(D1990 The Japanese Society of Pathology
Ca rcinoembryonic Antigen and Nonspecific Cross-react ing A nt igen in Medulla ry Carcinoma of the Thyroid
Kennichi Ka kudo', Hiroshi Takami2, Shoichi Katayama3, Nariaki Matsuura*, Shingo Kamoshida5, Akira Miyauchi6, and Keiichi Watanabe'
Carcinoembryonic antigen (CEA) and nonspecific crossreacting antigen (NCA) were studied immunohistochemically in formalin-fixed, paraffin-embedded tissues of 73 cases of medullary carcinoma of the thyroid (MTC) using 2 polyclonal antibodies (CEA antisera cross-reactive with or without NCA), 3 monoclonal antibodies recognizing epitopes only on CEA, and one monoclonal antibody against NCA. The staining patterns of the 5 antibodies against CEA in MTCs were not different, and they reacted with 86.3% of all cases. With regard to the effects of fixatives on the staining patterns, samples fixed with formalin or 4% paraformaldehyde demonstrated CEA immunoreactivity in both the cell membrane and cytoplasm. In Bouin-fixed tissue, the imrnunoreactivity was predominant on the cell membrane, whereas cytoplasmic positivity predominated in alcohol-fixed specimens. Thus the difference in fixatives used in previous studies does not appear to be a major reason for the difference in the reported incidence of CEApositive MTCs. It is concluded that CEA is still a useful tumor marker for MTC and that it is detectable only i n thyroid tumors originating from C cells, as seen i n our series. The epitope defined by monoclonal antibody F10688, present only on NCA, was found in 42.5% of a l l cases (49.2% of CEA-positive MTCs). The NCA immunoreactivity was located in the tumor cell cytoplasm as globular aggregates, which were also labeled for CEA. Acta Pathol Jpn 40: 261-266, 1990. Key words : Medullary carcinoma,
Received August 21, 1989. Accepted for publication December 11, 1989. 'Department of Pathology, Tokai University School of Medicine, Isehara. 2Department of Surgery, Teikyo University School of Medicine, Tokyo. Departments of 3Pathology and 4Surgery, Osaka University School of Medicine, Osaka. SPathology Laboratory, Kyodo Hospital, Isehara. 6Department of Surgery, Kagawa Medical School, Takamatsu. Mailing address : Kennichi Kakudo (Bza-), Department of Pathology, Tokai University School of Medicine, Bouseidai, Isehara, Kanagawa 259-1 1, Japan.
Carcinoembryonic antigen, antigen
INTRODUCTION Useful tumor markers have been reported in medullary carcinoma of the thyroid (MTC) by many authors. These include calcitonin, calcitonin gene-related peptide (CGRP) and other peptides, and carcinoembryonic antigen (CEA) (1-19). The incidence of CEA-positive MTCs has been reported to vary slightly from 77% to 100% (9-19). This difference is attributed to variations in specificity of the antibodies used, since CEA is a large glycoprotein and contains many different antigen epitopes (9-19). Nonspeclfic cross-reacting antigen (NCA) is also reported t o be expressed in all types of thyroid carcinoma, and most commercially available CEA antisera or monoclonal CEA antibodies cross-react with NCA (18, 20). In the present study, CEA and its related su bst a nce, NCA, were studied immunohistoc hemica IIy in 73 MTCs to clarify the real incidence of CEA- and/or NCA-positive MTCs using specific antibodies against CEA and NCA. The effects of fixatives on the immunohistochemical results were also tested to clarify the reason for the differences in the reported incidence of CEA-positive MTCs.
