Progress in cancer chemoprevention: perspectives on agent selection and short-term clinical intervention trials

The basic cancer-related chemical and biological sciences, pathology, and epidemiology have contributed to the understanding that antimutagenesis and antiproliferation are the important general mechanisms of chemoprevention and to the development of antimutagenic and anti-proliferative agents as potential chemopreventive drugs. These disciplines have also provided the biochemical and histopathological bases for identifying intermediate biomarkers that can be used as surrogate end points for cancer incidence in clinical chemoprevention trials and for selecting cohorts for these trials. Particularly important as histological biomarkers of cancer are the cytonuclear morphological and densitometric changes that define intraepithelial neoplasia (IEN). IEN changes are on the causal pathway to cancer. They may serve as target lesions in Phase II chemoprevention trials and as standards against which other earlier cellular and molecular biomarkers can be evaluated. Strategies for the clinical evaluation of chemopreventive agents have been defined for seven targets--colorectal, prostate, lung, breast, bladder, oral, and cervical cancers. Cohorts have been identified for short-term Phase II trials that investigate the effects of chemopreventive agents on IEN and on earlier biomarkers. Patients with adenomas serve as a cohort for trials in colon. One cohort for Phase II trials in prostate is patients with early stage cancers scheduled for prostatectomy; another is patients with prostatic intraepithelial neoplasia (without prostatic carcinoma). Patients treated for lung cancer are at high risk for bronchial dysplasia and second cancers; such patients are a cohort for Phase II trials in lung cancer. Presurgical breast cancer patients and patients with ductal or lobular carcinoma in situ are cohorts for studies in breast. Patients with superficial bladder cancers (Ta/T1 with or without carcinoma in situ) are cohorts for studies of chemoprevention in bladder, and patients with dysplastic oral leukoplakia are evaluated for chemoprevention of oral cancers. Cervical intraepithelial neoplasia is a prototype IEN, and patients with cervical intraepithelial neoplasia are a cohort for studies of cervical cancer.

Chemopreventive drug development: perspectives and progress

Chemoprevention drug development has the goal of identifying safe and effective chemopreventive agents for clinical use. Several distinctive strategies are pursued in developing chemopreventive agents: (a) identifying and validating predysplastic and early dysplastic lesions that can be used instead of cancers as endpoints for measuring chemopreventive activity; (b) identifying and testing candidate agents based on considerations of mechanisms of action; (c) evaluating combinations of agents with potential for maximizing efficacy and minimizing toxicity; and (d) applying a systematic methodology for identifying and ranking candidate agents at each stage of development to ensure discovery of the best agents and most effective use of available resources. This article discusses 22 drugs and three drug combinations which have reached an advanced stage of development as chemopreventive agents. The first generation of drugs are the most advanced, now being in Phase II and Phase III clinical trials. These drugs include several retinoids [vitamin A, 13-cis-retinoic acid, all-trans-N-(4-hydroxyphenyl)retinamide], calcium, beta-carotene, tamoxifen, and finasteride. The second generation drugs are those in Phase I clinical trials. From most to least advanced, these drugs are 2-difluoromethylornithine, sulindac, piroxicam, oltipraz, N-acetyl-I-cysteine, aspirin, ibuprofen, carbenoxolone, 18 beta-glycyrrhetinic acid, and the combination of 2-difluoromethylornithine with piroxicam. The third generation includes agents with significant evidence of chemopreventive activity in animal models. These agents are now in preclinical toxicity testing. They are S-allyl-I-cysteine, phenhexyl isothiocyanate, curcumin, ellagic acid, fumaric acid, fluasterone, and the combinations of all-trans-N-(4-hydroxyphenyl)retinamide with oltipraz and all-trans-N-(4-hydroxyphenyl) retinamide with tamoxifen.

Mechanistic considerations in chemopreventive drug development

This overview of the potential mechanisms of chemopreventive activity will provide the conceptual groundwork for chemopreventive drug discovery, leading to structure-activity and mechanistic studies that identify and evaluate new agents. Possible mechanisms of chemopreventive activity with examples of promising agents include carcinogen blocking activities such as inhibition of carcinogen uptake (calcium), inhibition of formation or activation of carcinogen (arylalkyl isothiocyanates, DHEA, NSAIDs, polyphenols), deactivation or detoxification of carcinogen (oltipraz, other GSH-enhancing agents), preventing carcinogen binding to DNA (oltipraz, polyphenols), and enhancing the level or fidelity of DNA repair (NAC, protease inhibitors). Chemopreventive antioxidant activities include scavenging reactive electrophiles (GSH-enhancing agents), scavenging oxygen radicals (polyphenols, vitamin E), and inhibiting arachidonic acid metabolism (glycyrrhetinic acid, NAC, NSAIDs, polyphenols, tamoxifen). Antiproliferation/antiprogression activities include modulation of signal transduction (glycyrrhetinic acid, NSAIDs, polyphenols, retinoids, tamoxifen), modulation of hormonal and growth factor activity (NSAIDs, retinoids, tamoxifen), inhibition of aberrant oncogene activity (genistein, NSAIDs, monoterpenes), inhibition of polyamine metabolism (DFMO, retinoids, tamoxifen), induction of terminal differentiation (calcium, retinoids, vitamin D3), restoration of immune response (NSAIDs, selenium, vitamin E), enhancing intercellular communication (carotenoids, retinoids), restoration of tumor suppressor function, induction of programmed cell death (apoptosis) (butyric acid, genistein, retinoids, tamoxifen), correction of DNA methylation imbalances (folic acid), inhibition of angiogenesis (genistein, retinoids, tamoxifen), inhibition of basement membrane degradation (protease inhibitors), and activation of antimetastasis genes. A systematic drug development program for chemopreventive agents is only possible with continuing research into mechanisms of action and thoughtful application of the mechanisms to new drug design and discovery. One approach is to construct pharmacological activity profiles for promising agents. These profiles are compared among the promising agents and with untested compounds to identify similarities. Classical structure-activity studies are used to find optimal agents (high efficacy with low toxicity) based on good lead agents. Studies evaluating tissue-specific and pharmacokinetic parameters are very important. A final approach is design of mechanism-based assays and identification of mechanism-based intermediate biomarkers for evaluation of chemopreventive efficacy.

