REPORTS ON SEMINARS

(1) Seminar on “Marine Natural Products and Cancer Chemoprevention”
This seminar was held on February 11, 12 and 13, 1987 at the East-West Center of the University of Hawaii in Honolulu, Hawaii. The organizers were Dr. Richard Moore, Department of Chemistry, University of Hawaii, Honolulu, and Dr. Hirota Fujiki, Cancer Prevention Division, National Cancer Center Research Institute, Tokyo. There were nine participants from the United States and six from Japan. The purpose of the meeting was to discuss and exchange ideas and information concerning the biology, chemistry and pharmacology of tumor promoters and inhibitors of tumor promotion from natural sources, in particular from field-collected marine organisms and cultured microalgae. The first day of the meeting was devoted principally to tumor promotion, while the second day was concerned with inhibition of tumor promotion. On the third day, a roundtable discussion was held to determine what new directions are needed in this area of research, to critically evaluate existing and new bioassays for the rapid detection of tumor promoters and inhibitors of tumor promotion, and to establish mechanisms for collaborative research efforts among United States and Japanese scientists.
Opening remarks, prepared by Dr. Takashi Sugimura, President of the National Cancer Center, Tokyo, and Japanese Coordinator of the Etiology Program Area of the U.S.-Japan Cooperative Cancer Research Program, were delivered by Dr. Hirota Fujiki.
Tumor Promotion: Chemistry and Biology
Dr. Hirota Fujiki (National Cancer Center Research Institute, Tokyo) presented a review of tumor-promoting natural products that have been studied in his laboratory. Essentially two distinct types of mouse skin tumor promoters exist. One type elicits its diverse biological effects in vitro and in vivo by initially binding to a receptor that activates protein kinase C. This receptor is commonly referred to as the phorbol ester receptor. Three chemically distinct classes of natural products have been found that bind to the same phorbol ester receptor in mouse skin, viz. the phorbol esters of which 12-0-tetradecanoyl-phorbol-13-acetate (TPA) is the most well known, the teleocidins and aplysiatoxins. Hence, these three classes of tumor promoters are referred to as TPA-type tumor promoters. The second type elicits its biological responses through mechanisms that do not involve activation of protein kinase C. Three natural products that belong to this second type, which can be called non-TPA-type tumor promoters, are palytoxin, thapsigargin and okadaic acid. The biological effects which are common to both the TPA and non-TPA-type tumor promoters that Dr. Fujiki studied are irritation of mouse ear, stimulation of arachidonic acid metabolism, and stimulation of super-oxide anion radical production.
Dr. Richard Moore (University of Hawaii, Honolulu, Hawaii) presented a brief review of the biology of hepatotoxic blue-green algae and the chemistry and pharmacology of the toxins that have been isolated from species that grow in the temperate regions of the world and are major public health and agricultural concerns. For example, nodulatoxin is a cyclic penta-peptide in Nodularia spumigina, a species which grows in estuaries and brackish-water seas and lakes, and the microcystins (synonymous with cyanoginosins) are cyclic heptapeptides in Microcystis aeruginosa, Anabaena flos-aquae and Oscillatoria agardhii, species that grow in freshwater reservoirs, lakes and ponds. Based on a study by Ian Falconer in Australia of the effects of toxin-containing drinking water on the liver and the internal organs of laboratory mice, Dr. Moore suspects that these toxins may be a new type of tumor promoter.
Dr. Moore also reported preliminary results on screening crude extracts of cultured blue-green algae for tumor-promoting activity using the HL-60 cell adhesion assay.
Dr. Mitsuru Hirota (National Cancer Center Research Institute, Tokyo) discussed the results of a structure-activity study on (-)-indolactam-V, the proposed intermediate in the biosynthesis of teleocidins A and B. The biological activity of indolactam-V has been found to be highest when the absolute configurations of the two amino acid residues (tryptophan and valine) are both L, i.e., 9S, 12S. Increasing the lipophilicity of the valine unit, e.g., replacing the isopropyl group with a t-butyl group (t-leucine), increases biological and biochemical activities related to tumor promoting activity, whereas decreasing the lipophilicity, e.g., replacing the isopropyl group with a methyl group (alanine) or a hydrogen (glycine), decreases biological and biochemical activities indicative of tumor promoting activity. Substitution of an acyl group at C-7 enhances activity, but substitution of an acyl group at C-2 or C-5 greatly diminishes the potency of (-)-indolactam-V. (-)-Indolactam-V and 7-substituted (-)-indolactam-Vs elaborate a twist conformation for the lactam ring, whereas the lactam ring of 5-acetyl-(-)-indolactam-V exists in a sofa conformation. For tumor promoting activity, it has been suggested that the lactam ring has a twist conformation.
Dr. Lawrence Levine (Brandeis University, Waltham, Massachusetts) discussed the stimulation of arachidonic acid as a possible mechanism of action of tumor promoters. He pointed out that the activities of several growth factors and the TPA-type tumor promoters with respect to arachidonic acid metabolism, are synergistic, as are the activities of the cells after interaction of the biologically active ligands with their specific receptors and the TPA-type tumor promoters. Turnover of PIP₃, stimulated by several growth factors and ligand-receptor interaction and the resultant bifurcation that arises by formation of both diacyl-glycerol, which activates protein kinase C, and 1,4,5-inositol-tris-phosphate, which mobilizes Ca²⁺, as well as the direct activation of protein kinase C by TPA and 1-oleoyl-2-acetyl-glycerol (OAG), probably accounts for their arachidonic acid metabolic activities. Unlike treatment with OAG and TPA, treatments with palytoxin and OAG or palytoxin and TPA or palytoxin and growth factors are synergistic with respect to arachidonic acid metabolism.
It is important to note that combinations of ouabain and growth factors, ouabain and TPA or ouabain and OAG resemble combinations of palytoxin and growth factors, palytoxin and TPA, or palytoxin and OAG in their synergistic arachidonic acid metabolism, suggesting that one of the effects0 of palytoxin is inhibition of the Na⁺, K⁺-ATPase.
Dr. Levine also discussed some immunochemistry of palytoxin and the properties of polyclonal antibodies to palytoxin from immunized rabbits. The serum from one rabbit blocked palytoxin’s stimulation of arachidonic acid metabolism in several cells, neutralized its hemolytic activity toward rat erythrocytes, and neutralized its lethal activities in mice.
Dr. Marsha Rosner (Massachusetts Institute of Technology, Cambridge, Massachusetts) reported on the regulation of the EGF receptor by palytoxin. Previous studies have shown that the EGF receptor is a sensitive responder to TPA-type tumor promoters. These agents regulate the action of the EGF receptor in cultured cells by inhibiting high affinity EGF binding and EGF-stimulated tyrosine kinase activity. The mechanism by which this modulation of the receptor occurs appears to be, at least in part, through direct phosphorylation of the receptor by protein kinase C. Thus, the EGF receptor system can be used to investigate the mechanistic action of these tumor promoters. Dr. Rosner’s group has now extended these studies to determine the effect of palytoxin, a non-TPA-type tumor promoter on the action of the EGF receptor. Palytoxin inhibits EGF binding to its receptor in at least two cell types. The inhibition appears to be independent of the level of protein kinase C in the cell, indicating the palytoxin can act on the receptor through a protein kinase C-independent mechanism. Unlike TPA-type tumor promoters which inhibit EGF binding in a rapid reversible and dose-dependent manner, palytoxin inhibits EGF binding with slower kinetics and is not readily reversible. The extent of inhibition of EGF binding reaches a maximal level almost independent of the initial dose of palytoxin, suggesting that palytoxin acts by disturbing ion fluxes.
Dr. Masami Suganuma (National Cancer Center Research Institute, Tokyo) reported on the tumor promoting activity of okadaic acid, a diarrhetic toxin isolated from the black sponge Halichondria okadai. Okadaic acid was found to be a non-TPA- type tumor promoter since it did not inhibit the specific binding of ³H-TPA to a mouse particulate fraction, nor did it activate protein kinase C isolated from mouse brain. Unlike palytoxin and thapsigargin, two other non-TPA-type tumor promoters, okadaic acid induced ornithine decarboxylase (ODC) activity in mouse skin. The ODC activity reached a maximum in 4-5 hours. Pretreatment of the skin with 13-cis-retinoic acid caused 98% inhibition of ODC induction by okadaic acid. A semisynthetic derivative, okadaic acid tetramethyl ether, did not induce ODC activity in mouse skin, suggesting that the hydroxyl groups of okadaic acid are necessary for ODC induction. Okadaic acid also induced histidine decarboxylase (HDC) activity as did thapsigargin; palytoxin, however, did not induce HDC activity. In a two-stage carcinogenesis experiment on mouse skin, okadaic acid exhibited tumor promoting activity as strong as teleocidin. Dr. Suganuma suspects that the three non-TPA-type tumor promoters, okadaic acid, palytoxin and thapsigargin, induce tumor promotion by different pathways.
Dr. Takeshi Yasumoto (Tohoku University, Sendai) presented a status report on his work with maitotoxin (MTOX), an exceedingly potent marine toxin isolated from the dinoflagellate Gambierdiscus toxicus. To date, 51 milligrams of pure MTOX have been isolated from 8000 liters of cultured G. toxicus. MTOX has two hydroxysulfate esters in the molecule and possesses at least one nitrogen positive to the Dragendorff reagent. The negative FAB mass spectrum suggests a molecular weight of 3424 for the disodium salt. The carbon-13 NMR spectrum indicates 159 carbon signals; none of these signals is assignable to acetal, ketal or carbonyl carbons. The proton NMR spectrum indicates the presence of 21 methyl groups, a conjugated diene with an exomethylene, a vinyl group and a disubstituted olefin. MTOX is highly lethal to mice (150 ng/kg) and fish, hemolyzes mouse blood cells, and activates Ca⁺⁺ channels. This toxin is being investigated by Dr. Fujiki’s group for possible tumor promoting properties.
