LncRNA‐AK137033 inhibits the osteogenic potential of adipose‐derived stem cells in diabetic osteoporosis by regulating Wnt signaling pathway via DNA methylation

Abstract Objectives Bone tissue engineering based on adipose‐derived stem cells (ASCs) is expected to become a new treatment for diabetic osteoporosis (DOP) patients with bone defects. However, compared with control ASCs (CON‐ASCs), osteogenic potential of DOP‐ASCs is decreased, which increased the difficulty of bone reconstruction in DOP patients. Moreover, the cause of the poor osteogenesis of ASCs in a hyperglycemic microenvironment has not been elucidated. Therefore, this study explored the molecular mechanism of the decline in the osteogenic potential of DOP‐ASCs from the perspective of epigenetics to provide a possible therapeutic target for bone repair in DOP patients with bone defects. Materials and methods An animal model of DOP was established in mice. CON‐ASCs and DOP‐ASCs were isolated from CON and DOP mice, respectively. AK137033 small interfering RNA (SiRNA) and an AK137033 overexpression plasmid were used to regulate the expression of AK137033 in CON‐ASCs and DOP‐ASCs in vitro. Lentiviruses that carried shRNA‐AK137033 or AK137033 cDNA were used to knockdown or overexpress AK137033, respectively, in CON‐ASCs and DOP‐ASCs in vivo. Hematoxylin and eosin (H&E), Masson's, alizarin red, and alkaline phosphatase (ALP) staining, micro‐computed tomography (Micro‐CT), flow cytometry, qPCR, western blotting, immunofluorescence, and bisulfite‐specific PCR (BSP) were used to analyze the functional changes of ASCs. Results The DOP mouse model was established successfully. Compared with CON‐ASCs, AK137033 expression, the DNA methylation level of the sFrp2 promoter region, Wnt signaling pathway markers, and the osteogenic differentiation potential were decreased in DOP‐ASCs. In vitro experiments showed that AK137033 silencing inhibited the Wnt signaling pathway and osteogenic ability of CON‐ASCs by reducing the DNA methylation level in the sFrp2 promoter region. Additionally, overexpression of AK137033 in DOP‐ASCs rescued these changes caused by DOP. Moreover, the same results were obtained in vivo. Conclusions LncRNA‐AK137033 inhibits the osteogenic potential of DOP‐ASCs by regulating the Wnt signaling pathway via modulating the DNA methylation level in the sFrp2 promoter region. This study provides an important reference to find new targets for the treatment of bone defects in DOP patients.


