The Effect of Polydeoxyribonucleotide Extracted from Salmon Sperm on the Restoration of Bisphosphonate-Related Osteonecrosis of the Jaw

Bisphosphonates (BPs) used for treating skeletal diseases can induce bisphosphonate-related osteonecrosis of the jaw (BRONJ). Despite much effort, effective remedies are yet to be established. In the present study, we investigated the feasibility of polydeoxyribonucleotide (PDRN) extracted from salmon sperm for the treatment of BRONJ, in a BRONJ-induced rat model. Compared with BRONJ-induced samples, PDRN-treated samples exhibited lower necrotic bone percentages and increased numbers of blood vessels and attached osteoclast production. Moreover, local administration of PDRN at a high concentration (8 mg/kg) remarkably resolved the osteonecrosis. Findings from this study suggest that local administration of PDRN at a specific concentration may be considered clinically for the management of BRONJ.


Introduction
Bisphosphonates (BPs) are one of the most widely used bone resorption inhibitor drugs for the management of skeletal diseases such as osteoporosis, Paget's disease, multiple myeloma, and complications from metastatic malignancy [1]. Despite the therapeutic effect of BPs on Figure 1 shows the experimental procedure for establishing a BRONJ rat model and the H&E-stained images of the control and experimental groups. For the BRONJ model, ovariectomy was carried out in normal rats, followed by tooth extractions. Zoledronic acid (ZA; 0.6 mg/mL; Sigma-Adrich, St. Louis, MO, USA) was intraperitoneally administered twice a week for 18 weeks following tooth extraction for each rat ( Figure 1A). Figure 1B shows the gross appearances of the tooth extraction sites (maxillae) in the Sham (control) vs. BRONJ-induced rats. PDRN was used to treat the BRONJ-induced rats twice a week for 20 days at doses of 2, 4, and 8 mg/kg.

Establishment of a BRONJ Rat Model and Gross Appearance of Tooth-Extracted Sites
Our results showed that full mucosal coverage of the defects was observed in the control group. In contrast, osteonecrosis with a yellowish-brown color due to the occurrence of BRONJ was observed in the BRONJ-induced group. Treatment with PDRN in the BRONJ-induced rats seemed to reduce the severity of the osteonecrosis.

Morphological Changes in Tooth-Extracted Sites
Morphological changes in the defects of the control, BRONJ-induced, and PDRN-treated rats were investigated through H&E staining (Figure 2A). After treatment with ZA every other day for 20 days, the control group (Sham) exhibited full mucosal coverage with lamella bone (LB), woven bone (WB), connective tissue (CT) and epithelial tissue (ET) in the extraction sites. In contrast, the BRONJ-induced samples showed incomplete mucosal coverage, along with fragmented connective tissue. The control group also exhibited normal osteogenesis at the site with no inflammatory infiltration, as seen in connective tissues of the extraction sites in BRONJ-induced rats. Similar to the control, PDRN-treated samples demonstrated resolution of BRONJ. In particular, a local injection of Figure 1. Bisphosphonate-related osteonecrosis of the jaw (BRONJ) rat model is made by the intraperitoneal administration of zoledronic acid (ZA) and is restored by a local injection of polydeoxyribonucleotide (PDRN). (A) The ZA was intraperitoneally administered after tooth extraction of ovary-and uterus-removed rats for 18 weeks twice a week. (B) For investigating the effect of PDRN on BRONJ restoration, total 10 groups were determined (n = 5~6 rats per group). Black opened squares and arrows indicate tooth extracted sites.

Morphological Changes in Tooth-Extracted Sites
Morphological changes in the defects of the control, BRONJ-induced, and PDRN-treated rats were investigated through H&E staining (Figure 2A). After treatment with ZA every other day for 20 days, the control group (Sham) exhibited full mucosal coverage with lamella bone (LB), woven bone (WB), connective tissue (CT) and epithelial tissue (ET) in the extraction sites. In contrast, the BRONJ-induced samples showed incomplete mucosal coverage, along with fragmented connective tissue. The control group also exhibited normal osteogenesis at the site with no inflammatory infiltration, as seen in connective tissues of the extraction sites in BRONJ-induced rats. Similar to the control, PDRN-treated samples demonstrated resolution of BRONJ. In particular, a local injection of PDRN of more than 4 mg/kg resulted in full mucosal coverage. In contrast, administration of ZA resulted in considerable necrotic cortical bone with empty osteocytic lacunae ( Figure 2B). Moreover, local treatment with PDRN at the defect sites of BRONJ-induced rats decreased necrotic bone formation.
Mar. Drugs 2018, 16, x 4 of 14 PDRN of more than 4 mg/kg resulted in full mucosal coverage. In contrast, administration of ZA resulted in considerable necrotic cortical bone with empty osteocytic lacunae ( Figure 2B). Moreover, local treatment with PDRN at the defect sites of BRONJ-induced rats decreased necrotic bone formation.  Percentage of necrotic bone was calculated using ImageScope software. This data were expressed as the mean ± standard deviation. Statistical analysis was carried out using one-way analysis of variance (ANOVA) (n = 6 rats per group; # p < 0.05).
The results of PDRN-treated groups were compared to those of control and BRONJ-induced groups.

