, , 435 new J'JK S g. 2., l v‘. '1'. 1‘”: cyan-4:: :' unwuy (l‘h‘fiwp r- ll‘ «(.5 v~ - w«. u» ll gm.“ 5 W2! ‘9’." -h u. ‘.r . .. z. “33;? ' garga 2:12;,- . . a»; .- tr. fizz‘i'fi‘gfir ‘ ; ‘53,. ‘a u. . ‘ 3‘ U 3... T'.. 3353;} "J" .1 .. -, I '9' ~ ” P4 " 'n"> " In. 1”,, T‘M‘t 1.5." , 7113::- :15? I v ‘ ‘ I I “ .. WEE. warm: 'r"4“"- " “@131: J! ' . TV’V " “g tin-pa A: '1” 3.249: n 0.. \ . r“ by“ . r ,‘. £3,315}? :3. ”3153‘s IGAN STAT will“ummumfillfislillfllgfil 3 1293 010 This is to certify that the dissertation entitled Sequencing PCR amplified ras genes to determine the importance of mutational activation of endogenous ras genes to the malignant transformation of human fibroblasts presented by Jeanette Marie Scheid has been accepted towards fulfillment of the requirements for Ph.D. degreein MlCY‘OblOlOQ)’ Major professor Datew MS U i: an Affirmative Action/Equal Opportunity Institution 0-12771 LIBRARY Michigan State Universlty PLACE IN RETURN BOX to romovo this checkout from your record. TO AVOID FINES return on or baton date duo. DATE DUE DATE DUE DATE DUE MSU Is An Minn-tin Adlai/Equal Opportunlty Inotltwon Wyn-9.1 SEMIEIUG PCR NI’LIFIEO RAS GEES TO OETERHIHE THE INTAKE (H7 HUTATIOHAL ACTIVATION OF EHOOGEHOUS RAS'GEHES TO THE HALIGHAHT TRANSFORMATION OF HOHAH FIBROBLASTS By Jeanette Marie Scheid A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Microbiology 1994 ABSTRACT SEMICI'B PCR MIFIEO RAS GENES TO DETERMINE THE INTAKE (I: MUTATIONAL ACTIVATION OF ENOOGENOOS RAS GENES TO THE MALIGNANT TRANSFORMATION OF HUMAN FIBROBLASTS By Jeanette Marie Scheid Activation of ras proto-oncogenes by mutation at specific codons has been documented in a variety of human malignancies. To determine the extent to which ras activation plays a role in thetmalignant transformation of human fibroblasts, I sequenced the coding region of the H-, K-, and N-ras genes from a series of human fibroblast cell lines, including normal diploid human fibroblasts, two morphologically normal non—tumorigenic infinite life span cell strains originated in this laboratory, spontaneous and carcinogen-induced malignant transformants of the latter, and cell lines derived from human fibrosarcomas. Direct sequencing of the PCR product was carried out using the dideoxy chain termination method. A total of twelve cell lines or strains were analyzed for ras mutations. 0f the eight that are malignant two fibrosarcoma-derived cell lines contained an activating ras mutation, a frequency of 25%. One mutant cell line, 8387, has a G to C transversion that causes an Ala to Pro amino acid substitution at codon 146 in one allele of the K-ras gene. This same mutation has been reported to activate K-ras in Hut-14 cells, an 8387 derivative. The other mutant cell line, HT1080, has a C to A transversion that causes a Gln to Lys amino acid substitution at codon 61 in one allele of the N-ras gene that has been reported previously. ‘The remaining cell lines or strains contain no mutations in ras. The mutation frequency in the fibroblastic cells tested in this study is consistent with one other report on ras activation in human sarcomas in the literature. My results show that although ras oncogenes can play a role in the development of fibrosarcomas in humans, some other changes must be able to substitute for ras activation in the malignant transformation of human fibroblasts. Also, methods of detecting ras activation that concentrate on the "classic" activating codons may miss activated genes. Further studies are planned to test for alterations elsewhere in the ras p21 signalling pathway using cells that contain only wild-type ras genes. This work is dedicated to my parents, Howard and Margaret Scheid, to my first role model, Mrs. Patricia Pregitzer, and to my partner, Cindy, for their caring and support. iv ACKNOHLEOGEMENTS The people from whom I received invaluable help and support throughout the pursuit of this degree are too numerous to count; however, I would like to specifically thank the people who played particularly important roles: My major professor, Dr. J. Justin McCormick, guided my work and unfailingly provided me with support and valuable advice in all areas of my professional life. He is a mentor in the truest sense, and I cannot adequately express my gratitude. Dr. Veronica M. Maher provided a wealth of knowledge, advice, energy, and support. She has been an inspiration as well as a role model. The members of my guidance committee, Drs. Richard Schwartz, Maria Patterson, and Margaret Jones provided significant thoughtful and helpful commentary during the course of our several meetings. Dr. Lynn Farquahr steered me through the minefields of the medical scientist training program (MSTP), and provided invaluable moral and financial support. Finally, all my colleagues in the Carcinogenesis Laboratory have unselfishly shared their knowledge and skills. However, I would especially like to thank Dr. Jia-ling Yang, Dr. Ruey-Hwa Chen, Dr. Glenn McGregor, Dr. David Reinhold, Dr. Rebecca Odenwaller, Dr. Chia Miao Mah, Chien Cheng Lin, Ann Ryan, Lonnie Milam, Jeremy Nray, Suzanne Kately- Kohler and Maie ElKassaby, for their help and support during the course of my studies. TABLE OF CONTENTS Page LIST OF TABLES ....................................................... x LIST'OF FIGURES ...................................................... xi INTRODUCTION ......................................................... 1 List of References ................................................. 4 CHAPTER 1. LITERATURE REVIEW 1. Theory of Multi-Step Carcinogenesis ......................... 5 A. In Vivo Evidence in Support of a Multi-Step Theory of Carcinogenesis .................. 6 1. Epidemiological ................................. 6 2. Experimental .................................... 7 3. Clinical ........................................ 8 B. In Vitro Evidence in Support of a Multi-Step Theory of Carcinogenesis .................. 9 1. Early Experiments ............................... 9 2. Current Experiments ............................. 11 a. Experiments using Syrian Hamster embryo cells .............................. 11 b. Experiments using human epithelial cells ..................................... 12 c. Experiments using human fibroblasts ....... 13 11. Genetic Basis for Malignant Transformation .................. 14 A. DNA Tumor Viruses .................................... 14 1. Adenovirus ...................................... 15 2. Papova Viruses .................................. 17 a. Polyoma virus ............................. 18 i. Large T antigen ..................... 18 ii. Middle T antigen .................... 19 iii. Small t antigen ..................... 19 vi b . SV40 ...................................... 20 i . Large T antigen ..................... 20 -Structure and function .......... 20 -Mechanism ....................... 21 ii. Small t antigen ..................... 23 c. Papilloma viruses ......................... 24 3. Herpes viruses .................................. 26 B. Retroviruses ......................................... 28 1. Transduction of an Activated Gene ............... 29 2. Insertional Activation .......................... 32 3. Alteration of a Gene Product .................... 34 4. Insertional Inactivation ........................ 38 5. Trans-Activation of Cellular Genes .............. 38 C. Transformation by Oncogenes .......................... 39 1. Dominant-Acting Oncogenes ....................... 40 a. Genes encoding membrane- associated proteins ....................... 40 b. Genes encoding nuclear proteins ........... 42 2. Tumor Suppressor Genes .......................... 44 a. The retinoblastoma susceptibility gene ....................... 45 b. The p53 gene .............................. 46 III. Epigenetic Contributions to Malignant Transformation ....... 47 A. Developmental and Environmental Theories ............. 47 8. Methylation .......................................... 49 C. Protein Kinase C ..................................... 50 IV. ras Proto-Dncogenes and Malignant Transformation ........... 51 A. General Properties of the ras Family ................. 51 1. ras Family ...................................... 51 2. Structural Properties of ras .................... 52 a. DNA ....................................... 52 b. Protein ................................... 53 3. Role of ras in Normal Cellular Function ......... 55 vii 8. Evidence for the Role of ras in Malignant Transformation ............................. 56 1. Evidence from Correlative Data .................. 56 a. Human tumor cells ......................... 56 b. Animal tumor cells ........................ 57 c. Cells in culture .......................... 58 2. Evidence from Insertion of Activated ras Genes ......... 7 .............................. 59 3. Evidence from Reversion of ras-Induced Transformation ...................... 60 C. Mechanisms of ras-Mediated Transformation ............ 62 1. Point Mutation-Induced Change in ras p21 Activity ................................ 62 2. C00peration between ras and Other Oncogenes ................................. 64 3. Alterations in ras Expression ................... 65 V. Methods for Determining the Presence of Activated ras Genes ........................................ 68 A. Early Methods ........................................ 68 8. Later Methods ........................................ 69 LIST OF REFERENCES ................................................... 74 CHAPTER II. SEQUENCING PCR-AMPLIFIED RAS GENES TO DETERMINE THEIR IMPORTANCE TO THE MALIGNANT TRANSFORMATION OF HUMAN FIBROBLASTS ............ 91 Abstract ....................................................... 92 Introduction ................................................... 94 Materials and Methods .......................................... 101 Routine Cell Culture ..................................... 101 RNA Extraction ........................................... 101 Preparation of first strand cDNA ......................... 102 Second strand cDNA synthesis and in vitro amplification using the polymerase chain reaction (PCR) .......................... 102 Direct sequencing of PCR-amplified cDNA .................. 103 Cycle sequencing ......................................... 106 viii Results ........................................................ 106 Determination of the status of the ras genes in cells from a series of f ibrosarcoma cell lines .................................. 106 Determination of the status of the ras genes in cells from the MSU 1 lineage ................ 107 Discussion ..................................................... 109 References ..................................................... 115 APPENDIX ............................... ' .............................. 120 A. Introduction ............................................. 120 B. Intermediates Between Growth Factor Receptors and ras p21 ...................... 121 C. Intermediates Between ras p21-GTP and the Nucleus .......................................... 123 1. Intermediates that Appear to Interact Directly with p21-GTP ..................... 123 2. An Intermediate that Interacts Directly'with the raf-l Protein .................... 126 0. Conclusions .............................................. 126 LIST OF REFERENCES ............................................. 129 ix LIST OF TABLES Table Page Chapter II 1. Cell lines derived from human fibrosarcomas ................... 96 2. Cell strains from the MSU-1 lineage ........................... 98 3. Summary of ras gene mutations ................................. 107 LIST OF FIGURES Figure Page Chapter II 1. MSU-1 lineage ................................................. 100 2. Oligonucleotide primers used for PCR and sequencing ........... 105 3. Point mutation in the Ki-ras gene of 8387 ..................... 108 Appendix 1. Consensus model of ras-mediated signalling pathway ............ 128 xi INTRODUCTION Malignant transformation of cells has been established through a variety of epidemiological and experimental studies to be a multi-step process. Through the accumulation of stable heritable changes an individual cell loses regulation of growth and differentiation, and gains the ability to invade into adjacent, normal tissue and to metastasize to distant body sites. The overall goal of investigators who study this process is to understand the nature of the events that result in transformation, and to find the means either to interrupt the process before a patient becomes ill with a malignancy, or to treat a malignant tumor once it has developed. In the course of decades of investigation, researchers have identified many genes coding for proteins that are involved in regulation of growth and differentiation in normal cells. These genes have been shown to be altered in characteristic ways in tumor cells, and therefore to encode proteins that no longer provide normal regulatory fUnction. These genes have been categorized into one of two groups: dominant-acting oncogenes and tumor suppressor genes. The dominant—acting oncogenes need only one of two alleles to be activated in order to have transforming activity. Within this group are genes encoding proteins that are associated with the plasma membrane (e.g., growth factors and growth factor receptors), genes that encode cytosolic proteins (e.g., protein kinases), and genes that encode nuclear proteins (e.g., DNA binding 1 2 proteins)(8ishop, 1983). The tumor suppressor genes exert their transforming effect only when both alleles are altered such that no functional protein is produced. The majority of the products of tumor suppressor genes are found in the nucleus, bind to DNA, and act as regulators of gene expression (Levine and Momand, 1991). The various ras genes are members of the membrane-associated dominant—acting oncogenes. They encode proteins classified as "G- proteins" because their activity is regulated by changes in tertiary structure brought about by binding to the guanine nucleotides GTP and GDP. The proteins encoded by these genes have a molecular weight of 21k0a, are anchored to the internal surface of the plasma membrane, and function as transducers of the mitogenic signals from protein growth factors (Barbacid, 1987). Ever since the discovery that the transforming gene in the acutely transforming Harvey'murine sarcoma virus had a normal cellular homolog (Parada et al., 1982), these genes, and their products, have been the subject of intense study. It is now known that activation of the ras genes by mutation at specific codons is a very common event in some types of human malignancy and a relatively rare event in others (Bos, 1989). Whether the source of this variability reflects differences between cell types in the process of malignant transformation, is a function of limited sensitivity of available methods, or is the result of some combination of the two, is unclear. This study was undertaken to determine the extent to which ras activation occurs in the malignant transformation of one type of cell, i.e., human fibroblasts. Human and rodent fibroblasts in culture are commonly used as model systems to study the process of malignant transformation, and are readily transformed by insertion of activated ras 3 genes. Therefore, it is important to determine whether activation of endogenous ras genes is commonly involved in the transformation of such cells. With the advent of polymerase chain reaction (PCR), it became possible to amplify specifically the entire coding region of the ras genes, and use this amplified fragment as a template for direct sequencing. Previous efforts to determine the frequency of ras activation have concentrated on the sequence of the specific codons that are known to cause transformation to focus formation in mouse NIH-3T3 cells. However, it is possible that activating mutations that do not cause transformation in that focus assay exist. Therefore, sequencing the entire coding region provides the most sensitive means to determine whether or not ras genes have been activated. This first chapter of this dissertation reviews the literature on the process and mechanisms involved in malignant transformation, and specifically the role of ras genes in that process. The second chapter is a paper that will be submitted to the Journal Molecular Carcinogenesis reporting the research I carried out for this dissertation and is written in the format required for that Journal. Finally, the Appendix section gives a review of recent data on the signal transduction pathway involving ras p21, and places the results of this thesis project in the context of these new data. LIST OF REFERENCES Barbacid. M. 1987. ras genes. Annu. Rev. Biochem. 56:779-827. Bishop, J. M. 1983. Cellular oncogenes and retroviruses. Annu. Rev. Biochem. 52:301-354. Dos, J. L. 1989. ras oncogenes in human cancer: a review. Cancer R e s . 49:4682-4689. Levine, A. J., and J. Monand. 1990. Tumor suppressor genes: the p53 and retinoblastoma sensitivity genes and gene products. Biochim. Biophys. Acta 1032:119-136. Parada, L. F., C. J. Tabin, C. Shih, and R. A. Neinberg. 1982. Human EJ bladder carcinoma oncogene is homologue of Harvey sarcoma virus r a 5 gene. Nature 297:474-478. CHAPTER 1 LITERATURE REVIEH 1. WW Cancer cells have been reported to differ from normal cells by many various characteristics. Among the’most fundamental of these are the loss of normal growth control and the acquisition of infinite life span leading to the ability of the malignant cells to invade into normal tissue and/or metastasize to distant body sites. The process by which normal cells become malignant is considered to consist of several steps. Through successive stable heritable changes and clonal expansions, a cell gains the ability to divide without requiring exogenous growth factors, to continue to divide without reaching senescence, and to express protein products such as proteolytic enzymes that allow invasion and metastasis. These and other changes collectively contribute to malignancy. This theory of multi-step carcinogenesis is supported by several kinds of studies, including epidemiological, experimental, and clinical. 6 A. In Vivo Evidence in Support. of a. MDlti-Stepi‘Theory' of Carcinogenesis I. Epideliological Peto (1977) studied epidemiological data on the incidence of several kinds of cancer. He was particularly interested the role of cigarette smoking in lung carcinoma. The data showed that a linear dose-response relationship exists between daily cigarette consumption and lung cancer incidence, and the extra incidence of lung cancer after an individual quits smoking stays approximately constant instead of continuing to rise. By studying these data in the context of several mathematical models, he found that the best fit was the multi-step model. He concluded that multi-step models would provide the most useful insights into the origins of cancer. In 1981 Moolgavkar and Knudson developed a two-stage model of carcinogenesis. The first stage is the result of an event leading to uncontrolled cell proliferation, and the second stage is a result of an event that occurs within one of the proliferating cells and gives rise to a malignant cell that grows into a clinically apparent tumor. They showed that this model was consistent with epidemiological data on age-specific cancer rates for children and adults, on smoking and lung cancer, and on heredity and exposure to hormones and radiation in breast cancer. In 1986 Kaldor et al. obtained and analyzed data on lung and nasal sinus cancer death rates in a group of nickel refinery workers. They concluded that calculations of excess mortality and risk are consistent with a multi-step model. In this model, type of hazard exposure, time of initial exposure, and duration of that exposure are variables that can be 7 used in calculating risk. Clearly, epidemiological methods have contributed to formulating multi-step models of malignant transformation. 2. Experilental Another source) of support for’ multi-step models of malignant transformation comes from the induction of cancer in experimental animals. A typical example is the 1964 study of Boutwell, who induced squamous cell carcinomas in the skin of susceptible mouse strains by treating once with a sub-carcinogenic dose of dimethylbenz(a)anthracene (DMBA), followed by repeated applications of the-non-carcinogenic agent, 12-0- tetradecanoylphorbol-13-acetate (TPA). Neither agent alone was sufficient to induce malignant transformation. Transformation followed a progressive course in which benign papillomas appeared initially, and foci of carcinoma developed from within the papillomas. In this experimental system, DMBA is called the initiator, since it induces a permanent change within a cell, and TPA is called a promoter, since it is required for the effect of this sub-carcinogenic dose to become apparent. To determine the mechanisms by which chemicals induce malignant transformation of mouse skin, Verma et al. (1982) induced tumors in mice by two methods. The first was the two-agent initiation/promotion regimen of Boutwell. The second was the single-agent regimen consisting of either a single treatment with a large dose of DMBA, or multiple treatments with an intermediate dose of DMBA (as compared to the dose used in the two- agent regimen). They compared the pattern of tumor formation induced by the two regimens, and found that the two-agent protocol results in the typical pattern of rapid papilloma induction and delayed appearance of foci of carcinoma. In contrast, the single-agent protocol results in 8 fewer papillomas appearing after long latent periods, but more rapid and frequent appearance of carcinomas. They also found that retinoic acid specifically inhibits tumor formation by the two-agent protocol but not by the single-agent protocol. 7,8-benzoflavone, an inhibitor of the metabolic activation of DMBA, inhibited tumor formation in the single- agent protocol; however, the effect of 7,8-benzoflavone on tunor induction by the two-agent protocol was not determined. Since the patterns of tumor formation and inhibition for the two regimens differ, these investigators concluded that the mechanisms by which DMBA and TPA promote tumor formation are different. In addition, they speculated that promotion by DMBA is due to a regenerative hyperplasia following the tissue damage that is known to be caused by this agent. 3. Clinical The induction of tumors in humans is clearly unethical. However, rare situations in which the susceptibility to a specific type of tumor is inherited offers the possibility of studying carcinogenesis in humans. The disease familial adenomatous polyposis (FAP) has a number of characteristics that make) it ideal for such studies. This is an autosomal-dominant genetic disease in which patients develop hundreds of adenomatous polyps in the colon and rectum. One or more of these polyps inevitably progresses to invasive colorectal carcinomas. Vogelstein et al. (1988) studied events associated with the progression of these tumors in two groups of patients, one with FAP and the other with sporadic colon carcinoma. They assayed the cells from individual polyps for four specific genetic changes that have a documented association with colorectal neoplasia, namely, the presence of mutationally activated ras 9 genes, and allelic deletions of chromosomes 5,17, and 18. They found that polyps displaying a more-transformed phenotype contained cells with more alterations than did cells from less-transformed polyps. The least- transformed adenomas typically had only undergone one change, most often ras activation, whereas carcinomas exhibited ras activation as well as deletions in all three of the chromosomes cited above. In the "ore- transformed adenomas an intermediate number of changes could be detected. From their data they concluded that 1) a number of changes are required for malignant transformation of colonic epithelial cells, 2) the specific changes they observed do not substitute for each other since in most transformed groups of cells, all four of the changes can often be found, and 3) some carcinogenic change(s) must exist that can substitute for the four they studied, sinceiall four alterations were not necessarily present in a carcinoma, and other changes have been reported. 8. In Vitro Evidence in Support of a Multi-Step Theory of Carcinogenesis 1. Early Experilents Early in vitro transformation studies indicated that carcinogenesis was not the complex process predicted by in vivo results. As early as 1941, Gey reported the occurrence of spontaneous malignant transformation in cultured rat fibroblasts. Earle (1943) reported similar results with mouse fibroblasts in culture. However, by the late 1960’s it was recognized that certain conditions of the existing in vitro culture methods (serum, light, oxygen levels) can induce transformation. Parshad et al. (1978) reported an experiment in which cultured mouse fibroblasts 10 were grown either in the presence or in the absence of fluorescent light. The cells grown in fluorescent light exhibited a significantly increased incidence of double strand chromosome breaks, and this damage could be decreased by the addition of reducing agents such as catalase and glutathione to the medium. The authors proposed that hydrogen peroxide produced in the medium exposed to light is responsible for the genetic damage. This could explain the frequency of so-called spontaneous transformation found by other researchers. To test this hypothesis, Sanford et al. (1979) studied the effects of fluorescent light and oxygen levels on the amount of DNA damage and the frequency of malignant transformation in cultured mouse fibroblasts. They showed that fibroblasts exposed to fluorescent light and atmospheric levels of oxygen exhibit more chromosome damage and an increased frequency of «alignant transformation compared to fibroblasts that are shielded from light and grown at lower oxygen concentration. It also appeared from early experiments that the malignant transformation of cells by infection with transforming viruses was a single step process. Halberstaedter et al. (1941) demonstrated that infection with a single agent, the Rous sarcoma virus, was sufficient to cause morphological transformation of chick fibroblasts in culture. Similar results were reported for infection of rat fibroblasts with the Harvey sarcoma virus. However, it is now recognized that what appeared to be a single event, i.e. viral infection, is in reality at least two events. The transforming viruses carry activated forms of cellular proto- oncogenes, for example, H-ras in the Harvey sarcoma virus (Parada et al., 1982). Also, the construction of these viruses places the genes under the transcriptional control of the viral long terminal repeats (LTR), 11 typically causing enhanced expression of the gene(s) carried by the virus compared to the expression levels that would be observed for the corresponding endogenous gene controlled by its endogenous promoter. 2. Current Experilents A series of in vitro experimental systems is currently being used to gain a clearer understanding of the events involved in malignant transformation. These include Syrian hamster embryo (SHE) cells, human epithelial cells, and human fibroblasts. ri n u m r ll In 1978 Barrett and Ts’o provided support for the multi-step theory of carcinogenesis in their work with cultured Syrian hamster embryo cells exposed to the chemical carcinogen, benzo(a)pyreneu Cells in bulk culture were’exposed to benzo(a)pyrene, then maintained in bulk culture conditions for up to seventy-five population doublings. Periodically during this time, cells from the bulk culture were cloned and then assayed for morphological transformation, fibrinolytic activity, and ability to grow in soft agar. Within two weeks of exposure to this agent, some cells became morphologically transformed (abnormal appearance and organization, increased cell density) and showed enhanced fibrinolytic activity. However, only after 32-75 population doublings post-exposure were some cells in the population found that displayed the ability to grow in soft agar (anchorage independence). Each of these populations that exhibited anchorage independence proved capable of forming tumors in test animals after a long latent period. The requirement for an extended period of growth after treatment in order to obtain malignant cells indicates that 12 subsequent events are needed to fully transform these cells. 