Lansoprazole Upregulates Polyubiquitination of the TNF Receptor-Associated Factor 6 and Facilitates Runx2-mediated Osteoblastogenesis

Abstract

The transcription factor, runt-related transcription factor 2 (Runx2), plays a pivotal role in the differentiation of the mesenchymal stem cells to the osteochondroblast lineages. We found by the drug repositioning strategy that a proton pump inhibitor, lansoprazole, enhances nuclear accumulation of Runx2 and induces osteoblastogenesis of human mesenchymal stromal cells. Systemic administration of lansoprazole to a rat femoral fracture model increased osteoblastogenesis. Dissection of signaling pathways revealed that lansoprazole activates a noncanonical bone morphogenic protein (BMP)-transforming growth factor-beta (TGF-β) activated kinase-1 (TAK1)-p38 mitogen-activated protein kinase (MAPK) pathway. We found by in cellulo ubiquitination studies that lansoprazole enhances polyubiquitination of the TNF receptor-associated factor 6 (TRAF6) and by in vitro ubiquitination studies that the enhanced polyubiquitination of TRAF6 is attributed to the blocking of a deubiquitination enzyme, cylindromatosis (CYLD). Structural modeling and site-directed mutagenesis of CYLD demonstrated that lansoprazole tightly fits in a pocket of CYLD where the C-terminal tail of ubiquitin lies. Lansoprazole is a potential therapeutic agent for enhancing osteoblastic differentiation.

Keywords: CYLD; Drug repositioning; Lansoprazole; Osteoblastogenesis; Runx2; TRAF6.

Supplemental Fig. 1

Effects of lansoprazole on the activity of human RUNX2 P1 promoter, nuclear accumulation of Runx2 in mouse and human mesenchymal stromal cells, and expressions of Runx2 gene in rat primary bone marrow cells. (a) Lansoprazole activates human RUNX2 P1 promoter in a dose-dependent manner. Firefly luciferase activities of BMP-2-stimulated C3H10T1/2 cells that stably express pGL4-hRunx2P1-luc2 (C3H10T1/2-hRunx2P1 cells), in which the human RUNX2 P1 promoter is fused to firefly luciferase cDNA. Luciferase activities are normalized to the mean of vehicle. Cells were treated with the indicated concentrations of lansoprazole for 1 d. *p < 0.001 by the Jonckheere–Terpstra trend test over the indicated concentration range of lansoprazole. (b) Runx2-specific immunofluorescence staining of C3H10T1/2 cells treated with lansoprazole for 3 d. Cells were pretreated with or without 100 ng ml^− 1 of recombinant BMP-2 for 6 d. Scale bar, 50 μm. (c) Runx2-specific immunofluorescence staining of human MSCs treated with lansoprazole for 3 d. Cells were pretreated with or without 100 ng ml^− 1 of recombinant BMP-2 for 3 d. Scale bar, 50 μm. (d) Lansoprazole upregulates the expression of Runx2 in rat bone marrow cells. Expression levels of endogenous Runx2 mRNA by real-time RT-PCR in rat bone marrow cells. Cells were treated with lansoprazole for 2 d. The relative expression levels are normalized to the mean of vehicle. A sequential protocol for culturing cells is shown at the bottom. Lower arrowheads indicate cell splitting. *p = 0.012 (Days 28–34) by the Jonckheere–Terpstra trend test above the evaluated concentration ranges of lansoprazole. NS, not significant.

Supplemental Fig. 2

Changes in body weights of vehicle- and lansoprazole-administered rats in a rat model of femoral fracture (n = 8 per group). No significance difference was observed between two groups by two-way repeated measures ANOVA.

Supplemental Fig. 3

Radiological and histomorphometrical analyses of the fractured and unfractured femurs in vehicle- and lansoprazole-administered rats. (a) Radiological assessment of total callus areas on the anteroposterior and lateral radiographs after a 4-week administration of lansoprazole. (b) Osteoclastic parameters at the injured metaphyses of the fractured femurs after a 4-week administration of lansoprazole. (c) Osteoblastic parameters at the intact metaphyses of the unfractured femurs after a 4-week administration of lansoprazole. (d) Histomorphometric parameters at the intact metaphyses of the unfractured femurs after a 12-week administration of lansoprazole. In (a–d), lanso., lansoprazole; and NS, not significant.

