~ SURGERYTODAY

Surgery Today Jpn J Surg (1993) 23:908-916

© Springer-Verlag 1993

The Relationship Between the Expression of Blood Group-Related Antigens and the Cell Proliferation of Pancreatic Carcinomas Induced by N-nitrosobis(2-oxopropyl)amine in Hamsters TADASHIKOBAYASHI,L2 EIJI UCHIDA,1 KOICHITAMURA,2 and NOBUAKIYAMANAKA2 1The FirstDepartmentof Surgeryand 2the FirstDepartmentof Pathology,NipponMedicalSchool,1-1-5Sendagi,Bunkyo-ku,Tokyo, 113Japan

Abstract: The relationship between expression of blood grouprelated antigens (BGRAs) A, B, H, and cell proliferation was investigated during pancreatic carcinogenesis and in transplanted pancreatic carcinomas induced by N-nitrosobis(2oxopropyl)amine in hamsters. Immunohistochemical staining was performed using monoclonal antibodies (MoAbs) against BGRAs A, B, H, and bromodeoxyuridine (BrdU). The labeling index (LI) was determined as the ratio of the number of BrdUlabeled cells to the total number of cells counted in each lesion. During carcinogenesis, the LI was observed to increase in line with the increase in the extent of atypism (P < 0.01). The mean LIs of the hyperplasia, atypical hyperplasia, and carcinoma were 0.32, 3.21, and 10.2, respectively. The mean LIs of transplanted carcinomas were higher than those of the original carcinomas (P < 0.01). The reactivity with each of the antibodies was determined using an arbitrary scoring system. Staining with MoAbs A and B (staining intensity; 1+ to 3+) appeared to be more intense than that with MoAb H (1+ to 2+) during carcinogenesis. Regarding the growth rate, which was very high, in the transplanted carcinomas, MoAb A reacted with all the cancer cells (4+), whereas, MoAbs B and H reacted with fewer cells (1+ to 3+). These results indicate that the A antigen in particular is associated with the cell proliferation of pancreas, especially with carcinoma induced in hamsters. Key Words: pancreatic carcinoma, blood group-related an-

tigens, bromodeoxyuridine, N-nitrosobis(2-oxopropyl)amine, Syrian golden hamster

Introduction

Pancreatic carcinomas induced by N-nitrosobis(2oxopropyl)amine (BOP) in Syrian golden hamster are of considerable interest because they are ductal Reprint requests to: T. Kobayashi (Received for publication on July 14, 1992; accepted on Jan. 8, 1993)

adenocarcinomas closely resembling those that constitute the most common malignant pancreatic neoplasms in man. 1-4 These induced cancers express some of the same antigens as the human tumors, such as blood group-related antigens (BGRAs). 5'6 The BGRAs of the ABO(H) system and the Lewis system are the major antigens in man and they are found in various epithelial cells as well as in blood cells. 7 As most cancers originate from epithelial cells, changes in BGRAs are of particular interest in tumor immunology) The BGRAs are a family of genetically and biosynthetically related carbohydrates that serve as allogenic, developmental and differentiation markers, the expression of which is under genetic control. 9 The BGRAs, in particular the A and B antigens, have been reported to be tumor-specific antigens of pancreatic carcinomas in hamsters. 5'1° It is well known that epidermal growth factor (EGF) and its receptor act biologically to stimulate cell proliferation in epidermal and epithelial tissues, la'lz The EGF receptor from a human vulval carcinoma cell line, A431, carries a variety of unusual carbohydrate sequences that are recognized by antibodies to BGRAs 13-a6 and there is a relationship between the epitopes of this receptor and BGRAs. a3,a7 However, the correlation between BGRAs and cell proliferation has not been investigated in vivo. In this study, a cell kinetic analysis using bromodeoxyuridine (BrdU) was performed in order to study cell proliferation in pancreatic carcinomas. The relationship between BGRAs and cell proliferation was investigated during pancreatic carcinogenesis and in transplanted pancreatic carcinomas induced by BOP in hamsters, using monoclonal antibodies (MoAbs) against BGRAs A, B, H and BrdU.

