Characterization of cells in salivary gland lesions by immunohistochemical identification of carcinoembryonic antigens

oral pathology Editor: JAMES J. SCIUBBA, D.M.D., Ph.D.

American Academy of Oral Pathology Department of Dentistry Long Island Jewish-Hillside Medical Center New Hyde Park, New York PI042

Characterization of cells in salivary gland lesions by immunohistochemical identification of carcinoembryonic antigens K. Tsukitani, D.D.S., K. Kobayashi, D.D.S., D.D.Sc., N. Murase, D.D.S., S. Stlmitomo, D.D.S., H. Mitani, D.D.S., and M. Mori, D.D.S., D.M.S., Gifu, Japan DEPARTMENT

OF ORAL

SURGERY,

GIFU COLLEGE

OF DENTISTRY

Immunohistochemical demonstration of carcinoembryonic antigen (CEA) was reported in normal salivary glands and in pathologic Uesions. Staining patterns of CEA and nonspecific cross-reacting antigen-absorbed CEA (NCAa-CEA) were compared. Normal salivary glands disclosed positive staining by CEA on border and luminal sides of acinar cells and occasionally in components secreted into ductal spaces with both antigens used. Chronic obstructive lesions displayed an intense CEA staining in ductlike structures and in material secreted into their lumina in the two antigens used. Pleomorphic adenoma exhibited varying intensities of CEA in neoplastic epithelial cells of ductal structures. In contrast, they showed a slight staining reaction to NCAa-CEA. Squamous-cell carcinomas showed a strong CEA reaction, whereas they showed no reaction or a trace reaction to NCAa-CEA. Positive staining to CEA in squamous neoplastic lesions was related to nonspecific reacting antigens. (ORAL SURG. ORAL MED. ORAL PATHOL. 59:595-599,

I

1985)

mmunohistochemical evaluation of CEA has been established for normal and inflamed parotid glands and parotid gland carcinomas,1-3as well as for various kinds of human neoplasms.4,5Caselitz and co-workers1.2have described immunohistochemical staining by CEA in parotid glands under normal, inflammatory, and neoplastic conditions. Previous immunohistochemical studies of CEA have shown a positive correlation between serum CEA levels and colonic adenocarcinoma. Radioimmunoassay for CEA has been used as a diagnostic and follow-up tool for this disease.4,6z7 The presence of CEA has been noted in many types of glandular tumor, as well as in nonneoplastic conditions, whose development may be related to nonspecific cross-reacting antigens.5,8That is, positive staining of immunohistochemically detectable CEA in tissue sections may

represent a false-positive reaction, and in order to immunologically identify true CEA, nonspecific cross-reacting antigen-absorbed CEA (NCAa-CEA) should be used. The present study demonstrates the presenceof NCAa-CEA in disease,including malignant conditions, of both major and minor salivary glands. MATERIALS Materials

AND METHODS

Major salivary gland specimens consisted of normal submandibular glands (26 specimens) and normal parotid glands (6 specimens), chronic obstructive adenitis (sialolithiasis, 17 specimens), and salivary gland tumors (38 pleomorphic adenomas, 1 epidermoid carcinoma). Specimens from 15 minor salivary glands were also examined. All specimens 595

96 Tsukitani et al. ~abte E. Comparison in distribution of CEA an NCAa-CEA in normal salivary glands and their lesions

I Normal salivary gland Acinar cell Striated and excretory duct Secreted material Obstructive adenitis Atrophic acini Ductlike structure Secreted material Pleomorphic adenoma Tumor cell Squamous metaplastic area Secreted material

CEA

1 NCAa-CEA

+2 -

i-1 -

+3

+1

+1”-+2 +I%+2 f3

+1 -l-l +1

Q-i-3 +4 +3

O--+-l o-.-I fl

The grade of staining intensity is arbitrarily divided into 0 to +4: +3 = strong; 0 = negative; t = trace: +l = weak; +2 = moderate; i4 = very strong.