Dr. Yasumoto also reported on the structure, distribution, and etiology of other polyoxygenated marine toxins. Palytoxin occurs in several crabs and fishes, and is responsible for many human intoxications. Okadaic acid, 35-methylokadaic acid, and 7-0-acyl-35-methylokadaic acid are accumulated in shellfish and cause a worldwide health hazard called diarrhetic shellfish poisoning. The dinoflagellates Dinophysis fortii, D. acuminata, and Prorecentrum lima were found to be the biogenetic origin of these polyether toxins. P. lima were also found to produce many diol esters of okadaic acid and 7-deoxyokadaic acid. A highly sensitive fluorometric HPLC method was developed to determine okadaic acid analogues in biological materials. The structure of toxic polyethers from scallops, viz. the pectenotoxins and yessotoxins, were presented.
Inhibition of Tumor Promotion
Dr. John Bertram (University of Hawaii, Honolulu, Hawaii) discussed the development of biological assays for determining the biochemical mechanisms of action of chemopreventive agents such as the retinoids. In his system of 10T1/2 mouse fibroblasts, retinoids are potent inhibitors of carcinogen-induced neoplastic transformation. Retinoids exert two actions on carcinogen-initiated 10T1/2 mouse fibroblasts, i.e., one where the initiated cell is stabilized in a non-neoplastic state and the other where the degree of cell/cell communication between growth inhibited cells is enhanced. Both of these actions have the effect of preventing the development of neoplastic transformed foci, and both of these actions are counteracted by certain classes of tumor promoters. The TPA-type tumor promoters enhance neoplastic transformation in 10T1/2 cells induced by 3-methylcholanthrene (MCA), but the non-TPA-type tumor promoter, palytoxin, does not.
Dr. Yasutoshi Muto (Gifu University School of Medicine, Gifu) reported on the antitumor activity of acyclic retinoids. A synthetic polyprenoic acid derivative, 3,7,11,15-tetramethyl-2,4,6,10,14-hexadecapentaenoic acid (E-5166), was found to exhibit a strong binding affinity to cellular binding proteins specific for acidic retinoids. Oral administration of E-5166 significantly reduced tumor incidences of experimental hepatomas induced by 3’-methyl-4-dimethylaminoazabenzene (3’-MeDAB) in rats as well as in “spontaneous” hepatoma-bearing mice (C3H/HeNCrj). Furthermore, skin papillomas in mice (number per animal) induced by DMBA treatment regressed significantly by administration of E-5166 through a stomach tube. Compound E-5166 proved to have an inhibitory activity on carcinogenesis of the liver and skin with a better therapeutic index for long-term use compared with other retinoids. The in vitro inhibitory effect of E-5166 on the human hepatocellular carcinoma cell line PLC/PRF/5 was also presented. With regard to the molecular mechanism of anti-tumor activity of E-5166, Dr. Muto’s work indicates that local deficiency of retinoids results exclusively in hepatomas.
Dr. Luigi De Luca (National Cancer Institute, Bethesda, Maryland) reported on the inhibition of tumor promotion through maintenance of normal differentiation with retinoids. Chemical respiratory carcinogens such as benzo[a]pyrene caused squamous metaplasia in cultural tracheas from male Syrian golden hamsters. The squamous metaplasia lesion was accompanied by the expression of a keratin protein having a molecular weight of 60,000 daltons. The expression of the 60 kD keratin was also observed in trachea obtained from hamsters kept 4 weeks on a vitamin A-deficient diet and cultured for 2 weeks in a retinoid-free medium. Other acidic and basic keratins were induced in the retinoid-deprived cultures. The studies indicated that vitamin A deficiency and carcinogen exposure elicit the expression of a similar differentiation program in the respiratory epithelium. Dr. De Luca also reported that tumors are not produced in the dorsal skin of vitamin A-deficient mice in a two-stage carcinogenesis experiment where the mice have been initiated with DMBA and subsequently treated with TPA. If retinoid is administered to the vitamin A-deficient mice along with the TPA applications, however, tumors are produced. This experiment indicated that retinoids in the diet are essential for the TPA-induced tumor promotion.
Dr. Ronald Merriman (Lilly Research Laboratories, Indianapolis, Indiana) reported on the chemopreventive action of sceptrin and oroidin. These two indole alkaloids that had been isolated by Walker and Faulkner at Scripps Institution of Oceanography from the Caribbean sponge Ageias sceptrin. Using cultured 10T1/2 cells, Dr. Merriman found that sceptrin inhibits the induction of ODC by TPA and the induction of malignant transformation by MCA. Oroidin was less active than sceptrin. Since adequate amounts of sceptrin and oroidin were not available for in vivo testing, several simpler compounds were synthesized for structure-activity studies. One synthetic compound, N,N’-di-(4-bromoprolyl)-1,4-butanediamine, Lilly No. 171325, was found to have activity in vitro equal to that of sceptrin. On a molar basis, sceptrin was found to be as potent as retinyl acetate in inhibiting the induction of malignant transformation in 10T1/2 cells by MCA. Subsequently, compound 171325 was found to inhibit the growth of urethane-induced pulmonary adenomas in strain A/J mice.
Dr. Ajit Verma (University of Wisconsin Clinical Cancer Center, Madison, Wisconsin) discussed the use of mouse skin for rapid screening of environmental tumor promoters and cancer chemopreventive agents. One of the properties ascribed to tumor promoters is the ability to induce ODC, the first and rate-limiting enzyme in the pathway to polyamine biosynthesis. Strong evidence for the role of ODC induction in skin tumor promotion by TPA has been provided by the finding that alpha-difluoromethyl-ornithine, a suicide inhibitor of ODC, when given in drinking water in conjunction with TPA, inhibits the formation of mouse skin tumors. Also, certain retinoids and inhibitors of arachidonic acid metabolism (e.g., indomethacin and nordihydroguaiaretic acid), which inhibit ODC induction by TPA, inhibit skin tumor promotion. These results indicate that induction of mouse epidermal ODC can be used to screen for tumor promoters and inhibition of TPA-induced ODC can be used to screen for anti-tumor promoters, at least of the TPA-type. Palmitoylcarnotine was found to be an inhibitor of TPA-induced ODC gene expression and tumor formation. Evidence, however, is accumulating that suggests that ODC induction is not a sufficient condition for tumor promotion. Dr. Verma also confirmed that 7-bromomethylbenz[a]anthracene is a tumor promoter.
Dr. William Fenical (Scripps Institution of Oceanography, University of California, La Jolla, California) discussed his work on the development of anti-inflammatory drugs from gorgonian corals found in the warm eastern Atlantic Ocean, in particular within the Caribbean Sea. The isolation and structure determination of the erythrolides from Erythropodium caribaeorum, pseudopterolide from Pseudopterogorgia acerosa, and the pseudopterosins and secopseudopterosins from P. elisabethae were discussed. All of these compounds are potent inhibitors of TPA-induced inflammation in mouse ear. None, however, appears to be an inhibitor of cyclooxygenase. Dr. Fenical pointed out that it is interesting that prostaglandins have also been found in gorgonian corals, for example, in Plexaura homomalla by the Weinheimer group.
Dr. Robert Jacobs (University of California, Santa Barbara, California) reported on the anti-inflammatory activity of manoalide, a sesterterpene isolated from the sponge Luffariella variabilis. The structure of manoalide was first reported by Dr. Paul Scheuer’s group at the University of Hawaii, but the unusual pharmacological properties of this compound were discovered by Dr. Jacobs, working in collaboration with Dr. John Faulkner’s group at Scripps Institution of Oceanography. Manoalide is a potent antagonist of TPA-induced inflammation and suppresses papilloma formation in a two-stage carcinogenesis experiment where mouse skin is treated with a single subcarcinogenic dose of 7,12-dimethylbenz[a]anthracene (DMBA) followed by multiple applications of TPA. Manoalide is a potent irreversible inactivator of phospholipase A₂ and only one of three that are known in the literature. Binding to lysine residues appears to be involved in the suicide inactivation of the enzyme.
Dr. Masaru Kobayashi (Hokkaido University, Sapporo) reported on the chemistry of sarcophytol A and related cembranoids from the soft coral Sarcophyton glaucum. Sarcophytol A is a simple monohydroxycembratetraene which was isolated about 10 years ago by Dr. Kobayashi. The S. g1aucum from Ishigaki Island, Okinawa, contains a large amount of this powerful anti-tumor promoter along with minor amounts of the corresponding acetate and sarcophytols B-E. Sarcophytol A could be converted to a 7,8-epoxide derivative which provided the starting material for the preparation of a tritium-labeled sarcophytol A for receptor binding and metabolic studies.
Dr. Hirota Fujiki reported on the anti-tumor promoting activity of sarcophytol A. This study was prompted by a report in the literature on the isolation of anti-tumor promoting cembranoids from cigarette smoke condensate. Sarcophytol A was found to inhibit teleocidin-induced hyperplasia and the tumor promoting activity of teleocidin in mouse skin at low concentrations. Interestingly, sarcophytol A did not inhibit the specific binding of ³H-TPA to phorbol ester receptors in cell membranes or the activation of protein kinase C in vitro. Sarcophytol B exhibited the same potency as sarcophytol A and sarcophytol A acetate was only slightly weaker in its tumor promoting activity. Compound Y and the corresponding methyl ester, which were derived from sarcophytol A by oxidatively opening the membrane ring, did not inhibit the tumor promoting activity of teleocidin. ³H-Sarcophytol A did not bind to an epidermal particulate fraction prepared from mouse skin. Most of the ³H-sarcophytol A applied to mouse skin was excreted rapidly in the urine and feces; very little accumulated in the liver and other internal organs. Sarcophytol A inhibits spontaneous mammary tumor formation in SHN mice.
Dr. Fujiki also reported on the anti-tumor promoting activities of N-(6-aminohexyl)-5chloro-1-napthalenesulfonamide (W-7), quercetin, glycyrrhetic acid, berberine sulfate, penta-O-galloyl-beta-D-glucose, and (-)-epigallocatechin gallate. Unlike sarcophytol A, however, three orders of magnitude more material are needed for activity.

