LMW-PTP is a positive regulator of tumor onset and growth

LMW-PTP is a positive regulator of tumor onset and growth

Paola Chiarugi1,3, Maria Letizia Taddei1, Nicola Schiavone2, Laura Papucci2, Elisa Giannoni1,Tania Fiaschi1, Sergio Capaccioli2, Giovanni Raugei*,1,3 and Giampietro Ramponi1,3

1Department of Biochemical Sciences of the University of Florence, viale Morgagni 50, 50134 Firenze, Italy; 2Department ofExperimental Pathology and Oncology, University of Florence, Italy

Low molecular weight protein tyrosine phosphatases(LMW-PTPs) are an enzyme family that plays a keyrole in cell proliferation control by dephosphorylating/inactivating both tyrosine kinase receptors (such asPDGF, insulin, and ephrin receptors) and dockingproteins (such, as b-catenin) endowed with both adhesionand transcriptional activity. Besides being a frequent eventin human tumors, overexpression of LMW-PTP has beenrecently demonstrated to be sufficient to induce neoplastictransformation. We recently demonstrated that over-expression of LMW-PTP strongly potentiates the stabi-lity of cell–cell contacts at the adherens junction level,which powerfully suggests that LMW-PTP may alsocontribute to cancer invasivity. Focusing on mechanismsby which LMW-PTP is involved in cancer onset andprogression, the emerging picture is that LMW-PTPstrongly increases fibronectin-mediated cell adhesion andmobility but, paradoxically, decreases cell proliferation.Nevertheless, LMW-PTP-transfected NIH3T3 fibro-blasts engrafted in nude mice induce the onset of largerfibrosarcomas, which are endowed with higher prolifera-tion activity as compared to mock-transfected controls.Quite opposite effects have been obtained with engraftedfibroblasts transfected with a dominant-negative form ofLMW-PTP. Notably, in sarcoma extracts, LMW-PTPoverexpression greatly influences the ephrin A2 (EphA2)but not PDGF receptor or b-catenin tyrosine phosphor-ylation. The high association of dephosphorylated EphA2overexpression with most human cancers and our observa-tion that cell growth stimulation by LMW-PTP over-expression is restricted to the in vivo model, stronglysuggest that LMW-PTP oncogenic potential is mediatedby its EphA2 tyrosine dephosphorylating activity.Oncogene (2004) 23, 3905–3914; doi:10.1038/sj.onc.1207508Published online 15 March 2004

Keywords: LMW-PTP; tumor onset; tumor progres-sion; ephrin receptors; cell adhesion

Introduction

Protein tyrosine phosphorylation generates the powerfulsignals necessary for the growth, migration, andinvasion of normal and malignant cells (van der Geeret al., 1994). A number of tyrosine kinases have beenassociated with cancer progression (van der Geer et al.,1994; Cance and Liu, 1995), and increased tyrosinekinase activity is a reliable marker of cancer progression(Lower et al., 1993; Kopreski et al., 1996). Growingevidence indicates that the contribution of phosphotyr-osine protein phosphatases (PTPs) to the control of cellphosphorylation is as relevant as that of protein tyrosinekinases. The PTPs superfamily is composed of almost 70enzymes that, despite very limited sequence similarity,share a common CX5R active site motif and an identicalcatalytic mechanism. On the basis of their function,structure, and sequence, PTPs can be classified into fourmain families: (1) tyrosine-specific phosphatases, (2)VH1-like dual specificity PTPs, (3) cdc25, and (4) lowmolecular weight phosphatases (Zhang, 2001).Low molecular weight protein tyrosine phosphatases

(LMW-PTP) are a group of 18-kDa enzymes that arewidely expressed (Raugei et al., 2002). LMW-PTP hasbeen shown to interact with several receptor tyrosinekinases and docking proteins, including platelet-derivedgrowth factor receptor (PDGF-R) (Chiarugi et al.,1995), ephrin A2 receptor (Eph2A) (Kikawa et al.,2002), and b-catenin (Taddei et al., 2002). LMW-PTPhas been largely considered a negative regulator ofgrowth factor-induced cell proliferation, although insome instances it functions as a positive device. Forexample, in v-Ha-Ras-transformed cells, LMW-PTPincreases the cell proliferation rate (Ramponi andStefani, 1997). Recently, Park et al. presented evidencethat Xenopus laevis LMW-PTP is a functional compo-nent of the fibroblast growth factor receptor-1 complexand acts as a positive regulator of the Ras MAP kinasepathway. In addition, using a loss-of-function strategyemploying antisense morpholino oligonucleotides, theyrevealed that Xenopus laevis LMW-PTP is required forfibroblast growth factor-induced mesoderm formationin ectodermal explants and during embryogenesis (Parket al., 2002). LMW-PTP has also been indicated as apositive regulator in the Eph receptor system, in whichthe recruitment of LMW-PTP to Eph receptor com-plexes was shown to be important for promotion of

Received 30 September 2003; revised 9 January 2004; accepted 13January 2004; Published online 15 March 2004

*Correspondence: G Raugei; E-mail: [email protected] for Research, Transfer and High Education ‘Study atMolecular and Clinical Level of Chronic, Inflammatory, Degenerativeand Neoplastic Disorders for the Development on Novel Therapies’

Oncogene (2004) 23, 3905–3914& 2004 Nature Publishing Group All rights reserved 0950-9232/04 $25.00

www.nature.com/onc

endothelial capillary-like assembly and cell adhesion(Stein et al., 1998). Mutation of the LMW-PTP bindingsite in Eph B1 causes failure of endothelial cells toadhere to fibronectin, suggesting that the interaction ofLMW-PTP with Eph B1 may be necessary for celladhesion and plays a positive role in this event (Steinet al., 1998). Finally, the role of LMW-PTP in celltransformation has been investigated by Kikawa et al.,reporting that LMW-PTP is overexpressed in manyoncogene-transformed or tumor-derived mammaryepithelial cells. In addition, the overexpression ofLMW-PTP is sufficient to confer transformation innontransformed epithelial cells and the oncogenicactivity of LMW-PTP in mammary-derived cells re-quires EphA2 signaling (Kikawa et al., 2002).Our present studies concern the role of LMW-PTP

during tumor onset and progression. We demonstratethat LMW-PTP, while negatively regulating growthfactor-mediated proliferation in in vitro NIH3T3 cellcultures, is a positive regulator of tumor onset andgrowth in in vivo animal models. Despite the fact thatLMW-PTP overexpressing cells are endowed withenhanced in vitro adhesion and mobility, their engraftsdid not lead to metastasis. In addition, we demonstratethat the oncogenic function of LMW-PTP in fibrosar-coma progression is linked to EphA2 regulation.