Strategy and planning for chemopreventive drug development: clinical development plans. Chemoprevention Branch and Agent Development Committee. National Cancer Institute

At the National Cancer Institute, Division of Cancer Prevention and Control, the Chemoprevention Branch and Agent Development Committee develop strategies for efficiently identifying, procuring, and advancing the most promising drugs into clinical trials. Scientific expertise is applied at each phase of development to critically review the testing methods and results, and to establish and apply criteria for evaluating the agents for further development. The Clinical Development Plan, prepared by the Chemoprevention Branch and the Agent Development Committee, is a summary of the status of the agent regarding evidence for safety and chemopreventive efficacy in preclinical and clinical studies. It also contains the strategy for further development of the drug that addresses pharmacodynamics, drug effect measurements, intermediate biomarkers for monitoring efficacy, toxicity, supply and formulation, regulatory approval, and proposed clinical trials. Sixteen Clinical Development Plans are presented here: N-acetyl-l-cysteine (NAC), aspirin, calcium, beta-carotene, 2-difluoromethylornithine (DFMO), DHEA analog 8354, 18 beta-glycyrrhetinic acid, N-(4-hydroxyphenyl)retinamide (4-HPR), ibuprofen, oltipraz, piroxicam, Proscar, sulindac, tamoxifen, vitamin D3 and analogs, and vitamin E. The objective of publishing these plans is to stimulate interest and thinking among the scientific community on the prospects for developing chemopreventive drugs.

Development of breast cancer chemopreventive drugs

Breast cancer is the second highest cause of cancer mortality (19%) estimated for U.S. women in 1993 and accounts for the highest proportion of new cancer cases (32%) in this population. The rate of documented cases increased during the early 1970s and again in 1980-87, probably due to early mammographic detection. Increased knowledge of personal risk may also have been a consideration; however, 60% of women diagnosed with breast cancer have no known risk factor(s), such as family history, early age at menarche, late age at menopause, nulliparity, late age at first live birth, socioeconomic status, contraceptive use, postmenopausal estrogen replacement, or high fat intake. To prevent cancer, one strategy undertaken by the NCI is cancer chemoprevention, or intervention with chemical agents at the precancer stage to halt or slow the carcinogenic process. An objective of the NCI, DCPC is to develop promising cancer chemopreventive chemical agents as drugs for human use. Briefly, the process begins with identification of potential agents (e.g., pharmaceuticals, natural products, minor dietary constituents) from surveillance and analysis of the literature and from in vitro prescreen assays. Data on both efficacy (i.e., biological activities that either directly or indirectly indicate inhibition of carcinogenesis) and toxicity are gathered these sources. Various criteria are used to select and prioritize agents for entry into the NCI, DCPC preclinical testing program. The program begins with battery of in vitro efficacy screens using both animal and human cells to select agents for further testing; agents positive in these assays are considered for further testing. In the assay used for breast cancer chemoprevention, 7,12-dimethylbenz(a)anthracene (DMBA)-induced mouse mammary organ culture, 64 chemicals have inhibited formation of hyperplastic alveolar-like nodules. A panel of organ-specific animal screening assays are then used to assess efficacy in vivo. Two assays relevant for breast cancer chemoprevention are inhibition of N-methyl-N-nitrosourea- and DMBA-induced rat mammary gland carcinogenesis. Of 89 agents tested, 29 have inhibited cancer incidence, multiplicity, or both in at least one of the mammary assays; 21 agents are currently on test. Highly promising agents are then placed in traditional preclinical toxicity tests performed in two species. Finally, the most promising and least toxic agents enter clinical trials. Phase I clinical trials are designed to investigate human dose-related safety and pharmacokinetics of the drug. Phase II trials are small scale, placebo-controlled studies designed to determine chemopreventive efficacy and optimal dosing regimens.(ABSTRACT TRUNCATED AT 400 WORDS)

Intraepithelial neoplasia, surrogate endpoint biomarkers, and cancer chemoprevention

Neoplasia is a progression of molecular, cellular, and tissue changes starting with a critical cell mutation and advancing by clonal evolution, involving further multiple mutations and expanding mutated clones. This process is characterized by five general stages: latency, focal growth of normal-appearing but disorganized cells, abnormal-appearing cells (dysplasia), microinvasion, and finally, metastasis. The two driving forces of neoplastic progression in an epithelium are mutagenesis and mitogenesis. These forces frequently occur concurrently, produced by exposure of the epithelium to environmental and endogenous mutagens and mitogens. The major strategy of chemoprevention is to block the effects of both mutagens and mitogens during the early stages of predysplasia and dysplasia. Surrogate endpoint biomarkers (SEBs) are tissue, cellular, and molecular changes that correlate with the later development of cancer. Because of the savings in cost, labor, and time, SEBs are urgently needed to replace the use of cancer incidence reduction as the endpoint for chemopreventive agent clinical trials. The advent of computer-assisted cytometry allows each of the seven basic criteria of dysplasia to be individually assayed as an SEB. Since the dysplastic changes that characterize intraepithelial neoplasia are embodied in the causal pathway to invasive neoplasia, they are already validated as predictors of cancer incidence. More attention should be paid to the quality control of SEB assays, including control of variation in cell composition of tissue samples, assay protocol, instrumentation used, and observer performance. The dose-response relationship between a known chemopreventive agent and the SEB should also be evaluated. The Division of Cancer Prevention and Control, National Cancer Institute, has begun a program to test chemopreventive agents in short-term Phase II clinical trials using dysplasia-based SEBs. The SEBs are assayed, when possible, by computerized cytometry. Trials are being conducted for oral leukoplakia, cutaneous actinic keratosis, superficial bladder cancer, pulmonary metaplasia/dysplasia, cervical dysplasia (CIN III), and adenomatous colonic polyps.

