A RUNX2 stabilization pathway mediates physiologic and pathologic bone formation

Abstract

The osteoblast differentiation capacity of skeletal stem cells (SSCs) must be tightly regulated, as inadequate bone formation results in low bone mass and skeletal fragility, and over-exuberant osteogenesis results in heterotopic ossification (HO) of soft tissues. RUNX2 is essential for tuning this balance, but the mechanisms of posttranslational control of RUNX2 remain to be fully elucidated. Here, we identify that a CK2/HAUSP pathway is a key regulator of RUNX2 stability, as Casein kinase 2 (CK2) phosphorylates RUNX2, recruiting the deubiquitinase herpesvirus-associated ubiquitin-specific protease (HAUSP), which stabilizes RUNX2 by diverting it away from ubiquitin-dependent proteasomal degradation. This pathway is important for both the commitment of SSCs to osteoprogenitors and their subsequent maturation. This CK2/HAUSP/RUNX2 pathway is also necessary for HO, as its inhibition blocked HO in multiple models. Collectively, active deubiquitination of RUNX2 is required for bone formation and this CK2/HAUSP deubiquitination pathway offers therapeutic opportunities for disorders of inappropriate mineralization.

Fig. 1. CK2 is required for BMSC osteoblast commitment.

a Immunohistochemistry (IHC) for RUNX2 in human HO tissue. IgG was used for negative staining. Scale bar, 100 μm. b–d Human BMSCs expressing control shRNA (shCtrl) or shRNAs targeting CSNK2A1, -2A2, or -2B (shCSNK2A1, shCSNK2A2, shCSNK2B) were cultured under osteogenic conditions. ALP activity was examined at day 7 (b), mineralization activity was assessed by alizarin red staining at day 14 (c), and expression of osteogenic genes was assessed by RT-PCR at day 12 (d) after osteogenic culture. b n = 14; c, n = 8; d, n = 4 biologically independent samples. shCtrl or shCSNK2B-expressing human BMSCs (e) and shCtrl or shCsnk2b-expressing C3H10T1/2 cells (f) were transfected with the RUNX2-responsive reporter gene (OG2-luc) and Renilla in the absence (e) or in the presence (f) of RUNX2 overexpression. Three days after osteogenic culture (e) or 2 days after transfection (f), OG2-luc activity was measured and normalized to a Renilla. (n = 3 biologically independent samples). g Diagram depicting kinetics of CK2 expression over the differentiation of human BMSCs. h, i shCtrl or shCSNK2B-epxressing human BMSCs were cultured under chondrogenic conditions for 21 days. After alcian blue staining, the size of chondrocyte pellets was measured (h). mRNA levels of chondrogenic genes were assessed by RT-PCR (i). Scale bar, 300 μm (h). h n = 5 (shCtrl) or 6 (shCSNK2B); i n = 4 biologically independent samples. j, k shCtrl or shCSNK2B-epxressing human BMSCs were cultured under adipogenic conditions for 12 days. Cells were stained with oil red O (j) and mRNA levels of adipogenic genes were assessed by RT-PCR (k). Scale bar, 100 μm (j). k n = 4 biologically independent samples. Data are representative of two (a) or three (b–f, h–k) independent experiments. Ordinary one-way ANOVA with Dunnett’s multiple comparisons test (a) and a two-tailed unpaired Student’s t test for comparing two groups (c–f, h, i, k; error bars, SD of biological replicates).

