Repression of TGF-B Signaling by the Oncogenic Protein SKI in Human Melanomas: Consequences for Proliferation, Survival, and Metastasis

Repression of TGF-b signaling by the oncogenic protein SKI in human melanomas: consequences for proliferation, survival, and metastasis

Estela E Medrano*,1

1Hufﬁngton Center on Aging and Departments of Molecular and Cellular Biology and Dermatology, Baylor College of Medicine, Houston, TX, USA

Transforming growth factor-b (TGF-b) has dual and transcription factors are also located in c-Ski, mouse paradoxical functions as a tumor suppressor and promoter Ski, and human SKI proteins. These domains include of tumor progression and metastasis. TGF-b-mediated from the amino-terminus: a proline-rich area, helix– growth inhibition is gradually lost during melanoma tumor loop–helix motifs, a cysteine/histidine-rich area, a region progression, but there are no measurable defects at the of basic amino acids, and finally a leucine zipper-like receptor level. Furthermore, melanoma cells release high domain (Figure 1). The amino-terminal residues 75–304 levels of TGF-b to the microenvironment, which upon are sufficient for the transforming activity of c-Ski activation induces matrix deposition, angiogenesis, survi- (Zheng et al., 1997a). Efficient homodimerization of Ski, val, and transition to more aggressive phenotypes. The mediated by a bipartite C-terminal domain consisting of SKI and SnoN protein family associate with and repress five tandemrepeats (TR) and a leucine zipper (LZ), the activity of Smad2, Smad3, and Smad4, three members correlates with efficient DNA binding and cellular of the TGF-b signaling pathway. SKI also facilitates cell- transformation (Sutrave et al., 1990; Colmenares et al., cycle progression by targeting the RB pathway by at least 1991; Heyman and Stavnezer, 1994; Zheng et al., two ways: it directly associates with RB and represses its 1997a, b). activity when expressed at high levels, and indirectly, it c-ski appears to be phosphorylated on serine residues, represses Smad-mediated induction of p21Waf-1. This and the carboxyl-terminal region is either the site of results in increased CDK2 activity, RB phosphorylation, phosphorylation or is required for phosphorylation and inactivation. Therefore, high levels of SKI result in (Sutrave et al., 1990). However, neither the c-ski kinases, lesions to the RB pathway in a manner similar to p16INK4a nor the consequences of the putative phosphorylations loss. SKI mRNA and protein levels dramatically increase have been studied in detail. Transfected c-Ski localizes during human melanoma tumor progression. In addition, to the nucleus of interphase cells (Colmenares et al., the SKI protein shifts from nuclear localization in 1989). This leads to the notion that c-Ski might function intraepidermal melanoma cells to nuclear and cytoplasmic in transcription. It is important to mention, though, that in invasive and metastatic melanomas. Here, I discuss the endogenous SKI in melanoma tumors can also localize basis for repression of intracellular TGF-b signaling by to the cytoplasm(Reed et al., 2001). In mitotic cells, SKI, some additional activities of this protein, and endogenous or transfected SKI is diffusely expressed propose that by disrupting multiple tumor suppressor throughout the cytoplasmwith a high intensity around pathways, SKI functions as a melanoma oncogene. the chromosomes. Such localization suggests that SKI Oncogene (2003) 22, 3123–3129. doi:10.1038/sj.onc.1206452 could also participate in chromatin condensation and/or chromosome segregation. This hypothesis is based on Keywords: melanocytes; melanoma; Ski; c-ski; Smad2; the association of SKI with protein complexes contain- Smad3; RB; p21Waf-1; SnoN ing histone deacetylases (Nomura et al., 1999). Over- expression of c-Ski by retroviral vectors induces transformation, anchorage independence, and muscle The Ski protein differentiation in avian cells (Colmenares et al., 1991). In turn, SKI dramatically accelerates cell-cycle progres- The ski oncogene was discovered in its viral form, v-ski, sion, and increases colony size and colony formation of as the transforming gene of the defective Sloan Ketering a human melanoma cell line with poor proliferative Virus (Stavnezer et al., 1981). The human SKI protein is capacity (UCD-Mel-N) (Medrano, unpublished results). 91% homologous to v-ski (Nomura et al., 1989). The These results correlate with the ability of SKI to repress chicken ski protein (c-ski) shares amino-acid sequences the retinoblastoma protein, a crucial regulator of the and biological activity with the c-myc oncogene (Stav- mammalian cell cycle (Tokitou et al., 1999). nezer et al., 1989). Putative structural domains found in SKI has been recently reassigned to chromosome 1 p36 (Colmenares et al., 2001; Kokura et al., 2001), a *Correspondence: EE Medrano, Baylor College of Medicine, One common region of alterations in human cancers includ- Baylor Plaza M320, Houston, TX 77030, USA; ing melanomas (Kaghad et al., 1997). The human SKI E-mail: [email protected] produces multiple forms of mRNAs, with two major Repression of TGF-b signaling EE Medrano 3124 Ski transforming activity (76-304)

