hsu SM, Raine L, Ranger H. A comparative study of the PAP method with avidin-biotin-complex mehtod for studying polypeptide hormones with radioimmunoassay antibodies. Am J Clin Path01 1981;75:734-8. 44. Hsu SM, Raine L, Ranger H. The use of avidin-biotinperoxidas techniques: a comparison between ABC and unlabeled antibody (PAP) procedures. J Histochem Cytochem 1981;29:577-80.

Reprint requests to:

Dr. Carol M. Stewart Division of Oral Medicine Department of Oral Diagnostic Sciences Box J-414 JHMHC University of Florida College of Dentistry Gainesville, FL 32610

435