Oncogene (2008) 27, 2728–2736 & 2008 Nature Publishing Group All rights reserved 0950-9232/08 $30.00 www.nature.com/onc ORIGINAL ARTICLE Activation of Rac1 and the exchange factor Vav3 are involved in NPM-ALKsignaling in anaplastic large cell lymphomas

A Colomba1,2, D Courilleau1,2, D Ramel1,2, DD Billadeau3,4, E Espinos1,2, G Delsol1,2, B Payrastre1,2 and F Gaits-Iacovoni1,2

1INSERM, U563, Centre de Physiopathologie de Toulouse Purpan, Toulouse, France; 2Universite´ Toulouse III Paul-Sabatier, IFR30, Toulouse, France; 3Division of Developmental Oncology Research, Mayo Clinic College of Medicine, Rochester, MN, USA and 4Department of Immunology, Mayo Clinic College of Medicine, Rochester, MN, USA

The majority of anaplastic large cell lymphomas (ALCLs) (Lamant et al., 1996; Duyster et al., 2001; Falini, 2001). express the nucleophosmin-anaplastic lymphoma kinase The constitutive activity of NPM-ALK (NPM-ALK) fusion , which is oncogenic due to its is responsible for malignant transformation of fibro- constitutive tyrosine kinase activity. Transformation by blasts and lymphoid cells and was shown to induce B- NPM-ALKnot only increases proliferation, but also and T-cell lymphomas in transgenic mice (Chiarle et al., modifies cell shape and motility in both lymphoid and 2003). fibroblastic cells. We report that the Rac1 GTPase, a NPM-ALK-induced transformation depends on the known cytoskeletal regulator, is activated by NPM-ALK activation of signaling pathways shared by many in ALCL cell lines (Karpas 299 and Cost) and transfected oncogenic tyrosine kinases. Pro-mitogenic functions cells (lymphoid Ba/F3 cells, NIH-3T3 fibroblasts). We include binding of adaptors, such as Shc, Grb2 and have identified Vav3 as one of the exchange factors IRS1, to regulators of the Erk pathway, phospholipase involved in Rac1 activation. Stimulation of Vav3 and Cg (PLCg), tyrosine and the proto- Rac1 by NPM-ALKis under the control of Src kinases. It oncogene pp60c-src (Fujimoto et al., 1996; Bai et al., involves formation of a signaling complex between NPM- 1998; Cussac et al., 2004; Honorat et al., 2006; Marzec ALK, pp60c-src, Lyn and Vav3, in which Vav3 associates et al., 2007). Antiapoptoticfunctionsare related to the with tyrosine 343 of NPM-ALKvia its SH2 domain. activation of the survival phosphatidylinositol 3-kinase Moreover, Vav3 is phosphorylated in NPM-ALKpositive (PI3K)/AKT pathway and of the Jak/STAT3–5 module biopsies from patients suffering from ALCL, demonstrat- (Bai et al., 2000; Amin et al., 2003; Chiarle et al., 2005). ing the pathological relevance of this observation. The use However, a hierarchy of downstream signaling, in of Vav3-specific shRNA and a dominant negative Rac1 terms of importance for transformation, diagnosis and mutant demonstrates the central role of GTPases in prognosis of the disease, remains to be established. With NPM-ALKelicited motility and invasion. this goal in mind, several large or medium scale Oncogene (2008) 27, 2728–2736; doi:10.1038/sj.onc.1210921; proteomicand transcriptomicstudies were undertaken published online 12 November 2007 by several groups (Lamant et al., 2006; Lim and Elenitoba-Johnson, 2006). New partners have been Keywords: anaplasticlarge celllymphomas; NPM- found that account for different functions of NPM- ALK; Rho GTPases; Vav3 ALK, such as the regulation of mRNA turnover (Fawal et al., 2006). Among others, regulating cell shape and cytoskeleton plasticity, two important features altered in transformed cells, were identified, leading to investigations of molecules classified as Introduction ‘cytoskeleton and motility regulators’ in the context of ALCLs (Crockett et al., 2004; Cussac et al., 2006). Anaplasticlarge celllymphomas (ALCLs), a subtype of Ambrogio et al. (2005) reported the association of high-grade non-Hodgkin’s lymphomas of T or null NPM-ALK with p130Crk associated substrate phenotype, are characterized by the aberrant expression (p130Cas) and described its role in actin depolymeriza- of the oncogenic fusion protein nucleophosmin-anaplas- tion and transformation. Along these lines, NPM-ALK ticlymphoma kinase (NPM-ALK) in 75% of cases expression alters fibroblasts shape dramatically. They display an elongated phenotype with extensions similar to what is observed in PC12 cells transformed by the Correspondence: Dr F Gaits-Iacovoni, INSERM, U563, Dpt d’Onco- native ALK receptor, indicating that the kinase activity ` se, Signalisation et Innovation the´ rapeutique, CHU Purpan, BP is responsible for the change in morphology. 3028, Toulouse cedex 3, Midi pyrenees 31024, France. E-mail: [email protected] Our group and others have described modifications of Received 11 May 2007; revised 27 September 2007; accepted 16 October the expression of regulators of the Rho GTPases 2007; published online 12 November 2007 (Crockett et al., 2004; Cussac et al., 2006). We observed Role of GTPases in NPM-ALK( þ ) ALCLs A Colomba et al 2729 the extinction of the GTPase inhibitor RhoGDI2 in the First, we took advantage of the small molecule inhibitor proteome of an NPM-ALK( þ ) cell line, an effect WHI-154 that inhibits the ALK kinase (Marzec et al., shown to correlate with higher metastatic activity and 2005). Inactivation of NPM-ALK resulted in a marked poor prognosis in bladder cancers (Theodorescu et al., decrease in Rac1 activation (Figure 1a). Second, we 2004; Cussac et al., 2006). These findings suggested that examined