Smad2(241) Smad3(441) Smad4(203-239)

61 69 126 136 172 205 242 296 398 493 509 572 629 666 708

Pro Zn AHs Zn Basic Ser/Thr α H αH728aa PML(261-330)PML

MeCP2(197-330)MeCP2 MeCP2(556-728)MeCP2

N-CoR(99-274) mSin3 mSin3

Figure 1 Domains of the human SKI protein. Pro: a proline-rich area; Zn: a leucine zipper-like domain; AHs: helix–loop–helix motifs; Basic: a region of basic amino acids; aH: a unique tandem repeats of alpha helical domains which may be involved in the dimerization of the SKI family through coiled-coil interactions. Arrowheads indicate three tandem repeats of 25 amino acids located at residues 572–645. SKI domains required for association with the proteins Smad2, Smad3, Smad4, PML, MeCP2, N-CoR, and mSin3 are indicated in the ﬁgure

species of 6.8 and 5.3 kb which are not due to alternative 1997), and it disrupts dorsal–ventral specification of splicing, within the coding region. The large message is neuroectodermand mesodermin zebrafish (Kaufman coexpressed at significantly higher levels than the small et al., 2000). At the cellular level, SKI represses TGF-b one (Namciu et al., 1994). Ski transcripts are detected in signaling (discussed in detail in the next sections), the the mouse embryo at 8.5–9.5 days postcoitum, during retinoblastoma protein RB (Tokitou et al., 1999), and migration of neural crest cells including melanocytes retinoic acid receptor signaling (Dahl et al., 1998a). and dorsal root ganglia (Lyons et al., 1994). Cell lines SKI is a component of the histone deacetylase that express SKI include neuroblastoma, carcinoma of complex that it is involved in transcriptional repression the vulva, stomach, thyroid, lung, prostate (Nomura (Nomura et al., 1999). SKI associates with the et al., 1989), megakaryocyte ((Namciu et al., 1994), and promyelocytic leukemia (PML) protein, and the cor- melanoma (Reed et al., 2001). epressors N-CoR, mSin3, and HDAC1. Such interac- During the isolation of human SKI,aSKI-related tion is required for transcriptional repression mediated novel gene (sno) was identified (Nomura et al., 1989). by the tumor suppressor Mad (Khan et al., 2001). Two sno cDNAs, SnoAandSnoN, representing alu- However, SKI directly associates to N-CoR/SMRT and containing and non-alu-containing sequences, respec- mSin3, but not with HDAC (Figure 1). The stoichio- tively, were discovered. SKI and SnoN share a large metry between N-CoR and SKI appears to be important region of homology in the amino terminus, and the for repression, as overexpression of SKI partly abro- biological activities of these related oncogenes appear to gates such function. SKI is also required for transcrip- be similar (Pearson-White and Crittenden, 1997). SnoN tional repression mediated by MeCP2, a member of the induces transformation and myogenic conversion of family of methyl-CpG-binding proteins (Figure 1) quail embryo cells when expressed at high levels (Boyer (Kokura et al., 2001). Except for the mouse studies, it et al., 1993). When coexpressed in vitro, c-ski and is important to mention, though, that the SKI activities c-SnoN preferentially formheterodimers. In vivo, c-ski described above were observed in vitro by overexpres- and c-SnoN formheterodimersthat bind the GTCTA- sion of SKI and other proteins. It remains to be GAC element. Tethered c-ski : Sno heterodimers that established how SKI alters the stoichiometry of a lack TR/LZ domains are more active in cellular variety of repressor complexes when it is endogenously transformation than either their monomeric counter- overexpressed. The only human tumors identified to parts, tethered ski : ski homodimers or full-length SnoN date that overexpress SKI are malignant melanomas and c-ski (Cohen et al., 1999). (Reed et al., 2001). Ski has been a lasting enigma for more than a decade. However, in the last four years work fromseveral laboratories has unraveled developmental and cell- autonomous pathways regulated by Ski. This protein Ski null mice: questions and unexplained paradoxes regulates neurulation in the mouse (Berk et al., 1997; Colmenares et al., 2001); overexpression of X-ski into SkiÀ/À mice die shortly after birth and present defects in Xenopus endodermal cells results in the formation of an neurulation, craniofacial patterning, and skeletal devel- ectopic neural-crest-like structure (Amaravadi et al., opment (Berk et al., 1997; Colmenares et al., 2001).

