IKKa controls canonical TGFb–SMAD signaling to regulate genes expressingSNAIL and SLUG during EMT in Panc1 cellsMartina Brandl, Barbara Seidler, Ferdinand Haller, Jerzy Adamski, Roland M. Schmid, Dieter Saur andGünter Schneider

Journal of Cell Science 126, 2747� 2013. Published by The Company of Biologists Ltddoi: 10.1242/jcs.134791

There was an error published in J. Cell Sci. 123, 4231-4239.

In Fig. 4A, the cRel siRNA western blot panel was inadvertently constructed using the wrong images and all three western blots showed

incorrect loading controls. The correct images and loading controls are shown in the figure below. The mistake in the figure did not

affect the conclusions of the paper.

We apologise for this mistake.

Fig. 4. TGFb-dependent downregulation of E-cadherin is NFkB-independent. (A) Panc1 cells were transfected with a control or RelA/p65-, RelB- or c-Rel-

specific siRNA. At 24 hours after the transfection, cells were treated with 10 ng/ml TGFb or were left as an untreated control in DMEM without FCS. After an

additional 48 hours, western blots detected RelA/p65, RelB or c-Rel and E-cadherin expression. The membrane was stripped and probed for a-tubulin or b-actin to

ensure equal protein loading. (B) Panc1 (upper graph) and MDA-MB231 cells (lower graph) cells were co-transfected with a control or IKKa-specific siRNA and

500 ng of the pGL3control-, NFkB- or SMAD-luciferase reporter gene constructs as indicated. At 24 hours after the transfection, cells were treated with 10 ng/ml

TGFb or were left as an untreated control. Luciferase activity was measured 6 hours after the TGFb treatment (Student’s t-test: *P,0.05 versus control).

(C) Panc1 (upper graph) and MDA-MB231 (lower graph) cells were transfected with a control or IKKa-specific siRNA. At 48 hours after the transfection, cells

were stimulated with TGFb (10 ng/ml) for 20 minutes and binding of SMAD3 and SMAD4 to a SMAD consensus oligonucleotide was detected using ABCD

assays. Input represents 5% of whole-cell extract of control siRNA-transfected cells. (D) Panc1 cells were treated as in C. Immunoprecipitation was performed

with an IKKa-specific antibody or pre-immune serum as a control. Western blots of immunoprecipitates were probed with antibodies against IKKa and SMAD3.

Input represents 5% of whole-cell extract of control siRNA-transfected Panc1 cells. (E) MDA-MB231 cells were transfected with a control or IKKa-specific

siRNA. At 24 hours after the transfection, cells were treated with 10 ng/ml TGFb or were left as an untreated control in DMEM without FCS. After an additional

48 hours, western blots detected IKKa expression. The membrane was stripped and probed for b-actin to ensure equal protein loading.

Author Correction 2747

http://dx.doi.org/10.1242/jcs.071100

Research Article 4231

IntroductionTransforming growth factor (TGF) controls various cellularprocesses, including cell proliferation, differentiation and apoptosis.Acting as a tumor suppressor in the early stages of epithelialcarcinogenesis, TGF promotes tumor progression in advancedstages by inducing tumor growth, epithelial to mesenchymaltransition (EMT), invasion, evasion of immune surveillance andmetastasis (Massague, 2008). The pleiotropic TGF signals throughthe serine/threonine receptor kinases type I (TGFBR1) and type II(TGFBR2). Binding of TGF induces heteromeric complexformation between TGFBR1 and TGFBR2, whereby the TGFBR2phosphorylates and activates TGFBR1 (Derynck and Zhang, 2003).TGFBR1 induces canonical SMAD signaling by thephosphorylation of the C-terminal SSXS motif of the receptor-regulated SMADs (R-SMADs) such as SMAD2 and SMAD3.Upon phosphorylation, R-SMADs complex with the common-partner SMAD (co-SMAD) SMAD4, translocate into the nucleus,and control transcription of target genes in cooperation with othertranscription factors (Ellenrieder, 2008).

TGF is a potent inducer of EMT (Heldin et al., 2009;Moustakas and Heldin, 2007). During the process of EMT,epithelial cells lose polarity and intercellular adhesions, andacquire a fibroblastoid phenotype. Subsequently, the transcriptomeand proteome change from an epithelial to a mesenchymal profile,leading to expression of mesenchymal markers such as vimentinand N-cadherin and to the loss of epithelial markers such as E-cadherin. This mesenchymal phenotype correlates with thecapacity of cells to migrate to distant organs and, therefore, with

the initiation of metastasis (Thiery et al., 2009). Severaltranscription factors, like the zinc finger transcription factorsSNAIL and SLUG, the ZEB family factors ZEB1 and ZEB2, andthe basic helix–loop–helix (bHLH) factors E47 and Twist, areknown to induce and maintain the mesenchymal phenotype(Peinado et al., 2007).

The contribution of NFB signaling to the initiation andprogression of solid cancers is clearly documented (Basseres andBaldwin, 2006; Karin et al., 2002). In mammals, five differentNFB members [RelA/p65, RelB, c-Rel, p50/p105 (NFB1) andp52/p100 (NFB2)] have been identified. These transcriptionfactors are controlled by the IB kinase (IKK) complex, whichcontains at least three core subunits: the two kinasesIKK/IKK1/CHUK and IKK/IKK2 as well as one regulatorysubunit IKK/NEMO (Hayden and Ghosh, 2008; Scheidereit,2006). Furthermore, excellent work using a combined in vitro/invivo EMT model of mammary carcinogenesis (the TGF-dependent EpH4/EpRas model) revealed the essential contributionof NFB to the induction of EMT, maintenance of themesenchymal phenotype, and metastasis in vivo (Huber et al.,2004a; Huber et al., 2004b). Consistently, the small-moleculeIKK inhibitor BI 5700 interfered with NFB-driven EMT andmetastasis (Huber et al., 2010). To investigate the role of NFBin a well-established TGF-dependent EMT model of Panc1 cells(Ellenrieder et al., 2001), we used RNA interference (RNAi).Interestingly, we observed in this model that TGF-induced EMTis controlled by IKK in an NFB-independent and SMAD-dependent manner.

Accepted 17 August 2010Journal of Cell Science 123, 4231-4239 © 2010. Published by The Company of Biologists Ltddoi:10.1242/jcs.071100

SummaryThe epithelial to mesenchymal transition (EMT) is a crucial step in tumor progression, and the TGF–SMAD signaling pathway is aninductor of EMT in many tumor types. One hallmark of EMT is downregulation of the adherens junction protein E-cadherin, a processmediated by transcription factors such as the zinc fingers SNAIL and SLUG. Here, we report that the catalytic IB kinase (IKK)subunit IKK is necessary for the silencing of E-cadherin in a Panc1 cell model of TGF–SMAD-mediated EMT, independently ofNFB. IKK regulates canonical TGF–SMAD signaling by interacting with SMAD3 and controlling SMAD complex formation onDNA. Furthermore, we demonstrate that the TGF–IKK–SMAD signaling pathway induces transcription of the genes encodingSNAIL and SLUG. In addition, we demonstrate that IKK also modulates canonical TGF–SMAD signaling in human MDA-MB231breast cancer cells, arguing for a more general impact of IKK on the control of TGF–SMAD signaling. Taken together, thesefindings indicate that IKK contributes to the tumor-promoting function of the TGF–SMAD signaling pathway in particular cancers.

