Transforming growth factor-β1 induces matrix metalloproteinase-9 and cell migration in astrocytes: roles of ROS-dependent ERK- and JNK-NF-κB pathways

Background Transforming growth factor-β (TGF-β) and matrix metalloproteinases (MMPs) are the multifunctional factors during diverse physiological and pathological processes including development, wound healing, proliferation, and cancer metastasis. Both TGF-β and MMPs have been shown to play crucial roles in brain pathological changes. Thus, we investigated the molecular mechanisms underlying TGF-β1-induced MMP-9 expression in brain astrocytes. Methods Rat brain astrocytes (RBA-1) were used. MMP-9 expression was analyzed by gelatin zymography and RT-PCR. The involvement of signaling molecules including MAPKs and NF-κB in the responses was investigated using pharmacological inhibitors and dominant negative mutants, determined by western blot and gene promoter assay. The functional activity of MMP-9 was evaluated by cell migration assay. Results Here we report that TGF-β1 induces MMP-9 expression and enzymatic activity via a TGF-β receptor-activated reactive oxygen species (ROS)-dependent signaling pathway. ROS production leads to activation of extracellular signal-regulated kinase 1/2 (ERK1/2) and c-Jun-N-terminal kinase (JNK) and then activation of the NF-κB transcription factor. Activated NF-κB turns on transcription of the MMP-9 gene. The rat MMP-9 promoter, containing a NF-κB cis-binding site, was identified as a crucial domain linking to TGF-β1 action. Conclusions Collectively, in RBA-1 cells, activation of ERK1/2- and JNK-NF-κB cascades by a ROS-dependent manner is essential for MMP-9 up-regulation/activation and cell migration induced by TGF-β1. These findings indicate a new regulatory pathway of TGF-β1 in regulating expression of MMP-9 in brain astrocytes, which is involved in physiological and pathological tissue remodeling of central nervous system.


