Journal of Immunological Methods, 8 (1975) 267--276 © North-Holland Publishing Company, Amsterdam -- Printed in The Netherlands

DETECTION OF CARCINOEMBRYONIC ANTIGEN IN TISSUE SECTIONS BY IMMUNOPEROXIDASE*

F. JAMES PRIMUS and REENIE H. WANG

Department of Immunology, Hoffmann -- La Roche Inc., Nutley, New Jersey 07110, U.S.A. and ROBERT M. S HAR K E Y and DAVID M. GOLDENBERG

Department of Pathology, University of Kentucky Medical Center, Lexington, Kentucky 40506, U.S.A.

A triple-bridge, indirect, immunoperoxidase method for detecting and localizing carcinoembryonic antigen (CEA) in tissue sections is described. By this technique, a cell-surface localization of CEA in colonic carcinoma and ovarian mucinous cystadenocarcinoma cells could be visualized. In the case of the colonic cancer, both the tumor from the descending colon and a metastasis to the skin gave positive peroxidase reactions for CEA. This immunocytochemical method for demonstrating the presence of CEA functioned in both frozen, ethanol-fixed and formalin-fixed, paraffin-embedded tissues, thus making it applicable for use with tissue sections conventionally prepared for light microscopy.

INTRODUCTION

Carcinoembryonic antigen (CEA) was originally detected in digestive tract cancers, b u t not in normal adult gastrointestinal tissues (Gold and Freedman, 1965). Likewise, after the development of a radioimmunoassay for CEA, it was initially reported that only patients with gastrointestinal malignancy had elevated levels of this antigen in their sera (Thomson et al., 1969). Subsequent studies have revealed that CEA or an immunochemically related substance(s) is present in increased quantities in the plasma or serum of patients with diverse kinds of cancer and in some benign conditions (Lo Gerfo et al., 1971; H o l y o k e et al., 1972; Laurence et al., 1972; Zamcheck et al., 1972; Hansen et al., 1974. At present, it appears that the CEA assay can provide useful information to supplement other conventional diagnostic procedures for certain kinds of cancer, and in following the course of disease in patients in w h o m elevated CEA levels were found prior to therapy (Holyoke et al., 1972; Hansen et al., 1974; Sorokin et al., 1974). * Reprint requests should be forwarded to Dr. D.M. Goldenberg, Department of Pathology, University of Kentucky Medical Center, Lexington, Kentucky 40506.

268 A relatively unexplored area is the detection and localization of CEA in tissue sections. The presence of CEA in such specimens would appear of particular importance in those patients with cancer who do not have elevated plasma CEA titers. For example, 50% to 80% of patients with established colon cancer limited to the bowel wall have normal plasma CEA values (Dhar et al., 1972; Lo Gerfo et al., 1972; Laurence et al., 1972; Mach et al., 1974), and 83% of patients with primary mammary cancer without metastases have normal plasma titers of CEA (Steward et al., 1974). Do these results indicate that circulating CEA reflects bulk of t u m o r present, rate of CEA biosynthesis by the tumor, metabolism of circulating antigen by the patient, access of the t u m o r to the circulation, or combinations of these factors? One approach to gaining an understanding of the relationship of CEA produced by a t u m o r to the neoplasm's stage of development and to the a m o u n t of CEA circulating in the patient's blood is to study the presence of the antigen in t u m o r specimens. We now report a relatively simple, convenient, and reproducible method for CEA detection and localization in conventionally prepared tissue sections. MATERIALS AND METHODS

Histological specimens Surgical biopsy specimens from colonic and ovarian cancers were divided for either fixation in 10% formalin or quick-freezing in tissue embedding medium (Tissue-Tek, Ames Co., Elkhart, Indiana) w i t h o u t prefixation. The formalin-fixed tissues were processed for embedding in paraffin blocks by the usual techniques. Sections of 4--6 p thickness were then cut from each specimen and placed on glass slides. The frozen sections were fixed in absolute ethanol at --70°C while the paraffin sections were deparaffinized with xylene and passed through graded dilutions of ethanol. Each slide was subsequently hydrated by washing in two 5-min changes of 0.01 M phosphate-buffered saline (PBS), pH 7.2. Sections from each specimen were also stained with hematoxylin and eosin for comparison purposes.