Results

LMW-PTP controls cell proliferation and motilityin vitro

Our main plan was to evaluate the influence of LMW-PTP on tumor onset and progression. The recentliterature presents LMW-PTP as a phosphatase able togive both positive and negative signals on cell growth,obviously causing an ambiguous background about itsphysiological role. The first point that we addressed wasthe in vitro behavior of LMW-PTP overexpressing cells.Previously reported data on PDGF-induced mitogenesisindicated LMW-PTP as a negative regulator of cellproliferation, although it positively influences integrin-mediated cell motility (Chiarugi et al., 1995, 2000b;Raugei et al., 2002). In addition, very recently it hasbeen reported that LMW-PTP increases cadherin-mediated cell–cell adhesions, leading to adherens junc-tion strengthening (Taddei et al., 2002). In order tofurther examine the role of LMW-PTP on in vitro cellproliferation, motility, and apoptosis, we overexpressedwild-type (wt) LMW-PTP and its dominant-negative(dn) mutant, carrying the substitution of the catalyticCys12 to Ser, in NIH3T3 cells. First, we analysedcytoskeleton motility of cells, either by fibronectin-coated dishes adhesion assay (Figure 1a) or their abilityto move towards chemoattractants in the serum usingthe Boyden chamber motility assay for 6 h (Figure 1b),indicating that overexpression of wtLMW-PTP leads toa strong increase of cell motility. Notably, the over-expression of dnLMW-PTP causes the opposite effect.Second, we analysed the chemoinvasivity of LMW-PTP

overexpressing cells by a Boyden chamber motility assaythrough the Matrigel support for 24 h (Figure 1c), andby a wound-healing assay, in which cells are allowed tofill the gap for 24 h (Figure 1d). Again the over-expression of wtLMW-PTP leads to a strong increaseof both chemoinvasivity and the ability to fill thewound, while the dn mutant induces the oppositephenomenon. These results validate the positive role ofLMW-PTP in the regulation of in vitro cell motility inresponse to serum, even through a physical barrier ofreconstructed basal lamina or through a rapid coloniza-tion of the wound. In agreement with their ability toescape from the basal lamina, we demonstrate herein, bygelatin zymography assay, that wtLMW-PTP expressingcells show enhanced matrix metalloproteinases (MMPs)production (Figure 1e).Second, to investigate the possibility that LMW-PTP

overexpression causes autocrine effects on cell growth,we analysed the proliferation on standard untransfectedNIH3T3 cells in conditioned media (supernatants ofmock transfected, wt or dnLMW-PTP overexpressingcells). We did not find any difference in the growth rateof control cells in wt and dnLMW-PTP conditionedmedia, thus excluding the possibility of the generationby LMW-PTP overexpression of an autocrine loopaffecting cell proliferation (Figure 2a).Third, since the enhanced apoptotic threshold, could

favor tumor growth independently from cell prolifera-tion rate by increasing cell survival (Borner, 1996; Evanand Vousden, 2001), we also analysed the susceptibilityto chemical hypoxia-induced apoptosis of wt anddnLMW-PTP overexpressing cells using the respiratorychain blocker Antimycin A as apoptotic stimulus. Ourdata (Figure 2b) show that the dnLMW-PTP expressingcells have a lower apoptotic threshold with respect tomock-transfected cells using 200 mM antimycin. Con-versely, we observed only a slight decrease of theapoptotic susceptibility of wtLMW-PTP compared tomock-transfected cells.The picture that emerges from all data both from

literature and reported here, summarized in Table 1, isthat LMW-PTP negatively affects cell proliferation inresponse to serum and positively influences the adhesiveand motility properties of cells, indicating that LMW-PTP may positively affect the attack and colonization ofconnective tissues.

LMW-PTP acts as a positive regulator of nude micesarcoma onset

In the light of these results, we have analysed the role ofLMW-PTP in in vivo tumor onset and progression.Because in vitro analyses of cell growth do not alwayspredict the in vivo tumorigenic potential, we planned toimplant wt and dnLMW-PTP overexpressing cells inathymic mice. Cells (1� 106) were subcutaneouslyinoculated into 30 nude mice (six per group) and causedthe presence of palpable tumors within 4 weeks(Figure 3a). Surprisingly, our results show thatwtLMW-PTP positively influences the formation of aprimary sarcoma in mice, while dnLMW-PTP produces

LMW-PTP positively regulates tumor growthP Chiarugi et al

3906

Oncogene

the opposite effect. In fact, mice inoculated withdnLMW-PTP expressing cells develop primary sarco-mas with a great delay compared to mock-transfected

cells. In addition to the rate of tumor onset, LMW-PTPinfluences the dimension of the tumor, as the medianvolume of resulting tumors was not related to the

