The Association of Notch2 and NF-κB Accelerates RANKL-Induced Osteoclastogenesis^▿

Animals.

This study was approved by the Council on Animal Care and Gene Transfer at Fukuoka Dental College. Male ddY mice (3 to 5 weeks old) were purchased from the Kyudo (Tosu, Saga, Japan). Mice were housed at the Animal Center of Fukuoka Dental College.

Reagents.

Recombinant human RANKL and M-CSF were purchased from PeproTech Inc. (Rocky Hill, NJ). γ-Secretase inhibitor X (L685,458) (GSI), a selective inhibitor of Notch signaling, was purchased from Calbiochem (La, Jolla, CA). Cleaved Notch1 (Val1744) and anti-Notch1 antibodies (C-20) were purchased from Cell Signaling (Beverly, MA) and Santa Cruz Biotechnology (Santa Cruz, CA), respectively. The anti-Notch2 antibody (C651.6DbHN) was purchased from Developmental Studies Hybridoma Bank (University of Iowa, Iowa City). The anti-RBPJκ antibody was purchased from the Institute of Immunology (Tokyo, Japan). Anti-NFATc1, -p65, -p50, -RANK, -IκBα, -histone deacetylase 1 (HDAC1), and -c-Fos antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). FLAG-fused human Jagged1 (Jagged-1 FL) was cloned and then purified as described previously (43, 62).

Cell culture.

Bone marrow-derived macrophages (BMMs) were prepared as osteoclast precursors from 3- to 5-week-old male ddY mice. Bone marrow cells obtained from the mouse tibia were suspended in 60-mm-diameter dishes for 16 h in the presence of M-CSF (50 ng/ml) in α-minimal essential medium containing 10% fetal bovine serum. Then, nonadherent cells were harvested and further cultured for 2 days with M-CSF (50 ng/ml). The adherent cells, most of which expressed macrophage-specific antigens such as Mac-1, Moma-2, and F4/80, were used as BMMs. BMMs were cultured for 3 days with RANKL (50 ng/ml). The macrophage cell line RAW 264.7 (RAW cells) (ATCC, Manassas, VA) was cultured for 3 days with RANKL (20 ng/ml). Cultures were fixed with 3.7% formaldehyde, and osteoclasts were detected by staining for tartrate-resistant acid phosphatase (TRAP). TRAP-positive multinucleated cells (MNCs) containing more than three nuclei were observed under a microscope and counted as osteoclasts. Soluble or immobilized Jagged1 was used as described elsewhere (43, 62).

Gene chip analysis.

RAW cells were stimulated for 3 days with or without RANKL (20 ng/ml). Total RNA from untreated or RANKL-treated RAW cells was extracted from three independent experiments using Trizol (Invitrogen, Carlsbad, CA). Total RNA (15 μg) was used for cDNA synthesis by reverse transcription followed by synthesis of biotinylated cRNA via in vitro transcription. After cRNA fragmentation, hybridization with a mouse U74Av2 GeneChip (Affymetrix, Santa Clara, CA) displaying probes for 12,000 mouse genes/expressed sequence tags was performed according to the manufacturer's protocol. Chips were washed, stained with streptavidin-phycoerythrin, and analyzed using a scanner and accompanying gene expression software GeneSpring (Agilent Technologies, Santa Clara, CA).

PCR analysis.

For reverse transcription-PCR (RT-PCR), total RNA prepared using Trizol was amplified using Superscript II and Taq polymerase (Invitrogen). Primer sequences were as described by Yamada et al. (62). The PCR conditions were as follows: 94°C (3 min), 60°C (2 min), and 72°C (3 min) for the primary cycle and 94°C (45 s), 60°C (1 min), and 72°C (1.5 min) for the following 30 cycles. The fluorescence of each PCR product was detected using an image analyzer (fluoro image analyzer FLA-2000F; Fuji Film, Tokyo, Japan).

Coprecipitation and immunoblotting.

Cells were lysed in TNT buffer (20 mM Tris-HCl, pH 7.5, 200 mM NaCl, 1% Triton X-100, 1 mM dithiothreitol) containing protease inhibitors (Roche, Basel, Switzerland). Cytosolic, membrane, and nuclear fraction proteins were extracted from cells using a Calbiochem Proteo Extract subcellular proteosome extraction kit (Merck KGaA, Darmstadt, Germany) by following the manufacturer's protocol. For coprecipitation experiments, nuclear extracts were incubated for 6 h at 4°C with anti-Notch2 antibody coupled to protein A/G-Sepharose beads. The immune complex was extensively washed with TNT buffer, and samples were boiled and analyzed by immunoblotting.

Protein content was measured with Pierce reagent by following the manufacturer's protocol. Twenty micrograms of protein was subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to a polyvinylidene difluoride membrane using 100 V for 1 h at 4°C. These membranes were then incubated with antibodies at 1:500 to 1:1,000 dilutions in 5% dry milk solution plus 0.01% azide overnight at 4°C. Subsequently, blots were washed in TTBS (10 mM Tris-HCl, 50 mM NaCl, 0.25% Tween 20) and incubated with a horseradish peroxidase-conjugated secondary antibody. The immunoreactive proteins were visualized using enhanced chemiluminescence (Amersham).

Immunofluorescence staining.

Cells were fixed with 4% paraformaldehyde in phosphate-buffered saline and incubated with 0.3% H₂O₂ in methanol for 30 min followed by 10% horse serum for 1 h. The slides were then incubated with primary antibodies for Notch1 and Notch2 overnight at 4°C. These primary antibodies were detected by goat anti-rabbit Alexa 488 and goat anti-rat Alexa 568 (Molecular Probes) and visualized by fluorescence microscopy (TMD 300; Nikon, Tokyo, Japan). After fluorescence analysis, cells were stained with TRAP and nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI).

shRNA cell lines.

For tetracycline-regulated expression of short hairpin RNAs (shRNAs) (Tet-on) that target Notch1 and Notch2 constructs, we designed four variant shRNAs to target the mRNA on a Block-iT RNA interference (RNAi) designer (https://rnaidesigner.invitrogen.com/rnaiexpress) (Invitrogen). First, double-stranded oligonucleotides that encoded the shRNAs were cloned into the pENTR/H1/TO vector using a Block-iT inducible H1 RNAi entry vector kit by following the manufacturer's protocol (Invitrogen). We screened the pENTR/H1/TO Notch1 and Notch2 shRNA clones by transfecting them into RAW cells stably expressing pcDNA6/TR. Transfection was performed using Lipofectamine 2000 (Invitrogen), and cells were maintained in 50 μg/ml blasticidin or 100 μg/ml zeocin (Invitrogen) for selection. Resistant cell lines were evaluated for their ability to attenuate Notch1 and Notch2 expression by RT-PCR and immunoblotting. Four of the tested constructs were then used to attenuate Notch1 and Notch2 shRNA-expressing cell lines (RAW-teton-shNotch1 and RAW-teton-shNotch2). Tetracycline-resistant pcDNA6/TR-positive clones were used as parent RAW cells.

RNAi assay.

The RNAi assay was described previously (32). Three small interference RNA (siRNA) molecules targeting RBPJκ (siRBPJκ) (Rbpsuh Stealth Select 3 RNAi [MSS208567, MSS208565, and MSS208566]) were purchased from Invitrogen. We used scrambled sequences as a negative control. We used immunoblotting to validate the silencing effect on RBPJκ.

DNA constructs.

Constructs that activate Notch1 (N1IC^ΔOP) and Notch2 (N2IC^ΔOP) with a myc tag in pCS2+6MT have been previously described (1, 16). The retroviral vector pMX-IRES-EGFP and Plat-E cells were kindly provided by T. Kitamura (Tokyo University, Japan) (27, 44). The constructs were subcloned from the EcoRI/XhoI site of N1IC^ΔOP/pCS2+6MT and N2IC^ΔOP/pCS2+6MT into the EcoRI/XhoI sites of pMX-IRES-EGFP, generating pMX-N1IC^ΔOP-IRES-EGFP and pMX-N2IC^ΔOP-IRES-EGFP, respectively. Retrovirus packaging was performed by transfecting the plasmids into Plat-E cells using Lipofectamine 2000 (Invitrogen). Supernatants of pMX-N1IC^ΔOP-IRES-EGFP- and pMX-N2IC^ΔOP-IRES-EGFP-transfected Plat-E cell culture medium were used to infect primary BMMs in the presence of 8 μg/ml Polybrene. The pCMX-N/RBP-Jκ construct was described elsewhere (20).

