Diagnostic Immunohistochemistry of Canine Round Cell Tumors

Vet. Pathol. 24495-499 (1987) Diagnostic Immunohistochemistry of Canine Round Cell Tumors G. E. SANDUSKY, W. W. CAIUTON, AND K. A. WIGHTMAN Lilly Research Laboratories, A Division of Eli Lilly and Company, Greenfield, IN, and School of Veterinary Science and Medicine, Purdue University, West Lafayette, IN Abstract. Sixty-five canine skin neoplasms studied using immunocytochemistry, included 22 histiocytomas, 18 amelanotic melanomas, 14 cutaneous lymphosarcomas, six mast cell tumors, and five transmissible venereal tumors. Formalin-fixed, paraffin-embedded sections were stained using the avidin-biotin-peroxidase complex (ABC) immunoperoxidase technique for reactivity with S- 100 protein, kappa and lambda immunoglobulin light chains, alpha- 1 -antitrypsin, alpha- 1 -antichymotrypsin, leukocyte common antigen (LCA), neuron-specific enolase, keratin, cytokeratin, muramidase, and vimentin. Detection of S- 100, kappa and lambda light chains, neuron-specific enolase, and vimentin were most useful for screening these neoplasms. None of the markers examined was consistent in staining histiocytomas. While reactivity of S- 100 (ten cases) and neuron-specific enolase (ten cases) was detected in some amelanotic melanomas, lambda light chain immunoglobulin (eight cases) was relatively consistent in cutaneous lymphomas. Mast cell neoplasms reacted with avidin and, therefore, were positive, even on negative control sections. Vimentin reacted strongly on all amelanotic melanomas and transmissible venereal tumors examined. These antibodies are helpful adjuncts in the differential diagnosis of canine skin tumors. Undifferentiated tumors of the skin and underlying subcutaneous tissue present a diagnostic challenge to the surgical pathologist. Both spindle cell tumors and small round cell tumors often pose problems of differential diagnosis when examined by light microscopy. However, immunohistochemistry has provided a technique with which polyclonal and/or monoclonal antibodies can be used to gain more diagnostic information about a variety of poorly differentiated neopla~ms.~~~~~,*j~j* During the past few years, a number of tumor markers for soft tissue neoplasms have been d e ~ ~ r i b e d. ~ Due ~ ~ ~ to ~ cross ~ ~ reactivity ~ ~ ~, ~ of ~ absolute, ~ ~ ~ methanol. ~, ~ ~ The, slides ~ ~ were covered with normal goat some antibodies and lack of staining of some tumors, a panel of selected commercially available antibodies was used, rather than reliance on one antibody, to identify common, undifferentiated skin tumors. Materials and Methods The neoplasms studied were obtained from the pathology files of Purdue University s oncology program. Tissues were fixed in 10% neutral buffered formalin, embedded in paraffin, sectioned at 4 pm, and stained with hematoxylin and eosin (HE). Two additional sections from each case were placed on one slide. One section was used as a negative control and the other was stained for immunoreactivity to the various tumor markers. Each tissue specimen was circled with a diamond pencil to prevent cross-contamination of antisera. Unstained sections of the tumors were deparaffinized and stained using the avidin-biotin complex method described by Hsu.I0 Table 1 lists the antibodies used in this study, their specificity for cells present in the skin and subcutaneous tissue, and the specific dilutions. The antigen against cytokeratin and vimentin was from Biogenex Corporation, Dublin, California. The rest of the primary antisera were from Dako Corporation, Santa Barbara, California. The biotinylated secondary antibody and ABC complex were from Vector Corporation, Burlingame, California. Polyclonal Antibody Technique Endogenous peroxidase activity was blocked by immersing the slides for 10 minutes in a 6% solution of H,O, made in serum and diluted 1 : 20 in modified phosphate buffered saline (PBS) for 60 minutes before incubation in the primary antibody. The incubation period for the various primary antisera was 1 hour, except for the cytokeratin, which was incubated overnight at 4 C, and vimentin, which was incubated for 2 hours. The biotinylated goat anti-rabbit secondary antisera and avidin-biotin-peroxidase complex were incubated for 30 minutes each at room temperature. Slides were washed for 10 minutes between each step in a modified PBS buffer. The chromagen, 3-amino-9-ethylcarbazole, was applied to the slides for 20 minutes at 37 C. The slides were then counterstained with Mayer s hematoxylin. Monoclonal Antibody Technique The technique used was identical to the above procedure except for using normal horse serum and biotinylated horse antimouse secondary antisera instead of rabbit reagents. Negative controls included substituting the primary anti- 495