Intraepithelial and postinvasive neoplasia as a stochastic continuum of clonal evolution, and its relationship to mechanisms of chemopreventive drug action

The progression of intraepithelial and postinvasive neoplasia depends on the occurrence of clonal evolution, defined as the continuous development of mutations and selective clonal expansions in the neoplastic cell population. The two continuously repeating events of clonal evolution, mutation and clonal expansion, occur at unpredictable times and locations. Therefore the neoplastic process is best characterized as a stochastic, i.e., probabilistic, continuum. The rate of intraepithelial neoplastic progression is continuously driven by the dosage level of exposure to mutagens and mitogens. For example, in chronic smokers the length of time before development of lung cancer depends on the number of cigarettes smoked per day. A commonly held misconception is that human carcinogenesis develops after an initial short period of mutation followed by a long period of stimulated proliferation (the multistage model). This incorrect idea derives from the sequential nature of the consecutive two- or three-step operational protocols imposed on experimental animal models by the experimenter. In reality, human carcinogenesis develops as the result of simultaneous and continuous exposure to mutagens and mitogens over the entire period of tumor development. A recent example is the finding that the intraepithelial neoplasia of colorectal adenomas continuously progresses through serial waves of mutation and clonal expansion. The rational design of chemopreventive agents should be based on blocking the two parameters which continuously drive neoplasia: mutagenesis and mitogenesis. In addition to blocking exposure, chemopreventive agents may act at many points during activation and DNA adduction of mutagens, or during stimulation of the proliferation signal pathway by mitogens. Based on the chemopreventive strategy of blocking mutagenesis and mitogenesis, chemopreventive agents are classed as either antimutagenic or antimitogenic. A third class, the antioxidants, are both antimutagenic and antimitogenic, and operate by the common mechanism of breaking free radical chain reactions initiated by reactive oxygen species. In the program of the Chemoprevention Investigational Studies Branch, Division of Cancer Prevention and Control, National Cancer Institute, preclinical development of antimutagens, antimitogens, and antioxidants is well under way, and some of these agents are highlighted here.

Selection criteria for breast cancer chemoprevention subjects

Early phase chemoprevention trials differ from standard therapeutic clinical trials because asymptomatic, healthy people are treated with a potentially toxic intervention for a prolonged period of time. Current subject selection protocols have relied upon epidemiological methods to identify high-risk individuals. Most available data provide risk estimates for various individual risk factors, but few have reported risk estimates for combinations of risk factors. Selection criteria for the large tamoxifen intervention trial (NSABP P1) were developed from the work of Gail et al. [1]. The Gail model takes into account non-genetic factors (e.g., nulliparity, age at menarche, preexisting pathological conditions) and genetic factors (family history). Using a lifetime risk of 10% of developing breast cancer as a standard to intervene, NSABP P1 uses the Gail algorithm to select pre- and postmenopausal women for a primary intervention trial. This approach has been criticized for being insufficiently selective (i.e., all women > or = 60 yrs), but appears to be the best available method to select subjects for a chemoprevention trial. Other approaches have been based on identification of very high-risk women with acknowledged pathologic conditions [lobular carcinoma in situ, ductal carcinoma in situ (DCIS)]. Attempting to use these proliferative lesions as pathologic endpoints for drug effect has not been attempted. DCIS as a risk factor for tamoxifen intervention was excluded because of controversies over its management and because of frequent difficulties in distinguishing microinvasive from non-invasive lesions. Women treated for early stage breast cancer (Stage I) may be subjects for early stage chemopreventive interventions.(ABSTRACT TRUNCATED AT 250 WORDS)

Development of chemopreventive agents for lung and upper aerodigestive tract cancers

The lung and upper aerodigestive tract (oral cavity, larynx, pharynx, upper esophagus) will harbor the greatest proportion (approximately 20%) of estimated new cancer cases in 1992. The estimated mortality rate is even higher (32%), which is reflected in a 5-year survival rate of only 7% and 12% for esophageal and lung cancer, respectively. Tobacco use appears to remain the major cause of aerodigestive cancers despite efforts at primary prevention--cessation of exposure. Another strategy to decrease this public health problem is secondary prevention or chemoprevention. Cancer chemoprevention is defined as intervention with chemical agents before invasion to halt or slow the carcinogenic process; potential agents may include minor dietary constituents and pharmaceuticals. The main objective of the Division of Cancer Prevention and Control (DCPC), National Cancer Institute, is to develop promising chemopreventive drugs for use in humans. The testing of cancer chemopreventives for efficacy in the clinic differs from that of cancer treatment drugs. Chemopreventive drug trials involve healthy target populations, and the endpoints are reduced cancer incidence or mortality, or increased latency, with no to minimal toxicity. The lung and upper aerodigestive tract represent a unique opportunity for intervention in this setting. Even with cessation of tobacco exposure, the risk of cancer in the entire epithelium remains high for years due to the "field cancerization" effect. Some of the first chemopreventive trials made use of this system due to the availability of a study population with a tissue at demonstrably high risk for malignant progression. Much of the evidence for chemopreventive efficacy is in the oral cavity because of the well-defined epithelial neoplastic progression, the existence of well-established preclinical models, and relative ease of tissue monitoring and sampling. In one of the first randomized trials, Hong and co-workers demonstrated that 13-cis-retinoic acid prevents the appearance of second primary tumors in patients previously treated for squamous cell carcinomas of the oral cavity and upper respiratory tract. Even using a high risk population, chemoprevention trials involve large sample sizes, lengthy duration and follow-up, and high cost. To circumvent these problems, the use of intermediate biomarkers as surrogate endpoints is being explored. Intermediate biomarkers are defined as biological alterations in tissue (histological, genetic, biochemical, proliferative, differentiation-related) occurring prior to cancer development. In the oral cavity, studies using modulation of a histological intermediate biomarker, dysplastic leukoplakia, as the endpoint have demonstrated response to a retinoid.(ABSTRACT TRUNCATED AT 400 WORDS)

Chemoprevention clinical trials

As part of a program to develop drugs which will delay or prevent cancer in humans, the Chemoprevention Branch, National Cancer Institute, National Institutes of Health, is sponsoring 12 Phase I and 22 Phase II and III clinical trials. Three agent classes are significantly advanced in the trials. These are the retinoids, including 13-cis-retinoic acid, retinol, and 4-hydroxyphenylretinamide (nine studies), beta-carotene (seven studies), and calcium compounds (three studies). In addition, six promising new compounds are in Phase I or Phase II trials. These are: piroxicam, ibuprofen, oltipraz (a dithiolthione), difluoromethylornithine, glycyrrhetinic acid, and N-acetylcysteine. Key concepts related to the development of cancer chemopreventive agents are (1) the need for long-term administration, (2) the need for oral route of administration, (3) the matching of toxic side effects to degree of cancer risk.

