Representative photomicrograph depicting extensive new bone and marrow growth beyond cut edge of defect in a group 5 specimen containing growth factor mixture. Arrows indicate cut edges of the defect. Note remaining beige-colored cement (Sanderson rapid bone stain and acid fuchsin, original magnification ×2.5).
Representative photomicrograph demonstrating extensive fibrous tissue growth with some remaining cement in a group 2 specimen. Arrows indicate cut edges of the defect (Sanderson rapid bone stain and acid fuchsin, original magnification ×2.5).
Volumes of mature bone, immature bone, osteoid, cement, marrow, and fibrous tissue for each experimental group.
Arosarena OA, Falk A, Malmgren L, Bookman L, Allen MJ, Schoonmaker J, Tatum S, Kellman R. Defect Repair in the Rat Mandible With Bone Morphogenic Proteins and Marrow Cells. Arch Facial Plast Surg. 2003;5(1):103-108. doi:
From the Division of Otolaryngology, Department of Surgery, University of Kentucky Medical Center, Lexington (Dr Arosarena); Department of Otolaryngology, Albany Medical College, Albany, NY (Dr Falk); and Departments of Otolaryngology (Drs Malmgren, Tatum, and Kellman and Ms Bookman) and Orthopedic Surgery (Dr Allen and Ms Schoonmaker), State University of New York Upstate Medical University at Syracuse, Syracuse.
Objective To investigate the ability of a bone growth factor mixture and bone marrow cells to repair a critical size defect of the rat mandibular body.
Design Prospective, randomized controlled trial.
Subjects Thirty-seven male Fischer rats.
Interventions Critical size defects 4 mm in diameter were created in the left mandibular bodies of the rats. The defects were filled with a bone marrow cell suspension (group 1), a synthetic bone matrix consisting of bovine collagen and calcium hydroxyapatite cement (group 2), the matrix and marrow cells (group 3), the matrix with 100 µg of bone growth factor mixture (group 4), or the matrix with bone growth factor mixture and marrow cells (group 5). Animals were killed after 8 weeks, and the nondemineralized specimens were processed histologically. Specimens from group 1 were not processed because there was no grossly appreciable bone regeneration. Stereologic techniques were used to determine and compare the volume fractions and volume estimates of mature bone, new bone, osteoid, marrow, remaining cement, and fibrous tissue in each defect.
Results Volumes of mature bone, new bone, and remaining cement did not differ significantly among the groups (P = .30 for mature bone, P = .17 for new bone, and P = .34 for cement). However, group 4 and 5 specimens contained significantly more osteoid and larger marrow spaces than did the group 2 and 3 specimens (P<.001 for both). The specimens in groups 2 and 3 contained significantly more fibrous tissue ingrowth than did those in groups 4 and 5 (P<.001).
Conclusion The synthetic bone substitute containing bone growth factor mixture was effective in stimulating new bone and osteoid development in the rat mandibular model.
RECONSTRUCTION OF craniofacial bony defects after trauma and ablative oncologic procedures, or in the repair of congenital anomalies, is a frequent surgical challenge. Autogenous bone grafting for reconstruction of craniofacial defects is limited by resorption, potential donor site morbidity, and, at times, insufficient quantities of donor bone that matches the structural and functional qualities of the defect. Homologous and heterologous bone grafts carry the risks of disease transmission, variable resorption, and potential activation of the host immune system.1 Although vascularized bone grafts carry less risk of resorption than free bone grafts and enable the transfer of larger quantities of tissue, they require increased technical expertise and additional surgical time. Refinements in tissue engineering techniques during the past decade, and in particular the regeneration of skeletal tissues, have fostered promise in the eventual application of these technologies in the treatment of human disease.
