The bone formation inhibitor sclerostin encoded by SOST binds in vitro to low-density lipoprotein receptor-related protein (LRP) 5/6 Wnt co-receptors, thereby inhibiting Wnt/β-catenin signaling, a central pathway of skeletal homeostasis. Lrp5/LRP5 deficiency results in osteoporosis-pseudoglioma (OPPG), whereas Sost/SOST deficiency induces lifelong bone gain in mice and humans. Here, we analyzed the bone phenotype of mice lacking Sost (Sost−/−), Lrp5 (Lrp5−/−), or both (Sost−/−;Lrp5−/−) to elucidate the mechanism of action of Sost in vivo. Sost deficiency–induced bone gain was significantly blunted in Sost−/−;Lrp5−/− mice. Yet the Lrp5 OPPG phenotype was fully rescued in Sost−/−;Lrp5−/− mice and most bone parameters were elevated relative to wild-type. To test whether the remaining bone increases in Sost−/−;Lrp5−/− animals depend on Lrp6, we treated wild-type, Sost−/−, and Sost−/−;Lrp5−/− mice with distinct Lrp6 function blocking antibodies. Selective blockage of Wnt1 class–mediated Lrp6 signaling reduced cancellous bone mass and density in wild-type mice. Surprisingly, it reversed the abnormal bone gain in Sost−/− and Sost−/−;Lrp5−/− mice to wild-type levels irrespective of enhancement or blockage of Wnt3a class-mediated Lrp6 activity. Thus, whereas Sost deficiency–induced bone anabolism partially requires Lrp5, it fully depends on Wnt1 class–induced Lrp6 activity. These findings indicate: first, that OPPG syndrome patients suffering from LRP5 loss-of-function should benefit from principles antagonizing SOST/sclerostin action; and second, that therapeutic WNT signaling inhibitors may stop the debilitating bone overgrowth in sclerosing disorders related to SOST deficiency, such as sclerosteosis, van Buchem disease, and autosomal dominant craniodiaphyseal dysplasia, which are rare disorders without viable treatment options. © 2014 American Society for Bone and Mineral Research.
Wnt/β-catenin signaling/canonical Wnt signaling plays a key role in many physiological and pathological conditions. In bone, Wnt/β-catenin signaling is required for commitment of mesenchymal stem cells to the osteogenic lineage, controls osteoblast precursor proliferation, differentiation, and survival, but also controls osteoclastic bone resorption.[1-4] It is initiated upon binding of secreted Wnt ligands to a dual-receptor complex formed by the frizzled (Fzd) receptor and the low-density lipoprotein receptor-related protein 5 or 6 (LRP5/6), triggering a cascade of downstream signaling events, which results in accumulation of the central Wnt signaling effector β-catenin and Wnt target gene expression.
Consistent with a central role of Wnt/β-catenin signaling in bone metabolism, LRP5 loss-of-function in humans causes the autosomal recessive osteoporosis-pseudoglioma (OPPG) syndrome characterized by ocular abnormalities and low bone mass due to decreased bone formation. This phenotype is recapitulated in Lrp5-deficient (Lrp5−/−) mice.[6, 7] Similarly, mutations in Lrp6/LRP6 as well as osteoblast-specific deletion of Lrp5 or Lrp6 have been found to lead to reduced bone mass and osteoporosis in mice and humans.[6, 8-10] Conversely, autosomal dominant gain-of-function mutations in Lrp5/LRP5 lead to high bone mass (HBM) in mice and humans as a result of increased bone formation.[11-13]
Loss-of-function mutations in genes encoding the secreted Wnt signaling antagonists Dickkopf1 (DKK1/Dkk1) and SOST/sclerostin also result in HBM-like phenotypes. Mutations in SOST cause the severe bone overgrowth disorders sclerosteosis,[15, 16] van Buchem disease (VBD),[17-19] and autosomal dominant craniodiaphyseal dysplasia (CDD). Likewise, Sost loss-of-function (Sost−/−) mice show elevated osteoblastic bone formation causing abnormal bone accumulation. Intriguingly, sclerostin and DKK1 binding to LRP5 is diminished in LRP5 gain-of-function mutants, providing a mechanistic link for the observed HBM phenotypes.[12, 22-25]
Sclerostin is a secreted glycoprotein, which in vitro causes inhibition of Wnt/β-catenin signaling[24, 26-28] by binding to the first of four YWTD-type β-propeller domains in the extracellular domain of LRP5/6, thereby disrupting the formation of the Wnt ligand-receptor complex.[29-31] In addition, a facilitator of sclerostin's inhibitory action on Wnt/β-catenin signaling, the LRP family member LRP4, was recently identified. Like SOST deficiency, loss-of-function of LRP4 results in bone overgrowth in humans. Given the crucial role of sclerostin as a negative bone-formation regulator, pharmacological interventions blocking its activity are under investigation to develop novel therapeutic agents for treatment of osteoporosis. Understanding SOST/sclerostin mode of action in vivo is thus of great interest to develop novel therapies for skeletal disorders. In particular, it remains to be shown in vivo whether SOST/sclerostin targets LRP5 and/or LRP6 or also interacts with other signaling molecules to exert its function as a bone-formation inhibitor.
