Mechanical Strain Induces Collagenase-3 (MMP-13) Expression in MC3T3-E1 Osteoblastic Cells*

Mechanical strain plays a crucial role in bone remodeling during growth and development and healing of bone besides systemic and local factors. One of the major factors involves in remodeling process is matrix metalloproteinases (MMPs) such as MMP-13 that has been shown to degrade the native interstitial collagens in several tissues. To study how mechanical strain affects extracellular matrix degradation by MMP-13, a biaxial strain was applied to MC3T3-E1 osteoblastic cells plated onto a collagen-coated flexible elastic membrane. The MMP-13 protein and mRNA expression were determined by Western blotting and reverse transcriptase-PCR, respectively. The zymographic activities of MMP-13 increased dramatically at 30 min, reached a peak by 2-fold at 1 h, and maintained up to 4 h. Moreover, the MMP-13 and c-fos mRNA expressed at 5 min, increased to 2.8- and 3-fold at 1 h, respectively, and gradually declined thereafter. Cycloheximide and actinomycin D did not inhibit the MMP-13 and c-fos mRNA expression, suggesting that such expression does not require de novo protein synthesis and not change their stabilities. To investigate which of the mitogen-activated protein kinase (MAPK) pathways involves in the expression of MMP-13, inhibitors such as PD98059, SB203580, and SP600125 were used. However, only PD98059 (an inhibitor of MEK1/2 activation) inhibited MMP-13 and c-fos gene expression; the result was further substantiated by transfecting with the dominant negative mutants of MEK1/2 (MEK K97R) and ERK2. Taken together, our results showed that mechanical strain induces the MMP-13 expression through MEK-ERK signaling pathway to regulate mechanical adaptation.

Bone is constantly remodeled throughout life to meet the functional demands of its physiological and mechanical environment (1)(2)(3)(4). Remodeling of bone is regulated by a wide variety of systemic and local factors such as 1␣,25-dihydeoxyvitamin D 3 , parathyroid hormone, calcitonin, sex hormones, prostaglandin, transforming growth factor-␤, bone morphogenetic protein, and insulin-like growth factors. It is also a well known fact during growth and in the adult skeleton, a moder-ate level of mechanical loading is considered essential for maintaining and adaptation to the physiological bone remodeling (4,5). In particular, mechanical loading on osteoblasts in vitro has been demonstrated to increase prostaglandin release (6), stimulate cell division (7), alter collagen synthesis (8), and promote collagenase activity (9).
The remodeling of the mineralized connective tissue involves coupling of the degradation of the extracellular matrix (ECM) 1 with the synthesis of new matrix components. Although osteoblasts have been implicated only in bone formation, in response to resorption stimulators, they may cease synthesis of collagen and start secreting neutral proteinases such as collagenase (10 -12). These collagenases may degrade the unmineralized osteoid layer covering bone surfaces, leading to the exposure of the mineralized matrix to osteoclasts. On the other hand, recent findings by Holliday et al. (13) suggested that collagenases enable osteoblasts to initiate bone resorption by generating collagen fragments, which in turn activate osteoclasts.
Matrix metalloproteinase (MMP)-13, one of the collagenase subgroup of MMPs, is a neutral proteinase capable of degrading native fibrillar collagens in the extracellular space (14,15). It cleaves type I, II, and III collagens at a specific site generating 3/4 N-terminal and 1/4 C-terminal fragments, which rapidly denature in physiologic temperature and become susceptible to degradation by other MMPs such as gelatinases. Furthermore, MMP-13 has been demonstrated to cleave type II collagen more readily than types I and III (16,17). Secreted as proenzyme, it consists of a small propeptide cleaved for activation, a catalytic domain, a linker, and a C-terminal hemopexinlike domain that may be involved in matrix binding and alignment of the catalytic domain at the site of enzymatic cleavage (15). The catalytic domain in mouse MMP-13 has 97% identity to rat and 93% identity to human (18). The induction of MMP-13 by growing dermal fibroblasts in three-dimensional type I collagen has been analyzed (19). The results showed that among the three distinct classes of mitogen-activated protein kinases (MAPKs), the p38 MAPK dominantly activates, whereas p42/p44 MAPK inhibits, the MMP-13 expression.