T. Kobayashi et al.: Pancreatic Carcinogenesis Materials and Methods

Animals A total of 104, 6-week-old, female Syrian golden hamsters (Shizuoka Laboratory Animal Center, Shizuoka), weighing about 100 g, were used. They were housed, five per plastic cage, under standardized conditions (temperature, 20 _+ 5°C; humidity, 40 _+ 10% ; light/dark cycle, 12h/12h) and given a commercially available basal diet, Oriental MF (Oriental Yeast, Tokyo) and water ad libitum.

Carcinogen N-nitrosobis(2-oxopropyl)amine (BOP; Iwaikagaku, Tokyo) was dissolved in 0.9% NaC1 at a concentration of 10 mg/ml immediately prior to injection.

Experimental Design Pancreatic Carcinogenesis Thirty-three hamsters received subcutaneous injections of BOP (10 mg/kg body weight) in 0.9% NaC1 solution once a week for up to 12 weeks. Three animals were sacrificed serially every 2 weeks from 5 to 25 weeks after initiation of BOP treatment. Autopsies were performed under ether anesthesia. The pancreatic tissues were removed from these hamsters and from untreated, age-matched (control) animals.

Transplanted Pancreatic Carcinomas Pancreatic carcinomas induced by BOP, which were taken from separate hamsters from the group described above, were minced with scissors and 1-mm cubes were transplanted subcutaneously using a trocar into the interscapular area of untreated hamsters. Before the transplanted tumors underwent central necrosis, the recipient hamsters were sacrificed 6 - 8 weeks after transplantation and a part of the tumor tissue was transplanted serially. Serial transplantations were performed in the same manner. The larger and smaller diameters of transplanted tumors were measured with calipers twice a week. The volume (V) of the tumor was calculated from the formula: V = (L x W2)/2, where L is the larger diameters and W is the smaller one, both in mm. Data on exponentially growing tumors were analyzed by the linear regression technique and the tumor doubling time (DT) was calculated by dividing the log 2 by the slope of the regression line. ~s

909 immunohistochemically and the labeling index (LI) was determined as the ratio of the number of BrdU-labeled cells to the total number of cells (more than 2,000 cells) counted in each lesion.

Histological Examination The pancreatic tissue specimens were removed and examined histologically. The tissues were fixed in 10% formalin, processed routinely and paraffin-embedded sections were processed for hematoxylin and eosin staining and immunohistochemistry. The pancreatic lesions were diagnosed according to the criteria described by Pour and Wilson. 3 The histological typing was compared with the antigen expression pattern of each lesion by examining immunostained serial sections under a light microscope. The cellular staining pattern was classified as Golgi, cytoplasmic granular, baso-lateral membrane, or luminal. 1° The reactivity with each of the antibodies in each lesion was determined using an arbitrary scoring system, in which the number of positively stained cells was estimated as follows: 0% ( - ) ; Positive cells _< 10% (1+); 10% < Positive cells _< 50% (2+); 50% < Positive cells _< 90% (3+); and 90% < Positive cells _< 100% (4+). In each lesion, the number of positive cells out of more than 2,000 cells were counted.

Immunohistochemical Procedures The anti-A, -B and -H (Dako, Santa Barbara, Calif.) and anti-BrdU (Becton Dickinson Immnocytometry System, Calif.) MoAbs were obtained commercially. Immunohistochemical staining was performed using a Vectastain ABC kit (Vector Laboratories, Burlingame, Calif.). The dilutions used were 1 : 100 for the anti-A, B, -H and 1:200 for the anti-BrdU MoAbs. For the negative control studies, the tissue specimens were processed similarly, except that mouse IgM was used instead of the primary antibodies.

Statistics The results were expressed as means _+ standard error (SE) and a statistical analysis was performed using Student's t-test. Differences at P < 0.05 were considered to be significant. Results

Cell Kinetics

Morphology

One hour before sacrifice, the hamsters were given an intraperitoneal injection of BrdU (40mg/kg body weight). The BrdU in the pancreatic tissues was detected

Focal hyperplasia of the ductal epithelium appeared during the 9th week of pancreatic carcinogenesis, and frequently showed mitotic activity by the 13th week.

