* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download molecular mechanisms of malignant transformation
Survey
Document related concepts
Transcript
Oncogene (2004) 23, 6484–6491 & 2004 Nature Publishing Group All rights reserved 0950-9232/04 $30.00 www.nature.com/onc Epidemiology and molecular pathology at crossroads to establish causation: molecular mechanisms of malignant transformation Maurizio Bocchetta*,1 and Michele Carbone1 1 Cardinal Bernardin Cancer Center, Loyola University Chicago, 2160 South First Ave, Maywood IL 60153, USA Epidemiology is a very reliable science for the identification of carcinogens. Epidemiological studies require that the effect, cancer in this case, has already occurred, when of course it would be more desirable to identify potential carcinogenic substances at an earlier stage before they have caused a large number of malignancies and thus become identifiable by epidemiological studies. In the past 30 years, molecular pathology (which includes chemistry, biochemistry, molecular biology, molecular virology, molecular genetics, epigenetics, genomics, proteomics, and other molecular-based approaches) has identified some key alterations that are required for cellular transformation and malignancy. Agents that specifically interfere with some of these mechanisms are suspected human carcinogens. It can be stated that tumor formation requires the following steps: (1) inactivation of Rb and p53 cellular pathways; (2) activation of Ras and/or other growth promoting pathways; (3) inactivation of phosphatase 2A that causes changes in the phosphorylation and activity of several cellular proteins; (4) evasion of apoptosis; (5) telomerase activation or alternative mechanisms of cellular immortalization; (6) angiogenic activity; and (7) the ability to invade surrounding tissues and to metastasize. Here, we review the molecular mechanisms of cellular transformation. The integration of this knowledge with classical epidemiology and animal studies should permit a more rapid and accurate identification of human carcinogens. Oncogene (2004) 23, 6484–6491. doi:10.1038/sj.onc.1207855 Keywords: tumor suppressor genes; senoscence; cell cycle checkpoints; cell immortalization; inflammation Introduction A carcinogen, or initiator, is a factor that causes the molecular damage that leads to cancer. A cocarcinogen is an agent that usually is not carcinogenic by itself but that enhances the activity of a carcinogen (it can also become carcinogenic at high doses or in the presence of *Correspondence: M Bocchetta; E-mail: [email protected] other cocarcinogens or predisposing factors). A tumor promoter is an agent that is not carcinogenic by itself but enhances the activity of an initiator when administered after the initiator. Epidemiology allows the determination of the overall effect of a given carcinogen, cocarcinogen, or promoter in the human population (for example, hepatitits B virus, HBV and hepatocellular carcinoma, HCC), but cannot prove causality in the individual tumor patient. Molecular pathology cannot determine the overall impact of a carcinogen in the population, but can at times prove causality in the individual tumor patient (such as the detection of highrisk human papillomavirus (HPV) in a cervical carcinoma biopsy). This is possible when molecular techniques have shown that the agent is required for transformation or malignant growth of human cells (such as antisense HPV strategies showing the requirement for the expression of HPV proteins for tumor cell growth), and when there is supportive experimental animal evidence. Ideally, epidemiology and molecular pathology information together with experimental evidence in animals should be available for the most reliable identification of human carcinogens. Here, we review the key molecular mechanisms of malignant transformation. Agents that specifically alter some of these mechanisms are those that pose a higher risk for tumor formation (Carbone et al., in press). In higher organisms, individual cells follow a strictly regulated life cycle, so that the architecture and functions of the various tissues can cooperate in harmony during ontogenesis and postnatal life. When cells are artificially removed from their natural context and cultured in vitro, they still behave as members of a regulated population. For example, mammalian fibroblasts stop proliferating when in contact with each other to form a monolayer of cells (this property is termed contact inhibition). Therefore, cells maintain their ‘social’ identity even when they are put in a context that is profoundly different from the body, and they do not proliferate in an unregulated fashion. Cells maintain also a limitation of their proliferative potential. For example, primary human fibroblasts can sustain up to 50 cell divisions before entering the so-called ‘senescence’ (Hayflick and Moorhead, 1961). Introduction into such cells of certain DNA viruses oncoproteins (such as the Simian Virus 40 –(SV40) large T antigen, or Tag) can extend the number of cell doublings up to 20 more cell division, then cells undergo the so-called Molecular mechanisms of malignant transformation M Bocchetta and M Carbone 6485 ‘crisis’, characterized by cell growth arrest and apoptosis (Shay et al., 1991). This emphasizes the fact that cells can tolerate a limited number of cell divisions even in the presence of gene products whose main function is to induce cells into proliferation. Only occasional cells escape crisis and develop into an immortal cell line (Bryan and Reddel, 1994). Malignant neoplasia is a group of different diseases that share some common characteristics: tumor cells are