Dietary nitrate improves jaw bone remodelling in zoledronate‐treated mice

Abstract Bisphosphonate‐related osteonecrosis of the jaw (BRONJ) is a serious complication that occurs in patients with osteoporosis or metastatic bone cancer treated with bisphosphonate. There is still no effective treatment and prevention strategy for BRONJ. Inorganic nitrate, which is abundant in green vegetables, has been reported to be protective in multiple diseases. To investigate the effects of dietary nitrate on BRONJ‐like lesions in mice, we utilized a well‐established mouse BRONJ model, in which tooth extraction was performed. Specifically, 4 mM sodium nitrate was administered in advance through drinking water to assess the short‐ and long‐term effects on BRONJ. Zoledronate injection could induce severe healing inhibition of the tooth extraction socket, while addition of pretreating dietary nitrate could alleviate the inhibition by reducing monocyte necrosis and inflammatory cytokines production. Mechanistically, nitrate intake increased plasma nitric oxide levels, which attenuated necroptosis of monocytes by downregulating lipid and lipid‐like molecule metabolism via a RIPK3 dependent pathway. Our findings revealed that dietary nitrate could inhibit monocyte necroptosis in BRONJ, regulate the bone immune microenvironment and promote bone remodelling after injury. This study contributes to the understanding of the immunopathogenesis of zoledronate and supports the feasibility of dietary nitrate for the clinical prevention of BRONJ.


| INTRODUCTION
Bisphosphonates-related osteonecrosis of the jaw (BRONJ) was firstly described in medical literature in 2003. 1 BRONJ, due to bone resorption inhibition, is reported as a side-effect of bisphosphonates which are used widely for treating osteoporosis, bone metastases in cancer and some other skeletal-related events. BRONJ is still one of the most common adverse events in stomatology, 2 with multiple clinical reports indicating that dental procedures such as routine tooth extraction markedly increased the incidence of BRONJ. Although the osteoclast dysfunction theory and systemic inflammation may be partially involved, 2,3 the pathogenesis of BRONJ is not fully understood. And there is no effective treatment for BRONJ for now.
Nitric oxide (NO) can quickly diffuse through the membrane, serving as an important signalling molecule in biological events. 4 Low concentrations of NO could regulate the polarization of macrophages and reduce the production of proinflammatory factors. 5 It has been reported that NO is involved in the differentiation and activation of osteoclasts and osteoblasts, promoting proliferation at low concentrations and inhibiting proliferation at high concentrations. 6 Inducible nitric oxide synthase (iNOS) plays an important role in the development and survival of osteoclasts in vivo. Only dysfunctional osteoclasts were induced in vitro by monocytes lacking iNOS. These results suggest that NO is essential for the differentiation and activation of osteoclasts. 6,7 However, the injection of bisphosphonates reduced iNOS expression, limiting the production of endogenous NO 8,9 and might aggravate the pathological changes. Therefore, the supplement of exogenous NO is essential. 10 Sodium nitrate is omnipresent in our daily diet, especially in green vegetables, such as spinach, lettuce and beetroot. Previous studies have shown that sodium nitrate had protective effects in systemic diseases which were achieved by regulating the immune response and promoting and maintaining cell growth after injury. [11][12][13] The molecular mechanisms behind it are complex and not fully understood. In the oral and gastrointestinal tract microenvironments, NO could be generated by commensal nitrate-reducing bacteria through the nitrate-nitrite-NO axis. 14 In contrast to NO synthases, the nitratenitrite-NO pathway is independent of L-arginine and is unaffected by bisphosphonates.
In this study, the high-prevalence of BRONJ in the animal model was used and modified as previously described. 15  Briefly, before tooth extraction, C57BL/6 J mice were intravenous injection of 600 μg/Kg zoledronate (Zometa, designated as Z; Novartis, Stein, Switzerland) as previously described. 15,16 No steroids were administered. The animals were grouped as follows: (1) the Ctl group, in which only the right maxillary first molars were removed; (2) Z group, which received zoledronate injection and tooth extraction; (3) Z + Nit group, which received zolendronate and 4 mM of sodium nitrate. The dose of 4 mM of sodium nitrate was repeated by our team for many times and was considered as the optimal concentration. 18 Sodium nitrate 4 mM was added to the drinking water of mice once a week before zoledronate injection and continued until euthanasia. The maxillary first molar was extracted after the zoledronate injection. After tooth extraction, a stereomicroscope was used to confirm whether dental roots of the extracted molar were fractured. Mice with fractured roots were excluded in this study since the fractured roots affected wound healing following tooth extraction.

