Conservation of Proto-Oncogenes
Homologues of proto-oncogenes have been found in all multicellular animals studied thus far, and their widespread distribution in nature indicates that their protein products have essential biologic roles. The more highly conserved domains of the protein are probably those that have a crucial structural and/or functional role, and characterization of their normal biochemical properties will provide insight into the contribution that an activated oncogene has to cell transformation. Understanding the mechanism of activation of each oncogene requires characterization of the proto-oncogene, a comparison of the changes that have occurred, and systematic testing of changes influencing the transforming potential.
There are essentially only three biochemical mechanisms by which proto-oncogenes act. One mechanism involves phosphorylation of proteins on serine, threonine, or tyrosine residues.41 Proteins of this class transfer phosphate groups from adenosine triphosphate (ATP) to the side chain of tyrosine, serine, or threonine residues. Phosphorylation serves two basic purposes in signal transduction. In many instances, it changes the conformation and activates the enzymatic kinase activity of the protein. Secondly, phosphorylation of tyrosine residues generates docking sites that recruit target proteins, which the activated kinase may phosphorylate. Thus, phosphorylation acts to potentiate signal transmission through the generation of complexes of signal-transducing molecules at specific sites in the cell where they are required to act. For example, activation of the catalytic activity of a receptor tyrosine kinase by its ligand leads to the formation of a complex of signaling proteins at the plasma membrane where the receptor is localized.
The second mechanism by which genes act to transmit signals is by GTPases.42 The prototype for this class of proteins is the ras gene family. In a similar manner to the kinase gene family, Ras proteins function as molecular switches that are turned of and on via a regulated GDP/GTP cycle. Ras proteins have been implicated as key intermediates that relay the signal from upstream tyrosine kinases to downstream serine-threonine kinase pathways. Some of the conventional heterotrimeric G proteins can also transform cells when altered.43,44 The third mechanism involves proteins that are localized in the nucleus. A large variety of proteins that control progress through the cell cycle and gene expression are encoded by proto-oncogenes, some of which may also be involved in DNA replication.45,46 Thus, the relaxation of requirements of transformed cells for growth factors could be mediated by an activated oncogene at multiple levels of the signal transduction pathway.
Growth factors are responsible for stimulating cells in a resting or G0 stage to enter the cell cycle. This mitogenic response occurs in two phases: a quiescent cell is stimulated to proceed into the G1 phase of the cell cycle by competence factors and then becomes committed to DNA synthesis by progression factors (Fig. 25-5).47 Successful transition through the G1 phase requires sustained growth factor stimulation over a period of several hours. This is followed by a critical phase where the presence of a progression factor such as insulin growth factor I is required in addition to the growth factor for successful progression through the cell cycle. This dual signal requirement may prevent accidental triggering of quiescent cells into the cell cycle by transient exposure to mitogenic growth factors. In some cell types, the absence of growth factor stimulation causes the rapid onset of programmed cell death (apoptosis).48 Therefore, inappropriate expression of a growth factor may result in a constant stimulation of cell growth in addition to a block in cell differentiation.
Combinatorial interactions of cyclins and cyclin-dependent kinases (cdks) during the cell cycle. Progression from G0 through the restriction point in G1 requires the continued presence of growth factors. This requirement is overcome by oncogenes, such as Myc, Ras, or Raf, or tyrosine kinases. Progression through G1 can be blocked by antimitogens, TGFβ, or the p53 or Rb tumor-suppressor genes. Cyclins D and E in complexes with cdks are required for progression through G1 and entry into S phase. Cyclins A and B form complexes with cdk2 later in the cell cycle and are involved in progression from G2 to M. The activity of cyclin-D complexes can be inhibited by p15 and p16, thus preventing the advance of the cell from G1 to S. Other inhibitors of cyclin-dependent kinases, p21 and p27, can act throughout the cell cycle. Cyclins and inhibitors found altered in human cancers are shown in grey.
There is now much evidence to support a role for growth factors and their receptors in the development of human malignancies. The first direct correlation of an oncogene with a growth factor was revealed from a computer-assisted comparison that showed that the amino acid sequence of the v-sis oncogene product was highly related to the B chain of platelet-derived growth factor (PDGF) (Table 25-1).49 PDGF is released from platelets during clotting and is recognized as an important serum mitogen required for mesenchymal cell growth in culture. Connective tissue tumors such as sarcomas and glioblastomas have been shown to express PDGF, whereas their normal tissue counterparts did not.50 Thus, in an autocrine fashion, the sarcoma and glial tumor cells appear to synthesize the mitogen to which they are normally responsive. On the other hand, no genetic alterations of the PDGF gene have yet been observed that would explain this synthesis. Until the mechanism underlying the expression of PDGF is clarified, it will be difficult to know whether this expression is a cause or an effect of neoplastic growth.
