Rac1 activity regulates proliferation of aggressive metastatic melanoma

E X P E R I M E N T A L C E L L R E S E A R C H 3 1 3 ( 2 0 0 7 ) 3 8 3 2 – 3 8 3 9

ava i l ab l e a t www.sc i enced i r ec t . com

www.e l sev i e r. com/ loca te /yexc r

Research Article

Rac1 activity regulates proliferation of aggressivemetastatic melanoma

Natalie N. Bauera,c,⁎, Yih-Wen Chena, Rajeev S. Samanta,Lalita A. Shevdea, Oystein Fodstada,b

aMitchell Cancer Institute, University of South Alabama, Mobile, AL 36688, USAbInstitute for Cancer Research, Rikshospitalet-Radiumhospitalet Medical Center, Faculty of Medicine, University of Oslo, NorwaycDepartment of Pharmacology, Center for Lung Biology, University of South Alabama, Mobile, AL 36688, USA

A R T I C L E I N F O R M A T I O N

⁎ Corresponding author. Department of PharmBlvd. Mobile, AL 36688, USA. Fax: +1 251 460

E-mail address: [email protected] (N.

0014-4827/$ – see front matter © 2007 Elsevidoi:10.1016/j.yexcr.2007.08.017

A B S T R A C T

Article Chronology:Received 19 March 2007Revised version received22 August 2007Accepted 23 August 2007Available online 28 August 2007

Molecular mechanisms underlying the different capacity of two in vivo selected humanmelanoma cell variants to form experimental metastases were studied. The doubling timesof the FEMX-I and FEMX-V cell sublines in vitro were 15 and 25 h, respectively. The invasivecapacity of FEMX-I cells was 8-fold higher than FEMX-V cells, and the time to formapproximately 10 mm s.c. tumors in nude mice was 21 versus 35 days. FEMX-I displayed aspindle-like formation in vitro, whereas FEMX-V cells had a rounded shape. Hence, weexamined known determinants of cell shape and proliferation, the small GTPases. The fourstudied showed equal expression in both cell types, but Rac1 activity was significantlydecreased in FEMX-V cells. Rac1 stimulates NFκB, and we found that endogenous NFκBactivity of FEMX-V cells was 2% of that of FEMX-I cells. Inhibition of Rac1 resulted in blockedNFκB activity. Specific inhibition of either Rac1 or NFκB significantly reduced proliferationand invasion of FEMX-I cells, the more pronounced effects observed with Rac1 inhibition.These data indicate that Rac1 activity in FEMX cells regulates cell proliferation and invasion,in part via its effect onNFκB, signifying Rac1 as a keymolecule inmelanoma progression andmetastasis.

© 2007 Elsevier Inc. All rights reserved.

Keywords:Rac1MelanomaProliferationNFκBSmall GTpaseMetastasis

Introduction

Melanoma is commonly diagnosed in patients at the prime oftheir lives, unlike many solid tumors which occur up to twodecades later [1]. Whereasmost primarymelanomas are curedby surgery alone, metastatic melanomas are resistant tosystemic therapy and are most often fatal. Clearly, increasedunderstanding of mechanisms determining melanoma pro-gression and metastasis is needed, and clinically relevantmodels of melanoma metastasis provide a means to study

acology, Center for Lung7452.N. Bauer).

er Inc. All rights reserved

aspects of the progression of this disease. We have previouslyreported on the metastatic phenotype of FEMX-I melanomacells in nude mice and rats [2]. The FEMX-I cells wereestablished from excision of s.c. and lymph node metastasesfollowing i.v. injection of FEMX cells into nude mice. Freshlyisolated cells were re-injected i.v. into other nude mice, andthis was repeatedly done for each new generation of metas-tases, and in parallel used for in vitro culture. The experimentswere initially designed to select FEMX cells with decreasedlatency for metastases to develop, however a decreased

Biology, University of South Alabama, MSB 3340, 307 N. University

.

3833E X P E R I M E N T A L C E L L R E S E A R C H 3 1 3 ( 2 0 0 7 ) 3 8 3 2 – 3 8 3 9

metastatic potential was revealed with each successivegeneration of metastases. The experimental cycles wererepeated to passage five, FEMX-V, which exhibited no metas-tases within 120 days [2,3]. These unique differences inphenotypic properties of two variant cell lines of the sameorigin provide a good model for examining mechanismsunderlying differences in characteristics of metastatic andnon-metastatic variants of one and the same cell line.

