AMP-activated Protein Kinase Suppresses Matrix Metalloproteinase-9 Expression in Mouse Embryonic Fibroblasts*

Matrix metalloproteinase-9 (MMP-9) plays a critical role in tissue remodeling under both physiological and pathological conditions. Although MMP-9 expression is low in most cells and is tightly controlled, the mechanism of its regulation is poorly understood. We utilized mouse embryonic fibroblasts (MEFs) that were nullizygous for the catalytic α subunit of AMP-activated protein kinase (AMPK), which is a key regulator of energy homeostasis, to identify AMPK as a suppressor of MMP-9 expression. Total AMPKα deletion significantly elevated MMP-9 expression compared with wild-type (WT) MEFs, whereas single knock-out of the isoforms AMPKα1 and AMPKα2 caused minimal change in the level of MMP-9 expression. The suppressive role of AMPK on MMP-9 expression was mediated through both its activity and presence. The AMPK activators 5-amino-4-imidazole carboxamide riboside and A769662 suppressed MMP-9 expression in WT MEFs, and AMPK inhibition by the overexpression of dominant negative (DN) AMPKα elevated MMP-9 expression. However, in AMPKα−/− MEFs transduced with DN AMPKα, MMP-9 expression was suppressed. AMPKα−/− MEFs showed increased phosphorylation of IκBα, expression of IκBα mRNA, nuclear localization of nuclear factor-κB (NF-κB), and DNA-binding activity of NF-κB compared with WT. Consistently, selective NF-κB inhibitors BMS345541 and SM7368 decreased MMP-9 expression in AMPKα−/− MEFs. Overall, our results suggest that both AMPKα isoforms suppress MMP-9 expression and that both the activity and presence of AMPKα contribute to its function as a regulator of MMP-9 expression by inhibiting the NF-κB pathway.

grades denatured collagens and native collagen type IV, which is a major component of the extracellular matrix (ECM) and basement membranes (1). Under normal circumstances, the degradation of the ECM by MMP-9 is a tightly controlled process involved in physiological wound healing and embryo development (1,2). Conversely, aberrant degradation of ECM by excess MMP-9 expression results in the pathologic destruction of connective tissue seen in cancer, arterial sclerosis, and rheumatoid arthritis (1,3). Therefore, under physiological conditions, regulated MMP-9 expression is low (1), but the mechanisms behind this are obscure.
AMP-activated protein kinase (AMPK) is a serine/threonine kinase, which regulates energy homeostasis and metabolic stress (4). AMPK acts as a sensor of cellular energy status and maintains the balance between ATP production and consumption. In mammals, AMPK exists as a heterotrimer with ␣, ␤, and ␥ subunits, each of which is encoded by two or three genes (␣1, ␣2, ␤1, ␤2, ␥1, ␥2, and ␥3). The ␣ subunit possesses catalytic activity, whereas the ␤ and ␥ subunits are regulatory and maintain the stability of the heterotrimer complex. The importance of AMPK␣ is illustrated by the fact that dual deficiency of AMPK␣1 and AMPK␣2 is embryonic lethal (5).
Recent evidence suggests that AMPK has a much wider range of functions, including the regulation of cell growth, cell proliferation, cell polarity, and autophagy (6,7). Because these functions are closely linked to the pathology of MMP-9-related diseases, including cancer, arterial sclerosis, and rheumatoid arthritis, we hypothesized that AMPK regulates MMP-9 expression. To address this, in the present study, we utilized AMPK␣-deficient mouse embryonic fibroblasts (MEFs) to investigate the effect of the genetic deletion and activation of AMPK on MMP-9 expression.
