The titanium particles used in the present study were obtained from Alfa Aesar; Thermo Fisher Scientific, Inc. and prepared as previously described (10,22,29). Briefly, the particles were incubated at 180°C for 12 h, and subsequently immersed in 75% ethanol at room temperature for 48 h to eliminate endotoxins. Endotoxin levels in particles were detected using a Limulus assay kit (Xiamen Bioendo Technology Co., Ltd.) to ensure that endotoxin levels were <0.02 EU/ml.

The cells used in the present study were cultured at 37°C with 5% CO₂. The murine osteoblast cell line, MC3T3-E1 (iCell Bioscience, Inc.) was cultured in α-minimum essential medium (Gibco; Thermo Fisher Scientific, Inc.) with 10% FBS (Gibco; Thermo Fisher Scientific, Inc.) and 100 U/ml penicillin/streptomycin (Gibco; Thermo Fisher Scientific, Inc.). For osteogenic differentiation, cells were cultured in the aforementioned growth medium with the addition of 50 µg/ml ascorbic acid (Sigma-Aldrich; Merck KGaA) and 10 mM β-glycerophosphate (Sigma-Aldrich; Merck KGaA). The prepared titanium particles were used at a concentration of 100 µg/ml. For titanium particle treatment groups, particles were cocultured with osteogenic induction medium to figure out the effect of titanium particles on osteoblasts in an osteogenic environment (10,22).

The RAW264.7 mouse macrophage cell line was purchased from American Type Culture Collection and cultured in DMEM (Gibco; Thermo Fisher Scientific, Inc.) with 10% FBS and 100 U/ml penicillin/streptomycin (Gibco; Thermo Fisher Scientific, Inc.). The supernatants of osteoblast cells with different treatments, including titanium particle treatment and the knockdown or overexpression of the Mettl3 gene, were harvested and then centrifuged at 500 × g at room temperature for 5 min to remove cell debris or titanium particles. Subsequently, the supernatant was mixed with culture medium at a ratio of 1:1 (conditioned medium) and then RAW264.7 cells were cultured in conditioned medium for the detection of osteoclastic differentiation. Cells cultured in growth medium were used as the negative control (NC), while cells treated with 50 ng/ml recombinant murine receptor activator of NF-κB ligand (RANKL; PeproTech, Inc.) were used as the positive control. The samples were collected at 2 and 3 days for reverse transcription-quantitative PCR (RT-qPCR) and western blotting, respectively.

For nucleotide binding oligomerization domain (NOD)1 pathway inhibition experiments, osteoblast cells were pretreated with 30 µM ML130 (MedChemExpress) and 10 µM WEHI-345 (Abmole Bioscience Inc.) at 37°C for 2 h (49,50).

Following co-culture with titanium particles for 24 h, osteoblast cells were collected and fixed with 2.5% glutaraldehyde in PBS at 4°C for 1 h. No stain was used as the stain would highlight various intracellular structures. Cells were fixed with 1% osmium tetroxide in H₂O for 2 h, followed by sequential dehydration with a graded ethanol series. Samples were embedded in epoxy resin at 60°C for 12 h, cut into 70–90-nm ultrathin sections and subsequently observed using a JEM-1200EX electron microscope (JEOL, Ltd.).

Total RNA was extracted from cells using TRIzol® reagent (Invitrogen; Thermo Fisher Scientific, Inc.). cDNA was synthesized using total RNA and HiScript III RT SuperMix (Vazyme Biotech Co., Ltd.) at 37°C for 15 h and then 85°C for 5 sec. ChamQ Universal SYBR qPCR Master Mix (Vazyme Biotech Co., Ltd.) and a Light Cycler 480 (Roche Diagnostics GmbH) were employed to amplify cDNA. The initial denaturation condition was 95°C for 10 sec. The thermocycling (40 cycles) conditions were as follows: 95°C for 10 sec and 60°C for 30 sec. mRNA levels were quantified using the 2^−ΔΔCq method and normalized to the internal reference gene GAPDH (51). The specific primers are listed in Table SI.

