How do mutated oncogenes and tumor suppressor genes cause cancer?

Medical Oncology (1998) 15, 20-26 9 1998 Stockton Press All rights reserved 0736-0118/98 $12.00 REVIEW How do mutated oncogenes and tumor suppressor genes cause cancer? Dan Grand6r Research Laboratory of Radiumhemmet, Karolinska Hospital, Stockholm, Sweden In recent decades we have been given insight into the process that transforms a normal cell into a malignant cancer cell. It has been recognised that malignant transformation occurs through successive mutations in specific cellular genes, leading to the activation of oncogenes and inactivation of tumor suppressor genes. The further study of these genes has generated much of its excitement from the convergence of experiments addressing the genetic basis of cancer, together with cellular pathways that normally control important cellular regulatory programmes. In the present review the context in which oncogenes and tumor suppressor genes normally function as key regulators of physiological processes such as proliferation, cell death/apoptosis, differentiation and senescence will be described, as well as how these cellular programmes become deregulated in cancer due to mutations. Keywords: oncogenes; tumor suppressor genes; proliferation; apoptosis; differentiation; senescence Introduction It is today clear that cancer is caused by specific mutations in specific key regulatory genes. Since the discovery of the first such genes some 30 years ago, we have seen remarkable progress in the definition of genetic lesions in malignancy, and furthermore our general understanding of cell biology has made us able to comprehend how these cancer genes normally function in the control of the non-malignant cell, and how this relates to the ability of these genes to cause malignant transformation when mutated. In normal tissues of :aaulticellular organisms the function of the single cell is tightly controlled due to inner and surrounding constraints, to maintain tissue homeostasis and function. During malignant transformation the cancer cell will gradually become more autonomous, and develop into an 'asocial citizen' of its tissue, growing in an uncontrolled manner at the Correspondence: Dr G Grand6r, Research Laboratory of Radiumhemmet, CCK, Karolinska Hospital, S-171 76 Stockholm, Sweden. Received 23 December 1997; accepted 2 January 1998 expense of the function of the normal tissue. More specifically, the cancer cell has obtained relaxed control of several normal cell regulatory mechanisms. The most obvious such function is proliferative control. In many types of malignancies (but not all), the tumor cells divide at a higher rate, often irrespective of the extracellular growth signals that usually control cell growth. 1 Although obvious today, it was not until recently that it was realised that tissue growth is not only regulated by the rate of cell division, but that the rate of cell death or apoptosis is another key factor in tissue homeostasis9 With this came the understanding that malignancy may also develop through reduced cell death rate. Furthermore, a common feature of cancer cells is their inability to differentiate terminally, which in normal tissues would lead to cessation of proliferation. Cell cultures derived from human tumors often divide indefinitely; furthermore, the number of population doublings a cell has to go through in order to become a clinically detectable tumor is very large. This is in sharp contrast to non-transformed cells which, with few exceptions, have a limited lifespan and will, after a defined number of population doublings, enter a non-

replicative but metabolically active state termed senescence. 3 This has led to the hypothesis that cellular senescence is a tumor suppressor mechanism which is abrogated in malignancy. 4 As previously mentioned a number of genes that are specifically mutated in malignant cell have been defined to date. They can generally be divided into two groups: oncogenes and tumor suppressor genes. Oncogenes take part in malignant transformation due to activating mutations such as amplification, small mutations or translocations. Due to their activating nature, mutations in these genes are usually dominant and only one allele of the gene needs to be affected. The function of tumor suppressor genes on the other hand seems to be to protect the normal cell from developing into a cancer cell, and a loss of their function leads to malignant transformation. As the majority of human genes are present in two copies in the genome, both alleles usually have to be inactivated for their tumor-suppressing activity to be lost. Recently a third category of cancer-related genes has been defined; namely the DNA repair genes. These are a group of genes that take part in the normal repair of DNA damage, and somatic or inherited loss of their function leads to an increased frequency of secondary mutations in oncogenes and tumor suppressor genes. 5 With the definition of the molecular background of a number of different inherited cancer syndromes, it has been shown that the increased cancer risk in these families may often be due to the inheritance of a mutated tumor suppressor gene. The present review will not give a detailed list of known cancer-related genes, but will rather describe the normal context in which these genes regulate physiological processes such as proliferation, cell death/ apoptosis, differentiation and senescence, and how they become deregulated in the malignant cell due to mutations in oncogenes and tumor suppressor genes. The review will not deal to any major extent with other factors that may be of importance in the malignant process, such as the mechanism behind the acquisition of metastasising potential or the interaction of tumor cells with surrounding stroma and blood vessels. Proliferation Our understanding of the regulation of eukaryotic cell division took a quantum leap some ten years ago, when it was realised that cell cycle progression is governed by sequential formation and activation of enzyme complexes consisting of a family of related cyclin-depend- ent serine-threonine kinases (cdks) and their cyclin partners (for a review see Ref. 6). At the various cell cycle phases, waves of