p130Cas, Crk-associated substrate (Cas), is an adaptor/scaffold protein that plays a central role in actin cytoskeletal reorganization. We previously reported that p130Cas is not tyrosine-phosphorylated in osteoclasts derived from Src-deficient mice, which are congenitally osteopetrotic, suggesting that p130Cas serves as a downstream molecule of c-Src and is involved in osteoclastic bone resorption. However, the physiological role of p130Cas in osteoclasts has not yet been confirmed because the p130Cas-deficient mice displayed embryonic lethality. Osteoclast-specific p130Cas conditional knockout (p130CasΔOCL–) mice exhibit a high bone mass phenotype caused by defect in multinucleation and cytoskeleton organization causing bone resorption deficiency. Bone marrow cells from p130CasΔOCL– mice were able to differentiate into osteoclasts and wild-type cells in vitro. However, osteoclasts from p130CasΔOCL– mice failed to form actin rings and resorb pits on dentine slices. Although the initial events of osteoclast attachment, such as β3-integrin or Src phosphorylation, were intact, the Rac1 activity that organizes the actin cytoskeleton was reduced, and its distribution was disrupted in p130CasΔOCL– osteoclasts. Dedicator of cytokinesis 5 (Dock5), a Rho family guanine nucleotide exchanger, failed to associate with Src or Pyk2 in osteoclasts in the absence of p130Cas. These results strongly indicate that p130Cas plays pivotal roles in osteoclastic bone resorption. © 2013 American Society for Bone and Mineral Research.
Osteoclasts are terminally differentiated multinucleated cells responsible for physiological and pathological bone resorption and thereby play an essential role in maintaining bone volume and homeostasis.[1, 2] Osteoclastic bone resorption consists of several important processes: the proliferation of osteoclast progenitors; the differentiation of these progenitors into mononuclear prefusion osteoclasts (pOCs); the fusion of pOCs into multinucleated osteoclasts; the attachment of osteoclasts to calcified tissues and the polarization of osteoclasts; the formation in osteoclasts of the ruffled border and the clear zone, followed by the secretion of acids and lysosomal enzymes into the space beneath the ruffled border; and apoptosis.
The molecular events of osteoclast differentiation have been well investigated. Receptor activator of nuclear factor-κB (NF-κB) ligand (RANKL) is an essential cytokine that induces the differentiation of monocyte/macrophage lineage cells into osteoclasts in the presence of macrophage colony-stimulating factor (M-CSF).[1, 2] RANKL induces osteoclastogenesis by the induction of nuclear factor of activated T cells c1 (NFATc1), which transcriptionally regulates most of the osteoclast-specific genes required for bone resorption, including Ctsk (which encodes cathepsin K), matrix metalloproteinase 9 (Mmp-9), and Clcn7 (which encodes chloride channel 7). Recent studies have shown that several transcription factors, such as NF-κB or c-Fos, are necessary for the induction of NFATc1; in contrast, interferon (IFN) regulatory factor-8 (IRF-8), v-maf musculoaponeurotic fibrosarcoma oncogene family protein B (MafB), and B lymphocyte-induced maturation protein 1 (Blimp1) negatively regulate NFATc1 activity.[4-7]
To organize highly polarized cytoplasmic structures such as ruffled borders and clear zones, osteoclasts must adhere to bone surface. The adhesion of osteoclasts to the bone surface induces cytoskeletal reorganization associated with activation. The recognition of extracellular matrix (ECM) components is, therefore, an important step in the initiation of osteoclast function. Several studies have demonstrated that αvβ3 integrins play a central role in osteoclast adhesion. We previously reported that integrin-mediated cell adhesion to ECM molecules, such as vitronectin, fibronectin, or type I collagen, induces the formation of a ring-like structure of F-actin dots at the cell periphery.[8, 9] This ring-like organization, the actin ring, is formed by the assembly of podosomes that precedes the formation of the sealing zone in bone-resorbing osteoclasts in vivo. Several lines of evidence have demonstrated that these actin rings are disrupted by various inhibitors of osteoclastic bone resorption such as calcitonin and bisphosphonates.[8, 10-13] Taken together, these data suggest that the formation of the actin ring in osteoclasts is closely related to the polarity and bone-resorbing activity of these cells.
Src tyrosine kinase is a major regulator of the integrin-mediated adhesions that are required for the efficient adhesion and spreading of a variety of cell types. Src-deficient fibroblasts exhibit reduced adhesion and delayed spreading compared with wild-type cells.[14, 15] Moreover, Src tyrosine kinase has been shown to be involved in the regulation of podosome dynamics. Src scaffolding activity supports the assembly of immature, stationary podosomes, while Src catalytic activity is essential for podosome maturation and turnover. In 1991, Soriano and colleagues reported that the targeted disruption of c-src in mice resulted in osteopetrosis. Osteoclasts from c-Src–deficient mice attached to bone surfaces but showed defects in ruffled-border formation and did not form actin rings due to the inhibition of integrin-mediated intracellular signaling. Furthermore, osteoclasts expressing constitutively active Src formed ectopic podosome rings with high turnover. Thus, Src and its substrates appear to be essential for podosome formation. However, the identity of the molecules downstream of c-Src in osteoclast activation remains unclear.