A plausible function of MMP-13 may be involved in situations where rapid and effective remodeling of collagenous ECM is required. Originally identified in human breast carcinoma, MMP-13 was also present with limited amount in normal chondrocytes and osteoblasts (17,20,21). It plays a significant physiological role during bone morphogenesis as in primary fetal ossification and during remodeling of the mature skeletal tissue (22,23). Moreover, MMP-13 is also implicated in excessive degradation of collagenous ECM in chronic inflammatory diseases such as osteoarthritic cartilage, rheumatoid synovium, chronic ulcer, intestinal ulcerations, and periodontitis and in malignant tumors such as breast carcinomas.
Osteoblasts have been shown to respond to mechanical loading of bone tissue by changes in enzyme activity and protein production. It remains unclear whether mechanical loading directly coupled to specific gene expression and protein synthesis without participation of circulating systemic or local regulatory factors. In the present study, we cultured MC3T3-E1 osteoblasts onto collagen-coated elastic membrane with a defined serum-free medium, then imposed mechanical strain in vitro to the adherent cells and later examined the induction of the MMP-13 expression.

MATERIALS AND METHODS
Murine osteoblast-like cell line, MC3T3-E1, a homogeneous source of non-transformed cell, was a gift from Dr. Hyun-Mo Ryoo (Department of Oral Biochemistry, Kyungpook National University, Taegu, Korea). Dominant negative MEK K97R was a gift from Dr. K. L. Guan (Department of Biological Chemistry, University of Michigan, Ann Harbor, MI). Dominant negative ERK2 was a generous gift from Dr. Melanie Cobb (Department of Pharmacology, University of Texas Southwestern Medical Center, Dallas, TX). Fetal bovine serum (FBS) and minimal essential medium-␣ (␣-MEM) were purchased from Invitrogen. The anti-MMP-13 mAb was purchased from NeoMarkers (Fremont, CA). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) antibody was from Biogenesis (Boumemouth, UK). The primer sets of the mouse c-fos (266 bp) and 18 S (495 bp) were purchased from Ambion (Austin, TX). PhosphoPlus p38, p42/p44 MAPK, and c-Jun N-terminal kinase (JNK) antibody kits were from Cell Signaling (Beverly, MA). PD98059, SB203580, and SP600125 were from Biomol (Plymouth Meeting, PA). Bicinchoninic acid protein assay kit was from Pierce. Enhanced chemiluminescence (ECL) Western blotting detection system and Hyperfilms were from Amersham Biosciences (Buckinghamshire, UK). Type I collagen, enzymes, and other chemicals were from Sigma.
In Vitro Equibiaxial Stretch Device-To deliver uniform and isotropic tensile strains to osteoblasts in the absence of shear, we used the equibiaxial stretch device (hereafter called biaxial stretcher), which was modified from the original work of Lee et al. (24). A biaxial stretcher is a cell culture chamber, the lid of which is obtained from a 35-mm cell culture dish in diameter, whereas the bottom, the silicone sheet, is suspended and secured by the screw base cylinder. The middle portion contains three concentric cylinders: an inner indenter ring, a membrane holder, and an outer screw top. The stretcher is passive when an indenter ring is resting on the elastic sheet. The degree of stretch is calibrated by the turns of the screw top, which in turn presses down the indenter ring. Biaxial strain was applied when the indenter ring was pressed against the elastic sheet toward the base (Fig. 1). It has been reported that tensile equibiaxial strains of 3 and 6% yield cell area ratios of 1.06 and 1.12, which are equivalent to the cell area changes produced by 10 and 20% uniaxial strain (25). In the present experiments, we used 8% stretch for the optimal expression of MAPKs (data not shown). There was no significance in release of lactate dehydrogenase in test versus control culture medium and no slippage of the strain cells from the collagen-coated membrane with a prolonged period of time. The assembled stretchers were H 2 O 2 gas-sterilized before use.