Background
Matrix metalloproteinases (MMPs) are a large family of zinc-dependent endopeptidases that play an important role in the turnover of extracellular matrix (ECM) and function in physiological and pathological processes [1]. In the central nervous system (CNS), MMPs, and MMP-9 especially, are implicated in development, morphogenesis, wounding healing, neurite outgrowth, and immune cell migration [2]. In addition, they also participate in the pathogenesis of several CNS diseases such as stroke, Alzheimer's disease, neuroinflammation, and malignant glioma [3]. Among members of the MMP family, MMP-9 has been shown to be elevated in various brain disorders [4][5][6]. Moreover, several pro-inflammatory mediators such as interleukin-1β (IL-1β), lipopolysaccharide, bradykinin (BK), and oxidized lowdensity lipoprotein (oxLDL) can induce MMP-9 expression and activity in cultured rat astrocytes [7][8][9][10], indicating that the expression and activation of MMP-9 may be regulated during brain injuries and inflammation.
Transforming growth factor-β (TGF-β) is a multifunctional cytokine that regulates a broad diversity of physiological and pathological processes, including tissue wound healing, inflammation, cell proliferation, differentiation, migration, and extracellualr matrix (ECM) synthesis [11][12][13]. Accordingly, TGF-β family members play an important role in early embryogenesis and in the homeostasis of adult tissues. However, several lines of evidence show that lack of coordination of TGF-βdependent signaling often leads to a number of human diseases, including fibrosis [14,15], cancer [16,17], and autoimmune diseases [18]. Moreover, TGF-β is a key immune system modulator, TGF-β1 especially, that may have both pro-and anti-inflammatory effects in immune system depending on the cell type (11)(12)(13). Within the CNS, all three isoforms of TGF-βs family, i.e. TGF-β1, -β2, and -β3, are produced by both glial and neural cells [19]. Previous reports have suggested a relationship between increased TGF-β1 levels and cerebral ischemic injury [20,21]. Following CNS injury, elevated TGF-β levels in astrocytes has been proven to be associated with astrocytic scar formation [22]. Emerging evidence has also demonstrated that TGF-β1 is a crucial mediator in the pathogenesis of several CNS disorders, such as in organization of glial scars in response to injury and in several neurodegenerative disorders [11,15,23].
TGF-βs binds to two serine/threonine kinase receptors which consist of TGF-βRI and TGF-βRII. When a ligand binds, TGF-βRII phosphorylates TGF-βRI and activates Smad-dependent intracellular signaling pathways and thus leads to expression of several genes [24][25][26]. In addition to activation of Smad-dependent pathways, TGF-β can affect several signal transduction pathways in a Smad-independent manner, such as mitogen-activated protein kinases (MAPKs), including extracellularsignal-related protein kinase (ERK1/2), p38 MAPK, and c-Jun N-terminal kinase (JNK) [12,25,27]. In human gingival and skin fibroblasts, both p38 MAPK and Smad3 cooperate in regulating TGF-β-induced MMP-13 expression, whereas ERK1/2 cooperates with Smad3 in regulating connective tissue growth factor expression [25,28,29]. Recently, increasing evidence has attributed the cellular damage in neurodegenerative disorders to oxidative stress that leads to generation of reactive oxygen species (ROS) that are responsible for brain inflammatory disorders and that have deleterious effects during CNS pathogenic processes [30][31][32]. TGF-β can stimulate ROS production, which participates in the expression of diverse genes, such as those for MMPs, in the processes of several human diseases like lung fibrosis [33,34]. However, very little information is available concerning the intracellular pathways involved in the effects of TGF-β1 in brain cells.
Recently, several studies have shown that TGF-β1 can up-regulate MMP-9 expression and activity in several cell types such as human skin [35] and corneal epithelial cells [36], implying a crucial role of TGF-β1 in the regulation of MMP-9 in tissue remodeling and wound healing during physiological and pathological processes. The MMP-9 expression is regulated by various mechanisms such as transcriptional and translational regulation in MMP-9 synthesis. The promoter of MMP-9 has been characterized to possess a series of functional enhancer element-binding sites, such as nuclear factor-B (NF-B) and activator protein-1 (AP-1), but not in MMP-2 promoter [37,38]. In RBA-1 cells, our previous studies have demonstrated that IL-1β and BK can upregulate MMP-9 expression via activation of NF-B [39,40]. However, the possibility of MAPKs and NF-B activation and their roles in MMP-9 up-regulation and cellular function (migration) induced by TGF-β1 in astrocytes (RBA-1 cells) are poorly defined.
In this study, we investigated the molecular mechanisms and the functional responses underlying TGF-β1induced MMP-9 expression in RBA-1 cells. These findings indicate that TGF-β1-induced MMP-9 expression via TGF-β receptors is mediated through a ROS-dependent activation of ERK1/2, JNK1/2, and NF-B pathway, finally leading to cell migration in RBA-1 cells. These results suggest that TGF-β1-induced astrocytic MMP-9 up-regulation might play a key role in physiological and pathological brain tissue remodeling such as wound healing and scar formation.

Rat brain astrocyte culture
RBA-1 cells were used throughout this study. This cell line originated from a primary astrocyte culture of neonatal rat cerebrum and naturally developed through successive cell passages [41]. Staining of RBA-1 with the astrocyte-specific marker, glial fibrillary acid protein (GFAP), showed nearly 95% positive staining. In this study, the RBA-1 cells within 40 passages were used that showed normal cellular morphological characteristics and had steady growth and proliferation in the monolayer system. Cells were cultured and treated as previously described [40]. Primary astrocyte cultures were prepared from the cortex of 6-day-old Sprague-Dawley rat pups as previously described [9]. The purity of primary astrocyte cultures was assessed with the astrocyte-specific marker, GFAP, showing over 95% GFAP-positive astrocytes. The cells were plated on 12well plates and 10-cm culture dishes for MMP gelatin zymography and RT-PCR, respectively. The culture medium was changed every 3 days.

MMP gelatin zymography
After TGF-β1 treatment, the culture medium was collected, mixed with equal amounts of non-reduced sample buffer, and electrophoresed on 10% SDS-polyacrylamide gels containing 1 mg/ml gelatin as a protease substrate. Following electrophoresis, gelatinolytic activity was determined as previously described [40]. Mixed human MMP-2 and MMP-9 standards (Chemicon, Temecula, CA, USA) were used as positive controls. Because cleaved MMPs were not reliably detectable, only proform zymogens were quantified. When inhibitors were used, they were added 1 h prior to the application of TGF-β1. Treatment of RBA-1 cells with the pharmacological inhibitors alone had no significant effect on cell viability determined by an XTT assay (data not shown).

