Recombinant mouse TWEAK, human bone morphogenetic protein (BMP)-2, mouse TNF-α, and mouse Fn14-Fc chimera were purchased from R&D Systems Inc. (Minneapolis, MN, USA). Anti-Fn14 monoclonal antibody (mAb) (ITEM-1) was generated in our laboratory as previously described 9. PD98059 and LY294002 were purchased from Calbiochem (San Diego, CA, USA), and Helenalin was purchased from BIOMOL International, L.P. (Plymouth Meeting, PA, USA).

MC3T3-E1, RAW264, ATDC5, and EL4 cells were purchased from Riken Cell Bank (Tsukuba, Japan). MC3T3-E1 cells were maintained in α-minimal essential medium (Invitrogen, Carlsbad, CA, USA) with L-glutamine supplemented with 10% fetal calf serum, 100 μg/ml streptomycin, and 100 units per ml penicillin G solution. Primary osteoblasts derived from healthy human femoral bone from a patient with rheumatoid arthritis were purchased from Cell Applications, Inc. (San Diego, CA, USA) and were maintained using the manufacturer's recommended growth medium and supplements.

Cells (1 × 10⁶) were incubated with anti-Fn14 mAb (ITEM-1) for 1 hour at 4°C, followed by phycoerythrin-labeled rabbit anti-mouse immunoglobulin G (IgG) antibody (BD Pharmingen, San Diego, CA, USA). After a washing with phosphate-buffered saline (PBS), the cells were analyzed on a FACSCalibur (BD Biosciences, San Jose, CA, USA), and the data were analyzed with the WinMDI program (The Scripps Research Institute, La Jolla, CA, USA).

Immunoblotting was performed as previously described 14 with specific antibodies for phosphorylated (Ser536)-NF-κB p65 (Cell Signaling Technology, Inc., Danvers, MA, USA), phosphorylated-Akt and Akt (Becton, Dickinson and Company, Franklin Lakes, NJ, USA), and phosphorylated-Erk p42/44 and Erk p42/44 (Cell Signaling Technology, Inc.).

Cells (1 × 10⁶) were stimulated with 100 ng/ml TWEAK for 48 hours. The amounts of several cytokines in the culture supernatants were then determined by using a cytokine protein array (TranSignal Mouse Cytokine Antibody Array 1.0; Panomics, Inc., Fremont, CA, USA) according to the manufacturer's instructions.

Cells (1 × 10⁶) were stimulated with 100 ng/ml TWEAK in the presence or absence of the indicated inhibitors for 48 hours. The amounts of RANTES (regulated upon activation, healthy T cell expressed and secreted) in the culture supernatants were then determined by using the mouse RANTES enzyme-linked immunosorbent assay (ELISA) kit (R&D Systems Inc.) according to the manufacturer's instructions.

Cells (5 × 10³) were cultured with or without 100 ng/ml TWEAK in the presence or absence of the indicated inhibitors in a flat-bottom 96-well microtiter plate. The number of viable cells in each well was then determined by the WST (water-soluble tetrazolium) assay with the Tetra Color ONE kit (Seikagaku Corporation, Tokyo, Japan) according to the manufacturer's instructions.

Cells (5 × 10³/well) were cultured with 100 ng/ml BMP-2 for 5 days in the presence or absence of the indicated inhibitors in a flat-bottom 96-well plate. Alkaline phosphatase (ALP) activity was determined by the TRACP & ALP double-stain kit (Takara Bio Inc., Shiga, Japan).

Reverse transcription-polymerase chain reaction (PCR) was performed as previously described 14. PCR amplification (osteocalcin: 94°C for 0.5 minutes, 58°C for 0.5 minutes, and 72°C for 1 minute [35 cycles]; β-actin: 94°C for 0.5 minutes, 56°C for 0.5 minutes, and 72°C for 1 minute [35 cycles]) was performed in a DNA engine cycler (MJ Research, Inc., Watertown, MA, USA). The PCR products were separated by 2.0% agarose gel electrophoresis and stained with 0.5 μg/ml ethidium bromide. The primers used were mouse osteocalcin (forward, 5'-GCAGCTTGGTGCACACCTAG-3'; reverse, 5'-GGAGCTGCTGTGACATCCAT-3') and mouse β-actin (forward, 5'-TGGAATCCTGTGGCATCCATGAAAC-3'; reverse, 5'-TAAAACGCAGCTCAGTAACAGTCCG-3').

Total RNA was extracted from cells (1 × 10⁶) or Balb/c mouse tissues. cDNA was then synthesised from 2 μg of total RNA by using the Reverse Transcriptase System (Applied Biosystems, Foster City, CA, USA). Quantitative PCR analysis was performed using the ABI7500 real-time PCR system (Applied Biosystems) according to the manufacturer's instructions. Primers and probes for mouse RANKL (receptor activation of nuclear factor-κB ligand), mouse TWEAK, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were purchased from Applied Biosystems. The ratio of each gene to that of GAPDH was calculated, and the value of 1.0 was assigned to cells that were incubated without transforming growth factor (TGF)-β1 or TWEAK.

Cells were grown on eight-well culture slides with 100 ng/ml TWEAK for 24 or 48 hours, washed with PBS, and fixed with 4% paraformaldehyde. After permeabilisation, slides were stained with goat polyclonal anti-RANKL antibody or control goat IgG (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA) diluted in PBS overnight at 4°C. After extensive washing, slides were incubated with rabbit polyclonal fluorescein isothiocyanate-conjugated anti-goat antibody (1:200) (DakoCytomation, Glostrup, Denmark) for 1 hour at room temperature. Images were acquired with a confocal microscopy (ECLIPSE E800; Nikon Corporation, Tokyo, Japan).

Mouse tail bones obtained from 4- to 6-week-old female Balb/c mice were dissected, fixed, decalcified, and embedded in paraffin according to conventional methods. Sections were then stained with goat anti-TWEAK antibody (Santa Cruz Biotechnology, Inc.) or control goat IgG (Santa Cruz Biotechnology, Inc.) through the use of the peroxidase-based VECTASTAIN ABC kits with DAB (diaminobenzidine) substrate (Vector Laboratories, Burlingame, CA, USA).

Data are represented as the mean ± standard deviation of triplicate samples. Statistical analysis was performed by the unpaired Student's t test. P < 0.05 was considered to be significant.

To investigate the effects of TWEAK on osteoblastic cells, we first examined whether a functional TWEAK receptor, Fn14, was expressed on MC3T3-E1 cells. As shown in Figure 1a, fluorescence-activated cell sorting analysis with anti-Fn14 mAb showed that Fn14 was significantly expressed on the surface of MC3T3-E1 cells. Moreover, TWEAK induced rapid phosphorylation of the NF-κB p65 subunit in MC3T3-E1 cells, which was abrogated by the addition of mouse Fn14-Fc chimera (Figure 1b).

Because MC3T3-E1 cells expressed Fn14, we next examined the effects of TWEAK on cytokine production by MC3T3-E1 cells by using a cytokine protein array. We found that TWEAK exclusively induced RANTES in this assay (Figure 2a). Using an ELISA, we confirmed that TWEAK significantly induced RANTES production by MC3T3-E1 cells in a dose-dependent manner and that the blockade of Fn14 with Fn14-Fc chimera inhibited TWEAK-induced RANTES production (Figure 2b). TWEAK did not affect cellular viability at the doses used in the experiments (Figure 2c).

