University of Southern California Dissertations and Theses

Analysis of multistep mammary tumorigenesis in Wnt1 transgenic mice: The role of fibroblast growth factor-8 in oncogenesis and apoptosis

(USC Thesis Other)

Analysis of multistep mammary tumorigenesis in Wnt1 transgenic mice: The role of fibroblast growth factor-8 in oncogenesis and apoptosis

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content INFORMATION TO USERS This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter 6 c e, while others may be from any type o f computer printer. The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely afreet reproduction. In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back o f the book. Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6” x 9” black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order. UMI A Bell & Howell Infonnation Company 300 North Zeeb Road, Ann Arbor MI 48106-1346 USA 313/761-4700 800/521-0600 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ANALYSIS OF MULTiSTEP MAMMARY TUMORIGENESIS IN WNT1 TRANSGENIC MICE; THE ROLE OF FIBROBLAST GROWTH FACTOR 8 IN ONCOGENESIS AND APOPTOSIS by Deepa Bhavani Shankar A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (Molecular Microbiology and Immunology) December 1997 Copyright 1997 Deepa Bhavani Shankar Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UMI Number; 9835059 UMI Microform 9835059 Copyright 1998, by UMI Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code. UMI 300 North Zeeb Road Ann Arbor, MI 48103 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UNIVERSITY OF SOUTHERN CALIFORNIA THE g r a d u a t e SCHOOL UNIVERSITY PARK LOS ANGELES, CALIFORNIA 90007 This dissertation, written by Deepa Bhavani Shankar under the direction of hac Dissertation Committee, and approved by all its members, has been presented to and accepted by The Graduate School, in partial fulfillment of re quirements for the degree of DOCTOR OF PHILOSOPHY ,CV2....Z .. j . . . ........... , of Graduate Studies Date [^Gcember 16, 1997 DISSERTATION COMMITTEE f /!> ' • /7 /auiirperson Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. This dissertation is dedicated to ail the women who have lost their lives to breast cancer. I would also like to dedicate this dissertation to the memory o f my father, who planted the seeds of scientific inquiry into my mind, and to my beautiful mother, who nurtured my dreams and made this possible. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Acknowledgem ents There are several people who have supported me through these years and have helped in making this possible. I wish to express my gratitude to them. First of all I wish to thank my mentor Dr. Greg Shackleford for his guidance, support, patience and specially for believing in my abilities. I would like to thank my committee members Dr. Yuen Kai Fung, and Dr. Pradip Roy-Burman for their guidance and timely advise. I also wish to thank all my teachers in the Departments of Microbiology and Biochemistry, specially Dr. Peter Vogt for inspiring me to work on oncogenes. My sincere gratitude also goes to Craig MacArthur. who helped me get started on the project and for the collaboration on characterization of Fgf8 (the work presented in Chapter 2 and part of Chapter 3). I wish to thank Jianwei Wang for his excellent technical guidance and for passing on the trade secrets to me, and Anat Epstein for her constant encouragement, help and generous use of her computer. I would like to express my thanks to the members of Dr. Comer's and Dr. Fung's laboratories for the use of their lab equipment. To all my friends; Thank you so much, this would not have been possible without your support and friendship. My heartfelt thanks to Suzan Imren for her encouragement and understanding, Ann Kapoun for all the pep talks and interesting and enlightening conversations (scientific and otherwise), Rocio Lopez-Diego for her honesty, constructive criticism, Spanish omelet's and for teaching me the nuances of camping, Sam Wong for all the coffee, lunches and advise, Paige Heaphy for making sure I had a life, Junqing Qian for the delicious food. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. and interesting conversations, and Mahalakshmi Krishnan for being a true friend. I wish to thank my mother, Raji B. Shankar for her love and constant support, my sister Nimi Kumar and brother in law Nirma! Kumar for putting up with me for all these years, caring for me like parents, and most importantly for sharing Ashik with me, Sanjay for all his warmth and understanding, and my little genius Ashik for being such a bundle of joy, and for making the stress of finishing up bearable. Last but not the least I wish to thank the breast cancer advocates for making breast cancer research an important issue, and the Department of Defense for funding this research. IV Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table of Contents PAGE Dedication il Acknowledgments iü Table of contents v List of Abbreviations vii List of Figures viii List of Tables x Abstract xi Chapter 1. Introduction 1 Chapter 2. Fgf8 activated by proviral Insertion, cooperates with the Wnt1 transgene in murine mammary tumorigenesis 3 4 Chapter 3. Demonstration of the oncogenic ability of Fgf8 6 4 Chapter 4. Apoptosis of mammary epithelial cells induced by fibroblast growth factors 9 5 Chapter 5. Long distance activation of int2/Fgf3 by proviral insertions in a putative common insertion locus for MMTV 131 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter 6. Summary and conclusions I 39 Epilogue ^44 References 145 VI Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. List of Abbreviations APC DCC EGF EGFR FGF Fgf8 FGFR LTR MMTV PDGF IGF WAP Adenosis polyposis coii Deleted in colon cancer Epidermal growth factor Epidermal growth factor receptor Fibroblast growth factor Fibroblast growth factor 8 Fibroblast growth factor receptor Long terminal repeat Mouse mammary tumor virus Flatlet derived growth factor Transforming growth factor Whey acidic protein VII Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. LIST OF FIGURES PAGE Chapter I Figure 1.1. Genetic changes associated with colorectal tumorigenesis. 8 Figure 1.2. The mechanisms of insertional activation of proto-oncogenes. 23 Figure 1.3. Acceleration of mammary tumorigenesis in MMTV infected Wnti transgenic mice. 31 Chapter II Figure 2.1. Schematic representation of exon trapping. 38 Figure 2.2. MMTV proviruses in tumors of Wnti transgenic mice. 44 Figure 2.3. MMTV proviral insertions into a new common integration locus. 47 Figure 2.4. Map of proviral insertions in the new common integration locus. 50 Figure 2.5. Transcriptional activation of Fgf8 by MMTV insertion in tumors from infected Wnti transgenic mice. 52 Figure 2.6. FgfS gene structure and protein sequences. 55 Figure 2.7. FgfS is normally expressed during embryogenesis and in adult ovaries and testes. 58 Chapter ill Figure 3.1. Structure of the FgfS protein isoforms. 66 Figure 3.2. Schematic representation of the expression vector pMIRB 73 Figure 3.3. NIH3T3 cells transfected with different FgfS isoform cDNAs display different morphologies. 74 Figure 3.4. Focus formation by M\RB-FgfSb on a monolayer of NIH3T3 cells. 76 VIII Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 3.5. Comparison of the deduced amino acid sequences of the three human FGF8 cDNA isoforms with those of mouse FgfS isoforms. 80 Figure 3.6. Morphological transformation of NIH3T3 cells by human FGFS isoforms. 83 Figure 3.7. Structure of the MMTV-Fg/8 transgene 84 Figure 3.8 . Expression of the FgfS transgene in mammary tumors and some adult mouse tissues 87 Figure 3.9 . Histopathology of the MMTV-Fg/8 mammary tumors. 88 Figure 3.10 Pathology of the ovaries from MMTV-Fg/8 transgenic mice. 90 Chapter IV Figure 4.1. Schematic representation of conditioned medium transfer. 101 Figure 4.2. Heparin-sulphate sepharose chromatography. 103 Figure 4.3. Overexpression of FgfSb in mammary epithelial cells (C57MG) results in apoptosis. 107 Figure 4.4. C57MG mammary epithelial cells treated with FgfSb containing conditioned medium undergo programmed cell death. 110 Figure 4.5. Binding of the FgfSb protein from the conditioned medium to heparin-sepharose affinity column abolishes the induction of apoptosis. 115 Figure 4.6. Neutralization of FGF2 induced apoptosis of mammary epithelial cells by an anti-human FGF2 antibody 119 Figure 4.7. Ectopic expression of BCL2 in mammary epithelial cells delays the onset of FGF induced apoptosis. 124 Chapter V Figure 5.1. Map of tumor 76 locus. 137 IX Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. List of Tables PAGE Chapter 1 Table 1.1. Oncogenes and Suppressor genes in breast cancer 5 Table 1.2. Summary of transgenic mice expressing oncogenes 16 implicated in breast cancer Chapter 3 Table 3.1. Properties of NIH3T3 cell lines expressing 78 FgfS isoform cDNAs Chapter 4 Table 4.1. Apoptosis of mammary epithelial cells induced by 117 members of the FGF family Table 4.2. FGFs induce apoptosis of several mammary epithelial cell lines but not of fibroblasts 121 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Deepa Bhavani Shankar Gregory M. Shackleford Analysis of Multistep Mammary Tumorigenesis in Wnti Transgenic Mice: The Role of Fibroblast Growth Factor 8 in Oncogenesis and Apoptosis. Retroviral Insertional mutagenesis of W nti transgenic mice by mouse mammary tumor virus (MMTV) was used as a model to study proto-oncogenes cooperate in multistep mammary tumorigenesis. By examination of the genes that are insertionally mutated in a clonal fashion by MMTV in tumors from infected mice, a new Wnti cooperating proto-oncogene was identified as fibroblast growth factor 8 {FgfS). Transcriptional activation of Fgf8 was observed in tumors with integrations near the gene . Further characterization of Fgf8 showed that normal expression was detectable only in midgestational embryos, and in adult in testis and ovary. Five coding exons were found in FgfS, and cDNAs encoding three different protein isoforms were isolated. Characterization of the oncogenic potentials of the three FgfS isoforms (FgfSa, 8b and 8c) in NIH3T3 cell transformation assays showed the isoform Fgf8b to be the most powerful transforming isoform; Fgf8a and 8c showed moderate transforming activity. Expression of FgfS in the mammary gland of transgenic mice resulted in generalized mammary gland hyperplasia followed by development of mammary tumors ranging from adenomas to adenocarcinomas. These results identify a third Fgf gene as a W nti collaborator in mammary tumorigenesis, and suggest that theH/hf and Fgf gene families cooperate effectively in this process. XI Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Analysis of the biological properties of Fgf8 isoforms in mammary epithelial cell lines in culture showed that FgfSb induces a transient morphological transformation and programmed cell death (apoptosis) in these cells. Several other members of the fibroblast growth factor family also induced apoptosis of mammary epithelial cells. Further characterization of the FGF-induced apoptosis pathway showed that overexpression of the anti-apoptotic gene BCL2 in these cells delayed the onset of apoptosis, suggesting that proteins of the BCL2 family may participate in FGF-induced apoptosis. These results describe a new property for FGFs and suggest that these growth factors may play important roles in regulating cell growth and death in development and cancer. XII Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter 1 Introduction Breast Cancer - A major public health problem The statistics of breast cancer indicate that it is the most common cancer among women and the second largest cause of death as a result of malignancy in women (Rich, 1997). During their lifetime, 12% of all women in the U.S. will be diagnosed as having breast cancer, and 3.5% will die of this disease (Rich, 1997). In 1997, about 184,300 women will be diagnosed with breast cancer and an estimated 44,300 women will die of breast cancer this year. Despite intense research and improvements in the overall treatment of cancer in the past thirty years, the mortality rate from breast cancer has been stable and without decline since 1950 (Rich, 1997). Part of the reason for the lack of decline in breast cancer mortality may be that the incidence of breast cancer has been steadily increasing over the last fifty years. At present, the lifetime risk of getting breast cancer for a woman living in the United States is 1 in 8. which is double the risk in 1940 (Rich, 1997). The reasons for this steady increase in the incidence of breast cancer are not known. These numbers demonstrate the magnitude of the problem. In recognition of the importance and magnitude of the problem posed by breast cancer and the need to develop effective solutions and cures for this widespread disease, breast cancer research funding has increased substantially during recent years. Breast cancer, like other malignancies, is thought to emerge as a result of a combination of genetic background and cumulative Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. environmental effects. The factor which most clearly puts a woman at high risk for developing breast cancer is a family history of pre menopausal bilateral breast cancer or a first degree relative with breast cancer. Other factors which increase her risk are old age, a history of cancer in one breast, fibrocystic disease of the breasts, and history of primary cancer in an ovary or the endometrium. The hormonal milieu of the woman as determined by age at first pregnancy, ovarian hormonal function, obesity, and age at menarche and menopause also affect the risk of breast cancer. Certain environmental factors such as socioeconomic class, country of birth, marital status, place of residence, and race are also associated with an increased risk for developing breast cancer (Weber and Garber, 1996). Biological Factors in Breast cancer Breast cancer, like all cancers is inherently a genetic disease. The principal genetic lesions observed Include gene amplifications, deletions, point mutations, loss of heterozygosity (LOH) and overall aneupioidy. The first observation that breast cancer was a familial disease was made by the ancient romans. In a few breast cancer prone families, women who carry the breast cancer susceptibility gene have an 85% risk of developing breast cancer during their lifetime (Weber and Garber, 1996). Much of the research in the area of breast cancer is aimed at uncovering the genes that predispose a woman to this disease and trying to understand how these genetic changes act at the molecular level to confer a high susceptibility to breast cancer. Many of the identified genetic alterations associated with cancer result in activation of Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. proto-oncogenes or inactivation of tumor suppressor genes. Even though several non-genetic mechanisms have been proposed for activation of oncogenes and inactivation of tumor suppressor genes, these are considered not of primary importance (Siamon et al., 1987). The normal functions of proto-oncogenes and tumor suppressor genes are positive and negative regulation of cell proliferation and cell death. Malignancy is the result of disruptions in the mechanisms that normally regulate the growth, division, and death of cells. Neoplastic transformation occurs when cells that should be quiescent divide; cells that should acquire special function remain immature: or cells that should undergo a preprogrammed cell death continue to live. The resulting uncontrolled multiplication of cells eventually leads to the growth of tumors which then invade adjacent tissues and metastasize to distant sites. Most oncogenes in breast cancer have been identified by detailed study of amplified and translocated chromosomal regions using cytogenetics. Point mutations are not a common mechanism for oncogenic activation in breast cancer. In contrast, tumor suppressor genes are characterized primarily by inactivating point mutations and LOH (Trent-J.M., 1985). A preview of the complexity of somatic alterations which occur during breast tumor progression was provided by earlier cytogenetic analysis of primary human breast tumor cells in culture (Callahan. 1996; Trent-J.M., 1985) The alterations include aneupioidy and rearrangements affecting chromosomes 1 q translocations), 6q (deletions and translocations), 7p (translocations, pericentric inversions, and iso chromosomes), and 11q (translocations). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Using comparative genome hybridization techniques to map regions of the genome with increased DNA sequence copy number (amplifications), 26 chromosomal subregions were found to be affected in primary human breast tumors and breast tumor cell lines (Kallioniemi, 1994). Genetic analysis of primary breast tumor DNAs has demonstrated amplification of known proto-oncogenes such as FLG (8p12), MYC{8q24), BEK/FGFR2 (10q24), 11q13 {FGF3, PRAD1/CYCUN D1, FGF4), 15q24-q25 {IGFR- 1/FES), and ERBB2 (17q12). However, it seems likely that amplification of other potential proto-oncogenes also contribute to malignant progression in breast cancer (Callahan, 1996; Kallioniemi, 1994) Loss of heterozygosity (LOH) represents another common genetic alteration in primary human breast tumors and occurs as a consequence of either interstitial deletions, chromosome loss, or aberrant mitotic recombinational events (Hollingsworth and Lee, 1991; Knudson, 1989). At least eleven chromosome arms have been frequently affected by LOH in breast tumors, including chromosomes Ip, 1q, 3p, 6q, 7q, lip , 13q, 16q, 17p, 17q, and 18q (Bishop, 1991; Hollingsworth and Lee, 1991 ; Kallioniemi, 1994; Lane and Benchimol, 1990). A major focus of research at present is to identify the target tumor suppressor genes affected by LOH or the oncogenes affected by DNA amplification. Many of the oncogenes implicated in breast cancer belong to several well studied families of cell surface growth factor receptors and their ligands such as the epidermal growth factor (EGF) receptor superfamily, the insulin-like growth factor (IGF-1) family and the fibroblast growth factor (FGF) family (Dickson, 1996). Other mutated genes which are Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table L I Qncogeaes.and Suppressor Genes in Breast Cancer ONSET OF FAMILIAL BREAST CANCER Suppressor gene (mutated or lost^ p53 (Li-Fraumeni syndrome^ Likely suppressor genes (mutated or lost) BRCAl (breast and ovarian cancer families) BRCA2 (breast cancer families) PROGRESSION OF BREAST CANCER Oncogenes (amplified) erbB-2 myc cyclin Suppressor genes (mutated or lost) cyclin E (?) Rb-l p53 Oncogene candidates ER PR EGFR C'Ras^ Suppressor gene candidates MTS-I nm23 integrin E-cadherin Brush I Adaptedfrom Diseases o f the Breast, (Lippman et a i, 1996). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. potentially involved in developing an invasive metastatic phenotype include: p53, a transcriptional activator with tumor suppressor properties and nm23, a possible metastasis suppressor (Davidson, 1989; Dickson, 1996). Levels of expression of some of these proteins in breast carcinoma have been shown to have prognostic significance, thereby influencing the decision of which type of treatment a patient will receive (Davidson, 1989; Dickson, 1996) (Table 1.1). The recently identified tumor suppressor genes BRCA1, and 2 are currently being investigated in several laboratories. BRC A1 is thought to be involved in genomic stability because of its interaction with RAD51 (Kinzler and Vogelstein, 1997; Scully et a!., 1997; Scully eta!., 1997). Phase-1 gene therapy trials with BRCAl seems to indicate that expression of a normal BRCA 1 allele in the ovaries by retroviral transduction has some therapeutic effect In ovarian cancers (Giannios and loannidou-Mouzaka, 1997; Katso et ai, 1997). Currently several different approaches are being employed to identify the genetic alterations that are involved in breast cancers, and one such approach is the use of animal models for the analysis of oncogenic cooperation in multistep tumorigenesis. Multistep tumorigenesis It is now generally accepted that cancer is a multistep process consisting of accumulation of different genetic alterations over a period of time. Statistical analysis of human cancer incidence to age indicate that multiple events are required and suggest that three to six steps are necessary for the generation of a tumor in a adult (Peto et ai, 1975). Molecular analysis of human cancers show that they typically arise from Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. a single cell (l;e., they are clonal), and that each cancer cell shows multiple genetic lesions, that include chromosomal translocations, gene amplifications, point mutations, transcriptional activation of normally silent genes, or inactivation of tumor suppressor genes. Evidence from experiments with primary rodent cells demonstrates that expression of two or more oncogenes is required for tumorigenic transformation (Hunter, 1991). These types of experiments led to the concept of oncogenic cooperation where, cooperation is defined as a situation in which a pair of oncogenes acting in concert can convert a single cell taken directly from the animal into an established tumor cell line, whereas neither oncogene alone has this capacity (Hunter, 1991). Over the last decade, much emphasis has been made towards understanding why multiple events are necessary for carcinogenesis and how the expression of oncogenes and loss of tumor suppressor genes cooperate in this process. One of the best studied examples of multistep tumorigenesis in human cancers is that of colorectal carcinomas, where it is possible to study a progressive series of disease states ranging from hyperplasia to metastatic carcinoma (Fearon and Vogelstein, 1990; Kinzler and Vogelstein, 1996). The events observed in colon carcinoma begins with a hyperplastic adenoma involving an allelic loss or mutation of the familial adenomatous polyposis gene (FAR), an activating mutation of the K-Ras gene, allelic loss of the tumor suppressor gene DCC, and the allelic loss of the p53 gene (Figure 1.1). In many instances, one p53 has been deleted, and the other is inactivated by point mutations (Lane and Benchimol, 1990; Nigro et al., 1989). On average these events occur in Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CD ■ D O Q . C 8 Q . ■ a CD < / > ( / ) 3 CD C 3 . = r CD 3 ■ o I c a O 3 ■ o O APC K-RAS DCC/DPG4/JV18 p53 Other changes? Normal Epithelium Dÿspïastu; • AG E • nkrmeqMk Adenoma Ca'rdiiiom'a Adenoma Adenaiiià & o c ■ o CD Figure 1.1 Genetic changes associated with colorectal tumorigenesis. A schematic representation of the multiple genetic events that occur during the development of colon cancer. (adapted from Kinzler and Vogelstein, 1996) ( / ) o' 3 00 the listed order, but this not an obligatory chain of events during colon carcinoma progression, rather it appears that It is the total accretion that is important (Fearon and Vogelstein, 1990). The number of detectable events increases in parallel with the stage of disease. Therefore, to understand the development and progression of cancer, it is imperative to identify not only the single mutations involved, but also synergistically acting groups of cancer genes. Animal model systems provide further support for the multistep nature of carcinogenesis. For example, analysis of carcinogen induced tumors has shown that carcinogens cause mutations in Ras genes and other proto-oncogenes; however, the initiating mutation is insufficient to cause carcinogenesis and one or more secondary events to convert the mutated cell into a tumor cell are required (Sukumar, 1990). Thus, in the rat mammary carcinoma model activated H-Ras is detected at very high frequencies, the initiating H-Ras mutation occurs within the first few hours of carcinogen administration, but subsequent tumor development is absolutely dependent on estrogen mediated secondary events (Kumar et al., 1990). The mouse skin carcinoma model is also an example for the multievent process of carcinogenesis (Balmain and Brown, 1988). The results from this experiment led to the two stage model of initiation and promotion, which is based on the observation that the application of a low non carcinogenic dose of an initiating carcinogen to mouse skin does not cause carcinoma unless succeeded by multiple treatments with a tumor promoter that by itself has no effect (Yuspa and Poirier, 1988). Molecular analysis of this process shows that activation of H-Ras gene is an early event, but this is not sufficient to cause the mutated cell to Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. progress, further treatment with tumor promoters like phorbol esters Is necessary for progression to carcinoma (Balmain and Brown, 1988). it Is also clear from animal systems that mutational events activating oncogenes do not have to occur In a particular order. Rather than any specific order of events being required. It Is the summation of these events that Is cardinal In tumorlgenesls. Experimental evidence for oncogenic cooperation There are a number of experimental systems In which cooperation between oncogenes can be observed. An early Indication that oncogenes can collaborate was the finding that certain combinations of oncogenes could transform primary rodent fibroblasts or kidney cells In culture, whereas either one of the genes by themselves were dormant (Palmlerl, 1989). Experiments with combinations of polyoma virus oncogenes (Ruley, 1983), followed by similar studies with cellular oncogenes, were the first evidence for oncogenic cooperation (Land et al., 1983). The specific combinations of oncogenes that were able to cooperate led to the concept that oncogene products that act In the nucleus cooperate best with those that act In the cytoplasm (Hunter, 1991). However this concept does not hold true In all Instances. Examples of cooperation between two nuclear oncoproteins or two cytoplasmic oncogenes exist (Reed etal., 1990). Compelling evidence for oncogene cooperation has also come from studies using transgenic mice (Cory and Adams, 1988; Hanahan, 1989; Varmus, 1989). Transgenic mice harboring Individual oncogenes under tissue specific or ubiquitous promoters generally show enhanced 1 0 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. level of tumorigenesls, however, depending on the system, the latency for development of tumors is often prolonged (Compere et al., 1988; Cory and Adams, 1988; Hanahan. 1989). In contrast when two appropriate strains of mice transgenic for two different oncogenes are crossed together, the tumor incidence is accelerated with a drop in latency (Donehower etal., 1995; Kwan etal., 1992). These studies illustrate that the oncogenes expressed in these mice can cooperate when expressed in an Fi mouse; and that the tumors found in the strains carrying a single oncogene are likely to require a second or perhaps multiple additional events. Therefore, transgenic studies allow definitively to test candidate oncogenes and tumor suppressor genes, or to see if a mutant protein interferes with wild-type functions. The transgene, in general, does not directly promote tumors but establishes a high predisposition, and emergence of a malignant clone requires genetic changes in a rare affected cell (Bishop, 1991). The two rate limiting steps in normal tumorigenesis, the initial mutation and the expansion of that clone, are suspended in a transgenic animal by expansion of the transgene throughout a cell compartment, but synergistic mutations are still required. Additionally, during the preneoplastic phase, any direct effects of the transgene on proliferation or differentiation can be determined, and the rules of oncogene cooperation can be explored as described before, by delivery of a second gene (Cory and Adams, 1988). This can be achieved by crossing of two lines of transgenic mice, each carrying a different activated proto-oncogene. For example, acceleration of tumor formation observed in the in the Wnt1/Fgf3 double II Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. transgenic mice in contrast to the sporadic nature of tumor formation in either Wnt1 or Fgf3 mice alone (Kwan et al., 1992; Muller et al., 1990; Tsukamoto et al., 1988). Another method is to look for activations of known proto-oncogenes in tumors from transgenic mice expressing another oncogene, for example, activated ras genes were sought and found in B cell lymphomas of transgenic mice expressing c-myc in their B cells (Alexander et ai, 1989). Although these models provide systems in which to test known cloned oncogenes for collaborative potential, they do not allow the discovery of novel proto-oncogenes involved in multistep tumorigenesis. Recently, insertional activation of cellular proto-oncogenes by retroviruses that lack viral oncogenes has been shown to be a useful means of identifying cellular genes, known or novel, that cooperate in tumorigenesis (Nusse, 1991). The sequential infection of chickens with two avian leukosis viruses, first in ovo with one virus, followed by infection of the chick with a virus of a different subgroup, causes metastatic potential compared with animals infected with either virus alone (Clurman and Hayward, 1989). Molecular analysis of these tumors has led to the discovery of a new proto-oncogene, bic, which is correlated with metastatic potential. Because retroviruses such as murine leukemia virus (MLV), primarily promote tumorigenesis by integration near a cellular proto oncogene affecting expression, proviral insertion provides a screen for genes that cooperate with the transgene to induce tumors. Using this strategy, Berns et al. infected Pim1 transgenic mice, which are predisposed to T cell lymphomas with MLV. T cell lymphomas indeed 12 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. developed at an accelerated rate compared with uninfected (or infected non-transgenics) due to activation of c-myc or N-myc by MLV insertional mutagenesis (van Lohuizen et ai, 1989). Conversely, MLV infection of c- myc transgenic mice accelerated either Pim1 or a gene at one of four novel loci {bmi-1, pal-1, bla-1 or emi-1) discovered as clonal insertion sites in these lymphomas (van Lohuizen at ai, 1989; van Lohuizen at ai, 1991 ). Thus, in the latter experiments, a total of five loci, four of which are apparently novel, were found to cooperate with c-m yc in B cell lymphomagenesis using this strategy of retroviral insertional mutagenesis in transgenic mice. A similar strategy has been used to identify proto-oncogenes that cooperate in mammary tumorigenesis using mouse mammary tumor virus (MMTV) infection of W nti transgenic mice (Lee at a!., 1995; MacArthur etal., 1995; Shackleford at ai, 1993). Since this model system forms the basis of my thesis rationale and hypothesis, it will be described in another section. The study of DNA tumor viruses and their transforming proteins, also strongly supports the idea that cooperation is an essential feature of oncogenesis. For instance, the demonstration that both polyoma virus large T (LT) and middle T (MT) antigens were required for the transformations of primary rat embryo fibroblasts (Rassoulzadegan at ai, 1982) indicates oncogenic cooperation. Even though SV40 large T by itself is sufficient to transform cells, it does so by its interaction with a critical set of cellular proteins. SV40 large T binds and inactivates both p53 and Rb (Marshall, 1991; Scheffner at ai, 1990). In the case of adenovirus, both transforming proteins El A and E1B bind Rb and p53 13 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. respectively (Whyte et al., 1988). Thus it is apparent that oncogenic cooperation is a common feature of carcinogenesis and is the premise for the multistep process of any type of cancer. Transgenic mouse models for breast cancer The genes thought to be involved in the genesis of human breast cancer encode growth factors, growth factor receptors, nuclear transcription factors, guanidine nucleotide binding proteins, cell cycle regulatory protein, tumor suppressors, etc (Clarke, 1996). Transgenic animal models have proven to be as useful in studying breast cancer as other cancers, and have yielded valuable information regarding breast transformation Three main mammary gland specific promoters that have been widely used to target specific oncogenes to the mammary gland are the MMTV long terminal repeat (LTR), the whey acidic protein promoter (WAP) and the 3 lactoglobulin promoter (Amundadottir at at., 1996; Clarke, 1996). MMTV is active in the virgin mammary gland with augmented expression during pregnancies adn lactation. The WAP and 3-lactoglobulin promoter are from milk specific genes causing expression of the transgenes in later stages of pregnancy and during lactation. A summary of several single transgenic lines established to study breast carcinoma is listed in Table 1.2. Some of these genes have been implicated in human breast carcinogenesis. These include cmyc, neu and p ra d l/C y d in D I, each of which has been found