1 1 ll Fewer examples of in vitro transformation exist for human epithelial cells systems owing to the historic difficulty in generating epithelial cell cultures. One problem was that serum-containing media promotes the overgrowth of fibroblasts when primary cultures are being cultivated. With the advent of serum-free medium, the establishment of epithelial cells became less complicated. Using such a medium, Rhim et al. (1985) found that successive infection with two different transforming viruses is required in order to fully transform cultured primary epidermal keratinocytes. Hhen these cells were infected with a hybrid of adenovirus 12 and simian virus 40 (SV40), they acquired an infinite life span and exhibited abnormal morphology, but they were not able to form tumors after injection into athymic mice. However, when these partially transformed cells were infected with the Kirsten murine sarcoma virus, they formed foci (localized areas of piled up cells), grew to a higher saturation density in culture, showed anchorage independence, and formed progressively growing tumors within three weeks of injection into athymic mice. Another example of multi-step transformation of human epithelial cells is a study by Stampfer and Bartley (1985), who exposed primary cultures of normal human manmary epithelial cells to the carcinogen benzo(a)pyreneu Following exposure, some cells displayed an extended life span and morphological changes. Two immortal cell lines arose from these extended life span cells. These two lines were dissimilar in morphology, karyotype and surface protein expression, but both were shown to be 13 progeny of the normal parent line. Although the immortal cells displayed some of the characteristics of malignant mammary cell lines, they formed colonies in soft agar only at a very low frequency and were not tumorigenic. In later work with these partially transformed cell lines, Stampfer and colleagues (Clark et al., 1988) reported that the presence of both the SV40 T antigen and ras oncogenes (inserted by retroviral infection) was required to complete the malignant transformation of the immortal cell lines. r i u fi r One example of transformation using cultured human fibroblasts is a study by Namba et al. (1986), who induced malignant transformation of human fibroblasts by. repeated exposure to “Co gamma radiation, followed by infection with the Harvey murine sarcoma virus. Repeated treatment with “Co resulted in immortal cells that had multiple karyotypic abnormalities, but were neither anchorage independent nor tumorigenic. However, when the imnortal fibroblasts were infected with the Harvey murine sarcoma virus, some became anchorage independent and tumorigenic in athymic mice. Studies using normal human fibroblasts in culture were also carried out in our laboratory. Normal diploid human fibroblasts with a finite life span were transfected with a plasmid carrying a selectable marker and an activated H-ras oncogene (Hurlin et al., 1987). The plasmid was engineered to allow over-expression of the oncogene. Transfected cells that expressed the oncogene were morphologically transformed, focus- forming, and anchorage independent. The transformed cell lines were expanded in culture for tumorigenicity studies. However, during this 14 expansion the cells showed decreased ras p21 expression and reversion to normal morphology. These revertants did not form foci or colonies in soft agar, nor were they tumorigenic. All the transformants senesced after 40- 50 population doublings as is typical for normal fibroblasts. Once an inlnortal fibroblast line had been generated (Morgan et al., 1991), transfection of this activated ras oncogene (Hurlin et al., 1989) or other ras oncogenes (Wilson et ial., 1990; Fry' et al., 1990) malignantly transformed these cells. The above studies provide strong evidence in support of the concept of multi-stepped carcinogenesis. III order to understand malignant transformation more completely, one must consider the changes that occur during the process. The next section of this chapter will outline the genetic mechanisms that play a role in malignant transformation. 11. W A. DNA Tulor Viruses Infection of permissive mammalian cells by members of the DNA tumor virus family typically results in lysis of the affected cell. Control of the lytic cycle is mediated by the proteins coded by viral genes that are transcribed early after infection (early region genes). They act to disrupt the normal pattern of protein expression in the host cell and to regulate the production of viral progeny. Hhen infection of non- permissive cells occurs, the host cells are not lysed, but on rare occasions may become transformed. Transformation is thought to be such a rare event because in order for the virus to permanently disrupt normal growth of the host, the early genes must continually express proteins. 15 This requires at minimum the integration of that 5’ regulatory region of the virus into the host genome such that the early protein(s) are expressed. The study of these viruses (adenovirus, papovavirus, and herpes virus) at the genetic and molecular level has revealed the identity of many of these transforming proteins, and has yielded an understanding of the mechanisms by which their interactions with the host proteins release the host cell from its normal growth regulation. 1. Adenovirus Theiadenovirus (Ad) family is comprised of multiple serotypes. None have been linked to cancer in humans, but each has the capacity to transform human and animal cells in culture and to induce tumors in animals. Three serotypes, Ad 2, 5, and 12, are similar in the sequence and organization of their early region genes. Several studies illustrate the role of the Ad early genes in transformation. In a series of experiments in which fragments of adenovirus genome were transfected into rodent cells in culture (reviewed by Petterson and Roberts, 1986), eight percent of the 5’ region, known as E1, was determined to contain the necessary'and sufficient transforming sequences. E1.has been divided into two genes, EIA and £18, each contributing uniquely to the process of malignant transformation. EIA encodes two major proteins, 243 and 249 amino acids respectively. These proteins are found in the nucleus of the cell, share amino and carboxyl terminal sequences, but differ in their central region by 43 amino acid residues that are unique to the larger protein. Domains of high sequence similarity among several serotypes have been identified and associated with specific transforming properties. Domains 1 and 2, 16 contained in both protein products, act to imortalize the host cell (perhaps in part as a result of binding to the product of the retinoblastoma susceptibility gene, RB (thte et al., 1988)), induce production of growth factors (also implicated by Quinlan et al., 1988 in immortalization), and cooperate with ras oncogenes to transform primary cultured cells to malignancy (Ruley, 1983). Domain 3 (unique to the 249 amino acid protein) is known to act as a transcriptional trans-activator for both viral and cellular genes. The EIB gene encodes two major polypeptides. The first, 179 amino acids in length (19kDa), is thought to act both as a growth factor and as a regulator’ of virion production. In studies of viruses carrying mutations in the region encoding this protein, Subramian et al. (1984) described a deletion mutant, cyt, that caused extensive destruction of human KB cells upon infection. This mutant was also significantly less- capable of inducing cellular transformation in primary rat embryo fibroblasts and in the established rat fibroblast cell line CREF, suggesting that this locus is important in regulating the balance between these two outcomes of infection. Further characterization of these mutants (Subramian et al., 1985) revealed that in the rare instances when the cyt mutant did transform cells, the transformants could not grow as well in low calcium medium as those transformed by wild-type virus. This suggested that the‘wild-type protein acts as a growth factor and decreased the need for the 1.0mM calcium typically utilized by those cells. Calcium is known to act as growth factor in normal fibroblasts (Schilz, 1988), whereas the ability of cells to grow in low calcium is shared by many transformed cells. The other major E18 protein, 495 amino acids in length (55k0a), is phosphorylated on serine and threonine, and has been shown to 17 bind the p53 gene product (Sarnow et al., 1982). The 618 gene alone is incapable of transforming cells, but can act to augment and cooperate‘with the products of EIA. 2. Papova Viruses The papova virus family is divided into two subfamily groupings, the polyoma viruses, including polyoma and SV40, and the papilloma viruses. Although many other viruses are included in the polyoma subfamily, polyoma and SV40 have been studied extensively with regard to transformation and will be used for illustrative purposes here. The genomic structure and transforming mechanisms of polyoma and SV40 are very similar. The transforming region of both viruses has been mapped to the early region. Byialternative splicing, the early gene gives rise to three mRNA species in polyoma (large T, middle T, and small t) and two mRNA species in SV40 (large T and small t). The products of these messages share amino terminal domains but have separate carboxyl terminal domains. In the course of lytic infection, these proteins mediate the expression of viral and host genes for the purpose of producing progeny virus. During transformation, they act to deregulate the host cell in two important aspects, immortalization and release from dependence on exogenous growth factors. 18 a.) Eg]ygma vjrus 1, Large | antigen Wild-type large T antigen of polyoma virus was shown by Rassoulizadegan et al., (1983) to be sufficient to establish continuous cultures from primary rodent fibroblasts. Also, cells that had been transfected with a plasmid expressing a temperature-sensitive mutant of large T antigen went into growth arrest at non-permissive temperatures, showing that expression of a functional large T was necessary for immortalization. Furthermore, using a sequential transfection protocol, fibroblast cell lines that were established by transfection with a vector containing the large T gene could be transformed further to focus formation by expression of middle T protein. Previous studies had shown only a transient transformation of primary fibroblast cultures by middle T antigen alone. Harlow and colleagues (Larose et al., 1991) recently showed that large T antigen binds the retinoblastoma susceptibility gene product (RB), and that mutants defective in R8 protein binding also lack the capacity to immortalize primary rat embryo fibroblasts. This provides a plausible explanation for the mechanism by which large T protein immortalizes primary cells. However, Larose et al. (1991) did find some specific mutants that retained some RB protein binding capacity but were still unable to im'rortalize primary rat embryo fibroblast cultures, suggesting that some other property of large T is also required to immortalize primary cells. 19 11, njgg1e I entigen Middle T appears to release an established cell from normal growth control. The phosphatidyl inositol/protein kinase C signal transduction pathway involving pp60"”“ is normally regulated by phosphorylation of carboxyl terminal residues, presumably as part of the response pathway to exogenous growth factor stimulus (Gould et al., 1988). Cheng et al. (1988) created several deletion and substitution mutants of pp60°*“ and found that phosphorylation of tyrosine 527 is critical for down-regulation of kinase activity. By co-immunoprecipitation they showed that middle I binds pp60”“5 and prevents phosphorylation of this particular residue. This change in phosphorylation pattern correlates with increased transforming potential, as measured by the induction of focus formation. This suggests that middle T stimulates growth by releasing the cell from its usual dependence on exogenous protein growth factors by binding to an intermediate in the signal transduction pathway, pp60c‘m. iii 1 n i n The role'of small t protein in polyoma-induced transformation is not yet clear. Noda et al. (1987) showed that small t antigen can cooperate with middle T antigen in transformation. In this study, two distinct vectors were created, one expressing middle T protein, one expressing small t protein. These vectors were transfected simultaneously into NIH- 3T3 cells and a strain of normal rat kidney (NRK) cells in culture. The transfectants showed increased growth rate, increased cell density, and increased growth in soft agar compared to those that had been transfected with the middle T-expressing vector alone. These properties, which are associated with the ability of cells to form tumors in test animals, were 20 not mediated by increased middle T expression or increased protein tyrosine kinase activity. Therefore, it appears that small t codes for a protein that contributes independently to transformation. :2.) SVQQ — r ur n n The 708 amino acid large T protein of SV40 has a complex structure that allows it to carry out a variety of activities during productive infection (regulation of the expression of both early and late viral genes), and transformation (interaction with cellular products). This protein is required for both initiation and maintenance of transformation induced by SV40. Genetic studies with mutated large T genes have identified domains of the large T protein that are sufficient to transform mammalian cells. Clayton et al. (1982) transfected the established fibroblast cell line Rat-1 with a series of recombinant plasmids carrying different portions of the gene encoding large T; .A plasmid encoding only the amino terminal half of the protein was capable of transforming the recipient cells to focus formation, but with only 1% of the frequency induced by complete large»T protein. Also, the foci induced by the truncated large T protein were morphologically different from those induced by intact large T protein. Other constructs encoding only portions of the amino terminal half of large T protein along with its upstream regulatory sequences could not transform the Rat-1 cells. These investigators concluded that although the amino terminal half alone of 21 large T protein is sufficient to induce transformation of Rat-1 cells, the transformation probably occurs by a mechanism distinct from that of intact large T protein. In 1987 de Ronde et al. created chimeras of the large T gene from SV40 and BKV (a member of the papova virus family that cannot transform human fibroblasts in vitro as can SV40) to determine the regions of the SV40 large T protein responsible for transformation of cultured human fibroblasts. They showed that the amino and carboxyl terminal halves of SV40 large T protein did not morphologically transform human fibroblasts, but did allow their overgrowth under conditions in which contact inhibition would normally occur. These data support the hypothesis that different domains of SV40 large T induce different transformed phenotypes. - n m The mechanism by which large T protein induces transformation is not entirely clear, although several investigators have identified cellular proteins that interact with large T protein. DeCaprio et al. (1988) showed that SV40 large T protein binds the product of the retinoblastoma susceptibility gene, p110-114. Binding occurs at residues 105-114 in the amino terminal half of large T protein. These residues are also required for the transformation by large T protein, supporting the hypothesis that this protein-protein interaction is critical for transformation. This same group of investigators (Ludlow et al., 1989) also showed that large T protein binds to an underphosphorylated form of p110-114, supporting the hypothesis that the underphosphorylated form of p110-114 acts as a tumor suppressor, a function inhibited by large T protein binding. Large T protein also binds the p53 protein (Linzer and Levine 1979), although the 22 role this plays in transformation remains unresolved. Deppert et al. (1989) studied the effect of two kinds of SV40 large T temperature mutants, one that reverts at non-permissive temperatures (N), and one that does not (A), on the stabilization of p53 protein (an event that is thought to contribute to transformation) and on transformation. They found that at non-permissive temperatures the A (non-reverting) type transformant remained transformed and maintained a stable pool of free p53 protein. However, although large T protein did not form a complex with p53 protein, it did continue to associate with chromatin and the nuclear matrix. In contrast, at non-permissive temperatures, the N (reverting) type transformant reverted, did not maintain a stable pool of p53 protein, and the large T protein could neither bind to p53 protein nor associate with chromatin and nuclear matrix. These investigators concluded that maintenance of transformation requires both large T protein activity and p53 protein stabilization. The transforming region of large T protein has also been shown to bind another human cellular protein, p118 (Ewen et al., 1989), but the normal function of this protein and the importance of its interaction with large T protein to transformation is unknown. Models currently exist to explain the mechanisms by which large T protein is involved in establishment or immortalization. Mechanisms by which SV40-transformed cells are released from external growth controls are less clear, especially since SV40 does not encode a middle T protein, as does polyoma. In 1988 Khandjian and Gauchat observed that early in abortive infection of mouse cells, SV40 stimulated several rounds of DNA synthesis and mitosis under growth factor-limiting conditions, even though only one round of viral transcription had occurred. This suggests that early in infection SV40 proteins stimulate the growth of the host cell. 23 It is possible that continued growth occurs in those cells that can integrate and continuously express early viral proteins. 11, final! 3 antigen The role of SV40 small t antigen in transformation is poorly understood. In some cell types the expression of large T protein alone is sufficient to transform, although expression of small t protein often increases the efficiency of transformation. In 1989 de Ronde et al. observed that focus formation is induced by large T protein alone, but that foci of morphologically transformed cells grow on a monolayer only when both large T protein and small t protein are expressed in cells. One explanation for this came from the work of Bikel et al. (1986), who observed that small t protein expression promotes the dissolution of actin cables in the cell. The domain of small t protein shown to be important is a twice-repeated Cys-X-Cys-Cys-X—X-Cys motif; Mutations in this region abolish the ability of small t protein to alter actin cable structure and markedly reduce the efficiency of abortive transformation in Balb/c-3T3 mice infected with this mutant: however, the mechanism behind the correlation between changes in cell architecture>and transformation is not known. Another possible role for small t protein in transformation is in the activation of intracellular metabolic intermediates by phosphorylation. Pallas et al. (1990) purified two peptides of 36kDa and 63kDa that associate with polyoma middle T protein and small t protein, as well as with SV40 small t protein. They identified these as the catalytic and regulatory subunits of protein phosphatase 2A (PP2A), one of the enzymes thought to play an important role in the regulation of metabolism. 24 Thus, an association between small t protein and PP2A may be another mechanism by which a cell infected with SV40 loses normal growth control. ll r Papilloma 'viruses, the third major' group classified with the papovavirus family, is comprised of multiple serotypes. Some are associated with benign growths (warts) whereas others are associated with malignant neoplasia. Although the close epidemiological association pointed to a causative role for papilloma viruses in these growths, it was difficult to perform the necessary studies to test this because tight host specificity of these viruses prohibited in vitro propagation of viral cultures. The application of molecular techniques enabled investigators to determine the genomic elements responsible for regulating the viral life cycle, and those involved in transformation. As with the other DNA tunor viruses, the papilloma virus genome is organized into an early region, which encodes regulatory proteins, and a late region, which encodes virion structural proteins. Seven open reading frames or genes (El-£7) have been identified within the early region. Some of these genes, e.g., £6 and E7, overlap and are capable of expressing several products. Of these seven early genes, E5, £6 and E7 have been shown to be particularly important for malignant transformation. For illustrative purposes, studies relating to two human papilloma virus (HPV) serotypes, i.e., 16 and 18, which are closely associated with human malignancies (zur Hausen, 1987), will be discussed. Baker et al. (1987) showed that the primary viral transcripts in two cervical carcinoma cell lines are derived from the £6 and £7 genes. Subsequently, Bedell et al. (1987) constructed plasmids containing the E6-E7 open reading frame (ORF) from the early 25 region of HPV 18 and transfected these constructs into NIH-3T3 and Rat-1 cells along with a selectable marker. Selected cells that also expressed the E6 and E7 proteins formed colonies in soft agar, whereas those transfected with either a control plasmid or a plasmid containing a deletion in the 66-57 ORF did not. The cells which formed colonies in agar were also tumorigenic in athymic mice. This experiment demonstrated the importance £6 and/or £7 to transformation of cultured fibroblasts. Mfinger et al. (1989) obtained similar results in cultured human keratinocytes. Although the mechanism by which transformation occurs is not understood, Dyson et al. (1989) showed that the E7 protein from HPV 16 forms a complex with the RB protein p110-114. Since the E7 protein contains a domain with sequence similarity to SV40 large T protein and the EIA proteins of adenovirus (Phelps et al., 1988), both of which are known to complex with the RB protein, an E7-RB protein complex might also be involved in transforming cells. The E6 protein has been shown to bind the p53 protein (Herness et al., 1990), as is also the case for the early region products from SV40 and adenovirus. In fact, studies by Scheffner et al. (1992) show that E6 protein binding promotes the degradation of wild-type p53 protein, whereas some mutant forms of p53 protein do not bind the E6 protein and are not targeted for degradation. A third open reading frame, £5, encodes a small (44 amino acid) protein that is located in the plasma membrane. Although no definitive transforming function has been assigned to HPV E5, expression of the E5 product from the closely related bovine papilloma virus (BPV) has been shown to be required for transformation of mouse cells (DiMaio et al., 1986). Green and Lowenstein (1987) showed that the BPV E5 protein can 26 stimulate DNA synthesis in resting cells. Thus it appears that it is through coordinate expression of a number of viral proteins that papilloma viruses induce cellular transformation. 3. Herpes Viruses The group of Herpes viruses that can infect humans is comprised of Herpes simplex virus (HSV) 1, 2, and 6, cytomegalovirus (CMV), Varicella zoster virus (VZV), and Epstein-Barr virus (EBV). Each has a complex genome that can take several isomeric forms. Typically, infection with the virus results in lysis of the affected cell; however, a latent infection can be established in which no virion production is observed, but low levels of some viral protein products are detected. By mechanisms that are not well understood, the latent virus can be reactivated to a lytic cycle. During reactivation the affected individual has clinically apparent symptoms. Several factors, including the ongoing work to determine the role of other DNA viruses in cellular transformation, the knowledge that members of the herpes family are capable of establishing latent infection in mammalian cells, and the proposed link between Herpes simplex and cervical cancer (Manservigi et al., 1986), led researchers to look for transforming regions in the genome of herpes viruses. Studies were complicated by the lack of non-permissive cells in which infection would occur without lysis. Even in cells infected with defective (non- lytic) virus, little of the viral genome remained, and that which did showed only limited sequence similarity to other herpes viruses, or to known transforming regions of the DNA tumor viruses. Transformation associated with Herpes infection will be discussed using Epstein-Barr virus as an example. 27 Of all the Herpes family, the clearest epidemiological link between infection and transformation is demonstrated by the relationship observed between EBV infection and Burkitt’s lymphoma or nasopharyngeal carcinoma. EBV is also capable of itmortalizing B lymphocytes in culture and of transforming established rodent cell lines. Two viral products have been determined to be responsible for transformation. These are Epstein-Barr virus nuclear antigen 2 (EBNA-2) and latent infection membrane protein (LMP). Dambaugh et al. (1986) transfected into NIH-3T3 and Rat-1 cells a series of expression vectors, carrying both a selectable marker and various regions of the EBV genome. Rat-1 cells transfected with a vector containing EBNA-2 grew better in low serum medium than cells transfected with the selectable marker alone. Ability to grow in low serum is a property commonly found in malignant cells. To elucidate the mechanisniby which this growth advantage occurs, Hang et al. (1987) infected an EBV- negative Burkitt’s lymphoma cell line with retroviral constructs containing genes coding for each of the 4 EBNA proteins, the EBV leader protein, or the EBV LMP and observed changes in the expression of B cell surface markers and growth. They found that the EBNA-2-containing construct stimulated the expression of C023, a surface marker of B cell activation that is thought to act by transducing the B cell growth factor signal. Calender et al. (1987) found no induction of the expression of this marker, nor of immortalization, in a strain of EBV that does not produce a functional EBNA-2. Hang et al. (1985) demonstrated that expression of the LMP in established rodent cells in culture results in transformation. NIH-3T3 cells expressing the LMP were morphologically altered, and LMP-expressing Rat-1 cells formed foci, colonies in soft agar, and progressively growing tumors in athymic mice. 28 Interestingly, latent EBV infection is common in the United States, but there is no obvious association with an increased incidence of either Burkitt’s lymphoma or nasopharyngeal carcinoma. Given the multi-stepped nature of malignant transformation, it seems likely that other genetic and/or environmental factors play a role in this disparity in clinical outcome based on geography. The knowledge obtained about the biology of this class of viruses has provided valuable insights into the mechanisms by which they cause disease. In the process, data have accumulated that support the notion that malignant transformation is a multi-stepped process. Furthermore, increased understanding of the interaction between viral and cellular products has laid the groundwork for a nore general understanding of cellular transformation. 8. Retroviruses The retroviral life cycle consists of infection of the host cell, synthesis of a complementary double-stranded DNA provirus from viral genomic RNA using reverse transcriptase encoded by the virus, and integration of the provirus into the host genome, followed by transcription, packaging and release of viral progeny. Retroviruses are considered to transform human and animal cells (reviewed by Bishop, 1983), and this transformation is thought to be effected by the same mechanisms that occur in the viral life cycle. One mechanism involves transduction and transfer of a cellular gene, for example, src, ras, and sis, to a new cell in which it is inappropriately expressed and/or exists in mutated form. A second mechanism is insertion of the provirus near to a gene whose expression is normally tightly regulated. This results in 29 inappropriate expression of that gene (e.g., myc, and myb). A third mechanism involves insertion of the provirus in such a position that it results in production of a chimeric protein consisting of both proviral and endogenous elements (e.g., erbB). A fourth mechanism is inactivation of an endogenous gene by insertion of proviral DNA into the gene itself. Examples include Friend spleen focus-forming virus (FSSV) and p53. Still another mechanism consists of regulatory elements within the viral genome that act to alter the expression of host genes (e.g., trane-activation by tat of HTLV, HIV). Understanding these mechanisms will provide insight not only into the biology of this virus family, but also into parallel endogenous cellular processes causing malignant transformation. 