Rac1 in cell lines of independent origin (NIH- upregulation of Rho GTPases might be of importance 3T3 fibroblasts and Ba/F3 lymphoid cells) that were during the progression of the disease. Recently, about 45 demonstrated to become transformed by stable expres- proteins were reported to interact with NPM-ALK, sion of ALK oncogenic fusions (Armstrong et al., 2004). including various Rho GTPase activating proteins NPM-ALK-dependent Rac1 activation was again ob- whose pattern of expression was also altered in a study served (Figures 1b and c). of the transcriptome of NPM-ALK( þ ) cells (Crockett Small GTPase activation requires guanosine exchange et al., 2004; Lamant et al., 2006). Rho GTPases mediate factors (GEFs), which can be regulated by PI3K many aspects of cell biology including proliferation, products and kinases of the Src family (Hall, 2005). regulation of the cell survival, polarity, adhesion, The p85 regulatory subunit of PI3K and pp60c-src were membrane trafficking and motility (Hall, 2005). The identified as downstream targets of NPM-ALK. Treat- high incidence of overexpression of some GTPases ment of serum and IL3-deprived Ba/F3 cells with 25 mM (RhoA, RhoC, Rac1, Rac3 and Cdc42) or their of LY294002 (PI3K inhibitor) or 2 mM of SU6656 regulators in human tumors suggests that GTPases play (indolinone inhibitor of Srckinases) abolished Rac1 a role in carcinogenesis (Sahai and Marshall, 2002). The activation (Figure 1c), showing that PI3K and Src most studied members of the family are RhoA, Rac1 kinases are important for NPM-ALK to signal to the and Cdc42, which exert their transformant effects by GTPase. We previously demonstrated that NPM-ALK regulating cell cycle progression via the cyclin-depen- is a substrate for pp60c-src (Cussac et al., 2004). Hence, dent kinases, and promoting migration and metastasis SU6656 treatment reduced NPM-ALK phosphoryla- through regulation of cytoskeleton dynamics (Sahai and tion, making it difficult to conclude on a direct role of Marshall, 2002). Rac1 and Cdc42 regulate actin poly- Src on Rac1 activation pathway (Figure 1c). To merization through the Arp2/3 complex with Rac1 overcome this, we used RNA silencing to target pp60c-src involved in the generation of motile structures and and Lyn, two Srckinases involved in hematological Cdc42 in the establishment of polarity. RhoA organizes malignancies (Cussac et al., 2004; Contri et al., 2005; stress fibers predominantly through its effector Rho Thompson et al., 2005), and found that both kinases kinase (Hall, 1998). In addition, Rho GTPase signaling could regulate NPM-ALK Y664 autophosphorylation was demonstrated to be necessary for the oncogenicity (not shown). Activation of PAK1 (p21-activated kinase), of other proteins, especially for oncogenes derived from a downstream target of Rac1, demonstrated the same receptor tyrosine kinases, such as EGFR, IGFR, MET pattern of regulation as the GTPase, as shown with or RET (Aznar et al., 2004; Titus et al., 2005). antibodies to the active phosphorylated form of the In this study, we demonstrate that Rac1 is activated in kinase, indicating the functional relevance of Rac1 NPM-ALK expressing cells and is regulated by PI3K activation in terms of downstream signaling (Figure 1c). and Srcfamily kinases. The Vav3 proto-oncogeneis Finally, evaluation of the activation status of RhoA involved in bridging NPM-ALK and Rac1. The NPM- and Cdc42 failed to demonstrate significant modifica- ALK chimera forms a multiprotein complex containing tions in ALCLs and transfected cells (Figures 1d and e). pp60c-src, Lyn and phosphorylated Vav3. Importantly, we observed activation of Vav3 in tumors from patients developing ALCL. Altering either Vav3 or the Rac The proto-oncogene Vav3 is activated downstream of pathway, by RNA interference or with dominant NPM-ALK negative mutants and toxins, blocked invasion by The Vav proto-oncogenes are the only GEFs with a NPM-ALK( þ ) cells. Altogether, our data demonstrate structural hallmark of signal transducer proteins repre- a critical role for Rho GTPases in ALCL. sented by the SH3-SH2-SH3 (Src homology 2 or 3) module at their C terminus. This unique feature suggests that they could act as nucleation points for multiple signaling complexes after being recruited by tyrosine Results kinase receptors (Bustelo, 2000; Hornstein et al., 2004). They are activated by phosphorylation by members of NPM-ALK activates the GTPase Rac1 via PI3K and Src the Srcfamily, and Vav1 and Vav3 were found to be Activation of the Rho GTPases was first studied in two associated with NPM-ALK partners, such as Grb2, Shc, NPM-ALK positive ALCL cell lines, Karpas 299 the p85 regulatory unit of PI3K, pp60c-src and PLCg (common type) and Cost (small cells, aggressive variant) (Bustelo, 2001). We therefore checked whether they (Falini et al., 1998; Lamant et al., 2004). Pull-down could be targets of NPM-ALK by studying the assays demonstrated that Rac1 was strongly activated in activating phosphorylation on Y174 of Vav1 and Y173 both cell lines (Figure 1a). NPM-ALK activation can be of Vav3 with specific antibodies. Although some monitored by its autophosphorylation on tyrosine 664. phosphorylation could be detected, inhibition of ALK We used two complementary approaches to determine did not affect the level of Vav1 Y174 phosphorylation, whether NPM-ALK was responsible for Rac1 activation. suggesting that NPM-ALK does not regulate Vav1 in

