c-Cbl and Cbl-b act redundantly to protect osteoclasts from apoptosis and to displace HDAC6 from beta-tubulin, stabilizing microtubules and podosomes

Abstract

c-Cbl and Cbl-b are highly conserved adaptor proteins that participate in integrin signaling, regulating cytoskeletal organization, motility, and bone resorption. Deletion of both c-Cbl and Cbl-b in mice leads to embryonic lethality, indicating that the two proteins perform essential redundant functions. To examine the redundant actions of c-Cbl and Cbl-b in osteoclasts, we depleted c-Cbl in Cbl-b(-/-) osteoclasts by using a short hairpin RNA. Depleting both Cbl proteins disrupted both the podosome belt and the microtubule network and decreased bone-resorbing activity. Stabilizing the microtubules with paclitaxel or inhibiting histone deacetylase 6 (HDAC6), which destabilizes microtubules by deacetylating beta-tubulin, protected both the microtubule network and the podosome belt. Examination of the mechanism involved demonstrated that the conserved four-helix bundle of c-Cbl's tyrosine kinase binding domain bound to beta-tubulin, and both c-Cbl and Cbl-b displaced HDAC6. In addition to the effects on microtubules and the podosome belt, depleting both Cbls significantly increased the levels of the proapoptotic protein Bim and apoptosis relative to the levels induced by eliminating either protein alone. Thus, both c-Cbl and Cbl-b promote bone resorption via the stabilization of microtubules, allowing the formation of the podosome belt in osteoclasts, and by promoting osteoclast survival.

Figure 1.

Depletion of both Cbl proteins leads to the loss of the podosome belt and decreased bone resorption. (A) c-Cbl^−/− and Cbl-b^−/− OCLs were generated in the presence of M-CSF and RANKL. Cbl-b^−/− OCLs were infected with control (Cntr)-shRNA or c-Cbl-shRNA adenovirus as described in Materials and Methods. OCLs were lysed and whole cell extracts (WCE) were analyzed for expression of c-Cbl, actin and GAPDH by Western blot. (B) c-Cbl^−/− and Cbl-b^−/− OCLs were generated by coculture with osteoblasts and the Cbl-b^−/− OCLs were infected with Cntr-shRNA and c-Cbl-shRNA as indicated. The OCLs were replated on the glass coverslips for 24 h and the presence of c-Cbl, Cbl-b (both green), and actin (red) analyzed by confocal immunofluorescence. Depleting both c-Cbl and Cbl-b led to the disruption of the podosome belt. Isolated podosomes colocalized with areas of low residual c-Cbl staining in some OCLs (bottom). (C) WT, c-Cbl^−/−, Cbl-b^−/−, and Cbl-b^−/− OCLs infected with Cntr-shRNA or c-Cbl-shRNA were replated onto dentin slices and cultured for 24 h. Resorption pits were stained and quantified. *p < 0.05 relative to WT; #p < 0.05 relative to Cbl-b^−/−. Depleting both c-Cbl and Cbl-b decreased the bone resorbing activity of OCLs.

Figure 2.

c-Cbl–depleted Cbl-b^−/− OCLs exhibit alterations in microtubule organization. (A) WT, Cbl-b^−/−, and c-Cbl-shRNA–infected Cbl-b^−/− OCLs were fixed and analyzed by confocal immunofluorescence for tubulin (green) and actin (red). The microtubular network of the doubly Cbl-depleted OCLs was disrupted. (B) The localization of c-Cbl (blue), actin (red), and tubulin (green) was analyzed in Cbl-b^−/− OCLs (top), and c-Cbl-shRNA–infected Cbl-b^−/− OCLs generated in the presence (middle) or absence (bottom) of 1 nM paclitaxel. Paclitaxel prevented the loss of both the peripheral MTs and the podosome belt.

Figure 3.

Depleting both c-Cbl and Cbl-b reduces tubulin acetylation, and inhibiting HDACs partially prevents the loss of the peripheral podosome belt. (A) WT, c-Cbl^−/−, Cbl-b^−/−, and c-Cbl-shRNA–infected Cbl-b^−/− OCLs were lysed and whole cell extracts (WCEs) were analyzed for c-Cbl, acetylated tubulin, and total tubulin by Western blot. GAPDH was blotted to confirm equal loading. The level of acetylated tubulin was decreased in the c-Cbl–depleted Cbl-b^−/− OCLs. (B) WT, c-Cbl^−/−, Cbl-b^−/−, and c-Cbl-shRNA–infected Cbl-b^−/− OCLs were fixed and analyzed by confocal immunofluorescence for c-Cbl (blue), actin (red), and acetylated tubulin (green). Acetylated tubulin was decreased in the doubly Cbl-depleted OCLs. (C) The mean pixel intensities of confocal images of Cbl-b^−/− OCLs treated as indicated and stained for acetyl-tubulin were determined in total cells (T), cell peripheries (P), and cell centers and normalized to the mean total pixel intensity of Cbl-b^−/− OCLs. Data are presented as mean ± SEM; significance was determined relative to the comparable area in the untreated Cbl-b^−/− OCLs; *p < 0.05 versus total uninfected Cbl-b^−/− fluorescence; †p < 0.001 versus peripheral uninfected Cbl-b^−/− fluorescence; ‡p = 0.0498 versus peripheral uninfected Cbl-b^−/− fluorescence. (D) Cbl-b^−/− OCLs and Cbl-b^−/− OCLs infected with c-Cbl-shRNA adenovirus in the presence or absence of 100 nM TSA were lysed, and WCEs were analyzed by Western blot for c-Cbl, acetylated tubulin and actin, confirming the inhibition of tubulin deacetylation by TSA. (E) Cbl-b^−/− (top) and Cbl-b^−/− OCLs infected with c-Cbl-shRNA in the presence (middle) or absence (bottom) of 100 nM TSA were fixed and analyzed by confocal immunofluorescence for c-Cbl (blue), actin (red), and acetylated tubulin (green). Inhibiting histone deacetylase with TSA prevented the loss of the podosome belt.

Figure 4.

c-Cbl and Cbl-b displace HDAC6 from β-tubulin. (A) 293VnR cells were transfected with 5 μg of myc-c-Cbl or HA-Cbl-b. The cells were lysed and immunoprecipitated with anti-β-tubulin antibody or normal IgG. Bound c-Cbl (left) and Cbl-b (right) were detected with anti-myc or anti-HA antibodies, respectively. Both c-Cbl and Cbl-b bound to β–tubulin. (B and C) 293VnR cells were transfected with 5 μg of empty vector or 2, 5, and 10 μg of myc-c-Cbl (B) or HA-Cbl-b (C). β-Tubulin was immunoprecipitated, and the amounts of HDAC6, c-Cbl, and Cbl-b in the immune complexes were quantified by Western blot. The expression of HDAC6, the Cbl proteins, and β-tubulin in WCEs were determined. Both c-Cbl and Cbl-b dose-dependently reduced the amount of β-tubulin–associated HDAC6. (D) Immobilized GST, GST-TKB, GST-50, GST-4H/EF, GST-4H, GST-EF/SH, and GST-SH fusion proteins were incubated with 500 μg of whole cell extract from 293VnR cells, washed, and eluted with SDS sample buffer. The bound proteins were resolved by electrophoresis on a 4–12% polyacrylamide gel and bound β-tubulin was detected by Western blot by using anti-β-tubulin antibody.

Figure 5.

Depleting both c-Cbl and Cbl-b increases OCL apoptosis. (A) WT, c-Cbl^−/−, Cbl-b^−/−, and Cbl-b^−/− OCLs infected with Cntr-shRNA or c-Cbl-shRNA adenovirus were lysed and whole cell extracts (WCEs) were analyzed for c-Cbl and Bim expression by Western blot. Actin was blotted to confirm equal loading. The absence of either c-Cbl or Cbl-b increased Bim levels slightly, but depleting OCLs of both c-Cbl and Cbl-b markedly increased the level of Bim protein. (B) WT, c-Cbl^−/−, Cbl-b^−/−, and Cbl-b^−/− OCLs infected with Cntr-shRNA or c-Cbl-shRNA were generated by culturing BMMs for 4 d M-CSF and RANKL as described in Materials and Methods, and then they were fixed and stained for TRAP and the multinucleated TRAP⁺ cells were counted. *p < 0.05 relative to WT; #p < 0.05 relative to Cbl-b^−/−. Depleting both c-Cbl and Cbl-b reduced the number of OCLs relative to the number when only Cbl-b was eliminated. (C) Apoptosis in WT, c-Cbl^−/−, Cbl-b^−/−, and Cbl-b^−/− OCLs infected with Cntr-shRNA or c-Cbl-shRNA in the presence or absence of 100 μM zVAD-FMK was quantified by measuring the fragmentation of 5-bromo-2′-deoxyuridine-labeled DNA by ELISA, as described in Materials and Methods, and normalizing the values to the number of OCLs present in the cultures. *p < 0.05 relative to WT; #p < 0.05 relative to Cbl-b^−/−. Depleting only Cbl-b increased apoptosis, in contrast to the lack of effect of deleting or depleting only c-Cbl. Depleting both c-Cbl and Cbl-b further increased the level of apoptosis. Inhibiting caspase activity reduced the amount of apoptosis.

Figure 6.

The increased apoptosis does not cause the disruption of the podosome belt in the c-Cbl–depleted Cbl-b^−/− OCLs. (A) Cbl-b^−/− OCLs were infected with Cbl-shRNA in the presence or absence of 100 μM zVAD-FMK. Cells were fixed and analyzed by confocal immunofluorescence for actin (red) and tubulin (green). The pan-caspase inhibitor zVAD-FMK failed to prevent the loss of the podosome belt. (B) Cbl-b^−/− OCLs infected with Cbl-shRNA were incubated with zVAD-FMK (100 μM), paclitaxel (1 nM), or TSA (100 nM) or vehicle as indicated. Apoptosis was quantified as in Figure 5. Paclitaxel and TSA increased apoptosis to similar levels in the WT, c-Cbl^−/−, and Cbl-b^−/− OCLs, despite the higher apoptosis in the untreated Cbl-b^−/− OCLs. Apoptosis in the c-Cbl-depleted Cbl-b^−/− OCLs increased by similar amounts relative to the Cbl-b^−/− OCLs, regardless of the treatment. *p < 0.05 versus matched untreated; †p < 0.05 versus untreated WT; ‡p < 0.05 versus untreated Cbl-b^−/−; §p < 0.05 versus paclitaxel-treated or TSA-treated WT, c-Cbl^−/−, and Cbl-b^−/−; ¶p < 0.05 versus untreated doubly Cbl-depleted.