Oncogene Repression of TGF-b signaling EE Medrano 3125 Perinatal lethality results fromexcencephaly caused by been proposed to be critical for the development of deep failed closure of the cranial neural tube during neurula- invasion and metastases in malignant melanoma (Reed tion. Apparently, no gross defects were detected in other et al., 1994). neural-crest derived cells. These results are consistent TGF-b signals are transduced by membrane serine/ with a model in which Ski activities are required for the threonine kinase ‘type I’ and ‘type II’ receptors (TbRI successful expansion of a subset of precursors in the and TbRII). Signaling by TGF-b receptors is mediated neuroepithelial and skeletal muscle lineages. In agree- by the Smads, a family of structurally related proteins. ment with these conclusions, SkiÀ/À melanocytes isolated Ligand-mediated receptor activation results in carboxy- fromthe skin of pups delivered by cesarian section at terminal phosphorylation of Smad2 and Smad3, forma- embryonic day 18.5 can be readily established in culture; tion of heterotrimeric complexes with the common indicating that Ski is not essential for melanocyte partner Smad4, nuclear translocation, and transcrip- migration, proliferation, or differentiation (Medrano tional activation of TGF-b target genes (reviewed by and Colmenares, unpublished). In turn, mice lacking Piek et al., 1999; Massague and Chen, 2000; Derynck Sno die at an early stage of embryogenesis, indicating et al., 2001). that Sno is required for blastocyst formation (Shinaga- In 1998, the Stavnezer lab using an in vitro binding wa et al., 2000). The Ski and SnoN heterozygous mice assay demonstrated that c-ski, in association with are an example of as yet unexplained paradoxes of Ski unknown proteins, binds to the DNA consensus and Sno functions. Although Ski and Sno are considered sequence GTCTAGAC (Nicol and Stavnezer et al., to be oncogenes, when challenged with carcinogens 1998). This group also found that c-ski represses heterozygous Sno+/À mice show increased number of transcription either through upstreamcopies of this lymphomas compared to wild-type mice (Shinagawa element or when brought to the promoter by a et al., 2000); a phenotype shared by SKI+/À animals heterologous DNA-binding domain (Nicol and Stavne- (Shinagawa et al., 2001). These results apparently zer, 1998). The GTCTAGAC sequence, which is bound contradict previous results showing that Ski is a to c-ski-containing proteins was subsequently identified transforming protein (Colmenares et al., 1991; Dahl as a Smad-binding element (SBE) containing the four et al., 1998a). However, the transforming activities of base pairs (50AGAC-30 or its reverse complement Ski and Sno might be restricted to lineages that express 50AGAC-3’) that are directly bound by Smad3 and basal levels of these proteins including skeletal muscle Smad4 proteins (Zawel et al., 1998). and neural crest. Shortly after, several groups including ours demonstrated, through a variety of approaches that included affinity chromatography, GST pull-downs, and yeast two-hybrid screening (Akiyoshi et al., 1999; Luo et al., c-Ski is a potent repressor of TGF-b signaling 1999; Sun et al., 1999a; Xu et al., 2000), that Ski associates with a multi-Smad complex (Figure 1) that Transforming growth factor b (TGF-b) is an ubiquitous specifically binds the SBE. SKI binds the MH2 domains cytokine with paradoxical functions in many tissues. of Smad2 and Smad3, which are virtually identical This topic has been extensively reviewed (Massague and (Figure 2) (Xu et al., 2000). Binding of Ski to Smad2/ Chen, 2000; Attisano et al., 2001; Derynck et al., 2001; Smad3 causes dissociation of the histone acetyltransfer- Moustakas et al., 2001; Piek and Roberts, 2001; Rooke ase p300 from the Smad2/3 complex and promotes and Crosier, 2001; Wong and Lai, 2001; ten Dijke et al., association with mSin3A and histone deacetylases 2002; Verrecchia and Mauviel, 2002), therefore, I will (Akiyoshi et al., 1999). HDAC-containing complexes only highlight some important findings and focus on are known to repress transcription by deacetylation of SKI overexpression, and its consequences for TGF-b histone residues, which in turn favor histone methyla- signaling, and melanoma tumor progression. tion and transcriptional silencing (reviewed by Rountree TGF-b functions as a negative growth regulator for et al., 2001; Baylin et al., 2001). epithelial cells fromdiverse tissues, but it is also a potent growth promoter for mesenchymal, smooth muscle cells, and chondrocytes. Tissue activities regulated by TGF-b Ski (218-467) include early development, cell-cycle control, differentiation, extracellular matrix deposition and remodeling, angiogenesis, chemotaxis, and induction of apoptosis Smad2 MH1 Linker MH2 (reviewed by Schuster and Krieglstein, 2002). Carcino- 1185 269 467 mas often secrete excess TGF-b and respond to it by enhanced invasion and metastasis (reviewed by Akhurst and Derynck, 2001). SkiSki (211(211-424)424) Melanocytes are exquisitely sensitive to exogenous TGF-b inhibition (Rodeck et al., 1994), but melanoma cells show various degrees of resistance (Krasagakis Smad3 MH1 LinkL ker MH2 et al., 1999). Melanoma cells also secrete substantial 1 144 220 424 amounts of TGF-b (Krasagakis et al., 1999), which correlates with the depth of invasion. Such an event has Figure 2 SKI binds to the MH2 domain of Smad2 and Smad3

