Regenerating mandibular bone using rhBMP-2: part 1 - immediate reconstruction of segmental mandibulectomies

Abstract

Objective

To describe a surgical technique utilizing a regenerative approach and internal fixation for immediate reconstruction of critical size bone defects following segmental mandibulectomy.

Study design

Prospective case series

Animals

Dogs (n=4) that had reconstruction following segmental mandibulectomy for treatment of malignant or benign tumors.

Methods

Using a combination of extraoral and intraoral approaches, a locking titanium plate was contoured to match the native mandible. Following segmental mandibulectomy, the plate was secured and a compression resistant matrix (CRM) infused with rhBMP-2, implanted in the defect. The implant was then covered with a soft tissue envelope followed by routine intraoral and extraoral closure.

Results

All dogs that had mandibular reconstruction healed with intact gingival covering over the mandibular defect and had immediate return to normal function and occlusion. Mineralized tissue formation was observed clinically within 2 weeks and solid cortical bone formation within 3 months. Computed tomographic findings at 3 months postoperatively demonstrated that the newly regenerated mandibular bone had ∼50% of the bone density and porosity compared to the contralateral side. No significant complications were noted.

Conclusion

Mandibular reconstruction using internal fixation and CRM infused with rhBMP-2 is an excellent solution for immediate reconstruction of segmental mandibulectomy defects in dogs.

Clinical Relevance

In dogs with a segmental mandibulectomy, reconstruction using rhBMP-2 and a CRM should be considered a viable surgical option.

Introduction

A common end result of critical size mandibular bone defects (i.e. an osseous defect that would not heal by bone formation during the lifetime of the animal) following segmental mandibulectomy is malocclusion due to mandibular drift.^1-7 Malocclusion can result in difficulty in eating and drinking, prehension and pain of the contralateral temporomandibular joint (TMJ).^1-3,8 While mandibular reconstruction represents the ideal solution, several aspects of this technique including the choice of graft material and matching anatomic geometry make this approach challenging.^9,10 Autologous bone grafts, bone graft substitutes, microvascular tissue transfer, and distraction osteogenesis are examples of the techniques available to address the problem.^4,9,11,12 However, these are still far from ideal due to donor site morbidity, scarce tissue availability, and limitation in graft size and contour.^9,13,14 Our group has previously investigated procedures to prevent mandibular drift following mandibulectomy.^1,2 First, mandibular rim excision with preservation of the ventral border was found to be a sound technique for small odontogenic or malignant tumors in medium and large dogs.¹ However, this technique is not advocated for more invasive tumors and in small dogs. Second, elastic training was found to be a viable option for preventing mandibular drift but required good client compliance and prevented mandibular drift in only half of the cases.² Therefore, in situations where segmental mandibulectomy is required, the ideal treatment should be anatomically-correct reconstruction of the mandible, potentially through bone regeneration, to allow proper biomechanics and thus a functional pain-free occlusion.^9,15

It has been over 40 years since Urist's pioneering work where he discovered the family of active compounds responsible for bone regeneration and named these bone morphogenetic proteins (BMPs).^16,17 Motivated by this, Sampath and Reddi created a bioassay for BMP based on the formation of ectopic bone.^18,19 Reddi proposed that BMPs are responsible for the initiation cascade of developmental events, in which progenitor cells are induced to differentiate into bone cells thus resulting in new bone formation.^20,21 Much work followed with the clinical use of recombinant human BMPs (rhBMPs) in the field of spinal fusion, fracture healing, and engineering of dental tissues.^22,23 This work resulted in FDA approval of two spinal fusion products consisting of either rhBMP-2 or rhBMP-7 delivered via absorption onto collagen matrices. ^18,22,24,25

