Many brain cells secrete MMPs (matrix metalloproteinases), and increased or misregulated MMP levels are found in neurodegenerative disorders. Here we report that MMP-3 transcription and protein secretion were increased in rat brain astrocytes stimulated with lipopolysaccharide, gangliosides or interferon-γ. Sequential deletion of the MMP-3 promoter revealed that sequences between −0.5 kb and the start codon were crucial for the transcriptional induction of MMP-3. In addition, experiments using pharmacological inhibitors of individual mitogen-activated protein kinases revealed that MMP-3 induction and promoter activity involved Jun N-terminal kinase, a representative upstream signal of AP-1 (activator protein-1). Sequence analyses of the region of the MMP-3 promoter 500 bp from the start codon indicated the presence of three AP-1 binding sequences. Among them, electrophoretic-mobility-shift assays as well as site-directed mutagenesis of individual AP-1 sequences revealed that distal and middle, but not proximal, sequences largely mediated its induction. Together, these results indicate that AP-1 could control MMP-3 induction in brain astrocytes and that its regulation through specific AP-1 elements could be exploited in the treatment of brain pathologies in which increased expression of MMP-3 plays crucial roles.
MMPs (matrix metalloproteinases) are zinc-dependent endopeptidases that are mainly involved in turnover and proteolytic degradation of the extracellular matrix [1,2]. So far, more than 25 members of the MMP family have been identified, and they can be divided into four groups on the basis of their domain structure, substrate specificity and cellular localization. These four groups are collagenases (MMP-1, -8 and -13), stromelysins (MMP-3, -7, -10, -11 and -12), gelatinases (MMP-2 and -9) and membrane-type MMPs . Among them, MMP-3 (stromelysin-1) plays an important role, because it has a broad substrate specificity and can activate other MMPs and procollagenase [4,5].
As in peripheral tissues, many brain cells can secrete MMP-3, the synthesis and repression of which depend on several inducible inflammatory stimulation or anti-inflammatory suppressive agents [6–10]. Increased or misregulated levels of MMP-3 have been reported in several acute and chronic neurodegenerative disorders, including stroke, Alzheimer's disease, Parkinson's disease and multiple sclerosis [9,11,12]. Similar to other secreted proteinases, the catalytic activity of MMP-3 is regulated at four stages, namely transcription, compartmentalization, proenzyme activation and enzyme inactivation by natural inhibitors [2,13]. The production of MMP-3 is mainly regulated at the level of transcription. The expression of MMP-3 in unstimulated cells is low, but it is induced by a variety of extracellular stimuli, including mitogenic growth factors, cytokines and contact with the extracellular matrix [14–16]. Owing to the similarities in structure across the MMP family, it is important to distinctively control transcription according to cell type or different stimulators. Although several transcription factors, including NF-κB (nuclear factor κB), the transcription factor ETS-1 (E26 transformation-specific-1) and AP-1 (activator protein-1), are reported to be involved in the transcriptional induction of MMP-3 [14,17–19], the exact mechanisms of transcriptional regulation under particular conditions are largely unknown.
In the present study we investigated the mechanisms involved in the transcriptional induction of MMP-3 in activated rat brain astrocytes. Sequential deletion of the promoter showed that sequences between −0.5 kb and the start codon, a region that contains three AP-1 binding sequences, are crucial in the regulation of LPS (lipopolysaccharide)-, ganglioside- and IFN-γ (interferon-γ)-induced MMP-3 transcription. Mutational studies as well as EMSAs (electrophoretic-mobility-shift assays) showed that two of these sequences (distal and middle sequences from the start codon) act through JNK (Jun N-terminal kinase) and are indispensable in MMP-3 induction in stimulated primary astrocytes. MMP-3 is thought to play a crucial role in the progression and aggravation of neurodegenerative disease [3,9,10]; thus efficient regulation of its expression through specific AP-1 signals could be exploited to control neurodegenerative diseases such as Alzheimer's disease and stroke.
