Parathyroid hormone signaling through low-density lipoprotein-related protein 6

Results

PTH induces β-catenin stabilization in osteoblasts

To determine whether PTH regulates expression of β-catenin, the effects of PTH on β-catenin levels in rat UMR-106 osteoblastic cells were examined. We found that PTH stimulated the transcription of a luciferase reporter bearing TCF/LEF-binding elements (Fig. 1A), and enhanced the abundance of β-catenin in the cytosol (Fig. 1B), whereas the unrelated peptide had no such effects (data not shown). Similarly, PTH enhanced the levels of β-catenin in the cytosol in a concentration- and time-dependent manner in both mouse calvarial primary preosteoblasts (Fig. 1C) and HEK 293 cells (Supplemental Fig. 1). β-Catenin accumulation in the cytosol induced by PTH is so rapid that the effect is unlikely to be mediated through synthesis of Wnt ligands or sensitization of Wnt-stimulated signaling. Indeed, Fz8CRD, a competitive inhibitor of the Wnt receptor Fz (Hsieh et al. 1999), inhibited Wnt3a-elevated, but not PTH-elevated, β-catenin level (Fig. 1D), thus excluding the possibility of the involvement of Wnts. To test whether PTH stimulates β-catenin in vivo, we analyzed the effects PTH (1–34) administered as a single dose to 5-mo-old rats. PTH (1–34) is a C-terminal-truncated synthetic analog of PTH with an anabolic effect on bone formation in humans (Potts et al. 1971; Tregear et al. 1973). Immunohistochemistry analysis of sections of the trabecular bone indicated that PTH induced expression of β-catenin in preosteoblasts and osteoblasts on the bone surface within hours (Fig. 1E,F). At 8 h after injection, positive staining of β-catenin was observed in most osteoblasts (99.08 ± 0.57%) at the metaphysis subjacent to the epiphyseal growth plates and ∼90.24 ± 0.68% of the osteoblasts at the diaphyseal bone marrow. Similar experiments were carried out using 2-mo-old male mice, and similar temporal β-catenin expression patterns were obtained in the mice injected with PTH (Supplemental Fig. 2).

Figure 1.

Activation of β-catenin signaling by PTH in osteoblast-like cells. (A) PTH stimulation of a luciferase reporter with TCF/LEF-binding elements (TCF4-Luc) in UMR-106 cells. Cells were transfected with TCF4-Luc plasmid and treated with control condition medium (CM) collected from culture medium of cells transfected with empty vector, CM-containing Wnt3a, and CM with 10 ^{M PTH (1–84). Luciferase activity was measured 8 h after transfection and normalized to internal controls as} Renilla luciferase units (RLU). (*) P < 0.01, n = 3. (B) PTH induced stabilization of β-catenin in UMR-106 cells. Cells were treated as described in A. Cytosolic and membrane fractions were prepared 1 h after treatment for detection of β-catenin levels by Western blotting analysis. (C) PTH-induced stabilization of β-catenin in mouse primary preosteoblasts. Mouse cavarial preosteoblasts were treated with vehicle (control), increasing dosages of PTH (1–84), or 50 ng/mL mouse recombinant Wnt3a. Cytosolic and membrane fractions were prepared 1 h after treatment for detection of β-catenin levels by Western blotting analysis. (D) PTH-elevated β-catenin level was not affected by Fz8CRD. Mouse cavarial preosteoblasts were treated with Wnt3a CM or 10 ^{M PTH (1–34) together with control CM or Fz8CRD CM for 1 h. Cytosolic and membrane fractions were prepared 1 h after treatment for detection of β-catenin levels by Western blotting analysis. (}E,F) Immunohistochemical analysis of β-catenin levels in femur sections from 5-mo-old male rats at the indicated time points after PTH (1–34) injection (40 μg/kg). Representative of sections immunohistochemically stained with antibody to β-catenin or control IgG and counterstained with hematoxylin viewed at lower power (top row) and higher power (middle and bottom rows). Metaphysis subjacent to the epiphyseal growth plates (middle row) or diaphyseal hematopoietic bone marrow (bottom row) were examined. (E) Red asterisks and green asterisks mark locations in the low-power images that are shown in the high-power fields below. βCatenin-positive osteoblasts were counted in a blinded fashion using OsteoMeasure Software (OsteoMetrics, Inc.) from three random high-power fields per specimens at metaphysis subjacent to diaphyseal hematopoietic bone marrow, and a total of six specimens in each group were used. (F) The quantification of β-catenin-positive osteoblasts is presented as percentage of total osteoblasts. (*) P < 0.005; (**) P <0.001 (in comparison with control), n = 6.

LRP6 forms a complex with PTH/PTH1R

The rapid enhancement of β-catenin protein levels in response to PTH treatment both in vitro and in vivo suggest that PTH may have a direct effect on the signaling components that promote the stabilization of β-catenin. Both LRP5 and LRP6 are key components in activating β-catenin signaling in canonical Wnt pathway. We attempted to examine whether these two receptors are also important in PTH-stimulated effects in osteoblasts. Recent studies reported that PTH anabolic effect was not affected in LRP5 KO mice (Sawakami et al. 2006; Iwaniec et al. 2007), indicating that LRP5 is not essential for the stimulatory effects of PTH on bone formation. We therefore focused on the function of LRP6 in PTH-activated signaling. We first tested whether inactivation of LRP6 would affect PTH-elevated β-catenin level by introducing siRNA complementary to lrp6 mRNA to the cells. Reduction of LRP6 (Fig. 2A) attenuated PTH-stimulated accumulation of β-catenin in the cytosol (Fig. 2B) and TCF/LEF luciferase activity (Fig. 2C). PTH-stimulated mRNA expressions of osteocalcin and RANKL, downstream target genes of PTH that are pertinent to osteoblast differentiation, were also inhibited by the siRNA (Fig. 2D,E). The results indicate that LRP6 is a critical mediator for PTH-induced β-catenin stabilization in osteoblasts.

Figure 2.

Formation of complexes of LRP6 with PTH–PTH1R. (A) LRP6-specific siRNA reduced the amount of LRP6 protein in HEK 293 cells as determined by Western blotting. siRNA directed against GFP was used as an siRNA control. (B) LRP6-specific siRNA reduced PTH-induced β-catenin stabilization in HEK293 cells as determined by Western blotting analysis. (C) LRP6-specific siRNA reduced PTH-stimulated TCF/LEF activity in UMR-106 cells as determined using a luciferase assay. (*) P < 0.01 (in comparison with control), n = 3; (n.s.) not significant (in comparison with control), n = 3. (D,E) Real-time PCR analysis of Osteocalcin (D) and RANKL (E) mRNA expression. C2C12 cells expressing siGFP (control) or siLRP6 together with PTH1R were treated with or without PTH (1–34) in osteogenic induction medium (100 nM ascorbic acid, 10 mM glycerophosphate, and 100 ng/mL BMP2) and harvested at day 3 for RNA extraction. (F) Co-IP of endogenous LRP6 with PTH1R in UMR-106 cells. Cells were serum deprived and treated with 10 ^{M PTH (1–84). The LRP6-associated PTH1R was determined separately by Western blotting of the anti-LRP6 immunoprecipitates. (WCL) Whole-cell lysates. (}G) PTH enhances binding of PTH1R to LRP6, but not LRP5. HEK 293 cells were transfected with VSVG-tagged LRP6 or HA-LRP5 together with PTH1R and treated with 10 ^{M PTH (1–84). The PTH1R-associated LRP5 or LRP6 was determined by Western blotting analysis of the anti-PTH1R immunoprecipitates. (WCL) Whole cell lysates. (}H) Ternary complex of LRP6, PTH, and PTH1R. HEK 293 cells were transfected with VSVG-tagged LRP6 and HA-PTH1R and treated with 10 ^{M PTH (1–84). The LRP6-associated PTH ligand was determined by Western blotting analysis of the anti-VSVG immunoprecipitates. (WCL) Whole cell lysates. (}I–K) PTH brings PTH1R and LRP6 into close proximity as demonstrated by FRET. (I) A photobleaching-based FRET (pbFRET) system was generated by transiently expressing two constructs in HEK293 cells in which CFP and YFP were fused at the C terminus of PTH1R and LRP6, respectively. The interactions of YFP-fused LRP6 with CFP-fused BMPRII or CFP-fused PTH1R with YFP-fused mLRP4T100 were also examined as controls. (J) Representative confocal imaging of the association of CFP-PTH1R with YFP-LRP6 at 5 min after PTH treatment in HEK293 cells by pbFRET. The total photobleached area (ROI_1) is marked with a green square. Quantification of fluorescent intensities of each chosen point within (ROI_2∼ROI_6) or outside of the marked bleached area (ROI_7∼ROI_9) by averaging fluorescence before and after the bleach was conducted. (K) Comparison of the FRET efficiencies (FRET Eff%) before and after photobleaching in the absence or presence of PTH. (*) P < 0.001, compare with unbleached, n = 6; (n.s.) not significant compare with unbleached. (L,M) Ventral injection of RNA for PTH (2 pg) plus PTH1R (50 pg) promotes LRP6 (200 pg)-induced axis duplication. (n) Numbers of embryos scored.

