(1) Seminar on “Genes Controlling Breast Tumor Behavior”
Naniwa Kaikan, Osaka, Japan
March 26 - 28, 1990
I. Present Status of Breast Cancer in Japan --
Dr. Haruo Sugano described the increasing incidence of both pre- and postmenopausal breast cancer, and discussed the various causes for these changes concluding that the most important effects were those of diet and life-style. Histopathologic characteristics of breast cancer in Japan do not differ much from Western countries, although there is a relatively high frequency of well-differentiated tumor types. In atomic bomb survivors, the incidence of
breast cancer is increasing parallel with radiation dose.
II. Growth Factors --
Dr. David Salomon discussed the estrogen-dependent TGF-alpha production and secretion in primary cultures of mammary carcinoma cells. Overproduction of this growth factor can lead to an autocrine-dependent transformation of the immortalized rodent and human mammary epithelial cells provided a sufficient complement of funcitonal EGF receptors are present.
Dr. Kent Osborne discussed the insulin-like growth factors I and II demonstrating that blockage of the IGF I receptor might offer a new treatment strategy to inhibit the growth of certain tumors. He also provided convincing data that blockage of the EGF receptor and the IGF I receptor does not inhibit estrogen-induced growth.
Dr. Anton Wellstein discussed the family of fibroblastic growth factors, and presented, data suggesting that the major role of these factors secreted from breast cancer cells is paracrine growth stimulatory activity.
Dr. Bunzo Sato described a novel androgen-induced growth factor which acts on SC-3 cells as an autocrine growth factor. It would appear that it belongs to the family of FGF growth factors.
Dr. Jun Shimazaki studied the expression of int-2 and hst-1 genes and believes that hst-1 plays a role in the secretion of FGF-like peptides which may explain the autonomous behavior of the Shionogi CS2 cell line.
III. Oncogenes and Tumor Suppressor Genes --
Dr. Robert Callahan described the expression of int-1, int-2, and int-3 in mouse mammary tumor cells. His laboratory has also studied the loss of heterozygosity at 8 sites on 6 chromosomes (1p, 1q, 3p, 11p, 13q, 17p, 17q, and 18q). He proposes that the progression to malignancy is partly a consequence of the collaborative effects of mutations at these sites.
Dr. Tadashi Yamamoto has studied the erb B-2 protein and, in particular, has studied mutants with truncated cytoplasrnic domains which seem to be constitutively activated. He also has preliminary evidence for a fragment of erb B-2 present in serum of cancer patients.
Dr. Yusuke Nakamura has also studied the loss of heterozygosity on numerous chromosomes with an extraordinarily large number of probes generated from his laboratory.
IV. Genes for Metastatis --
Dr. Mark Sobel presented an overview of the steps required for a cell to metastasize. He then concentrated his presentation on NM-23, a putative tumor suppressor gene which when expressed is associated with a lack of lymph node metastasis. He also described his work with laminin receptor which also seems to be an important requirement for metastasis.
Dr. Shun’ichiro Taniguchi presented convincing laboratory data on the role of the fos oncogene which enhances the metastatic potential and invasiveness of tumor cells.
Dr. Hitoshi Akedo described his innovative culture model for tumor cell invasion which can then be used as an assay for inhibitors of invasion. He also described the phenomena of exvasion which describes the movement of cells back through a gelated basement membrane into the original tumor nest.
V. Clinical Factors Predicting Tumor Behavior and Clinical Outocome --
Dr. William McGuire presented data on DNA flow cytometry, cathepsin D, stress responsive proteins, and oncogenes. He demonstrated that a combination of these factors could reliably separate axillary node-negative breast cancer patients into groups of low risk and high risk for recurrence and, thus, aid in the decision of whether to use adjuvant therapy.
Dr. Setsuo Hirohashi presented the experience of Japanese breast tumors and erb B-2. He found that amplification and expression were tightly correlated, and that erb B-2 was a significant predictor of relapse and survival in node-positive breast cancer.
Dr. Hiroki Koyama described his studies in axillary node-negative breast cancer patients in Japanese women and found that from a variety of factors studied, the most important was the mitotic index.