MATERIALS AND METHODS Seventy-three histologically proven cases of MTC were collected from various institutes and hospitals. Samples from the thyroid tumors and/or lymph node metastases were fixed in formalin and embedded in paraffin. In addition to formalin fixation, two of them were divided to be fixed in different fixatives (70% ethanol, Bouin and 4% paraformaldehyde) and embedded in
CEA in Medullary Thyroid Carcinoma (Kakudo et a / . )
paraffin to compare the effects of fixation on the immunohistochemical results. The paraffin sections were s u b jected to the following indirect immunoperoxidase met hod. The primary antibodies used were polyclonal anti-CEA antibody (Dako, Copenhagen, Denmark), polyclonal antibodies (Dako and ICN, Costa Mesa, USA) preabsorbed with lyophilized powdered spleen to prevent reactivity with NCA, monoclonal antibodies against CEA (CMO10, defining an epitope present only on CEA, Mochida, Tokyo, Japan ; F33-104, defining an epitope present only on CEA, Kyowa, Tokyo, Japan; Parlam 4, defining an epitope present only on CEA, ICN) and a monoclonal antibody against NCA (F106-88,defining an epitope present only on NCA, Kyowa) (21, 22). Polyclonal antibodies against calcitonin (Dako) and calcitonin generelated peptide (CGRP, Amersham, Buckinghamshire, UK) were used to confirm the histopathologic diagnosis of the tumors, and all 73 cases of MTC demonstrated positive immunoreactivity for CT and CGRP in their tumor cytoplasm. Horseradish peroxidase (HRP)labeled goat IgG Fab fragments against rabbit IgG, HRP-labeled rabbit IgG Fab fragments against sheep
polyvalent immunoglobulins and HRP-labeled rabbit IgG Fab fragments against mouse polyvalent immunoglobulins were prepared in our laboratory (23). The endogenous peroxidase activity was inactivated with 0.3% H,02 in methanol for 30 min. The immunoreactivity was visualized by the DAB reaction. The stained sections were examined after methyl green counterstaining at pH 4.0. Formalin-fixed, paraffin-embedded specimens of colon carcinoma served as a positive control, and phosphate-buffered saline instead of the primary antibody applied to MTC sections served as a negative control. All the negative control slides were completely unreactive. For comparison with MTCs, 5 cases of papillary carcinoma of the thyroid and 5 cases of follicular carcinoma were examined with a monoclonal antibody against CEA (CMOlO), a polyclonal antibody (Dako) preabsorbed with NCA, and also the F106-88 anti-NCA antibody.
Figure 1. CEA immunoreactivity after staining with monoclonal antibody CMOlO in 4% paraformaldehyde-fixed tissue. Diffuse, intensive reactivity is evident in both the cytoplasm and cell membrane (immunoperoxidase staining for monoclonal CEA).
Figure 2. NCA immunoreactivity after staining with monoclonal antibody F106-88 in 4% paraformaldehyde-fixed MTC. NCA immunoreaction deposits are evident as cytoplasmic globules in a few tumor cells (immunoperoxidase staining for NCA).
RESULTS Sixty--three (86.3%) of the 73 formalin-fixed and paraffin -embedded MTC specimens revealed positive im-
Acta Pathologica Japonica 40 (4) : 1990
Figure 3. CEA/NCA immunoreactivity after staining with polyclonal CEA antibody (Dako) in 4% paraformaldehyde-fixed tissue. Diffuse cytoplasmic as well as cell membrane positivity are evident (immunoperoxidase staining for polyclonal CEA).
Figure 4. CEA immunoreactivity after staining with polyclonal anti-CEA antibody preabsorbed with NCA in a 4% paraformaldehyde-fixed specimen. Diffuse cytoplasmic and cell membrane immunoreactivity are evident. Serial sections shown in Figs. 1 to 4 reveal essentially the same staining pattern for the 3 CEA antibodies (immunoperoxidase staining for CEA).