496 Sandusky, Carltc m, and Wightman Table 1. Antibodies and cell types evaluated in canine cutaneous neoplasms. Table 3. Results of immunocytochemical determinations on round cell tumors in the dog. Antibody Antibody Specificity Cellular Specificity Dilution S- 100 protein Melanocytes 1 : 100 Muramidase Histiocytes 1 : 100 Alpha- 1 -antitrypsin Histiocytes 1 : 100 Alpha- 1 -antichymo- Histiocytes 1 : 100 trypsin Kappa light chain B-lymphocytes 1 : 1,000 plasma cells Lambda light chain B-lymphocytes 1 : 1,000 plasma cells Keratin Epithelial cells 1 : 100 Cytokeratin (M)* Epithelial cells Prediluted Leukocyte common an- Leukocytes 1:20 tigen (M) Neuron-specific enolase Neuroendocrine 1 : 500 cells Vimentin (M) Mesenchymal cells 1 : 500 * M-monoclonal antibody. sera with nonimmune sera from the same species and, in addition, normal saline. Known positive controls were included for each marker. A normal tissue screen of 20 tissues was examined with each antibody to determine the specificity of the antibodies. All antibodies stained cellular structures in the dog, which were analogous to those stained in human tissues. Specimens in which 1% or less of the cells were stained were considered negative. Predigestion with 0.1% trypsin in 0.1% CaC1, solution was performed for 30 minutes on all tumors that failed to react with any primary antibodies tested. Results By the proper selection of four or five antibodies listed in Table 1, the basic cell type involved in the skin neoplasms examined was identified. False negative staining reactions seen require the use of more than one antibody for a positive diagnosis. In addition, the use of four or five tumor markers as a panel did exclude other tumors considered in differential diagnosis. Upon examination of the neoplasms by light microscopy, most were composed of spindle, round, or polygonal cells. All histiocytomas were easily distin- Table 2. Immunocytochemical results of staining of histiocytomas using macrophage markers. Antibody Specificity Histiocytomas ~ Ly soz y me 4/22* Alpha- 1 -antitrypsin 7/22 Alpha- 1 -antichymotrypsin 0/22 * Number positivehumber examined. Transmissi- Antibody Histiocytoma Lym- ble phoma Melanoma Vene- real Tumor s-100 0/22* 0/14 10/18 0/5 Kappa 0/22 3/14 0/18 0/5 Lambda 0/22 8/14 0/18 0/5 Cytokeratin NDt 0/14 0/18 0/5 Vimentin ND ND 18/18 5/5 Neuron-specific enolase O/ 18 ND 1 O/ 18 0/5 * Number positivehumber examined. t ND = not done. guished, based on routine histiologic examination. In the rest of the tumors studied, differential diagnoses included carcinomas, lymphomas, malignant melanomas, and mesenchymal tumors of fibroblast, smooth muscle, skeletal muscle, or neural origin. In a few of these cases, it was difficult to make a definitive diagnosis based on routine light microscopic examination. Histiocytomas were studied for the presence of muramidase (lysozome), alpha- 1 -antitrypsin, alpha- 1 -antichymotrypsin, kappa and lambda light chains, and S-100 protein (Tables 2, 3). No markers used in this panel were consistently positive for histiocytomas. Antibodies to S-100 protein, kappa and lambda light chains, and alpha- 1 -antichymotrypsin failed to react with any of the histiocytomas examined. Immunoreactivity of alpha- 1 -antitrypsin was strong in seven of 22 histiocytomas (Fig. 1) and failed to react in the remaining 16 tumors. Muramidase reacted moderately with four tumors in which the staining was intense in individual tumor cells; scattered positive histiocytes were positive in three additional tumors (Fig. 2). Diagnosis of a B-cell lymphoma was established on the basis of staining of neoplastic cells for kappa or lambda light chains. No immunoreactivity was seen in tumors examined for cytokeratin or S- 100 protein. Eleven cases were diagnosed as B-cell origin lymphosarcomas on the basis of a positive reaction of some or all of the tumor cells for kappa light chains (three cases) or lambda light chains (eight cases) (Fig. 3). Most of these tumors were positive for one marker (mainly lambda) and negative for the other. The quantity of cells that stained varied from less than 50% to nearly 100%. Staining was very intense and confined to the cytoplasm. Leukocyte common antigen failed to react on normal frozen canine lymph node and was unsatisfactory for use in the dog. All lymphosarcomas were negative with leucocyte common antigen (LCA). Ten of the 18 cases of malignant melanoma showed

