(1) Seminar on “Analysis of Tumor-associated Antigens of Digestive Organs with Monoclonal Antibodies and Their Clinical Application”
This seminar was held on February 21-23, 1985, at the La Jolla Village Inn in La Jolla, California. The organizers were Dr. Ralph A. Reisfeld, Stanford University, Stanford, California, and Dr. Akira Yachi, Department of Internal Medicine, Sapporo Medical College, Sapporo, Japan. There were seven participants each from the United States and from Japan. The purpose of the seminar was to discuss and exchange information on new approaches for immunodiagnosis and immunotherapy of gastrointestinal tract cancers.
After opening remarks by Dr. Ralph A. Reisfeld (Head and Member, Division of Tumor Immunology, Scripps Clinic and Research Foundation), Dr. Akira Yachi gave a talk on a circulating adenocarcinoma-associated antigen detected by monoclonal antibody. A hybridoma derived from murine myeloma cells fused with splenocytes of mice, which were immunized with lung adenocarcinoma line A549, secreted monoclonal antibody YH206 (IgM) that reacted with the antigenic determinant preferentially expressed on human adenocarcinomas. Interestingly, it reacted with the supernatant of the A549 line with ELISA, and Western blotting experiment revealed that the monoclonal antibody YH206 reacted with a single polypeptide of more than 330 kDa. Immunohistological studies demonstrated that this monoclonal antibody strongly reacted with adenocarcinomas and bronchoalveolar cell carcinomas of the lung but not with epidermoid carcinomas, large cell carcinomas, or small cell carcinomas. It stained the majority of stomach and pancreas cancers. Although it reacted with fetal lung, it did not stain the normal lung tissue. The determinant recognized by the antibody YH206 was considered to be an asialosyl carbohydrate chain.
Reversed passive hemagglutination test employing purified antibody YH206 revealed that this antibody could successfully detect the higher amount of circulating antigen in 35 of 67 (52.2%) sera from cancer patients, whereas it detected antigen in 3 of 30 (10.0%) sera from healthy controls.
Dr. A. C. Morgan (NeoRx Corporation, Seattle, Washington) reported that monoclonal antibodies to tumor-associated antigens (TAA) of human colorectal cancer were elicited by using immunosorbents of lectins combined with peripheral protein abstracts of xenografted colon adenocarcinoma. This method of immunization was compared with whole cells from surgical specimens and to crude membranes from xenografted tumors. The immunosorbent immunogens were found to be superior to the other immunogens in three ways: 1) the number of hybrids reactive with human colon tumor cells but not with lymphoid extracts; 2) the number of stable hybrids after cloning; and 3) the number of hybridoma clones reactive with tissue sections of colon tumors but not normal colonic mucosa. In addition, lectin immunosorbents elicited primarily IgG antibodies, especially IgG3 with almost 50% of the clones of interest reacting with seemingly less immunogenic glycoproteins. Monoclonal antibodies of all IgG isotypes were tested for their ability to mediate antibody-dependent cellular cytotoxicity (ADCC), using both mononuclear and large granular lymphocyte (LGL)-enriched preparations as effector cells. LGLs were the most efficient effector cells to mediate ADCC at effector:target cell ratios of 5:1. All monoclonal anti-bodies of IgG3 subclass showed substantial ADCC versus the LS180 colon carcinoma cell line. Occasional monoclonal antibodies of the IgG1 and IgG2a subclass also mediated ADCC but not to the same extent as IgG3 antibodies. The improved elicitation of monoclonal antibodies to TAA by the use of lectin immunosorbents and peripheral protein extracts has considerable potential for generating reagents useful in diagnosis and therapy of human tumors.
Dr. Tadashi Watanabe (Nagoya University School of Medicine, Nagoya) reported on the analysis of cell surface antigens of cancer of the stomach and radioimmunodetection of gastrointestinal tumors by mouse monoclonal antibodies.
Three gastric cancer cell lines derived from poorly differentiated adenocarcinoma were established in his laboratory and used to generate mouse monoclonal antibodies. The specificity of antigen detected by the monoclonal antibody was analyzed by mixed hemadsorption assay (MHA) on a panel of human cell lines and immunoperoxidase reactivity with frozen tissue sections.
Monoclonal antibody GC302 (IgG1) was generated by immunizing mice with NUGC3 and found to react with 7 of 7 gastric cancers and 3 of 3 colon cancers but not with 2 of 2 hepatocellular cancers, 3 of 3 melanomas, and 2 of 2 astrocytomas. In titers of direct tests, well-differentiated types of adenocarcinomas showed a high level of antigen expression, although this antibody was produced using poorly differentiated adenocarcinoma. In tissue sections GC302 did not react with normal adult gastric mucosa but reacted with mucosa in the stomach of fetus and the epithelial cells in intestinal metaplasia lesion. In tissues from cancers of the stomach, GC302 reacted with any type of gastric cancer, including poorly and well-differentiated adenocarcinoma. GC302 also reacted with normal mucosa of the intestine, colon, and rectum as well as cancers from these organs. Molecular weight was estimated to be 40,000 by Western blot analysis. From these results, characteristics of antigen detected by GC302 were suggested to be an oncofetal antigen in stomach or a differentiation antigen which divides the gastrointestinal tracts into foregut origin and mid or hindgut origin.