Total RNA extraction and RT-PCR analysis
For RT-PCR analysis of MMP-9 mRNA expression, total RNA was extracted from RBA-1 cells stimulated by TGF-β1 as previously described [40]. The cDNA obtained from 1 μg total RNA was used as a template for PCR amplification. Oligonucleotide primers were designed based on Genbank entries for rat MMP-9 and β-actin. The following primers were used for amplification reaction: for MMP-9: 5'-AGTTTGGTGTCGCG-GAGCAC-3' (sense), 5'-TACATGAGCGCTTCCGG CAC-3' (anti-sense); for β-actin: 5'-GAACCCTAAGGC-CAACCGTG-3' (sense), 5'-TGGCATAGAGGTCTT-TACGG-3' (anti-sense). The PCR amplification was performed in 30 cycles at 55°C, 30 s; 72°C, 1 min; 94°C, 30 s. PCR fragments were analyzed on 2% agarose 1X TAE gel containing ethidium bromide and their size was compared to a molecular weight markers. Amplification of β-actin, a relatively invariant internal reference RNA, was performed in parallel, and cDNA amounts were standardized to equivalent β-actin mRNA levels. These primer sets specifically recognize only the genes of interest as indicated by amplification of a single band of the expected size (754 bp for MMP-9 and 514 bp for β-actin) and direct sequence analysis of the PCR product.
Cell migration (wound healing) assay RBA-1 cells were grown to confluence in 6 well plates and starved with serum-free DMEM/F-12 medium for 24 h. The monolayer cells were manually scratched with a pipette tip to create extended and definite scratches in the center of the dishes with a bright and clear field. The detached cells were removed by washing the cells once with PBS. Serum-free DMEM/F-12 medium with or without TGF-β1 was added to each dish as indicated after pretreatment with the inhibitors for 1 h. Images of migratory cells from the scratched boundary were observed and acquired at 48 h with a digital camera and a light microscope (Olympus, Japan). The images shown represent one of three individual experiments.

Preparation of cell extracts and western blot analysis
Growth-arrested RBA-1 cells were incubated with TGF-β1 at 37°C for the indicated time intervals. The cells were washed with ice-cold PBS, scraped, and collected by centrifugation at 45,000 × g for 1 h at 4°C to yield the whole cell extract, as previously described [40]. Samples were denatured, subjected to SDS-PAGE using a 10% (w/v) running gel, and transferred to nitrocellulose membrane. Membranes were incubated overnight using an anti-phospho-ERK1/2, phospho-JNK1/2, phospho-p65, or GAPDH antibody. Membranes were washed with TTBS four times for 5 min each, incubated with a 1:2000 dilution of anti-rabbit horseradish peroxidase antibody for 1 h. The immunoreactive bands were detected by ECL reagents.

Measurement of intracellular ROS generation
The peroxide-sensitive fluorescent probe 2',7'-dichlorofluorescein diacetate (DCF-DA) was used to assess the generation of intracellular ROS [42] with minor modifications. RBA-1 cells in monolayers were incubated with RPMI-1640 supplemented with 5 μM DCF-DA for 45 min at 37°C. The supernatant was removed and replaced with fresh RPMI-1640 media before stimulation with TGF-β1. Relative fluorescence intensity was recorded over time (3 to 60 minutes) by using a fluorescent plate reader (Thermo, Appliskan) at an excitation wavelength of 485 nm and emission was measured at a wavelength of 530 nm.

