Virology 485 (2015) 481–491

Contents lists available at ScienceDirect

Virology

http://d0042-68

n CorrAlexandGerman

E-mchristinmbros@Brigitte.aakress@

journal homepage: www.elsevier.com/locate/yviro

Regulation of the tumor marker Fascin by the viral oncoprotein Taxof human T-cell leukemia virus type 1 (HTLV-1) depends on promoteractivation and on a promoter-independent mechanism

Caroline F. Mohr a, Christine Gross a, Matthias Bros b, Angelika B. Reske-Kunz b,Brigitte Biesinger a, Andrea K. Thoma-Kress a,n

a Institute of Clinical and Molecular Virology, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germanyb Department of Dermatology, University Medical Center, Johannes Gutenberg-University, Mainz, Germany

a r t i c l e i n f o

Article history:Received 10 June 2015Returned to author for revisions24 June 2015Accepted 24 August 2015Available online 9 September 2015

Keywords:Tumor virusHTLV-1OncogeneTaxATLFascinTranscription regulationPromoterNF‐kappa B (NF‐KB)PP2

x.doi.org/10.1016/j.virol.2015.08.02522/& 2015 Elsevier Inc. All rights reserved.

espondence to: Institute of Clinical and Mer-Universität Erlangen-Nürnberg, Schlossy. Tel.: þ49 9131 8526429; fax: þ49 9131 85ail addresses: caroline.mohr@viro.med.uni-erle.gross@viro.med.uni-erlangen.de (C. Gross),uni-mainz.de (M. Bros), A.Reske-Kunz@uni-mBiesinger-Zwosta@viro.med.uni-erlangen.de (viro.med.uni-erlangen.de (A.K. Thoma-Kress

a b s t r a c t

Adult T-cell leukemia/lymphoma is a highly infiltrative neoplasia of CD4þ T-lymphocytes that occurs inabout 5% of carriers infected with the deltaretrovirus human T-cell leukemia virus type 1 (HTLV-1). Theviral oncoprotein Tax perturbs cellular signaling pathways leading to upregulation of host cell factors,amongst them the actin-bundling protein Fascin, an invasion marker of several types of cancer. However,transcriptional regulation of Fascin by Tax is poorly understood. In this study, we identified a triple modeof transcriptional induction of Fascin by Tax, which requires (1) NF-κB-dependent promoter activation,(2) a Tax-responsive region in the Fascin promoter, and (3) a promoter-independent mechanism sensitiveto the Src family kinase inhibitor PP2. Thus, Tax regulates Fascin by a multitude of signals. Beyond, usingTax-expressing and virus-transformed lymphocytes as a model system, our study is the first to identifythe invasion marker Fascin as a novel target of PP2, an inhibitor of metastasis.

& 2015 Elsevier Inc. All rights reserved.

Introduction

Adult T-cell leukemia/ lymphoma (ATL) is an incurable andaggressive neoplasia of CD4þ T-lymphocytes that is well known forinfiltrations of leukemic cells into various tissues and organs likeskin, lungs, liver, gastrointestinal tract, central nervous system,lymph nodes, and bone. ATL occurs in about 5% of carriers infectedwith the tumor virus human T-cell leukemia virus type 1 (HTLV-1)(Ishitsuka and Tamura, 2014; Matsuoka and Jeang, 2007). HTLV-1 isa widespread delta-retrovirus that transforms CD4þ T-lymphocytes,a process mainly initiated by the viral Tax oncoprotein (Currer et al.,2012; Matsuoka and Jeang, 2007). Tax does not only activate viral

olecular Virology, Friedrich-garten 4, 91054 Erlangen,22101.angen.de (C.F. Mohr),

ainz.de (A.B. Reske-Kunz),B. Biesinger),).

transcription through the HTLV-1 promoter located in the longterminal repeat of the provirus, but it also induces cellular geneexpression through several signaling pathways including cyclic-AMP response element-binding protein (CREB), serum responsefactor, activator protein, and nuclear factor of kappa light chainenhancer in B cells (NF-κB) (Grassmann et al., 2005).

The actin-bundling protein Fascin has received attention as atumor marker since it is important for migration and invasion ofcancer cells (Hashimoto et al., 2011; Jayo and Parsons, 2010). Fascinis a globular protein that is expressed in endothelial, neuronal,mesenchymal, and mature dendritic cells, but not in lymphocytesfrom healthy individuals (Adams, 2004; Hashimoto et al., 2005;Jayo and Parsons, 2010; Ross et al., 2000). Functionally, Fascinstabilizes filamentous actin and is concentrated in podosomes,filopodia, and invadopodia during cell migration (Adams, 2004;Jayo and Parsons, 2010). Fascin expression is highly upregulated invarious types of cancer including breast, lung, colon, esophagus,nasopharyngeal, pancreatic, stomach, ovary, brain, and skin cancer,and in lymphomas/leukemias including Hodgkin's Lymphoma, andEpstein–Barr-virus associated lymphomas (Hashimoto et al., 2011;

Fig. 1. A 1.6 kb fragment of the Fascin promoter is responsive to Tax. (a) Scheme of luciferase reporter constructs with different truncations of the human Fascin promoter(phF). The start site of the truncation (left side) is shown as distance from the transcription start site (þ1). (b,c) Jurkat T-cells were cotransfected with different phFconstructs and pEFTax1 or the control vector pEF. After 48 h, cells were lysed. (b) Luciferase activity was measured in relative light units (RLU) and normalized on proteincontent. Results are shown as mean RLU/protein7standard error (SE) of three independent experiments performed in triplicate and were compared using a t-test (*,Po0.05). UT, untreated cells. (c) Immunoblot of Fascin, Tax-1 and ß-Actin (ActB).

C.F. Mohr et al. / Virology 485 (2015) 481–491482

Hashimoto et al., 2005; Jayo and Parsons, 2010; Krikelis et al.,2013). Fascin expression correlates with decreased overall anddisease-free survival, occurrence of metastasis, enhanced chemo-resistance, and aggressiveness in some carcinomas (Chen et al.,2010; Ghebeh et al., 2014; Tan et al., 2013), thus, representing anovel therapeutic target. In cancer tissue, Fascin is accumulated onthe leading edge of cancer cells (Hashimoto et al., 2011), con-tributes to self-seeding of cancer cells (Kim et al., 2009), andpromotes tumor progression in vivo (Schoumacher et al., 2014).

Fig. 2. NF-κB signals are important for Tax-mediated activation of the 1.6 kb fragment ofcarrying the 1.6 kb fragment of the human Fascin promoter (phF1.6) with putative bindNumbers give the distance from the transcription start point (þ1). The Tax-responsiveindicated luciferase reporter constructs. After 48 h, luciferase activity was measuredrespective control7SE. If not stated otherwise, three independent experiments werePo0.01). (b) Cells were cotransfected with pcTax1, the Tax mutants M22 (NF-κB-deficienthe luciferase reporter vectors HTLV-U3R (upper panel, black bars, representative resul(lower panel, n¼3). (c) Cells were transfected with phF1.6 (20 mg) and pEFTax1 (20 mg), ewith phF1.6 and Tax (filled with pEF to 50 mg) were treated for 48 h with 10 mM of thephF1.6 (20 mg) served as control. Luciferase assays (upper panel) and an immunoblot of Fawith phF1.6 and pEFTax1 or the control vector pEF. Tax-transfected cells were treated witto luciferase assays (upper panel) and immunoblotting (lower panel). (e) After transfectiml, in water) or TPA (20 nM, in DMSO) or were left untreated (UT). Mean values were norand β-Actin (ActB) of Jurkat cells 48 h after treatment with 20 ng/ml TNF, 20 nM TPA orwere treated with TNF (20 ng/ml) or TPA (20 nM). The percentage of CD14þ cells was dwith PE-Cy5. Mean values of three independent experiments7SE are shown.

Regulation of Fascin transcription varies between tissue, cell,and cancer type and is achieved via promoter activation, silencingelements, or via micro RNAs binding to the 3′ untranslated region(3′UTR) (Hashimoto et al., 2011; Ma and Machesky, 2014). Diversecytokines, receptors, signaling pathways, and transcription factorshave been implicated in transcriptional activation of Fascin incarcinomas, e.g. tumor necrosis factor (TNF), interleukin-6 (Snyderet al., 2014), human epidermal growth factor receptor 2 (Grotheyet al., 2000), CREB (Hashimoto et al., 2009), or hypoxia-inducible

the human Fascin promoter. (a) Schematic representation of the luciferase constructing sites for the transcription factors NFAT (green), CREB (blue), and NF-κB (red).promoter region (TRR) is indicated. (b–e) Jurkat T-cells were transfected with theand results are shown as mean fold induction of RLU/protein normalized to theperformed in triplicate and results were evaluated using a t-test (*, Po0.05; **,t), M47 (CREB-deficient), M7 (NF-κB- and CREB-deficient), or a control (pcDNA) andt), 4-1BB wt prom (upper panel, white bars, representative result), or with phF1.6ither together with pIκBα-DN (10 mg) or pcDNA (10 mg, control), or cells transfectedIKK-β inhibitor ACHP or DMSO (solvent). Transfection of pEF (30 mg) together withscin, Tax-1 and ß-Actin (ActB) (lower panel) are shown. (d) Cells were cotransfectedh 1 mg/ml Cyclosporin A (CsA), DMSO (solvent), or left untreated (UT), and subjectedon of reporter constructs pGL3-Basic or phF1.6, cells were treated with TNF (20 ng/malized to pGL3-Basic UT. The y-axis was cut at 140. (f) Immunoblot of Fascin, Tax-1transfection with 50 mg pEFTax1 or 50 mg pEF. (g) SVT35 CD14-reporter Jurkat cellsetected 48 h after transfection by flow cytometry using α-CD14 antibodies labeled

C.F. Mohr et al. / Virology 485 (2015) 481–491 483

factor-1 (Zhao et al., 2014). Nevertheless, Fascin regulation inleukemia is largely unknown. We could previously show thatFascin is strongly upregulated in ATL-derived cells and in CD4þ

T-cells transformed by HTLV-1 or by the viral oncoprotein Tax.Fascin contributes to the invasive potential of ATL-derived cells(Kress et al., 2011). We were the first to demonstrate that tran-scriptional regulation of Fascin in ATL depends on NF-κB signaling

(Kress et al., 2011), which was confirmed in breast cancer (Snyderet al., 2011). Beyond, we could show that NF-κB-dependentinduction of Fascin and its contribution to invasive migration arealso features of latent membrane protein 1 (LMP1), an oncoproteinencoded by Epstein–Barr virus (Mohr et al., 2014), suggesting thatderegulation of Fascin by viral oncoproteins is a commonmechanism of viral oncogenesis.

C.F. Mohr et al. / Virology 485 (2015) 481–491484

Details about the mechanism of Fascin induction by viraloncoproteins are still lacking. Here, we report that the viral Taxoncoprotein activates a 1.6 kb fragment of the human Fascin pro-moter. Our study shows for the first time that Tax-mediatedtranscriptional induction of Fascin requires both NF-κB-dependentpromoter activation and a promoter-independent signaling path-way that is inhibited by the Src family kinase (SFK) inhibitor PP2.Hence, this is the first study to implicate PP2, an inhibitor ofmetastasis, in repression of Fascin.

Results

A 1.6 kb fragment of the human Fascin promoter is regulated by Tax

Tax is a potent inducer of Fascin (Kress et al., 2011), however,the detailed mechanism of Fascin induction by Tax is unknown.Depending on cell and cancer type, several different regulatoryelements contribute to Fascin promoter activity (Hashimoto et al.,2011; Ma and Machesky, 2014). To evaluate whether the Fascinpromoter is activated by the viral oncoprotein Tax and in T-lym-phocytes, Jurkat T-cells were transfected with reporter constructs(phF) harboring a luciferase gene under the control of serial 5′deletions of the human Fascin promoter (Bros et al., 2003), rangingfrom 3.1 kb (phF3.1) to 0.12 kb in length (phF0.12) (Fig. 1a). Allconstructs contained the transcription start site (þ1) and the first123 bp upstream of the start site (þ123). Upon cotransfection ofan expression vector for Tax (pEFTax1) or a control vector (pEF),luciferase assays revealed only background activity upon analysisof the shortest promoter constructs phF0.12 and phF0.15, con-firming earlier observations in dendritic cells (Bros et al., 2003)(Fig. 1b). However, all constructs ranging from phF0.21 to phF3.1exhibited basal activity significantly exceeding the backgroundactivity. phF0.21 carried the smallest promoter fragment to showconsiderable reporter activity independent of the presence of Tax(Po0.05), thus, comprising a functional core promoter. Among allconstructs tested, only the activity of construct phF1.6 was sig-nificantly transactivated by Tax (Po0.05). Therefore, the regionranging from �1499 to �1325 was named Tax-responsive region(TRR). Fascin promoter constructs longer or shorter than phF1.6were not inducible by Tax despite sufficient amount of Tax proteinto induce expression of endogenous Fascin (Fig. 1c). Taken toge-ther, these data indicate that the Fascin promoter contains a corepromoter active in T-lymphocytes and a TRR containing positiveregulatory elements presumably surrounded by inhibitoryelements.

Tax-mediated activation of the Fascin promoter depends on NF-κB

To ascertain the signaling pathways that are important for Tax-mediated induction of the 1.6 kb fragment of the Fascin promoter(phF1.6), this region was analyzed for putative binding sites oftranscription factors that have been described in Tax-mediatedgene regulation including CREB, NF-κB, and nuclear factor ofactivated T-cells (NFAT) (Grassmann et al., 2005; Qu and Xiao,2011; Saggioro, 2011; Silbermann et al., 2008) (Fig. 2a). Bioinfor-matics revealed several putative binding sites for CREB, one in theTRR (�1367/�1344), two within the core promoter (�47/�22;þ50/þ73), and four other CREB binding sites at (�204/�184;-320/�300; �354/�333; �414/�394). Moreover, the humanFascin promoter harbors four binding sites for NF-κB transcriptionfactors between the TRR and the core promoter (�1072/�1048;�976/�962; �430/�411; �209/�194). Finally, a putative NFATbinding site is located within the TRR (�1480/�1460). To assessthe role of NF-κB and CREB signals in Tax-induced activation ofphF1.6, Tax mutants that functionally differentiate these two

pathways were analyzed (Fig. 2b) (Smith and Greene, 1990). M22is deficient in NF-κB signaling, M47 in activation of the HTLV-1promoter U3R via viral cAMP response element (v-CRE) sites, andM7 in both pathways (Fig. 2b). First, we checked the functionalityof the Tax-mutants in luciferase assays in Jurkat T-cells. Consistentwith a previous publication (Smith and Greene, 1990), M22 had aweaker effect on the HTLV-1 U3R than Tax1, while M47 and M7 didnot transactivate the U3R at all (Fig. 2b, upper panel, black bars).Analysis of the NF-κB-dependent promoter of the costimulatoryreceptor 4-1BB (Pichler et al., 2008) showed that the NF-κB-defi-cient mutants, M22 and M7, are indeed strongly impaired intransactivation of 4-1BB wt prom compared to Tax and M47(Fig. 2a, upper panel, white bars). Having confirmed the func-tionality of the Tax mutants, we measured their impact on acti-vation of phF1.6 (Fig. 2b, lower panel). M22 and M7 did not induceactivation of phF1.6 compared with wild type (wt) Tax (pcTax1)(Po0.01), confirming earlier observations of Fascin expression(Kress et al., 2011). In contrast, the CREB-deficient mutant M47could still induce activation of phF1.6 comparable to wt Tax,although it was impaired in transcriptional induction of Fascin(Kress et al., 2011), suggesting that promoter transactivation aloneis not sufficient to induce Fascin. In conclusion, NF-κB, but notCREB signals contribute to Tax-mediated transactivation of phF1.6.To substantiate these findings, Jurkat T-cells were transfected withphF1.6 together with Tax (pEFTax1) and, additionally, either achemical inhibitor of IKK-β, ACHP (10 mM, 48 h), was applied, orcells were cotransfected with pIκBα-DN encoding a dominantnegative inhibitor of the classical NF-κB pathway (Fig. 2c). IκBα-DN carries two point mutations at serines 32/36 to alanine,thereby competing with endogenous IκBα for phosphorylationand degradation (Voll et al., 2000). Both ACHP (Po0.05) andpIκBα-DN (Po0.01) significantly decreased Tax-induced phF1.6activity (Fig. 2c, upper panel), and both experimental conditionsreduced the amount of Tax-induced endogenous Fascin protein(Fig. 2c, lower panel). To control if Tax is dependent on NFAT toactivate phF1.6, Jurkat T-cells were cotransfected with phF1.6 andpEFTax1 in the presence of 1 mg/ml of the NFAT inhibitor cyclos-porin A (CsA). CsA did neither affect Tax-mediated activation ofphF1.6 (Fig. 2d, upper panel), nor did it affect Tax-inducedexpression of endogenous Fascin protein. In conclusion, use of Tax-mutants and specific inhibitors revealed that Tax depends on NF-κB signals to activate the 1.6 kb fragment of the Fascin promoter.

To decipher if NF-κB-inducing compounds alone activatephF1.6, Jurkat T-cells were transfected with phF1.6 and, addition-ally, the NF-κB inducer TNF (20 ng/ml) or 12-O-tetra-decanoylphorbol-13-acetate (TPA; 20 nM) was applied (Fig. 2e)and compared with a control (pGL3-Basic). Contrary to cellscotransfected with Tax (pEFTax1), both TNF and TPA did not acti-vate the 1.6 kb fragment of the Fascin promoter. In accordancewith the latter finding, TNF and TPA did not induce endogenousFascin protein expression in Jurkat T-cells (Fig. 2f) compared withcells transfected with Tax (pEFTax1). To show that NF-κB signalingis active in presence of the concentrations of TNF and TPA used,SVT35 Jurkat T-cells harboring an NF-κB-driven CD14 reporterwere analyzed (Fig. 2g) (Ting et al., 1996). Treatment with TNF orTPA induced CD14 expression in around 80% of the cells, therebyproving CD14 reporter activation and underlining the NF-κB-inducing activity of either compound. Taken together these find-ings clearly show that NF-κB signals are important for Tax-medi-ated activation of the 1.6 kb fragment of the Fascin promoter, butNF-κB signals alone are neither sufficient to induce the Fascinpromoter nor to induce Fascin protein expression.

Fig. 3. The isolated Tax-responsive promoter region (TRR) is inducible by Tax. (a) Scheme of the luciferase reporter constructs phF0.21 and phF-TRR, the latter carrying theTax-responsive promoter region (TRR) of the human Fascin promoter cloned upstream of the Fascin core promoter phF0.21. Putative transcription factor binding sites forNFAT (green) and CREB (blue) are shown. The arrows indicate the location of a putative Tax-response element (TRE) and numbers indicate distances from the transcriptionstart point of the full-length human Fascin promoter. (b) Jurkat T-cells were cotransfected with luciferase reporter constructs phF0.21 or phF-TRR in combination withpEFTax1 or pEF. After 48 h, luciferase activity was measured (upper panel). Results of three independent experiments performed in triplicates are shown as mean foldinduction of RLU/protein7SE and were compared to values of cells cotransfected with the mock vector pEF using a t-test (**, Po0.01). Immunoblots of Fascin, Tax-1 and β-Actin were performed (lower panel).

Fig. 4. PP2 inhibits Tax-mediated induction of Fascin, but not Tax-mediated promoter transactivation. (a–c) Jurkat T-cells were transfected with pEFTax1 (40 mg), replenishedwith 10 mg pEF, and coincubated with inhibitors ACHP, PP2, U0126, LY294002, Glivecs, DMSO (solvent control) or PP3 (negative control for PP2) for 48 h. Untreated cells (UT)or cells transfected with 50 mg pEF served as controls. (a) Viability was measured by flow cytometry using SYTOX AAdvanced and the percentage of SYTOX-negative cells(viable cells) is depicted. (b) Quantitative PCR (qPCR) of Fascin mRNA. Copy numbers were normalized to those of β-Actin (ActB) and thereafter normalized to relative Fascinexpression of UT. Mean values of fold induction7SE are shown and were compared using a t-test (n¼3; **, Po0.01). (c) Immunoblot of Fascin, Tax-1 and ActB of cellstreated with PP2 in comparison to controls. Values indicate Fascin protein expression as detected by densitometry and normalized on PP3-treated cells (set as 1). (d,e) JurkatT-cells were cotransfected with luciferase reporter plasmids (d) phF1.6 or (e) pNF-κB-Luc and pEFTax1 or a control plasmid (pEF). Cells were treated with 5 mM of PP3, PP2,the IKK-β inhibitor ACHP (positive control), or with DMSO (solvent control) for 48 h. Results are shown as mean RLU/protein7SE of three independent experimentsperformed in triplicate and were compared using a t-test (*, Po0.05).

C.F. Mohr et al. / Virology 485 (2015) 481–491 485

Fig. 5. Tax-mediated Fascin induction is independent of the Src family kinase Lck. (a,b) Lck-deficient JCaM1.6 cells were either transfected with pEFTax1 (30 mg), expressionplasmids pEF–Lck–myc (murine Lck; 20 mg), or pFJ-5′Lck (human Lck, 20 mg), or a combination of Tax and the Lck-expression plasmids. pEF served as negative control andwas used to adjust transfections to a total amount of 50 mg of DNA. Cells were harvested 48 h after transfection. (a) qPCR of Fascin mRNA. Copy numbers were normalized tothose of β-Actin (ActB) and thereafter normalized to relative Fascin expression of untreated cells. Mean values of fold induction7SE are shown (n¼3). (b) Immunoblot of Lck,Fascin, Tax-1 and ActB.

C.F. Mohr et al. / Virology 485 (2015) 481–491486

An isolated Tax-responsive promoter region is inducible by Tax

So far, phF1.6, which carries a TRR, is the only Fascin promoterconstruct identified to be inducible by Tax. Moreover, we observedthat neither NF-κB-inducing compounds nor the LMP1 oncopro-tein transactivate phF1.6 (data not shown), although the latter is apotent inducer of NF-κB and of Fascin (Mohr et al., 2014). Since thepromoter constructs longer or shorter than phF1.6 are not Tax-responsive, the question was raised, if the TRR alone confersresponsiveness to Tax similar to an enhancer element. For thispurpose, the TRR was analyzed in more detail and cloned into avector with the 0.21 bp core promoter of Fascin (Fig. 3a), therebygenerating phF-TRR, which still contains a putative CREB and NFATsite within the TRR, but lacks obvious NF-κB-responsive elements.Additionally, the core promoter contains two further putativeCREB sites (�47/�22; þ50/þ73). The putative CREB binding sitewithin the TRR (�1367/�1344) harbors at (�1360/�1353) inantisense orientation high similarity to a v-CRE, a Tax-responseelement (TRE; TGACGTTG). Additionally, we identified a palin-dromic TRE (TGACGTCA) (Giam and Xu, 1989) at (�38/�31)within the CRE site of the core promoter. The latter is also heavilyflanked by GC-rich regions, which is a characteristic of TREs in theHTLV-1-promoter (Lenzmeier et al., 1998; Lundblad et al., 1998). Tocontrol if the TRR is still inducible by Tax, phF-TRR was cotrans-fected with pEFTax1 in Jurkat T-cells and compared to phF0.21carrying the core promoter. The isolated Tax-responsive promoterregion was still responsive to Tax leading to 1.5 fold inductioncompared with mock (Po0.01) (Fig. 3b). Reduction of the Taxeffect relative to phF1.6 (Fig. 1b) may be attributed to the lack ofputative NF-κB sites in phF-TRR. Since phF0.21 was not inducibleby Tax, the TRE located in the TRR may contribute to Tax-inducedactivation of the human Fascin promoter. Thus, our data suggestthat both Tax functions, (1) activation of NF-κB (Fig. 2b and c) and(2) activation of the TRR (Fig. 3) are required for induction ofFascin.

Induction of Fascin by Tax is not only dependent on NF-κB, but alsoinhibited by the Src family kinase (SFK) inhibitor PP2 inT-lymphocytes

Since putative NF-κB-binding sites are absent from the TRR(Figs. 2A and 3A), and pharmaceutical activation of NF-κB was notsufficient to activate the Tax-responsive Fascin promoter phF1.6(Fig. 2e), a screening was conducted to identify additional signal-ing pathways that contribute to Tax-mediated activation of Fascin.Therefore, Tax-transfected Jurkat T-cells were cultured in the

presence of different signaling pathway inhibitors. The NF-κBinhibitor ACHP served as positive control for Fascin inhibition(Kress et al., 2011). PP2 is primarily an inhibitor of SFK signaling(Bain et al., 2007), while PP3 is an inactive form of PP2 and servedas negative control. Additionally, inhibitors of the mitogen acti-vated protein kinase (MAPK) pathway (U0126), the phosphatidy-linositol-3-kinase (PI3K) pathway (LY294002), and of Abl tyrosinekinases and of T-cell receptor-mediated T-cell activation (Glivec)(Lin et al., 2013; Seggewiss et al., 2005) were analyzed. None of thetested inhibitors affected cell viability compared with the solventcontrol DMSO (Fig. 4a). Among all applied inhibitors, only ACHPand PP2 significantly affected Fascin mRNA induction by Tax. ACHPled to a reduction from 13-fold to 1-fold (Fig. 4b, Po0.01), con-firming earlier observations (Kress et al., 2011). In the presence ofPP2, Fascin transcript levels decreased from 14-fold to 8-fold(Fig. 4b, Po0.01). The impact of PP2 on Tax-mediated Fascininduction was confirmed on protein level (Fig. 4c), but was notdose-dependent (data not shown). This reveals that Fascininduction by Tax is not only regulated by NF-κB, but also impairedby the SFK inhibitor PP2 in transiently transfected T-lymphocytes.

PP2 does not inhibit Tax-mediated activation of the 1.6 kb fragment ofthe Fascin promoter

As Tax-mediated activation of the 1.6 kb fragment of the Fascinpromoter is regulated by NF-κB, it remained to be determined, ifthe newly-discovered Fascin inhibitor PP2 is also involved inpromoter transactivation. Addition of PP2 (5 mM) to Jurkat T-cellstransfected with Tax and phF1.6 did not affect Tax-mediatedinduction of phF1.6 (Fig. 4d), while the IKK-β inhibitor ACHP(10 mM) diminished Tax-mediated transactivation of phF1.6(Po0.05). To exclude an indirect effect of PP2 on NF-κB signaling,luciferase assays with an NF-κB-dependent luciferase reporterplasmid were performed (Fig. 4e). NF-κB-dependent luciferaseactivity was induced by Tax and blocked by the NF-κB inhibitorACHP, while PP2 was not able to reduce Tax-mediated NF-κBreporter activity. These results indicate that PP2 has no effect onTax-induced NF-κB activity. Furthermore, Tax-mediated Fascinpromoter activation is dependent on NF-κB, but independent ofsignaling pathways inhibited by PP2.

Tax-mediated Fascin induction is independent of the Src family kinaseLck

PP2 blocks a specific spectrum of SFKs, and the strongest sup-pression is observed for lymphocyte-specific protein tyrosine

Fig. 6. PP2 inhibits Fascin expression in HTLV-1-/Tax-transformed T lymphocytes. (a,b) HTLV-1-transformed MT-2 cells (left panel) and Tax-transformed Tesi cells (rightpanel) were treated with the IKK-β inhibitor ACHP (2.5 mM, positive control) and increasing amounts of PP2 (5, 10, 25, 50 mM) for 48 h. PP3 (50 mM, negative control for PP2)and DMSO (solvent) served as controls. (a) Viability was measured by flow cytometry using LIVE/DEADs fixable far red stain and the percentage of LIVE/DEADs -negative cellsis depicted. Etoposide (15 mM) served as positive control for cell death. (b) qPCR of Fascin mRNA. Copy numbers were normalized to those of β-Actin (ActB). Relative copynumbers of Fascin in DMSO-treated cells were set to 100%. Mean values of % Fascin mRNA7SE were compared using a t-test (nZ4; *, Po0.05).

Fig. 7. Model of Tax-mediated regulation of Fascin. Tax-mediated transcriptionalinduction of Fascin requires NF-κB- and TRE-dependent promoter activation, and anovel, promoter-independent, but PP2-inhibitable signaling pathway.

C.F. Mohr et al. / Virology 485 (2015) 481–491 487

kinase (Lck) (Bain et al., 2007), the most upstream tyrosine kinasein T-cell receptor signaling. Since Lck is downregulated by Tax andin HTLV-1-infected cells (Koga et al., 1989; Lemasson et al., 1997),we used Lck-deficient Jurkat T-cells (JCaM1.6) (Straus and Weiss,1992) to test if Tax-mediated activation of Fascin expression occursindependent of Lck. We found that even in the absence of Lck, Taxwas able to induce Fascin mRNA (Fig. 5a) and protein (Fig. 5b).Reconstitution of Lck by cotransfection of expression plasmidsencoding mouse Lck (pEF-Lck) or human Lck (pFJ-5′Lck) did notalter the extent of Tax-induced Fascin expression in JCaM1.6 cells(Fig. 5a and b). Thus, the SFK Lck is not involved in Tax-mediatedFascin induction.

Regulation of Fascin in HTLV-1/Tax-transformed T-lymphocytes isinhibited by PP2

To test the impact of PP2 on Fascin expression in a more rele-vant context of HTLV-1 pathogenesis, HTLV-1 transformed T-lym-phocytes, MT-2 (Fig. 6, left panels), and Tax-transformed T-lym-phocytes, Tesi (Fig. 6, right panels), were treated with increasingconcentrations of PP2 for 48 h. Cell viability was not significantlyaffected by treatment with PP2, PP3, or the IKK-β inhibitor ACHPcompared with DMSO, but was decreased by the topoisomeraseinhibitor etoposide (Po0.05; Fig. 6a). ACHP, a known inhibitor ofFascin in these two cell systems (Kress et al., 2011), was used ascontrol. In both transformed T-cell lines, Fascin mRNA levels sig-nificantly decreased in the presence of 25 and 50 mM PP2 (Fig. 6b),while lower concentrations of PP2 had no significant effect. In MT-2 cells, Fascin diminished to 74% and 57% in presence of PP2(25 mM and 50 mM, respectively; Po0.05), and to 72% (Po0.05) inpresence of ACHP (2.5 mM). In Tesi cells, the same conditionsreduced Fascin to 50% and 40% (Po0.05) for PP2, and to 47% forACHP. While PP2 reduced Fascin upon transient expression of Taxin Jurkat T-cells already at low concentrations (5 mM; Fig. 4b andc), higher concentrations of PP2 (25 mM and 50 mM) were requiredin continuously Tax-expressing cell lines MT-2 and Tesi (Fig. 6b).This discrepancy may be attributed to differences in signalingintensities and pathways between cells expressing Tax transientlycompared to Tax-transformed cells expressing Tax continuously.Taken together, Fascin regulation in HTLV-1-/Tax-transformedT-lymphocytes is not only dependent on NF-κB, but also relies on anovel PP2-sensitive signaling pathway. Thus, Tax-induced Fascinexpression is regulated by NF-κB-dependent and TRR-dependentpromoter activation, and by a novel promoter-independent, butPP2-inhibitable signaling pathway (Fig. 7).

C.F. Mohr et al. / Virology 485 (2015) 481–491488

Discussion

Upregulation of the actin-bundling protein Fascin is a commonfeature of several types of cancer and is associated with poorprognosis, enhanced migration, and invasion (Hashimoto et al.,2011; Jayo and Parsons, 2010; Tan et al., 2013). However, regula-tion of Fascin transcription strongly varies between cell, tissue, andcancer types (Hashimoto et al., 2011; Ma and Machesky, 2014). Inthis study, we provide a molecular mechanism for the regulationof Fascin by the viral oncoprotein Tax. The significance of ourfindings lies in the identification of a triple mode of transcriptionalinduction of Fascin by Tax, which requires (1) NF-κB-dependentpromoter activation, (2) a Tax-responsive region (TRR) in theFascin promoter, and (3) a promoter-independent mechanism. Thelatter identifies Fascin as a novel target of PP2, an inhibitor ofmetastatic spread (Nam et al., 2002).

We identified a 1.6 kb fragment of the human Fascin promoter(phF1.6) upstream of the transcription start site that is inducibleby Tax. The Tax-specific effect on the Fascin promoter was lost inpromoter constructs shorter or longer than phF1.6, suggesting thatthis Tax-responsive promoter region (TRR; �1499/�1325) is sur-rounded by repressor elements, a fact that could also be demon-strated in mature DCs (Bros et al., 2003), proposing that there aremechanisms in Fascin regulation conserved between DCs and Tax-expressing T-cells. Tax-mediated activation of phF1.6 could beattributed to NF-κB signaling by use of NF-κB-deficient Taxmutants and by inhibitors of NF-κB. This is in line with thenecessity of NF-κB signaling for transcriptional induction of Fascinby Tax (Kress et al., 2011), although it cannot be excluded that NF-κB signaling contributes to strategies of Fascin regulation otherthan promoter transactivation. LMP1 also induces Fascin expres-sion depending on NF-κB (Mohr et al., 2014), but, contrary to Tax,LMP1 does not transactivate phF1.6 (data not shown). Beyond, TNFand TPA, potent inducers of NF-κB (Chang et al., 1994; Hoesel andSchmid, 2013), did not affect Fascin expression or transactivationof phF1.6 in Jurkat T-cells contrary to results obtained in other celltypes (Onodera et al., 2009; Snyder et al., 2014). Thus, despiteinduction of the same signaling pathway, the fine-tuning of tran-scriptional activation varies between different oncoproteins andcell systems.

The identification of a TRR with putative enhancer abilities andsimilarities to Tax-responsive elements (TRE) might explain theresponsiveness of the human Fascin promoter to Tax independent ofthe presence of NF-κB-responsive elements. The TRE is a 21 bp triplerepeat in the HTLV-1-promoter containing an octamer motif TGACG(T/A)(C/G)(T/A) homologous to cellular CRE sites (Giam and Xu,1989; Jeang et al., 1988). Tax activates viral transcription throughassociation with phosphorylated CREB and the adjacent minorgroove GC-rich DNA, which is usually lacking in cellular CREs (Kimet al., 2010). Thus far, only few CRE-dependent target genes of Taxhave been identified, e.g. cyclin D1 (Kim et al., 2010) and GEM, whichencodes a small GTP-binding protein (Chevalier et al., 2014). Wepreviously found that the CREB-deficient Tax mutant M47 isimpaired in transcriptional induction of Fascin (Kress et al., 2011).Since CREB acts as a positive regulator of Fascin in several othermetastatic human cancer cells (Hashimoto et al., 2009; Li et al.,2013), we cannot exclude a contribution of the CREB signalingpathway to Tax-mediated induction of Fascin. In contrast, we dis-covered that M47 can still induce activation of phF1.6 comparable towildtype Tax, suggesting that promoter transactivation alone is notsufficient to fully activate Fascin transcription. This is supported byour finding that a novel, promoter- and NF-κB-independent strategycontributes to Tax-mediated Fascin regulation.

PP2, a strong inhibitor of Src family kinases (SFK) (Geahlenet al., 2004), significantly impaired Tax-induced Fascin expressionin T-cells without affecting NF-κB signaling or the Fascin promoter

phF1.6. SFK activity is deregulated in many types of cancer, affectsnumerous cellular substrates, and contributes to viral transfor-mation (Katsch et al., 2012; Parsons and Parsons, 2004; Reynoldset al., 2014). Besides, Src, the prototype SFK, is important for cancerinvasion (Guarino, 2010). In human T-cells, PP2 inhibits the SFKsLck and FynT, and, as a consequence, tyrosine kinase-dependentT-cell proliferation (Geahlen et al., 2004). Upon HTLV-1 transfor-mation, SFK expression patterns switch from Lck and FynT to Lynand FynB (Shuh et al., 2011; Weil et al., 1999). Downstreampathways of SFK that can be inhibited by PP2 include the januskinase (Jak)/signal transducer and activator of transcription (STAT)pathway (Shuh et al., 2011). The Jak/STAT pathway is upregulatedin different types of cancer including ATL and represents apotential therapeutic target (Ju et al., 2011). In breast and gastriccancer, an NF-κB- and STAT-dependent mechanism of Fascininduction has already been described. Although TNF and inter-leukin-6 induce binding of a complex of NF-κB p50:p65 and STAT3to the Fascin promoter at position (�1095/�979) (Snyder et al.,2014, 2011; Yao et al., 2014), it is unlikely that this mechanism isresponsible for Tax-mediated induction of Fascin: (1) both STAT3and STAT5 are constitutively activated in HTLV-1-transformed cells(Migone et al., 1995), but Tax only induces STAT5 and not STAT3(Nakamura et al., 1999), and the latter is inactive in Tax-trans-formed lymphocytes (Fung et al., 2005; Grassmann et al., 1989).(2) Complexes of the SFK Lyn with Jak3 and Jak3:STAT5 have beenidentified in HTLV-1-transformed cells and their activity is inhib-ited by PP2, but Jak3:STAT3 complexes are lacking (Migone et al.,1995; Shuh et al., 2011). (3) The promoter construct phF1.2, whichharbors the STAT/NF-κB-binding site at (�1095/�979) is notresponsive to Tax. (4) Finally, PP2 does not impair Tax-mediatedactivation of phF1.6.

Apart from blocking SFKs, PP2 diminishes signaling of p38α/p38ß MAPK, casein kinase 1δ (CK1 δ), receptor-interacting protein2 (RIP2), cytoplasmic tyrosine kinase (CSK), and cyclin G-asso-ciated kinase (GAK) (Bain et al., 2007). Amongst those, p38 MAPKhas been implicated in transcriptional elongation (Kumari et al.,2013). Tax is a potent inducer of certain transcription elongationfactors (Mann et al., 2014), and our probe used in qPCR spans anexon junction between exon 4/5 in the 3′ part of the Fascin tran-script. Therefore, the inhibitory role of PP2 on Fascin transcriptscould be explained by deregulating transcriptional elongation ofFascin rather than promoter transactivation, which is unaffected byPP2. Interestingly, PP2 impaired migration and invasion of humangastric cancer cells, accompanied by an inhibition of p38 phos-phorylation (Ma et al., 2014).

Regulation of gene expression is frequently influenced by cis-acting enhancer/suppressor elements located within introns, inthe 3′UTR, or even in exons of a gene. Potential silencing elementslocated distal from the transcription start site (Bros et al., 2003;Hashimoto et al., 2009) and in the first exon of Fascin (Hou et al.,2009) have been proposed to be involved in Fascin gene regula-tion. Therefore, it is possible that PP2, or even NF-κB inhibitors,block Tax-induced pathways that impair these silencers. Beyond,several micro RNAs (miRs) bind to the 3′UTR of Fascin. miR-133b,miR-133a, miR-143, and miR145 are associated with negativeregulation of Fascin in different types of cancer, e.g. miR-145 inbreast (Gotte et al., 2010), bladder (Chiyomaru et al., 2010), andcolon cancer (Feng et al., 2014). Deregulation of miRs has also beenassociated with HTLV-1-pathogenesis (Pichler et al., 2008), andinterestingly, miR-145 is downregulated in PBMCs from aggres-sive-type ATL patients and associated with a worsened prognosis(Xia et al., 2014; Yeung et al., 2008). Thus, it is also conceivable thatTax affects the Fascin 3′UTR.

Our data shows that PP2 represses Fascin shed new light on theability of PP2 to impair the invasive capacity of a variety of can-cers: (1) in ATL, the docking protein Cas-L results in markedly

C.F. Mohr et al. / Virology 485 (2015) 481–491 489

enhanced motility of lymphocytes. Cas-L is inducible by Tax andheavily tyrosine phosphorylated by focal adhesion kinase and SFKs(Iwata et al., 2005), and thus, inhibited by PP2 (Tsai et al., 2014).(2) PP2 restores the E-cadherin/catenin cell adhesion system inhuman epithelial cancer cells and it reduces cancer metastasis inmice (Nam et al., 2002). This is in line with findings showing thatknockdown of Fascin significantly represses invasiveness ofhepatocellular carcinoma cells and induces E-cadherin expression(Hayashi et al., 2011). (3) In pancreatic ductal adenocarcinomacells, PP2 is a powerful inhibitor of transforming growth factor-β-induced cell invasion and inhibits induction of the invasion-asso-ciated matrix metalloproteinase-2 (MMP-2) and MMP-9 genes(Bartscht et al., 2012). Interestingly, Fascin overexpression pro-motes invasion by elevating MMP-2 expression (Zhao et al., 2014),providing the possibility that PP2 affects MMP-2 via Fascin. Thus,targeting Fascin by PP2 might be relevant to several types ofmetastatic cancer.

Taken together, we show for the first time that Tax-mediatedtranscriptional induction of Fascin requires NF-κB- and TRR-dependent promoter activation, and a promoter-independent, butPP2-sensitive signaling pathway. Our study is the first to implicatePP2, an inhibitor of metastatic spread (Bartscht et al., 2012; Namet al., 2002), in repression of Fascin in cancer.

Materials and methods

Cell culture

The T-cell lines MT-2 (HTLV-1 in vitro transformed) (Yoshidaet al., 1982), Tesi (Tax-transformed) (Schmitt et al., 1998), Jurkat(from acute lymphoblastic leukemia) (Schneider et al., 1977), and theJurkat-derived cell lines SVT35 (carrying an NF-κB-driven CD14reporter gene) (Ting et al., 1996), and JCaM1.6 (deficient for thelymphocyte-specific protein tyrosine kinase (lck) gene (exon 7)) (Strausand Weiss, 1992) were cultured in RPMI 1640 with glutamine(GIBCO, Life Technologies, Darmstadt, Germany), penicillin/strepto-mycin, and 20% fetal calf serum (FCS; Tesi) or 10% FCS (all other celllines). Media for Tesi and Jurkat cell lines contained 45% Panserin401 (PAN-Biotech, Aidenbach, Germany) in addition, and media forTesi cells were supplemented with 40 U/ml interleukin-2 (RocheDiagnostics GmbH, Mannheim, Germany). Cell lines were checkedfor integrity by DNA profiling of eight different and highly poly-morphic short tandem repeat loci (DSMZ, Braunschweig, Germany).

Plasmids and cloning

Luciferase reporter constructs of the human Fascin promoter(phF3.1) and truncations thereof (phF2.7–phF0.21) have beendescribed (Bros et al., 2003). The luciferase reporter construct phF-TRR was cloned using primers 5′-ATTTGGTACCGCACCATCA-CACCCAGCTAA-3′ with a KpnI restriction site (in bold) and 5′-ATTTGGGCCCGCCCCTGGCAACCCC-3′ with an ApaI restriction site(in bold) for amplification of the region (�1499 to �1325) ofphF3.1. Fragments were cloned into phF0.21 via KpnI/ApaI sites.Luciferase reporter vectors of the HTLV-1 promoter (HTLV-1 U3R)and of the human 4-1BB promoter (4-1BB wt prom) have beendescribed (Mann et al., 2014; Pichler et al., 2008). Expressionplasmids for wildtype (wt) Tax pEFTax1 (in pEFneo) and the con-trol plasmid pEFneo were kindly provided by Dr. Higuchi andDr. Fujii (Shoji et al., 2009). Expression plasmids pEF–Lck–myc formouse Lck (Witte et al., 2008) and pFJ-5′Lck for human Lck(Albrecht et al., 1999) with a 5′ deletion in the non-coding regionhave been described. Additionally, plasmids for wt-Tax driven by aCMV promoter (pcTax) (Rimsky et al., 1988), the NF-κB-deficientTax mutant M22, the CREB-deficient Tax mutant M47, and the

double-deficient Tax mutant M7 (Smith and Greene, 1990), adominant negative inhibitor of IkBα (pIkBα-DN) (Voll et al., 2000),pcDNA3.1 (Life Technologies GmbH, Darmstadt, Germany), and theluciferase-reporter vector without eukaryotic promoter andenhancer sequences pGL3-Basic (Promega, Mannheim, Germany)were used.

Inhibitors and NF-κB stimuli

The following inhibitors were dissolved in DMSO and used at theindicated concentrations for 48 h: IKK-β inhibitor ACHP (2-Amino-6-(2-(cyclopropylmethoxy)-6-hydroxyphenyl)-4-(4-piperidinyl)-3-pyridinecarbonitrile) (Merck, Darmstadt, Germany), Src familykinase inhibitor PP2 (4-Amino-5-(4-chlorophenyl)-7-(t-butyl)pyr-azolo[3,4-d]pyrimidine) (Merck), a negative control for PP2, PP3 (4-amino-7-phenylpyrazolo[3,4-d]pyramidine) (Merck), Bcr-Abl kinaseinhibitor Glivec (Gleevec, imatinib, STI571; Novartis, Nuremberg,Germany), MAPK/ERK kinase inhibitor U0126 (1,4-diamino-2,3-dicyano-1,4-bis[2-aminophenylthio] butadiene) (Cell SignalingTechnology (CST), Danvers, MA, USA); and phosphoinositide3-kinase inhibitor LY294002 (CST). Jurkat and SVT35 cells werestimulated for 48 h with TNF (20 ng/ml; Immunotools, Friesoythe,Germany) dissolved in water or 20 nM TPA (12-O-tetra-decanoylphorbol-13-acetate)/PMA (phorbol-12-myristate-13-acet-ate) (CST) dissolved in DMSO.

Transient transfection by electroporation

Transient transfection was performed as described (Mohr et al.,2014). Briefly, 1�107 Jurkat T-cells were transfected by electro-poration using a total amount of 50 μg plasmid DNA. To block NF-κB signaling, 10 μg of pIκBα-DN were used. Cells were harvested48 h after transfection. The amount of green fluorescent protein(GFP) expressed from transfected pmaxGFP plasmid (Amaxa/Lonza) was used to evaluate transfection efficiency by flow cyto-metry (not shown).

Luciferase assays

Jurkat T-cells (5�106 cells) were transfected in triplicates byelectroporation using 20 mg of reporter plasmid and 30 mg ofexpression plasmids and harvested 48 h after transfection. Luciferaseactivity was measured as described (Mann et al., 2014) using anOrion Microplate Luminometer (Berthold, Pforzheim, Germany). RLUwas normalized to protein concentration determined by Bradfordassay. Additionally, protein lysates were obtained as described(Mohr et al., 2014). Lysates were sonicated after the freeze-and-thawlysis for 3�20 s and processed for immunoblotting.

Immunoblots

Protein lysates were obtained 48 h after transfection and wes-tern blot was performed as described (Mohr et al., 2014). Immu-noblots were probed using rabbit polyclonal antibodies anti-Lck (BDBiosciences; Heidelberg, Germany), mouse monoclonal antibodiesanti-Fascin (55K-2; Dako Deutschland GmbH, Hamburg, Germany),anti-β-actin (ACTB; AC-15; Sigma), and mouse antibodies to Tax-1,(provided by B. Langton through the AIDS Research and ReferenceReagent Program, Division of AIDS, NIAID, NIH) (Langton et al.,1988). Secondary antibodies conjugated with horseradish perox-idase were obtained from GE Healthcare (Little Chalfont, UK). Atleast three independent experiments were performed and onerepresentative result is shown. Densitometry was performed usingAdvanced Image Data Analyser (AIDA Version 4.22.034, RaytestIsotopenmessgeräte GmbH, Straubenhardt, Germany).

C.F. Mohr et al. / Virology 485 (2015) 481–491490

Flow cytometry

Cell surface markers were stained with monoclonal antibodiesanti-CD14-PE-Cy5 (1 h; Immunotools, Germany). Cell death wasdetected using SYTOX AAdvanced (Pacific Blue™ Annexin V/SYTOXs AADvanced™ for flow cytometry, Life Technologies) or, iffixation in 2% paraformaldehyde was necessary, LIVE/DEADs fix-able far red stain (Life Technologies).

Quantitative real-time RT-PCR (qPCR)

Total cellular RNA was isolated and qPCR of Fascin and ß-Actin(ACTB) transcripts was performed as described (Mohr et al., 2014).Briefly, primers ACTB-fwd: 5′-CCTCGCCTTTGCCGA-3, and ACTB-rev: 5′-TGGTGCCTGGGGCG-3, and the FAM (6-carboxy-fluorescein)/TAMRA (tetramethylrhodamine)-labeled probe 5′-FAM-CCGCCGCCCGTCCACACCCGCC-3′-TAMRA were used. Forquantitation of Fascin, a TaqMan Gene Expression Assay (Hs0097-9631_g1) was used. Expression levels were calculated by inter-polation from standard curves generated from plasmids carryingthe respective target sequences and computing the mean of tri-plicate samples. Each sample was measured in at least three bio-logical replicates. ACTB was used for normalization.

Bioinformatics and statistics

For prediction of transcription factor binding sites at thehuman Fascin promoter, MATInspector (Genomatix, Munich,Germany) was used (Cartharius et al., 2005). For statistical analy-sis, Microsoft Office Excel was applied and t-tests (unpaired) wereperformed to detect significant differences.

Acknowledgments

This article is dedicated to the memory of Martina Kalmer(deceased).

This work was supported by the Deutsche Forschung-sgemeinschaft (DFG, SFB796 project C6) and by the Inter-disciplinary Center for Clinical Research (IZKF) at the UniversityHospital of the Friedrich-Alexander-Universität Erlangen-Nurem-berg (project J26).

Sarah Strobel is greatly acknowledged for graduate assistance.We thank Armin Ensser and Doris Jungnickl (Institute of Clinicaland Molecular Virology, Erlangen, Germany) for providingreagents.

References

Adams, J.C., 2004. Fascin protrusions in cell interactions. Trends Cardiovasc. Med.14, 221–226.

Albrecht, J.C., Friedrich, U., Kardinal, C., Koehn, J., Fleckenstein, B., Feller, S.M., Bie-singer, B., 1999. Herpesvirus ateles gene product Tio interacts with nonreceptorprotein tyrosine kinases. J. Virol. 73, 4631–4639.

Bain, J., Plater, L., Elliott, M., Shpiro, N., Hastie, C.J., McLauchlan, H., Klevernic, I.,Arthur, J.S., Alessi, D.R., Cohen, P., 2007. The selectivity of protein kinase inhi-bitors: a further update. Biochem. J. 408, 297–315.

Bartscht, T., Lehnert, H., Gieseler, F., Ungefroren, H., 2012. The Src family kinaseinhibitors PP2 and PP1 effectively block TGF-beta1-induced cell migration andinvasion in both established and primary carcinoma cells. Cancer Chemother.Pharmacol. 70, 221–230.

Bros, M., Ross, X.L., Pautz, A., Reske-Kunz, A.B., Ross, R., 2003. The human fascingene promoter is highly active in mature dendritic cells due to a stage-specificenhancer. J. Immunol. 171, 1825–1834.

Cartharius, K., Frech, K., Grote, K., Klocke, B., Haltmeier, M., Klingenhoff, A., Frisch,M., Bayerlein, M., Werner, T., 2005. MatInspector and beyond: promoter ana-lysis based on transcription factor binding sites. Bioinformatics 21, 2933–2942.

Chang, C.C., Zhang, J., Lombardi, L., Neri, A., la-Favera, R., 1994. Mechanism ofexpression and role in transcriptional control of the proto-oncogene NFKB-2/LYT-10. Oncogene 9, 923–933.

Chen, L., Yang, S., Jakoncic, J., Zhang, J.J., Huang, X.Y., 2010. Migrastatin analoguestarget fascin to block tumour metastasis. Nature 464, 1062–1066.

Chevalier, S.A., Turpin, J., Cachat, A., Afonso, P.V., Gessain, A., Brady, J.N., Pise-Masison, C.A., Mahieux, R., 2014. Gem-induced cytoskeleton remodelingincreases cellular migration of HTLV-1-infected cells, formation of infected-to-target T-cell conjugates and viral transmission. PLoS Pathog. 10, e1003917.

Chiyomaru, T., Enokida, H., Tatarano, S., Kawahara, K., Uchida, Y., Nishiyama, K.,Fujimura, L., Kikkawa, N., Seki, N., Nakagawa, M., 2010. miR-145 and miR-133afunction as tumour suppressors and directly regulate FSCN1 expression inbladder cancer. Br. J. Cancer 102, 883–891.

Currer, R., Van, D.R., Jaworski, E., Guendel, I., Sampey, G., Das, R., Narayanan, A.,Kashanchi, F., 2012. HTLV tax: a fascinating multifunctional co-regulator of viraland cellular pathways. Front. Microbiol. 3, 406.

Feng, Y., Zhu, J., Ou, C., Deng, Z., Chen, M., Huang, W., Li, L., 2014. MicroRNA-145inhibits tumour growth and metastasis in colorectal cancer by targeting fascin-1. Br. J. Cancer 110, 2300–2309.

Fung, M.M., Chu, Y.L., Fink, J.L., Wallace, A., McGuire, K.L., 2005. IL-2- and STAT5-regulated cytokine gene expression in cells expressing the Tax protein of HTLV-1. Oncogene 24, 4624–4633.

Geahlen, R.L., Handley, M.D., Harrison, M.L., 2004. Molecular interdiction of Src-family kinase signaling in hematopoietic cells. Oncogene 23, 8024–8032.

Ghebeh, H., Al-Khaldi, S., Olabi, S., Al-Dhfyan, A., Al-Mohanna, F., Barnawi, R., Tul-bah, A., Al-Tweigeri, T., Ajarim, D., Al-Alwan, M., 2014. Fascin is involved in thechemotherapeutic resistance of breast cancer cells predominantly via the PI3K/Akt pathway. Br. J. Cancer 111, 1552–1561.

Giam, C.Z., Xu, Y.L., 1989. HTLV-I tax gene product activates transcription via pre-existing cellular factors and cAMP responsive element. J. Biol. Chem. 264,15236–15241.

Gotte, M., Mohr, C., Koo, C.Y., Stock, C., Vaske, A.K., Viola, M., Ibrahim, S.A., Ped-dibhotla, S., Teng, Y.H., Low, J.Y., Ebnet, K., Kiesel, L., Yip, G.W., 2010. miR-145-dependent targeting of junctional adhesion molecule A and modulation offascin expression are associated with reduced breast cancer cell motility andinvasiveness. Oncogene 29, 6569–6580.

Grassmann, R., Aboud, M., Jeang, K.T., 2005. Molecular mechanisms of cellulartransformation by HTLV-1 Tax. Oncogene 24, 5976–5985.

Grassmann, R., Dengler, C., Muller-Fleckenstein, I., Fleckenstein, B., McGuire, K.,Dokhelar, M.C., Sodroski, J.G., Haseltine, W.A., 1989. Transformation to con-tinuous growth of primary human T lymphocytes by human T-cell leukemiavirus type I X-region genes transduced by a Herpesvirus saimiri vector. Proc.Natl. Acad. Sci. U. S. A. 86, 3351–3355.

Grothey, A., Hashizume, R., Ji, H., Tubb, B.E., Patrick Jr., C.W., Yu, D., Mooney, E.E.,McCrea, P.D., 2000. C-erbB-2/ HER-2 upregulates fascin, an actin-bundlingprotein associated with cell motility, in human breast cancer cell lines. Onco-gene 19, 4864–4875.

Guarino, M., 2010. Src signaling in cancer invasion. J. Cell Physiol. 223, 14–26.Hashimoto, Y., Kim, D.J., Adams, J.C., 2011. The roles of fascins in health and disease.

J. Pathol. 224, 289–300.Hashimoto, Y., Loftis, D.W., Adams, J.C., 2009. Fascin-1 promoter activity is regu-

lated by CREB and the aryl hydrocarbon receptor in human carcinoma cells.PLoS One 4, e5130.

Hashimoto, Y., Skacel, M., Adams, J.C., 2005. Roles of fascin in human carcinomamotility and signaling: prospects for a novel biomarker? Int. J. Biochem. CellBiol. 37, 1787–1804.

Hayashi, Y., Osanai, M., Lee, G.H., 2011. Fascin-1 expression correlates withrepression of E-cadherin expression in hepatocellular carcinoma cells andaugments their invasiveness in combination with matrix metalloproteinases.Cancer Sci. 102, 1228–1235.

Hoesel, B., Schmid, J.A., 2013. The complexity of NF-kappaB signaling in inflam-mation and cancer. Mol. Cancer 12, 86.

Hou, J., Guo, Z.Y., Xie, J.J., Li, E.M., Xu, L.Y., 2009. Fascin overexpression is regulatedby the transactivation of the promoter but not by its hypomethylation in eso-phageal squamous cell carcinoma. Mol. Med. Rep. 2, 843–849.

Ishitsuka, K., Tamura, K., 2014. Human T-cell leukaemia virus type I and adult T-cellleukaemia-lymphoma. Lancet Oncol. 15, e517–e526.

Iwata, S., Souta-Kuribara, A., Yamakawa, A., Sasaki, T., Shimizu, T., Hosono, O.,Kawasaki, H., Tanaka, H., Dang, N.H., Watanabe, T., Arima, N., Morimoto, C.,2005. HTLV-I Tax induces and associates with Crk-associated substrate lym-phocyte type (Cas-L). Oncogene 24, 1262–1271.

Jayo, A., Parsons, M., 2010. Fascin: a key regulator of cytoskeletal dynamics. Int. J.Biochem. Cell Biol. 42, 1614–1617.

Jeang, K.T., Boros, I., Brady, J., Radonovich, M., Khoury, G., 1988. Characterization ofcellular factors that interact with the human T-cell leukemia virus type I p40x-responsive 21-base-pair sequence. J. Virol. 62, 4499–4509.

Ju, W., Zhang, M., Jiang, J.K., Thomas, C.J., Oh, U., Bryant, B.R., Chen, J., Sato, N.,Tagaya, Y., Morris, J.C., Janik, J.E., Jacobson, S., Waldmann, T.A., 2011. CP-690,550,a therapeutic agent, inhibits cytokine-mediated Jak3 activation and prolifera-tion of T cells from patients with ATL and HAM/TSP. Blood 117, 1938–1946.

Katsch, K., de Jong, S.J., Schmidt, M., Muller-Fleckenstein, I., Fleckenstein, B.,Albrecht, J.C., Biesinger, B., 2012. Species restriction of Herpesvirus saimiri andHerpesvirus ateles: human lymphocyte transformation correlates with distinctsignaling properties of viral oncoproteins. Virus Res. 165, 179–189.

C.F. Mohr et al. / Virology 485 (2015) 481–491 491

Kim, M.Y., Oskarsson, T., Acharyya, S., Nguyen, D.X., Zhang, X.H., Norton, L., Mas-sague, J., 2009. Tumor self-seeding by circulating cancer cells. Cell 139,1315–1326.

Kim, Y.M., Geiger, T.R., Egan, D.I., Sharma, N., Nyborg, J.K., 2010. The HTLV-1 taxprotein cooperates with phosphorylated CREB, TORC2 and p300 to activateCRE-dependent cyclin D1 transcription. Oncogene 29, 2142–2152.

Koga, Y., Oh-Hori, N., Sato, H., Yamamoto, N., Kimura, G., Nomoto, K., 1989. Absenceof transcription of lck (lymphocyte specific protein tyrosine kinase) message inIL-2-independent, HTLV-I-transformed T cell lines. J. Immunol. 142, 4493–4499.

Kress, A.K., Kalmer, M., Rowan, A.G., Grassmann, R., Fleckenstein, B., 2011. Thetumor marker Fascin is strongly induced by the Tax oncoprotein of HTLV-1through NF-kappaB signals. Blood 117, 3609–3612.

Krikelis, D., Bobos, M., Karayannopoulou, G., Resiga, L., Chrysafi, S., Samantas, E.,Andreopoulos, D., Vassiliou, V., Ciuleanu, E., Fountzilas, G., 2013. Expressionprofiling of 21 biomolecules in locally advanced nasopharyngeal carcinomas ofCaucasian patients. BMC Clin. Pathol. 13, 1.

Kumari, G., Ulrich, T., Gaubatz, S., 2013. A role for p38 in transcriptional elongationof p21 (CIP1) in response to Aurora B inhibition. Cell Cycle 12, 2051–2060.

Langton, B., Sliwkowski, M., Tran, K., Knapp, S., Keitelmann, E., Smith, C., Wall-ingford, S., Liu, H.-L., Ralston, J., Brandis, J., Coates, S., 1988. Development andcharacterization of monoclonal antibodies to the HTLV-I Tax (P40X) protein.Med. Virol., 8.

Lemasson, I., Robert-Hebmann, V., Hamaia, S., Duc, D.M., Gazzolo, L., Devaux, C.,1997. Transrepression of lck gene expression by human T-cell leukemia virustype 1-encoded p40tax. J. Virol. 71, 1975–1983.

Lenzmeier, B.A., Giebler, H.A., Nyborg, J.K., 1998. Human T-cell leukemia virus type1 Tax requires direct access to DNA for recruitment of CREB binding protein tothe viral promoter. Mol. Cell Biol. 18, 721–731.

Li, D., Jin, L., Alesi, G.N., Kim, Y.M., Fan, J., Seo, J.H., Wang, D., Tucker, M., Gu, T.L., Lee,B.H., Taunton, J., Magliocca, K.R., Chen, Z.G., Shin, D.M., Khuri, F.R., Kang, S.,2013. The prometastatic ribosomal S6 kinase 2-cAMP response element-bind-ing protein (RSK2-CREB) signaling pathway up-regulates the actin-bindingprotein fascin-1 to promote tumor metastasis. J. Biol. Chem. 288, 32528–32538.

Lin, Y.L., Meng, Y., Jiang, W., Roux, B., 2013. Explaining why Gleevec is a specific andpotent inhibitor of Abl kinase. Proc. Natl. Acad. Sci. U. S. A. 110, 1664–1669.

Lundblad, J.R., Kwok, R.P., Laurance, M.E., Huang, M.S., Richards, J.P., Brennan, R.G.,Goodman, R.H., 1998. The human T-cell leukemia virus-1 transcriptional acti-vator Tax enhances cAMP-responsive element-binding protein (CREB) bindingactivity through interactions with the DNA minor groove. J. Biol. Chem. 273,19251–19259.

Ma, X.M., Shen, Z.H., Liu, Z.Y., Wang, F., Hai, L., Gao, L.T., Wang, H.S., 2014. Hepar-anase promotes human gastric cancer cells migration and invasion byincreasing Src and p38 phosphorylation expression. Int. J. Clin. Exp. Pathol. 7,5609–5621.

Ma, Y., Machesky, L.M., 2014. Fascin1 in carcinomas: its regulation and prognosticvalue. Int. J. Cancer. 10.1002/ijc.29260

Mann, M.C., Strobel, S., Fleckenstein, B., Kress, A.K., 2014. The transcription elon-gation factor ELL2 is specifically upregulated in HTLV-1-infected T-cells and isdependent on the viral oncoprotein tax. Virology 464–465C, 98–110.

Matsuoka, M., Jeang, K.T., 2007. Human T-cell leukaemia virus type 1 (HTLV-1)infectivity and cellular transformation. Nat. Rev. Cancer 7, 270–280.

Migone, T.S., Lin, J.X., Cereseto, A., Mulloy, J.C., O'Shea, J.J., Franchini, G., Leonard, W.J., 1995. Constitutively activated Jak-STAT pathway in T cells transformed withHTLV-I. Science 269, 79–81.

Mohr, C.F., Kalmer, M., Gross, C., Mann, M.C., Sterz, K.R., Kieser, A., Fleckenstein, B.,Kress, A.K., 2014. The tumor marker Fascin is induced by the Epstein-Barr virus-encoded oncoprotein LMP1 via NF-kappaB in lymphocytes and contributes totheir invasive migration. Cell Commun. Signal. 12, 46.

Nakamura, N., Fujii, M., Tsukahara, T., Arai, M., Ohashi, T., Wakao, H., Kannagi, M.,Yamamoto, N., 1999. Human T-cell leukemia virus type 1 Tax protein inducesthe expression of STAT1 and STAT5 genes in T-cells. Oncogene 18, 2667–2675.

Nam, J.S., Ino, Y., Sakamoto, M., Hirohashi, S., 2002. Src family kinase inhibitor PP2restores the E-cadherin/catenin cell adhesion system in human cancer cells andreduces cancer metastasis. Clin. Cancer Res. 8, 2430–2436.

Onodera, M., Zen, Y., Harada, K., Sato, Y., Ikeda, H., Itatsu, K., Sato, H., Ohta, T., Asaka,M., Nakanuma, Y., 2009. Fascin is involved in tumor necrosis factor-alpha-dependent production of MMP9 in cholangiocarcinoma. Lab. Investig. 89,1261–1274.

Parsons, S.J., Parsons, J.T., 2004. Src family kinases, key regulators of signal trans-duction. Oncogene 23, 7906–7909.

Pichler, K., Schneider, G., Grassmann, R., 2008. MicroRNA miR-146a and furtheroncogenesis-related cellular microRNAs are dysregulated in HTLV-1-trans-formed T lymphocytes. Retrovirology 5, 100.

Qu, Z., Xiao, G., 2011. Human T-cell lymphotropic virus: a model of NF-kappaB-associated tumorigenesis. Viruses 3, 714–749.

Reynolds, A.B., Kanner, S.B., Bouton, A.H., Schaller, M.D., Weed, S.A., Flynn, D.C.,Parsons, J.T., 2014. SRChing for the substrates of Src. Oncogene 33, 4537–4547.

Rimsky, L., Hauber, J., Dukovich, M., Malim, M.H., Langlois, A., Cullen, B.R., Greene,W.C., 1988. Functional replacement of the HIV-1 rev protein by the HTLV-1 rexprotein. Nature 335, 738–740.

Ross, R., Jonuleit, H., Bros, M., Ross, X.L., Yamashiro, S., Matsumura, F., Enk, A.H.,Knop, J., Reske-Kunz, A.B., 2000. Expression of the actin-bundling protein fascin

in cultured human dendritic cells correlates with dendritic morphology andcell differentiation. J. Investig. Dermatol. 115, 658–663.

Saggioro, D., 2011. Anti-apoptotic effect of Tax: an NF-kappaB path or a CREB way?Viruses 3, 1001–1014.

Schmitt, I., Rosin, O., Rohwer, P., Gossen, M., Grassmann, R., 1998. Stimulation ofcyclin-dependent kinase activity and G1- to S-phase transition in humanlymphocytes by the human T-cell leukemia/lymphotropic virus type 1 Taxprotein. J. Virol. 72, 633–640.

Schneider, U., Schwenk, H.U., Bornkamm, G., 1977. Characterization of EBV-genomenegative "null" and "T" cell lines derived from children with acute lympho-blastic leukemia and leukemic transformed non-Hodgkin lymphoma. Int. J.Cancer 19, 621–626.

Schoumacher, M., El-Marjou, F., Lae, M., Kambou, N., Louvard, D., Robine, S., Vig-njevic, D.M., 2014. Conditional expression of fascin increases tumor progressionin a mouse model of intestinal cancer. Eur. J. Cell Biol. 93, 388–395.

Seggewiss, R., Lore, K., Greiner, E., Magnusson, M.K., Price, D.A., Douek, D.C., Dunbar,C.E., Wiestner, A., 2005. Imatinib inhibits T-cell receptor-mediated T-cell pro-liferation and activation in a dose-dependent manner. Blood 105, 2473–2479.

Shoji, T., Higuchi, M., Kondo, R., Takahashi, M., Oie, M., Tanaka, Y., Aoyagi, Y., Fujii,M., 2009. Identification of a novel motif responsible for the distinctive trans-forming activity of human T-cell leukemia virus (HTLV) type 1 Tax1 proteinfrom HTLV-2 Tax2. Retrovirology 6, 83.

Shuh, M., Morse, B.A., Heidecker, G., Derse, D., 2011. Association of SRC-relatedkinase Lyn with the interleukin-2 receptor and its role in maintaining con-stitutive phosphorylation of JAK/STAT in human T-cell leukemia virus type1-transformed T cells. J. Virol. 85, 4623–4627.

Silbermann, K., Schneider, G., Grassmann, R., 2008. Stimulation of interleukin-13expression by human T-cell leukemia virus type 1 oncoprotein Tax via a duallyactive promoter element responsive to NF-kappaB and NFAT. J. Gen. Virol. 89,2788–2798.

Smith, M.R., Greene, W.C., 1990. Identification of HTLV-I tax trans-activator mutantsexhibiting novel transcriptional phenotypes. Genes Dev. 4, 1875–1885.

Snyder, M., Huang, J., Huang, X.Y., Zhang, J.J., 2014. A signal transducer and activatorof transcription 3 �Nuclear Factor kappaB (Stat3 �NFkappaB) complex isnecessary for the expression of fascin in metastatic breast cancer cells inresponse to interleukin (IL)-6 and tumor necrosis factor (TNF)-alpha. J. Biol.Chem. 289, 30082–30089.

Snyder, M., Huang, X.Y., Zhang, J.J., 2011. Signal transducers and activators oftranscription 3 (STAT3) directly regulates cytokine-induced fascin expressionand is required for breast cancer cell migration. J. Biol. Chem. 286,38886–38893.

Straus, D.B., Weiss, A., 1992. Genetic evidence for the involvement of the lck tyr-osine kinase in signal transduction through the T cell antigen receptor. Cell 70,585–593.

Tan, V.Y., Lewis, S.J., Adams, J.C., Martin, R.M., 2013. Association of fascin-1 withmortality, disease progression and metastasis in carcinomas: a systematicreview and meta-analysis. BMC Med. 11, 52.

Ting, A.T., Pimentel-Muinos, F.X., Seed, B., 1996. RIP mediates tumor necrosis factorreceptor 1 activation of NF-kappaB but not Fas/APO-1-initiated apoptosis.EMBO J. 15, 6189–6196.

Tsai, P.C., Chu, C.L., Fu, Y.S., Tseng, C.H., Chen, Y.L., Chang, L.S., Lin, S.R., 2014.Naphtho[1,2-b]furan-4,5-dione inhibits MDA-MB-231 cell migration and inva-sion by suppressing Src-mediated signaling pathways. Mol. Cell Biochem. 387,101–111.

Voll, R.E., Jimi, E., Phillips, R.J., Barber, D.F., Rincon, M., Hayday, A.C., Flavell, R.A.,Ghosh, S., 2000. NF-kappa B activation by the pre-T cell receptor serves as aselective survival signal in T lymphocyte development. Immunity 13, 677–689.

Weil, R., Levraud, J.P., Dodon, M.D., Bessia, C., Hazan, U., Kourilsky, P., Israel, A., 1999.Altered expression of tyrosine kinases of the Src and Syk families in humanT-cell leukemia virus type 1-infected T-cell lines. J. Virol. 73, 3709–3717.

Witte, V., Laffert, B., Gintschel, P., Krautkramer, E., Blume, K., Fackler, O.T., Baur, A.S.,2008. Induction of HIV transcription by Nef involves Lck activation and proteinkinase C theta raft recruitment leading to activation of ERK1/2 but not NF kappaB. J. Immunol. 181, 8425–8432.

Xia, H., Yamada, S., Aoyama, M., Sato, F., Masaki, A., Ge, Y., Ri, M., Ishida, T., Ueda, R.,Utsunomiya, A., Asai, K., Inagaki, H., 2014. Prognostic impact of microRNA-145down-regulation in adult T-cell leukemia/lymphoma. Hum. Pathol. 45,1192–1198.

Yao, J., Qian, C.J., Ye, B., Zhao, Z.Q., Wei, J., Liang, Y., Zhang, X., 2014. Signal trans-ducer and activator of transcription 3 signaling upregulates fascin via nuclearfactor-kappaB in gastric cancer: Implications in cell invasion and migration.Oncol. Lett. 7, 902–908.

Yeung, M.L., Yasunaga, J., Bennasser, Y., Dusetti, N., Harris, D., Ahmad, N., Matsuoka,M., Jeang, K.T., 2008. Roles for microRNAs, miR-93 and miR-130b, and tumorprotein 53-induced nuclear protein 1 tumor suppressor in cell growth dysre-gulation by human T-cell lymphotrophic virus 1. Cancer Res. 68, 8976–8985.

Yoshida, M., Miyoshi, I., Hinuma, Y., 1982. Isolation and characterization of retro-virus from cell lines of human adult T-cell leukemia and its implication in thedisease. Proc. Natl. Acad. Sci. U. S. A. 79, 2031–2035.

Zhao, X., Gao, S., Ren, H., Sun, W., Zhang, H., Sun, J., Yang, S., Hao, J., 2014. Hypoxia-inducible factor-1 promotes pancreatic ductal adenocarcinoma invasion andmetastasis by activating transcription of the actin-bundling protein fascin.Cancer Res. 74, 2455–2464.