(1) Seminar on “Human Cancer Genes: Growth Factor and Tumor-Suppressor Genes”
This seminar was held on January 22-24, 1990 at the Sheraton Coconut Beach Hotel, Kauai, Hawaii. The organizers were Dr. Stuart A. Aaronson, National Cancer Institute, Bethesda, Maryland, and Dr. Masaaki Terada, National Cancer Center Research Institute, Tokyo. There were 15 participants: eight from Japan, six from the United States and one from Canada. Dr. Stuart Aaronson welcomed the attendees and introduced Dr. Takashi Sugimura, Program Coordinator of the Etiology Program Area, U.S.-Japan Cooperative Cancer Research Program.
Dr. Takashi Sugimura (National Cancer Center, Tokyo) delivered the opening address, presenting an overview of contemporary research in the area of carcinogenesis.
Dr. Masaaki Terada reported on two oncogenes of the heparin-binding growth factor family, hst-1 and hst-2. Oncogenes and transforming genes are now classified into several distinct categories, but those encoding growth factors may best illustrate the functional significance of oncogenes in malignant transformation. Only one such oncogene, sis, had been nominated in this category until hst-1, or HSTF1 (heparin-binding secretory transforming factor 1) was added to the human gene nomenclature and significant homology of its product with fibroblast growth factors (FGFS) and the int-2 protein was demonstrated. These findings were followed by the discovery of several other members of the heparin-binding growth factor family (i.e., FGF5, hst-2/FGF6 and KGF). Along with evidence that suggests receptors of these growth factors also make up a rather large subfamily of tyrosine kinases, these findings have made heparin-binding growth factor-receptors of wide interest. Their affinity to heparin was suggested by sequence similarity of the het-1 amino acid sequence deduced from a cDNA clone with that of FGF, including its postulated heparin-binding domain. This was confirmed by efficient purification of a recombinant human hst-1 protein synthesized in silkworm cells by heparin affinity chromatography. The protein was a potent mitogen for a variety of cells, including human endothelial cells. When added exogenously to the culture medium of NIH/3T3 cells, the hst-1 protein induced a morphological change simulating typical transformation. The protein also supported anchorage-independent growth of NIH/3T3 and NRK-49F cells.
In contrast to FGF, which is known to be ubiquitous in tissue distribution, expression of hst-1 appears to be strictly suppressed in adult cells. Expression of the gene could not be detected in many cancer cells, including lung, stomach, breast and liver cancers; the only exceptions were germ cell tumors in vivo and in vitro. Mouse hst-1 gene was also found to be expressed at a certain stage of mouse embryogenesis. This embryo and germ cell tumor-specific pattern of expression is reminiscent of another member of the heparin-binding growth factor family, int-2. However, based on the fact that hst-1 and int-2 are differentially transcribed in F9 mouse teratocarcinoma cells in the course of in vitro differentiation, it should be stressed that these related oncogenes may have distinct functional roles during embryogenesis and, possibly, in carcinogenesis.
The hst-1 and int-2 genes are separated less than 35 kbp and 20 kbp on human and mouse genomes, respectively. This and other groups have found that the region of chromosome 11 at band q13.3, on which hst-1 and int-2 are located, was amplified in a variety of cancer cells, especially in more than 50% of squamous cell carcinomas of the esophagus and head and neck regions. Since expression of hst-1 or int-2 was not detected despite their amplification, Dr. Terada’s group is now in the process of genomic walking in this amplification unit to look for other candidate genes which may be involved in malignant transformation and/or progression of the cells with 11q13.3 amplification.
Dr. Terada’s group has cloned a close homologue of the hst-1 gene, designated hst-2. Recently, a partial genomic sequence of the FGF6 gene was reported by Dr. Burnbaum, and based on this it is presumed that FGF6 is identical to hst-2. The cosmid clone containing hst-2/FGF6 induced foci in NIH/3T3 cells when subjected to a defined medium culture assay employed to identify FGF5 by Dr. Goldfarb. Although the hst-2/FGF6 cosmid clone did not dramatically change the morphology of NIH/3T3 cells in 5% serum, the transfected cells made a rapidly growing tumor in nude mice. Another potent tansforming gene, fum, which was identified independently from a human stomach cancer xenograft, was found to be a human-mouse fusion gene generated in the process of transfection. The mouse portion of this hybrid gene was C-terminal two-thirds of the mouse hst-2/FGF6.
Dr. Stuart A. Aaronson (National Cancer Institute, Bethesda, Maryland) reported on growth factor activated pathways in the neoplastic process. Discoveries identifying the products of certain proto-oncogenes as growth factors or their receptors have provided strong evidence that proto-oncogenes play a central role in normal growth regulation. The altered or aberrant expression of such genes appears to be fundamental to steps that cause normal cells to become malignant. V-sis represents a prototype for the class of oncogenes that encode growth factors. The human sis proto-oncogene encodes the B chain of human platelet-derived growth factor (PDGF). Its expression by cells possessing PDGF receptors can induce uncontrolled growth associated with acquisition of the malignant phenotype. Dr. Aaronson’s group has utilized PDGF as a model to study mechanisms involved in autocrine growth stimulation and the intracellular signalling pathways triggered by growth factor receptors with tyrosine kinase activity. He described the isolation and characterization of cDNAs of two PDGF receptor genes whose products independently couple with mitogenic and chemotactic signalling pathways. Their activities appear to be differentially regulated by the different binding properties of three PDGF isoforms. He also summarized evidence concerning primary substrates of growth factor receptor kinases. These include PLC, the raf proto-oncogene product, a novel PI kinase, and the GTPase activating protein (GAP). In the case of GAP, which was recently identified as a receptor kinase substrate by Dr. Aaronson’s group, there appears to be specificity in that PDGF and epidermal growth factor (EGF) receptors, but not IGF-1 or FGF receptors, induce tyrosine phosphorylation of GAP in fibroblasts or epithelial cells.
Using a model system involving epithelial cells grown in chemically defined medium, Dr. Aaronson’s group has identified two major pathways required for cell proliferation. One is triggered by IGF-1 and can be replaced by high insulin concentration. The other involves EGF or any of the known members of the FGF family. Both sets of growth factors are required in order to support cell proliferation. Oncogenes which represent activated forms of gene products intimately involved in growth factor-triggered normal cell proliferation appear to specifically replace the requirements for only one of these major pathways, the “EGF” pathway. These findings suggest that the malignant phenotype in vivo involves selection for activation of genes in this major pathway, which must normally be growth limiting.
Finally, Dr. Aaronson described efforts to identify the sites of receptor activation and functional triggering of intracellular signalling pathways in autocrine transformation by PDGF-B. These findings have implications in efforts to intervene in the uncontrolled proliferation of neoplastic cells.
Dr. Minako Nagao (National Cancer Center Research Institute, Tokyo) described his findings on the ret oncogene. The ret proto-oncogene is predicted to encode two kinds of receptor-type tyrosine kinases. In adult rat tissues, expression of proto-ret is very low, and mRNA expression was detected using poly(A) + RNA only after very long exposure of the film by northern blot analysis. Compared with adult tissues, proto-ret expression was 20-fold higher in embryonal tissues at 9-11 days of gestation. In situ hybridization revealed that the proliferating trophoblasts of placenta specifically express proto-ret during mid-term gestation.
Expression of proto-ret is particularly high in neuroblastomas. In all human neuroblastoma cell lines examined, the ret proto-oncogene was expressed, although the levels of expression were different in respective cell lines. All dissected tissues of human neuroblatomas tested to date also expressed proto-ret. There was no correlation between the levels of proto-ret expression and clinical stage, histology, N-myc amplification or N-myc expression. Expression of proto-ret was enhanced after treating neuroblastoma cells with retinoic acids, in association with morphological changes. In a neuroblastoma cell line, Nagai, at least five species of mRNA are expressed. The size differences are due to alternative splicing and alternative poly(A) addition. From these five species of mRNA, two proteins, which are different in their c-terminals, were predicted to be encoded. No amino acid substitutions were found, as compared with the amino acid sequence encoded by a cDNA from a monocytic leukemia cell line, THP-1. Two types of cDNAs encoding two different types of proteins were expressed in NIH/3T3 cells under the MuLV-LTR promoter. No focus formation was observed with either cDNA. The physiological role of proto-ret remains to be determined.
Chromosome mapping by in situ hybridization using two overlapping cosmid clones locates proto-ret on chromosome 10q11.2. The locus of 10q11.2 is associated with the locus responsible for MEN2A which is associated with both medullary thyroid carcinoma and pheochromocytoma. Proto-ret is heterozygous and could be used for linkage analysis of MEN2A. This group found a rearranged form of the ret oncogene in a papillary thyroid carcinoma cell line, TPC-1, suggesting activation of proto-ret. Involvement of protein phosphatase 2A in malignant transformation of NIH/3T3 cells by activated ret was also presented.
Dr. Tadashi Yamamoto (University of Tokyo, Tokyo) presented data on the normal function of the protein-tyrosine kinases of the src family. A number of protein-tyrosine kinases have been identified in mammalian cells and are suggested to be important in growth and/or differentiation of cells. Approximately half of the protein-tyrosine kinases are integral membrane proteins and are in many cases receptors for polypeptide growth factors. The remaining protein-tyrosine kinases, typified by the 60-kDa c-src protein, are membrane-associated proteins but lack transmembrane and extracellular domains. These proteins are, therefore, called nonreceptor-type kinases. Genes encoding proteins of this group include src, yes, fgr, fyn, lck, lyn, hck and possibly tkl, all of which are members of the src family. Comparison of amino acid sequences of the protein products of the src gene family revealed that the sequences are highly conserved within this family, except for the amino-terminal regions of about 75 residues, which are variable.
To rationalize the presence of a set of genes whose products were structurally similar to p60src, it was assumed that each member of the family is expressed and that each function in a cell-type-specific manner. This possibility was examined by northern blot hybridization of RNAS from various tissues and by immunohistochemistry. The levels of lyn and c-fgr mRNAs were higher in fetal liver than in other tissues. The fyn mRNA was detected at a high level in fetal brain, whereas the yes mRNA was expressed ubiquitously. When RNAs from cancer cell lines were analyzed, patterns of expression of c-yes, lyn, fyn and c-fgr were different.
Since fetal liver is rich in hematopoietic cells, it was assumed that expression of c-fgr and lyn might be associated with hematopoiesis. In fact, expression of lyn mRNAwas confined mostly to the spleen of adult mice. Less expression of this mRNA was seen in the thymus. Further analysis of hematopoietic cell lines revealed that lyn was expressed in B but not T lymphocytes, except for HTLV-1-producer T cells. Consistent with these observations, western blotting analysis with anti-lyn gene product antibodies showed that the lyn gene product, p56lyn was expressed in tonsils, where B lymphocytes are located. Preliminary data suggested that the lyn gene product is involved in B lymphocyte activation.
Hybridization histochemistry with a c-fgr DNA probe showed that the c-fgr gene is expressed in large granular lymphocytes, granulocytes and monocytes/macrophages, which are important for natural immunity, in addition to B lymphocytes immortalized with Epstein-Barr virus. These data suggest a specific role of the c-fgr protein in the effector cells of natural immunity.
It was recently found that in addition to brain, fyn is expressed in T lymphocytes. Expression of the fyn gene was also found to be elevated in T lymphocytes of murine lymphoproliferative diseases, suggesting that the fyn gene product is important in T lymphocyte activation.
Oncogenes II- G Proteins and Other Signalling Pathways
Dr. Frank McCormick (Cetus Immune Corporation, Emeryville, California) described the properties of ras p21 GTPase activating protein (GAP). GAP is a 120-kd protein found in all mammalian cells. It acts catalytically to convert ras p21. GTP to ras p21.GDP, thus inactivating the p21 protein. Mutations that convert ras proto-oncogenes into oncogenes (position 12, 13, 61, etc.) allow p21 to escape from GAP action and thus exist in the GTP-bound state constitutively. Human GAP cDNAS Were expressed in insect cells using baculovirus vectors and GAP purified from these cells to 95% purity. This recombinant GAP is at least 10-fold more active than GAP purified from mammalian cells, possibly because of posttranslational modifications. To determine whether GAP is a substrate for phosphorylation by tyrosine kinases, it was co-expressed with v-src, c-src, c-abl (types I and IV), P210 and P185 bcr/abl proteins, EGF-R or insulin receptor. In most cases, tyrosine phosphorylation was observed. c-abl proteins, however, did not phosphorylate GAP. These proteins (characterized in collaboration with Owen Witte and Ann-Marie Pendergast) have tyrosine kinase activity when expressed on their own in insect cells, but have reduced activity when co-expressed with GAP. Furthermore, GAP forms stable complexes with c-abl proteins, and with P210 and P185 bcr/abl proteins.
In mammalian cells, GAP is a major substrate for phosphorylation by tyrosine kinase oncogenes. Phosphorylated GAP associates with two other tyrosine kinase substrates, designated P210 and P60. The identity of these proteins is unknown. Phosphorylated GAP also binds to PDGF-receptor, in vivo and in vitro. However, GAP does not bind to PDGF-receptor mutants that lack the kinase-insert domain, nor to PDGF-receptor in ras-transformed cells. Results indicate that binding of GAP to PDGF-receptor is a necessary part of mitogenic signalling.
Dr. Takao Sekiya (National Cancer Center Research Institute, Tokyo) discussed detection of point mutations of genes in human cancers by single-strand conformation polymorphism analysis of polymerase chain reaction products. A simple and sensitive method of DNA analysis on nucleotide substitutions, namely, single-strand conformation polymorphism analysis of polymerase chain reaction products (PCR-SSCP), was applied to detection of mutated ras genes in surgical specimens of human lung cancers. Of 129 tumors analyzed, 22 contained mutated ras genes. Of 66 adenocarcinomas analyzed, 14 contained activated c-Ki-ras2 genes (six with mutated codon 12, four with mutated codon 13, one with mutated codon 18, and three with mutated codon 61), one contained the c-Ha-rasl gene with mutated codon 61, and three contained N-ras genes (one with mutated codon 12, and two with mutated codon 61). Besides the adenocarcinomas, mutated ras genes were found in two of 36 squamous cell carcinomas (the c-Ha-rasl gene with mutated codon 61) and two of 14 large cell carcinomas (the c-Ki-ras2 gene with mutated codon 12). However, in eight small cell carcinomas and five adenosquamous cell carcinomas, no mutated ras genes were detected. These results revealed that activation of ras in human lung cancers is not frequent (17%), the preferential histological type having the mutated ras genes was adenocarcinoma (82%), and 73% of the point mutations were in the c-Ki-ras2 genes at codon 12, 13, 18 or 61.
The PCR-SSCP technique was also applied to analysis of the retinoblastoma (RB) gene in human tumors. With a combination of reverse transcription of mRNA and PCR-SSCP analysis of cDNAs, mRNA was examined in 10 human tumor cell lines. Abnormal RB gene transcripts were found in the giant cell lung carcinoma cell line Lu65. This method revealed two types of RB cDNA in the Lu65 cells, suggesting the presence of a minor mRNA species with a base substitution in the nucleotide sequence corresponding to exon 2 and a major species of mRNA without nucleotide sequences corresponding to exon 2. Introduction of a stop codon in the reading frame by the C to A transversion suggested inactivation of the RB gene product. As a very minor species, the transcript without the exon 2 sequence, probably due to an alternating splicing, was also observed in other human tumor cell lines and human placentas. This observation suggested possible physiological significance of the mRNA species.
