Regulation of MMP-9 (type IV collagenase) production and ...

Clinical & Experimental Metastasis 18: 245–252, 2000. © 2000 Kluwer Academic Publishers. Printed in the Netherlands.

245

Regulation of MMP-9 (type IV collagenase) production and invasiveness in gliomas by the extracellular signal-regulated kinase and jun amino-terminal kinase signaling cascades Sajani S. Lakka1 , Sushma L. Jasti2, Anthanassios P. Kyritsis2, W.K. Alfred Yung2 , Francis Ali-Osman1, Garth L. Nicolson3 & Jasti S. Rao1 Departments of 1 Neurosurgery and 2 Neuro-Oncology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas, USA; 3 Insititute for Molecular Medicine, Huntington Beach, California, USA Received 6 October 2000; accepted in revised form 12 October 2000

Key words: MMP-9, type IV collagenase, glioblastoma, JNK, ERK

Abstract Our previous studies have shown that MMP-9 levels are significantly elevated during the progression of human gliomas. In the current study, we examined the role of JNK- and ERK-dependent signaling modules in the regulation of MMP-9 production and the invasive behavior of the human glioblastoma cell line SNB19, in which JNK/ERK1 is constitutively activated. SNB19 cells that were transfected with dominant-negative JNK, MEKK, and ERK1 expression vectors showed reduced MMP-9 promoter activity. In addition, conditioned medium collected from SNB19 cells transfected with these expression vectors showed diminished MMP-9 activity in the presence of phorbol myristate acetate, as determined by gelatin zymography. The cotransfection of SNB19 cells with kinase-deficient c-raf also diminished MMP-9 promoter activity. Further, in the presence of a specific inhibitor of MEKK (PD098059), the Matrigel invasion assay showed the invasiveness of dominant-negative SNB19 cells transfected with dominant-negative JNK1 or ERK1 to be remarkably reduced. In conclusion, our studies showed for the first time that MMP-9 production and the invasive behavior of SNB19 cells are regulated by JNK- and ERK-dependent signaling modules and that interfering with either of the pathways reduces invasiveness.

Introduction Metalloproteinases constitute a large family of zincdependent proteases that are capable of degrading most extracellular matrix (ECM) macromolecules. One of the proteases that contributes to basement membrane dissolution is the collagen-degrading, 92kDa type IV collagenase commonly referred to as MMP-9 [1, 2]. This collagenase is secreted as an inactive precursor that is subsequently activated by the removal of 73 amino acids at the amino terminus of the metalloproteinase [3]. The activity of MMP9 can be blocked by the physiological inhibitors TIMP-1 and TIMP-2, which do so by forming noncovalent complexes with the metalloproteinase [4]. One of the first hints that this metalloproteinase is involved in tumor invasion was the finding that the release of MMP-9 correlated with the metastatic phenotype of transformed rat embryo cells [1]. This was followed by similar findings for rat 13762NF mammary adenocarcinoma [5]. However, the most compelling evidence of the involvement of MMP-9 in tumor invasion came from a study [4], which showed that the overproduction of this metCorrespondence to: Jasti S. Rao, Division of Cancer Biology, Department of Biomedical and Therapeutic Sciences, University of Illinois College of Medicine, P.O. Box 1649, Peoria, IL 61656, USA. Tel: +1-309-671-3445; Fax: +1-309-671-8403; E-mail: [email protected]

alloproteinase in nonmetastatic rat embryo cells conferred a metastatic phenotype on these cells. We [6, 7] and others [8–10] have reported that various protein components of the ECM play significant roles in the migration and invasion of glioma cells. We also have demonstrated that glioblastomas have higher levels of ECM components than do low-grade glioma or normal brain tissue [6]. In addition, the levels of metalloproteinases that degrade various ECM components; are higher during the progression of human gliomas and in glioma cell lines in culture [11–21]. The MMP-9 gene covers 13 exons spanning 7.7 kb and is transcribed into a 2.5-kb mRNA [22]. The 50 flanking sequence, which includes 670 nucleotides, contains the putative binding sites for AP-1, NF-kB, Sp1, and Ap-2 [22]. Stimulation of promoter activity by v-src in fibrosarcoma requires an intact AP-1 motif at position -79 and a GT motif located 54 nucleotides upstream of the transcriptional start site [23]. Stimulation of the MMP-9 gene by tumor necrosis factor-alpha is (TNF-α) is mediated partly through the NF-kB and Sp1 motifs located 600 and 558 nucleotides upstream of the transcriptional start site [22]. In addition, [24] have shown that the mutation of previously undescribed AP1 and PEA3 motifs located at −553 and −540, respectively, severely impairs the ability of ras to induce the MMP-9 gene in an ovarian cancer cell line. Thus, the cis elements of the promoter and the transacting factors regulating MMP-9 pro-