Following rinsing three times with PBS, cells were lysed using RIPA buffer (Beyotime Institute of Biotechnology) supplemented with protease and phosphatase inhibitors (CoWin Biosciences). Cell homogenates were obtained and centrifuged at 14,000 × g for 20 min at 4°C. The cell supernatant was obtained and a BCA protein assay kit (CoWin Biosciences) was used to determine the concentration of total proteins. Samples were subsequently incubated at 95°C for 10 min in SDS-PAGE sample loading buffer (Beyotime Institute of Biotechnology) for denaturation. 20 µg total protein samples were separated via 4–12% SDS-polyacrylamide gel electrophoresis and transferred to PVDF membranes (MilliporeSigma). Membranes were blocked in 5% skimmed milk powder supplemented with 1X TBS-1‰ Tween-20 (TBS-T) at room temperature for 1 h, and subsequently incubated with primary antibodies at 4°C for 12 h. All primary antibodies are listed in Table SII. Following rinsing with TBS-T, the membranes were incubated with the HRP goat anti-rabbit IgG secondary antibody (Beijing Emarbio Science & Technology Co., Ltd.) at room temperature for 1 h. Protein bands were visualized using ECL luminophore (MilliporeSigma) and an enhanced chemiluminescence detection system (Bio-Rad Laboratories, Inc.). Protein expression was semi-quantified using ImageJ software v1.53e (National Institutes of Health). For the MAPK pathway detection, vinculin (124 kDa) was used as internal reference as it has been widely used in other studies (52,53), to ensure there is no overlap with the molecular weight of the experimental proteins (40–54 kDa). In all other cases, β-actin (42 kDa) was used as the internal reference.

Cells were fixed with 4% paraformaldehyde for 15 min at room temperature and incubated with alizarin red solution (Wuhan Servicebio Technology Co., Ltd.) at room temperature for 10 min. Cells were incubated in PBS to remove excess stain. Matrix calcium depositions were imaged using a light microscope (Zeiss AG) and subsequently dissolved in 10% cetylpyridinium chloride (Sigma-Aldrich; Merck KGaA) at room temperature for 1 h. The absorbance was measured at 562 nm using Gene5 CHS 3.11 software (BioTek; Agilent Technologies, Inc.) on a microplate reader (BioTek; Agilent Technologies, Inc.).

Cell culture supernatants were collected from each group and centrifuged at 500 × g for 5 min at 4°C. ELISA kits (Beijing winter song Boye Biotechnology Co. Ltd.) were used to measure the secreted cytokine expression of RANKL (Mouse RANKL ELISA Kit; cat. no. DG30574M), osteoprotegerin (OPG; Mouse OPG ELISA Kit; cat. no. DG30295M), IL-6 (Mouse IL-6 ELISA Kit; cat. no. DG30754M), TNF-α (Mouse TNF-α ELISA Kit; cat. no. DG30048M) and IL-1β (Mouse IL-1β ELISA Kit; cat. no. DG94767Q) according to the manufacturer's protocol.

Cells were treated with 5 µg/ml actinomycin D (ACMEC biochemical) at 37°C to inhibit mRNA transcription. Total RNA was extracted at 0, 3 and 6 h. RT-qPCR was carried out as aforementioned to examine mRNA expression and calculate the degradation rate of target genes, which was presented as the relative expression of mRNA at each time point relative to that at the 0-h timepoint (54).

