Robert A. Weinberg, Ph.D. - US grants

The DNAs of a series of different chemically induced and spontaneously arising tumors have been applied to NIH3T3 mouse fibroblast cultures and observed to elicit foci of transformed cells. These transformed recipient cells thus have acquired a transmissible gene for transformation present in the donor cell from which the DNA was prepared. These cells can be placed in syngeneic mice in which they induce fibrosarcomas. Sera from these mice can then be used to immunoprecipitate antigens present in the transfected cells. One of these antigens may be the transforming protein encoded by the transforming gene. Such experiments have led to the detection of a protein induced specifically by the transforming genes of rat neuroblastomas. Another protein of 21,000 daltons is known to be associated with various human ras oncogenes and a third with the myc oncogene. The proposed work is designed to examine structural and enzymatic properties of these tumor proteins and how their expressions are regulated during cell growth and differentiation. (X)

Oncogenic transforming sequences, termed oncogenes, have been detected in the DNAs of a variety of human tumors. These oncogenes have been assayed experimentally by the procedure of gene transfer (transfection), since they are able to induce transformation of recipient cells. Several distinct human oncogenes have been characterized and/or isolated as molecular clones. Three of these oncogenes are related to oncogenes carried by rat Kirsten or Harvey sarcoma viruses. These oncogenes are consequently members of the ras gene family. Each oncogene appears to derive via somatic mutation from a normal cellular gene termed a proto-oncogene. The Kirsten ras-related oncogene, termed cKi-ras2, is of particular interest, since it has been implicated in a variety of commonly occurring human tumors including carcinomas of the colon and lung. Molecular isolation of this important oncogene has been impeded by its large size and by the presence of sequences that inhibit molecular cloning. An alternative approach for oncogene isolation is proposed here. Since the important lesions creating the oncogene appear to affect protein-encoding sequences, a strategy for the rapid retrieval of cDNA clones of these genes is proposed. This involves construction of a chimeric retrovirus vector designed to retrieve the desired Ki-ras2 sequences from human tumor cells or derived transfected cells. In addition, a second chimeric virus, carrying an intact Harvey ras (cHa-ras1) proto-oncogene will be used to experimentally generate a group of independently activated oncogenic alleles. These naturally existing (cKi-ras2) or experimentally created (cHa-ras1) oncogenic alleles will be compared to explore the hypothesis that they are functionally different from one another and have distinct effects on tumor cell phenotype and consequent progression of tumor growth. The experimental system will also be exploited to attempt activation of the proto-oncogenes by known carcinogens, thus exploring the interaction of these agents with an important cellular target.

The use of gene transfer and recombinant DNA techniques has made it possible to identify a number of oncogenes that are active in human tumors and are presumably responsible for the altered properties of the tumor cells. Workers have identified a large number of these cellular oncogenes as well as oncogenes of viral origin. A main goal of the proposed work is to understand how the actions of these various oncogenes may be integrated into a scheme that describes the growth regulatory networks of the cell. Some of these oncogenes have been found to collaborate with others in tumorigenic conversion of cells. This has allowed a classification of some into a functional group characterized by having properties similar to the ras oncogenes. The paradigm of the second group is the myc oncogene. Proposed work will undertake to analyze the nature of other oncogenes. A second scheme by which oncogenes may be understood stems from the principle that these genes confer on the cell an autonomy from normally required growth factors. Such oncogenes may provide the cell with its own growth factors. Proposed experiments will investigate whether a growth factor gene can be activated into an oncogene. Alternatively, growth factor autonomy may be achieved by alterations in a cellular growth factor receptor. Proposed experiments will examine an oncogene protein that appears to function as a growth factor receptor. Other experiments will study ras p21 proteins that may transmit growth stimuatory signals sent out by the receptors. Longer term experiments will examine tumor genes that enable cells to metastasize or escape immune surveillance.