p130Cas, Crk-associated substrate (Cas), is an adaptor/scaffold protein that plays a central role in actin cytoskeletal reorganization. We previously reported that a 130-kDa protein was tyrosine-phosphorylated in association with actin ring formation and identified this molecule as p130Cas. In osteoclasts derived from c-Src-deficient mice, p130Cas is not tyrosine-phosphorylated, suggesting that p130Cas serves as a downstream molecule of c-Src and is involved in osteoclastic bone resorption. However, the physiological role of p130Cas in osteoclasts has not yet been confirmed in vivo because p130Cas deficiency is embryonic lethal in mice. Herein, to examine the physiological role of p130Cas in osteoclastic bone resorption, we generated osteoclast-specific p130Cas conditional knockout (p130CasΔOCL–) mice and analyzed these mice.
Materials and Methods
Glutathione S-transferase (GST)-RANKL was kindly provided by Oriental Yeast Company, Ltd. (Shiga, Japan). Recombinant human M-CSF was purchased from PeproTech, Inc. (Rocky Hill, NJ, USA). Anti-p130Cas antibody (610271) was purchased from Becton-Dickinson (Franklin Lakes, NJ, USA). Anti-phosphorylated anti-c-Src (2101), anti-c-Src (2108), anti-phosphorylated-Pyk2 (3291), anti-Pyk2 (3291), anti-phosphorylated Paxillin (2541), anti-Paxillin (2542), and anti-Vav3 (2398) antibodies were purchased from Cell Signaling Technology (Beverly, MA, USA). Anti-v-Src (OP07T) antibody was purchased from CalBiochem (San Diego, CA, USA). Anti-Arp3 (sc-15390), anti-Dock5 (sc-98640), anti-FARP2 (sc-135125), anti-β3 integrin (sc-14009), and anti-Talin (sc-15336) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Anti-NFATc1 antibodies and anti-phosphorylated β3 integrin antibodies were purchased from Abcam (Cambridge, MA, USA). Anti-phosphotyrosine (4G10) antibodies were obtained from Upstate Biotechnology (Lake Placid, NY, USA). Anti-Rac1 (23A8) and anti-β-actin antibodies were purchased from Millipore (Billerica, MA, USA) and Sigma-Aldrich (St. Louis, MO, USA), respectively.
For tetracycline-regulated expression of short hairpin RNAs (shRNAs) (Tet-on) that target p130Cas constructs, we designed four variant shRNAs to target the mRNA on a Block-iT RNA interference (RNAi) designer (https://rnaidesigner.invitrogen.com/rnaiexpress) (Invitrogen, Carlsbad, CA, USA). 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 p130Cas shRNA clones by transfecting them into RAW 264.7 (RAW) cells, a mouse macrophage cell line (ATCC, Manassas, VA, USA), that stably expressed 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 p130Cas expression by immunoblotting. Four of the tested constructs were then used to attenuate p130Cas shRNA–expressing cell lines (RAW-teton-shCas#1 and RAW-teton-shCas#2). Tetracycline-resistant pcDNA6/TR-positive clones were used as parent RAW cells. The RAW cells were cultured for 3 days with RANKL (20 ng/mL). Bone marrow cells from 8-week-old male Casflox/flox or CasΔOCL– mice were suspended in 96-well plates for 3 days in the presence of M-CSF (100 ng/mL) in an α-minimal essential medium (α-MEM) containing 10% fetal bovine serum (FBS), 100 units/mL penicillin, and 100 µg/mL streptomycin. The cells were further cultured for 3 days with RANKL (100 ng/mL) in the presence of M-CSF (20 ng/mL). The cultures were fixed with 3.7% formaldehyde in phosphate buffered saline (PBS) for 10 minutes and permeated with 0.1% Triton X-100 in PBS for 5 minutes. The cells were stained for tartrate-resistant acid phosphatase (TRAP), a marker enzyme of osteoclasts, and with 0.3 mM rhodamine-conjugated phalloidin. TRAP-positive multinucleated cells (MNCs) containing more than three nuclei were observed under a microscope and counted as osteoclasts. The distribution of F-actin was detected by fluorescence microscopy (Biorevo, BZ-9000; Keyence, Tokyo, Japan).
The retroviral vector pMX-IRES-EGFP and Plat-E cells were kindly provided by Dr. T Kitamura (Tokyo University, Tokyo, Japan). The full-length and dominant-negative form of p130Cas, which lacks the SH3 domain, with a human influenza hemagglutinin (HA) tag were kindly provided by Dr. Sakai (Division of Metastasis and Invasion Signaling, National Cancer Center Research Institute, Tokyo Japan). Retrovirus packaging was performed by transfecting the plasmids into Plate-E cells using Genejuice (Merck, Darmstadt, Germany) transfection reagent. The supernatants of all retroviral vector-transfected Plat-E cells were used to infect RAW cells in the presence of 8 µg/mL of polybrene.
Immunoprecipitation and Western blotting
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). The protein content was measured with Pierce reagent, according to the manufacturer's protocol. An equal amount of protein was immunoprecipitated with the indicated antibodies for 2 hours, followed by the addition of protein G or A-Sepharose beads and incubation for another hour. The immunoprecipitates were washed three times with lysis buffer, extracted in sodium dodecyl sulfate (SDS) sample buffer, subjected to SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to a polyvinylidene difluoride membrane at 100 V for 1 hour at 4°C. These membranes were then incubated with antibodies at 1:500 to 1:1000 dilutions in 5% dry milk solution plus 0.01% azide overnight at 4°C. Subsequently, the 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).
Generation of osteoclast-specific Cas cKO mice
Knockout mice lacking the osteoclast-specific p130Cas gene (p130CasΔOCL–) were generated by crossing Cathepsin K-Cre knock-in mice and floxed-p130Cas mice (B6;129P2-p130Cas < tm2Homy >).
After euthanasia, the tibias and femurs were dissected, and the soft tissue was removed. All bones were subsequently fixed in a PBS-buffered glutaraldehyde (0.25%)-formalin (4%) fixative (pH 7.4) for 2 days (4°C) and washed with PBS for further studies. The bone mineral densities (BMDs) of the tibias and femurs were measured using peripheral quantitative computed tomography (pQCT; XCT Research SA+, Stratec Medizintechnik GmbH, Pforzheim, Germany), as described.[23, 24] Three-dimensional (3D) reconstruction images of proximal tibias were obtained using focal micro–computed tomography (µCT) (ScanXmate-E090; Comscan, Kanagawa, Japan), as described.[23, 24]
Histological preparation and bone histomorphometry
Tibias and femurs from the control or p130CasΔOCL– mice were embedded in mixtures of methyl methacrylate (MMA) and 2-hydroxyethyl methacrylate (GMA) resins, as described.[23, 24] Sagittal sections (4 µm) of the long bones were prepared. These sections were then stained with Alcian blue or TRAP and counterstained with methyl green. Osteoclasts were designated as the TRAP+ MNCs that contained more than three nuclei and were located on the bone surface. Standard bone histomorphometric analyses were performed in the secondary spongiosa of the tibias, starting at 0.3 mm distal from the proximal growth plate, to exclude the primary spongiosa with an image analyzing system (KS400; Carl Zeiss, Jena, Germany).
For electron microscopy, all bones were fixed in a PBS-buffered 2.5% glutaraldehyde fixative (pH 7.4) for 1 day at 4°C and then postfixed with 1% osmium tetroxide for 2 hours at room temperature. The tissues were then dehydrated in ascending concentrations of ethanol, passed through propylene oxide, and embedded in EPON 812 resin (EMJapan, Tokyo, Japan). The sections were cut on an ultramicrotome and examined on an electron microscope (H-7100; Hitachi, Hitachinaka, Japan).
For immunofluorescence analysis, bone marrow cells from 8-week-old male p130Casflox/flox or p130CasΔOCL– mice were plated onto sterile FBS-coated glass coverslips in the presence of M-CSF plus RANKL for 6 days. The cells were fixed with 4% paraformaldehyde in PBS for 15 minutes, blocked with 5% skim milk in PBS for 15 minutes at room temperature, and incubated with the indicated antibodies (1:100) for 30 minutes at 37°C. After extensive washing, the cells were incubated with Alexa Fluor 488-conjugated anti-rabbit immunoglobulin G (IgG), Alexa Fluor 488-conjugated anti-mouse IgG, or Alexa Fluor 594-conjugated anti-rabbit IgG (dilution 1:10,000; Invitrogen, Carlsbad, CA, USA) for 30 minutes at 37°C. The cells were then washed and mounted in Immunon (Lipshaw, Pittsburgh, PA, USA). The subcellular localization of the indicated proteins was determined by fluorescence microscopy (Biorevo, BZ-9000; Keyence). To visualize the nuclei, the cells were further stained with 4,6-diamidino-2-phenylindole (DAPI).
Rac1 activity measurement
Bone marrow cells from p130Casflox/flox or p130CasΔOCL– mice were cultured in the presence of M-CSF to induce macrophage formation (osteoclast precursors) and then further cultured in the presence of M-CSF and RANKL to induce osteoclast differentiation. Total cell lysates of macrophages and osteoclasts were prepared and then Rac1 activation was measured using the Rac G-Lisa Activation Assay Biochem kit (Cytoskeleton, Denver, CO, USA), according to the manufacturer's protocol.
Coculture and pit formation
Osteoblasts were obtained from the calvariae of newborn C57BL6 mice by digestion with 0.1% collagenase (Wako Pure Chemical, Osaka, Japan) and 0.2% dispase (Godo Shusei, Tokyo, Japan), and bone marrow cells obtained from 8-week-old male p130Casflox/flox or p130CasΔOCL– mice were cocultured in α-MEM containing 10% FBS, 1α,25-dihydroxyvitamin D3 (1 × 10−8 M) (Wako) and prostaglandin E2 (1 × 10−6 M) (Sigma) in 100-mm-diameter dishes coated with collagen gel (Nitta Gelatin, Osaka, Japan). Osteoclasts were removed from the dishes by treating them with 0.2% collagenase (Wako). Mature osteoclasts were seeded onto dentin slices (φ 4 mm; Kyushu-do, Kitakyushu, Japan). The slices were placed in 48-well plates containing α-MEM supplemented with 10% FBS and then cultured for 48 hours. The cells were scraped off the dentine slices, and the slices were stained with Mayer's hematoxylin. The area of each pit was measured from the photographs with Scion image software. In some experiments, mature osteoclasts were seeded onto dentin slices for 18 hours and then were fixed and stained for TRAP and F-actin. Rhodamine-conjugated phalloidin was used to stain for F-actin.
Real-time PCR analysis
Total RNA of osteoclasts or osteoblasts from p130Casflox/flox or p130CasΔOCL– mice was prepared with TRIzol reagent (Invitrogen). Two micrograms of total RNA was synthesized from first-strand cDNA using SuperScript II transcriptase and random primers (Invitrogen). Real-time PCR was performed with SYBR Green PCR Master Mix and the 7300 Real-time PCR system (Applied Biosystems, Foster City, CA, USA), according to the manufacturer's instructions. Samples were matched to a standard curve generated by amplifying serially diluted products under the same PCR conditions. GAPDH expression served as an internal control. The primer sequences were as follows: MMP-9, 5′-ctggacagccagacaactaaag-3′ (forward) and 5′-ctcgcggcaagtcttcagag-3′ (reverse); DC-STAMP, 5′-tcctccatgaacaaacagttccaa-3′ (forward) and 5′-agacgtggtttaggaatgcagctc-3′ (reverse); Cathepsin K, 5′-gggaagcaagcactggataa-3′ (forward) and 5′-ccgagccaagagagcatatc-3′ (reverse); Calcitonin receptor, 5′-ctccaacaaggtgcttggga-3′ (forward) and 5′-gssgcsgtsgstsgtcgcca-3′ (reverse); Carbonic anhydrase II, 5′-ggggatacagcaagcacaacg-3′ (forward) and 5′-ctcttggacgcagctttatca-3′ (reverse); Osteopontin, 5′-cgatgatgacgatggag-3′ (forward) and 5′-tggcatcaggatactgttcatc-3′ (reverse); Osteocalcin, 5′-aagcaggagggcaataaggt-3′ (forward) and 5′-tttgtaggcggtcttcaagc-3′ (reverse); and GAPDH, 5′-aactttggcattgtggaagg-3′ (forward) and 5′-acacattggggtaggaaca-3′ (reverse).
Comparisons were made using factorial ANOVA. When significant F values were detected, Fisher's projected least significant difference (PLSD) post hoc test was performed to compare each of the groups. The data were expressed as the mean ± SD; values of p < 0.05 were considered significant.
p130Cas is involved in actin ring formation in osteoclasts in vitro
We showed that tyrosine phosphorylation of p130Cas is involved in the organization of actin ring formation in osteoclasts and that the tyrosine phosphorylation of p130Cas is markedly reduced in osteoclasts derived from Src-deficient mice. To clarify the functional role of p130Cas in actin ring formation by osteoclasts, we performed p130Cas loss-of-function experiments using different methods. First, we generated RAW cell lines carrying two different tetracycline-inducible shRNAs that specifically target the expression of p130Cas. Knocking down p130Cas expression significantly reduced the number of osteoclasts with actin rings and the number of large multinucleated osteoclasts in RAW-teton-shp130Cas cells (Fig. 1A). Next, we constructed pMX-p130Cas and p130CasΔSH3, the latter of which lacks the SH3 domain (an important binding site of Src kinases). We then infected RAW cells with these constructs. Although the number of osteoclasts was similar, the ectopic expression of dominant-negative p130Cas (p130CasΔSH3), but not p130Cas in RAW cells, resulted in the inhibition of actin ring formation in osteoclasts (Fig. 1B). Smaller multinucleated osteoclasts were observed in p130CasΔSH3-expressing cells (Fig. 1B). These results suggest that p130Cas is required for actin ring formation and efficient multinucleation during osteoclastogenesis.
Osteoclast-specific p130Cas conditional knockout (p130CasΔOCL–) mice exhibit a high bone mass phenotype
Because p130Cas deficiency is embryonic lethal in mice, we generated osteoclast-specific Cas-deficient (Casfox/floxCtskCre/+: p130CasΔOCL–) mice by crossing p130Casflox/flox mice with CtskCre/+ mice to examine the physiological role of p130Cas in osteoclasts in vivo. The soft X-ray images and 3D microcomputed tomography images clearly indicate that the trabecular and cortical bone volume was greatly increased in p130CasΔOCL– male mice that were 8 and 32 weeks of age (Fig. 2A, B; Supplemental Fig. S1A) and in female mice that were 8 and 32 weeks of age (Supplemental Fig. S2A, C). The histomorphometric analysis demonstrated an increase in BMD, trabecular bone volume/tissue volume (BV/TV), trabecular number (TbN), and reduced trabecular separation (Tb.Sp) in p130CasΔOCL– mice compared with those of control mice (Fig. 2C; Supplemental Fig. S1B; Supplemental Fig. S2B, D). BV/TV and trabecular thickness (Tb.Th) was strongly reduced in 32-week-old female control mice compared with those of 8-week-old female control mice due to estrogen deficiency. However, both BV/TV and Tb.Th were almost comparable between 8 week-old and 32-week-old mice (Supplemental Fig. S2B, D). The failure to resorb the primary spongiosa resulted in an accumulation of “cartilaginous bars” encompassed within the bone matrix, as evident in the section stained with Alcian blue (Fig. 2D). Although osteoclast numbers appeared to be increased in the metaphyseal region, the bone marrow was abnormally filled with trabecular bone in p130CasΔOCL– mice (Fig. 2B, E, F; Supplemental Fig. S1A, C). Similar results were also obtained in female mice (Supplemental Fig. S2B, D). Furthermore, electron microscopy showed that whereas osteoclasts from control mice formed ruffled borders, osteoclasts from p130CasΔOCL– mice failed to form a ruffled border (Fig. 2G), suggesting that there was an increase in bone volume associated with a reduced osteoclast function but not osteoclast differentiation.
Bone formation parameters and osteoblast differentiation were not changed in p130CasΔOCL− mice
Inhibition of osteoclastic bone resorption might affect osteoblastic bone formation as a result of bone remodeling.[1, 2] Thus, we next examined osteoclastic parameters in control and p130CasΔOCL− mice. The osteoblastic parameters were unchanged in p130CasΔOCL– mice (Supplemental Fig. S3A, B; Supplemental Fig. S1D). Consistent with these results, the differentiation of osteoblasts induced by bone morphogenic protein 2 (BMP2) or β-glycerophosphate with ascorbic acid from p130CasΔOCL– mice was also similar to the differentiation, including alkaline phosphatase (ALP) activity, mineralization, and osteopontin and osteocalcin expression, observed in control mice (Supplemental Fig. S3C–E). The conditioned media from control or p130CasΔOCL− osteoclast cultures did not affect the BMP2-induced ALP activity of osteoblasts from controls or p130CasΔOCL− mice (Supplemental Fig. S3F). These data confirmed that increased bone was due to impairment of osteoclast function rather than osteoblast function.
Actin ring formation was inhibited in osteoclasts from p130CasΔOCL− mice
We next asked whether p130Cas deficiency impairs bone resorption in vitro. Bone marrow cells were treated with M-CSF for 3 days and then further cultured for 3 days in the presence of M-CSF and RANKL to induce osteoclast differentiation. The expression of p130Cas was diminished in p130CasΔOCL− cells 2 days after RANKL treatment, indicating that the p130Cas gene was absent during osteoclast differentiation (Fig. 3A). Although bone marrow cells from control mice differentiated into well-spread osteoclasts induced by RANKL, bone marrow cells from p130CasΔOCL− mice became TRAP+ MNCs but were incapable of spreading (Fig. 3B). Analysis of F-actin structures in osteoclasts from control and p130CasΔOCL− mice showed that complete and well-defined actin rings were observed in osteoclasts from control mice, whereas osteoclasts from p130CasΔOCL− mice formed diffuse or incomplete actin rings (Fig. 3B, C). Consistent with Fig. 1, p130Cas deficiency resulted in a decrease in the surface area covered by osteoclasts, and the average number of nuclei per osteoclast was significantly decreased (Fig. 3B, C).
To exclude the possibility that osteoclasts from p130CasΔOCL– mice fail to terminally differentiate, we examined the expression levels of mature osteoclast markers, such as DC-STAMP, Mmp-9, Ctsk, Calcitonin receptor, and Carbonic Anhydrase II. There was no difference in the expression levels of these genes when osteoclasts from p130CasΔOCL– mice and control mice were compared (Fig. 3D). Indeed, immunofluorescence analysis also showed that p130Cas was localized at perinuclear and peripheral regions and colocalized with the actin ring in osteoclasts from control mice. In contrast, in osteoclasts from p130CasΔOCL– mice, the expression of p130Cas was hardly detected, and actin rings did not form (Fig. 3E). Talin, another actin-related protein, colocalized with actin rings in control but not p130CasΔOCL– osteoclasts (Fig. 3E). These results further confirm that p130Cas is required for actin ring formation and efficient multinucleation.
CasL was not involved in the compensation for lack of p130Cas in osteoclasts
The Cas family consists of p130Cas, Crk-associated substrate in lymphocyte (CasL)/human enhancer of filamentation 1 (Hef1)/neural precursor cell-expressed, developmentally downregulated gene 9 (Nedd9), and embryonal Fyn substrate/Src-interacting protein (Efs), and these three proteins have high structural homology and conserved binding modulates and effector proteins. Thus, compensatory mechanisms may operate in vivo, and therefore p130CasΔOCL– mice may exhibit moderate bone phenotypes. To exclude the possibility that CasL might partially compensate p130Cas function in osteoclasts from p130CasΔOCL– mice, we crossed p130CasΔOCL– mice with mice lacking CasL, which is also expressed in osteoclasts, to generate p130CasΔOCL–/CasL-deficient mice. However, the bone phenotypes of p130CasΔOCL–/CasL-deficient mice and osteoclasts formed from p130CasΔOCL–/CasL-deficient mice were similar to those observed in p130CasΔOCL– mice in vivo (Supplemental Fig. S4A–C) and in vitro (Supplemental Fig. 4D, E), suggesting that CasL was not involved in the compensation for lack of p130Cas in osteoclasts. Furthermore, we could not rule out the possibility that other molecule(s) collaborating with p130Cas might regulate osteoclastic bone resorption.
p130Cas regulates actin rearrangement as a downstream signaling molecule of integrin
We next examined the actin cytoskeleton in osteoclasts from control or p130CasΔOCL– mice, because the arrangements of actin-based cytoskeletal structures are crucial for osteoclast development. Osteoclasts were plated on sterile FBS-coated glass coverslips in the presence of 10% FBS, and control osteoclasts began to form actin rings, the diameters of which increased with time, at the cell periphery (Fig. 4A). In contrast, the osteoclasts from p130CasΔOCL– mice displayed disturbed actin organization and impaired actin rings (Fig. 4A).
Integrin-mediated mechanical contact with membrane-matrix proteins is important for actin rings. Osteoclasts express several integrins as membrane matrix receptors, including αvβ1, α2β1, and αvβ3. In particular, αvβ3 integrin has been shown to be important for osteoclastic bone resorption. To examine the possibility that the impairment of actin ring formation in p130CasΔOCL– osteoclasts was due to an impairment of the integrin signals, we prepared suspended and adherent osteoclast preparations from control and p130CasΔOCL– mice to examine their integrin activation. The phosphorylation levels of β3 integrin were almost comparable between control and p130CasΔOCL– osteoclasts (Fig. 4B). We next examined whether p130Cas deficiency affected the signaling pathway crucial for osteoclast differentiation and function. The phosphorylation of c-Src, Pyk2, a major adhesion-dependent tyrosine kinase, and Paxillin, which is a key cytoskeleton-organizing adaptor protein, in osteoclasts from p130CasΔOCL– mice was similar to the phosphorylation observed in control mice (Fig. 4C). The expression levels of NFATc1 and Vav3, the latter of which is a Rho family guanine nucleotide exchanger that is essential for stimulated osteoclast activation, were similar to the levels observed in control mice (Fig. 4C).
p130Cas regulates Rac1 distribution and activity in osteoclasts
Rho guanosine triphosphatase (GTPase) organizes the actin cytoskeleton by cycling between active GTP-bound and inactive guanosine diphosphate (GDP)-bound states. In particular, the total lack of Rac in osteoclasts leads to severe osteopetrosis that is similar to that of p130CasΔOCL– mice through the inhibition of cytoskeletal rearrangement and multinucleation, but not of osteoclast differentiation. Therefore, we examined the Rac1 activity of p130CasΔOCL– mouse osteoclasts compared to that of osteoclasts from control mice. The Rac1 activity in osteoclasts from p130CasΔOCL– mice was lower than the Rac1 activity in control mice (Fig. 5A). Furthermore, Rac1 colocalized with dense actin rings in control osteoclasts, while Rac1 was diffusely distributed throughout the cytoplasm (Fig. 5B). Rac1 is known to have profound effects on the regulation of the actin cytoskeleton, especially on Arp2/3 activity. Arp3 colocalized with dense actin rings in control osteoclasts, and Arp3 was diffusely distributed throughout the cytoplasm in p130CasΔOCL– osteoclasts (Fig. 5C). Arp3 also colocalized with Rac1 in control osteoclasts. Again, Arp3 was diffusely distributed throughout the cytoplasm p130CasΔOCL– osteoclasts (Fig. 5C). These results suggest that p130Cas is necessary for actin ring formation through the recruitment of Arp3, as mediated by Rac1 activity.
Association of p130Cas with DOCK5 regulates Rac1 activity
To elucidate the regulatory mechanism of Rac1 activity in osteoclasts, we focused on guanine nucleotide exchanger factors (GEFs), which transform the inactive GDP-bound conformation of a GTPase into the active, GTP-bound conformation. Although there are more than 80 known or predicted Rho family GEFs, Dock5 is one of the candidate molecules known to participate in osteoclast adhesion and spreading. As described previously, Dock5 colocalized with dense actin rings and Rac1 in control cells (Fig. 6A). However, Dock5 was diffusely distributed throughout the cytoplasm in p130CasΔOCL– osteoclasts (Fig. 6A).
Because p130Cas stably associates with Src or Pyk2 and p130Cas makes a complex with Crk and Dock1, which is similar to Dock5, we then investigated how the absence of p130Cas affected the association of Dock5 with Src or Pyk2. When osteoclasts adhered to the culture dishes, Src or Pyk2 was phosphorylated and associated with p130Cas or Dock5 in control osteoclasts (Fig. 6B). Although either Src or Pyk2 was phosphorylated after attachment, Dock5 did not successfully associate with either Src or Pyk2 in p130CasΔOCL– osteoclasts (Fig. 6B). In the absence of p130Cas, Pyk2 failed to interact with Src, suggesting that the Src/Pyk2/p130Cas/Dock5 complex regulates actin ring formation in osteoclasts and that p130Cas is a key molecule in this complex.
Rescue of actin ring and pit formation in osteoclasts from p130CasΔOCL– mice by p130Cas re-expression
Osteoclasts from p130CasΔOCL– mice plated on dentine all showed similar morphological abnormalities, such as less spreading and a lack of actin rings (Fig. 7A). Notably, osteoclasts from p130CasΔOCL– mice did not excavate bone resorptive pits on dentin compared with control osteoclasts (Fig. 7A).
Finally, to determine whether the morphology and function of osteoclasts from p130CasΔOCL– mice could be rescued by the reintroduction of p130Cas, we retrovirally transduced bone marrow macrophages (BMMs) from p130CasΔOCL– mice with full-length p130Cas, p130CasΔSH3 tagged with green fluorescent protein (GFP) or GFP alone. The expression of full-length p130Cas, but not p130CasΔSH3, completely restored osteoclast morphology, actin ring formation, and pit formation, further supporting the hypothesis that p130Cas regulates osteoclast function (Fig. 7B, C). Although full-length p130Cas associated with Pyk2 or Dock5, p130CasΔSH3 failed to associate with neither Pyk2 nor Dock5 (Fig. 7C). These results further indicate that the Src/Pyk2/p130Cas/Dock5 complex regulates osteoclastic bone resorption.
Cell adhesion to the ECM occurs at specialized sites where adhesion receptors, mainly integrins, act as a bridge between the ECM and the actin cytoskeleton via a network of scaffold and signaling proteins. p130Cas functions as an actin-assembling molecule and plays vital roles in cell dynamics by stimulating integrins via surface receptors.[19, 26] Because extensive actin reorganization and cytoskeletal polarization are important steps for osteoclastic bone resorption, we sought to elucidate whether p130Cas may be involved in these processes by using gene transfer techniques and mouse osteoclasts that lack p130Cas. This study revealed that loss of p130Cas function of through shRNA or a dominant-negative form of p130Cas significantly suppressed actin ring formation in vitro and that mice lacking osteoclast p130Cas exhibited an osteopetrotic phenotype as a result of an impairment in bone resorption, even though osteoclasts existed in the metaphyseal region. Indeed, abundant cartilage remnants were detected in the trabeculae of 8-week-old mice. This abundance of cartilage remnants may result from the defective resorption of the growth plate cartilage and represents an important hallmark of osteopetrosis. Consistent with that phenotype defect of p130CasΔOCL– osteoclasts was found associated with a decrease in multinucleation, an irregular ruffled border structure and a cytoskeletal rearrangement inhibition.
Similar to c-Src, p130CasΔOCL– mice present osteoclast dysfunction associated with multinucleation and actin cytoskeleton defect. Nevertheless, osteopetrosis observed in p130CasΔOCL– mice is less severe than that observed in c-Src–deficient mice, which is probably due to the fact that defects in c-Src induce disruption of several pathways implicated in actin organization, including p130Cas, that would not be affected by the defect of p130Cas alone. Moreover, the lack of c-Src expression not only inhibited bone resorption but also stimulated osteoblast differentiation and bone formation, suggesting that osteogenic cells also contribute to the development of osteopetrotic phenotype in c-Src–deficient mice. However, osteoblasts from p130CasΔOCL– mice normally express p130Cas, and BMP2-induced, or β-glycerophosphate with ascorbic acid–induced osteoblast differentiation and bone formation was comparable to control mice in vitro and in vivo. Another possibility might relate to adaptive physiologic mechanisms that compensate for the low-resorbing activities of osteoclasts in vivo but that are not activated in the in vitro model of cultured osteoclasts.
Because a total lack of Rac in osteoclasts leads to severe osteopetrosis similar to that in p130CasΔOCL– mice by impairing osteoclast function and the p130Cas/Crk complex that is required for Rac activation,[39, 40] we examined how the absence of p130Cas might affect the activation of Rac1 in osteoclasts. Expectedly, Rac1 was diffusely distributed in cytoplasm and Rac1 activity in osteoclasts from p130CasΔOCL– mice was reduced compared with osteoclasts from control mice, suggesting that Rac serves as a downstream molecule of p130Cas for osteoclastic bone resorption. Moreover, actin dynamics resembling those found in the podosomes of osteoclasts are currently thought to be associated with the actin polymerization regulated by the Arp2/3 complex. The Arp2/3 complex is activated in other cell types by Cdc42 and Rac, albeit by distinct mechanisms. A previous study has shown that Arp3 is activated by GTPase in misplaced Rac-deficient osteoclasts, suggesting that Rac is required to localize the Arp2/3 complex to the actin ring. Indeed, Arp3 was localized to this ring structure in control osteoclasts, but Arp3 was diffusely distributed throughout the cytoplasm in p130CasΔOCL– osteoclasts. Thus, p130Cas organized the osteoclast cytoskeleton, at least in part, through the mediation of Arp3 by Rac1.
The activating factor of a Rho family consists of two distinct GEFs. One is a catalytic region, which has a Dbl homology (DH)–pleckstrin homology (PH) domain and is where the DH domain promotes the GDP-GTP exchange reaction of a Rho family; several proteins, such as Vav, Tiam, and FARP, have been identified as Rac-specific GEFs. Another group consists of a CDM (Ced5-Dock180-Mbc) family that lacks a DH domain but that instead has a catalytic site called DHR2. Vav3 is known as a Rho family GEF, and Vav3-deficient mice revealed an osteopetrotic phenotype through osteoclast function interference. In contrast to p130CasΔOCL– osteoclasts, Vav3-deficient osteoclasts express early markers of osteoclast differentiation, namely, TRAP, MMP9, and cathepsin K, but they lack expression of the calcitonin receptor, a mature marker of osteoclasts. Indeed, the M-CSF–stimulated and adhesion-stimulated activation of Rac1 is mediated by the Syk-Vav3 axis, suggesting that this signaling pathway is independent of p130Cas.
