Sclerostin Mediates Bone Response to Mechanical Unloading Through Antagonizing Wnt/β-Catenin Signaling

Authors

  • Chuwen Lin,

    1. Bio-X Center, Key Laboratory of Developmental Genetics and Neuropsychiatric Diseases, Ministry of Education, Shanghai Jiao Tong University, Shanghai, China
    2. Model Animal Research Center, MOE Key Laboratory of Model Animal for Disease Studies, Nanjing University, Nanjing, China
    3. Institute for Nutritional Sciences, SIBS, Chinese Academy of Sciences, Shanghai, China
    4. These authors contributed equally to this work
    Search for more papers by this author
  • Xuan Jiang,

    1. Model Animal Research Center, MOE Key Laboratory of Model Animal for Disease Studies, Nanjing University, Nanjing, China
    2. These authors contributed equally to this work
    Search for more papers by this author
  • Zhongquan Dai,

    1. These authors contributed equally to this work
    2. Laboratory of Space Cell and Molecular Biology, China Astronaut Research and Training Center, Beijing, China
    Search for more papers by this author
  • Xizhi Guo,

    1. Bio-X Center, Key Laboratory of Developmental Genetics and Neuropsychiatric Diseases, Ministry of Education, Shanghai Jiao Tong University, Shanghai, China
    Search for more papers by this author
  • Tujun Weng,

    1. Genetic Laboratory of Development and Disease, Institute of Biotechnology, Beijing, China
    Search for more papers by this author
  • Jun Wang,

    1. Orthopaedic Institute of PLA, Xijing Hospital, the Fourth Military Medical University, Xi'an, China
    Search for more papers by this author
  • Yinghui Li,

    1. Laboratory of Space Cell and Molecular Biology, China Astronaut Research and Training Center, Beijing, China
    Search for more papers by this author
  • Guoyin Feng,

    1. Bio-X Center, Key Laboratory of Developmental Genetics and Neuropsychiatric Diseases, Ministry of Education, Shanghai Jiao Tong University, Shanghai, China
    Search for more papers by this author
  • Xiang Gao PhD,

    Corresponding author
    1. Bio-X Center, Key Laboratory of Developmental Genetics and Neuropsychiatric Diseases, Ministry of Education, Shanghai Jiao Tong University, Shanghai, China
    2. Model Animal Research Center, MOE Key Laboratory of Model Animal for Disease Studies, Nanjing University, Nanjing, China
    • Bio-X Center, Shanghai Jiao Tong University, Hao Ran Building, 1954 Hua Shan Road, Shanghai 200030, China
    Search for more papers by this author
  • Lin He PhD

    Corresponding author
    1. Bio-X Center, Key Laboratory of Developmental Genetics and Neuropsychiatric Diseases, Ministry of Education, Shanghai Jiao Tong University, Shanghai, China
    2. Institute for Nutritional Sciences, SIBS, Chinese Academy of Sciences, Shanghai, China
    3. Institutes of Biomedical Sciences Fudan University, Shanghai, China
    • Bio-X Center, Shanghai Jiao Tong University, Hao Ran Building, 1954 Hua Shan Road, Shanghai 200030, China
    Search for more papers by this author

  • The authors state that they have no conflicts of interest

  • Published online on April 27, 2009

Abstract

Reduced mechanical stress leads to bone loss, as evidenced by disuse osteoporosis in bedridden patients and astronauts. Osteocytes have been identified as major cells responsible for mechanotransduction; however, the mechanism underlying the response of bone to mechanical unloading remains poorly understood. In this study, we found that mechanical unloading of wildtype mice caused decrease of Wnt/β-catenin signaling activity accompanied by upregulation of Sost. To further analyze the causal relationship among these events, Sost gene targeting mice were generated. We showed that sclerostin selectively inhibited Wnt/β-catenin in vivo, and sclerostin suppressed the activity of osteoblast and viability of osteoblasts and osteocytes. Interestingly, Sost−/− mice were resistant to mechanical unloading-induced bone loss. Reduction in bone formation in response to unloading was also abrogated in the mutant mice. Moreover, in contrast to wildtype mice, Wnt/β-catenin signaling was not altered by unloading in Sost−/− mice. Those data implied that sclerostin played an essential role in mediating bone response to mechanical unloading, likely through Wnt/β-catenin signaling. Our findings also indicated sclerostin is a promising target for preventing disuse osteoporosis.

INTRODUCTION

Mechanical loading is one of the crucial factors controlling bone mass.(1) Unloading (reduced mechanical stress on bone) leads to significant bone loss, as evidenced by disuse osteoporosis, which is a critical issue for bedridden patients and astronauts.(2,3) Although rapid progress has recently been achieved, the cellular and molecular mechanisms underlying the response of bone to unloading are just starting to emerge.

