Expression of matrix metalloproteinases in articular cartilage of temporomandibular and knee joints of mice during growth, maturation, and aging




This study examined the involvement of different matrix metalloproteinases (MMPs) in articular cartilage in the process of growth, maturation, and aging of mice, and compared the temporal changes in the expression of MMPs between temporomandibular joints (TMJ) and knee joints.


Homogenates of intact tibial plateau, femoral condyle, and TMJ condyle cartilages from animals of different ages were assessed for gelatinase (MMP-2 and MMP-9) activity by zymography. The messenger RNA (mRNA) expression of MMPs 1, 2, 3, 9, and 13 in tibial plateau cartilage was determined by semiquantitative reverse transcription–polymerase chain reaction, and immunohistochemistry was used to localize MMPs 2, 3, 9, and 13 in the knee joints and TMJ from mice of different ages.


The pattern of gelatinase (MMP-2 and MMP-9) activity and their protein expression as well as that of MMPs 3 and 13 varied with the age of the mouse, and differences in expression were observed between the knee and TMJ cartilage. The expression of mRNA for the MMPs in the tibial plateau was also age related.


This study demonstrated changes in the protein and mRNA expression of MMPs 2, 9, 3, and 13 during growth, maturation, and aging in mice. The temporal changes were characteristic of the joint, and distinct differences were observed between the TMJ and knee cartilage. The differences in temporospatial expression of MMPs between the knee joint and TMJ may be the result of differences in load and function of these joints. The information provided in this study contributes to a better understanding of the role of these MMPs in the maintenance and integrity of cartilage tissue.

Cartilage is composed of a small number of chondrocytes within an extracellular matrix of collagen and proteoglycans. In articular cartilage, the proteoglycans are mainly responsible for compressive stiffness, whereas the collagen fibrils provide the tensile strength and maintain the integrity of the cartilage (1). Type II collagen, located almost exclusively in cartilage, is the principal component of articular cartilage, while smaller amounts of minor collagens such as types VI, IX, X, and XI are also observed (2). The integrity of cartilage in normal, healthy adults is dependent on the steady state between the synthesis of the extracellular matrix by chondrocytes and the rate of its degradation. This extracellular matrix remodeling is a critical component of development and normal physiology (3). Any change in this homeostatic steady state rapidly affects the healthy function of the cartilage and may lead to excessive degradation, which is characteristic of the pathogenesis of osteoarthritis (OA) (4, 5).

Important to extracellular matrix remodeling is a unique protease family named the matrix metalloproteinases (MMPs), due to the presence of a zinc ion in their catalytic site. MMPs are a family of >20 structurally related proteolytic enzymes generally secreted as inactive proenzymes that undergo activation primarily in the extracellular milieu. These enzymes can collectively degrade all types of collagen, proteoglycans, and other components of the extracellular matrix (6). In general, MMPs are divided into subgroups according to their substrate specificity: collagenases (MMPs 1, 8, and 13) that degrade fibrillar collagen types I, II, and III, stromelysins (MMP-3 and MMP-11) that degrade proteoglycans and nonhelical regions of collagens, and gelatinases (MMP-2 and MMP-9) that have a high specificity for degraded collagen but are also capable of degrading collagen types IV, V, and XI which are resistant to the MMP collagenases and stromelysins (7). The collagenases and stromelysins show broader substrate specificity than do the gelatinases, while all 3 subgroups may degrade other extracellular matrix components as well. Collagenases initiate degradation of type II collagen by cleaving at a specific site within the triple helix, resulting in fragments that denature spontaneously at body temperature, thus becoming susceptible to degradation by gelatinases that have a higher affinity for denatured collagen (8, 9). It has been established that although individual enzymes have similar substrate specificities, their pattern of expression is often distinct and characteristic of a certain tissue and cell type.

