Osteopontin Is an Oncogenic Vav1– but Not Wild-Type Vav1– Responsive Gene: Implications for Fibroblast Transformation Vered Schapira, Galit Lazer, and Shulamit Katzav

Research Article

Osteopontin Is an Oncogenic Vav1– but not Wild-type Vav1– Responsive Gene: Implications for Fibroblast Transformation Vered Schapira, Galit Lazer, and Shulamit Katzav

The Hubert H. Humphrey Center for Experimental Medicine and Cancer Research, The Hebrew University-Hadassah Medical School, Jerusalem, Israel

Abstract proteins in this family, Vav2 and Vav3, are more widely expressed Mammalian wild-type Vav1 (wtVav1) encodes a specific GDP/ (3). There is good evidence that the Vav proteins play a central role GTP nucleotide exchange factorthat is exclusively expressedin in triggering the cytoskeletal reorganization required for many the hematopoietic system. Despite numerous studies, the physiologic processes by participating in several distinct pathways mechanism underlying transformation of fibroblasts by as a GDP/GTP nucleotide exchange factor (GEF) for Rho/Rac oncogenic Vav1 (oncVav1) is not well defined. We identified GTPases (4) and as scaffold proteins (3). Mammalian Vav proteins contain several characteristic structural motifs that enable their osteopontin, a markerfortumoraggressiveness, as an dbl oncVav1-inducible gene. Osteopontin is highly expressed in function as signal transducer proteins (2, 3). These include a oncVav1-transformed NIH3T3 cells (NIH/oncVav1) but is barely homology (DH) region that exhibits a GEF activity toward the Rho family GTPases, a pleckstrin homology domain that interacts with detected in NIH3T3 expressing wtVav1 (NIH/wtVav1) even following epidermal growth factor stimulation, which normal- polyphosphoinositides, a Src homology (SH) 2 and two SH3 ly induces osteopontin. Depleting oncVav1 in NIH/oncVav1 domains that mediate protein-protein interactions, a proline-rich using small interfering RNA led to a considerable decrease in motif that enables binding to SH3-containing proteins, and an osteopontin, whereas reducing osteopontin expression did not acidic-rich region and a calponin homology (CH) region at the NH2 affect oncVav1 expression, suggesting that oncVav1 operates terminus of the protein that function as actin-binding domain in upstream of osteopontin. Vav1-depleted NIH/oncVav1 cells, but other proteins. An additional important characteristic of the Vav not osteopontin-depleted NIH/oncVav1 cells, exhibited im- proteins is the close correlation between their activity as GEFs and paired extracellular signal-regulated kinase (ERK) and c-Jun their phosphorylation on tyrosine residues, which is regulated by NH -terminal kinase phosphorylation. Inhibition of ERK activation of cell surface receptors (5). 2 OncVav1 was activated in vitro by replacement of 67–amino phosphorylation in NIH/oncVav1 cells led to a decrease in osteopontin expression, implying that the elevated osteopontin acid residues of the Vav1 NH2 terminus (part of its CH region) expression in these cells is dependent on ERK phosphorylation. with 19–amino acid residues of pSV2neo sequences that was Vav1-depleted orosteopontin-depleted NIH/oncVav1 cells lost cotransfected with DNA of human esophageal carcinomas as a their tumorigenic properties as judged by the soft agar and selectable marker in the search for novel oncogenes (1). When invasion assays, although loss of osteopontin expression had a oncVav1-transformed NIH3T3 cells are injected i.v., they metas- tasize to mice lungs (6), suggesting that oncVav1 may contribute less dramatic effect. Suppression of Vav1 expression in NIH/ in vivo. oncVav1 cells led to reversion to ‘‘normal’’ morphology, to the aggressiveness of transformed cells In contrast, whereas when only osteopontin expression was diminished wild-type Vav1 (wtVav1) produces minimal transformation of cells retained their transformed morphology. This work NIH3T3 fibroblasts even when the protein is grossly overex- strongly supports a role for oncVav1 as a master oncogene pressed. Removal of the NH2-terminal sequences (65 residues), and provides clues to the molecular mechanism underlying mimicking its original mode of activation, was sufficient to induce oncVav1 transformation. (Cancer Res 2006; 66(12): 6183-91) Vav1 transformation (7, 8). Mutant Vav1 proteins that have lost a larger portion of the NH2 terminus (CH domain plus the acidic- rich region up to 186–amino acid residues; D1-186) exhibit a more Introduction pronounced transforming activity than oncVav1, which is missing Vav1 has been intensively studied by several investigators since it only 67 residues (9). was detected as an oncogene [oncogenic Vav1 (OncVav1)] in 1989 The mechanisms underlying transformation by Vav1 have been (1). These studies have provided us with several insights into the the subject of intensive research. Because Vav1 functions as a GEF role of the Vav proteins in mediating intracellular signals that and the morphology of the oncVav1-transformed cells is reminis- modulate cell growth and differentiation (2). However, many cent of cells transformed by the dbl oncogene or active versions of critical questions remain regarding the ways in which Vav1 can also Rho family GTPase (10), it seemed conceivable that oncVav1- act as a transforming protein. induced transformation is the result of deregulated GEF activity. Vav1, the first Vav family member isolated, is a hematopoietic Indeed, the mutant Vav1 protein (D1-186) shows constitutive, cell-specific signal transducer protein (1, 3). The other mammalian phosphorylation-independent exchange activity, correlated with its transforming activity (9). Nuclear magnetic resonance studies showed that Tyr174 within the acidic-rich region of Vav1 makes Note: V. Schapira and G. Lazer contributed equally to this work. several contacts with side chains of amino acids within the DH Requests for reprints: Shulamit Katzav, The Hubert H. Humphrey Center for 174 Experimental Medicine and Cancer Research, The Hebrew University-Hadassah domain (11). Following phosphorylation of Tyr , the tyrosine is Medical School, P.O. Box 12272, Jerusalem 91120, Israel. Phone: 972-2-6758350; Fax: released from its binding pocket, which allows access of the DH 972-2-6414583; E-mail: [email protected]. I2006 American Association for Cancer Research. domain to its substrate. Yet, the oncVav1 mutant (D1-66) that doi:10.1158/0008-5472.CAN-05-3735 retains the inhibitory residue Y174, the entire acidic domain, and www.aacrjournals.org 6183 Cancer Res 2006; 66: (12). June 15, 2006