910

T. Kobayashi et al.: Pancreatic Carcinogenesis

2.0

epithelium, hyperplasia, atypical hyperplasia, and carcinoma were 108, 86, 103, 126, and 92, respectively. The growth curves for the two types of transplanted tumors (No. 1 and No. 2) are shown in Fig. 1. The transplanted tumors grew rapidly after a latent period of about 7 days and two types of homologous transplanted pancreatic carcinomas were eventually obtained, of which one was passed successfully for 10 generations and the other for 5. The tumors grew as infiltrative nodules that invaded adjacent muscles and ulcerated the overlying skin, whereas, no metastases were found in the host hamsters. The transplanted tumors were moderately to well differentiated adenocarcinomas that were histologically identical to the original tumors and numerous mitoses were present. No significant histological changes among the passages were observed.

t 5 th gen, I I | st gen.

~

/ /llOthgen. iI

2l

/ / / 5 th gen.

I S / / 1 st gen..

~

III

,,',,;///

E~ 1.0

/,~'// z~z

/.~j 0

Cell Kinetics

7 14 21 28 35 42 49 56 Days after Transplantation

Fig. 1. Growth curves for the two types of transplanted tumors (No. 1 [solid line] and No. 2 [dashed line]), gem, Generation

By the 15th week, more progressive ductal changes had occurred and focal atypia was a common finding, multiple adenocarcinomas were observed during the 17th week, and during the 19th week the number of proliferating ductules with atypia had increased further. Isolated adenocarcinomas of various sizes were found by the 21st week and during the period from weeks 23 to 25, carcinomas developed, which were moderately to well differentiated adenocarcinomas with associated fibrous stroma. The number of lesions that were evaluated from the normal epithelium, non-atypical

The mean BrdU-LIs of the lesions during pancreatic carcinogenesis and the mean BrdU-LIs and DTs of the transplanted pancreatic carcinomas are shown in Table 1. When hamsters were sacrificed serially every 2 weeks from 5 to 25 weeks after initiation of BOP treatment, the mean LI values of the lesions (hyperplasia, atypical hyperplasia, and carcinoma) obtained during each period were coincidental. There were no significant differences between the mean LI of the normal epithelium in the control specimens and that of the non-atypical epithelium in the specimens from the BOP-treated hamsters. The mean LIs of the hyperplastic, atypical hyperplastic, and carcinoma tissue specimens were 0.32, 3.21, and 10.2, respectively. The LI increased as the degree of atypism increased and significant differences were found between the mean LIs of the non-atypical epithelium and hyperplasia (P < 0.05), those of hyperplasia and atypical

Table 1. The labeling indexes (LIs) of the lesions during pancreatic carcinogenesis and the LIs and doubling times (DTs) of the transplanted pancreatic carcinomas LI DT days Control (108 lesions) Normal epithelium Pancreatic carcinogenesis Non-atypical epithelium (86 lesions) (103 lesions) Hyperplasia (126 lesions) Atypical hyperplasia (92 lesions) Carcinoma Transplanted pancreatic carcinomas 1st generation 5th 10th 1st 5th

0.24 + 0.23 + 0.32 + 3.21 + 10.2 +

0'22a 1 NS 0.12 ~b o.12 ~c 1.19 3o 4.78 ~

18.7 + 2.53 26.4 _ 3.01 23.1 _+5.91 15.4 + 3.44 22.5 _+4.11

amean _+SE; up < 0.05; cp < 0.01; NS, not significant

c

5.11 +_ 1.5 4.60 + 1.9 4.65 _+2.6 4.56 + 2.3 3.95 --+3.3

T. Kobayashi et al.: Pancreatic Carcinogenesis

911

ig. 2a,b. Immunohistochemical staining ith anti-bromodeoxyuridine monoclonal ntibodies, a original carcinoma b transranted carcinoma. (H&E, x134)

hyperplasia (P < 0.01), and those of atypical hyperplasia and carcinoma (P < 0.01). The mean LI of the transplanted pancreatic carcinomas increased gradually, from 18.7 to 26.4, until the 5th generation after which it decreased and remained constant up to the 10th generation. A similar tendency was observed for the other type of transplanted tumor until the 5th generation. The distribution of the positive cells was not uniform in each tumor: The LI tended to be high in the peripheries and low in the central areas. The mean DTs of the volumes of the two types of transplanted tumors (No. 1 and No. 2) were 4.60-5.11 and 3.95-4.56 days, respectively. The mean LIs of both types of transplanted carcinoma were higher than those of the original carcinomas (P < 0.01 for both transplants, Fig. 2a,b).