Int-2, whose expression is activated by mouse mammary tumor virus (MMTV) proviral insertion in mouse mammary carcinomas, the KS3 oncogene identified in a Kaposi sarcomas, and HST, a transforming gene identified in a human stomach cancer by DNA transfection, are members of the basic or acidic fibroblast growth factor (FGF) family of related peptide mitogens. Basic FGFs are expressed by human melanoma cell lines but not by normal melanocytes that are dependent on bFGF to proliferate.51,52 Similarly, transforming growth factor α (TGFα) is frequently produced by carcinomas that express high levels of the EGFR and appears to function as an autocrine in this system through stimulation of the EGFR.53 Because many ligands and their receptors are not yet characterized, the contribution of autocrine growth stimulatory loops to human malignancies may be greater than is presently appreciated.
Oncogenes derived from growth factor receptors confer on the cells the ability to bypass the growth factor requirement, rendering cells growth factor independent. By far, the largest number of receptor-derived oncoproteins are derived from growth factor receptors that have tyrosine kinase activity.
Tyrosine Protein Kinases.
More than 40 different protein tyrosine kinases have been identified (Table 25-1).54 These kinases are subdivided into two main categories: those, such as the EGFR, spanning the plasma membrane and those located in the cytoplasm, many of which are associated with the plasma membrane, such as c-Src (Fig. 25-6). All of the protein tyrosine kinases have sequence homology over a region of approximately 300 amino acids that has been defined as the catalytic kinase domain 55 (Fig. 25-6). This domain is responsible for catalyzing the transfer of the phosphate group of ATP to tyrosine residues during trans- and autophosphorylation. This kinase domain is also homologous to the Raf, ERK, and Mos protein members, which have phosphorylation specificity for serine and threonine.56,57 Phosphorylation on tyrosine is a rare event in normal cells and accounts for only 0.05 percent of all protein phosphorylation, but tyrosine kinases regulate key events in signal transduction pathways that control cell shape and growth.58
Schematic comparison of structural features of cell surface growth factor receptor tyrosine kinases and membrane-associated tyrosine kinase oncogene products. The cytoplasmic tyrosine kinase domain is represented in grey (src homology 1 or SH-1), and SH-2 and SH-3 domains are indicated. A: Ligand binding to the epidermal growth factor receptor (EGFR) promotes receptor dimerization, activation of the kinase, and transphosphorylation of the receptor cytoplasmic domain on tyrosine residues, which then interact with SH-2 domain containing substrates and elicit an intracellular signal. The N- and C-terminal deletions, in addition to point mutations within critical domains of the molecule, that activate v-erbB are illustrated. Autocrine production of TGFβ in cells that express the EGFR results in constitutive activation of the EGFR. B: A single point mutation in the transmembrane domain of the Neu/HER2 oncogene product is sufficient for ligand-independent activation. This mutation promotes receptor dimerization and kinase activation in the absence of ligand. Similarly, the loss of a single cysteine residue in the extracellular domain of the Ret receptor in multiple endocrine neoplasia type-2A syndrome results in a constitutively activated receptor. This mutation frees a cysteine residue that is normally involved in intrareceptor disulphide bond formation and enhances receptor dimerization presumably by the formation of intermolecular disulphide bonds promoting constitutive activation of the receptor catalytic activity. C: The gene rearrangements that activate the Met and Trk and Ret receptor-derived oncogene products. The majority of receptor oncogenes activated by gene rearrangement are fused with a protein domain capable of protein-protein interaction and thus mediate dimerization and constitutive activation of the kinase in the absence of ligand. D: The complex formed between CD4 and the Src family kinase Lck. The protein domains through which Lck and CD4 interact are represented as open boxes in these molecules. The c-Src kinase is maintained in an inactive conformation through the interaction of a negative regulatory phosphotyrosine residue in the C-terminus of c-Src with the Src SH-2 domain. The v-src oncogene is generated following deletion of this negative regulatory C-terminal tyrosine residue such that the kinase is now in an unconstrained conformation and is constitutively active. Following activation of many growth factor receptor tyrosine kinases, the c-Src SH-2 domain interacts with a phosphorylated tyrosine residue on the receptor with a greater affinity than its C-terminal phosphotyrosine residue (see Fig. 25-8). This relieves the negative regulation of c-Src and activates the kinase. EGF = epidermal growth factor; HGF/SF = hepatocyte growth factor-scatter factor; NGF = nerve growth factor; GDNF = glial cell derived nerve growth factor, TGFβ = transforming growth factor β.