Osteonecrotic Formation in Tooth-Extracted Sites
To further investigate osteonecrotic bone formation in the defect sites of each sample, histological evaluation of empty osteocytic lacunae was conducted ( Figure 3A,B). The BRONJ-induced samples had a larger number of empty lacunae than the control group, due to increased osteonecrosis as a result of the ZA treatment. Compared with the BRONJ-induced samples, local treatment with PDRN decreased the number of empty lacunae. PDRN-treated samples exhibited a 2-fold lower percentage of necrotic bone than the BRONJ-induced samples.
Mar. Drugs 2018, 16, x 5 of 14 per group; # p < 0.05). The results of PDRN-treated groups were compared to those of control and BRONJ-induced groups.

Osteonecrotic Formation in Tooth-Extracted Sites
To further investigate osteonecrotic bone formation in the defect sites of each sample, histological evaluation of empty osteocytic lacunae was conducted ( Figure 3A,B). The BRONJinduced samples had a larger number of empty lacunae than the control group, due to increased osteonecrosis as a result of the ZA treatment. Compared with the BRONJ-induced samples, local treatment with PDRN decreased the number of empty lacunae. PDRN-treated samples exhibited a 2fold lower percentage of necrotic bone than the BRONJ-induced samples.  Statistical analysis was carried out using one-way analysis of variance (ANOVA) (n = 6 rats per group; # p < 0.05). The results of PDRN-treated groups were compared to those of control and BRONJ-induced groups. Figure 4 shows the size and number of blood vessels in newly formed bone in the control, BRONJ-induced, and PDRN-treated samples. Large-sized blood vessels were observed in the control and PDRN-treated samples ( Figure 4A). Moreover, the PDRN-treated samples exhibited an increased number of blood vessels compared to the control and BRONJ-induced samples. Figure 4 shows the size and number of blood vessels in newly formed bone in the control, BRONJ-induced, and PDRN-treated samples. Large-sized blood vessels were observed in the control and PDRN-treated samples ( Figure 4A). Moreover, the PDRN-treated samples exhibited an increased number of blood vessels compared to the control and BRONJ-induced samples.  Statistical analysis was carried out using one-way analysis of variance (ANOVA) (n = 6 rats per group; # p < 0.05). The results of PDRN-treated groups were compared to those of control and BRONJ-induced groups.

Number and Behavior of Osteoclasts in Tooth Extracted Sites
Tartrate-resistant acid phosphatase (TRAP) staining was conducted to evaluate the number of osteoclasts at the extraction sites ( Figure 5A,B). BRONJ-induced samples exhibited the smallest number of osteoclasts compared to the control and PDRN-treated samples. In the PDRN-treated samples, the PDRN concentration contributed to the increase of osteoclasts. Specifically, a treatment of 8 mg/kg PDRN remarkably increased osteoclast numbers. On average, the number of osteoclasts in 8 mg/kg PDRN-treated samples was 3.6, 2.9, and 1.8 times larger than the BRONJ-induced, 2 mg/kg PDRN-treated, and 4 mg/kg PDRN-treated samples, respectively. The numbers of attached and detached osteoclasts were further investigated ( Figure 6A,B). In the control, most of the osteoclasts were attached to the bone surfaces, indicating active bone remodeling. On the contrary, osteoclasts were predominantly detached from the bone surfaces in the test models. After PDRN treatment, the number of attached osteoclasts increased, while the number of detached osteoclasts decreased. In addition, even though the number of detached osteoclasts in the samples had no significance, the number of osteoclasts attached to the bone surfaces gradually increased with increasing PDRN concentrations. Furthermore, the number of attached osteoclasts in the 8 mg/kg PDRN-treated samples was similar to the control samples.

Bone Formation in Tooth Extracted Sites
Ex-vivo micro CT evaluation was carried out to confirm bone formation in the sockets after tooth extractions ( Figure 7A-C). The control (Sham) rats exhibited nearly complete healing of the tooth extraction sockets. On the other hand, incomplete bone formation was observed in the BRONJ animals. Interestingly, the PDRN-treated animals demonstrated recovery bone remodeling in the sockets.