To elucidate the in vivo mechanism of action of Sost, we generated Sost and Lrp5 double-deficient mice (Sost−/−;Lrp5−/−) and analyzed their bone phenotype relative to Sost−/− and Lrp5−/− single-mutant and wild-type littermates. Furthermore, we treated adult wild-type, Sost−/− and Sost−/−;Lrp5−/− mice with distinct Lrp6 function blocking anti-Lrp6 antibodies to assess the in vivo relevance of Lrp6 for Sost deficiency–induced bone anabolism. We provide some surprising insights into the mechanism of Wnt signaling in bone, implicating novel therapeutic opportunities for patients with OPPG and rare sclerosing bone disorders through use of selective Wnt signaling modulators.
Subjects and Methods
Sost−/− and Lrp5−/− mice were previously described and were crossed to generate wild-type, Sost−/−, Lrp5−/−, and Sost−/−;Lrp5−/− mice of mixed 129/Sv;C57BL/6 genetic background. For anti-Lrp6 antibody treatment experiments, C57BL/6J wild-type (Charles River Laboratories, Sulzfeld, Germany) and Sost−/− mice, which had been backcrossed to C57BL/6J genetic background, were used. All mice were kept in cages under standard laboratory conditions at 22°C and a 12-hour/12-hour light/dark cycle. Mice were fed a standard rodent diet containing 1.15% calcium, 0.85% phosphorus, and 1000 IU/kg vitamin D3 (3302; Provimi Kliba SA, Kaiseraugst, Switzerland) with food and water provided ad libitum. Procedures conformed to the Swiss federal law for animal protection controlled by the Basel-Stadt Cantonal Veterinary Office, Switzerland.
For longitudinal analyses, animals were placed laterally under inhalation narcosis (isoflurane, 2.5%) and the left leg was stretched and fixed. Cross-sectional bone parameters were measured in the proximal tibia metaphysis at monthly to bimonthly intervals. Ex vivo analyses were performed by measurement of six consecutive slices spaced equally along the axis of the left femur. An adapted Stratec-Norland XCT-2000 fitted with an Oxford (Oxford, UK) 50-µm X-ray tube (GTA6505M/LA) and a 0.5-mm-diameter collimator was used for all measurements with the following setup: voxel size = 0.07 mm × 0.07 mm × 0.4 mm; scan speed = 5 mm/s; 1 block; contour mode 1; peel mode 2; and cortical and inner thresholds = 350 mg/cm3.
Generation of three-dimensional (3D) bone reconstruction images and evaluation of bone structure parameters was accomplished by high-resolution ex vivo measurements using a vivaCT40 instrument (Scanco Medical AG, Brüttisellen, Switzerland; voxel size = 6 µm for anti-Lrp6 antibody experiments, otherwise voxel size = 10.5 µm). Thresholds of 225 for antibody experiments or 230 for genetic interaction studies were applied to determine the mineralized bone fraction. A Gaussian filter was used (σ = 0.7, support = 1) to remove noise. For antibody experiments, 60 or 65 vertebral and 60 or 80 femoral slices were evaluated; 85 slices were analyzed for genetic interaction studies.
Serum biomarker measurements
Animals were euthanized by terminal CO2 exposure and whole blood was collected by venipuncture for anti-Lrp6 antibody experiments or heart puncture for genetic interaction studies. Serum was prepared using clot activator centrifugation tubes (Sarstedt, Nümbrecht, Germany). Osteocalcin was quantified with an immunoradiometric assay kit (Immutopics, San Clemente, CA, USA) and tartrate-resistant acid phosphatase 5b (TRAP 5b) was determined by ELISA (Immunodiagnostic Systems, Boldon, UK). Serotonin was measured in serum from day-fasted animals by HPLC. Serum was suspended in 10 volumes (vol/wt) of 0.1 N perchloric acid/0.05% disodium EDTA/0.05% sodium metabisulfite and serotonin was extracted. Samples of 10-µL were applied onto a Beckman Ultrasphere 5-µm IP column (Beckman, Brea, CA, USA) and eluted serotonin was quantified electrochemically at 0.65 V.
Mice received fluorochrome markers applied subcutaneously 10 (20 mg/kg alizarin complexone; Merck, Whitehouse Station, NJ, USA) and 3 days prior to necropsy (30 mg/kg calcein; Fluka, Buchs, Switzerland). Bone samples were fixed for 24 hours in 4% phosphate-buffered paraformaldehyde, dehydrated and defatted at 4°C and embedded in methylmethacrylate. A set of 5-µm nonconsecutive longitudinal sections was cut in the frontal midbody plane per bone and animal. Fluorochrome marker–based dynamic bone parameters were determined using a Leica DM microscope fitted with a SONY DXC-950P camera and adapted Quantimet 600 software (Leica, Buffalo Grove, IL, USA). Microscopic images were digitized and evaluated semiautomatically on-screen (200-fold magnification) as described. Bone surface (BS), single-labeled bone surface (sLS) and double-labeled bone surface (dLS), and interlabel width were measured. Mineralizing surface [MS/BS = (dLS + sLS/2)/BS (%)] and mineral apposition rate (MAR) (corrected for section obliquity) were calculated and the daily bone formation rate (BFR) was derived. Osteoclast numbers per bone surface were measured on sections stained for TRAP 5b activity using a Merz grid. Bone histomorphometric nomenclature was applied as recommended by the American Society for Bone and Mineral Research (ASBMR).