Growth Factor Receptor Tyrosine Kinases.
Many growth factors mediate their effects by means of receptors with tyrosine kinase activity. These receptors have an extracellular ligand-binding domain, a single transmembrane domain, and an intracellular catalytic domain responsible for transducing the signal59 (Fig. 25-6). Binding of growth factors to cell surface receptors results in receptor dimerization and activation of their intrinsic tyrosine kinase, leading to intermolecular phosphorylation of each receptor on specific tyrosine residues.60 This results in the recruitment of signaling molecules containing Src homology 2 (SH2) domains, 61 which recognize, in a sequence specific manner, short peptide segments containing phosphotyrosine within activated receptors62-64 (Fig. 25-7). These include proteins with enzymatic activity, such as phospholipase Cγ (PLCγ), phosphatidylinositol 3-kinase (PI3K), and the GTPase-activating protein for p21Ras (p120GAP), and proteins that lack enzyme activity and function as adaptor proteins, such as Grb2. Collectively, they act to transmit signals that mediate the pleiotrophic responses to growth factors (reviewed in Pawson, 199565 and Hunter, 200066). PLCγ hydrolyzes phosphoinositols and thereby generates diacylglycerol and inositol-3-phosphate. PI3K phosphorylates phosphoinositides and generates putative second messengers for cytoskeletal rearrangements and cellular trafficking. By contrast, Grb2 lacks enzymatic activity and contains only SH2 and SH3 domains. SH3 domains also function as protein-protein interaction motifs and recognize proline-rich domains in other proteins.67 Grb2 serves as an adaptor that binds to phosphotyrosines in activated receptors via its SH2 domain and recruits the Ras nucleotide exchange factors Son of sevenless (mSos-1 and mSos-2) to the receptor via its SH3 domain. Sos catalyzes exchange of GDP bound to Ras for GTP and is considered to be the most important step in Ras activation.68,69 These proteins, through a series of protein-protein interactions, in part mediated through SH2 domains and SH3 domains, form signaling complexes downstream of receptor tyrosine kinases (reviewed by Pawson & Saxton, 199970). Many of these signaling complexes regulate the activity of serine/threonine kinases, which in turn regulate through phosphorylation the activity of transcription factors. A generalized scheme for receptor signaling is presented in Figure 25-7.
Substrates and mitogenic signaling for receptor tyrosine kinases (RTKs). Activation of receptor tyrosine kinases by the binding of growth factors stimulates cross phosphorylation of their kinase domains. Substrates containing SH-2 domains bind to phosphorylated tyrosine residues on activated receptors. Substrates shown to directly interact with and become activated by or phosphorylated by RTKs; GAP (GTPase-activating protein), PLCγ (phospholipase Cγ), PI-3 kinase (phosphatidylinositol 3′ kinase), and Src (tyrosine kinase) are represented. The Grb2 adaptor protein that lacks enzymatic activity acts as a bridge to translocate the Sos guanine nucleotide exchange protein to the plasma membrane, where it stimulates the exchange of GDP for GTP, activating the Ras protein (see in detail in Fig. 25-8). Activation of Ras translocates the serine threonine kinase Raf to membrane, where it is activated by an unknown mechanism. Activated Raf then activates kinases of the MAP kinase pathway that stimulate the phosphorylation of transcription factors and expression of the fos and jun transcription factors. Secondary responses involve the breakdown of phosphatidylinositol 4,5 bis-phosphate into diacylglycerol (DAG) and inositol triphosphate (InsP3), which stimulate protein kinase C and calcium release, respectively. Components of the signal transduction pathway that have been identified as oncogenes are shown in grey.
