TGF-β (transforming growth factor-β) induces a cytostatic response in most normal cell types. In cancer cells, however, it often promotes metastasis, and its high expression is correlated with poor prognosis. In the present study, we show that S100A4, a metastasis-associated protein, also called metastatin-1, can physically and functionally interact with Smad3, an important mediator of TGF-β signalling. In agreement with its known property, S100A4 binds to Smad3 in a Ca2+-dependent manner. The S100A4-binding site is located in the N-terminal region of Smad3. S100A4 can potentiate transcriptional activity of Smad3 and the related Smad2. When exogenously expressed in MCF10CA1a.cl1, an MCF10-derived breast cancer cell line, S100A4 increases TGF-β-induced MMP-9 (matrix metalloproteinase-9) expression. On the other hand, depletion of S100A4 by siRNA (small interfering RNA) from the MDA-MB231 cell line results in attenuation of MMP-9 induction by TGF-β. Consistent with these observations, S100A4 increases cell invasion ability induced by TGF-β in MCF10CA1a.cl1 cells, and depletion of the protein in MDA-MB-231 cells inhibits it. Because expression of both S100A4 and TGF-β is highly elevated in many types of malignant tumours, S100A4 and Smad3 may co-operatively increase metastatic activity of some types of cancer cells.

INTRODUCTION

TGF-β (transforming growth factor-β) is a prototypical member of a multifunctional cytokine family that regulates a wide variety of cellular functions [1]. It signals through two types of transmembrane serine/threonine kinase receptors, TβRI and TβRII [2,3]. On ligand binding, the two receptors form a complex, in which constitutively active TβRII phosphorylates and activates TβRI to phosphorylate Smad2 and 3, major intracellular effector proteins for TGF-β signalling. Phosphorylated Smad2/3 then form a complex with Smad4 and translocate into the nucleus. In the nucleus, Smad complexes interact with various factors to regulate gene expression [3]. Thus availability of different types of Smad-binding proteins in different cellular contexts can lead to diverse cell responses to the cytokine.

Growth inhibitory function is among the most important features of TGF-β, and it plays a central role in homoeostasis of normal cells. Accordingly, TGF-β is considered a tumour suppressor. The loss of a cytostatic response to TGF-β is indeed a hallmark of various kinds of cancer [4,5]. Once cancer is established, however, the same cytokine paradoxically favours tumour progression and metastasis through increasing activities of invasion, mitogenesis and angiogenesis [1,6]. Importantly, although genetic alterations of the TGF-β signalling components account for cancer development in some cases [5,7], this change of function often occurs with the main signalling components intact. This indicates that the same TGF-β signalling can result in opposite consequences in normal and cancer cells. Although what switches TGF-β from a tumour suppressor to a tumour promoter is not clear, it is obvious that TGF-β-mediated cell malignancy is regulated by multiple layers of molecular mechanisms [6,8].

Up-regulation of MMPs (matrix metalloproteinases) in cancer cells is a critical step for metastasis. MMPs promote cell invasion by degrading ECM (extracellular matrix) and regulating angiogenesis [9,10]. Among the MMPs, MMP-2 and MMP-9 play major roles in cancer cell metastasis [9]. Importantly, their expression is regulated by TGF-β [9]. In cancer cells, other signalling pathways or factors are integrated in TGF-β signalling for MMP regulation. RTKs (receptor tyrosine kinases) have been reported to co-operate with TGF-β for inducing the oncogenic property of epithelial cells in mouse models [1114]. At the molecular level, EGF (epidermal growth factor) synergizes with TGF-β for induction of some MMPs [12,13]. Moreover, a transcriptional co-activator, Cited {CBP [CREB (cAMP-response-element-binding protein)-binding protein]/p300-interacting transactivator with glutamic acid (E)/aspartic acid (D)-rich C-terminal domain} 2 has been shown to augment TGF-β-mediated up-regulation of MMP-9 and cell invasion in breast cancer cells [15]. This effect is through Smad2 and 3, indicating the pro-metastatic role of these transcription factors in cancer cells, and Cited2 provided an example that these Smads can increase their pro-metastatic activity through interaction with other cofactors.

S100A4, also known as Mts-1 (metastatin-1), is a member of the S100 family calcium-binding proteins [16]. It is highly expressed in various metastatic tumour cells and its expression correlates with poor prognosis in different types of cancer [1720]. For this reason, it is recognized as a good molecular marker for clinical prognosis [21,22]. Studies with mouse models have shown that S100A4 itself is not tumorigenic, but once a tumour is formed it enhances tumour invasiveness and metastasis [23,24]. At the molecular level, some protein targets of S100A4 have been identified. These include a tumour suppressor, p53, and non-muscle myosin. The binding of S100A4 on the C-terminal region of p53 interferes with its DNA-binding activity and transcriptional activity [25]. As a consequence, this interaction may indirectly contribute to the acquisition of an invasive phenotype [25,26]. Binding of S100A4 to non-muscle myosin IIA stabilizes its monomeric state and destabilizes myosin filaments, leading to increased cell motility [27,28]. The increased cell motility may contribute to metastasis. Thus, although the molecular mechanisms by which S100A4 induces a metastatic phenotype may not be fully understood, modulation of the function of its targets appears to be an important part of the role of S100A4.

In the present study, we have identified Smad3 as a new S100A4 target. Interaction of S100A4 with Smad3 augments TGF-β-mediated up-regulation of MMP-9, leading to an increase in cell invasion ability. The findings of the present study provide a new molecular mechanism for the regulation of metastatic ability by both TGF-β and S100A4.

EXPERIMENTAL

Constructs

Mammalian expression plasmids for 6×Myc–Smad3, Myc–Smad2 and HA (haemagglutinin)–TβRI have been described previously [29]. pCMV5-3×FLAG–p53 and c-Jun cDNA was kindly provided by Dr S. Hsu and Dr Y. Chen respectively [both at the NHRI (National Health Research Institutes, Taiwan)] The bacterial expression plasmid for the His-tag protein pET15b-S100A4 was a gift from Dr M. Takahashi (Division of Chemistry, Graduate School and Faculty of Science, Hokkaido University, Sapparo, Japan). pGEX4T-S100A4 was constructed by subcloning the S100A4 fragment from pET15b-S100A4. CS2-S100A4 was made by PCR-based subcloning. pGEX2T-Smad3 and 2, GAL4–Luc and GAL4–Smad3 were as described previously [29]. VP16–S100A4 and VP16–c-Jun were constructed by replacing the SgfI–PmeI barnase fragment with each cDNA in pFN10A (ACT) Flexi (Promega). pBABE-S100A4 was constructed by subcloning the S100A4 fragment into pBABE puro (Addgene).

Cell culture

MCF10CA1a.cl1 cells, a carcinoma cell line derived from the MCF10A breast epithelial cell line [30,31], were maintained in DMEM (Dulbecco's modified Eagle's medium)/F12 supplemented with 5% horse serum and 1% penicillin/streptomycin. HEK (human embryonic kidney)-293T cells were cultured in MEM (minimum essential medium), 10% FBS (fetal bovine serum) and 1% penicillin/streptomycin. Amphotrophic Phoenix cells (for retrovirus packaging) [32] and MDA-MB 231 were maintained in DMEM containing 10% FBS and 1% penicillin/streptomycin. Lipofectamine™ (Invitrogen) was used for transfection following the manufacturer's instructions.