498 Sandusky, Carlton, and Wightman positive immunoreactivity for neuron specific enolase (NSE) (Fig. 4). Only three of 18 melanomas failed to react with either S-100 or NSE. All 18 tumors reacted positively with vimentin. Immunostaining of tumor cells with NSE was weaker and more diffuse and homogeneous than staining for S- 100. Ten of 18 tumors were immunoreactive for S- 100, and staining was strong and more multifocal in most cases (Fig. 5). There was some variation in the staining pattern of S-100, since a few tumors were diffusely positive while other tumors had scattered positive strands of tumor cells. None of these tumors stained for kappa and lambda light chains or cytokeratin. Cytokeratin was positive in epithelial tissue around the tumors. Five mast cell tumors examined were positive using the avidin-biotin-peroxidase complex technique with no primary antibody. Granules in the cytoplasm were distinct and prominent with immunostaining. Transmissible venereal tumors were identified by the monoclonal antibody to vimentin. All five tumors examined stained positively for vimentin. Staining was diffuse, intense, and confined to cytoplasm ofthe round tumor cells (Fig. 6). No immunoreactivity in these tumors was seen using antibodies to cytokeratin, S- 100 protein, kappa and lambda light chains, and NSE. However, cytokeratin was stained in the normal overlying epithelium. Discussion In this study, by using both light microscopy and immunoperoxidase, a definitive tumor diagnosis was made in all cases. Previous studies have indicated that the majority of unclassifiable tumors can be correctly diagnosed by staining tumor tissue sections with a panel of antibodies directed against different cellular comp~nents.~.~j~j~ Muramidase, alpha- 1 -antitrypsin, and alpha- 1 -antichymotrypsin have been accepted as markers for benign and malignant histiocytes in man. L4~L7~19 In this study of cutaneous histiocytomas in the dog, none of the markers examined appeared to be a reliable marker for histiocytomas. This was consistent with our previous report on histiocytomas in the dog,2o and similar to a study in histiocytic disorders in the dog. Negative reactions could be due to loss of antigenicity during fixation and processing. Additionally, lysozyme may decrease with the loss of differentiation of tumor cells and may account for no staining with muramidase or the other markers. Kappa and lambda light chains have been excellent markers used for the detection of B-cell lymphosarcomas.i4 This study reflects previous results in man, and the majority of tumors examined were positive for the lambda light chain. A previous study in the rhesus monkey and the dog in our laboratories found that antibodies to kappa and lambda light chains stained only B-cell areas in normal lymph nodes.21 Negative reactions seen could be due to loss of immunoglobulin during fixation or processing. It is possible that some of these tumors are of T-cell lineage.i4 Currently, most T-cell markers and antibodies to leukocyte-common antigens are monoclonal antibodies that react with cell surface antigens, which do not survive formalin fixation. The monoclonal antibody to cytokeratin used in this study stains the major intermediate-sized filaments in epithelial tissue and was not expected to react on the tumors examined in this study.9 Either $100 or neuron specific enolase (NSE) reacted immunohistochemically in all but three of the melanomas examined. These results generally agree with those of others.12j6,22 Negative staining was probably due to over-fixation and processing of the tumors in formalin. All melanomas did stain with vimentin, and no immunoreactivity to keratin was seen. These results have been routinely observed in man.22 The combined use of keratin, vimentin, S-100, and NSE was useful in the identification of amelanotic melanomas in the dog. Mast cell tumors reacted positively on the negative controls examined. This has been seen by other investigators and is attributed to nonspecific staining of mast cells by avidin biotinylated peroxidase complexes (ABC). Nonspecific staining of mast cells can be eliminated by increasing the ph of the ABC solution above ph 9.0, and does not affect the previous binding of primary antibodies or the affinity of avidin to biotin. Both vimentin and cytokeratin, which are intermediate filaments, were of considerable value in the diagnosis of canine round cell tumors. Vimentin reacted positively on all melanomas and transmissible venereal tumors examined, while cytokeratin, which was negative, eliminated the diagnosis of anaplastic carcinoma in cutaneous lymphomas and transmissible venereal tumors. In the past, the diagnostic usefulness of intermediate filaments has been limited because they have been reliable only on frozen or alcohol-fixed tis- Most formalin-fixed tissue did not react with intermediate filaments.6 During the past 9 months, a cytokeratin and vimentin that are readily demonstrable (reactive) after formalin fixation have been found. These results, with some of the newer monoclonal antibodies to intermediate filaments, seem to indicate that these antibodies are sufficiently reactive on formalin-fixed tissue to be applied diagnostically. Findings in this study indicate that immunocytochemistry applied to tumors of unknown cellular origin is of great potential diagnostic benefit in surgical veterinary pathology. Immunocytochemical techniques used in conjunction with light microscopy will enhance the accuracy of definitive diagnoses of neoplasms.

