The study was approved by the Ethical Committee of the Okayama University Graduate School of Medicine, Dentistry, and Pharmaceutical Sciences (protocol number: 1949), including use of an advertisement poster in place of patient consent. Written consent was not acquired for this retrospective study. The authors had access to patients’ records prior to data anonymization. All the patients were examined and treated at Okayama University Hospital (Okayama, Japan) between 2000 and 2010, and the diagnosis was clinicopathologically confirmed aggressive invasive phenotype of lower gingival squamous cell carcinoma (n = 10). The surgically resected mandibles were collected as part of routine care by the authors. No patient had received chemotherapy and/or radiation therapy before surgery. All patients’ records/information were anonymised and de-identified prior to analysis. Sections from the deepest part of the invasion and the boundary between the tumor and the bone were evaluated primarily by light microscopic observation. The sections were deparaffinized, and then autoclaved in 0.2% citrate buffer for 15 min for antigen retrieval. Sections were incubated with 3% hydrogen peroxide for 30 min to block endogenous peroxidase activity. A primary anti-SHH (rabbit IgG), anti-Gli-2 (rabbit IgG) (Cell Signaling Technology, Danvers, MA), anti-Patched1 (rabbit IgG, Proteintech, Chicago IL) and anti-CD68 (mouse IgG, Dako, Glostrup, Denmark) was used for the immunohistochemical analysis. The specimens were incubated with a 1:100 dilution of the antibody overnight at 4°C, followed by 3 washes with TBS. The slides were then treated with a streptoavidin-biotin complex (Envision System Labeled Polymer, horse radish peroxidase (HRP); Dako, Carpinteria, CA) for 60 min at a dilution of 1:100. The immunoreaction was visualized using a 3,3'-diaminobenzidine (DAB) substrate-chromogen solution (Dako Cytomation Liquid DAB Substrate Chromogen System, Dako), and counterstaining was performed with hematoxylin. Finally, the sections were immersed in an ethanol and xylene bath and mounted for examination.

The murine macrophage RAW264.7 cells from the American Type Culture Collection (ATCC) were cultured in Dulbecco's Modified Eagle’s Medium (DMEM) supplemented with 10% heat-inactivated fetal bovine serum (FBS). CD11b⁺ bone marrow cells were cultured in α-Modified Eagle’s Medium (αMEM) with 10% FBS. Both cell types were cultured in an atmosphere of 10% CO₂ at 37°C.

Animals were housed in cages under pathogen-free conditions with humane care and anesthesia was performed using pentobarbital 150 mg /kg I.P. to minimize suffering. The experimental protocols for CD11b⁺ osteoclast progenitor cells from C57BL/6 mice (Charles River) was approved by the Ethics Review Committee for Animal Experimentation of the Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Sciences (protocol number: 2010291), and the animals for floxed Smoothened mice (Smotm^2Amc/J, The Jackson Laboratory) was approved for IACUC (Institutional Animal Care and Use Committee) of the Thomas Jefferson University (Authors' Former Institute). CD11b⁺ osteoclast progenitor cells were prepared from 5-week-old male C57BL/6 mice or floxed Smoothened mice. Bone marrow cells were washed twice in 20 ml of ice-cold phosphate-buffered saline (PBS) supplemented with 0.5% bovine serum albumin (Sigma, St. Louis, MO) and 2 mM EDTA (Sigma). The cell pellet was resuspended in 80 μl of buffer A per 10⁷ cells, and was incubated with 20 μl of anti-CD11b magnetic microbeads (Miltenyi Biotec Inc., Auburn, CA). The labeled cells were washed, resuspended in 500 μl of buffer A per 10⁸ cells, and applied onto MD depletion column (Miltenyi Biotec Inc., Bergisch Gladbach, Germany) placed in the magnetic field of a MidiMACS separation unit (Miltenyi Biotec Inc.). After washing, we removed the column from the separation unit and eluted CD11b⁺ cells. To study the direct role of hedgehog signaling in osteoclastogenesis, we removed smoothened gene in CD11b⁺ positive osteoclast precursor cells from floxed Smoothened mice by infecting Cre-recombinase encoding adenovirus (Ad-CMV-iCre, Vector Biolab, Malvem, PA). Briefly, freshly isolated CD11b⁺ positive cells were incubated with 100 pfu of Ad-CMV-Cre or control Ad-CMV-GFP (Vector Biolab) for 1 h in ice. Cells were resuspended every 15 min and then seeded onto the culture plates. Medium was replaced the next day. Adenovirus infection was confirmed at 70–90% 72 h after infection under fluorescence microscopy.

