BMP2-induced Apoptosis Is Mediated by Activation of the TAK1-p38 Kinase Pathway That Is Negatively Regulated by Smad6*

Bone morphogenetic protein 2 (BMP2), a member of the transforming growth factor-beta (TGF-β) superfamily, regulates a variety of cell fates and functions. At present, the molecular mechanism by which BMP2 induces apoptosis has not been fully elucidated. Here we propose a BMP2 signaling pathway that mediates apoptosis in mouse hybridoma MH60 cells whose growth is interleukin-6 (IL-6)-dependent. BMP2 dose-dependently induces apoptosis in MH60 cells even in the presence of IL-6. BMP2 has no inhibitory effect on the IL-6-induced tyrosine phosphorylation of STAT3, and the bcl-2 gene expression which is known to be regulated by STAT3, suggesting that BMP2-induced apoptosis is not attributed to alteration of the IL-6-mediated bcl-2pathway. We demonstrate that BMP2 induces activation of TGF-β-activated kinase (TAK1) and subsequent phosphorylation of p38 stress-activated protein kinase. In addition, forced expression of kinase-negative TAK1 in MH60 cells blocks BMP2-induced apoptosis. These results indicate that BMP2-induced apoptosis is mediated through the TAK1-p38 pathway in MH60 cells. We also show that MH60-derived transfectants expressing Smad6 are resistant to the apoptotic signal of BMP2. Interestingly, this ectopic expression of Smad6 blocks BMP2-induced TAK1 activation and p38 phosphorylation. Moreover, Smad6 can directly bind to TAK1. These findings suggest that Smad6 is likely to function as a negative regulator of the TAK1 pathway in the BMP2 signaling, in addition to the previously reported Smad pathway.

Bone morphogenetic protein-2 (BMP2), 1 a member of the transforming growth factor-␤ (TGF-␤) superfamily, signals through the heterotetrameric complex of type I and type II serine-threonine kinase receptors (1,2). Downstream of the receptor complex, at least two distinct intracellular pathways have been suggested for mediating inductive signals from the cell membrane to the nucleus.
One pathway involves a family of transcription factors collectively known as Smads. Smad1, Smad5, or Smad8 are phosphorylated by activated type I BMP2/4 receptors and are then associated with a common signaling mediator, Smad4. The resultant heteromeric Smad complex is translocated into the nucleus where they activate transcription (3)(4)(5). Another pathway is mitogen-activated protein kinase (MAPK) cascade initiated by TAK1 (for TGF-␤-activated kinase-1) (6). TAK1 was originally identified as a member of MAPK kinase kinase (MAPKKK) activated in response to TGF-␤ and BMP4 (6). More recently, it has been reported that TAK1 functions as a mediator of the MKK6-p38 pathway and the MKK7-JNK pathway (7)(8)(9). However, a regulatory mechanism of the signal transduction pathway mediated by TAK1 remains to be clarified.
After initial activation of the receptors, BMP2 elicits multiple effects ranging from cell differentiation to regulation of early embryogenesis. Because there are tens of different cytokines present simultaneously in vivo, it is sometimes the case that two different signal pathways could cause the synergistic or antagonistic interplay in common target cells. We have recently reported that BMP2 and leukemia inhibitory factor (LIF) act in synergy on cultured fetal neural progenitor cells to induce differentiation into astrocytes, whereas BMP2 or LIF alone do not induce astrocyte development under the same culture conditions (10). LIF is a member of the interleukin-6 (IL-6) family of cytokines, which shares membrane glycoprotein gp130 as a common subunit in the receptor complex (11). Ligand binding to the receptor triggers the dimerization of gp130, activating gp130-associated cytoplasmic tyrosine kinase in the Janus kinase family and a downstream transcription factor, signal transducer and activator of transcription-3 (STAT3) (11). In the case of synergistic effect caused by BMP2 and LIF, respective downstream transcription factors Smad1 and STAT3 form a signal-dependent complex bridged by p300 in the nucleus, thereby exerting the synergy between BMP2 and LIF in astrocyte differentiation (10). Moreover, another group has demonstrated the inhibitory cross-talk between the Smad cascade and the signal from a receptor tyrosine kinase: activation of epidermal growth factor receptor leads to MAPK activation, which then phosphorylates Smad1 on serine residues, thereby inhibiting Smad1 translocation into the nucleus (12).
