Epigenetic inhibition of Wnt pathway suppresses osteogenic differentiation of BMSCs during osteoporosis

1Osteoporosis is a common degenerative bone disease in aged population and postmenopausal females. It is characterized by decreased bone density and bone microarchitecture destruction. The fundamental mechanism of osteoporosis is the imbalance between bone formation and resorption. Bone marrow-derived mesenchymal stem cells (BMSCs), which are precursors of osteoblastic-lineage cells, play crucial role in bone formation. Recent studies of our group and others revealed that osteogenic differentiation capacity of BMSCs was impaired during osteoporosis–. Recovering the osteogenic capacities of BMSCs could reduce bone loss in osteoporosis,–, suggesting potential therapeutic strategy for osteoporosis. Understanding the mechanisms of BMSCs dysfunction has become a critical issue in osteoporosis research.

2Wnt signaling pathway, a pathway essential for skeletal development and homeostasis,, play a key role in regulating BMSCs differentiation. Canonical Wnt pathway could activate the expression of Runt-related transcription factor 2 (RUNX2) to promote osteogenic differentiation of BMSCs. Importantly, recent publications have reported that Wnt signaling was impaired in BMSCs of OVX mice. Rescuing Wnt genes expression could restore osteogenic differentiation capacity of BMSCs and attenuated bone loss,. However, the molecular mechanisms of Wnt signaling repression in osteoporosis is poorly understood.

3Emerging evidences revealed that the activation and repression of signaling pathways are controlled by histone acetylation modification,, a fundamental epigenetic mechanism. Histone acetylation is controlled by the balance between histone acetyltransferases (HATs) and histone deacetylases (HDACs). Recently, the role of HDACs in Wnt pathway regulation is extensively studied. For example, knockout of HDAC1, HDAC2, or HDAC3 impairs canonical Wnt signaling pathway,. HDAC1 could regulate β-catenin gene expression by binding to its promoter. HDAC6 suppresses β-catenin acetylation and degradation. On the other side, although HATs contribute to the precise regulation of gene expression,, the function of HATs in Wnt pathway remains elusive.

4In this study, we hypothesized that the dysregulation of Wnt signaling in BMSCs derived from OVX mice is related with histone acetylation modification. By screening the family of HATs, we tried to find out the HATs that affect Wnt signaling pathway and contribute to osteoporosis.

5We established an osteoporosis mice model by ovariectomy (OVX). Two months after OVX surgery, bone mass and bone formation were obviously decreased in osteoporotic femur, compared to sham-operated (sham) mice (Fig. 1a, b). The osteogenic differentiation capacity of BMSCs derived from OVX mice (OVX BMSCs) was significantly decreased, compared with Sham BMSCs (Fig. 1c, Supplementary Fig. S1). It has been demonstrated that disrupted hormones and cytokines in microenvironment lead to the dysfunction of osteoblastic-lineage cells. However, we found that the osteogenic defect of OVX BMSCs is maintained after several passages of in vitro culture (Fig. 1c), suggesting permanent changes in molecular signaling regulation.

6Previous reports showed that decreased levels of Wnt signaling, a key pathway regulating osteogenesis, led to bone formation defect of BMSCs during osteoporosis,. As reported, Wnt/β-catenin signaling was activated during osteogenesis (Supplementary Fig. S2A, S2B). Activating Wnt pathway by Wnt3a promoted osteogenic differentiation, while repressing Wnt signaling by DKK1 inhibited osteogenic differentiation of BMSCs (Supplementary Fig. S2C–S2F).

7We then compared Wnt signaling in sham and OVX BMSCs. Western blotting revealed that active beta-catenin levels in OVX BMSCs were lower than that in sham BMSCs (Fig. 2a). Real-time reverse transcription (RT)-PCR conformed that the expression of Axin2, a definite target of the Wnt/β-signaling, was decreased in OVX BMSCs, compared with Sham BMSCs (Fig. 2b). The expression of Wnt genes (Wnt1, Wnt6, Wnt10a, and Wnt10b) was also declined in OVX BMSCs (Fig. 2c).

8Histone acetylation modification, a key mechanism of epigenetic regulation, is closely related with the activation or suppression of Wnt signaling in tumor cells and hematopoietic stem cells,. As expected, CHIP analysis confirmed that the acetylation modifications of histone H3 on the promoters of Wnt genes was decreased in OVX BMSCs (Fig. 3a).

9Histone acetylation modification is controlled by deacetylases and acetyltransferases. It has been suggested that histone acetyltransferases play a role in bone development and diseases,. Therefore, we focused on HATs to elucidate the molecular mechanisms.

