Autologous platelet-rich fibrin stimulates canine periodontal regeneration

Platelet-rich fibrin (PRF) provides a scaffold for cell migration and growth factors for promoting wound healing and tissue regeneration. Here, we report using PRF in periodontal healing after open flap debridement (OFD) in canine periodontitis. A split-mouth design was performed in twenty dogs. Forty periodontitis surgical sites were randomly categorized into 2 groups; OFD alone and OFD with PRF treatment. Clinical parameters of periodontal pocket depth, gingival index, and the cemento-enamel junction-alveolar bone levels/root length ratio were improved in the OFD + PRF group. The OFD + PRF group also demonstrated a dramatically decreased inflammatory score compared with the OFD group. Collagen accumulation was improved in the OFD + PRF group at later time points compared with baseline. PRF application also significantly reduced inflammatory cytokine expression (TNFA and IL1B), and promoted the expression of collagen production-related genes (COL1A1, COL3A1, and TIMP1) and growth factors (PDGFB, TGFB1, and VEGFA). These findings suggest that PRF combined with OFD provides a new strategy to enhance the overall improvement of canine periodontitis treatment outcomes, especially in terms of inflammation and soft tissue healing. Therefore, PRF use in treating periodontitis could play an important role as a regenerative material to improve canine periodontitis treatment.

Periodontitis is an inflammatory disease of the periodontium comprising the gingiva, alveolar bone, periodontal ligament, and cementum 1 . The pathogenesis of periodontitis involves the interaction between host defense mechanisms and the dental biofilm. Triggered by microorganisms, periodontitis is caused by the chronic immune response, leading to inflammatory cytokine production that results in the destruction of the periodontium and subsequent manifestations of periodontitis 2,3 . In addition to its local impact on the periodontium and oral hygiene, periodontitis has a significant systemic effect. Previous reports have demonstrated that periodontitis is associated with obesity, diabetes, low birth weight, osteoarthritis, and cardiovascular disease 4,5 . In periodontitis treatment, the aim of the first phase is to eliminate the infectious source by removing plaque and calculus. Concomitantly, appropriate oral hygiene behavior is required to maintain a healthy periodontal status. In patients with periodontal destruction, periodontal regenerative treatments are commonly performed to restore the function of the periodontium 6,7 .
Various clinical modalities have been introduced as periodontal regenerative treatments. These treatments have focused on down-regulating inflammation, stimulating periodontal regeneration, and achieving optimal oral health 8 . Autogenous bone grafts are considered the best method for bony defect regeneration; however, this method has several disadvantages. Postoperative donor site morbidity, surgical complications, severe pain, and high cost are major limitations 9 . These disadvantages led to the development of novel tissue engineering therapies as autogenous graft substitutes. Among these therapies, bone substitutes, biomaterial scaffolds, and growth factors and have been evaluated in clinical studies over the past decade 10 . Several bone substitutes have been widely investigated. However, some limitations have been reported. The potential risk of cross-infection and immunological responses from the recipient are also important reasons for using natural transplants (allografts and xenografts). When using synthetic materials (alloplasts), limited periodontal regeneration has been Results PRF improved clinical periodontal parameters. The OFD and OFD + PRF groups exhibited a slightly higher PI compared with the control group. However, the difference between the groups was not significant (Fig. 1A). In addition, the MI results were not significantly different between these groups at the evaluated time points (Fig. 1B). The periodontal pockets were markedly deeper in the OFD and OFD + PRF groups at all-time points compared with the control oral healthy dogs (Fig. 1C). A trend of decreased periodontal pocket depth was observed in the OFD and OFD + PRF groups. However, the periodontal pocket depth in the OFD group was not significantly different between time points. In contrast, the OFD + PRF group demonstrated a significantly decreased periodontal pocket depth at day 14, day 21, and day 56 compared with day 7. The periodontal pocket depth was significantly lower in the OFD + PRF group at day 21 and 56 compared with the OFD group. A trend of decreased GI was noted in all groups (Fig. 1D). The OFD group exhibited a significantly decreased GI at day 21 and 56 compared with day 7. In contrast, the GI in the OFD + PRF group was significantly decreased beginning at day 14. Further, a significantly decreased GI was observed in the OFD + PRF group compared with the OFD group at day 14; however, there were no significant differences between these groups at the other time points evaluated. Representative images of the clinical gingival status in the OFD and OFD + PRF groups are illustrated in Fig. 1E-H. PRF did not affect alveolar bone gain. Representative radiographic images of each group are presented in Fig. 2A-F. A decreased CEJ-BL/root length ratio was observed in the OFD and OFD + PRF groups at day 21 and 56 day compared with baseline ( Fig. 2G). This ratio was slightly lower in the OFD + PRF group than that in the OFD group. However, these ratios were not significantly different.