amplified and overexpressed in many human tumors (Lammie and Peters, 1991; Mochotka et ai, 1989),. Genomic amplification accompanied by 1 4 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. overexpression of neu in human breast tumors has been correlated with decreased survival (Slamon et al„ 1987). The amplified genomic region containing prad1/CyclinD1 also includes linked genes Fgf3/Fgf4 that are involved in mouse mammary tumors (Laammie et al., 1991). Growth factors and their receptors. Transforming growth factor a {Tgf a) is a secreted growth factor that was originally detected in the culture media of certain retroviruses transformed firbroblasts. It is a member of the epidermal (EOF) growth factor family and binds to and activates the EGF receptor, the protein product of the proto-oncogene c-erbB (Massague, 1983). Tgfahas been found to be amplified in human breast cancer and its expression is frequently elevated. Numerous studies have shown that Tgfa mRNA and proteins are expressed in several human breast cancer cell lines and about 40-70% of primary human breast tumors, compared to 20% of normal/benign breast samples (Perroteau et al., 1986). Similarly, other members of this family of growth factors like amphiregulin and cripto-1 are preferentially expressed in human breast cancer cases compared to normal tissues (Amundadottir et a!., 1996). The EGFR itself is also expressed in about 30-50% of human breast cancers with an elevated expression associated with poor prognosis of the disease and a high degree of invasiveness (Amundadottir etal., 1996; Massague, 1983). Transgenic mice overexpressing Tgfa in the mammary gland develop generalized hyperplasia of the mammary gland followed by appearance of carcinomas with a long latency of 8-10 months (Jhappan et al., 1990; Sandgren et al., 1991). The tumors were stochastic and 15 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 1.2: Summary of transgenic mice expressing oncogenes implicated in breast cancer Transgene G row th F a cto rs receptors T g f a T g / u erbB2/neu neu N uclear Oncogenes cmyc cmvc Promoter a n d MT MMTV-LTR MMTV-LTR MMTV-LTR MMTV-LTR WAP V iral Oncogenes Polvoma middle MMTV-LTR T SV40 large T WAP Ras genes v-HcL-ras MMTV-LTR Ha-roj (activated) WAP In t genes WntI intl/FgfS int-3 (activated) G row th Suppressing genes p5i(mutated) Tgfp T g fp MMTV-LTR MMTV-LTR MMTV-LTR WAP WAP MMTV-LTR Cell genes Cvclin D1 phenotype Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Age of onset other aspects 7-14 months 7-9 months 5-10 months 3 months Adenocarcinoma Adenocarcinoma 7-14 months 4+ months Adenocarcinoma 5 weeks Adenocarcinoma 4-6 months Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma 4-10 months 10-12 months 4-7 months 10+ months 2-7 months cycle MMTV-LTR Adapted from (Amundadottir et al.. No tumors No tumors No tumors Adenocarcinoma 18 months 1996) métastasés synchronous tumor formation métastasés synchronous tumor formation métastasés frequent métastasés seen low frequency low frequency development^ arrest developmental arrest developmental arrest developmental arrest 16 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. arose predominantly in mice that had undergone multiple pregnancies (Sandgren etal., 1991). A number of studies have been performed using transgenic mice that overexpress erbB-2 or neu in the mouse mammary gland (Bouchard et ai., 1989). Female mice expressing erbB-2 developed tumors with a latency of 5-10 months and, in some cases, tumorigenesis was pregnancy dependent. The tumors metastasized with a high frequency, suggesting that elevated expression of neu can induce metastatic mammary gland tumors (Muller et al., 1988). cmyc Abnormalities in the cmyc proto-oncogene locus are strongly associated with a number of human malignancies. The gene is amplified in 25-30% of breast cancer cases and, in addition, is rearranged and overexpressed in many more cases. Furthermore, amplification of the cmyc gene has been shown to correlate with poor prognosis of the disease (Escot et al., 1986). MMTV-myc mice develop solitary mammary tumors at about 7-14 months of age after multiple pregnancies suggesting that additional genetic events are necessary for tumor onset (Stewart etal., 1984). In summary, myc overexpression in the mammary gland predisposes transgenic animals to mammary cancer, although this oncogene is not sufficient to induce tumors on its own. Ras: Augmented expression of Ras mRNA and proteins has been demonstrated in a very large number of human breast cancer cases 1 7 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Slamon et a!., 1984). Expression was not due to amplification, rearrangement or point mutations of ras genes (Theillet et al., 1986). A high frequency of rare ras alleles has been found in human breast cancer (Theillet et a!., 1986). Transgenic mice that overexpress ras in the mammary glands either from MMTV- LTR or WAP have been made. MMTV-ras female and male animals develop stochastic but sometimes multiple mammary tumors as early as five weeks of age. Metastasis were found in liver and lungs (Sinn et ai., 1987). Another study with MMTV-ras transgenic mice described similar results with mammary tumors arising in animals between 4-10 months of age. Thus, ectopic expression of ras, like most other oncogenes also predisposes to mammary cancer. Cell cvcle regulatory genes Cyclins were first identified because of their spectacular periodicity during cell divisions. Consistent with their roles in cell cycle progression, cyclins are also involved in many types of human cancers. The cyclins that are currently thought to be involved in human breast cancer are cyclin D and E. Cyclin E is amplified, its expression unregulated, and/or the stability of mRNA enhanced in many breast cancer cell lines (Keyomarsi et a!., 1994). The cyclin E protein is also overexpressed in human breast tumors (Keyomarsi etal., 1994). CyclinD I or prad1 is involved in moderating the mitogenic response to growth factors and hormones. About 15-20% of primary breast tumors show amplification of 11 band q13 on chromosome (Amundadottir etal., 1996). Of the known genes in this amplicon {int2, hsti, emsi and prad1), only p ra d i and ems1 are overexpressed in tumors with 11q13 amplification 1 8 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Amundadottir etal., 1996). Overexpression of CyclinDI mRNA is seen in about 45% of breast cancer samples. Transgenic mice with MMTV-LTR directing expression of CyclinDI to the mammary gland develop hyperplasias at young ages and adenocarcinomas after about 18 months. Multiple independent adenocarcinomas were observed in a number of transgenic mice (Wang et al., 1994). Like in all other cases, it is apparent that cyclinDI is involved but not sufficient to induce tumorigenesis when overexpressed in the mammary gland. Tumor suppressor genes The p53 gene is mutated in about 50% of breast cancer cases and its expression is associated with poor prognosis. Loss of normal p53 function is a very important genetic change in breast cancer (Amundadottir et al., 1996; Marshall, 1991). Mice deficient in one or both alleles of p53 develop lymphomas and sarcomas at very high frequencies, but mammary gland tumors occur infrequently (Donehower et al., 1995). Expression of mutated p53 in transgenic mice under the control of the WAP promoter caused an inhibition of both lobuloalveolar development and gene expression of milk proteins causing the inability of female transgenics to nurse their young. No mammary tumors were observed, but there appeared to be an increase in radiation-induced apoptosis in the gland, indicating that mutant p53 was behaving like the wild type protein (Donehower et al., 1995). Crossing of p53 deficient mice to Wnt1 transgenic animals showed that p53 was not required for Wnti mediated tumorigenesis, but that loss 19 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. of p53 function accelerates tumor formation in animals that express Wnt1 in their mammary glands (Donehower et al., 1995). These results indicate the cooperation between the oncogene Wnt1 and the tumor suppressor gene p53 in promoting tumorigenesis. The "/nf" oenes. The int genes were first discovered because of their frequent activations in tumors by insertion of MMTV. W nti was the first proto oncogene to be discovered in this way (Nusse and Varmus, 1982). Since then, molecular analysis of proviral integration sites in tumors has led to the discovery of a number of cellular proto-oncogenes - previously called int genes. An interesting feature of these is that they are all involved in short range signaling between cells; most of the mouse mammary oncogenes encode secreted molecules (Nusse, 1991). This includes several members FGF family and Wnt gene family (Shackleford et al., 1993). Another gene. int3, has homology to the Drosophila notch receptor (Robbins et al., 1992). Transgenic mice overexpressing Wnti or Fgf3 in their mammary glands, develop mammary gland hyperplasia and adenocarcinomas with a prolonged latency (Kwan et al., 1992; Tsukamoto et al., 1988). These results reiterate the need for secondary or additional events for carcinogenesis. Thus it is apparent from the various transgenic models for breast cancer described to date, that most of the implicated oncogenes by themselves are insufficient to induce frank neoplasia. Rather, they predispose the mammary gland for cancer development, and indicate 20 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. that other secondary events are necessary for progression to malignancy. These secondary events can be identified by 1 ) Crossing the singly transgenic animals and analyzing tumor latency in mice transgenic for both genes. However, this approach does not allow the detection of novel oncogenes or genes not implicated in breast cancer. 2) Retroviral infection of transgenic mice. In an attempt to identify novel or previously unidentified oncogenes, we have used retroviral insertional mutagenesis of Wnt1 transgenics by MMTV. MMTV and mammary tumors The mouse mammary tumor virus (MMTV) belongs to the group of slowly transforming retroviruses (B group), and induces mammary tumors in susceptible mice after a latency of one year (Nusse, 1991; Okumoto et al., 1990). Factors such as virulence of MMTV variants, histocompatibility genes of the host and the hormonal status of the animal influences the onset of mammary tumors and mammary tumor incidence (Michalides et al., 1983). The primary target for MMTV infection is the mammary gland. In the C3H mice, MMTV only transforms the alveolar cells, whereas in the GR strain, both the ductal and alveolar cells are transformed (Michalides et al., 1983). In addition, MMTV enhancers are also active in the salivary gland (Nusse, 1986; Nusse, 1991). MMTV is transmitted in two ways, the first horizontally, from the mother to the offspring, via the milk. Exogenous MMTV is transmitted in this way. The second method is genetically, via the germ cell. Several inbred strains of mice have some proviral DNA integrated in the germ 2 1 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. line. These endogenous MMTVs are transmitted to the offspring through the germ cells and hence are present in all mouse cells (Michalides et al., 1983). Both exogenous and endogenous MMTVs have been implicated in the development of mammary tumors. The mechanism of transformation by these two differently transmitted MMTVs may be identical. An obligate step in the replication of MMTV is the integration of the DNA copy of its proviral genome into the host cellular genome, an event that is potentially mutagenic for the host cell. Although there may be some favorable sites for integration into the chromatin, most evidence indicated that the integration process of retroviruses is essentially random and therefore every cellular gene is a potential target for mutagenesis by proviral insertion (Nusse, 1988; Peters, 1990). The insertion of proviral sequences could have different consequences for the cell depending on the target gene and where the integration occurs. Some integrations could have no effects, some could be deleterious, or some could impart a selective growth advantage to the cell. This could occur if a proto-oncogene is activated by the nearby insertion of a provirus. Activations usually occur by increased expression of the gene, either due to direct transcription from the inserted proviral promoter ("promoter insertion") (Figure 1.2) or due to activation of transcription from the target gene's promoter by the enhancer sequences in the MMTV LTR ("enhancer insertion"); activations may also occur by truncation of the gene product (Figure 1.2) (Gallahan etal., 1987; Nusse, 1986; Nusse, 1991; Robbins etal., 1992). This growth advantage leads to the clonal expansion of the cell, thus increasing the chances of subsequent proviral insertions leading to formation of a tumor. Because 22 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Promoter Insertion 3 ' L T R P r o m o t i o n 5‘ LTR 3'LTR R as 5 ' L T R P r o m o t i o n ( r e a d t h r o u g h a c t i v a t i o n ) 5' LTR 3D 3‘LTR us4t us U3% us "H Z R US Enhancer Insertion 3'LTR us % us 5‘ LTR US T U3 AAA S'LTR 3'LTR U3 R US % AAA Figure 1.2 The Mechanisms of Insertional Activation of Proto-oncogenes Arrows indicate the direction of transcription; The diagrams with AAA represent mature transcripts of the activated proto-oncogenes. Proto-oncogene □ Poly A signal ^ 23 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. retroviruses usually integrate rather close (-10 kilobases) to insertionaliy activated proto-oncogenes, examination of the cellular DMAs that flank clonal proviral insertions In tumor DNAs has led to the detection of activated proto-oncogenes. Over 40 different loci have been identified as common sites for proviral insertions in tumor cells (Peters, 1990). Only nine well characterized common insertion sites have been described for MMTV to date: intlA/VntI, int2/Fgf3, hst/Fgf4, Fgfd, int3/Notch4, intS, int6, Wnt3, and Wnt 10b (Clausse etal., 1993; Dickson, 1990; Dickson etal., 1984; Durgam and Tekmal, 1994; Gallahan and Callahan, 1987; Gallahan et al., 1987; Gray etal., 1986; Lee etal., 1995; MacArthur etal., 1995; Marchetti etal., 1995; Morris etal., 1990; Morris etal., 1991; Nusse etal., 1984; Nusse and Varmus, 1982; Peters et al., 1983; Peters et al., 1989; Robbins et al., 1992; Roelink et al., 1990). This could be due to the fact that MMTV has historically been difficult to work with. Until recently, it was very difficult to clone MMTV due to so-called poisonous sequence in its genome (Brookes et al., 1986). Shackleford at al. have circumvented this problem by constructing a fully biologically active MMTV provirus that is easily clonable in bacterial plasmids or phage vectors (Shackleford and Varmus, 1988). This new virus has made it possible to clone cellular sequences flanking the gag half of the proviral insertion. An interesting common property of the genes activated by MMTV is that they are all involved in short-range signaling between cells: most of the mouse mammary oncogenes encode secreted molecules or their receptors. They include several members of the Wnt family, and the fibroblast 24 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. growth factor (FGF) family (MacArthur et al., 1995; Shackleford etal., 1993) W ntgenes Wnt1 is one of the most common insertion site for MMTV (Nusse and Varmus, 1982). Since the discovery of Wnt1 and Wnt-3 as targets of MMTV, a number of Wnt genes have been cloned by homology from many different metazoan species (McMahon etal., 1992; Nusse and Varmus, 1992). In mice, and presumably in all vertebrates, there are as many as fourteen Wnt genes (Adamson et al., 1994; Christian et al., 1991; Christian etal., 1991; Christiansen etal., 1995; Gavin etal., 1990; Huguet et a!., 1994; Ku and Melton, 1993; LeJeune et al., 1995; Nusse et al., 1991; Nusse and Varmus, 1992; Roelink et al., 1990; Sidow, 1992; Wang and Shackleford), most of which are expressed in spatially and temporally restricted expression patterns during development (Gavin et ai., 1990). In non-vertebrate animals, Wnt genes are also found in families; in Drosophila four Wnt genes have been identified, among which wingless (wg) (Rijsewijk et al., 1987) is the best studied. The nematode C. elegans has three known Wnt homologs, including genes identified by a mutant phenotype (Nusse and Varmus, 1992; Shackleford etal., 1993). All Wnt genes encode proteins that have a signal sequence, a number of potential N-linked glycosylation sites and some 23 or 24 cysteine residues, most of which are spatially conserved between Wnt family members. Most Wnt proteins are about 350 to 380 amino acids long and show string sequence conservation over the entire length of the 25 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. protein. Usually, when overexpressed in tissue culture cells, W nt proteins are poorly secreted, with most of the protein accumulating in the endoplasmic reticulum or Golgi compartments (Papkoff and Schryver, 1990). Extracellular Wnt1 is present as a monomer of 42-44 kD and associated with compartments of the cell surface or the extracellular matrix. Extracellular forms of Wnt1 can be detected only when cells are grown in the presence of positively charged molecules such as heparin or suramin, which are thought to release Wnt1 from the cell surface or the extracellular matrix. Many other Wnt proteins exhibit the same characteristics when expressed in v/fro (Bradley and Brown, 1990). A limited number of mammalian cell lines, including mammary epithelial cells and PCI 2 cells, show changes in cell morphology upon overexpression of Wnti (Brown et al., 1986; Rijsewijk et al., 1987; Shackleford et al., 1993). These morphological changes are also seen as a paracrine effect in co-cultures. In these experiments, fibroblasts expressing Wnti can cause morphological transformation in neighboring Wnti responsive mammary epithelial cells, suggesting that Wnt proteins signal to neighboring cells (Nusse and Varmus, 1992). IVnf proteins may not have to be secreted in a cell free form to signal in a paracrine way, as forms of W nti tethered to the cell membrane display a variety of biological effects also seen with wild-type forms (Nusse and Varmus, 1992). In general, UVnf proteins act as secreted ligands, but presumably work as short-range signaling molecules, being sequestered by nearby cells or the extracellular matrix. Normal expression of the Wnt genes can be found in almost any adult tissue, although any single Wnt gene usually has a restricted 26 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. pattern of expression. W nt genes are also expressed during embryogenesis where they may be found expressed in the central nervous system or in developing limbs (Gavin etal., 1990; Shackleford and Varmus, 1987; Wilkinson etal., 1987). Several H/nf genes are known to be expressed in normal murine and breast tissue (Huguet etal., 1994; Weber-Hall et al., 1994), and studies of human breast tumors have shown that at least four Wnt genes are associated with human breast cancer (Huguet etal., 1994; lozzo etal., 1995; LeJeune et al., 1995). The identification of Wnt1 as a proto-oncogene was based on activation by MMTV (Nusse and Varmus, 1982), and transformation of normal mammary epithelial cells; C57MG (Brown etal., 1986; Rijsewijk et al., 1987). To prove the oncogenic potential of Wnt1, transgenic mice containing the Wnt1 gene under the control of an MMTV enhancer were generated. Both male and female transgenics developed mammary adenocarcinomas following a generalized mammary hyperplasia (Tsukamoto et al., 1988). The median latency of mammary tumor formation, in female mice was ~ 5 months. Males developed tumors less frequently and later in life. These data, the generalized hyperplasia coupled with the long latency and the sporadic nature of the tumor formation, suggest that Wnt1 contributes to but is not sufficient for mammary tumorigenesis in these mice. Activation of Wnti is probably an early event in the process of tumor formation. Therefore other events, presumably genetic, are necessary for tumor progression. 27 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. FGF genes The other major group of proto-oncogenes activated by MMTV In mammary tumors belong to the FGF family of growth factors; these include int2/Fgf3 (Peters etal., 1983), hst/Fgf4 (Peters at a!., 1989), FgfS (MacArthur et a/., 1995) and the receptor FGFR2. (unpublished). The fibroblast growth factors are a family of at least fifteen structurally related polypeptides (Aaronson et al., 1991; Abraham et al., 1986; de Lapeyriere et ai, 1990; Gospodarowicz et ai, 1987; Hébert et ai, 1990; Kelley et ai, 1992; Klagsbrun, 1990; Liscia et ai, 1989; MacArthur et ai, 1995; Yoshida et ai, 1988; Zhan et ai, 1988). Members of the FGF family vary in length, but are homologous to one another within a core of 120 amino acid residues (Basilico and Moscatelli, 1992). These FGFs have been implicated in diverse biological processes ranging from mesoderm induction to tumor formation (Doniach, 1995; Gospodarowicz et ai, 1987; Mason, 1994). Most of the members are secreted and exert their function through autocrine and paracrine mechanisms (Basilico and Moscatelli, 1992). The signals from these molecules are transduced through a family of transmembrane receptor tyrosine kinases, FGF receptors (FGFRs 1-4) (Ornitz et ai, 1996). In addition they are characterized by their affinity to bind glycosaminoglycans and cell surface proteoglycans. This interaction seems to be obligatory for FGFR activation (Mason, 1994; Yayon et ai, 1991). In vivo and in vitro studies show FGFs to modulate a variety of developmental processes. They are mitogens for cells of all three germ layers and have been implicated in mesoderm induction, vascularization/angiogenesis and development of muscle, teeth and limbs (Basilico and Moscatelli, 1992; Mason, 1994). They also induce 28 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. differentiation and stimulate neurite outgrowth (Rydel and Greene, 1987). Some of the FGFs have important roles in hair production, and epithelial branching morphogenesis (Montesano etal., 1991). FGFs are also involved in pathological processes such as wound healing and cancer. Several FGFs were initially identified and characterized as oncogenes {Fgf3, 4, 5, 6 and8) (Delli Bovi etal., 1987; MacArthur etal., 1995; Peters etal., 1989; Zhan etal., 1988), and most of them show transforming activity on NIH3T3 cells. Mouse Fgf3 and Fgf4 were initially identified as frequent targets of MMTV activations in murine carcinogenesis. Transgenic mice that carry the Fgf3 gene driven by the MMTV enhancer have been made. Expression of the Fgf3 gene in female transgenics results in pronounced mammary gland hyperplasia and infrequent development of mammary adenocarcinomas (Muller et al., 1990), indicating the need for additional genetic events for neoplastic growth. Fgf4/hst is the second most common gene to be mutated in the NIH3T3 transformation assay (Basilico and Moscatelli, 1992). Fgf3 and FgfS are not normally expressed in the mammary gland and their activation from a silent state seems to contribute to tumorigenesis. Fgf4 has been shown to be expressed during the ductal stage of mammary gland development(Coleman-Krnacik and Rosen, 1994). FgfS (also called androgen induced growth factor) was originally cloned from a murine mammary tumor cell line SC-3, in which it is androgen inducible (Tanaka ef al., 1992). FgfS cooperates with the oncogene Wnt1 in causing mammary tumors by transcriptional activation of the gene by MMTV from a silent state (MacArthur et al., 1995). Thus, 29 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. there seems to be a stong collaboration between the Wnt, and FGF family of growth factors. In sertio n al m u ta g e n e s is of Wnt1 tra n s g e n ic m ice, identification of cooperating oncogenes The kinetics of tumor formation in the Wnti transgenic mice clearly indicate that additional genetic events are necessary for tumorigenesis (Tsukamoto et al., 1988). In an attempt to identify genes that may be involved in the multistep tumorigenesis, we mutagenized these Wnt1 transgenic mice by infecting them with MMTV (Shackleford et al., 1993; Shackleford and Varmus, 1988). The strategy was that since MMTV transcriptionally activates proto-oncogenes by insertion of its own DNA near them (Nusse, 1991), one could possibly identify additional oncogenes that oncogenlcally cooperate with Wnt1. The advantage of this approach over other mutagenesis procedures is that tumors arising due to proviral insertions contain proviruses physically linked to the activated proto-oncogenes forming a molecular tag which permits easy identification and cloning of the activated genes (Nusse, 1991). Activation of the cooperating oncogene would therefore confer a growth advantage and would presumably produce a tumor composed mainly of cells that are clonally derived from the cell bearing the proviral insertion. Implicit in this hypothesis was the expectation of a reduction in tumor latency. As predicted, in MMTV infected Wnt1 transgenics, the median latency of tumor formation decreased from ~5 months to 2.5 months and the number of tumors per mouse increased (Shackleford at ai, 1993) (Figure 1.3). 30 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Breeding Females Norm (10) 80 - MMTV-Norm (40) u_ Wnt-1 (50) MMTV- Wnt-1 (39) 20 - 0 2 4 6 8 1 0 1 2 Age (mo) Figure 1.3. Acceleration of mammary tumorigenesis in MMTV infected W n ti transgenic mice. The percentage of mice in each group remaining free of palpable mammary tumors was plotted as a function of age. MMTV-infected animals received virus at 3-4 weeks of age. Abbreviations: Wnt-1, W nti transgenics; MMTV-Wnt-1, MMTV infected Wnti transgenics; norm, normal nontransgenics; MMTV-norm, MMTV infected normal nontransgenics. The number of animals in each group are shown in parentheses. 31 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Analysis of the tumor DNAs derived from infected Wnt1 transgenic mice by showed that at least 75 of 128 tumors (59%) contained clonal MMTV-specific proviruses (Shackleford et al., 1993). These tumors were examined for the insertional activation of proto-oncogenes known to be activated by MMTV: int2/Fgf3, hst/Fgf4, int3 and Wnt3. Approximately 45% of these tumors contained insertionaliy activated int2 and/or hst (Shackleford etal., 1993), indicating that these genes cooperate with Wnt1 in mammary tumorigenesis. Despite the detection of clonal tumor specific proviruses in the rest of the tumors (55%), activation of int-2, hst, int-3 or Wnt-3 was not detected (Shackleford etal., 1993). Therefore, these tumors (55%) are likely to have activations of other novel or unpredicted proto-oncogenes, which cooperate with Wnti in accelerating tumorigenesis. Thesis rationale and experimental hypothesis. The model system described above presents a straightforward strategy to identify oncogenes that participate at successive stages of multistep mammary tumorigenesis. The main hypothesis of this thesis project is that MMTV acts as an insertional mutagen to activate cooperating proto oncogenes in these animals which are predisposed to mammary tumors because of the overexpressed Wnti transgene. By analyzing the tumors derived from these animals for newly integrated proviruses and cloning and characterization of the genes activated by these insertions, both novel and unexpected oncogenes (that cooperate with Wnti) can be identified and their role in mammary tumorigenesis can be characterized. 32 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The specific goals of the thesis project were to; 1 ) Isolate and identify proto-oncogenes (novel and unexpected) insertionaliy activated by MMTV in tumors of infected Wnt1 transgenic mice. 2) Characterize, and analyze the oncogenic potential of the identified proto-oncogene(s). 33 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter 2 Fgf8, Activated By Proviral Insertion, Cooperates With The W nti Transgene In M urine M am m ary Tumorigenesis Introduction Ectopic expression of the Wnt1 proto-oncogene In the mammary glands of transgenic mice causes a diffuse mammary gland hyperplasia, followed by the stochastic development of adenocarcinomas In both males and females after a minimum latency of several months (Tsukamoto et al., 1988). These findings Indicate that activation of Wnt1 Is but one step in the multistep process of mammary tumorigenesis. Early evidence that fibroblast growth factors (FGFs) may be Involved In cooperating steps comes from the finding that mouse mammary tumor virus (MMTV) Insertionaliy activates both Wnt1 and Fgf3 In some tumors of Infected nontransgenic mice (Peters etal., 1986). The cooperation of Wnt1 and Fgf3 was confirmed with the generation of bitransgenic Wnt1/Fgf3 mice, which developed tumors earlier than either of the monotransgenic mice (Kwan et aL, 1992). Multistep murine mammary tumorigenesis was studied by Infection otW ntI transgenic mice with MMTV, expecting that Infection would hasten tumor formation In these mice by Insertional activation of proto oncogenes that oncogenlcally cooperate with Wnt1 (Shackleford et al., 1993). As a result, activated proto-oncogenes are "tagged" by the nearby provlrus, thus allowing their identification or isolation. MMTV infection of 34 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Wnt1 transgenic mice results in accelerated mammary tumorigenesis and increased numbers of tumors per animal (Shackleford et a!., 1993). Approximately 45% of the resulting tumors with clonal, tumor-specific MMTV integrations contain insertionally activated Fgf3 or Fgf4, suggesting that the Wnt1 gene product can cooperate with at least two members of the FGF family (Shackleford at al., 1993). Despite being able to detect clonal, tumor-specific MMTV proviruses in the remaining 55% of the tumors, insertional activations of previously identified targets for MMTV insertion mutations (Fgf3, Fgf4, int3, or Wnt3) were not detected in this group (Shackleford at al., 1993). To determine the identity of proto-oncogenes that may be activated in these tumors, the DNAs surrounding proviral integration sites were analyzed. In this chapter, I describe the cloning of a common integration locus for MMTV from these tumors. Materials and Methods Tumor Samples. All mammary tumors were derived from a previous study (Shackleford at a!., 1993) in which female Wnti transgenic mice (Tsukamoto at al., 1988) at 3-4 weeks of age were infected with MMTV produced from EH-2 cells (Shackleford and Varmus, 1988). Preparation of nucleic acids. Tumor DNAs were isolated as described (Shackleford and Varmus, 1988), except that serum separation tubes were used in the extractions (Thomas at al., 1989). Southern and northern blot analyses. DNAs (10 fig) were digested with restriction endonucleases, electrophoresed, and capillary 35 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. blotted to nylon membranes (Amersham Hybond-N) as previously described (Shackleford and Varmus, 1988). The blots were UV- crosslinked ((Church and Gilbert, 1984)), prehybridized for 1-2 hours at 65°C, and hybridized at 65°C with 32p.labeled DNA probes. The hybridization buffer was 0.5 M sodium phosphate (pH 7.2), 1 mM EDTA, 1% bovine serum albumin (Fraction V), 7% sodium dodecyl sulfate (SDS), and 15% (v/v) formamide. Blots were washed in 40 mM sodium phosphate (pH 7.2), 1% SDS, and 1 mM EDTA at 65°C. Blots were exposed to Kodak XAR-5 film with intensifying screens at -80°C. For northern blots, RNAs were resuspended in 50 mM N-(2- hydroxyethyl)piperazine-N'-(2-ethanesulfonic acid) (HEPES), pH 7.0, 10 mM sodium acetate, 1 mM EDTA, 0.25 mg/mL ethidium bromide, 0.66 M formaldehyde and 50% (v/v) formamide, denatured at 65°C for 5 minutes, and electrophoresed in 0.8% agarose gels with the same running buffer minus ethidium bromide and formamide at 30-45 V for 6-14 hours (Shackleford and Varmus, 1987). Following photography, the gel was capillary blotted overnight to nylon. Crosslinking, prehybridization, hybridization, washing, and exposure were the same as for DNA, except that the hybridization buffer contained 30% (v/v) formamide for northern blots. Blots were stripped of probes by incubation In boiling 0.1% SDS for 3-5 minutes, and were rehybridized as above. Preparation and screening of genom ic libraries. Following identification of a proviral-cellular junction fragment by Southern blotting, the fragment was size-selected by agarose electrophoresis and purified either by glass beads (Gene Clean II, Bio 101) or by agarose digestion (p-Agarase, New England Biolabs) and 36 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. isopropanol precipitation. The junction fragments were ligated into appropriate lambda phage vectors, depending on the size of the insert (A.Dash II for inserts 10-20 kb, and XZap II for inserts less than 10 kb, Stratagene). Phage were prepared using packaging extracts (Gigapack II Plus, Stratagene), then titered and screened on appropriate E. Coli hosts (XL-1 Blue for XZap libraries and P2392 for IDash libraries). Plaque lifts were performed as described (Sambrook eta!., 1989), and UV-crosslinked, hybridized, washed, and exposed as above for Southern blots. Putative positive clones were subjected to 2-3 rounds of screening to isolate clones. The inserts were obtained in Bluescript (Stratagene) plasmids, either by subcloning or by in vivo excision of IZap clones. Growth and transfection of Cos-7 cells. A 6 kb Sst II fragment containing most of Fgf8 (see Figure 2.6A) was subcloned into the "exon trap" vector, pML53ln (gift of Dr. M. Reth) (Auch and Reth, 1990), in both orientations. The resulting plasmids were transiently expressed in Cos-7 cells (American Type Culture Collection) by Lipofectin-mediated transfection (BRL-Gibco). Cos-7 cells were grown in Dulbecco's Modified Eagles Medium (DMEM) with 10% fetal calf serum (Sigma), penicillin and streptomycin. When the cells reached 70% confluence, the medium was changed to DMEM without serum and the DNA-lipofectin mixture was added according to the manufacturer's directions (BRL-Gibco). After 12 hours of incubation of the Lipofectin- DNA with the cells, serum was added back and the cells grown in DMEM with 10% fetal calf serum for an additional 48 hours. The cells were harvested and total cellular RNA prepared (Chomiczynski and Sacchi, 1987). 