used in this study were stained for CEA, while forty-one of these specimens were reacted with NCAa-CEA, and theresults were compared (Table I). ethods

Material obtained bvy surgical removal or biopsy was fixed in 10% formalin solution for 12 hours and embedded in paraffin. Four-micron paraffin sections were used for immunohistochemical detection by the indirect peroxidase method of CEA and NCAa-CEA as well as for routine hematoxylin and eosin staining. Immunohistochemical detection of CEA and NCAa-CEA. (1) Deparaffinized sections were rinsed with 0.05M TRIS buffer (pH 7.6) and immersed in a 0.3% hydrogen peroxide/ 100%methanol solution for 30 minutes in order to inactivate endogenousperoxidase. (2) The sections were immersed in a normal rabbit serum (1:20 dilution, Wheaton, U.S.A.) for 30 minutes. (3) They were next incubated with sheep anti-human CEA antiserum (1: 100 dilution, Serotee, U.K.) or anti-NCA absorbed sheep, anti-human CEA antiserum (1: 100 dilution, Serotec, U.K.) for 1 hour and then rinsed three times with phosphateuffered, saline solution (PBS) for 5 minutes. (4) They were subsequently incubated with horseradish peroxidase (HRP) conjugated rabbit anti-sheep IgG antiserum (1: 20 dilution, DAKOPATTS, Denmark) for 30 minutes and rinsed again. (5) Finally, they were incubated with TRIS buffer containing 0.3% 3,3-diaminobenzidine tetrahydrochloride (DAB, Dojindo Lab., Japan) with 0.05% hydrogen peroxide for 15 minutes. All reactions were performed in a moist chamber at 20” C.

Control sections. The same staining proce ut with normal sheep serum instead of anti for CEA or NCAa-CEA in step 3, were performed as a control test. Negative staining in all specimenswas noted. Trypsin digestion. Trypsin treatment prior to NCAa-CEA reaction was performed. The sections were treated with PBS containing 0.01% trypsin (Nakarai Chem. Ltd., Japan) for 15 minutes at 37” C prior to immunohistochemical staining of NCAa-CEA. ESULTS omparison NCAa-CEA

of distribution

between

CEA and

In the normal salivary gland, CEA staining was relatively stronger, including the background area, than NCAa-CEA staining. ~mmunohi§tochemical staining of CEA resembled NCAa-CEA staining (Figs. 1 and 2). Secretory material within the luminal cavities showed positive reactions to both CEA and NCAa-CEA. In the squamous area of pleomorphic adenoma and squamous-cell carcinoma, CEA staining was intense (Fig. 3), whereas Aa-CEA staining was negative or minimal (Fig. Effects of trypsin staining

digestion

prior to MCAa-CEA

Trypsin treatment prior to NCAa-CEA detection did not markedly change the staining pattern. Trypsin digestion prior to immunohistochemical reactions in tissue sections is useful since it decreases the nonspecific background uptake and enhances immunohistochemical staining probably by unmasking immunoreactive sites.9 ormal

salivary

glands

Immunohistochemical staining of NCAa-CEA in the normal major glands was detected along the intercellular borders and apical sides of the serous acinar cells (Fig. 2). Mucous cells of the submandibular gland indicated border staining in their basal areas and in the intercellular spaces. NCAa-CEA staining in mucous acinar cells was rather diffuse. Striated and excretory duct cells were devoid of NCAa-CEA reaction. Secreted material in the ductal lumina showed varying reactions. Xn the minor salivary glands, the mucous acinar cells and ductal components of the palatal glands displayed NCAaCEA staining in their luminal aspects only, while those of the tongue indicated positive staining along intercellular borders. Mucous acinar cells of the labial glands showed staining in their luminal regions.

CEA and NCAa-CEA

Volume 59 Number 6

in salivary gland lesions

597

Fig. 1,2. ‘Normal submandibular gland. Fig. 1. CE;A staining is moderately positive in apical side and intracellular border of acinous compartments, and
Obstructive

sialoadenitis

Cells of atrophic acini, ductlike spaces, and small multiple cystic structures showed intense NCAaCEA staining along luminal borders. Material in luminal spaces was also reactive to NCAa,-CEA.

