(1) Seminar on “Helicobactor and Gastric Cancer”
Chronic atrophic gastritis with intestinal metaplasia is thought to be a precancerous condition of gastric cancer on the basis of histological and epidemiological evidences. Although this concept has not been fully accepted, we agree that chronic atrophic gastritis and intestinal metaplasia have a close relation to gastric cancer. The causes of gastric atrophy have not been fully elucidated, but both environmental factors and age-dependent genetic factors have been suggested as candidates.
Since Helicobacter pylori (H.pylori) was isolated from the gastric mucosa of the patients with gastritis in 1983, the concept of human gastritis has changed rapidly and, the association of H.pylori infection with not only benign but also with malignant gastric diseases has been reported. In June 1994, the International Agency for Research in Cancer, an arm of the World Health Organization, reported that “H.pylori is a group 1 (definite) carcinogen” based on the epidemiological studies. However, the role of H.pylori in the mechanism of gastric cancer genesis is not clear. The development of primary gastric malignant lymphoma has also been linked with H.pylori. Besides the positive results in case-control studies, the recent evidence on the regression of MALT (mucosa associated lymphoid tissue) lymphoma produced by H.pylori eradication strongly supports their relation. However, this mechanism is not clear, either.
This seminar was held on February 12-14, 1996, at the Marriott Hotel, Maui, Hawaii, to discuss in depth a topic of current scientific significance on the association between Helicobacter and gastric malignant diseases. The organizers were Dr. Jerry M. Rice, National Cancer Institute, Frederick, and Dr. Daizo Saito, National Cancer Center Hospital, Tokyo. There were eight participants from the United States and six participants and two observers from Japan.
Dr. Rice welcomed the attendees and Dr. Takashi Sugimura, coordinator of the US-Japan Cancer Cooperative Research Program, outlined the development and the important role in cancer research of the US-Japan program up to date. Dr. Saito gave the introductory remarks and added that the studies concerning other kinds of Helicobacter species besides H.pylori might be useful to elucidate the association between H.pylori and gastric cancer.
The first speaker, Dr. Pelayo Correa provided an overview of the evolution of the rationale for the idea that H.pylori may lead to gastric neoplasia. Because the great majority of gastritis patients never develop neoplasias, he explained the H.pylori-associated gastritis at first. In most populations, H.pylori infection leads to nonatrophic gastritis, predominantly involving diffuse antral gastritis (DAG), the basic lesion seen in patients with duodenal ulcer, which has not been associated with increased risk for gastric carcinomas. On the other hand, in populations with high gastric cancer risk, H.pylori infection is associated with multifocal atrophic gastritis (MAG), which frequently advances to intestinal metaplasia, and dysplasia. Next, he showed that the gastric juice ascorbic acid concentration was significantly lower in black compared with white patients and blacks had higher prevalence of H.pylori infection, higher gastric pH, more severe acute and chronic inflammation of the gastric mucosa, and higher frequency of Lewis (a-b-; negative for both Lea and Leb) phenotype. He stated that these observations might help to explain the high incidence of gastric cancer among the black population in southern Louisiana. Furthermore, he reported on the data obtained from the case-control study conducted in New Orleans to elucidate causal association for H.pylori infection and chronic gastritis. Of the 321 patients included in this analysis, 240 patients (75%) were infected with H.pylori. A positive association was found between H.pylori infection and African-American ethnicity (odds ratio: OR = 3.64) as well as smoking (OR = 1.69). A strong interaction between race and smoking was found (OR = 6.32). Adequate intake of antioxidant micronutrients, especially vitamin C, was negatively associated with infection prevalence. Namely, the subjects in the highest quartile of vitamin C intake had an estimated relative risk of 0.38 compared with subjects in the lowest quartile, when the continuous dietary vitamine C indecies were categorized into 4 quartiles. H.pylori infection was strongly associated with an increased risk of atrophy (OR = 6.4) and intestinal metaplasia (OR = 4.7) of the gastric mucosa.
From these data, Dr. Correa concluded as follows: An etiologic hypothesis has been developed that postulates several etiologic factors acting at different points in the chain of events. Helicobacter pylori infection may play a role at three different points: (1) in the causation of chronic gastritis, (2) in interfering with the normal gastric secretion of ascorbic acid, and (3) in the attraction and activation of polymorphonuclear leucocytes (PMNs) and macrophages. It is now generally accepted that H.pylori is the main cause of chronic active gastrits and the attraction and activation of PMNs and macrophages are prominent features of H.pylori infection. Two events associated with chronic active gastritis may be implicated in the carcinogenic process, namely the state of hyperproliferation that potentiates the effects of carcinogens, and the damage to the foveolar epithelium that results in loss of the protection provided by mucin against luminal carcinogens. H.pylori gastritis may result in atrophy, and it may elevate of the gastric pH and decrease the concentration of ascorbic acid in the gastric lumen. Because the antioxidant effects of ascorbic acid are compromised [in the way], intake of vitamin C may play a role in preventing gastric atrophy. Although the role of H.pylori in inducing intestinal metaplasia in the case of atrophy is clear and strong in our study, the mechanism is still unknown. Although H.pylori has not been also shown to contain any mutagens, they may produce nitric oxide and hydroxyl radicals capable of inducing mutations in the DNA molecules.
Dr. Norio Matsukura discussed how H.pylori infection plays a role leading to gastric cancer before the appearance of intestinal metaplasia based on his study He investigated whether H.pylori infection is linked in the initial stage of the precancer-cancer sequence. He detected intestinal metaplasia in resected specimens after gastrectomy for gastric cancer by the Tes-Tape method biochemically and measured tissue IgA antibody against H.pylori by ELISA. Tissue H.pylori IgA antibody was positive in 6 of 19 (32%) specimens taken from complete and 2 of 7 (29%) incomplete types of intestinal metaplasia and was positive in 6 of 14 (43%) nonmetaplastic gastric mucosa from the antrum and 14 of 23 (61%) from the body. Duodenal mucosa and cancer tissue were positive for tissue IgA antibody in 1 of 6 (17%) and 0 of 17 (0%), respectively. That is, positive rate of tissue H.pylori IgA antibody was 60% in the mucosa of non-metaplastic gastric mucosa, 30% in the mucosa of intestinal metaplasia and was negative in cancer.
The first real evidence linking H.pylori and gastric cancer came in 1991 from three cohort (nested case-control) studies conducted by Forman (UK), Parsonnet (California), and Nomura (Hawaii). All three reached the same conclusions and the ORs for gastric cancer in the individuals with preceding H.pylori infection were 2.8, 3.6, and 6.0.
Dr. Haruhiko Fukuda reported the results of a case-control study in Tokyo conducted to confirm the effect of H.pylori infection on the risk of gastric cancer and discussed the underestimation of its risk in cross-sectional studies. Four hundred and ninety-nine gastric cancer patients were gastrectomized from 1989 to 1990 in the National Cancer Center Hospital. Of these, 428 patients without multiple gastric cancers or a history of gastrectomy, for whom stored serum samples were available, were selected for the potential case pool. Meanwhile, 1,925 cancer-free patients were selected out of 18,361 out-patients during the same period, in accordance with the following criteria: no cancerous lesions detected clinically in any organ, stored serum available, a visit within the previous 3 to 6 months, and no history of hospitalization in our hospital. Controls were matched to each case for sex, age (within 3 months), and date of blood sampling (within 3 months) until the available control pool was exhausted. He obtained 297 matched sets with at least one control, and up to 16 controls (297 cases and 786 controls). Of the 297 matched case-control sets, the stored serum samples from 282 cases and 767 controls were tested for the presence of anti-H.pylori lgG antibody (HM-CAP ELISA kit) and serum pepsinogen (PG) level (PG I and PG II RIABEAD). Matched analyses were performed using the Mantel-Haenszel test. Odds ratio and 95% confidence intervals (CI) were determined in logistic model. All calculations were performed using SAS software (SAS Institute Inc.). No significant association was observed in all set [OR= 1.04, 95% CI: 0.073-1.49]. In subgroup analysis, however, an association was suggested in females [OR=1.57], a younger population (<50years)[OR=1.86], early cancer [OR=1.53] and small cancer (<40mm)[OR=1.55]. These data suggested that H.pylori infection was a positive risk in one subgroup and a negative risk in others. If H.pylori infection is a risk factor in the early stage of cancer, it should be associated with a positive risk in the more advanced stages of cancer. Furthermore, a tendency for OR to decrease with the increase in age or cancer growth (depth of tumor invasion and tumor size) was observed. Therefore, these strange results can not be interpreted unless some factor(s) which decreases the antibody titer against H.pylori in proportion to cancer growth are taken into account. Some investigators have pointed out that extended atrophy may cause spontaneous disappearance of H.pylori from the stomach. Therefore, Dr. Fukuda hypothesized that mucosal atrophy extending in parallel with cancer growth causes us to underestimate the real cancer risk associated with H.pylori infection. Accordingly, he used the conditional logistic regression model to adjust for the negative effect of advanced gastric atrophy. After adjustment for the grade of gastric atrophy, gastric cancer patients showed a significant higher odds ratio of 1.69 (95% CI: 1.01-2.81). In particular, the intestinal type of cancer showed a strong association with H.pylori infection compared to the diffuse type (OR: 3.76 vs 1.14).
Dr. Fukuda concluded that H.pylori infection increases the risk of gastric cancer and the risk should be underestimated in studies with cross-sectional exposure, because of spontaneous disappearance of H.pylori due to extended mucosal atrophy. Furthermore, he pointed out that it may also explain the discrepancy between the high risk showed by some prospective studies in Western countries and no risk showed by some retrospective studies in developing countries.
Dr. Julie Pasonnet reported that H.pylori infection and a low PG I were associated with a marked increase in the risk of developing distal cancer (OR=10.0; p=0.08) on her study and pointed out the importance of the combination assay of PG I and anti-H.pylori IgG to survey the high-risk patients.
Although the Dr. Fukuda’s study and the previous three cohort studies strongly suggested a close relationship between H.pylori infection and gastric cancer, they did not provide direct evidence. Dr. Saito pointed out that, although the conclusion of WHO/JARC, i.e., “H.pylori is a Group 1 (definite) carcinogen” is based on these epidemiological studies, these are not intervention studies to demonstrate “causal link” but observation studies which affirm the “association” between them and a large-scale and long-term intervention trial in which H.pylori eradication is tried from young age and the long-term course after that is precisely followed-up is required, to clarify “whether H.pylori is a real risk factor of gastric cancer genesis.”
Dr. Saito reported on an on-going intervention trial in Japan, which is supported by a Grant-in-Aid for the Second Term Comprehensive 10-Year Strategy for Cancer Control from the Ministry of Health and Welfare, Japan. The present study is designed to compare the development and progress of gastric mucosal atrophy and incidence of gastric cancer in patients 20 to 59 years old without gastric localized lesions randomized into H.pylori eradicated or uneradicated group. In this trial, the patients with only ulcer scars are also excluded from the subject. For H.pylori eradication, a one week triple-drug therapy with Lansoprazole (30mg/day), Clarithromycin (400mg/day), and Amoxicillin (1,500mg/day) is used. When the evaluation after 6 months reveals the failure for eradication, the same regimen is used for 2 weeks. When it also fails, the combination therapy with other drugs (Plaunotol, Ecabeto-Na, Sofalcon, etc.) will be used. Some reports [Krause JR et al. Am J Gastroenterol 80: 978-982, 1985, Sloan DA et al. Am J Surg 145: 66-69, 1983] suggest carcinogenicity of Metronidazole which is used most frequently in US and Europe for H.pylori eradication. Accordingly, Dr. Saito considered its use in this trial as not appropriate. The basic method for H.pylori diagnosis uses endoscopic biopsy specimens taken from the greater curvature of the body and antrum of the stomach. One is a test by culture and another is microscopic test. Serum anti-H.pylori antibody is just for reference. The atrophic gastritis is evaluated endoscopically and histologically, with therapeutic regimen of the subjects masked. Serum pepsinogen level is also just for reference. The judgment of H.pylori infection and eradication and diagnosis of mucosal atrophy is done at one place. Next, he calculated how many samples are necessary for this study. Taking many factors, such as H.pylori infection rate in Japan, re-infection rate after eradication, eradication rate by the treatment, incidence and progression rate of atrophic gastritis, risk ratio of gastric carcinogenesis in H.pylori infected patients, and gastric cancer incidence in general population into account, he estimated the sample size. The required sample size is 102 to 736 cases when the progress of mucosal atrophy is endpoint and 2,342 to 7,458 cases when the comparison of gastric cancer incidence is endpoint. The evaluation on gastric mucosal atrophy will be made in the fourth year after the end of the enrollment period and the evaluation on gastric cancer after the end of the 8 year follow-up period. Finally, Dr. Saito described that if the present study, in which H.pylori-positive subjects are randomized into H.pylori eradicated or uneradicated group, can clarify the causal link between H.pylori infection and atrophic gastritis, generally considered to be precancerous condition of gastric cancer, and gastric cancer, it will lead to the establishment of preventive measures for gastric cancer with an extremely high mortality in Japan and be quite significant.
To conclude that H.pylori is a real risk factor of gastric cancer genesis, we have to wait for the results obtained from the above mentioned the long-term intervention study However, the strong association of H.pylori with atrophic gastritis and acceleration of proliferative cell kinetics in the gastric epithelial cells led to the speculation that this organism may well play a role in the carcinogenic process.
Dr. Toshio Fujioka reported on successful experimental infection of animals with H.pylori and its effects on cell proliferation. He observed the H.pylori-associated changes in the gastric mucosa of Japanese monkeys (Macaca fuscata:six infected and seven controls) up to 5 years after inoculation serially The mixture of bacterial strains used in his experiment were made up of OMU92189 and OMU92237 isolated from two patients with gastric ulcers, and of OMU92244 and OMU92070 isolated from two patients with duodenal ulcers. These strains were identified as H.pylori by morphological, biochemical, and DNA restriction endonuclease analysis. After inoculation with H.pylori, persistent colonization and gastritis were confirmed by culture and histological examination using the large biopsy specimens obtained by the endoscopic mucosal resection technique. The severity of gastritis was evaluated according to a scoring system by Rauws et al. [Gastroenterology 94: 33-40, 1988]. Endoscopic mucosal resections from the antral mucosa of both infected and control animals were repeated every year for histological examination and in-vitro Ki-67 immunohistochemical staining. The grandular height and the length of Ki-67 positive cells (proliferative zone) were used as markers of gastric mucosal atrophy and epithelial cell proliferation, respectively. Dr. Fujioka found that all of the infected animals revealed histological gastritis throughout the entire investigation period. In contrast, all the controls remained culture negative without histological evidence of gastritis. More in detail, a remarkable peak in the gastritis score was shown 1 week after inoculation and it decreased gradually. At all sampling times, however, the gastritis score of the infected group was significantly higher than that of the control group, confirming the persistence of histological gastritis throughout the 5 years. As atrophic change, it was reported that the pyloric glandular height was significantly lower (p<0.05) in the infected group than in the control group after 6 months of inoculation and the maximal decrease of glandular height was shown after 1.5 years of H.pylori infection (p<0.05,). On the other hand, cell proliferation in the antral mucosa of infected group was significantly accelerated throughout the entire observation period (p<0.01 and p<0.05 at 1 and 2-5 year periods, respectively). In contrast, no change was found in the control animals.
In his animal model, the glandular height in the antral mucosa significantly decreased and proliferative zone significantly increased in the infected animals as compared with controls. Based on these experimental results, Dr. Fujioka suggested that the gastric epithelial cell proliferation caused by H.pylori infection correlates closely to the progression of atrophic gastritis, and that it may explain the potential mechanism for a causal role in the chain of events leading to gastric carcinoma.
It is well known that H.pylori is highly diverse at the genetic level. The various strains share many structural and biochemical characteristics, but they are not all equally virulent. The cytotoxin is known to correspond to a protein of 87 kDa in which the gene has been cloned (vacA) and may be present in all strains. There is also a cytotoxin-associated protein (120-128kDa) in which the gene has also been cloned (cagA). It is present in about 60-80% of the strains. This cagA product is not the cytotoxin itself, but it is associated with vacuolating cytotoxin activity. In terms of heterogeneity, cagA is the first gene among the 12 or more cloned genes known and is found in most strains.
Dr. Parsonnet introduced at first the Hawaiian study by Blaser et al. arguing that infection by cagA+ H.pylori was associated with a double risk of gastric cancer. They matched 103 H.pylori-infected men who developed gastric cancer during a 21-year surveillence period with 103 H.pylori-infected men who did not develop gastric cancer and tested stored serum specimens from the patients and controls for the presence of serum IgG to the cagA product of H.pylori using ELISA. In the men with antibodies to cagA, the OR of developing gastric cancer was 1.9 (95% CI: 0.9-4.0); for intestinal type cancer of the distal stomach, the OR was 2.3 (95% CI: 1.0-5.2).
Next, Dr. Parsonnet reported on a nested case-control study in California which indicates cagA gene as a marker of increased risk of atrophy, metaplasia, and cancer. In that study, she evaluated whether cagA seropositivity (a marker of strain type), serum pepsinogen (PG) I or gastrin level (markerfrom H.pylori infected subjects were tested by ELISA for IgG antibodies against the cagA gene product of H.pylori, a marker for the type I H.pylori phenotype (cagA+). Sera had previously been tested for PG I ; the corpus atrophy was judged with low PG I (<50mg/ml). Stored sera were tested for gastrin using a commercial double antibody 125I RIA (DPC, Los Angeles, CA). After adjusting for age, sex, and race, infected cases were 3.2-fold more likely than infected controls to have antibodies to the cagA gene product (95% CI = 1.6-6.5). This association was significant for intestinal-(OR = 3.7, 95% CI = 1.7-8.3) but not diffuse-type cancers (OR- = 2.2, 95% CI = 0.8-6.4). Corpus atrophy significantly increased risk for both cancer types and lessened the magnitude of association between cagA and cancer by 25%. When analyzed as a continuous, long-transformed variable, there was no association between gastrin and cancer. Among persons with corpus atrophy, however, gastrin levels in the lowest quartile (<26.6 pg/ml) appeared to be protective against gastric cancer (OR = 0.1, p value for interaction = 0.08). In persons with normal PG levels, low gastrin was not protective against cancer (OR = 1.0). When compared to 70 uninfected subjects, infection with the Type II phenotype (cagA-) of H.pylori was only modestly and not significantly associated with cancer (OR = 2.0, 95%CI = 0.8-4.6). Neither educational level nor ABO blood group was associated with malignancy in infected persons.