37 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 2.1 Schematic Representation of Exon Trap Cloning Multiple cloning site RP2 exon RP3 exon r i intron bacterial origin SV 40 origin Insert flanking cellular DNA ri T Express vector in CQS cells and isolate RNA PCR primers cDNA synthesis using PCR B Clone and sequence trapped exons 38 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The RNA was used to generate an Fgf8 cDNA, using an "exon trap" protocol (Figure 2.1) (Auch and Reth, 1990). Generation of cDNAs by reverse transcription (RT) and polymerase chain reaction (PCR). Total cellular RNA from transiently transfected COS-7 cells was reverse transcribed in a buffer containing 50 mM Tris-HCI (pH 8.3), 75 mM KOI, 3 mM MgClg, with 10 mg/mL random hexamer primers and avian myeloblastosis virus reverse transcriptase (Stratagene) by incubating at room temperature for 10 minutes, followed by 37°C for 20 minutes and then 42°C for 90 minutes. The resulting cDNA was amplified by nested PCR with Amplitaq (Perkin Elm er, C etus), using an in itia l forw a rd p rim e r (5'-AAGCTCTCTACCTGGTGTGTGG-3‘) in the 5' insulin exon and an initial reverse primer (5 -CAGTGCCAAGGTCTGAAGGTCA-3') in the 3‘ insulin exon, followed by a nested forward primer in the 5' insulin exon (5'-GCGAAGTGGAGGATCCACAAG-3‘) and a nested reverse primer in the 3" insulin exon (5 -ACCCGGATCCAGTTGTGCCA-3'). Initial PCR conditions were 50 mM KCI, 1.5 mM MgCl2 and 40 cycles of the following; denaturing at 94°C for 1 minute, annealing at 59°C for 1 minute, and extending at 72°C for 1 minute. Nested PCR conditions were the same except that the annealing temperature was 50°C. The resulting Fgf8 cDNA product, obtained only from the plasmid with Fgf8\ n the correct orientation, was isolated from an agarose gel, cloned into a pBluescript "T-vector" (Marchuk et al., 1991), sequenced, and found to contain exons 2A, 3A, 4, and 5 up to the Sst II site. The portion of the coding region downstream of the Ssf II site in exon 5 was added by ligating a 160 bp Sst W/Sst I fragment. 39 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. We obtained cDNAs containing all of the known Fgf8 coding region (Tanaka et al., 1992) in 2 subsequent steps. First, upstream cDNAs (exons 1, 2A, and 3A; exons 1, 2A, and 3B; and exons 1, 2B, and 3A) were obtained by nested RT-PCR from murine testis RNA. The RNA was reverse transcribed as above. The first PCR reaction was carried out using a forward primer in Fgf8 exon 1 (5’-CGCAGCCCGACCTCCCTC- 3') and a reverse prim er in Fgf8 exon 3 (S'- GAGCTGATGCGTCACCAGGCT-3'). A nested Fgf8 exon 1 forward primer (5'-GGGAGGTTGGGGTTGTGG-3'), was used with the reverse Fgf8 exon 3 primer above, or with a reverse primer that spans exons 2B and 3A (5'-GTGGTGGGTGAGATGTGGGTG-3‘), in the nested PGR. PCR conditions with recombinant Pfu DNA Polymerase (Stratagene) for both the initial and nested steps were 10 mM KCI, 2 mM MgCl2 , and 40 cycles of the following: denaturing at 94°C for 1 minute, annealing at 58°G for 1 minute, and extending at 72°C for 1 minute. Next, these S' cDNA fragments (189 bp for exons 1, 2A, and 3A; 222 bp for exons 1, 2A, 3B; and 360 bp for exons 1, 2B, and 3A) were annealed with the “exon-trap" fragment and amplified by Pfu DNA polymerase with the forward primer (S'-GGCAGGTTGGGGTTGTGG-3') in exon 1 and a reverse primer in exon S (S'-GGAGTGGGGGTGGATTCGT-3') to give cDNAs containing the entire coding sequence of the 3 isoforms of Fgf8. The conditions for this step were 10 mM KCI, 2 mM MgCl2 , and 25 cycles of the following: denaturing at 94°G for 1 minute, annealing at S5°G for 1 minute, and extending at 72°G for 1 minute. These cDNAs were cloned into pBluescript "T-vectors" (Marchuk et al., 1991). Sequencing of 2 of these fragments showed that they correspond to the previously described 40 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. clones pSC17 (axons 1, 2A, 3B, 4, and 5;) and pSC15 (axons 1, 2B, 3A, 4, and 5) (Tanaka et a/., 1992). The third fragment Is a novel Isoform derived from exons 1, 2A, 3A, 4, and 5 (see Results). Probes and sequencing. DNA probes used for hybridization were isolated from plasmids by digestion with restriction endonucleases, agarose gel electrophoresis, and glass bead binding. Potential probes were tested for repetitive elements by blotting to nylon and hybridizing with 32p.|abeled mouse genomic DNA (Steinmetz et al., 1980). Unique probes were used on Southern and northern blots. In addition to the probes generated from the genomic clones and from PCRs described above, the following probes were used on blots; a 1.9 kb Pst UXhoi fragment (MMTV gag probe) and a 1.2 kb Bam HI fragment (MMTV env probe) (Shackleford and Varmus, 1988); a 475 bp murine acidic FGF/Ffiffi cDNA, a 475 bp murine basic FGF/FgGcDNA, a full-length 800 bp murine Fgf5 cDNA, and a 600 bp murine Fgf6 cDNA (Hébert et al., 1990); a 560 bp murine Fgf6 cDNA and a 1.6 kb Xba \/Xho\ murine 5' genomic Fgf6 fragment (de Lapeyriere et ai., 1990); a 693 bp rat KGF/FgfZcDNA (Van et a!., 1991); a 0.85 kb murine Fgf9 cDNA (gift of D. Omitz); and a 1.3 kb Pst I fragment containing rat GAPDH cDNA (Fort at al., 1985). Double stranded plasmids containing inserts were sequenced using Sequenase version 2.0 (USB). Both strands were sequenced using a combination of deletion subclones and Fg/8-specific primers. Nucleotide sequence accession number. The GenBank accession number of the cDNA encoding Fgfda is U18673. 41 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Results Detection and cloning of a new tum or-specific common insertion site for MMTV. In previous work, It was shown that MMTV infection of Wnt1 transgenic mice accelerates mammary tumorigenesis by insertional activation of Fgf3 and Fgf4 (Shackleford et al., 1993). However, approximately half of the tumors that were examined did not exhibit activation of any of the proto-oncogenes known to be affected by MMTV in tumors of normal mice (including int3 and Wnt3), despite the presence of clonal proviral insertions in these tumors (Shackleford et al., 1993). Therefore, these tumors were likely to harbor unpredicted or novel proto oncogenes activated by proviral insertion. To isolate such genes, we first sought to clone the proviral-cellular junction fragments from tumors that contained only one or two clonally integrated proviruses. Southern blot analysis of tumor DNAs using an MMTV gag probe identified several candidate junction fragments (Figure 2.2 and data not shown). The Xhol-cleaved junction fragments from tumors 95 and 111 (Figure 2.2) were cloned into lambda vectors and isolated by screening with an MMTV gag probe. Unique cellular DNAs flanking the MMTV proviruses in these clones were isolated and used as probes on Southern blots of DNAs from 80 tumors with clonal insertions to ask if other tumors had rearrangements due to proviral insertions in these loci. Data from these experiments suggested that the insertions in tumors 95 and 111 were actually in the same locus and, furthermore, that six other tumors also contained proviruses in this region (Figure 2.3 and 2.4 and 42 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure. 2.2. MMTV proviruses in tumors of Wnt1 tra n s g e n ic mice. (Upper) Genomic DNAs were isolated from normal mammary gland (c), or from mammary tumors (designated by tumor number) of MMTV- infected Wnt1 transgenic mice. DNAs were digested with Xho\, separated by agarose gel electrophoresis, capillary blotted to nylon membranes, and hybridized with a 32P-labeled MMTV gag probe. All DNAs have large gag fragments, indicative of endogenous retroviruses in these animals. Most of the tumor samples have additional fragment(s), indicating the presence of clonal, tumor-specific MMTV proviruses. DNA size markers (kb) are indicated at right. (Lower) Diagram showing the origin of the gag probe used in the upper panel and the 5’ proviral- cellular junction fragment it detects. MMTV DNA is represented with bold lines. Xh, Xho\. 43 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 2.2 Wnt-1 T ransgenic T u m o rs CO T - C J Î C N J C D C O O C D - ^ L n ^ c\j c o c o - ^ - < ^ c o r ^ r ^ c x 3 < j î ^ — 23 — 9.4 — 6.7 — 4.4 — 2.3 Xho \;gag Xh Xh y/— a gag 44 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. data not shown). Hybridization of these blots with MMTV probes confirmed that the rearrangements of this locus in affected tumors were due to proviral integrations (Figure 2.3 and 2.4 and data not shown). The apparent presence of two insertions in tumor 135 is discussed below. Detection of expressed genes in the common insertion region. In a search for active genes in this locus, probes from the cloned cellular DNAs described above, as well as from adjacent regions cloned subsequently, were made for use in northern blot analyses of tumor RNAs. Two genes were identified. A genomic clone that detected RNA from one of these genes was sequenced and found to be identical with the cDNA of androgen-induced growth factor (Tanaka et al., 1992), the eighth member of the FGF family. I will hereafter refer to this gene as Fgf8. The second gene in this locus, which we have designated Npm3 is novel. The deduced Npm3 product is unrelated to any oncoproteins, but it has sequence similarities to two nucleolar chaparone proteins, nucleoplasmin and B23/nucleophosmin (Borer et al., 1989: Dingwall and Laskey, 1990). A description of Npm3 beyond the RNA data below, is provided elsewhere (MacArthur and Shackleford, 1997). When small amounts of total cellular RNAs available from tumors were examined by northern blotting for expression of these genes, evidence for transcriptional activation of FgfS in all tumors with insertions in this locus (except tumor 104) were found; Fgf8 RNA was undetectable in tumors without insertions and in normal mammary gland (Figure 2.5 and data not shown). RT-PCR experiments confirmed the presence of Fgf8 RNA In tumor 104 and its absence in normal mammary gland (data not 45 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 2.3. MMTV proviral insertions into a new common integration locus. (A) Southern blots of DNAs from kidney as a control (c) and from mammary tumors (86, 104, 111, and 135). Rearrangements of this locus in tumors were detected following digestion of the DNAs with either Bgl II or Xba I (which cut within MMTV) and hybridizing the resulting blots with the cellular probe PH 0.8 (see Figure 2.4). Arrowheads identify fragments that anneal to both cellular (left panels) and MMTV gag (right panels) probes, thus demonstrating the location and orientation of the inserted pro virus. (B) Southern blots of DNAs from normal mammary gland as a control (c) and from mammary tumors (61, 63, 94, 95). Rearrangements of this locus in tumors were detected following digestion of the DNAs with Xho\ and hybridizing the resulting blots with the cellular probe XP1200 (see Figure 2.4). The orientation of the MMTV pro virus was determined as in (A) but using XP1200, gag and env probes. 46 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 2.3 C g ° L O e n _ S T - in o o f— e n 0 0 T — f — T — a a s M - 5 . 1 Bgl ii/PH 0.8 B g l Wgag C g ° in e n g ° L D C O 4 8 22 1 2 7,1 - 5.1 X b a I/P H 0 .8 X b a \/gag 47 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 2.3 (continued) B m ^cocf> ! -Il •«» - 9 . 4 - 6 . 7 # XP1200 gag env S S C S S I “ 5 ' — 4.1 XP1200 gag — 4.1 48 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 2.4 Map of provirai insertions in the new common Integration locus. This locus was mapped with restriction digests of lambda phage clones and plasmid subclones containing inserts from this locus, and by Southern blotting of genomic DNAs using probes from the locus. The lambda phage clones are indicated at the bottom of the figure. MMTV proviral insertions were mapped by Southern blotting using probes from this locus (XP 1200, PH 0.8, and PP 1.3) and from MMTV as in Figure 2.3. In some cases, the MMTV proviral insertion site was sequenced following amplification by PCR. Arrowheads denote the location and transcriptional orientation of the MMTV proviruses. B, Bam HI; Bg, Bgl\\\ RI, Eco RI; 82, Sst II; X, Xba I; Xh, Xho\. 49 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 2.4. 194 6H 63, 9 ? 104^ 1 1 1 ^ 135^ 86 MMTV INSERTIONS Npm3 Fgfd B B S J 1, , L 1 X Ri X X PROBES XP1200' S2 B B S2 ■ L j , , I, n i S2 1 _ 6 X RI Xh Bg Xh Xh R I 4 kb PH0.8 PP1.3 X-PHAGE CLONES 50 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. shown). In contrast, RNA levels of the constitutively active Npm3 gene do not appear to be significantly affected by the surrounding MMTV insertions (Figure 2.5). Thus the activated expression of Fgf8 in multiple tumors by clonal proviral insertions suggests that this gene is causally related to tumor formation. Fgf8 exon structure and alternative splicing. To determine the physical relationship of the proviral insertions to the Fgfd gene, the exon structure was established by sequencing portions of the genomic clones and comparing these with the Fgf8 cDNA sequences reported earlier (Tanaka et al., 1992). This revealed that Fgf8, as represented by these cDNAs, is approximately 6.5 kb in length and has at least 5 coding exons (Figure 2.6A), unlike other analyzed FGF genes which have only three coding exons (Abraham et a!., 1986; de Lapeyriere eta!., 1990; Gospodarowicz eta!., 1987; Kelley eta!., 1992; Moore et a/., 1986; Yoshida et a!., 1988; Zhan et a!., 1988). All of the insertions detected were mapped upstream of Fgf8, and all, except those in tumors 61 and 86, were in the opposite transcriptional orientation to Fgf8 (Figure 2.4 and 2.6A and data not shown). Two of the insertions (tumors 63 and 95) disrupted the Npm3 gene, which is located approximately five kb upstream of Fgf8 in the same orientation, and twoinsertions (tumors 61 and 94) were upstream of both genes (Figure 2.4 and data not shown). The comparison of Fgf8 genomic and cDNA sequences also showed the existence of at least two 5' splice sites in exon 2 and two 3' splice sites in exon 3. Alternative splicing between these sites could 51 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Normal T issues M am m ary Tumors ' ' - - + + + ^ + + + + ( Insertions m Fgf-8 locus F g f-8 0.37 0 55 0 36 N ub - 7 GAPDH - r " w g # # # # Figure 2.5 Transcriptional activation of FgfS by MMTV insertion in tumors from infected Wnti transgenic mice. Northern blot of total cellular RNA (5 |ig) from three normal tissues (adrenal, kidney, and mammary gland), two tumors without MMTV proviruses near FgfS (tumors 14 and 42), and eight tumors with MMTV insertions near FgfS. The blot was initially probed with a full length FgfSb cDNA (upper panel), then with a A/pm5cDNA (middle panel), and finally with a GAPDH cDNA (lower panel). Locations of the RNA size markers are shown at right. 52 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. potentially produce RNAs encoding four isoforms of the protein (Figure 2.6A-C). In "exon trap" experiments with an Sst II genomic fragment containing most of Fgf8 (see Materials and Methods), and in nested RT- PCR experiments using RNA from adult murine testis and from tumor 61, we amplified and cloned cDNAs of alternatively spliced mRNAs encoding three different isoforms: FGFSa, FGFSb, and FGFSc (Figure 2.6A-G). We did not obtain a cDNA containing both exons 2B and 3B, which would encode the fourth potential isoform (data not shown). All three cloned versions have the same predicted signal peptide sequence, and code for the same protein downstream of the alternative splice, but differ in the amino-terminal regions of the putative secreted products (Figure 2.60). Expression of Fgf8 in murine tissues. We performed northern blots of poly(A)+ RNA from normal murine tissues to determine the normal expression patterns of Fgf8 (Figure 2.7). Weak expression was detected in ovary and testis, but we did not detect Fgf8 RNA in normal mammary gland, or in any of ten other tissues (Figure 2.7). These data suggest that MMTV insertions in tumors activate this gene from a transcriptionally silent state. Since FGFs are frequently expressed during embryogenesis (Basilico and Moscatelli, 1992), we looked for expression of FgfSduring murine development by northern blot analysis of total cellular RNA from whole embryos staged by gestational age. We found high expression of Fgf8 at 10.5 days post conception (E10.5), the earliest time tested, which decreased to nearly undetectable levels by E l4.5 (Figure 2.7). We did not observe Fgfgexpression at E16.5 or E l7.5 (data not shown). 53 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 2.6. FgfS gene structure and protein sequences. (A) FgfS gene structure. The five exons of FgfS are depicted as boxes, with filled areas indicating the coding domains. Exons 2A and 2B differ in their 3' termini, and exons 3A and 3B have different 5' termini. Note that there are at least four potential alternatively spliced forms of FgfS mRNA (containing exons 2A-3A, exons 2A-3B, exons 2B-3A, or exons 2B-3B), all of which would maintain the same reading frame. Exon contents of cDNAs encoding three FgfS protein isoforms are indicated below the gene structure. FgfS sequences outside the jagged boundaries of exons 1 and 5 have not been determined. (B) Sequence of FgfS exon-intron junctions. Exon sequences are indicated by uppercase letters, and intron sequences by lowercase letters. Highly conserved nucleotides at the ends of introns are in bold. Three-letter amino acid abbreviations are above each codon. (C) Amino acid sequences of three FgfS isoforms. The predicted signal peptide sequence (Von Heijne, 1986) is underlined. Exon splice sites are indicated with bars. Note that the three isoforms differ only at the amino termini of the secreted forms. 54 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 2.6 ^ 1 0 4 ^ 1 1 1 <4135 EXONS 86 Fgf8 S2 I Xh 1 2A 3A I I 2B 3B Bg Xh S2 SI 4 - L É ^ - 4 5 1 kb ISOFORM Fgf-8a Fgf-8b Fgf-8c EXONS 1, 2A, 3A, 4, 5 1, 2A, 3B, 4, 5 1, 2B, 3A. 4, 5 55 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ■ o 1 8 Û . ■ o CD ( / ) o' 3 8 ■ o < 5 ' 3 CD 3 . 3 " CD CD ■ D O Q . C a O 3 T D O CD Q . " O CD C /) C /) B EXON J 2A 2B 3A /3B 4 C F g fS a F g fS b F g f 8 c F g fS a F g fS b F g f Sc Figure 2.6 (continued) 5' SPLICE SITE 3' SPLICE SITE EXON TGA GCTG CC'l'gtgayLac;cy LcLcccacagüG'1'G'ITGCAG 2A /2B C C A A G C C C A G gtaaggagcg.......... 111 tacacagCATGTGAGGG 3A C A C A C A G C G A gtgagtggag.......... tttcctaaag G T A A C T G T T C 38 G A C C C C T T C G gtaaggtgca.......... ccatctccagC G A A G C T C A T 4 A A T T G C C A A G gtgagctcca.......... ctacctgcagA G C A A C G G C A 5 1|2A MGSPRSALSCLLLHLLVLCLQAO------------------------------------------------------------------ 1|2A MGSPRSALSCLLLHLLVLCLQAO------------------------------------------------------------------ 1|2B M GSPRSALSCLLLHLLVLCLO AOVRSAAOKRGPGAGNPADTrGOGHEDRPFGQRSR 2A|3A ---------------------------------- HVREQSLVTDQLSR 2A|3B -------------------------------------------------------------- VTVQSSPNFTQHVREQSLVTDQLSR 2B|3A AGKNFTNPAPNYPEEGSKEQRDSVLPKVTQR----------------------HVREQSLVTDQLSR 3A/3B|4 F g f S a , b , c RLIRTYQLYSRTSGKHVQVLANKRINAMAEDGDPFAKLIVETDTFGSRVRVRGAET 4 | 5 F g f 8 a , b , c GLYICMNKKGKLIAKSNGKGKDCVFTEIVLENNYTALQNAKYEGWYMAFTRKGRPR F g f 8 a , b , c KGSKTRQHQREVHFMKRLPRGHHTTEQSLRFEFLNYPPFTRSLRGSQRTWAPEPR Discussion We previously showed that MMTV infection of Wnt1 transgenic mice with MMTV accelerates mammary tumorigenesis by insertionally activating Fgf3 and/or Fgf4, indicating that these two members of the FGF family can cooperate with Wnt1 in murine mammary tumorigenesis (Shackleford et al., 1993). By cloning and analyzing proviral insertion sites from several tumors that lack these activations, we show here that 8 of 80 tumors with clonal, tumor-specific MMTV proviruses had MMTV insertions in a genomic region that contains FgfS (Figure 2.3 and 2.4). Elevated levels of FgfS RNA were detected in affected tumors but not in other tumors or in normal mammary gland. (Figure 2.5). Analysis of the normal gene and its expression showed that FgfS contains at least five coding exons, can encode at least three different protein isoforms by alternative splicing, and that RNA was present in testis and ovaries of adults and in midgestational embryos. Together, these findings suggest that the overexpression of FgfS which accompanies MMTV insertion into this locus contributes to the formation of these tumors. Proviral Insertions. Studies of MMTV insertion sites in mammary tumors indicate that MMTV can transcriptionally activate an adjacent proto-oncogene from upstream or downstream locations in either transcriptional orientation relative to the gene, although the vast majority are classical "enhancer insertions" in which the S' LTR of the provirus is situated closest to the activated gene (Clausse et al., 1993; Nusse, 1991; Peters, 1990). All of the insertions described here are located upstream of FgfS; two of these (in tumors 61 and 86) were in the 57 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Fgf-8 - 6.2 - 3 . 9 - 2.8 - 1 . 9 - 0 . 8 7 - 0 . 5 6 GAPDH Figure 2.7 FgfS is normally expressed during embryogenesis and in adult ovaries and testes. Northern blots of normal adult tissue poly(A)+ RNAs (Br, brain; He, heart; Kl, kidney; Li, liver; Lu, lung; MG, mammary gland; Mu, muscle; Ov, ovaries; SG, salivary gland; Sp, spleen; Te, testis; Th, thymus; and Ut, uterus) and murine embryonic total cellular RNAs (embryo lanes labeled in days post conception). The FgfS and GAPDH probes are the same as in Figure 2.5. RNA size markers ,.<b) are denoted at right of each blot. 58 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. same orientation as Fgf8 (and Npm3). The location of the provirus in tumor 61 might predict a large fusion transcript with Npm3 starting from an MMTV LTR, but no aberrant Npm3 (or FgfS) RNAs were observed in this tumor. Thus it appears that the Fgf8 gene is activated by an enhancer insertion mechanism in tumor 61, despite the "promoter insertion" orientation of the pro virus. The identities of the additional sequences in the Fgf-8 transcripts larger than 1.9 kb observed in several tumors (e.g., tumors 63, 86 and 135 in Figure 2.5) are unknown. Tumors 61 and 94 are unusual in that a second gene, Npm3 is found between the proviral insertions and the activated FgfS gene. The basal activity of Npm3 does not seem to be significantly affected by the insertions, while FgfS is activated from an apparently quiescent state. Although proviral enhancers are known to activate genes from significant distances, to our knowledge, MMTV activation of the distal, but not the proximal, gene of a linked pair has not previously been observed. The apparent selective mechanism involved is unclear, but may be explained in part by the relatively high basal level of Npm3 expression, by possible toxicity of higher levels of Npm3 protein, or by a discriminating compatibility of MMTV enhancers with the FgfS promoter. Tumor 135 appears to contain two separate MMTV proviral insertions at similar locations in this locus, but not both insertions on a single chromosome (Figure 2.3A). It seems unlikely that a selective advantage for tumor cell growth would result from both copies of the gene being activated by proviral insertion, since, for example, we have never seen insertional activation of the endogenous Wnt1 gene in MMTV-infected Wnt1 transgenic mice (Shackleford et al., 1993). 59 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Although requirement for a threshold level of Fgf8 expression might explain the two insertions, alternative interpretations are that tumor 135 is oligoclonal, with tumor cells possessing one or the other insertion, or that a single insertion exists in a rearranged form in some tumor cells. In addition, tumor 135 has an insertion in the Fgf3/Fgf4 locus (Shackleford et al., 1993). This finding resurfaces the notion of possible cooperation among FGF family members (Shackleford et al., 1993), but oligoclonality could also explain the data. Gene activation by MMTV insertions. Almost all tumors with insertions in this locus displayed readily discernable activation of Fgf8 using small amouts of total RNA by northern analysis. Only tumor 104 required RT-PGR for detection. The cause for the wide variation in MMTV-activated Fgf8 gene expression is unknown, but such variation is common for genes activated by MMTV insertions 15, 31, 34, (Shackleford St a!., 1993). Due to the lack of obvious changes in Npm3 RNA levels in affected tumors, and to the interruption of its coding domain in two tumors, alterations in Npm3 expression by proviral insertions do not appear to contribute to tumorigenesis. The relationship of Npm3 to the nucleolar proteins nucleoplasmin and B23/nucleophosmin, two abundant and well-studied molecular chaperones with no known oncogenic potential (Borer eta!., 1989; Dingwall and Laskey, 1990), also does not immediately point to a role for Npm3 in tumorigenesis. Although the recently described fusion of nucleophosmin to the cytoplasmic tyrosine kinase domain of a transmembrane receptor in chromosomal translocations in some lymphomas is intriguing in this regard, nucleophosmin's probable role is to provide an active promoter 60 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. for the kinase (Bullrich etal., 1994; Morris et al., 1990). Together with the activated expression of Fgf8 shown here and the reported transforming potential of FgfSb cDNA in NIH3T3 cells (Kouhara et a!., 1994), it is likely that Fgf8, and not Npm3, is the gene pertinent to oncogenesis in this locus. Alternative splicing and gene expression. The transcribed portion of the Fgf8 gene is approximately 6.5 kb and is divided into at least five coding exons (Figure 2.6A), two more than the other examined FGF genes, (Abraham at a/., 1986; de Lapeyriere et a/., 1990; Gospodarowicz et al., 1987; Kelley et al., 1992; Moore et al., 1986; Yoshida et al., 1988; Zhan ef al., 1988). The first three exons of Fgf8 described here correspond to the first exon of other FGF genes. Fgf8 is also unusual in that it can be alternatively spliced between exons 2 and 3 to yield at least three protein isoforms which differ only in the amino terminal portion of the secreted mature protein. The two previously described cDNAs (Tanaka etal., 1992) were isolated, which code for the isoforms FGF8b and FGF8c, and a third cDNA that encodes the previously undescribed isoform FGF8a (Figure 2.6A-C). All three mRNA species were detected in both testis and tumor 61. These cDNA clones isolated will allow a comparison of the three isoforms for possible differences in, for example, their transforming potentials, tissue expression patterns, half-lives, or affinities to the various FGF receptors. Expression of Fgf8 was found in adult testis and ovary (Figure 2.7), but this expression is much lower than in the tumors with MMTV insertions near Fgf8 (Figure 2.5). The expression of Fgf8 during murine embryogenesis (Figure 2.7) suggests that this gene may be important in 61 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. embryonic development. Recent studies using in situ analysis of whole mount embryos and embryo sections confirm the expression of Fgf8 during murine embryogenesis (Heikinheimo etal., 1994). Definitive proof that Fgf8 expression is important in murine development will require creation and analysis of mice lacking a functional Fgf8 gene. Wnti and FgfS cooperation. Fgf8 is the third member of the FGF family to be identified as a Wnti collaborator mammary carcinogenesis. Surprisingly, only FGF genes have been found to be activated by MMTV in this transgenic model system (Shackleford et al., 1993). Assuming that many proto-oncogenes are potential targets for activating insertion mutations, this selectivity suggests that FGFs cooperate oncogenically with Wnti in an especially potent manner. It is possible that the signaling pathways of these two families function in a highly complementary fashion to deregulate mammary epithelial cell growth and proliferation. Altematively, activated FGFs could contribute to neoplasia in a less direct manner by stimulating angiogenesis (Folkman et al.. 1989: Kandel et al., 1991). However, it is notable that genes of the Wnt and FGF families can evidently collaborate in an experimental model of mesoderm induction (Christian et a!., 1992), suggesting that cooperation in tumorigenesis may derive from the normal interactions of these genes. The detection of a third activated FGF gene in these tumors prompted us to survey all other known FGF loci for insertions by Southern blot analysis, but no rearrangements in other FGF genes have been found thus far (Shankar and Shackleford, ). Given the propensity of this model system to activate FGF genes, we speculate that analyses of 62 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. other common insertion loci in the remaining tumors from this collection may reveal new members of this gene family or identify important downstream components in the FGF signal transduction pathway. 63 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter 3 Demonstration of the Oncogenic Ability of Fgf8 Introduction Oncogenesis is a multistep process involving the sequential acquisition of multiple genetic alterations. A Wnt1 transgenic mouse model has been used to study the multistep process of mammary tumorigenesis. Wnti transgenic mice were previously infected with mouse mammary tumor virus to accelerate tumorigenesis and "molecularly tag" proto-oncogenes that are activated in the resulting tumors (Shackleford at al., 1993). Using this approach, Fgf3 and Fgf4 were identified as Wnti collaborating proto-oncogenes (Shackleford at al., 1993), and subsequently cloned a genomic locus that contained MMTV insertions in the DNA from several mammary tumors of MMTV- infected Wnti transgenic mice (MacArthur at al., 1995). The activated gene in this locus was determined to be another member of the FGF family of growth factors, Fgfd (described in the previous chapter and (MacArthur era/., 1995). FgfS consists of at least six exons and codes for at least seven protein isoforms, due to alternative splicing of the primary transcript (Crossley and Martin, 1995; MacArthur at al., 1995; Tanaka at al., 1992) (Figure 3.1). RNAs for the FgfS isoforms are present during murine embryogenesis (Crossley and Martin, 1995; Heikinheimo at al., 1994; MacArthur at al., 1995; Ohuchi at al., 1994) and are detectable only in 64 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 3.1. Structure of FgfS protein Isoforms. FgfS consists of at least 6, exons which codes for potentially eight isoforms. Exons coding for regions common to all the mature FgfS isoforms are indicted as filled boxes (□ ). Exons that code for the shared signal peptide is indicated by (B ). The alternatively spliced exons are indicated by hatched boxes (□ ), and non coding exons by (□ ). The isoforms 8a, 8b, and 8c indicated, are characterized in this study. 