Bone Formation in Tooth Extracted Sites
Ex-vivo micro CT evaluation was carried out to confirm bone formation in the sockets after tooth extractions ( Figure 7A-C). The control (Sham) rats exhibited nearly complete healing of the tooth extraction sockets. On the other hand, incomplete bone formation was observed in the BRONJ animals. Interestingly, the PDRN-treated animals demonstrated recovery bone remodeling in the sockets. To quantitatively examine bone formation, the bone volume (BV) and bone volume/tissue volume (BV/TV) of each group was calculated ( Figure 7D,E). Comparing the BRONJ models, local treatment with PDRN increased BV and BV/TV in the sockets. BV and BV/TV gradually increased as a function of the PDRN concentration. In particular, treatment with 8 mg/kg PDRN demonstrated To quantitatively examine bone formation, the bone volume (BV) and bone volume/tissue volume (BV/TV) of each group was calculated ( Figure 7D,E). Comparing the BRONJ models, local treatment with PDRN increased BV and BV/TV in the sockets. BV and BV/TV gradually increased as a function of the PDRN concentration. In particular, treatment with 8 mg/kg PDRN demonstrated similar BV and BV/TV to the control. This indicated that local treatment using PDRN at an appropriate concentration is effective against bone necrosis in a BRONJ-like rat model.

Discussion
The administration of BPs for osteoporosis treatment often induces BRONJ because of their effects on osteocytes, through accumulation of alveolar bone in the jaw. For example, BPs inhibit angiogenesis by obstructing the migration of neutrophils, macrophages, and osteoclast progenitor cells in tooth extraction sites [5]. BPs also increases osteonecrosis by restricting mucosal healing and providing protection against bacterial infections, due to the inhibition of granulation tissue formation [13]. Due to the increased incidence of BRONJ, many studies have attempted to understand its pathogenesis and inhibit its occurrence. However, the cause remains unclear, and successful treatments are yet to be reported. Ikebe et al. [5] suggested two possible causes for BRONJ: (1) bone turnover rate, and (2) vulnerability of the jaw to bacterial infections. Appropriate methods to control bone turnover rates and promote complete mucosal coverage of defect sites are needed.
Only rat model has been widely used as BRONJ animal model because the animal has some merits of ease of care and handling, high reproduction, completed genome mapping. Therefore, many researchers have used the animal model as BRONJ model [14]. We also selected rat model for this study. The effectiveness may be related to PDRN's ability to induce complete mucosal healing and increase osteoclastic activity, through improvement of connective tissue formation and angiogenesis.
Among inflammatory cells, macrophages play an important role in the overall phases of wound healing, host defense, promotion and resolution of inflammation, removal of apoptotic cells, and support of cell proliferation and tissue recovery after injury [15]. Hence, for successful wound healing, appropriate function of the inflammatory cells is required in each phase [16]. Compared to the control and PDRN-treated samples, numerous inflammatory cells were observed in the BRONJ samples ( Figure 2). This is likely attributed to bacterial infection due to incomplete epithelial coverage. On the other hand, the control and PDRN-treated samples exhibited a small number of inflammatory cells. This indicated that PDRN led to proliferation and remodeling of the bone, after host defense mechanisms were triggered.
Angiogenesis and normal vascularization are essential factors for improving wound healing and tissue homeostasis, respectively. Vascular endothelial growth factor (VEGF) is an inducer of angiogenesis due to its highly specific mitogen for endothelial cells [17]. However, ZA, one of most widely used BPs, reduces the mRNA and protein expressions of VEGF and decreases serum levels of VEGF and other cytokines, such as interleukin-17, involved in angiogenesis [18]. In PDRN treatment for diabetes and burns, VEGF is produced by the actions on adenosis A2 receptors [19], which may contribute to increased angiogenesis. As expected, larger sized blood vessels were observed in the PDRN-treated samples compared to the BRONJ samples, due to this PDRN effect ( Figure 3A,B).
The loss of osteoclastic activity is likely to impair osteoblastic activity, causing irregular bone turnover, remodeling, and osteonecrosis [5]. Local treatment with PDRN may restore the activity of impaired osteoclasts, resulting in the resorption of necrotic bone and formation of new bone. Our results support these suggestions. In this study, we observed the involvement of attached osteoclasts in bone remodeling in PDRN-treated samples; in particular, the samples treated with 8 mg/kg PDRN. These histological results corresponded well to the results analyzed by micro CT (Figure 4). Newly formed bone was remarkably observed in PDRN-treated samples. Among the PDRN concentrations, a local injection of 8 mg/kg PDRN resulted in superior new bone formation.
The proper balance between osteoblast and osteoclast functions is vital in bone remodeling. Previous studies reported that the bone remodeling rate of cortical bone in the jaw is several times higher than in the cortex of the iliac crest of humans and in the tibial cortex in dogs [5,[20][21][22]. This bone remodeling is thought to be related to the bone turnover rate. BPs have a very high affinity for hydroxyapatite crystals in bony architecture through ionic interactions between two phosphoryl groups and calcium ions (Ca 2+ ). Therefore, they preferentially transfer to active bone remodeling sites and accelerate bone turnover [23,24]. BPs are known to inhibit mineral dissolution by specifically targeting farnesyl diphosphate synthase (FPP synthase) in osteoclasts. The functions of FPP synthase enzymes include regulation of cytoskeletal arrangement, vesicular trafficking, and membrane ruffling engagement in bone resorption. BPs inhibit FPP synthase activity, thereby preventing the activities of osteoclasts that destroy bones.