The expression and activity of MMPs is tightly regulated at different levels, comprising the transcriptional level, which involves growth factors and cytokines, and the translational and posttranslational levels, which include inhibition of activity via binding to specific endogenous tissue inhibitors of MMPs, named TIMPs (7, 10). Several MMPs are expressed in skeletal structures. These include interstitial collagenases 1 and 3 (MMP-1 and MMP-13, respectively) (11), gelatinases A and B (MMP-2 and MMP-9, respectively) (8), and stromelysin 1 (MMP-3) (12). The balance in the expression of MMPs to TIMPs in a specific microenvironment determines the course of extracellular matrix remodeling/degradation. Excess expression of MMPs in relation to TIMPs may lead to enhanced proteolysis and joint degradation.

The expression of MMPs and TIMPs during growth and aging of articular cartilage has rarely been described (13, 14). Most of the studies relating to the expression of MMPs in cartilage were conducted at different stages of embryogenesis (15, 16). A number of additional studies on the expression of MMPs in cartilage have been carried out in various models of OA in animals (17) and OA in humans (4, 5, 18, 19).

The aim of the present study was to monitor the messenger RNA (mRNA) levels of MMPs, to determine the activity and localize the expression of MMPs known to participate in the turnover of extracellular matrix in articular cartilage in joints of neonatal, young, matured, and aged mice.



Gelatin, Triton X-100, phenylmethylsulfonyl fluoride (PMSF), 1,10-phenanthroline, leupeptin, APMA, acrylamide/bis-acrylamide, Coomassie brilliant blue R250, agarose, and dNTPs were purchased from Sigma (St. Louis, MO), Moloney murine leukemia virus (MMLV) reverse transcriptase was from Amersham (Cleveland, OH), DNA Taq polymerase was from Roche (Mannheim, Germany), and DNase I, RNAguard, and oligo-hexamer primers were from Pharmacia (Peapack, NJ). Specific MMP primers were synthesized by Genset (Paris, France). RNA was isolated using TriReagent (Molecular Research Center, Cincinnati, OH). Protein concentration of cartilage homogenates was determined using the Bio-Rad protein assay (Bio-Rad, Munich, Germany).

Animals and cartilage harvest.

Female ICR mice that ranged in ages from newborn to 18 months old (1 day, 2 days, 1 week, 2 weeks, 1 month, 4 months, and 18 months old) were used in this study. Ethical guidelines for experimental investigations in animals were followed, and the Institutional Board from the Faculty of Medicine, Technion, Israel, approved the protocols.

Articular cartilage tissues from the temporomandibular joint (TMJ) condyle, femoral condyle, and tibial plateau (knee joint) were carefully dissected and cleaned to avoid any contamination from adjacent bone and ligaments. For RNA assays, tibial plateau cartilage was dissected and immediately transferred to liquid nitrogen. Cartilage samples were stored at −80°C until RNA extraction was performed.

Gelatin zymography.

Dissected intact cartilage was immediately homogenized in 0.5 ml homogenization buffer (50 mM Tris HCl, pH 7.4, 300 mM KCl, and 2.5 mM MgCl2) using Polytron PT (Kinematica, Luzern, Switzerland). Aliquots (10 μl) of cartilage homogenates containing 0.25–2 μg total protein were mixed with nonreducing sodium dodecyl sulfate–polyacrylamide gel electrophoresis sample buffer and electrophoresed in 10% polyacrylamide gels containing 1 mg/ml type A gelatin from porcine skin, at 20 mA constant current for 1.5–2 hours at 4°C. Following electrophoresis, gels were washed twice for 20 minutes with 2.5% Triton X-100 in Tris buffered saline (50 mM Tris HCl, pH 7.5, 150 mM NaCl) buffer to allow protein to renature. Gels were then incubated 20–40 hours at 37°C in substrate buffer (50 mM Tris HCl, pH 8, 10 mM CaCl2, and 0.02% NaN3), stained with Coomassie R250 for 30 minutes and destained for 1 hour. Gelatin-degrading enzymes were visualized as clear bands, indicating proteolysis of the substrate protein. Gels were scanned and analyzed by a Viber Lourmat (Torey, France) imaging system. Molecular weight markers (Bio-Rad) and positive controls, human MMP-2 and MMP-9 (Oncogen Research Products, Boston, MA), were used in all gels.

RNA isolation and reverse transcription–polymerase chain reaction (RT-PCR).