Monoclonal antibody GC401 (IgG2a), which is reactive with epithelium of gastrointestinal tracts, was also analyzed. The antigen detected by GC401 was widely distributed in mucosa of gastrointestinal tracts. Since GC302 could be used for the diagnosis of lymph node metastasis of gastric and colonic cancers, the localization of gastric cancer bearing in nude mice using 125I-radio-labeled GC302 was investigated.
Dr. Setsuo Hirohashi (National Cancer Center Research Institute, Tokyo) reported the carbohydrate antigen defined by a monoclonal antibody raised against a gastric cancer xenograft.
A monoclonal antibody, ST-4-39, was obtained using a human gastric cancer xenograft, ST-4, as an immunogen. Immunization was achieved by transferring immunocompetent mouse spleen cells into a nude mouse bearing ST-4. Hybridomas were produced with the spleen cells of the mouse rejecting the tumor and screened by immunohistochemical reactivity with cancers and normal tissues on formalin-fixed paraffin sections. ST-4-39 immunohistochemically reacted with various cancers including gastric, colorectal, and pancreatic cancers, as well as some normal tissues. ST-4-39 and NS-19-9 differed in immunohistochemical reactivity, although they reacted with some cancers and a few normal tissues in common.
PBS extracts of normal and cancer tissues were examined for the antigen reactive with ST-4-39 by sandwich enzyme immunoassay. The extractable antigen was detected in adenocarcinomas of colon, stomach, and lung, while it was detected only in the salivary gland and the trachea among normal tissues examined. Gel filtration analysis of the antigen indicated a molecular weight of 1 x 10 6, and the antigenic determinant was suggested to be a carbohydrate chain with terminal sialic acid by studies using periodic acid, neuraminidase, and pronase treatments. Furthermore, the ST-4-39 antigen affinity purified from two gastric cancer strains was shown to contain multiple carbohydrate determinants including sialyl-Lewisa and sialyl-Lewisx, suggesting the antigen to be a mucin.
Finally, the ST-4-39 antigen was detected in sera of patients with stomach cancer (18 of 28), pancreatic cancer (8 of 22), breast cancer (14 of 24), and colon cancer (6 of 28), while it was not detected in sera of 50 healthy donors.
Dr. Teruaki Sekine (National Cancer Research Institute, Tokyo) reported on the preparation of monoclonal antibodies to scirrhous gastric carcinoma and their use.
A monoclonal antibody reactive with scirrhous carcinoma (diffuse type of carcinoma) of the stomach was established. Human gastric carcinoma strain St-15 was used as an immunogen, and antibody activity was detected by ABC staining of paraffin sections. The monoclonal antibody named 2-70 reacted with 42 of 51 (82%) cases of gastric carcinoma. In 25 of these cases which had scirrhous carcinoma of the stomach, 22 (88%) cases reacted with the antibody. Colonic and pancreatic carcinomas also reacted, but weakly. The 2-70 antibody reacted with some normal tissues such as salivary gland and colonic epithelium. Treatment of both antigens derived from gastric carcinoma and colonic epithelium with neuraminidase slightly enhanced antigen titer, suggesting that sialic acid is not involved in the epitope. The molecular weights of these antigens were estimated to be over one million daltons by gel filtration. However, most of the antigen derived from gastric carcinoma was found to be soluble in 1 M perchloric acid, whereas the solubility of that from the colon was only about 10%. The antigen was detected in sera by EIA. Fifty percent of sera from various gastric cancer patients was shown to be positive.
Dr. Zenon Steplewski (Wistar Institute, Philadelphia, Pennsylvania) presented data indicating that selected mouse monoclonal antibodies of IgG2a and IgG3 isotype, combined with human effector cells such as monocytes, can effectively inhibit growth of human colorectal and pancreatic cancers in nude mouse model systems. In efforts designed to eventually optimize these effects in vivo, INF was found to stimulate expression of Fc receptors on monocytes as indicated by their improved in vitro reactivity in ADCC reactions. A selected monoclonal antibody, i.e., 17-1A (IgG2a), is already being tested in Phase I trials of pancreatic cancer patients, and efforts are underway to select isotype switch variants of monoclonal antibodies to produce new reagents for improved immunotherapy of pancreatic cancer.
Dr. Takesada Mori (Osaka University, Osaka) then reported on the epitopic differences between pancreas cancer-associated antigen (PCAA) (POA) and PCAAc revealed by monoclonal antibodies.
PCAA was isolated from ascites fluid of a pancreatic cancer patient by one of his colleagues (Shimano) in 1981, and later it was proved to be identical with Gelder's POA. The same antigen was also isolated from normal colonic mucosa by another colleague (Kitada) in 1984. Circulating PCAA levels in pancreatic cancer patients were frequently increased above mean +2 S.D. of those in healthy individuals.
Monoclonal antibodies were prepared against PCAAc (derived from colonic mucosa), and epitopic differences between PCAA and PCAAc were investigated. In radioimmunoassay, both PCAA and PCAAc were detected by type A antibodies, while PCAA was not detected at all by type C antibodies. Antigenic differences between PCAA from pancreatic cancer and PCAA from other cancers were also revealed. In immunohistological studies on normal tissues, type A antibodies were reactive to epithelial cells of various organs, while type C antibodies were reactive to goblet cells of small and large intestines, exclusively.
