Comparative Studies on Expression of Tumor-Associated Antigens in Human and Induced Pancreatic Cancer in Syrian Hamsters H i r o s h i E g a m i , t Y o s h i y u k i T a k i y a m a , 1 W i l l i a m G. C h a n c y , 2 Martin Cano, ~H i d e k i Fujii, ~ T s u t o m a Tomioka, t R i c h a r d M e t z g a r , ~ a n d P a r v i z M . P o u r *'f'4 tEppley Institute for Research in Cancer and AEied Diseases; 2Department of Biochemistry, University of Nebraska Medical Center (UNMC), 42nd Street and Dewey Avenue, Omaha, ArE 68105; 3Department of Microbiology and Immunology, Duke University Medical Center, Durham, NC 27710; and *Department of Pathology and Microbiology, UNMC, 42nd Street and Dewey A venue, Omaha, NE 68105
Summary The expression of blood-group-related antigens (BGRAs) in experimental primary pancreatic cancer induced by N-nitrosobis(2-oxopropyl)amine (BOP) treatment of Syrian hamsters and homologous subcutaneous transplants of this primary cancer in the celt line, PC-l, established from the primary cancer and intrapancreatic transplanted PC-t cells were studied by histochemical and biochemical methods. Human primary pancreatic cancer; the human pancreatic cancer cell line, HPAF; and its subclones, CD11 and CD18, also were studied on a comparative basis. Histochemical analysis of BGRAs demonstrated that A, B, H, Leb, LeX, LeY, and T antigen were expressed both in vivo and in vitro in hamster and human materials in similar patterns. However, Lea, CA 19-9 and sialylated Tn antigens were not found in hamster-derived tissues. SDS-PAGE and Western blotting procedures using anti-A antigen revealed similar major bands in the membrane fractions of both human and hamster pancreatic cells between 97 and 200 kdalton. Among other human pancreatic cancerassociated antigens, TAG-72, CA 125, and 17-1A were detected immunohistochemicatly in the hamster tumors both in vivo and in vitro, in a pattern similar to that seen in human pancreatic cancer. Tumor antigen DU-PAN-2, associated with human pancreatic cancer, was found infrequently in hamster pancreatic cancer specimens. These results indicate that the experimental hamster pancreatic cancer model provides a unique tool for investigating
*Author to whom all correspondence and reprint requests should be addressed. International Journal of Pancreatoiogy 91
9 1990 The Humane Press Inc.
92
Egami et al. antigenicity of pancreatic cancer, particularly in relation to diagnosis and therapy. Key Words: Tumor immunology; pancreatic cancer; human; hamster; blood group antigens.
INTRODUCTION Pancreatic cancer presents a dilemma diagnostically and therapeutically. Progress in the treatment of cancer o f the pancreas has been slow, in part because of the natural history of the disease with its silent course and terminal outcome. Fortunately, in recent years some pancreatic cancer models useful for understanding some aspects o f the disease have been developed (1). Although each of these models is important in its own right, the Syrian hamster model has been shown to be unique for studies concerned with the clinical aspects of the disease, because the induced cancer in this model mimics human pancreatic cancer in morphologic patterns and clinical presentations (2). Therefore, part of our studies on this model focused on immunologic aspects of the disease. Presently, no methods detect pancreatic cancer at the early, curable stage, mainly because initial symptoms are frequently vague and nonspecific, and sufficiently sensitive and specific diagnostic procedures for routine checkups are not available. To understand pancreatic cancer antigenicity and to develop more sensitive diagnostic or effective therapeutic methods, comparative in vivo and in vitro studies with human and hamster pancreatic cancers were performed in our laboratories. In this paper we describe part of our recent results. MATERIALS AND METHODS
Pancreatic Cancer Tissues (Solid Tumors) The following hamster pancreatic cancer tissues were used: (a) primary hamster pancreatic cancer induced by N-nitrosobis(2-oxopropyl)amine (BOP) (3), (b) subcutaneously transplanted primary hamster pancreatic cancer (4), and (c) intrapancreatic transplants of a hamster pancreatic cancer cell line (PC-l) as described (5). Human pancreatic cancer tissues were obtained by biopsy, surgery, or autopsy.
