Differential Regulation of Lipopolysaccharide-Induced Monocyte Matrix Metalloproteinase (MMP)-1 and MMP-9 by p38 and Extracellular Signal-Regulated Kinase 1/2 Mitogen-Activated Protein Kinases

Monocytes/macrophages are prominent cells at chronic inflammatory sites, such as the sites of arthritis, atherosclerosis, and periodontal disease, in which degradation of the connective tissue is believed to be a major contributing factor to the pathology associated with the disease. Destruction of the connective tissue in these lesions is believed to be due to the action of matrix metalloproteinases (MMPs).² MMPs are comprised of a family of enzymes that include interstitial collagenases, gelatinases, stromelysin, matrilysin, metalloelastase, and membrane-type MMPs (1, 2, 3). Collectively, MMPs can degrade all the components of the extracellular matrix. Stimulation of monocytes with LPS, a surface component of Gram-negative bacteria, induces a number of MMPs including two prominent monocyte MMPs: interstitial collagenase (MMP-1) and gelatinase B (MMP-9). MMP-1 cleaves fibrillar collagens, such as types I, II, and III, resulting in denatured collagens (gelatins) that are further degraded by MMP-9. Additionally, MMP-9 also degrades laminin and type IV collagen, components of the basement membrane. Thus, these enzymes are involved in the connective tissue loss associated with chronic inflammatory diseases as well as the migration of cells out of the blood stream and through the extracellular matrix.

It has been shown that the production of MMPs by human monocytes/macrophages following stimulation with agents such as Con A, LPS, or extracellular matrix components occurs, in large part, through a PGE₂-cAMP-dependent pathway (4, 5, 6, 7, 8). The early signaling events preceding PGE₂ and cAMP that are involved in regulating LPS-mediated monocyte MMP production are not well characterized. Within the last few years, numerous studies have shown that the mitogen-activated protein kinases (MAPKs) are involved in the regulation of MMPs by various cell types (9, 10). The MAPK family includes three major members: the extracellular signal-regulated kinases (ERKs), the c-Jun N-terminal kinase (JNK)/stress-activated protein kinases, and p38 (11, 12). The ERK1/2 pathway is activated by mitogens and phorbol esters and is sensitive to PD98059 inhibition of MAPK kinase 1/2. The JNK and p38 pathways are activated by cellular stress and inflammatory cytokines. The pyridinyl imidazole drug SB203580 has been shown to be an effective inhibitor of p38 activity. Although the role of MAPKs in the regulation of MMPs has been studied in many cell types, little is known about how MAPKs regulate the production of MMPs in monocytes.

In the present study, we have evaluated the role of MAPKs in the LPS-mediated signaling leading to the production of MMPs by monocytes. In this study, we report that ERK1/2 and p38 are essential for the induction of MMP-1 by LPS. Regulation of MMP-1 by p38 occurs through a PGE₂-dependent mechanism, whereas ERK1/2 controls MMP-1 through effects on transcription factors in the promoter that are regulated through a PGE₂-independent mechanism. In contrast, MMP-9 production is mainly regulated by ERK1/2.

The following Abs were used: anti-p38, anti-ERK1/2, anti-phospho-Elk, anti-mouse IgG-HRP, anti-rabbit IgG-HRP, anti-Goat IgG-HRP, anti-MAPK activated protein kinase-2 (MAPKAPK-2), anti-MMP-9 (Santa Cruz Biotechnology, Santa Cruz, CA), anti-MMP-1 (kindly provided by Dr. H. Birkedal-Hansen (National Institute of Dental and Craniofacial Research)) and anti-PGE₂ (Upstate Biotechnology, Waltham, MA). Additional reagents included the following: protein G-Sepharose, [γ-³²P]ATP, and [α-³²P]dCTP (Amersham Biosciences, Piscataway, NJ), 4-(4-fluorophenyl)-2-(4-methylsulfonylphenyl)-5-(4-pyridinyl) imidazole (SB203580), and PD98059 (Calbiochem-Novabiochem, San Diego, CA), TRIzol (Life Technologies, Rockville, MD), and QuikHyb (Stratagene, La Jolla, CA). LPS, dibutyryl cAMP (Bt₂cAMP), PGE₂, and indomethacin were purchased from Sigma-Aldrich (St. Louis, MO).