munoreactivity with polyclonal CEA antibodies as well as with monoclonal antibodies CMO10, Parlam 4 and F33104, with no significant difference in their staining patterns. In positive cases, clear DAB deposits were observed in the tumor cytoplasm and/or cell membrane, whereas in negative cases there was no equivocal staining. NCA immunoreactivity was observed in 31 cases (42.5%), all of which were simultaneously positive for CEA. The staining patterns for CEA were essentially the same among the 5 anti-CEA antibodies. CEA immunoreactivity in formalin-fixed or paraformaldehydefixed samples was observed predominantly on the cell membrane in addition to cytoplasmic reactivity, as shown in Fig. 1. On the other hand, NCA immunoreactivity was evident in the cytoplasm as globular deposits (Fig. 2). The polyclonal antibody against CEA (Dako) also demonstrated similar immunoreactlvity to the monoclonal antibodies (Fig. 3), and NCA-preabsorbed polyclonal CEA antibodies showed no difference in staining pattern (Fig. 4). The membranous staining pattern for CEA was more prominent in samples fixed in
Bouin solution (Fig. 5). NCA reactivity in Bouin-fixed samples was evident in the cytoplasm as globular inclusions similar to those in Fig. 2. Only some of the tumor cells had immunoreaction deposits for NCA, whereas most were stained for CEA. In ethanol-fixed samples, diffuse and intense CEA immunoreactivity was predominant in the cytoplasm (Fig. 7). NCA immunoreactivity was also evident in the cytoplasm (Fig. 8), but was more strongly expressed. NCA-positive cytoplasmic inclusions were also demonstrated in CEA-stained sections similar to those in Fig. 7, and thus these cytoplasmic inclusions seemed to possess both CEA and NCA immunoreactivity. As shown in the above figures, the four different fixatives employed produced significant differences in staining pattern, although the incidence of the CEA-positive cases did not vary among the 5 different anti-CEA antibodies used. Neither NCA nor CEA (CMO10 and polyclonal) immunoreactivity was found in the 5 papillary carcinomas and 5 follicular carcinomas.
CEA in Medullary Thyroid Carcinoma (Kakudo et a/.)
Figure 5. CEA immunoreactivity after staining with monoclonal anti-CEA antibody CMOlO in Bouin fixed MTC. The reaction deposits are more intense in the cell membrane than in the cytoplasm (immunoperoxidase staining for CEA).
DISCUSSION Carcinoembryonic antigen is a large glycoprotein which was originally isolated from colon carcinoma (20, 21). The specificity of antibodies against CEA varies because the antigen possesses many different epitopes including those of NCA (18, 20, 22, 24-26). It is also found in a high proportion of carcinomas in various other organs. CEA immunoreactivity has been detected in medullary (C cell) carcinoma of the thyroid at an incidence ranging from 77% to 100%(9-19). The difference in incidence is postulated to be due to differences in specificity of the antibodies used, some of which might cross-react with NCA (18). In this study the incidence of CEA positivity was 86.3% using 5 antibodies which were highly specific for CEA. This incidence was obtained using 73 cases of MTC, the largest number examined in any single previous study. We confirmed that MTCs do in fact express CEA immunoreactivity at high incidence, and that NCA immunoreactivity with polyclonal CEA antibodies accounts for only a small proportion of total immunoreactivity in this type of tumor. CEA thus remains a useful tumor
Figure 6. NCA immunoreactivity in Bow-fixed sample, a serial section of Fig. 5. The cytoplasmic immunoreactivity is evident as globular inclusions similar to those in Fig. 2 (immunoperoxidase staining for NCA).
marker, similar to calcitonin in patients with medullary carcinoma of the thyroid, even though there are a few CEA-negative MTCs. In this study, four different fixatives were examined for their effects on CEA staining in MTC. Although differences in the fixative used did not influence the incidence of CEA positivity in MTCs, the staining patterns varied significantly. As shown in Figs. 1, 3, 5 and 7, membrane immunoreactivity for CEA predominated in Bouin-fixed specimens, whereas diffuse cytoplasmic staining was observed in alcohol-fixed specimens. Cytoplasmic immunoreactivity might be masked in Bouin-fixed materials. Both forms of localization were evident in formalin-fixed or paraformaldehyde-fixed samples. Such light-microscopic localization was reported by DeLellis et a/. using the indirect immunofluorescence method in buffered formalin-fixed MTC specimens (11). An immunoelectron microscopic study of CEA in MTC would help to clarify the true subcellular localization of CEA. NCA immunoreactivity was found in 42.5% of MTCs (49.2% of CEA-positive MTCs), all of which demonstrated simultaneous CEA immunoreactivity. No cases posi-
Acta Pathologica Japonica 40 (4) : 1990
Figure 7. CEA immunorectivity after staining with monoclonal
anti-CEA antibody CMOlO in an ethanol-fixed specimen. The immunoreactivity in the cytoplasm is more diffuse and intense. Arrows indicate globular deposits, which are much clearer in Fig. 8 (immunoperoxidase staining for CEA).