MC3T3-E1 Osteoblastic Cells-The MC3T3-E1 cells were grown to subconfluence in ␣-MEM containing 10% FBS before plating to biaxial stretchers. At first, to promote cell attachment, elastic membranes of the biaxial stretcher were coated with a solution of 0.01% type 1 collagen overnight and dried. Then, MC3T3-E1 osteoblastic cells were plated at a density of 1 ϫ 10 5 cells/cm 2 to the collagen-coated sheet of the stretchers and grown to confluence. After changing to serum-free ␣-MEM medium for 24 h, quiescent adherent cells were stretched under testing condition. For experimental purposes, the specific inhibitors or blockers were added 1 h prior to stretching. Control cells were treated in an identical fashion as test cells except for stretching. Test and control experiments were carried out simultaneously with the same pool of cells in each experiment to match temperature, CO 2 content, and pH of the medium for the stretched and control cells.
Zymogram, Western Blot Analysis, and Immunoprecipitation-Un-less mentioned otherwise, protein concentrations were determined (26) with bovine serum albumin as the standard. In this session, the MMPs secreted into the culture medium by strain cells were collected and analyzed. Moreover, the anti-MMP-13 mAb was used to identify by Western blotting and to immunoprecipitate the MMP-13 molecule in the culture medium. Both fractions of immunoprecipitant and of supernatant were further subjected to zymographic analysis. Briefly, aliquots of the control, test medium, and fractions from immunoprecipitant and supernatant were electrophoresed on a 10% SDS-polyacrylamide gel containing 1.25% gelatin. Afterward, the gel was washed with 2.5% Triton X-100 to remove SDS, rinsed with 50 mM Tris-HCl, pH 7.5, and then incubated overnight at room temperature with the developing buffer (50 mM Tris-HCl, pH 7.5, 5 mM CaCl 2 , 1 M ZnCl 2 , 0.02% thimerosal, 1% Triton X-100). The zymographic activities were revealed by staining with 1% Coomassie Blue and, subsequently, destaining of the gel and were quantified by laser densitometry of the corresponding bands in the linear response range of the gelatin zymogram. RNA Isolation, Reverse Transcription (RT), and PCR-The adherent cells were harvested after strained for various times, and total RNA was isolated using TRIzol reagent (Invitrogen) according to the manufacturer's instructions and quantified by optical density. One g of total RNA was added to a RT reaction in RT buffer containing 20 mM Tris-HCl (pH 8.4), 50 mM KCl, and 2.5 mM MgCl 2 , 10 mM dNTPs, 0.1 M DTT, 0.5 mg oligo(dT) primer, 200 units of SuperScript II RT and Rnase H. Five l of cDNA from the RT was added directly to a 50-l PCR containing 20 mM Tris-HCl (pH 8.4), 50 mM KCl, 25 mM MgCl 2 , 10 mM dNTP, 2.5 unit of Taq DNA polymerase. The amplification conditions were as follows: 94°C/1 min, 62°C/1 min, 72°C/2 min (for MMP-13); 94°C/45 s, 60°C/45 s, 72°C/2 min (for c-fos); 94°C/1 min, 63°C/1 min, 72°C/1 min (for 18 S), all of which were amplified for 30 cycles. Oligonucleotide primers were designed to span at least one intron to detect any contaminating genomic DNA carried over from the RNA isolation step. Eighteen S ribosomal RNA and c-fos primer sequences have been described previously (Ambion), and MMP-13 primer sequences were derived from the mouse MMP-13 sequence (27) as follows: MMP-13, 5Ј-GGT CCC AAA CGA ACT TAA CTT ACA-3Ј (sense primer) and 5Ј-CCT TGA ACG TCA TCA TCA GGA AGC-3Ј (antisense primer), a total of 445 bp. Conditions were established so that PCR was stopped in the linear range, and the reaction products could be accurately quantified and compared. PCR products were electrophoresed on 1.5% agarose gels. Ethidium bromide staining of the bands corresponding to MMP-13, c-fos, and 18 S ribosomal RNA was photographed and digitized by using a Kodak DC290 Zoom digital camera. Density analysis  (24). The addition to the device was a circular screw base cylinder, which secured and suspended the elastic sheet. For ease of operation, the bottom of the sheet did not contact any surfaces of the working area such as the tissue culture hood or the incubator. Cells were plated and grown on the collagen-coated elastic sheet.
was performed using the UN-SCAN-IT gel program (Silk Scientific, Inc. Orem, UT). The levels of MMP-13 or c-fos mRNA were normalized to those of 18 S ribosomal RNA to correct for differences in loading and/or transferring.
