Gingival tissue of healthy individuals and patients with periodontitis was collected surgically. The healthy gingival tissue was extracted surgically from the third molars of healthy donors while gingival tissue of periodontitis was obtained from the inner wall of the periodontal pocket around the tooth which needs to be extracted because of severe periodontitis. Specifically, the healthy group should have no bleeding on probing (BOP) positive sites and show a probing depth (PD) < 3 mm. The severe periodontitis has the following features: probing depth (PD) > 6 mm、the attachment loss (AL) >5 mm and the bone absorption exceeds 1/2 of the root length. Informed consent was obtained from all donors. The study was approved by the Institutional Ethics Committee Board of the Guanghua School of Stomatology, Sun Yat-sen University (GHKQ-202211-K07-01).

Bone marrow-derived monocytes/macrophages (BMMs) were obtained from the tibia and femurs of 6-week-old male C57BL/6 mice (Animal Center of Sun Yat-sen University). Cells were seeded into 96‐well plates at a density of 1 × 10⁴/well, or 6‐well plates at a density of 5 × 10⁵/well in α-MEM (Gibco, Grand Island, NY, United States) containing 30 ng/mL M-CSF (Sino Biological Inc, Beijing, China) and 50 ng/ml RANKL (R&D Systems, Minneapolis, MN, United States) without or with CD40L (R&D Systems, Minneapolis, MN, United States) and TRAF-STOP (MedChemExpress, Shanghai, China). Cells treated with only M-CSF were used as controls. All cells were incubated in a 37°C incubator under an atmosphere with 5% CO₂.

Upon observation of multi-nucleated osteoclasts at 5-6 days of culture, TRAP staining was carried out to detect osteoclasts following the manufacturer’s instructions (TRAP staining kit, Sigma, MO, United States). TRAP‐positive cells with at least three nuclei are regarded as osteoclasts (Huang et al., 2022). TRAP-positive multinucleated cells (MNCs) were observed under the inverted microscope (Olympus IX71, Olympus, America) and counted by ImageJ (National Institutes of Health, Bethesda, MD, US)

BMMs were seeded in a 96-well plate as described above. After BMMs were differentiated into osteoclasts for 5, 6 days, cells were fixed in the 96-well plate with 4% PFA for 20 min at 37°C, followed by the permeabilization with 0.1% Triton X-100 in PBS for 10 min. After being blocked with 2% BSA for 20 min, cells were stained with Alexa Fluor 488 phalloidin (Beyotime, Shanghai, China) for 1 h at room temperature and then counterstained with DAPI for 10 min in the dark. Images of osteoclastic rings were visualized by fluorescence microscopy (Olympus IX71, Olympus, and America).

BMMs were seeded at a density of 1 × 10⁴/well onto bovine bone slices in a 96-well plate with three replicates for 8, 9 days. Later, the slices were harvested and brushed thoroughly to remove the cells after the formation of mature osteoclasts. The resorption pits on slices were observed and captured by the scanning electron microscope (Merlin Compact, Zeiss, and Germany) and the area of bone resorption was quantified by ImageJ software.

Total RNA was extracted from periodontal tissue and cells with the RNA-Quick purification kit (Yishan, Shanghai, and China) based on the instructions. Sample concentrations were determined using a NanoDrop ND-1000 spectrophotometer (Thermo Scientific, MA, and United States). cDNA synthesis was carried out using a PrimeScript™ RT Master Mix Kit (TaKaRa, Kyoto, Japan). qRT-PCR was performed using a Light Cycler 96 instrument (Roche, Basel, Switzerland) with SYBR Green I Master Mix (Roche, Basel, Switzerland). All target genes were normalized to the expression of GAPDH mRNA. The specific primers used for detecting mRNA are listed in Supplementary Table S1.

Protein from cells and tissues was extracted using radioimmunoprecipitation assay (RIPA) buffer (Millipore, MA, and United States) with protease and phosphatase inhibitors (Cwbio, Beijing, and China) on ice for 30 min. The nuclear fraction was acquired through nuclear and cytoplasmic protein extraction kit (Beyotime, Shanghai, and China) according to the manufacturer’s instruction. Ultrasonic homogenizers were used to disrupt the RIPA buffer-extracted lysates. Later, the supernatants were harvested by centrifugation for 25 min at 12000 rpm and their concentration was measured by a BCA protein assay kit (Cwbio, Beijing, and China). 4%–20% SDS-PAGE gel (SurePAGE™, GenScript, Nanjing, and China) was used to separate the proteins, which were then transferred to a polyvinylidene fluoride (PVDF) membrane (Millipore, MA, and United States). Blocking was performed with 5% BSA in TBST for 1 h, followed by incubation with indicated primary antibodies at 4°C overnight and secondary antibodies for 1 h at room temperature with mild shaking. Protein bands were visualized using the ECL system (ZEN-BIOSCIENCE, Chengdu, and China) and all images were analyzed with ImageJ software. The antibody information is listed in Supplementary Table S2.

