A Signal Transducer and Activator of Transcription 3·Nuclear Factor κB (Stat3·NFκB) Complex Is Necessary for the Expression of Fascin in Metastatic Breast Cancer Cells in Response to Interleukin (IL)-6 and Tumor Necrosis Factor (TNF)-α*

Background: IL-6/Stat3 promote breast cancer metastasis through regulation of the fascin gene. Results: In addition to IL-6, TNF-α induces binding of a Stat3·NFκB complex to the fascin promoter to induce transcription. Conclusion: Both NFκB and Stat3 are required for cytokine-induced fascin expression and cell migration. Significance: Identification of proteins critical for breast cancer metastasis will reveal drug targets. IL-6 mediated activation of Stat3 is a major signaling pathway in the process of breast cancer metastasis. One important mechanism by which the IL-6/Stat3 pathway promotes metastasis is through transcriptional regulation of the actin-bundling protein fascin. In this study, we further analyzed the transcriptional regulation of the fascin gene promoter. We show that in addition to IL-6, TNF-α increases Stat3 and NFκB binding to the fascin promoter to induce its expression. We also show that NFκB is required for Stat3 recruitment to the fascin promoter in response to IL-6. Furthermore, Stat3 and NFκB form a protein complex in response to cytokine stimulation. Finally, we demonstrate that an overlapping STAT/NFκB site in a highly conserved 160-bp region of the fascin promoter is sufficient and necessary to induce transcription in response to IL-6 and TNF-α.

IL-6 mediated activation of Stat3 is a major signaling pathway in the process of breast cancer metastasis. One important mechanism by which the IL-6/Stat3 pathway promotes metastasis is through transcriptional regulation of the actin-bundling protein fascin. In this study, we further analyzed the transcriptional regulation of the fascin gene promoter. We show that in addition to IL-6, TNF-␣ increases Stat3 and NFB binding to the fascin promoter to induce its expression. We also show that NFB is required for Stat3 recruitment to the fascin promoter in response to IL-6. Furthermore, Stat3 and NFB form a protein complex in response to cytokine stimulation. Finally, we demonstrate that an overlapping STAT/NFB site in a highly conserved 160-bp region of the fascin promoter is sufficient and necessary to induce transcription in response to IL-6 and TNF-␣.
The STAT (signal transducers and activators of transcription) proteins are a family of transcription factors that play various roles in cellular processes, including immune response, apoptosis, and oncogenesis (1)(2)(3)(4). STATs are members of the JAK-STAT signaling pathway, which is activated by growth factors and cytokines (1,2). In response to cytokine binding to its cell surface receptor, the receptor dimerizes and activates its associated Janus kinase (JAK). The JAK phosphorylates the cell surface receptor, which serves as a docking site for the STATs. Upon recruitment of the STATs to the phosphorylated receptor, STATs become phosphorylated by JAKs on a conserved tyrosine near the Src homology 2 domain of STAT (1,2). The phosphorylated STATs dimerize and translocate to the nucleus where they bind to the promoters of STAT target genes to form enhancersomes with other transcriptional co-activators to activate transcription (5,6).
Stat3 is one member of the STAT family that functions in a wide range of cellular processes, including wound healing, postnatal survival, stem cell renewal, and tumorigenesis (7)(8)(9)(10)(11)(12). Stat3 is activated by cytokines that bind to the gp130 receptor, including interleukin-6 (IL-6) and oncostatin M (13,14). Activated Stat3 has been associated with several cancers, including head and neck and breast cancer (7,15). In breast cancer, increased levels of IL-6 and activated Stat3 promote not only tumor growth but also metastasis (7,16,17). Recently, we have shown that one possible mechanism by which IL-6 and Stat3 promote metastasis is through transcriptional induction of the fascin gene (18).
Fascin is a highly conserved actin-bundling protein that localizes to microspikes and filopodia with functions in cell adhesion and motility (19). In mammalian cells, three isoforms of fascin exist (20). Fascin-1 (herein referred to as fascin) is expressed during embryonic development and in some normal adult cell types, including dendritic cells (20). Fascin-2 is retinal-specific, and fascin-3 is only found in the testes (21). Fascin expression increases significantly in many cancers, including metastatic gastric, colon, and breast cancer (19,20). Fascin is a key regulator of breast cancer metastasis. Knocking down fascin or inhibition of fascin with small chemical compounds block breast cancer metastasis (22). The mechanism by which fascin is up-regulated in cancer and the signaling pathways that regulate transcription of the fascin gene are not well understood, although the ␤-catenin pathway and the transcription factor cAMP-responsive element-binding protein (CREB) have been shown to be involved in controlling fascin expression in colon cancer (23)(24)(25).
