A novel interaction between insulin-like growth factor binding protein-6 and the vitamin D receptor inhibits the role of vitamin D3 in osteoblast differentiation

Molecular and Cellular Endocrinology 338 (2011) 84–92

Contents lists available at ScienceDirect

Molecular and Cellular Endocrinology journal homepage: www.elsevier.com/locate/mce

A novel interaction between insulin-like growth factor binding protein-6 and the vitamin D receptor inhibits the role of vitamin D3 in osteoblast differentiation夽 Jian Cui, Chunling Ma, Jia Qiu, Xiaoli Ma, Xin Wang, Hong Chen ∗ , Bingren Huang ∗ National Laboratory of Medical Molecular Biology, Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences & Peking Union Medical College, 5 Dong Dan San Tiao, Beijing 100005, China

a r t i c l e

i n f o

Article history: Received 25 July 2010 Received in revised form 25 January 2011 Accepted 10 March 2011 Keywords: IGFBP-6 VDR Osteoblast differentiation

a b s t r a c t Insulin-like growth factor binding protein-6 (IGFBP-6) is a secreted glycoprotein that reduces the bioavailability of IGFs. It has both IGF-dependent and -independent effects on cell growth, however the mechanisms responsible for its IGF-independent actions of IGFBP-6 are not fully understood. In previous studies, we have shown that recombinant IGFBP-6 can be internalized and translocated to the nucleus. The present study shows that IGFBP-6 interacts with the vitamin D receptor (VDR). Physical interactions between IGFBP-6 and the VDR were confirmed by GST pulldown and co-immunoprecipitation assays. We also determined that the interaction binding sites were on the C-terminal region of the VDR. This interaction can influence retinoid X receptor (RXR):VDR heterodimerization. Furthermore, immunofluorescence colocalization studies showed that IGFBP-6 colocalized with the VDR predominantly in the cell’s nucleus. Inductions of osteocalcin and growth hormone promoter activities by 1,25-dihydroxyvitamin D3 (1,25(OH)2 D3 ) were significantly decreased when cells were co-transfected with IGFBP-6 and the VDR compared with cells transfected with the VDR only. Moreover, we found that alkaline phosphatase activity (ALP, a general marker of osteoblast differentiation) was significantly decreased in osteoblast-like cells when they were transfected with IGFBP-6 in the presence of 1,25(OH)2 D3 . No obvious difference in ALP activity was observed when cells were transfected with IGFBP-6 and endogenous VDR was knocked down by siRNA. These results demonstrate that IGFBP-6 inhibits osteoblastic differentiation mediated by 1,25(OH)2 D3 and the VDR through interacting with the VDR and inhibiting its function. This is a novel mechanism for IGFBP-6. © 2011 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Insulin-like growth factor binding protein (IGFBP)-6 is a member of a family containing six high affinity IGFBPs that regulate IGF-II activity and also show IGF-independent effects (Firth and Baxter, 2002). IGFBP-6 is unique among the IGFBPs for its high IGFII binding specificity. It can inhibit IGF-II actions in vitro and in vivo by preventing IGF from binding to cell surface receptors (Firth and Baxter, 2002; Mohan and Baylink, 2002). Although IGFBP-6’s major action appears to be inhibiting IGF-II activity, a number of studies suggest that it may also have IGF-independent actions, such as growth inhibition and apoptosis (Sueoka et al., 2000; Gallicchio et al., 2001; Firth and Baxter, 2002; Mohan and Baylink, 2002). Studies on osteoblast differentiation show that overexpression of IGFBP-6 suppresses human and murine osteoblast differentiation

夽 Grant supporter: National Laboratory of Medical Molecular Biology. ∗ Corresponding authors. Tel.: +86 10 65296980; fax: +86 10 65228797. E-mail addresses: [email protected] (H. Chen), [email protected] (B. Huang). 0303-7207/$ – see front matter © 2011 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.mce.2011.03.011

(Strohbach et al., 2008). It has also been reported that all-trans retinoic acid (atRA), which inhibits osteoblast phenotypes and osteoblast differentiation gene expression markers, can increase the expression of IGFBP-6 in human osteoblasts because there are three retinoic acid response elements (RAREs) in the promoter of IGFBP-6 (Dailly et al., 2001; Yan et al., 2001). These studies suggest that IGFBP-6 may regulate osteoblast differentiation through unknown mechanisms. The vitamin D receptor (VDR) is a member of the nuclear receptor superfamily of ligand-activated transcription factors. It regulates the expression of target genes by binding to 1,25dihydroxyvitamin D3 (1,25-(OH)2 D3 ) (McDonnell et al., 1987; Haussler et al., 1998). The classic roles of VDR and its ligands are to stimulate calcium absorption in the intestines to maintain normal calcium levels and to indirectly regulate bone mineralization (van Driel et al., 2004). Studies using mice that lack the VDR demonstrate its essential functions in bone formation and metabolism. In growing animals, the primary role of the VDR is to maintain high levels of intestinal calcium absorption, and a loss of the VDR results in rickets (Panda et al., 2004). Transgenic mice that over-express the VDR in mature osteoblasts, which are the bone-forming cells that

J. Cui et al. / Molecular and Cellular Endocrinology 338 (2011) 84–92

deposit bone extracellular matrix have higher rates of bone formation (Gardiner et al., 2000). It has been reported that 1,25(OH)2 D3 promotes differentiation and maturation of the osteoblast (Gurlek and Kumar, 2001). Treating osteoblasts with 1,25(OH)2 D3 results in inhibiting proliferation, upregulating osteoblast-associated genes, such as osteocalcin and osteopontin, and stimulating calcium accumulation (Gurlek et al., 2002). We report here that IGFBP-6 interacts with the VDR by binding to its C-terminal region, which also contains the RXR heterodimerization motif and ligand-dependent transcriptional function region. Immunofluorescence colocalization and nuclear cytoplasmic fractionation studies showed that IGFBP-6 and the VDR co-localized predominantly to the nucleus. Further exploration revealed that IGFBP-6 inhibited osteocalcin and growth hormone promoter activities that are induced by 1,25(OH)2 D3 . Moreover, IGFBP-6 inhibited alkaline phosphatase (ALP) activity, which is a general marker of osteoblast differentiation in U2-OS and MG-63 cells. When the endogenous VDR was knocked down by siRNA in these cells, no obvious difference in the ALP activity was observed when these cells were transfected with IGFBP-6. Thus, our work suggests that IGFBP-6 could inhibit the role of 1,25(OH)2 D3 in osteoblast differentiation by entering the nucleus and binding to the VDR. This is a novel mechanism for IGFBP-6.

85

3 mM EGTA, 1 mM sodium orthovanadate, 1 mM DTT, 10 mM PNPP, and 10 ␮g/ml aprotinin) after being washed with PBS. Cell lysates were resolved by SDS-PAGE before being transferred to nitrocellulose membranes (Pall Corporation, Pensacola, FL). Nitrocellulose membranes were then incubated with 5% (w/v) nonfat dry milk in TBST washing buffer (20 mM Tris–HCl, pH 7.6, 150 mM NaCl, and 0.1% Tween 20) for 60 min at room temperature to block nonspecific protein binding. Primary antibodies (1:500) were diluted in washing buffer containing 5% (w/v) nonfat dry milk and then applied to the membranes overnight at 4 ◦ C. After being washed with TBST 3 times, the membranes were incubated with peroxidase-conjugated secondary antibodies for 60 min at room temperature and then washed again. Immunoreactive bands were visualized by Super Signal chemiluminescence (Pierce Chemical, Rockford, IL). 2.5. GST pulldown assays

Antibodies against IGFBP-3, IGFBP-6, VDR, RXR, GAPDH, Histone1, GST were from Santa Cruz Biotechnology (Santa Cruz, CA). 1,25(OH)2 D3 and antibodies against beta-actin, Flag-, HA- and Myc-tag were purchased from Sigma (Deisenhofen, Germany). Antibodies against LAMP-1 were purchased from Abcam (Cambridge, UK). Peroxidase-conjugated secondary antibodies, rhodamine (TRITC) and fluorescein isothiocyanate (FITC)-conjugated secondary antibodies were purchased from Zhongshan Goldenbridge Biotechnology Corporation (Beijing, China). 1,25(OH)2 D3 was dissolved in 95% ethanol by volume. 1,25(OH)2 D3 was dissolved in 95% ethanol by volume. Whenever 1,25(OH)2 D3 was used, 95% ethanol was also used so that the concentration of ethanol (<0.1%) was equal in all wells.

The GST-VDR fusion vector encoded the full-length VDR molecule. This fusion protein was produced in Escherichia coli BL21 (DE3) after isopropy-␤-dthiogalactoside (IPTG) induction for 10 h at room temperature. The bacteria were disrupted by sonication, and recombinant GST fusion proteins were captured from cell lysates using glutathione-Sepharose beads (Sigma, Deisenhofen, Germany). The purified GST-VDR that was bound to the beads was incubated with 293T cell lysates that were transiently transfected with Flag-IGFBP-6 and they were separated by centrifugation. The bound proteins were analyzed by SDS-PAGE followed by immunoblotting (IB) using an anti-Flag antibody. Experiments were repeated three times. In another series of experiments, the same volume of 293T cell lysates that were transfected with Flag-IGFBP-6 was incubated with decreasing amounts of GST-VDR (100, 50, 25, and 10 ␮g). The complexes were incubated and swirled at 4 ◦ C for 4 h. Bound IGFBP-6 was analyzed by IB with an anti-Flag antibody. In another series of experiments, GST-VDR was immobilized on glutathioneSepharose beads and incubated with an equal volume of 293T cell lysates expressing HA-IGFBP-6. IGF-II was added to the complex at concentrations of 0, 10 or 100 ng/ml at the same time. Then the complex was incubated and swirled at 4 ◦ C for 4 h. Bound IGFBP-6 was analyzed by SDS-PAGE followed by IB with an anti-IGFBP-6 antibody. Experiments were repeated three times. GST-VDR (0.4 nM) was immobilized on glutathione beads in another series of experiments. Equal amounts of GST-VDR were preincubated with IGFBP-6 (2 nM or 4 nM; expressed in Pichia pastoris and purified (Chen et al., 2007)) or with BSA (4 nM) for 1 h at 4 ◦ C. The same volume of 293T cell lysates, which was transiently transfected with plasmids expressing the RXR, was added to each sample and incubated and swirled at 4 ◦ C for 3 h. The amount of the RXR bound to GST-VDR was determined by IB using an anti-RXR antibody. Experiments were repeated three times.

2.2. Cell culture

2.6. Co-immunoprecipitation

The human cell lines kept in our lab. 293T, HeLa, U2-OS and MG-63 were maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% FBS, 100 units/mL penicillin, and 100 ␮g/mL streptomycin. DMEM and the antibiotics (penicillin/streptomycin) were purchased from Gibco Life Technologies (Karlsruhe, Germany). For each cell line, a cell stock was made 5–7 days after resuscitation. The continuous passage times were kept within 1–2 months.

Cells (293T) were transiently transfected with pCMV-Myc-VDR and pCMV-FlagIGFBP-6 or a control vector. After transfection for 24 h, cells were stimulated with 1,25(OH)2 D3 (10−8 M) for 24 h. The cells were then suspended and lysed in 50 mM Tris–HCl buffer (pH 7.5) with 0.5% Triton X-100 and protease inhibitors (1 mM phenylmethylsulfonyl fluoride, 1 ␮g/ml pepstatin, 1 ␮g/ml leupeptin, and 10 ␮g/ml aprotinin; Roche, Mannheim, Germany). An anti-Flag antibody (1 ␮g; Sigma, Deisenhofen, Germany) was added to each sample, and they were placed on a rotating stirrer at 4 ◦ C for 12 h. Then protein-G Sepharose 4 Fast Flow beads (30 ␮L, Sigma, Deisenhofen, Germany) were added to each sample according to the manufacturer’s protocol. The solutions were rotated overnight at 4 ◦ C, and the beads were then washed 3 times with Tris–HCl buffer (50 mM; pH 7.5) with 0.5% Triton X-100. In one experiment, the immune complexes were solubilized in SDS-PAGE loading buffer and separated on SDS-PAGE gels following IB to detect VDR in the complexes with a rabbit antihuman VDR antibody (Santa Cruz Biotechnology, Santa Cruz, CA). In a second experiment, the co-transfected cells were lysed, and a primary antibody for the Myc-tag was added to the cell lysates first. Next, the immune complexes were analyzed on SDS-PAGE gels following IB to detect IGFBP-6 in the complexes with a goat antihuman IGFBP-6 antibody (Santa Cruz Biotechnology, Santa Cruz, CA).

2. Materials and methods 2.1. Reagents

2.3. Transfection of plasmids and siRNAs The VDR, IGFBP-3 and pGL3-basic-GH (growth hormone) sequences were kindly provided by Prof. X. Peng (IIT Research Institute, Chicago), Prof. Werner Zwerschke (Cell Metabolism and Differentiation Research Group, Institute for Biomedical Aging Research of the Austrian Academy of Sciences, Innsbruck, Austria) and Prof. Norman Eberhardt (Department of Medicine/Division of Endocrinology, Minnesota), respectively. pCMV-Flag-IGFBP-6 (mature peptide form), pCMV-Flag-IGFBP-6-M, pCMV-Myc-VDR, pCMV-Flag-VDR115–427 , pCMV-Flag-VDR1–91 , pCMV-Flag-IGFBP-3, pCMV-HA-IGFBP-1, pGEX-6p-1-VDR, pGEX-6p-1-VDR115–427 , pGEX-6p-1-VDR1–91 , pGEX-6p-1-RXR, pGL3-basic-VDRE and pGL3-basic-Oste were constructed. IGFBP1, IGFBP-3 and IGFBP-6 were without signal peptide sequences (mature peptide forms). Both the VDR and the RXR used in the experiments were full-length sequences. pCMV-HA-IGFBP-6 was kept in our lab. Plasmids were transfected into cells using Lipofectamin 2000 (Invitrogen, Karlsruhe, Germany) or Fugene HD (Roche, Mannheim, Germany) according to the manufacturer’s protocols. The sequence of the siRNA oligonucleotide that was used to silence the VDR was 955-GUGCCAUUGAGGUCAUCAUTT-975. The irrelevant siRNA with random nucleotides was 5 -UUCUCCGAACGUGUCACGUTT-3 (negative control siRNA, N.C. siRNA). Both oligonucleotides were synthesized from the Genepharma Corporation (Shanghai, China). siRNAs were transfected into cells using Lipofectamine 2000 (Invitrogen), according to the manufacturer’s protocols. 2.4. Immunoblotting analyses Adherent cells were harvested with a cell scraper (Costars, Cambridge, MA) in ice-cold lysis buffer (0.5% NP-40, 20 mM Tris–HCl, pH 7.6, 250 mM NaCl, 3 mM EDTA,

2.7. Immunofluorescence Cells (293T and HeLa) were transfected with plasmids expressing pCMV-FlagIGFBP-6, pCMV-Myc-VDR or a pCMV vector. After transfection for 24 h, cells were stimulated with 1,25-(OH)2 D3 (10−8 M) for 12 h. The cells were then fixed with 4% (w/v) paraformaldehyde in PBS for 15 min and permeabilized with 0.3% Triton X100 in PBS for 15 min at room temperature. The nonspecific sites were blocked by incubating the cells with 1% bovine serum albumin in PBS for 60 min at room temperature. Cells were then rinsed in PBS containing 0.05% Tween-20 for 5 min and incubated overnight at 4 ◦ C with a VDR antibody (rabbit anti-human, 1:50), an IGFBP-6 antibody (goat anti-human, 1:50) or an anti-Flag-tag antibody (murine, 1:100), which were diluted in blocking buffer. After the cells were washed three times in PBS containing 0.05% Tween-20 for 10 min each, they were incubated with rhodamine (TRITC) or fluorescein isothiocyanate (FITC)-conjugated secondary anti-

86

J. Cui et al. / Molecular and Cellular Endocrinology 338 (2011) 84–92

bodies (1:100) for 60 min at room temperature. The cells were washed again as stated above, incubated with 4,6-diamidine-2-phenylindole (DAPI; 1 ␮g/mL), and sealed with coverslips. Images were acquired using a laser scanning confocal microscope (Leica, Deerfield, IL).

3. Results

2.8. Nuclear and cytoplasmic fractionations

To investigate the specificity of the VDR and IGFBP-6 bond, we performed GST pulldown experiments using recombinant fused VDR linked to glutathione S-transferase (GST). Unfused GST and fused GST-VDR proteins were immobilized onto beads and incubated with an equal volume of 293T cell lysates, which were transfected with pCMV-Flag-IGFBP-6. The bound protein was identified by IB using specific anti-Flag antibodies (Fig. 1A). Compared with GST alone, the GST-VDR protein bound IGFBP-6, and no IGFBP6 band was detected with GST. We used Flag-RXR or Flag-IGFBP-3 as a positive control (Fig. 1A) and found that RXR bound the VDR as a heterodimer (Sanchez-Martinez et al., 2008). IGFBP-3, which has a high homology to IGFBP-6, can also bind to the VDR to modulate its transcription (Schedlich et al., 2007). We used IGFBP-1 as a negative control (Fig. 1A), which cannot bind to the VDR (Schedlich et al., 2007). Equal volumes of 293T cell lysates were transiently transfected with pCMV-Flag-IGFBP-6 and incubated with decreasing amounts of the GST-VDR protein (100, 50, 25, and 10 ␮g). The bound IGFBP-6 was investigated by IB, and its levels appeared to decrease as the GST-VDR protein was decreased in a dose-dependent manner (Fig. 1B). To investigate the specificity of the VDR bound to IGFBP-6 in vivo, 293T cell lysates were transfected with Flag-IGFBP-6 and Myc-VDR, and they were subjected to immunoprecipitation. Using anti-Myc antibodies, we immunoprecipitated Myc-VDR and any interacting molecules from the lysates expressing Flag-IGFBP-6 with Myc-VDR or a Myc-vector control and then precipitated the complexes with protein G Sepharose beads (Fig. 1C). By performing an IB with an anti-IGFBP-6 antibody, IGFBP-6 was found in the immunoprecipitates of the cells that were co-transfected with Flag-IGFBP-6 and Myc-VDR, and no IGFBP-6 was detected in the control immunoprecipitates. We performed another immunoprecipitation experiment with anti-Flag antibodies and then performed an IB with antihuman VDR antibodies (Fig. 1D). A substantial amount of the VDR was detected when the VDR and IGFBP-6 were co-transfected in the 293T cells, which is indicative of an association between FlagIGFBP-6 and Myc-VDR proteins. To further confirm the association between IGFBP-6 and the VDR, a semi-endogenous co-immunoprecipitation experiment was performed with exogenous IGFBP-6 and endogenous VDR in HeLa cells. IGFBP-6 was immunoprecipitated with an anti-VDR antibody, and no IGFBP-6 was detected by IB when the cells were transfected with a Flag-vector control (Fig. 1E). These results indicate that the binding of IGFBP-6 and the VDR is specific.

For nuclear and cytoplasmic fractionations, cells (293T and HeLa) were transiently transfected with plasmids expressing Flag-IGFBP-6 or Myc-VDR. After transfection for 24 h, cells were stimulated with 1,25-(OH)2 D3 (10−8 M) for 24 h. The cells were harvested with EDTA-trypsin, centrifuged (1500 rpm, 6 min at 4 ◦ C), and washed twice with ice-cold phosphate-buffered saline (PBS). The cell pellets were lysed with buffer A (10 mM Tris–HCl, pH 7.5, 10 mM NaCl, 3 mM MgCl2 , and 0.5% NP-40) and swirled for 10 s. The cell lysates were incubated for 5 min on ice and centrifuged for 6 min at 1500 rpm and 4 ◦ C. The supernatants were considered the cytoplasmic fractions. The cellular pellets were resuspended with buffer A and treated again as described above to eliminate the cytoplasmic remainders. Finally, the cellular pellets were extracted with SDS-PAGE sample buffer for 5 min at 95 ◦ C and centrifuged (12,000 rpm, 5 min at 20 ◦ C), and the supernatants were considered the nuclear fractions. The protein GAPDH and lysosomal protein LAMP-1 (Chen et al., 1985) were used as cytoplasmic markers, and Histone 1 was used as a nuclear marker.

2.9. Reporter gene luciferase assay Three kinds of luciferase reporter genes were used in our experiments. The human osteocalcin promoter sequence between −986 and +22 (relative to the transcriptional starting point at +1) was amplified by PCR using genomic DNA as a template with the following primers: oligonucleotide primer 1: 5 -GGAAGATCTCAGGCTGGGATGTTCTGTAC-3 , and primer 2: 5 -CCCAAGCTTCCAGGAGTGTGAGGGCTCT-3 . These primers were designed for cloning into Bgl II and Hind III restriction sites in the promoter-less (minus) luciferase reporter vector pGL3-basic (Promega, Madison, WI) as pGL3-basic-Oste. The pGL3basic-VDREs luciferase reporter is a luciferase gene that is driven by two copies of the VDRE (5 -AGCTTCAGGTCAAGGAGGTCAGAGAGCTTCAGGTCAAGGAGGTCAGAGAGC3 ), and this sequence was inserted into the site between the Bgl II and Hind III restriction sites of the pGL3-basic vector. The pGL3-basic-GH luciferase reporter was also used, which contains a human growth hormone promoter. Cells were transiently co-transfected with one of the three luciferase reporter plasmids (0.5 ␮g; expressing firefly luciferase) containing the VDREs and an internal control reporter plasmid (0.005 ␮g; pRL-TK) per well in a 24-well plate. The pRL-TK vector expressing the Renilla luciferase gene was used to control for transfection efficiency. The cells were transiently co-transfected three times. Dual luciferase assays were performed with the dual luciferase reporter assay system (Promega Corp., Madison, WI). All reagents were prepared as described by the manufacturer. The 5× passive lysis buffer was supplied by the manufacturer and used for cell lysis. Briefly, after co-transfection with pRL-TK, one of the three luciferase reporter plasmids and the pCMV-Flag-IGFBP6 with or without pCMV-myc-VDR were combined. After transfection for 24 h, the cells were grown exponentially in the appropriate medium with or without 1,25(OH)2 D3 (10−8 M) for 24 h. The cells were then removed directly from culture and transferred to 1× passive lysis buffer (100 ␮L). After lysing for 10–15 s, an aliquot (10 ␮L) was used for luminescence measurements with a Turner Designs model TD-20/20 luminometer. The following steps were used for luminescence measurements: the firefly luciferase reagent (100 ␮L) was added to the test sample; a 10 s equilibration time occurred, and luminescence was measured with a 10 s integration time; the addition of Renilla luciferase reagent (100 ␮L) and firefly quenching; and a 10 s equilibration time and measurement of luminescence with a 10 s integration time. The data are represented as the ratio of firefly to Renilla luciferase activities (Fluc/Rluc). The results from at least three separate experiments were analyzed with Student’s t test.

2.10. Alkaline phosphatase assay Cells were seeded in 24-well plates at a density of 104 cells per well and transfected with plasmids expressing HA-IGFBP-6, Myc-VDR, vector control or siRNA against the VDR. After transfection for 24 h, the cells were treated with 1,25-(OH)2 D3 (10−8 M) or vehicle control for 24 h. Cells were washed with PBS and scraped and suspended with cold TXM-buffer [10 mM Tris–HCl (pH 7.4), 1 mM MgCl2 , 20 ␮M ZnCl2 and 0.1% Triton X-100]. After sonication, the cellular debris was removed by centrifugation, and the lysates were assayed for ALP activity with p-nitrophenylphosphate as a substrate (Sigma, Deisenhofen, Germany) (Schedlich et al., 2007).

2.11. Statistical analyses The data were expressed as the mean ± SD form at least three independent experiments. Student’s t test was used for statistical comparisons. Statistical significance was set at P < 0.01.

3.1. Verification and characterization of the interaction between IGFBP-6 and the VDR

3.2. Defining the VDR domains involved in its interaction with IGFBP-6 To define the VDR domains that were involved in these interactions with IGFBP-6, GST fusions proteins were produced and purified from bacteria. These fusion proteins included the following: (1) VDR1–91 : an N-terminal region containing an activation function-1 (AF-1) domain and a DNA binding domain (DBD) and (2) VDR115–427 : a C-terminal region with a ligand binding domain (LBD) and a ligand-dependent AF-2 domain (Fig. 2A). In these GST pulldown assays, IGFBP-3, which interacts with the VDR, was used as a positive control (Schedlich et al., 2007). Immobilized GST-VDR1–91 , GST-VDR115–427 , and GST-VDR were incubated in the same volume of 293T cell lysates, and they were transiently transfected with IGFBP-6 or IGFBP-3 at 4 ◦ C for 4 h. The bound IGFBP-6 or IGFBP-3 was determined by IB. Both IGFBP-6 and IGFBP3 were co-precipitated with GST-VDR115–427 and GST-VDR but

J. Cui et al. / Molecular and Cellular Endocrinology 338 (2011) 84–92

87

Fig. 1. Characterization of the interaction between IGFBP-6 and the VDR. (A) GST pull-down assays were performed with 293T cell lysates that were transiently transfected with Flag-IGFBP-6, Flag-IGFBP-3, Flag-RXR or HA-IGFBP-1. The input lanes represent 10% of the total cell extracts. (B) The same 293T cell lysate volumes that were transfected with Flag-IGFBP-6 were incubated with decreasing amounts of GST-VDR (100, 50, 25, and 10 ␮g) for 4 h at 4 ◦ C. The bound IGFBP-6 and an equal volume of the cell lysates (input) were detected by IB with an anti-Flag antibody. (C) Cells (293T) were transfected with Flag-IGFBP-6 with either a Myc vector control or Myc-VDR. Immunoprecipitations (IPs) were performed with an anti-Myc antibody, and immune complexes were analyzed by IB using anti-IGFBP-6. The amounts of transfected IGFBP-6 and the VDR in the total cell lysates (10% input) are shown. (D) Cells (293T) were transfected with Myc-VDR and either Flag-IGFBP-6 or a Flag-vector and stimulated with 1,25-(OH)2 D3 (10−8 M) for 24 h after transfection. IP was performed with an anti-Flag antibody, and immune complexes were analyzed by IB using anti-VDR. (E) Semi-endogenous IPs were conducted with HeLa cell lysates that were transiently transfected with Flag-IGFBP-6 or a Flag-vector. Cell lysates were incubated with an anti-VDR antibody, and immune complexes were analyzed by IB using anti-IGFBP-6. The endogenous VDR and transfected IGFBP-6 in the cell lysates (input) were detected by IB with anti-VDR and anti-IGFBP-6 antibodies, respectively.

not with GST-VDR1–91 (Fig. 2B). These results demonstrate that the VDR interacts with IGFBP-6 through its C-terminal region. To confirm these results, we constructed three plasmids expressing VDR1–91 , VDR115–427 and a full length VDR with a Flag-tag. Coimmunoprecipitation experiments were performed, and Flag-VDR was used as the positive control. IGFBP-6 was immunoprecipitated by the Flag antibody from the cell lysates transfected with HA-IGFBP-6 and Flag-VDR115–427 or Flag-VDR (full-length) but not with HA-IGFBP-6 and Flag-VDR1–91 or the Flag-vector (Fig. 2C). These data correspond to the results from the GST pulldown experiments. It has been reported that IGFBP-3 and IGFBP-5 can both interact with the VDR and modulate its transcription (Schedlich et al., 2007). We compared the sequences of six members and found that a “LXXLL” motif (L is leucine and X is any amino acid) that may mediate transcriptional co-activator binding to the nuclear receptors in IGFBP-3, IGFBP-5 and IGFBP-6 (Heery et al., 1997). Therefore,  we changed the “L94 R95 A96 L97 L98 in IGFBP-6 into “AAAAA” and named it “IGFBP-6-M”. IGFBP-3 was used as the positive control. Similar to IGFBP-3, both IGFBP-6 and IGFBP-6-M interacted with the VDR (Fig. 2D). This result demonstrated that the “LXXLL” motif in IGFBP-6 might be not essential for this interaction with the VDR.

3.3. IGF-II prevented the interaction between IGFBP-6 and the VDR, and IGFBP-6 may influence VDR:RXR heterodimerization IGFBP-6 binds to IGF-II with high affinity (Bach, 2005); therefore, we investigated whether IGF-II interferes with the interaction between IGFBP-6 and the VDR. IGF-II was able to compete with the VDR for binding to IGFBP-6, which reduced the VDR:IGFBP-6 interaction in a dose-dependent manner (Fig. 3A). As the concentrations of IGF-II increased from 0 ng/ml to 100 ng/ml, an obvious dose-dependent reduction was seen in the amount of bound IGFBP6. We next determined whether IGFBP-6 inhibited the formation of VDR–RXR heterodimers. In the absence of IGFBP-6, the VDR was co-precipitated by GST–RXR, which demonstrated that heterodimerization occurred (Fig. 3B). As the concentrations of IGFBP-6 increased from 0 nM to 4 nM, the amounts of the bound RXR was reduced. Therefore, IGFBP-6 may influence the heterodimerization between the VDR and the RXR. 3.4. Intracellular co-localization of IGFBP-6 and the VDR To investigate the subcellular distribution of IGFBP-6 and the VDR, we singly and transiently expressed Flag-IGFBP-6 or Myc-VDR

88

J. Cui et al. / Molecular and Cellular Endocrinology 338 (2011) 84–92

Fig. 2. Mapping of the VDR and IGFBP-6 binding domains. (A) Structure of human VDR. The schematic represents the different parts of the VDR molecule: the ligandindependent activation function-1 domain (AF-1) and DNA binding domain (DBD) are in the N-terminal region, and the ligand binding domain (LBD) and liganddependent activation function domain (AF-2) are in the C-terminal region. The VDR variant used in this study (VDR1–91 ) contains AF–1 and DBD, and variant VDR115–427 contains LBD and AF-2. (B) Cells (293T) were transfected with constructs expressing Flag-IGFBP-6 or Flag-IGFBP-3, which was the positive control. Bound IGFBP-6 or IGFBP-3 was analyzed by IB using an anti-Flag antibody. The GST and VDR protein variants fused with GST were determined by IB with an anti-GST antibody. (C) Cell lysates (293T) which were transfected with HA-IGFBP-6 and Flag-VDR1–91 , Flag-VDR115–427 , Flag-VDR or a Flag-vector were subjected to IP using an anti-Flag antibody. IGFBP-6 in the protein complexes was detected by IB with an anti-IGFBP-6 antibody. The amounts of transfected IGFBP-6 and VDR variants in the cell lysates (10% input) were also determined by IB using their relative antibodies. (D) Alignment of the LXXLL motif sequences in the six IGFBP family members. IGFBP-6-M contains a mutation in the LXXLL motif in which the motif was replaced by “AAAAA”. Cell lysates were immunoprecipitated with an anti-Flag antibody. The VDR in the protein complexes was analyzed by IB with an anti-VDR antibody. The amounts of transfected IGFBPs and VDR in the cell lysates (10% input) were also determined by IB using their relative antibodies.

in 293T and HeLa cells for immunofluorescence staining assays. Flag-IGFBP-6 and Myc-VDR were mainly localized in the nucleus of 293T and HeLa cells (Fig. 4A). To confirm whether the in vitro biochemical interaction between IGFBP-6 and the VDR was present in living cells, we compared their co-localizations in 293T and HeLa cells. Cells expressing Flag-IGFBP-6 and Myc-VDR were stimulated with 1,25-(OH)2 D3 (10−8 M) for 12 h before fixation, and they were detected using specific antibodies and fluorescent microscopy. Flag-IGFBP-6 (stained in green fluorescence) and Myc-VDR (stained in red fluorescence) were predominantly found in the nucleus (Fig. 4B). Also, both FlagIGFBP-6 and Myc-VDR co-localized to the nucleus in 293T and HeLa cells (see merged panel in Fig. 4B). The antibodies were specific because no signal was detected when they were replaced with mouse or rabbit IgG, respectively (data not shown). To confirm these observations, nuclear cytoplasmic fractionation was applied to 293T and HeLa cells that were transiently transfected with Flag-IGFBP-6 and Myc-VDR. As expected, the nuclear protein Histone 1 was retrieved in the nuclear fraction, and GAPDH and the lysosomal protein LAMP-1 were found exclusively in the cytoplasmic fraction, which suggests that a clean separation between the nuclear and cytoplasmic proteins was achieved (Fig. 4C). A small proportion of IGFBP-6 and the VDR was retained in the cytoplasm, but both localized predominantly to the cell nucleus. The results correspond to the results using immunofluorescence.

Fig. 3. IGF-II inhibited the VDR:IGFBP-6 interaction, and IGFBP-6 influenced the VDR:RXR heterodimerization. (A) Equal volumes of 293T cell lysates that expressed HA-IGFBP-6 were incubated with immobilized GST-VDR. IGF-II was added to the GST-VDR at the same time at concentrations of 0, 10 or 100 ng/ml. Bound IGFBP-6 was analyzed by IB with an anti-IGFBP-6 antibody. The amount of GST-VDR was also determined by IB using an anti-GST antibody. (B) Immobilized GST-VDR (0.4 nM) was preincubated with recombinant human IGFBP-6 (expressed in Pichia Pastoris and purified (Chen et al., 2007)) at 2 nM or 4 nM or with BSA (4 nM) for 1 h at 4 ◦ C. The same amount of 293T cell lysates that were transiently transfected with the plasmids expressing the RXR was added into each sample, incubated and swirled for 3 h at 4 ◦ C. The amount of the RXR that co-immunoprecipitated with GST-VDR was determined by IB using an anti-RXR antibody. The amounts of purified IGFBP-6 and GST-VDR were also determined by IB using anti-IGFBP-6 and anti-GST antibodies, respectively.

Taken together, these data suggest that IGFBP-6 colocalizes with the VDR in the nucleus where they interact with each other. 3.5. IGFBP-6 negatively regulates the ligand-induced transcriptional activity of the Vitamin D response element (VDRE) The VDR binds the vitamin D response elements (VDREs) with high affinity in the promoters of target genes. Canonical VDREs are a direct repeat of the 5 -AGG/TTCA-3 motif or a minor variation of this motif separated by three nucleotides (Nagpal et al., 2005). We investigated the ability of IGFBP-6 to modulate 1,25(OH)2 D3 induced transcription of the VDRE promoters. Cells (293T and HeLa) were transiently co-transfected with one kind of luciferase reporter construct (expressing the firefly luciferase gene), a pRL-TK vector (expressing the Renilla luciferase gene) and pCMV-Flag-IGFBP6 with or without pCMV-Myc-VDR. The cells were stimulated with 1,25(OH)2 D3 (10−8 M) or vehicle for 24 h. Cell lysates were harvested and assayed for luciferase activity. Using pGL3-basicVDREs and pRL-TK for the luciferase assays, IGFBP-6 significantly decreased the VDR mediated addition of reporter activity in the presence of 1,25(OH)2 D3 (Fig. 5A, 293T; Fig. 5B, HeLa). The ratios of firefly to Renilla luciferase activity (Fluc/Rluc) were 2.3 and 0.82 when the VDR was transfected singly, but the ratios were reduced to 1.7 and 0.61 when IGFBP-6 was co-transfected with the VDR in the presence of 1,25(OH)2 D3 (P < 0.01) in 293T and HeLa cells, respectively. Although the ratios were reduced from 0.69 to 0.50 and from 0.45 to 0.38 in the 293T or HeLa cells, respectively, no statistical differences were noted in the absence of 1,25(OH)2 D3 . To confirm that IGFBP-6 decreased the VDR’s transduction activity, we repeated the dual-luciferase assay using pGL3-basic-Oste (Fig. 5C and D) or pGL3-basic-GH (Fig. 5E and F) and pRL-TK. We observed similar decreases in luciferase activity when cells were co-transfected with IGFBP-6 and the VDR compared with the cells transfected with VDR only in the presence of 1,25(OH)2 D3 . Using pGL3-basic-Oste,

J. Cui et al. / Molecular and Cellular Endocrinology 338 (2011) 84–92

89

Fig. 4. Cellular co-localization of IGFBP-6 and the VDR. (A) Determination of IGFBP-6 and the VDR localization in 293T and HeLa cells. Cells (293T and HeLa) were transiently transfected with plasmids expressing Flag-IGFBP-6 or Myc-VDR. After transfection for 24 h, cells were stimulated with 1,25-(OH)2 D3 (10−8 M) for 24 h. Control experiments used the 293T and HeLa cells transfected with a Flag-vector or a Myc-vector (data not shown). (B) Cells (293T and HeLa) were transiently co-transfected with plasmids expressing Flag-IGFBP-6 and Myc-VDR. After transfection for 24 h, cells were stimulated with 1,25-(OH)2 D3 (10−8 M) for 24 h. The cells were fixed, permeabilized and incubated with anti-Flag and anti-VDR antibodies overnight at 4 ◦ C, and then with anti-mouse FITC-conjugated and anti-rabbit TRITC-conjugated IgG. In the immunofluorescence experiments, the nuclei were counterstained for DNA with 4’,6-diamidine2-phenylindole (DAPI). Control experiments used the 293T and HeLa cells transfected with a Flag-vector and a Myc-vector (data not shown). (C) Cells (293T and HeLa) co-transfected with constructs expressing IGFBP-6 and the VDR were treated with 1,25-(OH)2 D3 (10−8 M) for 24 h before extracting the nuclear and cytoplasmic fractions. The nuclear and cytoplasmic fractions were generated as described in Section 2. The subcellular distributions of IGFBP-6 and the VDR were determined by IB with GAPDH and lysosomal protein LAMP-1 as cytoplasm [C] markers and Histone 1 as a nucleus [N] marker.

the ratios were reduced from 14.9 to 12.8 in 293T cells (P < 0.01) (Fig. 5C) and from 9.3 to 7.6 in HeLa cells (P < 0.01) (Fig. 5D). Similarly, when using pGL3-basic-GH, the ratios were reduced from 5.3 to 3.4 (P < 0.01) in 293T cells (Fig. 5E) and from 0.22 to 0.16 (P < 0.01) in HeLa cells (Fig. 5F). In the absence of 1,25(OH)2 D3 , no statistical differences were observed between the cells co-transfected with IGFBP-6 and the VDR and cells transfected with VDR only. Together, our results suggest that IGFBP-6 reduces the transcriptional activity of the VDR in the presence of 1,25(OH)2 D3 .

3.6. Effect of IGFBP-6 on the induction of bone differentiation markers by 1,25(OH)2 D3 To further investigate the role of IGFBP-6 in modulating bone cell differentiation, we examined its effect on a bone differentiation maker, ALP, which could be induced by the VDR and 1,25(OH)2 D3 . The MG-63 cell line is considered to represent a population of undifferentiated human osteoblast-like cells and is a good model for examining the initial phases of human bone cell differentiation

90

J. Cui et al. / Molecular and Cellular Endocrinology 338 (2011) 84–92

Fig. 5. IGFBP-6 inhibited VDR transcriptional activity in the presence of 1,25(OH)2 D3 . 293T (left) and HeLa (right) cells were transiently co-transfected with plasmids (pCMV-Flag-IGFBP-6 with pCMV-Myc-VDR or the pCMV vector control), one kind of luciferase reporter plasmid (expressing the firefly luciferase gene) and the pRL-TK vector (expressing the Renilla luciferase gene). The pRL-TK was used at a concentration of 0.005 ␮g per well, and the other components were used at concentrations of 0.5 ␮g per well in 24-well plates. After a 24-h transfection, the cells were treated with 1,25-(OH)2 D3 (10−8 M) (black bars) or vehicle control (white bars) for 24 h, and cell lysates were assayed for dual-luciferase activity. (A and B) The luciferase reporter plasmid pGL3-basic-VDRE was used. (C and D) The luciferase reporter plasmid pGL3-basic-Oste was used. (E and F) The luciferase reporter plasmid pGL3-basic-GH was used. Each experimental condition was performed in triplicate and applied in three independent experiments. Values are expressed as the mean ± SD. * P < 0.01.

(Ryhanen et al., 2003; Schedlich et al., 2007). Therefore, we investigated the role of IGFBP-6 in modulating ALP activity in MG-63 cells and in another osteoblast cell line, U2-OS. MG-63 cells expressing human IGFBP-6 or vector control sequences were stimulated with 1,25(OH)2 D3 (10−8 M) or vehicle for 24 h, and cell lysates were assayed for ALP activity (Fig. 6A). ALP activity decreased from 0.63 to 0.58 when transfected with IGFBP-6 after 20 min in the absence of 1,25(OH)2 D3 . However, in the presence of 1,25(OH)2 D3 , IGFBP-6 decreased ALP activity from 0.88 to 0.68 after 20 min (P < 0.01). These results suggest that IGFBP6 down-regulates ALP activity in the presence of 1,25(OH)2 D3 . Also, we used siRNA against the VDR (siRNA-VDR) to determine whether the decreased ALP activity was associated with an interaction between IGFBP-6 and the VDR. Through IB, we successfully showed a silenced, endogenous VDR in MG-63 cells (Fig. 6B). MG63 cells were transfected with IGFBP-6, siRNA-VDR or a vector control (N.C. siRNA), and they were examined for ALP activity. A significant decrease in ALP activity was observed when cells were treated with IGFBP-6 compared with the vector control in the presence of 1,25(OH)2 D3 (P < 0.01; Fig. 6C). When the cells were transfected with siRNA-VDR, no obvious differences in ALP activities were observed between cells treated with IGFBP-6 or a vector control in the presence of 1,25(OH)2 D3 . Similarly, in the U2-OS cell line, IGFBP-6 inhibited ALP activity (P < 0.01), and IGFBP-6 had little effect when the endogenous VDR was knocked

down in the presence of 1,25(OH)2 D3 (Fig. 6D). Taken together, these data suggest that the influence of IGFBP-6 on ALP activity is due to its interaction with the VDR and an inhibition of VDR function. 4. Discussion The discovery of IGF-independent mechanisms of cell growth, differentiation and apoptosis by IGFBPs provides indirect evidence for the presence of IGFBP binding proteins (Sueoka et al., 2000; Yan et al., 2001). Previous work has shown that IGFBP-3 binds to RAR and inhibits the transcription of the retinoic acid response element (RARE) by atRA (Schedlich et al., 2004). IGFBP-5 interacts directly with the VDR and modulates the 1,25(OH)2 D3 response (Schedlich et al., 2007). These results suggest that the nuclear receptor may be an important part of the IGF-independent functions of the IGFBPs. In this study, we first demonstrated an interaction between IGFBP-6 and the VDR (Fig. 1), and we localized the IGFBP-6 binding sites to the C-terminal region of VDR (Fig. 2). IGFBP-6 and the VDR co-localized mainly to the nucleus (Fig. 4), and IGFBP-6 inhibited the formation of the VDR:RXR heterodimer and blocked the cellular response to 1,25(OH)2 D3 . Furthermore, we presented evidence that IGFBP-6 inhibits VDRE-mediated gene expression and represses ALP activity in osteoblast-like cells mediated by 1,25(OH)2 D3 (Figs. 5 and 6). Thus, IGFBP-6 may influence the VDR

J. Cui et al. / Molecular and Cellular Endocrinology 338 (2011) 84–92

91

Fig. 6. IGFBP-6 reduced the ALP activity increase in response to 1,25(OH)2 D3. (A) Cell lysates were prepared from MG-63 cells transiently transfected with constructs expressing IGFBP-6 [treated with 1,25(OH)2 D3 (10−8 M) (open triangle) or vehicle control (filled triangle) for 24 h after transfection] or vector control [treated with 1,25(OH)2 D3 (10−8 M) (open squares) or vehicle control (filled squares) for 24 h after transfection]. Lysates were assayed for ALP activity at 5, 9, 13, and 20 min. (B) MG-63 cells transfected with siRNA-VDR or N.C. siRNA. After transfection for 48 h, the cells were subjected to IB with an antibody against the VDR. The same cell lysates were used to detect the expression of beta-actin. MG-63 (C) and U2-OS (D) cells were transiently transfected with constructs expressing IGFBP-6 or vector control (0.5 ␮g per well) and siRNA-VDR or N.C. siRNA (1 ␮g per well) in a 24-well plate. After transfection for 24 h, cells were stimulated with 1,25(OH)2 D3 or vehicle control for 24 h. The cells were washed, scraped and suspended in TXM buffer, and the lysates were assayed for ALP activity. Each sample was expressed as the rate of ALP activity/mg protein. Values are expressed as the mean ± SD. * P < 0.01.

signaling pathway and modulate the differentiation of osteoblasts in response to 1,25(OH)2 D3 . The nuclear hormone receptor VDR is composed of a number of functional domains. The N-terminal region contains a ligand-independent activation function-1 (AF-1) domain and a DNA binding domain (DBD). The C-terminal region of the receptor contains a multifunctional domain harboring the ligand binding domain (LBD) and a ligand-dependent activation function domain (AF-2). LBD is responsible for heterodimerization with the RXR, which is an obligate partner of the VDR in mediating 1,25(OH)2 D3 ’s action (Shao and Lazar, 1999; Pinette et al., 2003). The VDR ligand 1,25(OH)2 D3 binds to the LBD of the VDR, which causes conformational changes and enhances the VDR–RXR heterodimer formation (Cheskis and Freedman, 1994). We showed that IGFBP-6 influences the interaction between the VDR and the RXR in vitro (Fig. 3). These findings are consistent with the observation that IGFBP-6 interacts with the C-terminal region of the VDR (Fig. 2). This ability of IGFBP-6 to compete with the RXR for heterodimerization with the VDR suggests an important role for IGFBP-6 in regulating nuclear hormone activity. In this way, IGFBP-6 could attenuate the ligand-induced activity of the VDR Coactivators that bind to nuclear receptors contain one or more of the consensus sequence LXXLL, which forms an amphipathic alpha-helix (Heery et al., 1997). This helix fits into the hydrophobic cleft of the ligand receptor (Shiau et al., 1998). Also, receptor-specific binding of coactivators containing the LXXLL motif is governed by amino acid residues flanking the binding site (Vanhooke et al., 2004). Nuclear receptor corepressors, such as Nco and SMRT, contain a ␸xx␸␸ motif (␸ is leucine or isoleucine and x is any amino acid), which is similar to a coactivator motif (Hu and Lazar, 1999; Nagy et al., 1999). This motif is also predicted to form an alpha helix, which is one turn longer than the one formed by the LXXLL motif (Teichert et al., 2009). The precise sequences of the cores and flanking sequences are critical determinants of receptor

preference (Hu and Lazar, 1999). However, in this study, the LXXLL motif in IGFBP-6 was not essential for its interaction with the VDR (Fig. 2D), suggesting that there may be another undetermined motif in nuclear receptor corepressors. It has been reported that the stably transfected, full length IGFBP-6 can be detected in both culture medium and cell lysates. Also, the full length IGFBP-6 was mostly concentrated in the nucleus (Iosef et al., 2008). In a previous study, we purified the recombinant human IGFBP-6 (rhIGFBP-6) in Pichia Pastoris (Chen et al., 2007), and we added a negative control without rhIGFBP-6 to investigate the basal expression of IGFBP-6 in MDA-MB-231 cells (Fig. S1A). We found that IGFBP-6 entered into the cells and concentrated mostly in the nucleus. The negative control showed that endogenous IGFBP-6 was too low to be detected. Also, we added FITC-rhIGFBP-6 into the culture medium of the MDA-MB-231 cells, and the cells were detected by confocal laser scan microscopy. We found that the green fluorescence was distributed in the whole cell and localized mainly in the nucleus (Fig. S1B). These two results suggest that rhIGFBP-6 could be translocated into the cell and the nucleus from the culture medium. We examined the impact of IGFBP-6 on the osteocalcin and growth hormone promoter transcription and expression activities that are mediated by the VDR and 1,25(OH)2 D3 . Both activities were induced by 1,25(OH)2 D3 , but they were significantly reduced when the cells were co-transfected with both IGFBP-6 and the VDR compared with the VDR-only transfected cells (Fig. 5C and D). In the presence of 1,25(OH)2 D3 , these results were strengthened by the repression of the transcriptional activity of the growth hormone promoter (Fig. 5E and F) and the promoter containing only VDREs when IGFBP-6 was co-transfected with the VDR compared to transfection with the VDR only (Fig. 5A and B). These results clearly demonstrate that IGFBP-6 was able to restrict the transcriptional activation activity of the VDR and exert an inhibitory effect on the cellular responsiveness to 1,25(OH)2 D3 . Interestingly, IGFBP-3

92

J. Cui et al. / Molecular and Cellular Endocrinology 338 (2011) 84–92

and IGFBP-5 have also been shown to block 1,25(OH)2 D3 induced transcriptional activity. We examined the effects of IGFBP-6 on the activity of a bone differentiation marker, ALP. A significant decrease in ALP activity was observed when the plasmids expressing IGFBP-6 were transiently transfected into both MG-63 and U2-OS cells in the presence of 1,25(OH)2 D3 (Fig. 6). When the endogenous VDR was knocked down by siRNA, no obvious changes in ALP activities were noted between cells transfected with an IGFBP-6 expression vector or a vector control. These results suggest that IGFBP-6 influences ALP activity due to its interaction with the VDR. It has been reported that the mature peptide forms of IGFBPs show similar functions to the full length forms (Bhattacharyya et al., 2006). Although the molecular mechanisms remain unclear, it was reasonable for us to use the mature peptide of IGFBP-6 in this study to demonstrate the interaction between the VDR and IGFBP-6. Perhaps after endogenous IGFBP-6 is secreted as a mature peptide into the medium and is not bound to IGF-I or IGF-II, it may have the chance to enter the cell again (Bhattacharyya et al., 2006). Together, this study represents a new paradigm in our understanding of the actions of IGFBP-6. The discovery of the VDR as an IGFBP-6 interacting protein represents further complexity into its modulation of cell growth and metabolism. In addition to the modulation of IGF activity, IGFBP-6 may also affect cell differentiation through several VDR-dependent mechanisms. The IGFBP-6:VDR interaction represents an interface of two previously unrelated signaling pathways and opens up a new direction in studying the cross-talk between growth factors and nuclear receptors. Acknowledgments We thank Prof. X. Peng (IIT Research Institute, Chicago), Prof. Werner Zwerschke (Cell Metabolism and Differentiation Research Group, Institute for Biomedical Aging Research of the Austrian Academy of Sciences, Innsbruck, Austria) and Prof. Norman Eberhardt (Department of Medicine/Division of Endocrinology, Minnesota) for giving us pcDNA3.1-hVDR, pI-IGFBP-3 and pGL3GH, respectively. We also thank Xu Liu (Central Laboratory, Chinese Academy of Medical Sciences & Peking Union Medical College, Beijing, China) for technical assistance in IF and DLR. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.mce.2011.03.011. References Bach, L.A., 2005. IGFBP-6 five years on; not so ‘forgotten’? Growth Horm. IGF Res. 15, 185–192. Bhattacharyya, N., Pechhold, K., Shahjee, H., Zappala, G., Elbi, C., Raaka, B., Wiench, M., Hong, J., Rechler, M.M., 2006. Nonsecreted insulin-like growth factor binding protein-3 (IGFBP-3) can induce apoptosis in human prostate cancer cells by IGF-independent mechanisms without being concentrated in the nucleus. J. Biol. Chem. 281, 24588–24601. Chen, J.W., Murphy, T.L., Willingham, M.C., Pastan, I., August, J.T., 1985. Identification of two lysosomal membrane glycoproteins. J. Cell Biol. 101, 85–95. Chen, Z., Chen, H., Wang, X., Ma, X., Huang, B., 2007. Expression, purification, and characterization of secreted recombinant human insulin-like growth factor-binding protein-6 in methylotrophic yeast Pichia pastoris. Protein Expr. Purif. 52, 239–248. Cheskis, B., Freedman, L.P., 1994. Ligand modulates the conversion of DNA-bound vitamin D3 receptor (VDR) homodimers into VDR-retinoid X receptor heterodimers. Mol. Cell Biol. 14, 3329–3338. Dailly, Y.P., Zhou, Y., Linkhart, T.A., Baylink, D.J., Strong, D.D., 2001. Structure and characterization of the human insulin-like growth factor binding protein (IGFBP)-6 promoter: identification of a functional retinoid response element. Biochim. Biophys. Acta 1518, 145–151.

Firth, S.M., Baxter, R.C., 2002. Cellular actions of the insulin-like growth factor binding proteins. Endocr. Rev. 23, 824–854. Gallicchio, M.A., Kneen, M., Hall, C., Scott, A.M., Bach, L.A., 2001. Overexpression of insulin-like growth factor binding protein-6 inhibits rhabdomyosarcoma growth in vivo. Int. J. Cancer 94, 645–651. Gardiner, E.M., Baldock, P.A., Thomas, G.P., Sims, N.A., Henderson, N.K., Hollis, B., White, C.P., Sunn, K.L., Morrison, N.A., Walsh, W.R., Eisman, J.A., 2000. Increased formation and decreased resorption of bone in mice with elevated vitamin D receptor in mature cells of the osteoblastic lineage. FASEB J. 14, 1908–1916. Gurlek, A., Kumar, R., 2001. Regulation of osteoblast growth by interactions between transforming growth factor-beta and 1alpha, 25-dihydroxyvitamin D3. Crit. Rev. Eukaryot. Gene Expr. 11, 299–317. Gurlek, A., Pittelkow, M.R., Kumar, R., 2002. Modulation of growth factor/cytokine synthesis and signaling by 1alpha, 25-dihydroxyvitamin D(3): implications in cell growth and differentiation. Endocr. Rev. 23, 763–786. Haussler, M.R., Whitfield, G.K., Haussler, C.A., Hsieh, J.C., Thompson, P.D., Selznick, S.H., Dominguez, C.E., Jurutka, P.W., 1998. The nuclear vitamin D receptor: biological and molecular regulatory properties revealed. J. Bone Miner. Res. 13, 325–349. Heery, D.M., Kalkhoven, E., Hoare, S., Parker, M.G., 1997. A signature motif in transcriptional co-activators mediates binding to nuclear receptors. Nature 387, 733–736. Hu, X., Lazar, M.A., 1999. The CoRNR motif controls the recruitment of corepressors by nuclear hormone receptors. Nature 402, 93–96. Iosef, C., Gkourasas, T., Jia, C.Y., Li, S.S., Han, V.K., 2008. A functional nuclear localization signal in insulin-like growth factor binding protein-6 mediates its nuclear import. Endocrinology 149, 1214–1226. McDonnell, D.P., Mangelsdorf, D.J., Pike, J.W., Haussler, M.R., O’Malley, B.W., 1987. Molecular cloning of complementary DNA encoding the avian receptor for vitamin D. Science 235, 1214–1217. Mohan, S., Baylink, D.J., 2002. IGF-binding proteins are multifunctional and act via IGF-dependent and -independent mechanisms. J Endocrinol 175, 19–31. Nagpal, S., Na, S., Rathnachalam, R., 2005. Noncalcemic actions of vitamin D receptor ligands. Endocr. Rev. 26, 662–687. Nagy, L., Kao, H.Y., Love, J.D., Li, C., Banayo, E., Gooch, J.T., Krishna, V., Chatterjee, K., Evans, R.M., Schwabe, J.W., 1999. Mechanism of corepressor binding and release from nuclear hormone receptors. Genes Dev. 13, 3209–3216. Panda, D.K., Miao, D., Bolivar, I., Li, J., Huo, R., Hendy, G.N., Goltzman, D., 2004. Inactivation of the 25-hydroxyvitamin D 1alpha-hydroxylase and vitamin D receptor demonstrates independent and interdependent effects of calcium and vitamin D on skeletal and mineral homeostasis. J. Biol. Chem. 279, 16754–16766. Pinette, K.V., Yee, Y.K., Amegadzie, B.Y., Nagpal, S., 2003. Vitamin D receptor as a drug discovery target. Mini. Rev. Med. Chem. 3, 193–204. Ryhanen, S., Jaaskelainen, T., Mahonen, A., Maenpaa, P.H., 2003. Inhibition of MG-63 cell cycle progression by synthetic vitamin D3 analogs mediated by p27, Cdk2, cyclin E, and the retinoblastoma protein. Biochem. Pharmacol. 66, 495–504. Sanchez-Martinez, R., Zambrano, A., Castillo, A.I., Aranda, A., 2008. Vitamin D-dependent recruitment of corepressors to vitamin D/retinoid X receptor heterodimers. Mol. Cell Biol. 28, 3817–3829. Schedlich, L.J., Muthukaruppan, A., O’Han, M.K., Baxter, R.C., 2007. Insulin-like growth factor binding protein-5 interacts with the vitamin D receptor and modulates the vitamin D response in osteoblasts. Mol. Endocrinol. 21, 2378–2390. Schedlich, L.J., O’Han, M.K., Leong, G.M., Baxter, R.C., 2004. Insulin-like growth factor binding protein-3 prevents retinoid receptor heterodimerization: implications for retinoic acid-sensitivity in human breast cancer cells. Biochem. Biophys. Res. Commun. 314, 83–88. Shao, D., Lazar, M.A., 1999. Modulating nuclear receptor function: may the phos be with you. J. Clin. Invest. 103, 1617–1618. Shiau, A.K., Barstad, D., Loria, P.M., Cheng, L., Kushner, P.J., Agard, D.A., Greene, G.L., 1998. The structural basis of estrogen receptor/coactivator recognition and the antagonism of this interaction by tamoxifen. Cell 95, 927–937. Strohbach, C., Kleinman, S., Linkhart, T., Amaar, Y., Chen, S.T., Mohan, S., Strong, D., 2008. Potential involvement of the interaction between insulin-like growth factor binding protein (IGFBP)-6 and LIM mineralization protein (LMP)-1 in regulating osteoblast differentiation. J. Cell Biochem. 104, 1890–1905. Sueoka, N., Lee, H.Y., Wiehle, S., Cristiano, R.J., Fang, B., Ji, L., Roth, J.A., Hong, W.K., Cohen, P., Kurie, J.M., 2000. Insulin-like growth factor binding protein-6 activates programmed cell death in non-small cell lung cancer cells. Oncogene 19, 4432–4436. Teichert, A., Arnold, L.A., Otieno, S., Oda, Y., Augustinaite, I., Geistlinger, T.R., Kriwacki, R.W., Guy, R.K., Bikle, D.D., 2009. Quantification of the vitamin D receptor-coregulator interaction. Biochemistry 48, 1454–1461. van Driel, M., Pols, H.A., van Leeuwen, J.P., 2004. Osteoblast differentiation and control by vitamin D and vitamin D metabolites. Curr. Pharm. Des. 10, 2535–2555. Vanhooke, J.L., Benning, M.M., Bauer, C.B., Pike, J.W., DeLuca, H.F., 2004. Molecular structure of the rat vitamin D receptor ligand binding domain complexed with 2-carbon-substituted vitamin D3 hormone analogues and a LXXLL-containing coactivator peptide. Biochemistry 43, 4101–4110. Yan, T., Wergedal, J., Zhou, Y., Mohan, S., Baylink, D.J., Strong, D.D., 2001. Inhibition of human osteoblast marker gene expression by retinoids is mediated in part by insulin-like growth factor binding protein-6. Growth Horm. IGF Res. 11, 368–377.