Healthy permanent premolars and third molars were collected from donors aged 18 to 25 for orthodontic reasons from the Department of Oral and Maxillofacial Surgery, Guanghua School of Stomatology, Sun Yat-sen University for approximately one year between March 2018 and April 2019. Only healthy teeth without carious disease or hyperemic pulp tissue were selected. A total of 128 teeth from 58 donors (29 males and 29 females) were obtained for dental pulp tissue isolation and cell culture. hDPCs were isolated and cultivated using an enzymatic method as described by Gronthos et al (22). After extraction, the teeth were washed with 70% ethanol and PBS (pH 7.4) and then split open to expose the pulp chamber. The dental pulp tissue was gently isolated with forceps and minced into small pieces, which were then digested in 3 mg/ml collagenase type I (Gibco; Thermo Fisher Scientific, Inc.) for 20 min at 37°C. Subsequently, the minced pulp tissue was cultured in DMEM containing 20% FBS, 100 mg/ml streptomycin and 100 U/ml penicillin (all purchased from Gibco; Thermo Fisher Scientific, Inc.) at 37°C with 5% CO₂. The medium was changed every 3 days. When the cells reached 80% confluence, they were detached using trypsin/EDTA (Gibco; Thermo Fisher Scientific, Inc.) and subcultured at a ratio of 1:2. Generally, 2–3 teeth from one donor were used for each primary culture. For each primary culture, ~10⁶ cells at the zero passage were obtained. All experiments were performed with cells from passages two or three. To avoid inter-individual variation, the experiments were performed at least three times for each sample and each experiment, and average data were generated. For each parameter, experiments were replicated three times each using donor cells from three samples, and average data for the three different cell types were obtained.

hDPCs were stimulated for the indicated times (0, 3, 6, 12 and 24 h) with 1 µg/ml purified Escherichia coli (E. coli) LPS (Sigma-Aldrich; Merck KGaA) at 37°C with 5% CO₂ (4,23). The blank controls were cells without LPS stimulation.

siRNA was used in hDPCs to knockdown DNMT1. A total of 3 siRNA sequences (Invitrogen; Thermo Fisher Scientific, Inc.) were designed to target the human DNMT1 gene. Before transfection, hDPCs were seeded in 6-well plates in 2 ml of α-MEM at 4×10⁵ cells/well containing 10% FBS for 24 h. After attachment overnight, hDPCs were then transfected with siRNA (50 nM) against DNMT1 or a nontargeting siRNA control using Lipofectamine^® 3000 (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocols. After incubation for 24 h, the media was changed, and DMEM supplemented with 10% FBS was added with or without 1 µg/ml E. coli LPS. All siRNA sequences are listed in Table I. siRNA #1 with the best interference effect was selected as the DNMT1 target sequence for the subsequent experiments.

Cells were lysed using TRIzol^® reagent (Invitrogen; Thermo Fisher Scientific, Inc.) following the manufacturer's protocols, and RNA was extracted and reverse transcribed into cDNA with a RevertAid First Strand cDNA Synthesis kit (Fermentas; Thermo Fisher Scientific, Inc.). PCR was performed using the complementary DNA as a template. SYBR-Green I (Roche Diagnostics) RT-qPCR results were detected by a LightCycler^® 480 thermal cycler. Thermal cycling conditions consisted of initial denaturation at 95°C for 5 min, followed by 45 cycles of 95°C for 10 sec, 65°C for 20 sec and 72°C for 30 sec. The relative results were normalized to the GAPDH mRNA levels (24). The primer sequences were designed using Primer Express Software v3.0.1 (Thermo Fisher Scientific, Inc.) and are listed in Table II.