Supplemental Fig. 4

Effects of lansoprazole on chondrogenesis and osteoclastogenesis during an early phase of fracture healing in a rat model of femoral fracture. (a) Representative images of cartilaginous callus formation at the fracture sites detected by alcian blue staining at 2 weeks after fracture (n = 4 per group). Dark blue signals show cartilage tissues. Scale bar, 5 mm. (b) Representative images of osteoclast formation at the fracture sites detected by TRAP staining at 2 weeks after fracture (n = 4 per group). Red signals show TRAP-positive cells. Scale bar, 5 mm. (a) and (b) are serial sections. (c) Morphometric analysis of (a) and (b). Total callus area was defined as the sum of bridging callus and the anchoring callus. Cartilaginous callus area was normalized to total callus area. The number of TRAP-positive cells was normalized to the fracture area, which was defined as the sum of the total callus area and the underlying cortical bone area. NS, not significant. *p < 0.02 by unpaired t-test.

Supplemental Fig. 5

TRAF6-anchored Lys63-linked autopolyubiquitination in cultured cells and effects of TRAF6-inhibitory peptide on lansoprazole-activated human RUNX2 P1 promoter activity. (a) Lansoprazole induced TRAF6-anchored polyubiquitination with BMP-2 induction in primary mouse calvarial osteoblasts. (b) The immunoprecipitated TRAF6 in Fig. 5c was immunoblotted with an antibody against Lys63-linked polyubiquitin chains. (c) The immunoprecipitated TRAF6 in Fig. 5d was immunoblotted with an antibody against Lys63-linked polyubiquitin chains. (d) The inhibitor peptide that specifically blocks binding of TRAF6 to BMPRI has no effect on lansoprazole-induced activation of RUNX2 P1 promoter in C3H10T1/2-hRunx2P1 cells. The cells were pretreated with the indicated amounts of the inhibitor peptide for 1 d and subsequently treated with 20 μM lansoprazole for 1 d. Statistical significance is determined by the Jonckheere–Terpstra trend test above the evaluated concentration range of lansoprazole. NS, not significant.

Supplemental Fig. 6

Ubiquitination ligase activity of recombinant TRAF6 and deubiquitination activity of recombinant wild-type and mutant CYLDs. (a) Purified recombinant Flag-tagged TRAF6 catalyzed TRAF6-anchored polyubiquitination. (b) Purified recombinant Flag-tagged wild-type CYLD exerted deubiquitination activity. (c) Purified recombinant Flag-tagged R758A-single-mutant CYLD retained deubiquitination activity. (d) Purified recombinant Flag-tagged R758A and F766A-double-mutant CYLD retained deubiquitination activity.

Supplemental Fig. 7

Changes in body weights of unoperated rats, which were administered with either vehicle or lansoprazole for 12 weeks (n = 4 per group). No significance difference was observed between two groups by two-way repeated measures ANOVA.

Fig. 1

Lansoprazole induces nuclear accumulation of Runx2 and activates its transcriptional activity. (a) Immunoblotting of Runx2 in the nuclear and cytosolic fractions as well as in the whole cell lysate of HOS cells after adding 20 μM lansoprazole for 2 d. The relative expression levels were densitometrically estimated using an essential splicing factor U2AF35 for the nuclear fraction and Gapdh for the cytosolic fraction and the whole cell lysate as a control. Lanso., lansoprazole. (b) Runx2-specific immunofluorescence staining of HOS cells treated with lansoprazole for 3 d. Scale bar, 50 μm. (c) Activity of Runx2-responsive luciferase reporter vector (6xOSE2-luc) in C3H10T1/2 cells. Cells were exposed to 20 μM of lansoprazole for 1 d without adding BMP-2. pCMV6-XL5-Runx2, human RUNX2 cDNA clone. (d) Expression levels of endogenous Spp1 mRNA by real-time RT-PCR in C3H10T1/2 cells. Cells were pretreated with 100 ng mL^− 1 of recombinant BMP-2 for 3 d, followed by incubation with the indicated concentrations of lansoprazole for 14 d. The relative expression levels were normalized to the mean of vehicle. (e) ALP activity of cell lysates of human bone marrow-derived mesenchymal cells. Cells were cultured in osteogenic medium for 19 d and lansoprazole was added from days 15 to 19. The relative activity was normalized to the mean of vehicle. (f) ALP staining of human bone marrow-derived mesenchymal cells, which were treated with the indicated concentrations of lansoprazole. Cells were cultured in osteogenic medium for 24 d and lansoprazole was added from days 18 to 24. Scale bar, 5 mm. In (c–e), the mean ± SD (n = 6) are indicated. *p < 0.0001 by unpaired t-test (c) and by the Jonckheere–Terpstra trend test over the indicated concentration range of lansoprazole (d and e). (d, e, and f) In the course of our analysis, 5–20 μM lansoprazole exhibits consistent dose–response effects, but data at concentrations less than 5 μM and more than 20 μM have not been omitted under the expectation that these concentrations will be of some help for future references.

