(1) Seminar on “Manifestation of Prostatic Cancer”
This seminar was held on February 18-20, 1991, at the Sir Francis Drake Hotel, San Francisco, California. The organizers were Dr. Nobuyuki Ito, Nagoya City University Medical School, Nagoya, and Dr. David G. Longfellow, National Cancer Institute, Bethesda, Maryland. There were six participants from Japan and eight from the United States. The purpose of the seminar was to discuss and exchange information on the various epidemiologic, etiologic, and biologic manifestations of multistage prostate cancer with particular emphasis on the similarities and differences seen between the expression of the disease in Japan and the United States.
In opening remarks by Drs. Ito and Longfellow, all participants were encouraged to explore and exchange fully on the diversity of topics to be covered and to enrich the discussions with their own experiences and viewpoints. It was noted that there has been a dramatic difference between the participating countries in the incidence of clinical prostate cancer. The low incidence seen in Japan, however, now seems to be approaching the U.S. rate with an approximate 30 -year time lag. This sets the stage for opportunity, collaboration and urgency for progress against this important cancer.
Dr. Longfellow then presented a brief review of the current activities in prostate cancer supported by the National Cancer Institute (NCI) and the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK). There are approximately 90 projects sponsored between the two institutes with grant instruments ranging from principal investigator initiated, typical RO1 grants to cancer centers and oncology groups. The FY91 expenditure exceeds $11 million (U.S.). The areas of current emphasis include cell and molecular biology, $4,038,246 (30), projects; immunology, $744,477 (4); therapy and treatment, $1,418,233 (10) (not including center and cooperative groups); epidemiology, $1,876,518 (11); diagnosis, $2,115,256 (16); and models, $735,236 (4). Approximately 70 of these projects are NCI supported. Dr. Longfellow then summarized the highlights of the study design for the upcoming Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial (PLCO) which is being developed by the NCI's Division of Cancer Prevention and Control. The prostate component of this trial is designed to determine in screenees aged 60-74 at entry whether screening with digital rectal examination plus serum prostate-specific antigen (PSA) can reduce mortality from prostate cancer.
Dr. Longfellow concluded with a report on the recent "NCI Roundtable on Prostate Cancer: Future Research Directions," which was held at the Tidewater Inn, Easton, Maryland on May 1-3, 1990. This organ systems program sponsored workshop developed lists of the ten most important topics meriting emphasis in the study of prostate cancer. The topics were developed under the categories of (1) Life and Death of a Prostate Cancer Cell; (2) Natural History, Pathology, and Prognostic Factors; (3) Risk Factors; and (4) Screening, Early Detection Diagnosis. The workshop report will be published in Cancer Research 51:2498-2505, May 1, 1991.
In the interest of having all meeting participants current with recent major events in the field, Dr. Gary Miller (University of Colorado, Denver, Colorado) then presented a summary of the “Third Tokyo Symposium on Prostate Cancer," Tokyo, Japan, on December 14-15, 1990, which was sponsored by the Japanese Foundation for Prostate Research and the Roswell Park Memorial Institute. He also reported on a satellite workshop on "Cellular and Molecular Biology of Prostatic Cancer Metastasis to the Bone," which preceded the Tokyo meeting at a conference center near Mount Fuji in Gotenba, Japan, on December 12-13, 1990. The Tokyo meeting will be published as a proceedings volume as has been done in the past. Upon completion of these reviews, the seminar continued with the scientific presentations.
Available data suggest a greater-than-100-fold difference between the lowest and highest incidence rates of prostate cancer, with high rates among U.S. blacks and low rates in Japan and other Asian countries. Both incidence rates and mortality rates have increased with time in most countries. These observations suggest an environmental component for the disease. Dr. Alice S. Whittemore (Stanford University School of Medicine, Stanford, California) described a collaborative, population-based case-control study of modifiable risk factors for prostate cancer among blacks, whites and Asians in Los Angeles, San Francisco, Hawaii, Vancouver and Toronto. Investigators in this study will use a common protocol and questionnaire to administer personal interviews to approximately 500 black patients, 500 white patients, and 500 Asian patients with histologically confirmed carcinoma of the prostate, and 1500 black, white and Asian population-based controls. Controls will be matched to cases on age, ethnicity and region of residence. The information requested from subjects will permit (a) testing of hypotheses relating to age- and ethnic-specific prostate cancer risks to diet, physical activity levels and body mass index; (b) examination of ethnic and age differences in distribution of etiologic factors and in relative and attributable risks for etiologic factors; (c) estimation of the fraction of interethnic rate differences attributable to differences in distributions of etiologic factors and to differences in relative risks for specific etiologic factors. Serum collected from a sub-sample of 600 control subjects will be analyzed for levels of sex hormones and prostate-specific antigen (PSA), a marker for prostate cancer and/or benign prostatic hypertrophy. These analyses will permit testing of hypotheses relating die~ physical activity and body size to serum hormones and to subclinical prostatic disease within and across ethnic subgroups in the population. They also will allow comparison of lifestyle characteristics of cancer cases to those of controls without abnormal PSA levels.
Dr. Whittemore also presented a hypothesis to explain many features of prostate cancer epidemiology. Specifically, each cell in low-grade prostate cancer tissue is at risk of transformation into a cell which produces a high-grade clinical cancer after a short period of growth. As a consequence, the volume of low-grade latent cancer tissue in prostates of men at any age determines their incidence rate for high-grade clinical cancer a few years later. Autopsy and incidence data for both whites and blacks support this conclusion with a growth period of about five years. The transformation rate is the same for blacks and whites, about.016 high-grade cancers per year per cc of low-grade latent cancer volume. The hypothesis explains the infrequent occurrence of clinical cancer despite the high prevalence of latent cancer, and the steep rise with age of clinical cancer incidence despite the slow rise with age of latent cancer prevalence. It suggests that low-grade cancer volume is a critical determinant of clinical cancer risk.
