Acceleration of bone regeneration by activating Wnt/β-catenin signalling pathway via lithium released from lithium chloride/calcium phosphate cement in osteoporosis

By virtue of its excellent bioactivity and osteoconductivity, calcium phosphate cement (CPC) has been applied extensively in bone engineering. Doping a trace element into CPC can change physical characteristics and enhance osteogenesis. The trace element lithium has been demonstrated to stimulate the proliferation and differentiation of osteoblasts. We investigated the fracture-healing effect of osteoporotic defects with lithium-doped calcium phosphate cement (Li/CPC) and the underlying mechanism. Li/CPC bodies immersed in simulated body fluid converted gradually to hydroxyapatite. Li/CPC extracts stimulated the proliferation and differentiation of osteoblasts upon release of lithium ions (Li+) at 25.35 ± 0.12 to 50.74 ± 0.13 mg/l through activation of the Wnt/β-catenin pathway in vitro. We also examined the effect of locally administered Li+ on defects in rat tibia between CPC and Li/CPC in vivo. Micro-computed tomography and histological staining showed that Li/CPC had better osteogenesis by increasing bone mass and promoting repair in defects compared with CPC (P < 0.05). Li/CPC also showed better osteoconductivity and osseointegration. These findings suggest that local release of Li+ from Li/CPC may accelerate bone regeneration from injury through activation of the Wnt/β-catenin pathway in osteoporosis.


Results
Composition and morphology of Li/CPC. At 1 day after immersion in simulated body fluid (SBF), the entangled structure of irregular crystals and irregular granular pattern were observed on the surfaces of cements under scanning electron microscopy (SEM), suggesting that the original materials were the major components of initial hardened cements. A typical flower-like structure was observed at 7 and 14 days, suggesting that HA with a poor crystalline structure had been formed gradually. Judging by the morphology of a plate-like structure at 28 days, the surface of cements was almost completely covered by precipitated HA (Fig. 1). In addition, X-ray diffraction (XRD) showed that the original material peaks were present and that the peak representing HA was absent after 1 day of immersion. This finding suggested that the composition of cements was initial materials [tetracalcium phosphate (TTCP), anhydrous dicalcium phosphate (DCPA)], and that HA did not form in the initial hardening of cements ( Fig. 2A). After immersion in SBF for 14 days, the intensity of peaks of TTCP and DCPA decreased gradually. Simultaneously, a broad peak at ≈ 32° was observed, suggesting HA formation, which was consistent with the morphology of the cements under SEM (Fig. 2B). Combined with the results shown above, which consistently reflected HA formation, we concluded that Li/CPC had favourable bioactivity in vitro to the same extent as that of CPC, a finding that is in accordance with other reports 23 . Morphology of MC3T3-E1 cells. attaching to cements was examined using SEM. MC3T3-E1 cells were observed on the cement surface in all groups at 1 day. They showed well-spread and stretched filopodia to anchor to scaffold surfaces. At 7 days, cells spread to most scaffold surfaces, connected to adjacent cells or stretched numerous pseudopodia to anchor to scaffold surfaces. At 14 days, scaffold surfaces were covered completely with MC3T3-E1 cells in all groups and connected to each other via plasma extensions. This finding suggested that the cell-compatibility of all groups was satisfied (Fig. 3).
Release of lithium ions (Li + ) from cement. Li + released from cement with a higher concentration of lithium in Li/CPC was greater in extracts of materials at the same time. Most Li + were released within 1 day and only a small amount was released after 7 days (Fig. 4A).
Proliferation of MC3T3-E1 cells. At 1 day, cell number was maintained at the same level in all groups.
Compared with CPC, the cell number in Li/CPC increased at 3, 5, and 7 days with Li + release at 25.35 ± 0.12 to 50.74 ± 0.13 mg/l with culture time. However, a too-high concentration of Li + (102.41 ± 0.11 mg/l) began to elicit cytotoxicity. Cell number in Li/CPC-200 was lower than that in Li/CPC-50 and Li/CPC-100, and the rate of cell proliferation in Li/CPC-200 was not significantly different compared with CPC (P > 0.05). The rate of cell proliferation on Li/CPC-100 was the highest of all groups tested (P < 0.05) (Fig. 4B).
Differentiation of MC3T3-E1 cells. At 7 days, alkaline phosphatase (ALP) activity in Li/CPC was significantly higher than that in CPC. Differentiation on Li/CPC-100 was highest (P < 0.05) in all groups tested, and differentiation in Li/CPC-200 was not significantly different compared with CPC (P > 0.05) (Fig. 4C). Osteogenic differentiation and mineralisation of Li/CPC increased as Li + were released at 25.35 ± 0.12 to 50.74 ± 0.13 mg/l, which reflected the calculation of ALP activity and staining with alizarin red (Fig. 4D,E). Figure 4F shows the cytoskeletons of MC3T3-E1 cells after immersion for 12 h in different extracts. Compared with CPC control, the attached cells for Li/CPC-50 and Li/CPC-100 extracts showed greater spreading and superior extension of filopodia, and greater focal adhesion via well-organised F-actin stress fibres (red filaments). However, the cytoskeletons of cells cultured with Li/CPC-200 were similar to those for CPC because the Li + concentration increased excessively, and there was slight deterioration of MC3T3-E1 cells.

