REPORTS ON SEMINARS

(1) Seminar on “Development of New Medium-Term Bioassays for Carcinogens”
This seminar was held on December 15-17, 1987 at the Sheraton Princess Kaiulani Hotel in Honolulu, Hawaii. The organizers were Dr. Jerrold M. Ward, National Cancer Institute, Frederick, Maryland, USA and Dr. Nobuykui Ito, Nagoya City University Medic School, Nagoya, Japan. There were seven participants from each nation. The purpose of the seminar was to discuss and exchange Information on the development of new medium-term bioassays from both nations. In addition, a few selected older medium-term bioassay systems were re-evaluated in light of recent experimental findings. These bioassays hopefully could be useful in detecting carcinogens and tumor promoters, which may be presumed carcinogens, in a relatively shorter period of time (weeks to months) than for the standard rodent 2-year carcinogen bioassay. Thus, both time and money could be saved.
In opening statements by Drs. Ward and Ito, emphasis was placed on the present availability of current medium-term bioassays and need for further development of selected models. Dr. Ito noted that as the number of environmental chemicals tested for mutagenicity has increased, it has become clear that mutagenicity results did not always show a direct correlation rodent carcinogenicity. He also pointed out that for two-year rodent carcinogenicity bioassays, costs have become very high and large numbers of animals must be used. Therefore, medium-term bioassays possibly could bridge the cost and time gap between mutagenicity tests and long-term bioassays. The present seminar was thought to be an excellent forum for exchange of information on current efforts in this field by scientists in the USA and Japan.

Overview and Definitions
Dr. James Huff, National Institute of Environmental Health Sciences, Research Triangle Park, North Carolina, reported on “The National Toxicology Program’s (NTP) Long-Term and Short-Term In Vivo Assays for Chemical Carcinogenesis”. For the long-term studies, he noted that about 7 years are required from chemical selection to reporting of the two-year carcinogenicity studies. Currently, the costs of a 2-year feeding study are about one million dollars and is close to two million dollars for an inhalation bioassay. This includes exposing both sexes of two species for 2 years. Nearly 350 chemicals have been evaluated for cancer-causing potential by the NTP and the National Cancer Institute since the early 1970s; close to 50% of these have been shown to be carcinogenic in at least one of the four sex-species experiments. Only about 40 (11%) caused cancer in each of the four experiments. The liver (especially of the mouse) was the most common target site followed by lung, mammary gland, and hematopoietic system.
Largely because of the high costs and extensive time required to evaluate a chemical for toxicologic and carcinogenic effects, the NTP has sought shorter in vivo (short and medium-term) assays for consideration to use in concert with in vitro assays. So far, these have included the Strain A mouse lung tumor assay, in vivo rat liver models (neonatal and partial hepatectomy), preweanling mouse liver model, mouse bone marrow cytogenetics assay, and peripheral blood erythrocyte micronucleus assay. The mouse lung tumor assay was discussed by Dr. Stoner (see below). The rat liver models are being further refined by NTP through contracts to research institutes. Chemicals are studied in two rat liver models at present and endpoints evaluated and compared to the long-term results. Sprague-Dawley rats were found to be most sensitive for use in studying the initiating or promoting mechanism of a specific chemical in liver. Initial results suggest that assay findings may depend on chemical class. The bone marrow assay revealed good correlation between the mutagenicity of the chemical in the Ames test and its carcinogenicity or positive micronucleus test. The preweanling mouse liver assay was valuable for producing liver tumors with activated oncogene mutations, which were often chemical specific.
Dr. Thomas J. Slaga, University of Texas Cancer Center, Science Park-Research Division, Smithville, Texas, discussed “Skin Initiation-Promotion and Carcinogenesis Assays.” After presenting basic definitions used in initiation-promotion assays, he agreed with seminar participants that almost all organ-specific promoters without genotoxic activity and detected in skin or other organ specific assays, were probably also carcinogenic in the same organ in 2-year studies, albeit frequently to a “weak” degree. This concept was thought to be very important because any medium-term bioassay could be used to detect promoting activity of a chemical, which would could be interpreted as evidence of carcinogenicity. The SENCAR inbred and outbred mouse skin tumor models were described. The advantages of these mice include increased sensitivity to skin tumor promotion and shorter latent periods for tumor development. Dr. Slaga reviewed a large data base of chemicals tested on skin for initiating or promoting activity. Initiators were defined as chemicals given at subcarcinogenic doses although only one pure initiator, a diol-epoxide, may have been found. Chemicals requiring metabolism in other organs may not reveal initiating or promoting activity in mouse skin, including nitrosamines, aromatic amines and liver tumor promoters.
Dr. Shigeru Yoshizawa, National Cancer Center Research Institute, Tokyo, described a “Short-Term Screening System for New Tumor Promoters.” Twenty-three new non-TPA type tumor promoters including palytoxin, thapsigargin, okadaic acid, dinophysistoxin-1 and staurosporine were found in assays designed to detect these new chemicals. These assays included skin irritation test on mouse ear, induction of histamine decarboxylase or ODC in mouse skin, stimulation of prostaglandin E2, arachidonic acid metabolism and superoxide anion radical production. The non-TPA type promoters did not inhibit TPA binding or induce protein kinase C or HL-60 cell adhesion, while TPA-type promoters did. They all promoted skin tumors, however, after skin initiation in CD-1 mice. Discussion concerned the biochemical basis for skin irritation in relation to tumor promotion. It was generally agreed that most skin promoters were irritants but not all irritants were promoters.

Rodent Liver Assays
Dr. Hiroyuki Tsuda, Nagoya City University Medical School, discussed “An Approach For A New Rapid Bioassay System For Hepatocarcinogens By Enhancement of DEN-Initiated Placental Form of Glutathione S-Transferase Positive Foci in Rat Liver.” The model system includes the injection of DEN (200 mg/kg) ip at 6 weeks of age, starting the test substance 2 weeks later, with partial hepatectomy at week 3 and continuous feeding of the test substance for an additional 5 weeks at which time the rats are sacrificed for evaluation of GST-P positive liver foci (number and area per unit area). Studies demonstrated that the GST-P positive foci progressed to carcinomas. In an extensive series of more than 130 chemicals known liver carcinogens were 89.7% Positive (26/29) in this assay irrespective of known mutagenicity while known non-liver carcinogens were seldom positive (2/21) and no known non-carcinogens were positive (0/25), indicating this assay system does not suffer from false positive results. Interestingly, 21.8% (12/55) of chemicals which have not been tested sufficiently (but are suspect carcinogens or important environmental chemicals) in 2-year bioassays were positive. The false negative known carcinogens included peroxlsomal proliferators, di(2-ethylhexyl)phthalate (DEHP) and clofibrate, and were due to negative staining for GST-P of the induced lesion. This assay offers a valuable model for screening a large number of chemicals in very short in vivo model (8 weeks).
Dr. James Popp, Chemical Industry Institute of Toxicology, Research Triangle Park, North Carolina, described “Comparative and Quantitative Aspects of Rat Liver Medium Term Assays.” He described and compared the rat liver models of Pitot et al., Shinozuka et al., Cayama et al., and Solt et al., noting that all model systems induced liver foci, although not equally, perhaps because of differences in the biochemical effects (P-450 induction) during the promotion phase. Several important differences and considerations were discussed which can affect interpretation of the findings for test chemicals in these assays. Liver lobe differences were described which sometimes depended on the chemical tested. It was noted that if both initiating and promoting activity could be demonstrated for a chemical in these assays, the chemical would surely be carcinogenic, although the presence of either activity may also be indicative of carcinogenic activity in 2-year bioassays. Studies on dinitrotoluene isomers revealed that minor contaminants were sometimes the more potent inducers of liver foci. Also, in a 15-week assay with the peroxisomal proliferator, Wy-14,643, only ATPase-deficient foci were promoted to grow larger but not increase in number. Thus, some promoters increased numbers of foci, while others promoted focus growth (perhaps, a late stage promoter). In order to properly evaluate these growth effects, stereology must be performed.
Dr. Jerrold M. Ward, National Cancer Institute, Frederick, Maryland, described “The Mouse Liver As An Endpoint In Initiation-Promotion Assays For Detecting Carcinogens.” Since the mouse liver was the major endpoint of 2-year bioassays conducted by the National Toxicology Program (NTP) in the United States, medium-term bioassays were developed for evaluation. The pathology of mouse liver tumors was reviewed in relation to long-term and medium-term assays. Injection of infant B6C3F1 mice (15 days of age) by experiments of Diwan et al. and others have shown that phenobarbital following DEN injection does not have promoting activity in mouse liver. Preliminary studies suggest that DEN affects phenobarbital (PB) pharmacokinetics, while other initiating agents do not. Thus, an assay with DEN injection at 4 weeks of age was successful in initiating foci promoted by PB. PB increased the numbers of DEN-initiated foci, while DEHP promoted growth of the foci, especially at the higher doses used. Using this assay which can detect promoted foci by 12-24 weeks, diazepam, oxazepam and BHA were shown to be promoters for mouse liver. Others have demonstrated that chlordane, heptachlor and carbon tetrachloride also have promoting activity. A discussion of the use of Zeiss (Pugh et al. and Campbell et al). stereology programs noted differences in mean values but not in interpretation of findings in comparison with control groups. The most effective marker was noted to be H&E stained liver foci. Other models using partial hepatectomy (PH) or chronic liver toxicity were reported. PH was noted to induce biliary hyperplasia because the gall bladder was removed. This problem and others may make PH of limited use in mice. Induction of hepatic enzymes by potential promoters and carcinogens was discussed as possible markers for these chemicals. In a 2-week assay, P450 induction could be shown for specific chemical classes which correlated with tumor promoting potential. Another new model was described whereby aged 12 month old male C3H mice (with high spontaneous incidence of liver tumors and presumably spontaneously initiated hepatocytes) were exposed to PB and other promoters without chemical initiation. The mice responded quickly within 12-24 weeks to induction of foci and adenomas. These medium-term mouse liver bioassays offer great promise for detecting liver carcinogens.