| INTRODUC TI ON
Diabetes mellitus (DM) is a systemic metabolic disease characterized by hyperglycemia. This systemic glucose metabolism disorder has a serious negative effect on the skeletal system by causing severe complications of the bone and joint system, ie, diabetic osteoporosis (DOP). 1,2 In addition to the hyperglycemic microenvironment, DOP patients are also characterized by bone microstructure damage, bone strength reduction, fracture susceptibility, and bone defects that are not easily healed. 3 For bone defects in DOP patients, the current treatment methods are not ideal. 4,5 With the rapid development of tissue engineering, bone tissue engineering, which includes scaffold materials, seed cells, and growth factors, is considered to be the most promising method for bone defect repair. 6 Adipose-derived stem cells (ASCs) are some of the most widely used seed cells in bone tissue engineering. However, our previous studies have shown that, compared with control adipose-derived stem cells (CON-ASCs), diabetic osteoporosis adipose-derived stem cells (DOP-ASCs) have less osteogenic potential, 7,8 which limits their application to the treatment of fractures and bone defects in DOP patients. Therefore, the molecular mechanism of the osteogenic decline of DOP-ASCs requires further exploration to find potential therapeutic targets for the treatment of bone defects in DOP patients.
Wnt signaling pathways are a group of multifunctional signal transduction pathways activated by the binding of the Wnt ligand to the cell membrane receptor. They participate in various physiological and pathological processes of cells, which include various bone tissue diseases such as osteoporosis and stem cell-related bone regeneration. 9 Activation of the Wnt signaling pathway in stem cells improves their ability of bone differentiation, whereas the inhibition of the Wnt signaling pathway reduces bone formation. 10,11 Our previous studies have shown that the decrease in the osteogenic potential of DOP-ASCs compared with CON-ASCs is related to the Wnt signaling pathway. 8 However, the molecular mechanism that underlies the regulation of the Wnt pathway for the osteogenic potential of DOP-ASCs is unclear.
Mammalian DNA methylation refers to methylation of the fifth carbon atom on cytosine in the CpG dinucleotide of DNA, which is catalyzed by four DNA methyltransferases, namely DNMT1, DNMT2, DNMT3a, and DNMT3b. It is generally believed that hypermethylation of DNA is related to the inhibition of gene expression, whereas DNA demethylation has the opposite effect. 12,13 Because the promoter regions of many genes contain high-density CpG dinucleotide aggregation regions, namely CpG islands, DNA methylation plays an important role in mammalian cell biology. 14,15 Recent studies have shown that DNA methylation may affect the multidirectional differentiation of stem cells by regulating the expression of specific genes, which results in various bone diseases that include osteoporosis and osteoarthritis. 16,17 Therefore, exploring DNA methylation provides a possibility to investigate novel molecular mechanisms of the osteogenic decline in DOP-ASCs.
Protein coding genes have been studied extensively, but they account for only 1.5% of the human genome. 18 Noncoding RNAs (ncRNAs), which account for up to 98% of the genome, were once ignored. These ncRNAs are divided into long noncoding RNAs (LncRNAs) and shortchain ncRNAs in accordance with the transcript length. LncRNAs are a class of ncRNAs with transcript lengths longer than 200 bases, which are localized in the nucleus or cytoplasm. 19,20 With the gradual deepening of the research on LncRNAs, many studies have reported that LncRNAs regulate gene expression at the epigenetic, transcriptional, and posttranscriptional levels, and have a potential regulatory effect on the cell fate of mesenchymal stem cells and the occurrence and development of specific diseases. 19,21,22 In particular, studies have confirmed that LncRNAs affect the expression of specific genes by regulating the DNA methylation level in their promoter region. [23][24][25] However, the molecular mechanism of LncRNAs in regulating the osteogenic differentiation and bone regeneration of DOP-ASCs is unclear. Therefore, an epigenetic mechanism, such as the regulation of LncRNAs or DNA methylation in the osteogenic ability of DOP-ASCs, has become a possible explanation for the decline in the osteogenic potential of DOP-ASCs.
In the present study, we subjected CON-ASCs and DOP-ASCs to mRNA/LncRNA expression profiling and MeDIP sequencing. The results showed a significant difference in the DNA methylation level of the promoter region of the Wnt signaling molecule sFrp2, which was related to LncRNA-AK137033. Subsequently, we conducted functional studies of sFrp2 and LncRNA-AK137033 in vivo and in vitro to explore the molecular mechanism that regulates bone differentiation of DOP-ASCs from the perspective of epigenetics.

| Diabetic osteoporosis animal model
All procedures related to animal experiments were reviewed and ap- Medical University. Fifty mice were randomly divided into CON and DOP groups. The CON group was fed an ordinary diet, and the DOP group was fed a high fat and sugar (HFS) diet. The HFS diet consisted of 65% standard chow, 10% fat, 20% sucrose, 2.5% cholesterol, and other necessary additives (BIOG, Beijing, China). The weight and blood glucose of mice were measured every week. After 4 weeks, 50 mice were fasted for 12 h, followed by injection of streptozotocin (STZ; Sigma, St Louis, USA) in the DOP group (140 mg/kg) or the same volume of citric acid-sodium citrate buffer (140 ml/kg) in the CON group. After the injection, CON mice were raised on standard chow and DOP mice were still raised on the HFS diet. Additionally, all mice were fed under appropriate conditions (20-25°C with 65%-80% humidity) with free access to drinking water and food. The blood glucose and body weight of each group of mice were recorded every 3-5 days. After 4 months of feeding, the diabetic osteoporosis model was established. Until the end of the study, the mortality rate of STZ-induced diabetic osteoporosis mice was 20%-33%.

| Hematoxylin and eosin (H&E), and Masson's staining
Samples were fixed in 10% paraformaldehyde for 24 h, decalcified in a decalcification solution for about 1 month (this step was omitted for soft tissue), washed with tap water for 48 h, dehydrated with an alcohol gradient, embedded in paraffin, and sectioned. The samples were then stained with H&E and Masson's dye.

