|
|
The Tumor Suppressor Laboratory
The proto-oncogenic RAS family of GTPases consists of at least four members: c-Ha-RAS, c-Ki-RAS, N-RAS and M-RAS. These small G-proteins are activated when in the GTP-bound state, initiating a downstream signalling cascade. Once the GTP is hydrolysed to GDP by the intrinsic GTPase activity of normal RAS with the aid of GAPs (GTPase activating proteins) such as NF1, the signalling cascade is again switched off. Activation of RAS occurs through stimulation of the cells by mitogenic growth factors, such as the EGF and PDGF receptors following binding of their respective ligands. The resultant phosphorylation at the key Tyr residues leads to the activation of specific GDP/GTP exchange factors such as SOS which replace GDP with GTP on RAS and hence to its activation. However, when these RAS proteins are oncogenically mutated, their GTPase activity is no longer stimulated by GAPs. As a consequence these mutants remain in the GTP-bound form, and thereby become constitutively activated. Such mutations are associated with more than 30% of all human cancers, notably more than 90% of pancreatic and 50% of colon cancers, and results in the activating of the downstream effectors such as the kinase RAF and PI-3 kinase in a growth factor independent manner. For the past decade our group has been studying the pathways downstream of RAS and we have discovered several distinct molecules that selectively block these. One class of such potential anti-RAS molecules are the minimal RAS-binding fragments of RAF and NFI (RAF81 and NF56). These molecules block the interaction of RAS with all downstream effectors. Another anti-RAS molecule is a large GTPase and GAP called p190-A. Its GAP domain attenuates the activity of Rho family G-proteins (Rho, Rac and CDC42) which are mainly activated by PI-3 kinase. A third class of anti-RAS molecules is a group of actin/PIP3-binding proteins such as NF2, HS1 and tensin which are controlled by both PIP3, the end-product of PI-3 kinase and Rac. Over-expression of p190-A or these actin-binding proteins blocks the oncogenic PI-3 kinase/Rac/CDC42 pathways. Furthermore, we have found two distinct actin-binding compounds, Chaetoglobosin K (CK) and MKT-077, which also suppress RAS transformation in vitro and in vivo respectively. These molecules exert their activity by capping and bundling actin filaments. Shortly after we and others found that the Rho family of GTPases are essential for RAS transformation, we focused our major efforts on novel, cell-permeable peptides or synthetic chemical compounds that could block the oncogenic pathways downstream of these GTPases. In particular, the pathway we chose was CDC42 and its effector PAK. PAK is a CDC42/Rac-activated Ser/Thr kinase that is essential for actin-based membrane ruffling. As we discuss in more detail in a following research report, we found that WR-ACK42, a CDC42-binding peptide, and WR-PAK18, a peptide that blocks the PAK-PIX interaction, suppresses RAS transformation in a cell culture system. Moreover, using chemical inhibitors specific for EGF receptor family members or the non-receptor Src family Tyr kinases, we found that oncogenic RAS required both autocrine pathways. The activation of PAK is essential for RAS-induced malignant transformation and requires pathways elicited by the ErbB1/EGF receptor, ErbB2/Neu and a Src family kinase. These findings suggest a new avenue for the treatment of RAS-associated cancers by blocking these downstream autocrine pathways. In fact, a combination of specific chemical inhibitors for these Tyr kinases such as PP1 and AG 879 almost completely suppresses the rapid growth of RAS-induced sarcomas in mouse models. Furthermore, in collaboration with an organic chemistry group at Yale, we have found the first PAK-specific chemical inhibitor that directly binds the unique ATP-binding pocket of PAK family kinases. Suppression
of RAS transformation by blocking Rho family GTPases-mediated pathways:
RAS
activates actin-based membrane ruffling through PI-3 kinase and RAC, stimulating
the production of PIP2 that uncaps the plus end of actin filament (F-actin),
and induces a rapid actin polymerization. To determine whether the PIP2-induced
uncapping of actin filament/membrane ruffling is essential for RAS transformation,
we have examined the effect of an F-actin capping drug called CK (discovered
by Horace Cutler in 1980) on RAS transformation. Like the over-expression
of an F-actin capping protein called tensin, treatment of RAS transformants
with 2 uM CK blocks their membrane ruffling and strongly suppress their
anchorage-independent growth in soft agar, suggesting that the uncapping
of actin filament is required for RAS transformation. Downstream
of RAC: F-actin bundling drug MKT-077 In
1998 we published that over-expression of an F-actin bundling protein
called HS1 suppresses RAS transformation. Interestingly, we found recently
that an anti-cancer drug called MKT-077, which was originally developed
as a photo-sensitizer by Tadao Shishido's group of Fuji-Photo Film, selectively
suppresses RAS transformation and binds two distinct proteins in RAS transformants.
