Alendronate Affects Cartilage Resorption by Regulating Vascular Endothelial Growth Factor Expression in Rats

Authors

  • J.H. Kang,

    1. Dental Science Research Institute, Second Stage Brain Korea, School of Dentistry, Chonnam National University, Gwangju, South Korea
    Search for more papers by this author
    • JH Kang and NK Choi contributed equally to this work.

  • N.K. Choi,

    1. Dental Science Research Institute, Second Stage Brain Korea, School of Dentistry, Chonnam National University, Gwangju, South Korea
    Search for more papers by this author
    • JH Kang and NK Choi contributed equally to this work.

  • S.J. Kang,

    1. Dental Science Research Institute, Second Stage Brain Korea, School of Dentistry, Chonnam National University, Gwangju, South Korea
    Search for more papers by this author
  • S.Y. Yang,

    1. Dental Science Research Institute, Second Stage Brain Korea, School of Dentistry, Chonnam National University, Gwangju, South Korea
    Search for more papers by this author
  • H.M. Ko,

    1. Dental Science Research Institute, Second Stage Brain Korea, School of Dentistry, Chonnam National University, Gwangju, South Korea
    Search for more papers by this author
  • J.Y. Jung,

    1. Dental Science Research Institute, Second Stage Brain Korea, School of Dentistry, Chonnam National University, Gwangju, South Korea
    Search for more papers by this author
  • M.S. Kim,

    1. Dental Science Research Institute, Second Stage Brain Korea, School of Dentistry, Chonnam National University, Gwangju, South Korea
    Search for more papers by this author
  • J.T. Koh,

    1. Dental Science Research Institute, Second Stage Brain Korea, School of Dentistry, Chonnam National University, Gwangju, South Korea
    Search for more papers by this author
  • W.J. Kim,

    1. Dental Science Research Institute, Second Stage Brain Korea, School of Dentistry, Chonnam National University, Gwangju, South Korea
    Search for more papers by this author
  • W.M. Oh,

    1. Dental Science Research Institute, Second Stage Brain Korea, School of Dentistry, Chonnam National University, Gwangju, South Korea
    Search for more papers by this author
  • B.Y. Kim,

    1. Department of Anatomy, School of Medicine, Chonnam National University, Gwangju, South Korea
    Search for more papers by this author
  • S.H. Kim

    Corresponding author
    1. Dental Science Research Institute, Second Stage Brain Korea, School of Dentistry, Chonnam National University, Gwangju, South Korea
    • Dental Science Research Institute, School of Dentistry, Chonnam National University, Yongbongdong, Gwangju 500-757, South Korea
    Search for more papers by this author
    • Fax: 82-62-530-4829


Abstract

This study was performed to determine effects of alendronate on the tibial proximal epiphyseal cartilage undergoing endochondral ossification and the expression of vascular endothelial growth factor (VEGF) from the cartilage. Alendronate was injected subcutaneously every other day in postnatal Day 1 Sprague Dawley rats. The rats were sacrificed 3, 5, 7, and 10 days after the first injection. The effect of alendronate treatment for 10 days was demonstrated from the morphological change that the area of the secondary ossification center in the epiphysis was significantly smaller in the alendronate group than that in the control group (P < 0.05). Strong immunoreactivity to VEGF was observed in the hypertrophied chondrocytes and some proliferating chondrocytes in the epiphyseal cartilage at postnatal Day 5 and was decreased after the alendronate treatment for 5 days. Immunoreactivity was observed in not only hypertrophied cells but also the peripheral cartilaginous matrix adjacent to the vascular canals invading into the central portion of the cartilage at postnatal Day 7. This reactivity was also reduced considerably by the alendronate treatment for 7 days. The level of VEGF expression was reduced by the alendronate treatment at both the transcription and translation levels. However, the transcriptional level of the flt-1 and flk-1 receptors was relatively unaltered by the treatment. These results suggest that VEGF expression is required for vascular invasion into the developing cartilage and alendronate can affect its resorption by downregulating VEGF expression. Anat Rec, 293:786–793, 2010. © 2010 Wiley-Liss, Inc.

