Scara3 regulates bone marrow mesenchymal stem cell fate switch between osteoblasts and adipocytes by promoting Foxo1

Abstract Objectives Scavenger receptor class A, member 3 (Scara3) was involved in adipogenesis. However, the effect of Scara3 on the switch between osteogenesis and adipogenesis of bone marrow mesenchymal stem cells (BMSCs) remains elusive. Materials and Methods The correlations between SCARA3 with the osteogenic‐related were analysed based on the GTEx database. The effects of Scara3 on osteogenic or adipogenic differentiation of BMSCs were evaluated by qPCR, Western blot (WB) and cell staining. The mechanisms of Scara3 regulating Foxo1 and autophagy were validated by co‐expression analysis, WB and immunofluorescence. In vivo, Scara3 adeno‐associated virus was injected into intra‐bone marrow of the aged mice and ovariectomized (OVX) mice whose phenotypes were confirmed by micro‐CT, calcein double labelling and immunochemistry (HE and OCN staining). Results SCARA3 was positively correlated with osteogenic‐related genes. Scara3 expression gradually decreased during adipogenesis but increased during osteogenesis. Moreover, the deletion of Scara3 favoured adipogenesis over osteogenesis, whereas overexpression of Scara3 significantly enhanced the osteogenesis at the expense of adipogenesis. Mechanistically, Scara3 controlled the cell fate by promoting Foxo1 expression and autophagy flux. In vivo, Scara3 promoted bone formation and reduced bone marrow fat accumulation in OVX mice. In the aged mice, Scara3 overexpression alleviated bone loss as well. Conclusions This study suggested that Scara3 regulated the switch between adipocyte and osteoblast differentiation, which represented a potential therapeutic target for bone loss and osteoporosis.


| INTRODUC TI ON
Osteoporosis is characterized by low bone mass and microarchitectural dysfunction. 1 Bone marrow mesenchymal stem cells (BMSCs) have potentials to differentiate into adipocytes and osteoblasts, whose cell fate was altered in osteoporotic individual. [2][3][4][5] Accumulated studies identified various master regulators associated with the adipogenic and osteogenic lineage commitment. 6,7 Some non-coding RNA, such as miR-188 and Bmncr, were declining with ageing, which both contributed to reduced bone mass and increased marrow adipose tissue (MAT) through controlling the BMSCs fate. 8,9 Other studies showed various protein modulators were involved in the cell fate switch of BMSCs, such as RB, PGC-1α and FOXP1. [10][11][12] However, the molecular network elucidating the bone-fat balance still needs further investigations.
SCARA3 encodes a macrophage scavenger receptor-like protein which is ubiquitously expressed in human tissues. 13 SCARA3 was reported to protect cells from oxidative stress-induced cell damage through removing oxidizing molecules or harmful products of oxidation. 13,14 Han et al 13 showed that the SCARA3 expression could be promoted upon the oxidative stress stimulation. However, the increase of SCARA3 expression diminished when there was excessive oxidative stress. Several disorders, such as obesity and multiple myeloma, characterized by a high level of oxidative stress, displayed a lower expression of SCARA3. 15,16 In addition, SCARA3 was recognized as a tumour suppressor-related gene because of its downregulation in prostate cancer tissues and involvement in cancers metastases and progression. 14,15 Through analysing a data set obtained from GEO database (GSE11 5068), 17 we previously found that the expression of Scara3 was downregulated in Ad-MSCs isolated from young mice than from old mice. 16 Moreover, our previous study showed that the SCARA3 gene was a critical regulator of adipogenic differentiation. 16 Balla et.al 18,19 demonstrated that SCARA3 transcription was significantly downregulated in osteoporosis bone tissue, but they did not further explored the functional roles of SCARA3 in bone formation. Based on these evidence, we investigated whether Scara3 played a role in controlling the bone-fat balance.
Autophagy, as a major catabolic process responsible for the degradation of damaged macromolecules and organelles, is now well-established as a regulator of bone physiology and bone-related disorders. [20][21][22][23][24] Previous studies showed that autophagy can affect the differentiation of osteoblasts and osteoclasts. 21 Moreover, aberrant autophagy could result in the dysfunction of cell fate in BMSCs. 20,25 Liu et.al 26 recently showed that autophagy receptor OPTN (optineurin) regulated bone-fat balance by regulating the cell fate switch.
However, it is unclear that if SCARA3 gene is involved in autophagy.
In this study, we revealed Scara3 as an important modulator that favoured osteogenic differentiation over adipogenic differentiation.
Scara3 plays an important role in autophagy signalling through regulating Foxo1. Moreover, we found that overexpression of Scara3 through intra-bone marrow injection of AAVs-Scara3 can lead to the increase of bone formation and decrease of bone marrow fat in ovariectomy (OVX) mice. Together, our study reported a new potential target to prevent and treat osteoporosis.