Multiple Organ Assays
Dr. Nobuyuki Ito, Nagoya City University Medical School, Nagoya, discussed “Development of A Medium-Term Multi-Organ Bioassay for Carcinogens By A Wide Spectrum Initiation Protocol.” Dr. Ito reviewed the preneoplastic and precancerous histopathologic lesions found in all rat organ systems including phenotypic markers for each lesion. These lesions and markers could be used in short-term and medium-term bioassays as experimental endpoints. Each was associated with tumor formation and was a valid representation for evaluation of the potential tumor promoting and carcinogenic potential of a test chemical. Three multi-organ combined initiation models were described. Generally, groups of 20 rats could be effectively used to evaluate the promoting or carcinogenic activity of test chemicals. In NMU Model, NMU (20 mg/ kg) was used as an initiator with 8 doses in 4 weeks, followed by the test chemical for 12 weeks. In Combined Model I, rats received DEN (100 mg/ kg) once at 6 weeks of age, then NMU (25 mg/ kg, 4 doses in 2 weeks) and finally DH PN (0.1% in the water for 2 weeks), followed by the test chemical for 14 weeks. In Combined Model II, rats received DHPN (1000 mg/ kg twice in one week), EHEN (1500 mg/ kg twice in one week) and DMAB (75 mg/ kg twice in one week) followed by the test chemical for 12 weeks. For all models, by 20-24 weeks, liver foci and preneoplastic lesions in thyroid, lung, forestomach, urinary bladder and esophagus were evaluated by quantitative methods. In these assays, all known carcinogens tested were always positive in the respective organotropic site. Discussion concerned the use of single agent initiators, including NMU and multi-organ histopathology required. These combined initiator models appear to provide promising assays for detecting carcinogens of unknown organ specificity.
Dr. Stan D. Vesselinovitch, University of Chicago, Chicago, discussed “The Infant Mouse-Liver System As Potential Medium-Term Carcinogenicity Bioassay Model.” The perinatal age was shown to be the most sensitive period during which various organs can undergo carcinogenesis. Leukemia originating in the thymus, kidney and liver tumors are most readily induced in infant mice by single injections of one of several carcinogens, especially alkylating agents, while lung tumors may be more readily induced in adult mice. Leukemias may be induced within 15-25 weeks after birth, while liver tumors may appear later. DEN is the most potent hepatocarcinogen in this infant model. The kinetics of DEN hepatocarcinogenesis was reviewed. It was noted that induction of liver foci alone should not be the basis for characterization of a test chemical as a carcinogen since only a small proportion of the foci may progress to carcinoma. The importance of stereology was also emphasized. This model should be of value for akylating agents, especially those with mutagenic activity.
Dr. Hideki Mori, Gifu University School of Medicine, Gifu, described “A Medium-Term Assay of Carcinogens or Carcinogenesis Modifiers in Hamster Models: Application of Early Changes in The Liver and Preneoplastic Mesenchymal Cells in the Stomach.” Hamster medium-term models may be of specific use in carcinogenesis assays. Certain species differences have been described. While initiation of hepatocarcinogenesis by DMN or MAM is evident, promotion by phenobarbital is not, despite hepatic enzyme induction similar to that induced by tumor promoters and shown by others. Typical hamster liver foci are readily induced and can be enhanced by carbon tetrachloride. It remains to be shown that other carcinogens or promoters may also enhance foci in hamster liver. In a new medium-term hamster stomach assay, MNNG, a gastric carcinogen, was shown to induce spindle cell proliferation in the submucosa prior to tumor induction. Gastric promoters, such as NaCl or Tween 60, enhanced the occurrence of the proliferation. This new system may prove of value for detecting direct-acting agents and modifiers of gastric carcinogenesis.

Stomach and Lung
Dr. Masae Tatematsu, Nagoya City University Medical School, Nagoya, described “A Trial of A Medium-Term Bioassay for Gastric Carcinogens Based On Quantitation of Putative Preneoplastic Pepsinogen-Decreased Pyloric Glands.” He described the occurrence of pepsinogen isozymes (Pg) 1-4 in different regions of the stomach and preferential deficiency of Pg 1 in pyloric glands during MNNG gastric carcinogenesis. Pepsinogen 1-decreased pyloric glands (PDPG) were found from week 10 using immunohistochemistry and were associated with carcinoma development. Of five rat strains studied, SD, WKY and Lewis were most susceptible PDPG was dose-dependent and associated with an increased labeling index as measured by double Brdu PDPG-immunohistochemistry. A gastric carcinogenesis model was developed using a single dose of MNNG (160 mg/ kg) ig, as initiator. Induction of PDPG was enhanced by ENNG from week 8, and by sodium taurocholate and sodium chloride, a promoter of gastric carcinogenesis, from week 16. In the MNNG model, catechol enhanced the development of PDPG, as well. The use of the phenotypic marker, PDPG, is a useful tool for evaluating gastric carcinogenesis and promotion.
Dr. Hiroko Ohgaki, National Cancer Center Research Institute, Tokyo, reported on “Cell Proliferation, A Key Factor Determining Susceptibility to Gastric Carcinogenesis In Rats.” A unique medium-term bioassay for gastric carcinogens or promoters involved F344 rats give MNNG, catechol, tauro-deoxycholic acid, taurocholic acid, or NaCl for only 4 weeks. At 4 weeks, the rats received Brdu and their stomachs were prepared for Brdu immunohistochemistry. The treatments increased the size of the proliferative zone of the pyloric epithelium of glandular stomach and number of cells labeled with Brdu. In order to study the possible mechanisms involved In susceptibility and resistance to gastric carcinogenesis in different rat strains, a study with [3H-methyl]MNNG revealed similar numbers of labeled cells in susceptible ACI rats and resistant Buffalo rats and FI hybrids. Double labeling, with [3H-methyl]MNNG and Brdu, however, demonstrated a greater number of cells labeled with Brdu and double Brdu labeled cells(hyperplasia) in ACI rats even only after 2 weeks of feeding of MNNG in the drinking water than in Buffalo and F1 rats. Thus, cell proliferation was a marker of susceptibility and tumor promoting or carcinogenic activity in the rat stomach.
Dr. Gary D. Stoner, Medical College of Ohio, Toledo, discussed “The Strain A Mouse Lung Tumor Bioassay.” In a review and re-evaluation of this classic medium-term bioassay, he reported on the previous mouse strain and dose response mouse strain studies and most recent experiments on oncogene activation in mouse lung tumors. Specific mutations of K-ras codons 12, 13, 61 or 117 have been found in association with specific carcinogens. In the 24-30 week Strain A mouse lung tumor bioassay, more than 350 chemicals from many chemical classes have been tested. The potency of chemicals increasing the numbers of lung tumors in this assay varies greatly, with the polycyclic aromatic hydrocarcbons, methylcholanthrene and DMBA being the most potent and the aminofluorenes and aromatic amines, the least potent. Chemotherapeutic agents, alkyl halides, and carbamates were positive sometimes depending on chemical structure relationships. In a validation study of the lung tumor assay for comparison with 2-year carcinogenesis bioassays by the National Toxicology Program, a poor correlation was found for many chemicals, especially for aromatic amines. Liver carcinogens were also less potent lung tumor inducers than might be expected. Dr. Stoner noted that this assay offers a cost effective method for screening chemicals of many classes with the consideration that false negatives can be seen.

Pancreas and Urinary Bladder
Dr. Daniel S. Longnecker, Dartmouth Medical School, Hanover, described “The Rat and Hamster Pancreas Models.” The azaserine rat model uses Wistar or Wistar/Lewis rats and atypical acinar cell nodules (AACN), precursors of acinar cell carcinomas, as the phenotypic endpoint in only 4 months after a single injection of carcinogen. The AACN in large defined areas of the pancreas are quantitatively measured by stereology (#/ cm³, focus size and volume %), especially the acidophilic foci which are the primary precursors of carcinoma. Using this assay, corn oil, soya flour, camostate and testosterone have been shown to be promoters of AACN in rats. The hamster model using the nitrosamine BOP at weeks 6,7 and 8 was reported. In addition to AACN, other phenotypic markers used as carcinoma precursors, especially of the pancreatic ducts, were described. These included tubular ductal complexes and intraductal hyperplasias. These lesions were increased by feeding of high lard diets. It was suggested that both rat and hamster models be used to screen possible pancreatic acinar cell tumor promoters. Some chemicals, including antioxidants and retinoids, have inhibited AACN in rats. Initiators may be detected because they induce the precursor lesions in the absence of azaserine or BOP.
Dr. Ryohei Hasegawa, National Institute of Hygienic Sciences, Tokyo, described “Rapid Bioassay Methods For Bladder Carcinogens and Promoters.” Bladder lesions were reviewed including the preneoplastic marker, papillary or nodular (PN) hyperplasia and papillomas. Histochemical markers were investigated and the H&E PN hyperplasia and papilloma were shown to be the most effective markers for carcinoma prediction. The standard medium-term model involves feeding of BBN in the water for 4 weeks followed by the test chemical for up to 32 weeks. Two new modifications included induction of urothelial proliferation by unilateral ureteric ligation (UL), partial cystectomy or uracil ingestion. Uracil induces bladder stone formation and epithelial hyperplasia. Several chemicals were shown to be bladder tumor promoters including BHA, sodium saccharin, sodium OPP, sodium L-ascorbate and DBN. With the use of uracil as a 3 week exposure during test chemical administration, the bioassay may only be of 20 week duration, a distinct advantage. This model offers advantages for testing of potential urinary tract carcinogens and promoters, especially chemicals with urinary metabolites.
Drs. Ward and Ito noted in their closing remarks that the seminar had been extremely fruitful. Much new data and models had been presented as well as a thorough re-review of previous models for each organ system. It was noted that several medium-term bioassays are now presently available for screening chemicals with unknown tumor promoting or carcinogenic activities. In particular, the multi-organ initiation, infant mouse, skin and Strain A lung tumor assays could be used for screening chemicals of unknown organ specificity. The liver models of rats and mice were well described and may be used for screening since the liver is the primary target site of environmental carcinogens in 2-year bioassays. The stomach, pancreas and urinary bladder assays could be used for selected chemicals. Cell proliferation as a marker for stomach and other organs should be further investigated. If a chemical with unknown carcinogenic or tumor promoting activity is detected as positive in one of the bioassay: described in the seminar, the chemical could be considered a suspect carcinogen and/or require further evaluation until making a final determination. At the least, a chemical which turns up positive in these assays is, at the minimum suspect as a promoting agent These points were discussed and noted for future studies

(2) Seminar on “Cancer Cell Membranes: Aberrant Glycosylation and Other Critical Molecular Events”
This seminar was held on February 15-18, 1988, at the Sheraton Makaha Resort Hotel, Makaha Valley, Oahu Island, Hawaii. The organizers were Drs. Sen-itiroh Hakomori (The Biomembrane Institute and University of Washington, Seattle, Washington, USA) and Setsuo Hirohashi (National Cancer Center Research Institute, Tokyo, Japan). The seminar was focused on aberrant glycosylation and other molecular changes expressed at the surface of cancer cells and defined by monoclonal antibodies. There were eight participants and seven observers from the US and the same numbers of participants and observers from Japan. This was the first seminar under the US-Japan Cooperative Cancer Research Program devoted to this topic, and led to many stimulating discussions and a productive exchange of ideas. Participants arrived on February 15 and enjoyed a no-host cocktail get-together followed by dinner on that evening.