Pancreatic Cancer Cell Lines (Cell Culture) A newly established hamster pancreatic cancer cell line, PC-1 (5), and a human pancreatic cancer cell line, HPAF (6), and its subctones, CD11 and CD18, were used.
Monoclonal Antibodies (MoAbs) MoAbs against the synthetic trisaccharides of blood groups A, B, and H determinant were obtained from DAKO Co., Santa Barbara, CA. Anti-Lewis
A ntigenicity of Pancreatic Cancer
93
related antigens were kindly provided by Z. Steplewski, Wistar Institute, PA; anti-T and anti-sialylated Tn antigen by T. Nujaim, E d m o n t o n University, Canada; OC125, CA 19-9, and CA 17-1A by Centocor Inc., Malvern, PA. DU-PAN-2 and B72o3 were generated and characterized as described (7,8). As control materials, MOPC-21 and MoAb UPC-10 (Litton Bionetics, Charleston, SC) and the culture supernatant of P3x63Ag 8 murine myeloma cell line were used as described (9,10).
Histology and Immunoperoxidase Procedures Paraffin-embedded tissues were cut into serial, 4-/xm-thick sections. One of these sections was stained with hematoxylin and eosin. Others were processed by immunoperoxidase procedures by Vectatin ABC Kit (Vector Laboratories, Burlingame, CA) as described (5). The cultured pancreatic cancer cell lines were pelleted and a small amount of 1% agarose (Sigma Chemical Co., St o Louis, MO) was added at 37 ~ The agar plug was dehydrated and processed for histology. Control slides were examined as follows: (a) tissues were incubated with phosphate buffered saline (PBS) instead of primary antibodies; (b) tissues were incubated with UPC-10, MOPC-21 or culture supernatant of P3x63Ag8 murine myeloma cell line instead of primary MoAbs.
Scoring The staining was scored on the percentage of cells showing staining. Samples were placed into one of the following categories, depending on the percentage of reactive cells: 0% (-); < 5 % (1 +); 5-30~ (2+); 30-70070 (3+); >70070 (4+).
Immunogold Techniques for Scanning Electron Microscopy Immunogold techniques were performed on culture cells as described (11) using Auro Probe EM goat-antimouse IgM (30-nm particle; Janssen, Piscataway, N J).
ELISA Methods An ELISA method was used as reported (5).
Preparation of Plasma Membrane Fractions Plasma membrane of cancer cells from each cell line, normal hamster pancreatic tissue and duodenal epithelial ceils (which express A antigen (12)), and red blood ceUs from a healthy blood-group-A individual were prepared as described (13). Protein contents were measured by micro BCA protein assay (Pierce Chemical Co., Rockford, IL) (14).
SDS-PA GE and Western Blotting Analysis Electrophoresis was carried out in 5-14% (w/v) gradient polyacrylamide gels, according to the method of Laemmli (15), and proteins were transferred
94
Egami et al. Table t Expression of Blood Group Antigens in Experimental Hamster Pancreatic Cancer Model
Blood group antigens Primary induced tumor Subcutaneously transplanted tumor Intrapancreatic transplanted tumor PC- 1 cells in vitro
A
B
H
Le ~
Le b
Le x
Ley
4+
3+
2+
-
1+
3+
4+
4+
3+
2+
-
1+
3+
4+
4+
3+
2+
-
1+
3+
4+
4+
3+
1+
-
4+
2+
2+
to polyvinylidine difluoride (PVDF) membrane (Immobilon, 0.45 #m pore size, Millipore Co., Bedford, MA). Immunoblotting was performed as described (16), using MoAbs, anti-A antibody, antimouse IgM, and alkalinephosphatase conjugated antigoat IgG (Sigma Chemical Co., St. Louis, MO). RESULTS
Histochemical Studies BGRAs The expression o f BGRAs in primary hamster pancreatic cancers, subcutaneously transplanted tumors, intrapancreatic transplants o f PC-1 cells, and PC-1 cells growing in culture, are summarized in Table 1. In solid tumors from hamsters, a strong expression of A antigen in diffuse cytoplasmic and glycocalyx pattern was observed in almost all cancer cells. B antigen was expressed in about 50% of the cells, mainly in the glycocalyx pattern; H antigen in 20% of the cell population, primarily in the cell cytoplasm; Le b in 5% o f the cells, in a grandular cytoplasmic pattern; Le ~ in 50% of the cells, in the glycocalyx form; and Ley in more than 70% of cancer cells, in both the cytoplasmic and glycocalyx patterns. The staining pattern of hamster pancreatic cancer tissues with each o f these MoAbs was similar to that observed in h u m a n tissue (9). In PC-1 cells in culture, A antigen was expressed in 100%, B in 50%, H in 5%, Le b in 90%, and Le x and LeY each in 20% of cancer cells. Le a antigen and Ca 19-9 could not be detected in any hamster-derived pancreatic cancer cells, either in vivo or in vitro. T antigen was found in a b o u t 30% o f hamster-derived cancer cells in vivo (solid tumors) and almost all PC-1 cells in vitro in a granular pattern similar
Antigenicity of Pancreatic Cancer
95
Fig. 1. Reactivity of OC125 with (a) hamster and (b) human in granular pattern. ABC method, x200, to that seen in human pancreatic cancer tissues (data not shown). The MoAb against the synthetic sialylated Tn antigen did not react with any pancreatic cancer ceils derived from either humans or hamsters.