Support is requested for a scientific meeting organized by the American Association for Cancer Research, Inc. (AACR). The conference is entitled "Negative Controls on Cell Growth and Their Breakdown during the Pathogenesis of Cancer" and will be held at the Chatham Bars Inn in Chatham, Massachusetts, on October 20-24, 1991. The Organizing Committee for this conference consists of Drs. Robert A. Weinberg, Chairperson; Arnold J. Levine, David M. Livingston, and Eric J. Stanbridge. The invited speakers are well-known investigators in this field. They will make 30- to 60-minute presentations. Approximately 150-175 other conference registrants will have the opportunity to present their own novel work during poster sessions. The major goals of the conference are the presentation of the latest discoveries in the areas of the molecular origins of cancer, negative growth controls, and tumor suppressor genes. While oncogenes act positively on cell growth, a diverse group of genes may act to repress or constrain cell growth. These growth-suppressing genes may become involved in tumor pathogenesis when they become inactivated. Once a cell is deprived of the services of one or more such genes, it may no longer be able to respond appropriately to the signals that would, in the context of normal tissue, dictate the shutdown of its growth. Accordingly, the end product of this gene inactivation may be very similar to that seen upon the deregulation of proto-oncogenes. Indeed, since gene inactivation is more readily achieved through random mutations than this deregulation/activation, it may well be the case that the inactivations of such genes play a more important role in cancer pathogenesis than the activation of oncogenes. For these reasons, the field of negative controls on cell growth will assume increasing importance in the ensuing years and by mid-decade will likely attract as much research interest as the oncogenes. Participants will be drawn from a variety of scientific areas and all levels of training from graduate students to senior basic scientists or clinical investigators. The conference will provide substantial opportunities for both formal and informal scientific exchange among all participants and speakers. It is anticipated that this exchange will enrich current research and suggest new conceptual insights into the rapidly developing field of negative growth control.

Work of the research groups listed below at the Whitehead Institute for Biomedical Research (WIBR) is a focused on the use of molecular biology and genetics to analyze complex biological problems in developmental biology and genetics, mammalian genetics, control of cell growth, gene therapy, replication of human pathogens, signal transduction pathways and the structure of biological molecules. With the exception of one small research group using computational processes to analyze molecular structure, all the research groups at the WIBR rely to greater or lesser extents on the analysis and manipulation of DNA molecules in order to identify, analyze and dissect the genes governing the biological systems of interest to them. The use of DNA sequencing has figured prominently in these analyses and manipulations. New genes are frequently being isolated and analyzed in detail by sequence analysis. Even more important, however, are the years of experiments that follow during which isolated genes are dissected, reconfigured, perturbed by site-directed mutagenesis, and then functionally studied. This re-engineering of isolated DNA segments makes intensive use of DNA sequencing to monitor continually the progress of the successive steps-involved in the reconstruction and reconfiguring of gene segments. Indeed, the progress of these manipulations is often determined by the rapidity with which reconstructed DNA segments can be sequenced in order to monitor success of the most recent manipulation. All this work has been performed manually since the founding of the WIBR in its various laboratories with the exception of a recently established automatic sequencing facility that is owned by the Human Genome Project. (Use of this facility has been and will be limited exclusively to those working directly on this project. The existing equipment of the Human Genome Project will shortly be removed from its current WIBR premises to a new, off-site facility.) During the past decade, automated sequencing technology has been developed to the stage that it is practical and results in very substantial increases in laboratory productivity. The use of automated sequencing instrumentation speeds up a factor of 5-10 the rate with which sequencing can be carried, and it is anticipated that the investment made in this equipment will be recovered in terms of improved productivity of the WIBR's NIH-sponsored research programs within 2-3 years' time. Thus, the experience provided by the Human Genome Project's WTBR facility (see above) informs us that with such equipment, we should be able to sequence nearly 500 templates per week (or approximately 15,000 nucleotides per week) with this instrument.

DESCRIPTION: (Applicant's Description) An understanding of the molecular mechanisms leading to human breast cancer development will come from insights into how the breast tissue develops normally. Such insights can be most easily obtained in the mouse breast, the development of which is governed by biochemical regulators very similar to those that operate to program human breast morphogenesis. Breast development is governed by bi-directional communication between the mammary epithelium and surrounding stroma. The present research examines three signalling pathways that appear to be essential for normal breast development and appear to be disrupted during cancer pathogenesis. First, estrogen acting through its receptor in normal mammary epithelial cells (MECs) does not appear to be mitogenic but has acquired such powers in many human breast carcinoma cells. The proposed research examines the possibility that these powers result from the inappropriate expression of a co-activator of the receptor that is not normally expressed in MECs. A second aspect of breast signalling involves the progesterone receptor, which is responsible for sidebranching of mammary ducts. The present research examines the possibility that this receptor, when activator by progesterone, causes expression of hepatocyte growth factor in nearby stroma; this factor then acts on MECs to induce ductal sidebranching. A third line of research examines the possibility that prolactin, acting through its receptor in MECs, is able to induce the production of the neuregulin growth factor in nearby stromal tissue; once released, the neuregulin causes MEC proliferation leading to the formation of alveoli. The latter pathway is often subverted in human mammary carcinoma cells that express the HER2 receptor of neuregulin.