The transcription factor NFB 2 has also been associated with tumorigenesis, apoptosis, inflammation, and cancer cell migration (26,27). NFB activity is controlled through shuttling between the cytoplasm and the nucleus. In response to stimulus by inflammatory factors such as tumor necrosis factor (TNF)-␣, increased levels of NFB accumulate in the nucleus (27)(28)(29). Interactions between Stat3 and NFB have been demonstrated previously (30,31). It has been shown that unphosphorylated Stat3 can be recruited to promoters through its interaction with NFB (30). The fascin promoter contains a 160-bp conserved region that contains overlapping STAT and NFB binding sites (18). We have shown previously that in response to IL-6 or oncostatin M treatment, Stat3 and NFB are recruited to the fascin promoter to increase transcription of the fascin gene and promote migration of breast cancer cells (18). Knocking down Stat3 or NFB blocks IL-6 or oncostatin M-mediated transcriptional expression of fascin. Furthermore, Stat3 is also required for breast cancer cell migration in response to IL-6 or oncostatin M treatment (18).
In this report, we further analyze the role of Stat3 and NFB in transcriptional regulation of the fascin promoter. We demonstrate that in addition to IL-6, TNF-␣ induces Stat3 and NFB binding to the fascin promoter in the MDA-MB-231 human breast cancer cells. NFB is necessary for TNF-␣-induced fascin expression and cell migration. In response to IL-6 treatment, binding of Stat3 to the fascin promoter requires NFB. Furthermore, we show that treatment of human breast cancer cells with IL-6 and TNF-␣ leads to the formation of a complex of Stat3, NFB p50, and NFB p65. Finally, we demonstrate that the 160-bp conserved region in the fascin promoter containing the overlapping STAT and NFB sites is sufficient to activate transcription in response to IL-6 and TNF-␣, and transcription driven by this conserved region is dependent on specific nucleotides shared by both the STAT and NFB binding sites.

EXPERIMENTAL PROCEDURES
Cell Culture, Reagents, and Plasmids-MDA-MB-231 cells were cultured in RPMI 1640 medium (Invitrogen) supplemented with 10% FBS. The Stat3, NFB p50, and p65 antibodies for ChIP and co-immunoprecipitation experiments were from Santa Cruz Biotechnology. Antibodies for Western blot analysis were anti-Stat3 (BD Transduction Laboratories), anti-tubulin (Sigma), anti-NFB p50 and p65, and anti-fascin (Santa Cruz Biotechnology). Ligands were used at the following concentrations: 40 ng/ml TNF-␣ (Millipore) and 20 ng/ml IL-6 combined with 200 ng/ml IL-6 R␣ (R&D Systems). PCR reactions for cloning used Pfu DNA polymerase (Agilent). The plasmid FasP-luc was constructed by amplifying a 160-bp conserved region with the primers 5Ј-ggggatccagaccttgtgggcagcctgt and 5Ј-ggggatccgggacattccctgcagacaccac from ChIP DNA. It was cloned into the BamHI site of pBluescript (Addgene). The fascin promoter conserved region was then PCR-amplified with the primers 5Ј-ggggatccaccttgtgggcagcctgt and 5Ј-atgcgctcgaggggacattccctgcagacaccac and cloned into the BamHI/ XhoI site of the plasmid tkluc.
Cell Migration Assays-Wound-healing assays were performed as described previously (26). Cells transfected with siRNAs were cultured to confluence in RPMI 1640 medium supplemented with 10% FBS in 24-multiwell plates. Wounds were made with a sterile pipette tip, and cells were washed with PBS and then cultured in indicated media for 24 h.
Luciferase Assays-Luciferase assays were performed using the Dual-Luciferase Reporter System (Promega). Luciferase results were represented as relative luciferase units, which are the luciferase activities standardized to Renilla luciferase (Promega).

Both IL-6 and TNF-␣ Induce Fascin mRNA Expression through Stat3 and NFB-
The actin-bundling protein fascin is up-regulated in some cancers and promotes metastasis (19,20). We have previously shown that in response to IL-6, Stat3 and NFB are recruited to the fascin promoter, and fascin expression increases to promote breast cancer cell migration (18). However, the role of TNF-␣ and NFB in fascin transcriptional regulation has not been fully analyzed. To determine whether TNF-␣ induces NFB to bind to the fascin promoter, ChIP analyses were performed in the human breast cancer cell line MDA-MB-231. In response to TNF-␣ treatment, there was an increase in NFB p50 and p65 binding to the fascin promoter compared with untreated cells (Fig. 1A, lanes 1 and 3). TNF-␣ treatment also caused an increase in Stat3 binding to the promoter (Fig. 1A, lanes 1 and 3). There was a stronger signal in NFB p50, p65, and Stat3 bound to the promoter in cells treated with IL-6 and TNF-␣ together (Fig. 1A, lanes 1 and 2).