Key words: EMT, IKK, NFB, Pancreatic cancer, SMAD, SNAIL, SLUG, TGF

IKK controls canonical TGF–SMAD signaling toregulate genes expressing SNAIL and SLUG duringEMT in Panc1 cellsMartina Brandl1, Barbara Seidler1, Ferdinand Haller2, Jerzy Adamski2, Roland M. Schmid1, Dieter Saur1 andGünter Schneider1,*1II. Medizinische Klinik, Technische Universität München, Ismaninger Strasse 22, 81675 München, Germany2Helmholtz Zentrum München, Genomanalysezentrum, Institut für Experimentelle Genetik, Ingolstädter Landstrasse 1, 85764 Neuherberg,Germany*Author for correspondence (guenter.schneider@lrz.tum.de)

Jour

nal o

f Cel

l Sci

ence

ResultsIKK controls downregulation of E-cadherin and EMT inTGF-treated Panc1 cellsTo further characterize the role of the IKK–NFB signalingpathway in TGF-dependent EMT, we used the established Panc1cell model (Ellenrieder et al., 2001) (Fig. 1A). TGF inducescharacteristic morphological changes (Fig. 1B), and downregulationof the epithelial marker E-cadherin at the protein (Fig. 1C), mRNA(Fig. 1D) and transcriptional levels (Fig. 1E). In addition,mesenchymal genes encoding proteins such as vimentin or N-cadherin are induced upon TGF treatment (Fig. 1C). Because theIKK-NFB signaling pathway was shown to contribute to EMT(Huber et al., 2004a), we used short interfering RNA (siRNA)targeting the catalytical IKK subunits IKK and IKK toinvestigate the contribution of the NFB signaling pathway toTGF-induced EMT. As shown in Fig. 2A, TGF-induceddownregulation of E-cadherin was distinctly diminished in Panc1cells transfected with siRNA specific for IKK. To test the influenceof IKK depletion on the TGF-dependent induction ofmesenchymal markers we probed for N-cadherin and vimentinexpression. As shown in Fig. 2C, the loss in ability of Panc1 cellstransfected with IKK-specific siRNA to downregulate E-cadherinupon TGF treatment was accompanied by impaired upregulationof vimentin and N-cadherin. Furthermore, IKK-depleted cellsremained more epithelial after TGF treatment (see supplementarymaterial Fig. S1). To control RNAi-dependent off-target effects,we stably transfected Panc1 cells with a control or IKK-specificshort hairpin RNA (shRNA) expression vector. Again, we observedimpaired TGF-dependent downregulation of E-cadherin andupregulation of vimentin in IKK-depleted cells (Fig. 2D).Consistent with protein expression levels, the TGF-mediated

inhibition of E-cadherin-encoding mRNA expression (Fig. 2E) andE-cadherin promoter activity (Fig. 2F) was dependent on IKK.By contrast, TGF-dependent downregulation of E-cadherin wasnot impaired in cells transfected with siRNA targeting IKK,demonstrating specificity (Fig. 2B).

IKK controls TGF-induced migration of Panc1 cellsOne hallmark of EMT is the acquisition of a migratory phenotype(Thiery et al., 2009). To investigate the influence of IKK towardsTGF-mediated increased migration, we performed Boydenchamber assays. TGF significantly increased the migratorycapacity of control siRNA-transfected Panc1 cells (Fig. 3A,B). Inline with impaired downregulation of E-cadherin, TGF-inducedmigration was significantly decreased in cells transfected withIKK-specific siRNA (Fig. 3A,B), suggesting that IKK controlsthe markers of EMT in our model system.

IKK controls SMAD signalingTo investigate the contribution of NFB in more detail, we usedsiRNA targeting the NFB subunits harboring transactivationdomains, and screened for downregulation of E-cadherin. Theknockdown of RelA/p65, RelB or c-Rel did not influence TGF-induced downregulation of E-cadherin in Panc1 cells (Fig. 4A).Correspondingly, TGF failed to activate a NFB-dependentluciferase reporter gene construct (Fig. 4B). Because IKK wasrecently demonstrated to influence SMAD signaling (Descargueset al., 2008a; Descargues et al., 2008b; Marinari et al., 2008), weinvestigated TGF-induced SMAD activation using a SMAD-dependent luciferase reporter gene construct. Here, a distinctlydecreased activation of the SMAD reporter was observed in cellstransfected with IKK-specific siRNA and treated with TGF

4232 Journal of Cell Science 123 (24)

Fig. 1. Panc1 model of TGF-dependent EMT.(A)Illustration of the TGF-dependent EMTmodel. Serum-starved Panc1 cells were treatedwith 10 ng/ml TGF to induce EMT.(B)Photomicrographs (original magnification,40�) of Panc1 cells left untreated or treated with10 ng/ml TGF for 48 hours. (C)Panc1 cells weretreated for 48 hours with 10 ng/ml TGF or wereleft as an untreated control. Western blotdetermined expression of E-cadherin, vimentin andN-cadherin. -tubulin serves as control for equalprotein loading. (D)Quantitative analysis of E-cadherin-encoding mRNA expression. Panc1 cellswere treated for 48 hours with 10 ng/ml TGF orwere left as an untreated control. Total RNA wasprepared, and mRNA levels were quantified usingreal-time PCR analysis and normalized to GAPDHexpression levels (Student’s t-test: *P

(Fig. 4B). To test whether IKK controls binding of SMADs toDNA we used avidin-biotin-complex DNA (ABCD) assays in cellstransfected with control and IKK-specific siRNAs. TGF inducesbinding of SMAD3 and SMAD4 to a SMAD consensusoligonucleotide (Fig. 4C). Binding of SMAD3 and SMAD4 to theSMAD consensus oligonucleotide was abolished in Panc1 cellstransfected with IKK-specific siRNA and treated with TGF(Fig. 4C). Furthermore, we detected a complex of IKK andSMAD3 in TGF-stimulated cells in immunoprecipitations (Fig.4D). In solution, no complex of IKK with SMAD4 was detectablein immunoprecipitations of TGF-stimulated Panc1 cells (data notshown). To investigate whether the modulation of the TGF–SMAD pathway by IKK is a more general molecular mechanism,we used human MDA-MB231 breast cancer cells. As in Panc1

cells, TGF-dependent induction of SMAD transcriptional activity(Fig. 4B) and TGF-induced binding of SMAD3 and SMAD4 toa SMAD-consensus oligonucleotide (Fig. 4C) was dependent onIKK in MDA-MB231 cells. IKK expression was not changedby the treatment of MDA-MB231 cells with TGF, and transfectionof the IKK-specific siRNA induced an efficient knockdowncompared to control siRNA-transfected cells (Fig. 4E). Together,the data suggest that IKK forms a complex with SMAD3 andcontrols binding of a SMAD complex to DNA.

