Oncology 1990;47:149-154
< 1990 S. Kargcr AG. Basel 0030 2414 90 0472 0149 S 2.75 0
Murine Monoclonal Antibodies to Human Pancreatic Cancer; Specificity and Sensitivity James G. Jakowatz a>
b ,
Karen Gand, Thomas R. UlicIT, Alan G. lVileb
a Surgical Service, Veterans Administration Medical Center, Long Beach, Calif.; b Division of Surgical Oncology and c Department of Pathology, University of California Medical Center. Orange, Calif.; d University of California/California College of Medicine, Irvine, Calif, USA
Key Words. Monoclonal antibodies • Pancreatic cancer Abstract. Pancreatic carcinoma (n = 7), pancreatitis tissue (n = 4), normal pancreas tissue (n = 5), colonic adenocarcinoma (n = 4) and in vitro human pancreatic cancer cell lines (n = 6) were studied with the murine monoclonal antibodies (MAbs) 3DS2A, AR1-28, AR2-20, Cal9-9 and CA17-1A to determine their immunohistologic specificity and sensitivity for use as radiolabeled diagnostic imaging agents. Using the avidinbiotin-immunoperoxidase staining technique, MAbs 3DS2A and AR1-28 stained 86 and 100% of pancreatic cancer specimens, respectively. MAbs 3DS2A and AR1- 28 are suitable agents for use as radiolabeled diagnostic imaging agents in patients with pancreatic cancer.
Adenocarcinoma of the pancreas can be difficult to diagnose preoperatively. Radiologic modalities are nonspecific, rendering an anatomic, rather than a his tologic, abnormality [1-4]. Radiologically guided needle aspirations are unsuccessful in as many as 50% of cases. Further, these procedures are associated with morbidity, tissue sampling error and reliance upon the cytologist’s experience and skill [5-9], Intraoperative biopsies are also associated with morbidity and a 50-75% false-positive biopsy rate due to sampling error [8, 10, 11], In most cases of intraoperative exami nation, the sampling error prevents the discrimination between pancreatic cancer and pancreatitis. Often the surgeon is left with a decision to perform a radical operation based on clinical judgement. Potentially, this clinical dilemma could be solved using monoclonal antibodies (MAbs) developed from tumor-associated antigens. MAbs for other tumors have been radiolabeled with 131I o r 11 'In and used for diagnostic imaging in patients with various malignan cies [12-15]. Murine MAbs directed against pancreatic adenocarcinoma have been developed by other inves tigators, from either human pancreatic or colon can-
cer cell lines [16-20]. Their potential use as imaging agents in pancreatic cancer has not been explored. The purpose of this study was to investigate the specificity and sensitivity of pancreatic-cancer-associated MAbs in histologic specimens of pancreatic cancer and hu man pancreatic cancer cell lines. These studies deter mined the most sensitive and specific Mab to use for radioimaging experiments in athymic mice bearing human pancreatic cancer cell line xenografts.
Materials and Methods Five MAbs, reactive with human pancreatic adenocarcinoma, were obtained from two investigators. CAI9-9 and 3DS2A are murine IgG MAbs of hybridomas whose antigenic source was the human colon cancer cell line SW1116. Both were obtained as ascites and diluted 1 : 1500 and 1 : 250, respectively, in phosphate-buffered saline (PBS). pH 7.4, for immunoperoxidase staining. CA17-1A is a murine IgG MAb of a hybridoma developed from the human colon cancer cell line SW1083. CAI7-1A, obtained as ascites, was affinitypurified on a Staph-protein A column (Affigcl. Bio-Rad, Richmond. Calif., USA). No reactivity with any pancreatic tissue was obtained with CA17-1A in the crude ascites form. The purified MAb was diluted at 1 : 50 in PBS. pH 7.4. AR1-28 and AR2-20 arc murine IgG MAbs whose hybridoma origins stem from the human pancre atic cancer cell line R WP-1. Again the crude ascites forms were used
Downloaded by: Karolinska Institutet, University Library 130.237.122.245 - 1/18/2019 1:54:02 AM
Introduction
150
Slides were evaluated by light microscopy for the presence and intensity of stain. Intensity of stain was graded as intense (3 + ), moderate (2 + ), weak (1 + ), trace (+ / —) and negative ( —).
Results Table 1 summarizes the results of the immunoreactivity of the MAbs for the tissue specimens. MAb CA19-9 intensely stained all pancreatic car cinomas but also stained interlobular ducts, in tralobular ducts and centroacinar cells of normal pan creatic tissue (whether adjacent to pancreatitis, pan creatic cancer or in pancreatic tissue from normal patients). Islet and acinar cells did not stain. MAb CA17-1A stained pancreatic tissue, car cinoma and non-neoplastic pancreatig structures with equal intensity. Three pancreatic cancer specimens were strongly positive, one was moderately positive and the remaining samples showed only trace or weak levels of reactivity. MAb CA17-1A stained all nonneoplastic elements of pancreatitis tissue with weak to moderate reactivity in variable fashion.
Tabic 1. Staining intensity of the MAbs for each tissue Tissue CA19-9 3DS2A AR1-■28 AR2-20 CAI7-1A Nonimmune mouse IgG Normal pancreas 1 3+ 2 3+ +/3 3+ +/4 3+ + /5 3+ +/Pancreatitis 1 3+ 2 3+ 3 3+ 4 3+
1+ 1+ 1+ + /-
Pancreatic cancer 1 3+ 1+ 2 3+ 3+ 3 3+ 2+ 4 3+ 3+ 5 3+ 3+ 6 3+ 3+ 7 3+ 3+
+ /+ /+ /-
1+ 1+ 1+ 1+
+ /-
1+ 2+ 2+
2+
+ /-
1+ 1+ 2+ 1+
2+ 3+ 3+ 2+ 3+ 3+ 2+
2+ 3+ 2+ 3+ 3+ 3+ 3+
3+
1+ 1+ 1+
3 + = Intense; 2 + == moderate; 1 + — = negative.