February 16 (morning session)
After brief opening remarks by Dr. Hirohashi, general information on the historical background of the US-Japan Cooperative Cancer Research Program was presented by Dr. Takashi Sugimura, President of the National Cancer Center, Tokyo, Japan, who established this binational seminar program over 15 years ago. Next, in his introductory remarks Dr. Hakomori described the significance of aberrant glycosylation expressed at the tumor cell surface in relation to general trends in cancer research. Some areas of studies that have aroused major interest in current cancer research are: (i) aberrant glycosylation expressed at the tumor cell surface; (ii) molecules defining tumor cell adhesion, the mechanism of which is important for understanding the metastatic potential of tumor cells; (iii) regulation of cell growth by growth factors and their receptors, particularly autocrine-dependent cell growth, the mechanism of which is important for understanding uncontrolled outgrowth of tumor cells; and (iv) transforming genes (oncogenes) triggering cascade reactions, which lead to the expression of a large variety of tumor cell phenotypes. This seminar was devoted to the first two topics, which have become recognized as increasingly important yet have not previously been the subject of a binational cancer seminar.
The regular program followed subsequently. The first section, chaired by Drs. Victor Ginsburg (National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, Maryland) and Yoshitaka Nagai (University of Tokyo School of Medicine, Tokyo), addressed tumor-associated aberrant glycosylation defined by monoclonal antibodies.
Although many tumor-associated antigens defined by monoclonal antibodies have been identified as glycolipids, many of the same epitopes are also carried by mucin-type glycoproteins, as initially described by Dr. John Magnani and his associates (National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, Maryland). Some epitopes are carried exclusively by mucin-type glycoproteins. Synthesis of a well-established epitope, sialyl Lea, expressed in glycoproteins, is competitively converted to Le^b, i.e., al 2 fucosylation precedes a2 3 sialylation of type 1 chain, in advance of al 4 fucosylation. Thus, the incidence of sialyl Le^a expression in tumors decreases. False negative cases in serum assays for sialyl Le^a in some patients with cancer may not only result from Le(a-b-) individuals, but may also result from partial conversion of the type 1 chain precursor to type 1 chain H and to Le^b. This suggestion, proposed by Dr. Magnani, may also be valid for false negative results for sialyl Le^x as a target antigen for serum diagnosis. Previously, Werkmeister et al. (J. Immunol. 135:689, 1985) observed that an IgM monoclonal antibody (LeoMel 3) directed to human melanoma could block IL-2-stimulated LAK cell killing. Dr. Magnani also presented data showing that LeoMel 3 antibody is directed to GD₃ ganglioside. Thus, LAK cells are capable of recognizing GD₃ at the target cells.
Currently, a number of monoclonal antibodies selected by preferential reaction with various human cancers over normal tissues have been shown to be directed to unknown carbohydrate epitopes carried by mucins. Dr. Ikuo Yamashina (Kyoto University, Kyoto) presented a general method for producing monoclonal antibodies directed to glycoprotein carbohydrates, i.e., immunization of mice with the tumor cells, followed by selection of hybridomas with the pronase-digested glycopeptide-coated on a plastic surface. In this way, he has been able to isolate a series of three classes of antibodies. Class I specifically defines sialyl Tn (NeuAca2 6GalNAcal Ser/Thr), class II reacts with both Tn (GaINAcal Ser/Thr) and sialyl Tn, and class III shows a reactivity similar to that of N-19-9, directed to a2 3sialyl Le^a, but also reacts with a2 3 type 1 chain without a2 4 fucosyl residue. He claimed that class II antibodies are suitable for imaging, while class III antibodies are suitable for serum diagnosis.
It has become increasingly clear that fucosylated and sialylated-fucosylated poly-N-acetyllactosamines constitute the major tumor-associated carbohydrate antigens occurring in a large variety of human cancers. One such antigen, sialyl Le^x-i, defined by monoclonal antibody FH6, was previously established as a good lung adenocarcinoma marker, and the level of the antigen has been shown to be useful as a diagnostic aid for various human cancers, particularly human lung cancer, as previously reported by Drs. Reiji Kannagi (Kyoto University School of Medicine, Kyoto, Japan) and Yasuo Fukushi (Tohoku University School of Medicine, Sendai, Japan). However, expression of the same antigen in fetal lung had not been investigated in detail. Dr. Kannagi presented data from systematic studies on oncodevelopmental changes in expression of Le^x, Le^y, and sialyl Le^x in lung tissue. Le^y antigen was found to be highly expressed in bronchial buds, while Le^x was found to be expressed in bronchiolar buds and sialyl Le^x in alveolar buds (terminal buds), although these antigens regressed and disappeared on completion of development of lung tissue. Thus, retro-ontogenesis of the expression of these antigens at a well-defined stage of lung tissue development has now beer visualized. Kannagi also demonstrated that sialyl Le^x-i seems to be a marker for human NK cells. This seemingly exciting result needs to be supported by functional data.
Diagnostic application and functional aspects of normal differentiation antigens defined by monoclonal antibodies were discussed by Dr. Setsuo Hirohashi (National Cancer Research Center Institute, Tokyo). The antibody NCC-CO-450, which reacts with normal carbohydrates associated with mucin-like molecules, has been shown to be useful for serological diagnosis of cancer patients. The data clearly indicate that an antigen showing higher levels in sera of cancer patients than in sera of control patients may not necessarily be tumor specific. Any normal tissue antigens could be useful as markers in serum diagnosis if they are not released into the bloodstream under normal conditions but are released into sera of cancer patients; e.g., exocrinous antigens in normal secretory tissue would become endocrinous in tumor tissue. Similar examples are acid phosphatase in prostate cancer and sialyl Le^a antigen in pancreatic cancer. Another antibody discussed by Dr. Hirohashi has been shown to disrupt cell-cell interaction in human squamous cell carcinoma. The target molecule seems to be acalcium-sensitive adhesion molecule similar to “cadherine “ reported by M. Takeuchi of Kyoto University.
The idea that tumor-associated epitope structures may not be unique, but that antibody specificity may be directed to such epitopes uniquely organized at the tumor cell surface, is supported by studies presented by Dr. Masahiko Watanabe (Keio University School of Medicine, Tokyo). The antibody NCC-ST-421, raised against a gastric cancer xenograft, belongs to the IgG3 subclass and recognizes Le^a as an epitope, but does not react with human erythrocytes regardless of Lewis blood group status. However, this antibody showed lysis of human cancer cell lines expressing Le^a antigen in the presence of human lymphocytes and serum. In vivo treatment with NCC-ST-421 administered intraperitoneally completely inhibited the growth of human cancer subcutaneous xenografts expressing Lea antigen in nude mice.
Five antigens, lactosylceramide, paragloboside, Le^x, dimeric Le^x, and sialyl dimeric Le^x, expressed in renal cancer and in fetal kidney tissues at various stages of development, have been studied employing monoclonal antibodies T₅A7, 1B2, FH2, FH4, and FH6, respectively, by Dr. Yasuo Fukushi (Tohoku University School of Medicine, Sendai). The expression of these antigens has been correlated with the histological grade of renal carcinoma. In addition, expression of Gal^$14GlcNAc defined by 1B2 and sialosyl difucosyl Le^x defined by FH6 was found to be related to prognosis of the patient. The cases expressing FH6 antigen showed a more favorable prognosis. Thus, examination of carbohydrate antigens in renal tumors may provide a good criterion for the degree of malignancy and prediction of patient prognosis.
Precursor accumulation is one of the most common carbohydrate changes occurring in various human cancers. Dr. Masao Iwamori (University of Tokyo, Tokyo) discussed a human monoclonal antibody derived from the lymph node lymphocytes of a patient with endometrial cancer. The antibody defining endometrial carcinoma and germinal proliferation was found to be directed to lactotetraosylcearmide, the precursor of various type 1 chain blood group antigens. To support this finding, he analyzed the glycolipid profile of fetal and adult intestine. The type 1 chain precursor (Lc₄Cer) and its sialylated derivative (IV³NeuAcLc₄Cer) were absent in adult intestine but were present in significant quantities in fetal intestine.
In view of the increasing importance of disialosyl gangliosides as melanoma- and neuroblastoma-associated antigens, Dr. Tadashi Tai (Tokyo Metropolitan Institute of Clinical Science, Tokyo) has scrutinized six monoclonal antibodies directed to GD₂ and GD₃ in order to define the fine specificities of the antibodies recognizing disialosyl structures. The first type of antibody reacts specifically with GD₂ with NeuAc2 8NeuAc residue. The second type reacts not only with GD₂ but also with several other gangliosides having the sequence NeuAc2 8NeuAc2 3Gal. The third type of antibody reacts with GD₂ with lactones. Thus, it is increasingly apparent that a group of antibodies may define a similar but slightly different structure having a common disialosyl residue.