Dr. Parsonnet concluded that, among persons infected was accompanied with the development of intestinal metaplasia. At the end of follow-up, 15 (62%) of the 24 cagA+ subjects had atrophic gastritis compared with 11 (32%) of the 34 cagA-subjects (P = 0.02; Fisher’s exact test, OR = 3.48; 95% CI = 1.02-12.18). Based on these data, he concluded that the infection with cagA+ H.pylori strains is associated with an increased risk for the eventual development of atrophic gastritis and intestinal metaplasia.
In addition, Dr. Blaser outlined the new picA and picotein and the IL-8 production was inhibited by knocking-out of these genes.
Furthermore, Dr. Blaser evaluated the hypothesis that early life acquisition of H.pylori is a risk factor for the development of gastric diseases. On earlier nested case-control studies of a cohort of Japanese-American men in Hawaii, evidence of H.pylori infection was associated with the development of gastric cancer or gastric or duodenal ulceration during the subsequent period, 1968-1989. The present analysis included increased risk of gastric (OR = 1.64) but not duodenal ulcers. He concluded that these data are consistent with the hypothesis that early life acquisition of H.pylori increases the risk of developing both gastric cancer and gastric ulcer but not duodenal ulcer.
Dr. Matsukura discussed different factors related to the outcome of H.pylori infection, such as the immune response of the host, the detoxifying enzymes, the molecular events, and the H.pylori strain diversity.
Exposure to N-nitivity of serum IgG antibody against H.pylori (OR = 1.25, 95% CI: 0.84-1.85) and specific genotypes of L-myc polymorphism (OR = 0.33, 95% CI: 0.59-2.99) were more commonly observed in gastric cancer cases, but without statistical significance. Specific genotypes of the CYP21EI Rsal and GSTMI gene deletion were not associated with gastric cancer. Dr. Matsukura concluded that atrophic mucosal change detected by serum markers was the most important risk factor for the gastric cancer. H.pylori infection and genetic polymorphisms of CYP21EI, L-myc and GSTMI gene deletion did not affect gastric carcinogenesis in this study.
Next, he investigated microsatellite instability in the mucosa of intestinal metaplasia detected by the Tes-Tape method and in the cancer tissue. DNA-replication errors (RERs) and loss of heterozygosity (LOH) were investigated in 3 microsatellite loci 2p (D2S123), 3p (D3S1067) and 17P (TP53). He found 3/43 (7%) of RERS and 20/43 (47%) of LOH in the gastric cancer tissue. However, microsatellite instability was found in low incidence (1/35 in complete and 1/16 in incomplete type of intestinal metaplasia) in the mucosa of intestinal metaplasia.
Finally, he reported on the strain diversity of H.pylori among gastric diseases in the Japanese population. He detected the cytotoxin gene (cagA and vacA) in the H. pylori DNA by PCR method and conducted an age and sex matched case-control study among the patients with chronic gastritis, gastric cancer, and peptic ulcer. He however could not find any significant difference among these 3 diseases in positivity of cytotoxin genes and indicating cytotoxin genes in H.pylori are not major factors for gastric carcinogenesis in Japan.
Dr. Toshiro Sugiyama focused on two genes, cagA gene and catalase gene, to investigate the H.pylori strain diversity. He designed two sets of primer to detect the non-repeated regions of cagA gene and PCR was performed. PCR products of cagA gene were detected in 58% of 24 chronic gastritis strains, in 70% of 10 gastric cancer strains, 90% of 19 gastric ulcer strains, and in 86% of 15 duodenal ulcer strains. CagA gene was significantly more present in gastric ulcer strains and duodenal ulcer strains. He performed an RFLP (restriction fragment length polymorphism) analysis of Hae-III digested H.pylori DNA with cagA as a probe to investigate the diversity between gastric ulcer and duodenal ulcer stains. He presented that 9 of 10 duodenal ulcer strains (90%) showed 3000 bp fragment band (cagA-l) but that 7 of 10 gastric ulcer strains (70%) had 900 bp fragment band (cagA-s) reported the probable exisence, unique restriction sites in the surrounding regions of cagA gene, which will be a useful marker to distinguish gastric ulcer strains.
Catalase of H.pylori is essential to escape peroxidation of neutrophils and to interact with neutrophils in the gastric mucosa. He had prepared the anti-H.pylori monoclonal antibody CP2, clinically used in immunohistochemistry and sero-diagnosis and cloned the cod by endoscopy and the endoscopic mucosal resection (EMR) is conducted in many intestinal type of small gastric cancer cases. Since the background gastric mucosa in such cases is characterized by chronic atrophic gastritis accompanied by advanced intestinal metaplasia, the need to pay attention to synchronous multiple cancer is well known. Also known is the incidence of metachronous multiple cancer at 5-10% within 5 years of follow-up in Japan. In other words, the residual gastric mucosa after EMR is a high-risk microenvironment regarding the incidence and growth of intestinal type of gastric cancer.
Dr. Naomi Uemura presented preliminary evidence of prevention of new gastric cancers after H.pylori eradication in patients with early gastric cancer resected endoscopically. He conducted H.pylori eradication treatment in 65 of the 132 H.pylori-infected patients (44-85 years old, mean age: 69 years old; 97 males and 35 females) whose gastric cancer was resected by EMR, and investigated the rved for 48 months at the longest and an average of 24 months to investigate and compare the occurrence of new gastric cancer in other regions after EMR. In the H.pylori eradicated 65 patients, Dr. Uemura observed not only the disappearance of neutrophil infiltration in the antrum and body of the stomach but also significant remission of intestinal metaplasia and detected no new gastric cancer. On the other hand, in 6 (9%) of the 67 uneradicated patients, he detected a new early stage intestinal type of gastric cancer endoscopically within 3 years after EMR. There was no difference between these groups in the factors including age, sex, histological type of gastric cancer, region, and the ratio of homochronic multiple cancer. In this way, he found that H.pylori eradication treatment improved neutrophil infiltration and intestinal metaplasia and the treatment might inhibit the growth of cancer in the initial stage of intestinal type of gastric cancer.
In conclusion, Dr. Uemura stressed the necessity of 1 new out-patient during the same period as the control pool. The control subjects had no cancer diagnosed clinically or no history of admission. Finally, he analyzed 30 age and sex matched case-control sets and showed the positive association between H.pylori and malignant lymphoma (OR = 2.76, 95% CI: 0.81-9.44). Next, he examined the histological, endoscopic, and genetical changes after H.pylori eradication in 9 cases of LG-MALT and 1 of atypical lymphoid proliferation (AL) as borderline lesion. They were treated together with Lansoprazole (30mg/day), Plaunotol (240mg/day), and Amoxicillin (1500mg/day) in combination for 2 weeks and the treatment with Lansoprazole and Plaunotol was extended for further 2 weeks. At all endoscopic examinations before and after the treatment, 2 biopsy specimens were taken from the greater curvature of the antrum and the body for H.pylori culture, at least 5 specimens from tumor area for histology and 2 for molecular genetic analysis. After the treatment, this procedure was repeated every 3 to 6 month during the follow-up period. Evaluation of H.pylori eradication was carried out by culture basically. In regard to histological evaluation, MALT lymphoma was defined as the lesions which have diffuse and abundant centrocyte-like cells infiltration and prominent lymphoepithelial lesions according to Wotherspoon and Isaacson’s criteria. Histological findings changed to chronic gastritis in 5 with LG-MALT and 1 with AL and endoscopic findings were improved in 4 out of these 6 cases. In 2 cases, monoclonal pattern of IgH rearrangement by PCR changed to polyclonal pattern after eradication. Although 1 with LG-MALT was worsened endoscopically and underwent gastrectomy, there was no lymph node metastasis and no involvement to other organs was seen. In addition, he reviewed the natural course of 8 MALT lymphomas [1 high-grade MALT (HGMALT), 7 LG-MALT] and 4 of AL, which had been followed as reactive lymphoid hyperplasia (RLH) with mean observation period of 3 years ranging from 2 months to 7.5 years, retrospectively. Although 4 cases of LG-MALT had no obvious histological changes, remaining 4 cases of MALT lymphomas and all of AL cases showed rapid histological change to chronic gastritis.
Based on these data, Dr. Ono presented that H.pylori infection is highly associated with the gastric MALT lymphoma and, therefore, H.pylori eradication would be a valuable approach for precise diagnosis and therapy of LG-MALT and lymphoproliferative disorders of the stomach. In addition, he pointed out that this favorable clinical behavior of gastric LG-MALT lymphomas indicated that they are not truly malignant neoplasm but “pseudolymphomas,” or that they are an acquired reaction to H.pylori infection. Finally, Dr. Ono proposed the guideline of clinical management of MALT lymphoma at this point. For H.pylori-infected patients who are suspected to have lymphoproliferative malignancy and who have no metastasis except to the stomach, H.pylori eradication should be performed. Then if endoscopic biopsied specimens reveal chronic gastritis histologically, it is recommended to continue the close follow-up. It is very difficult how we manage if MALT lymphoma remains in spite of the cure of H.pylori infection or if the patient is H.pylori-negative from the beginning. Certainly, since the natural course of MALT lymphoma is not well known, we have to give careful consideration to the relapse or systemic metastasis. We consider that it is worth attempting to continue close follow-up, because surgical gastrectomy deteriorates the quality of life. In the case of HG-MALT, the surgical treatment should be performed.
The development of effective immunization strategies to prevent or control H.pylori infection in humans has received increasing attention in the past few years. Exploration of this approach seems particularly worthwhile, since there is no evidence, to date, indicating that currently available antibiotic treatment protects individuals from infection. Moreover, the possible emergence of antibiotic-resistant strains of H.pylori lends attractiveness to the immune approach. Recently, a number of laboratories have begun to study the feasibility of prophylactic and therapeutic vaccination for the prevention and treatment of H.pylori infection using animal models. In addition, many Helicobacter species other than H.pylori have been isolated up to now; and studies have been conducted on these non-H.pylori species themselves, on the prevention of H.pylori infection using these species, and on the host factors in the association between H.pylori and human gastric disease.
Dr. Steven Czinn discussed the host response to Helicobacter infection and the status of vaccine development efforts. He examined whether an oral immunization protocol could prevent Helicobacter infection using H.felis infection of mice as a model system. Germ-free mice were orally immunized 4 times over 1 month with 2 mg of sonicated H.felis plus 10(g of cholera toxin (a known mucosal adjuvant). One week after the final immunization, control (non-immunized) and immunized animals were challenged with viable H.felis and sacrificed at various time intervals. Gastric biopsy specimens were tested for urease activity and cultured for the presence of H.felis. Serum and gastric and intestinal secretions were collected and assayed by ELISA for the presence of anti-H.felis antibodies. In this study, he found that a significant IgA and IgG anti-Helicobacter response was generated in serum and gastric and intestinal secretions in immunized mice as compared to non-immunized controls. Additionally significant protection [82% (28/34), vs 11% (4/36)] was observed when the mice were immunized prior to challenge with viable H.felis. He also showed the data regarding successful therapeutic immunization against Helicobacter infection. In that study, ferrets which had been chronically infected with H.mustelae were cured of Helicobacter infection (7/23) following oral immunization with H.pylori urease and cholera toxin. He described that if these experiments in mice and ferrets continue to yield consistent results, it could lead to immunotherapy for eradication of chronic H.pylori infection of humans.
Natural infection or immunization results in a systemic and mucosal immune response of equivalent magnitude. There appears to be a qualitative difference in the immune response following immunization, which results in protection. Specifically, antibodies against urease, and perhaps other antigens, appear to be prominent following immunization, and may play a key role in prevention or protection from Helicobacter infection. Furthermore, he added the usefulness of heat-labile toxin of enterotoxigenic Escherichia coli (LT) as an adjuvant for immunization. He indicated the necessity of further study to determine its role in protection and/or pathogenesis of Helicobacter infection.
Finally, Dr. Czinn described the cellular immune response following infection and immunization. He observed two types of cell-mediated immune responses depending on the nature of the antigen preparation. The first response is a Helicobacter-independent response, present in all experimental groups, which is directed toward antigens such as urease and heat shock proteins. The second is a Helicobacter-dependent cellular response restricted to mice previously exposed to Helicobacter antigens either by immunization or infection. This response was not seen in noninfected controls. The Helicobacter-dependent cellular response had a Th1 phenotype, as in either infected or immunized/challenged mice. From these data, he presented that Helicobacter infection and/or immunization stimulate a predominantly Th1 type, and this Th1 response promotes a delayed-type hypersensitivity response in the stomachs of mice. Furthermore, suppression of the Th1 response in the immunized mice unmasks the presence of a subpopulation of Th2 cells. However, he emphasized that further study is needed to determine its role in protection and/or pathogenesis of Helicobacter infection.
Dr. Rice reported on a new species of Helicobacter associated with a high incidence of hepatocellular neoplasms in infected animals. In the autumn of 1992, a novel form of chronic active hepatitis of unknown etiology was discovered in mice at the National Cancer Institute, Frederick Cancer Research and Development Center (NCI-FCRDC), Frederick, Md. A remarkably high incidence of hepatocellular neoplasms including both adenomas and carcinomas, often multiple, developed in affected animals by 64-77 weeks of age. The disease entity was originally identified in A/JCr mice that were untreated controls in a long-term toxicologic study. By the use of a special strain, Steiner’s modification of the Warthin-Starry stain, it was found by light microscopy that formalin-fixed sections of liver tissue from mice with hepatitis contained bodies of the size of bacteria that were distributed singly within hepatic parenchyma. The organism was identified as a novel species of the genus Helicobacter on the basis of its morphologic and biochemical characteristics and the base sequence of its 16S rRNA gene. The bacterium is motile and gram negative, 0.2 to 0.3(m in diameter, 1.5 to 5.0(m long, and curved spirally in shape, with one to several spirals. It has bipolar sheathed flagella, one at each end, but lacks the periplasmic fibers that envelop the bacterial cells in other mouse Helicobacter species. Because of its association with inflammatory and neoplastic liver disease, it was named Helicobacter hepaticus. Transmission electron microscopy has revealed that it is localized preferentially in bile canaliculi in infected livers; it has not been found within hepatocytes. It has been isolated thus far only from mice, ü@, and not from other rodent species including rats, guinea pigs, and Syrian hamsters. It appears to cause liver disease only in mice of certain strains, including A, DBA, and C3H, and to colonize the cecum and colon without invading the liver in certain other strains, notably C57BL/6. Male mice of susceptible strains are much more highly susceptible than females. While single-agent antibiotic therapy is ineffective in eradicating the bacteria from the gastrointestinal tract, either amoxicillin or tetracycline, in combination with metronidazole and bismuth, has been shown to eradicate H.hepaticus from the gastrointestinal tract when given by oral gavage for a period of two weeks.
In this way, the Helicobacter associated chronic active hepatitis represents a new model to study mechanisms of carcinogenesis by this genus of bacterium. Dr. Rice described H.hepaticus infection in mice constitutes the only other parallel association between a persistent bacterial infection and tumor development known to exist naturally and a study on the association between chronic infection of the liver by H.hepaticus and development of hepatocellular tumors in mice may yield insight into the role of H.pylori in human gastric cancer and lymphoma.
Dr. Lucy M. Anderson described oxidative damage to DNA in the livers of mice infected with H.hepaticus. To investigate possible mechanisms regarding liver carcinogenesis by H.hepaticus, as a starting point, she hypothesized that the liver tumors arising in the mice infected with H.hepaticus were caused by a product of the bacteria, either elaborated by them or formed by bacterial enzymes acting on an endogenous substrate; and/or by products of the responding liver or inflammatory cells. At first, she studied the genotoxic damage due to nitric oxide released during response to the infection caused genotoxic damage, either directly by deamination of 5-methylcytosine at CpG sites in DNA, or indirectly by nitrosation of endogenous substrates to generate carcinogenic nitrosamines. However, several tests of this hypothesis have given rather definitively negative results. The second hypothesis has been based on an argument that increased reactive oxygen species (ROS) have led to tumor development and oxidative damage to DNA used as a marker for increased ROS production. The promutagenic ROS product, 8-OH-dG, was found to be significantly elevated in the infected livers from an early stage, with a further increase as the hepatitis progressed. A significant 3-fold increase in 5-OH-5-methylhydantoin, a derivative of thymine, was also detected. This increased ROS could come from the hepatocytes, or from infection-stimulated Kupffer or inflammatory cells. Certain cytochrome P450 isoforms, which can be major sources of ROS, are known to increase in hepatocytes in other liver infection/tumorigenesis models. In a test for such changes, we found significant increases in P450s Cyp1a1 and 1a2 (by enzyme assay and immunohistochemistry) early in the infection, and in 2a5 at the time of tumor development. ROS have also been implicated as a part of the nongenotoxic mechanism of tumor promotion in other model systems, including liver tumorigenesis models. Dr. Anderson tested whether hepaticus infection can promote tumors initiated by a carcinogen. A/JCr male mice from infected and control groups were treated with N-nitrosodimethylamine (NDMA) as neonates or with N-nitrosodiethylamine (NDEA) as young adults to initiate liver tumors. The hepaticus infection had no effect on the NDEA-initiated tumors, but has been significantly promotive for the NDMA-initiated ones. In the infected group, 10/60 were moribund before 36 weeks, vs 0/60 controls (p = 0.001), and 8 of the 10 presented malignancies other than hepatic tumors, including 5 lymphomas. For the liver, at 36 weeks the multiplicity of liver tumors was significantly greater in infected mice (4.6ü}0.8) than controls (2.4ü}0.6), p = 0.04. These results suggest that the H.hepaticus infection is promotive not only for NDMA-initiated liver neoplasms, but possibly also for those in other tissues as well.
In conclusion, Dr. Anderson indicated that the findings thus far are consistent with responses to the H.hepaticus infection, involving hepatocytes and possibly blood cells, that promote liver tumorigenesis, with ROS as one of the etiologies.
Dr. Bruce E. Dunn reported on the induction of gastric dysplasia in p53-deficient mice inoculated with H.felis. Up to 50% of gastric carcinomas are associated with loss of function of p53 tumor suppressor gene, which constitutes the most common genetic defect identified to date in this malignancy. Mice lacking functional p53 tumor suppressor gene develop normally and show no histologic abnormalities of the stomach. He examined whether gastric epithelial cell hyperproliferation induced by Helicobacter infection coupled with lack of p53 tumor suppressor gene activity would lead to development of significant pre-malignant or malignant gastric alterations. The homozygous transgenic p53 gene-knockout (TSG-p53) mice were infected with H.felis (ATCC49179, S1 approximately 5 ü~109 CFU in 0.5ml) or with Brucella broth alone (controls). He confirmed H.felis infection and foci of active gastritis in TSG-p53 mice sacrificed as early as 17 days post-inoculation. In the remaining TSG-p53 mice sacrificed 60-130 days post-infection, all infected mice showed inflammation and H.felis within gastric antral glands. Some of the infected TSG-p53 mice developed focal atrophic changes and in several of which, foci of moderate dysplasia were identified. However, such alterations were not observed in uninfected TSG-p53 mice. In this study, he demonstrated that H.felis infection is capable of inducing gastric atrophy and dysplasia in at least some TSG-p53 mice as early as 60 days post-inoculation and emphasized that this is the first demonstration of a late stage pre-malignant gastric alteration such as dysplasia in response to experimental Helicobacter infection.