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 2006 American Association for Cancer Research. Cancer Research part of the CH domain (5) exhibits an exchange activity that is siRNA duplexes (100 pmol siRNAs) following the manufacturer’s instruc- phosphorylation dependent (9). tions (Invitrogen, Carlsbad, CA). To monitor transfection efficiency, green The possibility that GEF-independent mechanisms might also be fluorescent protein (GFP) plasmid was cotransfected. involved in Vav1 transformation has recently emerged. For Antibodies and Immunofluorescence Reagents and instance, the activity of Vav1 in GEF-independent pathways was 2+ Immunofluorescence Microscopy shown to be critical for its activity in T cells, including Ca The following antibodies were used in our studies: anti-Vav1 polyclonal mobilization (12, 13) and stimulation of the transcription factor antibodies were raised in rabbits against a specific peptide of Vav1 (residues nuclear factor of activated T cells (14). In addition, association of 528-541; 7); anti-Vav1 monoclonal antibody (mAb); anti-phosphotyrosine the SH2 (15) and SH3 (16) regions of oncVav1 (D1-66) with other mAb (4G10), anti-Rac, and anti–extracellular signal-regulated kinase (ERK) proteins was shown to be critical for transformation. Additionally, antibodies (Upstate Biotechnology, Lake Placid, NY); anti-phosphorylated phospholipase (PLC) g (Tyr783), anti-PLCg antibodies, anti-phosphorylated the NH2-terminal tail could mediate protein-protein interactions that inhibit the biological activity of Vav1 in the resting state and ERK mAbs, anti-phosphorylated c-Jun NH2-terminal kinase (JNK) mAbs, are relieved on activation. Indeed, we have recently described the and anti-phosphorylated JNK (Cell Signaling Technology, Beverly, MA); anti- osteopontin (R&D Systems, Minneapolis, MN); and anti-actin (Santa Cruz binding of Ly-GDI, another potential regulator of Rho GTPases, to Biotechnology, Santa Cruz, CA); donkey anti-goat and goat anti-rabbit Cy3- the NH2 terminus of Vav1 (13). The significance of this interaction conjugated antibodies (Jackson ImmunoResearch, West Grove, PA) were to the function of Vav1 in T cells has been shown (13). However, used in the immunofluorescence analysis. Immunofluorescence and because Ly-GDI is expressed only in the hematopoietic system, confocal microscopy analysis were done as described previously (13). other proteins may affect Vav transformation in NIH3T3 fibroblasts via association with the CH region. Thus, the mechanisms of Immunoprecipitation and Immunoblotting transformation by oncVav1 (D1-66) remain enigmatic. Cells lysis, immunoprecipitation, and immunoblotting were done as Here, we investigated whether oncVav1 (D1-66) up-regulates described previously (13). specific genes involved in cancer. We now report that osteopontin, RNA Interference in NIH/oncVav1 Cells a marker of tumor malignancy (17, 18), is overexpressed in cells OncVav1. Two siRNA targeting sequences were selected corresponding transformed by oncVav1 (D1-66) but repressed in cells expressing to the open reading frame of human Vav1 gene using an online design wtVav1. This study explored the role of osteopontin in oncVav1- program (Promega). These sequences included a synthetic double-stranded induced transformation and investigated the effects of oncVav1 oligonucleotide (5¶-TCGAGTGCCTATGCAGCGAGTTCgagtactgGAACTC- and osteopontin on various signaling pathways. GCTGCATAGGCATTTTT-3¶) that corresponds to a 19-nucleotide sequence from wtVav1 (nucleotide 979-998), which are separated by a 9-nucleotide linker (lowercase letters) from the reverse complement of the same Materials and Methods 19-nucleotide sequence, and an additional synthetic double-stranded oligonucleotide (5¶-TCGAGCGAGACAACGAGACACTGgagtactgCAGTGTC- Cell Culture TCGTTGTCTCGCTTTTT-3¶) that corresponds to wtVav1 (nucleotides NIH3T3, NIH/oncVav1 (1), and NIH/wtVav1 (7) cells were maintained in 1,217-1,236) and is composed of the same characteristics mentioned above. DMEM supplemented with 10% calf serum (Life Technologies, Grand Island, Both sequences used for siRNA plasmid construction were blasted to avoid NY). For treatment with epidermal growth factor (EGF), cells were grown to silencing other unrelated genes. These oligonucleotides were subcloned into subconfluence and then starved in DMEM containing 0.5 mg/mL insulin the Vector pSuppressor, which generates biologically active siRNAs from the and transferrin for 24 hours. Cells were then treated with DMEM containing U6 promoter (Imgenex, San Diego, CA), thus yielding si1Vav and si2Vav, 100 ng/mL murine EGF for the indicated time intervals. respectively. We used a circular control plasmid that contains a scrambled RNA Purification and Semiquantitative Real-time PCR sequence without significant homology to rat, mouse, or human gene Total RNA was purified using EZ-RNA kit according to the manufac- sequences [scrambled vector (scVector)]. The sequences of siRNA insert turer’s instructions (Biological Industries, Beit-Haemek, Israel). For reverse were verified by DNA sequencing. transcription, 2 Ag total RNA and 0.5 pmol/AL random primers (Promega, Osteopontin. siRNA duplexes were designed according to Teramoto Madison, WI) were added to each Moloney murine leukemia virus reverse et al. (19): OPNi-840 (sense GUUUCACAGCCACAAGGACdTdT and anti- transcriptase reaction (Promega). The primers used were as follows: sense GUCCUUGUGGCUGUGAAACdTdT) and the negative control, the 5¶-ATTTCCAGCTGTCCATTG-3¶ (forward) and 5¶-CTGAAACACCTCCAT- scrambled OPNi (sense CAGUACAACGCAUCUGGCAdTdT and antisense CTTG-3¶ (reverse) for Vav1 detection, 5¶-GATCGATAGTCAAGCAAGTT-3¶ UGCCAGAUGCGUUGUACUGdTdT). These siRNA duplexes are called OPNi (forward) and 5¶-GGTATAGTGATATAGACTGT-3¶ (reverse) for osteopontin and scOPN throughout, respectively. detection, and 5¶-CCAAGGTCATCCATGACAAC-3¶ (forward) and 5¶-GGCC- ATGAGGTCCACCACC-3¶ for glyceraldehyde-3-phosphate dehydrogenase Cell Invasion Assay (Matrigel Invasion Assay) (GAPDH) detection. We used the Matrigel invasion chambers (BD Biosciences, San Jose, CA) to assess migration (20). Briefly, a single-cell suspension of 200,000 cells per Quantitative Real-time PCR well was added to the upper chamber of a prehydrated membrane filter of Real-time PCR (RT-PCR) was done using assays on demand that contain 8-mm pore size detailed above. The lower chamber was filled with Taqman PCR master mix (Applied Biosystems, Foster City, CA) and the conditioned medium of semiconfluent NIH3T3 cells. Cell invasiveness was required primers. The following assays on demand were used: hOPN, determined at the lower surface of the filter after 24 hours following the #Hs00167093_m1; mOPN, #Mn00436767_m1; hVav1, #Hs00232108_m1; and recommended procedure by BD Biosciences. Noninvading cells on the h18sRNA, #Hs99999901_s1. Nontemplate control (contamination sample’s upper side of the filter and Matrigel were wiped off and migrating cells on reagents) and samples with 18SRNA for normalization were also used. The the reverse side of the filter were stained with DiffQuick, which consists of a analysis was done using the ABI Prism 7000 RT-PCR technology (Applied methanol fixative, buffered eosin, and phosphate-buffered azure B (Dade Biosystems). The assays were done thrice in triplicate. Behring, Deerfield, IL). Experiments were reproduced thrice in triplicate and counting was done on five different fields of each filter. Transfection The calcium phosphate precipitate method of transfection was used for Soft AgarAssay all transient transfections with Vav1 small interfering RNA (siRNA) Cells were transiently transfected as indicated (see Results), trypsinized, plasmids. LipofectAMINE 2000 was used for introduction of the osteopontin suspended in DMEM containing 0.4% agar and 10% calf serum, and plated

Cancer Res 2006; 66: (12). June 15, 2006 6184 www.aacrjournals.org

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 2006 American Association for Cancer Research. Vav1 and Osteopontin in Transformed Fibroblasts onto a bottom layer containing 0.8% agar. The cells were plated at a density of 1 Â 105 per well in a six-well plate, and colonies were counted 14 days later. Each experiment was done thrice in triplicate. ERK Inhibition NIH/oncVav1 and NIH/wtVav1 cells were grown to 80% confluence and then starved in DMEM containing 5 mg/mL insulin and transferrin for 24 hours. The cells were then incubated for 2 hours with 10 Amol/L UO126.