immortal and can proliferate indefinitely. Cancer cells do not stop proliferating when in contact with other cells, but rather infiltrate surrounding tissues. Malignant transformation also involves the capacity of cells to disseminate throughout the organism and form secondary colonies. Cancer cells arise from the malignant transformation of their normal counterparts. This implies that the latter have undergone a number of modifications leading to partial or complete loss of tissue identity. Malignant cells have become members of a different ‘tissue’ that has lost control of proliferation, an event ultimately leading to the death of the entire organism. In the last three decades a wealth of experimental data have revealed many of the molecular pathways that control cell proliferation and differentiation. The analysis of human cancers using genomic analysis has evidenced that the malignant phenotype can be achieved through different combinations of gene mutations. In other words, different combinations of mutations can cause the same type of tumor, although some genetic alterations are either characteristic or predominantly found in certain tumor types. In this section, we will not cover all the pathways that can eventually lead to the malignant transformation of the cell. We will rather discuss some central issues common to most (or all) tumors. In essence, a tumor cell requires gaining one or more chronic proliferative stimuli, evading senescence, acquiring immortality, and interacting with the surrounding tissues, and often the immune system. Gain of function in the pathway of malignant transformation: oncogenes Oncogenes were traditionally recognized as fragments of genetic material carried by viruses (both DNA viruses and retroviruses). Several retroviruses can induce certain types of tumors (mainly sarcomas) in birds, monkeys, mice, and cats. Some examples are represented by the Rous sarcoma virus, the Kirsten murine sarcoma virus, and the Avian erythoblastosis virus. These viruses can induce tumors because they express upon infection proteins that induce the host cell into proliferation. The identification of these viral oncogenes led to the discovery that normal cells contain homologue genes to these viral oncogenes. For this reason, the cellular counterparts of the viral oncogenes where termed protooncogenes. The latter encode cellular proteins that are generally related to the control of proliferation (growth factor receptors, proteins involved in signal transduc- tion, or transcription factors). The oncogene v-erb (encoded by the Avian leukosis virus) is a truncated version of the human EGF receptor. This truncated EGR receptor is constitutively active, and therefore cells expressing v-erb undergo a chronic EGF stimulation (Lax et al., 1985). An example of a viral oncogene coding for a critical component of growth signal transduction pathways is represented by Ha-MuSV (homologue to the cellular gene Ha-ras) (DeFeo et al., 1981), encoded by the Harvey murine sarcoma virus. An example of a viral oncogene encoding a transcription factor is provided by REV-T, homologous to a member of the NFkB protein family (Wilhelmsen et al., 1984; Lee et al., 1991), encoded by the Reticuloendotheliosis virus. The link between viral oncogenes and cellular proto-oncogenes is provided by the fact that mutations in the cellular gene can give rise to constitutively active version of the proto-oncogene that in turn behaves similarly to a viral oncogene. For example, substitution of the glycine at codon 12 of the Ki-ras gene produces a constitutively active, oncogenic ras allele, which is commonly found in human malignancies (Bos et al., 1987). Other cellular genes may become oncogenes as a result of chromosomal deletions or rearrangements, whether they have viral homologues or not. The deletion mutant TAN-1, coding for a truncated version of the Notch-1 gene (Aster et al., 1994), produces a constitutively active Notch-1 protein that is implicated in about 10% of human T-cell acute lymphoblastic leukemias (Ellisen et al., 1991). A hallmark of human chronic myeloid leukaemia is a 9;22 chromosome translocation that generates the so-called ‘Phyladelphia chromosome’. The junction of chromosome 9 and 22 produces a fusion of two genes, with the consequent creation of the bcr-abl oncogene (Hariharan and Adams, 1987). Another mechanism for the transformation of a normal cellular gene into an oncogene can be either overexpression or gene amplification. The best example of this phenomenon is provided by myc. Myc is a transcription factor that regulates the transcription of many downstream genes involved in cell cycle entry and DNA synthesis. Amplification of myc expression levels (due to different reasons, such as chromosomal instability or deregulated transcription) can result in aberrant myc-dependent transcription and consequent induction of the transformed phenotype (Collins and Groudine, 1982). In conclusion, cellular proto-oncogenes are molecules that normally regulate the cell’s response to growth factors and hormones (either as receptors, signal transduction molecules, or transcription factors). Upon gene mutations, which can be caused by environmental carcinogens, these molecules can be transformed into constitutively active factors thus mimicking a chronic stimulation by a growth signal. Although all cancer cells contain one or more active oncogenes, the expression of such molecules into normal cells does not usually lead to increased proliferation. For example, expression of an oncogenic ras allele in both human and mouse primary cells causes complete cell Oncogene Molecular mechanisms of malignant