Bisphosphonates have a similar chemical structure (P-C-P) to endogenous inorganic pyrophosphate (P-O-P). They accumulate selectively in hard tissue and inhibit bone resorption by the chelation of Ca2+ ions (Fleish,1998) and the direct inhibition of osteoclasts (Rowe et al.,1999; Reinholz et al.,2000; Rogers et al.,2000). Many clinical applications of these properties have been attempted in the treatment of osteoporosis, Paget disease, bone metastasis of tumor cells, etc. (Kyle,2000).

Vascular invasion into the cartilage is a characteristic phenomenon of endochondral ossification. Therefore, angiogenesis, which is the formation of a microvascular network, in the cartilage is believed to play a fundamental role in endochondral ossification and growth plate formation. The invasion produces vascular canals, which are accompanied by osteogenic mesenchymal cells, contributing to cartilaginous bone formation. As with other physiological and pathological processes, angiogenesis in the cartilage is finely modulated and controlled by the balance of molecules with an opposite biological action. Vascular endothelial growth factor (VEGF) is an angiogenic factor along with transforming growth factor beta and basic fibroblastic growth factor. VEGF is a dimeric glycoprotein, ranging in molecular mass from 17 to 22 kDa under reduced conditions. There are several isoforms of VEGF in mammals, resulting from alternative splicing. The soluble and cell matrix associated forms have similar biological activities (Ferrara et al.,1992; Dvorak et al.,1995). The specificity and binding affinity of the VEGF receptors, flt-1 (fms-like tyrosine kinase-1), and flk-1 (kinase domain region/flk-1, fetal liver kinase-1) to the extracellular matrix differ in the VEGF splice forms (Ferrara,1999).

Consistent with the endothelial mitogenicity, VEGF induces endothelial cell migration and proliferation both in culture and in vivo (Tischer et al.,1991). VEGF is released from chondrocytes for angiogenesis and increases the capillary permeability of the cartilage. In addition, the involvement of VEGF and its receptors in cartilage neovascularization for endochondral ossification in the long bone has been suggested (Carlevaro et al.,2000; Bluteau et al.,2007; Dai and Rabie,2007). Furthermore, VEGF is involved in the pathological neovascularization of the cartilage, such as in rheumatoid arthritis (Murata et al.,2008).

Many studies have examined the inhibitory effects of bisphosphonates on osteoclasts and the resulting inhibition of bone resorption. Despite the wide range of clinical applications of bisphosphonate, there are few reports on the inhibitory effects and mechanism of bisphosphonates on cartilage resorption (Kim et al.,2009b). Regarding the mechanism, we reported recently that alendronate, a second generation of bisphosphonates, inhibited the degradation of articular cartilage by downregulating the expression of ADAMT-4 and -5 (Kim et al.,2009a).

Bisphosphonates inhibit cartilage resorption (Lehmann et al.,2002; Kim et al.,2009a). Moreover, resedronate, a derivative of bisphosphonates, can impair bone formation by inhibiting angiogenesis (Cetinkaya et al.,2008). Indeed, the invasion of blood vessels is essential for cartilage resorption during endochondral bone formation. Neovascularization, which is induced by VEGF forms cartilage or vascular canals, resulting in the formation of an ossification center in the epiphyseal cartilage. Thus, it was hypothesized that bisphosphonates can affect cartilage resorption by regulating the expression of VEGF from cartilage cells. Although severe trauma and diseases, such as arthritis may cause changes to the cartilage, this tissue is more resistant to resorption and remodeling than bone. However, physiological resorption and remodeling of cartilage can be found during bone development only. Endochondral bone formation in the long bone addresses the invasion of blood vessels into cartilage, which is then resorbed and replaced with bone tissue, and can be a good model for examining the effects of chemicals on neovascularization.

This study was performed to test the hypothesis that bisphosphonates may inhibit cartilage resorption by downregulating VEGF.

