Michael J. F. Blumer, Department of Anatomy, Histology and Embryology, Division of Clinical and Functional Anatomy, Innsbruck Medical University, Müllerstrasse 59, A-6020 Innsbruck, Austria. T: + 43 512 900371120; F: + 43 512 900373112; E: firstname.lastname@example.org
During long bone development the original cartilaginous model in mammals is replaced by bone, but at the long bone endings the avascular articular cartilage remains. Before the articular cartilage attains structural maturity it undergoes reorganization, and molecules such as vascular endothelial growth factor (VEGF) and endostatin could be involved in this process. VEGF attracts blood vessels, whereas endostatin blocks their formation. The present study therefore focused on the spatio-temporal localization of these two molecules during the development of the articular cartilage. Furthermore, we investigated the distribution of the chondro/osteoclasts at the chondro–osseous junction of the articular cartilage with the subchondral bone. Mice served as our animal model, and we examined several postnatal stages of the femur starting with week (W) 4. Our results indicated that during the formation of the articular cartilage, VEGF and endostatin had an overlapping localization. The former molecule was, however, down-regulated, whereas the latter was uniformly intensely localized until W12. At the chondro-osseous junction, the number of tartrate-resistant acid phosphatase (TRAP)-positive chondro/osteoclasts declined with increasing age. Until W3 the articular cartilage was not well organized but at W8 it appeared structurally mature. At that time only a few TRAP cells were present, indicative of a low resorptive activity at the chondro–osseous junction. Subsequently, we examined the metaphyseal growth plate that is closed when skeletal maturity is attained. Within its hypertrophic zone, localization of endostatin and VEGF was observed until W6 and W8, respectively. At the chondro–osseous junction of the growth plate, chondro/osteoclasts remained numerous until W12 to allow for its complete resorption. According to former findings, VEGF is critical for a normal skeleton development, whereas endostatin has almost no effect on this process. In conclusion, our findings suggest that both VEGF and endostatin play a role in the structural reorganization of the articular cartilage and endostatin may be involved in the maintenance of its avascularity. In the growth plate, however, endostatin does not appear to counteract VEGF, allowing vascular invasion of hypertrophic cartilage and bone growth.
In mammals, the long bones arise from cartilaginous anlagen in a process called endochondral bone formation. This process is accompanied by disintegration of the cartilage matrix, which in turn will allow new blood vessels to invade the cartilage model. Initially, distinct matrix metalloproteinases (MMPs) dissolve type II collagen and aggrecan, then tartrate-resistant acid phosphatase (TRAP) attacks the mineralized components of the cartilage matrix (Vu et al. 1998; Holmbeck et al. 1999; Zhou et al. 2000; Davoli et al. 2001; Lee et al. 2001, 2009; Stickens et al. 2004; Álvarez et al. 2005; for review see Holmbeck & Szabova, 2006; Krane & Inada, 2008; Blumer et al. 2008a). The TRAP enzyme is released by polynucleated chondroclasts and osteoclasts, and activated by cathepsin K (Hayman et al. 1996, 2000; Ljusberg et al. 1999; Hollberg et al. 2002). Angiogenesis, the formation of new blood vessels from preexisting capillaries, is triggered by the vascular endothelial growth factor (VEGF), and with the vessels, bone-forming cells and the chondro- as well as osteoclasts are recruited. In doing so, the cartilage matrix is remodelled, and newly formed bone leads to the establishment of the diaphyseal primary ossification centre (POC) followed by the epiphyseal secondary ossification centre (SOC) (Gerber et al. 1999; Harper & Klagsbrun, 1999; Carlevaro et al. 2000; Maes et al. 2002, 2004; Blumer et al. 2007, 2008b; Grellier et al. 2009). In mice, epiphyseal vascularization and formation of the SOC start after birth, and VEGF is detected until day 20 in the hypertrophic chondrocytes located immediately beneath the articular cartilage (Blumer et al. 2007; Allerstorfer et al. 2010). Principally, the adult articular cartilage shows no evidence of hypertrophy (Binette et al. 1998). It is a stable avascular tissue, and VEGF is down-regulated during its development and nearly undetectable under normal conditions. Only in patients suffering from osteoarthritis is the angiogenic factor up-regulated in the joint system, and during the progression of this degenerative disease, vessels from the subchondral bone occasionally grow into the calcified zone of the articular cartilage (Pufe et al. 2001, 2004). For the maturation and maintenance of a healthy articular cartilage anti-angiogenic molecules are needed, and endostatin has been shown to be one of the factors that keep the joint system free of blood vessels (for review see Pufe et al. 2005). Endostatin is a proteolytic fragment of type XVIII collagen. It is able to block directly the interaction between VEGF and its receptor (Flk-1/KDR = VEGFR2), provoking a negative impact on the VEGF-mediated signalling events. This in turn inhibits endothelial cell proliferation and migration, and consequently angiogenesis under normal as well as pathological conditions (O’Reilly et al. 1997; Kim et al. 2002; Pufe et al. 2004; Sipola et al. 2006; for review see Zheng, 2009). In summary, as the articular cartilage develops, expression of VEGF obviously diminishes whereas synthesis of endostatin remains. However, data detailing the spatiotemporal localization pattern of the two counteracting molecules are lacking currently.
The formation of the metaphyseal vascular network associated with the development of the POC begins before birth, and VEGF, being expressed throughout postnatal development, has an important role in the maturation of the long bones (Carlevaro et al. 2000; Bluteau et al. 2007; Evans & Oberbauer, 2007). The vascularization of the metaphyseal growth plate is a finely balanced process (Pufe et al. 2004; Holmbeck & Szabova, 2006), thus we hypothesized that inhibitory factors, such as endostatin, might be involved in this event. We therefore examined whether endostatin is found in the metaphyseal growth plate.
The present study focused on the spatio-temporal localization of VEGF and endostatin during the postnatal development of the articular cartilage and the metaphyseal growth plate in the mouse femur. In addition, we analysed the distribution pattern of the chondro/osteoclasts at the interface of the articular cartilage with the subchondral bone as well as at the chondro–osseous junction of the metaphyseal growth plate with the POC.
Materials and methods
The mice (129SvEv) were obtained from the central laboratory animal facilities of the Innsbruck Medical University. Postnatal stages of week (W) 2, 3, 4, 6, 8 and 12 were investigated, with three animals per age group. The sex of the mice was not recorded. The mice were anaesthetized with CO2, and killed by cervical displacement. Subsequently, the legs were amputated, the soft tissue was carefully removed and the distal part of the right and left femur examined.
Tissue preparation for light (LM) and transmission electron microscopy (TEM)
The bones of the mice aged 3 and 4 weeks were fixed in 2.5% glutaraldehyde, 2% paraformaldehyde (PFA) buffered in sodium cacodylate (0.1 m), pH 7.4 for 4 h at room temperature and rinsed in the same buffer. They were postfixed in 0.5% osmium tetroxide, 1% potassium hexacyanoferrat III in distilled water overnight at 4 °C, rinsed and decalcified in 3% ascorbic acid in sodium chloride (0.15 m) for 2 days at 4 °C. This procedure was followed by dehydration in graded ethanol series and embedding in Spurr’s epoxy resin. Ribbons of consecutive semithin sections (2 μm) were cut on a Reichert Ultracut S microtome (Leica Microsystem, Wetzlar, Germany) with a histo-jumbo-diamond knife (Diatome, Biel, Switzerland) (Blumer et al. 2002) and stained with toluidine blue for 20 s at 60 °C. Complete series of ultrathin sections (80 nm) were cut on a Reichert Ultracut S microtome with an ultra-diamond knife, mounted on dioxan-formvar coated copper slot-grids and stained with an aqueous solution of uranyl acetate (1%) for 30 min at 20 °C followed by lead citrate for 3 min at 20 °C. The ultrathin sections were examined with a Philips CM 120 transmission electron microscope (FEI, Eindhoven, the Netherlands) equipped with a MORADA digital camera (Olympus SIS, Münster, Germany), and the software Olympus TEM Imaging Platform was used.