10Firstly, we identified the HATs controlling osteogenesis. Real-time RT-PCR was performed to screen HATs expression during osteogenic differentiation of BMSCs. Among 14 HATs, only GCN5 increased persistently after 1 h and 24 h of osteogenic induction (Fig. 3b). Western blotting confirmed the increase of GCN5 protein in BMSC during osteogenic differentiation (Fig. 3c). Accordingly, the acetylation levels of H3K9, a major target site of GCN5, were also elevated after BMSC differentiation (Fig. 3c).

11We then compared GCN5 as well as H3K9ac levels in sham and OVX BMSCs. Western blotting confirmed GCN5 and H3K9ac were decreased in OVX BMSCs (Fig. 3d). Moreover, GCN5 expression was decreased at the early stage of osteoporosis (Fig. 3e), suggesting that GCN5 is closely related with the dysfunction of OVX BMSCs.

12GCN5 has been found to play crucial roles in development and differentiation-. However, whether GCN5 is related with osteoporosis is not fully understood. First, we performed gain- and loss-of-function assays to confirm the role of GCN5 in osteogenic differentiation of BMSCs. Fourteen days after osteogenic induction, Alizarin red staining and real-time RT-PCR showed that overexpression of Gcn5 promoted osteogenic differentiation of BMSCs (Fig. 4a, b), while knockdown of Gcn5 inhibited osteogenic differentiation of BMSCs (Fig. 4c, d). To verify the in vitro findings, BMSCs overexpressed or knocked-down of Gcn5 were transplanted subcutaneously into nude mice. After 8 weeks, histological analysis revealed that overexpression of Gcn5 promoted bone formation of BMSCs (Fig. 4e, f), while knockdown of Gcn5 inhibited osteogenic differentiation of BMSCs in vivo (Fig. 4g, h).

13To explore whether declined GCN5 leads to differentiation defect of BMSCs, we restored GCN5 expression in OVX BMSCs. As expected, transfection of GCN5-lentiviral expression vector largely recovered the osteogenic capacity of OVX BMSCs (Fig. 4i, j).

14We next explored whether GCN5 regulates osteogenesis capacity of BMSCs through Wnt pathway. In vitro osteogenesis assay showed that DKK1, an inhibitor of Wnt signaling, abolished the positive effect of GCN5 on osteogenic capacities of BMSCs (Fig. 5a) and activation of Wnt pathway (Fig. 5b). Overexpression of Gcn5 obviously enhanced the expression of these Wnt genes (Fig. 5c). Gain- and loss-of-function assay confirmed that knockdown of GCN5 inhibited the activation of β-catenin, while overexpression of GCN5 promoted β-catenin activation (Fig. 5d).

15To identify whether GCN5 regulates Wnt signaling directly through epigenetic modulation, we assessed the histone acetylation modification. Western blot assay showed that H3K9ac levels in BMSCs were increased by GCN5 overexpression, whereas were repressed by knockdown of GCN5 (Fig. 5e). Accordingly, CHIP assay confirmed that overexpression of GCN5 increased the recruitments of GCN5 to the promoter regions of Wnt1, Wnt6, Wnt10a, and Wnt10b (Fig. 5f), and enhanced acetylation levels of H3K9ac on those promoters (Fig. 5g). These evidences indicated that GCN5 promoted Wnt genes transcriptions by increasing H3K9ac levels on their promoter regions.

16To further verify that recovering GCN5 expression could restore osteogenesis in vivo, we injected either empty vector or Gcn5- lentiviral expression vector into mouse femur bone marrow one week after OVX surgery. Six weeks after injection, micro-computed tomography (CT) scan showed that Gcn5-lentiviral vector administration largely prevented the bone mineral density (BMD) decline and trabecular bone loss in femurs after OVX (Fig. 6a). To confirm whether GCN5 prevented bone loss by enhancing bone formation, we performed calcein double labeling assay and found that injection of Gcn5-lentiviral vector attenuated the decline of bone formation in trabecular bone, but not cortical bone (Fig. 6b). The concentration of OCN, a biomarker of bone formation, was also enhanced in bone marrow after Gcn5 overexpression (Fig. 6c). Taken together, these results indicated that recovering GCN5 expression rescued bone formation in OVX mice.

17To investigate whether Gcn5-lentiviral vector injection recovered osteogenesis capacity of BMSCs, we isolated BMSCs from OVX mice 6 weeks after the lentivirus administration.

