Osteonecrosis development by tooth extraction in zoledronate treated mice is inhibited by active vitamin D analogues, anti-inflammatory agents or antibiotics

Invasive dental treatment such as tooth extraction following treatment with strong anti-bone resorptive agents, including bisphosphonates and denosumab, reportedly promotes osteonecrosis of the jaw (ONJ) at the extraction site, but strategies to prevent ONJ remain unclear. Here we show that in mice, administration of either active vitamin D analogues, antibiotics or anti-inflammatory agents can prevent ONJ development induced by tooth extraction during treatment with the bisphosphonate zoledronate. Specifically, tooth extraction during treatment with zoledronate induced osteonecrosis in mice, but administration of either 1,25(OH)2D3 or ED71, both active vitamin D analogues, significantly antagonized osteonecrosis development, even under continuous zoledronate treatment. 1,25(OH)2D3 or ED71 administration also significantly inhibited osteocyte apoptosis induced by tooth extraction and bisphosphonate treatment. Administration of either active vitamin D analogue significantly inhibited elevation of serum inflammatory cytokine levels in mice in response to injection of lipopolysaccharide, an infection mimetic. Furthermore, administration of either anti-inflammatory or antibiotic reagents significantly blocked ONJ development following tooth extraction and zoledronate treatment. These findings suggest that administration of active vitamin D, anti-inflammatory agents or antibiotics could prevent ONJ development induced by tooth extraction in patients treated with zoledronate.

vitamin D analogues in inhibiting ONJ development, we first cultured osteoclast progenitor cells from mouse bone marrow in the presence of M-CSF and RANKL in vitro and observed both formation of multi-nuclear TRAP-positive osteoclasts and significantly upregulated expression of the osteoclastic genes Cathepsin K (Ctsk), nuclear factor of activated T cells 1 (Nfatc1) and dendritic cell transmembrane protein (Dcstamp) (Fig. 2a, b). By contrast, M-CSF and RANKL treatment inhibited expression of the inflammatory cytokines TNFα (Tnfa), IL-1β (Il1b) or IL-6 (Il6) in osteoclast progenitors (Fig. 2c). Osteoclastogenesis of M-CSF and RANKL-treated cells was significantly inhibited by zoledronate, and zoledronate-treated cells differentiated into inflammatory cytokineexpressing cells (Fig. 2a-c). Elevation of inflammatory cytokine levels reportedly promotes osteonecrosis development 12,17 . Thus, we asked whether treatment with vitamin D analogues could reverse the zoledronate effects in osteoclast progenitors. However, inhibition of multi-nuclear TRAP-positive osteoclast formation and expression of osteoclastic genes by zoledronate was not rescued in the presence of active vitamin D analogues (Fig. 2a, b). Moreover, expression of inflammatory cytokines was significantly stimulated by zoledronate in osteoclast progenitors and enhanced rather than inhibited by either active vitamin D analogue (Fig. 2c).
Porphyromonas gingivalis (Pg) is a major oral bacteria and a pathogen in the case of periodontitis 27,28 . Osteoclastogenesis induced by combined M-CSF and RANKL treatment was inhibited by a Pg lysate or zoledronate, and inhibition by either one was not rescued by treatment with an active vitamin D analogue in vitro (Fig. 3a,  b). Finally, co-treatment with zoledronate and a Pg lysate induced expression of inflammatory cytokines, and that induction was not blocked by treatment with vitamin D analogues (Fig. 3c).

