Syk deficiency impairs osteoclast function but not differentiation in vitro. (A) BMMs derived from Syk^+/? or Syk^−/− chimeras were cultured with RANKL and MCSF for 5 d. Cells were stained for TRAP activity (red reaction product). (B) BMMs derived from Syk^+/? or Syk^−/− chimeras were cultured with RANKL and MCSF on dentin for 6 d. Actin ring formation was determined by immunofluorescence after FITC-phalloidin staining. (C) After 6 d, Syk^+/? or Syk^−/− osteoclasts were removed, and the dentin was stained with Coomassie brilliant blue to visualize resorption lacunae. (D) Mature Syk^−/− osteoclasts transduced with either WT Syk, kinase-inactive Syk^K402R, or pMX vector were generated on dentin for 5 d with RANKL and MCSF. Cells were stained for TRAP activity and/or with FITC-phalloidin. The expression of Syk was determined by immunoblot assay of TCL Syk^−/−. (E) BMMs were retrovirally transduced with WT Syk or empty vector. Syk^+/? and Syk^−/− cells serve as positive and negative controls, respectively. The cells were placed on bone slices and differentiated into osteoclasts by exposure to RANKL and MCSF. After 5 d, the osteoclasts were removed. Resorption pits were visualized by incubation of the specimen with 20 μg/ml peroxidase-conjugated wheat germ agglutinin. (F) Medium was collected and assayed for CTx concentration. Actin serves as a loading control. Error bars represent SD. *, P < 0.01; **, P < 0.001. (G) Syk^+/? or Syk^−/− BMMs were cultured with MCSF and RANKL for 3 d. The cells were then lifted and either maintained in suspension (S) or plated on vitronectin (A) for 30 min. GTP-bound Rac1 and total Rac1 expression were determined, and the immunoblot was densitometrically quantitated.

Bone density is increased in Syk^−/− embryos. (A) Skeletons of 18.5-d Syk^+/? and Syk^−/− embryos were fixed in 95% ethanol and stained with Alcian blue and Alizarin red. Syk^−/− forelimb and hindlimb bones are denser as indicated by increased Alizarin staining (arrows). (B) TRAP-stained (red reaction product) histological sections of day 18.5 Syk^+/? and Syk^−/− embryo femurs. Arrows indicate the trabecular network in Syk^−/− but not Syk^+/? marrow space.

Syk, c-Src, and αvβ3 form a complex in osteoclasts. (A) β3^−/− BMMs retrovirally transduced with hβ3WT were cultured with RANKL and MCSF for 5 d. Total cell lysate (TCL) was immunoprecipitated with anti-Syk or ant–c-Src antibodies or irrelevant IgG. Immunoprecipitates and TCLs were probed by Western blotting for β3-integrin, Syk, and c-Src content. (B and C) WT BMMs were cultured with MCSF and RANKL for 3 d. The cells were then lifted and either maintained in suspension (S) or plated on vitronectin (A) for 30 min. β3-integrin immunoprecipitates from each group of cells were probed by immunoblotting for β3-integrin and Syk content (B) or for c-Src content (C). (D) WT BMMs were cultured with MCSF alone or MCSF and RANKL for 3 or 5 d. TCL was immunoprecipitated with anti-Syk antibody followed by c-Src and Syk immunoblotting. (E) TCL from WT or β3^−/− osteoclasts was immunoprecipitated with anti-Syk (top) or anti–c-Src (bottom) antibodies. Immunoprecipitates and TCLs were immunoblotted for c-Src and Syk content. (F) β3^−/− BMMs were retrovirally transduced with hβ3 or pMX vector. The transductants were cultured in RANKL and MCSF for 5 d. Syk immunoprecipitates and TCLs were then immunoblotted for c-Src, Syk, or β3-integrin content. (G) WT preosteoclasts, which were generated by culturing BMMs in MCSF and RANKL for 3 d, were lifted and either plated on vitronectin (A) or maintained in suspension (S) for 30 min. c-Src and Syk content in Syk immunoprecipitates was determined by immunoblotting. (H) Suspended Syk^+/? or Syk^−/− preosteoclasts were plated on vitronectin. The percentage of spread cells was determined after 5 and 10 min. *, P < 0.001 versus Syk^+/?. Error bars represent SD.

c-Src mediates αvβ3-induced Syk activation in osteoclasts. (A and B) WT BMMs were cultured in RANKL and MCSF for 3 d. The preosteoclasts were lifted and preincubated in 10 μM piceatannol (Pice) or DMSO for 20 min. They were then plated on vitronectin (A) or maintained in suspension (S) for 30 min, after which tyrosine-phosphorylated Syk (A) or Vav3 (B) was determined by immunoblotting. (C) Syk^−/− BMMs transduced with WT Syk, Syk^R195A, Syk^K402R, or pMX vector were cultured in RANKL and MCSF for 3 d. The preosteoclasts were lifted and plated on vitronectin for 30 min, after which phosphorylated tyrosine in Syk immunoprecipitates and TCLs was determined by immunoblotting. (D) WT BMMs were cultured in RANKL and MCSF for 3 d. The preosteoclasts were then suspended and preincubated in 2 μM SU6656 (+) or DMSO (−) for 20 min. After 30-min adhesion to vitronectin, Syk immunoprecipitates and TCLs were immunoblotted for c-Src and Syk. (E) WT BMMs cultured in RANKL and MCSF for 3 d were lifted and preincubated for 20 min in 10 μM PP2, 2 μM SU6656, or DMSO for 20 min. The cells were then plated on vitronectin (A) or maintained in suspension (S) for 30 min, after which phosphorylated tyrosine in Syk immunoprecipitates was determined by immunoblotting. (F) WT or c-Src^−/− spleen cells were cultured in RANKL and MCSF for 3 d. The cells were lifted and either replated on vitronectin or maintained in suspension for 30 min, after which phosphorylated tyrosine in Syk immunoprecipitates was determined by immunoblotting. Syk, c-Src, and β-actin (loading control) content was also assessed in TCLs. (G) WT or Syk^−/− BMMs were cultured in RANKL and MCSF for 3 d. The cells were lifted and either replated on vitronectin or maintained in suspension for 30 min. The plated cells were then rinsed with PBS to remove nonadherent cells. c-Src^Y416 and β-actin (loading control) content was assessed by immunoblotting in TCLs of both suspended and adherent cells.

Model of Syk-mediated organization of the osteoclast cytoskeleton. c-Src constitutively associates with the β3-integrin subunit. αvβ3 is activated upon ligand (RGD) occupancy, which recruits Syk bound to the Y-phosphorylated ITAM proteins FcRγ and Dap12 in an SH2 domain–dependent manner to the integrin–c-Src complex. αvβ3 occupancy also activates c-Src by phosphorylating Y416. In turn, activated c-Src phosphorylates Syk. Activated Syk phosphorylates Vav3, which shuttles Rac to its GTP-bound form, thereby organizing the actin cytoskeleton.

Syk deficiency results in abnormal osteoclast function in vivo. (A) PTH(1–34) or vehicle was administered to Syk^+/? or Syk^−/− chimeras for 4 d. Calvariae were fixed and stained for TRAP activity, and osteoclast (OC) number was histomorphometrically determined, *, P < 0.05; **, P < 0.01; PTH versus buffer. (B) TRAP-stained histological sections of calvaria of PTH-treated Syk^+/? or Syk^−/− chimeric mice. Osteoclasts (red reaction product) of PTH-treated Syk^+/? mice reside in resorption lacunae and contain ruffled membranes (arrow), whereas most Syk^−/− osteoclasts are smaller and are not juxtaposed to the bone. (C) Percentage of bone surface juxtaposed to osteoclasts in control and PTH-treated Syk^+/? or Syk^−/− mice. *, P < 0.05 versus Syk^−/−; +, P < 0.05 versus buffer. (D) Serum was collected from vehicle and PTH-treated Syk^+/? and Syk^−/− chimeras. CTx was measured by ELISA. *, P < 0.01. Error bars represent SD.

Syk does not regulate osteoblast function. (A and B) Syk^+/? or Syk^−/− primary calvaria osteoblasts were cultured with differentiation media (α-MEM containing 50 μg/ml ascorbic acid and 2 mM β-glycerophosphate) for 20 d. Bone nodule formation was visualized by Alizarin red staining. (C) Syk^+/? or Syk^−/− primary calvarial osteoblasts were cultured in α-MEM media (con), 10 nM 1,25-dihydroxyvitamin D₃ (VitD), or 10 ng/ml TNF-α for 24 h. RANKL and OPG expression were analyzed by RT-PCR. Actin serves as loading control. (D) Syk^+/? or Syk^−/− primary calvarial osteoblasts were lysed, and Syk expression was determined by immunoblotting. WT macrophage (Mφ) lysate serves as a positive control.

The distal three amino acids of the β3-integrin cytoplasmic domain mediate Syk–c-Src association. (A) Amino acid sequences of human (h) β3-integrin constructs used in this study. Letters in red refer to substituted β1-integrin subunit residues. (B–D) β3^−/− BMMs transduced with hβ3WT or hβ3–hβ1(C3R) were cultured in the presence of RANKL and MCSF for 5 d. (B) Cells were stained for TRAP activity. (C) Syk immunoprecipitates and TCLs were immunoblotted for c-Src and Syk content. (D) β3-integrin immunoprecipitates and TCLs were immunoblotted for Syk and β3-integrin content. (E) β3^−/− BMMs transduced with hβ3WT or hβ3^Y747F/Y752F (hβ3 2Y-2F) were cultured in the presence of RANKL and MCSF for 5 d. β3-integrin immunoprecipitates and TCLs were immunoblotted for Syk and β3-integrin content. (F) β3^−/− BMMs transduced with hβ3WT, hβ3ΔC, hβ3Δ752–762, or hβ3–hβ1(C3R) were cultured in the presence of RANKL and MCSF for 5 d. β3-integrin immunoprecipitates and TCLs were immunoblotted for β3-integrin and c-Src content.

ITAM proteins mediate αvβ3-induced Syk activation in osteoclasts. (A–C) WT or Dap12^−/−/FcRγ^−/− (2KO) spleen cells were cultured in RANKL and MCSF for 3 d, after which the cells were stained for TRAP activity (A), analyzed for β3-integrin mRNA content by RT-PCR (B), and analyzed for c-Src expression by immunoblotting (C). (D and E) WT or Dap12^−/−/FcRγ^−/− (2KO) spleen cells were cultured in RANKL and MCSF for 3 d. The cells were lifted and either replated on vitronectin (A) or maintained in suspension (S) for 30 min, after which Syk immunoblots were probed for phosphotyrosine content (D) or TCLs were immunoblotted for phosphorylated Vav3 and c-Src^p-Y416 (E). (F) Syk^−/− BMMs transduced with WT Syk or Syk^R195A were cultured in RANKL and MCSF for 5 d. Syk immunoprecipitates and TCLs were then immunoblotted for c-Src and Syk content. (G) Syk^−/− BMMs were transduced with WT Syk, Syk^R195A, or empty vector (pMX). After 5-d exposure to RANKL and MCSF, the cells were stained for TRAP activity.