Oncogene Role of GTPases in NPM-ALK( þ ) ALCLs A Colomba et al 2730

Figure 1 Rac1 is activated downstream of NPM-ALK. Levels of active Rac1 were evaluated by pull-down experiments with GST- PBD as described in Materials and methods in (a) Karpas 299 and Cost cells treated for 1 h with 15 mM of the NPM-ALK inhibitor WHI-154, and (b) Control or NPM-ALK expressing NIH-3T3 cells, (c) Rac1 activation is dependent on PI3K and Src. Control or NPM-ALK expressing Ba/F3 cells were treated for 30 min with 25 mM of the PI3K inhibitor LY294002 or 2 mM of the Srcinhibitor SU6656. Active Rac1-GTP and total Rac1 in the extracts were assessed by western blotting with antibodies to Rac1. The activation of NPM-ALK was followed with antibodies to the phosphorylated Y664 (pNPM-ALK(Y664)) and NPM-ALK level with ALK antibody. Phosphorylation of RaceffectorsPAK1/2 was followed with anti-phospho-PAK1/2 (S199–204/S192–197) antibodies. a- Tubulin was used as loading control. (d) and (e) Karpas 299, Cost, Ba/F3 and NIH-3T3 cells were treated as above then levels of active Cdc42 and active RhoA were evaluated by pull-down experiments with GST-PBD and GST-RBD, respectively. Active Cdc42-GTP, RhoA-GTP, total Cdc42 and RhoA in the extracts were assessed by western blotting with antibodies to Cdc42 and RhoA. Data are representative of three to five experiments.

ALCLs (Figure 2). Conversely, Figure 3a shows that lymph nodes or peripheral blood lymphocytes, demon- NPM-ALK expression resulted in a robust phosphor- strating the pathophysiological relevance of this ob- ylation of Vav3 in both Ba/F3 and NIH-3T3 cells. The servation (Figure 3c). same observation was made in Karpas 299 and Cost In addition, we investigated whether Srckinases were cells where ALK inhibition decreased Vav3 phosphor- involved in Vav3 phosphorylation. Treatment of Ba/F3 ylation significantly (Figure 3b). We evaluated the status cells with SU6656 resulted in a decreased signal of Vav3 in protein extracts from four frozen lymph (Figure 4a, left panel). We then used RNA interference nodes from patients suffering from NPM-ALK( þ ) to target pp60c-src and Lyn. RNA silencing of pp60c-src ALCLs after immunohistochemistry analysis of the demonstrated that it was responsible for Vav3 phos- biopsies (not shown). Again, Vav3 phosphorylation phorylation in this model (Figure 4a, right panel). increased in tumor samples compared with control Finally, we transfected cells with the L211Q inactive

Oncogene Role of GTPases in NPM-ALK( þ ) ALCLs A Colomba et al 2731 mutant of Vav3 containing a point mutation in the levels were reduced, showing that functional Vav3 is exchange DH (Dbl homology) domain (Figure 4b). In required for Rac1 activation downstream of NPM- addition, we depleted Vav3 with specific short hairpin ALK. RNAs (shRNAs) (Figure 4c). In both cases, Rac1GTP The SH2 domain of Vav3 drives its association with phosphorylated NPM-ALK Immunoprecipitations with an antibody directed against ALK demonstrated that a complex containing phos- phorylated Vav3, pp60c-src and Lyn co-precipitated with the active oncogene (Figure 5a). In all cases, a decrease (but not abrogation) in the association is observed upon WHI-154 treatment suggesting that activation of NPM- ALK is important in the stabilization of the complex (Figure 5a). Figure 5b shows that the reciprocal immunoprecipitation of Vav3 also led to the same Figure 2 Modification of NPM-ALK activity does not affect results. Overexpression of Vav3 mutants in NPM-ALK phosphorylation of Vav1. The activating phosphorylation on Y174 transfected cells confirmed the association of NPM- of Vav1 was measured on Vav1 immunoprecipitates from Karpas ALK, Vav3 and active Src kinases in this model 299 and Cost cells treated for 1 h with 15 mM WHI-154. Proteins (Figure 5c). Indeed, the Y173F mutant that was shown present in the immunoprecipitates were analysed by western blotting with polyclonal anti-Vav1 and anti-pVav1 (Y174) anti- to adopt an open conformation facilitating access to bodies. The activation of NPM-ALK was followed with the anti- partners and to GTPases (Llorca et al., 2005), binds to pNPM-ALK(Y664) antibody, loading was assessed with the ALK NPM-ALK strongly (Figure 5c). Interestingly, when we antibody. Data are representative of two experiments. expressed the R697A mutant with a disabled phospho- tyrosine binding SH2 domain, no binding to NPM-ALK was observed, demonstrating that the SH2 domain is crucial for the interaction (Figure 5c).

Figure 4 Vav3 participates in Rac1 activation downstream of NPM-ALK. (a) NPM-ALK expressing Ba/F3 cells were treated for 30 min with 2 mM SU6656 or 1 h with 15 mM WHI-154 (left panel) or nucleofected with nothing (control) or 4 mM of siRNA targeting Figure 3 The exchange factor Vav3 is phosphorylated in NPM- pp60c-src or Lyn (right panel). Total cell lysates were immunoblotted ALK expressing cells and in tumors from patients. The activating with the indicated antibodies. Tubulin was used as a loading phosphorylation on Y173 of Vav3 was measured in total lysates control. (b) Control or NPM-ALK expressing NIH-3T3 cells were from (a) control or NPM-ALK expressing Ba/F3 cells or NIH-3T3 nontransfected (NT) or transfected with a plasmid expressing the cells, and (b) Karpas 299 and Cost cells treated for 1 h with 15 mM GEF inactive L211Q human Vav3 mutant. Active Rac1-GTP was WHI-154. Lysates were immunoblotted with anti-pVav3 (Y173), measured by pull-down experiments. Expression of the Vav3 anti-pNPM-ALK (Y664), anti-ALK or anti-Vav3 antibodies as mutant was assessed in total lysates with antibodies to human loading control. Data are representative of three experiments. (c) Vav3. (c) NPM-ALK expressing NIH-3T3 cells were not trans- Total lysates from two frozen control (LN) or four NPM-ALK fected (NT) or transfected with control vector (control) or a vector positive lymph nodes and peripheral blood lymphocytes (PBL) expressing short hairpin RNA (shRNA) to Vav3. After 40 h, levels from two healthy donors were immunoblotted with anti-pVav3 of Rac1-GTP were assessed by pull-downs. The level of Vav3 was (Y173) or anti-human Vav3 antibodies. measured with anti-mouse Vav3.