Tissue preparation for enzyme histochemistry and immunohistochemistry
The bones were fixed with 4% PFA in phosphate-buffered saline (PBS, 0.1 m) for 4 h, rinsed in the same buffer and decalcified as described before. The bones of mice aged 6, 8, and 12 weeks were, however, decalcified for up to 5 days. Subsequently, the samples were dehydrated in graded isopropanol and xylene series and embedded in paraffin. Serial sections (6–7 μm) were made on an HM 355S microtome (Microm, Walldorf, Germany) and collected on SuperFrost®Plus slides. For the histological purpose only, the sections were stained with haematoxylin/eosin (HE).
Enzyme histochemistry (HC)
To show TRAP activity, sections were deparaffinized, rinsed in PBS and incubated with a solution containing 50 mm sodium acetate (pH 5.2), 0.15% naphthol-AS-TR-phosphate, 50 mm sodium tartrate, and 0.1% Fast Red T.R. (Sigma Aldrich Chemie GmbH, Taufkirchen, Germany) for up to 1 h at room temperature. Subsequently, the sections were rinsed in PBS and counterstained with toluidine blue for 1 min at room temperature.
For cathepsin K, a rabbit anti-mouse (10 μg mL−1 in antibody diluents) (catalogue no. ab19027; Abcam, Cambridge UK), for type II collagen, a rabbit anti-human (1 : 250 in antibody diluents) (catalogue no. CL50211AP; Cedarlane, ON, Canada), and for vascular endothelial growth factor (VEGF), a rabbit anti-mouse antibody (1 : 200, 0.1% Triton X100 in antibody diluents) (catalogue no. sc-152; Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA) were used.
Histological sections were deparaffinized and rinsed in PBS. For antigen retrieval either citrate buffer (10 mm, pH = 6.0) and a microwave oven (400 W, 7 min for cathepsin K) or protease digestion (2 min for type II collagen and 4 min for VEGF) was used. Subsequently, endogenous peroxidase activity was blocked with 1% H2O2 in 30% methanol for 30 min in the dark. The following protocol comprised primary antibody incubation overnight at 4 °C, saturation of unspecific sites with 10% normal goat serum (NGS) for 20 min, and a 4-h incubation with a secondary antibody [goat anti-rabbit IgG/HRP-conjugated (1 : 250 for cathepsin K, and 1 : 1000 for VEGF and type II collagen) in antibody diluents] (catalogue no. P0448; DakoCytomation, Glostrup, Denmark) at room temperature. Afterwards the antigen-antibody complex was made visible by 0.05% 3,3-diaminobenzidine (DAB), 0.01% H2O2 in distilled water (10–15 min in the dark).
For endostatin, a goat anti-human biotinylated antibody (10 μg mL−1, 0.1% Triton X100 in antibody diluents) (catalogue no. BAF1098; R&D Systems, Minneapolis, MN, USA) was used. The sections were deparaffinized and rinsed, and protease digestion (4 min) was performed. After that, endogenous peroxidase was quenched with 3% hydrogen peroxidase, and subsequently the sections were incubated with the primary antibody overnight at 4 °C. Antibody site detection was performed using a goat kit (catalogue no. CTS008; R&D Systems), according to the manufacturer’s instructions. This procedure comprised the application of an anti-goat biotinylated secondary antibody (45 min) and a streptavidin-horseradish peroxidase (HRP) conjugated reagent for 30 min at room temperature. Visualization of the antigen–antibody complex was achieved by application of a DAB chromogen solution (5–10 min in the dark).
Negative controls were obtained by substituting the primary antibodies with 0.1% Triton X100 in antibody diluents (for VEGF and endostatin) or antibody diluents only (for type II collagen and cathepsin K). No negative control yielded any labelling.