65 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 3.1 EXON ISOFORM 8a 8b 8c 8d 8e 8f 8g 8h Protein Isoforms of Fgf8 F g fS o B I I m I I — in E n M « i w u ÜJU I Signal peptide B Common exons □ Alternatively spliced exons 66 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. the gonadal tissue of adult mice at low levels (MacArthur et al., 1995). These results and the Insertional activation of Fgf8 described in the previous chapter suggest that Fgfd is a normal embryonic gene that is oncogenic when overexpressed in adult mammary tissue. The significance of the different FgfS isoforms is not known; however their existence suggests that they have different biological properties. Overexpression of FgfS (Goldfarb et al., 1991), Fgf4 (Delli Bovi et al., 1987), FgfS (Zhan era/., 1988), FgfS (de Lapeyriere et al., 1990), and at least one Fg/B isoform, FgfSb (pSC17 (Kouhara et al., 1994; Tanaka et al., 1992), can transform NIH3T3 cells. Given the ability to observe a phenotype in NIH3T3 cells in response to FGFs, we decided to test the biological properties of FgfS isoforms in NIH3T3 cells. In a collaboration with Dr. Roy-Burman and coworkers (USC), cDNAs encoding three alternatively spliced isoforms of the human FGF8 were isolated gene from a prostatic carcinoma cell line. These clones, designated FGFSa, FGFSb and FGFSe, differ from each other at the NH2- terminal region of the mature proteins and share extensive nucleotide sequence homology to their murine counterparts (Crossley and Martin, 1995; MacArthur et al., 1995). FgfSa and 8b exhibit identical amino acid sequences to the corresponding mouse isoforms. FGFSe displays partial sequence variation from the corresponding mouse clone in the extra exon sequence found in this isoform in both species (Figure 3.5). There is extensive sequence diversity between the human and mouse Fgf genes in the 3' untranslated region of the mRNAs. Human FGF8 is normally expressed in fetal kidney, adult kidney, prostate, and testes. To 67 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. further demonstrate the oncogenicity of Fgf8, we tested the human FGF8 isoforms for differences in their NIH3T3 transforming potentials. Previous studies have shown that transgenic mice offer definitive tests for candidate oncogenes and tumor suppressor genes (refer to chapter 1). Several lines of transgenic mice have been generated that express various oncogenes in the mammary glands (Table 1.2). In the majority of cases, mammary tumors develop in these mice in a stochastic fashion with a latency of about 5-10 months, suggesting that the overexpression of the transgene contributes to, but is not sufficient for transformation. To further study the role of Fgf8 in transformation of the mammary gland, transgenic mice were created targeting the Fgf8 gene to the mammary gland under the control of the MMTV-LTR. Materials and Methods Cell Lines and Vectors. NIH3T3 cells (gift of R. Weinberg, Whitehead Institute, Cambridge, MA) were grown in DMEM supplemented with 10%FCS in humidified incubators with 5% CO2 at 37°C. Cells were passaged at subconfluence to avoid selecting for spontaneous transformants: any cultures with such variants were discarded. We used an expression vector, pMIRB (gift of D.M. Ornitz, Washington University, St. Louis, MO), that generates a bicistronic mRNA that has an internal ribosome entry site (1RES) for the downstream neomycin phosphoribosyl transferase gene, allowing translation of both transcription units (Figure 3). cDNAs encoding the three different Fgf8 isoforms (refer Chapter 2) were cloned into the upstream position of 68 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. pMIRB (Figure 3) using the EcoR\ and blunted Spel sites. The cDNAs were sequenced to confirm authenticity. For the human FGF8 assays, the cDNAs {FGFSa, FGF8b and FGFfle) were excised form the parental plasmid by EcoRI digestion and cloned into the corresponding site of the expression vector pcDNAS (Invitrogen). Transfection. The Fgf8 expression vectors (FgfSa, FgfSb, FgfSc) and the control pMIRB vector were used to transfect NIH3T3 cells using OptiMEM and Lipofectamine (Life technologies. Inc.). Briefly, 2-4 x 10^ cells were cultured on 6-well dishes (Corning-Costar). The next day, the cells were washed with OptiMem, and then placed with one ml of OptiMem. Expression vector DNA (MIRB, Fgf8a, Fgf8b, or Fgf8c; human FGF8a. 8b. 8e, or pcDNA; 2 ug ) was combined with 15 ^1 of Lipofectamine in 200 pi of OptiMem per manufacturer’s instructions. The DNA-Lipofectamine mixture was placed on the cells, and the cells were incubated for S-12 h at 37°C and 5% C02. One ml of DMEM with 10% FCS was added to the cells after the incubation period. Twenty-four hours after transfection, the cells were transferred to two 10 cm plates. One plate was grown to confluence for observation of focus formation. The other plate of cells were placed under selection in G418. The cells were grown in DMEM with 10% FCS containing 400 pg/ml Geneticin (G418; Life Technologies, Inc.). These cells were grown for 5-7 days. The G41S resistant colonies (G41Sr) were pooled from each transfection to create stable cell lines for transformation studies. All the transfections were done in triplicates. 69 Reproduced with perm issionofthe copyright owner. Further reproduction prohibited without permission. Transformation Assays. Focus formation. The unselected ceils from each transfection were grown to confluence, and observed for piling of cells (focus) among the uniform monolayer. Five to seven days past the day of confluence, the medium was removed and cells stained with crystal violet. The foci were observed and counted in the stained plates. Morphological transformation. About 60-75 G418r colonies were pooled from each transfection/selection. This pooling was done to minimize the possibility that a single clone that was morphologically transformed by an event unrelated to expression of the Fgf8 isoform would bias our results. The pooled cells were grown to confluence and were observed for changes in morphology, and documented by photomicroscopy. Anchorage independent growth. 1x10^ to 1x106 ceW s (MIRB, FgfSa, FgfSb, FgfSc; pcDNA, FGFSa, FGFSb, FGFSe) were embedded in 0.35% agar (Difco) in DMEM with 10% FCS, over a substratum of 0.5% agar, and incubated at 37°C. Visible colonies were counted after 14 days. Tumorigenicity in nude mice. Five week old nude mice {nu/nu; Jackson laboratory) were injected with 1x106 G41Sr cells in 0.1 ml of PBS subcutaneously into the flank. The animals were maintained in laminar cages under aseptic conditions. Following injection, the animals were observed every two days for the development of tumors. Animals were sacrificed when the tumor became 2 cm in its largest diameter. The tumors were used for isolating RNA and establishing post-tumor cell lines. Cell Proliferation assays. Multiple platings of G41S resistant cells (10<) from each transfection were plated in wells of 6-well dishes and grown at 70 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 37°C and 5% C02 in DMEM-10%FCS. Cells from each transfected/selected cell line were removed with trypsin and counted using a coulter counter. The doubling time was determined for cells in log phase. The saturation density for each cell line was determined when the cells reached confluence. The data were collected in duplicates on four occasions, both at the initial development of the cell lines and after several months in cryopreservation. Results are presented as the mean ±SE. Preparation and Screening of Genomic Library. The transgene from tumor 86 was cloned using the same strategy described in Chapter 2. The probes used to isolate the clone were the 2.0 kb genomic Fgf8 XE 2.0 fragment (see Figure 2.4), and the MMTV env probe (900 bp fragment). Microinjection. The injection of the DNA into eggs, and generation of transgenic mice were performed by the transgenic core facility (University of Southern California). Identification of Transgenic animals. DNA was extracted from 1 cm sections of the tails as described before. 10 mg of DNA was digested with Bgl -II and electrophoresed through a 0.6% agarose gel. The gel was blotted to nylon as described (see Chapter 2, Materials and Methods). The blots were hybridized to the 2.0 kb XE 2.0 (Figure 2.4, Figure 3.) and exposed to X-ray film as shown in Chapter 2. The XE 2.0 probe will detect both the endogenous Fgf8 fragment (22 kb) and a transgene specific (6.0 kb) fragment. 71 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Histology of tumors. The tumors from the transgenic animals were fixed in 10% formalin, embedded in paraffin, sectioned at lO^m and stained with hematoxylin and eosin. Results NIH3T3 ceils transfected with different mouse F g f8 cDNAs display different morphologies. In chapter 2, I described the identification and cloning of the cDNAs encoding three {FgfSa, FgfSb, FgfSc) of the seven known FgfS isoforms. To test the differences between their biological properties, we subcloned the corresponding cDNAs into the expression vector pMIRB (Figure 3.2). The majority of the amino acids encoded by these three FgfS isoform cDNAs are identical, but they differ in the region immediately following the signal peptide, i.e., the amino terminal portions of the mature secreted protein (Figure2.6; chapter 2.) (MacArthur ei al., 1995). The resulting vectors were used in transfection experiments of NIH3T3 cells, and stable cell lines were established after selection in G418. We chose to pool the selected G418r clones rather than individually clone them, so as to preclude inaccuracies due to clonal variation. The expression vector used, pMIRB, allowed this approach, since virtually all the selected G418r cells would express the Fgf8 isoform (Figure 3.2). NIH3T3 cells transfected with expression vectors coding for the Fgf8 isoforms and selected for G418 resistance display different morphologies (Figure 3.3). Cells transfected with pMIRB containing the 72 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CD ■ D O Q . C 8 Û . ■ D CD ( / ) ( / ) 8 5 c q ' 3" 3 CD CD ■ o I c a O 3 ■ o O & o c % ( / ) C O o' 3 Fg/X cDNA Rl Sp neoR MoMLV LTR 1RES SV40 splice/ poiy(A) Figure 3.2 Schematic representation of the expression vector pMIRB Transcription from the Moloney murine leukemia virus long terminal repeat (MoMLV-LTR), of the transfected plasmids result in a single transcript that uses the SV40 splicing and polyadenylation sequences. An internal ribosome entry site (1RES) allows translation of the sequences coding for G418 resistance (neoR). The Fgfd cDNAs were cloned into the EcoRI and Spel sites. & î ; Figure 3.3. NIH3T3 cells transfected with different FgfS isoform cDNAs display different morphologies. Pooled G418 resistant cells , transfected with pMIRB only (M), pMIRB containing the FgfSa cDNA (A), FgfSb cDNA (B), or the FgfSc cDNA (C) were grown as described in the Materials and Methods. 74 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. FgfSb cDNA (hereafter called 8b cells) displayed marked morphological transformation with an elongated spindle like shape (Figure 3.3). In contrast, cells transfected with the pMIRB containing the FgfSa cDNA (hereafter called 8a cells) or the pMIRB containing FgfSc cDNA (hereafter called 8c cells) displayed mild to moderate degrees of transformation, with 8c cells being more transforming than 8a (Figure 3.3). NIH3T3 cells transfected with pMIRB alone (MIRB cells) were morphologically identical to nontransfected NIH3T3 cells (Figure 3.3). These results suggest that the different Fgf8 isoforms have different potencies with respect to morphological transformation of NIH3T3 cells. NIH3T3 cells transfected with different FgfS isoforms show differences in focus formation. To test the NIH3T3 transfected cells expressing the different Fgf8 isoforms to be contact inhibited, we looked for the ability of the cells to for foci after the formation of a monolayer. A fraction of the transfected cells (see Materials and Methods) were grown to confluence and observed periodically for the formation of foci. The 8b cells, were not contact inhibited and started to form foci (piling of the cells). The cells within a focus was showed morphological transformation. Within 1-2 days, the cells in the entire plate became transformed, probably due to the secretion of the Fgf8b protein into the medium (Figure 3.4). 8a and 8c cells did not show any focus formation. They were contact inhibited and resembled MIRB cells in morphological appearance. 75 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5 > i l i . « ' ^ m Figure 3.4. Focus formation by MlRB-FgfSb on a monolayer of NIH3T3 cells. NIH3T3 cells were transfected with plasmids MIRB, MIRB- FgfSa, MIRB-FgfSb, M\HB-Fgf8c, and grown without selection for G418 resistance. The cells were allowed to remain at confluence for 2-3 days. The figure shows a typical focus in a dish transfected with FgfSb cDNA. No such foci were visible in the control (MIRB), FgfSa, or Fgf Sc transfected cells. 76 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. NIH3T3 cells producing different FgfS protein isoforms display different transforming properties in vitro and in vivo. The pooled G418r cells from each isoform transfection were compared In several biological assays. The doubling times of the cell lines were examined, and the 8a, 8b and 8c cell lines displayed slightly shorter doubling times when compared to the MIRB cells (16 h versus 18 h), but the difference was not statistically significant. In conditions of low serum ( 0.1% or 1% FCS) the MIRB and 8a cells died, while the 8c cells died in 0.1% and stopped growing in the 1% serum concentrations (Table 3.1). 8b cells on the other hand continued to thrive and displayed the same morphological transformation, suggesting that it was the most potent Fgf8 transforming protein. The saturation densities of the Fgf8 expressing cell lines were examined. At confluence, the number of 8a, 8b, and 8c cells was two, five, and four times the number of MIRB cells, respectively (Table 3.1). In soft agar clonagenicity assays, only the 8b cells were able to form soft agar colonies at an average frequency of 5% (Table 3.1). These in vitro assays of proliferation and transformation all suggest that Fgf8b is the more potent transforming protein isoform. In nude mice tumorigenecity assays (Table 3.1), tumors formed rapidly when 10® 8b cells were injected into nude mice, with all 10 animals possessing fibrosarcomas 2 cm or larger after 1 week. No tumors were seen in animals injected with 10® MIRB cells, even after 4 months of observation. Tumors were observed in two of the four animals injected with 10® 8a cells and three of the four animals injected with 10^ 8c cells, but the tumors in both the groups took 4-6 weeks to attain a size 77 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 3 CD ■ D O Q . C 3 Q . ■ O CD 3 c / ) C / î o' = 5 Table 3.1 Properties of NIH3T3 cells expressing FgfS Isoforms 2, CD 8 5 c 5 ' S i Cell line Growth in reduced serum {0.1%) Doubling time Saturation density (103 cells/cm2) Colonies in soft agar (% o f plated cells) Tu m origen ec ity in mude mice 3 CD -n c 3" CD MIRB No 18+2 12+2 0 0/4 3 ■ O 1 C & . o 3 FgfSa No 16+2 27+2 0 2/4 (6 wks) ■ O O S ' Q . r - H 3" O FgfSb Yes 16+2 67+2 5+2 10/10(1 wk) C s I. c/) o " FgfSc No 16+2 48+2 I) 3/4 (4-6 wks) 3 -vj 0 0 of 2 cm and were not detected in the first 3 weeks after the injection of the cells. Tumors were isolated from all the animals, and G418r cells from tumors from each of the Fgf8 isoform groups were reselected in culture. The morphology of the cells in culture after passage as tumor in the animals was identical to the morphology of the cells prior to passage as tumor. These results indicate that the morphology and expression of the transfected Fgf8 cDNAs are stable and are not altered by passage as tumors in nude mice and suggest that the production of the Fgf8 protein isoform is responsible for the observed phenotypes. Oncogenic potential of human FG F8 isoforms The human isoforms of FGF8 were isolated by reverse transcription-polymerase chain reaction (RT-PCR) from a human prostate tumor cell line. Three different isoforms were isolated which correspond to the murine Fgf8a, 8b and Be isoforms. The human FGFBa and FGFBb exhibit identical amino acid sequences to their murine counterparts, while FGFBe shows a partial variation from the corresponding murine isoform in the additional exon found in both species (Ghosh et al., 1996) (Figure 3.5). To address the biological effect of specific isoform expression, the cDNAs corresponding to the different isoforms {FGF8a, FGF8b, andFGF8c) were cloned into the eukaryotic expression vector pcDNA (Invitrogen). The corresponding plasmids together with the empty vector control were transfected into NIH3T3 cells and stable cell lines were established as described above, by selection in G418. Pooled G418 resistant clones were examined to avoid potential atypical representation 79 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Mouse FGFSe 1 Human FGFSe T _ Human FGFSa Human FGFSb Mouse FGFSb - M ouse FGFBe Human FGFBe 51 Human FGFBa 24 Human FGFBb 24 M ouse FGFBb 24 M ouse FGFBe 90 Human FGFBe 90 Human FGFBa 61 Human FGFBb 72 M ouse FGFBb ' ^ M ouse FGFBe 140 Human FGFBe 140 Human FGFBa 111 Human FGFBb 122 M ouse FGFBb 122 M ouse FGFBe 190 Humain FGFBe 190 Human FGFBa 151 Human FGFBb Mouse FGFBb I" ! M ouse FGFBe 1 Human FGFBe Human FGFBa Human FGFBb 1 M ouse FGFBb 1 M ouse FGFBe 51 Human FGFBe 51 Human FGFBa 24 Human FGFBb 24 M ouse FGFBb -•i M ouse FGFBe Human FGFBe 90 Human FGFBa 61 Human FGFBb 72 M ouse FGFBb 72 M ouse FGFBe 140 Human FGFBe 140 Human FGFBa Human FGFBb 122 Mouse FGFBb Mouse FGFBe 190 Human FGFBe 190 Human FGFBa 161 Human FGFBb 172 M ouse FGFBb 172 MGSPRSALSCLLLHLLVLCLQAQEGPGGGPALGREPTSLLRAGREPQGVS ...................................................................... R .................. LA. . F .......................... QQ HVREQSLVTDQLSRRLIRTYQLYSRTSGKHVQVLANK VTVQSSPNFTQ. .L A . 51 QQ- - HVREQSLVTDQLSRRLIRTYQLYSRTSGKHVQVLANK --'v ’T'/Q SSPNFTQ . RINAMAEDGLPFArtLI'ÆTDTFGSR'/RVRGAETGLYICI-lNKKGKLIAKSM Figure 3.5. Comparison of the deduced amino acid sequences of the three human FGF8 cDNA isoforms with those of mouse FgfS isoforms. The mouse isoforms correspond to FgfSb and FgfSe. identical aminoacids are indicated by dots. Note that following the predicted signal peptide sequence (positions 1-23), the three isoforms differ only at the amino termini of the secreted forms. Dashes, sequence gaps 80 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. due to clonal variation. Of all the different isolates of transfected NIH3T3 cells, the cells transfected with FGF8b cDNA displayed the most striking phenotypic changes (Figure 3.6). These cells exhibited marked morphological transformation with an elongated, spindle shape, in contrast to the flat morphology seen with the control transfectants receiving the vector alone. These cells also had a higher saturation density at confluence when compared to the control cells, indicating a loss of contact inhibition. NIH3T3 cells transfected with the FGFSa and FGFBe cDNAs were less morphologically transformed than the FGFSb cells, and were roughly equal in phenotype to each other (Figure 3.6). Like FGFBb cells the FGFBa and FGFBe cells displayed a lack of contact inhibition at confluence. These results correspond closely to those described above for mouse Fgfs. Since the deduced amino acid sequence of FgfBb is identical between mouse and human, it is not surprising that human FGFBb, like mouse FgfBb can induce morphological transformation in NIH3T3 cells. Mammary hyperplasia and adenocarcinomas in MMTV-FgfS transgenic mice To further assess the role of Fgf8 as an oncogene, we examined the consequences of overexpression of Fgf8 in transgenic mice. The Fgf8 transgene The transgene for the Fgf8 transgenic mice was directly cloned from a tumor (tumor B6) that had an MMTV insertion very close to the 5' end of the Fgf8 (Figure 2.4, Figure 3.7). This particular insertion is in a "promoter insertion" orientation, i.e., it is upstream from the initiation 81 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 3.6. Morphological transformation of NIH3T3 cells by human FGF8 isoform s. NIH3T3 cells were transfected with expression plasmids containing cDNAs of FGFSa, FGFBb, FGFBe or a control plasmid containing the vector pcDNA alone. G418-resistant colonies were pooled and grown to confluence to test for contact inhibition and morphology. All three cDNAs in the sense orientation produced loss of contact inhibition, and FGFBb also showed marked morphological transformation. 82 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 3.6. SUBCONFLUENT CONFLUENT E Al W m M 83 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 3 e u ■ o o Q . C g Q . T 3 CD C/) ( g o' 3 3 CD C p . CD - O O Q . C a O 3 ■ O O MMTV-Fgf8 Transgene B Xh Bg Xh B '§ ■ I 1 2a 2c 3a 1 . 4 , 5 2b 3b Ëi MMTV Fgf8 Figure 3.7. Structure of the MMTV-Fjgf» transgene. The transgene containing the entire FgfB genomic locusand the MMTV- LTR (3‘) was cloned from a mammary tumor(tumor 86). The six exons of FgfB are depicted as boxes, with the filled areas indicating the coding domains. % The alternatively spliced exons are indicated with hatched boxes. MMTV- LTR is indicated by w a shaded box. C/Î 2 codon in the same transcriptional orientation (Figure 1.2). This tumor expressed high levels of Fgf8 RNA (Figure 2.5), suggesting that the transgene is likely to express well. The transgene contains the 3' MMTV LTR, and the entire genomic Fgf8 gene in a single DNA fragment of -13 Kb. The DNA was size selected on an agarose gel and cloned into a lambda vector, and the resulting library was screened using both MMTV specific (env) and Fgf8 specific DNA fragments (XE 2.0) as probes. The 13 kb fragment was excised from this vector and used for microinjection of fertilized mouse eggs. Generation of Fgf8 transgenic mice Three transgenic founder animals were obtained that had comparable numbers of transgenic inserts by Southern blot analysis. These animals were bred to normal BALB/c mice in an attempt to generate independent lines. However, we were unable to generate lines from the two female founders (TG# 17, TG# 3), possibly because of their inability to nurse their offspring. An independent transgenic line was established from the other male founder (TG# 18). Expression of the transgene To ascertain the expression of the transgene in the mammary gland, salivary gland, and in the case of the males, testes (organs where MMTV promoter/enhancer is known to be active), northern analysis was performed on the RNAs from these tissues. Strong expression of FgfB mRNA was observed in the mammary glands from transgenics #17 and 3, while there was no expression seen in the salivary glands of these 85 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. animals (Figure 3.8). Fgf8 mRNA is not normally detected in the mammary glands of non-transgenic mice (northern analysis and RT- PCR), refer (chapter 2, and Figures 2.5 and 2.7. Expression of Fgf8 was not detected in any of the tissues from TG# 18, FI male or female offsprings (data not shown), indicating that despite the presence of the transgene we did not have a FgfB transgenic line that expresses it. MMTV-FgfS transgenic mice develop mammary gland hyperplasia and tumors The female Fgf8 transgenic mice (TG#17, #3), demonstrated mammary gland hyperplasia, and developed mammary tumors with a latency of 5-8 months. The incidence of tumor formation increased with the number of pregnancies. Tumor development was observed in several mammary glands of the transgenic animals. Histological analysis of the mammary glands from these animals showed pronounced ductal hyperplasia (Figure 3.9). The tumors on histological examination showed varying degrees of malignancy, ranging from benign adenomas to invasive ductal carcinomas (Figure 3.9). Analysis of the transgene expression in the tumors and a few tissues by northern analysis revealed high expression of Fgf8 in the mammary tumors when compared to the expression in the surrounding mammary gland or the mammary gland from non-transgenic animals (Figure 3.8, 2.7). Interestingly, a higher level of FgfB mRNA was also detected in the ovaries. Even though Fgf8 is known to be normally expressed in the ovaries of adult mice, the expression is quite low and cannot be detected 86 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Sg Mg Ov 1 2 Figure 3.8 Expression of the Fgf8 transgene in mammary tumors and some adult mouse tissues . Northern blot of total cellular RNA (10 mg) from three normal tissues (salivary gand (Sg, mammary gland (Mg), and ovary (Ov), and three tumors from FgfB transgenic mice (1,2, & 3). The blot was probed with a full length FgfSb cDNA (upper panel). Ethidium bromide stained photograph of the RNAs (lower panel). 87 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. & B Figure 3.9 Histopathology of the MMTV-FgfS mammary tumors. Histological section through fixed mammary gland tumors from MMTV-Fg/8 transgenic mice. A. benign adenoma (lOOx), B. invasive adenocarcinoma (100x). The sections were stained with hematoxylin and eosin. 88 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. using northern blots of total RNA (Chapter 2, Figure 2.7). Therefore, this result indicates that there is increased Fgf8 expression in the ovaries of the trangsenic mice, which presumably is due to transgene expression in that tissue. M MTV-FgfS mice develop stromal hyperplasia of the ovaries Because of the increased expression of Fgf8 observed in the ovaries, we examined the histology of the ovaries from these animals. Histological analysis demonstrated severe hyperplasia of the stroma surrounding the developing follicle and epithelium (Figure 3.10). This marked proliferation of the stromal fibroblasts, in consistent with the proliferation of the stromal cells observed in mammary gland. Discussion The existence of several FgfS isoforms suggests the possibility that they possess different biological functions. Our results show that three of the FgfS isoforms have different potencies for transformation of NIH3T3 cells. We have extended this study and analyzed all the seven isoforms for their differences in NIH3T3 transforming activities. Like FgfSb, isoforms that possess the extra amino acid sequences found in FgfSb but not FgfSa, are the most capable of transforming the NIH3T3 cells. FgfSb cells are morphologically transformed (Figure 3.3), clonagenic in soft agar (Table 3.1), and rapidly form tumors in nude mice. FgfSa and FgfSc cells show modest levels of transformation (Figure 3.3), are not clonagenic in soft agar, and are weakly tumorigenic in nude mice. RNase protection analyses indicate that the cell lines produce the correct 89 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 3.10 Pathology of the ovaries from MMTV-Fgffl transgenic mice. Histological section of an ovary from Fgf8 transgenic mouse showing extensive stromal hyperplasia (100x magnification). The sections were stained with hematoxylin and eosin. 90 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. FgfS isoform RNA, and suggest that the correct FgfB protein is produced by the cell lines. Hence the differences in potency of NIH3T3 cell transformation is not due to differences in the amount of FgfS isoform produced in the cell lines. Alternative isoforms of FGF2 and FGF3 exist due to alternative initiation sites (Acland et ai, 1990; Florkiewicz and Sommer, 1989), which results in the targeting of these isoforms to different cellular locations (cellular versus nuclear). In contrast, the three FgfB isoforms examined have identical signal peptides, are presumably secreted, and differ only at the amino termini of the mature peptides (MacArthur et a/., 1995). Assuming that the three FgfB isoforms are secreted, we would predict that their biological effects would relate to binding of FGF receptors (FGFR). Hence, the differences observed In the NIH3T3 cell transformation potency between FgfB isoforms suggest that the amino terminal differences in the FgfB isoforms result in the differential ability of the FgfB isoforms to bind to, or induce signals through, the FGFRs present on NIH3T3 cells. MacArthur at ai, have shown that the isoform FgfBb can activate the 'c' splice form of FGFR2, FGFR3, and FGFR4, while FgfBc can activate the 'c' splice forms of FGFR3 and FGFR4 (MacArthur at ai, 1995). FgfBa on the other hand does not activate any of the known FGFRs. These results support the hypothesis that the differences in the biological activities of the FgfB isoforms could be a result of differences in the receptor specificity of the isoforms. The temporal and spatial expression of F g f8 during postgastrulation mouse development show FgfB to be expressed in the 91 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ectoderm of the first branchial arch, nasal pits, limb buds, as well as in the neuroectoderm of the telencephalon, diencephalon., mesencephalon-metencephalon junction, and the infundibulum from days 9 to 13 of development. Analysis of the expression of the different isoforms during days 10-12 of mouse development detected RNAs of all the different isoform species (Crossley and Martin, 1995; Heikinheimo et a/., 1994; Ohuchi et al., 1994). These results suggest that splicing of the transcript is not regulated during development. If this is the case, then the specific interactions of isoforms with FGFRs would depend on the FGFRs present in the local environment of expresssion. The analysis of the differences in the biological activities of human isoforms on NIH3T3 cells reiterates the results seen with the mouse isoforms. The FGF 8b isoform is highly transforming, while FGFSa and FGFBe induce moderate degrees of transformation. As discussed before, the variation in the biological effects of the isoforms may be related to their specificity or binding affinity to FGF receptors. Considering the sequence similarities between the mouse and human FGF genes, we could speculate that the human FGFs also possibly activate the receptors activated by their mouse homologs. Therefore, to better understand the functional differences between the FGF isoforms, it will be necessary to analyze the isoform/receptor interactions and their signal transduction mechanisms more thoroughly. MMTV-FgfS transgenic mice. We generated transgenic mice expressing Fgf8 in their mammary gland under the control of MMTV-LTR. We cloned the transgene from an original tumor with a MMTV insertion in 92 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. this locus, to recapitulate the event that had occurred in the tumors. Analysis of the transgenic mice that expresses the transgene, showed the development mammary gland hyperplasia. Mammary tumors developed with a latency of 5 - 8 months, with the tumors ranging from benign adenomas to invasive ductal carcinomas. These results therefore indicate that Fgf8 is indeed an oncogene that is capable of transforming the mammary epithelium. However, the development of tumors after a prolonged period is consistent with a number of previously described studies. In these studies, different oncogenes have been expressed in the mammary gland of transgenic mice. The majority of these animals showed patterns of tumor development similar to Fgf8 transgenics: hyperplasia followed by adenocarcinomas after a prolonged latency (Chapter 1, and Table 1.2). These results suggest that overexpression of Fgf8 contributes to but is probably not sufficient to cause complete transformation. Overexpression of FgfB in the ovaries of MMTV-FgfS transgenic mice. We have demonstrated that the levels of expression in the transgenic ovaries was significantly higher than normal. Additionally, these animals developed stromal hyperplasia of their ovaries, indicating proliferation of the stromal fibroblasts in response to the mitogenic stimulus. The overexpression of Fgf8 in the ovaries is probably due to the targetting of the transgene by ovary-specific enhancers present in the genomic DNA used to construct the transgene. Another possibility is a secondary metastasis from the mammary tumors, but this seems less likely since both ovaries showed the same level of 93 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. hyperplasia. It Is interesting to note how Fgfd can stimulate proliferation of two different cell types. This could possibly be explained by the presence of different FGFreceptors in the two types of cells, and hence Fgf8 could possibly have a mitogenic effect on both epithelial as well as fibroblast cells. The appearance of mammary tumors along with the involvement of the ovaries, is reminiscent of the clinical features of some human breast cancers (Giannios and loannidou-Mouzaka, 1997). Since both these organs (breast and ovaries) are under the influence of hormonal variations, and there exists a correlation of hormonal status to cancer, it is reasonable to speculate that perhaps the ovarian hyperplasia is a result of metastasis from the primary breast tumor. Thus these transgenic mice could serve as an effective model system to study mammary tumorigenesis, and could be used to understand the relationship between hormones and growth factors in promoting cancer. FgfS and human cancers. The fact that Fgfd is a mammary oncogene in mice, and the detection of mRNA (human) expression in prostatic epithelial cell lines, presents a rationale for asking whether it could have a role in human cancers. One study of human prostate cancer showed a significant up-regulation of expression in high-grade prostate tumors, while the surrounding stroma and endothelium were negative for expression (Leung et al., 1996). Preliminary results of a study of human breast tumors also seem to show increased expression (Kapoun and Shackleford, unpublished). These results strongly suggest a role for FgfS in human malignancy, however more studies with larger sample sizes are necessary before any inference can be made. 94 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter 4 Apoptosis Of Mammary Epithelial Cells Induced By Fibroblast Growth Factors Introduction FgfS is the third member of the FGF family of growth factors to be activated by MMTV in the Wnt1 transgenic mammary tumorigenesis model. As described in the Chapter 3, FgfB encodes at least seven protein isoforms three of which (FgfSa, FgfBb, FgfSc) have been isolated and characterized (Figure 3.1)(Crossley and Martin, 1995; MacArthur et al., 1995). The FgfB isoforms exhibit differences in their NIH3T3 cell transforming potentials, with FgfBb (mouse and human) being the most potent transforming protein (MacArthur at a!., 1995). Transgenic mice expressing FgfS in their mammary glands from a MMTV-LTR develop mammary gland hyperplasia followed by the development of adenocarcinomas (Chapter 3). These studies {in vitro and in vivo) clearly demonstrate the oncogenic capability of FgfS, especially the isoform FgfBb. Since we cloned FgfS from mammary tumors, I was interested in testing the differences in the biological effects of FgfB protein isoforms on normal mammary epithelial cells. During this process, we serendipitously discovered that overexpression of FgfBb or stimulation of mammary epithelial cells with FgfBb protein resulted in programmed cell death (apoptosis) of the cells. This was indeed an intriguing finding. 