The multifunctional growth factors of the BMP family comprise over 20 distinct ligands and play an important role not only in bone formation and remodeling but also in development and regeneration after tissue damage.^26,27 Moreover, BMPs induce a plethora of different cellular effects ranging from stem cell maintenance, migration, differentiation, proliferation and apoptosis.²⁷ As these proteins also play important roles in various other processes unrelated to bone including iron and energy metabolism, and adipogenesis, Reddi proposed naming this family of growth factors body morphogenetic proteins.^26,28,29

Although mandibular reconstruction using titanium locking plates and rhBMP-2 delivered in a scaffold has been described in isolated case reports in human and veterinary medicine, we report this technique performed on a series of cases with additional considerations and refinement.^3,8-10 Specifically, we report our experience gained from applying a collagen and calcium ceramic matrix impregnated with rhBMP-2 to effect bone regeneration in four dogs undergoing reconstruction following segmental mandibulectomy.

Materials and Methods

Case recruitment

Cases requiring segmental mandibulectomy for odontogenic or malignant tumors that were presented to a university teaching hospital were recruited for this study. A signed informed consent was obtained from the clients. All cases had preoperative minimal data base (e.g., complete blood count, serum biochemistry, and urinalysis). The dogs were staged by means of abdominal ultrasound and thoracic radiography or computed tomography (CT).³⁰ Mandibular lymph nodes were fine-needle aspirated and submitted for cytological analysis. The dogs were evaluated at a regular intervals of 2, 4, 8, and 12 weeks postoperatively and then every 6 months for the duration of the reported follow-up period.

Scaffold and rhBMP-2 preparation

The CRM (collagen sponge with embedded granules of hydroxyapatite (HA) and tricalcium phosphate (TCP) (MasterGraft Matrix^® Medtronic, Memphis, TN) and rhBMP-2 (Pfizer, Cambridge, MA) were used in this study. The volume of the defect was measured in 3 dimensions and a sufficient amount of collagen sponge (i.e., to provide a half to three quarters of the mandibular height and a length 2 mm greater than the defect span) was measured. Fifteen minutes prior to implantation the sponge was infiltrated with a 0.5 mg/mL solution of rhBMP-2 at a volume corresponding to 50% of the volume of the prepared scaffold. For example, for a scaffold that was 5 cm in length, 1 cm mandibular width and 1.5 cm mandibular height (5 × 1 × 1.5 cm³), the total defect volume was 7.5 cm³; thus, 3.75 mL of the rhBMP-2 solution was used.

Surgical technique

Detailed description of preoperative oral care, mandibulectomy techniques and principles of internal fixation are beyond the scope of this report and can be viewed elsewhere.^{1,3,4,8,31,32} Briefly and specifically, the mandible was accessed via extra and intraoral approaches.³³ Following measurements and marking of the resection area (Figure 1a), a single titanium locking plate (3.0 mm, Synthes^® Maxillofacial, Paoli, PA) was contoured prior to the amputation, capturing the normal anatomic contour of the ventrolateral mandible. The plate was then secured to the bone with appropriate size titanium locking screws (Figure 1b). The plate and screws were removed and the segmental mandibulectomy was initiated extraorally and completed intraorally Then, resection of the mandible ensued with appropriate surgical margins and intraoral closure (Figure 1c), the plate was returned and secured to the mandible via the extraoral approach. The surgical site was copiously irrigated with sterile saline as following CRM implantation, irrigation is no longer possible. The soaked collagen sponge was then implanted in the defect to fit snugly and secured circumferentially with poliglecaprone-25 (Monocryl, Ethicon, Somerville, NJ) suture to prevent migration post-implantation (Figure 1d). A new surgical pack and new surgical gloves were then used for closure. The surrounding soft tissues were sutured around the plate and sponge to provide a soft tissue envelope. The subcutaneous tissues and skin were closed routinely.

Figure 1.

Intraoperative view demonstrating the extraoral approach and measurements and marking of the resection area (A). The plate is secured to the bone with appropriate size titanium locking screws and then removed (B). Following segmental mandibulectomy (C), the plate is returned and secured and the soaked collagen sponge is implanted in the defect (D).