February 17 (morning session)
The section on the functional significance of glycosylation was chaired by Drs. Garth Nicolson (M.D. Anderson Hospital and Tumor Institute, Houston, Texas) and Toshiaki Osawa (University of Tokyo, Tokyo).
A great deal of interest has been focused on gangliosides as protein kinase modulators, in relation to the physiological function of gangliosides in growth regulation and induction of differentiation. During the past few years it has become increasingly apparent that some gangliosides modify protein kinase activity. Two papers were discussed in relation to this phenomenon. Dr. Yoshitaka Nagai (University of Tokyo, Tokyo) previously found that GQ_1b induces differentiation of human neuroblastoma cells and promotes neurite outgrowth. The differentiation-inducing activity of GQ_1b is extremely high (order of a few nanomoles to observe inductive activity), and the structure of GQ_1b is strictly required. Furthermore, exogenous addition of GQ_1b promotes various protein kinase activities. Based on these previous findings, he has now observed that the target molecule(s) stimulated by GQ_1b seem to be 60 and 64 Kd proteins phosphorylated by cell surface protein kinases (called “ecto-Gg kinases”). Thus, a new type of carbohydrate-dependent biosignal transduction system is proposed.
Cell growth inhibition is induced in vitro by exogenous addition of GM₃ or GM₁, a phenomenon well established 15 years ago, which has been recently studied in view of growth factor receptor function. Dr. Senitiroh Hakomori and his associates (The Biomembrane Institute, University of Washington, Seattle, Washington) previously found that GM₁ and/ or GM₃, but not other ganglioside, inhibit PDGF-dependent as well as EGF-dependent cell growth and their receptor kinase activities. New data were presented on the effect of lyso-GM₃ and de-N-acetyl GM₃. The former inhibits EGF receptor kinase activity in a monophasic way, irrespective of detergent concentration. In contrast, the latter (de-N-acetyl-GM₃) showed a striking promoting activity of the EGF receptor activity, and exogenous addition of de-N-acetyl GM₃ promotes cell growth. Antibodies directed to de-N-acetyl GM₃ detected the presence of this compound in various types of tumor cells but not in normal cells.
The functional significance of glycoconjugates in defining malignancy, including metastatic potential, can be assessed by modulation of glycoconjugate synthesis by specific reagents. Along these lines, an important approach has been developed and was presented by Dr. Toshiaki Osawa (University of Tokyo, Tokyo) and his associates, who applied a modified sialic acid-nucleoside conjugate (KI-8110 compound)¹ that inhibits sialyltransferase for synthesis of NeuAca2 3Gal sequence in glycoproteins and glycolipids. The inhibition was particularly effective for synthesis of mucin-type glycans and long chain complex-type gangliosides. Interestingly, the compound inhibited the metastatic potential of NL-17 colon adenocarcinoma cells to lung, and inhibited tumor cell-dependent platelet aggregation. The compound also inhibited PDGF-dependent cell growth stimulation.
A possible correlation between metastatic potential and glycosylation pattern has been systematically studied by Dr. Tatsuro Irimura (M.D. Anderson Hospital and Tumor Institute, Houston, Texas) using more than 500 fresh tissue specimens from primary colonic carcinoma and metastatic deposits from more than 200 patients during the past 4 years. A positive correlation between a higher degree of sialylation and metastatic potential was clearly observed in many samples. He was able to detect a high molecular weight sialoglycoprotein that binds to WGA in metastatic deposits. He was also able to detect sialyl Le^x antigen defined by monoclonal antibody FH6 in metastatic deposits. Detection of a large amount of a mucin-like sialoglycoprotein with Mr 90,000 in primary tumors can be correlated with higher metastatic potential of colonic carcinomas in humans and in nude mice. These studies and others indicate that a high degree of sialylation, particularly in high molecular weight glycoproteins, may indeed correlate with high metastatic potential.
Cell surface glycoconjugates undergo characteristic changes during differentiation from immature precursors to functionally-active, mature cells. This general concept has been further reinforced by a clear demonstration of dramatic changes of 0-glycan biosynthesis associated with T lymphocyte activation, presented by Dr. Minoru Fukuda (La Jolla Cancer Research Foundation, La Jolla, California) and his associates, Friedrich Piller and Robert Fox. Resting T lymphocytes express disialotetrasaccharide (NeuAca2 3Gal!!!1 3[NeuAca26]GalNAcaSer/Thr) in leukosialin molecules, whereas activated human T cells contain the more complex structure NeuAca2 3Gal!!!l 3[NeuAca2 3Gal!!!l 4GlcNAc!!!l 6]GalNAcal Ser/Thr in the same leukosialin molecules. A dramatic shift in biosynthesis of simple to complex type 0-glycan in leukosialin is caused by a decrease of a2 6 sialyltransferase and enhancement of !!!l 6 GlcNAc transferase during T cell activation. Interestingly, cells with mature functions bear more complex glycans in leukosialin molecules. An important question remains open, i.e., whether leukosialin is simply a differentiation marker or is functionally closely associated with activating T cells.
Studies on enzymatic changes as a basis for the major aberrant glycosylation associated with human colonic cancer were presented by Dr. Eric Holmes (Pacific Northwest Research Foundation, Seattle, Washington). Accumulation of type 2 chain antigens such as Le^x, Le^y and sialyl Le^x in various human colonic cancer cell lines, and their absence in normal colonic cancer, have been assessed by immunoblotting of glycolipids with defined antibodies. The key enzyme for synthesis of an enhanced quantity of lacto-series glycans is!!!1 3 GlcNAc transferase, which catalyzes chain elongation. This enzyme was highly active in various.cancer cell lines and tissues, but barely detectable in a normal colonic cell line and in normal colonic mucosa. Activity of the enzyme was also detected on a histological basis by incubation of tissue sections prepared from frozen samples with substrate glycolipid in situ and UDP-Gal. The tumor sections were stained with antibody 1 B2 without preincubation with U DP-Gal and nLc₃, while normal tissues were not stained unless tissues were preincubated with nLc₃ and UDP-Gal. This technique can be widely applied in a large variety of tissues and could be important for in situ demonstration of transferases.
Polylactosamine antigens have been identified as the major determinants of embryoglycans expressed highly in preimplantation embryo as well as in teratocarcinomas by a series of studies previously made by Dr. Takashi Muramatsu and his associates (Kagoshima University, Kagoshima). In his presentation, he demonstrated that the sera of patients with germ cell carcinomas contain a polylactosamine embryoglycan, the epitope of which has been identified as Gala1 3Gal by endoglycosidase digestion. On the other hand, Dr. Muramatsu also presented data indicating that the protein core structure of the embryoglycan showed a great deal of homology with immunoglobulin and suggested that embryoglycan proteins belong to the immunoglobulin superfamily.

February 17 (evening session)
The evening session was devoted to studies on cell adhesion molecules and the possible application of such molecules in defining the metastatic potential of tumor cells. The section was chaired by Drs. George Martin (National Institute of Dental Research, Bethesda, Maryland) and Ikuo Yamashina (Kyoto University, Kyoto).
Dr. Kenneth Yamada (National Cancer Institute, Bethesda, Maryland) reviewed the domain structure of fibronectin, including the role of the R-G-D sequence in cell attachment. Recombinant fibronectin mutated to lack this sequence shows no cell attachment. However, he showed that another portion of the molecule separated by more than 50 amino acids from the R-G-D sequence is required for high affinity binding to cell surface receptors. Synthetic peptides containing the R-G-D sequence were shown to be able to inhibit the formation of lung lesions in animals injected with melanoma cells, perhaps by blocking the attachment of the cells to matrix proteins that contain this attachment motive.
Dr. George Martin (National Institute of Dental Research, Bethesda, Maryland) reviewed the interaction of malignant tumor cells with basement membranes. Since basement membranes surround all epithelial tissues and blood vessels, they create a significant barrier to the spread of tumors, but one that malignant cells can cross. Laminin is a key protein in the process, since malignant cells bind preferentially to laminin and this interaction triggers the production of degradation enzymes and the motility of the cells. Laminin has now been cloned and sequenced. A unique sequence, YIGSR (tyr-ile-gly-ser-arg), serves as a binding site for the laminin receptor on the tumor cells. This YIGSR peptide, particularly in cyclic form, shows antimetastatic activity in mice.
Although receptors for various adhesive proteins, i.e., fibronectin, thrombospondin, and von Wildebrand’s factor, have been well described as members of the “integrin family,” the receptor for laminin and collagen has been described as independent. The role of gangliosides and other glycolipids, partlcularly sulfatides, in cell adhesion has been investigated by Drs. Victor Ginsburg and David Roberts (National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, Maryland). Laminin, thrombospondin, and von Wildebrand’s factor bind spontaneously and with high affinity to sulfated glycolipids. It is this binding property that accounts for their ability to bind glutaraldehyde-fixed erythrocytes. Various kinetic data indicate that sulfated glycoconjugates participate in both laminin- and thrombospondin-mediated cell adhesion, but their mechanism of interaction are different.
Regarding the integrin family receptor recognition of the R-G-D-S sequence, Dr. David Cheresh (Scripps Clinic and Research Foundation, La Jolla, California) reported an elegant experiment using variants of human melanoma that fail to synthesize the a-chain protein or its mRNA but produce normal levels of the!!!-chain. In these cells, the!!!-chain does not reach the cell surface, but rather accumulates in the cytoplasm. Mutant M21-L cells, which lack the a-chain, are incapable of attaching to vitronectin, von Wildebrand’s factor, fibrinogen, or an R-G-D-S peptide, yet are capable of attaching normally to fibronectin, although the parent cells, which express both the a- and!!!-chains, are capable of binding to all cell adhesive proteins.
An extensive review supported by voluminous data on the selective metastatic potential of tumor cells as related to various molecular parameters was presented by Dr. Garth Nicolson (M.D. Anderson Hospital and Tumor Institute, Houston, Texas). Evidence from certain experimental tumor systems supports Paget’s “seed and soil” hypothesis on the nonrandom distributions of metastases, where the unique properties of the organ microenvironment determine the organ preference of metastasis. Multiple cell adhesion systems, differential invasion of basement membranes and organ tissues mediated by a family of degradative enzymes, and differential responses to organ-derived growth-stimulatory and -inhibitory factors appear to be important determinants in explaining the organ preference of metastasis. Tumor cell diversification and heterogeneity, host selection processes and increased production of tumor autocrine molecules that may modulate adhesion, invasion, and growth are important in promoting metastasis when neoplasms progress. These and other tumor and host cell properties may eventually be used to predict and explain the unique metastatic distributions of certain malignancies.