PRF reduced the inflammatory reaction and ameliorated fibrosis. The representative baseline histological images indicated that the OFD and OFD + PRF groups had an abundant inflammatory cell infiltration, mainly plasma cells and lymphocytes, suggesting chronic periodontitis (Fig. 3). The inflammatory score was significantly higher in the OFD and OFD + PRF groups than that of the control at baseline and day 14 ( Fig. 4). At day 14, inflammatory cells were still found in the OFD + PRF group (Fig. 3H,I). However, the OFD + PRF group inflammatory score at day 14 was significantly lower compared with the OFD group (Fig. 4). Collagen accumulation was observed using Masson's Trichrome staining. The OFD and OFD + PRF groups demonstrated loose and randomly arranged connective tissue at baseline (Fig. 5). However, at day 14, the OFD + PRF group exhibited dense and well-organized connective tissue with neovascularization.

PRF regulated gene expression related to inflammation and healing in periodontal tissues.
Two major pro-inflammatory cytokines, TNFA and IL1B, were chosen to evaluate the effect of PRF treatment on inflammation. The OFD + PRF group exhibited a significant decrease in TNFA and IL1B expression at day 7 and 14 compared with baseline (Fig. 6A,B). At day 14, the TNFA and IL1B mRNA levels were significantly lower in the OFD + PRF group than those in the OFD group.
COL1A1 mRNA expression was significantly lower in the OFD and OFD + PRF groups compared with the control at baseline (Fig. 6C). At day 7 and 14, the COL1A1 levels were slightly lower in the OFD and OFD + PRF groups than that of control; however, the difference was not significant. The OFD group exhibited lower COL3A1 and TIMP1 mRNA levels compared with the control and OFD + PRF groups at day 7 (Fig. 6D,E). In contrast, a significant increase in COL3A1 expression was observed at day 14 in the OFD and OFD + PRF groups compared with the control group (Fig. 6D). The OFD + PRF group demonstrated markedly increased TIMP1 expression at day 14 compared with the control and OFD groups (Fig. 6E). A similar expression pattern was observed for PDGFB mRNA expression (Fig. 6F). Further, the OFD + PRF group exhibited significantly higher TGFB1 mRNA expression than that of the control and OFD groups at day 7 and 14 ( Fig. 6G). Moreover, TGFB1 mRNA levels were markedly increased in the OFD and OFD + PRF groups at day 7 and 14 compared with baseline. Lastly, the VEGFA mRNA expression was evaluated (Fig. 6H). The OFD + PRF group demonstrated a significant increase PRF contained TGF-β1 and VEGF-A. The protein expression of TGF-β1 and VEGF-A in PRF was examined using ELISA. As shown in Fig. 7, TGF-β1 protein concentration (170.50 ± 15.24 mg/ml) in PRF was higher than the VEGF-A concentration (88.08 ± 10.32 mg/ml).