To ensure accurate measurement of in vivo mRNA levels, the RNA was isolated from intact tissue immediately after harvest. Total RNA was extracted from tibial plateau cartilage dissected from mice (6 animals at each time point; 50–80 mg total weight of tissue). Tissue was snap frozen and pulverized to a fine powder in its frozen state in the presence of liquid nitrogen using a mortar and pestle submerged in dry ice/ethanol. The powder was mixed with TriReagent and the supernatant was collected after centrifugation. The RNA dissolved in TriReagent was precipitated with isopropanol according to the manufacturer's instructions (Molecular Research Center). Precipitation was facilitated by the addition of high-salt solution (0.8M sodium citrate and 1.2M NaCl) to prevent possible proteoglycan contamination. Alcohol-washed RNA samples were treated with DNase I to completely remove residual DNA, and the enzyme was removed by a guanidine thiocyanate extraction step. RNA was ethanol precipitated and suspended in RNase-free water. The RNA purity and the yield were determined by spectrophotometry, measuring the absorbance of an aliquot at 260 nm and 280 nm.

Complementary DNA (cDNA) was prepared by RT at 37°C for 60 minutes in 50 μl reaction mixture containing 10 μg RNA, 400 units MMLV, RNAguard, and oligo-hexamer primers. The enzyme was heat inactivated and cDNA from each of the samples was amplified by PCR, using actin-specific primers, to assess RNA integrity. The PCR mix (25 μl volume) included reaction buffer (10 mM Tris HCl, 50 mM KCl, 1.5 mM MgCl2, 0.1% Triton X-100), 0.2 mM of each dNTP, 0.4 μM sense and antisense primers, and 1 unit DNA Taq polymerase. Thirty-five cycles of PCR amplification were performed, each consisting of denaturation at 94°C, annealing at 58°C, and extension at 75°C, in a thermal cycler (Biometra, Gottingen, Germany). PCR products were visualized by ultraviolet light following electrophoresis of products in ethidium bromide–stained 3% agarose gels. Standard molecular markers, negative controls (PCR mix without sample cDNA), and positive controls were run with each PCR assay.

Semiquantitative analysis of mRNA expression of MMPs.

Semiquantitation of mRNA expression of MMPs was performed as previously described (20), using murine-specific primer sequences for each of the MMPs examined (MMP-1, MMP-2, MMP-3, MMP-9, and MMP-13). Linear phase of the reaction was achieved by modifying the quantity of cDNA and the number of cycles for each of the genes separately. Three serial dilutions of cDNA were amplified in the PCR assay for each time point to confirm the linearity of the reaction (see Figure 4A). In addition, the cDNA samples from the animals of different ages were subjected to PCR amplification at the same time to unify the conditions as much as possible. PCR primers (designed using Gene Runner; Microsoft Corporation, Redmond, WA) and product sizes are described in Table 1. The PCR amplification mixture was as described above for actin, except for the use of specific primers and stringent annealing temperatures (47–68°C) for each MMP.

Figure 4.

Expression of mRNA of MMPs in tibial plateau cartilage. Reverse transcription–polymerase chain reaction (RT-PCR) analysis was performed for mRNA obtained from tibial plateau cartilage. Three different cDNA concentrations representing 3 different mRNA concentrations were amplified for samples from mice of different ages (1 and 2 days [1d and 2d, respectively], 1 and 2 weeks [1w and 2w, respectively], and 1, 4, and 18 months old [1m, 4m, and 18m, respectively]). PCR products, positive control (PC), and molecular size markers (M) were separated by electrophoresis. The optical densities of the PCR products were evaluated by computerized densitometry. A, Representative RT-PCR results for MMP-3, MMP-13, and actin. B, Histograms demonstrating the ratio between the densitometric assessment of each MMP and that of actin mRNA. Bars show the mean and SD of 3 assessments. See Figure 1 for other definitions.

Table 1. Primer pairs for semiquantitative polymerase chain reaction*
Gene evaluatedPrimersProduct size, basepairs
  • *

    MMP = matrix metalloproteinase.

Figure 1.

