LncRNA SNHG1 modulates adipogenic differentiation of BMSCs by promoting DNMT1 mediated Opg hypermethylation via interacting with PTBP1

Abstract Recent evidence indicates that the abnormal differentiation of bone marrow‐derived mesenchymal stem cells (BMSCs) plays a pivotal role in the pathogenesis of osteoporosis. LncRNA SNHG1 has been found to be associated with the differentiation ability of BMSCs. In this study, we aimed to elucidate the role of lncRNA SNHG1 and its associated pathway on the differentiation of BMSCs in osteoporosis. Mice that underwent bilateral ovariectomy (OVX) were used as models of osteoporosis. Induced osteogenic or adipogenic differentiation was performed in mouse BMSCs. Compared to sham animals, lncRNA SNHG1 expression was upregulated in OVX mice. Also, the in vitro expression of SNHG1 was increased in adipogenic BMSCs but decreased in osteogenic BMSCs. Moreover, overexpression of SNHG1 enhanced the adipogenic capacity of BMSCs but inhibited their osteogenic capacity as determined by oil red O, alizarin red, and alkaline phosphatase staining, while silencing of SNHG1 led to the opposite results. LncRNA SNHG1 interacting with the RNA‐binding polypyrimidine tract‐binding protein 1 (PTBP1) promoted osteoprotegerin (Opg) methylation and suppressed Opg expression via mediating DNA methyltransferase (DNMT) 1. Furthermore, Opg was showed to regulate BMSC differentiation. Knockdown of SNHG1 decreased the expressions of adipogenic related genes but increased that of osteogenic related genes. However, the knockdown of Opg partially reversed those effects. In summary, lncRNA SNHG1 upregulated the expression of DNMT1 via interacting with PTBP1, resulting in Opg hypermethylation and decreased Opg expression, which in turn enhanced BMSC adipogenic differentiation and contributed to osteoporosis.


| INTRODUC TI ON
Osteoporosis is a systemic skeletal disorder featured by less bone mass and weakened bone strength, resulting in bone fragility and occurs when bone resorption exceeds new bone formation. 1 Approximately 30% of females and 12% of males are affected sometime in their lives, especially among elderly and postmenopausal women. 2 The consequences of osteoporotic fractures can be serious, including disability, reduced dependence and even death. Osteoporosis has become a major and prevalent disease worldwide and significantly increased the burden of health care. 3 The aim of osteoporosis treatment is to restore bone homeostasis and maintain normal bone mass. Although there are a number of available agents for osteoporosis treatment in the market, concerns have been raised regarding the inherent side effects of their longterm use, prompting the exploration of novel therapeutic targets. 4 However, the pathophysiology of osteoporosis is multifactorial, including the imbalance between osteoblasts and osteoclasts, disrupted microarchitecture, and increased adipogenesis in the bone marrow, as well as changes in angiogenesis, oxidative stress and genetic/epigenetic factors. 5 Recent evidence indicates that the aberrant differentiation of bone marrow-derived mesenchymal stem cells (BMSCs) is one of the major contributing factors for the development of osteoporosis and therefore could be a potential target for the treatment of osteoporosis. 6 Long non-coding RNAs (lncRNAs) are a class of transcripts containing 200 or more nucleotides and have low or no capability of protein-coding. 7 It is conceivable that lncRNAs play important roles in gene expression on both transcriptional and posttranscriptional levels and regulating various biological functions and pathological implications, including those of the bone. 8 Indeed, accumulating evidence suggests that various lncRNAs are involved in bone homeostasis. For example, Huang et al. 9 demonstrated that overexpressed lncRNA H19 facilitated osteoblast differentiation of BMSCs and increased heterotopic bone formation, while knockdown of lncRNA H19 abrogated those effects. LncRNA LINC00311 has been found to induce osteoclast proliferation and inhibit osteoclast apoptosis in osteoporotic rats. 10 A recent study also discovered that the expression of lncRNA SNHG1 could inhibit the osteogenic differentiation of BMSCs. 11 Noncoding RNA sequences, including lncRNAs, exert their diverse functions in many cellular processes through their interactions with RNA-binding proteins (RBPs), where RBPs play multiple roles in post-transcriptional gene regulation. 12 RBPs are transcription factors involved in co-or post-transcriptional regulation of gene expression by affecting mRNA metabolism. 11 The polypyrimidine tract-binding protein 1 (PTBP1) is one of the RBPs that regulate mRNA decay stability and pre-mRNA splicing. 13 Apart from that, PTBP1 is also involved in the invasion of low malignancy cancer cells induced by BMSCs 14 and is associated with the differentiation of human adipose-derived MSCs toward definitive endoderm. 15 To date, the potential role of PTBP1 in the differentiation of BMSCs has not been reported. DNA methyltransferase (DNMT) 1 is an active participant of epigenetic modification and is responsible for the passive transmission of genomic methylation patterns via maintenance of DNA methylation during cell division. 16 Osteoprotegerin (Opg) is an important regulator during BMSC differentiation and adipogenesis. 17 As a potential target of DNMT1 and one of the key regulators of osteoclast differentiation and action, Opg functions as an inhibitor of the receptor activator of the nuclear factor kappa-B (RANK) signalling pathway by competitively inhibiting the interaction between Opg ligand and RANK on osteoclasts and their precursors. 18 The hypermethylation of Opg has also been found in patients with primary osteoporosis. 19 Moreover, previous studies indicated that ln-cRNA SNHG1 could bind to and regulate the expression of DNMT1 in certain diseases, such as cancers. 20,21 Therefore, we hypothesized that lncRNA SNHG1 could regulate the expression of DNMT1 and subsequently influence Opg methylation status, contributing to the occurrence of osteoporosis, but it needs further determination.
The aim of the present study was to elucidate the functional role of lncRNA SNHG1 in osteoporosis pathogenesis as well as the underlying mechanistic pathways.

