NSAIDs inhibit fracture repair, yet the mechanism behind this effect is unknown. It is recognised that NSAIDs impede tumour growth via an inhibition of angiogenesis, primarily via a COX-2 pathway. We propose that the inhibition of fracture repair is via a similar mechanism and have investigated this hypothesis using a murine fracture model. 225 animals were randomised into either treatment (rofecoxib) or control groups and underwent a standard open femoral fracture treated using an external fixator. Outcomes measures involved assessment of healing using radiographic, histolological and biomechanical means; and measurement of blood flow across the fracture gap using Laser Doppler Flowmetry. X-ray analysis showed a similar healing pattern in both groups, however at days 16 and 32 the NSAID group had significantly poorer healing. Histological analysis showed that controls healed quicker (significant at days 24 and 32); and had more bone but less cartilage at day 8. Biomechanical testing showed controls were statistically stronger and stiffer at day 32, while NSAID animals had a significantly greater rate of fixation failure, leading to loss of pin-bone osseointegration; this occurred primarily before day 16. There was no difference in blood flow between the groups on the day of surgery, and both groups exhibited a similar flow pattern; NSAID animals however, exhibited a lower median flow from day 4 onwards, which was significantly poorer at days 4, 16 and 24. Positive correlations were demonstrated between a higher blood flow and both the histological and radiographic results. While NSAIDs were seen to inhibit fracture repair in all outcome measures; and were also noted to decrease blood flow at the fracture, with strong negative correlations being noted between NSAID prescription and fracture repair; multiple regression analysis suggest that this negative effect of NSAIDs on healing is independent of its inhibitory action on blood flow. COX-2 inhibitors are marketed as having cleaner side effect profiles and prescribing is on the rise. Recently however some of the newer COX-2 specific inhibitors have been removed from the market as their seemingly clean side effect profile has come under scrutiny. We have demonstrated that the COX-2 specific inhibitor rofecoxib does has a significant negative effect on fracture repair; and as hypothesised that it also has a significant negative effect on blood flow at the fracture site. While these outcomes strongly correlate, the mechanism behind the effect remains to be elucidated, as we have also demonstrated that these modalities are independent of each other.
During bone development and repair, angiogenesis, osteogenesis and bone remodeling (resorption) are closely associated processes with some common mediators involved. BMPs, VEGF and other cytokines are released from bone during bone resorption. Recent study showed that VEGF caused a dose- and time-dependent increase in bone resorption in vitro and in vivo, and BMP-2 markedly enhanced osteoclast differentiation induced by sRANKL and M-CSF in mouse osteoclast culture system. The aim of this study was to further examine the effects of VEGF and BMP-2 on osteoclastogenesis using in vitro human osteoclast culture system. Mononuclear cells were isolated by Lympo-Prep density gradient centrifugation from bone marrow washouts in bone samples from patients undergone total hip replacement. Mononuclear cells were plated at a density of 1 x 106/cm2 in a T-75 flask with aMEM and 15% FCS. The first medium change was made at day 7, when the floating cells were collected from the withdrawn media by centrifugation, and plated in a separate flask. The non-adherent cells in the 2nd flask were harvested again 24 hours later in a similar fashion. The non-adherent cells were then cultured in 24-well plates or calcium phosphate (Ca-P) coated plates, with osteoclast-inducing media (OC media) containing sRANKL 30 ng/ml and M-CSF 30 ng/ml, media were changed every 4 days. After 4 days culture in OC media, rhBMP-2 (3, 30, 300 ng/ml) and VEGF (25 ng/ml) were added respectively or in combination to the cell culture, and the culture was kept for total 16 days. The number of TRAP positive multinuclear cells in each well and the resorptive pit areas on the Ca-P coated plates were calculated and compared. Osteoclastic cell phenotype was defined by expressing tartrate resistant acid phosphatase (TRAP), vitronectin receptor (VNR) and resorptive pit assay. By day 12–14, osteoclastic cells were found in all the experimental groups, they were positive for TRAP and VNR. The number of TRAP+ multinuclear cells were significantly reduced (p<
0.05, t-test) when rhBMP-2 (30 and 300 ng/ml) were present, and this was further reduced (p<
0.01) when rhVEGF was added together with rhBMP-2, comparing to the culture with OC media alone. Extensive lacunar resorption pits in the Ca-P coated plates were found in the culture treated with OC media and OC media with rhVEGF (25 ng/ml). The resorption pit areas were, however, significantly reduced when rhBMP-2 was added at 30 and 300 ng/ml with or without rhVEGF (25 ng/ml, p<
0.05, t-test). The presence of low concentration of rhBMP-2 (3 ng/ml) with VEGF had no effect on osteoclast number or the areas of resorption pit formation. In contrary to previous findings in the mouse osteoclast culture system, the present study had shown that the presence of rhBMP-2 at 30 and 300 ng/ml had strongly inhibited osteoclast differentiation and bone resorptive capability in the human osteoclast culture system, and the inhibition was further enhanced by the presence of rhVEGF. This study implies that VEGF and BMP-2 may be important, yet to be defined regulators, for osteoclastogenesis.