MATERIALS AND METHODS

Administration of Alendronate

The administration of alendronate {(4-amino-hydroxybutylidene) bisphosphonate} (MK-217®; Merck, Whitehouse Station, NJ) was performed according to a previous report (Kim et al.,2009a). Briefly, alendronate (1 mg/kg) was injected into the back of rat pups (Sprague-Dawley) every other day starting from postnatal Day 1 for a total experimental time of 3, 5, 7, and 10 days. For the control group, saline was injected in a similar manner to the same aged pups. The pups were housed in laboratory animal care-approved facilities. All procedures were performed according to the guidelines of the animal care committee of Chonnam National University.

Histomorphometric Analysis

Six rats were sacrificed 10 days after the first injection. The left proximal epiphyses of the developing tibias were isolated and fixed in 4% paraformaldehyde. They were decalcified in ethylene diamine tetra-acetic acid and processed routinely for embedding in paraffin. Serial sections were made to the mid portions of the epiphysis and stained with hematoxylin-eosin. Each mid-sagittal section with the maximum area of the secondary ossification center was selected from the six epiphyses. After obtaining digital microscopic images using an in situ image analysis system (Carl Zeiss Vision, Germany), the area of the center was measured using KS Lite image software. Also, the length of the proliferating and hypertrophied chondrocyte layers was measured, according to a previous report (Kim et al.,2009a). Three measurements at anterior, middle, and posterior regions from each section were averaged (Fig. 1). For the control, six left proximal epiphyses of the developing tibias were isolated and treated in the same way. The results were statistically analyzed using a Student t-test.

Figure 1.

Measurements of the secondary ossification center (SOC) and thicknesses of the chondrocyte layers in the tibial proximal epiphysis at postnatal Day 10. Three thickness measurements were performed and averaged at the anterior, the middle, and the posterior regions from the mid-sagittal sections. P = proliferating chondrocyte layer; H = hypertrophied chondrocyte layer; SOC = secondary ossification center; EZ = erosive zone of chondrocytes.

Immunohistochemistry

Immunohistochemical staining was carried out using Vectastain Elite ABC Kit (Vector Laboratories, Burlingame, CA). Purified mouse monoclonal anti-VEGF (Santa Cruz Biotech, Delaware, CA) was used as the primary antibody. As a negative control, normal serum was used instead of the primary antibody.

Mid-sagittal sections were deparaffinized with xylene and rinsed in PBS. The endogenous peroxidase activities were blocked in 0.3% H2O2. After incubating the sections in normal blocking serum to block the nonspecific reactions, they were reacted overnight with the primary antibody. They were subsequently washed with PBS and incubated in biotinylated secondary antibody for 2 hr. Finally, the sections were incubated in an avidine-biotin peroxidase complex for 30 min and developed with AEC for the optical microscopy observations. As the negative control, the primary antibody was substituted with normal serum.

RT-PCR

RT-PCR was carried out to quantify the relative expression of VEGF and flt-1 mRNA from the proximal epiphyses of the tibia. Five rats each were sacrificed at 3, 5, and 7 days after the first injection. For the control, the same number of rats for each experimental day was treated in the same manner. Five epiphyseal cartilages from each group were mixed together. The epiphyses all together for each day were homogenized in RNAse-free tubes. VEGF (Burchardt et al.,1999), flt-1 primers (Ishii et al.,2002), flk-1 (custom-designed, Genbank No. BC087029), and GAPDH (da Silva et al.,1994) were obtained from GenoTec (Daejeon, Korea). The sequences designed for the VEGF primer pairs were 5′ TGC ACC CAC GAC AGA AGG GGA 3′ and 5′ TCA CCG CCT TGG CTT GTC ACA 3′ for the forward and reverse, respectively, generating the expected PCR product of the 564, 492, 432, and 360 bp isoforms. The sequences for the flt-1 primer pairs designed were 5′ TGG GCA TAA AAC AGT CAA AGC 3′ and 5′ GAG AGT CAG CCA CCA CCA ATG 3′ forward and reverse, respectively, generating the expected PCR product of 363 bp. The sequences for the flk-1 primer pairs designed were 5′ AAG CTT GTC CTC AGG GCA TT 3′ and 5′ TAG CCC ACA TCC TCC ACA AA 3′ forward and reverse, respectively, generating the expected PCR product of 298 bp. The housekeeping gene GAPDH, which was used as a reference, was also amplified using the specific primer sequences of 5′ CCA TGG AGA AGG CTG GGG 3′ and 5′ CAA AGT TGT CAT GGA TGA CC 3′ forward and reverse, respectively, generating the expected product of 195 bp.