Dr. James R. Feramisco (University of California San Diego Cancer Center, La Jolla, California) reported on construction of mammalian cell lines with indicator genes driven by regulated promoters. The regulation of gene expression is a central feature of cell growth and differentiation. An important question is how information received at the cell surface is ultimately transmitted to the nucleus and how it causes changes in the pattern of gene expression. To study the molecular basis of this process, cell lines were established carrying marker genes under the control of known regulatory promoter elements. These cell lines are being used to investigate the effects of activating normal cellular second messenger systems or of microinjecting proteins hypothesized to function in signal transduction, such as oncogene products or subunits of enzymes.
Oncogenes III - Oncogenes Whose Products Target the Nucleus
Dr. Peter Vogt (University of Southern California, Los Angeles, California) presented his findings on the oncogenicity of jun. The jun gene codes for a 40 kD nuclear protein that is identical to a major component of the transcription factor AP-1. jun was originally recovered as a cell-derived insert in the genome of the avian sarcoma retrovirus ASV 17. It is the oncogenic effector of this virus. The chicken cellular jun gene in its normal genetic context is not oncogenic, while the viral jun gene shows high oncogenic activity. There are several structural differences between the two genes. The viral gene is fused at its 5’ end to viral gag sequences. A 27-amino acid deletion is located in the 5’ half of the viral jun coding domain, three amino acid substitutions occur in the 3’ half of the viral jun protein. The cellular jun mRNA carries signal sequences which mark it for rapid turnover.
A Rous sarcoma virus-based retroviral expression vector, RCAS, was used to examine the differences between viral and cellular jun proteins and to define the role of these differences in oncogenic transformation. Constructs were generated which express the viral jun protein, amino or carboxyl terminal deletions, recombinants between viral and cellular jun, or the unaltered cellular jun protein. These vectors were transfected into chicken embryo fibroblasts and were analyzed for jun expression and for oncogenic transformation. The results can be summarized as follows: deletion of the viral gag sequence significantly enhances oncogenic transformation by viral jun. Short amino terminal deletions of the jun coding sequences have no effect on the transforming ability of viral jun. Longer (approx. 100 amino acid) amino terminal deletions abolish oncogenic transformation. A carboxyl terminal deletion that may interfere with dimerization also inactivates the oncogenic potential of the viral jun protein. Cellular jun is very weakly transforming. This activity can be increased by replacing the amino terminal half of cellular jun with the corresponding half of the viral molecule. The recombinant is still significantly less oncogenic than wild-type viral jun. If, additionally, the 3’ untranslated region containing the instability signal of the cellular jun is deleted, then the recombinant acquires wild-type transforming activity. From these studies it was concluded that viral jun is highly oncogenic while cellular jun is only marginally oncogenic. The structural feature mainly responsible for the strong oncogenicity of viral jun is in the amino terminal half of the molecule, possibly the. 27-amino acid deletion, which could have an effect on the transcriptional activator function of jun. Besides this major contribution to oncogenicity, there appear to exist minor ones, possibly resulting from structural differences between viral and cellular jun in the carboxyl terminal region of the molecule.
Dr. Yoshiaki Ito (Kyoto University, Kyoto) reported on the nuclear oncogenes, c-jun and c-fos, as regulators of DNA replication. The products of nuclear oncogenes, c-jun and c-fos, are the components of the transcriptional activator, AP1, which represents one of the principal targets of signals elicited by growth factors or a tumor promoter, 12-0-tetradecanolphorbol-13-acetate (TPA). Deregulation of transcription in the cell has been considered as the mechanism of cell transformation by these genes. Polyomavirus (Py) DNA replication requires an enhancer activity in addition to the Py origin in cis and Py large T antigen in trans and is thought to mimic cellular DNA replication. We observed that overexpression of the c-jun and c-fos genes enhanced Py DNA replication dramatically by stimulating the function of the AP1 binding site.
These findings raise an intriguing question as to whether AP1 also participates in the regulation of cellular DNA replication. Since the most fundamental property of cancer cells is their uncontrolled growth, it is important to examine whether the oncogenes, c-jun and c-fos, are directly linked to uncontrolled DNA replication. It is worth noting, in that context, that the AP1 binding site is a well-known TPA-responsive element and that TPA induces cellular DNA synthesis in quiescent cells. Dr. Ito’s group recently showed that TPA activates Py DNA replication through the AP1 binding site. The possible involvement of the TPA-responsive element in the regulation of cellular DNA replication deserves rigorous investigation to understand the oncogenic processes induced by the c-jun and c-fos genes and the tumor promoting effect of TPA.
Dr. Mark Groudine (Fred Hutchinson Cancer Research Center, Seattle) presented a molecular analysis of the c-myc transcription elongation block-- implications for the generation of Burkitt’s lymphoma. The steady-state levels of c-myc RNA are controlled by several mechanisms including the rate of initiation of transcription, a block to transcription elongation, and stability of the cytoplasmic transcript. Burkitt’s lymphoma cells consistently exhibit c-myc chromosomal translocations mutations within and flanking the translocated, allele, a loss of the block to transcription elongation in c-myc exon 1 and a promoter shift to use of the upstream P1 promoter. To define the mechanism responsible for the loss of transcription elongation blockage and resulting c-myc deregulation in Burkitt’s lymphoma, Dr. Groudine’s group examined the transcription of normal and Burkitt’s lymphoma c-myc alleles introduced into murine cells and Xenopus oocyte germinal vesicles. The data suggest that although the mutations within and surrounding several Burkitt’s lymphoma c-myc alleles are not in themselves sufficient to abrogate the transcription elongation block, transcription initiation from the P2 promoter may be necessary to obtain the block to transcription elongation. As a direct test of the role of c-myc promoters in programming the block to transcription elongation, transcription patterns from in vitro mutagenized c-myc genes containing deletions of either the P1 or P2 promoter were analyzed. The results of these analyses confirm that P1-initiated c-myc transcripts do not terminate at discrete sites near the 3’ end of exon 1, whereas P2-initiated transcripts either terminate or read-through the transcription block signals. Therefore, overexpression and/or constitutive expression from the c-myc P1 promoter may contribute to increased read-through transcription in Burkitt’s lymphoma cells, and hence to aberrant expression patterns or levels of c-myc steady-state transcripts.
Dr. Masabumi Shibuya (University of Tokyo, Tokyo) described the analysis of activated oncogenes in human brain tumors. Some of the receptor-type tyrosine kinase genes, i.e., EGF receptor (EGFR) gene and c~rbB2/neu/HER2 gene, have frequently been found to be amplified in a variety of human tumors. However, the possibility of structural alteration of these genes in tumor tissues has not been extensively studied as yet. Dr. Shibuya's group recently reported that a few cases of human glioblastoma multiforme (GL-3, GL-5) carry a structural mutation of the EGFR gene, and the product of these mutated genes, a 140-kd receptor, showed a constitutive activation of tyrosine kinase.
In this study, the primary structure of these mutated EGFR genes, their levels of cDNA and genomic DNA were examined. EGFR cDNA was isolated from the GL-5 tumor which bore a mutated and amplified EGFR gene. Sequence analysis of the cDNA revealed that a large internal deletion of about 800 bp (from the amino acid residue #30 to #297); in other regions, including transmembrane and tyrosine kinase domains, predicted amino acids were identical to those of the normal EGFR. DNA fragments containing the junction point of deletion mutation in the GL-3 and GL-5 glioblastomas were isolated and the sequences compared with corresponding sequences in human placenta DNA The results were as follows: (1) In both GL-3 and GL-5, deletion mutation took place between the same two introns (#1 and #6) in the EGFR gene. (2) In each intron, the deletion points in GL-3 and GL-5 were not identical. (3) Inverted repeat sequences, not a homologous recombination, seemed to be involved in the recombination process. Using PCR with the oligonucleotides near junction points, another glioblastoma, GL-11, was shown to have a similar deletion mutation in the EGFR gene. These results suggest that a population of human brain tumors carry structurally altered and activated EGFR genes generated by a nonhomologous recombination mechanism. The transforming activity of the mutated EGFR gene was discussed.
In addition, one case of glioblastoma was found to carry 30- to 40-fold c-myc gene amplification. From this cellular DNA, a number of probes randomly distributed within the amplicon units were obtained. These probes may be useful for studying the mechanism and structure of c-myc gene amplification.
Tumor Suppressor Genes
Dr. Webster K. Cavenee (Ludwig Institute for Cancer Research, Montreal, Quebec) discussed the molecular genetics of human cancer predisposition and progression. The development of human cancer is generally thought to entail a series of events that cause a progressively more malignant phenotype. Such a hypothesis predicts that tumor cells of the ultimate stage will carry each of the events, cells of the penultimate stage will carry each of the events less the last one and so on. That is to say, a dissection of the pathway from a normal cell to a fully malignant tumor may be viewed as the unraveling of a nested set of aberrations. In experiments designed to elucidate these events, Dr. Cavenee’s group has compared genotypic combinations at genomic loci defined by restriction endonuclease recognition site variation in normal and tumor tissues from patients with various forms and at various stages of cancer. The first step, inherited predisposition, is best described for retinoblastorrra in which a recessive mutation of a locus residing in the 13q14 region of the genome is unmasked by aberrant, but specific, mitotic chromosomal segregation. A similar mechanism involving the distal short arm of chromosome 17 is apparent in astrocytic tumors and the event is shared by cells in each malignancy stage. This is distinct from a loss of heterozygosity for loci on chromosome 10 which is restricted to the ultimate stage, glioblastoma multiforme. These results suggest a genetic approach to defining degrees of tumor progression and a means for determining the genomic locations of genes involved in the pathway as a prelude to their molecular isolation and characterization.
Dr. Makoto Noda (Institute of Physical and Chemical Research, Ibaraki, Japan) presented a genetic analysis of the Krev-1 transformation suppressor gene. Krev-1 cDNA encodes a ras-related protein and exhibits an activity of inducing flat revertants at certain frequencies when introduced into a v-K-ras-transformed mouse NIH/3T3 cell line, DT. Toward understanding the mechanism of action of the Krev-1 protein, Dr. Noda’s group constructed a series of point mutants of Krev-1 cDNA as well as some chimeric genes composed of fragments of H-ras and K-rev-1 cDNAS, and tested their biological activities in DT and a human fibrosarcoma cell line, HTI080, known to harbor the activated N-ras gene. Results indicate that some of the residues within the putative guanine nucleotide-binding regions (Asp17 and As116), putative effector-binding domain (around residue 38), the unique Thr 61 residue and the putative acylation site (Cys181) are all essential for the transformation suppressor/tumor suppressor activiy. On the other hand, the regions surrounding Gly12 and Gln63 seem to play some regulatory roles, and certain substitutions at these positions are found to greatly increase the tumor-suppressing activity of Krev-1. A study with chimeric cDNAs revealed that the determinant for the suppressor activity resides in the N-terminal one-third of the K-rev-1-encoded polypeptide. The finding that a mutation such as Val12, which would be expected to decrease the GTPase activity of this ras-related protein, potentiated its biological activity raises the possibility that this protein may be regulated by an upstream negative-regulatory signal.
Dr. Edward Harlow (Cold Spring Harbor Laboratory, Cold Spring Harbor, New York) delivered the last presentation of the seminar, “Cellular targets for transformation by DNA tumor viruses.” Tumor suppressor genes ate genetic loci whose loss has been associated with tumor development. Because the inactivation of both alleles of these genes is a key feature in the genesis of certain tumors, it has been postulated that the protein products of tumor suppressor genes function in the negative regulation of cell proliferation. Tumor suppressor genes have been identified by genetic analysis either as loci associated with an inherited predisposition to certain tumors or by karyotic studies that have localized putative tumor suppressor genes to loci that show reduction to homozygosity or loss of heterozygosity during tumor development. Second, recent work from a number of laboratories has shown that the protein products of tumor suppressor genes often form protein/protein complexes with the transforming proteins of small DNA tumor viruses. The retinoblastoma gene, RB-1, is one of the best studied examples of the tumor suppressor genes. It was originally identified and cloned through its association with childhood retinoblastoma. However, it is also a key target for transformation by the oncogenes of several small DNA tumor viruses. The E1A proteins of adenovirus, the large T antigens of polyomaviruses, and E7 protein of papillomaviruses all bind to p105-RB. Genetic studies of all three viruses have shown that any mutation that destroys binding to p105-RB also destroys the ability of these proteins to transform cells, suggesting that interaction with the RB gene product is a key event in viral transformation. In addition to interacting with p105-RB, the adenovirus E1A proteins and the polyomavirus large T antigens also bind to other cellular proteins. One of these, a protein with a molecular weight of 107,000 daltons, 107K, binds to E1A and large T at the same amino acid region as p105-RB, suggesting that the 107K and p105-RB proteins may have structural similarities. These observations and other comparisons raise the question whether 107K may be functionally related to p105-RB. If so, the 107K protein may be another example of a product of a tumor suppressor gene interacting with the transforming proteins of small DNA tumor viruses.
Dr. Terada closed the meeting by thanking the participants for a very stimulating meeting. He had been impressed with the quality of the presentations and with the rapid progress in identifying the genes and molecular mechanisms involved in the conversion of normal cells to neoplasia. The insights being developed had relevance for the normal biology of growth and development as well as for cancer diagnosis, prognosis, prevention and control.
(2) Seminar on “Recent Advances in Research on Heterocyclic Amines'’
This seminar was held on March 19 and 20, 1990 at the Coco Palms Resort, Kauai, Hawaii. The organizers were Dr. Snorri S. Thorgeirsson, National Cancer Institute, Bethesda, Maryland, U.S.A. and Dr. Minako Nagao and Dr. Hiroyasu Esumi, National Cancer Center Research Institute, Tokyo, Japan. There were eight speakers from Japan and six from the United States. The purpose of the seminar was to exchange recent findings and views on recent advances in research on heterocyclic amines and to discuss uncertainties and scientific opportunities in this area.
Dr. Takashi Sugimura (National Cancer Center Research Institute, Tokyo, Japan) delivered the opening address to the meeting. He indicated that in 1928 Dr. Bauer had proposed that cancer cells arise from normal cells through somatic mutation. Data showing carcinogens are mutagens accumulated in the following decades. Based on these facts, Ames developed in 1974 sensitive bacterial strains (Ames Salmonella test) having the genes encoding error prone-repair enzymes as plasmids to detect chemical mutagens. By epidemiological approach, Drs. Peto and Doll, reported that following smoking, diet is a major cause of human cancers.
In 1976, Dr. Sugimura’s group, using the Ames Salmonella test, found strong mutagemcity in grilled fish and meat, and smoke produced during cooking meat or fish. In the following few years they determined chemical structures of mutagenic compounds in cooked meat and fish to be new compounds. Thus far, all compounds isolated as mutagens have a free amino group and arylnitrogen, and are referred to as heterocyclic amines. At Lawrence Livermore National Laboratory, several heterocyclic amines were also found. Carcinogenicity of Trp-P-1 and Trp-P-2 in mice was first reported by Matsukura et al. in 1981. All heterocyclic arnines examined so far were carcinogenic to mice and rats; and one, IQ, which has been evaluated in monkeys, was carcinogenic in that species.
Recent studies on the molecular biology of human cancer revealed that human cancer development is multistep and probably multifactorial. Therefore, carcinogenic data on young healthy animals is difficult to apply directly to human risk estimation. For human cancer development, many endogenous and exogenous factors may be involved. Based on these circumstances, Dr. Sugimura stressed the importance of clarifying the role of heterocyclic amines in human carcinogenesis since most people are exposed to a substantial amount of heterocyclic amines every day.
