HRBC Genomics Network project

Objectives

Breast cancer (BC) accounts for about than one fourth of all new cases of women's cancer and the mortality caused by these tumors has remained almost unchanged for the past 5 decades, becoming only second to lung cancer as a cause of tumor-related death.

The failure in eradicating these diseases is largely due to the lack of identification of a specific etiologic agent and of the molecular mechanisms responsible for BC initiation and progression. Clinical and experimental evidences, however, point to ovarian hormones, estrogens in particular, as the main causative agents of BC, and therapies aiming at interfering with the actions of these hormones are widely used for prevention and cure of BC. Indeed, adjuvant endocrine therapy based on estrogen ablation is very effective in reducing cancer recurrence and mortality. These therapies rely on the use of drugs affecting estrogen synthesis, metabolism or actions in tumor cells, and are most effective against hormone-responsive tumors, where estrogen is a primary growth promoting stimulus, whereas they generally fail when cancer growth is hormone-independent.

Hormone-responsive BCs are currently identified by the presence of ERs and of estrogen-regulated progesterone receptors (PRs). Comprehensive clinical trials have shown, however, that about one third of ER/PR positive tumors fail to respond to endocrine therapy, while about 10% of receptor negative ones do respond. Prognostic evaluation of hormone responsiveness based on ER expression in tumor cells, in fact, can be misled by the existence of subtle mutations of these proteins affecting their function but undetectable by routine protein expression assays. Furthermore, estrogen signaling in target cells is complex, involving at least two receptors (ER-alpha and -beta) and a host of other cellular components, as well as posttranslational modifications of the proteins themselves, e.g. a number of variables that can not be analyzed at present in surgical tumor samples to allow a more accurate evaluation of their hormonal responsivity status. Receptor-positive BCs are thus heterogeneous with respect to estrogen sensitivity and new molecular tools to allow discrimination of responsive from non responsive tumors are much needed and would find immediate and wide clinical applications.

The proposed research will focus on the problem of hormone responsiveness of BC, and the loss of this phenotype which might occur during genesis and progression of the tumor. To this aim, the studies will concentrate on identifying and characterizing new molecular markers of estrogen-dependent signaling in normal and mammary gland cells, in the form of a large number of novel genomic targets of estrogen and of hormone-related gene expression profiles, exploitable to achieve new molecular sub-typing of BC lesions and of practical use to improve clinical management of cancer patients (diagnostic and prognostic evaluation and prediction of responsiveness to endocrine-therapy, in particular).

Gene expression profiling with DNA microarrays and analysis of cancer cells proteome are promising new ways to achieve molecular classifications of tumors in clinically relevant subtypes and to identify new markers for immunophenotypic characterization of BC. Parallel analysis of the products (mRNAs and proteins) of large sets of estrogen-responsive genes in BC specimens, in fact, makes it possible to go beyond individual genes or markers (in this case ER and PgR), and to identify cancer-specific molecular 'signatures', that can help identify previously undefined tumor subtypes and, in the case of hormone-responsive tumors, to select more accurately patients that are more likely to respond to estrogen ablation therapy, while at the same time sparing them from unnecessary radical, or ultimately ineffective, treatments. Furthermore, the identification of new genes and pathways responsive to physiological stimuli that affect breast carcinogenesis and tumor progression might lead also to the identification of new means for BC prevention and treatment. These results can be best achieved when gene activity in a given cell type can be related to specific genetic or biological features.

For this reason, the identification of gene expression profiles associated with BC cells responses to relevant physiological stimuli has been proposed to be particularly important, as it could provide the framework necessary for interpreting the biological meaning of gene expression patterns observed in tumor samples. In the case of estrogen, this applies to the identification of hormone responsive gene sets useful to monitor hormone signaling status and their application to subtype BC specimens according to this functional phenotype. Hormone responsive human BC cell lines represent one of the best experimental models available to elucidate estrogen actions. Their intensive study led so far to the discovery of hormone-responsive genes, including for example identification of PRs as a downstream targets of estrogen, that were then exploited to improve clinical management of breast and other hormone-responsive cancers. The same applies to analysis of the effects of estrogen on rodent mammary gland transcriptome and proteome, in particular when the response to estrogen is compared between the mammary gland of wild-type and transgenic mice lacking one or more ER subtype (alpha and beta), or carrying trans-dominant oncogenes involved in breast carcinogenesis.

The research plan have thus been conceived to include also a careful investigation of all key aspects of estrogen signaling in mammary gland and BC cells. Genomics (gene expression profiling with cDNA microarrays), proteomics and bioinformatics approaches will be systematically applied to this aim, together with unique experimental tools available to the participating laboratories, e.g. transgenic mice (including ER-alpha and -beta knowk-outs) and engineered human BC cell lines useful for genetic dissection of estrogen-responsive pathways. This study will thus provide exploitable informations on the molecular bases of estrogen actions in BC cells, on the functions of the products encoded by the newly identified hormone-responsive genes and their biological roles in BC cells. At the same time, the possibility that some of such genes might represent novel and informative molecular markers for molecular and immunophenotypic characterization of breast tumors will be investigated by the clinical partners of the consortium, that will carry out morphological, cytogenetic, genomic and proteomic analysis of breast cancer biopsies, focusing mainly on the genes identified in the hormone-responsive cell lines and the other model systems (transgenic mice) described above, and search for correlations between expression or mutations of these genes and the clinical outcome of the disease.