Differential expression of gelatinases in articular cartilage of 1-month-old mice. Gelatin zymography was performed to determine the expression of gelatinases in homogenates (2 μg protein) prepared from articular cartilage of the tibial plateau (TP), femoral condyle (F), and temporomandibular joint condyle (TMJ) obtained from 1-month-old mice. Commercial matrix metalloproteinase 2 (MMP-2) (P-2) and MMP-9 (P-9) served as positive controls. Gelatinase activity typical of MMP-9 latent form (MMP-9L), MMP-2 latent form (MMP-2L), and MMP-2 active form (MMP-2A) was observed. The position of molecular weight markers (MW), electrophoresed in parallel, is denoted. A zymogram representative of 3 separate experiments is demonstrated.

The PCR products were fractionated on 3% agarose gels. Optical densities of resulting PCR products were measured by computerized video densitometry (Bio Imaging Gel documentation system; Dinco and Renium, Jerusalem, Israel) using Tina software (Raytest, Straubennhardt, Germany). The relative optical densities of the MMPs were calculated relative to actin product to correct for possible variance of sample preparation. Linear regression was determined for each time point for each gene to ensure linearity of results (r2 > 0.9 for all samples).


Dissected TMJ and knee joints were fixed in phosphate buffered saline (pH 7.4) containing 4% paraformaldehyde, decalcified in 10% EDTA, dehydrated, and embedded in paraffin and sectioned sagittally. Sections of 6 μm were processed for immunohistochemistry using UltraVision Mouse tissue detection system with horseradish peroxidase/aminoethylcarbazole (Lab Vision, Fremont, CA). The primary monoclonal antibodies used were against MMPs 2, 3, 9, and 13 (Oncogen Research Products). Immunostained sections were counterstained with hematoxylin.


Characterization of gelatinase activity in articular cartilage from temporomandibular and knee joints.

Gelatin zymography of the 3 different cartilage tissues (TMJ condyle, femoral condyle, and tibial plateau) from 1-month-old mice demonstrated the expression of both MMP-2 (latent form 72 kd, and active form 66 kd) and MMP-9 (mainly, latent form ∼97 kd) (Figure 1). Distinct differences were observed in the level of gelatinase activity between the 3 tissues examined. Densitometric assessment showed that the level of MMP-9 expression was the highest in the femoral condyle, being at least 5-fold higher than in the TMJ cartilage and 1.5-fold higher than in the tibial plateau cartilage. Differences were also observed in the level of MMP-2 active and latent forms when comparing the 3 cartilages. The highest level of latent-form MMP-2 was observed in the tibial plateau cartilage, while the highest level of active MMP-2 was found in the TMJ cartilage. In addition, the ratio between active and latent MMP-2 forms was higher (4.3-fold) in the TMJ compared with the other cartilages examined.

The properties of the gelatinases were characteristic of MMP-2 and MMP-9, since they were converted to active, lower molecular weight forms by APMA treatment (a known MMP activator) and their activity was completely abrogated by the presence of 2 mM 1,10-phenanthroline (zinc-chelating agent), but not by serine (PMSF) or aspartic (pepstatin A) proteinase inhibitors (Figure 2). Treatment of cartilage homogenates with APMA transformed the 97-kd gelatinase into an 82-kd molecule, and part of the 72-kd enzyme to lower molecular weight forms (Figure 2).

Figure 2.

Gelatinase activity in articular cartilage characteristic of MMP-2 and MMP-9. Homogenates prepared from the tibial plateau, femoral condyle, and temporomandibular joint condyle articular cartilage obtained from 3-month-old mice were treated with 1 mM APMA for 3 hours (3h) or overnight (ON) at 37°C and separated by electrophoresis on gels containing gelatin; nontreated homogenates and those containing DMSO (APMA solvent) served as controls. In addition, proteins from nontreated homogenates were separated by electrophoresis and gels were then incubated overnight in substrate buffer containing 2 mM 1,10-phenanthroline, 1 mM phenylmethylsulfonyl fluoride (PMSF), or pepstatin. Representative zymograms demonstrating MMP-9 latent and active forms as well as MMP-2 latent and active forms are demonstrated. See Figure 1 for other definitions.