Tumor-Associated Antigens Thirty percent of the cells of primary, subcutaneous, or intrapancreatic transplanted hamster tumors and of PC-I cells in tissue culture were immunoreactive with MoAb B72.3. The staining pattern was cytoplasmic and granular, similar to that observed in human pancreatic cancer ceils (10)o Normal hamster pancreatic tissues were not stained with B72.3. Antibody OC125 was immunoreactive with 50~ of cells of primary, subcutaneous, and intrapancreatic transplanted hamster cancer in a cytoplasmic and granular pattern. A similar pattern of reactivity was also seen in human pancreatic cancer tissue (Fig. 1). Less than half of the cultured PC-1 cells were immunoreactive with OC125. The staining pattern in these cells was also mostly cytoplasmic granular. In the normal hamster pancreatic tissues, no reactivity with OC125 was found. About 30070 of hamster-derived cancer cells, both in vivo and in vitro, were immunoreactive with CO 17-1A in cytoplasmic and granular form, comparable to those seen in human pancreatic cancer cells (Fig. 2). The normal pancreatic tissue CO17-1A showed positive reactivity with acinar cells in a granular cytoplasmic form. MoAb DU-PAN-2 was immunoreactive with a few tumor cells derived from hamsters in vivo but not in vitro.
ImmunogoM Scanning Electro Microscopy Many gold particles were detected on the cell surface of PC-I ceils, virtually covering the entire surface of the cells, including the microvilli (Fig. 3).
96
E g a r n i et al.
"
- <~........
h:.<
<
:i~{: : s ^..,
....~ , " ~
L
<~ . . . . . . . :.
"
D
Fig. 2. Reactivity of CO17-A with (a) hamster and (b) human in cytoplasmic granular form. ABC method, x 200.
Fig. 3. (a) Blood group A antigen on PC-1 cell surface, detected by scanning electron microscopy using gold-labeled antimouse IgMo (b) Reverse-phase view depicting gold particles as black grains.
97
Antigenicity of Pancreatic Cancer --
200
kd
--
97
kd
-
68
kd
43
kd
-
18
kd
--
14
kd
Fig. 4. Western blotting procedure with MoAb anti-A, using cell membrane fractions of hamster pancreatic cancer ceils (PC-l), culture medium of PC-1 cells, and two subclones of HPAF cells, CD-I 1 and CD-18. Cell fractions of the normal pancreatic cells were used as negative material. Duodenum was used as a positive control for the hamster celI line, and RBC was used as a positive control for the human cell lines.
EL1SA A higher titer of A-reactive substance, but not of other blood group-related substances, was detected in culture supernatant of PC-I cells (5), indicating that only A antigen was shed from the cells. In culture supernatants of H P A F , CD11, and CD18 cells, which also express A blood group antigen, no A-reactive substance was detected.