Figure 1 LMW-PTP overexpression increases motility and invasiveness in culture. Panel (a) Cell adhesion on fibronectin: 3� 104 cellswere seeded for the indicated time in a 96-well dish precoated with human fibronectin. Cell adhesion was evaluated as reported in theMaterials and Methods section. Panel (b) Cell motility assay: Migration of NIH-3T3 cells was assayed with the Transwell system ofCostar, equipped with 8-mm filters precoated with human type I collagen. The lower chamber was filled with medium supplementedwith 20% FCS. Cell motility was evaluated as reported in the Materials and Methods section. Panel (c) In vitro invasivity assay: Theinvasivity of cells was assessed with the Transwell system of Costar, equipped with 8-mm filters precoated with Matrigel. The lowerchamber was filled with medium supplemented with 20% FCS. The invasivity potential of cells was evaluated as reported in theMaterials and Methods section. Panel (d)Wound healing assay: Cells were allowed to adhere to dishes in complete medium, scored witha micropipette tip, and photographed either immediately or after 24 h to visualize incoming cells. Images were acquired with a digitalcamera under a phase-contrast microscope. Panel (e) MMP expression: Cells of the indicated type were allowed to grow for 48 h incomplete medium. The medium was collected and analysed for MMP activity by gelatin zymography. The result is a representative gelzymography; band intensities (proportional to the MMP activity on gelatin as substrate) were quantified from three independentexperiments with densitometric analysis, and the mean7s.d. is shown. The results are presented relative to mock-transfected cells. Theresults are normalized to the activity of MMPs of mock-transfected cells taken as a unit. Two clones of each transfected cell line (cl.Aand cl.B) were tested in all experiments. Mock-transfected cells were used as controls. The results are means of at least fourindependent experiments, each in triplicate


3907

Oncogene

number of implanted cells (Figure 3b). To confirm thisindication, NIH3T3 cell growth in vivo was assessed byquantifying the expression of proliferating cell nuclear

antigen (PCNA), a marker for actively dividing cells,including sarcoma cells (Albini et al., 2001; Liekenset al., 2001). PCNA expression was evaluated byWestern blot analysis using lysates from primary tumorsfrom three randomly selected mice from each group(Figure 3c). Normalization was achieved by antiactinimmunoblot of the same lysates (Figure 3d). PCNAexpression is perfectly correlated with the sarcomagrowth rate and volume of cells expressing LMW-PTP, reinforcing the idea of a positive function ofLMW-PTP in tumor onset.To assess the in vivo invasiveness potential of LMW-

PTP expressing cells, we checked for lung metastases innude mice explants. The necropsy did not reveal anymetastatic region in any of the analysed animals(n¼ 30). This finding was further confirmed by injectionof either wt or dnLMW-PTP expressing cells (n¼ 15) inthe tail vein of athymic mice. Again, we did not find anylung metastasis in animal explants.These findings suggest a positive role of LMW-PTP in

both tumor implantation and proliferation rate, butappear to exclude its involvement in the invasivecapacity and metastasis formation.

EphA2 is the tyrosine phosphorylated substrate ofLMW-PTP in developing sarcomas

To clarify the role of LMW-PTP as a positive regulatorof tumor onset and growth, we analysed the tyrosinephosphorylation levels of several LMW-PTP substratesin developing sarcomas. LMW-PTP has been shown tointeract with and dephosphorylate the PDGF-R, b-catenin, and EphA2 receptor (Chiarugi et al., 1995;Kikawa et al., 2002; Taddei et al., 2002). First, weevaluated the tyrosine phosphorylation level of PDGF-R in primary tumors from wt and dnLMW-PTPexpressing cells in the developing sarcomas (Figure 4a).Surprisingly, our findings reveal that in the developingsarcomas the PDGF receptor was not a preferentialLMW-PTP substrate. Neither wtLMW-PTP nordnLMW-PTP overexpression affected the phosphotyr-osine content of the PDGF receptor. In addition, thelevel of expression of PDGF receptor in sarcomas wasvery low in comparison to that of in vitro NIH3T3 cellsbefore injection in mice (data not shown).As an alternative possibility, we examined whether

LMW-PTP affected the phosphotyrosine levels of thecytoplasmic pool of b-catenin in the same tumor

Figure 2 Autocrine capacity and hypoxia-induced apoptosis in wtand dnLMW-PTP expressing cells. Two clones of each transfected cellline (cl.A and cl.B) were tested. Mock-transfected cells were used ascontrols. Panel (a): Cell growth in conditioned media from mocktransfected, wt, and dnLMW-PTP overexpressing cells is evaluated asreported in the Materials and Methods section. Panel (b): Apoptoticevents were detected and registered progressively by time-lapse videomicroscopy during 72h after application of the respiratory chainblocker Antimycin A (200mm). An apoptotic event was scored themoment the cell detached from the substrate, shrank, and blebbed.Each point is the mean7SE of three experiments

Table 1 General features of LMW-PTP overexpression effects on proliferation, motility and apoptosis

Effect WtLMW-PTP Dn LMW-PTP

Cell proliferation (Raugei et al., 2002) Decreased IncreasedIntegrin-mediated adhesion (this study and Chiarugi et al., 2000a) Increased DecreasedChemotaxis (this study and Chiarugi et al., 2000a) Increased DecreasedCadherin-mediated adhesion (Taddei et al., 2002) Increased DecreasedAutocrine loop (this study) None NoneHypoxia-induced apoptosis (this study) Slightly decreased IncreasedChemoinvasivity (this study) Increased DecreasedMMP secretion (this study) Increased Decreased

n.b.: Increases and decreases are with respect to mock-transfected cells


3908

Oncogene

samples (Figure 4b). Tyrosine phosphorylation of b-catenin leads to the disruption of contact with E-cadherin, leading to an increase in the free, uncomplexedcytoplasmic pool of the protein (Muller et al., 1999).Unlike the cadherin-associated b-catenin, the cytosolicb-catenin is known to be the sole pool that is tyrosinephosphorylated in the cell. Recent studies demonstratethat tyrosine phosphorylation of b-catenin is correlatedwith tumor formation and invasiveness (Sommers et al.,1994; Bonvini et al., 2001). We have already reportedthat in in vitro NIH3T3 cells the level of cytoplasmic b-catenin is negatively influenced by LMW-PTP expres-sion (Taddei et al., 2002). Unexpectedly, in developingsarcomas (although deriving from the same in vitro cellline) neither the tyrosine phosphorylation level nor thecontent of cytoplasmic b-catenin is influenced by LMW-PTP overexpression (Figure 4b).The possibility that EphA2 receptor could serve as a