Age-related changes in the activity of MMPs in the knee joints and TMJ.

The data presented in Figure 3 clearly demonstrate that the expression patterns of MMP-2 and MMP-9 in the 3 cartilages for the range of ages examined were distinctly different from each other. The profiles of gelatinolytic activity in the tibial plateau and femoral condyle in neonatal animals, as well as during maturation and aging, were similar, although both were different from that in the TMJ cartilage. In the tibial plateau and femoral condyle, the level of MMP-9 was low in neonatal animals and increased with the animals' age. The level of MMP-9 in the TMJ cartilage was relatively low in comparison with that in the knee joint cartilage and, contrary to that in the knee joint, although the level of MMP-9 increased during early growth (from 1 day to 2 weeks of age), it gradually decreased during maturation and aging. In all 3 cartilages, latent MMP-9 was the main form observed. In the TMJ cartilage, the levels of MMP-2, both the latent and active forms, were relatively higher than that of MMP-9, and slightly decreased following the age of 2 weeks. In the tibial plateau and femoral cartilage, the level of MMP-2 remained high from 1 week to 1 month of age and decreased during maturation and aging.

Figure 3.

Age-related changes in the activity of MMP-2 and MMP-9 in articular cartilage. Homogenates (1 μg protein) prepared from articular cartilage tibial plateau, femoral condyles, and temporomandibular condyles obtained from 1-day-old (1d), 1- and 2-week-old (1w and 2w, respectively), and 1-, 2-, 4-, and 18-month-old (1m, 2m, 4m, and 18m, respectively) mice were evaluated by gelatin zymography. Commercially available human MMP-2, MMP-9, and molecular weight markers were electrophoresed in parallel. Representative zymograms demonstrating MMP-9 latent and active forms as well as MMP-2 latent and active forms for each cartilage type, and the mean densitometric assessment, by optical density (O.D.), of 3 separate experiments are demonstrated. See Figure 1 for other definitions.

Semiquantitative analysis of MMP mRNA expression.

A wide profile of MMP mRNA expression was observed in the tibial plateau cartilage during development, maturation, and aging of mice. Figure 4A demonstrates that the assay developed was sensitive to the quantitative changes in the mRNA expression. The values exhibited in Figure 4B are the relative values of the expression of a particular MMP in relation to the housekeeping gene, actin, in order to compensate for possible variations in sample preparation.

In the tibial plateau cartilage, MMP-3 mRNA expression was extremely low until the age of 1 month, after which it continued to increase to 18 months. Similarly, MMP-13 mRNA expression age-dependently increased, reaching its highest level at 1 month, and although a slight decrease was observed in MMP-13 mRNA expression at 18 months, it was still 3.6-fold higher than at day 1. MMP-2 mRNA expression was low at birth, peaked at 2 weeks of age, and decreased slightly thereafter. MMP-9 transcripts were more abundant than those of MMP-2, since one-fifth of the cDNA was required to produce a similar level of RT-PCR product. The level of MMP-9 mRNA expression was low in newborns and also peaked at 2 weeks of age, then decreased with the age of the animals. No MMP-1 expression was observed at all evaluated time points.

Immunohistochemical localization of MMP-2 and MMP-9 (gelatinases), MMP-13 (collagenase 3), and MMP-3 (stromelysin 1).

In order to further assess the changes in the MMPs studied during growth and aging, the protein expression of MMPs 2, 3, 9, and 13 was immunolocalized in paraffin sections of knee joint and TMJ condyle from mice of different ages (newborn, 1 day, 2 weeks, and 1, 2, 4, and 18 months old). Representative sections of day-old, 2-week-old, and 18-month-old cartilage are shown in Figures 5 and 6. Immunohistologic assessment demonstrated that in the TMJ condyle, MMP-13 was constitutively, highly expressed in the 3 age groups examined (Figures 5A–C). The MMP-13 protein was localized mainly in the articular surfaces and chondroblastic zone, in the hypertrophic cells, and at the resorption front of the joint. A different pattern was observed in the knee joint (Figures 5D–F), where MMP-13 immunoreactivity displayed an increase from newborn to 2 weeks, and decreased gradually during maturation and aging. The strong positive immunoreactivity in the knee joint of 2-week-old animals could be detected mainly in the articular surface and chondroblastic zone and less in the mature hypertrophic chondrocytes. In the 18-month-old animals, the overall number of cells was reduced and MMP-13 expression was low.

