(1) Seminar on “Molecular Mechanism of Viral Carcinogenesis”
This seminar was held on January 26-28, 1989, at the Turtle Bay Hilton, Hawaii. The organizers were Dr. Mitsuaki Yoshida, Cancer Institute, Tokyo, and Dr. Peter M. Howley, National Cancer Institute, Bethesda, Maryland. There were seven participants from Japan and six from the United States. The purpose of the seminar was to discuss and exchange information on the role of viruses in cancer with the primary goal to exchange information on the mechanisms of viral-associated carcinogenesis and to prepare future cooperative efforts between scientists of the United States and Japan.
In opening remarks by Drs. Yoshida and Howley, each stressed the exciting new insights that have recently been provided by studies in viral oncology, and noted the importance of collaborative investigations between scientists of the two countries in most effectively exploiting these new leads.
Human Retroviruses
Dr. Mitsuaki Yoshida, Cancer Institute, Tokyo, presented a brief review of the molecular biology of human T cell leukemia virus type I (HTLV-1) and discussed two trans-acting regulators, tax and rex. The tax protein is a transcriptional activator which regulates an enhancer element in the viral LTR as well as a sequence located upstream of the interleukin-2 receptor!!
!(IL-2!!!) gene. By mutagenesis of the viral enhancer, he defined an essential sequence for the trans-activation. A cellular protein that binds to this sequence was identified by gel-retardation analysis which was likely to be involved in the trans-activation. The protein was shown to differ from a NF-KB-1ike protein that is required for trans-activation of IL-2R!!!gene expression. He thus concluded that lax protein can activate or induce at least two proteins, each interacting with an enhancer in the viral LTR or IL-2R!!!gene. He also discussed the mechanism of rex function. The rex protein is a regulator of RNA processing and accumulates intron-containing gag and env mRNAs, which are otherwise completely spliced into small tax/rex mRNA He identified that two cis-acting elements, a splice donor signal and a sequence of almost all the repeat segment from the 3' LTR, were responsible for this regulation. From these observations on tax and rex, he proposed a mechanism to explain why rex is specific to viral RNA, why expression of HTLV- 1 could be transient and why HTLV-1 proviral genomes are almost totally latent in vivo. Dr. Kunitada Shimotohno, National Cancer Center, Tokyo, also discussed two regulatory proteins of HTLV-1 and -2. He presented data to show that rex can regulate gene expression even without the 3’LTR sequence. The results were in contradiction with those of Dr. Yoshida who showed requirement for the 3' LTR. The magnitude of its effect, however, was much less than those reported by Dr. Yoshida. These differences between the results of Drs. Shimotohno and Yoshida were discussed in relation to the constructions, cells and assays used by the two investigators. Dr. Shimotohno obtained tax protein in large quantity by its expression with a baculovirus vector and found that the protein was phosphorylated. He also demonstrated phosphorylation of tax protein in HTLV-1-infected cell lines. The significance of this phosphorylation is currently not known.
Hepadnaviruses
Dr. Don E. Ganem, University of California, San Francisco, California, reviewed molecular aspects of hepatitis B virus replication comparing it with retrovirus replication. During HBV replication, three mRNA species are identified coding for its core protein/polymerase, surface envelope protein, and X protein, respectively. These three transcriptions are each initiated by an internal promoter. The largest 3.5 kb mRNA contains two open reading frames for core and pol protein in different frames partially overlapping each other. To analyze the mechanism of pol gene expression, he used the duck HBV DNA and introduced a termination codon at various positions surrounding the first AUG codon in the pol open reading frame. Termination codons upstream of the first AUG did not abolish pol protein expression in cultured cells, but termination codons downstream of the AUG completely blocked pol gene expression and viral replication in ducks. From more careful and detailed experiments, he concluded that the first AUG codon in the pol open reading frame is important for internal initiation of translation, thus ruling out a frameshift mechanism. He commented that the possibility of a minor spliced RNA species could not be excluded, although it was extremely unlikely. Dr. Ganem also discussed the function of the X gene of HBV, which is now known to be a transactivator of transcriptional promoters. He showed that X gene expression activated gene expressions of a wide variety of promoters in various cells. An HBV mutant with a defect in the X gene expresses the RNA very poorly, but this defect was complemented by coexpression of the X gene, indicating the importance of trans-activation of X protein in HBV replication.
Dr. Katsuro Koike, Cancer Institute, Tokyo, presented data suggesting that X protein function could be involved in cellular transformation. Several cell clones of NIH 3T3 transfected with X gene expression plasmid were classified into three groups of X gene expression: high, intermediate and low expressors. Cell clones expressing a high level of X mRNA showed transformed characteristics such as higher saturation density of cells, partially transformed morphologies and tumorigenic properties in nude mice. He also demonstrated that the c-myc gene promoter could be activated severalfold by the X protein. Interestingly, he also showed that a defect in the X gene greatly reduced expression of HBV RNA when assayed in HepG2 cells, but had no apparent effect in Huh-7 cells. He proposed the existence of some cellular factors(s) that mimics the X protein function in Huh7 cells, but not in HepG2 cells.
Dr. Takahiro Ochiya, Osaka University, Osaka, introduced his unique in vitro system for infection and replication of HBV. Human HBV can replicate only in humans and chimpanzees, and no in vitro system has been available for viral infection. He demonstrated that HBV can infect primary cultured cells obtained from human embryo liver, replicate and produce infectious particles. The original infectious virus was produced by cells transfected with three tandem repeats of HBV DNA. The establishment of infection and spreading of the virus were demonstrated by HBs and HBe antigens in the culture fluid, HBV specific RNA, and unintegrated HBV-DNA in recipient cells. However, the infected primary cells have not shown notable changes so far. He detected integrated HBV DNA within a few days after infection. The newly integrated viral genome contained almost the entire viral sequence and apparently integrated through two direct repeating sequences in the viral genome. The possible mechanism of the integration of HBV DNA was discussed and the usefulness of this in vitro system was suggested.