To clarify intracellular signaling pathways involved in TWEAK-induced RANTES production in MC3T3-E1 cells, we examined the effects of several inhibitors on TWEAK-induced RANTES production in MC3T3-E1 cells. Helenalin (an inhibitor of NF-κB) did not affect the TWEAK-induced RANTES production by MC3T3-E1 cells, and PD98059 (an inhibitor of the MAPK Erk pathway) marginally affected the TWEAK-induced RANTES production by MC3T3-E1 cells (Figure 3a). However, LY294002, an inhibitor of PI3K, significantly suppressed TWEAK-induced RANTES production (Figure 3a). Consistent with these findings, TWEAK induced phosphorylation of Akt, a downstream target of PI3K (Figure 3b). Phosphorylation of NF-κB subunit p65 and Erk p42/44 was also induced by TWEAK in MC3T3-E1 cells (Figures 1b and 3b). These results indicated that TWEAK activates several intracellular signaling pathways in MC3T3-E1 cells but that the PI3K-Akt pathway is involved in TWEAK-induced RANTES production in MC3T3-E1 cells.

BMPs are important members of the TGF-β superfamily and control osteogenesis 1516, and BMP-2 can induce differentiation of preosteoblastic MC3T3-E1 cells into osteoblastic phenotype as demonstrated by increased ALP activity and osteocalcin expression 17. We then investigated whether TWEAK affected BMP-2-induced differentiation of MC3T3-E1 cells. MC3T3-E1 cells cultured for 5 days in the presence of BMP-2 significantly developed ALP activity, which was inhibited by the addition of TWEAK (Figure 4a). Osteocalcin mRNA expression induced by BMP-2 was also inhibited by TWEAK (Figure 4b). The inhibitory effect of TWEAK on BMP-2-induced ALP activity in MC3T3-E1 cells was abrogated by the addition of Fn14-Fc chimera (Figure 4c). Furthermore, the inhibitory effect of TWEAK on BMP-2-induced ALP activity in MC3T3-E1 cells was also abrogated by the addition of PD98059 (Figure 4d). We confirmed that BMP-2 did not affect cell surface Fn14 expression on MC3T3-E1 cells (Figure 4e). These results indicate that TWEAK inhibited BMP-2-induced osteoblastic differentiation in MC3T3-E1 cells through the Fn14 and MAPK Erk pathways.

RANKL is an important ligand expressed on osteoblasts for osteoclast differentiation 18. We then investigated whether TWEAK affected RANKL expression in MC3T3-E1 cells. TWEAK-induced RANKL mRNA expression in MC3T3-E1 cells peaked at 120 minutes after the stimulation and was inhibited by Fn14-Fc chimera (Figure 5a). TWEAK-induced RANKL mRNA expression was also abrogated by PD98059, but not LY294002 (Figure 5a). An immunofluorescence study also showed that TWEAK induced RANKL expression at the protein level (Figure 5b). These results indicate that TWEAK induced RANKL expression in MC3T3-E1 cells through the Fn14 and MAPK Erk pathways.

Because we found that TWEAK affected various osteoblast functions, we then investigated whether TWEAK mRNA was expressed in various mouse tissues and mouse bone cell lines. We found that TWEAK mRNA expression was relatively high in the lung, liver, and bone marrow (Figure 6a). We also found that TWEAK mRNA was expressed in MC3T3-E1 cells and RAW264 cells, a mouse monocyte/osteoclast cell line (Figure 6b). Mouse chondrocyte cell line ATDC5 and mouse thymic tumour cell line EL4 did not express TWEAK mRNA (Figure 6b). Consistent with these findings, an immunohistochemical analysis with anti-TWEAK antibody showed that TWEAK immunoreactivity was detected in morphologically osteoblast- or osteocyte-like cells, but not in chondrocytes, in the mouse bone tissue (Figure 6c). Staining with control IgG did not show any immunoreactivity in the bone tissue (data not shown).

Finally, we investigated whether TWEAK had some effects on primary human osteoblasts. As shown in Figure 7a, cultured osteoblasts derived from healthy human femoral bone from a patient with rheumatoid arthritis expressed Fn14. TWEAK induced RANTES production, which was blocked with Fn14-Fc chimera (Figure 7b). We also found that TWEAK induced RANKL protein expression in human osteoblasts (Figure 7c).

These results suggest that TWEAK effects on osteoblasts may be relevant to the physiology of the bone.