Screening for chemopreventive (anticarcinogenic) compounds in rodents

The Chemoprevention Branch is testing dozens of candidate chemopreventive compounds in the following rodent model carcinogenesis systems: mouse skin papillomas, DMBA/TPA induced, rat mammary adenocarcinoma, DMBA and MNU induced, hamster tracheal squamous cell carcinoma, MNU induced, and lung adenocarcinoma, DEN induced, rat and mouse colon adenocarcinoma, AOM and MAM acetate induced, respectively, and mouse bladder carcinoma, hydroxy BBN induced. Significant chemopreventive (i.e., anticancer) effects have been produced with 4-hydroxy-phenylretinamide, difluoromethylornithine, piroxicam, oltipraz (a dithiolthione), calcium glucarate, N-acetylcysteine, beta-carotene, ibuprofen, dehydroepiandrosterone (DHEA) and a 16-fluoro DHEA analog, 8354, tamoxifen, glycyrrhetinic acid, molybdate, selenite, curcumin, and fumaric acid.

Natural history of intraepithelial neoplasia in humans with implications for cancer chemoprevention strategy

Intraepithelial neoplasia is of critical importance to the cancer chemoprevention field because it is a target condition for which drugs must be sought that will prevent its development or stop its progression. The term "dysplasia" refers to the morphological alterations that characterize intraepithelial neoplasia and according to many authors consists of seven basic morphological changes that occur in the majority of human epithelia, as well as in the epithelium of mouse skin papillomas induced by 7,12-dimethylbenz(a)anthracene and 12-O-tetradecanoylphorbol-13-acetate: increased nuclear size; altered nuclear shape; increased nuclear stain uptake; nuclear pleomorphism (increased variation in nuclear size, shape, and stain uptake); increased mitoses; abnormal mitoses; and disordered or absent maturation. Clonal evolution appears to begin early in the neoplastic process during intraepithelial neoplasia. Aneuploidy has been found during intraepithelial neoplasia in many human epithelia, and, in association with other forms of genetic instability, may provide the increase in genetically variant cells required for clonal evolution to occur. It is postulated that two major factors affecting the rate of progression of intraepithelial neoplasia are the cellular mutation rate, which is enhanced by environmental carcinogens, and the cellular proliferation rate, which is enhanced by agents that include sex hormones, inducers of chronic inflammation, and irritant chemicals which stimulate reactive hyperproliferation. A preferred chemoprevention strategy should consist of the development of drugs and drug combinations which will block mutagenic carcinogens or prevent epithelial hyperproliferation or its causes. Two examples of the induction of regression of intraepithelial neoplasia by chemopreventive drugs are the regression of oral leukoplakia produced by beta-carotene and the regression of colorectal polyps in patients with familial polyposis produced by sulindac. It is evident that there is a strong need for more research on the induction of regression of intraepithelial neoplasia with chemopreventive agents. There is also a critical need to identify and develop biomarkers that correlate with the appearance and regression of intraepithelial neoplasia.

Introductory remarks: development of chemopreventive agents for prostate cancer

The term "cancer chemoprevention" refers to the prevention of cancer by intervening with drugs prior to the malignant (i.e., invasive) stage of carcinogenesis. The development of chemopreventive drugs is the major objective of the Chemoprevention Branch at the National Cancer Institute. The testing of drugs for cancer chemoprevention differs from testing of those for cancer treatment. Chemopreventive drug trials involve healthy target populations, and the endpoints of reduced cancer incidence or mortality, reduced/eliminated precancerous lesions, or increased latency must be achieved with little or no drug toxicity. The design of cancer chemoprevention trials for prostate presents several problems, such as the age of the study population and undependable methods for detecting microscopic foci by sequential sampling. A major motivation for organizing this workshop is the development of strategies for the design of chemopreventive intervention trials for prostate cancer. One of the most difficult problems of chemoprevention drug testing is the necessity of lengthy trials due to the long developmental period of many cancers. This is especially true for prostate cancer. A major solution to the problem is the use of intermediate biomarkers, defined as morphological or molecular intraepithelial changes that can constitute short-term endpoints in chemoprevention clinical trials. They are categorized as histological, genetic, proliferation-related, and differentiation-related. Modulation of intermediate biomarkers, instead of cancer incidence, as trial endpoints would allow chemoprevention trials to be of shorter duration, to use fewer subjects, and to be of lower cost. Review of the current status of prostatic intermediate biomarkers, and methods for identifying and validating them, are also major reasons for convening this workshop.

When is intervention warranted?

A chemoprevention trial in prostate cancer would be a formidable but potentially rewarding study. The current status of knowledge of drug interactions with, biomarkers of, and even the natural history of prostate cancer is insufficient to study all levels of men at risk. Currently, the most promising group to study is group I--those men with a high probability of developing prostate cancer but who do not currently have evidence of the disease. This could be a placebo-controlled, prospective and randomized study with the endpoint being clinically-detected prostate cancer. In addition, much may be gained from short-term pilot studies of "chemo-active" agents on morphologic and other biomarkers of prostate cancer initiated immediately before surgical removal. It is hoped that such studies may provide rationale for future efforts directed at preventing progression of premalignant or early prostate cancer lesions.

