A major obstacle to preventive strategies for ovarian cancer is the lack of recognized modifiable risk factors. Age, a positive family history of ovarian cancer, having no children, lack of oral contraceptive use, and having intact uterine tubes are all known risk factors for ovarian cancer, but these are mostly not amenable to modification.

One of the difficulties in studying risk factors for ovarian cancer is that it is one of the most complex human cancers, with several different subtypes of disease recognized. There are at least five types of the most common form of ovarian cancer (epithelial cancer), and each has quite distinct biologic features. There is evidence that particular risk factors may influence the development of specific subtypes of ovarian cancer. Studies that do not consider the different subtypes of disease may compromise their ability to detect associations between important risk factors and ovarian cancer.

Ovarian tumors are highly variable at a genetic level. They display a wide range of DNA damage and a variety of genes are abnormally turned on or off. Recent exciting developments in technology plus great advances in the Human Genome Project promise the potential for novel classifications of cancers that are based upon fingerprints of gene activity. Identifying these molecular subtypes may be important to clarify the effects of risk factors. For example, a specific hormone-related risk factor may lead to particular genes being turned on or off in a recognizable pattern. Another form of variability when considering the effects of risk factors is that not all women are equally susceptible to environmental exposures. Common variation in genes involved in hormone pathways and genetic damage repair may modify the effects of risk factors, and thus, it is important to measure these genetic variants when studying risk factors for ovarian cancer.

We propose to establish a large multicenter study of over 1,000 women diagnosed with ovarian cancer and compare them to a similar number of women of the same age without ovarian cancer. All women will be interviewed using a comprehensive questionnaire regarding risk factors for ovarian cancer and they will also be asked to provide blood and urine samples. Women with ovarian cancer will also be asked for specimens of tumor tissue. With epidemiologists and molecular biologists working together, we will use this valuable resource to address new perspectives on potentially modifiable risk factors for ovarian cancer by considering important differences in disease subtypes and differences in susceptibility. In addition, we will ask new questions regarding novel molecular classifications of ovarian tumors and associate these with risk factors and clinical features. Only a large, specially designed study that integrates epidemiology and molecular biology can answer these more complex questions and we hope will provide valuable data on risk factors that can be modified, which will lead to relevant preventive strategies and a better understanding of the way ovarian tumors develop.

Project 1: Program Project Award

Molecular Epidemiology of Ovarian Cancer

Principal Investigator: David Bowtell

Project 1: Public Abstract

Ovarian cancer (OVCA) is one of the most complex human cancers, with many different types of disease recognized. Even among the more common forms of OVCA, serous and mucinous, there is a wide spectrum of clinical responses to treatment. Underlying these different forms of the disease and variable responses to treatment must be genetic differences that dictate the properties of the cancer cells. We have only identified a fraction of the genetic changes in OVCA, and we do not know how they interact with each other to drive OVCA cell growth.

Our total repertoire of genes, the genome, is the blueprint for building and operating a human. The selective activation of these genes dictates the properties of one tissue type versus another. We propose to use a new technology, DNA microarrays, to measure the activity of thousands of genes in OVCA samples simultaneously. While we expect each sample to have a unique fingerprint of gene activity, we anticipate that the patterns obtained will fall into groups where the fingerprints of the members are similar to each other. Since gene activity dictates the properties of cells, members of a given group should behave in similar ways. This is called molecular subtype analysis and the major focus of this grant application is to create a new classification of OVCA based at least in part on molecular subtypes.

Our study is based on banks of hundreds of frozen OVCA samples, plus a prospective collection of about another 1,000 cases over the course of the proposed work. These samples have extensive clinical and epidemiological information attached to them, such as response to treatment, oral contraceptive use, alcohol intake, number of children, etc. In addition, another project within this program will collect information on the frequency of specific variant genes in women with OVCA compared with controls. New molecular subtypes will be related to the epidemiological and genomic information to see whether there are any strong associations. For example, we may expect that carrying a certain variant gene or that exposure to certain potential carcinogens (talc use has been implicated in OVCA), will result in the development of particular molecular forms of OVCA.

The value of this work is substantial. Cancer is a disease of genes and if we are to understand the cause of OVCA and treat it more effectively, we must build a comprehensive picture of the genetic changes in different forms of the disease. This work is an exploration of the genetic landscape of OVCA, utilizing a very large number of samples with deep patient histories behind them. We expect this work to provide a genetic classification of OVCA that will provide insights into the growth and survival of cancer cells and help understand how a patient¿s genetic make up and environment may contribute to the development of OVCA.