Anti-Lrp6 antibody treatment
Three-month-old female mice were subjected to 1-month treatment with weekly intravenous injections of 10 mL/kg vehicle (PBS, pH 7.4; Invitrogen, Grand Island, NY, USA) or either of three anti-Lrp6 antibodies. Antibody A, corresponding to the described antibody A7-IgG, was administered once weekly at 3 mg/kg. Antibody B, equivalent to the reported antibody B2-IgG, was injected once weekly at 10 mg/kg. The biparatopic antibody A/B, corresponding to the published antibody A7/B2, was applied twice weekly at 3 mg/kg. These administered doses were chosen based on pharmacokinetics and pharmacodynamics studies to provide maximum exposure of these antibodies. Various doses were tested and the doses above were determined to be maximum efficacious doses for each antibody. Dosing schedule was chosen based on the half-life of these antibodies in circulation. Mice were assigned to treatment groups according to baseline cross-sectional total and cancellous bone mineral density (BMD) in the proximal tibia metaphysis as measured by pQCT as the first-rank and second-rank parameter, respectively, as well as body weight as the third-rank parameter to ensure comparable group means and standard variations at baseline to minimize experimental artifacts.
qPCR expression analyses
Total RNA isolated from cortical bone of femoral diaphyses by sequential enzymatic digestion as described was reverse transcribed with the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Grand Island, NY, USA). Gene expression was determined using an ABI Prism 7900HT sequence detection system, TaqMan Universal PCR Master Mix, and the following mouse TaqMan probes (Applied Biosystems): 18S (4310893E) for normalization; Dkk1 (Mm00438422_m1); Lrp5 (Mm01227479_m1); and Lrp6 (Mm00521783_m1).
All data represent the mean and SEM. Statistical analyses were performed with GraphPad Prism software using one-way analysis of variance (ANOVA) with Tukey's multiple comparisons post hoc test or two-way ANOVA followed by Sidak's or Tukey's multiple comparisons post hoc test as recommended by the software.
Reversal of Lrp5−/− osteopenia and blunting of Sost deficiency–induced bone gain in Sost−/−;Lrp5−/− mice
To assess whether Sost and Lrp5 genetically interact, we generated Sost;Lrp5 double-deficient mice. Sost−/−;Lrp5−/− mice were viable and fertile, showed no obvious phenotypical abnormalities and had a normal lifespan. Overall, body weight gain and tibial growth were comparable between wild-type, single-mutant, and double-mutant mice with no significant alterations compared to wild-type littermates (Supplemental Fig. S1A–D).
To determine the bone phenotypic consequences of Sost deletion in the Lrp5−/− genetic background we monitored bone gain in the proximal tibia by longitudinal pQCT during skeletal growth (2-month-old and 3-month-old mice), maturity (4-month-old mice), and aging (6-month-old mice). Sost−/− mice displayed progressive strong increases in cross-sectional total BMD and total bone mineral content (BMC) (about +25% to +115% over time) relative to wild-type (Fig. 1A–D). Lrp5−/− mice exhibited reduced cross-sectional total BMD and BMC compared to wild-type littermates during skeletal growth and maturity (about −10% to −30%; Fig. 1A–D). However, during aging, Lrp5 deficiency was not associated with significant differences relative to wild-type, with the exception of total BMC, which remained significantly decreased in females by 34% (Fig. 1A–D). Sost deficiency fully rescued the Lrp5−/− osteopenic phenotype to the extent that at most ages analyzed Sost−/−;Lrp5−/− mice had higher BMD and BMC than wild-type littermates (about +10% to +50% over time; Fig. 1A–D). However, the relative Sost deficiency–induced bone gain in the Lrp5−/− genetic background was smaller in Sost−/−;Lrp5−/− mice than in wild-type (about +25% to +85% over time compared to Lrp5−/− littermates; Fig. 1A–D), indicating blunting of Sost deficiency–induced bone anabolism. The overall changes in total BMD in mutant mice relative to wild-type animals are outlined in a summarizing table (Fig. 1E).
To determine bone microstructural changes in the axial skeleton, we performed micro–computed tomography (µCT) of lumbar vertebra L3 of 6-month-old mice. Sost−/− mice displayed elevated vertebral BMD and relative bone volume compared to wild-type (about +85% to +225%; Fig. 2A, B). Consistent with the minor differences in the tibia during aging, aged Lrp5−/− mice showed only nonsignificant reductions in vertebral BMD and BV/TV relative to wild-type (−7% to −10%; Fig. 2A, B). These were not only fully reverted in Sost−/−;Lrp5−/− mice, but also were increased compared to wild-type (about +45% to +115%; Fig. 2A, B). Conversely, compared to Sost−/− mice, combined deficiency of Sost and Lrp5 resulted in blunting of the relative Sost deficiency-induced increases in vertebral bone density and mass (about +55% to +130% relative to Lrp5−/− mice; Fig. 2A, B). This reduction in net cancellous bone gain was related to a reduction in the relative Sost deficiency–induced increase in trabecular number, but not thickness (Fig. 2C, D).