(2) Transacting Factors in Viral Carcinogenesis
This seminar was held on March 11-13, 1987 at the Park Hotel, Kyoto, Japan. The organizers were Dr. Mitsuaki Yoshida, Cancer Institute, Japanese Foundation for Cancer Research, Tokyo; Dr. Akira Hakura, Osaka University, Osaka; and Drs. Peter M. Howley and Robert Gallo, National Cancer Institute, Bethesda, Maryland.
There were six participants from the United States and ten participants from Japan. Originally, seven participants were planned from the United States but, unfortunately, Dr. Gallo was unable to participate. The purpose of the seminar was to discuss and exchange information on the molecular mechanisms involved in transcriptional transactivation of a group of viruses believed to be involved in carcinogenic progression. The focus of the seminar was on the molecular mechanisms of gene transactivation and the possible role of these functions in oncogenesis. In the opening remarks by Drs. Yoshida and Howley, each stressed the rapid progress being made in viral carcinogenesis and the likely important role that viral transactivating functions may play in the malignant process.
Dr. Peter Howley (National Cancer Institute, Bethesda, Maryland) discussed transcriptional transactivating factors in the regulation of papillomavirus gene expression. He discussed the bovine papillomavirus (BPV-1) and the role of the E2 gene product in the transactivation of enhancer elements. The E2 gene product is DNA binding protein and binds specifically to a sequence, ACC(N)₆GGT. Deletion mutagenesis of the long control region (LCR) revealed that these binding sites are critical components for E2 transactivation, implicating direct DNA binding by E2 to the target as critical for transactivation. Dr. Howley demonstrated that the E2 gene of HPV-16 similarly encodes a transactivating factor which targets the HPV-16 LCR. BPV-1 also encodes a transacting transcriptional repressor which can inhibit BPV-1 transformation. The repressor is encoded by the carboxy terminal domain of the E2 gene and is expressed from an alterative promoter located within the body of the E2 gene. Models for repression were discussed. A likely mechanism, however, involves competitive binding to the LCR enhancer elements since the E2 repressor protein is also a DNA binding protein which binds to the same sequences as the full E2 protein. It is of interest that in the cervical carcinomas containing HPV DNA, integration generally occurs in a manner to disrupt the E2 open reading frame. Thus, one possible event in HPV-associated carcinomas in humans may be the release of the viral expression from control by the E2 genes due to the integration disruption of the E2 genes.
Dr. Akira Hakura (Osaka University, Osaka) reported on transformation of rodent cells by HPV-16. He demonstrated that, like BPV-1, HPV-16 encodes a transactivating factor which can transactivate an element within the viral LCR. He preliminarily mapped this function to the E2 open reading frame of the HPV-16 genome. He inserted the early region of HPV-16 containing the E6 and E7 open reading frames into a mouse retrovirus vector, pZIPneo, and transfected it into psi2 cells. Construction with sense orientation induced transformed phenotypes of the cells; however, insertion in antisense orientation did not. The viruses released from these psi2 cells also included transformation on newly infected cells.
Dr. Masaaki Terada (National Cancer Center, Tokyo) reported on the transforming activity of human papillomavirus. He reported on the analysis of HPV DNA cell lines established from cervical cancers in Japan. Two of these lines, SKG I and SKG II, contained a low copy number of HPV-18 genomes integrated and two other lines, SKG IIIA and SKG IIIB, contained HPV-16 DNA integrated. In analysis of RNA in these cell lines, the E6/E7 region of the viral genomes was found to be expressed. Transcription was also studied for the HeLa cell line, which is a human cervical carcinoma cell line known to contain approximately ten copies of HPV-18 DNA. Two transcripts, 1.6 and 3.4 Kb, were expressed from a promoter upstream of the E6/E7 region and were hybrid RNAs containing human cellular sequences at their 3’ end. Using the cloned 3’ human cDNA sequences as a probe, he demonstrated that there was no expression of the sequence in the other cervical carcinoma cell lines. Dr. Terada also reported on the transforming activity of DNA from a human cervical adenocarcinoma which contained HPA-16 DNA. Total genomic DNA induced transformation on NIH 3T3 cells after transfection and the transforming DNAs were cloned into cosmids. These clones contained HPV-16 DNA with some flanking human sequences.
Dr. Shunsuke Ishii, RIKEN, Tsukuba, Japan, discussed the function of the myb gene product. The viral oncogene myb in AMV is a truncated version of its cellular counterpart, which is expressed in immature cells of hematopoietic cell lineage. He constructed an expression of plasmid of c-myb using the beta-actin promoter and transfected it into CV-1 cells. Expression of c-myb activated promoters of type-1 and -II collagens, EGF receptor and beta-actin genes in CAT constructs, but did not affect the SV40 early promoter or LTR promoter of bovine leukemia virus. On the other hand, the c-myb protein repressed expression of the C-Ha-ras 1 gene. He constructed a c-myb mutant truncated at the 5’ and 3’ ends of the coding sequence that mimics the v-myb gene. The truncated c-myb repressed the SV40 early promoter with enhancer, but normal c-myb did not. The effect of truncation as detected on the SV40 enhancer may give a clue to the mechanism of tumorigenesis induced by v-myb.
Dr. Mitsuaki Yoshida, Cancer Institute, Tokyo, discussed the transactivating factors of HTLV-I in the development of adult T-cell leukemia. The pX region of HTLV-I contains four small open reading frames. In his studies, he focused primarily on the most distal of these open reading frames which encodes a 40 kilodalton protein, p40X. This protein is found in the nucleus and functions to transactivate an element located in the viral LTR. The target for p40X in the LTR is between nt -322 and nt -55. The target is orientation-independent and contains three 21 base pair repeats. Transfection of p40X-expression plasmid into certain T-cell lines induced an efficient expression of interleukin-2 receptor (IL-2R). The IL-2 gene was also activated although it was weak. These activations of the cellular IL-2 receptor gene were specifically observed in certain T-cell lines such as Jurkat and HSB-2, but not in the other T-cell lines, CEM or Molt3. Dr. Yoshida suggested that these activations of IL-2 and IL-2R may account for unusual growth initiation of HTLV-I-infected T-cells and discussed possible early events in the development of adult T-cell leukemia. To explain an apparent contradiction that primary ATL cells do not express p40X, although they express IL-2R, he proposed a multi-step model for development of ATL.
Dr. Kunitada Shimotohno (National Cancer Center, Tokyo) discussed the transcriptional transactivation of HTLV-II gene expression. HTLV-II is quite analogous in organization to HTLV-I. The 21 base pair repeats are highly conserved between HTLV-I and HTLV-II and are located upstream of the TATA box in the U3 region of the LTR. The deletion of these 21 base pair repeats destroys the ability of HTLV-II LTR to serve as a target for the HTLV-II TAT transactivation. Replacement of these 21 base pair repeats restores transactivation function. In addition, the pX region of HTLV-II encodes two other proteins, 24 Kd and 26 Kd In size. Dr. Shimotohno then discussed how these proteins may be expressed.
The transactivational regulation of HTLV-I in replication was discussed by Dr. Motoharu Seiki of the Cancer Institute in Tokyo, Japan. The pX region of HTLV-I encodes a series of proteins. One open reading frame encodes the p40 or TAT protein. Another open reading frame encodes two proteins, pp27 and pp21. Expression of all of these proteins is from the same mRNA. Only p40X is required for transcriptional transactivation. He constructed defective proviruses containing defects in the pX sequences and cotransfected with a series of pX expression plasmids which can express each pX protein. He presented clear evidence that both p27X and p40X are required for the gag and envelope gene expression. With p40X alone, the proviral genome was transcribed, but only spliced mRNA was detected. Complimentation with p27X induced accumulation of unspliced mRNA. p27X, therefore, is involved in the posttranscriptional control of gag and env gene expression modulating either splicing, transport, or stability of the viral RNA, but not affecting the translation.
Dr. Flossie Wong-Staal (National Cancer Institute, Bethesda, Maryland) discussed trans-regulatory genes of HIV and AIDS pathogenesis. The genomic organization of HIV is slightly different from that of HTLV-I and HTLV-II. It is a member of the lentivirus class of retroviruses and contains a series of small open reading frames that encode factors that are involved in viral gene regulation. One of these, called SOR, encodes a p23 protein. Another, called 3’ ORF, encodes p27. Two other genes are involved in transcriptional regulation: TAT and the trans-regulator of splicing (TRS). By analysis of viral mutants and by in vitro transcription experiments using HeLa cell extracts, Dr. Wong-Staal was able to show that a major component of the TAT transactivation is at the transcriptional level. However, there also may be a translational component. TAT mutants do not make RNA. It was speculated that TAT may be a direct nucleic acid binding protein since it is proline-rich and contains the motif Cys-X-X-Cys three times. It was speculated that these motifs may be components of a metal binding finger. In addition, it contains a run of very basic amino acids. These features of the TAT protein are also predicted by the sequence for the STLV-III virus. The TRS mutants have full transactivation function but fail to make gag or envelope proteins. TRS mutants make only the small spliced RNAs and the overall amount of RNA made is increased. Dr. Wong-Staal presented evidence that the TRS gene may regulate expression of the gag and env genes by preventing the splicing of the larger messenger RNA which normally encodes the gag and env proteins.
Dr. Yoji Ikawa (RIKEN, Tsukuba) discussed the bovine leukemia virus, the X gene product, and their targeted genes. The bovine leukemia virus is analogous in structure to HTLV-I and HTLV-II and is a good animal model for studying those viruses. It also contains an X region and its pX gene product is a viral transcriptional transactivator. The gene product of the pX gene is a 38 Kd protein. He constructed a plasmid expressing only p38X and demonstrated that the p38X can activate the cognate LTR. He tried to demonstrate direct binding of p38X to the LTR sequence by gel retardation assay under various conditions; however, he could not clearly show a specific binding.
Dr. Arnie Berk (University of California at Los Angeles, Los Angeles, California) described studies on the mechanism of adeno E1a transactivation. This viral gene product can stimulate transcription by Pol III and Pol II indirectly by increasing the activity of a number of host transcription factors. The targets for Pol III genes include the adenovirus VA1 and VA2 genes. Polymerase II genes can also serve as targets, including the adenovirus E2, E3, and E4 genes. In addition, the adeno E1b promoter is a target for the E1a gene product. Evidence was presented that adeno E1a can work through a variety of host factors. It may, in some cases, work through the TATA box factor. The adeno E1a gene encodes two proteins through two different RNAs by differential splicing. Adeno E1a is not a DNA binding protein, although it is predicted to have metal binding fingers. Through the study of the adeno major late promoter in vitro, Dr. Berk was able to conclude that the adeno E1a protein does not have a direct effect on transcription, but rather works through activation of host factors. The effect of adeno E1a on Pol III gene expression is through the increase of TF-III C activity in infected cells.
Dr. Kei Fujinaga (Sapporo Medical College, Sapporo) discussed adenovirus transforming genes and trans-regulation. Adeno E1a proteins are involved in transcriptional transactivation, transcriptional repression, stimulation of DNA synthesis, stimulation of EGF, the modulation of class I histocompatibility genes, immortalization, and morphologic transformation. He achieved various mutagenes at the 5’ region of the E1A gene, which did not induce frame shift mutation, and examined the functions of these mutated E1A proteins in transformation of a rat cell line 3Y1 and transactivation of an E2 promoter-CAT construct. He could separate capacities of focus formation on 3Y1 cells and activation of the E2 promoter.
Dr. Joe Nevins (Rockefeller University, New York, New York) described transcriptional control by the adenovirus E1a oncogene. The E1a proteins are heavily phosphorylated. They cooperate with adeno E1b in transforming cells, can immortalize primary cells, can cooperate with ras to transform primary cells, and activate early viral transcription in the lytic viral infection. He used the adeno E2 promoter to score for adeno E1A transactivation. Using DNase footprinting and gel retardation techniques, he was able to map regions upstream of the CAP site of the adeno E2 promoter which is the target for the adeno E1a function. Two domains that are critical for the adeno E1a activation were found with the sequence TTTCGCGC. He was able to show, by treating with protein inhibitors such as cyclohexamide, that transcriptional activation through this target sequence does not require protein synthesis. By competition experiments, he was able to show that the factor involved in the E2 transcriptional transactivation cannot be competed by other adeno E1a responsive promoters Therefore, adeno E1a works through multiple factors. It can, however, be competed by the adeno E1a promoter region containing these repeated elements. Adeno E1a can also activate cellular genes including HSP70. In attempting to define what cellular factors are analogous to E1a, he has found that non-differentiated F9 teratocarcinoma cells contain an adeno E1a-like activity.
Dr. Yoshiaki Ito (Kyoto University, Kyoto) described analysis of transacting factors in embryonic cells using polyomavirus mutant enhancers. Dr. Ito has been using polyoma mutants to study factors in embryonic cells that are involved in transcriptional regulation. A series of mutant polyoma enhancer elements were selected for their ability to grow on differentiated F9 teratocarcinoma cells. His thesis is that the enhancer may be a target for negative regulatory factors and that the wild type polyoma enhancer is, therefore, negatively regulated within F9 cells. Mutants, therefore, that can be propagated in these differentiated F9 cells may have been knocked out for their ability to interact with these negative acting factors.
Dr. George Miller (Yale Medical School, New Haven, Connecticut) described Epstein-Barr viral replication activator proteins. Dr. Miller described the characterization of defective EBV virions present in one cellular subclone of the p3HR-I cell line which are capable of activating expression of latent EBV in lymphoid cells. He discussed the identification of the gene responsible for this activation, which he found in a DNA fragment (BamHI-WZhet), with 2.8 kpb. He discussed characterization of this DNA showing that there was an open reading frame encoding a protein within this fragment. His experiments are now directed at identifying the target for this gene product and at the regulation of the promoter directing expression of this factor which he calls ZEBRA.
Dr. Hiroyoshi Ariga (Tokyo University, Tokyo) described c-myc protein and SV40 T-antigen as DNA replication proteins. Dr. Ariga demonstrated in vitro replication of SV40 DNA with SV40 T-antigen. Using the same general approach and protocol, he then isolated mouse DNA sequences which are capable of replicating both in vitro and in vivo. Its replication is dependent on the presence of the c-myc gene product, as shown by inhibition with anti-myc antibodies. Using a McKay assay, he demonstrated that c-myc is a DNA binding protein. He used this binding to identify human DNA sequences which would bind to c-myc He is presently testing them for ARS activity in HL60 cells.
Dr. Bernard Roizman (University of Chicago, Chicago, Illinois) discussed transactivating factors in the regulation of herpesvirus gene expression. HSV-I has a complex pattern of gene expression with a series of gene products which both act positively and negatively to regulate viral gene expression. One of these factors, alpha-TIF, is packaged in a virion and induces expression of the alpha genes. A consensus sequence is required for TIF activation Alpha-TIF, however, is not a direct DNA-binding protein; it works indirectly through host proteins. ICP4 is a virally-encoded protein which is a DNA binding protein. It is ADP ribosylated and can regulate viral genes both positively and negatively. Dr. Roizman showed that the processed forms of ICP4 vary with their affinity for different cis-acting sequences. It is possible that these various processed forms have the alternative negative- and positive-acting functions. Alpha-TIF works by modifying or binding to host proteins. Chromatography has permitted the identification of at least two proteins, H-1 and H-2, with which alpha-TIF interacts. The SV40 enhancer competes for the H-1 factor and the metallothionine promoter competes for the H-2 protein.
Dr. Howley noted in his closing remarks the high quality of the scientific presentations and the excellent progress that has been made in transacting factors. The informal setting of the meeting was quite important in accomplishing a significant exchange of information. There was unanimous support among the participants that interactions between the United States and Japanese scientists should continue to be encouraged and that future joint seminars in related areas be promoted.