February 18 (morning session)
In the morning session, chaired by Drs. Masaru Taniguchi (Chiba University, Chiba) and Philip Livingston (Memorial Sloan-Kettering Institute, New York, New York), discussion was focused on the immunobiological significance of aberrant glycosylation, and clinical applications of antigens for immunotherapy. Despite the knowledge that many tumor-associated antigens defined by monoclonal antibodies are carbohydrates carried either by glycolipids or glycoproteins, little is known about immune response mechanisms. Many monoclonal antibodies directed to human cancer antigens were established after immunizing mice with human cancer cells. Thus, the approach may simply reflect the intensity of mice immune response to human tumor cells in such heteroimmune systems. Dr. Masaru Taniguchi (Chiba University School of Medicine, Chiba) immunized syngeneic C57/BL/6 mice with B16 melanoma and established monoclonal antibody M2590, which reacts specifically with melanoma cells. The antigen was identified as GM₃. His data indicated that anti-melanoma cytotoxic T cells (CTLs) are directed to a GM₃-protein complex, whereas anti-melanoma suppressor T cells (Ts) recognize NeuAc-GM₃ but not NeuGc-GM₃. The affinity of antibody M2590 for GM₃ lactone is 10 times higher than that for GM₃, and M2590 recognizes GM₃ at high but not at low density. Dr. Taniguchi also presented evidence that a soluble melanoma antigen, as well as normal GM₃ liposome, induces anti-melanoma Ts, which suppresses CTL generation and maintains self-tolerance to GM₃. It was emphasized that this Ts generation to self-GM₃ could be the major escape mechanism of melanoma from immune surveillance.
Some anti-GD₃ gangliosides with high affinity. such as R24. have been shown by Alan Houghton and associates (Sloan-Kettering Institute, New York, New York) to have a significant effect in causing regression of human melanoma. The immunobiological background of this phenomenon was studied by Dr. Sidney Welt (Sloan-Kettering Institute, New York, New York) using various melanoma cell lines and melanocytes. Melanoma cells binding low amounts of R24 were not lysed in R24-dependent immune reactions in both complement-dependent cytotoxicity (CDCC) and antibody-dependent effector cell cytotoxicity (ADCC) tests. In both processes, a threshold value of R24 molecules bound per cell is necessary to initiate the cytotoxic mechanism. Only those cells binding >10⁷ R24 molecules per cell showed >90% cell lysis. This situation is similar to that of GM₃ antigen expressed in mouse and human melanoma and recognized by antilactone antibody, as recently described by Nores et al. (J. Immunol. 139:3171, 1987). However, no clear effect was seen on established tumors in nude mice. Thus, the in vivo mouse model failed to predict the clinical pathological findings observed in trials of R24 in human patients, in which tumor regression was occasionally observed.
For possible development of a tumor vaccine, it is necessary to investigate effective methods to elicit antibody response that may prevent tumor development and progression. A great deal of effort has been paid to development of a method for active immunization of humans using gangliosides at the Sloan-Kettering Institute, New York. Dr. Philip Livingston (Sloan-Kettering Institute, New York, New York) presented data from a large-scale trial with more than 60 melanoma patients over the past several years. Since melanoma contains GM₂, GD₂, GD₃, and 9-0-acetyl-GD₃ gangliosides as the major antigens, melanoma patients were immunized with these gangliosides coated on carrier vehicles such as Salmonella minnesotae mutant R595, BCG, and liposomes. Immune responses to these gangliosides in terms of antibody production have been correlated with status of tumor progression. Antibody production was maximal in patients pre-treated with low-dose cyclophosphamide and vaccinated with GM₂-BCG. Out of 24 such patients, 19 made high titers of IgM and 8 made high titers of IgG antibody. Antibody reactivity was restricted to N-acetyl-GM₂, but in some cases there was also a cross-reaction with N-glycolyl-GM₂. Disease progression in antibody producers was significantly reduced. Interestingly, no immune response was observed against GD₃, GD₂, or 9-0-acetyl-GD₂.
Gangliosides have been implicated as modulators of the immune response of B and T lymphocytes, NK cells, and macrophages by increasing evidence compiled by a number of investigators. Studies reported by Dr. Stephan Ladisch (University of California at Los Angeles School of Medicine, Los Angeles, California) represent one such current line of study. He presented the following four items: (i) YAK lymphoma sheds gangliosides that modulate cellular immune response in vitro. (ii) Various lymphoma of AKR mice showed a correlation between ganglioside content, shedding ability, and tumorigenicity of cell lines. Highly tumorigenic SL12.3 cell lines shed specific gangliosides, and addition of a picomolar quantity of such gangliosides to poorly tumorigenic SL12.4 cells strongly enhanced the tumorigenicity. (iii) In a human system, macrophages are sensitive to suppression of their function by preincubation with gangliosides. (iv) Human neuroblastoma sheds a significant quantity of gangliosides, as shown in experimental tumor systems. All these findings support the hypothesis that tumor gangliosides are physiologically relevant modulators of immune function and tumor formation.
In various clinical trials of mouse monoclonal antibodies directed to human cancer, tumor regression was usually observed when anti-idiotype antibody or anti-anti-idiotype antibody was detected. In fact, an exciting approach in current tumor immunology is the use of anti-idiotype antibodies to invoke anti-tumor response. One such study on anti-idiotype antibodies was presented by Dr. Akira Yachi (Sapporo Medical College, Hokkaido). An anti-idiotype antibody directed to monoclonal antibody YH206 was useful in detecting the YH206-defined antigen in the sera of cancer patients. This antigen was a high-molecular weight asialo-carbohydrate associated with mucin-type glycoprotein. A second study on anti-idiotype antibodies was presented by Dr. Kozoh Imai (Sapporo University School of Medicine, Hokkaido) who established the anti-CEA antibody 5B3, which recognizes the carbohydrate moiety of CEA. These studies indicate that such anti-idiotype antibodies could be useful for fine analysis (i.e., tertiary structural analysis) of carbohydrate epitopes.
The section was concluded by an interesting presentation by Dr. Toshisuke Kawasaki (Kyoto University, Kyoto) indicating the novel possibility that human serum mannose-binding protein (MBP), a type of lectin, activates the complement system through a classical pathway. He showed that mannann-coated sheep erythrocytes were lysed, and E. coli containing high-mannose type glycoprotein were killed, by human serum MBP in the presence of human complement. Complement activation by the lectin was accomplished by mimicking the function of the complement component C1Q. Thus, the interaction of serum MBP with ligands on cell surfaces leads to the activation of the Clr2s2 complex. It is interesting to speculate that a tumor surveillance system involving such a mechanism could be operating against some tumor cells expressing high-mannose type glycoprotein or some other lectin with different carbohydrate binding specificity.

