Structure-Dependent Effects of Bisphosphonates on Inflammatory Responses in Cultured Neonatal Mouse Calvaria

Bisphosphonates (BPs) are classified into two groups, according to their side chain structures, as nitrogen-containing BPs (NBPs) and non-nitrogen-containing BPs (non-NBPs). In this study, we examined the effects of NBPs and non-NBPs on inflammatory responses, by quantifying the inflammatory mediators, prostaglandin E2 (PGE2) and nitric oxide (NO), in cultured neonatal mouse calvaria. All examined NBPs (pamidronate, alendronate, incadronate, risedronate, zoledronate) stimulated lipopolysaccharide (LPS)-induced PGE2 and NO production by upregulating COX-2 and iNOS mRNA expression, whereas non-NBPs (etidronate, clodronate, tiludronate) suppressed PGE2 and NO production, by downregulating gene expression. Additionally, [4-(methylthio) phenylthio] methane bisphosphonate (MPMBP), a novel non-NBP with an antioxidant methylthio phenylthio group in its side chain, exhibited the most potent anti-inflammatory activity among non-NBPs. Furthermore, results of immunohistochemistry showed that the nuclear translocation of NF-κB/p65 and tyrosine nitration of cytoplasmic protein were stimulated by zoledronate, while MPMBP inhibited these phenomena, by acting as a superoxide anion (O2−) scavenger. These findings indicate that MPMBP can act as an efficacious agent that causes fewer adverse effects in patients with inflammatory bone diseases, including periodontitis and rheumatoid arthritis.


Introduction
Bisphosphonates (BPs) are well recognized as potent inhibitors of osteoclastic bone resorption and have been successfully used for the treatment of various metabolic bone diseases, such as osteoporosis, Paget's disease, hypercalcemia, and osteolytic tumor-induced bone diseases, which cause excessive bone resorption [1]. All BPs are chemically characterized by a P-C-P bond, and therefore, share common properties (i.e., a strong binding affinity for hydroxyapatite in bone tissue [2] and potent anti-resorption pamidronate were purchased from Sigma-Aldrich Co. (St. Louis, MO, USA) and zoledronate was purchased from Novartis Pharma AG (Tokyo, Japan). Tiludronate and MPMBP were synthesized and supplied by Basic Research Laboratories at Toray Industries Inc., (Kamakura, Japan). NOC18 (1-hydroxy-2-oxo-3, 3-bis (2-aminoethyl)-1-thiazene) and SIN-1 (3-(4-morpholinyl) sydnonimine hydrochloride) were purchased from the Dojindo-Lab (Kumamoto, Japan). The animals were treated ethically in compliance with the commonly accepted '3Rs'-Replacement, Reduction, Refinement-in accordance with the protocol, which was approved by the "Regulations for animal experiments and related activities at Tohoku University (2008DnA-33)" and "Showa University Animal Care and Use Committee guidelines (#10076)."

Effects of NBPs and Non-NBPs on PGE 2 and NO Production in Calvaria
The effects of BPs (1-125 µM for etidronate, clodronate, and MPMBP or 0.1-62.5 µM for pamidronate, alendronate, risedronate and incadronate, or 0.2-25 µM for zoledronate) on the inflammatory responses in the calvaria were assessed by measuring PGE 2 and NO levels in culture media harvested at 24 h (preculture), 48 h (0-24 h), and 72 h (0-48 h) after treatment with BPs. The concentration of PGE 2 was determined using the enzyme immunoassay kit (EIA; Amersham Pharmacia Biotech UK Ltd., Buckinghamshire, UK). NO production was determined by measuring the end products of NO (i.e., nitrite (NO 2 − ) and nitrate (NO 3 − )), in the culture media by the Griess method, using an automated NO detector-HPLC system (ENO-10, Elcom, Kyoto, Japan). In brief, NO 2 − and NO 3 − released into the media, were separated by a reverse-phase separation column packed with polystyrene polymer, and then, NO 3 − was reduced to NO 2 − using a reduction column packed with copper-plated cadmium fillings. Samples containing NO 2 − were mixed with a Griess reagent, which resulted in the diazotization of sulfanilamide by NO 2 − and its subsequent coupling to N- (1-naphthyl) ethylene diamine, to form an azo dye, which was detected by measuring the absorbance at 540 nm.

Reverse Transcription-PCR (RT-PCR) Analyses of Cultured Calvaria
Total RNA was isolated from cultured calvaria using TRIzol (Invitrogen Ltd., Waltham, MA, USA). The cDNA was synthesized from 5 µg of total RNA using the SuperScript TM First-Strand Synthesis System for RT-PCR (Invitrogen Ltd., Waltham, MA, USA). RT-PCR was performed using a thermal cycler (iCycler, Bio-Rad Laboratories Inc., Tokyo, Japan). Quantitative RT-PCR (qPCR) was performed in a Thermal Cycler Dice Real Time System (TP-870; Takara Bio Inc., Shiga, Japan), using SYBR Premix Ex Taq Perfect Real Time (Takara Bio Inc., Shiga, Japan), according to the manufacturer's protocol. The sequence of the specific primers for the qPCR is listed in Table 1. The level of mRNA expression was normalized with that of β-actin.

Effects of Non-NBPs on the Generation of Superoxide Anions in the Xanthine-Xanthine Oxidase System
2-Methyl-6-phenyl-3,7-dihydroimidazo [1,2-a] pyrazin-3-one, a Cypridina luciferin analogue and a chemiluminescent probe (CLA, Tokyo Kasei Kogyo Co. Ltd., Tokyo, Japan) was dissolved in ultrapure water and stored at −20 • C until use. Diethylentriamine-N,N,N ,N",N"-pentaacetic acid (DTPA; Wako Pure Chemical Industries Ltd., Osaka, Japan) and various BPs were dissolved in Hanks' balanced buffered saline solution (HBSS; Gibco). The reaction mixture containing 1 µM CLA, 100 µM DTPA, 100 µM xanthine (Sigma Chemical Co., Saint Louis, MO, USA), and BPs was pre-incubated for 2 min at 37 • C. Then the reaction was started with the addition of 0.5 U/mL xanthine oxidase (Roche Diagnostics Co., Basel, Switzerland) to the reaction mixture at 37 • C. The amount of superoxide anion (O 2 − ) was determined by measuring the integral amount of light emission at 380 nm for 10 s with a luminometer (Luminescencer; Atto Bioscience and Biotechnology, Tokyo, Japan).

Statistical Analysis
Statistical analyses were performed using one-way analysis of variance (ANOVA) for comparison among all groups. The Tukey's honestly significant difference (HSD) test was performed for post hoc pair-wise comparisons after the ANOVA with software KaleidaGraph (HULINKS Inc., Tokyo, Japan). The data were from a minimum of four values in replicate experiments. A p-value less than 0.05 was considered statistically significant and only the significance level compared with the LPS-alone or control group was depicted, as indicated in the figure legends.

Effects of NBPs and Non-NBPs on PGE 2 Production in Cultured Calvaria
Since one of the major causes of BRONJ is persistent inflammation induced by host defense against bacterial infections in the oral cavity, we first explored whether NBPs and non-NBPs have differential effects on LPS-induced inflammatory responses, by measuring PGE 2 production. Calvaria from neonatal mice were pre-incubated in the presence of various NBPs (pamidronate, alendronate, incadronate, and risedronate; Figure 1B) or non-NBPs (etidronate, clodronate, tiludronate, and MPMBP; Figure 1B) for 24 h, to exclude the PGE 2 produced by mechanical stimulation, during the calvaria dissection process (Figure 2A,B).
Antioxidants 2020, 9, x FOR PEER REVIEW 5 of 19 differential effects on LPS-induced inflammatory responses, by measuring PGE2 production. Calvaria from neonatal mice were pre-incubated in the presence of various NBPs (pamidronate, alendronate, incadronate, and risedronate; Figure 1B) or non-NBPs (etidronate, clodronate, tiludronate, and MPMBP; Figure 1B) for 24 h, to exclude the PGE2 produced by mechanical stimulation, during the calvaria dissection process (Figure 2A,B). After changing the media to fresh media containing individual BPs, the calvaria were incubated for another 24 h ( Figure 2C,D) or 48 h ( Figure 2E,F). During the culture period, any toxic effects of BPs were observed under a microscope within the range of 0.1-62.5 μM of pamidronate, alendronate, incadronate, and risedronate, 0.2-25 μM of zoledronate, and 1-125 μM of etidronate, clodronate, tiludronate, and MPMBP ( Figure S1). Since the PGE2 production levels in the LPS-alone group were not significantly different across either different culture periods or different BP treatments, values were expressed as the ratio over the average value for the LPS-alone group in each experiment. As shown in Figure 2C,E, all the NBPs tested stimulated the LPS-induced PGE2 production in a timeand dose-dependent manner, indicating that NBPs might enhance inflammatory responses through  Figure S1). Since the PGE 2 production levels in the LPS-alone group were not significantly different across either different culture periods or different BP treatments, values were expressed as the ratio over the average value for the LPS-alone group in each experiment. As shown in Figure 2C,E, all the NBPs tested stimulated the LPS-induced PGE 2 production in a time-and dose-dependent manner, indicating that NBPs might enhance inflammatory responses through a mechanism involving accelerated PGE 2 production. On the contrary, non-NBPs inhibited LPS-induced PGE 2 production in a dose-dependent manner within the 48 h-long culture period ( Figure 2D,F). a mechanism involving accelerated PGE2 production. On the contrary, non-NBPs inhibited LPSinduced PGE2 production in a dose-dependent manner within the 48 h-long culture period ( Figure  2D,F). Next, we investigated the differential effects of zoledronate, the NBP known to exhibit the most potent antiresorptive activity and cause BRONJ most frequently, and MPMBP, a newly developed non-NBP with an antioxidant side chain, on PGE2 production in the same experimental settings. As expected, 25 μM of zoledronate significantly stimulated the basal level of PGE2 production by 90-and Next, we investigated the differential effects of zoledronate, the NBP known to exhibit the most potent antiresorptive activity and cause BRONJ most frequently, and MPMBP, a newly developed non-NBP with an antioxidant side chain, on PGE 2 production in the same experimental settings. As expected, 25 µM of zoledronate significantly stimulated the basal level of PGE 2 production by 90-and 83-fold after culturing for 24 h and 48 h, respectively, as compared to that observed for the control ( Figure 3A). Furthermore, in the concomitant presence of 10 µg/mL LPS, 5 µM and 25 µM of zoledronate stimulated PGE 2 production 165-and 402-fold at 24 h, and 384-and 1383-fold at 48 h of culture, respectively, as compared to that of the control ( Figure 3B). In contrast, 1-125 µM of MPMBP significantly inhibited LPS-induced PGE 2 production in a time-and dose-dependent manner ( Figure 3D). On the other hand, indomethacin, a nonselective COX-1 and COX-2 inhibitor, completely inhibited LPS-induced PGE 2 production before the end of the pre-culture period ( Figure 3D). The inhibitory effect of MPMBP on PGE 2 production, however, reached a maximum at 48 h after the pre-culture, indicating that MPMBP did not directly inhibit the activity of the PGE 2 synthesizing enzyme.
Antioxidants 2020, 9, x FOR PEER REVIEW 7 of 19 83-fold after culturing for 24 h and 48 h, respectively, as compared to that observed for the control ( Figure 3A). Furthermore, in the concomitant presence of 10 μg/mL LPS, 5 μM and 25 μM of zoledronate stimulated PGE2 production 165-and 402-fold at 24 h, and 384-and 1383-fold at 48 h of culture, respectively, as compared to that of the control ( Figure 3B). In contrast, 1-125 μM of MPMBP significantly inhibited LPS-induced PGE2 production in a time-and dose-dependent manner ( Figure  3D). On the other hand, indomethacin, a nonselective COX-1 and COX-2 inhibitor, completely inhibited LPS-induced PGE2 production before the end of the pre-culture period ( Figure 3D). The inhibitory effect of MPMBP on PGE2 production, however, reached a maximum at 48 h after the preculture, indicating that MPMBP did not directly inhibit the activity of the PGE2 synthesizing enzyme.

Effects of NBPs and Non-NBPs on COX-2 Expression in Cultured Calvaria
As shown in Figure 4A,B (middle panel), RT-PCR and Western blotting analyses showed that 1-100 μM MPMBP suppressed inducible COX-2 expression at both the mRNA and protein levels in cultured calvaria, whereas the constitutive COX-1 mRNA remained unaffected by MPMBP ( Figure  4A, upper panel). The results of qPCR analyses showed that inducible COX-2 mRNA was upregulated by NBPs (alendronate, risedronate, and zoledronate; Figure 4C) and downregulated by non-NBPs (clodronate, tiludronate, and MPMBP; Figure 4D), in the presence of 10 μg/mL of LPS in the cultured calvaria.

Effects of NBPs and Non-NBPs on COX-2 Expression in Cultured Calvaria
As shown in Figure 4A,B (middle panel), RT-PCR and Western blotting analyses showed that 1-100 µM MPMBP suppressed inducible COX-2 expression at both the mRNA and protein levels in cultured calvaria, whereas the constitutive COX-1 mRNA remained unaffected by MPMBP ( Figure 4A, upper panel). The results of qPCR analyses showed that inducible COX-2 mRNA was upregulated by NBPs (alendronate, risedronate, and zoledronate; Figure 4C) and downregulated by non-NBPs (clodronate, tiludronate, and MPMBP; Figure 4D), in the presence of 10 µg/mL of LPS in the cultured calvaria.

Effects of NBPs and Non-NBPs on the Production of NO and Expression of iNOS mRNA in Cultured Calvaria
To further elucidate the differential effects on inflammatory responses between NBPs and non-NBPs, we investigated the effects of NBPs and non-NBPs on LPS-induced NO production and iNOS mRNA expression, under the same experimental conditions as those of PGE2 production and COX-2 mRNA expression. A significant increase in the LPS-induced NO production ( Figure 5A), and a concomitant increase in iNOS expression ( Figure 5C), was observed in all the examined NBPs. On the contrary, MPMBP (5-125 μM) significantly suppressed LPS-induced NO production ( Figure 5B), with concomitant downregulation of iNOS mRNA ( Figure 5D), whereas nNOS and eNOS mRNA expression levels were assessed to be negligible (data not shown), suggesting that MPMBP prevented LPS-induced excessive inflammation. Moreover, tiludronate failed to inhibit LPS-induced NO production significantly at either tested concentration, even though iNOS mRNA expression was downregulated. Immunohistochemical analysis also showed that iNOS protein expression was suppressed by MPMBP ( Figure 5E, middle panel) and enhanced by zoledronate ( Figure 5E, lower panel), in the calvaria cultured under the same conditions as those of NO production and iNOS mRNA expression.

Effects of NBPs and Non-NBPs on the Production of NO and Expression of iNOS mRNA in Cultured Calvaria
To further elucidate the differential effects on inflammatory responses between NBPs and non-NBPs, we investigated the effects of NBPs and non-NBPs on LPS-induced NO production and iNOS mRNA expression, under the same experimental conditions as those of PGE 2 production and COX-2 mRNA expression. A significant increase in the LPS-induced NO production ( Figure 5A), and a concomitant increase in iNOS expression ( Figure 5C), was observed in all the examined NBPs. On the contrary, MPMBP (5-125 µM) significantly suppressed LPS-induced NO production ( Figure 5B), with concomitant downregulation of iNOS mRNA ( Figure 5D), whereas nNOS and eNOS mRNA expression levels were assessed to be negligible (data not shown), suggesting that MPMBP prevented LPS-induced excessive inflammation. Moreover, tiludronate failed to inhibit LPS-induced NO production significantly at either tested concentration, even though iNOS mRNA expression was downregulated. Immunohistochemical analysis also showed that iNOS protein expression was suppressed by MPMBP ( Figure 5E, middle panel) and enhanced by zoledronate ( Figure 5E, lower panel), in the calvaria cultured under the same conditions as those of NO production and iNOS mRNA expression.

Signaling Pathway through which MPMBP Inhibits the Production of Inflammatory Mediators
LPS is known to activate the transcription factors NF-κB and activator protein 1 (AP-1) through its binding to Toll-like receptor 2 or 4 (TLR2/4), and consequently upregulates a large number of genes, including COX-2 and iNOS, involved in inflammatory responses, in many cell types. Of these transcription factors, NF-κB has been reported to be essential for chronic inflammation [23], which is a hallmark feature of the BRONJ. To explore whether BPs affect the translocation of NF-κB to the nuclei, we analyzed the cultured calvaria by laser scanning confocal microscopy. Immunostaining for NF-κB/p65, an active subunit of NF-κB, revealed that 25.0 µM MPMBP inhibited the LPS-induced nuclear translocation of p65, which was clearly recognized as cytoplasmic staining for p65 (red, arrowheads in Figure 6Ah), and only nuclear staining with DAPI (blue) was observed in the merged image, in MPMBP-treated calvaria (arrowheads in Figure 6An). On the contrary, many nuclei appeared to be pink colored, as a result of the blue color of DAPI and red color of Alexa 594-labeled p65 in the merged images in ALPase-positive (green) cells, which were likely to be osteoblasts, in the LPS-alone (arrows in Figure 6Am), and LPS and zoledronate (arrows in Figure 6Ao) groups. Taken together, these results indicate that NF-κB inactivation in LPS-stimulated calvaria was one of possible mechanisms through which MPMBP prevented the inflammatory response.
To further investigate the mechanism by which MPMBP exerts suppressive effects on the inflammation process, we measured the levels of superoxide anion (O 2 − ) generated in the xanthinexanthine oxidase system, because NF-κB is known to be an oxidative stress-responsive transcription factor. As shown in Figure 6C

Discussion
The organ culture of neonatal mouse calvaria has some advantages over the single cell culture or co-culture systems. In the organ culture, both the cytoarchitecture of individual cells, including immune cells (neutrophils, lymphocytes, and macrophages), along with bone cells and the intercellular connections, are thought to be well retained within their natural environment. In addition, they are not influenced by systemic factors. By using this culture system, we investigated the association between the nitrogen-containing side chain of BPs and hyperinflammation levels, as measured by the production of PGE 2 and NO. The results showed that all the examined NBPs (pamidronate, alendronate, incadronate, risedronate, and zoledronate) enhanced LPS-induced inflammatory responses, while all non-NBPs (etidronate, clodronate, tiludronate, and MPMBP) suppressed inflammatory mediator production. Since the results of crystal structure analysis by Hosfield showed that the existence of the nitrogen atom in NBPs was essential for the inhibition of farnesyl pyrophosphate (FPP) synthase [25], a key enzyme in the mevalonate pathway, the mechanism for stimulating inflammation is thought to be identical among NBPs. On the other hand, the mechanisms explaining the anti-inflammatory effects of non-NBPs are likely to be different from each other, because non-NBPs do not share the common chemical structures in their side chains. It is generally accepted that non-NBPs act as cytotoxic ATP analogues in the osteoclasts; however, there is no evidence that MPMBP forms an ATP analogue, through which it inhibits osteoclastic bone resorption or other functions.
In the present study, we demonstrated that MPMBP exhibited the most potent anti-inflammatory properties against both the basal and LPS-induced production of PGE 2 and NO, by downregulating the expression of COX-2 and iNOS mRNA. Since NF-κB is well known to be a crucial transcription factor for inflammation-related genes, including COX-2 and iNOS, we observed the cultured calvaria by using confocal microscopy, in order to examine the contribution of NF-κB to the excessive production of PGE 2 and NO that leads to hyperinflammation. The results showed that the nuclei of an osteoblastic marker for ALP-positive cells were stained positive for NF-κB/p65 after LPS stimulation, indicating that NF-κB was activated in the osteoblasts in response to LPS. Under the same culture conditions, zoledronate stimulated the nuclear translocation of NF-κB/p65 in the calvaria, while MPMBP inhibited the nuclear translocation of NF-κB/p65, as evidenced by the cytoplasmic staining for inactive form of NF-κB/p65 in ALP-positive cells ( Figure 6A).
Calvaria used in this study contained several types of cells, including bone cells (osteoblasts, osteoclasts, and osteocytes), neutrophils, macrophages, fibroblasts, and osteoblast progenitor cells in the periosteum and bone marrow. Although the cell type which contributes to LPS-induced PGE 2 production in the cultured calvaria is thought not to be single, since PGE 2 was reported to be produced mainly by osteoblasts in bone [26], in response to various stimuli [27], our current results indicate that osteoblasts are considered to be one of cell types that produce PGE 2 through NF-κB activation in calvaria. Meanwhile, activated macrophages (M1 polarized) are also known to play a critical role in inflammatory diseases, through the excessive production of inflammatory mediators, including PGE 2 and NO, as well as pro-inflammatory cytokines [28][29][30], in which a transcription factor NF-κB is involved. Additionally, our recent investigation showed that MPMBP significantly inhibited Toll-like receptor 2 agonist-induced NF-κB/p65 activation in the macrophage-like cell line, J774.1 [31]. This evidence suggests that the activated macrophage might be a candidate for inflammation-responsive cell types in the calvaria and that MPMBP would be able to suppress inflammatory responses by acting against macrophages in the calvaria. When cells are stimulated with various pathogens, IκB is degraded post-phosphorylation, which leads to the activation of the oxidative stress-responsive transcription factor NF-κB. On the other hand, there is an alternate pathway, in which the tyrosine residue of IκB is nitrated by ONOO − , the extremely reactive molecule produced from NO and O 2 − [24,32]. This can cause NF-κB activation for a longer period than that observed for IκB phosphorylation [33], without the degradation of IκB [34]. The present study showed that zoledronate strongly enhanced the production of PGE 2 and NO; this might be attributable to the activation of NF-κB, which is concomitant with the increase in the cytoplasmic level of nitrotyrosine (Figure 6Ed). This finding indicates that the tyrosine nitration of IκB is thought to be a possible mechanism for NF-κB activation by zoledronate.
Although previous reports show that IκB is one of the candidates to be nitrated on its tyrosine [33,35,36], further investigations using immunoprecipitation and/or immunohistochemistry analyses in calvaria would enable us to identify the protein to which the tyrosine is nitrated by ONOO − . Taken together, these findings raise the possibility that once calvaria is stimulated with LPS, neutrophils generate a great amount of O 2 − , which can produce cytotoxic ONOO − , by reacting with the NO produced by osteoblasts and/or activated macrophages. As a consequence, ONOO − activates NF-κB, which results in the upregulation of inflammation-related genes, including COX-2 and iNOS in the osteoblasts and/or macrophages (Scheme 1).
Antioxidants 2020, 9, x FOR PEER REVIEW 14 of 19 Scheme 1. Proposed mechanism underlying the differential effects of N-containing zoledronate and non-N-containing MPMBP, on the inflammatory responses. The hypothesis is based on the experimental findings in the present study and published reports. We hypothesized that persistent inflammation leading to tissue damage might be caused by the signaling pathway as follows: (a) the LPS-stimulated production of O2 − by neutrophils; (b) peroxynitrite, which was produced from superoxide anions and NO, freely passes through the cell membrane; and (c) nitration of the tyrosine residue of IκB, which was reported to be a candidate for nitration target [33,35,36]; (d) translocation of activated NF-κB into the nucleus and thereby, (e) upregulation of COX-2 and iNOS mRNA expression, which, in turn, increases the production of (f) PGE2 and (g) NO. This pathway from (a) to (g) is a vicious cycle that leads to persistent inflammation.
Szabo et al. showed that the sites of peroxynitrite formation are spatially associated with sources of O2 − , reflecting its much shorter half-life and lower biomembrane permeability, owing to its negative charge, as compared to that of NO [37]. On the other hand, previous reports have shown that ONOO-was a potent nitrating agent with a short half-life (<1 s) [38], and both motor neuron apoptosis and nitrotyrosine were prevented either by the inhibition of NO synthesis or O2 − scavengers [39]. This indicates the significance of ONOO-as a biological effector molecule that can diffuse freely and rapidly across phospholipid membrane bilayers and react with target substrates [40]; this might support the hypothesis shown in Scheme 1. Furthermore, zoledronate accelerates this vicious cycle, Scheme 1. Proposed mechanism underlying the differential effects of N-containing zoledronate and non-N-containing MPMBP, on the inflammatory responses. The hypothesis is based on the experimental findings in the present study and published reports. We hypothesized that persistent inflammation leading to tissue damage might be caused by the signaling pathway as follows: (a) the LPS-stimulated production of O 2 − by neutrophils; (b) peroxynitrite, which was produced from superoxide anions and NO, freely passes through the cell membrane; and (c) nitration of the tyrosine residue of IκB, which was reported to be a candidate for nitration target [33,35,36]; (d) translocation of activated NF-κB into the nucleus and thereby, (e) upregulation of COX-2 and iNOS mRNA expression, which, in turn, increases the production of (f) PGE 2 and (g) NO. This pathway from (a) to (g) is a vicious cycle that leads to persistent inflammation.
Szabo et al. showed that the sites of peroxynitrite formation are spatially associated with sources of O 2 − , reflecting its much shorter half-life and lower biomembrane permeability, owing to its negative charge, as compared to that of NO [37]. On the other hand, previous reports have shown that ONOO − was a potent nitrating agent with a short half-life (<1 s) [38], and both motor neuron apoptosis and nitrotyrosine were prevented either by the inhibition of NO synthesis or O 2 − scavengers [39].
This indicates the significance of ONOO − as a biological effector molecule that can diffuse freely and rapidly across phospholipid membrane bilayers and react with target substrates [40]; this might support the hypothesis shown in Scheme 1. Furthermore, zoledronate accelerates this vicious cycle, which leads to uncontrolled persistent inflammation, while MPMBP may prevent tyrosine nitration via its O 2 − scavenging activity. Although the inhibitory effect of MPMBP on O 2 − production was not assessed in the cultured calvaria in this study, it is likely that MPMBP acts as an O 2 − scavenger in the calvaria, as otherwise a large amount of O 2 − would be produced by neutrophils in the presence of LPS. Since it was reported that the methylthio-residue could be s-demethylated by endogenous enzymes [41,42], it was possible that the thiol group (-SH) produced from the methylthio moiety of MPMBP functioned as an antioxidant, in a manner similar to that of SH-containing pyrrolidine dithiocarbamate (PDTC), a specific NF-κB inhibitor that was shown to inhibit the LPS-induced expression of COX-2 [43] and iNOS [44] mRNA. In contrast to antioxidant activity of MPMBP ( Figure 6C, left panel), none of the NBPs exhibited an O 2 − scavenging activity ( Figure 6C, right panel).
In addition, the association between long-term or high-dose NBPs and decreased plasma levels of antioxidants such as coenzyme Q10 (CoQ10), which serves as an endogenous antioxidant in every cell, has been demonstrated in postmenopausal women [45]. Furthermore, previous reports have demonstrated that gingival tissues from patients with periodontal disease showed a deficiency in CoQ10 [46,47], which supports this finding. Since the inhibition of FPP synthase by NBPs results in decreased levels of production of CoQ10 due to the shortage of precursor FPP [48,49], and no redistribution between organs occurs via circulation [50], it is likely that the reduced levels of antioxidants, such as CoQ10, could cause NBP-associated persistent inflammation, which leads to BRONJ.
Since BRONJ was first reported as a rare but serious adverse effect of BP by Marx in 2003 [51], the clinical use of BPs has been hampered at times, in spite of its beneficial effects on the risk of fracture in osteoporosis patients. Although the precise mechanisms for the onset of BRONJ remain unclear, accumulating evidence indicates that BRONJ is associated with invasive dental procedures, including teeth extraction, dental implant placement, periapical surgery, and periodontal surgery [11], in which a great number of neutrophils can be observed. This would generate a large amount of O 2 − , thereby leading to the production of ONOO − , which causes tissue damage. LPS, a pathogen used in the present study, is a major anaerobic bacterial component that causes inflammation in oral tissues, which are prone to be exposed to microorganisms, and a great number of pathological studies have suggested that BRONJ is an inflammation-associated process [52,53]. Thus, the stimulatory effects of NBPs on the inflammatory responses in alveolar bone and surrounding tissues might be related to the onset and development of BRONJ. Findings regarding the stimulatory effects of zoledronate on inflammation in this study, which are depicted in Scheme 1, are likely to explain the pathogenesis of BRONJ. The anti-inflammatory action of MPMBP was attributable to its potent antioxidant activity, as evidenced by the O 2 − scavenging action in both human PMN [13] and the xanthine-xanthine oxidase system ( Figure 6C). The consequent suppression of ONOO − production might result in a lower incidence of BRONJ, as compared with that of patients treated with NBPs. Further studies by using BRONJ animal model would be one of the strategies to examine the relevance of our hypothesis which is depicted in Scheme 1. In addition to BRONJ, the acute-phase response, another adverse effect of NBPs featured by influenza-like symptoms, has been shown to be attributed to the release of proinflammatory cytokines [10].
In the present study, we selected the dose of BPs in order to obtain comparable anti-resorptive activities among different BPs (i.e., between 0.1-62.5 µM and 1-125 µM for NBPs and non-NBPs, respectively). This would enable us to obtain the relevant estimates of the balance between their efficacy as anti-resorption agents and the safety of therapeutics. It is well known that BPs have extremely low oral bioavailability (at a range of 0.6 to 1.5% of the administered dose [54,55]) due to their negative charge. On the other hand, all BPs have a strong affinity for hydroxyapatite in bone tissue because of common P-C-P chemical structure; therefore, they can be retained in bone for a long time which enables them to act primarily on the bone. Rodan described the cellular and molecular processes in periodontitis, in which the accumulated bacteria caused the destruction of the periodontium and therefore of possible therapeutic targets, which were likely to be similar to those used in rheumatoid arthritis patients [56], an inflammatory bone disease. Furthermore, Sandhu showed that nitrotyrosine-containing proteins were found in essentially all synovia from rheumatoid arthritis and osteoarthritis patients [57]. Taken together with the previous findings, which demonstrated that MPMBP treatment prevented against the inflammation-related bone loss in animal models [14][15][16]18], it is possible that MPMBP can act as an effective anti-resorptive agent with less adverse effects, including BRONJ and acute phase response, as compared to most currently available NBPs.
Another problem associated with BRONJ is that fixed treatment strategies remain unknown at present [11]. Recently, Hokugo et al. showed a unique strategy to reverse the adverse effect of zoledronate in mice. It was observed that the local administration of a "rescue" BP other than zoledronate dramatically reduced the development of osteonecrosis symptoms [58]. Since the course of BRONJ is progressive and greatly influenced by exposure to surrounding microorganisms in the oral cavity, which leads to uncontrolled inflammation, MPMBP may be a useful compound as a "rescue" BP, due to the most potent antioxidant and anti-inflammatory activities among currently available BPs.

Conclusions
All the examined NBPs (pamidronate, alendronate, incadronate, risedronate, zoledronate) stimulated the LPS-induced production of PGE 2 and NO, by upregulating the expression of COX-2 and iNOS mRNA, whereas non-NBPs (etidronate, clodronate, tiludronate, MPMBP) suppressed production of PGE 2 and NO, by downregulating gene expression. Among non-NBPs, MPMBP exhibited the most potent anti-inflammatory activity via the inhibition of NF-κB/p65 nuclear translocation and tyrosine nitration of cytoplasmic proteins, by acting as an O 2 − scavenger. These findings indicate that MPMBP can act as an efficacious anti-resorptive agent exhibiting less adverse effects in inflammatory bone diseases, including periodontitis and rheumatoid arthritis.

Patents
The content of this manuscript partially correlates to that of our MPMBP patents (Japanese patent: 5655179 and European patent: 2340841).