Gelatin Zymography-Conditioned media from cultured cells were collected and subjected to gelatin zymography. After cells reached 90% confluence, they were rinsed twice, and the medium was replaced with serum-free medium with or without TNF-␣ (1-100 ng/ml). After 24-h incubation, the conditioned media were collected and concentrated 3-fold using an Ultrafree-MC centrifugal filter device (Millipore) with a 30,000-molecular mass cutoff. The amount of concentrated media was normalized to the amount of protein in the cell lysate, then loaded on a Zymogram 10% gel (Invitrogen). Recombinant mouse MMP-2 and MMP-9 were used as positive controls. After renaturing and developing the gels according to the manufacturer's instructions, gels were stained with Coomassie Brilliant Blue R-250 solution (Bio-Rad). The intensities of bands were quantified using ImageJ software.
Western Blotting-Western blotting was carried out according to standard protocols. Densitometric analysis of bands was performed using ImageJ software.
ELISA-Analysis of accumulated MMP-9 in cell culture medium was performed using a quantitative ELISA kit (R&D Systems). After cells reached 90% confluence, they were rinsed twice, and fresh DMEM with or without reagent was added. The media were collected 12 or 24 h later, and assays were conducted according to the manufacturer's instructions. Obtained values were normalized to cell lysate protein levels.
DNA-binding Activity-The DNA-binding activity of NF-B p50, p52, p65, and RelB was determined by the Trans AM TM NF-B family assay kit (Active Motif, Carlsbad, CA). Nuclear extracts were prepared as described above, and 15-g nuclear extracts were used for the detection of DNA binding following the manufacturer's protocol.
Real-time Quantitative RT-PCR (qRT-PCR)-Total RNA was harvested from cells using the RNeasy kit (Qiagen), and complementary DNA (cDNA) was generated with the First Strand cDNA synthesis kit (GE Healthcare) according to the manufacturer's instructions. Real-time PCR was carried out using the following mouse TaqMan gene expression assays (Applied Biosystems): AMPK␣1 (Mm01296695_m1), AMPK␣2 (Mm01264788_m1), MMP-9 (Mm00442991_m1), IB␣ (Mm00477798_m1), and ␤-actin (Mm00607939_s1). All reactions were prepared following the manufacturer's protocol and carried out using the StepOne TM Real-time PCR System (Applied Biosystems).
Adenovirus Vector Transduction-The adenovirus vector for the dominant negative (DN) form of AMPK␣2 (Ad-DN) with an inactivating mutation in the kinase domain (K45R substitution) has been described previously (10). The Ad-DN contained GFP as a marker, and the adenovirus vector 5 with GFP (Ad-GFP) (Vector BioLabs, Philadelphia, PA) was used as a control. MEFs were transduced with the adenovirus vectors at a multiplicity of infection of 300 for 48 h. The medium was then changed, and cell extracts and medium were harvested after 12 h. Under these conditions, the infection efficiency was Ͼ90%.
Statistical Analysis-All experiments were repeated a minimum of three times. All data were expressed as means Ϯ S.E. Statistical differences between two groups were analyzed by the unpaired Student's t test. Multiple group comparison was performed by one-way analysis of variance with Scheffe's test. Differences were considered significant at p Ͻ 0.05.

Deletion of Both AMPK␣1 and AMPK␣2 Isoforms Results in
Constitutive Expression of MMP-9 from MEFs-To study the role of AMPK in fibroblast expression of MMP-9, we utilized SV40-immortalized WT MEFs and MEFs that were nullizygous for both AMPK␣1 and AMPK␣2 subunits (AMPK␣ Ϫ/Ϫ MEFs). We examined the gelatinolytic activity of culture medium using gelatin zymography. As shown in Fig. 1A, the base-line gelatinolytic activity of pro-MMP-9 (under nonstimulatory conditions) of WT MEFs was barely detectable. By contrast, AMPK␣ Ϫ/Ϫ MEFs showed MMP-9 gelatinolytic activity that was 4.3 Ϯ 0.9 times higher than that of WT MEFs (p Ͻ 0.05). There was no significant difference in the gelatinolytic activity of MMP-2 between the two cell types. To determine the magnitude of AMPK deletion in fibroblast MMP-9 expression, we next treated both MEFs with the major inducible factor of MMP-9, TNF-␣ (1). TNF-␣ increased the MMP-9 gelatinolytic activity of both WT and AMPK␣ Ϫ/Ϫ MEFs in a dose-dependent manner (Fig. 1A). Notably, the MMP-9 gelatinolytic activity of unstimulated AMPK␣ Ϫ/Ϫ ⌴⌭Fs was equal to the maximum gelatinolytic activity of WT MEFs with TNF-␣ treatment. Furthermore, the MMP-9 gelatinolytic activity of AMPK␣ Ϫ/Ϫ ⌴⌭Fs with each dose of TNF-␣ was two to three times higher than that of WT MEFs treated in the same way. The results of gelatin zymography were confirmed by Western blotting (Fig.  1B). Collectively, these results demonstrate that deletion of both AMPK␣1 and AMPK␣2 subunits leads to constitutive expression of MMP-9 and augmentation of the effect of TNF-␣ on MMP-9 expression.