Dr. Tadao Kakizoe (National Cancer Center Hospital, Tokyo) then reported on a comparative study of pathology of clinical and latent prostatic carcinoma in Japanese males. Sixty-six cases of mainly stage C clinical prostate cancers treated by radical cystoprostatectomy or radical prostatectomy were compared with 45 cases of latent prostatic cancers. The latter specimens were obtained from autopsy samples at a hospital for geriatrics; cystoprostatectomy specimens for bladder cancer and autopsy samples for nonprostatic malignancy were obtained at the National Cancer Center. The specimens were examined in terms of the patient's age, Gleason grade and size of the cancer. Age, size and grade of clinical and latent prostate cancers are closely interrelated. Characteristic points of clinical cancer are that it arises between the ages of 50 and 77, has a high Gleason score and a large-sized tumor, while latent cancer is observed between the ages of 46 and 95 and has a small, low-grade tumor.
Based upon the theory that human cancer is the end-result of several successive cellular changes, it is reported that the logarithm of the death rate increased in direct proportion with the logarithm of age in 17 types of cancer, including prostate cancer. By comparing the age-specific incidence rates of latent cancer with those of clinical cancer of the prostate, there is another report that latent cancer was the result of a series of 3 events affecting the prostatic cells, whereas invasive, fatal clinical cancer was the result of an additional 4 or 5 cellular events. A similar finding also was reported in an animal experiment using ACI/Seg rats characterized by more than 80% of spontaneous microscopic prostate cancer by 36 months of age, while there was 16% of grossly manifest prostate cancer. Using this animal, Dr. Kakizoe's group conducted two experiments to compare the effect of high fat diet and low fat diet under isocaloric conditions, and the effect of high calorie diet and low calorie diet under isofat conditions. However, neither high fat diet nor high calorie diet affected the incidence of microscopic cancer in ACI/Seg rats. Multistep genetic alterations may be involved in development from latent to clinical prostate cancer in humans and animals.
Dr. Nobuyuki Ito (Nagoya City University Medical School, Nagoya), Co-chairman of the workshop, reported on the frequency of latent carcinomas of the prostate in aged cases and their relation to benign nodular hyperplasias and intraductal atypical hyperplasias using 303 prostates obtained at autopsy.
Prostate tissues were fixed in buffered formalin, serially step-sectioned vertical to the urethra at 3-mm intervals and examined microscopically. The cases ranged from 14 to 99 years old, numbers in the 70, 80 and 90 decades being 76, 86 and 27 cases, respectively. Overall incidence of latent cancer was 26.4% (80/303 cases). However, age-related increase in the incidence was evident; 7.7% in the 1950s and 59.3 % in the 1990s. Multiple development of latent cancer (more than 2 lesions) was noted in 46% of the cases. The proportion of well-differentiated carcinomas decreased with age, while that of poorly differentiated lesions increased. Among a total of 127 tumors, 80% were less than 10mm in diameter.
Intraductal atypical hyperplasias were found in 28.1% of the cases, an age- related increase again being found (37% in the 1990s). A tendency for atypical hyperplasia to be present more frequently in cases with than without cancer was evident. Benign nodular hyperplasias also increased in an age-related manner, from 5.1% in the 1950s to 89% in the 1990s. However, no clear relationship between development of latent cancer and nodular hyperplasia was found in individual patients. Ras oncogene product, P21 was immunohistochemically positive in 4 out of 16 cases (25%) of latent cancer and 4 out of 6 cases (67%) of clinical cancer. However, no mutation of Ha-ras was detected in 5 cases of latent cancer and 6 cases of clinical cancer by northern blot analysis. Thus the data revealed that latent cancer of the prostate develops at very high incidence in aged cases.
In order to clarify modifying factors of prostate carcinogenesis, a rat prostate cancer model given a carcinogen 3,2'-dimethyl-4-aminobiphenyl, was used. Rat strain-dependence and modification with testosterone, estrogen, and prolactin were found. No modification was seen with high fat (vegetable or animal), cholesterol, clofibrate, or antioxidants such as catechol, selenium, or!!
!-tocopherol. The workshop then considered the analysis of early stage prostatic cancer. The incidence of prostatic adenocarcinoma in the United States has recently surpassed that of lung cancer, making it the most common visceral cancer of adult males. Metastatic prostatic adenocarcinoma is still treated with palliative hormonal manipulation which has little impact on overall survival. Meaningful clinical intervention therefore, depends on detection of the disease at an earlier "curable" stage. Historical data regarding extension of malignant cells through the prostatic capsule suggests that survival time is adversely influenced once malignant cells have reached the periprostatic soft tissue. Potency sparing radical prostatectomy attempts to avoid the neurovascular bundles along the posterolateral margins of the gland. Prediction of which tumors have traversed the prostatic capsule, therefore, takes on increased importance.
Dr. Gary J. Miller (University of Colorado Health and Sciences Center, Denver, Colorado) and his colleagues have been studying the histology of early stage prostatic cancer. Histopathologic examination of radical prostatectomy specimens using step sections or “whole-mounts” provide a wealth of information regarding correlations with various clinical/pathological findings and should be considered a "gold standard." Representative sections are probably inadequate to assess capsular penetration since 6 or 8 such sections sample only approximately 20% of the average prostatic surface area. Whole-mount sections provide additional quantitative data regarding tumor volume and surface area of capsular penetration.
Dr. Miller and his group have studied over 100 radical prostatectomy specimens at 4 mm intervals using this technique. Their results do not confirm that grade and volume are directly correlated. Histologic grade is clearly related to capsule penetration. While more than 90% of tumors with Gleason's score sums of 2 or NPCP grade < or = II are confined to the prostate, only about 5 % of tumors with Gleason score sums > 7 or an NPCP grade of IV are still confined to the prostate. Volume also correlates with ability to penetrate through the prostatic capsule. Tumors with volumes ranging from 0.12 cc to 25.7 cc were found to have extracapsular extension. While only 16% of tumors with volumes < 1 cc have invaded through the capsule, 90% of those with volumes > 3 cc have done so. No tumors with volumes greater than 4.5 cc were confined within the prostatic capsule. In addition, in the 17% of patients who had seminal vesicle involvement, the mean Gleason sum was 7 and the mean tumor volume was 10.8 cc. Those patients without seminal vesicle invasion had a mean Gleason sum of 4 and a mean tumor volume of only 1.6 cc. Biopsy interpretation can be confounded by the fact that prostatic carcinoma is a multifocal disease. In the 105 patients they have studied, there are an average of 1.9 (range = 1 to 5) tumors per patient. DNA ploidy has been determined by image analysis on tumors of different size and grade for 19 patients. Large volume tumors of low or high histologic grade were found to contain significant ploidy abnormalities when compared to their low volume, low grade counterparts. Dr. Miller has also correlated whole-mount histopathology with transrectal ultrasonography (TRUS). Only approximately 50% of tumors identified in whole-mount sections were visualized during careful retrospective analyses of TRUS images. A number of benign processes such as cystic atrophy benign hyperplastic nodules and granulomatous prostatitis also produced "hypoechoic lesions." Using three-dimensional reconstruction they have compared the locations of tumors with their predicted positions based on sonographic findings. Those tumors which are “isoechoic” tend to have lower histologic grades and an epithelia/stromal ratio less than or equal to 1. Those tumors which image clearly, tend to have a higher epithelia/stromal ratio. Approximately 1/3 of prostatic carcinomas are limited to the anterior portion of the gland which is poorly evaluated by TRUS. Finally, a small number of cases in this series were found to have mixed echogenicity. Microfocal radiographs of slices from these regions revealed the presence of microcalcifications. These represent the necrotic debris in the lumens of comedocarcinomas.