We next analyzed the distal femur using pQCT. Again, the Lrp5−/− osteopenic phenotype was completely rescued, but there was a dramatic reduction of the relative Sost deficiency–induced bone gain in the cancellous and cortical bone compartment in Sost−/−;Lrp5−/− mice (Fig. 2E–G; Table 1). Although slightly less pronounced, overall comparable findings were present in femora of 3-month-old mice (Table 1).
|Bone structure parameter||Bone slice||Females||Males|
|Cn.BMD (mg/cm3)||S1||166.9 ± 17.2||437.5 ± 28.8**||113.6 ± 13.6++||212.1 ± 17.2++,xx||283.6 ± 33.5||395.9 ± 35.8||176.1 ± 17.7++||363.1 ± 33.4xx|
|S2||86.1 ± 5.7||479.5 ± 51.0**||39.5 ± 2.4++||51.7 ± 4.1++||160.8 ± 45.7||382.8 ± 33.3**||55.4 ± 5.3*,++||130.2 ± 13.0++|
|Ct.Th (mm)||S1||0.35 ± 0.02||0.74 ± 0.03**||0.30 ± 0.02++||0.44 ± 0.02++,xx||0.52 ± 0.07||0.76 ± 0.06*||0.32 ± 0.03*,++||0.65 ± 0.05xx|
|S2||0.28 ± 0.01||0.65 ± 0.05**||0.24 ± 0.00++||0.33 ± 0.01++,xx||0.36 ± 0.05||0.60 ± 0.03**||0.26 ± 0.01++||0.37 ± 0.01++,x|
|S4||0.31 ± 0.00||0.49 ± 0.01**||0.29 ± 0.00**,++||0.40 ± 0.00**,++,xx||0.36 ± 0.02||0.54 ± 0.01**||0.31 ± 0.00++||0.46 ± 0.01**,++,xx|
|Cn.BMD (mg/cm3)||S2||67.3 ± 8.0||611.7 ± 86.8**||45.0 ± 5.1++||81.2 ± 15.3++||121.9 ± 36.2||393.9 ± 55.5**||46.6 ± 4.9++||94.6 ± 22.7++|
|Ct.Th (mm)||S2||0.34 ± 0.01||1.00 ± 0.04**||0.30 ± 0.00++||0.42 ± 0.01++,x||0.32 ± 0.01||0.71 ± 0.06**||0.28 ± 0.00++||0.41 ± 0.02++|
|S4||0.35 ± 0.01||0.69 ± 0.01**||0.32 ± 0.00++||0.46 ± 0.02**,++,xx||0.34 ± 0.01||0.69 ± 0.01**||0.33 ± 0.00++||0.51 ± 0.02**,++,xx|
Blunting of Sost deficiency–induced increases in osteoblast function in absence of Lrp5
To determine the underlying mechanism of the reduced relative Sost deficiency–induced bone gain in Sost−/−;Lrp5−/− mice, we measured serum bone turnover markers and performed bone histomorphometry in 6-month-old mice. The bone resorption marker TRAP 5b was not significantly altered between genotypes (Table 2). Similarly, analysis of lumbar vertebra L3 did not reveal significant differences in osteoclast number between genotypes (Table 2). There was, however, a small decrease in Sost−/− females, which was also present in Sost−/−;Lrp5−/− mice of either sex. Consistent and more pronounced changes were observed in the distal femur metaphysis, indicating decreased bone resorption in aged Sost−/− and Sost−/−;Lrp5−/− mice relative to wild-type (Table 2).
|Bone resorption parameter||Females||Males|
|Serum TRAP 5b (U/L)||8.5 ± 1.1||8.1 ± 0.5||8.3 ± 0.4||7.5 ± 0.8||2.6 ± 0.3||3.4 ± 0.5||2.4 ± 1.0||2.6 ± 0.3|
|Oc.N/BS (1/mm)||3.1 ± 0.1||2.9 ± 0.3||3.8 ± 0.5||2.5 ± 0.3||2.7 ± 0.0||2.8 ± 0.2||3.1 ± 0.6||2.3 ± 0.2|
|Oc.N/BS (1/mm)||15.7 ± 2.5||2.9 ± 0.2**||14.0 ± 3.7+||6.2 ± 0.5*||9.9 ± 2.5||4.4 ± 0.7*||5.8 ± 1.2||5.3 ± 0.5*|
Next, we analyzed serum levels of the bone formation marker osteocalcin. Consistent with elevated bone formation, osteocalcin was increased in Sost−/−, whereas it was not significantly different in Lrp5−/− or Sost−/−; Lrp5−/− mice relative to wild-type littermates (Fig. 3A), suggesting that Sost deficiency–associated elevated bone formation might be normalized in Sost−/−;Lrp5−/− mice. We thus determined bone formation parameters in the distal femur and lumbar spine. Sost−/− mice displayed increased bone formation rates (BFRs) in the femur (females: +38%; males: +79%), whereas Lrp5 deficiency resulted in nonsignificantly altered BFR (females: −22%; males: +4%) compared to wild-type (Fig. 3B). Consistent with the CT results, the relative Sost deficiency–induced elevation in BFR was blunted in Sost−/−;Lrp5−/− mice (females: +26%; males: +12% relative to Lrp5−/− littermates; Fig. 3B). This was related to a loss of the relative Sost deficiency–induced increase in osteoblast function as reflected by higher MAR in Sost−/− compared to wild-type littermates (females: +17%; males: +35%), which remained unchanged in Sost−/−;Lrp5−/− mice (females: −4%; males: +3% relative to Lrp5−/− mice; Fig. 3C, D). Overall, similar observations were made in the lumbar spine (Fig. 3E–G).
Because sclerostin is known to not only bind to Lrp5, but also to the related Wnt co-receptor Lrp6, which might partially compensate for Lrp5 deficiency, we determined Lrp5/6 expression in femoral cortical bone using qPCR. Consistent with previously reported residual expression in Lrp5−/− mice resulting in expression of a truncated peptide, Lrp5 expression was still detectable in Lrp5−/− and Sost−/−;Lrp5−/− mice (Supplemental Fig. S2A). Interestingly, we observed a mild upregulation of Lrp6 in Lrp5−/− (females: +68%; males: +42%) and Sost−/−;Lrp5−/− mice (females: +40%; males: +33%) relative to wild-type (Supplemental Fig. S2B), whereas Lrp5/6 expression was unchanged in Sost−/− littermates (Supplemental Fig. S2).
