The human osteosarcoma cell line Saos-2 (11) and the human breast cancer cell line MCF-7 (12) were obtained from the American Type Culture Collection (Rockville, MD, USA). The normal human fibroblast cell line OUMS-24 (13) and the human osteosarcoma cell line HuO9 (14) were established in our department. Saos-2, MCF-7 and HuO9 were cultured in RPMI-1640 (Invitrogen, Carlsbad, CA, USA). OUMS-24 was maintained in Dulbecco's modified EM (Nissui, Tokyo, Japan). All the media were supplemented with 10% FBS (Invitrogen), 100 μg/ml kanamycin (Meiji Seika, Tokyo, Japan) and 0.5 μg/ml Fungizone (Invitrogen). Primary normal human keratinocytes (NHK) were purchased from Kurabo (Osaka, Japan) and were cultured in HuMedia-KB2 (Kurabo) with Human Keratinocyte Growth Supplement (Invitrogen). To determine the effect of S100A7 on cell growth, Saos-2 cells were inoculated at 1×10⁵ cells/well and cultured in a serum-free medium containing indicated recombinant proteins (1 μg/ml). The 10% FBS-containing medium was used as a positive control.

S100A7 (NM002963) cDNA was amplified by PCR and cloned into pGEX6P1 (GE Healthcare Bio-Sciences, Piscataway, NJ, USA). The nucleotide sequence of the cDNA was confirmed by DNA sequencing. Escherichia coli (BL21) cells were transformed by the vector and the recombinant GST-fusion protein was purified by glutathione-agarose affinity chromatography using a Sepharose 4B column (GE Healthcare Bio-Sciences) under conventional conditions. GST was released by cleaving with PreScission protease (GE Healthcare Bio-Sciences) and removed from the final preparations using the Sepharose 4B column. Recombinant GST protein was also prepared and used as a negative control.

Biotinylated recombinant S100A7 or GST protein (5 μg) was incubated with 5 μg of human RAGE Fc (R&D Systems, Minneapolis, MN, USA) and the complex of proteins was pulled down using streptavidine agarose (Invitrogen). After washing the agarose beads, the bound proteins were eluted and fractionated by SDS-PAGE. Streptavidine-HRP (R&D Systems) was used to detect the applied biotinylated S100A7 and GST proteins. Bound RAGE protein was detected by western blotting. Western blot analysis was performed under conventional conditions using 10 μg of protein extracts (15). The antibodies used were as follows: mouse anti-RAGE antibody (R&D Systems), mouse anti-tubulin antibody (Sigma, St. Louis, MO, USA) and HRP-linked anti-mouse IgG (Cell Signaling Technologies, Beverly, MA, USA).

siRNA against human RAGE (siGENOME SMART pool siRNA targeting AGER, Thermo Scientific Dharmacon, Lafayette, CO, USA) was transfected into cells using RNAi MAX reagent (Invitrogen). A control siRNA with no known mammalian homology (siGENONE non-targeting siRNA 1, Thermo Scientific Darmacon) was used as a negative control. The cells were incubated for 72 h and used for various assays.

The transmigration and invasion potentials of the cells were assayed in vitro under conditions similar to those described previously (16). Briefly, 50,000 osteosarcoma cells were inoculated into top wells of pre-coated Boyden chambers (pore size, 8 μm; BD Biosciences, Bedford, MA, USA). Following incubation, the filters were stained and the cells on the bottom surface were counted. For the transmigration assay, the chambers were coated with 2 μg of human fibronectin (Sigma) on the bottom and the incubation time was 8 h. For the invasion assay, 2 μg of human fibronectin and 10 μg of Matrigel (BD Biosciences) were applied onto the upper and lower sides, respectively, and the cells were incubated for 24 h. Mean values were obtained from 10 visual fields.

Saos-2 cells were treated with siRNA for 24 h, washed twice and incubated with 10 ml of Opti-MEM (Invitrogen) supplemented with either S100A7 or GST. The conditioned medium was collected and concentrated into 100 μl by centrifugation using a Centricon (Amicon Ultra-15 Ultracell-10k; Millipore, Billerica, MA, USA) and then mixed with 50 μl of 3X SDS sample buffer. The samples (10 μl) were applied onto an 8% SDS-polyacrylamide slab gel containing 0.5% gelatin (Sigma). Following electrophoresis, the gel was washed with water to remove the SDS, soaked in protein refolding buffer (2.5% Triton X-100, 10% glycerol, 0.5 mM CaCl₂, 100 mM NaCl, 50 mM Tris-HCl/pH 7.4) for 1 h and incubated in 50 mM Tris-HCl/pH 7.4 containing 0.5 mM CaCl₂ for 12 h at 37°C. The gel was then stained with 1% Coomassie brilliant blue (CBB) and further treated with 10% methanol and 5% acetic acid to destain it. The gelatinolytic activity was detected as clear bands on a blue background of the CBB-stained gel. The recombinant matrix metalloproteinase (MMP)-2 and -9 proteins (R&D Systems) were used as positive controls.

Each experiment was repeated a minimum of three times. The results are expressed as the mean ± SD. The Student's t-test was used to compare the two groups. P<0.05 was considered to indicate a statistically significant result.

We screened for S100 proteins that promote the migration of osteosarcoma cells. The cDNA of 20 S100 family members was cloned and GST-fused recombinant proteins were prepared. The proteins were applied to Saos-2 cells in the wound migration assay in vitro. Among the member proteins, S100A7 was identified as the most promising candidate (data not shown).

We confirmed the function of S100A7 in Saos-2 cells by the transmigration assay (Fig. 1). The number of cells that migrated was significantly increased by the S100A7 protein and the extent of the increase was dose-dependent. However, the growth of the Saos-2 cells was not affected by the addition of S100A7 protein (Fig. 2).

S100 proteins are recognized as damage-associated molecular pattern proteins (DAMPs) due to their release from damaged cells under conditions of cell stress (17). Certain DAMPs, including S100 proteins, are thought to be ligands for the multiligand receptor RAGE. RAGE is a type I transmembrane protein belonging to the immunoglobulin superfamily (18) and is involved in a broad range of inflammatory, degenerative and hyperproliferative diseases, including sepsis, rheumatoid arthritis, diabetic nephropathy/angiopathy, atherosclerosis, cancer and neurological disorders (19,20). The receptor is composed of an extracellular region, a hydrophobic transmembrane-spanning domain and a short cytoplasmic tail. The extracellular domain binds a number of ligands, including advanced glycation end products, high-mobility group box 1, S100 family proteins and amyloid fibrils (21). Wolf et al (9) revealed the interaction between S100A7 and RAGE by a competitive ligand binding assay. We performed a pull-down assay using recombinant RAGE and biotinylated S100A7 proteins (Fig. 3A). The RAGE protein was co-precipitated with biotinylated GST-S100A7 but not with biotinylated GST. The measurement of the surface plasmon resonance of these proteins also supported their interaction (data not shown). RAGE is expressed in various types of human cancer, including brain, breast, colon, lung, oral squamous cell and ovarian cancer (22). The expression of RAGE was examined in osteosarcoma cell lines, and the results showed that compared with normal fibroblasts and keratinocytes, which are known to express RAGE protein, osteosarcoma cells (Saos-2 and HuO9) expressed RAGE protein at high levels (Fig. 3B).

The functional involvement of RAGE in the S100A7-induced promotion of migration of osteosarcoma cells was also assessed. The expression of RAGE was successfully downregulated by the specific siRNA in Saos-2 cells (Fig. 4A). The downregulation of RAGE resulted in marked eradication of the S100A7-promoted transmigration of Saos-2 cells (Fig. 4B). Notably, RAGE downregulation suppressed not only the migration of S100A7-stimulated Saos-2 cells but also that of unstimulated cells, suggesting a potential role of endogenous S100A7 and/or other RAGE ligands in transmigration capacity.

The coordination of cell migration and the degradation of matrix proteins is crucial for the invasion of malignant cells. We therefore performed an invasion assay that mimicks the invasion of cancer cells passing through the basement membrane (23). The addition of S100A7 significantly increased the invasion capacity of Saos-2 and HuO9 osteosarcoma cells (Fig. 5A). The downregulation of RAGE by siRNA demonstrates that the invasion of Saos-2 cells also depends on the expression of RAGE (Fig. 5B).

MMPs are known to be involved in the invasion of various types of cancer cells, including osteosarcoma cells (24,25). We analyzed the activity of MMPs by gelatin zymography (Fig. 5C). The application of S100A7 to Saos-2 cells resulted in a marked enhancement of the activity of MMP-2 and MMP-9 (Fig. 5C). Collectively, our results indicate that the S100A7-RAGE signal transdution pathway is involved in the invasion of osteosarcoma cells.

S100A7 is known to promote breast cancer progression (26). However, S100A7 has been reported to act as a negative regulator of breast cancer invasion (27). Our data have demonstrated that S100A7 promoted the migration and invasion of osteosarcoma cells in vitro. A crucial question is whether S100A7 functions to promote the invasion and/or metastasis of osteosarcoma in vivo. Hiratsuka et al (28) revealed that S100A8 and S100A9 indirectly act on the lung to accelerate the migration of primary tumor cells to lung tissues. Andresen et al (29) reported that the expression of S100A7 was detected in the normal lung and that it was increased under pathological conditions. An increased production of S100A7 in the lung may enable circulating cancer cells to settle and grow invasively in the lung. Although further studies are needed, the S100A7-RAGE axis may be a new target for preventing the invasion and/or metastasis of osteosarcoma.