| Isolation and culture of CON-ASCs and DOP-ASCs
CON-ASCs and DOP-ASCs were obtained from inguinal subcutaneous adipose tissue of CON and DOP mice, respectively. An adipose block from the groin was cut into pieces of about 1 mm 3 and evenly spread in a T25 culture flask. After covering the cap tightly, the flask was turned upside down and incubated at 37°C for about 5 min.
Then, α-modified Eagle's medium (Hyclone, Pittsburgh, USA) with 10% FBS (Schaumburg, USA) was carefully added to the culture flask to completely immerse the tissue blocks. Subsequently, the tissue blocks were incubated at 37°C with 5% CO 2 for 7-10 days to obtain passage 0 cells. The cells were passaged at 80%-90% confluence, and passage 3 cells were used in experiments.

| Immunofluorescence staining
Immunofluorescence staining was used to analyze the relative expres-  Table 2.
For AK137033 overexpression, the AK137033 cDNA sequence was amplified and subcloned into a pcDNA3.1 vector. An empty pGFP3.1 vector that carried eGFP was used as the negative control.
SiRNA and the plasmid were transfected into ASCs using the   Table 3.
AK137033 overexpression or AK137033 knockdown lentiviruses were transfected into ASCs in accordance with the manufacturer's instructions for in vivo experiments.

| Preparation of ASC-seeded BCP scaffolds
Biphasic calcium phosphate (BCP) scaffolds were provided by Sichuan University Research Center. After autoclaving the scaffolds, ASCs infected with lentiviruses were seeded on BCP scaffolds at 2 × 10 5 /cm 2 and cultured in osteogenic induction medium for 48 h for subsequent in vivo experiments.

| Establishment of a critical-sized calvarial bone defect model in mice and implantation of ASCseeded BCP scaffolds in vivo
ASC-seeded BCP scaffolds were prepared as the following groups: BCP scaffold seeded with CON-ASCs (CON-B), BCP scaffold seeded with CON-ASCs transfected with the knockdown with 65%-80% humidity). SPSS 19.0 software was used for statistical analysis. Data were tested by the t-test or one-way analysis of variance. Each experiment was repeated at least three times. The results are expressed as the mean ± standard deviation (SD). Data were significantly different at p < 0.05.

| Successful establishment of the diabetic osteoporosis animal model
After STZ injection, we continuously observed changes in body weight and blood glucose in mice. The blood glucose levels of CON mice were <9 M; the blood glucose levels of mice in the DOP group were maintained at >16.8 M; and the weight of mice in the DOP group was lower than that in the CON group ( Figure 1A). Compared with the CON group, H&E and Masson's staining of the pancreas showed that the volume of islets tissue in the DOP group was smaller, vacuolar degeneration had occurred, and inflammatory cells had infiltrated around islets ( Figure 1B). The femurs of CON and DOP mice were stained with H&E and Masson, and subjected to micro-CT at 4 months after STZ injection. The results of histochemical staining showed that, compared with the CON group, bone trabeculae were fewer and disordered in the DOP group, and the bone cortex had become thinner ( Figure 1C). These results were confirmed by micro-CT analysis ( Figure 1D). Compared with CON mice, statistical analysis showed that Tb.BV/TV and Tb. Th were decreased, but Tb.BS/BV had increased in DOP mice ( Figure 1E). On the basis of the above results, we concluded that the DOP mice model was established successfully.

| AK137033 silencing inhibits the Wnt signaling pathway in CON-ASCs by reducing the DNA methylation level of the sFrp2 promoter region
Previous studies have shown that the inhibition of the Wnt signaling pathway in DOP-ASCs may be related to AK137033 and changes in the DNA methylation level in the sFrp2 promoter region, and AK137033 was highly expressed in CON-ASCs. Therefore, we silenced AK137033

| AK137033 silencing decreases the osteogenic ability of CON-ASCs cells
To explore changes in the osteogenic differentiation potential after silencing AK137033 in CON-ASCs, we performed immunofluorescence, alizarin red, and ALP staining. At 3 days after osteogenic induction, immunofluorescence staining showed that the expression of RUNX2 and OPN in the SiRNA group was decreased compared to that in B and NC groups ( Figure 5A, B). Alizarin red staining revealed fewer mineralized nodules in the SiRNA group than in B and NC groups after 14 days of osteogenic induction ( Figure 5C). At 3 and 5 days of osteogenic induction, ALP staining showed less alkaline phosphatase produced by the SiRNA group than that by B and NC groups ( Figure 5D, E). These results suggested that silencing AK137033 reduced the osteogenic differentiation potential of CON-ASCs.

