Bmal1- and Per2-Mediated Regulation of the Osteogenic Differentiation and Proliferation of Mouse BMSCs by Modulating the Wnt/β-catenin Pathway

Background Bmal1 and Per2 are the core components of the circadian clock genes(CCGs). Bmal1 -/- mice exhibit premature aging, as indicated by hypotrichosis and osteoporosis, with a loss of proliferation ability. The same occurs in Per2 -/- mice, albeit to a less severe degree. However, whether the effects of Bmal1 and Per2 on proliferation and osteogenic differentiation are synergistic or antagonistic remains unclear. Thus, our study aimed to explore the effects and specic mechanism. Materials and methods repress of the target genes β-catenin/TCF/LEF binding β-catenin in the nucleus. recent study found that Rev-erbα overexpression in BMSCs indirectly affected bone formation. These studies provide a clue that Rorα and Rev-erbα may act as mediators in the regulation of the Wnt/β-catenin pathway by Bmal1 or Per2. Our study showed that Bmal1 or Per2 could negatively regulate the expression of Rorα while upregulate the expression of Rev-erbα. The effect could be enhanced by simultaneous inhibition of Bmal1 and Per2 expression. Conversely, Rev-erbα suppressed osteogenic differentiation. Recent studies found that overexpression of Rorα in preosteoblasts promoted bone formation and osteoblast differentiation. In addition, Rorα reduced the destruction of bone tissue in human rheumatoid arthritis by inhibiting osteoclast differentiation. 41 Animal experiments also showed that overexpressing Rorα elevated the expression level of osteogenesis-related proteins such as ALP and OCN. 42.43 Rorα-knockout mice exhibited poor bone mineralization. These ndings are in accordance with our results, indicating that involved in the positive regulation of the osteogenic differentiation of BMSCs. It has been found that RORα expression represses the expression of bone sialoprotein (BSP) and dentin matrix protein (DMP-1) by impacting the downstream Wnt/β-catenin pathway, thus affecting the osteogenic of BMSCs and bone mineralization. Our results also showed that the trend


Introduction
Bone density and bone mass decrease during aging, which increases the risk of fracture and negatively affects the quality of life of adult . 1 Aging-related changes in the bone are derived from bone marrow mesenchymal stem cells (BMSCs), whose proliferation and osteogenic differentiation capacity decreases over time. 2 The roles of circadian clock genes in aging have recently attracted increasing attention. A related study showed that knockout of core circadian clock genes (CCGs) caused signi cant premature aging.
Rescuing gene expression reversed the premature aging changes. 3,4 The CCGs in mammals include brain and muscle Arnt-like protein-1 (Bmal1), period 2 (Per2), circadian locomotor output cycles kaput (Clock), cryptochrome (Cry), retinoid acid receptor-related orphan receptor Rorα and Rev-erbα, which regulate circadian rhythm through an oscillating expression system, establishing a negative feedback network. [5][6][7] In this feedback network, Bmal1-Clock dimer can bind to Per promoter 6 , which interferes with the transcription activity of Bmal1-Clock after translation, while Rorα can activate the transcription of Bmal1 and Rev-erbα can inhibit the transcription of Bmal1 7 . The disruption of the biological rhythms can cause aging. Some research has shown that Bmal1 −/− mice underwent premature aging, resulting in hypotrichosis and osteoporosis. Similar phenomena were found in Per2 −/− mice, albeit to a less severe degree. 8 Other studies also found that Bmal1 or Per2 can regulate the proliferation and osteogenic differentiation ability of BMSCs. [9][10][11] However, a consensus regarding whether their effects are synergistic or antagonistic has not yet been reached, 12 which makes our study particularly signi cant.
The Wnt/β-catenin pathway 13 plays a pivotal role in the regulation of proliferation and osteogenic differentiation of BMSCs by inducing the formation of osteoblasts and repressing the differentiation of osteoclasts. 14 Binding sites of Bmal1 are found in the coding sequence of many genes in the Wnt/βcatenin pathway, such as Wnt10a, β-catenin, and T-cell-speci c transcription factor-3 (TCF3). [15][16][17] However, whether there is some interaction between CCGs and the Wnt/β-catenin pathway remains poorly understood. Furthermore, Rorα and Rev-erbα not only constitute a closed loop with Bmal1 through mutual regulation but also regulate the downstream target genes of the Wnt/β-catenin pathway. 16,17 Therefore, Rorα and Rev-erbα may be critical factors in the regulation of the Wnt/β-catenin pathway by Bmal1 and Per2. In summary, we proposed the following hypothesis: Bmal1 and Per2 could regulate the aging of BMSCs by altering cell proliferation and osteogenic differentiation through the Wnt/β-catenin pathway. In addition, there might be some interaction between Bmal1 and Per2.
In this study, we altered the expression of Bmal1 and/or Per2 by virus transfection, then observed the expression of osteogenic marker genes such as alkaline phosphatase (Alp), osteocalcin (Ocn), and runtrelated transcription factor 2 (Runx2). We also detected the activity of ALP, the proliferation ability of BMSCs and the expression changes in Rorα and Rev-erbα as well as the core molecules of the Wnt/βcatenin pathway, including Wingless Int1 3a (Wnt3a ), Cell-myc1 (C-myc ), Axin2, β-catenin, T-cell-speci c transcription factor 1 (TCF1), and phosphorylated glycogen synthase kinase-3β (P-GSK-3 ), after altering the expression of Bmal1 and/or Per2. Finally, we identi ed the target genes and pathways of Bmal1 and Per2 by chromatin immunoprecipitation and high-throughput sequencing (ChIP-Seq) to verify the role of the Wnt/ β-catenin pathway and explore other possible mechanisms. Our research aimed to nd new targets for the treatment and prevention of bone senescence or osteoporosis.