Plasmid construction, transient transfection, and promoter activity assays
The dominant negative plasmids encoding ERK1 (ΔERK1), ERK2 (ΔERK2), p38 (Δp38), and JNK (ΔJNK) were kindly provided by Dr. K.L. Guan (Department of Biological Chemistry, University of Michigan, MI), Dr. J. Han (The Scripps Research Institute, La Jolla, CA, USA), and C.C. Chen (Department of Pharmacology, National Taiwan University, Taipei, Taiwan), respectively. The rat MMP-9 promoter was constructed as previously described [43] with some modifications. The upstream region (-1280 to +108) of the rat MMP-9 promoter was cloned into the pGL3-basic vector containing the luciferase reporter system. Introduction of a doublepoint mutation into the NF-B-binding site (B domain; GGAATTCC to GGAATTGG) to generate pGL-MMP-9-DB (mt-B-MMP-9) was performed using the following (forward) primer: 5'-GGGTTGCCCCGTGGA-ATTGGCCCAAATCCTGC-3' (corresponding to a region from -572 to -541). The underlined nucleotides indicate the positions of substituted bases. All plasmids were prepared by using QIAGEN plasmid DNA preparation kits. The MMP-9 promoter reporter constructs were transfected into RBA-1 cells using the Lipofetami-ne™RNAiMAX reagent according to the instructions of manufacture (Invitrogen, Carlsbad, CA). The transfection efficiency (~60%) was determined by transfection with enhanced EGFP. To assess promoter activity, cells were collected and disrupted by sonication in lysis buffer (25 mM Tris-phosphate, pH 7.8, 2 mM EDTA, 1% Triton X-100, and 10% glycerol). After centrifugation, aliquots of the supernatants were tested for luciferase activity using a luciferase assay system. Firefly luciferase activities were standardized to b-galactosidase activity.

Analysis of data
All data were estimated using GraphPad Prism Program (GraphPad, San Diego, CA, USA). Quantitative data were analyzed by one-way ANOVA followed by Tukey's honestly significant difference tests between individual groups. Data were expressed as mean ± SEM. A value of P < 0.05 was considered significant.

TGF-b1 induces de novo synthesis of MMP-9 and cell migration in RBA-1 cells
To investigate the effects of TGF-β1 on MMP-9 expression, RBA-1 cells were treated with various concentrations of TGF-β1 for the indicated time intervals. The condition media were collected and analyzed by gelatin zymography. As shown in Figure 1A, TGF-β1 induced MMP-9 expression in a time-and concentration-dependent manner. There was an apparent up-regulation within 16 h and sustained over 24 h. In contrast, the expression of MMP-2 was not significantly changed during incubation with TGF-β1 ( Figure 1A). To further examine whether the increase of MMP-9 expression by TGF-β1 resulted from the induction of MMP-9 mRNA expression, a RT-PCR analysis was performed. The data show that TGF-β1 time-dependently induced MMP-9 mRNA expression in RBA-1 cells ( Figure 1B), whereas the expression of a housekeeping gene β-actin (as an internal control) mRNA was not changed. There was a significant increase in MMP-9 mRNA within 4 h and sustained over 24 h during the period of observation. Moreover, to determine whether the TGF-β1-induced MMP-9 expression is dependent on de novo protein synthesis, the cells were exposed to TGF-β1 in the absence or presence of actinomycin D (Act. D) or cycloheximide (CHI) at a dose known to inhibit transcription or protein synthesis [44], respectively. The results show that TGF-β1-induced MMP-9 expression was significantly attenuated by pretreatment with either Act. D ( Figure 1C, upper panel) or CHI ( Figure 1C, lower panel) in a concentration-dependent manner. Moreover, TGF-β1-induced MMP-9 mRNA accumulation was attenuated by pretreatment with Act.D but not with CHI ( Figure 1D). Moreover, to demonstrate the functional activity of MMP-9 expression induced by TGF-β1, we evaluated in vitro cell migration of RBA-1 by a cell migration assay. After 48 h of TGF-β1 incubation, the images show that TGF-β1-enhanced cell migration was blocked by pretreatment with the inhibitor of MMP-2/9 activity (2/9i) ( Figure 1E), suggesting that up-regulation of MMP-9 and its activity are required for enhancing RBA-1 cell migration induced by TGF-β1.
TGF-b1 induces MMP-9 expression and cell migration via a TGF-b type I receptor SB431542, a selective inhibitor of TGF-β Type I receptor (TGF-βRI), has been shown to abrogate TGF-β1mediated expression of several genes in different cell types [45][46][47]. Thus, we examined whether TGF-β1 induced MMP-9 expression via TGF-βRI, a selective TGF-βRI antagonist SB431542 was used for this purpose. The data reveal that blockade of TGF-βRI by SB431542 attenuated both TGF-β1-induced MMP-9 protein ( Figure 2A) and mRNA ( Figure 2B) expression. Moreover, the involvement of TGF-βRI in TGF-β1induced cell migration was characterized by a cell migration assay. The image data show that pretreatment with SB431542 significantly attenuated TGF-β1enhanced cell migration ( Figure 2C). These results demonstrate that TGF-βRI-mediated MMP-9 induction is essential for enhancing RBA-1 cell migration.