| Bioinformatic analysis
Genotype-Tissue Expression (GTEx) program (https://www.gtexportal.org/) is a research project of normal tissue specific gene expression and regulation of comprehensive public database. We downloaded the raw genes expression data, ID conversion files, clinical phenotypes and GTF annotation file from the GTEx program for the co-expression calculation. Next, we integrated the gene expression and GTF annotation file with clinical phenotypes and then removed the data without relevant clinical information. Based on the fact that RUNX2, SP7 and BALAP are important regulatory genes in the process of osteogenic differentiation, we further analysed the correlation between SCARA3 gene and three key genes of osteogenic differentiation. Pearson's correlation (R) and P values were calculated. The P-value has been marked as '0' when it is less than BioGPS (http://biogps.org/#goto=welcome) is a free, extensible and customizable gene annotation portal that has resource for learning then gene and protein function. We extracted the expression data of Scara3 in osteoblasts at the day 5, day 14 and day 21 from Geneatlas MOE430, GCRMA.

| Mice
All mice with a C57/B6 background were purchased from Shanghai Slac Laboratory Animal Co. The 2-month-old female mice were performed the ovariectomy (OVX) surgery. Then, random allocation was taken to divide the mice after OVX. One week later, 10 11 vg AAV (adenoassociated virus) was injected into femur bone marrow of mice in the vehicle group. For the treatment group, 10 11 vg AAV was injected into femur bone marrow. 15 month-old male mice were also treated with control AAVs or AAVs-Scara3 at the same dosage using intra-bone marrow injection. Five mice were used for each group of each independent experiment. All protocols associated mice's care and experiments used in this study were reviewed and approved by the Animal Care and Use Committees of the Laboratory Animal Research Center at Xiangya Medical School of Central South University. All animals were maintained in a specific pathogen-free facility of the Laboratory Animal Research Center at Central South University, with free access to food and water prior to the initiation of experiments.

| Intra-bone marrow injection
Intra-bone marrow injection was carried out to deliver the AAV (10 11 vg) into mice. Briefly, mice were anaesthetized. Then, the hair near the knee joints were shaved. The knee joint was exposed using the microscissors to separate muscle tissue and using the tweezers to push the tendon to left side. Then, a 29-gauge insulin syringe was inserted into bone marrow cavity. 10 11 vg in 10-15 μL of AAVs was delivered into the bone marrow cavity. The muscle and skin were stitched.

| Ovariectomy
Ovariectomy surgery was performed as described before. 27 Five mice in each group. Briefly, mice were anaesthetized. Then, the hair on the waist were shaved. The kidney and the white adipose tissue on the kidney were seen after cut the skin and peritoneum. After removal of the ovaries, the oviducts were ligated and the peritoneum and skin were sealed.

| Calcein double labelling
The mice were treated with calcein at the dose of 0.5 mg/per mice (Sigma-Aldrich) using intraperitoneal injection. The first injection was carried out at 10th day before euthanasia, and the second injection was given at the second day before euthanasia. The femora were dissected and fixed in the 4% paraformaldehyde overnight.
Calcein double labelling was performed in the 5 μm longitudinal sections of undecalcified bone slice to evaluate MAR and BFR using Image-Pro Plus 6.0. Four randomly chosen visual fields in the distal metaphysis of the femur were measured to test trabecular bone formation in femora. The images were captured by a fluorescence microscope (Leica).