Cell culture
BMSCs were obtained from four-week-old C57BL/6 male mice. After passaging the cells twice, the thirdgeneration cells were cultured in α-MEM (HyClone, USA) with 10% fetal bovine serum, 100 U/ml penicillin and 100 mg/ml streptomycin and incubated at 37°C with an atmosphere of 95% relative humidity and 5% CO 2 . When the cells reached approximately 80% con uence, 0.25% trypsin in 0.01% ethylenediaminetetraacetic acid (EDTA) was used to detach the cells, which were passaged to the next generation. Mesenchymal stem cells (MSCs) were identi ed by immunohistochemistry as CD29(+), Stro-1(+), CD45(-) and CD34(-) cells. In our study, we conducted experiments with fth-generation cells.
Isolated BMSCs were seeded in 24-well plates at 5×10 5 cells per well in basic media. Bmal1 shRNA lentivirus (Arntl shRNA, purchased from Hanheng Bio, Shanghai, China) or Bmal1 empty lentivirus (PHBLV-U6-ZsGreen-Puro) was added to the cells at a multiplicity of infection (MOI) of 20 after the cell con uence reached 70%. The medium was replaced with α-MEM medium containing 2 µg/ml puromycin after another 24 hours. Green uorescence was observed under an inverted uorescence microscope (Olympus IX70, Japan). The cells were passaged with transfection e ciency over 80% and cryopreserved for the next experiments.
The 9 groups aforementioned were created through different transfection measures by adding both shRNA and empty vectors at an MOI of 20 and replacing the medium with basic medium after 24 hours of incubation. Red uorescence was observed under an inverted uorescence microscope 48 hours after transfection. Western blot analysis was also applied to calculate the e ciency of infection. In these analyses, anti-BMAL1 (diluted to 1:500), anti-PER2 (diluted to 1:500) and F-actin (diluted to 1:1000) were utilized as the primary antibodies, while goat anti-rabbit IgG (diluted to 1:1000) was utilized as the secondary antibody.

Methyl thiazolyl tetrazolium (MTT)-based detection of proliferative activity 20
After transfection with the corresponding lentivirus or adenovirus, cells in the 9 groups were plated and cultured in 96-well plates. Ten microliters of PBS solution containing 5 mg/ml MTT was added to the wells of each group and allowed to incubate for 24 hours. Six hours after incubation with MTT, 150 µl of dimethyl sulfoxide (DMSO) was added to each well, and the plates were vibrated on a table concentrator for 10 minutes. We measured the optical density at 490 nm (OD490) at 24, 48 and 72 hours.