Dr. James G. Fox also investigated host genetic factors and the influence of Helicobacter on cell proliferation in an inbred and p53 hemizygous mouse model. H.felis was inoculated into SPF C57BL/6 wild-type and p53 hemizygous mice that were followed up for 1 year and compared with uninfected controls of the same genotype histologically and by proliferating cell nuclear antigen (PCNA) staining and 5-brom0-2’-deoxyuridine (BrdU) analysis. Both wild-type and P53 hemizygous mice showed active chronic inflammation and marked mucosal hyperplasia 6 months after infection and severe adenomatous and cystic hyperplasia of the surface foveolar epithelium 1 year after infection compared with uninfected controls. BrdU uptake and PCNA staining were markedly increased in both sets of infected mice compared with controls. Infected p53 hemizygous mice had a higher proliferative index than the infected wild-type mice. He concluded that H.felis can induce a hypertrophic gastropathy in the C57BL/6 genotype; loss of one p53 allele, although insufficient to initiate carcinogenesis at 1 year, enhances the proliferative index, which may lead to an increased risk of cancer induction.
H.pylori is characterized by potent urease activity, thought to be located on the outer membrane of the bacterium, which is essential for survival at low pH. Dr. Dunn investigated the mechanisms whereby urease and HspB (a GroEL heat shock protein homolog) become associated with the bacterial surface in vitro. By cryo-immunoelectron microscopy, he assessed that urease and HspB are located strictly within cytoplasm in fresh log-phase cultures of H.pylori and become associated with the bacterial surface in late log-phase cultures: these cytoplasmic proteins are released by bacterial autolysis and become absorbed to the surface of intact bacteria, probably due to unique characteristics of the outer membrane. In addition, he confirmed this phenomenon in infected human gastric biopsy specimens using immunoelectron microscopic analysis.
Dr. Dunn concluded that such “altruistic lysis” in which autolysis of a fraction of the bacterial population presumably benefits the remaining viable bacteria, would: 1) allow absorption of the essential enzyme urease (and other cytoplasmic proteins) onto the surface of viable bacteria, thus promoting survival in the harsh gastric milieu; 2) help to explain how the non-invasive bacterium H.pylori can present virulent factors and antigens to the gastric mucosa and immune system; 3) potentially explain the inability of the immune system to eradicate H.pylori due, in part, to “immune overload” resulting from continual release of cytoplasmic proteins; and 4) potentially influence the choice of antigens for oral immunization against H.pylori in humans.
Dr. Fox reported on 13 formally named species of Helocobacter which are proven or suspected pathogens for the gastrointestinal tract. He presented 6 additional Helicobacter species isolated from the stomachs of various mammals, H.mustelae (ferreti; urease + ), H.acinonyx (cheetahs; urease + ), H.felis (cats, dogs, mouse; urease + ), H.heilmanii (human; urease + ), H.rappini (sheep; urease + ) and H.muridarum (mouse; urease + ) that caused various degrees of gastritis in their host. Also Helicobacter species have been isolated from the intestinal tracts of humans, animals, and birds. Two of them, H.muridarum and “H.rappini,” primarily colonize the ilea and ceca of rodents. “H.rappini” has been associated with abortion in sheep and H.bilis have been isolated from diseased livers of inbred and outbred mice. H.hepaticus isolated from A/JCr mice is linked to liver tumors as mentioned before. H.canis and H.pullorum isolated from dogs and chickens, respectively, have also been cultured from feces of diarrheic humans. H.pullorum is associated with hepatitis in chickens. H.fennelliae and H.cinaedi (previously classified as Campylobacters spp.) are two additional Helicobacters primarily associated with lower bowel disease in immunocompromised humans.
Dr. Fox pointed out that further research is required to definitively assess the significance of zoonotic risk, although several of these Helicobacter spp. can infect multiple mammalian hosts, including humans. More evidences should be accumulated to demonstrated that several Helicobacter spp. [play a pathogenic of the disease] and to identify important virulence determinants shared by these organisms and how they influence expression of disease induced by this expanding genera of pathogenic bacteria.
As described, the results of the latest studies in US and Japan on the relationship between H.pylori infection and the genesis of gastric cancer were presented and discussed. The knowledge obtained in this seminar will be extremely useful to clarify the relation between them and take measures for the primary prevention of gastric cancer in human and the results from the studies on Helicobacters other than H.pylori will be useful to clarify the mechanism of H.pylori with malignant diseases of the stomach. However, there remain many points still unidentified and the following problems to be solved were listed in the overall discussion. (1) Regarding the prevention and treatment of H.pylori infection; it is important to reveal the route of infection, particularly, to take measures for completely avoiding the infection through endoscopic instruments. It is to be desired that the effective treatment regimen, mainly using antibiotics, is established, but the development of new drugs without drug resisrance nor side effect is also necessary. The therapeutic vaccine and its clinical application should be further studied. (2) Regarding the relation of H.pylori with gastric cancer; although the results of large-scale intervention study are awaited, the examination of the strain diversity on the Helicobacter’s part, the relationship with genetic factors on the host’s part, and interrelationship with environmental factors, the development of animal model, the investigation on reversibility of atrophy and intestinal metaplasia by H.pylori eradication, the investigation on whether H.pylori eradication can prevent secondary gastric cancer genesis in patients receiving resection of gastric cancer, and the examination on the involvement of H.pylori infection in undifferentiated carcinoma genesis, etc. , are needed. (3) In lymphoproliferative disease, it is necessary to confirm whether MALT lymphoma regressed due to H.pylori eradication is really neoplasm or not. (4) Regarding other Helicobacter species; the development of animal models for elucidating pathogenesis, the future toxonomic refinement and the investigation of zoonotic potential are desired.
At the end of seminar, all the attendants reconfirmed Dr. Sugimura’s expectation that the American and Japanese should mutually exchange the H.pylori strain, DNA probe, serum, reprints etc, and also interchange personnel for further development of the Japanese-American collaborative studies on H.pylori study.
PARTICIPANTS
UNITED STATES
Dr. Jerry M. Rice
Chief, Laboratory of Comparative Carcinogenesis
NCI-Frederick Cancer Research and Development Center
Building 538 Room 205E Frederick, Maryland 21702-1201
Dr. Lucy Anderson
Chief, Perinatal Carcinogenesis Section, Laboratory of Comparative Carcinogenesis
NCI-Frederick Cancer Research and Development Center
Building 538 Room 205E Frederick, Maryland 21702-1201
Dr. Pelayo Correa
Professor, Department of Pathology
Louisiana State University Medical Center
1901 Perdido Street, New Orleans, Louisiana 70112
Dr. James G. Fox
Professor and Director, Division of Comparative Medicine
45-106 Massachusetts of Technology
37 Vassar Street, Cambridge, Massachusetts 02139
Dr. Julie Parsonnet
Assistant Professor of Medicine and Health Research and Policy, Division of Infectious Diseases
Stanford University School of Medicine
HRP Redwood Bldg. T225, Stanford, California 94305-5092
Dr. Steven Czinn
Division of Pediatric Gastroenterology
Rainbow Babies and Children’s Hospital
Cleveland, Ohio 441106
Dr. Martin Blaser
Division of Infectious Disease, Department of Medicine, Vanderbilt Univ. Sch. of Med.
1161 2lst Avenue South, Nashville, Tennessee 37232
Dr. Bruce E. Dunn
Pathology and Laboratory Medicine
Ciement J. Zablocki VA Hospital
5000 West National Avenue, Milwaukee, Wisconsin 53295
JAPAN
Dr. Daizo Saito
Head, Laboratory of Bacteriology and Immunology; Clinical Laboratory Division,
National Cancer Center Hospital, 1-1 Tsukiji 5-chome, Chuo-ku, Tokyo 104
Dr. Toshio Fujioka
Associate Professor, Second Department of Internal Medicine
Oita Medical University, 1-1 Idaigaoka, Hazamachiyou, Oita-gunn, Oita 879-55
Dr. Toshiro Sugiyama
Assistant Professor, Department of Laboratory Diagnosis
Sapporo Medical University School of Medicine
Nishi 16-chome, Minami 1-jo, Chuo-ku, Sapporo 060
Dr. Norio Matsukura
Assistant Professor, First Department of surgery
Nippon Medical School, 1-5 Senndaki, 1-chome, Bunnkyo-ku, Tokyo 113
Dr. Naomi Uemura
Chief, Division of Gastroenterology and Clinical Pathology
Kure Kyosai Hospital, 3-28 2-Chome, Nisi Chuo, Kure-shi, Hirosima 737
Dr. Haruhiko Fukuda
Staff Physician, First Division of Outpatient Clinic
National Cancer Center Hospital, 1-1 Tsukiji 5-chome, Chuo-ku, Tokyo 104
Observers
Dr. Takashi Sugimura
President Emeritus, National Cancer Center
1-1 Tsukiji 5-chome, Chuo-ku Tokyo 104
Dr. Hiroyuki Ono
Chief Resident, Endoscopy Division, National Cancer Center Hospital
1-1 Tsukiji 5-chome, Chuo-ku, Tokyo 104
(2) Seminar on “Stem Cells and Carcinogenesis”
A workshop on “Stem Cells and Carcinogenesis” was held in Oahu, Hawaii, December 1-2 1995, under the auspices of the US-Japan Cooperative Cancer Research Program. The organizers were Dr. Snorri S. Thorgeirsson of the National Cancer Institute, and Dr. Masae Tatematsu, Aichi Cancer Center Research Institute.
There were eight speakers from Japan and seven speakers from the United States. The purpose of the meeting was twofold. First, to discuss and exchange information on the cellular and molecular aspects of several stem cell systems and to probe the possible role that these systems might play in carcinogenesis, and secondly, to explore the possibility of increased research collaboration in this area between scientists in Japan and the United States.
The program ran for two full days. The first session was opened by Dr. Tatematsu who introduced the subject of stem cells by discussing a chimeric mouse model in which he has used staining patterns by a specific antibody to C3H strain specific antigen (CSA) and cell turnover in C3H/HeN BALB/C chimeric mouse tissues. Chimeric epithelial tissues were classified into two categories. In the first type, each unit is composed entirely of either positive or negative cells and displaying a rapid turnover, such as in gastric glands and intestinal crypts. The results suggest that: (1) there is one stem cell in each unit (gland); (2) pluripotent stem cells in units give rise to many types of cells; (3) cell proliferation occurs mainly in transit cells; and (4) stem cells are located at a “focal point.” In the second category, each unit is composed of a mixture of both positive and negative cells and there is a low cell turnover. This category includes the liver lobule, kidney tubules and pancreatic acinus, and the presence of active stem cells is unclear. Thus the staining patterns of such chimeric tissues indicates each cell is derived from its own type, suggesting that the presence and differentiational status of stem cells are different from tissues harboring active stem cells. In the latter type of tissues, stem cells are switched off under normal adult conditions and become reactivated only in the case of extreme tissue damage and possibly also during carcinogenesis.
Dr. Yukihiko Kitamura discussed the gain of function mutation of c-kit as a direct cause of hemopoietic malignancies and the loss of function mutation as an indirect cause of stomach papillomas. The stem cell factor (SCF) is essential for development of mast cells, erythrocytes, and melanocytes in both mice and rats. The receptor of SCF is the c-kit receptor for tyrosine kinase. Recently he found three different gain-of-function mutations of the c-kit in several tumor mast cell lines. Without SCF, the c-kit with the gain-of-function mutations was activated spontaneously. Two of them were the mutation at the juxtamembrane domain and the remaining one was the mutation at the tyrosine kinase domain. The magnitude of the activation was much greater in the tyrosine kinase domain mutation than in the juxtamembrane domain mutations. The introduction of cDNA encoding the tyrosine kinase domain mutation into the interleukin-3-dependent nontumorous mast cell line resulted in autonomous growth both in suspension cultures and in nude athymic mice. The cDNA of the tyrosine kinase domain mutation was also transferred into hematopoietic stem cells using the retroviral vector, and the stem cells were injected into mutant mice with loss-of-function mutations of the c-kit; leukemia of pre-B cell type developed in about half of the recipient mice. In addition, some of the transgenic mice with the tyrosine kinase domain mutation succumbed to pre-B cell type leukemia or lymphoma. Therefore, the gain-of-function mutations of c-kit appeared to be a direct cause of hematopietic malignancies. Fifteen years ago, Dr. Kitamura described the development of forestomach papillomas and chronic antral ulcers in mutant mice with loss-of-function mutations of the c-kit. He also demonstrated the bile reflux from the duodenum to the stomach in the mutant mice, but he could not correlate the bile reflux to the loss-of-function mutations of the c-kit. Since the essential role of the SCF/c-kit system for development of interstitial cells of Cajal was recently reported, he compared the number of interstitial cells of Cajal in the pylorus and the contraction activity of the pylorus between the mutant rats with loss of function mutations of the c-kit and the normal control rats. He found that the number of interstitial cells of Cajal was much greater in the normal control rats than in the mutant rats and that the contraction of the pylorus continued for significantly longer duration in normal control rats than in the mutant rats. The poor contraction of the pylorus appeared to result in the bile reflux from the duodenum to the stomach. Therefore, the loss of function mutations of the c-kit appeared to be an indirect cause of the stomach lesions in the mutant mice with loss of function mutations of the c-kit.
Dr. Tori Miki discussed his work on oncogenes encoding activators of the RHO family of small GTPases. The Rho family proteins, Rho, Rac and Cdc42, are involved in cytoskeletal organization. Using an expression cloning system with the capability of efficient plasmid rescue from mammalian cells, Dr. Miki has isolated two novel oncogenes, ect2 and ost, which contained a structural motif of regulators of the Rho family GTPases. The ect2 plasmid contained a 2.8-kb cDNA insert, the isolated cDNA is shown to have been activated by amino terminal truncation of the normal product. The Ect2 protein exhibits sequence similarity within a central core of 255 amino acids [DH (Dbl homology) domain] with the products of the breakpoint cluster gene, bcr, the yeast cell gene, CDC24, and the dbl oncogene. Each of these genes encodes regulatory molecules or effectors for Rho-like small GTP binding proteins. The baculovirus-expressed Ect2 protein was capable of highly specific binding to Rho and Rac proteins, while the Dbl product showed broader binding specificity to Rho family proteins. Recently, it was found that the regulatory domain of Ect2 is homologous to a yeast cell cycle regulator, Rad4/Cut5. Expression of ect2 mRNA was markedly induced by serum, suggesting a possible function in S phase of the cell cycle. The ost oncogene was also activated by truncation of its N-terminal domain and encodes a predicted protein of 100 kilodaltons containing the DH and PH (pleckstrin homology) domains. Purified Ost protein can catalyze guanine nucleotide exchange on RhoA and Cdc42. Ost did not detectably associate with RhoA or Cdc42, but interacted specifically with the GTP-bound form of Racl. These results implicate Ost as a critical regulatory component which links the signal transduction pathways that flow through Rac1, RhoA and Cdc42. Ost is mainly phosphorylated on serine and localized in the cytoplasm. Among the tissues examined, brain showed the highest expression of Ost especially in neurons and!!
!-tanycytes. Dr. Miki has recently found that Ost can potently activate SAPK/JNK, a novel MAP kinase pathway, through guanine nucleotide exchange on Cdc42.Dr. Jose Russo presented his work on the stem cells in mammary carcinogenesis. The study of the pathogenesis of mammary cancer in both humans and experimental models indicates that the neoplastic process originates in terminal ductal structures, the lobule type 1 (Lob1) in the human, and the terminal end bud (TEB) in the rodent mammary glands. These two types of structures give rise to adenocarcinomas, the most common type of mammary neoplasm. Lob1 and TEBs are considered to be equivalent in their malignant potential. Their high susceptibility to undergo neoplastic transformation is attributed to their high proliferative rate, which is maximal during early adulthood and the postpubertal period in women and rodents, respectively. Cell proliferation decreases progressively with aging in these species. A high rate of cell proliferation, is associated with greater binding of carcinogen to cellular DNA, and lower reparative capacity, both contributing to fixation of transformation, leading to a maximal tumorigenic response upon exposure to genotoxic agent. The binding of given carcinogens, such as 7,12-dimethylbenz(a)anthracene (DMBA). to the mammary epithelium is cell type selective, as well. In the TEB, which is composed of dark, intermediate and myoepithelial cells, intermediate cells are the most abundant, have the highest proliferative rate, and bind more effectively the carcinogen, becoming the source of intraductal proliferations, which progress to carcinoma in situ, and invasive carcinomas. A cell type similar to the TEB’s intermediate cell has been found in the undifferentiated Lob1 of the human breast. These cells also exhibit a high proliferative activity; they bind chemical carcinogens in vitro, expressing phenotypes indicative of neoplastic transformation. However, the “stem cell” of human breast cancer has not been definitively identified as yet. For identifying the breast cancer “stem cell” it is required to categorize in order of importance the role played by specific cell type, rate of cell proliferation, and location within a given compartment in the mammary parenchyma on the susceptibility of the organ to develop neoplasms, keeping in mind that the “stem cell” does not represent a static structure, but rather a dynamic target to given or still unknown carcinogens. The fact that physiological conditions such as pregnancy and aging modify the susceptibility of the breast to develop malignancies support this hypothesis. The higher incidence of ductal carcinomas observed in both nulliparous women and in young virgin rats treated with carcinogens are attributed to the presence of undifferentiated structures in the mammary tissue. The observations that a reduced cancer incidence is exhibited by women with a history of early first full-term pregnancy and by rodents treated with carcinogens after completion of a gestation, and the fact that the reproductive process induces full differentiation of the mammary gland, indicate that “differentiation” might play a leading role in modulating the fate of the “stem cell.” The understanding of the relationship between differentiation and cell proliferation requires an indepth study, since cell proliferation is a normal physiological mechanism of importance for function and tissue repair. The mammary gland epithelium proliferates cyclically in response to hormonal stimuli, and more continuously during pregnancy and lactation. Cells proliferating under these conditions belong to structures already primed by the complete cycle of differentiation induced by the first pregnancy. Dr. Russo hypothesized that these cells might represent a second type of “stem cell,” which upon exposure to any given carcinogen do not respond by expressing an aggressive malignant phenotype, but might develop neoplasms of low malignant potential. The parallels found between human and rodent mammary carcinogenesis validate the use of the rodent model system for testing the roles of differentiation and cell proliferation in the initiation of carcinogenesis and for developing mechanism based strategies for breast cancer prevention.