Dr. Richard S. Metzgar (Duke University, Durham, North Carolina) reported that a mouse monoclonal antibody directed to human pancreatic adenocarcinoma cells recognizes a heavily glycosylated mucin-like antigen (DU-PAN-2) that can be readily detected in body fluids by a competition radioimmunoassay (RIA). DU-PAN-2 is expressed on a variety of neoplastic and nonneoplastic cells and appears to be an oncofetal or differentiation type antigen. Measuring levels of DU-PAN-2 in the serum of patients with adenocarcinomas of pancreas, stomach, and gallbladder proved to be a good monitoring assay as high levels of this antigen found in blood correlated with tumor growth. RIAs measuring blood levels of DU-PAN-2 can accurately reflect the clinical course of the disease and may be a valuable adjunct to predict recurrent or progressive disease and to evaluate progress of therapy.
Dr. Vito Quaranta (Scripps Clinic and Research Foundation, La Jolla, California) described monoclonal antibodies recognizing antigens associated with carcinoma of the human exocrine pancreas. The association with this type of cancer was established on the basis of restricted reactivity with normal and neoplastic human tissues by immunohistochemistry. The antigen reactive with antibody 53-23 is a nonglycosylated protein of molecular weight 38,000, mostly present in the cytoplasm of pancreatic carcinoma cell lines and tumors. Anti-body S3-41 immunoprecipitates four components closely migrating in sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) gels at an approximate molecular weight of 140,000. Two of these components are glycosylated and not disulphide linked, but their molecular arrangement is not yet clear. The antigen is found associated mostly with the basal lamina of neoplastic epithelia. Both of these antibodies react with a variety of carcinoma of gastrointestinal tract origin and display some reactivity with normal pancreatic ductal cells. Dr. Quaranta stressed some of the problems associated with the development of antibodies to pancreatic tumors that stem from the scarcity of cell lines and surgical specimens, as well as from apparent antigen loss in some of the cell lines in culture. In this regard, he described three sublines isolated by Dr. Shama Kajiji from the pancreatic carcinoma line Colo 357. The cell lines are termed SG, FG, and FG-met2. Because of their stable growth characteristics in vitro and in vivo in athymic mice, they may constitute a useful model for human pancreatic carcinoma. In addition, the two FG sublines display a highly reproducible metastatic behavior in the nude mouse, whereas SG is unable to give rise to metastasis, so that the lines may be of importance at a more general level for studying the phenomenon of cancer metastasis.
Dr. R. A. Reisfeld (Scripps Clinic and Research Foundation, La Jolla California) provided some perspectives for immunotherapy of cancer by presenting data obtained by him and his colleagues, Drs. Cheresh, Honsik, and Walker, with monoclonal antibodies directed to malignant melanoma and lung adenocarcinoma. Three chemically well-characterized antigens proved to be excellent targets for effective destruction of human tumors in nude mice with two modalities, i.e., specific monoclonal antibodies combined with either immunological effector cells or conjugated with chemotherapeutic drugs. Specifically, monoclonal antibody 9.2.27 (IgG2a) directed to a chondroitin sulfate proteoglycan and monoclonal antibody MB3.6 (IgG3) reactive with the disialoganglioside GD3 both preferentially expressed on human melanoma cells, effectively destroyed well-established tumors via the effector cell mechanism. Monoclonal antibody KS1/4 (IgG2a) directed to glycoprotein of 40 kDa expressed on human lung adenocarcinoma effectively destroyed such tumors when this antibody was covalently conjugated with methotrexate and injected into nude mice bearing established human lung tumors. It is anticipated that these potential cancer therapy modalities will be tested in the clinic in the near future.
On February 23 Dr. Shinzo Nishi (Hokkaido University, Sapporo) reported on the development and application of monoclonal antibodies against alphafetoprotein (AFP) and carcinoembryonic antigen (CEA). He developed hybridomas which produced antibody against AFP by the fusing of mouse immune spleen cells and mouse myeloma cells. Six antibody preparations obtained had different affinity constants and were directed to at least four different epitopes on the AFP molecule. He could develop solid-phase sandwich RIA using some combination of the antibodies. Elevated serum AFP were detected not only in hepatoma but also other liver diseases and other digestive tract cancers. Monoclonal anti-AFPs were also used for the radioimmunodetection of hepatoma which produced AFP. Positive tumor images were scanned in four of nine cases. This positivity is almost the same as that obtained with polyclonal antibody (7 of 16). He also produced monoclonal antibody against CEA. Antibodies suitable for radioimmunodetection were screened by experiments on nude mice transplanted with human tumor. Clinical studies were carried out with eight cases using the Monoclonal antibodies, and half of them gave positive images. This positivity is almost the same as that obtained with polyclonal antibodies.
Dr. David M. Goldenberg (Center for Molecular Immunology, Newark, New Jersey) reported that conventional goat and murine monoclonal antibodies against CEA, AFP, human chorionic gonadotropin (HCG), prostatic acid phosphatase (PAP), colon-specific antigen-P (CSAp), and R. A. Reisfeld's melanoma-associated proteoglycan defined by monoclonal antibody 9.2.27 have a sensitivity of about 90% in CEA radioimmunodetection. In a consecutive series of 51 colorectal cancer patients, 11 cases of occult cancer, later confirmed by other methods, were found; a lead time of up to 40 weeks was achieved. Using CEA (primary and metastatic), radioimmunodetection revealed all of the tumors while CAT scans and NMR imaging disclosed 37.5% and 50%, respectively. Image resolution of radioimmunodetection ranges between 1.5- and 2.0-cm tumors. In the case of pancreatic carcinomas, use of single-photon emission tomography with a higher dose of radioactive antibodies to CEA appears to improve imaging results. Based on animal results with I-131 anti-CEA IgG used to treat GW-39 human colonic carcinoma xenografts, showing a retardation of tumor growth and an increase in survival time, a Phase I/II clinical trial of radioimmunotherapy has been initiated.