| Histochemistry and histomorphometry analyses
Femora were dissected from mice and fixed with 4% PFA (Paraformaldehyde) overnights, followed by the decalcification in 14% EDTA for 1 week. Then, decalcified bones were embedded in the paraffin and cut into 4μm-thick sections. To assess the capacity of bone formation, the OCN (Abcam) staining and HE staining were performed on the bone tissue slice according to the previous described method. 27 Briefly, for the OCN staining, bone sections were processed for antigen retrieval followed by the blocking by 3% BSA with Triton 100. Then, the bone sections were incubated by the primary antibody against osteocalcin (catalog M173; Takara) overnight at 4°C. Next, HRP-DAB cell from a tissue staining kit was used to detect the immunoactivity according to the manufacturer's instruction. For the HE staining, bone sections were conducted according to standard protocol. Briefly, sections were stained by haematoxylin for 30 seconds and eosin for 3 minutes. isolated from 1-month-old male C57/B6 mice according to the previous described method. 8,9 Briefly, mice were anaesthetized using 5 μL/g 5% pentobarbital sodium. Then, the femurs and tibias were dissected from mice and put in the PBS with 100 U/mL penicillin and 100 µg/mL streptomycin. After washing the bone tissues in the PBS with 100 U/mL penicillin and 100 µg/mL streptomycin three times, the both ends of bones were cut off and the bone marrow were flushed out through culture medium in syringe. Then, the cells were cultured in DMEM with 5% FCS supplement with 100 U/mL penicillin and 100 µg/mL streptomycin overnight. Next day, the nonadherent cells were removed using PBS to obtain purified BMSCs.
After 2 days of osteogenic differentiation, the cell lysates were washed using PBS followed by homogenized for ALP activity assay by spectrophotometric measurement of p-nitrophenol release using an Alkaline Phosphatase Assay Kit according to the manufacturer's instructions (P0321S, Beyotime).
After 7 days of osteogenic differentiation, alkaline phosphatase staining (ALP staining) was carried out to assess the capacity of mineralization as described before. 8 Briefly, the washed cells were fixed in 10% paraformaldehyde for 5 minutes. Then, cells were incubated in ALP incubation buffer (0.2 g barbital sodium, 0.4 g magnesium sulphate, 0.2 g calcium chloride and 0.3 g beta-glycerophosphate, 10 mmol/L β-glycerol phosphate and 50 mmol/L ascorbate-2-phosphate) at 37°C for 2 hours. Next, 2% calcium chloride was used to wash the cells and 2% cobaltous nitrate was used to incubate cells for 5 minutes. Then, cells were incubated in 1:80 ammonium sulphate for 10 seconds.
After 21 days of osteogenic differentiation, Alizarin Red staining was performed according to the manufacturer's instructions (MUBMX-90021; Cyagen Biosciences). Briefly, cells were washed using PBS three times followed by 4% paraformaldehyde for 30 minutes. After washed by PBS for three times, cells were stained in Alizarin red solution at 37°C for 5 minutes. The images were captured by the microscope.

| Adipogenic differentiation and oil red staining
To induce adipogenic differentiation of BMSCs, primary BMSCs with or without previous treatments were cultured in 6-well plates at 2.5 × 10 6 cells per well with the mesenchymal stem cell adipogenicinduced medium (MUBMX-90031; Cyagen Biosciences). The culture medium A and B were alternately used every 3 days.
Oil red staining was performed at the 12 days of adipogenic differentiation. Briefly, cells were washed using PBS three times followed by 4% paraformaldehyde for 5 minutes. After washed by PBS for three times, cells were stained in oil red solution at 37°C for 3 minutes (3 mL Oil red was dissolved in 2 mL PBS). The stained cells were observed using the microscope.