Human peripheral blood monocytes were obtained by leukapheresis of normal volunteers at the Department of Transfusion Medicine at the National Institutes of Health. The monocytes were purified by counter-flow centrifugal elutriation as previously described (13). Monocytes were enriched to >90% as determined by morphology, nonspecific esterase staining, and flow cytometry. Purified monocytes were cultured in DMEM (BioWhittaker, Walkersville, MD) supplemented with 2 mM l-glutamine (Mediatech, Herndon, VA) and 10 μg/ml gentamicin sulfate (BioWhittaker) at 37°C in a humidified atmosphere containing 5% CO₂. LPS, Bt₂cAMP, PGE₂, indomethacin, SB203580, or PD98059 were added to some of the cultures. Unless otherwise stated, monocytes were adhered for 30 min before the addition of reagents. Each experiment was repeated a minimum of three times with different donors.

For determination of the protein levels of MMP-1 and MMP-9 produced by monocytes, proteins in the supernatants of 48-h conditioned medium were precipitated with cold ethanol (final concentration of 60%) at −70°C as previously described (14). The proteins from equal portions of the conditioned medium were separated on a Novex 8–16% Tris-glycine polyacrylamide gel (Invitrogen, Carlsbad, CA) and then transferred onto a nitrocellulose membrane. The nitrocellulose membranes were blocked with 5% nonfat dry milk in a buffer containing 20 mM Tris-HCl (pH 7.5), 150 mM NaCl, and 0.1% Tween 20. The membranes were then incubated overnight with the MMP-1 or MMP-9 primary Abs. A peroxidase-conjugated secondary Ab was then added and the bound primary Ab recognized by the specific protein was detected by an ECL detection system (Pierce, Rockford, IL).

Membrane proteins for COX-2 protein determination were prepared as previously described (15). Equal amounts of cell membrane proteins (25 μg) were fractionated on a Novex 8–16% Tris-glycine polyacrylamide gel and then transferred to a nitrocellulose membrane. COX-2 was determined by Western blot analysis as described under Detection of MMP-1 and MMP-9 by Western blot analysis.

PGE₂ levels in the medium supernatants from monocyte cultures were determined by RIA as previously described (16).

Cells (20 × 10⁶) were lysed for 10 min on ice in 0.5 ml of lysis buffer (20 mM HEPES (pH 7.5), 10 mM EGTA, 40 mM β-glycerophosphate, 1% Nonidet P-40, 7.5 mM MgCl₂, 2 mM sodium orthovanadate, and 1 mM PMSF). Samples were clarified by centrifugation at 4°C for 10 min at 12,000 × g. ERK and p38 were immunoprecipitated by the addition of anti-ERK and anti-p38 Abs conjugated to agarose beads, respectively. The beads were washed twice with lysis buffer (20 mM HEPES (pH 7.5), 10 mM EGTA, 40 mM β-glycerophosphate, 1% Nonidet P-40, 7.5 mM MgCl₂, 2 mM sodium orthovanadate, and 1 mM PMSF) and twice with kinase assay buffer (12.5 mM MOPS (pH 7.5), 0.5 mM EGTA, 12.5 mM β-glycerophosphate, 7.5 mM MgCl₂, 0.5 mM sodium fluoride, and 0.5 mM vanadate). ERK activity was assayed using 1 μg of GST-Elk as substrate in kinase assay buffer with 200 μM ATP at 30°C. The reaction was terminated 30 min later with 2× Laemmli’s buffer, the reaction mixture was heated to 100°C for 4 min, and the beads were removed by centrifugation. Samples were separated by electrophoresis on a 10% Tris-glycine polyacrylamide gel (Invitrogen). The proteins were transferred to a nitrocellulose membrane and the phosphorylated substrate and MAPKs were analyzed by Western blot. For p38 activity, MAPKAPK-2 was immunoprecipitated with anti-MAPKAPK-2 Ab and protein G beads. Samples were assayed for kinase activity using 1 μg of GST-heat shock protein 27 (HSP27) as substrate in kinase assay buffer with 25 μM ATP and 10 μCi of [γ-³²P]ATP at 30°C. Phosphorylation of the substrate was detected on a PhosphorImager (ImageQuant, version 3.3; Molecular Dynamics, Sunnyvale, CA).