(3) Seminar on “Risk assessment of Environmental Carcinogenic Factors”
This seminar was held on March 12 and 13, 1987 at the Sheraton Waikiki in Honolulu, Hawaii. The organizers were Dr. Shigeaki Sato, National Cancer Center Research Institute, Tokyo; Dr. Yuzo Hayashi, National Institute of Hygienic Sciences, Tokyo; and Dr. Richard H. Adamson, National Cancer Institute, Bethesda, Maryland. There were eight speakers from Japan and five from the United States. The purpose of the seminar was to exchange recent findings and views on risk assessment of environmental carcinogenic tactors and to discuss uncertainties and scientific opportunities in this area.
Dr. Richard Adamson (National Cancer Institute, Bethesda, Maryland) presented an introduction and overview of risk assessment of environmental carcinogenic factors. Assessment of human cancer risk associated with some specified chemical exposure is a complicated scientific endeavor that requires careful review of all pertinent information by appropriately trained individuals. This process relies heavily on information derived from epidemiological, clinical, and long-term animal studies. Short-term test results and information on structure-activity relationships, comparative metabolism, pharmacokinetics and mechanisms of action also contribute, in varying degrees, to the assessment.
Four steps are typically involved in carcinogenic risk assessment. The first, which is often referred to as hazard identification, entails a qualitative evaluation of both the data bearing on an agent’s ability to product carcinogenic effects and the relevance of this information to humans. The second, exposure assessment, is concerned with the number of individuals who are likely to be exposed and with the types, magnitudes, and durations of the anticipated exposures. The third step, dose-response assessment, uses the information on carcinogenicity from the hazard identification phase together with mathematical modeling techniques to estimate the magnitude or an upper bound on the magnitude of the carcinogenic effect at any given dose level. Finally, the information from the first three steps is combined to characterize the carcinogenic risk associated with the expected human exposure to the compound of interest.
He identified five uncertainties in risk assessment that require further clarification: (1) description of the overall strengths and limitations of the data bearing on hazard identification; (2) characterization of the uncertainty associated with estimated exposure levels; for example, through the use of an upper bound on exposure, a range of probable exposure values, or upper and lower confidence limits on a specific exposure estimate; (3) estimation of the statistical uncertainty associated with the given risk projection; for example, through upper and lower confidence limits; (4) characterization of the variability introduced by selecting a particular low-dose extrapolation procedure; for example, by considering an envelope of risk estimates generated from a variety of plausible models; and (5) evaluation of the biologic variability re1ated to the use of a particular test organism and its extrapolation to humans.
A number of assumptions used in risk assessment that require further validation were also discussed.
A number of research needs for improving risk assessment in the areas of general toxicology, methodology in risk estimation and in epidemiology were identified.
Discussion of research needs for methodology in the estimates of risk included validation of assumptions used in risk assessment approaches and modeling (e.g., assumptions about interspecies differences, average dose, and low-dose extrapolation), and development of exposure assessment and dose-response assessment aimed at dealing with the prediction of the risk from human exposure to multiple chemicals via multiple media and routes, at varying levels, for less than a lifetime.
Research needs for improving risk assessment in epidemiology included conducting epidemiologic studies aimed at testing the accuracy of risk projections derived from animal data and modeling; conducting epidemiologic studies that incorporate biochemical and molecular probes to clarify exposures, precursor states, and mechanisms of action; conducting epidemiologic studies that evaluate interactions between toxic chemicals, life-style practices, and host susceptibility.
Dr. Dorothy Canter (National Toxicology Program, Bethesda, Maryland) discussed the hazard identification of chemical carcinogens by the National Toxicology Program (NTP). The NTP toxicology and carcinogenesis studies currently utilize Fischer 344/N rats and B6C3F₁ mice of both sexes, with 60 animals/group. The chemical nomination and selection process was outlined, and the various phases of the process were detailed, including the pre-study phase, the in-life study phase and the process of data analysis and interpretation. The five categories or levels of evidence of carcinogenicity used by the NTP include two categories for positive results (clear evidence and some evidence), one category for uncertain findings (equivocal evidence), one category for no observable effects (no evidence) and one category for experiments, that because of major flaws, cannot be evaluated for carcinogenesis (inadequate study). The categories do not refer to or imply either potency or mechanism.
Dr. Angelo Turturro (National Center for Toxicological Research, Jefferson, Arkansas), substituting for Dr. Ronald Hart, discussed contributions of mechanism of action studies to risk assessment. The current mechanistic studies relevant to risk assessment were divided into four areas: pharmacokinetics, DNA adducts, oncogenes and promotion.
A number of models for risk ASSESSMENT, proposed recently, incorporate pharmacokinetics. One of the most interesting has been developed by Dr. Mel Anderson. It assumes that the basic carcinogenic insult derives from the generation of radicals by the activating systems in the body. Dr. Anderson has had some success in modeling differences in species using a series of parameters such as partition coefficients. The physiologically based pharmacokinetic models, such as one developed by Dr. Perry Gehring, attempt to incorporate physiological phenomena, such as breathing rate, etc., into the model. These models are changing risk assessment. Incorporating such information when it is available is becoming standard practice.
One direct measure of agent interaction with the DNA is the induction of DNA lesions &, DNA adducts. The role of DNA lesion in toxicity is not a simple one. A few DNA lesions can induce a mutagenic change; many can kill a cell, making it unavailable to mutagenize. The biological role of DNA lesions in carcinogenesis is unclear. It is not a simple task to even correlate the presence of lesions with biological effect, thus demonstration that a lesion is responsible for a carcinogenic effect has been accomplished only for one agent, ultraviolet light. In the case of adducts, some can result in point mutations after DNA replication. Other observed consequences of damage, such as sister chromatid exchange indicate that gross chromosomal alterations may occur as a result of DNA adducts. Some adducts, because of their rapid disappearance, appear to be of no significance for carcinogenesis.
As indices of exposure, DNA adducts have the advantage of estimating the effects of absorption, pharmacokinetics, and metabolism in the tissue measured. An index of special value is the use of bladder cells sloughed off in the urine as a measure of exposure to bladder epithelial cells. However, because of availability of tissue, surrogates are frequently used. For instance, one is able to obtain few biopsies of liver to estimate exposure. Adducts in long-lived lymphocytes may be adequate substitutes. It should be realized that factors such as differential metabolism and local concentration can be important when surrogates are used.
Also, DNA repair and cell death are important when either surrogates or adduct concentration in the target tissue is evaluated. .
Using DNA adducts as estimators is slowly assuming an important role in comparing rodent-to-man exposure. By the use of these markers, much of the effect of variability of exposure in concentration and duration, etc., is expected to be addressed. When rat and man are compared, such variability is an important stumbling block. However, as a quantitative method to extrapolate from short-term tests, there are some problems. For example, there have been few demonstrations that a similar adduct is formed in short-term tests. Also, there are a number of compounds, such as asbestos, which do not seem to induce any adduct, and it is anticipated that phenomena such as these may be important in inducing carcinogenesis. By judicious use of these markers, one can attempt to adjust exposure to get a better estimate of effective dose.
The best recent evidence that DNA interaction can be important in carcinogenesis is derived from work using oncogenes. The cellular parents of these sequences, the protooncogenes, appear to be gene products important in cell regulation such as growth factors. An oncogene may be a mutated form of proto-oncogene. For instance, after exposure to a mammary carcinogen, methylnitrosourea (MNU), in rats, there appears to be a change in codon 12 of a ras Oncogene. Interestingly, in DMBA-induced tumors, a different codon, codon 61, seems to be the one affected. Alternatively, the oncogene can be inappropriately expressed, as c-myc is after translocation in leukemia lung carcinoma, or it can be complex products of splicing of various genes. There is also the probable presence of transposable elements in mice, as found in Drosophila, which leads to inappropriate expression of genetic material. Thus, different oncogenes appear to involve very different aspects of the interaction of organism and agent.
Some oncogenes appear to be very tissue specific, others more general. It is presently not clear which are appropriate for each tumor, which are obligatory, or which are merely correlative. However, the measurement of oncogenes may eventually displace the use of DNA adducts as a measure of biological effect for carcinogenesis when codified since they include not only the stimulus (a DNA alteration) but also a portion of the response (the induction of some biological change in the cell).
The mechanisms of promotion for carcinogenesis have been one of the most intensive areas of investigation in the last few years. One mechanism of promotion that has acquired special interest is induced cellular replication. Cellular replication seems to be a necessary part of the process of cancer formation. Its actual role is not totally clear, but the replication seems to convert some DNA alteration into a transmissible form and to stabilize its effects. However, most of the focus has been on replication as a promoting stimulus. Since many carcinogenic agents are cytotoxic, this frequently leads to hyperplasia, termed reactive hyperplasia, after exposure, especially to high doses. Hyperplasia can be a promoter. Determinations of mechanism of action, especially for a process such as hyperplasia, can be important since it is reasonable that doses at which frank hyperplasia does not occur would not be carcinogenic. Other promotion mechanisms may be the induction of hormone elaboration, induction of irritating stones, etc., which may or may not extrapolate to lower doses. For instance, while cell death is an all-or-none response and may occur at only a high dose, perhaps the induction of a growth factor (leading to carcinogenesis) may occur at any dose.