TGF-induced nuclear translocation of SMADs is notcontrolled by IKKTo gain insight into the mechanism by which IKK controlsSMAD activity and binding to DNA, we investigated nuclear

4233IKK and EMT

Fig. 2. IKK controls downregulation of E-cadherin in TGF-treated Panc1 cells. (A)Panc1 cells were transfected with a control or IKK-specific siRNA. At24 hours after the transfection, cells were treated with 10 ng/ml TGF or were left as an untreated control in DMEM without FCS. After an additional 48 hours,western blots detected IKK and E-cadherin expression. The membrane was stripped and probed for -actin to ensure equal protein loading. (B)Panc1 cells weretransfected with a control or IKK-specific siRNA and treated as for A. (C)Panc1 cells were transfected with a control or IKK-specific siRNA and treated as forA. Western blots detected IKK, E-cadherin, N-cadherin and vimentin expression. The membrane was stripped and probed for -actin to ensure equal proteinloading. (D)Panc1 cells were stably transfected with a control or IKK-specific shRNA expression vector and treated for 48 hours with 10 ng/ml TGF or wereleft as an untreated control in DMEM without FCS. Western blots detected IKK, E-cadherin and vimentin expression. The membrane was stripped and probed for-actin to ensure equal protein loading. (E)Panc1 cells were transfected with a control or IKK-specific siRNA. At 24 hours after the transfection, cells weretreated with 10 ng/ml TGF or were left as an untreated control in DMEM without FCS. After an additional 48 hours, E-cadherin mRNA levels were quantifiedusing real-time PCR analysis. The graph shows the TGF-mediated fold inhibition of E-cadherin-encoding mRNA expression of Panc1 cells transfected with acontrol or IKK-specific siRNA (Student’s t-test: *P

translocation of SMADs upon TGF treatment. Together withSMAD2 and SMAD3, IKK translocates into the nucleus ofTGF-treated Panc1 cells (Fig. 5A). TGF-induced translocationof SMAD3 was not impaired in Panc1 cells transfected with IKK-specific siRNA (Fig. 5B) nor in cells transfected with stable IKK-specific shRNA vector (Fig. 5C). Also, the TGF-induced nucleartranslocation of SMAD4 in MDA-MB231 cells occurred in IKK-depleted cells (supplementary material Fig. S2), which argues forIKK-independent nuclear translocation of SMADs.

TGF-induced IKK–SMAD signaling controls SNAIL andSLUG expressionBecause repression of E-cadherin depends on transcription factorslike SNAIL, SLUG, ZEB1, ZEB2, E47 or Twist, we investigatedregulation of these factors by the TGF signaling pathway. Intranscriptome profiles of TGF-treated Panc1 cells, no changes inexpression of TWIST, ZEB1, ZEB2 and E47 were observedcompared with untreated controls (data not shown). By contrast,TGF induced the expression of mRNAs encoding SNAIL andSLUG in control siRNA-transfected Panc1 cells (Fig. 6A, leftpanel), consistent with recent publications (Horiguchi et al., 2009;Takano et al., 2007). Induction of SNAIL and SLUG upon TGFtreatment was significantly impaired in cells transfected with IKK-specific siRNA (Fig. 6A). In addition, TGF-dependent inductionof SNAIL and SLUG is reduced in IKK-depleted MDA-MB231cells (Fig. 6A, right panel). Consistent with reduced binding of a

SMAD complex to a SMAD consensus oligonucleotide upon TGFtreatment in IKK-depleted cells (Fig. 4C), TGF-inducedexpression of classical SMAD-target genes, like PAI-1 or SMAD7,was dependent on IKK in Panc1 and MDA-MB231 cells(supplementary material Fig. S3). To demonstrate the dependencyof TGF-induced SNAIL and SLUG expression on SMADtranscription factors, we used SMAD4-, SMAD3- and SMAD2-specific siRNAs. TGF-dependent induction of mRNAs encodingSNAIL and SLUG was distinctly reduced in SMAD4- (Fig. 6B),SMAD3- (Fig. 6C) and SMAD2-specific (Fig. 6D) siRNA-transfected Panc1 cells compared to control siRNA-transfectedcells. Knockdown of SMAD4, SMAD3 and SMAD2 was controlledat the mRNA level (Fig. 6B–D). Consistently, knockdown ofSMAD2, SMAD3 as well as SMAD4 reduces TGF-induceddownregulation of E-cadherin, demonstrating the contribution ofall three SMADs to EMT in the Panc1 model (Fig. 6E).

IKK controls TGF-induced binding of SMADs to theSNAIL and SLUG promoterTo demonstrate direct binding of SMADs to the promoters ofSNAIL and SLUG we used chromatin immunoprecipitations (ChIP)and an antibody recognizing SMAD2 and SMAD3. For the SNAILpromoter, we used primers covering a region between –631 and–506 with respect to the transcriptional start of the human gene(Barbera et al., 2004). This region contains two copies of GC-richrepeats with high similarity to the Drosophila Mad-binding element,GCCGnCGc (Kim et al., 1997). SMAD3 binding to a GCCGnCGcelement of the bcl2 gene was recently demonstrated (Yang et al.,2006). Furthermore, the promoter alignment analysis tool ConTra(http://bioit.dmbr.ugent.be/ConTra/index.php) predicts binding ofhigh mobility group (HMG) proteins and SP1 to this region of theSNAIL promoter (Hooghe et al., 2008). As shown in Fig. 7A,TGF treatment induced binding of a SMAD2/3-containingcomplex to this SNAIL promoter region. Because we detected asimilar GC-rich element in the human SLUG promoter at position–785 from the ATG translation initiation site, we tested this SLUGpromoter region for binding of SMADs using primers covering–865 to –745 5� from the translational start. Again, TGF induceda SMAD2/3-containing complex to the 5� regulatory region of theSLUG gene in Panc1 cells (Fig. 7A). Because we observed a slightincrease of nuclear IKK upon TGF treatment of Panc1 cells(Fig. 5A) and IKK can function as a chromatin-associated kinase(Chariot, 2009), we tested whether IKK is recruited to the SNAILand SLUG promoters. Simultaneously with the binding ofSMAD2/3, we observed recruitment of IKK to the SNAIL aswell as to the SLUG promoter (Fig. 7A). No SMAD or IKKbinding upon TGF treatment was observed at the GAPDHpromoter, demonstrating specificity.

To investigate the dependency of TGF-induced SMAD2/3recruitment to the SNAIL and SLUG promoter regions on IKK,we repeated ChIP assays in IKK-depleted cells. Consistent withthe ABCD assays (Fig. 4C), recruitment of SMAD2/3 to SNAILand SLUG promoters was dependent on IKK and was notobserved in Panc1 cells transfected with either IKK-specificsiRNA (Fig. 7B) or with stable IKK-specific shRNA vector (Fig.7C) compared with controls. Again, no SMAD binding upon TGFtreatment was observed at the GAPDH promoter, demonstratingspecificity (Fig. 7B,C).

Together, these data demonstrate that IKK controls the bindingof a SMAD complex to the SNAIL and SLUG promoters inresponse to TGF.

4234 Journal of Cell Science 123 (24)

Fig. 3. TGF-induced migration of Panc1 cells depends on IKK.(A)Panc1 cells were transfected with a control or IKK-specific siRNA andstimulated 24 hours after the transfection with TGF (10 ng/ml) for anadditional 48 hours. Afterwards, 2.5�104 cells were seeded in the upper wellsof the multiwell insert system. Photomicrographs (original magnification,40�) of Giemsa-stained migrated Panc1 cells, treated as indicated, at 6 hoursafter seeding. (B)Panc1 cells were treated as in A. At 6 hours after seeding,Giemsa-stained migrated cells were counted (Student’s t-test: *P

DiscussionEMT is an important mechanism of tumor progression and anearly step in the metastatic cascade (Thiery et al., 2009). We nowdemonstrate that IKK controls canonical TGF–SMAD signalingin a pancreatic cancer EMT model to regulate the transcription ofthe zinc finger transcription factors SNAIL and SLUG, which areimportant contributors to EMT (Peinado et al., 2007) (Fig. 7D).This is consistent with the recent observations that SNAIL andSLUG are highly expressed in pancreatic cancer (Hotz et al.,2007). At the functional level, the contribution of SNAIL to therepression of E-cadherin and to EMT and metastasis in vivo hasbeen demonstrated in various pancreatic cancer models (Horiguchiet al., 2009; Takano et al., 2007; von Burstin et al., 2009).