Protein was extracted from hDPCs using RIPA buffer (Beyotime Institute of Biotechnology), and the concentrations were detected using a BCA Protein Assay kit (Beyotime Institute of Biotechnology). Proteins (30 µg) were separated using electrophoresis on 8% SDS-polyacrylamide gels and transferred to PVDF membranes (EMD Millipore). Next, the membranes were blocked with TBS-Tween 20 (20 mmol Tris-HCl, 150 mmol NaCl, 0.05% Tween-20) containing 5% BSA (Biofroxx; neoFroxx GmbH) for 1 h at room temperature. Then, the membranes were incubated with primary antibodies against DNMT1 (1:2,000; Abcam), IκB kinase αβ (IKKαβ), phosphorylated (p)-IKKαβ, p65, p-p65, IκBα, p-IκBα (1:1,000; NF-κB Pathway Sampler kit, 9936, Cell Signaling Technology, Inc.), p38, ERK, JNK (1:1,000; MAPK Family Antibody Sampler kit, 9926, Cell Signaling Technology, Inc.), p-p38, p-ERK, p-JNK (1:1,000; phospho-MAPK Family Antibody Sampler kit, 9910, Cell Signaling Technology, Inc.) and GAPDH (1:1,000; Cell Signaling Technology, Inc.) overnight at 4°C. After rinsing, the membranes were incubated with horseradish peroxidase-conjugated secondary antibodies (1:2,000; AQ160P and AP307P, EMD Millipore) at room temperature for 1 h. An enhanced chemiluminescence system (EMD Millipore) was used to visualize the antibody binding. The relative protein expression levels were normalized to that of the GAPDH gene, and the protein band densities were determined by ImageJ v1.47 software (National Institutes of Health). ReBlot Plus (EMD Millipore) was used to strip and re-probe with the p-antibodies for IκBα, p38, ERK and JNK to distinguish different target proteins when they share similar molecular weights with the total IκBα, p38, ERK and JNK, respectively, on the same membrane.

Human IL-6 ELISA kits (D6050, R&D Systems, Inc.) and Human IL-8 ELISA kits (D8000C, R&D Systems, Inc.) were used to analyze the culture supernatant protein concentrations of IL-6 and IL-8 collected after LPS stimulation for 6 h according to the manufacturer's protocols. A microplate reader (Tecan Safire microplate reader; Tecan Group, Ltd.) was used to evaluate the optical density (OD) values. Based on the standard solution concentration and corresponding OD value, sample concentrations were calculated.

DNA was extracted from hDPCs and fragmented to 200–500-bp fragments with a Bioruptor Waterbath Sonicator (8 cycles, 15 sec on/15 sec off, at the highest output level while cooling the tube to 1°C in a waterbath). Then, the DNA fragments were diluted to 700 µl with TE buffer (Invitrogen; Thermo Fisher Scientific, Inc.) with 60 µl Protein G Magnetic Beads (S1430S; New England BioLabs, Inc.) and denatured for 10 min at 94°C. Following denaturation, DNA was immunoprecipitated at 4°C overnight with an anti-5mC antibody (1:40, C02010031; Diagenode SA). Then it was incubated for 2 h with anti-IgG Magnetic Beads (S1430S; New England BioLabs, Inc.) at 4°C with agitation. The beads were trapped on a magnetic rack, the supernatant discarded, and washed three times with 1 ml 1XIP buffer [2 mM EDTA, 20 mM Tris (pH=8.0), 1% Triton X-100, 0.1% SDS, 150 mM NaCl] for 10 min at 4°C with agitation. Beads were then resuspended in 400 µl of Elution Buffer (50 mM Tris-HCl, pH=8.0; 10 mM EDTA, Ph=8.0; 1% SDS) with 10 µl of Proteinase K (Qiagen GmbH). IP with non-specific human IgG was measured as a negative control. After IP, the DNA samples were eluted using phenol-chloroform and precipitated using ethanol. After resuspending the precipitated samples in 10 µl Tris buffer, RT-qPCR was performed using 1 µl harvested DNA fragments. The primers designed for MeDIP-PCR are shown in Table III.

All experiments were carried out at least three times. The data were analyzed by the SPSS 20.0 software package (IBM Corp.) and are shown as the mean ± SD. Student's t-test was used to measure the differences between two groups. To evaluate the differences in multiple sets of data, one-way ANOVA or repeated-measures ANOVA with a post hoc Dunnett's test was performed. P<0.05 was considered statistically significant.