Dr. Goi Sakamoto described in detail the histological classification of breast cancer as described by the Japanese Breast Cancer Society. He compared the ten-year survival rates of Japanese patients with those from Vanderbilt University in the United States. He found that histological extension of tumor invasion is an important indicator for prognosis.
VI. Transgenic Breast Cancer in Mice --
Dr. Randall Moreadith presented a number of experiments where transgenic lines had been developed with mice carrying the int-2, c-erb B-2, ras, c-myc oncogenes. These oncogenes singly or in various combinations were able to produce breast tumors with a variety of histological patterns. In the case of int-2, benign prostatic hypertrophy was also developed suggesting an important new model for that disease process.
In summary, a tremendous amount of information was exchanged between the U.S. and Japanese scientists in the areas of growth factors, oncognes, metastatic genes, and the use of many of these factors predicting tumor behavior and clinical outcome. The meeting was an unequivocal success.
(2) Seminar on “Gastrointestinal Cancer; Ethnicity and Some Other High-Risk Groups”
This workshop was based on the observation that cancers in various parts of the aerodigestive tract differed markedly among ethnic groups in the United States with respect to the subsite and subtype of the cancer. The workshop, co-organized by R. W. Miller and H. Sugano, was held on March 10-11, 1990 at the Sheraton Waikiki Hotel in Honolulu.
Each of the five main ethnic groups in the United States has its own patch of tissue at unusually high risk, as reported by R. W. Miller (NCI, Bethesda). The differences were the same within each sex but males had substantially higher rates than females for each cancer subsite. The data were from the Surveillance, Epidemiology and End-Results (SEER) Program of NCI, 1973-1984.
As is well known, the Chinese in the U.S., as in South China from whence they came, have a high rate of nasopharyngeal carcinoma, and Blacks have the highest rates for squamous cell carcinoma of the mid-esophagus, which is less well known. Even more esoteric is the finding that the high-risk patch among Whites is the lower end of the esophagus. There, in Barrett's esophagus (replacement of the normal squamous cell lining with columnar epithelium) adenocarcinoma occurs. It is well known that high rates of stomach cancer occur among the Japanese, but not so well known out.side Japan is the subsite affected, the pyloric antrum. And finally, Native Americans (American Indians) have the highest risk for cancer of the gall bladder and bile ducts. The colon showed no dramatic difference by ethnic group, Whites and Blacks had the highest rates for cancer of the ascending colon.
Rates for cancer of the descending colon were highest among the Japanese, but not by much.
Cystic Finrosis of the Pancreas and Ileal Carcinoma
Miller also reported that three cases of adenocarcinoma of the ileum had been reported among persons who had survived cystic fibrosis of the pancreas to about 30 years of age. Cystic fibrosis is characterized by steatorrhea, as is non-tropical sprue which has been associated with adenocarcinoma or lymphoma of the jejunum. Possibly fecal mutagens in fatty bowel contents account for the high frequency of the adenocarcinomas at these rare sites.
Bloom Syndrome and Gastrointestinal Cancer:
Bloom syndrome, the main features of which are short stature, a sun-sensitive rash of the face, and a high frequency of sister-chromatid exchanges, is also associated with high risk of gastrointestinal cancers, as reported for the first time by J. German (New York Blood Center). These patients have a high frequency of leukemia and lymphoma during childhood, and carcinoma during early adulthood. Of those who have died among 132 studied, 75 percent succumbed to cancer. The eldest among them is only 46 years old. Seven have died of cancer of the colon, three of whom also had squamous cell carcinoma of the esophagus. One other member of the group had esophageal cancer and three had carcinoma of the tongue. Notably, stomach cancer has not yet been observed. Laboratory study of these cancer sites among persons with Bloom syndrome may lead to new understanding of their pathogenesis. Three of those with gastrointestinal cancer were Japanese. The possibility of using!!
!-carotene for chemoprevention was raised. Cancer Mortality among Koreans in Osaka:
The opportunity to study ethnic groups in Japan is limited. A relatively small group of Koreans has been studied, as reported by T. Hiyama (Center for Adult Disease, Osaka). Korean males in Osaka had significantly more deaths than Japanese, 1983-1987, from cancers of the liver and lung. Higher rates, though not statistically significant occurred among Korean males for cancers of the mouth, esophagus and pancreas. Among Korean females there was a significant excess of liver cancer, and a significant deficiency of cancer of the breast and lung, as compared with Japanese women. Presumably the differences are due to dissimilarities in life style; ie, smoking, drinking, and dietary.