Dr. Ryuichi Yatani (Mie University School of Medicine, Mie-ken. Japan) then discussed the geographic pathology of latent prostatic carcinoma. In terms of geographic distribution, there is considerable international variation in the frequency of prostatic carcinoma. According to cancer mortality statistics, the prostatic carcinoma mortality rate per 100,000 population in various countries is classifiable into three groups. The highest rate occurs in non-white groups in South Africa and the United States (U.S.), and the lowest rates are found among Asians. European countries and U.S. whites show an intermediate rate. From chronological aspects of cancer mortality rate; the U.S. shows no significant changes between 1960 and 1985. However, Japan has experienced considerable chronological increase in the corresponding 25 years.
Dr. Yatani has previously studied the frequency of latent prostatic carcinoma from 1965 to 1979 in five populations: U.S. blacks, U.S. whites, Colombians, Japanese living in Japan, and Japanese in Hawaii. The frequency of latent prostatic cancer, especially more aggressive or larger tumors, increased with increasing risk for clinical cancer. However, the geographic differences in the frequency of latent carcinoma are smaller than those of clinical tumors.
Since the previous study, they have examined another 660 prostate glands obtained at autopsy from 1982 to 1986 in Japan, and examined chronological changes in the frequency of latent carcinoma during the two periods. The age-specific frequency of latent carcinoma greatly increased from 22.5% in 1965-1979 to 34.6% in 1982-1986. This significant finding can be attributed to an increase in the frequency of less differentiated tumor (from 9.8% to 17.8%). The frequency of latent carcinoma and the well/moderately to poorly differentiated carcinoma ratio in the latter samples were comparable to those from U.S whites. Clinical cancer rates in Japan are still low when compared with those in the United States and Western European countries, but these findings may presage the time when these differences are greatly reduced.
The latent carcinomas arising from the inner portion of the prostate were compared with those from the outer portion. Among 1236 prostates examined from Japanese men, they selected 17 carcinomas which were small enough for the localization of the origin; all of these were located in the inner portion. Thirteen of the tumors were well differentiated in histologic type and did not reveal age-related tumor growth. There were no differences for ras-p21 oncogene products, HPA expression or tenascin. Mean Ag-NOR counts per nucleus were significantly lower in tumors arising from the inner portion than those from the outer portion. These results indicate that latent tumors originating from the inner portion of the prostate are less common, more differentiated, and probably less aggressive.
Future directions in our understanding of the biology of prostate cancer were then discussed by Dr. Donald S. Coffey (Johns Hopkins University School of Medicine, Baltimore, Maryland). Cancer is a disease of DNA organization and dynamic cell structure. Since the early studies of Virchow, it has been recognized that one of the primary histological characteristics of cancer is the wide variation observed in nuclear and cell shape. The pleomorphism is often associated with pleiotropism and genetic instability. It is believed that an of these tumor cell characteristics may be linked through the interaction of a tissue matrix system that connects structural elements from the extracellular matrix through the membrane, to the cytoskeleton and, finally, to the nuclear matrix that organizes the three-dimensional array of the DNA. Genetic instability develops cancer cell variants that express a wide range of tumor cell heterogeneity in relation to structure, drug response, and tumor properties. This permits the tumor to rapidly develop resistance to all known natural and synthetic drugs, whereas normal cells do not have this ability to produce resistance. Tumor cell heterogeneity provides a tremendous growth advantage to the tumor by producing this biological diversity that permits the tumor cells to escape a competitive environment that involves therapeutic manipulations. One possibility for genetic instability is that the DNA structure and function is altered by nuclear structural components such as the nuclear matrix. The large abnormality in nuclear structure and morphometry that is often associated with poor prognosis is determined by the underlying nuclear matrix structure that provides the framework and superstructure to the nucleus. The nuclear matrix is also the organizing center for the three-dimensional array of DNA and the fundamental basic unit of organization is the attachment to the matrix of over 50,000 DNA loops each anchored at their base to the nuclear matrix. Each of these loops contains approximately 60,000 base pairs of DNA and represents the replicon unit of DNA synthesis that is the unit amount of DNA that is synthesized between adjacent replicating forks. Replication is produced by these loops being reeled through fixed DNA replicating sites attached on the nuclear matrix at the base of each loop.
The nuclear matrix becomes a dynamic component of the nucleus and is capable of undergoing chemomechanical changes that reorganize the DNA during the formation of the chromosome in preparation for mitosis. The matrix is capable of interacting with important regulatory proteins such as transcription factors and steroid hormone receptors. Indeed, it is believed that during interphase, chemomechanical signals are transduced from the cell periphery to the nuclear DNA via the structural interacting tissue matrix system.