Dr. Okio Hino discussed the possible role of stem cells in kidney tumors. Cancer is a heritable disorder of somatic cells. Since cancers are derived from single cells, the target must be a cell that continues mitotic activity for a long enough time to sustain two or more events. Two kinds of such “stem” cells are known. The first is the embryonal cell that is critical in the generation of a tissue. The second kind of stem cell is found in renewal tissues, such as the surface epithelia of the digestive tract. Target cell numbers increase rapidly during early development, then vanish with differentiation (postmitotic tissues) or reach a steady state (renewal tissues). These are conditions under which the target cell number evidently changes. For embryonal tumors of children such as the Wilms’ tumor (nephroblastoma), the age-specific peaks in early life, the declines, and falls to such low levels that these tumors are never, or only very rarely, found in adults. The explanation for this phenomenon evidently lies in the observation that these tumors arise from cell types that are normally present in embryonic, fetal, or early postnatal tissue but thereafter characteristically differentiate into postmitotic cells. Thus, the nephroblast is a precursor cell for the nephrons in the kidney. The oncogenic event(s) must occur during the period when target cells are still present and before the postmitotic state is reached. N-Ethyl-N-nitrosourea (ENU)-induced transplacental renal carcinogenesis in the rat results primarily in Wilms’ tumors, apparently because primitive nephroblasts are the preferred target. The question is whether ENU-induced mutations in the fetal kidney would increase the number of adult-type renal cell carcinomas (Rcs) in the Eker rat, which is heterozygous for a mutation that predisposes to renal cell carcinoma. Surprisingly, renal cell tumors but no Wilms’ tumors began to appear from as early as one week after birth. Thus, the inheritance of a renal cell carcinoma mutation determines the specificity of tumor histology even during in utero carcinogenesis, where nephroblasts are the preferred target. The data thus support the hypothesis that the Wilms’ tumor gene may control differentiation of nephroblasts (metanephric stem cells) into renal tubular stem cells, while the Eker gene (RC gene) may control terminal differentiation of these stem cells but permit differentiation of nephroblasts of renal cells.
Dr. Benjamin F. Trump presented his work on stem cell populations in human and animal bronchial epithelium. Injury to and repair of the human bronchial epithelium is of great importance to the pathogenesis and treatment of several major human diseases including asthma, bronchitis, bronchogenic carcinoma, and cystic fibrosis. The bronchial epithelium in the human and experimental animals is pseudostratified, consisting of secretory cells, ciliated cells, and basal cells-all of which rest on the basement membrane. There is no additional morphologic cell type which has been designated as a “reserve” or “stem” cell. Of the normal bronchial cell population, only secretory and basal cells are capable of division; the ciliated cells appear to be terminally differentiated. The number of secretory cells varies with stimulation; following irritation, hyperplasia of secretory cells occurs, and the size of the secretory granules increases (goblet cell hyperplasia). With further injury, squamous or epidermoid metaplasia occurs-the epithelium usually becoming stratified. When human bronchial explants are grafted subcutaneously in nude mice, cells at the edge of the explant migrate outward, line a cavity, and then differentiate into a normal-appearing respiratory epithelium. Similar differentiation occurs when primary cultures of human bronchial epithelium are placed in denuded rat tracheas implanted subcutaneously in nude mice. Wounding studies in hamster tracheas show a similar result, in that the wounded area is repopulated by cells which differentiate into a normal appearing tracheobronchial epithelium. Recent studies employing retrovirus-mediated gene transfer to mark clones have indicated the presence of a cell type which is capable of extensive renewal and pluripotent development. While this and previous studies indicate that such cells are found in the basal and/or secretory cell populations, further characterization studies will be required to determine whether the entire bronchial epithelium can be reconstituted from one pluripotent cell and, if so, whether such a cell is to be found in the basal or secretory compartment.
Dr. Rebecca J. Morris discussed her work on epidermal stem cell as targets for cutaneous carcinogens, Epidermal stem cells long have been considered a target of carcinogenic chemicals, but neither the stem cells nor the target cells have been identified or isolated. Toward this goal, Dr. Morris has examined two-stage carcinogenesis in light of the stem cell model for cellular replacement in the epidermis and consider characteristics that may be useful in identifying and isolating the target cells. First, the similar tumor responses regardless of the interval between initiation and promotion indicates that the target cells normally persist in the epidermis for the life of the animal despite the continual turnover of the epidermis and the hair follicles. Hence, the slowly cycling (label-retaining) keratinocytes from the epidermis and hair follicles are potential targets. Label-retaining keratinocytes (LRC) are identified by light microscopic autoradiography following continuous labeling with [3H]thymidine. LRC are found after intervals of 8 weeks to be centrally located in the epidermal proliferative units, and after 14 months to be located on the ventral surface of the primary follicle. LRC are approximately 1% of the basal cells, are quiescent in vivo, are clonogenic in vitro, retain carcinogen-DNA adducts much longer than most of the basal cells, proliferate in response to treatment with the tumor promoter, 12-O-tetradecanoyl-phorbol-13-acetate (TPA) or to wounding, and are among the smallest and most dense of basal cells. Second, the results of carcinogenesis experiments have also indicated that the target cells necessarily have a high potential for proliferation relative to most of the proliferative population; they can form tumors. The keratinocyte colony forming units (kCFU) from the epidermis of normal and treated adult mice are consequently a quantifiable indicator of proliferative potential and another possible target. The kCFU are approximately 1% of the basal cells, are among the smallest and most dense of basal cells, remain constant in number throughout adult life, remain constant in number following tumor initiation, increase in number in response to tumor promotion, and are the in vivo LRC. Dr. Morris has also been exploring the concept of target cell quiescence in carcinogenesis experiments using topically applied 5-fluorouracil (5FU), an agent known to kill cycling, but not quiescent, cells. The tumor responses are surprisingly similar regardless of 5FU application even though 5FU causes severe damage to the interfollicular epidermis and lower follicles. If 5FU is applied after a short period of promotion, then the tumor responses are approximately one-half of the control values, indicating that the target cells can be killed when activated by TPA. Dr. Morris concluded that identifying and isolating the stem cells and the target cells of chemical carcinogens will lead to a better understanding of how growth and differentiation are regulated in normal and abnormal epidermis, and will lead to an understanding of what the neoplastic lesions are at the molecular level, how they are formed, and what activates their expression during tumor promotion.
Dr. Janardan K. Reddy discussed his work on liver stem cells in rat pancreas. The cellular integrity of various organs and epithelia is generally maintained by stem-cell proliferation. In tissues such as bone marrow, intestine, and skin, where there is rapid turnover of cells, stem cells, which are easily recognized, contribute to cell renewal. In organs such as liver and pancreas, where cell turnover under normal physiological conditions is somewhat imperceptible, the functionally fully differentiated cells retain the capacity to divide; therefore, the existence of stem cells in these organs is a matter of considerable debate. Mammalian pancreas, along with liver, develops from a common endodermal protrusion from the dorsal surface of the embryonic gut. During embryogenesis, especially in the early phase of pancreatic development, so-called protodifferentiated state, primitive duct cells appear first, and some of these are believed to function as transient stem cells from which acinar and islet cells originate. Postnatally, however, cells lining the ducts usually do not exhibit stem-cell properties because both the acinar and islet cells that are capable of DNA synthesis and mitosis participate in the replacement of senescent cells or a certain proportion of cells lost owing to injury Nevertheless, Dr. Reddy has hypothesized that bipotential stem cells are present within the ductal/periductal location in the adult pancreas and that their presence becomes apparent only under certain specific conditions. Partial obstruction of main pancreatic duct results in the proliferation of ductal/ductular cells, leading to the formation of new islets. Interestingly, the duct cells, which serve as progenitor cells for acinar cells during fetal development, appear to lose this capacity to differentiate into acinar cells under this experimental condition. Creation of marked expansion pressure on duct cells by inducing a global acinar cell necrosis results in an interesting phenomenon characterized by marked proliferation of oval cells, which then differentiate into hepatocyte phenotype. The pancreas of male F-344 rats depleted of copper by feeding a copper-deficient diet supplemented with 0.6% triethylenetetramine tetrahydrochloride, a chelating agent, shows acinar cell loss beginning at about 4 weeks and progressing to global loss of acinar cells by 8-9 weeks. The acinar cell loss induced by copper deficiency is through apoptosis, rather than by classical necrosis. Morphological studies by light and electron microscope showed randomly dispersed apoptotic bodies with characteristic nuclear features. Apoptotic index, which is about 0.2% at 4 weeks, increased to a maximum of 9.5% at 6 weeks. Copper deficiency does not appear to cause duct or islet cell injury. Loss of acinar cells is associated with appearance in the acinar cell depleted pancreas of “oval cells,” which display round to ovoid nucleus with scant cytoplasm. Ultrastructurally, oval cells resemble cells in embryonic pancreas present during protodifferentiation phase of development. Oval cells line duct-like structures and are also encountered in the periductal and peri-insular interstitium. The exact origin of these oval cells is not fully established; possible sources include duct cells or that they are derived from periductular stem cells that are dormant in the normal adult pancreas. Current evidence suggests that the (peri)ductular cells are the most likely progenitor cells for these oval cells proliferating in the pancreas of copper-depleted rats, because these are negative for!!
!-glutamyl-transpeptidase, whereas duct cells are strongly positive for this enzyme. Oval cell proliferation in pancreas is especially seen in the copper-deficiency model where there is global acinar cell loss. In situations where pancreaticotoxic agents cause only subtotal acinar cell necrosis, there is no oval cell proliferation. The reason for oval cell proliferation observed to date only in the copper-deficiency model is not entirely clear. This may be because of either expansion pressure imposed on duct/ periductal cells to populate pancreas that is devoid of cell mass and/or the direct effect of copper deficiency-induced molecular changes in these cells leading to oval cell proliferation. It is also not clear as to why these ductal/periductal derived oval cells do not differentiate into acinar cells after global acinar cell loss. Feeding of normal diet after 8-9 weeks of copper depletion results in the emergence of hepatocytes within the pancreas over the next few weeks. These cells are termed pancreatic hepatocytes. They appear either singly or as multicellular foci occupying 60-80% of volume in the pancreas devoid of exocrine acinar cells. Foci of hepatocytes of variable sizes are randomly distributed throughout the pancreas, and some of the clusters are located close to or surrounding the islets of Langerhans. Pancreatic hepatocytes are arranged in sheets without any lobular pattern. No sinusoidal vasculature is present between the hepatocytes at the light and electron microscopic levels. The cytological features of these hepatocytes is identical to the parenchymal cells of the liver. Bile canaliculi are fully developed between adjacent hepatocytes Immunofluorescence studies with domain-specific antibodies clearly demonstrated the presence of apical, lateral, and sinusoidal plasmalemmal domains in pancreatic hepatocytes. Once formed the pancreatic hepatocytes persist throughout the life-span of these rats. Pancreatic hepatocytes are also functionally fully differentiated. They synthesize albumin, respond to xenobiotics like normal liver cells, and contain liver-specific mitochondrial enzyme carbamoylphosphate synthase and uric acid-metabolizing peroxisomal urate oxidase. All pancreatic hepatocytes also contain ammonia-metabolizing enzyme glutamine synthetase. As in the normal liver parenchymal cells, a2u-globulin gene expression in pancreatic hepatocytes is regulated by male sex hormones. The rat pancreatic hepatocyte model that we developed and optimized offers several advantages for the search for liver stem cell over the hepatocyte oval cell system. First, there are no preexisting hepatocytes within the pancreas to confound the observations of lineage transition and differentiation. Second, the copper-deficiency model offers a window of opportunity to identify progenitor (stem) cells prior to the emergence of any mature hepatocytes. The earliest markers of liver specific transcription may be localized in these progenitor cells by in situ hybridization and by other methods. Third, the use of retroviral vector!!
!-galactosidase, as a marker gene for lineage analysis, will allow the fractionation of progenitor cells, and also permit analysis of their fate both in vivo and in vitro under varied conditions of growth factor and extracellular matrix interplay. Dr. Reddy concluded that the ductal system of the pancreas will be an effective system to transfect various genes (growth factors etc.) using adenoviral or retroviral vectors by infusing into the pancreatic duct of copper deficient rats, and study their influence on the lineage and differentiation potential of intraductal and periductal progenitor cells. Dr. Yoichi Konishi discussed his work on stem cells and pancreatic carcinogenesis in hamsters. The pancreas is thought to be a stable organ but it has been reported that hepatocytes appear in regenerating hamster pancreas, suggesting that stem cells may exist in the pancreas. He found that a 100% yield of hepatocytes which were stained positively with PAS for glycogen and with Stein iodine for bile pigments in the pancreas of female hamsters treated with twice repeated sequential administrations of ethionine together with a protein-free diet and then methionine for 10 weeks. Similarly, hepatocytes were also found in 40% hamsters receiving 20 mg/kg body weight of the pancreatic ductal carcinogen, N-nitrosobis(2-oxopropyl)amine (BOP) twice at the peak of pancreatic regeneration stimulated by methionine after ethionine-induced cell damage. However, BOP at doses of 30, 70 and 100 mg/kg administered before the occurrence of pancreatic regeneration resulted in dose-dependen inhibition of the appearance of pancreatic hepatocytes. Further, Makino et al. (Lab. Investig. 62: 552, 1990) studied electron microscopically the origin of pancreatic hepatocytes appearing in regenerating hamster pancreas and found trans-differentiation of ductular cells into hepatocytes which proliferate during pancreatic regeneration. In this model, there was no evidence supporting stem cell differentiation into hepatocytes in pancreas. Together with Dr. Reddy’s presentation, it is evident that hepatocytes appear in the pancreas of rats and hamsters under certain conditions. However, the possibility that the generation of pancreatic hepatocytes occurs via transdifferentiation needs to be considered.
He also presented a rapid production model for pancreatic duct adenocarcinomas in hamsters. In this model, he applied the principle of resistance of carcinogen-initiated cells to the cytotoxic and cytostatic effects of carcinogens during their expansion. The experimental protocol consisted of an initial subcutaneous injection of 70 mg/kg BOP followed 12 days later by 3 cycles of ethionine/methionine rescue-induced pancreatic regeneration while hamsters were maintained on a choline-deficient diet. Each cycle was followed two days later by injection of 20 mg/kg BOP, and the cycles were separated from one another by an interval of 10 days. The pathogenesis of pancreatic duct adenocarcinomas in this model appears to progress from duct epithelial cell hyperplasia and with cellular atypia with or without papillary growth to carcinomas in situ with a cribriform pattern, and finally to invasive carcinomas. It was found that K-ras mutation is an early event and p53 mutation is a late event during pancreatic duct carcinogenesis. No evidence was found suggesting that multipotential stem cells were involved in this experimental system. However, this model is very useful for studies on the mechanisms and modulation of pancreatic carcinogenesis since human pancreatic duct adenocarcinoma is an incurable, deadly disease.
Dr. Takanori Hattori presented studies on kinetics of stem cells and gastric carcinogenesis. By 3H-thymidine autoradiography (ARG), it was shown that cell proliferation for replacement of epithelial cells occurs at the isthmus (neck) region of gastric foveolae. Each foveola contains the proliferative cell zone which has one stem cell. Stem cells are pluripotent, giving rise to not only surface epithelial cells but also glandular cells. Stem cells continue to divide with a generation time of around 24 hrs in rodents, whereas little is known as to the control mechanism of cell proliferation and differentiation. From a cell kinetic point of view, Dr. Hattori showed modes of the development of gastric cancers, firstly about signet ring cell carcinomas and then adenocarcinomas. The materials were human small/minute cancers endoscopically-diagnosed, and those incidentally found in the mucosa of resected stomachs. In order to analyze the very initial stage of signet ring cell carcinoma development, he induced experimentally the signet ring cell carcinomas in dog stomachs by oral administration of N-ethyl-N’-nitro-N-nitrosoguanidine. Cell proliferation was analyzed by 3H-thymidine, ARG and PCNA-staining. Results of his studies showed that almost all gastric cancers arise at the neck region of glands. Signet ring cell carcinomas arise de novo from the stem cell of normal-looking glandular tubules. there is no precursor lesion for the development of this type of cancer. The cancer cells are at first confined to the neck region of glands, and then begin to infiltrate into the lamina propria by dripping from a tubule. In this process, it was suggested that dysfunction of cell adhesion molecules may be related to the dissociation of cancer cells. Next, Dr. Hattori showed the development of adenocarcinomas. It has been suggested that intestinal metaplasia may be related to this type of tumor. He studied the patterns of development of intestinal metaplasia by 3H-thymidine and ARG. It was shown that proliferative cells are confined to the lower part of the intestinalized glandular tubules. This indicates that the localization of proliferative cell zone becomes shifted from the neck to the bottom of glands when gastric/pyloric tubules change to intestinal types. The starting point of intestinal metaplasia is the neck region; the overall spreading of intestinal metaplasia in the gastric mucosa is an outcome of multifocal inductions of intestinal differentiation in the neck stem cell of each glandular tubule. Morphologically, two types of intestinal metaplasia, complete and incomplete, were characterized, and the latter was predominant in stomachs with adenocarcinomas. Histological studies revealed that the initial lesions of adenomas and adenocarcinomas were shown to be distributed around the neck region. It is likely that most gastric neoplasias occur at the neck stem cells, and intestinalized tubules do not necessarily precede the development of adenomas and adenocarcinomas. On the other hand, he also presented recent studies on the phenotypic expression of early gastric cancer cells showing that adenocarcinomas were not always of an intestinal-type from the beginning. About 1/3 of them were gastric-type, not related to intestinal neoplasia, 1/3 were intestinal, and the remaining 1/3 were a mixture of gastric and intestinal type. These findings indicate that more than half of the adenocarcinomas do arise from stem cells of ordinary gastric/pyloric glandular tubules, and only a few adenocarcinomas arise from stem cells of incompletely-intestinalized glandular tubules.