+ /-
2+ + /-
+ /-
3+ 3+ 2+ -
-
1+ - weak; + / — = trace;
Downloaded by: Karolinska Institutet, University Library 130.237.122.245 - 1/18/2019 1:54:02 AM
and diluted in PBS, pH 7.4, at 1 : 2000 and 1 : 1250. respectively. All dilutions were determined by immunoperoxidase checkerboard ti tration methods against pancreatic cancer tissue. Paraffin-embedded sections of human pancreatic cancer, pan creatitis tissue, normal pancreatic tissue and human colonic cancer obtained at the time of operation were utilized as target tissues. Surgical specimens were chosen over autopsy specimens to mini mize the associated antigen loss as a result of autodigestion. Pancreatic cancer specimens were obtained from resected pan creatic glands involved with primary adenocarcinoma. Normal pancreas tissue was obtained from glands resected for trauma where no evidence of malignancy existed. Pancreatitis tissue was obtained from biopsied or resected pancreas of patients with chronic pan creatitis. Colonic cancer samples were obtained from colectomy specimens containing both normal and malignant tissue. All paraf fin-embedded tissue was cut into 4-pm sections and applied to slides for immunoperoxidase staining. The human pancreatic cancer cell lines T3M4. HPAF, RWP-1, RWP-2. Capan-2 and Panc-Ca were obtained fresh in culture. All cell lines are established cultures maintained for at least 3 years in vitro using RPM1 1640 medium supplemented with 15% fetal calf scrum, penicillin, streptomycin and Hcpcs solution. The control cell line was normal, human C-77 T lymphocytes. In preparation for immunoperoxidase staining, the cells were harvested, washed 3 times and resuspended in PBS (pH 7.4). Resuspended cells (approx imately 1 x 107 cells in 15 ml PBS) underwent cytocentrifugation (Cryospin 2. Skondon. USA; 1500 rpm for 2 min) onto glued slides using two drops of the well-mixed cell suspension solutions. Rapid fixation of the mounted cells was accomplished with immersion in acetone for 2 min. To serve as negative control for the MAbs, nonimmune mouse serum (IgG purified. Coulter Immunology. USA) was used in place of the experimental MAbs in the immunoperoxidase assay, diluted I : 1000 in PBS. pH 7.4. For a nonspecific positive control of assay results, MAb to lymphocyte common antigen (LCA; Dako Santa Barbara. Calif.. USA), known to stain human lymphatic tissue, was reacted against paraffin sections of human tonsil. A standard im munoperoxidase assay using the avidin-biotin technique and 3-amino-9-ethyl carbazole (AEC) substrate coloring agent was carried out for each tissue, cell line and MAb combination. Slides contain ing the paraffin-embedded tissues were deparaffinized in sequential xylenes and rehydrated by immersion in graduated alcohols (ethanol) to water. Blocking of endogenous peroxidase activity was then accomplished in a 3% hydrogen peroxide-in-methanol solu tion for I h. Slides were washed twice in PBS (pH 7.4) for 10 min. F.xperimental MAbs. nonimmune mouse serum IgG and MAb to LCA were added to the tissue on the slides and incubated for I h at room temperature. Subsequent steps incorporated biotinylated rab bit antimouse sera (Dako). diluted 1 : 200 in PBS, for 30 min. avidin-biotin-peroxidasc complex (Dako) for 30 min. AF.C-3% hy drogen peroxide coloring solutions for 20 min, and counterstaining in hematoxylin and 2% ammonium chloride. Washing of the slides in rabbit sera (5% in PBS) preceded each reagent addition. Slides containing the cell lines were rehydrated in two separate PBS (pH 7.4) baths for 5 min each. To block endogenous peroxi dase activity, the slides were then washed in a 3% hydrogen peroxide-in-methanol bath for 30 45 min. Experimental MAbs. nonimmunc mouse serum IgG and MAb to LCA were added to the mounted cells and incubated for 1 h. The subsequent steps were identical to those used for the paraffin sections previously described.
Jakow atz/G an/U lich/W ile
151
Monoclonal Antibodies and Pancreatic Cancer
MAb 3DS2A stained 5 pancreatic cancer sections intensely. Two specimens stained with weak and moderate intensity, respectively. Sections of normal and pancreatic tissue showed at most trace staining in occasional interlobular ducts. Interestingly, MAb 3DS2A stained normal pancreas areas adjacent to pancreatic cancer with intense to moderate levels in interlobular, intralobular, centroacinar and acinar cells. This was not seen in the pancreatitis group where adjacent normal pancreatic tissue showed weak reac tivity solely in the interlobular ducts. MAb AR1-28 stained all malignant pancreatic tis sue at moderate to intense levels. In the normal-pan creas group, trace staining was seen in 3 specimens, while staining was absent in 2. Pancreatitis tissue sec tions stained variably with focally weak to moderate staining. Normal pancreas adjacent to malignant tis sue in the pancreatic cancer specimens showed vari able intense staining of interlobular ducts, in tralobular ducts, centroacinar cells and acinar cells. Islet cells did not stain. MAb AR2-20 stained 5 of 7 malignant pancreatic tissue specimens intensely and 2 with moderate in tensity. In the normal group, weak staining was seen in 4 specimens and no staining in 1. The location of stain in this group was evenly distributed among the interlobular/intralobular ducts and centroacinar/ acinar cells. Islets showed no reactivity. Normal pan creas adjacent to malignant tissue in the pancreatic cancer specimens showed intense staining evenly dis tributed among interlobular/intralobular ducts and centroacinar/acinar cells. This same result was seen in normal pancreas adjacent to pancreatitis tissue. The staining intensity of the MAbs for the human pancreatic cancer cell lines are summarized in table 2.
MAb CA19-9 stained 2 of the 6 cell lines, RWP-1 and Capan-2, intensely, while the remaining cell lines showed only trace to weak staining. Human T lym phocytes were nonreactive. MAb 3DS2A stained 3 cell lines, RWP-2. HPAF and Capan-2, intensely, while the remaining cell lines stained with moderate to trace intensity. Human T lymphocytes were nonreactive. MAb AR1-28 stained 3 cell lines, RWP-2, HPAF and T3M4, intensely, while 1 cell line, RWP-1, showed no reactivity. The remaining 2 lines, Panc-Ca and Capan-2, stained with moderate intensity. Human T lymphocytes were nonreactive. MAb AR2-20 stained 2 cell lines, RWP-2 and T3M4 intensely, but showed no reactivity to RWP-1 and Panc-Ca. HPAF and Capan-2 stained with moderate intensity. Human T lymphocytes were non reactive. MAb CA17-1A stained only 1 pancreatic cancer cell line, T3M4, intensely. RWP-1, RWP-2 and Ca pan-2 stained with moderate intensity, while Panc-Ca showed weak reactivity. However, MAb C al7-1A stained human T lymphocytes as intensely as was seen with anti-LCA. Anti-LCA showed trace reactivity with RWP-2 and Panc-Ca and no reactivity with the remaining pancre atic cancer cell lines. The control nonimmune mouse IgG was negative in all cases. Table 3 summarizes the results of MAb 3DS2A and MAb AR1-28 against tissue specimens of human co lonic adenocarcinoma. MAb 3DS2A stained all 4 specimens of colon cancer intensely. In addition, the normal mucosa adjacent to the malignant sections showed intense staining of the brush border of all specimens. MAb AR1-28 stained all 4 specimens of
Table 2. Staining intensity of the MAbs lor each cell line C'A 19-9
3DS2A
ARI-28
RWP-1 RWP-2 Panc-Ca IIPAF T3M4 Capan-2 T cells
3+ +/1+ + /1+ 3+ -
1+ 3+ 1+ 3+ 2+ 3+ —
3+ 2+ 3+ 3+ 2+ —
_
AR2-20
CA 17-1A
LCA
Nonimmune mouse IgG
—
2+ 2+ 1+ 2 -f 3+ 2+ 3+
_
—
+/+/-
-
-
-
3+
—
3+ -
2+ 3+ 2+ —
3 + = Intense; 2 + = moderate; 1 + = weak; + / — = trace; — = negative.
