Znt. J. Cancer (Pred. Oncol.): 69,35-37 (1996) 0 1996 Wiley-Liss, Inc. *This article is a US Government work and, as such, is in the public domain in the United States of America. Puoltcalo? of t i e mle.nattonal Un'on Aga.is! Cancer .-2 A EARLY DETECTION RESEARCH PROGRAM AT THE NCI Sudhir SRIVASTAVA' and Susan C. ROSSI Early Detection and Community Oncology Program, Division of Cancer Prevention and Control, National Cancer Institute, Bethesda, MD 20892, USA. The Early Detection Branch, Division of Cancer Prevention and Control, National Cancer Institute, has created a program called The Early Detection Research Network (EDRN). EDRN's mission is to support translational research leading to early detection of cancer. The objectives are to ( i ) establish a network of institutions with the facilities, resources, personnel and interest to undertake biomarker research in early cancer detection; (ii) advance the understandingof the molecular basis of tumorigenesis in relation to screening, early detection and risk assessment; (iii) identify potential biomarkers that can be used as outcome measures or as intermediate end-points for cancer screening studies and (iv) respond to late-breaking developments in the field of biomarkers in a timely fashion. This program has the singular purpose of studying biologic, molecular and genetic markers relevant to the early detection of prostate, colorectal, lung, head and neck, bladder and breast cancers. o 1996 Wiley-Liss,Inc. * One of the critical requirements in studying molecular progression is the availability of a well-defined set of tissues representing normal, pre-neoplastic, in situ and metastatic stages of disease. Each cancer is likely to have its own natural history of progression. Colorectal cancer has provided a model system for studying the underlying molecular mechanisms of carcinogenesis and cancer progression in general. The National Cancer Institute's (NCI) Early Detection Research Network (EDRN) has established a network of investigators charged with developing and maintaining a biorepository in conjunction with a historical data base of epidemiological and clinical information. The EDRN institutions set up a biorepository of well-characterized prospectively and serially biopsied pre-malignant, malignant and normal tissues as well as blood samples using standardized protocols for processing, storage and shipping. Medical histories, tumor stage/grade, patient therapies and clinical outcomes are indexed with collected tissues. The EDRN thus provides a reliable and comprehensive source of tissues for doing well-designed molecular progression studies. It is hoped that with the results of such studies one could distinguish a collection of useful markers for susceptibility, early detection and prognosis. The EDRN is designed to enable investigators to focus on critical research issues rather than the time-consuming task of setting up a high-quality biorepository. Colorectal cancer has provided a useful model for cancer research, but much remains to be done with respect to identification of markers which will be useful for genetic screening of high-risk families and the general population. EARLY DETECTION OF COLORECTAL CANCER Colorectal cancer is the second most common cause of cancer death in the United States (Miller et al., 1993). Most colorectal cancer cases are detected at advanced stages when surgical treatment, which continues to have disappointing results, appears to be the only option. Early detection appears to be the only viable option that could play a significant role in the secondary prevention of colon cancer. Current methods of screening include examination of stool for occult blood, air-contrast barium enema examination and endoscopy includ- ing rigid and flexible sigmoidoscopy and colonoscopy (Smart et al., 1991). Several studies have concluded that there is an increased incidence of colorectal cancer in persons with polyps, which varies with the relative position of the polyps (i.e., proximal to the sigmoid colon; Prager et al., 1974; Lotfi et al., 1986; Simons et al., 1992), pointing to the importance of monitoring early events in colon tumor development. This tumor type has been shown to progress from normal tissue to adenoma and carcinoma through accumulation of genetic alterations (Vogelstein et al., 1988). To date, we know of mutations which affect oncogenes such as