Project 2: Program Project Award

Determinants of Epithelial Ovarian Cancer¿by Histologic Subtype and Tumor Behavior

Principal Investigator: Penelope M. Webb

Project 2: Public Abstract

Ovarian cancer is the fifth most common cancer in women from developed countries like the United States and Australia. Although we know that development of ovarian cancer is related to reproductive factors, for example, taking the oral contraceptive pill and having a larger number of children reduce a woman¿s chances of getting ovarian cancer, we still do not know enough about what causes this fatal disease to be able to prevent it. Ovarian cancer is frequently diagnosed very late, so a woman¿s prospects for survival following a diagnosis of ovarian cancer can be very poor, thus increasing the need to prevent its occurrence in future generations of women.

Many previous studies have considered ovarian cancer to be a single disease; however, we know that there are several different types of ovarian cancer, and it is becoming increasingly clear that these types may be caused by different things. Studies that do not distinguish between the different types of ovarian cancer are therefore likely to miss factors that may only be associated with one specific type of ovarian cancer. We propose to analyze data from a large, multicenter Australian study of women with epithelial ovarian cancer and control women without cancer in order to evaluate potentially modifiable causes of the different types of ovarian cancer. For example, one of the specific factors we are interested in is the use of post-menopausal hormones. Hormone use is increasing, and many women now take these hormones for long periods. Recent data suggest that taking hormone replacement therapy may increase a woman¿s chances of developing ovarian cancer; however, the results are not clear because hormone use may not have the same effect for all types of ovarian cancer. It is important to clarify this association because a woman who is considering taking post-menopausal hormones needs to be properly informed as to the potential risks and benefits of this if she has intact ovaries at menopause. Other factors that we are particularly interested in are alcohol consumption and cigarette smoking as previous data concerning the effects of these factors on ovarian cancer risk have been conflicting, possibly because there has been little attempt to look at the different types of ovarian cancer separately.

We will have data from more than 1,000 women across Australia who have recently been diagnosed with ovarian cancer and an equal number of cancer-free control women from the Australian Ovarian Cancer Study (described in Core A, ¿Epidemiology¿). Detailed information about women¿s reproductive and medical history and behavioral factors obtained by interview will be available, and we will compare women with different types of ovarian cancer to the cancer-free women to identify factors that are more common in the cancer patients. We will also combine these new data with comparable data collected by us for a previous study of ovarian cancer (Survey of Women¿s Health). This will give a combined group of more than 1,800 women with cancer and more than 1,800 control women.

This data set will be the largest of its kind and will enable us to look in greater detail at the causes of ovarian cancer than has previously been possible. In addition, we will have access to blood and tumor (cases only) samples collected from the women allowing us to look at genetic (Project 3) and molecular (Project 1) factors that may important in the development of ovarian cancer. In the future, we plan specifically to combine the information from this project with that from Project 3 (low-risk genotypes) to evaluate the relationship between genes and environment (nature or nurture) in the development of the different types of ovarian cancer.

Project 3: Public Abstract

Having a mother, sister, daughter, or aunt with ovarian cancer substantially increases a woman¿s risk of developing ovarian cancer herself. Some of this risk is likely to arise because these women carry genes that increase their risk by two- to threefold. The best way to identify these genes is to compare carefully chosen candidate genes likely to play a role in ovarian cancer susceptibility in a large number of ovarian cancer cases compared with controls. Candidate genes can be predicted from the many hypotheses put forward for the causes of ovarian cancer. It is likely that subtle but common variants in these genes may impact the individual¿s risk of disease.

We plan to compare particular genetic variants in the DNA from about 1,500 ovarian cancer cases to that from about 1,800 control women who do not have ovarian cancer. We will focus on genes that are involved in hormone synthesis and action and in those that repair damage to DNA, which can occur in a cell for a variety of reasons. If these repair genes are not working efficiently, it is likely that damaged DNA will be maintained instead of repaired. If the damage has occurred in genes involved in growth and spread of cells, cancer may result. Similarly, variants in genes that result in an excess of androgen in a woman may increase her risk of ovarian cancer. We are particularly interested in finding out whether different genes cause different types of ovarian cancer and whether certain women carrying particular genetic variants may be more susceptible to particular lifestyle factors.

The long-term aim of this project is to identify women at increased risk for ovarian cancer on the basis of their genetic makeup because this will allow them to be targeted for early screening and detection and for intervention studies. Furthermore, understanding the genes that are important in ovarian cancer susceptibility is likely to increase our understanding of what biochemical pathways are altered in ovarian cancer. A better understanding of the biology is likely to lead to more directed therapies in the future.

Core Facility A: Program Project Award

Epidemiology

Principal Investigator: Adèle C. Green

Core Facility A: Public Abstract

Ovarian cancer is the fifth most common cancer in women from developed countries like the United States and Australia, and yet we still do not know enough about what causes this often fatal disease to be able to prevent it. Ovarian cancer is frequently diagnosed very late, so the prospects for survival for a woman following diagnosis can be very poor, thus increasing the need to prevent its occurrence in future generations of women. We have previously conducted a large study of ovarian cancer in Australia (the Survey of Women¿s Health, SWH). This has shed light on many aspects of ovarian cancer, and we now plan to conduct a second large study to build on and extend the knowledge we gained from the first study and, in particular, to identify options for prevention of this disease.