Figure 5.

Matrix metalloproteinase 3 (MMP-3) and MMP-13 expression in the knee joint and temporomandibular joint (TMJ) from mice of different ages. Immunohistochemistry was performed to determine MMP-13 (A–F) and MMP-3 (G–L) expression in the TMJ (A–C and G–I) and knee joint (D–F and J–L) from mice at the ages of 1 day, 2 weeks, and 18 months. Specific anti–MMP-13 and anti–MMP-3 binding was detected using biotin-labeled antibodies followed by peroxidase-conjugated streptavidin. The sections were counterstained with hematoxylin. Open arrow points to the hypertrophic zone, thick solid arrow to the articular surface, and thin solid arrow to the chondroblastic zone.

Figure 6.

MMP-2 and MMP-9 expression in the knee joint and TMJ from mice of different ages. Immunohistochemistry was performed to determine MMP-9 (A–F) and MMP-2 (G–L) expression in the TMJ (A–C and G–I) and knee joint (D–F and J–L) from mice at the ages of 1 day, 2 weeks, and 18 months. Specific anti–MMP-2 and anti–MMP-9 binding was detected using biotin-labeled antibodies followed by peroxidase-conjugated streptavidin. The sections were counterstained with hematoxylin. Open arrow points to the hypertrophic zone, thick solid arrow to the articular surface, and thin solid arrow to the chondroblastic zone. See Figure 5 for definitions.

Positive MMP-3 staining in the TMJ was observed in all ages evaluated (Figures 5G–I). MMP-3, unlike MMP-13, was confined only to the hypertrophic cells and the resorption front in newborns. In 2-week-old mice, positive immunoreactivity appeared along the articular surface in the chondroblastic zone and in the surrounding matrix. In 18-month-old mice, MMP-3 was observed in chondrocytes located in all regions of the cartilage. The expression pattern of MMP-3 in the knee joint was similar to that of MMP-13 (Figures 5J–L).

The general pattern of MMP-2 and MMP-9 immunoreactivity both in the TMJ and in the knee joint was similar to that observed for MMP-3 and MMP-13 (Figures 6A–L). Low immunoreactivity of MMP-2 and MMP-9 was observed in day-old animals both in the TMJ and in the knee joints, followed by a marked increase in 2-week-old mice. At this latter age, MMP-2, similar to MMP-3 and MMP-13, demonstrated relatively high immunoreactivity in both the TMJ and knee joint, whereas MMP-9 reactivity was lower in the TMJ compared with the knee joint. In sections from 18-month-old mice, immunoreactivity of MMP-2 and MMP-9 in the TMJ was localized along the articular surface. In the knee joint, MMP-2 and MMP-9 were also observed in the matrix.


This research studied and compared the involvement of several MMPs in the process of growth, maturation, and aging of articular cartilage from temporomandibular and knee joints of mice. In contrast to most previous research that examined isolated chondrocytes (21), synovial fluid (18, 22), or cultured explants (8, 16), the present study was performed on intact cartilage to better define the in vivo state. To the best of our knowledge, this study is the first to present information on mRNA levels of MMPs using RT-PCR methods, gelatinase activity using zymography, and localization of MMPs 2, 9, 3, and 13 by immunohistochemistry, at selected time points representing different stages of growth. The time points chosen represent the early stage of active growth (neonatal to 2 months old), maturation (4 months old), and aging (18 months old) in mice.