AccPower® RT PreMix (Bioneer, Daejeon, Korea) was used for reverse transcription. Briefly, a mixture of the total RNA and 200 pmol Oligo dT18 was incubated at 70°C for 5 min and transferred to an AccPower RT PreMix tube. The cDNA synthesis reaction was performed at 42°C for 1 hr, followed by incubation at 94°C for 5 min to inactivate the reverse transcriptase. AccPower® PCR PreMix (Bioneer, Daejeon, Korea) was used for the PCR reaction. The reactions were carried out in a Palm-Cycler thermocycler (Corbett Life Science, Sydney, Australia) using the following profile: denaturation for 30 sec at 95°C, annealing for 30 sec at 60°C and a 30 sec extension step at 72°C, followed by a final extension step of 10 min at 72°C. The products were resolved on 1.5% agarose gel and stained with ethidium bromide. The PCR product size was determined using a 100 bp marker (Takara, Otsu, Shiga, Japan). For the negative PCR control, the cDNA template was omitted.

Western Blot

Western blotting was used to quantify the relative expression of the VEGF protein from the proximal epiphyses of the tibia. Five rats each were sacrificed at 3 and 10 days after the first injection. For the control, five saline-administered rats were treated in a similar manner. Five epiphyseal cartilages from each group were mixed together. The protein was extracted using a Ready prep protein extraction kit (Bio-RAD, Hercules, CA) and quantified using Amersham GeneQuant Pro (Amersham-Pharmacia Biotech, Arlington Heights, IL). Briefly, after transferring the extracts to a Protran nitrocellulose membrane (Whatman GmbH, Dassel, Germany), the membrane was incubated overnight with 1:500 mouse monoclonal antibody raised against amino acids 1–140 of human VEGF (Santa Cruz Biotech, Delaware, CA) at 4°C. The purified mouse monoclonal primary antibody to β-actin (Sigma-Aldrich, ST Louis, MO) was used as a reference. The membrane was incubated with 1:3,000 anti-mouse horseradish peroxidase-conjugated secondary antibody (Cell Signaling Technology, Beverly, MA), and the bound antibodies were reacted with a Lumiglo reagent (Millipore, Billerica, MA). The reactants were visualized and photographed using a LAS 4000 mini loaded with a ImageReader LAS-4000 software (Fujifilm, Minatoku, Tokyo, Japan).

RESULTS

Histomorphometry

At postnatal Day 10 of development, the tibial proximal epiphysis showed the typical layers of resting, proliferating, and hypertrophied cartilage cells. Cells in erosive zone were being replaced with developing spongy bone. In addition, there was increased invasion of blood vessels into the central region of the epiphysis, forming a secondary ossification center (Fig. 2a1). The effect of alendronate on the resorption of cartilage was examined by administering alendronate to the postnatal Day 1 rats for 10 days and measuring area of the center and thickness of cartilage cell layers. The size of the area appeared to decrease. Furthermore, the thickness of the hypertrophied cell layer increased, whereas the proliferating cell layer did not appear to have been changed by the treatment (Fig. 2a2). The area in the alendronate-treated group was significantly smaller than that in the control group (1,000,630 ± 121,479 μm2 vs. 1,519,230 ± 157,735 μm2, respectively; P < 0.05) (Fig. 2b). The thickness of the proliferating chondrocyte layer in the alendronate-treated group was similar to that in the control group (458 ± 61 μm vs. 428 ± 87 μm, respectively). However, the hypertrophied chondrocyte layer was significantly thicker in the alendronate-treated group than in the control group (491 ± 105 μm vs. 324 ± 38 μm, respectively; P < 0.05) (Fig. 2c).