K-ras (Bos et al., 1987) and several tumor suppressor genes, including p53 (Hollstein et al., 1991),MCC (mutated in colorectal cancer; Kinzler et al., 1991), APC (adenomatous polyposis coli; Groden et al., 1991) and DCC (deleted in colon cancer; Fearon et al., 1990). Some of these genetic alterations are presently being tested for their potential to serve as clinical biomarkers of colon tumor development. For example, restriction fragment length polymorphism (RFLP) of chromosome 5q21-22 (for loss of theAPC gene) has been shown to be useful for pre-morbid diagnosis and counseling in familial adenomatous polyposis (FAP; Peterson et al., 1991). Although FAP is a rare disorder, occurring in only 1 in 500 persons, and accounts for less than 1%of colonic carcinomas, mutations of the APC gene occur in 60% of patients with colorectal carcinoma and are thought to be the earliest genetic abnormalities in the progression of colorectal carcinoma. The focus of molecular research has moved to a new class of genes that predisposes individuals to colon cancer. Hereditary non-polyposis colorectal cancer (HNPCC), also known as Lynch syndrome (Lynch et al., 1985, 1988), is found in as many as 1in 200 individuals in the Western world (Lynch et al., 1993) and has been the subject of intense study. Tumors manifesting this mutator phenotype have been termed replication errorpositive (RER+). A gene responsible for HNPCC has been identified on chromosome 2p16 (Leach et al., 1993), and because of its homology to the bacterial mutS and yeast MSH2 mismatch repair genes, it has been named hMSH2 (Leach et al., 1993).Avariety of other genes have now been implicated in HNPCC: hMLHl, hPMSl and hPMS2 (Papadopoulos et al., 1994;Bronner et al., 1994;Nicolaides et al., 1994).Mutations in any of these genes produce instability of micro-satellite sequences and have far-reaching implications that could affect large numbers of individuals, making them susceptible to various types of common epithelial malignancy. Thus, the potential for genetic screening and early intervention for those at high risk is enormous. With the discovery of 2 types of genetic marker, micro-satellite instability and the genetic susceptibility genes hMSH2, hMLHl, WMSl and hPMS2, it is now possible to define colon cancer risk for families and individuals. However, these markers must be validated in population-based studies prior to their application for mass screening. 'To whom correspondence should be addressed, at Early Detection Branch, Executive Plaza North, Room 305, Division of Cancer Prevention and Control, National Cancer Institute, National Institutes of Health, Bethesda, MD 20892, USA. Fax: (301) 402-0816. 36 SRIVASTAVA AND ROSS1 With the increasing number of hereditary genes identified and implicated in polyposis and non-polyposis types of colorectal cancer, it has now become possible to identify people pre-disposed to colon cancer through genetic screening. However, this will account for less than 15% of all colorectal carcinomas. Data accumulated thus far suggest that few of these mutant gene products by themselves or in combination may be fully predictive of cancer development. Thus, new potential biomarkers (Table I) need to be investigated. GENETIC SCREENING The presence of certain abnormal hereditary genes, as in familial adenomatous polyposis (Groden et al., 1991) and HNPCC (Leach et al., 1993), may indicate increased susceptibility to genotoxic (or epi-genotoxic) ambient exposures. Patients harboring these dominant genetic traits also carry an elevated risk for cancer. With these unique genetic characteristics, it would be possible to identify sub-populations and to classify individuals on the basis of potential susceptibility rather than simply on the basis of general risk factors. For example, direct sequence analysis of retinoblastoma (Rb), p53, DCC and mismatch repair genes could help predict a background risk of hereditary cancer and thereby identify a sub-population to target using available screening tests such as stool guaiac or sigmoidoscopy more effectively. Another area where biomarkers could have a significant impact is the selective screening of high-risk populations. For disease of very low prevalence, even screening tests with specificities as high as 95-98% would generate more falsepositives than true-positives. Positive predictive value is, in other words, highly