MATERIALS AND METHODS
LPS from Salmonella enterica and curcumin were purchased from Sigma Aldrich (St Louis, MO, U.S.A.). Bovine brain ganglioside mixture was purchased from Matreya (Pleasant Gap, PA, U.S.A.). IFN-γ and MAPK (mitogen-activated protein kinase) inhibitors (PD98059, SB203580 and SP600125) were purchased from Calbiochem (San Diego, CA, U.S.A.). Oligofectamine™, Lipofectamine™ Plus reagent, and β-Gal Staining Kit were purchased from Invitrogen Life Technologies (Carlsbad, CA, U.S.A.).
Cell culture and stimulation
Primary astrocytes were cultured from the cerebral cortices of 1-day-old Sprague–Dawley rats as described previously [20,21]. Briefly, the cortices were triturated into single cells in minimal essential medium (Sigma) containing 10% (v/v) fetal bovine serum (Hyclone) and were plated into 75-cm2 T-flasks. After 2–3 weeks, astrocytes were trypsinized and plated on to six-well plates or 60-mm-diameter dishes for use in subsequent experiments. We stimulated primary astrocytes with LPS, gangliosides or IFN-γ, all of which are well-known glial activators, as previously reported [22,23].
RT-PCR (reverse-transcription PCR) analysis
Total RNA was isolated using RNAzol B (Intron, Seoul, Korea) and cDNA was prepared using avian-myeloblastosis-virus reverse transcriptase (Promega, Madison, WI, U.S.A.). PCR involved 32 cycles of 94 °C for 30 s, 56–58 °C for 30–60 s and 68–72 °C for 90 s. Oligonucleotide primers were produced at Bioneer (Seoul, Korea). The PCR primer sequences were: forward 5′-ACCCAAATGGAGGAAAAACC-3′ and reverse 5′-CTTCAGCATTGGCTGAGTGA-3′ for MMP-3; forward 5′-TCCCTCAAGATTGTCAGCAA-3′ and reverse 5′-AGATCCACAACGGATACATT-3′ for GAPDH (glyceraldehyde-3-phosphate dehydrogenase).
Cells were washed twice with cold PBS and then lysed in ice-cold modified RIPA (radioimmunoprecipitation assay) buffer (50 mM Tris/HCl, pH 7.4, 1% Nonidet P40, 0.25% sodium deoxycholate, 150 mM NaCl and 1 mM Na3VO4) containing protease inhibitors (2 mM PMSF, 100 μg/ml leupeptin, 10 μg/ml pepstatin, 1 μg/ml aprotinin and 2 mM EDTA). The lysates were centrifuged at 12000 g for 10 min at 4 °C and the supernatant was collected. Conditioned media from control or stimulated rat astrocytes were concentrated using an Amicon Centriprep apparatus (Millipore, Billerica, MA, U.S.A.) at 6000 g for 90 min. Proteins were then precipitated with trichloroacetic acid (Sigma), separated by SDS/PAGE and transferred to nitrocellulose membranes. The membranes were incubated with primary antibodies against phospho-MAPKs [ERK (extracellular-signal-regulated kinase), JNK, and p38; Cell Signaling Technology Inc., Danvers, MA, U.S.A.] or MMP-3 (Chemicon, now part of Millipore), and peroxidase-conjugated secondary antibodies (Vector Laboratories, Burlingame, CA, U.S.A.). The proteins were then visualized using an enhanced-chemiluminescence system.