tive for NCA and negative for CEA were present in our series, thus the polyclonal antibody reacting with both CEA and NCA also showed the same incidence o f staining in the 7 3 MTCs, being essentially the same as that of more specific monoclonal antibodies. Detailed examinat io n of CEA-sta ined sections revealed immuno histochemical positivity on cytoplasmic inclusion bodies that were also observed in NCA-stained sections. Thus these inclusion bodies appeared t o contain both CEA and NCA immunoreactivities, as shown in Figs. 7 and 8. NCA is reported t o be stained in follicular cell tumors only at low incidence, and this has been considered a possible reason f o r the elevated serum CEA levels measured using antibodies reactive with both CEA and NCA in patients with papillary carcinoma or follicular carcinoma of the thyroid (18, 27, 28). However, neither anti-NCA (F106-88) nor anti-CEA antibodies (Dako polyclonal and CMO10) showed any positive reaction in our series of tumors originating from follicular cells. In the study by Schroder and Kloppel (18), the antibodies used, recognizing both CEA and NCA, might have detected different antigenic determinants o n NCA f r o m those
Figure 8. NCA immunoreactivity in an ethanol fixed specimen, a serial section of Fig 7 Cytoplasmic immunoreactivity with a few cytoplasmic globular inclusions (arrows) is shown
(immunoperoxidase staining for NCA).
recognized b y our anti-NCA antibody (F106-88) or other unknown epitopes which did not react with our antibody. Therefore it is concluded that CEA is detectable only in tumors originating f r o m C cells is in the thyroid. Acknowledgements : The authors thank Mr. Johbu ltoh for his technical help in microphotography, Ms. Yoshiko ltoh for immunohistochemistry and Ms. Yasuko lnagaki for secretarial work.
REFERENCES Tashjian AH Jr and Melvin KEW. Medullary carcinoma of the thyroid. Studies of thyrocalcitonin in plasma and tumor extracts. N Engl J Med 279: 279-283, 1968. Normann T, Johannessen JV, Gautvik KM, et a/. Medullary carcinoma of the thyroid-diagnostic problems. Cancer 38: 366-377, 1976. Sabate MI, Carpani M, Varndell IM, et a/. Calcitonin gene-related peptide in normal thyroid and medullary carcinoma of the thyroid (Abstract). J Pathol 142: ~ 2 3 9 1984. ,
CEA in Medullary Thyroid Carcinoma (Kakudo et a / . )
4. Morris HR, Panico M, Etienne T, e l a/. Isolation and characterization of human calcitonin gene-related p e p tide. Nature 308: 746 748, 1984. 5. Skrabanek P, Cannon D, Dempsy J, et a/. Substance P in medullary carcinoma of the thyroid. Experientia 35: 1259 1260, 1979. 6. Kakudo K and Vacca LL. lmmunohistochemical study of substance P like immunoreactivity in human thyroid and medullary carcinoma of the thyroid. J Submicrosc Cytol 15 : 563 568, 1983. 7. Sundler F, Almets J, Hakanson R, et a/. Somatostatin immunoreactive cells in medullary carcinoma of the thyroid. A m J Pathol 88: 381 386, 1977. 8. Kameya T, Shimosato Y, Adachi K, et a/. Immunohistochemical and ultrastructural analysis of medullary carcinoma of the thyroid. A m J P a t h o l 8 9 : 555-574, 1977. 9. lshikawa N and Hamada S. Association of medullary carcinoma of the thyroid with ca rcinoembryonic antigen. Br J Cancer 34: 111 115, 1976. 10. Issacson P and Judd MA. Carcinoembryonic antigen in medullary carcinoma of the thyroid. Lancet 2 : 1016 1017, 1976. 