Staining of the cytoskeleton proteins of MC3T3-E1 cells.
Expression of osteogenic genes. Osteogenesis-related gene expression of osteoblastic markers [collagen type I alpha 1 (Col1a1), bone gamma-carboxyglutamate protein (Bglap), osteoprotegerin (OPG), runt-related transcription factor 2 (Runx2), β -catenin] were detected after incubation for 3 and 7 days with MC3T3-E1 cells in different extracts. In general, gene expression was time-dependent. After 3 days, expression of osteoblastic  Wnt/β-catenin pathway. We examined the effect of Li+ released from Li/CPC on activation of Wnt/β -catenin pathway. As the amount of phosphorylated GSK-3β increased, the amount of phosphorylated β -catenin decreased. This action stabilised β -catenin aggregated in the cytoplasm, which translocated to the nucleus when Li + release was 25.35 ± 0.12 to 50.74 ± 0.13 mg/l (Fig. 5F,G). Li/CPC increased Runx2 expression significantly (P < 0.05) (Fig. 5H). Runx2 is essential for osteogenic differentiation as a Wnt/β -catenin/TCF target gene product. This result was similar to the result of Runx2 gene expression.
Rat model of osteoporosis. Histomorphometric parameters and micro-architectural properties of bone were analysed 3 months after surgery via micro-computed tomography (CT) and histological staining of tibia-tissue sections. Bone mineral density (BMD), relative bone volume (BV/TV), trabecular number (Tb.N), trabecular thickness (Tb.Th) and connectivity density (Conn.D) decreased 3 months after surgery relative to sham rats, whereas trabecular separation (Tb.Sp) increased (P < 0.05). Three-dimensional (3D) micro-CT showed that rats that had undergone ovariectomy (OVX) had significantly less trabecular bone compared with sham-operated rats. Compared with the sham-operated group, histology images of tibias at 3 months post-OVX showed significantly sparser bone trabecular in OVX rats, which was consistent with 3D micro-CT results (Fig. 6).

Micro-CT.
Bone defects were created by implantation of cylindrical material; such implantation was done in OVX rats. Micro-CT images and 3D computer models of tibial defects upon implantation were used to evaluate regenerated bone mass. At 4 weeks, a small amount of regenerated osseous tissue was found, and more extensive, newly formed bone occurred at 8 weeks (Fig. 7). Increased formation of new bone was detected around the Li/ CPC compared with CPC at 4 and 8 weeks. The BV/TV at different distances from the surface of Li/CPC was significantly higher than that for CPC at 4 and 8 weeks (P < 0.05), which suggested that Li/CPC had better capacity for bone regeneration at interfacial areas.
Histological staining. OVX rats that had undergone filling of bone defects were killed. Specimens of the proximal tibia were collected for staining [haemotoxylin and eosin (H&E), Giemsa]. At 4 weeks, fibrous tissue was seen to infiltrate into a small gap between cement and bone in CPC. However, the initial gap was occupied entirely by new bone in Li/CPC. At 8 weeks, regenerated osseous tissue anchored to the surface of the implant was found in CPC and Li/CPC groups. However, this regenerated osseous tissue penetrated into the initial outer surface around Li/CPC, which suggested that Li/CPC could accelerate bone regeneration (Fig. 8).