Fig. 2

Lansoprazole upregulates the expression of Runx2 and matrix calcium deposition. (a) Expression levels of endogenous Runx2 mRNA by real-time RT-PCR in BMP-2-pretreated C3H10T1/2 cells. Cells were treated with lansoprazole for 2 d. The relative expression levels were normalized to the mean of vehicle. (b) Expression levels of endogenous RUNX2 mRNA by real-time RT-PCR in HOS cells. Cells were treated with lansoprazole for 2 d. The relative expression levels were normalized to the mean of vehicle. (c) Expression levels of endogenous Runx2 protein in BMP-2-pretreated C3H10T1/2 cells. Cells were treated with lansoprazole for 3 d. Densitometric values were normalized to that of Gapdh and also to vehicle. (d) Expression levels of endogenous Runx2 protein in HOS cells. Cells were treated with lansoprazole for 3 d. Densitometric values were normalized to that of Gapdh and also to vehicle. (e) Expression levels of endogenous RUNX2 mRNA by real-time RT-PCR in human MSCs. The relative expression levels were normalized to the mean of vehicle. Sequential protocol of lansoprazole administration is indicated at the bottom. Lower arrowheads indicate the time points when we passaged cells. (f) Matrix calcium deposition by Alizarin red staining in human bone marrow-derived mesenchymal cells that were induced to differentiate immediately after isolation. Cells were cultured in osteogenic medium for 21 d and lansoprazole was added from days 15 to 21. Scale bars, 5 mm for dishes and 200 μm for magnified images. Note that the temporal protocol of human bone marrow-derived mesenchymal cells in (f) is different from that of human MSCs in (e), because the time courses of differentiation of these cells are different. (g) Matrix calcium deposition by Alizarin red staining in human bone marrow-derived mesenchymal cells that were induced to differentiate after proliferation for 23 d. Cells were proliferated in non-osteogenic medium for 23 d followed by additional culture in osteogenic medium up to day 44, and lansoprazole was added from days 24 to 44. In (f), cells were differentiated immediately after isolation, whereas in (g), cells were first proliferated for 23 d to increase the number of cells. Scale bars, 5 mm. In (a, b, and e), the relative expression levels were normalized to the mean of vehicle, and the mean ± SD (n = 6) are indicated. *p < 0.01 (a), *p < 0.0001 (b), *p < 0.01 (c, days 0–10), *p < 0.001 (c, days 11–17), and *p < 0.002 (c, days 18–24) by the Jonckheere–Terpstra trend test over the indicated concentration range of lansoprazole. In (c and d), the densitometric values were normalized to that of Gapdh and also to vehicle.

Fig. 3

Orally administered lansoprazole accelerates fracture healing in a rat model of femoral fracture. (a) Representative radiographs of vehicle- and lansoprazole-administered rats at 1, 2, 3, and 4 weeks after fracture. Formation of seamless bridging callus (black arrowhead) and incomplete formation of uniting callus (white arrow) are shown. The proportion of rats with the fracture union is indicated in the inset. LAT, lateral view, AP, anteroposterior view. (b) The number of rats showing the indicated number of seamless bridging calluses (left) and complete uniting calluses (right) (n = 6 per group). Bridging callus formation on 4 cortices was evaluated on the anteroposterior and lateral radiographs, and the number of calluses was counted (0 to 4). Uniting callus formation was defined by the disappearance of fracture line on the anteroposterior and lateral radiographs, and the number of calluses was counted (0 to 2). (c) Representative images of new bone formation within the anchoring callus assessed by double calcein labeling at 4 weeks after fracture. *p < 0.03 by unpaired t-test.

