METTL3 induces bone marrow mesenchymal stem cells osteogenic differentiation and migration through facilitating M1 macrophage differentiation

Abstract

Despite the crucial role of m⁶A methyltransferase METTL3 in multiple diseases onset and progression, there are still lacking hard evidence proving that METTL3 could affect macrophage polarization in the stage of bone repair. Here, we aimed to explore the potential involvement of METTL3 in bone repair through modulating macrophage polarization and decipher the underlying cellular/molecular mechanisms. Here we treated RAW 264.7 cells and BM-derived primary macrophages (BMDM) with lipopolysaccharide (LPS) to induce M1 differentiation. METTL3 expression was upregulated in pro-inflammatory macrophages (M1) as compared with macrophages (M0). And overexpression of METTL3 promoted the expression of IL-6 and iNOS secretion by M1 macrophage. In the coculture condition, M1 macrophages with forced expression of METTL3 significantly enhanced migration ability of BMSCs, and also remarkably facilitated osteogenesis ability of BMSCs; the opposite was true when expression of METTL3 was knockdown. In addition, the m⁶A-RIP microarray suggested that METTL3 silencing significantly reduce the m⁶A modification of DUSP14, HDAC5 and Nfam1. Furthermore, the findings showed that expression of HADC5 was downregulated in M1 macrophages with METTL3 knockdown, while the DUSP14 expression had slight change and Nfam1 expression was very low. In contrast, METTL3 overexpression promoted HDAC5 expression, indicating that HDAC5 is the critical target gene of METTL3. Under such a theme, we proposed that METTL3 overexpression might be a new approach of replacement therapy for the treatment of bone repair.

Introduction

Bone fracture healing is a multiply processes including three steps: bone regeneration, porotic stage and remodeling. Bone regeneration begins with the inflammatory reactions of macrophages which interacting with skeletal system via the regulation of bone marrow mesenchymal stem cells (BMSCs) function [1]. Excepting the promotion of inflammatory response occurrence and the local microorganisms removal, accumulating evidence has proved that macrophages make great contributions to bone regeneration due to the mediation of BMSCs osteogenic differentiation and migration ability [2,3]. Macrophages have at least two subtypes containing classically activated/inflammatory (M1) and alternatively activated regenerative subpopulations (M2), which regulated different cytokines [4,5]. M1 macrophage exhibits pro-inflammatory functions and secretes inflammatory cytokines such as IL-6, iNOS, TNF-α, IL-1α and IL-1β, while M2 macrophage expresses high anti-inflammatory cytokines including ARG1, IL-10 and so on [6]. Nevertheless, whether the ability to promote bone repair is attributed to M1 or M2 remains a matter of debate. In spite of this, several previous studies have shown that M1 but not M2 macrophage promotes BMSCs osteogenic differentiation and migration. M1 macrophage involved in bone repair is due to the syntheses of cytokines OSM, TNF-α and IL-6, which have been found to positively affect BMSCs osteogenic differentiation [7-9]. Moreover, there are additional evidence further support that M1 rather than M2 macrophage is more important for the osteogenesis of BMSCs [10,11].

Methyltransferase-like 3 (METTL3) is the m⁶A methyltransferase which binds to METTL14 and wilms tumour 1-associated protein (WTAP) to form a complex catalyzing the m⁶A modification, the most abundant epitranscriptomic modification in eukaryotic mRNAs [12]. M⁶A modification plays vital role in a variety of human disorders and diseases, such as tumor development, kidney injury, stem cell differentiation and myocardial ischemia [13-16]. Noteworthily, our previous study has shown that silence of METTL3 decreased m⁶A methylation levels and inhibited osteogenic differentiation of BMSCs and reduced bone mass [17]. However, the impact of macrophage polarization mediated by m⁶A modification on bone regeneration remains incompletely unclear.

Here, we have investigated that M1 macrophage modulated by METTL3 plays a critical role in BMSCs osteogenic differentiation and migration by regulating the expression of its critical target histone deacetylase 5 (HDAC5) through m⁶A-based post-transcriptional control.