dependent on the underlying prevalence, or population-based risk, of the disease. A potential way to circumvent this is to selectively screen high-risk populations, thus increasing pre-test prevalence. Some of the markers listed in Table I and some carcinogen adducts could possibly be used to improve upon this strategy. Other biological markers, particularly oncogene mutations in persons exposed to known carcinogens, may also significantly contribute to these efforts. The potential benefits of the markers will depend on their sensitivity and specificity. Before their large-scale applications, the markers must be validated; i.e., a biomarker must measure what it is intended to measure (Schatzkin et al., 1990). Validation studies in humans should include determinations of sensitivity, specificity and predictive value. Epidemiological applications of such markers will require carefully designed studies based on adequate sample size, control for salient confounders (such as age, sex and ethnic variability) and an understanding of the function of the marker. Eventually, with the early detection of cancer using biomarkers, it would be possible to determine whether the risk of rapid tumor progression was sufficient to justify the use of more invasive diagnostic tests and/or aggressive prophylactic or therapeutic interventions. Finally, the cost-effectiveness of the molecular approach to early detection of cancer would require a careful evaluation on a case-by-case basis. Since the molecular assays for a number TABLE 1 - POTENTIAL BIOMARKERS FOR COLON CANCER Marker SamDle ras mutations p53 or other suppressor gene mutations Ploidy Novel tyrosine kinases, telornerase hMSH2, hMLH1, PMS1, PMS2 Tissue, colonic effluent Tissue Tissue Tissue Tissue, colonic effluent of tumor markers are fairly specialized, the cost per assay is much less if tumor markers are assayed in large numbers in a centralized laboratory. FUTURE DIRECTIONS One of the objectives of the EDRN is to stimulate multiinstitutional, population-based collaborative studies and to encourage collaboration toward the common goals of (i) measuring the proportion of HNPCC cases among all colon cancer cases and (ii) piggybacking epidemiological data onto HNPCC cases to see if there are identifiable factors, including dietary habits, affecting the selective expression or nonexpression of HNPCC-related genes. It is hoped that the collaborative efforts of many institutions will provide the adequate sample size and develop a useful data base on family history and/or medical information on participants. A second objective relates to evolution of the basic research needed for translation to be successful in developing RER status for mass screening-ie., when and where to apply this technology. Developing automated sequencing technologies is of paramount importance for mass screening. We need to improve techniques to detect mutations within individual cells and to develop genotype/phenotype assays for mutational “hot spots” in genes. More efforts should be focused on understanding the natural history of tumor progression for developing intermediate end-points. Additional areas of high priority research for early detection of HNPCC include: (i) the development of a mismatch repair (MMR) gene mutational data base in HNPCC; (ii) the development of a mechanism to avoid reduplication of efforts; (iii) the development of consortium and international collaboration-q., BRCAl international collaboration,p53 mutational data base; (iv) the detection of technology priorities to include validation, automation and cost-effectiveness; (v) the development of family registries and epidemiological data bases; (vi) the effect of genes and environment on age at onset of tumors in those with HNPCC; (vii) the banking of patient materials for future studies and (viii) the frequency and penetrance of mutations in HNPCC and sporadic colon cancer. CONCLUSIONS In the era of technological advancement, one might see invasive procedures become almost unnecessary for early detection. We hope that the continuing developments in the area of molecular biology may provide sensitive and specific techniques for early diagnosis of cancer. For example, chromosomal aberrations in tumors may draw attention to potentially significant genetic sites in precursor lesions. Observations of genetic alterations, chromosomal breaks or altered DNA