Flow cytometry detection of BMSC proliferation
The cells from passage 5 in 9 groups were detached from the 6-well plates with trypsin and ethylenediaminetetraacetic acid. The samples were washed twice with PBS and centrifuged at 1000 rpm for 5 minutes. Then, the samples were placed in 70% alcohol and incubated at -20°C for at least 2 hours for xation. Next, they were centrifuged again to remove the alcohol and washed with PBS. Then, while shielded from light, they were stained with propidium iodide (PI), incubated in a water bath at 37°C for 30 minutes and then incubated with RNase at 4°C. The cell density was adjusted to approximately 1 10 6 cells/ml for detection. The following formula was used to calculate the S-phase fraction (SPF) and DNA After viral transfection, the cells in the 9 groups were cultured in 6-well plates with osteogenic induction medium (OS medium), which contained 12 mmol/L β-glycerophosphate, 1*10 − 8 mol/L dexamethasone, 0.05 mmol/L vitamin C and 10% fetal bovine serum. Approximately 1*10 5 cells were plated in each well.
The medium was changed after 2 days. The expression of osteogenesis-related genes, such as ALP, OCN, and Runx2, was detected after 7 days and 14 days. Total RNA was extracted with the Simply P total RNA extraction kit. To ensure high purity, assessment of the 20-µl reaction mixture products was performed with an ABI PRISM 7300 real-time system with glyceraldehyde-3-phosphate dehydrogenase (GAPDH) Factin as an internal reference. (Table 1)  To extract protein from passage 5 BMSCs cultured for 7 days and for 14 days, the cells were washed in ice-cold PBS 3 times and su ciently lysed in radioimmunoprecipitation assay (RIPA) lysis buffer. After centrifugation at 12000 rpm for 15 minutes at 4°C the liquid supernatant was collected for the protein concentration assessment using a bicinchoninic acid (BCA) protein assay kit. Sodium dodecyl sulfonatepolyacrylate gel electrophoresis (SDS-PAGE) was performed to separate the proteins, which were transferred to polyvinylidene uoride (PVDF) membranes. Then, the membranes were blocked and incubated with primary antibodies and then the secondary antibody. The chemiluminescence method was used to detect the immunoreactive proteins, and the band intensities were analyzed by a gel documentation system.

Analysis of ALP activity.
After BMSCs were transfected with the corresponding virus and allowed to undergo osteogenic induction for 7 days (with medium replacement every two days), we harvested and lysed the cells with 10 mM Tris-HCl and 0.1% Triton X-100 (pH 7.4). The supernatant obtained through centrifugation was stained according to the instructions of the ALP Detection Kit. ALP activity, expressed as units/L protein, was calculated by measuring the absorbance at 510 nm on a spectrophotometer.

Immunohistochemistry
The cells were lysed and adjusted to approximately 2 10 4 cells/ml in complete medium. The cells attached on the 12-well plates for 30 minutes were cultured in 37℃ with 5% CO 2 . The cell discs were washed in PBS twice and bathed in 4% Paraformaldehyde 20 minutes. Then the discs were soak in 0.5% Triton X-100 for 20 minutes and washed by PBS for three times. And then they were bathed in 3%H 2 O 2 for 15minutes. The membranes were incubated with primary antibodies and then the secondary antibody. Generally DAB stain, haematoxylin slightly stain, alcohol dehydration, xylene transparent and neutral balata xation.

Alizarin red staining
The BMSCs were cultured in OS medium after 7 days and 14 days and the cell density was adjusted to approximately 2 10 4 cells/ml in complete medium. The cells attached on the 24-well plates for 30 minutes were cultured in 37℃ with 5% CO 2 . The cell discs were washed in PBS twice and bathed in 4% Paraformaldehyde 20 minutes. Then the discs were stained by alizarin red for 30 minutes and washed by distilled water.

ChIP-Seq and data analysis
BMSCs were cross-linked with 1% formaldehyde at room temperature for 10 minutes, and the reaction was terminated by the addition of glycine. The collected cells were lysed (50 mM Tris-HCl, pH 7.4; 10 mM × × EDTA; 1% SDS; and protease inhibitor) and sonicated. Anti-BMAL1 and anti-PER2 antibodies were used to precipitate the corresponding DNA fragments. Bound DNA was incubated at 65°C overnight to reverse the cross-links. After the DNA purity and concentration was measured, the target gene was ampli ed by q-PCR. We performed high-throughput sequencing on an Illumina HiSeq platform and analyzed the relevant data.

Statistical analysis
The data are presented as the mean ± standard error of the mean. Signi cance was determined using one-way analysis of variance and the Student-Newman-Keuls (SNK) test for multiple comparisons with SPSS 19.0, with P < 0.05 as the cutoff.