Dr. Eiichi Tahara reviewed the molecular mechanism of stomach carcinogenesis. A number of molecular events are involved in stomach carcinogenesis. Among them, p53 mutation, c-met activation, genetic instability, CD44 intron expression, and telomerase activity are common events of both well differentiated or intestinal type and poorly differentiated or diffuse type of gastric cancers. Gene amplification and overexpression of cyclin E are frequently associated with both types of gastric cancer and correlates well with malignant progression. In addition to these genetic alterations, APC inactivation and K-ras mutation confer the development of the well differentiated gastric cancer. c-erbB-2 gene amplification, LOH at the DCC locus and the bcl-2 gene, and LOH on chromosome 1q facilitate the progression of the well differentiated gastric cancer. Interestingly, more than 30% of gastric intestinal metaplasia and adenomas harbor p53 mutation, CD44 abnormal transcripts and DNA replication error (RER). On the other hand, genetic changes or reduction of E-cadherin and catenins occur at the early stage of the poorly differentiated type. K-sam gene amplification is detected preferentially in the poorly differentiated type but not in the well differentiated type. Scirrhous type gastric cancer displays high frequency of RER. Therefore the particular combination of multiple gene changes seen in gastric cancer differs on the two types of gastric cancer, indicating that different genetic pathways exist for the well differentiated and the poorly differentiated adenocarcinoma of the stomach. In addition to these molecular events, the majority of gastric cancer express telomerase activity which may be responsible for cell immortality. Telomerase activity is detected in germ cells, cancer cells and hematopoietic or epithelial stem cells, but not in normal somatic cells. Of great importance is the observation that 23% of intestinal metaplasia and 50% of gastric adenomas express detectable telomerase activity, suggesting that telomerase activity has implications for early stage of stomach carcinogenesis. However, Dr. Hattori noted that several issues need to clarified: (1) is telomerase activity expressed in stem cells of gastric epithelium; (2) does telomerase activity in precancerous lesions correlate with genetic alterations in oncogenes or tumor suppressor genes; and (3) is telomerase activity the first step of pathogenesis of the gastric cancer?
Dr. Masae Tatematsu presented his work on stem cells and large intestinal tumors. Morphological analysis of isolated postnatal intestinal crypts in rats showed that crypts appear to reproduce themselves by a fission mechanism. The fission process begins in the region of the crypt base in the colonic crypts and proceeds until there are two separate crypts. Occasionally before completion of the separation, a second fission process starts on one or both sides of bifurcating crypts and triple or quadruple crypts are formed. By analysis of isolated aberrant crypts (ACF), the development of ACF was also revealed to be due to a fission mechanism. Initially, an indentation appears at the base of a single gland of an ACF and ACF consisting of 2, 3, or more, crypts then develop. The crypt base is the area where the functional stem cells are found in this system and it is probable that each fission event begins with a localized proliferation of stem cells. He also investigated clonality of colon preneoplastic and neoplastic lesions of C3H/HeN BALB/C chimeric mice treated with 1,2-dimethylhydrazine immunohistochemically using an antibody of CSA. Simple sequence length polymorphism (SSLP) analysis was also performed on DNA samples extracted from histological sections of adenocarcinomas of chimeric mouse. In normal colonic mucosa of the chimeras, each gland was composed entirely of either CSA positive or negative cells and no mixed glands were found, indicating that each gland in the adult mouse is derived from a single progenitor cell. It follows that the columnar cells and goblet cells in individual crypts are descended from multipotential single progenitor cells in each gland. Cells of all focal atypias, adenomas, and adenocarcinomas in chimeric mice were always found to be homogenous for one of the parental types, while comprising both columnar and goblet cell forms. However, among 119 adenomas, 12 contained cells of both parental types arranged in discrete areas, with intermingling limited to the junctions. SSLP analysis demonstrated DNAS extracted from CSA-positive and -negative tumors to exhibit the polymorphic patterns of C3H and BALB/C, respectively, while mixed CSA-positive and -negative tumors showed mixtures of both polymorphic types of DNAs. The results suggest that individual cancers are derived from single cells with multipotential capacities showing monoclonal growth, with apparently polyclonal tumors arising secondarily during progression, due to two or more lesions coalescing.
Dr. Patrice J. Morin discussed the role of APC in colorectal cancers. Most colorectal cancers are believed to arise from a preexisting adenoma, making the colon an attractive system for the study of neoplasia and tumor progression. In the current model, colorectal cancers occur due to a series of genetic changes, each of which leading to a more aggressive phenotype. Mutations can occur in three types of genes: oncogenes, tumor suppressor genes, and mutator genes. Mutations in the tumor suppressor gene adenomatous polyposis coli (APC) are responsible for FAP, a familial syndrome in which the patients develop hundreds of polyps, one or a few of which will progress to carcinomas. Moreover, somatic mutations of APC are an early, if not initiating event, in the development of colorectal cancers in the general population. Somatic APC mutations have been identified in the smallest dysplastic lesion analyzed. The unique role of APC mutations in tumor development in FAP patients and in the general population suggests that the APC gene plays a critical role in the maintenance of normal colonic mucosa. The physiological function of APC is still unknown, but the wild-type protein has been shown to associate with the microtubules and to be in a complex with!!
!- and!!!-catenin, suggesting a role for APC in cell adhesion. To further define the function of APC, Dr. Morin has established a system that enables him to inducibly express wild-type APC in a colon carcinoma cell line containing only mutant copies of the APC gene. Ongoing studies suggest that AFC coordinates the growth of colonic epithelial cells in response to extracellular contacts. Dr. Snorri S. Thorgeirsson presented his work on hepatic stem cells and liver carcinogenesis. Many of the properties of differentiated hepatocytes and of more poorly differentiated oval cells indicate that both types of cells can participate in the renewal of the hepatic parenchyma in rats. Bipotential stem cells participate in the embryogenesis of the liver, but not in the postnatal growth of the liver and not in the maintenance of hepatocyte or biliary epithelial cell populations in adult animals under normal conditions or after most types of induced cell loss. Turnover of hepatocytes and biliary epithelial cells is practically nil in the livers of adult animals; both hepatocyte birth and death occur at very low levels. Reparative renewal of the hepatocytes and biliary epithelial cell populations is effected by the proliferation of the residual differentiated cells of each type, without the need to activate stem cells and establish new lineages of differentiated cells. Certain extreme types of liver injury, however, can lead to the proliferation and accumulation of large populations of poorly differentiated cells, which appear to be produced by the activation of stem cells, and which can differentiate into hepatocytes, biliary epithelial cells, and other types of differentiated cells in favorable experimental circumstances. Additionally, small, poorly differentiated cells that possess some major properties of liver epithelial stem cells can be isolated from livers of rats and established in culture; when transplanted into livers of syngeneic animals, cells of some of these lines differentiate into hepatocytes. Furthermore, the tumors that result when transformed variants of these stem-like cells are transplanted into different sites of host animals express morphological differentiations, among them hepatocytic and biliary epithelial. These observations indicate that, rather than being a static system of cells whose only function is as a metabolic factory for the body, the liver is, in fact, potentially a kinetically dynamic tissue in which parenchymal cells and their precursors have wide and varied, but tightly regulated, options to proliferate and/or to differentiate. Although the mechanisms that regulate the proliferation and differentiation of liver epithelial cells are not clear as yet, the better understanding of the kinetic status and structural organization of cells that make up the liver provide a rational basis for the design and implementation of studies that may eventually lead to this knowledge. Dr. Thorgeirsson concluded that, the liver and its component cells, afford excellent tissue and cellular “reagents” with which to investigate further the controls of cell proliferation and differentiation, and the interaction of these two important cellular processes during normal and pathological replacement of cells.
Dr. Katsuhiro Ogawa discussed phenotypic changes of hepatocytes and bile duct cells during hepatic carcinogenesis and in culture. Hepatocytes and bile duct cells can express each other phenotype during chemically-induced hepatic carcinogenesis and in culture. Transition from bile duct cells to hepatocytes can be typically observed in 3’-Me-DAB-induced hepatic carcinogenesis. These bile duct cells, so-called oval cells, express intermediate morphology and biochemical properties between bile duct cells and hepatocytes. On the other hand, such change is not evident for bile duct cells induced by the Solt-Farber model in which proliferation of neoplastic nodule dominates, indicating that the reaction seems to be related to nature of hepatic damages induced by carcinogens. Normal hepatocytes can express properties of bile duct cells in culture. However, such changes are reversible, because the cultured hepatocytes are readily integrated into the hepatic plates after transplantation within the liver and show the properties of normal hepatocytes. When hepatocytes are cultured within a collagen gel matrix, they form bile duct like tubular structures within a relatively short period. Some soluble factors were demonstrated to be important for such phenomenon, but the morphogenic activity of the factors was not necessarily parallel to their growth stimulating effect. Dr. Ogawa concluded that both hepatocytes and bile duct can behave as stem cells under special conditions, that cells involved in stem cell reaction may exhibit the intermediate properties of bile duct cells and hepatocytes, and that gene expression of the cancers originated from these cells may be different from those originated from pure hepatocytes or pure bile duct cells.
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 complimentary between the U.S. and Japanese research in this area. New contacts and collaborations were developed in an atmosphere of international cooperation.
PARTICIPANTS
UNITED STATES
Dr. Snorri S. Thorgeirsson
Laboratory of Experimental Carcinogenesis
National Cancer Institute-37/3C28
37 Convent Drive MSC4255
Bethesda, Maryland 20892-4255
Dr. Toru Miki
Laboratory of Cellular & Molecular Biology
National Cancer Institute-37/1C05
37 Convent Drive MSC4255
Bethesda, Maryland 20892-4255
Dr. Patrice Morin
Howard Hughes Medical Institute
The Johns Hopkins Oncology Center
424 North Bond Street
Baltimore, Maryland 21030
Dr. Rebecca Morris
Lankenau Medical Research Center
100 Lancaster Avenue, west of City Line
Wynnewood, Pennsylvania 19096
Dr. Janardan K. Reddy
Department of Pathology
Northwestern University Medical School
303 E. Chicago Avenue
Chicago, Illinois 60611-3072
Dr. Jose Russo
Director, Breast Cancer Research Laboratory
Fox Chase Cancer Center
7701 Burholme Avenue
Philadelphia, Pennsylvania 19111
Dr. Benjamin F. Trump
Department of Pathology
University of Maryland School of Mediciine
10 S. Pine Street
Baltimore, Maryland 21201
JAPAN
Dr. Masae Tatematsu
Chief, Laboratory of Pathology
Aichi Cancer Center Research Institute
Kanokoden, Chikusa-ku
Nagoya 644 , Japan
Dr. Takanori Hattori
Professor, The First Department of Pathology
Shiga University of Medical Science
Seta, Tukiwa-cho
Otsu 520-21, Japan
Dr. Okio Hino
Chief, Department of Experimental Pathology
Cancer Institute
Kami-ikebukuro, Toshima-ku Tokyo 170, Japan
Dr. Yukihiko Kitamura
Professor, Department of Pathology
Osaka University Medical School
Yamadaoka, Suita
Osaka 565, Japan
Dr. Yoichi Konishi
Professor, Department of Oncological Pathology
Cancer Center
Nara Medical University
Shijo-cho, Kashihara 634
Nara, Japan
Dr. Katsuhiro Ogawa
Professor, The First Department of pathology
Asahikawa Medical College
Nishikagura, Asahikawa 078, Japan
Dr. Eiich Tahara
Professor, The First Department of Pathology
Hiroshima University School of Medicine
Kasumi, Minami-ku
Hiroshima 734, Japan
(3) Seminar on “The Use of Transgenic and Gene Knockout Animals in Carcinogenesis Studies”
There has been an explosion in the numbers of transgenic and gene knockout animals produced over the past ten years. Many of these are new animal models are of tremendous utility in cancer research. A meeting held at the end of November 1995, focused of the use of transgenic animals in carcinogenesis studies. This meeting was part of a series of US-Japan workshops co-sponsored by the National Cancer Institute of the National Institutes of Health in the US and the National Cancer Center of Japan. From seven to ten scientist from each country attended and presented data from their respective laboratories. Most of the talks covered studies on newly developed transgenic and gene knockout mice. There are three main classes of mice that are of use in carcinogenesis studies. These include 1) mice that are used to study mutagenesis of chemicals in vivo, the lacZ (MutaMouseTM)and the lacI (Big BlueR); 2) transgenic animals that overexpress oncogenes such as the TG.AC mouse line, the Em-pim-1 transgenic mouse, the TGF!!
!/c-myc double transgenic, the ada mouse and many others; 3) knockout mice that lack tumor suppresser genes such as p53, DNA repair genes including O6-methylguanine-DNA methyltransferase, and xenobiotic receptors such as the PPARa and the aryl hydrocarbon receptor that respond to non-genotoxic carcinogens. The transgenic/knockout mouse era
Dr. N. Ito presented a review of recent literature on tumorogenesis using gene knockout mice lacking expression of the apc gene, the homologue of the human apc gene, the p53 hemizygote and homozygote knockout, and transgenic mice expressing different genes including SV40 T-antigen, v-Ha-ras, c-Ha-ras and the hepatitis B vial genome. He noted that the number of papers dealing with transgenic and knockout mice appearing in the literature over the last decade has been increasing almost logrithmically, since 1985 when transgenic models began to appear and 1992 when a number of knockout lines began to be published. He noted, however, that in many papers where the mice present tumors, the histologies are frequently not clearly described. Analysis of tumorigenesis in these mice revealed that most of the tumors found in transgenic mice had similar histopathological features as those in non-transgenics. Tumors in transgenic mice did appear to grow more rapidly and be more invasive. In contrast to the colon tumors found in humans having the mutant apc gene, tumors in apc knockout mice appeared in the small intestine and had unique features as compared to small intestinal tumors in non-transgenic animals.
Dr. J. Ward reviewed the literature of gene knockout mice. Over 400 mice have been produced using this technology and a great majority are lethal, dying at implantation, placentation, and during organogenesis. Defects have been found in lung, kidney, brain, heart, and other organs. He also emphasized the need to carefully monitor the effects of genetic background citing the recent papers on mice lacking epidermal growth factor. In addition to their use in studying gene function and for development of models for gene therapy protocols, knockout and transgenic mice may be extremely valuable in the development of short-term bioassays for carcinogens and for determining the mechanism of toxicity and carcinogenesis by chemicals. Review of the literature indicates that transgenic models in which cancer-associated genes are over or under expressed generally exhibit early hyperplasia and cancer, sometimes showing spontaneous tumors in tissues not normally associated with spontaneous cancer in mice.
The commercialization and distribution of transgenic/knockout mice
Dr. D. Gulezian discussed the process of making transgenic mice widely available to the scientific community through commercialization. From the stand point of carcinogen bioassays, certain mouse lines have proven to yield tumors with a shortened latency and more uniform sensitivity using lower doses of carcinogens. Thus these mice might prove to be of extreme value to the chemical and pharmaceutical industries, and the regulatory agencies. For commercialization of a mouse line, three phases of development are required. Phase I is for genetic design, manipulation and expression characterization, and generally takes one to four years. Phase II is for colleague verification and basic research, usually taking up to five years while phase III is the stage for applied use and establishment of the mouse line as a commercially viable entity. It was also noted that trangenic lines should be backcrossed for 12 generations to establish a pure strain. The reproductive life span of the line should be accurately determined. Other procedures such as microbiological standardization, genetic stabilization and quality control are industry standards taking several years to complete before a mouse is distributed. The royalty hierarchy also must be dealt with for commercialization of a transgenic or knockout mouse. The original Cohen and Boyer recombinant DNA patent, and either awarded or filed patents covering microinjection and homologous recombination must be considered. Patents have been submitted for use of isogenic DNA in making knockouts and there are submitted and awarded patents covering sequences and use of selective genes. These are all relevant to commercialization process. The Oncomouse patent, which has broad claims and covers all oncogenes, was recently awarded. This patent may affect the commercialization of many transgenic and knockout mice used in cancer studies. Soon, a number of transgenic and knockout lines will be widely available for use in carcinogenesis studies and in testing chemical safety in commercial development.
Use of new mouse models for carcinogen bioassays: the p53 hemizygote
Dr. R. Tennant emphasized the large degree of strain and species specificity of responses to carcinogens in typical carcinogen bioassays. However, whole animals are critical for testing the carcinogenicity of chemicals. Transgenic and knockout models can be used to measure tumor induction rapidly and with fewer animals. Among the most widely used of the transgenic mice for carcinogen bioassays are the p53 hemizygous mouse and the TG.AC mouse. The p53 homozygous deficient mice develop lymphomas within two to three months and are therefore not suitable to most carcinogen bioassays. On the other hand, hemizygous p53 mice have proven to be useful. Using two known trans-species carcinogens, p-cresidine and 4-vinyl-1-cyclohexene diepoxide, these animals developed tumors much sooner than in the standard B6C3H F1 mice under the two year bioassay used by the National Toxicology Program. Most interestingly, tumors developed in the p53 hemizygote at the same sites as in the standard mouse strain. The nonmutagenic carcinogens N-methyl-O-acrylamide and reserpine were inactive in these mice as was the non-carcinogenic analog of p-cresidine, p-anisidine. The TG.AC mouse line, which expresses Ha-ras under control of the fetal globin promoter, was also examined. In this model, skin tumors can be visibly observed within 20 weeks of a dosing regimen using both genotoxic and non-genotoxic carcinogens. An ability discriminate between known carcinogens and noncarcinogens has been noted in preliminary experiments. It was therefore proposed that p53 hemizygous mice can be used to first test a suspect carcinogen. If the compound is not active, it can then be examined using the TG.AC mice to determine whether it is a non-genotoxic carcinogen. A matrix database using known carcinogens and non-carcinogens is currently under development at the National Toxicology Program.
Dr. Y. Gondo used two stage gene targeting to produce a mouse that expressed the lacZ gene under control of the natural p53 promoter. X-Gal staining to visualize b-galactocidase activity indicated that the p53 gene is ubiquitously expressed throughout mouse embryogenesis but was more prominently expressed in rapidly dividing or differentiating cells. Homozygous mice carrying the lacZ gene developed spontaneous lymphomas and died between 10 and 30 weeks of age as in the standard homozygous p53 knockout mouse. A transgenic mouse line was produced in which p53 was specifically expressed in the thymus and spleen by using the immunoglobulin heavy chain promoter to drive the p53 cDNA. These mice, designated Eµ-53(p53-/-), survived longer than p53-/- mice and developed very few malignant lymphomas. Instead, they had an increase in sarcomas and other tumor types that developed later than the 10 to 30 week period typical for lymphomas in p53-/- mice. The tumors seen in Em-53(p53-/-) are rarely found in p53-/- mice indicating that the lymphocyte-specific expression of p53 rescued the p53 knockout mice from developing lymphomas. It is believed that in the absence of early death from lymphomas, the p53 deficiency promoted development of tumors in other tissues.
