Liansheng Song

A 1-year study of osteoinduction in hydroxyapatite-derived biomaterials in an adult sheep model: part I.

The study presented here investigated hydroxyapatite biomaterials implanted in soft-tissue sites in adult sheep to determine whether these materials are osteoinductive and whether the rate of osteoinduction can be increased by manipulating the composition and porosity of the implants. For the study, 16.8-mm x 5-mm discs were prepared from mixtures of hydroxyapatite and beta-tricalcium phosphate. Five mixtures of hydroxyapatite-ceramic and hydroxyapatite-cement paste forms were studied: 100 percent hydroxyapatite-ceramic (Interpore), 60 percent hydroxyapatite-ceramic, 100 percent hydroxyapatite-cement paste, 60 percent hydroxyapatite-cement paste, and 20 percent hydroxyapatite-cement paste. Biomaterials were implanted in subcutaneous and intramuscular soft-tissue pockets in 10 adult sheep. Cranial bone grafts of equal dimension were implanted as controls. One year after implantation, the volume of all biomaterials and bone grafts was determined from a computed tomographic scan, and porosity and bone formation were determined using backscatter electron microscopy. Cranial bone and the 20 percent hydroxyapatite-cement paste implants demonstrated significant volume reduction in all sites after 1 year (p < 0.001). No significant difference in volume of the remaining four biomaterials was found. There was no significant change in pore size in the ceramic implants (range, 200 to 300 micro) and in the cement-paste implants containing 60 percent hydroxyapatite or more (range, 3 to 5 nm). Pore size in the cement-paste implants containing 20 percent hydroxyapatite increased significantly with resorption of the tricalcium-phosphate component, reaching a maximum of 200 to 300 micro in the periphery, where the greatest tricalcium-phosphate resorption had occurred. Both ceramic biomaterials demonstrated lamellar bone deposition within well-formed haversian systems through the entire depth of the implants, ranging from a mean of 6.6 percent to 11.7 percent. There was minimal bone formation in the cement-paste implants containing 60 percent hydroxyapatite or more. In contrast, cement-paste implants containing 20 percent hydroxyapatite demonstrated up to 10 percent bone replacement, which was greatest in the periphery of the implants where the greatest tricalcium-phosphate resorption had occurred. This study confirms the occurrence of true osteoinduction within hydroxyapatite-derived biomaterials, when examined using backscatter techniques. In this study, the rate of osteoinduction was greatest when a porous architecture was maintained, which was best achieved in ceramic rather than cement-paste forms of hydroxyapatite. Porosity and resultant bone formation in cement-paste implants can be improved by combining hydroxyapatite with a rapidly resorbing component, such as tricalcium phosphate.

Osteogenesis in calvarial defects: contribution of the dura, the pericranium, and the surrounding bone in adult versus infant animals.

Guided bone regeneration is a promising means for reconstructing bone defects in the cranium. The present study was performed to better define those factors that affect osteogenesis in the cranium. The authors studied a single animal model, investigating the contribution of the dura, the pericranium, and the adjacent calvarial bone in the process of calvarial regeneration in both mature and immature animals. Bilateral, 100-mm2, parietal calvariectomies were performed in immature (n = 16) and mature (n = 16) rabbits. Parietal defects were randomized to one of four groups depending on the differential blockade of the dura and/or the pericranium by expanded polytetrafluoroethylene membranes. Animals were humanely killed after 12 weeks, and histometric analysis was performed to quantitate the area of the original bone defect, new bone formation, and new bone density. Bone formation was quantified separately both at the periphery and in the center of the defects. Extrasite bone formation was also quantified both on the dural and on the pericranial sides of the barriers. Bone regeneration was incomplete in all groups over the 12-week study period, indicating that complete bone healing was not observed in any group. The dura was more osteogenic than the pericranium in mature and immature animals, as there was significantly more extrasite bone formed on the dural side in the double expanded polytetrafluoroethylene barrier groups. In both the dural and the double expanded polytetrafluoroethylene barrier groups, dural bone production was significantly greater in immature compared with mature animals. The dura appeared to be the source of central new bone, because dural blockade in the dural and double expanded polytetrafluoroethylene groups resulted in a significant decrease in central bone density in both mature and immature animals. Paradoxically, isolation of the pericranium in mature animals resulted in a significant reduction in total new bone area, whereas pericranial contact appeared to enhance peripheral new bone formation, with the control group having the greatest total new bone area. The present study establishes a model to quantitatively study the process of bone regeneration in calvarial defects and highlights differences in the contribution of the dura and pericranium to calvarial bone regeneration between infant and adult animals. On the basis of these findings, the authors propose that subsequent studies in which permeability of the expanded polytetrafluoroethylene membranes is altered to permit migration of osteoinductive proteins into the defect while blocking prolapse of adjacent soft tissues may help to make guided bone regeneration a realistic alternative for the repair of cranial defects.