A number of oncogenes encode mutant forms of receptor tyrosine kinases. These include receptors for known factors, such as the epidermal growth factor receptor (EGFR) (v-erbB), colony-stimulating factor-1 receptor (CSF-1R) (v-fms), hepatocyte growth/scatter factor receptor (met), nerve growth factor receptor (trk), stem cell factor receptor (SCFR) (kit), neuroregulin receptor (Neu/HER2), glial derived neurotopic factor (GDNF) receptor (ret), and receptor-like proteins with unknown ligands (ros) (Table 25-1). In the receptor-related oncogenes so far examined, the structural changes that activate the transforming potential appear to deregulate the receptor kinase activity, and these oncoproteins transform by delivering a continuous ligand-independent signal (reviewed by Rodrigves & Park, 1994).71 Receptors isolated as retrovirally transduced oncogenes, such as v-erbB (EGFR), v-fms (CSF-1R), and v-kit (SCFR) (Table 25-1), frequently sustain deletions of the extracellular ligand-binding domain in combination with other structural alterations, such as C-terminal deletions or mutations that remove negative regulatory domains and render the receptor catalytically active in the absence of ligand.71)
Mutations that promote ligand-independent dimerization represent a general mechanism for oncogenic activation of receptor tyrosine kinases. A single point mutation in the transmembrane domain of the Neu/HER2 oncogene product is sufficient for ligand-independent activation. This mutation promotes receptor dimerization and kinase activation in the absence of ligand 72,73 (Fig. 25-6). Similarly, the loss of a single cysteine residue in the extracellular domain of the Ret receptor in multiple endocrine neoplasia type 2A syndrome results in a constitutively activated receptor. This mutation frees a cysteine residue that is normally involved in intrareceptor disulphide bond formation and enhances receptor dimerization by the formation of intermolecular disulphide bonds promoting constitutive activation of the receptor catalytic activity.74,75 A similar mechanism has been demonstrated for the activation of the Neu/HER2 receptor in experimentally induced mammary neoplasias.76 Alternatively, activation of the Ret receptor in multiple endocrine neoplasia type 2B occurs by a single point mutation in the kinase domain that increases the basal kinase activity and alters the substrate specificity of the receptor, thus altering the signal transduction pathways activated by the receptor.77,78
A growing class of receptor tyrosine kinase oncogenes in human tumors, including TRK, MET, RET, and the platelet-derived growth factor receptor (PDGFR), are also rendered constitutively active following genomic rearrangements that juxtapose novel sequences derived from unrelated loci with the kinase domain of the receptor (Fig 25-7).79 The majority of receptor oncogenes activated by gene rearrangement are fused with a protein domain capable of protein-protein interaction and thus mediate dimerization and constitutive activation of the kinase in the absence of ligand.71 Receptor fusion oncogenes involving RET and TRK have been detected in papillary thyroid carcinomas.80-82 In these tumors, both RET (10q11.2)83 and TRK (1q21)83 are rearranged with loci from the same chromosome, such as H4-ret (D10S170, Ptc) 10q2183 and tropomyosin or tpr/trk (1q21-31).84 Thus, small intrachromosomal deletions or inversions that are not readily detected cytogenetically may be a common event in the oncogenic activation of these receptor kinases in human tumors.
In addition to structural rearrangements, growth factor receptors are frequently amplified and overexpressed in human tumors. The EGFR is overexpressed in squamous cell carcinomas and gliomas, 36,37 Neu/HER2 in adenocarcinomas of the breast stomach and ovary, 38-40 MET (hepatocyte growth factor receptor, HGFR) in human stomach and some colon carcinomas, 85,86 and bek, a member of the FGFR family, in human stomach carcinomas87 (Table 25-4).
Nonreceptor Protein Tyrosine Kinases.
The protein products of v-src, v-fes, v-fps, v-fgr, v-yes, and lck are associated with the plasma membrane but are not transmembrane proteins (Fig. 25-6 and Table 25-1). Many of these proteins have a myristilated N-terminal glycine residue that promotes association with the plasma membrane. The cytoplasmic tyrosine protein kinase domain is in the C-terminus of the protein, and all of the tyrosine kinase oncogene proteins are homologous in this region. The src subfamily has additional regions of homology not found in the receptor family. These regions include two additional domains named src homology 2 and 3 (SH-2 and SH-3) (with SH-1 defined as the kinase domain) (Fig. 25-6).88 As already discussed, the src homology 2 domain is highly conserved in proteins involved in signal transduction and recognizes phosphotyrosine residues, whereas the SH3 domain recognizes proline-rich motifs present in signaling proteins and cytoskeletal proteins.