Several studies have demonstrated the ability of bone morphogenic proteins to stimulate repair of critical size mandibular defects and effect alveolar augmentation in animal models.1-8 In this study, we investigated the ability of a bone growth factor mixture (GFm; Sulzer Biologics, Wheat Ridge, Colo) and bone marrow cells to repair a critical size defect in a rat mandibular body model. We hypothesized that the presence of bone marrow, which is a rich source of osteogenic precursor cells, would increase bone deposition.
Institutional guidelines for the humane use of laboratory animals were followed, and the Committee for the Humane Use of Animals, State University of New York Upstate Medical University, Syracuse, approved the study.
Thirty-seven syngenic Fischer 344 retired male breeder rats weighing an average ± SD of 417.9 ± 30.4 g were housed in the Department of Laboratory Animal Resources (DLAR), State University of New York Upstate Medical University, Syracuse, at a constant temperature of 24.5°C for 1 week before surgery. The animals were fed commercial rat chow and had access to food and water ad libitum. The rats were divided into 5 experimental groups: (1) those whose mandibular defects were filled with marrow cells (n = 9); (2) those whose mandibular defects were filled with a collagen (lyophilized bovine type I tendon collagen)–hydroxyapatite cement (HAC) (Bone Source; Stryker Leibinger, Portage, Mich) matrix (n = 5); (3) those whose defects were filled with the collagen-HAC matrix and bone marrow cells (n = 11); (4) those whose defects were filled with a matrix consisting of collagen treated with a bovine bone growth factor mixture (GFm) and HAC (n = 4); and (5) those whose defects were filled with the collagen-GFm-HAC matrix and marrow cells (n = 8).
The growth factor mixture is an osteogenic noncollagenous protein extract of bovine femurs that has been used to effect spinal fusion in several animal models.9-13 Its 2 major components are bone morphogenic protein 3 and transforming growth factor β2. Bone morphogenic proteins 2 through 7, as well as transforming growth factors β1 and β3, and fibroblast growth factor 1 are also present in the composite in much smaller quantities by mass (<1%). Other proteins present in the extract that probably do not contribute to tissue growth include ribosomal proteins S20, L6, and L32, and a protein related to histone H1.1. Each batch of the growth factor mixture is quality control tested for osteogenic potential via a subcutaneous injection model of bone induction in mice (James Benedict, PhD, Sulzer Biologics, written communication, April 3, 2001). An additional 9 animals were used as bone marrow cell donors.
The rats were anesthetized with a standard anesthetic cocktail consisting of a mixture of 5 mL of ketamine hydrochloride (100 mg/mL) and 0.25 mL of acepromazine maleate (10 mg/mL), administered at a dose of 0.2 mL/100 g of body weight intramuscularly. Surgery was performed with aseptic technique.
The rat marrow cell harvest was conducted according to the method of Connolly et al.14 The rat femurs were harvested, and the marrow was extracted by flushing the diaphyses with a heparinized isotonic sodium chloride solution to obtain a marrow plug. The cell aggregates were then reduced to a single cell suspension by means of serial passes through progressively smaller needles. Cell viability was assessed by trypan blue exclusion, and cells were concentrated to a density of 1010/mL. The rats used as marrow cell donors were killed at the conclusion of the harvest by lethal injection of pentobarbital sodium (100 mg/kg).
With the remaining animals, a linear incision was made through the skin, subcutaneous tissues, and masseter muscle, paralleling the inferior border of the left mandible. The buccal and lingual surfaces of the mandible were exposed with an elevator, and a 4 × 4-mm full-thickness defect was created in the body of the mandible, posterior to the root of the incisor. This ostectomy was performed with a high-speed drill and irrigation and did not interrupt mandibular continuity at the alveolus. The resulting defect was filled with marrow cells and/or the synthetic bone matrixes. The bone matrix was prepared in a sterile fashion with 25 mg of collagen, 100 mg of HAC, and 0.05 mL of sterile water per implant. In implants containing GFm, 100 µg of the growth factor mixture was used in each implant. The material was allowed to set in situ for 15 minutes. In groups in which marrow cells were used to repair the defect, 107 cells were seeded onto each implant once it had set, or into each defect. The marrow cell suspension was sufficiently viscous that it remained in the defect and was absorbed into the matrix. The surgical wounds were closed in 2 layers (periosteum-muscle layer and skin layer) with 4-0 chromic catgut sutures. Closure of the periosteum allowed the marrow suspension to remain in the defect. The animals were allowed to recover from anesthesia, and then returned to the Department of Laboratory Animal Resources for postoperative care where veterinarians supervised them. Buprenorphine hydrochloride, 0.1 mg/kg, was administered subcutaneously twice a day for the first 2 days postoperatively, and the rats were maintained on a diet of ground rat chow and water, to which they had access ad libitum.