18Real-time RT-PCR and western blot assay revealed that Gcn5-lentivirus injection enhanced both mRNA and protein levels of GCN5 in OVX BMSCs (Fig. 7a, b). Consistently, the protein level of H3K9ac in OVX BMSCs was also increased by Gcn5-lentivirus injection (Fig. 7b). Osteogenesis assay showed that overexpression of GCN5 largely recovered the osteogenic differentiation of OVX BMSCs (Fig. 7c, d). Consistent with the findings that GCN5 functions through Wnt pathway, the mRNA levels of Axin2 and Wnt genes were also enhanced in BMSCs after Gcn5-lentiviral vector injection (Fig. 7e, f). Consistently, H3ac levels in promoter regions of Wnt genes were also increased after overexpression of GCN5 (Fig. 7g). Taken together, the results demonstrated that local administration of Gcn5-lentiviral vector activated Wnt signaling to promote osteogenic differentiation of endogenous BMSCs.

19Dysfunction of BMSCs lead to bone formation defect in osteoporosis,,,,,. But its mechanism remains elusive. In this study, we found that an epigenetic mechanism that represses Wnt signaling to inhibit osteogenic differentiation of BMSCs during postmenopausal osteoporosis. The decrease of histone acetylation modification lead to persistent repression of Wnt signaling. By restoring histone acetylation levels in OVX BMSCs, the osteogenic capacities of BMSCs was recovered in osteoporotic mice. Our study revealed a novel mechanism of BMSCs dysfunction during osteoporosis and introduces a novel insight into the prevention of osteoporosis.

20Our findings indicated that Wnt signaling is precisely modulated by histone acetylation. In this study, we found that acetylation on histone declined in OVX MSCs. Histone acetylation lost the condensed chromatin and enhances gene transcription. Reduced histone acetylation limited the transcription of genes regulating BMSC functions. Consistently, the Wnt/β-catenin signaling, one of the crucial pathway for osteogenesis, was also repressed in OVX BMSCs. Restoring histone acetylation levels by GCN5 recovered expression of Wnt genes in OVX BMSCs, suggesting the crucial role of histone acetylation in controlling Wnt pathway. Besides acetylation, histone could be modified by various modifications. Further studies focused on histone epigenetic modifications would expand our understanding of the molecular mechanism of osteoporosis.

21Previous studies majorly focused on the function of HDACs in Wnt signaling,,. However, the role of HATs in Wnt signaling regulation remains elusive. By analyzing the family of HATs, we found that GCN5 played important roles in Wnt signaling regulation and BMSCs osteogenesis. Moreover, GCN5 controlled the expression of Wnt ligands through histone acetylation modification. Since Wnt signaling is one of the most important pathway involved in development, regeneration and diseases, it is interesting to investigate whether GCN5 play other physiological or pathological roles by regulating Wnt signaling.

22By injecting lentivirus-expressing GCN5 into femur of OVX mice, we restored bone volume and density in osteoporotic mice. After isolation of BMSCs from bone marrow, we confirmed that injection of GCN5 lentivirus recovered osteogenic capacity of OVX BMSCs, indicating that GCN5 recovered bone mass by regulating BMSCs. Since the lentivirus is not specific to BMSCs, it is possible that cells other than BMSCs, such as osteoclasts and immune cells are also involved in the recovery of osteoporosis. Based on previous reports, GCN5 upregulates T-cell activation and is required for iNKT cell development. Therefore, conditioned knockout of GCN5 in BMSCs might be needed to explore GCN5 function in vivo.

23Emerging studies revealed that GCN5, a highly conservative histone HAT, is involved in various diseases. GCN5 is widely studied in tumors and has been tested as therapeutic targets for cancer. We recently found that GCN5 controlled the pro-angiogenic factors secretion of BMSC through increasing VEGF expression. Taken together with this study, we revealed that GCN5 is a crucial factor regulating osteogenesis during bone remodeling. Importantly, recovering GCN5 expression rescued the dysfunction of BMSCs and attenuated bone loss during osteoporosis, indicating that GCN5 is a potential target for osteoporosis treatment.

24Eighty 8-week female C57BL/6 mice purchased from the Laboratory Animal Research Center of the Fourth Military Medical University were used for cell culture and osteoporosis model establishment. Twenty-two 6-week female immunocompromised nude mice (BALB/c) for subcutaneous transplantation and twenty-two 8-week female C3H mice used for femur injection of lentivirus were purchased from Vital River Laboratory Animal Technology Co. Ltd. (Beijing, China). All protocols have been approved by the Animal Care Committee of the Fourth Military Medical University and have met the NIH guidelines for care and use of laboratory animals in this study.