RAW264.7 cells were rinsed once with ice-cold phosphate buffered saline (PBS) and lysed in an ice-cold lysis buffer (50 mM Tris-HCl, pH 7.4, containing 150 mM NaCl, 1% Triton X-100, 1% NP-40, 10 mM NaF, 100 mM leupeptin, 2 mg/ml aprotinin, and 1 mM phenylmethyl sulfonyl fluoride). Cell lysates containing 10 μg of total protein in a lysis buffer were electrophoresed in 12% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) gels and the proteins were transferred to nylon membranes (Immobilon-P; Millipore Co.). The membranes were incubated with primary and secondary antibodies according to the ECL chemiluminescence protocol (RPN2109; Amersham Biosciences, Buckinghamshire, UK) to detect secondary antibody binding. Anti-ERK1/2 (rabbit IgG), pERK1/2 (Thy202/Tyr204, rabbit IgG), p38 (rabbit IgG), p-p38 (Tyr180/Tyr182, rabbit IgG) were purchased from Cell Signaling Technology Inc. (Danvers, MA). NFATc1 (mouse IgG), Cathepsin K (rabbit IgG), tartrate-resistant acid phosphatase (TRAP) (rabbit IgG) and actin (goat IgG) antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA) primary antibodies were used at a 1:200 dilution, and horseradish peroxidase-conjugated goat anti-rabbit antibodies or goat anti-mouse IgG were used as the secondary antibodies at a 1:1000 dilution. Image-J software (NIH, Bethesda, MD) was used to quantify band intensities on Immunoblots and the statistical analysis was performed by Bonferroni multiple comparisons tests.

RAW264.7 murine macrophage cells were treated with 50 ng/ml of recombinant mouse RANKL (PEPROTECH EC, London, UK) for 5 days CD11b⁺ bone marrow cells and bone marrow macrophages [21] were cultured with 30 ng/ml M-CSF, (R&D System, Minneapolis, MN) and 50 ng/ml RANKL in the presence or absence of SHH (R&D System, Minneapolis, MN) or Cyclopamine (Toronto Chemical Inc., Miltenyi Biotec Inc., Auburn, CA) for 8 days. The complete medium was changed every 2 days. The cells were then fixed and stained for TRAP (Sigma) and the number of TRAP-positive multinucleated cells (nuclear number >3) in each well was counted. For the osteoclast activity assay, CD11b⁺ bone marrow cells were seeded on dentin slices that had been placed in 96-well plates (1 slice/well) and were cultured for 5 days in αMEM containing RANKL (50 ng/ml) and M-CSF (30 ng/ml). Then the medium was changed to αMEM with or without RANKL (50 ng/ml) and SHH (500 ng/ml), and the cells were incubated for 7 days. A scanning electron microscope (SEM, VE9800; Keyence, Osaka, Japan) was used to examine the ultrastructure of the dentin slices. The area of pits on the slices was quantified by imaging analysis (Lumina Vision; Mitani Corporation, Osaka, Japan). The results were expressed as the mean ± SD of three cultures.

RAW264.7 cells were treated with 50 ng/ml RANKL for 3 days, and then cultured with or without 5 μM smoothened agonist (SAG) (Merck Millipore, Parmstadt, Germany) for 24 h. After the treatment with SAG, total RNAs were isolated using an miRNeasy Mini Kit (Qiagen, Valencia, CA). Genome-wide gene expression analysis was performed using SuperPrint G3 Mouse GE 8x60K 2 color (Agilent Technologies, Santa Clara, CA) according to the manufacturer’s protocol. The arrays were scanned and digitized by Agilent Feature Extraction 10.7.3.1 software (Agilent Technologies). The data was analyzed by GeneSpring GX v.13 (Agilent Technologies).

Total RNA was isolated according to the manufacturer’s protocol (RNeasy Minikit, Qiagen). After RNA reverse transcription, 1 μl of each cDNA preparation was PCR-amplified using the set of primers described below. The tachykinin receptor 3 primmer: 5’-TTGCAGTGGACAGGTATATG-3’ and 5’-GATGTGGTGGAGGCAGATTT-3’, NFATc1 primer: 5’- CAAGTCTCTTTCCCCGACATC-3’ and 5’- CTGCCTTCCGTCTCATAGTG-3’. The condition of quantitative real-time PCR (qPCR) were as follows: 40 cycles of 95°C for 30 seconds 60°C for 30 seconds, and 72°C for 30 seconds for tachykinin receptor 3 or 37 cycles of 95°C for 5 seconds and 62°C for 30 seconds for NFATc1. Melting curves were obtained, RT-qPCR was performed on the chromo4 (MJ Research, Waltham, MA, USA) using a SsoAdvanced Universal SYBR Green Supermix (Bio-Rad Laboratories, Inc. Philadelphia, PA), and the data was analyzed by using ΔΔCT methods.

Data were analyzed by using the unpaired Student’s t-test for analysis of two groups, Fisher’s protected least significant difference (Fisher’s PLSD) and Bonferroni for analysis of multiple groups. Results were expressed as the mean ± S.D. p < 0.05 was considered statistically significant.

Fig 1 shows representative microscopic images of invasive bone destruction observed in a patient with oral squamous cell carcinoma in the mandibular region. SHH, Patched, CD68 and Gli-2 were not detected in the normal lower gingival epithelium (Fig 1A). SHH was highly expressed in tumor cells that had invaded the bone matrix, and weakly expressed in osteoclasts (Fig 1B). On the other hand, the hedgehog receptor Patched was highly expressed in the osteoclasts and the progenitor cells (Fig 1B). CD68 was detected in both osteoclasts and progenitor cells (Fig 1B). Gli-2 was expressed in the progenitor cells, and weakly expressed in tumor cells (Fig 1B). We analyzed additional 9 cases of mandibular squamous cell carcinoma with invasive bone destruction and observed (i) intense SHH immunoreactivity in tumor cells, (ii) intense Patched and Gli-2 immunoreactivity in osteoclasts and progenitor cells in all cases (not shown).

Next, we studied the effect of hedgehog on osteoclastogenesis. To elucidate the effect of SHH in osteoclast formation, RAW264.7 murine macrophage cells, known as osteoclast precursor cells, were inoculated onto 96-wells plates at different cell densities and then treated with increasing concentrations of RANKL with or without SHH for 6 days. The concentration of SHH for osteoclastogenesis was determined 500 ng/ml by the previous our studies [19], and the osteoclast formation was evaluated by TRAP staining. SHH alone did not affect osteoclastogenesis, but greatly enhanced RANKL-induced osteoclastic cell differentiation (Fig 2A). Next, we evaluated the effect of hedgehog on CD11b⁺ osteoclast precursor cells. CD11b⁺ bone marrow cells were maintained in αMEM containing 10% FBS and 30 ng/ml M-CSF, and treated with or without 50 ng/ml RANKL and 500 ng/ml SHH for 6 days. The cells were then fixed and stained for TRAP activity, and TRAP cells containing 3 nuclei were counted as osteoclasts by microscopy (Fig 2B). As seen in RAW264.7 cells, treatment with SHH alone did not induce osteoclast formation compared to the control values in untreated cultures. As expected, RANKL treatment did induce formation of TRAP-positive and a small number of multinucleated osteoclastic cells. Strikingly, massive osteoclastogenesis was significantly observed following co-treatment with RANKL and SHH (Fig 2B). The data indicate that SHH acts directly on osteoclast progenitor cells and stimulates RANKL-induced osteoclastogenesis. To analyze the effect of SHH on osteoclastogenesis, CD11b⁺ bone marrow cells were treated with or without 500 ng/ml SHH and the hedgehog signaling inhibitor cyclopamine for 6 days. Treatment with cyclopamine significantly blocked the SHH-induced increase in osteoclastogenesis and brought it down to the basal level at 6 μM (Fig 2C). We also confirmed the significant dose dependent effect of SHH on the osteoclastogenesis by using primary isolated bone marrow macrophages (Fig 2D). To further confirm the role of SHH in osteoclast differentiation, we isolated CD11b⁺ bone marrow cells from floxed Smoothened mice, infected cells with either Cre recombinase encoding adenovirus or control adenovirus, and then compared the effect of SHH on RANKL-induced osteoclastogenesis (S3 Fig). As expected, stimulating effects of SHH on RANKL-induced osteoclastogenesis were almost completely abolished by ablation of Smoothened (S3 Fig). Then, we analyzed the effect of SHH on the bone resorption activity of mature osteoclasts. An osteoclast activity assay using CD11b⁺ bone marrow cells showed that SHH alone did not stimulate osteoclast resorption, but significantly enhanced the RANKL-induced bone resorptive activity of osteoclasts (Fig 3A and 3B).

To define the molecular mechanisms of the SHH effects on osteoclast formation, we examined the effects of SHH on the signaling pathways in RAW264.7 cells. RANKL stimulated the transient ERK and p38 phosphorylation at 15 min (Fig 4A). On the other hand, treatment of SHH with RANKL increased the sustained and synergistically promoted ERK phosphorylation for 30 and 60 min (P < 0.05) and transient p38 phosphorylation for 15 and 30 min (P < 0.01) compared with the single treatment of RANKL (Fig 4A). Total ERK and p38 levels remained constant after the treatments (Fig 4A). NFATc1 is a master regulator of RANKL-induced osteoclast formation and herein we investigated the effect of SHH on the NFATc1 expression with or without RANKL. Our results showed that the expression of NFATc1 in the presence of RANKL and SHH was significantly higher than in the presence of RANKL alone for 24 and 48h (Fig 4B). Immunoblot analysis showed that the TRAP and Cathepsin K protein expressions induced by RANKL in RAW264.7 cells were stimulated by SHH treatment for 4 days (Fig 4C). To confirm whether RAW264.7 cells and CD11b⁺ cells share similar mechanisms for SHH’s effect on osteoclast formation, Immunoblot and qPCR analysis were performed. SHH with RANKL synergistically promoted ERK and p38 phosphorylation for 30 and 60 min compared with the single treatment of RANKL in CD11b⁺ cells (S4A Fig). SHH significantly upregulated the relative NFATc1 mRNA levels in the presence of RANKL in CD11b⁺ cells (P < 0.01, S4B Fig).

Since SHH appears to play an important role in the differentiation and activation of osteoclasts, we compared the endogenous expression levels of genes between vehicle- or SAG-treated RAW264.7 cells by microarray analysis (S5 Fig). When compared with the vehicle-treated samples, SAG-treated samples showed an elevation of various pathways, most of which were G protein-coupled receptor (GPCR) pathways (S1 Fig). S2 Fig shows the genes that were changed more than three times relevant to the top 3 biological pathways depicted in pathway analysis in S1 Fig. The expression of tachykinin receptor 3 showed the highest upregulation, 9.8-fold over the control level, in the non-odorant GPCR pathways (S2 Fig). The expression of the chemokine receptor 3 (CCR3) gene in the GPCR2 class A rhodopsin-like pathways and olfactory receptor 498 in odorant GPCRs were also upregulated compared to the control levels (S2 Fig).