Although the impact of NFB towards EMT and metastasis is notentirely understood at the molecular level, there is clear experimentalevidence that NFB contributes to the process of EMT and metastasis(Min et al., 2008). In addition to the TGF-dependent EpH4/EpRas

model (Huber et al., 2004a; Huber et al., 2004b), it was demonstratedthat increased NFB activity maintains a mesenchymal phenotypein carcinogen-treated mammary tumor cells obtained from MMTV-c-Rel transgenic animals (Shin et al., 2006). Furthermore, RelBsustains an invasive phenotype of estrogen receptor (ER)-negativebreast cancers (Wang, X. et al., 2007). In contrast to these clear linksbetween NFB and metastasis and EMT, we observed neitheractivation of NFB by TGF nor the dependency of E-cadherindownregulation on RelA/p65, RelB and c-Rel, arguing that TGF-induced EMT is NFB-independent in our model system.Furthermore, the fact that we did not observe NFB activation byTGF treatment of SMAD4-positive Panc1 cells is consistent witha recent report in the SMAD4-negative pancreatic cancer cell lineBxPC3. Here, TGF-dependent activation of NFB is abolished bythe re-expression of SMAD4 (Chow et al., 2010). However, theSMAD4-dependent mechanism blocking TGF-dependent NFBactivation remains unclear at the molecular level.

4235IKK and EMT

Fig. 4. TGF-dependent downregulation of E-cadherin is NFB-independent. (A)Panc1 cells were transfected with a control or RelA/p65-, RelB- or c-Rel-specific siRNA. At 24 hours after the transfection, cells were treated with 10 ng/ml TGF or were left as an untreated control in DMEM without FCS. After anadditional 48 hours, western blots detected RelA/p65, RelB or c-Rel and E-cadherin expression. The membrane was stripped and probed for -actin to ensure equalprotein loading. (B)Panc1 (upper graph) and MDA-MB231 cells (lower graph) cells were co-transfected with a control or IKK-specific siRNA and 500 ng of thepGL3control-, NFB- or SMAD-luciferase reporter gene constructs as indicated. At 24 hours after the transfection, cells were treated with 10 ng/ml TGF or wereleft as an untreated control. Luciferase activity was measured 6 hours after the TGF treatment (Student’s t-test: *P

In addition to the regulation of NFB, IKK has several NFB-independent functions (Chariot, 2009). With respect to metastasis, itwas nicely demonstrated in prostate cancer in vivo, that nuclearactivated IKK is linked to a metastasis pathway that depends onthe repression of the metastasis suppressor Maspin (Serpinb5).Interestingly, IKK-controlled Maspin repression seems to be NFB-

independent in this model (Luo et al., 2007). Although the Maspinpromoter is bound by SMADs (Wang, S. E. et al., 2007), it ispresently unclear whether IKK engages SMADs to repress thegene in prostate cancer cells. Nevertheless, IKK is definitelyinvolved in the control of TGF–SMAD target genes (Descargueset al., 2008a; Descargues et al., 2008b; Marinari et al., 2008). In

4236 Journal of Cell Science 123 (24)

Fig. 5. TGF-induced nuclear translocation ofSMADs is not controlled by IKK. (A)Panc1 cellswere treated with 10 ng/ml TGF for 20 minutes.Western blots of cytoplasmatic and nuclear extractsshow the expression of IKK, SMAD2, SMAD3,HDAC1 and -actin. HDAC1 controls nuclearfractionation. (B)Immunocytochemistry demonstratingthe nuclear translocation of SMAD2/3 in Panc1 cellstransfected with a control or IKK-specific siRNA. At48 hours after transfection, cells were treated with TGFfor 20 minutes or were left as an untreated control. Cellswere stained with a SMAD2/3 primary antibodyfollowed by Alexa-Fluor-488-labeled secondaryantibodies (green). Nuclei were counterstained withDAPI (blue). Merged photomicrographs with a 40×original magnification. (C)Panc1 cells stably transfectedwith a control or IKK-specific shRNA vector weretreated for 20 minutes with TGF or were left as anuntreated control. Western blots of cytoplasmatic andnuclear extracts show the expression of SMAD3,HDAC1 and -actin.

Fig. 6. IKK controls the SMAD target genes encoding SNAIL and SLUG. (A)Panc1 (left graph) or MDA-MB231 cells (right graph) were transfected with acontrol or IKK-specific siRNA. At 48 hours after the transfection, cells were treated with 10 ng/ml TGF or were left as an untreated control in DMEM withoutFCS. At 6 hours after the treatment with TGF, the mRNA levels of SNAIL and SLUG were quantified using real-time PCR analysis and normalized to GAPDHexpression levels (Student’s t-test: *P

keratinocytes, IKK is needed to mediate TGF-dependent growthinhibition (Descargues et al., 2008b). At the molecular level, IKKis necessary for the expression of TGF–SMAD target genes, likeMad1 or Ovol1, to counteract c-myc-driven proliferation (Descargueset al., 2008b). These results were confirmed in squamous cellcarcinomas (SCCs), where IKK expression is lost especially incancers with a highly invasive phenotype (Liu et al., 2006; Maedaet al., 2007; Marinari et al., 2008; Van Waes et al., 2007). Consistently,we observed a clear dependency of canonical SMAD signaling onIKK in the TGF-dependent Panc1 EMT model and in humanMDA-MB231 breast cancer cells. TGF treatment induces formationof a complex between SMAD3 and IKK in Panc1 cells.Furthermore, we demonstrated that IKK controls binding of aSMAD complex to DNA in Panc1 and MDA-MB231 cells.Interestingly, we detected the EMT promoters SNAIL and SLUG astarget genes of the TGF–IKK–SMAD signaling pathway.Considering the role of SNAIL and SLUG for EMT and tumorprogression (Peinado et al., 2007; Thiery et al., 2009), our resultspoint to a tumor-promoting function of IKK for particular cancers.

In contrast to keratinocytes, we detected no upregulation of Mad1or Ovol1 in transcriptome profiles of TGF-treated Panc1 cells (datanot shown). Therefore, it awaits further detailed investigations todecipher target gene specificity of the TGF–IKK–SMAD signalingpathway. One possible explanation for the specificity of the pathwaymight be the fact that SMADs have low affinity for DNA andcooperate with other transcription factors to achieve high affinityand target gene selectivity. Accordingly, the outcome of a TGF

response varies depending on the SMAD cooperating transcriptionfactors present at the time of exposure (Massague, 2000).Interestingly, we detected TGF–IKK-regulated binding of aSMAD complex to SNAIL and SLUG promoter regionscharacterized by GC-rich elements. In addition, the promoteralignment analysis tool ConTra predicts binding of HMG proteinsand SP1 to both promoter regions (Hooghe et al., 2008). SP1 wasrecently demonstrated to control the TGF-dependent EMT of Panc1cells (Jungert et al., 2007). Here, SP1 cooperates with SMAD3 toinduce the vimentin-encoding gene in response to TGF (Jungert etal., 2007). This cooperativity, together with the observation thatTGF induces binding of SMAD3, SMAD4 and SP1 to a SP1consensus oligonucleotide in Panc1 cells (Jungert et al., 2006), mightindicate that SP1 confers affinity and specificity for TGF–IKK–SMAD signaling towards the SNAIL and SLUG promoters in ourmodel. In addition, HMGA2 was recently demonstrated to synergizewith SMADs to activate the SNAIL promoter during EMT (Thuaultet al., 2008). Because HMGA2 binds to a region from –880 to –517of the SNAIL promoter in pancreatic cancer cells (Watanabe et al.,2009) and contributes to SLUG transcription (Thuault et al., 2006),HMGA2 is a further candidate for mediating affinity and specificityfor TGF–IKK–SMAD signaling. Whether a stereospecificHMGA2/SP1/SMAD nucleoprotein complex is recruited to theSNAIL and SLUG promoters in response to TGF treatment awaitsfurther investigation.