Association of src family kinases with the plasma membrane is essential for their transforming activity. Mutation of the myristoylation signal in these proteins abolishes membrane association and transforming activity, indicating that their signal must be initiated at the plasma membrane, perhaps through interaction with other membrane-bound proteins.56,89 Oncogenic activation of src-like kinases as retrovirally transduced oncoproteins occurs through the acquisition of point mutations and/or deletion of negative regulatory protein domains located at the C-terminus. These alterations generate oncoproteins that phosphorylate cellular proteins on tyrosine residues in an unregulated fashion and thus deliver a continuous, rather than a regulated, signal. The src or other src-like kinases are essential components of mitogenic signaling pathways downstream from receptor tyrosine kinases.90 Moreover, src kinase activity is activated by receptor tyrosine kinases, 91 suggesting that src or src-like kinases would be activated in tumors where receptor tyrosine kinases are deregulated. Indeed, activation of c-src has been observed in mammary tumors in transgenic mice induced by an oncogenic Neu/HER2 receptor.92
The abl tyrosine kinase constitutes a separate family of nonreceptor tyrosine kinases that are localized to both the nucleus and the cytoplasm. The Bcr-Abl product has been implicated in the pathogenesis of greater than 95 percent of CML. In a manner similar to receptor fusion oncoproteins, Bcr mediates oligomerization of Bcr-Abl, promoting constitutive activation of the Abl kinase93 and association with downstream signaling molecules. Moreover, the Bcr-Abl oncoproteins are excluded from the nucleus, which may also prevent interactions with substrates that act to regulate cell growth negatively.94
Cytoplasmic Adaptor Proteins.
The discovery that some oncogenes encode adaptor proteins that lack contain only SH2 and SH3 domains but lack any catalytic activity has enabled a more complete understanding of the role of these proteins in signal transduction. For example, the v-crk oncogene product95 contains only SH2 and SH3 domains but causes an increase in tyrosine phosphorylated proteins in the cell.96 Crk and other SH-2/SH-3 domain-containing adaptor proteins bind to phosphorylated tyrosine residues on activated receptor tyrosine kinases or other proteins via their SH2 domains and bind to other proteins via their SH3 domains. In this manner, they bring together heteromeric protein complexes that allow the subsequent phosphorylation of proteins in this complex by the kinase.97,98 This phosphorylation event, in addition to dephosphorylation events, acts as a mechanism to relay the signal from the cell surface to the nucleus.
Proteins with GTPase Activity.
The role of proteins with GTPase activity in tumorigenesis was first identified through the discovery of the ras oncogenes that encode a previously unknown form of GTPase. Three ras gene family members designated c-Ha-ras, c-Ki-ras, and N-ras involved in malignant transformation have been identified by their presence in rapidly transforming retroviruses and by DNA transfection. In normal cells, members of the ras family have been highly conserved throughout evolution and encode cytoplasmic proteins of 21,000 daltons (p21ras).99 Ras proteins are posttranslationally targeted to the plasma membrane through a highly conserved sequence at their N- and C-termini. Membrane association is essential for function of Ras proteins. Certain domains in the Ras proteins are homologous to the subunit of trimeric G proteins, in regions involved in guanine nucleotide binding, and Ras proteins have been shown to bind guanine nucleotides (GTP and GDP).99 The model proposed for the p21Ras proto-oncogene product is that it exists in equilibrium between two conformations: active, with GTP bound; and inactive, with GDP bound (Fig. 25-8).
Model for regulation of the Ras p21 product and for the GTPase-activating protein (GAP) as a downstream effector and regulator of ras activity. Ras is localized to the inner aspect of the plasma membrane. The alternating relaxed (GDP bound) and activated (GTP bound) states of the p21 Ras protein are shown in normal cells. Conversion of GDP- to GTP-bound forms is the rate-limiting step. Binding of the Grb2 adaptor protein to a specific tyrosine-phosphorylated residue on an activated (growth factor stimulated) receptor tyrosine kinase translocates the Sos guanine nucleotide exchange factor to the plasma membrane, where it stimulates the exchange of GDP for GTP on Ras. Activation of Ras alters its conformation and enables it to interact with and recruit the Raf serine-threonine kinase to the membrane where it becomes activated by an unknown (not Ras) mechanism. Activation of Raf activates the downstream MAP kinase signaling pathway involved in the mitogenic response. In addition, activation of Ras stimulates changes in cell shape and motility mediated through Rho-like GTPase proteins that are part of the Ras superfamily of small GTPase proteins. Inactivation of Ras is in part controlled by the intrinsic intrinsic Ras GTPase, catalyzed by GTPase-activating proteins (GAP and NF1). Oncogenic p21 Ras proteins with mutations at amino acid positions 12, 13, 59, or 61 remain in their active GTP-bound states and constitutively activate downstream signaling pathways.