Oncogene Repression of TGF-b signaling EE Medrano 3126 SKI overexpression, localization, and function during In addition to SP-1, the Smad proteins functionally melanoma tumor progression cooperate with a diverse number of transcription factors in response to TGF-b. For example, the heteromeric Mammalian and avian melanocytes are targeted by the complex of Smad3 and Smad4 synergizes with c-Jun/c- transforming activity of Ski. Early studies showed that Fos at the AP-1 binding site of the collagenase I v-ski induces transformation of pigmented primary promoter to induce transcriptional activation in re- quail melanocytes (Barkas, 1986). During mouse devel- sponse to TGF-b (Qing et al., 2000). Cooperation opment, Ski is expressed at relatively high levels in between SP-1 and Smads is also repressed by SKI (Xu neural crest-derived cells (Lyons et al., 1994), and SKI et al., 2000). transcripts are expressed at high levels in human melanoma cell lines compared to normal melanocytes Downregulation of SKI restores TGF-b-mediated growth (Fumagalli et al., 1993). Overexpression of SKI in inhibition in human melanomas melanomas results from yet to be defined transcriptional and/or post-transcriptional events, as Southern blot Loss of TGF-b sensitivity is frequently observed in analysis did not find alterations in restriction enzyme tumors derived from cells that are otherwise sensitive to patterns or increase in gene dosage (Fumagalli et al., growth inhibition by this protein, and the extent of 1993). TGF-b resistance often correlates with metastatic Recently, we demonstrated that SKI is overexpressed progression (Stone et al., 1997). Although melanoma in vivo in all melanomas studied (n ¼ 44) (Reed et al., cells are highly resistant to the inhibitory activity of 2001). In preinvasive melanomas in situ, SKI was TGF-b, there are no measurable defects at the receptor observed predominantly in the nucleus of intraepider- level (Nishizuka, 1986; Holbrook and Fornace, 1991). mal melanoma cells. However, most of the primary However, downregulation of SKI protein levels by invasive melanomas demonstrated both nuclear and antisense SKI vectors restores TGF-b-mediated growth cytoplasmic localization of SKI, though nuclear labeling inhibition in melanomas (Reed et al., 2001). Therefore, was greater in the intraepidermal melanoma cells. In elevated expression of SKI is necessary and sufficient for melanoma metastasis, SKI localizes to both the nuclear curtailing the inhibitory activity of TGF-b. Thus, and cytoplasmic compartments, or predominantly to the deregulated expression of SKI represents a significant cytoplasmic compartment. Association of cytoplasmic event in the progression of melanomas in vivo. Together, SKI with Smad3 apparently prevents Smad3 nuclear these results suggest that manipulation of SKI levels translocation in response to TGF-b. Thus, the biological could be a prospective melanoma therapy. consequence of SKI/Smad3 interaction in the cytoplasm appears to be similar to NLS mutations in Smad3, as this mutant remains in the cytoplasm and functions as Secreted TGF-b1 increases matrix remodeling and dominant-negative inhibitor of TGF-b signaling (Xiao melanoma survival et al., 2000). Increased expression and activation of TGF-b1by tumor cells stimulates extracellular matrix deposition SKI represses induction of p21Waf-1 by TGF-b and chemoattraction of fibroblasts, promoting tumor growth and angiogenesis (Roberts et al., 1986; Zheng The Smad proteins regulate transcription of the cyclin- et al., 1997a; Derynck et al., 2001). TGF-b1 expression dependent kinase inhibitor p21Waf-1 (Pardali et al., 2000), by melanoma cells can modify the surrounding stroma which together with p15INK4b (Feng et al., 2002), mediate with increased production and deposition of extracel- the antiproliferative and tumor suppressor activity of lular matrix proteins, which favor increased metastatic TGF-b. Increased p21Waf-1 protein levels appear to be spread and survival (Berking et al., 2001). essential for TGF-b-mediated inhibition of the cyclin- TGF-b1 can also induce transdifferentiation of tumors dependent kinase CDK2 (reviewed by Moustakas et al., from epithelial or melanocytic origin to highly malignant 2002). The p21Waf-1 promoter contains a proximal Sp1- fibroblastoid tumors characterized by spindle-shaped binding site that confers Smad cooperativity and a distal cells (reviewed by Janji et al., 1999; Derynck et al., 2001). SBE (Sobballe and Herlyn, 1994). High levels of SKI By simultaneously repressing TGF-b-mediated growth prevent p21Waf-1 induction through a Smad-dependent inhibition and releasing high levels of this cytokine to the mechanism that involves transcriptional repression microenvironment, melanomas have developed a highly (Reed et al., 2001). Thus, elevated expression of SKI efficient, cell-autonomous pathway that stimulates pro- facilitates cell-cycle progression by targeting the RB liferation, invasion, and survival (Figure 3). For that, the pathway in at least two ways; high levels can directly Smads meet their repressor SKI. repress RB activity, and indirectly, increase CDK2 activity by repressing TGF-b-mediated induction of p21Waf-1. Importantly, such activities are reciprocal to Perspectives and future directions p16INK4a loss, an event associated with mouse and human melanoma formation and progression (reviewed These are exciting times for SKI and Sno fans. We have by Tietze and Chin, 2000; Bardeesy et al., 2001; Yang learned of numerous activities regulated by these et al., 2001). proteins, and yet, numerous unanswered questions

