C57BL/6 mice were purchased from CLEA Japan, Inc. and bred for experiments. The generation of ScxGFP transgenic and Scx ^Δ11/Δ11 strains has been previously reported (Sugimoto et al., 2013b; Shukunami et al., 2018). For detection of the activation of WNT/β-catenin pathway signaling, Axin2-CreERT2 (JAX stock #018867) and Rosa-CAG-LSL-tdTomato (RosaTomato) obtained from the Jackson Laboratory were bred with Runx2GFP or Col1GFP mice (Yang et al., 2019; Mizoguchi, 2024). All animal experimental protocols were approved by the Animal Care Committee of the Institute for Life and Medical Sciences, Kyoto University, and the Committee of Animal Experimentation, Hiroshima University, or Tokyo Dental College, and conformed to the institutional guidelines for vertebrate studies.

TALEN plasmids were constructed using the Platinum Gate TALEN Kit (Kit #1000000043, Addgene, Cambridge, MA, United States of America), as described previously (Sakuma et al., 2013). To prepare TALEN mRNA, TALEN plasmids mSostTALEN-B-L and mSostTALEN-B-R were linearized with SmaI and purified by phenol-chloroform extraction. mSostTALEN-B-L and -R mRNAs were synthesized, and a polyA tail was added using the mMESSAGE mMACHINE T7 ULTRA Kit (Ambion, Austin, TX, United States of America), according to the manufacturer’s instructions. After purification using the MEGAclear kit (Ambion, Austin, TX, United States of America), mSostTALEN-B-L and mSostTALEN-B-R mRNAs were microinjected into the cytoplasm of fertilized eggs obtained from C57BL/6 mice. The injected eggs were then transferred to the oviducts of pseudopregnant surrogate ICR female mice. Genomic DNA was extracted from the tail tips of the founder mice. A 444-bp fragment of exon 1, including recognition sites for TALENs, was amplified by PCR using primers (Sost_GTF1:5′-AAGGCAACCGTATCTAGGCTGG-3′; Sost_GTR1:5′-CCT​CCA​GGT​TCT​AAT​GCT​GTG​CTA​G-3′). The amplified fragments underwent direct sequencing using a BigDye Terminator Cycle Sequencing kit and an ABI 3100 Genetic Analyzer (Applied Biosystems, Foster City, CA, United States of America). Genomic DNA extracted from mouse ear pieces was subjected to PCR using a specific primer set (Sost_GTF3:5′-CCCGTGCCTCATCTGCCTACTTG-3’; Sost_GTR2:5′-TCTTCATCCCGTACCTTTGGC-3′), and the amplified fragments were analyzed using MultiNA (SHIMADZU).

The tibia was dissected from Sost ^Δ26/+ and Sost ^Δ26/Δ26 male mice at P120. The isolated tissue was homogenized in RIPA buffer containing Halt Protease Inhibitor Cocktail (Thermo Fisher Scientific) and Halt Phosphatase Inhibitor Single-Use Cocktail (Thermo Fisher Scientific). The concentrations of the tissue extracts were quantified using a BCA protein assay kit (Takara). Samples and Precision Plus Protein Dual Xtra Prestained protein Standards (Bio-Rad Laboratories) were electrophoresed on a 10% TGX Stain-Free gel (Bio-Rad Laboratories) and transferred to a polyvinylidene fluoride membrane (Bio-Rad Laboratories) using a Trans-Bio Turbo Transfer System (Bio-Rad Laboratories). The membrane was incubated with an anti-mouse SOST/Sclerostin (R&D Systems, AF1589, 1:500) antibody in Bullet Blocking One (Nacalai Tesque) and then an anti-GAPDH antibody (FUJIFILM Wako Pure Chemical Corporation, 1:2000), followed by incubation with horseradish peroxidase-conjugated anti-goat IgG or anti-mouse IgG. Peroxidase activity was detected using the SuperSignal West Pico Chemiluminescent Substrate (Thermo Fisher Scientific).