(3) Seminar on “Biology of Oncogenes”
Burgeoning new information being developed in the field of oncogenes, growth factors/ growth factor receptors and cancer, particularly in laboratories in the United States and Japan, provided strong impetus for this cooperative state-of-the-art forum. The sessions reflected the most impressive new observations in their respective fields and brought together leading scientists of both nations to discuss new concepts, approaches and proposed projects. One important objective was to identify areas of productive collaboration and to arrange for exchanges of reagents.
The seminar was held on February 17-18, 1988 under the auspices of the United States-Japan Cooperative Cancer Research Program at the Coco Palms Hotel, Kauai, Hawaii. There were 14 participants, eight from Japan, five from the United States and one from Canada.
Dr. Sugimura welcomed the participants. He followed with a precis of the U.S.-Japan Cooperative Cancer Research Program and expressed gratification at its success in fostering increased communication and collaboration between United States and Japanese scientists.
In the first session, Molecular Mechanisms of Oncogene Action I, chaired by Dr. Richard H. Adamson (National Cancer Institute, Bethesda, Maryland), Dr. Masaaki Terada (National Cancer Center Research Institute, Tokyo) presented information on hst, a transforming gene originally identified in the NIH/3T3 cell transfection assay in three different samples of DNAS from the noncancerous portion of stomach mucosa from patients with stomach cancer, and subsequently from three stomach cancers, three hepatomas and one colon cancer. The KS oncogene recently identified in another laboratory as a transforming gene in DNA from Kaposi’s sarcoma turned out to be hst. These data suggested that the hst gene is at present the most frequently found transforming gene among the non-ras group. The 206-amino acid sequence of the hst-encoded protein predicted a signal peptide and an N-linked glycosylation site. No ATP or GTP binding site was identified and the amino acid sequence was not homologous to those of either tyrosine or serine/ threonine protein kinases. These results indicated that the hst gene may encode a growth factor. The deduced amino acid sequence of the hst-encoded protein has 43%, 38% and 40% homology to human basic fibroblast growth factor (FGF) and human acidic FGF, and mouse int-2-encoded protein, respectively, in selected regions.
Dr. Stuart A. Aaronson (National Cancer Institute, Bethesda, Maryland) described growth factor-activated pathways in the neoplastic process. Discoveries identifying the products of certain proto-oncogenes as growth factors or their receptors have provided strong evidence that proto-oncogenes play a central role in normal growth regulation. The altered or aberrant expression of such genes appears to be fundamental to steps that convert normal cells to malignancy. One example, the human sis proto-oncogene, encodes one chain of human plateletderived growth factor (PDGF-2). Expression of this normal growth factor coding sequence in cells possessing PDGF receptors can induce uncontrolled growth associated with acquisition of the malignant phenotype. The human sis proto-oncogene is under strict regulation in cells such as normal fibroblasts and glial cells. In contrast, a large percentage of human fibrosarcomas and glioblastomas express sis transcripts, and sis/PDGF-2 gene products associated with such tumor cells had been identified. Thus, deregulation of sis/ PDGF-2 expression may play an important role in triggering the uncontrolled growth associated with these tumors. In addition, Dr. Aaronson’s group has investigated the effects of overexpression of the normal coding sequences for the epidermal growth factor (EGF) receptor and its related receptor-like gene, human erbB-2, on growth properties of NIH/3T3 cells and demonstrated that overexpression itself was an important mechanism by which a normal gene could be activated as an oncogene. Thus, overexpression, itself, may convert a gene for a normal growth factor receptor into an oncogene. Differences in the biological responses of the same cells to an overexpressed EGF receptor gene provide an approach toward mapping receptor sequences important in coupling with specific signal transduction pathways of the cell.
Dr. Tadashi Yamamoto (Institute of Medical Science, Tokyo) reported on new proto-oncogenes of the tyrosine-kinase family. Taking advantage of sequence similarities of the kinase domain, Dr. Yamamoto’s group has been able to identify and clone new members of the proto-oncogenes of the tyrosine-kinase family. These are the lyn,.fyn and the c-erbB-2 genes. Nucleotide sequence analysis suggested that both lyn and fyn encode tyrosine kinases of nonreceptor type, which are highly related to the protein products of the c-src, c-yes, c-fgr, lck and hck genes. Tests of fetal and embryo tissues demonstrated high expression of fyn in the brain and lyn in the liver, suggesting that they play roles in cells of the respective tissues in which found. Further analyses demonstrated expression of lyn in B lymphocytes. Possible transforming abilities of lyn and fyn are now under investigation.
In contrast with lyn and fyn, c-erb B-2 was found to encode a tyrosine kinase of the receptor type, a protein very similar to the EGF receptor. Tests revealed that amplification and/or overexpression of c-erb B-2 occurred frequently in adenocarcinomas of the stomach and breast and less frequently in astrocytomas, tumors of the brain. Studies to determine the molecular mechanism of cell transformation involving c-erbB-2 expression are currently underway.
In the second session, Transgenic Models for Oncogenes, chaired by Dr. Masabumi Shibuya (Institute of Medical Science. Tokyo). Dr. Douglas Hanahan (Cold Spring Harbor Laboratory, Cold Spring Harbor, New York) described aspects of tumor progression in transgenic mice. A general property of oncogenes has been revealed by their establishment in transgenic mice. In many cases the presence of a particular oncogene results in the heritable development of a specific tumor or spectrum of tumors. While the oncogene is clearly necessary to elicit these tumors, the patterns of tumor development in general suggest that additional events are necessary to manifest the malignant phenotype. Two examples of tumorigenesis in transgenic mice were presented, one induced by the bovine papillomavirus (BPV) I genome which elicits skin tumors in transgenic mice which inherit the integrated genome and the second by SV4O T antigen. In the former, there are two classes of pathology - abnormal skin and protuberant tumors. The abnormal skin is histologically described as a fibromatoses, or hyperplasia of the dermal fibroblasts. A more advanced (or acute) form of abnormal skin is described as aggressive fibromatoses. The protuberant tumors induced are fibrosarcomas. They are locally invasive (but not metastatic); they do not have a capsule typical of benign tumors, and thus show some features of a malignant tumor. Cultured fibroblasts derived from these three different pathologies show stable differences which appear to represent the state of tumor progression from which they were removed. The fibromatoses cells express low levels of BPV RNA, grow very slowly, are contact-inhibited, and are apparently not “immortal.” Aggressive fibromatoses cells grow rapidly, express higher levels of BPV RNA (and the E5 and E6 oncoproteins), are tumorigenic and immortal, but are still relatively contact-inhibited. Fibrosarcoma cells are more rounded, are immortal and tumorigenic, and form spontaneous dome-like foci in culture. The fibrosarcomas have levels of BPV DNA, RNA, and the E5 and E6 oncoproteins which are similar to those in the aggressive fibromatoses. Yet the fibrosarcoma cells show a clearly distinct phenotype in vivo and in vitro. This implies that other changes are necessary to produce a fibrosarcoma. Analyses of the chromosomes in these different stages of tumor progression suggest that specific changes in karyotype are associated with the fibrosarcomas and may well be causal to their development.
In the second example of tumor progression, targeted expression of SV4O T antigen to the pancreatic!!!cells results in the development of insulinomas. High level expression of T antigen does not perturb!!!cell development, and in the young mice, the mitotic index of!!!cells expressing both Tag and the cellular proto-oncogene p53 is normal. This is followed by the development of a proliferative capacity, which is manifested as a general hyperplasia of the several hundred focal clusters of!!!cells. Only a few of the hyperplastic islets progress into solid, highly vascularized tumors. Again, expression of the oncoprotein cannot distinguish these states m tumor progression, implying a requirement for secondary events.
Driven by the same Ig enhancer and SV4O T promoter, Dr. Shinichi Aizawa (Institute of Physical and Chemical Research [RIKEN], Ibaraki) reported that ras induced lung adenomatous tumors, myc induced pre-B cell lymphomas and the SV4O T gene induced a variety of tumors in transgenic mice. Different types of tumors developed in transgenic mice following the introduction of the entire coding region of ras, myc or SV4O large T gene (T) linked to the same regulatory unit, consisting of a human immunoglobulin gene enhancer (Ig) and SV4O early gene promoter (Tp) with a 21-bp repeat. Twelve transgenic mice harboring the intact T gene developed a variety of tumors including choroid plexus tumor, B cell lymphoma, histiocytic lymphoma, thymoma and others, suggesting that the Ig/Tp regulatory unit has transcriptional activity in these hetrologous tissues. With this regulatory unit, myc gene-induced pre-B cell lymphomas. Contrary to expectations, however, the mutated ras gene induced lung adenomatous tomors which were histologically comparable to adenocarcinomas in man. The tumors developed as early as four weeks after birth and the introduced ras gene was as efficiently expressed in both normal and neoplastic bronchiolo-alveolar epithelial cells as in normal lymphoid cells. An unidentified certain secondary event thus appears to be necessary for these ras-expressing cells to become neoplastic, as had been observed for rnyc by Leder et Al-l986. In a variety of tumors induced by Ig/ TP-T, on the other hand, T gene was expressed only in the tumor cells. Thus, the derepression of T gene in normal cells appears to be closely related to malignant change as had been reported in development of pancreatic acinar cell tumors by the T gene (Ornitz et Al-l985). Results suggest that ras and myc oncogenes penetrate differentially specific types of cells, while the SV4O T gene is tumorigenic in a variety of cell types.
In the third session, Molecular Mechanisms of Oncogene Action II, chaired by Dr. M. Terada, Dr. Raymond Erickson (Harvard University, Cambridge, Massachusetts) described molecular cloning of a serine-specific protein kinase activated by mitogenic stimuli. One of the most vigorous protein kinase activities in animal cells is detected by the phosphorylation of the 40S ribosomal subunit protein S6. The regulation of this enzyme activity is of interest because the phosphorylation of S6 is a highly conserved response of cells to a variety of mitogenic stimuli. Serum-deprived cultured cells undergo S6 phosphorylation in response to growth factors found in serum, oncogene products and tumor-promoting phorbol ester. Although these agents initially act through different receptors, the pathways apparently converge to activate a single enzyme, or a limited number of distinct enzymes.
A goal of these studies was to elucidate the mechanism of S6 kinase activation and to determine the relationship between various S6 protein kinases, taking into account that the initial events during activation could involve tyrosine phosphorylation, activation of protein kinase C or, in the case of progesterone, a steroid hormone receptor linked to the adenylate cyclase system. To date little information is available on the mechanism of activation of S6 kinase(s) in vivo, although in vitro S6KII can be inactivated with serine phosphatases and can be phosphorylated and partially activated by microtubule-associated protein kinase. Whether this activation represents a pathway that is operative in vivo is unclear, but it raises the possibility that the pathways of activation converge not at the S6 kinase but at an S6 kinase such as microtubule-associated protein kinase. Characterization of the relevant enzymes was necessary to describe the events in more detail. The molecular cloning of S6KII was undertaken in order to relate any findings on its activation to the structure of the molecule and to attempt to determine the number of proteins that may contribute to S6 phosphorylation in cells.
Results demonstrated that S6KII proteins have a very unusual structure when compared with previously studied protein kinases. They contain two apparent kinase domains, each similar to distinct protein kinases. These were described as were possible mechanisms of activation and the evolution of S6KII.
The next report, by Dr. M. Shibuya, dealt with characterization of cellular genes encoding protein kinases: tissue- and stage-specific expression of a tyrosine kinase gene (c-ros) and of a novel mammalian kinase gene related to the cdc28/cdc2 gene in yeast. Recent studies on oncogenes revealed that both tyrosine and serine/ threonine kinases have crucial roles in carcinogenesis. Furthermore, analysis of cell cycle-dependent mutants of yeast demonstrated that some serine/ threonine kinases such as cdc28/ cdc2 and cdc25 are directly involved in the control of cell cycle. Therefore, it appeared important to examine the structure and physiological functions of the protein kinase family in vertebrates for a basic understanding of cell proliferation. Two topics were explored: (1) the c-ros gene is the counterpart for v-ros of the UR2 sarcoma virus, which is suggested to encode a receptor-type tyrosine kinase. The physiological significance of the c-ros is not clear because of its extremely low expression in normal tissues. Results demonstrated expression of c-ros mRNA of large size in kidney and a few other tissues in a stage-specific manner. The cDNA encodes a receptor-type molecule which includes a long extracellular domain as well as transmembrane and kinase domains; c-ros mRNA of small size is expressed in testis but none of the cDNA obtained from testis contained a regular transmembrane domain in spite of the presence of a tyrosine kinase domain, suggesting an alternative splicing in this tissue; and comparison of the c-ros sequence with mcf3 (activated c-ros) indicated that no point mutation nor truncation at the C-terminus are required for its activation.
(2) Using v-ros DNA as a probe, this group isolated a human genomicDNA fragment encoding a protein kinase (ros-3). cDNA analysis of murine ros-3 revealed that this gene appears not to encode a tyrosine kinase but a protein kinase partly homologous to cdc28/ odc2 in yeast. ros-3 is distinct from a recent isolate, a human gene, cdc2Hs, functionally complementary to cdc2. It was also interesting to note that ros-3 is expressed in a highly tissue-specific manner.
Dr. Frank McCormick (Cetus Immune Corporation, Emeryville, California) presented evidence that GTPase-activating protein (GAP) may be a ras effector. Dr. McCormick and his collegues recently identified a cytoplasmic protein that greatly enhances the GTPase activity of N-ras p21 but does not affect oncogenic mutants. To identify the region of p21 with which this protein interacts, 20 H-ras p21 mutants were purified and tested for their ability to undergo stimulation of GTPase activity by GAP. It was found that mutations in p21 that did not affect molecular transformation by v-H-ras p21 did not affect GAP-mediated stimulation of p21 GTPase activity. Mutations at positions 12, 59 and 61 prevented stimulation of GTPase activity. Significantly, point mutations in the putative effector region of ras p21 (amino acids 35, 36 and 38) were insensitive to GAP. However, a point mutation at position 39, shown previously to have no effect on transformation, had no effect on GAP-p21 interaction. Results indicated that GAP may be the effector protein with which ras p21 interacts in a GTP-dependent manner and suggests a model in which the interaction of ras with its effector causes down-regulation by stimulating conversion of p21.GTP to inactive p21.GDP.
In the last paper of the session, Dr. M. Nagao (National Cancer Center Research Institute, Tokyo) described induction of flat revertants by benzamide from oncogene-transfectants. Benzamide is a potent inhibitor of a poly(ADP-ribose) polymerase. It induces sister chromatid exchanges, inhibits cell transformation by alkylating agents or x-ray irradiation, and potentiates some cancer chemotherapeutic agents. Dr. Nagao’s group studied the effect of benzamide on a ras transformant of the NIH/3T3 cell, al-l. A secondary transformant derived from a human bladder cancer T24, al-l, contains several copies of H-ras^T24. Cloned Al-l cells, cultured in 5mM benzamide for two weeks, became flat. Analysis of DNA from these cells showed that exogenous H-ras^T24 sequences were partially eliminated. Cloning of the latter cells revealed,them to be composed of two types of cell populations, one of flat morphology and no exogenous H-rasT24 sequence. The second showed transformed cell morphology and had the same copy number of H-rasT24 sequence as Al-l cells. On the other hand, endogenous H-ras sequence in human bladder carcinoma cell T24 was retained even after treatment with 5mM benzamide. Transformants induced by transfecting activated K-ras, N-ras, raf and ret also changed to flat morphology with benzamide. Loss of exogenous K-ras and N-ras from the transformants was demonstrated. Results indicat6d that exogenous sequences, at least tandemly repeated, were lost during culture in the presence of benzamide.
The fourth session, chaired by Dr. Richard Adamson, addressed Recessives in Human Malignancy. Dr. Jun Yokota (National Cancer Center Research Institute, Tokyo) described oncogene and chromosome abnormalities in human lung cancer. Various types of fresh lung tumors were examined for alterations of myc family oncogenes and loss of chromosomal heterozygosity. Seventy tumors obtained from 53 patients, included 17 small cell carcinomas (SCCs), 18 adenocarcinomas (AdCs), and 12 squamous cell carcinomas (SqCs). The heterogenous occurrence of myc family oncogene abnormalities in lung tumors was infrequent. In contrast, loss of chromosomal heterozygosity was detected very frequently and occurred consistently within tumors of the same individuals. In SCC, the incidence of allelic deletions at three different chromosomal loci was extremely high; loss of heterozygosity was detected on chromosome 3p in 7 of 7 patients, 13q in 10 of 11 patients, and 17 p in 5 of 5 patients. Deletions at these loci were observed even in the absence of clinical evidence of metastases. Furthermore, loss of heterozygosity on chromosomes 3p and 13q occurred prior to N-myc amplification and 11p deletion. Loss of heterozygosity on chromosome 3p was also detected with high frequency in AdCs. Results suggested that activation of myc family oncogenes occurs in the late stage of tumor progression, while loss of heterozygosity on chromosomes 3p, 13q and 17p occurs relatively early in the development of SCC. Therefore, activation of myc family oncogenes may influence the biological behavior of each tumor. In contrast, recessive genetic changes involving sequences on chromosome 3p, 13q and 17p may play important roles in the genesis of SCC, and those on chromosome 3p may also play an important role in the genesis of AdC.
Dr. Webster Cavanee (Ludwig Institute for Cancer Research, Montreal, Quebec) reported on recessive mutations and human cancer. The determination and comparison of genotypic combinations at genomic loci in constitutional and tumor tissues from patients with various types of cancer have defined the chromosomal locations of loci in which recessive mutations play a role in disease development. The predisposing nature of some of these mutant alleles is exemplified by studies of retinoblastoma and osteogenic sarcoma. These two clinically associated diseases share a pathogenetical causal predisposition mapping to 13q14. A similar mechanism involving 11p15.5 is involved in the development of the embryonal variant of rhabdomyosarcoma. Since the alveolar variant does not share this mechanism, its use as a differentially diagnostic indicator has been developed. Finally. genomic alteration for chromosome 10 is apparent in glioblastomas and mixed tumors of glioblastoma/astrocytoma grade III but not in homogeneous astrocytoma grades II or III, suggesting the definition of a locus involved in tumor progression and, perhaps, an approach to molecular genetic staging tumors.
The final scientific paper of the seminar, introduction of a normal human chromosome 11 suppresses the tumorigenicity of cervical tumor and rhabdomyosarcoma cell lines, was presented by Dr. Mitsuo Oshimura (Kanagawa Cancer Center, Yokohama). Stanbridge and his associates reported that the introduction of a single copy of a human chromosome 11 is sufficient to completely suppress tumorigenicity of cervical (HeLa) and Wilms’ (G401) tumor cells. In order to determine whether another human cervical tumor cell line and a rhabdomyosarcoma cell line would respond in the same way, Dr. Oshimura’s group performed chromosome transfer experiments via microcell fusion into SiHa cervical tumor cells containing human papillomavirus (HPV) type 16 and A204 Wilms’ tumor cells, after first isolating mouse A-9 cells containing a single human chromosome with integrated pSV2-neo plasmid DNA. Following introduction of chromosome(s) from A-9 cells containing the neo-marked chromosome 11 into SiHa cells via microcell fusion, they isolated 16 independent SiHa clones which were resistant to G418. All the parental SiHa cells contained three copies of chromosome 11, while 95-100% of the cells in 6 of the 16 G418-resistant microcell hybrids contained four or five copies. The parental cells formed tumors in nude mice 3-5 weeks after injections of 1 X 10⁷ cells, while the five clones which contained additional copies of chromosome 11 did not form tumors before 12 weeks. By contrast, introduction of chromosome 12 for the control experiment did not affect the tumorigenicity of SiHa cells. HPV16 gene expression decreased in most of the hybrids. The suppression of tumorigenicity in the hybrids was correlated with the regulatory expression of HPV16 early genes, whereas Ha-ras RNA levels were constant. Introduction of chromosome 11 into the rhabdomyosarcoma cell line, A204, also was found to suppress tumorigenicity. Results strongly suggested that the “suppressor” gene(s) on chromosome 11 play a crucial role in a step of the neoplastic development of the tumor lines tested.
Dr. Adamson’s closing remarks reflected the consensus satisfaction with the level of excellence represented in the presentations. He thanked the participants and expressed his enthusiasm for follow-up meetings in the near future.

