Enzymes from the matrix metalloproteinase (MMP) family play a crucial role in growth-plate vascularization and ossification via proteolytic cleavage and remodeling of the extracellular matrix. Their regulation in the growth plate is crucial for normal matrix assembly. Endochondral ossification, which takes place at the growth plates, is influenced by mechanical loading. Using an in vivo avian model for mechanical loading, we have found increased blood penetration into the growth plates of loaded chicks. The purpose of this work was to study the involvement of MMP-2, -3, -9, -13, and -16 in the growth plate's response to loading and in the catch-up growth resulting from load release. We found that mechanical loading, as well as release from load, upregulated MMP-2, -9, and -13 expressions. In contrast, MMP-3, associated with cartilage injuries, and its associated protein connective tissue growth factor (CTGF), were downregulated by the load. However, after release from load, MMP-3 was upregulated and CTGF levels were elevated and caught up with the control. MMP-3 and CTGF were also downregulated after 60 min of mechanical stretching in vitro. These results demonstrate the central role of MMPs in the growth plate's response to mechanical loading, as well as in the catch-up growth followed load release.
- connective tissue growth factor
- extracellular matrix
- matrix metalloproteinase
the long bones of the fetal skeleton develop through the process of endochondral ossification, in which cartilage serves as the initial skeletal element and is later replaced by bone (9). After birth, longitudinal growth occurs in the growth plates (57), which contain chondrocytes at different stages of differentiation, organized into several zones: resting, proliferative, prehypertrophic, and hypertrophic (16). During endochondral ossification, the cartilage, an avascular tissue, is gradually converted into a highly vascularized tissue, bone (42). This process is accompanied by changes in extracellular matrix (ECM) synthesis and remodeling mediated mainly by enzymes from the matrix metalloproteinase (MMP) family.
The MMPs are a family of zinc-dependent proteases (4) that include collagenases, stromelysins, and gelatinases (54). Their substrates are the ECM proteins (32, 37), and they are involved in proteolytic cleavage and remodeling of the cartilage (59) and regulation of angiogenesis (48). Twenty-four distinct MMPs have been cloned in humans, but only MMP-1, -2, -3, -9, -10, -13, and -14 have been shown to be involved in endochondral ossification (30, 62). Our laboratory has recently shown expression of the membrane-bound MMP-16 in the avian growth plate (50), but its significance to endochondral ossification is not yet clear.
Mechanical stimulation resulting from weight loading and muscle contraction plays an important role in bone remodeling (31, 36). MMPs are known to be affected by various strains; for instance in human vein cells, stationary strain significantly increased MMP-2, whereas cyclic strain decreased it. In those same cells, MMP-9 increased in response to stationary strain, but exhibited no response to cyclic strain (2). Static compression loading on rat caudal vertebrae elevated the expression of MMP-3, but not MMP-13 (6); similarly, in bovine synovial cells, MMP-3, but not MMP-1, increased after cyclic tensile strain (44). Our laboratory showed an increase in the expression of MMP-2, -9, and -13 in growth-plate chondrocytes of chicks after mechanical loading (45). However, despite accumulating data, the MMPs' role in the loading response remains unclear.
Among all of the MMPs, MMP-3 is particularly linked with pathological conditions. An increased level of synovial MMP-3 is usually found in acute knee injuries (55), together with increased IL-1β and TNF-α (23). Load-induced injury is associated with elevated MMP-3 levels in the articular cartilage (28), whereas physiological-range loads inhibit its expression in the flexor tendons (3). Moreover, load reduced MMP-3 synthesis in healthy, but not osteoarthritic, cartilage (34).
Connective tissue growth factor (CTGF/CCN2) is regulated by MMP-3 in two ways: 1) it cleaves CTGF protein into two ∼20-kDa fragments that are thought to have distinct functions (20); and 2) it acts intracellularly to enhance CTGF transcription by directly binding to the CTGF promoter (11). MMP-3 has been shown to upregulate the expression of CTGF in the human chondrocyte cell line HCS-2/8 (11). CTGF promotes proliferation, maturation, and hypertrophy of growth-plate chondrocytes during endochondral ossification (39). It acts as a central driver of cartilage and bone regeneration (26, 40). CTGF level is increased in HCS-2/8 cells after exposure to cyclic mechanical force, but it is reduced after exposure to cyclic tension force (41), indicating the importance of this protein in the response to different mechanical strains.