In the present study, we attempt to know the interaction of the signals mediated by IL-6 and those by BMP2 in mouse hybridoma MH60 cells whose growth is IL-6-dependent. We show here that BMP2 induces apoptosis in MH60 cells even in the presence of proliferative signals of IL-6. We show that phosphorylation of STAT3 and expression of its target gene bcl-2 are not affected by BMP2, suggesting that the cell death signaling pathway initiated by BMP2 is independent of the survival signaling by IL-6. The studies presented here further propose a mechanism of BMP2-induced apoptosis in MH60 cells, which appears to be mediated by the TAK1-p38 kinase pathway. In addition, we show that Smad6 unexpectedly blocks BMP2-induced apoptosis and suggest a mechanism by which Smad6 prevents cell death promoted by BMP2.
Cell Culture-Mouse Hybridoma MH60 cells were cultured in RPMI 1640 medium (Sigma) containing 10% fetal bovine serum, 2 ng/ml IL-6. COS7 cells were maintained in Dulbecco's modified Eagle's medium (Sigma) supplemented with 10% fetal bovine serum. To establish Smad6-overexpressing MH60 cell clones (MH60/Smad6), 20 g of FLAG-tagged Smad6 expression vector or mock vector was cotransfected with 1 g of pSV2-NEO using an electroporation method. Transfectants were selected by 500 g/ml G418, and single cell clones were obtained by limiting dilution. Expression of Smad6 was analyzed by Western blotting using anti-FLAG antibody (Sigma).
Cell Growth Assay-Cells (2 ϫ 10 4 cells per well in a 96-well plate) were cultured in 100 l of medium containing 2 ng/ml IL-6 and various concentrations of BMP2 for 40 h. The number of viable cells was then examined by a colorimetric assay using the WST-8 cell-counting kit (Wako).
Reverse Transcriptase-Polymerase Chain Reaction-First strand cDNAs were synthesized from 1 g of total RNA using superscript II (Life Technologies, Inc.). The respective first strand cDNAs were then used directly for amplification of bcl-2 and G3PDH genes by polymerase chain reaction (PCR). The PCR reaction was performed using cycling conditions appropriate for each gene.
DNA Fragmentation Assay-After the cultivation of cells in the presence of 50 ng/ml of BMP2 for 13 or 24 h, cells (5 ϫ 10 5 ) were lysed in a buffer containing 0.5% Triton X-100, 10 mM Tris, pH 7.4, and 10 mM EDTA. After treatment with RNase A and proteinase K, the size of DNA was analyzed by agarose gel electrophoresis.
To examine the p38 activation, MH60 cells were treated with BMP2 (20 ng/ml) for 10 or 20 min after starvation as described above. The cells were lysed with lysis buffer, and subjected to immunoblotting with p38 antibody (Santa Cruz Biotechnology) or phospho-specific p38 antibody (New England BioLabs), respectively.
For the coimmunoprecipitation assay, COS7 cells were transfected with expression constructs using a Trans-IT LT-1 polyamine transfection reagent (Mirus) according to the manufacturer's protocol. After 1 day in culture, cells were solubilized with lysis buffer. Lysates from transfected COS7 cells or MH60 transfectants were incubated with antibody to Myc (Genosys) or FLAG (Sigma) for 2 h, and immune complexes were then precipitated with magnetic protein G (BioMag) for 30 min. Precipitates and, in some cases, cell lysates were subjected to SDS-PAGE followed by immunoblotting with anti-FLAG, anti-Myc, an-ti-HA (Santa Cruz), or anti-TAK1 (Santa Cruz).
Immunofluorescence and Microscopic Observation of Apoptotic Cells-MH60 cells were transfected using SuperFect transfection reagent (Qiagen) according to the manufacturer's protocol. Cells were fixed with 3% paraformaldehyde in phosphate-buffered saline (PBS) for 10 min followed by washing with PBS. Cells were permeabilized with 0.1% Triton X-100 for 5 min, washed again, and incubated in blocking solution (3% bovine serum albumin in PBS) at 4°C for 2 h. Cells were incubated with monoclonal antibody against FLAG-epitope or polyclonal antibody against HA-epitope at 4°C for 1 h in PBS containing 2% bovine serum albumin. Cells were washed three times with PBS and incubated with fluorescein isothiocyanate-conjugated anti-mouse IgG (ImmunoResearch) or rhodamine-conjugated anti-rabbit IgG (Chemicon) for an additional 1 h. After three times washing, cells were counterstained with Hoechst 33258. Images were obtained using fluorescent microscopy (AX70, Olympus).

BMP2-induced Apoptosis in MH60 Cells-
We have previously demonstrated that BMP2 and LIF exert a synergistic effect on the differentiation of neural progenitor cells (10). In an attempt to know whether a similar cooperative mechanism takes place in other cell systems, we tested the effect of BMP2 on the growth of IL-6-dependent mouse hybridoma MH60 cells. As shown in Fig. 1, BMP2 dose-dependently suppressed growth of MH60 cells in the presence of IL-6. To know the mechanism of this growth suppression induced by BMP2, we first examined whether BMP2 alters the cell cycle distribution and/or induces apoptosis on MH60 cells cultured with IL-6. BMP2 treatment for 24 h resulted in the cell cycle arrest in the G 1 phase (data not shown) and genomic DNA fragmentation ( Fig.  2A). These observations indicate that treatment of BMP2 causes cell cycle arrest in the G 1 phase and apoptosis in MH60 cells.
Withdrawal of IL-6 for 24 h led to the apoptosis of MH60 cells ( Fig. 2A) without apparently inducing cell cycle arrest in the G 1 phase (data not shown) but with decreased expression of bcl-2 ( Fig. 2B). To examine whether the BMP2 signaling interferes with the IL-6 signaling pathway at the step of STAT3 activation, we analyzed tyrosine phosphorylation of STAT3 in the presence or absence of BMP2. Tyrosine phosphorylation of STAT3, which has been shown to be a prerequisite for STAT3 activation, was induced in MH60 cells by IL-6 stimulation, whereas it was not affected by BMP2 (Fig. 2C). STAT3 activation has been shown to be important for bcl-2 gene expression and eventually for cell survival by IL-6 (17). These data indicate that BMP2-induced apoptosis in MH60 cells is not caused by the inhibition of STAT3 activation or its downstream targets.
BMP2-induced Activation of p38 and TAK1 and Its Inhibition by Smad6 -The result described above that bcl-2 gene expression was not changed at the mRNA and protein levels (not shown) in response to BMP2 suggested the existence of a signaling cascade other than those involving bcl-2 regulation for the apoptotic response of MH60 cells. We then focused on p38 stress-activated protein kinase whose activation has been suggested to be involved in the induction of apoptosis independent of bcl-2 function, in response to various cytotoxic stresses such as UV and x-ray radiation, heat shock, and tumor necrosis factor-␣ stimulation (18 -22). We thus examined whether treatment of MH60 cells with BMP2 leads to activation of p38. As shown in Fig. 4A, the level of the activated form of p38, which can be detected by phospho-p38-specific antibody, increased within 5 min after treatment with BMP2. Because it is known that p38 lies downstream of the TAK1-MKK6 pathway (7), we next examined the activation of TAK1 kinase by immunocomplex kinase assay using a recombinant his-MKK6 protein as a substrate. Fig. 4B shows that treatment of MH60 cells with BMP2 significantly enhanced the protein kinase activity of TAK1, which coincided with p38 activation (Fig. 4A). These results indicate that BMP2 activates the TAK1-p38 pathway in MH60 cells.
To our surprise, in the Smad6-overexpressing MH60 cells (MH60/Smad6), which are resistant to BMP2-induced apoptosis, activation of p38 and TAK1 did not occur after BMP2 stimulation (Fig. 4, A and B). These results indicate that Smad6 inhibits the activation of the TAK1-p38 pathway in BMP2 signaling.
Essential Role of TAK1 Activation in the BMP2-induced Apoptosis in MH60 Cells-Catalytically inactive TAK1 (TAK1/ KN) is known to inhibit the TAK1 signaling pathway (6). To determine the involvement of TAK1 in BMP2-induced apoptosis, we transiently transfected expression vectors encoding FLAG-tagged TAK1/KN (FLAG-TAK1/KN) or wild type control ((FLAG-TAK1(WT)) into MH60 cells. Expression of these proteins could be identified by immunostaining with anti- FLAG   FIG. 2. BMP2 induced apoptosis in MH60 hybridomas but did not interfere with the IL-6 signal transduction pathway. MH60 cells were incubated with IL-6 (2 ng/ml) or BMP2 (50 ng/ml) in a combination as indicated above the figure. A, DNA purified after 24-h culture was analyzed for fragmentation in an agarose gel. B, total RNA prepared after 15-h culture was used for reverse transcriptase-PCR analysis with primers specific for bcl-2 and G3PDH. C, lysates prepared from cells stimulated with cytokines for 15 min were subjected to antiphosphotyrosine or anti-STAT3 immunoblotting. FIG. 4. BMP2 stimulated activation of the TAK1-p38 cascade, which was blocked by Smad6. A, MH60 cells or MH60/Smad6 cells were treated with BMP2 (50 ng/ml) for the indicated time periods. Cell lysates were subjected to immunoblotting with phospho-specific p38 antibody to detect the active form of p38 (upper panel) and subsequently with p38 antibody to analyze the level of endogenous p38 (lower panel). B, MH60 cells or MH60/Smad6 cells were treated with or without BMP2 for the indicated time periods. Cell lysates were immunoprecipitated with or without anti-TAK1 antibody. Protein kinase assay was then performed as described under "Experimental Procedures" with his-MKK6 as a substrate. The phosphorylated proteins were resolved by SDS-PAGE and visualized by autoradiography. The immunoprecipitates were also analyzed by SDS-PAGE and subsequent immunoblotting with anti-TAK1.
antibody. MH60 cells transfected with the above vectors were stimulated with BMP2 for 14 h, followed by staining for FLAG with the specific antibody. The nuclei were stained with the DNA dye bisbenzimide (Hoechst 33258), which detected apoptotic change characterized by chromatin condensation and nuclear fragmentation. As shown in Fig. 5, no apoptotic cells were detected in MH60 cells expressing FLAG-TAK1/KN even when treated with BMP2 (arrows in Fig. 5, A and B). Cells expressing negligible levels of FLAG-TAK1/KN underwent the apoptotic change (arrowheads in A and B). In contrast, MH60 cells transfected with FLAG-TAK1 (WT) exhibited the apoptotic change by treatment with BMP2, regardless of the expression levels of FLAG-TAK1(WT) (arrows in C and D, high expressing cells; arrowheads in C and D, low expressing cells). These results suggest that TAK1 is involved in mediating BMP2-induced apoptosis and that its endogenous expression level is sufficient.
It has previously been shown that kinase activity of TAK1 was increased in cells cotransfected with TAB1 vector (15). Furthermore, simultaneous expression of TAK1 and TAB1 was able to enhance transcription of reporter gene under the control of the promoter region of the TGF-␤-inducible PAI-1 gene even in the absence of TGF-␤ (15). In addition, coinjection of TAK1 and TAB1 mRNA into dorsal blastomeres caused ventralization of the Xenopus embryos, mimicking the effect of BMP2 and BMP4 (23). To investigate the effect of simultaneous expression of TAK1 and TAB1 in MH60 cells, we cotransfected both expression vectors into MH60 cells. As shown in Fig. 6, MH60 cells expressing both TAK1 and TAB1 genes exhibited clear apoptotic change even in the absence of BMP2, whereas cells expressing TAK1 or TAB1 alone or neither of the two exhibited no apoptotic change. These observations indicate that activated TAK1 is involved in induction of apoptosis in MH60 cells and mimics BMP2 effect.
Interaction of Smad6 and TAK1-As described above, Smad6 unexpectedly blocked activation of TAK1, which made us hypothesize that Smad6 can associate with TAK1 directly. To determine the target of Smad6 in the apoptotic pathway initiated by BMP2, we expressed these molecules in COS7 cells and analyzed physical interaction using coimmunoprecipitation assays. Lysates from COS7 cells expressing Myc-tagged Smad6 and FLAG-tagged TAK1 were immunoprecipitated with monoclonal antibody against FLAG or Myc and immunoblotted with anti-Myc or -FLAG antibody, respectively, to detect Smad6 or TAK1. As shown in Fig. 7A, Smad6 was coprecipitated with TAK1 (upper panel) and TAK1 was found in the Smad6 immunocomplex (lower panel). With a similar experimental procedure, binding of the endogenous TAK1 to Smad6 in the MH60/Smad6 transfectant clones was also observed (Fig.  7B). DISCUSSION Several recent studies have suggested that there is cross-talk between intracellular signaling pathways of different cytokines or growth factors. For example, epidermal growth factor and hepatocyte growth factor can antagonize the effects of BMP2 by inducing phosphorylation of Smad1 (12). On the other hand, LIF and BMP2 can elicit the synergistic interplay on differentiation of neural progenitor cells (10). In the later case, respective downstream transcription factors Smad1 and STAT3 form a complex bridged by p300 in the nucleus in a signal-dependent manner. In MH60 hybridoma cells, BMP2 opposes the proliferative effect of IL-6. However, our data suggest that BMP2 does not interfere with the IL-6 signal transduction pathway, partly because BMP2 does not alter phosphorylation of STAT3 and expression of bcl-2 gene induced by IL-6. Furthermore, we observed that BMP2 could induce apoptosis efficiently enough in MH60 cells regardless of the presence or absence of IL-6 (data not shown); therefore, BMP2 did not induce apoptosis in these cells more efficiently in the absence of IL-6. These results suggest that BMP2 promotes cell death by activating its own apoptotic pathway.
BMP2-induced apoptosis has been reported in vivo so far, for example, in rhombomeres 3 and 5 (24) and the interdigit field of the limb (25,26) in developing chickens. In addition, an earlier study demonstrated that BMP2 was capable of inducing apoptosis in cultured mouse hybridoma HS-72 cells whose growth is independent of IL-6 (27). Although the BMP2 signal transduction pathways and molecular nature of their components have been studied in detail, the signaling cascade by which BMP2 induces apoptosis both in vivo and in vitro are not fully understood.
BMP2 is a member of the TGF-␤ superfamily of cytokines whose signals have been suggested to be transduced from receptor serine/threonine kinases to the nucleus via at least two different signaling pathways involving transcription factors Smads and MAPKKK family kinase TAK1. TAK1 was originally identified as a TGF-␤-responsive MAPKKK (6). No member of the TGF-␤ superfamily other than TGF-␤ and BMP4 have been reported so far to directly activate TAK1 (6). It was later shown that injection of Xenopus TAK1 (xTAK1) and its activator xTAB1 with bcl-2 in early embryos caused ventral- FIG. 6. Simultaneous expression of TAK1 and TAB1 induced apoptosis, mimicking the effect of BMP2. MH60 cells were cotransfected with HA-tagged TAK1 and FLAG-tagged TAB1 expression vectors. After 45 h, expression of TAK1 was detected with anti-HA antibody and rhodamine-conjugated secondary antibody (red) (A), and expression of TAB1 was detected with anti-FLAG antibody and fluorescein isothiocyanate-conjugated secondary antibody (green) (B). Nuclear morphology was determined by staining with Hoechst dye (C). Only the MH60 cells expressing both TAK1 and TAB1 genes showed apoptotic morphology.
ization, mimicking the effect of BMP2 and BMP4 in Xenopus and implying that TAK1 participates in the BMP2 and BMP4 signaling pathway (15). In the present study, we show the first evidence that BMP2 could activate TAK1.