Three rats, 1 in group 2 and 2 in group 3, were killed 2 weeks postoperatively because of presumed wound infections. Postmortem examinations disclosed that 1 rat had developed a seroma and another a hematoma (both in group 3). Three rats, 1 in group 4 and the others in group 1, died several weeks after surgery without any apparent surgical complications. The remaining animals were killed 8 weeks postoperatively by lethal injection of pentobarbital, and the left hemimandibles were harvested. Most of these animals had continued to gain weight postoperatively, and this weight gain was significant (preoperative mean weight, 417.9 ± 30.4 g; postoperative mean weight, 437.0 ± 34.6 g; P<.001).
The hemimandibles were fixed in 10% neutral buffered formalin (Sigma-Aldrich Corp, St Louis, Mo) for 3 months. The specimens from group 1 animals were excluded from histologic processing and data analysis because there was no grossly appreciable bone regeneration. The remaining nondemineralized hemimandibles were dehydrated in graded ethanols and acetone under continuous negative pressure. They were similarly infiltrated with and embedded in polymethylmethacrylate. Sectioning was performed with a rotating diamond wafering saw (Buehler Ltd, Lake Bluff, Ill). The saw excursion was 650 µm for each section, with approximately half of this thickness being absorbed by the blade width. The sections were then mounted on plastic slides (Wasatch Histo Consultants, Inc, Winnemucca, Nev), ground to a thickness of 100 µm or less, and polished. The sections were stained with Sanderson rapid bone stain (Surgipath Medical Industries, Inc, Richmond, Ill) and counterstained with acid fuchsin. This staining combination afforded sufficient contrast to distinguish bone, which stained pink, from cement (beige), osteoid (deep blue), and fibrous tissue (light blue). Immature bone, which was irregular in architecture, was distinguished from mature lamellar bone.
The volume fractions and volumes of osteoid, remaining cement, mature bone, new bone, marrow, and fibrous tissue were determined for the entire defect by means of design-based stereologic techniques.15-16 These techniques provide a statistically unbiased, quantitative estimate of the 3-dimensional composition of the defect and do not depend on any assumption regarding the tissue geometry. At least 2 sections were analyzed for each specimen (mean ± SD, 4.3 ± 1.1), with the number of fields and field size varying with the size and orientation of the defect. The cut edges of the native mandible were used to define the defect.
Data collection was performed in a blinded fashion with a ×4 objective. A video camera (Optronics LX450A CCD; Optronics, Inc, Muskogee, Okla) interfaced to a workstation (Silicon Graphics, Inc, Mountain View, Calif) was used to project images onto a monitor (Silicon Graphics, Inc). The stereologic data were collected with custom software that implemented an automated uniform random sampling protocol using computer-interfaced x- and y-axis stepping motors. The uniform random sample fields were distributed in a regular lattice pattern that was randomly positioned with respect to the tissue and were defined by a graphic overlay that included a regular point-counting lattice consisting of 108 points, with a distance of 0.12 mm between points. The volume fraction estimates (V̂V Y, ref) were made by means of test-point counting16:
where PY is the number of lattice points hitting phase Y and Pref is the number of points hitting the reference space. The total volume of the defect (V̂) was determined by means of the Cavalieri technique15:
where T is the distance between parallel sample planes, a/p is the area associated with each point, and Pi is the number of test points intersecting the defect in the ith section. Calculated areas for missing sections were filled with values of the means of adjacent sections. Total volumes (cubic millimeters) were obtained by multiplying the volume fraction of the respective defect compartment by the total defect volume.