95 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. since this property has previously not been described for any of the known fibroblast growth factors. Apoptosis, or programmed cell death, is the process by which a cell will actively commit suicide under tightly regulated circumstances. The execution of this death program is associated with characteristic morphological and biochemical changes (Wyllie et al., 1980). During apoptosis, the nucleus and the cytoplasm condense, and the dying cell often fragments into membrane-bound apoptotic bodies that are rapidly phagocytosed and digested by macrophages or neighboring cells. In contrast, during necrosis, a pathological form of cell death that results from cellular injury, cells swell and lyse, thereby releasing cytoplasmic material which often triggers an inflammatory response (Novak, 1996; Vaux and Strasser, 1996). Apoptosis is associated with the activation of specific proteases (caspases), and nucleases that degrade the chromosomal DNA, first into large fragments (50-300 kilobases), and subsequently into multiples of 180 bp corresponding to nucleosomal multimers (Deiss et al., 1995; Evan et al., 1992; Garcia-Martinez et al., 1993; Gavrieli et al., 1992; Guchelaar et al., 1997; Haas-Kogan et al., 1995; Harrington et al., 1994; Heermeier et al., 1996; Hockenbery et al., 1991; Hoffman and Liebermann, 1994; Kim et al., 1997; Kim et al., 1995; Kyprianou and Isaacs, 1989; Leung et al., 1996; Li et al., 1996; Lu et al., 1995; Naik et aL, 1996; Oberhammer et aL, 1993; Raff, 1992; Raff et al., 1993; Rowan and Fisher, 1997; White, 1996; Wyllie et al., 1980) Detection of the broken ends of the DNA, the 180bp "ladder" seen upon electrophoresis of DNA from apoptotic cells, or reduction of the total DNA content , indicates that a dying cell's DNA is being degraded, and are 96 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. commonly used experimental methods for detecting apoptosis (Basnakian and James, 1994; Garcia-Martinez et a!., 1993; Gavrieli et al.. 1992; Oberhammer eta!., 1993). Apoptosis is a key biological regulatory mechanism that is conserved through evolution. It serves as a major process in sculpting the developing organism (Novak, 1996; Rowan and Fisher, 1997; Vaux and Strasser, 1996), as a major mechanism for the precise regulation of cell numbers (Raff, 1992; Raff et a/., 1993), and as a defense mechanism to remove unwanted and potentially dangerous cells, such as self reactive T cells (Guchelaar et a!., 1997; Jehn and Osborne, 1997), cells infected by viruses, and tumor cells. Many different signals that may originate either from within or outside a cell have been shown to influence the decision between life and death. These include lineage information, cellular damage inflicted by ionizing radiation or viral infection, extracellular survival factors, cell interactions, and hormones (Hoffman and Liebermann, 1994; Rowan and Fisher, 1997). These diverse signals may act either to suppress or promote the activation of the death program, and the same signal may have opposing effects on different cell types (Hoffman and Liebermann, 1994; Rowan and Fisher, 1997). In addition to the beneficial effects of programmed cell death, the inappropriate activation of apoptosis may cause or contribute to a variety of diseases, including autoimmune diseases, neurodegenerative diseases, and ischemia. Thus, too much or too little cell death can have catastrophic consequences (White, 1996). 97 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. In the last couple of years tremendous progress has been made in elucidating the process of apoptosis. This has led to the discovery of several intersecting pathways that are controlled by positive and negative regulators (Duckett and Thompson, 1997). A general scheme of events that occur during programmed cell death can be divided into phases. The earliest phase is the stimulus that provokes the apoptotic response. This may be an external signal delivered through surface receptors or may originate inside the cell from the action of a drug, toxin or radiation (Jehn and Osborne, 1997; Rowan and Fisher, 1997). The next phase includes detection of this signal or metabolic state and transduction of the signal. The effector phase is the third part of the cell death program and includes the proteases (caspases) that are activated during apoptosis as well as their positive and negative regulators (Bcl2 family) (Blandino and Strano, 1997; Dietrich, 1997; Guenal et al., 1997; Rowan and Fisher, 1997; Yuan, 1997). The fourth phase of cell death is the postmortem phase, in which the cell’s chromatin condenses and its DNA is degraded (Rowan and Fisher, 1997; White, 1996). Several gene products have been identified that either activate or suppress apoptosis. Both tumor suppressor genes like p53 and Rb (Berry et al., 1996; Duckett and Thompson, 1997; Haupt et al., 1997; Vaux and Strasser, 1996), and oncogenes {Bcl2, cmyc, El A, growth factors and receptors) play an active role in regulation of apoptosis. Growth factors, in general, are thought to be survival factors and are known to inhibit apoptosis (Chiarugi and Ruggiero, 1996; Harrington et al., 1994; Hoffman and Liebermann, 1994), Growth factor withdrawal 98 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. is a common mechanism of inducing ceil death (Merlo et a!., 1995). However, examples of growth factors and cytokines that are mitogenic for certain cell types but induce apoptosis of others exist. They include TG F/3, TNFa, EGF, PDGF, and y interferon (Armstrong et ai, 1994; Brabyn et ai, 1994; Deiss et ai, 1995; Hanahan, 1989; Kim et ai, 1997; Kim et ai, 1995; Kyprianou and Isaacs, 1989) FGFs, on the other hand have always been implicated as survival factors and are known to suppress apoptosis in several systems (Macias et ai, 1996; Raff, 1992; White, 1996). Our results describe a previously unidentified property for fibroblast growth factors. In addition to Fgf8b, several members of the FGF family of growth factors also exhibit this apoptosis inducing property. Overexpression of the apoptosis regulator BCL2 in these mammary epithelial cells delays the onset of FGF induced programmed cell death, indicating that this is probably a BCL2 dependent apoptotic pathway. Hence, from the view point of oncology,it makes it very interesting and important to study programmed cell death and the genes involved, so that one can manipulate it to control cancer and also other human diseases like neurodegenerative diseases. Materials and Methods Cell lines, Vectors, and Transfection. C57MG cells (obtained from A.M.C. Brown), MCF7, BMG, GR, SC115, and HC11 were grown in DMEM supplemented with 10% FGS and 10 ng/ml insulin. NIH3T3 cells and derivative cell lines (MIRB, 8a, 8b, 8c) were grown as described chapter 3. 99 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The pMIRB vector and the pMIRB vector containing the Fgfd cDNAs are the same as those described previously (Figure 3.2). The human BCL2 expression vector (pSFFV-Bcl2) was a generous gift of S. Korsemeyer. The control vector pSFFV was constructed by removal of the BCL2 cDNA (EcoRI digestion) and religation of the plasmid. All transfections were performed using the lipofectamine protocol and selected in G418 (Geneticin) as previously indicated (chapter 3). The BCL2 stable transfectants were cloned to generate independent cell lines, and each clone was analyzed for expression. A cell line of pooled BCL2 transfectants was also generated. Conditioned Medium Transfer. NIH3T3 cells expressing Fgfdb (8b cells), or the control vector (MIRB cells) were grown in DMEM supplemented with 10% FCS. When the cells were approximately 60- 70% confluent, the medium was changed to DMEM supplemented with 0.1% FCS and 0.1 (ig/ml insulin. Twenty-four and forty-eight hours after the shift, the conditioned medium was collected and either centrifuged or filtered through a cellulose acetate membrane (pore size lOum; Nalgene), and used to treat C57MG or other mammary epithelial cell lines (Figure 4.1). Hoechst Staining. Transfected or treated mammary epithelial cells were cultured on chambered slides (Nunc, Inc.), or grown on coverslips. Four days after treatment or shift to low serum, the cells were gently washed in PBS, and fixed with carnoyls fixative (3:1 methanol to acetic acid). After fixation, the cells were stained with 0.5 mg of the fluorochrome Hoechst 33528 (Sigma chemicals) for 30 minutes. The 100 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 4.1 Schematic Representation of Conditioned Medium Transfer NIH3T3 conditioned medium y / I C57M 3 APOPTOSIS: - Phase contrast microscopy: morphology - Hoechst 33258 staining : Nuclear condensation and fragmentation - Agarose gel electrophoresis (DNA): oiigonucleosomal laddering 101 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. slides and coverslips were gently washed with PBS (2x), mounted, and observed through fluorescent photo-microscopy (Nikon. UV-B filter). DNA Extraction and Analysis of DNA Fragmentation. DNA was extracted from 3-5 x 10® cells by incubation in 300 pi digestion buffer containing 100 mM NaCI, 10 mM Tris-HCI (pH 8.0), 25 mM EDTA, 0.5% sodium dodecyl sulfate, and 0.3 mg/ml proteinase K for 15 h at 50°C. The DNA was purified with phenol/chloroform and the aqueous phase was extracted twice with diethyl ether to remove residual phenol. The RNA was digested with 1 pg/ml DNase free RNase for 1 h at 37°C. The DNA was precipitated with 100% ethanol, washed with 70% ethanol, air dried, and resuspended in TE buffer (Tris-HCI, pH 8.0; 1 mM EDTA). Approximately 1 ug of DNA was subjected to electrophoresis on a 0.8% agarose gel and visualized with ethidium bromide staining under UV illumination. Heparin-Sepharose A ffinity Column. Fifty ml of the conditioned medium from MIRB or 8b cells were applied to pre-made heparin-sepharose columns (Pharmacia Corp.), equilibrated with 10 mM phosphate buffer containing 0.1 mM NaCI (pH 7.2). The flow-through was collected and stored at 4°C for later use. After a wash with 50 ml of the equilibrated buffer, proteins bound to the column were eluted with a gradient of 0.1 to 2.0 M NaCI in 10 mM phophate buffer. Each elution fraction was then dialyzed against 2 L of DMEM with no additions for 24 hours. The conditioned medium, flow-through, and elution fractions were used to treat C57MG cells (Figure 4.2). 102 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CD ■ D O Q. C g Q. ■ D CD C / ) C / ) Figure 4.2 Heparin-sulphate Sepharose Chromatography Vector FCF8b 8 ( O ' 3. 3 " CD CD " O O Q. O 3 ■ D O CD Q. ■ D CD ( / ) o' 3 Conditioned Medium (CM) NIH3T3 -G. r - £ -C I Hcparin-Sulphate Column U u u U i CM FT IM 1.5M 2M L / i Eluatcs CM FT IM 1.5M 2M (NaCI) o w Treat C57MG cells FGF Proteins, Heparin, A ntibodies and Treatm ent Protocol. Purified recombinant human FGF1, 2, 4, 5, 6, 7, and 9 were purchased from R&D systems, IL. The lyophilized proteins were reconstituted as per the manufacturer's instructions, and stored as aliquots of 10 M -g/m l (lOGx) stocks at -8G°C. Heparin (Sigma) was stored as a 10G ^g/ml (IGOx) stock solution at 4°C. The anti-human basic FGF antibody was obtained from R&D systems. The lyophilized antibody was reconstituted in sterile PBS and stored at -80°C. The antibody was used at a concentration of 5x the ND5G to get complete neutralization. The ND50 was determined using the manufacturer's instructions. The mammary epithelial cells were grown in DMEM with 10% FCS, and 10 pg/ml insulin. When the cells were about 40 - 60% confluent the medium was removed, the cells were washed and shifted to DMEM containing 0.1 % FCS and 0.1 pg/ml insulin. The FGF protein was added at a concentration of 100 ng/ml (or other concentrations depending on the experiment) with 1 ng/ml heparin. Results Overexpression of FgfSb, or treatment of the mammary epithelial cell line C57MG with FgfBb results in programmed cell death. Because we identified FgfS as an oncogenic collaborator of Wnt1 in mammary tumors, and the fact that it was initially cloned from a mammary carcinoma cell line (Tanaka et al., 1992), we sought to test the 104 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. biological and tumorigenic effects of Fgf8 protein isoforms on mammary epithelial cells. In order to determine the oncogenic potential of Fgf8 isoforms in mammary epithelial cells, the cDNAs encoding FgfSa, FgfSb and FgfSc isoforms of the protein were cloned into the expression vector MIRB and transfected Into a C57MG cells, and the cells were observed for the formation of discrete foci. The cells transfected with the FgfSb cDNA started to form foci on reaching confluence, indicating that the cells expressing the Fgf8b protein were not contact inhibited and hence continued to divide. However, the cells within the focus surprisingly showed classical signs of apoptotic death; the cells were rounding up and floating in the medium, and eventually all the ceils in the focus detached from the plastic leaving open spaces in the monolayer (Figure 4.3A). C57MG cells transfected with FgfSa or FgfSc did not form foci nor showed any morphological changes. In a parallel experiment, conditioned medium from NIH3T3 cells expressing FgfSb (Sb cells), and pMIRB (MIRB cells) was used as source of FgfS proteins. These cells have been previously shown to secrete biologically active FgfSb protein into the conditioned medium and to induce morphological transformation of NIH3T3 cells when treated with the 8b conditioned medium (MacArthur et al., 1995). The conditioned medium from Sb cells, or MIRB cells as a control, was collected from actively dividing cells, filtered and transferred onto C57MG cells at 40% confluence. Mammary epithelial cells treated with the FgfSb conditioned medium showed molecular transformation from a flat cuboidal 105 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 4.3. Overexpression of FgfSb in mammary epithelial cells (C57MG) results in apoptosis. A. Morphology of C57MG cells transfected with the expression vector pMIRB containing the FgfSb cDNA. The cells were not selected in G418, grown to confluence and observed for focus formation. The figure shows a typical focus in a dish transfected with FgfSb cDNA. Note the presence of dying cells leaving open spaces in the monolayer. B. Staining of the focus with Hoechst 33258, showed the presence of condensed and fragmented apoptotic nuclei within the focus in sharp contrast to the uniform nuclear morphology of the surrounding monolayer (100x magnification). 106 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 4.3 107 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. morphology to a more spindle shaped elongated morphology within 16 hours. By day 3 after treatment, many of the cells rounded up and detached from the plate and showed classical signs of cell death which was complete by day 4 (Figure 4.4A). Cells treated with MIRB cell conditioned medium remained flat and cuboidal. Based on these observations of cell death, we decided to determine the possibility of apoptosis (programmed cell death), since the morphology of the cells showed characteristic features of apoptotic cell death. To test for apoptosis. the treated cells and the foci were fixed and stained with Hoechst 33258 (a fluorochrome that stains the nuclei) to look for chromatin condensation and nuclear fragmentation. In addition, DNA was extracted from the treated cells and tested for DNA laddering by agarose gel electrophoresis. Cells transfected with the vector alone or treated with conditioned medium from cells expressing the vector alone were used as controls. Staining of the foci with Hoechst 33258 showed the presence of fragmented nuclei and condensed chromatin more intensely fluorescent than the normal nuclei of the cells in the monolayer (Figure 4.3B). Similarly, the nuclei of cells treated with the 8b conditioned medium showed classical apoptotic nuclear morphology, exhibiting varying degrees of chromatin condensation and nuclear fragmentation, in sharp contrast to the control cells (Figure 4.4B). DNA laddering experiments showed the cleavage of the DNA from the 8b treated cells into the classical ~180bp DNA ladders while the DNA from the control cells only showed the presence of high molecular weight genomic DNA (Figure 4.4C). DNA laddering is a very strong evidence for apoptosis. 108 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 4.4. C57MG mammary epithelial cells treated with FgfSb containing conditioned medium undergo programmed cell death. A. Morphology of C57MG ceils treated with either FgfSb containing conditioned medium or conditioned medium from cells expressing MIRB alone. Cells treated with FgfBb conditioned medium showed extensive cell death within 64-72 hours of treatment, in contrast to the MIRB treated cells which did not undergo any morphological changes. B. Staining of the C57MG cells treated with FgfSb or MIRB conditioned medium with Hoechst 3325S. The nuclei of the cells that received the FgfSb conditioned medium showed characteristic apoptotic morphology: condensation and fragmentation of the nucleus, while the cells treated with MIRB conditioned medium showed the typical uniform nuclear morphology of G57MG cells. C. Oiigonucleosomal laddering. DNA was extracted form the cells treated as described above and subjected to agarose gel electrophoresis. Note the presence of the characteristic 180 bp DNA "ladder" in the lanes containing DNA from FgfBb treated cells 109 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. £ 3 g 110 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. n 00 O) u_ "O 0 ) 3 g c 0 > Ü . 2 3 .P » G Q O C m 1 11 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 4.4 (continued) 12 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. These results suggest that stimulation of mammary epithelial cells with FgfSb protein signals the cells to enter the apoptotic pathway. Depletion of FgfSb from the conditioned medium by binding to heparin-sepharose, abrogates FgfSb induced apoptosis Since this was the first observation so far that FGFs could probably cause apoptosis of mammary epithelial cells, to better test our results, we partially purified FgfSb using a heparin-sepharose affinity column. It has been previously shown that FgfSb like other FGFs can bind heparin proteoglycans, and FgfS was Initially purified by binding to heparin. The main goal of this experiment was to deplete the conditioned medium of FgfSb by binding it to a heparin sulfate column and to partially purify the FgfSb protein by elution from the column. NIH3T3 cells were used as a control in this experiment to test the activity of the conditioned medium and elution fractions. Both cell lines were treated with conditioned medium (Sb, vector control), flow-through, and fractions 1 through 4 (0.5M, 1M. 1.5M and 2M NaCI concentrations). C57MG cells treated with the conditioned medium went through programmed cell death (tested by Hoechst and laddering). This activity was lost in the flow-through (conditioned medium after passing through the heparin-sepharose column), indicating that the factor inducing apoptosis bound heparin, and was also responsible for transformation of NIH3T3 cells, since the flow-through did not transform NIH3T3 cells. Both the transformation activity in NIH3T3 cells and apoptotic activity in C57MG cells was restored when the cells were treated with the 1.5M NaCI fraction from FgfSb but not from vector control. This fraction is most 113 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 4.5. Binding of the FgfSb protein from the conditioned medium to heparin-sepharose affinity column, abolishes the induction of apoptosis. The conditioned medium from either FgfSb or MIRB cells were subjected to heparin-sepharose affinity chromatography as described in Materials and Methods, upper panel. The C57MG cells treated with FgfSb conditioned medium (CM) showed apoptosis, this effect was abolished on passing through the affinity column (FT). This apoptotic activity was restored in the 1.5M NaCI fraction, lower panel. C57MG cells treated with control (MIRB) conditioned medium and elution fractions. No morphological changes were observed in these cells. 114 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. C J R z. I g z: « 0 U R z c a œ u < O Is. U o b t o U 115 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. likely to contain the FgfSb protein since these results corroborate the previous finding where the FgfSb protein was eluted at a concentration of 1.1M NaCI. These results suggest that the apoptosis induction is probably a direct consequence to the FGFSb signal. Several members of the FGF family of growth factors can induce apoptosis of mammary epithelial cells To test if this phenomenon was specific to only FgfSb or was also seen with any other FGF family member(s), we used purified FGF1, 2, 4, 5, 6, 7, and 9 (R&DSystems) on C57MG cells in a similar experiment. 100 ng/ml of the FGF protein was added to the cells at 40%-60% confluence with 0.1 [ig/ml of heparin. NIH3T3 cells were used as a control to assay the activity of the protein. Cells treated with FGF1, FGF2, FGF4, FGF6 and FGF9 showed morphological transformation and apoptotic cell death similar to FgfSb. FGF2, FGF4 and FGF9 showed this effect even in the absence of heparin. FGF5 did not show any changes at concentrations of 200 ng/ml. Cells treated with 500 ng/ml of FGF5 showed moderate transformation but no apoptosis, and cells treated with 1 pg/ml of the growth factor showed both transformation and apoptosis. The only tested growth factor that did not show any activity was FGF7. Even 1 [ig/ml of FGF7 did not induce any changes in morphology. Neutralization experiments using a polyclonal antibody against human FGF2 completely blocked the transformation and apoptosis seen in FGF2 treated C57MG cells (Figure 4.6). The cells treated with FGF2 that was preincubated with the antibody did not show any alteration in 116 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. TABLE 4.1. Apoptosis of mammary epithelial ceils is induced by several members of the FGF family FGF Transformation of Nffl3T3 cells Apoptosis of mammary epithelial cells FGFl FGF2 FGF4 FGF6 FGF8 FGF9 +++ +++ + + + + + + +++ +++ +++ +++ +++ +++ + + + +++ +++ +++ 117 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 4.6. Neutralization of FGF2 induced apoptosis of mammary epithelial cells by an anti-hum an FGF2 antibody. C57MG cells were treated with FGF2 protein, anti-FGF2 antibody, or FGF2 protein preincubated with the FGF2 antibody. The cells treated with the neutralized protein showed no evidence of transformation or apoptosis in contrast to the cells treated with the FGF2 protein alone. 18 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. >• il 1 1 o C M ■ * “ il c « + (O £ 3 g . ai c o o C M LL O U . 119 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. cellular and nuclear morphology, Indicating that the observed apoptosis was indeed a result of treatment with FGF2. FGFs can induce programmed cell death of several mammary epithelial cell lines. To address the possibility that the apoptotic effect was not a peculiarity of the cell line, we tested several other mammary cell lines (normal and tumor) for induction of apoptosis by FGF2, FGF4, FGF6 and FgfSb. Three cell lines (C127I, BMG, GR) resembled C57MG in all respects; cellular transformation followed by apoptosis. HC11 cells showed some degree of apoptosis but cell death was not complete as seen in the other cell lines. The human breast carcinoma cell line MCF-7 when treated with the different FGFs did not undergo programmed cell death. One explanation to this non-responsiveness to FGF signals is that MCF-7 cells are known to express high levels of the anti-apoptotic gene product BCL2. Therefore, BCL2 may be inhibiting or blocking the apoptotic signal provided by the FGF. The onset of FG F-induced apoptosis is delayed by overexpression of BCL2 The BCL2 protein is known to block apoptotic cell death induced during development and by external factors like radiation (Blandino and Strano, 1997; Guenal et ai, 1997; Kirshenbaum and de Moissac, 1997). It is also known to inhibit m yc induced apoptosis of fibroblasts (Bissonnette et ai, 1992; Evan et a i, 1992). However, BGL2 120 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 4.2 FGFs induce apoptosis of several mammary epithelial but not fibroblast cell lines Mammary Epithelial Cells Apoptosis C1271 +++ HC11 +/- BMG +++ GR +++ S115 MCF-7 Fibroblasts NIH3T3 Rat-1 121 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. independent apoptotic pathways are also known (Chinnaiyan and Dixit, 1996; Hoffman and Liebermann, 1994). To test if BCL2 could inhibit FGF induced apoptosis of mammary epithelial cells, C57MG cells were transfected with an expression vector containing the BCL2 cDNA. Stable clones expressing the BCL2 gene product were generated and treated with purified FGF2, FGF4, FGF6, and with FgfSb conditioned medium as described in the previous sections. Stable clones expressing the vector alone were used as controls. The control cells (clones transfected with vector alone) showed morphological changes within 12-16 hours after treatment. By 24 hours many cells started dying and detaching from the plate. By day 3 after treatment approximately 80% of the cells were apoptotic. In contrast, the BCL2 clones remained viable and were morphologically transformed from a flat cuboidal morphology to a more elongated spindle shaped appearance (Figure 4.7). No apoptosis was observed, suggesting that BCL2 can block the apoptotic signal from FGF. However, on days 4 and 5, when the control cells were completely dead, the BCL2 clones started showing signs of apoptosis, indicating that overexpression of BCL2 does not completely block apoptosis, but delays the onset. This lack of complete inhibition of apoptosis by BCL2 could be explained by the possibility of transient expression of the protein from the vector or poor stability of the protein. However, these results are similar to other reports which indictes that Bcl2 does not always completly block induction of apoptosis, but in many instances postpones or delays the onset (Guenal et al., 1997; Hockenbery et al., 1991; Stresser et al., 1996; Vaux and Stresser, 1996) 122 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 4.7. Ectopic expression of BCL2 in mammary epithelial cells delays the onset of FGF induced apoptosis. Three clones and a pool of C57MG cells transfected with an expression vector containing BCL2 cDNA were treated with several FGF proteins (treatment with FGF2 shown here). Note that the clones expressing BGL2 are morphologically transformed but show no evidence of apoptosis in contrast to the complete apoptosis seen on treatment of the control cells (cells transfected with the pSFFV vector). The pool of BCL2 expressing cells showed some apoptosis upon FGF treatment. 123 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. N T » 2 a !■ 0) c o o C M O 00 C M o c o o C M O O Q 124 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Discussion We have shown that the stimulation of mammary epithelial cells with either mouse FgfSb or human FGFl, 2, 4, 6, and 9. induces programmed cell death (apoptosis) (Figure 4.3 and 4.4). Mammary epithelial cells treated with conditioned medium depleted of FgfSb protein or with FGF protein neutralized with a specific antibody no longer undergo apoptosis, indicating that induction of apoptosis is a FGF-spacific response (Figure 4.5). In addition, several mammary epithelial cells (other than C57MG) exhibit the same response to the FGF signal. Therefore, cur data suggest that induction of apoptosis is a direct consequence of FGF stimulation and is not a cell line specific phenomenon. Further, we have also shown that overexpression of the BCL2 protein in these cells delays the onset of apoptosis and allows the progression of the mitogenic signal (observation of cellular transformation) (Figure 4.7). These results describe a new property for several members of the FGF growth factor family; induction of cell death, and reiterates the complexity and dynamic properties of these growth factors and their ability to modulate different responses in different cell types or micro environments (Mason, 1994). Fgfs and Apoptosis Fibroblast growth factors are a family of structurally related polypeptides with a wide range of functions. In vivo and in vitro studies show FGFs to modulate a variety of developmental processes. They are mitogens for cells of all three germ layers and have been implicated in 125 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. mesoderm induction, vascularization/angiogenesis and in the development of muscle, teeth and limbs (Doniach. 