The dogs were kept on soft food for 2 weeks after surgery and were administered ampicilin 20 mg/kg IV preoperatively and amoxicillin/clavulanic acid 20 mg/kg orally (Clavamox, Pfizer Animal Health, NY) twice daily for 2 weeks postoperatively. Finally, pain control was achieved by administration of opioids and non-steroidal anti-inflammatory medications for 7-14 days.

Diagnostic imaging

Radiographs of the mandibles were obtained using a digital radiography system (RapidStudy EDR6, Eklin Medical Systems, Sunnyvale, CA) immediately postoperatively and at regular intervals of 2, 4, 8, and 12 weeks after surgery. Radiographs were obtained at longer time points (e.g., 5-6 months) if indicated.

Transverse, 0.625-mm, collimated CT images of the mandibles were obtained for 2 patients 3 months after surgery using a LightSpeed 16 (GE Healthcare, Milwaukee, WI) CT scanner with kVp = 120 and auto-mA. All images were reconstructed using a bone filter. A CT calibration phantom containing 5 reference rods of known density (Mindworks Software, Inc.; San Francisco, CA) was included in the field of view during image acquisition.

CT images were evaluated qualitatively and quantitatively using DICOM viewing software (OsiriX v. 4.1.2 32-bit; Geneva, Switzerland) and data analysis software (MATLAB R2011a; Mathworks^®, Natick, MA). For quantitative measurements, four transverse CT images were selected at regular intervals along the length of the mandibular repair. The Hounsfield units (HU), bone density, and porosity were measured for the mandibular repair tissue using freeform regions of interest that excluded the tooth roots and mandibular canal. The four measurements were averaged to reduce error associated with measurement and image-to-image variability.

Results

All dogs had good physical condition and results of hematological, serum biochemical analysis and urinalysis were generally considered normal. One dog had preexisting lymphangiectesia and associated mild hypoproteinemia. Thoracic radiographs and abdominal ultrasonography performed during tumor staging revealed no abnormalities. No surgical complications occurred and no neoplastic cells were identified in the surgical margins of the submitted specimens.

Mandibulectomy

Overall, 4 dogs aged 8 – 9 years (mean 8.8 years) weighting 25 - 37 kg (mean 29 kg) were included in this report. The dogs had segmental mandibulectomy with a defect size of 42 mm – 60 mm (mean 50.5 mm) for the removal of squamous cell carcinoma (n=1) and canine acanthomatous ameloblastoma (n=3). The follow-up period was 15 – 22 months (mean 19 months).

Clinical evaluation

All dogs had appropriate occlusion immediately postoperatively and throughout the duration of the follow-up period. Besides restriction of heavy chewing (e.g., no raw hide chewing or rough play) for 3 months, all dogs returned to normal activity following surgery. At 2 weeks postoperatively, hard tissue spanning the entire defect site was palpable and covered by intact gingiva. Mild oozing from the intraoral incision site was noticeable at 2 weeks in all cases but completely resolved by the 4^th week. One dog had a small cystic lesion in the gingiva that spontaneously resolved after 4 weeks. At 4 weeks postoperatively, the defect felt completely solid and no abnormalities were noticed. At 2 and 3 months postoperatively, there was no recurrence of the tumors or fractures affecting the mandibles. For the remaining follow-up period, no abnormalities were noticed and no plate exposure through the mucosa or exuberant bone reactions were noted. Furthermore, all owners reported that the dogs had an excellent quality of life.

Radiological evaluation

Representative radiographs are presented in Figure 2. The radiographic opacity of the regenerated mandible increased from postoperative radiographs to 4 weeks after surgery. At 4 weeks, the margins of the implanted scaffold became smoother and demonstrated evidence of new bone connecting the implant to the adjacent mandible. Radiographs obtained 8 weeks postoperatively demonstrated that the implant material continued to increase in opacity and formed a mineralized union with the mandible. One dog exhibited a well defined, rounded radiolucency in part of the implant material on radiographs at 4 weeks after surgery, followed by progressive increase in opacity and formation of normal-appearing cortical bone along two-thirds of the dorsal margin of the previous defect by 24 weeks after surgery. No radiographic evidence of complications related to the bone plate and screws were observed.