DESCRIPTION (provided by applicant): The research proposed here derives from an unexpected convergence of two areas of current cancer research. One area describes the fact that carcinoma cells exploit certain cell-biological programs that are usually operative in normal embryonic development and function to convert epithelial cells to cells that have a more mesenchymal (connective tissue-like) appearance and behavior. This conversion is often termed an epithelial-mesenchymal transition (EMT) and, when appropriated by carcinoma cells, serves to impart to these cells critical powers, such as invasiveness and motility, both of which are essential for cancer cell metastasis. A second research area has addressed the existence of small minorities of self-renewing cancer cells that exist amid far larger numbers of non-self-renewing cancer cells within tumors. The minority cells, often termed cancer stem cells (CSCs), are capable of seeding new tumors, while the cancer cells forming the majority of cells within a tumor lack this ability. These CSCs mirror the behavior of the stem cells in normal tissues, which also constitute small minorities of the total cell populations in such tissues and, because of their self-renewing abilities, are also responsible for the regenerative abilities of these normal tissues. Recent research has revealed that when normal mammary epithelial cells (MECs) are induced to undergo an EMT, they acquire the self-renewing powers of normal epithelial stem cells. Analogously, when tumor cells are forced to undergo an EMT, they acquire far greater tumor-initiating ability and thus are likely to be CSCs. The ramifications of these recent findings will be explored in the proposed research. For example, there is reason to believe that an EMT suffices to convey carcinoma cells from a primary tumor to a site of distant metastasis. Some of this research will examine whether the resulting disseminated cells, having experienced an EMT, now gain the ability to form macroscopic metastases. This research will also reveal whether CSCs arise preferentially from normal cells, or whether they can be formed equally efficiently from normal non-stem cells. In addition, genetically altered mice will be constructed whose cells fluoresce in the event that they activate one of the several transcription factors that are known to program the EMT. The cells of these mice will make it possible to screen for the cell-to-cell signals that induce epithelial cells to undergo the EMT and induce non-stem cells to become stem cells. PUBLIC HEALTH RELEVANCE: While being responsible for ~90% of cancer related mortality, the process of cancer-cell metastasis remains poorly understood;many of its steps are now being revealed by studying cell-biological changes that occur during early embryonic development. These changes also impart to cancer cells an ability to regenerate themselves, yielding "cancer stem cells" that can seed new tumors. The ability to produce large numbers of such cancer stem cells, as described in the proposed research, may present an opportunity to develop novel anti-cancer drugs that are able to eliminate tumors by eliminating the stem cells that allow tumors to grow back after initially successful treatments.

DESCRIPTION: (Applicant's Description) The primary goal of this project is to develop an in vitro system for the transformation of normal human mammary epithelial cells (HMECs) into tumor cells with a defined set of genes. The development of such a system would mark the first time that this has been achieved and greatly enhance our understanding of the genetic requirements for transformation of HMECs in several respects. First, the in vitro transformation of HMECs would allow us to define the minimal requirements for the transformation of these cells and define which cellular signaling pathways must be disrupted for tumorigenic conversion. Second, it would enable us link each genetic mutation with a distinct role in the transformation process. Following the creation of this purely genetic system for the transformation of HMECs, we will refine the model to more accurately reflect the development of human breast cancer in two ways. First, we will replace the initial cohort of transforming genes with genes known to be involved in human breast cancer. This will allow us to classify the genes which have been found to be mutated in breast cancer over the past two decades into groups based on their causal role in the transformation process. Second, we will determine the requirements for tumor formation in the mammary stromal environment by injecting the transformed HMECs into the mouse mammary fat pad and assessing their tumorigenic potential. Lastly, we will investigate the genetic requirements for two of the later stages in breast cancer progression, invasion and metastasis. These requirements will be addressed using both a candidate gene approach and a retroviral cDNA expression screen in order to isolate invasion and metastasis-promoting genes.

The present proposal is a request for funds to purchase a phosphorimager to be used by eight laboratory groups at the Whitehead Institute for Biomedical Research, affiliated with the MIT Dept. of Biology, both in Cambridge, MA, this application being made under the aegis of the NCRR Shared Instrumentation Grant. The research programs in these laboratories, which involve in aggregate approximately 105 research personnel, address problems ranging from the characterizing of genes governing early chordate embryogenesis to the genetic determinants of oncogenesis and the genetics of the yeast MAP kinase signaling pathway. At present, these laboratories all rely heavily on the analysis of radiolabeled molecules, which have been detected over the past three decades through the use of radioautography, involving thereby the use of X-ray films and developers. Over the past five years, a new type of instrumentation has been perfected that enables radiolabeled molecules to be detector with greater sensitivity (and hence in far less time), with linear response measurement over 5 orders of magnitude, and without the use of X-ray film and associated developing reagents. In addition, and as an alternative, the new instrumentation makes it possible to avoid radiolabeling altogether and instead enables the researcher to detect molecules either labeled with non-radioactive fluorescent dyes or with chemiluminescent substrates. This new technology depends on the use of europium crystals to detected emitted radioactivity and photomultipliers to detect light signals emitted (by non radioactively labeled) biological samples. In addition to offering the prospect of dramatically decreased use of radiolabeled materials and attendant exposures to isotope experience experienced by laboratory personnel, this instrumentation offers the prospect of dramatic decreases in both the expenses and waste products that are entailed by the use of X-ray film and the associated chemicals needed to develop exposed film. The 8 laboratory groups which are prospective users of this instrument spend an aggregate $72,000 yearly on autoradiographic supplies and equipment, much of which expense would be obviated by the proposed instrument.