SDS-PA GE and Western Blotting The results o f SDS-PAGE and Western blotting o f membrane fractions prepared from the tumors and cell lines described above are shown in Fig. 4. M e m b r a n e proteins were separated by SDS-PAGE, transfered to immobilon, and prepared for reactivity with a M o A b to A antigen, as described above. Membranes of PC-1 cells exhibited four bands with molecular masses between 68 and 200 kdalton, including a m a j o r component with a molecular mass o f about 120 kgdalton. Five bands between 68 and 200 kdalton were observed in the culture supernatant o f PC-1 cells, but the bands lacked the major c o m p o n e n t seen in membrane fraction o f PC-1 cells. Normal hamster duodenal epithelium membranes contained A-reactive proteins in a diffuse pattern between 97 and 200 kdalton; normal hamster pancreas membranes did not react with the anti-A antibody. Membrane fractions of the h u m a n pancreatic cancer cell lines, CD11 and CD18, showed two bands between 68 and 120 kdalton, including a m a j o r component with the same molecular
98
Egami et al.
weight as that found in PC-1 cells. The membrane fraction of human red blood cells of blood group A exhibited a weak band with a molecular mass of about 35 kdalton. DISCUSSION BGRAs were expressed by hamster pancreatic cancer cells, both in vivo and in vitro, in a pattern similar to that found in the human material. In contrast to the human situation (9), no reactivity with anti-Le" and CO19-9 was seen in either normal or malignant pancreatic cells in hamsters. Blood group A antigen was the most common and consistent antigen in hamster-derived pancreatic cancer cells, as it is in pancreatic cancer patients of blood group A (9). Group A antigen was also found in the culture supernatant of PC-1 cells. However, other BGRAs present in pancreatic tumor cells are not found in the culture medium, indicating that only A antigen is shed from tumor cells. On the other hand, differences in the molecular weights of A-reactive material between the PC-I cells and culture medium could indicate differences in the nature of the membrane-bound and shed antigen. The similarities in the pattern of immunoblotting of pancreatic cancer cell membrane between hamster and human implies that human and hamster cancer cells produce antigens with a similar structure. The dissimilarities found in the migration patterns of immunoreactive material between PC-1 cells and duodenal cells, and between human cancer cells and red blood cells, may indicate distinct differences in the structure of the antigen produced by the normal and malignant cells. As in the other tumors, in which not only the antigen but also its precursor substance is produced in excess and accumulated (17), T antigen was found in both solid and cultured hamster cancer cells in a pattern similar to that observed in human tumors (data not shown). The lack of reactivity of antisialylated Tn and the positive staining of the same material from both hamsters and humans with B72.3, which presumably recognizes a sialylated Tn (18), indicate differences in the antigen recognition by these two MoAbs. Expression in hamster pancreatic cancer cells of TAG-72, a common antigen in human pancreatic cancer (10), is also of interest because the expression of TAG-72 disappears in cell culture of human pancreatic cancer cells (10). The rationale behind examining the expression of CA 125 by hamster pancreatic cancer ceils was based on our unpublished observation that this antigen was expressed in the normal human pancreatic cells belonging to blood group A. We also noticed that the antigen was present in more pancreatic cancers of blood group A or AB patients than cancers of blood group O or B patients, and was present in higher concentrations in the former groups (A or AB). Indeed, CA 125 was found in both the solid and cultured hamster pancreatic cancer in a pattern similar to that seen in humans. Also, the expression of 17-1A (a monoclonal antibody that is being used for immunotherapy of pancreatic cancer because of its antibody-dependent
A ntigenicity of Pancreatic Cancer
99