25Murine bone marrow mesenchymal stem cells (BMSCs) were flushed from long bones by using α-MEM medium supplemented with 20% FBS. After centrifuge, BMSCs were seeded into 10-cm dishes with α-MEM (Gibco BRL, Gaithersburg, MD, USA) supplemented with 20% FBS (Sijiqing, Hangzhou, China), 2 mM l-glutamine (Invitrogen, Carlsbad, CA, USA), 100 U/ml penicillin, and 100 mg/ml streptomycin (Invitrogen, Carlsbad, CA, USA). Cells were incubated in a humidified atmosphere of 5% CO2 at 37 °C. The culture medium was first changed after 24 h of seeding and followed by change every other day. After confluence, cells were passaged by using 0.25% trypsin. BMSCs at the first passage were used in all experiments.

26In all, 5 × 105 cells were seeded in each well of six-well plates. Then, cells were incubated with osteogenic-inducing medium supplemented with 100 μg/ml ascorbic acid, 2 mmol/l β-glycerophosphate and 10 nmol/l dexamethasone for 14 days. The media was changed every other day.

27The Alizarin red staining was conducted as previously reported. In brief, BMSCs were fixed with 60% isopropanol for 1 min at room temperature after washing with Phosphate-buffered saline(PBS) twice. Then, cells were washed twice by using ddH2O, followed by staining with 1% alizarin red (Sigma-Aldrich, St. Louis, MO, USA) for 10 min at room temperature. The quantification was performed by using 5% cetylpyridinium chloride to elute the staining. The absorbance was determined by a spectrophotometer at the wavelength of 562 nm. In order to exclude the influence of different cell numbers on the quantification, results were normalized to the concentration of protein.

28The CFU-OB assays were performed as previously described. Briefly, after whole bone marrow cells were flushed from long bones, one million cells were seeded in 6-cm dishes. After 48 h, the cells were incubated in osteogenic-inducing medium for 14 days. The cells were washed by PBS twice, fixed in 60% isopropanol for 1 min, washed twice with ddH2O, and stained with 1% Alizarin red (Sigma-Aldrich, St. Louis, MO, USA) for 10 min at room temperature. Colonies with > 50 cells were counted.

29Wild-type mouse GCN5 was amplified by PCR. NotI and KpnI were used to digest the amplicons before inserted into the Pent3c lentiviral vector. The sequences of the primers were: forward: CGGGGTACCATGGCGGAACCTTCCCAGGC; reverse: AAGGAAAAAAGCGGCCGCGGGACCAGGCTCAGGATGCTCA. The primers for GCN5 shRNA were: forward: CCGGGCTACCTACAAAGTCAATTATCTCGAGATAATTGACTTTGTAGGTAGCTTTTTG; reverse: AATTCAAAAAGCTACCTACAAAGTCAATTATCTCGAGATAATTGACTTTGTAGGTAG.

30Restriction enzymes AgeI and EcoRI were used to digest the PCR product before incorporated into the PLKO.1 vector (Invitrogen). Sanger sequencing was performed to verify the inserted fragments. To produce the lentivirus, 293T cell line was co-transfected with two packaging vectors (psPAX2 and pMD2.G) and two plasmid vectors. To get rid of the cell debris, supernatant was centrifuged at 1000 rpm for 10 min, followed by filtered through a 0.45-μm polyethersulfone low protein-binding filters. The titer of the lentivirus was 105 IFU/ml, assessed by a quantify kit (Clontech, CA). In all, 8 mg/ml polybrene (Sigma, St. Louis, MO) were used when virus was transfected into BMSCs. A vector inserted with scrambled GCN5 was used as a negative control.

31BMSCs were transfected with lentiviral vector expressing either Gcn5 or shRNA, as well as their negative control vectors. Three days after transfection, 2 × 106 BMSCs were mixed with 15 mg of hydroxyapatite/tricalcium phosphate (HA/TCP) ceramic particles (Sigma-Aldrich) and implanted into the dorsal subcutaneous space of NOD/SCID mice (Fourth Military Medical University Experimental Animal Center). Eight weeks after transplantation, the implants were harvested, fixed with 4% paraformaldehyde (PFA) and decalcified with buffered 10% EDTA (pH 7.4). Histological sections were stained with H&E or Masson’s Trichrome (BaSO Diagnostic Inc, Guangdong, China).