(1) actin and (2) a heat shock protein called HSC70/Mot-2. Furthermore,
we found that MKT-077 bundles actin filaments (and consequently blocks
membrane ruffling), and also blocks the interaction between Mot-2 and
the tumour suppressor p53. Renu Wadhwa's group published that Mot-2 inactivates
the transcription factor activity of p53 (inducing the CDK inhibitor p21/WAF1)
by keeping p53 in the cytoplasm. Her group then found that MKT-077 reactivates
p53 by blocking its interaction with Mot-2. Interestingly, RAS activates
CDKs by upregaulating the cyclin D1 gene (and therefore antagonizing p53
which inhibits CDKs through p21). Thus, MKT-077 suppresses RAS transformation
through two pathways. (1) blocking membrane ruffling and (2) reactivating
p53. ACKs are Tyr kinases that only bind CDC42 in the active (GTP bound) form. Using the CDC42-binding peptide of 42 amino acids derived from ACKs called ACK42 as a ligand specific for the CDC42/GTP complex, we provided the first direct evidence that RAS activates CDC42. Furthermore, we found that RAS transformation is suppressed by either over-expression of ACK42 in RAS transformants or treatment of RAS transformants with a cell-permeable derivative called WR-ACK42 (WR is a peptide vector of 16 amino acids that can deliver any small peptides into target cells). Both inhibitors block the interaction of CDC42 with downstream effectors, such as ACKs, PAKs and N-WASP (consequently blocking the CDC42-induced micro-spike formation). The anti-RAS cancer activity of ACK42 depends on its affinity for CDC42, and by replacing Arg 34 by Lys in the ACK42, we have created a mutant of ACK42 that binds CDC42 15 times more efficiently. Downstream
of RAC/CDC42: PAK18 (18 amino acid peptide of PAK) PAK is a Ser/Thr kinase that is activated by Rac or CDC42. Since both Rac and CDC42 are activated by RAS, RAS can activate PAK. However, PAK requires another protein called PIX for full activation. The SH3 domain of PIX binds a Pro-rich domain of 18 amino acids in PAK. This domain, called PAK18, can alone block both the PAK-PIX interaction, and RAS-induced activation of PAK (and membrane ruffling). We have found that a cell-permeable derivative of PAK18 (WR-PAK18) suppresses RAS transformation. In addition, oncogenic RAS mutants, such as v-Ha-RAS, require two independent autocrine pathways that involve the ErbB1/EGF receptor, ErbB2 and Src family kinases for both PAK activation and malignant transformation. This indicates that oncogenic RAS acts upstream of these Tyr kinases through the EGF family of ligands that are up-regulated by RAS. These findings are contrary to the conventional view that both ErbB family and Src family kinases act upstream of RAS and therefore open a new avenue for the treatment of RAS-associated cancers by inhibiting these Tyr kinases. Downstream
of RHO/RAC/CDC42: p190-A (GTPase & GAP) We have previously shown by expressing sense and anti-sense rat p190-A cDNAs in mouse fibroblasts, that p190-A is a tumour suppressor which blocks RAS trans-formation through its two independent domains. Both the N-terminal GTPase domain, and the C-terminal GAP domain that attenuates Rho family GTPases appear to be involved. To localize p190-A, we cloned the full-length human p190-A cDNA and, using this cDNA as a probe, Karen Arden's group at the UCSD Branch of the Institute mapped this gene to chromosome region 19q13.3. Interestingly, this region is often deleted in gliomas, astrocytomas and carcinomas, suggesting that dysfuntion or loss of p190-A might be responsible for the development of these cancers in human. Downstream
