REPORTS ON SEMINARS

(1) Seminar on "Nitric Oxide Synthase and Carcinogenesis"
A workshop on "Nitric Oxide Synthase and Carcinogenesis" was held in Williamsburg, Virginia on January 19 and 20, 1993 under the auspices of the U.S.-Japan Cooperative Cancer Research Program. The organizers were Dr. Hiroyasu Esumi of the Biochemistry Division, National Cancer Center Research Institute, Tokyo, Japan and Dr. Steven R. Tannenbaum, Department of Chemistry and Division of Toxicology, Massachusetts Institute of Technology, Cambridge, Massachusetts.
There were seven speakers from Japan and six speakers from the United States. The purpose of the meeting was to discuss and exchange information on the chemistry and biochemistry of nitric oxide and the Nitric Oxide Synthases (NOS) in relation to potential pathophysiological effects of nitric oxide and its reaction products. The opening remarks of Drs. Esumi and Tannenbaum emphasized the explosion in the literature on nitric oxide and its enormous physiological importance In December Science Magazine named nitric oxide as its "Molecule of the Year." Many biological effects are attributed to nitric oxide such as neural transmission, vasodilation, relaxation of smooth muscle and cytotoxic effects of macrophages and neutrophils. There are two classes of NOS constitutive and inducible Neurotransmission and vasodilation are conducted by the constitutive NOS but cytotoxicity of macrophages is conducted by the inducible NOS. In contrast to the constitutive NOS inducible NOS has been considered not to be regulated by calcium even though it has been shown to have an N-terminal calmodulin binding site.
Drs. Hiroyasu Esumi (National Cancer Center Research Institute, Tokyo) and Michael Marietta (University of Michigan, Ann Arbor) discussed isolation, purification and properties of NOS. NOS catalyzes the conversion of L-arginine to citrulline and NO• as shown below.

Dr. Marletta reviewed his work on the properties of the inducible mouse macrophage NOS. Co-substrates for the reaction include 0₂ and NADPH. In addition, there is a requirement for (6R)-tetrahydro-L-biopterin (H₂B). Two general classes of NOS have been purified, a constitutive form that is Ca²⁺ and calmodulin-dependent and an inducible form that is apparently independent ofCa²⁺ and calmodulin. NOS contains the coenzymes FAD and FMN and the amino acid sequence derived from the cDNA shows significant homology to NADPH-cytochrome P-450 reductase. The first step in the reaction yields N^G-hydroxyl-L-arginine. The further conversion of N^G-hydroxy-L-arginine to NO• and citrulline requires additional NADPH. NOS also contains a tightly bound H₄B that saturates at 1 mole H₄B per subunit. Evidence to date with pterin analogs shows that H₄B functions in a redox capacity during the reaction. Most recently NOS has been shown to contain a cytochrome P-450 type bound protoporphyrin IX heme that functions in catalysis. In addition, NOS oxidation of simple alkyl-substituted arginine analogs is consistent with P450 type catalysis. An overall picture of the NOS mechanism is emerging as well as further insight into the design of mechanism-based inhibitors, such as N^G-methylarginine.
Dr. Esumi's group bas identified and purified the inducible NOS from rat liver after treatment with Propionibacterium acnes and LPS. The enzyme has a strict requirement for its activity of calmodulin but not of calcium. Complementary DNA was cloned from the liver of rats treated as above. The longest open reading frame was 3441 bp encoding 1147 amino acids, being consistent with molecular weights of purified NOS. The cDNA was expressed in COS I cells and found to encode 125 kDa calmodulin dependent NOS, consistent with purified enzyme, Nucleotide sequence and deduced amino acid sequence were compared to those of mouse macrophage inducible NOS. Sequence conservations were 94% and 93% for amino acid and nucleotide levels, respectively, indicating identity to rat macrophage NOS. From structural analysis of the deduced amino acid sequence, putative binding sites for FMN, FAD and NADPH were found to be well conserved. Although the putative CaM binding site was not well conserved, hydrophobicity, the putative a-helix structure and surrounding basic amino acid cluster were conserved. Northern blot and western blot analyses revealed that LPS treatment induced NOS in various organs transcriptionally. Various cytokines including IFN-G, TNF, and IL-1 induce NOS in macrophages but this induction was abolished by cycloheximide, indicating newly synthesized protein is required for the induction of NOS mRNA.
Dr. Katsuei Shibuki (Laboratory for Neural Networks, Riken) presented his work on the function of NO• in the brain. Recently, involvement of NO• has been demonstrated in neural synaptic plasticity, which is responsible for learning and memory. In the cerebellum, simultaneous activation of two inputs (parallel and climbing fibers) evokes long-term depression (LTD) of the parallel fiber/Purkinje cell synapses. This LTD is the possible cellular mechanism for cerebellar motor learning. We have demonstrated that endogenous NO • release and the following cGMP production activated by climbing fiber stimulation are required for the induction of LTD. But the cellular consequence after the NO• cGMP signaling in LTD is not known. Marked NO•-induced cGMP elevation is observed in Bergmann glias which surround the parallel fiber/Purkinje cell synapses. The obvious NO• dependence of LTD in slice preparations is not observed in cultured Purkinje cells lacking peri-synaptic glias. These results raise the possibility that the consequence of NO•/cGMP signaling is to eliminate inherent glial suppression on neural plasticity. To test this hypothesis, rat cerebellar slices were treated with fluorocitrate (30 µM, min), a gliotoxic metabolic inhibitor. Although no marked changes were produced in synaptic responses of Purkinje cells following parallel or climbing fiber stimulation, NO• dependence of LTD was eliminated by the treatment: NO• was replaced by the fluorocitrate treatment in the induction of LTD. Another effect produced by the fluorocitrate treatment was facilitation of "giga-ohm sealing " between the membrane of Purkinje cell dendrites and the patch electrode used for intracellular recording. Similar facilitation of the giga sealing was evoked by sodium nitroprusside (1 mM). On the other hand, protamine (100 ng/ml), a basic heparin-binding protein, suppressed both the giga sealing and the induction of LTD. These results suggest the causal relationship between LTD and the membrane properties of Purkinje cells reflected in the giga sealing.
The development of drugs to control NOS and/or to trap NO• has become a key area of research. Dr. Hiroshi Maeda (Kumatomo University, School of Medicine, Kumatomo, Japan) presented his work on potential NO • traps. His group investigated the reactivity of stable radical compounds, imidazolineoxyl N-oxides such as 2-phenyl-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide (PTIO), carboxy-PTIO and carboxymethoxy-PTIO against NO• in both chemical and biological systems. By using electron spin resonance (ESR) spectroscopy, imidazolineoxyl N-oxides were found to react with NO• in a stoichiometric manuer (PTIO/NO• = 1-0) in a neutral solution (sodium phosphate buffer, pH 7.4) with rate constants of ~ 10⁴ M^-1 sec^-1, resulting in the generation of N0₂^-/NO₃^- and imidazolineoxyls such as 2-phenyl-4,4,5,5-tetramethylimidazoline-1-oxyl (PTI), or carboxy-/carboxymethoxy-PTI. Furthermore, the effects of imidazolineoxyl N-oxides on acetylcholine- or ATP-induced relaxation of the smooth muscle of rabbit aorta were tested. The vasorelaxations were markedly inhibited by all three imidazolineoxyl N-oxides. The inhibitory effects of carboxy-PTIO was almost twofold stronger than those of NO• synthesis inhibitors, N-nitro-L-arginine and N-monomethyl-L-arginine. Generation of EDRF/NO• - was identified by reacting the PTIO in aortic specimens and quantitating the reaction product with ESR spectroscopy. Thus, it was clarified that imidazolineoxyl N-oxides antagonize EDRF/NO• via a unique radical-radical reaction with NO•.
The remainder of the program dealt with the chemistry of NO• and its potential role in the pathogenesis of inflammation and cancer. Dr. Peter Ward (University of Michigan Medical School, Ann Arbor, Michigan) discussed his research on lung inflammatory injury. Deposition of either IgG or IgA immune complexes in rat lung results in acute injury to alveolar epithelial cells. In the case of IgG immune complexes, lung injury is both neutrophil and macrophage-dependent. Cytokines (TNF!!!and IL-1) are also involved in events leading to injury. Blocking of L-arginine by the intrapulmonary instillation of analogs of L-arginine (N^G-monomethyl-L-arginine, and N^G-nitro-L-arginine methyl ester) is highly protective. These protective effects can be monitored by reduced levels of N0₂^- and NO₃^- in the bronchoalveolar lavage fluids. The protective effects of the L-argiuine analogs are not associated with reduced recruitment of neutrophils into the alveolar compartment. While the source of products from L-arginine, such as NO • and ONOO-, is not known but may be phagocytic cells (e.g., pulmonary macrophages, neutrophils, etc.), the data indicate that metabolic products of L-arginine are tissue-toxic. In the case of IgA immune complex-induced lung injury, neutrophil recruitment does not occur and cytokine (TNF!!!and Il-1) production is not evident. Tissue damage appears to be related directly to toxic products from stimulated pulmonary macrophages. The intrapulmonary instillation of L-arginine analogs (of the type mentioned above) is also highly protective. These data indicate that, in neutrophil and/or macrophage-dependent lung injury, metabolic products of L-arginine are highly toxic to lung tissue.
Dr. William Pryor (Biodynamics Institute, Louisiana State University, Baton Rouge, Louisiana) discussed the chemistry of NO• and its reaction with O₂•- to form peroxynitrite (O₂NO^-). He reviewed earlier work on the oxidation of thiols by NO• and NO₂ and the inactivation of human antiproteinase inhibitor (a1PI). Inactivation is far more rapid with mixtures of H₂O₂ and NO₂ than with either agent alone. The inactivation involves the oxidation of a critical methionine residue. The reason for the more rapid oxidation by the combination of reagents is that H₂O₂ and nitrogen dioxide react to form peroxynitrate, HOONO₂, a potent oxidant. When a1PI is inactivated, a methionine residue at the active site is oxidized to methionine sulfoxide. The mechanism involved in the oxidation of a1PI by peroxynitrite does not involve free radicals. Rather, peroxynitrite appears to give HO•-like chemistry following protonation through two types of processes. Type 1 is unimolecular with E_a = 18 kcal/mole while type w is a bimolecular reaction with E_a of 8-10 kcal/mole. This suggests that peroxynitrite might diffuse a considerable distance before reacting with cellular nucleophiles. For this reason peroxynitrite is emerging as a major candidate for a key reactant related to NO•-induced cellular injury.
Inflammation is a vital consequence of the tissue injury that is inevitably associated with specific types of leukocyte infiltration. Dr. Kouji Matsushima (Kanazawa University, Kanazawa, Japan) spoke about his research on Interleukin 8 (IL 8), a novel heparin binding leukocyte chemotactic and activating factor. IL 8 is produced by various types of cells including monocytes/macrophages, dermal fibroblasts, keratinocytes, vascular endothelial cells, epithelial cells and mesangeal cells in response to cytokines (e.g., IL I and tumor necrosis factor), plant lectins, mitogens, viral and bacterial compounds (e.g., endotoxin, exotoxin, and transactivating proteins), heavy metals (e.g., mercury and cadmium), and superoxide. Molecular analysis of the enhancer region of the IL 8 gene has revealed that NF-IL 6(C/EBP) plus NF-KB or AP-1 plus NF-kB-like factor binding sites are essentially involved in the IL 8 gene activation. The production of IL 8 has been shown in various human inflammatory diseases such as rheumatoid arthritis, ulcerative colitis, Crohn's disease, psoriasis, neonatal infection, urinary tract infection and glomerulonephritis. Essential involvement of IL 8 has been established in causing neutrophil infiltration in experimental acute inflammation models in rabbits such as lipopolysaccharide-induced joint arthritis, skin inflammation, and lung reperfusion injury. On the other hand, tumors are often associated with leukocyte infiltration, and IL 8 and related cytokines produced at the tumor site may be responsible for leukocyte recruitment into the tumor site. Possible antitumor activity and anti-metastatic activity of IL 8 and IL 8 related molecules are under investigation.
The well-established relationship between infection with the liver fluke, Opistliorchis vivenini, and cholangiocarcinoma (CHCA), has recently attracted considerable attention for examination of the link between inflammation and cancer. Recent studies have shown a dramatic increase in risk (odds ratio 22.2) of CHCA among heavily infected individuals. Because of the chronic inflammation and high risk of cancer associated with this parasite, plus the ability to eliminate it with simple drug treatment, this is an ideal model to investigate the possible involvement of nitric oxide and N-nitrosation in carcinogenesis. Dr. Mitsuhiro Tsuda (National Institute of Hygienic Sciences, Tokyo, Japan) discussed the work of a scientific group including Thailand, Australia and Japan comparing endogenous production of nitric oxide (nitrate levels in plasma, saliva and urine) and nitrosation potential following proline loading of groups of men stratified by intensity of infection and individuals with varying hepatobiliary disease status. The 150 male subjects, aged 29-52, with no, moderate and heavy liver fluke infection (as determined by 2 egg count assessments and worm recovery following treatment) were drawn from endemic communities in Khon Kaen, Northeast Thailand. The subjects were also examined by ultrasound to assess hepatobiliary status. A second parallel study compares people with early cholangiocarcinoma with age-, sex-, and intensity-matched controls with normal and nonmalignant gall bladder disease as determined by ultrasonography. Urine, saliva and plasma samples were collected twice from the subjects, while they were maintained on low nitrate diet and refrained from smoking. The first sample collection was performed before curing of the infection by praziquantel treatment and ale second collection was performed 4 months post treatment. Levels of nitrate in the body fluid was determined by the Cd-reduction/Griess reagent method. The urinary N-nitrosamino acids, including N-nitrosoproline (NPRO), N-nitrosothiazolidine-4-carboxylic acid (N-nitrosothioproline, NTPRO) and 2-methyl derivatives of NTPRO were determined by gas chromatography-thermal energy analyzer (GC-TEA). Preliminary results indicate that endogenous formation of nitric oxide and NOC (measured by the levels of nitrate in the plasma, saliva and urine and by levels of N-nitrosamino acids in the urine) is higher among individuals with moderate and heavy infection with the parasites than those of uninfected controls and returned to normal 4 months post treatment. These results strongly suggest that endogenous production of nitrogen oxides is enhanced during liver fluke infection which may also contribute to an increase in the formation of NOC and hence perhaps be involved in cholangiocellular carcinogenesis.
Complementary studies were presented on an animal model of OV in the Syrian hamster by Dr. Katsumi Imaida (Nagoya City University Medical School, Nagoya, Japan). To test whether OV was a promoting agent, Syrian hamsters were initiated with a 20 mg/kg DMN i.p. injection 19 days prior to infection with 80 OV metacercaria. At week 45, 44% of the animals receiving both DMN and the parasite developed cholangiocarcinoma, while 93% had cholangiofibrosis and 4% demonstrated hepatocellular carcinomas. Animals receiving DMN alone developed no such lesions and only a 13% incidence of cholangiofibrosis was detected in the parasite group without any neoplastic lesion development. In animals infected with OV, fibrotic ductular proliferation was always observed and cystic bile ductules as well as acute and chronic proliferative cholangitis were prominent. These findings clearly revealed OV infection to enhance the development of cholangiocarcinomas in hamsters initiated with DMN presumably due to the associated chronic inflammation. The possibility of mechanical obstruction by OV as the main cause of its enhancement of cholangiocarcinogenesis in hamster was investigated. Syrian hamster livers were initiated with a single 20 mg/kg i.p. injection of DMN followed by complete ligation of the left extrahepatic bile duct 2 weeks subsequent to initiation. At the end of week 40, animals receiving DMN and ligation had a 60.9% incidence of cholangiofibrosis, 21.7% of mucous cystadenoma and 39. 1% of cholangiocarcinoma, whereas the group given DMN alone developed only a 5% incidence of cholangiofibrosis, with neither cystadenomas nor cholangiocarcinomas being observed in this group. No neoplastic lesions were observed in the ligation alone group. These results showed that bile duct proliferation resulting from ligation might play an important role in the promotion stage of cholangiocarcinogenesis.
Possible involvement of NOS activity in the enhancing effects of OV infection on hamster cholangiocarcinogenesis was examined. Syrian hamsters were infected with 50 or 100 metacercariae and at weeks 4 and 8, urine samples were collected over 24 hours and frozen, after adjustment of pH to 11 by NAOH, until analysis of NaNO₂ content. The measured levels of NaNO₂ were not significantly different between OV infected and control groups. At week 8, all hamsters were sacrificed and liver tissue was frozen for subsequent investigation of NOS activity by immunohistochemistry and in situ hybridization using an antibody to NOS and a cDNA, provided by Dr. H. Esumi, National Cancer Center Research Institute, Tokyo. No positive results have yet been gained concerning NOS activity in the hamster liver infected with OV.
These results clearly support the conclusion that inflection with OV is a very high risk factor for cholangiocarcinogenesis in both humans and rodents, but whether NOS contributes to the development of cholangiocellular lesions under the influence of fluke parasites remains to be elucidated.
Two speakers focused on the mutagenic properties of NO• and NO•-prodrugs. Dr. Steven Tannenbaum (Massachusetts Institute of Technology, Camibridge, Massachusetts) summarized some of the conditions under which NO• can exert its toxic effects:

1. NO• may diffuse into a target cell and interact with an iron prosthetic group. This may result in signalling as in GMP cyclase (positive effect) or in negative effects such as the quenching of a free-radical in ribonucleotide reductase or the destruction of a Fe-S center in aconitase or in the mitochondrial electron transport chain.
   
2. NO• may combine with O2 to form N2O3. Nucleophilic attack on N2O3 by S and N containing molecules will cause modification of essential sulfhydryl groups, formation of carcinogenic N-nitroso compounds, or diazotization of primary amines. When the amines are purines or pyrimidines, the result is modification of RNA or DNA bases via deamination or crosslinking. This will be discussed further.
   
3. NO• may combine with O2•- to form peroxynitrite anion.
   

The mutagenicity of NO• has been established in bacteria (2) and human cells (3). In tissues undergoing an inflammation reaction, both the infiltrating and resident cell populations produce a time-dependent mixture of nitrogen oxide radicals and oxygen radicals. These different radical species interact forming new radicals and oxygen radicals. These different radical species interact forming new reactive intermediates which may contribute to DNA damage. Although the flux of radicals per unit time is low, an inflammatory condition that continues for years becomes a significant risk factor for carcinogenic cell transformation. The relative contributions of various radical species to the carcinogenic process must be assessed through both chemical analysis of DNA and through genetic and molecular biological analysis of mutations in surviving exposed cells.
The MIT group has developed methods using GC-MS for the concurrent analysis of deamination and oxidative DNA damage. The deamination products of guanine and adenine measured are xanthin (X) and hypoxanthine (Hx). Oxidation of thymine is assayed via 5-hydroxymethyluracil (HU). Cultured muriune macrophages contain no detectable background for HU and a low background for X and Hx. Treatment with LPS causes elevation in all three modified bases, but addition of an NOS inhibitor, N^G-methylarginine brings the modifiied bases down to background levels. Thus, oxidative damage to DNA is potentiated by NOS via NO•. Similar experiments are being carried out in target cells.
In summary, the chemistry of DNA damage by NO^•, although potentially complex with respect to the overall mechanisms and the actual alterations of the nucleic acid, might arise from only two fundamental processes: reaction with oxygen to form nitrosating species and reaction with superoxide to form peroxynitrite. Chemical analysis of mutational spectra in key genes should test this hypothesis.
Dr. Larry Keefer (National Cancer Institute, Frederick Cancer Research and. Development Center, Frederick, Maryland) and his research group have launched a multifaceted research effort to assess the extent to which the genotoxic potential of nitric oxide (NO~ and its progenitors might contribute to cancer causation in people. Their efforts encompass chemical studies, experiments in bacteria and investigation in mammalian systems.
While "spontaneous" deaminauon of DNA is energetically quite unfavorable, the rate was vastly increased by exposure to NO• in aerobic buffers. When NO• was bubbled into a solution of 0.1 M phosphate containing 1-100 mM 2'-deoxycytidine until 1 mol of gas was absorbed per liter of buffer at pH 7.4, 2'-deoxyuridine (the deamination product of 2'-deoxycytidine) was detected to the extent of 3-5 altered bases per thousand. Similar results were seen for 2'-deoxyguanosine, -deoxyadenosine, and -deoxy-5-methylcytidine. as well as for the deoxynucleotides tested and for the cyiosine residues of calf thymus DNA.
To determine whether NO• could similarly damage DNA in the living cell, S. typhimurium TA 1535 was treated with DEA/NO and SPER/NO, two drugs that spontaneously release NO• on dissolution in physiological media and that have proven useful for the controlled biological delivery of NO•. Both were mutagenic. SPER/NO, a solid complex formed by reacting NO• with spermine, was the more potent. DNA from mutant colonies treated wuh SPER/NO was isolated and sequenced. Of 114 colonies analyzed, 113 contained C-->T transition mutations in the first or second position of the hisG46 (CCC) target codon.
Similar results were seen when this bacterial strain was exposed to the clinical antianginal agent, nitroglycerin, an important drug that is reductively converted to NO• as its active metabolite. The mutational spectrum for nitroglycerin was the same as that seen for SPER/NO. However, nitroglycerin proved inactive in six other strains of S. typhimurium (TA1975, TA102, TA1538, TA100, TA10ONR, and YG1026).
To learn how replication and repair pathways available in human cells might affect the mutagenic potential of NO •, thy exposed the shuttle vector, pSP189, which carries the E coli supF gene as a target for mutation, to gaseous NO• at pH 7.4. In a control study, the plasmid was separately treated with sodium nitrite at pH 7.4, 6.4, and 5.4. Treated plasmid was transfected into human Ad293 cells in which replication occurred. Plasmid was then extracted and transformed into the indicator E. coli strain, MLB 100. The mutational spectra seen when mutant plasmids were sequenced were very different for NO• vs. nitrite-treatment, with GC—>AT transitions being the predominant (44%) mutation seen in the nitrite-treated plasmid, while 60% of the lesions seen with nitric oxide were AT—>GC transitions. It is possible that this unexpected difference results from the production of oxidizing as well as nitrosating agents during autoxidation of NO•, while acidified nitrite is capable only of deamination. In a series of in vivo studies nitroglycerin, DEA/NO and SPER/NO are being tested for carcinogenicity.
Two speakers discussed the use of mutations in oncogenes as possible clues to the etiologies of human cancers. Carcinogenesis is now considered to be a multistage process accumulating multiple genetic alterations. Several mutations are often detected in one sample of clinical cancers. At this stage, it is very reasonable to ask whether or not a pattern of mutations in cancers reflects the etiological background.
Dr. Setsuo Hirohashi (National Cancer Center Research Institute, Tokyo, Japan) presented an analysis of a variety of types of human tumors. Human cancers can be roughly classified into two groups. In one group, cancer is developed in an otherwise normal organ. Sporadic adenoma and carcinoma of the colon, excluding those associated with ulcerative colitis, are surrounded by normal lookng colonic mucosa. Cholangiocarcinoma of Japanese patients are found in normal liver in most cases. In the other group, cancer is developed in a damaged organ, wbich shows chronic inflammation, cell death and regeneration. Gastric cancers, especially those of intestinal type, develop in mucosa affected by chronic atrophic gastritis. Chronic hepatitis and cirrhosis caused by infections of hepatitis viruses are seen in most cases of hepatocellular carcinoma. Furthermore chronic inflammation induced by infection of liver fluke (Opisthorchis viverrini) is the cause of a high incidence of cholangiocarcinoma in northeast Thailand. They examined the pattern of p53 mutations and some other genetic alterations in these cancers associated with and not associated with chronic inflammation. They also tried to analyze whether involvement of inflammation in the carcinogenic process confers types of mutations unique to cancer or not, but the result was not so simple. p53 mutations and other genetic alterations they detected in intestinal types of gastric cancers were very similar to those previously detected in colon cancers, although the etiological backgrounds of the two tumors were different. Mutations at CpG sites were very common in gastric cancer but rare in hepatocellular carcinomas, although those two tumors are considered to develop with a common basis of chronic inflammation.
When mutation of c-Ki-ras gene was analyzed, it was suggested that point mutation of c-Ki-ras is not involved in carcinogenesis in the presence of inflammation, e.g. cholangiocarcinoma associated with HBV or HCV inflection. Examples other than these tumors include carcinoma of the gallbladder, which is almost always associated with gallstones and chronic cholecystitis, and colon cancer associated with ulcerative colitis. On the other hand, point mutational activation of the c-Ki-ras gene seems to occur frequently in tumors that are not associated with inflammation. Examples of these tumors include cholangiocarcinoma in Japanese pancreatic adenocarcinoma and colorectal adenoma and adenocarcinoma. Activation of the c-Ki-ras gene is considered to be involved in the development of these cancer types.
According to Dr. Curtis C. Harris (National Cancer Institute, Bethesda, Maryland), mutations in the evolutionarily conserved codons of the p53 tumor suppressor gene are common in diverse types of human cancer. The p53 mutational spectrum differs among cancers of the colon, lung, esophagus, breast, liver, brain, reticuloendothelial tissues, and hemopoietic tissues. Transitions predominate in colon brain and lymphoid malignancies. Mutational hotspots at CpG dinucleotides in codons 175, 248, and 273 may reflect an endogenous mutagenic mechanism, i.e., deamination of 5-methylcyiosine to thymidine. The distribution of p53 mutations at these codons with CpG dinucleotides also varies among different tissue sites. Oxy-radicals may enhance the rate of deamination. G:C to T:A transversions are the most frequent substitutions observed in cancers of the lung, breast, stomach and liver, and are more likely to be due to bulky carcinogen-DNA adducts. p53 mutation and/or accumulation of posttranslationally modified p53 protein can be preinvasive events in bronchial, mammary, or esophageal carcinogenesis. Mutations at A:T base pairs are seen more frequently in esophageal than in other solid tumors. Most transversions in lung, breast, and esophageal carcinomas are dispersed among numerous conserved codons. In liver tumors in persons from geographic areas in which aflatoxin B₁ and hepatitis B virus are cancer risk factors, most mutations are at one nucleotide pair of codon 249. In geographic areas where hepatitis B and C viruses, but not aflatoxin B₁, are risk factors, the p53 mutations are distributed in numerous codons. One hypothesis concerning generation of liver cancers with this mutation is: (a) aflatoxin B₁ is metabolically activated to form the promutagenic N7dG adduct; and (b) enhanced cell proliferation due to chronic active viral hepatitis allows both fixation of the G to T transversion in codon 249 of p53 gene and selective clonal expansion of the cells containing this mutant p53 gene. The induction of skin carcinoma by ultraviolet light is indicated by the occurrence of p53 mutations at dipyrimidine sites including CC to IT double base changes. The p53 mutational spectrum in radon-associated lung cancer from uranium miners also differs from lung cancer caused by tobacco smoke alone. In summary, these differences in mutational spectrum among human cancer types suggest: (a) the etiological contributions of both exogenous and endogenous factors to human carcinogenesis (b) specific proliferative effects conferred by different mutant p53 genes in different human cell types; and (c) hypotheses for investigation.
Following the conclusion of the individual presentations there was an open discussion. All of the participants agreed that the meeting had been very productive and that there was a great deal of complementarity between the U.S. and Japanese research. New contacts were developed and old friendships were renewed in an atmosphere of international cooperation.

(2) Seminar on "Viral Regulatory Proteins and Host Factors Regulating Gene Expression"
This meeting was held January 26-27, 1993, at the Marriott Hotel in Maui Hawaii. The organizers were Dr. Peter Howley, National Cancer Institute, Bethesda, Maryland, and Dr. Mitsuaki Yoshida, University of Tokyo, Tokyo, Japan.