Expression of BGRAs The MoAb expression levels of the lesions during pancreatic carcinogenesis and of the transplanted pancreatic carcinomas are summarized in Table 2. None of the MoAbs reacted with normal ductal epithelium in the control specimens. Whereas, they reacted with nonatypical epithelium in the specimens from BOP-treated animals with a score of 1+. The MoAbs A and B reacted with the hyperplastic, atypical hyperplastic and carcinoma tissues with scores of 3+, although MoAb H reacted with comparatively fewer cells (1+ to 2+). No obvious histological differences between the stained and unstained cells were observed. The co-expression of MoAbs A and B or A, B, and H occurred in each lesion. The main binding site for all the MoAbs was the

912

T. Kobayashi et al.: Pancreatic Carcinogenesis

Table 2. The monoclonal antibody (MoAb)expression levelsa

of the lesions during pancreatic carcinogenesis and of the transplanted pancreatic carcinomas MoAbs A B H Control Normal epithelium (108 lesions) Pancreatic carcinogenesis Non-atypical epithelium (86 lesions) Hyperplasia (103 lesions) Atypical hyperplasia (126 lesions) Carcinoma (92 lesions) Transplanted pancreatic carcinomas 1st generation 5th 10th 1st 5th

--

--

--

1+ 3+ 3+ 3+

1+ 3+ 3+ 3+

1+ 1+ 2+ 2+

4+ 4+ 4+

2+ 3+ 1+

2+ 3+ 1+

4+ 4+

2+ 1+

1+ 2+

aIn each lesion, the number of positivelystainedcells was estimated to be as follows: 0% (-); positive cells -< 10% (1+); 10% < positive cells _<50% (2+); 50% < positivecells --<90% (3+); 90% < positivecells --<100% (4+)

luminal cell surface and occasionally was of a diffuse cytoplasmic nature (Fig. 3a-c). The reactivity was independent of the type of MoAb, although the staining with MoAbs A and B appeared to be more intense than that with MoAb H during pancreatic carcinogenesis. The MoAb A reacted with all the transplanted pancreatic carcinomas with a score of 4+. However, MoAbs B and H reacted with fewer cells (1+ to 3+). The MoAb A bound to the entire cytoplasm in a diffuse pattern. Whereas MoAbs B and H bound mainly to the luminal cell surface and only occasionally to the diffuse cytoplasm (Fig. 4a-c). The reactivity in the transplanted carcinomas was independent of the type of MoAb.

Discussion

In this study, the relationship between BGRAs and cell proliferation was investigated during pancreatic carcinogenesis and in transplanted pancreatic carcinomas induced by BOP in hamsters. We observed that the sequential changes of the pancreatic ductal epithelium during pancreatic carcinogenesis were hyperplasia, atypical hyperplasia and carcinoma and that no histological changes among passages were observed in transplanted pancreatic carcinomas, findings which are similar to those described previously.3,19-2~ We observed that the BrdU-LI increased in proportion to the increased extent of atypism during pancreatic carcinogenesis. Statistically significant differences were found between the LIs of the non-atypical epithelium