-
-
Downloaded by: Karolinska Institutet, University Library 130.237.122.245 - 1/18/2019 1:54:02 AM
Cell line
152 Table 3. Staining intensity of 3DS2A and AR1-28 for colon adenocarcinoma Colon cancer specimen
3DS2A
ARI-28
Mouse IgG
i 2 3 4
3+ 3+ 3+ 3+
3+ 3+ 3+ 3+
—
—
3DS2A also stained the brush border of the normal mucosa adjacent to the malignant tissue at a 3+ intensity. AR1-28 stained all normal mucosal cytoplasm and brush border adjacent to the malignant tissue at a 3 + intensity. 3 + = Intense; — = negative.
Table 4. Percentage of tissue specimens of moderate staining (2 + ) or greater MAb
CA19-9 3DS2A ARI-28 AR2-20 CAI7-IA
Tissue normal pancreas
pancreatitis
pancreatic cancer
5/5 0/5 0/5 0/5 2/5
4/4(100) 0/4 (0) 0/4 (0) 1/4 (25) 2/4 (50)
7/7 6/7 7/7 7/7 4/7
(100) (0) (0) (0) (40)
(100) (86) (100) (100) (57)
Numbers in parentheses are percentages.
colon cancer intensely, including the brush border and the epithelial cell cytoplasm of the normal mucosa adjacent to the malignant sections. Table 4 presents the staining results as percentage of specimens staining at moderate intensity levels and greater for benign and malignant pancreatic tissue. MAbs 3DS2A and AR1-28 showed no moderate or intense reactivity with normal pancreatic or chronic pancreatitis tissue while staining virtually all malig nant specimens.
Discussion In order to establish a clinical usefulness for radiolabeled MAbs directed against adenocarcinoma of the human pancreas, pre-radiolabeling studies of the sen
sitivity of human pancreatic-cancer-associated MAbs should be carried out in vitro. These studies will give direction and selection to further work involving the most ideal MAb to radiolabel for in vivo animal and human trials. We have compared the sensitivity and specificity of human pancreatic-cancer-associated MAbs against a battery of surgical paraffin sections of human normal pancreas, chronic pancreatitis tissue, pancreatic cancer, colon cancer and in vitro human pancreatic cancer cell lines. This was accomplished using the avidin-biotin-immunoperoxidase staining assay which is a well-established histologic technique of identifying tissue-associated antigens [21-23]. Our results show that MAbs 3DS2A and AR1-28 have the best combination of sensitivity and specificity for human pancreatic adenocarcinoma among the 5 MAbs tested. However, these 2 MAbs also showed intense reactivity toward human colon cancer tissues and weak to trace reactivity with normal pancreas and chronic pancreatitis tissue. MAb 3DS2A has not been characterized in this fashion against human pancreatic cancer. MAb AR1-28, however, has been previously characterized against pancreatic cancer cell lines, be nign and malignant pancreatic tissue and cancers of nonpancreatic origin by Chin et al. [24, 25]. They did not find any reactivity (as defined by immunofluores cence) to chronic pancreatitis tissue, to the majority of colon cancer specimens or to pancreatic cancer cell lines when AR1-28 was tested alone. This is in con trast to our results with respect to colon cancer tissue and the pancreatic cancer cell lines. Possible explana tions for these differences are the use of an immuno peroxidase rather than an immunofluorescence assay or the use of tissues with differing antigenic expres sion. Differing sensitivities and specificities result from inherent cross-reactivity of tumor-associated MAbs for normal tissue or malignant tissue from other gas trointestinal organs. Multiple factors, including tumor heterogeneity, similar embryonic origins and rapid cellular turnover with constant differentiation and dedifferentiation, are responsible for this cross-reac tivity [17, 24, 26]. Perhaps gastrointestinal malignan cies share tumor-specific antigens that are expressed in different amounts relative to histologic source. The possible solutions to this dilemma, which impact on clinical usefulness, are the relative range of expression of tumor-associated antigens and the use of multiple MAbs as a ‘cocktail’ agent used for immunodiagnosis [17, 19, 24, 26],
Downloaded by: Karolinska Institutet, University Library 130.237.122.245 - 1/18/2019 1:54:02 AM
Jakow atz/G an/U lich/W ile
153
M onoclonal Antibodies and Pancreatic Cancer
Conclusions The significance of this study is in the determina tion of a relative sensitivity of pancreatic-cancer-asso ciated MAbs for their malignant tissue targets. Once the histologic sensitivity is determined, limitation of specificity then becomes the primary clinical problem. MAbs AR1-28 and 3DS2A have excellent sensitivity for malignant pancreatic tissue, but will probably fail in their clinical usefulness when more than one gas trointestinal malignancy exists. When MAbs 3DS2A or AR1-28 are radiolabeled with either indium-111 or iodine-125, the intensity of uptake of each in malig nant versus benign pancreatic tissue should be suffi cient to resolve by nuclear scintiscanning. Preopera tive use of radiolabeled MAbs, sensitive to pancreatic cancer, may prove invaluable to the surgeon who must inform the patient of the probability of radical sur gery. Resolution of preoperative occult metastatic dis ease as well as follow-up scintiscan for recurrent dis ease postoperatively are further potential uses of MAb radiopharmaceuticals. Studies using radiolabeled AR1-28 and 3DS2A in the human tumor xenograftnude mouse model using the pancreatic cancer cell lines described are now underway in our laboratory.
Acknowledgements Wc arc indebted to Dr. Zenon Steplcwski (Wistar Institute. Philadelphia. Pa.. USA) for MAbs CAI9-9, CAI7-IA and 3DS2A. and Dr. J. Chin and Dr. F. Miller (Department ofPathology. State University of New York at Stony Brook School of Medicine, Stony Brook, N.Y.. USA) for MAbs ARI-28 and AR2-20. Wc also thank Dr. P.A. Meitner (Division of Medicine and Biology, Brown Uni versity, Providence. R.I.. USA) for the cell lines RWP-I and RWP2. Dr. R.S. Mctzgar (Duke University School of Medicine. Dur ham. N.C.. USA) for the cell lines HPAF and T3M4. Dr. George E. Moore (Division of Surgical Oncology, Denver General Hos pital. Denver. Col.. USA) for the cell line Panc-Ca and Dr. Gale S. Granger (Department of Molecular Biology. University of Califor nia, Irvine. Calif., USA) for the human C-77 T lymphocyte cell line.