Specific lentiviral vectors carrying short hairpin RNA (shRNA/sh) or control shRNA (GeneChem, Inc.) were used for Mettl3 knockdown. shRNA sequences are shown in Table SIII. Lentiviral vectors were used for Mettl3 overexpression, and blank vectors were prepared by Guangzhou IGE Biotechnology, Ltd. shRNA sequences were synthesized and Mettl3 overexpression products were amplified. The lentivirus transfer plasmid pLVX (Thermo Fisher Scientific, Inc.)-shRNA was used for recombinant plasmid construction and lentivirus production. And for Mettl3 overexpression, pCDH-CMV-MCS-EF1-copGFP-T2A-Puro vector (Thermo Fisher Scientific, Inc.) was employed and the Mettl3 coding region was inserted into the MCS sites of the vector. At 18 h before transfection, HEK 293T cells (Procell Life Science &Technology Co., Ltd.) were inoculated in six-well cell culture plates at a density of 4×10⁵ cells per well. Before transfection, 2 µg of the plasmid was diluted in 100 µl Opti-MEM medium (Gibco) without serum (tube A), and 6 µl polyethylenimine reagent was diluted in another 100 µl serum-free Opti-MEM (tube B). We then added the dilution in tube B into tube A and mixed it gently. The mixture was incubated at room temperature for 12 min and then added to the cell culture medium. A 3rd generation system including four-plasmid lentivirus, pLP/VSVG, pLP1, pLP2, and empty/recombinant pLVX vector (Thermo Fisher Scientific, Inc.), was applied at a molar ratio of 1:1:1:2. Cells were incubated with transfection reagent-plasmid mixture at 37°C for 6 h before replacement. At 24 h after replacement, cell culture medium containing lentiviral particles was collected. The above protocols were completed by company and we carried out the subsequent transfection and selection. In brief, lentiviruses were cultured with MC3T3-E1 cells at a MOI of 50 at 37°C for 12 h. Green fluorescence protein labeling was used to observe transfection using a light microscope (Zeiss AG). Cell lines with stable expression were selected using 6 µg/ml puromycin for 3 days and maintained using 2 µg/ml puromycin (Biosharp Life Sciences). For YTH domain family 2 (Ythdf2) knockdown, 50 nM small interfering RNA (si)Ythdf2 or NC (Guangzhou IGE Biotechnology, Ltd.), named siCtrl, was employed to transfect cells with the addition of Lipofectamine® 2000 Transfection reagent (Biosharp Life Sciences) at 37°C for 24 h (Table SIII). The transfected cells were employed for subsequent experimentations at 24 h after transfection. The transfection efficiency was determined using RT-qPCR and western blot analysis and at 24 and 48 h after the transfection, respectively.

Total RNA was extracted as aforementioned and incubated at 95°C for 3 min to disrupt secondary structures. Denatured total RNA was loaded onto nitrocellulose membranes and subsequently crosslinked to the membrane with UV light at 254 nm for 30 min. Unbound RNA was removed by washing with TBS-T. Western blot analysis was carried out as aforementioned using the anti-m⁶A antibody (Table SII).

The EpiQuik CUT&RUN m⁶A RNA Enrichment kit (EpigenTek Group, Inc.) was used for the MeRIP assay as previously described (55,56). Briefly, total RNA was collected and randomly fragmented. RNA fragments were immunoprecipitated with magnetic beads pre-coated with anti-immunoglobulin (IgG), anti-m⁶A or anti-Ythdf2 antibodies. Targeted RNA fragments were released, purified and eluted, followed by RT-qPCR as aforementioned. Relative enrichment was normalized to the input sample.

All data were representative of three independent experiments and presented as the mean ± standard deviation. Data were plotted and analyzed using GraphPad Prism v9.0 (Dotmatics). Comparisons between two groups were analyzed using an unpaired Student's t-test. Comparisons among multiple groups were analyzed using one-way ANOVA followed by Tukey's post hoc test and two-way ANOVA followed by Bonferroni's post hoc test. P<0.05 was considered to indicate a statistically significant difference.