kinase activity corresponding to different family members of the calks will propel cells through the periodic events of mitotic division by phosphorylation of key substrates. The activity of these kinases is positively and negatively regulated by various growth regulatory signals. Positive external mitogenic signals will come from stimulation of the cell by various growth factors through binding to their specific cell surface receptors. Growth factor stimulation will lead to activation of intracellular signalling systems such as membrane associated enzymes and transcription factors. These events will affect the cdks active in the G~ phase by increasing cyclin levels as well as changing the phosphorylative status of the cdks themselves, leading to full activation of these kinases and permitting further progress in the cell cycle (see Figure 1). One of the major substrates for the activated kinases in the G1 phase is the retinoblastoma protein (prb). When prb becomes phosphorylated by the kinases this will lead to release of a group of transcription factors of the E:F family that activate genes essential for S phase such as DNA polymerase alpha, and c-myc (see Figure 1). Several observations suggest that the phosphorylation of prb by the cyclin-cdk complexes is a rate-limiting step in Gj progression. After this point cells become committed to cell division irrespective of whether growth factors are present or not. 7 This event has previously been termed the 'restriction point'. The cdks can also be negatively regulated by the newly discovered cdk inhibitors (ckis). These latter inhibitory proteins have been found to be important effector molecules for internal and external growth inhibitory signals such as DNA damage, contact inhibition and the cell growth inhibitory cytokines TGFI3 and intcrferons. 6 As previously mentioned, uncontrolled growth is a common feature of malignant cells. Genes that control the decision to initiate a new round of DNA synthesis are attractive candidates for oncogenes and tumor suppressor genes, depending on whether they have a stimulatory or inhibitory role in the process. Indeed, several oncogenes and tumor suppressor genes have been found to participate directly in cell cycle regulation. 1"6 For example the malignant cell may have mutations in genes leading to an autonomous production of growth factors, constitutive activation of growth factor rcceptors or membrane-bound signalling proteins, leading to more or less autonomous growth. It has also become clear that malignant cells commonly carry 21

22 mutations in the very key cell cycle regulatory proteins such as overproduction of cyclin D1 or cdk 4 or inactivation of the cki p16. The inactivation of p16 seems to occur at high frequencies in a wide variety of malignancies such as leukaemias, melanomas, pancreatic cancer and bladder cancer. 8 Malignant cells may also have mutations in the more downstream effector molecules, such as inactivation of the retinoblastoma gene or overproduction of E2F responsive genes such as c-myc. As delineated above, cancer cells commonly have mutations at some level of the molecular machin- receptor g 1 growth factor -nucleus signa_.!jng proteins 1 S.phase\,,mu~mmlimm,,,~,n~,,~ DNA/ x'~. hinery Figure 1 Outline of cell cycle regulation in the Gz phase. Binding of growth factors to their specific receptors will lead to activation of membrane associated signalling enzymes such as ras, increase the levels of D-tcpe cyclins and activation of the cdk complex. Gt cdks will phosphorylate the retinoblastoma protein leading to the release of E2F transcription factors. E2F induces the transcription of S phase genes such as DNA polymerase-a and c-myc The cdk complexes can be negatively regulated by the cyclin kinase inhibitors as exemplified by p16. Shaded proteins are commonly mutated in malignant cells leading to relaxed control of,tell growth ery that regulates G1 progression into S phase, explaining increased and uncontrolled proliferation in malignancy (Figure 1). Apoptosis It is not until recently that the major importance of tightly regulated cell death in tissue homeostasis has been duly recognised. Apoptosis is the descriptive name given to the process of physiological programmed cell death in vertebrates (for a review see Ref. 2). During apoptosis a cell activates an intrinsic suicide machinery that results in cell death: its surface membrane begins to bleb, the cell shrinks, the chromatin becomes condensed and cleaved, and eventually the whole cell fragments into membrane-bound vesicles that are rapidly ingested by neighbouring cells. Apoptosis has been shown to be important in a number of physiological processes such as embryonic development, immune regulation and tissue homeostasis. Recently we have gained important knowledge in the molecular regulation of the apoptotic process. A number of stimuli have the capability of activating the apoptotic programme. These stimuli include both extracellular factors, such as the FAS antigen, TNE interferons and matrix attachments as well as intracellular events, as for example DNA damage (Figure 2). 2 However, not every cell that is exposed to an apoptotic signal undergoes apoptosis, which demonstrates the importance of intracellular modulators regulating the sensitivity of the cell. The Bcl2 family of proteins are the best known rheostats in regulating the cellular sensitivity to apoptosis. 9 The family members either promote cell survival (eg Bcl2) or augment programmed cell death (eg Bax). It is believed that the agonistic and antagonistic proteins of the Bcl2 family interact through hetero- or homodimerisation, and the relative protein amounts of pro- and anti-apoptotic Bcl2 family members will determined whether an apoptotic signal will result in cell death or not (Figure 2). Recently, an emerging family of cysteine proteases have been found to be essential downstream effectors in the actual killing of the apoptotic cell. In this family the nematode C. elegans protease Ced3 and its vertebrate homologue interleukin-l[3 converting enzyme (ICE) are the prototype cysteine proteases. 1~ All the members in the ICE/Ced3 protease family share some features: (1) They all induce apoptosis when overexpressed.