Dr. Hikoya Hayatsu (Okayama University) reported on the modifying effect of biological substances on the genotoxicity of heterocyclic amines. The genotoxic activity of heterocyclic amines can be demonstrated with a variety of biological systems: DNA, phage, bacteria, eukaryotic cells and whole organisms. By using these systems, Dr. Hayatsu has found several classes of antimutagenic biochemical agents against the actions of heterocyclic amines. Dr. Hayatsu discussed the suppressing effect of porphyrins and chlorophylls on the genotoxicity of heterocyclic amines particularly the toxicitv of Trp-P-2 and its metabolites.
In the Salmonella mutagenicity test, the activity of Trp-P-2 (NHOH), a metabolically activated form of Trp-P-2, was suppressed effectively by hemin, chlorophyll and their derivatives. Particularly strong suppression was observed with hemin and with iron chlorin e6. It was also found that hemin and iron chlorin e6 can mediate a rapid degradation of Trp-P-2 (NHOH).
With Drosophila melanogaster, the tests for wing-spots and DNA-repair showed the genotoxicity of Trp-P-2. In these tests, the mutagen was administered orally to the larvae so that they are better models than the bacterial systems to the human situation. Glu-P-1 MeIQx and other heterocyclic amines also gave positive responses in the DNA-re air test. When chlorophyll was supplemented to the mutagen-containing larvae-feed, the genotoxicity of Trp-P-2 was suppressed. The extent of suppression was dependent on the dose of the chlorophylls. The suppression by chlorophyllin was as effective as by chlorophyll. These inhibitions were also noted for Glu-P-1 and MeIQx.
Dr. Hayatsu showed data that indicated complexing of Trp-P-2 with chlorophyllin. It seems possible that the chlorophylls act as a trapping agent to the heterocyclic amines. However, chlorophyllin may act by inhibiting absorption of Trp-P-2 and other heterocyclic amines.
Dr. Elizabeth Snyderwine (National Cancer Institute, Bethesda, Maryland) spoke on the metabolic processing and disposition of 2-amino-3-methylimidazo[4,5-f]quinoline (IQ) in monkeys, and detection of heterocyclic arylamine (HAA) adducts in specific genes in vitro. Dr. Snyderwine presented results from studies examining the metabolic fate and disposition of IQ in two cynomolgus monkeys 14C-IQ (2.2 mol-120 Ci/kg, p.o.) was extensively metabolized in monkeys and excreted in urine and feces as metabolites. None of the parent compound was detected in urine or feces. In contrast, analysis of plasma by HPLC revealed the presence of only parent compound. Urine was the predominant route of excretion accounting for approximately 60% of the dose 72 hr after IQ administration. Fecal excretion accounted for only 9% of the dose during this same time period. The clearance of IQ from the blood was biphasic. A rapid clearance rate was observed during the first 8 hr and was followed by a much slower rate from 8 to 72 hr. Approximately 91% of the 1-hr peak blood level was cleared by 8 hr, while only an additional 7% was cleared from 8 to 72 hr. Concomitant with the clearance of the compound from blood, the urine showed a high rate of excretion of IQ metabolites during the first 8 hr followed by a slower excretion from 8 to 72 hr. Eight metabolites were found in the urine of monkeys at all time periods examined following IQ administration. IQ-5-O-glucuronide, IQ-sulfamate, IQ-5-sulfate and the N-demethylated metabolite of IQ have been identified in urine by comigration with synthetic standards on HPLC, u.v. spectra, and enzymatic and chemical analyses. At least three metabolites were found in feces and one major fecal metabolite was identified as IQ-sulfamate. By 72 hr after IQ administration, approximately 30% of the dose could not be accounted for in urine and feces indicating that a significant portion of the dose was retained. This retained amount may represent, in part, IQ covalently bound to tissue proteins or adducted to DNA. Dr. Snyderwine concluded that these results suggest that repeated daily exposure to IQ in foods may result in accumulation of the compound in the body.
During the second part of the presentation, Dr. Snyderwine showed results from studies examining HAA adducts in specific genes in vitro. Dr. Snyderwine has established methodology to detect these adducts of the food-derived HAAs in specific gene sequences in carcinogen-treated cultured cells and in DNA reacted in vitro with N-hydroxylamines. The technique involves treatment of isolated DNA with the bacterial repair enzyme uvrABC exonuclease followed by quantitation of damage by Southern hybridization. Dr. Snyderwine showed dose response of initial adduct formation and the rate of repair in the DHFR gene in CHO cells following N-OH-IQ exposure.
Dr. Takesi Kato (Osaka University, Osaka) dealt with the spectrum of mutations induced by Trp-P-2 in a cDNA of the human HPRT gene. Dr. Kato has used the pZipHprtNeo shuttle vector to determine the types of DNA sequence alterations induced by a potent carcinogen 3-amino- 1-methyl-5H-pyrido[4,3-b]indole (Trp-P-2). The shuttle vector contains a human cDNA hprt as the target gene and is stably integrated into a chromosome of the mouse cell line VH12. After Trp-P-2 treatment, 59 independent HPRT- mutant clones of VH12 were isolated and altered sequences of the mutant hprt- cDNA genes were determined by the dideoxy sequencing method. Mutations induced by Trp-P-2 comprised a variety of events: base substitutions, frameshifts, deletions and complex. Frameshifts were the most frequent mutational events (51%), and base substitutions were the next most frequent (30%) followed by deletions (14%). Examination of the DNA sequence context in the mutant genes revealed that approximately 70% of mutations induced by Trp-P-2 occurred at G:C sites and thymine residues were the suggested target for the remainder of the mutations. The results seem consistent with the known biochemical reaction of this chemical; in vivo, metabolically activated Trp-P-2 binds specifically to the C8 position of guanine residues in DNA to form C8G-Trp-P-2 adducts. As for molecular mechanisms, Dr. Kato showed that slippage misalignment could predict the generation of a large portion of Trp-P-2 induced mutations found in the cDNA gene, and suggested that slippage misalignment may be an important general mechanism of mutagenesis in eukaryotic cells.
Dr. Masakuni Degawa (Tohoku University, Sendai) reported on the induction of the cytochrome P-450 isozyme(s) with heterocyclic amines and other aromatic amines. Dr. Degawa demonstrated that administration of carcinogenic aromatic amines such as amino acid- and protein-pyrolysate components and 4-aminoazobenzene derivatives preferentially induced the P450IA2 isoform that is responsible for mutagenic activation of the carcinogen, itself, in the liver. Dr. Degawa assessed the species, sex and organ differences in the induction of the P-450 isozyme(s) responsible for mutagenic activation of the carcinogenic aromatic amines. The results indicated that the induction of the carcinogen activation enzyme is restricted in the liver but not in extrahepatic organs such as lung, kidney, small intestine or large intestine and that there are species and sex differences in hepatic microsomal carcinogen activation enzyme(s) among rats, mice, hamsters and guinea pigs. Dr. Degawa also reported that in some cases of the aromatic amine-induced hepatocarcinogenesis, the carcinogenic susceptibility of animals was found to correlate with the induction rate and/or total activity of the carcinogen activation by P-450.
Dr. Narayana Battula’s (National Cancer Institute, Bethesda, Maryland) presentation dealt with the genotoxicity of food-derived heterocyclic amines activated by cytochrome P450IA2 in mammalian cells. Dr. Battula noted that tests for evaluating potential genotoxicity and mutagenicity of xenobiotics and their extrapolation to man are more appropriate when carried out in mammalian cells than in non-mammalian cells. Most mammalian cells in continuous culture unfortunately do not retain the cytochromes P-450 dependent metabolic activation capabilities of the tissues of their origin. Therefore, several supplementary systems for activation have been used with mammalian test systems and they have been unsatisfactory since the activation mixture used was undefined. To meet the longstanding need for mammalian cell lines containing constitutive levels of P-450 enzymes which activate xenobiotics to forms that interact with cellular DNA, Dr. Battula has developed a prototype mammalian cell system constitutively expressing cytochrome P450IA2.
Dr. Battula has introduced cytochrome P450IA2 DNA into rat liver epithelial (RLE) cells. The introduced DNA directed the expression of the cytochrome P4501A2 enzyme. The cells constitutively expressing cytochrome P450IA2 were exposed to different concentrations of IQ, MeIQx and PhIP. DNAs extracted from the exposed cells were analyzed for DNA adducts by 32P-postlabeling assays. Cells exposed to IQ produced five specific adducts and there were no adducts in control cells. The formation of these adducts depended on time and concentration of the mutagen. The adducts formed in the clones were identical to those formed in rat and mouse liver after in vivo administration of IQ. The major adduct was identified as N-deoxy-guanosine-8-yl-IQ. In similar analysis MeIQx and PhIP showed two and three specific adducts each, respectively. Dr. Battula also showed that the formation of the adducts was inhibited by 7,8-benzoflavone, a known inhibitor of P450IA2 activity. Thus, these mutagens are metabolically activated by cytochrome P450IA2 and are genotoxic.
Dr. Battula has further examined the consequences of DNA-carcinogen adducts in these mammalian cell clones. As the target for mutations Dr. Battula chose to use the hypoxanthine-guanosine phosphoribosyl transferase (HPRT) gene which codes for the enzyme HPRT. Cells expressing P450IA2 were exposed to the food-mutagens at different doses and cultured in selective medium containing 6-thioguanine (6-TG). Biochemical analysis of the 6-TG resistant colonies for HPRT activity showed that all the clones either lacked HPRT activity or had much reduced levels, thus demonstrating that the method selects for true mutants. All of the tested food-derived carcinogens were mutagenic and the mutation frequency was dependent on the dose of the mutagen. Dr. Battula showed that PhIP, which constitutes more than 90% of the mass of bacterial mutagenesis in beef and is the least mutagenic in Ames bacterial mutagenesis assay, was more genotoxic and mutagenic than IQ and MeIQx in the RLE system. It was concluded that these unexpected results point to a major difference in mutagenic potencies between bacterial and mammalian assays and the importance of using mammalian test systems for mutagenic evaluation. These results clearly demonstrate that P450IA2 selectively metabolizes the food-derived mutagens to reactive forms which affect the chromosomal DNA in the mammalian cells in such a way that it produces a permanent genetic change. Thus, this prototype mammalian cell system is well defined for mutagenic studies and integrates the processes of metabolic activation, genotoxicity and mutagenesis.
Dr. Nobuyuki Ito’s (Nagoya City University Medical School, Nagoya) presentation was on demonstration of synergism between five heterocyclic amines in a rat liver carcinogenesis, medium-term bioassay system (DEN-PH model) and IQ dose response activity. Dr. Ito reported on the carcinogenic potential and modifying effects of combined treatment with heterocyclic amines produced as pyrolysates of amino acids, proteins and meat by utilizing his rat in vivo medium-term bioassay system for liver carcinogens.
Male F344 rats, aged 6 weeks old, were given a single i.p. injection of diethylnitrosamine (DEN, 200 mg/kg) and starting 2 weeks later received test compound(s) in the diet for 6 weeks. Control groups were given DEN or test compound(s) alone. All rats were subjected to 2/3 partial hepatectomy (PH) at week 3 and killed at week 8. Compounds and doses were as follows: 3-amino-1,4-dimethyl-5H-pyrido[4,3-b]indole (Trp-P-1, 0.015%), 2-aminodipyrido[1,2-a:3’,2’-d]imidazole (Glu-P-2, 0.05%),
2-amino-3-methylimidazo[4,5-f]quinoline (IQ, 0.03%),
2-amino-3,4-dimethylimidazo[4,5-f]quinoline (MeIQ, 0.03%),
2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline (MeIQx, 0.04%). These dose levels were reported as being tumorigenic in long-term carcinogenicity studies in rats. Additional groups were given 1/5 or 1/25 doses individually or in combination to allow assessment of the effects of simultaneous administration of all five chemicals.
Final body weights were lower, in a dose-related manner, in the Trp-P-1 treated groups and relative liver weights were increased with all chemicals at high doses. Numbers of glutathione S-transferase placental form (GST-P) positive liver cell foci were significantly increased with carcinogenic doses of all chemicals. Trp-P-1, IQ and MeIQ exerted positive influence even at the 1/5 dose level. Similar results were obtained for areas of foci with all chemicals except for Glu-P-2. MeIQ was also positive at the 1/5 dose. Synergism regarding induction of GST-P positive foci was observed; the values in both number and area for the group of animals given 5 chemicals together were significantly greater than the sum totals of individual 1/5 or 1/25 dose group data.
Dr. Ito also reported on the dose response for 2-amino-3-methylimidazo[4,5-f] quinoline (IQ) initiating activity in rat liver. The experimental system consisted of 6 week old male F344 rats receiving a 2-week dietary administration of IQ at doses of 0.1, 0.05 or 0.025% and were then maintained on diet supplemented with phenobarbital (PB, 0.05%) or 3’-methyl-4-dimethylaminoazoberLzene (3’-Me-DAB, 0.0024%) from week 3 until final sacrifice at week 86.
Dr. Ito showed that dietary administration of IQ resulted in dose-dependent development of hepatocellular carcinoma with both PB and 3’-Me-DAB promotion. Quantitation of GST-P positive liver cell foci revealed significant correlation of IQ concentration with lesion area. The prior treatment with IQ also resulted in significant development of thyroid tumors with PB, but not with 3’-Me-DAB promotion.
It was concluded that these findings demonstrated heterocyclic amines to posses liver carcinogenic activities as well as synergistic effects when administered at low dose levels. This is of major interest considering their potential importance as dietary contaminants.
Dr. Hiroyasu Esumi (National Cancer Center Research Institute, Tokyo) reported on the carcinogenicity of PhIP in CDF1 mice and metabolic aspects of PhIP. Dr. Esumi showed data from a long-term carcinogenesis experiment of PhIP in CDF1 mice using both males and females. The most frequent tumor observed was a lymphoma affecting the mesenteric lymph node and spleen in both sexes. The majority of the induced lymphomas had surface IgG upon immunofluorescence staining, indicating a B cell origin. In contrast to most other heterocyclic amines, PhIP was not carcinogenic to the liver under the present conditions. Histological examination showed that much less damage was caused in the liver. Approximately 70 times fewer DNA adducts were observed by 32P-postlabeling in the liver of rats administered PhIP when compared to IQ.
The metabolic processing of PhIP was examined by following 14C-labeled PhIP administered by gavage to rats at a dose of 0.6 mg/kg. Almost 70% of the administered PhIP was excreted into the feces in 48 hours. Most of the radioactivity was recovered as unchanged PhIP. Similarly the major radioactive peak in the bile was shown to be unchanged PhIP. The major urinary product was a sulfate conjugate as previously reported. Dr. Esumi also showed that PhIP feeding for 2 weeks resulted in induction of the P450IA2 isoform.
Dr. Richard Adamson’s (National Cancer Institute, Bethesda, Maryland) presentation was on carcinogenesis studies of aminoimidazoazaarenes in nonhuman primates. Dr. Adamson noted that the aminoimidazoazaarenes (AIAs) consist of imidazoquinolines, imidazoquinoxalines and imidazopyridines, which are formed under normal cooking conditions in cooked meat. Based on specific activity in the Ames/Salmonella assay, chemical structure, contribution to total mutational activity in cooked beef, type of DNA adducts formed, carcinogenic activity in rodents and availability of compound, one of the quinolines (IQ, 2-amino-3-methylimidazo[4,5-f]quinoline), one of the quinoxalines (MeIQX, 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline), and one of the pyridines (PhIP, 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine was selected for studies of carcinogenic activity in nonhuman primates, primarily cynomolgus monkeys. Dr. Adamson indicated that studies were underway on IQ and MeIQx and will shortly begin on PhIP.