Osteocytes, the most abundant cells in bone, are terminally differentiated osteoblasts that are embedded in mineralized matrix. Osteocytes are distributed throughout the bone tissue volume. They communicate with each other and with osteoblasts on the bone surface through their dendritic processes within the tunnels called canaliculi, thereby forming a neuron-like network throughout the skeleton.(2) Osteocytes have recently been shown to be the major cell population responsible for mechanotransduction (to sense the mechanical load and translate it into biochemical signaling), as targeted ablation of osteocyte diminishes bone response to unloading.(4) A number of important molecules were shown in vitro and in vivo to be modulated in osteocytes in response to increased mechanical stress, including ATP, NO, prostaglandin E2 (PGE-2), IGF-1, glutamate transporter, c-fos, ion channel, and intracellular calcium, but few of these responses have been shown to be a prerequisite for loading adaptation.(5,6) It is still poorly understood how, on sensing the mechanical stimulus, osteocytes translate the loading information into the biochemical signals to direct bone formation in osteoblasts.

Wnt/β-catenin signaling plays a key role in controlling bone mass through regulating multiple aspects including osteoblast differentiation and function.(7–9) Recently, its activity was shown to be enhanced on mechanical loading.(10) However, it remains unknown whether Wnt/β-catenin signaling is involved in the response of bone to unloading.

Sclerostin, encoded by the Sost gene, is a secreted cysteine-knot protein among the DAN family, which includes proteins antagonize BMP and Wnt signaling.(11) Sclerostin has emerged as a potent inhibitor of bone growth.(12–15) Sclerostin was originally identified as a BMP antagonist because of its cysteine-knot domain, which was shared by BMP antagonists, and its binding to BMP and potent inhibition on BMP-induced osteogenesis,(16,17) although it did not function as classical BMP antagonists did.(18) Later, sclerostin was shown to bind to low-density lipoprotein (LDL) receptor–related proteins 5 and 6 (Lrp5 and Lrp6) and inhibit Wnt/β-catenin signaling in vitro.(19,20) Thus, the mechanism of sclerostin's function remains to be elucidated in vivo. Importantly, Sost is nearly exclusively expressed in osteocytes in adult bone.(18,21) Moreover, Sost expression is responsive to mechanical stimulus.(22) These together suggest the involvement of sclerostin in bone mechanotransduction. However, the role of sclerostin in bone response to mechanical stress remains unknown.

In this study, we addressed the role of sclerostin and Wnt/β-catenin in mediating the response of bone to mechanical unloading. Wnt/β-catenin signaling was decreased in association with upregulation of Sost in wildtype (WT) mice in response to unloading. We showed by deleting Sost in vivo that sclerostin antagonized Wnt/β-catenin signaling. Under physiological loading, sclerostin secreted by osteocytes could suppress the activity and viability of osteoblasts in addition to viability of osteocytes. Interestingly, under unloading conditions, Sost−/− mice were resistant to unloading-induced bone loss. Reduction in bone formation in response to unloading was abrogated in Sost−/− mice. Furthermore, Wnt/β-catenin signaling in long bones from mutant mice was not altered by unloading. Taken together, our results strongly imply that sclerostin mediated the bone response to mechanical unloading through antagonizing Wnt/β-catenin signaling.

MATERIALS AND METHODS

Generation of Sost−/− mice

We ordered a 129/SvJ BAC containing the Sost gene from Sanger Institute (www.ensembl.org). We replaced the coding sequence (CDS) of Sost with a LacZ gene and a PGK-Neomysin cassette. The linearized targeting vector was electroporated into ES cells. After double selection, Sost-targeted ES clones were identified by Southern blot with 5′ and 3′ external probes (Fig. 1E). ES clones carrying the Sost-null allele were injected into C57BL/6J blastocysts to obtain male chimeric mice, which were crossed with C57BL/6J females.

Figure Figure 1.

(A–C) Wnt/β-catenin signaling was decreased in association with upregulation of Sost by unloading in WT mice. Real-time PCR was performed using cDNA from femora of 17-wk-old loaded and unloaded groups of WT mice. (A) Sost. (B) Lef-1. (C) Cyclin D1. *p < 0.05, **p < 0.01 (n = 4 per group). Gene expression was normalized by GAPDH. (D–G) Targeted disruption of the Sost gene. (D) Schematic illustrations of the WT allele, targeting vector, and mutant allele. Sost contains two exons (black boxes), and a genomic fragment comprising entire coding sequence was replaced with a LacZ-PGK-Neomycin (LacZ-Neo) cassette. A PCR fragment downstream of the 3′ BamHI site was used as 3′ probe. (E) Southern blot of G418 and gancyclovir-resistant ES clones. Genomic DNA from these ES clones was digested with BamHI and hybridized with a 3′ probe as indicated in D. Because of the replacement of CDS with a LacZ-Neo cassette that harbored a BamHI site, the labeled band shifted from 5.0 (WT allele) to 3.8 kb (mutant allele). (F) Absence of the Sost transcript in Sost−/− mice was shown by RT-PCR of femora cDNA using primers specific to Sost. β-actin was used as an internal control. (G) Disruption of sclerostin protein in Sost−/− was verified by Western blot using total protein lysates from Sost−/− and WT femora. β-actin was used as an internal control.