246 duction seem to differ depending on the cell type and the stimulus. Since MMP-9 expression requires DNA motifs recognized by transactivators, which themselves are substrates for the JNKs and ERKs, we designed this study to determine the role of JNK- and ERK-dependent signaling modules in regulating the production of this collagenase in human gliomas. Our results clearly demonstrated that MMP-9 production in the glioblastoma cell line SNB19 is regulated by JNK- and ERK-dependent signaling modules and that interfering with either of the pathways reduces invasiveness. Materials and methods Cell lines and cell culture SNB19 cell lines were purchased from the American Type Culture Corporation (Rockville, Maryland) and maintained in Dulbecco’s modified eagle’s medium (DMEM), F12 medium containing 10% fetal calf serum, and 100 µg/ml streptomycin, and 100 units/ml penicillin in a humidified atmosphere containing 5% CO2 at 37 ◦ C. DNA constructs The TAM-67 vector encoding a c-jun protein lacking the transactivation domain (amino acids 3-122 absent) of the molecule has been described elsewhere [25]. The raf c4 expression vector encodes a mutated c-raf-1 lacking the kinase domain of the serine threonine kinase [26]. The ERK1 construct encodes the ERK1 cDNA in which the conserved lysine at codon 71 was changed to arginine, thus impairing the catalytic efficiency of the enzyme [27]. The dominant negative JNK1 used herein includes the coding sequence in which threonine 183 and thyrosine 185 are substituted with alanine or phenylalanine, respectively [28]. The mutated MEKK cDNA harbors a lysine-to-methionine substitution at residue 432 and lacks the neo fragment from the amino terminus [29]. Transient transfections Glioblastoma cells were co-transfected at 70% confluency with 2 µg/ml MMP-9 reporter constructs, indicated amounts of other plasmids, and 1 µg/ml of an expression vector bearing the β-galactosidase gene with lipofectamine reagent in serum-free medium. This medium was replaced with serum containing medium after 5 h and allowed to incubate for 48 h. The cells were then harvested and lysed by repeated freeze-thaw cycles in a buffer containing 0.25 M Tris-C1 (pH 7.8). Transfection efficiencies were determined by the amount of β-galactosidase activity. Cells were lysed and assayed using a commercial firefly or Renilla luciferase assay system (Promega, Madison, Wisconsin) and luminometer (Monobite 2010; Analytical Luminescence Laboratory, San Diego, California).

S.S. Lakka et al. JNK activity assays JNK activity was determined as described elsewhere [30]. Cell extracts prepared in a buffer containing 1% Triton X100, 25 mM HEPES (pH 7.6), 300 mM NaCl, 1.5 mM MgCl2 , 0.2 mM EGTA, 0.5 mM DTT, 0.5% sodium deoxycholate, 0.1% SDS, 20 mM β-glycerophosphate, 20 µg/ml aprotinin, 20 µg/ml leupeptin, 1 mM PMSF, and 1 mM sodium vanadate were incubated at 4 ◦ C for 2 h with 1.5 µg of anti-JNK antibody Santa Cruz Biotechnology (Santa Cruz, California) (#sc-474) which is cross-reactive with mouse and human JNK1 and JNK2. Protein A agarose beads were then added to the mixture and incubated for an additional 1.5 hr at 4 ◦ C. The beads were rinsed several times as described elsewhere and then incubated in 30 µl of kinase buffer (12.5 mM MOPS- (pH 7.5) 7.5 mM MgCl2 , 12.5 mM β-glycerophosphate, 0.5 mM EGTA, 0.5 mM NaF, 0.5 mM sodium vanadate) with 1 µCi [γ 32 P]ATP, 2 µg GST-c-jun protein (amino acid residues 1–79 of c-jun), and 20 µm cold ATP at 30 ◦ C. The reaction was terminated 20 min later with 2× Laemmli’s buffer. The reaction mixture was heated to 100 ◦ C for 5 min and the beads removed by centrifugation. The supernatant (65 µl) was electrophoresed. The gel was dried and exposed to X-ray film. Gelatin zymography Protein (20 µg) from culture supernatants was denatured in the absence of reducing agent and electrophoresed in a 7.5% polyacrylamide gel containing 0.1% (w/v) gelatin. The gel was incubated at room temperature for 2 h in the presence of 2.5% Triton X-100 and subsequently at 37 ◦ C overnight in a buffer containing 10 mM CaCl2 , 0.15 M NaCl, and 50 mM Tris (pH 7.5). The gel was stained for protein with 0.25% Coomassie blue and photographed on a light box. Proteolysis was detected as a white zone in a dark field. Western blotting For immunoblotting studies, the ERK and JNK cell extracts generated in a phosphate-buffer saline buffer containing 1% NP40, 0.5% sodium deoxycholate, 0.1% SDS, and 1 µM PMSF (R1PA) were denatured in the presence of reducing agent and electrophoresed in a 10.5% SDS-PAGE gel. The resolved proteins were transferred to a nitrocellulose membrane. The membrane was then blocked with a solution containing 3.0% bovine serum albumin and incubated with antibodies for ERKs (#93, Santa Cruz Biotechnology) and JNK1, JNK2 (#571, Santa Cruz Biotechnology), and an antirabbit horseradish-peroxidase conjugate. Reactive proteins were visualized by electrochemiluminescence (Amersham, Arlington Heights, Illinois) according to the manufacturer’s recommendations. Matrigel invasion assays Invasion of SNB19 cells was determined in vitro by measuring the invasion of cells through Matrigel-coated (Collaborative Research, Inc., Boston, Massachusetts) transwell