Double labelling (TRAP and VEGF)
For double labelling, TRAP reaction was performed first, followed by IHC. The sections were counterstained with Gils’ haematoxylin, but some sections remained unstained. All sections were examined with a Zeiss Axioplan 2 (Zeiss, Oberkochen, Germany) and photographed as colour images using a Zeiss AxioCam HR and axiovision 4.1. software running on a Pentium 4 (Intel Inc. Santa Cruz, CA, USA) with WindowsXP (Microsoft Inc., Redmond, WA, USA).
Light microscopy, transmission electron microscopy and immunohistochemistry (type II collagen)
Histology of the articular cartilage
In mice at W2 a secondary ossification centre (SOC) was located in the distal epiphysis of the femur, and it expanded in radial direction during advancing development. In parallel, the height of the articular cartilage decreased (Fig. 1). At W2 and W3, the superficial (tangential) zone of the articular cartilage comprised several layers of flat elongated chondrocytes. They had a parallel alignment to the articular surface. This zone was followed by layers of roundish chondrocytes and an area of mineralized hypertrophic cartilage, invaded by numerous blood vessels that originated from the SOC (Figs 1A2 and 2A1,A2). At W4 the superficial zone was composed of two to three layers of flat elongated chondrocytes only. Oval-shaped cells, occasionally arranged in columns, followed this zone, but hypertrophic chondrocytes were not detected anymore (Fig. 2B1,B2). This structural organization of the articular cartilage did not significantly alter with increasing age (Fig. 1B2). A detailed description of the development of the mouse articular cartilage in long bones has been published by Hughes et al. (2005), who used both chemical and cryofixation techniques.
The growth plate was situated between the primary ossification centre (POC) and the SOC, the height of which markedly decreased during proceeding development (Fig. 1A1,B1).
Localization of TRAP-positive chondro/osteoclasts
Chondroclasts and osteoclasts are polynucleated cells having the same morphological features. The former are attached to the cartilage matrix, whereas the latter are in contact with the bone matrix. TRAP is a well-established histochemical marker to detect both cell types (Hayman et al. 2000). As we could not ascertain unambiguously whether the TRAP-positive cells were attached to the cartilage or the bone matrix we referred to them as chondro/osteoclasts in the present investigation. At W2 numerous chondro/osteoclasts were attached to the mineralized hypertrophic cartilage (Fig. 3A1). This distribution pattern was also observed at W4. At W8 the number of chondro/osteoclasts had considerably decreased, and at W12 only a few TRAP-positive cells were found below the articular cartilage (Fig. 3B1,C1). At the chondro–osseous junction of the metaphyseal growth plate with the POC, however, the number of the chondro/osteoclasts appeared not to alter throughout development. They remained numerous in this site, and as an example we showed their distribution at W12 (Fig. 3D1).
Localization of cathepsin K-positive chondro/osteoclasts
We intended to verify our data obtained from HC by performing IHC for cathepsin K, an additional marker for the chondro/osteoclasts (Ljusberg et al. 1999; Hollberg et al. 2002). Our findings were in line with the data obtained from HC and indicated that the cathepsin K-positive chondro/osteoclasts markedly decreased underneath the articular cartilage (Fig. 3A2–C2). A considerable number of chondro/osteoclasts, however, remained at the chondro–osseous junction of the metaphyseal growth plate with the POC, and as an example we demonstrated their distribution at W8 (Fig. 3D2).
Immunohistochemistry and histochemistry (double labelling)
Localization of VEGF- and TRAP-positive chondro/osteoclasts
To assess differences in the staining intensity, numerous sections were not counterstained with haematoxylin. The examination of numerous sections through the distal femoral epiphysis demonstrated that VEGF was present in the articular cartilage of all developmental stages. At W2, immunoreactivity for the growth factor was noted in the cells of the superficial zone, the roundish chondrocytes below this zone, and furthermore in the hypertrophic chondrocytes (Fig. 4A1,A2). At W4–12, VEGF was visible in the superficial layer and in numerous oval-shaped chondrocytes immediately below this layer but it was not observed in the deeper zones of the articular cartilage close to the subchondral bone (Fig. 4B1–D2). However, at W4, several chondrocytes near the vasculature of the SOC displayed staining for the growth factor (Fig. 4B2). Furthermore, our results showed that at W12 the staining intensity for the angiogenic factor had markedly decreased (Fig. 4D1,D2). TRAP-positive chondro/osteoclasts were encountered in all developmental stages (Fig. 4A2,B2,D2).