Tail suspension and sample preparation

For tail suspension, a tape was applied to the surface of the tail to set a metal clip. The end of the clip was fixed to an overhead bar, and the height of the bar was adjusted to maintain the mice at ∼30° head-down tilt with the hindlimbs elevated above the floor of the cage. The mice in the unloaded group were subjected to tail suspension for 14 days, and caged individually (for 17-wk-old female mice, n = 4∼6 per group, and for 7-wk-old female mice, n = 3 per group). Loaded control mice were also caged individually under the same conditions except for tail suspension. The mice were injected intraperitoneally with calcein (Sigma-Aldrich, St. Louis, MO, USA) at 20 mg/kg 7 and 2 days before death on day 14. After 2 wk of tail suspension, mice were killed by cervical dislocation. The femora and tibias were separated from adherent muscles and connective tissues other than the periosteum. One femur was used for extraction of total RNA, and the other femur was fixed in 4% paraformaldehyde (PFA) for 24 h, dehydrated, and stored in 70% ethanol for subsequent BMD and μCT analysis. Tibias were both fixed with 4% PFA for 24 h. One was decalcified in 14% EDTA for subsequent TUNEL assay. The other tibia was used for subsequent undecalcified section (plastic section).

TUNEL assay and bone histomorphometric analysis

In brief, the tibias, after decalcification, were dehydrated and embedded in paraffin. Samples were sectioned at 5 μm. To detected apoptosis, we used the Dead End Colorimetric TUNEL system (Promega, Madison, WI, USA) according to the manufacturer's instructions. The number of total apoptotic osteoblasts and osteocytes in the cortical bone was counted manually using microscopy, and the percentage of apoptotic cells was calculated.

Histomorphometric analyses were performed on the tibial metaphysis of both the unloaded and loaded groups of Sost−/− mice and WT mice. For each mouse, the tibia was embedded in methylmethacrylate. Five-micrometer plastic sections were stained using the modified Goldner trichrome technique, and sections adjacent to the stained sections were left unstained for fluorescent analyses. Histomorphometric analysis was performed according to standard protocols(23) using the OsteoMeasure Analysis System (Osteometrics, Atlanta, GA, USA).

Radiography and BMD measurement

Radiographic analyses were carried out using Philips Bucky Diagnost CS DR System (Philips Medical System, Best, The Netherlands). BMD of the whole body and femoral diaphysis was monitored in mice at 4, 6, 8, 12, 24, and 40 wk of age, using peripheral DXA (pDXA; PIXImus II; GE-Lunar, Madison, WI, USA). The BMD of total body was measured; the head was excluded. The region of interest (ROI) for the femur included the central 50% of the whole femur.

μCT

Geometric properties of femoral trabecular bone volume fraction and mid-diaphysis were evaluated using a high-resolution desktop μCT imaging system (eXplorer Locus SP; GE Healthcare). Scanning for the femur was started at 15% of the total femur length measured from the tip of femoral condyle and extended proximally for 100 slices. The area for trabecular analysis was outlined within the trabecular compartment with its closest and furthest edges at 0.3 and 1.5 mm to the growth plate of the proximal ends of the tibia, excluding the cortical and subcortical bone. Every 10 sections were outlined, and the intermediate sections were interpolated with the contour algorithm to create a volume of interest. The region of interest in cortical bone was chosen by MicroView, and bone analysis was performed.

Extraction of RNA and protein from bone

Femora of WT and Sost−/− mice were dissected free of soft tissues, and bone marrow was flushed away. Femora were crushed using a mortar and pestle in liquid nitrogen. RNA was extracted from the lysate using the TRizol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions.

For total protein extraction, the lysates were collected in 600 μl of ice-cold radioimmunoprecipitation assay (RIPA) buffer. For nuclear protein extraction, we isolated nuclear protein from the lysates using a Nuclear Extract Kit (ActiveMotif, Carlsbad, CA, USA). Proteins were quantified by the BCA assay (Pierce, Rockford, IL, USA).

Real-time PCR and Western blot

One microgram total RNA was reverse transcribed to cDNA with the use of superscript III (Invitrogen). Real-time PCR was performed using the ABI 7700 and EvaSYBR Green (Biotium, Hayward, CA, USA). The standard curve method of quantification was used to calculate the expression of target genes, which was relative to the housekeeping gene GAPDH. Experiments were repeated at least three times. The primer sequences are available on request.

Western blot was performed using the antibodies as followed: mouse antibody (Ab) to β-catenin (Zymed, San Francisco, CA, USA), to connexin 43 (Santa Cruz Biotechnology, Santa Cruz, CA, USA), to Bax (Santa Cruz), to Bcl-2 (Santa Cruz), to PHEX (obtained from Dr. Thomas P. Loisel), rabbit Ab to phosphor-β-catenin (Ser-33/Ser-37/Thr-41; Bioworld, Dublin, OH, USA), to Bcl-xL (PTGLAB, Chicago, IL, USA), and to pSMAD1/5/8 (CellSignaling, Beverly, MA, USA), and goat Ab to sclerostin (R&D Systems, Minneapolis, MN, USA).