Development of chemopreventive agents for bladder cancer

The term cancer chemoprevention refers to the prevention or prolongation of carcinogenesis by intervention with drugs prior to the malignant (i.e., invasive) stage. The development of chemopreventive drugs is the major objective of the Chemoprevention Branch of the National Cancer Institute. Neoplastic lesions of the urinary bladder present a unique opportunity for evaluating chemopreventive agents because of (1) the accessibility of the lesions to observation and biopsy, and (2) those patients who have been successfully treated for a primary lesion represent a population at unusually high risk for recurrence and/or progression. Although 70-80% of bladder cancers initially present as superficial, papillary transitional cell neoplasms with limited potential for invasion, the incidence of recurrence is high after resection (60-75%). Recurrent tumors are highly unpredictable, and may be of higher grade or stage (progression). Although recurrence is responsible for high treatment-related morbidity, progression represents the greatest potential for mortality. Thus, potential chemopreventive agents considered here would modulate bladder carcinogenesis from initiation of normal-appearing tissue through progression of superficial tumors. Clinical trials of chemopreventive drugs involve healthy target populations, and the endpoints are reduced cancer incidence or mortality, reduced/eliminated precancerous lesions or increased latency, with none to minimal toxicity. Since cancers may not appear for 20-30 years, two of the most difficult aspects of testing these drugs in intervention trials are the long observation periods and large study populations required to measure cancer incidence reduction. However, observing the regression or recurrence of superficial bladder lesions (TIS, T1, Ta) requires relatively short time periods. Thus, these lesions lend themselves to the investigation of intermediate biomarkers, defined as morphologic and/or molecular alterations in tissue between initiation and tumor invasion. It is hypothesized that modulation of one or more biomarkers would interrupt carcinogenesis and result in a decrease in cancer incidence. Thus, evaluation of biomarkers as surrogate endpoints would allow bladder trials to be of even shorter duration, use fewer subjects and be lower in cost. In addition, intermediate biomarkers could predict which superficial lesions (or normal-appearing tissue) have the greatest potential for neoplastic progression. Development of strategies for the design of intervention trials for bladder cancer and review of the current status of intermediate biomarkers in the bladder, and methods for their validation, are major objectives of this workshop.

The natural history of intraepithelial neoplasia: relevance to the search for intermediate endpoint biomarkers

The development of carcinomas, defined as invasive epithelial neoplasms, is preceded by a preinvasive stage termed intraepithelial neoplasia that typically lasts for years. Intraepithelial neoplasia is the target tissue for the action of chemopreventive agents and the site where biomarkers frequently develop. The term "dysplasia" refers to the morphological alterations that characterize intraepithelial neoplasia and, according to many authors, consists of seven basic changes that are the same for the majority of epithelia. These are increased nuclear size, abnormal nuclear shape, increased nuclear stain uptake, nuclear pleomorphism (increased variation in size, shape, and stain uptake), increased mitoses, abnormal mitoses, and disordered or absent differentiation. Clonal evolution appears to begin early in the neoplastic process during intraepithelial neoplasia. The use of intraepithelial neoplasia as an intermediate endpoint biomarker requires that effective chemopreventive agents cause it to regress. Two examples are the regression of dysplastic oral leukoplakia produced by beta-carotene and the regression of colonic polyps in familial polyposis patients following treatment with the nonsteroidal antiinflammatory drug sulindac. There is a critical need to identify and develop biomarkers that correlate with the appearance and regression of intraepithelial neoplasia.

Intermediate biomarkers of precancer and their application in chemoprevention

The Chemoprevention Branch of the National Cancer Institute has established a program for the development of safe and effective cancer chemopreventive agents. This program includes identification of new agents, testing for efficacy in vitro and in animals, studies in animals to model clinical use, and preclinical toxicity and metabolism evaluation. Ultimately, the most promising agents progress to clinical trials. The long period required for cancer onset presents a significant challenge to the design of clinical trials for chemoprevention. Phase III trials in which cancer reduction is the endpoint require large subject groups (tens of thousands) and follow-up duration of more than five years. Because of these requirements, the cost of such trials are high. The Chemoprevention Branch is addressing this challenge by expansion of the preclinical and Phase II clinical efficacy efforts to include intermediate biomarkers of carcinogenesis as study endpoints. The Chemoprevention Branch's studies focus on the development of biomarkers with high reliability and predictive value for cancer. Both single markers and batteries of complementary and parallel markers are evaluated. Among the criteria for biomarkers for chemoprevention studies are the following: (1) differential expression in normal and high risk tissue, (2) appearance early in carcinogenesis (the earlier a reliable biomarker appears, the greater is the chance for successful intervention with a chemopreventive agent), (3) high sensitivity, specificity, and accuracy relative to cancer, (4) ease of measurement (use of non-invasive techniques and small tissue samples is preferable), (5) demonstration of modulation by chemopreventive agents, and (6) correlation of modulation with decreased cancer incidence.

Study of postnatal effects of chemopreventive agents on offspring of ethylnitrosourea-induced transplacental carcinogenesis in rats. I. Influence of retinol acetate, alpha-tocopherol acetate, thiamine chloride, sodium selenite, and alpha-difluoromethylornithine

We studied the influence of the vitamins retinol acetate, alpha-tocopherol acetate and thiamine chloride; the antioxidant sodium selenite and an inhibitor of polyamine biosynthesis, alpha-difluoromethylornithine, on the offspring of transplacental carcinogenesis by ethylnitrosourea in rats. Ethylnitrosourea was given to pregnant rats as a single i.v. injection, at a dose of 75 mg/kg body wt. or 5.5 mg/kg body wt., on the 21st day after conception. Retinol, tocopherol or thiamine was added to the diet, and selenite and alpha-difluoromethylornithine to drinking water of the offspring throughout their postnatal life at moderate doses. In control groups, ethylnitrosourea induced tumors of brain, spinal cord, peripheral nervous system and kidneys in the offspring. alpha-Difluoromethylornithine exerted a slight inhibitory effect; this agent decreased the total tumor multiplicity and the multiplicity of peripheral nervous system tumors and also prolonged survival time. Retinol, tocopherol, thiamine and selenite did not influence the development of the transplacentally-induced tumors.