and hyperplasia, of hyperplasia and atypical hyperplasia and of atypical hyperplasia and carcinoma. These results indicate that the incidence of atypical cells in the S phase of the cell cycle increased as atypism increased and that the growth rate of atypical cells became high during pancreatic carcinogenesis. These results using BrdU are consistent with those reported by Levitt et al. 22 on the incorporation of 3H-thymidine into DNA of pancreatic ductal cells induced by the carcinogen, N-nitrosobis(2-hydroxypropyl)amine (DIPN), in hamsters. We observed that the mean LI of transplanted carcinomas was higher than that of the original carcinomas, which means that the incidence of tumor cells in the S phase was high in the transplanted carcinomas and also indicates that the growth rates of the transplanted pancreatic carcinomas are higher than those of the original carcinomas. 2°'21'23 Transplanted carcinomas may grow more rapidly as a result of the cell selection processes which occurs during tumor progression. 24'z5 Pancreatic carcinomas in hamsters express A, B, and H antigens5'6 and appear heterogeneically to be like human pancreatic carcinomas. 9'26 In this study, none of the MoAbs reacted with normal epithelium of control specimens. Expression of A, B, and H antigens was observed during pancreatic carcinogenesis and, moreover, the staining with MoAb A and B appeared to be more intense than that with MoAb H. Therefore, the expression of the A and B antigens differed significantly from that of the H antigen during pancreatic carcinogenesis, which indicates that the A and B antigens are tumor-specific antigens in hamsters. Takiyama et al. 1° detected A and B antigens in fetal, newborn and adult duodenal epithelia, from which the pancreatic tissue arises in hamsters. The A and B antigens are tumorrelated antigens in hamsters and their expression may represent the activation of a silent gene during carcinogenesis. 1° Tomioka et al. 27 reported that the A antigen appears to discriminate between the benign and malignant lesions better than the other BGRAs. The results of our study agree with those of these studies. We found that the A antigen was localized mainly on the cell membrane of the cancer cells during pancreatic carcinogenesis. Glycoproteins with blood group A specificity were observed, using sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting procedures, in the major component of the membrane fraction of a pancreatic cell line, which had a molecular mass of 120kDa. 28 However, marked differences between the expression of A and B, H antigens in transplanted carcinomas were observed. The MoAb A reacted with virtually all the transplanted carcinoma cells, whereas, MoAbs B and H reacted with fewer cancer cells. The binding site for MoAb A was the entire diffuse cytoplasm, but MoAbs B and H

T. Kobayashi et al.: Pancreatic Carcinogenesis

913

Fig. 3a-c. Photomicrographs show the immunohistochemical staining with antiA, B, and H monoclonal antibodies (Mo.Abs)in serial specimens of the original carcinoma, a Strong reactivity with antiA MoAb is evident showing a luminal pattern, b Strong reactivity with anti-B MoAb is evident showing a luminal pattern, c Weak reactivity with anti-H MoAb is evident showing a luminal pattern. (H&E, ×134)

Fig. 4a-c. Photomicrographs show the immunohistochemical staining with antiA, B, and H monoclonal antibodies (MoAbs) in serial specimens in transplanted carcinoma, a Strong reactivity with anti-A MoAb is evident showing a diffuse cytoplasmic pattern, b Weak reactivity with anti-B MoAb is evident showing a luminal pattern, c Weak reactivity with anti-H MoAb is evident showing a luminal pattern. (H&E, × 134) 914

T. Kobayashi et al.: Pancreatic Carcinogenesis bound mainly to the luminal cell surface. The expression of the A antigen is extremely strong in transplanted carcinomas, which have a very high growth rate. The antigenic determinants of B G R A s are composed of carbohydrate chains, which are formed by the addition of specific monosaccharides onto oligosaccharide backbones, and are carried on mucin as well as the glycoproteins and glycolipids of cell membranes.29 The A B H antigens are formed on Gal 1-3GlcNAc (type I) and Gal 1-4GlcNAc (type II) backbone structures (Gal: galactose, GlcNAc: N-acetylglucosamine). The H antigen is formed by the addition of fucose to the terminal galactose residues of the backbones, whereas the A and B antigens are formed by the addition of Nacetylgalactosamine or galactose, respectively, to the galactose residue of the H antigen. 29 The results of our study thus suggest that the subtle structural changes of B G R A s may alter their binding ability in pancreatic carcinomas in hamsters. The carbohydrate structure of the A antigen has been reported to be associated with the E G F receptor. 17,30 It is widely accepted that the E G F receptor plays an important role in the transduction of signals mediated by this growth factor, la After binding to its receptor, E G F induces a wide range of early and delayed biological responses that lead to D N A synthesis in the target cell. 11 The E G F receptor from a human vulval carcinoma cell line, A431, carries a variety of unusual carbohydrate sequences that are recognized by antibodies to B G R A s a3-16 and the A antigen has been reported to be expressed strongly on A431 cells) ° The mouse hybridoma antibody TL5, which precipitates the E G F receptor of the A431 cells, recognizes the carbohydrate structure of the A antigen. 31 This finding was confirmed in studies of the binding of the antibody to glycoproteins, the hemagglutination of erythrocytes with known B G R A activities and the inhibition of binding of the antibody to a radiolabeled A antigen-active glycoprotein by structurally defined oligosaccharides. 3° It is likely, therefore, that the structure of the A antigen is present as a terminal sequence on the carbohydrate chains of the glycoprotein receptor for E G F . 3° The results of this study show that the reactivity of A and B antigens increases in proportion to the increase in the growth rate of atypical cells during pancreatic carcinogenesis. Moreover, the A antigen is expressed in all cancer cells in transplanted pancreatic carcinomas, in which the growth rate is very high. Therefore, these results indicate that the A antigen, in particular, is associated with the cell proliferation of pancreas, especially with carcinoma induced in hamsters.

