Interdisciplinary Genetic Epidemiology (Past Initiative)

Five interdisciplinary studies in the genetic epidemiology of cancer were funded in 1999 in response to a Request for Applications (RFA). These projects, together with three related research projects also funded through the Clinical and Genetic Epidemiology Research Branch (CGERB), have formed a consortium to cooperate on common research themes and infrastructure needs. Summaries of these projects are provided below:

Background information

Genetic Epidemiology of Lung Cancer I: Gene Identification in High-Risk Families
Marshall W. Anderson, Ph.D.
University of Cincinnati
Department of Environmental Health Cincinnati, Ohio

The investigators hypothesize that there are specific genotypes that greatly increase the risk of developing lung cancer, through interaction with cigarette smoking and/or other environmental exposures. Accumulated evidence suggests that there are genetic susceptibility components in lung cancer, and that gene-environment interactions are important.

This research project is one of two parallel and complementary projects proposed by the Genetic Epidemiology of Lung Cancer Consortium (GELCC), a multi-investigator, interdisciplinary team whose goals are to identify lung cancer susceptibility gene(s), and to estimate gene-environment interaction in the etiology of the cancer.

The research team led by experts in genetic epidemiology, gene mapping, lung biology, and cancer molecular genetics, is focusing on the creation of a national resource of high-risk lung cancer families for use in linkage studies and identification of lung cancer susceptibility gene(s). The specific aims of this project are to:

1. Create a registry of high-risk families for genetic linkage studies of lung cancer. Using Consortium facilities, 11 U.S. sites will identify and enroll families who have at least 3 first-degree relatives with lung cancer in a research protocol. The researchers will screen an estimated 70,000 new lung cancer cases to identify 800 high-risk families, and develop the most informative 100 for linkage analysis. After obtaining informed consent from probands and family members, biosamples (blood, archival material, and buccal brushings), clinicopathologic, demographic, risk factor, and family history data will be collected, processed, and stored.
2. Identify a lung cancer susceptibility locus (loci) through linkage analysis of familial lung cancer pedigrees. Genotyping by Consortium core facilities and by the NIH-sponsored Center for Inherited Disease Research (CIDR) with genome wide and candidate region PCR markers will be done on high-risk lung cancer families identified from aim 1. Linkage analysis will be performed using the family history, risk factor, and genotype data.
3. If a susceptibility locus (loci) is localized and confirmed in aim 2, the researchers then will begin to identify and characterize the susceptibility gene(s). Positional cloning methods will be employed to identify and characterize novel genes. The longer term objective is to characterize the susceptibility gene(s) and develop ways to identify individuals who are predisposed to develop lung cancer. The combined scientific outcomes of the Consortium studies will provide a means to identify individuals and families at high risk for lung cancer, and elucidate a strategy for the prevention, control, and clinical management of the disease.

The consortium members are:

• University of Cincinnati, Ohio
• University of Colorado Cancer Center, Denver, Colo.
• Karmanos Cancer Institute, Wayne State University, Detroit, Mich.
• Louisiana State University Health Sciences Center, New Orleans, La.
• Mayo Clinic, Rochester, Minn.
  
• Medical College of Ohio, Toledo, Ohio
  
• Saccomanno Research Institute, St. Mary's Hospital, Grand Junction, Colo.

A Model for Genetic Susceptibility: Melanoma
Marianne Berwick, Ph.D., M.P.H.
Memorial Sloan-Kettering Cancer Center
Department of Epidemiology and Biostatistics
New York, N.Y.

Melanoma provides a unique model for studies of gene-gene and gene-environment interaction in the development of cancer. The cancer has several key features:

1. There is a major environmental cause of melanoma, exposure to solar ultraviolet (UV) radiation, which may account for as much as 90% of cases in populations of European origin.
2. Variants of, or mutations in, at least three classes of genes with very different functions may determine variation in melanoma risk-cell cycle genes, nucleotide excision repair genes, and genes relating to cutaneous pigmentation.
3. The latter two classes of genes are involved in protection against the effects of UV radiation.

These features permit the study of gene-gene and gene-environment interactions under circumstances largely free from confounding environmental factors, and with genes whose functions and relationships to the environmental factor are reasonably well understood.

The researchers hypothesize that there are wide variations among individuals in terms of susceptibility to melanoma. This variation is probably influenced in a heterogeneous manner by multiple susceptibility genes and sun exposure, which is the major known exogenous factor and may exert its influence interacting with these genes. The study will permit evaluation of the public health impact of genetic mutations and polymorphisms and their interaction with sun exposure, via estimation of relevant population parameters in a novel study design.

