Dkk1-induced inhibition of Wnt signaling in osteoblast differentiation is an underlying mechanism of bone loss in multiple myeloma

Bone 42 (2008) 669 – 680 www.elsevier.com/locate/bone

Dkk1-induced inhibition of Wnt signaling in osteoblast differentiation is an underlying mechanism of bone loss in multiple myeloma☆ Ya-Wei Qiang a,⁎, Bart Barlogie a , Stuart Rudikoff b , John D. Shaughnessy Jr. a,⁎ a

Myeloma Institute for Research and Therapy, University of Arkansas for Medical Sciences, Little Rock, AR, USA b Laboratory of Cellular and Molecular Biology, National Cancer Institute, NIH, Bethesda, MD, USA Received 15 June 2007; revised 19 November 2007; accepted 13 December 2007 Available online 27 December 2007

Abstract Expression of the Wnt signaling inhibitor, DKK1 by multiple myeloma cells is correlated with lytic bone disease in multiple myeloma. However, the mechanism(s) by which DKK1 contributes to this process is not clear. Herein, we analyzed the functional role of canonical Wnt signaling and Dkk1 inhibition of this pathway in bone morphogenic protein (BMP)-2-induced osteoblast differentiation. Osteoblast differentiation was measured by alkaline phosphatase (ALP) activity in murine (C2C12) and human pre-osteoblast (hFOB1.19) and osteoblast-like (Saos-2 and MG63) cell lines. Cytoplasmic β-catenin protein was separated by E-cadherin–GST pull-down assay and analyzed by Western blotting. A dominant negative form of β-catenin, Dkk1 and TCF reporter constructs were transfected into C2C12 cells. C2C12 cells were also transfected with siRNA specific to LRP5/6 to knockdown receptor expression. Canonical Wnt signaling was activated in these cell lines in response to Wnt3a as assessed by increased cytoplasmic, non-phosphorylated β-catenin and TCF/LEF transcription activity. Recombinant Dkk1 and plasma from MM patients containing high levels of Dkk1 blocked Wnt3a-induced β-catenin accumulation. Importantly, Dkk1 abrogated BMP-2 mediated osteoblast differentiation. The requirement for Wnt signaling in osteoblast differentiation was confirmed by the following observations: 1) overexpression of Dkk1 decreased endogenous β-catenin and ALP activity; 2) silencing of Wnt receptor mRNAs blocked ALP activity; and 3) a dominant negative form of β-catenin eliminated BMP-2-induced ALP activity. Furthermore, Wnt3a did not increase ALP activity nor did BMP-2 treatment result in β-catenin stabilization indicating that cooperation between these two pathways is required, but they are not co-regulated by either ligand. These studies have revealed that autocrine Wnt signaling in osteoblasts is necessary to promote BMP-2-mediated differentiation of pre-osteoblast cells, while Wnt signaling alone is not capable of inducing such differentiation. Dkk1 inhibits this process and may be a key factor regulating pre-osteoblast differentiation and myeloma bone disease. © 2007 Elsevier Inc. All rights reserved. Keywords: Wnt; Dkk1; BMP-2; Osteoblast; Bone disease; Myeloma

Introduction Multiple myeloma (MM) is characterized by bone marrow infiltration of malignant plasma cells where they closely interact Abbreviations: Dkk1, Dickkopf1; MM, multiple myeloma; BMP-2, bone morphogenetic Protein-2; ALP, alkaline phosphatase; DN-β-cat, dominant negative form of β-catenin; TCF/LEF, T cell factor/Lymphocyte enhance factor. ☆ Supported by grants CA97513 (Shaughnessy and Barlogie) from the National Cancer Institute and by Senior Research Award from the Multiple Myeloma Research Foundation (Qiang). ⁎ Corresponding authors. E-mail addresses: [email protected] (Y.-W. Qiang), [email protected] (J.D. Shaughnessy). 8756-3282/$ - see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.bone.2007.12.006

with osteoblasts and osteoclasts, triggering osteolytic bone lesions. In an effort to elucidate the mechanisms of bone destruction, considerable attention has been focused on increased bone resorption due to increased osteoclast formation and activity [1,2]. While bisphosphonates block bone resorption by inhibition of osteoclastogenesis, bone formation does not occur, and osteolytic bone lesions are not repaired [3]. Thus, it appears that impaired osteoblast function also plays a role in MM bone lesion development. A number of studies have reported fewer osteoblasts and decreased bone formation in MM patients with higher plasma cell infiltration [4,5] and unbalanced bone turnover is observed in MM patients [6]. Wnts are a family of 19 secreted glycoproteins that have a well-characterized role in stem cell maintenance and development

670

Y.-W. Qiang et al. / Bone 42 (2008) 669–680

[7]. Wnt signaling is also involved in bone formation [8] and likely in MM pathogenesis [9,10]. Many Wnt effects are mediated through β-catenin which plays a pivotal role in the ‘canonical’ Wnt signaling pathway [7]. Upon interaction of Wnt proteins with frizzled (Fz) receptors and low-density lipoprotein-receptorrelated protein 5/6 (LRP5/6) [11] co-receptors, β-catenin protein accumulates in the cytoplasm and translocates to the nucleus. Nuclear localization and association with T cell factors (TCF-1, -3, and -4) and lymphoid enhancer-binding factor1 (LEF1) lead to transcriptional activation of target genes that regulate many cellular processes, including cell cycle progression and differentiation [7]. In addition to regulating gene transcription, β-catenin performs a second major function as a structural adaptor protein at cell–cell junctions [12]. Thus, cellular levels and localization of β-catenin reflect a complex dynamics that regulates multiple cell functions. Emerging data suggest that Wnt signaling plays an essential role in normal bone biology and deregulation of this process contributes to bone disease [13,14]. A role for Wnt signaling in osteoblast biology was initially suggested by Bradbury et al. [15], however the first direct evidence came from observations that inactivating mutations of the LRP5 gene cause osteoporosis–pseudoglioma syndrome (OPPG) [16]. Subsequently, it was shown that a separate and distinct mutation in the same gene, presumably leading to inhibition of Dkk1 binding, results in high bone density [17,18]. Furthermore, deletion of LRP5 in a mouse model inhibited osteoblast differentiation [19]. Expression of Wnt10b in transgenic mice increased bone mass [20] and overexpression of Wnt7b and β-catenin in C3H10T1/2 cells induced osteoblast differentiation [21,22]. However, the exact mechanism by which Wnt signaling affects osteoblast differentiation remains to be elucidated. The canonical Wnt pathway is regulated by a large number of antagonists, including the Dkk family and secreted frizzledrelated proteins (sRFPs). To date, four Dkk proteins have been identified in mammals [23], among which Dkk1 and Dkk2 have been well characterized and found to act as antagonists to the canonical Wnt pathway by binding to LRP5/6 in combination with a second protein designated as Kremen [24,25]. Recently, in vivo experiments revealed that transgenic overexpression of Dkk1 under the control of the ColA1 promoter leads to decreased bone mass [26]. In agreement with this observation, deletion of Dkk1 expression in mouse osteoblasts results in an increase in bone formation and mass [27]. In relation to MM, we have previously suggested that elevated expression of Dkk1 by myeloma tumor cells is involved in bone lesion formation [28]. Interestingly, the role of DKK1 in promoting bone lesion development appears not limited to MM, but has also been indicated in prostate cancer [29]. Given the role of Wnt and Dkk1 in normal bone formation and disease, we sought to understand the role of canonical Wntβ-catenin signaling in osteoblast differentiation and to decipher the molecular mechanisms by which Dkk1 promotes osteolytic bone lesions in MM. In the present study, we demonstrate that canonical Wnt signaling in osteoblast precursors is increased in response to Wnt3a and that a steady state canonical Wnt signal is necessary for BMP-2-induced osteoblast differentiation. Importantly, Dkk1 from myeloma cells inhibits this process.