Discussion
The present study investigated the effect of PRF membranes in treating periodontitis as evaluated by various clinical, radiological, and histological parameters; and gene expression. We divided the experimental animals into 3 groups; sham operation in dogs with a healthy periodontium, OFD in dogs with periodontitis, and OFD + PRF treatment in dogs with periodontitis. In the control group, an elevated GI score was observed at day 7 due to the normal tissue response to the OFD procedure. The GI score in the control group returned to 0 at day 14, demonstrating that healthy gingiva heals rapidly. PRF treatment resulted in a decreased periodontal pocket depth and GI score compared with dogs treated with OFD alone. The present results corresponded with human studies that demonstrated that the PRF-treated group had a significantly reduced GI score and periodontal pocket depth compared with OFD only in intrabony defects [25][26][27][28] . Furthermore, the combination of an anorganic bovine bone mineral (ABBM) and PRF resulted in a reduced GI score and periodontal pocket depth compared with those treated with ABBM alone 29 . The present study also found that the PI between the OFD and OFD + PRF groups was not significantly different. Similarly, a previous report illustrated that PRF treatment with ABBM did not   PRF application resulted in a significant increase in radiographic density and bone fill in the PRF-treated group evaluated at 12 months post-operatively in humans 32 . Moreover, significantly higher new bone area percentages were observed in rat periodontal fenestration defects treated with PRF mixed with periodontal ligament stem cells at 2 months post-operatively as determined by histological analysis 33 . In contrast, the present study demonstrated that the alveolar bone gain was not significantly different when the PRF was applied after OFD. A previous study demonstrated that PRF treatment improved probing pocket depth, relative attachment level, and radiographic bone fill comparable to those treated with autologous bone grafts 34 . However, it was noted that the autologous bone grafts significantly promoted bone fill compared with PRF 34 . Correspondingly, a meta-analysis of PRF application as an adjuvant to open flap debridement demonstrates that PRF application increases the bone fill in intrabony defects 24 . However, the supportive evidence is not definitively conclusive 24 . Hence, additional clinical trials are required to conclude the effect of PRF on bone regeneration.
Various approaches for periodontal regeneration have been studied in several clinical investigations. Many studies observed positive results in hard tissue regeneration when using growth factors. A combination of rhBMP-2 and ACS was applied at the osteonecrotic maxillary/mandibular lesions induced by a bisphosphonate and in jaw reconstruction after tumor resection in human clinical cases. Radiographic evaluation demonstrated new bone formation by 3 or 4 months postoperatively 35,36 . Similar results were reported in a non-human primate animal model of distraction osteogenesis. This combination induced new bone formation at 3 months post-operation based on histological examination 37 . Another promising approach is using a titanium mesh. A report indicated that titanium mesh filled with autogenous bone and deproteinized anorganic bovine bone utilized as a barrier membrane enhanced alveolar ridge reconstruction 38 . Compared with our results, other approaches seem more effective compared with PRF alveolar bone regeneration. However, well-designed clinical trials and meta-analysis are required to confirm these findings.
To investigate the effect of PRF histologically, the inflammatory reaction and fibrosis score were assessed at baseline and day 14. We found that the OFD + PRF group had a significant reduction in the inflammatory reaction score. There are few studies concerning the effect of PRF on inflammation. Moreover, these studies focused on clinical and radiographic parameters to determine the efficacy of PRF in periodontal treatment, despite that histological assessment is one of the most accurate evaluation methods 26 . We hypothesized that the reduced inflammatory reaction score in the OFD + PRF group resulted from the various anti-inflammatory cytokines that www.nature.com/scientificreports www.nature.com/scientificreports/ are embedded in the PRF fibrin meshwork. Our results indicated that the collagen accumulation was not significantly different between the OFD and OFD + PRF groups. This observation might be attributed to the inadequate sensitivity of our histological evaluation. Other methods such as immunohistochemistry, wound healing assay, or fibroblast and collagen gene detection might generate different results.