Cells were harvested and suspended in 9×packaged cell vol. of a hypo-osmotic solution (10 mM Hepes, pH 7.9, 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM dithiothreitol and 0.5 mM PMSF) containing 0.5% Nonidet P40. Cells were centrifuged at 500 g for 10 min at 4 °C. The nuclear fractions were resuspended in a lysis buffer [20 mM Hepes, pH 7.9, 20% (v/v) glycerol, 0.4 M NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol and 1 mM PMSF], incubated on ice for 60 min with occasional gentle shaking and centrifuged at 12 000 g for 20 min. EMSA was performed for 30 min on ice in a volume of 20 μl containing 4 μg of nuclear protein extract in a reaction buffer containing 8.5 mM EDTA, 8.5 mM EGTA, 8% (v/v) glycerol, 0.1 mM ZnSO4, 50 μg/ml poly(dI-dC), 1 mM dithiothreitol, 0.3 mg/ml BSA, 6 mM MgCl2 and γ-32P-radiolabelled oligonucleotide probe (3×104 c.p.m.), with or without a 20-fold excess of unlabelled probe. In supershift experiments, protein extracts were incubated with 2 μg of antibodies against the transcriptional activators c-Fos and c-Jun (Santa Cruz Biotechnology) for 30 min prior to the addition of 32P-labelled probe. DNA–protein complexes were separated on 6% (w/v) polyacrylamide gels in Tris/glycine buffer, pH 8.3. The dried gels were exposed to X-ray film. Double-stranded oligonucleotides used were: forward 5′-TATGAGTCAGTTTGCGGGTG-3′ and reverse 5′-ATACTCAGTCAAACGCCCAC-5′ for distal AP-1; forward 5′-TGGAGATTAATCACCATTCG-3′ and reverse 5′-AGCTCTAATTAGTGGTAAGC-3′ for middle AP-1; and forward 5′-AGCTTTGACTTCTGGAAGTT-3′ and reverse 5′-TCGAAACTGAAGACCTTCAA-3′ for proximal AP-1 (Bioneer, Seoul, Korea). Radiolabelled probes were prepared with 40 μCi of [γ32P]ATP and T4 polynucleotide kinase (Promega, Madison, WI), and purified using Sephadex G-25 Quick Spin Columns (Roche, Mannheim, Germany).
Rat MMP-3 promoter reporter constructs
Three promoter constructs of rat MMP-3 were generated (with sizes of 2.0, 1.1, and 0.5 kb). The 5′-flanking region of the rat MMP-3 gene −(2000/+1) (2000 bp MMP-3 promoter region from start codon) was prepared using rat genomic DNA as a template. To manufacture the constructs, forward primers with a Mlu1 restriction site (underlined in the sequences below) and reverse primers with a Xho1 restriction site (indicated by broken lines in the sequences below) were used: forward 5′-ACGCGTATCTCCTTTTAATATATGGGGTAC-3′ for the 0.5-kb construct; forward 5′-ACGCGTCTCCTGAGAGAATCTGCCAGA-3′ for the 1.1-kb construct; forward 5′-ACGCGTGAACTACCTAAGGACCCAGCT-3′ for the 2.0-kb construct; and reverse 5′-CTCGAGCCTTTCATTTCCACTGGCTTCC-3′ for all three constructs. The PCR products were digested with Mlu1 and Xho1 restriction enzymes and subcloned into the Mlu1/Xho1-digested luciferase reporter plasmid pCR2.1 vector (Invitrogen, Carlsbad, CA, U.S.A.). The insert fragments were ligated into the pGL3 basic vector (Promega). The MMP-3 2.0 kb promoter-luciferase, MMP-3 1.1 kb promoter-luciferase and MMP-3 0.5 kb promoter-luciferase reporter vectors were generated and analysed using luciferase assays.
Transfection and luciferase assay
Cells plated on 35-mm-diameter dishes were transiently transfected using Lipofectamine™ Plus reagent (Invitrogen) as described previously . To normalize the variations in cell number and transfection efficiency, clones were co-transfected with pCMV-β-galactosidase for 12 h. Luciferase activity was measured using a luminometer (PerkinElmer Vitor3) and normalized by β-galactosidase activity (measured at A420).
PCR-primer mutagenesis was used to generate mutant constructs of the 0.5 kb promoter. Briefly, oligonucleotides with three altered bases were generated: 5′-ACTGTATTCAGCTTTGGTCTCTGGAAGTTCTTTG-3′ for distal AP-1 mutagenesis (dAP-1 MUT) and 5′-CATTTCCTGGAGATTAACAGCCATTCGCTTTGCA-3′ for middle AP-1 mutagenesis (mAP-1 MUT). The PCR products were cloned and the dAP-1 MUT and mAP-1 MUT fragments were cleaved using Xho1 and Mlu1 restriction enzymes. The mutant constructs were produced using the QuikChange® Site-Directed Mutagenesis Kit (Stratagene).