Oncogene Role of GTPases in NPM-ALK( þ ) ALCLs A Colomba et al 2732

Figure 5 Vav3 forms a complex with NPM-ALK and Src kinases. (a) NPM-ALK was immunoprecipitated with ALK1 antibodies from Karpas 299 and Cost cells treated for 1 h with 15 mM WHI-154. Proteins present in the immunoprecipitates were analysed by western blotting with the indicated antibodies. (b) Vav3 was immunoprecipitated from Karpas 299 and Cost cells with anti-Vav3 antibodies. Proteins in the complexes were analysed as described above. (c) Control or NPM-ALK expressing NIH-3T3 cells were nontransfected (NT), or transfected with wild-type human Vav3 (WT), active open Y173F Vav3, GEF inactive L211Q Vav3 or SH2 defective R697A mutants. Twenty-four hours after transfection, proteins in NPM-ALK immunoprecipitates were analysed by immunoblotting. Expression of the mutants in total lysates was assessed with anti-human Vav3 antibody. (d) HEK293 cells were transfected with vectors expressing wild-type Vav3 and NPM-ALK mutated on the indicated tyrosines. Twenty-four hours after transfection, proteins in NPM-ALK immunoprecipitates were analysed with the indicated antibodies. Data are representative of two to four experiments.

To determine which NPM-ALK tyrosine residue was Y338F, Y342F and Y343F, had reduced kinase activity required to bind Vav3, we co-transfected HEK293 cells (Duyster et al., 2001).The fact that two of them are still with wild-type Vav3 and different mutants of NPM- capable of binding Vav3 indicates that Y343 is probably ALK, in which targeted tyrosines were replaced by phosphorylated by another kinase still to identify. phenylalanines (Y338F, Y342F, Y343F, Y418F (dock- ing for pp60c-src) and Y664F (docking for PLCg) mutants). Our data indicate that among the tested Small GTPases drive NPM-ALK-induced migration mutants, only Y343F had an impact on Vav3 binding NPM-ALK was reported to affect cell morphology, (Figure 5d). pp60c-src is not a physical intermediate of adhesion and migration in various cell types. Those Vav3 and NPM-ALK association since mutation of its functions have clearly been attributed to GTPases of the (Y418F) had no effect on the association. Rho family in many normal cell types and in tumoral Interestingly, assessment of the intrinsictyrosine kinase cells (Sahai and Marshall, 2002; Hall, 2005). In invasion activity of NPM-ALK with the anti-pNPM-ALK(Y664) assays through 3D Matrigel chambers, transfection of antibody confirmed that three of the mutants, namely, adherent fibroblasts with the dominant negative mutant

Oncogene Role of GTPases in NPM-ALK( þ ) ALCLs A Colomba et al 2733 Rac1T17N completely blocked cellular migration of Racwere involved in BCR-ABL signaling in leukemia NIH-3T3 NPM-ALK cells (Figure 6a). Similarly, (Bassermann et al., 2002; Cho et al., 2005). Over- depletion of Vav3 by RNA interference also reduced expression of active Vav1 and Vav3 mutants have migration, confirming that Vav3/Rac1 signaling plays a phenotypes reminiscent of NPM-ALK transformed cells major role in mediating NPM-ALK effects on cellular (Movilla and Bustelo, 1999; Zeng et al., 2000; Hornstein migration and invasion in vitro (Figure 6b). et al., 2004). Although Vav2 was shown to act down- stream of tyrosine kinases such as the EGF and PDGF receptors, there are major differences between Vav2 and NPM-ALK signaling. Specifically, overexpression of Discussion Vav2 generates strong stress fibers while NPM-ALK cells are characterized by a disappearance of actin cables We report the PI3K- and Src-dependent activation of (Liu and Burridge, 2000; Ambrogio et al., 2005). Rac1 downstream of NPM-ALK in two ALCL cell lines Moreover, Vav2 was described as being more prone to (Karpas 299 and Cost), Ba/F3 and NIH-3T3 transfected activate RhoA than Vav1 and Vav3, which fits with its cells. To find the link between NPM-ALK and Rac, we action on the actin cytoskeleton but does not match the focused on the Vav family of GEFs (Vav1, Vav2 and pattern of NPM-ALK-induced GTPase activation since Vav3). In addition to their GEF activity, Vavs have the RhoA activity was fairly weak and not affected by the unique feature of acting as docking proteins, which oncogene. These observations do not favor a role of makes them excellent candidates for relaying NPM- Vav2 in NPM-ALK transformation. Moreover, we ALK functions (Bustelo, 2001). Indeed, these GEFs are demonstrated that no difference in Vav1 activating major regulators of lymphocyte function, and Vav1 and phosphorylation was observed when NPM-ALK

1200 Control GFP-Rac1T17N 1000 GFP-Rac1N17

GFP 800 NPM-ALK 600 pNPM-ALK (Y664) 700 Tubulin

Percent of migrating cells 200

0 Control NPM-ALK

1600 Control 1400 shRNAVav3

1200 Control shRNA Vav3 Vav3 1000 NPM-ALK 800 pNPM-ALK(Y664)

600 Tubulin

700 Percent of migrating cells

200

0

Control NPM-ALK Figure 6 Small GTPases are required for NPM-ALK-dependent migration in vitro.(a). Control or NPM-ALK expressing NIH-3T3 cells were transfected with pcDNA3-EGFP-Rac1T17N or the control vector. Thirty hours after transfection, cells were plated on Matrigel-coated inserts and the number of migrating GFP positive cells was evaluated as percentage of control cells. Lysates were immunoblotted with anti-GFP to control EGFP-Rac1T17N expression. (b) Cells were transfected with the short hairpin RNAs (shRNA) Vav3 targeting vector or with the control vector and analysed as above. Lysates were immunoblotted with anti-mouse Vav3 to control Vav3 silencing. Control, cells transfected with the control vector; shRNA Vav3, cells transfected with the Vav3 targeting vector. The levels and the activation of NPM-ALK were assessed with anti-ALK and anti-pNPM-ALK(Y664) antibodies. Data are expressed as mean±s.e.m.