cytotoxic effects (19)) highlights the unique usefulness of this model for studying antigenicity of pancreatic cancer, particularly for developing diagnostic and therapeutic methods. ACKNOWLEDGMENTS
This study was supported in part by NIH laboratory core grant No. CA36727, NIH grant 1R29 CA47980, and by American Cancer Society grant SIG-16. WGC is a recipient of an American Cancer Society Junior Faculty Research Award. REFERENCES 1 Longnecker DS, Wiebkin P, Schaeffer BK, and Roebuck BD. Experimental carcinogenesis in the pancreas. Int. Rev. Exp. Pathol. 1984; 26: 177-229. 2 Pour PM and Wilson RB. Experimental pancreas tumors. Cancer of the Pancreas. Moosa AR, ed., Baltimore: Williams and Wilkins Co. 1980; 37-158. 3 Pour PM, Runge RG, Birt D, Gingell R, Lawson T, Nagel D, Wallcave L, and Salmasi SZ. Current knowledge of pancreatic carcinogenesis in the hamster and its relevance to the human disease. Cancer 1981; 47: 1573-1587. 4 Takahashi M, Runge R, Donnelly T, and Pour PM. The morphologic and biologic patterns of chemically induced pancreatic adenocarcinoma in Syrian golden hamsters after homologous transplantation. Cancer Lett. 1979; 7: 127-133. 5 Egami H, Takiyama Y, Cano M, Houser WH, and Pour PM. Establishment of hamster pancreatic ductal carcinoma cell line (PC-l) producing blood group-related antigens. Carcinogenesis 1989; 10: 861-869. 6 Metzgar RS, Gaillard MT, Levin SJ, Tuck FL, Bosen EH, and Borowitz MJ. Antigens of human pancreatic adenocarcinoma defined by murine monoclonal antibodies. Cancer Res. 1982; 42: 601-608. 7 Metzgar RS, Rodriguez N, Finn O J, Lan MS, Dasch DS, and Seigler HF. Detection of a pancreatic cancer-associated antigen (DU-PAN-2 antigen) in serum and ascites of patients with adenocarcinoma. Proc. Natl. Acad. Sci. USA 1984; 81: 5242-5246. 8 Colcher D, Horan H, Nuti M, and Schlom J. A spectrum of monoclonal antibodies reactive with mammary tumor cells. Proc. Natl. Acad. Sci. USA 1981; 78: 3199-3203. 9 Pour PM, Tempero MA, Takasaki H, Uchida E, Takiyama Y, Burnett DA, and Steplewski Z. Expression of blood group-related antigens A, B, H, Lewis A, Lewis B, Lewis X, Lewis Y, and CAt9-9 in pancreatic cancer cells in comparison with the patient's blood group type. Cancer Res. 1988; 48: 5422-5426. 10 Takiyama H, Tempero MA, Takasaki H, Onda M, Tsuchiya R, Buchler M, Ness M, Colcher D, Schlom J, and Pour PM. Reactivity of CO t7-1A and B72.3 in benign and malignant pancreatic disease. Human Pathol. 1989; 20: 832-838. i1 Molday RS. Labeling of cell surface antigens for SEMo Immunocytochem. 1983; 2: 117-150. 12 Takiyama Y, Egami H, Pour PM. Blood group antigen expression in developing pancreas and in induced pancreatic cancer cells in Syrian hamsters. Carcinogenesis 1990 (in press). 13 Massague J and Czech MP. The subunit structures of two distinct receptors for insulinlike growth factors I and II and their relationship to the insulin receptor~ J. Biol. Chem. 1982; 257: 5038-5045. 14 Smith PK, Krohn RI, Hermanson GT, Mallia AK, Gartner FH, Provenzano MD, Fujimoto EK, Goeke NM, Olson BJ, and Klenk DC. Measurement of protein using bicinchoninic acid. Anal. Biochem. 1985; 150: 76-85.
100
Egami et alo
15 Laemmli UK. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970; 227: 680-685. 16 Mallison G, Martin PG, Anstee DJ, Tanner MJA, Meris AH, Tills D, and Sonneborn HH. Identification and partial characterization of the human erythrocyte membrane components that express the antigens of the LW blood-group system. Biochem. J. 1986; 234: 649-652. 17 Itzkowitz SH, Yuan M, Montgomery CK, Kjeldsen T, Takahashi HK, Bigbee WL, and Kim YS. Expression of Tn, sialosyl-Tn, and T antigens in human colon cancer. Cancer Res. 1989; 49: 197-204. 18 Kjeldsen T, Clausen H, Hirohashi S, Ogawa T, Iijima H, Hakomori S. Preparation and characterization of monoclonal antibodies directed to the tumor-associated O-linked sialosyt-2-6-N-acetylgalactosaminyl(sialosyl-Tn) epitope. Cancer Res. 1988; 48: 2214-2220. 19 Blottiere HM, Maurel C, and Douillard JY. Immune function of patients with gastrointestinal carcinoma after treatment with multiple infusions of rnonoclonal antibody 17-1A. Cancer Res. 1987; 47: 5238-5241.