Studies on early Xenopus embryos have demonstrated that the constitutively active form of xTAK1 or a combination of normal xTAK1 plus xTAB1 could induce not only ventral mesoderm but also cell death (23). Therefore, it has been presumed that activation of TAK1 has a potential to induce apoptosis in the Xenopus embryo. Here we demonstrated that cotransfection of TAB1 and TAK1 leads to the induction of apoptosis, mimicking the effect of BMP2. Because we have not examined whether Smad1 or Smad5 could induce apoptosis on MH60 cells, the possibility that Smad pathway is involved in the BMP2-induced apoptosis could not be completely excluded. However, BMP2-induced apoptosis was completely blocked in the MH60 cells expressing the kinase-defective form of TAK1, suggesting that TAK1 is essential for BMP2-induced apoptotic signaling. Treatment with BMP2 activates TAK1, which was accompanied by similar kinetics of p38 activation. Moriguchi et al. (7) have shown the existence of a kinase cascade consisting of TAK1-MKK6-p38. Other previous reports show that treatment of cells with ceramide (8), IL-1 (28), tumor necrosis factor-␣ (9), and TGF-␤ (29) results in activation of the TAK1-JNK cascade. JNK and p38, collectively known as stress-activated protein kinases, are involved in apoptosis, probably by inducing expression of a Fas ligand (30). Taken together, we conclude that BMP2-induced apoptosis is most likely mediated by the TAK1-p38 signal cascade.
Proteins Smad1, -2, -3, -5, and -8 have been shown to transduce signals, in cooperation with Smad4, of the TGF-␤ super-family, which is independent of the TAK1-MAPK cascade (3)(4)(5). Smad6, one of the two known inhibitory Smad species (Smad6 and Smad7), blocks the Smad pathway by associating with type I receptors to inhibit Smad1 and Smad2 phosphorylation (16) or by sequestration of phosphorylated Smad1 from Smad4 (31). Interestingly, an overexpression of Smad6 blocks BMP2-induced apoptosis in MH60 cells. In a similar approach, Ishisaki et al. (27) have previously shown that an overexpression of Smad6 suppressed BMP2-induced apoptosis in mouse B cell hybridoma HS-72 cells. They showed that Smad6 blocked BMP2-induced phosphorylation of Smad1/Smad5 in HS-72 cells, as has been demonstrated by other groups. However, because their studies did not explain how BMP2 could trigger apoptosis, it was not clear how Smad6 could prevent BMP2induced apoptosis. Here we clearly show that BMP2-induced apoptosis is dependent on the TAK1-p38 cascade in MH60 cells and that Smad6 prevents the activation of TAK1-p38 cascade induced by BMP2. Therefore, Smad6 is likely to function as a negative regulator of the TAK1 pathway in the apoptotic signaling of BMP2, in addition to the previously known negative regulatory function in the Smad pathway.
It has been reported that Smad6 binds to the type I receptors for the TGF-␤ superfamily cytokines and inhibits the Smad signaling pathway (16). It is important to note that our results show that Smad6 physically interacts with TAK1. This was confirmed in a COS7 cell expression system and also in MH60/ Smad6 transfectant clones, in the latter of which endogenous TAK1 was found to bind to Smad6. This suggests that Smad6 directly blocks the TAK1 activity by physical interaction. Xchromosome-linked inhibitor of apoptosis protein (XIAP) is a cytoplasmic molecule that is suggested to interact with BMP type I receptor (BMPR-I) and is involved in linking the BMP2stimulated receptor to TAK1 (14). The possibility that association of Smad6 with BMPR-I releases XIAP from BMPR-I and thereby unlinks the receptor to TAK1 dose not appear to be true, because association of BMPR-I with XIAP was not disrupted by binding of Smad6 to BMPR-I in COS7 cells (data not shown). Taken together with the previous finding that Smad6 is a negative regulator of the Smad signaling, our finding of the potential for Smad6 to also inhibit the TAK1-p38 pathway indicates that Smad6 serves as a general inhibitor of TGF-␤ superfamily cytokines.