Methods

Cell culturing and treatment

The mouse macrophage cell line RAW 264.7 was obtained from American Type Culture Collection (ATCC). The RAW 264.7 cell line was cultured in RPMI 1640 medium (Cat#C22400500BT, GIBCO, Suzhou, China) supplemented with 10% fetal bovine serum (Biological Industries, Israel) and 200U of penicillin/streptomycin (Cat#SV30010, HyClone, California, United States). Macrophage activation was performed using 100 ng/ml lipopolysaccharide (LPS) (Cat. No. L8880, Solarbio). BMSCs from C57BL/6 mice (MUBMX-01001) were purchased from Cyagen (Guangzhou, China) and cultured in BMSC complete medium (MUBMX-90011; Cyagen). For BMDMs, BMSCs were treated with 25 ng/ml M-SCF (Cat#426 ML; R&D Systems, USA) for 7 days, and then incubated with 100 ng/ml LPS for 24 h. All cell lines were maintained in an incubator at 37°C in an atmosphere containing 5% CO₂.

Quantitative real-time polymerase chain reaction (qRT-PCR)

QRT-PCR analysis was performed as described previously. Briefly, total RNA was extracted using TRIzol reagent (Life technologies Corporation) followed the manufacturer’s protocol. 500 ng was reverse transcribed to 10 μL cDNA using High Capacity cDNA Reverse Transcription Kit (Cat#00676299; Thermo Fisher Scientific, Waltham, USA). Detection were performed using 7500HT Fast Real-Time PCR System (Applied Biosystems) with SYBR Green PCR master Mix (Cat#31598800; Roche). Gene expression was normalized to endogenous GAPDH mRNA. The primers sequences used as follow: Forward (F)/Reverse (R) primer sequence (5’-3’); IL-6-F: CTGCAAGAGACTTCCATCCAG; IL-6-R: AGTGGTATAGACAGGTCTGTTGG; iNOS-F: CAGCGGAGTGACGGCAAACAT; iNOS-R: GCAAGACCAGAGGCAGCACATC; METTL3-F: CTTTCTACCCCATCTTGAGTG; METTL3-R: CCAACCTTCCGTAGTGATAGTC; Dusp 14-F: CCTTTCCTGCTTTCTTTCTG; Dusp 14-R: GCCCTTGATCTTCCTTAAAC; HDAC5-F: CAAGGCCTTGTGCATGCTGGGCTGG; HDAC5-R: CTGCTCCCGTAGCGCAGGGTCCATG; GAPDH-F: ACTGAGGACCAGGTTGTC; GAPDH-R: TGCTGTAGCCGTATTCATTG.

Western blot

Western blot analysis was performed as described previously. Briefly, washed the cells with PBS, and then lysed on ice for 1 h using RIPA buffer (Beyotime Biotechnology) followed by centrifugation at 13,500 g at 4°C for 15 min. Protein sample (µg) was separated on a polyacrylamide gel, transferred onto a nitrocellulose membrane. The membrane was blocked with 5% fat-free dry milk at room temperature for 1 h. Next, the membrane was incubated with rabbit anti-iNOS antibody (1:500, Cat#ab15323, Abcan), rabbit anti-METTL3 antibody (1:1000, Cat#A8370, ABclonal, Massachusetts, USA) a mouse anti-GAPDH antibody (1:500, Cat#abs830030, Absin) at 4°C for 24 h. All antibodies were diluted in PBS.

Cell transfection

Cells were transfected with siRNA using Lipofectamine RNA iMAX Reagent (MAN0007825, Invitrogen, Carlsbad, California, USA) according to the manufacturer’s instructions for 24 h. The sequences of siRNAs used are METTL3: sense 5’-GCUGCACUUCAGACGAAUUTT-3’ and antisense 5’-AAUUCGUCUGAAGUGCAGCTT-3’. Using METTL3-carrying plasmid which constructed by Cyagen to elevated the expression of METTL3 in cells. Cells were transfected with 1 μg METTL3 plasmid using X-tremeGENE HP DNA Transfection Reagent (Cat#26540900, Roche, Basel, Switzerland) for 48 h.