Studies of mechanisms relevant to promotion are especially important when considering agents of low potency since such large doses can be given that physiological alterations are evident in animal tests.’ But, it is a reasonable question to ask whenever a dose substantially larger than that encountered by man is used in a model system. These types of studies are being considered more and more when evaluating chemicals.
Dr. Yuzo Hayashi (National Institute of Hygienic Sciences, Tokyo) further discussed the use of animal experiments for quantitative risk estimation of environmental carcinogens. Two types of methods have been used or proposed for quantitative risk estimation of environmental carcinogens from animal experiments: the safety factor approach to allocation of ADI, acceptable daily intake and the estimation of relative potency index such as TD50 or VSD (virtually safe dose). The safety factor method rarely has been used for evaluation of carcinogenicity because of various statistical and biological problems. For example, the observation of no treatment-related effects at a given dose level (NOEL, no observed effect level) may depend on the number of animals exposed at that particular level. Also, this approach assumes the existence of a true population threshold below which no adverse effects can occur. The ADI is determined by dividing NOEL by the safety factor.
The TD₅₀ is a useful parameter for comparison of carcinogenic potencies among various compounds but may offer little insight into the relative risk at the low-dose level because this index ignores the shape or steepness of the dose-response curve. VSD is defined as a value corresponding to the dose level which can induce tumors at extremely low rates, such as 10^-6 or 10^-8. This value can be obtained by downward extrapolation of animal dose-response data by use of proper mathematical models.
Although VSD is regarded as an appropriate parameter for low-dose risk, many problems still exist in the process of estimation. The first is the selection of dose levels in long-term animal bioassays. For example, often in dose-response studies with strong carcinogens, tumors appear at high rates in all treated groups. In such cases, the dose-response curves exhibit a hyperlineality which results in too conservative estimates of VSD. In contrast, when tumors occur at very low rates in all treated groups, it may be difficult to select proper mathematical models on the basis of goodness-of-fit criteria. It should also be considered that the method for estimation of VSD does not include factors concerning the species variation as to the effects of test chemicals.
The carcinogenic risk of a given chemical (A) can be estimated using the dose-response data of a reference carcinogen. The reference carcinogen is defined as a compound which can induce tumors in animals in a manner similar to A in terms of target sites and possible mode of action. This method may provide information on the carcinogenic risk of A to humans, in cases where the reference carcinogen (R) is properly selected according to the following criteria: 1) the carcinogenicity of R to humans is known or 2) the carcinogenicity to humans is unknown for R but known for its related carcinogen, R’, and the carcinogenicity of both compounds is comparatively evaluated in animals.
Dr. Keiji Wakabayashi (National Cancer Center Research Institute, Tokyo) discussed the work he and his colleagues have performed on the presence of various carcinogens in food. The levels of carcinogens are mostly very low, and sensitive methods are required for their detection. Various sensitive methods for detecting different carcinogens, GC-TEA for N-nitrosamines, GC-MS for polycyclic aromatic hydrocarbons, HPLC-fluorimetry for aflatoxin B₁ and nitropyrenes and radioimmunoassay for aflatoxin B₁, are now available, and their distribution and amounts have been reported.
Recently, a new method consisting of simple partial purification and sensitive detection was developed for quantification of carcinogenic heterocyclic amines in heated materials such as cooked food and cigarette smoke condensates. The partial purification includes extraction with 0.1 N HCI and blue cotton treatment, 0.1 N HCl-CH₂Cl₂ partition and SEP-PAK silica cartridge fractionation, and the recoveries of spiked heterocyclic amines were 51-72% at this partial purification step. Heterocyclic amines in partially purified materials were then analyzed by a sensitive method of HPLC on two kinds of columns with electrochemical detection and fluorometry.
By using this method, heterocyclic amines in fried beef, broiled beef, chicken and mutton and food-grade beef extract were analyzed and their levels per gm of cooked food were found to be 0.19 ng for IQ, 0.64-3.10 ng for MeIQx, 0.12-0.81 ng for 4,8-DiMeIQx, 0.12-0.21 ng for Trp-P-1, 0.15-0.25 ng for Trp-P-2, 0.21-2.50 ng for A-alpha-C and 0.19 ng for MeA-alpha-C. Furthermore, five heterocyclic amines were detected in cigarette smoke condensate, and their concentrations per cigarette were 0.26 ng for IQ, 0.32 ng for Trp-P- l, 0.23 ng for Trp-P-2, l0.5 ng for A-alpha-C and 1.60 ng for MeA-alpha-C. Thus, the levels of carcinogenic heterocyclic amines in our environment is as low as the concentration of other carcinogens. Therefore, any single carcinogen seems to be insufficient to explain the development of human cancer.
Besides these heterocyclic amines, a new mutagen (PhIP), which is high in content but low in mutagenicity, was found in cooked food. At this time, compounds in such categories should be studied extensively for estimation of risk of cooked food to human beings, because there is no quantitative relationship between mutagenic and carcinogenic potencies of the heterocyclic amines described above.
Dr. Shigeaki Sato (National Cancer Center Research Institute, Tokyo) presented methods for measurement of the exposure of humans to various environmental carcinogens by measuring them or their metabolites in the serum, bile, urine or feces. However, except for some carcinogens such as aflatoxin B₁, procedures of purification and quantification of chemicals of interest in these samples are time consuming. Another method of dosimetry is the assay of chemical carcinogens bound to macromolecules, such as nucleic acid and proteins. Because most chemical carcinogens bind to these molecules, and modification of DNA by chemicals is regarded as the first event of chemical carcinogenesis, measurement of carcinogens bound to macromolecules is helpful to estimate the biologically effective exposure. Physico-chemical methods, as well as immunological methods using antibodies, have made possible the measurement of some carcinogens bound to DNA and proteins. Mutagenic and carcinogenic heterocyclic amines isolated from cooked foods and pyrolysates of amino acids, such as 2-amino-3-methylimidazo-[4,5-f] quinoline (IQ) and 2-amino-6-methyldipyrodo[1,2-a:3’,2’-d] imidazole (Glu-P-1), were found to bind to serum albumin and hemoglobin, as well as DNA, when given to rats. Monoclonal antibodies specifically reactive with IQ bound to serum albumins were produced, which may work for the immunological assay for this carcinogen-serum albumin adduct. Monoclonal antibodies against the adduct between DNA and 1-nitropyrene, an environmental mutagen, were also produced. Carcinogen-DNA adducts are also detectable in a very high sensitivity with the ³²P-postlabeling method. DNA adduct formation with heterocyclic amines, 1,6-dinitropyrene and other probable environmental carcinogens, was detected in animal experiments even at the level of one modified nucleotide per 10⁹ to 10¹⁰ nucleotides. Application of this method to human materials has been done. Mitomycin C, widely used as a chemotherapeutic agent for cancer patients, was found to form marked DNA adducts in the liver and leukocytes in the peripheral blood when administered to hepatoma patients before operation. These adducts remained in the liver even two months after the administration of this agent. Although the degree of DNA modification by chemicals does not necessarily correlate with their carcinogenicity, the level of adduct formation could be one of the parameters of their risk estimation for human cancer development.
Dr. Mimi Yu (University of Southern California Comprehensive Cancer Center, Los Angeles, California) summarized the contributions of epidemiology to risk assessment. There are several commonly used measures of disease frequency in a population. For etiologic investigation, the most desirable measure is the incidence rate, which is the number of new cases occurring in a given population during a specified period of time divided by the number of persons at risk for the disease in that population during the same period of time. Therefore, the incidence rate of a disease is a direct estimate of the probability, or risk, of developing the disease during a specified period of time. Currently, cancer incidence rates are known only for a small percentage of the world’s population.
The mortality rate is the ratio between the number of deaths in a given population during a specified period of time and the size of the population at the midpoint of that time interval. Point prevalence rate is the proportion of the population with a given disease at a particular time point.
There are still large areas of the world where information on incidence, mortality, or prevalence of cancer are missing. In some of these instances, examining frequencies of specific cancers as a percentage of all cancers in a hospital or an autopsy series have proved to be useful in uncovering unusual cancer patterns. However, these cancer frequencies can be highly biased, depending on whether the series under study is representative of all cancer cases from the general population.
International and national comparisons of cancer rates have shown large differences between different regions of the world, contributing to the current estimate that 80-90% of human cancers are due to extrinsic factors.
A number of cancers have shown changes in risk over time, providing conclusive evidence that extrinsic factors affect these types of cancer. Some examples are the current increase in lung cancer and the decrease in stomach cancer throughout most of the world.
Comparison of migrant populations to static populations in their homeland and host countries have provided evidence of the importance of environmental factors in the etiology of cancer. Some examples are the Japanese migrants in Hawaii acquiring the cancer risk pattern of local Caucasian residents and Indians in Fiji and South Africa losing their high risk for oral cancer.
Another useful tool in examining descriptive data is the analysis of birth-cohorts; that is, comparing the risk of disease in persons born within a particular period of time with persons born before and after that period. An example is the comparison of lung cancer rates in successive birth cohorts of U.S. men and women and relating these trends to changes in the prevalence of cigarette smoking.
Correlational (or ecological) studies are studies in which the rate of disease in a population is compared with the geographical or temporal distribution of suspected risk factors. For example, it has been shown that a close positive correlation exists between per capita consumption of dietary fat and female breast cancer, per capita meat consumption and colon cancer, and average level of aflatoxin contamination in foods and primary liver cancer.