There were a total of nine participants from Japan and six participants from the United States. The purpose of the meeting was to discuss and exchange information on the rapidly moving field of viral regulatory proteins and host factors involved in the regulation of gene expression. The opening remarks by Drs. Peter Howley and Mitsuaki Yoshida indicated that this was the third in a series of U.S.-Japan seminars which have focused on viral regulatory proteins, both DNA and RNA viruses. The study of viral cellular regulatory proteins has provided important insights into the deregulation of cell cycle control observed in cancer.
The six-session program with two or three speakers per session began with Dr. Inder Verma (The Salk Institute, San Diego). NF-!!!B was originally identified as a nuclear protein complex that specifically bound to the enhancer element of the immunoglobulin!!!light chain gene. NF-!!!B is now recognized to be a member of the superfamily that includes the oncogene rel and the Drosophila gene dorsal. All of these proteins share extensive amino acid homology at their amino terminal half, which includes the DNA binding and dimerization regions. These transcriptional factors are normally sequestered in the cytoplasm where they are inactive by their association with an inhibitor protein, IkB, which in response to external signals associates from the complex allowing nuclear translocation and transcriptional activation through binding to DNA. Dr. Verma discussed three IkB proteins which he has studied in his laboratory including: I!!!B!!!, I!!!B!!!, and bcl3. Dr. Verma discussed structure function studies with the NF-!!!B superfamily and the IkB proteins. The I!!!B proteins physically associate with specific kB proteins through conserved ankyrin repeat motifs. Post-translational modification of I!!!B proteins modulate their biologic activity. Dr. Verma also presented evidence that the NF!!!B family of proteins directly physically associate with one of the basal transcription factors called the TATA binding protein (TBP). This binding involves the N-terminal 50 amino acids of rel and is mediated through the basic region and helix-loop-helix region of TBP. It is believed that this physical interaction is involved in the transcriptional activation properties of these transcription factors. Evidence of complex formation between TBP and rel was shown in vivo.
Dr. Mitsuaki Yoshida (University of Tokyo, Tokyo) described his studies with the human retrovirus, HTLV-1, which is the known etiologic agent of adult T-cell leukemia (ATL) and tropical spastic paraparesis (TSP). Dr. Yoshida described the Tax regulatory gene encoded by HTLV-1 and showed that this gene is not only involved in regulating HTLV-1 expression, but also expression of a number of important cellular genes, including growth factor receptors, cytokines, proto-oncogenes, and genes involved in immune recognition. Through the activation of some of these cellular genes, Tax has been proposed to contribute to abnormal proliferation of infected T-cells and to production of various cytokines that are frequently associated with phenotypes of these diseases. The responsive element in the viral LTR includes the CRE in the 21-bp enhancer and that in the IL-2R!!!gene includes a binding site NF-!!!B. At least part of the transactivation property of the Tax protein for HTLV-1 appears to be involved in targeting NF!!!B. Tax protein is, itself, not a DNA binding protein and, thus, its ability to activate transcription is believed to be mediated through its specific interaction with proteins. To identify the cellular proteins that are involved in Tax-mediated transcriptional activation, Dr. Yoshida has identified the proteins that bind to the enhancer that mediate the Tax protein transactivation. So far, two leucine zipper proteins (CREB and CREM) and the NF!!!B precursor protein p105 have been identified as binding the Tax protein. Data with enhancer binding mutants of the viral control region strongly suggest that these protein interactions mediate the transactivation by the viral Tax protein.
Dr. Warner Greene (University of California at San Francisco) also talked about NF-!!!B and I!!!B. To investigate the molecular basis for the critical regulatory interaction between NF-!!!B and I!!!B, a series of human p65 mutants was identified that functionally segregated DNA binding, IkB-mediated inhibition, and I!!!B-induced nuclear exclusion of NF-!!!B. His studies also revealed an autoregulatory loop in T lymphocytes, whereby p65 governs the expression of I!!!B including both its rapid degradation after p65 release and its subsequent resynthesis which is mediated through p65 activated expression of the I!!!B gene.
Dr. James L. Manley (Columbia University, New York) discussed mammalian pre-mRNA splicing regulators. Control of pre-mRNA splicing plays an important role in the life cycle of many viruses. He described studies of cellular factors that can modulate this process. Using SV-40 and adenovirus early pre-mRNAs, he was able to purify and characterize those proteins that modulate alternative splicing in vitro. He described one such protein referred to as alternative splicing factor 1 (ASF-1).
The third session was moderated by Dr. Akira Hakura (University of Osaka). Masuo Yutsudo described experiments carried out with Dr. Akira Hakura. The repression of early gene expression of HPV-16, a recently cloned nuclear factor involved IL-6 expression (NF-IL6). NF-IL6 is a transactivating regulator of IL6 gene expression and is known to be involved in the regulation of several other cytokines and several acute phase protein genes He demonstrated that the non-coding region of the HPV-16 genome contains NF-IL6 binding sites and that NF-IL6 can repress transcription of the early genes of HPV-16. The relevance of this repression was unclear since NF-~L6 is expressed predominantly on lymphocytes and not in epithelial cells.
Dr. Peter M. Howley (National Cancer Institute, Bethesda, Maryland) described functional studies of the E6 gene encoded by the cancer-associated human papillomaviruses The E6 protein encoded by the high-risk HPVs combine the tumor suppressor gene product p53 and targets degradation. The binding of E6 to p53 requires an additional cellular factor which is designated E6 AP (E6-associated protein). He described studies from his laboratory in which the cellular gene for E6 AP has now been cloned and shown to have the following properties: (1) It associates with wild-type p53 in the presence of the HPV-16 E6 protein and simultaneously stimulates the association of E6 with p53; (2) it associates with the high-risk HPV-16 E6 proteins in the absence of p53; and (3) it induces the E6 ubiquitin-dependent degradation of p53 in vitro.
Dr. Nobuyuki Tanaka (Osaka University, Osaka, Japan) described studies on the anti-oncogenic and oncogenic potentials of two cellular genes called Interferon Regulatory Factor-1 (IRF-1) and IRF-2. These genes encode factors which are structurally related and bind to the same DNA sequence. They are involved in the regulation of interferon and interferon-inducible genes. IRF-1 and IRF-2 have antagonistic properties. Overexpression of IRF-2 in NIH 3T3 cells results in transformation and enhanced tumori genicity in nude mice. This phenotype can be reversed by the concomitant overexpression of the IRF-1 gene. It was proposed that IRF-1 may be a tumor suppressor gene. It is located on human chromosome 5q31. 1 which is a region frequently deleted in patients with leukemia and myelodysplasia.
Dr. Carol Prives (Columbia University, New York) described studies on the structure and function of the wild-type and mutant p53 proteins. She described studies that showed p53 exists in a tetrameric form. She demonstrated that p53 can function as a transcriptional activator in vitro when it binds to its recognition site. She presented data on the mapping of the binding domain of p53.
Dr. Shizuo Akira (Osaka University, Osaka) described his studies on NF-IL6, which was initially identified as a nuclear factor binding to a 14 base pair palindromic sequence with an IL- I responsive element in the human IL-6 gene. NF-IL6 is a member of the C/EBP transcription factor family and contains leucine zipper. He described studies in which the transactivating function of NF-IL6 could be stimulated by the co-expression of the oncogenic ras protein. The ras-dependent transcriptional activity is regulated by phosphorylation within the serine rich domain located just upstream of the DNA binding domain. This site can be specifically phosphorylated by mitogen-activated protein (MAP) kinases.
Dr. Kunitada Shimotohno (National Cancer Center Research Institute, Tokyo) discussed his studies on the hepatitis C virus and its association with human hepatocellular carcinoma. He discussed the specific association of this virus with human hepatocellular carcinoma and gave a detailed presentation on the molecular biology of this virus. He presented a characterization of the viral proteins encoded by the HCV genome. In addition, he discussed a feature of HCV which is the heterogeneity of various isolates made in Japan so far. He has identified two hypervariable regions within the putative envelope protein and speculated that this hypervariability, even observed within the same patient, may play an important role in persistent infection.
Dr. Yoshiaki Ito (Kyoto University, Kyoto) described a transcription factor involved in regulating the polyomavirus enhancer which is closely related to the Drosophila segmentation gene (runt) and the human leukemia gene AML1. PEBP2 is composed of at least two subunits, referred to as A and B. He has isolated cDNAs representing each of these subunits, and it is the!!!1 subunit which is highly homologous to the Drosophila segmentation gene runt. The B subunits showed no significant homology with any other known proteins. The alpha protein binds to DNA by itself, whereas the B proteins do not. The B proteins, however, significantly increase the DNA binding affinity of the A protein. He concluded that PEBP2 is important in early mouse development, in T cell specific gene expression, and in leukemogenesis.
Dr. Ed Harlow (Harvard Medical School, Boston, Massachusetts) discussed the retinoblastoma protein as a member of a family of proteins that interact with conserved regions 1 and 2 of the adenovirus E1A protein. In addition, he spoke about the E2F transcription factor which can bind to the retinoblastoma protein and which is released from the retinoblastoma protein by binding of RB to adenovirus E1A. The retinoblastoma protein is thought to be a negative regulator of transcription by binding to the E2F transcription factor and holding it in an inactive state. Dr. Harlow presented data that E2F is a member of a family of proteins and that individual members of this family of proteins may differentially complex the retinoblastoma protein, p107 and p130. He presented a model in which the different forms of E2F may act at different points in the cell cycle.
Dr. Tetsu Akiyama (Osaka University, Osaka) presented his studies on the retinoblastoma protein and its regulation of the cell cycle. The retinoblastoma protein is differentially phosphorylated during the cell cycle and has been implicated in cell cycle control. He has been examining cellular proteins that interact with the RB protein. A series of cDNAS encoding retinoblastoma-associated proteins have now been cloned and are being studied. The phosphorylation of the retinoblastoma protein through the cell cycle was also discussed.
Dr. Masami Horikoshi (University of Tokyo, Tokyo) presented studies on the TATA box-binding factor (TFIID). These studies were done in collaboration with Dr. Robert Roeder at Rockefeller University, TFIID is a member of the basal transcription factors involved in regulating promoter expression Dr Horikoshi's research has been focused on an understanding of the structure-function relationships of TFIID with the other members of the basal transcriptional machinery. His studies focused on recent advances on the structure function analysis of the multi-subunit form of TFIID.
Dr. Mitsuaki Yoshida provided closing comments in which he thanked all the participants for their open discussion of current research in their laboratories. The meeting was judged to be highly successful by all of the participants. The participants unanimously recommended that a follow-up seminar be organized in two years to focus again on viral regulatory factors and their interaction with host cell factors.