1. Transduction of an Activated Gene Doolittle et al. (1983) first proposed that the simian sarcoma virus (SSV) transforming protein p28""-" is encoded by the transduced form of the human gene encoding platelet-derived growth factor 8 chain (PDGF-B) based on finding a high degree of similarity between the two proteins. This was an. important discovery in light of prevailing knowledge that certain viruses had transforming capacity. Infection by a virus known to carry a cellular gene encoding a protein growth factor for mesenchymal cells provided a plausible explanation for the induction of sarcomas. Subsequent studies by Fry et al. (1986) established that insertion of v- sis into normal diploid human fibroblasts results in transformation as measured by focus formation and anchorage independence. However, the transformants do not acquire increased life span, nor are they tumorigenic in athymic mice. Pech et al. (1989) was able to induce fibrosarcomas in newborn mice by infection with a c-sis-containing retroviral vector that 30 overexpressed PDGF-B. However, these f ibrosarcomas were polyclonal rather than monoclonal, which supports the hypothesis that overexpression of PDGF-B is only one of several changes that together result in malignant transformation. Several studies were performed by' Aaronson and co-workers to determine the sequences in v-sis that are critical for transformation. The central region of p28“‘“- shows the highest level of sequence similarity to PDGF-B. Giese et al. (1987) showed that eight cysteine residues exist within the central region of p28""4. Of these eight residues, four could not be altered without loss of transforming function. The Cys residues in the central region are responsible for correct formation of intra-chain disulfide bonds thought to be critical for PDGF receptor recognition. Other work on this central region (Giese et al., 1990) demonstrated that these Cys residues do not actually lie within the minimal transforming region of p28"m (i.e. , the receptor binding domain), leading to the hypothesis that correct tertiary structure is required for function of PDGF-B. The minimal transforming domain was also shown by LaRochelle et al. (1990) to contain recognition specificity for the beta- type PDGF receptor. The structure/function relationships established by these investigations led to hypotheses regarding mechanisms by which PDGF-B stimulates transformation. 0f specific interest was whether PDGF-B needs to be expressed at the cell surface to interact with its receptor. To answer this question, studies were performed to determine the correlation between post—translational processing/transport of PDGF-B and its transforming ability. Hannick and Donoghue (1986) constructed vectors containing the gene coding for the membrane anchor domain of a G-protein 31 fused to v-sis. ‘They found that formation of correct quaternary structure (i.e., dimerization and glycosylation) is associated‘with transport of the protein product to the cell surface and that transport to the cell surface correlates with transforming ability. They concluded that transport to, and activity at, the cell surface is important for transformation by the v-sis product; however, they could not extrapolate the results in their experimental system to endogenous PDGF. Although there was little doubt that interaction between PDGF-B and its receptor was required for mitogenesis and transformation, debate continued about the location of this interaction. Data from Huang et al. (1984), showing that antibodies to PDGF-B could only partially inhibit transformation by SSV, led to the hypothesis that there are both extra- and intra-cellular locations for PDGF-receptor interactions. In subsequent work, Huang and Huang (1988) also studied the kinetics of PDGF receptor (PDGFR) processing and turnover in NIH-3T3 and normal rat kidney (NRK) cells in the presence and absence of PDGF. Their data indicated that few newly synthesized PDGFR molecules are transported to the cell surface prior to degradation. Moreover, suramin, which inhibits both cell surface and cytoplasmic PDGF-receptor interaction, altered the PDGFR turnover, whereas protamine, which acts at the cell surface'only, did not. They concluded that sis transformation occurs by an intracellular autocrine loop. This conclusion was supported by Bejcek et al. (1989) who transfected NRK cells with a v—sis mutant targeted to other endoplasmic reticulum (ER) and golgi. The transfected cells formed foci, and cell fractions from ER and golgi expressed v-sis product that was mitogenically active, but no sis product was detected outside the cells. Although the data supporting an intracellular autocrine loop 32 mechanism of transformation by sis are strong, the fact that anti-PDGF antibodies can inhibit transformation partially suggests that an extracellular component also exists. Lokeshwar et al. (1990) compared the kinetics of PDGF-B turnover and mitogenic activity. They showed that the mature PDGF-8 species (a 44k0a protein) interacts with some molecule'other than PDGFR on the cell surface, since treatment with suramin, which inhibits the binding of growth factors to their receptors, released free p44, whereas none was released by treatment with a specific PDGF receptor- binding inhibitor, protamine. Definitive determination of the mechanism (i.e., internal vs. external autocrine loop) by'which sis activity results in nfitogenesis and transformation does not yet exist. However, these studies have provided insight into the mechanisms of transformation mediated by expression of protein growth factors, and into one of the means by which retroviruses are capable of inducing transformation. 2. Insertional Activation The viral oncogenes myc and myb exemplify transformation by retroviral insertional activation. The products of these cellular genes have been shown to be located in the nucleus (myc by Donner et al., 1982; Alitalo et al., 1983; myb by Klempnauer et al., 1984) and to bind to DNA. Therefore, they are thought to act as transcriptional activators. In sequence comparison studies, Ralston and Bishop (1983) found regions of sequence similarity between myc, myb, and EIA of adenovirus. Overall sequence identity is 15-21%, but alignment studies showed clustered regions of increased similarity. Hydrophobicity profiles are also alike, further supporting the hypothesis that these proteins have a similar function in the cell. One possibility is that they, like the EM 33 proteins, act by releasing the cell from limited life span. In 1977, Bister et al. showed that the transforming region of the acute avian leukemia virus MC29 encodes a gag fusion product later found to be gag-myc. A similar product was identified as part of the defective viral particle released by chicken myeloblast cells transformed without helper virus (Duesberg et al., 1980). This product did not contain myc, but rather the related myb gene fused with the viral gag and pol genes. In 1981, Hayward et al. hypothesized that the slow and the acutely transforming retroviruses transform cells using different mechanisms. Specifically, the slow acting viruses transform by integration into the host genome, causing inappropriate expression of endogenous host genes, whereas the acutely transforming viruses transform by inserting an exogenous transforming gene. These investigators based their hypothesis in part on the'observations that protein-encoding segments of viral genome are often absent in infected cells. They screened several lymphomas and one lymphoma-derived cell line for the presence of oncogene mRNA, and found a significant increase in the expression ofimyc mRNA. ‘This increase in expression was correlated with the expression of sequences corresponding to the five- and three-prime ends of the viral genome (i.e. the LTR) and not of the viral structural genes (gag, env, and pol). Also, restriction enzyme digestion and subsequent analysis of myc and its flanking sequences showed that the provirus had integrated near to the endogenous myc gene. Hayward et al. (1981) concluded that insertional activation of myc is a common mechanism by which the slow-acting viruses transform their host cell. Data by Gauwerky et al. (1988) illustrate the applicability of the retroviral insertional activation model to the understanding of the 34 process of malignant transformation mediated by endogenous myc. They identified rearrangements involving activation of the myc gene that were associated with progression of malignancy in patients with pre-B-cell leukemia. They did not report data on changes in myc expression in the cells with rearrangements. Future experiments linking rearrangements in .myc in both viral and endogenous systems with changes in expression will be important to gaining a clearer understanding of the importance of this mechanism in malignant transformation. 3. Alteration of a Gene Product ' The oncogene v-erbB, isolated from avian erythroblastosis virus (AEV) is structurally similar to the class of viral oncogenes that includes src, fes and abl. These genes encode protein tyrosine kinases that presumably regulate cell growth and differentiation. Because of the knowniassociation between viral oncogenes and their cellular counterparts, Downward et al. (1984) compared the sequence of purified epidermal growth factor receptor (EGFR), which is also known to have tyrosine kinase activity, with that of the product of the v-erbB gene. The EGFR has been divided into three domains, an extracellular domain that binds the ligand (EGF), a transmembrane domain, and an intracellular domain that has both tyrosine kinase activity' and tyrosine residues that are sites for phosphorylation. ‘These investigators found high sequence identity between the v-erbB protein and portions of EGFR; however a striking difference*was the lack of the extracellular ligand binding domain in the v-erbB protein. Other alterations in v-erbB include a deletion in the carboxyl terminus- encoding region, a small internal deletion in the intracytoplasmic domain- encoding region, and several point mutations. 35 The regions of v-erbB that are vital to its transforming ability were determined by Hells and Bishop (1988). They created chimeric genes by substituting regions of v-erbB with corresponding fragments from the EGFR gene, c-erbBI. They found that loss of sequences encoding the amino terminus (ligand binding domain) is necessary and sufficient to cause malignant transformation (tumorigenesis) in Rat-1 and NIH-3T3 cells. Neither the~carboxyl region-encoding deletions nor point mutations present in v-erbB were found to be required for transformation, although these do increase the relative transforming efficiency and alter the host range of the constructs. Sequence and gene expression data from several AEV strains revealed the presence of retroviral gag and env genes in close proximity to the transduced erbB gene. Bishop and coworkers (Bruskin et al., 1990) found that the product of the v-erbB gene is a fusion protein containing several residues from the nearby gag gene. Vectors constructed to determine the effect of those>gag residues on transformation showed that the presence of six gag amino acids resulted in a hundred-fold increase in transforming activity in NIH-3T3 cells. The increase in transformation was not accompanied by increased expression of the fusion product. They speculated that the increase in transforming efficiency is the result of alterations in the post-translational processing of the fusion protein. Similar studies by Maihle et al. (1988) had shown that insertional activation of c-erbB’by avian leukosis virus (ALV) results in various erbB transcripts, including those coding for gag-erb and gag-env-erb fusion products. Vectors mimicking these fusion products were processed differently by the host cell, resulting in proteins with unique glycosylation patterns and cellular localization. Interestingly, those 36 products that were encoded by env-minus constructs were less likely to be found at the cell surface. 'They speculated that such differences in post- translational processing and subsequent cell localization have an effect on the transforming ability of fusion proteins. Lee and Hayman (1990) also discovered the importance of the gag-erb fusion product in experiments designed to test the role of the extracellular domain of v- erbB in transformation of chicken fibroblasts and erythroblasts. A vector was constructed to use the env gene to localize the protein to the cell membrane. However, the only stable transfectants found contained a recombinant gag-env-erbB fusion product, leading to the speculation that one function of the extra-membrane and trans-membrane domains of c-erbB is the localization of the product to the appropriate cellular location. In the case of v-erbB this function would be assumed by the viral gag and env genes. Early studies involving the intracytoplasmic domain of v-erbB had shown that alterations in the tyrosine residues typically available for phosphorylation did not abolish transforming capability, but did change the transforming efficiency and/or altered the host cell specificity for transformation. Raines et al. (1988) identified three viral strains that differed in erbB structure, specifically in the arrangement of tyrosine residues. One strain was only capable of inducing erythroblastosis; another induced both sarcomas and hemangiomas: and the third induced erythroblastosis and sarcomas. The cause of strain-specific cell type transformation was not clearly defined: however, the authors hypothesized that the erbB products produced by the three strains interact with downstream effectors unique to each cell type(s). They reasoned that further experiments to determine the substrates of the erbB protein 37 tyrosine kinase might yield insights into the mechanism of transformation in different cell types. Aaronson and colleagues (Di Fiore et al. 1990) compared the transforming capacity of c-erbBI and c-ersz in NIH-3T3 cells and 320 cells, a mouse hematopoietic cell line lacking both c-erbBI and c- erbBZ, and typically dependent on IL-3 for proliferation. c-erb82 is an altered form of c-erbB that is associated with malignant transformation of human cells, and has been shown to be amplified in mamary carcinoma and salivary gland carcinoma. Two 320 strains were established after transfection. One expressed high levels of the c-erbBl protein, the other expressed high levels of the c-erb82 protein. However, the c-erbBl protein was only phosphorylated when the cells were grown in the presence of EGF, whereas c-erbBZ transformants had a high endogenous level of c- erb82 protein phosphorylation even without EGF. However, even when c- erb82 was highly expressed and phosphorylated, IL-3 was required for anchorage independence, suggesting that the c-erb82 product is not capable of efficient interaction with downstream regulators of anchorage independence. NIH-3T3 cells transfected only with c-erbBI could be efficiently induced to anchorage independence with EGF stimulation, indicating that the c-erbBI protein effectively interacts with other growth-regulating elements in these cells. Upregulation of the tyrosine kinase activity of the erbB protein has also been shown to correlate with increased transforming efficiency. Segatto et al. (1988) generated mutants of c-erbBZ that were upregulated by association with an LTR or the SV40 promoter, encoded a protein with amino terminal deletions, or else with a point mutation at codon 659 (the codon shown to be altered in the mouse equivalent of the c-erbB gene). In each case, in vitro tyrosine kinase activity of the altered c-erbB product 38 correlated with transforming efficiency in NIH-3T3 cells. In view of their data, the authors suggested using tyrosine kinase activity as a screening tool for erbB mutations. They also speculated that further studies on cells that have increased tyrosine kinase activity would yield insights into the substrates for the tyrosine kinase that serve as downstream regulatory points for growth and differentiation. 4. Insertional Inactivation Another mechanism by which retroviral oncogenes act is inactivation of an endogenous gene. An example of this is the inactivation of the p53 protein by gene rearrangement in cells transformed with the Friend spleen focus forming viruses (FSSV). Monroe et al. (1988) reported the presence of a truncated form of the p53 protein in several FSSV-induced erythroleukemic cell lines. This altered 46kDa protein is formed by an in-frame deletion of one of the highly conserved regions of p53, and was shown to be incapable of binding with SV40 large T antigen. Two other rearrangements involving p53 were reported by Ben David et al. (1988). One is an internal deletion with a frameshift and the other was insertion of the viral genome into the p53 locus. Because of the presumed tumor suppressor function of the p53 protein, loss of a functional product as a result of retroviral infection is expected to contribute directly to the malignant transformation of the infected cells. 5. ItanseActivation of Cellular Genes Most retroviruses have been shown to transform through ale-acting mechanisms. Hith the advent of research into the lymphotrophic viruses, i.e., human T cell lymphotrophic virus (HTLV) and human immunodeficiency 39 virus (HIV), arena-acting transforming elements have been discovered. Nerenberg et al. (1987) showed that the expression of the tat protein (known to be a mans-acting element in the HTLV I genome) is highly associated with malignant transformation. They created transgenic mice carrying the tat gene flanked by a viral LTR and poly A site. This construct had previously been shown to result in the expression of a correct product that induces the expression of a reporter gene. Hhen the tat gene was expressed in muscle, soft tissue sarcomas formed. Gene expression was also seen in the thymus; however, in these mice thymic atrophy and death occurred. Further studies by this group (Vogel et al., 1988) used the HIV tat gene introduced into transgenic mice. In this case expression of the HIV tat was seen only in the skin, and Kaposi sarcoma- like tumors formed in most of the male transgenic animals. Interestingly, no transgenic females were affected. However, when the cells from the Kaposi sarcoma-like tumors were examined, expression of the tat gene was not observed. One explanation for this phenomenon is that tat expression is required early in the process of transformation, but is not needed at later stages. Another possibility is that the tat product activates the expression of regulatory genes (e.g. growth factors) in the cells adjoining the transformed cell, and thus has an indirect transforming effect. Further studies will be needed to determine the importance of trans-activation in malignant transformation. C. Transfer-ation by Oncogenes: Genetic alterations play an important role in malignant transformation. 'These alterations fall into two major categories, activation of dominant-acting proto-oncogenes, and inactivation of tumor 40 suppressor function requiring the loss of both alleles encoding that suppressor; An overview'of current understanding of both these categories will be briefly presented with examples. 1. Doudnant-Acting Oncogenes Activation of proto-oncogenes can occur by point mutations that lead to altered protein function, or by changes that lead to unregulated protein expression. Approximately forty such genes have been found, each encoding a protein critical to the regulation of growth and differentiation. Oncogene products are found at different levels of cellular growth pathways, namely membrane-associated proteins (growth factors and their receptors, membrane-bound signal transduction proteins, and protein kinases), cytosolic proteins (protein kinases, modulatory proteins), and nuclear proteins (DNA-binding proteins, transcriptional activators). Much of what is known about dominant—acting oncogenes has resulted from investigations into the transforming retroviruses, as discussed in the previous section. Because the acutely transforming retroviruses carry genes derived from endogenous cellular genes that have been altered during the course of their incorporation into the viral genome, and/or as viral genes evolved, they provide clues to how their endogenous counterparts act. n i r n - i r t in The prototype for understanding the relationship of viral-encoded transforming factors and their cellular counterparts is the src gene. ‘The virus that transforms chicken fibroblasts contains the src gene, first 41 described by Rous in 1911 (reviewed in Seemayer and Cavenee, 1989). The src gene encodes a 60kDa protein that associates with the cell membrane, has tyrosine kinase activity, and can be regulated by phosphorylation of tyrosine residues within its sequence. The v-src gene has been shown to have an endogenous cellular counterpart in several species. In 1983 Takeya and Hanufsa reported the sequence of the v-src honologue in chickens, and compared the sequence and structure to that of v-src. They found that the amino acid sequences vary at only 28 positions. The structure and sequence of c-src has been shown to be highly conserved among diverse species, suggesting that src plays a critical role in the normal function of the cell. In a comparison between human and chicken c- src, Anderson et al. (1985) showed a DNA sequence identity of approximately 90% and a amino acid sequence identity of approximately 98%. The mechanism by which proto-src becomes activated has been the subject of much ongoing research. The wild-type protein is known to be membrane bound, and to have low-level tyrosine kinase activity. Iba et al. (1985) compared several properties of v-src and c-src and found no difference in protein expression level or subcellular location. They did show that pp60*‘“‘has a much higher tyrosine kinase activity, and that the level of kinase activity correlates well with transforming capacity. In similar studies, Johnson et al. (1985) constructed plasmids containing v- src and c-src under the control of strong promoters to confirm that over- expression of the pp60°'m is insufficient to transform cells. They isolated NIH-3T3 c-src and v-src transfectants either by focus formation or by selection for a co-transfected selectable marker. In a few cases focus-forming, non-tumorigenic c-src transformants were found, and all of these expressed pp60°”-“ at five times the level of pp60"m. Even at these 42 high levels of expression, the kinase activity in the c-src transfectants did not reach that of the v-src transfectants. This supported the hypothesis that increased pp6023 tyrosine kinase activity is required for transformation. Coussens et al. (1985) showed that pp60°"“.and pp60"m could be distinguished not only by tyrosine kinase activity, but also by the pattern of proteins phosphorylated in vivo, suggesting that transformation by v-src involves subverting normal metabolic pathways to cause abnormal growth. Clearer understanding of transformation by activation of src awaits data regarding normal substrates for this tyrosine kinase, and their function in regulating cell growth. n i r i Oncogene products found in the nucleus and thought to act by regulating gene expression are exemplified by c-jun. This oncogene was first identified as the transforming agent in an avian sarcoma virus (ASV- 17). Cellular homologues have been identified in chicken, mouse~and human cells. Researchers found significant (44%) sequence identity between v- jun product and the yeast transcription activator GCN4. This transcription activator is similar to the cellular protein AP-l in amino acid sequence and DNA binding specificity (reviewed by Curran and Franza, 1988). Bohmann et al. (1987) performed several experiments that provided evidence that the c-jun protein and AP-l are related. First, antibodies raised against the v-jun protein cross-reacted with AP-l. Second, the v- jun and c-jun proteins had an overall sequence identity of 80%, with greatest similarity (only 2 amino acid changes) in the carboxyl terminal domain known to be responsible for DNA binding. Sequence similarity also held across several distantly-related species. Third, trypsin digest 43 fragments of AP-1 were easily matched with the putative sequence of the c- jun protein. Fourth, DNA binding footprints were essentially identical for purified AP-I and for the protein product encoded by a vector encoding the carboxyl terminal 133 codons of c-jun. It was noted; however, that even purified AP-l had several different MH species, leaving open the possibility that there is a family of DNA binding proteins with similar binding specificity. Some support for this hypothesis was obtained by finding more than one band when probing genomic DNA with v-jun under low- stringency conditions. Rauscher et al. (1988) identified a nuclear protein p39 as being AP-1, and also as the product of the c-jun gene. This protein is one of several that associate with another nuclear oncoprotein p62°fl“. AP—l and p39 extracts were shown to co-migrate, as does p39 that has been immunoprecipitated by antibodies to p62°'-'-”. Trypsin digestion fragments of p39 are identical to the protein produced by a vector encoding the c-jun protein. Sequential immunoprecipitation experiments also showed protein-protein interaction between the products of fos and jun. Lamph et al. (1988) explored the functional role of c-jun in mouse NIH-3T3 cells. They investigated the effect on the expression of mouse c- jun of treatment with serum and TPA, both of which are known to stimulate cell growth and to induce expression of fos. Expression of jun product was induced by both serum and TPA, and the time course of induction was very similar to that of induction of fos. The product encoded by their construct was also capable of mans-activation of a reporter gene, chloramphenicol acetyl transferase (CAT). Schfitte et al. (1989) tested the ability of wild-type, but deregulated, human c-jun to transform rat cells. Vectors were constructed that contained genomic c-jun and c-jun 44 cDNA in both sense and antisense orientation. These*were transfected into primary rat embryo cells and the immortal rat fibroblast cell line Rat 1. The Rat 1. line was transformed to anchorage independence and tumorigenicity when c-jun mRNA was highly expressed, i.e. at levels 10-15 times more that of endogenous jun mRNA. Primary rat cells could not be transformed by transfection of jun alone or by transfection of c-jun in combination with c-myc. However, when jun was co-transfected with activated H-ras gene, the transfectants showed increased focus formation. These focus—forming cells were capable of producing tumors in athymic mice. Transforming efficiency was augmented further by treatment of transfected rat embryo cells with the phorbol ester TPA. These studies highlight the complexity of pathways controlling cell growth and differentiation, and establish the importance of activation of dominant-acting proto-oncogenes in transformation of mamalian cells. Transformation induced by the activation of another family of genes, the ras genes will be discussed in subsequent sections of this review. 2. Tuner Suppressor Genes: The other general genetic mechanism contributing to the malignant transformation of cells is the inactivation of tumor suppressor genes. This class of genes, which includes the retinoblastoma susceptibility gene (R8), the p53 gene, the Hilms’ tumor sensitivity gene, and the gene associated with the development of colorectal carcinoma (MDMZ), has been studied extensively. However, as they are not directly related to the research topic of this thesis, consideration of their pr0perties will be limited to a brief discussion of the R8 and p53 genes (reviewed by Levine and Momand, 1990). 45 i l i l n Loss of both alleles of the R8 gene was first postulated by Knudson in 1971 after an examination of epidemiological patterns of retinoblastoma occurrence. Retinoblastoma is found in two distinct patterns, the first being the early appearance (i.e., average age 14 months) of bilateral retinal tumors in patients with a family history of retinoblastoma, the second being the later appearance (i.e., average age 30 months) of unilateral tumors in patients with no family history of this disease. Knudson hypothesized that in the first case, patients inherit a germ-line mutation that increases their risk of acquiring a second change in one of the altered retinal cells. This "two hit" hypothesis was substantiated when the Rb gene was localized to chromosome 13 (Yunis and Ramsay, 1978) and cloned (Lee et al., 1987). Restriction fragment length polymorphism (RFLP) analysis of DNA from retinoblastoma cells showed the loss of both alleles of the R8 gene (Cavenee et al., 1983; Friend et al., 1986). Furthermore, in vitro data showed that returning one wild-type allele to retinoblastoma cells reverses the malignant phenotype (Huang et al., 1988). This suggested that wild-type R8 is somehow involved in the regulation of cell growth. Further studies have shown interactions between the product of the RB gene and several of the transforming proteins of the DNA tumor viruses (thte et al., 1988; Larose et al., 1991 DeCaprio et al., 1988: Phelps et al., 1988), implicating the inactivation of the tumor suppressor function as the mechanism of transformation by these viruses. 