Preparation of Cell Extracts and Western Blot Analysis of MAPKs-Cells in control or in test (by 8% stretch) groups were incubated for various times before subjected to cell lysis as described previously (28). At the termination of mechanical stimulation, the osteoblastic cells were rapidly washed with ice-cold PBS twice and lysed on ice in 0.2 ml of lysis buffer containing 25 mM Tris-HCl (pH 7.4), 25 mM NaCl, 25 mM NaF, 25 mM Na 4 P 2 O 7 , 1 mM Na 2 VO 4 , 2.5 mM EGTA, 2.5 mM EDTA, and 0.05% Triton X-100, 0.5% Nonidet P-40, 0.5% SDS, 0.5% deoxycholate, and protease inhibitors such as 5 g/ml leupeptin, 5 g/ml aprotinin, and 1 mM phenylmethylsulfonyl fluoride. The lysates were centrifuged at 45,000 ϫ g for 1 h at 4°C to yield the cell extract. Equal amounts of samples were electrophoresed on 10% polyacrylamide gel and were then blotted to nitrocellulose membrane. Subsequently, the membrane was incubated at room temperature with 5% bovine serum albumin in TTBS (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.05% Tween 20) for 1 h. The total protein profile and the phosphorylated forms of the MAPKs were identified by Western blotting analysis using anti-phospho-p38, antiphospho-p42/p44, and anti-phospho-JNK, respectively. Briefly, membranes were incubated first with a 1:1000 diluted solution of specific anti-phospho-MAPK Ab or anti-GAPDH Ab and then with the second antibody (anti-rabbit horseradish peroxidase antibody in 1% bovine serum albumin/TTBS, 1:1500 dilution). Immunoreactive bands were visualized by an ECL system.
Plasmids and Transfection-The plasmids encoding K52R ERK2 and MEK K97R (dominant negative mutants of ERK2 and MEK1/2) were generously provided by Dr. M. Cobb (Department of Pharmacology, University of Texas Southwestern Medical Center, Dallas, TX) and by Dr. K. L. Guan (Department of Biological Chemistry, University of Michigan, Ann Harbor, MI), respectively. All plasmids were prepared by using Qiagen plasmid DNA preparation kits. For transfection, the amount of plasmid (0.7 g) was kept constant for each experiment. The DNA PLUS-LipofectAMINE reagent complex was prepared according to the instructions of the manufacturer (Invitrogen). The adherent MC3T3-E1 osteoblastic cells grown to 70% confluence were washed once with PBS and then with serum-free ␣-MEM. Subsequently, cells were transfected and incubated with plasmid in serum-free ␣-MEM (0.8 ml) and DNA PLUS-LipofectAMINE reagent (0.2 ml) at 37°C for 5 h and later with ␣-MEM medium (1 ml) containing 10% FBS overnight. After 24 h of transfection, cells were washed twice with PBS and maintained in ␣-MEM containing 10% FBS for additional 24 h. Before applying 8% stretch, cells were washed once with PBS and incubated with serumfree ␣-MEM for 24 h.
Statistics-Data were presented as mean Ϯ S.E. Statistical comparisons of control group with treated groups were carried out using the paired sample t test with p values corrected by the Bonferroni method. Comparisons among three or more groups were made by one-way analysis of variance followed by Dunnetts' modified t test for post hoc analysis. Significance was accepted at the p Ͻ 0.05 level.

RESULTS
Mechanical strain generated by physiologic demand may dictate local bony architecture through bone remodeling. MMP-13 has been proposed to participate actively in situations where rapid and effective remodeling of collagenous ECM is required. Hence, it is reasonable to investigate the molecular mechanism leading to the expression of the MMP-13 by applying mechanical strain to osteoblasts.
MMP-13 Secretion by MC3T3-E1 Osteoblastic Cells in Response to Mechanical Strain-MC3T3-E1 osteoblastic cells were plated to the collagen-coated elastic membrane of the stretcher, grown to confluence, conditioned in serum-free ␣-MEM medium for 24 h, and then subjected to test by 8% stretch. The media were collected at 30 and 45 min and run under SDS-PAGE. The anti-MMP-13 was included for the works of Western blotting to identify MMP-13 (Fig. 2, lane 1) and of immunoprecipitation (Fig. 2, lanes 4 -7). Moreover, to detect the activities of the MMP-13, zymogram was incorporated (Fig. 2, lanes 2-7).