Titanium particles engulfed into osteoblastic cells were observed using TEM (Fig. 1A). In the present study, RT-qPCR and western blot analysis were used to show early trends, and alizarin red staining was used to study later trends. The results of the present study demonstrated that titanium particle treatment significantly inhibited the expression of osteogenesis-associated markers and the formation of calcium nodules compared with those in the osteogenic induction group (Fig. 1B and C). Furthermore, titanium particles induced the mRNA expression of proinflammatory cytokines (Fig. 1D), and these results were verified via ELISA. To further investigate the regulatory role of titanium particles in osteoblast-osteoclast communication, which is critical for bone homeostasis (57), the supernatant of titanium-treated osteoblasts was harvested and co-cultured with preosteoclasts. As shown in Fig. 1E, mRNA and protein expression levels of markers associated with osteoclast differentiation of RAW264.7 cells were increased in the titanium particle treatment group (CM) compared with the NC group. Thus, the expression levels of additional soluble factors released from osteoblasts that mediate osteoblast-osteoclast communication were investigated. Notably, RANKL expression was significantly increased, leading to a higher ratio of RANKL/OPG (Fig. 1F).

To explore the role of m⁶A modification in titanium particle treatment of osteoblasts, m⁶A levels were evaluated. The results of the RNA dot blot analysis demonstrated that m⁶A levels were reduced in the titanium treatment group compared with the osteogenic induction group (Fig. 2A). Subsequently, expression levels of common methyltransferases and demethylases were further determined via RT-qPCR and western blotting (Fig. 2B and C). The expression levels of Mettl3, a key methyltransferase, were significantly decreased during titanium treatment, and these results were consistent with the measurement of m⁶A levels.

To explore the association between titanium particle-induced bioactivities and the reduced expression levels of Mettl3, Mettl3 knockdown was carried out in MC3T3-E1 cells, referred to as shMettl3 cells (Fig. 3A). The results of the present study demonstrated that Mettl3 mRNA and protein expression was significantly reduced following transfection compared with the NC or blank groups (Fig. 3B). The results of the RNA dot blot analysis revealed that m⁶A levels were lowest in the shMettl3 group compared with the control and titanium particle treatment groups (Fig. 3C). Notably, mRNA and protein expression levels of osteogenic markers were significantly reduced in the shMettl3 group (Fig. 3D). In addition, the alizarin red staining assay demonstrated consistent results, indicating that Mettl3 knockdown induced osteogenic inhibition comparable with that induced by titanium particles (Fig. 3D). Furthermore, Mettl3 knockdown increased the expression of proinflammatory cytokines (Fig. 3E). Mettl3 knockdown increased the RANKL expression and the ratio of RANKL/OPG compared with the control group (shCtrl osteoblasts without titanium particle) (Fig. 3F). Additionally, the supernatant of Mettl3-knockdown osteoblasts was harvested and co-cultured with RAW264.7 cells. mRNA and protein expression levels of markers associated with osteoclastogenesis were increased in RAW264.7 cells cultured with supernatant of Mettl3-knockdown osteoblasts compared with the control group (shCtrl osteoblasts without titanium particle) (Fig. 3G).

To further verify whether titanium particles mediated the aforementioned activities through reduced Mettl3 expression, a Mettl3-overexpression osteoblast cell line was constructed, and a blank vector was used as the NC. RT-qPCR and western blotting were performed to determine the transfection efficiency following overexpression (Fig. 4A). TEM analysis demonstrated that Mettl3 overexpression did not impact the engulfment capability of osteoblasts (Fig. 4B). Notably, m⁶A levels in the Mettl3 overexpression group were significantly increased compared with those in the blank group (Fig. 4C). The results of the present study also demonstrated that Mettl3 overexpression attenuated titanium-induced inhibition of osteogenesis (Fig. 4D). In addition, titanium-induced proinflammatory activities were also attenuated, as well as the ratio of RANKL/OPG (Fig. 4E and F). The aforementioned results in Fig. 1E indicated the regulatory role of titanium particles in osteoblast-osteoclast communication, while this indirect regulatory effect of titanium particles on osteoclastogenesis of RAW264.7 cells was also reversed by Mettl3 overexpression (Fig. 4G).