References 1 Savarino V, Mansi C, Bistolfi P. et al: Failure of new diagnostic aids in proving detection of pancreatic cancer at a resectable stage. Dig Dis Sci 1983;28:1078-1082. 2 Robbins AG. Gerzof SG. Pugatch RD: Newer imaging tech niques for diagnosis of pancreatic cancer. Semin Oncol 1979;6:332-333. 3 Wood RAB. Moossa AR. Blackstone MO. et al: Comparative value of four methods of investigating the pancreas. Surgery 1976;80:518-522. 4 Mackie CR. Lu CT, Noble HG, et al: Prospective evaluation of angiography in the diagnosis and management of patients sus pected of having pancreatic cancer. Ann Surg 1979;189:11-17. 5 Bodner E, Schwamberger K. Mikuz G: Cytological diagnosis of pancreatic tumors. World J Surg 1982;6:103-106. 6 Goldstein I1M, Zornoza J: Percutaneous transperitoncal aspira tion biopsy of pancreatic masses. Dig Dis Sci 1978;23:840-843. 7 Teplick SK. Haskin PH. Matsumoto T. et al: Interventional radiology of the biliary system and pancreas. Surg Clin North Am 1984; 64: 87 119. 8 Moossa AR. Altorki N: Pancreatic biopsy. Surg Clin North Am 1983;63:1205-1213. 9 Frccny PC, Ball TJ: F.ndoscopic retrograde cholangiopan creatography (F.RCP) and percutaneous transhepatic cholangi ography (PTC) in the evaluation of suspected pancreatic car cinoma: Diagnostic limitations and contemporary roles. Cancer 1981;47:1666-1678. 10 Ishe 1. Toregard B-M. Akerman M: Intraoperative fine needle aspiration cytology in pancreatic lesions. Ann Surg 1979; 190:732 734. 11 Lee Y-T: Tissue diagnosis for carcinoma of the pancreas and periampullary structures. Cancer 1982;49:1035 1039. 12 Murray JL, Rosenblum MG. Sobol RF: Radioimmunoimaging in malignant melanoma with Ill-indium-labelled monoclonal antibody 96.5. Cancer Res 1985;45:2376-2381. 13 Epenetos AA, Hooker G, Durbin H: Indium-111 labelled monoclonal antibody to placental alkaline phosphatase in the detection of neoplasms of testis, ovary and cervix. Lancet 1985;ii:350—353. 14 Mach JP. Chatal JF, Lumbroso JD. Tumor localization in
Downloaded by: Karolinska Institutet, University Library 130.237.122.245 - 1/18/2019 1:54:02 AM
MAbs 3DS2A and AR1-28 show a relative range of reactivity where staining intensity diminishes from malignant pancreatic tissue to chronic pancreatitis tissue to normal pancreatic tissue. Clinically, this dif ference in selectivity and sensitivity can be used to our advantage in identifying malignant or normal tissue of the same histologic origin. Since the usual clinical dilemma is a nonoperative diagnosis of benign or malignant pancreatic lesions. MAbs that have a great reactivity for malignant tissue, and a relatively low reactivity for benign or inflamed tissue, are potentially good radiolabeled targeting agents. Presuming the radiolabeled MAb will concentrate in malignant pan creatic lesions to a greater extent than in benign tissue, the differential intensity of nuclear images should be within the resolution of present scintillation cameras. This has already been shown in serum tests of patients with pancreatic cancer, pancreatitis and normal con trols using CA19-9 [27, 28]. Patients with pancreatic cancer had significantly higher levels of circulating antigen. Therefore, the presence of tumor-associated antigen was not as crucial as the relative level.
154
16
17
18
19
20
21 22
patients by radiolabelled monoclonal antibodies against colon carcinoma. Cancer Res 1983:43:5593 5599. Beatty JD, Duda RB. Williams LE, et al: Preoperative imaging of colorectal carcinoma with 1“ In-labeled anticarcinoembryonic antigen monoclonal antibody. Cancer Res 1986:46: 6494-6502. Maimonis PJ. Meitner PA, Kajiji SM: Detection of metastatic variants of human pancreatic cancer cells by a monoclonal antibody. Proc AACR 1984:25:63. Schmiegel WH. Kalthoff H. Arndt R: Monoclonal antibodydefined human pancreatic cancer-associated antigens. Cancer Res 1985;45:1402-1407. Metzgar RS, Gaillard MT; Levine SJ: Antigens of human pan creatic adenocarcinoma cells defined by murine monoclonal antibodies. Cancer Res 1982:42:601-608. Kajiji SM, Davceva B, Quaranta V: Six monoclonal antibodies to human pancreatic cancer antigens. Cancer Res 1987; 47:1367-1376. Chin J, Miller F: Indentification and localization of human pancreatic tumor-associated antigens by monoclonal antibodies to RWP-1 and RWP-2 cells. Cancer Res 1985:45:1723-1729. Taylor CR: Immunoperoxidase techniques: Practical and theo retical aspects. Arch Pathol Lab Med 1978;102:113-121. Hancock WW. Becker GJ, Atkins RL: A comparison of fixa tives and immunohistochemical technics for use with mono clonal antibodies to cell surface antigens. Am J Clin Pathol 1981:78:825-831.
23 Atkinson BF. Ernst CS, Steplewski Z, ct al: Gastrointestinal cancer-associated antigen in immunoperoxidase assay. Cancer Res 1982:42:4820-4823. 24 Chin J, Zuna R, Miller F: Reactivity of monoclonal anti-human pancreatic carcinoma antibodies AR2-20 and AR1-28 with tu mors of nonpancreatic origin. Am J Pathol 1987;126:183-193. 25 Chin J. Miller F, Lane B: Detection of human pancreatic adeno carcinomas by histochcmical staining with monoclonal anti body AR1-28. Diagn Immunol 1985;3:99-105. 26 Halpern SE, Dillman RO, Hagan PL: The problems and prom ise of monoclonal antitumor antibodies. Diagn Imag 1983;5:40. 27 Farini R. Fabris C. Bonvicini P. et al: CA19-9 in the differential diagnosis between pancreatic cancer and chronic pancreatitis. Eur J Cancer Clin Oncol 1985;4:429-432. 28 Mahvi DM, Meyers WC, Bast RC. el al: Carcinoma of the pancreas: Therapeutic efficacy as defined by a sérodiagnostic test utilizing a monoclonal antibody. Ann Surg 1985;202:440-445.