Luciferase activities of the MMP-3 promoter constructs were reported as means±S.D. They were analysed by one-way ANOVA and post-hoc comparisons (Student–Newman–Keuls) were performed using the Statistical Package for Social Sciences 10.0 (SPSS Inc., Chicago, IL, U.S.A.).
Increase in MMP-3 transcription and protein secretion by LPS, IFN-γ and gangliosides in primary astrocytes
To determine the involvement of MMP-3 in rat brain inflammation, we tested the effects of glial stimulators on MMP-3 transcription in cultured rat brain primary astrocytes. To examine MMP-3 transcript levels, we treated cells with LPS (100 ng/ml), gangliosides (50 μg/ml) or IFN-γ (10 units/ml) for 3 h, and the isolated total RNA was analysed by RT-PCR. MMP-3 transcription was induced with all three glial stimulators (Figure 1A). To measure the levels of MMP-3 secreted into the medium, cells were treated with LPS, gangliosides or IFN-γ (at the concentrations mentioned above) for 24 h, then the conditioned media were collected, concentrated and analysed by Western-blot assays using an antibody against MMP-3. Increased MMP-3 protein secretion was induced with all three glial stimulators, a finding consistent with the results from the RT-PCR-based analysis (Figure 1B).
MMP-3 transcript and protein secretions are increased by glial activators in rat primary astrocytes
Sequential deletion analysis of MMP-3 promoter
To identify the mechanisms involved in the transcriptional regulation of MMP-3, we first determined which regions of its promoter are necessary for stimulator-induced transcription, using sequential deletion of the promoter. We generated three MMP-3 promoter constructs, each of which contained the −2.0, the −1.1 or the −0.5 kb region (Figure 2A). Rat primary astrocytes were transfected with a luciferase reporter gene (pGL3) conjugated with one of the MMP-3 promoter regions. After cells were treated with LPS (100 ng/ml), gangliosides (50 μg/ml) or IFN-γ (10 units/ml) for 36 h, cell extracts were obtained and analysed using the luciferase assay. In cells transfected with the 0.5 kb and 2.0 kb reporter genes, treatment with the stimulators induced marked increases in MMP-3 transcription; however, MMP-3 transcription was not increased significantly in cells that had been transfected with the 1.1 kb promoter reporter construct (Figure 2B). These results indicate that the region −0.5 kb from the start codon is necessary for transcriptional induction of MMP-3. However, suppression of luciferase activity in cells transfected with the 1.1 kb sequence indicates that some factors bound to this region have negative effects on MMP-3 transcription. There was a small additional increase in luciferase activity in cells transfected with the 2.0 kb region compared with those transfected with the 0.5 kb region.
Induction of MMP-3 transcription is mainly regulated by the 0.5 kb promoter region
JNK involvement in MMP-3 transcriptional induction
We then investigated the signalling pathway involved in MMP-3 induction. MMP-3 expression has been reported to be regulated through MAPKs; hence, we tested the effects of pharmacological inhibitors of the MAPKs ERK, p38 and JNK on MMP-3 expression. Rat primary astrocytes were pretreated with one of the MAPK inhibitors for 1 h, stimulated with LPS (100 ng/ml), gangliosides (50 μg/ml) or IFN-γ (10 units/ml) for 3 or 36 h, and then analysed by RT-PCR or luciferase assays. RT-PCR-based experiments showed that MMP-3 induction was suppressed with SP600125 (a JNK inhibitor), but not with PD98059 or SB203580 (ERK and p38 inhibitors respectively) (Figure 3A). The results of the luciferase assays using the 2.0 kb sequences were in agreement with those from RT-PCR analyses, showing that transcriptional induction by stimulators was markedly suppressed by SP600125, but not by PD98059 or SB203580 (Figure 3B). With the 0.5 kb sequences, luciferase activities were also markedly suppressed by SP600125. In contrast with the results with the 2.0 kb sequence, luciferase activities were reduced slightly with PD98059 and SB203580 in the 0.5 kb-transfected cells (Figure 3C). Western-blot analyses using phospho-specific antibodies against individual MAPKs revealed that ERK and p38, in addition to JNK were activated in 15 or 30 min in LPS-, ganglioside- or IFN-γ-stimulated astrocytes (Figure 3D, panel a). The increase in phosphorylation was blocked with pharmacological inhibitors of respective MAPKs (Figure 3D, panel b).