Descriptive data have their limitations. First of all, measurement errors can lead to an apparent difference in risk between groups when none exists. Such measurement errors may stem from differences between groups in their access to, or utilization of medical services, their precision of diagnosis, or their completeness of reporting. When mortality rates are compared, an apparent difference between groups may be the result of differential rates of survival.
A more serious limitation of descriptive studies is their inability to control for other relevant factors when evaluating an association between exposure and disease. This is due to the fact that descriptive data are collected on populations, rather than on individuals, and groups of factors vary together in a population. Therefore, descriptive studies are useful in generating etiologic hypotheses, but they are, for the most part, not appropriate for hypothesis testing.
Analytical studies are used to test etiological hypotheses, and they fall into three broad categories: case-control approach, the cohort approach and intervention trials. Each of these categories was discussed in detail.
The most common measure of exposure and disease association used in analytical studies is the relative risk. It is the incidence rate in the exposed group divided by the incidence rate in the unexposed group. This parameter possesses a number of very desirable properties. First of all, its magnitude reflects the likelihood that an association is a causal one. In other words, the higher the relative risk, the more likely the association between exposure and disease is a causal one. The relative risk is also a sensitive indicator of disease specificity. That is, if a causal agent increases the risk of disease A but has no effect on risk of disease B, then the relative risk of developing disease A is greater than the relative risk of developing diseases A and B combined. Finally, the odds ratio, which is directly estimable from case-control data, is a good estimate of the relative risk when the disease under study is rare, and specific cancers are rare.
Dr. Yu also summarized the general strengths and limitations of epidemiological studies in assessing carcinogenic risks of environmental agents in humans. One obvious strength is that epidemiology directly evaluates the experience of human populations and their response to various environmental exposures and host factors. In addition, epidemiology provides a direct test of validity to any model that attempts to describe the process of cancer development.
Its limitations are that first it is difficult to evaluate the modest effect of low exposure to a carcinogen. Alternative explanations such as bias, measurement errors, chance, or confounding are hard to rule out when the exposure and disease association is a modest one.
Another common problem in epidemiology is exposure assessment. Often times, exposure cannot be measured directly and thus may be measured incorrectly. Random errors will deflate the relative risk toward one, while systematic errors will lead to invalid results.
Thirdly, the long latency period between exposure and most cancers means that the effect of agents newly introduced into the environment cannot be measured.
Finally, exposures to several suspected substances are sometimes closely correlated, making it extremely difficult to disentangle the individual effects.
Dr. Yu used an example of how the descriptive approach and the analytical approach were used to assess the hypothesis that nasopharyngeal cancer in Southeast China was due to a dietary factor; namely, salted fish.
Dr. Gary Flamm (Center for Food Safety and Applied Nutrition, Food and Drug Administration, Washington, D.C.) stated that risk assessment policy in the U.S., which incorporates quantitative estimates of risk, has been evolving since the late 1960’s. The impetus for this development is viewed as (1) the greatly increased sensitivity of analytical methods for measuring chemicals at very low levels and (2) the widely held scientific belief that many carcinogens (and mutagens) lack thresholds for biological activity below which there is no risk of cancer. During the early phases of the evolution of quantitative risk assessment for carcinogens, the scientific debate centered on two major issues: (1) Can estimates of risk be made conservatively enough to eliminate the danger of underestimating the human risk which might lead to the exposure of thousands or millions of people to “excessive” levels of cancer-causing compounds, and (2) What dose-response function or mathematical model best describes the relationship between dose and effect at levels far below those that can be measured? Both issues remain subjects of debate, but major concerns by scientists and policy makers has begun to shift toward (1) the need to ensure an appropriate consistency of approach, (2) the identification and separation of scientific assumptions from policy options, and (3) ways of validating scientific assumptions used in risk estimation. Even though major questions remain concerning the best approach and, in certain instances, the appropriateness of quantitative risk assessment, it is being applied on a daily basis to regulatory problems. The three questions above have been or are being pursued with great diligence by several agencies of the U.S. government, the U.S. National Academy of Sciences, committees of the U.S. Congress, and U.S. industries and industrial groups.
Dr. Suketami Tominaga (Aichi Cancer Research Institute, Nagoya) continued the discussion on the epidemiological approach to the evaluation of environmental carcinogenic factors. He stated that the primary prevention of cancer is important in cancer control. There are two major approaches to primary prevention of cancer: an epidemiological approach and an experimental approach.
The epidemiological approach to primary prevention of cancer requires (1) the identification of avoidable risk factors of cancer by conducting epidemiological studies and (2) avoidance of high risk factors and complement of low risk (protective) factors by lifestyle modification, etc.
He discussed a practical possibility of primary prevention of cancer in Japan. He stated that there are two major ways to estimate the potential of primary prevention of cancer: an attributable risk approach and a minimal cancer risk approach.
Based on calculation of attributable risk of cancer to smoking, it was estimated that about 10 percent of cancers could be prevented in Japan if the prevalence of smokers was reduced to 30-50% of the present level (from 65% to 30% in male adults and from 15% to 5% in female adults).
Diet has been regarded as the most important risk factor in the etiology of cancer. Attributable risk of cancer to diet was estimated to be at least about 10% in Japan.
Effects of air pollution on human cancer are not clear, but it was estimated that the effects of continuous exposure to diesel engine exhaust emissions would be equivalent to the effect of smoking 0.5-3 cigarettes per day (approximately 1/10 of the usual cigarette smoking).
Based on various estimates of the minimal cancer risk (male vs. female comparison, the time trend analysis of cancer mortality in Japan, etc.), it was estimated that about l0-20% of cancers could be prevented by lifestyle modification and control of other environmental factors.
From the present preliminary results on the practical possibility of primary prevention of cancer, it was concluded that about 30010 of cancer could be prevented in Japan if strong positive measures are taken.
Dr. Suminori Akiba (Radiation Effects Research Foundation, Hiroshima) presented the risk estimation of radiation-induced cancer based on data from atomic bomb survivors. Since 1950, the Radiation Effects Research Foundation (RERF) has followed the Life Span Study cohort, a fixed population of 82,000 atomic bomb survivors and 26,000 unexposed controls. Nearly all deaths among the cohort have been verified and their causes of death obtained from vital statistics death schedules. Incident cases of malignancy have also been ascertained by checking against tumor registry files in Hiroshima and Nagasaki. Since 1965 the tentative 1965 dose (T65D) has been used almost exclusively as dose estimation for A-bomb survivors, but it is now under revision. In most of this presentation, the T65 dose system will be used to estimate the risk of radiation-induced malignant tumors.
The incidence of radiation-related leukemia started to increase from 2-3 years following exposure, reached its peak in 5-6 years and then gradually declined. The excess risk may persist for 40 years or longer in the most heavily exposed persons. The most common types of radiation-induced leukemia were acute and chronic granulocytic leukemia among adults and children, and acute lymphocytic leukemia in children. Chronic lymphocytic leukemia has not been related to radiation.
Radiation-related solid tumors began to appear after the so-called “cancer age” was attained and has continued to increase with the increasing mortality of controls as they age. Thus far, statistically significant increased risk associated with radiation exposure has been seen for the following sites: breast, lung, colon, stomach, thyroid, ovary, bladder, and esophagus. Multiple myeloma has also been increased in exposed individuals. No specific histological type has been related to radiation exposure, except that undifferentiated cell carcinoma of the lung is more strongly associated with radiation exposure than other types.
Among several risk modifiers, age at time of bomb (age ATB) has been the strongest. In general, the younger age ATB is associated with higher cancer risk. For example, female breast cancer risk among the 0-9 age ATB group is about two times higher than that among other age groups. No marked sex difference in risk has been observed so far, except for lung and thyroid cancer. Lower risk of radiation-related lung cancer for men as compared with women can be attributed to the higher prevalence of smoking to men. An interactive effect between smoking and radiation on lung cancer was also discussed, along with possible biases in cancer risk estimation.
Dr. Kanehisa Morimoto (University of Tokyo, Tokyo) presented a talk on chromosome aberrations and sister chromatid exchanges as parameters for human risk of cancer development. Chromosome alterations, which are directly visible changes in the DNA, have close associations to cancer development. Among various types of chromosomal changes, chromosome-structural aberrations and sister chromatid exchanges (SCEs) are widely used indicators of chromosomal DNA damage. The former are efficiently induced by the chemical adducts and modifications of DNA, while the latter are produced directly by DNA-strand(s) breaks. The formation mechanisms of these two types of chromosome alterations are totally different.
Human lymphocyte cultures can be used for cytogenetic monitoring of human exposure to environmental carcinogens. Human lymphocytes can be obtained easily and carry information on both exposure doses and genetic effects of the mutagenic and carcinogenic factors. Although lymphocytes are in the resting G₀ stage in the body, the cultures soon contain cells that have divided several times following stimulation of cultures with phytohemagglutinin (PHA).