Results Osteopontin expression is elevated in NIH3T3 cells transformed by oncVav1. To explore the transforming potential of oncVav1, we investigated differences in gene expression in oncVav1 and wtVav1 cells by probing a Cancer Gene Array with cDNAs prepared from three populations of NIH3T3 cells: untransfected, stably transfected with oncVav1 (NIH/oncVav1), and stably transfected with wtVav1 (NIH/wtVav1). Using this approach, several genes exhibited different patterns of RNA expression in these cell populations (data not shown). Strikingly, osteopontin RNA was highly expressed in NIH/oncVav1 cells compared with NIH3T3 and NIH/wtVav1 cells. Osteopontin is a secreted adhesive glycoprotein with diverse functions and has been implicated in cancer development, progression, and metastasis (17, 18). In fact, osteopontin was recently used as a marker for the aggressiveness of tumors (17, 18). Semiquantitative RT-PCR confirmed that osteopontin is highly expressed in NIH/oncVav1 cells compared with NIH3T3 and NIH/ wtVav1 cells (Fig. 1A). Osteopontin was not detectable in NIH/ wtVav1 cells even after 30 cycles of PCR (Fig. 1A). The housekeeping gene, GAPDH, showed no significant differences in its RNA expression in the three lines used (Fig. 1A). RT-PCR results were compatible with the semiquantitative PCR results, showing osteopontin RNA expression 14 times higher in NIH/oncVav1 cells than in NIH/wtVav1 cells (data not shown). Similar results were obtained following transient transfection of NIH3T3, Jurkat T cells, Figure 1. Expression of osteopontin mRNA in noninduced NIH3T3, NIH/ and HCT (a human colon tumor cell line) with plasmids containing oncVav1, and NIH/wtVav1 cells (A)and following stimulation with EGF ( B and wtVav1 or oncVav1 (data not shown), attesting to the generality of C). A, RNA was prepared from the indicated established cell lines and used as templates in a semiquantitative RT-PCR assay done with osteopontin our observation. (OPN)–specific primers for 27 and 30 cycles. GAPDH was amplified on the same EGF stimulation of NIH3T3 cells induces osteopontin (21). As samples for 35 cycles. B and C, NIH3T3, NIH/oncVav1, and NIH/wtVav1 cells were either left untreated (0 hour)or treated with EGF for various time intervals expected, osteopontin mRNA was elevated in NIH3T3 cells (1, 6, and 24 hours). At the indicated times, cells were lysed and RNA was following 1 hour of EGF treatment and remained elevated after prepared and analyzed for osteopontin (B)and Vav1 ( C)RNA expression using 24 hours of treatment (Fig. 1B). In contrast, osteopontin mRNA semiquantitative RT-PCR. Thirty and 33 cycles were done for osteopontin and Vav1 (B and C)and 21 cycles were done for detection of GAPDH ( B and C). expression was very high in unstimulated starved NIH/oncVav1 Representative of three experiments. cells and did not increase following EGF stimulation (Fig. 1B). Osteopontin mRNA was not detectable in untreated NIH/wtVav1 cells, appearing only after 24 hours of stimulation with EGF NIH/oncVav1 cells with Vav1 siRNA plasmids also significantly (Fig. 1B). EGF had no effect on the expression of oncVav1 and reduced osteopontin mRNA expression (60-80%; Fig. 2A). Treat- wtVav1 (Fig. 1C). These results were supported by Western blot ment of NIH/oncVav1 cells with OPNi, which was reported analysis of osteopontin protein expression (data not shown). Thus, previously to be an effective siRNA duplex for down-regulating oncVav1 induces a dramatic increase in osteopontin expression osteopontin (19), led to a 70% reduction in osteopontin expression regardless of EGF stimulation. in these cells (Fig. 2A). No change in oncVav1 mRNA levels was Knocked down Vav1 expression in NIH/oncVav1: effects on noted in osteopontin-depleted NIH/oncVav1 cells (Fig. 2A). osteopontin expression and signal transduction cascades. To Western blot analysis shows that oncVav1 protein expression was determine whether the oncogenicity of oncVav1 is solely dependent considerably reduced in NIH/oncVav1 cells treated with either on osteopontin expression and/or on its participation in various si1Vav or si2Vav plasmids, whereas no change was seen in the level signal transduction cascades, we suppressed oncVav1 and osteo- of other proteins, such as actin or Rac (Fig. 2B). Osteopontin pontin in NIH/oncVav1 cells by using siRNA and analyzed the protein levels in NIH/oncVav1 cells were reduced following biological consequences of such treatments. RT-PCR analysis treatment with OPNi, si1Vav, or si2Vav but were unchanged in showed strikingly diminished expression of oncVav1 in NIH/ scOPN-treated cells (Fig. 2C). These results further establish the oncVav1 cells treated with si1Vav or si2Vav plasmids (78% and 58% link between oncVav1 and osteopontin mRNA and protein reduction below control levels, respectively; Fig. 2A). Treatment of expression and validate the use of Vav1 siRNA as a tool to www.aacrjournals.org 6185 Cancer Res 2006; 66: (12). June 15, 2006

oncVav1 cells treated with EGF (Fig. 3A). As expected, treatment with si1Vav or si2Vav dramatically reduced oncVav1 expression; the small amount of oncVav1 remaining seems to be tyrosine phosphorylated (Fig. 3A, lanes 5-8). Several reports clearly link wtVav1 to the activation of ERK, a known Ras-dependent kinase in various hematopoietic cell lineages, via a pathway involving the Ras GTPase, B-Raf, mitogen- activated protein kinase (MAPK) kinase (MEK) 1, and MEK2 kinases (22–24). Additionally, the MEK inhibitor PD90859 inhibited Vav1-induced activation of ERK (23). However, the activation of Ras and ERK appeared normal following stimulation of the T-cell receptor (TCR) in Vav1-null Jurkat T cells (25). The link between oncVav1 and ERK activation has thus far not been resolved. We analyzed ERK phosphorylation in NIH/oncVav1 cells treated with Vav1siRNA plasmids (Fig. 3B). ERK phosphorylation is increased