Data analysis was performed with BMDP statistical software (BMDP Statistical Software, Inc, Los Angeles, Calif). A standard factorial analysis of variance was used to analyze mean values for the defect volumes, volume fractions, and volumes of osteoid, cement, fibrous tissue, new immature bone, and mature bone. The first independent categorical variable was "BMP" (bone morphogenic protein), indicating the presence of the growth factor mixture, and the second independent categorical variable was "marrow cells." Levene test was used to test for the equality of group variability.17 For variables that did not satisfy the assumption of equal group variances, a Brown-Forsythe test was substituted for the standard 2-way analysis of variance.18-19 A Bonferroni post hoc multiple analysis test was performed with SYSTAT software (SYSTAT Software, Inc, Richmond, Calif) to determine which pairs differed significantly. P≤.05 was considered significant.
Volume fractions for new bone, osteoid, marrow, cement, and fibrous tissue are given in Table 1. Table 2 lists total defect volume, as well as volume estimates of new bone, osteoid, marrow, cement, and fibrous tissue.
Defect volumes did not differ significantly between the experimental groups (P = .11). Defect volume means and SDs for groups 2, 3, 4, and 5 were 52.47 ± 30.37 mm3, 42.02 ± 18.24 mm3, 67.03 ± 21.85 mm3, and 63.73 ± 12.84 mm3, respectively (Table 2).
Very little of the bone matrices was replaced by mature bone, and none of the specimens in group 2 contained mature bone. There was no significant difference between the groups (P = .30) (Table 2).
The specimens containing the growth factor mixture were characterized by defect healing with immature bone around the circumference of the defect that was contiguous with the cut edges of the native mandibular bone. This bony shell enclosed a cavity that was filled with varying amounts of fatty marrow and remaining implant (Figure 1). The specimens that did not contain the growth factor mixture formed less new bone that did not extend much beyond the native mandibular cut edges (Figure 2). Although the average new bone growth for the groups with the growth factor mixture was greater than that for those without the growth factor mixture, these groups did not differ significantly (P = .17). The estimated volume of total new bone for group 2 was 1.10 ± 1.5 mm3, that for group 3 was 2.20 ± 2.37 mm3, that for group 4 was 4.62 ± 3.55 mm3, and that for group 5 was 4.04 ± 2.27 mm3.
Osteoid volumes, however, were significantly greater in the specimens containing GFm (P<.001). Osteoid volumes were 0.08 ± 0.08 mm3 and 1.37 ± 1.51 mm3 for groups 2 and 3, respectively. For groups 4 and 5, these volumes were 6.79 ± 1.42 mm3 and 4.90 ± 2.68 mm3, respectively. Post hoc multiple analysis testing showed that group 4 differed significantly from groups 2 and 3 (P<.001 for both) and group 5 also differed significantly from groups 2 and 3 (P = .01 for both).
The development of large, mature fatty bone marrow spaces was marked in the groups containing the growth factor mixture, and this difference was significant (P<.001). This accounted for the increased graft size in these specimens, in that this new growth in most cases extended well beyond the lingual and buccal surfaces of the native mandible. The estimated volumes of marrow spaces for groups 4 and 5 were 6.05 ± 3.89 mm3 and 13.64 ± 3.98 mm3, respectively. The volumes for groups 2 and 3 were 0.15 ± 0.26 mm3 and 0.09 ± 0.15 mm3, respectively. Group 4 differed significantly from groups 3 and 5 (P = .03 and P<.001, respectively), and group 5 differed significantly from groups 2 and 3 (P<.001 for both).