(3) Seminar on "Topics in Carcinogenesis: Malignant Neoplasms Occurring Under Particular Circumstances"
This seminar was held on March 3-5, 1993, at the Maui Marriot Hotel, Lahaina, Maui, Hawaii. The organizers were Dr. Richard H. Adamson, National Cancer Institute, Bethesda, Maryland, and Dr. Yoichi Konishi, Nara Medical School, Nara. There were seven participants from both the United States and Japan. The purpose of the meeting was to present current work and discuss recent status of malignant neoplasms occurring under particular conditions. The first day of the meeting was devoted principally to carcinogenesis under particular conditions, carcinogenic mechanisms of non-genotoxic carcinogens, peroxisome proliferators and occupational cancer; the second day was concerned with malignant neoplasms under iatrogenic conditions and environmental tobacco smoke and lung cancer in non-smoking women; the third day was spent with general discussion and exchanging ideas individually among participants.
The opening remarks were delivered by Dr. Richard H. Adamson, who read a letter from Dr. Takashi Sugimura, who, unfortunately, could not attend the meeting.
Dr. Jerry M. Rice (National Cancer Institute, Bethesda, Maryland) discussed proliferative lesions that have occurred in humans at the sites of implantation of foreign bodies, in the context of the phenomenon of solid state carcinogenesis (foreign body oncogenesis) in animals. In humans, soft tissue sarcomas have occasionally been observed at the sites of permanent implantation of various flexible prosthetic devices (mostly vascular "grafts " of woven synthetic fiber textile materials) and at sites of other persistent foreign bodies, including surgical sponges, broken glass, bone wax, poorly absorbed liquids given as intramuscular injections; and rigid metallic objects including bullets, shrapnel and orthopedic prostheses such as artificial joints. Because of the very large number of breast implants that have been done for either corrective or augmentation mammoplasty during the last 20 years, this topic is of special current interest in the United States.
From studies in rodents, mostly rats and mice, several generalizations concerning foreign body oncogenesis have emerged. Virtually any persistent foreign body, of either natural or synthetic material, may elicit tumor development in rats or mice. Such neoplasms are always of connective tissue origin; they are generally osteo-, chondro-, or soft tissue sarcomas, with the single exception of the plasmacytomas that develop within the abdominal cavity in BALB/c mice that have been given either intraperitoneal injections of pristane or mineral oil, or intraperitoneal implantation of a plastic disc. Highly porous or finely powdered preparations of the same materials tend to be non-tumorigenic. The most active materials, irrespective of chemical composition, tend to be smooth objects that elicit a chronic, fibrotic, local tissue reaction that progresses to encapsulation of the implant.
Metallic foreign bodies often consist of alloys of carcinogenic metals such as nickel, chromium, and/or cadmium, and a chemical mechanism probably contributes to the oncogenic effect in such cases. The presentation by Dr. Rice dealt exclusively with non-metallic foreign bodies.
Experimentally, a marked pattern of interspecies variation in susceptibility to foreign body tumorigenesis exists and is directly correlated with the tendency of fibroblasts from a given species to become immortalized spontaneously on prolonged anchorage-dependent cultivation in vitro. Thus, rats and mice are highly susceptible, Syrian hamsters are less susceptible, and guinea pigs and chickens are refractory. Normal human fibroblasts are extremely resistant to immortalization in vitro, and only a very few cases of unequivocal sarcomas have developed in proximity to foreign bodies in humans. A cause-effect relationship between persistent foreign bodies and sarcomas in humans is thus still debatable. None of those sarcomas that have been reported in humans arose in a patient with a breast implant.
An additional difficulty in evaluating soft tissue responses to foreign.bodies is the phenomenon of pseudosarcoma in humans. Self-limited proliferative tissue reactions, including such entities as nodular fasciitis and myositis ossificans, can be mistaken histologically for sarcomas. Fibromatosis (desmoid tumor) resembles soft tissue sarcoma even more closely, but does not metastasize. It is generally accepted that pseudosarcomas may develop in humans in response to persistent foreign bodies, and in fact one such growth has been reported to have arisen in the fibrous capsule surrounding a breast implant. The clinical difficulty in distinguishing between pseudosarcoma and sarcoma is so great in some cases that a history of foreign body at the site of a growth (e.g., intramuscular injections of oil-based pharmaceutical preparations) has even recently been a significant criterion for electing conservative management of such cases, rather than the radical surgery that a diagnosis of sarcoma would justify. Because of this, reported cases of sarcoma in humans in connection with foreign bodies can be accepted with greatest confidence only in those cases in which metastasis has occurred.
The mechanism(s) of foreign body oncogenesis in animals remains obscure. There is ample evidence that, except in the case of metal alloys, classic genotoxic carcinogens of low molecular weight are not generally released from foreign bodies as a result of either a leaching process or decomposition of the foreign body. The process thus appears to be either a specific aspect of oncogenesis by non-genotoxic agents, or the result of an indirectly genotoxic process in which reactive agents are produced in response to the foreign body, but are not derived from it. There is evidence for the latter hypothesis from the inhibitory effects of antiinflammatory agents on the induction of intraabdominal plasmacytomas in BALB/c mice. Corticosteroids given before the foreign body, or indomethacin given concurrently with or following injection of mineral oil, markedly inhibit tumor development. Germ-free mice are also resistant. This evidence for a direct role of elements of the inflammatory process in at least one aspect of foreign body oncogenesis in rodents suggests a direction for further research on mechanisms of this process, and appears relevant to Dr Aozasa's report on sarcomas of the pleural cavity developing from long-standing empyema.
Dr. Yoichi Konishi (Nara Medical School, Nara) presented gene expression in rat transplantable osteosarcomas and malignant fibrous histocytomas (MFHs). In osteosarcomas, Northern blot analysis showed that c-Ha-ras, c-myc and NDPK/nm23 were expressed and the levels of c-Ha-ras and NDPK/nm23 expression increased in osteosarcoma with highly metastatic potentcy, suggesting that these two genes cooperated in the metastatic process of rat transplantable osteosarcoma. Transin expression was correlated with both c-fos and c-jun expression, suggesting that the transin gene expression might involve c-fos/c-jun conjugated at the AP- I site existing in the promoter region of transin gene. Bone Gla protein, a mature osteoblast phenotype, was expressed in osteosarcoma retaining osteoid formation. Lack of bone Gla protein mRNA expression was correlated with both c-fos and c-jun expression, suggesting that bone Gla protein expression might involve c-fos/c-jun conjugated at AP-site existing promoter region of bone gla protein gene in osteosarcoma. In MFHs, c-Ha-ras and NDPK/nm23 were expressed and transin expression was correlated with c-fos and c-jun expression. In both osteosarcomas and MFHs, Southern blot analysis showed no evidence for gene amplification and rearrangement of c-Ha-ras, c-myc, c-fos, transin and NDPK/nm23 were observed. These results suggest that the expression of these genes might be regulated at the transcriptional levels in those two particular malignant neoplasms, and further studies on other genetic alterations, such as point mutation of various genes, are required.
Dr. Tomoyuki Shirai (Nagoya City University Medical School, Nagoya) discussed reversibility of neoplastic cell proliferation induced by non-genotoxic carcinogens, butylated hydroxyanisole (BHA) and uracil, in the forestomach and urinary bladder of rats. He presented the mechanisms of forestomach carcinogenesis by BHA in which BHA-induced squamous cell hyperplasia is not primarily a regenerative response and may be caused by mitogenic effects of BHA which were studied by c-fos and c-jun expression followed by DNA synthesis and a very small amount of BHA-DNA adduct formation detected by the 32P-postlabeling method in the forestomach. Interestingly, epithelial hyperplasia induced by BHA regressed after cessation of BHA treatment. He also presented urinary bladder lesions in rats treated with uracil. Administration of uracil at the dietary level of 3% resulted in stones made by uracil itself and, subsequently, papillomatosis with active DNA synthesis followed by carcinomas. However, when uracil treatment was stopped, stone formation induced by uracil as well as papillomatosis of urinary bladder, epithelium disappeared suggesting that uracil-induced vigorous cell proliferation in the urothelium is due to direct mechanical irritation by calculi Those two results of continuous marked cell proliferation due to BHA or uracil which led to severe epithelial hyperplasias were essentially reversible.
Dr. James A. Popp (Chemical Industry Institute of Technology, Research Triangle Park, North Carolina) discussed peroxisome proliferation and cancer causation in animals.
Many chemicals including pharmaceuticals (most notably several important hypocholesterolemic agents); plasticizers, e. g., di-2-ethylhexylphthalate; solvents (such as trichloroethylene) and agricultural herbicides cause peroxisome proliferation in the livers of rodents. When these agents are given to rodents at high doses over long periods of time, liver tumors are generally found, although the incidence may be extremely variable depending on the chemical and dose administered. Despite the consistent hepatocarcinogenicity of rodent peroxisome proliferating agents, these chemicals have consistently been negative in a wide variety of genotoxic assays. The lack of genotoxicity and the association with peroxisome induction has resulted in the characterization of these agents as a special class of non-genotoxic rodent carcinogens. However, the relevance of the rodent liver tumor response for assessing potential risk to exposed humans is uncertain, particularly since dramatic species differences exist in the short-term response to these agents. It is particularly noteworthy that nonhuman primates and humans do not respond with peroxisome induction when exposed to these agents. Therefore, it is imperative that a better mechanistic understanding of tumor development following exposure to peroxisome proliferators be developed in rodents. Such a mechanistic understanding would allow for a scientifically based extrapolation of liver tumor responses in rodents for predicting potential effects in exposed humans.
The association of peroxisome induction with liver tumors in the rodent led to the hypothesis of a mechanistic linkage between the two responses. It is hypothesized that excess hydrogen peroxide is produced in the peroxisomes of rodent hepatocytes. Hydrogen peroxide that escapes degradation by peroxisomal catalase presumably leaks across the peroxisomal membrane resulting in oxidative injury in the cytoplasm of the hepatocyte. It is further hypothesized that peroxisomal derived hydrogen peroxide may cross the cytoplasm, eluding the degradative effects of cytoplasmic glutathione peroxidase and enter the nucleus where it results in oxidative injury to the DNA. This hypothesis has been evaluated in several laboratories around the world, resulting in experimental evidence that is both consistent and inconsistent with the hypothesized mechanism of carcinogenicity. Most notably, the evidence for oxidative injury in the cytoplasm is extensive but variable. While lipofuscin, a generally recognized marker of oxidative injury, is elevated very early in the course of exposure, biochemical markers of oxidative injury such as conjugated dienes are not elevated for more than 100 days after the onset of exposure. These results associated with the early induction of apoptosis have led to the suggestion that the early appearance of lipofuscin may be the result of apoptosis and not an indicator of early oxidative injury due to hydrogen peroxide derived from peroxisomes. Several references have addressed the appearance of 8-OH-deoxyguanosine appearing in the DNA of livers from rodents exposed to peroxisome proliferating agents. However, the origin of this oxidized base has not been clearly extablished at this time. Therefore, the presumed association of 8-OH-deoxyguanosine to peroxisome derived hydrogen peroxide is speculative at this time.
An alternative hypothesis for the mechanism of carcinogenicity is that peroxisome proliferating agents cause liver tumors in rodents through a potent promoting effect on spontaneously initiated cells in the rodent liver. Promoting activity has been demonstrated for many of the peroxisome proliferating agents, although a number of negative promotion studies have also been published. An analysis of this complex and confusing literature indicates that a number of negative studies are apparently the result of using an inappropriate marker for the identification of early preneoplastic foci. In these studies, gamma-glutamyltransferase (GGT) was used to identify and quantitate positively stained foci that are used as an indicator of promoting activity. However, since the publication of promotion studies based on GGT staining of foci, the lack of GGT staining of tumors and foci identified by other markers has been noted as a constant feature of lesions induced by peroxisome proliferators. Using tumors or foci identified by other histochemical markers as an endpoint most studies that have evaluated peroxisome proliferators for promotion activity have resulted in positive results. Indeed, in several cases the rate of proliferative lesion development and progression to tumors has been very rapid. The basis for this rapid development of tumors has become the basis for extensive investigation in recent years with efforts concentrated on the effect of peroxisome proliferators on hepatocyte replication. It has been known for many years that peroxisome proliferators cause a dramatic proliferative response in hepatocytes within the first several days of exposure. More recently, a chronic proliferative effect has been demonstrated for some but not all hepatocarcinogenic peroxisomal proliferators. The relevance of the proliferative effect in the entire population of hepatocytes for the eventual tumor formation is uncertain. However, the recently documented effects of peroxisome proliferators to dramatically enhance the replication of hepatocytes within specific populations of hepatocytes is undoubtedly important for the ultimate development of tumors. The replication rate of hepatocytes within foci is 60 times the replication rate of hepatocytes in adult control rat liver. This dramatic enhancement of replication within foci suggests that the peroxisome proliferating agent causes preferential growth of initiated cells, leading first to the development of preneoplastic foci and ultimately liver tumors. Current studies are addressing the mechanism of peroxisome proliferator enhanced cell replication. The recent identification of a peroxisome proliferator activated receptor provides an important opportunity for resolving the role of peroxisome induction and enhanced hepatocyte replication to hepatocarcinogenicity in the rodent. An understanding of the mechanism of carcinogenicity of peroxisome proliferators in the rodent will provide not only an interesting scientific advancement but will provide the scientific basis for quantitatively improving the assessment of risk of exposed humans.
