SGN and laricitrin were purchased from Extrasynthese (Genay, France); resveratrol, pterostilbene, piceatannol and myricetin were obtained from Sigma Chemical (St. Louis, MO, USA) (Fig. 1). These were dissolved in dimethylsulfoxide (DMSO) (Sigma Chemical) and stored at −20°C. Control cultures received the carrier solvent (0.1% DMSO). All other chemicals used were of the purest form available commercially.

Human lung adenocarcinoma A549 cells were obtained from the American Type Culture Collection (CCL-185) (Manassas, VA, USA) and cultured in F-12K medium containing 10% fetal bovine serum (FBS) (both from Gibco-BRL, Gaithersburg, MD, USA). CL1-5 human lung adenocarcinoma cells were generously provided by Dr Pan-Chyr Yang (Department of Internal Medicine, National Taiwan university Hospital) (33,34), and cultured in RPMI-1640 medium supplemented with 10% FBS and 1% penicillin-streptomycin (both from Gibco-BRL). Human primary osteoblasts were cultured in osteoblast growth medium (OBM) (both from Lonza, Walkersville, MD, USA). For conditioned medium (CM) collection, the A549 and CL1-5 cells were treated with vehicle (0.1% DMSO) or various concentrations of SGN for 12 h. After washing, fresh culture medium was added and cultured for another 24 h. The supernatants were collected, filtered (0.22-mm) and identified as A549-CM, SGN-A549-CM, CL1-5-CM and SGN-CL1-5-CM. Osteoblasts were cultured with A549-CM, SGN-A549-CM, CL1-5-CM or SGN-CL1-5-CM (20%) for another 24 h, then the supernatants were collected and filtered (0.22-mm). These supernatants were grouped as osteoblast-CM (OB-CM), A549-OB-CM, SGN-A549-OB-CM, CL1-5-OB-CM and SGN-CL1-5-OB-CM. All condition media were frozen and stored at −80°C, with a single thawing for study.

Human peripheral blood, obtained from healthy adult volunteers, was collected in syringes containing 1,000 U/ml of preservative-free heparin. Peripheral blood mononuclear cells (PBMCs) were isolated by density centrifugation using Ficoll-Hypaque, and re-suspended in RPMI-1640 medium supplemented with 10% heat-inactivated FBS. The PBMCs were then plated and incubated overnight at 37°C. CD14⁺ monocytes were isolated using CD14⁺ mAb-conjugated magnetic beads (MACS MicroBeads; Miltenyi Biotec, Ltd., Bergisch Gladbach, Germany) according to the manufacturer’s protocol. The purity of CD14⁺ cells was 94–97%. Monocytes were grown in culture medium containing vehicle or tested compounds preset at 200 ng/ml human M-CSF and 100 ng/ml human RANKL (R&D Systems, Minneapolis, MN, USA) for 14–21 days. The medium was replaced every 5 days. The Institutional Review Board of the Kaohsiung Medical University Chung-Ho Memorial Hospital approved the study protocol, and all participants provided written informed consent in accordance with the Declaration of Helsinki. Osteoclast formation was measured by quantifying cells positively stained by tartrate-resistant acid phosphatase (TRAP) (Sigma-Aldrich, St. Louis, MO, USA). Briefly, the cells were fixed with formaldehyde for 30 min and then stained with naphthol AS-BI phosphate and a tartrate solution for 1 h at 37°C, followed by counterstaining with a hematoxylin solution. Osteoclasts were determined to be TRAP-positive staining multi-nuclear (>3 nuclei) cells by light microscopy. The total number of TRAP-positive cells and the number of nuclei/TRAP-positive cell in each well were counted.

CD14⁺ monocytes were plated into a calcium phosphate apatite-coated 48-well plate for the bone resorption assay (Cosmo Bio Co., Ltd., Tokyo, Japan) in the same culture conditions as previously described. After a 14-day culture, each well was washed with saline. A solution of 5% sodium hypochlorite was left in the well for 5 min to detach the cells. The pit area in each well was determined by AlphaEaseFC software (Alpha Innotech Corporation, San Leandro, CA, USA).

CD14⁺ monocytes were pre-treated with vehicle control (0.1% DMSO) or SGN (20 µM) for 1 h or otherwise indicated times (time-dependent manner), and then M-CSF (200 ng/ml)/RANKL (100 ng/ml) was added for 0.5, 1 and 6 h. The expression of various proteins was assessed by immunoblotting. The cells were lysed on ice for 15 min by M-PER lysis reagent (Thermo Fisher Scientific, Waltham, MA, USA). The cell lysate was centrifuged at 14,000 × g for 15 min, and the supernatant fraction was collected for immunoblotting. Equivalent amounts of protein were resolved by SDS-PAGE (6–8%) and transferred to polyvinylidene difluoride membranes. After blocking for 1 h in 5% non-fat dry milk in TBS, the membrane was incubated with the primary Ab for 1–16 h (1 h for GAPDH; 16 h for phosphorylated AKT, mTOR, 4EBP1 and p70S6K; and total protein of AKT, mTOR, 4EBP1 and p70S6K). The membrane was then treated with the appropriate peroxidase-conjugated secondary Ab, and the immunoreactive proteins were detected using an ECL kit (Millipore, Billerica, MA, USA) according to the manufacturer’s instructions. All antibodies were obtained from Cell Signaling Technology (Beverly, MA, USA). To quantify immunoblot images on unsaturated bands, densitometric analysis was performed using AlphaEaseFC software.

M-CSF levels were assessed by the M-CSF ELISA kit (R&D Systems). The RANKL and OPG levels in the osteoblasts were quantified using the DuoSet ELISA (R&D Systems).

Data are expressed as means ± SD. Statistical comparisons were carried out using analysis of variance. Significant differences (p<0.05) between two test groups were analyzed by the Student’s t-test.

To explore whether the novel agents inhibit osteoclast development, the effects on osteoclast differentiation by various polyphenols, including resveratrol, piceatannol, pterostilbene, laricitrin, myricetin and SGN were assessed. Treatment of the CD14⁺ monocytes with M-CSF and RANKL caused formation of numerous TRAP-positive multi-nucleated osteoclasts (Fig. 2A), whereas osteoclast differentiation was significantly inhibited by resveratrol, piceatannol, pterostilbene, laricitrin, myricetin and SGN (20 µM). SGN showed the most significant inhibition of osteoclast differentiation in comparison with resveratrol, piceatannol, pterostilbene, laricitrin and myricetin. Consequently, we selected SGN as the model for exploring the mechanism of lung adenocarcinoma-related osteoclastogenesis. First, SGN treatment inhibited osteoclast differentiation in a dose-dependent manner (Fig. 2B). SGN treatment also substantially reduced bone resorption in a dose-dependent manner (Fig. 2C). These findings suggest that SGN inhibits osteoclast differentiation and reduces bone resorption activity in vitro.

Next, we evaluated the effects of SGN on lung adenocarcinoma-mediated osteoclastogenesis (Fig. 3A). The results showed that A549- and CL-1-5-CMs markedly promoted osteoclastogenesis (Fig. 3B). However, the osteoclastic effect induced by human lung adenocarcinoma A549- and CL1-5-CMs was markedly inhibited by SGN. Furthermore, A549- and CL1-5-CMs enhanced the bone resorption activity of the osteoclasts, yet this effect was also reduced by SGN (Fig. 3C). Both results indicate that SGN suppresses human lung adenocarcinoma A549 and CL1-5 cell-mediated osteoclast differentiation and bone resorption activity.

To clarify the molecular mechanism involved in SGN osteoclastogenesis, signaling transduction pathways were investigated. Treatment of M-CSF/RANKL increased the phosphorylation of AKT and mTOR. The downstream targets of mTOR, p70S6K and 4EBP1 were also activated by M-CSF/RANKL, suggesting that the AKT/mTOR signaling transduction pathway was activated during osteoclastogenesis. The osteoclastogenetic inducers (M-CSF and RANKL) of AKT/mTOR expression were suppressed by SGN (Fig. 4A). Furthermore, both A549- and CL1-5-CMs activated AKT/mTOR expression (Fig. 4B). Through treatment with both AKT inhibitor IV and mTOR inhibitor (rapamycin) we further supported the role of AKT/mTOR in osteoclast differentiation and bone resorption (Fig. 4C and D). Both AKT and mTOR inhibitors also suppressed A549 and CL1-5-mediated osteoclasteogenesis (Fig. 4C and D). In short, SGN not only inhibited M-CSF/RANKL-mediated AKT/mTOR activation, yet also reduced the reinforcing effect of lung adenocarcinoma on this signaling transduction pathway (Fig. 4B). These results indicate that SGN inhibits the AKT/mTOR signaling transduction pathway, which is critical in osteoclastogenesis induced by both M-CSF/RANKL and lung cancer cells.

Osteoblasts play a critical role in assisting bone metastasis by increasing M-CSF and RANKL expression and decreasing OPG expression (35). To verify this, we examined the effect of SGN on secretions of M-CSF, RANKL and OPG in osteoblasts by lung adenocarcinoma, A549- and CL1-5-CMs, which markedly increased M-CSF and RANKL expression in the osteoblasts (Fig. 5A and B). On the other hand, A549- and CL1-5-CMs decreased OPG expression in the osteoblasts (Fig. 5C). SGN attenuated the stimulatory effect of lung adenocarcinoma on the expression of M-CSF and RANKL, and reversed the effect of lung adenocarcinoma on the suppression of OPG expression in the osteoblasts (Fig. 5A–C).

The role of SGN in human lung adenocarcinoma-mediated interaction between osteoblasts and osteoclasts was further investigated (Fig. 6A). When compared with unstimulated osteoblasts, human lung adenocarcinoma A549 and CL1-5 cell-treated osteoblasts enhanced osteoclastogenesis. Although lung adenocarcinoma A549 and CL1-5 cells promoted osteoblast-mediated osteoclast differentiation and bone resorption, this effect was altered by SGN (Fig. 6B and C). These findings suggest that SGN inhibits the osteoblasts in lung adenocarci-noma-mediated osteolytic bone metastasis.