Viral and Cellular Oncogenes
Dr. Yoshiaki Ito, Institute of Viral Research, Kyoto University, Kyoto, discussed transcriptional factors involved in the regulation of polyoma virus enhancer. The polyoma virus enhancer is stimulated by treatment of cells with TPA or expression of the Ha-ras oncogene. This enhancer is inactive in undifferentiated cells such as F9 or PCC4 cells, but become active when these cells differentiate. Three complexes between polyoma enhancer and cellular proteins were detected by gel retardation assay. When the cells had been treated with an inhibitor of protein kinase C or the extract was treated with acid phosphatase, the fast migrating complex became the major component.
Purification of these cellular proteins on a DNA affinity column gave rise to two clusters (!!
!and!!
!) of several bands on gel electrophoresis. The components of!!!and!!!of bound to the enhancer at the same site, but gave different bands on gel retardation. A mixture of!!!and!!!components gave a new retardation band, suggesting the formation of hetero-dimers of these proteins. Based on these observations, he proposed a model in which a dimerization of enhancer proteins induces bending or folding of DNA molecules. Dr. Arnold J. Levine, Princeton University, Princeton, New Jersey, presented a review on Epstein-Barr virus discussing the putative oncogenes EBNA-1, -2, and -3 and LMP-1 and -2. In transgenic mice carrying a trans-activator gene EBNA-2 linked to the polyoma enhancer/promoter, no viral gene is expressed in vivo. However, when the spleen cells were taken out and activated with lipopolysaccharide, cell division continued much longer than that of cells without this transgene, suggesting that EBNA-2 may be associated with cell growth and immortalization. Transgenic mice with the LMP-1 gene directed by polyoma enhancer/promoter developed hyperkeratosis and hyperplasia of the epidermis. These mice expressed the LMP-1 gene in the epidermis, but not the dermis, and also showed abnormal expression of keratin in abnormal locations. The LMP-1 protein has six membrane-spanning domains and a large intracellular domain, like channel proteins bringing ions into cells. Thus, Dr. Levine speculated that disorganization of Ca++ ion transport may influence transcription of the keratin gene and result in disorganization of the epidermis. He also reported briefly results on transgenic mice carrying the Em-EBNA-1 gene that produces preB-cell lymphoma.
Dr. Kei Fujinaga, Sapporo Medical School, Sapporo, presented his studies on the E1A gene of adenovirus. Three domains of the E1A gene can be distinguished: domain 1 for the induction of DNA synthesis, domain 2 for transrepression, and domain 3 for trans-activation. He inserted or deleted 3 bases at various positions in the E1A coding sequence and assayed transforming activity in rat cells and assayed trans-activating activity of the adenovirus E3 promoter. Results suggested that domain 1 induced cellular DNA synthesis and was correlated with transformation, whereas domains 2 and 3 were associated with trans-repression and trans-activation and were not associated with transformation. He suggested that the transactivating capacity of E1A is not, therefore, associated with the transforming activity of adenovirus.
Dr. Edward Harlow, Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, discussed formation of a complex of the retinoblastoma (Rb) gene product with transforming proteins of DNA tumor viruses. He prepared a series of monoclonal antibodies to E1A protein recognizing different epitopes and co-immunoprecipitated proteins associated with the E1A protein from adenovirus transformed cells. One of the proteins which coprecipitated with these antibodies was identified as the Rb protein (p105-RB). Formation of a complex between E1A and Rb protein was also demonstrated using Rb protein translated in in vitro. The T antigens of SV40 and polyoma virus, E7 proteins of the human genital papillomaviruses, were also demonstrated to form complexes with the Rb protein. The Rb gene is a tumor suppressor gene involved in retinoblastoma; thus, it was proposed that the binding of T antigens inhibits the function of Rb protein and that this inhibition could be a general mechanism of oncogenesis by DNA tumor viruses. Supporting this idea, the Rb protein was shown to be bound with a nonphosphorylated form of SV40 T antigen during the G 1 to S phases of the cell cycle. He finally commented that the Rb gene is probably not the only gene responsible for transformation and showed evidence of a second protein (p107) which is also associated with AdE1A and SV40 T antigen.
Dr. David M. Livingston, Dana Farber Cancer Institute, Harvard Medical School, Boston, discussed the interaction of the Rb protein with SV40 T antigen and proposed the significance of this interaction in cell proliferation. The SV40 T antigen binds to the Rb protein and can be coprecipitated with the Rb protein. But the T antigens of several transformation-defective mutants do not bind to Rb protein. Thus, this binding correlated with and was suggested to be involved in cellular transformation. The Rb protein has several phosphorylated forms and gives a series of bands on gel electrophoresis. Dr. Livingston found that T antigen can bind to under- or unphosphorylated forms of the Rb protein, but not to the heavily phosphorylated forms. Furthermore, he found that almost all the unphosphorylated Rb protein did bind to T antigen in the presence of tenfold excess of T antigen and that phosphorylated Rb protein remained unbound. Based on these observations, he proposed a model in which unphosphorylated Rb protein is the active form in growth suppression and phosphorylation of Rb protein switches off the growth arrest. In other words, the function of Rb protein is controlled by phosphorylation and T antigen binding. Cells starved of serum showed the unphosphorylated form of Rb, and serum stimulation shifted most of the Rb protein into heavily phosphorylated forms. In synchronized cells, unphosphorylated Rb protein appeared transiently in the G1 to S phases of the cell cycle. The Rb protein was suggested to be cell-cycle dependent growth-suppressive factor inhibiting process from the G1 to S phase. Some amino acid homologies between SV40 T antigen and a cell-cycle dependent yeast factor, CDC25, were also discussed.