SV40 immortalization is known to inactivate the function of the cancer suppressor genes Rb and p53 (11,12). To exclude the effects of SV40 immortalization on MMP-9 expression, we cultured primary MEFs of WT and AMPK␣ Ϫ/Ϫ and analyzed the level of MMP-9 protein in their culture media by Western blotting and ELISA. Fig. 2A shows a representative result of confirmatory Western blotting for the deletion of AMPK␣ in AMPK␣ Ϫ/Ϫ MEFs. The ability of antibody to detect phosphorylated acetylcoenzyme A carboxylase (ACC), which is a downstream target of AMPK, was taken as an indicator of AMPK activation. In accordance with the results for SV40-immortalized MEFs, both Western blotting and the ELISA showed that MMP-9 expression from primary AMPK␣ Ϫ/Ϫ MEFs was significantly higher than that of primary WT MEFs (7.8 Ϯ 0.9 times higher in the ELISA, p Ͻ 0.01, Fig. 2, B and C). These results show that SV40 immortalization does not affect differences in MMP-9 expression between WT and AMPK␣ Ϫ/Ϫ MEFs. Hereafter, all experiments were performed with SV40immortalized MEFs unless otherwise noted.
Next, we investigated the effects of AMPK inhibition on MMP-9 expression by overexpressing the DN form of AMPK␣ in WT MEFs. Western blotting of Ad-DN-transduced WT MEF cell lysates showed a decrease in ACC phosphorylation, indicating inhibition of AMPK activity (Fig. 4E). The MMP-9 level of Ad-DN-transduced WT MEFs was 1.85 Ϯ 0.19 times higher than that of Ad-GFP-transduced WT MEFs (Fig. 4F). Taken together, these results indicate the importance of AMPK kinase activity for the inhibition of MMP-9 expression from MEFs.
Presence of AMPK Inhibits MMP-9 Expression from MEFs-It has been reported that protein kinase including AMPK can bind to different protein and regulates signal transduction independently of its kinase catalytic activity (19,20). This suggests that not only the kinase activity but also the presence of AMPK may play an important role in the regulation of MMP-9 expression. To address this, we transduced AMPK␣ Ϫ/Ϫ MEFs with the Ad-DN. Transduction was confirmed by Western blotting (Fig. 5A). Western blotting showed inhibition of ACC phosphorylation in Ad-GFP-and Ad-DN-transduced AMPK␣ Ϫ/Ϫ MEFs. It also revealed the up-regulation of AMPK␤1 in Ad-DN-transduced AMPK␣ Ϫ/Ϫ MEFs. Interestingly, ELISA revealed that the MMP-9 protein level of Ad-DNtransduced AMPK␣ Ϫ/Ϫ MEFs was significantly lower than that of Ad-GFP-transduced AMPK␣ Ϫ/Ϫ MEFs (p Ͻ 0.01, Fig. 5B). These results, in addition to the previous experiments mentioned above, indicate the importance of both the activity and the presence of AMPK in inhibiting MMP-9 expression in MEFs.