We then examined the possibility that LRP6 may form a ternary complex with PTH and PTH1R as it does with Wnt and Fz. Immunoprecipitation (IP) with antibodies to LRP6 from lysates of PTH-treated UMR-106 cells indicated that PTH1R formed a complex with endogenous LRP6 in response to PTH in a time-dependent manner (Fig. 2F). Unlike LRP6, PTH did not enhance the binding of LRP5 to PTH1R, although there is detectable binding in the absence of PTH (Fig. 2G), indicating that LRP5 may play a different role in PTH signaling. The presence of PTH ligand in the LRP6–PTH1R complex was also indicated by co-IP. The PTH ligand was immunoprecipitated by LRP6 only when both LRP6 and PTH1R were present (Fig. 2H), suggesting that PTH forms a ternary complex with LRP6 and PTH1R. Further evidence for the PTH–PTH1R–LRP6 complex formation was obtained from PTH-induced close association of PTH1R with LRP6 in cells by photobleaching-based fluorescence resonance energy transfer (FRET) (Fig. 2I–K). As shown in Figure 2J and K, PTH led to increased FRET efficiency between CFP-PTH1R and YFP-LRP6, but did not enhance the FRET efficiency in either YFP-LRP6 and CFP-BMPRII, BMP type II receptor (Cao and Chen 2005), or between CFP-PTH1R and YFP-mLRP4T100, another member of the low-density lipoprotein-related proteins family (Li et al. 2000). Thus, LRP6 specifically interacts with PTH1R upon PTH stimulation. The association of PTH1R with LRP6 is also supported by analysis of the model of LRP6-mediated secondary axis induction in Xenopus, in which PTH enhanced LRP6-induced secondary axis induction (Fig. 2L,M).

Extracellular domain of LRP6 interacts with PTH1R

To confirm and extend the studies of the LRP6 and PTH1R complex formation, we mapped the region of LRP6 required for its interaction with PTH1R. PTH1R was coexpressed in cells with LRP6, a truncated LRP6 containing the extracellular and transmembrane domains (LRP6N + T), or the transmembrane and intracellular domains (LRP6T + C) for IP assay. Binding of LRP6T + C to PTH1R could barely be detected, but the LRP6N + T associated with PTH1R as effectively as did full-length LRP6 (Fig. 3A). The presence of PTH in the LRP6N + T/PTH1R complex further suggested the formation of a ternary complex. Moreover, PTH-induced direct interaction of LRP6N with PTH1R on the cell surface was confirmed in an immunefluorescence colocalization assay. Immun-colocalization of LRP6N–IgG with PTH1R on the cell surface was increased from 22.8% to 82.3% with addition of PTH ligand (Fig. 3B [top two rows], C) whereas binding of IgG to PTH1R was barely detected (Fig. 3B [bottom two rows], C). Collectively, the results indicate that PTH induces formation of PTH1R/LRP6 complex through LRP6 extracellular domain.

Figure 3.

Extracellular domain of LRP6 interacts with PTH1R. (A) Interaction of the N-terminal domain of LRP6 with PTH1R. HEK 293 cells were transfected with indicated plasmids and treated with or without 10 ^{M PTH (1–84). IP was done to identify the interaction. (}B,C) Surface binding of LRP6N in the presence of PTH treatment. HEK 293 cells were transfected with HA-PTH1R and treated with CM containing human IgG or LRP6N-IgG for 1 h following PTH (1–84) treatment for 15 min. Cells were washed, fixed, and immunostained with either anti-HA (red) or anti-IgG (green). Nuclei were visualized using Hoechst 33342. (B) Representative images are shown. (C) IgG CM or LRP6N CM surface binding rates; i.e., the ratios of the number of cells showing in green to the number of cells showing in red were calculated. (D) Soluble LRP6N disrupts PTH1R binding to endogenous LRP6. UMR-106 cells were treated with control CM or LRP6N CM for 1 h followed by 10 ^{M PTH (1–84) treatment for another 1 h. Cells were washed with PBS and cell lysates were collected. The endogenous PTH1R-associated LRP6 was determined by Western blotting of the anti-PTH1R immunoprecipitates. PI, preimmune serum control. (}E) Inhibition of PTH-induced TCF4/LEF activation by soluble LRP6N. UMR-106 cells were transfected with TCF4-Luc plasmid and treated with control CM or LRP6N CM followed by PTH (1–84) treatment. Luciferase activity was measured and normalized to internal controls as Renilla luciferase units (RLU). (*) P < 0.01 (in comparison with control), n = 3; (n.s.) not significant (in comparison with control), n = 3. (F) LRP6N does not affect LiCl-induced TCF4/LEF activation. UMR-106 cells were transfected with TCF4-Luc plasmid and treated with control CM or LRP6N CM followed by 20 mM LiCl treatment. Luciferase activity was measured and normalized to internal controls as Renilla luciferase units (RLU). (*) P < 0.001 (in comparison with control), n = 3. (G) DKK1 reduced Wnt3a- or PTH-induced β-catenin stabilization in HEK293 cells as determined by Western blotting analysis. (H) Inhibition of PTH-induced TCF4/LEF activation by DKK1 and Sclerostin. UMR-106 cells were transfected with TCF4-Luc plasmid and treated with control CM, DKK1 CM, or Sclerostin CM followed by Wnt3a or PTH (1–84) treatment. Luciferase activity was measured and normalized to internal controls as Renilla luciferase units (RLU). (*) P < 0.01 (in comparison with Wnt3a or PTH treatment), n = 3.

We then examined whether LRP6N acts as a dominant-negative in PTH signaling through LRP6. As expected, LRP6N blocked the PTH-induced association of endogenous LRP6 with PTH1R (Fig. 3D). LRP6N inhibited PTH-, but not LiCl-induced TCF transcriptional activity (Fig. 3E,F). As LiCl directly inhibits GSK3 kinase in the cytoplasm to stabilize of β-catenin (Stambolic et al. 1996), our results indicate that LRP6N acts upstream of GSK3 and functions as a dominant-negative in the PTH-activated β-catenin signaling via binding to cell surface PTH1R. Furthermore, secreted proteins DKK1 and sclerostin, also binding to LRP6 at the extracellular domain (Bafico et al. 2001; B. Mao et al. 2001; Semenov et al. 2001, 2005; Li et al. 2005), disrupted PTH-induced β-catenin accumulation in the cytoplasm (Fig. 3G) and TCF/LEF luciferase activity (Fig. 3H). Thus, PTH-induced recruitment of LRP6 through its extracellular domain is essential in activation of β-catenin signaling pathway.

PTH induces phosphorylation of LRP6 and axin recruitment in osteoblasts

As phosphorylation of LRP6 at the PPPSP motifs plays a crucial role in activating downstream β-catenin signaling by Wnt (J. Mao et al. 2001; Tamai et al. 2004; Davidson et al. 2005; Zeng et al. 2005), we examined whether PTH induces phosphorylation of LRP6 at PPPSP motifs. Immunoprecipitated LRP6 from extracts of PTH-treated UMR-106 cells were monitored for their phosphorylation by Western blotting with an antibody that recognizes phosphorylated PPPSP motifs (Ab1490) (Tamai et al. 2004). PTH rapidly induced the phosphorylation of LRP6 at the PPPSP motifs (Fig. 4A). Phosphorylation of PPPSP motifs is required for axin recruitment from cytoplasm to LRP6 at cell membrane. PTH also increased axin1 level on cell membrane detected by cell fractionation assay in primary preosteoblasts (Fig. 4B). Consistently, PTH rapidly increased in the binding of axin to LRP6 by co-IP assays (Fig. 4C). Again, Fz8CRD, a competitive inhibitor of the Wnt receptor Fz (Hsieh et al. 1999), was used to exclude the possibility that these PTH effects are mediated through promotion of Wnts production or sensitization of Wnt-stimulated signaling. Fz8CRD inhibited Wnt3a-induced phosphorylation of LRP6 (Fig. 4D, lane 8), but did not inhibit the effect of PTH (Fig. 4D, lane 4). In contrast, LRP6N blocked PTH-stimulated LRP6 phosphorylation (Fig. 4E). The results indicate that PTH-induced formation of PTH1R–LRP6 complex through their extracellular domains is essential for phosphorylation of LRP6.