Papillomaviruses
Dr. Douglas Hanahan, Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, presented results on transgenic mice carrying the bovine papillomavirus (BPV-1) genome. Transgenic mice with a partial tandem repeat of the BPV genome showed abnormal skin, histologically characterized as fibromatosis, and also later developed fibrosarcomas. A11 cells in regions showing pathologic changes had high copy numbers of extrachromosomal viral DNA and expressed viral transcripts. In contrast, normal tissues showed neither extrachromosomal DNA nor viral transcripts, although all cells contained 4-5 copies of integrated viral DNA. No transcripts were detected during latency for 6-8 months. The abundance of viral transcripts was parallel with progression of disease from normal tissue to fibromatosis. Development of fibrosarcomas did not involve additional changes in viral gene expression. Instead, cellular events were associated with tumor progression. Early fibromatosis tissue usually, but not always, showed normal diploid or some aneuploidy, but there were no consistent chromosomal abnormalities. However, in aggressive fibrosarcoma tissues, trisomy of chromosome #8 and loss of #14 were frequently observed. The loss of chromosome #14 was particularly interesting because #14 is the human homologue of #13 and contains the Rb gene. Frequent gain of #8 was explained by a possible negative effect on a function of chromosome #14.
Dr. Masaaki Terada, National Cancer Center, Tokyo, discussed the roles of the E6 and E7 genes of HPV-16 and -18 in malignant transformation. He used two strategies in studies: characterization of cDNA clones of viral mRNA expressed in tumor cells and analysis of transforming genes detected in tumor cell DNA by transfection assay. Two cDNA clones isolated from HeLa cells contained the E6/E7 region of HPV-18 and human cellular sequences at their 3' end. These mRNAS retained three open reading frames, the original E6 and E7 plus spliced E6. Only E7 induced anchorage-independent growth of NIH 3T3 cells. Dr. Terada also showed that primary mouse keratinocytes could be immortalized by transfection of E7 cDNA. He derived NIH 3T3 transformants by transfection of DNA from human cervical cancer tissues and isolated cDNA clones from transformed NIH 3T3 cells. Most of these cDNA clones contained complete E7, but spliced or truncated E6 open reading frames. He suggested that the constant inactivation of E6 in these clones might be significant at some stages of tumor development. Expression of these cDNA clones could induce growth of NIH 3T3 cells in soft agar, but not tumorigenicity of these cells. From these observations, he concluded that E7 alone is sufficient to induce anchorage-independent growth of NIH 3T3 cells and that E6 might be required for tumorigenic conversion. He commented, however, that the effects of these genes might depend on the specific cellular test system. Dr. Terada also analyzed DNA abnormality in cervical cancer by restriction fragment length polymorphism analysis and found loss of heterozygosity at loci on chromosome 3p in all the cases of cervical cancers, regardless of presence or absence of HPV-16 or HPV-18 DNA. Involvement of a possible tumor suppressor gene on chromosome 3p was suggested in the development of cervical cancer as an additional step.
Dr. Akira Hakura, Institute for Microbial Diseases, Osaka University, Osaka, discussed the contributions of the E6 and E7 genes of HPV-16 to immortalization, transformation and tumorigenic conversion. The expressions of E6 and E7 convert NIH 3T3 cells to a transformed and tumorigenic state, but E7 expression does not induce tumorigenicity, although it induces anchorage-independent growth. On the other hand, when rat primary fibroblasts were used for assay, expression of E7 alone induced an extended life span of cells in tissue culture; and unlike in NIH 3T3 cells, even the coexpressions of E6 and E7 did not induce any apparent phenotypes of transformation. Prolonged cultivation of these immortalized rat cells resulted in spontaneous transformation. He termed this phenomenon "progression of transformation." No change in the copy number, expression or function of integrated viral DNA were detected before and after "progression." He suggested that cellular changes were due to severalfold increases of c-Ki-ras expression.
In the final talk, Dr. Peter M. Howley, National Cancer Institute, Bethesda, discussed the functions of E2 and E7 of BPV and HPV-16. The N-terminal and C-terminal regions of the E2 gene are fairly well conserved in all of the papillomaviruses that have been sequenced. The E2 gene of BPV expresses three proteins of 45, 31, and 28 Kda. The 45 Kda protein is expressed from the full E2 open reading frame (ORF), a transcriptional trans-activator containing both domains. The smaller 31 and 28 Kda proteins are trans-repressors and both contain only the C-terminal conserved domain (domain C). In vitro translation of RNA representative of different fragments of the E2 ORF and analysis of DNA binding capacity of their products showed that the domain C is essential and sufficient for DNA binding. Dimer formation of the domain C was observed, but cotranslation of the C domain was required for dimer formation, and mixing of proteins and DNA was not sufficient. This indicated that dimer formation can occur in the absence of DNA binding. The domain N was required for functional trans-activation and the essential region was defined. From these results, he concluded that E2 protein forms dimers through the C-terminal domain which also contains the specific E2 binding DNA activity. Domain N of the same E2 molecule interacts with another protein factors(s) for transcriptional trans-activation. E2 repression is dominant and mechanistically could be explained by (1) competitive binding and (2) subunit mixing by heterodimer formation. Dr. Howley also discussed the function of E7 of HPV-16. The amino terminus of E7 has remarkable homologies with adenovirus E1A in the region required for Rb protein binding. In fact, E7 protein binding to Rb protein was demonstrated, and possible involvement of the Rb gene in papillomavirus transformation was discussed. Dr. Howley also presented data showing that E6 and E7 are both required for transformation of human keratinocytes and that E7 alone is not sufficient.