Barett’s Esophagus and Adenocarcinoma
H. Garewal (Arizona Cancer Center, Tucson), stated that Barrett’s esophagus (BE) is a model premalignant lesion. In this disorder, the normal squamous cell lining of the lower esophagus is replaced by metaplastic columnar epithelium, which has 30-40 times greater cancer risk. BE can be readily distinguished histologically from the normal squamous cell lining of the esophagus, and is thus easily studied for its premalignant behaviour. In this way, an increase in ornithine decarboxylase (ODC) activity was found, especially when dysplasia was present, but is not related to polyamine levels. ODC activity is higher when BE tissue becomes dysplastic. Aneuploidy correlates with abnormalities detected by flow cytometry (FC), and is apparently of prognostic significance. The presence or absence of dysplasia and aneuploidy as indicated by FC has led to a classification, from type I (no dysplasia, normal FC), found in 76 percent of BE, to type 4 (both dysplasia and aneuploidy present), found in 4 percent of BE. BE cells studied in culture can show the response to chemopreventive drugs. Among other subjects, the effect of carcinogens remains to be studied.
Stomach Cancer:
Factors associated with high risk of stomach cancer are radiation exposure; blood group A, low socioeconomic status, diet and ethnicity said A. Nomura (Japan-Hawaii Cancer Study, Honolulu). Among the dietary suspects are high salt, high starch and N-nitroso compounds. Without intervention, stomach cancer rates have mysteriously declined progressively in many parts of the world. Studies of stomach cancer lag behind other gastrointestinal sites in identifying candidate gene loci.
K. Kawai (Kyoto Prefectural University) reported that his group has been using a dye-spraying technique to diagnose intestinal metaplasia endoscopically. This allows easy monitoring over a period of years. His group has surveyed one hospital in each of the 48 prefectures in Japan -- more than 20,000 patients, and determined the ratio of gastric to duodenal ulcers. The ratio ranged from 1.3 to 1.8 except in Okinawa, where it was low (0.83 in males and 0.83 in females, and the northernmost district, Hokkaido, where it was 3.09 for males, and 4.34 for females. The gastric/duodenal ulcer ratio correlates well as a marker .for atrophic gastritis, which may serve for international comparisons of atrophic gastritis as a precancerous condition.
In Japan two forms of gastric cancer occur: a) the differentiated or intestinal type, in which cancer cells form tubuli, and B) the undifferentiated or gastric type in which cancer cells do not form tubuli. In studies by H. Sugano (Cancer Institute, Tokyo), the histogenesis of microcarcinoma of the stomach (lesions less than 5 mm in their longest diameter), was clearly related to the surrounding mucosa. The poorly differentiated type was more common at younger ages and in women, whereas the intestinal type was more common in the elderly and among men. The poorly differentiated type is regarded as the basic (inherent) type, and the differentiated lesions are regarded as variable in frequency due to environmental influences.
G. N. Stemmermann (Kuakini Hospital, Honolulu) has become interested in ciliated cells of the stomach, which his group has found in the pyloric antrum, usually in the cystically dilated basal segments of glands, on the surface of which there is intestinal metaplasia. The ciliated cells are regarded not as metaplastic cells as others have thought, but as modified pyloric gland cells. Through ultrastructure studies these cells were found in 40 percent of 129 cancerous stomachs. The cells are associated with branching microvilli, as seen in pleural mesothelioma or epididymal cells, but not in the gastrointestinal tract. They have the immunohistologic and ultrastructural characteristics of antral epithelium. The authors concluded that these features are a reaction to stasis of mucus secretion and/or blocked migration. The close association between the cilia and long, branching compound microvilli suggests an origin in common.
Multiple alterations in growth factors and oncogenes during the progression of stomach cancer was reported by E. Tahara (Hiroshima University). Among other observations he noted that coexpression of the epidermal growth factor receptor and the estrogen receptor were found in 30 percent of scirrhous gastric carcinoma and reflect a high degree of malignancy. Also, restriction fragment length polymorphism analysis using “minisatellite” probes showed that loss of heterozygosity frequently occurs at chromosomes 1p, 5p and 12p in advanced gastric carcinoma. The progression and metastasis of the neoplasm is associated with the accumulation of multiple alterations in oncogenes, growth factors and tumor suppressor genes.