Anesthetized mice were perfused with 4% paraformaldehyde in phosphate-buffered saline (PFA/PBS) containing 16.6% or 20% sucrose. The specimens were fixed in 4% PFA/PBS containing 16.6% or 20% sucrose at 4°C for 1–3 h, embedded in SCEM or SCEM-L1 (Section-Lab). Undecalcified frozen sections at a thickness of 4 µm were obtained according to Kawamoto’s film method using TC-65 (Leica Microsystems) or SL-T35 (UF) (Section-Lab) tungsten carbide blades, and Cryofilm type 2C (9) or Cryofilm type 3 (16UF) (Section-Lab) (Kawamoto, 2003; Kawamoto and Kawamoto, 2021). After washing with ethanol and PBS, the sections were fixed in 4% PFA/PBS for 5 min and/or decalcified with 0.25 M ethylenediaminetetraacetic acid (pH 8.0). For the detection of GFP, Sox9, Sclerostin, CD31, Ocn, Col1, and Col2 (for P14), sections were treated with hyaluronidase (Sigma–Aldrich) at 37°C. Sections for the detection of Col2 (for P28) were treated with 1 μg/mL of protein kinase K. The sections were fixed in 4% PFA/PBS. The sections treated with hyaluronidase were boiled in 10 mM sodium citrate buffer (pH 6.0). For the detection of GFP, Sox9, Col10, Sclerostin, Ocn, and CD31, the sections were permeabilized in 0.2% Triton X-100 in PBS. The sections were incubated with primary antibodies for 16 h or overnight, washed, and then incubated with the appropriate secondary antibodies conjugated with Alexa Fluor 488 or 594 (Life Technologies, Cell Signaling Technology). Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) (Sigma–Aldrich). The primary antibodies used were anti-GFP (Nacalai Tesque, GF090R; 1:1000), anti-Sox9 (MILLIPORE, AB5535; 1:800), anti-Col1 (ROCKLAND, 600-401-103-0.1; 1:500), anti-Col2 (ROCKLAND, 600-401-104-0.1; 1:500), anti-Col10 (Abcam, ab260040; 1:250), anti-Mouse SOST/Sclerostin (R&D Systems, AF1589; 1:500), anti-Ocn (Takara, M173; 1:800), and anti-CD31 (BD, 553370; 1:2000). Images were captured using a Leica DMRXA microscope equipped with a Leica DFC310 FX camera (Leica Microsystems).

Intraperitoneal injection of Calcein (10 μg/g body weight) (DOJINDO, 348–00434) and Alizarin Complexone (30 μg/g body weight) (TOKYO CHEMICAL INDUSTRY CO., LTD., A3227) diluted in 2% NaHCO₃ was delivered to mice based on experimental designs. Labeled mice were anesthetized and perfused with 4% PFA/PBS containing 16.6% sucrose and fixed in 4% PFA/PBS containing 16.6% sucrose at 4°C for 2 h. Undecalcified frozen sections at a thickness of 4 µm were obtained according to Kawamoto’s film method (Kawamoto, 2003; Kawamoto and Kawamoto, 2021). Nuclei were counterstained with DAPI (Sigma–Aldrich), and the images were captured under a Leica DMRXA microscope equipped with a Leica DFC310 FX camera (Leica Microsystems).

For undecalcified frozen sections, anesthetized mice were perfused with 4% PFA/PBS containing 16.6% or 20% sucrose and fixed in 4% PFA/PBS containing 16.6% or 20% sucrose at 4°C for 2 or 3 h. Undecalcified frozen sections were stained with 0.05% toluidine blue (TB) solution at pH 4.1 (MUTO PURE CHEMICALS CO., LTD.) for 5 min, tartrate-resistant acid phosphatase (TRAP) staining with a TRAP/alkaline phosphatase (ALP) Stain Kit (FUJIFILM Wako Pure Chemical Corporation) for 30 min, or ALP staining with NBT/BCIP solution (Roche; 1:100) for 15 min, followed by Alizarin red (AR) staining prepared from 1% AR Solution at pH 6.3–6.4 (MUTO PURE CHEMICALS CO., LTD.) for 5 min.