Oncogene Role of GTPases in NPM-ALK( þ ) ALCLs A Colomba et al 2734 activity was challenged. Conversely, Vav3 is phosphory- tested Clostridium toxins (Petit et al., 2003; Genth et al., lated on the activating Y173 downstream of NPM-ALK 2006). Both Toxin B from Clostridium difficile (affects in ALCL and transfected cells as well as in lymph node Rac, RhoA and Cdc42) and LT-IP82 from Clostridium biopsies originating from patients suffering from NPM- sordellii (specific for Rac, Ras, Ral and Rap) affected ALK( þ ) lymphomas, suggesting that this pathway NPM-ALK-induced migration but also blocked prolif- could be a valid target in the human pathology. Because eration, demonstrating that the role of GTPases down- we demonstrated that it is ALK kinase activity that stream of NPM-ALK is broader than migration and drives Rac1 activation, it is likely that other transloca- invasion (Colomba, A and Gaits-Iacovoni, F., unpub- tions involving the ALK kinase domain would also lished). Altogether, our observations demonstrate a activate Rac1. A broader study will determine if central role for Rac1 in NPM-ALK positive lympho- activation of Vav3/Rac1 is a common feature of mas. These data illustrate how an oncogene, here NPM- neoplasia resulting from deregulated ALK activity. ALK, can potentiate its transforming activity by It is now well accepted that pathways regulated by recruiting proto-oncogenes (such as pp60c-src and Vav3) Rho GTPases are very important in human cancers. in its vicinity. Interestingly, both the common type and a Their role in tumorigenesis was first demonstrated in small cell variant of ALCL display the same pattern of fibroblasts overexpressing dominant positive forms of activation of Rho GTPases, indicating that targeting RhoA, Rac1 or Cdc42. Not only was their proliferation GTPases could be a therapeutic avenue valid to treat not deregulated, but they also induced lung metastasis when only the primary disease, but also prevent relapses of the subcutaneously grafted in mice. Since then, they were most aggressive subtypes. shown to be necessary for transformation evoked by oncogenes such as Ras or receptor tyrosine kinases (Sahai and Marshall, 2002; Titus et al., 2005). Beside Materials and methods their effects on cell proliferation, GTPases are central modulators of the actin and microtubule cytoskeleton, Reagents and antibodies cell adhesion and motility (Hall, 1998). Using an EGFP Peripheral lymph nodes biopsies were obtained from patients tagged version of the dominant negative mutant diagnosed for NPM-ALK( þ ) ALCLs at the Department of Rac1T17N, we demonstrated that small GTPases drove Pathology of Toulouse University Medical Center (France), NPM-ALK elicited migration. In fact, the balance after informed consent. The study was approved by the between Rac, Cdc42 and RhoA activities determines institutional review board of the Purpan’s Hospital, Toulouse, cell morphology and migration. In ALCLs and NPM- France. Cell culture reagents were purchased from Invitrogen ALK transfected cells, Rac1 activity was strong while (Carlsbad, CA, USA). SU6656 and WHI-154 were from Calbiochem (San Diego, CA, USA). All other chemicals were RhoA-GTP was difficult to detect. Indeed, it was from Sigma-Aldrich (St Louis, MO, USA). Antibodies used demonstrated that Rac activation can downregulate were ALK1 (DakoCytomation, Glostrup, Denmark); Rac1, RhoA, leading to the dissolution of focal adhesions and human Vav3 and mouse Vav3 (Upstate Biotechnology, Lake increasing cell motility (Rottner et al., 1999). We Placid, NY); a-tubulin (Sigma-Aldrich); ALK, pPAK1/2 observed a 10-fold decrease in the number of focal (S199–204/S192–197), pSrc(Y416), pALK (Y1604) that adhesions associated with the expression of NPM-ALK recognizes Y664 on NPM-ALK called pNPM-ALK in NIH-3T3 fibroblasts (not shown), the biological (Y664) (Cell Signaling Technologies, Beverly, MA, USA); impact of a cross-regulation between Rac1 and RhoA pVav3(Y173) (Biosource International); Vav1, pVav1(Y174), in ALCL is currently under investigation. Interestingly, Cdc42, RhoA, c-src (clone H-12), Lyn and GFP (Santa Cruz when we depleted cells for Vav3 using RNA inter- Biotechnologies Inc., Santa Cruz, CA, USA). Horseradish peroxidase-conjugated goat anti-mouse and goat anti-rabbit ference, we also observed a blockade of migration immunoglobulin antisera were from Promega (Madison, WI, through 3D Matrigel that was less intense than USA). Rac1T17N (66.7% inhibition for shVav3 versus 84.25% for Rac1T17N). This Rac1 mutant acts by sequestering the upstream Rho GEFs and thereby Cell culture, plasmids and transfection HEK293 and NIH-3T3 fibroblasts were maintained in inhibiting endogenous Rho GTPases. It then affects Dulbecco’s modified Eagle’s medium and Ba/F3 lymphoid more than one GEF/GTPase couple and its stronger cells in RPMI 1640 containing 2 ng mlÀ1 murine recombinant effect on migration suggests several GEFs act down- IL-3 (mrIL3) (R&D Systems Europe, Abingdon, UK). Media stream of NPM-ALK. Accordingly, the use of the Rac were supplemented with 10% fetal calf serum (FCS), GEF inhibitor NSC23766 (Gao et al., 2004), reported 100 U mlÀ1 penicillin, 100 mgmlÀ1 streptomycin and 0.5 mg mlÀ1 not to act on Vav, resulted in a blockade of migration geneticin (G418) for cells stably expressing NPM-ALK. similar to Vav3 depletion (61.28% of inhibition). We are Human ALCLs cell lines Karpas 299 and Cost were cultured currently investigating the synergy between different in Iscove’s modified Dulbecco’s medium supplemented with GEFs and their regulatory pathways in ALCLs. 15% FCS. Before immunoprecipitation, GTPases activation Recently, small GTPases have become potential and western blot studies, Karpas 299, Cost and Ba/F3 cells were serum and IL3 starved overnight and NIH-3T3 cells were candidates for anticancer therapy. Different types of maintained in 2% serum overnight. NIH-3T3 were transfected natural molecules and synthetic drugs that inactivate using Lipofectamine (Invitrogen) and HEK293 with Effectene GTPases or their effectors are available, some of them (Qiagen, Valencia, CA, USA) according to the manufacturer’s show striking antineoplasticor antimetastaticactivity instructions. Ba/F3 NPM-ALK cells were transiently (Aznar et al., 2004; Fritz and Kaina, 2006). We have nucleofected using the Amaxa technology (Amaxa, Koeln,

