Evaluating the bone‐regenerative role of the decellularized porcine bone xenograft in a canine extraction socket model

Abstract Objective To evaluate the efficacy of a novel decellularized porcine bone xenograft, produced by supercritical carbon dioxide extraction technology, on alveolar socket healing after tooth extraction compared to a commercially available deproteinized bovine bone (Bio‐Oss®). Materials and methods Nine dogs (about 18 months old and weighing between 20 kg and 30 kg) underwent extractions of lower second to fourth premolars, bilaterally. The dogs were randomly selected and allocated to the following groups: Group 1: control unfilled socket; Group 2: socket filled with decellularized porcine bone xenograft (ABCcolla®) and covered by a commercially available porcine collagen membrane (Bio‐Gide®); Group 3: socket filled with Bio‐Oss® and covered by Bio‐Gide® membrane. One dogs from each group was sacrificed at 4‐, 12‐, and 24‐week to evaluate the socket healing after tooth extraction. The mandible bone blocks were processed without decalcification and specimens were embedded in methyl methacrylate and subjected to histopathology analyses to evaluate the bone regeneration in the extraction sockets. Results At 24‐week after socket healing, ABCcolla® treated defects demonstrated significantly higher histopathology score in new bone formation and bone bridging, but significantly lower score in fluorescent labeling than those of the Bio‐Oss®. In the microphotographic examination, decellularized porcine bone xenograft showed similar characteristics of new bone formation to that of Bio‐Oss®. However, there was significantly less remnant implant materials in the decellularized porcine bone xenograft compared to the Bio‐Oss® group at 24‐week. Thus, the decellularized porcine bone graft seems to have promising bone regeneration properties similar to that of Bio‐Oss® with less remnant grafted material in a canine tooth extraction socket model. Conclusions Within the limits of the study, we concluded that ABCcolla® treated defects demonstrated significantly more new bone formation and better bone bridging, but less amount of fluorescent labeling than those of the Bio‐Oss® group. However, clinical studies in humans are recommended to confirm these findings.

Bio-Oss ® with less remnant grafted material in a canine tooth extraction socket model.