DESCRIPTION (provided by applicant): Metastasis is the process by which tumor cells spread from a primary site to distant organs and establish secondary tumors. Because metastases are difficult to detect and cure, these growths are the cause of approximately 90% of cancer mortafity. To form metastases, tumor cells need invade through the surrounding tissue, intravasate into the circulation. After traveling in the circulation, surviving cells extravasate into a distant site, and establish new malignant colonies (secondary growth). A key issue in elucidating the mechanisms of tumor metastasis is the question of whether distinct genetic or epigenetic changes are required for each of the individual steps of metastasis. For this reason, it has become crucial to identify the genes and gene products responsible for programming these changes and to dissect their functions. In this proposal, a mouse mammary tumor metastasis model and a human mammary tumor model system are investigated to search for genes that function as pleiotropically acting, central regulators of the metastatic cascade. Using DNA microarray technology, we have identified several candidate genes whose expression strongly correlates with the ability to accomplish specific steps of the metastatic process. Detailed validation and characterization will be performed in both systems to understand the roles these genes play in the metastatic process. Study of these metastasis-associated genes will open up a large number of experimental avenues and allow detailed dissection of the cellular and molecular mechanisms of this multi-step process.

(Revised Abstract) DESCRIPTION: This Program was begun in 1999 to foster the establishment of a consortium of laboratories in the Boston- Cambridge area that shared a common interest in elucidating the molecular mechanisms of breast cancer development. The proposed Program will continue the monthly meetings and annual retreat of all researchers involved in the research sponsored by this Program. The participating groups have specific interests in various of the molecular and biological steps that cause normal human mammary epithelial cells (MECs) to develop progressively into carcinoma cells. David Livingston's research addresses the BRCA1 and allied proteins, defective forms of which are associated with the destabilization of the genome of MECs; recent research indicates that the BRCA1 gene may function in yet other ways to prevent the initiation of breast cancers. Joan Brugge's research is concentrated on characterizing the molecular mechanisms by which certain aberrantly functioning genes are able to perturb the morphogenesis of the acini in the mammary epithelium, thereby triggering some of the earliest histological abnormalities in the human breast. Mark Ewen's work is focused on understanding the molecular mechanism by which cyclin D1 -- a protein that is frequently over-expressed in human breast cancer cells -- is able to derail the normal proliferation and differentiation of human MECs. Peter Sicinski's research is directed toward understanding how mitogenic signals (normal or oncogenic) impinging on the surface of a MEC are able to communicate with the cell cycle apparatus in the nucleus of such a cell, in part by inducing expression of D-type cyclins. The work of Myles Brown addresses a similarly important attribute of many advanced breast carcinomas -- the actions of the estrogen receptor (ER) and the mechanisms through which collaborating transcriptional regulators (i.e., co-activators) are able to mediate the activating and suppressive functions of the ligandbound ER. Finally, Robert Weinberg's research is directed at three distinct areas of breast cancer development -- the reconstruction of normal breast tissue in mice, the possible role of prolactin in creating localized mammary hyperplasias, and the actions of a transcription factor in imparting the attributes of invasiveness and metastasis to already-tumorigenic MECs.

PROJECT SUMMARY (See instructions): The Administrative Core of this Project shoulders several distinct responsibilities. The administrative tasks are carried out by an administrative assistant in the Principal Investigator's office -- that of Robert A. Weinberg. Most central is the task of receiving funds from the NIH and disbursing these funds to the various Projects. These funds are passed through the Sponsored Programs Office of the Whitehead Institute for Biomedical Research, and the Administrative Core ensures that the funds issued to the participating Projects are allocated appropriately. The three other institutions in which this Program takes place- the Dana-Farber Cancer Institute, Harvard Medical School and the Brighams and Women's Hospital - are paid as sub? conctactors by the Whitehead Institute for Biomedical Research. The Administrative Core is also responsible for organizing the monthly meetings of the Program, including scheduling, notifying all participants of scheduled meetings, inviting speakers, and organizing the yearly Retreat of this Program, which involves extensive planning for the transportation of guests, accomodations for all participants, arrangement of the agenda of speakers, procurement of food and audio visual equipment.