ChIP results of Stat3 and NFB p50 for IL-6, TNF-␣, and both TNF-␣ and Il-6 together were quantitated for comparison ( Fig. 1B) (18). For Stat3 binding, IL-6 induced a stronger signal than TNF-␣ (Fig. 1B). IL-6 and TNF-␣ together further increased the binding of Stat3 to the fascin promoter (Fig. 1B). For NFB p50, either IL-6 or TNF-␣ alone induced its binding to the fascin promoter; however, IL-6 and TNF-␣ together enhanced p50 binding significantly (Fig. 1B). These ChIP results demonstrate that either IL-6 or TNF-␣ induce Stat3 and NFB binding to the fascin promoter. However, simultaneous treatment of IL-6 and TNF-␣ can induce much stronger binding of these proteins to the promoter, possibly through increased stabilization of the complex.
To see whether these cytokines induced fascin mRNA expression, quantitative real-time RT-PCR analyses were performed for cells treated with TNF-␣. As demonstrated previously (18), IL-6 treatment alone induced fascin expression almost 4-fold after 30 min of treatment. Treatment of MDA-MB-231 cells with TNF-␣ alone induced fascin mRNA expression 3-fold after 30 min of treatment (Fig. 1C). Combined treatment of IL-6 plus TNF-␣ showed a similar increase in fascin expression to treatment with either cytokine alone at the 30-min time point. However, after 2 h of treatment with TNF-␣, either alone or with IL-6, fascin mRNA levels were higher compared with untreated or IL-6 treated cells (Fig. 1C). These results demonstrate that either TNF-␣ or IL-6 can induce expression of fascin through recruitment of Stat3 and NFB. However, TNF-␣ can induce a longer and more sustained transcriptional activation of the fascin gene.
NFkB Is Necessary for Fascin Expression and Binding of Stat3 to the Fascin Promoter in Response to Cytokines-We have shown previously that recruitment of NFB to the fascin promoter in response to IL-6 requires Stat3 (18). To determine whether Stat3 binding to the fascin promoter requires NFB, we performed ChIP assays in NFB siRNA knockdown cells. Two siRNAs targeting separate regions of NFB p50 were described previously (18). NFB p50 was efficiently knocked down in MDA-MB-231 cells compared with cells transfected with a control siRNA ( Fig. 2A). In cells with NFB p50 knockdown, no fascin mRNA was induced by IL-6 or TNF-␣ treatment (data not shown) (18) and fascin protein was not induced by TNF-␣ (Fig. 2B). To see whether there is any Stat3 binding at the fascin promoter in NFB p50 knockdown cells, ChIP analyses were performed in both NFB p50 knockdown and control cells. In response to IL-6, there was a significant increase in Stat3 and NFB p50 binding to the fascin promoter in control cells (Fig. 2C, lanes 1 and 2). In contrast, there was a significant decrease in Stat3 binding to the promoter in NFB p50 knockdown cells (Fig. 2C, lanes 3 and 4). There was no binding of NFB p50 in the knockdown cells (Fig. 2C, lanes 3 and 4).
To further demonstrate that NFB is necessary for fascin induction, NFB p65 was knocked down in MDA-MB-231 cells. In cells transfected with control siRNA, both fascin mRNA and protein were induced by TNF-␣, whereas in cells transfected with NFB p65 siRNA, fascin mRNA and protein were not induced in response to TNF-␣ (Fig. 3, A-C). Altogether, these results demonstrate that NFB is required for fascin expression and Stat3 recruitment to the fascin promoter in response to cytokine treatment.
NFB Is Required for IL-6 and TNF-␣-induced Cell Migration of MDA-MB-231 Cells-To determine whether NFB is also required for cytokine-induced cell migration, two rounds of siRNA knockdown of NFB p50 were performed to maintain the knock-down of fascin protein in the cells (Fig. 4). The cells transfected with control siRNA or NFB p50 siRNA were serum-starved for 10 h followed by wound healing assays. Cells did not migrate in medium without serum (Fig. 4, i and vi). In medium containing serum, both cells transfected with control siRNA or NFB p50 siRNA migrated well (Fig. 4, ii and vii). Cells transfected with control siRNA migrated efficiently in response to treatment with IL-6, TNF-␣, or IL-6 plus TNF-␣ (Fig. 4, iii-v). However, the cells transfected with NFB p50 siRNA did not migrate in any cytokine treated medium (Fig. 4, viii-x). Similar results were obtained with NFB p65 knockdown. These results demonstrate that NFB is necessary for MDA-MB-231 cells to migrate in response to IL-6 and TNF-␣ signaling.