To identify and reveal the presence of the MMP-13 molecule in the test medium, the large amount of 2 ml of sample medium was pooled, precipitated by trichloroacetic acid, and centrifuged. The pellet was subjected to Western blotting analysis. These results revealed that MMP-13 molecule, present in the test medium, was recognized by anti-MMP-13 mAb (Fig. 2, lane 1) at a molecular mass of 60 kDa as pro-MMP-13 form in agreement with previous reports (29). To differentiate the activities of the MMP-13 between the control and the test culture media, a small amount (200 l) of the sample medium was collected and examined under zymography. The results showed that the control medium contained trace amount of the zymographic activity (Fig. 2, lane 2). In contrast, when adherent cells were strained for 30 min, the activity of MMP-13 appeared in the test medium (Fig. 2, lane 3). Taken together, these data indicated that being present in the test medium, the MMP-13 itself matched with the co-migrated zymographic activity at the molecular mass of 60 kDa.
To ascertain that the zymographic activity is expressed exactly by MMP-13, the anti-MMP-13 mAb (as in Fig. 2, lane 1) was once again used to immunoprecipitate the molecule in the test medium (as in Fig. 2, lane 3). Both fractions of the pellet and the supernatant were run separately to localize the distribution of the zymographic activities of MMP-13. Only trace amounts of activity showed in the supernatant fraction (Fig. 2,  lanes 4 and 5), whereas the majority of the MMP-13 zymographic activities appeared in the pellet fraction (Fig. 2, lanes  6 and 7). Furthermore, the response of MC3T3-E1 osteoblasts to strain in vitro appeared to be time-dependent (Fig. 3). The zymogram revealed that the strain-induced activities of MMP-13 increased with time, were dramatically enhanced by 50% at 30 min, reached plateau by 1.8-fold increase at 1 h, and remained up to 4 h.  6 and 7) were compared.

Strain Induces Expression of MMP-13 and c-fos Genes in MC3T3-E1 Osteoblastic Cells-
The early appearance of the MMP-13 molecule in the test medium prompted us to examine the time-dependent manner of its gene expression by mechanical induction (Fig. 4A). In parallel, the inductive expression of c-fos gene, known as an immediate early mechanoresponsive gene, was included (Fig. 4B). It was clear that the induction of the MMP-13 and c-fos genes increased dramatically to 2.8-and 3-fold, respectively, within an hour by mechanical strain, tapered gradually afterward, yet was prolonged and maintained up to 2-fold even 4 h later. The inductive profiles of the mRNA level of MMP-13 (Fig. 4) ran in parallel with that of its zymographic activities (Fig. 3). Thus, the results confirmed that the early onset of the induced gene of the enzyme corresponded with its prompt appearance as the protein profile in the test medium. Furthermore, the findings that the inductive profile of the MMP-13 genes matched with that of the c-fos gene suggest that MMP-13 should also be recognized as an immediate early mechanoresponsive molecule.
To determine whether the effect of mechanical strain on mRNA expressions of c-fos and MMP-13 were dependent on de novo protein synthesis, adherent cultures of osteoblastic cells were stretched by 8% in the absence or presence of cycloheximide at doses known to inhibit protein synthesis (30). Treatment with cycloheximide alone did not affect the mRNA expressions of the c-fos and MMP-13. Moreover, the same results were produced by treatment with osteoblast stretching alone or in combination with cycloheximide, suggesting that MMP-13 (Fig. 5A) and c-fos (Fig. 5B) gene expression induced by mechanoinduction was a primary response through the activation of pre-existing transcriptional factors and was not dependent on de novo protein synthesis.
To determine whether effects of mechanical strain on c-fos and MMP-13 mRNA levels were due to changes in transcription stability, osteoblastic cells were exposed to control or strain for 1 h and then treated with RNA polymerase II inhibitor actinomycin D at 10 M (31), in the presence or absence of strain for 1, 2, 3, 5, and 9 h. The half-life of MMP-13 mRNA in transcriptionally arrested osteoblastic cells was 4 h in control and test cultures (Fig. 6A). Slope analysis indicated that con-trol and strain did not change the stability MMP-13 mRNA. Similar findings were obtained by examining the c-fos transcriptional mechanism (Fig. 6B).