1995; Mason, 1994). They also induce differentiation and stimulate neurite outgrowth. Some of the FGFs have important roles in hair production, epithelial branching morphogenesis (Montesano et al., 1991), and lung development. Therefore it is not surprising that FGFs could also induce apoptosis, a major pathway by which an organism develops several organs, maintains homeostasis, and guards against pathogens and cellular damage (Vaux and Strasser, 1996; White, 1996). A classical example of induction of apoptosis by growth factors that are known to be mitogens is transforming growth factor p (TGFp) (Kyprianou and Isaacs, 1989; Macias et a!., 1997) TGFp is known to suppress growth of mammary epithelial cells and induce apoptosis of hepatic cells(Kyprianou and Isaacs, 1989; Li et al., 1996). In addition, there have been reports of PDGF mediated apoptosis of a kidney cell line (Kim et a!., 1997; Kim et a!., 1995), EGF induced growth suppression and cell death (Armstrong et a!., 1994; Brabyn et a!., 1994), and induction of apoptosis by y Interferon (Deiss et a!., 1995). Therefore it is plausible that FGFs could also mediate similar responses in different cellular environments. An indirect evidence that FGFs could be involved in apoptosis comes from the study of the HIVI-Tat protein. HIVI-Tat protein has been reported to be a heparin binding molecule with homology to angiogenic factors FGFs, and VEGF (Albini at a!., 1996). Several investigators have shown HIVI-Tat to induce apoptosis of uninfected T- lymphocytes. Therefore, one could argue that since FGFs have a similar 126 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. structure and function (binding to heparin and angiogenesis), they might also induce apoptosis by similar mechanisms. It is also interesting that the FGF-induced cell death takes at least 64-72 hours after the initial signal. This suggests that although promoted by FGFs, the exact moment of commitment to apoptosis of any individual cell is probably determined by other unidentified factors. Previous studies have suggested three basic mechanisms that account for cell death: first, direct initiation of apoptosis by an external signal; second, death due to the absence of trophic factors; third apoptosis as a response to conflicting signals (Raff, 1992; White, 1996). An example of the third mechanism is the apoptosis induced by cmyc (Bissonnette et al., 1992; Evan et al., 1992). Constitutive expression of cmyc in fibroblasts combined with a block to cell proliferation (low serum conditions) results in apoptosis. Our finding that overexpression of FGFs in mammary cells induces cell death could also be a result of conflicting signals. Alternatively, induction of apoptosis could possibly be a normal function for FGFs in the development of the mammary gland. The differences in the spatial and temporal expression of the FGFs during different stages of mammary gland development and the fact that some FGFs do not induce apoptosis (FGF5, 7) allude to this. Role of FGFs in the development of the mammary gland Normal development of the mammary gland depends on a combination of systemic mammatrophic hormones as well as local cell cell interactions (Lund at al., 1996). These interactions are mediated by a variety of growth factors, including members of the TGFp, EGF, Wnt and 127 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. FGF families. The differential expression of these growth factors in the developing breast suggests that they may act in concert with systemic hormones during mammary gland development, pregnancy, lactation and involution (Jhappan et al., 1990; Li eta!., 1996; Lund et a!., 1996). During pregnancy, the mammary gland begins a fascinating cycle of lobulo-alveolar development and maturation that finally results in full functional differentiation resulting in production of the milk by the differentiated secretory epithelium during lactation. After cessation of lactation, the mammary gland undergoes remarkable morphological, ultrastructural, and biochemical changes, collectively known as the process of involution. This phase is characterized by extensive death of fully differentiated epithelial cells by apoptosis, followed by extracellular matrix remodeling and apoptosis of cells losing differentiated functions (Lund et al., 1996). In addition, alterations in the breast also occur during the menstrual cycle, and the defined epithelial cell number (after the proliferative phase) is maintained by programmed cell death (Li et al., 1996; Lund et al., 1996). Thus, it is apparent that several stages of mammary gland development as well as phases of the normal breast during hormonal cycles, are characterized by apoptosis. Several growth factors secreted by the epithelium and the stroma have been identified to function together with the hormones in regulating these complex developmental processes. Both TGFp and FGFs have been shown in vivo to regulate ductal development, and are implicated in the mesenchymal-epithelial interactions during glandular development (Li et al., 1996; Lund et al., 1996; McLeskey et al., 1996). Therefore, it is reasonable to speculate 1 28 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. that FGFs could also play a role in the molecular mechanisms involved during involution. One such mechanism could be induction of apoptosis. FGFs: tumorigenesis versus apoptosis What would be the evolutionary sense in having FGFs or other proto-oncogenes {cmyc) drive both proliferation and programmed cell death, two apparently contradictory processes? The role of FGFs as mitogens is clearly established (Mason, 1994). It is also evident that FGFs are sufficient for cell proliferation and are proto-oncogenes (Basilico and Moscatelli, 1992). This is potentially very dangerous because any deregulation or misexpression of FGFs would become oncogenic. If, however, in addition to regulating genes mediating cellular proliferation, they (FGFs) could also activate genes mediating apoptosis, then all the cells expressing or receiving FGF signals, would be primed for cell death. Successful proliferation would presumably only occur if apoptosis was actively inhibited by, perhaps, activation of complimentary signal transduction pathways or perception of other cooperative signals. This once again reiterates the concept of oncogene cooperation and multistep tumorigenesis. Apoptosis and multistep mammary tumorigenesis The specific aim of this thesis project was to analyze multistep mammary tumorigenesis and oncogenic cooperation. Our finding of FGF induced apoptosis could possibly explain one mechanism of oncogenic cooperation. In a tumor environment there are multiple events that have taken place all of them having an end result of uncontrolled cellular 129 Reproduced with permission o f^ c o p y r ig h t owner. Further reproduction prohibited without permission. proliferation. The study of MMTV-mycMMTV-ras double transgenic mice, clearly demonstrates this type of cooperation (Hundley et al., 1997). Analysis of the apoptosis occurring in tumors from MMTV-myc or MMTV- myc,/MMTV-ras double transgenic mice showed a remarkable decrease in the number of cells undergoing programmed cell death in tumors of the latte rmice. These results indicate that expression of the proto- oncogene ras in the mammary gland partially rescues the myc induced apoptosis. Therefore, in our system, FGFs can provide a mitogenic stimulus if its apoptotic effect is blocked or suppressed by other survival signals to the cells. These survival signals could be provided by other cooperating events, one of which could possibly be Wnt expression, or expression of other genes regulated by Wnts. This hypothesis seems plausible since FGFs and Wnts are very strong collaborators, both in mammary tumorigenesis and in several developmental systems (Christian eta!., 1991; Doniach, 1995; MacArthur eta!., 1995). Hence, it seems reasonable that one mechanism of cooperation could possible involve regulation of programmed cell death. Thus our results suggest a potentially significant role for FGFs in regulating cell growth and death in normal and pathological processes. Further analysis of the apoptotic property of FGFs, in particular the signal transduction pathway, and the identification of modulators of FGF induced apoptosis, will provide opportunities for the design of rational therapies for cancer. 130 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter 5 Long distance activation o f int2/Fgf3 by M M TV proviral insertions Introduction Infection of Wnt1 transgenic mice with the MMTV, results in accelerated mammary tumorigenesis and increased numbers of tumors per animal (Shackleford at al., 1993). Approximately 45% of the resulting tumors with clonal, tumor-specific MMTV integrations contain insertionally activated Fgf3 or Fgf4, suggesting that the Wnt-1 gene product can cooperate with at least two members of the FGF family (Shackleford eta!., 1993). Despite the identification of clonal, tumor-specific MMTV proviruses in the remaining 55% of the tumors, we did not detect insertional activation of any previously identified targets for MMTV insertion mutations (Fgf3, Fgf4, int3, or Wnt3) in this group (Shackleford at a/., 1993). To determine the identity of proto-oncogenes that may be activated in these tumors, we have analyzed the DNAs surrounding proviral integration sites in several tumors. The results form one such analysis identified FgfS as a new common insertion locus, and 8 of 80 tumors analyzed had MMTV insertions within this locus (10%) (MacArthur at al., 1995). Therefore, the remaining 45% of tumors with clonal tumor specific proviruses probably had activations of other novel or unexpected genes. In this chapter, I will describe the cloning and identification of the gene(s) activated by MMTV in some of these tumors. 131 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Materials and Methods Tumor Samples. All mammary tumors were derived from a previous study (Shackleford et al., 1993) in which female Wnt-1 transgenic mice (Tsukamoto at al., 1988) were infected at 3-4 weeks of age with MMTV produced from EH-2 cells (Shackleford and Varmus, 1988). Preparation of nucleic acids. Tumor DNAs were isolated as described (Shackleford and Varmus, 1988), except that serum separation tubes were used in the extractions (Thomas et al., 1989). Southern blot analyses. Preparation and screening of genomic libraries. DMA extraction. Southern blotting, and construction of genomic libraries have been described in chapter 2. Mapping to the mouse chromosome. The locus was mapped to the chromosome using the Jackson Laboratory Backcross DNA Panel Map service (Rowe eta/., 1994). Briefly, a restriction length polymorphism (RFLP) was identified between the two mouse strains M. spretus and C57BL/6. This RFLP was traced through the entire BSS reciprocal cross panel by Southern analysis. The results from this analysis were sent to Jackson Laboratory and, through analysis of the database the locus was mapped to the mouse chromosome. Probes and sequencing. DNA probes used for hybridization were isolated from plasmids by digestion with restriction endonucleases, agarose gel electrophoresis, and glass bead binding. Potential probes were tested for repetitive elements by blotting to nylon and hybridizing with 32p.|abeled mouse genomic DNA (Steinmetz et al., 1980). Unique probes were used on Southern blots. In addition to the probes 132 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. generated from the genomic clones the following probes were used on blots: a 1.9 kb Pst\!Xho\ fragment (MMTV gag probe); a 1.2 kb Bam HI fragment (MMTV en y probe) (Shackleford and Varmus, 1988); a 0.95 kb PstUSst I LTR fragment (MMTV-LTR probe); a1.3 kb murine Cyclin D1 cDNA; 3.0 kbH19cDNA. Results Detection and cloning of a new tum or-specific common insertion site for MMTV. In previous work, we showed that MMTV infection of Wnt1 transgenic mice accelerates mammary tumorigenesis by insertional activation of Fgf3 and Fgf4 (Shackleford et al., 1993), and Fgfd. (MacArthur at al., 1995). However, approximately 45% of the tumors examined did not show proviral insertions near any of the proto oncogenes known to be affected by MMTV in tumors of normal mice (including int3 and Wnt3), nor the newly identified target FgfS, despite the presence of clonal proviral insertions in these tumors (Shackleford at al., 1993). We reasoned that these tumors may harbor unpredicted or novel proto-oncogenes activated by proviral insertion. To isolate such genes, we first sought to clone the proviral-cellular junction fragments from tumors that contained only one or two clonally integrated proviruses. Southern blot analysis of tumor DNAs using an MMTV gag probe identified several candidate junction fragments (Figure 2.2 and data not shown). The Xhol-cleaved junction fragment from tumor 42, and a Sst I fragment from tumor 76 (Figure 2.2) were cloned into lambda vectors and 133 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. isolated by screening with an MMTV gag, or LTR probe. Unique cellular DNAs flanking the MMTV proviruses in these clones were isolated and used as probes on Southern blots of DNAs from 80 tumors with clonal insertions to ask if other tumors had rearrangements due to proviral insertions in these loci. Data from these experiments suggested that the insertions in tumors 42 and another tumor #14 (cloned by 0. Tuason) were actually in the same locus and, furthermore, that eight other tumors also contained proviruses in this region. Since this locus was being characterized by another student, further analysis of tumor 42 insertion was not pursued. Screening of the tumor DNA panel with a tumor 76 specific cellular probe showed MMTV proviral insertions within this locus in 12 additional tumors. Hybridization of these blots with MMTV probes confirmed that the rearrangements of this locus in affected tumors were due to proviral integrations (Figure 5.1, data not shown). Detection of expressed genes in the common insertion region. In a search for active genes in this locus, we made probes from the cloned cellular DNAs described above, as well as from adjacent regions cloned subsequently, for use in northern blot analyses of tumor RNAs. No specific transcript was seen hybridizing to any of the repeat-free cellular probes. Additionally, exon trapping procedure was performed in order to look for transcribed regions within the cloned cellular fragment. Analysis of the PCR products after the exon trap procedure detected only the 100 bp insulin fragment indicating that the splicing occurred between the second and third insulin exons (control) but no exon was trapped from the cloned region. 134 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chromosomal mapping of tumor 76 locus. In the process of identifying the gene activated by proviral insertions in this locus, I decided to map this locus to the mouse chromosome, hoping to provide reasonable a determination whether the cloned fragment was located near a previously identified MMTV insertion site, and/or whether any known candidate genes were present in the region. The locus was mapped to the distal region of mouse chromosome? using the Jackson Laboratory Backcross DNA Panel Map service (Rowe et al., 1994). Unfortunately, this locus was linked to the int2/Fgf3 locus suggesting that this was a likely candidate to be activated by MMTV proviral insertions. However, previous analysis of these tumors did not detect MMTV insertions within the int2/Fgf3 locus (Shackleford et a!., 1993). This suggests that either these insertions are activating other genes that are linked to this locus, or that these insertions are long range activations of int2/Fgf3. Several genes (cyclinDI, Igf2, H-19, hst/Fgf4) are located near int2/Fgf3 on mouse chromosome-7 and seem to be good candidates for MMTV activations (Rowe et a!., 1994). The corresponding human region (llq IS ) has been implicated in breast tumors (Kallioniemi, 1994). Cyclin D1 has been implicated in tumorigenesis, and overexpression of CyclinDI in the mammary gland using a MMTV promoter results in hyperplasia and tumor formation (Wang et ai, 1994). H-19 has been shown to function as a tumor suppressor (Mao et al., 1993). Since H-19 is a maternally imprinted gene (Hao et al., 1993), MMTV insertions into this locus could possibly knock out the functional copy of the gene and 135 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. thereby contribute to mammary tumorigenesis. Hst/Fgf4 is a previously described target for MMTV activations (Peters etal., 1989; Shackleford et al., 1993). To test if there were activations in these other genes, Southern analysis was performed using DNAs from tumors with MMTV insertions. Southern blots were hybridized to probes containing coding sequences of the following genes: Cyclin D1, H-19, IGF-2, and hst. Results from these experiments detected no rearrangements in any of the tumors analyzed. Given the propensity of MMTV to activate FGF genes in this system, together with the inability to identify any other candidate genes, Fgf3 appears to be the most likely candidate to be activated by these MMTV insertions. Long range activations of genes by MMTV and MLV have been described (Nusse and Varmus, 1982). Long range activations of int2/Fgf3 by MMTV insertions -20kb upstream of the gene has been previously reported (Peters et ai, 1989). Comparison of the restriction maps between my lambda clones and the lambda clones from their study (Peters et al.) indicate that the 5' cluster of insertions in the tumor 76 locus and the cloned region, overlaps to the insertions and lambda clones previously published (Peters at ai, 1989). This further supports the idea of long range activations. Northern analysis of some of the tumor RNAs probed with an int2/Fgf3 probe showed activation of the gene in the tumors and no expression in mammary gland controls, proving that int2/Fgf3 is the gene activated by MMTV in these tumors (data not shown). The insertion in the tumor 77 locus seems to be the most distant activation of a gene by MMTV insertion to be described. 136 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. % s c e a > * g G I ■ o SI) G C ▼ ▼ c « > X ■ :▼ T ! V ;V:V 3 i " 3 t •il s a 7 3 .8 E i S > .X * N I n I « 1 137 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Discussion An apparent common insertion locus was identified for MMTV in infected Wnt1 transgenic mice. Chromosome mapping procedures mapped this locus to the distal arm of the mouse chromosome?, a region linked to the Fgf3 locus. Further characterization of this locus revealed that the proviral insertions in this locus were activating the Fgf3 proto oncogene over a distance of 15-21 kb. The insertion in tumor 77 is at least 21 kb S' to the Fgf3 gene, the longest range described to date. The other five prime cluster of proviral insertions (Figure 5.1) overlap with the insertions in the previously described tumors MT40, and MT42 (Peters et al., 1989). These activations were probably missed in the previous study (Shackleford at a!.. 1993) because the preliminary screen for proviral insertions in int2/Fgf3 was based on Southern analysis using cDNA probes. These analyses are limited by the number of restriction sites present for any given enzyme and the number of DNA probes available to cover large distances. Hence, it is very possible to miss long range insertions with these types of analyses. The identification of FgfS activations in twelve additional tumors, increases the percentage of tumors with Fgf3 activations in the Wnti transgenic system. The fact that every activation identified in this model system seems to be a Fgf gene or Fgf receptor, reiterates the strong cooperation between Fgfs and Wnts. Interestingly, in the system where Fgf3 transgenic mice were infected with MMTV, all the identified activations were in IVnf genes (Lee etal., 1995), thus corroborating our findings. 138 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter 6 Summary and Conclusions Breast cancer, like all cancers, Is a multistep process Involving the sequential acquisition of genetic alterations over a period of time. Studying this process In humans Is a prolonged and arduous task; therefore, animal models are a desirable alternative. We have used a W nti transgenic mouse model to study the multiple genetic events In mammary cancer development. These mice express Wnt1, a secreted growth factor like protein. In the mammary gland and develop mammary tumors with a prolonged latency. These transgenic mice were Infected with the mouse mammary tumor virus (MMTV) to accelerate tumorigenesis and to molecularly tag proto-oncogenes that are activated In the resulting tumors and that cooperate with W nti In mammary tumorigenesis. Using this approach. Transcriptional activation of two fibroblast growth factor (FGF) genes Fgf3 and Fgf4 were found In these tumors Indicating their strong collaboration with Wnti. By examination of the tumors that lack activation of these genes (and other genes that are usual targets of MMTV Insertions), we Identified a common Insertion locus for MMTV and determined the activated gene In this locus to be another member of the FGF family, FgfS. FgfS Is transcrlptlonaly activated In 50% of the tumors from Infected Wnti transgenic mice, In comparison to the lack of Fgf8 RNA In other tumors and mammary tissues, suggesting a strong oncogenic cooperation between FgfS and Wnti In mammary tumorigenesis. Further 139 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. characterization of the Fgf8 gene showed that normal expression was detectable only in testis and ovary (by poly(A)+ RNA northern analysis of 13 tissues) of adult mice, and in midgestational embryos. Fgf8 consists of at least six exons and codes for at least seven protein isoforms which differ at their mature amino termini, due to alternative splicing. We isolated cDNA clones for three of these isoforms (FgfSa, b and c) using exon trapping and reverse transcription polymerase chain reaction. In order to test for differences in the biological/oncogenic potencies of the three isoforms, NIH3T3 cells were stably transfected with expression vectors containing the three different cDNAs, and cell lines expressing the three different protein isoforms were generated. These stably transfected cells were compared in several biological assays. Our results showed that NIH3T3 cells expressing FgfSb were strongly transformed; they become morphologically altered, clonagenic in soft agar and tumorigenic in nude mice. In contrast, cells expressing FgfSa and FgfSc showed subtle morphological changes with FgfSc showing stronger changes than FgfSa. These results indicate that there are differences in the oncogenic potencies of the FgfS isoforms and suggest that these isoforms may have different functions in vivo. Similarly NIH3T3 cells trasnsfected with the human FGFS isoforms FGFSa, FGFSb, and FGFSe showed differences in their biological activities, with FGFSb being the most potent of all the isoforms in NIH3T3 transformation assays. Targetted expression of the Fgf8 gene in the mammary gland using a MMTV-LTR promoter/enhancer resulted in transformation of the mammary gland: the animals developed generalized hyperplasia, 140 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. followed by development of tumors that ranged from benign adenomas to invasive adenocarcinomas. Interestingly, the female mice also developed stromal hyperplasia of the ovaries. These results provide further evidence for the oncogenic capability of Fgf8. During the analysis of the oncogenic potential of Fgf8 in mammary epithelial cells, we serendipitously discovered that stimulation of a normal mammary epithelial cell line C57MG with the conditioned medium containing the secreted FGFSb protein resulted in transformation of the cells followed by complete cell death due to apoptosis, as evidenced by nuclear condensation and fragmentation and oiigonucleosomal laddering. Ablation of the protein from the conditioned medium by binding it to heparin-sepharose abolished this apoptotic activity. Treatment of the cells with the partially purified FgfSb protein restored this effect. Since this was a previously undescribed property of FGFs, we tested to see if other members of the FGF family of growth factors could induce programmed cell death of mammary epithelial cells. Cells treated with purified human FGF1, 2, 4, 6, and 9 (100 ng/ml) went through morphological transformation followed by apoptotic cell death similar to FgfSb, indicating that this property was not unique to FgfSb alone but was common to other FGF family members. Several other mammary epithelial cell lines show the same response to FGF stimulation. To further prove the apoptosis inducing property of FGFs, we performed neutralization experiments with an anti-FGF2 antibody. Cells treated with FGF2 that was preincubated with the antibody did not undergo programmed cell death, in contrast to the cells treated with FGF2 protein alone, confirming that the apoptosis of mammary epithelial 141 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. cells is a result of the FGF signal. Overexpression of the anti-apoptotic gene BCL2 in these ceils delayed the onset of apoptosis induced by FGFs. These results describe a new property for FGFs and suggest that these growth factors play very important roles in regulating cell growth and death both in normal development as well as in pathological conditions like cancer. Our results also reiterate the concept of oncogenic cooperation in cancer, where we can speculate that the death signal from the overexpression of FGFs could be blocked or overcome by survival signals from other gene products (oncogenic cooperation) in the tumor environment thus allowing cellular proliferation and tumorigenesis. In conclusion, we have identified Fgf8 as the third FGF gene to be activated by MMTV insertions in this Wnt1 transgenic system, indicating the strong cooperation between the FGF and Wnt families in oncogenesis. Interestingly, this cooperation between Wnts and FGFs is also seen during embryonic development: mesoderm induction (Christian et a/., 1992), formation of the mesencephelon and metencephalon, and antero-posterior patterning (Doniach, 1995). This study, identified Fgfd as an oncogenic collaborator of Wnti, and also showed that Fgf8 had at least seven different protein isoforms, which differed in their bilogical properties. The differences in the oncogenic capabilities of Fgf8 isoforms was demonstrated both in vitro (cell culture assays) and in vivo (transgenic mice). Finally, I also described the identification of a new property for FGFs: induction of programmed cell death. Thus, the Wnt1 mammary tumorigenesis model system has been successful in realizing my experimental goals 142 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. described in chapter 1. However, despite the success with Fgf8, we were unable to identify any novel or other genes (other than Fgfs and FGFR2) activated in these tumors. Therefore, this system has to be taken one step further to identify other cooperating events In the multistep process of mouse mammary tumorigenesis. This can possibly be accomplished by infecting transgenic mice (with MMTV) that overexpress both Fgf and Wnt in their mammary gland, and then identifying the proviral insertion sites and activated proto-oncogenes. This process will probably circumvent the activations of either Wnts or Fgfs and therefore new activations can be identified. Some of the caveats of this system have been mainly technical. The identification of long range activations of Fgf3 or Fgf8 had been missed because of the use of Southern blotting as the main method for screening the tumors. Northern analyses combined with chromosomal mapping could have avoided this problem. However, since the RNA from these tumors were limited, this could not be done. In future, approaches like reverse northerns for screening the tumors and chromosomal mapping could possibly be more helpful in picking up long distance activations. 143 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Epilogue Breast cancer is a debilitating and often fatal disease associated with severe psycho-social trauma. Despite the apparent strides and the technological advances made in understanding cancer, the incidence of breast cancer in the world is still on a rise. Hence, it is imperative that the exact genetic and epigenetic factors involved in breast cancer be well understood so that better therapeutic strategies can be devised. Studies such as ours, identifying the oncogenes involved in breast cancer and analyzing the multiple genetic alterations involved, will be significant in determining the prognosis of the disease, as well as in designing rational therapies that specifically alter or inhibit the function of oncogenes. However, to achieve these goals, it will take the concerted efforts of the scientific and medical communities, along with the cooperation of the general public, to work towards curing if not eradicating this disease that afflicts our mothers and daughters. 144 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. References Aaronson, S., Abraham, J., Baird, A., Basillco, C., Birnbaum, D., Bohlen, P., Burgess, W., Dickson, C., Fiddes, J., Goldfarb, M., Klagsbrun, M., Maciag, T., Martin, G., Peters, G., Rubin, J., Thomas, K., Terada, M., and Yoshida, T. (1991). The fibroblast growth factor family; nomenclature meeting report and recommendations. Ann. N.Y. Acad. Sci. 638, xiii-xvi. Abraham, J. A., Whang, J. L , Tumolo, A., Mergia, A., and Friedman, J. (1986). Human basic fibroblast growth factor: nucleotide sequence and genomic organization. EMBO J. 5, 2523-2528. Acland, P., Dixon, M., Peters, G., and Dickson, C. (1990). Sub cellular fate of the int-2 oncoprotein is determined by choice of initiation codon. Nature 343, 662-665. Albini, A., Benelli, R., Presta, M., Rusnati, M., Ziche, M., Rubartelli, A., Paglialunga, G., Bussolino, F., and Noonan, D. (1996). HIV-tat protein is a heparin-binding angiogenic growth factor. Oncogene 12, 289-297. Alexander, W. S., Bernard, O., Gory, S., and Adams, J. M. (1989). Lymphomagenesis in Em-myc transgenic mice can involve ras mutations. Oncogene 4, 575-581. Amundadottir, L. T., Merlino, G., and Dickson, R. B. (1996). Transgenic mouse models of breast cancer. Breast Cancer Research and Treatment 39, 119-135. Armstrong, D. K., Kaufmann, S. H., Ottaviano, Y. L., Furuya, Y., Buckley, J. A., Isaacs, J. T., and Davidson, N. E. (1994). Epidermal growth factor- mediated apoptosis of MDA-MB-468 human breast cancer cells. Cancer Res. 54, 5280-5283. Auch, D., and Reth, M. (1990). Exon trap cloning: using PCR to rapidly detect and clone exons from genomic DNA fragments. Nucleic Acids Res. 18, 6743-6744. 145 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Balmain, A., and Brown, K. (1988). Oncogene activation in chemical carcinogenesis. Adv. Cancer Res. 51, 147-182. Basilico, 0., and Moscatelli, D. (1992). The FGF family of growth factors and oncogenes. Adv. Cancer Res. 59, 115-165. Basnakian, A. G., and James, S. J. (1994). A rapid and sensitive assay for the detection of DNA fragmentation during early phases of apoptosis. Nucleic Acids Res. 22, 2714-2715. Berry, D. E., Lu, Y., Schmidt, B., Fallon, P. G., O’Connell, C., Hu, S. X., Xu, H. J., and Blanck, G. (1996). Retinoblastoma protein inhibits IFN-gamma induced apoptosis. Oncogene 12, 1809-1819. Bishop, J. M. (1991). Molecular themes in oncogenesis. Cell 64, 235- 248. Bissonnette, R. P., Echeverri, F., Mahboubi, A., and Green, D. R. (1992). Apoptotic cell death induced by c-myc is inhibited by bcl-2. Nature 359, 552-554. Blandino, G., and Strano, S. (1997). BCL-2: the pendulum of the cell fate. J Exp. Clin. Cancer Res. 16, 3-10. Borer, R. A., Lehner, C. F., Eppenberger, H. M., and Nigg, E. A. (1989). Major nucleolar proteins shuttle between nucleus and cytoplasm. Cell 56, 379-390. Bouchard, L., Lamarre, L., Tremblay, P. J., and Jolicoeur, P. (1989). Stochastic appearance of mammary tumors on transgenic mice carrying the MMTV/cneu oncogene. Cell 57, 931-936. Brabyn, C. J., Franks, D. J., and Kleine, L. P. (1994). Long-term exposure to epidermal growth factor results in apoptosis in T51B cells. Biochem. Cell Biol. 72, 429-438. 146 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Bradley, R. S., and Brown, A. M. C. (1990). The proto-oncogene int-l encodes a secreted protein associated with the extracellular matrix. EMBO J. 9, 1569-1575. Brookes, S., Placzek, M., Moore, R., Dixon, M., Dickson, C., and Peters, G. (1986). Insertion elements and transitions in cloned mouse mammary tumor virus DNA: further delineation of the poison sequences. Nucleic Acids Res. 14, 8231-8245. Brown, A. M. 0., Wildin, R. S., Prendergast, T. J., and Varmus, H. E. (1986). A retrovirus vector expressing the putative mammary oncogene int-1 causes partial transformation of a mammary epithelial cell line. Cell 46, 1001-1009. Bullrich, P., Morris, S. W., Hummel, M., Pileri, S., Stein, H., and Croce, C. M. (1994). Nucleophosmin {NPM) gene rearrangements in Ki-1-positive lymphomas. Cancer Res. 54, 2873-2877. Callahan, R. (1996). MMTV-induced mutations in mouse mammary tumors: Their potential relevance to human breast cancer. Breast Cancer Research and Treatment 39, 33-44. Chiarugi, V., and Ruggiero, M. (1996). Role of three cancer "master genes" p53, bcl2 and c-myc on the apoptotic process. Tumori. 82, 205- 209. Chinnaiyan, A. M., and Dixit, V. M. (1996). The cell-death machine. Curr. Biol. 6, 555-562. Chomczynski, P., and Sacchi, N. (1987). Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal. Biochem. 162, 156-159. Christian, J. L, McMahon, J. A., McMahon, A. P., and Moon, R. T. (1991). Xwnt-8, a Xenopus IVnf-f/Znf-f-related gene responsive to mesoderm- inducing growth factors, may play a role in ventral mesodermal patteming during embryogenesis. Development 111, 1045-1055. 147 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Christian, J. L. Oisen, D. J., and Moon, R. T. (1992). XM/ht-6 modifies the character of mesoderm induced by bFGF in isolated Xenopus ectoderm. EMBO J. 11, 33-41. Church, G. M., and Gilbert, W. (1984). Genomic Sequencing. Proc. Natl. Acad. Sci. USA 81, 1991-1995. Clarke, R. (1996). Animal models of breast cancer. In Diseases of the breast, J. R. Harris, Lippman, M.E., Morrow,M., Heilman,S., ed. (Philadelphia; Lippcott-Raven). Clausse, N., Baines, Moore, R., Brookes, S., Dickson, C., and Peters, G. (1993). Activation of both Wnt-1 and Fgf-3 by insertion of mouse mammary tumor virus downstream in the reverse orientation: a reappraisal of the enhancer insertion model. Virology 194, 157-165. Clurman, B. E., and Hayward, W. S. (1989). Multiple proto-oncogene activations in avian leukosis virus-induced lymphomas: evidence for stage specific events. Moi. Cell. Biol. 9, 2657-2664. Coleman-Krnacik, S., and Rosen, J. M. (1994). Differential temporal and spatial gene expression of fibroblast growth factor family members during mouse mammary gland development. Mol. Endocrinol 8, 218-29. Compere, S. J., Baldacci, P., and Jaenisch, R. (1988). Oncogenes in transgenic mice. Biochim. Biophys. Acta 948, 129-149. Cory, S. J., and Adams, J. M. (1988). Transgenic mice and oncogenesis. Annu. Rev. Immunol. 6, 25-48. Crossley, P. H., and Martin, G. R. (1995). The mouse Fgf8 gene encodes a family of polypeptides and is expressed in regions that direct outgrowth and patterning in the developing embryo. Development 121, 439-451. Davidson, N. E. (1989). Biology of breast cancer and its clinical implications. Curr. Opin. Oncol 1 , 269-76. 148 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. de Lapeyriere, O., Rosnet, O., Benharroch, D., Raybaud, F., Marchetto, S., Planche, J., Galland, F., Mattel, M.-G., Copeland, N. G., Jenkins, N. A., Couller, F., and Birnbaum, D. (1990). Structure, chromosome mapping and expression of the murine Fgf-6 gene. Oncogene 5. 