Reduction of Sost deficiency–induced bone gain in absence of Wnt1 class–mediated Lrp6 signaling in the axial and appendicular skeleton
To address the role of Lrp6 in Sost deficiency–induced bone anabolism, we subjected wild-type and Sost−/− mice during late-stage skeletal growth to treatment with either of three distinct anti-Lrp6 antibodies with different modes of action. Anti-Lrp6 antibody A was previously shown to bind to the first YWTD-type β-propeller in the extracellular domain of Lrp6, thereby selectively blocking Lrp6 signaling induced by binding of Wnt1 class ligands including Wnt1/2/6/7a/7b/9a/10a/10b, while potentiating Wnt3-mediated and Wnt3a-mediated Lrp6 activity. Conversely, anti-Lrp6 antibody B binds to the third β-propeller and exerts the reverse pattern of function. Finally, the engineered biparatopic antibody A/B potently inhibits both Wnt1 and Wnt3/Wnt3a class–mediated Lrp6 signaling (Fig. 4F). Analysis of the proximal tibia by pQCT revealed that blocking Wnt1 class–mediated Lrp6 activity with antibody A strongly reduced the abnormal increase in cross-sectional total BMD in Sost−/− mice by 63% over the 4-week treatment period (Fig. 4A). Consistently, Sost deficiency–induced gain in cross-sectional cancellous BMD was diminished by 73% (Fig. 4B) and expansion of cortical thickness was reduced by 67% (Fig. 4C). Moreover, blocking Wnt1 class–mediated Lrp6 activity in Sost−/− mice not only robustly decreased Sost deficiency–induced enhanced bone gain, but resulted in bone gain that was comparable to the normal age-related late-stage skeletal growth in vehicle-treated wild-type mice; with exception of cancellous BMD gain, which had already stopped at this age, but was present at a reduced level in antibody A–treated Sost−/− mice (Fig. 4B). In contrast, normal skeletal growth–related bone gain in wild-type mice was mildly, but nonsignificantly reduced by inhibiting Wnt1 class–mediated Lrp6 signaling (Fig. 4A–C). Importantly, the reduction in tibial bone growth in wild-type and Sost−/− mice occurred irrespective of enhancement or blockage of Wnt3a class–mediated Lrp6 signaling (Fig. 4A–C). Consistently, promoting Wnt1 class–mediated Lrp6 activity in absence of Wnt3a class–mediated Lrp6 signaling induced by antibody B treatment resulted in a dramatic further enhancement of the abnormal bone increases in Sost−/− mice by about 3.2-fold in both bone compartments (Fig. 4A–C, E). Likewise, normal bone growth in wild-type mice was also significantly augmented in both bone compartments by antibody B treatment (Fig. 4A–C). However, the relative increases compared to vehicle-treated mice were about 30% smaller than in Sost−/− mice.
Next, we performed detailed bone structure analyses in the axial and appendicular skeleton using µCT. Consistent with the changes observed in the tibia, blocking Wnt1 class–mediated Lrp6 activity in Sost−/− and wild-type mice following either treatment with antibody A or A/B resulted in reductions of BMD and relative bone volume in the spine and smaller nonsignificant decreases in the femur (Table 3; Fig. 4D). These decreases in cancellous bone density and mass were associated with selective reductions in trabecular number (Table 3). Conversely, stimulating Wnt1 class–mediated Lrp6 activity in presence of blocked Wnt3a–mediated Lrp6 signaling by antibody B treatment further enhanced the abnormal vertebral and femoral bone gain in Sost−/− mice by about 1.3-fold to 2-fold for most parameters analyzed (Table 3; Fig. 4D). The same treatment in wild-type mice caused only nonsignificant changes.