Materials and methods Cells and cell culture Mouse pluripotent mesenchymal precursor cell line C2C12 and the human osteoblast cell line hFOB1.19 were purchased from America Type Culture Collection (Manassas, VA). C2C12, MG63, Saos-2, and 293T cells were cultured in Dulbecco's Modified Eagle Medium (DMEM) (Invitrogen, Carlsbad, CA) containing 10% heat-inactivated FBS, penicillin (100 U/ml), streptomycin (100 mg/ ml), and 4 mM L-glutamine. Cells were maintained at 37 °C and humidified with 95% air and 5% CO2 for cell culture. hFOB1.19 was cultured in a 1:1 mixture of Ham's F12 and DMEM with 10% FBS in the presence of 0.3 mg/ml G418.

Constructs and transfectants To generate dominant negative (DN)-β-catenin stable clones, C2C12 cells were transfected with pcDNA4 vector or a vector containing DN-β-catenin cDNA [30] using Lipofectamine (Invitrogen) following the manufacturer's instructions. After transfection, stable clones were generated by growing the cells in DMEM containing 10% FBS in the presence of Neocin (1 mg/ml) for two weeks. Stable Dkk1 and Dkk2 expressing clones generated in C2C12 and OPM-2 cells were previously described [31].

Preparation of conditioned medium Conditioned medium (CM) containing Wnt3a, Dkk1, Dkk2 and/or appropriate control constructs was prepared as previously described [31]. Dkk1 and Dkk2 proteins in CM were detected by immunoblotting using anti-V5 and anti-Dkk1 antibodies by ELISA. The supernatant from the culture medium was concentrated five-fold by using a YM-30 column [31]. Dkk1 levels in bone marrow plasma from MM patients were detected by ELISA.

Immunoblotting analysis Cells were incubated in MEM, Wnt3a CM, control CM, or with recombinant Wnt3a for the indicated times. For inhibition studies, cells were pretreated with purified recombinant Dkk1 at the indicated concentrations or Dkk1 CM, Dkk2 CM, or bone marrow plasma with low or high concentrations of Dkk1 protein for 1 h. Following treatment, cells were lysed as described [32]. Cell lysates were separated by SDS-PAGE and transferred to Immobilon polyvinylidene difluoride membranes (Millipore, Bedford, MA). Immunoblotting was performed using the indicated antibodies.

GST–E-cadherin binding assay The GST–E-cadherin binding assay was performed as described [33]. Briefly, the β-catenin binding site of E-cadherin as a GST-fusion protein was purified using GST beads. GST–E-cadherin was used to precipitate uncomplexed β-catenin present in 500 μg of cell lysate. Precipitated β-catenin was detected by immunoblotting using a β-catenin monoclonal antibody. Non-phosphorylated β-catenin was detected with a monoclonal antibody specific for β-catenin dephosphorylated at residues of 27–37 (Alexis, San Diego, CA).

Enzyme-linked immunosorbent assay Microtiter plates were coated with 50 µl of anti-Dkk1 antibody (R&D Systems, Minneapolis, MN) according to the manufacturer's recommendations. Bone marrow plasma (1:50) in dilution buffer was added and incubated overnight at 4 °C. Plates were washed and incubated with biotinylated goat anti-human Dkk1 IgG (R&D Systems, Minneapolis, MN) followed by streptavidin–horseradish peroxidase (Vector Laboratories), according to the manufacturer's recommendations.

Luciferase reporter gene assay Cells plated at 5 × 104 per well in a 12-well plate were transiently co-transfected with 1 microgram/ml of either TOPflash, FOPflash [34], or Cbfa-1-Luc (kindly

Y.-W. Qiang et al. / Bone 42 (2008) 669–680 Table 1 Mouse frizzled oligonucleotide primers for RT-PCR Primer Orientation Nucleotide sequence (5′ to 3′)

Nucleotide position

Fz1F Fz1R Fz2F Fz2R Fz3F Fz3R Fz4F Fz4R Fz5F Fz5R Fz6F Fz6R Fz7F Fz7R Fz8F Fz8R Fz9F Fz9R Fz10F Fz10R

2530–2551 2854–2832 3077–3098 3537–3514 2286–2309 2607–2584 1394–1417 1865–1842 1568–1589 2058–2038 635–655 1001–977 1843–1862 2181–2158 1225–1247 1667–1645 998–1021 1371–1349 528–550 834–811

Sense Anti-sense Sense Anti-sense Sense Anti-sense Sense Anti-sense Sense Anti-sense Sense Anti-sense Sense Anti-sense Sense Anti-sense Sense Anti-sense Sense Anti-sense

ATGTGTATGTGCGTGTGGACCG GGGAGATGCTGAAGGAAATGACC AAATAGGTTGGGTTGGAGGGAG AAACAGGAGAGACGGTTGAGAGCG TATTGAGGAGGATGGAACCAGTGC CAAAGCAGTCACCACACATAGAGG TAGTGGATGCCGATGAACTGACTG TTCCCCCTCTTCTCTCTCTTTACC ACATTCGCCACCTTCTGGATTG TTTTGGTTGCCCACATAGCAG AATGGACACTTTTGGCATCCG CTCTGGGTATCTGAATCGTCTAACG AAGGGGGAAACTGCGGTATG TCTCTCTCTCTGCTGGTCTCAACC TCCATCTGGTGGGTAATCCTGTC CGGTTGTGCTGCTCATAGAAAAG CGCCCGATTATCTTCCTTTCTATG TAGCAGAGCCCAGTCAGTTCATC CCAACAAGAACGACCCCAACTAC AAGAAGCACAGCACGGACCAGATG

provided by Dr. Ying Zhang, NCI, NIH) and 50 ng of pSV-β-galactosidase vector to normalize for transfection efficiency using Lipofectamine according to the manufacturer's instructions (Invitrogen). Following transfection, cells were exposed to Wnt3a CM or control CM for 24 h prior to luciferase assay. Luciferase activity was measured as previously described [31].

Alkaline phosphatase (ALP) assay Cells were cultured in DMEM with 2% horse serum including either BMP-2 (200 ng/ml), Wnt3a CM, BPM-2 plus Dkk1, or BMP-2 plus Wnt3a CM for 72 or 96 h followed by lysis in 150 μl of lysis buffer (20 mM Tris–HCl, pH 8 and 150 mM NaCl, 0.2% NP40). ALP activity was measured using ALP kit (Diagnostic Chemical Limited, Exton, PA) according to the manufacturer's instructions. Absorbance of the of samples was determined with a Spectra Max340 Microplate Spectrophotometer (Molecular Devices, Sunnyvale, CA) at 402 nm. Cell lysates were analyzed for protein content with using the micro-BCA assay kit (Pierce, Rockford, IL).

671

detection system (Applied Biosystems, Foster City, CA). The reaction mixture contained 1 μl of cDNA, dedicated buffers with specific primers and probes (5′-labeled by 6-carboxy-fluorescein and 3′-labeled by 1-carboxy-teteramethyrhdamine), and DNA polymerase in a total 20 μl volume. Following 2 min incubation at 50 °C and 10 min incubation at 95 °C for denaturing, the reaction was subjected to 40cycle amplifications at 95 °C for 15 s to denature and at 60 °C for 1 min for annealing/ extension. Each cDNA sample was analyzed in triplicate in parallel with GAPDH as a control. Changes in mRNA concentration were determined by subtracting the CT (threshold cycle) of target gene from the CTof GAPDH (Δ=CT gene−CT GAPDH). The mean of Δ control was subtracted from the ΔSiLRP5/6 reaction (mean Δ control −ΔSiLRP5/6=e). The difference was calculated as 2e by the 2−DDCT [35].

RNA interference Chemical synthesis of siRNA specific to LRP5/6, GFP, and control siRNA were purchased from Qiagen (Valencia, CA). The siRNA were transiently transfected into C2C12 using Lipofectamine according to the manufacturer's instructions (Invitrogen). RNA was isolated after 24, 48 or 72 h then subjected to RT-PCR or qPCR for determination of efficacy of target gene silencing.

Statistical analysis Statistical significance of differences between experimental groups was analyzed by a Student's t-test using the Microsoft Excel software statistical package. Significant p values were less than 0.05 by two-tailed test.

Results Expression of Wnt receptors and co-receptors in OB cells RT-PCR was used to evaluate the presence of Wnt receptor mRNA in C2C12, hFO1.19, and two human osteoblast-like cells lines, MG63 and Saos-2. Analysis using primers for all Fz family members (Fig. 1) revealed expression of Fzs1, 2, 4, 5, 6, 7, 8, and 9 with relatively higher levels of Fzs1, 6, and 7 in C2C12 cells. Similar expression of multiple Fzs was observed in the human lines. Fz3 was expressed in all human lines, but not C2C12, while Fz6 was expressed in C2C12 but none of the human lines. LRP5 and LRP6 were identified in all mouse and human lines. Thus, both components of functional Wnt

RT-PCR analysis First strand cDNA synthesis was performed as previously described [31]. All PCR reactions began with a first cycle at 95 °C for 3 min and a final cycle at 72 °C for 10 min with an additional 35 cycles at 94 °C/30 s, 60 °C/45 s, and 72 °C/1 min. Primer sequences for the indicated human genes are as described [31]. Primers, including Fz (Table 1), TCF and Dkk (Table 2) were designed using a primer pair program in the MacVector (Accelrys Software Inc, San Diego, CA) software based on gene sequences from the NIH Gene Bank (www. ncbi.nlm.nih.gov).

Sub-cloning of PCR fragments and DNA sequence analysis PCR fragments were subcloned using TOPO-TA cloning vector according to the manufacturer's instructions (Invitrogen) and sequence analysis performed as previously described [31]. Data analysis was performed using MacVector software and comparisons made with NCBI BLAST (http://www.ncbi.nlm.nih. gov/blast/).

Real-time quantitative PCR One microgram of total RNA was reverse transcribed into total cDNA. Quantitative PCR (qPCR) was performed using an ABI Prism 7000 sequence

Table 2 Mouse TCF and Dkk oligonucleotide primers for RT-PCR Primer

Orientation Nucleotide sequence (5′ to 3′)

Nucleotide position

TCF1F TCF1R TCF3F TCF3R TCF4F TCF4R LEF1F LEF1R Dkk1F Dkk1R DKk2F DKK2R Dkk3F Dkk3R Dkk4F Dkk4R

Sense Anti-sense Sense Anti-sense Sense Anti-sense Sense Anti-sense Sense Anti-sense Sense Anti-sense Sense Anti-sense Sense Anti-sense

197–219 629–607 208–229 509–488 96–115 561–540 288–309 567–546 1501–1525 1797–1776 1593–1613 1839–1819 1511–1533 1937–1914 697–717 839–817

ACGAACATTTCAGCAGTCCACAC GCATTGAGGGGTTTCTTGATGAC CAACGAATCGGAGAATCAGAGC ATGGCGACCTTGTGTCCTTGAC TGCTGGTGGGTGAAAAATGC CTTGAGGGTTTGTCTGCTCTGG TTCTCTTTTTCTCCCCTCCCCC AAACCTCTCCACGGATTCCTCG ACATTCGCCACCTTCTGGATTG GCAAAAGCACCAACCACACTTG AATGCGGAAGAATGAGGGATG TGCCAATCTGAAGGAAATGCC CGTGGACTTGGCAAAATGTAACC GAGCACTGGCTTTCAGAGGTATTG AAGCCCCAGAAATCTTCCAGC TGAACACAACAACAAGTCCCGTG

672

Y.-W. Qiang et al. / Bone 42 (2008) 669–680

(Fig. 2C). Similar results of multiple TCF/LEF family member expression were observed in the human lines although LEF1 appeared not to be expressed in MG63 (Fig. 2D). Transient transfection of cells with TOPflash reporter constructs containing binding sites for the β-catenin/TCF/LEF transcription complex resulted in a significant increase in luciferase activity in the presence of Wnt3a CM. In contrast, no transcriptional activation was observed in control cells treated with L929 CM or cells transfected with FOPflash containing mutant TCF/LEF binding sites (Fig. 2E). Taken together, these results indicate that a canonical Wnt signaling pathway is functional in preosteoblast cells. Dkk1 and MM patient sera inhibit Wnt3a induced β-catenin in pre-osteoblasts

Fig. 1. Frizzled (Fz) and LRP5/6 mRNAs are expressed in osteoblast cells. Total RNA was extracted from the indicated cell lines. RT-PCR was performed with mouse (A) and human (B) primers specific for Fz1 through 10 and LRP5/6 (mouse, C; human, D), respectively. GAPDH was included as control.

receptors are expressed in OB cell lines and the presence of multiple receptors is likely. Canonical Wnt signaling is activated in pre-osteoblast cell lines Having demonstrated the presence of Fz and LRP receptor gene expression, we sought to determine whether a functional canonical Wnt/β-catenin pathway was present by first examining the status of downstream β-catenin. Because osteoblasts express high levels of cadherin proteins (including β-catenin) [36], especially in the form of membrane-bound protein [12], we took advantage of the GST–E-cadherin binding assay to separate cytosolic, free (uncomplexed) β-catenin from the membrane-bound form. Examination of Wnt3a treatment effects revealed significant increases of free β-catenin appearing in a time-dependent manner (Fig. 2A) in all cell lines. Increases in βcatenin levels were apparent at 8 h and remained elevated for 24 h. Increases in β-catenin correlate with the presence of the non-phosphorylated, transcriptionally active form of the protein [37] as determined by analysis with antibodies specific to nonphosphorylated β-catenin (Fig. 2B). RT-PCR analysis of TCF/LEF family members revealed expression of TCF1, 3, 4, and LEF1 mRNA in C2C12 cells

To examine the effect of Dkk1 in OBs, cells were incubated with increasing amounts of Dkk1 prior to Wnt3a treatment. As shown in Fig. 3A, pretreatment with Dkk1 led to dosedependent inhibition of Wnt3a-increased, non-phosphorylated β-catenin in C2C12, hFOB1.19 and Saos-2 cells. Dkk1 inhibited non-phosphorylated β-catenin at concentrations beginning at 25 ng/ml with maximal inhibition at 50 ng/ml in C2C12. In hFOB1.19 and Saos-2 pronounced inhibition was observed at the lowest concentration tested (5 ng/ml) which changed only moderately with increasing concentration. Similar results were observed in analysis of total free β-catenin (data not shown). To determine whether Dkk1 expression by MM cells might have similar effects on functional Wnt signaling in the bone marrow microenvironment, the following experiments were performed. First, stable Dkk1-expressing clones were generated in the OPM2 MM cell line and the conditioned media containing Dkk1 used to determine the effect on Wnt3a-induced stabilization of β-catenin in C2C12 cells. The presence of Dkk1 protein in cell lysates from stable OPM-2 clones was determined by Western blot analysis with anti-V5 antibody (Fig. 3B). Dkk1 CM from OPM-2/Dkk1-expressing clones inhibited accumulation of uncomplexed, non-phosphorylated β-catenin in C2C12 cells (Fig. 3C). Furthermore, bone morrow serum from four MM patients containing over 100 ng/ml of Dkk1 (designated H1 to H4) similarly inhibited β-catenin (Fig. 3D). These results suggest that Dkk1 produced from MM cells and MM bone marrow can negatively regulate Wnt signaling in pre-osteoblasts. Dkk1 inhibits BMP-2-induced ALP activity Given the characterization of a functional Wnt signaling pathway in mouse and human pre- and osteoblast-like cells, experiments were undertaken to identify the biological effects associated with this pathway. C2C12 cells were selected as a model because of the following reasons. First, they undergo preosteoblast differentiation in the presence of BMP-2 [38]. Second, they express less Dkk1 mRNA and protein, compared with other human cell lines and primal human mesenchymal cells (data not shown) and finally they react to addition of Dkk1 more sensitively than that other human lines (Fig. 3A). As expected, BMP-2 treatment led to increased ALP activity

Y.-W. Qiang et al. / Bone 42 (2008) 669–680

673

Fig. 2. Canonical Wnt signaling is functional in osteoblast (OB) cells. OB cells were treated with Wnt3a CM or control CM for the indicated time and cell lysate harvested. Lysate protein was subjected to GST–E-cadherin assay and immunoblotting analysis using anti-β-catenin antibody (A) and anti-non-phosphorylated form (B) as in Materials and methods. RT-PCR was performed using specific primers for the indicated mouse (C) and human (D) genes in the indicated cell lines. C2C12 cells transfected using wild type (TOPflash) or mutant (FOPflash) LEF/TCF reporter luciferase constructs and pSV-b-galactosidase vector (transfection efficiency internal control) (E) were treated with Wnt3a CM or control CM prior to determination of luciferase activity. Results are shown as mean ± SD (n = 3) and are representative of three independent experiments. ⁎⁎p b 0.01 versus control.

whereas Wnt3a alone had no effect (Fig. 4A). C2C12 cells treated with both Wnt3a and BMP-2 demonstrated no obvious increase in ALP activity over BMP-2 alone indicating a lack of synergy between these factors in ALP production. Thus, exogenous Wnt3a alone is not sufficient to induce C2C12 differentiation. However, since Wnt3a mRNA (data not shown) and abundant steady state levels of β-catenin were detected in C2C12 cells (Figs. 2 and 3), the possibility of a cooperative role between BMP and Wnt pathways in C2C12 differentiation was next examined using Dkk1 protein to block canonical Wnt-β-catenin signaling. Interestingly, pretreatment of C2C12 cells with 100 ng/ ml of Dkk1 significantly inhibited ALP activity in the presence or absence of Wnt3a (Fig. 4A). Comparable studies in the human cell lines hFOB1.19 and Saos-2 cells produced similar results (Fig. 4B, C). These findings suggest that an autocrine β-catenin loop is critical to BMP-2-mediated pre-osteoblast differentiation.

An autocrine Wnt loop is required for differentiation of C2C12 cells To address the role of autocrine Wnt-β-catenin signaling in preosteoblast differentiation, C2C12 cells were transfected with a dominant negative (DN)-β-catenin construct which lacks TCF/ LEF binding sites and inhibits transcriptional activity [30]. DN-βcatenin expression was confirmed by Western blot analysis with the anti-X-press antibody (Fig. 5A) and DN-β-catenin clones exhibited significantly reduced Wnt3a induced luciferase activity (Fig. 5B). Importantly, marked reductions in ALP activity were observed in DN-β-catenin C2C12 cells in the presence of BMP-2 (Fig. 5C, D). These results suggest that endogenous canonical βcatenin signaling contributes to BMP-2-induced differentiation in C2C12 cells. To further confirm the effect of Dkk1 on osteoblast differentiation, we ectopically expressed Dkk1 and Dkk2 in C2C12 cells. Overexpression of both Dkk1 and Dkk2 inhibited

674

Y.-W. Qiang et al. / Bone 42 (2008) 669–680

in Fig. 6A, LRP5 and LRP6 mRNA expression in C2C12 cells was inhibited by the homologous (siLRP5 and siLRP6), but not by heterogeneous siRNA (control siRNA). Furthermore, inhibition of either LRP5 or 6 by their respective siLRNA's occurred in a dose-dependent manner as determined by qPCR (Fig. 6B, C) and neither reduced the expression of the other (not shown). Treatment of siLRP5- and siLRP6-expressing cells with BMP-2 resulted in significant reductions in ALP activity compared with cells expressing control siRNA (Fig. 6D) indicating that LRP coreceptors are likely required for pre-osteoblast differentiation.

Fig. 3. Dkk-1 and MM patient sera inhibit Wnt3a induced accumulation of β-catenin in OB cells. (A) The indicated cells were treated with recombinant Dkk-1 at the indicated concentrations and then stimulated with Wnt3a or control CM. Proteins in cell lysate were subjected to GST–E-cadherin assay and immunoblotting analysis using anti-non-phosphorylated β-catenin as described in Materials and methods. (B) Dkk1 protein in cell lysate (upper panel) and CM (middle panel) from OPM-2 clones expressing PEF6/V5-His-TOPO-Dkk-1 (Dkk1) or vector (EV) was confirmed by immunoblotting blotting with anti-V5 antibody as described in Materials and methods. (C) C2C12 cells were incubated with Dkk1 or EV CM from OPM-2 transfected cells then treated with Wnt3a or control CM. Proteins in lysates were analyzed as in Fig. 2 using the indicated antibodies. (D) C2C12 cells were incubated with sera from MM patients containing low (L1) or high (H1–4) concentrations of Dkk1 protein. Cells were then treated with Wnt3a or control CM and lysates prepared and analyzed as in C.

endogenous non-phosphorylated beta-catenin (Fig. 5E). Yet, exposure of these clones to medium containing 200 ng/ml of BMP-2 significantly inhibited the ALP activity in both Dkk1 and Dkk2 clones, compared with cells expressing empty vector (Fig. 5F). These results suggest that Dkk1 and Dkk2 inhibit mesenchymal cell differentiation by inhibition of autocrine Wntbeta-catenin pool in pre-osteoblasts. Inhibition of ALP activity by siRNA specifically targeting LRP5/6 To determine the specific role of Wnt receptors in mesenchymal cell differentiation in vitro, we used siRNA specific for LRP5 and LRP6 to down regulate gene expression. As shown

Fig. 4. Dkk1 inhibits BMP-2 induced alkaline phosphatase activity. C2C12 (A), Saos-2 (B), and hFOB1.19 (C) cells were cultured for 72 h in DMEM with 2% horse serum containing Wnt3a, BMP-2, or Dkk1 either alone or in the indicated combinations. Cells were lysed and ALP activity measured and normalized to protein concentration. Data represent the mean ± SD (n = 3). ⁎p b 0.05, ⁎⁎p b 0.01, ⁎⁎⁎p b 0.001, and ⁎⁎⁎⁎p b 0.00001 versus control or BMP2 versus BMP-2 plus DKK1.

Y.-W. Qiang et al. / Bone 42 (2008) 669–680

675

Fig. 5. Blocking the canonical β-catenin pathway inhibits TCF/LEF mediated transcription and BMP-2 induced alkaline phosphatase activity. (A) C2C12 cells were transfected with dominant negative β-catenin (DN-β-Cat) or control (pcDNA4his) constructs and the expressions were confirmed by the indicated antibodies by immunoblotting analysis. (B) The positive clones were co-transfected with wild type (TOPflash) or mutant (FOPflash) LEF/TCF reporter luciferase constructs. After transfection, the cells were treated and subjected to luciferase assay as in Fig. 2 E. C2C12 pcDNAhis or DN-β-Cat clone #4 (C) and #5 (D) cells were cultured in DMEM containing 2% horse serum or 100 ng/ml of BMP-2 for 72 h after which ALP activity was measured. (E) C2C12 cells were transfected with empty vector or Dkk1- and Dkk2expressing vectors and the expression of the proteins were determined as in Fig. 3B. The positive clones and control vector were treated with BMP-2 and subjected to ALP activity assay as in D. Results are shown as mean ± SD (n = 3) and are representative of three independent experiments (F). ⁎⁎p b 0.01 and ⁎⁎⁎p b 0.001 versus control.

676

Y.-W. Qiang et al. / Bone 42 (2008) 669–680

Fig. 6. Silencing LRP5/6 mRNA blocks BMP-2 induced alkaline phosphatase activity. C2C12 cells were transiently transfected with 1.0 μg/ml (A) or serial concentration of siRNA specific for LRP5 (B) or LRP6 (C). Forty eight hour after transfection RNA was isolated and subjected to RT-PCR (A) or qPCR (B and C) as described in Materials and methods. (D) The cells transfected with 0.25 or 0.5 μg/ml of siRNA specific for LRP5 or LRP6 were cultured in medium containing 100 ng/ ml of BMP-2 in DMEM. Cells were lysed after 72 h and alkaline phosphatase activity determined. Data represent the mean ± SD (n = 3) of representative experiments. ⁎p b 0.05, ⁎⁎p b 0.01, and ⁎⁎⁎p b 0.001 versus control.

Dkk1 inhibition of pre-osteoblast differentiation is independent of the Smad/Cbaf1/Runx2 pathway To address the question of whether the Wnt and BMP-2 pathways were involved in cross-regulation and co-regulation of downstream elements and/or if Dkk1 directly interferes with BMP-2 signaling to inhibit osteoblast differentiation, we first analyzed the effects of BMP-2 on β-catenin stabilization. As

shown in Fig. 7A, treatment of all pre-osteoblast cell lines with BMP-2 for 8, 24, and 48 h did not result in changes in β-catenin as compared to Wnt3a treated controls. BMP-2 alone did not induce TCT/LEF transcriptional activity, nor did it synergize Wnt3a-stimulated TCF/LEF transcriptional activity as determined by luciferase activity in C2C12 cells transfected with TOPflash constructs (Fig. S1). Thus, BMP-2 did not activate the Wnt-β-catenin pathway at the β-catenin and TCF/LEF

Y.-W. Qiang et al. / Bone 42 (2008) 669–680

677

the Wnt3 role in osteoblast differentiation is independent of Cbfa-1/ Run2 transcriptional activity. Taken together, these results suggest that Wnt does not activate downstream elements in the BMP-2 pathway, BMP-2 does not activate downstream elements in the Wnt pathway, and Dkk1 does not directly inhibit BMP-2 signaling. Discussion

Fig. 7. Wnt3a mediated increase in β-catenin is independent of BMP-2. (A) C2C12, hFOB1.19, MG63, and Saos-2 cells were treated with Wnt3a CM, control CM or 100 ng/ml of BMP2. Lysate protein was subjected to GST–Ecadherin assay and immunoblotting analysis by anti-β-catenin antibody as described in Materials and methods. (B) 50 μg aliquots of protein from cell lysates were resolved on 8% SDS-PAGE and analyzed with the indicated antibodies.

levels. Experiments were next implemented to determine whether Wnt and Dkk1 directly regulate BMP-2 signaling. First, BMPR-I and -II were immunoprecipitated from C2C12 cell lysates followed by blotting with antibodies to LRP5 and LRP6. Complexes of LRP5/6 and BMPR-I/-II were not found following either Wnt3a or Dkk1 treatment (not shown). Similar results were obtained in 293T cells transiently transfected with plasmid containing LRP5/6 cDNA. To ascertain whether Wnt3a can activate BMP-2 downstream targets, Smad phosphorylation was analyzed using antibodies specific to p-Smad1-S463/465. Wnt3a did not induce phosphorylation of Smad1 in C2C12 cells (Fig. 7B) nor Smads5 and 8 as assessed with antibodies to p-Smad5-S463/ 463, and p-Smad8-S426/428 (not shown) in contrast to BMP-2 controls. Because BMP-2 also increases Smad6 gene expression, which serves as inhibitor of BMP-2 signaling [39,40], we examined if Wnt and Dkk1 affect its expression. As expected, BMP-2 induced increase in Smad6 mRNA in a time-dependent manner in C2C12 cells as measured by qPCR analysis (Fig. S2). However, neither Wnt3a alone nor Dkk1 affected BMP-2-induced Smad6 gene expression. Finally, we investigated the effect of Wnt3a and Dkk1 on transcriptional activity of Cbfa-1/Runx2 by transiently transfecting C2C12 with a luciferase reporter construct, Cbfa-1-Luc and treating with Wnt3a and Dkk1. Cell lysates assayed for luciferase revealed no change in activity (data not shown) indicating

A critical element to understanding the osteolytic process in myeloma is to elucidate the soluble factors released by myeloma cells and the associated biochemical pathways that drive unbalanced bone remodeling. Substantial attention has been focused on the role of the RANK/OPG signaling axis, which regulates osteoclastogenesis [1,2]. However, it has also become apparent that myeloma cells also inhibit bone formation by inhibiting osteoblast differentiation [41,42]. Recent studies have suggested that increased Dkk1 protein in MM is associated with the bone remodeling process [28]. Although Wnt signaling plays an important role in regulating osteoblast development [21,22], and elevated expression of a Wnt inhibitor, Dkk1, by MM cells affords a mechanistic hypothesis, a functional analysis of how this factor might contribute to lytic bone disease has not been established. The present studies confirm a role for Wnts in osteoblast differentiation and describe a molecular mechanism by which Dkk1 likely inhibits this differentiation and contributes to myeloma bone disease. Initial experiments (Figs. 1 and 2) demonstrated that mouse and human pluripotent mesenchymal cell lines are capable of transducing a canonical Wnt signal. These cells expressed mRNA corresponding to multiple Wnt receptors and the LRP5/6 co-receptors, suggesting pre-osteoblasts are capable of interacting with Wnt ligand. Treatment with exogenous Wnt3a led to enhanced activation of a functional signaling pathway as evidenced by increases of both uncomplexed and transcriptionally active forms of β-catenin in the cytosol. Accumulation of β-catenin in the cytoplasm leads to nuclear translocation and binding of TCF/LEF family members to form complexes capable of activating transcription. RT-PCR analysis (Fig. 2) revealed that all members of the TCF family are expressed in mouse C2C12 cells, and TCF1, 4, and LEF1 are expressed in human osteoblast-like cell lines. Moreover, functional activation of these transcription factors following Wnt3a treatment was demonstrated using a luciferase reporter construct. It has previously been reported that exogenous Dkk1 blocks Wnt3a-induced stabilization of β-catenin and inhibits activation of the canonical Wnt pathway in MM cells [9,31]. In the present studies (Fig. 3), it was observed that Dkk1 can similarly regulate Wnt-induced stabilization of β-catenin in mouse and human preosteoblasts as Dkk1 protein produced by MM cell lines stably expressing the protein, or MM patient bone marrow sera containing high levels of Dkk1, inhibited stabilization of β-catenin by Wnt3a. It is also of note that Dkk1 completely attenuated Wnt3a-induced TCF/LEF transcriptional activity in mouse preosteoblasts. Conflicting results exist as to the role of Dkk2 and other Dkk family members on pre-osteoblast differentiation [43,44]. Analysis of mRNA expression in the tested cell lines (not shown) revealed relative expression of Dkk family mRNAs in the following order: Dkk3 N Dkk2 N Dkk1= Dkk4. Treatment of pre-osteoblast

678

Y.-W. Qiang et al. / Bone 42 (2008) 669–680

cells with supernatants containing Dkk2 inhibited β-catenin stabilization (not shown) similar to that observed with Dkk1. In addition, over-expression of Dkk1 or Dkk2 directly in C2C12 cells reduced endogenous β-catenin levels. Although osteoblasts clearly respond to Wnt3a by enhanced activation of the canonical Wnt/β-catenin pathway, Wnt3a alone had no apparent effect on differentiation of osteoblast precursors as reflected by ALP production. Surprisingly, Dkk1 (or Dkk2, not shown) significantly blocked BMP-2-induced differentiation (Fig. 4). These results indicate that an autocrine canonical Wnt signal present in OB precursor cells is necessary for BMP2-induced differentiation. This conclusion is supported by several additional observations. First, abundant expression of multiple Wnt mRNAs was observed in all cell lines (unpublished data), in addition to Fz and LRP5/6 receptors (Fig. 1). Second, silencing of LRP5 or LRP6 mRNA expression completely abrogated BMP-2-induced ALP activity (Fig. 6). This result is consistent with previous reports showing that lack of LRP5 [19] and LRP6 [45] reduced bone formation and osteoblast differentiation in a mouse model. Third, higher levels of uncomplexed and transcriptionally active β-catenin protein exist in these cells relative to osteoclasts, MM cells and other cell types (data not shown). Fourth, blockage of endogenous βcatenin by a dominant negative β-catenin construct significantly inhibited BMP-2-induced C2C12 differentiation (Fig. 5) in agreement with data indicating that over-expression of active βcatenin increased bone mass [46]. Finally, constitutive expression of Dkk1 (or Dkk2, not shown) led to, not only reduced endogenous levels of uncomplexed and transcriptionally active β-catenin (data not shown), but also decreased ALP activity following BMP-2 treatment. These results are consistent with previous in vivo studies showing that deletion of Dkk1 leads to increased bone formation and bone mass [27,44], and overexpression of Dkk1 in transgenic mice results in osteopenia [26]. The present study provides direct evidence that endogenous βcatenin levels in mesenchymal cells are necessary for osteoblast differentiation, and Dkk1 from MM cells inhibits this process. Since Dkk1 can block BMP-2 induced ALP production and Wnt signaling is necessary to induce differentiation, the possibility exists for cross-regulation between these two pathways. Experiments to test this hypothesis revealed that BMP-2 treatment alone did not induce increased levels of β-catenin over steady state (Fig. 7), nor did it affect TCF/LEF transcriptional activity, nor did Wnt3a increase ALP activity (Fig. 4). Furthermore, no association was observed between Dkk1 and BMP-2 receptors, and neither Wnt3a nor Dkk1 treatment led to activation of Smads1, 5, and 8 (important downstream targets of BMP-2 activation that play a pivotal role in BMP-2-induced mesenchymal cell differentiation) [47]. Moreover, Cbfa-1/Runx2 transcription factor activity and increase in Smad6 mRNA induced by BMP treatment did not change in response to Wnt or DKK1, further indicating that Wnt-signaling does not activate the BMP pathway. However, the fact that Dkk1 did inhibit the canonical Wnt pathway and BMP-2-induced osteoblast differentiation in the present study suggests that there is indeed co-regulation between BMP-2 and Wnt signaling of osteoblast differentiation. The nature of this co-regulation is the focus of active investigation.