Figure 4.

PTH-induced phosphorylation of LRP6 and axin recruitment. (A) UMR-106 cells were serum deprived and treated with 10 ^{M PTH (1–84). Phosphorylated endogenous LRP6 was identified by Western blotting analysis of the anti-LRP6 immunoprecipitates with Ab1490. (}B) PTH induced the recruitment of axin1 to cell membrane in mouse primary preosteoblasts. Cells were treated with vehicle or 10 ^{M PTH (1–84) for 30 min. Membrane fractions were prepared for detection of axin1 and LRP6 levels by Western blotting analysis. (}C) LRP6–axin binding in HEK 293 cells. Cells were transfected with PTH1R, HA-tagged axin, and VSVG-tagged LRP6 and treated with 10 ^{M PTH (1–84). The axin-associated LRP6 was determined by Western blotting of the anti-VSVG immunoprecipitates. (}D) Fz8CRD does not inhibit PTH-induced LRP6 phosphorylation. HEK293 (for lanes 1–4) or MEF (for lanes 5–8) cells were transfected with VSVG-tagged LRP6, treated with control CM or Fz8CRD CM for 2 h, and then with 10 ^{M PTH (1–84) or Wnt3a. The phosphorylated LRP6 was detected by Western blotting analysis of the anti-VSVG immunoprecipitates with Ab1490. (}E) UMR-106 cells were serum deprived, treated with control CM or LRP6N CM for 1 h, and then with 10 ^{M PTH (1–84) for another 15 min. Phosphorylated endogenous LRP6 was identified by Western blotting analysis of the anti-LRP6 immunoprecipitates with Ab1490. (}F,G) Immunohistochemical analysis of phosphorylated LRP6 levels in femur sections from 5-mo-old male rats at the indicated time points after PTH (1–34) injection (40 μg/kg). Representative of sections immunohistochemically stained with antibody to total LRP6 (top row) and phosphorylated LRP6 (Ab1490, middle and bottom rows), and counterstained with hematoxylin viewed at lower power (middle row) and higher power (bottom row). The metaphyseal area of distal femurs was examined. (F) Double asterisks mark locations in the low-power images that are shown in the high-power fields below. The phosphorylated LRP6-positive osteoblasts were counted in a blinded fashion using OsteoMeasure Software (OsteoMetrics, Inc.) from three random high-power fields per specimens at metaphysis subjacent to diaphyseal hematopoietic bone marrow, and a total of six specimens in each group were used. (G) The quantification of phosphorylated LRP6-positive osteoblasts is presented as percentage of total osteoblasts. (*) P < 0.005; (**) P < 0.001 (in comparison with control), n = 6.

Because the levels of β-catenin were increased in osteoblasts of rats with injection of a single dose of PTH (Fig. 1E,F), we tested whether the amounts of phosphorylated LRP6 were enhanced in osteoblasts from the same tissue. Immunostaining with an antibody specific for the phosphorylated PPPSP demonstrated that PTH-stimulated phosphorylation of LRP6 in preosteoblasts or osteoblasts at the surface of trabecular bone (Fig. 4F, second and third rows), whereas the amount of total LRP6 protein remained unchanged (Fig. 4F, first row). The temporal pattern of phosphorylation of the PPPSP motifs was similar to that of β-catenin (cf Figs. 1E, ​,4F4F [second and third rows] and Figs. 1F, ​,4G).4G). Thus, PTH increases the abundance of β-catenin in osteoblasts in vivo through phosphorylation of LRP6.

PKA is required in PTH-, but not in Wnt-activated LRP6–β-catenin signaling

As the activation of β-catenin signaling by PTH in osteoblasts seems to be independent of Wnt, we attempted to investigate the mechanism responsible for the PTH effects. PTH activates cAMP-dependent PKA, which is sufficient for initiation of signals mediating PTH action in osteoblasts. We assessed whether PKA participates in PTH-activated LRP6–β-catenin signaling. Binding of intact PTH (1–84) or PTH (1–34) to PTH1R activates PKA. However, the native C-terminal fragments of PTH bind PTH1R but do not activate PKA (Kronenberg et al. 1998; Gensure et al. 2005; Murray et al. 2005). The C-terminal fragments PTH (7–84) and PTH (39–84) were much less effective than PTH (1–84) in activating β-catenin signaling (Fig. 5A), altering the stability of β-catenin (Fig. 5B), and inducing axin–LRP6 binding (Fig. 5C). These results suggest that cAMP–PKA activation is involved in activation of LRP6. The minimum effects induced by PTH (7–84) and PTH (39–84) (Fig. 5A–C) may be mediated by other signaling components than PKA as the unrelated peptide of the similar length did not exert such effect (data not shown). PKI-(14–22), a specific inhibitor of PKA-directed phosphorylation, inhibited PTH-induced LRP6 phosphorylation (Fig. 5D). Moreover, the PKA inhibitors, PKI-(14–22) and H89 reduced the binding of axin to LRP6 (Fig. 5E), β-catenin stabilization (Fig. 5F, lane 3), and β-catenin-dependent transcription activity (Fig. 5G), further indicating that PKA activity is essential for PTH-activated LRP6–β-catenin signaling. However, H89 did not affect Wnt3a-stimulated LRP6 phosphorylation (Fig. 5H), β-catenin stabilization (Fig. 5F, lane 5), and β-catenin-dependent transcription activity (Fig. 5I). The results provide further evidence to support the concept that both PTH and Wnt activate LRP6–β-catenin signaling in osteoblasts, but do so through distinct pathways.

Figure 5.

PKA mediates PTH-, but not Wnt3a-activated LRP6–β-catenin signaling. (A) Failure of PTH C-terminal ligands to stimulate TCF4/LEF activity. UMR-106 cells were transfected with TCF4-Luc plasmid, treated with 10 ^{M PTH ligands and subject to a luciferase assay. (*)} P < 0.05 (in comparison with control), n = 3; (**) P < 0.005 (in comparison with control), n = 3. (B) Failure of PTH C-terminal ligands to stabilize β-catenin. UMR-106 cells were treated with PTH ligands for 1 h. Cytosolic fractions were prepared for detection of β-catenin protein by Western blot analysis. (C) Effect of PTH C-terminal ligands on LRP6–axin binding. Cells were transfected with indicated plasmids and treated with PTH ligands. The LRP6-associated axin was determined by Western blotting of the anti-VSVG immunoprecipitates. (WCL) Whole-cell lysates. (D) Inhibition of PTH-induced LRP6 phosphorylation by PKA inhibitor. Cells were transfected with VSVG-LRP6 and PTH1R, metabolically labeled with [^{P] phosphate, and treated with PKI-(14–22) for 1 h before adding PTH (1–34) for another 15 min. VSVG-LRP6 was immunoprecipitated from cell lysates. Proteins were resolved by SDS-PAGE and visualized by autoradiography. (}E) Inhibition of the binding of axin with LRP6 by PKA inhibitors. Cells were transfected with indicated plasmids and pretreated with H89 or PKI-(14–22) for 1 h before adding PTH (1–84) for another 30 min. The LRP6-associated axin was determined as in C. (F) Inhibition of PTH-induced, but not Wnt3a-induced β-catenin stabilization by PKA inhibitor. UMR-106 cells were treated with 10 ^{M PTH (1–34) or 50 ng/mL recombinant mouse Wnt3a (rmWnt3a) together with vehicle (control) or H89. Cytosolic fractions were prepared for detection of β-catenin protein by Western blot analysis. (}G) Inhibition of PTH-activated TCF4/LEF activity by PKA inhibitors. UMR-106 cells were transfected with TCF4-Luc plasmid, treated with 10 ^{M PTH together with vehicle (control), H89, or PKI-(14–22), and subject to a luciferase assay. (*)} P < 0.005 (in comparison with PTH treatment group), I = 3. (H) Failure of PKA inhibitor to inhibit Wnt3a-induced LRP6 phosphorylation. MEFs were transfected with VSVG-LRP6 and pretreated with increasing doses of H89 for 1 h before adding control CM or Wnt3a CM for another 30 min. The phosphorylated LRP6 was detected by Western blotting analysis of the anti-VSVG immunoprecipitates with Ab1490. (I) Failure of PKA inhibitors to inhibit Wnt3a-stimulated TCF4/LEF activity. UMR-106 cells were transfected with TCF4-Luc plasmid, treated with control CM (Con) or Wnt3a CM (Wnt3a) together with vehicle, H89, or PKI-(14–22) and subject to a luciferase assay. (n.s.) No significance (in comparison with Wnt3a group), n = 3. (J) Schematic model of LRP6–β-catenin pathway activation in response to PTH. Upon PTH ligand binding to its receptor PTH1R, LRP6 are recruited and form complexes with PTH/PTH1R, thus positioning LRP6 in close proximity with PTH1R. In parallel, PKA is activated downstream from PTH1R and mediate the phosphorylation of LRP6, which leads to the recruitment of axin and stabilization of β-catenin.