Constitutive Activation of the NF-B Pathway Is Involved in Up-regulating MMP-9 Expression by AMPK␣ Ϫ/Ϫ MEFs-MMP-9 expression is largely controlled by transcription of the gene, although mRNA stability and translational efficiency also play a role in regulating protein levels (1). Regulation of transcription is achieved via a regulatory sequence containing binding sites for activator protein 1 (AP-1), NF-B, Sp1, and PEA3/ Ets (21)(22)(23). To investigate which of these transcription factors are involved in the up-regulation of MMP-9 expression in AMPK␣ Ϫ/Ϫ MEFs, we used Western blotting to examine the effects of AMPK␣ deletion on the phosphorylation of c-Jun, major subunits of AP-1, and its upstream MAPKs (i.e. ERK, JNK, and p38 MAPK), and the nuclear translocation of c-Jun and NF-B. As shown in Fig. 6A, the phosphorylation of JNK was increased in AMPK␣ Ϫ/Ϫ ⌴EFs, whereas there was no change in the phosphorylation of ERK. The phosphorylation of p38 MAPK was down-regulated in AMPK␣ Ϫ/Ϫ ⌴⌭Fs. Nuclear extracts from WT and AMPK␣ Ϫ/Ϫ ⌴⌭Fs showed an increase in the phosphorylation and the nuclear translocation of c-Jun, indicating that it is activated through the activation of JNK (Fig.  6B). Western blotting of the nuclear extract also revealed nuclear translocation of NF-〉 p65 and p52 in AMPK␣ Ϫ/Ϫ ⌴⌭Fs, indicating that both canonical and noncanonical NF-B pathways are activated in AMPK␣ Ϫ/Ϫ ⌴EFs (Fig. 6B).
To determine whether the activation of c-Jun and NF-B pathways is responsible for the up-regulation of MMP-9 expression in AMPK␣ Ϫ/Ϫ MEFs, we treated them with IKK inhibitor (BMS345541), NF-B inhibitor (SM7368), or JNK inhibitor (SP600125) and evaluated MMP-9 protein levels by ELISA. Treatment with both BMS345541 and SM7368 significantly suppressed MMP-9 expression by 51 and 34%, respectively (both p Ͻ 0.01), whereas application of SP600125 did not. This indicates that the NF-B pathway is partially responsible for the up-regulation of MMP-9 expression in AMPK␣ Ϫ/Ϫ MEFs (Fig. 6C). AMPK␣ deletion (comparing WT versus AMPK␣ Ϫ/Ϫ MEFs) increases MMP-9 expression by about   (Fig. 6C), but as shown previously, activation of endogenous AMPK␣ only suppresses MMP-9 in WT MEFs by about 50% (Fig. 4, B and D). This would suggest that endogenous AMPK␣ activity is sufficient to suppress MMP-9 expression tonically. Indeed, when NF-B inhibitors BMS345541 and/or SM7368 were given in WT MEFs, they markedly suppressed endogenous MMP-9 expression by 71 and 61%, respectively (both p Ͻ 0.01, Fig. 6C). Together, these results highlight the importance of the AMPK␣/NF-B pathway in MMP-9 expression.

5-fold
To investigate whether the AMPK␣ deletion causes the constitutive activation of the NF-〉 pathway in MEFs, we next investigated the phosphorylation and mRNA expression of IB␣ (p-IB␣). Western blotting showed that IB␣ phosphorylation was markedly higher in AMPK␣ Ϫ/Ϫ MEFs than in WT MEFs (Fig. 6D). The expression level of IB␣ mRNA in AMPK␣ Ϫ/Ϫ MEFs was also significantly higher than in WT MEFs (p Ͻ 0.01, Fig. 6E). As IB␣ is degraded by the ubiquitin process after phosphorylation (24 -26), these findings suggest that IB␣ might undergo chronic degradation and resynthesis in AMPK␣ Ϫ/Ϫ ⌴⌭Fs and that the AMPK␣ deletion causes constitutive activation of the NF-〉 pathway in MEFs (27).