Involvement of ROS-dependent ERK1/2 and JNK1/2 pathways in TGF-b1-induced MMP-9 expression
Recently, several reports have demonstrated that increasing ROS production contributes to expression of several genes such as MMP-9 in different cell types [50][51][52]. To examine whether ROS participated in TGF-β1-induced MMP-9 expression, cells were pretreated with N-acetyl cysteine (NAC, an ROS scavenger) for 1 h and then incubated with TGF-β1 for 16 h. Our results show that pretreatment with NAC reduced TGF-β1induced MMP-9 expression ( Figure 5A) and its mRNA accumulation ( Figure 5B), implying that ROS may contribute to induction of MMP-9 by TGF-β1 in RBA-1 cells. To determine whether generation of ROS was involved in TGF-β1-induced MMP-9 expression in RBA-1 cells, a fluorescent probe DCF-DA was used to determine the generation of ROS in these cells. RBA-1 cells were labeled with DCF-DA, incubated with TGF-β1 for the indicated time intervals, and the fluorescence intensity (relative DCF fluorescence) was measured at 485-nm excitation and 530-nm emission. The data reveal that TGF-β1 stimulated intracellular ROS generation in a time-dependent manner with a maximal response within 10 min and sustained over 60 min ( Figure 5C). Furthermore, TGF-β1-stimulated ROS generation was markedly attenuated by pretreatment with NAC ( Figure 5C), demonstrating that NAC is an efficient ROS scavenger. Next, to determine whether TGF-β1-induced MAPK phosphorylation occurs via a ROS-dependent pathway, we pretreated cells with NAC for 1 h and then incubated them with TGF-β1 for 10 min (for ERK1/2 phosphorylation) or 4 h (for JNK1/ 2 phosphorylation). These results show that pretreatment with NAC (100 μM) significantly reduced TGF-β1-stimulated phosphorylation of ERK1/2 and JNK1/2 in RBA-1 cells ( Figure 5D). In addition, the role of ROS in TGF-β1-induced cell migration was assessed by a cell migration assay. The imaging data show that TGF-β1induced cell migration is attenuated by pretreatment with NAC ( Figure 5E). Furthermore, to demonstrate the direct role of ROS in MMP-9 up-regulation, cells were directly exposed to various concentrations (0.1 and 1 mM) of H 2 O 2 or to combination of 1 mM of H 2 O 2 and 15 ng/ml of TGF-β1 for 24 h. The data show that exposure of cells to H 2 O 2 concentration-dependently induced MMP-9 expression which was blocked by pretreatment with NAC ( Figure 5F), suggesting that ROS play a critical role in up-regulation of MMP-9 in RBA-1 cells. These results suggest that ROS-dependent ERK1/2 and JNK1/2 cascades may contribute to TGF-β1-induced MMP-9 expression and cell migration in RBA-1 cells.