Regulation of MMP-9 (type IV collagenase) production

247

Figure 1. Repression of MMP-9 promoter activity by a transactivation domain-lacking c-jun. Glioblastoma cells were transiently co-transfected with various amounts of an expression vector encoding a transactivating domain lacking the c-jun protein (TAM-67) or an empty CMV vector with a luciferase reporter gene driven by 670 nucleotides of the MMP-9 regulatory sequence. Each value is the mean ± SD of the results of 5 different experiments (∗ P < 0.001).

Figure 2. Repression of MMP-9 promoter activity by a mutated ERK1 expression vector. (A) Equal amounts of extracted protein from SNB19 cells treated with PMA were subjected to SDS-PAGE and the proteins transferred to nitrocellulose. The filter was probed with an antibody that cross-reacts with human and murine ERK1 and ERK2, and the bands were visualized by ECL. (B) SNB19 cells were transfected with 6 µg of the MMP-9-driven luciferase reporter gene and different concentrations of mutated ERK1 expression vectors or empty vectors (pCEP4), and assays were performed to determine the luciferase activity. Each value is the mean ± SD of the result of 5 different experiments (∗ P < 0.001). (C) Extracted protein from SNB19 cells/serum-free conditioned medium treated with PMA in the presence of the ERK-1 mt and empty vector PCEP4 was subjected to SDS-PAGE containing 0.1% gelatin.

inserts (Costar, Cambridge, Massachusetts) using a previously described procedure [31]. Briefly, transwell inserts with an 8-µm pore size were be coated with Matrigel in cold serum-free DMEM at a final concentration of 0.78 mg/ml. Cells were trypsinized and washed three times with serumfree medium, after which 1 × 106 cells/ml were added in triplicate wells in the presence of various concentrations of PD098059 and ERK1 mt. The lower chamber contained serum-free medium as the chemoattractant. After a 48-h incubation, the filters were stained by Hema stain according to the manufacturer’s recommendations. The cells adhering to the lower surface of the filter were counted under a light microscope. Cells from 10 non-overlapping fields from triplicate filters were counted and the average was used taken for analysis.

Results Repression of MMP-9 promoter activity by a transactivation domain-lacking c-jun To investigate the role of AP-1-binding transcription factors in the regulation of MMP-9 expression, we co-transfected glioblastoma cells with a luciferase reporter gene fused to 670 base pairs of the MMP-9 regulatory sequence (Figure 1) and an expression construct (TAM-67) encoding a c-jun that lacks amino acids 3-to 122. The resultant mutated protein inhibits AP-1-dependent gene expression through a quenching mechanism by inhibiting the function of endogenous Jun and/or fos proteins. Expression of the TAM-67 construct caused significant repression of the MMP-9 promoter activity in a dose-dependent (∗ P < 0.001), which was specific since the empty expression vectors cytomegalovirus (vectors) had no effect on MMP-9 promoter activity (Figure 1).

248 Repression of MMP-9 expression by a mutant ERK-1 expression construct To determine whether the c-raf-ERK pathway is activated in glioblastoma cells, we first examined the activation status of the ERK subgroup of MAPK on glioblastoma cells (SNB19) that produced large amounts of MMP-9 using western blotting. The SNB19 cells were treated with PMA (50 ng/ml), and the cell extracts were subjected to 10% SDS-PAGE, transferred onto nitrocellulose membranes, and probed with an anti-ERK antibody that reacts with both human and murine ERK1 and ERK2. The results of this experiment indicated that SNB19 cells contain constitutively activated ERK1 and ERK2, the amounts of which are higher in the presence of phorbol myristate acetate (PMA) (Figure 2A). We then determined the effect of a kinase-deficient ERK1 [27] on MMP-9 promoter activity and its expression. In this experiments, SNB19 cells were co-transfected with the MMP-9-driven luciferase reporter gene and varying amounts of an expression vector encoding mutant ERK1, or the empty vector (pCEP4). A two-fold molar excess of the ERK1 mutant expression vector was significantly (∗ P < 0.001) inhibited MMP-9 promoter activity by 80%, whereas a four-fold molar excess of the empty vector had no effect on MMP-9 promoter activity (Figure 2B). Figure 2C shows that MMP-9 enzyme activity was inhibited by a kinase-deficient ERK1 even in the presence of PMA, whereas a four-fold molar excess of the empty vector had no effect on MMP-9 enzyme activity. Reduction in MMP-9 enzyme activity and ERK levels by an MEK1-specific inhibitor (PD098059) Activation of the ERKs has been shown to be achieved through MEK1. To determine whether MMP-9 expression was regulated by this MEKI inhibitor, SNB19 cells were treated with different concentrations of an MEK1 inhibitor (PD098059) and the MMP-9 activity and ERK protein by gelatin zymography and western blotting. MMP-9 activity determined (Figure 3A) was present in SNB19 conditioned medium in the presence of PMA but was decreased in a dose-dependent manner with increasing concentrations of PD098059. Figure 3B shows the presence of ERK1 and ERK2 in SNB19 cells, which was increased in the presence of PMA. Conversely, ERK1 and ERK2 levels were decreased in a dose-dependent manner with increasing concentrations of PD098059. Repression of the MMP-9 promoter by a kinase-deficient c-raf We performed two experiments to determine whether c-raf was involved in the activation of the MMP-9 promoter. In the first experiment we transiently co-transfected SNB19 with an MMP-9 promoter-driven luciferase reporter gene and increasing amounts of an expression vector encoding a mutant c-raf (raf C4 mt). The co-transfection of SNB19 cells with raf C4 mt and the MMP-9 promoter-driven luciferase reporter gene led to a dose-dependent reduction in