Figure 2.

Radiographs of the reconstructed mandible immediately postoperatively (A) and at two weeks (B), four weeks (C), eight weeks (D), and nine months (E) after surgery. Note the progressive increase in opacity and smoother margins of the implant material from postoperative to nine months. The radiographs also demonstrate progressive narrowing and opacification of the gap between the caudal aspect of the implant material and native mandible (black arrowheads) and formation of smooth, bridging, mineral opacity repair tissue at the ventral aspect of the junction between the implant and native mandible (white arrowheads). Nine months after surgery, no gap is visible between the implant and native mandible and dense bone resembling normal cortex spans the entire dorsal aspect of the previous ostectomy site.

On CT images, there was radiologic evidence of new bone formation with complete integration of the implant material with the native mandible (Figure 3). Density and porosity of the repair tissue and contralateral mandible varied widely between patients. In the 2 patients, the repair tissues achieved 46 to 54% of the density of the contralateral mandible (3 months after surgery). Moreover, in 1 dog the regenerated bone exhibited similar to slightly greater porosity (1.1 times) compared to the contralateral mandible and in the 2^nd patient, the porosity of the repair tissue was much less (0.4 times) than that of the contralateral mandible.

Figure 3.

Sagittal reconstructed computed tomographic images of the reconstructed mandible three months following reconstruction of segmental mandibulectomy in two dogs. The approximate borders of the native mandibulectomy are indicated by the white arrowheads. Note the evidence of new bone formation with complete integration with the native mandible.

Discussion

The present report provides a case series on the use of rhBMP-2 delivered via adsorption into a CRM for regenerating bone across large critical size mandibular defects in dogs. Moreover, this report comprises a surgical, clinical, and radiological experience with the use of rhBMP-2 in mandibular reconstruction. Importantly, this report emphasizes the benefits of incorporating regenerative medicine with veterinary oral surgery for mandibular reconstruction.

Overall, this combined surgical and regenerative strategy resulted in a rapid return to normal function. This was because the surgical approach allowed the correct reconstruction of normal anatomy and occlusion, and bone regeneration restored proper biomechanics. Palpable bone quickly formed using the CRM infused with the appropriate dosage of rhBMP-2. By 3 months this tissue radiographically approximated the density of native bone and appeared well-integrated. Histologically, previous reports confirmed CRM infused with rhBMP-2 to result in well-mineralized trabecular bone reflective of healthy bone turnover and remodeling.^3,34,35 The results of this study agree with several human case reports demonstrating that successful reconstruction of critical size mandibular defects could be achieved without the use of autograft or other form of bone grafts.^9,10,15 In experimental animal studies that used the same regenerative system (i.e., CRM and rhBMP-2), successful spinal fusion and mandibular reconstruction in non-human primates, dogs, and rabbits due to robust formation of bone approximating native tissue was observed.^25,34,36

The therapeutic outcome following the use of rhBMP-2 critically depends on the delivery vehicle, quantity, concentration and time of application.^37,38 The use of rhBMP-2 without a carrier is contraindicated and the selection of the matrix used for delivery must be carefully considered.³⁹ In this study and others, the CRM proved to be appropriate for the delivery and release of rhBMP-2 at the defect site.^3,9,15,25 With regards to the concentration, a study that evaluated the application of rhBMP-2 in a rat critical bone defect model has found that the degree of bone formation is dose dependent.^25,35,36,40 However, increasing the dose of rhBMP-2 beyond a certain threshold concentration does not improve bone quality, and may promote lower quality bone and invoke a detrimental inflammatory response.³⁵ In this study we used a uniform dose of 0.5 mg/ml with a 50% soak volume and bone approximating native geometry and density formed within the critical size defect and was well integrated to adjacent native tissue. However, in cases where a higher dosage of rhBMP-2 was applied in dogs there was initial excessive bone formation but this resolved within several months.^8,32,34 Although we did not evaluate a series of concentrations, we conclude that the dose generally used in this study is clinically appropriate.