46 mm The p53 gene has been localized to the short arm of chromosome 17 (Benchimol et al., 1985; Isobe et al., 1986). Loss of activity of its product has been implicated in a ‘variety' of tumors. However, an understanding of the role of p53 in malignant transformation has been complicated. Early experiments by Rotter (1983) showing the overexpression of the p53 protein in several transformed mouse cell lines, as well as data showing that mutated p53 can cooperate with other oncogenes in transformation (Hinds et al., 1989), led to speculation that p53 acts similarly to the dominant oncogenes. In contrast, studies exemplified by Finlay et al. (1989), have shown that expression of wild- type p53 protein suppresses transformation induced by ras and adenovirus EIA or by ras and mutant p53 in rat embryo fibroblasts. Additional investigation into the properties of mutant p53 protein have shown that it has an altered conformation (Finlay et al., 1988), altered oligomer formation (Kraiss et al., 1988; Milner et al., 1991; Stenger et ai., 1992), altered protein-protein interaction (e.g., binding the heat shock protein encoded by hsc70 as shown by Finlay et al., 1988), and an altered half life (Finlay et al., 1988). These data support the hypothesis that mutant p53 protein subunits, when interacting with wild-type p53 protein subunits, inactivate the function of the wild-type protein in the transformed cell (Eliyahu et al., 1988). This hypothesis is consistent with a model in which the normal tumor-suppressing function of p53 protein is lost during malignant transformation. Further support that loss of normal p53 protein function is required for the transformation of some cells comes from data showing that the transforming proteins of the DNA tumor viruses interact with p53 (Linzer and Levine, 1979; Sarnow et al., 47 1982; Herness et al., 1990), and that the Friend leukemia virus can inactivate p53 (Ben David at al., 1988). Future studies of this class of genes will undoubtedly lead to significant insights into the normal regulation of cell growth and differentiation, as well as the derangement of these processes in the course of malignant transformation. 111- WWW Early contributions to understanding malignant transformation were based on genetic models supported by experiments showing that transformation is induced by the acutely transforming RNA tumor viruses. The transforming agents in these viruses were shown to be transduced cellular genes, as discussed in earlier sections of this review. As evidence accumulated supporting the multi-stepped theory of transformation, it became clear that not all of the changes occurring in malignant transformation were necessarily genomic alterations. Specifically, tumor promoters seemed to work by a different mechanism. As knowledge of malignant transformation has developed, theories have been advanced to explain in vivo and in vitro data that call into question the universality of mutation-induced carcinogenesis. A. Developlental and Environ-ental Theories In 1985, Rubin criticized theories of carcinogenesis that included only genetic events. In his view, even the inclusion of epigenetic phenomena such as DNA methylation was insufficient to account for the complexity of the transformation process. He maintained that consideration of developmental issues and cellular organization is 48 required in order to account for several experimental findings in the field of carcinogenesis. First, he cited data showing that transplanted stem cells either differentiate or become transformed depending on the area to which they were transplanted, as well as studies describing tumor heterogeneity. He claimed that these data could not be explained by a mutation-based model, but rather that they supported the view that the phenotype of cells depends upon the surrounding environment, sometimes referred to as "field effects". Second, he discussed studies showing the progressive nature of transformation. Some early-passage cells taken from tumors have a non-metastatic, slow-growing phenotype and near-diploid genotype, but with time and serial passage in culture, cells become metastatic, aggressively-growing, and karyotypically more abnormal. He interpreted this as evidence that the tumor cells respond to a particular environment, either in culture or in a test animal. Finally, he cited experiments showing that cells are differentially susceptible to chemically induced transformation depending on their surrounding environment, (e.g. within an organ system or in monolayer culture), and sparse or dense distribution in a culture dish. His explanation for these data was that the surrounding environment in which a cell exists is critical to its behavior, and in particular its transformability. Although it is important to evaluate results carefully in the context of an experimental system, the findings Rubin cited to refute the importance of genetic phenomena to transformation can, in fact, be explained by genetic mechanisms. In each case, there are known genetic mechanisms that would allow cells in the surrounding environment to modulate gene expression within the target cell. Such pathways include growth factors and their receptors, second-messenger systems, basement 49 membrane-protease interactions, alteration of gene expression by the action of nuclear transcription factors, and activity of tumor suppression factors. Selective conditions would allow for expansion of a specific cell type and/or genetic change. Similarly, altered regulation followed by subsequent rounds of selection would result in progressive changes in cell phenotype. B. Methylation One area of epigenetic control of cell function thought to play a role in transformation is DNA methylation-mediated alteration of gene expression. In a review of this subject, Holliday (1987) first made a case for the heritability of methylation patterns, as heritability is an important consideration in multi-step carcinogenesis. Primary evidence included data showing that the X chromosome that is inactivated in each generation can be traced stably to either maternal or paternal origin. This process is called chromosomal imprinting. Second, there is an increased incidence of the development of tunors in the offspring of animals exposed to carcinogenic agents. The rate of mutation required to account for this increase is unreasonably high. Another possible explanation is a heritable alteration in methylation patterns which would predispose to tunor formation. One example that suggests a role for methylation in transformation is the interspecies difference in the rate of carcinogenesis. Rates of cancer development in rodents are much greater than that of humans, even though their cells can be shown to have similar mutation rates. Holliday argues that differential maintenance of methylation explains the difference. Although our knowledge of the mechanisms by which methylation is maintained and/or altered has increased 50 markedly since 1987, specific mechanisms by which methylation regulates gene expression still remain to be elucidated. Further studies are needed to understand how normal cellular processes are controlled by DNA methylation, and the role that methylation plays in the malignant transformation of cells. C. Protein Kinase C Alteration in second—messenger systems as a result of protein kinase C (PKC) activity is another example of an epigenetic phenomenon that can play a role in the process of transformation (c.f., review by Heinstein, 1991). Heinstein emphasizes the importance of epigenetic changes in malignant transformation, given that-non-genotoxic chemicals are known to be involved in the induction of transformation, and that the behavior of transformed cells can be modulated by non-genotoxic agents (e.g. hormones). He considers that since normal growth and differentiation of cells is presumably regulated by both genetic and epigenetic factors, one can postulate that alterations in both types of factors are required for malignant transformation. Data from experiments performed in his laboratory showed that a high expression level of PKC in rat fibroblasts is associated with an increased susceptibility to transformation by the T24 H-ras oncogene (Hsiao et al., 1989), and that overexpression of PKC in cultured rat liver epithelial cells results in altered expression of the myc protein, phorbin, and ornithine decarboxylase (Hsieh et al., 1989). He and co-workers also performed studies implicating alterations of PKC activity in the transformation of colon epithelium (Guillem et al., 1987). The concept of malignant transformation as a multi-stepped process resulting from alterations in the complex pathways controlling normal 51 cellular growth and development has been well documented. One general approach to gaining insights into these processes is to consider the cellular processes which are altered in the process of transformation. Attempts to define the likely mechanisms for these alterations should involve investigations of both genetic and epigenetic phenomena. The explosion of data in the field of oncogenes has certainly contributed to the understanding of transformation. .Similar investigations into the role of epigenetic phenomena in malignant transformation might also yield significant insights. IV. A. General Properties of the ras Fanily I. ras Falily The ras proto-oncogenes are a family of genes within the superfamily of genes encoding guanine nucleotide-binding proteins. These genes encode proteins involved in many aspects of cell growth and regulation (reviewed by Valencia et al., 1991). H-, K- and N-ras are the most widely studied in this family because they have been implicated in the process of malignant transformation since the discovery of the Harvey rat sarcoma virus and Kirsten sarcoma virus. However, the identity and functions of the protein products of other members in this gene family have recently become clearer. For example, sharing approximately 30% sequence identity to the ras genes are the genes of the Rho family and the Rab family. The former have been implicated in maintenance of the cytoskeleton, and the latter appear to be involved in the regulation of secretory functions 52 (reviewed by Hall, 1990). 2. Structural Properties of ras LLMA The ras genes were discovered when the transforming agent in the Harvey Murine sarcoma virus was identified (Harvey, 1964). A similar gene was identified in the Kirsten Murine sarcoma virus (Kirsten and Mayer, 1967). Later these genes were found to have human cellular homologues (Parada et al., 1982; Ellis et al., 1981; McGrath et al., 1983). N-ras has been identified within a human neural cell tumor (Shimizu et al., 1983) and human sarcomas (Hall et al., 1983), and has not been shown to have a viral counterpart. H-ras is located on chromosome 11 ( McBride et al., 1982; Popescu et al., 1985), K-ras on chromosome 12 (Popescu et al., 1985), and N-ras on chromosome 1 (Popescu et al., 1985). In addition, in human cells there are processed pseudogenes for H- and K-ras that are not expressed and are thought to have arisen by insertion of a processed mRNA into the germline (McGrath et al., 1983; Miyoshi et al., 1984). The general structure of the three major ras genes has been determined. In addition, the sequence of the coding regions of K- and N- ras genes and of the entire H-ras gene is known. In 1983 the genomic organization and sequence of the H-ras gene was reported from three laboratories (Reddy, Capon et al., and Fasano et al.). The gene is contained within a 6.6kb BamHI restriction fragment and expresses a transcript of approximately 1kb containing four translated exons and untranslated flanking regions. The protein product has 189 amino acids and a molecular weight of 21kDa. McCoy and Heinberg (1986) found that K-ras 53 has a different organization, in that it occupies over 50kbp and has a total of five coding regions with two forms of exon 4, designated a and b. Because of alternate splicing, two different transcripts are present in the cell. However, over ninety percent of the transcripts contain only exon 4b. The remainder contain both 4a and 4b, but because a stop codon at the 5’ end of 4b is recognized, only 4a is translated. Translation of exon 4b results in a 188 amino acid protein, whereas 189 amino acids are encoded by the transcript containing exon 4a. In either case, the protein has a molecular weight of 21kDa. McGrath et al. (1983) determined the sequences of the coding region of the K-ras proto-oncogene (K-ras 2) and of the entire K-ras pseudogene (K-ras 1). The N-ras gene from the neuroblastoma cell line SK—N-SH was cloned by Taparowsky et al. (1983). They identified four coding exons within approximately 15kbp of DNA, encoding a protein of 189 amino acids. Hall and Brown (1985) reported similar studies on cDNA clones derived from transcripts expressed by the human fibrosarcoma cell line HT1080. ‘They found that HT1080 cells express two different transcripts, one 4.3kb in length, containing seven exons, and another 2kb in length, also containing seven exons. In both cases only four exons encode protein, resulting in a 189 amino acid product. Pr in Overall, the amino acid sequence identity among the three human ras gene products is greater than 90% (Valencia et al., 1991). The greatest divergence occurs in the carboxyl terminal domain encoded by exon 4. Furthermore, there is significant sequence similarity' across widely divergent species. The implication of this level of sequence identity is that the amino terminal portion of the protein is responsible for some 54 function that is critical in normal cells. The p21 protein has been divided into several functional domains each defined by structural elements. Five regions in particular have been shown to be important for the binding of the guanine nucleotides and for interaction*with the GTPase activating protein GAP (purified by'Trahey'and McCormick 1987), as well as downstream effector molecules (reviewed by Bourne et al., 1991). Determination of the structure of the ras proteins has resulted in a better understanding? of their function and their role in malignant transformation. Hierenga and Hal (1983) compared the structure of wild- type H-ras p21 and the activated form encoded by the EJ and T24 H-ras genes. They showed that any amino acid other than glycine at position 12 results in changes in guanine nucleotide binding. In similar studies, Pai et al. (1989) determined the structure of H-ras p21 and deduced the changes induced by mutations at the codons known to activate oncogenic activity. In every case these mutations are predicted to affect the guanine nucleotide binding or the GTPase activity (both intrinsic and GAP- activated), lending support to the hypothesis that appropriate regulation of ras p21 is dependent on wild-type structure. Similarly, Krengel et al. (1990) showed that activating mutations impair normal guanine nucleotide binding and GTP hydrolysis. Further evidence of the importance of p21 structure came from studies by Alonso et al. (1990), who found that NIH- 3T3 cells transformed by mutant ras have impaired coupling to phospholipase C in response to PDGF stimulus, whereas they show enhanced coupling in response to bradykinin. Cellular localization is also important to the function of ras p21. In 1980, Hillingham et al. showed by immunocytochemistry that the H-ras product is associated with the internal surface of the cell membrane. 55 This association occurs as a result of post-translational modification (reviewed by Grand and Owen, 1991), involving a carboxyl terminal tetrapeptide, CAAX (Cys, aliphatic amino acid, aliphatic amino acid, any amino acid). Subsequent work by Der and colleagues (Jackson et al., 1990) showed that post-translational modification is essential for transformation by K-ras. 3. Role of ras in Nor-a1 Cellular Function ras gene expression is found in most cell types. In a study of ras p21 expression in human cells, Furth et al. (1987) detected the presence of p21 in almost all tissues. Furthermore, they found increased levels of expression in actively dividing cells, for example, squamous epithelium near the basal layer and colon crypt cells. Increased expression was also found in cell types with specialized membrane function, for example, neural cells and pancreatic secretory cells. Increased H-ras expression also occurs in liver cells after partial hepatectomy and during carcinogen-induced regeneration (Goyette et al., 1983). These data support the hypothesis that ras expression plays an important role in regulation of cell division. It has been unclear whether there is any coordination of the expression of the various ras genes. Support for this notion comes from data by Nieto et al. (1987) in which H-ras p21 expression was shown to be increased in rat mammary gland during gestation‘whereas the level of K-ras expression remained constant. In contrast, Miller et al. (1983) found that although some proto-oncogenesiare specifically expressed at different stages of mouse development, H— and K-ras are both stably expressed throughout embryonic development and post-natally. 56 The role of ras in normal growth and development remains unclear; however, in individual cells there is support that ras expression is required for cell division. Mulcahy et al. (1985) microinjected anti ras p21 antibody into cells that had been transformed by microinjection of large amounts of p21. The cells that received the antibody could not enter the S phase of the cell cycle, indicating that ras expression is required for cells to cycle. 8. Evidence for the Role of ras in Malignant Transfer-ation 1. Evidence frol Correlative Data fl.) flunnn tuner gel IS The study of human tumor cells has provided strong evidence that ras activation plays a role in several tumor types. In some cases the importance of ras in malignant transformation can be inferred from the high incidence of ras activation in cells from a common origin. For example, Forrester et al. (1987) reported a 39% incidence of K-ras mutations in human colon tumor cells by the RNAse mismatch method, and Smit et al. (1988) found a 93% incidence of ras activation in pancreatic adenocarcinoma. On the other hand, in other tissues there is a weaker association between ras activation and malignancy. For example, Neri et al. (1988) reported N-ras activation in only 18% of a series of acute lymphoblastic leukemias and no ras activation in non-Hodgkin lymphoma. Although associative data implies, but does not prove, causality, one study provides stronger evidence that ras activation plays a causative role in transformation. Senn et al. (1988) found N-ras mutations in 57 malignant cells from five of eighteen patients with acute non-lymphocytic leukemia. These patients were treated, and some achieved remission. Hhen the patients were re-assessed, the cells of the patients who had achieved remission had only wild-type N-ras. In contrast, the cells of one patient who did not achieve remission continued to show the same N-ras mutation that was detected prior to treatment. Taken together these data provide strong evidence for the role of ras activation in the genesis of certain tumors. h.1_An1mal_tqu:_cclls Experiments in carcinogen-treated animals also provide strong supporting evidence that ras activation can play a causative role in tumor development. One such study reported by You et al. (1989) compared K-ras mutations in spontaneous and chemically induced lung tumors from the strain A mouse. Mice of this strain have increased susceptibility to the formation of lung tumors. In cells from spontaneous and carcinogen- induced tumors, the incidence of ras activation was high (83-100%); however, there was a difference in the location of the activating mutation depending on whether the tumor was spontaneous or carcinogen-induced. Specifically, cells from spontaneous tumors showed a mix of mutations at codon 12 and codon 61 whereas cells from all methylnitrosourea-induced tumors showed mutations at codon 12, and 90% of cells from ethyl carbamate-induced tumors were mutated at codon 61. These data indicated the importance of ras activation in animal tumors, linked tumor induction by carcinogen treatment with K-ras activation, and added to existing data on H-ras mutation in DMBA-induced mouse mammary carcinogenesis (Dandekar et al., 1986) and polycyclic aromatic hydrocarbon-induced skin tumors in 58 mouse (Bizub et al., 1986). Experiments using carcinogen-treated animal models have also provided some information regarding the timing of ras activation in the process of malignant transformation. In 1988 Pelling et al. studied ras mutations in mouse skin tumors that arose in animals treated only with a tumor promoter, in contrast to other studies in which animals had been exposed to DMBA, a mutagen, followed by treatment with a promoting agent. They found an H-ras mutation in more than 70% of benign and malignant tumors. Similar data were reported by Barbacid and colleagues (Kumar et al., 1990) in a mouse mammary tumor model. Mutations in H- and K-ras were detected as early as two weeks after initiating carcinogen treatment and several months before neoplastic cells were observed. A reasonable conclusion from these studies is that ras activation is an early change in malignant transformation and that cells containing this change can be targeted by other transforming agents such as tumor promoters. l u u Another line of correlative evidence for involvement of ras activation in malignant transformation has come from carcinogen treatment of both human and animal cells in culture. One such study by Miyamoto et al. (1990) showed that methyl nitrosourea-treated mouse mammary epithelial cells contained specific mutations in the K-ras gene. Furthermore, mutations could be detected in partially transformed, as well as in malignant cells, again supporting the notion that ras activation is an early event in the neoplastic process. Heissman et al. (1989) performed similar experiments in a mouse keratinocyte cell line. This cell line is epidermal growth factor(EGF)-dependent.and can be induced to differentiate 59 terminally by the addition of calcium to the growth medium. These characteristics. allowed the investigators to select EGF-independent, calcium—resistant transformants after treatment. In each case, the cells that were selected by their ability to differentiate following EGF stimulation contained ras mutations, implicating ras mutations iniabnormal growth and differentiation. 2. Evidence frol Insertion of Activated ras Genes More direct evidence that ras plays a role in the malignant transformation comes from experiments in which transformation is brought about by insertion of activated ras genes. One example is from work reported by 805 and colleagues (Vousden et al., 1986), in which plasmids containing wild-type H-ras were treated with mutagens. The resulting activation of the ras gene was evidenced by focus formation in NIH-3T3 cells transfected with the treated plasmids. Hhen the activated DNA was analyzed, point mutations were found at codon 12 and codon 61. The nature of the mutations was consistent with that typically found in DNA treated with the specific mutagenic agents used. This study serves as a bridge between correlative studies and transformation by the insertion of activated ras into cells. Other in vitro transformation studies strengthen support for the role of ras activation in multi-step transformation. For example, Harper et al. (1986) were able to complete the transformation of mouse papillomas by insertion of theiactivated H-ras gene isolated from the human EJ bladder carcinoma cell line. Similar studies on human fibroblasts in our laboratory have been discussed in earlier sections of this review (Hurlin et al., 1989). In this case, an immortal human cell line that expresses a transfected v-myc gene (Morgan 60 et al., 1991) was transformed to tumorigenicity after transfection of the T24 (EJ bladder carcinoma) H-ras oncogene, whereas a finite life span cell line, although morphologically transformed, eventually senesced, and did not form tumors (Hurlin et al.,1987). Others working with human fibroblasts have reported similar results using transfection with N-ras (Kinsella et al., 1990), suggesting that c-myc-induced immortality is a prerequisite to obtaining full transformation. Evidence of the function of activated ras genes in bringing about the tumorigenic state was obtained by Kasid et al. (1985), who transfected a human breast cancer cell line with the viral H-ras gene. Prior to transfection, the cells, though fully transformed, were dependent on estrogen in order to form tumors. The ras-expressing transfectants no longer required estrogen to grow rapidly in culture, were not significantly inhibited by anti-estrogen compounds, and no longer required estrogen supplementation to form tumors in test animals. This suggests that the presence of activated ras bypasses the normal estrogen-regulated growth control pathways in breast cells. 3. Evidence fron Reversion of ras-Induced Transfer-ation A third avenue by which the importance of ras activation to malignant transformation and the mechanisms involved have been determined is from observations of ras-transformed cells that have reverted to a less-transformed phenotype. Exemplifying this approach is a study by Schaefer et al. (1988) in which genomic DNA from human placenta was transfected into ras-transfected tumorigenic cells from rats. Some~of the transfectants displayed partial reversion as evidenced by decreased growth in soft agar, morphological reversion, and decreased tumorigenic a /_;. 61 potential. The reversion did not occur by a decrease in the expression of the activated ras gene. Although the investigators were able to identify the fragment of DNA that conferred the reverted phenotype, more specific identification was not reported. A repressor of K-ras-induced transformation was identified through a series of experiments involving NIH-3T3 cells transformed by v-K-ras.- Noda et a1. (1989) found that reversion to a non-transformed state was mediated by plasmid DNA introduced into the cells. The gene responsible (termed Krev-l) was shown by Kitayama et al. (1989) to encode a 184 amino acid protein with 50% sequence identity to K-ras, as well as similar guanine nucleotide binding and GTPase activity. By performing site specific mutagenesis studies, Zhang et al. (1990) localized the sequences required for reversion to the effector domain (residues 32-40), leading to the speculation that Krev-l acts by competing with K-ras for downstream targets. Another example of reversion was reported by Shindo-Okada et al. (1989). Their studies focused on identifying chemotherapeutic agents by their ability to cause reversion of transformation in H-ras-transformed NIH-3T3 cells. One compound, the antibiotic azatyrosine, caused a high frequency of reversion in both ras- and raf- transformed NIH-3T3 cells without causing significant growth inhibition of normal NIH-3T3 cells. The~antibiotic did not interfere with the expression of the transfected activated ras gene. Furthermore, if the agent was removed, reversion was stable for at least three months. The mechanism by which this antibiotic acts was not determined. Because of the high frequency of reversion the authors concluded that genetic mutation was unlikely. Since reversion was attained in ras- and ref-transformed cells one can speculate that azatyrosine acts within the signal transduction pathway that involves the 62 normal counterparts of the ras and raf genes. C. Hechanisl of ras-Mediated Transfer-ation 1. Point Mutation-Induced Changes in ras p21 Activity Insights into the mechanisms by which activated ras genes function in the process of malignant transformation can be gained by investigating the biochemical activities of ras gene products in their wild-type form, their activated form, or those that have been altered in such a way that specific functions are lost. The conserved functions of these proteins are guanine nucleotide binding, GTPase activity, effector interaction, and membrane localization. ‘Therefore, it is reasonable to hypothesize that an alteration in DNA that causes the protein to lose one or more of these functions would lead to activation or inactivation of the ability of ras to cause transformation. Clanton et al. (1986) investigated the guanine nucleotide binding capacity and transforming activity of ras gene products. They created several vectors, each containing a specific mutation within the v-H-ras gene that was predicted to alter guanine nucleotide binding. Only two of the mutants, one that substituted lysine for asparagine at codon 116, and another that substituted tyrosine for asparagine at codon 116, showed loss of guanine nucleotide binding. These two mutants were also found to be deficient in autokinase activity (a property of the activated viral ras oncogenes), and were unable to transform NIH-3T3 cells to focus formation. The investigators concluded that guanine nucleotide binding is required for transforming function of ras-encoded products. In contrast, Der et al. (1986), created H-ras mutants at codons 61, 116, 117, or 119 that showed decreased guanine 63 nucleotide binding, but no change in transforming capacity. However, the mutants they tested contained different amino acid substitutions than were in Clanton’s study, and were not deficient in autokinase activity. It is reasonable to conclude from these two studies that specific amino acid changes interfere differently with protein function, and that absolute loss of guanine nucleotide binding is required in order to lose transforming activity. It is established that ras p21 cycles between its GTP-bound ("on state") and GDP-bound ("off state"), and that the steady state ratio of GTP- to GDP-bound p21 is greater in actively dividing cells than in quiescent cells (Haubruck and McCormick, 1991). Mutations in ras that result in decreased p21 GTPase'activity (and therefore increased GTP-bound p21) might be expected to cause transformation. The earliest reported example of an association between GTPase activity and transformation is by Sweet et al. (1984), who found that the product of the T24 ras gene has only one-tenth the GTPase activity of wild-type ras p21. Manne et al. (1985) compared the GTPase activity of wild-type and activated ras p21 mutants using a highly sensitive assay that employed bacterially expressed, purified p21. All of the activated products showed decreased GTPase activity. However, data by Der et al. (1986) showed that although alteration in normal GTPase activity may be necessary for transformation, it is not sufficient, since they could not establish a direct correlation between GTPase activity and transforming capacity. Clearer understanding of the mechanism by which GTPase activity is decreased in activated ras p21 came from work by McCormick and colleagues (Adari et al., 1988) who showed that the'GAP protein (identified by Trahey and McCormick in 1987) interacts with p21 at its effector binding domain, 64 and that ras genes containing activating mutations encode products that are resistant to the GAP-mediated increase in GTPase activity. However, Srivastava et al. (1989) showed that certain deletion mutants of ras that abolish interaction with the antibody ‘Y13-259, also show decreased interaction with GAP and have diminished transforming capacity. This result runs counter to expectations if GAP is the only protein which interacts with the ras p21 effector binding domain, since one would expect loss of GAP interaction to lead to a decrease in GTPase activity and augment transforming capacity. Evidence of other effector interactions comes from data showing increased inositol phospholipid breakdown in cells containing activated ras (Hancock et al., 1988). Recent data have also shown that ras p21 interacts with other proteins at the effector binding domain. The identity of these so-called "downstream" molecules and their role in transformation have only recently been reported and will be discussed in the appendix section of this dissertation. 