In one series of experiments, a simple combination of autoradiography, to determine when a cell synthesized DNA, and sister chromatid differential staining, to determine how many times a cell divided, was developed to follow the proliferating fate of human lymphocytes in PHA-stimulated cultures, because the results of cytogenetic studies can be greatly affected by the number of cell divisions before sampling. Cells were incubated continuously with 5-bromodeoxyuridine (BRDU) and pulsE1abeled with 0.1 Ci/ml ³H-thymidine (dThD) at various times following stimulation with PHA. The cells were then harvested at 4 hour intervals up to 72 hours, and the percentages of labeled metaphases was determined separately in the first, second, or third division cells. Studies also examined the distribution of the first, second, and third division cells from 72-hour cultures exposed to BRDU for various times before fixation. The data showed that the cycling cells, whether they began cycling at earlier or later times after stimulation, had about the same generation times of 12-14 hours. This indicates that the heterogeneity of different generation metaphases seen in short-term lymphocyte cultures is due mainly to the difference in the times when cells began cell cycling in response to PHA.
Another set of experiments was performed to quantitatively compare the results of in vitro studies using PHA-stimulated human lymphocytes and those of in vivo animal studies. As model environmental carcinogens, cigarette smoke tar (CST) and diesel engine emission tar (DET) were examined. When human lymphocytes were treated with various concentrations of CST or DET, it was found that CST was a 2 to 3 times stronger inducer of SCES compared to DET. When these data were compared with the results of animal experiments for subcutaneous tumor induction conducted by Prof, K. Takemoto which demonstrate that the same CST sample is a stronger inducer of the tumors in C57BL mice than the same DET sample, it seemed that the SCE assay in human lymphocytes might quantitatively predict the tumorigenicity.
Because many cancers are caused by long-term, unhealthy lifestyles, studies investigating the correlations between SCE frequencies in the peripheral lymphocytes and lifestyle of the blood donors were performed. One hundred fifty people, aged 40 years or older who took health check-ups at a health center, were the subjects of the study. The following healthy lifestyle or “good health practices’ were examined: (1) not smoking, (2) not drinking too much alcohol, (3) performing physical exercise regularly, (4) sleeping more than 6 hours per night, (5) having nutritionally balanced meals, (6) not snacking, (7) eating breakfast every day, and (8) not having too much perceived stress. The persons were categorized into three groups: good, moderate, and poor lifestyle practices. Mean frequencies of baseline SCES in lymphocytes from men with poor lifestyles were shown to be significantly higher than those in cells from men having good lifestyles. The former cells also showed a hypersensitivity to the induction of SECs by mitomycin-C (3 x 10^-8 M, 72 hours) treatment compared to the latter cells.
Further experiments were performed to examine whether lymphocytes from persons having unhealthy lifestyles might show a higher susceptibility to the chemical’s inhibitory effect of repair of radiation-induced chromosome breakage, since the environment contains various possible repair inhibitors. Lymphocytes were exposed to beta-rays emitted from tritiated water or to gamma-rays from ¹³⁷Cs along with cytosine arabinoside (ara-C) treatment for 4 h immediately before PHA-stimulation of cultures. Dicentric and ring chromosomes were examined in 52-h cultures, and the enhancing ratios by ara-C of the aberration frequency were calculated. The results showed that unhealthy lifestyles could make the cells more sensitive to ara-C’s enhancement of radiation-induced chromosome aberrations.
Dr. Shoji Fukushima (Nagoya City University, Nagoya) discussed the risk estimation from tumor promoters. Recently, an in vivo rapid assay system for liver carcinogens and promoters was designed utilizing the two-step concept of liver carcinogenesis. Assay of the carcinogenic potential of 102 different compounds was carried out using this rapid bioassay system. Rats were initially given a single dose (200 mg/kg) of diethylnitrosoamine (DEN) intraperitoneally and, starting 2 weeks later, were treated with test compounds for 6 weeks and then sacrificed, all rats being subjected to two-thirds partial hepatectomy at week 3. Promoting potential was judged by comparing the number and area of induced glutathione S-transferase placental formpositive (GST-P⁺) foci in the liver with those of the corresponding control group given DEN alone. Of the liver carcinogens, 10 out of 11 (90.9%) mutagenic, and 11 out of 13 (84.6%) nonmutagenic compounds gave positive results. Carcinogens other than the hepatocarcinogens gave less positive results (mean 15.4%). Not one of the compounds reported as noncarcinogenic was positive, suggesting that the assay system does not suffer from the disadvantage of false-positive results. This system also provided information concerning the inhibitory potential of compounds.
Thus, the dose-dependent effects of three antioxidants on modification of liver carcinogenesis were investigated using this rapid bioassay system. The results indicated that butylated hydroxyanisole (BHA) showed a dose-dependent decrease of GST-P⁺ foci, indicating the inhibitory potential of BHA, whereas butylated hydroxytoluene (BHT) and ethoxyquin did not.
Other studies were made on the dose-response of promoters in induction of preneoplastic or neoplastic lesions of the urinary bladder in rats pretreated with N-butyl-N-(4-hydroxybutyl)nitrosamine. Results demonstrated that NaHC0₃, with or without L-ascorbic acid, sodium saccharin, BHA and BHT, but not ethoxyquin, increased the developments of the lesions of the urinary bladder and the biologic response showed dose-response effects.
Thus, the data from these systems provide important information for risk estimation of tumor promotion.
Dr. Curtis C. Travis (Oak Ridge National Laboratory, Oak Ridge, Tennessee) spoke about biologically based modeling in risk assessment. While prediction of human risk due to chemical exposure should be based on human data, adequate human data are rarely available. Consequently, experimental animal data are the common alternatives upon which analyses of risk are based. Because of gaps in our current scientific understanding of the cancer-causing process, risk assessment requires the use of a series of judgmental decisions concerning unresolved issues. The major assumptions arise from the necessity to extrapolate experimental results (1) across species from rats or mice to humans; (2) from the high-dose regions to which animals are exposed in the laboratory, to the low-dose regions to which humans are exposed in the environment; and (3) across routes of administration. Development of tools and methodologies which can help evaluate the scientific bases of these assumptions will reduce the uncertainties in the risk assessment process. Pharmacokinetic and pharmacodynamic models represent two such tools.
The cancer process can be separated into a pharmacokinetic phase and a pharmacodynamic phase. The former relates applied dose to effective dose at target tissue, while the latter relates effective dose with biological effect. A recent development in the cancer risk area is the advent of biologically based pharmacokinetic and pharmacodynamic models. Relying on actual physiological parameters, such as body weight, cardiac output, breathing rates, tissue volumes, etc., to describe the metabolic process, pharmacokinetic models can predict chemical transport and metabolism across routes of administration, across species, and through temporal variations in exposure. A chief advantage of the biologically based model is that by simply changing the physiological parameters, the same model can be used to describe the dynamics of chemical transport and metabolism in mice, rats, and humans. These models allow for estimation of metabolized dose to target tissue and provide insight into questions of interspecies extrapolation. Such an ability will vastly improve the risk assessment process.
Biologically based pharmacodynamic models relate fundamental cellular processes to the epidemiology of cancer in animal and human populations. The models are based on the assumption that cancer is a two-stage process, and incorporate data on genetic mutation frequencies and cell turnover dynamics. When calibrated with biological data, these models will provide data on background probabilities of initiation and transformation per cell division, and provide quantitative estimates of age-specific tumor rates. They also provide a tool to investigate the relationship between hepatocellular proliferative lesions (foci) and hepatocellular carcinomas. They provide insights as to why B6C3F₁ mice have a higher background liver cancer incidence than do Fischer 344/N rats. Such models will increase our understanding of the actions of a pure promoter and may provide insight into the question of the existence of a threshold for pure promoters.
Biologically based models provide a scientific foundation for evaluating the various assumptions used in risk assessment and have a good possibility of producing more realistic estimates of risk associated with environmental carcinogens.
Dr. Chihiro Hirotsu (Faculty of Engineering, University of Tokyo, Tokyo) discussed a biostatistical method for evaluation of environmental carcinogens. Recently, various statistical methods have been developed for detecting and assessing the environmental carcinogenic risk factors. Dr. Hirotsu’s report added another method based on the cumulative chi-squared statistic, which can be used for various purposes including: (1) testing for the equality of kappa binomial populations against a simple ordered alternative and (2) testing goodness-of-fit of a statistical model, such as the dose-response curve or the proportional hazards model, assuming a systematic departure from the model.
For the first problem, there is a well-known linear trend test by Cochran and Armitage. While the linear test is most efficient if the assumed logistic model is correct or is asymptotically efficient if a monotonic, locally linear response curve can be assumed, a more robust trend test might be useful where less is known for the underlying response curve. The trend test, based on: the cumulative chi-squared statistic, assumes only the monotonic relation and should, therefore, work well for the wide range of monotonic dose responses.
For the second problem, it was pointed out that the omnibus test statistic, such as the usual goodness-of-fit chi-square, is not appropriate for testing a systematic departure from the assumed model. On the other hand, an approach to parameterize the departure by a few parameters is too restricted. The cumulative chi-squared statistic is expected to give a good compromise between the two methods. A related problem is the goodness-of-fit of a dose-response curve at extremely low-dose levels. Another concern discussed was the optimal allocation of samples for the model discrimination.
In his closing remarks, Dr. Adamson noted that the informal setting of the meeting was very useful for risk assessment information exchange between the two countries. He also stated that the quality of the scientific presentations was excellent and that the area of risk assessment continued to develop and become more precise and more widely applicable.