Anesthetized male mice were perfused with PBS, and the Achilles tendon entheses were dissected. Specimens were prepared from Sost ^+/+ or Sost ^Δ26/Δ26 mice perfused with PBS, embedded in SCEM (Section-Lab), frozen in hexane (FUJIFILM Wako Pure Chemical Corporation). Undecalcified frozen sections at a thickness of 20 µm were obtained according to Kawamoto’s film method (Kawamoto, 2003; Kawamoto and Kawamoto, 2021). The cell nuclei were stained with Hoechst 33342. For the AFM-based tissue indentation (Ichijo et al., 2022), a JPK BioAFM NanoWizard 3 (Bruker Nano GmbH) was employed. The AFM system was mounted on a bright-field fluorescence microscope (IX81; Evident Co.). AFM cantilevers (TL-CONT; spring constant 0.2 N/m; Nanoworld AG) were modified with glass beads with a diameter of 10 μm and calibrated using the thermal noise method (Butt and Jaschke 1995). To identify the tendon, fibrocartilage, and bone regions, Hoechst-stained tissue sections were observed by IX81 microscope. In particular, AR-stained sections from the same tissue were used to classify the unmineralized and mineralized fibrocartilage regions. For AFM-based indentation, the piezo displacement speed and the sampling rate were set as 3 μm/s and 4,000 Hz, respectively. The obtained indentation force (F) versus depth (h) curve was smoothed using a moving average of 10 datum points before and after each averaging point. Stiffness [nN/μm] was estimated by linear regression for the (F, h) datum points within an indentation depth range of 0 nm ≤ h ≤ 50 nm.

Mice anesthetized at P28 and P120 were perfused with 4% PFA/PBS or PBS and soaked in 99.5% ethanol (Wako, 057–00456). Mice were analyzed by InspeXio SMX-90CT Plus (SHIMADZU) in 99.5% ethanol (Wako, 057–00456) with a 90 kV source voltage and 110 µA source current and a resolution of 0.026 mm/pixel (n = 4 or 6 biological replicates for each group). Two- and three-dimensional reconstructions were performed using Amira 3D software version 2021.1 (Thermo Fisher Scientific). Bone mineral density was calculated from the equation derived from the least-squares method with five plots using hydroxyapatite phantom (RATOC, No06-U5D1mmH) with 100, 200, 300, 400, and 500 mg/cm³ separately scanned on the same day under the same conditions as the samples.

All statistical analyses were performed using Microsoft Excel or GraphPad Prism, version 9 (GraphPad Software, LLC). Data are presented as mean ± SD. Comparisons were performed using the unpaired t-test (for cortical bone thickness and bone mineral density), unpaired t-test with Welch’s correction (for trabecular bone volume/tissue volume), or Mann–Whitney U test (for AFM-based tissue indentation) to determine significance between groups. The level of significance was set at p < 0.05.