Osteogenesis in cranial defects: Reassessment of the concept of critical size and the expression of TGF-β isoforms

Transforming growth factor-betas (TGF-&bgr;) have been demonstrated to be upregulated during osteoblast function in vitro and during cranial suture fusion in vivo. The authors hypothesized that spontaneous reossification of calvarial defects was also associated with upregulation of TGF-&bgr;. The present study was designed to (1) evaluate the concept of a critical-size defect within the calvaria in an adult guinea pig model and (2) investigate the association between the reossification of calvarial defects and TGF-&bgr; upregulation. Paired circular parietal defects with diameters of 3 and 5 mm and single parietal defects with diameters of 8 or 12 mm were made in 45 six-month-old skeletally mature guinea pigs. Three animals per defect size were killed after survival periods of 3 days, 1 week, 4 weeks, 8 weeks, or 12 weeks. New bone ingrowth was evaluated by assessing for linear closure by a traditional linear method and by a modified cross-sectional area method using an image analysis system in which the thickness of new bone was taken into account. Immunohistochemistry was performed using rabbit polyclonal antibodies to localize TGF-&bgr;1, -&bgr;2, and -&bgr;3. All specimens were photographed, and the intensity of immunostaining was graded based on subjective photographic assessment by three independent reviewers. No defect demonstrated any measurable bone replacement after a survival period of 3 days. All 3- and 5-mm defects were completely reossified after 12 weeks based on the linear analysis of new bone, indicating these defects to be less than critical size. However, new bone formation in the 5-mm defects never exceeded a mean of 40 percent by cross-sectional area of new bone. Percent of new bone formation by cross-sectional area was significantly higher within 3-mm defects than in all larger defects 4 weeks after the craniotomy, reaching a mean of 89 percent new bone by 12 weeks. Persistent gaps were noted on linear analysis of the 8- and 12-mm wounds by 12 weeks, and mean percent new bone by cross-sectional area remained below 30 percent. Immunolocalization demonstrated osteogenic fronts at the advancing bone edge and the endocranial side, in which the osteoblasts stained strongly for all isoforms of TGF-&bgr;. The intensity of osteoblast expression waned considerably after the majority of the defect had reossified. These data indicate that histometric analysis based on cross-sectional area more accurately reflects the osteogenic potential of a cranial defect than does linear inspection of defect closure. Although the interpretation of immunolocalization studies is highly subjective, independent assessment by three reviewers indicates that isoforms of TGF-&bgr; were upregulated during a limited “window” of time corresponding to the period of active calvarial reossification, and expression of TGF-&bgr; corresponded to osteoblast activity within osteogenic fronts. (Plast. Reconstr. Surg. 106: 360, 2000.)

Biomechanical evaluation of titanium, biodegradable plate and screw, and cyanoacrylate glue fixation systems in craniofacial surgery.

&NA; Choice of appropriate fixation after reduction of displaced bone fragments or advancement of osteotomized segments requires knowledge of the maximal force to which these segments can be subjected. The present study was performed to obtain a biomechanical comparison of a variety of resorbable fixation systems as an alternative to metal plates and screws. Sheep cadaver parietal bone segments were osteotomized and fixed with one of six methods of fixation: (A) titanium plates and screws consisting of (1) miniplates and 2.0‐mm‐diameter screws; (2) midface plates and 1.5‐mm screws; (3) microplates and 1.0‐mm screws; (B) resorbable systems consisting of combinations of butyl‐2cyanoacrylate glue and biodegradable polylactic acid/polyglycolic acid copolymer plates and 2.0‐mm screws as follows: (1) direct glue fixation of segments; (2) resorbable plates fixed to bone segments with cyanoacrylate glue; (3) resorbable plates fixed with resorbable screws. Compression testing was performed upon bone segments advanced and fixed across a central gap, and distraction testing was performed on bone segments fixed in direct contact. Force to failure in both distraction and compression was significantly greater in bone segments fixed with titanium miniplates than with any other method of fixation. Segments fixed with plates and screws, either nonresorbable or resorbable, achieved stronger fixation in distraction than in compression for all plate sizes tested. Resorbable plate and screw fixation was as strong as standard titanium midface and microplating systems in distraction, and stronger than the latter techniques in compression. With compressive forces of relapse, fixation with glue and resorbable plates was as strong as standard titanium midface and microplating systems. However, with distractive forces of relapse, glue fixation of either the bone segments or resorbable plates was weaker than both titanium and resorbable alternatives in which plates and screws were used. These findings may have direct impact on the choice of fixation devices used to support osteotomized or fractured bone segments, which are subjected to persistent muscular and soft‐tissue pull.