Figure 2.

(a1) The hypertrophied chondrocyte layer is being resorbed and replaced by the bone. There was an invasion of blood vessels into the central region of the epiphysis forming a secondary ossification center (SOC) at postnatal Day 10. (a2) To determine the effect of alendronate on the resorption of cartilage, alendronate was administered for 10 days to postnatal Day 1 rats. After the treatment, the area of the center decreased. In addition, the thickness of the hypertrophied chondrocyte layer increased, whereas the thickness of the proliferating chondrocyte layer is relatively unaffected. (b) The area of the secondary ossification center in the tibial proximal epiphysis in the alendronate group was significantly smaller than that in the control group (*P < 0.05). (c) The hypertrophied chondrocyte layer became significantly thicker than that of the control group as a result of the alendronate treatment (*P< 0.05). However, the thickness of the proliferating chondrocyte layer was unaffected by the treatment. The values (mean ± SD) were acquired from six independent mid-sagittal sections from six independent experiments. HC = hypertrophied chondrocytes; PC = proliferating chondrocytes; EZ = erosive zone of chondrocytes; C = control group; A = alenronate group.

Immunohistochemical Findings

To demonstrate, which layer of cartilage cells express VEGF for angiogenesis and the effects of alendronate on its expression, immunohistochemical staining was performed. The tibial proximal epiphysis at postnatal Day 3 of development showed also the typical chondrocyte layers. However, there were no traces of a vascular canal into the cartilaginous matrix (data not shown). At postnatal Day 5, the tibial proximal epiphysis showed vascular invasions from the perichondrium into the very peripheral region of the epiphysis. Strong immunoreactivity to VEGF was detected from the most hypertrophied chondrocytes and some proliferating chondrocytes (Fig. 3a). However, the immunoreactivity generally became weaker by the alendronate treatment for 5 days compared to those in the control group (Fig. 3b). At postnatal Day 7, vascular invasions into the central region for the formation of the secondary ossification center were observed. Not only many hypertrophied cells but also peripheral region of the cartilaginous matrix of the burgeoning ossification center was strongly immunoreactive to VEGF (Fig. 3c). However, most vascular invasion in the group treated with alendronate for 7 days remained at the peripheral region of the epiphysis. This suggested that alendronate might prevent the invasion of blood vessels into the central portion of the epiphysis. The immunoreactivity in the chondrocytes adjacent the invasions was also reduced remarkably by the alendronate treatment (Fig. 3d). No reactivity was observed in the negative control (Fig. 3e).

Figure 3.

(a) Strong immunoreactivity to VEGF was observed in most hypertrophied chondrocytes and some proliferating chondrocytes at postnatal Day 5. (b) The immunoreactivity became generally weaker by the alendronate treatment for 5 days. (c) A vascular canal invading into the central region for the formation of a secondary ossification center is observed at postnatal Day 7. Many hypertrophied cells and peripheral cartilaginous matrix (*) adjacent to the canal showed strong immunoreactivity to VEGF. (d) A vascular canal remained at the peripheral region of the epiphysis. Hypertrophied chondrocytes and cartilaginous matrix (*) adjacent to the canal showed reduced immunoreactivity in the 7 day alendronate treated rats. (e) The negative control omitting the primary antibody did not show any reactivity. EZ = erosive zone of chondrocytes; HC = hypertrophied chondrocytes; PC = proliferating chondrocytes; VC = vascular canal.

RT-PCR

At least three splice forms of VEGF mRNA, 564, 492, and 360 bp, were detected from the chondrocytes by RT-PCR. The 564 bp splice form only first appeared at Day 3 of development. At Day 5 of development, the splice forms of 564, 492, and 360 bp were detected, and the expression pattern was maintained at Day 7. By the alendronate treatment for 3 days, the transcriptional level of VEGF mRNA was not changed. However, the transcriptional levels of 492 and 360 bp isoforms were downregulated by the alendronate treatment for 5 and 7 days (Fig. 4).