NF-B is required for TGF-b1-induced MMP-9 expression and cell migration in RBA-1 cells
Recent findings have suggested that NF-B is a fundamental transcription factor for induction of several genes such as MMP-9 in astrocytes [9,53]. Moreover, as shown in Figures 1C and 1D, we found that TGF-β1 induces MMP-9 expression at the transcriptional level. The MMP-9 gene promoter with potential binding elements is required for recognition of transcription factors including NF-B [37]. On the other hand, the NF-B family is considered to be an essential regulator of both cellular and inflammatory activities [54]. In astrocytes, TGF-β1 has been shown to stimulate NF-B activation, associated with astrocyte activation during CNS injury [55]. Thus, we examined whether NF-B was required for induction of MMP-9 by TGF-β1 in RBA-1 cells. First, cells were pretreated with the selective NF-B inhibitors, helenalin and Bay11-7082, which block activation of NF-B signaling [56], and then incubated with TGF-β1 for 16 h. The zymographic data show that pretreatment with either helenalin or Bay11-7082 significantly attenuated TGF-β1-induced MMP-9 expression ( Figure 6A) and mRNA accumulation ( Figure 6B), suggesting the involvement of NF-B in TGF-β1-induced MMP-9 expression in RBA-1 cells. To further ensure that activation of NF-B is involved in signaling stimulated by TGF-β1, the phosphorylation of NF-B p65 was determined by western blot using an anti-phospho-p65 NF-B antibody. As shown in Figure 6C, TGF-β1 stimulated phosphorylation of NF-B p65 in a timedependent manner, which was inhibited by pretreatment  (Figure 6D), indicating that TGF-β1-stimulated NF-B signaling is mediated through ROS-dependent ERK1/2 and JNK1/2 cascades in RBA-1 cells. Furthermore, the cell migratory images show that pretreatment with Bay11-7082 inhibited TGF-β1induced RBA-1 cell migration ( Figure 6E). These results demonstrate that NF-B is necessary for TGF-β1induced MMP-9 expression and cell migration in RBA-1 cells.

Involvement of NF-B binding site in regulation of the rat MMP-9 promoter by TGF-b1
We have found that TGF-β1 stimulates activation of NF-B. Next, we examined whether the binding of NF-B to its promoter binding element is essential for TGF-β1-induced MMP-9 gene regulation. The rat MMP-9 promoter luciferase reporter was constructed and its activity was evaluated by a promoter-luciferase activity assay. The rat MMP-9 promoter was constructed into a pGL3-basic vector containing a luciferase reporter system (as illustrated in Figure 7A, upper part; pGL-MMP-9-Luc), which possesses several putative recognition elements for a variety of transcription factors including NF-B family. Thus, to determine the effect of TGF-β1 on the MMP-9 promoter activity, cells were transfected with a pGL-MMP-9-Luc construct and then incubated with TGF-β1 for the indicated time intervals. As shown in Figure 7A, TGF-β1 increased the MMP-9 promoter activity in a time-dependent manner. A maximal response was obtained within 16 h, which was significantly inhibited by pretreatment with the inhibitor of TGF-βRI (SB431542), MEK1/2 (U0126), JNK1/2 (SP600125), NF-B (Bay11-7082), or an antioxidant (NAC) ( Figure 7B). To further ensure that NF-B mediated TGF-β1-induced MMP-9 promoter activity through binding to their regulatory elements within the MMP-9 promoter region, wild-type (WT) MMP-9 promoter, mutated by a single-point mutation of the B binding site (mt-B-MMP-9), was constructed (as indicated in Figure 7C, upper part). As shown in Figure 7C, TGF-β1-stimulated MMP-9 promoter activity was significantly attenuated in RBA-1 cells transfected with mt-B-MMP-9, indicating that the B element is essential for TGF-β1-induced MMP-9 promoter activity. These results further confirm that TGF-β1 induces MMP-9 promoter activity via enhanced NF-B binding to the B element of the MMP-9 promoter in RBA-1 cells.
Finally, using rat primary cultured astrocytes, we also demonstrated that TGF-β1 induces MMP-9 expression in a time-dependent manner ( Figure 8A, upper panel). The condition media were immunoprecipitated with an anti-MMP-9 antibody and analyzed by western blot. As shown in Figure 8A, TGF-β1 induced expression of MMP-9 protein, but not MMP-2 protein, and release into medium (lower panel), indicating that TGF-β1 also induces MMP-9 protein expression and activation in rat primary cultured astrocytes. In addition, pretreatment of rat primary cultured astrocytes with various inhibitors used in RBA-1 cells also significant attenuated TGF-β1induced MMP-9 expression ( Figure 8B). These data demonstrate that, as in RBA-1 cells, TGF-β1-induced MMP-9 expression is also mediated through the same signaling pathways in rat primary culture astrocytes.