A 1-year study of osteoinduction in hydroxyapatite-derived biomaterials in an adult sheep model: part II. Bioengineering implants to optimize bone replacement in reconstruction of cranial defects.

The present study investigated hydroxyapatite biomaterials implanted in critical-size defects in the calvaria of adult sheep to determine the optimal bioengineering of hydroxyapatite composites to facilitate bone ingrowth into these materials. Five calvarial defects measuring 16.8 mm in diameter were made in each of 10 adult sheep. Three defects were filled with cement paste composites of hydroxyapatite and beta-tricalcium phosphate as follows: (1) 100 percent hydroxyapatite-cement paste, (2) 60 percent hydroxyapatite-cement paste, and (3) 20 percent hydroxyapatite-cement paste. One defect was filled with a ceramic composite containing 60 percent hydroxyapatite-ceramic, and the fifth defect remained unfilled. One year after implantation, the volume of all biomaterials was determined by computed tomography, and porosity and bone replacement were determined using backscatter electron microscopy. Computed tomography-based volumetric assessment 1 year after implantation demonstrated that none of the unfilled cranial defects closed over the 1-year period, confirming that these were critical-size defects. There was a significant increase in volume in both the cement paste and ceramic implants containing 60 percent hydroxyapatite (p < 0.01). There was no significant change in volume of the remaining cement paste biomaterials. Analysis of specimens by backscatter electron microscopy demonstrated mean bone replacement of 4.8 +/- 1.4 percent (mean +/- SEM) in 100 percent hydroxyapatite-cement paste, 11.2 +/- 2.3 percent in 60 percent hydroxyapatite-cement paste, and 28.5 +/- 4.5 percent in 20 percent hydroxyapatite-cement paste. There was an inverse correlation between the concentration of hydroxyapatite and the amount of bone replacement in the cement paste for each composite tested (p < 0.01). Bone replacement in 60 percent hydroxyapatite-ceramic composite (13.6 +/- 2.0 percent) was not significantly different from that in 60 percent hydroxyapatite-cement paste. Of note is that the ceramic composite contained macropores (200 to 300 microm) that did not change in size over the 1-year period. All cement paste composites initially contained micropores (3 to 5 nm), which remained unchanged in 100 percent hydroxyapatite-cement paste. Cement paste implants containing increased tricalcium phosphate demonstrated a corresponding increase in macropores following resorption of the tricalcium phosphate component. Bone replacement occurred within the macropores of these implants. In conclusion, there was no significant bone ingrowth into pure hydroxyapatite-cement paste (Bone Source, Stryker-Leibinger Inc., Dallas, Texas) in the present study. The introduction of macropores in a biomaterial can optimize bone ingrowth for reconstruction of critical-size defects in calvaria. This was demonstrated in both the ceramic composite of hydroxyapatite tested and the cement paste composites of hydroxyapatite by increasing the composition of a rapidly resorbing component such as beta-tricalcium phosphate.

A 1-year Study of Hydroxyapatite-derived Biomaterials in an Adult Sheep Model: Iii. Comparison with Autogenous Bone Graft for Facial Augmentation

Background: The present study investigates onlay bone grafts and implants in a large-animal (sheep) model to determine whether there are composite biomaterials that can maximize long-term facial augmentation when compared with conventional bone grafts. Methods: Facial augmentation was performed in 10 adult sheep. First, 16.8 × 5-mm disks were prepared from autogenous calvarial bone, hydroxyapatite ceramic, ceramic composite of 60 percent hydroxyapatite and 40 percent β-tricalcium phosphate (60 percent hydroxyapatite ceramic), and hydroxyapatite cement paste. Facial recipient sites were the body of the mandible (depository), the maxillary region (resorptive), and the frontal bone (depository). The volume of all bone grafts and implants was determined using computed tomographic scans, and the amount of bone formation was measured by means of backscatter electron microscopy 1 year postimplantation. Results: Cranial bone graft demonstrated a highly significant reduction in volume in all sites studied. Other than a slight decrease in volume of hydroxyapatite cement paste disks applied to the maxillary region, there was no significant change in volume of the biomaterials implanted in any of the remaining recipient sites. Bone replacement was greatest in hydroxyapatite ceramic (23.9 percent) followed by 60 percent hydroxyapatite ceramic (16.4 percent) and least with hydroxyapatite cement paste (4.2 percent). Minimal differences in bone replacement were noted between recipient sites. Conclusions: This study demonstrates that the volume maintenance of onlay hydroxyapatite composites is highly predictable, whereas that of cranial bone graft is unpredictable. Minimal differences were seen in bone replacement within biomaterials between “depository” and “resorptive” facial recipient sites. Ceramic forms of onlay hydroxyapatite implants demonstrated significantly greater bone replacement than did the cement paste forms of hydroxyapatite.