Dr. Stewart Sell (University of Texas, Houston, Texas) presented an analysis of the morphologic and molecular alterations that occur in the livers of rats exposed to chemical carcinogens which reveals new cell populations and new gene products. Using AFP as one marker, it was possible to recognize several new cell types, including stem cells, oval cells, duct cells, and nodular cells. Autoradiographic studies indicate that there is proliferation of each of these cell types following hepatocarcinogen exposure. An effort to analyze the "premalignant" stage of hepatocarcinogenesis included: 1) development of monoclonal antibodies that recognize epitopes on different cell populations; 2) use of probes from the DNA of carcinogen-treated liver cells to identify new species of mRNA present in tumor and premalignant cells, but not in normal adult or fetal liver; and 3) application of alloantigenic markers and a P to Fl transplant system to identify and isolate donor cells from a carcinogen-treated animal after transplantation. It is anticipated that this multisystem will aid in gaining a better understanding of mechanisms by which carcinogens cause cancer.
Dr. Kohzoh Imai (Sapporo Medical College, Sapporo) reported on the monoclonal antibodies to hepatoma and their therapeutic application.
A monoclonal antibody (IgG3) to human hepatoma cell line c-Hc-4 showed ADCC activity against c-Hc-4 in vitro. Intraperitoneal injection of this antibody into nude mice unexpectedly suppressed the growth of the hepatoma cells which were transplanted subcutaneously. Monoclonal antibody MT008 (IgG3) to human colon carcinoma line, which reacts with the CEA molecule by itself, did not show any suppressive effect upon tumor growth in nude mice. Therefore, this antibody was conjugated with the toxic agent purothionin, the latter being a low molecular weight protein from barley flour that is especially toxic to dividing cells. The conjugate did not react with cultured gastric carcinoma cells KATO III, but reacted with cultured colon carcinoma cells to some extent similar to that of unconjugated antibody. Moreover, when the conjugate was given to nude mice which were transplanted intraperitoneally with a human colon carcinoma line, it significantly prolonged their survival.
(2) Seminar on "Oncogenes and Experimental Carcinogenesis"
This seminar was held on March 12-14, 1985, at the East-West Center of the University of Hawaii in Honolulu, Hawaii. The organizers were Dr. Masaaki Terada, National Cancer Center Research Institute, Tokyo, Japan, and Dr. Stuart A. Aaronson, National Cancer Institute, Bethesda, Maryland, USA. There were seven participants each from Japan and the United States. The purpose of the seminar was to discuss and exchange information on the molecular mechanisms underlying the development of cancer, with the primary goal being to exchange information on the role of oncogenes in carcinogenesis and to prepare future cooperative efforts between scientists of the United States and Japan, including exchanges of patient materials and reagents, most notably cloned oncogene probes.
In opening remarks by Drs. Terada and Aaronson, each stressed the rapid strides being made in developing important insights into the malignant processes, as well as the importance of collaborative investigations between scientists of the two countries in most effectively exploiting these new leads.
Oncogenes and Chemical Carcinogenesis
Dr. George Vande Woude (Litton Bionetics, Inc., Basic Research Program, National Cancer Institute-Frederick Cancer Research Facility, Frederick, Maryland) addressed mechanisms of activation of transforming genes. He reported that the nucleotide sequence of the Moloney murine sarcoma virus strain HT-1 (HT1MSV) differed from the cellular mos gene in three positions, but these were silent changes, and the amino acid sequence of the v-mos and c-mosm open reading frames was identical. The study of the mos oncogene product had been hampered due to extremely low levels of expression of the protein in acutely infected or morphologically transformed cells. To circumvent that problem, his group had overproduced the mos HT1MSV (equivalent to vor c-mosm) in E. coli under the control of phage!!
!promoter (pL). The E. coli p40mos thus obtained was partially purified and examined for several biochemical activities. Their findings demonstrated that the p40mos binds ATP, the analog (p-fluorosul-fonylbenzoyl[14C]adenosine), and exhibited ATPase activity. Dr. Vande Woude reported further that the human homolog of the Mo-MSV oncogene mos was ~70% conserved at both the nucleic acid and protein level. His laboratory had previously described v-mos/c-mosh hybrid constructs which contained either 5' or 3' c-mosh sequences and which transformed NIH-3T3 cells at reduced efficiencies. They now reported that constructs in which an MSV-derived LTR sequence was inserted with ~50 bp of the start of the conserved c-mosh open frame was able to morphologically transform NIH-3T3 cells at low efficiency. The foci contained small numbers of morphologically transformed cells and were only easily observed in the presence of dexamethasone, which reduced the overgrowth of the normal 3T3 cell background. Human mos-transformed cells formed tumors in nude mice at low efficiency and only small colonies in agar suspension. The transforming activity of c-mosh was decreased as the LTR was moved further 5’ to the start of the mos homologous region. Excluding a possible species-specific response, the results suggested that evolutionary divergence had resulted in a reduction but not the elimination of the transforming potential of c-mosh.
Dr. Vande Woude also described the isolation of a new transforming gene, met, from a human osteosarcoma-derived cell line (HOS) transformed with N-methyl-N'-nitro-N- nitrosoguanidine (MNNG). By molecularly cloning the met gene, it was possible to map it to chromosome 7 (7p11.4-7qter), and by this criterion and by direct hybridization it was shown to be unrelated to known oncogenes. The presence of met on chromosome 7 was of interest because several different types of human neoplasia are associated with alterations in chromosome 7, including most subtypes of acute nonlymphocytic leukemia. Further characterization of the met transforming gene and its comparison with the non-transforming allele was expected to provide insight into one of the ways in which chemical carcinogens activate proto-oncogenes in human cells.