Stat3, NFB p50, and NFB p65 Form a Complex-We have shown that in response to IL-6 and TNF-␣, Stat3 and members of the NFB complex interact with the fascin promoter using ChIP analyses. To further analyze the interactions between Stat3 and NFB in vivo, co-immunoprecipitation experiments were performed on nuclear extracts from cells treated with both IL-6 and TNF-␣. Co-immunoprecipitation experiments with the Stat3 antibody showed that NFB p50 and p65 coimmunoprecipitated with Stat3 (Fig. 5, lane 2). The NFB p50 antibody co-immunoprecipitated both Stat3 and NFB p65 (Fig. 5, lane 3), and the NFB p65 antibody co-immunoprecipitated both Stat3 and NFB p50 (Fig. 5, lane 4). An IgG antibody was used as a negative control (Fig. 5, lane 5). Stat3 and NFB A.   also can be co-immunoprecipitated in untreated MDA-MB-231 cells; however, the interactions were much weaker because NFB was present at lower levels in untreated nuclear extracts (data not shown). Altogether, these results demonstrate that Stat3, NFB p50, and NFB p65 form a protein complex in the nucleus of MDA-MB-231 cells treated with IL-6 and TNF-␣.

A Conserved Region in the Fascin Promoter Containing the STAT/NFB Sites Is Sufficient to Induce Transcription in
Response to IL-6 and TNF-␣-There is an evolutionarily conserved 160-bp region in the fascin promoter that contains overlapping STAT and NFB sites (Fig. 6A) (18). To further analyze the role of this conserved region in transcriptional activation, we subcloned it into a luciferase reporter (FasP-luc) (Fig. 6A). Luciferase assays were performed in MDA-MB-231 cells transiently trans-fected with either the empty luciferase vector (tkluc) as a negative control, M67 (which contains four Stat3-binding sites (35)) as a positive control, and FasP-luc. In response to IL-6 treatment, there was an ϳ8-fold induction of relative luciferase units for M67 compared with untreated cells (Fig. 6B). IL-6 treatment did not cause a significant increase in relative luciferase units for the empty vector tkluc (Fig. 6B). Cells transfected with the FasP-luc contruct showed a ϳ10-fold increase in relative luciferase units compared with untreated cells (Fig. 6B).
Once it was determined that IL-6 could activate the FasP-luc reporter containing the conserved 160-bp region, it was necessary to determine whether TNF-␣ could activate transcription of the FasP-luc reporter. Cells transfected with the empty vector tkluc did not show any transcriptional activation when treated with IL-6, TNF-␣, or both IL-6 and TNF-␣ together (Fig. 6C). In response to TNF-␣ treatment, there was a significant induction of luciferase in cells transfected with the FasP-luc reporter compared with untreated cells (Fig. 6C). Luciferase levels were similar to those induced by IL-6 alone (Fig. 6C). When cells transfected with FasP were treated with both cytokines together, luciferase was induced at levels similar to treatment with either cytokine alone (Fig. 6C). These results demonstrate that a 160-bp conserved region of the fascin promoter is sufficient to activate transcription in response to IL-6 or TNF-␣.
The Overlapping STAT/NFB site in the FasP-luc Contruct Is Necessary for Transcriptional Activation in Response to IL-6 and TNF-␣-To determine whether the overlapping sequences of the STAT/NFB site are required for cytokine-induced tran-  scriptional activation, point mutations of nucleotides important for Stat3 and NFB binding were introduced into the FasPluc contruct (Fig. 7A). Luciferase assays were performed for MDA-MB-231 cells transiently transfected with tkluc, FasPluc, and the constructs containing the indicated point mutations. Cells transfected with the wild-type FasP-luc construct showed a significant increase in relative luciferase units (ϳ8fold) compared with untreated cells (Fig. 7B). In contrast, the constructs containing the point mutations G1086A and G1085A showed an induction of only ϳ2-fold compared with untreated cells (Fig. 7B). The construct containing the point mutation G1084A showed an induction of ϳ4-fold, whereas the AA1082CC mutation did not show significant induction in response to IL-6 ( Fig. 7B). Luciferase assays were performed to determine whether these mutations had an effect in response to TNF-␣ treatment. When the cells were treated with TNF-␣ alone, there was a significant increase in luciferase expression for the FasP-luc construct compared with untreated cells (Fig. 7C). There was no induction in response to TNF-␣ for the G1086A, G1085A, or AA1082CC mutations (Fig. 7C). The G1084A mutation did not significantly affect luciferase expression (Fig. 7C).