IQ was administered orally to monkeys as they became one year of age at a dose of 10 or 20 mg/kg 5 times per week. Each dose was administered to a group of 20 monkeys. The oldest monkeys have been on the compound approximately five years. Thus far, tumors have been induced in two monkeys at the 10 mg/kg level and in nine monkeys at the 20 mg/kg level. The average latent period was greater for the two monkeys at the 10 mg/kg level than those at the 20 mg/kg level. At both dose levels all the tumors induced were primary hepatocellular carcinoma.
Similar to IQ, MeIQx was administered orally to monkeys as they became one year of age at a dose of 10 or 20 mg/kg 5 times per week. Each dose is being administered to a group of 10 monkeys. The oldest monkeys have been on the compound 15 months. Thus far no tumors have occurred in these monkeys.
Dr. Adamson also reported on recent studies on the use of 32P-postlabeling assay to monitor adducts in white blood cells of monkeys fed IQ, MeIQx and PhIP. IQ adducts are routinely detected in the monkeys given either the low or the high dose. However, MeIQx adducts have so far not been detected in the white blood cells. Dr. Adamson showed very preliminary data on the distribution of PhIP adducts in the monkey tissues. The adduct pattern is similar to that found in rodents by Japanese investigators; namely, relatively high levels were found in the heart, pancreas, and salivary gland as well as in white blood cells and intermediate levels in the liver.
Dr. Unnur Thorgeirsson (National Cancer Institute, Bethesda, Maryland) reported on IQ-induced hepatocellular carcinoma in nonhuman primates: autopsy findings. At autopsy, the extent of hepatocellular carcinoma (HCC) ranged from a single nodule of 2.2 cm in one animal to massive multinodular involvement in another. The first monkey was autopsied in 1988 after receiving 20 mg/kg of IQ for almost two and a half years. In this animal 80% of the liver w is replaced with multiple necrotic nodules of HCC which had invaded the stomach and diaphragm and metastasized to the lungs and hilar lymph node. Another monkey in the 10 mg/kg group had gross evidence of lung metastases. Two cases had microscopic evidence of vascular invasion in the liver and tumor thrombi in the pulmonary vessels. Microscopically, most of the tumor nodules demonstrated a trabecular pattern of HCC, with less frequent occurrence of a diffuse pattern and pseudoglandular formation. For the most part the IQ-induced HCC was well to moderately well differentiated with relatively little desmoplastic response. In a few instances there was a great variability in the appearance of the individual nodules within the same liver, ranging from well-differentiated HCC with rare mitotic figures to desmoplastic tumor nodules with highly atypical spindle cell morphology. There was extensive lymphocytic infiltration within and around some of the tumor nodules. Presence of hyperplastic nodules was detected in rare instances. The most striking feature in sections of the non-neoplastic liver was the evidence of clear cell foci, ranging from few cells to several mm in diameter. These cells had distinct cell boundaries, abundant clear cytoplasm, and small dark eccentric nuclei. Special stains demonstrated an abundance of glycogen within the clear cells. Other findings included mild lymphocytic infiltrate of the portal areas.
Besides the HCC with pulmonary metastases, the only notable findings were that of lymphoid hyperplasia. It was most prominent in mesenteric lymph nodes, but was also seen in other abdominal and hilar nodes. Prominent submucosal lymphoid aggregates were observed throughout the colon in almost all of the cases. These appeared as whitish round 1-2 mm mucosal elevations. Microscopically, the submucosal colonic aggregates and the enlarged lymph nodes presented as lymphoid hyperplasia with prominent follicles of variable size with pronounced mitotic activity.
Dr. Thorgeirsson concluded that the study has unequivocally demonstrated that IQ is a potent liver carcinogen in nonhuman primates.
Dr. Yuzo Hayashi (National Institute of Hygienic Sciences, Tokyo) spoke on the risk assessment of heterocyclic amines in cooked food. Dr. Hayashi indicated that the purpose of his presentation was to illustrate a possible approach to cancer risk assessment in man by daily intake of heterocyclic amines from food. Hazard identification: Long-term feeding of heterocyclic amines are known to cause cancers in experimental animals by direct genotoxic effects on their target cells. Biochemical studies indicate that heterocyclic amines are metabolized to genotoxic derivatives by cytochrome P-450 and other enzymes such as sulfotransferase and acetyltransferase known to ubiquitously occur in various species. Therefore, it is likely to assume that heterocyclic amines possess the potential to cause cancers in man. Dr. Hayashi indicated that using the best exposure and dose-response assessments provided only limited mathematical models applicable for low dose extrapolation. However, he concluded that the upper limit estimate of tumor risk by daily intake of heterocyclic amines from food for lifetime is around the order of 10-5.
Dr. Minako Nagao’s (National Cancer Center Research Institute, Tokyo, Japan) presentation involved two topics: (1) analysis of MeIQx-DNA adducts in rats and humans, and (2) actiated oncogenes of tumors induced by heterocyclic amines. Dr. Nagao noted that heterocyclic amines are converted to proximate or ultimate forms by cytochrome P450IA2. A large difference in the amounts of cytochrome P450IA2 and heterocyclic amine metabolizing activity between individual humans has been reported. She then proposed that analysis of heterocyclic amine-DNA adduct levels would give us some important information to estimate the risk of heterocyclic amines in human carcinogenesis. MeIQx, the concentration of which is the highest in food except for PhIP, was analyzed on DNA-adduct formation by the 32P-postlabeling method. Three major adducts were detected in the DNA modified in vitro by MeIQx and rat liver microsomes, after hydrolysis with nuclease P1; two were identified as guanine adducts. The same three adducts were also detected as major modified nucleotides in vivo. Dr. Nagao showed that adduct levels of rat liver, pancreas and colon increased with an increase of MeIQx concentration in the diet, and adducts were detected after 4 weeks of continuous feeding of 0.1 ppm MeIQx. The levels of adducts also increased with an increase of administration period. Dr. Nagao also showed that the expression of the MDR gene was increased after feeding 400 ppm MeIQx for 8-12 weeks.
Dr. Nagao presented data from human liver samples showing similar but not identical adducts (spots) as observed for the MeIQx C8 adduct.
The second topic of Dr. Nagao’s presentation dealt with activated oncogenes of tumors induced by heterocyclic amines. By NIH 3T3 cell transfection assay, DNA of one of five rat hepatocellular carcinomas (HCCs) induced by IQ produced a primary transformant in which c-Ha-ras was activated. Selective oligonucleotide hybridization or direct sequencing was performed on PCR amplified sequences of the other 16 rat hepatic tumors induced by IQ, and two were shown to have mutations in the exon 1 of c-Ha-ras. These two mutations were located in the codon 13. In the case of tumors induced by MeIQx, c-Ki-ras mutation was found in one of ten HCCs. No mutation was found in exon 1or 2 of either c-Ha-ras or N-ras in these 10 HCCs. Dr. Nagao concluded from these results that involvement of ras activation would be rare in hepatocarcinogenesis of rats induced by IQ or MeIQx. Further, no ras mutations in exon 1 or 2 was found in a monkey HCC induced by IQ, by direct sequencing of DNA amplified by PCR. Nevertheless, 4 of 7 and 2 of 5 Zymbal gland squamous cell carcinomas induced by IQ and MeIQx, respectively, possessed mutation in c-Ha-ras. c-Ha-ras mutation may play an important role in Zymbal gland carcinogenesis.
Dr. Nagao presented data showing elevated c-raf expression in hepatic tumors of rats induced by IQ. However, she found no gross changes of c-raf genomic sequence or minor changes in the critical region of c-raf in 30 hepatic tumors of high c-raf expressors Dr. Nagao reported that in these hepatic tumors, protein phosphatase 2A mRNA was also highly expressed. Dr. Nagao also showed that okadaic acid, a protein phosphatase inhibitor, can inhibit transformation with raf and ret II in NIH 3T3 cells, indicating a possible involvement of protein phosphatase in carcinogenesis.
Dr. James Felton’s (Lawrence Livermore National Laboratory, University of California, Livermore, California) presentation dealt with heterocyclic amines in cooked food: identification, human monitoring and low level detection. Dr. Felton noted that the cooking of food derived from muscle generates heterocyclic amines that are very potent bacterial mutagens. All of the compounds tested have proven to be carcinogens following feeding to mice, rats and monkeys. Presently 12 mutagenic compounds have been found in cooked protein containing food derived from beef, pork, chicken, or fish consumed in the Western diet. Only 6 of these compounds have been definitely assigned a chemical structure. Specifically designed monoclonal antibodies are being used to quantify these mutagens in various foods and to help isolate unknown mutagens and metabolites. These antibodies can detect nanogram amounts each of
2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP),
2-amino-3-methylimidazo[4,5-f]quinoline (IQ), and
2-amino-3,4-dimethylimidazo-[4,5-f]quinoxaline (MeIQX) in the meat matrix, but specific quantitation requires a prior HPLC separation before use of antibodies for identification.
Dr. Felton discussed the analysis of the DNA binding of heterocyclic amines by the very powerful method of accelerator mass spectroscopy. With this method, Dr. Felton has been able to measure one DNA adduct per 1011 nucleotides, the most sensitive measurement of this type made to date. The binding appears to be linear even to this low level of adducts. The method is so sensitive that human DNA binding and pharmacokinetic studies should be possible because the radiation from a 14C dose will be so low it will not contribute significantly over the background abundance of 14C in the body.
Dr. Felton showed that human subjects fed fried ground beef as their sole source of cooked protein revealed high Ames/Salmonella activity in their urine. Approximately 50% of the activity required acid treatment to release the conjugates. Antibody-positive material was also found in the urine from these individuals, although correlation and quantification with known AIA standards is difficult. After consumption of the fried beef diet, blood from splenectomized individuals was used for micronuclei analysis. The majority showed an increase over control and followed the kinetics of the exposure, but folate levels may be indicated as a major factor in the response.
Dr. Felton also showed that CHO cells containing the P450IA2(P3) gene were much more effective in activating PhIP than IQ. Similarly, Dr. Felton showed that repair-deficient CHO cells containing the P3 gene are extremely sensitive to toxic effects of PhIP.
It was felt by Dr. Adamson, Dr. Sugimura and all the participants that the quality of the presentations had been of high quality and new insights into the carcinogenicity of the heterocyclic amines and their relevance for human cancer had been further clarified.
(3) Seminar on “Multifactorial Etiology and Multistep Development of Hepatocellular Carcinoma”
This seminar was held on March 21 and 22, 1990 at the Coco Palms Resort, Lihue, Kauai, Hawaii. The organizers were Dr. Setsuo Hirohashi, National Cancer Center Research Institute, Tokyo, Japan and Dr. Dennis P.H. Hsieh, University of California, Davis, California, U.S.A. There were seven speakers from Japan and eight from the United States. The purpose of the seminar was to exchange recent findings and views on the multifactorial etiology of hepatocellular carcinoma and the stepwise development of the disease with emphasis on aflatoxin, hepatitis B virus, and related agents as etiological factors. This seminar was immediately preceded by one on “Recent Advances in Research on Heterocyclic Amines” organized by Drs. Minako Nagao and Snorri Thorgeirsson. The participants of each seminar were invited to attend the other seminar, resulting in most of the two groups attending both seminars. This arrangement was especially desirable and beneficial to the participants because the two seminars complemented each other very well.
Multifactorial Etiology of Human Hepatocellular Carcinoma
To set a stage for the seminar, Dennis P.H. Hsieh (University of California at Davis, Davis, California) led off with a presentation on “Aflatoxin and the Multifactorial Etiology of Human Hepatocellular Carcinoma (HCC)” by using the prototype hepatocarcinogen aflatoxin B1 (AFB1) as an example to illustrate the multifactorial etiology of HCC. Aflatoxin is a family of carcinogenic mycotoxins and their animal metabolites. Th9 parent compounds are produced by toxigenic strains of Aspergillus flavus and Aspergillus parasiticus which can invade and grow in a wide variety of food commodities such as corn, peanuts, and cottonseed. The word “aflatoxin” is usually represented by aflatoxin B1 (AFB1) which is the principal and also the most potent member of this family of foodborne carcinogens. Because of the widespread occurrence of AFB1 in food commodities, human exposure to this naturally occurring carcinogen is difficult to avoid. Since its discovery in the early 1960’s, AFB1 has been known as an extremely potent hepatocarcinogen in the rat, rainbow trout, and other laboratory animals. It has been associated with the high incidence of primary HCC in some populations in African and Southeast Asian countries, based on the results of a number of epidemiological studies conducted in Kenya, Mozambique, Swaziland, Transkei, and Thailand. However, due to uncertainties in risk and exposure data and the possible involvement of hepatitis B viral infection as an etiological factor, aflatoxin was classified by the International Agency for Research on Cancer (IARC) as a “probable human carcinogen” in its reviews in 1972, 1976, and 1982. Additional data on the correlation of incidence of HCC and occurrence of aflatoxin in food commodities finally enabled IARC to conclude that aflatoxin is a human carcinogen in 1987. This IARC decision has been challenged by some investigators based on 1) the high uncertainties in the dose and response databases, 2) the wide species differences in the sensitivity to aflatoxin carcinogenicity, and 3) the strong evidence that hepatitis B virus (HBV) is a causative agent of HCC. The uncertainty about the involvement of aflatoxin in HCC is supported by the observations that (1) there is no correlation between HCC incidence and exposure to aflatoxin found in the U.S.A., Hong Kong, Taiwan, and some regions in China; (2) in regions of high exposure only subsets of populations are at high risk, and (3) aflatoxin is not found in the diet in regions of Alaska where HCC incidence is high. The argument that HBV is the causative agent of HCC is supported by the following facts: (1) the odds ratio for HCC in hepatitis B surface antigen (HBsAg) carriers in Taiwan is as high as 223 (compared to 10 for lung cancer in cigarette smokers in the U.S.); (2) in areas of high HCC incidence, 80% of the disease cases involve chronic HBV infection; and (3) induction of HCC by HBV-like viruses has been demonstrated in woodchucks, Peking ducks, and treeshrews as animal models. The wide species difference in animal sensitivity to aflatoxin carcinogenicity has long been demonstrated by carcinogenicity bioassays in different animal species and by the difference in the activity of in vitro activation of AFB1 to a mutagenic metabolite by liver samples of humans and other animals. In 1989, the published result of a cohort study conducted in Guangxi, China strongly supported the aflatoxin theory of HCC. In this study, four areas of different HCC incidences were chosen to compare correlations between the disease and HBV infection or aflatoxin exposure as a risk factor. While a very strong linear dose-response relationship was observed between HCC incidence and aflatoxin exposure, no correlation was found between the disease and HBV infection. Based on this and the previous epidemiological studies mentioned above, a TD50 was estimated to be 0.005 mg/kg/day and 0.132 mg/kg/day, respectively, corresponding to a virtually safe dose of 0.264 ng/kg/day and 0.001 ng/kg/day, respectively, at the 10-6 risk level. A recent toxicological study has revealed that although human liver does not activate AFB1 as efficiently as rat liver, it contains very little glutathione S transferase as compared to rat liver to remove the active form of AFB1, indicating that humans may not be less sensitive to aflatoxin carcinogenicity than rats. Also, aflatoxin-DNA adducts have been consistently found in the liver samples of HCC patients in Czechoslovakia, Taiwan, and Japan. The currently accepted multistep theory of HCC development suggests that both initiators and promoters are required for development of the disease. Exposure to genotoxic agents such as aflatoxin, heterocyclic amines, pyrrolizidine alkaloids, and nitrosamines will increase the mutagen load in the human body and will increase the probability of initiation of carcinogenesis. Similarly, exposure to promoters such as aflatoxin, hepatitis viruses, and other hepatotoxins will expedite the transformation of the initiated and altered cells to cancerous cells. Clearly, HCC is not caused by any single agent, but is the result of insults by different combinations of multiple factors. Accurate identification of critical risk factors is important in cancer prevention in that the return of investment can only be maximized by controlling the right factors.