Active vitamin D analogues inhibit osteocyte apoptosis induced by zoledronate administration and tooth extraction.
To define mechanisms underlying inhibition of ONJ by vitamin D analogues, we analyzed osteocyte apoptosis using TUNEL staining (Fig. 4). Wild-type mice were administered zoledronate once a week for 2 weeks before tooth extraction. Four days after tooth extraction, osteocyte apoptosis was induced by a combination of zoledronate administration and tooth extraction in mice (Fig. 4) www.nature.com/scientificreports/ however, osteocyte apoptosis as indicated by TUNEL-positivity was significantly inhibited by co-administration of either active vitamin D analogue in the presence of continuing zoledronate treatment (Fig. 4). Serum TNFα and IL-6 levels as analyzed by ELISA were significantly elevated two hours after administration of lipopolysaccharide (LPS), a bacterial infection mimetic, to wild-type mice, and such elevation was significantly inhibited by a co-administration of either of the two active vitamin D analogues (Fig. 5a, b). Thus, either analogue has anti-inflammatory activity. In all experiments, 8-week-old female C57BL/6 mice received subcutaneous injection of zoledronate (500 μg/kg) once a week. Vehicle (ethanol: 0.05 µl/kg/day), ED71 or 1,25(OH) 2 D 3 (VD3) (0.05 µg/kg/day) was intraperitoneally injected twice a week for 2 weeks before extraction and afterwards twice a week. (b) Two weeks after the first injection when mice were 10 weeks old, the right first and second molars in mandible were extracted. Six weeks after extraction, mandibles were removed, stained with HE, observed microscopically, and the percent of empty lacunae among all lacunae was evaluated. Scale bars = 100 μm (upper) and 20 μm (lower) panels. (c) Data represents mean relative proportion of empty lacunae among all lacunae in bone ± SD (each with n = 5, ***P < 0.001). Representative data are shown of at least two independent and identical experiments, each with n = 5. www.nature.com/scientificreports/

Administration of either anti-inflammatory drugs or antibiotics inhibits ONJ induced by combined zoledronate treatment and tooth extraction in vivo.
To determine whether elevated inflammatory cytokine levels function in induction of ONJ in vivo, we administered the anti-inflammatory drug loxoprofen intraperitoneally to mice 1 day before tooth extraction and then daily for 6 days (Fig. 6a). Concomitantly, zoledronate was administered once a week for 2 weeks before and 6 weeks after tooth extraction, and then, ONJ development was analyzed (Fig. 6a). ONJ development, as assessed by formation of empty lacunae after tooth extraction, was significantly blocked by loxoprofen administration even in the presence of ongoing zoledronate . Data represent mean Ctsk, Dcstamp, Nfatc1, Tnfa, Il6 or Il1b expression relative to Actb ± SD (each with n = 3, **P < 0.01; ***P < 0.001 by ANOVA). Representative data of at least two independent experiments are shown. www.nature.com/scientificreports/ treatment (Fig. 6b). Similarly, administration of meloxicam, another anti-inflammatory drug (Fig. 6a), significantly blocked ONJ development induced by combined zoledronate and tooth extraction (Fig. 6c). Bacterial infection also increases levels of inflammatory cytokines 29,30 , and infectious osteomyelitis reportedly promotes osteocyte apoptosis 12 . Others report the presence of periodontitis, even in mice maintained in specific pathogen-free (SPF) conditions 31 . Thus, we administered the antibiotic Amoxicillin hydrate (AMPC) to mice that treated with zoledronate 1 day before tooth extraction and then continued Amoxicillin administration daily for 6 weeks after extraction, while continuing zoledronate treatment (Fig. 7a). ONJ development, as with or without 0.2 μg/ml zoledronate (Zol) with or without either ED71(10 -6 M) or VD3 (10 -6 M) and/ or Porphyromonas gingivalis extract (Pg). Osteoclast formation was evaluated by TRAP staining (a) or by quantitative RT-PCR to analyze expression of indicated markers. Tnfa, Il6, and Il1b expression in indicated groups of progenitor cells, as evaluated by quantitative RT-PCR (b, c). Scale bar = 100 μm. Data represent mean indicated transcript levels relative to Actb ± SD (each with n = 3, *P < 0.05; ***P < 0.001; NS, not significant, by ANOVA). Representative data are shown of at least two independent and identical experiments each with n = 3. www.nature.com/scientificreports/ measured by formation of empty lacunae, was significantly inhibited by AMPC administration 6 weeks after tooth extraction (Fig. 7b). A condition known as severely suppressed bone turnover (SSBT) is a reported risk factor for ONJ development 32,33 . Teriparatide, a recombinant parathyroid hormone (1-34), stimulates bone turnover and can reportedly counteract or even reverse ONJ induced by anti-resorptive agents following invasive dental treatment in humans 34 . To analyze the effects of teriparatide in inhibiting ONJ development, teriparatide was administered three times a week from two weeks before and 6 weeks after tooth extraction concomitantly with zoledronate as shown in Fig. 8. However, in our model, teriparatide administration did not inhibit ONJ development induced by zoledronate and tooth extraction (Fig. 8a, b).