823-831. Deiss, L. P., Feinstein, E., Berissi, H., Cohen, O., and Kimchi, A. (1995). Identification of a novel serine/threonine kinase and a novel 15-kD protein as potential mediators of the gamma interferon-induced cell death. Genes Dev. 9, 15-30. Delli Bovi, P., Curatola, A. M., Kern, F. G., Greco, A., Ittman, A., and Basilico, C. (1987). An oncogene isolated by transfection of Kaposi's sarcoma DNA encodes a growth factor that is a member of the Fgf family. Cell 50, 729-37. Dickson, C. (1990). Role of the int-genes in murine mammary tumor development and implications for human breast cancer. Int J Cancer SuppI 5. 51-4. Dickson, C., Smith, R., Brookes, S., and Peters, G. (1984). Tumorigenesis by mouse mammary tumor virus; proviral activation of a cellular gene in the common integration region int-2. Cell 37, 529-536. Dickson. R., Lippman, M.E. (1996). Oncogenes and suppressor genes. In Diseases of the breast, J. R. Harris, Lippman, M.E., Morrow,M., Heilman,S., ed. (Philadelphia: Lippcott-Raven). Dietrich, J. B. (1997). [Apoptosis and anti-apoptosis genes in the Bcl-2 family]. Arch. Physiol. Biochem. 105, 125-35. Dingwall, C., and Laskey, R. A. (1990). Nucleoplasmin: the archetypal molecular chaperone. Semin. Cell Biol. 7 , 11-17. Donehower, L. A., Godley, L. A., Aldaz, C. M., Pyle, R., Shi, Y. P., Pinkel, D., Gray, J., Bradley, A., Medina, D., and Varmus, H. E. (1995). Deficiency of p53 accelerates mammary tumorigenesis in Wnt-1 transgenic mice and promotes chromosomal instability. Genes Dev.. 9, 882-895. 149 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Doniach, T. (1995). Basic FGF as an inducer of anteroposterior neural pattern. Cell 83, 1067-70. Duckett, 0. S., and Thompson, C. B. (1997). The control and execution of programmed cell death: an update. Biochim Biophys Acta 1332, R45-7. Durgam, V. R., and Tekmal, R. R. (1994). The nature and expression of int-5, a novel MMTV integration locus gene in carcinogen-induced mammary tumors. Cancer Lett. 87, 179-186. Escot, C., Theillet, C., Lidereau, R., Spy rates, F., Champeme, M., Gest, J., and Callahan, R. (1986). Genetic alteration of teh cmyc proto-oncogene in human primary breast carcinomas. Proc. Natl. Acad. Sci, USA 83, 4834-38. Evan, G. I., Wyllie, A. H., Gilbert, C. S., Littlewood, T. D., Land, H., Brooks, M., Waters, C. M.. Penn, L. Z., and Hancock, D. C. (1992). Induction of apoptosis in fibroblasts by c-myc protein. Cell 69, 119-28. Fearon, E. R., and Vogelstein, B. (1990). A Genetic Model for Colorectal Tumorigenesis. Cell 61, 759-767. Florkiewicz, R. Z., and Sommer, A. (1989). Human basic fibroblast growth factor gene encodes four polypeptides: three initiate translation from the non-AUG codons. Proc. Natl. Acad. Sci., USA, 86, 3978-81. Folkman, J., Watson, K., Ingber, D., and Hanahan, D. (1989). Induction of angiogenesis during the transition from hyperplasia to neoplasia. Nature 339, 58-61. Fort, P., Marty, L., Piechaczyk, M., El Sabrouty, S., Dani, C., Jeanteur, P., and Blanchard, J. M. (1985). Various rat adult tissues express only one major mRNA species from the glyceraldehyde-3-phosphate- dehydrogenase multigenic family. Nucl. Acids Res. 13, 1431-1442. Gallahan, D., and Callahan, R. (1987). Mammary tumorigenesis in feral mice: identification of a new int locus in mouse mammary tumor virus (Czech ll)-induced mammary tumors. J Virol 61, 66-74. 150 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Gallahan, D., Kozak, C., and Callahan, R. (1987). A new common Integration region {int-3) for mouse mammary tumor virus on mouse chromosome 17. J. Virol. 61, 218-220. Garcia-Martinez, V., Macias, D., Ganan, Y., Garcia-Lobo, J. M., Francia, M. V., Femandez-Teran, M. A., and Hurle, J. M. (1993). Intemucleosomal DNA fragmentation and programmed cell death (apoptosis) in the interdigital tissue of the embryonic chick leg bud. J Cell Sci 106, 201- 208. Gavin, B. J., McMahon, J. A., and McMahon, A. P. (1990). Expression of multiple novel tVnMZ/nM-related genes during fetal and adult mouse development. Genes Dev. 4, 2319-2332. Gavrieli, Y., Sherman, Y., and Ben-Sasson, S. A. (1992). Identification of programmed cell death in situ via specific labeling of nuclear DNA fragmentation. J Cell Biol 119, 493-501. Ghosh, A. K., Shankar, D. B., Shackleford, G. M., Wu, K., T’Ang, A., Miller, G. J., Zheng, J., and Roy-Burman, P. (1996). Molecular cloning and characterization of human FGF8 alternative messenger RNA forms. Cell Growth Differ 7, 1425-34. Giannios, J.. and loannidou-Mouzaka, L. (1997). Molecular aspects of breast and ovarian cancer Eur J Gynaecol Oncol 18, 387-93. Goldfarb, M., Deed, R., MacAllen, D., Walther, W., Dickson, C., and Peters, G. (1991). cell transformation by Int2-a member of the fibroblast growth factor family. Oncogene 6, 65-71. Gospodarowicz, D., Neufeld, G., and Schweigerer, L. (1987). Fibroblast growth factor: structural and biological properties. J. Cell. Physiol. SuppI. 5, 15-26. Gray, D. A., Jackson, D. P., Percy, D. H., and Morris, V. L. (1986). Activation of int-1 and int-2 loci in GR mammary tumors. Virol. 154, 271- 278. 151 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Guchelaar. H. J.. Vermes, A., Vermes, I., and Haanen, C. (1997). Apoptosis: molecular mechanisms and implications for cancer chemotherapy. Pharm World Sci 19, 119-25. Guenal, I., Sidoti-de Fraisse, C., Gaumer, S., and Mignotte, B. (1997). Bcl-2 and Hsp27 act at different levels to suppress programmed cell death. Oncogene 15, 347-60. Haas-Kogan, D. A., Kogan, S. C., Levi, D., Dazin, P., T'Ang, A., Fung, Y. K., and Israel, M. A. (1995). Inhibition of apoptosis by the retinoblastoma gene product. EMBO J 14, 461 -72. Hanahan, D. (1989). Transgenic mice as probes into complex systems. Science 246, 1265-1275. Hao, Y., Crenshaw, T., Moulton, T., Newcomb, E., and Tycko, B. (1993). Tumour-suppressor activity of H I9 RNA. Nature 365, 764-767. Harrington, E. A., Bennett, M. R., Fanidi, A., and Evan, G. I. (1994). c-Myc- induced apoptosis in fibroblasts is inhibited by specific cytokines. Embo J 13, 3286-95. Haupt, Y., Rowan. S., Shaulian, E., Kazaz, A., Vousden, K., and Oren, M. (1997). p53 mediated apoptosis in HeLa cells: transcription dependent and independent mechanisms. Leukemia 7 7 , 337-9. Hébert, J. M., Basilico, C., Goldfarb, M., Haub, O., and Martin, G. (1990). Isolation of cDNAs encoding four mouse FGF family members and characterization of their expression patterns during embryogenesis. Devel. Biol. 138, 454-463. Heermeier, K., Benedict, M., Li, M., Furth, P., Nunez, G., and Hennighausen, L. (1996). Bax and Bcl-xs are induced at the onset of apoptosis in involuting mammary epithelial cells. Mech Dev. 56, 197-207. 152 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Heikinheimo, M., Lawshé, A., Shackleford, G. M., Wilson, D. B., and MacArthur, C. A. (1994). Fgf-8 expression in the post-gastrulation mouse is localized to the developing face, limbs and central nervous system. Mech. Dev. 48, 129-138. Hockenbery, D. M., Zutter, M., Hickey, W., Nahm, M., and Korsmeyer, S. J. (1991). BCL2 protein is topographically restricted in tissues characterized by apoptotic cell death. Proc. Natl. Acad. Sci. U S A .88, 6961-5. Hoffman, B., and Liebermann, D. A. (1994). Molecular controls of apoptosis: differentiation/growth arrest primary response genes, proto oncogenes, and tumor suppressor genes as positive & negative modulators. Oncogene 9, 1807-12. Hollingsworth, R., E, and Lee, W., H. (1991). Tumor suppressor genes: new prospect for cancer research. J. Natl. Cancer Inst. 83, 91-6. Huguet, E. L., McMahon, J. A., McMahon, A. P., Bicknell, R., and Harris, A. L. (1994). Differential expression of human Wnt gene-2, gene-3, gene-4 and gene-7b in human breast cell lines and normal and disease states of human breast tissue. Cancer Res. 54, 2615-2621. Hundley, J. E., Koester, S. K., Troyer, D. A., Hilsenbeck, S. G., Barrington, R. E., and Windle, J. J. (1997). Differential regulation of cell cycle characteristics and apoptosis in MMTV-myc and MMTV-ras mouse mammary tumors. Cancer Res 57, 600-3. Hunter, T. (1991). Cooperation between oncogenes. Cell 64, 249-270. lozzo, R. V., Eichstetter, I., and Danielson, K. G. (1995). Aberrant expression of the growth factor WNT-Sa in human malignancy. Cancer Res. 55, 3495-3499. Jehn, B. M., and Osbome, B. A. (1997). Gene regulation associated with apoptosis. Crit Rev Eukaryot Gene Expr 7, 179-93. 153 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Jhappan, C., Stable, C., Harkins, R. N., Fausto, N., Smith, G., H., and Merlino, G. T. (1990). TGF alpha overexpression in transgenic mice induces liver neoplasia and abnormal development of the mammary gland and pancreas. Cell 61, 1137-46. Kallioniemi, A., Kallioniemi, 0-P., Piper, J., Tanner,M., Stokke, T., Chen, L., Smith, H. S., Pinkel, D., Gray,J.W., and Waldman, P.M. (1994). Detection and mapping of amplified DNA sequences in breast cancer by comparative genomic mapping. Proc. Natl. Acad. Sci, USA 91, 2156- 2160. Kandel, J., Bossy-Wetzel, E., Radvanyi, F., Klagsbrun, M., Folkman, J., and Hanahan, D. (1991). Neovascularization is associated with a switch to the export of bFGF in the multistep development of fibrosarcoma. Cell 66, 1095-1104. Katso, R. M., Manek, S., O'Byrne, K., Playford, M. P., Le Meuth, V., and Ganesan, T. S. (1997). Molecular approaches to diagnosis and management of ovarian cancer. Cancer Metastasis Rev. 16, 81-107. Kelley, M. J., Pech, M., Seuanez, H. N., Rubin, J. S., OBrien, S. J., and Aaronson, S. A. (1992). Emergence of the keratinocyte growth factor multigene family during the great ape radiation. Proc. Natl. Acad. Sci. USA 89, 9287-9291- Keyomarsi, K., O'Leary, N., Molnar, G., Lees, E., Fingert, H. J., and Pardee, A. B. (1994). Cyclin E, a potential prognostic marker for breast cancer. Cancer research 54, 380-385. Kim, H. E., Han, S. J., Kasza, T., Han, R., Choi, H. S., Palmer, K. C., and Kim, H. R. (1997). Platelet-derived growth factor (PDGF)-signaling mediates radiation- induced apoptosis in human prostate cancer cells with loss of p53 function. Int J Radiat Oncol Biol Phys 39, 731-6. Kim, H. R., Upadhyay, S., Li, G., Palmer, K. C., and Deuel, T. F. (1995). Platelet-derived growth factor induces apoptosis in growth-arrested murine fibroblasts. Proc Natl Acad Sci U S A 92, 9500-4. 154 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Kinzler, K. W., and Vogelstein, B. (1997). Cancer-susceptibillty genes. Gatekeepers and caretakers [news; comment]. Nature 386, 761, 763. Kinzler, K. W., and Vogelstein, B. (1996). Lessons from hereditary colorectal cancer. Cell 87, 159-170. Kirshenbaum, L. A., and de Moissac, D. (1997). The bcl-2 gene product prevents programmed cell death of ventricular myocytes. Circulation 96, 1580-5. Klagsbrun, M. (1990). The affinity of fibroblast growth factors (FGFs) for heparin: FGF-heparan sulfate interactions in cells and extracellular matrix. Curr. Opinion Cell Biol. 2, 857-863. Knudson, A., G (1989). Hereditary cancers: clues to mechanisms of carcinogenesis. Br. J. Cancer 59, 661-6. Kouhara, H., Koga, M., Kasayama, S., Tanaka, A., Kishimoto, T., and Sato, B. (1994). Transforming activity of a newly cloned androgen- induced growth factor. Oncogene 9, 455-462. Kumar, R., Sukumar, S., and Barbacid, M. (1990). Activation of ras oncogenes preceding the onset of neoplasia. Science 248, 1101-1104. Kwan, H. C., Pecenka, V., Tsukamoto, A. S., Parslow, T. G.. Guzman, R., Lin, T. P., Muller, W. J., Lee, F. S., Leder, P., and Varmus, H. E. (1992). Transgenes expressing the Wnt-1 and int-2 proto-oncogenes cooperate during mammary carcinogenesis in doubly transgenic mice. Mol. Cell. Biol. 12, 147-154. Kyprianou, N., and Isaacs. J. T. (1989). Expression of transforming growth factor-beta in the rat ventral prostate during castration-induced programmed cell death. Mol. Endocrinol. 3, 1515-22. Laammie, G. A., FantI, V., Smith, R., Schuuring, E., Brookes, S., Michalides, R., and Peters, G. (1991). D118287, a putative oncogene on chromosome 11q13, is amplified and expressed in squamous cell and mammary carcinomas and linked to BCL1. Oncogene 6, 439-443. 155 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Lammie, G. A., and Peters, G. (1991). Chromosome 11q13 abnormalities in human cancer. Cancer ceils 3, 413-416. Land, H., Parada, L. P., and Weinberg, R. A. (1983). Tumorigenic conversion of primary embryo fibroblasts requires at least two cooperating oncogenes. Nature 304, 596-602. Lane, D. P., and Benchimol, S. (1990). p53: oncogene or antioncogene? Genes Dev. 4, 1 -8. Lee, F. S., Lane, T. P., Kuo, A., Shackleford, G. M., and Leder, P. (1995). Insertional mutagenesis identifies a member of the Wnt gene family as a candidate oncogene in the mammary epithelium of int-2/Pgf-3 transgenic mice. Proc. Nat.l Acad. Sci. USA. 92, 2268-2272. LeJeune, S., Huguet, E. L., Hamby, A., Poulsom, P., and Harris, A. L. (1995). WNT5-alpha cloning, expression and up-regulation in human primary breast cancers. Clin. Cancer Res. 1 , 215-222. Leung, H. Y., Dickson, C., Robson, C. N., and Neal, D. E. (1996). Over expression of fibroblast growth factor-8 in human prostate cancer. Oncogene 12, 1833-1835. Li, M., Hu. J.. Heermeier, K., Hennighausen, L., and Purth, P. A. (1996). Apoptosis and remodeling of mammary gland tissue during involution proceeds through p53-independent pathways. Cell Growth Differ 7, 13- 20. Liscia, D. S., Merlo, G. R., Garrett, C., French, D., Mariani, C. R., and Callahan, R. (1989). Expression of int-2 mRNA in human tumors amplified at the int-2 locus. Oncogene 4, 1219-1224. Lu, P. J., Lu, Q. L., Rughetti, A., and Taylor-Papadimitriou, J. (1995). bcl-2 overexpression inhibits cell death and promotes the morphogenesis, but not tumorigenesis of human mammary epithelial cells. J Cell Biol 129, 1363-1378. 156 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Lund, L. R.. Romer, J., Thomasset, N., Solberg, H., Pyke, C., Bissell, M. J., Dano, K., and Werb, Z. (1996). Two distinct phases of apoptosis in mammary gland involution: proteinase-independent and -dependent pathways. Development 122, 181-193. MacArthur, C. A., Lawshe, A., Shankar, D. B., Heikinheimo, M., and Shackleford, G. M. (1995). Fgf8 isoforms differ in NIH3T3 cell transforming potential. Cell, Growth and Diff. 6, 817-825. MacArthur, C. A., Lawshe, A., Xu, J., Santos-Ocampo, S., Heikinheimo, M., Chellaiah, A. T., and Ornitz, D. M. (1995). FGF-8 isoforms activate receptor splice forms that are expressed in mesenchymal regions of mouse development. Development 121, 3603-3613. MacArthur. 0. A., and Shackleford, G. M. (1997). Npm3: A novel, widely expressed gene encoding a protein related to teh molecular chaperones nucleoplasmin and nucleophosmin. Genomics 42, 137-140. MacArthur, 0. A., Shankar, D. B., and Shackleford, G. M. (1995). Fgf-8, activated by proviral insertion, cooperates with the Wnt-1 transgene in murine mammary tumorigenesis. J. Virol. 69, 2501-2507. Macias, D., Ganan, Y., Ros, M. A., and Hurle, J. M. (1996). In vivo inhibition of programmed cell death by local administration of FGF-2 and FGF-4 in the interdigital areas of the embryonic chick leg bud. An at. Embryol. (Beri) 193, 533-541. Macias, D., Ganan, Y., Sampath, T. K., Piedra, M. E., Ros, M. A., and Hurle, J. M. (1997). Role of BMP-2 and OP-1 (BMP-7) in programmed cell death and skeletogenesis during chick limb development. Development 124, 1109-17. Marchetti, A., Buttitta, F., Miyazaki, S., Gallahan, D., Smith, G. H., and Callahan, R. (1995). lnt-6, a highly conserved, widely expressed gene, is mutated by mouse mammary tumor virus in mammary preneoplasia. J. Virol. 69, 1932-1938. 157 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Marchuk, D., Drumm, M., Saulino, A., and Collins, F. S. (1991). Construction of T-vectors, a rapid and general system for direct cloning of unmodified PCR products. Nucl. Acids Res. 19, 1154. Marshall, C. J. (1991). Tumor suppressor genes. Cell 64, 313-326. Mason, I. J. (1994). The ins and outs of fibroblast growth factors. Cell 78, 547-552. Massague, J. (1983). Epidermal growth factor like transforming factor. J. Biol. Chem. 258, 13614-13620. McLeskey, S. W., Zhang, L , Kharbanda, S., Kurebayashi, J., Lippman, M. E., Dickson, R. B., and Kern, F. G. (1996). Fibroblast growth factor overexpressing breast carcinoma cells as models of angiogenesis and metastasis. Breast Cancer Res Treat 39, 103-117. McMahon, A. P., Gavin, B. J., Parr, B., Bradley, A., and McMahon, J. A. (1992). The Wnt family of cell signalling molecules in postimplantation development of the mouse. Ciba Found. Symp. 165, 199-212. Merlo, G. R., Basolo, F., Fiore, L., Duboc, L., and Hynes, N. E. (1995). p53-dependent and p53-independent activation of apoptosis In mammary epithelial cells reveals a survival function of EOF and insulin. J Cell Biol 128, 1185-96. Michalides. R., Van Goyen, A., and Nusse, R. (1983). Mouse mammary tumor virus expression and mammary tumor development. Volume 106 (Heidelberg: Springer-Verlag Berlin). Mochotka, S. V., Garrett, C. V., Schwartz, A. M., and Callahan, R. (1989). Amplification of the Proto-oncogenes int2, cerbB2, and cmyc in human breast cancer. Clin. Chim. Acta 184, 207-211. Montesano, R., Schaller, G., and Orel, L. (1991). Induction of epithelial tubular morphogenesis In vitro by fibroblast- derived soluble factors. Cell 66, 697-711. 158 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Moore, R., Casey, G., Brookes, S., Dixon, M., Peters, G., and Dickson, C. (1986). Sequence, topography and protein coding potential of mouse int- 2: a putative oncogene activated by mouse mammary tumour virus. EMBO J. 5, 919-924. Morris, D. W., Barry, P. A., Bradshaw, H. D. J., and Cardiff, R. D. (1990). Insertion mutation of the int-1 and int-2 loci by mouse mammary tumor virus in premalignant and malignant neoplasms from the GR mouse strain. J. Virol. 64, 1794-1802. Morris, V. L., Rao, T. R., Kozak, C. A., Gray, D. A., Lee Chan, E. C. M., Cornell, T. J.. Taylor, 0. B., Jones, R. P., and McGrath, 0. M. (1991). Characterization of Int-5, a locus associated with early events in mammary carcinogenesis. Oncogene Res. 5, 53-63. Muller, W. J., Lee, F. S., Dickson, C., Peters, G., Pattengale, P., and Leder, P. (1990). The int-2 gene product acts as an epithelial growth factor in transgenic mice. EMBO J. 9, 907-913. Muller, W. J., Sinn, E., Pattengale, P., Wallace. R., and Leder, P. (1988). Single step induction of mammary adenocarcinomas in transgenic mice bearing the activated cneu oncogene. Cell 54, 105-115. Naik, P., Karrim, J., and Hanahan, D. (1996). The rise and fall of apoptosis during multistage tumorigenesis: down- modulation contributes to tumor progression from angiogenic progenitors. Genes Dev. 10, 2105-2116. Nigro, J., M, Baker, S., J, Harris, C., c, and Vogelstein, B. (1989). Mutations in p53 occur in diverse tumor types. Nature 342, 705-708. Novak, J. (1996). Programmed cell death [letter; comment]. Science 274, 20 - 21 . Nusse, R. (1988). The activation of cellular oncogenes by proviral insertion in murine mammary tumors. In Breast Cancer: Cellular and Molecular Biology, M. E. Lippman and R. Dickson, eds. (Boston: Kluwer Academic Publishers), pp. 283-306. 159 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Nusse, R. (1986). The activation of cellular oncogenes by retroviral insertion. Trends Genet. 2, 244-247. Nusse, R. (1991). Insertional mutagenesis in mouse mammary tumorigenesis. In Retroviral Insertion and Oncogene Activation, H. J. Kung and P. K. Vogt, eds. (New York: Springer-Verlag), pp. 43-65. Nusse, R., van Goyen, A., Cox, D., Fung, Y. K. T., and Varmus, H. E. (1984). Mode of proviral activation of a putative mammary oncogene (/nf- 1 ) on mouse chromosome 15. Nature (London) 307, 131-136. Nusse, R., and Varmus, H. E. (1982). Many tumors induced by the mouse mammary tumor virus contain a provirus integrated in the same region of the host genome. Cell 31, 99-109. Nusse, R., and Varmus, H. E. (1992). M/nf genes. Cell 69, 1073-1087. Oberhammer, P., Wilson, J. W., Dive, C., Morris, I. □., Hickman, J. A., Wakeling, A. E., Walker, P. R., and Sikorska, M. (1993). Apoptotic death in epithelial cells: cleavage of DNA to 300 and/or 50 kb fragments prior to or in the absence of internucleosomal fragmentation. EMBO J. 12, 3679- 3684. Ohuchi, H., Yoshioka, H., tanaka. A., Kawakami, Y.. Nohno, T., and Noji, S. (1994). Involvement of Androgen-induced growth factor gene in mouse embryogenesis and morphogenesis. Biochem. Biophys. Res. Commun. 204, 882-888. Okumoto, M., Nishikawa, R., Iwai, M., Iwai, Y., Takamori, Y., Niwa, 0., and Yokoro, K. (1990). Lack of evidence for the involvement of type-C and type-B retroviruses in radiation leukemogenesis of NFS mice. Radiat. Res. 121, 267-73. Ornitz, D. M., Xu, J., Colvin, J. S., McEwen, D. G., MacArthur, C. A., Coulier, P., Gao, G., and Goldfarb, M. (1996). Receptor specificity of the fibroblast growth factor family. J .Biol. Chem. 271, 15292-7. 160 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Palmieri, S. (1989). Oncogene requirements for tumorlgenlcity: cooperative effects between retroviral oncogenes. Curr. Topics Microbiol. Immunol. 748,43-91. Papkoff, J., and Schryver, B. (1990). Secreted int-1 protein is associated with the cell surface. Mol. Cell. Biol. 10, 2723-2730. Perroteau, I., Salomen, D., Kidwell, W., Hazarika, P., Pardue, R., and Tam, J. (1986). Immunological detection and quantitation of alpha transforming growth factors in human breast carcinoma cells. Breast Cancer Res. Treat. 7, 201-210. Peters, G. (1990). Oncogenes at viral integration sites. Cell Growth & Development 7 , 503-510. Peters, G., Brookes, S., Placzek, M., Schuermann, M., Michalides, R., and Dickson, C. (1989). A putative int domain for mouse mammary tumor virus on mouse chromosome 7 is a S' extension of int-2. J. of Virol.83, 1448-1450. Peters, G., Brookes, S., Smith, R., and Dickson, C. (1983). Tumorigenesis by mouse mammaiy tumor virus: evidence for a common region for provirus integration in mammary tumors. Cell 33, 367-377. Peters, G., Brookes, S., Smith, R., Placzek, M., and Dickson, C. (1989). The mouse homolog of the hst/k-FGF gene is adjacent to int-2 and is activated by proviral insertion in some virally induced mammary tumors. Proc. Natl. Acad. Sci. U S A 88, 5678-82. Peters, G., Lee, A. E., and Dickson, C. (1986). Concerted activation of two potential proto-oncogenes in carcinomas induced by mouse mammary tumour virus. Nature 320, 628-631. Peto, R., Roe, F. J. C., Lee, P. N., Levy, L., and Clack, J. (1975). Cancer and aging in mice and men. Br. J. Cancer 32, 411-426. 161 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Raff, M. C. (1992). Social controls on cell survival and cell death. Nature 356, 397-400. Raff, M. C., Barres, B. A., Burne, J. F., Coles, H. S., Ishizaki, Y., and Jacobson, M. D. (1993). Programmed cell death and the control of cell survival: lessons from the nervous system. Science 262, 695-700. Rassoulzadegan, M., Cowle, A., Carr, A., Glaichenhaus, N., Kamen, R., and Cuzin, F. (1982). The roles of individual polyomavirus early proteins in oncogenic transformation. Nature 300, 713-718. Reed, J., C., Haider, S., Croce, C., M., and Cuddey, M., P. (1990). Complementation by BCL2 and c-HA-RAS oncogenes in malignant transformation of rat embryo fibroblasts. Mol. Cell. Biol. 10, 4370-4374. Rich, I. M. (1997). Status of breast cancer research today. In "Era of Hope" The Department of Defense Breast Cancer Research Program (Washington D.C.) Rijsewijk, F., Schuermann, M., Wagenaar, E., Parren, P., Weigel, D., and Nusse, R. (1987). The Drosophila homologue of the mouse mammary oncogene /nf-1 is identical to the segment polarity gene wingless. Cell 50, 649-657. Rijsewijk, F., van Deemter, L., Wagenaar, E., Sonnenberg, A., and Nusse, R. (1987). Transfection of the int-1 mammary oncogene in cubiodal RAC mammary cell line results in morphological transformation and tumorigenicity. EMBO J. 6, 127-131. Robbins, J., Blondel, B. J., Gallahan, D„ and Callahan, R. (1992). Mouse mammary tumor gene int-3 : a member of the notch gene family transforms mammary epithelial cells. J. Virol. 66, 2594-2599. Roelink, H., Wagenaar, E., Lopes da Silva, S., and Nusse, R. (1990). Wnt-3, a gene activated by proviral insertion in mouse mammary tumors, is homologous to int-1/Wnt-1 and is normally expressed in mouse embryos and adult brain. Proc. Natl. Acad. Sci. USA 87, 4519-4523. 162 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Rowan, S., and Fisher, D. E. (1997). Mechanisms of apoptotic cell death. Leukemia 11. 457-465. Rowe, L. B.. Nadeau, J. H., Turner, R., Frankel, W. N., Letts, V. A.. Eppig, J. T., Ko, M. S. H., Thurston, S. J., and Birkenmeir, E. H. (1994). Maps from two interspecific backcross DNA panels available as a community genetic mapping resource. Mammalian Genome 5, 1-11. Ruley, H. E. (1983). Adenovirus early region 1A enables viral and cellular transforming genes to transform primary cells in culture. Nature 304, 602- 606. Rydel, R. E., and Greene, L. A. (1987). Acidic and basic fibroblast growth factors promote stable neurite outgrowth and neuronal differentiation in cultures of PCI 2 cells. Neuron7, 3639-3653. Sambrook, J.. Fritsch, E. F., and Maniatis, T. (1989). Molecular cloning: a laboratory manual. Second Edition (Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press). Sandgren, E. P., Luetteke, N. 0., Palmiter, R. D., Brinster, R. L., and Lee, D. C. (1991). overexpression of TGF alpha in transgenic mice: induction of epithelial hyperplasia, pancreatic metaplasia, adn carcinoma of the breast. Cell 61, 1121-1135. Scheffner, M.. Werness, B. A., Huibregtse, J. M., Levine, A. J., and Howley, P. M. (1990). The E6 oncoprotein encoded by human papillomavirus types 16 and 18 promotes the degradation of p53. Cell 63, 1129-1136. Scully, R., Chen, J., Ochs, R. L., Keegan, K., Hoekstra, M., Feunteun, J., and Livingston, D. M. (1997). Dynamic changes of BRCA1 subnuclear location and phosphorylation state are initiated by DNA damage. Cell 90, 425-435. Scully, R., Chen, J., Plug, A., Xiao, Y., Weaver, D., Feunteun, J., Ashley, T., and Livingston, D. M. (1997). Association of BRCA1 with Rad51 in mitotic and meiotic cells. Cell 88, 265-75. 163 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Shackleford, G. M., MacArthur, C. A., Kwan, H. C., and Varmus, H. E. (1993). Meuse mammary tumor virus infection accelerates mammary carcinogenesis in Wnt-^ transgenic mice by insertional activation of int- 2/Fgf-3 and hst/Fgf-A. Proc. Natl. Acad. Sci., USA 90, 740-744. Shackleford, G. M., Shivakumar, S., Shiue, L, Mason, J., Kenyon, C., and Varmus, H. E. (1993). Two ivnf genes in Caenorhabditis elegans. Oncogene 8, 1857-1864. Shackleford, G. M., and Varmus, H. E. (1988). Construction of a clonable, infectious, and tumorigenic mouse mammary tumor virus provirus and a derivative genetic vector. Proc. Natl. Acad. Sci. USA 85, 9655-9659. Shackleford, G. M., and Varmus, H. E. (1987). Expression of the proto oncogene int-^ is restricted to postmeiotic male germ cells and the neural tube of mid-gestational embryos. Cell 50, 89-95. Shackleford, G. M., Willert, K., Wang, J., and Varmus, H. E. (1993). The Wnt-1 proto-oncogene induces changes in morphology, gene expression, and growth factor responsiveness in P C I2 cells. Neuron 11, 865-875. Shankar, D. B., and Shackleford, G. M. unpublished results. . Sinn, E.. Muller, W., Pattengaie, P., Tepler, I., Wallace, R., and Leder, P. (1987). Coexpression of MMTV/v-Ha-ras and MMTV/c-myc genes in transgenic mice; synergistic action of oncogenes in vivo. Cell 49, 465- 475. Slamon, D. J., Clark, G. M., Wong, S. G., Levin, W. J., Ullrich, A., and McGuire, W. L. (1987). Human breast cancer: Correlation of relapse and survival with amplification of HER2/neu oncogene. Science 235, 177- 179. Slamon, D. J., Dekemion, J. B., Verma, 1 . M., and Cline, M. J. (1984). Expression of cellular oncogenes in human malignancies. Science 224, 256-262. 164 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Steinmetz, M., Hocht, J., Schnell, H., Gebhard, W., and Zachau, H. (1980). Cloning of V region fragments from mouse liver DNA and localization of repetitive DNA sequences in the vicinity of immunoglobulin gene segments. Nucl. Acids Res. 8, 1721-1729. Stewart, T. A., Pattengaie, P. K., and Leder, P. (1984). Spontaneous mammary adenocarcinomas in transgenic mice that carry and express MMTV-myc fusion genes. Cell 38, 627-237. Strasser, A., O'Connor, L., Huang, D. C., O'Reilly, L. A., Stanley, M. L., Bath, M. L., Adams, J. M., Cory, S., and Harris, A. W. (1996). Lessons from bcl-2 transgenic mice for immunology, cancer biology and cell death research. Behring Inst Mitt, 101-117. Sukumar, S. (1990). An experimental analysis of cancer; role of ras oncogenes in multistep carcinogenesis. Cancer cells 2, 199-204. Tanaka, A., Miyamoto, K., Minamino, N., Takeda, M., Sato, B., Matsuo, H., and Matsumoto, K. (1992). Cloning and characterization of an androgen- induced growth factor essential for the androgen-dependent growth of mouse mammary carcinoma cells. Proc. Natl. Acad. Sci, USA 89, 8928- 8932. Theillet, C., Liderau, R., Escot, C., Hutzell, P., Brunet, M., Gest, J., Schlom, J., and Callahan, R. (1986). Loss of a c-H-ras allele and aggressive human primary breast carcinomas. Cancer Res. 46, 4476-4781. Thomas, S. M., Moreno, R. P., and Tilzer, L. L. (1989). DNA extraction with organic solvents in gel barrier tubes. Nuc. Acids Res. 17, 5411. Trent-J.M. (1985). Cytogenteic and molecular biological alterations in breast cancer; a review. Breast cancer research treatment 5, 221 -229. Tsukamoto, A. S., GrosschedI, R., Guzman, R. C., Parslow, T., and Varmus, H. E. (1988). Expression of the /nf-1 gene in transgenic mice is associated with mammary gland hyperplasia and adenocarcinomas in male and female mice. Cell 55, 619-625. 165 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. van Lohuizen, M., Verbeek, S., Krimpenfort, P., Domen, J., Saris, C., Radaszkiewicz, T., and Berns, A. (1989). Predisposition to iymphomagenesis in p/m-1 transgenic mice: cooperation with c-myc and N-myc in murine leukemia virus-induced tumors. Cell 56, 673-682. van Lohuizen, M., Verbeek, S., Scheijen, B., Wientjens, E., van der Gulden, H., and Bems, A. (1991). Identification of cooperating oncogenes in Em-myc transgenic mice by provirus tagging. Cell 65, 737-752. Varmus, H. (1989). Transgenic mice and host cell mutants resistant to transformation as model systems for identifying multiple components in oncogenesis. Ciba. Found. Symp. 142, 20-6. Vaux, D. L., and Strasser, A. (1996). The molecular biology of apoptosis. Proc. Natl. Acad. Sci. USA 93, 2239-2244. Wang, T. C., Cardiff, R., Zukerberg, L., Lees, E., Arnold, A., and Schmidt, E. V. (1994). Mammary hyperplasia and carcinoma in MMTV-cuclin D1 transgenic mice. Nature 366, 669-671. Weber, L. M., and Garber, J. E. (1996). Familial Breast cancer. In Diseases of the breast, J. R. Harris, Lippman, M.E., Morrow,M., Heilman,S., ed. (Philadelphia: Lippcott-Raven), pp. 168-185. Weber-Hall, S. J., Phippard, D. J., Niemeyer, 0. C., and Dale, T. C. (1994). Developmental and hormonal regulation of Wnt gene expression in the mouse mammary gland. Differentiation 57, 205-214. White, E. (1996). Life, death, and the pursuit of apoptosis. Genes Dev. 10, 1-15. Whyte, P., Buchovich, J. J., Horowitz, J. M., Friend, S. H., Raybuck, M., Weinberg, R. A., and Harlow, E. (1988). Association between an oncogene and an anti-oncogene: the adenovirus El A proteins bind to the retinoblastoma gene product. Nature 334, 124-129. 166 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Wilkinson, D. G., Bailes, J. A., and McMahon, D. G. (1987). Expression of the proto-oncogene /nf-1 is restricted to specific neural cells in the developing mouse embryo. Cell 50, 79-88. Wyllie, A. H., Kerr, J. F., and Currie, A. R. (1980). Cell death: the significance of apoptosis. Int. Rev. Cytol. 68, 251-306. Van, G., Nikolanopoulos, S., Wang, P., and McKeehan, W. L. (1991). Sequence of rat keratinocyte growth factor (heparin-binding growth factor type 7). Cell. Dev. Biol. 27A, 437-438. Yayon, A., Klagsbrun, M., Esko, J. □., Leder, P., and Omitz, D. M. (1991). Cell surface, heparin-like molecules are required for binding of basic fibroblast growth factor to its high affinity receptor. Cell 64, 841-8. Yoshida, T., Miyagawa, K., Odagiri, H., Sakamoto, H., Little, P. F., Terada, M., and Sugimura, T. p988). Genomic sequence of hst, a transforming gene encoding a protein homologous to fibroblast growth factors and the /nf-2-encoded protein. Proc. Natl. Acad. Sci. USA 84, 7305-7309. Yuan, J. (1997). Genetic control of cellular suicide. Reprod. Toxicol. 11, 377-384. Yuspa, S. H., and Poirier, M. C. (1988). Chemical carcinogenesis: from animal models to molecular models in one decade. Adv. Cancer Res. 50, 25-70. Zhan, X., Bates, 8., Hu, X. G., and Goldfarb, M. (1988). The human FGF-5 oncogene encodes a novel protein related to fibroblast growth factors. Mol. Cell. Biol. 8, 3487-3495. 167 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. IMAGE EVALUATION TEST TARGET ( Q A - 3 ) 1 . 0 i.i 1 . 2 5 2.8 1 .4 m 2.2 2.0 .8 1 . 6 1 5 0 m m < 9 . / / o 7 / j /APPLIED A IN/HGE . Inc 1653 East Main Street Rochester. NY 14609 USA Phone; 716/482-0300 — Fax: 716/288-5969 e 1 9 9 3. Applied im ag e. Inc.. All Rights Reserved Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Identification of oncogenes cooperating in murine mammary tumorigenesis and transgenic mouse models of breast cancer