Monocyte total RNA was isolated with TRIzol reagent (Invitrogen) according to the instructions of the manufacturer. RNA (10 μg) was electrophoresed on denaturing formaldehyde/1.2% agarose gels with 0.41 M formaldehyde and then capillary transferred to a Hybond N membrane and fixed by UV cross-linking. cDNA probes for MMP-1, MMP-9, and GAPDH were labeled with [α-³²P]dCTP by random priming. Prehybridization and hybridization of MMP-1, MMP-9, and GADPH probes were conducted at 65°C in QuickHyb solution (Stratagene). The membranes were washed two times in 2× SSC and 0.1% SDS for 15 min at room temperature and once in 0.1× SSC and 0.1% SDS for 30 min at 65°C.

The double-stranded oligonucleotides used for EMSA were designed to be specifically complementary to the binding sites in MMP-1 and MMP-9 promoters. For the MMP-1 promoter, the oligonucleotides used were the following: SP 1 (-1029), 5′-GGCAAGGGGTGGGGAGTTATC-3′; CREB (-2017), 5′-TTCCAGGGTGACGTCTTAGGC-3′; and AP1 (-69), 5′-CGCTTGATGAGTCAGCCGGAA-3′. For the MMP-9 promoter, the oligonucleotides used were the following: SP 1 (-73), 5′-AAGGAGGGGTGGGGTCACAGG-3′; NF-κB p65 (-619), 5′-CCCCAGTGGAATTCCCCAGCC-3′; NF-κB p50 (-347), 5′-GGGATGGGGGATCCCTCCAGC-3′; and PEA3-AP1 (-79), 5′-GCAGGGAGAGGAAGCTGAGTCAAAGA-3′. Nuclear proteins were extracted from monocytes that were incubated in the presence or absence of LPS for 8 h. EMSA was performed using equal amounts of nuclear extracts (5 μg) that were incubated with ³²P-labeled oligonucleotides.

To study the role of MAPKs in the regulation of monocyte MMPs, we first determined the effective doses of MAPK inhibitors on MAPK activity. SB203580, an inhibitor that blocks the ability of activated p38 to phosphorylate downstream substrates, was used in this study. MAPKAPK-2, a specific physiological substrate for p38, was immunoprecipitated, and its activity was analyzed using HSP27 as a substrate. Activation of p38 by LPS was demonstrated by the increase in the ability of MAPKAPK-2 to phosphorylate HSP27 (Fig. 1). Addition of SB203580 resulted in a dose-dependent inhibition of MAPKAPK-2 phosphorylation of HSP27. SB203580, at a concentration of 10 μM, reduced the phosphorylation of HSP27 by MAPKAPK-2 immunoprecipitated from LPS-stimulated cells to a level below that of control monocytes.

We next investigated the effect of MAPK inhibitors on ERK activity. Because several studies have shown that there is potential cross-talk between p38 and ERK (17), we examined the effects of the ERK inhibitor, PD98059, and the p38 inhibitor, SB203580, on ERK activity. Various concentrations of these inhibitors were added to monocytes in the presence or absence of LPS, and ERK was immunoprecipitated and tested for its ability to phosphorylate Elk. LPS treatment of the monocytes resulted in a significant activation of ERK as shown by the phosphorylation of Elk that was inhibited by 10 μM PD98059 (Fig. 2). In contrast, SB203580 caused a significant dose-dependent increase in the phosphorylation of Elk by ERK. Abs against ERK confirmed that equal amounts of enzyme were used in each reaction (Fig. 2). These findings indicate that p38 can function as a negative regulator of ERK activation. However, unlike the affect of SB203580 on ERK phosphorylation, PD98059 did not influence the phosphorylation or protein levels of p38 (data not shown).

To assess the role of MAPKs on the LPS-mediated signal transduction pathway in monocytes that leads to MMP production, we looked at downstream events that may be affected by MAPKs. Previous studies have shown that monocyte MMP production occurs, at least in part, through a PGE₂-cAMP-dependent pathway (4, 5, 6, 7, 8). An example of this is shown in Fig. 3, in which indomethacin, a PG inhibitor, suppresses the LPS-induced mRNA and protein levels of MMP-1 and MMP-9. Whereas there is an almost complete inhibition of MMP-1 by indomethacin, MMP-9 is only partially inhibited. Exogenously added PGE₂ or Bt₂cAMP reversed the inhibition of MMPs by indomethacin. Thus, LPS induction of monocyte MMP-1 is tightly regulated by a PGE₂-cAMP-dependent pathway, whereas MMP-9 is partially controlled by this pathway.