An tisera Goat anti-CEA antiserum was prepared by biweekly injections of 500 #g purified CEA (Newman et al., 1974), coupled with 1 mg methylated bovine serum albumin (Sigma Chemical Co., St. Louis, M o . ) a n d emulsified in an equal volume of complete Freund's adjuvant (Difco, Detroit, Mich.). The antigen was injected subcutaneously into 2 separate sites for 2 months. Crossreacting antibody against colon carcinoma, antigen-III, CCA-III (Newman et al., 1974), in the antiserum was removed by adsorption onto CCAIII-containing immunoadsorbents according to procedures described elsewhere (Primus et al., 1975). Likewise, the same anti-CEA antiserum served as

269

a control for specificity by passing it over a CEA-containing immunoadsorbent, thus producing a CEA-neutralized antiserum. Both of these procedures exhaustively removed CCA-III-crossreacting or CEA-specific antibodies. Rabbit antiserum to goat IgG was prepared by biweekly subcutaneous injections of 2 mg IgG (Miles Laboratories, Kankakee, Ill.) emulsified in an equal volume of complete Freund's adjuvant over a course of 2 months. Goat antiserum to horseradish peroxidase (Sigma Type VI) was prepared in a similar fashion. These antisera were used without further modification.

Immunoperoxidase reaction The immunoperoxidase technique of Mason et al. (1969), as modified by Hsu et al. (1972), was used to demonstrate CEA. This triple-bridge method is depicted schematically in fig. 1, and consists of goat anti-CEA antibody (crossadsorbed with CCA-III) reacted with the antigen-containing tissue, the rabbit anti-goat IgG b o u n d to the previous antibody, and goat anti-peroxidase which is likewise b o u n d to the rabbit anti-goat IgG antibody. As a control for the anti-CEA antibody step, we always reacted a duplicate section on each slide with the CEA-neutralized antibody in order to evaluate background or endogenous peroxidase activity. The exact procedure carried out was as follows: one to 2 drops of antibody to CEA diluted 1 : 300 in PBS were added to the tissue sections and incubated in a moist chamber at 37°C for 10 min. A similar incubation was made with CEA-neutralized anti-

HORSERADISH PERC . . . . . . .

RABE IgG

I;LLL- 5UH~ALt

ANTIGEN ( C E A )

Fig. 1. Simplified schematic representation of triple-bridge, indirect immunoperoxidase reaction for CEA cell-surface marker, in which any valency depicted for antibody or antigen is not intended.

270

body substituted for the specific antiserum on the adjacent control section. The sections were then washed in PBS for two 5-min changes, incubated with rabbit antibody to goat IgG, 1 : 50 dilution, for 10 min at 37°C, washed in PBS, again incubated with goat anti-peroxidase antibody (1 : 100 dilution) for 10 min at 37°C, and then washed in PBS. The sections were treated with horseradish peroxidase (Sigma Type VI, 100 pg/ml PBS) at 37°C for 20 min, washed, and the histochemical reaction developed by incubating at room temperature for 20 min in 0.05 M Tris buffer, pH 7.6, containing 0.075% 3,3'-diaminobenzidine (Sigma Chemical Co., St. Louis, Mo.) (w/v) and 0.05% H2 02 (w/v), according to the method of Graham and Karnovsky (1966). The sections were thereafter washed in PBS, dehydrated with graded dilutions of ethanol, treated with xylene, and finally mounted under a coverslip with Permount.

Fig. 2. I m m u n o p e r o x i d a s e r e a c t i o n for C E A in c o l o n c a r c i n o m a s p e c i m e n fixed in form a l i n a n d e m b e d d e d in paraffin. M a g n i f i c a t i o n x 750.