Dr. Frank J. Gonzalez (National Cancer Institute, Bethesda, Maryland) discussed characterization of the human peroxisome proliferation-activated receptor. Peroxisome proliferators are a diverse group of chemicals which includes hypolipidemic drugs, herbicides, leukotriene antagonists and plasticizers. Two major categories of peroxisome proliferator chemicals play a significant role in human society. The first, the fibrate class of the hypolipidemic drugs, have been found effective at reducing the levels of triglycerides and cholesterol in humans suffering from hyperlipidemia, a major risk factor for coronary heart disease. The second category relates to phthalate ester plasticizers used in the production of highly versatile flexible vinyl plastics. Peroxisome proliferators seem to affect most mammalian species that have been tested. They induce hepatomegaly resulting from liver hyperplasia and an increase in size and number of peroxisomes. Neverthless, based on the hypolipidemic drug dose required to produce recognizable peroxisome proliferation, mouse and rat are considered highly responsive to these agents and develop hepatocellular carcinoma following long-term drug administration, hamsters have intermediate responses and guinea pigs, marmosets and other nonhuman primates are weakly responsive.
Examination of human liver biopsy samples revealed either no differences between fibrate-treated and control groups or small changes in the density of mitochondria and peroxisomes but no toxic or carcinogenic effect. Only one study indicated an increased morbidity of patients receiving clofibrate. The carcinogenic effect in rodents together with the limited fibrate studies in humans raised the argument that humans treated with fibrate drugs could be at risk for cancer development.
Elucidation of the mechanism by which peroxisome proliferators induce carcinogenesis is a prerequisite for assessing the health risk to humans in the pharmaceutical use of fibrates. So far, the mechanism of action of clofibrate and similar drugs is unknown. Peroxisome proliferators are termed non-genotoxic carcinogens since they fail to directly cause DNA damage. The oxidative stress hypothesis suggests that enhanced production of hydrogen peroxide by increased peroxisomal fatty acid 6-oxidation overcomes the ability of H₂O₂-degrading enzymes and results in a slow accumulation of oxidative damage to the genome. Alternatively, it was suggested that peroxisome proliferators promote spontaneously-initiated foci in liver.
Studies on long-term effects of two peroxisome proliferators inducing high and low carcinogenic activity, respectively, indicated that factors other than oxidative injury may be important in the carcinogenicity of such chemicals. Speculation on the existence of specific receptors for mediating the action of peroxisome proliferators has recently been supported by the isolation of new members of the steroid hormone receptor superfamily which are activated by peroxisome proliferators: the mouse peroxisome proliferator activated receptor (PPAR), the Xenopus laevis PPAR and the rat PPAR. Moreover, it has been shown that PPARs are capable of activating the promoter of the rat acyl coenzyme A oxidase gene, the key enzyme of peroxisomal fatty acid!!!-oxidation, and mediates the induction of rabbit CYP4A6, a cytochrome P450 fatty acid!!!-hydroxylase, by clofibric acid. This activation is carried out through specific response elements that have recently been identified.
Our approach has been to examine whether differences between human and mouse PPAR could account for the marked differences in responsiveness to peroxisome proliferators displayed by both species and perhaps to predict the effects of peroxisome proliferators on human hepatocarcinogenesis. To achieve this goal, the human PPAR was isolated from a human liver cDNA library and compared its structure and ability to activate the ACO response element with that of mouse PPAR.
The human peroxisome proliferator activated receptor, hPPAR was cloned from a human liver cDNA library. The cDNA exhibited 85% and 91% DNA and deduced amino acid sequence identity with mouse PPAR, respectively. The hPPAR gene was mapped on human chromosome 22 slightly telomeric to a linkage group of 6 genes and genetic markers that are located in the general region 22ql2-ql3.1. Cotransfection assays of mouse Hepa 1 cells were used to roughly compare the ability of hPPAR and mPPAR expressed cDNAs to trans-activate the acyl CoA oxidase (ACO) PPAR response element located 5' upstream to the minimal thymidine kinase promoter driving expression of the chloramphenicol acetyl transferase (CAT) reporter gene. Both receptors elicited a response with the prototypical peroxisome proliferators nafenopin, clofibrate and WY-14,643. Moreover, using co-transfection assays in which the CAT reporter plasmid contained the CYP4A6 gene response element rather than the ACO element, it was shown that hPPAR is capable of very efficiently trans-activating a second PPAR response element. These results indicate that the PPAR is present in humans in a form that is functional and can trans-activate response elements derived from two different genes, the rat ACO and the rabbit CYP4A6.
Dr. Seiji Naito (Kyushu University, Fukuoka) presented a talk on urothelial carcinoma in dyestuff workers exposed to aromatic amines in a certain chemical factory in Kitakyushu, Japan. The production of aromatic amines of benzidine, 2-naphthylamine and dianisidine was started in 1936, 1948 and 1949 and continued until 1971, 1959 and 1971 respectively. He found 75 (17.1%) urothelial carcinoma cases out of 438 workers by the end of 1991 with an average latent period of 24 years and 5 months. Sixty-one (23.3%) cases out of 262 workers were exposed to benzidine, 12 (12.8%) of 94 workers were exposed to 2-naphthylamine, none of 60 workers were exposed to 1-naphthylamine, none of 2 workers were exposed to dianisidine, and 2 (10%) of 20 workers were exposed to the plural compounds. Smoking did not increase the incidence of urothelial carcinomas. He also stated that even in the past 5 years, urothelial carcinomas were newly developed in 8 workers and that follow-up studies, including early detection by urine cytology and treatment by transurethral surgery, are strongly required for the workers exposed to aromatic amines. Except for urothelial carcinoma, liver cell carcinoma was found in 5 (1.1% ) cases among workers.
Dr. G. Iris Obrams (National Cancer Institute, Bethesda, Maryland) discussed AIDS-associated cancers. What is interesting about cancer in AIDS? The human immunodeficiency virus (HIV) is itself not an oncogenic virus--it is cytopathic and does not directly cause transformation. So we need to ask what is known about HIV infection that may increase susceptibility to agents that cause cancer, since we observe that overall approximately one-third of HIV-infected persons develop cancer.
The World Health Organization estimates that 5 to 10 million persons are infected with HIV worldwide. Hundreds of thousands of mV-infected people are now approaching the median time to AIDS development. Infected individuals are in general young, with an age at AIDS diagnosis in the mid-thirties for both men and women. It is well known that HIV is transmitted sexually, with exposure to blood, and perinatally from an infected mother. Although detectable HIV viremia declines markedly after the acute syndrome of HIV infection, HIV causes gradual depletion of T lymphocytes: HIV clearly affects cell-mediated immunity. It is well established that the CD4 membrane glycoprotein of the T helper lymphocyte acts as the receptor for HIV in human cells. The profound immunodeficiency found in AIDS is principally due to depletion of these CD4 lymphocytes, but a wide range of immunological abnormalities are associated with HIV infection.
With iatrogenic immune deficiency, specific types of tumors have long been known to occur with increased incidence. Among renal transplant recipients, skin cancer, non-Hodgkn's lymphomas, Kaposi's sarcoma, and possibly cervical and vulvar/perineal cancers are increased in the data reported to the Cincinnati Transplant Tumor Registry. Of other cancers, there are small numbers, but there are suggestions of increases in melanoma and primary liver cancer. The most common tumors of the general population (carcinomas of the lung, breast, colon) have not shown an increase in transplant recipients. Hodgkin's disease as well is not increased. There are similarities between the tumor types that are generally thought to be virus-associated and the tumors occurring with immunosuppression. This suggests that it may be viral oncogenesis that the immune system protects us against.
Increases in cancer incidence among HIV-infected persons have been striking. Significant elevations in relative risk are so far limited to Kaposi's sarcoma (KS) and B-cell non-Hodgkin' s lymphoma (NHL). The risk for KS among HIV-infected individuals is over 1000-fold that in the general population, and the risk for NHL may be 50- to 100-fold.
The most common neoplasm associated with HIV infection is KS. KS is a multifocal tumor of the skin, which can affect other organs as well. Many persons have lesions in different stages of development. The histology (proliferation of vascular channels and spindle cells) is more consistent with reactive hyperplasia than with malignancy. A variety of cells have been proposed as the cell of origin, most commonly a lymphatic cell of vascular endothelium. KS in AIDS patients differs from the disease seen prior to AIDS and is similar to that seen in other conditions of immunosuppression: clinically, the disease can be aggressive with involvement of mucosal surfaces, visceral organs, and lymph nodes. Male homosexuals or bisexuals are the HIV transmission group with by far the highest incidence of KS, which is seen in 20 to 25% of this group. The difference between risk groups suggests that immunosuppression alone is not the entire explanation for this disease in AIDS. The geographic distribution of KS in the U.S. is highest in the foci of the HIV epidemic; no other HIV complication has a similar geographic distribution Among the few women with KS the source of exposure to HIV was a bisexual man rather than a drug abuser, suggesting that a "KS agent," perhaps a causative infectious agent, is acquired by sexual contact with a member of the groups at highest risk for KS. KS cases tend to have more sex partners, have sex partners from areas where KS is common, and have more exposure to feces during sexual activity than HIV-infected individuals who do not develop KS. This may implicate agents such as enteric viruses in KS. KS cases among heterosexuals and children a especially frequent in Africa and the Caribbean, which may reflect a higher prevalence of symptom-free carriers of the KS agent. The epidemiology of KS is consistent with a disease caused by a widely disseminated, clinically inapparent transmissible agent that is expressed in persons with immunosuppression. Viruses have been looked for in KS, but no associations with known viruses have been detected to date.
Various growth-promoting angiogenic factors are secreted by HIV-infected cells, including IL-6 and the HIV tat gene product (tat protein) that may cause endothelial cell proliferation and promote KS. These are being studied for their roles in KS; for example, mice transfected with the HIV tat gene can develop KS-like tumors. KS cells themselves secrete substances that stimulate their own growth, such as interleukins. Interestingly, KS in HIV infection and other immunosuppressed conditions has been halted and even reversed by withdrawal of steroids or other immunosuppressive drugs, consistent with a growth factor-driven process.
Numerous reports describe the significantly increased risk in AIDS of NHL the second most common cancer in HIV infection. The relative risk for males in their thirties who are HIV-infected is over 50-fold, at a median age that is much younger than for non-HIV-related NHL. The histologic types of NHL in AIDS differ markedly from those in non-AIDS patients, with over 60% in AIDS being high-grade B-cell NHL. Primary CNS lymphoma is common enough to be recognized as an AIDS-defining condition. A large proportion of HIV-associated NHLs have extranodal involvement and advanced disease at diagnosis.
NHL is usually diagnosed after the onset of other manifestations of HIV infection; some cases remain silent and are diagnosed only at postmortem examination. The risk of NHL increases as immunodeficiency worsens. Prolongation of life with HIV infection in the setting of continued severe immunosuppression may well lead to the development of more NHL with time, as this NCI data seems to suggest: after one year on AZT, about 25% of a group of patients had been diagnosed with lymphoma.
The epidemiology of NHL does not provide clear etiologic leads. Unlike KS, there is no predilection for any particular AIDS risk group for this disease. There is no systematic geographic variation in HIV NHL. As with KS, no evidence of HIV DNA has been found in NHL specimens. The role of Epstein-Barr virus (EBV) in the pathogenesis of NHL has been proposed by many investigators. Activation of EBV among carriers of this virus is well recognized in the setting of HIV infection, with increased numbers of circulating EBV-infected B cells. EBV DNA-specific sequences have been found in about 1/3 of NHL tumor specimens. EBV has been found in the tumor specimens of all patients in a series of AIDS brain lymphomas. EBV infects B lymphocytes and stimulates a polyclonal B cell proliferation that normally is controlled by EBV-specific cytotoxic T lymphocytes and possibly by natural killer cells. However, since these mechanisms are often defective in mV infection, the proliferation of infected cells can continue unchecked. A cytogenetic event then may have a higher probability of occurring, perhaps due to factors yet to be identified, with the emergence of a monoclonal tumor. This theory is supported by reports that AIDS patients with polyclonal B-cell proliferations progressed within months to clinically malignant monoclonal B-cell lymphoma. However, this is likely to explain only the 1/3 of HIV NHL associated with EBV infection, so there is no single cause.