| Animals
All animals were housed and handled in accordance with the Guide for the care and use of laboratory animals. All animal studies were approved by the Experimental Animal Ethics Committee of Ningbo University (2020-106). The mouse model of oestrogen deficiencyinduced osteoporosis was established as previously described. 22 Female C57BL/6 mice weighing 20-25 g at 11-12 weeks of age were housed under standard conditions with a 12 h light/dark cycle. The temperature and the humidity were controlled at 20-25°C and 40%-70% respectively. Mice had free access to standard chow and water.
Before randomization, all animals were acclimatized to the environment for at least 5 days. Subsequently, mice were randomized to undergo bilateral ovariectomy (OVX) or sham surgery (sham). After 8 weeks, the mouse femurs were collected after euthanasia and fixed with paraformaldehyde for 48 h. Blood samples were collected and bone tissues were harvested. Samples were then frozen immediately in liquid nitrogen and stored at −80°C for further experiments.

| Clinical tissues collection
In this study, six pairs of bone tissues were harvested from osteoporotic patients and non-osteoporotic patients (controls) underwent routine therapeutic surgery in our hospital and instantly stored at −80°C for subsequent analyses. The experimental protocols were reviewed and approved by the Ethics Committee of HwaMei Hospital, University of Chinese Academy of Sciences (PJ-NBEY-KY-2019-081-01) and all of the methods in this study were in accordance with the approved guidelines. All participating patients gave written informed consent before any study procedures occurred.

