Advanced glycation end products inhibit the osteogenic differentiation potential of adipose‐derived stem cells by modulating Wnt/β‐catenin signalling pathway via DNA methylation

Abstract Objectives Advanced glycation end products (AGEs) are considered a cause of diabetic osteoporosis. Although adipose‐derived stem cells (ASCs) are widely used in the research of bone regeneration, the mechanisms of the osteogenic differentiation of ASCs from diabetic osteoporosis model remain unclear. This work aimed to investigate the influence and the molecular mechanisms of AGEs on the osteogenic potential of ASCs. Materials and methods Enzyme‐linked immunosorbent assay was used to measure the change of AGEs in diabetic osteoporotic and control C57BL/6 mice. ASCs were obtained from the inguinal fat of C57BL/6 mice. AGEs, 5‐aza2′‐deoxycytidine (5‐aza‐dC) and DKK‐1 were used to treat ASCs. Real‐time cell analysis and cell counting kit‐8 were used to monitor the proliferation of ASCs within and without AGEs. Real‐time PCR, Western blot and Immunofluorescence were used to analyse the genes and proteins expression of osteogenic factors, DNA methylation factors and Wnt/β‐catenin signalling pathway among the different groups. Results The AGEs and DNA methylation were increased in the adipose and bone tissue of the diabetic osteoporosis group. Untreated ASCs had higher cell proliferation activity than AGEs‐treatment group. The expression levels of osteogenic genes, Opn and Runx2, were lower, and mineralized nodules were less in AGEs‐treatment group. Meanwhile, DNA methylation was increased, and the Wnt signalling pathway markers, including β‐Catenin, Lef1 and P‐GSK‐3β, were inhibited. After treatment with 5‐aza‐dC, the osteogenic differentiation capacity of ASCs in the AGEs environment was restored and the Wnt signalling pathway was activated during this process. Conclusions Advanced glycation end products inhibit the osteogenic differentiation ability of ASCs by activating DNA methylation and inhibiting Wnt/β‐catenin pathway in vitro. Therefore, DNA methylation may be promising targets for the bone regeneration of ASCs with diabetic osteoporosis.


| INTRODUC TI ON
Diabetes and osteoporosis are common diseases in aged populations, and numerous studies have demonstrated a close relationship between diabetes and osteoporosis. [1][2][3] Diabetic osteoporosis is one of the most important complications of diabetes mellitus in the skeletal system and is a systemic metabolic bone disease accompanied by osteopenia and destruction of bone microstructure leading to a high risk of fractures. The incidence of diabetic osteoporosis occurs in about 50% of diabetic patients, and the risk of osteoporotic fractures is significantly increased in patients with diabetes compared with healthy subjects. 1 Recent studies showed that the increased incidence of fractures in diabetic patients is associated with the long pathogenic effects of diabetes. Now, many researchers focus on the issues of reducing bone damage and promoting bone regeneration in diabetic patients.
Promising bioengineering technologies, such as tissue engineering, which can repair bone defects with stem cells, may provide novel tools for reconstructive surgery. Recent studies reported that adipose-derived stem cells (ASCs) have been widely used in bone tissue engineering and regenerative medicine research. [4][5][6][7][8][9] ASCs are ideal stem cells because they can be harvested easily from adipose tissue. Moreover, they have the capacity to differentiate into osteoblastic-like cells, extensive proliferative ability and low immunogenicity. [10][11][12] Diabetes mellitus can seriously affect the metabolism of bone tissue. Roy, B's research showed that diabetes leads to the dysfunction of osteoblasts and osteoclasts, and have negative effects on the differentiation of stem cells into bone cells. 13 The osteogenic potential of ASCs might be destroyed in the high glucose and inflammatory environment caused by diabetes.
Advanced glycation end products (AGEs) are the end products of glucose, fructose, glucose 6-phosphate and proteins via a non-enzymatic and non-reversible glycosylation reaction. The formation and deposition of AGEs accelerate during ageing, inflammation and especially diabetes. 1, [14][15][16][17][18] Schwartz et al and Ding et al 16,17 found that AGEs significantly inhibit the proliferation of osteoblasts but increased the activity of osteoclast, which eventually leads to reduced bone strength, osteoporosis and fractures. Therefore, AGEs are associated with the development and progression of diabetic osteoporosis. However, there is no clarity concerning the mechanisms whereby AGEs lead to diabetic osteoporosis. 1,[14][15][16][17][18] DNA methylation refers to the methylation of CpG dinucleotides on cytosine dinucleotide 5-carbon atoms under the action of DNA methyltransferase using S-methionine as the donor of the methyl group. Recent studies show that hypermethylation of a DNA sequence is associated with the suppression of gene expression and demethylation of DNA is contrary to the effect of DNA methylation, which leads to an increase in gene expression. [19][20][21][22] DNA methylation modification affects the normal differentiation function of cells and has a close relationship with the occurrence, development and treatment strategies of diseases. Studies showed that DNA methylation modification during stem cell differentiation is apparent in a variety of bone diseases, including osteoporosis and osteoarthritis. [23][24][25] In the process of osteogenic differentiation of stem cells, DNA methylation is a mechanism involved in regulating the expression of osteogenic marker genes, and the expression of stem cell differentiation-related genes is inhibited because of the high promoter methylation levels. However, there are only a few reports describing how AGEs regulate DNA methylation in ASCs.
Based on literatures and our previous work, we put forward that AGEs might have negative impact on the physiological functions of ASCs. The aims of this study are to explore the effects of AGEs on the proliferation and function of ASCs and to assess whether DNA methylation plays a key role in these processes during diabetic osteoporosis.