In his concluding remarks, Dr. Arnold Levine, Princeton University, Princeton, New Jersey, discussed recent experimental evidence that indicates that P53 is not an oncogene as originally proposed. Instead, the wild type P53 gene is a tumor suppressor gene. Mutations, however, over a wide part of the gene result in transdominant activation of an oncogene activity. Like the retinoblastoma gene product, P53 can also form a complex with viral oncoproteins. Thus far, this has been demonstrated with SV40 large T-antigen and the adenovirus EIB protein. Dr. Levine postulated that the transforming proteins of the DNA tumor viruses are likely to prove a rich source of substrates for identifying other cellular proteins which are tumor suppressor gene products. These will be critical proteins involved in cellular proliferation and differentiation.
Drs. Howley and Yoshida closed the meeting noting that many excellent papers on different viruses with different aspects were presented. The variety made the information exchanges attractive and efficient. The rapid progress in understanding molecular mechanisms of human viral carcinogenesis was remarkable. In the US-Japan meeting held two years ago, the viral genes associated with cellular transformation were discussed. In this meeting, the interactions of viral gene products with cellular components were the major topic on DNA tumor vnuses and also on HrILV- I. The interaction of the tumor suppressor gene products and transforming proteins of DNA tumor viruses was particularly exciting. All the participants agreed that the exciting progress in the field of viral carcinogenesis would certainly lead to better understanding of the carcinogenic process in human cancers. Exchange of information in this rapidly developing field between the United States and Japan is now and will continue to be essential for efficient promotion of this field. The participants expressed enthusiasm for continuing this series of seminars on viral carcinogenesis to encourage future cooperative efforts between scientists of the United States and Japan. Support was expressed for having the meeting again in two years and Hawaii was suggested as the possible meeting site. Drs. Yoshida and Howley agreed that it would be worthwhile to organize another seminar in a year or two. There was no additional business and the meeting adjourned at 12:00 noon.
(2) Seminar on “Multiple Primary Cancer”
This seminar was held on February 16-17, 1989, in Hawaii. The organizers were Dr. Curtis Harris, National Cancer Institute, Bethesda, Maryland, and Dr. Shaw Watanabe, National Cancer Center, Tokyo, Japan.
Chromosome deletion is now considered to be a trigger of various malignant neoplasms, not only of mesenchymal but also epithelial origin. The recent development of methodology in molecular biology made it possible to generate a new field of research, molecular epidemiology, by combining epidemiology and molecular genetics. Hereditary cancer has been one of the focuses of molecular epidemiology, but multiple primary malignant neoplasms or multiple primary cancers (MPC) also provide good research opportunities, because the occurrence of two or more cancers in a particular person is considered to be a result of interaction between genetic susceptibility and environmental carcinogenesis. In addition, according to the increased number of long-term survivors, late iatrogenic complications like MPC have become an important problem in recent years. Early detection of cancer by advanced screening methods has contributed to clarification of the natural history of cancer from etiologic standpoints.
Current status of MPC was summarized at the meeting. In addition, each organ-specific presentation from both epidemiology and molecular genetics focused on both etiological agents that cause cancers at multiple sites and genetic factors that predispose individuals to these carcinogens.
Childhood Cancer and Genetic Epidemiology
A specific chromosomal deletion and its direct involvement in carcinogenesis was first reported from study of retinoblastoma. Recent studies have revealed that the same gene, RB1, was deleted in osteosarcoma, a frequent second tumor in retinoblastoma patients. Dr. Y. Tsunematsu, National Children's Hospital, Tokyo, described cases of MPC in Japan between 1962-1988 with estimation of incidence from incidence and mortality statistics. The observed/expected (O/E) ratio was 10.8, and the estimated incidence of MPC in children with primary cancer was 105.8/100,000. MPC in retinoblastoma patients occurred in 23--9 osteosarcoma, 5 soft part sarcomas, 3 acute lymphoblastic leukemia and 5 carcinomas. Twenty-nine out of 85 MPC cases were synchronous, a high frequency of neurogenic tumors was noted, and frequent association was present between brain tumors, Wilms' tumors, leukemia, lymphoma and/or neuroblastoma. Combination of Wilms' tumors and rhabdomyosarcoma was found in five cases. This combination has not as yet been reported in the U.S.A About two-thirds of the heterochronous MPC, including another 33 heterochronous cases, seemed to be related to radiation therapy, but only 6 out of 12 secondary leukemia cases had received previous chemotherapy. The combination of MPC suggested the presence of a common susceptibility gene in tumorigenic mechanism.
Dr. AP. Feinberg, Howard Hughes Medical Institute, University of Michigan, Ann Arbor, Michigan, reported that when one performs careful chromosomal mapping of sporadic Wilms' tumors, the common region of overlap that would contain the recessive tumor gene is actually on 11p15 and not on 11p13, the site of the deletion in WAGR syndrome. In support of the existence of a tumor gene on 11p15, he has identified two Beckwith-Wiedemann's syndrome families with linkage to 11p15. These families include obligate heterozygotes for the gene, but show no phenotypic manifestations of the disease. Thus, a family could be carrying a mutation representing one hit of a common recessive tumor gene, but with low penetrance for malignancy. If there are a large number of such loci, then they could easily contribute significantly to the attributable risk for cancer in the population.