Discussion
MMPs contribute to a wide range of biological activities in several CNS diseases, such as stroke, Alzheimer's disease, and malignant glioma [3]. Among MMPs, MMP-9 expression and activation have been shown to be predominantly elevated by various brain injuries [4,6], suggesting that MMP-9 may be a critical molecule in the degradation of ECM and in the pathophysiology of many brain diseases. Another gelatinase, gelatinase A (MMP-2, 72 kDa), is constitutively expressed and its expression is usually not inducible in several cell types including brain cells. Moreover, TGF-β and related peptides are simultaneously produced and released following injury to the human CNS [48,57]. Despite an obviously important role of TGF-β in brain trauma and diseases, the processes by which TGF-β is implicated in astrocytic functions are not completely understood. A well-established rat astroglial cell line (RBA-1) is derived from dissociated cultures of normal neonatal rat brain tissues [41]. According to various analyses in previous studies, the properties of RBA-1 cells are similar to those of normal astrocytes [41]. Thus, we used a culture model of RBA-1 cells to investigate the mechanisms underlying TGF-β1-induced MMP-9 expression and cellular functional responses. These results suggest that in RBA-1 cells, activation of ROS-dependent ERK1/2 and JNK1/2 linking to NF-B, mediated through a TGF-β receptor, is essential for TGF-β1-induced MMP-9 gene expression and cell migration. However, previous studies have demonstrated that MMP-2 can be up-regulated by some stimuli such as TGF-β, but usually participates in development of cancer including growth, invasion, and metastasis [25,58].
Abnormal regulation of MAPKs might be implicated in several CNS disorders [8]. Moreover, TGF-β1 has been reported to act as a multifunctional factor through activation of MAPK cascades in different cell types [19,25,34]. In the present study, we found that ERK1/2 and JNK1/2 are required for MMP-9 expression, since RBA-1 cells transfected with dominant negative ERK1 (ΔERK1), ERK2 (ΔERK2) or JNK (ΔJNK) plasmid led to down-regulation of MMP-9 ( Figures 3E and 4E). These results are consistent with the MMP-9 expression and secretion through ERK1/2 in rat cortical astrocytes [8,40,59] and the induction of MMP-9 by oxidized lowdensity lipoprotein via ERK1/2 and JNK1/2 pathways in RBA-1 cells [10]. Our results are consistent with MMP-9 expression through ERK1/2 in transformed keratinocytes [60]. Previously, many reports have indicated that long-term activation of MAPKs may participate in regulating some cellular functions such as gene expression and cell survival [61,62]. Consistent with these reports, our data show that TGF-β1 stimulated JNK1/2 phosphorylation with a maximal response observed within 4 h ( Figure 4D), suggesting that long-term phosphorylation of JNK1/2 by TGF-β1 may play a sustained role in up-regulation of MMP-9 in RBA-1 cells. Moreover, we have also demonstrated that either p38 MAPK inhibitor SB202190 or dominant negative mutant (Δp38) have no effect on TGF-β1-induced MMP-9 expression (Figures 4A, C, and 4E). However, recent reports have also indicated that TGF-β-induced MMP-9 expression is mediated through activation of p38 MAPK, but not ERK1/2, in MCF10A human breast epithelial cells [63] and in human glioblastoma cells [64]. The different results may be due to diverse cell types and experimental conditions. ROS have been shown to exert a key role in the physiological functions and pathological processes [65][66][67]. In the brain, ROS also extend to the control of vascular tone which is tightly modulated by metabolic activity within neurons [68]. Moreover, increasing oxidative stress (i.e. ROS production) by diverse stimuli can regulate the expression of inflammatory genes linked to pathogenesis of human CNS disorders [65][66][67][68][69]. Recently, increasing evidence attributes the cellular damage in neurodegenerative disorders such as AD to oxidative stress that is due to generation of free radicals implicated in brain inflammatory disorders [30,32]. The effects of TGF-β on ROS generation have been reported to be involved in pathogenesis of tumor progression, connective tissue degradation, and lung disease [33,34,70]. In this study, we found that TGF-β1-induced MMP-9 expression is mediated through ROS generation, since pretreatment with ROS scavenger NAC significantly attenuated