THP-1 cells (Wuhan Pricella Biotechnology Co., Ltd., China) and RA-FLSs (Cellverse Bioscience Technology Co., Ltd., China) were resuscitated, passaged, and cryopreserved before experiments. For resuscitation, cells were obtained from liquid nitrogen, rapidly thawed in a water bath at 37°C, and then seeded for culture. During passaging, adherent RA-FLSs were passaged at 1:2 after trypsinization, while THP-1 cells were directly passaged at 1:2. THP-1 cells were seeded into 6-well plates at 1 × 10^6, treated with 100 ng/mL phorbol-12-myristate-13-acetate (PMA) overnight, and induced into M0 macrophages after 72 h. The induced M0 macrophages were co-cultured with treated RA-FLSs.

RA-FLSs cells were resuscitated and cultured in DMEM, whereas THP-1 cells were resuscitated and maintained in RPMI-1640. The complete DMEM medium consisted of 89% DMEM basal medium, 10% fetal bovine serum (FBS, v/v), and 1% penicillin-streptomycin double antibiotic solution (v/v). The complete RPMI-1640 medium had the same composition as DMEM complete medium, except the basal medium was replaced with RPMI-1640. The cryopreservation solution was prepared by combining 50% complete medium, 40% FBS, and 10% cell-grade dimethyl sulfoxide (DMSO, v/v). RA-FLSs cells were passaged at a 1:2 to 1:3 ratio and subcultured every 2–3 days upon reaching confluence, while THP-1 cells were passaged at a 1:2 ratio and subcultured every 2 days when confluent.

After resuscitation, passaging, and cryopreservation, RA-FLSs were treated according to groups for different research purposes and then co-cultured with macrophages. As for basic treatment, some cells were treated without additional interventions as a blank control group, while some cells were treated with 20 ng/mL TNF-α for 48 h to mimic the enhanced inflammatory microenvironment in RA. To investigate the effect of exosome deficiency, 10 μM GW4869 was added to TNF-α-treated RA-FLSs (24 h) to inhibit exosome release. To clarify the function of WTAP protein, the WTAP overexpression vector was introduced into RA-FLSs using transfection reagents, with empty vectors as the negative control (NC). Small interfering RNA (siRNA) targeting WTAP was designed for transfection to knock down WTAP expression, with non-specific siRNA as the control. Finally, rescue experiments were performed to verify the role of m⁶A modification in the regulation of circ-CBLB by WTAP. Wild-type (WT) and mutant (MUT) circ-CBLB plasmids were separately transfected into cells. Additionally, these two types of plasmids were further co-transfected with WTAP overexpression vectors to observe whether the effect of WTAP on circ-CBLB was altered by m⁶A modification site mutation. Each of the above groups contained 3 biological replicates (n = 3).

IL-6, TNF-α, IL-13, and IL-10 levels were measured as instructed in the manuals of corresponding ELISA kits (Ruixin Biotech, China).

Samples were centrifuged at 5,000 rpm for 10 min, and the supernatant was filtered through a 0.22 μm filter. EP Solution was added to the supernatant at ratios of 3:1 and 3:2, followed by 1 h of incubation at 4°C. After centrifugation at 12,000 rpm (4°C, 15 min), the supernatant was discarded. The precipitate was resuspended with 500–1,000 μL of PS Buffer through repeated pipetting, incubated at 37°C for 20 min, and further pipetted to obtain exosomes.

Exosomes were stained with the PKH26 Red Fluorescent Cell Membrane Labeling Kit (Solarbio, China). Exosome pellets were resuspended in 100 μL of 1× diluent, mixed with 2× staining solution, and incubated at 25°C for 2–5 min. THP-1 cells were seeded into 6-well plates, treated with 100 ng/mL PMA for 48 h, and incubated with labeled exosomes (90 μg/mL) for 24 h. After liquids were aspirated, cells were washed with PBS twice. Finally, images were captured under a microscope (400×).

Cell precipitates were lysed with 1 mL of TRIzol, and RNA was extracted through sequential addition of chloroform and isopropanol, centrifugation, and incubation. RNA was dissolved in DEPC water and stored. Reverse transcription was performed with a reverse transcription kit to generate cDNA, which was stored at -20°C. With the use of 2× SYBR Green qPCR Master Mix, specific primers, and cDNA, Fluorescence quantitative PCR was conducted on a StepOne Plus system (ABI, USA) using the Relative Quantification Study method. Relative expression was calculated with the 2^−ΔΔCt method. Primer sequences are provided in Supplementary Materials .