| Plasmid construction and transfection
The full-length lncRNA SNHG1 sequences were amplified by PCR and then inserted into the pcDNA 3.1 vector (Life Technologies) to establish a vector overexpressing SNHG1 (pcDNA-SNHG1) according to the manufacturer's instructions. The empty vector was served as a negative control (NC). Using the same method as above, PTBP1 and Opg were cloned and termed as pcDNA-PTBP1 and pcDNA-Opg respectively. Short-hairpin (sh) RNAs directed against lncRNA SNHG1, PTBP1, Opg and sh-negative control (sh-NC) were obtained from GenePharma. Cell transfection was performed using Lipofectamine 3000 (Life Technologies). Briefly, BMSCs grown to 70% confluence in alpha-MEM medium were incubated with Lipofectamine 3000 reagent and a pcDNA or an sh-RNA for 48 h before culturing in adipogenic or osteogenic induction medium. For in vivo transfection, lenti-sh-SNHG1 and/or sh-Opg were injected into the circulation via the tail vein. Mice were finally euthanized after continuous injection for 3 weeks.

| Haematoxylin and eosin staining
The mouse femurs were fixed in phosphate-buffered 10% paraformaldehyde for 24 h followed by decalcification with 8% formic acid at 4°C under continuous shaking. Dehydrated tissues were then embedded in paraffin and sliced into 5 µm sections for haematoxylin and eosin (HE) staining. Haematoxylin and eosin staining was conducted according to routine protocols. Briefly, after deparaffinization and rehydration, 5 μm longitudinal sections were stained with haematoxylin solution for 5 min followed by 5 dips in 1% acid ethanol (1% HCl in 70% ethanol) and then rinsed in distilled water. Then the sections were stained with eosin solution for 3 min and followed by dehydration with graded alcohol and clearing in xylene. At least 5 fields were selected on a random basis and photographed under light microscopy (Nikon) for histological evaluation.

| ORO and ARS staining
Oil red O staining and ARS staining were performed 14 days after osteogenic and adipogenic differentiation of BMSCs as previously described. 23 The cell culture medium was discarded and the cells were

| Alkaline phosphatase (ALP) staining and activity
Alkaline phosphatase staining was performed as described previously on day 14 following osteogenic differentiation of BMSCs. 24 In brief, cells were washed with PBS and fixed by using 4% paraformaldehyde for 10 min and then incubated in ALP solution (CWBIO) for staining for 20 min protected from light at RT. After three times wash, calcium mineralization pictures were obtained using an inverted optical microscope. After incubation with 10 mM p-nitrophenyl phosphate (Sigma) for 15 min, ALP activity was calculated at 420 nm by a microplate reader.

| Fluorescence in situ hybridization (FISH) analysis
Biotin-labelled sense and antisense SNHG1 probes and the corresponding control oligo for FISH analysis were obtained from GenePharma and FISH assays were performed as previously described. 25 For immobilization, cells were treated with 4% paraformaldehyde for 20 min at RT. After washing with PBS, the cells were prehybridized with a hybridization solution. Then, the cells were incubated with the SNHG1 probes overnight at 37°C. An anti-Biotin-Cy3 antibody (C5585; Sigma) was used for signal detection. Cell nuclei were counterstained with DAPI to assess the nuclear morphology.
Samples were incubated with diluted DAPI solution at a concentration between 0.5 µg/ml in PBS for 5 min at RT in the dark. The immunofluorescence was measured by using an IX73 microscope (Olympus).

| RNA pull down assay
The biotinylated RNA pull-down assay was performed as previously described in order to determine the interaction between lncRNA

| The RNA immunoprecipitation assay
The RNA immunoprecipitation (RIP) assay was conducted by using a commercial EZ-Magna RIP™ Kit (Millipore) according to the manufacturer's instructions. In brief, cells (5 × 10 6 ) were treated with RIP lysis buffer at 4°C for 30 min and then whole-cell lysates were incubated with RIP buffer containing magnetic beads conjugated to antibodies against Ago2 (Millipore) or anti-PTBP1 antibody (ab133734; Abcam) or anti-SNRP70 antibody (ab51266; Abcam) or immunoglobulin (Ig) G (Abcam) for 2 h at RT with gentle shaking. The coprecipitated RNAs associated with PTBP1 were extracted with the Trizol reagent (Life Technologies) and analysed by quantitative real-time PCR (qPCR).
Enrichment associated with SNRP70 and species-matched normal IgG served as positive and negative RIP controls respectively. Total RNA was considered as input controls.