The alteration in cell and nuclear structure that is the basis for much of tumor pathology is only a "freeze frame" view of a far more dramatic and dynamic process occurring within cells. Pathologists primarily observed dead cells which have lost their dynamic properties. In contrast, when live tumor cells are recorded in time-lapse cinematography, they show a tremendous amount of dynamic properties including cell motion and motility. It is now possible to quantitate the various types of cell motility and to essentially provide a mathematical analysis of the shape and movements of individual cells. These measurements have been correlated with the metastatic ability of many types of tumor cells, including the Dunning prostatic adenocarcinoma rat model. It is possible to demonstrate that this increased cell motility that is associated with an increased metastatic property can be produced by transfecting cells with oncogenes that are involved in the cell signalling pathway such as the src or mutated ras oncogenes. In addition, it is observed that many growth factors also have the ability to induce rapid cell ruffling and lamellipodial extension that quickly down regulates as the cells become refractory to the growth factor; however, when a ras mutated oncogene is placed within the cell, the ruffling and lamellipodial extension continue but do not down regulate and constant motion appears to be induced. This increase in cell motion is accompanied by the ability of the cell to exhibit motility and translocate through tissues. Cell motility is one of the required parameters, but is not in itself sufficient for metastatic ability. In the 1960s, Dr. Sumner Wood, a pathologist at the Johns Hopkins University, was able to demonstrate tumor cell motility within the ear of a living rabbit by using time-lapse cinematography. It now appears that this motility is an essential component of metastatic tumor cells and can be induced by specific oncogenes and growth factors.
The changes in cell motility are also accompanied by a decreased negative charge on the membrane of the cell that can be measured by using cell electrophoresis techniques. This cell surface charge is measured as a zeta potential and reflects the effective surface charge in a moving boundary as the cell is electrophoresed through a capillary tube with electrodes applied at each end. It is also possible to measure this potential using cationic polymers containing fluorescent probes that bind to the negative charge on the cell surface and that can be measured by flow cytometric techniques. Although direct correlations are observed, it is still unknown what produces this increased membrane negative charge and exactly how it is related to cell motility and metastatic properties of the cell.
Dr. Coffey discussed the fact that many factors are required to predict the biologlcal response of an individual tumor. The relationship between the anaplastic properties of tumors, the total DNA content and ploidy, and how this is related to tumor cell heterogeneity and genetic instability were discussed. This is reflected in tumor grade and in static samples, and may be related to quantitative tumor morphometric analysis such as nuclear shape factors. The stage and tumor volume are related to the tumor growth time and the balance between cell replication and cell death. Presently there is much activity in the area of growth factors and the induction of cell replication and just minor progress has been made in understanding the cycle of cell death. Cell death is a programmed event and requires the activation of specific genes. New developments in prostate tumor growth and cell death were also discussed.
The overall metastatic potential of the cancer cell involves cell motility, degradative enzymes that allow the cell to escape its local environment, and specific genes that appear to be related to establishing growth. There is presently much progress being made in the area of studying cell adhesion molecules and the loss of uvomorulin and other adhesion molecules. Cell adhesion loss is required for the cells to lose their cell social interaction and binding, both important factors in the metastatic process.
The workshop then considered cytological and morphometric analysis of small foci of prostate cancer in humans. The natural history of the individual prostate cancer focus is not known, but the prevalence of cancer at autopsy greatly exceeds that of clinically apparent prostate cancer. It has been proposed that all small prostate cancers are early clinical cancers which progressively dedifferentiate with growth to gain malignant potential. Others propose a multistep concept of prostate cancer initiation and progression with some small cancers already having attained full malignant potential. The development of DNA ploidy analysis by static cytomety has enabled examination of individual cancer foci, and the associations between the ploidy value of each cancer and the pathological features predictive of aggressive malignant behavior are being explored by Dr. Peter T. Scardino, Baylor College of Medicine, Houston, Texas and his collaborators in the Department of Urology and the Department of Pathology at Baylor.
He reported on the results of examination of 30 radical prostatectomy specimens with early stage prostate cancer (3 A1, 9 A2, 12 B1, 6 B2). Sixty-three separate cancers were identified and the volume, grade, zone of origin, presence of extracapsular extension (ECE) and seminal vesicle invasion (SVI) were determined. Image analysis of Feulgen-stained tissue sections was used to determine ploidy. Overall, 62% of the cancers were diploid and 38% non-diploid. Ploidy correlated with volume: all cancers < 0.02 cc were diploid, 26% of foci 0.02 - 2.0 cc and 82% of foci > 2.0 cc were non-diploid. There were 16 cancers of transition zone origin ranging in size from 0.02 - 12.1 cc and only 1 (7.3 cc) was non-diploid. There were 47 cancers of peripheral zone origin (range 0.01 - 18.98) and 23 (49%) were non-diploid. Eight of the 23 non-diploid peripheral zone cancers were < 1.0 cc and 2 were only 0.03 cc. These small non-diploid cancers showed other features of aggressiveness; all were moderately differentiated (Gleason sum 5-6) and 3 had complete ECE.
This group has previously reported that, volume for volume, peripheral zone cancers have a greater malignant potential than transition zone cancers. DNA ploidy analysis further substantiates this difference. Cancers originating in the peripheral zone of the prostate acquire malignant features at smaller volume than cancers of transition zone origin. Many small peripheral zone cancers were non-diploid, while all small and moderate sized transition zone cancers were diploid. These results support a multistep theory of initiation and progression of prostate cancer in which tumors acquire full malignant potential at widely varying volumes.
Dr. Yoichi Arai (Kyoto University Faculty of Medicine, Kyoto) has been investigating the prognostic significance of prostate specific antigen (PSA) in endocrine treatment for prostate cancer in an effort to predict response. Dr. Arai was unable to attend the workshop; however, his talk was presented by Dr. Tadashi Itakizoe of the National Cancer Center Hospital, Tokyo. There are two groups of prostate cancer patients, one that responds well to endocrine therapy and one that does not. In patients with advanced prostatic cancer, both local tumor and skeletal metastasis are frequently composed of a heterogenous population of cells that differ in their androgen dependency. Such tumor cell diversity alters responses to therapy of both metastatic and local structures, thus critical problems arise in treatment and monitoring. There is a need for a simple test to evaluate the response to endocrine therapy.
It has been reported that initial decrease in the volume of the primary tumor evaluated by transrectal ultrasonometry is a parameter that can predict progression. But it still is difficult to identify prostatic carcinoma within the gland. Transrectal ultrasonometry of the prostate still does not offer certain important clinical information because systemic progression is not well correlated with local recurrence.