Dr. Minako Nagao (National Cancer Center Research Institute, Tokyo) reported on the transforming genes of hepatomas and sarcomas in rats induced by environmental carcinogens IQ and 1,8-DNP. Rats fed on a diet containing 0.03% of 2-amino-3-methylimidazo[4,5-f] quinoline (IQ), a mutagen in broiled fish and meat, developed hepatocellular carcinomas, adenocarcinomas in the intestine, and squamous cell carcinomas in the Zymbal gland. Transforming genes of two original hepatocellular carcinomas were examined by transfecting tumor DNAs into NIH-3T3 cells. To transformants (IQ4-1 and IQ4-2) from a hepatocellular carcinoma (IQ4) and two transformants (IQ7-1 and IQ7-2) from a hepatocellular carcinoma (IQ7) were obtained. By transfecting 60µg of DNAs of IQ4-1 and IQ4-2, 30 and 378 secondary transformants, respectively, were induced. The transforming gene of IQ4-2 was identified to be Ha-ras. The transforming gene of IQ4-1 was neither a member of the ras family nor related to erbB oncogenes.
Dr. Nagao further described studies with 1,8-Dinitropyrene (1,8-DNP), a super mutagen in Salmonella, which is contained in diesel exhaust. All 10 rats which had received 4 mg of 1,8-DNP by subcutaneous injection developed sarcomas. Oncogenes of seven original sarcomas and one first transplant were examined by the transfection assay. DNAs of original sarcomas, 1,8-DNP-1, 1,8-DNP-2, 1,8-DNP-3, and 1,8-DNP-7, induced morphological transformants of NIH-3T3 cells by transfection. The primary transformant of 1,8-DNP-2 and its seven secondary transformants contained rat Ki-ras sequences.
Dr. Steven Reynolds (National Institute of Environmental Health Sciences, Research Triangle Park, North Carolina) described oncogene activation in spontaneous and chemically induced mouse liver cancer. He noted that a relatively low percentage of C3H x C57B1 mice, approximately 5%, spontaneously developed tumors of the liver. DNAs from both spontaneous mouse hepatocellular adenomas and spontaneous mouse hepatocellular carcinomas were capable of inducing morphological transformation of NIH-3T3 fibroblasts. Approximately 30% (3 of 10) of the hepatocellular adenomas and 77% (10 of 13) of the hepatocellular carcinomas scored as positive by DNA transfection. The transforming genes of the spontaneous adenoma transfectants were identified as Ha-ras in each case. The transforming genes of the spontaneous carcinoma transfectants were identified as Ha-ras in each case. The transforming genes of the spontaneous carcinoma transfectants were identified as Ha-ras in 8 of 10 transfectants and as a non-ras gene in the remaining two transfectants.
When C3H x C57B1 mice were fed on a diet containing furfural, an aldehyde derivative of furan used in the foundry industry, approximately 100% of the treated mice developed hepatocellular carcinoma. DNAs from 86% (6 of 7) of the chemically induced hepatocellular carcinomas were capable of morphologically transforming NIH-3T3 cells. The transforming genes of the furfural-induced hepatocellular carcinoma transfectants were identified as Ha-ras in three of the six transfectants, Ki-ras in one of the six transfectants, and as a non-ras gene in two of the six transfectants.
Human Oncogenes
Dr. Kenji Shimizu (Department of Biology, Faculty of Science, Kyushu University, Fukuoka) described novel transforming genes identified in human stomach cancer. His group had found that one of the high molecular weight DNAs from surgically removed stomach cancers gave rise to foci of NIH-3T3 cells upon transfection at very low efficiency. Five independently obtained secondary transformants retained common six EcoRI fragments containing human Alu sequences. By Alu screening of a phage library constructed with DNA of a secondary transformant, Dr. Shimizu's group had molecularly cloned almost the entire region of this transforming gene. The overall size of the gene was approximately 65 kbp. The restriction endonuclease map of this gene was clearly different from those of known human active transforming genes. Furthermore, the gene exhibited no homology to 14 viral oncogenes tested. However, tests for homology to raf, fos, rel, ski, fms, and ets had not yet been performed.
As another approach to access, Dr. Shimizu analyzed DNAs of six human cell lines established from stomach cancers. He found that a single 11-kbp EcoRI fragment, which had weak homology to Ki-ras, was amplified (5- to 10-fold) in three out of the six cell lines. When these DNAs were used as donors for transfection, slowly growing foci appeared, and cells from these foci seemed to carry a similar 11-kbp EcoRI fragment. This 11-kbp fragment appeared to be a distal member of the ras gene family. Molecular cloning of this fragment is now in progress.
Dr. Masaaki Terada reported further on transforming genes from human stomach cancers. Stomach cancers are the most common type of neoplasm, not only in Japan but in the world. His group succeeded in identifying transforming genes in human stomach cancers by transfection assay using NIH-3T3 cells.