32The lentivirus-expressing GCN5 and the control vector were concentrated into 5–10*105 transducing units/ml by using the Lenti-X™ Concentrator (Cat. Nos. 631231 & 631232) before the intra-femoral injection. The injection of lentivirus was conducted 1 week after the OVX surgery was performed. After anesthetization, a small incision was made over the kneecap to allow the 22-gauge needle to pierce into the femur. The mice were either injected with 20 ul lentivirus-expressing GCN5 or empty vector. After the injection, the incision was sutured carefully to recover leg movement of the mice. Analgesic drug was added into water after the surgery and feed the mice for 1 week.

33The mice were intraperitoneally injected with 15 mg/kg of calcein (Sigma-Aldrich) 15 and 3 days prior to sacrifice. The femora were then fixed in 4% PFA for 24 h and embedded in methylmethacrylate without being decalcified. The samples were sectioned using a Sliding Microtome, and the slices were observed and photographed under a fluorescence microscope. The dynamic bone formation rate was measured by mineral apposition rate (MAR) and bone formation rate per unit of bone surface (BFR/BS). The calculation was performed in areas randomly selected from five fields in distal metaphysis of femur and periosteum.

34The femora were fixed in 4% phosphate-buffered PFA after the mice were sacrificed. The Locus SP Pre-Clinical Specimen micro-CT system (GE Healthcare, USA) was used to scan the samples with 8-mm resolution, 50 kV tube voltage and a 0.1 mA tube current. The region of interest (ROI) was selected for the area from 0.5 mm to 1.5 mm below the growth plate in the distal ends of femur. The ROI of cortical bone was conducted in the region of the middle of the femur.

35Total RNA was extracted by TRIzol reagent (Invitrogen Life Technology, Carlsbad, CA). One microgram of total RNA was used for reverse transcription. The process was performed by using a PrimeScript RT reagent kit (Takara Bio, Shiga, Japan). Real-time RT-PCR was conducted by using a Quantitative SYBR Green Kit (Toyobo, Osaka, Japan) and was detected by ABI Prism 7500 Sequence Detection System (Applied Biosystems, Foster City, CA). The Supplementary Table S1 shows the primer sequences. The PCR program was set as: 94 °C, 3 min; 94 °C, 15 s; 60 °C, 30 s, for 40 cycles. β-actin was used as control. The results were calculated by ΔΔ−ct method.

36ChIP assay was performed following manufacturer’s instructions of a ChIP assay kit (Millipore, Catalog 17-371, Billerica, MA). The antibody against GCN5 (Santa Cruz Biotechnology, sc-365321 × ), H3K9ac (Abcam, ab10812) and H3ac (Millipore, 06-599) were used. The mouse IgG and anti-RNA Polymerase II (Millipore) were used as negative control and the positive control, respectively. The precipitated DNA samples were assessed by real-time PCR. Primers were designed from the region of 500-bp distance to the transcription start sites. The primers of Wnt1 were: forward GGTTAGCCTGTCAGCTCTTTGC; reverse TCAGGGCCACAGACAATGAG; The primers for Wnt6 were: forward CTTCCTTCCCCCAAAGAAATG; reverse GTCCAACAGCTCTTCCCTACCTATC; The primers for Wnt10a were: forward TCCTTCCTTCACCCTCTGCAT; reverse TAGTGTCTAAGGGTTCTACCCCAAGT. The primers for Wnt10b were: forward TCCAAAGGAAAGGTTTTAGCCATA; reverse CCCACCCTTCCTGCTGAA. The values of results were normalized to the input value.

37RIPA buffer (Beyotime) with protease inhibitor were used to extract protein from BMSCs. BCA quantitative assay (Beyotime) was used to measure the protein concentration. Twenty microgram protein was separated by 4 – 12% Tris-glycine sodium dodecyl sulfate-polyacrylamide gel (Invitrogen Life Technology), followed by transferred onto a polyvinylidene fluoride(PVDF) membrane (Millipore). Blocking was performed by using 5% BSA for 2 h at room temperature. The membrane was incubated with primary antibodies overnight at 4 °C.

38Secondary antibodies (Cowin Biotech Co, Beijing, China) were incubated for 2 h at room temperature. The protein accumulation was detected by Western-light Chemiluminescent Detection System (Peiqing, Shanghai, China). Antibodies against GCN5 (Santa Cruz Biotechnology, sc-365321 × ), H3K9ac (Abcam, ab10812), active β-catenin (Millipore, 05-665), GAPDH (Cowin Biotech, CW0100), and β-actin (Cowin Biotech, CW0096A) were used.

39Data were showed as means ± standard deviation (SD). Statistical analyses were used by using the Prism software (GraphPad, La Jolla, CA). Statistical differences were assessed by t-test or one-way analysis of variance. P < 0.05 was considered as significant.