Dentin matrix protein (DMP-1), a phosphoprotein highly linked to dentin formation, has recently been reported to have an important role in skeletal development. Previously we reported that adult mice lacking the gene for DMP-1 exhibit the characteristics of chondrodysplasia, osteoarthritis, and showed severe defects in mineralization. DMP-1 knock-out (KO) mice display a profound defect in mineralization, and this is not due to a systemic defect in calcium/phosphate metabolism because serum levels of calcium and phosphate are similar to those in the wild-type mice. Although KO neonates and newborns appear normal, upon closer examination, these animals exhibit skeletal abnormalities, which include delayed secondary ossification and impaired bone remodelling. Heterozygous DMP-1 (H) mice however, show no apparent differences to the wild-type mice. In this study, biomechanical assessment tests of bones from DMP-1 KO mice were performed. Fifteen heterozygous, H, (DMP-1 +/−) and 15 KO, (DMP-1 −/−) male mice were produced and used in this study. At 1, 3 and 7.5 months of age, the mice were sacrificed and 4–5 ulnae from each animal group were harvested and stored in 70% ethanol solution. Volumetric density (BMD) measurements of the intact ulnae were performed using peripheral quantitative computed tomography (XCT960M; Stratec, Pforzheim, Germany) and Norland Stratec software version 5.10. One millimetre thick slices were scanned at a distance of 1 mm under the articular cartilage surface of the elbow as identified by the scout view of the CT scan. BMD of the corticalis and subcortical bone were recorded. Cross-sectional area measurements were also made at the mid-diaphysis of the ulnae. Biomechanical tests were performed in 3-point bending, with supports 3.5 mm apart at a rate of 3 mm/min (Lloyd Instruments Ltd, UK). The ultimate load, yield load and stiffness were determined from the load-displacement curves. All data were analysed using Mann-Whitney U tests (SPSS, Version 9, Chicago, Illinois). Differences were considered significant at p <
0.05. Density studies revealed that H mice had higher BMD than KO mice at all ages (p <
0.001). In the H and KO mice, the cortical BMD peaked at 3 and 7.5 months, respectively. At 1 month, the mean cross-sectional areas of the ulnae were larger in H mice compared to KO mice (0.50 mm2 Vs 0.33 mm2). However at 7.5 months of age, the reverse was observed (H = 0.75 mm2 and KO = 0.98 mm2). Biomechanically, stiffness increased with age at a higher rate in H mice than KO mice. Significant differences were observed at 3 months (p<
0.01) and 7.5 months (p<
0.05) between the two animal groups. There were no significant differences between stiffness values at 1 month. This study has demonstrated that DMP-1 deficiency leads to:
severely compromised bone mineralization; poor biomechanical properties of the long bone; and delayed bone development and remodelling. In conjugation with previous findings that DMP-1 plays important roles in the early developmental stage of bone through its effects on osteogenic gene expression of Cbfa1, Col I, and Col II and regulating vascular invasion, the current study may suggest another important role for DMP-1 as a regulator for skeletal mechanostasis.
The potential importance of bone morphogenic proteins (BMPs) to improve fracture healing is of great interest to orthopaedic surgeons. Although the complex mechanisms leading from the presence of local BMP (either endogenous or exogenous) to form bone is increasingly understood, however most appropriate time to administer exogenous BMP has yet to be elucidated. The purpose of this study was to investigate when BMP may be administered to a fracture arena in order to best improve fracture healing. Forty mice were randomised into 4 groups; (group I) control, treated at day 0 with placebo; (groups II, III and IV) treated with BMP at days 0, 4 and 8, respectively. All animals underwent a previously validated surgical procedure involving the creation of an open femoral fracture which is stabilised using a 4 pin external fixator. Thirty microlitres of bovine serum albumin (BSA) alone was used in group I, and the other groups (II, III and IV) were treated with a combination of the BSA and 2.5 microgrames of rhBMP-2. The BSA and rhBMP were injected through a lateral approach immediately after operation, or at 4, or 8 days postoperatively. At days 0, 8, 16 and 22, sequential radiographs were taken using a digital x-ray machine and at day 22 all animals were sacrificed. Both femora were harvested and assessed biomechanically in 3-point bending prior to fixation for histological evaluation. All data were analysed using Mann-Whitney U tests (SPSS, Version 9, Chicago, Illinois) and differences were considered significant at p <
0.05. X-ray analysis indicated that healing of fractures treated with BMP at day 0(group II) or day 4(group III) was significantly greater than that at both days 16 and 22 (p <
0.05) than those animals in placebo (group I) and BMP day 8(group V) treatment groups. Although the administration of BMP at day 4 seemed to cause more bone formation than treatment at day 0, no significant difference were observed. There were no differences between group IV and group I. Biomechanically, group III exhibited ultimate load values closest to the contralateral unoperated femora followed by group II, then IV and finally the control group I. Significant differences (p <
0.05) were observed between the control animals (group I) and both groups II and III. Qualitative histology suggested that at 22 days after surgery, only groups II and III had healed with woven bone. Group I and group IV had considerable amounts of fibrous tissue and cartilage at the fracture gap. This study suggests that a single percutaneous injection of BMP has a positive effect on fracture healing in this model, when prescribed between the time of injury (day 0) and 4 days. Data suggests that the most effective timing of delivery of BMP may not be at the time of surgery but actually in the early healing phase. The day 4 time point in the mouse model is likely to equate to that of 7–10 days in larger animals or humans. This suggests that current human treatment practices may require further investigation in order to elucidate the most appropriate time of delivery for these important proteins. This work may negate the current requirements for carrier products and large doses of these expensive drugs.