Oncogene Repression of TGF-b signaling EE Medrano 3127 TGF-b and related factors induce apoptosis in a Stroma Remodeling, Survival, variety of tissues. However, melanoma tumors are Invasion, Angiogenesis, Metastasis notoriously resistant to apoptosis mediated by diverse insults (reviewed by Serrone and Hersey, 1999). The death-associated protein kinase (DAP-kinase) is a Active TGF-β calcium/calmodulin-regulated serine/threonine kinase that localizes to the cytoskeleton, participates in several apoptotic processes (reviewed by Kogel et al., 2001) and is a tumor suppressor (Kissil and Kimchi et al., 1998; RR* Latent TGF-β Levy-Strumpf and Kimchi, 1998; Raveh and Kimchi, 2001). Is SKI involved in the resistance to apoptosis of Melanoma Cell Smad2,Smad3, melanoma cells? Probably yes, as expression of the Smad4 DAP-kinase is transcriptionally induced by TGF-b through canonical SBE elements in its promoter (Jang et al., 2001). We also know little regarding protein complex N-CoR mSin3 Ski formation and activity between endogenous SKI and members of its own family such as SnoN, and with HDAC Transcriptional Smad2 Smad3 Repression of Genes proteins that might interfere with its activity such as the Involved in SKI-interacting protein Skip (Dahl et al., 1998b). Skip is Smad4 Growth Inhibition particularly interesting as its overexpression partially and Apoptosis overcomes Ski/Sno-dependent repression. However, Ski/Sno overexpression attenuates Skip augmentation Figure 3 Repressing the good of TGF-b signaling leaves the bad of TGF-b-dependent transcription (Leong et al., 2001). for melanoma tumor progression. RR* indicates functional Considering both, resistant to TGF-b and SKI over- melanoma TbRI and TbRII receptors (Rodeck et al., 1994) expression, we predict that SKI counteracts Skip activity in malignant melanoma tumors. remain. For example, what signals upregulate SKI Another intriguing SKI partner is PML. This protein transcription in melanomas (Fumagalli et al., 1993). Is interacts with RB, and is involved in the induction of such an event necessary and sufficient for melanoma growth arrest and senescence induced by oncogenic tumor progression as the pattern of SKI protein RAS (reviewed by Salomoni and Pandolfi, 2002). Could expression appears to indicate (Reed et al., 2001)? Is high levels of SKI prevent PML and p53 induction by SKI protein stability also altered? The anaphase- RAS and overcome senescence? promoting complex is a ubiquitin ligase required for Finally, in addition to repression, c-ski can activate the destruction of SnoN, a pathway regulated by TGF-b transcription fromsometissue-specific and viral en- (Sun et al., 1999b; Bonni et al., 2001; Stroschein et al., hancers and promoters (Engert et al., 1995; Ishikawa 2001; Wan et al., 2001). Considering the homology with et al., 1997; Kelder et al., 1997). Therefore, c-Ski can act SnoN, it is likely that SKI levels are also similarly as activator or repressor of transcription depending on regulated. However, once again, most of these studies protein-protein interactions dictated by the promoter had used transfected proteins, therefore, we know very and physiological context. What we presently know little on how SKI stability is regulated in melanomas. about SKI in melanomas is sufficient to catalogue it as High levels of SKI displayed by all cells within a an oncogene. Unraveling additional SKI functions melanoma tissue section suggest that SKI might be more through global approaches including microarrays and stable or differentially regulated in these tumors. proteomics should help determine whether this protein Furthermore, cytoplasmic and nuclear SKI might dis- is within the ‘gut apparatus’ that drives melanocyte play different half-lives if they localize to different transformation and/or tumor progression. protein complexes. Initial experiments in synchronized human melanoma cell lines indicate that although endogenous SKI levels consistently peak at mitosis, Acknowledgements G1 cells of different lines show variable amounts of this I thank Ed Stavnezer, Xin-Hua Feng, for critical comments protein (Medrano, unpublished results). Additional and suggestions, and the contribution of colleagues and past studies should determine whether such variability and present members of my laboratory to the work discussed depends on the distribution of SKI between the in this review. I also thank the National Cancer Institute for cytoplasmic and nuclear compartments. research support.