In this study, we examined changes in the expression of several MMPs in vivo in the growth plates of fast-growing young chicks following mild mechanical loading. We harnessed the loads on the chicks without surgical procedures or dramatic alterations in their environmental conditions (45, 46). MMP expression was evaluated by in situ hybridization and real-time PCR analyses. We also examined changes in CTGF expression in response to mild load. Little is known about the effect of loading on the growth plate of young, growing animals. The results of mechanical stress studies on bones in their rapid elongation phase could be relevant to intense sports activities during adolescence, the act of bearing heavy loads, and, possibly, childhood obesity.
MATERIALS AND METHODS
Radiolabeled UTP (35S 1,000 Ci/mmol) was purchased from the Radio-chemical Center (Amersham, UK). SYBR Green I was purchased from ABgene (Epsom, UK). pGEM T Easy kit was purchased from Promega (Madison, WI). ECL was purchased from Santa Cruz Biotechnology (Santa Cruz, CA).
One-day-old broiler chicks (Cobb) were obtained from a commercial hatchery (Brown Hatcheries, Hod Hasharon, Israel), raised in constant-temperature battery brooders at 34°C, and fed an age-appropriate diet ad libitum, according to National Research Council recommendations. Before each experiment, the chicks were given a 24-h period for adaptation. The experiment was approved by the committee of ethics in animal experimentation.
Loading and release model.
Two-day-old male chicks (n = 110) were divided into two randomized groups: “load” (55 chicks) and “control” (55 chicks); the load group was further divided into two separate groups, which were harnessed for 2 or 4 days with small bags (2.5 × 4 cm) filled with sand, weighing 10% of their body weight (BAG 2d and 4d, respectively), as previously described (45). Chicks in the control group were also divided by this criterion and were raised under the same conditions, but without artificial loads (CTRL 2d and 4d, respectively). At 4 days of age, chicks from the BAG 2d and CTRL 2d groups were killed (15 from each group). At 6 days of age, chicks from the BAG 4d and CTRL 4d groups were killed (20 from each group), while the other 20 chicks from the load group were released from loading and allowed to grow for a further 5 days free of load; these chicks were killed on day 11, together with the remaining 20 load-free control chicks (REL-LOAD 11d and CTRL 11d, respectively). At each time point, proximal tibia samples were processed for histology or RNA (46).
Histological staining and in situ hybridization of growth-plate sections.
Growth plates were fixed overnight in 4% paraformaldehyde (Sigma) in PBS at 4°C. The samples were dehydrated in graded ethanol solutions, cleared in chloroform, and embedded in Paraplast, and 5-μm sections were prepared. Slides were stained with Safranin-O to analyze proteoglycan distribution. The sections were placed in Mayer's hematoxylin for 2 min and rinsed in water. Next, the slides were placed in Fast Green FCF Yellowish for 10 min and rinsed in 1% acetic acid. Finally, the slides were placed in Safranin-O for 30 min, and rinsed in water and 95% ethanol. Hybridizations were performed as described by Reich et al. (45). The sections were deparaffinized in xylene, rehydrated through a graded series of ethanol solutions, rinsed in distilled water (5 min), and incubated in 2× saline sodium citrate at 55°C for 30 min. The sections were then rinsed in distilled water and treated with proteinase K (10 μg/ml in 0.2 M Tris · HCl, 5 mM EDTA, pH 7.5) for 10 min. After digestion, slides were rinsed with distilled water, fixed in 10% formaldehyde in PBS, blocked in 0.2% glycine, rinsed in distilled water, rapidly dehydrated through graded ethanol solutions, and air-dried for several hours. The sections were then hybridized with 35S-labeled anti-sense RNA probes (10 ng) for MMP-2, -3, -13, or -16 (see primers in Table 1). Radioactive signals were intensified using emulsion (Eastman Kodak, Rochester, NY): the sections were incubated with the emulsion solution for 1 mo, in the dark, at room temperature (45). No signal was observed in any of the hybridizations with sense probes, which were used as controls.