cell death inducers Bcl-2 protein family rheostat imlmnmnm... DNA damage TNF FasL (~ ~1 r? / ICE/CED-3 protease cascade cell death Recently a direct link between the FAS/FASL and TNF signalling and the ICE protease cascade was found. In this system, a limited number of proteins transmit the signal from the plasma membrane receptors to the proteases in the cytoplasm through proteinprotein interactions, using the conserved motives 'death domain' and 'death effector domain'. 11 The finding that the Bcl2 oncogene functions in preventing apoptosis instead of promoting proliferation not only established a new class of oncogenes, but also revealed the fact that malignant cells often exhibit decreased sensitivity to apoptotic signals leading to prolonged survival. 2 Furthermore, it seems that most chemotherapeutic agents and radiotherapy utilised in anticancer treatment act through induction of apoptosis in the malignant cells. 2 It has also been realised that the p53 tumor suppressor gene is involved in the apoptotic response to several types of DNA-damaging agents, including chemotherapy and radiotherapy. The p53 gene is the hitherto most commonly mutated gene in human cancer, as abberations in this gene is estimated to occur in more than 50% of all tumors. Tumor cells with p53 mutation thus have a reduced susceptibility to apoptosis induced by several stimuli, including agents used in anticancer therapy. 12 Several other traditional oncogenes and tumor suppressor genes, such as c-myc and the retinoblastoma gene, have also been implicated in regulating cell death 2 even though the mechanism behind this is less clear. 23 Figure 2 A model of the molecular regulation of apoptosis. The cell receives apoptotic signals from exogenous stimuli such as TNF and FasL or internal stimuli such as DNA damage. The relative ratios of pro- and anti-apoptotic Bcl-2 family members will determine whether an apoptotic signal will result in cell death or not. Apoptotic signals result in the activation of the ICE/CED-3 cystein protease cascade, presumably cleaving essential cellular substrates leading to cell death. Shaded proteins such as Bcl-2 and p53 are commonly mutated in malignant cells leading to a decreased cellular sensitivity to apoptosis (2) They are all produced as proenzymes and have substrate specificity for cleavage after an aspartate residue. Several different classes of substrate have been identified including the pro-proteascs themselves. Poly (ADP-ribose) polymerase (PARP) and DNA-dependent protein kinase (DNA-PK) are two proteins involved in DNA repair, which are cleaved by the ICE proteases, but also structural proteins such as lamins and actin are substrates for these enzymes. However, other unknown important targets probably exist. Differentiation The ability of cells to differentiate in an orderly and controlled manner is of major importance for multicellular organisms, where all specialised cells are derived from a single totipotent cell. In comparison with our comprehension of cellular proliferation and apoptosis, our understanding of the differentiation process is less clear, possibly reflecting a greater complcxity in its regulation and a larger variability between tissues. One process that has been studied in some detail is haematopoiesis and will therefore be further discussed as an example in this context. All types of mature specialised blood cells are believed to originate from a small subset of ancestor cells. During haematopoiesis immature multipotent stem cells undergo progressive restriction of lineage potential to give rise to mature, terminally differentiated functional blood cells, which eventually undergo programmed cell death (Figure 3). 13 This process is regulated by the intricate cooperation between exogenous and endogenous molecules. The exogenous molecules include different cytokines,

- 24 D Granddr whose cellular response is activated by binding to specific receptors. These receptors transform the cytokine message into cellular signals that ultimately result in altered DNA transcription, thus leading to the activation of different genetic programmes and progressive differentiation. At the centre of this transcriptional regulation induced by the external factors reside endogenous cell-specific proteins known as transcription factors. These proteins bind to short stretches of DNA in a sequence-specific manner and thus alter the transcription of specific genes in a positive or negative manner. Some of these transcription factors have been found to coordinate the expression of many different genes involved in lineage specific genetic programmes of differentiation, and are thus thought of as master or key transcriptional regulalors (Figure 3). In