11 DeLellis RA, Rule AH, Spiler I, et a/. Calcitonin and carcinoembryonic antigen as tumor mqrkers in medullary thyroid carcinoma. Am J Clin Pathol 70: 587 594, 1978. 1 2 Lloyd RV, Sisson JC, and Marangos PJ. Calcitonin, carcinoembryonic antigen and neuron specific enolase in medullary thyroid carcinoma. An immunohistochemical study Cancer 51 : 2234-2239, 1983. 13 Mendelsohn G, Wells SA Jr, and Baylin SB. Relationship of tissue carcinoembryonic antigen and calcitonin to tumor virulence in medullary thyroid carcinoma. An immunohistochemical study in early, localized and virulent disseminated stages of disease. Cancer 54 : 657-662, 1984. 14 Holm R, Sobrinho Simoes M, Nesland JM, et a/. Medullary carcinoma of the thyroid gland: An immunocytochemical study. Ultrastruct Pathol 8 : 2541, 1985. 15 Uribe M, Grimes M, Fenoglio Preiser CM, and Feind C. Medullary carcinoma of the thyroid gland. Clinical, pathological, and immuohistochemical features with review of the literature. A m J Surg Pathol 9 : 577594, 1985. 16 Krisch K, Krisch I, Horvat G, et a/. The value of immunohistochemistry in medullary thyroid carcinoma : A systematic study of 30 cases. Histopathology 9 : 1077-1089, 1985.
17. Sikri KL, Varndell IM, Hamid QA, et a/. Medullary carcinoma of the thyroid. An immunocytochemical and histochemical study of 25 cases using eight separate markers. Cancer 56: 2481-2491, 1985. 18. Schroder S and Kloppel G. Carcinoembryonic antigen and non specific cross reacting antigen in thyroid cancer. An immunocytochemlcal study using polyclonal and monoclonal antibodies. A m J Surg Pathol 11: 100 108, 1987. 19. Takami H, Bessho T, Kameya T, et a/. Immunohistochemical study of medullary thyroid carcinoma : Relationship of clinical features to prognostic factors in 36 patients. World J Surg 12: 572-579, 1988. 20. Kuroki M, Arakawa F, Higuchi H, et a/. Epitope mapping of the carcinoembryonic antigen by monoclonal antibodies and establishment of a new improved radioimmunoassay system. Jpn J Cancer Res (Gann) 78: 386-396, 1987. 21. Kuroki M. Personal cornrnunicat/on. 22. Tsutsumi Y. Carcinoembryonic antigens in colon carcinoma. lmmunohistochemical approach. Byori to Rinsho 6 : 889-999, 1988 ( i n Japanese). 2 3. Willson M B and Nakane PK. Recent developments in the periodate method of conjugating horseradish peroxidase (HRPO) to antibodies. In Knapp W, Holubar K, and Wick G, eds. lmmunofluorescence and related s t a i n i n g techniques. Elesevier,/North H o l l a n d Biomedical Press, New York, Amsterdam, Oxford, 1978 : 215-225. 2 4. Gold P and Freedman SO. Demonstration of tumor specific antigens in human colonic carcinomata by immunological tolerance and absorption techniques. J Exp Med 121 : 439-462, 1965. 25. Tsutsumi Y, Nagura H, and Watanabe K. Immunohtstochemical observation of carcinoembryonic antigen (CEA) and CEA-related substances in normal and neoplastic pancreas. Pitfalls and caveats in CEA immunohistochemistry. A m J Clin Pathol 82 : 535 542, 1984. 26. Kuroki M, Koga Y, and Matsuoka Y. Purification and characterization of carcinoembryonic antigen-related antigen in normal adult feces. Cancer Res 41 : 713720, 1981. 27. Cimitan M, Busnardo B, Girelli ME, e t a / . Carcinoembryonic antigen in thyroid cancer. J Endocrinol Invest 2 : 241-245, 1979. 28. Wilson NW, Pambakian H, Richardson TC, et a/. Epithelial markers i n thyroid carcinoma: An immunoperoxidase study. Histopathology 10 : 81 5829, 1986.