FARP2 is a Dbl family GEF that is specific for Rac1. The RacGEF domain-deleted forms of FARP2 (ΔGEF-FARP2)-expressing or FARP2-deficient cells were similar in the context of defective osteoclast multinucleation and bone resorbing activities, both of which were observed in p130CasΔOCL– osteoclasts. FARP2 colocalized with dense actin rings and Rac1 in control cells but was diffusely distributed throughout the cytoplasm in p130CasΔOCL– osteoclasts (Supplemental Fig. S5). However, we did not observe an interaction between FARP2 and p130Cas (data not shown), and ΔGEF-FARP2–expressing osteoclasts exhibited enhanced adhesion activities against the ECM, which is an opposite phenotype of p130CasΔOCL– osteoclasts (data not shown). Thus, a possible hypothesis is that p130Cas and FARP2 regulate Rac1 activation at distinct locations in osteoclasts and at different phases of the bone resorbing cycle.
Among the CDM family, Dock5 was strongly induced during RANKL-stimulated osteoclastogenesis, and osteoclasts lacking Dock5 demonstrated impaired actin ring formation that can be explained by perturbed Rac1 activity and the phosphorylation of p130Cas. Dock5 colocalized with dense actin rings and Rac1 in control cells, but Dock5 was diffusely distributed throughout the cytoplasm in p130CasΔOCL– osteoclasts. Indeed, Dock5 associates with Src or Pyk2 in the presence of p130Cas, whereas Dock5 dissociates with Src or Pyk2 in the absence of p130Cas. These results suggest that p130Cas and Dock5 are functionally linked in the formation of the actin ring.
p130Cas serves as a docking protein for multiple SH2 and SH3 domain-containing molecules, and p130Cas is stably associated with Pyk2 via the SH3 domain of p130Cas and the C-terminal domain of Pyk2. Src phosphorylates tyrosine residues of YXXP motif of p130Cas, and the phosphorylation of Pyk2 or p130Cas was markedly reduced in Src-deficient osteoclasts. Sawada and colleagues, reported that p130Cas acts as a primary mechanosensor in which force induces a mechanical extension of this molecule, leading to its tyrosine phosphorylation and activation of downstream signaling. Thus, we examined the relationship between Src, Pyk2, and p130Cas during actin ring formation of osteoclasts. In the absence of p130Cas, both Src and Pyk2 were phosphorylated with levels similar to those of the controls after the cell's attachment to the ECM, but the association of Src with Pyk2 was disrupted in the absence of p130Cas (Figs. 4C, 6B). Immunofluorescence staining also showed that Src colocalized with Pyk2 at focal adhesion contact, even in the absence of p130Cas, when cells were attached to the ECM (Supplemental Fig. S6, arrowheads). Src colocalized with Pyk2 as a ring structure in the periphery of control osteoclasts, but both Src and Pyk2 diffusely distributed throughout the cytoplasm in p130CasΔOCL–osteoclasts (Supplemental Fig. S6). These results may have been observed because the integrin signal phosphorylates Src and Pyk2 independently from p130Cas in the initial step of osteoclast attachment, and p130Cas then connects Src with Pyk2 to maintain the Src/Pyk2/p130Cas/Dock5 complex during actin rearrangement (Fig. 8). Consistent with previous results,[44, 47] p130CasΔSH3 failed to associate with Pyk2 and Dock5 or to rescue actin ring formation in osteoclasts from p130CasΔOCL– mice. The phosphorylation of p130Cas was reduced in p130CasΔSH3-expressing cells, suggesting that Src and Pyk2 functionally collaborate in the phosphorylation of p130Cas.
Taken as a whole, the Src/Pyk2/p130Cas/Dock5 complex plays an important role in osteoclast activation. An in-depth understanding of the regulatory mechanism of osteoclastic bone resorption by p130Cas should help provide a molecular basis for future therapeutic strategies for bone disease.
All authors state that they have no conflicts of interest.
This work was supported by grants from the Ministry of Education, Culture, Sports, Science and Technology of Japan (23390424 to EJ; 23593039 to MK; 24890214 to K. Osawa), grant-in-aid from Chugai Pharmaceutical Co., LTD. (to EJ), and the Joint Usage/Research Center (RIRBM), Hiroshima University (to EJ and H Honda).
Authors' roles: YN, K Osawa, HF, and H Hikiji performed most experiments. YT, KA, MT, and K Ohya performed the radiological assessments, histology preparation, bone histomorphometry, and electron microscopy. YS prepared the samples for electron microscopy and supervised the immunofluorescence staining. HY provided the GST-RANKL. SS and MK provided CasL-deficient mice. SK and H Honda provided Cathepsin K-Cre knock-in mice and floxed-p130Cas mice, respectively. KM and IN reviewed the intermediate draft. EJ designed the study, performed the literature review, prepared the initial and final versions of the paper, and submitted the document. We thank Drs. Shizuko Ichinose and Yuriko Sakamaki in Research Center for Medical and Dental Sciences, Tokyo Medical and Dental University for help with the TEM observation, and Dr. Anne Blangy (CRBM, Montpellier, France) for helpful suggestions for Dock5 immunofluorescence staining.