RNA was extracted with the same procedure as RT-qPCR. m⁶A-containing RNA fragments were enriched with the EpiQuik CUT&RUN m6A RNA Enrichment Kit (Epigen Tek, USA). Enriched RNA was released, purified, eluted, and stored at -20°C. Reverse transcription and qPCR were carried out as described above. Primer sequences are provided in Supplementary Materials .

RA-FLSs (2 × 10^7) were washed with PBS and lysed with polysome lysis buffer containing Protease inhibitor and RNase inhibitor. DNA was removed with DNase. Protein A/G beads were equilibrated with polysome lysis buffer. Lysates were divided into IP, IgG, and Input groups. IP and IgG samples were incubated with 3 μg of WTAP antibody (Proteintech, USA) or IgG antibody, respectively, followed by incubation with equilibrated beads. After multiple washes with Washing Buffer and DTT, RNA was eluted with polysome elution buffer, DTT, and proteinase K. Eluted RNA was extracted with TRIzol/chloroform, precipitated, and stored. Reverse transcription and qPCR were performed as described above. Primer sequences are provided in Supplementary Materials .

Cells were incubated with CD11B and CD86 antibodies (15 min, room temperature, dark), permeabilized, and then stained with CD206 antibodies (15 min, room temperature, dark). After washing, cells were resuspended in PBS containing 5% FBS, filtered using a 200 mesh filter, and analyzed on a NovoCyte flow cytometer (Agilent, USA). Data were processed using NovoExpress software.

Total protein was extracted with RIPA lysis buffer containing PMSF. Proteins were separated via SDS-PAGE (5% stacking gel, 10% separating gel) and transferred to PVDF membranes. Membranes were blocked with Western blocking solutions, incubated with primary antibodies overnight, washed, and incubated with secondary antibodies. Protein bands were visualized with an automated exposure system and analyzed with Image J software. Antibody details are presented in Supplementary Materials .

Biotin-labeled hsa_circ-CBLB and NC probes were synthesized by BersinBio (China). Cells (2 × 10^7) were lysed with RIP buffer containing Protease inhibitor. Nucleic acids were removed using DNase. Agarose beads were pre-washed, and lysates were divided into Input, RPD, and NC groups. Biotin-labeled probes (1 μg) were denatured, ice-bathed, and folded with RNA structure buffer. Streptavidin magnetic beads were bound to RNA probes, mixed with cell lysates, RNase inhibitor, and Yeast tRNA, and washed. Proteins were eluted with Protein elution buffer and DTT. Products were analyzed with rapid silver staining and Western blot. Silver staining involved SDS-PAGE gel preparation, electrophoresis, fixation, sensitization, staining, and development.

Cells were treated with 4 μM Actinomycin D (AbMole, USA) for 0 h, 2 h, 4 h, and 8 h. RNA was extracted at each time point, and circ-CBLB expression was quantified with qPCR.

Cells (1 × 10^6) were seeded into 6-well plates. At 70% cell confluence, siRNA-WTAP, pcDNA3.1-WTAP, WT-circ-CBLB, or MUT-circ-CBLB was mixed with Lipo8000TM transfection reagents (Beyotime, China) in serum-free medium, incubated, and added with cells. After 4 h, the medium was replaced with complete medium. Cells were harvested after 48 h for qPCR detection of transfection efficiency. For co-transfection, pcDNA3.1-WTAP/pcDNA3.1-NC was co-transfected with WT-circ-CBLB/MUT-circ-CBLB into cells as described above.

Cells were permeabilized with 200 μL of Saponin (15 min, room temperature), washed with PBS, blocked with 3% BSA (20 min), and incubated with primary antibodies (4°C, overnight). Following cell incubation with secondary antibodies, nuclei were counterstained with DAPI, and slides were imaged under a fluorescence microscope. Semi-quantitative analysis was performed with Image J. Antibody details are listed in Supplementary Materials .

All data in our study were analyzed with GraphPad Prism 4.0. All experimental procedures were repeated at least 3 times. All results were presented as mean ± standard deviation unless otherwise specified. Statistical comparisons were conducted with the one-way analysis of variance or the independent samples two-tailed Student’s t-test. Differences at P < 0.05 were considered statistically significant: * P < 0.05, ** P < 0.01.

As exhibited in Figure 1A , the typical cup-shaped particles with a diameter of 50−200 nm were observed under the transmission electron microscope, consistent with typical exosome features (19). Western blot demonstrated the presence of exosome-specific marker proteins CD63 and CD81 in both RA-FLS + M0 and RA-FLS + M0 + TNF-α groups ( Figure 1B ). After exosomes were labeled with PKH26, the uptake of exosomes by macrophages was successfully observed under the fluorescence microscope ( Figure 1C ).