|Bone structure parameter||Wild-type||Sost−/−|
|Vehicle||Antibody A||Biparatopic antibody A/B||Antibody B||Vehicle||Antibody A||Biparatopic antibody A/B||Antibody B|
|Cn.BMD (mg/cm3)||274.9 ± 33.6||172.6 ± 12.2||142.5 ± 4.4*||295.5 ± 8.5x||579.0 ± 40.4††||435.9 ± 38.5*,††||435.9 ± 107.6††||1076.6 ± 19.0**,++,xx,††|
|BV/TV (%)||23.5 ± 3.3||14.2 ± 1.1||12.1 ± 0.4||25.1 ± 0.8+||53.9 ± 4.1††||38.0 ± 3.6**,††||38.3 ± 10.7*,††||96.5 ± 1.0**,++,xx,††|
|Tb.N (1/mm)||6.8 ± 0.4||5.0 ± 0.1*||4.5 ± 0.1**||6.7 ± 0.1x||11.2 ± 0.6††||7.5 ± 0.3**,††||7.8 ± 1.4**,††||9.9 ± 0.6+,††|
|Tb.Th (μm)||39.3 ± 3.1||48.1 ± 2.7||44.7 ± 1.6||41.0 ± 0.5||65.7 ± 3.4††||67.8 ± 3.1||67.1 ± 6.9||97.1 ± 15.1**,++,xx,††|
|Cn.BMD (mg/cm3)||113.5 ± 8.5||79.9 ± 6.2||66.5 ± 8.5||163.2 ± 6.1+,xx||358.0 ± 14.0††||294.2 ± 27.1††||277.5 ± 99.5††||729.6 ± 20.0**,++,xx,††|
|Tb.N (1/mm)||4.4 ± 0.1||3.4 ± 0.1**||3.4 ± 0.1**||4.8 ± 0.1++,xx||5.5 ± 0.2††||4.6 ± 0.2**,††||4.7 ± 0.7††||8.7 ± 0.3**,++,xx,††|
|Tb.Th (μm)||39.0 ± 0.8||55.6 ± 1.4**||57.8 ± 1.9**||41.0 ± 0.7++,xx||83.2 ± 2.8††||86.6 ± 3.0††||78.7 ± 13.8††||134.6 ± 2.9**,++,xx,††|
|Total BMD (mg/cm3)||360.9 ± 29.2||344.8 ± 2.9||291.6 ± 11.4||388.5 ± 8.4x||549.0 ± 29.1††||517.0 ± 24.1††||481.2 ± 69.0††||877.4 ± 11.6**,++,xx,††|
|Cortical BMD (mg/cm3)||943.3 ± 9.8||941.3 ± 3.9||931.7 ± 9.0||933.2 ± 6.2||986.7 ± 12.9†||1006.9 ± 10.6††||1008.0 ± 12.0††||1018.2 ± 22.2††|
|Ct.Th (mm)||0.12 ± 0.00||0.13 ± 0.00||0.11 ± 0.00||0.11 ± 0.00||0.15 ± 0.01††||0.16 ± 0.01†||0.14 ± 0.01†||0.20 ± 0.01**,++,xx,††|
We then determined serum osteocalcin levels. Treatment with antibodies blocking Wnt1 class–mediated Lrp6 activity suppressed osteocalcin concentration weakly, but nonsignificantly by 20% to 30% relative to vehicle-treated mice of either genotype (Supplemental Fig. S3). Conversely, antibody B treatment nonsignificantly increased osteocalcin by 25% in Sost−/− mice, whereas it was nonsignificantly reduced by 31% in wild-type mice (Supplemental Fig. S3).
Loss of Sost deficiency–induced bone gain in absence of Lrp5 function and Wnt1 class–mediated Lrp6 signaling
To determine whether the residual excessive bone gain in Sost−/−;Lrp5−/− mice resulted from signaling through Lrp6, we subjected Sost−/−;Lrp5−/− and wild-type female mice to treatment with anti-Lrp6 antibody A. As observed before, normal late-stage skeletal growth–related increases in tibial total BMD and cortical thickness were not altered in antibody A–treated wild-type mice, whereas cancellous BMD was significantly reduced (Fig. 5A–C, F). In contrast, the remaining Sost deficiency–induced abnormal gain in tibial total BMD and cortical thickness was abolished in Sost−/−;Lrp5−/− mice upon antibody A treatment, resulting in comparable levels as in wild-type (Fig. 5A, C, F). Similarly, Sost deficiency–associated increase in cancellous BMD was fully suppressed in Sost−/−;Lrp5−/− mice following antibody A treatment (Fig. 5B). We then analyzed serum osteocalcin concentration as a measure of osteoblastic bone formation. The residual Sost deficiency–induced increase in osteocalcin was significantly suppressed by almost 50% in Sost−/−;Lrp5−/− mice to the level present in antibody A–treated wild-type mice (Fig. 5D).
Next, we performed µCT analyses of the axial and appendicular skeleton, comparing the effect of 4-week antibody A treatment in wild-type, Sost−/−, and Sost−/−;Lrp5−/− mice. In line with the tibial changes, the remaining Sost deficiency–induced vertebral and femoral cancellous bone gain was lost in Sost−/−;Lrp5−/− mice, resulting in comparable bone mass and density as in wild-type mice, which also showed significant decreases in these parameters compared to vehicle-treated mice (Table 4). Blocking Wnt1 class–mediated Lrp6 activity in wild-type mice did not alter total femoral BMD, but caused significant reductions of about 15% in Sost−/− and Sost−/−;Lrp5−/− mice (Table 4).