transformation M Bocchetta and M Carbone 6486 growth arrest and senescence (Serrano et al., 1997). Same effects are obtained when constitutively active MEK, a critical component of the mytogen-activated protein-kinase (MAP-K) cascade, is expressed in these cells (MAP-K is one of the major downstream effectors of ras signaling, see Figure 1) (Lin et al., 1998). This indicates that cells have mechanisms in place to respond to the deregulated expression of an oncogene by eliminating that cell. This is just one of the many mechanisms that prevent deregulated cell growth, which can lead to tumor formation. When the same oncogenes are aberrantly expressed in premalignant (or malignant) cells, because those cells have lost the ability to arrest cell growth, the oncogene is able to further drive the expansion of the aberrant cell clone. Oncogenes encoded by DNA tumor viruses are a different class of oncogenes. DNA tumor viruses include members of very different families of viruses. Their genome can be large and encode numerous viral proteins, as in the case of Herpes viruses, or their chromosome can be very small and code for a limited number of proteins, as in the case of HPV and SV40 (Butel, 2000). One common requirement of DNA viruses is to induce the host cell to enter S phase, because these viruses rely on the cellular machinery of DNA replication to different extents for the amplifica- tion of their genome. Different viral proteins that target central functions controlling the G1–S checkpoint achieve this task. DNA viruses’ oncogenes do not have cellular homologues, and are often multifunctional. The SV40 Tag achieves the highest degree of multifunctionality. SV40 is a small virus encoding three capsid proteins, a small ancillary protein of uncertain biological function, and three oncoproteins, namely Tag, the small t antigen (or tag), and the 17 kDa protein (Rundell and Parakati, 2001). The Tag is the major SV40 oncoprotein, because it targets a number of cellular proteins and transactivates a number of cellular promoters. The net result of these activities is the simultaneous inactivation of the two major cellular G1– S checkpoints and the induction of a number of cellular pathways that induce proliferation (Ali and DeCaprio, 2001). Overall, the common characteristic of DNA tumor viruses’ oncogenes is that they do not require mutations to exert their transforming functions, they do not have cellular homologues, and that they are often multifunctional. The latter is the reason why the introduction, for instance, of the Tag into a primary cell does not cause senescence or apoptosis, as often caused by deregulated expression of cellular protooncogenes. This is because Tag targets numerous cell pathways simultaneously. The induction of cellular senescence and cell cycle checkpoints: tumor suppressors Figure 1 Representative schematic of the principal mediators of the ras signaling pathway. The pathway has been divided into four major branches, but these branches should not be considered as independent; rather they are activated in concert, and overlap extensively. For instance, MEK-K1 is involved in the phosphorylation of Jun N-terminal kinase-kinase (JNK-K), and of inhibitors of NFkB (IkB), and is also involved in the activation of extracellular regulated kinase (MEK and ERK). The Raf-1 serine–threonine protein kinase associates with ras upon ras activation and phosphorylates a number of downstream mediators. The end result of the activation of the ras signaling pathway is the activation of several transcription factors that promote cell growth, survival, angiogenesis Oncogene The activation of a cellular oncogene represents a gain of functions. The mutation of one allele of a protooncogene is almost invariably dominant over its wildtype counterparts. The genetic analysis of human cancer has revealed that loss of function is also a common characteristic in the process of malignant transition. Since the loss of both copies of certain genes is often required for tumor formation, these mutations are considered to be recessive (the remaining wild-type allele is dominant over the mutated one), and these genes were named tumor suppressor genes. The reason for the denomination resides in the need for the complete loss of them for tumors to arise; one copy of these genes is still able to suppress malignant transformation. In reality, this is an oversimplification, since haploinsufficiency of certain tumor suppressor genes is compatible with tumor formation (Kwabi-Addo et al., 2001; McLaughlin and Jacks, 2002; Magee et al., 2003). Tumor suppressors include proteins with very different activities. Some are transcription factors, such as p53 (Chene, 2003). Others are proteins that through their interaction with other proteins can either inhibit the activity of kinases involved in cell cycle progression, such as p16INK4A and p21WAF-1 (Malumbres and Barbacid, 2001), or affect the cellular localization of other proteins, such as the members of the 14-3-3 protein family (Hermeking, 2003). Some tumor suppressors are kinases that activate a number of other kinases and transcription factors, such as members of Molecular mechanisms of malignant transformation M Bocchetta and M Carbone 6487 the PI3K-like protein kinases (PIKKs) family (Shiloh, 2003). These include ATM (ataxia-telangiectasia mutated), ATR (ataxia telangiectasia-mutated and Rad3related), and other members. These kinases are critically involved in the signaling of DNA damage and can trigger cell cycle arrest throughout G1 to the G2/M checkpoint. The list of all known tumor suppressor gene products and the description of the intricate interconnections among