SEMINAR AGENDA AND PARTICIPANTS

(1) SEMINAR ON MARINE NATURAL PRODUCTS AND CANCER CHEMOPREVENTION
Honolulu, Hawaii, February 11-13, 1987

AGENDA

PARTICIPANTS

UNITED STATES

Dr. John Bertram
Cancer Research Center of Hawaii
University of Hawaii
1236 Lauhala Street
Honolulu, HI 96813

Dr. Luigi M. De Luca
Laboratory of Cellular Carcinogenesis and Tumor Promotion
National Cancer Institute
Bethesda, MD 20892

Dr. William Fenical
Scripps Institution of Oceanography
University of California
Santa Barbara, CA 93106

Dr. Lawrence Levine
Department of Biochemistry
Brandeis University
Waltham. MA 02254

Dr. Ron L. Merriman
Lilly Research Laboratories
307 East McCarty Street
Indianapolis. IN 46285

Dr. Richard E. Moore
Department of Chemistry
University of Hawaii
2545 The Mall
Honolulu, HI 96822

Dr. Marsha R. Rosner
Department of Applied Biological Sciences, E18-564
Massachusetts Institute of Technology
Cambridge, MA 02139

Dr. Ajit K. Verma
Department of Human Oncology
University of Wisconsin Center for Health Sciences
Madison, WI 53792

JAPAN

Dr. Hirota Fujiki
Cancer Prevention Division
National Cancer Center Research Institute
5-1-1 Tsukiji
Chuo-ku, Tokyo 104

Dr. Mitsuru Hirota
Cancer Prevention Division
National Cancer Center Research Institute
5-1-1 Tsukiji
Chuo-ku, Tokyo 104

Dr. Masaru Kobayashi
Faculty of Pharmaceutical Sciences
Hokkaido University
Kita-12, Sapporo 060

Dr. Yasutoshi Muto
Medical Faculty
Gifu University
40 Tsukasa-cho, Gifu-shi 500

Dr. Masami Suganuma
Cancer Prevention Division
National Cancer Center Research Institute
5-1-1 Tsukiji
Chuo-ku, Tokyo 104

Dr. Takashi Sugimura
National Cancer Center
5-1-1 Tsukiji
Chuo-ku, Tokyo 104

Dr. Takeshi Yasumoto
Department of Agriculture
Tokoku University
1 - 1 Amamiya-cho
Tsutsumidori, Sendai 980



(2) TRANSACTING FACTORS IN VIRAL CARCINOGENESIS
Kyoto, Japan, March 11-13, 1987

AGENDA

PARTICIPANTS

UNITED STATES

Dr. Arnold Bek
Department of Microbiology
University of California
Life Science Building
Los Angeles, CA 90024

Dr. Peter Howley
Laboratory of Tumor Virus Biology
National Cancer Institute
Building 41, Room D201
Bethesda, MD 20892

Dr. George Miller
Pediatric Infectious Diseases
Yale Medical School Institute
LSOG Building, Room 420
New Haven, CT 06610

Dr. Joe Nevins
Laboratory of Molecular Cell Biology
Rockefeller University
1230 York Avenue
New York, NY 10021

Dr. Bernard Roizman
Department of Molecular Genetics and Cell Biology
University of Chicago
Chicago, IL 60637

Dr. Flossie Wong-Staal
Laboratory of Tumor Cell Biology
National Cancer Institute
Building 37, Room 6B04
Bethesda, MD 20892

JAPAN

Dr. Hiroyoshi Ariga
Institute for Medical Science
University of Tokyo
Minato-ku, Tokyo 108

Dr. Kei Fuginaga
Cancer Research Institute
Sapporo Medical College
Sapporo 060

Dr. Akira Hakura
Department of Tumor Viruses
Institute for Microbial Diseases
Osaka University
Suita-shi, Osaka 660

Dr. Yogi Ikawa
The Institute of Physical and Chemical Research
(RIKEN)
Tsukuba, Ibaragi 305

Dr. Shunsuke Ishii
The Institute of Physical and Chemical Research
(RIKEN)
Tsukuba, Ibaragi 305

Dr. Yoshiaki Ito
Institute for Virus Research
Kyoto University
Sakyo-ku, Kyoto 606

Dr. Motoharu Seiki
Cancer Institute
Kami-Ikebukuro, Tohima-ku
Tokyo 1 70

Dr. Kunitada Shimotohno
National Cancer Center Research Institute
Tsukiji
Chuo-ku, Tokyo 104

Dr. Masaaki Terada
National Cancer Center Research Institute
Tsukiji
Chuo-ku, Tokyo 104

Dr. Mitsuaki Yoshida
Cancer Institute
Kami-Ikebukuro
Toshima-ku, Tokyo 170



(3) RISK ASSESSMENT OF ENVIRONMENTAL CARCINOGENIC FACTORS
Honolulu, Hawaii, March 12-13, 1987

PROGRAM