Objective response of bony lesions is assessed by skeletal scintiscanning which is sensitive but of poor specificity. Osteocalcin (OC) is a small protein, unique to bone and tooth dentine, which is synthesized in osteoblasts. This group studied 63 patients with newly diagnosed prostate cancer and showed that serum OC was elevated in stage D2 patients. OC levels changed at various rates after the start of endocrine therapy and the response seemed to correlate with the interval free of progression.
PSA is highly sensitive to endocrine therapy and is considered to be sensitive to overall changes in response. PSA was used to predict response to endocrine therapy. A total of 50 patients with newly diagnosed prostate cancer were studied, 4 with stage B disease, 11 with stage C, and 35 with stage D disease. The endocrine therapy included bilateral orchiectomy, diethylstilbestrol diphosphate and LH-RH analogue. ALI of the 50 patients responded to endocrine therapy. When the initial changes in PSA with treatment were correlated with the time to disease progression, the patients who had a decrease in their PSA levels of 90% or more at one month survived significantly longer, free of disease progression (p < 0.01). The patients whose PSA level fell to below 1.5 ng/ml by 6 months had significantly longer intervals free of progression (p < 0.01). When initial changes in PSA level were correlated with Gleason histological grade or tumor stage, no differences in the rate of decrease in PSA level were seen. The lack of correlation of decrease in PSA with any of these prognostic factors indicates that the predictive value of changes in PSA is independent of these well-established parameters.
These data suggest that PSA is a simple and objective parameter with which to predict response to endocrine therapy. Further, the phenomenology related to PSA indicates the heterogenous population of tumor cells.
The workshop then went on to consider the interaction of stromal and epithelial cells as a factor in tumor growth and progression. Dr. Leland W.K. Chung (University of Texas, M.D. Anderson Cancer Center, Houston) reported on work in his laboratory on stromal-epithelial interaction in prostate tumor growth and progression. To understand the mechanism of cellular interaction between stroma and epithelium during prostatic tumor growth and progression, his laboratory has developed several relevant cell lines from human and rat tissues to address the following questions: (1) What is the cellular basis for androgen-induced epithelial growth? (2) Can soluble growth factors (GFs) or extracellular matrix (ECM) alone confer signals between stroma and epithelium? (3) Is there differential expression of GFS or ECM by the relevant cell types? Their results showed that: (1) An androgen receptor positive but nontumorigenic fibroblast cell line derived from fetal urogenital sinus (rUGM) can confer androgen-induced growth response to androgen receptor negative human prostate (PC-3) and bladder (WH) epithelial cells in vitro. These results indicate that androgen-responsive fibroblasts may be the direct targets for androgen-induced growth responses seen in the prostate gland. (2) Non-dialyzable soluble GFs can stimulate in a reciprocal manner the growth of epithelial and fibroblast cell lines in vitro. rUGM conditioned media (CM) or ECM alone are capable of stimulating the growth of nontumorigenic human prostatic epithelial LNCaP cells in vivo. Moreover, a bone fibroblast cell line, its CM, or ECM also stimulated LNCaP tumor growth in vivo. These results suggest that close interaction between GFs and ECM may contribute to prostate tumor growth and progression in vivo. (3) Northern blot analysis of total RNA extracted from fibroblast and epithelial cell lines indicates that while epithelial cells contributed GFs (!!
!FGF,!!
!FGF, TGF!!!, TGF!!!, EGF), fibroblasts contributed ECM (fibronectin, collagen 1, 3, and 4) that may be the basis for the molecular mechanism of stromal-epithelial interaction, tumor growth and progression. The potential implications of these results on human benign prostatic hyperplasia and prostate cancer progression were discussed. Animal models of prostate carcinogenesis have particular potential for studying mechanisms and suspected risk factors of prostate cancer development. Dr. Maarten Bosland (New York University Medical Center) reported that prostate cancer can be induced in Wistar (WO) rats with a 5-15% incidence by a single injection of the carcinogen N-methyl-N-nitrosourea (MNU) following hormonal stimulation of prostatic epithelial DNA replication. These tumors are slow-growing adenocarcinomas with a variable degree of differentiation that develop in the dorsolateral prostate lobe and the coagnlating gland ( = anterior prostate). These tumors metastasize, have elevated plasma acid phosphatase, and cause urinary obstruction. Chronic treatment with a low dose of testosterone (T), given from two weeks after MNU injection onwards, results in elevation of mean plasma levels of T from 1.9 to 6.0 ng/ml; this is well within the normal range (0.6-8.7 ng/ml; n = 69). This treatment increased the incidence of dorsolateral prostate carcinomas to 50% and the carcinoma incidence of all accessory sex glands combined to 80%. Preliminary data indicate that when this dose of T is halved, not resulting in a measurable elevation of plasma T levels, there is still a marked increase in carcinoma incidence in comparison to animals that did not receive T treatment. Chronic treatment with only T (no MNU) also leads to a low incidence of dorsolateral prostate carcinomas (10-15%). Other investigators have found similar synergistic effects of exposure to MNU (Pollard and associates) or other carcinogens (Pour and coworkers) and chronic T treatment. These observations indicate that T is a tumor promoter as well as a carcinogen for the rat prostate at very low doses. Interestingly, some human populations that are at high risk for prostate cancer, i.e., U.S. black men, have been reported to have higher circulating T levels than lower risk populations, e.g., U.S. white men.
When Lewis (LEW) rats were used instead of Wistar rats, MNU did not induce prostate carcinomas, T induced a 0-4% incidence, and the combination of MNU and low doses of T resulted in 7-11% incidence. Thus, there are marked differences in the susceptibility to prostate cancer induction between rat strains. Human prostatic cancer is known to occur more frequently in some families than in others, suggesting a genetic component of prostate cancer risk.
Seven of 10 rat prostatic carcinomas induced by MNU and T contained activating G to A mutations in the second position of codon 12 of the K-ras protooncogene. This finding is consistent with the known mutagenic mechanism of MNU and this mutation is thus the likely first step in the multistage process of carcinogenesis in this prostate cancer system. Some human prostate carcinomas have been reported to contain K-ras oncogenes that are mutationally activated by the same G to A transition.
Combined chronic exposure to low doses of T and estradio1-17 leads to the development in the dorsolateral prostate of dysplasia in acinar and ductal epithelium within four months, and adenocarcinomas of ducts within 10-12 months in 100% of treated rats. This indicates that estrogens can play a role in rat prostatic carcinogenesis.