Dr. S. Fukushima evaluated the use of p53 hemizygous male mice in the study of bladder carcinogenesis using N-butyl-N-(4-hydroxybutyl) nitrosamine (BBN) administered in drinking water. The p53 mice were considerably more sensitive, yielding bladder tumors at a dose of BNN that was at least three times the dose required in normal mice. The overall frequencies of bladder carcinomas and the p53 mutations were found in both bladder carcinomas and dysplastic lesions but the frequencies of mutations were not different between the p53 hemizygous and normal mice. In the p53 hemizygous animals, a higher degree of cell proliferation and induction of a hyperplastic response in the bladder epithelium, as determined by scanning electron microscopy, were noted as compared to normal mice. In addition, only the knockout mice had the hemangiomas and hemangiosarcomas. The former tumors are rare in normal mice. Studies using transgenic mice overexpressing the 06-methylguanine-DNA methyltransferase gene (ada mice, see below) revealed no difference from normal mice in BBN-induced bladder cancer.
The Em-pim-1 and human c-Ha-ras oncogene transgenics
Dr. R. Storer described results with the Eµ-pim-1 transgenic mouse which overexpresses the pim-1 oncogene in lymphoid tissue. The pim-1 product exhibits strong cooperativity c-myc and may play a regulatory role in apoptosis. The mice are predisposed to lymphoma development with about 10% of them spontaneously developing tumors in 120 days. They have also been shown to have a markedly decreased latency period for development of induced tumors with rapid appearance of T cell lymphomas after treatment with the direct acting carcinogen N-ethyl-N-nitrosourea (ENU). However, further studies using genotoxic procarcinogens such as 2-acetylaminofluorene, N-nitrosodiethylamine (DEN) and 1,2-dichloroethane, revealed only small but statistically-significant increases in the incidence of malignant lymphomas in short term (38-40 weeks) studies. Benzene, a known lymphomagen in mice, failed to produce any increases in lymphoma incidence in Eµ-pim-1 transgenic mice. A high frequency of hemangiosarcomas were found in DEN-treated mice. Analysis of exons 5-8 of p53 in lymphomas produced in the Eµ-pim-1 mice revealed only a low percentage of tumors with mutations (4%) in contrast with a higher frequency (14%) found in carcinogen-induced hemangiosacomas suggesting that the p53 gene alterations may play a more significant role in carcinogenesis in endothelial cells as opposed to lymphoid cells in this model.
Dr. T. Kamataki reported on studies of the CYP3A7, a cytochrome P450 that is highly expressed in the liver of human fetus. CYP3A7 metabolizes dehydroepiandrostone-sulfate in a pathway that leads to the synthesis of estriol. The enzyme is also capable of metabolizing aflatoxin B1 (AFB) and a number of food-derived carcinogens. Transgenic mice were produced that express CYP3A7; one expresses the enzyme only in the liver while the other produces it kidney and other tissues but not in liver. Higher levels of DNA damage were found in mice expressing the P450 in liver after AFB administration than a strain that does not express hepatic CYP3A7. This result confirms that CYP3A7 can activate AFB to a DNA-binding metabolite in an intact animal model. Other studies were described using p53 knockout mice. Hepatocytes isolated from these mice were found to be immortal in culture as apposed to those from mice expressing p53. Crossing the p53 mice with the CYP3A7 transgenic to produce a p53-/-/CYP3A7 mouse line. Hepatocytes from these mice, are immortal and can potentially be used for toxicology studies.
Dr. K. Imaida presented results on induced carcinogenesis in mice carrying the human c-Ha-ras oncogene encoding p21. Treatment with DEN resulted in forestomach squamous cell carcinomas but these tumors had no detectable ras gene mutations. In non-transgenic mice, the typical target organs for DEN are liver and colon but tumors in these tissues were not detected. Administration of 6-nitrochrysene (6NC) resulted in a high incidence of forestomach tumors including papillary lesions, papillomas, and squamous cell carcinomas. Lung adenocarcimas were also found at a high icidence in the c-Ha-ras animals treated with either 6NC or urethane. In contrast to the forestomach tumors, ras gene mutations were observed in the lung adenocarcinomas and these mutations bore similarities to those found in human lung cancers. Overexpression of the p21 was not detected in any of the induced tumor types and no mutations were found in the endogenous mouse Ha-ras of Ki-ras genes. Mutations in codon 61 were found in squamous cell carcinomas but not in adenocarcinomas in the mouse model. The c-Ha-ras transgenic mouse model may be useful in the study of lung carcinogenesis.
New models for study of hepatocarcinogenesis
Dr. S. Thorgeirsson described studies using transgenic mice to determine the role of c-myc and hepatocyte growth factor (HGF)in hepatocarcinogenesis. c-Myc, which is amplified or up regulated in 90% of liver tumors, is a downstream component of the tyrosine kinase receptor cascade that is stimulated by growth factors. TGF-!!
!is also overexpressed and it probably enhances c-myc carcinogenesis, possibly by interacting with cyclin D1, a target of c-myc that is associated with an upregulation of cell division. The role of HGF is controversial; its expression is increased in hepatocellular carcinomas and in chronic liver disease but it inhibits the growth of liver tumor cells in culture and the growth of preneoplastic foci in vivo but stimulates the grown of normal hepatocytes. Transgenic mice were produced in which c-myc and HGF were placed under control of the serum albumin promoter, a promoter/enhancer that is efficiently expressed in liver. The TGF-!!!gene was placed under control of the metallothionine I promoter. Liver adenomas and carcinomas were produced in mice expressing c-myc. The latency of expression was markedly reduced in a TGF!!!/c-myc double transgenic. In mice expressing the HGF, only adenomas were found in older animals. In a c-myc/HGF double transgenic, a lower cell proliferation and delay in the neaoplastic process was noted as compared to the c-myc single transgenic mice. Tumor promotion by phenobarbital was also inhibited in the double transgenic. These data establish in an intact animal model that HGF acts as a liver tumor suppressor. In vivo gene mutation analyses using MutaMouseTM and Big BlueRTM
Dr. N. Gorelick described transgenic mice that are used to measure gene mutations in vivo. The mice were developed by insertion of tandem copies of the a 1 shuttle vector that carries either lacZ (MutaMouseTM) or the lacI (Big BlueR) gene. These systems are thought to be more relevant than the in vitro systems such as the Ames test and L5178Y mouse lymphoma gene mutation assay because it is known that a number of endogenous factors affect gene mutations in an intact animal. These include cell proliferation, age, tissue and species specificity, genetic susceptibility and metabolism and metabolite transport. However, the question remain as to whether mutations detected in these models are actually reflective of relevant mutations in endogenous genes. For example, the target genes in these new mouse models are not expressed and are highly methylated. In addition, these systems do not detect large DNA deletions. To address this issue, mutations in the lacI gene of BigBlueTM were compared with mutations in the endogenous hypoxanthine phosphoribosyl transferase (hprt) gene. In splenic cells, background mutant were found to be about ten-fold higher in the lacI gene whereas ethylnitrosourea(ENU)-induced mutant frequencies were similar in the two genes. The difference in background mutant frequencies might be due to the high occurrence of CpG methylation-dependent mutations in the lacI gene. Mutation spectra were also analyzed and found to be similar in the two genes in ENU-treated animals DNA damage in indirect gene comparison (i.e., in different animals). Mutation spectra were also consistent with ENU-induced mutations in lacI in Big BlueTM ininduced by 7,12dimethylbenz[a]anthracene are similar to those produced by this carcinogen in the ras oncogene, and ultraviolet light mutation spectra in lacI are similar to those in the p53 gene in human skin cancers thus suggesting that the mutations detected in the lacI mouse model are relevant to carcinogenesis. About 35 chemicals tested to date demonstrate good predictivity in the Big Blue™, but exceptions still exist that must be investigated.
Dr. T. Ushijima described a series of experiments designed to determine the relationship between DNA adducts, mutation frequencies, cell proliferation and carcinogenicity in various tissues of BigBlueR mice treated with 2-amin0-3,4-dimethylimidazo[4,5-f]quinoline (MeIQ), a food-derived carcinogen. Mutation frequencies were highest in colon >bone marrow> liver. The highest cell proliferation was found in the bone morrow and the highest DNA adduct levels were in the liver. MeIQ causes liver and colon tumors in mice but does not cause lymphoid tumors, despite the finding of mutation frequencies in bone marrow that are similar to those in liver. These data indicate that mutation frequencies are not correlated with carcinogenicity in different tissues. Among the reasons for this discrepancy are 1) different numbers of mutations are required for tumorigenesis in different tissues, 2) cell turnover or apoptosis levels determine the frequency of tumorigenesis and differ between tissues, 3) the mutation spectra of MeIQ is appropriate for one tissue but not another, 4) there are only a few target cells in a particular tissue and they are not adequately represented in the total tissue samples harvested to determine DNA adduct levels and cell proliferation. Analysis of mutational spectra of the lacI gene revealed that the bone morrow mutations differed from those found in the susceptible liver and colon tissues. The most common mutations were G to C transversions at positions in which the G was flanked by a 5’ G and a 3’ C. Indeed this is a frequent Ha-ras mutation site at codon 13 found in MelQ-induced tumors of the forestomach and Zymbal gland in rats.
Mouse models for study of DNA repair and carcinogenesis
Dr. T. Ishikawa sought to determine the role of DNA repair in carcinogenesis by developing the ada mouse which expressed the E. coli O6-methlyguanine-DNA methyltransferase (MT) under control of the metallothionein I promoter. The gene was expressed at a high level in liver after administration of zinc. Mice express only one-ten of level of MT than do humans. The ada mice develop significantly less adenomas and hepatocellular carcinomas than do control mice after administration of DEN or N-nitrosodimethylamine (DMN) thus demonstrating the important role of DNA repair in the carcinogenesis process. The role of nucleotide excision repair in chemical carcinogensis was investigated using knockout mice deficient in the Xeroderma pigmentosum A gene product (XPA). These mice demonstrated enhanced skin ulcers followed by papillomas after receiving a single dose of 7,12-dimethylbenz[a]anthracene and tetradecanoylphorbol acetate treatment. A 100% incidence of tumors was found after 16 weeks of treatment of these mice whereas the heterozygous and wild-type mice yield zero or only a few papillomas under similar treatment conditions. These results provide the first direct evidence that DNA repair protects mice from DNA damage and carcinogenesis by chemicals.
Dr. T. Tsuzuki described the characteristics of a mouse lacking the O6-methylguanine-DNA methyltransferase. The DNA repair methyltransferase transfers the methyl group from O6-methylguanine and O4-methylthymine in DNA to a cysteine residue of its own active site. The gene was cloned, found to contain five exons spanning about 150 kbp, and used to produce a mouse lacking the methyltransferase. The mice are viable but have lower body weight and show growth retardation. Administration of the alkylating agent methylnitrosourea (MNU) resulted in early death of the knockout mice. These dose of MNU did not lead to death in normal mice. Marked decreases in the size of spleen and thymus, low levels of bone marrow stem cells and bacterial colonies in the liver were noted upon necropsy of MNU-treated knockout mice. These data establish that O6-methylguanine and O4-methylthymine are lethal lesions in DNA and that the methyltransferase has an essential role in protection of tissues from alkylating agents.
Xenobiotic receptor knockout mice
Dr. F. Gonzalez described the phenotype of mice that lack expression of the peroxisome proliferator-activated receptor!!
!(PPAR!!!). PPAR!!!is thought to mediate the biological effects of a class of chemicals called peroxisome proliferators that include hyperlipidemic drugs, plasticizers, herbicides, pesticides, and the endogenous steroid dehydroepiandrosterone-sulfate. These chemicals cause an increased number of peroxisomes in the liver and to a lesser extent in kidney. Chronic administration to rodents results in hepatocellular carcinomas. Peroxisome proliferators have not been found to induce peroxisomes in liver of humans and primates. They have also not been shown to be associated with increased cancers in humans. The mechanism of carcinogenesis of these non-genotoxic chemicals is not understood. The PPAR!!!deficient mice to not exhibit peroxisome proliferation or target gene inductions when administered the most potent proliferator WY-14,643. Efforts are underway to determine whether these mice are susceptible to the hepatocarcinogenic effects of peroxisome proliferators. These studies should yield insights in to the mechanism of carcinogenicity of non-genotoxic carcinogens. The phenotype of a mouse lacking the aryl hydrocarbon receptor (AHR) was described by J. Ward. This receptor is responsible for the toxic and carcinogenic effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) and polycyclic aromatic hydrocarbons. The animals lacking expression of AHR had a markedly deficient immune system and liver abnormalities and were resistant to the acute toxicity of TCDD. No spontaneous tumors have been found in these animals but a number of interesting lesions never before seen in mice were found in various epithelial tissues. The mice tend to die before the age of one year. These studies indicate that the AHR is required for health and longevity and most of the toxic effects of dioxins. The mechanisms by which the many abnormal phenotypes are generated are under investigation.
Transgenic mice to study the mechanism of gene regulation by the tumor promoter phenobarbital
Dr. C. Omiecinski presented a series of experiments designed to determine the mechanism of induction of cytochrome P450 genes by phenobarbital. The mechanism of this induction response is not well understood. Similar to the peroxisome proliferators, a large number of structurally diverse chemicals induce P450s and other genes and stimulate proliferation of hepatic endoplasmic reticulum. No receptor for these compounds has been identified. CYP2B1 and CYP2B2 gene expression can be stimulated in primary hepatocyte cultures, but trans-activation assays using reporter genes have not been successful in identifying cis-acting DNA elements. Transgenic mice were produced using rat CYP2B2 genes containing various lengths of upstream and downstream DNA. Mice containing 20 kbp of upstream DNA were responsive to phenobarbital administration similar to the endogenous rat gene. To further delineate the DNA important for induction, chloramphenical acetyltransferase (CAT gene) constructs were prepared having various lengths of upstream DNA. Transgenic mice were prepared using these constructs and a phenobarbital response was found with a - 2.3 kbp construct and not with a - 1.7 kbp construct. These data suggest the existence of a cis-acting element lying between - 2.3 and - 1.7 kbp upstream of the transcription start site of CYP2B2 that is required for phenobarbital induction. Data were also presented showing that cAMP is an important negative regulator of induction in primary rat liver hepatocytes.
The above brief summaries are a snapshot of the many areas of investigation into transgenic and knockout mice. The number of new and interesting mice will no doubt continue to expand. These animal models will be of use to industrial and regulatory applications for chemical testing for carcinogen prediction and to basic research into the mechanisms of carcinogenesis.
PARTICIPANTS
UNITED STATES
Dr. Frank J. Gonzalez
Laboratory of Molecular Carcinogenesis
NCI
Bldg. 37, Room 3E24
Bethesda, MD 20892-4255
Dr. Jerrold M. Ward
Office of Laboratory Animal Science
NCI, FCRDC, FVC 201
Frederick, MD 21702
Dr. Snorri S. Thorgeirsson
Laboratory of Experimental Carcinogenesis
NCI
Bldg. 37, Room 3C28
Bethesda, MD 20892-4255
Dr. Nancy J. Gorelick
The Proctor & Gamble Company
Miami Valley Laboratory
Cincinnati, OH 45253-8707
Dr. Curtis J. Omiecinski
University of Washington
Dept. of Envionmental Health
4225 Rossevelt Way, NE #100
Seattle, WA 98105-6099
Dr. Donna Gulezian
Taconic Transgenics Models & Services
Madison, CT 06433
Dr. Raymond W. Tennant
Laboratory of Enviornmental Carcinogenesis & Mutagenesis
National Institute of Enviornmental Health Sciences
Research Triangle Park, NC 27709
Dr. Richard Storer
Dept. of Safety Assessment
Merck Research Laboratories
WP45-311
West Point, PA 19486
JAPAN
Dr. Takatoshi Ishikawa
Dept. of Pathology
Faculty of Medicine
The University of Tokyo
7-3-1 Hongo, Bunkyo-ku
Tokyo 113, Japan
Dr. Toshikazu Ushijima
National Cancer Center Research Institute
5-1-1 Tsukiji, Chuo-ku
Tokyo 104, Japan
Dr. Tetsuya Kamataki
Hokkaido University, Sapporo
Division of Drug Metabolism
Dr. Nobuyuki Ito
Nagoya City University
l Kawasumi
Mizuho-cho, Mizuho-ku
Nagoya 467, Japan
Dr. Shoji Fukushima
Dept. of Pathology
Osaka City University Medical School
1-4-54 Asahi-machi
Abeno-ku
Osaka 545, Japan
Dr. Katsumi Imaida
Dept. of Experimental Pathology
Nagoya City University
1 Kawasumi
Mizuho-cho, Mizuho-ku
Nagaoya 467, Japan
Dr. Yoichi Gondo
Medical Institute of Bioregulation
Kyushu University
Maidashi, Higashi-ku
Fukuoka 612
Japan
(4) Seminar on “Molecular Targets and their Function for Virally-Induced Proliferation of Cells”
ORGANIZERS:
The organizers of this seminar were
Dr. Peter Howley, Harvard Medical School, Boston, Massachusetts
Dr. Mitsuaki Yoshida, University of Tokyo, Tokyo, Japan.
The seminar took place in
Maui Prince Hotel, Makena Resort, Kihei, Hawaii
January 24-26, 1996.
There were a total of seven participants from Japan and eight participants from the United States. The purpose of the meeting was to discuss and exchange information on the field of molecular targets and their function for the viral inducement of cell proliferation. The opening remarks by Drs. Peter Howley and Mitsuaki Yoshida indicated that this was the fourth in a series of US-Japan seminars which have focused on viral regulatory proteins. The study of viral cellular regulatory proteins has provided important insights into the deregulation of cell cycle control observed in cancer.
The four-session program had four speakers per session. Session One, “Human Retroviruses,” was moderated by Drs. Bernard Roizman and Takeharu Nishimoto. The session began with Dr. Mitsuaki Yoshida who discussed a human retrovirus, HTLV-1, which causes adult T-cell leukemia (ATL) and is also associated with tropical spastic paraparesis (HAM/TSP). The virus has two regulatory genes, one codes for the Tax protein and activates the viral and cellular gene expression enhancing the transcription. Furthermore, Tax can immortalize normal human T cells, transform rodent fibroblasts and induce solid tumors in its transgenic mice, therefore has been thought to play critical roles in the viral replication and in development of ATL. As direct targets of Tax, cellular transcription factors are first identified. These are CREB or CREM protein that binds to the 21-bp enhance of HTLV-1 LTR, NF-!!