We also observed that the difference in phenotype includedcell shape. Data suggest that members of the Rho family ofGTPases, Rac1 and RhoA, which are known to control cellshapemay regulate transcription factor nuclear factor kappa B(NFκB) activity in order to stimulate cellular proliferation [4–8]and activation of Rac1, specifically, induces NFκB DNA bindingandactivity [9]. It has also recently been shown that expressionof NFκB is up-regulated in human metastatic melanomabiopsies as compared to normal melanocytes and benignintradermal nevi [10,11]. Based on the findings in melanomaand preliminary in vitro data on phenotypic differencesbetween the twocell variants,we sought to determinewhetherthere was a shift in NFκB activities between FEMX-I and theFEMX-V generations of cells and whether this was associatedwith upstream activation of Rac1. We found significantdifferences between the FEMX cell variants in both Rac1 andNFκB activities directly related to the differences in prolifera-tion and invasion. Collectively, the data indicate an importantrole for Rac1 signaling in regulating NFκB activity and itsfunction in melanoma progression and metastasis.

Materials and methods

Cell culture, trypan blue exclusion and WST assay

Both FEMX-I and FEMX-V were cultured in RPMI (MediatechInc. Herndon, VA) supplemented with 10%FBS (AtlantaBiologicals, Lawrenceville, GA) at 37 °C in a 5% CO2 humidifiedincubator. Cells were used between passages 3 and 15 andtested routinely for mycoplasma contamination. For trypanblue proliferation assays cells were seeded at the same densityon day 0. Each subsequent day one plate of FEMX-I and ofFEMX-V was dispersed with Cellstripper (Mediatech Inc.)aliquoted with equal volume of trypan blue (Sigma-Aldrich,St. Louis, MO) and counted for a total of 4 days. For WST-1proliferation assays we followed manufacturer's protocols(Roche Applied Science, Indianapolis, IN). For cells treatedwith either Rac inhibitor (20 μM; Calbiochem, San Diego, CA) orNFκB activity inhibitor (5 μM; Calbiochem) the treatment wasapplied on day 1 two hours prior to WST-1 reagent. WST-1reagent was incubated for 4 h and read using a DynexTechnologies Plate Reader at 450 nm.

Invasion assays

In vitro invasion assayswere performed using Falcon transwellinvasion chamber with Matrigel-coated-8 μm-pore PET mem-brane invasion chamber inserts (BD Biosciences, San Jose, CA).The invasion chambers were rehydrated with serum-freeRPMI 1640 medium per manufacturer's directions. The lowerchambers, providing a chemotactic gradient, were filled with

1 ml RPMI 1640 medium with 10% FBS and 1 μg/ml offibronectin. Following 24–48 h of incubation, the cells on theupper side of the membrane were gently removed with acotton swab. The invasive cells on the lower surface of themembrane were stained 0.2% crystal violet in 30% ethanol for1 h before washing. The number of cells in 8–12 randomlyselected microscopic fields (20×) was counted under anOlympus CKX31 microscope (Olympus America Corp., CenterValley, PA).

F-actin staining

Cells of each typewere seeded on chambered slides and grownovernight. Following aspiration of media, cells were fixed with4% paraformaldehyde, washed with PBS and permeabilizedwith 0.1% Triton X-100 (Sigma-Aldrich). Slides were thenwashed rapidly with 0.1% Bovine Serum Albumin in PBS andstained with Alexafluor 488 phalloidin (Invitrogen, Carlsbad,CA) for 20 min at room temp. Chambers were then removed,slides coverslipped using Vectashieldmountingmedia (VectorLaboratories, Burlingame, CA) and stored at 4 °C until viewingin a Nikon confocal microscope with Ultraview RS-3 confocalimaging system.

In vivo tumor formation studies

Cells (passage 7) at 80–90% confluence were detached with2 mM EDTA solution, washed with cold CMF-DPBS, countedusing a hemacytometer, resuspended in ice-cold HBSS to afinal concentration of 1×107 cells/ml and 100 μl (1×106 cells)was injected intradermally into the scapular region of femaleathymic nude mice (Harlan, Indianapolis, IN). Tumor size wasmeasuredweekly andmean tumor diameterwas calculated bytaking the square root of the product of orthogonal measure-ments. When the mean tumor diameter reached 1.0–1.2 cm,the mice were killed by cervical dislocation under anesthesia.Animalsweremaintained under the guidelines of theNationalInstitute of Health and the University of South Alabama. Allprotocols were approved by Institutional Animal Care and UseCommittee. Food and water were provided ad libitum [12].