Fig. 2. CK2 is required for bone formation during skeletal development.

a Csnk2b mRNA levels in the hindlimbs (femur and tibia) of E17.5 Csnk2b^fl/fl and Csnk2b^Prx1 embryos. (n = 4). b, c Alizarin red/alcian blue staining of skeletal preparations of E17.5 Csnk2b^fl/fl and Csnk2b^Prx1 embryos. Scale bar, 1 mm. Safranin O staining of humeri (d) and femurs (e) of P0 Csnk2b^fl/fl and Csnk2b^Prx1 pups. Scale bars, 250 μm (left) and 50 μm (right, enlarged one). f, g Csnk2b mRNA levels in SSCs (CD45⁻Ter119⁻Tie2⁻αV-Int⁺Thy1⁻6C3⁻CD105⁻CD200⁺) isolated from E17.5 Csnk2b^fl/fl and Csnk2b^Prx1 embryos (f). Frequency of SSCs within the population of total skeletal cells (CD45⁻Ter119⁻Tie2⁻αV-Int⁺) (g). f n = 4; g, n = 6 (Csnk2b^fl/fl) or 5 (Csnk2b^Prx1)). h–j SSCs (CD45⁻Ter119⁻Tie2⁻αV-Int⁺Thy1⁻6C3⁻CD105⁻CD200⁺) isolated from E17.5 Csnk2b^fl/fl and Csnk2b^Prx1 embryonic limbs were transplanted beneath the kidney capsule. MicroCT analysis shows 3D-reconstruction (h) and quantification (i) of bone mass in the kidney capsule. BV bone volume. Histologic sections of kidney capsule were stained with H&E (j, left) or Von Kossa (j, right). The arrow highlights the ectopic bone. Scale bars, 200 μm (h); 100 μm (j). i n = 7 or 14. Data are representative of three (a–e, h, j) independent experiments or are pooled from two experiments (f, g, i). A two-tailed unpaired Student’s t test for comparing two groups (a, f, g, i; error bars, SD of biological replicates).

Fig. 3. CK2 is required for maturation of osteoprogenitors in vitro and in vivo.

a–d COBs isolated from P5 Csnk2b^fl/fl pups were infected with lentiviruses expressing vector (WT) or Cre recombinase (Csnk2b KO) and cultured under osteogenic conditions. mRNA levels of Csnk2b (a) and osteogenic genes (d) were measured by RT-PCR at day 12. ALP activity (b, left) and staining (b, right) or alizarin red staining (c) were performed at day 7 or 18, respectively. Scale bar, 250 μm (b). a, d, n = 4; b, c, n = 6. e Csnk2b mRNA levels in P0 calvaria of Csnk2b^fl/fl, Osx-Cre, and Csnk2b^Osx pups. (n = 4). f Alizarin red/alcian blue staining of skeletal preparations of calvaria (top) and clavicles (bottom) obtained from P10 Csnk2b^fl/fl, Osx-Cre, and Csnk2b^Osx pups. Scale bar, 2 mm. g MicroCT analysis shows 3D-reconstruction of calvaria from 2-month-old Csnk2b^fl/fl, Osx-Cre, and Csnk2b^Osx male mice. The arrows indicate hypomineralization areas. Scale bar, 2 mm. h Radiographic images of clavicles from 2-month-old Osx-Cre and Csnk2b^Osx male mice. The arrow indicates defective sternoclavicular ossification. Scale bar, 2 mm. i, j MicroCT analysis of femurs from 2-month-old Csnk2b^fl/fl, Osx-Cre, and Csnk2b^Osx male mice. Quantification (i) and 3D-reconstruction (j) are displayed. Trabecular bone volume/total volume Tb. BV/TV trabecular thickness, Tb.Th trabecular number per cubic millimeter, Tb.N, and C.Th cortical thickness. Scale bar, 1 mm (j). i Tb. BV/TV, Tb.Th, Tb.N; n = 12 (Csnk2b^fl/fl), 7 (Osx-Cre) or 5 (Csnk2b^Osx), C.Th; n = 13 (Csnk2b^fl/fl), 5 (Osx-Cre) or 5 (Csnk2b^Osx)). k H&E-stained longitudinal sections of femurs from 2-month-old Osx-Cre and Csnk2b^Osx male mice. Scale bar, 50 μm. l, m Histomorphometric analysis of femurs from 2-month-old Osx-Cre and Csnk2b^Osx female mice. Images of calcein/alizarin red-labeled sections of the femur (l) and quantification (m) are displayed. Scale bar, 50 μm (l). BFR/BS bone formation rate per bone surface, MAR mineral reposition rate, Ob.S/BS osteoblast surface per bone surface. (n = 9 (Osx-Cre) or 8 (Csnk2b^Osx)). Similar skeletal phenotypes were observed in both male and female mice (f–m). Data are representative of three (a–h, j–l) independent experiments or are pooled from two experiments (i, m). A two-tailed unpaired Student’s t test for comparing two groups (a–e, i, m; error bars, SD of biological replicates).