methylation could identify a precursor or pre-malignant condition. Loss of DNA methylation is found in very early stages of colon tumor progression (Goelz et al., 1985) and has been shown to lead to mitotic disjunction by inhibiting chromosomal condensation (Schmid et al., 1990). The near-term benefit of studying biornarkers may lie in the improvement of survival by detecting cancer cases at an early stage rather than as potential targets for therapeutic intervention using tools that do not yet exist. Curative modalities already exist for most tumors if discovered prior to the development of invasive or metastatic potential. Oncogenes, for example, may be used as biomarkers for early diagnosis of cancer or may provide important prognostic information. Many have already been shown to be associated with poor survival. For example, over-expression of c-myc in head and neck squamous cell carcinoma has been associated with EARLY DETECTION RESEARCH PROGRAM shorter survival compared with those with low levels of c-myc protein (Field et a/.. 1989). A correlation has been found between an altered c-myc gene and a poor short-term prognosis in 121 patients with breast cancer (Varley et al., 1987). c-myc gene over-expression has also been shown to be correlated with breast cancers of poor prognosis (GuCrin et al., 1988). In the longer term, interventions to stop or reverse tumor progression may also improve survival and prognosis, but these await new developments. A note of caution is important: The use of genetic and biological markers in a cancer prevention study implies 2 assumptions: (i) the marker-indicated cancer is likely to occur and ( i i ) reduction of the marker is synonymous with control of the disease. However, these assumptions may not be true for many of the markers and events discussed above; thus, each biomarker must be validated. Biomarkers may or may not be disease-specific and may or may not lie directly along the causal pathway. If not on the causal pathway, an intervention may successfully block its expression but fail to block the progression to malignancy; there may also be more than one causal pathway. Intermediate markers may occur early or late on the causal pathway so that a given successful intervention may not change an intermediate marker, while another would, 37 depending on where in the cascade of genetic events each marker occurred. Finally, we do not know the spontaneous regression rate of molecular events, for it is conceivable that some events occur and disappear. None of these caveats is likely to be insurmountable, but they must be taken into account as future studies are used to validate intermediate end-points. Therefore, prospective studies on individual cancer sites correlating molecular and biochemical changes with clinical disease characteristics and patient outcome and demographics should be carried out. These points highlight the importance of establishing a bank of normal cells and tissues and of pre-malignant and malignant lesions to identify potential cellular and molecular markers for early detection. Studies on the patterns of biochemical, cellular, genetic and molecular events on such prospectively collected cells and tissues may identify surrogate end-points for prevention and a “candidate marker” for the detection of early lesions. Matching these markers with an associated prospective data base of demographic information, exposure to potential carcinogens and risk factors on the subjects from whom specimens have been obtained may aid in developing a new generation of screening and early detection modalities. REFERENCES Bos. J.L., FEARON, E.R., HAMILTON, S.R., VERLAAN-DE VRIES.M., colorectal cancer: an updated review. Gastroenterology, 104,1535-1549 VAN BOOM,J.H., VANDER EB,A.J. and VOGELSTEIN, B., Prevalence of (1993). ras gene mutations in human colorectal cancers. Nature (Lond.), 327, LYNCH,H.T., WATSON,P., LANSPA,S.J., MARCUS,J., SMYRK,T., 293-297 (1987). FITZGIBBONS, R.J., JR., KRIEGLER, M. and LYNCH, J.F., Natural history BRONNER, C.E., BAKER,S.M., MORRISON, P.T., WARREN,G., SMITH, of colorectal cancer in hereditary nonpolyposis colorectal cancer C., LIPFORD,J. and (Lynch syndromes I and 11). Dis. Colon Rectum, 31,372-377 (1988). L.G., LESCOE,M.K., KANE, M., EARABINO, LINDBLOM, A,, Mutation in the DNA mismatch repair gene homo- MILLER,B.A., RIES, L.A.G., HANKEY, B.F., KOSARY,C.L., FEUER, logue