271 RESULTS

The t w o t u m o r specimens examined consisted of an adenocarcinoma of the descending colon and a mucinous cystadenocarcinoma of the ovary. The colonic carcinoma had metastasized to the the skin, from which a specimen was also available. Paraffin sections of the colonic carcinoma revealed a membrane staining for CEA on the cells of the colon specimen (fig. 2) and of the skin metastasis (fig. 3) with the immunoperoxidase reaction. Under higher magnification, it appeared that the t u m o r acini stained preferentially at the apical border of the cells (fig. 4). The control reaction, using CEAneutralized antibody, did n o t reveal any peroxidase activity. Frozen sections of these specimens reacted similarly. The ovarian mucinous cystadenocarcinoma, which was rapidly frozen and then fixed in cold ethanol, revealed surface staining of the malignant cells

Fig. 3. I m m u n o p e r o x i d a s e r e a c t i o n for C E A in c o l o n c a r c i n o m a m e t a s t a t i c to skin ( f i x e d in f o r m a l i n a n d e m b e d d e d in paraffin). M a g n i f i c a t i o n × 300.

272

Fig. 4. Peroxidase-labelling

of CEA

from section

of fig. 3. Magnification

X

750.

and groups of cells (fig, 5). No reaction could be seen in the stromal tissue or when other sections were treated with the CEA-neutralized antibody instead of the specific antibody preparation. Formalin-fixed, paraffin-embedded sections also revealed the presence of CEA by immunoperoxidase, but predominantly the cells comprising the cystic borders. A normal adult ovary and a colon, both obtained as autopsy specimens, were treated similarly to demonstrate CEA immunocytochemically. In both cases we failed to visualize a peroxidase reaction.

273

Fig. 5. Immunoperoxidase reaction for CEA in ovarian mucinous adenocarcinoma specimen after rapid freezing and fixation in cold ethanol. Magnification x 750.

DISCUSSION This study demonstrates the feasibility of detecting and localizing CEA in t u m o r tissue specimens by a relatively simple i m m u n o c y t o c h e m i c a l technique. Since low levels of CEA have been r eport ed in normal h u m a n colon by radioimmunoassay (Martin and Martin, 1970; Rosai et al., 1972), it would seem th at our i m m u n o p e r o x i d a s e reaction is n o t as sensitive in detecting minute amo u n ts of CEA in normal colonic mucosa or in stromal c o m p o n e n t s surrounding colonic or ovarian carcinoma cells. In the specimens examined, CEA was demonstrable bot h in frozen, ethanol-fixed and in formalin-fixed, paraffin-embedded sections. Thus, it would appear t hat tissues conventionally processed f o r hi s t opat hol ogy could be utilized for det ect i on of CEA.

274

Although the presence of CEA in colonic carcinoma is to be expected, it is interesting that the antigen was detected in a mucinous cystadenocarcinoma of the ovary, thus confirming other immunochemical data recently obtained (Van Nagell et al., 1975). The demonstration of CEA in a colonic cancer metastasis to the skin suggests another possible application of this method; namely, to assist in the identification of the primary neoplasm in cases of metastatic cancer in which the primary tumor is not obvious. CEA was found to be localized predominantly on the cell surface by means of the triple-bridge immunoperoxidase reaction, thus confirming immunofluorescent studies (Gold et al., 1968; Goldenberg et al., 1972). The immunofluorescent technique for antigen localization, however, has a number of disadvantages. (1) Naturally occurring autofluorescence can interfere with the specific reaction, (2) fading with time occurs and only a temporary result is available (thus requiring photographic documentation), (3) a darkfield ultraviolet microscope is required, and (4) the method cannot be adapted for ultrastructural studies. The immunoperoxidase reaction used in this study obviates these limitations. Moreover, it has been shown in other problems that the peroxidase-labelled antibody method is as specific as the immunofluorescent technique (Nakane and Pierce, 1967; Murphy et al., 1973). Whether or not they are equally sensitive must await further investigation.