Acknowledgments. We wish to thank Prof. Masahiko Onda for his valuable criticism and help in preparing this manuscript.

915

References 1. Pour P, Althoff J, Kruger FW, Mohr U (1977) A potent pancreatic carcinogen in Syrian hamsters: N-nitrosobis(2-oxopropyl)amine. J Natl Cancer Inst 58:1449-1453 2. Pour PM, Runge RG, Birt D, Gingell R, Lawson T, Nagel D, Wallcave L, Salmasi SZ (1981) Current knowledge of pancreatic carcinogenesis in the hamster and its relevance to the human disease. Cancer 47:1573-1587 3. Pour PM, Wilson RB (1980) Experimental tumors of the pancreas. In: Moosa AR (ed) Tumors of the pancreas. Williams and Wilkins, Baltimore, pp 37-158 4. Takahashi M, Pour P, Althoff J, Donnelly T (1977) Sequential alteration of the pancreas during carcinogenesisin Syrianhamsters by N-nitrosobis(2-oxopropyl)amine.Cancer Res 37:4602-4607 5. Pour PM, Uchida E, Burnett DA, Steplewski Z (1986) Bloodgroup antigen expression during pancreatic cancer induction in hamsters. Int J Pancreatol 1:327-340 6. Pour PM (1988) Cell differentiation during pancreatic carcinogenesis. Scand J Gastroenterol 23[Suppl 151]:123-130 7. Glynn LE, Holborow EJ (1959) Distribution of blood-group substances in human tissue. Br Med Bull 15:150-153 8. Hakomori S, Kannagi R (1983) Glycosphingolipids as tumorassociated and differentiation markers. J Natl Cancer Inst 71: 231-251 9. Coon JS, Weinstein RS (1986) Blood group-related antigens as markers of malignant potential and heterogeneity in human carcinomas. Human Pathol 17:1089-1106 10. Takiyama Y, Egami H, Pour PM (1990) Blood group antigen expression in developing pancreas and in induced pancreatic cancer cells in Syrian hamsters. Carcinogenesis 11:577-582 11. Carpenter G, Cohen S (1979) Epidermal growth factor. Ann Rev Biochem 48:193-216 12. Cohen S, Carpenter G, King L Jr (1980) Epidermal growth factor-receptor-protein kinase interactions. J Biol Chem 255: 4834-4842 13. Gooi HC, Hounsell EF, Lax I, Kris RM, Libermann TA, Schlessinger J, Sato JD, Kawamoto T, Mendelsohn J, Feizi T (1985) The carbohydrate specificitiesof the monoclonalantibodies 29.1,455 and 3C1B12 to the epidermal growth factor receptor of A431 cells. Biosci Rep 5:83-94 14. Gooi HC, Picard JK, Hounsell EF, Gregoriou M, Rees AR, Feizi T (1985) Monoclonal antibody (EGF/G49) reactive with the epidermal growth factor receptor of A431 cells recognizes the blood group ALeu and ALeystructures. Mol Immuno122:689-693 15. Parker PJ, Young S, Gullick WJ, Mayes ELV, Bannett P, Waterfield MD (1984) Monoclonal antibodies against the human epidermal growth factor receptor from A431 cells. J Biol Chem 259:9906-9912 16. Pendu JL, Fredman P, Richter ND, Magnani JL, Willingham MC, Pastan I, Oriol R, Ginsberg V (1985) Monoclonal antibody 101 that precipitates the glycoprotein receptor for epidermal growth factor is directed against the Y antigen, not the H type 1 antigen. Carbohydr Res 141:347-349 17. Basu A, Murthy U, Rodeck U, Herlyn M, Mattes L, Das M (1987) Presence of tumor-associated antigens in epidermal growth factor receptors. Cancer Res 47:2531-2536 18. Isaac JT, Coffey DS (1981) Adaptation versus selection as the mechanism responsible for the relapse of prostatic cancer to androgen ablation therapy as studied in the Dunning R-3327-H adenocarcinoma. Cancer Res 41:5070-5075 19. Kozuka S, Sassa R, Taki T, Masamoto K, Nagasawa S, Saga S, Hasegawa K, Takeuchi M (1979) Relation of pancreatic duct hyperplasia to carcinoma. Cancer 43:1418-1428 20. Mattern J, Haag D, Wayss K, Volm M (1981) Growth kinetics of human lung tumors in nude mice. lExp Cell Biol 49:34-40 21. Kubota T, Nakada M, Tsuyuki K, Inada T, Asanuma F, Ishibiki K, Abe O (1986) Cell kinetics and chemosensitivity of human

