Transforming growth factor-{beta} coordinately induces suppressor of cytokine signaling 3 and leukemia inhibitory factor to suppress osteoclast apoptosis

Abstract

Local release of TGF-beta during times of high bone turnover leads to elevated levels within the bone microenvironment, and we have shown that TGF-beta suppresses osteoclast apoptosis. Therefore, understanding the influences of TGF-beta on bone resorbing osteoclasts is critical to the design of therapies to reduce excess bone loss. Here we investigated the mechanisms by which TGF-beta sustains suppression of osteoclast apoptosis. We found TGF-beta rapidly increased leukemia inhibitory factor (LIF) expression and secretion by phosphorylated mothers against decapentaplegic-dependent and -independent signaling pathways. TGF-beta also induced suppressor of cytokine signaling 3 (SOCS3) expression, which was required for TGF-beta or LIF to promote osteoclast survival by. Blocking LIF or SOCS3 blocked TGF-beta promotion of osteoclast survival, confirming that LIF and SOCS3 expression are necessary for TGF-beta-mediated suppression of osteoclast apoptosis. Investigation of the mechanisms by which LIF promotes osteoclast survival revealed that LIF-induced expression of Bcl-X(L) and repressed Bcl-2 interacting domain expression by activating MAPK kinase, AKT, and nuclear factor-kappaB pathways. Suppression of Janus kinase/signal transducer and activator of transcription signaling further increased Bcl-X(L) expression and enhanced osteoclast survival, supporting that this pathway is not involved in prosurvival effects of TGF-beta and LIF. These data show that TGF-beta coordinately induces LIF and SOCS3 to promote prosurvival signaling. This alters the ratio of prosurvival Bcl2 family member Bcl-X(L) to proapoptotic family member Bcl-2 interacting domain, leading to prolonged osteoclast survival.

Figure 1

Influences of TGF-β treatment on Bcl-2 family member expression. Mature osteoclasts were treated with 2 ng/ml TGF-β for 1 or 24 h before harvest and real-time PCR analysis for BclX_L (A) and Bid (B). Data are the mean of three replicates from a single marrow preparation from three mice. The experiment was repeated four times with separate mice and these results are representative of these experiments. *, P < 0.05.

Figure 2

TGF-β stimulates LIF expression. A, Osteoclasts were treated with vehicle or 2 ng/ml TGF-β for 1 or 24 h. RNA was analyzed by real-time PCR. Data are the mean of three replicates from a single marrow preparation from three mice. The experiment was repeated four times with separate mice and these results are representative of these experiments. *, P < 0.05. B, Osteoclasts were adenovirally infected with empty vector or a construct composed of 274 bp upstream of the LIF start site linked to a luciferase reporter. After 24 h, osteoclasts were treated with TGF-β for 2 d before lysis, normalization for protein levels, and luciferase assay. *, P < 0.05. Data are the mean of three replicates from a single marrow preparation from three mice. The experiment was repeated three times with separate mice and these results are representative of these experiments. C, Osteoclasts were treated for 8 or 24 h with vehicle or 2 ng/ml TGF-β and the conditioned media collected. LIF protein was assayed using R&D’s mouse LIF Quantikine kit. Data are the mean of three replicates from a single marrow preparation from one mouse. The experiment was repeated three times with separate mice and these results are representative of these experiments *, P < 0.05.

Figure 3

LIF mediates TGF-β support of osteoclast survival and regulates BclX_L and Bid expression. A, Osteoclasts were treated with vehicle (BSA), 2 ng/ml TGF-β, or 0.1 μg/ml anti-LIF blocking antibody for 8 h. Cells were analyzed for apoptosis by analysis of chromatin condensation. Arrows indicate viable osteoclasts, and asterisks indicate apoptotic osteoclasts with condensed chromatin. Data are the mean of three replicates from a single marrow preparation from one mouse. The experiment was repeated three times with separate mice and these results are representative of these experiments. B, Quantitation of osteoclasts pictured in A. Cultures that were treated with BSA or TGF-β, 50 ng/ml LIF or anti-LIF blocking antibody for 8 h. Data are the mean of six replicates from a single marrow preparation from one mouse. The experiment was repeated four times with separate mice and these results are representative of these experiments *, P < 0.05 compared with vehicle treatment; ƒ, P < 0.05 compared with BSA treatment. C–E, Mature osteoclasts were treated with vehicle or 50 ng/ml LIF before RNA harvest and real-time PCR analysis. Data in E are for 1 h treatment. Results of 24 h treatment were similar. Data are the mean of three replicates from a single marrow preparation from three mice. The experiment was repeated three times with separate mice and these results are representative of these experiments. *, P < 0.05.

Figure 4

TGF-β induces SOCS3 to promote osteoclast survival. A, Osteoclasts were treated with vehicle or 2 ng/ml TGF-β, or 50 ng/ml LIF for 1 or 24 h. RNA was harvested, processed, and analyzed by real-time PCR. Data are the mean of three replicates from a single marrow preparation from three mice. The experiment was repeated three times with separate mice and these results are representative of these experiments. *, P < 0.05. B, Osteoclasts were adenovirally infected with empty vector or a construct composed of 940 bp upstream of the SOCS3 start site linked to a luciferase reporter. After 24 h, osteoclasts were treated with TGF-β for 2 d before lysis, normalization for protein levels, and luciferase assay. Data are the mean of three replicates from a single marrow preparation from two mice. The experiment was repeated three times with separate mice and these results are representative of these experiments. *, P < 0.05. C, Osteoclasts were treated with siRNA as indicated for 40 h before vehicle, 2 ng/ml TGF-β, or 50 ng/ml LIF treatment for 8 h. Cells were fixed and analyzed for apoptosis as above. Data are the mean of six replicates from a single marrow preparation from one mouse. The experiment was repeated three times with separate mice and these results are representative of these experiments. *, P < 0.05 compared with vehicle treatment; ƒ, P < 0.05 compared with nontargeting siRNA treatment. D, Osteoclasts were treated as in C with nontargeting or siRNA targeting SOCS3 before LIF treatment. RNA was harvested in triplicate replicates from a single marrow preparation from one mouse for real-time PCR analysis. Data are the mean of three replicates from a single marrow preparation from three mice. The experiment was repeated three times with separate mice and these results are representative of these experiments.

Figure 5

TGF-β stimulates LIF and SOCS3 expression via NF-κB and SMAD. A, Mature osteoclasts were treated with dimethyl sulfoxide (DMSO), the NF-κB inhibitor pyrrolinedithio carbamate (PDTC) (25 μm), or specific inhibitor of Smad3 (SIS3) (20 μm) for 1 h before treatment. B, Mature osteoclasts were infected with the indicated adenovirus 20 h before treatment. A and B, Vehicle or 2 ng/ml TGF-β were added to the indicated samples for an additional 1 h. RNA was harvested and analyzed by real-time PCR for LIF and SOCS3 expression. *, P < 0.05 to vehicle treatment; ƒ, <0.05 to DMSO (A) or vector (B) treatment. Data are the mean of three replicates from a single marrow preparation from three mice. The experiment was repeated three times with separate mice and these results are representative of these experiments.

Figure 6

LIF stimulates activation of osteoclast signaling to modulate BclX_L and Bid expression. A, Osteoclasts were serum starved for 1 h before 50 ng/ml LIF addition for 0, 5, 10, 15, or 30 min. Cells were examined by Western blotting. The experiment was repeated three times with separate mice and these results are representative of these experiments. B, Mature osteoclasts were pretreated with dimethyl sulfoxide (DMSO), the JAK inhibitor AG-490 (10 μm), or the JAK/STAT inhibitor WP-1034 (STATi; 6 μm) as indicated for 1 h. Vehicle (VEH), 2 ng/ml TGF-β, or 50 ng/ml LIF were added for 8 h before fixing and analyzed for apoptosis as above. Data are the mean of six replicates. The experiment was repeated three times with separate mice and these results are representative of these experiments. *, P < 0.05 compared with vehicle treatment; ƒ, P < 0.05 compared with DMSO treatment. C and D, Mature osteoclasts were pretreated with DMSO, the MEK inhibitor UO126 (10 μm), the AKT inhibitor AKT IV (1 μm), the NF-κB inhibitor pyrrolinedithio carbamate (PDTC), AG-490 (JAKi), or WP-1034 (STATi) as indicated for 1 h. BSA (VEH) or 50 ng/ml LIF were added to the indicated samples for 60 min. RNA was harvested and analyzed by real-time PCR. Data are the mean of three replicates from a single marrow preparation from one mouse. The experiment was repeated three times with separate mice and these results are representative of these experiments. *, P < 0.05.

Figure 7

Model of TGF-β influences on osteoclast survival. TGF-β binding to its receptor complexes (RII) activates SMAD and AKT/NF-κB signaling to increase LIF and SOCS3 expression. LIF protein is secreted, binding to its receptor. SOCS3 suppresses JAK/STAT signaling to reduce proapoptotic responses including suppression of Bcl-X_L, whereas LIF activates prosurvival AKT/NF-κB and MEK signaling to increase Bcl-X_L and suppress Bid expression.