Retroviral infection using Phoenix cells has been described previously [29]. MCF10CA1a.cl1 cells transduced with pBABE vector or pBABE-S100A4 were selected with 5 μg/ml puromycin.

siRNA (small interfering RNA) for S100A4 was purchased from Invitrogen. MDA-MB 231 cells were transfected with siRNA using Lipofectamine™ RNAi Max (Invitrogen) for 48 h.

GST (glutathione transferase) pulldown assay and co-immunoprecipitation

GST proteins were expressed in bacterial BL21(DE3) cells. Protein on the GST–Sepharose beads was quantified and adjusted so that an equal amount of protein was on the beads (Figure 1A, right-hand panel).

S100A4 can interact with Smad3 in vitro

Figure 1
S100A4 can interact with Smad3 in vitro

(A) Smad3 and Smad2 bind to S100A4 in a Ca2+-dependent manner in a GST-pulldown assay. HEK-293T cells were transfected with 6×Myc-tagged Smad2, Smad3 or 3×FLAG-tagged p53. The cell lysate was subject to a GST-pulldown assay with GST–S100A4, or GST as a control. EGTA or CaCl2 to 1 mM was included throughout the procedure. Smad proteins and p53 recovered on the beads were detected by Western blot analysis with anti-Myc (left-hand panel) and anti-FLAG (middle panel) antibodies respectively. For input lanes, 2% of the lysate used for the precipitation was applied. Right-hand panel: the amounts and purity of GST and GST–S100A4 on the beads were examined. The molecular mass in kDa is indicated on the right-hand side of the gel. (B) Smad3 and S100A4 interact at the endogenous protein level. Cell lysates of MDA-MB-231 cells (with or without TGF-β treatment) were immunoprecipitated (IP) with an anti-Smad3 antibody (left-hand panel) or anti-S100A4 antibody (right-hand panel). Precipitated proteins were detected by Western blot analysis (WB). Control IgG was used as a negative control (con). CaCl2 was included at 1 mM throughout the procedure in all samples except for the EGTA sample. Co-IP, co-immunoprecipitation. (C) Smad3 binds to S100A4 in a BIAcore assay. His–S100A4 was immobilized on the BIAcore sensor chip. The same amount of GST or GST–Smad3 (40 μl of 174 nM) was applied to the BIAcore instrument with buffer containing 1 mM CaCl2. Sensorgrams are shown for GST and GST–Smad3. The curve of GST returns to baseline (no binding), whereas that of GST–Smad3 stays above the baseline (binding). (D) Smad3 binds to S100A4 in a Ca2+-sensitive manner in a BIAcore assay. GST–Smad3 binding to immobilized S100A4 was examined with buffer containing 1 mM CaCl2 or 1 mM EGTA. RU, resonance units.

Figure 1
S100A4 can interact with Smad3 in vitro

(A) Smad3 and Smad2 bind to S100A4 in a Ca2+-dependent manner in a GST-pulldown assay. HEK-293T cells were transfected with 6×Myc-tagged Smad2, Smad3 or 3×FLAG-tagged p53. The cell lysate was subject to a GST-pulldown assay with GST–S100A4, or GST as a control. EGTA or CaCl2 to 1 mM was included throughout the procedure. Smad proteins and p53 recovered on the beads were detected by Western blot analysis with anti-Myc (left-hand panel) and anti-FLAG (middle panel) antibodies respectively. For input lanes, 2% of the lysate used for the precipitation was applied. Right-hand panel: the amounts and purity of GST and GST–S100A4 on the beads were examined. The molecular mass in kDa is indicated on the right-hand side of the gel. (B) Smad3 and S100A4 interact at the endogenous protein level. Cell lysates of MDA-MB-231 cells (with or without TGF-β treatment) were immunoprecipitated (IP) with an anti-Smad3 antibody (left-hand panel) or anti-S100A4 antibody (right-hand panel). Precipitated proteins were detected by Western blot analysis (WB). Control IgG was used as a negative control (con). CaCl2 was included at 1 mM throughout the procedure in all samples except for the EGTA sample. Co-IP, co-immunoprecipitation. (C) Smad3 binds to S100A4 in a BIAcore assay. His–S100A4 was immobilized on the BIAcore sensor chip. The same amount of GST or GST–Smad3 (40 μl of 174 nM) was applied to the BIAcore instrument with buffer containing 1 mM CaCl2. Sensorgrams are shown for GST and GST–Smad3. The curve of GST returns to baseline (no binding), whereas that of GST–Smad3 stays above the baseline (binding). (D) Smad3 binds to S100A4 in a Ca2+-sensitive manner in a BIAcore assay. GST–Smad3 binding to immobilized S100A4 was examined with buffer containing 1 mM CaCl2 or 1 mM EGTA. RU, resonance units.

HEK-293T cells were transfected with Myc–Smad3, Myc–Smad2 or FLAG–p53, and were treated or untreated with 500 pM TGF-β (Peprotech) for 1 h. Cell lysate was prepared in TNTE buffer [10 mM Tris/HCl (pH 7.8), 150 mM NaCl and 1% Nonidet P40], 10 mM sodium pyrophosphate, 25 mM NaF, 10 mM β-glycerophosphate, 1 mM DTT (dithiothreitol), 5 μg/ml RNase A and 1×Complete protease inhibitor cocktail (Roche). GST or GST–S100A4 beads were incubated with 300 μg of lysate for 2 h in the presence of either 1 mM EGTA or CaCl2. Beads were washed with the same buffer. Proteins remaining on the beads were visualized by Western blot analysis with an anti-Myc or anti-FLAG antibody (Sigma).

For co-immunoprecipitation of endogenous Smad3 and S100A4, MDA-MB-231 cells were treated and lysed as described above. Lysate (750 μg) was pre-cleared with Protein A/G beads (Pierce) and mixed with a mouse monoclonal anti-Smad3 antibody (4D5, used at 1:100; Abnova) or 3 μg of a rabbit polyclonal anti-S100A4 antibody (Dako Cytomation) and incubated at 4 °C overnight. Protein A/G beads were added and incubated for a further 2 h. The beads were washed, and eluted samples were subjected to Western blot analysis. The anti-S100A4 antibody described above was used for S100A4 detection. The mouse monoclonal antibody 4D5 was used to detect Smad3 following the precipitation by the anti-S100A4 antibody, whereas a rabbit monoclonal anti-Smad3 antibody (C67H9; Cell Signaling) was used for the detection of precipitated protein by 4D5.

BIAcore interaction analysis

A BIAcore assay was carried out following the manufacturer's instructions (Biosensor). His–S100A4 (17.5 ng≈1.75 pmol) was immobilized on a BIAcore Sensor Chip CM5 using an amine coupling kit. The chip was equilibrated with running buffer [50 mM Hepes (pH 7.2) and 100 mM NaCl] either in the presence of Ca2+ (1 mM CaCl2) or its absence (1 mM EGTA). An appropriate amount of GST proteins in 40 μl were injected into the BIAcore instrument. Sensorgrams (time against relative resonance units, where 1 resonance unit is equivalent to l pg of protein/mm2 on the sensor surface) were recorded and analysed using the accompanying software, BIAevaluation. For calculation of kinetic parameters, sensorgrams using a BSA-coated chip were subtracted from those using His–S100A4.