Attempts at understanding the detailed molecular mechanisms of breast cancer pathogenesis have been frustrated in part by the lack of experimental models that adequately recapitulate the biological attributes of normal and neoplastic human breast tissues. This shortcoming has inspired the construction of human mammary gland tissue in immunocompromised mice that appears to function in a fashion that closely resembles that of the normal human mammary gland. With the availability of this model, the precise biological and morphogenetic actions of various genes that are thought to participate in human breast pathogenesis can be examined. The mechanisms that lead to the hyperplasias seen in the human breast epithelium are still poorly understood. Prolactin (Prl) of pituitary origin is normally involved in promoting the development of the mammary gland late in pregnancy. Research to be conducted examines how Prl is able to elicit histopathological alterations in the mammary epithelium that represent early steps of tumor progression. A genetically altered mouse model that has been developed will prove highly useful in elucidating the biochemical changes in mammary epithelial cells (MECs) that are directly induced by the ligand-activated Prl receptor. A third line of research addresses the molecular mechanisms that enable tumorigenic but non-invasive MECs to acquire the traits that enable such cells to invade and metastasize. Three transcription factors (TFs) have been characterized that are capable of activating the epithelial-tomesenchymal transformation (EMT) that appears to be closely linked to the acquired ability of MECs to invade and thereafter to metastasize. Detailed research will be focused here specifically on the Goosecoid TF. On the one hand, expression of this TF can be induced in epithelial cells by TGF, a protein that has frequently been associated with the invasive front of primary tumors. On the other, Goosecoid is capable of programming many of the changes associated with the EMT, including loss of epithelial markers, acquisition of mesenchymal markers, and increased motility. The functional relationships between Goosecoid and two other TFs that function analogously will be explored in an attempt to determine whether a single signaling circuit involving these three TFs is responsible for programming all types of cancer-associated invasiveness, or whether different TFs are associated with distinct invasion-controlling regulatory circuits in human MECs.

Carcinomas comprise about 90% of all human tumors and arise through the genetic alteration of epithelial cells. Many human carcinomas, including those of the breast, consist of neoplastic epithelial cells intermingled with a complex mixture of cells, including fibroblasts, myofibroblasts, endothelial cells and immune cells, which collectively form the tumor stroma. The supporting stroma of a carcinoma is distinct from that found in the adjacent normal tissue, and as the carcinoma expands, these stromal cells are recruited and induced to proliferate. The presence of stromal cell types within carcinomas is a powerful determinant of the biology of a tumor through its ability to have profound influences on the growth, survival, invasiveness, and metastatic powers of the neoplastic epithelial cells. The presence of these cells has been demonstrated by our laboratory and others to have an essential role in supporting tumor progression; however, the mechanisms of stromal recruitment and how these cells contribute to tumor growth are not well understood. The proposed studies are directed at elucidating the recruitment of stromal cells and the mechanisms by which they influence tumor progression by accomplishing the following: A1. Determining the origins of the tumor-associated stromal cells; A2. Determining the nature of the signals used to recruit stromal precursor cells; A3. Characterizing the heterotypic signals released by recruited stromal cells that influence invasion and metastasis by carcinoma cells; A4. Determining the mechanisms by which mesenchymal stem cells can enhance metastatic ability.

Attempts at understanding the detailed molecular mechanisms of breast cancer pathogenesis have been frustrated in part by the lack of experimental models that adequately recapitulate the biological attributes of normal and neoplastic human breast tissues. This shortcoming has inspired the construction of human mammary gland tissue in immunocompromised mice that appears to function in a fashion that closely resembles that of the normal human mammary gland. With the availability of this model, the precise biological and morphogenetic actions of various genes that are thought to participate in human breast pathogenesis can be examined. The mechanisms that lead to the hyperplasias seen in the human breast epithelium are still poorly understood. Prolactin (Prl) of pituitary origin is normally involved in promoting the development of the mammary gland late in pregnancy. Research to be conducted examines how Prl is able to elicit histopathological alterations in the mammary epithelium that represent early steps of tumor progression. A genetically altered mouse model that has been developed will prove highly useful in elucidating the biochemical changes in mammary epithelial cells (MECs) that are directly induced by the ligand-activated Prl receptor. A third line of research addresses the molecular mechanisms that enable tumorigenic but non-invasive MECs to acquire the traits that enable such cells to invade and metastasize. Three transcription factors (TFs) have been characterized that are capable of activating the epithelial-tomesenchymal transformation (EMT) that appears to be closely linked to the acquired ability of MECs to invade and thereafter to metastasize. Detailed research will be focused here specifically on the Goosecoid TF. On the one hand, expression of this TF can be induced in epithelial cells by TGF, a protein that has frequently been associated with the invasive front of primary tumors. On the other, Goosecoid is capable of programming many of the changes associated with the EMT, including loss of epithelial markers, acquisition of mesenchymal markers, and increased motility. The functional relationships between Goosecoid and two other TFs that function analogously will be explored in an attempt to determine whether a single signaling circuit involving these three TFs is responsible for programming all types of cancer-associated invasiveness, or whether different TFs are associated with distinct invasion-controlling regulatory circuits in human MECs.