Figure 4.

The mRNA level of VEGF splice forms were determined by RT-PCR. At Day 3, only the 564 spliced form was demonstrated, whereas the other forms appeared at Days 5 and 7. The expression of the VEGF isoforms was reduced by the alendronate treatment at Days 5 and 7. The product size was confirmed using a 100 bp marker (M).

flt-1 and flk-1, VEGF receptors, were also expressed over the entire experimental period. Both flt-1 and flk-1 mRNA expressions were weak at postnatal Day 3, but increased with time. The level of their expression was not changed by the alendronate treatment for 3, 5, and 7 days (Fig. 5).

Figure 5.

The mRNA level of flt-1 and flk-1 receptors at postnatal Days 3, 5, and 7 were determined by RT-PCR. (a) The expressions of flt-1 receptor increase with time. The expressions of flt-1 were unaltered by the alendronate treatment for 3, 5, and 7 days. (b) The expressions of flk-1 receptor also increase slightly with time. Also, its expressions were unaltered by the treatment. The product size was confirmed using a 100 bp marker (M).

Western Blot

Expression of the VEGF protein was also examined in the proximal epiphysis of the tibia by Western blotting. The size of the VEGF detected was ∼15–30 kDa, as expected. VGEF expression in the control group increased from Day 3 to 10, as expected from the RT-PCR section. The level of the VEGF protein was not changed significantly after 3 day alendronate treatment, but decreased remarkably after a treatment for 10 days (Fig. 6).

Figure 6.

The VEGF protein expression was determined in the proximal epiphysis by Western blotting. This molecule was detected at 15–30 kDa in size. The expression increased from Day 3 to 10. The expression of VEGF was reduced significantly by the alendronate treatment for 10 days.

DISCUSSION

Use of bisphosphonates has been attempted not only to prevent the resorption of cartilage in arthritic diseases (Van Offel et al.,2005; Felson and Kim,2007) but also to control growth of articular cartilage (Evans et al.,2003; Kimura et al.,2008). Our preliminary study exhibited that the vascular canals invaded into the cartilaginous matrix from the periphery at ∼7 days and the secondary ossification center in the rat tibia began to form before postnatal Day 10. Therefore, the relatively early age of the experimental animals was chosen to examine the ossification process and effects of bisphosphonate on this process. The levels of VEGF and its receptors expressions were determined at the earlier period of the experiment and the morphological changes as a result of the gene expression were determined at Day 10, the last day of the experiment.

According to Evans et al. (2003), the oral dose recommended for the treatment of human osteoporosis patients is 10 mg/day, which is equivalent to 1 mg/kg/week. The amount of alendronate administered in the present study was 3.5 mg/kg/week, which can be regarded as being a relatively high dose. This relatively high dose in the present study was expected to cause morphological changes in the cartilage, even though the in vivo concentration of alendronate might be lower than the amount administered. The present result showed a decrease in the size of the ossification center and was coincident with reports showing that high doses of alendronate (>2.5 mg/kg/week) inhibited long bone growth in a mouse model (Evans et al.,2003) and affected mineralization of the cartilage at the osseo-chondral junction (Kim et al.,2009b). Furthermore, the present alendronate treatment increased the thickness of the hypertrophied cell layer, even though it did not affect the proliferating cartilage cell layer. These effects of alendronate on the growing cartilage should be evaluated carefully for clinical use because the drug might affect the modeling of the cartilage at this high concentration.

In this study, several VEGF mRNA isoforms were detected in the chondrocytes by RT-PCR; 360, 492, and 564 bp. VEGF primers were designed to include all the known VEGF isoforms according to Burchardt et al. (1999), who reported four isoforms, 564, 492, 432, and 360 bp in the rat penis. This difference in expression of the VEGF isoforms might be due to the different tissue type and function between the cartilage and penis.