James G. Jakowatz, MD Division of Surgical Oncology University of California Irvine Medical Center 101 The City Drive South Building 25, Route 81 Orange, CA 92668 (USA)
Downloaded by: Karolinska Institutet, University Library 130.237.122.245 - 1/18/2019 1:54:02 AM
15
Jakow atz/G an/U lich/W ile
< 1990 S. Kargcr AG. Basel 0030 2414 90 0472 0149 S 2.75 0
Murine Monoclonal Antibodies to Human Pancreatic Cancer; Specificity and Sensitivity James G. Jakowatz a>
b ,
Karen Gand, Thomas R. UlicIT, Alan G. lVileb
a Surgical Service, Veterans Administration Medical Center, Long Beach, Calif.; b Division of Surgical Oncology and c Department of Pathology, University of California Medical Center. Orange, Calif.; d University of California/California College of Medicine, Irvine, Calif, USA
Key Words. Monoclonal antibodies • Pancreatic cancer Abstract. Pancreatic carcinoma (n = 7), pancreatitis tissue (n = 4), normal pancreas tissue (n = 5), colonic adenocarcinoma (n = 4) and in vitro human pancreatic cancer cell lines (n = 6) were studied with the murine monoclonal antibodies (MAbs) 3DS2A, AR1-28, AR2-20, Cal9-9 and CA17-1A to determine their immunohistologic specificity and sensitivity for use as radiolabeled diagnostic imaging agents. Using the avidinbiotin-immunoperoxidase staining technique, MAbs 3DS2A and AR1-28 stained 86 and 100% of pancreatic cancer specimens, respectively. MAbs 3DS2A and AR1- 28 are suitable agents for use as radiolabeled diagnostic imaging agents in patients with pancreatic cancer.
Adenocarcinoma of the pancreas can be difficult to diagnose preoperatively. Radiologic modalities are nonspecific, rendering an anatomic, rather than a his tologic, abnormality [1-4]. Radiologically guided needle aspirations are unsuccessful in as many as 50% of cases. Further, these procedures are associated with morbidity, tissue sampling error and reliance upon the cytologist’s experience and skill [5-9], Intraoperative biopsies are also associated with morbidity and a 50-75% false-positive biopsy rate due to sampling error [8, 10, 11], In most cases of intraoperative exami nation, the sampling error prevents the discrimination between pancreatic cancer and pancreatitis. Often the surgeon is left with a decision to perform a radical operation based on clinical judgement. Potentially, this clinical dilemma could be solved using monoclonal antibodies (MAbs) developed from tumor-associated antigens. MAbs for other tumors have been radiolabeled with 131I o r 11 'In and used for diagnostic imaging in patients with various malignan cies [12-15]. Murine MAbs directed against pancreatic adenocarcinoma have been developed by other inves tigators, from either human pancreatic or colon can-
cer cell lines [16-20]. Their potential use as imaging agents in pancreatic cancer has not been explored. The purpose of this study was to investigate the specificity and sensitivity of pancreatic-cancer-associated MAbs in histologic specimens of pancreatic cancer and hu man pancreatic cancer cell lines. These studies deter mined the most sensitive and specific Mab to use for radioimaging experiments in athymic mice bearing human pancreatic cancer cell line xenografts.
Materials and Methods Five MAbs, reactive with human pancreatic adenocarcinoma, were obtained from two investigators. CAI9-9 and 3DS2A are murine IgG MAbs of hybridomas whose antigenic source was the human colon cancer cell line SW1116. Both were obtained as ascites and diluted 1 : 1500 and 1 : 250, respectively, in phosphate-buffered saline (PBS). pH 7.4, for immunoperoxidase staining. CA17-1A is a murine IgG MAb of a hybridoma developed from the human colon cancer cell line SW1083. CAI7-1A, obtained as ascites, was affinitypurified on a Staph-protein A column (Affigcl. Bio-Rad, Richmond. Calif., USA). No reactivity with any pancreatic tissue was obtained with CA17-1A in the crude ascites form. The purified MAb was diluted at 1 : 50 in PBS. pH 7.4. AR1-28 and AR2-20 arc murine IgG MAbs whose hybridoma origins stem from the human pancre atic cancer cell line R WP-1. Again the crude ascites forms were used
Downloaded by: Karolinska Institutet, University Library 130.237.122.245 - 1/18/2019 1:54:02 AM
Introduction
150
Slides were evaluated by light microscopy for the presence and intensity of stain. Intensity of stain was graded as intense (3 + ), moderate (2 + ), weak (1 + ), trace (+ / —) and negative ( —).
Results Table 1 summarizes the results of the immunoreactivity of the MAbs for the tissue specimens. MAb CA19-9 intensely stained all pancreatic car cinomas but also stained interlobular ducts, in tralobular ducts and centroacinar cells of normal pan creatic tissue (whether adjacent to pancreatitis, pan creatic cancer or in pancreatic tissue from normal patients). Islet and acinar cells did not stain. MAb CA17-1A stained pancreatic tissue, car cinoma and non-neoplastic pancreatig structures with equal intensity. Three pancreatic cancer specimens were strongly positive, one was moderately positive and the remaining samples showed only trace or weak levels of reactivity. MAb CA17-1A stained all nonneoplastic elements of pancreatitis tissue with weak to moderate reactivity in variable fashion.
Tabic 1. Staining intensity of the MAbs for each tissue Tissue CA19-9 3DS2A AR1-■28 AR2-20 CAI7-1A Nonimmune mouse IgG Normal pancreas 1 3+ 2 3+ +/3 3+ +/4 3+ + /5 3+ +/Pancreatitis 1 3+ 2 3+ 3 3+ 4 3+
1+ 1+ 1+ + /-
Pancreatic cancer 1 3+ 1+ 2 3+ 3+ 3 3+ 2+ 4 3+ 3+ 5 3+ 3+ 6 3+ 3+ 7 3+ 3+
+ /+ /+ /-
1+ 1+ 1+ 1+
+ /-
1+ 2+ 2+
2+
+ /-
1+ 1+ 2+ 1+
2+ 3+ 3+ 2+ 3+ 3+ 2+
2+ 3+ 2+ 3+ 3+ 3+ 3+
3+
1+ 1+ 1+
3 + = Intense; 2 + == moderate; 1 + — = negative.