To confirm activation of the NF-B pathways further, we next evaluated the binding activity of nuclear extracts to NF-〉 subunits of p65, p52, p50, and RelB by ELISA. We found that the binding activities of all NF-B subunits were up-regulated in AMPK␣ Ϫ/Ϫ MEFs (p Ͻ 0.01, Fig. 6F), indicating that both canonical and noncanonical NF-B pathways are activated. Collectively, these results suggest that the AMPK deletion leads to the constitutive activation of the NF-B pathway, which is at least partly responsible for the up-regulation of MMP-9 expression.

DISCUSSION
MMP-9 plays a critical role in tissue remodeling under both physiological and pathological conditions. Its expression is low in most cells and is tightly controlled. Although many factors have been identified as stimulators of MMP-9 expression, only integrin ␣1 and transgelin have been shown to inhibit MMP-9 expression in normal cells under physiological conditions (28, 29). Here, we utilized MEFs that were nullizygous for the catalytic ␣ subunit of AMPK to identify AMPK as a negative regulator of MMP-9 expression, thus adding it to the short list of known MMP-9 repressors.
Using the gene knock-out system, Pozzi et al. (28) and Nair et al. (29) reported that integrin ␣1 and transgelin are negative regulators of MMP-9 expression by demonstrating over-expression of MMP-9 in normal lung endothelial cells and uterine epithelial cells from knock-out mice. Other groups have shown that proteins such as kisspeptin-1 (KiSS-1), PKCs, heme oxygenase 1, reversion-inducing cysteine-rich protein with kazal motifs (RECK), and caveolin-1 down-regulate MMP-9 expression (30 -35). However, the latter studies used cell lines such as HT-1080, prostate cancer cells, pancreatic cancer cells, and NMuMG which, in contrast to normal cells and MEFs, express MMP-9 constitutively at high levels. Constitutive up-regulation of MMP-9 in these cells is considered to be due to constitutive activation of MAPKs by oncogenic transformation of Ras (36 -38). Even if these proteins can decrease MMP-9 expression in Ras-transformed cell lines, it does not follow that they are responsible for the basal suppression of MMP-9 expression in non-Ras-transformed cells.
We observed that deletion of both AMPK␣1 and ␣2 isoforms in primary or immortalized MEFs led to a significant up-regulation of MMP-9 expression (Figs. 1, A and B, and 2, B and C). However, single deletion of AMPK␣1 or ␣2 did not cause overexpression of MMP-9 to the levels seen in MEFs that are nullizygous for both AMPK␣1 and ␣2 (Fig. 3, B and C). Notably, the MMP-9 expression level of AMPK␣1 Ϫ/Ϫ MEFs, which is lacking a majority of AMPK␣ (Fig. 3A), was only 20% of that of AMPK␣ Ϫ/Ϫ MEFs. These results suggest that AMPK␣1 and AMPK␣2 both function to inhibit MMP-9 expression, and that even a small amount of AMPK␣ can potently inhibit MMP-9 expression.