In contrast with the clearly excessive incidence of both KS and NHL in association with HIV infection, the association with other malignancies has not yet been established. Why is it that many types of cancer, including virus-associated neoplasms, have not emerged despite profound immunodeficiency? One reason may be that the latent period for many cancers exceeds the 10-year period in which AIDS has been a major epidemic. In iatrogenic immunosuppression, the median latency of many tumors occurring is 8 years or more, and we are only now approaching this time in the course of the HIV epidemic.
Cervical cancer has been reported by several centers to be occurring in young, HIV infected women. It is not clear, however, that cervical cancer is increased in AIDS. While the numbers of women with AIDS are increasing rapidly, it does not appear that cervical cancer has been increasing in areas with high risk groups. However, the reports of a possible relationship between progression of cervical cancer and mV infection remain of interest because of the close association between several strains of human papillomavirus (HPV) and cervical neoplasia. For this reason, the Centers for Disease Control and Prevention has just expanded the AIDS definition to include invasive cervical cancer. If HPV reactivation occurs with immune system perturbations which is likely there may be increased chances for integration and malignancy development. In this context it is notable that recent data show that HIV-positive women are up to fivefold more likely to have HPV infection than HIV-negative women.
Anal cancer in homosexual men antedates the HIV epidemic, but there is a sharper rate of increase among homosexual men since 1987 reported from San Francisco. In one series of stage IV HIV-infected men m San Francisco, 15% were found to have anal intraepithelial neoplasia. Up to 100% of these tumors have been found to contain HPV. HPV infection is up to tenfold more common with HIV infection than in HIV-negative homosexual men.
Another cancer reported in AIDS patients is Hodgkn's disease (HD) It is as yet unclear whether HIV infection leads to HD or modifies the course of the disease. A large proportion of HD cases in AIDS present with advanced disease, extranodal involvement, and a relative increase in the mixed cellularity subtype--which may be the type most clearly related to EBV infection.
Another tumor mentioned as possibly increased among HIV patients is hepatocellular carcinoma. The majority of HIV-infected individuals in some series are inflected with hepatitis B virus (HBV), which has been associated with hepatoma, and loss of immunity has been shown to result in reactivation of HBV.
Tumors of smooth muscle, such as leiomyomata and leiomyosarcomata, have been described in a number of children with HIV infection. These are not associated with other forms of immunodeficiency. The hypothesis that is proposed is that a growth factor or factors elaborated during HIV infection play a role m the development of these tumors.
The study of HIV-associated cancers continues to be of much interest and very challenging. No single mechanism can account for the development of HIV associated malignancies. Of the malignancies that are associated with HIV infection or are suspected to have such an association, most are also thought to have viral etiologies. All of the candidate viruses that have gained attention cause latent infections that are normally kept in check by cellular immunity. With the progressive depletion of CD4 cells that is the hallmark of HIV infection, accompanied by many other immune changes, an individual 's latent infections may escape normal controls. Increased viral replication may then increase risk of malignancy. In addition, other mechanisms involving growth factors and cytokines that may well play a role in the development of KS, and possibly smooth muscle tumors as well, need to be further defined. The National Cancer Institute is supporting a wide range of studies of HIV infection and cancer. Future directions will include expanding surveillance for malignancies in groups at risk to clarify whether other cancers emerge, evaluating treatments for these aggressive tumors, and continuing the analytic and etiologic studies into the pathogenesis of these malignancies: for example, to clarify the role EBV and other viruses play in the development of lymphomas, and to continue the search for the proposed KS agent. The information gained through study of AIDS associated malignancies may provide insights into more general mechanisms of carcinogenesis, particularly the role of altered immunity in the development of specific malignancies.
Dr. Dennis Deapen (University of Southern California School of Medicine Los Angeles, California) presented his research on augmentation mammoplasty and breast cancer The potential health risks associated with breast implants have received considerable attention during the last two years. Yet, prior to this study, no evaluation of cancer risk among human breast implant recipients had been conducted.
In 1976-1980, he established a retrospective cohort of 3,112 non-reconstructive breast implant patients who received their implants between 1959 and 1980. These patients were collected from the records of 35 board-certified plastic surgeons; all non-Spanish and non-Asian-surnamed white female patients who resided in Los Angeles County at the time of implantation were included. Patients with prophylactic subcutaneous mastectomy or breast cancer were excluded. Also collected from the medical records were identifying and demographic information details of the devices and surgical procedures and information on subsequent visits.
Due to privacy concerns, the patients were not contacted. Rather, they were followed through use of public records (i.e., driver's and vehicle license records, voter registration rolls, property tax rolls, marriage and divorce records, death records, records of civil actions and through two large multistate credit bureaus. All name changes and alternate birthdates and social security numbers were recorded Residence in or out of the Los Angeles County coverage area was documented.
Incident cancers occurring in the study population were detected by searching the records of the Los Angeles County Cancer Surveillance Program (CSP), the complete, population-based cancer registry covering the Los Angeles County. The registry was searched for cohort members by means of computerized matching, using all available name birthdate and social security number information. Each potential match was manually judged to be an exact, possible or nonmatch. For exact and possible matches, the CSP and implant patients records were further examined for evidence of match or nonmatch.
The cancer incidence reporting was complete for the years 1972- 1986. For each study subject, the post-implant cancer follow-up period was determined, commencing with the date of first implant or January 1, 1972, whichever was later and ending the date of last continual residence in the county, date of death or December 31, 1986, whichever was earliest.
Population-based age-, race- and socioeconomic-specific cancer incidence rates for the county population 1972- 1986 were calculated. The number of cancer cases expected to occur in the breast implant cohort was calculated by multiplying these incidence rates by the age-specific person-years of follow-up of the implant patients, Standardized incidence ratios (SIR) were calculated as the number of observed cancers divided by the number of expected cancers, multiplied by 100%.
The median age of the patients at implant was 31.4 years and the median interval of follow-up was 10.6 years (range 0.1 to 31.7 years). Silicone-filled implants were used in 79% of the patients, 9% were saline and 14% were other or unknown. Sixty-nine of the devices were polyurethane-coated.
Twenty-one breast cancers were observed among implant patients as compared to 31.7 expected (SIR = 66%, 95% confidence limits (CL): 41%, 101%). The median age at implant of the 21 breast cancer cases was 40.5 years with a median age at diagnosis of 45.8. The median interval from implant to diagnosis was 8.1 years. Since the median age at implant of the cancer cases was 9 years older than the cohort as a whole, we separately estimated the risk in the 572 women who were 40 years and older at implant. Fourteen breast cancers were expected and 11 were observed (SIR = 78%, CL: 39%, 140%). Risk among patients with the longest implant exposure was also examined. For the 1,947 patients with 10 or more years post-implant, 27.4 breast cancers were expected and 18 observed (SIR = 66%, CL: 39%, 104%).
The distribution of stage at diagnosis among the breast cancer patients was found to be virtually identical to the Los Angeles population figures for white women of the same age, with the following observed for implant patients and population figures, respectively: in situ 7.1% and 5.3%, localized 50.2% and 57.9%, and nonlocal 42.7% and 36.8%.
Finding 33% fewer breast cancer cases than expected prompted the investigation to the question whether their detection methods were complete. Assuming that the method would be no more likely to miss breast cancers than cancers of any other site, all other cancers, excluding nonmelanoma skin cancers and cervical cancer in situ were examined. While 50.0 cancers were expected, 45 were observed (SIR 90%, CL: 66%o, 120%) -- a nonsignificant difference.
Thus, a nonsignificant deficit of breast cancer was found among 3,112 breast implant patients followed an average of 10.6 years. No increase was observed in either older patients or those with longer exposure. Although the devices might be expected to hinder both palpation and radiologic screening, the stage at diagnosis of the breast cancers was found to be no different than expected. No cancer of any site was observed among the 69 patients who received polyurethane-coated implants.
Currently, cancer risk in this cohort is again being estimated for the period 1987 - 1991. These results will be available in late 1993 and are expected to increase the median period of exposure to 15 years, with several hundred patients 20 years post-implantation, allowing assessment of long-term risk.
Dr. Hiroyuki Tsuda (Fujita Health University School of Medicine, Toyota) reported on the occurrence and histogenic analysis of renal cell tumors in long-term hemodialysis patients. He collected 28 autopsy cases which received hemodialysis (20 males aged 56.7 and 8 females aged 63.5) and histopathologically analyzed renal lesions. Total incidences of neoplastic lesions were: renal cell carcinoma. 8/28 (28.6%); renal cell adenoma, 4 (14.3%); epithelial hyperplasias, 18 (64.3%); dysplastic tubules, 12 (42.9%). The incidence of renal cell carcinomas reached 75% (6/8) in the patients who received hemodialysis more than 10 years and histologically, those carcinomas were mainly papillary type carcinoma. Histogenetically, it was suggested that an unusually high incidence of papillary type tumors may be related to cystic change and calcium and occasional oxalate crystal deposition associated with hemodialysis. He added some experimental results of renal carcinogenesis in rats and insisted that papillary tumor developed from distal segment or collecting duct.
Dr. Hiroshi Tanooka (National Cancer Center Research Institute, Tokyo) discussed the Japanese thorotrast study and the effects of radon on cancer development in humans. Thorotrast and radon are known as!!!-radiation. Thorotrast was used for angiography and deposits in the liver and induces liver tumors, especially cholangiocarcinomas. In Japanese studies, Dr. Tanooka emphasized the importance of selecting a proper control group for obtaining relative risk values. The relative risk for liver cancer from the German thorotrast study is 155, while the relative risk in the Japanese thorotrast study is 52.2 which was obtained from proper control data set up by Dr. Mori. The different values between Germany and Japan are thought to be due to a low spontaneous liver cancer incidence in Germans. Radon hot springs are located in a town called Misasa, located in Tottori-ken, Japan. Misasa provides substantial dosimetry data and enough numbers of a stable population for an epidemiological survey. Residents of Misasa used high amounts of radon-containing water until completion of central water supply in 1983. The average radon concentration in spring water is 437 KB₆/m³. Mortality records (1952-88) were analyzed in an age- and sex- matched manner. The SMR values for Misasa vs. Japan as a whole were 0.538 and 0.463 for all cancers in males and females. The Poissan regression analysis showed significantly lower mortalities in Misasa for cancers of all sites, including stomach. Lung and colorectal cancer mortalities were also lower but not statistically significant. Those two studies indicate that the risk of cancer from!!!-radiation is quite high at a high dose rate, e.g., from thorotrast, but is low at a low dose rate as from radon in hot springs. The underlying mechanism is thought to be due to efficient repair at a low dose rate or "hormesis " in the whole body system.
Dr. Katsuyuki Aozasa (Nara Medical University, Nara) reported the development of sarcomas of the pleural cavity from long-standing empyema patients. He found 3 (2.2%) cases of malignant lymphoma in the plural cavity among 154 patients with chronic tuberculous empyema in a certain hospital in Osaka from 1971 to 1985. This incidence of malignant lymphoma indicated an intimate relationship between these two diseases based on the survey of Annual Pathological Autopsy Cases in Japan during 1978 to 1987, etc. He extended the studies to a nationwide survey of Japanese hospitals and collected 37 cases with pleural lymphomas and found that the artificial pneumothorax left chronic non-healing inflammation in the pleural cavity, which may be a causal factor for pleural lymphoma. He also added that there were two cases (1.5 % for 100 patients with empyema) of soft tissue sarcomas affecting the pleural cavity and stated that chronic empyema may also be the predisposing lesion for soft tissue sarcomas and malignant fibrous histiocytoma.
Dr. Elizabeth T. H. Fontham (Louisiana State University, New Orleans, Louisiana) presented studies on environmental tobacco smoke and lung cancer in nonsmoking women. In January 1993, the U.S. Environmental Protection Agency (EPA) issued a report on the respiratory health effects of passive smoking in which they concluded that environmental tobacco smoke (ETS) is a human lung carcinogen, responsible for approximately 3000 lung cancer deaths per year in U.S. nonsmokers (1).
Weight of evidence analysis for lung cancer risk was developed according to EPA guidelines for risk assessment which included evaluation of nine criteria. Briefly they are as follows:

1. Biologic plausibility. The components of ETS have been demonstrated in nonsmokers. Various components can be and have been both detected and quantified. The same carcinogens present in mainstream smoke (MS) are also present in ETS, although there are quantitative differences and potential differences in phase distribution. There is ample scientific documentation that active smoking is carcinogenic in humans in the lung and at other sites, and the nonthreshold dose-response relation in active smoking supports biologic plausibility.
   
2. Supporting evidence exists from animal bioassays and genotoxicity experiments. Although there are no lifetime inhalation studies of ETS, sidestream smoke (SS) condensates have been found to be carcinogenic in intrapulmonary implantations and skin painting experiments. Both MS and SS are positive in genotoxicity experiments.
   
3. Consistency of response is observed. An increased risk of lung cancer in female never smokers ever exposed to spousal ETS has been observed in 24 of 30 studies worldwide, including all four cohort studies.
   
4. Broad-based body of data is available. With the exception of two areas of China notable for high levels of indoor air pollution, studies from eight countries show consistent, modest increases in lung cancer associated with spousal ETS exposure.
   
5. There is an upward trend in exposure-response. Although many of the studies had small sample sizes and low statistical power, 10 of 14 studies with sufficient data demonstrated statistically significant increased risk of lung cancer with increasing levels of ETS exposure.
   
6. Increased risk is detectable at environmental exposure levels. Excess lung cancer risk in nonsmoking women from exposure to spousal ETS is large enough to be observed even for all exposure levels combined (i.e., ever/never exposed). More often than not high exposure settings, i.e., occupational settings or high-dose experimental animal studies, are required to demonstrate carcinogenic response.
   
7. Effects remain after adjustment for potential upward bias. Since smoking causes lung cancer, a misclassified smoker has a greater chance of being a case than control. Smokers tend to live with smokers; therefore, misclassification of current or ex-smokers as nonsmokers would tend to bias the risk associated with living with a smoker upward. After adjustment of the relative risk estimates downward to reduce the potential for misclassification bias on a study by study basis, the summary relative risks for the U. S. and other countries remain statistically significant.
   
8. Strong associations are observed at the highest exposure levels. The highest relative risks are observed in the highest exposure categories, and 9 of 1 6 studies report statistically significant elevations despite small sample size and low power.
   
9. Confounding cannot explain the association. Whenever a risk factor yields a relative risk estimate of less than 2.0 there is always concern that the observed association is a result of confounding. Numerous potential confounding factors have been examined and none appear to explain the observed relative risk. The dose-response relation argues against major confounding.