2. Cooperation between ras and Other Oncogenes Because malignant transformation is a process requiring many steps, it is not surprising that multiple genetic alterations are required, and that these changes are cooperative. 'This hypothesis has been supported by several experiments. For example, Reed et al. (1990) co-transfected the oncogenes bclz (thought to encode a protein with GTP-binding activity) and H-ras into rat embryo fibroblasts. The transfectants were transformed as evidenced by focus formation, growth in soft agar and tumorigenicity. They also performed a series of co-transfections, including ras and EIA, ras and c-myc, bcl-Z and EIA, and bcl-Z and c-myc. Co-transfectants expressing ras were malignantly transformed, whereas those expressing bcl- 65 2 were not, suggesting that although ras and bcl-2 have similar biochemical properties, they act in distinct cellular regulatory pathways. The importance of myc and ras cooperation was shown by Sklar et al. in 1991. In these studies ras-transformed tumorigenic NIH-3T3 cells transfected with a vector containing anti-sense c-myc were found to revert to a less tumorigenic, less transformed phenotype. Transfection of anti- sense» myb and jun did not produce reversion, suggesting that ras transformation occurs through a myc-mediated pathway. However, Schfitte et al. (1989) showed that c-H—ras and c-jun can cooperate to transform primary rat embryo cells. Binétruy et al. (1991) showed that expression of an activated H-ras caused increased transcriptional activity of the c- jun protein, associated with its hyperphosphorylation. They concluded that expression of activated ras initiates a protein kinase cascade resulting in a hyperphosphorylated, more active c-jun protein. Further studies are needed to form meaningful conclusions about the role of these oncogenes and the means by which their products cooperate to produce transformation. 3. Alterations in ras Expression There is evidence that the transforming capacity of the ras oncogenes can be altered by increased expression as well as by point mutation as is true of several other oncogenes. Duesberg and colleagues (Chakraborty et al., 1991) showed that even a wild-type ras gene is capable of transforming the mouse fibroblast line C3H/10Ti/2, although the presence of activating mutations greatly enhances transforming capacity. It should be noted; however, that this cell line has an infinite life span and is transformed with a frequency as high as 3% after carcinogen 66 treatment. Therefore, it is likely that other genetic changes in that cell line are cooperating with increased ras expression to cause transformation. Similar experiments performed by Ricketts and Levinson (1988) showed that over-expression of the wild-type ras gene can partially transform Rat-1 cells, but only when the level of expression are very high. Several additional studies have shown that transfection of activated ras genes, either under the control of an inducible promoter (Seyama et al. 1988), or under selective conditions leading to gene amplification (Sistonen et al. 1987), transforms recipient cells to a greater or lesser extent, depending on the level of expression achieved. Investigators have also examined the effect of alterations in the level of expression of both wild-type and mutant endogenous ras. For example, Cohen and Levinson (1988) studied the T24 H-ras gene and showed that in addition to the mutation at codon 12 known to result in activation, there is an additional mutation in the last intron. This latter mutation results in a ten-fold increase in expression over that ordinarily found in cells. Hithout that mutation, the T24 gene shows decreased transforming efficiency. Comparison of this region of the T24 H-ras gene to that of H- ras genes from a panel of human genomic DNA showed that this alteration is not typically found in H-ras genes, leading to the speculation that the intron sequence'alteration is a second somatic event that is important for full transforming capacity. In subsequent experiments to determine the role of the altered sequence, Cohen et al. (1989) found that the mutation abolishes a splice site, and that this results in an altered transcript encoding a truncated 19kDa protein. This altered transcript also exists in cells with wild-type H-ras gene, but is present only in low levels. Since the truncated protein lacks the normal carboxyl terminus required L/—. 67 for normal activity, it is thought to be non-functional and to suppress the expression of the normal protein. Loss of this splice site is expected to result in an increase in p21 expression. In the case of T24, this over-expression should result in transformation. Hashimoto-Gotoh et al. (1988) have reported other regulatory elements in the first intron of H-ras that appear to be conserved regions found in several proto- oncogenes, leading to the speculation that the expression of these proto- oncogenes is tightly regulated, and that derangements in this regulation contributes to transformation. Coordination of expression of the different ras genes is also thought to play a role in ras-mediated transformation. Bizub et al. (1987) studied the expression of H- and K-ras in epithelial cells at several different stages of transformation and found an abundance of expression of H-ras. In view of many other reports of the frequency of H- ras activation in chemically-induced mouse skin tumors, these investigators hypothesized that cells in which one of the ras genes is preferentially expressed are a target for transformation by agents that can mutate that particular ras gene. In support of this idea, they also cite unpublished data on the increased expression of N-ras, and the increased frequency' of mutations in that gene, that are found in hematopoietic.malignancies. In contrast, Campisi et al. (1984) showed that although K-ras was mutated in a mouse fibroblast cell line transformed by chemical carcinogens, cell cycle regulated expression of K-ras was not lost, whereas normal regulation of myc expression was altered in the transformed cells. Clearly more work needs to be done to gain a clearer understanding of the normal control of ras gene expression and the role of altered regulation of expression in malignant transformation. 68 V. Methods for Deternining the Presence of Activated ras Genes A. Early Methods Early methods for detecting cellular changes associated with malignant transformation were based on the rationale that malignant transformation had a genetic basis, and that certain in vitro phenotypic changes correlated with transformation. However, the specific genetic changes responsible’for transformation had not been identified. ‘Therefore one approach was to transfer DNA from a transformed cell line into test cells, observe them for the known transformation-associated phenotypes, isolate the DNA from the cells that had acquired this phenotype, and search for the specific oncogene that was responsible. Heinberg and colleagues (Shih et al., 1979) reported such an experiment, using DNA from rat cells transformed by carcinogen treatment. The DNA was transferred into a mouse cell line, NIH-3T3, known to take up large quantities of DNA and to exhibit specific transformation-associated changes in its normal growth. Some of the recipient cells were, indeed, transformed, forming dense piled-up areas of cell overgrowth termed foci. This assay was subsequently used to identify the presence of transforming genetic material. As methods developed to identify specific regions of DNA (for example Southern analysis and probing with known fragments of DNA) it became clear that ras genes bearing specific mutations were capable of transforming NIH-3T3 to focus formation. The NIH—3T3 focus assay, and variants utilizing transformation to tumorigenicity (Blair et al., 1982) were used to identify specific activating ras mutations in a number of tumor cells derived from various human tissues. These mmtations are located at codons 12, 13, 59, 60, 61, 117, and 119. This method has some 69 limitations that have decreased its utility for detecting mutations in ras genes. First, its basis is the ability of a ras gene carrying a mutation to cause a recognizable phenotypic change, i.e.. focus formation. Therefore, any mutation that is capable of activating the ras gene, but that does not result in focus formation would be missed. This was documented by Higashi et al. (1990), who showed that a mutation at codon 146 in the K-ras gene resulted in no significant increase in focus formation in NIH-3T3 cells, but did efficiently transform a hamster cell line. Secondly, this method is a screening tool, and even when focus formation is induced, the DNA fragment responsible must be identified by other methods. Finally the method is cumbersome and time-consuming, and as more sophisticated approaches have evolved, this method has been used less frequently. 8. Later Methods The invention of the polymerase chain reaction (PCR) method of amplifying specific segments of DNA, which is capable of synthesizing large quantities of DNA from a very small starting sample (often as little as one cell), has resulted in a number of new techniques to search for the presence of mutations in ras genes. Among the most commonly used is mismatch Oligonucleotide hybridization. In this method, a fragment of DNA containing a sequence of interest is specifically amplified, and probed with one of several radiolabeled oligonucleotides designed to match either the wild-type sequence, or one of the possible mutants. By utilizing highly stringent hybridization conditions, the sequence of the amplified fragment can be inferred from the Oligonucleotide with which it can still hybridize. Mutations in ras genes of many human tumor cells 70 have been identified by this method. In one example, Vogelstein and colleagues (Sidransky et al., 1992) were able to identify K-ras mutations in sloughed cells obtained from stool of patients with colon cancer. These mutations corresponded to the K-ras mutations observed in the cells taken from colon adenocarcinoma removed from those patients. Sequence analysis can also be performed on amplified DNA. Sequence determination has the additional benefit of identifying wild-type and mutant fragments simultaneously. In one example, Mane et al. (1990) determined the sequence of an N-ras fragment from patients with acute 'myelogenous leukemia. They showed that a mutation in the amplified DNA could be detected by sequencing even when the mutant DNA had been diluted ten-fold with the corresponding wild-type DNA. Some PCR-related methods combine the amplification with analytical methods. These include the use of amplification primers that are designed to anneal specifically to DNA containing a mutation (Ehlen and Dubeau, 1989), as well as primers designed to form a specific restriction endonuclease recognition site when annealing to target DNA containing a specific mutation (Kumar and Barbacid, 1988). Mutations of ras genes can also be detected at the level of RNA or protein expression. These methods rely on specificity of association of an RNA probe for wild-type or mutated ras messenger RNA, or on the specific interaction between an anti-p21 antibody and wild-type or mutant protein. An example of RNA detection was reported by Lang et al. in 1989 who used radiolabeled probes to detect H-ras message inueitn, avoiding the additional step of extraction of RNA from the cells of interest. One advantage of such a method is that it allows the screening of large numbers of cells for a specific mutation. Also, detection at the mRNA 71 level indicates that mutant genes are being expressed and, therefore, are likely to have biological function. An example of detection of activated ras p21 utilizing monoclonal antibodies to detect mutant p21 was reported by La Vecchio et al. (1990). These investigators generated two monoclonal antibodies to a synthetic peptide corresponding to residues 5-16 of ras p21 that contained a specific alteration (substitution of aspartate) at codon 13. One of these monoclonal antibodies specifically immunoprecipitated cellular ras p21 containing this altered amino acid. This led to the conclusion that this monoclonal antibody would be useful for rapid detection of ras p21 activated by the presence of aspartate at residue 13. Investigators have also reported that some mutant ras proteins are detectable by alteration of electrophoretic mobility of ras protein immunoprecipitates. However, Andeol et al. (1988) showed that some cell lines known to contain mutant activated ras did not express an abnormal protein detectable by an altered immunoprecipitation pattern. ‘Therefore, this method cannot be'expected to find all activating mutations. All of the above methods for the detection of activated ras rely on previous data associating mutation at specific codons with activation of the oncogene. Therefore, mutations which have not been identified by the early functional assays (for example focus formation), are not routinely searched for, and may be missed by these specific methods. Until recently, only sequence determination of the entire coding region could provide data that were not limited by this historical bias. A very new screening method is capable of detecting the presence of a single point mutation in an amplified fragment and can conveniently be used to screen the entire coding region. This method, single strand conformation 72 polymorphism (SSCP), was shown by Suzuki et al. (1990) to be sensitive in detecting single nucleotide alterations in one strand of an amplified fragment in a series of ras genes from a number of cell lines known to contain activating point mutations. This method does not specify the nature or the precise location of the mutation, but does provide promise as a screening method to search large numbers of cells for mutations throughout the coding region. Any cells found to contain a mutation by this method can then be analyzed further using sequencing of the relevant fragment to determine the nature and exact location of the mutation. In general, several issues need to be considered when determining the most appropriate method by'which ras genetmutations are detected. One question is the type of material to be tested. For example, if the samples are to come from archival tissue, a method would need to be based on detection of DNA rather than RNA or protein since it is much more stable. Also, a method would need to be able to differentiate between malignant and normal cells in the sample. If the sample consists of a homogeneous cell line, the question of separating normal and transformed tissues need not be considered. Another issue is whether the study is intended to screen samples for possible mutations, or whether the effect of specific alterations on function is being investigated. In the former case, a more general method for detecting mutations (such as sequence determination of the coding region) might be expected to be most sensitive. In the latter case, methods designed to detect mutations in specific codons would be sufficient. In the study that I carried out for this dissertation, the purpose was to determine the role of ras activation in a specific subset of cells. Therefore, the determination of the sequence of the entire coding region was considered to provide the most 73 sensitive analysis. It was also judged to be feasible because the number of cell lines to be analyzed was relatively small. LIST OF REFERENCES Adari, H., D. R. Lowy, B. M. Hilli-sen, C. J. Der, and F. McCorIick. 1988. Guanosine triphosphateiactivating protein (GAP) interacts with the p21 ras effector binding domain. Science 240:518-521. Alitalo, K., G. Ralsay, J. M. Bishop, S. D. Pfeifer, H. H.Colby, and A. D. Levinson. 1983. Identification of nuclear proteins encoded by viral and cellular myc oncogenes. Nature 306:274-277. Alonso, 1., S. Srivastava, and E. Santos. 1990. Alterations of G-protein coupling function in phosphoinositide signaling pathways of cells transformed by ras and other membrane-associated and cytoplasmic oncogenes. Mol. Cell. Biol. 10:13117-3124. Andeol, V., P. C. Nardeaux, L. Daya-Grosjean, O. Brison, J. Cebrian, and H. Suarez. 1988. Both N-ras and c-myc are activated in the SHAC human stomach fibrosarcoma cell line. Int. J. Cancer 41:732-737. Anderson, S. K.. C. P. Gibbs, A. Tanaka, H—J. Kung, and D. J. Fujita. 1985. Human cellular src gene: nucleotide sequence and derived amino acid sequence of the region coding for the carboxy-terminal two-thirds of pp60°' '“5. Mol. Cell. Biol. 5:1122-1129. Baker, C. C., H. C. Phelps, V. Lindgren, M. J. Braun, M. A. Gonda, and P. H. Howley. 1987. Structural and transcriptional analysis of human papillomavirus type 16 sequences in cervical carcinoma cell lines. J. Virol. 61:962-971. Barrett, J. C., and P. O. Ts’o. 1978. Evidence for the progressive nature of neoplastic transformation in vitro. Proc. Nat. Acad. Sci. USA 75:3761- 3765. Bedell, M. A., K. H. Jones, and L. A. Lailins. 1987. The E6-E7 region of human papillomavirus type 18 is sufficient for transformation of NIH-3T3 and Rat-1 cells. J. Virol. 61:3635-3640. Bejcek, B. E., D. Y. Li, and T. F. Deuel. 1989. Transformation by v-sis occurs by an internal autoactivation mechanism. Science 245:1496-1499. Ben David, Y., V. R. Prideaux, V. Chow, S. Benchilol, and A. Bernstein. 1988. Inactivation of the p53 oncogene by internal deletion or retroviral integration in erythroleukemic cell lines induced by Friend leukemia virus. Oncogene 3:179-185. 74 75 Benchilol, S., P. Lad) , L. V. Crawford, 0. Sheer, T. 8. Shows, G. A. P. Bruns, and J. Peacock. 1985. Transformation associated p53 protein is encoded by a gene on chromosome 17. Somatic. Cell. Mol. Genet. 11:505-509. Bikel, I., H. Manon, E. L. Brown, J. Boltax, A. Mounzer, and D. M. Livingston. 1986. The t-unique coding domain is important to the transformation maintenance function of the Simian Virus 40 small t antigen. Mol. Cell. Biol. 6:1172-1178. Binétruy, 8., T. Sleal, and M. Karin. 1991. Ha-ras augments c-jun activity . and stimulates phosphorylation of its activation domain. Nature 351:122- 127. Bishop, J. M. 1983. Cellular Oncogenes and retroviruses. Annu Rev Biochem 52:301-354. Bister, K., M. J. Hay-an, and P. K. Vogt. 1977. Defectiveness of avian myelocytomatosis virus MC29: isolation of long-term nonproducer cultures and analysis of virus-specific polypeptide synthesis. Virology 82:431-448. Bizui, D., A. M. Hood, and A. M. Skalka. 1986. Mutagenesis of the Ha-ras oncogene in mouse skin tumors induced by polycyclic aromatic hydrocarbons. Proc. Nat. Acad. Sci. USA 83:6048-6052. Bizub, D., E. P. Heiler, A. Felix, R. Chizzonite, A. Hood, A. M. Skalka, D. Slater, T. H. Aldrich, and M. E. Furth. 1987. Antisera to the variable region of ras oncogene proteins, and specific detection of H-ras expgession in an experimental model of chemical carcinogenesis. Oncogene :1 1-142. Blair, D. G., C. S. Cooper, M. K. Oskarsson, L. A. Eader, and G. F. Vande Houde. 1982. New method for detecting cellular transforming genes. Science 218:1122-1125. Bohnnn, D., T. J. Dos, A. Adam, T. Nishilura, P. K. Vogt, and R. Tjian. 1987. Human Proto-oncogene c-jun encodes a DNA binding protein with structural and functional properties of transcription factor AP-l. Science 238:1386-1392. Bourne, H. R., D. A. Sanders, and F. McCorIick. 1991. The GTPase superfamily: conserved structure and molecular mechanism. Nature 349:117- 127. Boutwell, R. K. 1964. Some biological aspects of skin carcinogenesis. Prog. Exp. Tumor. Res. 4:207-250. Bruskin, A., J. Jackson, J. M. Bishop, D. J. McCarley, and R. C. Schatzlan. 1990. Six amino acids from the retroviral gene gag greatly enhance the transforming potential of the oncogene v-erb-B. Oncogene 5:15- 24. 76 Calender, A., M. Billaud, J-P. Aubry, J. Banchereau, M. Vuillaue, and G. M. Lenoir. 1987. Epstein-Barr virus (EBV) induces expression of B-cell activation markers on in vitro infection of EBV-negative 8-lymphoma cells. Proc. Nat. Acad. Sci. USA 84:8060-8064. Camisi, J., H. E. Gray, A. B. Pardee, M. Dean, and G. E. Sonenshein. 1984. Cell-cycle control of c-myc but not c-ras expression is lost following chemical transformation. Cell 36:241-247. m, n. J., E. '0 cm. A. D. LEV‘IIISUI, P. "o seam. Il'ld n. V. Goeddel. 1983. Complete nucleotide sequences of the T24 human bladder carcinoma oncogene and its normal homologue. Nature 302:33-37. Cavenee, H. K., T. P. Dryja, R. A. Phillips , H. F. Benedict, R. Godbout, B. L. Gallie, A. L. Murphree, L. C. Strong, and R. L. Hhite. 1983. Expression of recessive alleles by chromosomal mechanisms in retinoblastoma. Nature 305:779-784. Chakraborty, A. K., K. Cichutelg, and P. H. Duesberg. 1991. Transforming function of proto-ras genes depends on heterologous promoters and is enhanced by specific point mutations. Proc. Nat. Acad. USA 88:2217-2221. , S. H., H. Piwnica-Horls, R. H. Harvey, T. M. Roberts, and A. E. Slith. 1988. The carboxy terminus of pp60°fl5 is a regulatory domain and is involved in complex formation with the middle-T antigen of polyomavirus. Mol. Cell. Biol. 8:1736-1747. Clanton, D. J., S. Hattori, and T. V. Shih. 1986. Mutations of the ras gene product p21 that abolish guanine nucleotide binding. Proc. Nat. Acad. Sci. USA 83:5076-5080. Clark, R., M. R. Stamfer, R. Milley, E. O’Rourke, K. H. Halen, M. Kriegler, J. Kopplin and F. McCorIick. 1988. Transformation of human mammary epithelial cells by oncogenic retroviruses. Cancer Res. 48:4689- 4694. Clayton, C. E., D. Murphy, M. Lovett, and P. H. J. Rigby. 1982. A fragment of the SV40 large T-antigen gene transforms. Nature 299:59-65. Cohen, J. D., and A. D. Levinson. 1988. A point mutation in the last intron responsible for increased expression and transforming activity of the c-Ha-ras oncogene. Nature 334:119-124. Cohen, J. 8., S. D. Broz, and A. D. Levinson. 1989. Expression of the H- ras proto-oncogene is controlled by alternative splicing. Cell 58:461-472. Coussens, P. M., J. A. Cooper, T. Hunter, and D. Shalloway. 1985. Restriction of the in vitro and in vivo tyrosine protein kinase activities of pp60"m relative to pp60"m. Mol. Cell. Biol. :2753-2763. Curran, T., and B. R. Franza. 1988. Fos and jun: the AP-l connection. Cell 55:395-397. 77 Dalbaugh, T., F. Hang, K. Hennessy, E. Headland, A. Rickinson, and E. Kieff. 1986. Expression of the Epstein-Barr virus nuclear protein 2 in rodent cells. J. Virol. 59:453-462. Dandekar, S., S. Sukular, H. Zarbl, L. J. T. Young, and R. D. Cardiff. 1986. Specific activation of the cellular Harvey-ras oncogene in gimethylbenzanthracene-induced mouse mammary tumors. Mol. Cell. Biol. :4104-4108. de Ronde, A., C. J. A. Sol, M. E. MacDonald, M. Koot, J. ter Schegget, A. van Strien, E. Routers and J. van der Noordaa. 1987. Two domains within the early coding region of SV40 involved in the transformation of human fibroblasts. Intervirology 28:221-231. W10. J. A., J. "o LINN", Jo F1996, J'Yo Sh“. c."o "m. "-“o In.” E. Marsillo E. Paucha and D. M. Livingston. 1988. SV40 large tumor antigen forms a specific complex with the product of the retinoblastoma susceptibility gene. Cell 54:275-283. Deppert, H., T. Stein-ayer, and H. Richter. 1989. Cooperation of SV40 large T antigen and the cellular protein p53 in maintenance of cell transformation. Oncogene 4:1103-1110. Der, C. J., T. Finkel, G. M. and Cooper. 1986. Biological and biochemical properties of human rasfl genes mutated at codon 61. Cell 44:167-176. Der, C. J., B-T. Pan, and G. M. Cooper. 1986. ras" mutants deficient in GTP binding. MOl. Cell. Biol. 5:3291-3294. de Ronde, A., C. J. A. Sol, A. van Strien, J. ter Schegget, and J. van der Hoordaa. 1989. The SV40 small t antigen is essential for the morphological transformation of human fibroblasts. Virology 171:260-263. Di Fiore, P. P., D. Segatto, H. G. Taylor, S. A. Aaronson, and J. H. Pierce. 1990. EGF receptor and erbB-2 tyrosine kinase domains confer cell specificity for mitogenic signaling. Science 248:79-83. DiMaio, D., D. Guralski, and J. T. Schiller. 1986. Translation of Open reading frame E5 of bovine papillomavirus is required for its transforming activity. Proc. Nat. Acad. Sci. USA 83:1797-1801. Donner, P., I. Greiser-Hilke, and K. Moelling. 1982. Nuclear localization and DNA binding of the transforming gene product of avian myelocytomatosis virus. Nature 296:262-266. Doolittle, R. E., M. H. Hunkapiller, L. E. Hood, S. G. Devare, K. C. Robbins, S. A. Aaronson, and H. N. Atoniades. 1983. Simian sarcoma virus onc gene, v-sis, is derived from the gene (or genes) encoding a platelet- derived growth factor. Science 221:275-277. 78 Downward, J., V. Varden, E. Hayes, G. Scrace, N. Totty, P. Stockwell, A. Ulrich, J. Schlessinger, and M. D. Haterfield. 1984. Close similarity of epidermal growth factor receptor and v-erb-B oncogene protein sequences. Nature 307:521-527. Duesberg, P. H., K. Bister, and C. Moscovici. 1980. Genetic structure of avian myeloblastosis virus, released from transformed myeloblasts as a defective virus particle. Proc. Nat. Acad. Sci. USA 77:5120-5124. Dyson, H., P. M. Howley, K. Hunger, and E. Harlow. 1989. The human papilloma virus-16 E7 oncoprotein is able to bind to the retinoblastoma gene product. Science 243:934-936. Earle, H. R. 1943. Production of malignancy in vitro. IV. The mouse fibroblast cultures and changes seen in the living cells. J. Nat. Cancer Inst. 4:165-212. Ehlen, T., and L. Dibeau. 1989. Detection of ras point mutations by polymerase chain reaction using mutation-specific inosine-containing oligonucleotide primers. Biochem. Biophys. Res. Commun. 160:441-447. Eliyahu, D., M. Goldfinger, O. Pinhasi-Kiflii, G. Shaulsky, V. Skurnik, N. Arai, V. Rotter, and M. Oren. 1988. Meth A fibrosarcoma cells express two transforming mutant p53 species. Oncogene 3:313-321. E1113. R. "o. no ”90, To '0 SI'IflI . "o A. m. H. A. Tom, "0 Tsuchida, D. R. Lowy, and E. M. Scolnick. 1981. The p21 src of Harvey and Kirsten sarcoma viruses originate from divergent members of a family of normal vertebrate genes. Nature 292:506-511. Ewen, M. E., J. H. Ludlow, E. Marsillio, J. A. DeCaprio, R. C. Millikan, S. M. Cheng, E. Paucha, and D. M. Livingston. 