Discussion
CPC has been used for synthetic bone grafts in bone engineering because of its excellent bioactivity and osteoconductivity 25-28 . Li/CPC was manufactured by doping lithium chloride onto CPC. After immersion, testing of the hardened cement body was done to investigate bioactivity in vitro. With an increase in immersion time, cement morphology was changed gradually to the typical morphology seen in HA. This change was in accordance with the XRD-peak characteristics of HA, suggesting formation of a poorly crystalline apatite. HA has good biocompatibility and bioactivity 29 . Its chemical composition and crystalline structures are similar to those of apatite in human bone. Also, the molar ratio of calcium: phosphorus in HA is 1.67, which is close to that of human bone 30 . An exchange reaction of calcium ions in HA can occur with the carboxyl group present in amino acids, proteins, and organic acids, and new bone can be combined with implant material to provide a good growth interface for bone cells 31,32 . In addition, MC3T3-E1 cells exposed to cements also showed good biocompatibility at 14 days.
The Wnt/β -catenin pathway has an essential role in regeneration of osseous tissue by stimulation of the proliferation and differentiation of osteoblasts 33,34 . If signalling is blocked, phosphorylation of β -catenin is induced by a protein complex consisting of axin, adenomatous polyposis coli, and GSK-3β 6. Lithium chloride can activate the Wnt/β -catenin pathway by inhibition of GSK-3β activity to stimulate the proliferation and differentiation of osteoblasts 35 .
The present study showed linear release of Li + in the medium. That is, the lithium in Li/CPC is directly proportional to Li + in the medium. The cell proliferation seen in Li/CPC-50 and Li/CPC-100 was better than that in Li/CPC-200 and CPC, which was in accordance with the results of cell-cytoskeleton staining. Early osteogenic differentiation (as reflected by ALP level) in Li/CPC-50 and Li/CPC-100 was better than that for Li/CPC-200 and CPC. As a marker of osteogenic differentiation, high expression of ALP suggests that a certain concentration of lithium doped to CPC favoured osteoblastic differentiation. Alizarin-red staining showed more calcium nodules, suggesting that a lithium concentration of 50-100 mM doped to CPC also favours osteoblastic mineralisation.
Studies on the Wnt/β -catenin pathway demonstrated that Li/CPC-50 and Li/CPC-100 increased the amount of phosphorylated GSK-3β significantly and decreased the amount of phosphorylated β -catenin. Then, stabilised β -catenin aggregated in the cytoplasm and translocated to the nucleus compared with Li/CPC-200 and CPC 36 . Reverse transcription-polymerase chain reaction (RT-PCR) showed that expression of osteogenesis-related genes was upregulated significantly in Li/CPC-50 and Li/CPC-100 compared with Li/CPC-200 and CPC, data that are in accordance with expression of the Wnt/β -catenin/TCF target gene product Runx2. These findings strongly suggest that the positive effect of Li + release from Li/CPC at 25.35 ± 0.12 to 50.74 ± 0.13 mg/l on the proliferation, differentiation and mineralisation of MC3T3-E1 cells is associated with activation of the Wnt/β -catenin pathway (Fig. 9).
Osteoporosis is a systemic metabolic disease characterised by reduction of bone density, which leads to osteoporotic fractures. In osteoporosis, bone-formation capacity is weakened or bone-resorption capacity is enhanced relatively 37,38 . CPC has been used for treatment of non-load-bearing bone defects by virtue of its excellent bioactivity, biocompatibility and osteoconductivity, but it does not have special advantages for osteoporosis treatment. A GSK-3β inhibitor can activate the Wnt/β -catenin pathway to improve the microenvironment of bone cells 7,9,39 . Lithium had been used widely for the treatment of mental disorders (e.g., depression, mania) but increased bone mass and decreased bone transformation has been observed in these patients 40,41 . Hence, lithium is considered a candidate drug for osteoporosis treatment.
Li/CPC improved the proliferation, differentiation and biomineralisation of MC3T3-E1 cells. This action is achieved by stimulation of specific cellular responses at the molecular level upon Li + release. In our study, female Sprague-Dawley rats underwent bilateral OVX. This osteoporotic rat model was established at 3 months post-OVX, which is in accordance with previous reports 42,43 . Micro-CT showed that Li/CPC enhanced osteogenesis significantly and elicited greater bone formation with higher BV/TV at the periphery of the implant material compared with CPC, indicating superior stimulation of osteogenesis in vivo. Early endochondral ossification and formation of new bone at 4 weeks, and penetration of regenerated osseous tissue into the initial outer surface at 8 weeks postoperatively in Li/CPC, suggested better osteoconductivity and osseointegration. We studied the effect of lithium-doped and undoped CPC on local osteogenesis. A sham-operated control group was not created and studied. Thus, we did not know the effect of bone fracture treated with or without implanted materials. The band density was quantified using ImageJ software and data from three independent experiments were presented (CPC as the control group, the values were expressed as the mean ± SD, n = 3). Statistical analyses were done using Students't-test. *p < 0.05 was considered significant.