Fig. 4

Orally administered lansoprazole accelerated osteoblastic differentiation in a rat model of femoral fracture. (a) Undecalcified bone histology of the fracture sites in vehicle- and lansoprazole-administered rats (Villanueva Goldner staining) 4 weeks after fracture. Green signals show calcified bones. Red signals show granulation tissues spanning edges of osteotomy. Note newly formed bone (arrows) in the lansoprazole-treated rat. Scale bar, 1 mm (upper) and 200 μm (lower). (b) Histomorphometric analyses of undecalcified bone histology of the fracture sites showing enhanced bone formation in lansoprazole-treated rat femurs. The mean ± SD (n = 8) are indicated. *p < 0.02 (BV/TV), *p < 0.04 (Ob.S/BS), and *p < 0.05 (OS/BS) by unpaired t-test. Lanso., lansoprazole. (c) Undecalcified bone histology of the metaphyses of the fracture femur in vehicle- and lansoprazole-administered rats (Villanueva Goldner stain) 4 weeks after fracture. Red signals show osteoid bones. Note increased osteoid formation in the lansoprazole-treated rat. Scale bar, 100 μm. (d) Histomorphometric analyses of undecalcified bone histology of the metaphyses of the fractured femur. The mean ± SD (n = 8) are indicated. *p < 0.04 (OV/BV), *p < 0.02 (OS/BS) and *p < 0.05 (N.Oc/BS) by unpaired t-test. Lanso., lansoprazole; and NS, not significant.

Fig. 5

Lansoprazole upregulates the noncanonical BMP–TAK1–p38 MAPK pathway. (a) Expression levels of endogenous Runx2 mRNA by real-time RT-PCR in native C3H10T1/2 cells without BMP-2 pretreatment. Cells were treated with the indicated concentrations of lansoprazole for 2 d. (b) Expression levels of endogenous Runx2 mRNA by real-time RT-PCR in native C3H10T1/2 cells pretreated with 400 ng mL^− 1 of noggin for 2 d. Cells were treated with the indicated concentrations of lansoprazole for 2 d. (c) Effects of BMP signaling-associated protein kinase inhibitors on lansoprazole-induced increases of RUNX2 P1 promoter activity. C3H10T1/2 cells that stably express pGL4-hRunx2P1-luc2 (C3H10T1/2-hRunx2P1 cells) without BMP-2 pretreatment were treated with the indicated kinase inhibitors for 1 h and subsequently treated with 20 μM lansoprazole for 1 d. In addition, we expected future clinical application of lansoprazole, and we decided to use 20 μM in later analyses. The luciferase activities were normalized to those without lansoprazole and the kinase inhibitors and also to the protein amount of cell lysates. BMPRI kinase inhibitor, dorsomorphin (1 μM); Erk1/2 inhibitor, U0126 (20 μM); p38 inhibitor, SB203580 (20 μM); JNK inhibitor, SP600125 (10 μM); and TAK1 inhibitor, 5Z-7-oxozeaenol (1 μM). *p < 0.0001 by unpaired t-test. (d) Phosphorylation of BMP signaling-associated proteins in human MSCs. Relative levels indicate densitometric ratios of phosphorylated to total proteins normalized to the ratio without lansoprazole. (e) Effects of TAK1 and p38 MAPK inhibitors on lansoprazole-induced increases of ALP activity in human MSCs. *p < 0.02 and **p < 0.0003 by unpaired t-test. Human MSCs (LONZA) were cultured in the osteogenic medium with or without 20 μM lansoprazole, 100 nM TAK1 inhibitor (5Z-7-oxozeaenol), and 5 μM p38 MAPK inhibitor (SB203580) for 14 d. (f) Lansoprazole activates interactions of Runx2 and CBP. HOS cells were treated with or without 20 μM of lansoprazole for 2 d, and immunoprecipitated with anti-Runx2 antibody. Immunoprecipitated CBP was visualized by immunoblotting. In (a–c), the mean ± SD (n = 6) are indicated. NS, not significant by the Jonckheere–Terpstra trend test (a and b).