Figure 2. siRNA-mediated knockdown of oncVav1 (A)and osteopontin RNA (A)and protein ( B and C)expression in NIH/oncVav1 cells. NIH/oncVav1 cells were transiently transfected with either an empty vector, two siRNA plasmids directed against Vav1-specific sequences (si1Vav and si2Vav), or a scrambled (scOPN)or an effective OPNi siRNA duplex. Microarray analysis ( A)and Western blotting (B and C)were done after 48 hours. A, RNA prepared from the transfected cell lysates was tested by RT-PCR for expression of oncVav1 and osteopontin and normalized by the expression of 18S RNA in each sample. Experiments were done in triplicate and reproduced thrice. Bars, SE. B, NIH/ oncVav1 cells transiently transfected with a scVector (lane 1), an empty vector (Vector; lane 2), and si1Vav and si2Vav plasmids (si1Vav and si2Vav; lanes 3 and 4)were lysed 48 hours following transfection. Cell lysates were separated on SDS-PAGE and immunoblotted with anti-Vav1, anti-actin, and anti-Rac antibodies. Similar results were obtained in three independent experiments. C, a similar experiment to (B)was done using scrambled (scOPN)and effective OPNi siRNA duplexes. Cell lysates were separated on SDS-PAGE and immunoblotted with anti-osteopontin and anti-actin antibodies. Similar results were obtained in three independent experiments. Figure 3. Signaling pathways following siRNA-mediated knockdown of oncVav1 in NIH/oncVav1 cells stimulated with EGF. A, NIH3T3 cells (lanes 1 and 2)and NIH/oncVav1 cells (lanes 3 and 4)were transiently transfected with an empty drastically reduce the expression of osteopontin and oncVav1. vector. NIH/oncVav1 cells were also transiently transfected with si1Vav (lanes 5 and 6)and si2Vav ( lanes 7 and 8)plasmids. Cells were either nonstimulated Moreover, these results clearly indicate that oncVav1 operates (À; lanes 1, 3, 5, and 7)or stimulated with EGF (+; lanes 2, 4, 6, and 8). upstream of osteopontin. Cell lysates were immunoprecipitated with anti-Vav antibodies, and proteins were resolved on SDS-PAGE and then immunoblotted with either To assess how reduced oncVav1 and osteopontin expression anti-phosphotyrosine (top)or anti-Vav antibodies ( bottom). B to D, cell lysates affect signal transduction cascades in NIH/oncVav1 cells, we were prepared as in (A), except that NIH/oncVav1 cells were also transfected analyzed the level of oncVav1 tyrosine phosphorylation following with a scVector. Proteins were separated on SDS-PAGE and immunoblotted with either anti-phosphorylated ERK mAb (B, pERK, top)or anti-ERK mAbs EGF treatment of cells transfected with unrelated vector, si1Vav, or (B, bottom), anti-phosphorylated JNK mAb (C, pJNK, top)or anti-JNK mAbs si2Vav (Fig. 3A). oncVav1 was tyrosine phosphorylated in NIH/ (B, bottom), or anti-PLCg mAb (D, pPLCc,top)or anti-PLC g mAbs (D, bottom).

Cancer Res 2006; 66: (12). June 15, 2006 6186 www.aacrjournals.org

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 2006 American Association for Cancer Research. Vav1 and Osteopontin in Transformed Fibroblasts following stimulation of NIH3T3 cells with EGF and in EGF-treated NIH/oncVav1 cells transfected with a scVector or an unrelated vector (Fig. 3B). This increase is eliminated in cells transfected with either si1Vav or si2Vav despite the substantial amount of ERK protein present in these samples (Fig. 3B). Our results clearly show that oncVav1 is required for ERK phosphorylation in NIH/oncVav1 cells following EGF treatment. Mammalian wtVav1 has been shown to activate JNK in various cells, such as fibroblasts (26), mast cells (27), and T cells (28). We analyzed JNK phosphorylation in NIH/oncVav1 cells with or without knockdown of oncVav1. JNK phosphorylation was inhibited in NIH/oncVav1 cells following transfection with oncVav1 siRNA, although total JNK protein levels remained high (Fig. 3C). Vav1 controls Ca2+ flux by participating in the regulation of PLCg activation in the hematopoietic system (12, 13). However, it is not known if oncVav1 plays a role in PLCg phosphorylation in fibroblasts. We found that treatment with either si1Vav1 or si2Vav1 had no effect on PLCg1 phosphorylation (Fig. 3D). Thus, oncVav1 is critical for ERK and JNK phosphorylation but not for PLCg phosphorylation in NIH/oncVav1 cells (Fig. 3). Osteopontin gene expression in NIH/oncVav1 is affected by phosphorylation of ERK. We next examined the possibility that the up-regulation of osteopontin in NIH/oncVav1 cells might stem from the function of oncVav1 as a signal transducer protein. ERK phosphorylation is elevated to a greater extent in EGF-stimulated NIH/oncVav1 cells than in NIH/wtVav1 cells (Fig. 4A). Moreover, a basal level of phosphorylated ERK is noted in unstimulated NIH/ oncVav1 cells but not in NIH/wtVav1 cells (Fig. 4A). To examine whether increased phosphorylation of ERK in NIH/oncVav1 cells is responsible for the observed elevation in osteopontin expression, we treated NIH/wtVav1 and NIH/oncVav1 cells with a MEK inhibitor (UO126) and analyzed the levels of ERK phosphorylation and osteopontin protein expression (Fig. 4A and B). Treatment Figure 4. Osteopontin gene expression in NIH/oncVav1 is affected by with UO126 inhibited ERK phosphorylation in both NIH/wtVav1 phosphorylation of ERK (A and B), but its depletion does not affect ERK A phosphorylation (C). A, NIH/wtVav1 cells (lanes 1-4)and NIH/oncVav1 cells and NIH/oncVav1 cells (Fig. 4 ). Concomitantly, osteopontin (lanes 5-8)were either nonstimulated ( À; lanes 1, 3, 5, and 7)or stimulated with expression was markedly reduced in NIH/oncVav1 cells (Fig. 4A). EGF for 1 hour (+; lanes 2, 4, 6, and 8). Cells were either treated for two hours The level of osteopontin in NIH/wtVav1 cells remained barely with UO126 before EGF stimulation (+; lanes 3, 4, 7, and 8)or remained nontreated (À; lanes 1, 2, 5, and 6). Cell lysates were resolved on SDS-PAGE detectable (Fig. 4A). Analysis of osteopontin gene expression by RT- and then immunoblotted with either anti-pERK (top), anti-ERK (middle),or PCR confirms that UO126 inhibits osteopontin expression in NIH/ anti-osteopontin antibodies (bottom). Representative of one of two experiments. B, densitometry was done on blots from two separate experiments. Results oncVav1 (data not shown). are relative expression of either pERK/ERK or osteopontin/ERK. Bars, SE. The finding that UO126 blocked the expression of osteopontin in C, NIH3T3 cells (lanes 1 and 2)and NIH/oncVav1 cells ( lanes 3-8)were NIH/oncVav1 cells indicates that expression of this gene is coupled transiently transfected with an empty vector (lanes 1-4), scOPN (lanes 5 and 6), and OPNi siRNA duplexes (lanes 7 and 8). Cells were either nonstimulated to the ERK pathway. To further substantiate these results, we (À; lanes 1, 3, 5, and 7)or stimulated with EGF (+; lanes 2, 4, 6, and 8). examined whether signaling events, such as ERK phosphorylation, Cell lysates were resolved on SDS-PAGE and then immunoblotted with either are compromised by osteopontin depletion (Fig. 4C). Following anti-phosphorylated ERK mAb (top)or anti-ERK mAbs ( bottom). EGF stimulation, ERK was phosphorylated in NIH/oncVav1 cells transfected with either an empty vector and in cells treated with suffices for osteopontin overexpression (mRNA and protein levels) scOPN or OPNi (Fig. 4C). Thus, a reduction in osteopontin in cultured NIH/oncVav1 cells. expression did not alter ERK phosphorylation in contrast to the Tumorigenic properties of NIH/oncVav1 cells depleted of loss of ERK phosphorylation seen following reduction of oncVav1 oncVav1 orosteopontin. NIH/oncVav1 cells exhibit transformed expression (Fig. 3B). Osteopontin depletion had no effect on JNK morphology: the cells are round with many cytoplasmic and PLCg phosphorylation following EGF stimulation with the extensions. To determine whether this morphology is dependent same paradigm shown in Fig. 4C (data not shown). on oncVav1 expression, we transfected NIH3T3 and NIH/oncVav1 Taken together, the reduction in osteopontin expression due to cells with a GFP expression plasmid and either an empty inhibition of ERK phosphorylation (Fig. 4A and B) and the fact that expression plasmid (vector) or the plasmid carrying si1Vav1 and depletion of osteopontin does not affect ERK phosphorylation analyzed expression of oncVav1 and GFP and the morphology of (Fig. 4C) clearly indicate that expression of osteopontin can be the transfected cells (Fig. 5A). oncVav1 accumulated in the coupled to the ERK pathway. Because osteopontin expression in cytoplasm of NIH/oncVav1 cells but was not found in NIH3T3 NIH/oncVav1 cells is not markedly induced following EGF cells (Fig. 5). Following transfection with si1Vav1 plasmid, no stimulation (Figs. 2 and 4), it is conceivable that the basal level expression of oncVav1 was noted in NIH/oncVav1 cells express- of activation of ERK in these cells, but not in NIH/wtVav1 cells, ing GFP. Notably, these cells also underwent a dramatic change www.aacrjournals.org 6187 Cancer Res 2006; 66: (12). June 15, 2006