S.S. Lakka et al. luciferase activity significantly (∗ P < 0.001), with a twofold molar excess of raf-C4 mt producing an 80% reduction in luciferase activity (Figure 4A). The empty expression vector had no effect on MMP-9 promoter activity. In the second experiment, we assessed the effect of the dominant-negative c-raf expression vector on the activity of a CAT reporter gene driven either by three tandem AP-1 reporters upstream of a thymidine kinase minimal promoter (3 × AP-1 pBL CAT) or by the thymidine kinase minimal promoter (pBL CAT). The dose-dependent repression of the AP-1-driven promoter was significant (∗ P < 0.001), evident in cells made to produce the raf c4 mt but not in cells with the empty expression vector (Figure 4B). Repression of MMP-9 by CMV vector encoding a mutated JNK1 JNK proteins immunoprecipitated with an anti-JNK1 and JNK2 were highly efficient at phosphorylating a GST-c-jun (amino acids 1–79) fusion protein, indicating that JNKs and JNK are activated in SNB19 cells (Figure 5A). JNK activity was increased in the presence of PMA. SNB19 were then co-transfected with the MMP-9 promoter-driven luciferase reporter gene and a CMV vector encoding a mutated JNK (JNK mt), in which threonine 183 and tyrosine 185 were substituted for an alanine and an phenylalanine, respectively, thus preventing activation of the JNK [28]. The dominantnegative JNK repressed significantly the (∗ P < 0.001) MMP-9 promoter in a dose-dependent manner (Figure 5B). A four-fold molar excess of the JNK mt over the reporter construct reduced luciferase activity by about 80%, but this did not occur in the cells with the empty expression vector. Repression of MMP-9 promoter and enzyme activity by expression of a dominant-negative MEKK Since our preliminary data indicated that MMP-9 promoter activity in SNB19 cells was partly a function of an AP-1 binding site (at −79) and since dominant-negative mutants of c-raf and the ERKs did not completely abolish the activity of the MMP-9 promoter, we undertook experiments to clarify the contribution of the MEKK-JNK signaling module in the production of this metalloproteinase. To determine the role of MEKK in the constitutive expression of MMP-9, we conducted parallel experiments with a dominant-negative MEKK expression construct in SNB19 cells. Figure 6A shows that MMP-9 activity was inhibited by the mutant MEKK vector even in the presence of PMA. However, a four-fold molar excess of the empty vector had no effect on MMP-9 activity. In particular, Figure 6B shows an approximate 85% decrease in the MMP-9 promoter activity in response to a four-fold molar excess of the mutant MEKK vector compared with activity in response to the empty expression vector (SRα vector). There was a dose-dependent repression of MMP-9 promoter activity in response to an increase on the concentration of the mutant MEKK vector significantly (∗ P < 0.001).


249

Figure 3. Effect of MAPK inhibitor (PD098059) on MMP-9 expression and activity of ERKs, in SNB19 cells. (A) SNB19 cells were treated with PMA (50 µg/ml) and the indicated concentrations of PD098059. Conditioned medium containing an equal amount of protein was run on SDS-PAGE gels containing 0.1% gelatin. (B) Cell extracts were prepared by lysing the cells in RIPA buffer, electrophoresing the extracts on 10% SDS-PAGE gels, transferring them to nitrocellulose membranes, and probing them with anti-ERK antibody, as described in ‘Materials and methods’.