Dr. W.K. Cavenee, Ludwig Institute for Cancer Research Royal Victoria Hospital, Montreal, discussed the work of his group, particularly those of H. Scrable, in analyzing tumors associated with Beckwith-Wiedemann's syndrome. These are primarily Wilms' tumor, hepatoblastoma, adrenal carcinoma and rhabdomyosarcoma. The latter was most intensively described and is a tumor of skeletal muscle or its mesenchymal precursor cells, which exhibits differential age peaks, anatomical preferences, responses to treatment, tendencies to metastasize, and histological appearances. These tumors are grouped into embryonal and alveolar subtypes according to their morphology and degree of cellular differentiation. Histologically, embryonal rhabdomyosarcoma resembles fetal muscle of about 8 to 12 weeks of gestation, whereas alveolar rhabdomyosarcoma histologically resembles a more mature stage of fetal muscle development. The close resemblance of these tumor variants to fetal muscle of distinct developmental ages suggests the possibility that each subtype might represent clones of human muscle cells arrested in vivo at distinct developmental stages and that the extant spectrum of tumors might allow the molecular dissection of the normal differentiation process. As the initial step in this approach, he reported experiments designed to provide conclusive molecular definition of variants of rhabdomyosarcoma. DNA was extracted from peripheral blood and tumors from eleven patients with rhabdomyosarcoma, digested with restriction endonucleases, and probed with various recombinant DNA probes. The data showed that eight of the tumors specifically lost constitutional heterozygosity for chromosome 11, whereas three did not. A retrospective comparison of these findings with the histopathological reports showed a perfect correlation between such loss of alleles and the embryonal form of the tumor, and also maintenance of heterozygosity and the alveolar subtype. Thus, they identified a major molecular genetic distinction between these subtypic variants of rhabdomyosarcoma whose normal developing muscle cognates are about a week apart in gestation. The identification of a mechanism specific to the embryonal variant may provide an entry into the nature of the chromosome 11 locus and the developmental boundary defined by its loss. Further, these tumors were unique among all the soft tissue sarcomas in their expression of the MyoD1 gene. In combination, these studies may represent the initial stages of a genetic approach to differential histopathology and may also lead to an explanation for the specific types of MPC in this syndrome.
Dr. B.R. Seizinger, Massachusetts General Hospital, Boston, Massachusetts, provided evidence that the same genes that cause von Hippel-Lindau's (VHL) disease and neurofibromatosis type 2 (NF-2) are also responsible for tumorigenesis of the much more prevalent sporadic counterparts of NF-2 and VHL-associated neoplasias by the "reverse genetics" approach. Frequency of NF- 1 (von Recklinghausen's disease) is 1/4000, while that of NF-2 (bilateral acoustic neurinoma) is 1/100,000 children. The genes causing NF-1 (von Recklinghausen's NF), NF-2 and VHL map to chromosome 17q12-22 (5), 22q, and 3p21-26. Fibrosarcoma in Recklinghausen's syndrome showed loss of heterozygosity at loci on 17p but not on 17q. Putative tumor suppressor gene for VHL was present in chromosome 3p21-26; it is also the frequent deletion site in the renal cell carcinoma.
Dr. R. Takahashi, Center for Biotechnology, Baylor College of Medicine, The Woodlands, Texas, discussed the RB-1 gene on chromosome 13ql4, its mRNA and gene product. Their group found two alternative translational sites in the Rb gene; one yields pp110-116RB and p110RB, the other pp98-104RB and p98RB. Homozygous internal deletion yielded a truncated transcript with resultant truncated protein. Structural changes and abnormal expression of the Rb-1 gene was detected either in primary tumor cells from breast cancer patients or fibroblasts from hereditary breast cancer cases, implying the lack of correlation between Rb-1 gene changes and hereditary occurrence of the breast cancer as a primary event.
Lung Cancer
Dr. J. Yokota, National Cancer Center, Tokyo, reported application of molecular genetics to MPC study, including assessment of genetic changes, such as point mutation amplification, rearrangement in oncogenes, translocation and deletion in chromosomes, and integration of viral genes in the tumor DNA. Chromosomal deletion at 3p, 13q and 17p, detected by DNA probes, was present in almost all small cell lung cancers. Rb-1 gene product was also not produced. However, non-small cell lung cancer usually expresses mRNA of Rb-1 gene and produced the protein. Allelic deletions seemed to have some cancer site specificity, e.g., 5q, 17p, 18q in colon cancer; 1lp, 13q, 17p in breast cancer; 3p in uterine cancer; 14q, 1p in neuroblastoma; and 10 in glioblastoma. One case with renal cell cancer and esophageal cancer showed a different pattern of 3p deletion.
Dr. C.C. Harris, National Cancer Institute, Bethesda, Maryland, reported on the role of oncogenes and tumor suppressor genes in human lung carcinogenesis. Overexpression of c-myc, L-myc, N-myc, c-raf-1 and c-jun/AP1-1 genes was frequently found in small cell lung cancer, while mutant ras proto-oncogenes were found in non-small cell carcinomas. The transfection of ras (H-, N-, or Ki) oncogene into the human bronchial cells caused both a defect in the signal transduction pathway triggered by inducers of terminal squamous differentiation and neoplastic transformation. Fifty-four non-small cell bronchogenic carcinomas were studied for allelic deletion with 13 polymorphic markers. Loss of heterozygosity in 17p was more frequent in squamous cell carcinoma (89 percent) than adenocarcinoma (18 percent) or large cell carcinoma. Chromosomes 11p13 and 11p15.5 were also frequently deleted. Introduction of chromosome 11 with microcell-cell hybrid method into a squamous cell carcinoma cell line showed suppressor activity in these hybrids.
For lung cancer, many environmental factors are known to have an etiological role; their effects are often considered in the context of a multistage model of carcinogenesis. J. Samet, University of New Mexico, Albuquerque, New Mexico, addressed the potential role of cigarette smoking and radon in causing MPC of the lung. In the past, studies of MPC have documented the excess risk of a smoking-associated site. The MPC approach might be used to assess determinants of susceptibility to cigarette smoke. In a longitudinal approach, follow-up of persons with a first cancer and comparison of persons developing a second cancer with those not developing a second cancer, in spite of comparable exposure, may improve understanding of the determinants of susceptibility. However, the study on MPC does not appear to be an informative approach for studying radon and lung cancer.