Oncogene Role of GTPases in NPM-ALK( þ ) ALCLs A Colomba et al 2735 Germany). Briefly, 2 Â 106 cells were nucleofected using the deoxycholate, 0.1% SDS, 1 mM PMSF, 50 mM NaF, 1 mM c-src À1 À1 Amaxa solution V, 4 mM of siRNA smartpool pp60 or Na3VO4,10mgml leupeptin and 2 mgml aprotinin) and Lyn (Dharmacon Inc., Lafayette, CO, USA) and the program clarified lysates were incubated with 30 mg GST-PBD (PAK1 X-01 following the Amaxa guidelines. Cells were lysed 48 h Binding Domain for Rac1 and Cdc42) or GST-RBD post nucleofection. pcDNA3 containing wild-type NPM-ALK (Rhotekin Binding Domain for RhoA) bound to glutathione or the Y338F, Y342F, Y343F, Y418F and Y664F mutants sepharose at 4 1C for 30 min. Beads were washed with 50 mM were already described (Duyster et al., 2001; Cussac et al., Tris-base pH 7.4, 150 mM NaCl, 1 mM MgCl2,5mM EGTA, 2004). pcDNA3-EGFP-Rac1T17N was a generous gift from 1% triton X-100 and GTP-bound GTPases eluted with Dr K Hahn (University of North Carolina, Chapel Hill, Laemmli buffer and subjected to SDS–PAGE followed by NC, USA). pCI2.F.hVav3.WT, pCI2.F.hVav3.Y173F, pCI2.F. western blotting. hVav3.L211Q, pCDNA3.F.hVav3R697A, and pCMS3.H1P and pCMS3.H1P.shVav3 that contain a separate transcrip- tional cassette driving Green Fluorescent Protein (GFP) Cell invasion assay expression allowing easy identification of transfected cells To assess the role of Rac1 and Vav3 in migration, NIH-3T3 were already described (Zakaria et al., 2004; Charvet et al., cells were transfected with pCMS3.H1P (expressing GFP), 2005). pCMS3.H1P.shVav3 (expressing GFP and shRNA targeting Vav3) or pcDNA3.EGFP.Rac1T17N. After 30 h, cells were Cell lysis, immunoprecipitation and immunoblotting seeded in 24-well plates on biocoated Matrigel Invasion Total proteins were extracted with lysis buffer (50 mM Tris- Chambers that consisted in a 8 mm-size pore filter coated with base pH 8, 150 mM NaCl, 5 mM EGTA, 1% Nonidet P-40, a reconstituted basal membrane matrix (Becton Dickinson, À1 Mountain View, CA, USA). Migration proceeded for 17 h at 1mM PMSF, 25 mM NaF, 2 mM Na3VO4,10mgml leupeptin and 2 mgmlÀ1 aprotinin). For lysates from NPM-ALK positive 37 1C. Then, cells on the upper side of the filters were scrapped or negative lymph nodes from ALCLs patients, frozen tissues with a cotton swab and cells positive for GFP were scored. A were sonicated in detergent-free lysis buffer, cleared by correction index was applied to the raw values, it corresponds centrifugation and proteins in the supernatant quantitated to the growth of the various transfectants over 17 h. Results with the Bio-Rad protein assay (Bio-Rad, Munich, Germany). are expressed as percentages of control cells expressing GFP For immunoprecipitations, clarified homogenates were incu- only. bated overnight at 4 1C with suited antibodies and a mix of protein A/G sepharose beads. After washes, proteins were eluted with Laemmli buffer and analysed by SDS–PAGE Acknowledgements followed by western blotting on Immobilon-P membranes (Millipore, Billerica, MA, USA). Immunoreactive bands were We thank Dr Popoff for Clostridium lethal toxins. We are detected by chemiluminescence with the SuperSignal detection grateful to Dr H Tronche` re, Dr S Manenti, Dr MP Gratacap, system (Pierce Chemical Co, Rockford, IL, USA). Dr C Racaud-Sultan and Dr M Plantavid for helpful discussions. AC and DR were financed by the ‘Ministe` re de GTPases pull-down assays la Recherche et de la Technologie’ and the ‘Association pour la The amounts of GTP-bound active Rac1, Cdc42 or RhoA Recherche sur le Cancer’. This work was supported by grants were determinated by pull-down as previously described from the INSERM, ARC, ARECA, La Ligue contre le (Benard and Bokoch, 2002). Briefly, cells were lysed in ice- Cancer, the ‘Cance´ ropoˆ le Grand Sud-Ouest’ and the ‘Institut cold lysis buffer (50 mM Tris-base pH 7.4, 500 mM NaCl, National du Cancer’ (INCa), the Re´ gion Midi-Pyre´ ne´ es and 10 mM MgCl2, 2.5 mM EGTA, 1% Triton X-100, 0.5% sodium the ‘Poˆ le de Compe´ titivite´ Cancer-Bio Sante´ ’.