Figure 4. Repression of MMP-9 promoter activity by a dominant-negative c-raf. Cells were co-transfected with a luciferase reporter gene fused to either the MMP-9 promoter (MMP-9 luciferase; (A) or a thymidine kinase minimal promoter (pPBL CAT). (B) Flanked by 3 AP-1 tandem repeats (3×AP-1 pBL CAT) and the indicated amounts of a kinase domain-lacking c-raf (raf c4) or an equivalent amount of an empty expression vector. 2X indicates a two-fold molar excess of raf c4 with respect to the luciferase reporter gene. Cell extracts were normalized for the difference in transfection efficiency and luciferase activity. Each value is the mean ± SD of the result of five different experiments (∗ P < 0.001)

Figure 5. A. JNK kinase activity. Equal amounts (100 µg) of extracted protein were incubated with an anti-JNK antibody. Protein A agarose beads were subsequently added to the mixture after which the beads were rinsed and then incubated with [ϒ 32 -P]ATP-aGST-c-jun (1–79) protein and 20 µM unlabeled ATP. The mixture was heated to 100 ◦ C and centrifuged, and the supernatant was electrophoresed in a 12% polyacrylamide gel. The gel was dried and subjected to autoradiography. B. Repression of MMP-9 promoter activity by of dominant-negative JNK vectors. SNB19 cells were co-transfected with a luciferase reporter gene fused to the MMP-9 promoter (MMP-9 luciferase), the indicated amounts of a kinase-inactive JNK, and its respective empty vector (CMV vector). 2× indicates a two-fold molar excess of the expression vector with respect to the luciferase reporter gene to determine the cell extracts, normalized for differences in the transfection efficiency, were assayed to determine the luciferase activity. Each value is the mean ± SD the results of 5 different experiments (∗ P < 0.001).

250

S.S. Lakka et al.

Figure 6. Repression of MMP-9 enzyme activity and promoter activity by the dominant-negative MEKK vectors. A. Protein extracted from SNB19 cells/serum free medium treated with PMA in the presence of the MEKK-mt and empty vector (SRα vector) was subjected to SDS-PAGE containing 0.1% gelatin. B. SNB19 cells were co-transfected with a luciferase reporter gene fused to the MMP-9 promoter (MMP-9 luciferase) and the indicated amounts of a kinase-inactive MEKK and its respective empty vector (SRα vector). 2× indicates a two-fold molar excess of the expression vector with respect to the luciferase reporter gene. Cell extracts, normalized for differences in the transfection efficiency, were assayed to determine the luciferase activity. Each value is the mean ± SD the reults of 5 different experiments (∗ P < 0.001).

Effect of dominant-negative ERK-1, JNK1, and PD098059 on Matrigel invasion by SNB19 cells Since our data suggested that MMP-9 is regulated by an ERK- and JNK-dependent mechanism, we investigated whether interfering with these pathways influences the invasive capacity of this glioma cell line. Because earlier studies from our laboratory showed that SNB19 cells are highly invasive in vitro, SNB19 cells were transfected with dominant-negative JNK1, ERK1, and MEKK inhibitor and their invasiveness determined in vitro. Figure 7A shows that cell staining was much less intense in the presence of dominant-negative JNK1, ERK1 mt and PD098059 (5 µM and 10 µM) compared with staining in the parental cell line. The percentage of invasion of SNB19 cells in the presence of dominant-negative JNK1, ERK1 mt and PD098059 (5 µM and 10 µM) was 25%, 20%, 15% and 10% respectively (Figure 7B), of that in the parental cell lines and empty vector PCEP4 was considered as 100%. The percent of invasion was significantly decreased in the presence of dominant negative JNK1, ERK1 mt and PD098059 when compared to parental and vector control (∗ P < 0.001). Discussion To our knowledge, this is the first study to demonstrate the role of these two protein serine-threonine kinases, in the regulation of MMP-9 in a glioma cell line. This 92-kDa-type IV collagenase plays a critical role in various physiological and pathological conditions such as tissue remodeling, reproduction, morphogenesis, various connective tissue diseases, and cancer cell invasion and metastasis. Although the increased production of this metalloproteinase in response to cytokines such as TNF-α and EGF is triggered by increased promoter activity, the role of transcriptional activation of this gene in a glioma cell line has not previously been determined. We found that a CAT reporter gene driven by a full-length MMP-9 promoter was strongly activated in the SNB19 glioblastoma cell line compared with