To determine whether MAPKs regulated monocyte MMPs through a PGE₂-cAMP-dependent pathway, MAPK inhibitors were added to LPS-stimulated monocytes in the presence or absence of PGE₂ or Bt₂cAMP. Addition of PD98059 to monocytes resulted in a significant inhibition of LPS-induced MMP-1 and MMP-9, as demonstrated at the mRNA and protein level (Fig. 4). However, neither MMP-1 nor MMP-9 were restored by the addition of PGE₂ or Bt₂cAMP. In contrast, inhibition of p38 by SB203580 resulted in suppression of LPS-induced MMP-1 but enhanced MMP-9 (Fig. 5). However, unlike PD98059, the inhibition of MMP-1 mRNA and protein by SB203580 was reversed by PGE₂ or Bt₂cAMP. Additionally, the enhancement of MMP-9 by SB203580 treatment was increased further by the addition of PGE₂.

Restoration of MMP-1 production by PGE₂ in SB203580-treated monocytes indicated that MAPKs may be involved in the potential regulation of COX-2 and the subsequent production of PGE₂. This possibility was examined by determining the effect of p38 or ERK inhibition on LPS induction of COX-2 and PGE₂. As shown in Fig. 6, LPS stimulates significant levels of COX-2 and PGE₂. Inhibition of p38 by SB203580 resulted in an almost complete inhibition of COX-2 and PGE₂ even at a low concentration of l μM. PD98059 also resulted in a significant but not total inhibition of COX-2 and PGE₂ at a concentration of 10 μM. Thus, while both MAPK pathways can regulate PG synthesis, the inability to restore MMP production in PD98059-treated monocytes indicates that ERK regulates MMP production through an additional PG-independent pathway.

The ability of PGE₂ to restore MMP-1 in SB203580-treated monocytes but not in monocytes cultured with PD98059 indicated that p38 and ERK1/2 may differentially regulate transcription factors in the promoters of MMP-1 and MMP-9. To examine this possibility, we designed complementary oligonucleotides to selected binding sites in the promoters of MMP-1 and MMP-9. PD98059 treatment resulted in the inhibition of DNA binding of complementary oligonucleotides of AP 1 and SP1 for MMP-1, as well as those for AP1 and NF-κB of MMP-9 (Fig. 7). In contrast, SB203580 only inhibited the binding of the SP 1 oligonucleotide for the MMP-1 promoter and actually increased the binding of the NF-κB p50 complementary oligonucleotide for MMP-9.

Because CREB is regulated by PGE₂, which restored MMP-1 in SB203580-treated monocytes, we also examined the binding of a CREB oligonucleotide designed to bind to the MMP-1 promoter. Stimulation of monocytes increased CREB binding, which was inhibited by SB203580 (Fig. 8). Moreover, this inhibition was reversed by PGE₂, which was in agreement with the restoration of MMP-1 in SB203580-treated monocytes by PGE₂. In contrast to MMP-1, inhibition of p38 resulted in an increase in MMP-9, as previously shown in Fig. 4. A possible explanation for this is that the increase in phosphorylation of ERK1/2 by SB203580, as shown in Fig. 2, resulted in enhanced binding of the complementary NF-κB to the MMP-9 promoter (Fig. 7). This was demonstrated again by the data in Fig. 9, in which SB203580 increased NF-κB p65 binding, and this binding was further enhanced by PGE₂, which was in agreement with the increase in MMP-9 protein as shown in Fig. 4. Thus, the ability of SB203580 to enhance NF-κB binding was demonstrated by both the p50 (Fig. 7) and p65 NF-κB (Fig. 9) oligonucleotides.

These findings demonstrate that LPS stimulation of monocyte MMP-1 and MMP-9 is differentially regulated by p38 and ERK1/2. Inhibitors of either pathway suppressed MMP-1 at the mRNA and protein level, whereas MMP-9 was inhibited only by the ERK1/2 inhibitor PD98059. In contrast to MMP-1, MMP-9 mRNA and protein were enhanced by SB203580, which also increased phosphorylation of ERK1/2. These data suggest that MMP-9 is primarily regulated through ERK1/2 and that p38 serves as a negative regulator of ERK1/2. These findings are in agreement with previous results with regard to cross-talk between MAPKs, which have shown the suppression of ERK1/2 phosphorylation by p38 (18, 19).