References

Ambrogio C, Voena C, Manazza AD, Piva R, Riera L, Barberis L Bassermann F, Jahn T, Miething C, Seipel P, Bai RY, Coutinho S et al. (2005). p130Cas mediates the transforming properties of the et al. (2002). Association of Bcr-Abl with the proto-oncogene Vav is anaplasticlymphoma kinase. Blood 106: 3907–3916. implicated in activation of the Rac-1 pathway. J Biol Chem 277: Amin HM, Medeiros LJ, Ma Y, Feretzaki M, Das P, Leventaki V 12437–12445. et al. (2003). Inhibition of JAK3 induces and decreases Benard V, Bokoch GM. (2002). Assay of Cdc42, Rac, and Rho anaplasticlymphoma kinase activityin anaplasticlarge cell GTPase activation by affinity methods. Methods Enzymol 345: 349– lymphoma. Oncogene 22: 5399–5407. 359. Armstrong F, Duplantier MM, Trempat P, Hieblot C, Lamant L, Bustelo XR. (2000). Regulatory and signaling properties of the Vav Espinos E et al. (2004). Differential effects of X-ALK fusion family. Mol Cell Biol 20: 1461–1477. proteins on proliferation, transformation, and invasion properties of Bustelo XR. (2001). Vav proteins, adaptors and cell signaling. NIH3T3 cells. Oncogene 23: 6071–6082. Oncogene 20: 6372–6381. Aznar S, Fernandez-Valeron P, Espina C, Lacal JC. (2004). Rho Charvet C, Canonigo AJ, Billadeau DD, Altman A. (2005). Membrane GTPases: potential candidates for anticancer therapy. Cancer Lett localization and function of Vav3 in T cells depend on its association 206: 181–191. with the adapter SLP-76. J Biol Chem 280: 15289–15299. Bai RY, Dieter P, Peschel C, Morris SW, Duyster J. (1998). Chiarle R, Gong JZ, Guasparri I, Pesci A, Cai J, Liu J et al. (2003). Nucleophosmin-anaplastic lymphoma kinase of large-cell anaplastic NPM-ALK transgenic mice spontaneously develop T-cell lympho- lymphoma is a constitutively active tyrosine kinase that utilizes mas and plasma cell tumors. Blood 101: 1919–1927. phospholipase C-gamma to mediate its mitogenicity. Mol Cell Biol Chiarle R, Simmons WJ, Cai H, Dhall G, Zamo A, Raz R et al. (2005). 18: 6951–6961. Stat3 is required for ALK-mediated lymphomagenesis and provides Bai RY, Ouyang T, Miething C, Morris SW, Peschel C, Duyster J. a possible therapeutictarget. Nat Med 11: 623–629. (2000). Nucleophosmin-anaplastic lymphoma kinase associated with Cho YJ, Zhang B, Kaartinen V, Haataja L, de Curtis I, Groffen J et al. anaplastic large-cell lymphoma activates the phosphatidylinositol 3- (2005). Generation of rac3 null mutant mice: role of Rac3 in Bcr/ kinase/Akt antiapoptoticsignaling pathway. Blood 96: 4319–4327. Abl-caused lymphoblastic leukemia. Mol Cell Biol 25: 5777–5785.