| Materials
Advanced glycation end product (AGE)-BSA was purchased from BIOVISION (San Francisco, USA). Osteogenic differentiation medium used for C57BL/6 mouse adipose-derived stem cells (ASCs) was obtained from CYAGEN (Shanghai, China). The 5-bromo-4chloro-3-indolyl phosphate (BCIP)/nitro blue tetrazolium (NBT) alkaline phosphatase (ALP) colour development kit was purchased from BEYOTIME (Jiangsu, China). Primary antibodies were purchased from Abcam (Cambridge, UK), and secondary antibodies were provided by Thermo Fisher Scientific (MA, USA). in vitro. Therefore, DNA methylation may be promising targets for the bone regeneration of ASCs with diabetic osteoporosis. in general), and the negative control group was injected with the same volume of sodium citrate buffer (1 mL/Kg). The fasting blood glucose of mice was monitored every 3 days after injection. When the fasting blood glucose of the test group remained higher than 11.1 mmol/L, the mice were considered diabetic.

| ELISA
Furthermore, we randomly selected three diabetic mice and negative control mice for ELISA. The expression levels of AGEs in adipose tissue and bone tissue of the two groups were detected by sandwich enzymelinked immunosorbent assay (ELISA). The samples were added to the enzyme-labelled well, which were pre-labelled with anti-AGE monoclonal antibodies, incubated, washed and labelled with biotinylated anti-AGE antibody. Streptomycin-HRP was added to washed wells and developed in the dark. The reaction was stopped with acid, and the optical density was measured at 450 nm using a spectrophotometer.

| Micro-CT analysis
Femurs were dissected from diabetic and control mice at week 24.
Samples were fixed with 4% paraformaldehyde and analysed using a SCANCO Medical CT-40 (SCANCO Medical, Switzerland). The whole femur was scanned completely. The main observation area was about 1 cm from the distal femur to the epiphysis.

| Immunohistochemistry
The femur and tibia were dissected from 18-week-old mice.
Immunohistochemical analyses were carried out according to the manufacturer's recommended protocol. The samples were incu-

| Isolation and culture of ASCs, and the identification and characterization procedure of isolated ASCs
Adipose-derived stem cells were collected from the subcutaneous fat of the inguinal sites. The adipose tissue was finely cut into smaller  Sections of the cell balls were observed and imaged through an optical microscope.
The second-passage ASCs (5 × 10 4 cells per 1 mL) were seeded into 96-and 16-well plates. ASCs were cultured for 24 hours in 5% CO 2 at 37°C; then, ASCs were treated with different concentrations of AGEs (20, 40, 80 and 160 µg/mL). Cell proliferation was measured by CCK-8 for 24, 48 and 96 hours and monitored by RTCA for 24 hours. Then, the optical densities of CCK-8 were measured at 450 nm by a microplate reader (Spectra Thermo, Switzerland), and data were analysed using the provided RTCA software.