In a large, international population-based case-control study covering a wide range of latitudes, the researchers will:

1. determine the relative risk for developing melanoma due to germline mutations and polymorphisms in the cell cycle genes CDKN2A and CDK4;
2. determine the relative risk for developing melanoma due to polymorphisms in the melanocortin receptor gene (MC1R), a major pigmentary gene;
3. determine the relative risk for developing melanoma due to allelic variation in the DNA repair genes known as the nucleotide excision repair genes (NER), which specialize in removing DNA damage due to UV radiation; and
4. analyze the interactions among genetic variants that are associated with the development of melanoma and their associations with solar UV radiation.

The researchers will use a novel epidemiological design, a population-based case-control study in which the controls are subjects with incident primary melanoma and the cases are subjects diagnosed with a second or higher order primary melanoma. This design offers substantially greater statistical power to test the hypotheses than a classical case-control study with general population non-diseased controls (Begg and Berwick, 1997) and, in principle, higher participation rates by both cases and controls. The design has substantial potential for application to other areas of cancer susceptibility.

Breast Cancer, Radiation Exposure, and the ATM Gene
Jonine L. Bernstein, Ph.D., M.S.
Mount Sinai School of Medicine New York, N.Y.
and
Risk of Cancer in A-T Families
Robert W. Haile, Dr.P.H
University of Southern California
Norris Comprehensive Cancer Center
Los Angeles, Calif.

The principal investigators of these two multisite, interdisciplinary studies are conducting studies of ATM (mutated in ataxia telangiectasia (A-T)) that complement each other and provide a strong empirical basis for estimating cancer risk in ATM heterozygotes, and assessing the role of an ATM -radiation interaction in the etiology of breast cancer. Several collaborators are involved in both studies, which will facilitate communication and coordination of activities.

Together, the studies represent a comprehensive, concerted approach to the question of cancer risk and A-T, and should provide informative data on:

• the risk of cancer in ATM heterozygotes,
• the frequency and spectrum of mutations (including truncating and missense mutations) in breast cancer, and
• whether radiation exposure modifies the risk of cancer in ATM heterozygotes.

More specifically, Dr. Bernstein and colleagues are investigating gene-environment interactions in the etiology of breast cancer. They are establishing a repository of epidemiologic risk factor information and biologic specimens from 900 women with asynchronous bilateral breast cancer and 2,700 women with unilateral breast cancer, who will be ascertained through 6 population-based tumor registries in the United States and Denmark. All subjects will be interviewed using a structured questionnaire, and blood samples will be collected for genetic analyses.

The initial plan for using this repository is to examine the interaction of radiation exposure, the ATM gene, and breast cancer. Ionizing radiation is known to be a breast carcinogen, and recent studies suggest that the ATM gene may increase susceptibility to radiation-induced breast cancer. The hypothesis is that women who are ATM gene carriers, and who have received radiation therapy as part of breast conservation treatment, are at especially high risk of developing second primary contralateral breast cancer.

The researchers will also provide descriptive statistics on the prevalence of the ATM gene in this large population-based sample of women. ATM heterozygosity will be assessed through an efficient staged approach appropriate for analysis of this complex gene. For subjects who received radiation therapy, radiation scatter dose to the contralateral breast will be reconstructed.

This unique repository will be critical for future interdisciplinary investigations into the mechanisms and nature of gene-gene and gene-environment interactions influencing susceptibility to breast cancer. The study of second primaries presents a particularly promising context in which to disentangle the complex interactions among hormonal, genetic, and environmental factors influencing breast carcinogenesis because any important etiological factors (e.g., genetic abnormalities) will be more prevalent among women who already have breast cancer than in the general population. Further, the rising incidence of breast cancer, coupled with improved survival, has placed an increased number of women at risk for second primary breast cancers, making this an issue of public health importance.

Dr. Robert Haile and colleagues are determining whether individuals who are heterozygotes for a mutation in the gene recently identified as a cause of ataxia-telangiectasia (A-T) have a higher than expected risk of cancer, with a primary focus on breast cancer. Specific aims are to:

1. Collaborate with investigators in Canada, Costa Rica, Germany, Israel, Italy, Poland, Turkey, and the United States who have already identified families with the ATM gene to: a) obtain a reported history of cancer in parents, grandparents, and aunts and uncles of A-T cases; b) verify these reported cancers; and c) to complete genetic testing for mutations in ATM in these families, if not already available. Participating centers have collectively identified 588 A-T families to date.
2. Conduct statistical analysis to determine whether ATM heterozygotes in these families have an excess risk of cancer (greater than that expected for the appropriate source population), with a primary interest in breast cancer. The researchers will also initiate exploratory analyses addressing the issue of heterogeneity of risk between subgroups defined by country of origin and type of mutation. (They anticipate that an analysis combining all known data sets will be necessary to address questions of heterogeneity in an informative manner and propose the beginnings of such an effort in this study.)