In the past 10 years, components of signal transduction pathways have been elucidated and place Ras as a crucial regulator of cell shape, motility, and growth downstream from growth factor receptors (Fig. 25-7). Activation of Ras is coupled to ligand stimulation of growth factor receptors and is mediated by a guanine nucleotide exchange factor (GEF) (Sos). In the case of receptor tyrosine kinases, activation of Ras is mediated by binding of the Grb2 adaptor molecule via its SH2 domain to a specific phosphorylated tyrosine residue on the receptor. Grb2 pulls along the Sos protein, thus localizing it to the plasma membrane where its substrate Ras is localized. Sos then stimulates the exchange of GDP for GTP on Ras, converting Ras from an inactive to an active form100,101 able to interact with an effector/substrate molecule(s)102 (Fig. 25-8). Conversely, the inactivation of Ras is mediated in part by the intrinsic GTPase activity of Ras. Usually this activity is low; however, the GTPase activity is stimulated by a GTPase-activating protein GAP, which converts the active GTP-bound form of Ras into the inactive GDP-bound form.103-105 GAP also contains an SH2 domain and is recruited by activated receptor tyrosine kinases to the plasma membrane.
The activation and inactivation of Ras proteins are carefully orchestrated. The conversion of Ras to the GTP-bound state enables it to interact with other proteins that function as downstream effectors for Ras. One effector for Ras is the serine-threonine kinase Raf. Activation of Ras recruits the Raf kinase to the membrane where it is activated. In turn, Raf activates a linear signaling pathway involving a series of mitogen activated protein (MAP) kinases that culminates in the expression and activation of transcription factors Fos and Jun (Fig. 25-7). These in turn form the AP1 transcription factor that induces transcription of the c-MYC transcription factor, which in turn regulates genes whose products control cell cycle progression, culminating in one round of DNA replication and cell division.
The Raf protein kinase was independently isolated as a retrovirally transduced oncoprotein v-raf (Table 25-1). Inhibition of the Raf signaling pathway by specific inhibitors or through the use of dominant negatively interfering mutants106 blocks transformation of fibroblasts in culture by an oncogenic Ras protein. In addition to Raf, transformation of cells by an oncogenic Ras also requires the activity of members of the Rho family of GTPases: Rho and Rac.107,108 Members of the Rho family of GTPases are involved in rearrangements in the actin cytoskeleton and are thought to regulate the morphologic changes in cell shape associated with transformation of cells by Ras.109,110
The RAS oncogenes have been identified in a variety of tumors. As discussed previously, the oncogenic forms of Ras differ from their normal counterparts by mutations that result in amino acid substitutions at positions 12, 13, or 61 in the phosphate-binding domain of the protein (reviewed by Barbacid111). These oncogenic Ras proteins are locked in their active GTP-bound state through an increased exchange of GDP for GTP or through an inability to interact with or be dephosphorylated by GAPs.112,113 They therefore have a reduced requirement for GDP/GTP exchange factors and no longer require activation by the Sos exchange factor.