DESCRIPTION (provided by applicant): Much attention has focused in recent years on the promise of using embryonic stem (ES) cells and their differentiated progeny to regenerate various adult tissues. We believe, however, that for the foreseeable future, a far more viable and attainable strategy will derive, at least in the case of epithelial tissues, from experimental protocols designed to induce the reprogramming of differentiated adult cells into committed progenitors and stem cells, which can then serve as vehicles for the regeneration of adult tissues depleted by aging or by various pathological processes. For this purpose, we intend to develop a cocktail of signaling agents that can be applied to differentiated epithelial cells in order to reprogram them to lineage-committed epithelial stem cells. Our strategy is to re-activate a latent developmental program, termed the Epithelial-Mesenchymal Transition (EMT). We have previously shown that induction into EMT confers on differentiated mammary epithelial cells many of the traits associated with epithelial stem cells. Importantly, we believe that our rationale is applicable to a wide range of differentiated epithelial cells, including stratified squamous epithelia, such as the mucosa of the mouth. EMTs effect key morphogenetic steps during various steps of embryonic development. In general, research has focused on the ability of EMTs to switch cellular fate from epithelial to mesenchymal, thereby causing profound changes in cell morphology and function. However, our recent findings indicate a novel physiological role for EMTs in stem-cell self-renewal and thus in tissue homeostasis. Reinforcing this association between the products of EMT and the stem-cell state is the finding that putative human and mouse mammary stem cells have been shown to possess mesenchymal traits, further linking EMT to stem cell biology. Importantly, the close parallels between EMT and the stem-cell state do not appear to be an idiosyncrasy of the mammary epithelium: emerging evidence suggests an essential role for the EMT-inducing Transcription Factors (TF) Snail and Slug in skin morphogenesis. During the course of development, epithelial-mesenchymal transitions (EMTs) are induced and coordinated by contextual signals received by individual cells from adjacent cell layers. Thus, the convergence of several paracrine signals on an embryonic epithelial cell induces that epithelial cell to undergo an EMT. Based on these observations, we have hypothesized that the initial triggering of an EMT program within a cell depends on concomitant stimulation of that cell by a defined set of such paracrine signals, and that maintenance of the resulting mesenchymal/stem-cell state depends subsequently on the activation of autocrine signaling loops involving the same factors that previously triggered entrance into the EMT. Indeed, our preliminary data are in concordance with this model, underscoring the importance of the extracellular signaling environment for regulating EMTs, and reinforcing our rationale to develop an experimental protocol with a cocktail of signaling agents as a means to dedifferentiate and reprogram differentiated epithelial cells to a stem cell-like state. In the longer term, we intend to develop a protocol for inducing the controlled dedifferentiation of a number of distinct adult epithelial cell types into their corresponding stem cells. Such protocols have the potential of generating large numbers of committed epithelial stem cells that hold promise for the regeneration of a variety of depleted adult epithelial tissues. PUBLIC HEALTH RELEVANCE: Much attention has focused in recent years on the promise of using embryonic stem (ES) cells to regenerate various adult tissues. We believe, however, that for the foreseeable future, a far more viable and attainable strategy will derive from experimental protocols designed to induce the reprogramming of differentiated adult cells into committed progenitors and stem cells. These reprogrammed cells can then serve as vehicles for the regeneration of adult tissues depleted by aging or by various pathological processes. For this purpose, we intend to develop a cocktail of signaling agents that can be applied to in vitro cultures of differentiated epithelial cells in order to reprogram them to committed epithelial stem cells via activation of a latent developmental program. We believe that this strategy can be applied to many epithelial tissues, such as the skin or mouth mucosa.