Cell Counting Kit-8 (CCK-8) test was carried out to assess the effect of TRAF -STOP on cell viability. Briefly, BMMs were seeded in 96-well plates at a density of 1 × 10⁴/well in triplicate. After 24 h of incubation, various TRAF-STOP concentrations (0–20 µM) were added to BMM cells for 48 h. Each well was incubated for an additional 2 h with CCK-8 solution (Dojindo Molecular Technology, Kumamoto, Japan) at 37°C under 5% CO₂ and humidified atmosphere. Cell viability was determined by a microplate reader (BioTek, Vermont, United States) at 450 nm.

PLGA-PEG-PLGA triblock copolymers were synthesized through ring-opening copolymerization of D, L-lactide (D, L-LA), and glycolide (GA) with Poly (ethylene glycol) (PEG) as the macroinitiator and Sn (Oct)2 as the catalyst. PEG was purchased from Sigma. D, L-LA, and GA were purchased from Aladdin Reagent (Shanghai) Co., LTD. Later, the crude products were dissolved in cold water (4°C–8°C), and the polymer solution was heated to 80°C to precipitate the polymer. The precipitated polymer was separated from the supernatant by pouring. And the product was freeze-dried.

The composition and structure of the copolymers were examined by 1H NMR spectrum, which was performed on a 400 MHz Bruker spectrometer using CDCl3 as a solvent. Molecular weights (MW) of the copolymer were evaluated by gel permeation chromatography (GPC, Waters) with tetrahydrofuran as eluting solvent at a flow rate of 1.0 mL/min. The MW was calibrated with polystyrene standard. Additionally, the inverted vial technique was used to investigate the sol-gel phase transition. PLGA-PEG-PLGA polymer structure has a certain proportion of hydrophilic and hydrophobic groups. At low temperature, the hydrophilic groups and water molecules dissolve in water due to the strong carbon-oxygen hydrogen bonding force between PLGA-PEG-PLGA molecular chains, at which time PLGA-PEG-PLGA polymer is in solution state. With the increase of temperature, some of the hydrogen bonding forces gradually weakened, while the hydrophobic forces in the PLGA-PEG-PLGA polymer chains continuously increased. When reaching a certain temperature, the polymer chains aggregated with each other and a bulk phase transition occurred, and the PLGA-PEG-PLGA polymer was in the gel state. Therefore, we examined whether the PLGA-PEG-PLGA polymer is temperature-sensitive by the inverted tube method.

All animals used in this study were male C57BL/6J mice aged 6–8 weeks purchased from Sun Yat-sen University (Guangzhou, China). The animal experiments were approved by the Ethics Committee of Sun Yat-sen University (Document No. 2021001578). Mice were randomly and equally divided into three groups (n = 5) as follows: (a) PBS injection; (b) TRAF-STOP injection; (c) TRAF-STOP + hydrogel injection. PBS was used as vehicle control. 5-0 thread was ligated at the submarginal position around the bilateral maxillary second molars and kept for 10 days to induce periodontitis (Abe and Hajishengallis, 2013; Marchesan et al., 2018). For injection of medication, medical sterile insulin syringe (320312) from BD was utilized. The insulin needle is suited for local injection because it is smaller with a diameter of 0.33 mm. Drugs were injected into the central palatal gingiva of the maxillary second molar after ligature removal every 3 days for 1 month.

Maxillae from experimental mice were collected and scanned with high-resolution micro-CT (Scano Micro-CT, μCT50, Switzerland). Image acquisition was performed at 70 kV, 114 mA, and 10-um increments. RadiAnt DICOM Viewer (Medixant, Poznan, Poland) was used to reconstruct three-dimensional images and measure the distance of the cementoenamel junction to the alveolar bone crest (CEJ-ABC distance) for the evaluation of bone loss.