Measurement of growth plate width and blood vessel number.
The growth plate width was measured from the top of the proliferative zone to the bottom of the hypertrophic zone at three different points along the plate center, averaged for each plate, and then averaged with measurements from other plate samples. Blood vessels number at the chondroosseous junction (identified as the cross section at the bottom of the hypertrophic zone) were counted and averaged in four different slides from four chicks in each group.
RNA isolation and reverse transcription.
To prepare the RNA, proximal tibia growth plates (pooled from 10 chicks) or primary cell culture were isolated and prepared with RNeasy Maxi kit (Qiagen), according to the manufacturer's protocol. Total RNA (1 μg) was reverse transcribed in a final volume of 20 μl with the Reverse-RT kit (ABgene) using oligo-dT/hexamer primers, at reaction temperatures of 42°C for 1 h and 75°C for 10 min.
cDNA (1 μl) was used for real-time RT-PCR using the fluorescent dye SYBR Green I (Absolute QPCR SYBR Green Mix, ABgene) to monitor DNA synthesis, using specific primers for chicken MMP-3 and CTGF. Gene expression was normalized to the housekeeping gene Gallus gallus ribosomal 18S (see primer list in Table 2).
The PCR was carried out in an ABI Prism 7300 system (Applied Biosystems, Foster City, CA) using the following cycling protocol: 95°C denaturation for 15 min, followed by 40 cycles of 95°C denaturation (15 s), 60°C annealing (30 s), and 72°C extension (30 s). The amplified PCR product was analyzed with ABI Prism 7300 software (Applied Biosystems). At the end of the real-time PCR, a melting curve was determined to verify the presence of a single amplicon (46).
Chondrocyte cell culture.
Epiphyseal growth-plate chondrocytes were isolated from a pool of ten 1-day-old chicks by collagenase treatment, as described previously (35). The cells were cultured as a monolayer of proliferative cells, as monitored by the expression of collagen type II mRNA and the cells' polygonal morphology (35, 56, 61). Before each experiment, cells were detached with trypsin, seeded, grown for a few days in 10% FBS DMEM to confluence, and monitored for their proliferative state. Only early passages of polygonal chondrocytes were used for experiments.
Mechanical stimulation of chick primary growth-plate chondrocytes.
Mechanical stimulation achieved by longitudinal cell stretching was due to the irreversible deformation of the cell-attachment surface of the culture dish, as described previously (49). Specifically, the cell cultures were grown in DMEM supplemented with 10% FBS on 6-cm plastic tissue culture dishes for 5–6 days, to confluence. The mechanical longitudinal stretch was achieved by applying an orthodontic screw device attached to two pieces of solid acrylic resin. The whole device was fixed to the outer surface of the culture dish with epoxy resin. With the screw, it was possible to deform the dish by 0.1–0.05% of its length, thereby producing a unilateral force of ∼10 kg/cm2. Cells were subjected to the mechanical stretch for 0, 15, 30, and 60 min. RNA was extracted after 6 h of recovery.
Western blot analysis.
Whole growth plates were mashed, and total protein lysate was collected. Protein concentration was measured using a bicinchoninic acid protein assay reagent kit (Pierce Biotechnology). Lysates (30 μg protein) were separated by 10% SDS-PAGE, transferred to nitrocellulose membranes, incubated overnight with anti-CTGF antibody (SC-14939, Santa Cruz, CA) at 4°C, followed by incubation with peroxidase-conjugated donkey-anti-goat secondary antibody, and detected with ECL.
Statistical comparisons between the groups were analyzed by t-test. The statistical analyses were carried out with JMP software (SAS Institute, Cary, NC). Differences between the groups of P < 0.05 were considered statistically significant.
Mechanical load does not cause injury in the growth plate.