+ /-
2+ + /-
+ /-
3+ 3+ 2+ -
-
1+ - weak; + / — = trace;
Downloaded by: Karolinska Institutet, University Library 130.237.122.245 - 1/18/2019 1:54:02 AM
and diluted in PBS, pH 7.4, at 1 : 2000 and 1 : 1250. respectively. All dilutions were determined by immunoperoxidase checkerboard ti tration methods against pancreatic cancer tissue. Paraffin-embedded sections of human pancreatic cancer, pan creatitis tissue, normal pancreatic tissue and human colonic cancer obtained at the time of operation were utilized as target tissues. Surgical specimens were chosen over autopsy specimens to mini mize the associated antigen loss as a result of autodigestion. Pancreatic cancer specimens were obtained from resected pan creatic glands involved with primary adenocarcinoma. Normal pancreas tissue was obtained from glands resected for trauma where no evidence of malignancy existed. Pancreatitis tissue was obtained from biopsied or resected pancreas of patients with chronic pan creatitis. Colonic cancer samples were obtained from colectomy specimens containing both normal and malignant tissue. All paraf fin-embedded tissue was cut into 4-pm sections and applied to slides for immunoperoxidase staining. The human pancreatic cancer cell lines T3M4. HPAF, RWP-1, RWP-2. Capan-2 and Panc-Ca were obtained fresh in culture. All cell lines are established cultures maintained for at least 3 years in vitro using RPM1 1640 medium supplemented with 15% fetal calf scrum, penicillin, streptomycin and Hcpcs solution. The control cell line was normal, human C-77 T lymphocytes. In preparation for immunoperoxidase staining, the cells were harvested, washed 3 times and resuspended in PBS (pH 7.4). Resuspended cells (approx imately 1 x 107 cells in 15 ml PBS) underwent cytocentrifugation (Cryospin 2. Skondon. USA; 1500 rpm for 2 min) onto glued slides using two drops of the well-mixed cell suspension solutions. Rapid fixation of the mounted cells was accomplished with immersion in acetone for 2 min. To serve as negative control for the MAbs, nonimmune mouse serum (IgG purified. Coulter Immunology. USA) was used in place of the experimental MAbs in the immunoperoxidase assay, diluted I : 1000 in PBS. pH 7.4. For a nonspecific positive control of assay results, MAb to lymphocyte common antigen (LCA; Dako Santa Barbara. Calif.. USA), known to stain human lymphatic tissue, was reacted against paraffin sections of human tonsil. A standard im munoperoxidase assay using the avidin-biotin technique and 3-amino-9-ethyl carbazole (AEC) substrate coloring agent was carried out for each tissue, cell line and MAb combination. Slides contain ing the paraffin-embedded tissues were deparaffinized in sequential xylenes and rehydrated by immersion in graduated alcohols (ethanol) to water. Blocking of endogenous peroxidase activity was then accomplished in a 3% hydrogen peroxide-in-methanol solu tion for I h. Slides were washed twice in PBS (pH 7.4) for 10 min. F.xperimental MAbs. nonimmune mouse serum IgG and MAb to LCA were added to the tissue on the slides and incubated for I h at room temperature. Subsequent steps incorporated biotinylated rab bit antimouse sera (Dako). diluted 1 : 200 in PBS, for 30 min. avidin-biotin-peroxidasc complex (Dako) for 30 min. AF.C-3% hy drogen peroxide coloring solutions for 20 min, and counterstaining in hematoxylin and 2% ammonium chloride. Washing of the slides in rabbit sera (5% in PBS) preceded each reagent addition. Slides containing the cell lines were rehydrated in two separate PBS (pH 7.4) baths for 5 min each. To block endogenous peroxi dase activity, the slides were then washed in a 3% hydrogen peroxide-in-methanol bath for 30 45 min. Experimental MAbs. nonimmunc mouse serum IgG and MAb to LCA were added to the mounted cells and incubated for 1 h. The subsequent steps were identical to those used for the paraffin sections previously described.
Jakow atz/G an/U lich/W ile
151
Monoclonal Antibodies and Pancreatic Cancer
MAb 3DS2A stained 5 pancreatic cancer sections intensely. Two specimens stained with weak and moderate intensity, respectively. Sections of normal and pancreatic tissue showed at most trace staining in occasional interlobular ducts. Interestingly, MAb 3DS2A stained normal pancreas areas adjacent to pancreatic cancer with intense to moderate levels in interlobular, intralobular, centroacinar and acinar cells. This was not seen in the pancreatitis group where adjacent normal pancreatic tissue showed weak reac tivity solely in the interlobular ducts. MAb AR1-28 stained all malignant pancreatic tis sue at moderate to intense levels. In the normal-pan creas group, trace staining was seen in 3 specimens, while staining was absent in 2. Pancreatitis tissue sec tions stained variably with focally weak to moderate staining. Normal pancreas adjacent to malignant tis sue in the pancreatic cancer specimens showed vari able intense staining of interlobular ducts, in tralobular ducts, centroacinar cells and acinar cells. Islet cells did not stain. MAb AR2-20 stained 5 of 7 malignant pancreatic tissue specimens intensely and 2 with moderate in tensity. In the normal group, weak staining was seen in 4 specimens and no staining in 1. The location of stain in this group was evenly distributed among the interlobular/intralobular ducts and centroacinar/ acinar cells. Islets showed no reactivity. Normal pan creas adjacent to malignant tissue in the pancreatic cancer specimens showed intense staining evenly dis tributed among interlobular/intralobular ducts and centroacinar/acinar cells. This same result was seen in normal pancreas adjacent to pancreatitis tissue. The staining intensity of the MAbs for the human pancreatic cancer cell lines are summarized in table 2.
MAb CA19-9 stained 2 of the 6 cell lines, RWP-1 and Capan-2, intensely, while the remaining cell lines showed only trace to weak staining. Human T lym phocytes were nonreactive. MAb 3DS2A stained 3 cell lines, RWP-2. HPAF and Capan-2, intensely, while the remaining cell lines stained with moderate to trace intensity. Human T lymphocytes were nonreactive. MAb AR1-28 stained 3 cell lines, RWP-2, HPAF and T3M4, intensely, while 1 cell line, RWP-1, showed no reactivity. The remaining 2 lines, Panc-Ca and Capan-2, stained with moderate intensity. Human T lymphocytes were nonreactive. MAb AR2-20 stained 2 cell lines, RWP-2 and T3M4 intensely, but showed no reactivity to RWP-1 and Panc-Ca. HPAF and Capan-2 stained with moderate intensity. Human T lymphocytes were non reactive. MAb CA17-1A stained only 1 pancreatic cancer cell line, T3M4, intensely. RWP-1, RWP-2 and Ca pan-2 stained with moderate intensity, while Panc-Ca showed weak reactivity. However, MAb C al7-1A stained human T lymphocytes as intensely as was seen with anti-LCA. Anti-LCA showed trace reactivity with RWP-2 and Panc-Ca and no reactivity with the remaining pancre atic cancer cell lines. The control nonimmune mouse IgG was negative in all cases. Table 3 summarizes the results of MAb 3DS2A and MAb AR1-28 against tissue specimens of human co lonic adenocarcinoma. MAb 3DS2A stained all 4 specimens of colon cancer intensely. In addition, the normal mucosa adjacent to the malignant sections showed intense staining of the brush border of all specimens. MAb AR1-28 stained all 4 specimens of
Table 2. Staining intensity of the MAbs lor each cell line C'A 19-9
3DS2A
ARI-28
RWP-1 RWP-2 Panc-Ca IIPAF T3M4 Capan-2 T cells
3+ +/1+ + /1+ 3+ -
1+ 3+ 1+ 3+ 2+ 3+ —
3+ 2+ 3+ 3+ 2+ —
_
AR2-20
CA 17-1A
LCA
Nonimmune mouse IgG
—
2+ 2+ 1+ 2 -f 3+ 2+ 3+
_
—
+/+/-
-
-
-
3+
—
3+ -
2+ 3+ 2+ —
3 + = Intense; 2 + = moderate; 1 + = weak; + / — = trace; — = negative.