Fig. 4. CK2-induced phosphorylation is required for RUNX2 stabilization in osteoblasts.

a–c Csnk2b-sufficient or -deficient COBs were infected with lentiviruses expressing vector- or RUNX2 and cultured under osteogenic conditions for 6 days. ALP activity (a), mRNA levels of osteogenic gene (b), and protein levels of RUNX2 (c) were assessed. a n = 6; b, n = 4 biologically independent samples. d shCtrl or shCsnk2b-expressing C3H10T1/2 cells were transfected with Myc-RUNX2. 2 days after transfection, cells were treated with 10 μM MG132 for 6 h, lysed, immunoprecipitated with anti-Myc conjugated agarose, and immunoblotted with anti-ubiquitin antibody. Protein (e) and mRNA (f) levels of Runx2 in P0 Csnk2b^fl/fl and Csnk2b^Prx1 limbs (top) and P0 Csnk2b^fl/fl and Csnk2b^Osx calvaria (bottom). f n = 4 biologically independent samples). g Wild-type COBs were lysed, immunoprecipitated with anti-IgG control or anti-RUNX2 antibody and protein G-conjugated agarose, and immunoblotted with the indicated antibodies. Asterisk indicates CSNK2B; double asterisks indicate RUNX2; WCL whole cell lysate. h The kinase activity of CK2 was assessed by a cell-free kinase assay using recombinant CK2 (rCK2) and RUNX2 (rRUNX2). i A diagram depicting CK2-induced phosphorylation sites on RUNX2 as determined by phospho-mass spectrometry. j C3H10T1/2 cells were transfected with vector, RUNX2-WT or RUNX2-Tri-A mutant along with OG2-luc and Renilla. 2 days after transfection, OG2-luc activity was measured and normalized to Renilla. (n = 6). k Wild-type COBs were infected with lentiviruses expressing vector, RUNX2-WT or RUNX2-Tri-A mutant, cultured under osteogenic conditions for 6 days, and ALP activity was measured. (n = 6). RUNX2-WT or Tri-A mutant was transfected into C3H10T1/2 cells and 2 days later, protein (l) and mRNA (m) levels of RUNX2 were assessed. m n = 4. n HEK293T cells were transfected with vector, RUNX2-WT or RUNX2-Tri-A mutant along with His-ubiquitin. 2 days after transfection, cells were treated with 10 μM MG132 for 6 h, lysed, immunoprecipitated with Ni-NTA agarose, and immunoblotted with anti-RUNX2 antibody. Data are representative of three (a–h, j–n) independent experiments. A two-tailed unpaired Student’s t test for comparing two groups (a, b, f, m) and ordinary one-way ANOVA with Sidak’s multiple comparisons test (j, k; error bars, SD of biological replicates).

Fig. 5. HAUSP regulates osteoblast differentiation by controlling RUNX2 stability.