DNAs were isolated from 21 stomach cancer tissues, 17 metastatic lymph nodes, and 21 noncancerous stomach mucosa, which were obtained at the time of surgery. His group had also tested DNAs from 14 stomach cancers transplanted into nude mice. Two samples of DNAs from stomach cancers transplanted into nude nice gave transformed foci upon transfection to NIH-3T3 cells. These primary transformants contained Alu sequences and were highly tumorigenic to nude mice. Southern blot hybridization analysis of DNAs from these transformants revealed that transforming genes were not c-Ha-ras, c-Ki-ras, or N-ras. DNAs from the secondary transformants were analyzed for the presence of Alu sequences. It was found that at least one of the transforming genes was not met, mcf2, mcf3, or Blym. One of the transforming genes was being cloned.
Dr. Terada also described amplification of c-myc and N-ras genes in stomach cancers.
Dr. Carlo Croce (Wistar Institute, Philadelphia, Pennsylvania) presented data on the molecular genetics of B cell neoplasia. He and his associates used the "chromosome walking" approach to clone the t(11;14) and t(14;18) break-points from human cell neoplasms. Interestingly, in the chromosome breakpoints on chromosome 11 in B cell malignancies carrying the t(11;14), translocation occurred within 0.9 kb. Sequence analysis of the breakpoints and of the homologous legion on normal chromosome 11 suggested that the chromosomal translocations were the result of mistakes during the process of V-D-J joining. They also analyzed the t(14;18) chromosomal translocations occurring in follicular lymphomas. The results of this analysis indicated that a similar mechanism was involved in the t(14 ;18) chromosome translocation. They identified a gene, for which they proposed the name bcl-2, which is located at band q21 of chromosome 18 and which is activated as a result of the chromosome translocations leading to malignant transformation.
Dr. Stuart A. Aaronson (National Cancer Institute, Bethesda, Maryland) discussed oncogenes of hematopoietic cells. He noted that investigations of acute transforming retroviruses had led to important insights concerning a small group of cellular genes with transforming potential. Such transduced cellular onc genes conferred to the virus properties essential for the induction and maintenance of the transformed state. Accumulating evidence indicated that these cellular genes could also be activated as transforming genes in human tumors by mechanisms completely independent of retrovirus involvement. His group had demonstrated activation of members of the ras family in a wide variety of human hematopoietic tumors. N-ras oncogenes appeared to be preferentially activated. More than 50% of AMLs tested, as well as diverse lymphomas and leukemias, had been shown to contain this oncogene.
He also described recent studies of his colleague, Alessandra Eva, in which she found that DNA of a human diffuse B cell lymphoma induced in an unusual transformed focus on NIH-3T3 cells. The transforming gene was serially transmissible, conferred the neo plastic phenotype to NIH-3T3 cells, and appeared to be larger than 20 kbp by analysis of transfectants for conserved human DNA sequences. From a cosmid recombinant DNA library of third-cycle transfectants, overlapping clones spanning 80 kbp of cellular DNA had been isolated. One clone, which contained a 45-kbp insert comprised entirely of human sequences, had been shown to be biologically active, with a specific transforming activity of 650 focus-forming units/pmol. By restriction mapping and hybridization analysis, this human transforming gene, designated dil, was shown to be unrelated to any previously reported oncogene.
Dr. Tsuneya Ohno (Jikei University School of Medicine, Tokyo) reported on a human breast carcinoma antigen and mammary neoplasias. Earlier studies had shown that the T47D human breast carcinoma cell contained an antigen which crossreacted with the 52 kDa envelope glycoprotein (gp52) of the mouse mammary tumor virus (MMTV). This cell line released particles which had the physical and biochemical characteristics of retroviruses. Particles recovered from cultures of [3H] uridine-labeled T47D clone 11 had a density of 1.18 g/ml and contained 60-70 S and 35 S RNAs associated with reverse transcriptase activity. The production of these particles was steroid dependent. By RIA, antigens crossreactive with the gp52 were found to be associated with these particles and were found in the media. This antigen, immunologically related to the gp52 of MMTV, was also detected in paraffin sections of human breast cancer by immunocytochemistry. The fact that it was the polypeptide rather than the polysaccharide portion of gp52 that was responsible for the immunological reactivity with the human breast cancer antigen added additional significance to the biological similarities between human and murine mammary neoplasias.
In attempting to produce monoclonal antibody that reacted with the MMTV gp52 and the human breast carcinoma antigen, most strains of mice, e.g., BALB/c, RIII, were found to be immunologically tolerant to gp52. To overcome this problem, immunization was performed in mice with genetic autoimmune diseases; specifically, strains NZB x NZW, F1 and MRL/1. Splenic lymphocytes of immunized mice with MMTV or gp52 of MMTV were then fused with murine myeloma cells (NS-1). As a result of screening of immunoglobulin reactivities and double cloning, IS monoclonal antibodies were chosen that demonstrated reactivities with MMTV and human mammary tumor cell line T47D and/or MCF-7. Further screening of the positive hybridoma was done by indirect immunoperoxidase staining of paraffin sections. Two hybridomas whose supernatants stained the MCF-7 and T47D cells and human breast carcinoma were cloned for further characterization. Tests to detect the human breast cancer antigen on nitrocellulose paper after SDS-PAGE were performed to define antibody reactivity. Both anti-bodies recognized one protein, weighing approximately 50 to 52 kDa, on MMTV and human materials.