Epidemiological studies have suggested that dietary factors, particularly a high intake of fat and protein, are associated with increased prostate cancer risk. The influence of fat type (lard vs. sunflower seed oil) and amount (5 vs. weight 20%) and amount of protein (17 vs. weight 344%) in isocaloric semipurified diets on induction of prostate cancer by MNU or MNU + T was examined in Wistar rats. None of these dietary variables had a systematic effect on prostatic carcinogenesis, regardless of T treatment. These findings suggest that dietary fat and protein, if at all capable of modifying prostatic carcinogenesis, must act in concert with other critical determinants of prostate cancer risk.
Tomoyuki Shirai (Nagoya City University Medical School, Nagoya) reported on animal model work on the development of prostate carcinomas in rats and the promoting effects of testosterone. The potent carcinogen, 3,2'-dimethy1-4-aminobiphenyl (DMAB), was found to exert tumorigenicity in the prostate of F344 rats. In the first experimental series, administration of DMAB was synchronized with the peak of hormone dependent DNA synthesis in an attempt to enhance neoplastic development in this organ. Repeated subcutaneous (s.c.) injections of DMAB, at a dose of 50 mg/kg body weight (b.w.), were synchronized with peaks of prostate cell proliferation stimulated by testicular androgen following incomplete atrophy of the gland caused by three weeks of dietary ethinyl estradiol (EE) treatment. This regimen was repeated 10 times at five-week intervals. Microscopic carcinomas of the ventral prostate were found in 85.7% of rats by week 60, as compared to only 5.2% of animals given DMAB alone. However, when carcinogen exposure was synchronized to the cell proliferation phase after administration of methyltestosterone to orchiectomized rats, no carcinomas developed in the prostate. In a second experiment, rats were given the cyclic dietary treatment with EE and basal diet 10, 5, or 3 times, respectively, and during each basal diet period, single s.c. injections of DMAB were given at doses of 50, 100, and 167 mg/kg b.w., respectively, so that each group received the same total exposure to carcinogen. At the end of the experiment (week 60), the incidences of prostate carcinomas were 58.6, 45.0 and 25.9%, respectively.
In a third experiment, DMAB was given s.c. at a dose of 50 mg/kg b.w. at two-week intervals for a total of 10 times to animals which were implanted s.c. with testosterone propionate (TP)-containing silastic tubes. After cessation of carcinogen administration, rats received further continuous TP implant exposure to the end of the experiment (week 60). Invasive desmoplastic adenocarcinomas developed in the dorsolateral prostate, seminal vesicles and coagulating glands at an overall incidence of 84.2%. Metastasis was evident in the abdominal cavity, liver or lung of some animals. None of the control groups had equivalent tumors. The development of ventral carcinomas was not enhanced b treatment with TP.
The data thus clearly showed that carcinogen administration during periods of cell proliferation is very effective for induction of ventral prostate carcinomas and that TP can exert strong promoting effects on tumor development and progression in the dorso-lateral prostate as well as in the seminal vesicles. Experimental approaches utilizing these two findings should yield good animal models for understanding induction of carcinomas of the prostate.
Dr. Jun Shimazaki (Chiba University, Chiba-shi, Japan) reported on studies of the progression of an androgen-dependent tumor to an autonomous one. These studies are stimulated by the clinical observation that approximately 80% of advanced prostatic cancer patients show a response to endocrine therapy at the start of treatment but more than half of the responders gradually lose responsiveness. When grade is compared at pretreatment and at relapsed state, an increase in the Gleason pattern is observed in many cases. This may imply that clonal change is an important factor for the progression of tumors from endocrine therapy-responsive to endocrine-resistant. To clarify the progression further, a study was performed using Shionogi Carcinoma 115 cells (SC 115) and an androgen-dependent mouse tumor, and its autonomous subline, Chiba Subline 2 (CS 2).
At transplantation of SC 115 to mice, castration will abolish proliferation of the tumor, but in medium or large tumors, castration only retarded or did not influence tumor growth respectively. Although SC 115 tumors of large size seemed to be autonomous, the next generation of tumor maintained androgen sensitivity, indicating that SC 115 scarcely changed sensitivity to androgen under various conditions. These results suggested mass-dependent autonomous growth of androgen-dependent tumors.
CS 2 was obtained during serial transplantation of SC 115, and is distinguished by cell size, amount of androgen receptor, and karyotype, but both tumors share common marker chromosomes. When mixed tumor composed of SC 115 and CS 2 was transplanted in mice CS 2 showed a predominant growth over SC 115, and this may be partly explained by a large necrotic tendency of SC 115. Cloned cell lines from SC 115 and CS 2 were obtained. Although SC 115 did not grow in serum-free medium without androgen SC 115 could proliferate in mixed culture with CS 2 even in the absence of androgen, suggesting stimulation to SC 115 by CS 2. A growth factor which stimulated growth of SC 115 and CS 2 was partially purified from serum-free conditioned medium of cultures with CS 2. The factor was a fibroblast growth factor-like peptide with molecular weight of approximately 50,000. Suramin reversibly inhibited the growth of SC 115 and CS 2 in a dose dependent manner. Growth promoting activity evoked by conditioned medium of CS 2 was inhibited by suramin, but suramin had no effect on growth factor production of CS 2, suggesting that suramin inhibited the growth of SC 115 and CS 2 by blocking the binding of growth factors to their receptors.
Hybridization of genomic DNA with long terminal repeat (LTR) of mouse mammary tumor virus (MMTV) revealed the same pattern of restriction fragments, showing the same integration of MMTV. Expression of transforming oncogene, hst-1, was found in CS 2 but not SC 115. Southern blot analysis revealed that there seemed to be some DNA rearrangements around hst-1 gene in CS 2. Therefore, CS 2 may acquire the ability of autonomous growth through expression of hst-1. In conclusion, through complicated autocrine and paracrine growth control, progression occurs in mixed clones, and gradually adaptation may occur among them.