In contrast to the lack of cross-regulation described above, other studies have suggested that BMP-2 increases endogenous Wnt mRNA expression to promote increased ALP activity [22,48]. Furthermore, BMP-2 and a β-catenin mutant with constitutive transcriptional activity (ΔN151) synergized to stimulate ALP activity, osteocalcin gene expression, and matrix mineralization [49]. However, in the present experiments, BMP-2 did not induce increased stabilization of cytosolic, free β-catenin in mouse and human pre-osteoblast cells, nor BMP-2 alone was able to induce TCF/LEF transcriptal activity, nor did BMP-2 synergyze with Wnt3a-a induced TCF/LEF activity indicating that, under these conditions, BMP-2 is unlikely to alter baseline or steady state levels of Wnt signaling sufficient, and required, for BMP-2 induced differentiation. It appears that cross talk between BMP-2 and a canonical Wnt pathway does not occur at, or above, the analyzed downstream targets of each signaling pathway. Consistent with this hypothesis, Nakashima et al. have previously reported that BMP-2 alone failed to increase TCF/LEF activity [50] although these two pathways are required for pre-osteoblast differentiation. Moreover, Mbalaviele et al. provided in vivo evidence that BMP-2 does not influence TCF/LEF activity related to that activated by β-catenin mutant (ΔN151) in C3H10T1/2 cells [49]. However, we cannot exclude the possibility that interactions between Wnt and BMP2 signaling pathways occur through alternate cascades as yet unidentified. In fact, in other systems, it has been reported that βcatenin and LEF/TCF form complexes with Smads [51]. Additionally, Wnt3a and BMP-2 [52] can induce expression of the ID2 gene and both induced MSX1 gene expression [53]. Further studies will be required to clarify these differences. Since it is thought that either LRP5 or LRP6 can act redundantly, why did we observe an almost complete loss of BMP-2 induced ALP activity, when we silenced only one of the two? The reasons for this phenomenon are not clear at the moment. It may be that the amount of LRP5 or LRP6 on the cell surface tightly regulates Wnt signaling required for BMP-2 induced ALP. Alternatively, while required, either of them alone is not sufficient. Indeed, Wnt-1 induced TCF/LEF transcriptional activity was almost completely blocked in fibroblast of LRP6 null mice in the presence of LRP5 [45]. On the other hand, expression of loss-function of LRP5 mutant almost completely abolished BMP-2 induced ALP activity in the presence of LRP6 in ST2 pluripotent bone marrow stromal cells [16]. Another explanation could be that the trafficking of LRP5 and LRP6 on the cell surface might regulate Wnt signaling. This is consistent with recent studies that show that R-Spondin-1 regulates the cell surface levels of LRP6 by interfering with Dkk1/Kremenmediated internalization of LRP6 [54]. Further studies will be needed to distinguish these hypotheses. In conclusion the above studies have revealed that autocrine Wnt signaling in osteoblasts is necessary to promote BMP-2mediated differentiation of pre-osteoblast cells, while Wnt signaling alone is not capable of inducing such differentiation. Dkk1 inhibits this process and may be a key factor regulating preosteoblast differentiation. These observations raise the possibility that Dkk1 could be an important molecular target for novel therapeutic approaches to modulate myeloma bone disease.

Y.-W. Qiang et al. / Bone 42 (2008) 669–680

Acknowledgments The authors would like to thank Drs Jeffery Rubin and Ying Zang in The Laboratory of Cellular and Molecular Biology, NIC, NIH for providing reagents. We would like to acknowledge the technical assistance of Yu Chen, Nathan Brown and Owen Stephens. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.bone.2007.12.006. References [1] Pearse RN, Sordillo EM, Yaccoby S, Wong BR, Liau DF, Colman N, et al. Multiple myeloma disrupts the TRANCE/osteoprotegerin cytokine axis to trigger bone destruction and promote tumor progression. Proc Natl Acad Sci U S A 2001;98(20):11581–6. [2] Sezer O, Heider U, Zavrski I, Kuhne CA, Hofbauer LC. RANK ligand and osteoprotegerin in myeloma bone disease. Blood 2003;101(6):2094–8. [3] Berenson JR. Zoledronic acid in cancer patients with bone metastases: results of Phase I and II trials. Semin Oncol 2001;28(2 Suppl 6):25–34. [4] Bataille R, Chappard D, Marcelli C, Dessauw P, Sany J, Baldet P, et al. Mechanisms of bone destruction in multiple myeloma: the importance of an unbalanced process in determining the severity of lytic bone disease. J Clin Oncol 1989;7(12):1909–14. [5] Taube T, Beneton MN, McCloskey EV, Rogers S, Greaves M, Kanis JA. Abnormal bone remodelling in patients with myelomatosis and normal biochemical indices of bone resorption. Eur J Haematol 1992;49(4):192–8. [6] Hashimoto T, Abe M, Oshima T, Shibata H, Ozaki S, Inoue D, et al. Ability of myeloma cells to secrete macrophage inflammatory protein (MIP)1alpha and MIP-1beta correlates with lytic bone lesions in patients with multiple myeloma. Br J Haematol 2004;125(1):38–41. [7] Nusse R. Wnt signaling in disease and in development. Cell Res 2005;15 (1):28–32. [8] Westendorf JJ, Kahler RA, Schroeder TM. Wnt signaling in osteoblasts and bone diseases. Gene 2004;341:19–39. [9] Qiang Y, Rudikoff S. Wnt signaling in B and T lymphocytes. Front Biosci 2004;9:1000–10. [10] Qiang Y, Walsh K, Yao L, Kedei N, Blumberg P, Rubin J, et al. Wnts induce migration and invasion of myeloma plasma cells. Blood 2005;106:1786–93. [11] He X, Semenov M, Tamai K, Zeng X. LDL receptor-related proteins 5 and 6 in Wnt/beta-catenin signaling: arrows point the way. Development 2004;131(8):1663–77. [12] Nelson WJ, Nusse R. Convergence of Wnt, beta-catenin, and cadherin pathways. Science 2004;303(5663):1483–7. [13] Holmen SL, Zylstra CR, Mukherjee A, Sigler RE, Faugere MC, Bouxsein ML, et al. Essential role of beta-catenin in postnatal bone acquisition. J Biol Chem 2005;280(22):21162–8. [14] Stewart JP, Shaughnessy Jr JD. Role of osteoblast suppression in multiple myeloma. J Cell Biochem 2006;98(1):1–13. [15] Bradbury JM, Niemeyer CC, Dale TC, Edwards PA. Alterations of the growth characteristics of the fibroblast cell line C3H 10T1/2 by members of the Wnt gene family. Oncogene 1994;9(9):2597–603. [16] Gong Y, Slee RB, Fukai N, Rawadi G, Roman-Roman S, Reginato AM, et al. LDL receptor-related protein 5 (LRP5) affects bone accrual and eye development. Cell 2001;107(4):513–23. [17] Boyden LM, Mao J, Belsky J, Mitzner L, Farhi A, Mitnick MA, et al. High bone density due to a mutation in LDL-receptor-related protein 5. N Engl J Med 2002;346(20):1513–21. [18] Little RD, Carulli JP, Del Mastro RG, Dupuis J, Osborne M, Folz C, et al. A mutation in the LDL receptor-related protein 5 gene results in the autosomal dominant high-bone-mass trait. Am J Hum Genet 2002;70(1):11–9.