them is beyond the scope of this section, and the reader is referred to a number of excellent reviews, some of which have been mentioned above. One general feature of tumor suppressors is the participation in the establishment of critical checkpoints during G1/S and G2/M transitions. Tumor suppressors become activated and halt the cell cycle in the presence of aberrant proliferative stimuli, DNA damage, and other ‘out-of context’ biological situations. The cell cycle is arrested to allow the cell to ‘fix’ the problem. If such problem cannot be corrected, the cell enters into senescence or initiates apoptosis. This is the reason why all tumor cells contain mutations in a number of tumor suppressor genes that ultimately disrupt the two major tumor suppressor pathways of the cell: the p53 and Retinoblastoma pathways (Hahn and Weinberg, 2002). p53 activates the G1/S checkpoint mainly by inducing the cyclin-dependent kinase (CDK) inhibitor p21WAF. Alternatively, p53 can induce apoptosis because it regulates the transcription of proapoptotic proteins. Retinoblastoma family member activate the G1/S checkpoint by binding to and inhibiting the E2F transcription factors (the latter positively regulate the transcription of a wide number of genes responsible for cell cycle progression). Retinoblastoma proteins can be inactivated through phosphorylation by CDK (mainly CDK4 and 6) bound to G1 cyclins (cyclins D and E members). The tumor suppressor p16INK4A (similarly to p21WAF) can inhibit CDK–cyclin complexes, therefore re-establishing the G1/S checkpoint (see Figure 2). As mentioned above, introduction of oncogenic ras into primary human cells causes senescence. This phenomenon is accompanied by a dramatic increase of the p16INK4A CDK inhibitor (Wei et al., 2001). The CDK inhibitor p16INK4A binds to CDK4 and CDK6 complexed to D-type cyclins displacing the latter from their association with CDKs. The result of this action is the inactivation of CDKs, which in turn are not able to phosphorylate Retinoblastoma family members and, as a consequence, the cell remains blocked in G1 (Malumbres and Barbacid, 2001). These data underscore the importance of the Retinoblastoma tumor suppressor pathway in the establishment of senescence in human cells. However, p53 is also considered to contribute to senescence in human cells (Macip et al., 2003; Horner et al., 2004), especially in the process of replicative senescence (see below). Loss of p53 is important not only to evade senescence. Tumor cells are characterized by a high degree of chromosomal instability (Rangarajan and Weinberg, 2003). In such environment, an intact p53 pathway would invariably cause G1 arrest or p53-dependent apoptosis. Therefore, it is not surprising that p53 itself is lost in about 60% of all human cancers Figure 2 Representative schematic of the two major checkpoint of the cell cycle and involvement of some of the major tumor suppressor gene products. The ATM and ATR kinases are involved in the signaling of DNA damage. These kinases phosphorylate a number of downstream mediators, including p53 and Chk (Checkpoint kinases 1 and 2). Phosphorylation of p53 activates this transcription factor that in turn promotes synthesis of p21WAF, which inhibits the cyclin/CDK complexes, thus activating the G1/S checkpoint. If DNA damage cannot be repaired, p53 promotes apoptosis. p53 is also involved in the establishment of the G2/M checkpoint. The tumor suppressor p16INK4A is involved in inhibition of CDKs, thus contributing to the G1/S checkpoint. The checkpoint kinases halt the cell cycle by phosphorylating (thus inactivating) cdc25 protein family members. Cdc25 are phosphatases that activated cdc2 complexed with cyclin B. The latter complex is the mytosis-promoting factor (MPF), which is the complex that regulates cell entry into the M phase (Levine et al., 1994). More recent data indicated that tumor suppressor genes are not exclusively proteins involved in checkpoint control, but that different proteins having functions apparently unrelated to the control of the cell cycle can nevertheless be important for cancer prevention. One good example is provided by dyskerin, a protein involved in the modification of uridine residues of ribosomal RNA into pseudouridine and therefore important for ribosomal biogenesis. Dyskerin-deficient mice display signs of human dyskeratosis congenita, a disease characterized by premature aging and cancer susceptibility, and are markedly susceptible to cancer development (Ruggero et al., 2003). Mutations (both point mutations and deletions) are not the only way through which a tumor suppressor gene can be inactivated. In recent years, the importance of epigenetic changes in the establishment of the malignant phenotype has been underscored. Some tumor suppressor genes are not transcribed because of methylation of their promoter. Common examples of tumor suppressor genes that are inactivated because Oncogene Molecular mechanisms of malignant transformation M Bocchetta and M Carbone 6488 of hypermethylation of their promoter are represented by those encoded by the INK4A locus (Myohanen et al., 1998) and RASSF1A (Agathanggelou et al., 2001; Dammann et al., 2001). The expression of tumor suppressors can be lost because of differential splicing. One example is provided by IG20, a splice variant of MADD (MAPK activating death domain-containing protein). IG20 is a protein interacting with the tumor necrosis factor alpha (TNFa) receptor (Al-Zoubi et al., 2001) and TNF-related apoptosis-inducing ligand) (TRAIL) receptor (Efimova et al., 2003). Through its activities, IG20 sensitizes the cell to apoptotic stimuli and induces the secretion