Keiji Wakabayashi (National Cancer Center Research Institute, Tokyo) reported on the “Quantitation of Aflatoxins and Heterocyclic Amines in Various Kinds of Foods” to assess the exposure of Japanese populations to these two classes of hepatocarcinogens. Aflatoxins in food samples were extracted with chloroform-water and determined by TLC and HPLC. From April 1986 to March 1989, more than 2000 samples of cereals, treenuts, spices, and other commodities obtained from market places in Tokyo were analyzed. The frequency and concentrations of contamination were both rather low except for a small number of imported pistachios and spices. No aflatoxins were detected in samples of food produced in Japan. An average aflatoxin B1 intake rate of 0.1 g/day/person was estimated for people in Tokyo based on the annual consumption rates of the contaminated commodities. This aflatoxin intake rate is 20,000-fold less than that in the Inhambane Province, Mozambique, even though the incidences of HCC in these two areas are almost comparable. This discrepancy suggests that some etiological factors other than aflatoxins are responsible for the high HCC incidence in Japan. Heterocyclic amines (HCA) were quantified by HPLC with electrochemical or fluorometric detectors. The amounts of 9 HCA were determined in various kinds of cooked meat and fish. HCA were also determined from the urine samples of 10 healthy volunteers and two patients receiving parenteral alimentation. The urinary excretion of HCA was 0.03 - 47.1 ng/24 hr for healthy subjects, but HCA were not detected in urine from these patients. Based on the analytical data, the ratios of TD50 to daily intake of aflatoxin B1, MeIQx (a heterocyclic amine), and N-nitrosodimethylamine were estimated. The respective values indicate that exposure to these three well-known hepatocarcinogens cannot sufficiently explain the high incidence of HCC in Tokyo. DNA adducts of MeIQx were determined using 32P-postlabeling techniques in the liver of rats given diets containing MeIQx at the concentration range of 0.4 to 400 ppm for 1 to 12 weeks. The detection of MeIQx-DNA adducts indicated that MeIQx has an initiating activity with no threshold. Cigarette smoking is another source of HCA exposure. Airborne HCA in cooking places were not monitored, but it is anticipated to be an occupational safety problem. Mycotoxins other than aflatoxins were not surveyed in this study, but the contribution of mycotoxins to HCC in Japan does not seem important. The results of the present study clearly indicate that other etiological factors are involved in HCC in addition to exposure to aflatoxins and HCA.
Epidemiology
T. Colin Campbell (Cornell University, Ithaca, New York) discussed “Dietary Protein, Aflatoxin, and HBV in Human Liver Cancer” based on previous and current large-scale epidemiological studies in China. The conclusion by IARC that there is sufficient evidence to classify aflatoxin as a human carcinogen is largely based on five epidemiological studies conducted in Asia and Africa involving 17 survey sites (reviewed by Van Rensburg, 1985) and one study in China involving 4 survey sites (Yeh et al., 1986). However, an ongoing comprehensive study covering 48 counties in China did not show an aflatoxin-HCC association. In the survey conducted in 1983 to 1984 (Chen et al., 1990), mortality rates of cancer of seven sites, including HCC, were determined. Exposure assessment for a large number of variables was performed for two villages in each county. Four methods were used: (1) analysis of food samples randomly collected in each county, (2) a 3-day dietary survey of households, (3) questionnaires completed by 50 people (half of each sex) per village between the ages of 35 to 64, and (4) analysis of blood and urine samples from representative donors. The parameters of interest included nutrients, antioxidants, viral antigens, hormone status, aflatoxin metabolites, heavy metals, pesticides, nitrosamine markers, smoking characteristics, and other dietary and lifestyle factors. Special attention was directed to the possible relationships of HCC with the prevalence of HB surface antigen carriers, chemical contaminants in food samples, aflatoxin metabolites in urine samples, biochemical indicators of nutritional status in blood samples, consumption of alcohol, moldy products, corn, and a variety of other foods. Aflatoxin exposure (urinary excretion) ranged about 600-fold, HCC mortality rates 39-fold, and HBsAg + carrier prevalence 28-fold. HCC mortality was unrelated to aflatoxin intake (r= -0.17), but was positively correlated with HBsAg + prevalence (p), plasma cholesterol (p), frequency of liquor consumption (p), and mean daily intake of cadmium from foods of plant origin (p). Multiple regression analyses for various combinations of risk factors showed that aflatoxin exposure consistently remained unassociated with HCC mortality regardless of variable adjustment. Also, no association was found between HCC mortality and the proportion of people consuming corn or moldy peanuts. In contrast, association of HCC mortality with HBsAg +, plasma cholesterol, and cadmium intake remained, regardless of model specification, while the association with liquor consumption was markedly attenuated (insignificant) with adjustment for plasma cholesterol. The correlation between HCC and plasma cholesterol level suggests that fat and animal protein intake may be important contributing factors to the etiology of HCC. The discrepancy in the correlation of HCC and aflatoxin exposure between the rest four areas of the previous studies: (1) HBsAg + status was not considered in the previous studies, (2) estimates of aflatoxin intake were very crude, (3) HCC mortality rates may not be accurate, and (4) the number of surveyed sites and the size of populations analyzed were very limited.
Shoichiro Tsugane (National Cancer Center Research Institute, Tokyo) gave a general overview of the epidemiology of HCC in Japan. He pointed out that the number of liver cancer deaths in Japan was approximately 22,000 in 1988 (16,000 in males and 6,000 in females), which is the third leading cause of cancer deaths in males and the fifth in females. There is a distinct difference between males and females in the recent trend of age standardized mortality rates. The incidence in males has increased steadily since 1975, whereas the incidence in females has slightly decreased. Japan is among the countries in the world where the mortality rate due to liver cancer is relatively high. Within the uniracial population of Japan, there are geographic differences in this mortality rate. The western part of Japan, such as the Osaka and Kyushu areas, has higher SMR (standardized mortality ratio) in liver cancer. High mortality from liver cancer in the Kyushu area seems to be partially related to the high proportion of HBsAg carriers, but the mortality in Okinawa is very low, despite a high prevalence of HBsAg carriers. In two recent case-control studies in Osaka and Fukuoka, it was estimated that 22% of HCC patients are HBsAg carriers, 14% had a history of blood transfusion, and 22% to 33% are heavy alcohol drinkers. Another risk factor, non-A non-B hepatitis, was not analyzed in these studies. In the past, HBsAg was detected in the majority of HCC patients, but the proportion has been decreasing. The mortality rate of HBsAg-positive HCC has been stable during the last 5 years, therefore the recent increase of HCC in males has to be explained by other causes, such as alcohol consumption, infection by a non-A non-B hepatitis virus and/or other factors. A study was recently conducted of the epidemiologic background of 20 HCC male patients, aged 20 to 40, whose tumors were histologically examined at the National Cancer Center since 1975. HBsAg was found in 75% (15/20) of this group, a rate remarkably high even compared with HCC patients of 50 to 60 years of age. Alcohol consumption and history of blood transfusion did not seem to be related to the etiology in these patients. This result suggests that HBsAg-related HCC may develop in early life, whereas HBsAg unrelated HCC may have an older age of onset. A recent survey of the cancer mortality among Japanese immigrants in Sao Paulo, Brazil, indicated that liver cancer mortality in this population is as low as that of the general population in Sao Paulo, even among first generation immigrants. The same tendency was found in mortality from chronic liver diseases. The carrier rate of HBsAg was 1.6%, a value comparable to that seen in Japan, indicating that dietary and other environmental factors in Sao Paulo may act to prevent the development of chronic hepatitis and liver cancer, or that environmental factors in Japan may favor the development of chronic hepatitis or HCC.
Anne P. Lanier (Center for Disease Control, Anchorage, Alaska) reported on “Hepatitis B as an Etiologic Factor in Hepatocellular Carcinoma.” In Alaska, cancer incidence data are available for Alaska Natives since 1969. Among liver cancer patients, one-third are diagnosed in people under 30 years of age. There is a high degree of regional and community clustering in the distribution of liver cancer. Five multicase families accounted for nearly one-third of the cases seen in the last 20 years. Very few patients had cirrhosis. Chronic alcohol consumption does not seem to be a significant etiological factor. Results of tests for 12 different mycotoxins in samples of frequently consumed foods were negative. HBV on the other hand is a predominant factor associated with the disease. Eighty percent of HCC patients are HBsAg carriers, while only 3% of the overall Alaskan Native population are positive for HBsAg. Testing for the antibody of HCV in HBsAg negative liver cancer patients has given negative results. A prospective study of 1,400 HBsAg+ carriers followed for 7,815 carrier years disclosed an annual HCC incidence of 387 per 100,000 for men and 63 per 100,000 for women. Studies in Taiwan (annual HCC incidence of 495 per 100,000) and Alaska both show increasing incidence of HCC in HBsAg + carriers with increasing age. In the study in Alaska, the incidence rose from 103 per 100,000 in male carriers aged 0-9 to 1,190 in male carriers over 50. In multiple case families, HBsAg carriers are at 15-fold increased risk for developing liver cancer over that of HBsAg positive residents of the same region suggesting that other factors in addition to HBV are also important. Primary prevention of HBV infection through immunization was initiated in Alaska Natives in one high-risk area in 1981 and statewide in 1985. In addition, a secondary prevention program has been directed towards identification of early resectable tumors in HBsAg+ carriers. A prospective study since 1982 of 1,300 HBsAg+ carriers tested for alpha-fetoprotein (AFP) every six months has identified resectable cases of HCC in this cohort. Only one patient developed HCC with an AFP level less than the cutoff of 25 ng/ml.
Human Hepatocarcinogenesis
Tatsuo Miyamura (National Institute of Health, Tokyo) elaborated on “The Relationship between Hepatitis C Virus Infection and Hepatocellular Carcinoma.” For more than 15 years it has been speculated that the agent which causes non-A, non-B hepatitis is a small, enveloped RNA virus (hepatitis C virus, HCV). Recently cDNA of this virus has been cloned and, by expressing viral polypeptides in recombinant yeast, a simple and convenient assay system has been established to detect circulating antibody specific for the virus. Among 23 well-characterized non-A, non-B hepatitis patients surveyed, 15 of the 17 with chronic infection and 1 out of the 6 with acute infection were positive for anti-HCV antibody. Among the 15 who were anti-HCV antibody positive, 2 were carriers before blood transfusion, while 13 acquired the marker after blood transfusion. The presence of anti-HCV antibody in healthy blood donors (about 1.2%) suggests that the carrier rate of healthy individuals in Japan should be examined. A nurse attending HCC patients developed the disease after a needle accident and was found to be positive for anti-HCV, indicating that HCV is transmissible by needle sticks as well as by blood transfusion. In order to investigate the possible role of HCV infection in the development of HCC, a number of serum samples from HCC patients in Japan were assayed for the presence of the HCV antibody. A high prevalence of HCV antibody (70%) was detected among HCC patients who had no serum markers of HBV infection. The antibody was also shown to be present among patients with HCC originally thought to be due to very high alcohol consumption. He also presented the characterization of HCV cDNA fragments isolated from a single healthy Japanese HCV carrier whose blood had been retrospectively shown to contain infectious HCV by HCV antibody assay. Using these cDNAs, studies are in progress to analyze the relationship of HCV infection and the development of HCC at the molecular level. The existence of HCV genome in liver tissues of anti-HCV positive HCC patients was investigated by the sensitive reverse transcription/polymerase chain reaction method. Amplified cDNA sequences of HCV were detected in either cancerous or non-cancerous portions of liver tissues when the primers of nonstructural region 3 were used. These patients received blood transfusions over 20 years ago. These results suggest that persistent infection of HCV in liver has an important role in the development of HCC.
Boris H. Ruebner (University of California, Davis) gave a review on “Inherited Liver Diseases and Liver Cancer.” A relatively small proportion of HCCs develop in patients with a great variety of inherited diseases involving the liver. While most of these diseases appear to be associated with HCC, the degree of risk appears to vary greatly. The incidence of HCC in certain inherited diseases is associated with a wide variety of intrinsic and extrinsic factors. Hepatoblastoma is a tumor of infants associated with congenital malformations and genetic anomalies, such as familial polyposis. The tumor consists of embryonal and fetal type liver cells and possibly also of immature hepatic stem cells. Glycogen storage disease type I is associated with hepatocellular adenomas, but not the other types of glycogen storage disease. No cirrhosis is seen in this disease. In type IV there is a 100% incidence of cirrhosis, but no tumors have been reported so far. Idiopathic hemochromatosis and Wilson’s disease are both relatively common inherited disorders involving excessive storage of metals--iron in the case of hemochromatosis, copper in the case of Wilson's disease. While the incidence of HCC is high in the case of hemochromatosis, it is very low in Wilson’s disease, although a high incidence of cirrhosis is seen in both diseases. Galactosemia and tyrosinemia are autosomal recessive disorders of infants caused by deficiencies of different enzymes. Cirrhosis is seen in both diseases. In tyrosinemia, the incidence of HCC is 40% by the age of two years, while it is much lower in the case of galactosemia. Differences in the incidence of HCC are also found in the porphyrias and alpha-1-antitrypsin deficiency. Factors that may be responsible for these differences in the incidence of HCC include (1) chromosomal localization of the responsible gene, (2) severity of the associated liver disease (assessed as incidence of cirrhosis), and (3) possible association of environmental carcinogens. Based on the association of these liver diseases with HCC, one may conclude that (1) different liver diseases are associated with different types of liver tumors, (2) cirrhosis is not a prerequisite of HCC neoplasms, (3) not all the agents that stimulate liver cells to enter a cell cycle will cause HCC, and (4) mechanistic studies on the involvement of metabolic and genetic lesions are of great interest in the elucidation of the etiology of HCC.
