JOURNAL OF BONE AND MINERAL RESEARCH Volume 22, Number 12, 2007 Published online on August 13, 2007;

doi: 10.1359/JBMR.070802 2007 American Society for Bone and Mineral Research

Bone Regeneration Is Regulated by Wnt Signaling

Jae-Beom Kim,1,2 Philipp Leucht,1,2,3 Kentson Lam,1 Cynthia Luppen,1 Derk Ten Berge,4 Roel Nusse,4 and Jill A Helms1

ABSTRACT: Tissue regeneration is increasingly viewed as reactivation of a developmental process that, when misappropriated, can lead to malignant growth. Therefore, understanding the molecular and cellular pathways that govern tissue regeneration provides a glimpse into normal development as well as insights into pathological conditions such as cancer. Herein, we studied the role of Wnt signaling in skeletal tissue regeneration. Introduction: Some adult tissues have the ability to regenerate, and among these, bone is one of the most remarkable. Bone exhibits a persistent, lifelong capacity to reform after injury, and continual bone regeneration is a prerequisite to maintaining bone mass and density. Even slight perturbations in bone regeneration can have profound consequences, as exemplified by conditions such as osteoporosis and delayed skeletal repair. Here, our goal was to determine the role of Wnts in adult bone regeneration. Materials and Methods: Using TOPgal reporter mice, we found that damage to the skeleton instigated Wnt reporter activity, specifically at the site of injury. We used a skeletal injury model to show that Wnt inhibition, achieved through adenoviral expression of Dkk1 in the adult skeleton, prevented the differentiation of osteoprogenitor cells. Results: As a result, injury-induced bone regeneration was reduced by 84% compared with controls. Constitutive activation of the Wnt pathway resulting from a mutation in the Lrp5 Wnt co-receptor results in high bone mass, but our experiments showed that this same point mutation caused a delay in bone regeneration. In these transgenic mice, osteoprogenitor cells in the injury site were maintained in a proliferative state and differentiation into osteoblasts was delayed. Conclusions: When considered together, these data provide a framework for understanding the roles of Wnt signaling in adult bone regeneration and suggest a feasible approach to treating clinical conditions where enhanced bone formation is desired. J Bone Miner Res 2007;22:19131923. Published online on August 13, 2007; doi: 10.1359/JBMR.070802 Key words: Dkk, Lrp5, osteoblast, differentiation, repair, regeneration

INTRODUCTION
LTHOUGH HUMANS CANNOT grow new limbs like salamanders, the molecular machinery for regeneration seems to be an elemental part of our genetic makeup. The prevailing opinion in regenerative medicine is that the genes responsible for regeneration have for some reason fallen into disuse and that they may be jump-started by the selective activation of key molecular pathways. Here, we study one such pathway regulated by Wnt proteins. Wnt ligands are secreted molecules that bind to cell surface receptors encoded by the Frizzled and low-density lipoprotein receptor-related proteins (LRPs). Once bound, the ligands initiate a cascade of intracellular events that eventually lead to the transcription of target genes through

The authors state that they have no conflicts of interest.

the nuclear activity of -catenin and the DNA binding protein TCF.(13) The majority of studies have focused on -catenindependent (canonical) Wnt signaling, but accumulating evidence indicates that Wnts can also participate in -cateninindependent processes. There remains only a cursory understanding of how cells integrate and respond to Wnts, especially in an in vivo setting, and it was in this context that we initiated out study of the contribution of Wnt signaling to adult bone repair. Wnts are involved in a wide variety of cellular decisions associated with the program of osteogenesis. For example, Wnts regulate the expression level of sox9,(4) which influences the commitment of mesenchymal progenitor cells to a skeletogenic fate.(57) Wnts influence the differentiation of cells, into either osteoblasts(815) or chondrocytes.(1623) In adult animals, there is abundant evidence that Wnt signaling regulates bone mass.(24) For example, mutations in

1 Department of Surgery, Division of Plastic and Reconstructive Surgery, Stanford University, Stanford, California, USA; 2These authors contributed equally to this work; 3Department of Trauma, Hand and Reconstructive Surgery, University of Frankfurt/Main, Frankfurt, Germany; 4Department of Developmental Biology and Howard Hughes Medical Institute, Stanford University School of Medicine, Stanford, California, USA.

1913

1914 the human Wnt co-receptor LRP5 are associated with several high bone mass syndromes, including osteopetrosis type I, and endosteal hyperostosis or autosomal dominant osteosclerosis, as well as a low bone mass disease, osteoporosis-pseudoglioma.(2528) Increased production of the Wnt inhibitor Dkk1 is associated with multiple myeloma, a disease that has increased bone resorption as one of its distinguishing features.(29) Despite this strong association between Wnt signaling and bone disease, it has been difficult to precisely pinpoint how perturbations in this pathway culminate in aberrant bone remodeling. The difficulty stems in part from the temporal separation between the cause (e.g., a gene mutation present at birth) and the effect (abnormal bone mass that becomes evident later in life). This temporal separation permits other factors (e.g., humoral, environmental) to intervene and either exacerbate or ameliorate the underlying skeletal condition. Unlike the life-long process of bone remodeling, injuryinduced bone regeneration occurs within a compressed time frame and in a precise location. The cellular and molecular machinery, however, is identical between bone remodeling and bone regeneration. When an animal sustains skeletal damage, the basic challenge it faces is to detect the injury and heal the defect as rapidly as possible. This is achieved by the recruitment of skeletal progenitor cells to the injury site; once a sufficient population of progenitor cells has been generated, cells decelerate their proliferation and differentiate into osteoblasts, which secrete a matrix that eventually becomes mineralized. Osteoclasts remodel the matrix and thus restore the original shape of the bone. In a murine model, this entire process, from initial injury to new bone formation, takes 1 wk.(30) We exploited the spatiotemporal parameters of a skeletal injury model to gain a better understanding of the function of Wnt signaling in bone regeneration. Here, we report that damage to the skeleton upregulates Wnt signaling specifically at the injury site and prompts bone marrowderived cells to respond to this endogenous Wnt signal. If Wnt signaling is inhibited, bone marrowderived cells in the injury site do not differentiate into osteoblasts, and bone regeneration is terminated. If Wnt signaling is constitutively activated, cells in the injury site show heightened proliferation but once again do not differentiate into osteoblasts, and bone regeneration is temporarily delayed. These data extend our understanding of Wnt functions in the process of bone regeneration and suggest therapeutic strategies that may accelerate skeletal tissue regeneration.