Mammary Neoplasms; Stromal Cells

The acquisition of highly malignant traits by carcinoma cells enables their invasion and metastatic dissemination. In many, and possibly all carcinomas, these traits are conferred by a cell-biological program termed the epithelial-mesenchymal transition (EMT). The EMT program is induced in carcinoma cells, usually reversibly, by contextual signals that they receive, largely if not entirely from the adjacent reactive stroma that is present in many highly progressed tumors. While the intracellular signaling circuits that enable activation of EMT programs have come into view, the nature of the extracellular, contextual signals that activate these programs in carcinoma cells is still obscure as are the cellular origins of these signals. This Indicates that the ultimate governors of malignant progression - the stromal cells and the signals that they release - are still poorly understood. The proposed research derives from the convergence of two lines of research in the Pi's laboratory. First. Twist - an EMT-inducing transcription factor - is responsible for the malignant behavior of highly aggressive 4T1 mouse mammary carcinoma cells; its expression is induced in vivo, ostensibly by signals that the implanted 4T1 cells receive from the stromal cells that they recruit. Second, these carcinoma cells induce splenomegaly in tumor-bearing syngeneic mouse hosts. Rather than being an epiphenomenon of tumor pathogenesis, this splenomegaly plays a causal role in 4T1 tumor formation, since in the absence of a functional spleen (achieved by splenectomy), the expression of the Twist and Snail EMT-TFs and their ability to spawn metastases are all significantly reduced. Hence, the local malignant state of these tumors, and likely many other tumors, is controlled by systemic factors, in this case myeloid cells recruited from the spleen. These dynamics suggest that certain carcinomas, such as the 4T1 tumors studied here, induce splenomegaly; the myeloid cells that collect In the spleen may then be recruited via the circulation into the tumor-associated stroma, where these cells proceed to release heterotypic signals that induce an EMT in the 4T1 cells, enabling the latter to metastasize. Importantly for the proposed research, the spleens of these mice should represent a rich source of the myeloid cells that participate in the induction of the EMT and thus in the triggering of metastatic relapse. The research that is proposed will examine these cells as well as others that are recruited into the tumor-associated stroma in order to develop a list of the functionally important, malignancy-promoting cells in the tumor-associated stroma and to uncover the EMT-inducing paracrine signals that they release.

DESCRIPTION (provided by applicant): About 15-20% of breast cancer patients fall under the basal-like category, which represent a diverse subtype that is characterized by tumors that are more aggressive conferring poor prognosis. Deaths from these carcinomas result from metastatic spread of the disease to distant sites and from therapeutic resistance, which results in the relapse of cancers into more aggressive forms that are difficult to contain. Both these properties are attributed to their cellular heterogeneity that arises through various mechanisms including clonal evolution and the presence of cancer stem cells (CSCs). The epithelial-to-mesenchymal transition (EMT) is one program that we have shown to be responsible for the generation of cells that have CSC-like properties. Our current proposal aims to induce differentiation of these CSCs through the induction of a mesenchymal- to-epithelial transition (MET). To do this, we carried out a screen to identify compounds that are capable of inducing the transcription of E-cadherin, a hallmark of the epithelial/non-CSC state, in cells that are more mesenchymal/CSC-like. Through this screen we identified Forskolin, an activator of cAMP, to be able to induce E-cadherin transcription and a reversion of the mesenchymal/CSCs to a more benign epithelial state. Through this proposal we aim to uncover the mechanism by which cAMP-elevating agents are able to induce an MET by complete characterization of the essential downstream components of signaling, namely Protein Kinase A (PKA) and its downstream substrates. We also aim to uncover specific GPCR- ligand pairs that could modulate cAMP levels thereby serving to maintain the epithelial state. Through extensive in vivo studies, we plan to develop a novel targeting strategy that combines the use of cAMP- elevating agents as adjuvants to chemotherapeutic drugs. This would serve as a means of depleting CSCs through cAMP-mediated differentiation as well as non-CSCs by hemotherapy in basal-like breast cancers.

This research is focused on the molecular mechanisms that govern whether a primary breast cancer will or will not become aggressive, metastatic and thus life threatening. At present, the ability to predict such aggressiveness is imperfect, in that there is great inter-individual variability in the behavior of a group of tumors that are all classified together in one of the major subgroups of breast cancer and thus predicted to share a common prognosis. At present, this inability to generate accurate predictions of the future behavior of individual breast cancers leads to aggressive treatment of the great majority of diagnosed tumors, when only a minority are destined to become life threatening. The research describes three major determinants of malignant progression of breast cancer cells and how they conspire to generate aggressive behavior. These are (i) the ability of carcinoma cells to release pro-inflammatory signals; (ii) the reciprocal responses of nearby mesenchymal stem cells within the stroma of tumors to these carcinoma-derived signals, resulting secondarily in the release of signals that have the potential of inducing carcinoma cells to move from an epithelial(benign) to mesenchymal (malignant) state; and (iii) the propensity of the carcinoma cells to respond to these stroma-derived signals by undergoing this shift in differentiation state, thereby acquiring highly aggressive characteristics. The propensity of cancer cells to move from an epithelial/benign to a mesenchymal/malignant state appears to be governed by the state of the chromatin associated with the gene that encodes ZEB1, which functions as the key molecular governor of the epithelial vs. mesenchymal states. Examining the chromatin configuration of this gene- more specifically the covalent modifications of the histones associated with the promoter of this gene - holds the promise of revealing the proclivity of a breast cancer cell to activate its program of malignant conversion, often termed the epithelial-mesenchymal transition.