| qRT-PCR analysis
For analysis of mRNA expression, RNA of cultured cells from six-well-plate was extracted using 1ml TRIzol reagent. 1000 ng of RNA was reversetranscribed into first-strand cDNA using the Reverse Transcription Kit (Accurate Biology). Primers were designed in UCSC database or acquired from ORIGEN database. All primers have been blasted and tested their efficiency. SYBR Green PCR Master Mix (Takara) was used to perform qPCR. mRNA expression was normalized to the reference gene Gapdh.

| Immunofluorescence
For immunofluorescence analysis, after the different treatment, the

| Statistical analysis
Data were imported into Excel and scaled and normalized to appropriate controls. Unpaired, two-tailed Student's t test was performed for the comparisons of two groups. One-way ANOVA was performed for the comparison for multiple groups.

| Ethics statement
The animal study was reviewed and approved by Xiangya Hospital of Central South University of ethics committee.

| Scara3 is a critical modulator involved in adipogenesis and osteogenesis of BMSCs
Bone marrow mesenchymal stem cells have multiple potential capacities to differentiate into osteoblasts and adipocytes, which is involved in the fat-bone balance. 8,29,30 To identify the key modulators of adipogenic and osteogenic differentiation, we firstly analysed the correlation between the differentially expressed genes with osteogenic-related and adipogenic-related genes. Based on our previous analysis suggesting that SCARA3 was highly negatively correlated with the expression of adipogenic-related genes, 16 we then investigated its correlation with osteogenic-related genes. The bioinformatics analysis revealed that SCARA3 is positively correlated with osteogenic-related genes, such as RUNX2 (Pearson's r = .36, P = 0), SP7 (Pearson's r = .36, P = 0) and BGLAP (Pearson's r = .45, P = 0) in 7858 types of tissues ( Figure 1A-C).
As expected, we found that the expression of Scara3 was remarkably downregulated in BMSCs during adipogenesis ( Figure 1D,E). RT-PCR and WB results also verified that Scara3 expression was upregulated during osteogenesis ( Figure 1F,G, Figure S1).

| The deficiency of Scara3 inhibited osteogenic differentiation and enhanced adipogenic differentiation
To investigate the role of Scara3 in the cell fate switch of BMSCs, we silenced the expression of Scara3 in the primary BMSCs using the Scara3 siRNA. The knockdown efficiency measured by RT-PCR showed the Scara3 siRNA has been successfully transfected in BMSCs during osteogenesis (Figure 2A). Then, BMSCs with a deficiency of Scara3 were cultured in osteogenic differentiation medium or adipogenic differentiation medium. The RT-PCR results showed that the osteogenic markers, such as Bglap, Alpl, Sp7 and Runx2, were significantly downregulated at both the RNA and protein levels in response to osteogenic induction ( Figure 2B-F). The ALP activity also suggested that the osteogenic differentiation ability was impaired by the knockdown of the Scara3 gene ( Figure 2G). The reduced mineralization was consistently validated by the ALP staining and Alizarin red staining ( Figure 2H-J). However, in response to adipogenic induction, BMSCs with Scara3 deficiency showed higher expression of the adipogenic markers, including Fabp4, Pparg, Cebpa and AdipoQ.

| Scara3 regulated autophagy and controlled cell fates via Foxo1
Autophagy plays a crucial role in bone hemostasis. 20,22 To investigate whether Scara3 can regulate autophagy signalling, we firstly ensured that Scara3 was overexpressed in the primary BMSCs both at mRNA with Scara3 plasmid compared to the control BMSCs ( Figure 4G,H).

On contrary, a less number of LC3 punches in BMSCs interfered with
Scara3 siRNA were observed than in control BMSCs ( Figure 4I,J).
It is reported that FOXO1 can stimulate bone formation and restrain adipogenesis. 31-33 FOXO1, as an autophagy inducer, can promote autophagy signalling as well. 34 Then, we explored that if Scara3 was associated with Foxos-mediated effects. We analysed a positive correlation between SCARA3 and FOXO1 in 7858 types of tissues based on GTEx database. Surprisingly, the Pearson correlation was up to 0.47 ( Figure 5A). In accordance with the positive correlation, the overexpression of Scara3 in BMSCs enhanced the expression of FOXO1 ( Figure 5B,C). When compared to control BMSCs, immunofluorescence images displayed that FOXO1 tended to nuclear accumulation in BMSCs with overexpression of Scara3 ( Figure 5D). In consistent with these vitro results, we found that FOXO1 expression was elevated in the 15-month aged mice and OVX mice which were injected with AAVs-scara3 into intra-bone marrow ( Figure 5E,F).