Multiple GAP proteins that function to switch the Ras signal off have been identified. The p120 GAP and neurofibromin (NF1) were the first to be discovered. The p120 GAP appears to control the response of Ras to growth factor stimulation, whereas NF1 appears to control basal Ras activity.114,115 In humans, loss of the GAP protein NF1 results in the disease neurofibromatosis type 1. Aspects of the disease can be explained in terms of elevated Ras-GTP and are thought to be the result of the loss of neurofibromin GAP activity.116,117
Support for Rho-like GTPases in cell transformation and tumorigenesis is also provided by the discovery that multiple independently isolated oncogenes act as exchange factors for Rho-type GTP-binding proteins. Dbl, Vav, Ect-2, Ost, Tiam, Lbc, Lfc, and Dbs were discovered by gene transfer methods by virtue of their ability to transform fibroblasts in culture118 (Table 25-1). Tiam-1, which appears to directly influence the invasive capacity of T-lymphoma cells, was identified adjacent to a proviral insertion site in retrovirally induced invasive T-lymphoma varients.119 The DBL oncogene was the first member of this family to be identified.120 Amino acid sequence analysis revealed that Dbl shared homology with a yeast cell division cycle protein Cdc 24, which is an exchange protein for a yeast Rho-like small GTP-binding protein.121 The homology was restricted to the domain of Cdc24 responsible for the GEF activity, and this result led to the discovery that Dbl is an exchange factor for a mammalian Rho-like protein (Cdc42).121 All of the oncogenic exchange proteins for small GTP-binding proteins that have been identified, contain a similar domain that is now referred to as a Dbl homology domain. The Dbl domain is essential for the transforming activity of this class of oncogenes, suggesting that their exchange activity for Rho-like GTP-binding proteins is essential for transformation. A full characterization of the mechanism of oncogenic activation of this family of oncoproteins has not been achieved, but the deletion of N-terminal sequences of Dbl or Vav result in oncogenic activation.122,123 These deletions may remove a negative regulatory domain that normally acts to regulate the GEF activities of these proteins. The current thinking is that Dbl and related proteins activate Rho-like GTP-binding proteins that play important roles in mediating various cytoskeletal reorganizations in cells. Unlike the signaling cascade in which the Ras GEF Sos participates, which binds to the adaptor protein Grb2 and translocates Sos to cell surface receptor tyrosine kinases to activate Ras in response to growth factors, the signaling complexes coupling Dbl-like GEFs to upstream components remain elusive.
Dbl-like GEFs act as exchange factors for members of the Rho family of small GTPases. These proteins are structurally related to the Ras family of GTPases. Members of the Rho family of small GTPases Rho, Rac, and Cdc42 are important regulators of the actin cytoskeleton. Microinjection of activated proteins into serum-starved fibroblasts demonstrates that Rho stimulates the organization of actin stress fibers, Rac stimulates the formation of lamellipodia, and Cdc42 induces filopodial formation.124 Because of their effects on motile cytoskeletal structures, it is not hard to envision that inappropriate activation of Rho family of GTPases during tumorigenesis could affect whether a cell maintains a differentiated morphology or acquires a motile, invasive phenotype.
Cytoplasmic Serine-Threonine Protein Kinases
The serine-threonine kinases studied so far are soluble cytoplasmic proteins. This class includes the Mos, Cot, Pim-1, and Raf oncogenes in addition to protein kinase C (PKC) (Table 25-1). Oncogenic forms of the v-raf serine kinase have lost N-terminal regulatory sequences that lead to constitutive activation of the kinase activity and mitogenic MAP kinase pathway (Fig. 25-7). Phosphorylation of the c-Raf protein kinase is normally tightly regulated. Raf kinase activity is rapidly elevated when resting cells are stimulated by mitogens125 (Fig. 25-7) or by another member of the serine-threonine kinase family, PKC. Several tumor promoters act via stimulation of the PKC family126 and mediate activation of the Raf signaling pathway. Although a mutant form of PKCα has been detected as an oncogene and although overexpression of PKC can affect the growth of cells in culture, 127 mutant PKC enzymes are rare in human cancers.
The products of oncogenes and proto-oncogenes localized to the nucleus are directly implicated in the control of gene expression involved in cellular proliferation and differentiation. Many of these have been shown to act as transcription factors and appear to be constitutively activated forms of their normal cellular counterparts (reviewed by Lewin, 1991128). For example, a complex between c-Jun129 and c-Fos130 corresponds to the mammalian transcription factor AP-l, which interacts following phorbol ester treatment or serum stimulation of cells with specific promoter elements to stimulate gene transcription.131,132 The oncogenic Jun and Fos transcription factors carry mutations that lead to loss of negative regulatory elements, and these factors are now constitutively active.133 In addition to loss of negative regulatory domains, some oncogenic transcription factors lose positive effector domains, resulting in dominant negative proteins that appear to prevent expression of genes required for cell differentiation.134
Since many nuclear oncogenes have been implicated in trans-activating and/or trans-repressing gene expression, it is possible that alteration of these genes either directly (activated c-myc, v-jun, or v-erbA) or indirectly (e.g., induction of their expression by an activated growth factor receptor) may lead to an imbalance in the delicate network of gene expression that regulates cell differentiation and growth control. Consistent with the hypothesis that nuclear oncogenes have central roles in events involved in cellular proliferation, the proto-oncogene forms of these genes are normally expressed in a variety of cell types during proliferation and have RNA and protein products with short half-lives. Because of the lability of the RNA and protein products, changes in transcription could lead to relatively rapid fluctuations in the steady-state levels of RNA and protein.133 For example, c-fos and c-myc are expressed in replicating cells, but their expression is negligible in quiescent cells or during terminal differentiation.134,135 When quiescent murine fibroblasts are stimulated with serum or growth factors to enter the Gl phase of growth, a transient increase in the levels of c-myc, c-fos, c-jun, and c-myb is observed (Fig. 25-5).133-135 It is now accepted that these proto-oncogenes are required for cells to transit from a resting state (G0) to a state in which proliferation can proceed (Gl) (c-fos, c-jun, and c-myc) and to traverse specific points in the cell cycle.