Oncogene Role of GTPases in NPM-ALK( þ ) ALCLs A Colomba et al 2736 Contri A, Brunati AM, Trentin L, Cabrelle A, Miorin M, Cesaro L lymphoma-cell line (COST) from a ‘small-cell variant’ of ALCL. et al. (2005). Chronic lymphocytic leukemia B cells contain Leukemia 18: 1693–1698. anomalous Lyn tyrosine kinase, a putative contribution to defective Lamant L, Meggetto F, al Saati T, Brugieres L, de Paillerets BB, apoptosis. J Clin Invest 115: 369–378. Dastugue N et al. (1996). High incidence of the t(2;5)(p23;q35) Crockett DK, Lin Z, Elenitoba-Johnson KS, Lim MS. (2004). translocation in anaplastic large cell lymphoma and its lack of Identification of NPM-ALK interacting proteins by tandem mass detection in Hodgkin’s disease. Comparison of cytogenetic analysis, spectrometry. Oncogene 23: 2617–2629. reverse transcriptase-polymerase chain reaction, and P-80 immu- CussacD, Greenland C, RocheS, Bai RY, Duyster J, Morris SW et al. nostaining. Blood 87: 284–291. (2004). Nucleophosmin-anaplastic lymphoma kinase of anaplastic Lim MS, Elenitoba-Johnson KS. (2006). Mass spectrometry-based large-cell lymphoma recruits, activates, and uses pp60c-src to proteomicstudies of human anaplasticlarge celllymphoma. Mol mediate its mitogenicity. Blood 103: 1464–1471. Cell Proteomics 5: 1787–1798. CussacD, Pichereaux C, Colomba A, Capilla F, Pont F, Gaits- Liu BP, Burridge K. (2000). Vav2 activates Rac1, Cdc42, and RhoA Iacovoni F et al. (2006). Proteomicanalysis of anaplasticlymphoma downstream from growth factor receptors but not beta1 integrins. cell lines: identification of potential tumour markers. Proteomics 6: Mol Cell Biol 20: 7160–7169. 3210–3222. Llorca O, Arias-Palomo E, Zugaza JL, Bustelo XR. (2005). Global Duyster J, Bai RY, Morris SW. (2001). Translocations involving conformational rearrangements during the activation of the GDP/ anaplasticlymphoma kinase (ALK). Oncogene 20: 5623–5637. GTP exchange factor Vav3. EMBO J 24: 1330–1340. Falini B. (2001). Anaplastic large cell lymphoma: pathological, MarzecM, KasprzyckaM, Liu X, Raghunath PN, Wlodarski P, molecular and clinical features. Br J Haematol 114: 741–760. Wasik MA. (2007). Oncogenic tyrosine kinase NPM/ALK induces Falini B, Bigerna B, Fizzotti M, Pulford K, Pileri SA, Delsol G et al. activation of the MEK/ERK signaling pathway independently of c- (1998). ALK expression defines a distinct group of T/null Raf. Oncogene 26: 813–821. lymphomas (‘ALK lymphomas’) with a wide morphological MarzecM, KasprzyckaM, Ptasznik A, Wlodarski P, Zhang Q, Odum spectrum. Am J Pathol 153: 875–886. N et al. (2005). Inhibition of ALK enzymatic activity in T-cell Fawal M, Armstrong F, Ollier S, Dupont H, Touriol C, Monsarrat B lymphoma cells induces apoptosis and suppresses proliferation and et al. (2006). A ‘liaison dangereuse’ between AUF1/hnRNPD and STAT3 phosphorylation independently of Jak3. Lab Invest 85: the oncogenic tyrosine kinase NPM-ALK. Blood 108: 2780–2788. 1544–1554. Fritz G, Kaina B. (2006). Rho GTPases: promising cellular targets for Movilla N, Bustelo XR. (1999). Biological and regulatory properties of novel anticancer drugs. Curr Cancer Drug Targets 6: 1–14. Vav-3, a new member of the Vav family of oncoproteins. Mol Cell Fujimoto J, Shiota M, Iwahara T, Seki N, Satoh H, Mori S et al. Biol 19: 7870–7885. (1996). Characterization of the transforming activity of p80, a Petit P, Breard J, Montalescot V, El Hadj NB, Levade T, Popoff M hyperphosphorylated protein in a Ki-1 lymphoma cell line with et al. (2003). Lethal toxin from Clostridium sordellii induces chromosomal translocation t(2;5). Proc Natl Acad Sci USA 93: apoptotic cell death by disruption of mitochondrial homeostasis in 4181–4186. HL-60 cells. Cell Microbiol 5: 761–771. Gao Y, Dickerson JB, Guo F, Zheng J, Zheng Y. (2004). Rational Rottner K, Hall A, Small JV. (1999). Interplay between Racand Rho design and characterization of a Rac GTPase-specific small molecule in the control of substrate contact dynamics. Curr Biol 9: 640–648. inhibitor. Proc Natl Acad Sci USA 101: 7618–7623. Sahai E, Marshall CJ. (2002). RHO-GTPases and cancer. Nat Rev Genth H, Huelsenbeck J, Hartmann B, Hofmann F, Just I, Gerhard R. Cancer 2: 133–142. (2006). Cellular stability of Rho-GTPases glucosylated by Clostri- Theodorescu D, Sapinoso LM, Conaway MR, Oxford G, Hampton dium difficile toxin B. FEBS Lett 580: 3565–3569. GM, Frierson Jr HF. (2004). Reduced expression of metastasis Hall A. (1998). Rho GTPases and the actin cytoskeleton. Science 279: suppressor RhoGDI2 is associated with decreased survival for 509–514. patients with bladder cancer. Clin Cancer Res 10: 3800–3806. Hall A. (2005). Rho GTPases and the control of cell behaviour. Thompson MA, Stumph J, Henrickson SE, Rosenwald A, Wang Q, Biochem Soc Trans 33: 891–895. Olson S et al. (2005). Differential gene expression in anaplastic Honorat JF, Ragab A, Lamant L, Delsol G, Ragab-Thomas J. (2006). lymphoma kinase-positive and anaplasticlymphoma kinase-nega- SHP1 tyrosine negatively regulates NPM-ALK tyr- tive anaplasticlarge celllymphomas. Hum Pathol 36: 494–504. osine kinase signaling. Blood 107: 4130–4138. Titus B, Schwartz MA, Theodorescu D. (2005). Rho proteins in Hornstein I, Alcover A, Katzav S. (2004). Vav proteins, masters of the cell migration and metastasis. Crit Rev Eukaryot Gene Expr 15: world of cytoskeleton organization. Cell Signal 16: 1–11. 103–114. Lamant L, De Reynies A, Duplantier MM, Rickman DS, Sabourdy F, Zakaria S, Gomez TS, Savoy DN, McAdam S, Turner M, Abraham Giuriato S et al. (2006). Gene expression profiling of systemic RT et al. (2004). Differential regulation of TCR-mediated gene anaplastic large cell lymphoma reveals differences depending on transcription by Vav family members. J Exp Med 199: 429–434. ALK status and two distinct morphological ALK+ subtypes. Blood Zeng L, Sachdev P, Yan L, Chan JL, Trenkle T, McClelland M et al. 109: 2156–2164. (2000). Vav3 mediates receptor protein tyrosine kinase signaling, Lamant L, Espinos E, Duplantier M, Dastugue N, Robert A, Allouche regulates GTPase activity, modulates cell morphology, and induces M et al. (2004). Establishment of a novel anaplasticlarge-cell cell transformation. Mol Cell Biol 20: 9212–9224.

Oncogene