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 2006 American Association for Cancer Research. Cancer Research in cell shape, reverting to a nontransformed phenotype similar to NIH/oncVav1 cells exhibited a 45% reduction in invasion compared NIH3T3 cells (Fig. 5A). Transfection with si2Vav1 plasmid had with scOPN-treated cells (Fig. 6C and D), suggesting that similar effects (data not shown). Treatment of NIH/oncVav1 cells osteopontin is only one part of the mechanism by which oncVav1 with OPNi had no effect on the morphology of NIH/oncVav1 controls tumorigenesis. cells (Fig. 5B). NIH/oncVav1 cells are transformed cells, which have been shown to be metastatic (6). We tested whether NIH/oncVav1 cells Discussion depleted of oncVav1 or osteopontin retain their transforming We have shown here that oncVav1, but not wtVav1, up-regulates properties as manifested by growth in agar (Fig. 6A and B) and the osteopontin gene and explored the contribution of osteopontin Matrigel invasion (Fig. 6C and D). The growth in agar assay has to oncVav1-mediated transformation. Osteopontin is considered a been used many years as a justifiable assay to test the marker of the malignant phenotype (17, 18). Elevated levels of transforming properties of cells. NIH/oncVav1 cells transfected osteopontin in a wide variety of malignant cells and in the with si1Vav1 exhibit greatly reduced ability to grow in agar surrounding stroma of numerous human cancers suggest that it compared with NIH/oncVav1 cells transfected with the scVector may play a key role in tumorigenesis (17, 18). Currently, increased (Fig. 6A and B). OPNi-treated NIH/oncVav1 cells also exhibited a osteopontin expression is associated with tumor invasion, pro- considerable reduction in growth compared with the scOPN- gression, or metastasis in cancers of the breast, stomach, lung, treated cells (Fig. 6A and B). The size of the colonies of NIH/ prostate, liver, and colon (17, 18). Osteopontin is not expressed in oncVav1 treated with si1Vav1 was smaller than the ones observed the normal matrix of most tissues, other than mineralized tissues, in cells treated OPNi (Fig. 6B). but it is synthesized by cells in response to various stimuli and thus The in vitro Matrigel invasion assay is commonly considered to is considered to be a cytokine (29). be a measurement of the first step of tumor cell invasion (20). NIH/ Increased expression of osteopontin has also been found in cells oncVav1 cells transfected with si1Vav1 and si2Vav1 plasmids have transformed by other oncogenes. Elevated osteopontin expression greatly reduced ability to invade Matrigel (82% for si1Vav1 and 70% was noted in NIH3T3 fibroblasts transformed by ras (30), and these for si2Vav1), with levels of invasiveness similar to the non- ras-transformed fibroblasts require osteopontin for growth (31). transformed parent NIH3T3 cells (Fig. 6C and D). OPNi-treated The v-src oncogene stimulates osteopontin expression in murine

Figure 5. Morphology and expression of oncVav1 (A and B)and osteopontin ( B)in NIH/oncVav1 cells transfected with specific Vav1 plasmids and osteopontin siRNA duplexes. A, NIH3T3 and NIH/oncVav1 cells were cotransfected with an empty vector or Vav1-specific siRNA plasmids together with the GFP. Cells were permeabilized and stained with anti-Vav antibodies followed by Cy5-conjugated anti-rabbit antibodies (red). Cells were then analyzed by confocal microscopy. Green, GFP; Merge, an overlay of Vav and GFP staining; DIC, differential interference contrast image. B, a similar experiment to (A), except that NIH/oncVav1 cells were treated with scrambled (scOPN)and OPNi siRNA duplexes. Cells were stained as detailed above and analyzed for oncVav1 or osteopontin expression. More than one cell is presented in each panel.