Statistical evaluations

Data were expressed as the mean ± SE. Statistical analysis was performed by Student's t-test or ANOVA. p < 0.05 was considered statistically significant.

RESULTS

Wnt/β-catenin signaling was decreased in association with upregulation of Sost in response to mechanical unloading

To assess a potential role of Wnt/β-catenin in the bone response to unloading, 17-wk-old WT mice were subjected to tail suspension for 2 wk. Mice under normal loading conditions served as controls. The femora were isolated for RNA extraction and subsequent real-time PCR. In consistence with other reports, Sost expression was increased by unloading (Fig. 1A). Intriguingly, expression of Lef-1 and Cyclin D1, two targets of Wnt/β-catenin signaling,(24) were both decreased by unloading (Figs. 1B and 1C). Thus, Wnt/β-catenin signaling was decreased in response to unloading, accompanied by upregulation of Sost. This finding strongly suggested a role of sclerostin in bone response to unloading, probably through regulating Wnt/β-catenin signaling.

Targeted disruption of Sost gene in mice

Upregulation of Sost accompanied by a decrease of Wnt/β-catenin signaling under mechanical unloading suggested a possible mechanism that osteocytes control bone growth in response to unloading through adjusting sclerostin level and thus modulating Wnt/β-catenin signaling activity. To verify this hypothesis would require clarifying the mechanism of sclerostin function and investigating the response of mice to unloading in the absence of sclerostin. For this purpose, we generated Sost-deficient mice by homologous recombination in embryonic stem (ES) cells. The CDS of Sost was replaced by a LacZ-PGK-Neo cassette (Fig. 1D). Targeted clones were screened by Southern blot (Fig. 1E). Heterozygous crosses produced litters with expected mendelian ratios of WT (Sost+/+), heterozygous (Sost+/−), and homozygous mutant (Sost−/−) pups. Sost−/− mice had a normal morphological appearance and were viable and fertile.

To confirm disruption of sclerostin in Sost−/− mice, RNA and protein was extracted from femora of 2-mo-old Sost−/− and WT littermates, and no Sost transcript or sclerostin protein was detected by RT-PCR or Western blot in Sost−/− femurs (Figs. 1F and 1G). Thus, Sost−/− mice were functionally null for sclerostin.

Sost−/− mice exhibited high bone mass

Bone densitometry showed that the total body and femoral BMD of Sost−/− mice was elevated compared with WT mice from as early as 4 wk and throughout life (Figs. 2A and 2B). Radiographic analysis of 2-mo-old mice confirmed increased bone mass in Sost−/− mice compared with WT littermates (data not shown). Accordingly, μCT showed that the bone mass of both trabecular and cortical bone was markedly increased in Sost−/− mice (Fig. 2C), as reported in the other Sost-null mice.(14)

Figure Figure 2.

Sost−/− mice exhibited high bone mass postnatally but normal bone development. (A) Total body BMD and (B) femoral BMD in WT and Sost−/− mice female mice, measured at 4, 6, 8, 12, 24, and 40 wk of age. (n = 3–8 per group). (C) 3D μCT image of trabecular (bottom) and cortical bone (top) of femora from 12-wk-old male Sost−/− and WT. (D) H&E staining on sections of developing tibias at E14.5. (E) Alizarin red and Alcian blue staining of WT (+/+) and Sost−/− (−/−) mice at P1. Head (left) and forelimb (right) are shown. **p < 0.01, ***p < 0.001.

Sost is expressed in perichondrium during embryogenesis.(25) However, analysis on section of tibia at E14.5 and skeleton staining of P1 pups showed the bone development in Sost−/− embryo was normal (Fig. 2D and 2E). Thus, sclerostin might be more important in the mature skeleton that was exposed to greater mechanical strain than the developing skeleton.

Wnt/β-catenin signaling was upregulated in Sost−/− mice

Normally, the mouse skeleton bears mechanical stress at the physiological level. Mature osteocytes that are buried deep in matrix secrete sclerostin to control bone mass.(21) Sost-null mice provided an excellent model to explore the mechanism of sclerostin function, whereby osteocytes control bone mass physiologically.

To explore the molecular mechanism of sclerostin function, we performed Western blot using total protein and nuclear fraction of total protein, which was extracted from femora of 7-wk-old Sost−/− mice and WT littermates. To evaluate Wnt/β-catenin signaling activity, we first examined the level of β-catenin in the nuclear fraction of the femur protein, because nuclear localization of β-catenin is a crux of activation of Wnt/β-catenin signaling.(7) Nuclear accumulation of β-catenin was increased in the Sost−/− femur (Fig. 3A). Conversely, phosphorylated β-catenin level (the form targeted for degradation) was decreased in the Sost−/− femur (Fig. 3B). Moreover, expression of connexin 43, a target of Wnt/β-catenin signaling,(26) was increased in the Sost−/− femur (Fig. 3C). Finally, real-time PCR showed that expression of Lef-1, also a target of Wnt/β-catenin signaling, was consistently upregulated in the 7-wk-old Sost−/− femur (Fig. 3E). Together, these results showed that Wnt/β-catenin signaling was upregulated in the absence of sclerostin.

Figure Figure 3.

Wnt/β-catenin signaling was enhanced in Sost−/− mice. (A–D) Western blot analysis of protein expression using total protein or nuclear fraction of total protein extracted from femora of 7-wk-old Sost−/− and WT littermates. (A) Increased nuclear localization of β-catenin in Sost−/−. Nuclear fraction of total protein was used. β-actin was used as an internal control. (B–D) Total protein, β-actin, or α-tubulin was used as an internal control. (B) Decreased phosphorylated β-catenin in Sost−/−. (C) Increased Connexin 43 in Sost−/−. (D) No change in phosphorylated SMAD1/5/8. (E) Real-time PCR, showing Lef-1 was upregulated in Sost−/−. (F) Real-time PCR, showing no change in mRNA level of Id-1. Gene expression was normalized by GAPDH.

Sclerostin was originally designated as an antagonist of BMP signaling for its binding with BMPs and its inhibition on BMP-induced osteogenesis in in vitro experiments.(16,17) Although more recent evidence showed it was more likely to be an antagonist of Wnt/β-catenin signaling, no in vivo data were reported to exclude the first hypothesis. To rule out the possibility that the high bone mass phenotype in Sost−/− mice was attributable to BMP signaling alteration, phosphorylation of SMAD1/5/8, which was a central event during activation of BMP signaling, was assessed by Western blot. There was no significant difference in the level of phosphorylated SMAD1/5/8 between WT and Sost−/− mice (Fig. 3D). Correspondingly, expression of BMP signaling target Id-1(27) was unchanged in Sost−/− bone (Fig. 3F).

Taken together, these results strongly implied that sclerostin was an endogenous antagonist to Wnt/β-catenin signaling rather than to BMP signaling.

Osteocyte and osteoblast apoptosis was attenuated and osteoblast activity was elevated in Sost−/− mice

Osteocytes were hypothesized to deliver sclerostin to the osteoblasts on the bone surface, which were effector cells actively synthesizing bone matrix. Thus, to explore the mechanism of sclerostin function at a cellular level under normal loading, we need to examine the conditions of both osteocytes and osteoblasts in Sost−/− mice. We first assessed the number of osteocytes by examining the level of PHEX and DMP1, markers of osteocyte,(28,29) using protein or RNA extracted from femora from 7-wk-old mice. The level of PHEX protein and DMP1 mRNA was both increased in Sost−/− femora, shown by Western blot and real-time PCR, respectively (Figs. 4A and 4B). This result may suggest an increased number of osteocytes or alternatively altered activity of osteocytes in Sost−/− bone. Subsequent direct quantification of osteocyte number showed that, whereas total osteocyte number was increased in Sost−/− cortical bone (data not shown), there was no significant difference in osteocyte density (osteocyte number/mm2 bone area) between Sost−/− and WT (Fig. 4C). Thus, the increase in osteocyte number should be caused by an increase in bone tissue.

Figure Figure 4.

Apoptosis of osteoblasts and osteocytes was attenuated and osteoblastic activity was enhanced in Sost−/− (−/−) compared with WT (+/+). (A, B, D, I, and J) Western blot (A and D) and real-time PCR (B, I, and J) using total protein or RNA extracted from 7-wk-old mouse femora. (A) Increased PHEX (phosphate-regulating gene with homologies to endopeptidases on the X chromosome) in Sost−/−; β-actin was used as an internal control. (B) Increased DMP1 (dentin matrix protein 1) expression in Sost−/−. (C) Osteocyte density was not altered in Sost−/− cortical bone. (D) Decreased ratio of Bax/Bcl-2 and increased Bcl-xL in Sost−/−. β-actin was used as an internal control. (E and F) Prevalence of apoptosis of osteocytes (E) and osteoblasts (F) was decreased in cortical bone of Sost−/− tibias. (G) TUNEL staining (brown) showed apoptotic osteocytes (arrowhead) and apoptotic osteoblasts (arrow) in cortical bone of 7-wk-old WT and Sost−/− mice. Hematoxylin stained for nucleus (blue). Osteocytes and osteoblasts were indicated by blank arrowhead and blank arrow, respectively. (H) Increased Ob. S/BS (osteoblast surface per bone surface) in Sost−/−. (I) Real-time PCR, showing increased collagen I (collagen type I) expression in Sost−/−. (J) Real-time PCR, showing increased osteocalcin expression in Sost−/−. Gene expression was normalized by GAPDH. *p < 0.05, **p < 0.01 (n = 3 per group). NS, no significance.

Change in apoptosis could alter the life span of cells and thus change the duration of their function. To test the level of apoptosis in bone cells, expression levels of Bax (pro-apoptotic factor), Bcl-2, and Bcl-xL (anti-apoptotic factors) were examined by Western blot. In Sost−/− mice, the ratio of Bax/Bcl-2 was decreased, and the level of Bcl-xL was increased (Fig. 4D), indicating overall apoptosis in bone cells was reduced. To determine the cell types of which apoptosis was altered, TUNEL was performed on decalcified sections of 7-wk-old tibias. It showed in Sost−/− mice that the osteocyte apoptosis was significantly decreased, and osteoblast apoptosis was also impaired (Fig. 4G, and the quantitative result was shown in Figs. 4E and 4F). These data were consistent with the previous observation that sclerostin promoted osteoblastic apoptosis in vitro.(30)

Bone histomorphometric analysis showed increased osteoblast surface (Ob.S/BS) in Sost−/− (Fig. 4H). Moreover, dynamic bone histomorphometric analysis showed that matrix apposition rate (MAR) and bone formation rate (BFR/BS), indicators of osteoblast activity, were elevated in 17-wk-old Sost−/− mice compared with WT (data not shown). In addition, real-time PCR showed increased expression of collagen type I (collagen I) and osteocalcin in 7-wk-old Sost−/− femora, indicating elevated activity of Sost−/− osteoblasts (Figs. 4I and 4J).

To understand whether the enhancement in osteoblast activity was cell autonomous, we performed an alkaline phosphatase (ALP) activity assay on the primary calvarial osteoblasts isolated from P3 Sost−/− and WT littermates. Primary Sost−/− osteoblasts showed no difference in function from WT (data not shown). Thus, the postnatal enhancement in Sost−/− osteoblast activity was caused by ablation of an inhibitory effect of sclerostin, which was normally secreted from osteocytes but not inherent to osteoblast itself.

Taken together, the function of osteoblasts was enhanced, and the apoptosis of osteoblasts and osteocytes was attenuated in Sost−/− mice. Therefore, physiologically sclerostin secreted by osteocytes suppressed the activity of osteoblasts as well as viability of osteoblasts and osteocytes.

Unloading-induced bone loss was prevented in Sost−/− mice

To directly study the role of sclerostin in response of bone to mechanical unloading, 17- and 7-wk-old female Sost−/− and WT mice were subjected to tail suspension for 2 wk. Mice on the ground under normal loading conditions (Sost−/− loaded and WT loaded) served as controls. Baseline body weight levels were similar between Sost−/− and WT mice (data not shown). The femora and tibias of the mice were isolated for subsequent bone densitometry, μCT analysis, bone histomorphometric analysis, and real-time PCR. As expected, unloading significantly reduced BMD of femora in WT. Remarkably, no significant reduction was observed in the Sost−/− mice subjected to unloading (Sost−/− unloaded) compared with Sost−/− loaded mice (Fig. 5A). The absence of unloading-induced decrease in BMD was similarly observed in 7-wk-old Sost−/− mice (Fig. 5B). μCT scanning on the primary trabecular fraction of femora of 17-wk-old mice showed that 3D bone volume (BV/TV), which was markedly decreased in WT bone, was not altered in Sost−/− mice (Figs. 5C and 5D). Correspondingly, the bone microstructure parameters including trabecular bone thickness (Tb.Th) and number (N.Tb/BS) and connection density (Conn.Dens) were not altered in Sost−/− in contrast to the significant change in WT mice (Figs. 5E–5G). Also, the similar abrogation in the reduction of cortical bone area and thickness was detected in Sost−/− mice (Figs. 5H and 5I). These observations clearly showed that Sost−/− mice were resistant to the bone loss induced by unloading.

Figure Figure 5.

Sost deficiency suppressed unloading-induced bone loss. (A) Femoral BMD from groups of 17- (A) and 7-wk-old (B) WT and Sost−/− mice after 2 wk of normally loading (LOAD) or unloading (UNLO). (C) Representative 3D μCT images of the distal metaphyseal regions of femora from loaded and unloaded group of WT and Sost−/−. (D–G) Quantification of the 3D image in C with microstructural parameters. (D) 3D BV/TV (bone volume per tissue volume). (E) Tb.N (trabecular number). (F) Tb.Th (trabecular thickness). (G) Conn.Dens (connection density). (H and I) Microstructural parameters of cortical bone in loaded and unloaded groups of WT and Sost−/− mice. (H) Ct.Ar (cortical area). (I) Ct.Th (cortical thickness). *p < 0.05,**p < 0.01,***p < 0.001 (n = 4–6 per group). NS, no significance.

To understand the cellular basis of the resistance to unloading-induced bone loss in Sost−/− mice, dynamic bone histomorphometric analysis was conducted based on calcein double labeling. The analysis provided in vivo evaluation of osteoblastic activity with respect to accumulation of bone mass. MAR and BFR/BS were markedly decreased in WT by unloading (Figs. 6A–6C), indicating a decrease in osteoblastic activity in WT mice during unloading. In contrast, BFR and MAR in Sost−/− mice were not altered by unloading. Thus, the reduction in bone formation in response to unloading was abolished in Sost−/− mice.

Figure Figure 6.

(A–C) Sost deficiency prevented unloading-induced reduction in osteoblastic activity. (A) Representative image of calcein bands from loaded and unloaded groups of WT and Sost−/− mice. Calcein bands were visualized to obtain histomorphometric parameters. Double arrowheaded lines showed the distance between two calcein bands. (B) MAR (matrix apposition rate) in each group. (C) BFR/BS (bone formation rate per bone surface) in each group. (D–F) Real-time PCR analysis of targets of Wnt/β-catenin signaling using RNA extracted from femora of 17- (D and E) and 7-wk-old (F) loaded and unloaded groups of WT and Sost−/− mice. Gene expression was normalized by GAPDH. (D) Lef1. (E) Cyclin D1. (F) Lef1. *p < 0.05, **p < 0.01. NS, no significance.

To test whether the different change in function between Sost−/− and WT osteoblasts responding to unloading was osteoblast autonomous or not, the Sost−/− and WT primary osteoblasts were cultured under simulated microgravity (in vitro unloading), which mimicked low fluid shear stress and 3D freedom. Sost−/− primary osteoblasts showed a similar response to absence of shear stress as that from WT under simulated microgravity (data not shown). Therefore, the distinguished response to unloading between Sost−/− and WT osteoblasts was not osteoblast autonomous.

Wnt/β-catenin signaling was not altered by unloading in Sost−/− mice

To obtain insights into the molecular mechanism of sclerostin action in mechanotransduction, real-time PCR was performed using RNA extracted from femora of 17- and 7-wk-old mice subjected to tail-suspension assay. The targets of Wnt/β-catenin signaling Lef-1 and Cyclin D1(24) were significantly downregulated in WT by unloading. In contrast, their expression was not altered in Sost−/− mice (Figs. 6D–6F), indicating the Wnt/β-catenin signaling was not decreased by unloading in Sost−/− mice. Thus, unleashing Wnt/β-catenin signaling was underlying the abolishment of unloading-induced bone loss in Sost−/− mice. It implied that Wnt/β-catenin was a key signaling pathway mediating the effect of unloading.

It should be noted that, in 17-wk-old mice, there was no difference in Wnt/β-catenin signaling activity between WT and Sost−/− in long bone, as shown by real-time PCR (Figs. 6D and 6E). It might be that Wnt/β-catenin signaling was not activated in 17-wk-old long bones, or alternatively, other antagonists of Wnt/β-catenin signaling functioned instead of sclerostin at that time.

DISCUSSION

In this study, we found Wnt/β-catenin signaling was decreased in association with increased Sost expression in response to unloading in WT mice. Moreover, we showed sclerostin inhibited Wnt/β-catenin signaling in vivo. Under normal loading, sclerostin secreted by osteocytes suppresses the function and viability of osteoblasts as well as the viability of osteocytes to inhibit bone formation. Intriguingly, in the absence of sclerostin, decrease in Wnt/β-catenin signaling in response to unloading was prevented, and, at the cellular level, unloading-induced reduction in osteoblast activity was diminished. As a result, the unloading-induced bone loss was inhibited. Therefore, sclerostin played an essential role in mediating bone response to mechanical unloading, likely through antagonizing Wnt/β-catenin signaling; Wnt/β-catenin signaling was a key signaling mediating the effect of unloading.

Sclerostin is a mechanical sensor secreted by osteocytes to act on osteoblasts on bone surface

Skeleton unloading, which is often caused by immobilization and spaceflight, significantly decreases bone formation and bone mass. Rapid progress has been made in understanding the mechanism of mechanotransduction. However, it remains elusive how osteocytes translate the loading information into biochemical signals to direct bone formation. Sclerostin is one potential factor for mediating the regulation of bone mass by mechanical stimulus given its expression pattern.(18,21,22)

We generated Sost−/− mice, which exhibited high bone mass. However, Sost−/− embryos showed no bone defect despite Sost expression in embryogenesis (Fig. S1), suggesting the main function of sclerostin may lie in the mature skeleton that was exposed to greater mechanical strain. Interestingly, in skeletal unloading, Sost−/− mice were resistant to unloading-induced bone loss. Inconsistent with this result, reduction in bone formation in response to unloading was inhibited in Sost−/−. Our results clearly showed that sclerostin mediated the control of bone mass by mechanical stress.

Therefore, sclerostin is likely the long sought after molecule by which osteocytes convey the mechanical stimulus condition to the effector cells on bone surface.(21,31)

Sclerostin and Wnt/β-catenin signaling

The mechanism of sclerostin inhibition on bone growth was unsettled before. Sclerostin was originally regarded as an antagonist of BMP signaling, but later, two groups showed that sclerostin bound to Lrp5 and Lrp6 and inhibited Wnt/β-catenin signaling in vitro.(19,20) However, the conclusion that sclerostin acted as an inhibitor of Wnt/β-catenin signaling came from only in vitro analysis, lacking in vivo evidence. In this study, we showed Wnt/β-catenin signaling activity was enhanced in Sost−/− bone, whereas BMP signaling was unchanged. Our results, together with previous data that sclerostin bound to Lrp5/Lrp6 and inhibited Wnt/β-catenin signaling in vitro,(19,20) strongly implied sclerostin was an antagonist of Wnt/β-catenin signaling rather than an inhibitor of BMP signaling. However, it should be noted that, taken individually, our evidence was indirect, and to finally determine the direct inhibition of sclerostin on Wnt/β-catenin signaling, mice deficient both in β-catenin in osteoblasts and in Sost may provide more direct evidence.

Recently, a notable study reported that duodenum-derived Lrp5 controlled bone formation by inhibiting serotonin synthesis.(32) Their data did not necessarily contradict the effect of sclerostin on Wnt/β-catenin. Previous studies have shown that Lrp6, the closest homolog of Lrp5, was required for Wnt/β-catenin signaling and was a better candidate as co-receptor of Wnt/β-catenin signaling than Lrp5.(33,34) As an indirect instance, the mechanism of Dkk1 antagonism on Wnt/β-catenin signaling was identified originally through its binding with Lrp6.(35,36) Sclerostin bound to both Lrp5 and Lrp6, and particularly, it was able to disrupt Wnt-induced Frizzled–Lrp6 complex formation.(20) In addition, sclerostin protein expression was relatively restricted in bone postnatally, unlike Lrp5.(16,18,21,37) Therefore, sclerostin could inhibit Wnt/β-catenin signaling through binding to Lrp6.

Wnt/β-catenin signaling in bone response to unloading

Wnt/β-catenin signaling critically regulates bone mass.(8,9,38) Recently, Wnt/β-catenin signaling was shown to be involved in the bone response to mechanical loading.(10,39) However, a role of Wnt/β-catenin in bone response to unloading remains unknown. In this study, we showed Wnt/β-catenin signaling was physiologically responsive to mechanical unloading, because it was suppressed by unloading in WT mice (Figs. 1B and 1C). Moreover, we showed uncompromised Wnt/β-catenin signaling was underlying the resistance of Sost−/− mice to unloading-induced bone loss (Figs. 6D–6F). Therefore, Wnt/β-catenin signaling seemed to be a key local signaling pathway mediating the effect of mechanical unloading.

Although sclerostin was regarded as an inhibitor of Wnt/β-catenin signaling, their relationship in mediating bone response to unloading was first examined in our study. Among the Wnt antagonists expressed in bone, sclerostin and Dkk1 were both responsive to mechanical stimuli.(22) Whereas the Sost gene was exclusively expressed in mature osteocytes,(21) Dkk1 was more widely distributed. Interestingly, in the absence of sclerostin, Wnt/β-catenin signaling activity was not altered by unloading; thus, sclerostin seemed to play a dominant role in antagonizing Wnt/β-catenin signaling in the aspect of mechanotransduction.

Taken together, we proposed that, in response to unloading, an increased amount of sclerostin secreted by osteocytes inhibited Wnt/β-catenin signaling, thus inducing apoptosis and suppressing osteoblast activity, which decreased the bone formation; as a result, bone mass decreased (Fig. 7).

Figure Figure 7.

Proposed model of the function of sclerostin and Wnt/β-catenin signaling under mechanical unloading. In response to unloading, osteocytes secrete more sclerostin to osteoblasts, which inhibits Wnt/β-catenin signaling; as a result, the osteoblast activity is suppressed, and apoptosis of osteoblasts and osteocytes is induced. Thus, bone formation is greatly inhibited, leading to bone loss. OB, osteoblast; OC, osteocyte.

Sclerostin and apoptosis of osteoblast lineage cells

The promotion of sclerostin on osteoblast and osteocyte apoptosis was in line with a previous report that sclerostin induced the apoptosis of human osteoblastic cells in vitro.(30) This function of sclerostin might be dependent on its antagonism to Wnt/β-catenin signaling. Increased Wnt/β-catenin signaling is well known to be associated with decreased apoptosis.(40) It was shown that Wnt signaling inhibited osteoblastic cell death in vitro.(41) Accordingly, deletion of a Wnt antagonist sFRP1 prevented the apoptosis of osteoblast lineage cells in vivo and in vitro.(42,43)

In conclusion, our results showed that sclerostin played an essential role in mediating the bone response to mechanical unloading, likely through antagonizing Wnt/β-catenin signaling; Wnt/β-catenin signaling seemed to be an important signaling pathway, mediating the effect of unloading. Thus, disrupting endogenous sclerostin could be beneficial in preventing unloading-induced bone loss in spaceflight or bed rest.

Acknowledgements

This work was partially funded by a number of grants (2006AA02A407, 2006CB910600, 2006BAI05A05, 2007CB947300), the Knowledge Innovation Program of the Chinese Academy of Sciences (KSCX2-YW-R-01) and the Shanghai Leading Academic Discipline Project (B205), and grants for X.G. from NSFC (32825024) and MOST (2005CB522501, 2006CB943500).

Ancillary