Both the collected human gingiva and the maxillae were fixed in 4% paraformaldehyde (PFA) for 24 h and the latter were decalcified in 0.5 M EDTA for 4 weeks. After being embedded in paraffin, all samples were sectioned to 4–5 μm, followed by hematoxylin and eosin (H&E), tartrate-resistant acid phosphatase (TRAP) staining, immunohistochemistry (IHC) staining as well as immunofluorescence (IF) staining. H&E staining was used to assess bone loss by showing the distance between the cementoenamel junction and alveolar bone crest (CEJ-ABC distance). The TRAP staining was performed using a commercial TRAP kit (TRAP staining kit, Sigma, MO, United States). TRAP-positive cells are stained as dark purple multinucleated cells located on the bone perimeter within a resorption lacuna. Immunohistochemistry (IHC) staining was performed with DAB Detection Kit (Polymer) kit (Gene Tech, Shanghai, China) according to the manufacturer’s instructions. The primary antibodies used were the NFATc1 antibody (1:100, A1539, ABclonal, Wuhan, China), CD40 (1:100, A0218, ABclonal, Wuhan, China), TRAF6 (1:100, A0973, ABclonal, Wuhan, China). After removal of the secondary antibody, slides were dehydrated, hematoxylin immersed, differentiated, and mounted with neutral gum and image. As for IF staining, sections were counterstained with DAPI at room temperature for 5 min and samples were visualized using a confocal laser scanning microscope (LSM780, Carl Zeiss Microscopy, Oberkochen, Germany).

All data are indicated as the mean ± standard deviation. Student’s t-tests were used for the comparison of two groups while one-way ANOVA was used for the comparison of three or more groups. p < 0.05 was considered statistically significant. All statistical analyses were performed using GraphPad Prism.

To investigate the variable expression of CD40 and TRAF6 in healthy individuals and patients with periodontitis, we harvested the gingival samples separately. Initially, to ensure the reliability of the samples, inflammatory cytokines including IL-1β, IL-6, and TNF-α of each group were evaluated by RT-qPCR, all of which are significantly upregulated in periodontitis (Figure 1A). Meanwhile, RT-qPCR data revealed approximately threefold higher CD40 and TRAF6 gene expression in periodontitis compared with the healthy group (Figure 1B). Consistently, immunofluorescence staining also exhibited higher expression of CD40 and TRAF6 in the gingiva with periodontitis (Figure 1C). Western blotting results further confirmed the upregulation of CD40 and TRAF6, both of which showed a nearly 1.5-fold increase in the periodontitis group (Figures 1D, E). Taken together, these data indicated that the CD40-TRAF6 signaling pathway may play a significant role in periodontitis.

To investigate the effects of TRAF-STOP on osteoclastogenesis, BMMs in a medium containing RANKL and M-CSF were exposed to CD40L in the presence and absence of TRAF-STOP. TRAP staining exhibited that CD40L successfully promoted the differentiation of osteoclasts, as indicated by the increase in the number and size of TRAP + cells (Figure 2A). However, the addition of TRAF-STOP attenuated osteoclast formation. The number of TRAP-positive cells declined from 243.1/well (without TRAF-STOP) to 61.6/well (with 1 µM TRAF-STOP). Moreover, both the expression of NFATc1 in cell lysate and nuclear fraction detected by Western blot were consistent with the results above (Figures 2B, C). Similarly, RT-qPCR data revealed the striking upregulation of osteoclast-specific marker genes expression including NFATc1, c-Fos, TRAP, and MMP9 after the treatment of CD40L, which was reversed by TRAF-STOP (Figure 2D). In the meantime, we found that CD40L treatment dramatically increased NF-κB activation (the phosphorylation of P65) with a percentage of seventy by Western blotting, whereas TRAF-STOP treatment resulted in a 60% decrease in NF-κB activation (Figure 2E). Furthermore, the CCK8 assay showed TRAF-STOP had no cytotoxic effects on BMMs at concentrations below 20uM (Figure 2F). Interestingly, we found that TRAF-STOP could inhibit RANKL-induced osteoclast development in the absence of CD40L, and this impact was more pronounced in the group including CD40L as well (Figure 2G). These data suggest that TRAF-STOP treatment in BMMs inhibits NFATc1 expression and NF-κB activation, which reduces osteoclast formation.