MMP-3 transcription and promoter activity are inhibited by JNK inhibitors
Sequence analyses of MMP-3 promoter
To determine which transcription factors are involved in MMP-3 transcriptional induction downstream of JNK, we first analysed the MMP-3 promoter. Analysis of the 500 bases flanking the 5′-end of MMP-3 for transcription-factor binding sites indicated the presence of three AP-1 elements that could be associated with the JNK pathway: distal (−389 to −399), middle (−247 to −237), and proximal (−132 to −122) elements, according to their distances from the start codon (Figure 4).
The 0.5 kb promoter sequences of MMP-3 contain three AP-1 elements
NF (nuclear factor)-binding activities of three AP-1 elements in MMP-3 promoter
To investigate the function of the AP-1 elements in the transcriptional induction of MMP-3, we tested the NF-binding activity of individual AP-1 elements in the 0.5 kb promoter region using three distinct AP-1 primers radiolabelled with γ-32P, as described in the Materials and methods section. Rat primary astrocytes were treated with LPS (100 ng/ml), gangliosides (50 μg/ml) or IFN-γ (10 units/ml) for the times indicated, and the nuclear extracts were isolated and analysed by EMSA. The NF-binding activities of the distal (Figure 5A, panel a) and middle AP-1 elements (Figure 5A, panel b) were increased after the indicated times in LPS-, ganglioside and IFN-γ-stimulated cells. By contrast, the NF-binding activity of the proximal element was not increased with any of the three stimulators (Figure 5A, panel c). A supershift assay to assess the LPS-induced NF-binding activity of distal AP-1 oligonucleotides showed that c-Fos and c-Jun competitively inhibit NF binding to the distal AP-1 element (Figure 5A, panel d). A competition assay with the middle AP-1 element showed the same results as with the distal element (results not shown). These results indicate that the c-Fos–c-Jun heterodimer complex binds to the middle and distal AP-1 element of MMP-3. To investigate whether the NF-binding activities of the distal and middle AP-1 elements are regulated by JNK activation, we used EMSA to analyse the distal and middle AP-1 elements in the presence or absence of SP600125. NF-binding activities on these two specific AP-1 elements, which were increased with stimulators, were reduced with JNK inhibitors (Figure 5B). We suggest that the distal and middle AP-1 sites within the 500 bp promoter are important for MMP-3 induction and that these binding activities are mediated by JNK activation.
NF-binding activities of distal and middle AP-1 elements of MMP-3 0.5 kb promoter are increased by glial activators
Site-directed mutagenesis of distal and middle AP-1 elements in MMP-3 promoter
To confirm that the distal and middle AP-1 elements of the 0.5 kb promoter region are necessary for induction of MMP-3, we performed two experiments: one using individual AP-1 mutants and one using pharmacological inhibitors of the AP-1 signal. For mutational analysis, we generated constructs comprising an AP-1 mutant ligated to the 2 kb or 0.5 kb promoter (Figure 6A). Cells were transiently transfected with wild-type or mutant luciferase constructs and incubated with LPS, gangliosides or IFN-γ for 36 h. Cell lysates were analysed using the luciferase assay. Although luciferase activities of cells transfected with the wild-type 2 or 0.5 kb promoters were generally increased with LPS, ganglioside and IFN-γ treatment, the activities of cells transfected with either AP-1 mutant ligated to the 2.0 (Figure 6B) or 0.5 kb promoter were not increased (Figure 6C). These results confirmed that MMP-3 expression requires both distal and middle AP-1 elements of the 0.5 kb promoter region.
Site-directed mutations of AP-1 elements fail to induce MMP-3 promoter activity
The findings of the present study shed light on the mechanisms involved in transcriptional induction of MMP-3 by LPS, gangliosides and IFN-γ in cultured rat brain astrocytes. Previous studies mainly used fibroblasts and hepatocytes in which DNA elements located in the −700 bp region are known to have important roles in the regulation of MMP-3 expression [18,24–26]. Several transcription-factor-binding sequences in the −700 bp region are known to function in constitutive and inducible transcription of MMP-3, but the precise regulatory mechanisms, and differences among tissues or stimulators, have not yet been determined.