a Ingenuity pathway analysis of proteins that interact with RUNX2-WT and RUNX2-Tri-A mutant. b–d Human BMSCs expressing control shRNAs or shRNAs targeting the indicated DUBs were cultured under osteogenic conditions for 7 days and ALP activity was assessed (b). Mineralization (c) and mRNA levels of osteogenic genes (d) in human BMSCs expressing control (shCtrl) or HAUSP-targeting shRNAs (shHAUSP) were assessed by alizarin red staining at day 18 and by RT-PCR at day 12, respectively. b n = 14; d, n = 4 biologically independent samples. Protein (e) and mRNA (f) levels of RUNX2 in shCtrl or shHAUSP-expressing human BMSCs. f n = 4 biologically independent samples). g RUNX2 and His-ubiquitin were transfected into HEK293T cells along with different concentrations of HAUSP. 2 days after transfection, cells were treated with 10 μM MG132 for 6 h, lysed, immunoprecipitated with Ni-NTA agarose, and immunoblotted with anti-RUNX2 antibody. h Cell-free deubiquitination analysis using recombinant HAUSP (rHAUSP) and ubiquitinated Flag-RUNX2 proteins. Flag-RUNX2 and HA-ubiquitin were co-transfected into HEK293T cells and ubiquitinated Flag-RUNX2 proteins were obtained by immunoprecipitation with anti-Flag-conjugated agarose. i HEK293T cells were transfected with RUNX2-WT or RUNX2-Tri-A mutant along with HA-HAUSP and treated with 10 μM MG132 for 6 h prior to lysis. HA-HAUSP immunoprecipitates were immunoblotted with the indicated antibodies. j Cell-free interaction analysis using recombinant RUNX2 (rRUNX2), CK2 (rCK2), and HAUSP (rHAUSP). Recombinant proteins were incubated in the absence or presence of ATP for 30 min, immunoprecipitated with anti-RUNX2 antibody and protein G-conjugated agarose, and immunoblotted with the indicated antibodies. k RUNX2-WT or RUNX2-Tri-A mutant was transfected into HEK293T cells along with different concentrations of HAUSP and immunoblotted with the indicated antibodies. l Schematic diagram depicting a posttranslational regulation of RUNX2 by the CK2/HAUSP pathway in osteoblasts. Data are representative of three (b–k) independent experiments. Ordinary one-way ANOVA with Dunnett’s multiple comparisons test (b) and a two-tailed unpaired Student’s t test for comparing two groups (d, f; error bars, SD of biological replicates).

Fig. 6. HAUSP is required for maturation of osteoprogenitors in vitro and in vivo.

a–e COBs isolated from P5 Hausp^fl/fl pups were infected with lentiviruses expressing vector (WT) or Cre recombinase (Hausp KO) and cultured under osteogenic conditions. RT-PCR analysis to measure mRNA levels of Hausp and osteogenic genes (a) and ALP activity (b, left) and staining (b, right) were performed 7 days after the culture. Mineralization was assessed by alizarin red staining 16 days after the culture (c). Protein (d) and mRNA (e) levels of RUNX2 were assessed in these COBs. Scale bar, 100 μm (b). a, e n = 4; b, c n = 9. f shCtrl or shHausp-expressing C3H10T1/2 cells were transfected with OG2-luc and Renilla in the presence or absence of RUNX2 overexpression. Two days after transfection, OG2-luc activity was measured and normalized to Renilla. (n = 6 biologically independent samples). Hausp-sufficient or -deficient COBs were infected with lentiviruses expressing vector or RUNX2, cultured under osteogenic conditions for 7 days, and ALP activity (g) and mRNA levels of Bglap2 (h) were assessed. g n = 6; h n = 4 biologically independent samples). i Hausp mRNA levels in P0 Hausp^fl/fl, Osx-Cre, and Hausp^+/Osx calvaria. (n = 4). j, l, m MicroCT analysis shows 3D-reconstruction of calvaria (j) and femurs (l) from 2-month-old Hausp^fl/fl, Osx-Cre and Hausp^+/Osx male mice. Quantification of femoral bone mass is displayed (m). The arrows indicate hypomineralization areas. Scale bars, 2 mm (j); 500 μm (l). m n = 10 (Hausp^fl/fl), 5 (Osx-Cre) or 10 (Hausp^+/Osx). k Alizarin red/alcian blue staining of skeletal preparations of clavicles obtained from 2-month-old Hausp^fl/fl, Osx-Cre and Hausp^+/Osx male mice. Scale bar, 2 mm. n H&E-stained longitudinal sections of femurs from 2-month-old Osx-Cre and Hausp^+/Osx male mice. Scale bar, 50 μm. Alizarin red/alcian blue staining of skeletal preparations of calvaria (top) and clavicles (bottom) from P10 Osx-Cre, Csnk2b^+/Osx, Hausp^+/Osx, and Csnk2b^+/Osx;Hausp^+/Osx mice (o) and Osx-Cre, Runx2^+/Osx, Runx2^+/Osx;Csnk2b^+/Osx, and Runx2^+/Osx;Hausp^+/Osx mice (p). Scale bars, 2 mm (o, p). Data are representative of three (a–l, n–p) independent experiments or are pooled from two experiments (m). A two-tailed unpaired Student’s t test for comparing two groups (a–c, e–i, m; error bars, SD of biological replicates).