GI Tract Cancer
Dr. S. Watanabe, National Cancer Center, Tokyo, reported an increasing trend of MPC during 1962-1988 in the National Cancer Center Hospital. Synchronous MPC contributed to the significantly increased O/E ratio in most organs. Site relationship between the two cancers revealed a down stream association in the gastrointestinal tract; e.g., oropharynx to esophagus, esophagus to stomach, stomach to colon and rectum, colon to rectum. Secondary leukemia is well known to be induced by alkylating agents. Analyses of second hematopoietic malignancies (40 malignant lymphomas and 11 leukemias) revealed an excess occurrence between 1-4 and 5-9 year intervals after the initial cancer. Original sites of lymphomas were the tonsil (21.6 percent), stomach (16.2 percent), other extranodal tissues (32.4 percent), and only 29.7 percent in the lymph node. Such a high frequency of extranodal lymphomas suggested that the initiated lymphocytes by the same carcinogen with epithelial cells were already present in these sites and that the treatment had promoter effects to progress to malignant lymphoma. Lymphoma and leukemia occurred in patients without chemotherapy or radiation who were related to Li-Fraumeni syndrome.
Dr. L.A. Cannon-Albright, University of Utah, Salt Lake City, Utah, presented a model for common inherited susceptibilities to premalignant lesions which are hypothesized to underlie cancer occurrence. The model was investigated for adenomatous polyps and colorectal cancers. Analysis of polyp and cancer occurrence in 670 individuals in 34 pedigrees strongly supported the dominant inheritance of a susceptibility to colorectal adenomas and cancers, with a gene frequency of 19 percent. According to the most likely genetic model, adenomatous polyps and colorectal cancers occur primarily in genetically susceptible persons. These results suggest that an inherited susceptibility to colonic adenomatous polyps and colorectal cancer is common and that it is responsible for the majority of colonic neoplasms observed clinically. This evidence of an inherited susceptibility to a cancer with well-recognized environmental risk factors interact in the formation and transformation of polyps. It is suggested that many of the occurrences of nonfamilial cancer are due to predisposing genes which cause precursor lesions. Because the conversion of these lesions to cancer is a relatively low probability event, the cancers themselves do not demonstrate a clear Mendelian pattern.
Dr. K. Ushio, National Cancer Center, Tokyo, reported 49 cases of familial polyposis coli in 26 families. Adenomatosis coli syndrome is regarded as a systemic disease with a high frequency of tumor throughout the body; intramandibular osteomas (86.8 percent), duodenal polyp (71.9 percent), gastric polyps (56.8 percent), bone lesions other than mandibular bone (50 percent), soft tissue tumor (27 percent), mesenteric desmoid tumor (13.5 percent) and odontoma (13.2 percent). A wide variety of tumorous lesions derived from the ecto-, meso- and endoderm characterized the syndrome. Hereditary gastrointestinal polyposis, such as Peutz-Jeghers' syndrome, juvenile polyposis, Turcot's syndrome, familial polyposis coli, Gardner's syndrome, cancer family syndrome and familial aggregative colorectal cancer are all examples of multiple- site tumorigenesis. The incidence of these tumors and associated anomalies were analyzed from the standpoint of aging.
Dr. A.P. Feinberg also reviewed the role of genetic alterations in multistep carcinogenesis. With Dr. G. Thomas of the Institute Curie, Paris, he observed loss of heterozygosity on chromosome 5 in 1/3 of colorectal cancers, while allelic deletion was more frequent on chromosomes 17 and 18, occurring in 2/3 of the cases. In colorectal cancers from familial postposis coli and Lynch syndrome patients, allelic losses were not detected in small adenomas. One-third of colorectal cancers showed point mutation at the 12th or 13th codon of c-Ki-ras gene. All colorectal carcinomas and premalignant adenomas showed dramatic changes in DNA methylation of 1/3 of examined single-copy genes, but obviously the level of DNA methylation was increased in the normal colonic mucosa of patients with Lynch syndrome. In multistep carcinogenesis, hypomethylation was considered to be the first observed genetic change associated with adenomas. Roughly coincident with the development of carcinomas, Ki-ras mutations and genetic losses on chromosomes 17 and 18 appear. Losses on chromosomes 17 and 18 may also occur late in tumor progression.
Methodology and Miscellaneous
Dr. S. Hirohashi, National Cancer Center Research Institute, Tokyo, reported multistage, multicentric development of human hepatocellular carcinoma by analysis of hepatitis B virus DNA. Southern blot analysis of DNA samples from two adenomatous hyperplasia lesions (AH) showed discrete bands that hybridized to cloned hepatitis B virus DNA. DNA from the small area of overt hepatocellular carcinoma showed the same restriction pattern as those of DNA from the surrounding AH, indicating that the two contiguous lesions were derived from an identical clone and that the overt hepatocellular carcinoma had resulted from progression of AH. Some hepatomas of multinodular type were shown to be multicentric in origin also by analysis of hepatitis B viral DNA, so such multiplicity caused severe problems to surgical resection for complete cures.
Dr. N. Yamaguchi, University of Occupational and Environmental Health, Kitakyushu, described mathematical problems in the analysis of MPC. He reviewed the association study, prevalence method and incidence method for evaluating an excess of MPC occurrence in a particular population with several examples, and proposed the necessity of a bio-indicator study for clarifying risks of MPC. The incidence method was based upon the person-year method. The prevalence method was effective for estimating occult or latent cancer.