substrate for LMW-PTP in fibrosarcoma cells has beeninvestigated. The tyrosine phosphorylation level ofEphA2 receptor was evaluated in developing sarcomas(Figure 4c). The results show that EphA2 tyrosinephosphorylation level was dramatically decreased by theoverexpression of wtLMW-PTP, and enhanced byexpression of dnLMW-PTP, as indicated by bothstraight Western blot and immunoprecipitation ana-lyses, suggesting that EphA2 receptor may be asubstrate of LMW-PTP in vivo. It should be noted thatin in vitro cell cultures the level of EphA2 is almostundetectable (data not shown), while is greatly enhancedin in vivo tumor explants. Finally, it has been reportedthat the function of EphA2 in epithelial cells is regulatedby E-cadherin expression (Zantek et al., 1999) and thatthe increases of E-cadherin in NIH3T3 induce theexpression of EphA2 (Orsulic and Kemler 2000). In thislight, we have analysed the level of E-cadherin inNIH3T3 cells (either expressing wt or dnLMW-PTP)and in tumor extracts. Although we did not finddetectable levels of E-cadherin in in vitro NIH3T3cultures, we found high E-cadherin expression levels(Figure 4d) in tumor extracts. In particular, we observeda good correlation between E-cadherin and EphA2expression level in wtLMW-PTP-derived tumors, show-ing increased expression of both proteins, and indnLMW-PTP-derived tumors, showing the oppositephenomenon.Altogether these data suggest that LMW-PTP may

have in vivo and in vitro differential substrates, EphA2being its preferential target in developing sarcomas andPDGF receptor and/or b-catenin in NIH3T3 in in vitrocultures.

Discussion

A popular paradigm suggests that a balance betweentyrosine kinase and phosphatase activities determinesthe cellular levels of protein tyrosine phosphorylationand thereby governs cellular decisions regarding growth,survival, and invasiveness. This model predicts that

Figure 3 Wt and dnLMW-PTP NIH3T3 sarcomas in nude mice.Tumorigenicity assays were performed by subcutaneous injectionwith 1� 106 NIH3T3 cells either expressing wt or dnLMW-PTP.Injected mice were observed every 3 days. Subcutaneous tumorswere grossly visible at the site of injection after 2–3 weeks. Animalswere examined twice weekly and killed 7 weeks later. Each cell linewas tested for tumorigenicity in six different mice. Panel (a): Thenumber of developed sarcomas in each analysed group is reportedand the mean7s.d. is shown. Panel (b): The volume of the tumor ineach analysed group is reported and the mean7s.d. is shown. Panel(c): PCNA expression: three randomly selected sarcomas from eachgroup were used for the anti-PCNA immunoblot analysis ofactively dividing cells. A representative immunoblot is shown.Equalization was checked by anti-a-actin immunoblot analysis ofthe same lysates from sarcomas (panel (d))


3909

Oncogene

tyrosine kinases are oncogenic, whereas tyrosine phos-phatases negatively regulate neoplastic transformation.Although this paradigm has generally been sustained bythe recognition of oncogenic tyrosine kinases, emergingevidence reveals a more complex interplay betweentyrosine kinases and phosphatases. For example,CAAX-PTP has been recently implicated as an onco-gene (Cates et al., 1996). Moreover, the enzymaticactivity of Src family kinases is released by RPTPa-mediated dephosphorylation of critical tyrosine 527(Cobb and Parsons, 1993; Liu and Pawson, 1994).There is a great abundance of biochemical findings

concerning the in vitro LMW-PTP role, but its in vivofunction is less understood. By means of wt anddnLMW-PTP overexpression, we and others suggestboth negative and positive roles for LMW-PTP. Inparticular, in in vitro cell cultures LMW-PTP negativelyinfluences PDGF-R (Chiarugi et al., 1995), fibroblastgrowth factor receptor (Rigacci et al., 1999), epidermalgrowth factor receptor (Ramponi et al., 1989) andinsulin receptor signaling (Chiarugi et al., 1997), andpositively affects EphA2 (Kikawa et al., 2002) and

EphB1 signaling (Stein et al., 1998). Similar to LMW-PTP, SHP-2 has been proposed for both a positive andnegative role in signal transduction (Ronnstrand andHeldin, 2001). For example, mutation of PDGF-RTyr1009, namely the SHP2 binding site, impairs thedownregulation of PDGF-R (Lechleider et al., 1993),suggesting a negative role for this phosphatase. How-ever, SHP-2 has also been shown to act as a positivecomponent of PDGF signaling, where mutations in theSHP-2 binding sites inhibited chemotaxis and reducedPDGF-induced Ras and MAP kinase activation (Ronn-strand et al., 1999). To note, an in vitro study has shownthat LMW-PTP binds to and dephosphorylates aphosphopeptide containing Tyr1009 derived fromPDGFR-b (Bucciantini et al., 1998), suggesting thatSHP-2 and LMW-PTP may share a common substrate.In addition, both SHP-2 and LMW-PTP have beenshown to play a positive role in FGF-mediatedmesoderm induction during Xenopus laevis development(O’Reilly et al., 2000; Park et al., 2002).To investigate the in vivo role of LMW-PTP, we

analysed the effect of wt and dnLMW-PTP overexpres-

Figure 4 Tyrosine phosphorylation level of PDGF-R, b-catenin, and EphA2 in nude mice sarcomas. Three randomly selectedsarcomas from each group were used for the analysis of the tyrosine phosphorylation level of PDGF-R, b-catenin, and EphA2. Panel(a): PDGF-R was immunoprecipitated from lysates and the antiphosphotyrosine immunoblot is shown. The membrane was reprobedwith anti-PDGF-R antibodies for normalization. Panel (b): b-Catenin was immunoprecipitated from lysates and the antipho-sphotyrosine immunoblot is shown. The membrane was reprobed with b-catenin antibodies for normalization. Panel (c): EphA2 wasimmunoprecipitated from lysates and the antiphosphotyrosine immunoblot is shown. EphA2 was either quantified by reprobing thesame filter with anti-EphA2 antibodies or by straight Western blot of total lysates. Panel (d): E-cadherin immunoblot of tumor extractsin representative Western blots of at least three independent experiments are shown