hMLHl is associated with hereditary non-polyposis colon E.J., BORING, C.C. and EDWARDS, B.K.. Section I: Overview. In: SEER cancer. Nature (Lond.),368,258-261 (1994). cancer statistics review: 1973-1990, National Cancer Institute, NIH FEARON, E.R., CHO,K.R., NIGRO,J.M., KERN,S.E., SIMONS. J.W., Publ. 93-2789, Bethesda, MD (1993). S.R., PREISINGER, A.C., THOMAS,G., NICOLAIDES, RUPPERT,J.M., HAMILTON, N.C. and 17 OTHERS, Mutations of two PMS homologues KINZLER, K.W. and VOGELSTEIN, B., Identification of a chromosome in hereditary nonpolyposis colon cancer. Nature (Lond.), 371, 75-80 18q gene that is altered in colorectal cancers. Science, 247, 49-56 (1994). (1990). PAPADOPOULOS, N., NICOLAIDES, N.C., WEI, Y.F., RUBEN,S.M.. FIELD, J.K., SPANDIDOS. D.A., STELL.P.M., VAUGHAN, E.D., EVAN, CARTER,K.C., ROSEN,C.A., HASELTINE, W.A., FLEISCHMANN, R.D., G.I. and MOORE,J.P., Elevated expression of the c-myc oncoprotein FRASER,C.M. and ADAMS,M.D., Mutation of a mutL homolog in correlates with poor prognosis in head and neck squamous cell hereditary colon cancer. Science, 263,1625-1629 (1994). carcinoma. Oncogene. 4,1463-1468 (1989). PETERSEN, G.M., SLACK.J. and NAKAMURA, Y., Screening guidelines GOELZ,S.E., VOGELSTEIN, B., HAMILTON, S.R. and FEINBERG, A.P., and premorbid diagnosis of familial adenomatous polyposis using Hypomethylation of DNA from benign and malignant human colon linkage. Gastroenterology.100,1658-1664 (1991). neoplasm. Science, 228,187-190 (1985). PRAGER,E.O., SWINTON, N.W., YOUNG,J.L., VEIDENHEIMER, M.C. GRODEN, J. and 22 OTHERS, Identification and characterization of the and CORMAN, M.L., Follow-up study of patients with benign mucosal familial adenomatous polyposis coli gene. Cell, 66,589-600 (1991). polyps discovered by proctosigmoidoscopy. Dis. Colon Rectum, 17, GUERIN,M., BARROIS. M., TERRIER, M.J., SPIELMA”,M. and RIOU, 322-324 (1974). G., Overexpression of either c-myc or c-erb B-2inerc protooncogenes in SCHATZKIN, A,, FREEDMAN, L.. SCHIFFMAN, M. and DAWSEY,S., human breast carcinoma: correlation with poor prognosis. Oncogene Validation of intermediate end points in cancer research. J. nut. Res., 3,21-23 (1988). Cancerbist., 82, 1746-1752 (1990). HOLLSTEIN, M., SIDRANSKY, D., VOGELSTEIN, B. and HARRIS, C.C., p53 SCHMID, M., HAFF,T. and GRUNER, D., 5-azacytidine-induced undermutations in human cancers. Science, 253,49-53 (1991). condensation in human chromosomes. Hum. Genet., 67, 257-263 KINZLER, K.W., NILBERT,M.C.. VOGELSTEIN, B., BRYAN, T.M., LEVY, (1990). D.B., SMITH,K.J., PREISINGER, A.C., HAMILTON, S.R., HEDGE,P. and SIMONS,B.D., MORRISON, A.S., LEV,R. and VERHOEK-OFTEDAHL, W., MARKHAM, A., Identification of a gene located at chromosome 5q21 Relationship of polyps to cancer of the large intestine. J. nut. Cancer that is mutated in colorectal cancers. Science, 251,1366-1370 (1991). Qsi., 84,962-966 (1992). LEACH,F.S., NICOLAIDES, N.C., PAPADOPOULOS, N., LIU. B., JEN.J., SMART, C.R., CHU,K.C., CONLEY, V.L., HENSON, D.E.. POMMERENKE, PARSONS,R.,PELTOMAKI, P., SISTONEN, P., AALTONEN, L.A. and F.P. and SRIVASTAVA, S., Cancer screening and early detection. In: J.F. NYSTROM-LAHTI, M., Mutations of a mutS homolog in hereditary Holland, E. Frei 111, R.C. Bast, Jr., D.W., Kufe, D.L. Morton and R.R. nonpolyposis colorectal cancer. Cdl, 75,1215-1235 (1993). Weichselbaum (eds.), Cancer medicine, pp. 408-431, Lea and Febiger. LOTFI,A.M., SPENCER, R.J., TLSTRUP. D.M. and MELTON,L.J., 111, Philadelphia (1991). Colorectal polyps and the risk of subsequent carcinoma. M q o Clin. VARLEY, J.M., SWALLOW, J.E., BREAMMAR, W.J., WHI-ITAKER. J.L. and Proc., 61,337-343 (1986). WALKEK, R.A., Alterations to either c-erb-B-2 (neu) or c-myc protoonLYNCH. H.T., SCHUELKE, G.S., WELLS,I.C., CHENG,S.C., KIMBERLING, cogenes in breast carcinoma correlate with poor short term prognosis. W.J., BISCONE, K.A., LYNCH, J.F. and DANES, B.S., Hereditary ovarian Oncogene, 1,423-430 (1987). carcinoma. Biomarker studies. Cancer, 55,410415 (1985). VOGELSTEIN, B., FEARSON, E.R., HAMILTON, S.R., KERN,S.E., PREISLYNCH,H.T.. SMYRK, T.C., WATSON,P., LANSPA,S.J., LYNCH,J.F., TNGER, A.C., LEPPERT, M., NAKAMURA, Y., WHITE,R., SMITS,A.M. LYNCH,P.M., CAVALIERI, R.J. and BOLAND, C.R., Genetics, natural and Bos, J.L., Genetic alterations during colorectal-tumor develophistory, tumor spectrum, and pathology of hereditary nonpolyposis ment. New Engl. J. Med., 319,525-532 (1988).