(4) Seminar on “Glutathione S-Transferases and Carcinogenesis”
This seminar was held on March 23-24, 1988 at the Sheraton Makaha Hotel, Oahu, Hawaii. The organizers were Dr. Kiyomi Sato, Second Department of Biochemistry, Hirosaki University School of Medicine, Hirosaki, Japan, and Dr. Henry C. Pitot, McArdle Laboratory for Cancer Research, University of Wisconsin, Madison, Wisconsin, USA. There were six participants each from Japan and the United States. The purpose of this seminar was to discuss the properties, regulation, genetic expression, and role of glutathione S-transferases (GST) and related functions in carcinogenesis, drug resistance, and normal tissue expression. The program was informal, allowing considerable time for discussion and interaction between the scientists.

March 23, 1988 - Drs. Sato, Muramatsu, Pickett, Fahl, Saijo, and Tew
Dr. Kiyomi Sato (Hirosaki University School of Medicine, Hirosaki) opened the presentations with a discussion of the historical discovery by his laboratory of the placental form of glutathione S-transferase and its appearance during the early stages of hepatocarcinogenesis in the rat. The regulation of the expression of the gene for this enzyme appeared to differ not only in neoplasms of the rat, but also in the two different sexes. The enzyme can be induced by the administration of lead nitrate maximally at 3 days after administration in essentially all hepatocytes. Using the enzyme as a marker during carcinogenesis by the Solt-Farber model, identification of single hepatocytes containing the placental form of glutathione S-transferase could be identified by immunohistochemistry. The numbers of such individual cells, as well as “mini-foci” consisting of only a few such cells, change during the process of carcinogenesis peaking early and then either remaining constant or increasing slowly as neoplasms later appeared. Only those hepatocarcinogens inducing genetic damage also caused the appearance of single cell and mini-foci alterations. The human form of this enzyme appears to be a good marker in human cancer, except liver, as well as in several examples of premalignant conditions, e.g., colon and cervix. Among its other functions, the placental form of glutathione S-transferase may catalyze the destruction of lipid peroxides in cells and cell membranes.
Dr. Masami Muramatsu (University of Tokyo, Tokyo) described the early experiments from his laboratory which identified the increased levels of the placental form of glutathione S-transferase in two-dimensional electrophoresis gels of normal rat liver, neoplastic nodules, and carcinomas. The appearance of a spot in nodules and tumors was demonstrated to be the placental form of glutathione S-transferase. Isolation of the protein from this spot and sequence determination revealed that its N terminus was not blocked, and on the basis of such amino acid sequence data, oligonucleotides were synthesized and cDNA clones of the mRNA as well as clones of the gene for this enzyme were isolated. The gene consists of seven exons with the initiator codon being split between the first and second exon. Using the “CAT” assay and transfection, both enhancer and “silencer” sequences upstream of the cap site could be identified, at least one of which is quite similar to a tetradecanoylphorbol acetate (TPA) responsive element, whereas another was an SV4O-like enhancer. Gel retardation studies indicated that the AP-l protein appears to interact with the TPA-responsive element, thus possibly governing the level of expression of this enzyme.
Dr. Cecil Pickett (Merck Sharpe and Dohme Laboratories, Rahway, New Jersey) described his studies on several forms of glutathione S-transferase, especially the Ya, Yb, and Yc subunit species. The diversity and similarities of the various mRNAs of these subunits was described. Nuclear runoff experiments demonstrated the increased level of transcription of the Ya subunit following 3-methylcholanthrene administration. Transfection of constructs containing the 5’ sequence upstream of the Ya gene linked to CAT gave some inducible preparations depending on the cell type and presence of the Ah receptor. From these studies Dr. Pickett concluded that the Ah receptor-ligand may act directly on the 5’ sequence of the Ya subunit or on the 5’ flanking sequence of a “regulatory gene” indicated by the fact that cycloheximide inhibits induction of mRNA in some experiments. In other studies utilizing the Solt-Farber model, nodules were found to contain high levels of quinone reductase and microsomal epoxide hydrolase. Through the use of specific restriction enzymes, it appeared that these genes were hypomethylated, thus possibly accounting for the high levels of expression in these lesions. During the discussion, Dr. Muramatsu indicated that a similar situation was found with respect to the placental form of glutathione S-transferase in foci, nodules, and tumors.
Dr. William Fahl (McArdle Laboratory for Cancer Research, Madison, Wisconsin) reported on the transfection of a number of constructs containing glutathione S-transferase cDNA and SV4O promoter electrophorated into monkey cells in culture. Following transfection of a construct including the Ya subunit, all cells containing the construct became more resistant to the toxicity of the anti-diolepoxide of benzo[a]pyrene. In a preliminary experiment, a transgenic mouse was produced with a construct of the Ya subunit cDNA and an SV4O early promoter. Expression was noted in both the liver and the kidney but not in other tissues.
Dr. Nagahiro Saijo (National Cancer Center Hospital, Tokyo) described studies on the mechanism of resistance of cells to cisplatin and its analogs. In some cases the P-glycoprotein product of the mdr gene was involved in the resistance of cells from small cell lung carcinomas, whereas greater degrees of drug resistance were associated with high levels of the placental form of gluathione S-transferase. In general, however, the patterns of resistance were quite varied, but the inherent resistance to cisplatin was not related to the mdr gene. Such resistance may be correlated with increased levels of the placental form of glutathione S-transferase in resistant cells, but acquired resistance was not related to these levels. In the discussion, Dr. Tew pointed out that DNA repair may also be involved as a general mechanism of the resistance to such drugs, since topoisomerase II seems to be involved in such resistant mechanisms.
Dr. Kenneth Tew (Fox Chase Cancer Center, Philadelphia, Pennsylvania) described aspects of acquired and induced drug resistance, especially to alkylating agents and the involvement of glutathione and glutathione S-transferases. In a specific model system of Walker carcinoma 256 cells, a high level of resistance was correlated with a marked increase in glutathione S-transferase of the Yc form. Revertant cells showed a lowered level of this enzyme. Resistance was also correlated with the appearance of a new form of glutathione S-transferase similar to the Yc subunit. In many drug resistant tumors one may see dramatic increases in the levels of various types of glutathione S-transferase, especially in the placental form. Thus, if one could inhibit the activity of this latter enzyme, the resistance might be reversed. In studies in vitro, and to some extent in vivo, such inhibitors, including ethacrynic acid and piriprost, appeared to be effective as predicted. It was also pointed out that glutathione S-transferase is involved in prostaglandin synthesis in the reduction of hydroperoxides formed in the synthesis of various prostaglandins. Using a commercially available affinity column, levels of glutathione S-transferase in human tissues were also described, as well as the appearance of a new molecular species isolated by the column present in normal and malignant human colon.