3910

Oncogene

sion on tumor onset and progression. The present dataled to two major conclusions: (i) LMW-PTP in vivo is apositive regulator of both tumor onset and development;(ii) the oncogenic potential of the phosphatase is linkedto EphA2 and not to PDGF-R or b-catenin depho-sphorylation. Our results strongly suggest that the sameenzyme, namely a phosphotyrosine phosphatase, mayhave different in vivo and in vitro performances/substrates. In particular, in vitro overexpression ofwtLMW-PTP in NIH3T3 cells causes a strong reductionof proliferation in response to serum growth factors,together with an enhancement of both the adhesiveproperties of cells, that is integrin- and cadherin-mediated cell adhesion, and cell motility, that ischemotaxis towards growth factors and the ability ofwounds to regenerate. Conversely, engrafted LMW-PTPtransfected/overexpressing NIH3T3 cells lead to theonset of larger and more numerous sarcomas comparedto mock-transfected cells. Nevertheless, NIH3T3 over-expression does not affect the ability of primary tumorsto form lung metastases.Cancer arises when a population of cells gain the

ability to inappropriately grow and survive. Thesebiological behaviors result from genetic and/or environ-mental abnormalities triggering specific signals andpromoting cell proliferation and survival (Hunter,2000). Growing evidence from experimental and clinicalstudies points to the fundamental, pathophysiologic roleof several factors in the growth of solid tumors, that is,(i) the self-production of growth factors which maysustain tumor growth, (ii) susceptibility to hypoxia-mediated apoptosis which structurally and functionallydisturbs microcirculation, causing deterioration of thediffusion geometry and tumor-associated anemia, (iii)alterations in the adhesive properties of tumor cells,often correlated to progression to tumor malignancy(Comoglio and Trusolino, 2002). Our data suggest thatthe overexpression of LMW-PTP does not affect theautocrine capacity of NIH3T3 cells, thus excluding thatthe oncogenic potential of the phosphatase may belinked to growth factor overproduction. Conversely, it isimplied that the effect of LMW-PTP on hypoxia-mediated apoptosis may, to some extent, correlate withthe positive effect of the phosphatase in tumor onset andprogression. In regard to the third point, the phenotypiceffect of LMW-PTP on cell adhesive properties appearsto be quite remarkable, that is, LMW-PTP positivelyaffects both cell–cell and/or cell–matrix adhesion. It isoften reported that tumor cells show a decrease in cell–cell and/or cell–matrix adhesion, although it is notknown whether these alterations are causally involved intumor progression or if they are only the epiphenomenaof tumor cell phenotypic changes (Cavallaro andChristofori, 2001). In addition, the involvement ofcadherin-mediated adherens junctions is well documen-ted for the spreading of tumor cells from the primarysite to distant organs and the dissemination ofmetastases (Beavon, 2000). Conversely, besides tumormalignancy, the role of cell adhesion in the regulation ofprimary tumor onset and in the proliferation rate ofcancer cells has not been strongly addressed. In this

light, it is possible that the enhancement of the adhesiveproperties of cells leads to a facility in contacting theprimary site of growing tumors. In this view, LMW-PTPmay improve tumor implantation and growth, while notaffecting the in vivo invasiveness properties of LMW-PTP overexpressing cells. In agreement with thisspeculation, wtLMW-PTP overexpressing cells are ableto induce fibrosarcoma onset and growth but not itsmetastasis, despite the in vitro invasivity and MMPproduction induced by LMW-PTP overexpression.Indeed, metastasis is known to be the result of anintricate and coordinated interplay between differentfactors including cell adhesion, proteolysis, migration,angiogenesis, responsiveness to local growth, andsurvival factors (for a recent review see Comoglioet al., 2002). Since a multitude of conditions must beexhaustively satisfied for metastasis onset, it is notsurprising that LMW-PTP overexpression, althoughaffecting cell motility and adhesion, is not able per seto complete this multistep process leading to metastaticspread.Our data demonstrate that LMW-PTP may interact

and dephosphorylate different substrates in vivo or invitro. We previously reported that in in vitro NIH3T3cells PDGF-R and b-catenin are efficiently depho-sphorylated by the phosphatase, thus correlating withdecreased PDGF-mediated proliferation rate and in-creased cadherin/b-catenin cell adhesion, respectively(Chiarugi et al., 1997; Taddei et al., 2002). On the otherhand, in fibrosarcomas neither PDGF-R nor b-cateninare dephosphorylated by LMW-PTP which is converselyactive on EphA2 receptor. The EphA2 receptor isoverexpressed in a large number of cancers, includingbreast, prostate, and lung carcinomas together withmelanomas and sarcomas (Easty et al., 1995; Zelinskiet al., 2001; Varelias et al., 2002). In addition to itsoverexpression, EphA2 is functionally altered in trans-formed cells by regulating its tyrosine phosphorylationlevel. In particular, EphA2 is prominently tyrosinephosphorylated in nontransformed cells, while it ismainly dephosphorylated in transformed cells (Zanteket al., 1999, 2001; Zelinski et al., 2001). LMW-PTP hasalready been reported to interact with EphA2 inepithelial cells cultured in vitro. In these cells, theoverexpression of LMW-PTP is sufficient to confertransformation, the oncogenic activities of the phospha-tase being dependent on altered EphA2 tyrosinephosphorylation (Kikawa et al., 2002). In this context,our data underline that, in vivo, EphA2 dephosphoryla-tion represents the main feature of LMW-PTP activity,likely attributing to PDGF-R and b-catenin depho-sphorylation a role in different circumstances. We thinkthat this behavior is mainly due to differences between invivo and in vitro models and that the effect of a givenprotein, in this case a phosphatase, may vary dependingupon the physiological circumstances and differentsubstrata availability.We stress that the EphA2 and E-cadherin expression

levels in wtLMW-PTP and dnLMW-PTP developingsarcomas (Figure 4c and d) are correlated with theirtumor onset and growth, in agreement with previously