Our laboratory, and others, have previously reported that mechanical loading reduces the width of the growth plate (6, 45). To examine whether our loading model generates an injury response in the growth plate, characterized, in part, by cartilage proteoglycans depletion, we used Safranin-O staining. Reduced and diffused red color in this staining is known to be used as a marker for inflammation and injury of cartilage (8, 47). Safranin-O staining revealed a uniform presence of proteoglycans throughout the articular cartilage (data not shown) and the growth plates in the loaded group, which was not different from the control (Fig. 1). These results suggest that, in our model, the mechanical load applied to the chicks did not cause injury in the growth plate. A reduction in growth-plate width, together with an increased number of vessels penetrating the growth plate from the metaphysic, was observed, as previously reported (15.9 ± 2.8 and 22.2 ± 2.3 in the control and load group, respectively) (45).
Mechanical load upregulates the expression of MMP-2 and MMP-13 with no effect on MMP-16.
Different types of strains influence the expression of MMPs in different tissues and cells (2, 6, 44, 45). We examined the effect of mechanical loading on MMP-2, MMP-13, and MMP-16 expression patterns and localization in the chick's proximal tibia by in situ hybridization analysis. MMP-2 expression was detected in the compact bone, the perichondrium and the proliferative zone, as well as in cells surrounding the blood vessels of the epiphysis and the bone (Fig. 2, A and B).Four days of mechanical load increased MMP-2 expression around the blood vessels, and this was particularly evident in the proliferative zone (Fig. 2, A and B).
In an earlier study, we showed upregulation of MMP-13 expression in the growth plate in response to 4 days of loading. In this study, MMP-13 expression was localized to cells surrounding the blood vessels of the bone, as well as the epiphysis, proliferative, and hypertrophic zones of the growth plate and in hypertrophic chondrocytes. MMP-13 expression was upregulated after 4 days, mainly in the chondroosseous junction (Fig. 2G).
MMP-16 was detected in the epiphysis and in cells surrounding its blood vessels, in the reserve zone of the growth plate, in cells surrounding the blood vessels penetrating the hypertrophic zone, but not in hypertrophic chondrocytes, and in the compact bone. Mechanical load did not affect MMP-16 expression (Fig. 2, C and D). Overlay of MMP-2 and MMP-16 shows colocalization of those two genes in cells surrounding the blood vessels penetrating the hypertrophic zone and in the compact bone, but not in the proliferative zone (Fig. 2, E and F).
Mechanical load downregulates the expression of MMP-3 and CTGF in vivo and in vitro.
MMP-3 expression and localization were studied by in situ hybridization. Its expression was detected in hypertrophic chondrocytes (Fig. 3, A–C).No difference was found in the expression of MMP-3 after 2 days of mechanical loading, but, after 4 days of load, its expression was reduced (Fig. 3, B and C). Quantitative PCR also showed a reduction in MMP-3 expression level after 4 days of loading (Fig. 3E).
CTGF expression and activity are regulated by MMP-3 (11, 20). This protein is known to be involved in endochondral ossification (39) and plays a critical role during bone regeneration (26). We found that, like MMP-3, CTGF expression is also downregulated after 4 days of mechanical load (Fig. 3F). Its protein level was significantly reduced after 2 days of load, as well as after 4 days (Fig. 3G).
Next, we examined the effect of mechanical stretch on MMP expression in primary chondrocytes isolated from 1-day-old chicks. The load applied in vivo is the end result of combined types of loading, including compression, tension, stretching, etc. The in vitro stretching protocol mimics the stretching load applied on the bone. The cells were stretched for different periods of time (0, 15, 30, 60 min), after which they were given 6 h of recovery to allow mRNA synthesis. The expressions of MMP-2, -3, -9, and -13 were quantified by real-time PCR. The mechanical stretching had no effect on the levels of MMP-2, -9, and -13 (data not shown). However, the expression of MMP-3 was significantly reduced by 60 min of stretching (Fig. 4A Mechanical stretching also downregulated CTGF expression after 60 min, in a manner similar to MMP-3 (Fig. 4B).
Release from load increases blood penetration into the growth plate and upregulates the expression of MMP-2, -9, -13, and -3 and CTGF.