Although the cement volumes did not differ significantly between the groups (P = .34 for remaining cement), the specimens without the growth factor mixture demonstrated extensive fibrous tissue ingrowth, with a fibrous tissue capsule surrounding the implants (P<.001 for fibrous tissue) (Figure 2). Cement volumes were 11.48 ± 7.85 mm3 for group 2, 10.55 ± 4.07 mm3 for group 3, 19.35 ± 12.28 mm3 for group 4, and 14.22 ± 7.69 mm3 for group 5. Fibrous tissue volumes were 24.25 ± 14.61 mm3, 14.64 ± 5.97 mm3, 12.60 ± 8.45 mm3, and 5.40 ± 3.54 mm3 for groups 2, 3, 4, and 5, respectively. Group 2 differed significantly from group 5 (P<.001). Figure 3 graphically depicts the volumes of the defect, mature bone, new bone, osteoid, marrow, and fibrous tissue for each experimental group.
The use of synthetic bone substitutes for repair of skeletal defects dates back to the 19th century with the use of plaster of paris to fill tuberculous cavities in long bones.20 The biocompatibility, osteoconductive, and osseointegrative properties of calcium-based alloplasts have been recognized for many years.21 Calcium hydroxyapatite has been in use for many years as an implant for treatment of long-bone fractures and alveolar augmentation.21-22 Although calcium hydroxyapatite is very slowly replaced by bone, the process of in vivo replacement of calcium phosphate–based alloplasts has been reported to have been accelerated by the addition of bone growth factors to the materials.21, 23
The bone morphogenic proteins (except for bone morphogenic protein 1) are members of the transforming growth factor β superfamily of polypeptide growth factors.1 Approximately 20 BMPs have been identified and cloned. Recombinant BMPs have been shown to induce new bone formation in critical size mandibular defects in canine alveolar bone and rat mandibular angle models.
A number of carriers for the BMPs have been used in several studies, including bovine microfibrillar collagen, bovine collagen–ceramic calcium hydroxyapatite composite alloplasts, porous ceramic calcium hydroxyapatite disks, and synthetic bioabsorbable polymers.2-8,24-25 The use of an implant containing collagen, recombinant BMP-2, and ceramic calcium hydroxyapatite beads was believed to result in improved alveolar ridge augmentation as compared with a collagen carrier and BMP-2 without the ceramic hydroxyapatite beads, demonstrating the ability of the calcium phosphate–containing implant to maintain bone volume.25 Toriumi and Robertson1 demonstrated the osteoinductive ability of recombinant BMP-2 to repair a full-thickness critical size segmental mandibular body defect in the canine model. This is the first study, to our knowledge, that used calcium HAC and a bone growth factor mixture (GFm) in a critical size defect in a rat mandibular body model. This synthetic bone substitute has the advantage of being easily molded.
Most of the specimens in our experimental groups contained relatively large amounts of remaining implant (volume fractions were 20.76% ± 3.32%, 26.90% ± 6.17%, 26.72% ± 10.26%, and 21.38% ± 9.29% for groups 2, 3, 4, and 5, respectively; Table 1), and these volume fractions were relatively consistent between groups. It is unclear whether the large amount of remaining cement in the defects was due to insufficient healing time or to isolation of the growth factor from osteoprogenitor cells by the HAC. The addition of collagen to the cement matrix was designed to enhance cellular infiltration. Extension of the postoperative healing time may have allowed for more complete resorption of the cement, and this is consistent with the findings of Breitbart et al,23 who used osteogenin (decalcified bone matrix) and tricalcium phosphate to repair rabbit calvarial defects. At 1 month, these investigators found 0.21% bone ingrowth and 38.76% tricalcium phosphate; at 3 months, 8.85% bone ingrowth and 32.86% tricalcium phosphate; and at 6 months, 22.33% bone ingrowth and 27.25% tricalcium phosphate. Differences in bone ingrowth and remaining implant material in that study did not reach significance until 6 months.