Dr. Norio Nonomura (Osaka University Medical School, Osaka) reported on his work concerning the molecular mechanism of androgen-dependent growth of Shionogi carcinoma 115 cells (SC 115). The growth of some cancer cells has been well known to be regulated by steroid hormones. Recently, it has been recognized that the effect of steroid hormones on the growth of hormone-dependent cancer cells might be mediated through specific polypeptide growth factor(s). SC 115 has been established for more than 20 years as an androgen-responsive mouse mammary tumor. The growth of an androgen-dependent cloned cell line (SC 3) from SC 115 was markedly stimulated in serum-free conditions by a physiological concentration of testosterone. Another cloned cell line (SC 4) whose growth was unaffected by androgen, proliferated well only in high cell density condition. Analysis of growth-promoting activity in the conditioned medium from androgen-stimulated SC 3 cells revealed that the growth-stimulatory effect of androgen on SC 3 cells might be mediated through an androgen-induced growth factor in an autocrine fashion. This growth factor proved to be fibroblast growth factor (FGF)-like from its high affinity to heparin-Sepharose. The growth of androgen-unresponsive SC 4 cells is stimulated by this androgen-induced FGF-like growth factor in a paracrine fashion with simultaneous morphological alteration from epithelial to spindle shape. High concentrations of glucocorticoid have been known to stimulate the growth of SC 115 to a lesser extent than androgen. Antibody (IgG) against basic fibroblast growth factor markedly inhibited the growth stimulation induced by androgen, glucocorticoid, and basic fibroblast growth factor to an equal extent (70%). These facts suggest that the growth-stimulatory effect of a high concentration of glucocorticoid may be mediated through the same or similar androgen-induced FGF-like growth factor. The fact that SC 3 cells have cell-surface FGF receptors suggests that this androgen-induced growth factor may exert its stimulatory effect on SC 3 cells through FGF receptors in an autocrine fashion. Among various growth factors analyzed, only transforming growth factor (TGF-
!!!)- inhibited the growth of SC 3 cells. TGF-!!!seemed to suppress the production of androgen-induced FGF-like growth factor. This serum-free culture system using SC 115 is considered to be a very useful model for experiments on the growth of androgen-dependent cancer such as prostatic cancer.Dr. Donald J. Tindall (Mayo Clinic Foundation, Rochester, Minnesota) then reported on recent work in his laboratory on the molecular mechanisms of androgen action in prostate cancer. Androgens play an important role in the development and secretory properties of the prostate. However, few genes in the human prostate have been described to be regulated directly by androgens. Prostate-specific antigen (PSA) is a member of the kallikrein gene family and is expressed exclusively in human prostatic epithelial cells. PSA protein has been shown to exhibit proteolytic activity that is responsible for liquefaction of semen. The importance of PSA, in addition to its role in reproductive physiology, lies in its usefulness as a tumor marker in the management of prostate cancer. Interestingly, a gene closely related to PSA, human glandular kallikrein-1 (hGK-1), has also been found to be expressed only in the human prostate. PSA and hGK-1 have similar gene structures and are located adjacent to each other on the same chromosome. In addition, there is a high degree of homology in both nucleotide and amino acid sequences. Comparative analysis of these two genes should provide insight into the relationship of their evolution, regulation of expression, and biological functions. They examined by Northern blot analysis whether PSA and hGK-1 mRNA are under androgenic control and found that both PSA and hGK-1 mRNA in LNCaP cells were inducible upon addition of mibolerone (3 nM) or dihydrotesterone (DHT) (10 nM) in culture medium. The addition of diethylstilbestrol (50 nM) or dexamethasone (50 nM) did not induce either PSA or hGK-1 mRNA in LNCaP cells. Furthermore, like PSA, hGK-1 mRNA could be repressed by an antiandrogen, hydroxyflutamide, in the presence of DHT. To further assess whether the androgen-mediated induction of these two genes is in part due to transcriptional activation~ a truncated form of bacterially-expressed human androgen receptor (hAR) was used to study its interaction with putative steroid response elements (SRE) from these kallikrein genes. Using a gel retardation assay, it was demonstrated that a 15 bp region in the 5'-flanking region of the PSA gene formed a complex with the truncated hAR. Moreover, preliminary data obtained from transfection experiments using a PSA 5'-flanking DNA fragment linked to a CAT reporter gene showed that the putative SRE may confer the androgen-mediated transactivation. These data suggest that both PSA and hGK-1 genes in LNCaP cells are directly regulated by androgens. Further elucidation of androgen receptor-induced transactivation of the human prostate-specific kallikrein genes is in progress.
The seminar concluded with an enthusiastic discussion by all participants regarding optimism for newly recognized research opportunities in prostate cancer. Dr. Ito summed up the seminar by noting that this was not an "ending" but a 'beginning" for a very valuable interaction. All participants strongly encouraged a repeat of this meeting in a few years.
(2) Seminar on “Hepatitis C Virus, A Causative Infectious Agent of Non-A, Non-B Hepatitis: Prevalence and Structure”
This conference was held on March 11-12, 1991, on Kauai, Hawaii. There were eight participants from the United States and fourteen participants from Japan. The conference was opened with introductory remarks by Dr. Richard Adamson, National Cancer Institute, Bethesda, Maryland; Dr. Kenichi Kobayashi, Kanazawa University, Kanazawa, Japan; and Dr. Edward Tabor, National Cancer Institute, Bethesda, Maryland.
There has been a high level of interest in the possible role of hepatitis C virus (HCV) as a cause of hepatocellular carcinoma (HCC). This has been due in part to the recognition that about 60% of cases of HCC in Japan occur in patients with HCV infection, and the recent cloning and molecular analysis of HCV have provided the tools to study its relationship to HCC. This conference was designed to bring together many of the leading researchers working on HCV in Japan and in the United States. It was hoped that the exchange of information about recent discoveries would lead to rapid advances in our understanding of the causes of HCC.
Dr. Edward Tabor (National Cancer Institute, Bethesda, Maryland) presented the current evidence that HCV has a role in the etiology of HCC. Although HCC and the hepatitis B virus (HBV) are closely associated, some HCC patients have no evidence of HBV infection. Several studies in Japan have shown that the number of such patients doubled over the past 25 years. HCC in patients without detectable HBV infection has been documented in other countries too. Seventy-five to one hundred percent of HCC patients with no serologic markers of HBV lack HBV DNA in HCC tissue by hybridization studies (except in studies from France), and 50-90% lack HBV DNA in HCC tissue by polymerase chain reaction (PCR).