The retroviruses that have transduced myc, myb, and fos express these genes in infected cells at levels higher than their cellular counterparts and in a nonregulated manner. Similarly, the amplification of the c-MYC locus in human tumors or the rearranged c-MYC locus in Burkitt lymphomas is no longer subjected to control, and these genes are expressed constitutively. Thus, the unregulated and/or ectopic expression of these genes in a differentiated cell substitutes for the growth factor requirement for quiescent cells to enter G1 and provides a constant proliferative signal in the absence of growth factors.
The identification of new oncogenes and tumor-suppressor genes can lead to the delineation of new signaling pathways involved in cancer. The recent dissection of the function of the adenomatous polyposis coli tumor suppressor and β-catenin has identified a critical role for these proteins in tumorigenesis. β-Catenin is found in two distinct multiprotein complexes in the cell. One complex located at the plasma membrane couples cadherins (calcium-dependent adhesion molecules) with the actin cytoskeleton, stabilizing cell-cell adhesion.136 The other complex, which contains the adenomatosis polyposis coli (APC) protein, a serine-threonine kinase (glycogen synthase kinase 3β), and another protein called Axin, targets β-catenin for degradation.137,138 In the absence of functional APC, β-catenin is not degraded and free β-catenin acts to enhance transcription through its interaction with a transcription factor: LEF1/TCF. This process is normally regulated by extracellular signals. Mutations in the APC gene that cause loss of function are found in multiple human tumors, thus stabilizing β-catenin and enhancing transcription.139 Defects in the APC gene are responsible for inherited and sporadic forms of colon cancer and may account for up to 80 percent of the cancers in this tissue, implicating β-catenin-dependent signals as a key event in colon carcinogenesis.140 Recently, mutations have also been identified in β-catenin that prevent its degradation induced by a wild-type APC protein.141-143 These have been identified in human melanomas, hepatocellular carcinomas, cancers of the uterine and ovarian endometrium, and a small subset of colon cancers. Where studied, mutations in β-catenin and APC are mutually exclusive (reviewed by Polakis, 1999144). The APC gene is targeted for mutation in human colon cancer far more frequently than is β-catenin. This might relate to the specific dietary or environmental insults that give rise to these mutations. In contrast in experimentally induced intestinal tumors induced in rats, the majority contained mutations in β-catenin but not APC.145 Once activated, β-catenin can constitutively interact with multiple different targets, thus sending persistent signals that override normal cell growth control. Activation of gene transcription through LEF/TCF is probably involved in the cancer process, as the proto-oncogene myc has recently been identified as one of the targets of the APC pathway.146
In addition to protein kinases acting as signal relays important for the control of cell growth and tumorigenesis, lipid kinases have emerged as controlling many cellular processes.147,148 One subfamily of lipid kinases include phosphoinositide 3-kinases (PI3′K), which catalyze the addition of a phosphate molecule specifically to the 3-position of the inositol ring of phosphoinositides.149 Phosphatidylinositol 3,4,5-phosphate acts as a ligand for some proteins that contain a phospholipid-binding PH domain.150 One of these is a serine-threonine kinase (PDK1) that activates the Akt kinase.151 Activation of Akt results in the phosphorylation of multiple proteins, some of which act to suppress apoptosis152 whereas others stimulate a mitogenic response (Fig. 25-7).153 Notably, the PI3′K and Akt gene β were identified as retroviral oncoproteins and the P13′K gene is amplified in human ovarian cancer.154 In addition to enzymes that phosphorylate lipids, an enzyme called PTEN was recently identified that dephosphorylates PI-3,4,5. This acts to antagonize the activity of PI3′K, and loss of a functional PTEN gene has been demonstrated in multiple human tumors (reviewed by Cantley & Neel, 1999154).