Identification of candidate cancer chemopreventive agents and their evaluation in animal models and human clinical trials: a review

A search of the literature using National Library of Medicine databases and individual cancer journal articles yielded over 500 compounds with published chemopreventive activity in animals. From these, an initial 16 agents or agent combinations have been evaluated in the following animal tumor models: mouse skin papillomas/carcinomas induced by 7,12-dimethylbenz(a)anthracene/12-O-tetradecanoylphorbol-13-acetate; rat breast adenocarcinoma induced by N-methyl-N-nitrosourea or 7,12-dimethylbenz(a)anthracene; hamster lung carcinoma induced by N-methyl-N-nitrosourea or diethylnitrosamine; mouse bladder papillary carcinoma induced by N-butyl-N-(4-hydroxybutyl)nitrosamine; and rat and mouse colon cancer induced by azoxymethane/methylazoxymethanol acetate. Some of the most interesting positive results observed include 4-hydroxyphenyl retinamide plus tamoxifen in breast cancer, piroxicam in colon cancer, dimethylfluoroornithine in breast and bladder cancer, oltipraz in lung cancer, dehydroepiandrosterone in colon cancer, and molybdate in bladder cancer. Eighteen human intervention trials in progress are described that involve the following agents: beta-carotene (eight trials). Retinol/retinoic acid (seven trials), vitamins C and E (three trials), 4-hydroxyphenyl retinamide (one trial), piroxicam (one trial), and calcium (one trial). By organ site these studies involve cancer of the lung (six studies), skin (five studies), colon (four studies), breast (one study), and uterine cervix (two studies).

Delayed hypersensitivity reactions to virus-augmented syngeneic and allogeneic tumor associated antigens

Delayed hypersensitivity (DH) reactions to intact MSV-induced tumor cells or their crude membranes (CM) were studied across the H-2 histocompatibility barrier, using the radioisotopic footpad assay. Allogeneic as well as syngeneic tumor cells or their CM produced significant DH reactions in MSV-immune mice. CM from these cells after infection with influenza virus induced a stronger tumor-specific DH reaction in immune mice than did CM from uninfected cells. These results indicate that H-2 histocompatibility is not always required for induction of a cell-mediated immune response to tumor associated antigens and support the feasibility of a tumor specific skin test in humans using CM of autologous or allogeneic cultured human tumor cells.

Cytomorphologic evaluation of the neoplastic potential of 28 cell culture lines by a panel of diagnostic cytopathologists

A panel of 7 diagnostic cytopathologists, i.e., physicians trained to diagnose the malignant potential of human cells in Papanicolaou-stained smears, was asked to evaluate two sets of microscope slides of stained coverslip preparations of 28 cell culture lines, 15 of which were neoplastic. Slide Set I consisted of 13 pairs of cell lines, one member of each pair being nontumorigenic and the other tumorigenic; the lines were of mouse (9 pairs), rat (3 pairs), and human (1 pair) origin. Slide Set II contained 4 human lines: one lung cancer, one melanoma, and two fibroblast lines. Of a total of 114 diagnostic decisions by the panel, 88 were correct (66/86, 77%) in choosing which member of a pair was neoplastic and 22 were correct (22/28, 79%) in choosing whether a given individual human line was or was not neoplastic. Two members of the panel were correct more frequently, with 16/17 (94%) correct diagnoses, each. Five nuclear morphologic criteria of malignancy used by cytopathologists were prominent in the tumorigenic lines: altered chromatin pattern characterized by increasing size of chromatin granules and chromatin clumping, sharp angularity of large nucleolar and/or chromocenter borders with spicule formation (pointed projection), irregular parachromatin clearing (increase in the clarity of the clear spaces between chromatin threads, granules and clumps), uneven thickness of chromatin at the nuclear border, and variability in nuclear size and shape from cell to cell. These markers of neoplastic transformation, when added to those previously reported, should increase overall accuracy in the diagnosis of neoplastic transformation of mammalian cells in culture.

Breast cancer skin test antigens of increased sensitivity prepared from vesicular stomatitis virus-infected tumor cells

Crude membrane (CM) extracts were prepared from five cultured breast tumor lines (MDA-MB-157, MDA-MB-231, ZR75-1, HS0578T, and MCF-7) which had been infected with vesicular stomatitis virus (VSV) to augment their antigenicity. In skin test trials, CM extracts of uninfected MCF-7 cells elicited positive response in 0 of 13 (0%) tests in breast cancer patients, while VSV-MCF-7 elicited positive responses in 11 of the same 13 patients (84.6%). CM extracts of VSV-ZR75-1 and VSV-MCF-7 elicited greater delayed hypersensitivity responses (mm induration at 48 hr) in breast cancer patients than in patients with lung carcinoma or melanoma. Although the sensitivity of VSV-ZR75-1 was too low (ten of 28, or 35.7% of tests positive) to be useful as a skin test antigen, VSV-MCF-7 elicited positive responses in 30 of 38 (78.9%) tests in breast cancer patients, as compared to two of 15 (13.3%) and two of 13 (15.4%) of tests in patients with lung carcinoma and melanoma, respectively. The "virus-augmented" CM extract of cultured MCF-7 cells exhibited markedly greater sensitivity as compared to control MCF-7 extracts (P less than .005), with a high degree of specificity for breast cancer patients as compared to patients with the other neoplasms (P less than .00001). The results of skin testing with VSV-MCF-7 CM extracts demonstrated antigenic cross-reactivity with a large number of breast cancer patients, a finding of great importance for any potential immunotherapy and/or immunodiagnosis.

Anticancer drugs: effect on the cloning of Raji lymphoma cells in soft agar

The effect of 11 anticancer drugs on the ability of Raji lymphoma cells to form colonies in soft agar was determined with the use of both a 1-hour and a continuous drug exposure. Three distinct patterns of drug sensitivities were observed: a) Dactinomycin, adriamycin, bleomycin, mitomycin C, vincristine, and cis-platinum II all produced a dose-dependent reduction in colony formation following a 1-hour exposure, which was further augmented by a continuous exposure to the drugs; b) the antimetabolites (methotrexate, beta-cytosine arabinoside, and 5-fluorouracil) and pentamethylmelamine had no suppressive effects on colony formation with a 1-hour exposure, but they produced marked cytotoxicity with continuous drug exposure; and c) L-phenylalanine mustard had the same degree of colony suppression with both a 1-hour and a continuous drug exposure. Preincubation of Raji cells with an enzyme mixture (DNase + pronase + collagenase) did not alter the degree of colony suppression observed with the anticancer drugs. These results indicate that continuous drug exposure should be compared to a 1-hour drug incubation to determine in vitro drug sensitivities of fresh human tumors in the soft agar clonogenic assay, because the 1-hour drug exposure may not identify certain drugs that are potentially clinically active.