A new sensitive bioassay for transforming genes based on the tumorigenicity of cotransfected NIH-3T3 cells in nude mice, which differed substantially from the NIH-3T3 focus assay, made it possible to detect the transfer of two transforming genes, about 9 kb in size, from the DNA of the T47D cell line. This line had been established from the pleural effusion of a patient with intraductal and invasive carcinoma of the breast. It had not yet been determined whether these transforming genes were associated with known oncogenes or abnormalities in T47D cells. Results from Western blot analysis with polyclonal antibody against MMTV on cell extracts of NIH-3T3 transformants with DNA from T47D cells revealed an extra band of approximately 43 kDa which had also been identified on the original T47D cell extracts.
Human ras Oncogenes
Dr. Arthur D. Levinson (Genentech, Inc., South San Francisco, California) discussed ras oncogenes and their encoded products. The mutational activation of ras proto-oncogenes associated with various tumors is well documented. To define the consequences of activating mutations on p21 polypeptide function, a variety of normal and activated p21 in E. coli and yeast were expressed. While activating mutations were found not to alter GTP binding, all such mutations examined impaired the GTPase activity intrinsic to p21, indicating that this impairment was critical for the transforming properties of the protein.
It was found that p21 was processed in yeast in a manner closely resembling that in E. coli. The protein was synthesized as a soluble precursor, which matured to a form with an increased electrophoretic mobility accompanying its translocation to a membrane component. Profound phenotypic changes in yeast expressing p21 were observed. In view of the homologies between p21 and G proteins involved in mediating the activity of adenylate cyclase, adenylate cyclase activities in yeast-producing p21 were studied and found to be elevated, although it was not established that the effect was direct.
In addition, evidence was presented that the 3' sequences of the Ha-ras transcriptional unit, mapped to the extended repeated segments, strongly affected the transformation potential of the gene by modulating the efficiency with which this gene was expressed.
Dr. Takao Sekiya (National Cancer Center Research Institute, Tokyo) presented work on the transforming activity of recombinant DNAs carrying the c-Ha-ras-1 oncogene. The c-Ha-ras-1 gene with a single point mutation at the sixty-first codon was isolated from the amplified allele of human melanoma tissue which had been maintained in nude mice. To characterize activation and expression of the gene, the DNA fragment cloned was modified in several ways, and the activity of the resultant DNAs for transformation of NIH-3T3 cells was analyzed.
(1) A chimeric gene having two point mutations at codons 12 and 61 was constructed by joining the corresponding regions from melanoma and bladder c-Ha-ras-1 genes. Transfection experiments revealed that the transforming activity of the gene with a single mutation was not affected by introduction of the second mutation. (2) A DNA fragment carrying all four exons of the c-Ha-ras-1 gene and the surrounding regions of 620 and 600 bp beyond the first and fourth codon, respectively, was revealed to lack transforming activity. The activity was restored by joining the MO-MuLV LTR to the fragment at either end. Therefore, the fragment appeared to lack some element(s) required for expression of the gene. To identify the region containing the element(s), plasmid clones having further upstream or downstream regions in different lengths were constructed. Analysis of their transforming activity was in progress.
Dr. Yasuhito Yuasa (Institute of Medical Science, university of Tokyo, Tokyo) reported on a transforming c-K-ras-2 gene of a human colon carcinoma cell line from a patient with hereditary adenomatosis of the colon and rectum (ACR). ACR is inherited as an autosomal dominant trait. Carcinoma of the large bowel arises in virtually all untreated ACR patients. DNAs of normal colonic mucosa, adenomas, and carcinomas of hereditary ACR patients were transfected onto NIH-3T3 cells to detect oncogenes associated with ACR. No transformation was observed with normal mucosa (two cases), adenoma (four cases), or carcinoma (three cases) DNAs. However, the DNA of a colon carcinoma cell line (KMS-4) from a patient with hereditary ACR was able to transform NIH-3T3 cells.
In view of evidence relating transforming DNA sequences of a number of human tumors to ras oncogenes, KMS-4 transfectant DNAs were analyzed for sequences homologous to H-, K-, and N-ras genes. Neither H-ras nor N-ras probes detected any bands other than their endogenous mouse-related bands. However, when KMS-4 transfectant DNAs were analyzed using K-ras-specific probes, each exhibited additional hybridizing bands not observed in mouse DNA. These results indicated that the transforming gene of the KMS-4 cells was an activated c-K-ras-2 gene.
Recent studies had shown that ras oncogenes could be activated in human tumors as a result of point mutations in either the twelfth codon in the first exon or sixty-first codon in the second exon of the coding sequences. To characterize the c-K-ras-2 oncogene associated with KMS-4 cells, two EcoRI fragments containing the first and second exons, respectively, were cloned using phage!!
!gtWES!!!B. Sequence analysis of these fragments were discussed. Myc and Other Oncogenes
Dr. Masabumi Shibuya (Institute of Medical Science, University of Tokyo, Tokyo) discussed amplification of cellular oncogenes in human gastric cancer. He introduced the subject by noting that gastric adenocarcinoma was one of the most frequent malignancies in Japan and some other countries. Of 17 gastric cancers maintained in nude mice, three were found to carry amplified c-myc genes. The degree of amplification was 30-fold (Shiraishi strain), 20-fold (SC-2 strain), and 5-fold (NS-3 strain), respectively. In two cases showing 20- to 30-fold amplification, double minute chromosomes had been observed in karyotype analysis.
In this study, the following points were examined: (1) Was gene amplification associated with qualitative changes in c-myc exon DNA? (2) Did the level of c-myc RNA directly correlate with the degree of amplification? (3) How important was c-myc amplification for tumor formation?