Results
3.1 Altered expression of Bmal1 and/or Per2 at the mRNA and protein levels after transfection with viral vectors.
Green and red uorescence protein expression were detected by uorescence microcopy, con rming the knockdown of Bmal1 and the knockdown/overexpression of Per2 ( Figure 1A). The protein expression level was in accordance with gene expression level for each gene ( Figure 1B, C).

The growth curve
The growth curve( Figure 2A)according to the OD490 values obtained through the MTT assay indicated the proliferation rates of the different groups. The proliferative activity of cells overexpressing Per2 increased compared to that of control cells, and accordingly, the survival rate of this group (group 5) was the highest among all groups. When we inhibited Bmal1 (group 2), Per2 (group 4) or both (group 7), the proliferation rate decreased signi cantly compared with that in the groups transfected with an empty viral vector (groups 3, 6, and 9).
3.3 The proliferation rates of different groups evaluated by ow cytometry.
The percentage of cells distributed in the G0 phase, G1 phase, G2 phase, S phase and M phase re ects the cell activity to some degree. The ow cytometry results( Figure 2B) showed that cells overexpressing Per2 had the greatest proliferative activity, with the highest SPF value and PI staining rate. Cells with knockdown of Bmal1, Per2 or both exhibited a lower proliferation rate than their matched control groups.
The difference in the expression of Wnt-3a between the Bmal1 and Per2 knockdown group and the matched empty vector group was signi cant, speci cally showing that suppressing Bmal1 and Per2 simultaneously increased the levels of Wnt-3a, c-myc1 and axin2. Knockdown of Bmal1 alone increased the expression of c-myc1 and axin2, and the effect on axin2 was increased with combined knockdown of Per2 ( Figure 3A).

Expression level of Wnt/β-catenin signaling pathway-related proteins.
It was concluded that the levels of β-catenin, TCF-1, and P-GSK-3β ( Figure 3B)increased under four different conditions in our experiment: Bmal1 knockdown, Per2 knockdown, Bmal1 knockdown/Per2 knockdown and Bmal1 knockdown/Per2 overexpression. The expression level in the Bmal1 knockdown/Per2 overexpression group was higher than that in the Per2 overexpression group, and the difference was statistically signi cant. The probable mechanism is that the expression of Per2 is promoted when Bmal1 is suppressed. The effects of Per2 overexpression were strengthened in the combination group, leading to increased expression of proteins related to the Wnt/β-catenin signaling pathway.
3.6 Changes in osteogenic marker gene expression in each group after osteogenic induction.
Our previous research explored the effect of Bmal1 knockdown, Per2 knockdown and Bmal1 knockdown/Per2 knockdown on the osteogenic differentiation ability of BMSCs; we found that Bmal1 and Per2 negatively and synergistically regulated this process. 18 Therefore, we designed an experiment to assess the osteogenic differentiation capability of BMSCs in different groups, including Bmal1 knockdown, Per2 overexpression, and Bmal1 knockdown/Per2 overexpression groups, to verify the previous results.
3.6.1 ALP activity in each group after 7 days of osteogenic induction.
In the Bmal1 knockdown group, ALP activity, an early osteogenic differentiation marker, was enhanced compared with that in the empty control group ( Figure 4D) but declined in the Per2 overexpression group.
The activity of ALP in the Bmal1 knockdown/Per2 overexpression group was also elevated and was even higher than that in the Bmal1 knockdown group. There were no signi cant differences between the control groups (group 1, group 3, group 6, and group 9), which shows that viral transfection had no impact on ALP activity.
3.6.2 Expression of osteogenic differentiation-related genes.
The mRNA expression levels of Alp, Runx2 and Ocn in the Bmal1 knockdown group exceeded those in the control group but were lower than those in the Bmal1 knockdown/Per2 overexpression group. Bmal1 knockdown/Per2 knockdown had signi cantly stronger effects than empty control transfection after 7 days of osteogenic induction ( Figure 4A-C). The differences in Ocn expression in each group detected 14 days later were in accordance with the result observed after 7 days of osteogenic induction ( Figure 5A). The protein expression levels were still in accordance with the mRNA expression levels at these time points.
3.7 Changes in Rorα and Rev-erbα expression after alteration of Bmal1 and Per2 expression in BMSCs.