Enedoscopic Screening for Breast Cancer:
As a guest speaker, G. Sakamoto (Cancer Institute, Tokyo) dramatically described the diagnosis of breast cancer by intraductal endoscopy and biopsy. The probe-like endoscope is threaded deep into the ducts of the breast, and watching is not for the squeamish. Screening by this method detected 5 nonpalpable breast tumors among 467 women.
Biliary Tract Cancer in Japan:
From S. Tominaga (Aichi Cancer Center) we learned that the rate of biliary tract cancer is more than twice as high in Japan than in the U.S., and the rates among Japanese-Americans are intermediate. The rates in Japan are highest in northern Honshu. A case-control study was made in Niigata Prefecture of 109 patients with cancer of the gall bladder and 84 with bile-duct cancer. High-risk factors for gall- bladder cancer included a past history of biliary tract disease, a family history of cholelithiasis and a preference for oily foods. For biliary tract cancer high-risk factors included a past history of biliary disease, family history of cerebral vascular accident, a thin body habitus and a low caloric intake.
Laboratory Tests for the Prevention of Colon Cancer:
W. R. Bruce (Ontario Cancer Institute) described five short-term tests of causes and hypotheses concerning cancer of the colon. 1) The Salmonella test for fecal mutagens identified the highly mutagenic, presumably carcinogenic fecapentaenes, which on further testing proved not to be carcinogenic. 2) A search for an effect on the colon itself revealed that some carcinogens induce micronuclei, along with other nuclear aberrations. Extracts of feces of persons at risk for colon cancer revealed that they induced micronuclei, the major active responsible agent being 4-cholestene-3-one. Findings relating to animal or human carcinogenicity have not yet been reported. 3) Boluses of fat given to mice caused loss of cells followed by a wave of proliferation in the colonic epithelium. Fat with meals did not produce the same effect. The effect of fat boluses is inhibited by calcium, which suggests that calcium deficiency enhances carcinogenicity. 4) Colon carcinogens cause the formation of a small number of aberrant crypts, which can be readily recognized microscopically by staining with methylene blue. Initiators and promotors of colon cancer can be recognized in this way. It is not yet possible to seek aberrant crypt foci or microadenomas in the human, but the appearance of macroscopic adenomas can be scored. Patients who have had polypectomies are being placed on dietary clinical trials to evaluate the frequency of polyps subsequently. The time needed to perform these 5 tests is 2-400 days, as compared with more than 4000 days for a human cancer to proceed from initiation through progression. Two hypotheses have converging lines of evidence: 1) cooked protein produced potent mutagens that the crypt analysis assay suggests may be promotors or initiators of colon cancer. 2) High levels of dietary fat promote colon carcinogenesis.
Molecular Genitics of Colorectal Cancer:
A. P. Feinberg (Howard Hughes Medical Institute, Ann Arbor) reported that his group is observing several genetic alterations in colorectal carcinoma. Loss of heterozygosity has been seen on three chromosomes 5, 17 and 18, often within the same tumor. Non-random association of these gene alterations with one another but not with mutations in c-Ki-ras suggest that altered gene dosage rather than a simple Mendelian model is involved in the multistep genesis of colorectal cancer. Chromosomal instability leads to multiple gene losses which are associated, but may not be the primary cause of the neoplasia or its progression. This chromosomal instability apparently distinguishes left-sided from right-sided colorectal cancer. A much higher rate of allelic loss and hyperploidy occurs in left-sided tumors. Feinberg’s group is also studying altered DNA methylation in multistep transformation. Nude mice were treated with an agent that causes hypomethylation of DNA, and, from in vitro studies of genes in cells from colonies, comparisons are being made of pretransformant, transformant and non-transformed cells.