|Bone structure parameter||Wild-type||Sost−/−||Sost−/−;Lrp5−/−|
|Vehicle||Antibody A||Vehicle||Antibody A||Vehicle||Antibody A|
|Cn.BMD (mg/ccm)||276.8 ± 6.1||178.8 ± 13.6**||600.2 ± 12.1††||405.4 ± 34.8**,††||442.2 ± 8.1++,††||192.4 ± 14.0**,++|
|BV/TV (%)||23.4 ± 0.6||15.2 ± 1.2**||55.7 ± 1.3††||35.5 ± 3.4**,††||38.5 ± 0.8++,††||15.7 ± 1.2**,++|
|Tb.N (1/mm)||6.8 ± 0.1||5.3 ± 0.2**||11.1 ± 0.3††||8.0 ± 0.4**,††||7.3 ± 0.1++||5.6 ± 0.2**,++|
|Tb.Th (μm)||38.4 ± 0.3||44.0 ± 0.6*||69.9 ± 0.9††||59.1 ± 2.6**,††||66.5 ± 0.9††||46.5 ± 1.4**,++|
|Cn.BMD (mg/ccm)||100.9 ± 4.3||57.3 ± 6.9**||397.5 ± 8.1††||305.4 ± 13.1**,††||133.4 ± 9.6++,†||102.1 ± 8.5*,++,††|
|BV/TV (%)||7.4 ± 0.3||4.6 ± 0.5||32.1 ± 0.8††||24.4 ± 1.2**,††||9.3 ± 0.9++||7.4 ± 0.7++,†|
|Tb.N (1/mm)||4.8 ± 0.1||4.4 ± 0.1*||5.9 ± 0.1††||5.5 ± 0.1*,††||4.4 ± 0.1++,†||4.5 ± 0.1++|
|Tb.Th (μm)||35.1 ± 0.4||55.0 ± 2.5**||88.6 ± 1.8††||75.7 ± 2.6**,††||66.3 ± 2.1++,††||62.2 ± 1.4++,†|
|Total BMD (mg/ccm)||303.4 ± 4.0||302.2 ± 9.7||620.5 ± 7.7††||521.1 ± 12.9**,††||444.5 ± 6.9++,††||396.9 ± 10.6**,++,††|
|Cortical BMD (mg/ccm)||952.2 ± 2.4||965.7 ± 7.8||1022.1 ± 6.5††||1007.8 ± 9.2††||1032.4 ± 7.1††||1021.0 ± 7.1††|
|Ct.Th (mm)||0.11 ± 0.00||0.12 ± 0.00**||0.16 ± 0.00††||0.15 ± 0.00*,††||0.16 ± 0.00††||0.14 ± 0.00**,††|
Finally, to further address the relative differences in sensitivity between genotypes toward altered Wnt1 class–mediated versus Wnt3a class–mediated Lrp6 signaling, we measured expression of the Wnt signaling inhibitor Dkk1 in femoral cortical bone in response to vehicle or antibody A treatment. In contrast to sclerostin, which selectively binds to the first YWTD-type β-propeller of LRP6, Dkk1 also interacts with the third β-propeller, antagonizing both Wnt1 class–induced and Wnt3a class–induced Lrp6 activity.[38, 39] Interestingly, in vehicle-treated animals Dkk1 expression was strongly upregulated in Sost−/− and Sost−/−;Lrp5−/− mice compared to wild-type (Fig. 5E). Treatment with anti-Lrp6 antibody A did not significantly alter Dkk1 expression in either genotype, although it was slightly reduced in Sost−/− and Sost−/−;Lrp5−/− mice, whereas it remained unchanged in wild-type animals (Fig. 5E).
Here, we analyzed the role of the Wnt co-receptors Lrp5/6 in Sost deficiency–induced bone anabolism in vivo and report some surprising insights into the mechanism of Wnt signaling in the skeleton. Whereas the relative Sost deficiency–induced enhanced bone gain was blunted in Sost−/−;Lrp5−/− mice, it was fully normalized by pharmacological blockage of Wnt1 ligand class–mediated Lrp6 activity, indicating that Sost deficiency–induced bone anabolism in vivo requires Lrp5 and Lrp6 function. Because Sost deficiency–induced bone gain was equally reverted to wild-type levels in both Sost−/− and Sost−/−;Lrp5−/− mice by blockage of Wnt1 class–induced Lrp6 signaling, we conclude that Lrp6-induced Wnt/β-catenin signaling constitutes the major pathway through which Sost/sclerostin exerts its function in vivo.
In line with this interpretation, previous data indicate that sclerostin also targets factors other than Lrp5 or that compensatory mechanisms exist to functionally substitute for loss of Lrp5. Interestingly, mice lacking one or both alleles of Lrp5 and just one copy of Lrp6 constitutively display gene dosage–dependent limb deformities and decreased BMD, suggesting that Lrp5/6 serve partially redundant functions in embryonic as well as adult bone. In addition, single and combined osteoblast-specific deletion of Lrp5 and Lrp6 in mice revealed overlapping functions of both receptors in osteoblasts and postnatal bone acquisition. Similarly, Sost and Lrp6 have been shown to genetically interact during skeletal development and recent in vitro data have identified LRP6 as the main sclerostin receptor in mesenchymal cells, supporting our findings that Sost deficiency–induced bone anabolism largely depends on Wnt1 ligand–induced Lrp6 activity.
Sost deficiency–induced bone gain was reversed to levels of normal late-stage skeletal growth by blockage of Wnt1 class–induced Lrp6 activity in the cortical but not cancellous bone compartment, because cancellous bone gain, although reduced, remained slightly elevated compared to wild-type mice. Whether additional pathways such as the bone morphogenetic protein (BMP) signaling pathway[16, 28, 42, 43] or other targets interacting with sclerostin[44, 45] contribute to Sost deficiency–induced bone anabolism in the cancellous bone compartment remains to be addressed.
Given that the Lrp5 loss-of-function bone phenotype was suggested to be caused by a bone-independent role of Lrp5 controlling serotonin biosynthesis in the duodenum, we also determined serum serotonin levels. Opposed to previously reported elevated serum serotonin in the present Lrp5−/− mouse model, we failed to detect differences between genotypes (Supplemental Fig. S4), consistent with recent data demonstrating a local role of Lrp5 in bone.