Pleomorphic

adenoma

Immunohistochemical localization of NCAa-CEA in epithelial salivary gland tumors was evident in the ductal, microcystic, and glandular areas with variable degrees of staining intensity (Figs. 5 and 6).

598 Tmkitani et al.

Fig. Fig. their Fig. Fig. Fig.

Oral Surg. June. 1985

NCAa-CEA staining in pieomorphic adenoma. 5. NCAa-CEA reactions arc found to be moderately intense in duct-like tumor cells and intense in secreted material. 6. Secretory material generally shows moderate to strong staining. 7. Weak staining of neoplastic epithelial cells is located along the lumen border. 8. Positive reactions are weak at the border of the duct-like structures.

Generally NCAa-CEA was not detectable in the area containing spindle-shaped tumor cells (Figs. 7 and 8). Epithelial cells of salivary gland pleomorphic adenomas were not always positively stained for NCAa-CEA, whereas components secreted into the luminal cavities were usually positive (Fig. S). NCAa-CEA reaction in pleomorphic adenomas showed a high degree of irregularity in distribution. In some of the neoplastic epithelium, very strong NCAa-CEA staining was found (Fig. 6). However, tumors of similar histologic type indicated no NCAa-CEA staining or only a trace. Squamous cells and squamous cell carcinomas were generally devoid of NCAa-CEA staining (Fig. 4). There was no correlation between histologic types of pleamorphic

adenoma and immunohistochemical CEA.

distribution

of

lscussloN Since CEA is an established specific antigenic marker of some malignant tumors, it is helpful in establishing the diagnosis of malignant tumors in the gastrointestinal tract.4,8*io-” The immunohistochemical distribution of CEA has been described for various kinds of neoplastic lesion, including salivary gland tumors. I-3 Elevated CEA levels in the blood of patients with certain malignant tumors have been described, with concentrations of CEA detected compared by radioimmunoassay and immunohistochemical methods.6x‘3I63I7 Immunohistochemical dis-

Volume 59 Number 6

CEA and NCAa-CEA

in salivary gland lesions

599

tribution of CEA in normal parotid salivary glands occurred at the borders among acinar cells and in the luminal aspects of acinous components.‘,2 This border staining of CEA was generally confined.to serous acinar cells in major and minor salivary glands. Mucous acinar cells showed very weak and diffuse staining for CEA and positive staining in the apical regions. CEA in obstructive adenitis of the submandibular gland was present, and it increased in staining in secretory compartments of luminal spaces. This finding suggests that an increased amount of CEA on the luminal side or luminal spacesmight be derived from the blood. The biologic significance of the presence of CEA in epithelial compartments, such as salivary gland tumors,‘” mammary carcinomas,r3 gastrointestinal adenocarcinomas,“, I2 and sweat gland carcinomas,14* l5 is unclear, although occurrences, of CEA have been reported as markers of neoplastic cells. CEA has been (detected by immunohistochemical and radioimmunoassay techniques in normeoplastic conditions and in normal human saliva.lH CEA in such locations and tissues may originate in serum. In this regard, CEA in salivary glands seems to be a nonspecific marker for identifying malignancy irrespective of the use of labeled NCAa-CEA antiserum. Several types of nonspecific cross-reacting antigens are related to specific detection of CEA, that is, normal glycoproteins, fetal sulfoglycoprotein antigen, and CEA-associated protein.5,8 In the present experiment, immunohistochemically detectable CEA in normal salivary glands and in nonneoplastic lesions of the glands was compared to serum levels of CEA and NCAa-CEA. Strong staining in squamous cells and inflammatory cells was not a specific CEA reaction. Positive CEA staining may be related to a false-positive reaction, which can and should be verified by using NCAa-CEA serum in a confirmatory sense.