BMSC migration analysis

BMSC migration assay was performed with a 24-well transwell chamber (porous polycarbonate membrane with a pore size of 8 μm, Corning Costar). RAW 264.7 cells were cultured in the lower chambers and transfected with siRNA for 24 h and plasmid for 48 h to allow M1 differentiation. BMSCs were seeded in the upper chambers. Co-cultured for 24 h. The membranes were stained with 0.1% crystal violet (Beyotime Biotechnology, China). Six fields were randomly selected and counted using a light microscopy (ECLIPSE TS100, Nikon).

Osteogenic differentiation and Alizarin Red S (ARS) staining

Induced BMSCs into osteoblasts was performed as described previously. Briefly, BMSCs were seeded onto a 6-well plate coated with gelatin (GLT-11301; Cyagen) at a cell density of 2×10⁴ cells/cm². When the cell confluence reached 60-70%, 2 mL osteogenic differentiation medium (MUBMX-90021; Cyagen) was carefully added to the plate. Differentiation was induced for 7 days, and fresh culture medium was replaced every 3 days. BMSCs were fixed with 4% formaldehyde at RT for 30 min. Cells were washed with PBS three times and then stained with 40 mM Alizarin Red S solution (S0141; Cyagen) for 5 min. Six fields were randomly selected using a light microscopy and calculated using ImageJ software (National Institutes of Health, NIH, USA).

Statistical analysis

Data are expressed as mean ± SEM. Statistical analysis was performed using GraphPad Prism7 software and analyzed with Student’s t-test (two-tailed). All experiments were independently repeated at least three times. *P < 0.05; **P < 0.01; ***P < 0.001.

Results

Impact of METTL3 in macrophages polarization

To determine the effects of METTL3 mediated m⁶A modification in the polarization towards M1 macrophages in bone regeneration, we treated a macrophage cell line RAW 264.7 (M0 macrophage) with lipopolysaccharide (LPS) at 100 ng/μl to induce M1 macrophage (Figure 1A). QRT-PCR and western blot results showed the increased expression of M1 macrophage makers iNOS and IL-6, which indicating the M1 differentiation of the RAW 264.7 cells (Figure 1B, 1C and Supplementary Figure 4). We next explored the expression level of METTL3 in M0 and M1 macrophage respectively. Western blot analysis showed that METTL3 protein level was significantly higher in the M1 macrophage compared to M0 macrophage (Figure 1D and Supplementary Figure 4). To further investigate the role of METTL3 in macrophage differentiation, we conducted loss-of-function studies in RAW 264.7 cells using METTL3 siRNA. Knockdown efficiency of METTL3 mRNA and protein were detected by qRT-PCR and western blot analysis (Figure 1E and Supplementary Figure 4). Consistently, METTL3 silence decreased iNOS and IL-6 mRNA expression level in the presence of LPS, indicating that METTL3 knockdown inhibits the M1 macrophage differentiation (Figure 1F).

Figure 1.

METTL3 play a critical role in M1 macrophage polarization of RAW 264.7 cells. A. The schematic diagram presenting the treatment procedure for M1 macrophage polarization of RAW 264.7 cells. B. qRT-PCR results showing the mRNA expression level of M1 macrophage makers IL-6 and iNOS in M0 and M1 macrophages. C. Western blot showing the protein expression level of iNOS. D. METTL3 protein expression increased markedly in M1 macrophages compared to the M0 macrophage. E. The efficiency of METTL3 knockdown in RAW 264.7 cells confirmed by qRT-PCR and Western blot. F. qRT-PCR showing that METTL3 silencing decreased the mRNA level of IL-6 and iNOS in M1 macrophages. Data are analyzed with Student’s t-test (two-tailed) and expressed as mean ± SEM. ***P < 0.001.