PTH induces β-catenin stabilization in osteoblasts

To determine whether PTH regulates expression of β-catenin, the effects of PTH on β-catenin levels in rat UMR-106 osteoblastic cells were examined. We found that PTH stimulated the transcription of a luciferase reporter bearing TCF/LEF-binding elements (Fig. 1A), and enhanced the abundance of β-catenin in the cytosol (Fig. 1B), whereas the unrelated peptide had no such effects (data not shown). Similarly, PTH enhanced the levels of β-catenin in the cytosol in a concentration- and time-dependent manner in both mouse calvarial primary preosteoblasts (Fig. 1C) and HEK 293 cells (Supplemental Fig. 1). β-Catenin accumulation in the cytosol induced by PTH is so rapid that the effect is unlikely to be mediated through synthesis of Wnt ligands or sensitization of Wnt-stimulated signaling. Indeed, Fz8CRD, a competitive inhibitor of the Wnt receptor Fz (Hsieh et al. 1999), inhibited Wnt3a-elevated, but not PTH-elevated, β-catenin level (Fig. 1D), thus excluding the possibility of the involvement of Wnts. To test whether PTH stimulates β-catenin in vivo, we analyzed the effects PTH (1–34) administered as a single dose to 5-mo-old rats. PTH (1–34) is a C-terminal-truncated synthetic analog of PTH with an anabolic effect on bone formation in humans (Potts et al. 1971; Tregear et al. 1973). Immunohistochemistry analysis of sections of the trabecular bone indicated that PTH induced expression of β-catenin in preosteoblasts and osteoblasts on the bone surface within hours (Fig. 1E,F). At 8 h after injection, positive staining of β-catenin was observed in most osteoblasts (99.08 ± 0.57%) at the metaphysis subjacent to the epiphyseal growth plates and ∼90.24 ± 0.68% of the osteoblasts at the diaphyseal bone marrow. Similar experiments were carried out using 2-mo-old male mice, and similar temporal β-catenin expression patterns were obtained in the mice injected with PTH (Supplemental Fig. 2).

Figure 1.

Activation of β-catenin signaling by PTH in osteoblast-like cells. (A) PTH stimulation of a luciferase reporter with TCF/LEF-binding elements (TCF4-Luc) in UMR-106 cells. Cells were transfected with TCF4-Luc plasmid and treated with control condition medium (CM) collected from culture medium of cells transfected with empty vector, CM-containing Wnt3a, and CM with 10 ^{M PTH (1–84). Luciferase activity was measured 8 h after transfection and normalized to internal controls as} Renilla luciferase units (RLU). (*) P < 0.01, n = 3. (B) PTH induced stabilization of β-catenin in UMR-106 cells. Cells were treated as described in A. Cytosolic and membrane fractions were prepared 1 h after treatment for detection of β-catenin levels by Western blotting analysis. (C) PTH-induced stabilization of β-catenin in mouse primary preosteoblasts. Mouse cavarial preosteoblasts were treated with vehicle (control), increasing dosages of PTH (1–84), or 50 ng/mL mouse recombinant Wnt3a. Cytosolic and membrane fractions were prepared 1 h after treatment for detection of β-catenin levels by Western blotting analysis. (D) PTH-elevated β-catenin level was not affected by Fz8CRD. Mouse cavarial preosteoblasts were treated with Wnt3a CM or 10 ^{M PTH (1–34) together with control CM or Fz8CRD CM for 1 h. Cytosolic and membrane fractions were prepared 1 h after treatment for detection of β-catenin levels by Western blotting analysis. (}E,F) Immunohistochemical analysis of β-catenin levels in femur sections from 5-mo-old male rats at the indicated time points after PTH (1–34) injection (40 μg/kg). Representative of sections immunohistochemically stained with antibody to β-catenin or control IgG and counterstained with hematoxylin viewed at lower power (top row) and higher power (middle and bottom rows). Metaphysis subjacent to the epiphyseal growth plates (middle row) or diaphyseal hematopoietic bone marrow (bottom row) were examined. (E) Red asterisks and green asterisks mark locations in the low-power images that are shown in the high-power fields below. βCatenin-positive osteoblasts were counted in a blinded fashion using OsteoMeasure Software (OsteoMetrics, Inc.) from three random high-power fields per specimens at metaphysis subjacent to diaphyseal hematopoietic bone marrow, and a total of six specimens in each group were used. (F) The quantification of β-catenin-positive osteoblasts is presented as percentage of total osteoblasts. (*) P < 0.005; (**) P <0.001 (in comparison with control), n = 6.

LRP6 forms a complex with PTH/PTH1R

The rapid enhancement of β-catenin protein levels in response to PTH treatment both in vitro and in vivo suggest that PTH may have a direct effect on the signaling components that promote the stabilization of β-catenin. Both LRP5 and LRP6 are key components in activating β-catenin signaling in canonical Wnt pathway. We attempted to examine whether these two receptors are also important in PTH-stimulated effects in osteoblasts. Recent studies reported that PTH anabolic effect was not affected in LRP5 KO mice (Sawakami et al. 2006; Iwaniec et al. 2007), indicating that LRP5 is not essential for the stimulatory effects of PTH on bone formation. We therefore focused on the function of LRP6 in PTH-activated signaling. We first tested whether inactivation of LRP6 would affect PTH-elevated β-catenin level by introducing siRNA complementary to lrp6 mRNA to the cells. Reduction of LRP6 (Fig. 2A) attenuated PTH-stimulated accumulation of β-catenin in the cytosol (Fig. 2B) and TCF/LEF luciferase activity (Fig. 2C). PTH-stimulated mRNA expressions of osteocalcin and RANKL, downstream target genes of PTH that are pertinent to osteoblast differentiation, were also inhibited by the siRNA (Fig. 2D,E). The results indicate that LRP6 is a critical mediator for PTH-induced β-catenin stabilization in osteoblasts.

Figure 2.

Formation of complexes of LRP6 with PTH–PTH1R. (A) LRP6-specific siRNA reduced the amount of LRP6 protein in HEK 293 cells as determined by Western blotting. siRNA directed against GFP was used as an siRNA control. (B) LRP6-specific siRNA reduced PTH-induced β-catenin stabilization in HEK293 cells as determined by Western blotting analysis. (C) LRP6-specific siRNA reduced PTH-stimulated TCF/LEF activity in UMR-106 cells as determined using a luciferase assay. (*) P < 0.01 (in comparison with control), n = 3; (n.s.) not significant (in comparison with control), n = 3. (D,E) Real-time PCR analysis of Osteocalcin (D) and RANKL (E) mRNA expression. C2C12 cells expressing siGFP (control) or siLRP6 together with PTH1R were treated with or without PTH (1–34) in osteogenic induction medium (100 nM ascorbic acid, 10 mM glycerophosphate, and 100 ng/mL BMP2) and harvested at day 3 for RNA extraction. (F) Co-IP of endogenous LRP6 with PTH1R in UMR-106 cells. Cells were serum deprived and treated with 10 ^{M PTH (1–84). The LRP6-associated PTH1R was determined separately by Western blotting of the anti-LRP6 immunoprecipitates. (WCL) Whole-cell lysates. (}G) PTH enhances binding of PTH1R to LRP6, but not LRP5. HEK 293 cells were transfected with VSVG-tagged LRP6 or HA-LRP5 together with PTH1R and treated with 10 ^{M PTH (1–84). The PTH1R-associated LRP5 or LRP6 was determined by Western blotting analysis of the anti-PTH1R immunoprecipitates. (WCL) Whole cell lysates. (}H) Ternary complex of LRP6, PTH, and PTH1R. HEK 293 cells were transfected with VSVG-tagged LRP6 and HA-PTH1R and treated with 10 ^{M PTH (1–84). The LRP6-associated PTH ligand was determined by Western blotting analysis of the anti-VSVG immunoprecipitates. (WCL) Whole cell lysates. (}I–K) PTH brings PTH1R and LRP6 into close proximity as demonstrated by FRET. (I) A photobleaching-based FRET (pbFRET) system was generated by transiently expressing two constructs in HEK293 cells in which CFP and YFP were fused at the C terminus of PTH1R and LRP6, respectively. The interactions of YFP-fused LRP6 with CFP-fused BMPRII or CFP-fused PTH1R with YFP-fused mLRP4T100 were also examined as controls. (J) Representative confocal imaging of the association of CFP-PTH1R with YFP-LRP6 at 5 min after PTH treatment in HEK293 cells by pbFRET. The total photobleached area (ROI_1) is marked with a green square. Quantification of fluorescent intensities of each chosen point within (ROI_2∼ROI_6) or outside of the marked bleached area (ROI_7∼ROI_9) by averaging fluorescence before and after the bleach was conducted. (K) Comparison of the FRET efficiencies (FRET Eff%) before and after photobleaching in the absence or presence of PTH. (*) P < 0.001, compare with unbleached, n = 6; (n.s.) not significant compare with unbleached. (L,M) Ventral injection of RNA for PTH (2 pg) plus PTH1R (50 pg) promotes LRP6 (200 pg)-induced axis duplication. (n) Numbers of embryos scored.