The results of our study using brain astrocytes were consistent with those of a previous study, demonstrating that the sequences −500 bp from the start site are necessary for MMP-3 transcriptional induction by LPS, gangliosides or IFN-γ. The 500 bp sequence flanking the 5′-end of MMP-3 contains three AP-1-responsive elements. The proximal AP-1 element is a consensus sequence for AP-1, and the distal and middle AP-1 sequences are putative AP-1 elements. Our results show that the distal and middle AP-1 responsive elements, but not the proximal AP-1 element, are crucial for transcriptional induction of MMP-3 by LPS, gangliosides or IFN- γ, as shown by EMSA- and mutation-based analyses. Previous results have shown that consensus proximal AP-1 sequences are mainly involved in basal induction of MMP-3, whereas ETS-1 and other factors are involved in transcriptional induction by PMA or interleukins [18,27]. Although AP-1 was first described as an important transcription factor responsive to PMA, there is increasing evidence to suggest that it is involved in transcriptional responses to other stimuli [17,28,29]. As we have shown, AP-1 elements are involved in MMP-3 induction by LPS, gangliosides and IFN-γ in primary astrocytes. These results also highlight tissue-specific differential transcriptional regulation.
Despite crucial roles for the 500 bp region in transcriptional induction in activated astrocytes, further increases in transcription with full promoter sequences indicate that additional factors, which interact with sequences beyond −1.1 kb, are involved in the transcriptional induction of MMP-3. Human single-nucleotide-polymorphism studies have shown that the stromelysin IFN-responsive element or the stromelysin platelet-derived-growth-factor-responsive element located beyond the 1.1 kb region are involved in MMP-3 regulation by mediation of NF-κB or transcription factor 20 [6,14]. Although further studies are needed, the sequences beyond −1.1 kb could further increase MMP-3 transcription in brain astrocytes, owing to interactions with middle and/or distal AP-1-responsive regions in the 500 bp region. Considering similar structural domains across MMP family members, the interplay among transcription factors could critically regulate their transcription in different cell types and situations.
MMP-3 can degrade the extracellular matrix or activate other pro-MMPs such as MMP-1, -9 and -13, and it also has important roles in cellular functions involving cell movement, angiogenesis and organ morphogenesis [30,31]. Moreover, its expression is up-regulated in brain diseases such as stroke, Alzheimer's disease and Parkinson's disease, as well as in cancer and rheumatoid arthritis [9,32–34], indicating that it has pathogenic roles in these diseases. Release of MMP-3 from apoptotic neurons could activate microglia, thereby inducing cytokine release . In addition, β-amyloid, which is a component of senile plaques in Alzheimer's disease, induces the expression of MMP-3 and several other metalloproteinases . In primary astrocytes, MMP-3 expression and secretion are also increased by LPS, gangliosides and IFN-γ, which are well-known activators of brain glial cells. This indicates that increased MMP-3 induction and secretion from activated glial cells could trigger or aggravate neuronal damage and thereby play a role in brain pathologies in which inflammatory reactions due to activated glial cells are known to be involved.
Recently, beneficial roles of MMP-3 in delayed cortical responses after stroke have also been reported, indicating differential functions of MMP-3 depending on the sequence of events or conditions . Therefore, tight regulation of MMP-3 expression is critical in cellular functions, and AP-1 could be a useful target to control MMP-3 induction.
In summary, we have demonstrated that induction of MMP-3 is mainly achieved through distal and middle AP-1 sequences in the 0.5 kb promoter region. Therefore specific AP-1 elements could be exploited for the control of several brain pathologies in which increased MMP-3 levels trigger and aggravate neuronal damage.
This work was supported by a Korea Science and Engineering Foundation grant (no. R13-2003-019) funded by the Korean Government Ministry of Science and Technology.
Jun N-terminal kinase
mitogen-activated protein kinase
nuclear factor κB
These two authors contributed equally to the work.