Involvement of p42/p44 MAPK in Strain-induced MMP-13 Gene Expression in MC3T3-E1
Osteoblastic Cells-Three distinct signaling pathways of MAPKs have been identified. They are p42/p44 MAPK cascade, which preferentially regulates cell growth and differentiation, JNK, p38 MAPK cascades, which function mainly in stress responses like inflammation and apoptosis (32). To determine whether mechanical strain stimulated MMP-13 expression through activation of these MAPK cascades, the cells were preincubated with specific MAPK inhibitors such as PD98059 (for MEK1/2, upstream of p42/p44 MAPK), SB203580 (for p38 MAPK), and SP600125 (for JNK), respectively. As shown in Fig. 7, A and B, mechanical strain stimulated phosphorylation of p42/p44 MAPK, p38, and JNK-1 and JNK-2 in MC3T3-E1 osteoblastic cells determined by Western blotting using antiserum reactive with anti-phospho-p42/p44-MAPK, p38, JNK-1, and JNK-2 antibodies. The phosphorylation of these MAPKs was inhibited by pretreatment of MC3T3-E1 osteoblastic cells with respective inhibitors. The membranes were then stripped, and the amounts of GAPDH were determined with the use of an antibody that recognized GAPDH as an indicative of protein loading in each well. There was no significant difference in the amount of GAPDH among these samples. The results showed that the activities of these three MAPKs were all stimulated by the mechanical strain and were abolished by the inclusion of the specific inhibitors.
To determine which of the MAPK cascades may mediate the expression of the c-fos and MMP-13 genes induced by mechanical strain, the cells were preincubated with specific MAPK inhibitors. At the mRNA level, the amounts of the expressed c-fos and MMP-13 genes were determined by RT-PCR (Fig. 7, C  and D). Only the MEK1/2 inhibitor, PD98059, was able to diminish the expression of both c-fos and MMP-13 genes in MC3T3-E1 osteoblastic cells. SB203580 and SP600125 had no significant effect on these responses. The results clearly pointed out that the MEK/p42/p44 MAPK pathway mediated the mechanical strain-induced expression of c-fos and MMP-13 genes. To substantiate these findings, we transfected MC3T3-E1 osteoblastic cells with the dominant negative mutants of both MEK1/2 (MEK K97R) and ERK2 to specifically block the MEK1/2-p42/p44 MAPK signaling pathway leading to the expression of both genes. As shown in Fig. 8, transfection with MEK K97R and dominant negative ERK2 plasmids suppressed the mechanical strain-stimulated expression of MMP-13 and c-fos genes, confirming the MEK1/2-p42/p44 MAPK pathway involved in the expression of the c-fos and MMP-13 genes in response to mechanical strain. DISCUSSION Our results demonstrate that MMP-13 fulfills the following criteria as a candidate of the mechanical strain-induced adaptive molecule in osteoblastic cells in vitro. First, the secreted molecule of the MMP-13 was obtained from the test culture medium. This molecule qualified as the MMP-13 properties in that it appeared as a latent form with molecular mass of 60 kDa, reacted with antibodies raised against anti-MMP-13 mAb, and co-migrated with the zymographic activities (Figs. 2 and  3). The identity of the molecule and the co-migrated zymographic activities were further verified by immunoprecipitation against anti-MMP-13 mAb to separate supernatant and pellet fractions from the test culture medium. The results showed that the activities were depleted in the supernatant fraction by anti-MMP-13 mAb and were recovered in the pellet fraction, confirming the secretion of MMP-13 from the cells stimulated by mechanical strain. Furthermore, when mechanical strain FIG. 3. Effect of strain on secretion of MMP-13 from MC3T3-E1 osteoblastic cells. Cells were plated at a field density of 1 ϫ 10 5 cells/cm2 on a collagen-coated elastic membrane and grew to confluence. Culture medium was changed to serum-free ␣-MEM medium (1 ml/chamber) for 24 h. After treatment with or without 8% strain, media were removed, and cells were harvested at the times indicated. Equal amounts of protein in the media were loaded, and the MMP-13 activities were assessed by zymogram after SDS-PAGE on a 10% gel and quantified by densitometry of the corresponding bands in the linear response range of the zymogram as described under "Materials and Methods." *, p Ͻ 0.05, as compared with the cells without strain.