The gene expression of Rorα
In the four experimental groups, Bmal1 knockdown, Per2 knockdown, Bmal1 knockdown/Per2 knockdown and Bmal1 knockdown/Per2 overexpression, the expression of Rorα was remarkably higher than that in the corresponding empty control groups; in the Per2 overexpression group, the outcome was the opposite ( Figure 5B). Inhibiting both Bmal1 and Per2 had a greater effect than inhibiting Per2 alone. To our surprise, the effect in the Bmal1 knockdown/Per2 knockdown and Bmal1 knockdown/Per2 overexpression groups was greater than that in the Bmal1 knockdown group. There were no signi cant differences among the blank control group and the other control groups transfected with empty vectors, indicating that viral transfection exerted no impact on the expression of Rorα. The protein and gene expression trends of Rorα were consistent.

The gene expression of Rev-erbα
In the four experimental groups, Bmal1 knockdown, Per2 knockdown, Bmal1 knockdown/Per2 knockdown and Bmal1 knockdown/Per2 overexpression, the expression of Rev-erbα decreased compared with that in the blank control group, but the difference between the Per2 knockdown group and the control group was not signi cant ( Figure 5C). In addition, knockdown of both Bmal1 and Per2 led to a much lower expression level of Rev-erbα than did knockdown of Bmal1 or Per2 alone, verifying their synergistic effects. Overexpression of Per2 increased the expression of Rev-erbα, which contrasted with the effect of combined Bmal1 knockdown/Per2 overexpression. The protein expression trend of Rev-erbα in these groups was in accordance the mRNA expression trend. (Figure 5D,5E) 3.8 Immunohistochemistry blank control, Bmal1 knockdown, Bmal1 empty vector control, Per2 knockdown, Per2 overexpression, Per2 empty vector control, Bmal1 knockdown and Per2 knockdown, Bmal1 knockdown and Per2 overexpression, Bmal1 empty vector and Per2 empty vector control.
The ALP activity of Bmal1 knockdown group was enhanced compared with the control group, which indicated that Bmal1 had negative effect on osteogenic differentiation of BMSCs. The ALP activity decreased in Per2 overexpression group, which demonstrated Per2 could inhibit the osteogenic differentiation of BMSCs. Whether Per2 was overexpressed or suppressed, Bmal1 knockdown would facilitate bone mineralization.

Alizarin red staining
The results of alizarin red staining shows that the mineralized nodules were increased in Bmal1 knockdown and Per2 knockdown group compared with blank control. While inhibiting Bmal1 and Per2 simultaneously would enhance the positive effect. Furthermore, both Bmal1 and Per2 were notably enriched in the mTOR signaling pathway, 22 Hippo signaling pathway and ubiquitin-mediated proteolysis pathway (P<0.05, Figure 6C, 6D), which means that the speci c regulatory mechanisms may be coming into view.