Luciferase asssays were performed to determine whether combined treatment with IL-6 plus TNF-␣ would restore expression of any of the mutants. Treatment with both cytokines induced luciferase expression of the FasP-luc contruct to levels similar to treatment with either cytokine alone (Fig. 7D). Treatment with both IL-6 and TNF-␣ together showed some induction with the G1086A and G1085A constructs that was similar to results obtained with IL-6 treatment alone (Fig. 7D). For the G1084A construct, treatment with both cytokines induced luciferase expression levels similar to the wild-type FasP-luc construct (Fig. 7D). Although the AA1082CC construct was not induced with either cytokine alone, there was some induction of luciferase expression with both together (Fig.  7D). These results, taken together, show that the overlapping sequences in the STAT and NFB sites in the 160-bp conserved region of the fascin promoter are required for efficient transcriptional activation by IL-6 and TNF-␣.

DISCUSSION
The cytokine IL-6 has long been associated with inflammation and metastatic cancers including breast cancer (36 -39). IL-6 induces activation of Stat3, which is also associated with  breast cancer metastasis (16,17). We have shown previously that through induction by IL-6, Stat3 along with NFB bind to the fascin promoter to increase its expression and promote breast cancer cell migration (18). In this work, we further explored the roles of Stat3, NFB, and the TNF-␣ signaling pathway in fascin gene regulation. Similar to IL-6/Stat3, TNF-␣/NFB has been linked to inflammation and cancer (28). This study demonstrates that in response to TNF-␣, NFB binds to the fascin promoter and increases its expression. Compared with IL-6 induction of fascin, TNF-␣ induces a longer, more sustained induction of the fascin gene. It is interesting that IL-6 and TNF-␣ treatment together does not lead to an additive effect of both cytokines. Despite an increase in Stat3 and NFB binding to the fascin promoter based on ChIP analyses, treatment with both cytokines induced mRNA expression levels similar to either cytokine alone. It has been shown that MDA-MB-231 cells express and constitutively secrete IL-6 and TNF-␣ (40), which probably results in a high basal level of fascin because there is always a presence of both cytokines in the media. If these cytokines were not secreted into the medium, a more dramatic induction of fascin expression in response to simultaneous treatment of IL-6 and TNF-␣ might be observed. However, further treatment with high levels of cytokines can induce more cell migration, which is dependent on both Stat3 (18) and NFB (Fig. 4), suggesting that in vivo, local inflammation caused by the growth of a tumor may result in high levels of inflammatory cytokines such as IL-6 and TNF-␣ in the tumor micro-environment, which would then lead to tumor cell migration and metastasis.
NFB is required for Stat3 binding to the fascin promoter in response to IL-6 treatment (Fig. 2), and NFB binding requires Stat3 as shown in our previous studies (18). Neither transcription factor can bind to the fascin promoter separately. Co-immunoprecipitation data demonstrates that in response to IL-6 and TNF-␣, Stat3, NFB p50, and NFB p65 form a complex in the nucleus (Fig. 5). This suggests that for either Stat3 or NFB to bind to the fascin promoter, the formation of the Stat3⅐NFB complex is a prerequisite step. Therefore, Stat3 and NFB are co-dependent in their binding to the fascin promoter in response to cytokine treatment.
We had previously identified a 160-bp region of the fascin promoter that is highly conserved in both the human and mouse genes (18). Furthermore, this conserved region contains overlapping STAT and NFB sites. Using a luciferase reporter, we isolated this region and showed that it is sufficient for transcriptional activation in response to IL-6 and TNF-␣. We further demonstrated that the overlapping sequences are required for IL-6 and TNF-␣ induced transcription. It is interesting to note that the STAT site deviates from the consensus at the first three nucleotides (TTC to TGT) (41). It is possible that binding of NFB to the overlapping NFB site serves as a stabilizer for Stat3 binding to the sequence, as well as recruitment of other co-activators in the transcriptional complex.
Cancer metastasis is a multistep process that involves many cellular factors and signaling pathways (42). Fascin has been shown to be a key regulator of breast cancer metastasis (22). Identification of Stat3 and NFB as two important regulators of IL-6 and TNF-␣-induced fascin expression is a major step in understanding the process of metastasis. Further analysis of the interaction of these two transcription factors could potentially lead to identification of drug targets to block breast cancer metastasis.