Of interest is the ability of PGE₂ or Bt₂cAMP to reverse the inhibition of MMP-1 mRNA and protein by SB203580. These findings show that p38 MAPK mainly regulates LPS-induced monocyte MMP production through its effect on COX-2 and the subsequent production of PGE₂ required for MMP synthesis. In contrast, even though the ERK1/2 inhibitor PD98059, like SB203580, decreases COX-2 and PGE₂, its inhibition of MMP production cannot be reversed by PGE₂. Thus, suppression of monocyte MMPs by PD98059 occurs not only by decreasing PGE₂ synthesis but also through additional mechanisms that most likely involve direct modulation of the transcription factors that are required in conjunction with PGE₂ to induce MMP production.

The regulation of PGE₂ by p38 and ERK1/2 may occur through several mechanisms. Both p38 and ERK1/2 have been reported to phosphorylate cytosolic phospholipase A₂ in macrophages and thus contribute to the release of arachidonic acid for the generation of PGE₂ (20, 21). Moreover, the inhibition of both p38 and ERK1/2 is more effective in blocking arachidonic acid release than the inhibition of only one of these MAPKs. Therefore, it has been suggested that the phosphorylation sites or the particular amino acid affected by these MAPKs may be different. In addition to the regulation of PGE₂ through effects on phospholipase, MAPKs can also affect COX-2 activity as has been shown by the ability of p38 to regulate the transcription and stability of COX-2 mRNA (22, 23). In our study, the specific p38 inhibitor SB203580 was more effective in inhibiting monocyte COX-2 and PGE₂ than was PD98059.

In addition to their effect on PG production, MAPKs, either through direct or indirect interaction with transcription factors, regulate various biological functions including MMPs. The MAPK pathways invoked by biological stimulants and the interaction with transcription factors are most likely determined by the cell type and the specific agonists (9, 24, 25). An additional layer of diversity occurs due to differential regulation of various members of the MMPs by MAPKs, as well as their effect on stabilization of mRNA (26). Studies have shown that, in skin fibroblasts, depending on the stimulant, ERK1/2, JNK, and p38 MAPK are all involved in the regulation of MMPs (9). Additionally, p38 has been shown to induce the stabilization of fibroblast MMP-1 and MMP-3 mRNAs, while ERK1/2 is involved in the transcription of these MMPs (26).

Our findings with LPS-stimulated monocytes also show that both the p38 and ERK1/2 pathways are involved in the regulation of MMP-1 production by monocytes. However, MMP-9 is primarily regulated through the ERK1/2 pathway. To determine whether p38 and ERK1/2 differentially regulated monocyte MMP-1 and MMP-9 through interaction with specific transcription factors, we assessed the binding of selected oligonucleotides with consensus sequences found in the promoters of these MMPs to nuclear extracts from monocytes. Inhibition of the ERK1/2 pathway resulted in suppression of all of the transcription factors examined. In contrast, of the transcription factors evaluated, only SP 1 and CREB binding was suppressed by inhibition of the p38 pathway. Moreover, the inhibition of CREB binding by SB203580 treatment was reversed by PGE₂, as was the case for MMP-1. These data indicate that p38 regulates MMP-1 mainly through a PGE₂-dependent pathway involving the activation of CREB. In contrast to MMP-1, SB203580-enhanced MMP-9 production correlated with an increase in binding of the NF-κB p50 and NF-κB p65 oligonucleotides that are complementary to an NF-κB binding site in the MMP-9 promoter. Thus, by inhibiting p38 and its negative regulatory effect on ERK1/2, phosphorylation of NF-κB is increased, resulting in enhanced MMP-9 production. Moreover, the SB203580-mediated increase in MMP-9 was further enhanced by the addition of PGE₂, indicating that the increase in NF-κB activity in the MMP-9 promoter was, at least in part, regulated by PGE₂.

In summary, these findings demonstrate that the LPS-mediated signal transduction leading to the production of MMPs by monocytes involves a complex interaction between the MAPKs that results in a differential regulation of MMP-1 and MMP-9. Understanding this interaction may provide insight into appropriate therapeutic intervention in various disease states.

We thank Drs. Nancy Francis and Nancy Vazquez-Maldonado for their critical review of the manuscript.