leukaemia as well as other malignancies, the malignant cells commonly have lost the capacity to Lymfocytes ~6 stemcell progenitor tlus; - - GM-CSF o ~ Erythrocytes "O'S'A"YI / Macrophages Figure 3 Haematopoietic differentiation and some of its important molecular regulators. Exogenous signals such as cytokines (denoted below the arrows) and endogenous proteins in the form of key transcription factors (denoted within the arrows), will have positive or negative roles in the commitment of cells to maturation along a specific lineage. Some of the transcription factors known to be mutated during malignant transformation are indicated by italics differentiate, and become frozen at a certain stage of differentiation. The pathomolecular background to this is not entirely known, but in acute leukaemias a recurrent theme of genetic alteration is the occurrence of translocations involving known differentiation regulating transcription factors (Figure 3). These translocations usually lead to the formation of chimeric proteins which either abrogate the function of the transcription factor or lead to activation at an inappropriate timepoint, 13 presumably perturbing the normal differentiation programme of the affected cell. It has furthermore been suggested that abrogation of the differentiation blockage in malignant cells could lead to reversion into a non or less malignant state. 14 In different in vitro systems, mainly utilising leukaemic cell lines, a wide variety of compounds have indeed demonstrated differentiating activity with the induction of mature non-proliferating cells. 13A4 Despite the progress with in vitro systems, little has been achieved in vivo. To date, only patients with acute promyelocytic leukaemia have come to benefit from this approach. This disease is characterised by a balanced translocation between chromosomes 15 and 17, producing a chimeric protein fusing the retinoic acid receptor RARer to the transcription factor PML. In the majority of these patients treatment with retinoic acid leads to complete remission with the appearance of mature cells containing the t(15;17). However, retinoic acid treatment does not cure acute promyelocytic leukaemia, as the disease usually relapses in the absence of consolidating therapy. Senescence As described by Hayflick more than 30 years ago, human cells have a limited lifespan and will, after a defined number of population doublings, enter a nonreplicative but metabolically active state termed senescence. 3 During replicative senescence, cells arrest in the GI phase of the cell cycle. With the discovery of a number of proteins normally regulating the cell cycle and G1 progression, some of the molecular events involved in replicative senescence have been identified. 4 It has recently been demonstrated that senescent cells express increased levels of the previously mentioned ckis p16, p15, p21. 4"15 The p16 and p15 proteins accumulate as a consequence of the increasing number of population doublings, eventually leading to senescence, t5 Another feature shared by senescent cells is telomere shortening. The telomeres are repetitive sequences at the ends of chromosomes, and their shortening has been suggested as a candidate for a cell division 'counting' mechanism in normal somatic cells

of higher organisms, 16 as the telomere length decreases with increasing population doublings. It has been proposed that during successive rounds of replication, progressive loss of telomere length is sensed as DNA damage and causes cells to exit the cell cycle. Senescence has been suggested to be an important tumor suppressive mechanism, and there is substantial evidence to support this idea. 4 First, senescence prevents cells from acquiring the multiple mutations that are needed for malignant transformation. Indeed, many, if not most, malignant tumors contain cells that have an extended or indefinite division potential. During malignant transformation there seems to be a selection for cells that can bypass senescence. Secondly, several oncogenes, both cellular and viral, act at least in part by extending the replicative lifespan. Furthermore, the p16 and p15 ckis that have been shown to accumulate during normal senescence, are commonly inactivated in many human tumors, possibly allowing the malignant cells to escape the senescence programme, s Malignant cells have often also acquired an ability to maintain telomere length through the expression of the enzyme telomerase which adds essential telomeric sequences to maintain the ends of chromosomes. 