Molecular Epidemiology of Colorectal Cancer
Stephen Bernard Gruber, M.D., Ph.D., M.P.H.
University of Michigan Health System
Department of Internal Medicine
Ann Arbor, Mich.

This population-based case-control study is examining the contribution of genetic sequence variation and environmental factors to the development of colorectal cancer. Colorectal cancer is the second leading cause of cancer death in the United States, and is the most common cancer in Israel. The majority of cases of colorectal cancer are diagnosed in individuals who do not have identifiable risk factors other than age. Major susceptibility genes are likely to account for less than 25% of all colorectal cancers, but new evidence supports the existence of low-penetrance susceptibility alleles which may play an important role in the population dynamics of this complex disease. A novel cancer susceptibility allele, APC I1307K , has been identified in 6% of individuals of Ashkenazi Jewish descent and appears to double the risk of colorectal cancer.

The broad objective of the study is to examine how environmental factors may modify genetic risk. This collaborative study between the University of Michigan and the National Center for Cancer Control of Kupat Holim Clalit (Israel's largest health care provider) will identify 2,100 cases of incident colorectal cancer and 2,100 age-, sex-, and geographically matched controls in Northern Israel. Interviews and food frequency questionnaires will assess epidemiological risk factors, and histopathologic and molecular studies will provide data for understanding the genetic basis of colorectal cancer. The advantage of studying colorectal cancer in Israel is that this population is one of the few in the world where a known cancer susceptibility allele occurs at a high frequency, thus facilitating study of how inherited risk for the disease may be modified. The specific aims are to:

1. Measure the risks of developing colorectal cancer associated with the APC I1307K allele.
2. Identify and measure potential effect modification of genetic and environmental risks in the pathogenesis of the cancer.
3. Analyze colorectal cancer tumor specimens in order to characterize the somatic mutational fingerprint in a defined region of the APC tumor suppressor gene, and understand what influences the transformation of normal colonic epithelium to colorectal cancer.
4. Establish a resource for further epidemiologic studies and genome screening to map novel, low-penetrance colorectal cancer genes.

Susceptibility Genes for Human Colon Neoplasia
Sanford Markowitz, M.D., Ph.D.
Case Western Reserve University and University Hospitals
Ireland Cancer Center School of Medicine
Cleveland, Ohio

The ultimate goal of this project is the identification of a gene whose polymorphisms within the American population determine the risk of developing colon polyps and cancers among average adults. This will be accomplished by identifying and collecting DNA from a cohort of 300 pairs of siblings in which the two members of each sibship have each developed colon neoplasia by age 65 or younger. Using the powerful genetic algorithm of affected sibling pair linkage analysis, the researchers will use this cohort of concordantly affected sibling pairs to examine the entire human genome to identify loci linked to colon neoplasia susceptibility in adults.

Colon cancer is the second most common cause of cancer death among adult Americans. These cancers develop from premalignant precursor lesions known as adenomatous colon polyps. Multiple epidemiological studies have demonstrated that once one member of a family has developed an adenomatous colon polyp, his or her siblings are at markedly elevated risk for developing both colon adenomas and colon cancers. These studies are consistent with 20% of all Americans carrying a dominant "colon neoplasia susceptibility" gene with 40% penetrance by age 60.

An apparently rare, but illustrative, example of one such colon neoplasia susceptibility variant is the finding that 6% of Ashkenazi Jews carry the I1307K APC variant whose DNA sequence is predisposed to in adult life, undergo second hit somatic mutations that inactivate APC tumor suppressor activity. This I1307K APC variant is not present outside the Ashkenazi population. It is thus a pressing question to answer whether other APC gene variants exist and are the missing colon neoplasia susceptibility genes in the broader population.

Moreover, variant murine alleles of each COX2 , sPLA2 , and DNMT genes can suppress or promote neoplasias in mice that are genetically predisposed to intestinal tumor development. It is thus a pressing question to answer whether human variants of these genes are present as major colon neoplasia susceptibility genes in humans.