Mammalian two-hybrid assay and Smad-responsive reporter assay

For the mammalian two-hybrid assay, HEK-293T cells on a 12-well plate were transfected with pRLTK (2 ng), HA–TβRI (12.5 ng), GAL4–Luc (100 ng) and GAL4—Smad3 (80 ng), along with 80 ng of pFN10A vector, VP16–S100A4 or VP16–c-Jun. The barnase fragment had been deleted from the pFN10A vector because of its toxicity. At 24 h post-transfection, cells were treated or left untreated with 500 pM TGF-β for 20 h and subjected to a luciferase assay using the dual reporter assay system (Promega). The Smad-responsive reporter gene assay has been described previously [29].

RT (reverse transcription)-qPCR (quantitative PCR)

Retrovirus-infected MCF10CA1a.cl1 cells in low-serum medium (1% horse serum) were treated with 500 pM TGF-β or left untreated for 7 h. Total RNA was prepared and primary cDNA was synthesized using the high-capacity RT kit from Applied Biosystems. qPCR was carried out using the Taqman gene expression assays kit (Applied Biosystems) with the ABIPrism 7700 system. β-Actin was used as an internal reference. For MDA-MB 231 cells, siRNA-transfected cells were subjected to the procedure above, except that TGF-β treatment was for 8 h.

MMP-9 zymography and Western blot analysis of conditioned medium

Retrovirus-infected MCF10CA1a.cl1 cells or siRNA-transfected MDA-MB 231 cells in 60 mm plates were treated with TGF-β in 2 ml of low-serum medium for 24 h. Non-reducing SDS sample buffer was added to the conditioned medium. After a 10 min incubation at room temperature (25 °C), an appropriate amount of the sample was separated by SDS/PAGE (7.5% gels containing 1 mg/ml gelatin). The gel was washed for 20 min (or longer) successively with Buffer 1 (3 mM sodium azide and 2.5% Triton X-100), Buffer 2 [3 mM sodium azide, 2.5% Triton X-100 and 50 mM Tris/HCl (pH 7.5)], Buffer 3 [3 mM sodium azide, 2.5% Triton X-100, 50 mM Tris/HCl (pH 7.5), 5 mM CaCl2 and 1 μM ZnCl2] and Buffer 4 [3 mM sodium azide and 50 mM Tris/HCl (pH 7.5)]. The gel was then incubated in fresh Buffer 4 for 5–6 h at 37 °C. The gel was Coomassie-Blue-stained for negative-stained bands that represent MMP-9 activity.

For MMP-9 Western blot analysis, conditioned medium was concentrated 10-fold and SDS sample buffer was added to 1× (reducing buffer). The sample was subjected to Western blot analysis for MMP-9 protein.

Cell invasion assay

Retrovirus-infected MCF10CA1a.cl1 cells were treated or left untreated with 500 pM TGF-β for 24 h in DMEM/F12 containing 1% horse serum. The cells were trypsinized and resuspended in DMEM/F12 and 0.1% BSA at 8×105 cells/ml. TGF-β (at a concentration of 500 pM) was kept included for those cells that had been treated with the cytokine. A 250 μl aliquot of the suspension was applied into the inner chamber of a cell invasion plate assembly in a QCM 24-well cell invasion assay kit (Chemicon International). The lower chamber contained 500 μl of the medium with 5% horse serum, with or without TGF-β. The cells were incubated at 37 °C for 30 h. Invaded cells were fluorimetrically detected according to the manufacturer's instruction.

siRNA-transfected MDA-MB-231cells were treated as described above. Cell suspension (2×105 cells/300 μl of DMEM and 0.1% BSA with or without 500 pM TGF-β) was applied to the Millicell hanging cell culture insert (8 μm pore size, Millipore) that had been pre-coated with Matrigel™ (BD Biosciences). The inserts were placed in the 24-well plate with DMEM and 10% FBS with or without 500 pM TGF-β, and incubated for 24 h. Unmigrated cells on the upper side were removed using a cotton swab and cells that had migrated to the other side of the membrane were stained with 0.1% Crystal Violet, 20% ethanol and 1% formaldehyde. The number of migrated cells was counted for quantification.

RESULTS AND DISCUSSION

Smad3 and Smad2 can physically interact with S100A4

A possible explanation for the conflicting dual role of TGF-β signalling in cancer progression is that the signal transduction pathway functionally has two arms (tumour suppressive and promoting) [6]. In normal cells, the tumour-suppressive arm dominates the other so the net response is growth inhibition. In advanced cancer cells, the components for the tumour-suppressive arm are selectively inactivated, so the tumour-promoting arm can take the advantage [6]. However, the components in the tumour-promoting arm equipped in normal cells may not be sufficient to drive cancer cells into full malignancy. From this idea, we hypothesized that cancer cells may utilize additional components whose expression is elevated in cancer cells to modulate the TGF-β signal for more aggressive behaviour.

MAPKs (mitogen-activated protein kinases) and Akt, which are highly activated in cancer cells, together with TGF-β mediate the EMT (epithelial–mesenchymal transition), a prerequisite step for the rest of the metastatic behaviour of cells [11,12,33]. In cancer cells, the canonical TGF-β/Smad pathway also mediates pro-metastatic signals [3436]. On the basis of the hypothesis above, we sought proteins that are highly expressed in tumours and could interact with Smad protein(s). One of the candidates for such proteins is S100A4, a member of a large family of Ca2+-binding proteins (the S100 family). Interestingly, Smads have been reported to interact with calmodulin, another type of Ca2+-binding protein [37,38]. Since some Ca2+-binding proteins can share a similar binding target motif [39,40], we tested whether Smad proteins and S100A4 could physically interact.

The left-hand panel of Figure 1(A) shows a result from a GST-pulldown assay. Myc-tagged Smad2 or Smad3 was expressed in HEK-293T cells and precipitated by GST–S100A4-bound beads (Figure 1A, right-hand panel). When CaCl2 was included in the buffer, these proteins were recovered on the beads. However, when Ca2+ was depleted (using EGTA), this binding was abolished. Beads with GST alone (Figure 1A, right-hand panel) failed to precipitate Smad proteins, even in the presence of CaCl2, indicating that the binding is mediated by S100A4. In this assay, TGF-β treatment of cells did not significantly affect the interaction. One of the known targets of S100A4 is p53 [25]. Thus FLAG-tagged p53 was included in the assay as a positive control (Figure 1A, middle panel). Like Smads, p53 had Ca2+-dependent binding to GST–S100A4, and the assay result was similar to those with Smads overall.

Because of the low-affinity nature of the interaction (see below and [41]), immunoprecipitation by the Smad3 antibody from MDA-MB-231 cells yielded a low amount of endogenous S100A4 (Figure 1B, left-hand panel). However, the co-immunoprecipitation of S100A4 is specific to the Smad3 antibody, as control IgG failed to precipitate the protein. Moreover, the co-immunoprecipitation of S100A4 by the Smad3 antibody is also Ca2+-dependent, consistent with the GST-pulldown assay (Figure 1A). Similar results were obtained in the reciprocal co-immunoprecipitation by the S100A4 antibody (Figure 1B, right-hand panel).

To analyse in vitro Smad3–S100A4 binding in more detail, we performed a BIAcore assay, which can measure low-affinity interactions more quantitatively. His–S100A4 was immobilized on a biosensor chip and GST proteins were applied to the system (Figure 1C). Although GST protein did not show binding to the immobilized S100A4, GST–Smad3 showed relatively weak, but nonetheless significant, binding. In the BIAcore assay, the binding was also Ca2+-dependent (Figure 1D). The kinetic parameters for the binding were determined [Kd, (2.1±0.53)×10−7 M; kd, (1.18±0.17)×10−3 s−1; and ka, (5.81±1.15)×103 M−1·s−1; values are the means±S.D. for six measurements, four to six different concentrations of the sample were applied in each measurement] and Kd was found to be in the submicromolar range (~ 0.21 μM). The submicromolar range of Kd explained the relatively low recovery in the GST-pulldown assay and co-immunoprecipitation assay; but these assays and the BIAcore assay together established the in vitro interaction between Smad3 and S100A4.

Smad3 can interact with S100A4 in the cell

We next examined, using a mammalian two-hybrid assay, whether Smad3 and S100A4 could interact in the cell (Figure 2). GAL4–Smad3 was co-transfected with VP16, VP16–S100A4 or VP16–c-Jun along with a GAL4–Luc construct. The luciferase activity was significantly higher with VP16–S100A4 than with VP16 alone, both in the absence and in the presence of TGF-β. We included c-Jun as it had been reported to interact with Smad3 using the same type of assay [42]. VP16–c-Jun showed a similar result to that of VP16–S100A4, except that the interaction seemed stronger. For both proteins, binding was increased by TGF-β treatment. This suggested that, like c-Jun [43], the S100A4 interaction with Smad3 might be regulated by TGF-β in the cell.

S100A4 interacts with Smad3 in the cell

Figure 2
S100A4 interacts with Smad3 in the cell

HEK-293T cells were transfected with pRLTK, HA–TβRI, GAL4–Luc and GAL4–Smad3, along with pFN10A vector, VP16–S100A4 or VP16–c-Jun. Cells were treated or not with TGF-β for 20 h and the luciferase activity was measured. The luciferase activity shown is the mean±S.D. for three independent experiments.

Figure 2
S100A4 interacts with Smad3 in the cell

HEK-293T cells were transfected with pRLTK, HA–TβRI, GAL4–Luc and GAL4–Smad3, along with pFN10A vector, VP16–S100A4 or VP16–c-Jun. Cells were treated or not with TGF-β for 20 h and the luciferase activity was measured. The luciferase activity shown is the mean±S.D. for three independent experiments.

S100A4 binds to the N-terminal region of Smad3

There are two calmodulin-binding sites in Smad2 and they are located in its N-terminal domain (MH-1 domain) [37,38]. When the entire MH-1 domain was deleted from GST–Smad3 (Δ142; Figure 3A), the binding was abolished (Figure 3B). This result indicated that the S100A4-binding site(s) in Smad3 was also in its MH-1 domain. In Smad2, one calmodulin-binding site is at its N-terminus and the other is located in amino acids 117–124 (equivalent to amino acids 86–93 of Smad3) [38]. On the basis of this information, we deleted the first nine and 93 amino acids of GST–Smad3 (Δ9 and Δ93; Figure 3A). Deletion of the N-terminal nine amino acids impaired the binding, but significant binding still remained (Figure 3B). Further deletion to 93 amino acids almost abolished the binding (Figure 3B). This result suggested that the S100A4-binding region in Smad3 might overlap with that for calmodulin [38]. Consistent with this, the MH-1 domain alone (δLC) bound to S100A4. The binding of Smad3δLC is even stronger than that of the parental protein (Figure 3B), suggesting that another region of Smad3 may be inhibitory to the binding. Interestingly, Thr8, a threonine residue in the region, is a phosphorylation site for CDKs (cyclin-dependent kinases) and MAPKs [29,44,45]. We are currently investigating whether the phosphorylation of Thr8 has any role in the interaction.

S100A4 binds to the N-terminal region of Smad3

Figure 3
S100A4 binds to the N-terminal region of Smad3

(A) Smad3 deletion mutants. Left-hand panel: schematic diagram of Smad3 deletion mutants. Three major domains of Smad3, MH-1, Linker and MH-2, are marked. The size of the domains is not necessarily proportional to their actual size. Thr8 (T8), the phosphorylation site by MAPKs and CDKs at the N-terminus, is indicated. The name of each mutant and its region is shown. Right-hand panel, GST-fusion proteins used in the present study were examined by SDS/PAGE for their amount and purity. The molecular mass in kDa is indicated on the right-hand side of the gel. (B) BIAcore assay for Smad3 deletion mutants. The same amount (40 μl of 150 nM) of GST–Smad3 (WT, wild-type) and its deletion mutants, Δ9, Δ93, Δ142 and ΔLC, were applied to the BIAcore instrument as described in Figure 1(B) in the presence of 1 mM CaCl2. The sensorgram of each mutant is shown as indicated. RU, resonance units.

Figure 3
S100A4 binds to the N-terminal region of Smad3

(A) Smad3 deletion mutants. Left-hand panel: schematic diagram of Smad3 deletion mutants. Three major domains of Smad3, MH-1, Linker and MH-2, are marked. The size of the domains is not necessarily proportional to their actual size. Thr8 (T8), the phosphorylation site by MAPKs and CDKs at the N-terminus, is indicated. The name of each mutant and its region is shown. Right-hand panel, GST-fusion proteins used in the present study were examined by SDS/PAGE for their amount and purity. The molecular mass in kDa is indicated on the right-hand side of the gel. (B) BIAcore assay for Smad3 deletion mutants. The same amount (40 μl of 150 nM) of GST–Smad3 (WT, wild-type) and its deletion mutants, Δ9, Δ93, Δ142 and ΔLC, were applied to the BIAcore instrument as described in Figure 1(B) in the presence of 1 mM CaCl2. The sensorgram of each mutant is shown as indicated. RU, resonance units.

S100A4 can increase Smad transcriptional activity

The physical interaction between Smad3 and S100A4 suggested its functional and physiological significance. To evaluate the effect of S100A4 on Smad function, we performed a reporter assay with two luciferase constructs with a Smad-responsive promoter, synthetic 4×SBE–Luc and A3-Luc/FAST-1 [44]. 4×SBE–Luc was combined with Smad3, and A3-Luc/FAST-1 combined with Smad2. In both cases, inclusion of S100A4 significantly increased luciferase expression compared with the control (Figure 4). This result suggested that S100A4 could potentially increase Smad transcriptional activity.

S100A4 can increase transcription of Smad-responsive promoters

Figure 4
S100A4 can increase transcription of Smad-responsive promoters

(A) S100A4 increases Smad2-mediated transcription. HEK-293T cells were co-transfected with pRLTK, A3–Luc, Myc–Smad2 or FAST-1, along with vector or S100A4. Cells were treated or not with TGF-β for 24 h before being harvested for the luciferase assay. Values are means±S.D. for four independent assays. (B) S100A4 increases Smad3-mediated transcription. HEK-293T cells were co-transfected with pRLTK, 4×SBE–Luc or Myc–Smad3, along with vector or S100A4. Cells were treated with TGF-β for 24 h before being harvested for the luciferase assay. Values are means±S.D. for four independent assays.

Figure 4
S100A4 can increase transcription of Smad-responsive promoters

(A) S100A4 increases Smad2-mediated transcription. HEK-293T cells were co-transfected with pRLTK, A3–Luc, Myc–Smad2 or FAST-1, along with vector or S100A4. Cells were treated or not with TGF-β for 24 h before being harvested for the luciferase assay. Values are means±S.D. for four independent assays. (B) S100A4 increases Smad3-mediated transcription. HEK-293T cells were co-transfected with pRLTK, 4×SBE–Luc or Myc–Smad3, along with vector or S100A4. Cells were treated with TGF-β for 24 h before being harvested for the luciferase assay. Values are means±S.D. for four independent assays.

S100A4 co-operates with the TGF-β signal for MMP-9 expression

In advanced cancer cells, Smad2 and Smad3 function as pro-metastatic factors [35]. From the results described above, we hypothesized that the S100A4–Smad interaction in cancer cells might stimulate the pro-metastatic potential of the Smads. MMP-9 plays important roles in metastasis and the TGF-β/Smad pathway regulates its expression [9,15]. On the basis of this hypothesis, we tested the effect of S100A4 on MMP-9 expression. MCF10CA1a.cl1 is a high-grade carcinoma cell line derived from a MCF10A1 breast epithelial cell line [30]. The contribution of the TGF-β/Smad2/3 pathway to the metastatic ability of this cell line has been demonstrated [31,34]. Moreover, levels of S100A4 are low in this cell line (Figure 5A, vector lane). From these aspects, this cell line was ideal for examining the effect of S100A4 on the TGF-β/Smad-mediated metastatic response by exogenously expressing the protein.

S100A4 augments MMP-9 up-regulation by TGF-β in breast cancer cell lines

Figure 5
S100A4 augments MMP-9 up-regulation by TGF-β in breast cancer cell lines

(A) Exogenous expression of S100A4 in MCF10CA1a.cl1 cells. MCF10CA1a.cl1 cells were transduced with pBABE-S100A4. Post-puromycin selection, expression levels of S100A4 were examined by Western blot analysis (upper panel). The actin level was also examined (lower panel). (B) TGF-β-induced MMP-9 gene up-regulation is augmented by S100A4. Cells from (A) were cultured in medium with 1% horse serum overnight before treatment with TGF-β for 7 h. Total RNA was prepared and subjected to qPCR for the MMP-9 gene. Values are means±S.D. for four independent experiments. (C) TGF-β-induced MMP-9 activity/protein is augmented by S100A4. Cells from (A) were cultured in medium with 1% horse serum overnight before treatment with TGF-β for 24 h. Conditioned medium was applied to gelatin-containing SDS/PAGE under non-reducing conditions or regular SDS/PAGE (see the Experimental section). MMP-9 activity in the medium was visualized by negative Coomassie-Blue-staining (upper panel). MMP-9 protein was detected by Western blot analysis (lower panel). (D) S100A4 does not affect the level of Cip1 (p21) gene. The same total RNA as in (B) was examined by qPCR for the Cip1 gene. Values are means±S.D. for four independent experiments. (E) Depletion of S100A4 from MDA-MB231 cells. MDA-MB231 cells were transfected with either control or S100A4 siRNA. Endogenous S100A4 (upper panel) and actin (lower panel) levels were examined by Western blot analysis. (F) TGF-β-induced MMP-9 gene up-regulation is compromised in S100A4-depleted MDA-MB231 cells. siRNA-transfected cells were cultured in low serum medium (1% FBS) for 15 h. Cells were then treated or not with 500 pM TGF-β for 8 h. Total RNA was prepared and subjected to qPCR for the MMP-9 gene. Values are means±S.D. for four independent experiments. (G) TGF-β-induced MMP-9 activity/protein is reduced in S100A4-depleted MDA-MB231 cells. siRNA-transfected cells were cultured in medium with 1% FBS overnight before being treated or not with TGF-β for 24 h. Conditioned medium was applied to gelatin-containing polyacrylamide gels and separated by SDS/PAGE under non-reducing conditions. MMP-9 activity in the medium was detected by negative Coomassie-Blue-staining (upper panel). MMP-9 protein was detected by Western blot analysis (lower panel)

Figure 5
S100A4 augments MMP-9 up-regulation by TGF-β in breast cancer cell lines

(A) Exogenous expression of S100A4 in MCF10CA1a.cl1 cells. MCF10CA1a.cl1 cells were transduced with pBABE-S100A4. Post-puromycin selection, expression levels of S100A4 were examined by Western blot analysis (upper panel). The actin level was also examined (lower panel). (B) TGF-β-induced MMP-9 gene up-regulation is augmented by S100A4. Cells from (A) were cultured in medium with 1% horse serum overnight before treatment with TGF-β for 7 h. Total RNA was prepared and subjected to qPCR for the MMP-9 gene. Values are means±S.D. for four independent experiments. (C) TGF-β-induced MMP-9 activity/protein is augmented by S100A4. Cells from (A) were cultured in medium with 1% horse serum overnight before treatment with TGF-β for 24 h. Conditioned medium was applied to gelatin-containing SDS/PAGE under non-reducing conditions or regular SDS/PAGE (see the Experimental section). MMP-9 activity in the medium was visualized by negative Coomassie-Blue-staining (upper panel). MMP-9 protein was detected by Western blot analysis (lower panel). (D) S100A4 does not affect the level of Cip1 (p21) gene. The same total RNA as in (B) was examined by qPCR for the Cip1 gene. Values are means±S.D. for four independent experiments. (E) Depletion of S100A4 from MDA-MB231 cells. MDA-MB231 cells were transfected with either control or S100A4 siRNA. Endogenous S100A4 (upper panel) and actin (lower panel) levels were examined by Western blot analysis. (F) TGF-β-induced MMP-9 gene up-regulation is compromised in S100A4-depleted MDA-MB231 cells. siRNA-transfected cells were cultured in low serum medium (1% FBS) for 15 h. Cells were then treated or not with 500 pM TGF-β for 8 h. Total RNA was prepared and subjected to qPCR for the MMP-9 gene. Values are means±S.D. for four independent experiments. (G) TGF-β-induced MMP-9 activity/protein is reduced in S100A4-depleted MDA-MB231 cells. siRNA-transfected cells were cultured in medium with 1% FBS overnight before being treated or not with TGF-β for 24 h. Conditioned medium was applied to gelatin-containing polyacrylamide gels and separated by SDS/PAGE under non-reducing conditions. MMP-9 activity in the medium was detected by negative Coomassie-Blue-staining (upper panel). MMP-9 protein was detected by Western blot analysis (lower panel)

Figure 5(A) shows the expression of S100A4 in MCF10CA1a.cl1 cells retrovirally transduced with vector or the S100A4 expression construct. The level of expression was comparable with that of MDA-MB 231 cells, a highly metastatic cell line that overexpresses S100A4 (results not shown).

The cells were treated with TGF-β and levels of MMP-9 mRNA were quantified by RT–qPCR (Figure 5B). MMP-9 mRNA was up-regulated by TGF-β treatment. S100A4 itself does not significantly increase the mRNA level. However, when S100A4-expressing cells were treated with TGF-β, the mRNA level was ~ 3-fold higher than that of controls. Consistently, zymography and Western blot analysis also showed a much higher level of secreted MMP-9 from TGF-β-treated S100A4-expressing cells (Figure 5C). These results suggested a positive effect of the Smad–S100A4 interaction on MMP-9 gene expression. Cip1 (p21) is another TGF-β-responsive gene. However, in the MCF10CA1a.cl1 cell line, this gene does not respond to TGF-β. In this circumstance (i.e. when a gene had lost TGF-β responsiveness), S100A4 had no effect (Figure 5D). To further strengthen the observations above, we depleted endogenous S100A4 from MDA-MB231 cells by siRNA (Figure 5E). Control and S100A4-depleted cells were treated with TGF-β, and MMP-9 levels were measured by RT-qPCR, zymography and Western blot analysis (Figures 5F and 5G). The results showed that when S100A4 is knocked down, induction of MMP-9 mRNA, activity and protein by TGF-β was greatly attenuated.

Taken together, these results suggested the involvement of S100A4 in TGF-β-stimulated MMP-9 regulation.

S100A4 can increase TGF-β-stimulated cell invasion

From the effect of S100A4 on MMP-9 up-regulation by TGF-β, we anticipated that S100A4 could have an influence on the TGF-β-stimulated invasion ability of the cancer cells. To quantitatively analyse the invasion ability, we performed an in vitro invasion assay (Figure 6A). Invasion of MCF10CA1a.cl1 cells with the control vector was stimulated by TGF-β treatment. This stimulation is mediated by the Smad pathway [34]. S100A4-overexpressing cells showed a higher invasion ability compared with the controls. When S100A4-overexpressing cells were treated with TGF-β, invasion was potentiated to a level much higher than either of S100A4-overexpressing cells alone or controls.

S100A4 enhances TGF-β-stimulated invasion of breast cancer cells

Figure 6
S100A4 enhances TGF-β-stimulated invasion of breast cancer cells

(A) S100A4 enhances TGF-β-stimulated invasion of MCF10CA1a.cl1 cells. Cells transduced by vector or S100A4 were treated with TGF-β for 30 h in medium with 1% horse serum. Cell invasion was fluorimetrically measured by using the QCM 24-well cell invasion assay kit (Chemicon). Relative cell invasion is expressed in RFUs (relative fluorescent units). Values are means±S.D. for three independent experiments. (B) Depletion of S100A4 from MDA-MB-231 cells attenuates TGF-β-induced invasion. Cells transduced by vector or S100A4 were treated with TGF-β for 30 h in medium with 1% FBS. Cell invasion was measured by using Matrigel™-coated cell culture inserts. Invaded cells were stained and counted. Values are means±S.D. for three independent experiments (right-hand panel). The left-hand panel shows cell staining from a representative assay.

Figure 6
S100A4 enhances TGF-β-stimulated invasion of breast cancer cells

(A) S100A4 enhances TGF-β-stimulated invasion of MCF10CA1a.cl1 cells. Cells transduced by vector or S100A4 were treated with TGF-β for 30 h in medium with 1% horse serum. Cell invasion was fluorimetrically measured by using the QCM 24-well cell invasion assay kit (Chemicon). Relative cell invasion is expressed in RFUs (relative fluorescent units). Values are means±S.D. for three independent experiments. (B) Depletion of S100A4 from MDA-MB-231 cells attenuates TGF-β-induced invasion. Cells transduced by vector or S100A4 were treated with TGF-β for 30 h in medium with 1% FBS. Cell invasion was measured by using Matrigel™-coated cell culture inserts. Invaded cells were stained and counted. Values are means±S.D. for three independent experiments (right-hand panel). The left-hand panel shows cell staining from a representative assay.

We also examined the effect of S100A4 depletion on TGF-β-stimulated cell invasion in MDA-MB-231 cells. This cell line shows strong induction of invasion by TGF-β. In this assay, we stained invaded cells for visualization (Figure 6B, left-hand panel). As anticipated, depletion of S100A4 resulted in attenuation of the invasion ability of the cells (Figure 6B). This result suggested that S100A4, through an interaction with Smad proteins, could synergistically increase the cancer cell invasion.

In the present study, we have demonstrated a physical interaction of the two proteins in four different systems (Figures 1 and 2). The affinity of the binding was relatively low (Kd~ 0.21 μM). A low affinity to non-muscle myosin, another target of S100A4, has been reported [41]. In a situation where S100A4 is highly overexpressed (i.e. in some cancer cells), considerable interaction between S100A4 and the target proteins would still occur, even with the low affinity. Moreover, a higher Ca2+ level has been noted due to an alteration of Ca2+ homoeostasis in cancer cells [46]. Consistent with this notion, their interaction in the cell was detected in a mammalian two-hybrid assay, where S100A4 was exogenously overexpressed (Figure 2). Importantly, the expression level of endogenous S100A4 in the MDA-MB 231 breast cancer cell line can be comparable with that of the exogenous protein (results not shown). Thus this interaction would occur in cancer cells overexpressing S100A4.

In conclusion, we have provided a possible mechanism of how TGF-β and Smad proteins become stronger metastatic-promoting factors in some advanced tumours. Overexpression of S100A4 in cancer cell lines strongly correlates with p53 status (deletion or mutation) [25]. Although the detailed mechanism is unknown, this strongly suggests that wild-type p53 keeps cells from overexpressing S100A4. A p53 mutation may actively co-operate with the TGF-β signal for a more aggressive behaviour of cancer cells. Recently, Adorno et al. [43] have reported that the p53 mutant observed in many types of cancer enpowers TGF-β-induced metastasis through the Smad pathway. Their observations and the results of our present study suggest that p53 status may be a key to TGF-β-mediated malignancy, and influences pro-metastatic activity of TGF-β at multiple levels. A p53 mutant, through direct interaction with Smads, on one hand down-regulates metastasis inhibitory genes such as Sharp-1 and Cyclin G2 [43]; on the other hand it turns on S100A4 expression to increase TGF-β/Smad-mediated cell invasion ability.

It is possible that, in addition to S100A4, some other proteins overexpressed in cancer cells may also modulate Smad function. Finding such additional Smad partners in cancer will establish a model in which TGF-β/Smad signalling promotes the progression of advanced tumours and may provide important therapeutic implications.

Abbreviations

     
  • CDK

    cyclin-dependent kinase

  •  
  • Cited

    CBP [CREB (cAMP-response-element-binding protein)-binding protein]/p300-interacting transactivator with glutamic acid (E)/aspartic acid (D)-rich C-terminal domain

  •  
  • DMEM

    Dulbecco's modified Eagle's medium

  •  
  • FBS

    fetal bovine serum

  •  
  • GST

    glutathione transferase

  •  
  • HA

    haemagglutinin

  •  
  • HEK

    human embryonic kidney

  •  
  • MAPK

    mitogen-activated protein kinase

  •  
  • MMP

    matrix metalloproteinase

  •  
  • qPCR

    quantitative PCR

  •  
  • RT

    reverse transcription

  •  
  • siRNA

    small interfering RNA

  •  
  • TGF-β

    transforming growth factor-β

  •  
  • TβR

    TGF-β receptor

AUTHOR CONTRIBUTION

Isao Matsuura directed the project, planned, designed and conducted the experiments, and prepared the Figures and the manuscript. Chen-Yu Lai conducted the BIAcore assay experiments. Keng-Nan Chiang conducted the cloning experiments.

We thank Dr G. Liou for technical support. We thank B.L. Houston for helpful discussion and support.

FUNDING

This work was supported by the National Health Research Institutes [grant numbers MG-096-PP-12, MG-097-PP-12, MG-098-PP-12 (to I.M.)].

References

References
1
Siegel
P. M.
Massague
J.
Cytostatic and apoptotic actions of TGF-β in homeostasis and cancer
Nat. Rev. Cancer
2003
, vol. 
3
 (pg. 
807
-
821
)
2
Attisano
L.
Wrana
J. L.
Smads as transcriptional co-modulators
Curr. Opin. Cell Biol.
2000
, vol. 
12
 (pg. 
235
-
243
)
3
Shi
Y.
Massague
J.
Mechanisms of TGF-β signaling from cell membrane to the nucleus
Cell
2003
, vol. 
113
 (pg. 
685
-
700
)
4
Derynck
R.
Akhust
R. J.
Balman
A.
TGF-β signaling in tumor suppression and cancer progression
Nat. Genet.
2001
, vol. 
29
 (pg. 
117
-
129
)
5
Massague
J.
Blain
S. W.
Lo
R. S.
TGF-β signaling in growth control, cancer and heritable disorders
Cell
2000
, vol. 
103
 (pg. 
295
-
309
)
6
Massague
J.
TGF-β in cancer
Cell
2008
, vol. 
134
 (pg. 
215
-
229
)
7
Levy
L.
Hill
C. S.
Alterations in components of the TGF-β superfamily signaling pathways in human cancer
Cytokine Growth Factor Rev.
2006
, vol. 
17
 (pg. 
41
-
58
)
8
Pardali
K.
Moustakas
A.
Actions of TGF-β as tumor suppressor and prometastatic factor in human cancer
Biochem. Biophys. Acta
2006
, vol. 
1775
 (pg. 
21
-
62
)
9
Deryugina
E. I.
Quigley
J. P.
Matrix metalloproteinases and tumor metastasis
Cancer Metastasis Rev.
2006
, vol. 
25
 (pg. 
9
-
34
)
10
Yoon
S.-O.
Park
S.-J.
Yun
C.-H.
Chung
A.-S.
Roles of matrix metalloproteinases in tumor metastasis and angiogenesis
J. Biochem. Mol. Biol.
2003
, vol. 
36
 (pg. 
128
-
137
)
11
Seton-Rogers
S. E.
Lu
Y.
Hines
L. M.
Koundinya
M.
LaBaer
J.
Muthuswamy
S. K.
Brugge
J. S.
Cooperation of the ErbB2 receptor and transforming growth factor-β in induction of migration and invasion in mammary epithelial cells
Proc. Natl. Acad. Sci. U.S.A.
2004
, vol. 
101
 (pg. 
1257
-
1262
)
12
Uttamsingh
S.
Bao
X.
Nguyen
K. T.
Bhanot
M.
Gong
J.
Chan
J. L.
Liu
F.
Chu
T. T.
Wang
L.-H.
Synergistic effect between EGF and TGF-β1 in inducing oncogenic properties of intestinal epithelial cells
Oncogene
2007
, vol. 
27
 (pg. 
2626
-
2634
)
13
Tian
Y.-C.
Chen
Y.-C.
Chang
C.-T.
Hung
C.-C.
Wu
M.-S.
Phillips
A.
Yang
C.-W.
Epidermal growth factor and transforming growth factor-β1 enhance HK-2 cell migration though a synergistic increase of matrix metalloproteinase and sustained activation of ERK signaling pathway
Exp. Cell Res.
2007
, vol. 
313
 (pg. 
2367
-
2377
)
14
Muraoka
R. S.
Koh
Y.
Roebuck
L. R.
Sanders
M. E.
Brantley-Sieders
D.
Gorska
A. E.
Moses
H. L.
Arteaga
C. L.
Increased malignancy of Neu-induced mammary tumors overexpressing active transforming growth factor-β1
Mol. Cell. Biol.
2003
, vol. 
23
 (pg. 
8691
-
8703
)
15
Chou
Y.-T.
Wang
H.
Chen
Y.
Danielpour
D.
Yang
Y-C.
Cited2 modulates TGF-β-mediated upregulation of MMP9
Oncogene
2006
, vol. 
25
 (pg. 
5547
-
5560
)
16
Marenholz
I.
Heizman
C. W.
Fritz
G.
S100 proteins in mouse and man: from evolution to function and pathology (including an update of the nomenclature)
Biochem. Biophys. Res. Commun.
2004
, vol. 
19
 (pg. 
739
-
744
)
17
Rudland
P. S.
Platt-Higgins
A. M.
Renshaw
C. A.
West
C. R.
Winstanley
J. H. R.
Robertson
L.
Barracolough
R.
Prognostic significance of the metastasis-inducing protein S100A4 (p9Ka) in human breast cancer
Cancer Res.
2000
, vol. 
60
 (pg. 
1595
-
1603
)
18
Gongoll
S.
Peters
G.
Mengel
M.
Piso
P.
Klempnauer
J.
Kreipe
H.
von Wasielewski
R.
Prognostic significance of calcium-binding protein S100A4 in colorectal cancer
Gastroenterology
2002
, vol. 
123
 (pg. 
1478
-
1484
)
19
Davies
B. R.
O'Donnell
M.
Durkan
G. C.
Rudland
P. S.
Barraclough
R.
Neal
D. E.
Mellon
J. K.
Expression of S100A4 protein is associated with metastasis and reduced survival in human bladder cancer
J. Pathol.
2002
, vol. 
196
 (pg. 
292
-
299
)
20
Ninomiya
I.
Ohta
T.
Fushida
S.
Endo
Y.
Hashimoto
T.
Yagi
M.
Fujimura
T.
Nishimura
G.
Tani
T.
Shimizu
K.
, et al. 
Increased expression of S100A4 and its prognostic significance in esophageal squamous cell carcinoma
Int. J. Oncol.
2001
, vol. 
18
 (pg. 
715
-
720
)
21
Platt-Higgins
A. M.
Renshaw
C. A.
West
C. R.
Winstanley
J. H.
De Silva Rudland
S.
Barracolough
R.
Rudland
P. S.
Comparison of the metastasis-inducing protein S100A4 (p9ka) with other prognostic markers in human breast cancer
Int. J. Cancer
2000
, vol. 
89
 (pg. 
198
-
208
)
22
Lee
W. Y.
Su
W. C.
Lin
P. W.
Guo
H. R.
Chang
T. W.
Chen
H. H.
Expression of S100A4 and Met: potential predictors for metastasis and survival in early-stage breast cancer
Oncology
2004
, vol. 
66
 (pg. 
429
-
438
)
23
Ambartsumian
N. S.
Grigorian
M. S.
Larsen
I. F.
Karlstorm
O.
Sidenius
N.
Rygaard
J.
Georgiev
G.
Lukanidin
E.
Metastasis of mammary carcinomas in GRS/A hybrid mice transgenic for the mts1 gene
Oncogene
1996
, vol. 
13
 (pg. 
1621
-
1630
)
24
Davies
M. P.
Rudland
P. S.
Robertson
L.
Parry
E. W.
Jolicoeur
P.
Barraclough
R.
Expression of the calcium-binding protein S100A4 (p9Ka) in MMTV-neu transgenic mice induces metastasis of mammary tumors
Oncogene
1996
, vol. 
13
 (pg. 
1631
-
1637
)
25
Grigorian
M.
Andersen
S.
Tulchinsky
E.
Kriajevska
M.
Carlberg
C.
Kruse
C.
Cohn
M.
Ambartsumian
N.
Christensen
A.
Selivanova
G.
Lukanidin
E.
Tumor suppressor p53 protein is a new target for the metastasis-associated Mts1/S100A4 protein
J. Biol. Chem.
2001
, vol. 
276
 (pg. 
22699
-
22708
)
26
Garrett
S. C.
Varney
K. M.
Weber
D. J.
Bresnick
A. R.
S100A4, a mediator of metastasis
J. Biol. Chem.
2006
, vol. 
281
 (pg. 
677
-
680
)
27
Li
Z. H.
Spektor
A.
Varlamova
O.
Bresnick
A. R.
Mts1 regulates the assembly of nonmuscle myosin-IIA
Biochemistry
2003
, vol. 
42
 (pg. 
14258
-
14266
)
28
Li
Z.
Bresnick
A. R.
The S100A4 metastasis factor regulates cellular motility via a direct interaction with myosin-IIA
Cancer Res.
2006
, vol. 
66
 (pg. 
5173
-
5180
)
29
Matsuura
I.
Denissova
N. G.
Wang
G.
He
D.
Long
J.
Liu
F.
Cyclin-dependent kinases regulates antiproliferative function of Smads
Nature
2004
, vol. 
430
 (pg. 
226
-
231
)
30
Santner
S. J.
Dawson
P. J.
Tait
L.
Soule
H. D.
Eliason
J.
Mohamed
A. N.
Wolman
S. R.
Heppner
G. H.
Miller
F. R.
Malignant MCF10CA1 cell lines derived from premalignant human breast epithelial MCF10AT cells
Breast Cancer Res. Treat.
2001
, vol. 
65
 (pg. 
101
-
110
)
31
Tang
B.
Vu
M.
Booker
T.
Santner
S. J.
Miller
F. R.
Anver
M. R.
Wakefield
L. M.
TGF-β switches from tumor suppressor to prometastatic factor in a model of breast cancer progression
J. Clin. Invest.
2003
, vol. 
112
 (pg. 
1116
-
1124
)
32
Swift
S.
Lorens
J.
Achacoso
P.
Nolan
G. P.
Coligan
J. E.
Bierer
B. E.
Margulies
D. H.
Shevach
E. M.
Strober
W.
Coics
R.
Rapid production of retroviruses for efficient gene delivery to mammalian cells using 293T cell-based systems
Current Protocols in Immunology
1999
Supplement 31
New York
John Wiley & Sons
 