Cancer Res 2006; 66: (12). June 15, 2006 6188 www.aacrjournals.org

Figure 6. Tumorigenic properties of NIH/oncVav1 cells transfected with Vav1 and osteopontin-specific siRNA plasmids: Growth in soft agar (A and B)and invasiveness (C and D).A,NIH/oncVav1 cells treated as indicated were trypsinized 48 hours following transfection, suspended in DMEM containing 0.4% agar and 10% calf serum, and plated onto a bottom layer containing 0.8% agar. Cells were plated at a density of 1 Â 105 per well in a six-well plate, and the number of colonies was counted 14 days later. Representative of three experiments in triplicates. Bars, SE. B, four representative colonies grown in agar for each treatment described in (A). C, NIH3T3, NIH/oncVav1, and NIH/oncVav1 cells transfected with Vav1 and osteopontin-specific siRNA plasmids or duplexes were assayed by the in vitro Matrigel assay that enables the estimation of invasiveness of cells through ECM. Representative of five experiments in duplicates. D, a representative staining of the invading cells described in (A). cells (32) and osteopontin synthesis is suppressed in cells that do transforming activity correlated with the activation of the c-Jun, not express c-src (33), suggesting that the wild-type gene may also Elk-1, and nuclear factor-nB (NF-nB) transcription factors as well as play a role in normal expression of osteopontin. It was recently increased transcription from the cyclin D1 promoter. They also reported that osteopontin expression is also induced in a model of showed that wtVav1 has little, if any, effect on any of these activities. murine leukemia caused by Bcr-Abl (34). These studies show that a However, this study also indicated that the oncVav1 used in our study variety of oncogenes can induce osteopontin, but none of them (oncVav1D1-66) does not induce cyclin D1 (9). Fernandez-Zapico directly compared osteopontin expression in the presence of the et al. reported a different effect of wtVav1 on cyclin D1 expression, corresponding proto-oncogenes. We show here not only that showing that wtVav1 up-regulates cyclin D1 in pancreatic tumors oncVav1 stimulates the expression of osteopontin but also that (35). This discrepancy might result from differences in the wtVav1 depresses its expression. This result raises the possibility experimental systems. Our observation that oncVav1 and wtVav1 that osteopontin might contribute to the transforming capability have opposite effects on osteopontin expression is unique. of oncVav1. Several mechanisms might explain the oncVav1-mediated up- Our demonstration that osteopontin is up-regulated in NIH/ regulation of osteopontin. One possibility is that this induction is oncVav1 cells is the first report of an effect of oncVav1 on direct, via an effect of oncVav1 on the osteopontin promoter as was downstream genes. Recently, several groups reported elevated shown for ras-mediated osteopontin overexpression (36). However, expression of cyclin D1 due to the presence of a different oncVav1 this option was ruled out because we could not detect any induction mutant (D1-186). Abe et al. (9) reported that the increase in Vav1 by oncVav1 of the osteopontin promoter plasmids (data not shown). www.aacrjournals.org 6189 Cancer Res 2006; 66: (12). June 15, 2006

The other possibility is that the osteopontin up-regulation in NIH/ GTPases, wtVav1 participates in their regulation, mainly manifested oncVav1 cells stems from changes in signaling pathways, such as by cytoskeleton reorganization (44). For example, Vav1 is involved ERK, in these cells. Indeed, treatment of NIH/oncVav1 cells with the in the transmission of migratory signals in T cells (45). Ablation of specific MEK1/2 inhibitor UO126 resulted in a marked decrease oncVav1 by Vav1-specific siRNA reverted the morphology of the in osteopontin expression along with down-regulation of ERK transformed NIH/oncVav1 cells, whereas ablation of osteopontin phosphorylation (Fig. 4). Osteopontin expression in NIH/wtVav1 expression had a minimal effect on cell morphology (Fig. 5), cells was barely detectable even before the treatment with UO126. suggesting that osteopontin does not contribute to cytoskeleton These data suggest that sustained ERK phosphorylation, such reorganization but to another process involved in motility. Similarly, as that observed in NIH/oncVav1 cells, is required to induce it was shown that the regulation of endothelial cell motility in the osteopontin (Fig. 4). These results are further supported by the fact presence of osteopontin is Rac independent (46). that ablation of osteopontin in NIH/oncVav1 cells does not change Osteopontin and Vav1 might be linked through integrin the level of EGF-induced ERK activation (Fig. 4C). Thus, in our signaling. wtVav1 plays a critical role as a downstream signal experimental system, ERK phosphorylation takes place upstream of transducer protein following integrin engagement (47, 48). h3 h osteopontin expression. The MAPK signaling cascade has also been Tyrosine phosphorylation and av 3-mediated adhesion are re- implicated in osteopontin regulation in various cell systems, such as quired for Vav1 association and Rho activation in leukocytes (47). HL-60 cells (37), JB6 cells (38), and human osteoblastic cells (39), and Additionally, Vav proteins were shown to play an essential role in response to injury of rat arterial muscle cells (40). The association coupling h2 to Rho GTPases and regulating multiple integrin- between ERK phosphorylation and osteopontin expression was induced events important in leukocyte adhesion and phagocytosis shown by the use of both MEK inhibitors (34, 37, 38, 40) and a (48). Osteopontin contains a RGD tripeptide sequence that can be h h h dominant-negative ERK plasmid (39). Interestingly, the dependence bound by av 3, av 1, and av 5 integrins on the cell surface (49). of osteopontin up-regulation on ERK activation was shown in cells The binding of integrins to osteopontin results in distinct transformed by another oncogene, Bcr-Abl, via a cascade, including functional consequences, including cell adhesion and cell migra- Ras, phosphatidylinositol 3-kinase, a-protein kinase C, Raf-1, and tion (50). Osteopontin also interacts with cellular receptor(s) other MEK (34). Together, these data suggest that activation of the MAPK than integrins, such as CD44, through domains that do not contain cascade is a crucial component in the regulation of osteopontin the RGD sequence (51, 52). CD44 has been implicated in a wide expression regardless of the stimulus. It is plausible that additional variety of cellular processes, including cell-cell and cell-matrix signaling cascades activated by oncVav1 in NIH3T3 cells also interactions, cell migration, metastasis, lymphocyte homing, and contribute to osteopontin up-regulation. T-cell activation (53). In NIH3T3 cells transformed by H-RasV12, Numerous studies have implicated wtVav1 in ERK (22–24), JNK osteopontin mainly interacts with CD44 and not with integrins (26–28) and PLCg (12, 13) phosphorylation, but most of these (19). This interaction contributes to the malignancy of these cells studies were done in hematopoietic cells, where wtVav1 is (19). It is conceivable that osteopontin functions in a similar physiologically active. Little was known thus far regarding the manner in NIH/oncVav1 cells, enhancing the effect of oncVav1 involvement of oncVav1 in these pathways in fibroblasts. Our studies transforming properties via a feed-forward mechanism. clearly show a role for oncVav1 in ERK and JNK phosphorylation Alternatively, the contribution of osteopontin to the metastatic but not in PLCg phosphorylation (Fig. 3). oncVav1 probably process could be a result of its ability to regulate the activity of influences ERK phosphorylation in a different manner than wtVav1 several extracellular matrix (ECM)–degrading proteins (54, 55). in T cells. In T cells, wtVav1 transduces TCR signals to Ras by Such enzymes play a key role in events, which enable cancer cells controlling the membrane recruitment of RasGRP1 and Sos1 and to invade into the surrounding environment (54, 55). Philip et al. Sos2 through the transmembrane adapter protein LAT (22). How- have shown that osteopontin up-regulates pro-matrix metal- ever, as several of the hematopoietic signaling proteins, including loproteinase-2 expression in a NF-nB-dependent fashion during LAT, are absent in fibroblasts, oncVav1 might function in these cells ECM invasion (54). Osteopontin also increases cell invasiveness in through its association with GRB2 (41) or through additional human mammary carcinoma through stimulation of urokinase unknown mechanisms. The lack of involvement of oncVav1 in PLCg plasminogen activator (55). The cell-matrix adhesive and cytokine- phosphorylation in fibroblasts (Fig. 3) is in agreement with our signaling properties of osteopontin provide mechanisms for previous study in T cells (13), pointing to a cardinal difference in mediating these invasive processes (54, 55). signaling pathways between oncVav1 and wtVav1. In summary, our studies reveal that osteopontin plays a role in Depleting oncVav1 in NIH/oncVav1 cells drastically reduces their the molecular mechanism that underlies transformation by ability to grow in soft agar and migrate through Matrigel (Fig. 6), a oncVav1 but that additional pathways are involved that do not model for tumor invasiveness. The fact that osteopontin ablation include osteopontin. More studies are necessary to identify also reduces the growth in soft agar and migration of NIH/oncVav1 additional genes that are differentially regulated by oncVav1 and cells, although to a lesser extent (Fig. 6), attests to the importance wtVav1. This is particularly relevant in light of recent reports that of osteopontin expression for tumorigenesis and cell migration, a Vav1 is an important gene in mammalian neoplasms (35, 56). critical step in tumor invasion and metastasis. Reorganization of the actin cytoskeleton is the primary mechanism of cell motility and is Acknowledgments essential for most types of cell migration. Actin reorganization is Received 10/17/2005; revised 2/15/2006; accepted 4/5/2006. regulated by Rho family small GTPases, such as Rho, Rac, and Cdc42 Grant support: Israel Science Foundation (599/04) and the Israel Cancer (42, 43). These proteins alternate between an inactive GDP-bound Association in memory of Profs. Dafna and Dov Israeli. The costs of publication of this article were defrayed in part by the payment of page state and an active GTP-bound state. When the protein is in the charges. This article must therefore be hereby marked advertisement in accordance active GTP-bound conformation, it interacts with effector proteins with 18 U.S.C. Section 1734 solely to indicate this fact. We thank Dr. Susan Lewis for editing the article, Y. Markson for help with the that propagate further signaling events that lead ultimately to the figures, Dr. M. Tarshish for assistance with the confocal microscopy, Dr. R. Bar-Shavit desired biological responses. As a known GEF for the Rho/Rac for the HCT cell line, and Dr. N. Hijaya for the osteopontin promoter plasmids.