of interleukin (IL)-6, thus suppressing tumor cell growth (BS Prabhakar, personal communication). In certain tumors, IG20 expression could be lost because of perturbed splicing of the IG20 gene primary transcript leading to overexpression of an IG20 splice variant (DENN-SV) that induces cell proliferation and resistance to apoptosis (Efimova et al., 2004). Immortality is acquired at chromosomes’ ends The ends of all eucaryotes’ chromosome are organized into structures termed telomeres. Human telomeres are 5–15 Kbp long, and contain tandem repeats of the sequence 50 -TTAGGG-30 (Moyzis et al., 1988). Telomeres end in a 30 overhang (about 300 nucleotides long) that invades the double-stranded portion of telomeres to form a lariat structure also known as T loop (Griffith et al., 1999). Several proteins are bound to the T loop and contribute to its stability and/or signal whether such structure is perturbed. Telomere’s end is structured this way because a terminal DNA sequence of a putative linear chromosome would be detected as a doublestrand break in DNA and it would trigger DNA repair response, including activation of a cell cycle checkpoint and/or apoptosis. Furthermore, unprotected chromosomal ends would lead to end-to-end chromosomal fusion, thus leading to mitosis catastrophe (Goytisolo and Blasco, 2002). One characteristic of DNA replication is that the ends of linear DNA molecules are progressively shortened at each round of replication because DNA polymerase uses RNA primers that are degraded after elongation, a phenomenon known as ‘end replication problem’ (Watson, 1972). Bacteria solve this problem by having a circular genome, while bacteriophages with a linear chromosome arrange it in tandem concatemers, therefore limiting the importance of terminal DNA sequences. In most normal somatic cells, instead chromosomes shorten at each round of cell replication (Harley et al., 1990). This progressive erosion of chromosomes ends ultimately results in progressive shortening of telomeres. When telomeres shorten beyond a critical threshold, they start to become structurally dysfunctional. This process eventually leads to activation of the p53 signaling pathway with consequent establishment of replicative senescence or Oncogene crisis (Chin et al., 1999). Therefore, telomeres work as a cellular hour glass that determines how many replication rounds a cell can afford. Germinal cells and most of tumor cells evade this phenomenon (and therefore achieve immortality) through the activity of a riboprotein complex called telomerase. This enzyme is a reverse transcriptase that adds TTAGGG repeats to the telomeres, thus preventing their shortening and the consequent induction of senescence and apoptosis. Telomerase has an RNA component (TERC) that is universally expressed in human cells, and functions as template for the protein component of telomerase to synthesize TTAGGG tandem repeats (Goytisolo and Blasco, 2002). The protein component of telomerase (TERT, the reverse transcriptase) is not expressed (or expressed at insufficient levels) in most somatic cells. For this reason, somatic cells cannot restore the length of telomeres during cell proliferation and undergo replicative senescence and crisis. As a result of mutations, or of viral infection in certain cells (Foddis et al., 2002), tumor cells regain sufficient amounts of TERT expression and thus can proliferate indefinitely. Some have questioned whether certain cancers (especially breast cancer) need to become immortal in order to become clinically evident (Reddel, 2000). This issue was raised from the observation that breast cancer is a very fastidious tissue to obtain cell lines from, arguing against the case that these tumors are truly made of a uniformly immortal cell population. However, more than 85% of human cancers (including breast cancer) have readily measurable telomerase activity (Shay and Bacchetti, 1997). Many of those cancers in which telomerase activity is not detected have an alternative process to preserve telomeres length called ALT (alternative lengthening of telomeres; Bryan et al., 1997). ALT cancer cells are characterized by unusually long telomeres, and the process underlying the ALT phenotype is believed to be nonreciprocal recombination between telomeres of different chromosomes (reviewed in Neumann and Reddel, 2002). All these evidences indicate that cell immortalization is a necessary event for malignant growth. Dual role in cancer pathogenesis: the immune system Researchers are increasingly focusing on the interplay between cancer cells and the surrounding stromal tissues. It appears more and more evident that the microenvironment surrounding cancer cells plays a pivotal role in tumor promotion, vascularization. and metastatization (Coussens and Werb, 2002). The physiological rearrangements of connective tissue are largely under the control of the various members of the immune system (Schreiber, in press). The fact that the immune system could have played a major role in tumor containment and, at the same time, promotion was long known. In the late 19th century, it was shown that a minority of terminally ill cancer patients showed partial or complete remission of their malignancy when Molecular mechanisms of malignant transformation M Bocchetta and M Carbone 6489 injected with a mixture of bacterial toxins (the so-called Coley extract) in order to cause a potent inflammatory response (Nauts et al., 1953). These data indicated that a massive inflammatory stimulus can trigger a generalized reorganization of the immune systems capable of activating a response against the tumor cells (even if the primary stimulus was not