In hyaline cartilage, Sox9⁺ chondrocytes produce Col2 and Acan and then mature to become hypertrophic chondrocytes, synthesizing Col10 prior to mineralization (Kozhemyakina et al., 2015). Using Kawamoto’s film method for sectioning undecalcified hard tissues, we compared the expression of these cartilage markers in the calcaneus and their insertion sites of the Achilles tendon by immunostaining (Figure 1). Expansion of the unmineralized fibrocartilage and underlying hyaline cartilage at P7 was visualized using TB staining (Figure 1A). Col1 was co-expressed with Scx in the tendon and fibrocartilage (Figure 1B). Col2 was detected in both the epiphyseal hyaline cartilage and fibrocartilage, whereas Scx was expressed in the upper portion of the Col2⁺ fibrocartilage near the tendon (Figure 1C). Only a small number of ALP⁺ cells were observed at the junction between the unmineralized fibrocartilage and epiphyseal hyaline cartilage (Figure 1D). At P14, hypertrophic chondrocytes were observed in the epiphyseal hyaline cartilage beneath the fibrocartilage (Figure 1E). ALP activity was high in osteoblasts, fibrochondrocytes, and chondrocytes, except in the resting hyaline cartilage (Figure 1F). Chondroclasts/osteoclasts positive for TRAP were found in the mineralized hypertrophic cartilage of the growth plate and primary spongiosa (Figure 1G). Intense Col2 staining was detected in Sox9⁺ cartilage and cartilage remnants around the chondro-osseous junction (Figures 1H–J), whereas Col1 was co-expressed with Scx in the fibrocartilage, tendon, and primary spongiosa (Figure 1K). At P18, Sox9 was expressed in proliferating and resting chondrocytes (Figures 1L,M). In the hyaline cartilage, hypertrophic/mineralized chondrocytes strongly expressed Col10, whereas its expression in the mineralized fibrocartilage was low (Figures 1N,O). Therefore, mineralizing chondrocytes are divided into two distinct groups: Col2⁺/Col10⁺⁺⁺/Col1^– hypertrophic chondrocytes in the epiphyseal hyaline cartilage and Col2⁺⁺/Col10⁺/Col1⁺⁺ fibrochondrocytes. We also observed that sclerostin was expressed in mature fibrochondrocytes at P28 (Figure 1P). Calcein labeling and ALP/AR staining revealed that the mineralization front consisting of ALP⁺ cells extended towards the midsubstance of the Achilles tendon (Figures 1Q,R).

We examined the expression profile of sclerostin during the fibrocartilaginous enthesis (Figure 2). At P14, most fibrochondrocytes were present in the mineralized region adjacent to the mineralized hyaline cartilage, which consisted of hypertrophic chondrocytes (Figures 2A,B). ALP⁺ cells were observed in both the unmineralized and mineralized regions (Figure 2B). Osteocalcin (Ocn) and sclerostin were not detected in either fibrocartilage or hyaline cartilage at P14 (Figure 2C). At P22, a secondary ossification center appeared in the epiphysis, where the mineralized hyaline cartilage was invaded by blood vessels and gradually replaced by bone (Figures 2D,E). More intense ALP staining was observed in the unmineralized and mineralized fibrocartilage, hyaline cartilage, and subchondral bone (Figure 2E). Sclerostin⁺ cells were observed in the Ocn-expressing mineralized fibrocartilage at P22 (Figure 2F). The size of fibrochondrocytes in the mineralized fibrocartilage was much smaller than that of the hypertrophic chondrocytes in the mineralized hyaline cartilage (Figures 2B,E). By P45, the epiphyseal mineralized hyaline cartilage was replaced with bone, and the entheseal mineralized fibrocartilage expanded further (Figures 2G,H). More sclerostin-expressing cells were observed in the OCN-deposited mineralized fibrocartilage (Figure 2I). At P84, the expansion of the unmineralized fibrocartilage above the mineralized fibrocartilage was more evident, in association with a decrease in ALP⁺ cells (Figures 2J,K); however, the plantaris tendon was still ALP⁺ (Figure 2K). Sclerostin expression largely overlapped with Ocn expression in the fibrochondrocytes (Figure 2L). In the epiphysis, the hyaline cartilage was completely replaced by bone (Figures 2H, K), and rapid mineralization of the fibrocartilage without significant cellular hypertrophy was followed by the expansion of the unmineralized fibrocartilage (Figures 2J,K). The expansion of mineralized fibrocartilage is guided by ALP⁺ cells at the mineralization front. This was followed by the expansion of unmineralized fibrocartilage after a decrease in the number of ALP⁺ cells. We then examined the activation of canonical Wnt signaling in the fibrocartilaginous enthesis using 4-week old Axin2-CreERT2;RosaTomato mice with Runx2GFP or Col1GFP reporters. For induction of Cre-recombinase, tamoxifen diet was given for 5 days and then sacrificed at P43 (Figures 2M–O) or P45 (Supplementary Figure S1). Axin2-lineage cells visualized by Tomato expression were found in fibrocartilage as well as subchondral bone and tendon (Figures 2M–O and Supplementary Figures S1B–D). These results suggest that sclerostin is an excellent marker for mature fibrochondrocytes located in the mineralized fibrocartilage adjacent to the subchondral bone, and that activation of canonical Wnt signaling occurs in fibrochondrocytes of the developing enthesis.

The Achilles tendon of Scx-deficient mice were defective and negative for Tnmd, a mature tenocyte marker (Shukunami et al., 2006; Yoshimoto et al., 2017; Shukunami et al., 2018). The loss of Scx leads to defective tendon and enthesis formation, resulting in impaired mechanical outcomes (Killian and Thomopoulos, 2016; Yoshimoto et al., 2017). We analyzed the changes in sclerostin expression together with cartilage and blood vessel markers in the defective Achilles tendon enthesis of Scx ^Δ11/Δ11 mice (Shukunami et al., 2018).

At P14 in Scx ^+/+ mice, the columnar fibrochondrocytes of the enthesis were small, whereas the chondrocytes in the epiphyseal hyaline cartilage of the calcaneus became hypertrophic (Figure 3A). Metachromatic staining with TB was weak in fibrocartilage and strong in hyaline cartilage (Figures 3A,B). In association with defective formation of the Achilles tendon in Scx ^Δ11/Δ11 mice, both fibrocartilaginous enthesis formation and maturation of epiphyseal hyaline cartilage were defective (Figures 3A,B).

At P28, the Achilles tendon enthesis was convex and mineralized in Scx ^+/+ mice, but rounded and unmineralized in Scx ^Δ11/Δ11 mice (Figures 3C–H). Fibrochondrocytes in Scx ^+/+ mice were arranged in a column along the collagen fibers connected to the Achilles tendon (Figure 3C). However, the layer of fibrocartilage with irregularly aligned fibrochondrocytes was thin and unmineralized in Scx ^Δ11/Δ11 mice (Figure 3D). In Scx ^+/+ mice, the replacement of mineralized fibrocartilage with bone was observed in the secondary ossification center (Figures 3C,E,G). In contrast, in Scx ^Δ11/Δ11 mice, cellular hypertrophy and mineralization of epiphyseal hyaline cartilage were delayed, and vascular invasion did not occur (Figures 3D,F,H). At P102, enthesis and epiphyseal bone formation were complete in Scx ^Δ11/+ mice (Figure 3I), while the immature epiphysis was covered with thin ALP⁺ cells in Scx ^Δ11/Δ11 mice (Figure 3J).

We then analyzed Sclerostin, Ocn, CD31 (a marker of vascular endothelial cells), and Col2 localization in Scx ^+/+ or Scx ^Δ11/Δ11 mice at P28 (Figures 3K–R). Wild-type sclerostin⁺ fibrochondrocytes co-expressed Ocn in mineralized entheses that were negative for CD31 (Figures 3M,O,Q) but positive for Col2 (Figure 3K). In Scx ^Δ11/Δ11 mice, vascular invasion did not occur in the epiphyseal cartilage, and sclerostin was faintly co-expressed with Ocn and Col2 (Figures 3L,N,P,R).

Micro-CT imaging at P28 revealed that the enthesis of Scx ^Δ11/Δ11 mice was round and undermineralized compared with Scx ^Δ11/+ mice (Figures 3S,T). These results suggest that mechanical stimulation is essential for the proper development of fibrocartilaginous enthesis, and that sclerostin expression in the avascular mineralized fibrocartilage is closely associated with fibrochondrocyte maturation.

To elucidate the in vivo role of sclerostin in the fibrocartilaginous enthesis, we investigated Sost-deficient mice generated using Platinum TALENs (Sakuma et al., 2013) (Figure 4; Supplementary Figure S2). We designed the TALEN recognition sequences to be within exon 1 of the Sost locus (Figure 4A; Supplementary Figure S2A), so that most of sclerostin would be lost due to a frameshift mutation causing a premature stop codon after creation of a double stranded break by TALENs. Of 40 newborn mice obtained, 16 pups with deletion mutation and 1 pup with an insertion mutation were identified by genotyping and direct sequencing of the amplified DNA. We have established two lines each with a 26-base pair (bp) or 2-bp deletion, both resulting in frameshift causing a premature stop codon shortly downstream (Figure 4A and Supplementary Figures S2A,B, S3A,B). For genotyping of a line with a 26-bp, the wild-type allele corresponds to the upper 211-bp band and the mutant allele to the lower 185-bp band (Figure 4B). Loss of sclerostin expression in Sost ^Δ26/Δ26 mice at P120 was confirmed by Western blotting (Figure 4C; Supplementary Figure S4). Sclerostin localization was observed in the fibrocartilage and bone of Sost ^+/+ mice (Figure 4D) but absent in Sost ^Δ26/Δ26 mice (Figure 4E) at P90.

At P28, the acceleration of replacing mineralized hyaline cartilage with bone was evident in Sost ^Δ26/Δ26 mice compared to Sost ^Δ26/+ mice (Figures 5A–D). To analyze the direction and mineral apposition rate during enthesis formation, we injected two different fluorescent mineralization labels (Calcein and Alizarin complexone) at P17 and P24 in Sost ^Δ26/+ or Sost ^Δ26/Δ26 mice, euthanizes at P25 (Figure 5E). Mineral apposition occurred from the bottom of the enthesis towards the tendon midsubstance, with the mineral apposition rate in Sost ^Δ26/Δ26 mice comparable to that in Sost ^Δ26/+ mice (Figures 5F,G). At P120, while the overall number of ALP⁺ cells decreased, more ALP⁺ cells were observed in both the fibrocartilage and bone of Sost ^Δ26/Δ26 mice compared to Sost ^+/+ mice (Figures 5H–K). Micro-CT images at P120 revealed that the fibrocartilaginous enthesis in Sost ^Δ26/Δ26 mice was comparable to Sost ^+/+ (Figures 5L,M), yet mineralization was enhanced in Sost ^Δ26/Δ26 mice (Figure 5O) compared to Sost ^+/+ mice (Figure 5N). Increased bone mineral density was also observed in Sost ^Δ2/Δ2 mice compared to Sost ^Δ2/+ mice (Supplementary Figure S3).

For quantitative assessment, we scanned the calcaneus of Sost ^+/+ mice (Figure 6A) and Sost ^Δ26/Δ26 mice (Figure 6B) at P120, segmenting the cortex from the trabeculae based on image structure and intensity (Figures 6C,D). Cortical bone thickness (Figure 6E) and trabecular bone volume/tissue volume (Figure 6F) of the calcaneus were significantly higher in Sost ^Δ26/Δ26 mice than in Sost ^+/+ mice. To analyze the mineralization of the fibrocartilaginous enthesis, we extracted an image of the calcaneus epiphysis, including subchondral bone and mineralized fibrocartilage of Sost ^+/+ (Figures 6G,H) or Sost ^Δ26/Δ26 mice (Figures 6I,J) at P120, demonstrating significantly increased bone mineral density in Sost ^Δ26/Δ26 mice compared to Sost ^+/+ mice (Figure 6K).

To assess the stiffness of the enthesis in Sost ^+/+ and Sost ^Δ26/Δ26 mice at P120, we conducted tissue indentation experiments using AFM (Figure 7). Tendon, Unmineralized/mineralized fibrocartilage and bone were identified using Hoechst staining under a fluorescence microscope (Figure 7A). Both unmineralized and mineralized fibrocartilage exhibited higher stiffness in Sost ^Δ26/Δ26 mice, with larger stiffness values, whereas the tendon and bone regions did not show significant differences (Figure 7B).

These findings suggest that Sost/sclerostin plays a crucial role in regulating the degree of mineralization, contributing to the modulation of the fibrocartilaginous enthesis gradient and maintaining mechanical tissue integrity.