In addition to Panc1 cells, we observed that TGF–SMADsignaling is modulated by IKK in the mesenchymal human breast

4237IKK and EMT

Fig. 7. IKK controls binding of a SMAD complex to the SNAILand SLUG promoters. (A)Panc1 cells were treated with 10 ng/mlTGF for 20 minutes or were left as an untreated control. Chromatin ofPanc1 cells was immunoprecipitated with IKK- or SMAD2/3-specificantibodies or with IgG as a negative control. Precipitated DNA or 10%of the chromatin input was amplified with gene-specific primers forSNAIL, SLUG or GAPDH as indicated. (B)Panc1 cells weretransfected with a control or IKK-specific siRNA. At 48 hours afterthe transfection, cells were treated with 10 ng/ml TGF for 20 minutesor were left as an untreated control. Chromatin of Panc1 cells wasimmunoprecipitated with a SMAD2/3-specific antibody or with IgG asa negative control. Precipitated DNA or 10% of the chromatin input wasamplified with gene-specific primers for SNAIL, SLUG or GAPDH asindicated. (C)Panc1 cells stably transfected with a control or IKK-specific shRNA vector were treated for 20 minutes with TGF or wereleft as an untreated control. ChIP assays were carried out as for B.(D)Illustration of IKK-dependent TGF–SMAD signaling, drivingSNAIL and SLUG expression and EMT in Panc1 cells.

Jour

nal o

f Cel

l Sci

ence

cancer cell line MDA-MB231. Consistent with a recent report(Smith et al., 2009), we observed TGF-dependent induction ofSNAIL in MDA-MB231 cells. Similarly, we detected that TGFactivates expression of SLUG, PAI-1 and SMAD7 in an IKK-dependent fashion. Furthermore, TGF-induced SMAD binding toDNA is controlled by IKK. Because TGF signaling in MDA-MB231 cells is implicated in diverse processes like invasion (Farinaet al., 1998; Matsuura et al., 2010), metastasis (Yin et al., 1999),angiogenesis (Safina et al., 2007) or autophagy (Kiyono et al.,2009) our observations might argue that the described TGF–IKK–SMAD signaling controls molecular functions beyond EMT.However, this hypothesis needs further experimental validation.

In summary, we demonstrate that TGF–IKK–SMAD signalingdrives SNAIL and SLUG expression during EMT of Panc1 cells,pointing to a novel tumor-promoting pathway controlled by IKK.

Materials and MethodsCell culture, transfection, luciferase assays, siRNAs, plasmids and reagentsThe pancreatic cancer cell lines Panc1 was cultivated in DMEM supplemented with10% fetal calf serum (FCS) and 1% (w/v) penicillin/streptomycin (Invitrogen, Karlsruhe,Germany) as recently described (Fritsche et al., 2009; Schild et al., 2009). MDA-MB231 cells were a kind gift from Oliver Krämer (Friedrich-Schiller University, Jena,Germany) and cultivated in DMEM supplemented with 10% FCS and 1% (w/v)penicillin/streptomycin. TGF1 was purchased from PeproTech (Hamburg, Germany).Untreated controls received vehicle alone. Double-stranded siRNAs were transfectedat a final concentration of 50 nM using oligofectamine (Invitrogen, Karlsruhe, Germany)according to the manufacturer’s protocol. siRNAs were purchased from Eurofins(Ebersberg, Germany). Target sequences of the used siRNAs were: control siRNA, 5�-CAGTCGCGTTTGCGACTGGdtdt-3�; IKK siRNA, 5�-GTCTTGTCGC -CTAGAGCTAdtdt-3�; IKK siRNA, 5�-ATGTCATCCGATGGCACAAdtdt-3�;RelA/p65 siRNA, 5�-GATCAATG GCTACACAGGAdtdt-3�; RelB siRNA, 5�-GACTGCACCGACGGCATCTdtdt-3�; c-Rel siRNA, 5�-GAACG -CAGACCTTTGTTTTdtdt-3�; SMAD2 siRNA, 5�-GTCCCATGA AAAGA CT -TAAdtdt-3�; SMAD3 siRNA, 5�-GGAGAAATGGTGCGAGAAGdtdt-3�; and SMAD4siRNA, 5�-GGTCTTTGATTTGCGTCAGdtdt-3�. The 3xNFB luciferase reporterwas as described (Häussler et al., 2005). Stable transfection of Panc1 cells was carriedout as described (Schneider et al., 2006). The SureSilencing shRNA plasmid waspurchased from SuperArray Bioscience (Frederick, MD). The shRNA target sequenceswere: control shRNA, 5�-GGAATCTCATTCGATGCATAC-3� and IKK shRNA, 5�-AGCGTGCCATTGATCTATATA-3�. The SMAD luciferase reporter (pSMAD-Luc)contains the 5�-AGTATGTCTAGACTGAAGTATG TCTAGACTGAAG TATG -TCTAGACTGA-3� SMAD-binding site between the NheI and BglII restriction sitesof pTL-Luc (Panomics, Santa Clara, CA) and was purchased from BioCat (Heidelberg,Germany). pGL3control and phRG-B vectors were purchased from Promega(Mannheim, Germany). The 5� regulatory region of the murine E-cadherin gene wasamplified by PCR using the following primers: E-cadherin-FW, 5�-TAGGAAGCTGGGAAG-3� and E-cadherin-RV 5�-CTCGGGTGCGGTCG-3�. Thefragment corresponds to –174 to +94 of the murine E-cadherin promoter, containingthe two proximal E-boxes. The resulting fragment was blunt-end cloned into SmaIopened phRG-B (Promega). All plasmids were verified by sequencing. Transfectionsof the luciferase reporter genes (500 ng per well) were performed using FuGene6(Roche Applied Science, Mannheim, Germany) according to the manufacturer’sprotocol in 12-well plates. After stimulation with TGF1 for indicated time points,cells were incubated in lysis buffer (Promega), harvested, and cleared by centrifugation.Lysates were normalized for protein content. Luciferase activity was determined in aLB 9501 luminometer (Berthold, Bad Wildbad, Germany) using a renilla luciferaseassay system (Promega).

Western blotting and immunoprecipitationWhole-cell lysates were prepared and western blots and immunoprecipitations werecarried out as recently described (Fritsche et al., 2009; Schneider et al., 2010;Schneider et al., 2006). Nuclear extracts were prepared as described (Häussler et al.,2005). The following antibodies were used: HDAC1 (Millipore, Schwalbach,Germany), SMAD2/3 (FL-425, sc-8332), SMAD4 (B-8, sc-7966), RelA/p65 (C-20,sc-372), -tubulin (B-7, sc-5286), RelB (C-19, sc-226), c-Rel (C, sc-71), vimentin(VI-REI1, sc-51721) (Santa Cruz Biotechnology, Santa Cruz, CA), IKK (CellSignaling, Danvers, MA), E-cadherin and N-cadherin (BD Biosciences, Heidelberg,Germany), IKK (clone 10AG2) (Biomol, Hamburg, Germany) and -actin (Sigma-Aldrich, Munich, Germany). Western blots were scanned and quantified usingOdyssey Infrared Imaging System (LI-COR Biosciences, Bad Homburg, Germany).