Conclusions
The present study suggests that local application of Li/CPC to osteoporotic fractures can accelerate bone regeneration by activation of the Wnt/β -catenin pathway via lithium released from bioactive materials. Li/CPC maintained the bioactivity and biocompatibility of CPC, but also showed better osteoconductivity and osseointegration than CPC. Hence, Li/CPC may be a novel biomedical filling material for treatment of orthopaedic diseases with reduced bone volume, such as osteoporosis.

Materials and Methods
Preparation of Li/CPC. Li/CPC was prepared according to a method described previously 23 . TTCP (Sigma-Aldrich, Saint Louis, MO, USA) and DCPA (Sigma-Aldrich) were mixed at a ratio of 1:1 as a powder phase. Lithium chloride (0, 50, 100 and 200 mM; Sigma-Aldrich) was added to 20% wt of citric acid as a liquid phase (pH 4). Powder and liquid phases were mixed at a ratio of 0.3 ml/g and placed in a cylindrical mould to prepare the cement paste. A diameter of 12 mm and a height of 2 mm were used for material and cell-culture experiments, whereas 2.5 mm and 4 mm, respectively, were employed for animal studies.   was observed under SEM (SUPRA 55; Zeiss, Oberkochen, Germany). Some cements were ground into powders and used for detection of the constitution and structure of the material phase by XRD (D8 Advance; Bruker, Billerica, MA, USA) using Cu Kα (γ = 1.5406 Å) radiation in step-scan mode (2θ = 0.02° per step).
Li + release from cements. Liquid extracts were prepared according to the methods described above. Li + concentrations in cell culture media at 1, 3, 5 and 7 days were determined by inductively coupled plasma-atomic emission spectroscopy (JY2000-2; Horiba Jobin Yvon, Kyoto, Japan).
Assay to measure ALP activity. MC3T3-E1 cells (2 × 10 4 ) were added to 24-well plates. After culture in cell-differentiated solutions for 7 days, supernatants were removed gently and the plates washed carefully with PBS. Cells on plates were homogenised in 0.2% Triton X-100 (200 μ l). ALP activity was tested using an ALP Assay kit (Beyotime, Jiangsu, China). p-nitrophenol production was detected by monitoring OD using a microplate reader (Multiskan ™ GO; Thermo Scientific) at 405 nm. The obtained value of OD was compared with the value from a standard curve of a series of diluted concentrations of p-nitrophenol in lysis buffer. The result of ALP activity was normalised by total protein content, which was measured with a Bicinchoninic Acid (BCA) Protein Assay kit (Thermo Scientific). The result was expressed as micromoles of p-nitrophenol formed per minute per microgram of total protein (1 μ mol/min/mg protein).
Staining with alizarin red. MC3T3-E1 cells (1 × 10 5 ) were added to six-well plates and cultured with cell-differentiated solutions for 21 days. The capacity for cell mineralisation was measured using Alizarin Red Staining kits (Sigma-Aldrich USA), which indicated cell mineralisation by changing the colour of the substrate to red. Images were obtained under light microscopy (IX71; Olympus).
Staining of cytoskeleton proteins. MC3T3-E1 cells (2 × 10 4 ) were added to glass coverslips in 24-well plates. After 24 h, extracts were added to each well to replace the culture medium for an additional 12 h. Then, cells were fixed with 4% paraformaldehyde for 10 min. Cells were permeabilised in permeabilisation buffer (0.2% Triton X-100) for 5 min. Then, coverslips were placed on a piece of Parafilm ® in a humid chamber and 200 μ l of 100 nM rhodamine phalloidin (Cytoskeleton, Denver, CO, USA) added, followed by incubation in the dark for 30 min at room temperature. DNA was counterstained for 30 s with 200 μ l of 4′,6-diamidino-2-phenylindole (DAPI; 100 nM; Dojindo) in PBS. Coverslips were rinsed in PBS and inverted on a drop of Antifade mounting media on a glass slide. Excess media were removed gently with a tissue and sealed on each side with nail polish. Images were taken with a fluorescence microscope (BX53; Olympus) to observe the cytoskeleton protein F-β -actin. Gene expression by real-time RT-PCR. Expression of osteogenesis-related genes was evaluated using RT-PCR. MC3T3-E1 cells (5 × 10 4 ) were added to 12-well plates, and cultured by cell-differentiated solutions for 3 and 7 days. Total RNA was isolated using TRIzol ® reagent (Invitrogen) and reverse transcription for mRNAs carried out using a Transcriptor First Strand cDNA Synthesis kit (Thermo Scientific) according to manufacturer instructions. Relative mRNA expression was determined using a SYBR Green qPCR kit (Toyobo, Osaka, Japan) and employing β -actin as the reference control. Expression of target mRNA was calculated from delta-delta Ct values. β -actin was used as an internal control. Primer sequences are listed in Table 1.