specific for the tumor itself). On the other hand, it is also known that cancer arises preferentially within a context of chronic inflammation. Currently, about 15% of all human tumors are attributable to infectious agents, and inflammation is a critical component of these chronic infections (Parkin et al., 1999). Experimental evidences have underscored that in the early stages of cancer, the presence of consistent intratumoral lymphocyte infiltrations are strongly correlated with reduced metastasis formation and prolonged patient survival (Clark et al., 1989; Naito et al., 1998). Indeed, both innate and adaptive immunity can fight cancer development. Innate immunity cells decipher pattern-recognition of receptors and other transmembrane proteins expressed by cancer cells and directly target them. The main cellular players in this mechanism of recognition are natural killer cells (NK). The latter recognize stress-related proteins on the surface of cancer cells (Diefenbach and Raulet, 2003). NK can also recognize cells with decreased expression of HMC class I, a common feature in cancer cells (Karre, 2002). Dendritic cells (DC) also play an essential role. They can phagocytose apoptotic cancer cells, a process eventually leading to activation of adaptive immunity against cancer cell-specific antigens. DC also phagocytose heat-shock proteins (HSP) released by necrotic cancer cells. HSP released in the extracellular space activate inflammation (Basu et al., 2001). Adaptive immune system targets tumor cells indirectly. Tumor cells potentially express a number of tumor-specific antigens (deriving from mutations of wild-type proteins), and therefore should be easy targets for CD4 þ and CD8 þ T cells. However, cancer cells decrease the level of expression of major histocompatibility complex (MHC) class I molecules and that of critical immuno costimulatory proteins such as the member of the B7 protein family, which are required for efficient Tdependent response (Sharpe and Freeman, 2002). This problem is overcame by DC that, after phagocytosis of tumor cells, debris migrate to lymph nodes and activate a tumor-specific T response (Banchereau and Steinman, 1998). A limitation of this process (termed ‘cross priming’) is that DC infiltrating cancer tissues are often immature, a process mediated by a quite complicated interplay between tumor cells and immune cells in which cytokines play a major role (see below). The importance of the immune system in the protection against tumor formation is underscored also by a number of studies conducted in experimental animals. Mice defective for interferon g (IFN-g, a cytokine produced mainly by T cells and NK and, to a lesser extent, by DC and macrophages), or for other cytokines important for IFN-g production display an increased susceptibility to chemical carcinogenesis (Kaplan et al., 1998; Smyth et al., 2000). IFN-g is a multifactorial cytokine that probably mediates a wide variety of cell effects, but one possible mechanism to explain these data is that IFN-g can upregulate MHC class I molecules and therefore render cancer cells easier targets for T cell-mediated targeting (Shankaran et al., 2001). Other experiments conducted in immunocompromized mice showed a protective role for the immune system against cancer, although other mice models, especially those deficient in the production of certain cytokines yielded opposite effects. These apparently paradoxical results indicate that in cancer biology it is impossible to envision the role of the immune system in just one direction. Despite a wide variety of possible mechanisms of self-defense, the immune system has obviously lost its battle against cancer when the disease finally becomes clinically manifest. The principle of boosting this compromised immune system following a wide variety of strategies in order to fight clinically evident cancer has been the central hypothesis of Cancer Immunology. A number of studies are presently in progress to produce therapeutic cancer vaccines, or to infuse genetically modified immune cells into cancer patients to trigger a cancerspecific immune response. However, immune depressed patients (e.g. AIDS patients, patients who are immunocompromised following organ transplant, etc.) do not develop an excess of malignancies other than those that are virally induced (e.g. EBV-related lymphomas, Kaposi sarcoma, etc.). This indicates that the immune system has a much more limited role in cancer prevention than originally hypothesized and that this role is mainly limited to malignancies that express unique antigens, such as viral proteins. Moreover, also for virally induced cancers, the immune system can promote cancer growth, such as in the case of Hepatitis C infection that leads to chronic inflammation, which in turn can result in the development of HCC (Carbone et al., in press). As stated above, cancer and inflammation have been correlated. A number of pathologies characterized by chronic inflammation are strikingly associated with a higher cancer incidence. Individuals who have taken aspirin and other cyclooxygenases inhibitors over a period of several years have a 40–50% reduced risk of developing colon cancer (Garcia-Rodriguez and HuertaAlvarez, 2001). There are a number of reasons why inflammation could play a landscaping role for cancer development, mainly due to the action of proinflammatory cytokines secreted by various component of the innate immune system, including macrophages, mast cells, neutrophils, eosinophils, and other members. However, tumor-associated macrophages are increasingly implicated as possible tumor-promoting cells. These macrophages produce a wide array