REFERENCES Dhar, P., T. Moore, N. Zamcheck and H. Kupchik, 1972, J. Am. Med. Assoc. 221, 31. Gold, P. and S.O. Freedman, 1965, J. Exp. Med. 212,439. Gold, P., J.M. Gold and S.O. Freedman, 1968, Cancer Res. 28, 1331. Goldenberg, D.M., R.A. Pavia, H.A. Hansen and J.P. Vandevoorde, 1972, Nature New Biol. 239, 189. Graham, R.C., Jr. and M.J. Karnovsky, 1966, J. Histochem. Cytochem. 14, 291. Hansen, H.J., J.J. Snyder, E. Miller, J.P. Vandevoorde, O.N. Miller, L.R. Hines and J.J. Burns, 1974, Human Pathology 5, 139. Holyoke, D., G. Reynoso and T.N. Chu, 1972, Ann. Surg. 176, 559. Hsu, K.G., E.A. Zimmerman, L. Rudin, P. Lo Gerfo, S. Bennett and M. Tannenbaum, 1972, in.: Proc. Second Conf. Embryonic and Fetal Antigens in Cancer, Vol. 2 (N.G. Anderson and J.H. Coggin, Jr., eds.), U.S. Dept. of Commerce, Springfield, Va., 147. Laurence, D.J.R., U. Stevens, R. Bettelheim, D. Darcy, C. Leese, C. Turberville, P. Alexander, E.W. Johns and A.M. Neville, 1972, Brit. Med. J. 3,605. Lo Gerfo, P., J. Krupey and H.J. Hansen, 1971, New Engl. J. Med. 285, 138. Lo Gerfo, P., F. Lo Gerfo, F. Herter, H.G. Barker and H.J. Hansen, 1972, Amer. J. Surgery 123, 127. Mach, J-P., G. Pusztszeri, D. MacDonald and M-M. Bertholet, 1974, in: The Role of Immunological Factors in Viral and Oncogenic Processes; Proc. VII Miles Internat. Symp., (R.F. Beers, Jr., R. Carmichael Tilghman and E.G. Bassett, eds.), The Johns Hopkins Press, 369. Martin, F. and M.S. Martin, 1970, Int. J. Cancer 6, 352. Mason, T.E., R.F. Phifer, S.S. Spicer, R.A. Swallow and R.B. Dreskin, 1969, J. Histochem. Cytochem. 17,563.

275 Murphy, W.M., S.D. Deodhar and L.P. Cawley, 1973, Clin. Chem. 19, 1370. Nakane, P.K. and G.B. Pierce, Jr., 1967, J. Histochem. Cytochem. 14, 929. Newman, E.S., S.E. Petras, A. Georgiadis and H.J. Hansen, 1974, Cancer Res. 34, 2125. Primus, F.J., E.S. Newman and H.J. Hansen, 1975, J. Immunol., submitted for publication. Rosai, J., T.W. Tillack and V.T. Marchesi, 1972, Int. J. Cancer 10, 357. Steward, A.M., D. Nixon, N. Zamcheck and A. Aisenberg, 1974, Cancer 33, 1246. Sorokin, J.J., P.H. Sugarbaker, N. Zamcheck, M. Pisick, H.S. Kupchik and F.D. Moore, 1974, J. Am. Med. Assoc. 228, 49. Thomson, D.M.P., J. Krupey, S.O. Freedman and P. Gold, 1969, Proc. Natl. Acad. Sci. (Wash.), 64, 161. Van Nagell, J.R., Jr., Q.A. Pletsch and D.M. Goldenberg, 1975, Cancer Res., in press. Zamcheck, N., T.L. Moore, P. Dahr and H. Kupchik, 1972, New Engl. J. Med. 286, 83.