We then performed gain-of-function studies to confirm whether METTL3 overexpression develop M1 macrophage differentiation. RAW 264.7 cells were transfected with vector and METTL3 plasmid for 24 h and then treated with LPS (100 ng/μl) for 24 h. qRT-PCR and western blot analysis tested the transfection efficiency (Figure 2A and Supplementary Figure 4). As expected, forced expression METTL3 promoted the expression of M1 makers iNOS and IL-6 obviously (Figure 2B).

Figure 2.

Overexpression of METTL3 promotes the polarization of M1 macrophage. (A) Verification of METTL3 overexpression at the mRNA level (left) and protein level (right) in RAW 264.7 cells. (B) qRT-PCR showing the mRNA expression level of IL-6 and iNOS increased by METTL3 forced expression in RAW 264.7 cells. Data are analyzed with Student’s t-test (two-tailed) and expressed as mean ± SEM. **P < 0.005, ***P < 0.001.

Additionally, we also treated bone marrow derived macrophages (BMDM) with LPS (100 ng/μl) to induce M1 macrophage, which were induced from BMSCs in the presence of MSC-F (25 ng/μl) as descripted in Figure 3A. M1 macrophage were characterized by the expression of IL-6 and iNOS in response to LPS stimulation for 24 h (Figure 3B). Consistent with what has mentioned previously, METTL3 expression was also increased in M1 macrophage polarized by BMDM (Figure 3C). To further demonstrate the METTL3-dependent effect on M1 macrophage polarization, we performed loss-of-function and gain-of-function assays. Knock-down or over-expression efficiency of METTL3 in BMDM were detected by qRT-PCR (Figure 3D, 3E). M1 macrophage maker IL-6 mRNA expression was slight reduced with METTL3 knock-down (Figure 3D). Oppositely, METTL3 over-expression promoted IL-6 and iNOS mRNA expression, which mean the induction of M1 macrophage polarization (Figure 3F). In short, METTL3 plays the promotion role in differentiation of M1 macrophages both for RAW 264.7 and BMDM cells.

Figure 3.

METTL3 promotes M1 macrophage polarization of BMDM. A. Schematic diagram presenting the procedure for the induction BMDM (M0 macrophage) from BMSCs and the polarization of M1 macrophages from BMDM. B. qRT-PCR results showing the mRNA expression level of IL-6 and iNOS in M0 and M1 macrophages. C. qRT-PCR results showing that METTL3 increased in M1 macrophage compared with M0 macrophage. D. qRT-PCR showing the verification of METTL3 knockdown in BMDM (left) and the expression level of IL-6 decreased by with METTL3 knockdown in M1 macrophage (right). E. The verification of METTL3 overexpression in BMDM detected by qRT-PCR. F. The mRNA expression level of IL-6 and iNOS increased with METTL3 overexpression. Data are analyzed with Student’s t-test (two-tailed) and expressed as mean ± SEM. *P < 0.05, ***P < 0.001.

M1 macrophage differentiation facilitated by METTL3 arguments osteogenesis differentiation of bone marrow mesenchymal stem cells (BMSCs)

At the remodeling stage of bone fracture, macrophages secrete a repertoire of inflammatory and chemotactic mediators such as TNF-α, IL-1, IL-6, CCL2 to recruit BMSCs to the fracture site [18]. Additionally, M1 macrophages could enhance mineralization by BMSCs and bone healing through the pro-inflammation cytokines secretion involving OSM, TNF-α and IL-6 [7-9]. To further investigate the interaction between bone formation and macrophage polarization mediated by METTL3, we performed trans-well assays to evaluate the function of M1 phenotype macrophage on the migration of BMSCs. After transfected with METTL3 siRNA or METTL3 plasmid, we cultured RAW 264.7 cells (lower well) with BMSCs (upper well) in the trans-well systems with the M1 differentiation induced condition as descripted in Figure 4A. Our dates demonstrated that M1 macrophage with METTL3 knockdown inhibited BMSCs invasion ability compared with normal M1 macrophage. Oppositely M1 macrophage with METTL3 overexpression promoted BMSCs migration ability significantly (Figure 4B, 4C). To further elucidate the in vivo significance of M1 polarization mediated by METTL3 in BMSCs osteogenic differentiation, we examined whether the osteogenic ability was changed when BMSCs cultured with the osteogenic induction medium which having RAW 264.7 cells supernatant culture medium. Alizarin red S (ARS) staining results illustrated that the supernatant culture medium of RAW 264.7 cells treated with LPS 24 h after METTL3 siRNA transfection observably inhibited the formation of osteogenic nodules compared to the siNC group (Figure 4D, 4E). Conversely, osteogenic nodules were dramatically in METTL3 overexpression group. Collectively, M1 macrophage polarization accelerated by METTL3 plays a key role in controlling the osteogenic differentiation and migration of BMSCs in vitro.