PDF

A Study Of The Mechanism Of Lead Chromate Induced Neoplastic Transformation Of C3H/10T1/2 Mouse Embryo Cells

PDF

Characterization of the interactions between cellular proteins and hepatitis C virus core protein.

PDF

FGF8, isoform b, is an etiological factor in prostate tumorigenesis

PDF

Biochemical factors determining tumor response to 5,10-dideazatetrahydrofolate, a folate antimetabolite inhibitory to de novo purine biosynthesis

PDF

Isolation, Separation, And Structural Identification Of Streptomyces Quinones

PDF

Development of replication -competent retroviral vectors for efficient, targeted gene therapy of cancer

PDF

Analysis of the function and regulation of the 94kDa glycoprotein of the glucose-regulated protein family.

PDF

A POU gene expressed in the early sea urchin embryo: Structure, evolution and requirement for embryonic development

PDF

Characterization of the physiological ligand and function of a novel opsin RGR from the retinal pigment epithelium

PDF

Wnt factors: Genes, functions and cooperative signaling in mammary cell transformation

PDF

Oncogenic changes and induction of differentiation in pediatric astrocytic cell lines

PDF

Immunogenicity of acute lymphoblastic leukemia: In vivo and in vitro studies of Bcr -Abl specific immune responses

PDF

Studies on the regulation of p53 and investigation of functional significance of interactions between p53 and its binding proteins

PDF

Changes of DNA methylation patterns during human development and transformation.

PDF

Studies On Antigens Associated With Wilm'S Tumor (Nephroblastoma)

PDF

Analysis of Sonic hedgehog/Patched-1 downstream genes in embryonic feather morphogenesis and the development of novel biotechnologies thereof

PDF

The Human 3 Beta-Hydroxysteroid Dehydrogenase Type Ii Gene And Genetic Predisposition To Prostate Cancer

PDF

Anomalies at NGX6 locus: Potential involvement in feline lymphomas

PDF

Functional analysis of MSX2 and its role in skull patterning

Asset Metadata

Creator Bhavani Shankar, Deepa (author)

Core Title Analysis of multistep mammary tumorigenesis in Wnt1 transgenic mice: The role of fibroblast growth factor-8 in oncogenesis and apoptosis

Degree Doctor of Philosophy

Degree Program Molecular Microbiology and Immunology

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag biology, cell,biology, microbiology,biology, molecular,health sciences, oncology,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Shackleford, Greg (committee chair), Fung, Yuen Kai (committee member), Roy-Burman, Pradip (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c17-313190

Unique identifier UC11353861

Identifier 9835059.pdf (filename),usctheses-c17-313190 (legacy record id)

Legacy Identifier 9835059.pdf

Dmrecord 313190

Document Type Dissertation

Rights Bhavani Shankar, Deepa

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA