Transforming Growth Factor-B1 Induces Tissue Inhibitor of Metalloproteinase-1 Expression via Activation of Extracellular Signal-Regulated Kinase and Sp1 in Human Fibrosarcoma Cells

The net balance of matrix metalloproteinases (MMP) and tissue inhibitor of metalloproteinases (TIMP) system has been known to be a key factor in tumor cell invasion. In the present study, we investigated the molecular mechanisms of anti-invasive and antimigrative activity of transforming growth factor (TGF)-B1 on HT1080 human fibrosarcoma cells. In in vitro Matrigel invasion and Transwell migration assays, TGF-B1 dose-dependently inhibited the invasion and migration of HT1080 cells, respectively. Gelatin zymography, Western blot, and real-time PCR analysis showed that TGF-B1 enhanced the expression and secretion of MMP-2, TIMP-1, and, to a lesser degree, MMP-9 but not membrane type 1-MMP and TIMP-2. The addition of recombinant TIMP-1 protein reduced the Matrigel invasion and Transwell migration of HT1080 cells, similar to TGF-B1. Because augmentation of TIMP-1 might be the major factor for the anti-invasive and antimigrative activity of TGF-B1, we investigated possible molecular mechanisms responsible for the expression of TIMP-1 induced by TGF-B1. Treatment of HT1080 cells with TGF-B1 rapidly phosphorylated three mitogen-activated protein kinases [MAPK; extracellular signal-regulated kinase 1/2 (ERK1/2), p38, and c-Jun NH2-terminal kinase] and Akt. Among these kinases, the inhibition of only ERK1/2 pathway by PD98059, a specific inhibitor of MAPK/ERK kinase(MEK)-1, and transfection of dominant-negative MEK 1 effectively blocked the TIMP-1 induction by TGF-B1. Mithramycin, a specific inhibitor of Sp1 transcription factor, but not curcumin, an inhibitor of activator protein-1, and transfection of Sp1 small interfering RNA significantly inhibited the TGF-B1-induced expression of TIMP-1. In addition, electrophoretic mobility shift assay showed that TGF-B1 up-regulated Sp1 DNA-binding activity, and PD98059 and mithramycin effectively inhibited these events. Finally, pretreatment of HT1080 cells with PD98059 and mithramycin, but not curcumin, restored the invasive activity of these cells. Taken together, these data suggest that TGF-B1 modulates the net balance of the MMPs/TIMPs the systems in HT1080 cells for antiinvasion and antimigration by augmenting TIMP-1 through ERK1/2 pathway and Sp1 transcription factor. (Mol Cancer Res 2006;4(3):209–20) Introduction Tumor cell invasion is a multistep process, involving several interactions between tumor cells and extracellular matrix (ECM), particularly a directed proteolysis of ECM, such as basement membranes required for intravasation and extravasation of tumor cells (1). Proteolytic degradation of ECM components involves several proteases, such as matrix metalloproteinases (MMP), plasminogen activators, and serine proteases secreted by invading tumor cells (2-5). Among the ECM-degrading proteases, however, much attention have been focused on the MMP family, especially MMP-2 and MMP-9, and many evidences have shown that elevated levels of MMP-2 and MMP-9 are well correlated with an invasive phenotype of cancer cells (6). The extracellular activities of these enzymes are inhibited, when they are complexed with endogenous inhibitors, such as tissue inhibitor of metalloproteinase (TIMP)-1 and TIMP-2. Consequently, an imbalance between MMP and TIMP activities results in an excessive ECM degradation necessary for tumor invasion and metastasis (7, 8). TIMPs are multifunctional proteins that can inhibit the catalytic activity of MMPs, thus maintaining ECM homeostasis (9). Besides acting as proteinase inhibitors, TIMPs also modulate cell growth, apoptosis, and angiogenesis (9). Currently, four Received 8/19/05; revised 1/20/06; accepted 1/30/06. Grant support: National Nuclear R&D Program of Ministry of Science and Technology (Seoul, Korea). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Note: Supplementary data for this article are available at Molecular Cancer Research Online (http://mcr.aacrjournals.org/). H-J. Kwak and M-J. Park contributed equally to this work. Requests for reprints: Seok-Il Hong, Laboratory of Functional Genomics, Korea Institute of Radiological and Medical Sciences, 215-4 Nowon-Ku, GongneungDong, Seoul 139-706, Korea. Phone: 82-2-970-1260; Fax: 82-2-970-2402. E-mail: [email protected] Copyright D 2006 American Association for Cancer Research. doi:10.1158/1541-7786.MCR-05-0140 Mol Cancer Res 2006;4(3). March 2006 209 on April 20, 2017. © 2006 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from members of the TIMP family have been characterized: TIMP-1, TIMP-2, TIMP-3, and TIMP-4 (10, 11). Among these, TIMP-1 is a glycosylated protein with molecular mass ranging from 28.5 to 34 kDa, depending on the degree of glycosylation, and is secreted in a soluble form by different cell types and forms a 1:1 complex with MMP-9 (9). Tumor invasion and metastasis are negatively regulated by TIMP-1 in different types of tumor cell lines (12-16), and down-regulation of TIMP-1 expression or mutated TIMP-1 is associated with increased invasiveness (17, 18). Recent transgenic studies also showed that endogenous overexpression or suppression of TIMP-1 inhibited or enhanced the tumor growth in several cell types, respectively (19-21). In addition, TIMP-1 has been shown to exert an antiangiogenic activity both in vitro and in vivo (22, 23). TIMP-1 promoter contains several 12-O -tetradecanoylphorbol-13-acetate – responsive elements, such as activator protein-1 (AP-1), Sp1, and Ets sites, and functional interaction of these transcription factors has been shown to induce synergistic transcriptional activation of the TIMP-1 promoter. Consequently, contrary to TIMP-2, TIMP-1 expression is up-regulated by several factors, including phorbol esters, interleukin-6, transforming growth factor (TGF)-h1, epidermal growth factor, oncostatin M, or erythropoietin (9, 24, 25). TGF-h, 25-kDa homodimeric protein with autocrine and paracrine activities, is a multifunctional protein that has been implicated in various physiologic and pathologic conditions, including cell cycle control, carcinogenesis in human skin, and invasion and metastasis of carcinoma cells (26). In tumor progression, TGF-h1 has been shown to have dual function, either decreasing or promoting cancer invasion, depending on the type and the progression stage of tumor (27, 28). Consistent with proinvasive or anti-invasive roles in cancer progression, TGF-h1 stimulates tumor invasion by up-regulating integrin, MMP-2, MMP-9, and urokinasetype plasminogen activator expression (29-33) while suppresses it by down-regulating plasminogen activator systems, MMPs, and plasmin-dependent ECM degradation through stimulating the expression of TIMP-1 and plasminogen activator inhibitor-1 (34-37). TGF-h1 signaling is initiated via binding of ligand to its serine/threonine kinase receptor type II and then type I on the cell surface (38). On activation, the type I and II receptors interact to form hetero-oligomers, which then activate receptor-associated Smads, Smad2 and Smad3. The receptorassociated Smads then bind to Smad4, translocate to the nucleus, and ultimately regulate transcription and expression of target genes. Therefore, Smad proteins are the best known intracellular substrates for signal transmitter of TGF-h1 receptor activation (38). In addition to the Smad-mediated signaling, TGF-h1 also can signal through mitogen-activated protein kinase (MAPK) pathways, including extracellular signal-regulated kinase 1/2 (ERK 1/2) and p38 (39-41). Several studies have shown that MAPK pathways are involved in the TGF-h1-mediated modulation of ECMdegrading proteases and their inhibitors in the process of cellular invasion (34, 42-44). However, there is little information available about signaling mechanisms of TGFh1-mediated TIMP-1 expression and anti-invasion of tumor cells. In the present study, we investigated the role of TGF-h1 in the regulation of MMPs and TIMPs expression in HT1080, a cellular model of invasive fibrosarcoma cells, and observed that treatment of HT1080 cells with TGF-h1 significantly upregulated TIMP-1 expression and blocked invasion and migration of the cells and that ERK1/2 and Sp1 signaling pathways were crucially involved in these events. Results TGF-b1 Inhibited the Invasion and Migration of HT1080 Cells In vitro First, we examined the effect of TGF-h1 on invasiveness of HT1080 cells using Matrigel invasion assay system. As shown in Fig. 1A, invasion of HT1080 cells was significant and was reduced dose-dependently by treatment with TGF-h1. Next, the effects of TGF-h1 on the migration of HT1080 cells were determined by Transwell migration assay. In line with the results from the above invasion assays, TGF-h1 also inhibited the migration of HT1080 cells in a dose-dependent manner (Fig. 1B). No distinctive cellular morphologic changes, which are typically associated with apoptosis, such as cell detachment, rounding, or chromosomal fragmentation, were detected under the experimental condition (data not shown). In addition, significant cytotoxic effect was not detected by 3-(4,5dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assays in the serum-free medium following up to 100 ng/mL TGF-h1 treatment for 12 and 24 hours (Supplementary Fig. S1), indicating that the anti-invasive activity of TGF-h1 was not due to cytotoxic effect of this cytokine. Augmentation of TIMP-1 Might Be Responsible for the Anti-Invasive and Antimigrative Activity of TGF-b1 in HT1080 Cells Tumor cell invasiveness is strongly associated with the balance between ECM-degrading proteases and their inhibitors, especially MMPs and TIMPs. Therefore, to investigate the effects of TGF-h1 on MMP-2, MMP-9, membrane type 1-MMP (MT1-MMP), TIMP-1, and TIMP-2 induction in HT1080 cells, we measured the expression of these molecules in the conditioned medium or cell lysates by using gelatin zymography analysis and Western blotting. As shown in Fig. 2A and B, treatment of HT1080 cells with TGF-h1 significantly increased the secretion of MMP-2 and TIMP-1, but not MT1-MMP and TIMP-2. TGF-h1 also increased the secretion of MMP-9 but to a lesser degree. Because TGF-h1 modulated the secretion of MMP-2 and TIMP-1 significantly, we further investigated the effect of TGF-h1 on the mRNA expression of these molecules by using quantitative real-time PCR method. As shown in Fig. 2C, TGF-h1 significantly augmented MMP-2 and TIMP-1 mRNA levels in HT1080 cells. Enhanced MMP-2 and TIMP-1 mRNA peaked at 3 and 6 hours, respectively, and declined thereafter. Because TGF-h1 significantly suppressed the invasion and migration of HT1080 cells, we wondered whether TIMP-1 secretion by TGF-h1 might be critically involved in the inhibition. Therefore, we did Matrigel invasion and Transwell migration assays in the presence or absence of recombinant TIMP-1 and found that TIMP-1 dose-dependently suppressed Kwak et al. Mol Cancer Res 2006;4(3). March 2006 210 on April 20, 2017. © 2006 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from the invasion and migration of HT1080 cells (Fig. 2D and E). However, the addition of TIMP-1 up to 100 ng/mL did not completely inhibit invasion (Supplementary Fig. S2), indicating that TIMP-1 is not the only invasion suppressor in this system. The above data indicate that TGF-h1 regulates the expression of MMP-2 and TIMP-1 at the transcriptional level and that augmentation of TIMP-1 might be a key event in modulating the balance of MMPs and TIMPs expression against invasion and migration of HT1080 cells. To further confirm that TIMP-1 was the direct mediator of the effects of TGF-h1 on the inhibition of invasion, endogenous expression of TIMP-1 genes in HT1080 cells was silenced by TIMP-1 small interfering RNA (siRNA), and efficient and specific knockdown was observed by Western blot analysis of cell lysates (Fig. 2F). Transfection of HT1080 cells with TIMP-1 siRNA reduced TGF-h1-mediated inhibition of invasion compared with control siRNA-transfected cells, whereas TIMP-1 siRNA transfection alone slightly up-regulated the invasion of these cells (Fig. 2G). These results indicate that TIMP-1 is the critical invasion and migration suppressor in HT1080 cells, at least in part. TGF-b1 Induced Expression of MMP-2 and TIMP-1 through p38 and c-Jun NH2-terminal kinase and ERK1/2 Signaling Pathway, Respectively To elucidate signaling mechanism possibly involved in antiinvasive action of TGF-h1, we next examined whether TGF-h1 influenced MAPKs and Akt signaling pathways in HT1080 cells. As shown in Fig. 3A, treatment of the cells with TGF-h1 strongly up-regulated ERK1/2 and Akt phosphorylation at 15 minutes, lasting up to 180 minutes. TGF-h1 also increased p38 and c-Jun NH2-terminal kinase (JNK) phosphorylation, but the effect was weak compared with ERK1/2 and Akt phosphorylation. To further ascertain the involvement of the activation of these kinases in the expression of MMPs and TIMPs, the cells were treated with TGF-h1 in the presence or absence of highly specific inhibitors of MAPK/ERK kinase (MEK)-1 (PD98059), p38 (SB202190), JNK (SP600125), and phosphatidylinositol-3 kinase (PI3K; LY294002). Gelatin zymography, Western blot, and real-time PCR analyses clearly showed that pretreatment of the cells with SB202190 and SP600125, but not PD98059 and LY294002, reduced the increase of MMP-2 induced by TGF-h1, whereas only PD98059 significantly FIGURE 1. Effect of TGF-h1 on the invasion and migration of HT1080 cells. Chemotaxic cells were coated with Matrigel (100 Ag/cm) on the upper side of filter for invasion assay or with type I collagen (100 Ag/cm) on the underside of filter for migration assay, and 2 10 HT1080 cells cultured in the presence or absence of TGF-h1 at indicated concentrations were placed in the upper well. Invasion and migration of the cells were determined by measuring the ability to pass through a layer of Matrigel-coated and intact filters, respectively. Detailed procedures are described in Materials and Methods. Top, HT1080 cells invaded (A) and migrated (B) cells under the membrane after 12 hours (top ); bottom, invasion (A) and migration (B) were determined by counting cells in four microscopic fields per sample. Bars, SD. *, P < 0.05 versus control. TGF-h1 Signaling for Induction of TIMP-1 Mol Cancer Res 2006;4(3). March 2006 211 on April 20, 2017. © 2006 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from blocked the increase of TIMP-1 by TGF-h1 (Fig. 3B-D). Because the above data strongly implicated the involvement of ERK1/2 signaling pathway in TGF-h1-induced TIMP-1 expression, we further investigated the role of ERK1/2 pathway in TGF-h1-stimulated TIMP-1 expression by using dominantnegative (DN) mutants of MAPKs. Thus, we transiently FIGURE 2. Effect of TGF-h1 on MMP and TIMP expression and in vitro invasion and migration stimulated with TGF-h1 or TIMP-1 protein. HT1080 cells were treated with various concentrations of TGF-h1. The conditioned serum-free medium was collected for analysis after 18 hours. A. Gelatin zymography analysis of conditioned medium from HT1080 cells to detect the secretion of MMP-2 and MMP-9. B. Western blot analysis of conditioned medium and membrane extracts to detect MMP-2, MMP-9, TIMP-1, TIMP-2, and MT1-MMP. C. Quantitative real-time PCR analysis of total RNA isolated from HT1080 cells treated with TGF-h1 (10 ng/mL) for indicated time points. Quantification of real-time PCR analysis. Columns, average for each sample normalized with the amount of GAPDH; bars, SD. *, P < 0.05 versus control; TGF-h1. Matrigel invasion (D) and Transwell migration (E) of HT1080 cells cultured in the absence or presence of TGF-h1 (10 ng/mL) or TIMP-1 (0.1 or 1 ng/mL). Left, HT1080 cells invaded and migrated under the membrane after 12 hours; right, invasion and migration rates were determined by counting cells in four microscopic fields per sample. Bars, SD. *, P < 0.05 versus control; TGF-h1; **, P < 0.05 versus control; TIMP-1. F. Western blot analysis of HT1080 cells after transfection with TIMP-1 (200 nmol/L) or control siRNA for 24 hours. G. Matrigel invasion assay of siRNA-transfected cells cultured in the absence or presence of TGF-h1 (10 ng/mL) for an additional 12 hours. Whole-cell lysates were analyzed for the detection of TIMP-1 by Western blot. Invasiveness was determined by counting cells in four microscopic fields per sample. Bars, SD. *, P < 0.05 versus control; TGF-h1 in control siRNA-transfected cells; **, P < 0.05 versus TGF-h1; TGF-h1 in TIMP-1 siRNA-transfected cells. Kwak et al. Mol Cancer Res 2006;4(3). March 2006 212 on April 20, 2017. © 2006 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from transfected DN mutant plasmids of MEK-1 (the immediate upstream kinase of ERK1/2) and retroviruses for DN form of p38 and JNK into HT1080 cells for 24 hours. After another 24 hours of incubation with TGF-h1 (10 ng/mL) in serum-free medium, we measured the TIMP-1 in the conditioned medium by using gelatin zymography and Western blot analyses. Consistent with the results obtained with the pharmacologic inhibitors, DN-MEK-1 abrogated TGF-h1-induced TIMP-1 expression (Fig. 3E). On the other hand, transduction of DNp38 and DN-JNK retroviruses significantly suppressed MMP-2 but not TIMP-1 expression by TGF-h1 (Fig. 3F). These results suggest that ERK1/2 pathway is involved in the regulation of TGF-h1-induced TIMP-1 expression, whereas p38 and JNK pathway is involved in the regulation of MMP-2 expression in HT1080 cells, indicating that TGF-h1 regulates the expression of MMP-2 and TIMP-1 in HT1080 cells through different mechanisms. TGF-b1-Induced Sp1 Activation Was Involved in TIMP-1 Expression Several transcriptional factors, including AP-1, Sp1, and 12O-tetradecanoylphorbol-13-acetate–responsive element, regulate the expressions of TIMP-1 (45, 46). Therefore, to identify the possible transcription factor involved in TGF-h1-induced TIMP-1 expression, we tested the effect of curcumin and mithramycin, an inhibitor of AP-1 and Sp1, respectively, on the event. As shown in Fig. 4A, pretreatment of the cells with mithramycin, but not curcumin, significantly reduced the secretion of TIMP-1 induced by TGF-h1. Mithramycin also inhibited the induction of TIMP-1 mRNA by TGF-h1 (Fig. 4B). Furthermore, transfection of Sp1 siRNA significantly reduced the both Sp1 expression and TIMP-1 secretion induced by TGF-h1 as well as basal levels (Fig. 4C). Next, we investigated the involvement of Sp1 DNA-binding activity in the TIMP-1 expression. Because TGF-h1-mediated TIMP-1 expression was blocked by ERK1/2 inhibition, as shown above, it is likely that the inhibitory effect of PD98059 on TGF-h1induced TIMP-1 expression might have been mediated, at least in part, through down-regulation of Sp1 DNA-binding activity. To examine this possibility, we did electrophoretic mobility shift assay and found that TGF-h1 strongly increased Sp1 DNA-binding activities and that pretreatment of HT1080 cells with PD98059 and mithramycin effectively inhibited the TGFh1-induced Sp1 DNA-binding activities in a dose-dependent manner (Fig. 5A). Specificity of the Sp1 band was confirmed by detecting supershift band, when the nuclear proteins were preincubated with Sp1 antibodies (Fig. 5A). TGF-h1 also induced the AP-1 DNA-binding activity and dose-dependently blocked the event by pretreatment with PD98059 and curcumin (Fig. 5B). These results together strongly suggest that TGF-h1mediated TIMP-1 expression in HT1080 cells is regulated dominantly by Sp1 transcriptional activity, which is mediated by ERK1/2. ERK1/2 Pathway and Sp1 Transcription Factor Were Involved in TGF-b1-Induced Anti-Invasion and Antimigration of HT1080 Cells As shown above, because ERK1/2 and Sp1 signaling pathways were critical for induction of TIMP-1 in HT1080 cells by TGF-h1, we next examined whether the inhibition of these pathways restored the invasive and migrative activity of these cells. As expected, treatment of HT1080 cells with PD98059 and mithramycin, but not SB202190 and SP600125, ameliorated the TGF-h1-mediated suppression of HT1080 cell invasion (Fig. 6A and B) and migration (Fig. 6D and E). However, treatment of the cells with curcumin together with TGF-h1 more suppressed the invasion (Fig. 6C) and migration (Fig. 6F) of HT1080 cells. These additional inhibitory effects of HT1080 cell invasion and migration by curcumin were not due to cytotoxicity, because curcumin up to 10 Amol/L did not show any toxic effect on HT1080 cells in 3-(4,5-dimethylthiazol-2-yl)-2,5diphenyltetrazolium bromide assay (Supplementary Fig. S3A). In addition, flow cytometric analysis revealed no increase in the proportion of Annexin V–positive cells and propidium iodide–positive cells with up to 10 Amol/L curcumin (Supplementary Fig. S3B). However, significant cytotoxicity was detected with >35 Amol/L concentration of curcumin (Supplementary Fig. S3B). Therefore, these data strongly indicate that ERK1/2 and Sp1 signaling pathways are important for the induction of TIMP-1 and suppression of invasion of HT1080 cells by TGF-h1. Discussion Invasion of tumor cells through ECM and basement membrane barriers is a crucial step in tumor dissemination and metastasis (47, 48). Proteolysis of ECM results from an imbalance between activators and inhibitors of proteinases, which are controlled by various cytokines or growth factors, including TGF-h1. Tumor cell invasion and migration activities are greatly increased by TGF-h1, leading to a more malignant phenotype (31, 34). On the other hand, several studies showed that TGF-h1 is a potent growth suppressor of different normal and tumor cell types (27, 32). Induction of TIMP-1 and inhibition of tumor cells invasion by TGF-h1 have been described previously in an invasive tumor cell line HT1080 (37); however, the regulation of these events by TGF-h1 has not been investigated. In the present study, therefore, we attempted to identify TGF-h1induced signaling pathway that is involved in the regulation of TIMP-1 expression and invasion and migration by HT1080 cells and showed that TGF-h1 suppressed the invasion and migration of HT1080 cells by modulating the balance of ECM-degrading proteases and their inhibitor proteins. Specifically, TGF-h1 up-regulated MMP-2, MMP9, and TIMP-1 expression, and augmentation of TIMP-1 by TGF-h1 might be responsible for suppressing invasion and migration of these cells. TGF-h1 induced TIMP-1 at both mRNA and protein levels in a dose-dependent manner. Furthermore, TGF-h1-induced TIMP-1 expression and antiinvasion and antimigration were regulated through the ERK1/ 2 signaling pathway and Sp1 transcription factor, suggesting the possible mechanism of anti-invasive and antimigrative action of TGF-h1 in HT1080 cells. Invasion of tumor cells is determined by the net activity of ECM-degrading proteases and their inhibitors. Several reports showed that TGF-h1 stimulates tumor invasion by TGF-h1 Signaling for Induction of TIMP-1 Mol Cancer Res 2006;4(3). March 2006 213 on April 20, 2017. © 2006 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from up-regulating integrin, MMP-2, MMP-9, and urokinase-type plasminogen activator expression (29-33), while suppresses it by down-regulating plasminogen activator systems, MMPs, and plasmin-dependent ECM degradation through stimulating the expression of TIMP-1 and plasminogen activator inhibitor-1 (34-37). TGF-h1 suppression of invasion of HT1080 human fibrosarcoma cells seems to be largely due to the induction of TIMP-1 expression (9, 37). Consistent with the results of Kubota et al. (37), our data also showed that TGF-h1 upregulated the expression and secretion of MMP-2, MMP-9, and TIMP-1 but not MT1-MMP and TIMP-2. Although TGF-h1 induced the expression of MMP-2 and MMP-9, the net invasive and migrative activity of HT1080 cells was down-regulated due to the induction of TIMP-1 by TGF-h1. Indeed, the addition of recombinant TIMP-1 reduced the invasion and migration of these cells. In other tumor cell systems, it has also been reported that reconstitution of Smad3 restores the TGF-h1-induced TIMP-1 expression and results in inhibiting the invasion of FIGURE 3. Effect of various kinase inhibitors on the secretion of MMP-2 and TIMP-1 by TGF-h1 in HT1080 cells. A. Western blot analysis of HT1080 cells stimulated with TGF-h1 (10 ng/mL) for indicated times. Cell extracts (50 Ag each) were resolved by SDS-PAGE and probed with phosphorylated ERK1/2 (p-ERK1/2), phosphorylated p38 (p-p38 ), phosphorylated JNK (p-JNK ), and phosphorylated Akt (p-Akt ) antibodies. To verify equal loading, the blots were probed with actin. B to D. HT1080 cells were pretreated with various kinase inhibitors, SB202190 (SB ; p38 inhibitor, 10 Amol/L), PD98059 (PD ; MEK-1 inhibitor, 25 Amol/L), SP600125 (SP ; NK inhibitor, 10 Amol/L), and LY294002 (LY ; PI3K inhibitor, 5 Amol/L) for 1 hour before stimulation with TGF-h1 (10 ng/mL). B. Conditioned serum-free medium after 18 hours was collected for gelatin zymography to detect MMP-2. C. Western blot to detect MMP-2, TIMP-1, and TIMP-2. D. Quantitative real-time PCR analysis of total RNA isolated from HT1080 cells treated with various kinase inhibitors to detect the expression of MMP-2 and TIMP-1. Quantification by real-time PCR analysis. Columns, average for each sample normalized with the amount of GAPDH; bars, SD. *, P < 0.05 versus control; TGF-h1; **, P < 0.05 versus TGF-h1. Western blot analysis of cell lysates and conditioned medium from HT1080 cells transfected with DN mutant plasmids of MEK-1 (E) and retroviruses for DN form of p38 and JNK (F), respectively, in the presence or absence of TGF-h1 (10 ng/mL). CTL, control. Kwak et al. Mol Cancer Res 2006;4(3). March 2006 214 on April 20, 2017. © 2006 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from human choriocarcinoma cells (35). On the contrary, TGF-h1 inhibited invasiveness of normal trophoblast by up-regulating the TIMP-1 level, but not in JAR and JEG-2 choriocarcinoma cells, which show very low level of TIMP-1 expression (49). Therefore, TIMP-1 induction by TGF-h1 is an important mechanism for inhibiting invasion of choriocarcinoma, at least in part. In addition, human hepatoma cells (HepG2) also produce TIMP-1 in response to TGF-h1 (50). Therefore, TGFh1-induced TIMP-1 production might be critical mechanism for the suppression of invasion in some tumor cells, such as fibrosarcoma, choriocarcinoma, and hepatoma, although further studies are needed to uncover the critical role of TIMP-1 in invasiveness of these tumors. Several external stimuli, such as growth factors, phorbol esters, and cytokines, induce TIMP-1 expression in various cell types (9). However, the regulation mechanism of TIMP-1 is not well understood. Tong et al. reported that ERK1/2 and p38 pathways are required for maximal regulation of TIMP-1 in murine fibroblast by oncostatin M (24), and Kadri et al. showed that erythropoietin-induced TIMP-1 expression and secretion are mediated by ERK1/2 and PI3K/Akt signaling pathway (25). In line with these studies, we also examined the involvement of MAPK and PI3K/Akt pathways in the regulation of TIMP-1 by TGF-h1 in HT1080 cells. TGF-h1 phosphorylated three MAPKs (ERK1/2, p38, and JNK) and Akt, a PI3K downstream effector molecule. By using specific kinase inhibitors of these kinases, we found that the activation of ERK1/2 pathway was involved in TGF-h1-induced TIMP-1 expression and secretion in HT1080 cells. Actually, PD98059, a specific inhibitor of MEK, which is upstream kinase of ERK1/2, but not SB202190, SP600125, and LY294002, specific inhibitors of p38, JNK, and PI3K, respectively, could effectively inhibit TIMP-1 expression and secretion induced by TGF-h1 in these cells. Furthermore, transient transfection of DN-MEK-1, but not DN-p38 and DNJNK, abrogated the induction of TIMP-1 by TGF-h1. To our best knowledge, this is the first report to show that induction of TIMP-1 by TGF-h1 is regulated by ERK1/2 signaling pathway. In addition, Matrigel invasion and Transwell migration assays showed that blockade of ERK1/2 pathway by PD98059 restored the invasion and migration of HT1080 cells from the invasion and migration suppressive effect of TGF-h1. Therefore, in HT1080 cells, ERK1/2 pathway plays an important role in TIMP-1 expression following the suppression of invasion and migration by TGF-h1. Regulation of TIMP-1 gene expression is mainly mediated at the transcription level and involves the activation of several well-known transcription factors, including those belonging to AP-1, Sp1, and Pea3/Ets families (45). A recent report shows that the up-regulation of TIMP-1 promoter activity by TGFh1 requires the AP-1 response element (51). In other stimulatory systems, such as oncostatin M and erythropoietin, the dependence of TIMP-1 expression on AP-1 has also been described (25, 49). In the present study, we examined which transcription factor was related with the expression of TIMP1 and found that mithramycin, a specific Sp1 inhibitor, but not curcumin, which is known as an AP-1 inhibitor, significantly inhibited TGF-h1-induced TIMP-1 expression. In addition, transfection of siRNA for Sp1 dramatically diminished the TIMP-1 expression, and TGF-h1 increased the Sp1 DNA-binding activity and PD98059 and mithramycin effectively suppressed this increase. The inhibitory effect of PD98059 on Sp1 DNA-binding activity implies that the ERK1/2 signal pathway is critical for TGF-h1-induced Sp1 activation and TIMP-1 expression. We cannot, however, rule out the possible involvement of AP-1, the well-known transcription factor for TIMP-1 expression (25, 51, 52), and other transcription factors. Nevertheless, it is certain that Sp1 transcription factor is dominantly involved in the induction of TIMP-1 by TGF-h1 in these cell systems, partly at least. Therefore, the above observations led us to conclude that TGF-h1 increases TIMP-1 expression through ERK1/2 and Sp1 signaling pathways. FIGURE 4. Involvement of Sp1 transcription factor in TGF-h1mediated TIMP-1 induction in HT1080 cells. A. Western blot analysis of conditioned medium to detect the secretion of TIMP-1. Cells were pretreated with or without curcumin and mithramycin for 1 hour at the indicated concentrations and then stimulated with TGF-h1 (10 ng/mL) for 18 hours in serum-free medium. B. Quantitative real-time PCR analysis of total RNA isolated from HT1080 cells treated with or without mithramycin in the presence or absence of TGF-h1 (10 ng/mL). Quantification by realtime PCR analysis. Columns, average for each sample normalized with the amount of GAPDH; bars, SD. *, P < 0.05 versus control; TGF-h1; **, P < 0.05 versus TGF-h1; TGF-h1 + mithramycin. M, mithramycin. C. Western blot analysis of HT1080 cells to detect TIMP-1 and Sp1. Cells were transfected with Sp1 siRNA (200 nmol/L) or control siRNA for 24 hours and treated with TGF-h1 (10 ng/mL) for an additional 24 hours. Conditioned medium and whole-cell lysates were analyzed by Western blot for the detection of TIMP-1 and Sp1 proteins, respectively. TGF-h1 Signaling for Induction of TIMP-1 Mol Cancer Res 2006;4(3). March 2006 215 on April 20, 2017. © 2006 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from In summary, the data described above showed that TGF-h1 inhibited pericellular proteolysis by inducing the expression of TIMP-1 in HT1080 cells. Mechanistic analyses showed that ERK1/2 activation of TGF-h1 regulated the levels of TIMP-1 expression and this regulation was largely mediated by the regulation of transcription factor, Sp1. These results provide important clues in the regulatory mechanisms underlying TGF-h1-induced inhibition of tumor cell invasion and migration. Materials and Methods Reagents and Cell Culture Recombinant human TGF-h1 was purchased from Biosource Co. (Camarillo, CA). Antibodies against MT1-MMP, MMP-2, MMP-9, TIMP-1, and TIMP-2 were obtained from Calbiochem (La Jolla, CA), and antibodies against phosphorylated ERK1/2, p38, JNK, and Akt were from New England Biolabs (Beverly, MA). The following kinase inhibitors were purchased from Calbiochem and used in this study: PD98059 (ERK1/2), SB202190 (p38), SP600125 (JNK), and LY294002 (PI3K). Matrigel was purchased from Becton Dickinson (Bedford, MA), and gelatin and type I collagen were purchased from Sigma Chemical Co. (St. Louis, MO). HT1080 cells were grown in DMEM (Life Technologies, Grand Island, NY) supplemented with 2 mmol/L glutamine, 1 mmol/L sodium pyruvate, 100 units/mL penicillin, 100 Ag/mL streptomycin (Life Technologies), and 10% heat-inactivated fetal bovine serum (Life Technologies) in a humidified incubator containing 5% CO2 at 37jC. FIGURE 5. Effect of MEK-1 and Sp1 inhibitors on TGF-h1-induced DNA-binding activity of Sp1 and AP-1 in HT1080 cells. Cells were grown to 70% confluence in DMEM containing 10% fetal bovine serum, and medium was changed to serum-free medium. Cells were stimulated with TGF-h1 (10 ng/mL) in the presence or absence of MEK inhibitor, PD98059, mithramycin, or curcumin for 4 hours. Nuclear extracts from the cells were subjected to gel shift assay with Sp1 (A) and AP-1 (B) consensus oligonucleotides as described in Materials and Methods. For the detection of Sp1 supershift band, nuclear extracts from TGF-h1-treated cells were incubated with anti-Sp1 antibody for 30 minutes on ice and analyzed by electrophoretic mobility shift assay. Negative control, P-labeled Sp1 (A) or AP-1 (B) oligonucleotides without HeLa nuclear extract; Positive control, nuclear extract of HeLa cells incubated with P-labeled Sp1 (A) and AP-1 (B) oligonucleotides; SS, supershift. Kwak et al. Mol Cancer Res 2006;4(3). March 2006 216 on April 20, 2017. © 2006 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from Transfection and Transduction DN-MEK-1 plasmids (generous gift from Dr. Su-Jae Lee, Korea Institute of Radiological and Medical Sciences, Seoul, Korea) were transfected in HT1080 cells, using Effectene transfection reagent (Qiagen, Valencia, CA), by following the supplier’s instructions. For retrovirus transduction, cells were infected with retrovirus harboring DN-p38 and DN-JNK-1 (generous gift from Dr. Su-Jae Lee) for 90 minutes. After 24 hours, transfected and transduced cells were washed twice with serum-free medium and incubated with or without TGFh1 (10 ng/mL) in the same medium. The infection and selection of the target cells by puromycin were done as described previously (51). The siRNA against Sp1 and TIMP1 were purchased from Santa Cruz Biotechnology (Santa Cruz, CA), and transfections of cells with siRNA molecules were done using RNAiFect transfection reagent (Qiagen) as described by the manufacturer. Zymography Production of MMPs by HT1080 cells was analyzed by gelatin zymography. HT1080 cells in subconfluent culture condition (f70-80% confluent) were washed and replenished with serum-free DMEM and incubated with or without indicated concentrations of TGF-h1 in the presence or absence of various kinase inhibitors for 18 hours. Serum-free conditioned medium (20 AL) was mixed with SDS sample buffer without heating or reduction and applied to 10% polyacrylamide gels copolymerized with 1 mg/mL gelatin. After electrophoresis, gels were washed in 2.5% (v/v) Triton X-100 for 2 hours at room temperature to remove SDS, rinsed twice with water, and then incubated in developing buffer [50 mmol/ L Tris-HCl (pH 7.5), 5 mmol/L CaCl2, 1 Amol/L ZnCl2, and 0.1 NaN3] for 18 hours at 37jC. Subsequently, gels were fixed and stained with 10% 2propanol and 10% acetic acid containing 0.5% Coomassie blue R250. The protease activity was visualized as clear bands within the stained gel. Invasion and Migration Assays Matrigel invasion assays were done using modified Boyden chambers with polycarbonate Nucleopore membrane (Corning, Corning, NY). Precoated filters (6.5 mm in diameter, 8-Am pore size, Matrigel 100 Ag/cm) were rehydrated with 100 AL medium. Then, 2 10 cells in 200 AL serum-free DMEM FIGURE 6. Effect of PD98059, mithramycin, and curcumin on the anti-invasion of HT1080 cells induced by TGF-h1. Cells were pretreated for 1 hour with various kinase inhibitors (A and D), mithramycin (B and E), or curcumin (C and F) in the presence or absence of TGF-h1 (10 ng/mL). Left, HT1080 cells invaded (A-C) and migrated (D-F) under the membrane after 12 hours; right, invasion and migration rates were determined by counting cells in four microscopic fields per sample. T, TGF-h1; M, mithramycin; C, curcumin. Bars, SD. *, P < 0.05 versus control; **, P < 0.05 versus TGF-h1. TGF-h1 Signaling for Induction of TIMP-1 Mol Cancer Res 2006;4(3). March 2006 217 on April 20, 2017. © 2006 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from supplemented with 0.1% bovine serum albumin were seeded into the upper part of each chamber, whereas the lower compartments were filled with 1 mL of the same medium mentioned above in the presence or absence of TGF-h1 (10 ng/mL) with or without specific kinase inhibitors or TIMP-1. Following incubation for 18 hours at 37jC, noninvaded cells on the upper surface of the filter were wiped out with a cotton swab, and the invaded cells on the lower surface of the filter were fixed and stained with Diff-Quick kit (Fisher Scientific, Pittsburgh, PA). Invasiveness was determined by counting cells in five microscopic fields per well, and the extent of invasion was expressed as an average number of cells per microscopic field. Transwell migration assays were done using the same procedure as for invasion assay, except coating the underside of filters with type I collagen (100 Ag/cm). Western Blot Analysis HT1080 cells were lysed in lysis buffer [20 mmol/L TrisHCl (pH 7.4), 150 mmol/L NaCl, 1 mmol/L EDTA, 1 mmol/L EGTA, 1% Triton X-100, 2.5 mmol/L sodium pyrophosphate, 1 mmol/L h-glycerophosphate, 1 mmol/L Na3VO4, 1 Ag/mL leupeptin, and 1 mmol/L phenylmethylsulfonyl fluoride]. After a brief sonication, the lysates were clarified by centrifugation at 12,000 g for 10 minutes at 4jC, and protein content in the supernatant was measured by the Bradford method. An aliquot (30-50 Ag protein/lane) of the total protein was separated by 10% or 12% SDS-PAGE and blotted to nitrocellulose transfer membrane (0.2 Am; Amersham, Arlington Heights, IL). The membrane was blocked with 5% nonfat skim milk in TBST [20 mmol/L Tris-HCl (pH 7.6), 137 mmol/L NaCl, and 0.01% Tween 20] for 1 hour at room temperature followed by incubation with the primary antibodies. After extensive washing with TBST, the membrane was reprobed with horseradish peroxidase–linked anti-rabbit immunoglobulin, at 1:3,000 diluted in 5% nonfat skim milk in TBST, for 40 minutes at room temperature. Immunoblots were visualized by enhanced chemiluminescence (Amersham) according to the manufacturer’s protocol. To measure the MMP-2, MMP-9, TIMP-1, and TIMP-2 proteins secreted, the conditioned medium from each sample was also subjected to protein analysis. For this purpose, culture medium in each tissue culture dish was collected and concentrated 10-fold using a Centricon 10 microconcentrator (Amicon, Beverly, MA). The concentrates were subjected to