We then examined the possibility that LRP6 may form a ternary complex with PTH and PTH1R as it does with Wnt and Fz. Immunoprecipitation (IP) with antibodies to LRP6 from lysates of PTH-treated UMR-106 cells indicated that PTH1R formed a complex with endogenous LRP6 in response to PTH in a time-dependent manner (Fig. 2F). Unlike LRP6, PTH did not enhance the binding of LRP5 to PTH1R, although there is detectable binding in the absence of PTH (Fig. 2G), indicating that LRP5 may play a different role in PTH signaling. The presence of PTH ligand in the LRP6–PTH1R complex was also indicated by co-IP. The PTH ligand was immunoprecipitated by LRP6 only when both LRP6 and PTH1R were present (Fig. 2H), suggesting that PTH forms a ternary complex with LRP6 and PTH1R. Further evidence for the PTH–PTH1R–LRP6 complex formation was obtained from PTH-induced close association of PTH1R with LRP6 in cells by photobleaching-based fluorescence resonance energy transfer (FRET) (Fig. 2I–K). As shown in Figure 2J and K, PTH led to increased FRET efficiency between CFP-PTH1R and YFP-LRP6, but did not enhance the FRET efficiency in either YFP-LRP6 and CFP-BMPRII, BMP type II receptor (Cao and Chen 2005), or between CFP-PTH1R and YFP-mLRP4T100, another member of the low-density lipoprotein-related proteins family (Li et al. 2000). Thus, LRP6 specifically interacts with PTH1R upon PTH stimulation. The association of PTH1R with LRP6 is also supported by analysis of the model of LRP6-mediated secondary axis induction in Xenopus, in which PTH enhanced LRP6-induced secondary axis induction (Fig. 2L,M).

Extracellular domain of LRP6 interacts with PTH1R

To confirm and extend the studies of the LRP6 and PTH1R complex formation, we mapped the region of LRP6 required for its interaction with PTH1R. PTH1R was coexpressed in cells with LRP6, a truncated LRP6 containing the extracellular and transmembrane domains (LRP6N + T), or the transmembrane and intracellular domains (LRP6T + C) for IP assay. Binding of LRP6T + C to PTH1R could barely be detected, but the LRP6N + T associated with PTH1R as effectively as did full-length LRP6 (Fig. 3A). The presence of PTH in the LRP6N + T/PTH1R complex further suggested the formation of a ternary complex. Moreover, PTH-induced direct interaction of LRP6N with PTH1R on the cell surface was confirmed in an immunefluorescence colocalization assay. Immun-colocalization of LRP6N–IgG with PTH1R on the cell surface was increased from 22.8% to 82.3% with addition of PTH ligand (Fig. 3B [top two rows], C) whereas binding of IgG to PTH1R was barely detected (Fig. 3B [bottom two rows], C). Collectively, the results indicate that PTH induces formation of PTH1R/LRP6 complex through LRP6 extracellular domain.

Figure 3.

Extracellular domain of LRP6 interacts with PTH1R. (A) Interaction of the N-terminal domain of LRP6 with PTH1R. HEK 293 cells were transfected with indicated plasmids and treated with or without 10 ^{M PTH (1–84). IP was done to identify the interaction. (}B,C) Surface binding of LRP6N in the presence of PTH treatment. HEK 293 cells were transfected with HA-PTH1R and treated with CM containing human IgG or LRP6N-IgG for 1 h following PTH (1–84) treatment for 15 min. Cells were washed, fixed, and immunostained with either anti-HA (red) or anti-IgG (green). Nuclei were visualized using Hoechst 33342. (B) Representative images are shown. (C) IgG CM or LRP6N CM surface binding rates; i.e., the ratios of the number of cells showing in green to the number of cells showing in red were calculated. (D) Soluble LRP6N disrupts PTH1R binding to endogenous LRP6. UMR-106 cells were treated with control CM or LRP6N CM for 1 h followed by 10 ^{M PTH (1–84) treatment for another 1 h. Cells were washed with PBS and cell lysates were collected. The endogenous PTH1R-associated LRP6 was determined by Western blotting of the anti-PTH1R immunoprecipitates. PI, preimmune serum control. (}E) Inhibition of PTH-induced TCF4/LEF activation by soluble LRP6N. UMR-106 cells were transfected with TCF4-Luc plasmid and treated with control CM or LRP6N CM followed by PTH (1–84) treatment. Luciferase activity was measured and normalized to internal controls as Renilla luciferase units (RLU). (*) P < 0.01 (in comparison with control), n = 3; (n.s.) not significant (in comparison with control), n = 3. (F) LRP6N does not affect LiCl-induced TCF4/LEF activation. UMR-106 cells were transfected with TCF4-Luc plasmid and treated with control CM or LRP6N CM followed by 20 mM LiCl treatment. Luciferase activity was measured and normalized to internal controls as Renilla luciferase units (RLU). (*) P < 0.001 (in comparison with control), n = 3. (G) DKK1 reduced Wnt3a- or PTH-induced β-catenin stabilization in HEK293 cells as determined by Western blotting analysis. (H) Inhibition of PTH-induced TCF4/LEF activation by DKK1 and Sclerostin. UMR-106 cells were transfected with TCF4-Luc plasmid and treated with control CM, DKK1 CM, or Sclerostin CM followed by Wnt3a or PTH (1–84) treatment. Luciferase activity was measured and normalized to internal controls as Renilla luciferase units (RLU). (*) P < 0.01 (in comparison with Wnt3a or PTH treatment), n = 3.

We then examined whether LRP6N acts as a dominant-negative in PTH signaling through LRP6. As expected, LRP6N blocked the PTH-induced association of endogenous LRP6 with PTH1R (Fig. 3D). LRP6N inhibited PTH-, but not LiCl-induced TCF transcriptional activity (Fig. 3E,F). As LiCl directly inhibits GSK3 kinase in the cytoplasm to stabilize of β-catenin (Stambolic et al. 1996), our results indicate that LRP6N acts upstream of GSK3 and functions as a dominant-negative in the PTH-activated β-catenin signaling via binding to cell surface PTH1R. Furthermore, secreted proteins DKK1 and sclerostin, also binding to LRP6 at the extracellular domain (Bafico et al. 2001; B. Mao et al. 2001; Semenov et al. 2001, 2005; Li et al. 2005), disrupted PTH-induced β-catenin accumulation in the cytoplasm (Fig. 3G) and TCF/LEF luciferase activity (Fig. 3H). Thus, PTH-induced recruitment of LRP6 through its extracellular domain is essential in activation of β-catenin signaling pathway.

PTH induces phosphorylation of LRP6 and axin recruitment in osteoblasts

As phosphorylation of LRP6 at the PPPSP motifs plays a crucial role in activating downstream β-catenin signaling by Wnt (J. Mao et al. 2001; Tamai et al. 2004; Davidson et al. 2005; Zeng et al. 2005), we examined whether PTH induces phosphorylation of LRP6 at PPPSP motifs. Immunoprecipitated LRP6 from extracts of PTH-treated UMR-106 cells were monitored for their phosphorylation by Western blotting with an antibody that recognizes phosphorylated PPPSP motifs (Ab1490) (Tamai et al. 2004). PTH rapidly induced the phosphorylation of LRP6 at the PPPSP motifs (Fig. 4A). Phosphorylation of PPPSP motifs is required for axin recruitment from cytoplasm to LRP6 at cell membrane. PTH also increased axin1 level on cell membrane detected by cell fractionation assay in primary preosteoblasts (Fig. 4B). Consistently, PTH rapidly increased in the binding of axin to LRP6 by co-IP assays (Fig. 4C). Again, Fz8CRD, a competitive inhibitor of the Wnt receptor Fz (Hsieh et al. 1999), was used to exclude the possibility that these PTH effects are mediated through promotion of Wnts production or sensitization of Wnt-stimulated signaling. Fz8CRD inhibited Wnt3a-induced phosphorylation of LRP6 (Fig. 4D, lane 8), but did not inhibit the effect of PTH (Fig. 4D, lane 4). In contrast, LRP6N blocked PTH-stimulated LRP6 phosphorylation (Fig. 4E). The results indicate that PTH-induced formation of PTH1R–LRP6 complex through their extracellular domains is essential for phosphorylation of LRP6.