Conclusions:
Within the limits of the study, we concluded that ABCcolla ® treated defects demonstrated significantly more new bone formation and better bone bridging, but less amount of fluorescent labeling than those of the Bio-Oss ® group. However, clinical studies in humans are recommended to confirm these findings. The key processes of healing in an extraction socket have been well documented in both dogs and humans (Amler, 1969;Cardaropoli et al., 2005;Pietrokovski & Massler, 1967;Trombelli et al., 2008), although bone remodeling has been reported to be three to five times quicker in dogs than in humans. The healing process of an extraction socket can be divided into three sequential phases including inflammatory, proliferative and modeling/remodeling (M. G. Araujo et al., 2015). In general, following tooth extraction, most of the socket is filled with a blood clot during the first 3 days of healing. Ultimately the blood clot is replaced by granulation tissue comprised of inflammatory cells, vascular sprouts, and immature fibroblasts. The granulation tissue is steadily replaced by a temporary connective tissue matrix containing immature collagen fibers and cells.
Osteoid formation begins as early as 1 week after tissue remodeling.
Immature woven bone can be identified as early as 2 weeks after tooth extraction and remains in the socket for several weeks. As bone remodeling and modeling progress, the woven bone is replaced with lamellar bone or bone marrow and the socket walls undergo bone resorption. Consequently, bone remodeling results in the reduction of both alveolar ridge width and height.
The extent and magnitude of ridge resorption post-extraction typically result in a certain degree of bone loss in both horizontal and vertical directions. It has been demonstrated that loss of alveolar ridge height and width is an irreversible phenomenon after tooth extraction (M. G. , 2009Schropp et al., 2003).
A systematic review investigating the magnitude of human ridge dimensional changes following tooth extraction has revealed that a horizontal bone loss of 29%-63% and a vertical bone loss of 11%-22% occur 6 months after tooth extraction (Tan et al., 2012). It is expected that half of the original alveolar ridge width will be lost in 6 months after tooth extraction. Furthermore, extensive resorption of the alveolar ridge may have a significant impact on implant-supported therapy (Seibert & Salama, 1996). Therefore, alveolar ridge preservation (also known as socket preservation) after tooth extraction often leads to a decrease in post-extraction dimensional loss. Therefore, alveolar ridge preservation is an effective therapy to prevent or mini- Xenogeneic bone replacement materials are obtained from nonhuman species and have been widely used in clinical periodontal regeneration (Sheikh et al., 2017). Processing of the xenogeneic bone tissue must be conducted to warrant the elimination of immunogenic components and pathogens before its clinical use. Different methods have been applied in the processing of non-human tissue, and innovative supercritical carbon dioxide (SCCO 2 ) technology is one of those methods. The regenerative property of the porcine grafting material can be improved by a decellularization process using SCCO 2 extraction technology (Huang et al., 2017). Using CO₂ as a solvent in the SCCO₂ technique has many advantages. CO₂ is natural, safe, nontoxic, non-corrosive, non-flammable, easily accessible, and cost effective (Fages et al., 1994). There is no solvent residue or off-odors, and supercritical CO₂ easily solubilizes hydrocarbons including oil and lipids for their removal. However, polar molecules such as amino acids and proteins are preserved. The use of SCCO 2 in sterilizing amniotic membrane-derived tissue grafts while simultaneously preserving their biological characteristics has been well-documented (Wehmeyer et al., 2015). Furthermore, SCCO 2 has been used for bone delipidating, because it possesses a good solvent capacity for lipids (Fages et al., 1994). In particular, the mechanical properties of bone can be enhanced due to cross-linking of collagen chains. A significant reduction in the antigenicity by SCCO 2 without compromising the strength of bone, similar to native bone, has been reported (You et al., 2018).
With this theoretic background, it was hypothesized that the decellularized porcine bone xenograft, created by SCCO 2 extraction technology, exhibits comparable outcome to that of a commercial deproteinized bovine bone (Bio-Oss ® ) in socket healing of an extracted tooth in dog mandible. The purpose of this study was to evaluate and compare bone regeneration in canine dental sockets when using decellularized porcine bone xenograft or Bio-Oss ® .
2 | MATERIALS AND METHODS 2.1 | Preparation of decellularized porcine bone (ABCcolla ® dental bone graft) Porcine bones were purchased from Tissue Source, LLC (Indiana, USA). The bone was cleaned of residual tissues and washed with phosphate-buffered saline (PBS). The porcine femur was prepared by first sectioning into circular discs using a mechanical saw, followed by cleaning in an ultrasonic bath. Then, the cleaned circular bone sections were ground into granules (0.25-1 mm) using a mechanical grinder before undergoing the SCCO 2 decelluarization process. The bone was placed on a tissue holder, which was then inserted into a SCCO 2 vessel system (Helix SFE Version R3U, Applied Separations Inc., Allentown, PA) containing 30 ml 60% ethanol. The SCCO 2 system was then operated at 350 bar and 45 C for 80 min to produce decellularized porcine bone (ABCcolla ® , ACRO Biomedical Co. Ltd., Kaohsiung, Taiwan). FDA approved and commercially available Bio-Oss ® and Bio-Gide ® (porcine collagen membrane) were purchased from Geistlich Pharma North America Inc. (Princeton, NJ).

| Surgical procedures
Extraction of mandibular premolars was performed under local and general anaesthesia. Surgical extraction of mandibular premolars (P2 to P4) was performed under general anaesthesia with Atropine (0.4 mg/10 kg body weight, i.m.) and Zoletil (10 mg/kg, i.m.) 30 min before surgery. The dogs were placed on a heating pad, intubated and inhaled with 1%-3% isoflurane and monitored during the surgery.
After disinfection of the surgical site with 10% povidone-iodine solution/1% titratable iodine, 2% lidocaine HCl with epinephrine 1:100,000 was administered by local infiltration at the buccal and lingual aspects. An intra-sulcular incision was made around the neck of all four premolars with two vertical releasing incisions placed on buccal aspect extending beyond the muco-ginigval junction. Full-thickness mucoperiosteal flaps were elevated on both buccal and lingual aspects to expose the crestal bone. The three premolars (P2 to P4) on each side of the mandible were hemisected vertically with a dental carbide bur to separate all teeth into single root segments. Extraction was then carried out carefully with the use of forceps and elevators without damaging the alveolar cortical bone. All extraction sockets were checked visually after irrigation with 0.9% saline solution.
Nine dogs were randomly divided into three groups: Group 1: control (unfilled); Group 2: the socket was filled with ABCcolla ® dental bone graft and covered by Bio-Gide ® membrane; Group 3: the socket was filled with Bio-Oss ® and covered by Bio-Gide ® membrane. The flaps were released by periosteal incision in order to mobilize the flap to achieve a tension-free primary wound closure. The soft tissues were closed and sutured with absorbable polyglactin 910 (Vicryl 4-0, Ethicon™; Johnson & Johnson Medical Devices Companies, New Brunswick, NJ), in which the post-extraction sockets were completely covered with the mobilized flaps for the control group, and the Bio-Gide ® membranes were mainly covered by the flaps for the other two groups. Hydration was maintained by intravenous infusion with 0.9% NaCl solution during the whole surgical procedure. Postoperative pain management and wound care were provided until the animals were returned to activities ad libitum. Softened dog food was provided during the first-week post-surgery and wound healing was monitored closely to avoid infection.