Preliminary research carried out by participant laboratories led already to the identification of about 500 genes whose expression is regulated by estrogen in human BC cells. Such genes represent the starting point of the planned research, and their number is expected to increase significantly as the search for hormone-responsive genes will be expended to include a much larger fraction of the genome. They will be subject of study in all selected experimental models, and in BC biopsies, by complementary technological approaches (genomics, proteomics and bioinformatics). Tumor resistance to endocrine therapy develops in about one third of BC cases, as tumor cells become unsensitive to drugs interfering with the actions of estrogens (antiestrogens). Resistance to one type of endocrine therapy (treatment with the non-steroidal antiestrogen tamoxifen, for example) is not always accompanied by resistance to another endocrine-disrupting treatments (with steroidal antiestrogens, such as ICI 182,780). Where individual differences on drug sensitivity or metabolism can be excluded, these facts suggest that each class of antiestrogens is likely to have specific targets within the estrogen signalling pathway.

Indeed, significant differences in the cellular and molecular actions of steroidal and non-steroidal antiestrogens have been identified in BC cells in culture, while binding of each of these compounds induce specific conformational changes on ERs structure. The identification of more effective antiestrogens, and the ability to accurately predict the efficacy of these drugs, are a necessary pre-requisite for the establishment of therapeutic protocols tailored to specific features of the patient and her disease. This, however, first requires that more knowledge of the estrogen-dependent and -independent gene networks affected by antiestrogens in BC cells are identified, that the expression and activity of such genes are investigated in tumor samples and the results are confronted with the clinical outcome of the disease.

This project will thus aim at acquiring also new knowledge of the genetic and molecular pathways mediating the growth-inhibitory and pro-apoptotic actions of antiestrogens in cancer cells. The same experimental approaches and technologies describe above will be applied in these studies, while they will benefit also from an investigation of the 'in vivo' effects of antiestrogens in the transcriptome and proteome of primary BC lesions, that will be carried out by one of the partners. The failure rates of the current treatments for BC prompted the search for new therapies against these tumors. A particularly promising avenue, that aims at targeting other members of the nuclear receptors family (retinoid/rexinoid receptors -RARs/RXRs, in particular, but also peroxisome proliferator-activated receptor -gamma and vitamin D3 receptors), has been proposed and is currently being clinically tested not only for therapy but also for prevention of BC. Fenretinide (4-hydroxypnenyl retinamide) for example, a synthetic retinoic acid derivative that decreases the occurrence of a second breast malignancy in premenopausal women, is already undergoing clinical testing in large BC prevention trials. Also, drugs targeting downstream components of nuclear receptor-responsive pathways (histone-modifying enzymes in particular) are being conceived and tested for breast and other cancers treatment.

The development of new drugs targeting the RAR/RXR-responsive signaling pathways still requires, however, more understanding of their effects on mammary gland and BC cells proliferation, differentiation and apoptosis, and of the underlying molecular and genomic bases. One of the participating laboratories will thus focus its research on the analysis of the mechanism of action of drugs affecting nuclear receptors activity on the genome. This will be carried out by investigating the effects of RAR/RXR ligands in the activity of estrogen-responsive gene networks and of the effects of drugs affecting histones acetylation status on estrogen and retinoid/rexinoid signaling in hormone-responsive BC cells.

Competitive post-genome research requires building up and merging expertise and resources in genetics, genomics, proteomics and bioinformatics, to be able to attain significant achievements and exploit them to practical ends. Despite the fact that this area of research, stemming essentially from completion of the Human Genome Project, is relatively novel and still unexplored, resources and competence can be found in the Italian scientific community. Almost invariably, however, these are scattered between different laboratories and Institutions, a fragmentation that represent an obstacle toward achieving the critical mass essential to pursue and implement truly novel and far-reaching post-genome research, particularly in the field of cancer research.

One the primary objectives of this project, and the driving force beyond its conception and design, has been to bring together scientific excellence and competence in estrogen and nuclear receptors action, clinical management, pathology and cytogenetics of BC as well as in key areas of post-genomics (proteomics and bioinformatics, in particular), and to focus them toward the common goals of increasing the knowledge base on hormone-responsive BCs and expanding the scientific and technical capabilities of the participating Institutions and scientists in post-genome research. BC was selected for its high incidence and negative impact on the patients, and for the problems it raises for clinical oncologists. The hormone-responsive phenotype of this disease was selected, instead, to focus the research on a specific and well defined biological problem, avoiding the dispersion which might result from a too broad and less defined scope, the risk of a generic study on the transcriptomes and proteomes of breast tumors, for example. The same logic underlines the decision to study predominantly estrogen-responsive gene networks, as these can be first identified in well controlled and reliable model systems (hormone-responsive cell lines and tissues) and it is well known and demonstrated how they are most likely to include key genes for understanding the biological, clinical and pharmacological behavior of hormone-responsive BC cells and the differences that distinguishes them from hormone-unresponsive ones.

A post-genome research project should also aim at providing new data on differences in genome activity between specific cell types and on the expression, regulation and nature of the products of a relatively large number of genes. In our opinion, this can be best achieved by focusing on a set of genes that are part of a functional network (estrogen response in this case) and that are thus likely to undergo similar regulation and to have complementary roles in the cell (control of cell cycle progression, cell survival and differentiation status, in this case).

Finally, the decision to concentrate the transcriptome and proteome studies on the same set of genes will make it possible to implement a common microarray platform, to be shared between all laboratories involved in gene expression profiling experiments of human cell lines and BC specimens, thereby reducing the costs and experimental variables of such screenings, to rationalize proteome and bioinformatic analyses and to build a truly useful database.