To detect the effect of LPS on the inflammatory reaction in hDPCs, hDPCs were stimulated with LPS at a concentration of 1 µg/ml for the indicated times. As illustrated in Fig. 1A and B, compared with the control group, IL-6 and IL-8 mRNA and protein expression was significantly increased by LPS. IL-6 and IL-8 expression was upregulated and peaked after 3 h, which was followed by a gradual decrease. The levels of DNMT1 mRNA were significantly reduced within 24 h after treatment with LPS. DNMT1 protein expression also decreased, with the most significant change at 3 h (Fig. 1C and D). Moreover, the mRNA expression of DNMT3a and DNMT3b did not change significantly before or after LPS treatment (Fig. 1).

Our preliminary study found that 5-Aza-CdR, a DNMT inhibitor, can increase the secretion of inflammatory cytokines in LPS-stimulated hDPCs, and among the upregulated cytokines, IL-6 and IL-8 experienced the greatest increase (unpublished data). To investigate the effect of DNMT1-dependent methylation on inflammatory cytokine production in hDPCs stimulated with LPS, the IL-6 and IL-8 expression levels after DNMT1 knockdown in hDPCs transfected with siRNAs were measured. As shown in Fig. 2A and B, after DNMT1 siRNA (#1, #2 and #3) interference, DNMT1 mRNA expression levels were significantly reduced when compared to the negative control group. These data were further confirmed by western blotting, which showed a reduction in the protein expression. Among the DNMT1 siRNAs, the siRNA #1 group showed the best interference effect at ~72% (Fig. 2B). Therefore, siRNA #1 was selected as the DNMT1 target sequence for the subsequent experiments.

IL-6 and IL-8 gene expression levels were then measured in cells stimulated by LPS after DNMT1 depletion (Fig. 2C). The results showed that IL-6 and IL-8 mRNA expression within 24 h after LPS stimulation was notably higher in the DNMT1 knockdown group compared with the control group. In LPS-inflamed hDPCs, the protein levels of IL-6 and IL-8 were also significantly increased after DNMT1 knockdown (Fig. 2D).

One of the most important signaling pathways that influences inflammatory cytokine production in inflammation induced by LPS is the NF-κB signaling pathway (25). By means of western blotting, the phosphorylation levels of three crucial proteins of the NF-κB signaling pathway were examined (IKKα/β, p65 and IκBα) to determine whether DNMT1 is engaged in NF-κB pathway activation. As illustrated in Fig. 3A and B, DNMT1 knockdown significantly increased the phosphorylation of IKKα/β at 15 and 30 min after LPS treatment in hDPCs. The p65 and IκBα phosphorylation levels also increased at several time points, but there was no significant difference.

Another vital signaling transduction pathway involved in inflammation in the LPS-related inflammatory response is the MAPK signaling pathway (26). The phosphorylation levels of three key proteins in the MAPK signaling pathway were assessed (p38, ERK1/2 and JNK) to determine whether DNMT1 plays an important role in MAPK signaling pathway activation. As illustrated in Fig. 4A and B, after DNMT1 knockdown in LPS-inflamed hDPCs, the p38 phosphorylation level was increased, while both p-ERK and p-JNK levels were not significantly altered.

DNA methylation can regulate the occurrence and progression of inflammatory responses by modulating the methylation levels of inflammation-related cytokines and signaling molecule promoters (27). TRAF6 and MyD88 are key intracellular signal transducers of LPS-induced signaling pathways (28). To identify whether the methylation of IL-6, IL-8, TRAF6 and MyD88 was regulated through DNMT1, the levels of 5mC present at their gene promoters were examined by means of MeDIP-PCR. The results illustrated that the levels of 5mC at the IL-6 and TRAF6 promoters decreased notably in LPS-stimulated hDPCs after DNMT1 knockdown. However, no significant change was observed in the 5mC levels of the IL-8 and MyD88 promoters (Fig. 5). These experimental results indicated that DNMT1 can modulate the methylation of IL-6 and TRAF6 in hDPCs stimulated by LPS.