activity in low-grade gliomas. The converse observation that MMP-9 promoter activity was reduced substantially by the co-expression of a construct (TAM-67) that interferes with the function of endogenous Fos and Jun proteins convincingly suggests that Ap-1-binding transcription factors are required for the MMP-9 promoter to be produced in the SNB19 cell line. The synthesis of c-fos is regulated by both the JNK-dependent [32] and ERK-dependent [33] signaling modules via the phosphorylation of p62tef Elk . It has been shown that the serum response element in the fos promoter represents a point of convergence of the JNKand ERK-dependent signaling modules [32]. Increased cfos production drives the formation of jun-fos heterodimers, which are more stable than the pre-existing jun-jun homodimers. Such increased dimer stability results in higher levels of Ap-1 DNA-binding activity. JNK/SAPKs are involved in c-jun induction through the phosphorylation of c-jun and ATF2 which forms a hetrerodimer and increases their transcriptional activity, therby enhancing c-jun transcription and hence c-jun synthesis [34, 35]. The newly synthesized c-jun may combine with newly synthesized c-fos or other proteins such as ATF2 form homodimers, all of which can contribute to increased AP-1 activity. In addition to stimulating of cjun synthesis, the JNK/SAPKs contribute to elevating AP-1 activity by phosphorylating the activation domain of c-jun, thereby enhancing its transcriptional activity. Our studies do not, however, rule out the possibility that the regulation of MMP-9 by ERK and JNK signaling pathways requires the binding of transcription factors to the non-AP-1 motifs necessary for optimal promoter activation in glioblastoma cells. Nevertheless, the observation that MMP-9 promoter activity was reduced substantially by either by co-expression of a construct (TAM-67), which interferes with the ability of endogenous FoS and Jun proteins to transactivate AP-1 controlled genes. The observation that MMP-9 was clearly regulated by JNK1 was both surprising and very interesting. Although JNK1 activity in response to stress is associated with the induction of cell-cycle arrest and apoptosis [36, 37], recent


251

Figure 7. Matrigel invasion assay. SNB19 cells were transfected with dominant-negative ERK1 and JNK1-mt vectors for 5 h. The cells were trypsinized after 12–14 h and allowed to invade through Matrigel-coated 8-µm transwells for 24 h. The cells in the upper chamber were treated with various concentrations of PD098059. The cells in the lower chamber were then stained (A) and quantitated (B) as described in ‘Materials and methods’. Each value is the mean ± SD of 5 different experiments (∗ P < 0.001).

studies have shown that the duration of this MAPK activity can also determine the physiological response of a cell. For example, the transient induction of JNKs triggers a growth enhancement signal, whereas persistent activity induced by ultraviolet light or γ -irradiation induces apoptosis [38,39] have shown that continuous JNK activity confers a growth advantage on EGFRvIII transfectants. It has been reported [40] that the induction of prostaglandin synthase 2 gene expression by v-src was mediated by activation of JNK1 and c-jun transcription. Studies by other groups have shown that treatments with IL-1 [41] and TNF-α [22] also cause an increase in JNK activity. This suggests that the ability of these cytokines to stimulate the production of this collagenase is due to this MAPK subset. Signal transduction pathways distinct from the JNK MAPK cascade may also participate in the activation of this collagenase. This possibility is supported by our finding PD098059, a specific inhibitor of MEK1, inhibited the production of MMP-9 in SNB19 cells. Similarly transient transfection with kinase-deficient dominant-negative ERK1 vectors also decreased MMP-9 promoter activity, suggesting a role for ERK1 in the activation of MMP-9. This also abolished the ability of these cells to invade Matrigel. Transient transfection of the kinase-inactive raf, which is an immediate upstream activator of ERK, also decreased promoter activity, clearly suggesting a role for the Raf/MEK/ERK cascade in the regulation of MMP-9 activity. The finding that multiple signaling pathways are involved in the induction of a single molecule is not new. Earlier studies have shown that all three MAPK cascades are involved in MMP-9 regulation in the ovarian cancer cell line UM-SCC-1 [42, 43]. Recently, it has been reported that using dominant-negative mutants uPA and uPAR production requires ERK1, JNK, and p38 signaling cascades [44]. ERK1, ERK2 and p38 MAPK are also necessary for the signal transduction of basic fibroblast growth factor in endothelial cells [45].

Substantial evidence has come from out and other laboratories showing that MMP-9 plays a major role in glioma invasion. Moreover, there is a positive correlation between increases in MMP-9 levels and the progression of gliomas [12, 13, 46, 47]. Because of the importance of MMP-9 in cancer invasion and metastasis, our study suggests that both ERK and JNK signaling pathaway may be a potential therapeutic target for anticancer treatment,

Acknowledgements We thank Lydia Soto for preparing the manuscript and Beth Notzon for manuscript review. This work was supported by National Cancer Institute grant CA-56792 (to JSR).

References 1.

2.

3.

4.

5.

6.