Effects of transforming growth factor-beta and mechanical strain on osteoblast cell counts: an in vitro model for distraction osteogenesis.

Factors known to regulate bone production during distraction osteogenesis include mechanical strain on bone forming cells and up‐regulation of transforming growth factor‐&bgr; (TGF‐&bgr;) during the distraction, or strain phase of distraction osteogenesis. In the present study, an in vitro model was used to evaluate the functional effect of exogenous TGF‐&bgr;1 on mitogenesis in murine‐derived MC3T3 osteoblasts during the period of active mechanical strain. The first hypothesis to be tested was that mitogenic suppression of MC3T3 osteoblasts by TGF‐&bgr;1 is further enhanced when these cells are also subjected to mechanical strain. To test this hypothesis, MC3T3 osteoblasts were seeded on flexible and rigid membranes. These were subjected to cyclic, vacuum‐induced strain, simulating physiologic stress loads. After 24 hours, all cells were transferred to media containing TGF‐&bgr;1, and strain was continued for an additional 48 hours. The study was repeated by using two doses of TGF‐&bgr;1. This study demonstrated that final cell counts were significantly decreased in the presence of TGF‐&bgr;1 in both the nonstrained and strained groups (p < 0.0001). The final cell count in the strained group was significantly less than that in the nonstrained group (p < 0.0001) for both concentrations of TGF‐&bgr;1 tested, confirming the initial hypothesis. The second hypothesis to be tested was that alteration in the mitogenic response of MC3T3 osteoblasts after strain is not directly due to autocrine factors produced by the strained osteoblasts. To test this hypothesis, a proliferation assay was performed on nonconfluent MC3T3 osteoblasts by using conditioned media collected from strained and nonstrained osteoblasts. This study demonstrated no significant differences in cell counts after addition of conditioned media collected from strained versus nonstrained cells, confirming the latter hypothesis. The present study demonstrates the functional significance of mechanical strain on osteoblast cell counts. Furthermore, this may help to explain the temporal relationship observed during the early distraction (strain) phase of distraction osteogenesis in rodent models in which peak up‐regulation of TGF‐&bgr;1 gene expression correlates with peak suppression of osteoblast function as measured by gene expression of extracellular matrix proteins. (Plast. Reconstr. Surg. 105: 130, 2000.)

Long-term remodeling of vascularized and nonvascularized onlay bone grafts: a macroscopic and microscopic analysis.

The present study was performed to compare vascularized and nonvascularized onlay bone grafts to investigate the potential effect of graft-to-recipient bed orientation on long-term bone remodeling and changes in thickness and microarchitectural patterns of remodeling within the bone grafts. In two groups of 10 rabbits each, bone grafts were raised bilaterally from the supraorbital processes and placed subperiosteally on the zygomatic arch. The bone grafts were oriented parallel to the zygomatic arch on one side and perpendicular to the arch on the contralateral side. In the first group, vascularized bone grafts were transferred based on the auricularis anterior muscle, and in the second group nonvascularized bone grafts were transferred. Fluorochrome markers were injected during the last 3 months of animal survival, and animals were killed either 6 or 12 months postoperatively. The nonvascularized augmented zygoma showed no significant change in thickness 6 months after bone graft placement and a significant decrease in thickness 1 year after graft placement (p < 0.01). The vascularized augmented zygoma showed a slight but statistically significant decrease in thickness 6 months after graft placement (p < 0.003), with no significant difference relative to its initial thickness 1 year after graft placement. In animals killed 6 months after bone graft placement, both the rate of remodeling and the bone deposition rate measured during the last 3 months of survival were significantly higher in the vascularized bone grafts compared with their nonvascularized counterparts (p < 0.02). By 1 year postoperatively, there were no significant differences in thickness, mineral apposition rate, or osteon density between bone grafts oriented perpendicular and parallel to the zygomatic arch. These findings indicate that the vascularity of a bone graft has a significant effect on long-term thickness and histomorphometric parameters of bone remodeling, whereas the direction of placement of a subperiosteal graft relative to the recipient bed has minimal effect on these parameters. In vascularized bone grafts, both bone remodeling and deposition are accelerated during the initial period following graft placement. Continued bone deposition renders vascularized grafts better suited for the long-term maintenance of thickness and contour relative to nonvascularized grafts.