A previous study reported that PRF prepared from beagle dogs contained TGF-β1 at a concentration of approximately 64 ng/ml 39 . However, our study demonstrated a higher TGF-β1 concentration in PRF. This difference might be due to the different preparation procedures and variation in experimental animals. Further, it has been shown that plasma and platelet-rich plasma contained lower concentrations of TGF-β1 and VEGF-A than the present study 40,41 . Therefore, this could imply that PRF contains a higher growth factor concentration compared with plasma and platelet-rich plasma due to its effective growth factor entrapment ability.
To understand the mechanism of PRF in periodontal healing, we evaluated the expression of various genes associated with inflammation and periodontal wound healing. Our results revealed that the OFD + PRF group had a significant upregulation in TGFB1, PDGFB, VEGFA, and COL3A1 expression, while the TNFA and IL1B mRNA levels were downregulated. TGF-β1, PDGF-B, and VEGF are key growth factors found in PRF. TGF-β1  www.nature.com/scientificreports www.nature.com/scientificreports/ significantly participates in wound healing processes, including immune cell modulation, stimulating osteoblast proliferation, and promoting collagen synthesis 42 . PDGF-B is a member of the PDGF family and is a powerful chemoattractant, angiogenesis mediator, and potent activator of mesenchymal lineage cell migration and proliferation. PDGF-B mainly functions in the inflammatory and proliferative phases of wound healing 43 . VEGF acts as an endothelial mitogen, chemotactic agent, is angiogenic, and induces epithelialization and collagen deposition. VEGF predominately functions in the inflammatory and proliferative process of wound healing 44 . Therefore, high expression of these cytokines would increase the entire wound healing process. PRF has been shown to increase angiogenesis in guided-bone regeneration of cranial defects in rabbits 45 . The VEGF expression observed via immunostaining was higher in the group receiving xenogenic bone combined with PRF compared with the xenogenic bone alone group 45 .
TIMP-1, COL1A1, and COL3A1 mainly participate in the proliferative and remodeling phases of periodontal wound healing. TIMP functions as a matrix metalloproteinase (MMP) inhibitor, and plays an important role in the remodeling phase by inhibiting extracellular matrix and collagen breakdown 46 . Collagen is a basic component of the periodontium. Type III collagen is predominately synthesized during the initial phase of wound healing and is then gradually replaced with type I collagen 2-3 weeks after the initiation of wound healing 47 . We found that TIMP1 and COL3A1 mRNA expression was significantly higher in the OFD + PRF group. Similarly, it has been shown that PRF combined with periodontal ligament stem cells and jaw bone mesenchymal stem cell sheets exhibited higher COL1A1 and COL3A1 expression compared with control 48 . Based on their function and expression in our study, we concluded that their high expression in the OFD + PRF group in conjunction with the growth factors mentioned above accelerated periodontal healing. The expression of COL1A1 and COL3A1 mRNA in the OFD + PRF group correlates with our histological observation of dense collagen fiber accumulation at day 14.
TNF-α and IL-1β are key pro-inflammatory cytokines that are involved in the pathogenesis of periodontitis and decrease wound healing and regeneration 49 . High production of these cytokines stimulates other inflammatory mediators, tissue destruction by MMP induction, and bone resorption by stimulating osteoclast activity 50,51 . Therefore, high expression would result in increased inflammation and tissue destruction. Our results revealed that the OFD group presented higher expression of these cytokines, which was related to the higher GI and inflammatory reaction score compared with OFD + PRF group. Deeper periodontal pockets might also be associated with our cytokine expression results.
Based on our overall results, PRF potentiates wound healing and diminishes the inflammatory response. The upregulation of TGFB1, PDGFB, VEGFA, TIMP1, COL1A1, and COL3A1 could act as chemo-attractants for other immune cells and fibroblasts, inhibiting extracellular matrix degradation, promoting angiogenesis, inducing cell proliferation, and stimulating collagen and extracellular matrix synthesis. Likewise, stimulating inflammatory mediator production, MMP expression, matrix producing cell apoptosis, and osteoclast activity might also be diminished due to the downregulated TNFA and IL1B expression. These PRF effects can explain the increased periodontal attachment gain, decreased gingivitis, and decreased histological inflammatory score found in our study.