Figure 4.

PTH-induced phosphorylation of LRP6 and axin recruitment. (A) UMR-106 cells were serum deprived and treated with 10 ^{M PTH (1–84). Phosphorylated endogenous LRP6 was identified by Western blotting analysis of the anti-LRP6 immunoprecipitates with Ab1490. (}B) PTH induced the recruitment of axin1 to cell membrane in mouse primary preosteoblasts. Cells were treated with vehicle or 10 ^{M PTH (1–84) for 30 min. Membrane fractions were prepared for detection of axin1 and LRP6 levels by Western blotting analysis. (}C) LRP6–axin binding in HEK 293 cells. Cells were transfected with PTH1R, HA-tagged axin, and VSVG-tagged LRP6 and treated with 10 ^{M PTH (1–84). The axin-associated LRP6 was determined by Western blotting of the anti-VSVG immunoprecipitates. (}D) Fz8CRD does not inhibit PTH-induced LRP6 phosphorylation. HEK293 (for lanes 1–4) or MEF (for lanes 5–8) cells were transfected with VSVG-tagged LRP6, treated with control CM or Fz8CRD CM for 2 h, and then with 10 ^{M PTH (1–84) or Wnt3a. The phosphorylated LRP6 was detected by Western blotting analysis of the anti-VSVG immunoprecipitates with Ab1490. (}E) UMR-106 cells were serum deprived, treated with control CM or LRP6N CM for 1 h, and then with 10 ^{M PTH (1–84) for another 15 min. Phosphorylated endogenous LRP6 was identified by Western blotting analysis of the anti-LRP6 immunoprecipitates with Ab1490. (}F,G) Immunohistochemical analysis of phosphorylated LRP6 levels in femur sections from 5-mo-old male rats at the indicated time points after PTH (1–34) injection (40 μg/kg). Representative of sections immunohistochemically stained with antibody to total LRP6 (top row) and phosphorylated LRP6 (Ab1490, middle and bottom rows), and counterstained with hematoxylin viewed at lower power (middle row) and higher power (bottom row). The metaphyseal area of distal femurs was examined. (F) Double asterisks mark locations in the low-power images that are shown in the high-power fields below. The phosphorylated LRP6-positive osteoblasts were counted in a blinded fashion using OsteoMeasure Software (OsteoMetrics, Inc.) from three random high-power fields per specimens at metaphysis subjacent to diaphyseal hematopoietic bone marrow, and a total of six specimens in each group were used. (G) The quantification of phosphorylated LRP6-positive osteoblasts is presented as percentage of total osteoblasts. (*) P < 0.005; (**) P < 0.001 (in comparison with control), n = 6.

Because the levels of β-catenin were increased in osteoblasts of rats with injection of a single dose of PTH (Fig. 1E,F), we tested whether the amounts of phosphorylated LRP6 were enhanced in osteoblasts from the same tissue. Immunostaining with an antibody specific for the phosphorylated PPPSP demonstrated that PTH-stimulated phosphorylation of LRP6 in preosteoblasts or osteoblasts at the surface of trabecular bone (Fig. 4F, second and third rows), whereas the amount of total LRP6 protein remained unchanged (Fig. 4F, first row). The temporal pattern of phosphorylation of the PPPSP motifs was similar to that of β-catenin (cf Figs. 1E, ​,4F4F [second and third rows] and Figs. 1F, ​,4G).4G). Thus, PTH increases the abundance of β-catenin in osteoblasts in vivo through phosphorylation of LRP6.

PKA is required in PTH-, but not in Wnt-activated LRP6–β-catenin signaling

As the activation of β-catenin signaling by PTH in osteoblasts seems to be independent of Wnt, we attempted to investigate the mechanism responsible for the PTH effects. PTH activates cAMP-dependent PKA, which is sufficient for initiation of signals mediating PTH action in osteoblasts. We assessed whether PKA participates in PTH-activated LRP6–β-catenin signaling. Binding of intact PTH (1–84) or PTH (1–34) to PTH1R activates PKA. However, the native C-terminal fragments of PTH bind PTH1R but do not activate PKA (Kronenberg et al. 1998; Gensure et al. 2005; Murray et al. 2005). The C-terminal fragments PTH (7–84) and PTH (39–84) were much less effective than PTH (1–84) in activating β-catenin signaling (Fig. 5A), altering the stability of β-catenin (Fig. 5B), and inducing axin–LRP6 binding (Fig. 5C). These results suggest that cAMP–PKA activation is involved in activation of LRP6. The minimum effects induced by PTH (7–84) and PTH (39–84) (Fig. 5A–C) may be mediated by other signaling components than PKA as the unrelated peptide of the similar length did not exert such effect (data not shown). PKI-(14–22), a specific inhibitor of PKA-directed phosphorylation, inhibited PTH-induced LRP6 phosphorylation (Fig. 5D). Moreover, the PKA inhibitors, PKI-(14–22) and H89 reduced the binding of axin to LRP6 (Fig. 5E), β-catenin stabilization (Fig. 5F, lane 3), and β-catenin-dependent transcription activity (Fig. 5G), further indicating that PKA activity is essential for PTH-activated LRP6–β-catenin signaling. However, H89 did not affect Wnt3a-stimulated LRP6 phosphorylation (Fig. 5H), β-catenin stabilization (Fig. 5F, lane 5), and β-catenin-dependent transcription activity (Fig. 5I). The results provide further evidence to support the concept that both PTH and Wnt activate LRP6–β-catenin signaling in osteoblasts, but do so through distinct pathways.

Figure 5.

PKA mediates PTH-, but not Wnt3a-activated LRP6–β-catenin signaling. (A) Failure of PTH C-terminal ligands to stimulate TCF4/LEF activity. UMR-106 cells were transfected with TCF4-Luc plasmid, treated with 10 ^{M PTH ligands and subject to a luciferase assay. (*)} P < 0.05 (in comparison with control), n = 3; (**) P < 0.005 (in comparison with control), n = 3. (B) Failure of PTH C-terminal ligands to stabilize β-catenin. UMR-106 cells were treated with PTH ligands for 1 h. Cytosolic fractions were prepared for detection of β-catenin protein by Western blot analysis. (C) Effect of PTH C-terminal ligands on LRP6–axin binding. Cells were transfected with indicated plasmids and treated with PTH ligands. The LRP6-associated axin was determined by Western blotting of the anti-VSVG immunoprecipitates. (WCL) Whole-cell lysates. (D) Inhibition of PTH-induced LRP6 phosphorylation by PKA inhibitor. Cells were transfected with VSVG-LRP6 and PTH1R, metabolically labeled with [^{P] phosphate, and treated with PKI-(14–22) for 1 h before adding PTH (1–34) for another 15 min. VSVG-LRP6 was immunoprecipitated from cell lysates. Proteins were resolved by SDS-PAGE and visualized by autoradiography. (}E) Inhibition of the binding of axin with LRP6 by PKA inhibitors. Cells were transfected with indicated plasmids and pretreated with H89 or PKI-(14–22) for 1 h before adding PTH (1–84) for another 30 min. The LRP6-associated axin was determined as in C. (F) Inhibition of PTH-induced, but not Wnt3a-induced β-catenin stabilization by PKA inhibitor. UMR-106 cells were treated with 10 ^{M PTH (1–34) or 50 ng/mL recombinant mouse Wnt3a (rmWnt3a) together with vehicle (control) or H89. Cytosolic fractions were prepared for detection of β-catenin protein by Western blot analysis. (}G) Inhibition of PTH-activated TCF4/LEF activity by PKA inhibitors. UMR-106 cells were transfected with TCF4-Luc plasmid, treated with 10 ^{M PTH together with vehicle (control), H89, or PKI-(14–22), and subject to a luciferase assay. (*)} P < 0.005 (in comparison with PTH treatment group), I = 3. (H) Failure of PKA inhibitor to inhibit Wnt3a-induced LRP6 phosphorylation. MEFs were transfected with VSVG-LRP6 and pretreated with increasing doses of H89 for 1 h before adding control CM or Wnt3a CM for another 30 min. The phosphorylated LRP6 was detected by Western blotting analysis of the anti-VSVG immunoprecipitates with Ab1490. (I) Failure of PKA inhibitors to inhibit Wnt3a-stimulated TCF4/LEF activity. UMR-106 cells were transfected with TCF4-Luc plasmid, treated with control CM (Con) or Wnt3a CM (Wnt3a) together with vehicle, H89, or PKI-(14–22) and subject to a luciferase assay. (n.s.) No significance (in comparison with Wnt3a group), n = 3. (J) Schematic model of LRP6–β-catenin pathway activation in response to PTH. Upon PTH ligand binding to its receptor PTH1R, LRP6 are recruited and form complexes with PTH/PTH1R, thus positioning LRP6 in close proximity with PTH1R. In parallel, PKA is activated downstream from PTH1R and mediate the phosphorylation of LRP6, which leads to the recruitment of axin and stabilization of β-catenin.