Quantitative reverse-transcriptase PCRTotal RNA was isolated from pancreatic carcinoma cell lines using the RNeasy kit(Qiagen, Hilden, Germany) following the manufacturer’s instructions. Quantitative

mRNA analyses were performed as previously described using real-time PCRanalysis (TaqMan, PE Applied Biosystems, Norwalk CT) (Fritsche et al., 2009;Schneider et al., 2006). Expression was normalized to GAPDH expression levels.Primer sequences were: SNAIL-FW, 5�-CCCAATCGGAAGCCTAACTA-3�;SNAIL-RV, 5�-CAGGACAGAGTCCCAGATGAG-3�; SLUG-FW, 5�-ATGCATATTCGGACCCACAC-3�; SLUG-RV, 5�-GCAGATGAGCCC -TCAGATTT-3�; SMAD2-FW, 5�-TCAGTTCCGCCTCCAATC-3�; SMAD2-RV,5�-CAAGCCACGCTAGGAAAAC-3�; SMAD3-FW, 5�-CCACGCAGAAC -GTCAACA-3�; SMAD3-RV, 5�-TTGAAGGCGAACTCACACAG-3�; SMAD4-FW,5�-ACAAGGTGGAGAGAGTGAAACA-3�; SMAD4-RV, 5�-CTCCAGAGAC -GGGCATAGAT-3�; PAI-1-FW, 5�-AATCAGACGGCAGCACTGTCT-3�; PAI-1-RV,5�-GGCAGTTCCAGGATGTCGTAGT-3�; Smad7-FW, 5�-TGCTCCCATCCT -GTGTGTTAAG-3�; Smad7-RV, 5�-TCAGCCTAGGATGGTACCTTGG-3�;GAPDH-FW, 5�-CGTGGAAGGACTCATGACCA-3�; and GAPDH-RV, 5�-GCCATCACGCCACAGTTTC-3�.

Chromatin immunoprecipitation assaysChromatin immunoprecipitation (ChIP) assays were performed as recently described(Fritsche et al., 2009; Schneider et al., 2010; Schneider et al., 2006). An equalamount of chromatin (50–100 g) was used for each precipitation. The followingantibodies were used: SMAD2/3 (FL-425, sc-8332) and control IgG from SantaCruz Biotechnology (Heidelberg, Germany) and IKK from BD Biosciences(Heidelberg, Germany). One-twentieth of the precipitated chromatin was used foreach PCR reaction. To ensure linearity, 28–38 cycles were performed, and onerepresentative result is shown. Sequences of the promoter specific primers were:GAPDH-FW, 5�-AGCTCAGGCCTCAAGACCTT-3�; GAPDH-RV, 5�-AAGAAGATGCGGCTGACTGT-3�; SNAIL-FW, 5�-CGCTCCGTAAA CACT -GGATAA-3�; SNAIL-RV, 5�-GAAGCGAGGAAAGGGACAC-3�; SLUG-FW,5�-GCCTG CCTTTAGAGGGCTAC-3�; and SLUG-RV, 5�-TGCGCTA -CTCAGGGCTTC-3�.

Avidin-biotin-complex DNA assayA total of 1000 g of untreated or TGF-treated Panc1 whole-cell extract wasincubated with 3 g biotinylated double-stranded oligonucleotides for 4 hours withconstant rotation at 4°C in a total volume of 400 l immunoprecipitation (IP) buffer(50 mM HEPES, 150 mM NaCl, 1 mM EDTA, 0.5% NP-40, 10% glycerol, 1 mMdithiothreitol, 1 mM phenylmethysulfonylfluoride and 5 mM NaF) as described(Schild et al., 2009). After addition of 50 l equilibrated streptavidin agarose beads(Invitrogen), incubation was continued overnight at 4°C on a rotator. Beads werecollected by centrifugation and washed repeatedly with IP buffer. Afterwards, theprecipitated beads were boiled in Laemmli sample buffer and proteins were separatedvia SDS-PAGE. SMAD3 and SMAD4 were detected by western blot. The following3� biotinylated oligonucleotides were used: SMAD-s, 5�-TCGA GAGCCAG -ACAAAAAGCCAGACATTTAGCCAGACAC-3� and SMAD-as, 5�-GTGTCT -GGCTAAATGTCTGGCTTTTTGTCTGGCTCTCGA-3�.

Boyden chamber assayCell migration was determined using a Boyden chamber assay. HTS 24-MultiwellInsert Systems were purchased from BD Biosciences (Bedford, VA). Panc1 cellswere transfected with control or IKK-specific siRNAs and stimulated with TGF(10 ng/ml) for 48 hours. Afterwards, 2.5�104 cells were seeded in serum-freeDMEM in the upper wells of the multiwell insert system. The lower wells were filledwith 500 l DMEM supplemented with 10% fetal bovine serum. After 6 hours in anincubator (5% CO2, 37°C) cells on the upper surface of the filter were removed witha cotton-tipped swab. Migrated cells on the lower side of the membrane were fixedwith methanol, stained with Giemsa solution (Merck, Darmstadt, Germany) andcounted.

ImmunocytochemistryCells were washed twice with PBS and fixed for fluorescence microscopy byincubation for 10 minutes at 37°C in PBS containing 3.7% formaldehyde. Fordetection using specific antibodies, cells were permeabilized with 0.5% Triton X-100 in PBS for 5 minutes at room temperature. After washing the cells twice withPBS, unspecific antibody binding sites were blocked by incubation of cells in 3%BSA (in PBS) for 30 minutes at room temperature. The following antibodies wereused: SMAD2/3 (Fl-425, sc-8332), SMAD4 (B-8, sc-7966) antibody (Santa CruzBiotechnology) and Alexa Fluor 488 goat anti-rabbit IgG (Invitrogen) as secondaryantibody. For counterstaining of the nucleus, fixed cells were incubated for 2 minutesin medium containing 600 nM DAPI-solution (Biomol). Cells were washed twice inPBS, mounted on slides with VectaShield mounting medium (Vector Labs, Grünberg,Germany) and examined with an Axiophot epifluorescence microscope (Zeiss, Jena,Germany) using an 40� oil immersion objective. Optical data were collected withAxioVision Rel. 4.6 (Zeiss, Jena, Germany).

Statistical methodsAll data were obtained from at least three independent experiments performed intriplicate, and the results presented as mean and standard error of the mean (s.e.m.).To demonstrate statistical significance, a two-tailed Student’s t-test was used.* denotes a P-value of at least 0.05.

4238 Journal of Cell Science 123 (24)

Jour

nal o

f Cel

l Sci

ence

We thank Kerstin Hoffmann, Tatjana Netz and Birgit Kohnke-Ertelfor excellent technical support and Oliver Krämer for providing MDA-MB231 cells. This work was founded by DFG SFB456 (to G.S.) andDeutsche Krebshilfe (to D.S.).

Supplementary material available online athttp://jcs.biologists.org/cgi/content/full/123/24/4231/DC1

ReferencesBarbera, M. J., Puig, I., Dominguez, D., Julien-Grille, S., Guaita-Esteruelas, S., Peiro,

S., Baulida, J., Franci, C., Dedhar, S., Larue, L. et al. (2004). Regulation of Snailtranscription during epithelial to mesenchymal transition of tumor cells. Oncogene 23,7345-7354.