3911

Oncogene

reported data showing that E-cadherin is directlyrequired for EphA2 function (Orsulic and Kemler,2000). These observations further strengthen our ideaof wtLMW-PTP as a positive regulator of cell transfor-mation.In addition to the rate of tumor onset, LMW-PTP

influences the dimension of the tumor, as the medianvolume of resulting tumors is not related to the numberof implanted cells. These findings indicate that in vivoLMW-PTP positively influences both the ability oftransformed cells to adhere to the tumor primary siteand their proliferation rate, as indicated by PCNAanalysis. This is in agreement with the dephosphoryla-tion of EphA2 by LMW-PTP in developing sarcomas.In fact, in contrast to other RTKs, activation andtyrosine phosphorylation of Eph receptors fails topromote cell proliferation (Brambilla et al., 1995; Bruceet al., 1999), but initiates signaling pathways that exertantimitogenic functions. Hence, the LMW-PTP-mediated EphA2 dephosphorylation leads to attenua-tion of antiproliferative activity, in agreement withrecent studies showing that unphosphorylated EphA2functions as a powerful oncoprotein (Zantek et al., 1999;Zelinski et al., 2001; Carles-Kinch et al., 2002).Accordingly, Kikawa et al. (2002) reported thatLMW-PTP-overexpressing MCF10A cells acquire atransformed phenotype when cultured in three-dimen-sional basement membrane, such as Matrigel, and gainthe ability to colonize soft agar but show a reduced rateof monolayer cell growth. This is in agreement with thedecreased growth rate we found in LMW-PTP over-expressing NIH3T3 fibroblasts (Chiarugi et al., 1995;Raugei et al., 2002). Consistent with this, we underlinethat transformation of epithelial cells often leads to adecreased monolayer cell growth rate and that the mostaggressive variants of them in vivo, show the slowestgrowth in monolayer cultures (Zantek et al., 2001;Zelinski et al., 2001). These outcomes may have seriousimplications when using proliferation assays in non-transformed cell lines as predictive of transformed cellbehavior.In summary, our current studies recognize LMW-PTP

as a positive regulator of tumor growth in vivo, linkingthe biochemical and biological actions of LMW-PTP toEphA2 dephosphorylation. We stressed that thesefindings have important significance as they definitelydislodge LMW-PTP from the negative PTP family. It isnoteworthy that only two in vivo investigations out ofmany in vitro studies concerning LMW-PTP negativefunction, that is, this study and observations on Xenopuslaevis development (Park et al., 2002), are in agreementthat LMW-PTP is a positive regulator.

Materials and methods

Cell lines and reagents

NIH3T3 were from ATCC. Monoclonal antibodies specific forproliferating cell nuclear antigen (PCNA) and b-catenin werepurchased from Transduction Laboratories. Monoclonalantibodies specific for phosphotyrosine (4G10) and EphA2

(clone D7) were purchased from Upstate Biotechnology Inc.Polyclonal antibodies against LMW-PTP were prepared asdescribed elsewhere (Chiarugi et al., 1995).

Cell culture and transfections

NIH3T3 cells were routinely cultured in Dulbecco’s modifiedEagle’s medium supplemented with 10% fetal calf serum in a5% CO2 humidified atmosphere. A measure of 10 mg ofpRcCMV-wtLMW-PTP or pRcCMV-C12S-LMW-PTP0 ex-pressing the dn Cys-12 to Ser mutant of LMW-PTP (Chiarugiet al., 1995) were transfected in NIH3T3 cells using the calciumphosphate method. Stable transfected clonal cell lines wereisolated by selection with G418 (400mg/ml). For in vitro and invivo studies, we selected two clones of NIH3T3 cells thatequally express wtLMW-PTP (A and B) or dnLMW-PTP(A and B).

Immunoprecipitation and Western blot analysis

NIH3T3 cells (1� 106) were seeded in 10-cm plates inDulbecco’s modified Eagle’s medium supplemented with10% fetal calf serum. Cells were serum-starved for 24 h beforereceiving 20% FCS. Cells were then lysed for 20min on ice in500 ml of RIPA lysis buffer (50mM Tris-HCl, pH 7.5, 150mMNaCl, 0.1% SDS, 0.5% sodium deoxycholate, 1% Nonidet P-40, 2mM EGTA, 1mM sodium orthovanadate, 1mM phenyl-methanesulfonyl fluoride, 10 mg/ml aprotinin, and 10 mg/mlleupeptin). For the solid tumors, 100mg of each sample washomogenized in lysis buffer with a Dounce homogenizer.Lysates were clarified by centrifugation and immunoprecipi-tated for 4 h at 41C with 1–2 mg of the specific antibodies.Immunocomplexes were collected on protein A-Sepharose,separated by SDS–PAGE, and transferred onto nitrocellulose.Immunoblots were incubated in 3% bovine serum albumin,10mM Tris-HCl, pH 7.5, 1mM EDTA, and 0.1% Tween 20 for1 h at room temperature, probed with specific antibodies andthen with secondary antibodies conjugated with horseradishperoxidase, washed, and developed with the EnhancedChemiluminescence kit.