March 24, 1988 Drs. Watabe, Ito, Kensler, Reddy, Konishi, and Pitot
Dr. Tadashi Watabe (Tokyo College of Pharmacy, Tokyo) described a number of elegant experiments on the stereochemistry of hydration and glutathione S-conjugation of mutagenic/carcinogenic epoxide enantiomers by hepatic epoxide hydrolase and glutathione S-transferases, respectively. In general, the benzylic carbon of the epoxide was acted on by the transferase following ring opening. The enantiomers of the 7,8-epoxide of 9,10-dihydrobenzo[a]pyrene were acted on differently by different glutathione S-transferases, which was in turn correlated with the level of the various isozymes in both normal and neoplastic liver. Similarly, with styrene oxide the changes in conjugate distribution were evident between liver and neoplastic and neoplastic nodules. The importance of methylation of polycyclic aromatic hydrocarbons for their carcinogenicity was described. In these forms the proximate carcinogen appears to be the hydroxymethyl form with subsequent sulfation of the hydroxy group resulting in a mutagenic/carcinogenic form. Hepatic cytosolic glutathione S-transferases can detoxify the sulfate form at a fast rate, thus leading to an inhibition of its carcinogenic/mutagenic effects on the liver.
Dr. Nobuyuki Ito (Nagoya City University Medical School, Nagoya) described a number of situations utilizing the placental form of glutathione S-transferase as a marker in rat hepatocarcinogenesis. Monitoring both focal lesions and neoplasms, dose response curves with several carcinogens were described indicating a significant correlation between the two types of lesions. A model system developed in his laboratory gave a good correlation between hepatocarcinogens and positive effects in the system. False positives appeared to be minimal, and the system could identify inhibitory effects of chemicals on the carcinogenic process. Three-dimensional reconstructions of focal lesions and nodules demonstrated a good correlation between the two- and three-dimensional analyses. Serial sections of large nodules revealed different shapes at different levels of section which could result in spurious calculations.
Dr. Thomas Kensler (Johns Hopkins School of Hygiene and Public Health, Baltimore, Maryland) described the effect of several antioxidants on aflatoxin B₁ carcinogenesis in rats showing an effective inhibition of this process when the antioxidant was administered simultaneously with the carcinogen. Such an inhibition was associated with an induction of glutathione S-transferase, increased conjugates and metabolites of aflatoxin B₁ in the bile and urine, as well as a marked decrease, at least initially, in the DNA adducts formed with the carcinogen. However, the qualitative distribution of adducts was the same whether the antioxidant was present with the carcinogen or not. In more recent studies, a structurally different antioxidant, oltipraz, was seen to induce high levels of several forms of glutathione S-transferase and to protect against the formation of aflatoxin-induced focal lesions containing the placental form of. glutathione S-transferase or gamma-glutamyltranspeptidase as markers.
Dr. Janardan Reddy (North Western University Medical School, Chicago, Illinois) described his pioneering studies on the carcinogenicity of peroxisome proliferating agents. These agents exhibit no DNA-damaging effect but do markedly increase the intracellular levels of peroxisomal enzymes. Neoplasms and focal lesions induced by these agents do not contain the placental form of glutathione S-transferase or gamma-glutamyltranspeptidase as markers. Most enzymes found within the peroxisomes are induced, but not catalase. A 68 kilodalton protein which has the characteristics of a receptor for peroxisome proliferating agents has been isolated, and preliminary experiments suggest that some DNA damage can be identified on chronic administration of these agents. In the discussion it was pointed out that phenobarbital and the peroxisome proliferating agents do not appear to have additive effects in tumor promotion, and peroxisomal proteins induced by these agents contain no obvious signal sequence.
Dr. Yoichi Konishi (Nara Medical College, Nara) described extensive studies on the histochemical levels of the placental form of glutathione S-transferase in rodent and human neoplasms. Squamous cell lung carcinomas in rat and human stained for this enzyme, but poorly differentiated lesions did not. In the hamster, preneoplastic lesions in the pancreas, but not those of acinar cells, were scored by this marker, whereas most ductular adenocarcmomas of the pancreas in the human also expressed this gene product. In human stomach carcinomas signet ring cell tumors did not express this enzyme, whereas most others did. Furthermore, in fetal development but not in the adult stomach, except the fundus, this enzyme is expressed, suggesting that the expression in neoplasms is of oncofetal character.
Dr. Henry C. Pitot (McArdle Laboratory for Cancer Research, Madison, Wisconsin) compared the expression of the placental form of glutathione S-transferase and gamma-glutamyltranspeptidase in multistage carcinogenesis in the liver. The instability of focal lesions promoted by phenobarbital was demonstrated, and the relative phenotypic stability of the placental form of glutathione S-transferase as a marker as compared to gamma-glutamyltranspeptidase was also shown. The latter enzyme was a good marker for the stage of promotion for phenobarbital but not by peroxisome proliferating agents or C.I. solvent Yellow 14, a commercial dye. Cereal-based crude diets induced gamma-glutamyltranspeptidase in the livers of some strains of rats, but when semi-purified diets were used, no such enzyme change occurred. Crude diets appeared to contain inhibitors of initiation by diethylnitrosamine and enhancers of promotion by phenobarbital. The appearance of gamma-glutamyltranspeptidase by histochemical staining is reflected by an increased level of the mRNA of the gene product as judged by Northern blot analysis of gamma-glutamyltranspeptidase-positive cells isolated from livers of animals undergoing multistage hepatocarcinogenesis and also by in situ hybridization. The complete nucleotide sequence of the cDNA for human and rat gamma-glutamyltranspeptidase were compared and shown to have a very high degree of homology, approximately 85% at the amino acid level and better than 90% at the nucleotide level.
In conclusion, the participants felt that the meeting was extremely successful, with each speaker having ample time for presentation and considerable discussion following each paper. Information exchange was active by all participants. From the data presented, it is clear that glutathione S-transferases play a significant role in all stages of neoplastic development and thus should come under consideration both in cancer prevention and in cancer therapy.

I₅-fluoro-2’3-isopropylidene-5’-O-(4-N-acetyl-2,4dideoxy-3,6,7,8-tetra-O-acetyl-1-methoxycarbonyl-D-glycero-a-D-galacto-pyranosyl)uridine



SEMINAR AGENDA AND PARTICIPANTS

(1) SEMINAR ON “DEVELOPMENT OF NEW MEDIUM-TERM BIOASSAYS FOR CARCINOGENS”
Honolulu, Hawaii, December 15-17, 1987

AGENDA

PARTICIPANTS

UNITED STATES

Dr. James Huff
Toxicology Research and Testing Program
National Toxicology Program
National Institute of Environmental Health Sciences
Research Triangle Park, NC 27709

Dr. Daniel S. Longnecker
Department of Pathology
Dartmouth Medical School
Hanover, NH 03755

Dr. Gary D. Stoner
Department of Pathology
Medical College of Ohio
3000 Arlington Avenue
Building 3, Room 202
Toledo, OH 43614

Dr. Stan D. Vesselinovitch
Department of Radiology & Pathology
University of Chicago
5841 Maryland Avenue
Chicago, IL 60637

Dr. James A. Popp
Experimental Pathology & Toxicology
Chemical Industry Institute of Toxicology
Research Triangle Park, NC 27709

Dr. Thomas J. Slaga
University of Texas Cancer Center
Science Park-Research Division
P.O. Box Drawer #179
Smithville, TX 78957

Dr. Jerrold M. Ward
Tumor Pathology & Pathogenesis Section
Laboratory of Comparative Carcinogenesis
Division of Cancer Etiology
National Cancer Institute
Frederick, MD 21701

JAPAN

Dr. Ryohei Hasegawa
Department of Pathology
National Institute of Hygienic Sciences, Tokyo

Dr. Nobuyuki Ito
Department of Pathology
Nagoya City University Medical School, Nagoya

Dr. Hideki Mori
Department of Pathology
Gifu University School of Medicine, Gifu

Dr. Hiroko Ohgaki
Biochemistry Division
National Cancer Center Research Institute Tokyo

Dr. Masae Tatematsu
Department of Pathology
Nagoya City Univeristy Medical School, Nagoya

Dr. Hiroyuki Tsuda
Department of Pathology
Nagoya City University Medical School, Nagoya

Dr. Shigeru Yoshizawa
Cancer Prevention Division
National Cancer Center Research Institute, Tokyo



(2) CANCER CELL MEMBRANES: ABERRANT GLYCOSYLATION AND OTHER CRITICAL MOLECULAR EVENTS
Makaha Valley, Oahu Island, Hawaii
February 15-18, 1988

AGENDA

PARTICIPANTS
UNITED STATES

Sen-itiroh Hakomori
The Biomembrane Institute
201 Elliott Ave. W.
Seattle, WA 98119

John Magnani
Laboratory of Structural Biology
National Institute of Diabetes and Digestive and Kidney Diseases
Bldg. 4, Rm. 337
Bethesda, MD 20892
George Martin

Laboratory of Developmental Biology
National Institute of Dental Research
Bldg. 30, Rm. 414
Bethesda, MD 20892

Tatsuro Irimura
Department of Tumor Biology
M.D. Anderson Hospital
6723 Bertner
Houston, TX 77030

David Cheresh
Department of Immunology
Scripps Clinic and Research Foundation
La Jolla, CA 92037

Kenneth Yamada
Membrane Biochemistry Section
Laboratory of Molecular Biology
NCI, NIH, Bldg. 36, Rm. 1D32
Bethesda, MD 20892

Sidney Welt
Sloan-Kettering Institute
1275 York Avenue
New York, NY 10021

Stephan Ladisch
Department of Pediatrics
UCLA School of Medicine
Los Angleles, CA 90024

OBSERVERS
UNITED STATES

Phillip O. Livingston
Sloan-Kettering Institute
1275 York Avenue
New York, NY 10021

Victor Ginsburg
Laboratory of Structural Biology
NIDDK, NIH, Bldg. 4, Rm. 337
Bethesda, MD 20892

Garth Nicolson
Department of Tumor Biology
M.D. Anderson Hospital
6723 Bertner
Houston, TX 77030

Eric Holmes
Pacific Northwest Research Foundation
l102 Columbia Street
Seattle, WA 98104

Gary Ostrander
Pacific Northwest Research Foundation
1102 Columbia Street
Seattle, WA 98104

Minoru Fukuda
La Jolla Cancer Research Foundation
10901 N. Torrey Pines RD.
La Jolla, CA 92037

PARTICIPANTS
JAPAN

Ikuo Yamashina
Professor and Chairman
Faculty of Pharmaceutical Sciences
Kyoto University
Shimoadachi-cho, Sakyo-ku
Kyoto 606, Japan

Yoshitaka Nagai
Professor and Chairman
2nd Department of Biochemistry
University of Tokyo
7-3-1 Hongo, Bunkyo-ku
Tokyo 113, Japan

Toshiaki Osawa
Professor and Chairman
Department of Immunochemistry and Toxicology
Faculty of Pharmaceutical Sciences
University of Tokyo
7-3-1 Hongo, Bunkyo-ku
Tokyo 113, Japan

Reiji Kannagi
Kyoto University School of Medicine
Sakyo-ku, Kyoto 606, Japan

Masaru Taniguchi
54 Shyogoin-Kawaramachi
Professor and Chairman
Department of Immunology
Chiba University School of Mdicine
1-8-1 Inohana
Chiba 281, Japan

Setsuo Hirohashi
Head, Department of Pathology
National Cancer Research Institute
5-1-1 Tsukiji, Chuo-ku
Tokyo 104, Japan

Akira Yachi
Professor and Chairman
Department of Medicine
Sapporo University School of Medicine
South 1st St., West 17th Ave. Chuo-ku, Sapporo
Hokkaido 060, Japan

Takashi Muramatsu
Professor and Chairman
Department of Biochemistry
Kagoshima Univ. School of Medicine
1208-1 Ubusuki-machi
Kagoshima 890, Japan

OBSERVERS
JAPAN

Kazo Imai
Assistant Professor of Medicine
Sapporo Univ. School of Medicine
South 1st St., West 17th Ave.
Chuo-ku, Sapporo
Hokkaido 060, Japan

Masao Iwamori
2nd Department of Biochemistry
University of Tokyo
7-3-1 hongo, Bunkyo-ku
Tokyo 113, Japan

Tadashi Tai
Tokyo Metropolitan Institute of Medical Sciences
3-18-22 honkomagome Bunkyo-ku
Tokyo 113, Japan

Masahiko Watanabe
Department of Surgery
Keio University School of Medicine
35 Shinanomachi, Shinjuku-ku
Tokyo 160, Japan

Yasuo Fukushi
Department of Urology
Tohoku University School of Medicine
1-1, Seiryo-cho Sendai, Japan

Toshisuke Kawasaki
Associate Professor
Faculty of Pharmaceutical Sciences
Kyoto University
Shimoadachi-cho, Sakyo-ku
Kyoto 606, Japan



(3) SEMINAR ON BIOLOGY OF ONCOGENES
Coco Palms, kauai, Hawaii, February 17-18, 1988

PROGRAM