CD86 and CD206 expression was measured with flow cytometry. The results showed that the M1/M2 ratio in the RA-FLS + M0 + TNF-α group was significantly higher than that in the RA-FLS + M0 group ( Figures 2A, B ). To verify whether the shift in macrophage polarization was caused by differences in exosome expression, we added the exosome inhibitor GW4869 to cells in the RA-FLS + M0 and RA-FLS + M0 + TNF-α groups, respectively, and then analyzed the expression of macrophage markers. The results revealed that the addition of GW4869 increased the M1/M2 ratio in the RA-FLS + M0 and RA-FLS + M0 + TNF-α groups ( Figures 2C, D ). The levels of IL-6, TNF-α, IL-13, and IL-10 in the supernatants of RA-FLSs co-cultured with macrophages were tested with ELISA. As found in Figures 2E-H , the levels of pro-inflammatory cytokines IL-6 and TNF-α were markedly higher and the levels of anti-inflammatory cytokines IL-13 and IL-10 were lower in the RA-FLS + M0 + TNF-α group than in the RA-FLS + M0 group, whereas GW4869 addition increased the levels of IL-6 and TNF-α and lowered the levels of IL-13 and IL-10 in the RA-FLS + M0 + TNF-α group. Images under the fluorescence microscope exhibited reductions in the uptake of exosomes by macrophages after the addition of GW4869 ( Figure 2I ). Taken together, macrophages were polarized toward the M1 phenotype after exosome inhibition.

Circ-CBLB expression in RA-FLSs, exosomes, and macrophages was examined, which displayed that circ-CBLB expression in RA-FLSs, exosomes, and macrophages was substantially lower in the RA-FLS + M0 + TNF-α group than in the RA-FLS + M0 group ( Figure 3A-C ). The addition of GW4869 did not significantly change circ-CBLB expression in RA-FLSs but declined circ-CBLB expression in exosomes and macrophages of the RA-FLS + M0 and RA-FLS + M0 + TNF-α groups ( Figures 3D-F ).

MeRIP-qPCR and RT-qPCR were used to test the m⁶A modification and expression of circ-CBLB in RA-FLSs, respectively. The results exhibited that the m⁶A modification level of circ-CBLB in the RA-FLS + M0 + TNF-α group was obviously higher than that in the RA-FLS + M0 group, accompanied by lower circ-CBLB expression ( Figures 3G, H ).

RT-qPCR was performed to measure the expression of m⁶A modification-related key enzymes including METTL3, METTL14, WTAP, FTO, and ALKBH5. As depicted in Figure 4A , the RA-FLS + M0 + TNF-α group had prominently higher METTL14, WTAP, and FTO expression than the RA-FLS + M0 group. Further Western blot and protein concentration assay confirmed that compared with those in the RA-FLS + M0 group, RA-FLSs in the RA-FLS + M0 + TNF-α group had increased levels of METTL14, WTAP, and FTO ( Figures 4B, C ), consistent with the results of immunofluorescence ( Figures 4D-I ).

RNA pull-down and RIP assays demonstrated the mutual binding of WTAP protein and circ-CBLB ( Figure 5A, B ). The enrichment of circ-CBLB in cells was tested with RIP-qPCR, which showed a significant enrichment trend in the IP group as compared with the IgG group ( Figure 5C ).

To ascertain the effect of WTAP on macrophage polarization via the m⁶A modification of circ-CBLB in RA-FLSs, we constructed WTAP silencing and overexpression models and detected the m⁶A modification level of circ-CBLB in RA-FLSs using MeRIP-qPCR. The results displayed significantly higher m⁶A modification levels in the RA-FLS + M0 + TNF-α group than in the RA-FLS + M0 group. m⁶A modification levels were insignificantly different among the RA-FLS + M0 + TNF-α + pcDNA3.1-NC, RA-FLS + M0 + TNF-α + si-NC, and RA-FLS + M0 + TNF-α groups. m⁶A modification levels was markedly higher in the RA-FLS + M0 + TNF-α + pcDNA3.1-WTAP group but lower in the RA-FLS + M0 + TNF-α + si-WTAP group than in the RA-FLS + M0 + TNF-α group ( Figure 6A ). Altogether, WTAP overexpression enhanced the m⁶A modification level of circ-CBLB in RA-FLSs.