Osteoporosis model in rats.
Animal experiments were undertaken in accordance with guidelines outlined by the Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences (SIAT-IRB-160304-YYS-PHB-A0220) in Shenzhen, China, and approved by the Committee on the use of Live Animals in Teaching and Research of the same institute. Bilateral OVX was conducted on 6-month-old female Sprague-Dawley rats according methods described previously 43 . The BMD, BV/TV, Tb.N, Tb.Th, Tb.Sp and Conn.D of trabecular bone at proximal tibiae was analysed by micro-CT (1176; SkyScan, Kontich, Belgium) 3 months post-OVX.

Material implantation.
A second surgical procedure was carried out 3 months after establishment of the OVX model to create bone defects at the medial aspect of the tibial shaft, below the tibial plateau, bilaterally. Routine shaving and aseptic procedures were done. An incision was made to expose the tibia, creating a bone defect (diameter, 2.5 mm; depth, 4 mm) which was filled with Li/CPC-100 or CPC. The incision was closed with sutures. Animals were killed 4 and 8 weeks postoperatively.

Micro-CT.
Was done to evaluate the ultrastructure and morphology of defects. Raw images were reconstructed in 3D and converted to binary images with adaptive local thresholding. The new BV/TV at various distances from the cement surface (6 and 12 pixels; 1 pixel ≈ 18 μ m) was calculated. H&E staining. Specimens of proximal tibia were fixed in 4% paraformaldehyde, decalcified in 10% ethylenediamine tetraacetic acid, and dehydrated in a graded series of ethanol solutions (70, 80, 90 and 100%). Specimens were embedded in paraffin and sectioned (thickness, 5 μ m). H&E staining was used to observe regenerated osseous tissue and monitor specific tissue responses to implanted materials under light microscopy (BX53; Olympus). Five sections of each specimen were produced. Giemsa staining. Specimens were fixed in 4% paraformaldehyde and dehydrated in a graded series of ethanol solutions (70, 95, and 100%). Specimens were immersed in xylene and embedded in methyl methacrylate (Merck, Kenilworth, NJ, USA). Specimens were cut into sections (thickness ≈ 300 mm) by a hard tissue microtome (310 CP Band System; Exakt, Norderstedt, Germany). Five sections of each specimen were produced. Then, a micro grinding system (310 CP; Exakt) was applied to polish the sections down to a thickness of 30-40 mm. Giemsa staining (Merck) was used to observe regenerated osseous tissue under light microscopy (BX53; Olympus).

Statistical analyses.
Data are the mean ± SD of triplicate experiments. Statistical analyses were done using Students't-test. P < 0.05 was considered significant. SPSS v21.0 (IBM, Armonk, NY, USA) was used in the study.