Scientific Reports
Overall, our results suggest that high levels of inflammatory cytokines promote ONJ development following zoledronate treatment and tooth extraction, and that ONJ is inhibited by anti-inflammatory agents, such as active vitamin D analogues, anti-inflammatory drugs or antibiotics.

Discussion
Although rare, ONJ limits activity of daily living and worsens quality of life. Thus preventive treatments are needed. ONJ develops due to an inflammatory cytokine storm induced by anti-resorptive agents, local infection, tooth extraction or any combination of these conditions 17,35,36 . We show here that ONJ can be prevented in mice by either active vitamin D analogues, anti-inflammatory drugs or antibiotics, all of which inhibit inflammatory conditions.
We previously showed that inflammatory cytokine expression is inhibited by RANKL treatment of cultured osteoclast progenitor cells but stimulated by treatment with anti-resorptive agents or bacterial infection, even in the presence of RANKL 17 . Inflammatory cytokine levels, which likely underlie ONJ development, also increase following invasive dental treatment, such as tooth extraction, or oral bacterial infection 17 . As a result, those Osteocyte apoptosis is significantly blocked by VD3 and ED71. (a, b) C57BL/6 mice were administered zoledronate (500 μg/kg) or vehicle (PBS) once a week for 2 weeks and ED71 (0.5 µg/kg) or 1,25(OH) 2 D 3 (VD3) (0.5 µg/kg) intraperitoneally twice a week and then the right first and second molars in the mandibles were extracted. Four days later, bone sections from mice with zoledronate + tooth extraction (Extraction), VD3 + tooth extraction (Extraction + VD3) or ED71 + Zoledronate + (Zoledronate + ED71) were prepared and labeled with Biotin-dUTP using terminal deoxynucleotidyl transferase (TdT), followed by Avidin-DTAF as TUNEL staining (TUNEL) to identify apoptotic cells. Nuclei were visualized by DAPI. Sections were observed under a fluorescence microscope. (c) Data showing mean percentage (%) of TUNEL-positive relative whole osteocytes in bone ± SD (each with n = 5, ***P < 0.001). Representative data are shown of at least two independent and identical experiments, each with n = 5. Bar = 100 µm. www.nature.com/scientificreports/ cytokine levels likely reach or exceed threshold levels necessary to induce osteocyte apoptosis 17 . Indeed, here we demonstrate that administration of anti-inflammatory drugs or antibiotics inhibits ONJ development induced by zoledronate and tooth extraction. We also demonstrate that active vitamin D analogues inhibit ONJ in this context as well as expression of inflammatory cytokines. Administration of anti-resorptive agents strongly inhibits bone turnover and can lead to SSBT, a risk factor for ONJ development 32,33 . Administration of a bone turnover stimulator, teriparatide, is reportedly effective in antagonizing or treating ONJ development in humans 34,37 . Teriparatide was also reportedly effective in treating infectious osteomyelitis likely by promoting bone turnover 38 . However, we found that teriparatide treatment in mice did not antagonize ONJ development (Fig. 8), suggesting that it is a cytokine storm rather than inhibition of bone turnover that underlies ONJ development.
Vitamin D has diverse effects on bone biology and anti-inflammation. Active vitamin D analogues are frequently co-administered with anti-osteoporotic drugs such as bisphosphonates, denosumab or romosozumab, and may mobilize calcium to increase bone mass and inhibit hypocalcemia potentially caused by bone massincreasing agents. Active vitamin D stimulates RANKL expression in osteoblastic cells and promotes osteoclast differentiation 39 . Moreover, active vitamin D analogues reportedly inhibit osteoclastogenesis induced by RANKL by downregulating c-Fos, a transcription factor essential for osteoclast differentiation 40 . Here, we show that inflammatory cytokine expression induced by LPS was significantly inhibited by active vitamin D analogues, and this activity is likely protective against ONJ development. However, a large proportion of people exhibit vitamin D deficiency 41 and therefore are potentially at risk for ONJ development. Thus, we conclude that for these individuals, taking active vitamin D analogues or vitamin D supplements would be effective in decreasing the risk of ONJ development. Maternal serum was collected 2 h after LPS injection. Serum TNF-α (a), IL-6 (b) were measured using ELISA. All data were expressed as means ± SD of six samples from 5 different mice (n = 5 each, **P < 0.01; ***P < 0.001). Representative data are shown of at least two independent and identical experiments, each with n = 5. www.nature.com/scientificreports/ Infectious osteomyelitis is also a risk factor for osteocyte apoptosis 12 . We show that inflammatory cytokine expression was stimulated by infectious conditions in mice. Thus, immediate tooth extraction may be considered preferable to prevent ONJ. Patients treated with strong anti-resorptive agents are occasionally advised to stop taking those agents before and after invasive dental treatment such as tooth extraction due to the risk of ONJ development 42 . However, discontinuation of anti-resorptive agents reportedly reduces bone mass and may increase risk of fragility fractures in patients 43,44 . Our data indicate that ONJ development initiated by zoledronate  , c) Loxoprofen sodium hydrate and Meloxicam (15 mg/kg/day), a non-steroidal anti-inflammatory drug, or a Vehicle (control) water was administered by intraperitoneally a day before and 6 days after surgery for a total of 1 week. Six weeks after extraction, alveolar bone sections from the right mandible were prepared and stained with HE, and the proportion of empty lacunae among all lacunae was calculated. Scale bars = 100 (upper) or 20 μm (lower) panels. Data shows mean percentage (%) of empty lacunae ± SD (n = 5 each, ***P < 0.001, by a Mann-Whitney test). Representative data are shown of at least two independent and identical experiments, each with n = 5. www.nature.com/scientificreports/ treatment and tooth extraction is inhibited by either active vitamin D analogues, anti-inflammatory agents or antibiotics, even in the presence of zoledronate treatment.
In conclusion, our data in mice suggest a way to inhibit ONJ development induced by zoledronate and tooth extraction without discontinuation of zoledronate treatment.