Surprisingly, the presence of the growth factor mixture in the current study did not result in significantly greater new bone formation, as has been demonstrated with the use of BMP-2 and BMP-7 in mandibular defect repair.1-8 Although the group 4 and 5 specimens grossly demonstrated more new bone formation than the other 2 groups, and the average new bone volumes were greater in groups 4 and 5 (4.62 ± 3.55 mm3 and 4.04 ± 2.27 mm3 vs 1.10 ± 1.5 mm3 and 2.20 ± 2.37 mm3, respectively), the averages between the groups only approached statistical significance (P = .17). Again, extension of the postoperative healing time may have borne out significant differences in new bone formation because the groups containing the growth factor mixture had significantly more osteoid present (P<.001). However, Linde and Hedner2 demonstrated complete bony bridging of a similar defect in the rat mandible after 12 days; Yoshida et al6 and Zellin and Linde3 demonstrated increased bony ingrowth in comparison with control subjects after 21 and 24 days, respectively; and Higuchi et al8 demonstrated increased bony ingrowth after 4 weeks. All of these studies used BMP-2, which may explain differences in healing times. Similarly, King et al5 were able to demonstrate increased alveolar bone ingrowth with BMP-2 after 10 days in the rat periodontal defect model, while Wikesjö et al7 had similar results after 8 weeks in the canine alveolar bone model. Giannobile et al4 were able to effect improved alveolar bone healing with bone morphogenic protein 7 after 8 weeks in the canine alveolar bone model.
Inadequate healing time may also explain the relative scarcity of mature bone in the specimens in the current study. The cut edges of the native mandible were difficult to distinguish in some sections, so that areas of new mature bone may have been undercounted. Fluorochrome labeling would have helped to make this distinction.
The development of large marrow spaces in the specimens containing the growth factor mixture accounted for the increased graft size in these specimens, as this new growth in most cases extended well beyond the lingual and buccal surfaces of the native mandible. The development of large marrow spaces has not been reported in any of the previous studies that used bone morphogenic proteins for defect repair or alveolar augmentation in the mandible. However, the fibrous tissue infiltration and replacement of the implants that did not contain the bone growth factor has been reported with all of the carriers used in these studies.2-8,23-25
The rationale behind seeding of the implants in groups 3 and 5 with bone marrow cells was to determine whether the rate of bone deposition and matrix resorption could be influenced by the addition to the microenvironment of osteogenic precursor cells that were responsive to bone growth factors. In the groups without the growth factor mixture, the presence of bone marrow cells did not affect any of the measured variables. In groups 4 and 5, the presence of bone marrow cells resulted in larger marrow spaces, but had no effect on the other variables.
Few studies have evaluated new bone growth in mandibular defects implanted with bone substitutes containing growth factors with statistically unbiased, 3-dimensional morphometric analyses, and no studies have evaluated the effectiveness of the use of bone marrow cells in the repair of mandibular defects. We found that the presence of the growth factor mixture resulted in the formation of significantly larger volumes of osteoid in the specimens implanted with this growth factor mixture. The defects filled with the GFm-containing implants also demonstrated the development of large, mature marrow spaces, which has not been previously reported. The presence of bone marrow cells did not appear to affect the rate of new bone formation, nor that of implant resorption, although it did result in larger marrow spaces in the GFm-implanted specimens. These results indicate that a synthetic bone substitute containing calcium HAC and a bone growth factor mixture is effective in the repair of mandibular defects.
Corresponding author and reprints: Oneida A. Arosarena, MD, Division of Otolaryngology, Department of Surgery, University of Kentucky Medical Center, 800 Rose St, Room C236, Lexington, KY 40536-0293 (e-mail address: email@example.com).
Accepted for publication December 3, 2001.
We thank Sulzer Biologics for graciously providing the GFm growth factor mixture used in this study.