Bernhard EJ, Muschel RJ, Hughes EN. Mr 92,000 gelatinase release corelates with the metastatic phenotype in transformed rat embryo cells. Cancer Res 1990; 50: 3872–7. Bernhard EJ, Gruber SB, Muschel RJ. Direct evidence linking expression of matrix metalloproteinase 9 (92-kDa gelatinase/collagenase) to the metastatic phenotype in transformed rat embryo cells. Proc Natl Acad Sci USA 1994; 91: 4293–7. Wilhelm SM, Collier IE, Marmer BL et al. SV40-transformed human lung fibroblasts secrete a 92-kDa type IV collagenase which is identical to that secreted by normal human macrophages. J Biol Chem 1989; 264: 17213–21. Goldberg G, Strongin A, Collier IE et al. Interaction of 92-kDa type IV collagenase with the tissue inhibitor of metalloproteinase prevents dimerization, complex formation with interstitial collagenase, and activation of the proenzyme with stromelysin. J Biol Chem 1992; 267: 4583–91. Nakajima M, Welch DR, Wynn DM et al. Serum and Plasma Mr 92, 000 progelatinase levels correlate with spontaneous metastasis of rat 13762NF mammary adenosarcoma. Cancer Res 1993; 53: 5802–7. Chintala SK, Sawaya R, Fuller GN et al. Immunohistolocalization of ECM proteins in human glioblastoma cells in vitro. Cancer Lett 1996; 101: 107–14.

252 7. Chintala SK, Sawaya R, Gokaslan ZL et al. The effect of type III collagen on migration and invasion of human glioblastoma cells in vitro. Cancer Lett 1996; 102: 57–63. 8. Reith A, Bjervig R, Rucklidge GJ. Laminin: A potential inhibitor of rat glioma cell invasion in vitro. Anticancer Res 1994; 14: 1071–6. 9. Koochekpour S, Merzak A, Pilkington GJ. Growth factors and gangliosides stimulate laminin production by human glioma cells in vitro. Neurosci Lett 1995; 186: 53–6. 10. Friedlander DR, Zagzag D, Grondahl-Hansen J et al. Migration of brain tumor cells on extracellular matrix proteins in vitro correlates with tumor type and grade and involves α5 and β1 integrins. Cancer Res 1996; 56: 1939–47. 11. Apodaca G, Rutka JT, Bouhana K et al. Expression of metalloproteinases and metalloproteinase inhibitors by fetal astrocytes and glioma cells. Cancer Res 1990; 50: 2322–9. 12. Rao JS, Steck PA, Mohanam S et al. Elevated levels of Mr. 92,000 type IV collagenase in human brain tumors. Cancer Res 1993; 53: 2208–11. 13. Rao JS, Yamamoto M, Mohanam S et al. Expression and localization of a 92 kDa type IV collagenase in human astrocytomas in vivo. Clin Exp Metastasis 1996; 14: 12–8. 14. Miyagi N, Kato S, Terasaki M et al. Fibroblast growth factor-2 and -9 regulate proliferation and production of matrix metalloproteinase in human gliomas. Inter J Oncol 1998; 12: 1085–90. 15. Sawaya R, Yamamoto M, Mohanam S et al. Expression and localization of 72-kDa type IV collagenase in human malignant gliomas in vivo. Clin Exp Metastasis 1996; 14: 35–42. 16. Sawaya R, Go Y, Kyritsis AP et al. Elevated levels of Mr. 92,000 type IV collagenase during tumor growth in vivo. Biochem Biophys Acta 1998; 251: 623–36. 17. Tonn JC, Kerkau S, Hanke A et al. Effect of synthetic matrixmetalloproteinase inhibitors on invasive capacity and proliferation of human malignant gliomas in vitro. Intern J Cancer 1999; 80: 764–72. 18. Rossi M, Rooprai HK, Maidment SL et al. The influence of sequential, in vitro passages on secretion of matrix metalloproteinases by human brain tumor cells. Anticancer Res 1996; 16: 121–8. 19. Uhm JH, Dooley NP, Villemure JG et al. Glioma invasion in vitro: regulation of matrix metalloprotease-2 and protein kinase C. Clin Exp Metastasis 1996; 14: 421–33. 20. Uhm JH, Dooley NP, Villemure JG et al. Mechanisms of glioma invasion: role of matrix-metalloproteinases. Can J Neuro Sci 1997; 24: 3–15. 21. Yoshida D, Piepmeir JM, Bergenheim T et al. Suppression of matrix metalloproteinase-2-mediated cell invasion in U87MG, human glioma cells by anti-microtubule agent. Br J Cancer 1998; 77: 21–5. 22. Sato H, Seiki M. Regulatory mechanism of 92-kDa type IV collagenase gene expression which is associated with invasiveness of tumor cells. Oncogene 1993; 8: 395–405. 23. Sato H, Kita M, Seiki M. v-Src activates the expression of 92-kDa type IV collagenase gene through the AP-1 site and the GT box homologous to retinoblastoma control elements. J Biol Chem 1993; 268: 23460–8. 24. Gum R, Lengyel E, Juarez J et al. Stimulation of 92 kDa type collagenase promoter activity by ras is MEK1-independent and requires multiple transcription factor binding sites including closely spaced PEA3/ets and AP1 sequences. J Biol Chem 1996; 271: 10672–82. 25. Brown PH, Alani R, Preis LH et al. Suppression of oncogene-induced transformation by a deletion mutant of c-jun. Oncogene 1993; 8: 877– 86. 26. Bruder JT, Heidecker G, Rapp UR. Serum-, TPA- and Ras-induced expression from Ap-1/Ets-driven promoters requires Raf-1 kinase. Genes Dev 1992; 6: 545–56. 27. Frost JA, Geppert TD, Cobb MH et al. A requirement for extracellular signal-regulated kinase (ERK) function in the activation of AP-1 by HA-ras, phorbol 2 myristat 13 acetate and serum. Proc Natl Acad Sci USA 1996; 91: 3844–8.