Materials and methods
Animal model. C57BL/6 background wild-type mice were purchased from Sankyo Labo Service (Tokyo, Japan). Mice were maintained under specific pathogen-free (SPF) conditions in animal facilities certified by the Keio University Institutional Animal Care and Use Committee, and animal protocols were approved by that committee. Mice were housed up to 5 per cage and kept on a 12 h light/dark cycle. Sterile distilled water and a standard diet (CLEA Rodent Diet CE-2, Japan) were available ad libitum. All mouse studies were performed in accordance with Institutional Guidelines on Animal Experimentation at Keio University of The Keio University Institutional Animal Care and Use Committee. Various doses of zoledronate administered to promote ONJ development have been reported in mouse studies: some utilized a single injection at doses from 0.1 to 0.54 mg/kg before tooth extraction 24,25,45,46 , and others administered zoledronate ranging from 0.1 to 0.6 mg/ kg/week before and after tooth extraction 17,26,[47][48][49][50][51][52][53][54][55] . Also, zoledronate has been administered to mice either intravenously 48,52 , intra-peritoneally or subcutaneously 47,49,50,56 . In general, higher doses of zoledronate are administered to mice than to humans 17,26,45,50,56 , as bone turnover in mice is reportedly faster than in humans 57 . Here, we performed experiments using our established ONJ model 17 . Briefly, eight-week old wild-type mice received zoledronate (500 µg/kg) or vehicle (ethanol or PBS) once a week. Two weeks later, when mice were 10 weeks old, the right first and second molars in the mandible were extracted. All mice received a mixture of ketamine (100 mg/ kg) and xylazine (10 mg/kg) by intraperitoneal injection for anesthesia. Groups of mice were administered either ED71 (0.05 µg/kg/day) or 1,25(OH) 2 D 3 (VD3) (0.05 µg/kg/day) subcutaneously twice a week for 2 weeks before and 6 weeks after extraction. Other mice were administered either Amoxicillin hydrate (50 mg/ kg/day) in drinking water or normal water for the week before extraction and continuing for six more weeks. Hydrate, an antibacterial drug, or normal drinking water was administered orally a day before and 6 weeks after surgery for a total of 6 weeks and 1 day. Six weeks after extraction, alveolar bone sections from the mandible were prepared and stained with HE, and the proportion of empty lacunae among all lacunae was calculated. Scale bars = 100 (upper) or 20 μm (lower) panels. Data shows mean percentage (%) of empty lacunae ± SD (n = 5 each, ***P < 0.001, by a Mann-Whitney test). Representative data are shown of at least two independent and identical experiments, each with n = 5. www.nature.com/scientificreports/ Other mice were administered Loxoprofen sodium hydrate or Meloxicam (15 mg/kg/day), or Vehicle (control) intraperitoneally the day before extraction and then thereafter daily for 6 days. Some mice were injected PTH (80 μg/kg) or PBS intraperitoneally twice a week for 2 weeks before extraction and afterwards twice a week, as previously described 12 . Euthanasia was performed 6 weeks after tooth extraction by cervical dislocation under anesthesia with ketamine hydrochloride (75 mg/kg). No dermatological problems were detected at the site of subcutaneous zoledronate injection (data not shown). Mice that underwent tooth extraction did not exhibit body wight loss (data not shown). All methods are carried out in accordance with the ARRIVE guidelines. Histopathological and fluorescent immunohistochemical analysis. Mouse mandibles were removed and decalcified in 10% EDTA, pH7.4, before embedding. At the time of mandible removal, mice exhibited fistulas as pinholes in the mucosa at the site of tooth extraction (Fig. S1), and the jaw bone could be touched through the fistula by a periodontal probe. Paraffin-embedded mandible sections were deparaffinized and rehydrated in a graded ethanol series. Hematoxylin and eosin (HE) staining was performed according to standard methods. For each fluorescent immunohistochemistry assay, sections were subjected to microwave treatment for 10 min in 10 mM citrate buffer solution (pH 6.0) for antigen retrieval, as described 58 . After blocking with 3% BSA in PBS for 1 h, sections were stained using a MEBSTAIN Apoptosis TUNEL Kit Direct (Medical & Biological Laboratories Co., Ltd., Nagoya, Japan). Nuclei were visualized by DAPI (#D1306 1:750; Wako Pure Chemicals Industries, Osaka, Japan). Empty lacunae located just beneath the extracted tooth were detected in H&E-stained mandible sections (Fig. S2). The proportion of empty lacunae was calculated relative to total (empty + non-empty) lacunae.