After 4 days of loading, a group of chicks were released from the load and were raised for another 5 days (to 11 days of age). Our laboratory has previously reported catch-up growth in these chicks (46), and this growth resulted in thicker growth plates than those of the nonloaded chicks that served as controls (Fig. 5A).The catch-up growth was also associated with increased blood penetration into the growth plate: 23.66 ± 1.5 and 27.33 ± 1.1 vessels per section in the control and released groups, respectively (P < 0.05) (Fig. 5B). However, no signs of injury and inflammation were evident, as no cartilage destruction or reduced color was noticed in Safranin-O staining (Fig. 5A). Together with the increase in blood penetration, a quantitative increase in the expression of MMP-2, -9, and -13 was observed in the released chicks (Fig. 5C). MMP-3 expression, which, in contrast to the other MMPs checked, was downregulated by the loading, inverted its expression pattern, resulting in upregulation after release from load (Fig. 5D). CTGF followed the same pattern as MMP-3, i.e., downregulation in the loaded growth plates, and, in the released growth plates, it was elevated and caught up with the control in both mRNA (Fig. 5D) and protein (Fig. 5E) levels.
In this work, we studied the events leading to the increased blood penetration observed in the growth plates of weight-loaded chicks. We used a model for generating mechanical load on chicks, which involved harnessing 2-day-old chicks with small bags filled with sand, weighing 10% of their body weight. This magnitude of mechanical load results in decreased growth-plate width after 4 days, together with increased blood penetration (45). To understand the mechanism underlying this response, we examined the expression of different members of the MMP family known to be involved in endochondral ossification: MMP-2, -3, -9, and -13 (30, 62), as well as MMP-16, a membrane-bound MMP, which is expressed in human (19) and avian growth plates (50). Since the MMPs are matrix-degrading enzymes (32, 37, 59), which are involved in the regulation of angiogenesis (48), they are natural candidates for involvement in the increased blood penetration induced by loading. Two days of mechanical loading did not change the expression of any of the observed MMPs. However, after 4 days, changes were observed in both their expression and localization.
The expression of MMP-2, a member of the gelatinase subfamily, increased after 4 days of mechanical load, and in situ hybridization analysis showed the transcript increase to be mainly around the blood vessels and in the proliferative zone. MMP-2 is known to be expressed by osteoclasts (19) and endothelial cells (27); thus its expression in cells surrounding blood vessels associated with endothelial cells and osteoclasts suggests that the upregulation of this MMP plays a role in the increased vascularization of the growth plate after loading.
The second gelatinase, MMP-9, is a key regulator of growth-plate angiogenesis and apoptosis of hypertrophic chondrocytes (60). Its action results in the release of angiogenic vascular endothelial growth factor and other growth factors from the matrix (12, 51) and the degradation of type II collagen. Our laboratory has previously shown that 4 days of mechanical load upregulates this transcript, which becomes intensified in the chondroosseous junction (45). We speculate that the increase in MMP-9 is one of the primary regulators of the increased vascularization, as knocking out MMP-9 in mice resulted in delayed vascularization and lengthening of the growth plate (60), while, in our model, the growth plate became narrower as a result of the load.
MMP-13 works synergistically with MMP-9 to degrade the ECM proteins (12, 52). MMP-13 knockout mice also showed delayed vascularization, lengthening of the hypertrophic zone, and irregular arrangement of the chondrocyte columns (21). In our study, the mechanical load upregulated MMP-13 expression, mainly in the chondroosseous junction. Moreover, we found that the mechanical load leads to highly arranged chondrocyte columns (46), which can also be explained by the increased MMP-13 expression. Taken together, upregulation of MMP-2, -9, and -13 may mediate the load-induced narrowing of the growth plate and play a significant role in the increased vascularization. The MMPs are known to activate one another; hence their colocalization is biologically important. For example, MMP-2 is a pro-MMP-13 activator (7): it is expressed in the compact bone and may activate pro-MMP-13 from the hypertrophic chondrocytes adjacent to it. Likewise, MMP-16 is a pro-MMP-2 activator (25). We show that these two genes are colocalized in cells surrounding the blood vessels penetrating the hypertrophic zone and in the compact bone, suggesting a role for MMP-16 in pro-MMP-2 activation in those sites. Interestingly, MMP-16 is not expressed in the proliferative zone, whereas MMP-2 is highly expressed in this zone, and its expression is upregulated by the load, whereas MMP-16 is not affected by the load. These results suggest that MMP-2, in the proliferative zone, is involved in the growth plate response to load in a non-MMP-16 dependent pathway. Probably the activation of MMP-2 in the proliferative zone is mediated through other mechanisms (19, 43).