!B family proteins including p50, p65, p52 and c-Rel that bind to NF-!!!B site of IL-2Ra and serum response factor (SRF) that binds to SRE of the c-fos proto-oncogene. These indirect bindings of Tax protein to the specific enhancer DNA sequences are a general mechanism of the trans-activation of specific transcription. In transactivation of NF-!!!B, Tax uses another strategy to transactivate the transcription through NF-!!!B binding site. This mechanism includes the binding of Tax to I!!!B, inhibitors of NF-!!!B proteins, including I!!!B!!!,!!!and!!!. Tax binding I!!!B!!!induces destabilization of complex of NF-!!!B/I!!!B!!!consequently translocating the released NF-!!!B into nucleus. Such effect of Tax on IKBa protein was demonstrated in all T cell lines infected with HTLV-1, and also in primary leukemic cells from ATL patients. Interestingly, all five cases so far tested showed the same results, that is increased expression of I!!!B mRNA resulted in by destabilization of NF-!!!B/I!!!Ba complex. In addition to the transcriptional activation, Dr. Yoshida found p16INK4A, a cell cycle inhibitor, as another direct target of Tax binding. It is known that p16INK4A inhibits Cdk4 kinase which phosphorylates Rb protein and induces cell cycle arrest at G1 phase. Tax binding to p16INK4A suppressed the inhibitory activity of p16INK4A and resulted in an activation of Cdk4. Tax also counteracted with p16INK4A which arrests cells at G1 and induced efficient cell cycle promotion. Therefore, it implies that Tax can induce abnormal cell growth directly affecting a cell cycle regulator. Dr. Yoshida emphasized that the retrovirus HTLV-1 uses the same signaling Rb pathways as DNA tumor viruses and its significance was also discussed. Dr. John Brady described his studies with the Tax1 protein of the human T-cell lymphotropic virus type-I which plays a key role in viral replication, transformation, and gene regulation, His laboratory has cloned and characterized Tax1 binding protein, TRX, by screening a T-cell!!
!gt11 library using bacterially expressed, purified Tax1 protein. TRX anti-peptide antibodies recognized a 28 kD protein in western blot analysis of MT4, C81, Jurkat, CEM and human PBL extracts. TRX protein expression was not detected in non-lymphocyte cell lines. TRX mRNA is ubiquitously expressed in lymphocyte and non-lymphocyte cells, suggesting that TRX protein expression is regulated at a post-transcriptional level. An amino acid search of GenBank reveals TRX to be a novel protein, with approximately 30% sequence homology to the conserved cyclin box region of the human cyclins. Cell cycle studies revealed that TRX forms a complex with cyclin B and its catalytic partner, cdc2. To further study this interaction, Jurkat cells were blocked at the G2/M border using nocodazole, a mitotic spindle formation inhibitor. Extracts were prepared from cells withdrawn at specific time points following the removal of nocodazole. Coimmunoprecipitations were performed using these extracts and anti-cdc2 or anti-cyclin B antibodies, followed by western blot analysis using anti-TRX sera. In these assays, the association of TRX with the mitotic cdc2/cyclin B complex was shown to be cell cycle specific. Interestingly, it was observed that the level of the TRX protein does not fluctuate during the cell cycle. Experiments are in progress to determine the role of TRX in the regulation and/or localization of the cdc2/cyclin B complex of lymphocytes. The Tax1 protein of HTLV-I activates transcription from the viral LTR as well as several cellular growth regulatory genes and cytokines. Tax1 has been shown to activate transcription through association with various enhancer-binding proteins (CREB/ATF, NF-!!!B, SRF). It has not been clearly established whether Tax1 also interacts directly with basal transcription factors. Based on several independent assays, Dr. Brady reports a physical and functional interaction between Tax1 and the 35 kD (!!!) subunit of the basal transcription factor TFIIA. First, HeLa whole-cell extracts were depleted of TFIIA when passed over a Tax1 affinity column. Second, in vitro-translated TFIIA interacted specifically with a GST-Tax1 fusion protein. The interaction of Tax1 and TFIIA was significantly greater than that observed with Tax1 and TBP, ATF1 or ATF2. Third, the laboratory observed in vivo the Tax1-TFIIA protein association using the yeast two-hybrid system and a coimmunoprecipitation assay from Tax1-expressing cells. In the two-hybrid system, a Tax1 mutant deleted of the carboxy terminal 78 amino acids (amino acids 276-353), a region containing a intrinsic activation domain, failed to interact with TFIIA. Finally, cotransfection with both the p55 (!!) and p12 (!!!) subunits of TFIIA increased Tax1 transactivation in a human T-cell line. In addition, Tax1 transactivation was not observed in TFIIA-depleted extracts, suggesting that TFIIA is directly required for Tax1 transactivation. Analysis of the interaction of Tax1 with TFIIA will provide important insights into the pleiotropic mechanisms by which Tax1 regulates transcription of viral and cellular promoters.Dr. Kazuo Sugamura also talked about the HTLV-I Tax protein. HTLV-I-Tax induces expression of various cellular genes as well as its own viral gene. He presented two target gene products of HTLV-Tax transactivator, OX40 and the IL-2 receptor g chain. OX40, originally identified as an activated T cell marker, belongs to the TNF receptor family. Its ligand was recently discovered to be an identical molecule to gp34 which Dr. Sugamura had molecularly cloned as one of the target gene products of Tax. Although OX40 has been shown to have growth signal transducing ability, Dr. Sugamura demonstrated that not only OX40 but also OX40 ligand, gp34, has ability to transduce intracellular growth signals in normal antigen-stimulated T cells. Furthermore, he found that not only gp34 but also OX40 can be transcriptionally activated by Tax, suggesting that the OX40/OX40 ligand system may be involved in autonomous growth stimulation of HTLV-I-infected T cells. In fact, IL-2-dependent growth of HTLV-I-infected ILT-Myj cells was partially inhibited by addition of anti-gp34 monoclonal antibody, indicating that the OX40/OX40 ligand contributes to growth promotion of ILT-Myj cells. The IL-2 receptor g chain is indispensable for formation of the functional receptor complexes for not only IL-2 but also IL-4, IL-7, Il-9 and IL15. Consequently, the IL-2 receptor g chain is now called the common g (gc) chain. Although expression of the IL-2 receptor g chain has been considered to be constitutive in various hematopoietic cell populations including T cells, Dr. Sugamura demonstrated that expression of the IL-2 receptor g chain is downregulated by IL-2 stimulation in IL-2-responsive T cell lines. However, IL-2 responsive T cells lines expressing Tax showed no such downregulation, suggesting that the IL-2 receptor g chain gene is a target of Tax transactivator. In fact, the Sagamura laboratory demonstrated that the promoter region of IL-2 receptor g chain gene is responsive to Tax for its transcriptional transactivation. These observations may explain the success of establishing IL-2-dependently from normal human T cell lines when they are infected with HLTV-1.
Dr. Warner Greene ended the first session with a discussion of the type I human T cell Leukemia virus (HTLV-I). HTLV-I has been etiologically associated with the development of the adult T cell leukemia (ATL) as well as degenerative neurological syndrome termed tropical spastic paraparesis (TSP) or HTLV-I associated myelopathy (HAM). HTLV-I encodes a potent transactivator protein termed Tax that appears to play an important role in the process of T cell immortalization. HTLV-I Tax has been shown to transform cells in vitro and to induce non-lymphoid malignancies when introduced as a transgene into mice. While altering the expression of many cellular genes including interleukin-2 and the a subunit of the interleukin-2 receptor, Tax does not directly bind to DNA. Rather, this viral protein appears to act by altering the activity of various host transcription factors leading to abnormal cellular gene expression and amplified activity of the HTLV-I LTR. One family of host transcription factors whose expression is deregulated by HTLV-I Tax includes NF-!!
!B/Rel. These transcription factors are post-translationally regulated by their assembly with a second family of inhibitory proteins termed IkB which serve to sequester the NF-!!!B/Rel complexes in the cytoplasm. Upon cellular activation, IkBa is phosphorylated by one or more unidentified kinases on two N-terminal serine residues. This modification in turn triggers ubiquitination of the inhibitor which targets the protein for degradation in the proteasome. This proteolytic event liberates the NF-!!!B heterodimer, permitting its rapid translocation into the nucleus where it binds to its cognate enhancer elements Of note the p105 precursor of the NF-!!!B p50 subunit is also post-translationally processed in the proteasome. Indeed, this is the first example of a targeted incomplete cleavage mediated through the proteasome. In order to further explore the mechanism by which Tax activates nuclear expression of NF-B, the yeast-two hybrid assay was employed to identify cellular gene products capable of interacting with Tax. Screening of a cDNA library prepared from activated peripheral blood mononuclear cell mRNA identified one Tax interacting cellular gene product as the HsN3 subunit of the 26S proteasome. This physical interaction was confirmed in mammalian COS cells by cotransfection and coimmunoprecipitaton of radiolabeled Tax and HsN3 proteins utilizing antibodies specific for either protein component. To explore the function consequences of this interaction, a cDNA expression vector encoding HsN3 was cotransfected with Tax and a kB luciferase reporter plasmid into COS cells. HsN3 produced a dose related inhibition of !!!B specific gene expression. In contrast, cotransfection of HsN3 with the glucocorticord receptor did not inhibit dexamethasone directed activation of a glucocorticord sensitive gene. Together these findings raise the intriguing possibility that physical association of the HsN3 proteasome subunit with HTLV-I Tax coupled with the independent interaction of Tax with either p100 or p65 I!!!B!!!targets these cytoplasmic NF-!!!B/Rel complexes to the proteasome for pathologic processing. It is also possible that Tax interaction with the HsN3 subunit may alter the composition of the two seven membered rings of the proteasome in a manner that facilitates I!!!B!!!degradation or p100/-105 processing. Session Two, “Human Herpesviruses,” was moderated by Drs. Masanori Hatakeyama and Joseph Nevins. The Session began with Dr. Takeharu Nishimoto. A thermosensitive (ts) mutation is a main avenue to clarify the cell cycle regulation. Dr. Nishimoto has isolated a series of ts cell-growth mutants from the BHK21 cell line derived from golden hamsters which can normally grow at 33.5°C, but not at 39.5°C (Nishimoto and Basilico, 1978; Nishimoto et al., 1982). Since then, Dr. Nishimoto has been cloning human genes complementing these hamster ts mutants by DNA mediated gene transfer (DMGT)-method. So far Dr. Nishimoto has cloned RCC1 (Ohtsubo et al., 1987), CCG1/TAF250 (Hayashida et al., 1994), RPS4X (Watanabe et al., 1993), and DAD1 (Nakashima et al., 1993). Dr. Nishimoto has cloned a human gene which complements a G1 ts mutant, tSBN67 cell line that was arrested in G0/G1 at 33.5°C, nonpermissive temperature. Under such conditions, early G1 genes; Fra-I and c-myc was not expressed, while c-jun was normally expressed, similar to serum arrested tsBN67 cells. The tsBN67 cell line was proved to have a single base change in HCF gene. Due to the mutation, tsBN67-HCF could not make a stable complex with VP16 and so, the VP16 mediated gene expression was abolished in tsBN67 cells. HCF is required for the expression of immediate early genes in order to start virus proliferation. In this context, it is interesting to note that HCF is required for the progression from G0 to G1 transition in order to initiate a new cell cycle, even in host cells as well as virus proliferation. Although one can not exclude the possibility that HCF by itself has some function required for the cell cycle, it is reasonable to assume that HCF functions as a co-factor for Oct-1, a ubiquitously expressed protein. In this context, cells might have an unknown factor corresponding to VP16, which Dr. Nishimoto temporarily designated as cellular VP16 (c-VP16). It is still obscure how cells enter, or exit the G0 phase. The phenotype of tSBN67 cells suggests that HCF functions as a fine regulator for the G0/G1 transition, while the Nishimoto laboratory could not exclude the possibility that it is specific for the allele of tsBN67 mutation. Regarding this assumption, the identification of c-VP16 is particularly interesting, since it may clarify the molecular mechanism of the G0/G1 transition.
Dr. Bernard Roizman described the HSV-1!!
!134.5 gene, which plays a key role in pathogenesis in that its product enables the virus to replicate in the central nervous system of mice. In human neuroblastoma cells infected with!!!134.5 minus (!!!134.5-) viruses, a premature shutoff of protein synthesis induced by cellular factors precludes viral replication. The cellular response is triggered by events occurring after the onset of viral DNA synthesis. The carboxyl terminus of!!!134.5 is the only domain necessary to block the cell response. This domain is homologous to the corresponding domain of GADD34 proteins (murine, hamster, or human) induced on differentiation, growth arrest (e.g. caused by serum deprivation), or during repair of damaged DNA. In his presentation, Dr. Roizman reported the following: (1) The murine GADD34 carboxyl terminus can substitute for that of !!!134.5 protein. The functional equivalence of these homologous domains suggests that GADD34 and !!!134.5 carboxyl termini may interact with identical proteins during HSV infection and during growth arrest caused by differentiation or other causes and that HSV “borrowed” this domain during its evolution. (2) The cessation of protein synthesis in !!!134.5-infected cells correlates with complete phosphorylation of the a subunit of the eIF-2 translation initiation factor. In addition, the PKR kinase becomes phosphorylated and a novel M190,000 phosphorylated protein coprecipitates with the PKR kinase from cell lysates of !!!134.5-infected cells. Although PKRT is activated in cells infected with wild type virus or with viruses carrying deletions in the 5’ terminus of the !!!134.5 gene, the a subunit is not extensively phosphorylated and the M190,000 phosphoprotein is not coprecipitated with PKR kinase, !!!134.5 protein and its homologs may act to preclude shutoff of protein synthesis, and, ultimately, premature cell death. Dr. Kenzo Takada discussed the Epstein-Barr virus (EBV), which is detected in several human cancers, such as Burkitts’s lymphoma (BL) , nasopharyngeal carcinoma (NPC), gastric cancer, and peripheral T-cell lymphoma. However, the role of EBV in the development of these cancers is still controversial. EBV transforms primary lymphocytes into blasts that can proliferate indefinitely. Such EBV-transformed cells maintain the entire viral genome in a plasmid form and express 12 EBV genes encompassing two small non-polyadenylated RNA known as EBER1 and EBER2, six nuclear antigens (EBNA1, EBNA2, EBNA3A, EBNA3B, EBNA3C, and EBNA-LP), three membrane proteins (LMP1, LMP2A, and LMP2B), and the transcripts for the BamHI A region, termed BARF0. In contrast to the EBV expression in transformed lymphocytes, in Burkitt’s lymphoma and gastric cancer EBNAI is expressed, but other EBNAs and LMPs are not expressed, and the expression of LMP1 is limited to about half of cases. It is known that EBNAs other than EBNA1 become targets of cytotoxic lymphocytes (CTL). Therefore,!!
!134.5is important for tumor cells not to express these antigens in order to survive under CTL surveillance. However, the restricted EBV expression in tumor cells raises the question about the pathogenic role of EBV in EBV-associated malignancies, because the EBNA and LMP1 are known to play particularly important roles in the initiation and maintenance of transformation. The possibility has not been excluded that EBV is the passenger virus, or a factor which plays roles on the development of tumor but is not required on its maintenance. It has been difficult to examine this, since we not have an in vitro system of producing tumor cell-type EBV infection. The Akata cell line is derived from an EBV-positive BL from a Japanese patient, and has a unique property to retain tumor cell type EBV expression after long-term cultivation in vitro. Initially, Akata cells were 100% positive for EBNA by the immunofiuorescence assay. However, after serial passage for about 2 years, the Takada laboratory noted that EBV DNA is lost from part of cells. Isolation of EBV-positive and -negative Akata cell clones with the identical origin made it possible to examine the effects of EBV in B cells. The results indicate that malignant phenotypes of BL, such as growth in low serum, anchorage-independent growth in soft agar and tumorigenicity in nude mice, are dependent on the presence of EBV genomes. These observations may be extended to the pathogenic role of EBV in other EBV-associated tumors, such as NPC, gastric cancer, and peripheral T-cell tumors, and underline the oncogenic function of EBV in human cancer. When EBV was re-infected to EBV-negative Akata cells, cells acquired malignant phenotypes. On the other hand, when the EBNAI gene was transfected to EBV-negative Akata cells, cells did not become malignant. Therefore, EBNA1 is not responsible for the maintenance of malignant phenotypes. BARF0 or unidentified proteins may be responsible for it. Dr. Thomas Shenk concluded the session by describing human cytomegalovirus-infected fibroblasts. These fibroblasts are resistant to the induction of apoptosis by superinfection with a mutant adenovirus lacking the E1B 19 kDa gene that normally causes an E1A protein-mediated apoptotic response. Two cytomegalovirus gene products were identified that block apoptosis. The IE1 and IE2 proteins each inhibit the induction of apoptosis by tumor necrosis factor a or by the E1B 19 kDa-deficient adenovirus but not by irradiation with ultraviolet light. The IE1 and IE2 proteins share a common amino-terminal sequence but deletion analysis indicates that this domain is not responsible for their abilities to antagonize cell death. Rather, the anti-apoptosis function resides primarily to the carboxy-terminal side of the shared domain in both proteins. Dr. Shenk suspects that the two proteins function differently to block apoptosis, and their mechanisms of action are currently under investigation. The cytomegalovirus proteins also cooperate with the adenovirus E1A proteins to transform baby rat kidney cells, presumably by blocking cell death. The ability to block apoptosis and to cooperate in the immortalization of primary cells could be key elements of the mechanisms by which the virus replicates and persists, influencing the course of HCMV pathogenesis in its infected host.
The next day, Session Three covered “Adenoviruses and Cell Cycle Control,” and was moderated by Drs. Thomas Shenk and Yoshiaki Ito. Dr. Joseph Nevins began with a discussion about how since the initial studies of retinoblastoma tumor suppressor gene function, most attention has focused on the role of this protein in controlling cell growth during the G1 to S phase transition. Considerable insight into the action of Rb has been gained from the study of the DNA tumor virus oncoproteins that disrupt the function of Rb. Such studies have led to the identification of the cellular transcription factor E2F as a target for the action of Rb as a growth suppressor. Indeed, various studies now suggest that the E2F transcription factor is an integral part of the growth regulatory network which controls the progression of cells from G0 and early G1 into the S phase of the cell cycle and when perturbed, can lead to the development of cancer. The ability of Rb to function as a growth suppressor and to regulate the activity of E2F is controlled by phosphorylation, likely mediated by the G1 cyclin-dependent kinases. Numerous experiments have defined a critical role for the G1 cyclins and associated kinases in allowing a normal progression of cells from a quiescent state, through G1, and into S phase. The Nevins group now finds that G1 cyclin-dependent kinase activity is critical for the accumulation of E2F activity late in G1. Moreover, E2F1 over-expression can bypass a G1 arrest caused by the inhibition of G1 cyclin-dependent kinase activity, consistent with E2F activation being an important consequence of the action of these kinases. E2F1 also overcomes a G1 block due to!!