| qPCR assay
Animal bone tissues were harvested from OVX and sham mice.
Human bone samples were collected from patients diagnosed with osteoporosis and control patients. Total RNA was extracted from BMSCs and bone tissues by using Trizol (Life Technologies) method.
Cells or tissues were homogenized in Trizol and extracted with chloroform (0.2 ml/ml Trizol). Isopropanol (0.5 ml/ml Trizol) was used to precipitate the RNA from its aqueous phase. RNA was pelleted by centrifugation at 7500 g and 4°C for 5 min, washed in 75% ethanol,

| Subcellular fractionation
The separation of the nuclear and cytosolic fractions was performed using the NE-PER Nuclear and Cytoplasmic Extraction Reagents (Thermo Scientific) according to the manufacturer's instructions.
Briefly, cells were harvested and incubated with ice-cold CER I on ice for 10 min. Then, ice-cold CER II was added, followed by incubation on ice for another 5 min. After centrifugation at about 4000 g for 5 min, the supernatant was transferred to a new tube. This was the cytoplasmic fraction. The pellet was suspended with ice-cold NER and vortexed on the highest setting for 15 s every 10 min for a total of 40 min. After centrifugation at about 4000 g for 5 min, the supernatant was collected as the nuclear fraction. Total RNA was isolated from cytoplasmic and nuclear extracts. qPCR was performed to assess the relative proportion in the nuclear and cytoplasmic fractions.
U6 and GAPDH were served as nuclear and cytoplasmic markers, respectively.

| Western blotting
Whole-cell and tissue lysates were prepared by using the radioimmunoprecipitation assay (RIPA) buffer supplemented with pro-

| Statistical analysis
In all in vitro studies, experiments were conducted at least three replications with at least three independent samples each time.
In in vivo studies there were six mice per group for all animal studies. The data were presented as mean ± standard deviation (SD

| LncRNA SNHG1 is associated with BMSC differentiation in osteoporosis
In order to elucidate the role of lncRNA SNHG1 and its relationship with BMSC differentiation in osteoporosis, we first looked into the expression of lncRNA SNHG1 in in vivo mouse models and osteoporosis patients. The qPCR revealed that the expression of lncRNA SNHG1 was significantly upregulated in bone tissues and blood serums of OVX mice, when compared to sham control mice ( Figure 1A,B). Similarly, we observed increased lncRNA SNHG1 expressions in tissues of osteoporosis patients than those of control patients ( Figure 1C). We further investigated lncRNA

| LncRNA SNHG1 enhances BMSC adipogenic differentiation but inhibits BMSC osteogenic differentiation
We explored the role of lncRNA SNHG1 in BMSC differentiation in  (Figure 2D), but inhibited ALP activity with a smaller and less intense stained area after ALP staining ( Figure 2G). Taken together, our observations indicate that lncRNA SNHG1 regulates BMSC differentiation by enhancing adipogenic differentiation but inhibiting osteogenic differentiation.

| LncRNA SNHG1 interacts with PTBP1 and promotes the expression of PTBP1
In order to better understand the function and mechanism of lncRNA SNHG1, the FISH and qPCR were conducted to examine specific subcellular localization and cell fractionation of lncRNA SNHG1 in cells cultured in normal medium. The FISH showed that lncRNA SNHG1 was widely localized in the cell nuclear ( Figure 4A). Consistently, the qPCR revealed that lncRNA SNHG1 largely displayed a nuclear distribution (>60%) ( Figure 4B). To determine the potential interaction of lncRNA SNHG1 with PTBP1, we performed RNA pull-down assays in BMSCs using in vitro synthesized biotinylated sense and antisense SNHG1 RNAs ( Figure 4C). After incubation with BMSC extracts, western blot showed that PTBP1 was identified as a binding partner of lncRNA SNHG1 ( Figure 4D). Alternatively, the RIP assay was performed using a PTBP1 antibody to detect lncRNA SNHG1 in the PTBP1 precipitates. We found that PTBP1 antibody significantly precipitated lncRNA SNHG1 compared with the