Epithelial ovarian cancer is often considered a single disease, but there are actually several different types of ovarian tumors, and it is becoming increasingly clear that these may develop in different ways. Studies that look at all of these different types together may not be able to identify factors that only cause some types of cancer. For example, some studies have found that use of hormone replacement therapy (HRT) increases a woman¿s risk of getting ovarian cancer but other studies have found no association. Our preliminary data from the SWH suggest that use of HRT is associated with ovarian cancer but only one particular type (endometrioid ovarian tumors). This could explain the inconsistent results from other groups who have considered all types of ovarian cancer together. Similarly, we know that development of ovarian cancer is related to reproductive factors, for example, taking the oral contraceptive pill and having a larger number of children reduce a woman¿s chances of getting ovarian cancer, but it now seems likely that these factors are only protective against certain types of tumor (non-mucinous ovarian cancers).

It also seems increasingly likely that genetic factors must play a role in the development of ovarian cancer and that some women will, because of their genetic makeup, be inherently more susceptible to potential cancer-causing factors than others. For example, some women may, because of their genes, be better able to repair damage to their genetic material. These women may then be less susceptible to the adverse effects of compounds or exposures, such as radiation, that are known to damage genetic material.

A narrow approach to ovarian cancer will inevitably be limited in terms of the information it can provide. We therefore propose to conduct a multicenter, multidisciplinary study (the Australian Ovarian Cancer Study) that is large enough to be able to look at the different types of ovarian cancer separately and that will maximize resources by integrating a wide range of epidemiologic, molecular, and genetic tools in a concerted effort to increase our understanding of what causes some women to develop ovarian cancer.

We will do this by identifying a group of more than 1,000 women across Australia who have recently been diagnosed with ovarian cancer and a comparable group of women without ovarian cancer. With the permission of the doctors of the women with cancer and the consent of the women themselves, we will interview them to collect information about their past reproductive history and their behavior patterns. We will also collect blood and urine samples from all women and tumor samples from women with ovarian cancer. We will then compare the women with and without cancer to identify environmental and genetic factors that are more (or less) common among the women with cancer.

Importantly, we will also be able to combine these new data with data from our previous study (SWH), thereby creating the largest ovarian cancer resource of its kind. This combined resource will comprise epidemiological data from more than 1,800 women with ovarian cancer and 1,800 without cancer, with blood and tumor samples from more than half of the women. This will put us in a unique position to be able to investigate both genetic and environmental factors that cause the different types of ovarian cancer.

Core Facility B: Program Project Award

Biospecimens

Principal Investigator: David Bowtell

Core Facility B: Public Abstract

Ovarian cancer is one of the most complex human cancers, with many different types of disease recognized. Even among the more common forms of ovarian cancer, serous and mucinous, there is a wide spectrum of clinical responses to treatment. Underlying these different forms of the disease and variable responses to treatment must be genetic differences that dictate the properties of the cancer cells. We have only identified a fraction of the genetic changes in ovarian cancer, and we do not know how they interact with each other to drive ovarian cancer cell growth. We also know little about the genetic and environmental causes that trigger ovarian cancer. Although several genes have been identified that greatly increase the risk of developing ovarian cancer, only a small percentage of ovarian cancer cases are clearly inherited. The bulk of sporadic disease is thought to be dependent on an intersection between a person¿s genetic makeup, particularly genes that have a smaller influence, and ill-defined environmental effects.

We have designed three studies to investigate the cause genetic and environmental causes of ovarian cancer and the genetic forms the disease takes once it develops. In order to address these challenging questions, we need a collaborative, large-scale program to generate the statistical power needed to make clear conclusions. Our studies are based on a network of Australian gynecological-oncologists and researchers. Through this network, we will have immediate access to banks of ~860 frozen ovarian cancer samples for immediate use in genetic studies, plus a prospective collection of about another 600 surgical samples, and about 1,000 blood and urine samples over the course of the proposed work. The latter samples will have extensive clinical and epidemiological information attached to them, such as oral contraceptive use, alcohol intake, number of children, as well as pathological information.

The value of this work is substantial. Cancer is a disease of genes, and if we are to understand the cause of ovarian cancer and treat it more effectively, we must build a comprehensive picture of the genetic changes in different forms of the disease. This work is an exploration of the genetic landscape of ovarian cancer, utilizing a very large number of samples with deep patient histories behind them. We expect this work to provide a genetic classification of ovarian cancer that will provide insights into the growth and survival of cancer cells and to understand how a patient¿s genetic makeup and environment may have contributed to the development of ovarian cancer.