Because the transcription of Smad6 mRNA is induced by TGF-␤ superfamily cytokines (32), Smad6 may completely block the signaling by the TGF-␤ superfamily in an autoregulatory negative-feedback manner. Recently, involvement of TAK1 in the signal transduction of cytokines outside the TGF-␤ superfamily has been suggested, for instance in the NF-Binducing kinase-IB kinase cascade in the IL-1 signaling pathway (28) and in the NF-B essential modulator like kinase MAPK-like cascade in the Wnt signaling pathway (33,34). Therefore, our finding suggests that antagonistic cross-talk exists between BMP2 and other cytokines such as IL-1 and Wnt.
The molecular mechanism by which Smad6 inhibits TAK1 kinase activation remains to be elucidated. TAK1 is known to be activated by binding with TAB1, and it has been proposed that TAB1 binding to TAK1 induces an activating conformational change (15). We observed that Smad6 interacted with TAK1 but led to little disruption of the TAK1⅐TAB1 complex (not shown), implying that Smad6 may not compete with TAB1 for the binding to TAK1 but may block the catalytic site, or alternatively, interaction of Smad6 with the TAK1⅐TAB1 complex could cause the conformational change of TAK1 back to its inactive form again.
Our present study and others (27) show BMP2-induced ap-

FIG. 7. Smad6 and TAK1 interacted in vivo.
A, Myc-tagged Smad6 was expressed together with FLAG-tagged TAK1 in COS7 cells. Cell extracts were subjected to immunoprecipitation (IP) with anti-FLAG or anti-Myc antibody. Precipitates or lysates were separated by SDS-PAGE and analyzed by immunoblotting with anti-Myc antibody to detect Smad6 (upper panel) or anti-FLAG to detect TAK1 (lower panel). B, MH60/Smad6 transfectants and parental cells were lysed with lysis buffer and subjected to immunoprecipitation with anti-FLAG antibody to pull down Smad6. The precipitates were subjected to SDS-PAGE and subsequent immunoblotting with anti-TAK1 antibody to detect Smad6-TAK1 interaction. optosis in the mouse hybridomas that were derived from plasmacytomas, the tumors of terminally differentiated B lineage cells. We are in the process of testing the effect of BMP2 on humanequivalent tumors, i.e. multiple myeloma. It is of much interest to note that (i) growth of myelomas are often IL-6-dependent; (ii) myeloma cells produce IL-6, IL-6 receptor (IL-6R), and soluble IL-6R (35); and (iii) patients with multiple myeloma often exhibit osteoporosis, presumably because of promotion of osteoclast development by IL-6 and soluble IL-6R produced by myeloma cells. BMP2 is known to play an important role in inducing the bone formation. In the bone marrow, BMP2 produced by osteoblasts is suggested to be stored in a latent form bound to bone matrix as in the case of TGF-␤ (36,37). When bone resorption is initiated by osteoclasts, BMP2 is released from the deposit in the calcified bone matrix, and thereafter it stimulates the differentiation of osteoblasts. In addition, osteoclasts activated by IL-6 and soluble IL-6R, which are produced by myelomas, may be growth-inhibited by the released BMP2. Because the bone marrow provides an important hematopoietic microenvironment, bone formation and hematopoiesis are closely related there. For instance, Kajkenova et al. (38) has previously shown that myeloid progenitors and IL-6 production were significantly increased in the bone marrow of osteoblastogenesis-defective mouse. Taken together, BMP2 might act as a regulatory factor for maintaining a balance between bone formation and hematopoiesis in the bone marrow.