Y. Nakamura (Cancer Institute, Tokyo) spoke on his group’s search for the gene responsible for familial adenomatous polyposis (FAP). FAP is a genetic disorder, and thus differs from colorectal cancer in general, as discussed by Feinberg. The gene for FAP is on the distal arm of chromosome 5. Previous studies showed that two markers closely flanked the FAP gene. New studies by Nakamura’s group make use of somatic cell hybrids which contain a small piece of chromosome 5 as the only human component. Forty cosmid clones were isolated that span a 7 cM segment which should contain the FAP gene. From genetic linkage analysis using large FAP families, the maximum segment has been further narrowed to about 5 cM. By pulsed-field gel analysis a continuous physical map of the segment was made with 20 cosmid clones, 7 of which have been eliminated because they have shown no recombination, with a lod score of 2.7-11.8. (These laboratory studies are greatly aided by the availability of specimens from a large registry of patients in Japan with familial polyposis.)
The multiplicity of oncogenes related to cancer of the digestive tract, as studied at the National Cancer Center was summarized by T. Sugimura. They include coamplification of hst-1 and int-2 oncogenes in more than 50 percent of esophageal cancers, amplification of erbB-2 in well-differentiated adenocarcinomas of the stomach, and of K-sam in poorly differentiated gastric adenocarcinomas. Ki-ras activation was rare in stomach cancer, but has frequently been found in pancreatic cancer. When these tumors are highly malignant, Ki-ras and c-myc are amplified. Ki-ras activation was rare in hepatocellular carcinoma, but frequent in cholangiocarcinoma. Such knowledge is crucial. Sugimura added, for new developments in prevention, diagnosis, predicting prognosis and treatment.
Distinguishing Chemical from Spontaneous Tumo
Two methods have been used by W. W. Nichols (Merck, Sharp and Dohme, West Point, PA) to distinguish chemically induced from spontaneous liver tumors in CD-1 mice. In one method, PCR-amplified tumor DNA was sequenced in regions encompassing codons 12-13 and codons 59-61 of three genes (c-Ha-ras, c-Ki-rasand c-N-ras). Three chemicals carcinogens were used: DMBA, AAB and AAF. There was a low frequency (3/32) of ras mutations in spontaneous CD-1 liver tumors, but, among the three groups with tumors induced by chemicals, a higher frequency of ras activation was observed: DMBA 22/34, AAB 5/9, and AAF 3/8). Other studies suggested that ras mutations in spontaneous CD-1 liver tumors may be a late event in tumor progression, but may occur relatively early in the progression of chemically induced tumors. The other method uses DNA fingerprinting. Each tumor was assayed for genomic rearrangements by comparing its DNA fingerprint with the DNA fingerprint of its normal tissue counterpart. Changes in the bands of (chromosomal) minisatellites were sought and more frequently found in chemically induced than in spontaneous CD-1 liver tumors. The total band changes per number of tumors studied were 2/25 for as compared with 13/29 for those induced by DMBA, 6/8 for AAF, and 2/9 for AAB.
In Perspective:
It is at times difficult to judge the merits of a workshop. One participant noted that the information presented built progressively to a climax, from population studies to molecular biology. Undoubtedly the best feature of this workshop was that it brought together people with interests in common but who had not met before. Several new collaborations are being planned among those engaged in laboratory research.
The clinical observations from the first part of the meeting did not cross-connect with the laboratory discussions as well as they might have. The Japanese were interested in Barrett’s esophagus, which is apparently rare in Japan. Its relationship to gastroesophageal reflux may have a counterpart in the intestinalization of the pyloric antrum, a precursor of gastric cancer which occurs with the highest rate at this subsite. The discovery in the laboratory that cancers of the left side of the colon are “genetic” whereas those of the right side are not, provides the logic for epidemiologic grouping of colon cancers; i.e., dividing them at the splenic flexure. Laboratory research on fecal fat in carcinogenesis of the large bowel appears to be relevant to the epidemiologic observations that small bowel carcinomas, rare as they are, occur excessively in two diseases with fatty feces due to defective absorption. The idea was raised that chemoprevention be considered for persons with Bloom syndrome, who die young usually of cancer, often of the gastrointestinal tract. The results should be apparent within a few years. Perhaps there was more of an interaction than was realized at the time.
SEMINAR AGENDA AND PARTICIPANTS
(1) Seminar on “GENES CONTROLLING BREAST TUMOR BEHAVIOR”
Naniwa Kaikan, Osaka, Japan
March 26 -28, 1990
AGENDA