was applied, MMP-13 expression was enhanced with time. Second, the expressions of the MMP-13 mRNA in parallel with that of the c-fos mRNA was time-dependent, stable, and did not involve de novo protein synthesis (Figs. 4 -6). Finally, the signaling pathway leading to the gene expression of both MMP-13 and c-fos was blocked by the MEK1/2 inhibitor PD98059, spe-cifically by transfection with the dominant negative mutants of MEK1/2 and ERK2 (Figs. 7 and 8). Since all the experiments were performed in serum-free condition, and the adherent cells were synchronized overnight to a resting state, the response of the osteoblasts would be attributed to the strain generated by the biaxial stretchers. Thus, our results support the notion that the mechanical strain specifically induces MMP-13 expression in osteoblasts, and the expression is mediated by the MEK1/2-p42/p44 MAPK signaling pathway.
The induction of MMP-13 has been demonstrated to be mediated through cell-matrix interactions by growing periodontal ligament cells in two-dimensional collagen (33) or by growing dermal fibroblasts in three-dimensional collagen (19). Accordingly, in our biaxial stretcher system where osteoblasts were grown on collagen-coated elastic membrane, a basal level of MMP-13 activity appeared. Interaction of cells with collagen matrix could lead to activation of a signaling pathway such as p42/p44 MAPK. It has been reported that the activation of p42/p44 MAPK can be found in osteoblasts grown in twodimensional collagen (34) as well as in fibroblasts inhabited in two-and three-dimensional collagen (35). However, in cases where dermal fibroblasts were grown in three-dimensional collagen, the induction of MMP-13 in collagen gel showed otherwise. It was inhibited by p42/p44 MAPK (19). Instead, the p38 MAPK was reported to play a dominant role in the induction of MMP-13. Our approach to the MMP-13 induction is somewhat different from previous work of dissecting signaling through cellular contact with collagen. In a biaxial stretcher system, cells are grown to confluence on collagen-coated flexible membranes, changed to serum-free condition, mechanical strain imposed to these cells, and then the cells are examined to determine which signaling molecule(s) mediate MMP-13 induction by the use of the specific MAPK inhibitors such as PD98059, SB203580, and SP600125. These results showed that all three known MAPKs, including p42/p44 MAPK, p38 MAPK, and JNK, were activated. However, the induction of the MMP-13 gene by the participating MAPKs was quietly selective and specific. We noticed that only p42/p44 MAPK mediated the MMP-13 expression induced by mechanical strain, similar to the activation by contact of osteoblasts with two-dimensional collagen. The activated p38 MAPK and JNK may be involved in a signal relaying other molecules yet to be defined. Even though the design of the experiments varies, it remains to be determined whether the differences in the signal activation of p38 MAPK or p42/p44 MAPK lie in the origin of cells or in the dimension of collagen. It appears that osteoblasts strategically adopt a straight forward protocol of using p42/p44 MAPK signaling pathway to deal with the contact of cells with collagen and the mechanical load. This notion was also reflected by the pattern of the c-fos induction from our results and others (36). Moreover, in osteoblasts, the induction of the c-fos gene has been shown to be dependent either on a protein kinase C-or protein kinase A-mediated signaling pathway via gravitational loading at large or mild magnitude stress, respectively (37,38). It remains to be determined which of the upstream signaling molecules mediate the expressions of c-fos and/or MMP-13 genes.
The c-fos gene is the best studied member of the cellular immediate-early genes whose transcription is activated rapidly and transiently within minutes of growth factor stimulation. Since the c-fos gene has been reported to be induced by a uniaxial stretcher system (39), we examined the c-fos induction by our biaxial stretcher system. The results revealed that the c-fos induction by biaxial strain behaved similar to that by uniaxial strain. Apparently the c-fos induction becomes a common mechanoresponsive pattern despite the origin of the strain, be it static or dynamic, or generated by stretching (40), vibration, or fluid flow (41)(42)(43). Such phenomenon may leave us to ponder whether a single application of strain above cellular activity is sufficient by osteoblasts to initiate c-fos-related cellular responses. What also interested us were the findings that the induction of MMP-13, a phenotype, runs in parallel with that of c-fos, a protooncogene, on the basis that early expressions of both genes were time-dependent, stable, did not involve de novo protein synthesis, and shared common signaling p42/p44 MAPK. These findings indicated that the regulations of both genes in response to mechanical strain were not at the post-transcriptional level. On the other hand, their immediate appearances may lie in the inherent AP-1 binding site in the promoter regions of both genes (44). Nevertheless, one may learn from these coordinated efforts that osteoblasts, when stimulated by mechanical strain, regulate mechanical adaptation promptly and simultaneously, not only intracellularly but extracellularly.