Canine Round Cell Tumor Histochemistry 499 1 8 9 10 11 References Altmannsberger M, Osborn M, Schauer A, Weber K Antibodies to different intermediate filament proteins. Cell type-specific markers on paraffin-embedded human tissues. Lab Invest 45427-434, 1981 Battifora H: Recent progress in the immunohistochemistry of solid tumors. Semin Diagn Pathol 1:251-271, 1984 Bussolati G, Gugliotta P: Nonspecific staining of mast cells by avidin-biotin-peroxidase complexes (ABC). J Histochem Cytochem 31:1419-1421, 1983 Denk H, Krepler R, Artlieb V, Gabbiani G, Rungger- Brandle E, Leoncini P, Franke WW: Proteins of intermediate filaments. An immunohistochemical and biochemical approach to the classification of soft tissue tumors. Am J Pathol 110:193-208, 1983 DuBoulay CEH: Immunohistochemistry of soft tissue tumors: a review. J Pathol 146:77-94, 1985 Erlandson R A: Diagnostic immunochemistry of human tumors. Am J Surg Pathol8:615-624, 1984 Gabbiani G, Kapanci Y, Barazzone P, Franke WW: Immunochemical identification of intermediate-sized filaments in human neoplastic cells. A diagnostic aid for the surgical pathologist. Am J Pathol104:206-216, 1981 Gatter KC, Alcock C, Heryet A, Mason DY: Clinical importance of analyzing malignant tumors of uncertain origin with immunohistological techniques. Lancet 1302-1305, 1985 Gown AM, Vogel AM: Anti-intermediate filament monoclonal antibodies: tissue-specific tools in tumor diagnosis. Surv Synth Pathol Res 3:369-385, 1984 Hsu S-M, Raine L, Fanger H: Use of avidin-biotinperoxidase complex (ABC) in immunoperoxidase techniques: a comparison between ABC and unlabeled antibody (PAP) procedures. J Histochem Cytochem 29: 577-580, 1981 Issacson PG, Jones D B: Immunohistochemical differentiation between histiocytic and lymphoid neoplasms. Histochem J 15621-635, 1983 12 Kahn HJ, Marks A, Thom H, Baumal R: Role of antibody to S-100 protein in diagnostic pathology. Am J Clin Pathol 79:341-347, 1982 13 Lewis RE, Johnson WW, Cruse JM: Pitfalls and caveats in the-methodology for immunoperoxidase staining in surgical pathologic diagnosis. Surv Synth Pathol Res 1:134-152, 1983 14 Mir R, Kahn LB, Selzer G: Immunohistochemistry of primary gastrointestinal lymphomas: a study of 76 cases. Histopathology 10:391-403, 1986 15 Moore PF Utilization of cytoplasmic lysozyme immunoreactivity as a histiocytic marker in canine histiocytic disorders. Vet Pathol23:757-762, 1986 16 Nakajima T, Kameya T, Watanabe S, Hirota T, Sato Y, Shimosato Y: An immunoperoxidase study of S-100 protein distribution in normal and neoplastic tissues. Am J Surg Pathol 6:715-727, 1982 17 Permanetter W, Meister P: Distribution of lysozyme (muramidase) and a, -antichymotrypsin in normal and neoplastic epithelial tissues: a survey. Acta Histochem 74~173-179, 1984 18 Roholl PJM, DeJong ASH, Ramaekers FCS: Application of markers in the diagnosis of soft tissue tumours. Histopathology 9: 1019-1035, 1985 19 Roholl PJM, Kleyne J, Elbers H, van der Vegt MCD, Albus-Lutter CH, Van Unnik JAM: Characterization of tumour cells in malignant fibrous histiocytomas and other soft tissue tumours in comparison with malignant histiocytes. I. Immunohistochemical study on paraffin sections. J Pathol 147:87-95, 1985 20 Sandusky GE, Carlton WW, Wightman KA: Diagnostic immunohistochemistry of canine skin tumors. Am Assoc Vet Lab Diagn, 29:83-92, 1986 21 Sandusky GE, Horton PJ, Wightman KA Use of monoclonal antibodies to human lymphocytes to identify lymphocyte subsets in lymph nodes of the rhesus monkey and the dog. J Med Primatol15:44 1-45 1,1986 22 Springall DR, Gu J, Cocchia D, Mitchetti F, Levene A, Levene MM, Marangos PJ, Bloom SR, Polak JM: The value of S-100 immunostaining as a diagnostic tool in human malignant melanomas. Virchows Arch (Pathol Anat) 400:33 1-343, 1983 Request reprints from Dr. G. E. Sandusky, Lilly Research Laboratories, PO Box 708, Greenfield, IN 46 140 (USA).