Transfections and luciferase activity

FEMX-I and FEMX-V cells were seeded at the same density on96-well tissue culture plates and incubated overnight. Theywere washed with PBS and transfected with either pNFκB-Lucor pTAL-Luc as a negative control (Clontech, Palo Alto, CA)using the Lipofectamine 2000 transfection reagent (Invitro-gen) per manufacturer's instructions. Following 24 h incuba-tion the transfection media was removed, cells washed withPBS and media replaced with complete media. Following afurther 24 h incubation cells were lysed and analyzed forluciferase activity using the Promega luciferase assay systemper manufacturer's protocols and read on a TD-20/20 lumin-ometer (Turner Biosystems, Sunnyvale, CA).

Dual transfections of FEMX-I cells were performed asdescribed above with the addition of Rac1 (T17N substitution;N17Rac; Upstate Biotechnology, Lake Placid, NY). Cells wereallowed the same washes and incubation times as above, fol-lowedbyassessmentwith thePromega luciferaseassay system.

3834 E X P E R I M E N T A L C E L L R E S E A R C H 3 1 3 ( 2 0 0 7 ) 3 8 3 2 – 3 8 3 9

Analysis of p65 translocation

FEMX-I and FEMX-V cells were grown to 80% confluence on100mmdishes. Cytoplasmic and nuclear fractions were isolatedwith the Pierce NE-PER nuclear and cytoplasmic extractionreagents per the included protocols (Pierce Biotechnology, Rock-ford, IL). Protein determinations of extracts were done with RC/DC Protein assay (Biorad, Hercules, CA). Equal concentrations ofprotein were run on SDS-PAGE and western analysis performedwith a p65 specific antibody (Cell Signaling Technology, Boston,MA).

siRNA transfections

siRNA against either Rac1 or a Scrambled control (Santa CruzBiotechnology, Santa Cruz, CA) were transfected into FEMX-Icells using Lipofectamine RNAiMax (Invitrogen, Carlsbad, CA)per the manufacturer's reverse transfection protocol. Theywere allowed to transfect from 24–72 h dependent on theassay following transfection. Knockdown of Rac1 was con-firmed using western analysis for Rac1 as described below.

Endogenous small GTPase expression and activity

Each cell type was seeded and grown to 80% confluence on100 mm dishes. Total cell lysates were collected in modifiedRIPA buffer (1% IGEPAL CA-630, 0.5% Sodium deoxycholate,0.1% SDS, and 1% Triton X-100; Sigma-Aldrich) and protein

Fig. 1 – In vitro and in vivo proliferation of FEMX-I and FEMX-V cellsFEMX-I cells grew at a two-fold higher capacity than FEMX-V cells.time for FEMX-I cells was 15 h versus 24 h for FEMX-V (n=4, *P<0.0demonstrates that FEMX-I formed approximately 10 mm tumors wgroup, *P<0.05). (D) FEMX-V cells invade matrigel coated transwells

analyzed as above. Equal protein loading was assured bywestern analysis for β-actin. Immunoblots were performedwith specific antibodies for either RhoA, Rac1, or CDC42(Upstate Biotechnology) or RhoC (Santa Cruz Biotechnology,Santa Cruz, CA). For activity assays cells were lysed with lysisbuffer supplied in the RhoA or Rac1 activity assay kit (UpstateBiotechnology). Protein was determined by Bradford assay(Biorad). Equal protein concentrations were used for westernanalysis of total expression (30 μg) as well as for testing theactivity (1 mg) following the manufacturer's protocols.Densitometry of the immunoblots was performed with thedigital analysis software AlphaEase®FC image analysissoftware.

Statistical analysis

All experiments were repeated 3 times unless otherwiseindicated. Results are expressed as mean±S.E.M. All statisticswere performedwith GraphPad Prism4 software and Student'st test or one-way ANOVA applied where necessary. A value ofPb0.05 was considered statistically significant.

Results

Initially, we compared the growth rates of FEMX-I and FEMX-Vcells in vitro and in vivo. As determined by trypanblue exclusionthe calculated doubling time was 15 h for FEMX-I cells

. (A) Trypan blue exclusion proliferation assays determined that(B) WST proliferation assays showed that the mean doubling5). (C) Mean tumor diameter was routinely monitored andithin 21 days compared to 35 days for FEMX-V cells (n=8 perat only 8% the capacity of FEMX-I cells (n=3–4, *P<0.05).

3835E X P E R I M E N T A L C E L L R E S E A R C H 3 1 3 ( 2 0 0 7 ) 3 8 3 2 – 3 8 3 9

compared to 24 h for FEMX-V (Fig. 1A, n=4, ⁎Pb0.05). A WSTproliferation assay confirmed these data (Fig. 1B), demonstrat-ing a significant difference in proliferation rates between thetwo cell variants in vitro (n=4, ⁎P=0.0035). The tumorigenicityof FEMX-I and FEMX-V cells in nude mice injected s.c. with1×106 of either cell type showed that the FEMX-I cellsgenerated approximately 10 mm tumors within 21 dayscompared to 35 days for FEMX-V (Fig. 1C, n=8 per group,⁎Pb0.05). Both cell types formed a palpable tumor within7 days, however the FEMX-I tumors grew more rapidly thanFEMX-V cells resulting in a steeper slope of the growth curve.Morphology of excised tumors revealed that both cell typesexhibited prominent nucleoli, which is expected inmelanoma,multiple mitotic figures and abundant cytoplasm (Fig. 2B).Tumors generated from each cell type also stained positive forthe melanoma marker HMB45 (data not shown). As a furtherindex of difference in the cell variants aggressiveness weexamined invasive capacity. FEMX-I cells invaded at an 8-fold

Fig. 2 – In vitro and in vivo morphology. (A) FEMX-I cells, under stand generate significant fillapodia and lamellapodia structures amore rounded structure with a tendency to grow in clusters withshows that both tumors have abundant cytoplasm and promine

higher capacity than the less aggressive FEMX-V cells (Fig. 1D,n=3–4, ⁎Pb0.05).

The more aggressive FEMX-I cells appeared to have adistinctly different phenotype in vitro compared to FEMX-Vcells. Staining of the f-actin cytoskeleton of both cell typesshowed that FEMX-I cells displayed amore spindle-like shape,formation of invadapodia, and seemed to grow well indepen-dently, whereas the FEMX-V cells tended to grow in roundedformations of three ormore cells with significant cortical actinrim staining (Fig. 2A), consistent with their decreased prolif-erative capacity.

Since cell shape is dictated by the cytoskeleton these datasuggested that a difference between the two cell variants mayinvolve cytoskeletal regulating proteins [13]. The Rho GTPasesare a family of proteins involved in a multitude of cellularfunctions, but most notably cytoskeletal regulation for move-ment and proliferation. Based on the phenotypic analysis ofFEMX-I and FEMX-V cellswe examined expression ofmembers

andard in vitro conditions, display a spindle-like morphologys illustrated by staining of f-actin. FEMX-V cells exhibitstrong cortical rim f-actin staining. (B) Tumor morphology

nt nucleoli, as is characteristic of melanoma.

Fig. 3 – SmallGTPaseexpressionandactivation inFEMX-IandFEMX-Vcells. (A)Totalproteinexpressionof thesmallGTPasesRhoA,RhoC, Rac1, and CDC42 was equivalent in FEMX-I and FEMX-V cells (I=FEMX-I, V=FEMX-V). (B) Immunoblot analysis of activeRhoA was the same in the two cell variants, whereas, FEMX-V cells have less endogenous active Rac1 than FEMX-I cells. GTPstimulation of lysates followed by western analysis of active Rac1 showed the FEMX-V protein was functional. (C) Densitometryconfirmed total expression among the small GTPases was equivalent in FEMX-I and FEMX-V cells and that GTP-stimulated Rac1activity was the same in both cell types, whereas endogenous activity was less in FEMX-V vs. FEMX-I cells (A.U.=arbitrary units).

3836 E X P E R I M E N T A L C E L L R E S E A R C H 3 1 3 ( 2 0 0 7 ) 3 8 3 2 – 3 8 3 9

of theRhoGTPase family [8,14,15]. Immunoblots for expressionof RhoA, RhoC, Rac1, and CDC42, after loading identicalconcentrations of protein, revealed that each was expressedat the same level in FEMX-I and FEMX-V cells (Fig. 3A).Therefore, protein expression of small GTPases did not appearto be responsible for the difference in cell phenotype andproliferation between the two cell variants.

Although there was no difference in RhoA and Rac1expression levels, we questioned whether the endogenousactivity of these small GTPases was the same in FEMX-I andFEMX-V cells [16,17]. Using specific pull-down assays for activeRhoAandRac1,we found thatwhereas therewasno differencein the level of active RhoA in FEMX-I and FEMX-V cells, Rac1activity differed significantly and was almost absent in FEMX-V cells (Fig. 3B). To determine whether Rac1 was functional inFEMX-V cells we performed the pull-down assay after loadingeach of the variant cell lysates with a non-hydrolyzable GTP

analog and, and found that Rac1 could be activated to reach anequal level in both cell types (Fig. 3B). These data demonstratethat FEMX-V cells express functional Rac1; however it is notendogenously active.

Since Rac1 has been shown to function upstreamof NFκB inregulation of proliferation [6,8] and several recent studies havesuggested that NFκB is constitutively expressed and activatedin malignant melanoma [18–21] we examined whether NFκBactivity differed between FEMX-I and FEMX-V cells. Both celltypes were transfected with NFκB-Luc and assayed forendogenous NFκB-activated luciferase activity. As illustratedin Fig. 4A, FEMX-V displayed only 2% of the constitutive NFκBactivity of FEMX-I cells (n=4, ⁎Pb005), suggesting that NFkBmay be implicated in regulating the difference in proliferationand invasion between the two melanoma cell variants.

In order for NFκB to be active the p65 subunit of the tran-scription factor must be translocated from the cytosol to the

3837E X P E R I M E N T A L C E L L R E S E A R C H 3 1 3 ( 2 0 0 7 ) 3 8 3 2 – 3 8 3 9

nucleus [19]. We therefore isolated cytoplasmic and nuclearextracts from FEMX-I and FEMX-V cells, and western analysisof p65 showed a significantly higher nuclear translocation inFEMX-I as compared to FEMX-V cells (Fig. 4A, inset). Thissupports a role for increased NFκB activity in melanomaproliferation and invasion, and that this is downregulated inFEMX-V cells. To determine Rac1's influence on NFκB activity,FEMX-I cells were co-transfectedwith dominant negative Rac1and NFkB-luc and assayed for NFκB-regulated luciferaseactivity. Inhibition of Rac1 with the dominant negative-Rac1decreased NFκB activity by 67% as compared to controls,

indicating that Rac1 plays a regulatory role in NFκB activity ofFEMX cell variants (Fig. 4A; n=3, ⁎Pb0.05). To more specificallydisrupt Rac1/NFκB signaling FEMX-I cells were transfectedwith siRNA for Rac1. Control cells were treated with transfec-tion reagent alone or scrambled siRNA. We found that Rac1siRNA decreased expression of Rac1 in FEMX-I cells andinhibited NFκB luciferase activity by 30%, whereas neithertransfection reagent nor scrambled siRNA prevented activa-tion of NFκB-luciferase (data not shown). Combined these datasuggest significant control over NFκB activity by Rac1 inmelanoma.

Since both NFκB and Rac1 activity were decreased in FEMX-Vcells and inhibitionof Rac1blockedNFκBactivation,we sought todetermine whether inhibition of NFκB or Rac1 would decreasethe proliferation and invasion of the more aggressive FEMX-Icells. FEMX-I cells were seeded for a proliferation assay andtreated once with an NFκB activity inhibitor, Rac1 inhibitor, ortransfected with Rac1 siRNA. We found that inhibition of NFκBincreased the FEMX-I cell doubling time to 22 h, close to thedoubling time of FEMX-V cells (Fig. 4B; n=3, ⁎Pb0.05; 5 μM). Incomparison, inhibition of Rac1 by either method increased thedoubling time of FEMX-I cells 3-fold (Fig. 4B; n=4, ⁎Pb0.05; 20 μM).We also examined the effects of inhibiting Rac1 or NFκB oninvasive capacity. Inhibiting the activity of NFκB blockedinvasionofFEMX-I cellsby59%,whereasblockingRac1activationinhibited the FEMX-I cell invasion by nearly 75% (Fig. 4C, n=3–4,⁎Pb0.05) and preventing expression of Rac1 with siRNA signifi-cantly inhibited the invasive capacity of FEMX-I. These datashow that Rac1 regulates NFκB activity and suggests that thedifference in cell proliferation and invasiveness between themelanoma cell variants is related to downregulation of Rac1stimulation in the less aggressive FEMX-V melanoma cells.

Discussion

Metastatic melanoma is resistant to therapy and most oftenfatal. In spite of the wealth of new data the outcome ofmetastatic disease is still dismal and further elucidation ofmechanisms of metastasis and factors mediating the

Fig. 4 – Regulation of NFκB activity by Rac1 in FEMX-I andFEMX-V cells. (A) FEMX-I and FEMX-V were transfected withNFκB-luc and FEMX-V cells exhibited only 2% of theendogenous NFκB activity of FEMX-V cells (n=4, *P<0.05).Inset: western analysis of nuclear extracts shows that p65 ismore abundant in the nucleus of FEMX-I cells than FEMX-Vcells (n=3). FEMX-I cells dually transfected with dominantnegative Rac1 (N17Rac; DN-Rac) and NFκB-Luc had only 37%of endogenous NFκB activity and siRNA against Rac1inhibited NFκB by 30% compared to FEMX-I cells (n=4,*P<0.05). (B) Treatment of FEMX-I cells with Rac inhibitor(20 μM, single dose) or siRNA for Rac1 caused a 3-foldincrease in doubling time (n=3, *P<0.05), whereas FEMX-Icells treated with NFκB activity inhibitor (5 μM, single dose)showed a 1.5 fold increased doubling time (n=4, *P<0.05).(C) Inhibition of Rac1 (20 μM, single dose) with inhibitor orsiRNA or NFkB (5 μM) prevented FEMX-I invasion throughmatrigel coated transwells (n=3–4, *P<0.05).

3838 E X P E R I M E N T A L C E L L R E S E A R C H 3 1 3 ( 2 0 0 7 ) 3 8 3 2 – 3 8 3 9

resistance are clearly needed. We have developed severalhumanmelanomametastasismodels (2) and have shown thatwhereas FEMX-I melanoma cells aggressively metastasized innudemice the variant FEMX-V cells did not [2]. This differencein metastatic capacity provided us with a setting to analyzeunderlying molecular mechanisms [2,3,22]. Our studies showthat the decreased proliferation and invasive capacity ofFEMX-V cells, ultimately leading to metastasis, is controlledvia reduced Rac1 activity, in part via regulation of NFκB.

Oneof our first observationswas thedifferencebetween thecell variants in in vitro proliferative capacity. Thus, theproliferation rate of cells from the first generation of metas-tases, FEMX-I, was twice that of the fifth generation isolates,FEMX-V. These datawere indicative of a potential difference intumorigenic abilities and the differences in metastatic aggres-siveness which had been previously observed (3) were alsoreflected in the significantly decreased invasiveness of FEMX-V cells. Phenotypically, the FEMX-I cells were spindle-like andexhibited invadapodia typical of an aggressive and invasivecell, whereas the FEMX-V cells presentedwith amore roundedappearance and generally proliferated in clusters of three ormore cells. Hence, the metastatic aggressiveness of FEMX cellvariants correlated with both their phenotypes and theirinvasive capacity in the in vitro functional assays.

Our observations of the distinctiveness of the FEMXphenotypes led us to examine determinants of cell shapewhich are also implicated in induction of metastasis, tumor-igenicity, and overall cell proliferation, the small GTPases[23,24]. The small GTPases RhoA and RhoC have been stronglyimplicated in invasion, metastasis, and proliferation of breastcancer, and CDC42 and Rac1 in melanoma [14,25]. Our datarevealed that whereas each of these small GTPases wasexpressed at equal levels in FEMX-I and FEMX-V cells, Rac1had significantly decreased activity in the less aggressive of thetwo cell lines. Rac1 expression and signaling is a requirementof cellular proliferation in normal adult cells and completedisruption of Rac1 signaling during development is lethal[26,27]. Rac1 was shown to be fully functional but not endo-genously active in FEMX-V cells. We interpret these data tomean that although Rac1 expression is the same, thedecreased signaling in FEMX-V cells directly relates to theirdecreased rate of proliferation and less invasive capacity ascompared to the more aggressive FEMX-I cells. We surmisedfrom these data that Rac1 activity might regulate the distinctphenotypic expression as well as the proliferation andinvasion of the cell variants.

Rac1 has several downstream targets, but has recently beenshown to induce NFκB DNA binding activity, and studies com-paring the protein expression of NFκB in normal melanocytes,benign intradermal nevus and metastatic melanoma biopsiesshowed increased NFkB expression in nevi and melanoma ascompared to normal melanocytes [4,5,9,11]. These data alsosuggested an increased activity of the overexpressed proteinsin metastatic melanoma as compared to the intradermal nevi[11]. We found that endogenous NFκB activity was decreasedsignificantly in the less aggressive FEMX-V cells and thatinhibitionofNFκBdecreased theproliferationof FEMX-I cells tobecome nearly identical to that of FEMX-V. Inhibition of NFκBalso impeded invasion of FEMX-I cells. These data show thatNFκB plays a significant regulatory role in proliferation and

invasion of aggressive FEMX-Imelanoma cells whichmaywelltranslate to increased capacity to metastasize. Further experi-ments revealed that inhibition of Rac1 blockedNFκB activity aswell as proliferation and invasion. Notably the effects of Rac1inhibition on invasion and proliferation were even morepronounced than that observed with the NFkB inhibitor,suggesting Rac1 signaling plays a greater role in metastaticdisease than its regulation of NFκB alone. Rac1, as a cytoskel-etal regulator may have been expected to influence invasivecapacity significantly, but its role in proliferative capacityoutside of NFκB signaling will warrant significant furtherstudy. Although previous investigations have implicatedeither Rac1 and/or NFκB in tumor progression, proliferation,or invasion, our results are the first, to our knowledge, thatdemonstrate the link between the two in human melanoma.Moreover, the data suggest that Rac1 signaling has consider-able influence over proliferation and invasion as compared toNFκB. Further studies will be needed to address the mechan-isms responsible for the putative NFκB independent activity ofRac1. However, the present data provide new insight in animportant aspect of melanoma biology and presents Rac1 as aputative key molecule in melanoma metastasis.

Acknowledgments

Wewould like to thankMs. GinaCapley, Dr. Justin Campbell, andMr. Grant Shell for their technical assistance.Wewould also liketo thank Dr. Judy King for her pathology expertise.

R E F E R E N C E S

[1] S.J. O'Day, C.J. Kim, D.S. Reintgen, Metastatic melanoma:chemotherapy to biochemotherapy, Cancer Control 9 (2002)31–38.

[2] O. Fodstad, I. Kjonniksen, Microenvironment revisited: timefor reappraisal of some prevailing concepts of cancermetastasis, J. Cell. Biochem. 56 (1994) 23–28.

[3] O. Fodstad, I. Kjonniksen, S. Aamdal, J.M. Nesland, M.R. Boyd,A. Pihl, Extrapulmonary, tissue-specific metastasis formationin nude mice injected with FEMX-I human melanoma cells,Cancer Res. 48 (1988) 4382–4388.

[4] R. Perona, S. Montaner, L. Saniger, I. Sanchez-Perez, R. Bravo,J.C. Lacal, Activation of the nuclear factor-kappaB by Rho,CDC42, and Rac-1 proteins, Genes Dev. 11 (1997) 463–475.

[5] S. Montaner, R. Perona, L. Saniger, J.C. Lacal, Multiplesignalling pathways lead to the activation of the nuclearfactor kappaB by the Rho family of GTPases, J. Biol. Chem. 273(1998) 12779–12785.

[6] P. Matos, P. Jordan, Expression of Rac1b stimulatesNF-kappaB-mediated cell survival and G1/S progression,Exp. Cell Res. 305 (2005) 292–299.

[7] S.A. Benitah, P.F. Valeron, J.C. Lacal, ROCK and nuclearfactor-kappaB-dependent activation of cyclooxygenase-2 byRho GTPases: effects on tumor growth and therapeuticconsequences, Mol. Biol. Cell 14 (2003) 3041–3054.

[8] K.A. Kwei, J.S. Finch, J. Ranger-Moore, G.T. Bowden, The role ofRac1 in maintaining malignant phenotype of mouse skintumor cells, Cancer Lett. 231 (2006) 326–338.

[9] P.J. Reddig, D. Xu, R.L. Juliano, Regulation of p21-activatedkinase-independent Rac1 signal transduction by nischarin,J. Biol. Chem. 280 (2005) 30994–31002.

3839E X P E R I M E N T A L C E L L R E S E A R C H 3 1 3 ( 2 0 0 7 ) 3 8 3 2 – 3 8 3 9

[10] K.I. Amiri, A. Richmond, Role of nuclear factor-kappa B inmelanoma, Cancer Metastasis Rev. 24 (2005) 301–313.

[11] S.E. McNulty, R. del Rosario, D. Cen, F.L. Meyskens Jr., S. Yang,Comparative expression of NFkappaB proteins inmelanocytes of normal skin vs. benign intradermal naevusand human metastatic melanoma biopsies, Pigment Cell Res.17 (2004) 173–180.

[12] L.A. Shevde, R.S. Samant, S.F. Goldberg, T. Sikaneta, A.Alessandrini, H.J. Donahue, D.T. Mauger, D.R. Welch,Suppression of human melanoma metastasis by themetastasis suppressor gene, BRMS1, Exp. Cell Res. 273 (2002)229–239.

[13] R.A. Bartolome, B.G. Galvez, N. Longo, F. Baleux, G.N. VanMuijen, P. Sanchez-Mateos, A.G. Arroyo, J. Teixido, Stromalcell-derived factor-1alpha promotes melanoma cell invasionacross basement membranes involving stimulation ofmembrane-type 1 matrix metalloproteinase and Rho GTPaseactivities, Cancer Res. 64 (2004) 2534–2543.

[14] J.Y. Pille, C. Denoyelle, J. Varet, J.R. Bertrand, J. Soria,P. Opolon, H. Lu, L.L. Pritchard, J.P. Vannier, C. Malvy, C.Soria, H. Li, Anti-RhoA and anti-RhoC siRNAs inhibit theproliferation and invasiveness of MDA-MB-231 breastcancer cells in vitro and in vivo, Molec. Ther. 11 (2005)267–274.

[15] C.F. Welsh, Rho GTPases as key transducers of proliferativesignals in g1 cell cycle regulation, Breast Cancer Res. Treat. 84(2004) 33–42.

[16] N.L. Tran, W.S. McDonough, B.A. Savitch, S.P. Fortin, J.A.Winkles, M. Symons, M. Nakada, H.E. Cunliffe, G. Hostetter,D.B. Hoelzinger, J.L. Rennert, J.S. Michaelson, L.C. Burkly, C.A.Lipinski, J.C. Loftus, L. Mariani, M.E. Berens, Increasedfibroblast growth factor-inducible 14 expression levelspromote glioma cell invasion via Rac1 and nuclearfactor-{kappa}B and correlate with poor patient outcome,Cancer Res. 66 (2006) 9535–9542.

[17] B. Wojciak-Stothard, A.J. Ridley, Shear stress-inducedendothelial cell polarization is mediated by Rho and Rac butnot Cdc42 or PI 3-kinases, J. Cell Biol. 161 (2003) 429–439.

[18] Y. Ueda, A. Richmond, NF-kappaB activation in melanoma,Pigment Cell Res. 19 (2006) 112–124.

[19] P. Viatour, M.P. Merville, V. Bours, A. Chariot, Phosphorylationof NF-kappaB and IkappaB proteins: implications in cancerand inflammation, Trends Biochem. Sci. 30 (2005) 43–52.

[20] J. Yang, A. Richmond, Constitutive IkappaB kinase activitycorrelates with nuclear factor-kappaB activation in humanmelanoma cells, Cancer Res. 61 (2001) 4901–4909.

[21] A.J. Miller, M.C. Mihm Jr., Melanoma, N. Engl. J. Med. 355 (2006)51–65.

[22] O. Fodstad, Tumorigenicity and dissemination of humantumors in congenitally immune-deficient mice, J. Natl.Cancer Inst. 83 (1991) 1419–1420.

[23] L. del Peso, R. Hernandez-Alcoceba, N. Embade, A. Carnero,P. Esteve, C. Paje, J.C. Lacal, Rho proteins induce metastaticproperties in vivo, Oncogene 15 (1997) 3047–3057.

[24] S. Pervaiz, J. Cao, O.S. Chao, Y.Y. Chin, M.V. Clement,Activation of the RacGTPase inhibits apoptosis in humantumor cells, Oncogene 20 (2001) 6263–6268.

[25] I.D. Jung, J. Lee, S.Y. Yun, C.G. Park, W.S. Choi, H.W. Lee, O.H.Choi, J.W. Han, H.Y. Lee, Cdc42 and Rac1 are necessary forautotaxin-induced tumor cell motility in A2058 melanomacells, FEBS Lett. 532 (2002) 351–356.

[26] S. Patil, M. Bunderson, J. Wilham, S.M. Black, Important rolefor Rac1 in regulating reactive oxygen species generation andpulmonary arterial smooth muscle cell growth, Am. J.Physiol., Lung Cell. Mol. Physiol. 287 (2004) L1314–L1322.

[27] L. Wei, K. Imanaka-Yoshida, L. Wang, S. Zhan, M.D.Schneider, F.J. DeMayo, R.J. Schwartz, Inhibition of Rho familyGTPases by Rho GDP dissociation inhibitor disrupts cardiacmorphogenesis and inhibits cardiomyocyte proliferation,Development 129 (2002) 1705–1714.