Antibody to HCV (anti-HCV), using assays to detect antibodies to the C-100 nonstructural antigen (anti-C-100), has been reported in 54-69% of HCC patients in Japan and Europe without serologic markers of HBV. However, all of these studies were conducted using stored serum, in most cases stored from 5-30 years; false-positive results often occur when these assays are used to test stored serum~ or when they are used to test serum with high globulin levels, as may occur in patients with cirrhosis. Thus, these findings will have to be re-evaluated using the improved "second generation" of assays that are expected to become generally available during the next year.
There have been seven well-documented reports of patients who were prospectively followed and whose chronic non-A, non-B hepatitis progressed to HCC. In one of these, the progression to HCC was documented in five biopsies and an autopsy, progressing through unresolved viral hepatitis, chronic persistent hepatitis, chronic active hepatitis, cirrhosis, and HCC. The role of cirrhosis in HCV-associated HCC (found in 86-100% of anti-HCV-positive HCC patients) is an important issue because of the need to establish a mechanism by which HCV could contribute to the development of HCC, since according to currently available data HCV does not become integrated in the host cell. Dr. Amy Weiner (Chiron Corporation. Emeryville, California) stated that no HCV cDNA has been detectable by PCR in the livers of chimpanzees with prolonged courses of chronic HCV infection. Dr. Izumi Saito (National Institute of Health, Tokyo) said that no HCV cDNA could be detected by PCR in three human HCCs studied in his laboratory.
Dr. Kenichi Kobayashi (Kanazawa University, Kanazawa) discussed the etiologies of HCC in Japan. Anti-HCV (anti-C-100) can be detected in 77% of patients with HCC. Most of the HCC patients with anti-HCV can be shown to have active HCV infection based on the detection of HCV RNA in their serum by PCR. (Anti-HCV could be detected in 1.1% of 363,428 normal blood donors, ranging from 0.36% among donors age 16-20 to 2.78% among those age 51-64.)
In Japan, it appears that there may be an interaction of HCV and alcoholic liver disease in the etiology of HCC. Alcoholic liver disease may progress to cirrhosis, and HCV may cause alcoholic cirrhosis to progress to HCC. Over a five-year period HCC develops in 35% of anti-HCV-positive patients with alcoholic cirrhosis.
An unexplored area is the role of dual infections by HCV and HBV in some HCC patients. Dr. Kobayashi and his coworkers studied 14 patients whose serum contained both anti-HCV and HBsAg. Since the detection of anti-HCV (anti-C-100) can indicate either ongoing or prior HCV infection, the presence of dual infections was confirmed in 8/14 of these patients by detection of HCV RNA in their serum by PCR.
Dr. Daniel Bradley (Centers for Disease Control, Atlanta, Georgia) described the physical-chemical characteristics and structure of HCV. HCV is a positive-sense RNA virus, whose genome consists of one open reading frame. It has similarity to the flaviviruses and pestiviruses in overall genomic structure and in phenotypic characteristics. It is 10 kb in length, smaller than most of the flaviviruses. It has not been visualized by electron microscopy; it is known that an HCV preparation remains infectious after passing through 50nm filter, indicating its small size
The HCV genome includes regions coding for the virus core (c) and envelope (e), and five nonstructural coding regions (NSl to NS5). NS3 codes for protease and helicase, NS4 codes for a membrane binding function as well as C-100, and NS5 codes for an RNA-dependent RNA polymerase, homologous to that of the related flaviviruses and pestiviruses. Subgenomic RNA messages have been detected from both the 5'- and 3'- ends of the virus; at present it is not known whether these are defective virus particles, such as those found in other flaviviruses and pestiviruses (ratio of 100:1 for yellow fever virus).
Dr. Kunitada Shimotohno (National Cancer Center, Tokyo) described a group of HCV isolates from Japan (prototype "HCV-J") that differ from the prototype isolate studied in the United States (by the Chiron Corporation). HCV-J has been sequenced almost in its entirety. Unlike the results reported for the United States isolate, no poly-A tail was seen on HCV-J. At the nucleotide level, a 14-17% difference was seen between the United States strain and HCV-J, compared to a 2.5-11% difference among 19 isolates from Japan, based on an analysis of 275 nucleotides in the NS5 region. This kind of variation was seen in analyses of other regions of the genome as well.
The United States strain also has been found in some Japanese patients. In four Japanese hemophiliac patients, the United States strain was found in three and HCV-J in one; none had both. Presumably the United States strain was acquired from infusions of antihemophilic factor concentrates manufactured in the United States prior to the use of heat treatment to inactivate viruses.
Hypervariability was found from one clone to another in the two envelope proteins, gp35 and gp70. Similar variability has been found in virus isolates obtained from a single patient after the passage of time. Such variability could be a mechanism that permits the virus to persist in the infected host.
Dr. Shimotohno described a new enzyme immunoassay to detect antibodies to the core protein of HCV (anti-core), and studies to compare it with assays to detect antibodies to the C-100 region of NS4. Of 200 serum samples from patients with non-A, non-B hepatitis, 141 were positive for anti-C- 100 (139 of whom were also positive for anti-core), 39 were positive for anti-core alone, and 55 were negative by both assays (some of these were positive for HCV RNA by PCR). He suggested that the anti-C-100 assay only detects antibody that appears late in the course of infection, whereas the anti-core assay detects an antibody that often appears earlier in the course of infection.
Dr. Izumi Saito (National Institute of Health, Tokyo) discussed several "second generation" assays for detecting antibodies to HCV, including antibodies to the product of a cDNA segment of the N-terminal half of the NS-1 gene ("anti-N1519"), antibodies to an envelope protein and antibodies to the core protein (~anti-p22"). Tests for anti-p22 appeared to detect acute infections earlier than the anti-C-100 assay, detecting antibodies as early as the time of first elevation of serum alanine aminotranferase (ALT) in acute infection; thus this assay might be better than anti-C-100 for screening blood donors.
Dr. Amy Weiner (Chiron Corporation, Emeryville, California) discussed the gene structure and function of HCV. HCV appears to be a non-integrating virus; no DNA intermediates have been found. Although HCV is a 9.5 kb positive-strand RNA virus, negative strands have also been identified by PCR, but their existence has not been explained. The gene structure and protein products are: