Bone Density Ligand, Sclerostin, Directly ... - Wiley Online Library

JOURNAL OF BONE AND MINERAL RESEARCH Volume 21, Number 11, 2006 Published online on August 21, 2006; doi: 10.1359/JBMR.060810 © 2006 American Society for Bone and Mineral Research

Bone Density Ligand, Sclerostin, Directly Interacts With LRP5 but Not LRP5G171V to Modulate Wnt Activity Debra L Ellies,1,6 Beth Viviano,2 John McCarthy,1 Jean-Philippe Rey,1 Nobue Itasaki,3 Scott Saunders,2 and Robb Krumlauf1,4,5

ABSTRACT: We compared and contrasted the mechanism of action for the cysteine knot protein subfamily, Wise and Sost (Sclerostin). Our data suggest that functional interactions between Sost or Wise and LRP5/ LRP6 have the potential to regulate bone deposition by modulating the Wnt pathway. Introduction: The human disease sclerosteosis exhibits an increase in bone mass thought to be caused by hyperactive osteoblasts. Sclerostin, SOST, the gene affected in this disease, has been postulated to exert its activity by functioning as a BMP antagonist. However, recent evidence indicates that SOST is highly related to Wise, which can also modulate the Wnt pathway by binding to LRP5 and LRP6. Materials and Methods: For this study, we used cell culture to test the BMP and Wnt activity function of both Wise and Sost. In addition, we used Xenopus in vivo Wnt assays along with Xenopus in vitro Wnt assays to support our cell culture results. Epitope tagged cell supernatants containing either Sost or soluble mutant or wildtype LRP5/LRP6 were used for immunoprecipitation. Sost immunoprecipitation results were confirmed in vivo using cell culture. Finally, to support our in vitro data, we co-localized Sost, Wise, LRP5, and LRP6 in mouse long bone sections. Results: In this study, we report in vitro and in vivo evidence to show that Sost physically interacts with Lrp5 and Lrp6 and inhibits the canonical Wnt signaling pathway. Furthermore, using in vitro and in vivo assays, we showed that a variant of LRP5 (LRP5G171V) known to cause the human high bone mass (HBM) trait and a homologous change in LRP6 (LRP6G158V) abolished protein interactions with Sost. We used variants of Sost amino acids to further identify the contact points between Sost and LRP6. In Xenopus and mammalian cell culture assays, we showed that SOST is able to attenuate Wnt signaling and that this attenuation can be rescued by the addition of ␣-Sost antibodies or by the introduction of single amino acid substitution that alter its binding to LRP6. Sost differs from Wise in that it is unable to stimulate Wnt signaling. Using immunohistochemistry, we found that Sost and Wise are co-localized to osteoblasts, along with LRP5 and LRP6. Conclusions: Our data suggest that functional interactions between Sost or Wise and LRPs have the potential to regulate bone deposition by modulating Wnt signaling. J Bone Miner Res 2006;21:1738–1749. Published online on August 21, 2006; doi: 10.1359/JBMR.060810 Key words: sclerostin, Wise, long bone, LRP5, LRP6, cysteine knot

INTRODUCTION

O

FTEN REFERRED TO as the “silent disease,” osteoporosis is an asymptomatic skeletal disorder characterized by low bone mass that exposes sufferers to a greatly increased risk of bone fractures. The disease is caused by a disruption of the fine equilibrium between the activity of osteoblast cells responsible for bone deposition and osteoclast cells for bone resorption.(1,2) Bone deposition is regulated in part by the BMP pathway. However, several lines of recent evi-

The authors state that they have no conflicts of interest.

dence has lead to the idea that Wnt signaling plays a key role in bone deposition.(3–8) Positional cloning of the gene associated with dominant human high bone mass (HBM) trait revealed that the LRP5 co-receptor for Wnt signaling is mutated and serves as a “bone deposition thermostat” during normal development.(5) HBM is the result of a G171V point mutation in LRP5 and is associated with a five times increase in bone mass.(5,9) In addition, a loss of LRP5 function results in osteoporosis pseudoglioma (OPPG) syndrome that is characterized by a juvenile onset of decreased bone mass.(10) Thus, LRP5 and the modulation of the Wnt signaling pathway is an important component in the mechanisms that regulate bone deposition and homeostasis.

1 Stowers Institute for Medical Research, Kansas City, Missouri, USA; 2Department of Pediatrics, Washington University Medical School, St Louis, Missouri, USA; 3National Institute for Medical Research, The Ridgeway, Mill Hill, London, United Kingdom; 4Department of Anatomy and Cell Biology, Kansas University Medical School, Kansas City, Kansas, USA; 5Department of Oral Biology, School of Dentistry, University of Missouri at Kansas City, Kansas City, Missouri, USA; 6Present address: Osteogenex Inc., Kansas City, Missouri, USA.

1738

SOST INTERACTS WITH LRP5 TO MODULATE WNT ACTIVITY

1739

FIG. 1. Characterization of Wise/SOST family. (A) Alignment of cysteine knot containing proteins into subfamilies. Yellow shows the conserved cysteine residues, green shows other conserved AA blocks, and red shows extra cysteine resides. At the bottom is the consensus cysteine knot core arrangement. Sequences shown start at “gi” sequence denoted numbers. (B) Phylogenetic tree of cysteine knot containing proteins. Red dots depict significant branches within the tree. The numbers adjacent to the tree branches connecting the family members show the percentage of chance occurrences for each branch, according to Phylip phylogenetic software. Note that Wise/SOST exist in a branch between the CCN and DAN families. (C) BMP4 and BMP6 inhibition assays for Wise and Sost using ATDC-5 cells. Wise does not inhibit the action of BMP6 and weakly blocks BMP4 at high concentrations. Conversely Sost inhibits the action of BMP6 but not that of BMP4.

Our work and that of other groups has shown that there are an increasing number of molecules (CTGF, Wise, Dkk) capable of binding LRP5 (and LRP6) and modulating activity of the Wnt pathway.(4,11–13) In particular, we identified a novel secreted cysteine-knot protein, Wise, and showed that it functions as a context dependent activator and inhibitor of the Wnt pathway using in vivo assays.(4) Wise shares homology with a known human bone density regulator, SOST/Sclerostin. Deficiency of SOST in humans leads to a skeletal disorder characterized by progressive bone overgrowth considered to be caused by hyperactive osteoblasts.(1,14) Sost deletion in mice leads to massive high BMD, and gain of Sost function leads to low bone mass.(15,16) Because Wise and Sost seem to be members of a family of proteins (CCN, Dan, VWF, Norrin, Mucin, and Slits) some of which have the ability to bind BMP ligands, published studies have speculated that increase and decrease in BMD related to alterations of Sost are likely to be mediated through inhibition of the BMP pathway.(14,15,17) In this study, we provide compelling evidence using in vitro and in vivo assays that both Wise and Sost bind normal LRP5 and LRP6 but not to variants carrying the HMB mutations and that they are capable of inhibiting Wnt signaling. In combination with immunohistochemical analyses of protein distribution patterns in P0 bone cell types, our data suggest that Sost and Wise modulate bone deposition by affecting components of the Wnt pathway in osteoblast cells.

MATERIALS AND METHODS Isolation of mouse sost Sost was isolated from a mouse d11 cDNA library using touchdown PCR (1 cycle, denature 92°C 2⬘; 5 cycles, 92°C 1⬘, anneal 68°C 1⬘, extend 72°C 2⬘; 30 cycles, 92°C 1⬘, anneal 64°C 1⬘, extend 72°C 2⬘; 1 cycle, extension 72°C 10⬘); forward primer, 5⬘-CGTGCCTCATCTGCCTACTTGTGCA-3⬘; reverse primer, 5⬘-GAAGTCCTTGAGCTCCGACTGGTTGTG-3⬘. The full-length mouse Sost cDNA was created using successive PCR reactions based on GI: 13161022. Phylogenetic tree: Sost and Wise cysteine knot protein sequences were Blasted (NCBI), and all significant sequences were isolated. The cysteine knots from all sequences were manually aligned using software T-Coffee (Fig. 1A) and analyzed with Phylip bootstrap neighbor joining methods. Xenopus assays: Capped RNA synthesis used 5 ␮g linear DNA template (Sost, Noggin, ␤-catenin, Wnt8, Wise) suspended in transcription buffer (P118B; Promega), 10 mM dithiothraitol (DTT), 1 mM dNTPs, 0.5 mM m7G(5⬘)ppp (5⬘)G Cap analog (NEN 514045), Rnasin, and polymerase as previously described.(4) Amounts of mRNA used for microinjection in animal cap assays: 250 pg Noggin; 300, 600, and 900 pg Sost; and 5 pg Wnt8. RNA for ventral marginal zones: 300 pg Sost, 100 pg Wnt8, and 200 pg ␤-catenin.

1740

ATDC-5 cell culture assays for BMP activity Cells, at low passage, were maintained at subconfluency in DMEM/F12 media supplemented with 10% heat inactivated FCS, 100 units/ml penicillin, and 100 mg/ml streptomycin. In 96-well plates (Corning), inhibitors were diluted to a final volume of 100 ␮l/well. A constant amount of BMP (R&D Systems) in 50 ␮l was added to each well and incubated for 1 h at 37°C. ATDC-5 cells were counted and plated at 2 × 105 cells/ml and added at 50 ␮l/well. Ten microliters of heparin was added to the plate containing BMP4 at a final concentration of 2 ␮g/ml. Ten microliters of L-ascorbic acid was added to the plate containing BMP6 at a final concentration of 50 ␮g/ml. ATDC-5 cells were counted and plated in the wells at 2 × 105 cells/ml. Cells were incubated for 3 days at 37°C. Cell layers were washed twice with PBS and lysed in 200 ␮l of 0.15 M NaCl, 3 mM NaHCO3, and 0.1% Triton X-100 at pH 9.3. Care was taken to maintain the cell layers. Cell layers were incubated at 37°C for 30 minutes. Twenty microliters of each sample was incubated with 1 mg/ml of p-nitrophenol phosphate (Sigma) in 1 M diethanolamine (Sigma), with 0.5 mM MgCl2 at a pH of 9.8, and incubated at 22°C for 8 minutes. Reaction was stopped by the addition of 0.5 N NaOH. Optical density was measured at 405 nm.

Cell-based Wnt Luciferase reporter assay The 293 cell line was grown to ∼60% confluence and transfected with both TopFlash or FlopFlash (Upstate) and Renilla control vector using a 1:4 ratio of Fugene 6 (Roche). After a 3-h incubation, the transfection reagent was siphoned off and replaced with a mixture containing; 60% total volume Wnt3a conditioned media from day 10 stably transfected Mouse L-cells (ATCC) and 30% total volume of Sost conditioned media (∼3 mg total protein in conditioned media) from day 1 transiently transfected 293 cells. In addition to Wnt3a and Sost, a 2% FCS supplement was added to allow cells to grow to confluency. Blocking reagents were added in addition to this mixture: 100 ␮g of 1A12 LRP5/6 antibody (cat. AB2572; Abcam) mouse monoclonal to synthetic human “DTGTDRIEVTR” peptide. These residues are completely conserved between human and mouse and between LRP5 and LRP6. This antibody recognizes both the full-length (180–190 kDa) and proteolytic fragment (130–140 kDa) of LRP6 and LRP5 and undiluted full-length murine Sost Mab hybridoma supernatant at 30% total well volume (Mab 4G10 or 4A6). The cells were incubated overnight and assayed for reporter activity the next day using Promega’s Dual Luciferase Kit. All luciferase results were normalized to Renilla controls and were all done in triplicate, and significance was assessed using two-tailed Student t-tests. Statistically significant findings (p < 0.05) are represented by asterisks.

Generation of Sost and Wise antibodies Wise antibodies: Wise peptide polyclonal supernatants were generated by Zymed using rabbits that were immunized using the conserved “GTKYWSRRSSQEWR” region of the chick Wise protein. The polyclonal chick Wise

ELLIES ET AL. peptide antibody cross-reacts with the murine protein. Wise monoclonal and polyclonal antibodies against the fulllength chick protein sequence were generated by Genovac. Genovac used the cDNA encoding the full-length chick Wise to generate the rat monoclonal antibody and rabbit polyclonal antibody. Sost antibodies: Monoclonals were generated by Genovac by immunizing a rat using sequences from full-length murine Sost cDNA. Clones were screened for blocking ability using the above cell luciferase assay. Mabs were tested using immunohistochemistry with proper controls (with preimmune and without primary antibody), in vivo blocking ability, and Western blot. Two clones were found to detect Sost both in mouse bone sections and Western blot and to block Sost function in vivo: monoclonal (Mab) antibodies 4G10 and 4A6 (Fig. 6A).

Immunoprecipitation assay The 293 cells were transfected with LRP5, LRP6, Sost, or Wise, using a 1 ␮g DNA:4 ␮l ratio of Fugene 6 (Roche). pCS2 + LRP5, containing only the extracellular (soluble) portion of the human sequence between EcoRI-Xba1 sites, was epitope tagged with either IgG-FC or Myc. The murine pCS2 + Sost has been modified to have a Kozak sequence and at the 3⬘ end a Flag tag. Wise-Flag and the soluble LRP6-IgG-FC are described in Itasaki et al.(4) pCS2 + an IgG-FC insert was the vector control. The supernatants were collected on days 1–3. The protein supernatants were concentrated using Amicon Ultra columns (Millipore) at specific molecular weight cut-offs. A Bradford protein concentration was taken on the concentrated cell supernatants. Equal total amounts of protein supernatant were run on a Western blot to estimate Sost protein amounts. Using these data, ∼50 ␮g of total protein amounts were added to each immunoprecipitation (IP). IP was carried out overnight at 4°C using 1× TBS (containing 136 mM NaCl and 3 mM KCl) with 0.4% Triton-X and washed with 1× TBST (containing 0.2% Triton-X). Anti-Flag M2 affinity gel (Sigma) was used for the WiseFlag and SostFlag IPs. To quantify, equal amounts of Sost and LRP protein were verified using Western blots. Immunoprecipitation carried out as stated above. The resulting Western blot was digitally captured, and the inputs and IP proteins on the digitally captured image were quantified using ImageQuant Software.

In vivo protein binding The 293 cells were transfected with Fugene 6 (Roche) using the full-length transmembrane (tm) versions of LRP5 and LRP5G171V. After transfection, SostFlag conditioned media, quantified using a Western blot, was added to the LRP-transfected cells and cultured overnight. The following day, the cells were fixed with 4% paraformaldehyde (PFA) for 30 minutes at room temperature. A primary mouse anti-Flag M2 antibody (cat. F1804; Sigma) was used at 200 ng/␮l and a fluorescent conjugated secondary (1: 10,000 dilution from Zymed).

Histology and immunohistochemistry Femurs from P0 C57BL6 mice were harvested and fixed in 4% PFA for 2 days at 4°C and decalcified in 10% EDTA

SOST INTERACTS WITH LRP5 TO MODULATE WNT ACTIVITY at 4°C for 10 days or until no calcium was visible on X-ray film. Six-micrometer alternating serial cryosections were taken and stained for either Wise, Sost, or LRP5/6 using polyclonal antibodies generated against the mouse antigen for Sost, the chicken antigen for Wise and antipeptide Wise polyclonal, and a commercially available 1A12 (cat. AB2572; Abcam) for LRP5/6. The secondary antibodies used were all from Molecular Probes and were used at a concentration of 1:300; Wise/Sost used Rat Alex-488, and LRP5/6 used mouse Alex-568. All antibodies were run adjacent to samples run without primary antibody or with preimmune serum alone.

In situ hybridization Radioactive in situ: C57B6J embryos were harvested at day 16.5, and 14-␮m cryostat sections were collected on Probe-On Plus slides and stored at −80°C with desiccant before hybridization. Sections were equilibrated to room temperature, fixed in 4% paraformaldehyde in PBS for 20 minutes, rinsed twice in PBS with 0.1 M glycine and once in PBS alone, acetylated in 0.1 M triethanolamine (pH 8.0) and 0.25% acetic anhydride for 10 minutes, rinsed twice more in PBS, and dehydrated. Hybridization was carried out using [35S]labeled antisense probes. Slides were dipped in Kodak NTB-2 liquid emulsion diluted 1:1 with distilled water and exposed for 21–28 days at −80°C and stained with H&E. Whole mount in situ: The methods are as described in Ellies et al.(18) Riboprobes are as follows: cDNA for avian Wise contained 770 bp of coding and some of the 3⬘UTR, the plasmid was linearized with XhoI, and the antisense was transcribed using T7 RNA polymerase. The avian cDNA for Six3 was linearized using EcoRI and transcribed using T3 RNA polymerase. Avian cDNA for Sox3 was linearized using Nde1 and transcribed using T3 RNA polymerase. Avian cDNA for HNF3␤ was linearized using HindIII and transcribed with SP6 RNA polymerase. Mouse Wise and Sost cDNA contains 250 bp of 3⬘UTR; these were both linearized using EcoRI and transcribed with T7 RNA polymerase.

RESULTS Wise and Sost subfamily of cysteine knot proteins Wise is closely related to the human protein Sclerostin, SOST. WISE maps to human chromosome 7p21.1, 10.6 Mb downstream of the HOXA cluster, and SOST maps to chromosome 17q21.31, 5 Mb downstream of the HOXB complex. Both loci have a similar structure and, in combination with their linkage to HOX complexes, seem to have arisen by duplication and divergence from a common ancestral chromosome region. The cysteine knot domains in Wise and SOST are homologous to a region found in the DANfamily of BMP antagonists. This similarity has been used to postulate its role as a BMP inhibitor in the human disease sclerosteosis.(14,15) However, we have previously reported that Wise and CCN-family members CTGF and Cyr61 function as a Wnt pathway antagonists.(4,11,19) The sequence similarity between Wise and Sost opens the possi-

1741

bility that Sost may be functioning to regulate BMD through its ability to modulate Wnt signaling. To study these issues, we compared Wise and Sost and tested the ability of Sost to modulate/interact with the Wnt pathway. The Wise and Sost loci have the same two exon gene structure and encode predicted proteins that are 48% homologous. We used the cysteine knot motifs encoded by the second exon of Wise and Sost as a basis for comparison with other cysteine knot–containing proteins (Fig. 1A). There is significant similarity with the cysteine knot motifs from DAN (Cerberus, DAN, Gremlin, Caronte), CCN (NOV, CTGF, Cyr61), Slit, and Mucin family members (Figs. 1A and 1B). There is also homology to the cysteine motifs in individual genes, such as Von Willderbrand Factor (VFW) and Norrie Disease Protein (NDP). Using these homologies, we generated a phylogenetic tree that divides these proteins into distinct subfamilies based on the number and position of cysteine residues and other conserved amino acids in the motif (Fig. 1B). They all contain a consensus organization of eight core cysteine residues and one glycine residue (between cys-3 and cys-4). In contrast to Wise and Sost, the other subfamilies contain one to two additional cysteine (cys) residues between cys-4 and cys-5 of the core motif (Fig. 1A). The cysteine knot domain is the only recognizable motif found in the Wise/Sost and DAN subfamilies, whereas members of the other groups contain multiple motifs, such as insulin binding, Von Willderbrand, and TSP1 domains.(20–22) The Wise and Sost subfamily is most closely related to the DAN and CCN subgroups (Fig. 1B), and a common feature of Dan and CCN proteins is that they selectively bind and inhibit BMP signaling.(23–30,15) In cell culture models, using alkaline phosphatase (ALP) activity as a late marker for BMP-mediated osteoblast differentiation, we and others have found that Sost inhibits BMP6 but not BMP4 (Fig. 1C).(15,30) In our ATDC-5 cell assay, Wise did not inhibit BMP6 and only weakly influenced BMP4dependent activity at high concentration (Fig. 1C). However, Wise has been reported to have a preference for inhibiting BMP7, BMP2, and BMP4, and more weakly, BMP6, in other cells.(28,29) Therefore, despite the significant homology between Wise and Sost, they are not identical in their functional ability to inhibit BMP signaling.

Sost and Wise seem to be found in complimentary cell types It seems that even though Wise and Sost arose by duplication of an ancestral gene, some aspects of their function and gene expression have diverged in avian and murine embryos. In the avian, Wise is first detected in the prechordal plate at HH4 somites (s), similar to Dan family members (Figs. 2A and 2B), along with expression in the surface ectoderm apposing the neural tube (Figs. 2C and 2G). The expression of the prechordal plate diminishes by HH9s, whereas the surface ectoderm expression remains until the neural crest has finished emigrating and the apposing neural folds have fused. At HH stage 15, Wise is found in pituitary anlagen and in the epithelium ventral to thickened Six3 expressing neural epithelium (Figs. 2D and

1742

ELLIES ET AL.