679

[19] Kato M, Patel MS, Levasseur R, Lobov I, Chang BH, Glass II DA, et al. Cbfa1-independent decrease in osteoblast proliferation, osteopenia, and persistent embryonic eye vascularization in mice deficient in Lrp5, a Wnt coreceptor. J Cell Biol 2002;157(2):303–14. [20] Bennett CN, Longo KA, Wright WS, Suva LJ, Lane TF, Hankenson KD, et al. Regulation of osteoblastogenesis and bone mass by Wnt10b. Proc Natl Acad Sci U S A 2005;102(9):3324–9. [21] Hu H, Hilton MJ, Tu X, Yu K, Ornitz DM, Long F. Sequential roles of Hedgehog and Wnt signaling in osteoblast development. Development 2005;132(1):49–60. [22] Rawadi G, Vayssiere B, Dunn F, Baron R, Roman-Roman S. BMP-2 controls alkaline phosphatase expression and osteoblast mineralization by a Wnt autocrine loop. J Bone Miner Res 2003;18(10):1842–53. [23] Kawano Y, Kypta R. Secreted antagonists of the Wnt signalling pathway. J Cell Sci 2003;116(Pt 13):2627–34. [24] Mao B, Wu W, Li Y, Hoppe D, Stannek P, Glinka A, Niehrs C. LDLreceptor-related protein 6 is a receptor for Dickkopf proteins. Nature 2001;411:321–5. [25] Mao B, Wu W, Davidson G, Marhold J, Li M, Mechler BM, et al. Kremen proteins are Dickkopf receptors that regulate Wnt/beta-catenin signalling. Nature 2002;417(6889):664–7. [26] Li J, Sarosi I, Cattley RC, Pretorius J, Asuncion F, Grisanti M, et al. Dkk1mediated inhibition of Wnt signaling in bone results in osteopenia. Bone 2006;39(4):754–66. [27] Morvan F, Boulukos K, Clement-Lacroix P, Roman Roman S, Suc-Royer I, Vayssiere B, et al. Deletion of a single allele of the Dkk1 gene leads to an increase in bone formation and bone mass. J Bone Miner Res 2006;21 (6):934–45. [28] Tian E, Zhan F, Walker R, Rasmussen E, Ma Y, Barlogie B, et al. The role of the Wnt-signaling antagonist DKK1 in the development of osteolytic lesions in multiple myeloma. N Engl J Med 2003;349(26):2483–94. [29] Hall CL, Bafico A, Dai J, Aaronson SA, Keller ET. Prostate cancer cells promote osteoblastic bone metastases through Wnts. Cancer Res 2005;65 (17):7554–60. [30] Chung EJ, Hwang SG, Nguyen P, Lee S, Kim JS, Kim JW, et al. Regulation of leukemic cell adhesion, proliferation, and survival by betacatenin. Blood 2002;100(3):982–90. [31] Qiang Y, Endo Y, Rubin J, Rudikoff S. Wnt signaling in B-cell neoplasia. Oncogene 2003;22:1536–45. [32] Qiang Y, Kopantzev E, Rudikoff S. Insulinlike growth factor-I signaling in multiple myeloma: downstream elements, functional correlates, and pathway cross-talk. Blood 2002;99:4138–46. [33] Bafico A, Gazit A, Wu-Morgan SS, Yaniv A, Aaronson SA. Characterization of Wnt-1 and Wnt-2 induced growth alterations and signaling pathways in NIH3T3 fibroblasts. Oncogene 1998;16(21):2819–25. [34] Korinek V, Barker N, Morin PJ, van Wichen D, de Weger R, Kinzler KW, et al. Constitutive transcriptional activation by a beta-catenin-Tcf complex in APC−/− colon carcinoma. Science 1997;275(5307):1784–7. [35] Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) method. Methods 2001;25(4):402–8. [36] Cheng SL, Lecanda F, Davidson MK, Warlow PM, Zhang SF, Zhang L, et al. Human osteoblasts express a repertoire of cadherins, which are critical for BMP-2-induced osteogenic differentiation. J Bone Miner Res 1998;13(4):633–44. [37] van Noort M, Meeldijk J, van der Zee R, Destree O, Clevers H. Wnt signaling controls the phosphorylation status of beta-catenin. J Biol Chem 2002;277(20):17901–5. [38] Nishimura R, Kato Y, Chen D, Harris SE, Mundy GR, Yoneda T. Smad5 and DPC4 are key molecules in mediating BMP-2-induced osteoblastic differentiation of the pluripotent mesenchymal precursor cell line C2C12. J Biol Chem 1998;273(4):1872–9. [39] Itoh F, Asao H, Sugamura K, Heldin CH, ten Dijke P, Itoh S. Promoting bone morphogenetic protein signaling through negative regulation of inhibitory Smads. EMBO J 2001;20(15):4132–42. [40] Wang Q, Wei X, Zhu T, Zhang M, Shen R, Xing L, et al. Bone morphogenetic protein 2 activates Smad6 gene transcription through bonespecific transcription factor Runx2. J Biol Chem 2007;282(14):10742–8.

680

Y.-W. Qiang et al. / Bone 42 (2008) 669–680

[41] Giuliani N, Rizzoli V, Roodman GD. Multiple myeloma bone disease: pathophysiology of osteoblast inhibition. Blood 2006;108:3992–6. [42] Oshima T, Abe M, Asano J, Hara T, Kitazoe K, Sekimoto E, et al. Myeloma cells suppress bone formation by secreting a soluble Wnt inhibitor, sFRP-2. Blood 2005;106(9):3160–5. [43] Li X, Liu P, Liu W, Maye P, Zhang J, Zhang Y, et al. Dkk2 has a role in terminal osteoblast differentiation and mineralized matrix formation. Nat Genet 2005;37(9):945–52. [44] van der Horst G, van der Werf SM, Farih-Sips H, van Bezooijen RL, Lowik CW, Karperien M. Downregulation of Wnt signaling by increased expression of Dickkopf-1 and -2 is a prerequisite for late-stage osteoblast differentiation of KS483 cells. J Bone Miner Res 2005;20(10):1867–77. [45] Kokubu C, Heinzmann U, Kokubu T, Sakai N, Kubota T, Kawai M, et al. Skeletal defects in ringelschwanz mutant mice reveal that Lrp6 is required for proper somitogenesis and osteogenesis. Development 2004;131 (21):5469–80. [46] Glass II DA, Bialek P, Ahn JD, Starbuck M, Patel MS, Clevers H, et al. Canonical Wnt signaling in differentiated osteoblasts controls osteoclast differentiation. Dev Cell 2005;8(5):751–64. [47] Canalis E, Economides AN, Gazzerro E. Bone morphogenetic proteins, their antagonists, and the skeleton. Endocr Rev 2003;24(2):218–35.

[48] Chen Y, Whetstone HC, Youn A, Nadesan P, Chow EC, Lin AC, et al. Beta-catenin signaling pathway is crucial for bone morphogenetic protein 2 to induce new bone formation. J Biol Chem 2006;282:526–33. [49] Mbalaviele G, Sheikh S, Stains JP, Salazar VS, Cheng SL, Chen D, et al. Beta-catenin and BMP-2 synergize to promote osteoblast differentiation and new bone formation. J Cell Biochem 2005;94(2):403–18. [50] Nakashima A, Katagiri T, Tamura M. Cross-talk between Wnt and bone morphogenetic protein 2 (BMP-2) signaling in differentiation pathway of C2C12 myoblasts. J Biol Chem 2005;280(45):37660–8. [51] Hu MC, Rosenblum ND. Smad1, beta-catenin and Tcf4 associate in a molecular complex with the Myc promoter in dysplastic renal tissue and cooperate to control Myc transcription. Development 2005;132(1):215–25. [52] Willert J, Epping M, Pollack JR, Brown PO, Nusse R. A transcriptional response to Wnt protein in human embryonic carcinoma cells. BMC Dev Biol 2002;2:8–15. [53] Binato R, Alvarez Martinez CE, Pizzatti L, Robert B, Abdelhay E. SMAD 8 binding to mice Msx1 basal promoter is required for transcriptional activation. Biochem J 2006;393(Pt 1):141–50. [54] Binnerts ME, Kim KA, Bright JM, Patel SM, Tran K, Zhou M, et al. R-Spondin1 regulates Wnt signaling by inhibiting internalization of LRP6. Proc Natl Acad Sci U S A 2007;104(37):14700–5.