Fig. 7. CK2/HAUSP pathway is required for acquired HO development.

a H&E-stained sections (left) and immunohistochemistry (right) for CSNK2A and HAUSP in human HO tissue. Scale bar, 50 μm. b Achilles tenotomy was performed at 3-month-old male mice. Eight hours after the injury, Achilles tendon was dissected and mRNA levels of Csnk2a1, and Hausp were measured by RT-PCR. None, non-tenotomized Achilles tendon; HO, tenotomized Achilles tendon. (n = 4 (Csnk2a1) or 6 (Hausp)). c–e Quadriceps muscle injury by an aluminum ball drop was performed in 3-month-old Csnk2b^fl/fl, Osx-Cre, and Csnk2b^Osx male mice, followed by injection of a mixture of rBMP2/7 and matrigel to the injured muscle. HO was assessed by microCT and histology 3 weeks post injury. 3D-reconstruction (c) and quantification (d) and H&E-stained sections of HO areas in the muscle (e) are displayed. The arrows indicate HO in muscle. B bone, M muscle, C cartilage, BM bone marrow. Scale bars, 1 mm (c); 100 μm (e). d n = 7 (Csnk2b^fl/fl), 7 (Osx-Cre) or 4 (Csnk2b^Osx). f–h Achilles tenotomy was performed in 3-month-old Csnk2b^fl/fl, Osx-Cre, and Csnk2b^Osx male mice following with a remote burn injury on the back using a heated aluminum block. HO was assessed by microCT and histology 8 weeks post injury. 3D-reconstruction (f) and quantification (g) and H&E-stained sections of HO areas in the boxes of the microCT images (h) are displayed. None non-tenotomized leg, A Achilles tendon, BM bone marrow, B bone. Scale bars, 1 mm (f); 100 μm (h). g n = 11 (Csnk2b^fl/fl), 7 (Osx-Cre) or 6 (Csnk2b^Osx). i–l 3-month-old wild-type male mice were daily treated with DMSO (Veh) or an inhibitor of CK2 (i-CK2, 2.5 mg/kg) or HAUSP (i-HAUSP, 2 mg/kg) via intraperitoneal (i.p.) injection one day after muscle injury and BMP2/7-matrigel injection (i, j) or burn injury and Achilles tenotomy (k, l). HO was assessed by microCT analysis 3 weeks (i, j) or 8 weeks (k, l) post injury. 3D-reconstruction (i, k) and quantification (j, l) are displayed. Scale bars, 1 mm (i, k). j, n = 7 (DMSO), 7 (i-CK2), or 6 (i-HAUSP); l n = 5 (DMSO), 6 (i-CK2) or 7 (i-HAUSP). Data are representative of three (a–c, e, f, h, i, k) independent experiments or are pooled from two experiments (d, g, j, l). A two-tailed unpaired Student’s t test for comparing two groups (b, d, g) and ordinary one-way ANOVA with Dunnett’s multiple comparisons test (j, l; error bars, SD of biological replicates).