FIG. 2. Sost and Wise are found in complimentary cell types. Avian wholemount expression of Wise. (A) 4s embryo showing Sox3 expression in the neural tube (red) and Wise (purple) in the surface ectoderm apposing the trunk neural tube (arrow) and in the prechordal plate (arrow). (B) Sagittal section of a 4s embryo showing Wise expression (arrow) in the prechordal plate (purple) and HNF3␤ in the ventral neural tube (VNT). (C) Wise expression in the surface ectoderm apposing the neural tube (NT) of a 4s embryo (arrow, purple), and its protein distribution seen in G. (D) Low-power magnification of the pituitary anlagen area, which is blown up in E. (E) Thickened neuroepithelium (ne) is positive for pituitary marker Six3 that overlies Wise expression (purple) within the ventral epithelium (arrow) at HH stage 15. (F) Hair follicle placodes express Wise at HH stage 35 (arrow), and its protein is found in center of the placode H. Murine expression of Wise (I, M, and P) and Sost (J, N, and Q). (I) Wise (purple) is found in mouse otic vesicle and branchial arches, BA1–2, at stages 11dpc. (J) Sost (red) is also found in the same otic vesicle location as wise at 10 dpc. Sost can be seen in the dorsal ventricular layer (dvl) of the otic vesicle (otic epithelium, oe). (L and O) H&E phase of areas shown in M–Q. m, molar; I, incisor; tb, trabecular bone. (L–Q) All taken at 16.5 dpc. Wise is expressed in molar inner enamel epithelium (M) and incisor (P), whereas Sost is expressed in odontoblasts (N) and osteoblasts of the trabecular bone (N). Scale bar measurements; A, 100 ␮m; B, 46 ␮m; C, 100 ␮m; D, 100 ␮m; E, 40 ␮m; F, 550 ␮m; G, 100 ␮m; H, 600 ␮m; I, 150 ␮m; J, 22.5 ␮m; L–Q, 1200 ␮m.

2E). Wise is also found to be expressed in the hair follicle placodes at HH stage 35 (Figs. 2F and 2H). In the avian, we see transient expression of Wise early during development of many organs that are known to have Wnt–BMP interplay. Whereas there are common sites of expression, many aspects of the expression of Wise is substantially different in the mouse. Wise is not expressed in the murine prechordal plate, surface ectoderm apposing the neural tube, or pituitary anlagen. It is expressed from 9.5 to 11 dpc in the branchial arches (edges of BA1–BA2), areas known to give rise to developing teeth (Fig. 2I). Later, at 16.5 dpc, Wise is found in the molar inner enamel epithelium (Fig. 2M) and incisors (Fig. 2P). Wise is also expressed in the dorsal ventricular layer of the otic vesicle from 9.5 to 10.5 dpc (Fig. 2I) and in the hair follicles as seen in the avian expression pattern for Wise. Sost is expressed in the same cell type and stage as Wise for the otic vesicle (Fig. 2J). However, Sost differs from Wise in that it is not expressed in the branchial arches, but it is expressed in complementary cell types in the more differentiated tooth. Sost is found in the polarized odontoblasts (Fig. 2N), which give rise to the adult alveolar

bone. Both Wise and Sost are transcribed at E14.5 developing murine osteoblasts during bone development (Figs. 3Q–3S).

Co-localization of Sost, Wise, and LRP5/6 proteins in postnatal mouse bone Recent evidence has shown that Sost physically interacts with Lrp5 and Lrp6, and human mutations affecting either SOST or LRP5 have profound effects on bone deposition. We thus wanted to establish whether interactions between Wise, Sost, LRP5, and LRP6 are functionally relevant to bone deposition by analyzing protein distribution in relevant bone cell types by immunohistochemistry (Fig. 3). We used mouse bones at P0 because they contain most bone/chondrocyte cell types, with the exception of bone lining cells and few osteocytes. At P0, we found that Sost and Wise proteins are localized to the nucleus (nuclear membrane) and membrane of hypertrophic and prehypertrophic chondrocytes in the growth plate (Figs. 3A–3G and 3M–3P). Wise also displays punctuate cellular staining in proliferating chondrocytes and osteoblasts at 2 weeks postnatally (Fig. 3H). LRP5 and LRP6 were detected using the


1743

FIG. 3. Wise, Sost, and LRP5/6 localization in murine bone. (A–D) Sost immunolocalization at P0. (A) H&E-stained representative section for B. (B) Sost murine polyclonal antibody at P0. Black or white arrow in all panels represents ossification front between chondrocytes and osteoblasts. To the left of arrow are the hypertrophic chondrocytes and to the right the trabecular network consisting of osteoblast cells. (C) Nomarski image of chondrocytes in growth plate, representative for D. (D) Sost polyclonal antibody localized to cytoplasmic membrane and nuclear membrane (yellow arrowhead). (E–H) Wise immunolocalization at P0 (E and F ; 2 weeks, G and H). (E) Nomarski image for F. (F) Wise polyclonal antibody at P0. (G) Wise polyclonal antibody (green) and Dapi nuclear stain (Blue). Yellow arrows point to Wise nuclear staining. (H) Wise polyclonal antibody showing punctate staining in proliferating chondrocytes. (I–L) LRP5 and LRP6 immunolocalization at P0. (I) H&E representative sections for J. (J) 1A12 antibody staining both LRP5 and 6. (K) 1A12 staining (red) in conjunction with Dapi nuclear stain (blue). (L) Co-localization of Sost (green) with 1A12 LRP5/6 antibody staining (yellow arrow) in hypertrophic chondrocyte cell. (M–P) Nomarski, Sost polycloal antibody, Dapi nuclear stain, merged Dapi (blue) and Sost (green) images. Yellow arrowheads show overlapping nuclear staining (membrane) at P0. (Q–S) In situ mRNA detection in sections through ulna of a mouse bone at E14.5 dpc. Wise mRNA is found in osteoblasts and pre-osteoblasts (flat osteoblast cells). Sost mRNA is only found in pre-osteoblasts (flat osteoblasts) at E14.5 dpc. (W) ALP staining at E14.5 showing localization of osteoblast cells producing ALP. Scale bars measurements: A and B, 257.5 ␮m; C and D, 33.75 ␮m; E and F, 128.75 ␮m; G, 33.75 ␮m; H, 32.25 ␮m; I and J, 257.5 ␮m; K and L, 128.75 ␮m; M–P, 33.75 ␮m, Q–S, 257.5 ␮m.

commercial 1A12 antibody that recognizes the second YWTD repeat in both LRP5 and LRP6. LRP5/6 is localized to osteoblast membranes and hypertrophic chondrocyte membranes (Figs. 3I–3K). Interestingly, not all cells within one layer are positive for Sost, Wise, or LRP5/6, possibly indicating that another mechanism is at play to titrate the amount of Wnt signaling during bone development. Double labeling reveals that Wise, Sost, and LRP5/6 co-localize to the hypertrophic chondrocytes at P0 (Fig. 3L). All antibodies were initially tested alongside samples with only preimmune and samples without primary antibody. This expression analysis shows that the proteins are present in the relevant cell types and stages consistent with their putative roles in modulating bone deposition.

Sost inhibits Wnt signaling in Xenopus assays Wise has the dual ability to stimulate or inhibit Wnt signaling by directly binding to the LRP6 co-receptor.(4) This raises the question of whether Sost has similar abilities to modulate Wnt signaling. Wise RNA is able to activate the canonical Wnt pathway and induce engrailed-2 expression(4) and other markers(31) in a manner similar to that of Wnt8 using Xenopus animal cap assays (Fig. 4C). To study Wnt inhibitory potential of Sost, we used a different assay for Wnt signaling. Injection of Wnt8 RNA into developing Xenopus embryos induces a duplicate body axis and expression of the direct target genes Siamois and Xnr3 (Figs. 4A and 4B).(32,33) As had been previously reported by Se-

1744

ELLIES ET AL.

FIG. 4. Sost inhibits Wnt signaling. (A) Xenopus secondary axis formation assay with control mock saline, Wnt8 or Wnt8 + Sost–injected RNA. Sost inhibits the Wnt8induced secondary axis formation. (B) Assay for the immediate early Wnt8 response genes Siamois and Xnr3. Sost blocks the ability of Wnt8 to induce Siamois and Xnr3. Intracellular stimulation of the canonical Wnt pathway using ␤-catenin is unaffected by Sost. EF1␣ serves as a loading control. (C) Animal cap assay for engrailed-2 expression. Sost in combination with Noggin is unable to induce engrailed-2(en-2) expression. In contrast, both Wise and Wnt8 induce engrailed-2.

menov et al.,(3) we also observed that Sost RNA inhibits the ability of Wnt8 to induce the formation of a duplicate axis and blocks activation of the Siamois and Xnr3 target genes (Fig. 4B). However, we also observed that Sost does not completely rescue (Fig. 4A) the duplicated axis. In addition, injection of Sost RNA over a wide range of concentrations does not induce engrailed-2 expression, indicating that it does not possess the ability to activate Wnt signaling like Wise (Fig. 4C). Like Wise, Sost RNA does not block intracellular stimulation of the Wnt pathway using a stabilized form of ␤-catenin (Fig. 4B), suggesting that like Wise it works extracellularly to inhibit Wnt activity.(4) These experiments show that both Wise and Sost can inhibit Wnt signaling but only Wise activates the pathway.

Sost binds to LRP5 and LRP6 but not to the human HBM LRP5G171V variant We next studied the mechanisms by which Sost inhibits Wnt signaling. Wise can inhibit Wnt activity by binding to the N-terminal region of LRP6, which contains the first two of four EGF/YWTD propeller repeats.(4) Recently, Li et al.(34) and Semenov et al.(3) (2005) showed that Sost physically interacts with the first two YWTD domains of Lrp5 and Lrp6. Using biochemical assays with epitope tagged versions of Sost and soluble forms of LRP5 or LRP6 from cell supernatants, we also found that SostFLAG is able to immunoprecipitate LRP5myc or LRP6IgG as detected by Western blotting (Fig. 5B). A five times HBM trait in humans has been mapped to LRP5 and found to correspond to a G171V substitution in the first YWTD propeller repeat.(5) This domain is highly conserved between LRP5 and LRP6 (Fig. 5A). In light of the increased bone mass in sclerosteosis (SOST−/−) and the finding that Sost physically interacts with LRP5 and LRP6, we studied the effects that a LRP5G171V change and an analogous LRP6G158V change would have on interactions with Sost. Transient transfection of either LRP5 or LRP5G171V resulted in similar amounts of secreted LRP protein in the media (Fig. 5D). Interestingly, our mutant substitution in this region of both LRP5 and LRP6 abolishes binding with Sost in vitro (Fig.

5B). Furthermore, transfecting constructs expressing the wildtype and variant proteins into mammalian cells also reveals that the G171V mutation in LRP5, and analogous LRP6 mutation, abolishes Sost interaction in vivo (Fig. 5C). These results show that the HBM phenotype seen in human sclerosteosis is the result of its inability to interact with LRP5, which is similar to the human LRP5G171V HBM phenotype. This suggests that there may be a conformational change in LRP5 and LRP6 that alters the ability of Sost to interact and modulate their activity.

Sost and requirements for attenuation of Wnt reporter activity in cell culture To complement the Xenopus Wnt assays and characterize Sost–LRP interactions in modulating activity, we used the Wnt Topflash reporter system.(35) This assay links three inverted Wnt-responsive TCF regulatory elements to a Luciferase reporter gene to monitor the activation of a Wntdependent target in mammalian cell culture (Fig. 6). Using an alternate class 1 Wnt ligand, Wnt3a, we showed an induction of Wnt reporter activity by ∼3-fold (Fig. 6A, lane 2), and this induction is dependent on LRP5 and/or LRP6 as an ␣LRP-antibody (Fig. 6A,1A12) against these proteins reduces reporter activity to basal levels (Fig. 6, lane 4). The addition of Sost conditioned media attenuates Wnt3amediated induction by 40% (Fig. 6A, lane 3), and the specificity of this inhibition is shown using specific monoclonal antibodies (4G10 and 4A6) directed against murine Sost, which completely restores reporter activity (Fig. 6, lane 5).

Functional characterization of LRP5 and LRP5G171V variant in Wnt reporter assays To determine the role of LRP5G171V mutation on LRP5 function, we generated full-length transmembrane (tm) versions of both LRP5 and LRP5G171V and co-transfected 293 cells to study their effect on Wnt activity (Fig. 6B). Based on the inability of Dkk1 to inhibit Wnt signaling in cell culture assays using the LRP5G171V variant, it has been widely assumed that the LRP5G171V mutation represents a gain-of-function change in LRP5.(9) Transfection of LRP5


1745

FIG. 5. Sost binding to LRP5/6 but not to LRP5G171V. (A) Amino acid alignment between LRP5 and LRP6 of area encompassing G171V mutation. Conserved amino acids shown in yellow and semiconserved in green. (B) Immunoprecipitation assay showing that Sost can pull down soluble forms of LRP5 or LRP6 but not LRP5G171V or LRP6G158V from cell supernatants. (C) In vivo functional evidence that Sost can not bind to LRP5 G171V . Cells were transected with transmembrane forms of either LRP5 or LRP5G171V, and conditioned media containing SostFlag was added to the cells. Immunohistochemistry reveals that SostFlag can be detected on many more cell surfaces of the LRP5 transfected cells than the LRP5G171V cell surfaces. (D) Western blot showing the abundance of secreted LRP protein (day 1 [D1], day 2 [D2], etc.) in the condition media taken from cells that were transiently transfected with either soluble LRP5 or LRP5G171V.

or the LRP5G171V variant alone do not activate the Wnt reporter, whereas we find that Wnt3a conditioned media dramatically stimulates reporter activity (Fig. 6B, lanes 1–3). Interestingly, in the presence of the Wnt3a conditioned media, transfected LRP5 functioned as a Wnt inhibitor (Fig. 6B, lane 4), presumably by titrating out the available Wnt3a from the rate limiting amounts of endogenous Frizzled. We find that the LRP5G171V variant is much less effective at inhibiting Wnt3a-dependent activity (Fig. 6B, lane 5). Under these same conditions (Wnt3a and LRP5G171V), the addition of Sost leads to a three time reduction in reporter activity (Fig. 6B, lane 6). Because we have shown that Sost is unable to bind to Lrp5G171V (Fig. 5), this implies that the inhibition arises because of an effect of Sost on endogenous LRPs and Frizzled. These results are consistent with a model whereby the changes induced in LRP5G171V variant alter its ability to bind Wnt3a and/or inhibitors such as Sost, Wise, and Dkk1. To address whether regions of the Sost protein other than the cysteine knot are important for interaction of with

LRP6 and in vivo activity, we used site-directed mutagenesis. By comparing protein sequences of various species of Wise and Sost, we noted a conserved domain downstream of the leader sequence and potential N-glycosylation site. We introduced similarly charged amino acid substitutions (M1:GGR to AVS) within this region of the Sost backbone (Fig. 7A). The M1 protein variant of Sost acted like the wildtype version of Sost in its stability during freeze thaws, ability to be secreted, and its abundance (Fig. 7D). This M1 variant displayed reduced binding to LRP6 relative to wildtype Sost (Fig. 7B). Furthermore, this mutation has functional implications on its ability to modulate Wnt signaling, because it displays reduced inhibition of the Wnt Topflash reporter expression (p < 0.05; Fig. 7C).

DISCUSSION In this study, we characterized interactions between Sost and LRP5 or LRP6 and showed the functional ability of Sost to inhibit Wnt signaling in vivo, which has important

1746

FIG. 6. Sost attenuates Wnt activity by 40% and ␣-Sost Mab restores this Sost inhibition. The 293 cells were co-transfected using Topflash and incubated with conditioned media containing Wnt3a, Sost, ␣-LRP5/6, and/or ␣-Sost Mab. Wnt3a conditioned media (cm) activated Topflash luciferase activity 3-fold (lane 2) and ␣-LRP5/6 significantly abolished Wnt3a activity to basal levels (lane 4). Sost-cm significantly reduced Wnt3a activity by 40% (lane 3), and ␣-Sost was able to significantly restore the Wnt3a activity (lane 5). All experiments were done in triplicate and normalized to Renilla controls. (B) 293 cells were co-transfected using Topflash and transmembrane (tm) versions of either LRP5 or LRP5G171V. The transfected cells were incubated with conditioned media containing Wnt3a and/or Sost. Wnt3a-cm was able to activate the Topflash reporter (lane 1), and the transfection of LRP5tm caused a loss of this Wnt3a activity (lane 4). Transfection of LRP5G171Vtm with Wnt3a-cm showed a loss of LRP5 inhibitory effect (lane 5). Addition of Sost-cm to the LRP5G171Vtm– transfected cells caused the expected Sost attenuation of the Wnt3a activity (lane 6), even though Sost was unable to bind to LRP5G171Vtm. This suggests that Sost is attenuating endogenous LRP5 Wnt3a activity and that LRP5G171V may be unable to bind to Wnt and thus not acting as a gain of LRP5 function.

implications in understanding the regulation of bone deposition. It has been proposed that the recessive HBM phenotype seen in patients with sclerosteosis is caused by the potential role of SOST as a BMP inhibitor.(14,15,17,36) A much higher concentration of Sost is required for its role as a BMP antagonist than its ability to inhibit Wnt-mediated ALP induction,(36) suggesting Sost may have a higher affinity for the Wnt/LRP pathway. Winkler et al.(15) reported that Sost preferentially inhibits BMPs and indirectly impacts the Wnt pathway because of sclerostin’s inability to alter ␤-catenin protein levels and thus stability. We inter-

ELLIES ET AL. pret the ␤-catenin result of Winkler et al.(15) to be caused by Sost’s partial inhibitory effect on the Wnt pathway. Our data are consistent with the idea that Sost acts as a direct modulator of Wnt activity, acting to partially (40%) inhibit Wnt signaling. These results are important because recent studies have implicated Wnt signaling as a key component of bone formation and homeostasis.(6–8) This focuses attention on the need to understand the molecular mechanisms that modulate the activity of the Wnt pathway. In vitro, Sost binds equally well to both LRP5 and LRP6, and genetic studies in mice with single and compound mutants have shown that Lrp5 and Lrp6 have both unique and overlapping functions.(37,38) Therefore, Sost has the potential to modulate Wnt signaling in a number of tissues and developmental processes in addition to its role in bone. This study shows that Sost is an effective direct inhibitor of the Wnt pathway, like Wise, CTGF, and Cyr61. In light of the significant similarity with the cysteine motifs from DAN (Cerberus, DAN, Gremlin, Caronte), CCN (NOV, Cef10, CTGF, Cyr61), Slit, and Mucin family members and individual genes, such as VFW and NDP (Fig. 1B), this raises the possibility that many other members of this superfamily may have unanticipated roles in regulating the Wnt pathway, extracellularly through LRPs or intracellularly through GSK3. Whereas there are similarities in the inhibitory activity of Wise and Sost, our data show that there are also differences in their regulatory potential. Wise is able to activate Wnt signaling in Xenopus assays, whereas Sost lacks this ability (Fig. 4C). In axial duplication experiments, we also observed that Sost is unable to completely restore a wildtype trunk (tail) (Fig. 4A), unlike Wise and CTGF. Differences between Sost and Wise or CTGF in modulating Wnt activity might arise from the nature of their interactions with LRP5 or LRP6. Competitive binding and functional studies have yet to be done to establish if they bind to the same or different regions of LRP and if simultaneous interactions occur. The multiple YWTD propeller and EGF domains provide an opportunity for multiple protein interactions; hence, it will be essential to understand the functional consequences of the combined activity of these modulators in controlling the Wnt pathway. By showing that Sost can bind to LRP5 or LRP6 and modulate Wnt signaling, we have uncovered a means by which Sost can control bone formation and homeostasis. In light of these findings, the human sclerosteosis disease, seen on loss of SOST function, is likely to arise because of decreased inhibition of the Wnt pathway. Similarly, our finding that Sost does not physically interact with the LRP5G171V variant seen in the human HBM trait(5) or with the analogous LRP6G158V amino acid substitution (Fig. 5B) is consistent with the idea that it is no longer able to inhibit Wnt signaling mediated by this LRP5 variant. The LRP5G171V mutation seems to act as a gain of Wnt function. However, it is unknown if Wnt is able to interact with LRP5 harboring the G171V mutation. Therefore, the LRP5G171V mutation in the first YWTD propeller domain seems to cause a change in the structure or conformation leading to a loss of binding to Sost or Wnt3a, which normally interact with this domain. In addition, there may be a


1747

FIG. 7. Mapping of Sost–LRP6 interaction domains. (A) Alignment of predicted mouse protein sequences for Wise and Sost. Shaded marks position of the conserved cysteine residues, the leader sequence cleavage site is denoted with a horizontal bar, the intron exon junction with a vertical line, the cleavage signal with an *, and a putative Nglycosylation with an N-Gly. The horizontal box depicts the conserved domain outside the cysteine knot that contains the N-Gly and M1 regions. The vertical box indicates the GGR to AVS amino acid variant of Sost (M1) used for LRP6 binding assay. (B) Relative binding of Sost variant M1 to LRP6. M1 has reduced LRP6 binding ability. (C) In vivo functional relevance of Sost variant LRP6 relevant binding. Sost M1 variant has significantly lowered ability to bind to LRP6 in vivo. (D) Western blot of wildtype Sost (from media of D1 or D2) and the M1 variant of Sost subjected to freeze thaws. Lane demarcated using asterisks shows the abundance of Sost M1 when it was not subjected to a freeze thaw.

more global change in the structure of LRP5 because it has also lost the ability to be inhibited by Dkk1,(9) which interacts with the third YWTD propeller domain. Further evidence for the role of LRP5 and LRP6 and Wnt signaling in regulating BMD comes from human and mouse loss-offunction mutations in LRP5(10,39) and the mouse hypomorphic ringelschwanz Lrp6R886W allele.(40) LRP5 is expressed in osteoblasts,(5) and recent evidence has shown that an osteoblast-specific loss of ␤-catenin results in low bone mass.(41) The Lrp6R886W allele contains a mutation in the third YWTD propeller repeat, which corresponds to the binding region for Dkks.(42) Combined, our results underscore the idea that Sost and Wise represent a new subfamily of cysteine knot proteins with the dual potential to exert their regulatory functions through modulation of both the Wnt and BMP signaling pathways.

ACKNOWLEDGMENTS The authors thank Arcady Mushegian for protein bioinformatics, Teri Johnson for immunohistological expertise, Eric Schuenemann for site-directed mutagenesis, and Ron and Joan Conaway and members of the Conaway Laboratory for reagents and advice on protein analysis. We also thank members of the Krumlauf Laboratory for stimulating discussions. This work was supported by SIMR funds awarded to RK; NIMR support to NI and RK; and NIH Grants DK56063 and HD39952 and a March of Dimes Birth Defects Foundation Grant to SS.

REFERENCES 1. Balemans W, Ebeling M, Patel N, Van Hul E, Olson P, Dioszegi M, Lacza C, Wuyts W, Van Den Ende J, Willems P, Paes-Alves AF, Hill S, Bueno M, Ramos FJ, Tacconi P, Dik-

1748

2.

3. 4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

kers FG, Stratakis C, Lindpaintner K, Vickery B, Foernzler D, Van Hul W 2001 Increased bone density in sclerosteosis is due to the deficiency of a novel secreted protein (SOST). Hum Mol Genet 10:537–543. van Bezooijen RL, Roelen BA, Visser A, van der Wee-Pals L, de Wilt E, Karperien M, Hamersma H, Papapoulos SE, ten Dijke P, Lowik CW 2004 Sclerostin is an osteocyte-expressed negative regulator of bone formation, but not a classical BMP antagonist. J Exp Med 199:805–814. Semenov M, Tamai K, He X 2005 SOST is a ligand for LRP5/ LRP6 and a WNT signaling inhibitor. J Biol Chem 280:26770– 26775. Itasaki N, Jones CM, Mercurio S, Rowe A, Domingos PM, Smith JC, Krumlauf R 2003 Wise, a context-dependent activator and inhibitor of Wnt signalling. Development 130:4295– 4305. Little RD, Carulli JP, Del Mastro RG, Dupuis J, Osborne M, Folz C, Manning SP, Swain PM, Zhao SC, Eustace B, Lappe MM, Spitzer L, Zweier S, Braunschweiger K, Benchekroun Y, Hu X, Adair R, Chee L, FitzGerald MG, Tulig C, Caruso A, Tzellas N, Bawa A, Franklin B, McGuire S, Nogues X, Gong G, Allen KM, Anisowicz A, Morales AJ, Lomedico PT, Recker SM, Van Eerdewegh P, Recker RR, Johnson ML 2002 A mutation in the LDL receptor-related protein 5 gene results in the autosomal dominant high-bone-mass trait. Am J Hum Genet 70:11–19. Glass DA II, Bialek P, Ahn JD, Starbuck M, Patel MS, Clevers H, Taketo MM, Long F, McMahon AP, Lang RA, Karsenty G 2005 Canonical Wnt signaling in differentiated osteoblasts controls osteoclast differentiation. Dev Cell 8:751–764. Hill TP, Spater D, Taketo MM, Birchmeier W, Hartmann C 2005 Canonical Wnt/beta-catenin signaling prevents osteoblasts from differentiating into chondrocytes. Dev Cell 8:727– 738. Day TF, Guo X, Garrett-Beal L, Yang Y 2005 Wnt/betacatenin signaling in mesenchymal progenitors controls osteoblast and chondrocyte differentiation during vertebrate skeletogenesis. Dev Cell 8:739–750. Boyden LM, Mao J, Belsky J, Mitzner L, Farhi A, Mitnick MA, Wu D, Insogna K, Lifton RP 2002 High bone density due to a mutation in LDL-receptor-related protein 5. N Engl J Med 346:1513–1521. Gong Y, Slee RB, Fukai N, Rawadi G, Roman-Roman S, Reginato AM, Wang H, Cundy T, Glorieux FH, Lev D, Zacharin M, Oexle K, Marcelino J, Suwairi W, Heeger S, Sabatakos G, Apte S, Adkins WN, Allgrove J, Arslan-Kirchner M, Batch JA, Beighton P, Black GC, Boles RG, Boon LM, Borrone C, Brunner HG, Carle GF, Dallapiccola B, De Paepe A, Floege B, Halfhide ML, Hall B, Hennekam RC, Hirose T, Jans A, Juppner H, Kim CA, Keppler-Noreuil K, Kohlschuetter A, LaCombe D, Lambert M, Lemyre E, Letteboer T, Peltonen L, Ramesar RS, Romanengo M, Somer H, Steichen-Gersdorf E, Steinmann B, Sullivan B, Superti-Furga A, Swoboda W, van den Boogaard MJ, Van Hul W, Vikkula M, Votruba M, Zabel B, Garcia T, Baron R, Olsen BR, Warman ML 2001 LDL receptor-related protein 5 (LRP5) affects bone accrual and eye development. Cell 107:513–523. Mercurio S, Latinkic B, Itasaki N, Krumlauf R, Smith JC 2004 Connective-tissue growth factor modulates WNT signalling and interacts with the WNT receptor complex. Development 131:2137–2147. Bafico A, Liu G, Yaniv A, Gazit A, Aaronson SA 2001 Novel mechanism of Wnt signalling inhibition mediated by Dickkopf-1 interaction with LRP6/Arrow. Nat Cell Biol 3:683–686. Mao B, Wu W, Li Y, Hoppe D, Stannek P, Glinka A, Niehrs C 2001 LDL-receptor-related protein 6 is a receptor for Dickkopf proteins. Nature 411:321–325. Brunkow ME, Gardner JC, Van Ness J, Paeper BW, Kovacevich BR, Proll S, Skonier JE, Zhao L, Sabo PJ, Fu Y, Alisch RS, Gillett L, Colbert T, Tacconi P, Galas D, Hamersma H, Beighton P, Mulligan J 2001 Bone dysplasia sclerosteosis results from loss of the SOST gene product, a novel cystine knotcontaining protein. Am J Hum Genet 68:577–589.

ELLIES ET AL. 15. Winkler DG, Sutherland MK, Geoghegan JC, Yu C, Hayes T, Skonier JE, Shpektor D, Jonas M, Kovacevich BR, StaehlingHampton K, Appleby M, Brunkow ME, Latham JA 2003 Osteocyte control of bone formation via sclerostin, a novel BMP antagonist. EMBO J 22:6267–6276. 16. Loots GG, Kneissel M, Keller H, Baptist M, Chang J, Collette NM, Ovcharenko D, Plajzer-Frick I, Rubin EM 2005 Genomic deletion of a long-range bone enhancer misregulates sclerostin in Van Buchem disease. Genome Res 15:928–935. 17. Balemans W, Foernzler D, Parsons C, Ebeling M, Thompson A, Reid DM, Lindpaintner K, Ralston SH, Van Hul W 2002 Lack of association between the SOST gene and bone mineral density in perimenopausal women: Analysis of five polymorphisms. Bone 31:515–519. 18. Ellies DL, Church V, Francis-West P, Lumsden A 2000 The WNT antagonist cSFRP2 modulates programmed cell death in the developing hindbrain. Development 127:5285–5295. 19. Latinkic BV, Mercurio S, Bennett B, Hirst EM, Xu Q, Lau LF, Mohun TJ, Smith JC 2003 Xenopus Cyr61 regulates gastrulation movements and modulates Wnt signalling. Development 130:2429–2441. 20. Perbal B 2001 NOV (nephroblastoma overexpressed) and the CCN family of genes: Structural and functional issues. Mol Pathol 54:57–79. 21. Bork P 1993 The modular architecture of a new family of growth regulators related to connective tissue growth factor. FEBS Lett 327:125–130. 22. Meitinger T, Meindl A, Bork P, Rost B, Sander C, Haasemann M, Murken J 1993 Molecular modelling of the Norrie disease protein predicts a cystine knot growth factor tertiary structure. Nat Genet 5:376–380. 23. Piccolo S, Agius E, Leyns L, Bhattacharyya S, Grunz H, Bouwmeester T, De Robertis EM 1999 The head inducer cerberus is a multifunctional antagonist of Nodal, BMP and Wnt signals. Nature 397:707–710. 24. Hsu DR, Economides AN, Wang X, Eimon PM, Harland RM 1998 The Xenopus dorsalizing factor Gremlin identifies a novel family of secreted proteins that antagonize BMP activities. Mol Cell 1:673–683. 25. Stanley E, Biben C, Kotecha S, Fabri L, Tajbakhsh S, Wang CC, Hatzistavrou T, Roberts B, Drinkwater C, Lah M, Buckingham M, Hilton D, Nash A, Mohun T, Harvey RP 1998 DAN is a secreted glycoprotein related to Xenopus cerberus. Mech Dev 77:173–184. 26. Yokouchi Y, Vogan KJ, Pearse RV II, Tabin CJ 1999 Antagonistic signaling by Caronte, a novel Cerberus-related gene, establishes left-right asymmetric gene expression. Cell 98:573– 583. 27. Abreu JG, Ketpura NI, Reversade B, De Robertis EM 2002 Connective-tissue growth factor (CTGF) modulates cell signalling by BMP and TGF-beta. Nat Cell Biol 4:599–604. 28. Yanagita M, Oka M, Watabe T, Iguchi H, Niida A, Takahashi S, Akiyama T, Miyazono K, Yanagisawa M, Sakurai T 2004 USAG-1: A bone morphogenetic protein antagonist abundantly expressed in the kidney. Biochem Biophys Res Commun 316:490–500. 29. Laurikkala J, Kassai Y, Pakkasjarvi L, Thesleff I, Itoh N 2003 Identification of a secreted BMP antagonist, ectodin, integrating BMP, FGF, and SHH signals from the tooth enamel knot. Dev Biol 264:91–105. 30. Kusu N, Laurikkala J, Imanishi M, Usui H, Konishi M, Miyake A, Thesleff I, Itoh N 2003 Sclerostin is a novel secreted osteoclast-derived bone morphogenetic protein antagonist with unique ligand specificity. J Biol Chem 278:24113–24117. 31. O’Shaughnessy RF, Yeo W, Gautier J, Jahoda CA, Christiano AM 2004 The WNT signalling modulator, Wise, is expressed in an interaction-dependent manner during hair-follicle cycling. J Invest Dermatol 123:613–621. 32. Smith WC, Harland RM 1991 Injected Xwnt-8 RNA acts early in Xenopus embryos to promote formation of a vegetal dorsalizing center. Cell 67:753–765.

SOST INTERACTS WITH LRP5 TO MODULATE WNT ACTIVITY 33. Sokol S, Christian JL, Moon RT, Melton DA 1991 Injected Wnt RNA induces a complete body axis in Xenopus embryos. Cell 67:741–752. 34. Li X, Zhang Y, Kang H, Liu W, Liu P, Zhang J, Harris SE, Wu D 2005 Sclerostin binds to LRP5/6 and antagonizes canonical Wnt signaling. J Biol Chem 280:19883–19887. 35. Korinek V, Barker N, Willert K, Molenaar M, Roose J, Wagenaar G, Markman M, Lamers W, Destree O, Clevers H 1998 Two members of the Tcf family implicated in Wnt/beta-catenin signaling during embryogenesis in the mouse. Mol Cell Biol 18:1248–1256. 36. Winkler DG, Sutherland MS, Ojala E, Turcott E, Geoghegan JC, Shpektor D, Skonier JE, Yu C, Latham JA 2005 Sclerostin Inhibition of Wnt-3a-induced C3H10T1/2 Cell Differentiation Is Indirect and Mediated by Bone Morphogenetic Proteins. J Biol Chem 280:2498–2502. 37. Kelly OG, Pinson KI, Skarnes WC 2004 The Wnt co-receptors Lrp5 and Lrp6 are essential for gastrulation in mice. Development 131:2803–2815. 38. Holmen SL, Giambernardi TA, Zylstra CR, Buckner-Berghuis BD, Resau JH, Hess JF, Glatt V, Bouxsein ML, Ai M, Warman ML, Williams BO 2004 Decreased BMD and limb deformities in mice carrying mutations in both Lrp5 and Lrp6. J Bone Miner Res 19:2033–2040. 39. Kato M, Patel MS, Levasseur R, Lobov I, Chang BH, Glass DA II, Hartmann C, Li L, Hwang TH, Brayton CF, Lang RA, Karsenty G, Chan L 2002 Cbfa1-independent decrease in os-

1749

teoblast proliferation, osteopenia, and persistent embryonic eye vascularization in mice deficient in Lrp5, a Wnt coreceptor. J Cell Biol 157:303–314. 40. Kokubu C, Heinzmann U, Kokubu T, Sakai N, Kubota T, Kawai M, Wahl MB, Galceran J, Grosschedl R, Ozono K, Imai K 2004 Skeletal defects in ringelschwanz mutant mice reveal that Lrp6 is required for proper somitogenesis and osteogenesis. Development 131:5469–5480. 41. Hu H, Hilton MJ, Tu X, Yu K, Ornitz DM, Long F 2005 Sequential roles of Hedgehog and Wnt signaling in osteoblast development. Development 132:49–60. 42. Glinka A, Wu W, Delius H, Monaghan AP, Blumenstock C, Niehrs C 1998 Dickkopf-1 is a member of a new family of secreted proteins and functions in head induction. Nature 391:357–362.

Address reprint requests to: Robb Krumlauf, PhD Stowers Institute 1000 50th Street Kansas City, MO 64110, USA E-mail: [email protected] Received in original form Novmeber 21, 2005; revised form July 2, 2006; accepted August 11, 2006.