S.S. Lakka et al. 28.

29.

30.

31.

32.

33.

34.

35.

36. 37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

Derijard B, Hibi M, Wu I et al. JNK1: A protein kinase stimulated by uv light and Ha-ras that binds and phosphorylates the c-jun activation domain. Cell 1994; 76: 1025–37. Minden A, Lin A, McMohan M et al. Differential activation of ERK and JNK mitogen-activated protein kinases by raf-1 and MEKK. Science 1994; 266: 1719–22. Coso OA, Chiariello M, Yu JC et al. The small GTP-binding proteins Rac1 and Cdc42 regulate the activity of the JNK/SAPK signaling pathway. Cell 1995; 81: 1137–46. Mohanam S, Sawaya R, McCutcheon I et al. Modulation of in vitro invasion of human glioblastoma cells by urokinase-type plasminogen activator receptor antibody. Cancer Res 1993; 53: 4143–7. Whitmarsh AJ, Shore P, Sharrocks AD et al. Integration of MAP kinase signal transduction pathways at the serum response element. Science 1995; 269: 403–7. Gille H, Sharrocks AD, Shaw PE. Phosphorylation of transcription factor p62tcf by MAP kinase stimulates ternary complex formation at c-fos promoter. Nature 1992; 358: 414–7. Gupta S, Campbell D, Derijard B et al. Transcription factor ATF2 regulation by the JNK signal transduction pathway. Science 1995; 267: 389–93. Van Dam H, Wilhelm D, Herr I et al. ATF2 is preferentially activated by stress-activated protein kinases to mediate c-Jun induction in response to genotoxic agents. EMBOJ 1995; 4: 1798–811. Xia Z, Dickens M, Raingeaud J et al. Opposing effects of ERK and JNK-p38 MAP Kinases on apoptosis. Science 1995; 270: 1326–31. Verheij M, Bose R, Lin XH et al. Requirement for ceramide initiated SAPK/JNK signaling in stress-induced apoptosis. Nature 1996; 380: 75–9. Chen YR, Wang X, Templeton D et al. The role of c-Jun-terminal kinase JNK in apoptosis induced by ultraviolet C and ϒ radiation. J Biol Chem 1996; 271: 31929–36. Antonyak MA, Moscatello DK, Wong AJ. Constitutive activation of c-Jun N-terminal kinase by a mutant epidermal growth factor receptor. J Biol Chem 1998; 273: 2817–22. Xie W, Herschman HR. v-src induced prostglandin synthase 2 gene expression by activation of the c-Jun terminal kinase and the c-Jun transcription factor. J Biol Chem 1995; 270: 27622–8. Hurwitz A, Dushnik M, Mayer M et al. Cytokine-mediated regulation of rat ovarian funciton: interleukin-1 stimulates the accumulation of a 92-kilodlton gelatinase. Endocrinology 1993; 132: 2709–14. Gum R, Wang H, Lengyel E et al. Regulation of 92 KDa type IV collagenase exopresion by the Jun aminoterminal kinase and the extracellular signal-regulated kinase-dependent signaling cascades. Oncogene 1997; 14: 1481–73. Simon C, Goepfert H, Boyd D. Inhibition of the p38 mitogenactivated protein kinase by SB203580 blocks PMA-induced Mr 92,000 type IV collagenase secretion and in vitro invasion. Cancer Res 1998; 58: 1135–9. Huang S, New L, Pan Z et al. Urokinase plasminogen activator/urokinase-specific surface receptor expression and matrix invasion by breast cancer cells requires constitutive p38alpha mitogenactivated protein kinase activity. J Biol Chem 2000; 275: 12266–72. Tanaka K, Abe M, Sato Y. Role of Extracellular signal-regulated kinase 1/2 and p38 mitogen-activated protein kinase in the signal transduction of basic fibroblast growth factor in endothelial cells during angiogenesis. Jpn J Cancer Res 1999; 90: 647–54. Lampert K, Machein U, Machein MR et al. Expression of matrix metalloproteinases and their tissue inhibitors in human brain tumors. Am J Pathol 1998; 153: 429–37. Forsyth PA, Wong H, Liang TD et al. Gelatinase-A (MMP-2), gelatinase-B (MMP-9) and membrane type matrix metalloproteinase1 (MT1-MMP) are involved in different aspects of the pathophysiology of malignant gliomas. Br J Cancer 1999; 79: 1828–35.