Figure 8.
Osteonecrosis development is not effectively blocked by teriparatide. C57BL/6 mice were administered zoledronate for 2 weeks. PTH (80 μg/kg) or PBS was intraperitoneally injected twice a week for 2 weeks before extraction and afterwards twice a week. One week after the first injection when mice were 10 weeks old, the right first and second molars in mandible were extracted. Six weeks after extraction, mandibles were removed, stained with HE, observed microscopically (a), and the proportion of empty versus whole lacunae was calculated (b). Scale bars = 100 (upper) or 20 μm (lower panels). Data shows the mean percentage (%) of empty versus whole lacunae ± SD (n = 5 each, NS, not significant). Representative data are shown of at least two independent and identical experiments, each with n = 5.  were used to measure mouse TNF-α and IL-6 in maternal sera according to the manufacturer's protocol and using a multiple plate analyzer (Cytation 5, BioTek Instruments, Inc., Vermont, US).

Statistical analysis.
All quantified data were expressed as means ± SD. Statistical significance of differences between groups was evaluated using Student's t test, a Mann-Whitney U test or a one-way analysis of variance (ANOVA) using statistical software (version 25; SPSS Inc., Chicago, IL, USA) (*P < 0.05; **P < 0.01; ***P < 0.001; NS, not significant, throughout the paper).