ChIP.

Chromatin immunoprecipitation (ChIP) was performed with a ChIP assay kit (Upstate Biotechnology) according to the manufacturer's instructions using antibodies against NFATc1, p50, p65, c-Fos, Notch2, RBPJκ, and normal immunoglobulin G. The purified DNA was analyzed by PCR using primers that detect sequences containing the NFATc1-P1 promoter, 5′-CCGGGACGCCCATGCAATCTGTTAGTAATT-3′ (sense) and 5′-GCGGGTGCCCTGAGAAAGCTACTCTCCCTT-3′ (antisense), and the Hes-1 promoter, 5′-CACATCCTCTTTACCTTGTTCCCTC-3′ (sense) and 5′-CCCTTTGCTAACTCTTTCCTCTGG-3′ (antisense).

Luciferase reporter assay.

pNFATc1P1 0.8 kb-Luc (pNFATc1-pr-luc) was constructed by inserting the pGL3 basic promoter vector according to the method of Asagiri et al. (4). The RBPJκ site mutation in the pNFATc1P1 promoter (pNFATc1-RBPJκ mut-pr-luc) was constructed by PCR amplification using the following primers: 5′-ATTTAGCGGGAGGGGAATTTCC-3′ (sense) and 5′-GGAAATTCCCCTCCCGCTAAAT-3′ (antisense); the following primers were used to generate a mutation of the NF-κΒ binding site in the pNFATc1P1 promoter (pNFATc1-RBPJκ mut-pr-luc): 5′-CGGGATGGGAATTTCGTTACTCCC-3′ (sense) and 5′-GGGAGTAACGAAATTCCCATCCCG-3′ (antisense). The plasmids were transfected into RAW cells using Lipofectamine 2000 (Invitrogen), and luciferase activity was measured with the dual-luciferase reporter assay system (Promega, Madison, WI).

Data analysis.

Data are shown as mean values ± standard errors of means (SEM; n = number of culture wells). All experiments were performed at least three times, and similar results were obtained. Statistical differences were analyzed using one-way analysis of variance, and P values less than 0.05 were considered significant.

Notch is expressed and activated during osteoclast differentiation.

By microarray screening to identify mRNAs that were differentially expressed in untreated versus RANKL-treated RAW cells, a macrophage cell line that differentiates into osteoclasts in response to RANKL, we found that several Notch family members and Notch targets were upregulated following RANKL stimulation, concomitant with the expression of osteoclast differentiation markers (Table 1). We confirmed these results for RANKL-stimulated BMMs using semiquantitative RT-PCR and found that expression of Notch2 mRNA was increased within 24 h after RANKL stimulation (Fig. 1A) together with a weak induction of Notch1. The expression of the Hes-1 gene, a Notch target gene, was similar to that of Notch2, suggesting that Notch signaling was activated during osteoclast differentiation induced by RANKL. On the other hand, other Notch family members such as Notch3 and Notch4 were undetectable in these cells (Fig. 1A). We next examined the expression and cellular localization of Notch1 and Notch2 proteins in RANKL-treated BMMs undergoing terminal differentiation of multinucleated osteoclasts. By Western blotting and immunofluorescence, we detected the presence of the intracellular active Notch2 form, N2ICD, at 24 h, and N2ICD was still present after 3 days of RANKL treatment (Fig. 1B and C). To further demonstrate that Notch2 was translocated into the nucleus following RANKL stimulation, we obtained membrane, cytoplasmic, and nuclear fractions from BMM cells and performed Western blot analysis with anti-Notch2 antibody. N2ICD was observed only in the nuclear fractions after RANKL stimulation (Fig. 1D). RANKL, IκBα, and HDAC1 were used to determine the purities of the different cellular fractions. In addition, total levels of Notch2 protein in both membrane and nuclear fractions were significantly increased after the RANKL stimulation compared to those in untreated cells. These results indicated that RANKL induced Notch2 protein expression and activated Notch signaling during osteoclastogenesis.

GSI suppresses RANKL-induced osteoclastogenesis.

To further examine the role of Notch2 in osteoclast differentiation, we added GSI, which blocks Notch signaling, to bone marrow cell cultures together with RANKL and M-CSF. GSI inhibited RANKL-induced osteoclastogenesis from bone marrow cells and the expression of Hes-1 in a dose-dependent manner (Fig. 2A). It was previously shown that osteoclast differentiation consists of at least two steps: (i) the proliferation of osteoclast progenitors and their differentiation into osteoclast precursors (BMMs) induced by M-SCF and (ii) the subsequent differentiation of osteoclast precursors into osteoclasts induced by RANKL (54). We examined which step was affected by GSI treatment. Incubation with GSI together with M-CSF for 3 days and then further culture with RANKL gave rise to the generation of numerous osteoclasts (Fig. 2B), whereas expression of Hes-1 did not change (Fig. 2B). However, when added to the culture together with RANKL, GSI strongly inhibited osteoclast formation together with the expression of Hes-1 (Fig. 2C). These findings suggest that Notch signaling is essential during the second wave of osteoclastogenesis, which is mediated by RANKL.

Inhibition of RANKL-induced osteoclastogenesis by silencing Notch2 mRNA.

We next examined whether specific loss of function of Notch1 or Notch2 homologues affects osteoclast formation. Since both Notch1- and Notch2-deficient mice die at the early embryonic stage (37, 53), there is little information about the physiological role of these proteins in bone development. Thus, we generated RAW cell lines carrying tetracycline-inducible shRNA that specifically targets the expression of Notch1 or Notch2 (RAW-teton-shNotch1 or -shNotch2). We confirmed that endogenous Notch1 and Notch2 expression was suppressed in RAW-teton-shNotch1 and -shNotch2 cells in response to tetracycline (Fig. 3A and B, right). We found that RANKL-induced differentiation was not affected in RAW-shNotch1 cells compared to the control cells (Fig. 3A). In contrast, knocking down Notch2 expression significantly inhibited osteoclast formation in RAW-teton-shNotch2 cells (Fig. 3B).

The active form of Notch2 positively regulates RANKL-induced osteoclastogenesis.

To further investigate the role of Notch signaling, we examined the effects of expressing the active forms of Notch1 or Notch2 in BMMs. We constructed pMX-N1IC^ΔOP-IRES-GFP and pMX-N2IC^ΔOP-IRES-EGFP, which were engineered to express both active forms of Notch1 or Notch2 and green fluorescence protein (GFP) and infected BMMs with the different constructs. Infection efficiency was monitored by measuring GFP expression (Fig. 4A), and expression of ectopic Notch1 or Notch2 was detected by immunostaining with the anti-Notch1 and -Notch2 intracellular domains (ICDs) or the anti-myc tag antibodies (Fig. 4B). Ectopic expression of activated Notch2 but not Notch1 in BMMs resulted in efficient induction of osteoclast differentiation in a RANKL dose-dependent manner (Fig. 4C).

Expression of Notch ligands in osteoclast precursors.

Previous data showed RANKL induced not only Notch2 expression but also activation of Notch signaling (Fig. 1B to D), thus suggesting that RANKL regulates Notch ligand expression. Therefore, we examined the expression of Notch ligands in RANKL-treated BMMs and found that Jagged1 expression gradually increased from day 1 (1.23-fold) to day 3 (3.21-fold) of treatment (Fig. 5A). To determine whether Jagged1 was responsible for activating Notch2 during osteoclast differentiation, we incubated BMMs or RAW-teton-shNotch2 cells with either soluble recombinant FLAG-Jagged1 or Jagged1 immobilized on the culture plate. We found that Jagged1 enhanced RANKL-induced osteoclastogenesis in both culture conditions (Fig. 5B). As a control, addition of tetracycline inhibited RANKL-dependent osteoclast formation from RAW-teton-shNotch2 (Fig. 5C), but not RAW-teton-shNotch1, cells (data not shown) incubated with FLAG-Jagged1. Even though we cannot exclude the possibility that other RANKL-induced effects also affect Notch activation, these results strongly suggest that Jagged1 activates Notch signaling via Notch2 and that this activation enhanced the osteoclast formation induced by RANKL.

Notch2 modulates RANKL-induced osteoclastogenesis through NFATc1.

To investigate the molecular mechanisms by which Notch signaling enhances osteoclast formation induced by RANKL, we examined the expression of NFATc1, a master regulator of osteoclastogenesis (4, 55). NFATc1 was induced in BMMs at 24 h after RANKL stimulation (Fig. 6A). Activation of Notch signaling by Jagged1 or by ectopic expression of N2ICD enhances NFATc1 expression (Fig. 6A). Conversely, inhibition of Notch signaling by GSI or shRNA for Notch2 strongly suppressed NFATc1 expression (Fig. 6A). These results suggest that Notch modulates RANKL-induced NFATc1 expression. Analysis of the NFATc1 promoter revealed the presence of a putative RBPJκ binding site overlapping an NF-κB site (Fig. 6B). ChIP analysis showed that Notch2 was recruited to the NFATc1 promoter after 30 min of RANKL stimulation, reached a maximum level at 1 h, and declined after 6 h of treatment (Fig. 6C). In contrast, we found that RBPJκ was in the NFATc1 promoter before RANKL stimulation and that RBPJκ was undetectable at 30 min and 3 h after RANKL stimulation, coinciding with Notch2 recruitment. It has been recently reported that RANKL-induced NF-κB activity is important for the initial induction of NFATc1 in the early stage of osteoclastogenesis and that c-Fos recruitment to the NFATc1 promoter is important in the autoamplification phase of NFATc1 induction (4). Consistent with this, we found that both p50 and p65 NF-κB subunits were recruited to the NFATc1 promoter within 30 min after RANKL stimulation while NFATc1 and c-Fos were recruited to the promoter after 6 h (Fig. 6C). These data suggest the possibility that NF-κB and Notch2 cooperate in the activation of NFATc1 expression. In agreement with this, coimmunoprecipitation experiments demonstrated that Notch2 physically interacts with p65 in the nucleus in a RANKL-dependent manner (Fig. 6D). Next, we generated luciferase reporter constructs containing specific mutations in the NF-κB (NF-κB-mut) or RBPJκ (RBPJκ-mut) binding sites of the NFATc1 promoter. Cotransfection of Notch2 with p65 enhanced NFATc1 promoter activity compared with transfection with Notch2 alone in both the wild-type and the RBPJκ-mut constructs (Fig. 6E). In contrast, Notch2 alone or in combination with p65 failed to activate the NF-κB-mut promoter (Fig. 6E), suggesting that p65 plays an essential role in Notch2-mediated activation of the NFATc1 gene. To further study the putative role of RBPJκ in the regulation of NFATc1 expression, we transfected RBPJκ together with Notch2 or Notch2/p65 and then measured luciferase activity. In both conditions RBPJκ strongly suppressed NFATc1 promoter activity (Fig. 6F), suggesting that RBPJκ may compete with p65 for DNA binding. We next examined the binding of endogenous p65 and RBPJκ to the NFATc1 promoter in the absence of Notch2 using RAW-teton-shNotch2 cells. Interestingly, we found that RBPJκ constitutively bound the NFATc1 promoter in these conditions (Fig. 6G). In a reciprocal experiment, knocking down RBPJκ expression using specific siRNA slightly enhanced binding of p65 and Notch2 to the NFATc1 promoter (Fig. 6H). These results indicate that NFATc1 is downstream of both Notch2 and NF-κB during RANKL-mediated osteoclast differentiation.

Altogether, these results demonstrate an essential role for Notch2 in the induction of the terminal osteoclast differentiation.