When blood vessels invade the cartilaginous matrix during endochondral ossification of the long bones, they accompany perivascular mesenchymal cells, which in turn differentiate into osteoblasts to secrete the osteoid matrix. Neovascularization in the cartilage is modulated delicately by a variety of molecules that are secreted from cartilage cells at different differentiation stages (Moses et al.,1992). VEGF is considered to be a factor for neovascularization and chondrocyte differentiation (Carlevaro et al.,2000; Garcia-Ramirez, et al.,2000; Peterson et al.,2002). In this study, immunohistochemistry, Western blotting, and RT-PCR demonstrated the presence of VEGF in the epiphyseal cartilage. The development of a secondary ossification center in the epiphyseal cartilage provides a good model for examining neovascularization. In the present study, immunohistochemical studies were performed from the epiphysis at different developmental stages to determine the localization of VEGF. This in vivo study demonstrated that hypertrophied and some proliferating chondrocytes expressed VEGF. However, resting cartilage cells showed very weak expression, which supports a previous report (Carlevaro et al.,2000).

Neither vascular invasions from the perichondrium nor the secondary ossification center in the epiphysis were observed at postnatal Day 3. However, the invasions burgeoned from the periphery of the epiphysis at Day 5, became deeper at Day 7 and formed a definite secondary ossification center at Day 10. The chondrocytes became hypertrophied along the pathway of the invasions. These morphological findings are consistent with the results showing that mRNA levels of VEGF at postnatal Day 3 were lower than those at Days 5 and 7. This difference in the expression of the VEGF isoforms may reflect their role in the postnatal growth of the long bone, as suggested by Evans and Oberbauer (2007). In addition, these results suggested that VEGF from hypertrophied chondrocytes may contribute to the formation of vascular canals in the epiphysis and concur with a report showing that VEGF is expressed in hypertrophied chondrocytes for the formation of the bone marrow cavity (Alvarez et al.,2005).

In this study, VEGF expression was reduced by the relatively short term treatment of alendronate at both mRNA and protein levels, whereas the transcriptional level of flk-1 and flt-1 receptors were not changed by the treatment. In addition, the immunoreactivity to VEGF in the matrix and chondrocytes was also reduced by this treatment. These early changes in VEGF are in agreement with a report showing a significant reduction in Type I collagen C-telopeptide, the level of which correlated with the VEGF level, occurring just 1 day after a single infusion of zolendronic acid (Santini et al.,2006). These results strongly suggested that alendronate may inhibit neovascularization in cartilage by inhibiting the invasions of vascular canals. This inhibition reflected the decreased size of the secondary ossification center at postnatal Day 10. Therefore, alendronate may inhibit cartilage resorption by inhibiting the expression of VEGF from chondrocytes.

The possibility of VEGF in the abnormal neovascularization of cartilage has been suggested. Microvascular endothelial cells induced by VEGF express the mRNA for both flt-1 and flk-1 indicating that VEGF plays a role in rheumatoid arthritis (Fava et al.,1994, Nakahara et al.,2003; Murata et al.,2008). Considering the wide range of clinical uses of bisphosphonate, this study might provide a rationale for its application to tumor angiogenesis and joint diseases. The inhibition of vascularization in the cartilage by alendronate indicates its dual advantages in the management of pathological ossification of the cartilage. First, it inhibits the invasion of vessels to carry osteogeneic cells and osteoblasts. In addition, it inhibits the formation of osteoclasts and chondroclasts for further abnormal bone destruction. However, the use of bisphosphonate in the developing bone should be considered carefully because vascular invasions are essential for normal endochondral bone development.

Resorption of the matrix is needed for vascular invasions into the cartilaginous matrix. Indeed, it was reported that alendronate inhibited the degradation of the matrix by downregulating the expression of ADAMT-4 and -5 (Kim et al.,2009a). Therefore, the inhibitory effects of bisphosphonate on vascular invasion into the cartilaginous matrix should be studied further in terms of matrix resorption. Furthermore, this study examined the short and early exposure to alendronate. Similar effects of this exposure in humans would be unlikely. However, this experimental model may be valuable for examining the effects of bisphosphonate on cartilaginous bone development.

Ancillary