| Flow cytometric analyses of tissue and peripheral blood
The tissue around the tooth sockets and peripheral blood were collected during the inflammatory phase of tissue 1 week after tooth extraction. Peripheral blood was processed by lysing the red blood cells. The tissue was digestion, centrifugation and filtration. Finally, the single cell suspension was stained for immune cell analysis. The cells were incubated with purified anti-mouse CD45, anti-mouse F4/80, anti-mouse CD11b, anti-mouse CD11c, anti-mouse Ly6C, anti-mouse Ly6G and anti-mouse CD3 (Biolegend) at 4 C for 30 min.
Stained cells were tested on FACS Symphony (BD Biosciences), and data were analysed with FlowJo software.

| Non-targeted metabolomic profiling for monocytes from blood
The blood from the mice was collected by cardiac puncture. Then, monocytes were isolated by Ficoll-Paque density gradient centrifugation and EasySep Mouse Monocyte Isolation Kit (Stemcell Technologies) according to the manufacturer's instructions.
Monocytes were ground using a tissue crusher, followed by sonication. After centrifugation and filtering, 20 μl of the supernatant was collected for liquid chromatography-mass spectrometry (LC-MS) analysis. Metabolome profiling was performed using a Thermo Scientific Dionex ΜltiMate 3000 Rapid Separation LC system coupled with a Q Exactive hybrid quadrupole Orbitrap mass spectrometer (Thermo Scientific). Metabolites were analysed in the ESI+ and ESI-modes. Data were collected and processed using Progenesis QI 2.3 (Waters Corporation, Milford, MA) software. Data analysis was performed by Beijing Qinglian Biotech, Co., Ltd.

| Bone marrow-derived monocyte isolation and culture conditions
Bone marrow-derived monocytes were obtained from 8-week-old C57BL6J mice and seeded in 6-well plates in RPMI 1640 medium containing 10% foetal bovine serum and 10 ng/ml murine M-CSF at 37 C under 5% CO 2 . On the third day, 10 μM of zoledronate 19 was added to the medium. Exogenous NO and NO scavenger were provided by 10 μM sodium nitroprusside (SNP) and 100 μM Carboxy-PTIO (CP), respectively. After 24 h, monocytes were detected.

| Microcomputed tomography (microCT) assessment
Occlusal view images of tooth extraction sockets were taken just after euthanasia with a stereomicroscope. Then, the maxilla was dissected, fixed with 4% paraformaldehyde, and scanned by SKyScan1276

| Validation of BRONJ-like lesions by histomorphometry
After fixation of right maxillae, demineralization was performed with 10% EDTA at 4 C for 2 months. Demineralized maxillae were paraffin-embedded and sectioned at a thickness of 5 μm. HE, TRAP and Masson's trichrome staining were carried out. ROIs in the tooth extraction sockets were determined as follows: the area surrounded by the line linking the tops of the mesial and distal alveolar ridges, and the curve parallel to the line within 50 μm of the outer line of the laminar dura in the hard tissue of tooth extraction sockets. Moreover, the areas of connective tissue, excluding epithelium above the line linking the tops of the mesial and distal alveolar ridges, were also determined as ROIs in the soft tissue of tooth extraction sockets ( Figure S2). The parameters were histomorphometrically assessed to validate impaired tooth extraction socket healing showing human-like BRONJ conditions, as previously described. 20 24 The plasma and the supernatant from tooth sockets tissue were collected. Before assay, samples were filtered using 10,000 MW filters and diluted. Total Nitric Oxide and Nitrate/ Nitrite Parameter Assay Kit (R&D) was employed to determine the concentration of NOx.

| Cytokines detection of tooth extraction sockets
The multiple cytokine levels of sockets were measured using the LEGENDplex mouse inflammation panel (740446, BioLegend, USA).
According to standard experimental procedures provided by the manufacturer, tissues were separated and weighed for homogenization, and then diluted in a 96-well plate with Assay Buffer. Tissue homogenates were centrifuged at 12,000 rpm for 10 min to collect the supernatant from tooth sockets tissue. 25 μl mixed beads were added to each well and incubated for 2 h at room temperature.
After washing three times with 1 Â Wash Buffer, 25 μl of Detection Antibodies was added to each well and incubated for 1 h. After that, 25 μl SA-PE was added to each well and incubated for 30 min.
Then the prepared samples were tested using FACS Symphony (BD LSRFortessa), and the data were analysed by LEGEND plex software v8.0 (Biolegend).

| Flow cytometry analysis of intracellular NO levels and prodium iodide (PI)-labelled monocyte
Monocytes were collected and stimulated as described as above.
After 24 h, cells were collected, and incubated with intracellular NO probe (Beyotime), PI (Beyotime), respectively. All samples were acquired on a FACS LSRFortessa flow cytometer (BD Biosciences), and data were analysed by FlowJo software.

| Cell viability assay
The cell viability was examined using a 3-(4,5-DimethylthiaZA-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay (Beyotime), which determined a formazan product due to the viable mitochondria in active cells. The media with 0.2% MTT solution were added and incubated at 37 C for 4 h. Cellular reduction of the MTT tetrazolium ring resulted in the formation of a purple water-insoluble deposit of formazan crystals. The medium was discarded, and the reaction was stopped with dimethylsulfoxide and glycine buffer. The coloured product was transferred to a 96-well plate and analysed by measuring the absorbance at 620 nm (A620).

| Statistical analyses
All statistical analyses were performed using GraphPad Prism 9 or the R (http://www.R-project.org). Unpaired Student's t test or one-way ANOVA followed by Tukey's multiple comparison test was to evaluate differences between groups. Data are expressed as mean ± SD. Statistical analyses and graphical representation of data were carried out with GraphPad Prism 7.0. p < 0.05 was considered to be statistically significant.

| Dietary nitrate improved tooth extraction sockets osseous wound healing after zoledronate administration in mice
To investigate the effects of dietary nitrate on jawbone remodelling in zoledronate-treated mice, the BRONJ model was established, as shown in Figure 1A. There was no significant difference in water intake between the groups ( Figure S1). The supplement of 4 mM nitrate in the drinking water significantly increased the NOx content in plasma ( Figure 1B) and the socket surrounding tissues ( Figure 1C).
Analysis was carried out after dividing the regions of interest (ROIs) of the tissue ( Figure S2). MicroCT showed that the alveolar bone of the Ctl group was healing at 3 weeks after tooth extraction, but there was no significant difference in bone mineral density between the alveolar tissue and the surrounding tissue at 5 weeks. So we chose 5 weeks after extraction as the long-term observation point ( Figure S3).
MicroCT analysis ( Figure 1D) revealed that 5 weeks after tooth extraction, Z + Nit group had significantly improved bone healing than that of Z group, which was confirmed by the quantitative analy-  Figure 2K).
In addition, bone metabolism markers in plasma were also investigated, and a significantly higher level of alkaline phosphatase (ALP) was detected in the Z + Nit group compared with that in the other two groups. However, no significant difference in phosphorus or calcium ion content was observed in the plasma ( Figure S5). The overall results indicated that dietary nitrate balanced the dynamics of osteogenesis and osteoclasts.