Biomechanical and histologic alteration of facial recipient bone after reconstruction with autogenous bone grafts and alloplastic implants: a 1-year study.

&NA; Potential alteration of the underlying recipient bone resulting from a graft or implant has significant clinical relevance. The present study was designed to evaluate the biomechanical and histologic alteration of facial recipient bone with autogenous bone graft and alloplastic implants over a 1‐year period. The bilateral arches of 15 rabbits were randomized between four groups: (1) control (n = 6), subperiosteal exposure of the zygomatic arch was made; (2) onlay (n = 12), bone graft was placed as an onlay to the zygomatic arch; (3) inlay (n = 6), bone graft was placed as an inlay within the zygomatic arch; (4) implant (n = 6), a stainless steel plate was placed as an onlay to the zygomatic arch. Animals were killed 1 year after grafting. In the onlay groups, all steel implants and half of the onlay bone grafts (n = 6) were separated from the zygomatic arch; the remaining onlay bone grafts (n = 6) were left on the zygomatic arch. Three‐point breaking strength was measured through the center of the graft/implant site on the zygomatic arch, followed by histologic evaluation and histometric assessment of residual bone density. The findings demonstrated no difference in the breaking strength per unit bone area between the control zygomatic arch group and the onlay group in which the bone graft was left in place. Breaking strength of the zygomatic arch in the former two groups was significantly greater than that in either group in which the onlay bone graft or implant had been removed, and was also greater than the breaking strength in that group in which inlay bone had been placed (p < 0.05). Histologic assessment showed full‐thickness conversion in architecture of the zygomatic arch from compact to woven bone beneath onlays of either autogenous bone graft or steel implant; histometric assessment demonstrated an accompanying decrease in bone density in the latter groups relative to the control zygoma (p < 0.05). We conclude that onlay autogenous bone graft and alloplastic implants to the facial skeleton induce transformation of both graft and recipient bone from compact to woven architecture, accompanied by a reduction in bone density. The biomechanical strength of recipient facial bone is significantly weakened if an onlay bone graft or implant is removed. Weakening occurs per unit area of remaining bone, and is therefore independent of any thinning that may occur within the recipient bone because of graft/implant placement. These findings may impact upon decisions to augment stress‐bearing regions of the facial skeleton with bone graft or implants, particularly if the graft/implant may eventually require removal. (Plast. Reconstr. Surg. 101: 1561, 1998.)

Regulation of osteogenesis and survival within bone grafts to the calvaria: The effect of the dura versus the pericranium

Background: The present study evaluates the isolated role of dura and pericranium in the survival of fresh (osteoblasts viable) and frozen (osteoblasts nonviable) bone grafts. Methods: Bilateral craniectomies were performed in 48 mature rabbits. On one side, bone was replaced immediately; on the contralateral side, it was flash-frozen before replacement. Animals were randomized into four groups by placement of Silastic barriers adjacent to bone grafts, as follows: (1) control (no barriers); (2) dural barrier; (3) pericranial barrier; and (4) double (dural and pericranial) barriers. Fluorescein labels were injected at specified intervals, with animals euthanized after 1 or 10 weeks. Results: After 1 week, fresh grafts without dural barriers demonstrated greater fluorescein labeling on the dural than on the pericranial surface (p < 0.05); in contrast, fresh grafts without pericranial barriers had no statistical difference in fluorescein labeling between pericranial and dural surfaces. After 10 weeks, the new bone area was greater in fresh than in frozen grafts (p < 0.05). Total new bone area and dural-side new bone were greater in grafts without dural barriers (p < 0.001); this was not seen in grafts without pericranial barriers. Pericranial new bone was greatest in fresh grafts without a pericranial barrier (p < 0.001); this was not seen in frozen grafts. Conclusions: The dura and pericranium each contributed to osteogenesis, although dural contact was more effective. Maintenance of dural contact enhanced osteogenesis through the entire graft, whereas pericranial contact enhanced osteogenesis only on the pericranial surface of fresh grafts. These data suggest dura is largely responsible for cranial graft survival.