KIM ET AL. ing -galactosidase gene expression. We used both strains to identify Wnt-responding cells in the injury sites.(32)

Generation of skeletal injuries

We found that all of the steps comprising fracture healing are recapitulated by generating a simple transcortical defect, consisting of a 1.0-mm hole drilled through a single cortex of the tibia. The healing response in this defect is equivalent to a stabilized fracture and has a number of advantages over other models of closed and open fractures; first, morbidity to the animal is considerably reduced by limiting the scope of the skeletal injury. Second, the repair callus is much smaller and more spatially organized, which facilitates histomorphometric analyses. Third, trauma to surrounding tissues and infections are minimal in this model of skeletal tissue regeneration. All procedures were approved by the Stanford Committee on Animal Research. Skeletally mature (1012 wk of age, male) mice were used for all studies. After anesthesia and analgesia, the right leg was shaved, and the skin was cleansed. An incision was made over the proximal medial diaphysis, and the anterior tibial muscle was divided until the medial surface of the tibia was exposed; the periosteum was preserved. Using a 1.0-mm drill bit, a hole was created that penetrated one cortex. The region was irrigated, and the skin was closed using a nonabsorbable suture. During the recovery from anesthesia, all mice were kept under heating lamps to maintain constant body temperature. Their ability to reach food and water, their surgical site, activity, weight, appearance, and behavior were monitored. Mice were killed at multiple time-points that represented the inflammatory, hard callus, and remodeling phases of healing.

Molecular and cellular assays

Under RNase-free conditions, tibias were harvested, the skin and outer layers of muscle were removed, and the tissues were washed in 1 PBS at 4C and fixed in 4% paraformaldehyde. Tissues were decalcified in 19% EDTA for 1014 days at 4C and prepared for paraffin embedding. Paraffin embedding followed our standard protocol,(33) and sections were generated at an 8- m thickness. In situ hybridization: The relevant digoxigenin-labeled mRNA antisense probes were prepared from cDNA templates for wnt3a, dkk2, wnt2b, runx2, collagen type I, collagen type II, sox9, and osteocalcin. Sections were dewaxed, treated with proteinase K, and incubated in hybridization buffer containing the riboprobe. Probe was added at an approximate concentration of 0.25 g/ml. Stringency washes of saline sodium citrate solution were done at 52C and were further washed in maleic acid buffer with 1% Tween20. Slides were treated with anti-digoxigenin antibody (Roche). For color detection, slides were incubated in Nitro blue tetrazolium chloride and 5-bromo-4-chloro-3indolyl phosphate (Roche). After developing, the slides were coverslipped with an aqueous mounting medium. Immunohistochemistry: In general, tissue sections were dewaxed followed by immersion in H2O2/PBS, washed in

MATERIALS AND METHODS Wnt reporter mice

BATgal and TOPgal transgenic mice were used as in vivo reporters of Wnt responsiveness. BATgal reporter mice carry a transgene that contains seven TCF binding motifs and the Xenopus siamois promoter.(31) The TOPgal reporter cassette contains three copies of the TCF motif CCTTTGATC upstream of a minimal c-fos promoter driv-

BONE REGENERATION IS REGULATED BY WNT SIGNALING PBS, incubated in Ficin (Zymed), treated with 0.1 M glycine, washed further, and blocked in ovalbumin (Worthington) and 1% whole donkey IgG (Jackson ImmunoResearch). Appropriate primary antibody was added and incubated overnight at 4C and washed in PBS. Samples were incubated with peroxidase-conjugated secondary antibody (Jackson ImmunoResearch) for 1 h, and a DAB substrate kit (Vector Laboratories) was used to develop the color reaction. Some commonly used antibodies included proliferating cell nuclear antigen (PCNA; Zymed) and platelet endothelial cell adhesion molecule 1 (PECAM-1; BD Biosciences). For TRACP staining, tissue sections were dewaxed and treated with a TRACP staining kit (Sigma). Quantification of cell proliferation: Quantification of PCNA+ cells was done using NIH Image J software. Four animals were used for each condition, Ad-Fc or Ad-Dkk1 treatment and 10 random PCNA-stained tissue sections were evaluated for each condition. Histology: Pentachrome and Aniline blue staining were performed as described.(34) Slides were mounted with Permount after dehydration in a series of ethanol and xylene. -galactosidase detection: Cells responsive to Wnt signaling express -galactosidase, which can be detected by Xgal staining. For Xgal staining, tissues were fixed with 0.2% glutaraldehyde for 15 min and stained with Xgal overnight at 37C. To detect the Xgal-positive cells in the adenovirustreated samples, whole mount Xgal-stained tissues were processed into paraffin blocks and sectioned. Xgal staining on cryo-sections was performed in a similar fashion to whole mount.

1915

within and among treatment groups were calculated based on these averages.

Adenovirus-mediated inhibition of Wnt signaling

All adenoviral constructs were generated previously.(35) Adenoviruses expressing the soluble Wnt antagonist, Dkk1, and the murine IgG2 Fc fragment (Ad-Fc) were infected into 293 cells at MOI 1 in T175 tissue culture flasks (7.5 106 pfu). After 2 days, cells were collected, lysed, and precipitated by centrifugation. The supernatant was subjected to CsCl gradient centrifugation at 35,000 rpm for 20 h, and the adenovirus band was harvested. After sucrose dialysis, the purified adenovirus was aliquoted and stored at 80C. Wnt inhibition was achieved by either tail vein or local, subcutaneous injection of Ad-Dkk1 and the control Ad-Fc, and injuries were generated either immediately (in the case of tail vein injections) or after 48 h (in the case of local injections).

Western blot and RT-PCR

Calluses were collected from wildtype (CD1) and BATgal mice 24 h after surgery, total RNA was extracted, and 1 g of RNA was subjected to RT-PCR using a galactosidase primer set. Western blot procedure was performed using a peroxidase-conjugated goat anti-mouse IgG2a (Southern Biotech) as described previously.(35) Signals were detected using enhanced chemiluminescence (Amersham).