The PRF preparation procedure used in our study differed from that of most studies. First, the present study collected 4 mL of autologous venous blood. Thus, the PRF characteristics in the present study may differ from other studies using 10 mL of blood. The 4 mL volume was chosen due to the limited periodontal defect size in the canine periodontitis model. Our preliminary study demonstrated that the application of PRF collected from 10 mL autologous blood volume in the defect resulted in gingival flap dehiscence at day 1 or 3 post-operatively. This gingival flap dehiscence could affect the healing processes and compromise the interpretation of our results. In addition, it has been shown that platelet and leukocyte distribution in the PRF membrane is different at specific regions 52,53 . Therefore, PRF prepared from 10 mL blood and cutting a relatively large portion out to fit in the canine defect may not contain the same components as did our PRF, causing less effects, in the present study. Further, PRF preparation from 5 mL blood has previously been reported and utilized in both human and dogs [54][55][56] . Therefore, we decided to use 4 ml of blood, which resulted in a relatively appropriate PRF size that fit the defect size and gingival flap in our canine model. Second, the centrifugation method was modified from the conventional method. In the present study, reduced relative centrifugal forces (RCF) and time was used. Reducing the RCF results in increased cell populations in the PRF collected from the top one-third layer, whereas the high centrifugation forces used in the previous platelet-rich fibrin preparation protocol shift the cell populations to the bottom of the tubes. In addition, a lower centrifugation time reduces cell pull-down by centrifugation forces, which increases the cell populations in the platelet-rich fibrin matrix 57 . The decreased RCF and time resulted in significant increases in platelet cell numbers, monocyte/macrophage behavior, and growth factor release compared with other preparation methods 41,57,58 . Moreover, in terms of tissue regeneration, human gingival fibroblasts demonstrated significantly increased migration and proliferation when cultured with PFR acquired using decreased centrifugation speed and time 57 . Other studies additionally revealed that reduced RCF contributed to increased leukocytes and platelets gain in PRF matrices [59][60][61] . Third, the centrifugation machine and the blood collection tubes are different between ours and other studies. As reported previously, different apparatuses utilized in PRF preparation may influence the PRF quality 62 . These differences would affect the RCF at the PRF clot (referred to as RCF-clot), the RCF at the shortest distance from the rotor (referred to as PRF-min), and the RCF at the largest distance from the rotor (referred to as PRF-max) 62 . As stated in the materials and methods section, 1,300 rpm for 8 min (RCF-clot = 164 g) was used in our study. This RCF-clot value was slightly different from those in another study; however, it was similar to the low-speed concept PRF that is produced using a low centrifugation speed (approximately 200 g RCF-max and 130 g RCF-clot) 57,58,63 . Another consideration is that the tube utilized for PRF preparation is crucial 58 . In the present study, a glass tube was used. Generally, platelets can interact with a glass surface, resulting in coagulation activation during centrifugation. This interaction leads to the formation of a solid PRF matrix composed of a fibrin network entrapping platelets, leukocytes, plasma proteins, and growth Scientific RepoRtS | (2020) 10:1850 | https://doi.org/10.1038/s41598-020-58732-x www.nature.com/scientificreports www.nature.com/scientificreports/ factors 64 . Moreover, a PRF clot made from a glass tube forms and retracts from the tube wall faster compared with a plastic tube 65 . Due to the differences between our and other studies, we measured the concentration of two key cytokines, TGF-β1 and VEGF-A to ensure that an adequate amount of these growth factors was present in each PRF membrane. We found that the mean concentration of TGF-β1 was slightly higher compared with a previous report 39 . Therefore, we assume that our PRF contained an adequate amount of growth factors.