Tables 1 and 2 display the summary relative risks by country for overexposed nonsmokers and nonsmokers exposed in the highest category, respectively. For the United States, these estimates are 1.19 (90% confidence interval 1.04-1.35) for ever-exposed and 1.38 (90% confidence interval 1.13-1.70) at the highest exposure level. Worldwide, an 80% excess risk is observed in the highest category.

Two large U.S. studies have been published since the preparation of the EPA report (2,3). Because they are similar in size to the largest U.S. study (4), their findings could have had an impact on the U.S. summary risk estimates had they been included. Brownson et al. (2) reported findings from a population based case-control study of Missouri female nonsmokers which included 434 cases and 1166 controls who were considered lifetime nonsmokers. Interviews were conducted by telephone. No increased risk was observed for ever-exposed to spousal ETS, O.R. = 1.0 (0.8-1.2); however, the confidence interval does include 1.19, the summary U.S. risk estimate for ever-exposed. The highest exposure category in this study was defined as greater than 40 pack-years. The relative risk estimate, 1.3 (1.0-1.7), was quite close to the U.S. high-exposure summary estimate of 1.38. Stockwell et al. (3) conducted a population-based case-control study of lung cancer in central Florida. Their study included 210 female lifetime nonsmokers with lung cancer and 301 lifetime nonsmokers as controls. Cases and controls were interviewed either in person, by telephone or by mail. The estimated relative risks for spousal ETS exposures in this study are among the highest reported for U.S. studies. The adjusted odds ratio for ever-exposed to spousal ETS was 1.6 (95% C.I. 0.8-3.0). A significant positive trend (p=0.025) was observed for household smoke-years of exposure during adulthood. The risk estimate at the highest level of 40 or more smoke-years was 2.4 (1.1-5.3).
In addition to these two new studies, ongoing analysis of a large multicenter U.S. case-control study was discussed (4). Two additional years of data, accrued after the original report which was included in the EPA analysis, support the earlier findings. A 70% excess risk of lung cancer is observed for the highest spousal pack-year exposure category (p for trend = 0.05), and statistically significant increased risks are observed for ever-exposed to spousal ETS, ETS in the workplace and t,-TS in social settings of approximately 25%, 45% and 50%, respectively.
The findings of these new reports suggest that the conclusions of the EPA report were both reasonable and timely. They provide additional support for the finding that environmental tobacco smoke increases risk of lung cancer in nonsmokers.

1. Respiratory Health Effects of Passive Smoking: Lung Cancer and Other Disorders. United States Environmental Protection Agency. Washington. D.C., 1992.
   
2. Brownson, R.C., Alavanja, M.C.R., Hock, E.T., and Loy, T.S. Passive smoking and lung cancer in nonsmoking women. Am J Pub Health 82 (11): 1525-1530, 1992.
   
3. Stockwell, H.G., Goldman, P.L., Lipman, G.H., Moss, C.I., Armstrong, A.W., Pinkham, P.A., Candelora, E.C., and Brusa, M.R. Environmental tobacco smoke and lung cancer risk in nonsmoking women. J. Natl. Cancer Inst. 84: 1417-1422, 1992.
   
4. Fontham, E.T., Correa, P., Wu-Williams, A., et. al. Lung cancer in nonsmoking women: A multicenter case-control study. Cancer Epidemiol, Biomarkers & Prev. 1: 35-43, 1991.

Dr. Eiji Tsuchiya (Cancer Institute, Tokyo) reported on a follow-up study and possible carcinogenesis of lung cancer in chromate workers in a certain area of Japan. He followed 84 male ex-chromate workers from a chromate factory since 1975 and so far has found 12 cases of lung cancer. Seven out of 12 cases were multiple lung cancer and the average exposure period was 19 years. The average latent period was 39 years. Smoking might also be involved in the induction of lung cancer. Histologically, by the numbers of tumors including multiple lung cancer cases, 19 were squamous cell carcinoma, 4 were small cell carcinoma and I was large cell carcinoma. Only one case of multiple lung cancer was squamous cell and small cell carcinomas and 6 cases showed the same histology in multiple carcinomas. Mostly, tumors were of the hilar type. The carcinogenesis study showed that squamous cell carcinoma appeared to develop from precancerous lesions located often at the bifurcation of the bronchus where the chromium content was particularly high. Time-trend data revealed that some atypical lesions develop into cancer, but most atypical lesions surprisingly showed regression.


SEMINAR AGENDA AND PARTICIPANTS

(1) Seminar on "Nitric Oxide Synthesis and Carcinogenesis
January 19-20, 1993, The Williamsburg Hospitality House, Williamsburg, VA

AGENDA

Tuesday, January 19
9:00 a.m.	Introduction	H. Esumi & S. Tannenbaum
9:10 a.m.	Biochemistry of NO•Synthases	M. A. Marletta
9:55 a.m.	Structure and Function of NO•Synthases	H. Esumi
10:40 a.m.	Break
11:00 a.m.	Cellular Consequence after NO•Stimulation in Neural Cell	K. Shibuki
11:45 a.m.	Role of L-Arginine Products in Lung Inflammatory Injury	P. Ward
12:30 p.m.	Lunch
1:45 p.m.	Nitric Oxide and Cigarette Smoke Toxicity. Oxidation of Methionine, a Deca Peptide Containing Methionine and!!!-l-Protease Inhibitor	W. Pryor
2:30 p.m.	Role of IL-8 in Inflammation	K. Matsushima
3:15 p.m.	Break
3:30 p.m.	Endogenous Formation of Nitrate and Nitrosamines in Humans Associated with OV Infection	M. Tsuda
4:15 p.m.	Open Discussion
6:30 p.m.	Dinner
Wednesday, January 20
8:30 a.m.	Chemistry of DNA Damage Induced by NO•	S. Tannenbaum
9:15 a.m.	A New NO•Antagonist Imidazoline-N-oxide-oxyl	H. Maeda
10:00 a.m.	Break
10:20 a.m.	Genotoxicity of NO• and its Progenitor Drugs	L. Keefer
11:05 a.m.	Role of NO• in Carcinogenesis in Animal Model	T. Shirai
11:50 a.m.	Discussion
12: 10 p.m.	Lunch
1:30 p.m.	Mutations Induced by NO•	G. N. Wogan
2:15 p.m.	p53 Mutations in Human Cancers from Various Organs	S. Hirohashi
3:00 p.m	Hypothesis - Generating Clues to Cancer Etiology and Mechanisms	C. Harris
3:45 p.m.	Open Discussion
4:15 p.m.	Adjourn

PARTICIPANTS

UNITED STATES

Steven R. Tannenbaum, Ph.D.
Professor of Chemistry and Toxicology
Massachusetts Institute of Technology
77 Massachusetts Ave., 56-311
Cambridge, NM 02139-4307

Peter A. Ward, Ph.D.
Godfrey D. Stobbe Professor of Pathology
Chairman, Department of Pathology
University of Michigan Medical School
1301 Catherine Street
Ann Arbor, MI 48109-0602

Michael A. Marletta, Ph.D.
John G. Searle Professor of Medicinal
Chemistry and Phalmacognosy
College of Pharmacy
The University of Michigan
428 Church Street
Ann Arbor, MI 48109-1065

William Pryor, Ph.D.
Director, Biodynamics Institute
Louisiana State University
711 Choppin Drive
Baton Rouge, IA 70806

Curtis C. Harris, Ph.D.
Associate Chief
Laboratory of Human Carcinogenesis
National Cancer Institute
Bldg. 37, Rm. 2AO7
Bethesda, MD 20892

Larry K. Keefer, Ph.D.
Chief, Chemistry Section
Laboratory of Comparative Carcinogenesis
NCI-Frederick Cancer Research and Development Center, Bldg. 538 Rm. 205-E
Frederick, MD 21701

JAPAN

Hiroyasu Esumi, Ph.D
Deputy Director, Research Insitute and Chief, Biochemistry Division
National Cancer Center Research Institute
1-1, Tsukiji 5-chome, Chuo-ku, Tokyo 104

Hiroshi Maeda, Ph.D.
Department of Microbiology
Kumamoto University School of Medicine
Kumamoto 860

Katsuaki Shibuki, Ph.D.
Laboratory for Neural Networks
RIKEN

Setsuo Hirohashi, Ph.D.
Pathology Division
National Cancer Center Research Institute
1-1, Tsukiji 5-chome, Chuo-ku, Tokyo 104

Kouji Matsushima, Ph.D.
Departnent of Pharmacology
Cancer Research Institute Kanazawa University
13-1 Takara-Machi, Kanazawa 920

Tomoyuki Shirai, Ph.D.
First Department of Pathology
Nagoya City University Medical School
Kawasumi, Mizuho-cho, Mizuho-ku, Nagoya 467

Mitsubiro Tsuda, Ph.D.
Department of Toxicology
National Institute of Hygienic Sciences, Tokyo



(2) Seminar on "Viral Regulatory Proteins and Host Factors Regulating Gene Expression"
MAUI, HAWAII, JANUARY 26-27, 1993

AGENDA

PARTICIPANTS

JAPAN

Mitsuaki Yoshida
Institute of Medical Science
University of Tokyo
4-6-1 Shirokanedai, Minato-ku, Tokyo 108

Dr. Yoshiaki Itoh
Institute of Virus Research
University of Kyoto
Shougoin Kawahara-cho, Sakyo-ku, Kyoto 606

Dr. Akira Hakura
Research Institute of Microbial Diseases
Osaka University
3-1 Yamadaoka, Suita 565, Osaka

Dr. Tetsu Akiyama
Research Institute of Microbial Diseases
Osaka University
3-1 Yamadaoka, Suita 565, Osaka

Dr. Masami Horikoshi
Institute of Applied Microbiology
University of Tokyo
1-1-1 Yayoi, Bunkyo-ku, Tokyo 113

Dr. Kunitada Shimotohno
National Cancer Center
5-1-1 Tsukiji, Chuo-ku, Tokyo 104

Dr. Shizuo Akira
Institute for Molecular and Cellular Biology
Osaka University
1-3 Yamadaoka, Suita 565, Osaka

Dr. Masuo Yutsudo
Associate Professor
Osaka University
3-1 Yamadaoka, Osaka

Dr. Nobuyuki Tanaka
Assistant Professor
Osaka University
3-1 Yamadaoka, Osaka

UNITED STATES

Dr. Peter M. Howley
Laboratory of Tumor Virus Biology
National Cancer Institute
Building 41, Room C-111
Bethesda, MD 20892

Dr. Carol Prives
Department of Biological Sciences
713 Fairchild Center
Columbia University
New York, NY 10027

Dr. James Manley
Department of Biological Sciences
713 Fairchild Center
Columbia University
New York, NY 10027

Dr. Ed Harlow
MGH Cancer Center
MGH East-74149 13th Street
Charlestown, MA 02129

Dr. Inder Verma
The Salk Institute
Tumor Virus Laboratory
P.O. Box 85800
San Diego, CA 92138-9216

Dr. Warner Green
Gladstone Institute of Virology
San Francisco General Hospital
P.O. Box 419100
1001 Potrero Avenue
Building 3, Room 545
San Francisco, CA 94141-9100



(3) Seminar on "Topics in Carcinogenesis: Malignant Neoplasms Occurring Under Particular Circumstances"
March 3-5, 1993, Maui Marriott, Lahaina, Maui, Hawaii

PROGRAM