!134.5irradiation and leads to an apparent complete replication of the cellular genome and entry into mitosis. This E2F1-mediated induction of S phase and mitosis is not accompanied by the rise in either cyclin D-associated kinase activity or cdk2 kinase activity that normally occurs during the G1 phase of the cell cycle. From these results, Dr. Nevins concludes that one key function for G1 cyclin-dependent kinase activity is the activation of E2F1, that the accumulation of E2F activity may be sufficient to allow initiation and completion of S phase, but that additional events, including G1 cyclin kinase activity, are likely necessary for a normal proliferative event. In addition to the approaches afforded by mammalian cell culture experiments, the Nevins laboratory has utilized a Drosophilia homologue of the E2F1 gene to further explore the role of E2F activity in promoting normal cell cycle progression, and the relationship of these events to cell differentiation and cell survival. The laboratory finds that ectopic expression of E2F1 is sufficient to stimulate many cells in the larval imaginal discs to enter S phase, cells that would not normally enter S phase. Coincident with the S phase induction, large numbers of cells undergo cell death with characteristics of apoptosis. At least two exceptions to this pattern can be found. Cells within the morphogenetic furrow of the eye disc do not enter a premature S phase following E2F1 induction and do not undergo cell death. Rather, these cells continue their synchronous progression towards a normal S phase. In addition, cells that have exited the cell cycle and acquired a terminally differentiated accumulation of E2F activity is sufficient to initiate S phase, that certain cells are resistant to the deleterious effects of E2F1 mis-expression, and that successful completion of the cell cycle and cell survival requires activities in addition to E2F that likely constitute the various pathways other than the replication of the cellular genome that are essential for a productive proliferative event. Dr. Manasori Hatakeyama discussed the retinoblastoma gene product (pRB), which constrains cell growth by preventing cell cycle progression from G1 to S phases. The pRB function appears to be negatively controlled by multiple serine/threonine phosphorylations provoked in late G1 phase of the cell cycle. The cell-cycle dependent phosphorylation of pRB can be faithfully reproduced by expressing human pRB in the yeast Saccharomyces cerevisiae. As is the case with mammalian cells, the phosphorylation requires an intact pRB pocket domain which is involved in pRB-target molecule interaction. Phosphopeptide mapping analysis of pRB expressed in yeast showed that biochemical details of the hyperphosphorylation is basically indistinguishable from that shown in mammalian cells. Phosphorylation of pRB in yeast requires two known cyclin-dependent kinases (CDKs), Cdc28 and Pho85, whose activities are totally dependent on their interactions with cyclins. Expression of pRB in a series of isogenic yeast strains lacking various combinations of endogenous G1 cyclins (Cln1 + Cln3 or Cln2 + Cln3) are required for the pRB hyperphosphorylation by CDKs. Intriguingly, mammalian cyclin D1 and cyclin E can complement the function of Cln2 and Cln3, respectively, in the induction of pRB kinase activity in yeast. Recent genetic studies have shown that activity of Pho85 kinase in late G1 is regulated by newly identified G1 cyclin-like molecules denoted Pc11 or Pc12. Importantly, expression of Pc11 or Pc12 in late G1 is totally dependent on the active Cln3-Cdc28 kinase. This implicates that pRB hyperphosphorylation in yeast is mediated (at least) by two distinct cyclin-CDK complexes, Cln1/2-Cdc28 and Pc11/2-Pho85, the latter kinase activity is positively regulated by Cln3-Cdc28 Consistent with the idea that multiple cyclin-CDK complexes cooperate in pRB inactivation, pRB hyperphosphorylation can be reconstituted in Pho85(-) cells only when human cyclin D1-CDK4 and cyclin E-CDK2 are simultaneously expressed in the yeast cell. Requirement of multiple cyclin-CDK kinase activities in pRB hyperphosphorylation/ inactivation implies that a variety of growth-related signals impinge on pRB. When phosphorylation status of pRB reaches to a certain threshold, the protein is suddenly inactivated. In the absence of functional pRB, cells are now able to proceed biochemical process that leads to G1-to S-phase transition of cell cycle.
Dr. Eileen White described regulation of programmed cell death (apoptosis) by the adenovirus E1A and E1B oncogenes, which is important for sustaining a productive infection in human cells and for transforming rodent cells. The E1A proteins initiate cellular proliferation which causes p53 accumulation and apoptosis. The E1B gene encodes overlapping, redundant functions to suppress p53-dependent apoptosis, the 19K and 55K proteins. The E1B 55K protein complexes with and inhibits p53 directly, whereas the E1B 19K protein is the adenovirus functional equivalent and homologue of Bcl-2. Expression of the E1B 19K, E1B 55K, or Bcl-2 proteins permits E1A expression and subsequent growth deregulation to occur unimpeded by cell death. Without inhibition of p53-dependent apoptosis by E1B or Bcl-2, transformation of rodent cells is rare and premature death of productively infected human host cells impairs virus yield. IN cells driven into apoptosis by E1A and p53, E1B 19K and Bcl-2 protein expression prevent apoptosis but the growth arrest function of p53 remains intact. Thus, the apoptotic and growth arrest functions of p53 are separable, and Bcl-2 expression may represent a cellular mechanism for controlling the activity of p53. p53 is a transcription factor that can both activate and repress transcription. Mutant p53 containing a crippled activation domain is unable to induce apoptosis. p53 must therefore induce apoptosis by either activating the transcription of death genes such as bax or by repressing the transcription of survival genes such as bcl-2. Bax mRNA and protein levels were dramatically elevated by conformational change of p53 to the wild-type form. Furthermore, bax expression induced apoptosis in cells expressing mutant p53 indicating that Bax acts downstream of p53 and is alone sufficient to induce apoptosis. To establish the mechanism of regulation of apoptosis, E1B 19K interacting cellular proteins have been identified using the yeast two-hybrid system. One 19K binding protein is Bax, the previously identified functional antagonist of Bcl-2, and another was NBK (natural born killer). E1B 19K expression prevents Bax-induced apoptosis suggesting that the 19K protein acts similarly to Bcl-2 by binding to the same death promoter Bax to prevent apoptosis. Thus, p53 transcriptionally actives bax transcription which then induces apoptosis unless the Bcl-2 or E1B 19K proteins are present to bind to and inhibit the action of Bax. Residues 50 to 78 of Bax containing a conserved region designated Bcl-2 homology region 3 (BH3) were sufficient for specific binding to both the E1B 19K and Bcl-2 proteins. The Bax-E1B 19K interaction was detectable in vitro and in lysates from mammalian cells, and Bax expression antagonized E1B 19K protein function. In cells where p53 was mutant, Bax expression induced apoptosis suggesting that Bax was sufficient for apoptosis and acted downstream of p53. p53 may simultaneously activate the transcription of genes required for both growth arrest (p21/Waf-1/Cip-1) and death (bax) , and E1B 19K and Bcl-2 may act distally and function through interaction with an antagonism of Bax to prevent apoptosis. With the death pathway disabled, induction of growth arrest by p53 can then be manifested. NBK functioned similarly to Bax in that NBK expression antagonized the ability of the 19K protein to block apoptosis and could induce apoptosis in cells expressing mutant p53. NBK possess a BH3 domain but not sequences similar to BHI or BH2, suggesting that it associates with the 19K protein by a similar means as that of Bax. NBK and 19K associate in vitro and colocalize in mammalian cells. Thus, NBK represents a novel death regulator. The biochemical function of NBK and Bax in the regulation of apoptosis is under investigation.
Dr. Kyosuke Nagata concluded the session with a discussion of how the adenovirus (Ad) genome in virions and in infected cells at early stages of infection exits in the form of the complex of Ad DNA and viral basic proteins (Ad core). DNA replication of the Ad core is dependent on the host factor designated template activating factor-I (TAF-I) in addition to factors required for replication of the naked genome. The Nagata laboratory has purified TAF-I as 41 (TAF-I!!
!) and 39 (TAF-I!!!) kDa polypeptides from HeLa cells. TAF-I stimulated not only the replication of the Ad genome but also the transcription of viral genes when the Ad core was used as template in the cell-free systems. These observations suggest that in addition to factors essential for replication or transcription on naked DNA template, the factor such as TAF-I is required for replication or transcription from the template in a chromatin-like structure. Cloning and nucleotide sequence analysis of CDNAS encoding TAF-I revealed that the TAF-I!!!corresponds to the protein encoded by the set gene, which is the part of the putative oncogene associated with acute undifferentiated leukemia when translated to the can gene. TAF-Ia protein contained the same amino acid sequence as TAF-I!!!except that short amino terminal regions differ in both proteins. A particular feature of TAF-I proteins was the presence of a long acidic tail in the carboxyl terminal region, which is though to be an essential part of the SET-CAN fusion protein. Studies with recombinant TAF-I proteins indicated that the carboxyl acidic region of the TAF-I protein is required for the TAF-I activity in the cell-free replication and transcription of the Ad core. It is proposed that TAF-I causes the dissociation or the conformational change of core proteins, thereby activating replication and transcription from the Ad core. TAF-I has the significant amino acid similarity to nucleosome assembly protein-I (NAP-I), which is involved in generation of the chromatin structure. It was shown that TAF-I can be substituted by NAP-I in activation of the cell-free Ad core transcription system. Of interest was that TAF-I is shown to have the NAP-I activity. These observations suggest that this type of the molecular chaperone has the dual function, i.e., facilitating assembly of the chromatin structure and perturbing the chromatin structure. Studies of the mechanism of stimulation of the Ad core DNA replication and transcription by TAF-I as well as CAN equipped with TAF-I/SET will contribute to an understanding of the intrinsic function of TAF-I and the leukemogenic function of the fusion protein.Session Four, “Papovaviruses,” was moderated by Drs. Kazuo Sugamura and John Brady. Dr. Peter Howley began the session with a discussion of “Cellular Growth Proliferation and Suppression Medicated by HPV Gene Products.” Carcinogenic progression of human papillomavirus-infected cells is often associated with integration of the viral genome in a manner which results in the loss of expression of the viral regulatory protein E2. One function of E2 is the regulation of expression of the viral oncogenes, E6 and E7. Introduction of the bovine papillomavirus (BPV1) E2 transactivator (E2-TA) into HeLa cells, an HPV-18 positive cervical carcinoma cell line results in growth arrest. Dr. Howley describes studies from his laboratory in which they have found that the HPV-16 and HPV-18 E2 proteins share with BPV E2 the ability to suppress HeLa cell growth. This property was, surprisingly, not observed for a shorter form of E2 called the transrepressor which is also encoded by the E2 gene of BPV. Analysis of various mutant E2 proteins for growth suppression revealed a requirement for the intact transactivation and DNA binding domains. A HeLa cell line (HeLa-tsE2) which expressed a mutant E2 protein which was functional only at the permissive temperature, permitted an analysis of the molecular and cellular consequences of E2 expression. Data from his laboratory indicated that one mechanism by which E2 suppresses cell growth is through the repression of E6 and E7 transcription, thereby enabling the cellular targets of E6 and E7 to resume regulation of the cell cycle. The consequence of this is the stabilization of the p53 protein and induction of the cyclin-dependent kinase inhibitor p21, and an inhibition of cyclins as shown for cyclin E. Dr. Howley also described a structure-function study of the HPV 16 E2 transactivation domain, which revealed that the transcriptional activation and DNA replication function of E2 can be separated. The papillomavirus E2 proteins play key roles in viral replication, both as regulators of viral transcription and as auxiliary factors that act with E1 in viral DNA replication. Using an alanine scanning mutagenesis approach, single amino acid substitution mutants were made in the E2 transactiviation domain and assayed for various E2 functions. Analysis of these stably expressed mutants revealed that the transcriptional activation and replication activities of HPV 16 E2 could be dissociated. The data further suggested that the ability of E2 to complex with E1 is essential for E2 to function as an auxiliary replication factor.
Dr. Yoshiaki Ito described Polyomavirus enhancer binding protein 2 (PEBP2), also called Core Binding Factor (CBF), is composed of two subunits, a and b. One of the gene encoding the a subunit is AML1 which is a very frequent target of chromosome translocations associated with human leukemia. Neither the a nor the b subunit of PEBP2 have intrinsic transactivation domains. However, PEBP2 can stimulate transcription when the expression of the reporter plasmid is under the control of, for example, the minimal T cell receptor b (TCRb) enhancer which contains three PEBP2 and two Ets sites. In this enhancer element, PEBP2 functionally cooperates with Ets-1 and this activity requires the region of AML1 downstream of the Runt domain (the DNA binding and heterodimerization domain). AML1 physical associates with Ets-1 at the Runt domain and the C-terminal regional downstream of the Runt domain. The chimeric protein AML1/ETO(MTG8) generated as a result of t(8;21) does not transactivate the reporter plasmid containing the TCRb enhancer. It interferes with the function of AML1. Our working hypothesis is that PEBP2 stabilizes Ets-1 on DNA by the Runt domain and possibly modulates the transactivation domain of Ets-1 by the C-terminal region. By doing so, PEBP2 may contribute to form much higher ordered structure of the enhancer (enhanceosome) required for regulation of transcription. The chimeric protein may destroy the higher ordered structure and, therefore, it interferes with the transcription activation mediated by AML1. Dr. Ito’s laboratory is currently working to prove this hypothesis. PEBP2 has also been characterized with other assays. Polyomavirus DNA replication requires an enhancer. PEBP2 can stimulate replication alone in this case. Furthermore, a portion of AML1 linked to the DNA binding domain of GAL4 also is able to stimulate replication. By mutation analyses, small regions of AML1 have been identified which are responsible for the activity Dr. Ito hopes to obtain mutants which are perfectly normal for transcription activation but lack the ability to stimulate replication. These mutants will be very useful to ask whether the ability to stimulate polyomavirus DNA replication is required for some biological activities, such as induction of cell differentiation known to be associated with PEBP2. Eventually Dr. Ito hopes to answer whether the activities of some transcription factors detected only by the polyomavirus DNA replication assay are reflecting more general role of transcription factors in cellular DNA replication.
Dr. Thomas Roberts concluded the session with a discussion of middle T antigen (MT), which is the oncogene encoded by murine polyoma virus. MT transforms cells by mimicking a constitutively activated growth factor receptor. In the process middle T binds to a dozen or more cellular proteins. These include: the 63kd and 36kd subunits of protein phosphatase 2A, 27kd and 29kd forms of the 14-3-3 proteins, hsc70, the src family members src fyn and yes, the 85kd and 110kd subunits of P13 kinase, the 45kd and 52kd forms of shc, and GRB2. His talk first described how the MT complex is assembled and then went on to detail recent findings on the workings of various components of the complex. In particular three points will be discussed: Insights into P13 kinase function revealed by the construction and expression of a constitutively activated form of the 110kd catalytic subunit; new genetic studies on the shc binding site on MT; and finally, the. unexpected implications of heat shock protein binding for the virus.
Closing remarks were made by Drs. Howley and Yoshida in which they 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, who unanimously recommended that a follow-up seminar be organized in two years, again on viral regulatory factors and their interactions with host cellular factors
PARTICIPANTS
UNITED STATES
Dr. John Brady
Virus Tumor Biology Section
Laboratory of Molecular Virology
National Cancer Institute
Building 41, Room B403
Bethesda, MD 20892
Tel. 301-496-6201 Fax: 301-496-4951
E-mail: bradyj@dcc41.nci.nih.gov
Dr. Warner Greene
The J. David Gladstone Institutes
Gladstone Institutes of Virology and Immunology
P.O. Box 419100
San Francisco, CA 94141-9100
Tel. 415-695-3800 Fax: 415-826-1817
Dr. Peter M. Howley
Department of Pathology
Harvard Medical School
200 Longwood Avenue/D2-629
Boston, MA 02115
Tel: 617-432-2884 Fax: 617-432-2882
E-mail: phowley@warren.med.harvard.edu
Dr. Joseph R. Nevins
Department of Genetics
HHMI Research Laboratories
Duke University Medical Center
P.O. Box 3054
Durham, NC 27710-3054
Tel. 919-684-2748 Fax: 919-681-8973
Dr. Thomas Roberts
Dana-Farber Cancer Institute
44 Binney Street
Boston, MA 02115
Tel. 617-632-3049 Fax: 617-632-4770
Dr. Bernard Roizman
Viral Oncology Laboratories
University of Chicago
910 East 58th Street
Chicago, IL 60637
Tel. 312-702-1898 Fax: 312-702-3791
Dr. Thomas Shenk
Department of Molecular Biology
Lewis Thomas Laboratory
Princeton University
Princeton, NJ 08544-1014
Tel. 609-258-5992 Fax: 609-258-1704
Dr. Eileen White
Center for Advanced Biotechnology and Medicine
679 Hoes Lane
Piscataway, NJ 08854-5838
Tel. 908-235-5329 Fax: 908-235-5318
E-mail: ewhite@mbcl.rutgers.edu
JAPAN
Dr. Masanori Hatakeyama
Department of Viral Oncology
Cancer Institute
1-37-1 Kami-Ikebukuro, Toshima-ku, Tokyo 170
Tel: 03-5394-3880 Fax: 03-5394-3816
Dr. Yoshiaki Ito
Department of Viral Oncology
Institute of Viral Research
Kyoto University
Shogoin Kawara-cho, Sakyo-ku, Kyoto 606
Tel: 075-751-4028 Fax: 075-752-3232
Dr. Kyosuke Nagata
Department of Biomolecular Engineering
Faculty of Bioscience and Biotechnology
Tokyo Institute of Technology
4259 Nagatsuda-cho, Midori-ku, Yokohama 227
Tel: 045-924-5798 Fax: 045-924-5940
Dr. Takeharu Nishimoto
Department of Molecular Biology
Graduate School of Medical Science
Kyushu University
3-1-1 Maide, Higashi-ku, Fukuoka City 812
Tel: 092-641-1151 (ext. 3471) Fax: 092-632-2373
Dr. Kazuo Sugamura
Department of Microbiology
Tohoku University School of Medicine
2-1 Seiryo-cho, Aoba-ku, Sendai 980
Tel: 022-273-9073 Fax: 022-273-2787
Dr. Kenzo Takada
Cancer Institute
Hokkaido University School of Medicine
7 Nishi, Kita 15 zyo, Kita-ku, Sapporo 060
Tel: 011-716-2111 (ext. 5071) Fax: 011-717-1128
Dr. Mitsuaki Yoshida
Department of Cellular and Molecular Biology
Institute of Medical Science
The University of Tokyo
4-6-1 Shirokanedai, Minato-ku, Tokyo 108
Tel: 03-5449-5275 Fax: 03-5449-5421