| Postoperative care
Animals were observed throughout the postoperative period and received the analgesic buprenorphine (0.02 mg/kg, i.m.) for two additional doses at about 8 hours apart for 3 days. Antibiotic injection of Ceftriaxone (50 mg/kg, i.m.) was given at end of surgery and twice a day thereafter for 3 days.

| Animals sacrifice
Tetracycline-calcein double-labeling was used in this study to assess the new bone growth rate. All animals received calcein (10 mg/kg, s.c.) and Tetracycline (10 mg/kg, i.v.) injection once at 2 and 9 days, respectively, prior to sacrifice, to be able to visualize new mineralizing tissue under fluorescent microscopy. The animals were humanely euthanized by an overdose of pentobarbital (100 mg/kg) at the end of 4-, 12-, and 24-week (n = 3, with three per group and one at each time point per group) after the surgery. At necropsy, the wound healing was examined and photographed. Dog mandibles were microradiographed in a digital X-Ray machine (Faxitron model MX20; Faxitron Bioptics, LLC., Tucson, AZ) to document the defect locations and orientations.

| Biomechanical analysis
The healed extraction socket from both sides of the mandible was sectioned to generate two rectangular bone blocks. Each bone block contained the healed extraction socket of P2, P3, and P4 region. The biomechanical testing was performed using an MTS testing machine (MTS 858 Mini Bionix ® II Biomaterials Testing System, Eden Prairie, MN). Only bone blocks collected from dogs euthanized at 24-week were tested for the compression analysis. The specimen was placed on a specially designed platform with a self-aligning function to ensure vertical compression. The sampling frequency was set at 500 Hz with a probe diameter of 2.0 mm. A pre-load of 40 N with 30 s of accommodation time followed by a continuous and progressive load at a rate of 2 mm/min was applied. The first peak force detected during the test was recorded as the ultimate strength. The displacement versus force was recorded to calculate the maximal load.
Maximal load in Newtons, stiffness in Newtons/mm, and energy to the maximal load in millijoules were measured and recorded in the biomechanical analysis of the bone blocks.

| Histological preparation and histopathological analysis
All samples were immersed in 10% neutral buffered formalin for histo- Histopathologic scoring is a tool by which semi-quantitative analysis of bone healing in tissue-engineer-treated defects can be achieved (Han et al., 2018). A histological scoring system (Table 1) Table 1

| Statistical analysis
A non-parametric Mann-Whitney test with the Bonferroni correction were used to compare the difference between groups. All data were presented as means ± standard error (SE). Statistical analysis was performed by using SPSS software, version 17 (IBM Corp., Armonk, NY).
p value <0.05 was considered statistically significant. The significance level after Bonferroni correction was set at α < 0.017 (α = p/k, where p = 0.05, k is the number of groups = 3).

| RESULTS
In the post-operative period, the dogs did not show any complications. The postoperative wound healing around the surgical sites in all the dogs was normal and no dog needed any intensive care. The specimen tissues were well preserved at the sacrifice and free from inflammation.

| Biomechanical analysis at 24-week
The results of the biomechanical analysis, including maximal load, stiffness, and energy to maximal load, were expressed as the means ± standard error in Figure 1. ABCcolla ® treated sites did not show any significant (p > 0.05) differences compared to Bio-Oss ® treated defects in maximal load and the energy to maximal load. However, the unfilled control sites showed a significant increase in maximal load than those of ABCcolla ® and Bio-Oss ® treated defects (Figure 1(a); p < 0.01) and significant greater in the energy to maximal load than that of ABCcolla ® (Figure 1(c); p < 0.01). In addition, the stiffness of the Bio-Oss ® treated defects were significantly decreased compared to the ABCcolla ® treated defects and control; whereas, no significant difference was found between ABCcolla ® treated defects and unfilled control sites (Figure 1(b)). The significant differences between groups in maximal load, stiffness, and energy to maximal load mentioned above were not changed even after a Bonferroni correction. Our results demonstrated that ABCcolla ® and Bio-Oss ® treated defects showed similar characteristics in both maximal load and energy to maximal load; however, ABCcolla ® treated defects exhibited higher stiffness than that of Bio-Oss ® treated sites.