1989. An N-terminal transformation-governing sequence of SV40 large T antigen contributes to :36 binding of both p110“ and a second cellular protein, p120. Cell :257-267. Fasano, D., E. Taparowsky, J. Fiddes, M. Higler, and M. Goldfarb. 1983. Sequence and structure'of the coding region of the human H-ras-1 gene from T24 bladder carcinoma cells. J. Mol. Appl. Genet. 2:173-180. Finlay, C. A., P. H. Hinds, and A. J. Levine. 1989. The p53 proto-oncogene can act as a suppressor of transformation. Cell 57:1083-1093. Finlay, C. A., P. H. Hinds, Tefl. Tan, D. Eliyahu, M. Oren, and A. J. Levine. 1988. Activating mutations for transformation by p53 produce a gene product that forms an hsc70-p53 complex with an altered half-life. Mol. Cell. Biol. 8:531-539. Forrester, K., C. Alloguera, K. Han, H. E. Grizzle, and M. Perucho. 1987. Detection of high incidence of K-ras oncogenes during human colon tumorigenesis. Nature 327:298-303. 79 Friend, S. H., R. Bernards, S. Rogelj, R. A. Heinberg, J. M. Rapaport, D. M. Albert, and T. P. Dryja. 1986. A human DNA segment with properties of the gene that predisposes to retinoblastoma and osteosarcoma. Nature 323:643-646. Fry, D. G., L. D. Milan, V. M. Maher, and J. J. McCorIick. 1986. Transformation of diploid human fibroblasts by DNA transfection with the v-sis oncogene. J. Cell. Physiol. 128:313-321. Fry, D. G., L. D. Milan, J. E. Dillberger, V. M. Maher, and J. J. McCorlick. 1990. Malignant transformation of an infinite life span human fibroblast cell strain by transfection with v-Ki-ras. Oncogene 5:1415- 1418. Furth, M. E., T. H. Aldrich, and C. Cordon-Cardo. 1987. Expression of ras proto-oncogene proteins in normal human tissues. Oncogene 1:47-58. Gauwerky, C. E., F. G. Haluska, V. Tsujiloto, P. C. Howell, and C. M. Croce. 1988. Evolution of B-cell malignancy: pre-B-cell leukemia resulting from MYC activation in a B-cell neoplasm with a rearranged 8CL2 gene. Proc. Nat. Acad. Sci. USA 85:8548-8552. Gey, 6.0. 1941. Cytological and cultural observations on transplantable rat sarcomata produced by the inoculation of altered normal cells maintained in continuous culture. Cancer Res. 1:737. Giese, M. A., K. C. Robbins, and S. A. Aaronson. 1987. The role of individual cysteine residues in the structure and function of the v-sis gene product. Science 236:1315-1318. Giese, N. A., H. J. Larochelle, M. Mabeiroff, K. C. Robbins, and S. A. Aaronson. 1990. A small v-sis/platelet-derived growth factor (PDGF) 8- protein domain in which subtle conformational changes abrogate PDGF receptor interaction and transforming activity. Mol. Cell. Biol. 10:5496- 5501. Gould, K.,L., and T. Hunter. 1988. Platelet-derived growth factor induces multisite phosphorylation of pp60””3 and increases its protein-tyrosine kinase activity. Mol. Cell. Biol. :3345-3356. Goyette, M., C. J. Petropoulos, P. R. Shank, and N. Fausto. 1983. Expression of a cellular oncogene during liver regeneration. Science 219:510-512. Grand, R. J. A., and D. Owen. 1991. The biochemistry of ras p21. Biochem. J. 279:609-631. Green, M., and P. M. Loewenstein. 1987. Demonstration that a chemically synthesized BPVI oncoprotein and its c-terminal domain function to induce cellular DNA synthesis. Cell 51:795-802. 80 Guillel, J. G., A. C. O’Brian, C. J. Fitzer, K. A. Forde, P. LoGerfo, M. Treat, I. B. Heinstein. 1987. Altered levels of protein kinase C and Ca”- dependent protein kinases in human colon carcinomas. Cancer Res. 47:2036- 2039. Halberstaedter, L., L. Doljanski, and E. Tenenbaul. 1941. Experiments on the cancerization of cells in vitro by means of Rous sarcoma agent. Br. J. Exper. Path. 22:179-187. Hall, A. 1990. The cellular functions of small GTP-binding proteins. Science 249:635-640. Hall, A., and R. Brown. 1985. Human N-ras: cDNA cloning and gene structure. Nucleic Acids Res. 13:5255-5268. Hall, A., C. J. Marshall, N. K. Spurr, and R. A. Heiss. 1983. Identification of transforming gene in two human sarcoma cell lines as a new member of the ras gene family located on chromosome 1. Nature 303:396- 400. Hancock, J. F., C. J. Marshall, I. A. McKay, 1988. Mutant but not normal p21 ras elevates inositol phospholipid breakdown in two different cell systems. Oncogene 3:187-193. Hannink, M., and D. J. Donoghue. 1986. Cell surface expression of membrane-anchored v-sis gene products: glycosylation is not required for cell surface transport. J. Cell. Biol. 103:2311-2322. Harper, J. R., D. R. Roop, and S. H. Yuspa. 1986. Transfection of the EJ ras gene into keratinocytes derived from carcinogen-induced mouse papillomas causes malignant progression. Mol. Cell. Biol. 6:3144-3149. Harvey, J. J. 1964. An unidentified virus which causes the rapid production of tumors in mice. Nature 204:1104-1105. Hashiloto-Gotoh, T., R. Kikuno, M. Takahashi, and H. Honkawa. 1988. Possible role of the first intron of c-Ha-ras in gene expression: anti- cancer elements in oncogenes. Anticancer Res. 8:851-860. Haubruck, H., and F. McCorIick. 1991. Ras p21: effects and regulation. Biochim. Biophys. Acta 1072:215-229. Hayward, ii. 5., B. G. Mel, and S. M. Astrin. 1981. Activation of a cellular one gene by promoter insertion in ALV-induced lymphoid leukosis. Nature 290:475-480. "19”.", To. “0 5‘5“, F. SIIZI‘T. J. "1305'", To “M1. $0-1. lell. T. Mori, and T. Kakunaga. 1990. Hamster cell line suitable for transfection assay of transforming genes. Proc. Nat. Acad. Sci. USA 87:2409-2413. Hinds, P., C. Finlay, and A. J. Levine. 1989. Mutation is required to activate the p53 gene for cooperation with the La; oncogene and transformation. J. Virol. :739-746. 81 Holliday, R. 1987. The inheritance of epigenetic defects. Science 238:163- 170. Hsiao, H.-L. H., G. M. Housey, M. D. Johnson, and I. B. Heinstein. 1989. Cells that overproduce protein kinase C are more susceptible to transformation by an activated H-Lae oncogene. Mol. Cell. Biol. 9:2641- 2647. Hsieh, L. L., S. Hoshina, and I. D. Heinstein. 1989. Phenotypic effects of oIerexpression of PKC,1 in rat liver epithelial cells. J. Cell. Biochem. 4 :179-188. Huang, S. S., and J. S. Huang. 1988. Rapid turnover of the platelet- derived growth factor receptor in sis-transformed cells and reversal by suramin. J. Biol. Chem. 263:12608-12618. “Mg ".4. So. JQ'KO Yee. Jo'Yo Sh”. PO-L. Chm, R. “Stein. To Fried-am, E. V.-H. P. Lee, and H.-H. Lee. 1988. Suppression of the neoplastic phenotype by replacement of the RB gene in human cancer cells. Science 242:1563-1566. Huang, J. S., S. S. Huang, and T. F. Deuel. 1984. Transforming protein of simian sarcoma virus stimulates autocrine growth of SSV-transformed cells through PDGF cell-surface receptors. Cell 39:79-87. Hurlin, P. J., D. G. Fry, V. M. Maher, and J. J. McCorlick. 1987. Morphological transformation, focus formation, and anchorage independence induced in diploid human fibroblasts by expression of a transfected H-ras oncogene. Cancer Res. 47:5752-5757. Hurlin, P. J., V. M. Maher, and J. J. McCorIick. 1989. Malignant transformation of human fibroblasts caused by expression of a transfected T24 HRAS oncogene. Proc. Nat. Acad. Sci. USA 86:187-191. Iba, H., F. R. Cross, E. A. Garber, and H. Hanafusa. 1985. Low level of cellular protein phosphorylation by nontransforming overproduced p60"”5. Mol. Cell. Biol. 5:1058-1066. Isobe, M., B. S. Elanuel, D. Givol, M. Oren, and C. M. Croce. 1986. Localization of gene for human p53 tumour antigen to band 17p13. Nature 320:84-85. Jackson, J. H., C. G. Cochrane, J. R. Bourne, P. A. Solski, J. E. Buss, and C. J. Der. 1990. Farnesol modification of Kirsten-ras exon 48 protein is essential for transformation. Proc. Nat. Acad. Sci. USA 87:3042-3046. Johnson, P. J., P. M. Coussens, A. V. Danko, and D. Shalloway. 1985. Overexpressed pp60°151 can induce focus formation without complete transformation of NIH 3T3 cells. Mol. Cell. Biol. 5:1073-1083. Kaldor, J., J. Peto, D. Easton, R. Doll, C. Her-on, and L. Morgan. 1986. Models for respiratory cancer in nickel refinery workers. J. Nat. Cancer Inst. 77:841-848. 82 Kasid, A., M. E. Lippnan, A. G. Papageorge, D. R. Lowy, and E. P. Gelnan. 1985. Gene for human insulin receptor: localization to site on chromosome 19 involved in pre-B-cell leukemia. Science 228:725-728. Khandijian, E. H., and J.-F. Gauchat. 1988. One single burst of SV40 gene expression induces cell proliferation in transforming infection of mouse cells. Oncogene 3:195-200. Kinsella, A. R., L. Fiszer-Maliszewska, E. L. D. Mitchell, Y. Guo, M. Fox, and D. Scott. 1990. Introduction of the activated N-ras oncogene into human fibroblasts by retroviral vector induces morphological transformation and tumorigenicity. Carcinogenesis 11:1803-1809. Kirsten, H. H., and L. A. Mayer. 1967. Morphologic responses to a murine erythroblastosis virus. J. Nat. Cancer Inst. 39:311-335. Kitayana, H., V. Suginoto, T. Matsuzaki, Y. Ikawa, and M. Moda. 1989. A ras-related gene with transformation suppressor activity. Cell 56:77-84. Kleqinauer, K.-H., G. Synonds, G. I. Evan, and J. M. Bishop. 1984. Subcellular localization of proteins encoded by oncogenes of avian myeloblastosis virus and avian leukemia virus E26 and by the chicken c-myb gene. Cell 37:537-547. Knudsen, A. G. 1971. Mutation and cancer: statistical study of retinoblastoma. Proc. Nat. Acad. Sci. USA 68:820-823. Kraiss, S., A. Ouaiser, M. Oren, and M. Montenarh. 1988. Oligomerization of oncoprotein p53. J. Virol. 62:4737-4744. Krengel, U., I. Schlichting, A. Scherer, R. Schunann, M. Frech, J. John, H. Kabsch, E. Pai, and A.Hittinghofer. 1990. Three-dimensional structures of H-ras mutants: molecular basis for their inability to function as signal switch molecules. Cell 62:539-548. Kunar, R., and M. Barbacid. 1988. Oncogene detection at the single cell level. Oncogene 3:647-651. Kunar, R., S. Sukunar, and M. Barbacid. 1990. Activation of ras oncogenes preceding the onset of neoplasia. Science 248:1101-1104. La vecchio, J., P. J. Haner, S. C. Mg, K. L. Trinpe, and H. P. Carney. 1990. Characterization of monoclonal antibodies specific to the activated ras p21 with aspartic acid at position 13. Oncogene 5:1173-1178. Lawn, H. H., P. Hansley, P. Sassone-Corsi, and I. M. Verna. 1988. Induction of proto-oncogene JUN/AP-l by serum and TPA. Nature 334:629-631. Lang, S. M., R. G. Morris, K. A. Rose, and A. H. Hyllie. 1989. Detection of human Ha-ras transcripts in tumor cells by in situ hybridization. Anticancer Res. 9:805-810. 83 LIROCherIe, '0 JO. “0 Giese. "o m-Simff’ K. c. “Wins. m S. A. Aaronson. 1990. Molecular localization of the transforming and secretory properties of PDGF A and PDGF 8. Science 248:1541-1544. Larose, A., N. Dyson, M. Sullivan, E. Harlow, and M. Bastin. 1991. Polyomavirus large T mutants affected in retinoblastoma protein binding are defective in immortalization. J. Virol. 65:2308-2313. Lee, F., and M. J. Hay-an. 1990. Analysis of the role off the extracellular domain of the v-erbB oncogene in cellular transformation. Oncogene 5:1165-1172. L”. "a "a. R. “Stein, F. m. c. J. Ym. J'Yo Shag m E. Y. H. P. Lee. 1987. Human retinoblastoma susceptibility gene: cloning, identification, and sequence. Gene 235:1394-1399. Levine, A. J., and J. Monand. 1990. Tumor suppressor genes: the p53 and Ietinoblastoma sensitivity'genes and gene products. Biochim. Biophys. Acta 032:119-136. Linzer, D. I. H., and A. J. Levine. 1979. Characterization of a 54K dalton cellular SV40 tumor antigen present in SV40-transformed cells and uninfected embryonal carcinoma cells. Cell 17:43-52. Lokeshwar, V. D., S. S. Huang, and J. S. Huang. 1990. Intracellular turnover, novel secretion, and mitogenically-active intracellular forms of v-sis gene product in simian sarcoma virus-transformed cells. J. Biol. Chem. 265:1665-1675. Ludlow, J. H., J. A. DeCaprio, C-M. Huang, H-H. Lee, E. Paucha, and D. M. Livingston. 1989. SV40 large T antigen binds preferentially to an underphosphorylated member of the retinoblastoma susceptibility gene product family. Cell 56:57-65. Maihle, N. J., M. A. Raines, T. H. Flickinger, and H-J. Kung. 1988. Proviral insertional activation of c-erbB: differential processing of the protein products arising from two alternate transcripts. Mol. Cell. Biol. :4868-4876. Mane, S. M., S. J. Meltzer, J. C. Gutheil, V. Kapil, E. J. Lee, and S. H. Heedlenan. 1990. RAS gene activation in acute myelogenous leukemia: analysis by in vitro amplification and DNA base sequence determination. Genes Chromosomes and Cancer 2:71-77. Manne, V., E. Bekesi, and H-F. Kung. 1985. Ha-ras proteins exhibit GTPase activity: point mutations that activate Ha-ras gene products result in decreased GTPase activity. Proc. Nat. Acad. Sci. USA 82:376-380. Manservigi, R., E. Cassai, L. P. Deiss, 0. Di Luca, V. Segala, and N. Frenkel. 1986. Sequences homologous to two separate transforming regions of herpes simplex virus DNA are linked in two human genital tumors. Virology 155:192-201. 84 McBride, 0. H., D. C. Swan, E. Santos, M. Barbacid, S. R. Ironick, and S. A. Aaronson. 1982. Localization of the normal allele of T24 human bladder carcinoma oncogene to chromosome 11. Nature 300:773-774. McCoy, M. S., and R. A. Heinberg. 1986. A human Ki-ras oncogene encodes two transforming p21 proteins. Mol. Cell. Biol. 6:1326-1328. W‘th. J. P., D. J. cam. 00 "0 “1th. E. Y. cm. P. H. mm. n. v. Goeddel, and A. D. Levinson. 1983. Structure and organization of the human Ki-ras proto-oncogene and a related processed pseudogene. Nature 304:501- 506. Milner, J., E. A. Medcalf, and A. C. Cook. 1991. Tumor suppressor p53: analysis of wild-typeiand mutant p53 complexes. Mol. Cell. Biol. 11:12-19. Miyanoto, S., S. Suki-ar, R. C. Guznan, R. C. Osborn, and S. Mandi. 1990. Transforming c-Ki-ras mutation is a preneoplastic event in mouse mammary carcinogenesis induced in vitro by N-methyl-N-nitrosourea. Mol. Cell. Biol. 10:1593-1599. Miyoshi, J., M. Kaginoto, E. Soeda, and Y. Sakaki. 1984. The human c-Ha- rasZ is a processed pseudogene inactivated by numerous base substitutions. Nucleic Acids Res. 12:1821-1828. Moolgavkar, S. H., and A. G. Knudsen. 1981. Mutation and cancer: A model for human carcinogenesis. J. Nat. Cancer Inst. 66:1037-1052. Morgan, T. L., D. Yang, D. G. Fry, P. J. Hurlin, S. K. Kohler, V. M. Maher, and J. J. McCornick. 1991. Characteristics of an infinite life span diploid human fibroblast cell strain and a near-diploid strain arising from a clone of cells expressing a transfected v-myc oncogene. Exp. Cell. Res. 197:125-136. Mulcahey, L. S., M. R. Snith, and D. H. Stacey. 1985. Requirement for ras proto-oncogene function during serum-stimulated growth of NIH-3T3 cells. Nature 313:241-243. Muller, R., n. J. Slanon, E. n. Adanson, J. n. Trenblay, D. Muller, n. J. Cline, and I. M. Verna. 1983. Transcription of c-onc genes c-ras“ and c- fms during mouse development. Mol. Cell. Biol. :1062-1069. Manger, K., v. c. Phelps, v. Bubb, P. n. Howley, and R. Schlegel. 1939. The E6 and E7 genes of the human papillomavirus type 16 together are necessary and sufficient for transformation of primary human keratinocytes. J. Virol. 63:4417-4421. Munroe, D. G., B. Rovinski, A. Bernstein, and S. Benchinol. 1988. Loss of a highly conserved domain on p53 as a result of gene deletion during Friend virus-induced erythroleukemia. Oncogene 2:621-624. Nadia, M., K. Nishitani, F. Fukushina, T. Kinoto, and K. Nose. 1986. Multistep process of neoplastic transformation of normal human fibroblasts by “Co gamma rays and Harvey sarcoma viruses. Int. J. Cancer. 37:419-423. 85 , M., S. H. Hinrichs, R. K. Reynolds, 6. Khoury, and G. Jay. 1987. The tat gene of human T-lymphotropic virus type 1 induces mesenchymal tumors in transgenic mice. Science 237:1324-1329. Neri, A., D. M. Knowles, A. Greco, F. McCornick, and R. Dalla-Favera. 1988. Analysis of RAS oncogene mutations in human lymphoid malignancies. Proc. Nat. Acad. Sci. USA 85:9268-9272. Nieto, A., P. K. Dasha, H. L. Nakhasi, and R. Dhar. 1987. Differential regulation of c-Ha-ras and c-Ki-ras gene expression in rat mammary gland. Carcinogenesis 8:1955-1958. Moda, M., H. Kitayana, T. Matsuzaki, V. Suginoto, H. mayana, R. H. Bassin, and Y. Ikawa. 1989. Detection of genes with a potential for suppressing the transformed phenotype associated with activated ras genes. Proc. Nat. Acad. Sci. USA 86:162-166. Noda, T., M. Satake, Y. Yanaguchi, and Y. Ito. 1987. Cooperation of middle and small T antigens of polyomavirus in transformation of established fibroblast and epitheliallike cell lines. J. Virol. 61:2253-2263. Pai, EL. F., IL. Kabsch, IL. Krengel, KL (L. Holnes, iL. John, tuwl A. Hittinghofer. 1989. Structure’of the guanine-nucleotide-binding domain of £2: Ha-ras oncogene product p21 in the triphosphate conformation. Nature :209-214. Pallas, D. C., L. K. Shahrik, B. L. Martin, S. Jaspers, T. B. Miller, D. L. Brautigan, and T. M. Roberts. 1990. Polyoma small and middle I antigens and SV40 small t antigen form stable complexes with protein phosphatase 2A. Cell 60:167-176. Parada, L. F., C. J. Tabin, C. Shih, and R. A. Heinberg. 1982. Human EJ bladder carcinoma oncogene is homologue of Harvey sarcoma virus ras gene. Nature 297:474-478. Parshad, R., K. K. Sanford, G. M. Jones, and R. E. Tarone. 1978. Fluorescent light-induced chromosome damage and its prevention in mouse cells in culture. Proc. Nat. Acad. Sci. USA 75:1830-1833. Pech, M., A. Gazit, P. Arnstein, and S. A. Aaronson. 1989. Generation of fibrosarcomas in vivo'by'a retrovirus that expresses the normal 8 chain of platelet-derived growth factor and mimics the alternative splice pattern of the v-sis oncogene. Proc. Nat. Acad. Sci. USA 86:2693-2697. Pelling, J. C., R. Meades, and J. Strawhecker. 1988. Epidermal papillomas and carcinomas induced in uninitiated mouse skin by tumor promoters alone contain a point mutation in the 615t codon of the Ha-ras oncogene. Carcinogenesis 9:665-667. Peto, R. 1977. Epidemiology. multistage models, and short-term mutagenicity tests, p. 1403-1428. In: H. H. Hiatt, J. D. Hatson, and J. A. Hinsten (ed), Origins of Human Cancer. Cold Spring Harbor. 86 Petterson, U., and R. J. Roberts. 1986. Adenovirus gene expression and replication: a historical review. p. 37-57. In: M. Botchan, T. Grodzicker, and P. A. Sharp (ed.), Cancer Cells 4: DNA Tumor Viruses Control of Gene Expression and Replication. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. Phelps, H. C., C. L. Yee, K. Monger, and P. M. Howley. 1988. The human papillomavirus type 16 £7 gene encodes transactivation and transformation functions similar to those of adenovirus £1A. Cell 53:539-547. Popescu, N. C., S. C. Msbaugh, J. A. DiPaolo, S. R. Tronick, S. A. Aaronson, and D. C. Swan. 1985. Chromosomal localization of three human {:5 genes by in situ molecular hybridization. Somatic Cell. Mol. Genet. :149-155. Quinlan, M. P., P. thte, and T. Grodzicker. 1988. Growth factor induction by the adenovirus type 5 EIA 125 protein is required for immortalization of primary epithelial cells. Mol. Cell. Biol. 8:3191-3203. Raines, M. A., N. J. Maihle, C. Moscovici. M. G. Moscovici. and H-J. Kung. 1988. Molecular characterization of three erbB transducing viruses generated during avian leukosis virus-induced erythroleukemia: extensive internal deletion near the kinase domain activates the fibrosarcoma- and hemangioma-inducing potentials of erbB. J. Virol. 62:2444-2452. Ralston, R., and J. M. Bishop. 1983. The protein products of the myc and myb oncogenes and adenovirus Ela and structurally related. Nature 306:803- 806. Rassoulzadegan, M., Z. Naghashfar, A. Cowie, A. Carr, M. Grisoni, R. Kanen, and R. Cuzin. 1983. Expression of the large T protein of polyoma virus promotes the establishment in culture of "normal" rodent fibroblast cell lines. Proc. Nat. Acad. Sci. USA 80:4354-4358. "USCIICT' III. F. J., n. R. m. T. CW“. T. J. 805, P. K. th. D. Bohnann, R. Tjian. and B. R. Franza, Jr.. 1988. Fos-associated protein p39 is the product of the jun proto-oncogene. Science 240:1010-1016. Reddy, E. P. 1983. Nucleotide sequence analysis of the T24 human bladder carcinoma oncogene. Science 220:1061-1063. Reed, J. C., S. Haldar, C. M. Croce, and M. P. Cuddy. 1990. Complementation by BCLZ and C-HA-RAS oncogenes in malignant transformation of rat embryo fibroblasts. Mol. Cell. Biol. 10:4370-4374. Rhilg Jo So. a. J“. P. msulfl, F. "0 Price. K. K. S‘nford. m S. A. Aaronson. 1985. Neoplastic transformation of human epidermal keratinocytes by A012-SV4O and Kirsten sarcoma virus. Science 227:1250-1252. Ricketts, M. H., and A. D. Levinson. 1988. High-level expression of c-H- rasl fails to fully transform Rat-1 cells. Mol. Cell. Biol. 8:1460-1468. 87 Rotter, V. 1983. p53, a transformation-related cellular-encoded protein, can be used as a biochemical marker for the deletion of primary mouse tumor cells. Proc. Nat. Acad. Sci. USA 80:2613-2617. Rubin, H. 1985. Cancer as a dynamic developmental disorder. Cancer Res. 45:2935-2942. Ruley, H. E. 1983. Adenovirus early region 1A enables viral and cellular transforming genes to transform primary cells in culture. Nature 304:602- 606. Sanford, K. K., R. Parshad, G. Jones, S. Mandlenan, C. Garrison, and F. Price. 1979. Role of photosensitization and oxygen in chromosome stability and "spontaneous" malignant transformation in culture. J. Nat. Cancer Inst. 63:1245-1255. Sarnow, P., Y. S. Ho, J. Hillians, and A. J. Levine. 1982. Adenovirus Elb- 58kd tumor antigen and SV40 large tumor antigen are physically associated with the same 54 kd cellular protein in transformed cells. Cell 28:387- 394. Schaefer, R., J. Iyer, E. Iten, and A. C. Mirkko. 1988. Partial reversion of the transformed phenotype in HRAS-transfected tumorigenic cells by transfer of a human gene. Proc. Nat. Acad. Sci. USA 85:1590-1594. Scheffner, M., T. Takahashi, J. M. Huibregtse, J. D. Minna, and P. M. Howley. 1992. Interaction of the human papillomavirus type 16 E6 oncoprotein with wild-type and mutant human p53 proteins. J. Virol. 66:5100-5105. Schfitte, J., J. D. Minna, and M. J. Birrer. 1989. Deregulated expression of human c-jun transforms primary rat embryo cells in cooperation with an activated c-Ha-ras gene and transforms Rat-1a cells as a single gene. Proc. Nat. Acad. Sci. USA 86:2257-2261. Seenayer, T. A., and H. K. Cavenee. 1989. Biology of disease molecular mechanisms of oncogenesis. Lab. Invest. 60:585-599. Segatto, D., C. R. King, J. H. Pierce, P. P. Di Fiore, and S. A. Aaronson. 1988. Different structural alterations upregulate in vitro tyrosine kinase activity and transforming potency of the erbB-2 gene. Mol. Cell. Biol. 8:5570-5574. Senn, H.-P., C. Tran-Thang, A. Hodnar-Filipowicz, J. Jiricny, M. Fopp, A. Gratwohl, E. Singer, H. Heber, and Ch. Moroni. 1988. Mutation analysis of the N-ras proto-oncogene in active and remission phase of human acute leukemias. Int. J. Cancer 41:59-64. Seyana, T., A. K. Godwin, M. DePietro, T. S. Hinokur, R. M. Lebovitz and M. H. Liebernan. 1988. In vitro and in vivo regulation of liver epithelial cells carrying a metallothionein-rasT24 fusion gene. Mol. Carcinog. 1:89- 95. 88 Shih, C., Shilo B-Z., M. P. Goldfarb, A. Dannenberg, and R. A. Heinberg. 1979. Passage of phenotypes of chemically transformed cells via transfection of DNA and chromatin. Proc. Nat. Acad. Sci. USA 76:5714-5718. Shilz, R. J. 1988. Ph.D. thesis. Michigan State University, East Lansing. Shinizu, K., M. Goldfarb, M. Perucho. and M. Higler. 1983. Isolation and preliminary characterization of the transforming gene of a human neuroblastoma cell line. Proc. Nat. Acad. Sci. USA 80:383-387. Shindo-Okada, H., D. Makabe, H. Nagahara, and S. Hishimra. 1989. Permanent conversion of mouse'and human cells transformed by activated ras or raf genes to apparently normal cells by treatment with the antibiotic azatyrosine. Mol. Carcinog. 2:159-167. Sidransky, D., T. Tokino, S. R. Hanilton, K. K. Kinzler, B. Levin, P. Frost, and B. Vogelstein. 1992. Identification of ras oncogene mutations in the stool of patients with curable colorectal tumors. Science 256:102- 105. Sistonen, L., J. Keski-Oja, I. Ulnanen, E. Holtta, B-J. Hikgren, and K. Alitalo. 1987. Dose effects of transfected c-Ha-ras‘“1 1’ oncogene in transformed cell clones. Exp. Cell. Res. 168:518-530. Sklar, M. D., E. Thonpson, M. J. Helsh, M. Liebert, J. Harney, H. B. Grossnan, M. Snith, and E. V. Prochownik. 1990. Depletion of c-myc with specific antisense sequences reverses the transformed phenotype in ras oncogene-transformed NIH-3T3 cells. Mol. Cell. Biol. 11:3699-3710. Slit. V. T., A. J. "0 mt. A. "a "n SIltS. 6. Jo Flew. c. Jo Cornelisse, and J. L. Bos. 1988. KRAS codon 12 mutations occur very frequently in pancreatic adenocarcinoma. Nucleic Acids Res. 16:7773-7782. Srivastava, S. K., A. Di Donato, and J. C. Lacal. 1989. H-ras mutants lacking the epitope for the neutralizing monoclonal antibody Y13-259 show decreased biological activity and are deficient in GTPase-activating protein interaction. Mol. Cell. Biol. 9:1779-1783. Stanpfer, M. R., and J. C. Bartley. 1985. Induction of transformation and continuous cell lines from normal human mammary epithelial cells after exposure to benzo[a]pyrene. Proc. Nat. Acad. Sci. USA 82:2394-2398. Stenger, J. E., Mayr G. A., K. Mann, and P. Tegtneyer. 1992. Formation of stable p53 homotetramers and multiples of tetramers. Mol. Carcinog. 5:102- 106. Subranian, T., M. Kuppuswany, and G. Chinnadurai. 1985. An adenovirus 2- coded tumor antigen located on the endoplasmic reticulum and nuclear envelope is required for growth of transformed cells in Ca”-deficient media. Mol. Cell. Biol. 5:3297-3300. 89 Subranian, T., M. Kuppuswany, S. Mak, and G. Chinnadurai. 1984. Adenovirus cyt‘ locus, which controls cells transformation and tumorigenicity, is an allele of lpf locus, which codes for a 19-kilodalton tumor antigen. J. Virol. 52:336-343. Suzuki, V., M. Orita, M. Shiraishi, K. Hayashi, and T. Sekiya. 1990. Detection of ras gene mutations in human lung cancers by single-strand conformation polymorphism analysis of polymerase chain reaction products. Oncogene 5:1037-1043. Sweet, R. H., S. Vokoyana, T. Kanata, J. R. Feranisco, M. Rosetterg, and M. Gross. 1984. The product of ras is a GTPase and the T24 oncogene mutant is deficient in this activity. Nature 311:273-275. Takeya, T., and M. Hanafusa. 1983. Structure and sequence of the cellular gene homologous to the RSV src gene and the mechanism for generating the transforming virus. Cell 32:881-890. Taparowsky, E., K. Shinizu, M. Goldfarb, and M. Higler. 1983. Structure and activation of the human N-ras gene. Cell 34:581-586. Trahey, M., and F. McCornick. 1987. A cytoplasmic protein stimulates gggmal N-ras p21 GTPase, but does not affect oncogenic mutants. Science :542-545. Valencia, A., P. Chardin, A. Hittinghofer, and C. Sander. 1991. The ras protein family: evolutionary tree and role of conserved amino acids. Biochemistry 30:4637-4647. Verna, A. K., E. A. Conrad, and R. K. Boutwell. 1982. Differential effects of retinoic acid and 7,8-benzoflavone on the induction of mouse skin tumors by the complete carcinogenesis process and by the initiation- promotion regimen. Cancer Res. 42:3519-3525. Vogel, J., S. H. Hinrichs, R. K. Reynolds, P. A. Luciw, and G. Jay. 1988. The HIV tat gene induces dermal lesions resembling Kaposi’s sarcoma in transgenic mice. Nature 335:606-611. Vogelstein, B, E. R. Rearon, S. R. Hanilton, S. E. Kern, A. C. Preisinger, M. Leppert, Y. Nakalira, R. Hhite, A. M. M. Snits, and J. L. Bos. 1988. Giesetic alterations during colorectal-tumor development. N. Engl. J. Med. 3 :525-537. Vousden, K. H., J. L. Bos, C. J. Marshall, and D. M. Phillips. 1986. Mutations activating human c-Ha-rasl protooncogene (HRASI) induced by chemical carcinogens and depurination. Proc. Nat. Acad. USA 83:1222-1226. Hang, F., C. 0. Gregory, M. Rowe, A. B. Rickinson, D. Hang, M. Birkenbach, H. Kikutani, T. Kishinoto, and E. Kieff. 1987. Epstein-Barr virus nuclear antigen 2 specifically induces expression of the B-cell activation antigen C023. Proc. Nat. Acad. Sci. USA 84:3452-3456. 90 Hang, D., D. Liebowitz, and E. Kieff. 1985. An EBV membrane protein expressed in immortalized lymphocytes transforms established rodent cells. Cell 43:831-840. Heinstein, I. B. 1991. Nonmutagenic mechanisms in carcinogenesis: role of protein kinase C in signal transduction and growth control. Environ. Health Perspect. 93:175-179. Heissnan, B. E., M. J. Arboleda, K. A. Kinoto, M. J. MacDonald, R. Ochiai, and C. J. Der. 1989. Activation of ras oncogenes in chemically transformed BALB/MK-2 mouse keratinocytes. Mol. Carcinog. 2:150-158. Hells, A., and J. M. Bishop. 1988. Genetic determinants of neoplastic transformation by the retroviral oncogene v-erbB. Proc. Nat. Acad. Sci. USA 85:7597-7601. Herness, B. A., A. J. Levine, and P. M. Howley. 1990. Association of human papillomavirus types 16 and 18 E6 proteins with p53. Science 248:76-79. thte, P., K. J. Buchkovich, J. M. Horowitz, S. H. Friend, M. Raybuck, R. A. Heinberg, and E. Harlow. 1988. Association between an oncogene and anti-oncogene: the adenovirus EIA proteins bind to the retinoblastoma gene product. Nature 334:124-129. Hierenga, R. K., and H. G. J. Hol. 1983. Predicted nucleotide-binding properties of p21 protein and its cancer-associated variant. Nature 302:842-844. Hillinghan, M. C., I. Pastan , T. Y. Shih, and E. M. Scolnick. 