Basseres, D. S. and Baldwin, A. S. (2006). Nuclear factor-kappaB and inhibitor of kappaBkinase pathways in oncogenic initiation and progression. Oncogene 25, 6817-6830.

Chariot, A. (2009). The NF-kappaB-independent functions of IKK subunits in immunityand cancer. Trends Cell Biol. 19, 404-413.

Chow, J. Y., Ban, M., Wu, H. L., Nguyen, F., Huang, M., Chung, H., Dong, H. andCarethers, J. M. (2010). TGF-beta downregulates PTEN via activation of NF-kappaB inpancreatic cancer cells. Am. J. Physiol. Gastrointest. Liver Physiol. 298, G275-G282.

Derynck, R. and Zhang, Y. E. (2003). Smad-dependent and Smad-independent pathwaysin TGF-beta family signalling. Nature 425, 577-584.

Descargues, P., Sil, A. K. and Karin, M. (2008a). IKKalpha, a critical regulator of epidermaldifferentiation and a suppressor of skin cancer. EMBO J. 27, 2639-2647.

Descargues, P., Sil, A. K., Sano, Y., Korchynskyi, O., Han, G., Owens, P., Wang, X. J.and Karin, M. (2008b). IKKalpha is a critical coregulator of a Smad4-independentTGFbeta-Smad2/3 signaling pathway that controls keratinocyte differentiation. Proc.Natl. Acad. Sci. USA 105, 2487-2492.

Ellenrieder, V. (2008). TGFbeta regulated gene expression by Smads and Sp1/KLF-liketranscription factors in cancer. Anticancer Res. 28, 1531-1539.

Ellenrieder, V., Hendler, S. F., Boeck, W., Seufferlein, T., Menke, A., Ruhland, C., Adler,G. and Gress, T. M. (2001). Transforming growth factor beta1 treatment leads to anepithelial-mesenchymal transdifferentiation of pancreatic cancer cells requiring extracellularsignal-regulated kinase 2 activation. Cancer Res. 61, 4222-4228.

Farina, A. R., Coppa, A., Tiberio, A., Tacconelli, A., Turco, A., Colletta, G., Gulino, A.and Mackay, A. R. (1998). Transforming growth factor-beta1 enhances the invasivenessof human MDA-MB-231 breast cancer cells by up-regulating urokinase activity. Int. J.Cancer 75, 721-730.

Fritsche, P., Seidler, B., Schüler, S., Schnieke, A., Göttlicher, M., Schmid, R. M., Saur,D. and Schneider, G. (2009). HDAC2 mediates therapeutic resistance of pancreaticcancer cells via the BH3-only protein NOXA. Gut 58, 1399-1409.

Häussler, U., von Wichert, G., Schmid, R. M., Keller, F. and Schneider, G. (2005).Epidermal growth factor activates nuclear factor-kappaB in human proximal tubule cells.Am. J. Physiol. Renal Physiol. 289, F808-F815.

Hayden, M. S. and Ghosh, S. (2008). Shared principles in NF-kappaB signaling. Cell 132,344-362.

Heldin, C. H., Landstrom, M. and Moustakas, A. (2009). Mechanism of TGF-betasignaling to growth arrest, apoptosis, and epithelial-mesenchymal transition. Curr. Opin.Cell Biol. 21, 166-176.

Hooghe, B., Hulpiau, P., van Roy, F. and De Bleser, P. (2008). ConTra: a promoteralignment analysis tool for identification of transcription factor binding sites acrossspecies. Nucleic Acids Res. 36, W128-W132.

Horiguchi, K., Shirakihara, T., Nakano, A., Imamura, T., Miyazono, K. and Saitoh, M.(2009). Role of Ras signaling in the induction of snail by transforming growth factor-beta.J. Biol. Chem. 284, 245-253.

Hotz, B., Arndt, M., Dullat, S., Bhargava, S., Buhr, H. J. and Hotz, H. G. (2007).Epithelial to mesenchymal transition: expression of the regulators snail, slug, and twist inpancreatic cancer. Clin. Cancer Res. 13, 4769-4776.

Huber, M. A., Azoitei, N., Baumann, B., Grunert, S., Sommer, A., Pehamberger, H.,Kraut, N., Beug, H. and Wirth, T. (2004a). NF-kappaB is essential for epithelial-mesenchymal transition and metastasis in a model of breast cancer progression. J. Clin.Invest. 114, 569-581.

Huber, M. A., Beug, H. and Wirth, T. (2004b). Epithelial-mesenchymal transition: NF-kappaB takes center stage. Cell Cycle 3, 1477-1480.

Huber, M. A., Maier, H. J., Alacakaptan, M., Wiedemann, E., Braunger, J., Boehmelt,G., Madwed, J. B., Young, E. R. R., Marshall, D. R., Pehamberger, H. et al. (2010).BI 5700, a selective chemical inhibitor of IB kinase 2, specifically suppresses epithelial-mesenchymal transition and metastasis in mouse models of tumor progression. GenesCancer 1, 101-114.

Jungert, K., Buck, A., Buchholz, M., Wagner, M., Adler, G., Gress, T. M. and Ellenrieder,V. (2006). Smad-Sp1 complexes mediate TGFbeta-induced early transcription of oncogenicSmad7 in pancreatic cancer cells. Carcinogenesis 27, 2392-2401.

Jungert, K., Buck, A., von Wichert, G., Adler, G., König, A., Buchholz, M., Gress, T. M.and Ellenrieder, V. (2007). Sp1 is required for transforming growth factor-beta-inducedmesenchymal transition and migration in pancreatic cancer cells. Cancer Res. 67, 1563-1570.

Karin, M., Cao, Y., Greten, F. R. and Li, Z. W. (2002). NF-kappaB in cancer: frominnocent bystander to major culprit. Nat. Rev. Cancer 2, 301-310.

Kim, J., Johnson, K., Chen, H. J., Carroll, S. and Laughon, A. (1997). Drosophila Madbinds to DNA and directly mediates activation of vestigial by Decapentaplegic. Nature388, 304-308.

Kiyono, K., Suzuki, H. I., Matsuyama, H., Morishita, Y., Komuro, A., Kano, M. R.,Sugimoto, K. and Miyazono, K. (2009). Autophagy is activated by TGF-beta andpotentiates TGF-beta-mediated growth inhibition in human hepatocellular carcinomacells. Cancer Res. 69, 8844-8852.

Liu, B., Park, E., Zhu, F., Bustos, T., Liu, J., Shen, J., Fischer, S. M. and Hu, Y. (2006).A critical role for I kappaB kinase alpha in the development of human and mousesquamous cell carcinomas. Proc. Natl. Acad. Sci. USA 103, 17202-17207.

Luo, J. L., Tan, W., Ricono, J. M., Korchynskyi, O., Zhang, M., Gonias, S. L., Cheresh,D. A. and Karin, M. (2007). Nuclear cytokine-activated IKKalpha controls prostatecancer metastasis by repressing Maspin. Nature 446, 690-694.

Maeda, G., Chiba, T., Kawashiri, S., Satoh, T. and Imai, K. (2007). Epigenetic inactivationof IkappaB Kinase-alpha in oral carcinomas and tumor progression. Clin. Cancer Res. 13,5041-5047.