The present study was conducted in a canine periodontitis model, which has some limitations. Although canine periodontal anatomy and its physiological mechanisms are well described, this limitation is of concern. Hence, research methodology and a periodontal defect model in dogs should be further developed to use as a predictable translatable animal model prior to human clinical trials 66 . To confirm the observations in our study, long-term and large-scale studies should be performed. Moreover, other cytokines should be investigated, especially those involving in the initiation and progression of periodontal diseases. This would help to elucidate the potential mechanism of PRF in periodontal applications. However, within the limitation of the present study, we found that PRF improves clinical outcomes, accelerates wound healing, and reduces the inflammatory response. PRF could be a novel alternative modality for periodontitis management in human and dogs. However, additional randomized control trials are required to ensure the positive effect of PRF in clinical application.

Materials and Methods
Experimental animals. Experimental animals were obtained from the Small Animal Teaching Hospital at Chulalongkorn University with the following inclusion criteria: 1) Mesocephalic dogs aged between 8 months to 5 years, 2) Healthy dogs based on physical examinations and laboratory tests, 3) Maxillary 4th premolars and mandibular 1st molars with periodontal pocket depth (PPD) between 3-5 mm, and 4) No periodontal complications, e.g. fractured teeth 67 . Dogs with a history of receiving anti-inflammatory medicine within 30 d, having systemic/metabolic/immunosuppressive illness, insufficient platelet count (<20,000/mm 3 ), teeth with alveolar bone loss over 75%, or a mobility index grade 2 or higher were excluded from the study 68 . Dogs with healthy oral status served as the controls (Control group, n = 5). Forty periodontitis sites were identified based on the inclusion criteria and divided using a split mouth clinical design into 2 groups; open-flap debridement (OFD group, n = 20) and OFD with PRF treatment (OFD + PRF group, n = 20). Informed consent was obtained from the owners and the experiments were conducted in accordance with the guidelines for animal welfare of experimental animals and approved by the Chulalongkorn University Animal Care and Use Committee, Pathumwan, Bangkok, Thailand (#1831017) Surgical procedure. Each surgical procedure was performed under general anesthesia using intramuscular injection with 0.02 mg/kg acepromazine (2 mg/ml, Vetranquil; CEVA Sante Animal, France) and 0.3 mg/kg morphine (10 mg/ml) as a premedication. General anesthesia was induced with 2-4 mg/kg propofol (10 mg/ml, Lipuro 2%; Braun, Germany) and inhalation with 2% isofurane was used for anesthetic maintenance. Local anesthesia of the maxilla and mandible was obtained with 0.5% bupivacaine. Cefazolin (22 mg/kg) was given as a prophylaxic antibiotic 69 . Full mouth dental scaling and polishing were performed. The studied sites underwent OFD via the oft-modified Widman flap technique (MWF) (Fig. 8A-E). The procedure comprised an internal bevel incision, mucoperiosteal flap reflection, intrasulcular incision, and horizontal incision along the alveolar crest. Root planning was performed using ultrasonic instruments and Gracey curettes (Hu-Friedy Mfg Co. Inc., Platelet-rich fibrin preparation and administration. The PRF membranes were prepared based on a previous protocol with slight modification 57 . Four mL of autologous blood was collected from the jugular vein and kept in a sterile 10 ml glass tube (16 mmx100 mm, Becton, Dickinson and Company, Franklin Lakes, NJ, USA). PRF membranes were produced using 1,300 rpm for 8 min (RCF-clot = 164 g, RCF-max = 303 g) using a Kubota4000 centrifugation machine (Japan) at a 45° rotor angulation with a radius of 87 mm at the clot and 160 mm at the max. The PRF clot formed between the acellular plasma (upper layer) and red blood cell base (bottom layer) was harvested 1 mm below the interface with the red blood cell layer to maximize platelet quantity 52,71 (Fig. 8F,G). The PRF membrane was compressed in a gauze and then cut into 4 × 5 mm pieces with a scalpel blade, and placed in the surgical sites (Fig. 8H). The serum exudate collected from the compression was used for graft material hydration, surgical site rinse, and autologous graft storage. In the OFD + PRF group, the PRF membrane was positioned over the denuded root surface just below the cemento-enamel junction (CEJ) and mucoperiosteal flap closure was performed in the same manner as for the OFD group 25,72 (Fig. 2D). In the control group, a sham operation was performed in dogs with a healthy periodontium.