Discussion
Osteoporosis due to aging has become a common bone and metabolic disease. Osteoporosis is mainly caused by disturbed proliferation and osteogenic differentiation of BMSCs. In our previous study, Bmal1 and Per2, the CCGs, were found to affect proliferation and osteogenic differentiation.2 2 As a supplement, in the present study we found that Bmal1 or Per2 knockdown resulted in G1-phase cell cycle arrest, giving rise to the loss of proliferation ability in BMSCs, which was the most well-established aging feature. During this process, DNA was not produced and the cell did not proceed to the S phase, while suppressing Bmal1 and Per2 simultaneously generated stronger effects than inhibiting one of them separately. The results supported our hypothesis that Bmal1 and Per2 engaged in cross talk during cell proliferation.
Some research demonstrated that Bmal1 −/− mice had increased osteogenesis and bone-formation. 23 And some indicated that Bmal1 negatively affected mineral apposition rate of mice femur. 24 In consistence, our results found that the osteogenic differentiation of BMSCs could be enhanced by inhibiting Bmal1. In our previous research, the capability of osteogenesis was elevated after inhibition of Per2. However, related research or projects in regard to the effect of Per2 on osteogenesis have not been reported. In this study, the expression levels of Alp, Runx2, and Ocn, which were important markers in osteogenic differentiation of BMSCs, decreased after Per2 overexpression. Our work further validated that Per2 performed a negative effect on osteogenesis.
The oscillatory mechanisms underlying the CCGs has been unraveled. CLOCK and BMAL1, as a heterodimeric complex of two transcriptional activators, could bind to E-box enhancer elements and thereby activated the expression of Per1, Per2. 25 Some research demonstrated Per2, Cry1 mutant mice displayed an altered expression of genes regulated by BMAL1. 26 It also has been reported that Per2 could bind to BMAL1 and interact with speci c sites of BMAL1-CLOCK heterodimer. In the present study, Bmal1 played a synergistic effect with Per2 in the regulation of BMSCs might be explained by the mechanism mentioned above. Our previous study showed that the expression of Bmal1 changed with inhibition of Wnt/β-Catenin pathway. It also demonstrated that Wnt/β-Catenin pathway played a vital role in proliferation and osteogenesis differentiation of BMSCs. Recent research showed that Bmal1 had a synergistic effect on the aging of BMSCs directly or indirectly through the Wnt/β-Catenin pathway. 22 Furthermore, we conducted ChIP-Seq and KEGG analysis as a powerful tool to explore whether there were targets of Bmal1 and Per2 in the Wnt/β-catenin pathway. Surprisingly, enrichment of Bmal1 or Per2 in the Wnt/β-catenin pathway was not signi cantly high, while enrichment of Hippo and mTOR signaling pathway and ubiquitin-proteasome proteolytic system was high, hence we hypothesized that the regulatory effects of Bmal1 and Per2 were achieved through multiple pathways crosslinked with Wnt/βcatenin pathway. As vital downstream molecules of the Hippo signaling pathway 29-31 , YAP/TAZ isolate βcatenin from the destruction complex (APC, Axin, Gsk-3β, and β-catenin) and escape from the cytoplasm to the nucleus, which enhances the expression of downstream target genes and activates the Wnt/βcatenin pathway. 32.33 Some studies have shown that disturbing mTOR suppressed the growth of normal cells and tumor cells induced by the Wnt/β-catenin signaling pathway. Changes in mTOR in uenced the expression of the Wnt/β-catenin signaling pathway to some degree. 34191919191919191919 The ubiquitinproteasome proteolytic system plays an important role in the regulation of cell growth, proliferation, aging, and apoptosis. β-Catenin is degraded through its reaction with the E3 ubiquitin enzyme complex, 35 which blocks the Wnt/β-catenin signaling pathway. In summary, the Wnt/β-catenin pathway crosstalks with the Hippo signaling pathway, mTOR and the ubiquitin-proteasome proteolytic system. Therefore, although direct enrichment of Bmal1 and Per2 in the Wnt/β-catenin signaling pathway was not found, we still could speculate that Bmal1 and Per2 regulated the Wnt/β-catenin pathway through other pathways, such as the Hippo, mTOR, and ubiquitin-proteasome pathways.
On the other hand, CLOCK-BMAL1/PER-CRY constitute a core feedback pathway, and the CLOCK-BMAL1 heterodimer activates the transcription of Rorα and Rev-erbα. RORα activates the transcription of Bmal1, while REV-ERBα downregulates BMAL1 through glucose synthesis kinase 3 (GSK-3β)-mediated phosphorylation. 36 Thus, BMAL1, RORα and REV-ERBα constitute a closed loop that controls the stability of the CCG system. GSK-3β, a critical component of the Wnt/β-catenin pathway, plays a negative role in the activation of this pathway by phosphorylating and promoting the degradation of β-catenin. 37 Phosphorylated RORα can also repress the expression of the downstream target genes of βcatenin/TCF/LEF by binding to β-catenin in the nucleus. 38 A recent study found that Rev-erbα overexpression in BMSCs indirectly affected bone formation. 39 These studies provide a clue that Rorα and Rev-erbα may act as mediators in the regulation of the Wnt/β-catenin pathway by Bmal1 or Per2. Our study showed that Bmal1 or Per2 could negatively regulate the expression of Rorα while upregulate the expression of Rev-erbα. The effect could be enhanced by simultaneous inhibition of Bmal1 and Per2 expression. Conversely, Rev-erbα suppressed osteogenic differentiation. Recent studies found that overexpression of Rorα in preosteoblasts promoted bone formation and osteoblast differentiation. 40 In addition, Rorα reduced the destruction of bone tissue in human rheumatoid arthritis by inhibiting osteoclast differentiation. 41 Animal experiments also showed that overexpressing Rorα elevated the expression level of osteogenesis-related proteins such as ALP and OCN. 42.43 Rorα-knockout mice exhibited poor bone mineralization. 44.45 These ndings are in accordance with our results, indicating that Rorα is involved in the positive regulation of the osteogenic differentiation of BMSCs. It has been found that decreasing RORα expression represses the expression of bone sialoprotein (BSP) and dentin matrix protein 1 (DMP-1) by impacting the downstream Wnt/β-catenin pathway, thus affecting the osteogenic differentiation of BMSCs and bone mineralization. 45 Our results also showed that the expression trend of Rorα was in accordance with the trend of Wnt/β-catenin pathway activation after alteration of Bmal1 and Per2. Therefore, Bmal1 and Per2 might negatively regulate the osteogenic differentiation of BMSCs through repression of the Wnt/β-catenin pathway by downregulating Rorα. Many studies have found that overexpression of Rorα can suppress the expression of β-catenin in the nucleus and inhibit tumor growth, cancer cell proliferation and invasion by negatively regulating Wnt/β-catenin signaling pathway activity. 46.47 This seems to contradict our hypothesis that Rorα can promote osteogenic differentiation by activating the Wnt/β-catenin pathway. However, the understanding of the complexity of the Wnt/β-catenin signaling network is ever growing, as it has multiple ligands and receptors, which may have cell-, tissue-, or stage-speci c effects. 48.49 Therefore, further research is needed to clarify the relationship between Rorα and the Wnt/β-catenin pathway. 50 Our previous study found that Rev-erbα could play a negative role in BMSCs during the late stage of osteogenesis by inhibiting the secretion of bone sialoprotein (BSP).
Overexpression of Rev-erbα inhibited the proliferation of BMSCs, and activating the Wnt/β-catenin pathway could partially reverse the inhibition of cell proliferation caused by Rev-erbα overexpression. This also supports that Rev-erbα participates in the negative regulation of the osteogenic differentiation and proliferation of BMSCs and has some interaction with the Wnt/β-catenin pathway. Our results showed that after alteration of Bmal1 and Per2 expression, the expression trend of Rev-erbα was opposite to that of the Wnt/β-catenin pathway activators. However, this conclusion needs to be veri ed by observing the changes in Wnt/β-catenin signaling molecules after alteration of Rorα and Rev-erbα expression, which will be the goal of our follow-up research.   Figure 1 The