We hypothesised that in response to fracture, some integral osteoblasts are recruited via the circulation from remote bone marrow sites.
All animals had bone marrow harvested from their right tibia by saline flush. The mononuclear cells were isolated and culture-expanded in osteogenic medium for 3 weeks. Fluorescent reporter molecules were incorporated into the cell membranes, 24 hours prior to re-implantation of the cells into the fracture model. A 3 mm ulnar defect was preformed in all the animals. In groups I–III this was established 48 hours prior to cell re-implantation. The animals were sacrificed at least 3 weeks after fracture surgery. Representative samples of the fracture callous, lung, liver, spleen and kidney were harvested from all animals and cryo-sectioned. Using confocal microscopy, the labelled cells were expressed as the average in 5 high power fields for each solid tissue. In addition, cyto-spins were made from blood and marrow and the cell number expressed as a percentage of the total cells.
In all sections, these labelled cells appeared on trabecular surfaces in an osteoblastic fashion, but occasionally they were surrounded by osteoid, corresponding to osteocytes. A small number of labelled cells were found in the blood, bone marrow, lung, liver and spleen of all animals in groups I–III. No labelled cells were identified in the kidney tissue.
The thrombin-related peptide, TP508, is a synthetic 23 amino acid peptide, which represents the receptor binding domain of thrombin. TP508 mimics thrombin by interacting with receptors on cells involved in tissue repair. TP508 has been shown to enhance revascularization of injured tissue, and promote soft tissue wound healing, cartilage repair, and fracture repair. The aim of this study is to (1) test the effect of TP508 on bone regeneration during distraction osteogenesis; (2) study the chemotactic effect of TP508 on human osteoblasts. Unilateral tibial osteoectomies were performed and stabilized with MX100 Orthofix lengthener in 5 male adult NZW rabbits. After 7 days, distraction was initiated at rates of 1.4 mm / day for 6 days. TP508 (100 μg/ml, n=2; 10 μg/ml, n=1) or saline (300 μl, n=2) was injected into the osteotomy / lengthening gap at days 1, 7 and 14 post surgery. Animals were sacrificed at 2 weeks after leg lengthening. Bone formation in the regenerate was assessed by radiography, quantitative computed tomography (pQCT) and histology. For chemotaxis studies, MG63 cells were cultured on glass cover slips for three days, and then inverted onto a Dunn chamber slide and sealed with dental wax. Gradients of TP508 (1, 10, 100 μg/ml) were added to the outer well and plain medium to the inner well. A sequence of images of the cells between the wells was taken via a CCD camera for 9 hours at interval of 10 minutes. Movements of individual cells were tracked and statistically analysed by a specially written Macro program. The Rayleigh test for unimodal clustering was used to determine the directional chemotactic movements. The radiographic evaluation indicated a significant increase in new bone in the distraction regenerate in the TP508 treated groups at 1 and 2 weeks. pQCT images at 2 weeks demonstrated more advanced bone formation in the TP508 treated animals compared to the control. The mean total bone mineral density (BMD) of the regenerate, obtained from 3 slices was significantly greater (p = 0.019, t-test) in the TP508 treated group (BMD = 479.20 +/− 35.57 mg/ccm) than that in the saline control group (BMD = 355 +/− 2.83 mg/ccm). The histological evaluation supported the radiographic and the pQCT results. For chemotaxis study, no directional movements of the cells were found in the controls, whereas the MG63 cells were strongly chemotactic to TP508 at 1, 10 and 100 μg/ml concentrations. This preliminary study shows that administration of TP508 enhances bone formation during distraction osteogenesis in the rabbit. The findings also show that TP508 has a chemotactic effect on osteoblasts, consistent with the effect of TP508 on fracture repair. A large animal study is in the process to confirm these findings and explore the underlying mechanisms.
That fracture callus with overlying muscle crush would contain raised expression of acute inflammatory cytokines (IL-1β, IL-6 and TNF-α). That application of locally applied blocking antibodies to these inflammatory cytokines might negate excessive cytokine release and modulate fracture healing in this model.
Injection of anti-TNF-α antibodies into MC mice caused more new bone formation on day 16 (p=0.03) and day 24 (p=0.06), stiffer calluses at day 24 (p=0.01) and faster fracture gap obliteration at day 16 (p=0.05) and day 24 (p=0.001). IL-1β blockade had slightly less effect, more new bone formationd ay 16 (p=0.01) and day 24 (p=0.03), slightly stiffer (p=0.08), but no significant difference in fracture gap obliteration from controls.
In relation to the conduct of this study, one or more of the authors is in receipt of a research grant from a non-commercial source.