| Osteogenic induction
When first-passage ASCs reached a fusion of 90%, second-passage ASCs were seeded into 6-well plates at a density of 2 × 10 4 cells/cm.

| Alkaline phosphatase and Alizarin red staining
Cells were treated as previously described. After osteogenic induction for 7 days, we used a BCIP/NBT alkaline phosphatase colour

| Immunofluorescence and confocal laser scanning microscopy
Cells were treated as previously described. After osteogenic induction for 7 days, the medium was discarded and cells were rinsed three times with PBS. Cells were fixed with 4% paraformaldehyde solution for 20 minutes, permeabilized with 0.5% TritonX-100 for 20 minutes and blocked with 5% sheep serum at 37°C for 1 hour. Cells were washed with PBS three times after each step. Then, Runx2 (1:100; Abcam, Cambridge, UK), 5-Methylcytosine (5-mC; 1:100; Active Motif, Carlsbad, North America) rabbit monoclonal antibodies were incubated with the cell samples overnight at 4°C, and an appropriate fluorescence-conjugated secondary labelled anti-rabbit antibody (1:500; BEYOTIME, Shanghai, China) was subsequently added to bind with the primary antibody. DAPI was used to stain the nucleus, and phalloidine was applied to stain the cytoskeleton. Finally, images were captured using a confocal laser microscope (TCS SP8; Leica, WETZLAR, Germany).

| Extraction of RNA and semi-quantitative polymerase chain reaction
Expression levels of the following genes were detected by semiquantitative PCR after ASCs were treated as described before osteopontin (OPN), runt-related transcription factor 2 (Runx2), alkaline phosphatase (ALP), glycogen syntheses kinase (GSK), lymphatic enhancement factor-1 (Lef-1) and β-catenin. The primer sequences used to detect the relevant genes are displayed in (Table 1) Image-Pro plus 6.0 (Media Cybernetics, Rockville, MD, USA) was used to detect the optical density of each band.

| Western blot assay
The treated ASCs were rinsed with cold PBS three times, and total proteins were harvested with a cell protein extraction reagent (KEYGEN Biotech, Nanjing, China). The collected proteins were then mixed with loading buffer at a ratio of 4:1 (V/V) and boiled for 5 minutes.
The proteins were separated on a 10% SDS-PAGE, and proteins were subsequently transferred onto a polyvinylidene fluoride membrane.
After blocking with 5% skim milk for 1 hour at 37°C, the membranes were incubated with primary rabbit monoclonal antibodies specific to the following targets: OPN, Runx2, Lef-1, GSK, cyclin D and β-catenin (Abcam, Cambridge, UK). Then, each band was incubated with a secondary labelled anti-rabbit antibody (BEYOTIME, Shanghai, China), and the results were visualized using an ECL chemiluminescence detection system (Bio-Rad, Hercules, CA, USA).

| Statistical analysis
All experiments were performed in triplicate and reproduced at least three times. Student's t test or one-way ANOVA followed by Duncan's multiple range tests was used for statistical analysis.
Statistical analysis was completed using SPSS 18.0.

| Diabetic osteoporosis mice
We established a model of diabetic osteoporosis in C57BL/6 mice.
There were significant differences between the diabetic osteoporosis group and healthy controls. In the diabetic osteoporosis group, blood glucose levels were maintained above 20 mmol/mL, but the body weight of the mice was slightly lower than the control group ( Figure 1A

| Cell proliferation
CCK-8 results showed that AGEs inhibit the proliferation of ASCs.
The ASCs were still highly proliferative (80% activity) following treatment with low concentrations of AGEs (20 or 40 μg/mL) for 24, 48 and 96 hours while treatment with high concentrations of AGEs (80 or 160 μg/mL) for 24, 48 and 96 hour significantly inhibited cell proliferation, particularly at 160 μg/mL AGEs for 96 hours (50% activity) (Figure 2A). The control BSA at various concentrations did not have a significant effect on cell viability ( Figure 2C). RTCA ( Figure 2B) was consistent with the results of the CCK-8 assays, demonstrating that AGEs inhibited the proliferation of ASCs in a dose-dependent manner.