We also examined the metaphyseal growth plate where VEGF has been shown to be expressed postnatally (Evans & Oberbauer, 2007). Our results revealed distinct labelling for VEGF in the hypertrophic chondrocytes in mice aged 2 and 4 weeks, respectively. In addition, the proliferating chondrocytes displayed a faint staining that, however, was exclusively noted at W2 (Fig. 5A,B). At W6 the staining intensity decreased within the hypertrophic chondrocytes, and at W8 only a small portion of these cells was slightly VEGF-positive (Fig. 5C,D). At W12 no labelling for VEGF was noted (data not shown).
The skeletal muscle served as a positive control because it was shown that muscle fibres expressed VEGF during normal development (Gu et al. 2006; Allerstorfer et al. 2010). Our results demonstrated a distinct localization of VEGF throughout development but the staining was more intense immediately near the perichondral bone of the diaphysis (Fig. 5E,F).
Localization of endostatin
To assess differences in the staining intensity, some sections were not counterstained with haematoxylin. For localization of endostatin we chose the subsequent sections from this very region where VEGF labelling could be observed. Strong immunoreactivity for endostatin was evident within the articular cartilage throughout development. At W2 endostatin was encountered in the cells of the superficial zone as well as the roundish chondrocytes underneath this zone, and furthermore in the hypertrophic chondrocytes (Fig. 6A1,A2). At W4–12, endostatin was seen in the superficial layer and in many oval-shaped chondrocytes immediately below this layer but chondrocytes of the deeper zones did not display any labelling (Fig. 6B1–D2). At W4, however, several chondrocytes near the vasculature of the SOC revealed staining for endostatin (Fig. 6B2). Contrary to our findings obtained from VEGF immunolabelling, the staining intensity for endostatin did not alter during proceeding development but was rather strong at all stages (Fig. 6A1–D2).
We furthermore examined the metaphyseal growth plate. By W2, a signal for endostatin was observed in the hypertrophic chondrocytes as well as in several proliferating chondrocytes (Fig. 7A). At W4 and 6, respectively, the protein could likewise be detected in the hypertrophic chondrocytes but not in the proliferating chondrocytes (Fig. 7B,C). The staining intensity was uniformly strong between W2 and 6 (Fig. 7A–C). At W8, however, labelling for endostatin had practically vanished (Fig. 7D).