Fig. 6

Lansoprazole upregulates TRAF6 polyubiquitination. (a) Activity of NF-κB-responsive-element (RE) luciferase reporter vector in C3H10T1/2 cells. Lanso., lansoprazole (20 μM); TNF-α (1 ng mL^− 1). *p < 0.0001 by unpaired t-test. (b) Effects of a dominant negative TRAF6 (pDeNy-hTRAF6) on a lansoprazole-induced increase of the RUNX2 P1 promoter activity. A relative luciferase activity was normalized to that of vehicle, and a difference to the vehicle is shown by the mean and SD (n = 6). Lanso., lansoprazole. (c) Lansoprazole enhanced TRAF6-anchored polyubiquitination with BMP-2 induction in HEK293 cells. (d) Lansoprazole induced TRAF6-anchored polyubiquitination without BMP-2 induction in HEK293 cells.

Fig. 7

Lansoprazole stably fits in a pocket of CYLD and attenuates its deubiquitination activity in vitro. (a) Lansoprazole was incapable of enhancing TRAF6-anchored autopolyubiquitination in vitro. (b) Lansoprazole suppressed the deubiquitination activity of CYLD. (c) Simulated crystal structure of CYLD (PDB id: 2VHF) bound to ubiquitin (blue) (PDB id: 1NBF). The C-terminal tail of ubiquitin extends to the active site (yellow). Lansoprazole (green) fits into a pocket to cross the C-terminal tail of ubiquitin. (d) Enlarged image of a pocket where lansoprazole (sticks) fits. Note that lansoprazole crosses the C-terminal tail of ubiquitin (blue). R758 (red) and F766 (purple) are artificially mutated to alanine to deform the lansoprazole pocket. (e) The wild-type CYLD retained the suppressive effects of lansoprazole on CYLD. (f) The R758A-single-mutant CYLD partially abrogated the suppressive effects of lansoprazole on CYLD. (g) The R758A and F766A-double-mutant CYLD completely abolished the suppressive effects of lansoprazole on CYLD.

Fig. 8

Proposed model of lansoprazole-induced activation of Runx2. Lansoprazole activates TRAF6-catalyzed Lys63-linked autopolyubiquitination of TRAF6 itself by inhibiting a specific deubiquitination enzyme, CYLD. Lys63-linked polyubiquitinated chains of TRAF6, in turn, trigger autophosphorylation of TAK1. Unlike Lys48-linked polyubiquitination in protein degradation, Lys63-linked polyubiquitination of TRAF6 activates TRAF6. The TAK1–p38 MAPK pathway, a noncanonical BMP pathway, then activates transcriptional activity of Runx2, which requires recruitment of a transcriptional coactivator, CBP. The binding of BMP ligand to BMPRI and BMPRII markedly potentiates lansoprazole-induced activation of Runx2.

Fig. 9

Lansoprazole activates osteoclast formation and bone resorption activity. (a) Effects of lansoprazole on osteoclast formation and maturation. RAW 264.7 cells were cultured for 4 d with or without lansoprazole in the presence of 100 ng mL^− 1 RANKL. TRAP-positive cells with more than 3 nuclei were counted as osteoclasts. The numbers of total and mature osteoclasts are shown on the top. Representative TRAP-stained cell images are shown at the bottom. †p < 0.0000001, *p < 0.0005, **p < 0.04, and ***p < 0.0001 by one-way ANOVA with post-hoc Dunnett analysis. Scale bar, 1 mm. (b) Effects of lansoprazole on osteoclast function. RAW 264.7 were cultured on a calcium phosphate-coated plate with or without lansoprazole in the presence of 100 ng mL^− 1 RANKL for 7 d. Sizes of the eroded areas are shown on the top. Representative images of the resorption pits are shown at the bottom. Scale bar, 2 mm. *p < 0.04, **p < 0.03, and ***p < 0.02 by one-way ANOVA with post-hoc Dunnett analysis.

Fig. 10

Systemic administration of lansoprazole enhances osteoclastogenesis and induces osteomalacia-like change at the metaphyses of unfractured rat femurs. (a) Histomorphometric analyses of undecalcified histology in rats that were administered with lansoprazole for 4 weeks. The mean ± SD (n = 4) are indicated. *p < 0.05 (ES/BS) and *p < 0.03 (Oc.S/BS) by unpaired t-test. Lanso., lansoprazole; and NS, not significant. (b) Histomorphometric analyses of undecalcified histology in rats that were administered with lansoprazole for 12 weeks. The mean ± SD (n = 4) are indicated. *p < 0.02 (Tb.Th) and *p < 0.04 (O.Th) by unpaired t-test. Lanso., lansoprazole; and NS, not significant.