of cytokines, proteases, angiogenic factors, all of which may play an important role in tumor progression and dissemination (Schoppmann et al., 2002). Additionally, tumor-associated macrophages secrete the proinflammatory chemokine TNFa (Torisu et al., 2000). This cytokine, through induction of the NFkB signaling pathway, may provide surrounding cells with critical survival Oncogene Molecular mechanisms of malignant transformation M Bocchetta and M Carbone 6490 stimuli in the presence of mutagenic molecules, such as reactive oxygen species (ROS) and other potentially mutagenic molecules produced during infection (Maeda and Akaike, 1998). Another powerful cytokine produced by macrophages and T cells is macrophage migration inhibitory factor (MIF). MIF inhibits p53 transcriptional activity (Hudson et al., 1999). Therefore, this cytokine may promote extended lifespan, accumulation of mutations due to impaired p53 function, inhibition of p53-mediated apoptosis. This situation may explain why inflammation may be fertile ground for cancer development. Inhibition of p53 functions, paralleled by NFkB activation, provide cells with survival and proliferative stimuli in the absence of the major tumor suppressor gene product of the cell. Potentially mutagenic substances produced during infection may introduce mutations into cellular protooncogenes, thus allowing the mutated cell to survive the activities of an active oncogene. Furthermore, as a response to pro-inflammatory cytokines, mast cells release a number of proteases, including heparin, heparanase, matrix metalloproteinases and serine proteases, all thought to play a key role in tissue dynamics and in metastatic dissemination of cancer cells (Balkwill and Mantovani, 2001). A role for certain proinflammatory cytokines in tumor formation has been demonstrated in experimental animals. TNFa knockout mice are resistant to skin cancer induced by different chemical carcinogens (Moore et al., 1999), while mice homozygous for the IL-6-null allele (/) are resistant to plasma cell tumor development (Hilbert et al., 1995). Other proinflammatory cytokines, such as colony-stimulating factor 1 (C-SF1) (Lin et al., 2001) and IL-1 (Voronov et al., 2003), proved to be required for tumor formation in experimental animals. In conclusion, the immune system is one of the most important lines of defense for tumor formation for malignancies that express unique antigens, such as viral proteins. The immune system can also promote tumor growth through the stimulatory and mutagenic effects of cytokines released by inflammatory cells (Wogan et al., in press; Schreiber, in press). Conclusions Our knowledge about the molecular mechanisms required for tumor formation has tremendously improved during the past 40 years. Key molecular mechanisms required for malignancy have been identified (Carbone et al., in press). It can be easily predicted that new oncogenes will be identified in the future; however, it can be predicted that these new ‘cancer genes’ will affect those pathways required for tumor formation described above. Thus, the apparent complexity of tumor formation can be reduced to a much simpler picture. Regardless of the specific mechanism by which a given gene pathway is altered, carcinogenesis requires the activation or inactivation of the pathways we have discussed in this review. Cancer is a multifactorial event in which numerous alterations contribute to the emergence of the malignant cell. Tumor cells per se can be innocuous unless additional factors such as tumor promoters stimulate cell growth (hormones, chronic inflammation, etc.). Moreover, tumor cells that express viral antigens or some other type of unique antigen can be eliminated by the immune system. Therefore, malignant tumor growth is a dynamic process in which it is difficult to identify a unique event that caused the process. Rather, it is possible to identify numerous events that together contributed to malignancy. This information should allow scientists to identify those agents that are more likely to cause or to contribute to cancer before they have caused a substantial number of cancer deaths detectable by classical epidemiology. Thus, it is hoped that the integration of molecular pathology with classical epidemiological and animal studies will lead to a more rapid and accurate identification of human carcinogens. Acknowledgements Dr Carbone’s work is supported by grants from the National Cancer Institute, USA, RO1CA92657; the American Cancer Society, RSG-04-029-01-CCE, and by the Cancer Research Foundation of America. References Agathanggelou A, Honorio S, Macartney DP, Martinez A, Dallol A, Rader J, Fullwood P, Chauhan A, Walker R, Shaw JA, Hosoe S, Lerman MI, Minna JD, Maher ER and Latif F. (2001). Oncogene, 20, 1509–1518. Ali SH and DeCaprio JA. (2001). Semin. Cancer Biol., 11, 15–23. Al-Zoubi AM, Efimova EV, Kaithamana S, Martinez O, ElIdrissi Mel-A, Dogan RE and Prabhakar BS. (2001). J. Biol. Chem., 276, 47202–47211. Aster J, Pear W, Hasserjian R, Erba H, Davi F, Luo B, Scott M, Baltimore D and Sklar J. (1994). Cold Spring Harb. Symp. Quant. Biol., LIX, 125–136. Balkwill F and Mantovani A. (2001). Lancet, 357, 539–545. Banchereau J and Steinman RM. (1998). Nature, 392, 245–252. Basu S, Binder RJ, Ramalingam T and Srivastava PK. (2001). Immunity, 14, 303–313. Oncogene Bos JL, Fearon ER, Hamilton SR, Verlaan-de Vries M, van Boom JH, van der Eb AJ and Vogelstein B. (1987). Nature, 327, 293–297. Bryan TM, Englezou A, Dalla-Pozza L, Dunham MA and Reddel RR. (1997). Nat. Med., 3, 1271–1274. Bryan TM and Reddel RR. (1994). Crit. Rev. Oncog., 5, 331– 357. Butel J. (2000). Carcinogenesis, 21, 