Figure 4.

M1 macrophage stimulated by METTL3 promotes BMSCs migration and osteogenic differentiation ability. A. Schematic diagram showing the procedure of trans-well assays with RAW 264.7 cells and BMSCs co-cultured to evaluated the impact of M1 macrophage mediated by METTTL3 on BMSCs migration ability. B. Crystal violet staining representative image of trans-well assays. (bar: 100 μm). C. Bar graph representing the quantification of invasive cells (n=8). D. ARS staining shows effects of supernatant from different RAW 264.7 or not on osteogenic differentiation of BMSCs. PM: proliferation media; OM: osteogenic media (black bar: 2 mm; white bar: 50 μm). E. Typical examples of ARS staining showing the effects of M1 macrophage with METTL3 silencing or overexpression on BMSCs osteogenic differentiation ability. Data are analyzed with Student’s t-test (two-tailed) and expressed as mean ± SEM. ***P < 0.001.

M⁶A-immunoprecipitation (RIP) microarray assays to identify potential targets of METTL3

Considering that BMDM is derived from BMSCs with the treatment of M-SCF, we used our previously published m⁶A RNA immunoprecipitation (RIP) microarray data [17] to identify the target mRNAs of METTL3 in siRNA negative control (siNC)- and siMETTL3-transfected BMSCs. Among 41,187 mRNAs detected, we identified 9,438 mRNAs m⁶A modification decrease in METTL3 knockdown cells relative to those in the NC control cells (Figure 5A). Among these, there are 6 mRNAs which are the most strongly m⁶A-mehylated (level > 0.4 in NC group) and they methylation was markedly decreased upon knockdown of METTL3 (> 3-fold decrease) (Table 1). Notably, dual-specificity phosphatase 14 (Dusp14), histone deacetylase 5 (HDAC5) and NFAT activating protein with ITAM motif 1 (Nfam1) are related to inflammation cytokine secretion and bone formation. To probe the mechanism by which m⁶A methylation of Dusp14, HDAC5 and Nfam1 affects its processing in vivo, we sought to identify the position of the methylation site within the Dusp14, HDAC5 and Nfam1 mRNA [19-23]. Interestingly, they all carry potential m⁶A modification sites according to a sequence-based SRAMP m⁶A modification site prediction web (http://www.cuilab.cn/sramp) (Figure 5B and Supplementary Figures 1, 2, 3). Therefore, we asked whether they are regulated by METTL3 during the processing of BMDM differentiation from BMSCs. Previous reports have revealed that Dusp14 alleviated inflammation by reduced the expression of IL-1β and TNF-α and protect bone loss via AMPKa-dependent manner [19,20]. Dusp14 mRNA expression was significant down-regulated in BMDM and reversed by METTL3 silence, while there is no obvious change being detected with METTL3 overexpression (Figure 5C). Interestingly, the expression of HDAC5, a class IIa HDAC, was downregulated by METTL3 knockdown (Figure 5D). Previous studies have suggested a significant role of HDAC5 in osteogenic differentiation and osteoporosis progress [24,25]. Li et al. have proofed that HDAC5 deficiency mice presented poor bone mineral density [26]. As expected, forced overexpression obviously reduced HDAC5 mRNA expression. Besides, the mRNA expression of Nfam1 is too low to be detected. Collectively, METTL3 mediated BMDM function in BMSCs osteogenesis and migration via targeting HDAC5.