Cancer Res 2006; 66: (12). June 15, 2006 6190 www.aacrjournals.org

References high and low human metastatic variants. Cancer Lett mediate phorbol ester-induced osteopontin expression. 1987;38:137–47. Int J Biochem Cell Biol 2002;34:1142–51. 1. Katzav S, Martin-Zanca D, Barbacid M. vav, a novel 21. Medico E, Gentile A, Lo Celso C, et al. Osteopontin is 39. Lai CF, Chaudhary L, Fausto A, et al. Erk is essential human oncogene derived from a locus ubiquitously an autocrine mediator of hepatocyte growth factor- for growth, differentiation, integrin expression, and cell expressed in hematopoietic cells. EMBO J 1989;8: induced invasive growth. Cancer Res 2001;61:5861–8. function in human osteoblastic cells. J Biol Chem 2001; 2283–90. 22. Reynolds LF, de Bettignies C, Norton T, et al. Vav1 276:14443–50. 2. Tybulewicz VL. Vav-family proteins in T-cell signaling. transduces T cell receptor signals to the activation of the 40. Moses S, Franzen A, Lovdahl C, et al. Injury-induced Curr Opin Immunol 2005;17:267–74. Ras/ERK pathway via LAT, Sos, and RasGRP1. J Biol osteopontin gene expression in rat arterial smooth 3. Bustelo XR. Regulatory and signaling properties of the Chem 2004;279:18239–46. muscle cells is dependent on mitogen-activated protein Vav family. Mol Cell Biol 2000;20:1461–77. 23. Villalba M, Hernandez J, Deckert M, et al. Vav kinases ERK1/ERK2. Arch Biochem Biophys 2001;396: 4. Hornstein I, Alcover A, Katzav S. Vav proteins, masters modulation of the Ras/MEK/ERK signaling pathway 133–7. of the world of cytoskeleton organization. Cell Signal plays a role in NFAT activation and CD69 up-regulation. 41. Ye ZS, Baltimore D. Binding of Vav to Grb2 through 2004;16:1–11. Eur J Immunol 2000;30:1587–96. dimerization of Src homology 3 domains. Proc Natl 5. Crespo P, Schuebel KE, Ostrom AA, et al. Phosphotyr- 24. Costello PS, Walters AE, Mee PJ, et al. The Rho-family Acad Sci U S A 1994;91:12629–33. osine-dependent activation of Rac-1 GDP/GTP ex- GTP exchange factor Vav is a critical transducer of T cell 42. Etienne-Manneville S, Hall A. Rho GTPases in cell change by the vav proto-oncogene product. Nature receptor signals to the calcium, ERK, and NF-nB biology. Nature 2002;420:629–35. 1997;385:169–72. pathways. Proc Natl Acad Sci U S A 1999;96:3035–40. 43. Van Aelst L, D’Souza-Schorey C. Rho GTPases and 6. del Peso L, Hernandez-Alcoceba R, Embade N, et al. 25. Cao Y, Janssen EM, Duncan AW, et al. Pleiotropic signaling networks. Genes Dev 1997;11:2295–322. Rho proteins induce metastatic properties in vivo. defects in TCR signaling in a Vav-1-null Jurkat T-cell 44. Olson MF, Pasteris NG, Gorski JL, et al. Faciogenital Oncogene 1997;15:3047–57. line. EMBO J 2002;21:4809–19. dysplasia protein (FGD1) and Vav, two related proteins 7. Katzav S, Cleveland JL, Heslop HE, et al. Loss of the 26. Crespo P, Bustelo XR, Aaronson DS, et al. Rac-1 required for normal embryonic development, are amino-terminal helix-loop-helix domain of the vav dependent stimulation of the JNK/SAPK signaling upstream regulators of Rho GTPases. Curr Biol 1996;6: proto-oncogene activates its transforming potential. pathway by Vav. Oncogene 1996;13:455–60. 1628–33. Mol Cell Biol 1991;11:1912–20. 27. Song JS, Haleem-Smith H, Arudchandran R, et al. 45. Vicente-Manzanares M, Cruz-Adalia A, Martin- 8. Coppola J, Bryant S, Koda T, et al. Mechanism of Tyrosine phosphorylation of Vav stimulates IL-6 pro- Cofreces NB, et al. Control of lymphocyte shape and activation of the vav protooncogene. Cell Growth Differ duction in mast cells by a Rac/c-Jun N-terminal kinase- the chemotactic response by the GTP exchange factor 1991;2:95–105. dependent pathway. J Immunol 1999;163:802–10. Vav. Blood 2005;105:3026–34. 9. Abe K, Whitehead IP, O’Bryan JP, et al. Involvement of 28. Salojin KV, Zhang J, Delovitch TL. TCR and CD28 are 46. Soga N, Connolly JO, Chellaiah M, et al. Rac regulates NH(2)-terminal sequences in the negative regulation of coupled via ZAP-70 to the activation of the Vav/Rac-1-/ vascular endothelial growth factor stimulated motility. Vav signaling and transforming activity. J Biol Chem PAK-1/p38 MAPK signaling pathway. J Immunol 1999; Cell Commun Adhes 2001;8:1–13. 1999;274:30410–8. 163:844–53. 47. Gao C, Schaefer E, Lakkis M, et al. h3 tyrosine h 10. Khosravi-Far R, Chrzanowska-Wodnicka M, Solski 29. Denhardt DT, Noda M. Osteopontin expression and phosphorylation and av 3-mediated adhesion are re- PA, et al. Dbl and Vav mediate transformation via function: role in bone remodeling. J Cell Biochem Suppl quired for Vav1 association and Rho activation in mitogen-activated protein kinase pathways that are 1998;30–1:92–102. leukocytes. J Biol Chem 2005;280:15422–9. distinct from those activated by oncogenic Ras. Mol Cell 30. Craig AM, Nemir M, Mukherjee BB, et al. Identifica- 48. Gakidis MA, Cullere X, Olson T, et al. Vav GEFs are h Biol 1994;14:6848–57. tion of the major phosphoprotein secreted by many required for 2 integrin-dependent functions of neu- 11. Aghazadeh B, Lowry WE, Huang XY, et al. Structural rodent cell lines as 2ar/osteopontin: enhanced expres- trophils. J Cell Biol 2004;166:273–82. basis for relief of autoinhibition of the Dbl homology sion in H-ras-transformed 3T3 cells. Biochem Biophys 49. Hu DD, Lin ECK, Kovach NL, et al. A biochem- domain of proto-oncogene Vav by tyrosine phosphory- Res Commun 1988;157:166–73. ical characterization of the binding of osteopontin lation. Cell 2000;102:625–33. 