-
-
Downloaded by: Karolinska Institutet, University Library 130.237.122.245 - 1/18/2019 1:54:02 AM
Cell line
152 Table 3. Staining intensity of 3DS2A and AR1-28 for colon adenocarcinoma Colon cancer specimen
3DS2A
ARI-28
Mouse IgG
i 2 3 4
3+ 3+ 3+ 3+
3+ 3+ 3+ 3+
—
—
3DS2A also stained the brush border of the normal mucosa adjacent to the malignant tissue at a 3+ intensity. AR1-28 stained all normal mucosal cytoplasm and brush border adjacent to the malignant tissue at a 3 + intensity. 3 + = Intense; — = negative.
Table 4. Percentage of tissue specimens of moderate staining (2 + ) or greater MAb
CA19-9 3DS2A ARI-28 AR2-20 CAI7-IA
Tissue normal pancreas
pancreatitis
pancreatic cancer
5/5 0/5 0/5 0/5 2/5
4/4(100) 0/4 (0) 0/4 (0) 1/4 (25) 2/4 (50)
7/7 6/7 7/7 7/7 4/7
(100) (0) (0) (0) (40)
(100) (86) (100) (100) (57)
Numbers in parentheses are percentages.
colon cancer intensely, including the brush border and the epithelial cell cytoplasm of the normal mucosa adjacent to the malignant sections. Table 4 presents the staining results as percentage of specimens staining at moderate intensity levels and greater for benign and malignant pancreatic tissue. MAbs 3DS2A and AR1-28 showed no moderate or intense reactivity with normal pancreatic or chronic pancreatitis tissue while staining virtually all malig nant specimens.
Discussion In order to establish a clinical usefulness for radiolabeled MAbs directed against adenocarcinoma of the human pancreas, pre-radiolabeling studies of the sen
sitivity of human pancreatic-cancer-associated MAbs should be carried out in vitro. These studies will give direction and selection to further work involving the most ideal MAb to radiolabel for in vivo animal and human trials. We have compared the sensitivity and specificity of human pancreatic-cancer-associated MAbs against a battery of surgical paraffin sections of human normal pancreas, chronic pancreatitis tissue, pancreatic cancer, colon cancer and in vitro human pancreatic cancer cell lines. This was accomplished using the avidin-biotin-immunoperoxidase staining assay which is a well-established histologic technique of identifying tissue-associated antigens [21-23]. Our results show that MAbs 3DS2A and AR1-28 have the best combination of sensitivity and specificity for human pancreatic adenocarcinoma among the 5 MAbs tested. However, these 2 MAbs also showed intense reactivity toward human colon cancer tissues and weak to trace reactivity with normal pancreas and chronic pancreatitis tissue. MAb 3DS2A has not been characterized in this fashion against human pancreatic cancer. MAb AR1-28, however, has been previously characterized against pancreatic cancer cell lines, be nign and malignant pancreatic tissue and cancers of nonpancreatic origin by Chin et al. [24, 25]. They did not find any reactivity (as defined by immunofluores cence) to chronic pancreatitis tissue, to the majority of colon cancer specimens or to pancreatic cancer cell lines when AR1-28 was tested alone. This is in con trast to our results with respect to colon cancer tissue and the pancreatic cancer cell lines. Possible explana tions for these differences are the use of an immuno peroxidase rather than an immunofluorescence assay or the use of tissues with differing antigenic expres sion. Differing sensitivities and specificities result from inherent cross-reactivity of tumor-associated MAbs for normal tissue or malignant tissue from other gas trointestinal organs. Multiple factors, including tumor heterogeneity, similar embryonic origins and rapid cellular turnover with constant differentiation and dedifferentiation, are responsible for this cross-reac tivity [17, 24, 26]. Perhaps gastrointestinal malignan cies share tumor-specific antigens that are expressed in different amounts relative to histologic source. The possible solutions to this dilemma, which impact on clinical usefulness, are the relative range of expression of tumor-associated antigens and the use of multiple MAbs as a ‘cocktail’ agent used for immunodiagnosis [17, 19, 24, 26],
Downloaded by: Karolinska Institutet, University Library 130.237.122.245 - 1/18/2019 1:54:02 AM
Jakow atz/G an/U lich/W ile
153
M onoclonal Antibodies and Pancreatic Cancer
Conclusions The significance of this study is in the determina tion of a relative sensitivity of pancreatic-cancer-asso ciated MAbs for their malignant tissue targets. Once the histologic sensitivity is determined, limitation of specificity then becomes the primary clinical problem. MAbs AR1-28 and 3DS2A have excellent sensitivity for malignant pancreatic tissue, but will probably fail in their clinical usefulness when more than one gas trointestinal malignancy exists. When MAbs 3DS2A or AR1-28 are radiolabeled with either indium-111 or iodine-125, the intensity of uptake of each in malig nant versus benign pancreatic tissue should be suffi cient to resolve by nuclear scintiscanning. Preopera tive use of radiolabeled MAbs, sensitive to pancreatic cancer, may prove invaluable to the surgeon who must inform the patient of the probability of radical sur gery. Resolution of preoperative occult metastatic dis ease as well as follow-up scintiscan for recurrent dis ease postoperatively are further potential uses of MAb radiopharmaceuticals. Studies using radiolabeled AR1-28 and 3DS2A in the human tumor xenograftnude mouse model using the pancreatic cancer cell lines described are now underway in our laboratory.
Acknowledgements Wc arc indebted to Dr. Zenon Steplcwski (Wistar Institute. Philadelphia. Pa.. USA) for MAbs CAI9-9, CAI7-IA and 3DS2A. and Dr. J. Chin and Dr. F. Miller (Department ofPathology. State University of New York at Stony Brook School of Medicine, Stony Brook, N.Y.. USA) for MAbs ARI-28 and AR2-20. Wc also thank Dr. P.A. Meitner (Division of Medicine and Biology, Brown Uni versity, Providence. R.I.. USA) for the cell lines RWP-I and RWP2. Dr. R.S. Mctzgar (Duke University School of Medicine. Dur ham. N.C.. USA) for the cell lines HPAF and T3M4. Dr. George E. Moore (Division of Surgical Oncology, Denver General Hos pital. Denver. Col.. USA) for the cell line Panc-Ca and Dr. Gale S. Granger (Department of Molecular Biology. University of Califor nia, Irvine. Calif., USA) for the human C-77 T lymphocyte cell line.