The specific aims of this proposal are accordingly to:

1. identify and obtain DNA from 300 affected sibling pairs in which both sibs are concordant for a history of colon neoplasia;
2. exclude from this cohort any sibships with known familial colon cancer syndromes (FAP , HNPCC , or APC-I1307K carriage);
3. test four candidate susceptibility genes (APC , COX2 , sPLA2 , DNMT ) for linkage to colon neoplasia development among these sibships via an affected sibling pair linkage analysis; and
4. identify novel colon neoplasia susceptibility loci by similarly testing for linkage each of 350 microsatellite markers that span regions encompassing the entire human genome.

Gene-environment Interactions: The Odyssey Cohort
Kathy J. Helzlsouer, M.D., M.H.S.
Johns Hopkins School of Medicine
Baltimore, Md.

Polymorphisms of many genes controlling metabolism of xenobiotic compounds and DNA repair processes are associated with susceptibility to cancer as well as other chronic diseases. Studies are needed that are more holistic in design and provide insights into the net effect on health of having specific genotypes.

The researchers will conduct a population-based cohort study to examine the broad impact of genetic variation in candidate genes and their interaction with environmental exposures on cancer incidence and survival specifically, and health and aging more generally. Participants (N=8395) of two blood and data specimen banks CLUE I (1974) and CLUE II (1989) comprise the study cohort (Odyssey Cohort). The cohort has been followed prospectively for 24 years, and information on environmental factors such as smoking, education, and housing are available as far back as 1963. DNA will be extracted from buffy coat specimens obtained from participants in 1989 and stored at -70 C.

The researchers will investigate polymorphisms in genes coding for enzymes that:

1. metabolize nutrients and hormones (MTHFR, VDR, CYP17, CYP1B1, COMT, CYP3A4);
2. metabolize carcinogens (ADH, GSTM1, NAT1, ans NAT2, NOS, CYP1A1, CYP1B1, EH); and
3. control DNA repair processes (XRCC1, XRCC3, XPD).

In this cohort, power is greater than 90% to detect a relative risk of 2 for main effects of genotypes on mortality and cancer incidence, and at least 80% to detect gene-environment and gene-gene interactions of 2-fold for the major cancer sites (such as breast, prostate, and colorectal) and 3- to 5-fold for less common cancers (such as endometrial and bladder cancer).

As additional polymorphisms in candidate genes with potential relevance to the major health outcomes are identified, the researchers will be able to investigate their impact on health and survival.

Gene and Gene-environment Interactions in Ovarian Cancer
Alice S. Whittemore, Ph.D., M.A.
Stanford University
Stanford, Calif.

Mutations in the genes BRCA1 and BRCA2 are thought to account for at least 70% of families with multiple cancers of the ovary and breast. Little is known about other predisposing genes in the remaining breast/ovarian cancer families, or in families with multiple ovarian cancers with little or no breast cancer. It is now feasible to identify other predisposing genes for ovarian cancer. Critical to this effort is obtaining DNA from many large ovarian cancer families. Because such families are rare, the Gilda Radner Familial Ovarian Cancer Registry is an important resource for such study. This registry currently comprises records of more than 1,000 U.S. families reporting two or more ovarian cancer diagnoses.

This research project is identifying 300 of the families, each containing 2 or more verified ovarian cancers in first- or second-degree relatives, for linkage analysis and study of gene-environment interactions. Affected and unaffected members of each identified family will be asked to provide epidemiological and pedigree data and biological specimens for genetic analysis. The researchers will classify the families as:

1. segregating mutations of BRCA1,
2. segregating mutations of other genes (BRCA2, MSH2, MLH1), or
3. containing no known mutations of these genes. To date, the researchers have identified 156 participating families and genotyped 96 of them for BRCA1 and for part of BRCA2. Thirty of 96 genotyped families (31%) segregate BRCA1 mutations, and only one family has been found to segregate BRCA2 mutation.

The goals are to:

1. pool families of type (iii) with similar families from the United Kingdom to map new genes associated with familial ovarian cancer; and
2. among women with BRCA1 mutations, evaluate associations relating ovarian cancer to epidemiological attributes such as parity, oral contraceptive use, and hormone replacement therapy.

Use of this large registry to map predisposing genes for ovarian cancer and characterize families segregating different types of mutations offer these benefits: It will allow interested women at genetically high risk to be targeted for screening and intervention. It will allow classification of ovarian cancers into genetic subtypes, thereby facilitating the study of etiologic factors and the planning of strategies for screening and prevention. The registry will also provide a large database for examining lifestyle factors that modify ovarian cancer risk in carriers of mutations.