| LncRNA SNHG1 interacts with PTBP1 and promotes the expression of DNMT1
Previous studies suggested that RNA-binding proteins could affect DNMT stability and regulate DNA methylation. 28,29 We were particularly interested in DNMT1, which is one of the key enzymes maintaining DNA methylation. 30 In order to illustrate the potential roles

| Opg regulates cell differentiation of BMSCs
The role of Opg in BMSC differentiation was examined in BMSCs treated with osteogenic or adipogenic inducing media for 14 days.  Figure 8D).
Thus, we conclude that SNHG1 plays a role in the osteoporotic changes in the in vivo model of osteoporosis, which may be mediated by Opg.

| DISCUSS ION
As one of the epigenetic regulators, lncRNAs play important roles in gene expression and multiple biological processes. 32 Several studies suggest that lncRNAs are implicated in bone remolding by affecting the proliferation and function of osteoblasts and osteoclasts, as well as the differentiation of BMSCs. 33 For example, lncRNA ANCR had been demonstrated to inhibit osteoblast differentiation and was essential to maintain osteoblasts in an undifferentiated state. 34 On the contrary, lncRNA H19 promoted osteoblast differentiation of BMSCs, as evidenced by increased expressions of Runx2 and ALP. 9 In the present study, lncRNA SNHG1 was found to enhance BMSC adipogenic differentiation but inhibit BMSC osteogenic differentiation, implicating its role in osteoporosis. Indeed, SNHG1 has been known to be a novel oncogenic lncRNA aberrantly expressed in a number of cancers and was linked to cell growth, migration and invasion. 35 with lncRNAs, such as H19, HCG22 and MACC1-AS1. [38][39][40] The interaction between lncRNA SNHG1 and PTBP1 had also been confirmed by RNA pull-down and immunoprecipitation assays in our study. Moreover, by interacting with PTBP1, lncRNA SNHG1 upregulated the expression of DNMT1, which in turn promoted Opg hypermethylation. DNA methylation through adding a CH3 methyl group to cytosine by a DNMT is considered a long-term, relatively stable epigenetic modification, resulting in inhibition of transcription and downregulation of target genes. 41 In this regard, accumulating evidence suggested that the methylation status of Opg probably functioned as a "main switch" in the pathogenesis of osteoporosis. 19,42 The observations of Behera et al. 43  The regulatory function of Opg on BMSC differentiation and adipogenesis has also been suggested by a series of studies on transgenic animals. 44,45 In Opg knock-out mice, mice developed early-onset osteoporosis characterized by increased trabecular porosity and adipocyte accumulation in the bone marrow, 17 while administration of Opg protein effectively reversed the osteoporotic bone phenotype presented in Opg-deficient mice. 44 Well agree with that, we found that Opg overexpression inhibited adipogenesis but promoted osteogenesis in BMSCs than cells transfect with control vector. Interestingly, we also found that lncRNA SNHG1 regulated the expression of Opg in osteoporotic mice; knockdown of SNHG1 improved osteoporotic changes, while simultaneous knockdown of Opg could partially reverse the beneficial effects of SNHG1 silencing on the loss of bone mass. However, in our study, there was lack of comparisons of gene and protein expressions during the duration of treatments. Additionally, MicroCT (or µCT) analysis was not performed to support the function of SNHG1 in osteoporosis. These were limitations of the present study.
In sum, our study demonstrated that lncRNA SNHG1 upregu-

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest. Writing-review & editing (equal).

E TH I C A L A PPROVA L
All animal procedures were performed in accordance with the guidelines by the National Institutes of Health Guide for the care and use of laboratory animals. All protocols were approved by the Animal Ethics Committee of our hospital. The clinical study was approved by the Ethics Committee of our hospital, and written informed consent was obtained from all participants.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data generated or analysed during this study are included in this published article.