In Vivo Animal Study
New Zealand White rats (about 4 kg, n = 40) were used for evaluating the effect of PDRN (Pharma Research Products, Seongnam, Korea) on BRONJ. To induce osteoporosis, the ovary and uterus of each rat were removed by peritoneotomy, and the rats were reared for 2 months. Then, 3-aminopropionitrile fumarate salt solution (15 g/3.5 L; Sigma-Aldrich, St. Louis, MO, USA) was dissolved in the drinking water. After 15 days of this treatment, the first and second teeth in the left maxillary bone were extracted using a dental explorer. After the extraction, ZA (0.6 mL) was intraperitoneally administered to each BRONJ-induced rat for 18 weeks, twice a week. Additionally, dexamethasone (5 mg/kg) was administrated for the last two weeks, twice a week, because the drug accelerates the occurrence of BRONJ. Afterwards, three kinds of specific concentrations of PDRN (2 mg/kg, 4 mg/kg and 8 mg/kg; Rejuvenex ® , chain lengths ranging from 50 bp to 2000 bp, PHARMARESEARCH PRODUCTS, Seongnam, Korea) were directly injected to soft tissues near to the defected sites. Six sections on each sample were prepared to investigate the effect of PDRN on BRONJ therapy both macroscopically and histologically.

In Vivo Animal Study Approval
The animal study was approved by the Institutional Animal Care and Use Committee (IACUC) of Kyung Hee University (KHUASP(SE)-16-063).

Histological Evaluations
The harvested specimens were fixed in 4% paraformaldehyde at 4 • C for 48 h, and then decalcified in a 10% EDTA solution at room temperature (RT), for 4 to 6 weeks. After dehydration with gradient ethanol, the specimens were degreased in xylene and then embedded in paraffin. Sections (5 µm thick) were analyzed histochemically with hematoxylin-eosin (H&E) and Masson's trichrome (MT) staining according to the manufacturer's instructions. Osteoclasts were identified by tartrate resistant acid phosphatase (TRAP) staining using an acid phosphatase leukocyte kit (Sigma-Adrich, St. Louis, MO, USA), following the protocols recommended by the manufacturer. Five uniformly spaced H&E-stained slides of the extraction sites from each specimen were scanned digitally with the ScanScope slide scanner (Nikon Instruments Inc.; Melville, NY, USA). An area of interest (~4.0 mm 2 ) located within 2.0 mm of the second molar was selected. The necrotic bone, defined as any region that contained three or more empty lacunae per 1000 µm 2 , was marked. The number of empty lacunae was determined by calculating its number per 1 mm 2 . The total areas of necrotic bone were analyzed for each slide, using the ImageScope software. An average value of the necrotic bone area was calculated using five slides per rat. The percentage of necrotic bone area over total bone area was also calculated. The number and surface of osteoclasts were determined by examining TRAP activity as TRAP-positive multinucleated cells. They were calculated as osteoclast number per bone surface perimeter and as a percentage of osteoclast perimeter to bone surface perimeter, respectively.

Statistical Analysis
All quantitative data were expressed as the mean ± standard deviation. One-way analysis of variance (ANOVA) using SPSS software (SPSS Inc., Chicago, IL, USA) was used for statistical analysis. A value of # p < 0.05 was considered statistically significant.

Conclusions
This study demonstrated the effect of a local PDRN injection on BRONJ treatment by modifying the occurrence of angiogenesis, the completion of full mucosal coverage, and the migration and activity of osteoclasts in defect sockets. Furthermore, it was found that the therapeutic effect of PDRN is concentration-dependent. Consequently, a local injection of PDRN may potentially be an effective therapeutic modality for BRONJ treatment, because it has been clinically shown to be effective in soft tissue regeneration.