A well-formed podosome actin belt was regarded as the mature osteoclast actin cytoskeleton, reflecting osteoclastic bone resorption capacity (Leightner et al., 2020). Thus, immunofluorescence staining of mature osteoclasts was carried out to evaluate the impact of CD40L and TRAF-STOP on the function of the osteoclast (Figures 3A, C). CD40L stimulated the larger and clearer actin belt compared to RANKL-only treated controls. Interestingly, the size and amount of actin ring were decreased under the treatment of TRAF-STOP and only cluster-like organization was detected. As for the pit formation assay (Figures 3B, D), the treatment of CD40L led to 1.7-fold more resorption pits, while TRAF-STOP reduced pit resorption on bovine slices caused by CD40L and RANKL-stimulated osteoclasts, further confirming the inhibitory effect of TRAF-STOP on bone resorption. Collectively, these results indicate TRAF-STOP represses the bone resorption of osteoclasts in vitro.

The PLGA-PEG-PLGA triblock copolymer was synthesized as mentioned above. PLGA-PEG-PLGA hydrogel was in a liquid state at 25°C and converted into non-flowing hydrogels upon exposure to the incubator at 37°C (Figure 4A). The ¹H NMR spectra and GPC were used to establish the copolymer’s chemical composition and molecular weight. As demonstrated in Figure 4C, the ¹H NMR spectrum exhibited characteristic proton signal peaks of the copolymer, and the signals appearing at 5.20 and 4.80 ppm showed LA/GA molar ratio was 13/1. The polymer dispersity index (PDI) determined by GPC read 1.17, and the average MW value of the synthesized copolymer was 6640 (Figure 4B), indicating the sample was successfully synthesized as desired.

The prospective role of TRAF-STOP in the progression of periodontitis in vivo was investigated in the ligature-induced periodontitis model. Distance between CEJ (cement–enamel junction) and ABC (alveolar bone crest) was used to evaluate the degree of bone resorption (Xin et al., 2022). The CEJ-ABC distance was found to be nearly 2-fold higher in the periodontitis group compared to the control group (Supplementary Figure S1), showing that our periodontitis model was successfully established. Micro-CT-based quantitative analyses of the alveolar bone indicated that TRAF-STOP partially alleviated ligature-induced bone resorption and the TRAF-STOP/hydrogel-treated group exhibited an approximately 1.5-fold decrease in distance compared to the PBS group (Figure 5A). Additionally, H&E staining of the PBS-injection group exhibited inflammatory cell infiltration, and severe bone loss, all of which were the manifestations of alveolar bone resorption (Figure 5B). Moreover, the multinucleated giant cell was detected in PBS group, which was recognized as the hallmark of chronic inflammation (McNally and Anderson, 2011). However, this situation was dramatically attenuated by TRAF-STOP, especially in the TRAF-STOP/hydrogel-treated group. Moreover, IL-1β in periodontal tissue was shown the lowest expression level in the TRAF-STOP/hydrogel treatment group (Figure 5C), which was consistent with the mRNA levels of IL-1β, IL-6, and TNF-α (Figure 5D), suggesting the combination of TRAF-STOP and hydrogel might be more effective than TRAF-STOP alone for the treatment of periodontitis.

The reduction of osteoclastogenesis was confirmed by qRT-PCR analysis (Figure 6A). Both experimental groups exhibited a lower expression of osteoclast genes encoding NFATc1, ATP6v0d2, DC-STAMP, and MMP9 compared to the control group, with the greatest reduction in TRAF-STOP/hydrogel. Later, TRAP staining was performed to investigate whether TRAF-STOP/hydrogel prevents ligature-induced bone loss through the inhibition of osteoclastogenic activity in vivo (Figures 6B, F). The number of TRAP-positive multinucleated cells was significantly decreased in TRAF-STOP-injected mice compared to the PBS-injected group, from 9.67 per field of view to 5.67 per field of view. Meanwhile, the numbers of osteoclasts in the periodontium (average of 2 per field of view) were fewest in the TRAF-STOP/hydrogel-treated group. Moreover, the histochemical analysis demonstrated that the expression of NFATc1, the key transcriptional regulator involved in osteoclastogenesis, was distinctly downregulated by TRAF-STOP, and the addition of hydrogel amplified the effect (Figure 6C). In addition, the expression level of the osteoclast marker protein NFATc1 detected by Western blotting was consistent with the above observations, with a statistically 2.6-fold decrease in the TRAF-STOP group and a 3.3-fold reduction in the TRAF-STOP/hydrogel group compared to the PBS group (Figures 6D, G). Furthermore, Western blotting results revealed that the phosphorylation of P65 was visibly inhibited upon exposure to TRAF-STOP and further diminished in the TRAF-STOP/hydrogel group, suggesting that the osteoprotective effect of TRAF-STOP/hydrogel is attributed to the inhibition of NF-κB-mediated osteoclast formation and development (Figures 6E, H).