Clinical evaluation. The plaque index (PI), gingival index (GI), mobility index (MI), and periodontal pocket depth were determined as described by Löe 73 and Laster, et al. 74 . For determining the PI, iC plaque ® (iM3, Australia) was used to stain the accumulated plaque as a pink layer on the tooth surface. The GI was evaluated based on the presence of gingival inflammation on the mesial, distal, buccal, and lingual surfaces. The periodontal pocket depth was recorded as the mean measurement of 6 areas around each surgical site (mesio-buccal, mid-buccal, disto-buccal, mesio-lingual, mid-lingual, and disto-lingual) using a William's probe. The MI was based on tooth mobility using an explorer. The criteria for each parameter are shown in Table 1.
Intra-oral radiographic evaluation. Intra-oral radiographs were taken with a CR7 Vet Dental X-ray unit (iM3, Australia). Alveolar bone loss was assessed by direct measurements of the distance between the CEJ and the alveolar bone level (BL). The distance was measured at three points; the mesial, middle, and distal aspects of each tooth. The root length was also measured for calculating the CEJ-BL/root length ratio 75,76 . Histological analysis. Tissue samples measuring approximately 3 × 5 mm were collected from the middle buccal area via the MWF technique to orient and contain the pocket epithelium, oral epithelium, and connective tissue in the same section 77 . The samples were fixed with 10% neutral buffered formalin, processed through a graded series of ethanol, and embedded in paraffin. The sections were obtained at a 5 μm thickness and stained with hematoxylin-eosin (H&E) and Masson's trichrome. The inflammatory reaction and fibrosis were evaluated. For the inflammation scoring, the inflammatory cells (neutrophils, lymphocytes, plasma cells, and eosinophils) were identified and individually counted as follows: Not present (0), Mild (1), Moderate (2), and Severe (3). The sum of the total scores of each classification were divided into three grades of inflammatory reaction; mild inflammation (score 0-3), moderate inflammation (score 4-6), and sever inflammation (score ≥ 7) 78 .
Enzyme-linked immunosorbent assay. The PRF membranes were cut into small pieces with a scalpel blade and homogenized using disposable homogenizers (BioMasher II, Nippi, Inc., Tokyo, Japan). After centrifuging at 3000 rpm for 10 min at 4 °C, the supernatant was collected and the concentration of TGF-β1 and VEGF-A  Polymerase chain reaction. Gingival tissue samples were harvested from the mid-buccal area. Total cellular RNA was extracted using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) and treated with RNase-Free DNase I (Qiagen, Netherland) to remove any genomic DNA. The RNA integrity and amount were evaluated using a Nanodrop spectrophotometer (Thermo Scientific, USA). Complementary DNA transcription was performed from one microgram of total RNA using a reverse transcriptase ImPromII kit (Promega, Madison, WI, USA). Real-time polymerase chain reaction was performed using a FastStart Essential DNA Green Master kit (Roche Diagnostic, USA) on a MiniOpticon real-time PCR system (Bio-Rad, USA). The data were analyzed using the 2 −ΔΔCt method. Target gene expression value was normalized to ACTB expression values and then normalized to the expression in the periodontal healthy dog control group. The primer sequences used are shown in Table 2.
Statistical analyses. The clinical, intra-oral radiographic, histological, and cytokine expression data are expressed as mean ± S.D. The intergroup and intragroup comparisons of specific parameters were evaluated using two-way ANOVA followed by the Bonferroni-type multiple t-test. The data were quantitatively analyzed using SPSS version 22 for Windows program (Version 22, IBM, US). A significant difference was considered when P < 0.05.

Gene
Oligonucleotide sequences Annealing temperature (°C) Amplified product size (bp) Accession no.