Unit 10.28
33
Bakin
A. V.
Tomlinson
A. K.
Bhowmick
N. A.
Moses
H. L.
Arteaga
C. L.
Phosphatidylinositol 3-kinase function is required for transforming growth factor- mediated epithelial to mesenchymal transition and cell migration
J. Biol. Chem.
2000
, vol. 
275
 (pg. 
36803
-
36810
)
34
Tian
F.
Byfield
S. D.
Parks
W. T.
Yoo
S.
Felici
A.
Tang
B.
Piek
E.
Wakefield
L. M.
Roberts
A. B.
Reduction in Smad2/3 signaling enhances tumorigenesis but suppresses metastasis of breast cancer cell lines
Cancer Res.
2003
, vol. 
63
 (pg. 
8284
-
8292
)
35
Kang
Y.
He
W.
Tulley
S.
Gupta
G. P.
Serganova
I.
Chen
C. R.
Manova-Todorova
K.
Blasberg
R.
Gerald
W. L.
Massague
J.
Breast cancer bone metastasis mediated by the Smad tumor suppressor pathway
Proc. Natl. Acad. Sci. U.S.A.
2005
, vol. 
102
 (pg. 
13909
-
13914
)
36
Deckers
M.
van Dinther
M.
Buijs
J.
Que
I.
Lowik
C.
van der Pluijm
G.
ten Dijke
P.
The tumor suppressor Smad4 is required for transforming growth factor-β-induced epithelial to mesenchymal transition and bone metastasis of breast cancer cells
Cancer Res.
2006
, vol. 
66
 (pg. 
2202
-
2209
)
37
Zimmerman
C. M.
Kariapper
M. S.
Mathews
L. S.
Smad proteins physically interact with calmodulin
J. Biol. Chem.
1998
, vol. 
273
 (pg. 
677
-
680
)
38
Scherer
A.
Graff
J. M.
Calmodulin differentially modulates Smad1 and Smad2 signaling
J. Biol. Chem.
2000
, vol. 
275
 (pg. 
41430
-
41438
)
39
Prosser
B. L.
Wright
N. T.
Hernãndez-Ochoa
E. O.
Varney
K. M.
Liu
Y.
Olojo
R. O.
Zimmer
D. B.
Weber
D. J.
Schneider
M. F.
S100A1 binds to the calmodulin-binding site of ryanodine receptor and modulates skeletal muscle excitation-contraction coupling
J. Biol. Chem.
2008
, vol. 
283
 (pg. 
5046
-
5057
)
40
Wright
N. T.
Prosser
B. L.
Varney
K. M.
Zimmer
D. B.
Schneider
M. F.
Weber
D. J.
S100A1 and calmodulin compete for the same binding site on ryanodine receptor
J. Biol. Chem.
2008
, vol. 
283
 (pg. 
26676
-
26683
)
41
Chen
H.
Fernig
D. G.
Rudland
P. S.
Sparks
A.
Wilkinson
M. C.
Barraclough
R.
Binding to intracellular targets of the metastasis-inducing protein, S100A4 (p9Ka)
Biochem. Biophys. Res. Commun.
2001
, vol. 
286
 (pg. 
1212
-
1217
)
42
Zhang
Y.
Feng
X.-H.
Derynck
R.
Smad3 and Smad4 corporate with c-Jun/c-Fos to mediate TGF-β-induced transcription
Nature
1998
, vol. 
394
 (pg. 
909
-
913
)
43
Adorno
M.
Cordenonsi
M.
Dupont
S.
Wong
C.
Hann
B.
Solari
A.
Bobisse
S.
Rondina
M. B.
Guzzardo
V.
Parenti
A. R.
, et al. 
A mutant-p53/Smad complex opposes p63 to empower TGF-β-induced metastasis
Cell
2009
, vol. 
137
 (pg. 
87
-
98
)
44
Wang
G.
Matsuura
I.
He
D.
Liu
F.
TGF-β-inducible phosphorylation of Smad3
J. Biol. Chem.
2009
, vol. 
284
 (pg. 
9663
-
9673
)
45
Funaba
M.
Zimmerman
C. M.
Mathews
L. S.
Modulation of Smad2-mediated signaling by extracellular signal-regulated kinase
J. Biol. Chem.
2002
, vol. 
277
 (pg. 
41361
-
41368
)
46
Lee
W. J.
Monteith
G. R.
Roberts-Thomson
S. J.
Calcium transport and signaling in the mammary gland: targets for breast cancer
Biochem. Biophys. Acta
2006
, vol. 
1765
 (pg. 
235
-
255
)