916

22. 23. 24. 25. 26.

27.

T. Kobayashi et al.: Pancreatic Carcinogenesis carcinomas serially transplanted into nude mice. Jpn J Cancer Res (Gann) 77:502-507 Levitt MH, Harris CC, Squire R, Springer S, Wenk M, Mollelo C, Thomas D, Kingsbury E, Newkirk C (1977) Experimental pancreatic carcinogenesis. Am J Pathol 88:5-28 Ikeuchi S, Shimosato Y, Kameya T, Watanabe S, Kakegawa T, Abe O (1980) Cell kinetic study of human cancers transplanted to nude mice. Exp Cell Biol 48:218-228 Noble RL (1977) Hormonal control.of growth and progression in tumors of Nb rats and a theory of action. Cancer Res 37:82-94 Shimosato Y, Kameya T, Nagai K, Hirohashi S, Koide T, Hayashi H, Nomura T (1976) Transplantation of human tumors in nude mice. J Natl Cancer Inst 56:1251-1260 Nakasaki H, Mitomi T, Noto T, Ogoshi K, Hanaue H, Tanaka Y, Makuuchi H, Clausen H, Hakomori S (1989) Mosaicism in the expression of tumor-associated carbohydrate antigens in human colonic and gastric cancers. Cancer Res 49:3662-3669 Tomioka T, Fujii H, Egami H, Takiyama Y, Pour PM (1991)

28.

29. 30.

31.

Correlation between morphology and blood group-related antigen expression in pancreatic tumors induced in Syrian hamsters. Carcinogenesis 12:441-447 Egami H, Chaney WG, Takiyama Y, Pour PM (1991) Subcellular localization of blood group A substance produced by pancreatic adenocarcinoma induced in hamsters by N-nitrosobis(2oxopropyl)amine (BOP) and by its cell line (PC-l). Carcinogenesis 12:509-514 Watkins WM (1980) Biochemistry and genetics of the ABO, Lewis and blood group systems. Adv Hum Genet 10:379385 Gooi HC, Schlessinger J, Lax I, Yarden Y, Libermann TA, Feizi T (1983) Monoclonal antibody reactive with the human epidermal growth factor receptor recognizes the blood group A antigen. Biosci Rep 3:1045-1052 Schreiber AB, Libermann TA, Lax I, Yarden Y, Schlessinger J (1983) Biological role of epidermal growth factor-receptor clustering. J Biol Chem 258:846-583