References

Akhurst RJ and Derynck R. (2001). Trends Cell Biol., 11, S44– Akiyoshi S, Inoue H, Hanai J, Kusanagi K, Nemoto N, S51. Miyazono K and Kawabata M. (1999). J. Biol. Chem., 274, 35269–35277.

Oncogene Repression of TGF-b signaling EE Medrano 3128 Amaravadi LS, Neff AW, Sleeman JP and Smith RC. (1997). Leong GM, Subramaniam N, Figueroa J, Flanagan JL, Dev. Biol., 192, 392–404. Hayman MJ, Eisman JA and Kouzmenko AP. (2001). J. Attisano L and Tuen Lee-Hoeflich S. (2001). Genome Biol., 2, Biol. Chem., 276, 18243–18248. REVIEWS3010. Levy-Strumpf N and Kimchi A. (1998). Oncogene, 17, 3331– Bardeesy N, Bastian BC, Hezel A, Pinkel D, DePinho RA and 3340. Chin L. (2001). Mol. Cell Biol., 21, 2144–2153. Luo K, Stroschein SL, Wang W, Chen D, Martens E, Zhou S Barkas A. (1986). Ph. D. dissertation (New York University, and Zhou Q. (1999). Genes Dev., 13, 2196–2206. New York). Lyons GE, Micales BK, Herr MJ, Horrigan SK, Namciu S, Baylin SB, Esteller M, Rountree MR, Bachman KE, Schuebel Shardy D and Stavnezer E. (1994). Dev. Dyn., 201, 354–365. K and Herman JG. (2001). Hum. Mol. Genet., 10, 687–692. Massague J and Chen YG. (2000). Genes Dev., 14, 627–644. Berk M, Desai SY, Heyman HC and Colmenares C. (1997). Moustakas A, Pardali K, Gaal A and Heldin CH. (2002). Genes Dev., 11, 2029–2039. Immunol. Lett., 82, 85–91. Berking C, Takemoto R, Schaider H, Showe L, Satyamoorthy Moustakas A, Souchelnytskyi S and Heldin CH. (2001). J. Cell K, Robbins P and Herlyn M. (2001). Cancer Res., 61, 8306– Sci., 114, 4359–4369. 8316. Namciu S, Lieberman MA and Stavnezer E. (1994). Oncogene, Bonni S, Wang HR, Causing CG, Kavsak P, Stroschein SL, 9, 1407–1416. Luo K and Wrana JL. (2001). Nat. Cell Biol., 3, 587–595. Nicol R and Stavnezer E. (1998). J. Biol. Chem., 273, 3588– Boyer PL, Colmenares C, Stavnezer E and Hughes SH. (1993). 3597. Oncogene, 8, 457–466. Nishizuka Y. (1986). Science, 233, 305–312. Cohen SB, Zheng G, Heyman HC and Stavnezer E. (1999). Nomura N, Sasamoto S, Ishii S, Date T, Matsui M and Nucleic Acids Res., 27, 1006–1014. Ishizaki R. (1989). Nucleic Acid Res., 17, 5489–5500. Colmenares C, Heilstedt HA, Shaffer LG, Schwartz S, Berk M, Nomura T, Khan MM, Kaul SC, Dong HD, Wadhwa R, Murray JC and Stavnezer E. (2001). Nat. Genet., 30, 106–109. Colmenares C, Kohno I and Ishii S. (1999). Genes Dev., 13, Colmenares C, Sutrave P, Hughes SH and Stavnezer E. (1991). 412–423. J. Virol., 65, 4929–4935. Pardali K, Kurisaki A, Moren A, ten Dijke P, Kardassis D, Dahl R, Kieslinger M, Beug H and Hayman MJ. (1998a). and Moustakas A. (2000). The Journal of Biological Proc. Natl. Acad. Sci. USA, 95, 11187–11192. Chemistry, 275, 29244–29256. Dahl R, Wani B and Hayman MJ. (1998b). Oncogene, 16, Pearson-White S and Crittenden R. (1997). Nucleic Acids Res., 1579–1586. 25, 2930–2937. Derynck R, Akhurst RJ and Balmain A. (2001). Nat. Genet., Piek E, Heldin CH and ten Dijke P. (1999). FASEB J., 13, 29, 117–129. 