Osteoblasts and osteoclasts participate actively in the resorption and apposition phases of bone (re)modeling. When subjected to chronic intermittent strain by hydrostatic compression or "Flexcell" stretch, however, they take time to synthesize and secrete molecules needed in the apposition phase. Their phenotypic genes such as alkaline phosphatase, type I collagen, osteopontin, and osteocalcin require days before appearance (45,46). When osteoblasts were under dynamic uniform biaxial strain for a 2-h period, osteopontin genes were reported to appear by 3 h and continue up to 9 h (47). As for the time required to carry on the post-translational modification of the osteopontin, for example, we have shown that it takes 10 -30 min to metabolically label the phosphorylated form of the molecule before secretion (48). In normal tissue, collagen secreted from rat incisor odontoblasts required 30 and 60 min (49). These radiographic results were confirmed in biochemical analyses by Dimuzio and Veis (50), who found a 45-min minimum secretion time. In contrast, MMP-13, regarded as an immediate-early mechanoresponsive molecule, synthesized and secreted by strained osteoblasts, reached the ECM space rapidly. The results showed that it took a minimum of 15 min for MMP-13 secretion to appear even though the secreted enzyme was in its latent form with a molecular mass of 60 kDa (Fig. 2). Moreover, in a serum-free condition but with a single static strain, there was no delay in synthesis and delivery of the molecule. The quick response and short secretion time for MMP-13 suggest that an immediate action be taken by strain osteoblasts to participate in the resorption phase of matrix (re)modeling.
The inactive zymogens require activation before attaining catalytic capacity. The active MMP-13 cleaves the undenatured triple-helical collagen molecule into characteristic 3/4 N-terminal and 1/4 C-terminal fragments at a specific site containing the collagenase-susceptible Gly-Leu and Gly-Ile peptide bonds (51). The cleaved fragments denature spontaneously and are further digested by other proteolytic enzymes such as gelatinase. Ultimately, the digested products in turn activate osteoclasts (13). Being the only interstitial collagenase in the mouse, the enzyme may be designed to degrade a wide variety of substrates such as cartilage aggrecan (52) and the telopeptide region of type I collagen (53). Since it is involved in tissue remodeling and/or matrix degradation, the enzyme is widely present in normal cells such as osteoblasts, fibroblasts, and smooth muscle cells, in chondrocytes in cases of osteoarthritis or rheumatoid arthritis, as well as in pathological tissue such as breast carcinoma cells. During development and morphogenesis, the in situ hybridization analyses of later embryonic mouse tissue revealed the expression of MMP-13 (54) and gelatinase B (55) in the onset of bone formation.
Bone architecture is constantly under the influence of the non-mechanical (i.e. systemic and local factors) regulation and particularly the mechanical (or functional) loading (2). The process of bone (re)modeling under mechanical loading may repair fatigue damage and determine bone strength (1,3,4). Moreover, active (re)modeling is demanded to keep pace with the orthodontic tooth movement where the moderate or higher levels of mechanical loading were generated (56). Apparently, such (re)modelings require the participation of the osteoblasts and osteoclasts to carry out the task of matrix degradation in conjunction with new matrix synthesis. A coupling mechanism between osteoblasts and osteoclasts in initiating matrix degradation may be exemplified at the molecular level by the active MMP-13 molecule discussed previously. Here, in this study we were able to demonstrate that mechanical strain induces specifically the MMP-13 expression by osteoblasts. Accordingly, preliminary information regarding the role of osteoblasts under mechanical loading in bone matrix (re)modeling may be delineated as such that the strained osteoblasts secrete MMP-13 to cleave intact collagen, the digested products of which may in turn activate osteoclasts.