CysLTR1 antagonist (Montelukast) and leukotriene C₄ were purchased from Cayman Chemical Company, (Ann Arbor, MI, USA). P2Y12 antagonist (MRS2395) and 2-(Methylthio)-ADP were purchased from Santa Cruz Biotechnology, Inc. (Dallas, TX, USA). Antibodies against phospho-PLC_γ2, phospho-Akt, phospho-ERK, ERK, and c-Fos were purchased fromCell Signaling Technology, Inc., (Danvers, MA, USA). Antibody against NFATc1 was purchased from Santa Cruz Biotechnology, Inc. Antibody against phospho-CREB was purchased from EMD Millipore (Billerica, MA, USA). Antibody against β-actin was purchased from Sigma-Aldrich; Merck KGaA (Darmstadt, Germany). All other reagents were from Sigma-Aldrich; Merck KGaA.

Male ICR (4- to 6-week-old) mice were purchased from Samtako Inc., (Osan, Korea). Female C57BL/6J (6- to 7-week-old) mice were purchased from DBL Inc., (Umsung, Korea). All mice were maintained in the animal facility of the Sookmyung Women's University in a 12:12 h light-dark cycle at 21±2°C, and were allowed food and water ad libitum. All experiments were performed in accordance with institutional guidelines approved by the Sookmyung Women's University Care and Use Committee.

Mouse bone marrow cells were obtained from the long bones of 4- to 6-week-old male mice. Bone marrow cells were cultured in α-MEM (10% FBS), supplemented with M-CSF (10 ng/ml; R&D Systems, Inc., Minneapolis, MN, USA) for 12 h to separate adherent and non-adherent cells. The non-adherent cells were then harvested and cultured with M-CSF (30 ng/ml). After 3 days of culture, the floating cells were removed and the attached cells used as bone marrow-derived macrophages (BMMs). BMMs (1×10⁴) were plated in 96-well plates and differentiated into OCs by further treatment with M-CSF (30 ng/ml) and RANKL (100 ng/ml). Cells were then fixed and permeabilized with an equal volume mixture of acetone and ethanol for 30 s and then treated with tartrate resistant acid phosphatase (TRAP) staining solution (0.01% naphthol AS-MX phosphate (Sigma-Aldrich; Merck KGaA) and 0.06% Fast Red Violet LB salt (Sigma-Aldrich; Merck KGaA) in 50 mM sodium tartrate dehydrate and 45 mM sodium acetate (pH 5.0)). TRAP-positive multinucleated cells (>3 nuclei/cell) were counted as mature osteoclasts.

Total RNA was purified with easy-BLUE (iNtRON Biotechnology, Seoul, Korea), and cDNA was prepared from 5 µg of RNA using Revert Aid^™ First-Strand cDNA Synthesis kit (Fermentas; Thermo Fisher Scientific, Inc., Waltham, MA, USA). RT-qPCR reactions were performed in a total volume of 20 µl using SYBR⁻Green PCR Master mix (Applied Biosystems; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. Thermocycling was performed using a 7500 real-time PCR System (Applied Biosystems; Thermo Fisher Scientific, Inc.) with the following conditions: initial hold at 95°C for 10 min, followed by 40 cycles of denaturation at 95°C for 15 sec, annealing at 58°C, and extension at 60°C for 1 min. Specific primer sequences for RT-qPCR were designed as follows: calcitonin receptor (CTR), 5′-tttcaagaaccttagctgccagag-3′ (forward), 5′-caa ggc acg gac aat gtt gag aag-3′ (reverse); cathepsin K, 5′-ctt cca ata cgt gca gca ga-3′ (forward), 5′-acg cac caa tat ctt gca cc-3′ (reverse); Atp6v0d2, 5′-tca gat ctc ttc aag gct gtg ctg-3′ (forward), 5′-gtg cca aat gag ttc aga gtg atg-3′ (reverse); DC-STAMP, 5′-tgg aag ttc act tga aac tac gtg-3′ (forward), 5′-ctc ggt ttc ccg tca gcc tctc tc-3′ (reverse); αv-integrin, 5′-cct cag aga ggg aga tgt tca cac-3′ (forward), 5′-aac tgc caa gat gat cac cca cac-3′ (reverse); β3-integrin, 5′-gat gac atc gag cag gtg aaa gag-3′ (forward), 5′-ccg gtc atg aat ggt gat gag tag-3′ (reverse); GAPDH, 5′-tgc acc acc aac tgc tta gc-3′ (forward), 5′-ggc atg gac tgt ggt cat gag-3′ (reverse). Data were analyzed using 7500 System Sequence Detection Software v2.0 (Applied Biosystems; Thermo Fisher Scientific, Inc.). An index mRNA level was assessed using a threshold cycle (Cq) value and normalized against GAPDH expression. qPCR results were calculated using the 2^−ΔΔCq method (15).

Total cell lysates were separated by SDS-PAGE and transferred onto Immobilon-P membranes (EMD Millipore). The membranes were blocked with 5% non-fat-milk in PBS-T, and then immunostained with anti-phospho-PLCγ2 (1:1,000), anti-phospho-Akt (1:1,000), anti-phospho CREB (1:1,000), anti-phospho ERK (1:1,000), anti-ERK (1:1,000), anti-c-Fos (1:1,000), anti-NFATc1 (1:200), or anti-β-actin (1:4,000), followed by secondary horseradish peroxidase-conjugated antibody (1:5,000). The membranes were developed using an enhanced chemiluminescence detection kit (GE Healthcare Life Sciences, Little Chalfont, UK).

Cells were washed twice with PBS and fixed with 10% formalin for 5 min. Cells were permeabilized with EtOH/Acetone (1:1) for 1 min at room temperature, the solvent removed, and the cells dried. Actin rings were stained with rhodamine-conjugated phalloidin (Molecular Probes; Thermo Fisher Scientific, Inc.) overnight in PBS at 4°C in the dark. After washing the cells twice with PBS, the number of actin rings was counted using a fluorescence microscope (Olympus Corporation, Tokyo, Japan).

BMMs were differentiated on dentin slices with M-CSF and RANKL (PeproTech, Inc., Rocky Hill, NJ, USA) for 3 days, and treated with montelukast or MRS2395 for 4 days. The cells were removed from the dentin slice by wiping the surface, and then the slices were stained with toluidine blue (1 µg/ml; J.T. Baker Chemical Co., Phillipsburg, NJ, USA). The number of pits formed by bone resorption on the dentin slices was counted.

To generate retroviral stocks, PMX-puro GFP (GFP-vector) or PMSCV-GFP NFATc1 (NFATc1) was transfected into the packaging cell line platinum-E (Plat-E). Viral supernatant was collected from culture media at 48 h after transfection using Lipofectamine 2000 (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's instructions. For infection with retroviruses, BMMs were incubated with the viral supernatant (4 ml/dish), polybrene (10 µg/ml, Santa Cruz Biotechnology, Inc.), and M-CSF (30 ng/ml) for 2 days, and selected by puromycin (2 µg/ml, Sigma-Aldrich) for an additional 48 h.

BMMs were plated on 48-well plates at a density of 3.5×10⁴ cells/well with 30 ng/ml M-CSF. After 24 h, cells were transfected with 40 nM mouse P2Y12 on-target siRNAs (sc-151960; Santa Cruz Biotechnology, Inc.) using Lipofectamine 2000 (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's instructions. The control contained 40 nM non-targeting siRNA (Qiagen GmbH, Hilden, Germany). The transfection was performed in 2.5 ml of serum free medium for 6 h; the cells were then cultured for 3 days in complete medium containing 30 ng/ml M-CSF and 100 ng/ml RANKL to form osteoclasts.

ICR mice (5-week-old) were i.p. injected with vehicle (EtOH) or montelukast (5 mg/kg) every day. The day after the first injection of montelukast, LPS (0.5 mg) or PBS was directly injected on the calvarium. Six days after LPS injection, mice were euthanized and the calvaria were extracted (n=5 per each group). Calvaria were fixed in 4% paraformaldehyde for 24 h at 4°C, and then stained for TRAP. Image analysis was performed by using ImageJ software (v1.32; National Institutes of Health, Bethesda, MD, USA) according to the manufacturer's protocol.

Female mice (9-week-old, C57BL/6J) were used for ovariectomy. After 5 days, vehicle (1% methylcellulose) or montelukast (5 mg/kg) was orally administered daily for 12 weeks (n=7 per each group). Femurs were extracted and stored in 70% ethanol at 4°C until analysis. Two-dimensional measurement was performed with a micro-CT scanner and associated analysis software (Model 1172; Bruker microCT, Kontich, Belgium) at 9-µm voxel size. Image acquisition was performed at 35 kV of energy and 220 mA of intensity (590 ms of integration time). The threshold was set to segment the bone from the background, and the same threshold setting was used for all samples.

Eight-week-old male ICR mice were orally administered vehicle (1% methylcellulose) or montelukast (5 mg/kg) daily. After a day, mice were subjected to hind limb unloading for 14 days. Hind limb unloading was conducted by applying a tape to the surface of the hind limb to set a metal clip (16). The other end of the clip was fixed to an overhead bar. The height of the bar was adjusted to maintain the mice at 30° head down tilt with the hind limbs elevated above the floor of the cage. Loaded control mice were also housed under the same conditions except for hind limb unloading for the same duration (n=7 per each group). Femurs were extracted and stored in 70% ethanol at 4°C and analyzed by micro-CT as described above.

Data are presented as the mean ± standard deviation from at least three independent experiments. The statistical significance of differences between groups was determined using Student's t-test and analysis of variance with a least significant difference post hoc test. P<0.05 was considered to indicate a statistically significant difference.

RANKL is essential and sufficient for the differentiation of OC precursors into mature OCs in the presence of M-CSF (3,4). Thus, we examined the effects of CysLTR1 antagonist on RANKL-induced OC formation from bone-marrow-derived macrophages (BMMs). We used montelukast since it is a commonly prescribed CysLTR1 antagonist and is generally considered a safe drug with the occurrence of few adverse drug reactions (17). When BMMs were incubated with M-CSF and RANKL for 4 days, numerous TRAP-positive multinucleated OCs were generated. Treatment of the same cultures with montelukast completely suppressed OC formation (Fig. 1A). Consistently, the mRNA expression levels of various OC markers such as calcitonin receptor (CTR), cathepsin K, Atp6v0d2, DC-STAMP, αv-integrin, and β3-integrin were increased in cultures treated with M-CSF and RANKL for 4 days. This effect was dramatically suppressed by montelukast treatment (Fig. 1B).

Since montelukast is known to block the effects of CysLTs by competitive binding to CysLTR1 (11), we next investigated whether addition of CysLT rescues the inhibitory effect of montelukast on OC formation. Suppression of osteoclastogenesis by montelukast was partly rescued by addition of exogenous LTC₄, suggesting the partial involvement of CysLTR1 in the inhibition of OC formation by montelukast (Fig. 1C).

We next investigated the effect of montelukast on mature OCs. Mature OCs exhibit highly polarized morphological features, which is an essential event to initiate bone resorption (18). When mature OCs were incubated with RANKL, a clear cytoplasm and a smooth periphery were observed in polarized OCs. Conversely, the presence of montelukast induced morphological changes such as a contracted cytoplasm and an irregular cell periphery within 24 h (Fig. 1D). When we further observed the morphology of OCs after removal of montelukast from the culture medium, the number of mature OCs with a clear cytoplasm and a smooth periphery increased along with the removal of montelukast (Fig. 1D). Since actin rings are believed to be one of the markers of polarized OCs (3), we examined whether the effect of montelukast on actin ring formation was involved in the morphological changes of OCs. F-actin staining showed that montelukast efficiently disrupted the formation of actin rings in mature OCs, which was recovered after removal of montelukast (Fig. 1E). These data suggest that montelukast reversibly induces the disruption of actin rings, causing reversible morphological changes of mature OCs. To investigate whether the effect of montelukast on the morphology of OCs could be reflected in the osteoclastic activity, we performed an in vitro resorption pit assay using dentine slices. Many resorption pits were generated by RANKL-treated OCs (Fig. 1F). In contrast, montelukast treatment strongly inhibited the formation of resorption pits by RANKL-treated OCs (Fig. 1F). Together, these results suggest that the CysLTR1 antagonist exerts inhibitory effects on the morphology of OCs, which leads to reduced bone resorption.

The NFATc1 pathway plays a critical and fundamental role in OC development and the lack of NFATc1 arrests osteoclastogenesis (8). Thus, we investigated the effect of montelukast on the NFATc1 pathway. Montelukast abolished RANKL-induced NFATc1 expression in BMMs (Fig. 2A). We next investigated whether the effect of montelukast was rescued by the overexpression of NFATc1. The NFATc1-transduced BMMs were cultured with M-CSF and RANKL in the absence or presence of montelukast for 3 days. As shown in Fig. 2B, suppression of osteoclastogenesis by montelukast was efficiently overcome by the forced expression of NFATc1, suggesting that the anti-osteoclastogenic effect of montelukast is mainly due to a reduction in NFATc1 expression.

To identify the molecular mechanism underlying the inhibitory effects of montelukast on osteoclastogenesis, we next examined the effects of montelukast on early signaling pathways such as ERK, Akt, and/or phospholipase Cγ2 (PLCγ2), which activate the NFATc1 pathway (5–7).

As shown in Fig. 2C-E, phosphorylation of ERK, Akt, and PLCγ2 was observed 15 min after RANKL treatment, but suppressed by pretreatment with montelukast. PLCγ2 phosphorylation increases intracellular calcium levels, which in turn activates calcium-dependent transcriptional factors such as cAMP response element-binding protein (CREB) (7,19). Since CREB activation plays an essential role in NFATc1 induction during osteoclastogenesis (9,10), we further examined whether montelukast inhibits CREB activation. RANKL stimulation led to CREB phosphorylation, and montelukast interfered with this process (Fig. 2F). Taken together, these data suggest that montelukast regulates NFATc1 via ERK, Akt, and PLCγ2 signaling pathways.

Since the inhibition of osteoclastogenesis by montelukast was only partially rescued by LTC₄ (Fig. 1C), we speculated that the inhibitory effect of montelukast on OC formation is likely to involve mechanisms other than CysLTR1 blockage. In an effort to identify an alternative receptor for montelukast, we found that the addition of adenosine diphosphate (ADP), a P2Y12 agonist, also rescued the blockade of OC formation by montelukast (Fig. 3A). These data suggest that the inhibitory effect of montelukast on osteoclastogenesis was partly due to its suppressive actions on P2Y12 receptor. To confirm the regulatory role of P2Y12 on OC formation, we next investigated the involvement P2Y12 receptor in OC differentiation. MRS2395, a P2Y12 antagonist, decreased the formation of TRAP-positive OCs in a dose-dependent manner (Fig. 3B). The suppression of osteoclastogenesis by MRS2395 was successfully rescued by addition of exogenous ADP, suggesting the involvement of P2Y12 in the inhibition of OC formation by MRS2395 (Fig. 3C). Similar inhibitory effects on osteoclastogenesis were also obtained by using P2Y12-specific siRNAs (Fig. 3D). Accordingly, the RANKL-induced expression of various OC markers was suppressed by MRS2395 treatment (Fig. 3E). Furthermore, MRS2395 significantly inhibited the resorption pit formation activity of OCs in dentine slices (Fig. 3F).

We next examined the effect of P2Y12 blockade on RANKL-induced NFATc1 expression. Pharmacological P2Y12 inhibition by MRS2395 or down-regulation of P2Y12 by siRNA decreased RANKL-induced NFATc1 expression (Fig. 4A and B). In addition, suppression of osteoclastogenesis by MRS2395 was successfully overcome by the forced expression of NFATc1 (Fig. 4C). MRS2395 abolished RANKL-induced activation of ERK and Akt, as montelukast did (Fig. 4D and E). Taken together, these data suggest that P2Y12 is involved in OC differentiation, and the inhibitory effect of montelukast on OC formation is partly due to P2Y12 blockade.

LPS is produced by gram-negative bacteria and induces the secretion of TNF and other inflammatory cytokines, leading to systemic inflammatory response. Injection of LPS in animals can mimic OC-mediated bone destruction that occurs under sepsis (20). To further investigate the anti-osteoclastogenic effect of the CysLTR1 antagonist in vivo, we injected LPS into the supracalvarial region of mice with or without montelukast. TRAP staining of whole calvariae showed that LPS dramatically increased OC numbers (Fig. 5A and B). In parallel with its anti-osteoclastogenic effect in vitro, montelukast notably reduced LPS-induced OC formation (Fig. 5A and B). These data suggest that the CysLTR1 antagonist has an inhibitory effect on inflammation-induced osteoclastogenesis in vivo.

To further explore the therapeutic effects of montelukast on pathological bone loss, we next used the ovariectomized (OVX) or hind limb unloading mouse model to mimic menopause- or disuse-induced osteoporosis. The µCT analysis demonstrated that OVX significantly decreased the values of bone mineral density (BMD), bone volume/tissue volume (BV/TV), trabecular number (Tb.N), and trabecular thickness (Tb.Th), and increased the values of trabecular separation (Tb.Sp) compared with the sham operation. In contrast, in OVX mice, montelukast treatment significantly inhibited the OVX-induced bone loss assessed by measuring these parameters (Fig. 5C). In addition, the higher structure model index (SMI) number, an indicator of increased fragility, in OVX mice was significantly decreased by montelukast treatment (Fig. 5C). Similar results were demonstrated by a two-dimensional visualization of the femoral area (Fig. 5D). Furthermore, µCT analysis for cortical bone confirmed that OVX significantly decreased the values of cortical area (Ct.Ar), cortical thickness (Ct.TH), and BMD compared with the sham operation (Fig. 5E). By treating montelukast, OVX-induced cortical bone loss was suppressed (Fig. 5E). Montelukast also showed the therapeutic effects in disuse osteoporosis. Two weeks after hind limb unloading, the bone loss was less severe in montelukast-treated mice than that in the vehicle-treated mice (Fig. 6A). Montelukast-treated mice showed higher BMD compared to vehicle-treated mice (Fig. 6B). In addition, unloading-induced structural bone alterations, including decreases in BV/TV, Tb.Th, and Tb.N and increases in Tb.Sp and SMI were abolished by montelukast treatment (Fig. 6B). Furthermore, µCT analysis for cortical bone confirmed that unloading condition significantly decreased the values of cortical area (Ct.Ar), cortical thickness (Ct.TH), and BMD compared with the vehicle group mice. By treating montelukast, unloading-induced cortical bone loss was suppressed (Fig. 6C). Taken together, these results suggest that the CysLTR1 antagonist prevented OVX- or hind limb unloading-induced bone destruction in vivo.