To explore the downstream signaling molecules affected by Mettl3, the canonical BMP-Smad signaling pathway was evaluated. The results of RT-qPCR analysis demonstrated that titanium particle treatment increased Smad7 and Smurf1 expression in osteoblasts, while Mettl3 overexpression attenuated the upregulated expression of Smad7 and Smurf1 induced by titanium particle treatment (Fig. 5A). The results of western blotting further demonstrated that titanium particle treatment inhibited the phosphorylation levels of Smad1/5/9 and induced the increased expression of Smad7 and Smurf1, and these reactions induced by titanium particle treatment were also reversed by Mettl3 overexpression (Fig. 5B). The results of the present study also demonstrated that both titanium particle treatment and Mettl3 knockdown increased the mRNA stability of these two genes, leading to increased expression levels (Fig. 5C). Furthermore, MeRIP-qPCR analysis revealed that the m⁶A enrichment of Smad7 and Smurf1 transcripts was significantly reduced in the titanium particle treatment and Mettl3 knockdown groups compared with the control group (Fig. 5D).

The results of western blotting revealed that titanium particle treatment induced the activation of the MAPK and NF-κB pathways, while Mettl3 overexpression attenuated this activation. For p38, it seemed that the titanium particle induced its activation while the Mettl3 overexpression failed to attenuated this activation. (Fig. 6A). The expression changes of key components of the NLR signaling pathway were also evaluated. The results of the present study demonstrated that titanium particle treatment increased the expression and the mRNA half-life of NOD1 and receptor interacting serine/threonine kinase 2 (RIPK2) in osteoblasts. However, the mRNA stability and expression of NOD2 were not enhanced (Fig. 6B and C). MeRIP-qPCR analysis demonstrated that NOD1 and RIPK2 transcripts functioned as the target of Mettl3, as titanium particle treatment and Mettl3 knockdown both significantly reduced the m⁶A enrichment of these two transcripts compared with the control group (Fig. 6D).

To further verify the role of the NLR1 signaling pathway in the titanium particle-induced inflammatory response, cells were pretreated with ML130 (NOD1 inhibitor) and WEHI-345 (RIPK2 inhibitor) to block the NOD1 pathway since titanium particles can act on both NOD1 and RIPK2 mRNA based on the aforementioned results. In Fig. 7A, without the addition of inhibitors, titanium particle induced the activation of the MAPK and NF-κB signaling pathways, and the activation of these two pathways was significantly inhibited following treatment with both inhibitors from the comparison between titanium particle treatment without inhibitors group and titanium particle treatment with inhibitors group (Fig. 7A). In addition, inhibition of the NOD1 pathway attenuated the titanium particle-induced increased expression of IL-6, TNF-α, IL-1β and RANKL (Fig. 7B). Collectively, these data suggested that titanium particles may target the NOD1 signaling pathway to regulate subsequent inflammatory responses.

m⁶A readers recognize m⁶A modifications and then mediate regulatory effects, including RNA translation, decay and splicing (41). The aforementioned results (Figs. 5D and 6D) revealed that downregulated Mettl3 levels led to the enhancement of mRNA stabilities of Smad7, Smurf1, NOD1 and RIPK2, and thus, it was hypothesized that it was Ythdf2 that exerted the effects. Firstly, siYthdf2 was used to knockdown its expression (Fig. 8A). The mRNA and protein expression levels of Smad7, Smurf1, NOD1 and RIPK2 were increased in the Ythdf2 knockdown group compared with the control group (Fig. 8B). The mRNA stabilities of these four mRNAs were also enhanced in Ythdf2 knockdown cells at 3 and 6 h, while the stability of NOD at 6 h didn't show statistical difference (Fig. 8C), suggesting that Ythdf2 could affect the expression of these four mRNAs by regulating their stabilities. In Fig. 8D, we measured the relative expressions of these four mRNAs precipitated by Ythdf2, which were shown in anti-Ythdf2 column. Notably, titanium particle treatment reduced the expression levels of these mRNAs compared with control group, indicating that the binding between Ythdf2 and mRNAs was significantly decreased following titanium particle treatment (Fig. 8D).