PRomotion of smooth surface tumorigenicity by phorbol myristate acetate in Balb/3T3 cells and Balb/3T3 T proadipocytes

In vitro exposure of Balb/3T3 cells to phorbol and phorbol myristate acetate significantly increased their tumorigenic potential when implanted sc on smooth surface plates into syngeneic mice. This finding supports the hypothesis that many so-called normal cell lines may actually represent initiated cells that can be induced to become tumorigenic following exposure to promoting agents. Since the tumorigenic potential of many tissues in vivo is inversely proportional to the state of differentiation of the stem cell populations undergoing transformation, we assayed the relative tumorigenicity of Balb/3T3 T proadipocytes, which can differentiate in culture, and Balb/3T3 cells, which cannot differentiate in culture. The effects of tumor-promoting agents on these cells were also tested. Plate-implanted Balb/3T3 T proadipocytes were markedly less tumorigenic than Balb/3T3 cells, and Balb/3T3 T proadipocytes were not sensitive to the promotional effects of phorbol myristate acetate.

The plate-implant as a bioassay for the neoplastic potential of cultured cells

One objective of this study was to test the plate implant as an in vivo bioassay for the neoplastic state with pairs of cell lines that were non-tumorigenic and tumorigenic when inoculated in suspension at various sites. Each pair originated from a common cell pool or from one cell derived from normal rat or mouse embryonic tissue. The implantation of cells attached to polycarbonate plates was compared with the subcutaneous injection of comparable numbers of cells in suspension or the intraocular injection of 10(5) cells. In addition, some implants of glass helices with attached cells were made into syngeneic mice and compared with injection of comparable cell numbers intramuscularly into X-irradiated hosts. A total of 10 lines or clones were tested on plates. Of four lines that were non-tumorigenic when inoculated in suspension either subcutaneously or intraocularly, two were tumorigenic when implanted subcutaneously attached to plastic plates. The remaining two are the first permanent rodent lines that have not been tumorigc on plastic plates. Of the remaining six lines, all were tumorigenic on plates and two produced sarcomas at a higher frequency on plates than in suspension. Cytologic diagnoses of the cells implanted on plates, subcutaneously and intraocularly, correlated best with the results from plate implants. Of three mouse lines implanted on helices, two were equally tumorigenic in suspension and on helices, whereas the third was tumorigenic on helices only. The plate implant was more satisfactory technically than the helices and proved to be one of the most sensitive bioassays for the neoplastic state of cultured cells.

Plate-induced tumors of BALB/3T3 cells exhibiting foci of differentiation into pericytes, chondrocytes, and fibroblasts

The BALB/3T3 clone A31 mouse embryo cell line has been used by many investigators as a model "normal" "fibroblast" line for a variety in vitro studies. It has been shown, however, that these cells are not "normal" because they will produce tumors within 2-4 months if 3 x 10(4) cells are implanted subcutaneously in BALB/c mice attached to 0.2 x 5 x 10-mm plastic plates. Previous studies also suggested that these cells were not fibroblasts because they gave rise tumors with the characteristics of vascular endothelium not fibroblasts. We now report that BALB/3T3 (clone A31), BALB/3T3-T, a proadipocyte subclone of clone A31 cells, and six recent subclones of BALB/3T3-T cells show additional differentiation patterns were tumors derived by implantation of these cells attached to plastic plates are examined. Differentiation into pericytes, chondrocytes, and fibroblasts was observed. We conclude that the BALB/3T3 clone A31 cell line and related lines are multipotent mesenchymal cells which are capable of differentiation into a variety of cell types.

Temperature-dependent alteration in immunogenicity of tumor-associated transplantation antigen monitored via paraformaldehyde fixation

Paraformaldehyde-fixed mKSA tumor cells of BALB/c mice were shown to retain tumor-associated transplantation antigen (TATA) activity to a degree comparable to that of X-ray-inactivated tumor cells under the optimal conditions of 1% paraformaldehyde for 30 min at 37 degrees. Unexpectedly, the TATA activity of cells fixed below 10 degrees was greatly reduced. This temperature effect was reversible. TATA activity was restored if the cells were returned to 37 degrees before fixation. Fixation at all temperatures for longer than 2 hr or at paraformaldehyde concentrations greater than 1% also caused a decrease in immunogenicity. Spleen cells from mice immunized with tumor cells fixed at 37 degrees were able to more effectively neutralize tumor growth in the Winn assay compared with those from mice immunized with cells fixed at 0 degrees. Immunization with paraformaldehyde-fixed tumor cells was completely specific. Mice immunized with an antigenically unrelated tumor were not rendered immune to tumor challenge. Fixed tumor cells could be stored for at least 1 month without loss of TATA activity.

"Sontaneous" neoplastic transformation in vitro: a form of foreign body (smooth surface) tumorigenesis

Explants of subcutaneous connective tissue from adult BALB/c mice into plastic petri dishes were serially subcultured and tested for tumorigenicity in two ways: by the subcutaneous implantation of cells attached to plastic plates (1 by 5 by 10 millimeters), and by the subcutaneous injection of cells suspended in saline. Cells grown in vitro for 18 or more days before being implanted attached to a plastic plate (2.4 x 10(4) to 3.4 x 10(5) cells per plate) formed tumors after 24 to 79 weeks. The latent period before tumor appearance correlated inversely with the time spent by the cells in tissue culture. Cells inoculated in saline suspension (10 to 100 times the above number per plate) did not form tumors until after 84 days in vitro; plates alone did not induce tumor formation within more than 1 1/2 years of implantation. The tumors arising from the plate-attached cells were transplantable without plates and histologically appeared to be undifferentiated sarcomas. It is well established that smooth-surfaced foreign bodies, regardless of their chemical composition, will produce sarcomas when transplanted subcutaneously in rodents. We interpret our data, particularly the decrease in tumor latent period with time spent in tissue culture, as indicating that a smooth surface was acting as a carcinogen first in vitro (the surface of the tissue culture dish) and then in vivo (the surface of the plastic plate).