| Histopathology analysis
The mean values of histopathological score in each parameter and their statistical comparison were shown in

| DISCUSSIONS
Bone substitutes, such as porcine bone xenografts, have been demonstrated to be an excellent grafting material for bone augmentation procedures. Clinical studies using cortical porcine bone showed excellent results and promising clinical application (Scarano et al., 2011).
Previously, we have demonstrated that the regenerative ability of porcine grafting material can be improved by a decellularization process using SCCO 2 extraction technology (Huang et al., 2017). By employing this novel SCCO 2 extraction technology, the cellular components of porcine corneas were efficiently removed, and the biomechanical properties of the scaffold were well preserved. In addition, the acellular corneas exhibited no immunological reactions, displayed good biocompatibility and long-term stability. We concluded that acellular porcine cornea developed by SCCO 2 extraction technology was a suitable scaffold for corneal wound healing (Huang et al., 2017). In the F I G U R E 1 Biomechanical analysis of extraction socket in dog mandible after 24 weeks of healing. The biomechanical parameters included (a) Maximal load, (b) Stiffness, and (c) Energy to maximal load. The black, gray, and white color bars were used to represent ABCcolla ® , Bio-Oss ® , and control groups, respectively. Results were expressed as the means ± standard error (n = 6). *: p < 0.05, **: p < 0.01-significantly different between ABCcolla ® or Bio-Oss ® and unfilled control; #: p < 0.05 -significantly different between ABCcolla ® and Bio-Oss ® ; @: Bonferroni corrected p < 0.017, @@: Bonferroni corrected p < 0.0033 -significantly different between ABCcolla ® or Bio-Oss ® and unfilled control; $: Bonferroni corrected p < 0.017 -significantly different between ABCcolla ® and Bio-Oss ® F I G U R E 2 Microradiographs of the mandible taken at 4-, 12-, and 24-week after tooth extraction. Representative Xray photographs of SCCO 2 decellularized bone graft on bone regeneration in a canine extraction socket model at 4-week, 12-week, and 24-week after healing T A B L E 2 Histopathological findings during the healing of the extraction socket at various post-operative intervals  The present preclinical study compared two similar inorganic bone grafts from different animal species. It has been reported that the placement of a barrier membrane greatly improved the potential for bone regeneration in a surgically created defect (Nakajima et al., 2007). In a preclinical study, M. Kim, Kim, et al. (2008)  dog were all assigned in one treatment modality. It has been reported that clustered randomization could result in loss of efficiency, particularly in terms of effective sample size (Hauck et al., 1991). Nevertheless, it may also create some issues if all the six extraction sockets were randomly assigned to three different treatment groups. The defect sizes of the P2 to P4 extraction sockets were different and may have an impact on bone regeneration. This is the rationale that we select cluster randomization by allocate treatment to individual dog instead of extraction site. Lastly, caution should be taken when interpreting the mean data in the histopathology scores because the ordinal numbers from 0 to 4 were used in the scoring system.
Although, the ordinal method is commonly used by pathologists for the evaluation of a medical device, it does make the data interpretation more challangings (Gibson-Corley et al., 2013). In the ordinal type of data, samples were allocated to a category exhibiting an ordered progression in degree of severity, such as from score 0 (absent) to 4 (extensive) in this study. The problem of the ordinal data is that they are separated by unequal intervals. For example, a score of 4 is not necessarily 4 times as more as a score of 1. Furthermore, the interval between a score of 1 and 2 is not necessarily the same as the interval between a score of 2 and 3. Therefore, the use of these type of scores in mathematical calculation is considered statistically inappropriate F I G U R E 4 Representative photomicrographs of Masson's trichrome staining of the extraction sockets after implantation of either ABCcolla® or Bio-Oss® at 24-week after socket healing. (a) ABCcolla®; (b) 8X higher magnification of yellow square in (a); (c) Bio-Oss®; and (d) 8X higher magnification of yellow square in (c). Scale bars are 1.0 mm in (a) and (c); whereas, 125 μm in (b) and (d) (Green et al., 2014). It has been reported that it is necessary for the histopathological slides examined and graded blindly by a pathologist to minimize observational bias (Gibson-Corley et al., 2013). It is also recommended that 4-5 score levels may be an ideal range to increase detection and repeatability in a scoring system (Shackelford et al., 2002). In order to frequently produce meaningful and valid scoring data in the histopathological analysis, we have followed the key principals as mentioned above for the development of semiquantitative scoring system via histologic examination in this study. Moreover, a Bonferroni correction was applied to recheck the status of significant difference since this study contained three comparison groups.
The significant differences were not altered even after a Bonferroni Correction.

| CONCLUSIONS
Within the limits of the study, we concluded that ABCcolla ® treated defects demonstrated significantly more new bone formation and better bone bridging, but less amount of fluorescent labeling than those of the Bio-Oss ® group. However, clinical studies in humans are recommended to confirm these findings.