DESCRIPTION (provided by applicant): The epithelial-mesenchymal transition (EMT) is a complex cell-biological program that operates during the progression of carcinoma cells to high-grade malignancy, conferring on these cells many of the attributes associated with aggressive tumors, including the ability to disseminate to distant sites and to seed metastatic colonies. This program is orchestrated by a series of pleiotropically acting master transcription factors (EMT- TFs) that organize the complex changes in gene expression causing the replacement of a large cohort of epithelial cell proteins with those associated with the mesenchymal cell state. A major, critical level of control required for expression of the aggressive mesenchymal state is poorly understood however: the precursors of many of the mRNAs whose expression changes during the EMT also undergo alternative splicing (AS) that confer on resulting mature, processed mRNAs altered properties, including changes in stability, protein-coding information, and responsiveness to microRNA-mediated inhibition. The current fragmentary insights into the effects of AS on the execution of the EMT program make it impossible to form a reasonably complete understanding of how this critical cell-biological program is effected. The proposed research will begin by enumerating the hundreds of AS events that occur in response to several alternative mechanisms of inducing an EMT program both in cultured cells and in a living tissue. Having done so, bioinformatics algorithms will be employed to determine the sequences adjacent to involved splice sites. Thereafter, using the known nucleotide-recognizing properties of the large array of already-characterized RNA- binding, splice-regulating proteins, predictions will be made by these algorithms about the identities of the splice-regulators that are likely to b responsible for the observed large-scale shifts in AS occurring during passage through an EMT. This experimental strategy should yield the identities of key regulators of AS that are likely to b as important functionally as the EMT-TFs in executing the EMT program. Experimental tests designed to functionally test the candidacies of these AS factors will be performed. These tests will gauge whether the forced or blocked expression of these factors affect execution of critical components of the EMT program, and whether, as predicted, such imposed changes in AS factor expression affect the production of key EMT-associated proteins, i.e., proteins that play key roles in the expression of the epithelial versus mesenchymal cell phenotypes observed during malignant progression. This work also has the potential to identify novel biomarkers of the EMT program that are applicable, for example, for the detections of stem cells in a variety of epithelial tissues.

Project Summary/Abstract: Much of current molecular oncology research is grounded on the paradigm that the behavior of cancer cells and the tumors that they form is dictated by the mutant genomes that these cells have acquired en route to entering into neoplastic/cancerous states. Implicit in this work is the notion that non-genetic mechanisms (that are not directly influenced by DNA sequence) ? notably changes organized by non-genetic (i.e., epigenetic) regulatory circuits operating within a cancer cell ? play a secondary role in dictating its behavior. In fact, the majority of cancer cell traits may actually be governed by non-genetic regulatory programs that operate in a relatively stable fashion to ensure that many of these traits are transmitted heritably from normal cells-of-origin through multiple cancer cell generations during the course of multi-step tumor development. In addition, activation within carcinoma cells of the cell-biologic program termed the epithelial-mesenchymal transition (EMT) creates a second dimension of cancer cell behavior that is not dictated by the cells' DNA sequences, acting by converting relatively benign carcinoma cells to malignant derivatives. These two non- genetic regulatory mechanisms exert profound effects on the malignant progression of cancer cells and their responsiveness to various forms of therapy. However, relatively little is known about how these non- genetically determined traits of cancer cells are acquired and disrupted during tumor formation. Super-enhancers (SEs) represent large aggregations of transcription factors (TFs) associated with a relatively small number (several hundred) of gene promoters in both normal cells and cancer cells. These SEs are responsible for orchestrating the differentiation programs of a variety of normal cell types that enables them to become tissue-specialized cell types. The proposed research examines the disruptive effects of specific oncogene-encoded proteins (and/or loss of tumor suppressor gene proteins) on the spectrum of SEs within experimentally transformed human cells from a variety of normal cell lineages, with the goal of understanding which SEs survive disruption and continue to influence the behavior of resulting tumorigenic cells (with respect to their responsiveness to therapy and their malignant traits), and which are altered, resulting in the acquisition of novel cancer cell traits. As a complement to these analyses will be an examination of the SEs associated with the EMT programs activated in various types of transformed cells, these being activated either through the induced expression of EMT-inducing TFs or the exposure of transformed cells to known EMT-inducing growth factors and cytokines. By surveying SEs in normal cell types and their neoplastic derivatives, the proposed work will generate an experimental platform that will enable many researchers to finally elucidate at the molecular level why and how various distinct subtypes of human carcinomas naturally exhibit an array of key biological traits or do so in response to imposed cytotoxic therapies.