Dr. O. Hiyama, Osaka Adult Diseases, Osaka, reported a significant excess of secondary lung cancer in patients with uterine cervical cancer from the cancer registry in Osaka. The common statistical problem from the registry data was the underestimate of MPC cases due to insufficient follow-up and overestimate of the expected number of MPC which weakened the detection power of MPC. Histology of all but one of the 14 secondary lung cancers was squamous cell carcinoma; thus, the common etiology, like human papillomavirus, should be considered. Low association of stomach cancer with uterine cancer was also noteworthy, and statistical problems in calculating the expected number of MPC were discussed.
Recent observations of leukemia occurrence after treatment with growth hormone (GH) were presented by Dr. S. Watanabe. Although the total dose of GH was not related to the occurrence of leukemia, apparent promoter effect of GH was shown in one chronic myeloid leukemia case. Almost half of the leukemia cases reported in the world used GH to treat secondary hypopituitarism that was due to the treatment of the initial cancer. Therefore, the therapeutic use of GH should be strictly limited to prevent the enhanced risk of second tumors by other growth factors, such as colony stimulating factor, fibroblast growth factor, and others. Study of common fragile sites on chromosomes to predict the predisposition to cancer in particular individuals was introduced.
Comments
The loss, deletion, or inactivation of a putative "tumor suppressor" gene or "antio-oncogene" has been recognized in several types of cancer, and it appears to be one of the most fundamental mechanisms of tumorigenesis in both sporadic and hereditary forms of cancer. A cascade in carcinogenesis involving the accumulation of both the genetic and epigenetic lesions, such as hypomethylation, mutated proto-oncogenes, and inactivation of tumor suppressor genes leading to the development of clinical cancer seems to be a reasonable hypothesis.
Allelic deletion and genetic linkage analyses offer successful strategies to identify genetic loci that are critical in the multistage process of carcinogenesis and tumor progression.
Insights into the pathogenetic mechanisms of these disease genes and their corresponding proteins should have important implications not only for individuals with these hereditary tumor syndromes, but for the much greater number of cancer patients in the general population. MPC is a good example for study on the etiology of carcinogenesis because multiple occurrence of cancer may be influenced by the combined effects of inherited predisposition and exposure to carcinogens.
(3) Seminar on “Fundamental and Clinical Aspects of Pancreatic Cancer”
This seminar was held on February 23 and 24, 1989, at the Turtle Bay Hilton Hotel, Oahu, Hawaii. The organizers were Dr. Dante G. Scarpelli, Northwestern University Medical School, Chicago, Illinois, and Dr. Yoichi Konishi, Cancer Center Nara Medical School, Nara. There were seven participants from both the United States and Japan. The purpose of the meeting was to present current work and exchange ideas on experimental pancreatic cancer, and the early diagnosis and treatment of human pancreatic carcinoma. The first day of the meeting was devoted principally to a consideration of pancreatic carcinogenesis and its modulation by chemical and hormonal factors; the second day was concerned with studies of a transgenic mouse model of acinar cell cancer, tumor markers, and various therapeutic approaches to the treatment of pancreatic cancer.
Opening remarks were delivered by Dr. Yoichi Konishi at the time when the official funeral for the late Japanese emperor, Showa Tenno, was in progress in Tokyo; a minute of silence was observed in his honored memory before the sessions began.
Dr. Daniel Longnecker, Dartmouth University Medical School, Hanover, New Hampshire, opened the session with a brief review of recent advances in pancreatic carcinogenesis in which he highlighted details of the utility of azaserine-induced acinar cell cancer in the rat, and ductal adenocarcinoma of the hamster induced by N-nitrosobis(2-oxopropyl)amine. Although research involving these models has contributed significantly to our understanding of the development and biology of pancreatic cancer, there are still many basic questions that remain to be elucidated. It is now clearly established that the induction of azaserine-induced acinar cell neoplasia is enhanced by the amount of unsaturated fat in the diet; and the growth of acinar cell nodules is enhanced by testosterone, CCK caerulein and bombesin and inhibited by estrogen and tamoxifen. Such manipulations in this rat model should permit the study of how various molecular species enhance or inhibit the neoplastic transformation of acinar cells.
A dog model of pancreatic ductal adenocarcinoma has been recently developed, which because of its size, can be employed in research on more effective diagnostic and therapeutic strategies applicable to humans. Transgenic mouse models of acinar cell cancer bearing a transforming gene such as SV40 T antigen flanked by promoter sequences of the trypsin or elastase genes promise to be valuable models for genetic and molecular studies.
Dr. Dante Scarpelli, Northwestern University Medical School, Chicago, Illinois, reported on comparative metabolism of N-nitrosamine pancreatic carcinogens in the hamster and rat. He described in vivo and in vitro studies of the metabolism of N-nitrosobis(2-oxopropyl)amine (BOP) and N-nitroso(2-hydroxypropyl)(2 oxopropyl)amine (HPOP) in hamster, a highly susceptible species, and the rat, an insensitive one. In each specie, hepatic reductive metabolism of each carcinogen was significantly greater than that of pancreas; in hamster it was increased sixfold, and in rat, fourfold. Reduction of BOP to HPOP was most rapid in hamster liver due to its exclusive content of a NADH-dependent ketone reductase, and its high ATP-mediated uptake of BOP. The most significant species difference encountered was the high capacity of hamster liver to conjugate HPOP to a sulfate ester which breaks down in an aqueous medium and is capable of alkylating DNA. In contrast, rat liver conjugates HPOP to glucuronic acid and excretes it rapidly in urine. BOP is completely removed from the circulation two hours following injection; however, its reduction products were still detectable in blood six hours after its administration. A comparison of urinary metabolites of hamsters treated with BOP and HPOP showed that while levels of products derived from activation reactions were comparable, those of HPOP and BHP and their conjugates were always greater following the administration of HPOP than of BOP. This indicates that only a fraction of BOP is reduced to HPOP and most of the metabolism of the two carcinogens, ranging from 60 to 80 percent, does not involve a common intermediate. Analysis of urine, blood, bile and pancreatic juice of hamsters treated with 40 mg/kg [1- 14C]BOP showed that conjugates of BOP and HPOP reduction products are concentrated in the urine and are not secreted in bile or accumulated in tissues, while both carcinogens are concentrated in bile and pancreatic juice in addition to their urinary excretion. Because of the noncontinuous flow of pancreatic juice, duct cells are exposed to relatively high concentrations of BOP and HPOP for extended periods of time, which may contribute to their neoplastic transformation.