Discussion

The data reported here demonstrate that PTH activates β-catenin signaling in osteoblasts in vitro and in vivo by direct recruitment of LRP6 to PTH/PTH1R complex. PTH ligand induces direct interaction of the extracellular domain of LRP6 with PTH1R at the cell surface. As a result, the PPPSP motifs of LRP6 are phosphorylated and axin is recruited to the phosphorylated PPPSP motifs, leading to stabilization of β-catenin (Fig. 5J). The extracellular domain of LRP6, which inhibits LRP6–PTH1R complex formation, blocks LRP6 phosphorylation and β-catenin activation. In addition, DKK1 and sclerostin, antagonists of LRP6 by direct binding to the extracellular domain of LRP6, inhibit PTH-induced axin–LRP6 binding (Supplemental Fig. 3). Thus, PTH-induced recruitment of LRP6 to PTH1R is essential in activation of β-catenin signaling. This PTH-activated LRP6–β-catenin pathway is likely a direct effect rather than via a Wnt ligand-dependent process because the phosphorylation of LRP6 by PTH is rapid. In addition, Fz8CRD, the competitive inhibitor of the Wnt receptor Fz, is not able to block the action of PTH.

LRP6 is a well-recognized coreceptor for Wnt, and its function in osteoblasts has been considered primarily in terms of its effect on Wnt signaling (Baron et al. 2006; Balemans and Van 2007; Glass and Karsenty 2007). It would be of considerable interest to further determine the roles of LRP6 and β-catenin in PTH-mediated effects on the osteoblasts function. We found that knockdown of LRP6 blocked PTH-stimulated phosphorylation of CREB at Ser ^{and ERK (Supplemental Fig. 4), the known direct PKA or PKC targets (}Tyson et al. 1999; Gesty-Palmer et al. 2006), and importantly, inhibited PTH-stimulated mRNA expression of osteocalcin and RANKL, osteoblast differentiation markers (Fig. 2D,E). Moreover, we analyzed mouse models of long-term administration of PTH, which has paradoxical effects in that PTH stimulates bone formation when injected daily, but causes severe bone loss if continually infused (Supplemental Fig. 5A,B). Intermittent, but not continuous, administration of PTH causes phosphorylation of LRP6 and stabilization of β-catenin in osteoblasts (Supplemental Fig. 5C). In addition to stabilization of β-catenin by PTH and Wnts, TGFβ also stimulates proliferation of osteoprogenitors through stabilization of β-catenin (Jian et al. 2006). Thus, it appears that β-catenin is one of the common mediators of osteoblastic bone formation induced by different extracellular signals.

PTH is also known to inhibit osteocyte expression of sclerostin (Bellido et al. 2005; Keller and Kneissel 2005; Leupin et al. 2007), which is an inhibitor of Wnt-activated β-catenin signaling via directly binding to LRP6 and a negative regulator of bone formation (Winkler et al. 2003; van Bezooijen et al. 2004; Li et al. 2008). Sclerostin also inhibits PTH-induced axin–LRP6 binding, implying that suppression of sclerostin by PTH would increase the availability of LRP6 to facilitate PTH signaling in a positive feedback fashion. The activation of PKA by PTH may occur in parallel to mediate the phosphorylation of LRP6 and axin recruitment for activation of β-catenin (Fig. 5J). Sequence analysis reveals four PKA consensus sites in the cytoplasmic domain of LRP6. Our results demonstrate that LRP6 can be directly phosphorylated by PKA catalytic subunit (Supplemental Fig. 6), and PKA inhibitor H89 inhibited PTH-induced phosphorylation of LRP6 as well as its binding to axin (Fig. 5D,E). These data suggest that the functional binding of LRP6 to axin is PKA phosphorylation-dependent. Therefore, both recruitment of LRP6 to PTH1R and phosphorylation of LRP6 by PKA are involved in PTH-induced stabilization of β-catenin. The fact that PTH-stimulated bone formation can still occur in LRP5-deficient mice establishes that LRP5 alone is not essential for the stimulatory effects of PTH on bone formation (Sawakami et al. 2006; Iwaniec et al. 2007). In our studies, PTH directly phosphorylates the intracellular domain of LRP6, but not LRP5 (Supplemental Fig. 6). Forskolin, a PKA agonist, induced binding of axin to LRP6, but not to LRP5 (Supplemental Fig. 7), further indicating that LRP6 may act differently from LRP5 in PTH actions in osteoblasts. The distinct roles of these coreceptors in PTH function and whether their mutations cause bone defects by altering PTH signaling remain to be investigated.

Materials and methods

FRET procedure

PTH1R and BMPRII cDNAs were cloned into ECFP-N1, and LRP6 and mLRP4T100 cDNAs were cloned into EYFP-N1 (Clontech) expression vectors. These vectors were modified by site-directed mutagenesis that prevents the self-dimerization (Bhatia et al. 2005). CFP and YFP were fused at the C termini of the receptors. Because CFP-PTH1R or CFP-BMPRII (the fluorescent FRET donors) is quenched when in the proximity of YFP-LRP6 or YFP-mLRP4T100 (the acceptors), FRET efficiency can be measured by comparing donor fluorescence pre- and post-photobleaching of the acceptor. An increase in donor fluorescence after acceptor photobleaching indicates that donor and acceptor fluorophores were within FRET range. HEK293 cells on coverslips in 35-mm dishes were cotransfected with 0.1 μg of each plasmid. Cells were observed using Leica TCS SP II AOBS laser-scanning confocal microscope. An excitation wavelength of 405 nm and an emission range of 416–492 nm, and an excitation wavelength of 514 nm and an emission range of 525–600 nm were used to acquire images of CFP and YFP, respectively. YFP was photobleached by using full power of the 514 nm line for 1–2 min. An image of CFP and YFP fluorescence after photobleaching was obtained by using the respective filter sets. Images were representatives of three experiments. The FRET efficiencies were calculated according to

cDNA constructs

PTH1R tagged with HA was subcloned into pCDNA3.1. cDNAs from human LRP5 (J. Mao et al. 2001) and LRP6 (Tamai et al. 2000) tagged with HA and VSVG were subcloned into pCMV and pCS2^{, respectively. LRP6N + T (LRP6 N-terminal plus the transmembrane domain) and LRP6T + C (LRP6 transmembrane domain plus C-terminal) tagged with VSVG were subcloned into pCS2 + . LRP6N + 1479m, LRP6N + 1490m, LRP6N + 1493m, and LRP6N + 1496m were generated by mutagenesis of either the serine (at aa1490 or aa1496) or threonin (at aa1479 or aa1493) to alanine. LRP6N-IgG was generated by fusing the LRP6 extracellular domain with IgG (}Tamai et al. 2000). si-GFP (Wan et al. 2005) and si-LRP6 plasmids were generated using a BS/U6 vector. Briefly, a 22-nucleotide (nt) oligo (oligo 1) corresponding to nucleotides 2981–3002 of the human LRP6 coding region was first inserted into the BS/U6 vector digested with ApaI (blunted) and HindIII. The inverted motif that contains the 6-nt spacer and five Ts (oligo 2) was then subcloned into the HindIII and EcoRI sites of the intermediate plasmid to generate BS/U6/LRP5/6.

Primary osteoblast isolation and culture

Osteoblasts were isolated by digestion of calvaria of newborn mice as decribed (Wang et al. 2007). Briefly, calvaria were incubated with 10 mL of digestion solution containing 1.8 mg/mL of collagenase type I (Worthington Biochemical Corp.) for 15 min at 37°C under constant agitation. The supernatant was then harvested, replaced with fresh collagenase, and the digestion repeated an additional four times. Digestion solutions containing the osteoblasts were pooled together. After centrifugation, osteoblasts were obtained and cultured in α-MEM containing 10% FBS, and 1% penicillin/streptomycin at 37°C in a humidified incubator supplied with 5% CO₂.