| AK137033 overexpression activates the Wnt signaling pathway in DOP-ASCs by increasing the DNA methylation level of the sFrp2 promoter region
At 3 days after osteogenic induction, compared with B and NC groups, qPCR and western blotting showed that the mRNA and protein levels of Wnt signaling pathway markers and downstream osteogenesis-related molecules were increased in the AK137033 plasmid (OE) group ( Figure 6A, B). Similar results were obtained after 5 days of osteogenic induction ( Figure 6C, D). Additionally, BSP results showed that the DNA methylation level of the sFrp2 promoter region in the OE group was increased after AK137033 overexpression ( Figure 6E). To activate the Wnt signaling pathway in DOP-ASCs, we transfected a plasmid that carried the AK137033 cDNA sequence into DOP-ASCs ( Figure 6F). Taken together, these observations indicated that AK137033 overexpression activated the Wnt signaling pathway in DOP-ASCs by increasing the DNA methylation level of the sFrp2 promoter region.
F I G U R E 3 AK137033 and a change in the methylation levels of the sFrp2 promoter region are involved in the regulation of the osteogenic differentiation potential of DOP-ASCs. A, Western blot analysis of sFrp2, p-GSK-3β, β-catenin, RUNX2, and OPN in CON-ASCs and DOP-ASCs. B, qPCR analysis of AK137033, sFrp2, β-Catenin, Runx2, and Opn in CON-ASCs and DOP-ASCs. C, ALP and alizarin red staining revealed less alkaline phosphatase activity and fewer mineralized nodules in the DOP group than those in the CON group. D, MeDIP sequencing showed that the DNA methylation level of the sFrp2 promoter region in DOP-ASCs was significantly higher than that in CON-ASCs (genomic coordinates: chr3, 83765880-83766230). E, Meth Primer analysis showed a large amount of CpG islands (CGIs) in the sFrp2 promoter region (genomic coordinates: chr3, 83765880-83766230). F, BSP confirmed that the methylation level in the sFrp2 promoter region (genomic coordinates: chr3, 83765880-83766230) of CON-ASCs was higher than that in DOP-ASCs. G, Correlation analysis of AK137033 and the DNA methylation level of the sFrp2 promoter region in CON-ASCs and DOP-ASCs. Data shown as the mean ± SD (n ≥ 3), *p < 0.05; **p < 0.01; ***p < 0.001

| AK137033 overexpression increases the osteogenic ability of DOP-ASCs
To investigate changes in the osteogenic differentiation potential of DOP-ASCs after overexpression of AK137033, immunofluorescence, alizarin red, and ALP staining were performed. After 3 days of osteogenic induction, immunofluorescence staining showed that the expression of RUNX2 and OPN in the OE group was increased compared with that in B and NC groups ( Figure 7A, B). Alizarin red staining showed more mineralized nodules in the OE group than those in B and NC F I G U R E 4 AK137033 silencing inhibits the Wnt signaling pathway in CON-ASCs by reducing the DNA methylation level of the sFrp2 promoter region. A, B, The mRNA and protein levels of Wnt signaling pathway makers and osteogenesis-related molecules in the SiRNA group were decreased after AK137033 silencing in CON-ASCs (osteoinduction for 3 days). C, D, After AK137033 was silenced in CON-ASCs, the mRNA and protein levels of Wnt signaling pathway markers and osteogenesis-related molecules were decreased in the SiRNA group (osteoinduction for 6 days). E, BSP results revealed that the DNA methylation level of the sFrp2 promoter region in the SiRNA group was downregulated compared with that in B and NC groups. Data shown as the mean ± SD (n ≥ 3), *p < 0.05; **p < 0.01; ***p < 0.001 groups at 14 days of osteogenesis ( Figure 7C). ALP staining revealed higher production of alkaline phosphatase in the OE group than that in B and NC groups at 3 and 5 days of osteogenic induction ( Figure 7D, E).
These changes indicated that overexpression of AK137033 enhanced the osteogenic differentiation potential of DOP-ASCs.