Histomorphometric analyses
Tibias were collected on postsurgical day 6 to quantify the amount of new bone generated in response to injury. All tibias were embedded in paraffin and sectioned longitudinally. Four to 10 animals were used for each condition (i.e., Ad-Fc, Ad-Dkk1; see text for details). The 1.0-mm circular monocortical defect was represented across 40 tissue sections, each of which was 8 m thick. Of those 40 sections, we used a minimum of 8 sections to quantify the amount of new bone. All the tissue sections were stained with Aniline blue, which labels osteoid matrix. The sections were photographed using a Leica digital imaging system at the same magnification (5 objective). The resulting digital images were analyzed with Adobe Photoshop CS2 software. We chose a fixed, rectangular region of interest (ROI) that in all images corresponded to 106 pixels. The injury site was always represented inside this ROI by manually placing the box in the correct position on each image. The aniline bluepositive pixels were partially automated by using the magic wand tool set to a color tolerance of 60. This tolerance setting resulted in highlighted pixels with a range of blue that corresponded precisely with the histological appearance of osseous tissue in the aniline blue stained sections. Cortical surfaces, or bone fragments resulting form the drill injury, were manually deselected. The total number of aniline bluepositive pixels for each section was recorded. The pixel counts from individual sections were averaged for each tibia sample, and the differences

RESULTS Wnt signaling is active in the intact skeleton

To gain an appreciation for the level of endogenous Wnt signaling in the intact skeleton, we examined adult, skeletally mature mice for the expression of Wnt pathway components. We found that osteocytes expressed Wnts 3a, 5a, 5b, 11, Frizzled 4 (Fzd4), and Dickkopf 2 (Dkk2), whereas cells in the endosteum and bone marrow expressed Wnts 2b, 3a, 5a, 11, Dkk2, and Wnt inhibitory factor (Figs. 1A 1C; Table 1). These data showed that Wnt signaling is active in the mature skeleton, which is consistent with a role for Wnt signaling in the maintenance of adult bone mass.(10,36) They do not, however, indicate which cells are responding to the Wnt ligands, because in some cases, activators and repressors of Wnt signaling were co-expressed (Table 1, and data not shown). Therefore, we examined the intact skeleton of TOPgal and BATgal transgenic mice. These mice carry a -galactosidase reporter gene under the control of multiple copies of the Tcf/Lef binding site.(31,32) Consequently, in these mice, the spatiotemporal pattern of Xgal staining is a reflection of Wnt responsiveness within a tissue.(31,32) To confirm the sensitivity of our Xgal staining procedure in mineralized tissues, we first examined the adult growth plate, where Wnt signaling is known to regulate bone formation.(37) We found that hypertrophic chondrocytes and trabecular osteoblasts in the TOPgal growth plate were Xgal positive (n 10; Fig. 1D), which verified the sensi-

1916

KIM ET AL.

FIG. 1. Wnt signaling in the intact and injured skeleton. Components of the Wnt pathway, including (A) Wnt3a and (B) Dkk2, are expressed in periosteum (po) and cortical bone (cb), whereas other genes including (C) Wnt2b are expressed solely in periosteum. These expression patterns and others are summarized in Table 1. (D) Hypertrophic chondrocytes (hc) and trabecular osteoblasts (tb) in the TOPgal growth plate are Xgal positive. (E) Reporter activity is also seen in osteocytes of the cortical bone (cb). (F) A subpopulation of bone marrow cells (bm) is Xgal positive. (G) To study the role of Wnt signaling in skeletal repair an injury model is used, which consists of a 1.0mm hole that penetrates one cortex and enters into the bone marrow (bm) cavity, but leaves the second cortex intact. (H) Newly deposited osteoid matrix, which stains light blue, is readily observable by postsurgical d6 in the injury site (is). (I) RT-PCR shows that, in BATgal mice, the reporter -galactosidase is expressed within 48 h of injury (red arrow). (J) Detection of -galactosidase by Xgal staining shows that Wnt responsive cells are tightly confined to the injury site (is) at postsurgical day 3 (dotted line circumscribes the cortical bone). (K) Doublestaining for Xgal and the inflammatory cell marker Gr-1 shows no co-localization of Wnt responsiveness and inflammation. (LN) In situ hybridization for components of the Wnt pathway shows their widespread expression throughout the injury site. Wnt3a, Dkk2, and Wnt2b showed representative expression patterns; other gene expression patterns are summarized in Table 1. Scale bar in AF: 100 m; in H: 300 m; in JN: 200 m.

tivity of the reporter system. Osteocytes embedded in the bone cortex also stained for Xgal (n 10; Fig. 1E), as did a very small percentage of cells in the bone marrow (n 10; Fig. 1F). Together with the expression data, these findings indicated that Wnt signaling was functional, albeit at reduced levels, in the intact skeleton.

Damage to the skeleton upregulates Wnt signaling at the site of injury

Our next objective was to determine the relationship between Wnt signaling and skeletal repair. We found that creating a fracture is not necessary to stimulate skeletal

BONE REGENERATION IS REGULATED BY WNT SIGNALING

TABLE 1. EXPRESSION PATTERNS
OF

1917
IN THE INTACT AND INJURED

WNT COMPONENTS

TIBIAS

Expression domains Bone marrow Gene Wnt 2b Wnt 3a Wnt 7a Wnt 5a Wnt 5b Wnt 11 Frizzled 4 FrzB Dickkopf 2 Wnt inhibitory factor Periosteum + + + + + + + + + + Osteocytes + + + + + + Endosteum + + + + + + Injury site + + + + + + + + + Adjacent to injury + + + + + + Growth plate chondrocytes + + + + + + + + +

Expression patterns of Wnt ligands and related proteins in intact and injured tibias. In situ hybridizations were performed for each gene listed, and the presence or absence of a hybridization signal was recorded. The intact skeletal elements and the injury sites were subdivided into periosteum, endosteum, and bone marrow.

repair. A simple transcortical defect, created by drilling a 1.0-mm hole through a single cortex in the tibia, induces the same healing response as a fracture (Fig. 1G). This model has a number of advantages over previous models of closed and open fractures; first, pain and suffering to the animal is considerably reduced by limiting the size of the skeletal injury. Second, the repair callus is much smaller and less amorphous, which facilitates histomorphometric analyses. Third, trauma to surrounding tissues and infections are all but nonexistent in this simplified, more humane model of skeletal tissue regeneration. As with all other skeletal injuries, the skeletal defect we created heals by the recruitment of skeletal progenitor cells to the injury site. In this model, progenitor cells directly differentiate into osteoblasts and deposit a mineralized matrix.(30,38) Osteoclasts remodel the bony matrix until the shape of the tibia is restored. In a monocortical defect, bone regeneration occurs exclusively through intramembranous ossification without an intervening cartilage intermediate. Adjacent to the injury site, reparative cartilage does form but this endochondral ossification response is strictly limited to the periosteum. Because of the highly ordered manner in which bone regeneration occurs in this model system, we can precisely pinpoint when steps of the reparative process are disrupted. For example, we determined that in skeletally mature (10 12 wk old), male wildtype mice, newly regenerated osteoid matrix is first detectable in the injury site by postsurgical day 6 (n 15; Fig. 1H). We generated the same tibial defects in TOPgal and BATgal mice and examined the injury sites over the course of the next 7 days for evidence of Wnt reporter activity. We used both types of reporter mice to ensure that the pattern of Xgal staining, and therefore Wnt responsiveness, was equivalent in the two reporter strains. The early (e.g., days 1 and 2) injury callus is composed primarily of a fibrin clot, whose amorphous material properties complicate tissue sectioning and histological assessment. Therefore, instead of using standard immunohistochemical analyses, we isolated RNA from 24- and 48-h

injury calluses created in reporter mice and performed RTPCR for expression of the -galactosidase reporter gene. Within the first 24 h of skeletal injury, -galactosidase expression was dramatically increased over baseline, and it remained elevated in the 48-h injury callus (Fig. 1I and data not shown). These data indicate that Wnt signaling was initiated within the first 24 h after injury. By day 3 after injury, the callus becomes more organized and can be analyzed in tissue sections. We used Xgal staining on tissue sections and found that reporter activity was highly localized to the injury site (n 6; Fig. 1J). At this early stage of healing, the injury site is populated largely by bone marrowderived cells and inflammatory cells.(39) We performed immunohistochemistry for the neutrophil marker Gr-1 to determine whether the Wnt responsive cells were inflammatory in origin but found no specific colocalization of Xgal and Gr-1positive cells (Fig. 1K). These results suggested that the Wnt responsive cells were not inflammatory in origin. At later stages of bone repair, when inflammation ended, we were still able to detect Xgalpositive cells, again suggesting that Wnt reporter activity was not confined to inflammatory cells. We did, however, determine that multiple Wnt ligands, along with Wnt cofactors, receptors, and Wnt antagonists, were widely expressed in and near the injury site (Figs. 1L1N; Table 1), which implicated Wnt signaling in the process of bone healing. We also analyzed later time-points in the healing process of reporter activity and found Xgal staining for at least 28 days after injury.

Inhibiting Wnt signaling in vivo

Thus far, our data showed that components of the Wnt pathway are expressed in the right time and place to play a role in bone regeneration, because Wnt signaling is upregulated within 24 h of damage to the skeleton; and Wnt responsive cells are localized to the injury site (Fig. 1). Given these data, we tested whether perturbations in Wnt signaling affected the program of skeletal repair. To inhibit Wnt signaling in vivo, we injected adenovirus expressing the

1918 soluble Wnt antagonist Dkk1 (Ad-Dkk1), into TOPgal mice. Dkk 1 and 2 proteins both act as antagonists, but depending on the cellular context, Dkk2 can also act as an agonist.(40,41) Because of this complication, and because overexpression of Dkk1 has been used in a number of in vivo and in vitro contexts to produce a conditional, reversible inhibition in Wnt signaling in the adult animal,(35,42,43) we used the same strategy. Adenovirus expressing the Fc portion of mouse immunoglobulin gene (Ad-Fc) served as a control. Immediately after infection, skeletal defects were generated in the tibia of TOPgal mice. Two days later, we collected blood samples through retro-orbital phlebotomy and performed Western blot analyses to determine the levels of protein produced from the control Fc adenovirus and from the Dkk-1 adenovirus. In all animals tested, we found that high levels of the Fc fragment and Dkk-1 were produced from the adenoviral constructs (Figs. 2A and 2B, insets; each lane represents a separate animal). Whole mount Xgal staining showed that Ad-Dkk1 treatment effectively reduced Wnt signaling throughout the skeleton, most notably in the growth plate (Figs. 2A and 2B). Six days after the injuries were generated and viruses delivered, the tibias were collected. We used Xgal staining once again to show that Ad-Dkk1 treatment abolished reporter activity at the injury site (Figs. 2C and 2D). Thus, intravenous Ad-Dkk1, delivered at the time of injury, proved to be a useful method for reducing or eliminating the contribution of canonical Wnt signaling to the process of skeletal repair. We used histomorphometric measurements, performed on representative tissue sections throughout the injury sites, to assess how Ad-Dkk1 affected the process of bone regeneration. Aniline blue staining was used to identify new osteoid matrix within the injury site (Figs. 2E and 2F; dotted lines demarcate new bone formation). The amount of osteoid matrix was quantified (Figs. 2G and 2H). Using this quantitative approach, we found that Ad-Dkk1 treatment reduced bone regeneration by 84% (control Ad-Fc, n 7; Ad-Dkk1, n 10; Fig. 2I). When high levels of infection are achieved, Ad-Dkk1 can have deleterious systemic effects,(35) which we theorized might indirectly affect the reparative process by disrupting the metabolism of the mouse. Therefore, we also tested whether local delivery of Ad-Dkk1 blocked bone regeneration to the same extent as systemic Ad-Dkk1. In these cases, Ad-Dkk1 or Ad-Fc was injected into the musculature surrounding the tibias, and 48 h later, a skeletal defect was generated in this same location. Xgal staining of the injury site confirmed once again that Ad-Dkk1 effectively blocked Wnt-induced reporter activity (data not shown), and histological analyses confirmed that local Ad-Dkk1 blocked bone regeneration as effectively as systemic Ad-Dkk1 injection (Figs. 2J and 2K). Our next series of experiments were aimed at determining the mechanisms underlying the Dkk-1mediated arrest in bone regeneration. In previous studies, we have shown that a disruption in angiogenesis can delay bone regeneration.(44) When we evaluated endothelial cell invasion into the injury site, however, we did not detect any differences in

KIM ET AL. the extent or distribution of PECAM immunostaining (AdFc, n 4; Ad-Dkk1, n 4; Figs. 3A and 3B). Therefore, Ad-Dkk1 treatment did not seem to curtail injury-induced angiogenesis. We also used TUNEL staining to evaluate Ad-Dkk1 injury sites, reasoning that Wnt inhibition might lead to an increase in programmed cell death. Once again, Ad-Dkk1 and Ad-Fc tibias showed equivalent TUNEL staining (Ad-Fc, n 4; Ad-Dkk1, n 4; data not shown). Therefore, it was unlikely that Ad-Dkk1 hampered bone regeneration by increasing cell death in the injury site. Because Wnt signaling influences the commitment of progenitor cells to an osteogenic fate, we examined cells in the Ad-Dkk1 and Ad-Fc injury sites for evidence of osteoblast differentiation. On postsurgical day 4, cells in the AdFc injury site strongly expressed the osteogenic genes runx2 and collagen type I, indicating their imminent differentiation into osteoblasts.(4547) In Ad-Dkk1 injury sites, runx2 expression was essentially undetectable (Figs. 3C and 3D and data not shown). We also found that Ad-Fc injury sites showed robust alkaline phosphatase activity, whereas this same activity that characterizes osteoblast differentiation(4850) was nonexistent in Ad-Dkk1 injury sites (Figs. 3E and 3F). Together, these data showed that Dkk1 effectively blocked canonical Wnt signaling in the adult skeleton. As a consequence, bone marrowderived cells in the injury site failed to initiate runx2 expression. Mineralization was halted, and ultimately, the program of adult bone regeneration was arrested.

Constitutive Wnt activation increases skeletal progenitor cell proliferation but delays bone regeneration
Having shown that inhibition of Wnt signaling was detrimental to bone regeneration, we next explored whether activation of Wnt signaling would be advantageous to bone healing. A gain-of-function mutation in the Wnt coreceptor Lrp5 is associated with a high bone mass phenotype(10,15,5154); therefore, we evaluated bone regeneration in transgenic mice harboring the same mutation (i.e., Lrp5G171V mice). Skeletal defects were generated in Lrp5-G171V mice, and samples were analyzed on postsurgical day 6. The first and most obvious phenotype we found was a conspicuous absence of new osteoid matrix in the Lrp5 injury sites compared with wildtype controls (control, n 6 Lrp5-G171V, n 6; Figs. 4A and 4B). We confirmed this absence of new osteoid with Aniline blue staining (Figs. 4C and 4D). Activated Wnt signaling is oftentimes associated with increased cell proliferation(21,55,56); therefore, we used immunohistochemical detection of PCNA to examine the repair sites. In Lrp5-G171V mice, PCNA immunostaining was consistently elevated in the periosteum (n 6; Figs. 4E and 4F), the injury site, and throughout the bone marrow, relative the same locations in wildtype mice. Another possible explanation for the lack of bony regenerate in the Lrp5G171V injury sites would be a perturbation in boneresorbing osteoclast activity, but TRACP activity appeared to be the same in the mutant and wildtype samples (Figs. 4G and 4H).

BONE REGENERATION IS REGULATED BY WNT SIGNALING

1919

FIG. 2. Inhibition of Wnt signaling by Dkk1 arrests bone regeneration. TOPgal mice were used for this study, where skeletal injury was generated in the tibia immediately after the systemic administration of a control virus, Ad-Fc, and a virus expressing the soluble Wnt inhibitor, Dkk1. (A) Xgal staining shows that the pattern of -galactosidase reporter activity is unperturbed in the growth plates (gp) of the tibia (ti) and fibula (fi) after systemic delivery of Ad-Fc. (B) Systemic delivery of Ad-Dkk1 nearly abolishes -galactosidase activity in the growth plates of TOPgal mice. Insets: Western blots in multiple animals (numbered lanes) confirm that high levels of the Fc fragment and Dkk1 protein are achieved within 48 h of adenoviral delivery. (C) Xgal staining of tissue sections shows the characteristic pattern of Wnt-responsiveness in the injury site (is) after Ad-Fc injection. (D) The injury site after Ad-Dkk1 treatment is absent of Wntresponding cells. (E) Tissue sections from an Ad-Fc injury site on postsurgical day 6 showed evidence of new bone matrix, which stains blue (dotted line circumscribes the regenerate). (F) In the Ad-Dkk1 injury site, there is very little new bone (dotted line). (I) The amount of new bone was quantified using histomorphometric measurements, which shows that relative to (G) control, AdFc injury sites, (H) Ad-Dkk1 injury sites have an 84% drop in bone regeneration. Local delivery of Ad-Fc (J) and Ad-Dkk1 (K) results in a similar response, with normal bone formation in the control site and reduced osteogenesis in the site of Wnt inhibition. Scale bars in A and B: 1 mm; in C and D: 200 m; in EK: 300 m.

We also examined Lrp5-G171 mice for changes in alkaline phosphatase activity. In wildtype mice, alkaline phosphatase activity was localized to the injury site, indicating robust osteoblast differentiation in response to skeletal damage (Fig. 4I), but in Lrp5-G171V mice, alkaline phosphatase activity was restricted to adjacent, uninjured bone and was nonexistent in the injury site itself (Fig. 4J). In addition, the expression levels of the pro-osteogenic genes

sox9, runx2, and osteocalcin were all reduced in Lrp5 injury sites relative to wildtype littermates (Fig. 4). Together these data indicated that constitutively active Wnt signaling in Lrp5-G171V mice resulted in an increase in cell proliferation at the injury site, which ultimately delayed osteoblast differentiation. These observations are similar to previously reported in vitro findings(21); however, our in vivo model system allowed us to examine the longer-term effects of

1920

KIM ET AL.

FIG. 3. Bone regeneration is halted because of a Dkk1-mediated arrest in osteoblast differentiation. (A) On postsurgical day 4, PECAM immunopositive endothelial cells fill the Ad-Fc injury site (is), showing the typical angiogenic response after skeletal damage. The cut edge of the cortical bone (cb) is circumscribed by a dotted line in all panels. (B) Ad-Dkk1 treatment does not alter PECAM staining to a notable degree. (C) Cells in the Ad-Fc injury site begin to differentiate into osteoblasts, as indicated by the widespread expression of runx2. (D) Runx2 expression is undetectable in the Ad-Dkk1 injury site. (E) Ad-Fc injury sites show robust alkaline phosphatase activity, indicating the onset of mineralization of the osteoid matrix. (F) There is no evidence of alkaline phosphatase activity in the injury site of AdDkk1treated animals. Scale bar in AF: 100 m.

activated Wnt signaling, and these findings were unexpected: When we examined Lrp5-G171V mice on postsurgical day 14, the amount of bone regeneration was equivalent between Lrp5-G171V mice and wildtype mice (data not shown). Thus, constitutive activation of the Wnt pathway caused an initial delay in bone regeneration, by temporarily inhibiting the differentiation of cells into osteoblasts. However, this postponement was only transient; eventually, the effect of activated Wnt signaling in Lrp5G171V mice resulted in robust bone regeneration.

DISCUSSION
Our objective was to gain insights into the roles of Wnt signaling in adult bone formation and remodeling. Our data indicate that damage to the skeleton triggers a reparative response, and one pathway involved in this response is mediated by Wnts (Fig. 1; Table 1). In response to skeletal damage, progenitor cells at the injury site begin to proliferate. Our data revealed that at least one of the pathways involved in this injury-induced proliferation is mediated by Wnt signaling. Proliferation is reduced when Wnt signaling is blocked, and proliferation is increased when Wnt signaling is constitutively active. Once a sufficient cell population has been generated,

skeletal progenitors begin to differentiate into osteoblasts. Although we do not have clues into how this transition between proliferation and differentiation occurs, our data showed that Wnt signaling is a critical mediator of osteoblast differentiation in an injury site. All skeletal progenitor cells express sox9, and those that are destined for an osteogenic lineage express runx2 as well.(5,17,57) Our in vivo study showed that Wnt signaling is required for runx2 expression in the injury site and that in the absence of Wnt signaling, bone regeneration is disrupted. We recognize that the Wnt-dependent proliferative effect might be used to an advantage in a therapeutic effort to stimulate bone formation. The molecular machinery for Wnt signaling is functional in the very early injury site, which suggests that transient exposure of bone marrow derived cells to a Wnt protein may have a pro-osteogenic effect. In other reparative contexts, Wnt proteins stimulate stem cell self-renewal.(58,59) We are interested in determining whether Wnt proteins may be able to act in a similar fashion, to induce marrow-derived stem cell proliferation and self-renewal, during bone regeneration and thus hasten the repair of skeletal injuries. Another question that remains to be addressed is whether Wnt proteins can be used to enhance the inherent osteogenic potential of bone marrow. When the preexisting

BONE REGENERATION IS REGULATED BY WNT SIGNALING

1921

FIG. 4. Constitutive Wnt activation stimulates cell proliferation but delays bone regeneration. (A) Pentachrome staining of a day 6 injury site (is) in a wildtype animal shows cortical bridging of newly deposited bone matrix (blue-green). (B) In the LRP5-G171V mutant animals, the defect is lacking new bone matrix, note the high bone mass phenotype in the periphery with intramedullary trabeculae. (C) Aniline blue staining labels the newly deposited osteoid in the injury site on postsurgical day 6 in wildtype animals. (D) Identical injuries in LRP5-G171V mice exhibit less new bone on day 6. (E) Only few PCNA immunopositive cells are present in wildtype the injury site. (F) The injury site of LRP5-G171V mice is occupied by multiple proliferating cells. (G) TRACP staining in wildtype mice at day 6 shows abundant osteoclastic activity in the area of newly deposited bone. (H) The LRP5 mutant injury site exhibits modest TRACP activity in the vicinity of the defect. (I) Alkaline phosphatase activity is exuberant in the wildtype injury site, (J) but almost absent in the LRP5-G171V mice. The inset shows alkaline phosphatase activity in an uninjured LRP mutant tibia. (K) In situ hybridization for osteochondroprogenitor genes reveals moderate levels of sox9 in the wildtype injury site but (L) undetectable expression in the LRP5-G171V injury. (M) Runx2 is also expressed in wildtype injury sites but (N) runx2 is expressed at nearly undetectable levels in the LRP5-G171V injury site. (O) Wildtype cells express the osteoblast marker osteocalcin, whereas (P) in the LRP5-G171V injury site, osteocalcin expression is very low. Scale bar in AD: 300 m; in EP: 200 m.

bone stock is inadequate, cancellous bone containing marrow is frequently harvested and mixed with bone substitutes (i.e., demineralized bone matrix) in an attempt to stimulate bone formation. This clinical procedure is predicated on the assumption that bone marrow contains endogenous factors that stimulate osteogenesis in the host.(60,61) Our data suggest that Wnt signaling is required for cell proliferation within marrow-derived cells; therefore, pretreatment of the bone marrow with purified Wnt proteins may be able to stimulate the osteogenic capacity of a patients bone marrow. This hypothesis remains to be tested, but the recent ability to purify Wnt proteins(58) will undoubtedly facilitate such studies. In summary, our study provides a foundation for understanding the program of adult tissue regeneration, using the skeleton as a model system. When we understand how to control the choice between cell self-renewal, proliferation, and differentiation, theoretically we will have the ability to expand what might be a limited population of adult progenitor cells, induce their timely differentiation, and thus restore the function of a tissue or organ. Our data suggest that the Wnt pathway is among the most attractive targets for such therapeutic interventions to treat skeletal diseases and injuries.

ACKNOWLEDGMENTS
This work was supported by funds from the NIH R01 DE012462, FA9550-04-1-0075 (JAH), and RO1 AR45989 (JBK), the Howard Hughes Medical Institution (RN), and the Northern California Arthritis Foundation (JBK).

REFERENCES
1. Clevers H 2004 Wnt signaling: Ig-norrin the dogma. Curr Biol 14:R436R437. 2. Nelson WJ, Nusse R 2004 Convergence of Wnt, beta-catenin, and cadherin pathways. Science 303:14831487. 3. Gordon MD, Nusse R 2006 Wnt signaling: Multiple pathways, multiple receptors, and multiple transcription factors. J Biol Chem 281:2242922433. 4. Gregory CA, Gunn WG, Reyes E, Smolarz AJ, Munoz J, Spees JL, Prockop DJ 2005 How Wnt signaling affects bone repair by mesenchymal stem cells from the bone marrow. Ann NY Acad Sci 1049:97106. 5. Akiyama Ddagger H, Kim Ddagger JE, Nakashima K, Balmes G, Iwai N, Deng JM, Zhang Z, Martin JF, Behringer RR, Nakamura T, de Crombrugghe B 2005 Osteo-chondroprogenitor cells are derived from Sox9 expressing precursors. Proc Natl Acad Sci USA 102:1466514670. 6. Akiyama H, Lyons JP, Mori-Akiyama Y, Yang X, Zhang R, Zhang Z, Deng JM, Taketo MM, Nakamura T, Behringer RR,

1922
McCrea PD, de Crombrugghe B 2004 Interactions between Sox9 and beta-catenin control chondrocyte differentiation. Genes Dev 18:10721087. Mori-Akiyama Y, Akiyama H, Rowitch DH, de Crombrugghe B 2003 Sox9 is required for determination of the chondrogenic cell lineage in the cranial neural crest. Proc Natl Acad Sci USA 100:93609365. Bennett CN, Longo KA, Wright WS, Suva LJ, Lane TF, Hankenson KD, MacDougald OA 2005 Regulation of osteoblastogenesis and bone mass by Wnt10b. Proc Natl Acad Sci USA 102:33243329. Luo Q, Kang Q, Si W, Jiang W, Park JK, Peng Y, Li X, Luu HH, Luo J, Montag AG, Haydon RC, He TC 2004 Connective tissue growth factor (CTGF) is regulated by Wnt and BMP signaling in osteoblast differentiation of mesenchymal stem cells. J Biol Chem 279:5595855968. Westendorf JJ, Kahler RA, Schroeder TM 2004 Wnt signaling in osteoblasts and bone diseases. Gene 341:1939. De Boer J, Wang HJ, Van Blitterswijk C 2004 Effects of Wnt signaling on proliferation and differentiation of human mesenchymal stem cells. Tissue Eng 10:393401. Bain G, Muller T, Wang X, Papkoff J 2003 Activated betacatenin induces osteoblast differentiation of C3H10T1/2 cells and participates in BMP2 mediated signal transduction. Biochem Biophys Res Commun 301:8491. Kahler RA, Westendorf JJ 2003 Lymphoid enhancer factor-1 and beta-catenin inhibit Runx2-dependent transcriptional activation of the osteocalcin promoter. J Biol Chem 278:11937 11944. 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 osteoblast proliferation, osteopenia, and persistent embryonic eye vascularization in mice deficient in Lrp5, a Wnt coreceptor. J Cell Biol 157:303314. Babij P, Zhao W, Small C, Kharode Y, Yaworsky PJ, Bouxsein ML, Reddy PS, Bodine PV, Robinson JA, Bhat B, Marzolf J, Moran RA, Bex F 2003 High bone mass in mice expressing a mutant LRP5 gene. J Bone Miner Res 18:960974. 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 8:739750. Rodda SJ, McMahon AP 2006 Distinct roles for Hedgehog and canonical Wnt signaling in specification, differentiation and maintenance of osteoblast progenitors. Development 133:32313244. Hartmann C, Tabin CJ 2000 Dual roles of Wnt signaling during chondrogenesis in the chicken limb. Development 127:3141 3159. Daumer KM, Tufan AC, Tuan RS 2004 Long-term in vitro analysis of limb cartilage development: Involvement of Wnt signaling. J Cell Biochem 93:526541. Tufan AC, Daumer KM, Tuan RS 2002 Frizzled-7 and limb mesenchymal chondrogenesis: Effect of misexpression and involvement of N-cadherin. Dev Dyn 223:241253. Boland GM, Perkins G, Hall DJ, Tuan RS 2004 Wnt 3a promotes proliferation and suppresses osteogenic differentiation of adult human mesenchymal stem cells. J Cell Biochem 93:12101230. Tuli R, Tuli S, Nandi S, Huang X, Manner PA, Hozack WJ, Danielson KG, Hall DJ, Tuan RS 2003 Transforming growth factor-beta-mediated chondrogenesis of human mesenchymal progenitor cells involves N-cadherin and mitogen-activated protein kinase and Wnt signaling cross-talk. J Biol Chem 278:4122741236. Tuan RS 2003 Cellular signaling in developmental chondrogenesis: N-cadherin, Wnts, and BMP-2. J Bone Joint Surg Am 85(Suppl 2):137141. Hartmann C 2006 A Wnt canon orchestrating osteoblastogenesis. Trends Cell Biol 16:151158. Cheung WM, Jin LY, Smith DK, Cheung PT, Kwan EY, Low L, Kung AW 2006 A family with osteoporosis pseudoglioma

KIM ET AL.
syndrome due to compound heterozygosity of two novel mutations in the LRP5 gene. Bone 39:470476. Kwee ML, Balemans W, Cleiren E, Gille JJ, Van Der Blij F, Sepers JM, Van Hul W 2005 An autosomal dominant high bone mass phenotype in association with craniosynostosis in an extended family is caused by an LRP5 missense mutation. J Bone Miner Res 20:12541260. Henriksen K, Gram J, Hoegh-Andersen P, Jemtland R, Ueland T, Dziegiel MH, Schaller S, Bollerslev J, Karsdal MA 2005 Osteoclasts from patients with autosomal dominant osteopetrosis type I caused by a T253I mutation in low-density lipoprotein receptor-related protein 5 are normal in vitro, but have decreased resorption capacity in vivo. Am J Pathol 167:13411348. Van Wesenbeeck L, Cleiren E, Gram J, Beals RK, Benichou O, Scopelliti D, Key L, Renton T, Bartels C, Gong Y, Warman ML, De Vernejoul MC, Bollerslev J, Van Hul W 2003 Six novel missense mutations in the LDL receptor-related protein 5 (LRP5) gene in different conditions with an increased bone density. Am J Hum Genet 72:763771. Sezer O 2005 Myeloma bone disease. Hematology 10(Suppl 1):1924. Colnot C, Romero DM, Huang S, Helms JA 2005 Mechanisms of action of demineralized bone matrix in the repair of cortical bone defects. Clin Orthop Relat Res 435:6978. Maretto S, Cordenonsi M, Dupont S, Braghetta P, Broccoli V, Hassan AB, Volpin D, Bressan GM, Piccolo S 2003 Mapping Wnt/beta-catenin signaling during mouse development and in colorectal tumors. Proc Natl Acad Sci USA 100:32993304. DasGupta R, Fuchs E 1999 Multiple roles for activated LEF/ TCF transcription complexes during hair follicle development and differentiation. Development 126:45574568. Albrecht UEG, Helms JA, Lin H 1997 Visualization of gene expression patterns by in situ hybridization. In: Daston GP (ed.) Molecular and Cellular Methods in Developmental Toxicology. CRC Press, Boca Raton, FL, USA, pp. 2348. Sheehan DC, Hrapchak BB 1980 Theory and Practice of Histotechnology, 2nd ed. Batelle Press, Columbus, OH, USA. Kuhnert F, Davis CR, Wang HT, Chu P, Lee M, Yuan J, Nusse R, Kuo CJ 2004 Essential requirement for Wnt signaling in proliferation of adult small intestine and colon revealed by adenoviral expression of Dickkopf-1. Proc Natl Acad Sci USA 101:266271. Krishnan V, Bryant HU, Macdougald OA 2006 Regulation of bone mass by Wnt signaling. J Clin Invest 116:12021209. Andrade AC, Nilsson O, Barnes KM, Baron J 2007 Wnt gene expression in the post-natal growth plate: Regulation with chondrocyte differentiation. Bone 40:13611369. Thompson Z, Miclau T, Hu D, Helms JA 2002 A model for intramembranous ossification during fracture healing. J Orthop Res 20:10911098. Leucht P, Lam K, Kim JB, Mackanos M, Simanovskii DM, Longaker MT, Contag C, Schwettman HA, Helms JA 2007 Accelerated bone repair after plasma laser corticotomies. Ann Surg 246:140150. Mao B, Wu W, Davidson G, Marhold J, Li M, Mechler BM, Delius H, Hoppe D, Stannek P, Walter C, Glinka A, Niehrs C 2002 Kremen proteins are Dickkopf receptors that regulate Wnt/beta-catenin signalling. Nature 417:664667. Mao B, Niehrs C 2003 Kremen2 modulates Dickkopf2 activity during Wnt/LRP6 signaling. Gene 302:179183. Osakada F, Ooto S, Akagi T, Mandai M, Akaike A, Takahashi M 2007 Wnt signaling promotes regeneration in the retina of adult mammals. J Neurosci 27:42104219. Ito M, Yang Z, Andl T, Cui C, Kim N, Millar SE, Cotsarelis G 2007 Wnt-dependent de novo hair follicle regeneration in adult mouse skin after wounding. Nature 447:316320. Colnot C, Thompson Z, Miclau T, Werb Z, Helms JA 2003 Altered fracture repair in the absence of MMP9. Development 130:41234133. Karsenty G 2001 Minireview: Transcriptional control of osteoblast differentiation. Endocrinology 142:27312733.

26.

7.

8.

27.

9.

28.

10. 11. 12.

29. 30.

13.

31.

14.

32.

33.

15.

34. 35.

16.

17.

36. 37.

18. 19. 20. 21.

38.

39.

40.

22.

41. 42.

43.

23.

44.

24. 25.

45.

BONE REGENERATION IS REGULATED BY WNT SIGNALING

46. Komori T 2002 Runx2, a multifunctional transcription factor in skeletal development. J Cell Biochem 87:18. 47. Dacic S, Kalajzic I, Visnjic D, Lichtler AC, Rowe DW 2001 Col1a1-driven transgenic markers of osteoblast lineage progression. J Bone Miner Res 16:12281236. 48. Komori T, Yagi H, Nomura S, Yamaguchi A, Sasaki K, Deguchi K, Shimizu Y, Bronson RT, Gao YH, Inada M, Sato M, Okamoto R, Kitamura Y, Yoshiki S, Kishimoto T 1997 Targeted disruption of Cbfa1 results in a complete lack of bone formation owing to maturational arrest of osteoblasts. Cell 89:755764. 49. Otto F, Thornell AP, Crompton T, Denzel A, Gilmour KC, Rosewell IR, Stamp GW, Beddington RS, Mundlos S, Olsen BR, Selby PB, Owen MJ 1997 Cbfa1, a candidate gene for cleidocranial dysplasia syndrome, is essential for osteoblast differentiation and bone development. Cell 89:765771. 50. Farley JR, Baylink DJ 1986 Skeletal alkaline phosphatase activity as a bone formation index in vitro. Metabolism 35:563 571. 51. Tamai K, Semenov M, Kato Y, Spokony R, Liu C, Katsuyama Y, Hess F, Saint-Jeannet JP, He X 2000 LDL-receptor-related proteins in Wnt signal transduction. Nature 407:530535. 52. 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 107:513523. 53. Semenov MV, He X 2006 LRP5 mutations linked to high bone mass diseases cause reduced LRP5 binding and inhibition by SOST. J Biol Chem. 54. Johnson ML, Harnish K, Nusse R, Van Hul W 2004 LRP5 and Wnt signaling: A union made for bone. J Bone Miner Res 19:17491757. 55. Masckauchan TN, Shawber CJ, Funahashi Y, Li CM, Kitajew-

1923

56.

57.

58.

59.

60.

61.

ski J 2005 Wnt/beta-catenin signaling induces proliferation, survival and interleukin-8 in human endothelial cells. Angiogenesis 8:4351. Hu H, Hilton MJ, Tu X, Yu K, Ornitz DM, Long F 2005 Sequential roles of Hedgehog and Wnt signaling in osteoblast development. Development 132:4960. 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. Willert K, Brown JD, Danenberg E, Duncan AW, Weissman IL, Reya T, Yates JR III, Nusse R 2003 Wnt proteins are lipid-modified and can act as stem cell growth factors. Nature 423:448452. Reya T, Duncan AW, Ailles L, Domen J, Scherer DC, Willert K, Hintz L, Nusse R, Weissman IL 2003 A role for Wnt signalling in self-renewal of haematopoietic stem cells. Nature 423:409414. Colter DC, Sekiya I, Prockop DJ 2001 Identification of a subpopulation of rapidly self-renewing and multipotential adult stem cells in colonies of human marrow stromal cells. Proc Natl Acad Sci USA 98:78417845. Horwitz EM, Prockop DJ, Fitzpatrick LA, Koo WW, Gordon PL, Neel M, Sussman M, Orchard P, Marx JC, Pyeritz RE, Brenner MK 1999 Transplantability and therapeutic effects of bone marrow-derived mesenchymal cells in children with osteogenesis imperfecta. Nat Med 5:309313.

Address reprint requests to: Jill A Helms, DDS, PhD Department of Plastic and Reconstructive Surgery Stanford University 257 Campus Drive Stanford, CA 94305-5148, USA E-mail: [email protected]
Received in original form January 30, 2007; revised form June 12, 2007; accepted August 3, 2007.