According to RT-qPCR data, there was substantially lower circ-CBLB expression in the RA-FLS + M0 + TNF-α group than in the RA-FLS + M0 group, with no differences in circ-CBLB expression among the RA-FLS + M0 + TNF-α + pcDNA3.1-NC, RA-FLS + M0 + TNF-α + si-NC, and RA-FLS + M0 + TNF-α groups. Compared to the RA-FLS + M0 + TNF-α group, the RA-FLS + M0 + TNF-α + pcDNA3.1-WTAP group had lower circ-CBLB expression, whilst the RA-FLS + M0 + TNF-α + si-WTAP group had higher circ-CBLB expression ( Figure 6B ). The expression of circ-CBLB in macrophages was concordant with that in RA-FLSs ( Figure 6C ). Collectively, overexpression of WTAP decreased the expression of circ-CBLB.

To further delve into the specific mechanism by which WTAP affects circ-CBLB expression, we examined the half-life of circ-CBLB in RA-FLSs under WTAP silencing or overexpression using the actinomycin D assay. As observed in Figure 6D , the decay rate of circ-CBLB was the lowest in the RA-FLS + M0 group and the highest in the RA-FLS + M0 + TNF-α + pcDNA3.1-WTAP group, and there were no differences in the decay rate among the RA-FLS + M0 + TNF-α + pcDNA3.1-NC, RA-FLS + M0 + TNF-α + si-NC, and RA-FLS + M0 + TNF-α groups. The decay rate of the RA-FLS + M0 + TNF-α + si-WTAP group was higher than that of the RA-FLS + M0 group and lower than that of the RA-FLS + M0 + TNF-α group.

According to flow cytometry results, M1 polarization of macrophages was enhanced in the RA-FLS + M0 + TNF-α + pcDNA3.1-WTAP group and repressed in the RA-FLS + M0 + TNF-α + si-WTAP group ( Figures 7A, B ). PKH26 labeling of exosomes demonstrated that exosomes were taken up by macrophages ( Figure 7C ).

MeRIP-qPCR and RT-qPCR were used to examine the m⁶A modification level and expression of circ-CBLB in RA-FLSs, respectively. The results uncovered that the m⁶A modification level of circ-CBLB was decreased and the expression of circ-CBLB was increased in the RA-FLS + TNF-α + M0 + pcDNA3.1-NC + MUT-circ-CBLB group (the MUT group) versus the RA-FLS + TNF-α + M0 + pcDNA3.1-NC + WT-circ-CBLB group (the WT group) ( Figures 8A, B ).

The total RNA of RA-FLSs was extracted with TRIzol, and WTAP expression was tested with RT-qPCR. As exhibited in Figure 8C , no significant difference was observed in WTAP expression between the WT and MUT groups, and WTAP expression was substantially higher in RA-FLS + TNF-α + M0 + pcDNA3.1-WTAP + WT-circ-CBLB (the pcDNA3.1-WTAP + WT group) and RA-FLS + TNF-α + M0 + pcDNA3.1-WTAP + MUT-circ-CBLB groups (the pcDNA3.1-WTAP + MUT group) than in the WT and MUT groups, respectively. Western blot results were similar to the results of RT-qPCR ( Figures 8D, E ). According to MeRIP-qPCR results, the m⁶A modification level was lower in the MUT group than in the WT group, higher in the pcDNA3.1-WTAP + WT group than in the WT group, and higher in the pcDNA3.1-WTAP + MUT group than in the MUT group ( Figure 8F ). RT-qPCR results showed that circ-CBLB expression in RA-FLSs was higher in the MUT group than in the WT group, whilst circ-CBLB expression was lower in the pcDNA3.1-WTAP + WT and pcDNA3.1-WTAP + MUT groups when compared with the WT and MUT groups, respectively ( Figure 8G ).

Flow cytometry results exhibited that the M1/M2 ratio in the MUT group was lower than that in the WT group, whereas the M1/M2 ratios in the pcDNA3.1-WTAP + WT and pcDNA3.1-WTAP + MUT groups were higher than those in the WT and MUT groups, respectively ( Figures 9A, B ). As observed in ELISA results, IL-6 and TNF-α levels in the supernatant of macrophages co-cultured with RA-FLSs were elevated in the pcDNA3.1-WTAP + MUT group compared to the MUT group, with reduced levels of IL-10 and IL-13 ( Figures 9C-F ).

RT-qPCR results unveiled higher circ-CBLB expression in the MUT group than in the WT group but lower circ-CBLB expression in the pcDNA3.1-WTAP + WT and pcDNA3.1-WTAP + MUT groups than in the WT and MUT groups, respectively ( Figure 9G ). Fluorescence microscopy images validated the uptake of exosomes by macrophages ( Figure 9H ).