Cell culture, conditioned media, transfection, and luciferase reporter assays

HEK293, UMR-106, and mouse embryonic fibroblast (MEF) cells were maintained in DMEM with 10% FCS. Mouse Wnt3a conditioned medium (Wnt3a CM) was produced from mouse L cells stably transfected with mouse Wnt3a (American Type Culture Collection) and control conditioned medium (Control CM) was from nontransfected L cells. IgG, LRP6N-IgG, DKK1, Sclerostin, VSVG-LRP6N and Myc-Fz8CRD conditioned media were produced from HEK 293 cells transfected with the individual plasmids. Transfections were carried out using lipofectamine reagent (Invitrogen). Luciferase assays were carried out in either UMR-106 or HEK 293 cells as described previously (Wan et al. 2005), with 0.3 μg of TCF-Luc reporter plasmid plus 50 ng of Renilla luciferase plasmid (internal control) per well in the 12-well plate. Experiments were repeated at least three times with triplicate for each experiment.

Cell fractionation, co-IP, and Western blot analysis

Cells were harvested in cavitation buffer (5 mM HEPES at pH 7.4, 3 mM MgCl₂, 1 mM EGTA, 250 mM sucrose) containing protease and phosphatase inhibitors and homogenized by nitrogen cavitation (200 p.s.i., for 5 min) in a cell disruption bomb (Parr Instrument Co.). The cell homogenate was centrifuged twice at 700g for 10 min to pellet the nuclei. The supernatant was further centrifuged at 100,000g (Beckman SW50.1 rotor) for 1 h to separate the membrane and cytosol fractions, and the resulting membrane pellet was washed three times with cavitation buffer before use in the assays (Zhang et al. 1999). IP and Western blot analysis of cell lysates were performed as described previously (Wan et al. 2005).

Cell suface binding by immunofluorescence colocalization assay

Cells were transfected with HA-PTH1R and treated with IgG conditioned medium or LRP6N-IgG conditioned medium for 1 h followed by PTH (1–84) treatment for 15 min. Cells were then washed with PBS, fixed with 4% paraformaldehyde, permeabilized with 0.1% Triton X-100, and incubated with primary antibody followed by incubation with chromophore-conjugated secondary antibody. Digital pictures were taken using an Olympus, IX TRINOC camera fitted to an Olympus, IX70 Inverted Research Microscope (Olympus) with objective lenses of Hoffman Modulation Contrast, HMC 10 LWD PL FL, 0.3NA ∞/1, (OPTICS, Inc.) at room temperature, and processed using MagnaFire SP imaging software (Optronics). A Zeiss TCs SP2 system was used for confocal imaging. The ratios of the number of cells showing in green to the number of cells showing in red were calculated. For each treatment, 100 cells on each of three different slides were analyzed.

Metabolic ^{P labeling and in vivo phosphorylation assays}

Cells were transfected with expression plasmids and were washed twice with phosphate-free DMEM containing 2% dialyzed fetal calf serum, incubated in the same medium for 4 h, and then labeled with 1 mCi/mL [^{P]orthophosphate (Perkin-Elmer) for an additional 2 h. The P-labeled cells were then washed with ice-cold PBS and lysed with radioimmunoprecipitation assay buffer. VSVG-LRP6 was immunoprecipitated with anti-VSVG, and the resultant precipitates were separated by 8.5% SDS-PAGE. Gels were dried and exposed to Biomax} M_r or MS film (Eastman Kodak Co.). After autoradiographic analysis, dried gels were rehydrated with transfer buffer, and transferred onto PVDF membranes. For equal loading confirmation, the transfected VSVG-LRP6 was visualized by the ECL ^{Western blotting detection system (Amersham Biosciences).}

Quantitative real-time PCR

Cells were homogenized using Trizol reagent (Invitrogen), and total RNA was extracted according to the manufacturer's protocol. cDNA was produced and quantitative real-time PCR were performed in an iCycler real-time PCR machine using iQ SYBR Green supermix (Bio-Rad). Primers are as follows: GAPDH (forward, 5′-GGGTGTGAACCACGAGAAAT-3′; reverse, 5′-CCT TCCACAATGCCAAAGTT-3′), Osteocalcin (forward, 5′-CT TGGTGCACACCTAGCAGA-3′; reverse, 5′-CTCCCTCATGT GTTGTCCCT-3′), and RANKL (forward, 5′- CCAAGATCTC TAACATGACG-3′; reverse, 5′-CACCATCAGCTGAAGATA GT-3′). The quantity of RANKL and Osteocalcin mRNA in each sample was normalized using the C_T (threshold cycle) value obtained for the GAPDH mRNA amplifications.

FRET procedure

PTH1R and BMPRII cDNAs were cloned into ECFP-N1, and LRP6 and mLRP4T100 cDNAs were cloned into EYFP-N1 (Clontech) expression vectors. These vectors were modified by site-directed mutagenesis that prevents the self-dimerization (Bhatia et al. 2005). CFP and YFP were fused at the C termini of the receptors. Because CFP-PTH1R or CFP-BMPRII (the fluorescent FRET donors) is quenched when in the proximity of YFP-LRP6 or YFP-mLRP4T100 (the acceptors), FRET efficiency can be measured by comparing donor fluorescence pre- and post-photobleaching of the acceptor. An increase in donor fluorescence after acceptor photobleaching indicates that donor and acceptor fluorophores were within FRET range. HEK293 cells on coverslips in 35-mm dishes were cotransfected with 0.1 μg of each plasmid. Cells were observed using Leica TCS SP II AOBS laser-scanning confocal microscope. An excitation wavelength of 405 nm and an emission range of 416–492 nm, and an excitation wavelength of 514 nm and an emission range of 525–600 nm were used to acquire images of CFP and YFP, respectively. YFP was photobleached by using full power of the 514 nm line for 1–2 min. An image of CFP and YFP fluorescence after photobleaching was obtained by using the respective filter sets. Images were representatives of three experiments. The FRET efficiencies were calculated according to

Xenopus embryo manipulation

RNAs for microinjection were synthesized using SP6 mMessage mMachine in vitro transcription kit (Ambion). RNAs were injected into the marginal zone region of two ventral blastomeres of four-cell stage embryos, and the phenotype of the embryos was observed at the tadpole stages. The doses of RNAs used were 200 pg LRP6, 2 pg PTH, and 50 pg PTH1R.

Animals

The experimental protocol was reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) of University of Alabama at Birmingham. For the experiments in which rats or mice were administered PTH as single-dose injection, 5-mo-old male Sprague Dawley rats (Charles River Laboratories) or 2-mo-old male C57BL/J6 mice (The Jackson Laboratory) (six per group) were administered a single dose of either vehicle (1 mM acetic acid in sterile PBS) or PTH (1–34) (Bachem, Inc.) at 40 μg/kg in a volume of 100 μl. In the mouse model, mouse recombinant Wnt3a (R&amp;D Systems) was injected at 25 μg/kg in a volume of 100 μL. All treatments were through bolus intravenous injection via the tail vein. Rats/mice were sacrificed at 0.5, 2, 8, and 24 h after injection.

Immunohistochemical analysis of the bone tissue

Formalin-fixed femur or tibia tissue sections of 5 μm thickness from rats or mice were processed with antigen retrieval and hydrogen peroxide treatment prior to incubation with primary monoclonal antibody specific for β-catenin (BD Biosciences), goat polyclonal antibody sclerostin (R&amp;D Systems), or rabbit polyclonal phosphorylated LRP6 (Ab1490) for 1 h at room temperature or overnight at 4°C. Negative controls were obtained by replacing the primary antibodies with irrelevant control isotype IgG. Antibody detection was accomplished using the biotin-streptavidin horseradish peroxidase (for β-catenin and sclerostin) or alkaline phosphatase (for Ab1490) (EnVision System, Dako). β-Catenin and sclerostin staining was based on peroxidase (HRP) using DAB as chromogen. Phospho-LRP6 staining was based on alkaline phosphotase (AP) using Permanent Red as chromogen. The sections were then counterstained with hematoxylin. Isotype-matched negative control antibodies (R&amp;D Systems) were used under the same conditions. Osteoblasts/preosteoblasts were observed at the bone surface with large, spherical, and basal mononucleus. Only those specimens in which >10% of the cells were stained were considered as positive. In the rat model, numbers of total osteoblasts and numbers of β-catenin- or p-LRP6-positive osteoblasts were counted in three random high-power fields at metaphysis subjacent to the epiphyseal growth plates or the diaphyseal hematopoietic bone marrow per specimen, and a total of six specimens in each group were used. In the mouse model, numbers of total osteoblasts and numbers of β-catenin-positive osteoblasts were counted in three random high-power fields in a 2-mm square, 1 mm distal to the lowest point of the growth plate in the secondary spongiosa. Numbers of total osteocytes and numbers of sclerostin-positive osteocytes were counted in three random high-power fields per trabecular bone section or cortical bone section, and a total of six specimens in each group were used.

Statistical analysis

Data were analyzed using Student’s t-test and are expressed as the mean ± SEM.