Amplified c-myc DNA was molecularly cloned from Shiraishi cells, and the coding region was sequenced. One base change was found at the third letter for a codon in the second exon, but the deduced amino acid sequence was essentially the same as that of the normal c-myc gene. Total cellular RNA was extracted from several gastric cancers and the level of c-myc RNA examined by Northern blot analysis. The Shiraishi cells, derived from a rapidly growing and poorly differentiated adenocarcinoma, showed 10- to 20-fold higher levels of c-myc RNA. On the other hand, SC-2 cells, derived from a slowly growing and more differentiated type of gastric cancer, showed a level of c-myc RNA lower than that expected from the degree of gene amplification. The authors suggested that the decrease of c-myc RNA in SC-2 was due to differentiation and reduction of cell division of tumor cells. A possible role of the transcriptionally activated c-myc gene for cell transformation was tested using a rat embryo fibroblast system. The results obtained were consistent with the "two step theory," described by Weinberg's group, but the fibroblasts transformed by LTR-myc with the activated ras gene were more malignant in vivo than those reported.
Dr. Takis Papas (National Cancer Institute, Bethesda, Maryland) reported on myc-related oncogenes and their gene products. The authors had elucidated the structure of the cloned 5.2-kb integrated DNA provirus of MH2 and a major portion (3.5 kb) of the 3' region had been sequenced. The complete genetic structure had been determined and found to be 5'-&Mac198;gag (1.9 kb)-mht (1.2 kb)-myc (1.3 kb)-[noncoding] c-region (0.2 kb)-3'. Aside from the genetic elements shared in common with other nondefective avian retroviruses, the mht gene was found to be a unique oncogene sequence. Hybridizations revealed that mht was derived from a normal cellular gene distinct from the proto-myc family of oncogenes. The &Mac198;gag-mht formed a hybrid gene, containing a 2682-nucleotide contiguous reading frame with a stop codon near the 3, end. The 3' region of mht was 90% sequence-related to the onc-specific raf sequence of the MSV 3611 and 95% homologous at the deduced amino acid sequence level, the closest homology determined so far of the 19 known onc sequences. Using a unique expression vector, pJL6, the carboxy terminal portion of the avian MC29 v-myc oncogene had been cloned and expressed as a fusion protein in bacteria as more than 10% of the total cellular protein. This expression product enabled the preparation of antibodies raised against it; these antibodies immunoprecipitated the MC29 oncogene product P110gag-myc. Employing the same expression system, an Ha-MSV p21ras oncogene product prepared as a fusion protein produced more than 10% of the total protein. This protein exhibited GDP binding activity and was capable of autophosphorylation, similar to authentic Ha-MSV p21ras. Two reciprocal recombinant sites between c-myc and the immunoglobulin heavy chain µregion in a Burkitt lymphoma had been sequenced and characterized, showing that the onc gene had been interrupted within its first intron region and joined to the immunoglobulin heavy chain µ switch region.
Dr. Lewis T. Williams (Howard Hughes Medical Institute, University of California, San Francisco, California) reported on the role of platelet-derived growth factor (PDGF) receptors in normal and transformed cells. The interaction of PDGF with its receptors initiates a number of cellular responses including activation of tyrosine kinase, release of diacylglycerol from phosphotidylinositol, and the stimulation of transcription of the c-myc and c-fos genes. The PDGF receptor was purified from BALB/c 3T3 cells using antiphosphotyrosine antibodies that recognize phosphotyrosine present on PDGF-stimulated receptors. The purified 180 kDa receptor is a glycoprotein that binds [125I]PDGF to intact 3T3 cells. When activated by PDGF, the receptor rapidly stimulated protein kinase C activity, presumably as a result of the release of diacylglycerol from phosphotidylinositol. After regulation of protein kinase C activity by chronic exposure to phorbol esters, the stimulation of the myc gene was dramatically diminished, although the ability of PDGF to stimulate DNA synthesis and cell proliferation was unaffected. Thus, there is a component of the action of PDGF that is independent of the protein kinase C and myc gene pathways.
Cells transformed by the simian sarcoma virus (SSV) released peptides that could bind to PDGF receptors, activate PDGF receptor tyrosine kinase, and stimulate DNA synthesis. These peptides were presumably encoded by the sis oncogene. Normal cells cocultured with SSV-transformed cells lost their PDGF receptors due to receptor internalization and metabolism (down regulation). Cultured SSV-transformed cells had no detectable PDGF receptors. However, receptors on their cells could be unmasked by treating the cells with suramin to dissociate the PDGF-like peptides from the receptor sites, thereby preventing receptor down regulation. Thus, PDGF receptors on SSV-transformed cells were activated and then down regulated by endogenous PDGF-like material produced by these cells.
Drs. Aaronson and Terada noted in their closing remarks the enormous progress that has occurred in only the past 2 or 3 years in focusing such diverse fields as carcinogenesis, growth factor research, and tumor virology on common questions related to the neoplastic process. Each stressed the high quality of the presentations during the seminar, as well as the excellent opportunity for informal discussions among the participants. Finally, each expressed the hope that the discussions initiated at the meeting would foster continued meaningful interchanges of ideas and collaborations.
There being no further business, the meeting was adjourned at 12:00 noon, March 14, 1985.
SEMINAR AGENDA AND PARTICIPANTS
(1) SEMINAR ON ANALYSIS OF TUMOR-ASSOCIATED ANTIGES OF DIGESTIVE ORGANS WITH MONOCLONAL ANTIBODIES AND THEIR CLINICAL APPLICATION
La Jolla, California, February 21-23, 1985
AGENDA