of RAC: NF2 (suppressor of JNK kinase) We have previously shown that over-expression of the tumour suppressor NF2 (either full-length, N-terminal half or C-terminal half) suppresses RAS transformation. Although we found that the N-terminal half (NF2-N) binds actin filament and PIP2, the target of the C-terminal half (NF2-C) remained unknown. As part of a collaboration with Christoph Renner, during his time visiting our branch, we found that the tumour suppressor GPS2 is a binding partner of NF2-C using a yeast two-hybrid system. Interestingly, GPS2 has previously been shown to suppress RAS transformation by blocking the Rac-induced activation of JNK kinase. Subsequently, we found that NF2-C also blocks RAS/Rac-induced activation of JNK kinase, suggesting that the GPS2-NF2 complex is likely responsible for the regulation of JNK kinase. Anti-RAS
cancer potential of PP1, a specific SRC family kinase inhibitor The interaction between PAK (p21-activated Ser/Thr kinase) and PIX is required for v-Ha-RAS malignant transformation and PAK activation. Oncogenic RAS activates PAK kinase through an autocrine mechanism. This pathway involves at least two distinct receptor kinases: the EGF receptor (ErbB1) and ErbB2. Both ErbB1 and ErbB2 activate SRC family kinases, which phosphorylate Cat, another binding partner of PIX. To determine whether Src family kinases are required for RAS-induced PAK activation and malignant transformation, we studied the effect of PP1, a specific Src family kinase inhibitor, on the anchorage-dependent growth of normal and v-Ha-RAS transformed NIH 3T3 fibroblasts, PAK activation and anchorage-independent growth of RAS transformed cells. In addition, the development of RAS-induced sarcomas in nude mice was also evaluated. We found that PP1 (10nM) inhibits PAK activity in RAS transformed cells by 60%. Although PP1 has no effect on the anchorage-dependent growth of RAS-transformed cells, it significantly inhibits the anchorage-independent growth in soft-agar, as well as the RAS-induced sarcomas in mice. These findings suggest that PP1 and other Src family kinase inhibitors may potentially be useful for the treatment of RAS-associated cancers. PAK
activation and malignant transformation by oncogenic RAS via ErbB family-dependent
autocrine pathways Oncogenic RAS causes dramatic changes in the actin cytoskeleton organization. It induces the disruption of stress fibers and induces membrane ruffling and micro-spike formation. Rho family GTPases, including Rho, Rac and CDC42, have been shown to regulate actin-cytoskeletal organization and RAS activates Rac through the PI-3 kinase pathway. Both Rac and CDC42 are required for RAS transformation of NIH3T3 cells. The
PAK Ser/Thr kinase family is activated by both Rac and CDC42 following
binding and phosphorylation. Activation of PAK causes accumulation of
F-actin, the formation of membrane ruffles/lamelipodia and microspikes/filopodia.
PAK binds the SH3 domain of PIX (PAK-interacting exchange protein) which
has been reported to activate Rac. An 18 amino acid proline rich motif
(residues 186-203, PAK18) of PAK is required for PAK to bind PIX. The
PAK-PIX interaction is required in both Rac-induced membrane ruffling
in PC12 cells and PAK kinase activation in COS-1 cells.
|
||||||||||||||||||||||||||||||||