Figure 5.

M⁶A-immunoprecipitation (RIP) microarray identification of METTL3 targets. A. Schematic illustration of m⁶A-RIP microarray analysis for RNA methylation flowchart showing the procedures for searching the METTL3 targets in BMSCs (left). Overview of totally different expression mRNA and the most obviously different expression mRNA in m⁶A MeRIP-Seq (right). B. Potentially m⁶A sites of Dusp14, HDAC5 and Nfam1 mRNA predicted by SRAMP program. C. The effects of METTL3 on the Dusp14 mRNA expression level in BMDM. D. The effects of METTL3 on the HDAC5 mRNA expression level in BMDM. Data are analyzed with Student’s t-test (two-tailed) and expressed as mean ± SEM. **P < 0.005, ***P < 0.001.

Table 1.

mRNA with METTL3-mediated m⁶A methylation

Gene Symbol	siNC (IP)	siMETTL3 (IP)	siNC (Supernatant)	siMETTL3 (Supernatant)	siNC (Methylation Level)	siMETTL3 (Methylation Level)	Fold change (siMETTL3/siNC)
Dusp14	3.73	-1.30	1.22	1.50	0.85	0.13	6.78
Slc17a7	-2.69	-7.45	-4.14	-4.94	0.73	0.15	4.89
Kcnj14	3.64	-3.63	2.70	-1.24	0.66	0.16	4.12
Hdac5	4.44	-3.46	3.79	-1.38	0.61	0.19	3.20
Mpc1	3.75	-3.43	3.18	-0.77	0.60	0.14	4.38
Nfam1	-4.45	-4.90	-4.33	-2.18	0.48	0.13	3.61

Discussion

In the present study, we demonstrated that methyltransferase-like 3 (METTL3), a m⁶A methyltransferase, participates M1 polarization of macrophages and further induces BMSCs migration and osteogenic differentiation, which are the critical beneficial factors for bone repair. We found that METTL3 expression was positively correlated with M1 polarization of macrophages. Moreover, M1 macrophage polarization promoted by METTL3 forced expression could obviously improve the migration and osteogenesis ability of BMSCs benefiting to the bone formation. In contrast, down-regulated of METTL3 produced the opposite effects. Moreover, METTL3 silence significantly reduced the m⁶A modification of HDAC5, which has been reported to be a favorable factor for osteoporosis treatment [26]. Meanwhile, METTL3 depletion attenuated the expression of HDAC5 during the M1 polarization of macrophages. Collectively, our study demonstrates the functional importance of METTL3 in bone repair via targeting HDAC5 and modulating M1 macrophage polarization.

Plenty of studies have demonstrated the strong association between m⁶A modification and bone-related diseases such as osteoporosis, osteoarthritis and osteosarcoma. We have shown that METTL3 regulates BMSCs osteogenic differentiation via m⁶A-based direct and indirect of RUNX2 [17]. In this previous work, we found that the expression of METTL3 were very low both in the bone tissues of osteoporosis patients and ovariectomized mice, increased dramatically during osteogenic differentiation in vitro. Therefore, it is possible that the increased expression of METTL3 may also contribute to the progression of bone repair.