Figure 4
The mRNA and protein expression of ALP (Alkaline phosphatase), Runx2 (Runt related transcription factor2) and OCN (Osteocalcin) in BMSCs 7 days after osteogenic induction. A, ALP mRNA expression and intensity; B, OCN mRNA expression and intensity; C, Runx2 mRNA expression and intensity. D, The result of Western blotting. GAPDH (Glyceraldehyde-3-phosphate dehydrogenase) was the internal reference. Densitometry was utilized to quantify the protein expression levels, which were determined by the ratio of ALP, RUNX2 and OCN to GAPDH. E, The activity of ALP in 7 groups after 7 days of osteogenic induction. Data are presented as the means±SD (n=3). Asterisks indicate the signi cant difference of a group compared to the level of the control groups: group 1, 3, 6 and 9 (**P<0.01, and ***P<0.001). Pound signs represent the signi cant difference between two groups (##P<0.01 and ###P<0.001).

Figure 5
A, mRNA and protein expression of OCN (Osteocalcin) in 7 groups after 14 days of osteogenic induction. B,C, The mRNA and protein expression of Rorα (Retinoid acid receptor related orphan receptor α) and Reverbα after viral transfection of the BMSCs. The Western blotting results of Rorα and Rev-erbα expression among all groups showing the same trend as the expression of mRNA. F-actin was the internal reference.
D,E, Western blot results. Densitometry was utilized to quantify the protein expression levels, which were determined by the ratio of RORα and REV-ERBα, respectively, to F-actin. Data are presented as the means±SD (n=3). Asterisks indicate the statistical difference of a group compared to the level in the control groups: group 1, 3, 6 and 9 (*P<0.05, **P<0.01, ***P<0.001). Pound signs represent the statistical difference between two groups (#P<0.05, ##P<0.01, ###P<0.001).

Figure 6
The immunohistochemistry results of ALP. A-I represent 9 groups respectively.

Figure 7
The alizarin red staining results after osteogenic induction 14 days . A-I represent 9 groups respectively.

Figure 8
The alizarin red staining results after osteogenic induction 21 days. A-I represent 9 groups respectively.