Setsuo Hirohashi (National Cancer Center Research Institute, Tokyo) reviewed the “Mechanism of Multistep Development of Human HCC.” He indicated that it is now clear that most HCC develops in patients with chronic liver diseases, such as cirrhosis or chronic hepatitis, caused by infection of HBV or HCV. However, more than a decade is required for the development of clinically diagnosable HCC after viral infection. Little is known about what is happening in liver during this long period. Recently, small HCCs, that are small nodular lesions grossly and microscopically difficult to differentiate from HCC, have been detected with increasing frequency and have been surgically resected. By the morphological and genetic analyses of these lesions, it has become evident that many HCCs develop in a stepwise manner through clonal expansion and subclonal progression. In chronically injured livers, small nodular lesions are formed by the clonal expansion of hepatocytes. These lesions were classified as adenomatous hyperplasia (AH) and early HCC (eHCC). AH is a nodular lesion showing hypercellularity of hepatocytes without any discernible structural abnormality, and is considered to be a precancerous lesion. eHCC corresponds to in situ or microinvasive HCC with no or minimal destruction of the pre-existing lobular or psuedolobular liver structure. When the size of nodules exceeds 1.5 cm, smaller nodular areas of definite HCC with neovascularization and higher cell proliferative activity are often detected within the nodules. This morphologic appearance, named “nodule-in-nodule,” corresponds to the subclonal progression of HCC. Intrahepatic metastasis seems to develop from this stage. HCCs show further steps of progression and finally form tumors with the ability to metastasize to distant organs and kill the patients. By restriction fragment length polymorphism analysis, it was found that allele loss of chromosome 16 is frequent in poorly differentiated HCCs and HCCs of large size and with metastasis. It is likely that allele loss of chromosome 16 is a late event responsible for some stage of progression. It is clear that infection of hepatitis viruses, HBV and HCV, is the main initial event for human hepatocarcinogenesis at least in Japan. However, the multiple steps leading to clinical cancers cannot be explained by viral infection alone, and suggest a multifactorial etiology for HCC.
Richard H. Adamson (National Cancer Institute, Bethesda, Maryland) reported on “Hepatocarcinogenesis in Nonhuman Primates.” Nonhuman primates, though more expensive, have many advantages over rodents as an animal model for chemical carcinogenicirvy testing. Their long lifespan is conducive to evaluation at low doses. Their large size allows various diagnostic procedures to be performed. If raised from birth, they are easy to handle. They are susceptible to viral diseases similar to humans. Their metabolic activation of many foreign compounds is similar to humans and they have a very low spontaneous tumor incidence rate. At NCI, clean colonies of rhesus, cynomolgus, and African monkeys were bred so that their offspring could be used in carcinogenicity testing. The objectives of testing with nonhuman primates are to obtain comparative data on rodent carcinogens, to evaluate suspect human carcinogens, to develop markers for early diagnosis for human cancer, and to offer a better system for anti-tumor drug evaluation. Several compounds were evaluated in nonhuman primates for their potential to induce neoplasms, especially HCC. The compounds evaluated for their ability to induce HCC can be divided into three groups: food contaminants, model rodent carcinogens, and various nitrosamines. Of the food contaminants, aflatoxin B1 (AFB1), sterigmatocystin, methylazoxymethanol arcetate, and 2-amino-3-methylimidazo-(4,5-f)quinoline (IQ) induced HCC. None of the model rodent carcinogens tested consistently induced HCC in rhesus and cynomolgus monkeys. Three nitrosamines induced HCC and one nitrosamine, diethylnitrosamine (DEN), is a predictable and potent inducer of HCC and is useful for establishing a nonhuman primate model for numerous oncological studies. AFB1, a potent hepatotoxin and hepatocarcinogen in the rat, is a relatively weak hepatocarcinogen in monkeys, with an average latency period of 14 years. AFB1, however, induces cirrhosis of the liver, tumors in the gall bladder and pancreas, and angiosarcoma of liver and osteosarcoma. IQ has rather low hepatotoxicity, does not induce cirrhosis or toxic hepatitis, but it induces HCC quite effectively. DEN is very effective in inducing HCC and gives the most reproducible results. The average latency period for HCC produced by DEN in cynomolgus monkeys is 2 years at a total oral dose of 7 g per animal.
Jerome B. Zeldis (University of California, Davis) summarized the current understanding of “How Does Viral Hepatitis Cause Human Hepatocellular Carcinoma?” Malignant transformation is the result of complex processes in which factors that regulate cell growth become aberrant. Factors that promote cells to remain in the replicative state include mutations, deletions, and overexpression of various proto-oncogenes, inactivation of anti-oncogenes, activation of growth factors, demethylation of DNA and alteration of gene regulation. The adult hepatocytes usually are in Go. They are stimulated to enter the cell cycle of replication when injury has occurred. Thus, disease processes that cause chronic injury to hepatocytes are associated with continued hepatocyte growth and the potential for malignant cytogenetic changes. Chronic infections with HBV and HCV viruses are associated with increased risk of HCC. The inflammatory response to continued viral infection releases factors that are mutagenic in the Ames assay. Thus, the main reason chronic infections with HBV and HCV are associated with HCC may be that these agents promote hepatocytes to remain in the cell cycle. Equally intriguing are properties of hepadnaviruses that may relate to their role in hepatocarcinogenesis. HBV can integrate in any chromosome of the human genome. This is associated with deletions, duplications, and inversions of both human and viral genome. Often more than one chromosome will have an HBV integration. Once integrated, the modified viral DNA may act as a mobile genetic element and integrate in other parts of the human genome. The promoter for the envelope protein (HBsAg) is constitutively expressed. If this portion of the integrated virus is intact, the cell may secrete HBsAg or accumulate large (Pre-S1, Pre-S2, s) HBsAg protein. Not only may the expression of viral proteins serve as a stimulus for continued immunological injury, but accumulation of large HBsAg protein may serve as a stimulus for mit0~es resulting in malignant cytogenetic changes. Transgenic mice transformed with the large HBsAg protein will eventually develop HCC, probably by the latter mechanism. The multiple viral integration sites in the human genome can cause a number of potentially malignant changes to occur. The HBV integrations can serve as sites for illegitimate recombination between chromosomes. This may result in apposition or deletion of cell growth regulatory elements. Furthermore, HBV contains a number of cis and trans regulatory elements that can affect host gene transcription. The X protein and a protein produced by truncation of the pre-S2-s gene are both trans transcriptional activators. HBV also contains steroid-independent and steroid-responsive enhancer elements that affect gene expression. How all the above viral-induced effects synergize with those caused by chemical carcinogens in producing HCC is not yet known. An intriguing aspect of the problem is that the continued destruction of mature hepatocytes may stimulate oval cells (hepatocyte progenitors) to clonally repopulate the hepatic acinus. Malignant changes in the oval cells may lead to hepatocarcinogenesis.
Experimwntal Hepatocarcinogenesis and Mechanisms
Shoji Fukushima (Nagoya City University Medical School, Nagoya) reported on the results of their “Analysis of Enzyme Phenotype in Preneoplastic and Neoplastic Lesions of Rat Liver.” The concept of three-step carcinogenesis, consisting of initiation, promotion and progression in the rat liver, is now well accepted. During the stepwise hepatocarcinogenesis, foci of altered hepatocytes develop into hyperplastic nodules (neoplastic nodules) and hepatocellular carcinomas. Recently, much attention is being focused on changes in the expression of enzyme phenotypes of these lesions in the rat liver as influenced by exposure to promoting or inhibitory agents. In Experiment 1, male F344 rats were initially given a single dose (200 mg/kg) of diethylnitrosamine (DEN) intraperitoneally and 2 weeks later were treated with test compounds for 6 weeks; at week 3 they were subjected to two-thirds partial hepatectomy (PH). At week 8 the experiment was terminated. The compounds tested were 2-amino-3-methylimidazo(4,5-f)quinoline (IQ 0.05% in diet), phenobarbital (PB, 0.05% in diet) and butylated hydroxyanisole (BHA, 1% in diet). Markers analyzed were glutathione S-transferase placental form (GST-P), glucose-6-phosphate dehydrogenase (G6PD), adenosine triphosphatase (ATPase) and gamma-glutamyltranspeptidase (gamma-GT). A second set of rats which received D and PH only were used as control f EN s or comparison. Quantitative values (number/cm2) of all different enzyme-altered foci of the liver were increased in IQ and PB groups, when compared to the control group. The numbers of GST-P, G6PD- and ATPase-altered foci, but not gamma-GT foci, were greater in the IQ group than in the PB group. BHA clearly inhibited the induction of these phenotypes. Integration of enzyme-altered foci indicated that the number of foci decreased as the number of integrated enzyme alterations increased. The foci with 2 simultaneously altered enzymes occurred frequently in all treated groups. Most of the foci in the control group contained only one altered enzyme In Experiment 2, hyperplastic nodules and hepatocellular carcinomas (HCCs) were induced in male F344 rats by administration of DEN followed by 0.01% of 2-acetylaminofluorene (30-58 week). The rats were subjected to two-thirds partial hepatectomy at week 3. Another group was treated for 4 weeks with 0.05% N-ethyl-N-hydroxyethylnitrosamine. The experiment was terminated at week 60. The liver lesions were examined for the four marker enzymes mentioned above and glucose-6-phosphatase (G6Pase). The percentage of hyperplastic nodules increased with the number of enzymes altered in concert. For example, 55% had all five enzymes altered, 35% had 4 enzymes altered, and only 4% had only one enzyme altered. In HCCs, the percentages of these catagories were similar to those of hyperplastic nodules, being 50%, 30%, and 6%, respectively. In addition, labeling indices of bromodeoxyuridine (BrdU) were increased with the number of different enzymes in concert in HCCs, but not in hyperplastic nodules. Metastatic foci of HCC in lungs and subcutaneous transplanted tumors of HCC in nude rats were found to have all five enzyme phenotypes and high values of BrdU labeling indices. Alpha-fetoprotein (AFP), as shown by immunohistochemistry, seems to be another useful marker of malignant change in hepatocarcinogenesis. This marker was not found in the control group, but it was found in 38 to 44% of altered enzyme foci and hyperplastic nodules and in 93% of HCCs. Cell to cell communication, as shown by the presence of gap junction protein 32, also played an important role in malignant change. The staining intensity for this gap junction protein was strong in liver tissue from control rats, weak in foci and nodules, and negative in HCCs and transplanted tumors. These results indicate that, in rats, an increase in integrated enzyme alteration is associated with neoplastic and preneoplastic progression of the lesions Modifying agents may cause an increase or decrease in the induction of enzyme-altered lesions. GST-P appears to be the most reliable enzyme marker of phenotypic expression. Elevated Brd. U incorporation, appearance of AFP, and inhibition of cell to cell communrcatron are all induced during the progression of hepatocarcinogenesis.
Snorri S. Thorgeirsson (National Cancer Institute, Bethesda, Maryland) reported on “Cellular Molecular Aspects of Chemical Hepatocarcinogenesis.” Currently, there are a number of in vivo models in which the multistage process of carcinogenesis can be studied. Chemically induced hepatocarcinogenesis in the rat represents one of the most extensively used systems to study both cellular and molecular aspects of neoplastic development. One of the earliest cellular responses in chemical hepatocarcinogenesis is the appearance of small basophilic cells commonly referred to as oval cells. The fate of these oval cells, particularly as it relates to the precursor-product relationship to normal hepatocytes and to the histogenesis of hepatocellular carcinoma, has been a matter of considerable controversy. His group has recently shown that the oval cells can, under certain conditions, become precursor cells for hepatocytes, and thereby define a facultative stem cell compartment in the liver. Using the modified Solt and Farber protocol, in which the initiating agent DEN was omitted and the replication of normal hepatocytes was blocked by 2-AAF, partial hepatectomy induced very active proliferation of oval cells in the periductal and portal space. TGF-beta1 and other growth factors may be involved in the commitment of oval cells to differentiation into hepatocytes, bile duct and epithelial cells, or hepatic neoplastic cells. Cell lines which are the progeny of oval cells were very sensitive to growth inhibition by TGF-beta1; but if these cells were transformed by AFB1, or oncogenes, they lost this sensitivity. Using a complete Solt and Farber protocol, in which both regeneration of liver and clonal expansion of initiated cells are stimulated, TGF-beta1 was found only in the early stage of oval cell development. The cells then became TGF-beta1-negative and started to produce OV6, AFP, albumin, and GST-P markers. During the entire carcinogenesis process, TGF-beta1 is highly expressed in nonparenchymal cells, latent TGF-beta1 in nodular state and mature TGF-beta1 in full-blown carcinoma state, but TGF-beta1 is not found in nodules nor hepatocarcinomas. The role of TGF-beta1, if any, in the development of hepatocarcinomas therefore appeared to be a paracrine, rather than autocrine phenomenon.
Michio Mori (Sapporo Medical College, Sapporo) reported on the “Etiology of Hepatocellular Carcinoma in a New Mutant (LEC) Rat Developing Spontaneous Hepatitis.” In experimental hepatocarcinogenesis in rats, after the cessation of expodure to carcinogens or promoters, about half of the initiated cell populations would lose their preneoplastic phenotypical markers. In order to study the fate of the reverting initiated cells, a genotypical marker was introduced for detection. F1 hybrid rats of albumin-less NAR rats and Sprague-Dawley rats were raised; these hybrid rats are capable of producing albumin in hepatocytes and accepting albuminless hepatocytes with minimal rejection. Initiated hepatocyte populations from NAR rats were developed by the Solt-Farber protocol and were transfused into the F1 hybrid rats. The receptor F1 rats were subjected to PH and AAF treatment for selective growth of the initiated NAR hepatocytes. Albumin immunohistochemistry indicated that two weeks after transplantation, 99.6% of albumin-negative cell colonies are positive in at least one preneoplastic marker. Only initiated cells could proliferate under the selection pressure of the host liver. Two months after transplantation, half of the albumin-negative colonies became free of phenotypical preneoplastic markers while remaining as monoclonal colonies. HCC often develops in association with chronic liver diseases such as chronic hepatitis and liver cirrhosis. The LEC rat is a new mutant strain established recently which suffers from spontaneous hepatitis inherited by an autosomal recessive gene. The LEC rats therefore provide an animal model useful for the analysis of the role of chronic liver diseases in the development of hepatocellular carcinoma. Spontaneous hepatitis manifests itself suddenly in adult LEC rats 3 to 4 months after birth and about 40% of the rats die of hepatic failure within a week after the onset of hepatitis. Glutathione S-transferase-positive hepatocellular foci appear in the livers of remaining rats at about 5 months after birth. Chronic hepatitis continues thereafter and HCC appears at around I year of age. In an attempt to learn whether LEC rats are sensitive to carcinogens, a comparison was made between LEC and LEA rats (hepatitis free controls) for the numbers of induced glutathione S-transferase-positive foci in the liver using a modified procedure of Solt and Farber’s protocol. Approximately 15 times more foci were found in LEC rat liver than LEA liver, indicating that the LEC rats are highly sensitive to chemical carcinogens. Comparison was also made between the proliferative activities of LEC rat hepatocytes before and after the onset of chronic hepatitis, to assess whether chronic hepatitis is responsible for the development of HCC. It was shown that hepatocytes of LEC rats after onset of chronic hepatitis did not respond to growth stimulation such as partial hepatectomy in vivo and EGF administration in vitro, whereas the hepatocytes before the onset of hepatitis responded well. These results suggest that endogenous proliferative stimuli produced as a consequence of continuous loss of hepatocytes due to chronic hepatitis may act as a potent promoter for the development of HCC in LEC rats.