Increased vascularization can result from both normal physiological and pathological processes. Normal physiological loading is vital for many processes, such as bone formation and remodeling, angiogenesis, stimulation of chondrocyte metabolism, and cartilage adaptation to the demands imposed by the body (31, 36). However, excess loading can result in injury. Most studies demonstrate the effect of load in the context of articular cartilage, in which abnormal or damaging loading results in tissue breakdown or upregulation of proinflammatory cytokines, which are known risk factors for the subsequent development of secondary osteoarthritis (15, 18, 24). In our loading model, no proteoglycan depletion was observed by Safranin-O staining, in either the area of the growth plate or in the articular surface, indicating that the loading did not cause any cartilage damage or breakdown. It is important to note that, in large-profile gene expression studies that we performed with RNA isolated from growth plates of loaded vs. nonloaded chicks, no differences in proinflammatory cytokines were observed (unpublished data). Furthermore, MMP-3, which is known to be upregulated by IL-1β in pathological conditions of the joint (22, 23, 55) and to be elevated in load-induced injuries (28), was downregulated in our model. Taken together, these results suggest that our loading model does not impose any injury to the growth plate, and that the observed increased vascularization is a physiological response.
MMP-3 is the most potent proteoglycanase among all MMPs; it has broad substrate specificity, including proteoglycans, casein, fibronectin, and collagens type III, IV, and V (1, 58). We show that, unlike MMP-2 and -9, MMP-3 is highly expressed in the hypertrophic chondrocytes. This suggests a role for MMP-3 in the maintenance of the cartilage itself and not in the vascularization process. Moreover, it may offer an explanation for the fact that MMP-2, -9, and -13 are upregulated by the load, while MMP-3 is downregulated by it. One possibility is that MMP-3 mediates its action on chondrocytes through CTGF.
CTGF is expressed in the prehypertrophic zone of the growth plate (26). It regulates chondrocyte proliferation and differentiation by forming a complex with BMP-2 (29). In postfracture regenerating cartilage, its expression is enhanced in the hypertrophic and proliferative chondrocytes (38), and it is regulated by MMP-3 (11, 20). CTGF has been shown to be involved in the response to mechanical load in chondrocytes (41), and it was downregulated in our model. It might be that MMP-3 is working through CTGF to mediate chondrocyte proliferation and differentiation and might be connected to the narrowing of the growth plate induced by the load.
The term catch-up growth refers to the phenomenon in which longitudinal growth velocity is transiently accelerated above the statistical limits of normality for age, after the removal of a growth-inhibiting condition (33). It has been observed in humans and animals, and its underlying pathogenic mechanism is not clear (5, 14). We were the first to show that release from load is followed by a catch-up growth process that results in thickened growth plates relative to those of a control group (46). Interestingly, both load and release from load increased growth plate vascularization, a rate-limiting step in endochondral ossification. However, the mechanism underlying the two processes is different; while in both processes MMP-2, -9, and -13 were upregulated, the load caused a reduction in MMP-3, whereas the release from load induced upregulation of this gene. The increase in MMP-3 may reflect a stress condition in the growth plate as a result of the accelerated growth. Even-Zohar et al. (13) showed an increase in hypoxia-inducible factor 1α (HIF-1α) during nutrition-induced catch-up growth in the proximal tibial epiphyseal growth plates (13). HIF-1 signaling responds to decreased oxygen tension; however, HIF-1 signaling can also be regulated by other mechanisms, including hormones, growth factors (17), and MMP activation (10). Components of its signaling pathway are involved in wound healing (53). Based on this information, we propose that the process of catch-up growth may resemble that of wound healing.
This research was supported by The Israel Science Foundation (Grant no. 292/07).
None of the authors have a conflict of interest.
We thank Svetlana Penn for technical help.
- Copyright © 2010 the American Physiological Society