References 1 Savarino V, Mansi C, Bistolfi P. et al: Failure of new diagnostic aids in proving detection of pancreatic cancer at a resectable stage. Dig Dis Sci 1983;28:1078-1082. 2 Robbins AG. Gerzof SG. Pugatch RD: Newer imaging tech niques for diagnosis of pancreatic cancer. Semin Oncol 1979;6:332-333. 3 Wood RAB. Moossa AR. Blackstone MO. et al: Comparative value of four methods of investigating the pancreas. Surgery 1976;80:518-522. 4 Mackie CR. Lu CT, Noble HG, et al: Prospective evaluation of angiography in the diagnosis and management of patients sus pected of having pancreatic cancer. Ann Surg 1979;189:11-17. 5 Bodner E, Schwamberger K. Mikuz G: Cytological diagnosis of pancreatic tumors. World J Surg 1982;6:103-106. 6 Goldstein I1M, Zornoza J: Percutaneous transperitoncal aspira tion biopsy of pancreatic masses. Dig Dis Sci 1978;23:840-843. 7 Teplick SK. Haskin PH. Matsumoto T. et al: Interventional radiology of the biliary system and pancreas. Surg Clin North Am 1984; 64: 87 119. 8 Moossa AR. Altorki N: Pancreatic biopsy. Surg Clin North Am 1983;63:1205-1213. 9 Frccny PC, Ball TJ: F.ndoscopic retrograde cholangiopan creatography (F.RCP) and percutaneous transhepatic cholangi ography (PTC) in the evaluation of suspected pancreatic car cinoma: Diagnostic limitations and contemporary roles. Cancer 1981;47:1666-1678. 10 Ishe 1. Toregard B-M. Akerman M: Intraoperative fine needle aspiration cytology in pancreatic lesions. Ann Surg 1979; 190:732 734. 11 Lee Y-T: Tissue diagnosis for carcinoma of the pancreas and periampullary structures. Cancer 1982;49:1035 1039. 12 Murray JL, Rosenblum MG. Sobol RF: Radioimmunoimaging in malignant melanoma with Ill-indium-labelled monoclonal antibody 96.5. Cancer Res 1985;45:2376-2381. 13 Epenetos AA, Hooker G, Durbin H: Indium-111 labelled monoclonal antibody to placental alkaline phosphatase in the detection of neoplasms of testis, ovary and cervix. Lancet 1985;ii:350—353. 14 Mach JP. Chatal JF, Lumbroso JD. Tumor localization in
Downloaded by: Karolinska Institutet, University Library 130.237.122.245 - 1/18/2019 1:54:02 AM
MAbs 3DS2A and AR1-28 show a relative range of reactivity where staining intensity diminishes from malignant pancreatic tissue to chronic pancreatitis tissue to normal pancreatic tissue. Clinically, this dif ference in selectivity and sensitivity can be used to our advantage in identifying malignant or normal tissue of the same histologic origin. Since the usual clinical dilemma is a nonoperative diagnosis of benign or malignant pancreatic lesions. MAbs that have a great reactivity for malignant tissue, and a relatively low reactivity for benign or inflamed tissue, are potentially good radiolabeled targeting agents. Presuming the radiolabeled MAb will concentrate in malignant pan creatic lesions to a greater extent than in benign tissue, the differential intensity of nuclear images should be within the resolution of present scintillation cameras. This has already been shown in serum tests of patients with pancreatic cancer, pancreatitis and normal con trols using CA19-9 [27, 28]. Patients with pancreatic cancer had significantly higher levels of circulating antigen. Therefore, the presence of tumor-associated antigen was not as crucial as the relative level.
154
16
17
18
19
20
21 22
patients by radiolabelled monoclonal antibodies against colon carcinoma. Cancer Res 1983:43:5593 5599. Beatty JD, Duda RB. Williams LE, et al: Preoperative imaging of colorectal carcinoma with 1“ In-labeled anticarcinoembryonic antigen monoclonal antibody. Cancer Res 1986:46: 6494-6502. Maimonis PJ. Meitner PA, Kajiji SM: Detection of metastatic variants of human pancreatic cancer cells by a monoclonal antibody. Proc AACR 1984:25:63. Schmiegel WH. Kalthoff H. Arndt R: Monoclonal antibodydefined human pancreatic cancer-associated antigens. Cancer Res 1985;45:1402-1407. Metzgar RS, Gaillard MT; Levine SJ: Antigens of human pan creatic adenocarcinoma cells defined by murine monoclonal antibodies. Cancer Res 1982:42:601-608. Kajiji SM, Davceva B, Quaranta V: Six monoclonal antibodies to human pancreatic cancer antigens. Cancer Res 1987; 47:1367-1376. Chin J, Miller F: Indentification and localization of human pancreatic tumor-associated antigens by monoclonal antibodies to RWP-1 and RWP-2 cells. Cancer Res 1985:45:1723-1729. Taylor CR: Immunoperoxidase techniques: Practical and theo retical aspects. Arch Pathol Lab Med 1978;102:113-121. Hancock WW. Becker GJ, Atkins RL: A comparison of fixa tives and immunohistochemical technics for use with mono clonal antibodies to cell surface antigens. Am J Clin Pathol 1981:78:825-831.
23 Atkinson BF. Ernst CS, Steplewski Z, ct al: Gastrointestinal cancer-associated antigen in immunoperoxidase assay. Cancer Res 1982:42:4820-4823. 24 Chin J, Zuna R, Miller F: Reactivity of monoclonal anti-human pancreatic carcinoma antibodies AR2-20 and AR1-28 with tu mors of nonpancreatic origin. Am J Pathol 1987;126:183-193. 25 Chin J. Miller F, Lane B: Detection of human pancreatic adeno carcinomas by histochcmical staining with monoclonal anti body AR1-28. Diagn Immunol 1985;3:99-105. 26 Halpern SE, Dillman RO, Hagan PL: The problems and prom ise of monoclonal antitumor antibodies. Diagn Imag 1983;5:40. 27 Farini R. Fabris C. Bonvicini P. et al: CA19-9 in the differential diagnosis between pancreatic cancer and chronic pancreatitis. Eur J Cancer Clin Oncol 1985;4:429-432. 28 Mahvi DM, Meyers WC, Bast RC. el al: Carcinoma of the pancreas: Therapeutic efficacy as defined by a sérodiagnostic test utilizing a monoclonal antibody. Ann Surg 1985;202:440-445.
James G. Jakowatz, MD Division of Surgical Oncology University of California Irvine Medical Center 101 The City Drive South Building 25, Route 81 Orange, CA 92668 (USA)
Downloaded by: Karolinska Institutet, University Library 130.237.122.245 - 1/18/2019 1:54:02 AM
15
Jakow atz/G an/U lich/W ile