| Injection of AAV-Scara3 into bone marrow cavity alleviated bone loss and MAT accumulation in OVX-induced osteoporotic mice
Our results indicated that Scara3 promoted osteogenesis while All data are expressed as mean ± SD. *P < .05, **P < .01, ***P < .001, Student's t test osteoporosis. 35 Then, these mice were treated with AAVs-Scara3 through intra-bone marrow injection. Immunofluorescence images displayed that Scara3 was significantly overexpressed in femur, especially to the bone surface ( Figure 7A). The BMSCs from mice overexpressing Scara3 also showed a higher level of Scara3 expression than the control group ( Figure 7B,C). The micro-CT images exhibited that the bone obtained from AAV-Scara3 treated mice has a stronger potential for bone formation when compared to those from control mice ( Figure 7D). The quantita- N) was found in OVX mice injected by AAVs-Scara3 ( Figure 7E-H).

| Injection of AAV-Scara3 into bone marrow mitigated bone loss in aged mice
Notably, we revealed a higher expression of Scara3 in young mice than in old mice at mRNA level and protein level ( Figure 8A-C).
In parallel, we also observed the bone phenotypes in aged mice caused by overexpression of Scara3. The 15-month aged mice were obtained through bilaterally intra-bone injection of AAVs

| D ISCUSS I ON
In this study, we revealed that Scara3 regulated the autophagy signal- SCARA3 is reported to play a role in the antioxidant process. 13,15,36 our recent study revealed a higher level of SCARA3 expression of Ad-MSCs in young mice than in aged mice, which was based on the bioinformatics analysis on a public data set. 16  The expression of SCARA3 gene was previously reported to be associated with bone mass in osteoporotic women bone tissue. 19 The downregulation of SCARA3 transcription in osteoporotic bone implied that SCARA3 might be a key modulator to restore bone loss.
Thus, we investigated the role of Scara3 in bone formation in vivo.
[Correction added on 26 July 2021, after first online publication: Figure 8 was a duplicate of Figure 7. It has now been corrected.] tested in BMSCs isolated from femurs of different groups which confirmed that Scara3 was overexpressed in BMSCs. In consistent with our hypothesis, OVX mice with overexpression of Scara3 exhibited an increase of adipocytes and a decrease of osteoblasts, which suggested that Scara3 attenuated bone loss and MAT deposition.
Similarly, in the aged mice, the bone loss was alleviated by Scara3.
Even though various studies found MAT accumulation in aged mice, some studies showed the absence of adipocytes in bone marrow of 12-month wild-type mice and over 1-year TR4 +/+ mice. 45,46 The variance of adipocytes numbers in bone marrow might be related to the different lifestyle of diet, exercise or individuals difference.
In this study, we did not notice obvious changes of adipocytes by AAVs-Scara3 in aged mice, which might be because of the absence of adipocytes in the control aged group. Together, based on the in vitro and in vivo experiments, this study revealed that SCARA3 is a potential target to prevent bone loss and bone adiposity.

ACK N OWLED G EM ENTS
This work was supported by the grant of General Program of the National Natural Science Foundation of China (No.81770877).

CO N FLI C T O F I NTE R E S T
The authors declare that they have no competing interests.

AUTH O R CO NTR I B UTI O N S
Hui Peng designed the study, and Hui Peng and Peng Chen per-

DATA AVA I L A B I L I T Y S TAT E M E N T
The raw data to analyse the genes co-expression are available in the GTEx database (https://www.gtexp ortal.org/). The raw data for Scara3 expression in osteoblasts at different stages are available in the BioGPS database (http://biogps.org/#goto=welcome). Other data that support the findings of this study are available within the article or available from the authors upon request.