Bone repair requires the involvement of multiple cells including osteoblast, osteoclast, monocyte such as preosteoclast and macrophage. Osteal macrophages (OsteoMacs) are a special subtype of macrophage residing in bone tissues which play vast and substantial roles in bone biology due to their key function in bone formation and remodeling [27]. Current data indicates that polarization of macrophages toward either M1 or M2 subset phenotype could affect bone formation process directly or indirectly by their inflammation factor secretion [28]. Inflammatory cytokines secreted by macrophages could enhance BMSC mobilization to injured sites and osteogenic differentiation [29]. In the early steps of bone formation, the capabilities of BMSCs to display osteogenic differentiation and migrate to the fracture site play a crucial role [30]. Noticeably, METTL3 has reported to have important function in macrophage polarization, which could facilitate M1 macrophage polarization through the methylation of STAT1 mRNA [31,32]. Nevertheless, it was still a puzzle whether m⁶A methyltransferase METTL3 exerted effects on bone repair before the present study. In this study, we investigate that the expression of METTL3 increase with LPS treatment in macrophages. Besides, forced expression of METTL3 facilitates M1 macrophages polarization. Furthermore, we for the first time revealed that overexpression of METTL3 in M1 macrophages could accelerate the osteogenic differentiation of BMSCs when co-cultured with BMSCs. Consisting with our results, Yang et al. have proved that M1 macrophages secretes higher soluble factors oncostatin M (OSM) levels to promote the osteogenic differentiation of co-cultured BMSCs at the early time [33]. Meanwhile, inflammatory cytokines secreted by M1 macrophages, including TNF-α, IL-6, OSM, MCP-1, MIP-1α and IL-8 can accelerate the migration of BMSCs [34,35]. Nevertheless, we have to point out that we do not discover the influence of METTL3 on M2 polarization in this study. To date there has been little agreement on the function of these two macrophage subtypes in osteogenic differentiation and migration ability of BMSCs. Recent studies have confirmed that M1 macrophages are the main immune cells at the first 3 days of bone injury, which will be gradually replaced by M2 macrophages. Additionally, Schlundt et al. discovered that the enhancement of M2 phenotype in macrophages at bone injury sites through IL-4 and IL-13 secretion promoted bone formation during the 3 weeks of investigation period [36]. Although not ignoring the function of M1 macrophages, these results suppose a more important function of M2 macrophages in bone repair [36]. Not with standing its limitation, this study does suggest METTL3 is beneficial to bone formation via modulating M1 macrophage differentiation. We will further explore the exact role of METTL3 in M2 macrophage differentiation in the future research.

Our m⁶A-RIP-microarray demonstrated that the m⁶A methylation modification levels of DUSP14, HDAC5 and Nfam1 were significantly decreased with METTL3 silencing, which have been reported to be related with bone formation or macrophage polarization modulation. Dusp14 is an atypical member of the DUSP proteins family, which has been verified to participate in a variety of pathogenesis, including oxidative stress and apoptosis and suppress various types of inflammatory cells [37-39]. Lei et al. have demonstrated that Dusp14 can prevent osteoclasts differentiation via elevating M-CSF/RANKL-elicited inflammation and apoptosis, which has an adverse impact on osteoporosis development. Although, in our study, we found that Dusp14 mRNA expression significant down-regulated in M1 macrophage polarization process which was reversed by METTL3 silence, there is no obvious change being detected with METTL3 overexpression. HDAC5, a class II HDACs, has been reported to participate the macrophage polarization [40] and has a promotion role in osteogenesis of BMSCs [22,26]. Therefore, we hypothesized that M1 macrophage polarization promoted by METTL3 induce BMSCs osteogenesis and migration ability may by the means of expression changes of HDAC5. According to our results, knockdown of METTL3 reduced and overexpression of METTL3 enhanced the expression of HDAC in BMDM, which is positive collection with pro-inflammation cytokines (IL-6, iNOS). NFAM1 signaling is favorable for osteoclast formation and bone resorption activity in Paget’s disease of bone [23]. However, the expression of Nfam1 was too low to be detected by qRT-RCR in M1 macrophage induced from BMDM with or without METTL3 siRNA or vector transfection. Thus, it is possible that the proliferation of M1 macrophages relies on the METTL3 signaling which is more responsive to LPS treatment, and the positive target of METTL3 is HDAC5.

In summary, our results suggest that METTL3 is a bone repair stimulating factor that could promote osteogenic differentiation and migration ability of BMSCs, working at least partially by enhancing the differentiation of M1 macrophage from M0 macrophage. We have also demonstrated that METTL3 silencing directly reduce the m⁶A methylation of HDAC5 and down-regulated the mRNA expression of HDAC5, which has been reported to increase osteogenic genes expression and attenuate osteoporosis process [25]. Thus, METTL3 signaling might be considered a molecular target for the development of new strategies for the treatment of bone repair.

Disclosure of conflict of interest

None.