Cell adhesion assay

Cell adhesion was assessed as described elsewhere (Chiarugiet al., 2000b). Briefly, 3� 104 cells were serum starved for 24 hand then seeded for the indicated time in a 96-well dishprecoated for 2 h with 10mg/ml of human fibronectin andwashed twice with PBS, in medium containing 20% FCS. Celladhesion was stopped by removing the medium and by theaddition of a 0.5% crystal violet solution in 20% methanol.After 5min of staining, the fixed cells were washed with PBSand solubilized with 200 ml/well of 0.1M sodium citrate, pH4.2. The absorbance at 595 nm was evaluated using amicroplate reader. All cell adhesion assays were performed intriplicate.

Cell motility and chemoinvasivity assays

Cell motility (Chiarugi et al., 2000b) and chemoinvasivity(Yagel et al., 1989) were assessed as described elsewhere.Briefly, migration of NIH-3T3 cells was assayed with theTranswell system of Costar, equipped with 8-mm porepolyvinylpyrrolidone-free polycarbonate filters (diameter,13mm) precoated with human type I collagen (20 mg/ml) andplaced between the chemoattractant (lower chamber) and theupper chamber. The lower chamber was filled with mediumsupplemented with 20% FCS. Serum-free medium culturedcells were suspended by trypsinization, and 1.5� 105 cells in


3912

Oncogene

200ml were added to the top wells and incubated at 371C in 5%CO2 for 6 h. After incubation, the cells attached to the upperside but not migrated through the filter were mechanicallyremoved using cotton swabs. For the in vitro invasion assaychambers were prepared at 1 : 2 dilution of Matrigel (BectonDickinson). Cells (5� 104) were loaded on top of the Matrigellayer for 24 h at 371C. The filters were fixed in 96% methanoland stained with Diff Quick staining solutions. Chemotaxisor chemoinvasivity were evaluated by counting the cells thathad migrated to the lower surface of the polycarbonate filters.The number of cells in six randomly chosen fields wasdetermined for each filter, and the counts were averaged(mean7s.d.).

Autocrine cell growth assay

NIH-3T3 cells (2� 104) were seeded in 24-multiwell plates andserum starved for 24 h before receiving conditioned mediafrom 48 h subcultured mock transfected, wt and dn over-expressing cells. Fresh conditioned medium was added dailyfor 3 days. Cellular growth was stopped by removing themedium and by the addition of a 0.5% crystal violet solutionin 20% methanol. After 5min of staining the fixed cells werewashed with PBS and solubilized with 200 ml/well of 0.1Msodium citrate, pH 4.2. The absorbance at 595 nm wasevaluated using a microplate reader.

Quantification of cumulative apoptotic events

The respiratory chain blocker Antimycin A was used at aconcentration of 200 mM, previously reported to commit cellsto apoptosis (Formigli et al., 2000). The cumulative apoptoticevents were scored for 72 h by the time-lapse videomicroscopyusing a Zeiss inverted phase-contrast microscope equippedwith a � 10 objective, Panasonic CCD cameras, and JVCBR9030 time-lapse video recorders (Papucci et al., 2002). Aftercell detachment from the substrate, apoptotic deaths werecounted the moment the cell became fully shrunken andblebbing started.

Zymography

Metalloprotease zymography was assessed as described else-where, with minor modifications (Chiarugi et al., 2000b).Briefly, culture medium from confluent cell monolayers wascollected, and 20 ml was added to Laemmli sample buffer andrun on an 8% SDS gel containing 0.1% gelatin. Afterelectrophoresis the gel was washed twice with 2.5% TritonX-100 and once with reaction buffer (50mM Tris-HCl, pH 7.5,200mM NaCl, 5mM CaCl2. The gel was incubated overnightat 371C with freshly added reaction buffer and stained withLaemmli Coomassie blue solution.

Wound-healing assay

Cells were seeded in 10% FCS in culture dishes coated with25 mg/ml fibronectin. After 4 h of incubation, dishes werescored with a sterile 200-ml micropipette tip and photographed.After 24 h, the wounds were photographed again to visualizeincoming cells.

Tumorigenicity assay

Tumorigenicity was assayed by injection of 105 cells into nudemice (NIH nu/nu, female, and 6 weeks of age) subcutaneouslyover the lateral thorax. Animals were examined twice weeklyand killed 7 weeks later. Lung metastases and the neoplasticnature of local tumors were confirmed by histologic examina-tion.

AcknowledgementsThis work was supported by the Italian Association for CancerResearch (AIRC), by the Ministero Italiano della Universita eRicerca (MIUR-PRIN 2001 to SC and 2002 to GR), byMinistero della Salute (to SC), in part by ConsorzioInteruniversitario Biotecnologie (to GR), in part by Cassa diRisparmio di Firenze (to PC and SC), and in part by ConsiglioNazionale delle Ricerche (Progetto Finalizzato Oncologia toSC). The authors are very grateful to Professor Mary Forrestfor accurate revision of the paper.

References

Albini A, Morini M, D’Agostini F, Ferrari N, Campelli F,Arena G, Noonan DM, Pesce C and De Flora S. (2001).Cancer Res., 61, 8171–8178.

Beavon IR. (2000). Eur. J. Cancer, 36, 1607–1620.Bonvini P, An WG, Rosolen A, Nguyen P, Trepel J, GarciadH, Dunach M and Neckers LM. (2001). Cancer Res., 61,1671–1677.

Borner C. (1996. J. Biol. Chem., 271, 12695–12698.Brambilla R, Schnapp A, Casagranda F, Labrador JP,Bergemann AD, Flanagan JG, Pasquale EB and Klein R.(1995). EMBO J., 14, 3116–3126.

Bruce V, Olivieri G, Eickelberg O and Miescher GC. (1999).Brain Res., 821, 169–176.

Bucciantini M, Stefani M, Taddei N, Chiti F, Rigacci S andRamponi G. (1998). FEBS Lett., 422, 213–217.

Cance WG and Liu ET. (1995). Breast Cancer Res. Treat., 35,105–114.