| In vivo verification of AK137033 in regulating the osteogenic ability of ASCs
To further assess the osteogenic role of AK137033 in vivo, we  Figure 8A, B). Next, we prepared ASC-seeded BCP scaffolds. Scanning electron microscopy and fluorescence microscopy showed that CON-ASCs and DOP-ASCs adhered to the surface and pores of BCP scaffolds ( Figure 8C).
Subsequently, we implanted ASC-seeded BCP scaffolds into a critical-sized calvarial bone defect model in CON and DOP mice ( Figure 8D). Among them, CON mice were implanted with a BCP scaffold seeded with CON-ASCs (CON-B), BCP scaffold seeded with CON-ASCs transfected with the knockdown lentivirus vector (CON-NC), and BCP scaffold seeded with CON-ASCs with AK137033 knockdown reversed the decrease in bone formation caused by DOP ( Figure 9A, B). In CON mice, histochemical staining revealed that the amount of F I G U R E 6 AK137033 overexpression activates the Wnt signaling pathway in DOP-ASCs by increasing the DNA methylation level of the sFrp2 promoter region. A, B, The mRNA and protein levels of Wnt signaling pathway makers and osteogenesis-related molecules in the OE group were increased after AK137033 overexpression in DOP-ASCs (osteoinduction for 3 days). C, D, After AK137033 overexpression in DOP-ASCs, the mRNA and protein levels of Wnt signaling pathway markers and osteogenesis-related molecules were increased in the OE group (osteoinduction for 6 days). E, BSP results revealed that the DNA methylation level of the sFrp2 promoter region in the OE group was upregulated compared to that in B and NC groups. F, Cellular uptake of NC and OE plasmids by DOP-ASCs after treatment for 48 h. Data shown as the mean ± SD (n ≥ 3), *p < 0.05; **p < 0.01; ***p < 0.001 fibrotic and mineralized new bone in the shRNA group was less than that in B and NC groups. In DOP mice, the bone formation ability of the LVRNA group was rescued compared to that in B and NC groups ( Figure 9C, D). On the basis of these results, we concluded that cer. 47 However, no study has revealed the specific role of sFrp2 in regulating the osteogenic potential of DOP-ASCs through F I G U R E 8 Preparation of ASC-seeded BCP scaffolds and the critical-sized calvarial bone defect mouse model. A, B, Transfection efficiency verified by qPCR and western blot analysis after CON-ASCs were infected with the shRNA lentivirus, and DOP-ASCs were infected with LVRNA. C, Scanning electron microscopy and fluorescence microscopy of prepared CON-ASC-seeded BCP scaffolds and DOP-ASC-seeded BCP scaffolds. D, Establishment of the criticalsized calvarial bone defect mouse model and implantation of ASC-seeded BCP scaffolds. Data shown as the mean ± SD (n ≥ 3), *p < 0.05; **p < 0.01; ***p < 0.001 F I G U R E 9 Evaluation of skull repair at 8 weeks after implantation. A, B, Micro-CT scanning showed that the knockout of AK137033 decreased Tb.BV/TV and Tb. Th, while Tb.BS/BV was increased in CON mice. In DOP mice, overexpression of AK137033 reversed the decreases in Tb.BV/TV and Tb. Th, and the increased Tb.BS/BV caused by DOP. C, D, H&E and Masson's staining showed that, in CON mice, the amount of fibrotic and mineralized new bone in the shRNA group was less than that in B and NC groups. In DOP mice, the bone formation ability of the LVRNA group was rescued compared with B and NC groups. Data shown as the mean ± SD (n ≥ 3), *p < 0.05; **p < 0.01; ***p < 0.001 the Wnt signaling pathway. In the present study, we performed mRNA/LncRNA expression profiling and MeDIP sequencing of CON-ASCs and DOP-ASCs. The results showed a significant difference in the DNA methylation level of the sFrp2 promoter region in the two groups, which was related to LncRNA-AK137033.
Moreover, we verified the AK137033 expression level, Wnt signaling pathway difference, and osteogenic differentiation potential LncRNAs are a kind of long-chain RNA (more than 500 nt) that lacks a protein-coding ability. 48,49 There is increasing evidence of the roles of LncRNAs in many important biological processes that include gene transcription, mRNA shearing, cell cycle control, epigenetic regulation, and cellular immunity. Therefore, we conclude that AK137033 regulates osteogenesis of ASCs in vivo.

In summary, in vitro and in vivo experiments demonstrated that
LncRNA-AK137033 regulates the osteogenic potential of DOP-ASCs by modulating the Wnt signaling pathway via DNA methylation in the sFrp2 promoter region. Our study provides an epigenetic explanation for the regulatory mechanism of the osteogenic potential in DOP-ASCs and is an important reference for the treatment of bone defects in DOP patients.

CO N FLI C T O F I NTR E S T
The authors declare that there are no competing interests.

AUTH O R CO NTR I B UTI O N S
All authors have made important contributions to this study. Xiaoxiao Cai designed the study and revised the manuscript. Jingang Xiao conceived and initiated the study, analyzed the data, and provided funding. All authors have read and approved the manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data included in this article can be obtained from corresponding author upon reasonable requirements.