In mammals, the bony ends of the synovial joints are lined with the articular cartilage that undergoes structural reorganization during the postnatal growth phase (Hunziker et al. 2007). In doing so, the overall height of the articular cartilage decreases, and this could be also observed in mice. Our present findings revealed that until W3, layers of hypertrophic chondrocytes were present, and the morphology of the articular cartilage resembled that of the femur of a 1 month-old rabbit. In both animal models, the articular cartilage is not well structured at this point of time, and numerous blood vessels as well as chondro/osteoclasts invade the mineralized hypertrophic layers indicative of an intense tissue resorption and epiphyseal bone formation. According to Hunziker et al. (2007) the articular cartilage represents a surface growth plate with a high level of activity at this developmental time. In rabbits, the activity of the vascular invasion front almost ceases after 3 months when the articular cartilage has achieved its mature structural organization. In mice, a high resorptive activity was noted until W4, and it subsequently decreased as evidenced by the low number of TRAP cells in W8 mice. We therefore assume that at that time the articular cartilage has attained its structural maturity. During the postnatal period all zones of the immature articular cartilage are subjected to complete resorption except the most superficial one that functions as a pool for the stem cells. After activation of this reservoir, new tissue is formed leading to the establishment of the mature articular cartilage (Hunziker et al. 2007). In the present investigation, VEGF was noted throughout the postnatal period but its protein localization declined with increasing age (Fig. 4). Likewise, in the adult articular cartilage of humans, the growth factor is largely down-regulated (Pufe et al. 2001). VEGF attracts blood vessels essential for coupling cartilage resorption and bone formation (Gerber et al. 1999; Harper & Klagsbrun, 1999; Kishimoto et al. 2006). Nevertheless, angiogenesis is a fine-tuned process controlled by both stimulating and inhibiting molecules to enable a proper maturation of an avascular tissue (Fukai et al. 2002; Pufe et al. 2004). Endostatin blocks angiogenesis and is expressed in the adult articular cartilage of humans and rats (O’Reilly et al. 1997; Kim et al. 2002; Pufe et al. 2004; Sipola et al. 2006; for review see Zheng, 2009). Our data are consistent with these findings; we could localize the anti-angiogenic factor in mice, and its expression was uniform and high until W12 (Fig. 6). Coupled with previous findings, our observations suggest that VEGF as well as endostatin are initially essential for a controlled remodelling of the articular cartilage. After attainment of its structural maturity, VEGF is down-regulated but endostatin localization remains strong, thus indicating a role in the maintenance of an avascular tissue.
Contrary to the articular cartilage, the metaphyseal growth plate is replaced by bone, and VEGF co-ordinates this process. In normal bone formation, expression of VEGF starts before birth in the hypertrophic zone and is retained postnatally (Gerber et al. 1999; Carlevaro et al. 2000; Maes et al. 2002, 2004). In mice, the gene for VEGF encodes at least three spliced isoforms having different temporal localizations in the course of increasing age. Whereas VEGF120 isoform expression remains uniformly constant until W11, isoform VEGF164 and VEGF188 transcription decrease, before their initial expression patterns are restored. VEGF protein expression correlates with these findings (Evans & Oberbauer, 2007). Our results are principally in line with these observations but we could show that VEGF protein expression dropped with increasing age and finished before mice had reached an age of 12 weeks (Fig. 5). Contrary to the decline of the chondro/osteoclasts during the establishment of the joint system, their number did not alter at the metaphyseal chondro–osseous junction (Fig. 3D1,D2). They were still abundant in 12-week-old mice and presumably play a role in the complete resorption of the growth plate and hence cessation of the bones lengthening. VEGF apparently co-ordinates these processes, and its importance for a normal development of the skeleton was shown after its having been blocked, giving rise to a suppression of blood vessel invasion and a reduced recruitment of the chondro/osteoclasts. In the absence of VEGF the long bones are shortened, despite the growth plate expansion (Gerber et al. 1999; Harper & Klagsbrun, 1999). Endostatin, according to our findings, was present until W6 but was not detected anymore starting from W8 (Fig. 7). In the human growth plate it is also encountered but in somewhat different areas, namely, within the proliferative but not in the hypertrophic zone (Pufe et al. 2004). Mice lacking collagen XVIII and its fragment endostatin (Col18a1−/− mice) show a slight though passing delay in bone development, and consequently endostatin seems not to be critical for a normal skeletal development (Sipola et al. 2009).
In conclusion, our results suggest that both VEGF and endostatin play a role in a controlled structural reorganization of the articular cartilage. After attainment of its mature architecture, endostatin keeps articular cartilage free of blood vessels. Furthermore, VEGF obviously has a major effect on the postnatal development of the long bones; endostatin, however, seems not to be involved in this process.
This study was supported by the Austrian Science Fund (FWF) (grant P20758-B05). The authors thank Prof. VMD. H. Dietrich for helping to rear the animals. We specially thank Mag. H. Hübl her valuable assistance in the laboratory and C. Siemon and Travieso for carefully reading and correcting the manuscript.