Carles-Kinch K, Kilpatrick KE, Stewart JC and Kinch MS.(2002). Cancer Res., 62, 2840–2847.

Cates CA, Michael RL, Stayrook KR, Harvey KA, Burke YD,Randall SK, Crowell PL and Crowell DN. (1996). CancerLett., 110, 49–55.

Cavallaro U and Christofori G. (2001). Biochim. Biophys.Acta, 1552, 39–45.

Chiarugi P, Cirri P, Marra F, Raugei G, Camici G, Manao Gand Ramponi G. (1997). Biochem. Biophys. Res. Commun.,238, 676–682.

Chiarugi P, Cirri P, Raugei G, Camici G, Dolfi F, Berti A andRamponi G. (1995). FEBS Lett., 372, 49–53.

Chiarugi P, Cirri P, Taddei L, Giannoni E, Camici G, ManaoG, Raugei G and Ramponi G. (2000a). J. Biol. Chem., 275,4640–4646.

Chiarugi P, Taddei ML, Cirri P, Talini D, Buricchi F, CamiciG, Manao G, Raugei G and Ramponi G. (2000b). J. Biol.Chem., 275, 37619–37627.

Cobb BS and Parsons JT. (1993). Oncogene, 8, 2897–2903.Comoglio PM and Trusolino L. (2002). J. Clin. Invest, 109,857–862.

Easty DJ, Guthrie BA, Maung K, Farr CJ, Lindberg RA,Toso RJ, Herlyn M and Bennett DC. (1995). Cancer Res.,55, 2528–2532.

Evan GI and Vousden KH. (2001). Nature, 411, 342–348.Formigli L, Papucci L, Tani A, Schiavone N, Tempestini A,Orlandini GE, Capaccioli S and Orlandini SZ. (2000). J. CellPhysiol., 182, 41–49.

Hunter T. (2000). Cell, 100, 113–127.Kikawa KD, Vidale DR, Van Etten RL and Kinch MS.(2002). J. Biol. Chem., 277, 39274–39279.


3913

Oncogene

Kopreski MS, Witters L, Brennan Jr WA, Buckwalter EA,Chinchilli VM, Demers LM and Lipton A. (1996). Anti-cancer Res., 16, 3037–3041.

Lechleider RJ, Sugimoto S, Bennett AM, Kashishian AS,Cooper JA, Shoelson SE, Walsh CT and Neel. (1993). J.Biol. Chem., 268, 21478–21481.

Liekens S, Verbeken E, De Clercq E and Neyts J. (2001). Int. J.Cancer, 92, 161–167.

Liu X and Pawson T. (1994).Recent Prog. Horm. Res., 49, 149–160.Lower EE, Franco RS, Miller MA and Martelo OJ. (1993).

Breast Cancer Res. Treat., 26, 217–224.Muller T, Choidas A, Reichmann E and Ullrich A. (1999). J.

Biol. Chem., 274, 10173–10183.O’Reilly AM, Pluskey S, Shoelson SE and Neel BG. (2000).

Mol. Cell Biol., 20, 299–311.Orsulic S and Kemler R. (2000). J. Cell Sci., 113, 1793–1802.Papucci L, Schiavone N, Donnini M, Lapucci A, Luzi E,Tempestini A, Witort E and Capaccioli S. (2002). AntisenseNucleic Acid Drug Dev., 12, 21–31.

Park EK, Warner N, Mood K, Pawson T and Daar IO. (2002).Mol. Cell Biol., 22, 3404–3414.

Ramponi G and Stefani M. (1997). Biochim. Biophys. Acta,1341, 137–156.

Ramponi G, Manao G, Camici G, Cappugi G, Ruggiero Mand Bottaro DP. (1989). FEBS Lett., 250, 469–473.

Raugei G, Ramponi G and Chiarugi P. (2002). Cell Mol. LifeSci., 59, 941–949.

Rigacci S, Rovida E, Bagnoli S, Dello SP and Berti A. (1999).FEBS Lett., 459, 191–194.

Ronnstrand L and Heldin CH. (2001). Int. J. Cancer, 91,

757–762.Ronnstrand L, Arvidsson AK, Kallin A, Rorsman C, HellmanU, Engstrom U, Wernstedt C and Heldin CH. (1999).Oncogene, 18, 3696–3702.

Sommers CL, Gelmann EP, Kemler R, Cowin P and ByersSW. (1994). Cancer Res., 54, 3544–3552.

Stein E, Lane AA, Cerretti DP, Schoecklmann HO, SchroffAD, Van Etten RL and Daniel TO. (1998). Genes Dev., 12,667–678.

Taddei ML, Chiarugi P, Cirri P, Buricchi F, Fiaschi T,Giannoni E, Talini D, Cozzi G, Formigli L, Raugei andRamponi G. (2002). Cancer Res., 62, 6489–6499.

van der Geer P, Hunter T and Lindberg RA. (1994). Annu.Rev. Cell Biol., 10, 251–337.

Varelias A, Koblar SA, Cowled PA, Carter CD and Clayer M.(2002). Cancer, 95, 862–869.

Yagel S, Warner AH, Nellans HN, Lala PK, Waghorne C andDenhardt DT. (1989). Cancer Res., 49, 3553–3557.

Zantek ND, Azimi M, Fedor-Chaiken M, Wang B, Bracken-bury R and Kinch MS. (1999). Cell Growth Differ., 10,

629–638.Zantek ND, Walker-Daniels J, Stewart J, Hansen RK,Robinson D, Miao H, Wang B, Kung HJ, Bissell MJand Kinch MS. (2001). Clin. Cancer Res., 7,

3640–3648.Zelinski DP, Zantek ND, Stewart JC, Irizarry AR and KinchMS. (2001). Cancer Res., 61, 2301–2306.

Zhang ZY. (2001). Curr. Opin. Chem. Biol., 5, 416–423.


3914

Oncogene

LMW-PTP is a positive regulator of tumor onset and growth

Documents