TGF-β1-induced responses ( Figures  5A-C). The role of ROS in TGF-β1-induced ERK1/2 and JNK1/2 phosphorylation was further confirmed by pretreatment with NAC ( Figure 5D), suggesting that ROS-dependent activation of ERK1/2 and JNK1/2 is involved in TGF-β1-induced MMP-9 expression in RBA-1 cells. Consistently, many reports have also shown that MAPKs are the down-stream signaling molecules regulated by ROS [34,70]. In addition, we demonstrated that ROS participates in up-regulation of MMP-9 by direct exposure of RBA-1 cells to H 2 O 2 ( Figure 5F). Herein we are the first to establish that intracellular ROS generation contributes to up-regulation of MMP-9 induced by TGF-β1 in RBA-1 cells.
NF-B is a well-known redox-regulated transcription factor for expression of genes induced by diverse stress signals, including mutagenic, oxidative, and hypoxic stresses associated with physiological and pathological events. Our results reveal that TGF-β1-induced MMP-9 expression via NF-B phosphorylation, is mediated through ROS-dependent ERK1/2 and JNK1/2 cascades in RBA-1 cells (Figure 6). The requirement of NF-B . After overnight cotransfection and incubation with TGF-β1 for 16 h, promoter activities of different MMP-9-promoter constructs were measured as relative MMP-9 promoter activity to bgalactosidase. The relative increase in MMP-9 promoter activity induced by TGF-β1 normalized to that of un-stimulated cells is indicated as fold increase. Data are expressed as mean ± SEM of at least three independent experiments (n = 3). *P < 0.05; # P < 0.01, as compared with the cells exposed to vehicle (A) or TGF-β1 (B, C) alone. signaling for MMP-9 induction has been confirmed by in vitro and in vivo studies [40,53], which demonstrate a relationship between MMP-9 expression and enhancing cell motility [9,10] and tumor invasion [53]. In RBA-1 cells and human U87 astrocytoma cells, ERK1/2 has been suggested to be necessary for NF-B activation [40,71]. In addition, accumulating evidence also indicates that TGF-β1-triggered urokinase up-regulation and promotion of invasion is mediated through an ERK1/2-dependent, but not p38 MAPK, activation of NF-B in human ovarian cancer cells [72]. Our previous study of RBA-1 cells has indicated that up-regulation of MMP-9 by BK is mediated through an ERK1/2-dependent NF-B pathway [40]. Recently, the JNK/NF-B cascade has also been shown to participate in TGF-β1induced MMP-9 expression in corneal epithelial cells [36]. These data imply that different MAPK members are differentially involved in NF-B activation in various cell types. These studies are consistent with our presented results in RBA-1 cells challenged with TGF-β1.
Cell migration is essential for the organization and maintenance of tissue integrity and plays a role in embryonic development, wound healing, inflammation, and invasiveness through ECM [73,74]. It has been reported that ROS, MAPKs, and NF-B are involved in MMP-9 up-regulation, which is crucial for regulating cell motility in different cell types [9,56,75,76]. In this study, we demonstrated that TGF-β1-enhanced cell migration is mediated through up-regulation of MMP-9 protein and activity ( Figure 1E) via TGF-β receptor and ROS-dependent NF-B cascade ( Figures 2C, 5E, and 6E). Moreover, to rule out the possibility of cell proliferation in TGF-β1-induced cell migration, hydroxyurea, an inhibitor of DNA synthesis [77], was used to prevent proliferation of astrocytes during the period of observation in the migration (wound healing) assay. Therefore, these results suggest that up-regulation of MMP-9 by TGF-β1 is essential for enhancing migration of RBA-1 cells.

Conclusion
In the study, we have demonstrated that TGF-β1 directly induces MMP-9 expression via TGF-β receptor, ROS-dependent activation of ERK1/2 and JNK1/2, and transcription factor NF-B pathway, which results in the promotion of cell migration in RBA-1 cells. Based on observations from the literature and on our findings, Figure 8C depicts a model for the molecular mechanisms underlying TGF-β1-induced MMP-9 expression and migration of RBA-1 cells. These findings imply that TGF-β1 might play a critical role in the processes of wound healing and scar formation after brain injuries and diseases. Pharmacological approaches suggest that targeting MMP-9 and their upstream signaling components may yield useful therapeutic targets for the treatment of brain injury, tumors, and inflammatory diseases.