| Dietary nitrate regulated the immune microenvironment of tooth extraction sockets and the surrounding tissues
Since the monocytes played an important role in initiating and maintaining BRONJ, 25 Figure 3E-H).
F I G U R E 2 Effects of zoledronate therapy and nitrate administration on osseous healing 1 week post-extraction. (A) Schematic overview of study procedures: 4 mM nitrate was added to the drinking water of C57/BL6 mice 1 week before injection of zoledronate up to 1 week after extraction of the right upper first molar (Z + Nit). In the negative control group, there was no injection of zoledronate; only the teeth were extracted (Ctl). In the positive control group, tooth were extracted after zoledronate injection (Z). Mice were sacrificed 1 weeks after extraction, and plasma and maxilla were harvested. (B) Appearance of mucosal healing at the extraction sites and (C) the incidence rate of the delayed healing socket post-zoledronate and nitrate administration (black line: wounds, M2: second molar, bar = 2 mm). (D) Representative HE-stained images of tooth extraction sockets of roots (yellow line: tooth extraction sockets; areas surrounded by red line: necrotic bone with empty lacunae; bar = 500 μm). (E), The number of osteocyte lacunae without nuclear stain were quantified as empty lacunae. (F) Areas containing 10 empty lacunae were quantified as necrotic and are presented as a percentage of the total bone present. (G) Representative trichrome-stained images of tooth extraction sockets (bar = 500 μm) and (H) the blue-stained bone trabeculae were quantified. (I, J) Representative TRAP-stained images (black line: tooth extraction sockets; bar = 100 μm). The number of osteoclasts was significantly increased in the Z + Nit group versus the Z group. (K) Relative contents of CTX-1 and Trap-5b in the homogenate of extracted tooth sockets and surrounding tissue (normalized to Ctl group). n = 10/group. Data are mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, ns, not significant.  . n = 5/group. Data are mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, ns, not significant.

| Dietary nitrate induced a distinct metabolomic pattern of monocyte
After zoledronate injection, monocyte function could be affected and regulated by intracellular metabolism, resulting in a cytokine storm. 25,26 Since NO has a very powerful immune-regulating function, 5  classification analysis of differential metabolites between groups. (G) Pathway enrichment analysis of differential metabolites that were used for visualization. n = 4/group. Data are mean ± SD. *p < 0.05, ***p < 0.001, ****p < 0.0001, ns, not significant. groups (Table S1). The PCA scoring plot showed distinct separation in the Ctl, Z and Z + Nit groups ( Figure 4C).