1980. Localization of the src gene product of the Harvey strain of MSV to plasma membrane of transformed cells by electron microscopic immunocytochemistry. Cell 19:1005-1014. Hilson, D. M., D. Yang, J. E. Dillberger, S. E. Dietrich, V. M. Maher, J. J. McCornick. 1990. Malignant transformation of human fibroblasts by a transfected N-ras oncogene. Cancer Res. 50:5587-5593. You, H., U. Candrian, R. R. Maronpot, G. D. Stoner, and M. H. Anderson. 1989. Activation of the Ki-ras protooncogene in spontaneously occurring and chemically induced lung tumors of the strain A mouse. Proc. Nat. Acad. Sci. USA 86:3070-3074. Yunis, J. J., and N. Ransay. 1978. Retinoblastoma and subband deletion of chromosome 13. Am. J. Dis. Child. 132:161-163. Zhang, K., M. Noda, H. C. Vass, A. G. Papageorge, and D. R. Louy. 1990. Identification of small clusters of divergent amino acids that mediate the opposing effects of ras and Krev-l. Science 249:162-165. zur Hausen, H. 1987. Papillomaviruses in human cancer. Cancer 59:1692- 1696. CHAPTER II Sequencing PCR-anplified ras Genes to Deternine their Inportance to the Malignant Transfornation of Hunan Fibroblasts Jeanette M. Scheid’, Veronica M. Maher“, and J. Justin McCormick1 ‘Carcinogenesis Laboratory, Fee Hall Department of Microbiology and Department of Biochemistry Michigan State University, East Lansing, MI 48824-1316 91 Activation of ras proto-oncogenes by mutation at specific codons has been documented in a variety of human malignancies. To determine the extent to which ras activation plays a role in the malignant transformation of human fibroblasts, we sequenced the coding region of the H-, K-, and N-ras genes from a series of human fibroblast cell lines, including normal diploid human fibroblasts, two morphologically normal non-tumorigenic infinite life span cell strains originated in this laboratory, spontaneous and carcinogen-induced malignant transformants of the latter, and cell lines derived from human fibrosarcomas. Direct sequencing of the PCR product was carried out using the dideoxy chain termination method. .A total of twelve cell lines or strains were analyzed for ras mutations. Of the eight that are malignant, two fibrosarcoma- derived cell lines contained an activating ras mutation. One mutant cell line, 8387, has a G to C transversion that causes an Ala to Pro amino acid substitution at codon 146 in one allele of the Ki-ras gene. This same mutation has been reported to activate Ki-ras in Hut-14 cells, an 8387 derivative. The other mutant cell line, HT1080, has a C to A transversion that causes a Gln to Lys amino acid substitution at codon 61 in one allele of the N-ras gene that has been reported previously. The remaining cell lines or strains contain no mutations in ras. The mutation frequency in the fibroblastic cells tested in this study is consistent with one other ° report on ras activation in human sarcomas in the literatureu Our results 92 93 show that although ras oncogenes can play a role in the development of fibrosarcomas in humans, some other changes must be able to substitute for ras activation in the malignant transformation of human fibroblasts. Also, methods of detecting ras activation that concentrate on the "classic" activating codons may'miss activated genes. Further studies are planned to test for alterations elsewhere in the ras p21 signalling pathway using cells that contain only wild-type ras genes. INTRODUCTION It is well established that malignant transformation of human cells occurs by a multi-step process in which an individual cell accumulates a number of stable heritable changes that together result in the acquisition of characteristics such as immortality, autonomous growth, and metastatic capacity. The steps in this process have been linked to alterations in a number of genes. Among these, the ras family of proto-oncogenes, the first of which was discovered as the cellular homolog of the transforming gene of the Harvey sarcoma virus [1], has been widely studied. The three major ras genes, Ha-, Ki-, and N-ras, encode 21kDa proteins that associate with the internal surface of the cell membrane, bind guanine nucleotides, and have an intrinsic GTPase activity. Studies have shown that ras p21 cycles between its GTP-bound ("on" state) and GDP-bound ("off" state) forms (reviewed by [2]); however, the steady-state ratio of GTP- and GDP- bound p21 is greater in actively dividing cells than in quiescent cells. This observation, along with data showing the requirement for ras expression in serum-stimulated cell division [3], supports the hypothesis that wild-type ras p21 is involved in transduction of mitogenic signals from protein growth factors (reviewed by [4]). Hhen mutated at specific codons (most commonly at 12, 13, 59, 60, 61, and 117 [5]) ras proto-oncogenes become activated oncogenes. Several lines of evidence point to a role for activated ras oncogenes in malignant transformation. First, mutated ras genes have been shown to be present in 94 95 a number of malignant human cells [6]. Second, ras activation has been associated with induction of malignant transformation of animal cells in culture by chemical carcinogens [7] and radiation [8]. Third, the introduction of activated ras genes into cells in culture has been shown to induce transformation in animal [9] and human [10] model systems. Finally, loss of expression of activated ras genes has been shown to result in reversion to a less-transformed phenotype [11-13]. Cultured human and animal fibroblasts are commonly used as a model system in which to study the process of malignant transformation, primarily because the relative ease of propagation has allowed their growth and phenotypic properties to be well defined. In this and other laboratories, ras oncogenes have been established as transforming agents when transfected into normal diploid human fibroblasts [14] and immortal human fibroblasts [15-17]. However, there is a paucity of data on the role of endogenous ras genes in malignant transformation of human fibroblasts. Studies have implicated ras activation in some human mesenchymal cell malignancy (e.g. myeloid disorders [6]), but in one study on a series of sarcomas, no ras mutations were found [6]. Hith the development of more sophisticated methods for detection of ras activation, data on the incidence of ras activation in many cell types has changed; either showing an increased incidence, as in ovarian carcinoma [18, 19] or a decreased incidence, as in bladder cancer [20]. It is reasonable to hypothesize that the incidence of ras activation in sarcoma cells is also different than reported previously. To gain a clearer understanding of the role of ras genes in the malignant transformation in human fibroblasts, we have searched for ras activation in two series of human fibroblast-derived cell lines. The 96 first series is comprised of established cell lines from human fibrosarcomas (Table 1). By working with a pure population of cultured cells one avoids the of problem separating normal and malignant cells. Although one might be concerned about the possibility that long-term propagation in culture has resulted in selection artifacts, these cell lines have remained relatively stable in terms of chromosome content and phenotype. Furthermore, as all but one continue to be tumorigenic, they most likely contain the changes that originally resulted in their transformation. Table 1. Cell lines derived from human fibrosarcomas Cell line Origin Original_3e£erenee__ SL84 neonatal foreskin' [21] HT1080 fibrosarcoma’ [22] SH982 fibrosarcoma‘ [23] 8387 fibrosarcoma [23] NCI fibrosarcoma [23] VIP:FT spontaneous malignant [24] transformant Table 1. Fibrosarcoma-derived cell lines used for determination of ras mutation. 'Normal cell line used as a negative control. *Known to carry an activating N-ras mutation [25]. * Originally derived from a human fibrosarcoma, but no longer tumorigenic. 97 The second series (listed in Table 2) is comprised of a number of cell strains derived from an immortal cell strain generated in this laboratory [26]. This series includes the immortal near-diploid non- malignant parent human fibroblast strain, and several malignant transformants derived from that immortal strain that have occurred either spontaneously or as a result of treatment with chemical carcinogens. The origin and characteristics of the individual strains used in this study are outlined in Figure 1. ‘The specific cell strains from the lineage that were studied were chosen because they possess characteristics in culture that are similar to those of cells transformed by transfection of activated ras, because of evidence from immunoprecipitation of p21 that their ras genes are altered (McCormick, unpublished observation), or because they exhibit changes in growth factor requirements that are hypothesized to result from ras activation. Since this series forms a lineage, it could also allow us to determine whether ras activation is an early or late event in malignant transformation. He examined these various cell strains for ras activation by determining whether or not mutations were present in the coding region of the Ha-, Ki-, and N-ras genes. Direct sequencing of PCR-amplified cDNA was performed, since sequencing provides information on the location and kinds of mutations present in a cell, and also indicates whether one or both alleles are altered. He considered it important to determine the sequence of the entire coding region, particularly in view of recent developments in describing the complexity of ras-mediated 98 Table 2. Cell strains from the MSU-1 lineage Tumor Unusual flame Treatment Selection Morphology Features MSU-1.1 v-myc life span NA' -2 marker chromosomes’ -slight growth advantage MSU-1.2 low Ca”’ clonal growth NA novel phospho- protein (PDGFRi?) CSVO.20/ BPDE' soft agar spindle STI cell L55I-3T MNNG' -focus formation round altered p21 -soft agar cell immuno- precipitation pattern SDS32AC ENU‘ morphology round altered p21 4MT/T2 cell immuno- precipitation pattern DY6:16. BPDE focus formation giant 3CIOC/T2 cell Table 2. Cell strains from the>MSU-I lineage used for determination of ras mutations. ' NA, non-applicable since this strain does not form tumors. ' This parental strain in the series has an infinite life span in culture, contains two marker chromosomes, and grows slowly in serum-free medium, but in all other respects it is similar to normal human fibroblasts. * PDGFR, platelet-derived growth factor receptor. ' BPDE, (t)-7858a— dihydroxy-9o,10a-epoxy-7,8,9,10-tetrahydrobenzola]pyrene (a reactive metabolite of benzo[a]pyrene). ' MNNG, N-methyl-N’-nitro-N- nitrosoguanidine. * ENU, ethylnitrosourea. 99 Figure 1. MSU-1 Lineage Members of the MSU-1 lineage of cell strains used in this study. In addition to the cell strains shown here, there are several spontaneous, oncogene-transformed and carcinogen-induced malignant transformants of MSU-1.1. See text and Table 2 for more complete descriptions and references pertaining to the specific strains used in this study. 100 2.32: . :3 22...... momma: FquS— ._. 930E Phw\ou.o>w0 32.283955... 2o:&oo.oo::EE_ MDQm 5n no.3? an oo.2.o cocoocooouc. .203 5320 .833 2.39.: 2.32: 2.32: «not: 232532: :00 253 :00 2.59. :oo 0:39. «.7392 thoowomdto>o hon—mm.— whxh2vo._m> ._O BEES .Eo—BUV QTDw—Z 223823.: 23.-.,“ 3.8.85: 5.2.... 22a... .952. 70.. . 101 signalling pathways [27]. The so-called classic activating codons have been determined by their ability to transform NIH-3T3 cells to focus formation. It is possible that mutations at other locations, for instance in the effector domain, are also important, but these would not be detected by methods concentrating only on the codons that have previously been reported to cause activation. Our results indicate that ras activation is not required for the malignant transformation of human fibroblasts. Further studies are needed to determine what changes substitute for ras activation in these cells. MATERIALS AND lETlKIIS Routine Cell Culture Cells were grown routinely in 5% CO, humidified air at 37°C on plastic in Eagle’s minimal essential medium (Gibco, Gaithersburg, MD) supplemented with 0.2 mil aspartic acid, 0.2 mM serine, 1.0 mM Na pyruvate, and with 10% supplemented calf serum (HyClone, Logan, UT), penicillin (100 U/ml)-streptomycin (100 ug/ml), and hydrocortisone (10 ug/ml). RNA Extraction Total RNA was extracted by the acid guanidinium thiocyanate-phenol- chloroform method [28]. Briefly, cells in log-phase growth were rinsed with ice-cold PBS. The denaturing solution was added (4 M guanidinium thiocyanate, 25 mM Na citrate pH 7.0, 0.5% Na lauryl sarcosine, and 0.1 M 2-mercaptoethanol) and the cell lysate was scraped into a centrifuge tube. To the lysate, 2 M Na acetate pH 4.0, water-saturated phenol, and chloroformzisoamyl alcohol (24:1) were added in succession, with mixing at 102 each step. The suspension was centrifuged for 10 min. at 5,000 rpm. The aqueous top layer containing the extracted RNA was removed and precipitated with an equal volume of isopropanol at -20°C for at least one hour (typically overnight). The RNA was centrifuged for 20 min. at 5,000 rpm to form a pellet, washed with 75% ethanol and resuspended in water. The amount of RNA was determined from the 00 um“. RNA was distributed into 80-100 ug aliquots, vacuum-dried and stored at -20°C until use. Preparation of first strand cDNA Total RNA (20 ug) was dissolved in 5 ul cDNA cocktail (50 mM Tris- HCl pH 8.3, 75 mM KCl, 10 mM DTT, 3 mM MgCl,, 0.5 mM each of dATP, dCTP, dGTP, dTTP (Boehringer Mannheim Biochemicals, Indianapolis, IN), 100 ng oligo dT primer (Pharmacia, Piscataway, NJ), 15 U RNAse inhibitor (Promega, Madison, HI), and 200 U M-MLV reverse transcriptase (Bethesda Research Laboratories, Gaithersburg, MD)) and incubated at 37°C for 1 hr. The samples were either amplified immediately or stored at -20°C for up to one week. Second strand cDNA synthesis and in vitro mlification using the Polynerase Chain Reaction (PCR) PCR amplification of the cDNA was carried out using a modification of the method of Saiki et al. [29]. 45 ul of PCR cocktail (10 mM Tris-HCl pH 8.3, 50 mM KCl, 2.5 mM MgCl,, 0.001% gelatin, 0.4 mM each of dATP, dGTP, dCTP, dTTP (Boehringer Mannheim Biochemicals, Indianapolis, IN), 50 ng each of 5' and 3’ PCR primers specific to the flanking sequences of the coding region of Ha-, Ki-, or N-ras (Fig. 2), and 2.5 U Taq polymerase (Perkin-Elmer Cetus Corp., Norwalk, CT)) was added to 5 ul of single- 103 stranded cDNA preparation. Cycling was performed in a Perkin-Elmer Cetus (Norwalk, CT) thermal cycler as follows: initial denaturation 94°C for 5 min. followed by 30 cycles amplification (94°C 1 min., 60°C 1 min., 72°C 2 min.) and finished with 72°C for 7 min. to complete extension. After the first 30 cycles, unincorporated nucleotides and primers were removed using Centricon-30 microconcentrators (Amicon, Danvers, MA). 10% of the concentrated sample (5 ul) was amplified a second 30 cycles under the same conditions using a new set of primers internal to those used for the first 30 cycles. After the second round of 30 cycles, the unused primers and nucleotides were again removed in a Centricon 30 microconcentrator (Amicon, Danvers, MA) and stored at -20°C prior to direct sequencing. Amplification was verified by agarose mini-gel and ethidium bromide staining. Direct sequencing of PCR-anlified cDNA Sequencing of amplified cDNA was carried out using a modification of the dideoxy chain termination method [33]. 100 ng sequencing primers (Fig. 2) were end-labeled with 20 U T4 polynucleotide kinase (New England Biolabs, Beverly, MA) and 100 uCi y-“P-dATP (Dupont New England Nuclear, Boston, MA) according to United States Biochemical Corporation (USB, Cleveland, OH) protocols for Sequenase version 2.0. Approximately 100 ng of amplified double-stranded DNA was annealed to 2 pmol end-labeled primer and reactions carried out with 3.25 U Sequenase version 2.0. Samples were separated on 8% polyacrylamide, 7 M urea gels. Gels were exposed to autoradiograph film (Kodak, Rochester, NY) for 12-36 hours. Sequences were compared to published sequences of Ha-, Ki-, or N-ras. Mutations were verified by repeating sequence determination with a different primer. 104 Figure 2: Oligonucleotide primers used for PCR and sequencing. For concentrations used see~Materials and Methods. Numbers in parentheses refer to the location of that primer in the gene or transcript: for Ha-ras [30]; for Ki-ras [31]; for N-ras [32]. Primer KSEQZ spans the exon 1-exon 2 splice site. Underlined bases indicate where the sequence of the primer differs from the published sequence. These changes were made to have a BamHI recognition site. Ha-ras: Ki-ras: H'T‘OS 2 Figure 2. 105 PCB_£r1me::. HPR1 (forward) 5'-6CCCGAATTQGCAGGTGGGGCAGGA-3’ (1603) (1627) HPRZ (reverse) » 5'-AAGGGAI§CTGCTGACCGCAGGCCA-3' (3457) (3433) HPR3 (forward, nested) 5’-GGAGGACCCCGGGCCGCAGG-3’ (1637) (1656) HPR4 (reverse, nested) 5’-CCGGCAGGGGCGGGGAGCCG-3' (3427) (3408) KPR1 (forward) 5'-GTACTGGTGGAGTATTTGAT-3' (6282) (6301) KPRZ (reverse) 5’-ACAGGCATTGCTAGTTCAAA-3' (37501) (37482) KPR3 (forward, nested) 5’-AGTGTATTAACCTTATGTGT-3’ (6302) (6321) KPR4 (reverse, nested) 5'-CTGGGAATACTGGCACTTCG-3’ (21352) (21364) NPRS (forward) 5'-ACTCGTGGTTCGGAGGCCCA-3’ (-205) (-186) NPR6 (reverse) 5’-TTGTACTAAACTACTGAGAG-3’ (724) (705) NPRI (forward, nested) 5’-AGGCGCCTGGCAGCCGACTG-3' (-165) (-146) NPR2 (reverse, nested) 5'-AGCAGGAGCTTCTCTGTGAG-3' (694) (685) NPR3 (forward, nested) 5'-CCGGTCTGTGGTCCTAAATC-3’ (~65) (-46) NPR4 (reverse, nested) 5’-TACTTCTCCTCCAGGAAGTC-3' (654) (635) NPR7 (forward, nested) 5’-TGTCCAAAGCAGAGGCAGTG-3' {-45) {-26) NPR8 (reverse, nested) 5’-AGGACCAGGGTGTCAGTGCA-3’ (634) (615) Senuencinn_2:1mers_______ HSEOI (forward) 5'-TGCCTGTTGGACA-3' (2087) (2099) HSEOZ (forward) 5’-GAACAAGTGTGAC-3’ (2435) (2447) HSE03 (reverse) 5'-CTCAAAAGACTTG-3' (2209) (2197) KSEOl (reverse) 5'-TCGAGAATATCCA-3’ (19528) (19516) KSEOZ (forward) 5’-TAGAG/GATTCCTA-3’ (6483) (19476) KSEO3 (forward) 5’ -AGAGTTAAGGACT-3’ (21352) (21364) NSEOI (reverse) 5'-CCTGTCCTCATGT-3' (223) (211) NSEOZ (forward) 5'-TGGACATACTGGA-3’ (257) (269) NSEO3 (forward) 5'-AAGTGTGATTTGC-3' (349) (361) NSEO4 (reverse) 5'-ATACACAGAGGAA-3' (243) (231) 106 Cycle sequencing: The 5’ 200 base pairs of the N-ras cDNA was sequenced using the "fmol" sequencing kit (Promega, Madison, HI) according to the manufacturers instructions and with 32P end-labeled primers (Dupont NEN, Boston, MA). Specific cycling conditions varied with the length and GC content of the various primers. RESULTS Deternination of the status of the ras genes in cells fron a series of fibrosarcona cell lines: He sequenced the entire coding region of the Ha-, Ki-, and N-ras genes of these cell lines. Two mutations were seen in this series of cell lines. The first, a mutation in N-ras codon 61 of HT1080 cells, was reported previously and is known to cause activation [25]. The second, a mutation in K-ras codon 146 of 8387 (Figure 3), has never been reported in this cell line, or at this location. However, a literature search performed after this mutation was discovered revealed that Hut-14 cells, which are derived from the 8387 cell line [34], contain the same mutation. The gene containing this mutation does not induce focus formation in NIH- 3T3 cells, but efficiently transforms an immortal hamster cell line [35]. No other amino acid-altering mutations in any of the three ras genes were found among the fibrosarcoma cell lines examined. 107 Deternination of the status of the ras genes in cells fron the MSU 1.1 lineage: He also determined the sequences of the coding regions for the Ha-, Ki-, and N-ras genes in the cell strains derived from the MSU 1 lineage. No mutations that alter the amino acids encoded by these genes were observed. The overall frequency of ras mutations observed in these two series of cell strains or lines was 2/12 (Table 3). The frequency of ras mutations in the cell lines or strains that are capable of forming malignant tumors is 2/8 (Table 3). Therefore, although activation of ras by point mutation can be observed in the malignant transformation of human fibroblasts, some other pathway to malignancy must exist. Table 3. Summary of ras gene mutations Mutations/cell line Qell_group Ha-ras Ki-nas N-gas Fibrosarcoma- 0/5 1/5 1/5 derived (8387 codon 146 (HT1080 codon 61 GCA>CCA CCA>CAA Ala>Pro) Gln>Lys) MSU-1 lineage 0/7 0/7 0/7 Total malignant 0/8 1/8 1/8 cell lines/strains Total cell 0/12 1/12 1/12 lines/strains 108 (3(3.A T K'LQS codon 146 8387 Figure 3. Point mutation in the Ki-ras gene of 8387. This photograph shows the G to C transversion at codon 146 in one allele of the Ki-ras gene seen in cells from the 8387 cell line. 109 DISCUSSION This study was undertaken to determine the extent to which the malignant transformation of human fibroblasts in vivo and in vitro is associated with the mutational activation of a ras gene. Sequencing the coding regions of the Ha-, Ki-, and N-ras genes from a total of eight malignant cell strains or cell lines derived from fibrosarcomas revealed only two mutations, one in N-ras and one in Ki-ras. Mutations were found only in fibrosarcoma-derived cell lines. It is somewhat surprising that none of the malignant transformants within the MSU-1 lineage contained ras mutations, especially given the characteristics of these cells. Two of the malignant cell strains are known to have similarly altered p21 immunoprecipitation patterns (unpublished data). A third malignant cell strain forms giant cell tumors very similar to those formed by'a malignant derivative of MSU 1.1 that was transformed by an over-expressed and activated N-ras gene [16] (unpublished observations). A fourth malignant strain has a typical spindle cell morphology, but its non-tumorigenic precursor has been shown to contain two unique species of phosphoproteins [36]. One of these phosphoproteins is considered to be the platelet- derived growth factor receptor, which is known to stimulate ras-mediated signal transduction [37]. The frequency of ras mutations seen in our study is similar to that reported in the few studies on ras activation and fibroblast transformation. For example, Tahira et al. [38], found Ki-ras mutations in cells from one of a total of seven fibrosarcomas from dinitropyrene- treated rats. Similarly, Hilke et al. [39] reported Ha-ras mutations in the cells from four of twelve human sarcoma biopsies. These studies show 110 that there is an association between ras activation and malignant transformation of fibroblasts. These correlative data, taken with the numerous studies showing that insertion of activated ras transforms fibroblasts in culture, lend support to the hypothesis that ras genes can play a role in the] malignant transformation of human fibroblasts. However, some other change(s) must also play a role since the majority of fibrosarcomas that were tested did not contain mutated ras genes. It should be pointed out that one of the mutations that we identified was located in a codon that is not one of the "classic" codons typically tested. This shows the importance of determining the sequence of the entire coding region of these genes. This is, of course, a labor- intensive task. However, it is now possible to use single stranded conformation polymorphism to rapidly identify regions in genes where point mutations are located. The advent of such rapid, sensitive techniques makes it feasible to screen the coding region of the three ras genes from a large number of cell strains for exons that contain such a mutation and then determine the sequence only of these regions. A number of investigators have attempted, with varying degrees of success, to establish a cause-effect relationship between ras activation and malignant transformation. One study [40] showed an increased frequency of N-ras mutations in cells from malignant skin tumors from xeroderma pigmentosum (XP) patients compared to cells from malignant skin tumors from normal patients. XP patients have an inherited defect in the ability to repair DNA damage caused by ultraviolet light and are prone to a high frequency of skin tumors. The N-ras mutations found in the cells used in that study were consistent with those typically seen in cells exposed to UV, lending support to a cause-effect relationship between UV 111 damage and malignant transformation. This cause-effect relationship was called into question by Nakazawa et al.[41]. These investigators tested cells from DMBA-induced BALB/c3T3 foci for an A to T transversion in Ha- ras at codon 61 that is commonly found in cells from DMBA-induced tumors. None of the 30 independently cloned focus-forming cell strains had that particular mutation. However, these investigators showed that DMBA can induce A to T mutations in the Ha-ras genes of BALB/c3T3 cells at a frequency of 1 in 10‘. These data support these investigors’ hypothesis that whether or not a carcinogen-induced mutation contributes to malignant transformation depends on the type of cell in which the mutation occurs. Devereux et al. [42] compared sensitive and resistant strains of mice for the concentration of DNA adducts in lung cells when these mice were treated with two different carcinogens, for the frequency of lung tumors induced by ‘these two carcinogens, and for the 'frequency' of Ki-ras activation in the cells from these carcinogen-induced lung tumors. These investigators found large differences both in the frequency of induced tumors and in the frequency of Ki-ras mutations between the two strains of mice that could not be explained by the small differences in carcinogen- induced adducts between the two strains. This finding also supports the hypothesis that transformation is affected by factors in addition to carcinogen exposure. Thus, the relationship between carcinogen-induced DNA damage and malignant transformation is complicated by such factors as the ability of the carcinogen to damage DNA, the type of damage done to the target gene, the ability of the cell to repair the damage to this gene before it can be converted into a mutation, and the susceptibility of the cell or tissue to transformation as a result of a mutation in that target gene. 112 A major difficulty faced by investigators trying to establish a cause-effect relationship between carcinogen-induced ras activation and induction of tumors in animals is the relatively high frequency of spontaneous tumors in animal studies [43, 44]. Hhen the frequency of carcinogen-induced tumors is only 2-3 fold above background, interpretation of data is difficult because of the high chance of analyzing spontaneous tumors and attributing the results to carcinogen treatment. An example of spontaneous mutagenesis in the Ha-ras gene is that of Hoffmann et al. [45] who showed that mutations at codons 12 and 13 can result when DNA synthesis by DNA polymerase alpha is interrupted by secondary structure of the DNA at that site. Our finding, and that of others, that activating mutations in ras genes occur in some malignant human fibroblasts, leads to speculation about the role, if any, that activated ras genes play in the observed malignant transformation. Mutational activation of ras genes has been implicated in the acquisition of metastatic capacity by rodent fibroblasts. One study [46] identified an association between expression of activated Ha-ras in cultured rat embryo fibroblasts, altered expression of the surface marker CD44 in those cells, and the induction of metastasis by the cells that expressed the altered marker. Studies in our laboratory with human fibrosarcomas also support this association. A series of cell lines or strains that over-express mutated ras and form tumors were shown to possess increased secreted and receptor-bound plasminogen activator activity [47] and to exhibit metastatic activity when injected into the tail vein of athymic mice [48]. However, cell strains that did not contain an activated ras gene and/or did not over-express ras (e.g. HT1080) were found among the cell strains that possess increased 113 plasminogen activator activity and form metastatic tumors after tail vein injection. Therefore, over-expression of activated ras genes is associated with metastasis; however, other changes must also be able to confer this capacity in some cells. The molecular mechanisms by which ras mutation mediates transformation have been the subject of research for many years. Early studies showed that certain nuclear transcriptional factors were required for ras-induced transformation [49], indicating the existence'of a pathway connecting ras p21 and alterations in gene expression. Recent work has shown that ras p21 cooperates with the product of the src gene to transform hamster fibroblasts [50] and human colon epithelial cells [51]. Still more recently research has identified the components of a signal transduction pathway involving ras [52-56]. Future investigations should provide insight into the nature of the interaction between these various proteins. In the cell lines and strains that we showed in the present study did not contain mutations of the ras genes, but were malignant, various changes might have occurred. For instance, it has been shown in our laboratory that growth factor expression is altered in several of the fibrosarcoma-derived cell lines [23]. It is also possible that oncogenes unrelated to ras have been activated. However, since exogenous activated ras genes are efficient transforming agents in fibroblasts, it is possible that the cell lines and strains with only'wild-type ras genes have changes in other components of the ras-mediated signal transduction pathway that are substituting for mutational activation of ras. Studies measuring the ratio of GTP-bound and GDP-bound ras p21 are planned, because this measurement will likely provide information that will aid in determining 114 which, if any, of the ras p21-mediated signal transduction pathway components is activated in these cells. 