Isolated acinar and islet cells activate BOP and HPOP to intermediates which are mutagenic for S. typhimurium TA 1535. At least part of this activity is localized in the S-9 and microsomal fractions. Although pure isolates of acinar and ductular cells metabolize both compounds at rates only about 100- to 1000-fold slower, respectively, than by hepatocytes, they generate active forms of carcinogen more proximal to the target cell. If the liver plays a dominant role in N-nitrosamine-induced pancreatic carcinogenesis, the carcinogenic intermediates generated must remain active following their release from the liver and arrive at the pancreas rapidly. While the question of stability of BOP and HPOP must remain open at present, physiologic studies with the dye, lissamine green, showed a liver to pancreas circulation time of only 2 seconds, so at least one of the conditions in favor of the importance of the liver in pancreatic carcinogenesis appears to exist.
Dr. Takatoshi Ishikawa, Cancer Institute, Tokyo, discussed species differences in adduct formation and O6 -methyltransferase activity by the pancreatic carcinogen 4-hydroxyaminoquinoline-1-oxide (4-HAQO). He reported on the immunohistochemical demonstration of 4-hydroxyaminoquinoline l-oxide (4HAQO)-DNA adducts in rats using a specific antibody specie and organ differences in 4-HAQO binding to DNA, and O6-methyltransferase activity in the pancreas and liver from various species including man. 4-HAQO-DNA adducts were localized in the nuclei of acinar cells in mouse pancreas, whereas staining of duct and islet cell nuclei was much weaker. The lowest limit of detection of 4-HAQO-DNA adducts by immunochemical staining was 2-5 adducts per 106 nucleotides. DNA binding levels correlated well with the reported species organ specificity of 4-HAQO tumorigenesis; high levels were present in the pancreas but not in the liver of rats, 06-Methyltransferase activity of the pancreas was lower than that of the liver in mouse and rat, and preliminary data showed that it was also low in human pancreas.
Dr. Yoichi Konishi, Nara Medical College, Nara, described the modification of pancreatic carcinogenesis by endogenous and exogenous stimuli in hamsters. The experimental protocol used in this investigation was that of classical initiation and promotion where N-nitrosobis(2-hydroxypropyl)amine (BHP) or N-nitrosobis(2-oxopropyl)amine (BOP) was used as an initiator followed by test substances in an effort to detemine whether they had promoting activity. Both secondary bile acids of deoxycholic acid (DCA) at a dose of 0.1 percent and lithocholic acid (LCA) at a level of 0.5 percent in the diet enhanced the induction of pancreatic carcinoma. In contrast, 0.5 percent clofibrate and 2 percent butylated hydroxyanisole (BHA) administered in the diet were inhibitory. A rapid model for pancreatic carcinogenesis in hamsters was described which incorporated the principle of selection by resistance to cytotoxicity demonstrated earlier for liver carcinogenesis in rats. Adenocarcinomas were induced in 84 percent of treated hamsters in 10 weeks when initiated with BOP (70 mg/kg body weight) combined with a choline-deficient diet and pancreatic regeneration (induced by D1-ethionine followed by methionine rescue). The hamsters were then again returned to treatment of 20 mg/kg BOP and basal diet and this regimen was applied 3 times.
Dr. Yuzo Hayashi, National Institute of Hygienic Sciences, Tokyo, reported on the effect of modifying factors on experimental pancreatic carcinogenesis. A single injection of 4-hydroxyaminoquinoline-1-oxide (4-HAQO) to rats results in a high incidence of eosinophilic nodules, acinar cell adenomas, and basophilic degenerative acinar cell foci which mimic an age-associated lesion of the rat pancreas. The administration of testosterone propionate or soybean trypsin inhibitor (SBTI), enhanced the occurrence of acinar cell adenoma in 4-HAQO-initiated rats, while feeding of SBTI inhibited the appearance of basophilic foci. In addition it was shown that consecutive SC injections of a synthetic long-acting ACTH preparation in 4-HAQO-initiated rats promoted the development of adrenocortical adenoma and suppressed the occurrence of peliosis adrenalis, an adrenocortical lesion known to occur in aging rats. Based on these findings, it is postulated that the administration of trophic factors may affect target cells in essentially two different ways, namely by enhancing tumor development while suppressing the occurrence of age-associated degenerative disorders.
Dr. Longnecker, Dartmouth University Medical School, Hanover, New Hampshire, discussed hormonal influences in pancreatic carcinogenesis. Epidemiologic evidence shows that the age-adjusted incidence of pancreatic cancers in humans is higher in males than in females. A similar predominance is noted in several rat models of pancreatic carcinogenesis. Using protocols of four-months duration to study the early phase of carcinogenesis, it has been shown that males develop more foci than females. The number of preneoplastic lesions is regarded as an indicator of initiation and the rate of growth of such lesions is used to detect promotion or inhibition of carcinogenesis.
Treatment with sex hormones, and several GI peptide hormones was begun after azaserine injection in order to focus on nodule growth rather than initiation. Results summarized in the Table below show that these hormones can modulate growth of lesions.