| AGEs inhibit ASCs osteogenic differentiation potential
Alkaline phosphatase activity, an early osteogenic differentiation marker, was measured using a BCIP/NBT alkaline phosphatase colour development kit at day 7 during osteogenic differentiation. was 0.82-fold and 0.68-fold less than the control group after 4 days, respectively. After 7 days, OPN was 0.61-fold and 0.14fold less and RNUX2 was 0.63-fold and 0.47-fold less than controls ( Figure 3J,L). Figure 3 demonstrates that AGEs inhibit ASCs osteogenic differentiation potential in a dose-dependent manner.

| AGEs increase DNA methylation in ASCs
The main product of DNA methylation in ASCs, 5-mC, was detected using immunofluorescence after treatment with 20 and 40 μg/mL AGEs for 4 days. We found that with increasing AGEs concentration, the expression of 5-mC increased ( Figure 4A). The statistical analysis of immunofluorescence images confirmed that the expression of 5-mC in ASCs treated with 40 μg/mL AGEs was expressed as much as 1.12-fold compared with the control group ( Figure 4B).
The mRNA expression of Dnmt1, Dnmt3a and Dnmt3b was analysed using semi-quantitative PCR and quantitative RT-PCR. As shown in Figure 3C, treatment with 20 or 40 μg/mL AGEs resulted in a significant increase in expression of DNMTs ( Figure 4C). The statistical analysis further confirmed that Dnmt1 increased by 1.84-fold, Dnmt3a increased by 2.18-fold and Dnmt3b increased by 1.05-fold ( Figure 4D). Furthermore, the results of the quantitative RT-PCR analysis ( Figure 4E) were consistent with the semi-quantitative PCR analysis, demonstrating that AGEs increase the DNA methylation level in ASCs, particularly Dnmt1 (increased 2.12-fold) and Dnmt3a (increased 2.58-fold). Taken together, these results demonstrated that AGEs increased the DNA methylation level in ASCs.

| DNMT inhibitor 5-aza-dC rescues the osteogenic differentiation capacity of ASCs treated with AGEs
Subsequently, we explored the association between DNA methylation and the osteogenic differentiation potential of ASCs. The DNA methylation inhibitor 5-aza-dC inhibited Dnmt1 during cell replication and prevented the methylation of new chains, leading to demethylation. 19 Alizarin red staining showed that treatment of ASCs with 40 μg/mL AGEs resulted in fewer mineralized nodules than in the control group ( Figure 5A), consistent with previous results. In contrast, ASCs incubated with 1 μm of the DNMT inhibitor, 5-aza-dC, produced more mineralized nodules than the control group ( Figure 5A).
We also found that there were more mineralized nodules after treatment with 40 μg/mL AGEs and 1 µm 5-aza-dC than in the AGEs group

| The canonical Wnt signalling pathway is involved in 5-aza-dC-induced osteogenic differentiation in ASCs treated with AGEs
The canonical Wnt signalling pathway plays a crucial role in osteogenesis. Therefore, we investigated the effect on Wnt/β-catenin signalling on the osteogenic differentiation of ASCs. We detected the mRNA and protein expression levels of Wnt signalling molecules to investigate the role of AGEs and 5-aza-dC in osteogenesis in ASCs. After ASCs were incubated with osteogenic medium for 4 days, they were treated with AGEs and 5-aza-dC, the gene expression of β-catenin, Lef1 and Fzd6 was determined using quantitative RT-PCR analysis. As shown in Figure 6, treatment of ASCs with AGEs led to the largest decrease in the mRNA expression levels of the canonical Wnt signalling molecules compared with the other groups; β-catenin decreased by 0.55-fold, Lef1 decreased by 0.45-fold, and Fzd6 decreased by 0.86-fold. (Figure 6D-F). While ASCs exposed to It has been widely reported that DNA methylation levels are associated with bone diseases in the diabetic microenvironment. [18][19][20][21][22][23][24] In a recent study, Zhang et al 19

CO N FLI C T O F I NTE R E S T S
The authors declare that there is no conflict of interests regarding the publication of this paper.

AUTH O R CO NTR I B UTI O N
All authors contributed to research concept. Yong Li and Lang Wang

DATA AVA I L A B I L I T Y S TAT E M E N T
The data which support research results are available from the corresponding author upon reasonable request.