10. 11. REFERENCES . Parada LF, Tabin CJ, Shih C, Heinberg RA. Human EJ bladder carcinoma oncogene is homologue of Harvey sarcoma virus ras gene. Nature 297:474-478, 1982. . Haubruck H, McCormick F. Ras p21: effects and regulation. Biochim Biophys Acta 1072:215-229, 1991. . Mulcahy LS, Smith MR, Stacey DH. Requirement for ras proto- oncogene function during serum-stimulated growth of NIH-3T3 cells. Nature 313:241-243, 1985. . Santos E, Nebreda AR. Structural and functional properties of ras proteins. FASEB J 3:2151-2163, 1989. . Barbacid M. ras genes. Annu Rev Biochem 56:779-827, 1987. . Bos JL. ras oncogenes in human cancer: a review. Cancer Res 49:4682-4689, 1989. . Hang Y, Hang Y, Stoner G, You M. ras mutations in 2- acetylaminofluorene-induced lung and liver tumors from C3H/HeJ and (C3HxA/J)Fl mice. Cancer Res 53:1620-1624, 1993. . Chen CH, Endlich B, Salavati R, Ling CC. Presence of point mutations in the N-ras gene in radiation-transformed rat embryo cells. Cancer Res 53:1511-1515, 1993. . Greenhalgh DA, Rothnagel JA, Quintanilla MI, et al. Induction of epidermal hyperplasia, hyperkeratosis, and papillomas in transgenic mice by a targeted v-Ha-ras oncogene. Mol Carcinog 7:99-110, 1993. Basolo F, Elliot J, Tait L, et al. Transformation of human breast epithelial cells by c-Ha-ras oncogene. Mol Carcinog 4:25-35, 1991. Mukhopadhyay T, Tainsky M, Cavender AC, Roth JA. Specific inhibition of K-ras expression and tumorigenicity of lung cancer cells by antisense RNA. Cancer Res 51:1744-1748, 1991. 115 an... 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 116 Georges RN, Mukhopadhyay T, Zhang Y, Yen N, Roth JA. Prevention of orthotopic human lung cancer growth by intratracheal instillation of a retroviral antisense K-ras construct. Cancer Res 53:1743- 1746, 1993. Paterson H, Reeves 8, Brown R, et al. Activated N-ras controls the transformed phenotype of HT1080 human fibrosarcoma cells. Cell 51:803-812, 1987. Hurlin PJ, Fry 0G, Maher VM, McCormick JJ. Morphological transformation, focus formation, and anchorage independence induced in diploid human fibroblasts by expression of a transfected H-ras oncogene. Cancer Res 47:5752-5757, 1987. Hurlin PJ, Maher VM, McCormick JJ. Malignant transformation of human fibroblasts caused by expression of a transfected T24 HRAS oncogene. Proc Natl Acad Sci USA 86:187-191, 1989. Hilson DM, Yang 0, Dillberger JE, Dietrich SE, Maher VM, McCormick JJ. Malignant transformation of human fibroblasts by a transfected N-ras oncogene. Cancer Res 50:5587-5593, 1990. Fry 0G, Milam LD, Dillberger JE, Maher VM, McCormick JJ. Malignant transformation of an infinite life span human fibroblast cell strain by transfection with v-Ki-ras. Oncogene 5:1415-1418, 1987. Mok SC, Bell DA, Knapp RC, et al. Mutation of K-ras protooncogene in human ovarian epithelial tumors of borderline malignancy. Cancer Res 53:1489-1492, 1993. Teneriello MG, Ebina M, Linnoila RI, et al. p53 and Ki-ras gene mutations in epithelial ovarian neoplasms. Cancer Res 53:3103- 3108, 1993. Knowles MA, Hilliamson M. Mutation of H-ras is infrequent in bladder cancer: conformation by single-strand conformation polymorphism analysis, designed restriction fragment length polymorphism analysis, and direct sequencing. Cancer Res 53:1330139, 1993. McCormick JJ, Maher VM. Measurement of colony-forming ability and mutagenesis in diploid human cells. In: Friedberg EC and Hanawalt PA (eds), DNA Repair A Laboratory Manual of Research Procedures, Vol 1, Marcel Dekker, New York, 1981, pp. 501-521. Rasheed S, Nelson-Rees HA, Toth EM, Arnstein P, Gardner MB. Characterization of a newly-derived human sarcoma cell line (HT1080). Cancer 33:1027-1032, 1974. Schilz R. Ph.D. Dissertation. Michigan State University. East Lansing, 1989. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 117 Mukherji B, MacAlister TJ, Guha A, Gillies CG, Jeffers DC, Slocum SK. Spontaneous in vitro transformation of human fibroblasts. J Natl Cancer Inst 73:583-587. 1984. Brown R, Marshall CJ, Pennie 56, Hall A. Mechanism of activation of an N-ras gene in the human cell fibrosarcoma cell line HT1080. Embo J 3:1321-1326, 1984. Morgan TL, Yang 0, Fry DG, et al. Characteristics of an infinite life span diploid human fibroblast cell strain and a near-diploid strain arising from a clone of cells expressing a transfected v- myc oncogene. Exp Cell Res 197:125-136, 1991. Marx J. Forging a path to the nucleus (Research News). Science 260:1588-1590, 1993. Chomczynski P., Sacchi N. Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal Biochem 162:156-159, 1987. Saiki RK, Gelfand DH, Stoffel S, et al. Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science 239:487-491, 1988. Reddy EP. Nucleotide sequence analysis of the T24 human bladder carcinoma oncogene. Science 220:1061-1063, 1983. McGrath JP, Capon DJ, Smith DH, et al. Structure and organization of the human Ki-ras proto-oncogene and a related processed pseudogene. Nature 304:501-506, 1983. Hall A, Brown R. Human N-ras: cDNA cloning and structure. Nucleic Acids Res 13:5255-5268, 1985. Sanger F, Nicklen S, Coulson AR. DNA sequencing with chain- terminating inhibitors. Proc Natl Acad Sci USA 74:5463-5467, 1977. McCormick JJ, Yang 0, Maher VM, et al. The HuT series of 'carcinogen-transformed’ human fibroblast cell lines are derived from the human fibrosarcoma cell line 8387. Carcinogenesis 9:2073- 2079, 1988. Higashi T, Sasai H, Suzuki F, et al. Hamster cell line suitable for transfection assay of transforming genes. Proc Natl Acad Sci USA 87:2409-2413, 1990. Pan C-H. M. S. Thesis. Michigan State University. East Lansing, 1990. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 118 Gibbs J8, Marshall MS, Scolnick EM, Dixon RAF, Vogel US. Modulation of guanine nucleotides bound to Ras in NIH-3T3 cells by oncogenes, growth factors, and the GTPase activating protein (GAP). J Biol Chem 265:20437-20442, 1990. Tahira T, Hayashi K, Ochiai M, Tsuchida N, Nagao M, Sugimura T. Structure of the c-Ki-ras gene in a rat fibrosarcoma induced by 1,8-dinitropyrene. Mol Cell Biol 6:1349-1351, 1986. Hilke H, Maillet M, Robinson R. H-ras-l point mutations in soft tissue sarcomas. Mod Pathol 6:129-132, 1993. mutation frequency in ras genes of skin tumors isolated from DNA repair deficient xeroderma pigmentosum patients. Cancer Res 53:1625-1629, 1993. Daya-Grosjean L, Robert C, Drougard C, Suarez H, Sarasin A. High L Nakazawa H, Aguelon A-M, Yamasaki H. Relationship between chemically induced Ha-ras mutation and transformation of BALB/c 3T3 cells: evidence for chemical-specific activation and cell-type specific recruitment of oncogene in transformation. Mol Carcinog 3:202-209, 1990. Devereux TR Belinsky SA, Maronpot RR, et al. Comparison of pulmonary O -methylguanine DNA adduct levels and Ki-ras activation in lung tumors from resistant and susceptible mouse strains. Mol Carcinog 8:177-185, 1993. Ohgaki H, Kleihues P, Hard GC. Ki-ras mutations in spontaneous and chemically induced renal tumors of the rat. Mol Carcinog 4:455- 459, 1991. Herzog CR, Schut HAJ, Maronpot RR, You M, ras mutations in 2- amino-3-methylimidazo-[4.5-flquinoline-induced tumors in the CDF1 mouse. Mol Carcinog 8:202-207, 1993. Hoffmann J-S, Fry M, Ji J, Hilliams KJ, Loeb LA. Codons 12 and 13 of H-ras protooncogene interrupt the progression of DNA synthesis catalyzed by DNA polymerase a . Cancer Res 53:2895-2900, 1993. Hofmann M, Rudy H, Gfinthert U, et al. A link between ras and metastatic behavior of tumor cells: ras induces CD44 promoter activity and leads to low-level expression of metastatic-specific variants of CD44 in CREF cells. Cancer Res 53:1516-1521, 1993. Jankun J, Maher VM, McCormick JJ. Malignant transformation of human fibroblasts correlates with increased activity of receptor- bound plasminogen activator. Cancer Res 51:1221-1226, 1991. Dillberger JH. Ph.D. Dissertation. Michigan State University. East Lansing, 1989. 49. 50. 51. 52. 53. 54. 55. 56. 119 Schonthal A, Herrlich P, Rahmsdorf HJ, Ponta H. Requirement for fos gene expression in the transcriptional activation of collagenase by other oncogenes and phorbol esters. Cell 54:325- 334, 1988. Topol LZ, Kisseljova NP, Gutierrez ML, et al. Modulation of pp60” "‘ and pp60°‘" expression in Rous sarcoma virus-transformed hamster fibroblasts transfected with activated N-ras. Mol Carcinog 8:167-176, 1993. Delage S, Chastre E, Empereur S, et al. Increased protein kinase C a expression in human colonic caco-2 cells after insertion of human Ha-ras or polyoma virus middle I oncogenes. Cancer Res 53:2762-2770, 1993. DeClue JE, Stone JC, Blanchard RA, et al. A ras effector domain i mutant which is temperature sensitive for cellular transformation: interactions with GTPase-activating protein and NF-1. Mol Cell Biol 11:3132-3138, 1991. Molloy CJ, Fleming TP, Bottaro DP, Cuadrado A, Aaronson SA. Platelet-derived growth factor stimulation of GTPase-activating protein tyrosine phosphorylation in control and c-H-ras-expressing NIH-3T3 cells correlates with p21m activation. Mol Cell Biol 12:3903-3909, 1992. Satoh T, Endo M, Nakafuku M, Akiyama T, Yamamoto T, Kaziro Y. Accumulation of p21“"oGTP in response to stimulation with epidermal growth factor and oncogene products with tyrosine kinase activity. Proc Natl Acad Sci USA 87:7926-7929, 1990. Buday L, Downward J. Epidermal growth factor regulates p21"‘ through the formation of a complex of receptor, Grb2 adapter protein, and Sos nucleotide exchange factor. Cell 73,611-620, 1993. Zhang X, Settleman J, Kyrikis JM, et al. Normal and oncogenic p21"° proteins bind to the amino-terminal regulatory domain of c- Raf-l. Nature 364:308-313, 1993. APPENDIX APPENDIX Recent Developnents in ras Signalling Pathways A. Introduction Evidence has been accumulating since the first discovery of the ras genes that the proteins encoded by these genes are critical for regulation of normal growth and for the process of malignant transformation. However, though data linked ras activation with transformation, investigators gained little insight into the molecular mechanisms by which ras p21 functions. Early indications that ras p21 functions in a growth factor/mitogenic response pathway came from the data by Mulcahy et al. (1985) who showed that functional p21 is required in order for DNA synthesis to occur, but advances in knowledge about ras p21-mediated signalling pathways haveronly recently provided information about the role of these pathways in control of growth and cellular responses to extracellular mitogenic signals. A consensus model for the ras-mediated signal transduction pathway has been outlined recently in Seienee (Marx, 1993), as shown in Figure 1. I will summarize these very recent findings, most of which were reported after I had completed the bulk of the research for this dissertation, because this will allow me to place the results I obtained in the context of this new knowledge, and suggest additional experiments that could be done to gain a better understanding of the role of ras in normal growth and malignant transformation of human fibroblasts. 120 121 B. Internediates between Growth Factor Receptors and ras p21 The structural properties of ras p21 (attachment to the inner surface of the plasma membrane, guanine nucleotide binding, low level GTPase activity, presence of an effector binding domain) are all consistent with its putative function as a signal transduction molecule. However, it was unclear how p21 is triggered by mitogenic signals such as protein growth factors. Since ras p21 is known to cycle between the active GTP-bound and the inactive GDP-bound forms, it was reasonable to hypothesize that signalling would involve an increase in the GTP-bound form of p21 (Haubruck and McCormick, 1991). Research on signalling mechanisms in several non-mammalian organisms (Drosophila melanogaster, Caenorhabditis elegans, saccharomyces cerevisiae, saccharomyces pombe), all of which have genes very similar to ras, has identified proteins that act as intermediaries between protein growth factor receptors and ras p21. The mammalian counterparts of these intermediary proteins appear to have a similar function (Feig, 1993). At least three related sets of intermediary proteins have been identified. One set appears to be present in numerous cell types, whereas the others are expressed only in specific cell types. Since the ubiquitous set of proteins have been studied the most widely, only these will be reviewed. One of these proteins, encoded in humans by the hsosl gene and in rodents by the.msosl gene, is a guanine nucleotide releasing protein that causes GDP to be released from p21 (Simon et al., 1991; Jones et al., 1991; Bonfini et al., 1992; Bowtell et al., 1992; Hei et al., 1992; Shou et al., 1992). p21 has a much greater affinity for GTP than for GDP. Therefore, once the GDP is released, ras p21 binds to GTP and becomes activated. The other protein, encoded by the grb2 gene, has no catalytic domain (Clark et al., 1992; Lowenstein et al., 122 1992; Simon et al., 1993: Olivier et al., 1993: Buday and Downward, 1993), but does have two types of domains that function as adaptors, i.e., SHZ and SH3. Evidence suggests that the SH2 domain binds to the phosphotyrosine residues of growth factor receptors such as the epidermal growth factor receptor (EGFR), and the SH3 domains interact with the product of the hsosl gene (Rozakis-Adcock et al., 1993). Evidence for the proposed interactions between growth factor receptors, grb2 adaptor proteins, sos nucleotide exchange factors and ras p21 proteins comes from several sources. Li et al. (1993) showed that when expressed in human kidney 293 cells that have been stimulated with epidermal growth factor (EGF), the human grb2 protein and the hsosl protein interact. Subsequent work reported from the same laboratory (Gale et al., 1993) using a variant of NIH-3T3 cells that over-expresses the EGFR, showed that an increase in grb2 protein activity results in an increase in the relative amount of GTP-bound p21 (p21-GTP), that this increase is caused by release of GDP rather than by a decrease in GTPase activity, and that the increase in p21-GTP is associated with increased protein kinase activity in the stimulated cells. These investigators also showed that the sos1 protein and the grb2 protein form a complex even in non-stimulated cells, but that the complex associates with the EGFR only when the cells have been stimulated with EGF. Rozakis-Adcock et al. (1993), using a human grb2 protein with a mutated SH3 domain, showed that the SH3 domain is required for grb2 to bind to msosl protein. Using a similar approach, these investigators also showed that the carboxyl terminal domain of the msos protein binds to grb2. This carboxyl terminal domain has sequence similarity to a protein that binds the SH3 domains of the src and abl proteins. That these intermediary proteins 123 have been highly conserved across divergent species gives an indication of their importance to normal cell function. In addition, Egan et al. (1993) showed that over-expression of the Drosophila sosl protein can transform NIH-3T3 cells, lending support to the hypothesis that this part of the signal transduction pathway is also important to malignant transformation. C. Internediates between ras p21-GTP and the Nucleus 1. Internediates that Appear to Interact Directly with p21-GTP Events that occur after ras p21 has bound GTP have also been the subject of extensive research. The consensus of these investigations is that the interaction between p21-GTP and one or more cytoplasmic protein kinases initiates a phosphorylation cascade that results in the'activation of nuclear transcription factors and triggers expression of the genes that control mitogenesis. The following paragraphs describe some examples of research in this area. Morrison et al. (1988) showed that transformation of NIH-3T3 cells with oncogenes encoding membrane-bound proteins (src, fms, sis, ras, and polyoma middle T) resulted in cells capable of phosphorylating the raf-l protein, a serine/threonine kinase that is the product of the raf—l gene, the mammalian homologue of the v-raf oncogene of the murine sarcoma virus 3611 (Rapp et al., 1983). This led to the hypothesis that the raf—l protein interacts with ras p21 and other membrane-bound oncogene products as part of a signal transduction pathway. However, Reed et al. (1991) failed to observe phosphorylation of the raf-l protein in response to expression of activated H-ras in NIH-3T3 cells unless the cells had also been stimulated with a protein growth factor. In their study, cells that 124 expressed the v-src protein did not require additional growth factor stimulation to phosphorylate the raf-l protein. This led the authors to conclude that an activated ras protein, unlike the src protein, cannot initiate a mfitogenic signal, but can only amplify signals provided by protein growth factors. However, it should be noted that although the cells used in this study by Reed et al. (1991) had been transformed by an activated ras gene, they were not able to cycle without the addition of exogenous growth factors. In many other experimental transformation systems, cells transformed by transfection of ras are growth factor independent, and in some cases are fully transformed. One explanation for the results of Reed et al. (1991) is that in their system, some component of the ras p21-mediated pathway prior to the raf-l protein was not functioning. Another approach to determining whether there is interaction between ras p21 and the protein kinases in the signal transduction pathway, is to transfect cells with a gene encoding a dominant-negative'mutant of ras p21 and to observe the effect that mutant ras p21 has on the phosphorylation state and/or the kinase activity of the protein kinases. An example of this approach is the work of Hood et al. (1992) who transformed human pheochromocytoma cells with a dominant-negative ras mutant whose expression could be controlled with dexamethasone. Expression of this non-functioning mutant p21 abolished the phosphorylation of the raf-l protein and other cytosolic protein kinases that normally occurs in response to stimulation with protein growth factors. This indicates that the presence of functioning ras p21 is critical for transmission of the mitogenic signal. The block was not always complete, but whether this reflected the presence of a ras-independent pathway or merely the 125 inability of mutant ras p21 to fully compete with endogenous ras p21 was not determined. Hood et al. (1992) also showed that increased phosphorylation of the raf-l protein occurred when the cells expressed an oncogenic form of ras. Thomas et al. (1992) used this same approach to show that functional ras p21 is required to stimulate the activity of another serine/threonine kinase, mitogen-activated protein kinase (MAPK). These investigators also showed that the ras p21-induced increase in MAPK activity resulted from phosphorylation of MAPK, not from its over- expression. Hork by Harne et al. (1993) also linked phosphorylation of the raf-l protein to interaction with p21-GTP. In addition, they showed that the effector domain of ras p21 interacts with the amino terminal domain of the raf-l protein. Moodie et al. (1993) studied the association between ras p21 and the raf-l protein using an in vitro system in which ras p21 immobilized on silicon beads was used to probe rat brain cytosol extracts. These investigators showed that.wild-type p21-GTP strongly bound the raf-1 protein and stimulated its kinase activity whereas wild-type p21-GDP bound raf-l poorly and did not stimulate its kinase activity. Interestingly, the GDP-bound formiof an oncogenic variant of ras p21, with valine instead of glycine at codon 12, bound poorly to the raf-l protein but nonetheless strongly stimulated raf-l protein kinase activity. This led to the hypothesis that oncogenic ras p21 has an altered structure that allows constitutive interaction with the raf-l protein or with other intermediates in the pathway. 126 2. An Internediate that Interacts Directly with the raf-l protein A study providing strong evidence for the interaction between Raf1 and another protein in the chain, MAPK kinase (MAPKK), was reported by Hughes et al. (1993). Since MAPKK has a high degree of sequence similarity to the saccharomyces pombe mating factor encoded by the byrl gene, these investigators attempted to complement a byrl-minus mutant with the mammalian MAPKK protein. They could not successfully complement the mutant unless the raf-l protein‘was expressed along with MAPKK. By mixing cellular extracts and testing for MAPKK activity they showed that its kinase activity was dependent on the raf-l protein. Enhancement of MAPKK activity was the result of increased phosphorylation of MAPKK, not increased protein expression. 0. Conclusions Several problems in the ras p21-mediated signalling pathway remain unresolved. One example is that the specific reactions between the intermediates in the pathway, the kinds of control mechanisms that are present, and the interactions, if any, between various pathways are not yet defined. A second example is that the role in this pathway of ras p21 itself needs further clarification, for instance, the differences in wild- type and activated ras p21 and the proteins with which they interact. Studies to resolve these problems must be carried out with attention to cell biology, in particular to the ability of the cell to divide under the conditions of the experiment, and to the transformation status of the cells. The results to date of studies that I have recently been carrying out indicate that the cycling characteristics of a cell line are closely related to its ras p21 GTP/GDP ratio, to the nature of its ras genes, and 127 to the level of expression of its ras genes. However, the precise relationship of these variables to one another and to cell transformation is not entirely clear. Experiments are planned to determine whether the cell lines and strains that I showed do not have an activated ras gene do have an alteration in one or more of the other components in the ras- mediated pathway that is contributing to transformation. One approach will be to determine the p21 GTP/GDP ratio in cells that contain two wild- type ras alleles. It is possible that alterations in either grb2 or 5051 result in constitutive activation of p21, and if so, this would be expected to result in an increase in GTP-bound ras p21. in contrast, alteration of one of the protein kinases that normally interacts with ras p21-GTP would not be expected to be reflected in an increase in GTP-bound ras p21. Such information, along with what is known about ras genes from research in this laboratory and from data in the literature on the ras- mediated signalling pathways, should yield insights into malignant transformation of human fibroblasts. 128 Figure 1. Consensus model of the ras-mediated signalling pathway (Reprinted with permission of AAAS. Marx, J. Forging a path to the nucleus [Research News]. Science 260:1589, 6/11/93, °AAAS.) LIST OF REFERENCES Bonfini, L.,-C. A. Knrlevich, C. Dasgupta, and U. Banerjee. 1992. The son of sevenless gene product: a putative activator of ras. Science 255:603- 606. - Bowtell, D., P. Fe, M. Sinon, and P. Senior. 1992. Identification of murine homologues of the Drosophi la son of sevenless gene: potential activators of ras. Proc. Nat. Acad. Sci. USA 89:6511-6515. Buday, L., and J. Downward. 1993. Epidermal growth factor regulates p21m through the formation of a complex of receptor, grb2 adapter protein, and sos nucleotide exchange factor. Cell 73:611-620. Clark, S. G., M. J. Stern, and H. R. Hervitz. 1992. C. elegans cell- ;ignalling gene sem-5 encodes a protein with SH2 and SH3 domains. Nature 5 :340-344. Egan, S. E., B. H. Giddings, M. H. Brooks, L. Buday, A. M. Sizeland, and R. A. Heinberg. 1993. Association of Sos Ras exchange protein with Grb2 is implicated in tyrosine kinase signal transduction and transformation. Nature 363:45-51. Feig, L. 1993. The many roads that lead to Ras. Science 260:767-768. (Perspectives.) “Ie’ "o "o. 5. m1“. E. J. Lmsulfl, So ”1631”". m D. Bar- Sagi. 1993. Grb2 mediates the EGF-dependent activation of guanine nucleotide exchange on Ras. Nature 363: 88-92. Haubruck, H., and F. McCornick. 1991. Ras p21: effects and regulation. Biochim. Biophys. Acta 1072:215-229. Hughes, D. A., A. Ashwerth, and C. J. Marshall. 1993. Complementation of byrI in fission yeast by'mammalian MAP kinase kinase requires coexpression of Raf kinase. Nature 364:349-352. Jones, 5., M-L. Vignais, and J. R. Breach. 1991. The 00625 protein of Saccharomyces cerevisiae promotes exchange of guanine nucleotides bound to Ras. Mol. Cell. Biol. 11:2641-2646. Li, H., A. Batzer, R. Daly, V. Yajnik, E. Skolnik, P. Chardin, D. Bar- Sagi, B. Margolis, and J. Schlessinger. 1993. Guanine-nucleotide-releasing factor hSosl binds to Grb2 and links receptor tyrosine kinases to Ras signalling. Nature 363:85-88. 129 130 Lowenstein, E. J., R. J. Only, A. G. Batzer, H. Li, B. Margolis, R. La-ers, A. Ullrich, E. Y. Skolnik, D. Bar-Sagi. and J. Schlessinger. 1992. The SH2 and SH3 domain-containing protein GR82 links receptor tyrosine kinases to ras signaling. Cell 70:431-442. Marx, J. 1993. Forging a path to the nucleus. Science 260:1588-1590. (Research News.) Moodie, S. A., 8. M. Hillusen, M. J. Heber, and A. Holfnan. 1993. Complexes of Ras GTP with Raf- 1 and mitogen-activated protein kinase kinase. Science 260: 1658-1661. Morrison, 0. K., D. R. Knplan, U. Rapp. and T. M. Roberts. 1988. Signal transduction from membrane to cytoplasm: growth factors and membrane-bound oncogene products increase Raf-1 phosphorylation and associated protein kinase activity. Proc. Nat. Acad. Sci. USA 85:8855-8859. Mulcahy, L.S., M.R. Snith, and D.H. Stacey. 1985. Requirement for ras proto-oncogene function during serum-stimulated growth of NIH-3T3 cells. Nature 313:241-243. Olivier, J. P., T. Rube, M. Henkeneyer, B. Dickson, G. lbanalu, B. Margolis, J. Schlessinger. E. Hafen, and T. Pawsen. 1993. A Drosophila SH2-SH3 adaptor protein implicated in coupling the sevenless tyrosine kinase to an activator of ras guanine nucleotide exchange, sos. Cell 7 :179-191. Rapp. U. R., M. D. Goldsbereugh, G. E. Mark, T. I. Bonner, J. Greffen, F. H. Reynolds, Jr., and J. R. Stephenson. 1983. Structure and biological activity of v-raf, a unique oncogene transduced by a retrovirus. Proc. Nat. Acad. Sci. USA 80: 4218-4222. Reed, J. C., S. Yun, M. P. Cuddy, B. C. Turner, and U. R. Rapp. 1991. Differential regulation of the p72-74 RAF-1 kinase in 3T3 fibroblasts expressing ras or src oncogenes. Cell Growth and Differentiation 2:235- 243. ROZ‘k‘S- . "o. R. FernIey. J. "l“, T. P‘”«'. CM 0. “£11. 1993. The SH2 and SH3 domains of mammalian Grb2 couple the EGF receptor to the Ras activator mSosl. Nature 363:83-85. Shou, C., C. L. Farnswerth, B. G. Heel, L. A. Feig. 1992. Molecular cloning of cDNAs encoding a guanine-nucleotide-releasing factor for Ras p21. Nature 358:351-354. Sinon, M. A., D. D. L. Bowtell, G. S. Dodson, T. R. Laverty, and G. M. Rubin. 1991. Rasl and a putative guanine nucleotide exchange factor perform crucial steps in signaling by the Sevenless protein tyrosine kinase. Cell 67:701-716. Sinon, M. A., G. S. Dodson, and G. M. Rtbin. 1993. An SH3- SH2- SH3 protein is required for p213m activation and binds to sevenless and sos proteins in vitro. Cell 73:169-177. 131 Thenas, S. M., H. DeMarco, G. D’Arcangele, S. Halegoua, and J. S. Brugge. 1992. Ras is essential for nerve growth factor- and phorbol ester-induced tyrosine phosphorylation of MAP kinases. Cell 68:1031-1040. Harne, P. H., P. R. Viciana, and J. Downward. 1993. Direct interaction of Ras and the amino-terminal region of Raf-1 in vitro. Nature 364:352-355. Hei, It”. R. IL. Mosteller, In. Sanyal, EL. Gonzales, IL. McKinney, (L. Dasgupta, P. Li, B-X. Liu, and D. Break. 1992. Identification of a mammalian gene structurally and functionally related to the £0025 gene of Saccharomyces cerevisiae. Proc. Nat. Acad. Sci. USA 89:7100-7104. Heed, K. H., C. Sarnecki, T. M. Roberts, and J. Blenis. 1992. ras mediates nerve growth factor receptor modulation of three signal-transducing protein kinases: MAP kinase, Raf-1, and RSK. Cell 68:1041-1050. "‘mmlmll