16 Other Deregulated Processes in Malignant Cells Although it is not the scope of the present review to cover other functions that may be altered in malignant cells and their molecular background, one such area, metastasising potential, will be mentioned briefly. The majority of cancer-related deaths are not due to growth of the primary tumor, but rather because of metastasising disease. The molecular background of metastasis is still unclear, even though some mechanisms have been proposed, such as changes in cell adhesion molecules and expression of proteases. 17 Alterations in the above mentioned oncogenes and tumor suppressor genes leading to transformation of cells exhibiting undifferentiated features, autonomous growth and decreased sensitivity to apoptotic signals, probably also create cells with an ability to reside in other environments than the original organ. This is also exemplified by a host of animal experiments, in which normal cells transfected with a limited number of oncogenes gain the ability to form tumors at various sites in the animal. Other important functions aiding in the malignant transformation is the ability of the malignant cells to interact with the surrounding stroma and vasculature, but these functions will not be further discussed. Current Perspectives and Future Directions During recent decades we have seen remarkable progress in our understanding of the events that transform a normal cell into a malignant cell. It is now clear that malignant transformation is caused by specific mutations in oncogenes and tumor suppressor genes. We are also starting to comprehend how these mutations lead to the malignant phenotype, as we begin to understand the way in which these genes interact with cellular programmes such as control of proliferation, apoptosis, differentiation and senescence, and the way that mutations in oncogenes and tumor suppressor genes lead to a relaxed control of these processes. To date, this increase in our basic knowledge of malignant transformation has unfortunately been of little help to the vast majority of patients with cancer. Hopefully, we are now starting to obtain the tools to utilise this knowledge in designing more efficient and specific anticancer therapies. The current understanding of the malignant process has, for example, formed the basis of a number of current clinical pilot studies, based on gene-therapeutic approaches. In the future we will probably be able to design chemicals and peptides that can specifically interfere with the functions that are dysregulated in malignant cells. Although no clear data concerning the beneficial effects of such novel treatments exist to date, some studies have shown promising results, giving clear hope for the future. Acknowledgements Supported by grants from the Swedish Cancer Society and the Cancer Society of Stockholm. The author is a recipient of a fellowship from the Stockholm County Council. References 1 Palmero I, Peters G. Perturbation of cell cycle regulators in human cancer. Cancer Surv 1996; 27: 351-367. 2 WyUie AH. Apoptosis and carcinogenesis. Eur J Cell Biol 1997; 73: 189-197. 3 Hayflick L. The limited in vitro lifetime of human diploid strains. Exp Cell Res 1965; 37: 614-636. 4 Campisi J. Rcplicative senescence: An old live's tale?. Cell 1996; 84: 497-500. 5 Cleaver JE. It was a very good year for DNA repair. Cell 1994; 76: 14. 25

26 6 Grafia X, Reddy EP. Cell cycle control in mammalian cells: role of cyclins, cyclin dependent kinases (CDKs), growth suppressor genes and cyclin-dependent kinase inhibitors (CKIs). Oncogene 1995; 11: 211-219. 7 Pardee AB. G-I events and regulation of cell proliferation. Science 1989; 246: 603~:~08. 8 Kamb A, Gruis NA, Weaver-Feldhaus J. A cell cycle regulator potentially involved in genesis of many tumor types. Science 1994; 264: 436--440. 9 Yang E, Korsmeyer SJ. Molecular thanatopsis: A discourse on the BCL2 family and cell death. Blood 1996; 88: 386--401. 10 Whyte M. ICE/CED-3 proteases in apoptosis. Cell Biol 1996; 6: 245-248. 11 Fraser A, Evan G. A license to kill. Cell 1996; 85: 781-784. 12 Vogelstein B, Kinzler KW. p53 function and dysfunction. Cell 1992; 70: 523-526. 13 Olsson I, Bergh G, Ehinger M, Gullberg U. Cell differentiation in acute myeloid leukaemia. Eur J Hematol 1997; 57: 1-16. 14 Sachs L. Control of normal cell differentiation and the phenotypic reversion of malignancy in myeloid leukaemia. Nature 1978; 274: 535. 15 Alcorta DA. et al. Involvement of the cyclin-dependent kinase inhibitor p16 (INK4a) in replicative senescence of normal human fibroblasts. Proc Nat Acad Sci USA 1996; 93: 13742-13747. 16 Harley CB, Villeponteau B. Telomeres and telomerase in aging and cancer. Current Opinion in Genetics & Development 1995; 5: 249-255. 17 Jones JL, Walker RA. Cell-cell and ceu-stromal interactions in breast cancer invasion and metastasis. International Journal of Oncology 1997; 11: 609--616.