Distinct temporal patterns of gelatinase (MMP-2 and MMP-9) expression/activity in the 3 articular cartilage types examined during growth, maturation, and aging were observed by zymography. MMP-9 expression was barely detected in neonatal and early stages of growth in the tibial plateau and femoral condyle of the knee joint, while its level increased during later phases of development. In the mandibular condyle cartilage, although the overall level of MMP-9 was lower compared with that in the knee joint cartilage, its level was higher in younger animals compared with older ones. Most of the MMP-9 expression in both joints was found to be the latent form. Contrary to the pattern seen with MMP-9, both the active and the latent forms of MMP-2 were detected and their pattern of expression was similar in the 3 articular cartilages examined, increasing from birth to 1 month of age and decreasing thereafter. The immunohistochemistry assays supported the temporal changes observed in the gelatinase activity. Furthermore, the assays associated these MMPs with chondrocytes in different regions of the TMJ and knee joint cartilage, suggesting that these cells produce the MMPs at early stages of development and maturation. As observed in the zymography assay, the general level of MMP-9 expression in 2-week-old animals was lower than that of MMP-2 in the TMJ. In 18-month-old animals, the gelatinases in the TMJ were localized mainly in cells in the articular surface, while in the knee joints, these MMPs were found mainly in the extracellular matrix.

Messenger RNA analysis, performed on tibial plateau cartilage, demonstrated that this cartilage expressed MMP-2 and MMP-9 transcripts in all ages of mice examined. The temporal mRNA expression profiles of MMP-2 and MMP-9 were similar to each other, in contrast to the protein expression of these MMPs. Only trace expression of message was detected in newborns (1 day) and their levels were up-regulated during early growth, reaching a peak at 2 weeks of age and decreasing thereafter. The decrease in MMP-2 mRNA during maturation and aging paralleled the decrease in MMP-2 protein expression in this tissue. Unlike MMP-2, the expression of MMP-9 protein markedly increased during aging, whereas its mRNA level decreased, suggesting posttranscriptional modulation of MMP-9.

The present study, consistent with previously published data, demonstrated that although MMP-9 and MMP-2 share similar structural and substrate specificity, the regulation of these 2 enzymes is not coordinated and different factors are involved in their activation and activity (23–25). MMP-2 expression was more prominent at the early stages of growth, which is consistent with its housekeeping function of normal turnover of cartilage (7, 26), while the findings for MMP-9 in knee joint cartilage support the previously suggested role for this protein in the removal of denatured collagen fragments that increase with deterioration of the cartilage in aging joints (7). Age-related elevation of proinflammatory cytokines such as interleukin-1, reported to be found in the synovial fluid of aging human joints, may up-regulate the expression of MMP-9 in these animals (27). Previous studies (8) on human articular cartilage demonstrated MMP-9 mRNA and protein expression only in OA knee joint cartilage, while MMP-2 mRNA and protein were detected in both normal and OA cartilage, which is consistent with our present findings.

Differences in temporal as well as in intertissue mRNA and protein expression were also detected for MMP-3 and MMP-13. Similar to the expression of the gelatinases in TMJ cartilage, both MMP-3 and MMP-13 were prominent throughout postnatal development and aging, while in the knee joint, the protein level, as assessed by immunohistochemistry, was low in newborn and aged animals although high in 2-week-old animals. At this latter age, these MMPs appeared to be produced by chondrocytes in different regions of the TMJ and knee joint cartilage. At 2 weeks, MMP-3 protein expression was confined to immature chondrocytes and the matrix of the articular surface in both joints. In contrast, MMP-13 expression was observed in hypertrophic cells and at the resorption front of the mandibular condyle. It has been well established that MMP-13 is expressed in skeletal tissues, where it participates in both collagen and proteoglycan degradation in hypertrophic chondrocyte calcifying matrix (28–30). This MMP may be associated with resorption and bone formation via endochondral ossification at the cartilage–bone interface and with the appearance of OA lesions (21, 31, 32).