Gang-Hong Lee (Cancer Institute, Tokyo) described “Multi-stage Carcinogenesis in Mouse Liver.” Three topics on hepatocarcinogenesis in the mouse system were presented: The first topic dealt with interstrain differences in the sensitivity to initiation and promotion stimuli. C3H/He mice are known for their high incidence of spontaneous hepatoma and C57BL/6 and BALB mice for their low incidence. The three mouse strains were used to compare their sensitivity to carcinogenic agents in a two-stage hepatocarcinogenesis protocol. Male mice were injected i.p. with diethylnitrosamine (DEN) after partial hepatectomy and then fed either basal diet or diet containing phenobarbital (PB), clofibrate (CF) or ethynylestradiol (EE) for 20 weeks. In the groups receiving DEN alone, the number and size of enzyme-altered islands (EAIS) in the livers of C3H/He were far larger than in the livers of other strains. Under promoting pressure of PB, the growth of EAIs in C3H/He and BALB/cA was accelerated markedly, but those in C57BL/6 were only slightly accelerated. Interestingly, in BALB/cA the number of EAIS was much smaller than in C3H/He in spite of their good sensitivity to PB, suggesting that BALB/cA was refractory to the initiation process by DEN. The promoting effect of CF could only be demonstrated in C3H/He, while EE inhibited the development of EAIS in all the strains. These results indicate that there are interstrain differences in the two-stage hepatocarcinogenesis among mice with different genetic backgrounds, in either initiation or promotion, or in both processes. These differences may largely depend on the strain-specific nature of the hepatocyte itself rather than strain-specific environmental factors surrounding the hepatocytes, because in C3H/He-C57BL/6 chimeras C3H/He hepatocytes were selectively initiated by DEN. The second topic concerned spontaneous carcinogenesis in the culture of “normal” C3H/He hepatocytes. Using primary cultures of mouse hepatocytes, it was demonstrated that colony-forming, immortal epithelial cells (MLECs) existed in young adult C3H/He mouse livers. These cells were shown to possess the capability of albumin and/or alpha-fetoprotein production. Their colony formation activity was enhanced by the addition of PB to the medium. Since colony formation such as that seen with C3H/He cells was rarely seen with cells from C57BL/6, the MLECs were presumed to represent precursors for spontaneous HCC. More than 10 clonal MLEC lines were established from C3H/He mouse livers to examine the evidence of ras activation. No evidence was seen in any of these lines. However, transfection of an activated human c-H-ras oncogene conferred tumorigenicity on originally non-tumorigenic MLEC lines, suggesting that ras activation can cause malignant progression in precancerous liver cells in vivo. The third topic concerned the multi-step aspect of hepatocarcinogenesis in the SV40 T antigen-introduced transgenic mouse. A transgenic mouse line harboring SV40 T antigen gene conjugated with rat albumin promoter, established by Aizawa’s group, was used in the study of ht patocarcinogenesis. In this mouse the T antigen gene is expressed specifically in hepatocytes from the late embryonic stage onwards and all the animals die approximately 6 months after birth with multiple hepatocellular tumors. Although this finding strongly suggested that the T antigen may be a causal agent for hepatocellular transformation in vivo, a sequential histologic study revealed multiple processes and eventually only a small proportion of the hepatocytes underwent neoplastic alteration. This clearly indicates that even in this model, multiple events other than T antigen expression are involved in the tumorigenesis. One such event may be mutational activation of a cellular oncogene, since 40% of hepatocellular carcinomas developing in the transgenic mice contained an activating point mutation within codon 61 of c-H-ras.
George Bailey (Oregon State University, Corvallis, Oregon) reported on “Hepatocarcinogenesis and its Dietary Modulation in the Sensitive Trout Model.” The rainbow trout has long been recognized as a highly sensitive animal model for aflatoxin-induced HCC. One basis for this sensitivity is its small body size and rapid organ growth during initiation. Thus, a single microinjection of 0.4 ng of AFB1 in 1 µl of vehicle into trout embryos can induce 30% tumor risk nine months later. A biochemical basis for sensitivity also exists: trout have been shown to efficiently convert AFB1 to its activated 8,9-epoxide, to lack efficient glutathione transferase detoxication of this molecule, and to be high]y deficient at excision repair of the 8,9-dihydro-8-(N7-guanyl)-9-hydroxyaflatoxin B1 adduct These deficiencies may not exist with respect to other kinds of carcinogens. Trout and rats reportedly show equal tumorigenic response per unit liver AFB1-DNA adduct present during carcinogen treatment, which argues that the special sensitivity of trout is restricted to the initiation phase of AFB1 carcinogenesis. For support of this, DNA from AFB1-initiated liver tumors in trout, like the rat, have been shown to contain sequences capable of transforming NIH/3T3 cells to the oncogenic phenotype. Partial nucleotide sequences of two expressed trout ras proto-oncogenes have been determined to be highly homologous to the mammalian genes. As in the rat model, one of the trout ras genes was shown to carry activating point mutations in codon 12 or 13 guanines in a high preparation of AFB1-initiated hepatic tumors. These results demonstrate a molecular signature for AFB1 initiation that is highly conserved among vertebrates. It is therefore nnportant to compare human liver tumor DNA from areas of high and low endemic AFB1 exposure for evidence of such ras mutations. Over three dozen mammalian carcinogens have been assessed in the trout model and about 75% are carcinogenic, including mycotoxins, nitroso-compounds, and polynuclear aromatic hydrocarbons. In the use of embryo models, the water solubility of the chernical is an important factor of consideration. Water soluble compounds such as heterocyclic amines tend to wash out of the embryos after injection, resulting in reduction of doses. Several tumor types have been described involving liver, kidney, stomach, swim bladder, and other organs, but hepatocellular carcinoma is the predominant response in the trout. The organoselectivity of mammalian carcinogens often differs in trout, in some cases because the analogous organ (e.g., mammary gland, lung, forestomach, colon) is absent. Current indications are that oval cells, not hepatocytes, are the target cells for trout hepatocellular carcinoma. Trout are susceptible to dietary modulation of liver cancer by a range of promoters (e.g., hormones, high protein) and inhibitors (e.g., flavones, indoles). Comparative mechanism studies in trout have strengthened the phylogenetic basis for extrapolation of animal studies to humans. Thus, current results provide a mechanism for inhibition of AFB1 carcinogenesis by indole-3-carbinol (I3C) that clearly should apply to man. The active anticarcinogenic forms is not parent I3C, but dimeric and trimeric I3C derivatives formed in the stomach. Tumor studies have also shown that this and other dietary factors can act either as inhibitors or promoters, depending on the protocol chosen. There is a clear need to develop approaches for quantifying potential promotional risk versus anticarcinogenic benefit for such ambivalent modulators, a challenge that the low-cost trout model is helping to meet. These statistically demanding studies, designed to determine TD50 and ED50, have required as many as 10,000 individuals per experiment, making cost a significant deterrent for rodent models. Current results indicate that I3C is able to inhibit DNA modification by AFB1 and tumorigenesis as well as to promote AFB1-induced hepatocarcinogenesis depending upon whether it is given b f e ore or after animals are exposed to the carcinogen. When AFB1 and I3C are given together chronically, the overall effect is a reduced tumor response. These results also suggest (but do not prove) that at the normal human consumption rate, this cruciferous vegetable component may provide a protective benefit rather than promotional risk to populations at high risk from daily AFB1 exposure.
(4) Seminar on “Molecular Mechanisms of Initiation, Promotion and Progression”
This seminar was held on March 28 and 29, 1990 at the Senri Hankyu Hotel in Osaka, Japan. The organizers of this seminar were Dr. William F. Benedict, Center for Biotechnology, Baylor College of Medicine, The Woodlands, Texas; Dr. Curtis C. Harris, Laboratory of Human Carcinogenesis, National Cancer Institute, Bethesda, Maryland; and Dr. Hirota Fujiki, Cancer Prevention Division, National Cancer Center Research Institute, Tokyo, Japan. Based on the friendly support of both the U.S. and Japanese scientists, this U.S.-Japan seminar had been planned to commemorate the late Prof. Takeo Kakunaga and was dedicated to him. Following the original dates of the U.S.-Japan seminar, “The Takeo Kakunaga Memorial Lectures” were held on March 30. Dr. Kumao Toyoshima, Institute of Medical Science, University of Tokyo, was in charge of the memorial portion. and had additionally invited several scientists from the United States and Europe, who were well acquainted with the late Dr. Kakunaga. In addition to the speakers, a large number of observers from the University of Osaka and other institutions attended the seminar. The purpose of the seminar was to discuss and exchange ideas and information on a molecular level concerning initiation, promotion and progression of cancer. The first day of the meeting was devoted to the topics of DNA damage, carcinogen biomonitoring, and mechanisms of action of tumor promoters. The second day was concerned with gene expression, oncogenes, phosphorylation in transformation, suppressor genes, chromosome deletion and gene alteration. On the third day, three main talks and two shorter talks were presented at 'The Takeo Kakunaga Memorial Lectures.” Dr. Richard H. Adamson, Division of Cancer Etiology, National Cancer Institute, Bethesda, Maryland, summarized the extended three day meeting in the concluding remarks.
The opening comments were delivered by Dr. Hirota Fujiki and Dr. William F. Benedict. Dr. Fujiki stressed that this U.S.-Japan seminar, different from the usual U.S.-Japan seminar, was sort of an alumni meeting of NCI, because many U.S. speakers had come from NCI and many Japanese who have been at NCI attended. Therefore, the seminar also provided the participants with the opportunity to communicate with each other and renew their friendships. Furthermore, this seminar provided a significant occasion for global discussion among the invited speakers, most of whom are very much involved in editorial review and the stimulating competition of the scientific journals, including those related to carcinogenesis. Dr. Benedict, as one of the U.S. organizers of the U.S.-Japan seminar, stated that he deeply appreciated Dr. Toyoshima's efforts to make the seminar a more extended (international) meeting under the same title, and that a large number of the observers were able to be welcomed. Dr. Takashi Sugimura, National Cancer Center, Tokyo, described in his opening remarks that an idea to formally plan “The Takeo Kakunaga Memorial Lectures” came both from U.S. and Japanese scientists. He mentioned Dr. Kakunaga’s scientific career in Japan and the United States, and his personal touch as scientist and friend. A reprint of the Editorial on “Takeo Kakunaga in Memorial” appeared in Molecular Carcinogenesis, 1: 213-215, 1989, and was attached to each of the seminar abstracts.
Dr. Stuart H. Yuspa (National Cancer Institute, Bethesda, MD) discussed the molecular mechanisms of initiation, tumor promotion and progression. In the model of chemically induced mouse skin tumors, multiple benign squamous papillomas precede the development of an occasional squamous cell carcinoma. The incidence of carcinomas can be substantially enhanced by treating papilloma-bearing mice with mutagens such as urethane, nitroquinoline-N-oxide, or cisplatinum, suggesting that a distinct genetic event is responsible for malignant conversion. Cells expressing a benign phenotype can be obtained by introducing the v-ras oncogene into primary epidermal cells or by culturing cells from benign tumors induced by chemical carcinogens in vivo. Transfection studies of these cells reveal that transforming constructs of the fos oncogene induce malignant conversion, whereas myc and adenovirus E1A do not. Malignant tumors induced by fos transfection do not express differentiation-specific epidermal markers and secrete transin and urokinase, proteases characteristic of malignant skin tumors. Introduction of v-ras and v-fos oncogenes into cultured normal epidermal cells is sufficient to produce the malignant phenotype. Alone the v-fos oncogene does not detectably alter the normal phenotype of recipient cells. These studies implicate a limited number of genetic changes is sufficient to produce squamous malignancies.
Dr. John B. Little (Harvard School of Public Health, Boston, Massachusetts) and coworkers are investigating the molecular changes associated with the induction of mutations by physical and chemical agents in autosomal genes in a human lymphoblastoid cell line. Most radiation-induced mutants showed large-scale genetic changes involving loss of the entire tk gene and frequently extending to other loci on chromosome 17. They are currently developing two systems to examine the role of genetic recombination between alleles in this process. Initially, they are analyzing recombination and gene conversion events which occur during reversion of tk-/- cell lines to tk +/-. The nature and location of the frameshift mutations in parental cells and revertant clones are determined by exon specific polymerase chain reaction (PCR) amplification and direct sequencing, while polymorphic markers are analyzed by Southern blotting. In preliminary studies, one revertant line appears to have become homozygous at either end of the tk gene, while retaining one heterozygous frameshift. This finding is consistent with the involvement of recombination in the reversion event. In other experiments, they are examining mitotic rearrangements at VNTR loci in mutant clones by Southern hybridization with a series of VNTR probes on different chromosomes. Preliminary results indicate the. generation of new alleles in at least one of 60 radiation-induced mutant clones, whereas no new alleles were observed in 100 spontaneous mutants. Dr. Little presented a hypothesis that genetic recombination may play an important role in radiation mutagenesis, and perhaps be an important early event in neoplastic transformation.
Dr. R. Colin Garner (University of York, Heslington, York, The United Kingdom) reviewed the status of human carcinogen biomonitoring at the DNA level. Recent technical developments have enabled laboratory scientists to start to examine humans for DNA damage in vitro after exposure to caicinogens at environmental concentrations. Using both immunological and 32P-postlabeling techniques, Dr. Garner has examined human tissue DNA obtained at surgery or autopsy for the presence of DNA adducts and has found with the nuclease P1 modification of the postlabeling method that lung, pancreas, cervix, bladder, liver, stomach but not breast contain DNA adducts. Adducts in the lung are cigarette smoking related and show a dose-response relationship. Cigarette smoking associated adducts were also found in the pancreas, bladder, cervix and stomach. In addition, cervical DNA also contained in addition non-smoking related adducts which appeared as discrete chromatographic spots which appeared to be common from individual to individual. Adduct immunoconcentration of lung DNA using an anti-benzo[a]pyrene diol epoxide DNA antibody and subsequent postlabeling revealed several discrete chromatographic spots which were localized in the area associated with polycyclic aromatic hydrocarbon DNA adducts. Dr. Garner discussed the implication of these findings for human carcinogen risk assessment.
Dr. Kenneth H. Kraemer (National Cancer Institute, Bethesda, Maryland) discussed molecular and clinical studies of xeroderma pigmentosum (XP). XP is a rare disorder with marked sun sensitivity and increased frequency of cancer of the sun-exposed portions of the skin and eyes. Cultured cells from XP patients are hypersensitive to the cytotoxic and mutagenic effects of ultraviolet (UV) radiation and have defective DNA repair. Studies with UV- treated plasmid shuttle vectors indicate that the XP cells introduce more mutations into the vectors than do normal cells but have an altered spectrum of base substitution mutations.
Treatment of XP patients with an oral retinoid (I3Cis retinoic acid) resulted in a large reduction in the frequency of new skin cancers. Despite marked toxicity, this study demonstrates the feasibility of cancer chemoprevention in humans.
Dr. Peter Cerutti (Swiss Institute for Experimental Cancer Research, Lausanne, Switzerland) discussed oxidant carcinogenesis. While oxidants are toxic, they also trigger pathophysiological reactions which resemble those induced by growth and differentiation factors. Since oxidants are ubiquitous, they may represent “natural” tumor promoters. For example, active oxygen causes S6-phosphorylation, the activation and translocation of protein kinase C and the transcriptional activation of growth competence genes in mouse epidermal cells. For the modulation of gene expression by oxidants and other clastogenic carcinogens, the post-translational poly ADP-ribosylation of chromosomal proteins plays a role. It establishes a unique link between DNA damage and epigenetic regulatory processes. Following exposure to the clastogenic carcinogens, active oxygen and N-methyl-N’-nitro-N-nitrosoguanidine (MNNG), ADPR-transferase served as major poly ADPR-acceptor in mouse epidermal cells. Inactivation of ADPR-transferase may prevent excessive poly ADP-ribosylation. Topoisomerase I served as minor acceptor. Its inactivation might retard DNA replication and allow more time for DNA repair. Poly ADP-ribosylation also occurred at histones and, therefore, may modulate local chromatin conformation. Active oxygen and MNNG induce the immediate early genes c-fos and c-myc. The participation of poly ADP-ribosylation in the induction mechanism is suggested since the inhibition of ADPR-transferase by benzamid suppressed the transcriptional induction of c-fos and the consecutive increase in Fos-protein. Poly ADP-ribosylation of transcription factors and structural chromosomal proteins may alter protein/protein and DNA/protein interactions which are regulating the expression of immediate early genes.
