Research ArticleBone

Wnt Proteins Promote Bone Regeneration

See allHide authors and affiliations

Science Translational Medicine  28 Apr 2010:
Vol. 2, Issue 29, pp. 29ra30
DOI: 10.1126/scitranslmed.3000231


The Wnt signaling pathway plays a central role in bone development and homeostasis. In most cases, Wnt ligands promote bone growth, which has led to speculation that Wnt factors could be used to stimulate bone healing. We gained insights into the mechanism by which Wnt signaling regulates adult bone repair through the use of the mouse strain Axin2LacZ/LacZ in which the cellular response to Wnt is increased. We found that bone healing after injury is accelerated in Axin2LacZ/LacZ mice, a consequence of more robust proliferation and earlier differentiation of skeletal stem and progenitor cells. In parallel, we devised a biochemical strategy to increase the duration and strength of Wnt signaling at the sites of skeletal injury. Purified Wnt3a was packaged in liposomal vesicles and delivered to skeletal defects, where it stimulated the proliferation of skeletal progenitor cells and accelerated their differentiation into osteoblasts, cells responsible for bone growth. The end result was faster bone regeneration. Because Wnt signaling is conserved in mammalian tissue repair, this protein-based approach may have widespread applications in regenerative medicine.


In the skeleton, adult skeletal stem cells serve two essential functions. First, they are responsible for generating osteocytes that maintain bone mass and bone volume during physiologic bone turnover (1). Second, they are essential for replacing missing bone after skeletal disease or injury (2). In both scenarios, Wnt signals regulate skeletal stem and progenitor cell behavior and control bone formation. Wnt proteins are lipid-modified growth factors that are secreted from one cell and transported to adjacent cells, perhaps in association with membranes (3, 4). Upon binding to a cell surface receptor complex, Wnt signals initiate an intracellular cascade that results in tissue-specific activation of target gene transcription.

Reductions in Wnt signaling, by overexpression of Wnt antagonists, deficiency of Wnt ligands, or mutations in genes encoding Wnt receptors, cause bone loss (57). Conversely, enhancement of Wnt signaling results in increased bone volume, abnormal bone density (that is, hyperostosis), and pathological thickening of bone, a condition known as sclerosing bone dysplasia (811). Wnt signals achieve these effects by promoting bone-forming osteoblast activity (12, 13), by inhibiting bone-resorbing osteoclast function (14, 15), and by driving the differentiation of many types of multipotent stem cells toward an osteoblast cell fate (1618). These data suggest that Wnt proteins could be used as therapeutic agents to stimulate bone formation after injury or disease. The use of Wnts for bone regeneration, however, has not been directly tested because of practical difficulties in administering the highly hydrophobic and insoluble Wnt proteins. Soluble agonists of the Wnt pathway have been identified, but none act in the absence of Wnt proteins (1922).

There is another difficulty that needs to be resolved before Wnts can be used as therapeutic agents: Wnt signals regulate multiple phases of the skeletogenic program ranging from proliferation and commitment of skeletal progenitor cells to the maturation of chondrocytes and osteoblasts (2). Accordingly, the optimal time for Wnt pathway stimulation requires a detailed understanding of the function of endogenous Wnt signaling during the entire reparative process. For example, we demonstrated that skeletal injury triggers a rapid increase in Wnt signaling at the site of damage (5) and that Wnt signals are necessary for bone healing (5, 23), but whether exogenous Wnt signals are sufficient to promote bone healing is not known.


Mutating a negative feedback regulator potentiates Wnt signaling in a ligand-dependent manner

To determine how Wnt signals influence bone homeostasis, we used a strain of mice in which LacZ was inserted into the Axin2 gene (24). Axin2, a target of Wnt signaling (24), encodes a ligand-dependent, negative feedback regulator that dampens Wnt signaling in a cell-autonomous manner. The LacZ insert inactivates Axin2 (24), and thus, homozygous mutant mice (Axin2LacZ/LacZ) provide a genetic model in which to assess how loss of a negative Wnt regulator affects bone homeostasis and bone repair.

We first evaluated how the Axin2 mutation influenced Wnt signaling in osteoprogenitor cells in culture. To measure Wnt pathway activity, we used a T cell factor (Tcf)–dependent luciferase reporter construct, which was infected (through a lentiviral vector) into primary osteoprogenitor cells isolated from Axin2LacZ/LacZ, Axin2LacZ/+, and wild-type mice. In the unstimulated state, Axin2LacZ/LacZ cells had the same baseline luciferase reporter activity as wild-type and Axin2LacZ/+ cells (Fig. 1A). After brief exposure to Wnt protein, Axin2LacZ/LacZ cells showed higher luciferase reporter activity than either wild-type or Axin2LacZ/+ cells (Fig. 1A). The duration of Wnt signaling was also extended in Axin2LacZ/LacZ cells: 70 hours after withdrawing Wnt protein, wild-type and Axin2LacZ/+ cells had returned to near-baseline responsiveness, whereas Axin2LacZ/LacZ cells maintained ~40% of their maximum Wnt responsiveness [Fig. 1A; see also (25)]. Thus, Axin2 repressed Wnt signaling when cells had been exposed to Wnt ligands, and removing the Axin2 gene resulted in an amplified Wnt response in these cells.

Fig. 1

Wnt signaling is amplified in Axin2LacZ/LacZ mice. (A) In vitro analysis of wild-type (WT), Axin2LacZ/+, and Axin2LacZ/LacZ preosteoblasts transduced with a 7xTcf-Luc Wnt-activated reporter construct. Luciferase activity was monitored for 70 hours. Cells were exposed to Wnt3a protein from 0 to 4 hours (green arrow), after which time the Wnt was removed and medium was replaced (red arrow). Data are presented as the mean ± SEM. Lines represent modeled exponential decay from peak Wnt responsiveness at t = 25 hours, shown in red for Axin2LacZ/LacZ, blue for Axin2LacZ/+, and green for WT cells. (B) μCT reconstruction to detect structural differences in WT and Axin2LacZ/LacZ bones. (C) Bone volume fraction (BVF) (n = 5) and bone mineral density (BMD) (n = 5) calculated from μCT data ± SEM (fig. S2); ROI denotes entire skeletal element (tibia or femur). (D and E) WT (D) and Axin2LacZ/LacZ (E) cortical bones stained with pentachrome. cb, cortical bone; en, endosteum; po, periosteum. (F) ALP activity in femurs and tibiae of WT mice is significantly lower than in femurs and tibiae of Axin2LacZ/LacZ mice (n = 6 for each genotype; P = 0.017 for femurs, P = 0.0043 for tibiae). (G and H) Distribution of ALP activity in bones from WT (n = 4) (G) and Axin2LacZ/LacZ (n = 3) (H) mice. (I) TRAP activity in femurs and tibiae from WT mice is not significantly different than in femurs and tibiae of Axin2LacZ/LacZ mice (n = 6 for each genotype). (J and K) Distribution of TRAP activity in bones from WT (n = 3) (J) and Axin2LacZ/LacZ (n = 9) (K) mice. Scale bar, 100 μm.

Axin2 deletion does not cause pathologic bone accrual

Despite the increased response of Axin2LacZ/LacZ osteoprogenitor cells to an acute Wnt exposure, Axin2LacZ/LacZ mice had bones with a normal cortical thickness. Three-dimensional reconstruction with micro–computed tomography (μCT) demonstrated that wild-type and Axin2LacZ/LacZ mice had equivalent bone area and bone volume and equivalent bone mineral density (n = 5 for each condition; Fig. 1, B and C, and fig. S2). Histological analyses verified that bones from wild-type and Axin2LacZ/LacZ mice had similar cortical anatomy (Fig. 1, D and E, n ≥ 12 for each genotype). Notwithstanding this equivalence, Axin2LacZ/LacZ mice exhibited an increase in alkaline phosphatase (ALP) activity. Within the osteogenic cell lineage, ALP activity is a marker of osteoprogenitor cell differentiation (26, 27), and we found elevated ALP activity both in bone extracts (Fig. 1F) and on the endosteal bone surfaces of Axin2LacZ/LacZ mice relative to wild-type controls (Fig. 1, G and H).

Using tartrate-resistant acid phosphatase (TRAP) activity as a marker for bone-resorbing osteoclast activity (28), we found no differences in either the amount of TRAP activity (Fig. 1I) or the distribution (Fig. 1, J and K) of osteoclasts in wild-type or Axin2LacZ/LacZ bones. These data indicated that osteoprogenitor cells from Axin2LacZ/LacZ mice showed an earlier commitment toward the osteoblastic phenotype (29) when they were compared to osteoprogenitor cells from wild-type mice.

A prolonged Wnt response in Axin2 mutant mice results in accelerated skeletal healing

The normal bone morphology of the Axin2LacZ/LacZ mice stands out against the pathologically thickened bones of mice carrying activating mutations in other components of Wnt signaling (9). This normal bone morphology may be attributable to functional compensation of Axin2 by its homolog Axin (30), but it could also mean that Axin2LacZ/LacZ cells behave the same as wild-type cells in the absence of an acute Wnt stimulus (for example, Fig. 1A). One method to induce an acute Wnt stimulus in vivo is to injure the skeleton (5). Using Western blot analysis, we showed that both activated and total β-catenin were expressed at higher levels in injured bone relative to intact bone (Fig. 2A). We verified that the Wnt pathway was elevated in response to injury using quantitative reverse transcription polymerase chain reaction (qRT-PCR): The Wnt target Axin2 was not significantly up-regulated by injury in wild-type mice (Fig. 2, B and C), but injury caused a significant increase in messenger RNA (mRNA) transcripts for Axin2 (exon 1) in Axin2LacZ/LacZ mice (Fig. 2B). Using X-galactosidase (X-Gal) staining, we found that both the distribution and the number of X-Gal–positive cells in Axin2LacZ/LacZ mice were increased after injury (Fig. 2, D and E). This injury-induced increase in Wnt signaling was restricted to the area of damage; we detected no obvious changes in Wnt signaling at sites distant to the injury (fig. S3).

Fig. 2

Injury increases Wnt signaling and skeletal progenitor cell proliferation in Axin2LacZ/LacZ mice. (A) Immunoblot for total β-catenin and activated β-catenin protein, with quantification of bands shown normalized against β-actin protein. (B) Validation of injury-induced Wnt pathway up-regulation using qRT-PCR for exon 1 of the Axin2 gene. Injury sites in Axin2LacZ/LacZ mice show a significant increase in mRNA transcripts for Axin2 (exon 1) compared to sham-treated controls (#) (n = 6; P < 0.05) and to WT mice (*). (C) Representative gel from RT-PCR analyses using primers to detect Axin2, Axin2 (exon 1), and β-actin in Axin2LacZ/LacZ and WT intact and injured bones. (D and E) X-Gal staining of the endosteum of Axin2LacZ/LacZ intact bones (D) and injured bones (n = 6 for each condition) (E). (F and G) PCNA immunostaining identifies proliferating cells in the injury sites (is) of WT (n = 4) (F) and Axin2LacZ/LacZ (n = 5) (G) mice. (H) Quantification of cell proliferation in injury sites of WT and Axin2LacZ/LacZ mice. Data represent mean ± SD (P < 0.01). Scale bar, 100 μm.

This region of increased Wnt signaling coincided with enhanced cell proliferation. Using 5-bromo-2′-deoxyuridine (BrdU) incorporation and immunostaining for the proliferation markers proliferating cell nuclear antigen (PCNA) and Ki67 (31), we found more dividing cells in injury sites than in intact bones (fig. S4). In addition, we found more dividing cells in the Axin2LacZ/LacZ injury sites than in wild-type sites (Fig. 2, F to H, and fig. S4). The proliferating cells were largely restricted to the injury site (fig. S4).

Bone healing was accelerated in Axin2LacZ/LacZ mice. Compared to wild-type mice, the osteoblast markers Runx2, Collagen type I, and Osteocalcin were expressed sooner in the injury sites of Axin2LacZ/LacZ mice (Fig. 3, A to F). Relative to wild-type controls, ALP activity was also detectable at an earlier time point in injury sites of Axin2LacZ/LacZ mice (Fig. 3, G and H). By day 7, Axin2LacZ/LacZ injury sites were filled with new bone regenerate, far in advance of new bone generation in injuries of wild-type bones (Fig. 3, I and J). We found no differences in the pattern or level of TRAP activity within the injury sites of Axin2LacZ/LacZ and wild-type mice (fig. S5). The regenerative advantage conferred by the Axin2LacZ/LacZ mutation was maintained for at least 2 weeks (Fig. 3K), and the appearance of the Axin2LacZ/LacZ bony regenerate was indistinguishable from the bony regenerate in wild-type mice at later stages of healing (fig. S6).

Fig. 3

Loss of Axin2 accelerates bone regeneration. (A to F) Three days after injury, Runx2 expression is evaluated by in situ hybridization in the injury sites of WT (n = 6) (A) and Axin2LacZ/LacZ (n = 6) (B) mice. Collagen type I expression in the injury sites of WT (n = 5) (C) and Axin2LacZ/LacZ (n = 6) (D) mice. Osteocalcin expression in the injury sites of WT (n = 3) (E) and Axin2LacZ/LacZ (n = 6) (F) mice. (G to J) After 7 days, ALP activity marks the onset of osteoprogenitor cell differentiation in the injury sites of WT (n = 3) (G) and Axin2LacZ/LacZ (n = 6) (H) mice. After 7 days, aniline blue staining indicates new osteoid matrix in WT (n = 3) (I) and Axin2LacZ/LacZ (n = 6) (J) mice. (K) Histomorphometric measurements are used to quantify new bone formation on day 7 (P < 0.05) and day 14 (WT, n = 6; Axin2LacZ/LacZ, n = 14; P < 0.01). Scale bars, 100 μm.

Cumulatively, our data demonstrated that injury transiently activated endogenous Wnt signaling and that Axin2LacZ/LacZ cells responded more robustly to this Wnt stimulus (Fig. 1). In the injury site, Axin2LacZ/LacZ cells showed enhanced proliferation (Fig. 2) and also adopted an osteoblastic fate sooner than wild-type cells, which resulted in a robust regenerative response and faster skeletal healing (Fig. 3). This regenerative response was self-limiting: Within 3 weeks of injury, endogenous Wnt signaling gradually returned to a preinjury level (fig. S7). Collectively, our data raised the possibility that early, transient exposure to additional Wnt signals could enhance cell proliferation and accelerate bone healing after skeletal damage.

Liposomal Wnt enhances bone tissue regeneration

To test whether an exogenous Wnt stimulus could promote bone regeneration, we packaged the purified Wnt protein into liposomal vesicles (32) and delivered them, or control (PBS) liposomes, to the skeletal injury site by injection.

Skeletal defects treated with liposomal Wnt healed faster. We found that the onset of mineralization (Fig. 4, A and B) and new osteoid deposition (Fig. 4, C and D) was accelerated by liposomal Wnt treatment. Histomorphometric measurements demonstrated that within 3 days of a single treatment, injury sites treated with liposomal Wnt had 3.5 times as much new bone as injury sites treated with PBS (Fig. 4E). We compared the pro-osteogenic effects of liposomal Wnt with bare Wnt protein and found that only liposomal Wnt effectively enhanced bone regeneration: Defects treated with bare Wnt showed the same amount of bone regeneration as those treated with PBS (Fig. 4, F to H). Liposomal Wnt-treated injuries also underwent the final stages of remodeling sooner than PBS-treated defects, resulting in more organized, mature bone matrix at earlier time points (Fig. 4, I and J).

Fig. 4

Liposomal Wnt treatment accelerates bone regeneration. Skeletal defects were generated in WT CD1 mice and treated with PBS liposomes (10 μl) of liposomal Wnt3a (10-μl volume with an effective Wnt3a concentration of 0.5 μg/ml). (A and B) After 2 days, ALP activity detects the onset of osteoprogenitor cell commitment to an osteoblastic lineage in PBS (A) and liposomal Wnt3a (L-Wnt3a)–treated (B) injuries. (C and D) After 3 days, aniline blue staining identifies new osteoid matrix in injury sites (is) treated with PBS (n = 5) (C) and liposomal Wnt3a (D) (n = 6). (E) Histomorphometric measurements (see Materials and Methods) demonstrate a factor of 3.5 increase in new bone in liposomal Wnt3a–treated sites. (F to H) After 3 days, pentachrome staining identifies new bone matrix (greenish yellow) in injury sites treated with liposomal PBS (L-PBS) (n = 5) (F), liposomal Wnt3a (n = 6) (G), or bare Wnt3a protein (n = 4) (H). (I and J) After 28 days, aniline blue staining illustrates the remodeled bone matrix in injuries treated with PBS (n = 5) (I) and liposomal Wnt3a (J). Boxed areas indicate former injury site. Scale bar, 100 μm.

We gained mechanistic insights into the basis for the pro-osteogenic effects of liposomal Wnt by carrying out a series of molecular and cellular analyses at earlier time points during the healing process. In the first 24 hours after injection, we saw no apparent histological differences between injuries treated with liposomal PBS and those treated with liposomal Wnt (Fig. 5, A and B), but by using X-Gal staining, we detected an obvious increase in the number of Wnt-responding cells after liposomal Wnt treatment (Fig. 5, C and D). Liposomal Wnt-treated sites also had significantly more proliferating cells than those treated with liposomal PBS (n = 6 for each condition; Fig. 5, E and F, and fig. S8). Cells in injury sites treated with liposomal Wnt also expressed the osteogenic genes Runx2, Collagen type I, and Osteocalcin sooner than did cells in injury sites treated with PBS (Fig. 5, G to L). Bisphosphonates promote bone accrual primarily by inhibiting osteoclast function (33), but by using TRAP staining as an indicator of osteoclast activity, we found that liposomal Wnt did not enhance bone accrual by blocking bone resorption (Fig. 5, M and N). Thus, the bone-promoting effects of Wnt liposomal treatment were achieved via the proliferation of skeletal progenitor cells, which rapidly adopted an osteogenic fate and gave rise to abundant new bone regenerate specifically in the injury site.

Fig. 5

Liposomal Wnt stimulates skeletal stem or progenitor cell proliferation and enhances osteogenic differentiation. (A and B) Pentachrome staining is used to assess organization of the injury site 1 day after treatment with liposomal PBS (A) or liposomal Wnt3a (B). (C and D) X-Gal staining on similar sections reveals Wnt-responsive cells in the injury sites treated with liposomal PBS (n = 6) (C) or liposomal Wnt3a (n = 6) (D). (E and F) BrdU incorporation indicates skeletal stem or progenitor cell proliferation in injury sites 1 day after treatment with liposomal PBS (n = 6) (E) or liposomal Wnt3a (n = 6) (F). (G and H) Runx2 expression is evaluated in the injury site by in situ hybridization 2 days after delivering liposomal PBS (n = 4) (G) or liposomal Wnt3a (n = 7) (H). (I and J) Collagen type I is evaluated in injury sites 2 days after delivering liposomal PBS (n = 4) (I) or liposomal Wnt3a (n = 7) (J). (K and L) Osteocalcin expression is evaluated in injury sites 3 days after treatment with liposomal PBS (n = 5) (K) or liposomal Wnt3a (n = 5) (L). (M and N) TRAP activity reveals osteoclast function in injury sites 3 days after treatment with liposomal PBS (n = 4) (M) or liposomal Wnt3a (n = 7) (N). Scale bars, 100 μm [(A), (B), (E), and (F)], 10 μm (all other panels).


In this and in previous work, we have shown that Wnt-responsive cells are located on the endosteal surfaces of bone, where they constitute a pool of skeletal stem or progenitor cells that participate in bone homeostasis (Fig. 2) (34). Skeletal trauma activates endogenous Wnt signaling and leads to an expansion in this skeletal stem or progenitor pool. Eventually, skeletal progenitor cells express the osteogenic genes Runx2 and Collagen type I, which indicates their commitment to an osteogenic fate (Fig. 3).

If Wnt signaling is blocked, then both the proliferative effect and osteogenic commitment are impeded in skeletal progenitor cells (5, 23). Here, we have shown that increasing the Wnt signal has an opposite, and beneficial, effect on bone repair. We demonstrate that loss of the negative regulator Axin2 results in a prolonged, amplified Wnt signal, which has a potent effect on cells in the injury site, causing them to proliferate to a greater degree. Axin2LacZ/LacZ skeletal progenitor cells eventually cease their proliferation and differentiate into osteoblasts.

The transient nature of the Wnt signal in Axin2LacZ/LacZ mice is important because other mutations (for example, Lrp5G171V, glycogen synthase kinase 3, activated β-catenin) that either produce constitutive activation of the Wnt pathway or disrupt other signaling pathways in addition to Wnt (17, 3537) maintain skeletal stem or progenitor cells in a proliferative state, delaying their differentiation into osteoblasts and impeding the repair process (5). Thus, a balance between osteoprogenitor proliferation and osteoblast differentiation is required for timely bone regeneration, and Wnt signaling is a key mediator of this equilibrium.

On the basis of these findings, we developed a biochemical approach to transiently increase Wnt signaling after injury. Previously, the use of Wnts as bioactive reagents had been limited because of inherent difficulties in purifying and delivering the hydrophobic protein in vivo. By packaging the lipid-modified Wnt into lipid vesicles (32), we could test the therapeutic efficacy of Wnts. When we supplied liposomal Wnt after the initial inflammatory period, cells in the injury site proliferated more and accelerated their differentiation into osteoblasts. Thus, liposomal Wnt treatment effectively prolonged the endogenous Wnt signal that is initiated by injury and, in doing so, accelerated bone healing.

Bone morphogenetic proteins (BMPs) have been proposed as bone healing reagents (38). Recombinant BMP-2 (rBMP-2) treatment induces cartilage formation, and thus, repair takes place via endochondral ossification (39). In contrast, the same injuries treated with liposomal Wnt heal via intramembranous ossification. In both humans (40) and our mouse model (41), rBMP-2 treatment induces heterotopic ossifications; in contrast, we have never observed ectopic bone formation associated with liposomal Wnt3a treatment. Therefore, liposomal Wnt and rBMPs differentially affect the process of skeletal repair. BMP and Wnt signals are integrated at the level of Smad1 (42), and in skeletal injury, rBMP-2 both induces phosphorylation of Smad1/Smad5/Smad8 and represses β-catenin–dependent Wnt signaling (41). As a consequence, Wnt-dependent Sox9 inhibition (18) is relieved and skeletal stem or progenitor cells in the injured periosteum adopt a chondrogenic lineage (41). Recombinant BMP-2 treatment also represses Wnt signaling in cells that populate the injured bone marrow cavity (41). In vivo, these skeletal stem or progenitor cells have only one fate—to differentiate into osteoblasts. By repressing Wnt signaling in this cell population, bone formation is inhibited within the marrow cavity. Collectively, these experimental results are in keeping with clinical observations associated with rBMPs (43).

Cells in the injured periosteum and bone marrow cavity are responsive to a liposomal Wnt stimulus but, beyond the fact that they can differentiate into osteoblasts, the identities of these cells are still unknown. Multiple cell populations contribute to the bone healing process, including inflammatory cells, muscle satellite cells, pericytes, and angioblasts. Most of these cell types migrate into the wound sites during the first 72 hours after injury and participate to varying degrees in bone regeneration. Many of these populations contain Wnt-responsive cells (4447) and could be targets for the liposomal Wnt signal. A genetic strategy whereby the progeny of Axin2-expressing cells in the injury site could be permanently labeled and then followed throughout the bone healing program would provide valuable insights.

In conclusion, we have demonstrated that liposomal Wnt delivered to an injury site expedites the healing process. We elucidate the basis for their pro-osteogenic effects by showing that liposomal Wnt stimulates skeletal stem or progenitor cell proliferation and also accelerates the differentiation of progenitor cells into mature osteoblasts. In invertebrates, Wnts regulate tissue regeneration (48, 49), and in vertebrates, Wnts are implicated in the repair of multiple organs and tissues (5055). Given the highly conserved nature of the Wnt pathway, and its critical role in maintaining embryonic and adult stem cell self-renewal and proliferation (56), liposomal Wnts may have therapeutic applications far beyond skeletal tissue applications.

Materials and Methods

In vitro quantification of Wnt responsiveness

A DNA fragment containing seven Tcf/lymphoid enhancer factor–binding sites, the minimal promoter, and the 5′ untranslated region of the pSuperTOPflash reporter plasmid (57) was amplified by PCR and inserted upstream of the firefly luciferase (FFluc) gene in the self-inactivating lentivirus TOP-FFluc (58). The 7xTcf-FFluc//SV40 viral construct also contained a SV40-puro cassette for puromycin selection of transfected cells. Lentivirus was produced by transient transfection in 293T cells.

Primary calvarial osteoprogenitor cells were harvested from postnatal day 2 Axin2LacZ/+, Axin2LacZ/LacZ, and wild-type (WT) mice. Passage 1 cells were transfected with 7xTcf-FFluc//SV40-puro lentivirus for 24 hours at a concentration of one transfecting agent per cell. Cultures were treated with puromycin (4 μg/ml) for 2 days to select for transfected cells. Cells were trypsinized, plated in 96-well plates, and treated with purified Wnt3a protein (100 ng/ml) to examine the Wnt responsiveness. Cells were collected at different time points after treatment, and luciferase activity was measured with Luc-Screen Firefly Luciferase Reporter Gene Assay System (Applied Biosystems). Linear regression was calculated with Excel.

In vivo Wnt responsiveness

Axin2LacZ/+ mice were bred to generate Axin2LacZ/LacZ offspring (24). TOPgal transgenic mice were also used (Fig. 5, C and D) as bred as described (59). The LacZ product, β-galactosidase, was detected by X-Gal staining. Tissues were embedded in optimum cutting temperature compound followed by cryosectioning, then fixed with 0.2% glutaraldehyde for 15 min, and stained with X-Gal overnight at 37°C.

μCT analysis

After anesthesia, male, 3-month-old WT, Axin2LacZ/+, and Axin2LacZ/LacZ mice were subjected to μCT analysis with a GE Medical Systems eXplore RS MicroCT System (General Electric Healthcare). Mice were scanned in a prone position at 98-μm resolution. Tape was used to secure limbs and minimize distortion from breath motion. Individual CT slices were reconstructed with GE reconstruction software, and data were analyzed with GE MicroView. A “phantom” density standard containing air, water, and hydroxyapatite (synthetic bone) was scanned alongside each mouse to calibrate and standardize the density of the scanned tissue. MicroView software was used to calculate bone volume fraction and bone mineral density. A standard region of interest (ROI) was specified for either tibiae or femurs of all mice before analysis and was not changed among samples.

Generation of skeletal injuries

All procedures followed protocols approved by the Stanford Committee on Animal Research. Adult mice (males, between 3 and 5 months old) were anaesthetized with an intraperitoneal injection of ketamine-xylazine. A 5-mm incision was made over the anterior-proximal tibia, and the tibial surface was exposed while carefully preserving the periosteum. A 1.0-mm hole was drilled through the anterior cortex with a high-speed dental drill. Wounds were closed with size 6-0 vicryl sutures. After surgery, mice received subcutaneous injections of carprofen for analgesia and were allowed to ambulate freely. Mice were euthanized at days 1, 2, 3, 4, 5, 6, 7, 14, and 28 after surgery.

Western blot and qRT-PCR

Forty-eight hours after skeletal injury (or sham surgery, where access to the bone was made without producing a monocortical defect), tissues were collected by making two transverse cuts 1.0 mm proximal and distal to the injury site. Extreme care was taken to ensure that the same segment of the tibiae was isolated from injured and sham controls.

For gene expression analyses, tissues were homogenized in TRIzol (Invitrogen), and RNA was isolated with RNeasy mini column (Qiagen). Reverse transcription was performed with SuperScript III First-Strand Synthesis SuperMix for qRT-PCR (Invitrogen). Quantitative PCR reactions were performed and monitored with StepOnePlus Real-Time PCR System. Normalized expression levels reported were calculated on the basis of differences between threshold cycles for the gene of interest and the housekeeping gene β-actin. The following primer sets were used: β-actin, 5′-GGAATGGGTCAGAAGGACTC-3′ (sense) and 5′-CATGTCGTCCCAGTTGGTAA-3′ (antisense) (110). In Axin2LacZ/LacZ mice, exon 2 is replaced by the LacZ gene, which results in a truncated, nonfunctional Axin2 protein (24). Exon 1 is unaffected by the LacZ insertion, and therefore, the following primers were used to detect Axin2 expression in Axin2LacZ/LacZ mice: 5′-TTGATAAGGTCCTGGCAACTC-3′ (sense) and 5′-GCGAACGGCTGCTTATTT-3′ (antisense).

For Western blot analyses, the isolated tissue was snap-frozen with liquid nitrogen and homogenized. Proteins were lysed in buffer containing 50 mM tris-HCl (pH 7.4), 150 mM NaCl, 1% NP-40, and 0.1% SDS supplemented with protease inhibitor cocktail (Sigma) and phosphatase inhibitors (1 mM NaF and 1 mM Na3VO4). Proteins were fractionated by 10% SDS–polyacrylamide gel electrophoresis, transferred to nitrocellulose membrane, and incubated with respective primary antibodies. Bound primary antibodies were detected with horseradish peroxidase–conjugated secondary antibodies, visualized by enhanced chemiluminescence (Amersham), and autoradiographed. Rabbit polyclonal total β-catenin (Themo Fisher Scientific), mouse monoclonal antibodies to activated β-catenin, which is dephosphorylated on Ser37 or Thr41 (60), and pan–β-actin (Themo Fisher Scientific) were used.

To quantify the bands obtained via Western blots and RT-PCR, we performed densitometric analysis by using the public domain Java image processing program ImageJ (developed at the National Institutes of Health). The software calculates the area and pixel value statistics of defined selections on the image. Relative change was calculated by between the target proteins and β-actin, which served as an internal control.

Cell quantification

To quantify PCNA immunostaining, we used 4′,6-diamidino-2-phenylindole (DAPI) to counterstain cell nuclei. Using Adobe Photoshop, we used the magic wand tool to highlight brown cells (tolerance = 30). Nonspecific brown areas were manually deselected. These highlighted cells were cut out. Publicly available ImageJ was used to convert these pictures to binary, and the “Analyze Particles” function was used to count discrete particles. To quantify DAPI, we used ImageJ to count discrete nuclei. The “watershed” function was used to approximate tightly packed nuclei. Data were expressed as positive cells per total cells.

X-Gal–positive cells were quantified with ImageJ (see above). For intact bones, X-Gal–positive cells were counted along a 3-mm section of the endosteum and then normalized against the cross-sectional area of the marrow cavity. To quantify X-Gal cells in the tibial injury, we took an image of the total cross-sectional area of the injury and counted the total number of blue cells within the injury site and normalized against the total cross-sectional area of the injury.

Molecular and cellular assays

Under ribonuclease-free conditions, tibiae were harvested, the skin and outer layers of muscle were removed, and the tissues were washed in 1× phosphate-buffered saline (PBS) at 4°C and then fixed in 4% paraformaldehyde. Tissues were decalcified in 19% EDTA for 10 to 14 days at 4°C and then prepared for paraffin embedding. Paraffin embedding followed standard protocols, and sections were generated at an 8-μm thickness. For in situ hybridization, the relevant digoxigenin-labeled mRNA antisense probes were prepared from complementary DNA templates for Runx2, Collagen type I, and Osteocalcin. Sections were dewaxed, treated with proteinase K, and incubated in hybridization buffer containing the relevant RNA probe. Probe was added at an approximate concentration of 0.25 μg/ml. Stringency washes of saline sodium citrate solution were done at 52°C and further washed in maleic acid buffer with 1% Tween 20. Slides were treated with an antibody to digoxigenin (Roche). For color detection, slides were incubated in nitro blue tetrazolium chloride (Roche) and 5-bromo-4-chloro-3-indolyl phosphate (Roche). After developing, the slides were coverslipped with aqueous mounting medium. For immunostaining, tissue sections were dewaxed followed by immersion in H2O2-PBS, washed in PBS, incubated in ficin (Zymed), treated with 0.1 M glycine, washed further, and then blocked in ovalbumin (Worthington) and 1% whole donkey immunoglobulin G (Jackson ImmunoResearch). Appropriate primary antibody was added and incubated overnight at 4°C and then washed in PBS. Samples were incubated with peroxidase-conjugated secondary antibody (Jackson ImmunoResearch) for an hour, and a DAB substrate kit (Vector Laboratories) was used to develop the color reaction. Some commonly used antibodies include PCNA (Zymed) and platelet endothelial cell adhesion molecule 1 (BD Biosciences). For TRAP staining, tissue sections were dewaxed and then treated with a TRAP staining kit (Sigma).

For whole-bone ALP and TRAP activity, tibiae and femurs from WT and Axin2LacZ/LacZ mice were collected immediately after euthanasia. The bone was snap-frozen and homogenized in 400 μl of 0.1% Triton X-100 solution. Sample ALP and TRAP activity was determined with an enzymatic assay on the basis of hydrolysis of p-nitrophenyl phosphate (pNP-PO4) to pNP. Briefly, for ALP activity, aliquots of the sample homogenate (5 μl) were added to 95 μl of 10 mM pNP-PO4 in 0.1 M Na2CO3 and 2 mM MgCl2 (pH 10.0). The samples were then incubated at 37°C for 15 min, and the reaction was terminated by addition of 100 μl of 0.1 N NaOH. For TRAP activity, aliquots of the sample (5 μl) were added to 95 μl of 0.1 M pNP-PO4 in 0.2 M sodium citrate, 0.2 M sodium chloride, and 80 mM sodium l-(+)-tartrate. The samples were then incubated at 37°C for 1 hour, and the reaction was terminated by addition of 100 μl of 1 N NaOH. Absorbance for both assays was measured at 410 nm with a microplate reader.

Histology and histomorphometric analyses

Pentachrome and aniline blue staining were performed; slides were mounted with Permount after dehydration in a series of ethanol and xylene. To quantify new bone, we represented the 1.0-mm circular monocortical defect across about forty 8-μm-thick tissue sections. Of those 40 sections, we used a minimum of 8 sections to quantify the amount of aniline blue–stained new osteoid matrix. Tissue sections were photographed with a Leica digital imaging system (5× objective). The resulting digital images were analyzed with Adobe Photoshop CS2 software. We chose a fixed, rectangular 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. Aniline blue–positive pixels were automatically selected with 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 new osteoid tissue in the aniline blue–stained sections. Cortical surfaces, or bone fragments resulting from the drill injury, were manually deselected. The total number of aniline blue–positive pixels for each section was then recorded. The pixel counts from individual sections were averaged for each tibia sample, and the differences within and among treatment groups were calculated on the basis of these averages.

Liposomal preparation and delivery

Liposomal Wnt3a was prepared as described (32). Briefly, 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC; Sigma) in chloroform was dried to a thin film in a 10-ml round bottom flask. Purified Wnt3a with a concentration of 1 to 1.3 μg/ml was mixed with dried DMPC. The lipid-Wnt3a solution was extruded 40 times through a 100- to 200-nm polycarbonate membrane in a thermobarrel extruder, keeping the temperature constant at 30° to 32°C (Avanti Polar Lipids). The supernatant was removed and the liposome pellet was resuspended in 1× Dulbecco’s modified Eagle’s medium (Mediatech). Liposomes were stored at 4°C and used within 10 days of preparation. The liposomal preparation had an effective Wnt3a concentration of 0.5 μg/ml, and a single (10 μl) dose of this solution was delivered to the injury site by injection.

Monocortical tibial defects were treated with liposomal Wnt3a preparation by injecting 10 μl of liposomal Wnt3a into the injury site on postsurgical day 3.

Statistical analyses

Results are presented as the mean ± SD, with n equal to the number of samples analyzed. Both Student’s t test and nonparametric Wilcoxon test were used to test for significant differences between data sets. Significance was attained at P < 0.05, and all statistical analyses were performed with the JMP software (SAS).

Supplementary Material

Fig. S1. Constitutively active Wnt signaling, but not Axin2LacZ/LacZ mutation, causes pathological bone accrual.

Fig. S2. Bone volume fraction, bone mineral density, and cortical bone thickness in wild-type and Axin2LacZ/LacZ mice.

Fig. S3. Wnt signaling is elevated only at the site of injury.

Fig. S4. Cell proliferation is enhanced in Axin2LacZ/LacZ mice at the site of injury.

Fig. S5. TRAP activity is similar between wild-type and Axin2LacZ/LacZ mice.

Fig. S6. The bony regenerate in Axin2LacZ/LacZ mice appears sooner but is histologically indistinguishable from wild-type mice.

Fig. S7. Wnt signaling is transiently elevated by skeletal injury.

Fig. S8. Liposomal Wnt3a stimulates cell proliferation in skeletal injury sites.


  • * These authors contributed equally to this work.

  • Received June 26, 2009.
  • Accepted April 9, 2010.

References and Notes

  1. Acknowledgments: We thank S. Rooker for technical assistance with the experiments with wild-type and Axin2LacZ/LacZ mice and L. Zhao for the fabrication of liposomal Wnt3a. Funding: CIRM TR1-01249 and FA9550-04-1-0075. S.M. was supported by a Howard Hughes Medical Institute medical scholars fellowship. J.A.H. is a Hagey Faculty Scholar. Author contributions: P.L.: experiments associated with liposomal Wnt3a in the bone healing experiments; S.M.: experiments associated with injuries in Axin2LacZ/+ and Axin2LacZ/LacZ mice; J.J.: in vitro and in vivo analyses on Axin2LacZ/+ and Axin2LacZ/LacZ mice; B.L.: Western analyses; Y.A.Z. provided Axin2LacZ/+ and Axin2LacZ/LacZ mice; C.F.: development of lentiviral Wnt reporter construct; R.N. contributed to experimental design, discussions, data interpretation, and writing of the study; J.A.H. contributed to the experimental design, data interpretation, analyses, and writing of the study. Competing interests: A patent application has been filed by J.A.H., R.N., and P.L. for the composition, production, and therapeutic use for Wnt liposomes. U.S. Patent Application Serial No. 12/074,766 for “WNT Compositions and Methods of Use Thereof.”
View Abstract

Navigate This Article