MMP-3 and MMP-13 mRNA expression was detected at all time points, as has been demonstrated by others (29, 33). The mRNA levels of MMP-13, from young and matured animals, reflected the protein levels in the knee joints. Both the mRNA and protein levels were low at birth and peaked at 2 weeks. However, while MMP-13 mRNA remained high and constant during later growth and aging, only a low level of immunostaining was observed in the knee joint. This may partly be due to its secretion into the synovial fluid or to posttranscriptional regulation of MMP-13. Differences were also observed in the protein and mRNA levels of MMP-3 in the tibial plateau. While MMP-3 mRNA increased with age in the tibial plateau, little protein was observed in aged knee joints. Future studies characterizing the focal mRNA expression of this MMP may elucidate the relationship between MMP-3 protein and mRNA levels. MMP-3 mRNA has been associated with OA (34, 35). Although no histologic findings of OA were observed in the specific knee joint sections of 18-month-old mice stained for MMP-3, the elevated MMP-3 mRNA may be indicative of OA, as observed in additional sections demonstrating MMP-2 and MMP-9 in mice of the same age (see Figure 6).

No evidence of MMP-1 mRNA expression was observed in the present study. This is surprising, in view of the fact that MMP-1 collagenolytic activity has been demonstrated in many species, including humans (33, 36). Recently, a murine MMP with similar structure and functional activity to that of human MMP-1, expressed in trophoblast cells during embryogenesis, was reported (37). Although we used primers specific for this new murine “MMP-1–like gene,” we failed to detect any mRNA transcription in mice cartilage at all stages of development examined.

Structural and histologic changes in cartilage from articular joints at different stages of growth and aging have been well documented. During intensive skeletal growth (in neonatal and young animals), joint cartilage is highly hypertrophic, actively involved in endochondral ossification. Following skeletal maturation (7–8 weeks), the articular cartilage gradually decreases, becomes less hypertrophied, and is engaged mainly in tissue maintenance. Aging leads to a further reduction in size and hypertrophy, and signs of degradation and fibrillation are observed. A fine balance between the continuous process of degradation and synthesis of extracellular matrix components maintains the integrity of normal cartilage. In adults, matrix turnover in articular cartilage is extremely slow, in comparison with that in the early stages of growth and maturation when the metabolic rate is substantially higher to allow for growth as well as homeostatic remodeling (18, 38).

The activity of MMPs is essential for the physiologic remodeling of cartilage and other tissue (3). Thus, the presently observed temporospatial differences in the expression of MMPs during early growth and aging are to be expected. Changes in MMP-13 accompanying rapid morphologic changes in the rat femoral neck after birth have been reported (39). In addition, in chicken sterna, MMP-13 expression was associated with hypertrophied cartilage. The latter research also implicated MMP-2 in the activation of MMP-13 during chondrocyte maturation, and the combined expression of both proteinases was found to be essential for subsequent calcification of the cartilage matrix.

The variation observed in the expression of MMPs during growth and aging of the knee joints and TMJ can be related to the differences in the function of these 2 joints. Weight-bearing forces are imposed on knee joints, in contrast to shear-bearing forces that are more prominent on TMJs. The TMJ is one of the most active joints in the body, and with the advancement of age, OA lesions were shown to spontaneously develop in this joint at a relatively younger age in comparison with that in the knee joints (27, 40). The relationship between mechanical load, the collagen network, and proteoglycan content of articular cartilage has long been established (41, 42). Studies conducted by Jin et al (43) on the response of cultured chondrocytes to shear stress indicated enhanced expression of MMP-9, while an earlier study by Smith et al (44) provided evidence that fluid shear forces affect expression of MMPs 2, 3, and 13 by chondrocyte monolayers. Using an in vitro cartilage-loading model, Blain et al (45) demonstrated that mechanical stimuli, such as cyclic loading, up-regulated both the synthesis and activation of pro–MMP-2 and pro–MMP-9, while having little influence on the expression of TIMPs. Taken together, this information combined with our results implies that mechanical load can affect the homeostasis of cartilage tissue through its effect on the expression of MMPs and TIMPs. Thus, the age-related differences in the expression of the MMPs presently examined could reflect the differential demands imposed on the cartilage during various stages of growth and aging.

In summary, the present study shows that the expression and activity of MMPs in the articular cartilage of the knee joint and TMJ reflect developmental processes of growth, maturation, and aging, and suggests that their activity may be affected by load and functional differences in these joints. This information provides important knowledge in the understanding of the role of MMPs in tissue maintenance and integrity. Further research in this field may contribute to the therapeutic use of MMP inhibitors to prevent and repair pathologic destruction of joint cartilage.