405–426. Carbone M, Klein G, Gruber J and Wong M: Cancer Res. (in press). Chene P. (2003). Nat. Rev. Cancer, 3, 102–109. Chin L, Artandi SE, Shen Q, Tam A, Lee SL, Gottlieb GJ, Greider CW and DePinho RA. (1999). Cell, 97, 527–538. Clark Jr WH, Elder DE, Guerry IV D, Braitman LE, Trock BJ, Schultz D, Synnestvedt M and Halpern AC. (1989). J. Natl. Cancer Inst., 81, 1893–1904. Collins S and Groudine M. (1982). Nature, 298, 679–681. Molecular mechanisms of malignant transformation M Bocchetta and M Carbone 6491 Coussens LM and Werb Z. (2002). Nature, 420, 860–867. Dammann R, Yang G and Pfeifer GP. (2001). Cancer Res., 61, 3105–3109. DeFeo D, Gonda MA, Young HA, Chang EH, Lowy DR, Scolnick EM and Ellis RW. (1981). Proc. Natl. Acad. Sci. USA, 78, 3328–3332. Diefenbach A and Raulet DH. (2003). Curr. Opin. Immunol., 15, 37–44. Efimova E, Martinez O, Lokshin A, Arima T and Prabhakar BS. (2003). Cancer Res., 63, 8768–8776. Efimova EV, Al-Zoubi AM, Martinez O, Kaithamana S, Lu S, Arima T and Prabhakar BS. (2004). Oncogene, 23, 1076– 1087. Ellisen LW, Bird J, West DC, Soreng AL, Reynolds TC, Smith SD and Sklar J. (1991). Cell, 66, 649–661. Foddis R, De Rienzo A, Broccoli D, Bocchetta M, Stekala E, Rizzo P, Tosolini A, Grobelny JV, Jhanwar SC, Pass HI, Testa JR and Carbone M. (2002). Oncogene, 21, 1434–1442. Garcia-Rodriguez LA and Huerta-Alvarez C. (2001). Epidemiology, 12, 88–93. Goytisolo FA and Blasco MA. (2002). Oncogene, 21, 584–591. Griffith JD, Comeau L, Rosenfield S, Stansel RM, Bianchi A, Moss H and de Lange T. (1999). Cell, 97, 503–514. Hahn WC and Weinberg RA. (2002). Nat. Rev. Cancer, 2, 331–341. Harley CB, Futcher AB and Greider CW. (1990). Nature, 345, 458–460. Hariharan IK and Adams JM. (1987). EMBO J., 6, 115–119. Hayflick L and Moorhead PS. (1961). Exp. Cell. Res., 25, 585–621. Hermeking H. (2003). Nat. Rev. Cancer, 3, 931–943. Hilbert DM, Kopf M, Mock BA, Kohler G and Rudikoff S. (1995). J. Exp. Med., 182, 243–248. Horner SM, DeFilippis RA, Manuelidis L and DiMaio D. (2004). J. Virol., 78, 4063–4073. Hudson JD, Shoaibi MA, Maestro R, Carnero A, Hannon GJ and Beach DH. (1999). J. Exp. Med., 190, 1375–1382. Kaplan DH, Shankaran V, Dighe AS, Stockert E, Aguet M, Old LJ and Schreiber RD. (1998). Proc. Natl. Acad. Sci. USA, 95, 7556–7561. Karre K. (2002). Scand. J. Immunol., 55, 221–228. Kwabi-Addo B, Giri D, Schmidt K, Podsypanina K, Parsons R, Greenberg N and Ittmann M. (2001). Proc. Natl. Acad. Sci. USA, 98, 11563–11568. Lax I, Kris R, Sasson I, Ullrich A, Hayman MJ, Beug H and Schlessinger J. (1985). EMBO J., 4, 3179–3182. Lee JH, Li YC, Doerre S, Sista P, Ballard DW, Greene WC and Franza Jr BR. (1991). Oncogene, 6, 665–667. Levine AJ, Perry ME, Chang A, Silver A, Dittmer D, Wu M and Welsh D. (1994). Br. J. Cancer, 69, 409–416. Lin AW, Barradas M, Stone JC, van Aelst L, Serrano M and Lowe SW. (1998). Genes Dev., 12, 3008–3019. Lin EY, Nguyen AV, Russell RG and Pollard JW. (2001). J. Exp. Med., 193, 727–740. Macip S, Igarashi M, Berggren P, Yu J, Lee SW and Aaronson SA. (2003). Mol. Cell. Biol., 23, 8576–8585. Maeda H and Akaike T. (1998). Biochemistry, 63, 854–865. Magee JA, Abdulkadir SA and Milbrandt J. (2003). Cancer Cell, 3, 273–283. Malumbres M and Barbacid M. (2001). Nat. Rev. Cancer, 1, 222–231. McLaughlin ME and Jacks T. (2002). Cancer Cell, 1, 408–410. Moore RJ, Owens DM, Stamp G, Arnott C, Burke F, East N, Holdsworth H, Turner L, Rollins B, Pasparakis M, Kollias G and Balkwill F. (1999). Nat. Med., 5, 828–831. Moyzis RK, Buckingham JM, Cram LS, Dani M, Deaven LL, Jones MD, Meyne J, Ratliff RL and Wu JR. (1988). Proc. Natl. Acad. Sci. USA, 85, 6622–6626. Myohanen SK, Baylin SB and Herman JG. (1998). Cancer Res., 58, 591–593. Naito Y, Saito K, Shiiba K, Ohuchi A, Saigenji K, Nagura H and Ohtani H. (1998). Cancer Res., 58, 3491–3494. Nauts H, Fowler G and Bogatko F. (1953). Acta Med. Scand., 144, S1–S103. Neumann AA and Reddel RR. (2002). Nat. Rev. Cancer, 2, 879–884. Parkin DM, Pisani P, Munoz N and Ferlay J. (1999). Infections and Human Cancer Newton R, Beral V and Weiss RA (eds). Cold Spring Harbor Laboratory Press: Cold Spring Harbor, NY. Rangarajan A and Weinberg RA. (2003). Nat. Rev. Cancer, 3, 952–959. Reddel RR. (2000). Carcinogenesis, 21, 477–484. Ruggero D, Grisendi S, Piazza F, Rego E, Mari F, Rao PH, Cordon-Cardo C and Pandolfi PP. (2003). Science, 299, 259– 262. Rundell K and Parakati R. (2001). Semin. Cancer Biol., 11, 5–13. Schoppmann SF, Birner P, Stockl J, Kalt R, Ullrich R, Caucig C, Kriehuber E, Nagy K, Alitalo K and Kerjaschki D. (2002). Am. J. Pathol., 161, 947–956. Schreiber H: Semin. Cancer Biol. (in press). Serrano M, Lin AW, McCurrach ME, Beach D and Lowe SW. (1997). Cell, 88, 593–602. Shankaran V, Ikeda H, Bruce AT, White JM, Swanson PE, Old LJ and Schreiber RD. (2001). Nature, 410, 1107–1111. Sharpe AH and Freeman GJ. (2002). Nat. Rev. Immunol., 2, 116–126. Shay JW and Bacchetti S. (1997). Eur. J. Cancer, 33, 787–791. Shay JW, Wright WE and Werbin H. (1991). Biochim. Biophys. Acta, 1072, 1–7. Shiloh Y. (2003). Nat. Rev. Cancer, 3, 155–168. Smyth MJ, Taniguchi M and Street SE. (2000). J. Immunol., 165, 2665–2670. Torisu H, Ono M, Kiryu H, Furue M, Ohmoto Y, Nakayama J, Nishioka Y, Sone S and Kuwano M. (2000). Int. J. Cancer, 85, 182–188. Voronov E, Shouval DS, Krelin Y, Cagnano E, Benharroch D, Iwakura Y, Dinarello CA and Apte RN. (2003). Proc. Natl. Acad. Sci. USA, 100, 2645–2650. Watson JD. (1972). Nat. New Biol., 239, 197–201. Wei W, Hemmer RM and Sedivy JM. (2001). Mol. Cell. Biol., 21, 6748–6757. Wilhelmsen KC, Eggleton K and Temin HM. (1984). J. Virol., 52, 172–182. Wogan GN, Hecht SS, Felton JS, Conney AH and Loeb LA: Sem. Cancer Biol. (in press). Oncogene