Previous experimental inflammation animal models and cancer cell lines have been used to show that the broad and nonspecific AMPK activators metformin and AICAR decrease MMP-9 expression; however, base-line MMP-9 expression in these artificial models was already up-regulated, and metformin and AICAR can have many off-target effects (39,40). Similarly, here we found that the activity of AMPK was partially responsible for the regulation of basal MMP-9 levels. However, the magnitude of the effect of AMPK activity manipulation on MMP-9 expression was not as large as the effect seen by deletion of both AMPK␣ isoforms (compare Fig. 3B and Fig. 4, B, D,  and F). This led us to investigate whether not only the activity but also the presence of AMPK␣ is important for inhibition of basal MMP-9 expression. Bronner et al. reported that AMPK binds to PPAR␣ and co-activates PPAR␣-mediated transcription, independently of its catalytic activity (19). Indeed, transduction with kinase-dead AMPK␣ (DN AMPK␣) suppressed MMP-9 expression in AMPK␣ Ϫ/Ϫ MEFs (Fig. 5B). Although the reduction rate in this experiment (45%, Fig. 5B) was not as drastic as we expected, we think that one possible reason for this can be the limited infection efficiency of Ad-DN AMPK. Interestingly, AMPK␣ Ϫ/Ϫ MEFs transduced with Ad-DN showed a greater increase in AMPK␤1 protein levels than Ad-GFP-transduced AMPK␣ Ϫ/Ϫ MEFs (Fig. 5A). This can be explained by the importance of heterotrimeric complex in the stabilization of individual AMPK subunit. AMPK is stable as a heterotrimer complex, whereas each of its subunits is subject to an increased turnover rate and is depleted from the cell when not associated with others (41)(42)(43). Therefore, the observed higher expression of AMPK␤1 in AMPK␣ Ϫ/Ϫ MEFs indicates that the transduction of the Ad-DN resulted in an increase in the AMPK heterotrimer complex in the cell. Recently, it was reported that the AMPK␤ subunit is not only a scaffold that assembles ␣ and ␥ subunits, but also determines the subcellular localization and substrate specificity of the AMPK heterotrimer complex (44,45). Thus, understanding the function of the AMPK heterotrimer complex in the regulation of MMP-9 is important, and further study is required to achieve this.
We found that AMPK suppresses MMP-9 expression by inhibiting the NF-B pathway in MEFs. The gelatin zymography result showed that the AMPK␣ deletion did not affect MMP-2 expression in MEFs (Fig. 1A). These findings are consistent with the fact that the regulatory sequence of MMP-2 does not contain a NF-B-binding site (46). Furthermore, the nuclear localization and NF-B DNA-binding activity results suggest that activation of both the canonical and the noncanonical NF-B pathways is responsive to overexpression of MMP-9 in AMPK␣ Ϫ/Ϫ MEFs (Fig. 6, B, C, and F). Wang et al. reported that deletion of AMPK␣2 results in the constitutive activation of NF-B in mouse aortic endothelial cells (47). In addition, many studies have reported that the activation of AMPK inhibits the NF-B pathway (48 -51). Although the mechanism of AMPK inhibition of the NF-B pathway is not fully elucidated, there are at least two main possibilities (Fig. 7).
The first possibility is that AMPK inhibits IKK-dependent IB␣ phosphorylation. Phosphorylation of IB␣ by the upstream kinase IKK is essential for NF-B nuclear translocation. Our results demonstrate that the AMPK␣ deletion led to the phosphorylation of IB␣ and nuclear translocation of NF-B, indicating that AMPK targets IKK activity or upstream FIGURE 7. Proposed model for the mechanism by which AMPK suppresses MMP-9 expression. AMPK negatively regulates MMP-9 expression by inhibiting the NF-B pathway. Our results suggest at least two possibilities for the mechanism: 1) AMPK inhibits IKK activity or upstream proteins of IKK, such as TAK1, TAB1-3, and NEMO; and 2) AMPK directly inhibits the DNA-binding activity of NF-B.
The second possibility is that AMPK directly inhibits the DNA-binding activity of NF-B. It has been reported that the AMPK␣1 isoform is primarily cytoplasmic, whereas AMPK␣2 is predominantly nuclear and plays a role in transcriptional regulation (58,59). Indeed, Katerelos et al. reported the possibility that AMPK reduces the NF-B DNA-binding activity in bovine aortic endothelial cells (51). Further study is required to determine the target of AMPK for interference with the NF-B pathway.
In conclusion, we identified AMPK as a novel negative regulator of MMP-9 expression in MEFs under physiological conditions. Recently, ample evidence indicates the importance of AMPK in the pathogenesis of cancer and arterial sclerosis (6, 7, 60 -62). Because MMP-9 plays an important role in these diseases, in the invasion and metastasis of cancer cells and the rupture of atheromatous plaques, our findings might provide fundamental insights not only into the regulatory mechanism of MMP-9 expression and the function of AMPK, but also into the pathogenesis of these diseases.