LRP5/6 are highly homologous Wnt co-receptors of the LRP receptor family that are coexpressed in many developing and adult tissues with overlapping function. Structural data revealed that sclerostin binds to LRP6 through a conserved interaction motif present in the first YWTD-type β-propeller of the extracellular domain of LRP5/6, thereby causing inhibition of Wnt1 but not Wnt3a class–induced signaling.[29-31] Consequently, inhibiting Lrp6 activity induced by Wnt1 class ligands including Wnt1/2/6/7a/7b/9a/10a/10b was sufficient to prevent Sost deficiency-induced elevated bone gain in adult Sost−/− and Sost−/−;Lrp5−/− mice. This effect was irrespective of simultaneously potentiated or blocked Wnt3a class-induced Lrp6 activity. Conversely, potentiating Wnt1 class–mediated Lrp6 signaling while blocking Wnt3a class–mediated Lrp6 signaling further dramatically increased Sost deficiency–induced bone gain in adult Sost−/− mice, suggesting that Wnt1 class–induced Lrp6 activity might be rate-limiting for Sost deficiency–induced bone anabolism, whereas Wnt3a class–mediated Lrp6 signaling is dispensable.
In wild-type mice, Wnt1 class–mediated Lrp6 signaling also appears to be of greater relevance for adult late-stage skeletal growth–related bone gain than Wnt3a class–mediated Lrp6 activity, because activation or blockage of Wnt3a ligand class–dependent Lrp6 signaling in the context of blocked Wnt1 class–mediated Lrp6 activity resulted in comparable bone phenotypes. In contrast, although inhibiting Wnt1 class–induced Lrp6 activity mostly induced only small nonsignificant reductions in bone gain, its activation was sufficient to enhance adult tibial bone acquisition. In line with Wnt1 class ligands playing a crucial role in adult bone homeostasis, Wnt1, Wnt7b, and Wnt10b have been implicated in bone mass regulation in mice and humans.[26, 49-56] Nevertheless, WNT3a polymorphisms are associated with human BMD variation and rare inherited variants of WNT3a with reduced signaling activity have been identified in patients with childhood-onset primary osteoporosis.
Given the importance of Wnt/β-catenin signaling for adult bone homeostasis, various therapeutic approaches are being developed to target this bone anabolic pathway, aiming at increasing bone mass in conditions of bone fragility.[33, 59] Novel therapeutic agents such as neutralizing antisclerostin antibodies are in clinical development with promising results in a first-in-human study. Our data suggest that patients suffering from LRP5 loss-of-function causing the debilitating disease OPPG with severely decreased skeletal mass and strength should benefit from therapeutic principles blocking SOST/sclerostin activity given that Lrp5 deficiency–induced osteopenia was fully reversed upon Sost loss-of-function. There is a clear medical need in this rare genetic disorder of decreased BMD and elevated fracture risk because only some, but not all, OPPG patients respond to currently available treatment options.[61, 62]
In addition, our data suggest that patients suffering from severe bone overgrowth associated with decreased SOST function, such as in sclerosteosis, VBD, and CDD, might profit from therapeutic WNT signaling antagonists, given that the abnormal bone gain in Sost−/− mice was normalized upon blockage of Wnt1 class–mediated Lrp6 activity. Currently, the only available treatment for SOST-related sclerosing bone dysplasias consists in repeated surgical intervention for bone removal, illustrating the medical need in these rare diseases. Similarly, therapeutic WNT signaling inhibitors might be beneficial for treatment of other rare skeletal hyperostosis disorders, including bone dysplasias related to LRP4 loss-of-function mutations, which disrupt the LRP4-sclerostin interaction and associated LRP4 facilitator effect on sclerostin function. Several approaches to inhibit WNT/β-catenin signaling are presently being investigated, some of which, after promising preclinical results, have progressed to early clinical development for treatment of cancer-related indications. These range from function-blocking anti-LRP6 antibodies[30, 63] to Fzd receptor antagonists such as the anti-Fzd antibody vantictumab; to small molecule inhibitors of porcupine, a membrane-bound O-acyltransferase required for WNT ligand secretion and activity; and tankyrase inhibitors causing axin stabilizing and β-catenin degradation. However, given the pleiotropic nature of WNT/β-catenin signaling, treatment-related adverse effects need to be taken into account, especially in non-oncogenic clinical settings. Nevertheless, careful assessment of safety and therapeutic potential of WNT signaling antagonists in sclerosing bone dysplasias might reveal therapeutic windows and uncover novel therapeutic strategies to overcome the medical need in these devastating skeletal diseases.
In conclusion, our data highlight the potential for novel drug targeting approaches in rare skeletal disorders ranging from severe skeletal fragility to sclerosing bone overgrowth through selective modulation of canonical WNT signaling. These approaches should also be beneficial in the context of other medical conditions with much less severe phenotypes, such as postmenopausal osteoporosis, given the crucial role of this potent bone anabolic pathway for skeletal homeostasis.
All authors except JHG and CC are employees of the Novartis Institutes for BioMedical Research.
This study was supported by the Novartis Institutes for BioMedical Research Education Office Postdoctoral Fellowship Program (MKC, IK). We thank H. Anklin, M. Flückiger, P. Ingold, H. Jeker, M. Merdes, M. Pegurri, R. Schumpf, A. Studer, and A. Venturiere for excellent technical assistance.
Authors' roles: Study design: MKC, IK, DJ, SAE, FC, and MK. Data collection: MKC, IK, HK, JHG, CC, and CH. Data analysis: MKC, IK, HK, JHG, and CC. Data interpretation: MKC, IK, and MK. Drafting manuscript: MKC. Revising manuscript content: IK, and MK. Approving final version of manuscript: MKC, IK, HK, JHG, CC, DJ, SAE, FC, CH, and MK. MKC, IK, and MK take responsibility for the integrity of the data analyses.