| NO pretreatment exerted a protective effect on monocytes induced by zoledronate in vitro
Although the non-targeted metabolomic analysis revealed significant metabolic differences between the groups, the pathogenesis of the alteration in lipids and lipid-like molecules of monocytes induced by zoledronate and NO pretreatment is unknown. To investigate how zoledronate could influence monocyte viability, monocytes were isolated from C57/B6 mice and treated with 10 μM zoledronate. SNP -a common NO donor, provided exogenous NO in the culture. An NO scavenger CP was also used in this study to block NO in monocytes. 27 As shown in Figure 5A, monocytes ( Figure S8A) were pretreated with 10 μM SNP or SNP + CP for 12 h and followed by 24 h Z stimulation. DAF-FM DA was used to detect intracellular NO levels. NO levels were significantly reduced after 24 h zoledronate stimulation. Meanwhile, pretreatment with the NO donor effectively increased intracellular NO levels ( Figure 5B, C).
The protective effect of NO on monocytes was confirmed by measuring cell viability using the MTT assay ( Figures 5D and S8B, C).  The relationship between bone marrow stromal cells, osteoblasts and osteoclasts is a dynamic process in bone formation and remodelling.
Excessive inhibition of osteoclasts and their precursors by monocytes can affect osteogenic activity in a dose-dependent manner. 32,33 The F I G U R E 6 Proposed model: dietary nitrate could reduce monocyte necrosis stimulated by zoledronate by producing NO in vivo, thus regulating immunity and promoting bone remodelling after injury. recovery of osteoclasts by nitrate may contribute to bone remodelling. These processes involve the activation of bone turnover coupled with osteoblasts and osteoclasts. We found that zoledronate inhibited the proliferation and differentiation potential of monocytes, the precursors of osteoclasts, resulting in monocyte death in the tooth extraction sockets. However, injection of zoledronate increased the frequency of mature monocytes in peripheral blood. These findings suggest that BRONJ is not only a local lesion in the maxillary bone, but mechanically and more importantly, a systemic deterioration of the monocyte-macrophage system caused by bisphosphonates.
Necroptosis is a non-apoptotic form of regulated cell death that is dependent on RIPK3 and MLKL protein phosphorylation. Phosphorylated RIPK3 activiated the pore-forming protein MLKL, facilitating its oligomerization and subsequent translocation to the plasma membrane, resulting in plasma membrane rupture and lipid molecule changes. 34 Various classes of lipids in eukaryotic cells, such as glycerophospholipids, 35 sphingolipids, 36 and fatty acids, 37 42 However, cost, uncertain efficacy, potential side effects, tolerance issues, immunogenicity and/or ethical issues have delayed the translation of these therapeutic strategies to the clinical setting. The short-acting effect of organic nitrates, intolerance and side effects-such as headache, bradycardia, flushing, nausea and vomiting-limit the use of these organic nitrates. Inorganic nitrate is a hydrophilic salt, produced endogenously and exogenously. It does not undergo first-pass metabolism by the liver and is readily absorbed by the intestinal mucosa. Furthermore, inorganic nitrate lacks immunogenicity, is not subject to tolerance, is non-toxic and does not induce significant side effects at regular doses. 43 Based on our previous studies, 12,13 4 mM nitrate supplementation alone has no significant effect on the body without other challenging. Therefore, inorganic nitrate may have wide clinical applications.
In the present study, we found that the proliferation and differentiation potential of monocytes in tooth extraction was inhibited by zoledronate. In peripheral blood, however, the number of monocytes increased. To further investigate the effect of zoledronate on monocytes, we conducted in vitro experiments. With zoledronate stimulation, the NO levels in monocytes decreased significantly, while SNP, which releases NO, increased the viability and decreased the impairment of monocytes caused by zoledronate. Furthermore, with NO loading, the phosphorylated RIPK3 driven necroptosis pathway was inhibited, alleviating MLKL oligomerization, as verified by western blotting. To further investigate the effects of zoledronate on monocytes and the protection mechanisms of the nitrate/NO axis on BRONJ in vitro, CP was used in the cell culture system. Necroptosis could be activated in monocytes by zoledronate stimulation, while SNP loading could down-regulate this effect. CP could reverse the protective effects of SNP on zoledronate-treated monocytes. Our data suggested that NO, which was derived from nitrate, could protect monocytes from zoledronate-induced necrosis.
In summary, our data revealed a novel regulatory mechanism of the metabolomic signature on innate immune responses induced by zoledronate. Furthermore, we found that dietary nitrate can regulate monocytes stimulated by zoledronate by producing NO in vivo, thus regulating immunity and promoting bone remodelling after injury.
These data might contribute to the understanding of the immunopathogenesis of zoledronate and support the feasibility of dietary nitrate for the clinical prevention of BRONJ.

AUTHOR CONTRIBUTIONS
Wen Pan and Jianyu Gu contributed to conception, design, data acquisition, analysis and interpretation, drafted the manuscript; Shihan Xu, Chunmei Zhang, Jinsong Wang contributed to data acquisition and performed all statistical analyses; Songlin Wang and Junji Xu contributed to conception, design, interpretation and critically revised the manuscript. All authors gave final approval and agree to be accountable for all aspects of the work.