Not only is the dose of rhBMP-2 critical to obtain bone formation, there must be appropriate cells and these cells must have the ability to respond to the cytokine. Thus, the success of rhBMP-2 application in our approach was due to the presence of appropriate stem cells in the local environment and their ability to differentiate into bone forming cells.⁹ Although, it is accepted that with increasing age the amount of stem cells available decrease,^34,41 the osteogenic capabilities of rhBMP-2 are not negatively affected by increasing age.³⁴ In agreement with this, we observed excellent clinical outcome suggesting that the presence and osteogenic ability of the resident stem cells in middle to older age dogs is sufficient.

In one case we observed a radiographic variation in remodeling in which a rounded, radiolucent bone void was observed, but resolved within 5 months postoperatively. One study histologically confirmed similar appearing voids to be fatty marrow instead of normal trabecular bone structure.³⁵ This phenomenon could be explained by the fact that at the molecular level, BMP-2 can induce adipogenesis in addition to, or instead of, osteogenesis through activation of transcription factor peroxisome proliferator-activated receptor gamma (PPARγ), a key regulator of adipocyte commitment.^{29,35,35,42-45} PPARγ activation leads bone marrow stem cells to differentiate to adipocytes rather then osteocytes and once this occurs, the osteogenic program is suppressed.^35,46,47

In all cases we observed mild oozing from the oral incision site, possibly due to underlying mild inflammation at 2 weeks post-surgery. This completely resolved by the 4-week recheck examination. Short term BMP-induced inflammation is commonly reported beginning on the 3rd day postoperatively, peaks at 1 week, and typically resolves 2 to 3 weeks postoperatively.^35,48,49 This response is expected as rhBMP-2 is known to be chemotactic for inflammatory cells including mono- and polymorphonuclear cells and osteoclasts-like cells.^35,50 We conclude that the general dose and method of application of the rhBMP-2 in our study, although resulted initially in minimal inflammation and mild oozing, is clinically appropriate as it resolved spontaneously by the 4th week.

Earlier reports described plate exposure through the mucosa.^3,8 In an attempt to re-establish the alveolar margin, these cases used more than one plate to buttress the defect. Plate exposure on the dorsal aspect was resolved by plate removal and did not negatively affect the long term excellent outcome.^3,8 However, using a single larger plate (e.g., 3.0 mm), this complication did not occur. Therefore, to avoid this complication we recommend that a single 3-mm titanium locking plate, placed at the ventrolateral aspect of the mandibular border be used. This approach avoids iatrogenic damage to the teeth roots, is sufficient to buttress the defect, does not result in plate failure, and avoids plate exposure through the mucosa, as the overlying soft tissues are not subject to mastication. Therefore implant removal was unnecessary and would likely be difficult and traumatic, given the osteointegration of titanium plate and screws.

In this study immediate reconstruction was performed, given the likelihood of achieving tumor-free surgical margins, based on preoperative planning. In case of more extensive tumors, it may be prudent to stage the procedure as the use of rhBMP-2 in a tumor-laden site would be contraindicated. The pre-contoured locking plate could maintain the occlusion until the tumor-free margins are histologically confirmed. However, this would require a second re-entry surgery to place the CRM with rhBMP-2.

In conclusion, the combined surgical and regenerative methodology reported here achieved predictable, timely reconstruction of critical size bone defects using a CRM with rhBMP-2. The use of rhBMP-2 should not be taken lightly as this is a very potent molecule that has wide-ranging functions and versatility and is dose dependent.^26,35 Finally, incorporating regenerative technology into veterinary oral surgery provides exciting possibilities that eliminate or minimize the morbidity associated with bone grafting and allow for a quick return to normal function.