3. Caselitz J, Seifert G, and Jaup T: Tumor antigens in neoplasms of the human parotid gland. J Oral Path01 11: 374-386, 1982. 4. Egan ML, Go VLM: Carcinoembryonic antigen in gastrointestinal secretions. In Herverman RB, Mclntire KR (editors): Immunodiagnosis of cancer, Basel, 1979, Marcel Dekker, Inc., part 1, pp. 304-321. 5. von Kleist S, Burtin P: Antigens cross-reacting with CEA. In Herverman RB, McIntire KR (editors): Immunodiagnosis of cancer, Basel, 1979, Marcel Dekker, Inc., part 1, pp. 322342. 6. Kuhajda FP, Offutt LE, Mendelsohn G: The distribution of carcinoembryo~nic antigen in breast carcinoma: diagnostic and nroenostic imnlications. Cancer 52: 1257- 1264. 1983. I. Gold P, Shustkr J, Freedman SO: Carcinoembryonic antigen (CEA) in clinical medicine: historical perspectives, pitfalls and projections. Cancer 42: 1399-1405, 1978. 8. Burtin P, von Kleist S, Sabine MC, King M: Immunohistochemical localization of carcinoembryonic antigen and nonspecific cross-reacting antigen in gastrointestinal normal and tumoral tissues. Cancer Res 33: 3299-3305, 1973. 9. Huang S-N, Minassian H, More JD: Application of immunofluorescent staining on paraffin sections improved by trypsin digestion. Lab Invest 35: 383-340, 1976. 10. Isaacson P, Judd MA: Carcinoembryonic antigen (CEA) in the normal human small intestine: a light and electron microscopic study. Gut 18: 786-791, 1977. of 11. Wiley EL, Mendelsohn G, Eggleson JC: Distribution carcinoembryonic antigens and blood group substances in adenocarcinoma of the colon. Lab Invest 44: 507-5 13, 198 1. 12. Nagura H, Tsutsumi Y, Shioda Y, Watanabe K: Immunohistochemistry of gastric carcinoma and associated diseases: novel distribution of gastric cancer cells. J Histochem Cytothem 31: 193-198, 1983. 13. Walker RA: Demonstration of carcinoembryonic antigen in human breast carcinomas by the immunoperoxidase technique. J Clin Path01 33: 356-360, 1980. 14. Penneys NS, Nadji M, Morales A: Carcinoembryonic antigen in benign sweat gland tumors. Arch Dermatol 118: 225-227, 1982. 15. Penneys NS, Nadji M, Ziegela-Weissman J, Ketabchi Morales AR: Carcinoembryonic antigen in sweat gland carcinomas. Cancer 50: 1605-1611, 1982. 16. Goldenberg DM, Sharkey RM, Primus FJ: Carcinoembryonic antigen in histopathology: immunoperoxidase staining of conventional tissue. J Nat1 Cancer Inst 57: 1 l-22, 1976. 17. Goldenberg DM, Sharkey RM, Primus FJ: Immunohistochemical detection of carcinoembryonic antigen in conventional histopathology specimens. Cancer 42: 1546-1553, 1978. antigen in normal 18. Martin F, Devant J: Carcinoembryonic human saliva. J Nat1 Cancer Inst 50: 1375-1379, 1973.

REFERENCES

Reprint requests to: Dr. K. Tsukitani Department of Oral Surgery Gifu Collene of Dentistry 1851-l Hoiumi, Hozumi-cho Motosu-gun, Gifu-pref. Sol-02 Japan

1. Caselitz J, Seifert G, Jaup T: Presence of carcinoembryonic antigen (CEA) in the normal and inflamed parotid gland. J Cancer Res Clin Oncol 100: 205-211, 1981. 2. Caselitz J, Jaup T, Seifert G: Immunohistochemical detection of carcinoembryonic antigen (CEA) in parotid gland carcinomas. Virchows Arch [Pathol Anat] 394: 49-60, 1981.