31. Wu Y, Denhardt DT, Rittling SR. Osteopontin is to integrin vh3,vh1 and vh5. J Biol Chem 1995;270: 12. Manetz TS, Gonzalez-Espinosa C, Arudchandran R, required for full expression of the transformed pheno- 2632–8. et al. Vav1 regulates phospholipase Cg activation and type by the ras oncogene. Br J Cancer 2000;83:156–63. 50. Liaw L, Skinner MP, Ranies EW, et al. The adhesive calcium responses in mast cells. Mol Cell Biol 2001;21: 32. Tezuka K, Denhardt DT, Rodan GA, et al. Stimulation of and migratory effects of osteopontin are mediated via h 3763–74. mouse osteopontin promoter by v-Src is mediated by a distinct cell surface integrins: role of v 3 in smooth 13. Groysman M, Hornstein I, Alcover A, et al. Vav1 and CCAAT box-binding factor. J Biol Chem 1996;271:22713–7. muscle migration to osteopontin in vitro. J Clin Invest Ly-GDI two regulators of Rho GTPases, function cooper- 33. Chackalaparampil I, Peri A, Nemir M, et al. Cells 1995;95:713–24. atively as signal transducers in T cell antigen receptor- in vivo and in vitro from osteopetrotic mice homozy- 51. Katagiri Y, Murakami M, Mori K, et al. Non-RGD induced pathways. J Biol Chem 2002;277:50121–30. gous for c-src disruption show suppression of synthesis domains of osteopontin promote cell adhesion without 14. Wu J, Katzav S, Weiss A. A functional T-cell receptor of osteopontin, a multifunctional extracellular matrix involving v integrins. J Cell Biochem 1996;62:123–31. signaling pathway is required for p95vav activity. Mol protein. Oncogene 1996;12:1457–67. 52. Weber GF, Ashkar S, Glimcher MJ, et al. Receptor- Cell Biol 1995;15:4337–46. 34. Hickey FB, England K, Cotter TG. Bcr-Abl regulates ligand interaction between CD44 and osteopontin 15. Katzav S. Single point mutations in the SH2 domain osteopontin transcription via Ras, PI-3K, aPKC, Raf-1, (Eta-1). Science 1996;271:509–12. impair the transforming potential of vav and fail to and MEK. J Leukoc Biol 2005;78:289–300. 53. Herrlich P, Zoller M, Pals ST, et al. CD44 splice activate proto-vav. Oncogene 1993;8:1757–63. 35. Fernandez-Zapico ME, Gonzalez-Paz NC, Weiss E, variants: metastasis meet lymphocytes. Immunol Today 16. Groysman M, Nagano M, Shaanan B, et al. et al. Ectopic expression of VAV1 reveals an unexpected 1993;14:395–9. Mutagenic analysis of Vav reveals that an intact SH3 role in pancreatic cancer tumorigenesis. Cancer Cell 54. Philip S, Bulbule A, Kundu GC. Osteopontin stimulates domain is required for transformation. Oncogene 1998; 2005;7:39–49. tumor growth and activation of promatrix metallopro- 17:1597–606. 36. Denhardt DT, Mistretta D, Chambers AF, et al. teinase-2 through nuclear factor-nB-mediated induction 17. Rittling SR, Chambers AF. Role of osteopontin in Transcriptional regulation of osteopontin and the of membrane type 1 matrix metalloproteinase in murine tumour progression. Br J Cancer 2004;90:1877–81. metastatic phenotype: evidence for a Ras-activated melanoma cells. J Biol Chem 2001;276:44926–35. 18. Wai PY, Kuo PC. The role of osteopontin in tumor enhancer in the human OPN promoter. Clin Exp 55. Tuck AB, Hota C, Chambers AF. Osteopontin (OPN)- metastasis. J Surg Res 2004;121:228–41. Metastasis 2003;20:77–4. induced increase in human mammary epithelial cell 19. Teramoto H, Castellone MD, Malek RL, et al. 37. Atkins KB, Simpson RU, Somerman MJ, et al. invasiveness is urokinase (uPA)-dependent. Breast Autocrine activation of an osteopontin-CD44-Rac path- Stimulation of osteopontin mRNA expression in HL-60 Cancer Res Treat 2001;70:197–204. way enhances invasion and transformation by H- cells is independent of differentiation. Arch Biochem 56. Hornstein I, Pikarsky E, Groysman M, et al. The RasV12. Oncogene 2005;24:489–501. Biophys 1997;343:157–63. haematopoietic specific signal transducer Vav1 is 20. Hendrix MJ, Seftor EA, Seftor RE, et al. A simple 38. Chang PL, Tucker MA, Hicks PH, et al. Novel protein expressed in a subset of human neuroblastomas. quantitative assay for studying the invasive potential of kinase C isoforms and mitogen-activated kinase kinase J Pathol 2003;199:526–33.

www.aacrjournals.org 6191 Cancer Res 2006; 66: (12). June 15, 2006

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 2006 American Association for Cancer Research. Osteopontin Is an Oncogenic Vav1− but not Wild-type Vav1− Responsive Gene: Implications for Fibroblast Transformation

Vered Schapira, Galit Lazer and Shulamit Katzav

Cancer Res 2006;66:6183-6191.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/66/12/6183

Cited articles This article cites 56 articles, 23 of which you can access for free at: http://cancerres.aacrjournals.org/content/66/12/6183.full#ref-list-1

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/66/12/6183. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.