2105–2124. Engert JC, Servaes S, Sutrave P, Hughes SH and Rosenthal N. Piek E and Roberts AB. (2001). Adv. Cancer Res., 83, 1–54. (1995). Nucleic Acids Res., 23, 2988–2994. Qing J, Zhang Y and Derynck R. (2000). J. Biol. Chem., 275, Feng XH, Liang YY, Liang M, Zhai W and Lin X. (2002). 38802–38812. Mol. Cell, 9, 133–143. Raveh T and Kimchi A. (2001). Exp. Cell Res., 264, Fumagalli S, Doneda L, Nomura N and Larizza L. (1993). 185–192. Melanoma Res., 3, 23–27. Reed JA, Bales E, Xu W, Okan NA, Bandyopadhyay D and Heyman HC and Stavnezer E. (1994). J. Biol. Chem., 269, Medrano EE. (2001). Cancer Res., 61, 8074–8078. 26996–27003. Reed JA, McNutt NS, Prieto VG and Albino AP. (1994). Am. Holbrook NJ and Fornace AJ. (1991). N. Biol., 3, 825–833. J. Pathol., 145, 97–104. Ichikawa K, Nagase T, Ishii S, Asano A and Mimura N. Roberts AB, Sporn MB, Assoian RK, Smith JM, Roche NS, (1997). Biochem. J., 328, 607–613. Wakefield LM, Heine UI, Liotta LA, Falanga V and Kehrl Jang CW, Chen CH, Chen CC, Chen JJ, Su YH and Chen RH. JH. (1986). Proc. Natl. Acad. Sci. USA, 83, 4167–4171. (2001). Nat. Cell Biol. Rodeck U, Bossler A, Graeven U, Fox FE, Nowell PC, Janji B, Melchior C, Gouon V, Vallar L and Kieffer N. (1999). Knabbe C and Kari C. (1994). Cancer Res., 54, 575–581. Int. J. Cancer, 83, 255–262. Rooke HM and Crosier KE. (2001). Pathology, 33, 73–84. Kaghad M, Bonnet H, Yang A, Creancier L, Biscan JC, Rountree MR, Bachman KE, Herman JG and Baylin SB. Valent A, Minty A, Chalon P, Lelias JM, Dumont X, (2001). Oncogene, 20, 3156–3165. Ferrara P, McKeon F and Caput D. (1997). Cell, 90, 809– Salomoni P and Pandolfi PP. (2002). Cell , 108, 165–170. 819. Schuster N and Krieglstein K. (2002). Cell Tissue Res., 307, 1– Kaufman CD, Martinez-Rodriguez G and Hackett Jr PB. 14. (2000). Mech. Dev., 95, 147–162. Serrone L and Hersey P. (1999). Melanoma Res., 9, 51–58. Kelder B, Richmond C, Stavnezer E, List EO and Kopchick Shinagawa T, Dong HD, Xu M, Maekawa T and Ishii S. JJ. (1997). Gene, 202, 15–21. (2000). EMBO J., 19, 2280–2291. Khan MM, Nomura T, Kim H, Kaul SC, Wadhwa R, Shinagawa T, Nomura T, Colmenares C, Ohira M, Nakaga- Shinagawa T, Ichikawa-Iwata E, Zhong S, Pandolfi PP and wara A and Ishii S. (2001). Oncogene, 20, 8100–8108. Ishii S. (2001). Mol. Cell, 7, 1233–1243. Sobballe PW and Herlyn M. (1994). Melanoma Res., 4, 213– Kissil JL and Kimchi A. (1998). Mol. Med. Today, 4, 268–274. 223. Kogel D, Prehn JH and Scheidtmann KH. (2001). BioEssays, Stavnezer E, Brodeur D and Brennan LA. (1989). Mol. Cell 23, 352–358. Biol., 9, 4038–4045. Kokura K, Kaul SC, Wadhwa R, Nomura T, Khan MM, Stavnezer E, Gerhard DS, Binari RC and Balazs I. (1981). J. Shinagawa T, Yasukawa T, Colmenares C and Ishii S. Virol., 39, 920–934. (2001). J. Biol. Chem., 276, 34115–34121. Stone JG, Spirling LI and Richardson MK. (1997). J. Cell Sci., Krasagakis K, Kruger-Krasagakes S, Fimmel S, Eberle J, 110, 1673–1682. Tholke D, von der OM, Mansmann U and Orfanos CE. Stroschein SL, Bonni S, Wrana JL and Luo K. (2001). Genes (1999). J. Cell Physiol, 178, 179–187. Dev., 15, 2822–2836.