Augmented immunogenicity of tumor cell membranes produced by infection with influenza virus as compared to Moloney sarcoma virus

The tumor-associated transplantation antigens (TATA) of crude membrane extracts from SV40-transformed BALB/3T3 tumor cells lytically infected with influenza virus were markedly more immunogenic than were extracts from uninfected cells measured either by the ability to induce heightened resistance to tumorgraft challenge or by heightened lymphocyte-mediated cytotoxicity against tumor cells in vitro. When intact tumor cells (as opposed to membrane extracts) were productively infected with Moloney sarcoma virus, they were made so immunogenic that they would only grow in X-irradiated syngeneic animals. Yet crude membrane extracts from the Moloney sarcoma virus-infected tumor cells showed no increase in TATA activity analogous to that seen after infection with influenza virus. Thus, influenza virus augmentation of tumor membrane TATA may operate by a different mechanism than does the oncornavirus augmentation of intact tumor cell TATA reported by others. It appears that Moloney sarcoma virus and possibly other oncornaviruses cannot be used to augment the TATA activity of tumor cell membranes in the same way that other surface-budding viruses can.

Neoplasms produced from C3H/10T 1/2 cells attached to plastic plates; saturation density, anchorage dependence and serum requirement of in vitro lines correlated with growth aggressiveness in vivo

The C3H/10T 1/2 embryo cell line, which is nontumorigenic when inoculated subcutaneously in saline suspension, produces tumors when implanted subcutaneously attached to 1 X 5 X 10 mm plastic plates. Under these in vivo conditions there is direct selection for "spontaneous" transformants that have undergone the specific cellular alterations required for neoplastic behavior. This is in contrast to the conventional situation where transformants are obtained in vitro and are only secondarily tested in vivo for neoplastic behavior. Early passages of cell lines from four different C3H/10T 1/2 tumors explanted back in culture were quantitatively examined for tumorigenicity and for alteration in the properties of density inhibition, anchorage dependence, serum requirement, and plasminogen activator production. A fairly consistent quantitative relationship was found between the degree of growth aggressiveness in vivo and the degree of expression of these phenotypic markers of the transformed state in vitro during early passages of the cell lines after tumor explantation.

Melanoma skin test antigens of improved sensitivity prepared from vesicular stomatitis virus-infected tumor cells

Crude membrane (CM) extracts from three different cultured human melanoma lines that were "virus-augmented" (infected with vesicular stomatitis virus (VSV) and subsequently inactivated by ultraviolet light) produced positive skin tests in 17 of 20 (85%), 11 of 20 (55%), and 13 of 18 (72%) tests, respectively, performed in 20 melanoma patients. Identical CM extracts from the same melanoma lines that had not been infected with VSV gave positive skin tests in 2 of 20 (10%), 4 of 20 (20%), and 2 of 18 (11%) tests, respectively, performed in the 20 melanoma patients, and no positive tests in the control patients. The 3 virus-augmented extracts were positive in only 2 of 18 (11%), 0 of 18 (0%), and 1 of 17 (6%) control subjects, respectively. The controls consisted of six normal volunteers and 12 patients with cancers other than melanoma. The "virus-augmented" CM extracts thus exhibited markedly greater sensitivity without significant loss of specificity as compared to nonvirus augmented extracts when used as tumor-specific melaonma skin test antigens.

Immunological study of two stocks of Moloney sarcoma virus producing regressor and progressor tumors in C57BL/6 mice

The immune response to progressor murine sarcoma virus (P-MuSV)2 and regressor murine sarcoma virus (R-MuSV) was analyzed by the radioisotopic footpad assay (FPA) for delayed hypersensitivity (DH) and by the membrane immunofluorescence assay for anti-tumor antibodies. In mice injected with P-MuSV, the footpad response remained at a low level for 7 weeks, while membrane immunofluorescence antibody titers reached their maximum levels after 5 weeks. By contrast, in mice injected with R-MuSV, the footpad response rose to a maximum by 4 weeks and membrane immunofluorescence antibody titers reached a maximum after 4 weeks. The adoptive footpad response reached its highest level in 2 weeks. A most striking difference between the P-MuSV and the R-MuSV systems was the lower immunogenicity of tumor cells induced by the P-MuSV as measured by the FPA. Considerable differences between the two stocks were also observed in oncogenic activity and tumor growth patterns. The P-MuSV contained one-tenth the amount of helper Moloney leukemia virus found in R-MuSV. Addition of more helper Moloney leukemia virus to the P-MuSV did not produce regressor tumors.

Independence of normal phenotypic properties and cell surface receptor mobility in variant cell lines

We report the use of three classes of variants from the long-established, malignantly transformed LM cell line to demonstrate that the apparent mobility of cell surface receptors need not be dependent on the expression of the transformed phenotype in vitro.

Common tumor rejection antigens in methylcholanthrene-induced squamous cell carcinomas of mice detected by tumor protection and a radioisotopic footpad assay

Seven transplantable lines of squamous cell carcinoma of the skin initiated in BALB/c mice by skin painting with methycholanthrene were systematically tested for cross-reactivity of their tumor rejection antigens in a 7 X 7 matrix. As determined by decreased tumor frequency after tumor cell challenge, each line was immunogenic against and/or immunosensitive to at least one and usually more than one of the other lines. A radioisotopic footpad assay for delayed hypersensitivity against viable tumor cells confirmed the cross-reactivity shown by tumor rejection. More than two antigens appeared to be present in the lines. Tests for C-type viruses were positive in all tumors; those for polyoma virus were negative. Whether the uniform presence of C-type viruses can account for the number and variety of antigens found, or whether the tumor rejection antigens are independent of virus expression, remains an open question. The finding of cross-reacting tumor rejection antigens in methylcholanthrene-induced squamous cell carcinomas encourages prospects for the development of more broadly applicable immunodiagnostic and immunotherapeutic reagents.