Dr. Hisao Ueyama (Shiga University of Medical Science, Ohtsu, Japan) first found a protein which is present in bovine serum and inhibits the activity of debromoaplysiatoxin, a tumor promoter of the aplysiatoxin class, but not that of aplysiatoxin. The protein with a molecular weight of 65 kDa interacts specifically with debromoaplysiatoxin, and results in inhibiting the activation of protein kinase C by debromoaplysiatoxin and the receptor binding of debromoaplysiatoxin to the phorbol ester receptors. Dr. Ueyama succeeded in characterizing the protein as 1-acid glycoprotein, the physiological role of which is unknown. The presence of a specific binding protein for debromoaplysiatoxin explained how debromoaplysiatoxin had weaker biological activity in cell-culture systems than aplysiatoxin, although a tumor promoting activity of debromoaplysiatoxin is as strong as that of aplysiatoxin in mouse skin.
Dr. Toshio Kuroki (University of Tokyo, Tokyo) identified the two target proteins, p34 and p40, of protein kinase C in tumors of mouse skin as well as in BALB/Mk-2 keratinocytes. The p40 was creatine phosphokinase B, the activity of which was increased by phosphorylation activated by protein kinase C. He thought that ATP homeostasis regulated by protein kinase C might be involved in tumor promotion by TPA In addition, he reported that mRNA of metallothionein, urokinase, osteopantin and TGF-!
!!were constitutively expressed in tumors on mouse skin, indicating that transcriptional control is altered during the process of tumor promotion. Dr. Hirota Fujiki (National Cancer Center Research Institute, Tokyo) presented the new mechanisms of action of the okadaic acid class tumor promoters, okadaic acid, dinophysistoxin-1 (35-methylokadaic acid) and calyculin A The inhibition of protein phosphatases 1 and 2A by the okadaic acid class compounds results in the apparent “activation” of protein kinases, which seems to be one of the essential processes of tumor promotion. The mechanism of action of okadaic acid was also shown to promote carcinogenesis in rat glandular stomach initiated with N-methyl-N’-nitro-N-nitrosoguanidine. Therefore, the okadaic acid pathway of tumor promotion through inhibition of protein phosphatases broke away from the traditional concept of tumor promotion by TPA, that is, organ specificity and tissue specificity. The same pathway was also indicated by hepatotoxic monocyclic polypeptides, microcystins and nodularin, which are thought to be liver tumor promoters. Dr. Fujiki presented the okadaic acid pathway as a general mechanism for tumor promotion in various organs.
Dr. Yoshiaki Ito (Institute for Virus Research, Kyoto University, Kyoto) reported the enhancing effect of TPA on polyomavirus DNA replication. Treatment of cells with TPA enhanced polyomavirus DNA replication by activating the function of the AP1 binding site, to which the heterodimeric transcriptional activating factor, consisting of the products of nuclear protooncogenes, c-jun and c-fos, binds. Enhancement of the DNA replication by TPA is not primarily due to an increase of large T antigen. Dr. Ito reported that overexpression of exogenously added c-jun and c-fos genes strongly activated polyomavirus DNA replication through the AP1 binding site. This is the first report showing the direct involvement of oncogene products in the regulation of virus DNA replication.
Dr. Peter K. Vogt (University of Southern California School of Medicine, Los Angeles, California) discussed the oncogenicity of jun, a major component of the AP1 transcription factor. Dimerization with fos is a prerequisite for DNA binding, and DNA binding in turn appears to be required for transcriptional activation. There are several distinct differences between the viral jun protein and its cellular counterpart derived from chicken cells. One of these differences involves an amino acid substitution that removes a serine phosphorylation site. Dr. Vogt has used a Rous sarcoma virus based retroviral expression vector to examine the oncogenic potential of viral jun, of cellular jun, of reciprocal recombinants between viral and cellular jun and of deletion mutants in the viral jun. The results of these studies show that viral jun is highly oncogenic even without the gag sequences, while cellular jun is marginally oncogenic. The structural feature responsible for the strong oncogenicity of viral jun is the amino terminal half of the molecule, possibly the 27 amino acid deletion which resides in the transcriptional activator region of jun and appears to remove a negative control element of transactivation.
Dr. Noriyuki Sagata (Laboratory of Molecular Genetics, Kurume University, Fukuoka, Japan) discussed studies done in collaboration with Dr. George Vande Woude on the regulation of the induction of meiotic maturation and maturation promoting factor (MPF) in oocytes. The induction of meiotic maturation is dependent on the activation of MPF, a universal autocatalytic activity that drives the transition from G2 through M phase. Two integral components of this regulatory activity are the p34cdc2 proteins and the cyclin proteins. They have found that the mos proto-oncogene product is necessary for the normal progression of both Xenopus and mouse oocytes through meiotic maturation. Recently, it has been shown that mos protein is an essential component of cytostatic factor (CSF), the activity responsible for stabilizing MPF (Sagata et al. Nature, in press). In Xenopus oocytes the induction of meiotic maturation and the activation of MPF by both insulin and progesterone are dependent on the presence of the mos gene product. However, injections of synthetic cyclin AXe RNA or MPF cytoplasmic extracts are both able to induce meiotic maturation in the absence of mos product. While the injection of synthetic mos RNA in fully grown Xenopus oocytes can result in the induction of meiotic maturation and the activation of MPF, this induction can be blocked by the cAMP-dependent protein kinase catalytic subunit.
Dr. Hidesaburo Hanafusa (The Rockefeller University, New York, New York) discussed the phosphorylation of cellular proteins as a critical step in transformation by the crk oncogene. The SH2 and SH3 domains of the nonreceptor tyrosine kinases were originally defined as sequences conserved among these kinases, in a region amino-terminal to the catalytic domain. More recently, sequences with the SH2 and SH3 motifs have been found in a variety of proteins, including phospholipase C, GAP (ras GTPase activator protein), alpha-spectrin and viral gag-crk protein. The last protein is the transforming protein encoded by avian sarcoma virus CT10. The gag-crk protein has no homology to any known catalytic domain of protein kinases, but its expression causes increased phosphorylation of tyrosine residues of several proteins in infected cells. Evidence has been obtained that the crk protein associates with an endogenous tyrosine kinase activity and with the major phosphotyrosine-containing proteins in CT10-infected cells. Mutational analysis indicated that (i) both the SH2 and SH3 blocks are required for the association with cellular proteins and for cell transformation, and (ii) increased cellular phosphotyrosine correlates with transformation. It appears that the crk oncogene modulates protein-tyrosine phosphorylation and transforms through a novel mechanism.
Dr. H. Yamasaki (International Agency for Research on Cancer (Lyon, France) discussed altered gene expression and function of gap junctions during carcinogenesis and its possible role in tumor promotion and tumor suppression. Alterations of gap-junctional intercellula communication (GJIC) of transformed cells in culture can be divided into two classes: (i) loss or decrease of GJIC among transformed cells (e.g;, transformed mouse epidermal cells rat liver cells), and (ii) loss of communication with surrounding normal counterparts (e.g., oncogene- or chemically-transformed fibroblasts, certain rat liver epithelial cell lines). In both cases, transformed cells reduce their communication with their surrounding normal cells. Expression of certain transfected oncogenes can also be suppressed by GJIC with surrounding normal cells, whereas oncogene expression itself does not appear to alter the homologous GJIC capacity. Results from the examination of connexin (gap junction protein) gene expression in liver tumors suggest that in the rat liver, connexin (connexin 32 gene expression is reduced in hyperplastic nodules as well as in carcinomas, whereas heart-type connexin (connexin 43) gene expression appears in human hepatocellular carcinomas. A recent study with mouse epidermal cell lines suggests that a cell adhesion molecule, i.e., E-cadherin, plays a critical role in the regulation of GJIC. These results suggest that GJIC can be regulated at various levels and this is reflected in various types of changes observed in tumor cells.
Dr. Shun’ichiro Taniguchi (Kyushu University, Fukuoka) reported the biological roles of fos and actin genes in relation to cancer metastasis. Transfer of v-fos gene into the src-transformed 3Y1 cell line augmented lung metastasis, mainly through an increase in invasiveness and motility, which was associated with increases in lysosomal enzyme activity, cathepsin L and in expression of cytoskeletal proteins. Dr. Taniguchi found a new type of actin (m) present in mouse B16 melanoma cells. This m cDNA was analyzed and cloned. Transfer of m cDNA into highly metastatic B16-F10 cells decreased the invasiveness and metastasis. He concluded that the m gene is different from known -actin.
Dr. Jun Yokota (National Cancer Center Research Institute, Tokyo) reported loss of heterozygosity on chromosomes. Most small-cell lung carcinoma are associated with simultaneous loss of heterozygosity on chromosomes 3, 13, and 17. The common regions of loss of heterozygosity on chromosomes 3, 13 and 17 were 3p14-p24.1, the RB locus on 13q14 and the p53 locus on 17p13, respectively. Therefore, it was supposed that the RB gene and the p53 gene are inactivated in small-cell lung carcinomas as the net result of the loss of heterozygosity on these two chromosomes. Dr. Yokota concluded that at least six genetic alterations are necessary to convert a normal cell into a small-cell lung carcinoma cell.
Dr. Manfred F. Rajewsky (West German Cancer Center, Essen, Federal Republic of Germany) discussed ethylnitrosourea (EtNU) induced tumorigenic conversion of neuroectodermal cells in the developing rat nervous system: DNA modification and repair, and early cell lineage-specific gene alterations. The selective induction of malignant tumors by EtNU in the developing brain and peripheral nervous system of the rat is well suited as a model for exploring the significance of specific gene alterations in defined cell lineages, in relation to different steps in the process of malignant transformation and tumor progression. The neurooncogenic effect of EtNU is strongly dependent on the developmental stage of the target cell system at the time of carcinogen exposure. The resulting tumors are almost exclusively of neuroectodermal origin, and the relative proportions of different tumor types and localizations vary with the developmental “window” chosen for the EtNU-pulse. Subsets of neural (precursor) cells exposed to EtNU at distinct stages of development can be isolated for molecular analyses by electronic cell sorting, using monoclonal antibodies (Mab) directed against lineage-specific cell surface antigens.
Dr. William F. Benedict (Baylor College of Medicine, The Woodlands, Texas) discussed the retinoblastoma (Rb) gene and its role as an initiation and/or progression factor in several human cancers. The Rb susceptibility gene has now been implicated in not only the development of retinoblastoma as well as second malignancies which arise in patients with the hereditary form of retinoblastoma, but also is an important factor in some of the most common human malignancies such as those of the breast and lung. It still remains a critical issue whether the loss of Rb function is causally related to tumor formation or whether it is instead involved in the progression of a given tumor. To distinguish between the role of Rb gene loss as an initiation versus a progression factor in each of these tumor types, Dr. Benedict is examining the presence or absence of the Rb protein at the single cell level in primary tumors and metastasls. For the initial studies, a purified high-affinity polyclonal antibody was utilized which was raised against a synthetic peptide corresponding to a portion of exon 10 of the deduced Rb cDNA Using this material, it is possible to routinely obtain strongly positive nuclear staining for normal and tumor cells which contain what is apparently a normal Rb protein. However, in malignant cells from the same tissue type which are known to be missing the Rb gene, no nuclear staining is observed. More recently, Dr. Benedict has been able to examine nuclear staining using primary frozen tissue. For example, in initial experiments, it was found that each primary retinoblastoma examined was missing the Rb protein in every cell. This was the expected result since Dr. Benedict and others believe that a loss of Rb function is the basis for retinoblastoma development and that this loss is an initiation event.
Dr. Curtis C. Harris (National Cancer Institute, Bethesda, Maryland) discussed oncogenes and tumor suppressor genes in human lung carcinogenesis. Six families of activated protooncogenes, ras, raf, fur, jun, neu and myc have so far been associated with human lung cancer. Human bronchial epithelial cells in vitro are being used to investigate the functional role of these specific oncogenes and growth regulatory genes in carcinogenesis and tumor progression. Transfection of the SV40 T antigen gene leads to nontumorigenic cell lines that have a nearly normal pathway of terminal squamous differentiation and can be transformed to malignant cells by transfected Ha-ras, N-ras or Ki-ras. The combination of transfected c-myc and c-raf-1 will also cause transformation of human bronchial epithelial cells to neoplastic cells that exhibit some phenotypic traits found in small cell carcinomas. These and other results indicate that proto-oncogenes dysregulate pathways of growth and differentiation of human bronchial epithelial cells and play an important role in human lung carcinogenesis. Allelic deletion and somatic cell hybrid analyses are being used to identify the chromosomal localization of putative tumor suppressor genes. Mutations in p53, a putative tumor suppressor gene on 17p, were frequently found in human lung cancers. Allelic sequence deletion analysis of chromosome 11 revealed two commonly deleted regions (11p13 and 11p15.5). Somatic cell hybrids between normal human bronchial epithelial cells and Hut292DM, a lung carcinoma cell line, had a finite lifespan in vitro and were nontumorigenic in athymic nude mice. These studies suggest that tumor suppressor genes may play a dominant role in lung carcinogenesis.
Two young Japanese scientists who had worked with the late Prof. Kakunaga at the Department of Oncogene Research, Research Institute for Microbial Diseases, Osaka University, presented their research briefly. Dr. Takashi Miwa reported on the present status of the study on the human actin gene family. At least six actin isoforms are identified in mammalian cells; four muscle types and two non-muscle types. Their structures, organization and evolutionary origins were discussed. Dr. Jun Miyoshi established the SHOK cell line from Syrian hamster embryonic cells as indicator cells for detecting cellular oncogenes previously overlooked in the DNA transfection assay using NIH 3T3 cells. The new transforming gene, designated the cot gene, was isolated from a human thyroid carcinoma cell line. The cot gene product is a protein with 415 amino acids, similar to serine/threonine protein kinases. Dr. Miyoshi stressed that DNA transfection and focus-forming assay using SHOK cells would be an alternative method for isolating unknown oncogenes.
Dr. Kumao Toyoshima (University of Tokyo, Tokyo) reviewed the eight protooncogenes encoding cytoplasmic protein tyrosine kinases, which are closely related to p60Proto-src. Oncogenic activation of these genes might be caused by mutations, such as of the tyrosine residue at C-terminal position, at a defined position in the kinase domain, and in the modulatory domain. The expressions of the protooncogenes of the src-family are tissue specific and unique for each gene, suggesting the presence of a functional allotment for each gene product. For example, lck gene is associated with CD4 or CD8 antigens in T cells. Dr. Toyoshima emphasized that the functional allotments of lck, lyn and fun genes were significantly expressed in the hematopoietic system.
In closing remarks, Dr. Adamson noted the productive interactions between Japanese, European and U.S. scientists and reviewed the seminal role of the late Dr. Takao Kakunaga in the carcinogenesis research and his contribution to the U.S. - Japan seminar.
SEMINAR AGENDA AND PARTICIPANTS
(1) HUMAN CANCER GENES: GROWTH FACTOR AND TUMOR-SUPPRESSOR GENES
January 22-24, 1990 Kapaa, Kauai, Hawaii
AGENDA