Oncogene Repression of TGF-b signaling EE Medrano 3129 Sun Y, Liu X, Eaton EN, Lane WS, Lodish HF and Weinberg Wong SF and Lai LC. (2001). Pathology, 33, 85–92. RA. (1999a). Mol. Cell, 4, 499–509. Xiao Z, Liu X and Lodish HF. (2000). J. Biol. Chem., 275, Sun Y, Liu X, Ng-Eaton E, Lodish HF and Weinberg RA. 23425–23428. (1999b). Proc. Natl. Acad. Sci. USA, 96, 12442–12447. Xu W, Angelis K, Danielpour D, Haddad MM, Bischof O, Sutrave P, Copeland TD, Showalter SD and Hughes SH. Campisi J, Stavnezer E and Medrano EE. (2000). Proc. Natl. (1990). Mol. Cell Biol., 10, 3137–3144. Acad. Sci. USA, 97, 5924–5929. ten Dijke P, Goumans MJ, Itoh F and Itoh S. (2002). J. Cell Yang FC, Merlino G and Chin L. (2001). Semin. Cancer Biol., Physiol., 191, 1–16. 11, 261–268. Tietze MK and Chin L. (2000). Mol. Med. Today, 6, 408–410. Zawel L, Dai JL, Buckhaults P, Zhou S, Kinzler KW, Tokitou F, Nomura T, Khan MM, Kaul SC, Wadhwa R, Vogelstein B and Kern SE. (1998). Mol. Cell, 1, 611–617. Yasukawa T, Kohno I and Ishii S. (1999). J. Biol. Chem., Zheng G, Blumenthal KM, Ji Y, Shardy DL, Cohen SB and 274, 4485–4488. Stavnezer E. (1997b). J. Biol. Chem., 272, 31855–31864. Verrecchia F and Mauviel A. (2002). J. Invest Dermatol., 118, Zheng G, Teumer J, Colmenares C, Richmond C and 211–215. Stavnezer E. (1997a). Oncogene, 15, 459–471. Wan Y, Liu X and Kirschner MW. (2001). Mol. Cell, 8, 1027– 1039.

Oncogene