Marinari, B., Moretti, F., Botti, E., Giustizieri, M. L., Descargues, P., Giunta, A., Stolfi,C., Ballaro, C., Papoutsaki, M., Alema, S. et al. (2008). The tumor suppressor activityof IKKalpha in stratified epithelia is exerted in part via the TGF-beta antiproliferativepathway. Proc. Natl. Acad. Sci. USA 105, 17091-17096.

Massague, J. (2000). How cells read TGF-beta signals. Nat. Rev. Mol. Cell Biol. 1, 169-178.Massague, J. (2008). TGFbeta in Cancer. Cell 134, 215-230.Matsuura, I., Lai, C. Y. and Chiang, K. N. (2010). Functional interaction between Smad3

and S100A4 (metastatin-1) for TGF-beta-mediated cancer cell invasiveness. Biochem. J.426, 327-335.

Min, C., Eddy, S. F., Sherr, D. H. and Sonenshein, G. E. (2008). NF-kappaB and epithelialto mesenchymal transition of cancer. J. Cell Biochem. 104, 733-744.

Moustakas, A. and Heldin, C. H. (2007). Signaling networks guiding epithelial-mesenchymaltransitions during embryogenesis and cancer progression. Cancer Sci. 98, 1512-1520.

Peinado, H., Olmeda, D. and Cano, A. (2007). Snail, Zeb and bHLH factors in tumourprogression: an alliance against the epithelial phenotype? Nat. Rev. Cancer 7, 415-428.

Safina, A., Vandette, E. and Bakin, A. V. (2007). ALK5 promotes tumor angiogenesis byupregulating matrix metalloproteinase-9 in tumor cells. Oncogene 26, 2407-2422.

Scheidereit, C. (2006). IkappaB kinase complexes: gateways to NF-kappaB activation andtranscription. Oncogene 25, 6685-6705.

Schild, C., Wirth, M., Reichert, M., Schmid, R. M., Saur, D. and Schneider, G. (2009).PI3K signaling maintains c-myc expression to regulate transcription of E2F1 in pancreaticcancer cells. Mol. Carcinog. 48, 1149-1158.

Schneider, G., Saur, D., Siveke, J. T., Fritsch, R., Greten, F. R. and Schmid, R. M.(2006). IKKalpha controls p52/RelB at the skp2 gene promoter to regulate G1- to S-phaseprogression. EMBO J. 25, 3801-3812.

Schneider, G., Henrich, A., Greiner, G., Wolf, V., Lovas, A., Wagner, M., Wieczorek, M.,Reichardt, S., von Werder, A., Schmid, R. M. et al. (2010). Crosstalk between stimulatedNF-kB and the tumor suppressor p53. Oncogene 29, 2795-2806.

Shin, S. R., Sanchez-Velar, N., Sherr, D. H. and Sonenshein, G. E. (2006). 7,12-dimethylbenz(a)anthracene treatment of a c-rel mouse mammary tumor cell line inducesepithelial to mesenchymal transition via activation of nuclear factor-kappaB. Cancer Res.66, 2570-2575.

Smith, A. P., Verrecchia, A., Faga, G., Doni, M., Perna, D., Martinato, F., Guccione, E.and Amati, B. (2009). A positive role for Myc in TGFbeta-induced Snail transcriptionand epithelial-to-mesenchymal transition. Oncogene 28, 422-430.

Takano, S., Kanai, F., Jazag, A., Ijichi, H., Yao, J., Ogawa, H., Enomoto, N., Omata, M.and Nakao, A. (2007). Smad4 is essential for down-regulation of E-cadherin induced byTGF-beta in pancreatic cancer cell line PANC-1. J. Biochem. 141, 345-351.

Thiery, J. P., Acloque, H., Huang, R. Y. and Nieto, M. A. (2009). Epithelial-mesenchymaltransitions in development and disease. Cell 139, 871-890.

Thuault, S., Valcourt, U., Petersen, M., Manfioletti, G., Heldin, C. H. and Moustakas,A. (2006). Transforming growth factor-beta employs HMGA2 to elicit epithelial-mesenchymal transition. J. Cell Biol. 174, 175-183.

Thuault, S., Tan, E. J., Peinado, H., Cano, A., Heldin, C. H. and Moustakas, A. (2008).HMGA2 and Smads co-regulate SNAIL1 expression during induction of epithelial-to-mesenchymal transition. J. Biol. Chem. 283, 33437-33446.

Van Waes, C., Yu, M., Nottingham, L. and Karin, M. (2007). Inhibitor-kappaB kinase intumor promotion and suppression during progression of squamous cell carcinoma. Clin.Cancer Res. 13, 4956-4959.

von Burstin, J., Eser, S., Paul, M. C., Seidler, B., Brandl, M., Messer, M., von Werder,A., Schmidt, A., Mages, J., Pagel, P. et al. (2009). E-cadherin regulates metastasis ofpancreatic cancer in vivo and is suppressed by a SNAIL/HDAC1/HDAC2 repressorcomplex. Gastroenterology 137, 361-371, 371 e1-5.

Wang, S. E., Narasanna, A., Whitell, C. W., Wu, F. Y., Friedman, D. B. and Arteaga, C.L. (2007). Convergence of p53 and transforming growth factor beta (TGFbeta) signalingon activating expression of the tumor suppressor gene maspin in mammary epithelialcells. J. Biol. Chem. 282, 5661-5669.

Wang, X., Belguise, K., Kersual, N., Kirsch, K. H., Mineva, N. D., Galtier, F., Chalbos,D. and Sonenshein, G. E. (2007). Oestrogen signalling inhibits invasive phenotype byrepressing RelB and its target BCL2. Nat. Cell Biol. 9, 470-478.

Watanabe, S., Ueda, Y., Akaboshi, S., Hino, Y., Sekita, Y. and Nakao, M. (2009).HMGA2 maintains oncogenic RAS-induced epithelial-mesenchymal transition in humanpancreatic cancer cells. Am. J. Pathol. 174, 854-868.

Yang, Y. A., Zhang, G. M., Feigenbaum, L. and Zhang, Y. E. (2006). Smad3 reducessusceptibility to hepatocarcinoma by sensitizing hepatocytes to apoptosis throughdownregulation of Bcl-2. Cancer Cell 9, 445-457.

Yin, J. J., Selander, K., Chirgwin, J. M., Dallas, M., Grubbs, B. G., Wieser, R.,Massague, J., Mundy, G. R. and Guise, T. A. (1999). TGF-beta signaling blockadeinhibits PTHrP secretion by breast cancer cells and bone metastases development. J. Clin.Invest. 103, 197-206.

4239IKK and EMT

Jour

nal o

f Cel

l Sci

ence

SummaryKey words: EMT, IKK, NFkB, Pancreatic cancer, SMAD, SNAIL, SLUG,IntroductionResultsIKKa controls downregulation of E-cadherin and EMT in TGFb-treated Panc1IKKa controls TGFb-induced migration of Panc1 cellsIKKa controls SMAD signalingTGFb-induced nuclear translocation of SMADs is not controlled by IKKaTGFb-induced IKKa-SMAD signaling controls SNAIL and SLUG expressionIKKa controls TGFb-induced binding of SMADs to the SNAIL and

Fig. 1.Fig. 2.Fig. 3.DiscussionFig. 4.Fig. 5.Fig. 6.Fig. 7.Materials and MethodsCell culture, transfection, luciferase assays, siRNAs, plasmids and reagentsWestern blotting and immunoprecipitationQuantitative reverse-transcriptase PCRChromatin immunoprecipitation assaysAvidin-biotin-complex DNA assayBoyden chamber assayImmunocytochemistryStatistical methods

Supplementary materialReferences2747.pdfFig 1