Isolation and characterization of DPSCs

Healthy and intact premolars for orthodontic therapy from young adults aged 18 to 25 were employed to isolate DPSCs under approved guidelines set by the Ethics Committee of the School of Stomatology, China Medical University, Shenyang, China (G2018009). The donors provided written informed consent, granting permission to use their extracted dental tissues for experimental purposes. In brief, the extracted teeth were exposed to phosphate-buffered saline (PBS) supplemented with 2% penicillin (200 U/mL) and streptomycin (0.2 mg/mL) (PS, HyClone, SV30010, USA). Subsequently, pulp tissue fragments were obtained by mechanically disrupting the pulp cavity structure and further treated with 4 mg/mL type II neutral proteinase (Roche, 04942078001, Switzerland) and 3 mg/mL type I collagenase (Invitrogen, 17100017, USA). DPSCs were cultured in α-MEM medium (Gibco, C12571500BT, USA) supplemented with 20% fetal bovine serum (FBS, Clark, FB15015, USA) and 1% PS (penicillin: 100 U/mL and streptomycin: 0.1 mg/mL) until the primary cells migrated from the tissue fragments. Cells in passages three to six were used in subsequent experiments. The stemness capabilities of DPSCs were confirmed through flow cytometric analysis. Furthermore, their diverse differentiation potential was validated using alizarin red S (ARS) staining (Sigma, A5533, USA), oil red O staining (Beyotime, C0158S, China), and alcian blue chondrocyte staining (Cyagen, ALCB-10001, China) after induction 21 days as previously described [31, 32].

Cell treatment and transfection

DPSCs were cultured in α-MEM supplemented with 10% FBS and 1% PS. When the cell confluence reached approximately 40–50%, transfection was performed following the manufacturer's instructions (Genechem, HiTransG, China). Briefly, DPSCs were transfected with lentiviruses at a multiplicity of infection (MOI) of 25, and the culture medium was replaced after 12 h. To establish stable cell strains, DPSCs were exposed to a culture medium containing 2 μg/mL puromycin for 48 h. The surviving DPSCs were further cultured and utilized in subsequent experiments.

To assess autophagic flux in DPSCs during odontoblastic differentiation, 5 μM rapamycin (Rapa, APExBIO, A8167, USA), or 50 nM bafilomycin A1 (Baf A1, APExBIO, A8627, USA) was added to the odontoblastic medium (OM). DPSCs were cultured following the cycle of “2-day treatment with autophagy inducers or inhibitors 3-day culture without autophagy modifiers”.

To achieve the stable overexpression of ATF4 or BNIP3, as well as the knockdown of BNIP3, lentiviruses carrying oe-ATF4 (Ubi-MCS-3FLAG-SV40-EGFP-IRES-puromycin), oe-BNIP3 (Ubi-MCS-3FLAG-SV40-puromycin), kd-BNIP3 (hU6-MCS-CMV-Puromycin), and their respective negative controls were procured from Genechem, China. The ATF4-WT vector/ATF4-mutant vector, siBNIP3, or shKPNB1 with its negative control respectively were transfected using the jetPRIME transfection reagent (Polyplus, 101,000,046, France). The target sequences of BNIP3 small interference were (RIBOBIO, China):

1#-ACACGAGGTCATGAAGAA; 2#-GTTCCAGCCTCGGTTCTA; 3#-GAACTGCACTTCAGCATA.

For the evaluation of mitophagy, DPSCs were co-transfected with adenovirus Ad-GFP-LC3 and HBAD-Mito-dsRed (Hanbio, China). Mitophagy was visualized and recorded using a laser confocal microscope.

Alkaline phosphatase (ALP) and ARS staining

To initiate odontoblastic differentiation, DPSCs were cultured in DMEM (Gibco, C11995500BT, USA) supplemented with 10% FBS, 1% PS, 50 μg/mL L-ascorbic acid (Sigma, A4403, USA), 10 mM β-glycerophosphate (Sigma, G9422, USA), and 10 nM dexamethasone (Sigma, D4902, USA). The induction medium was replaced every three days. After 7 days of induction, the cells were fixed with 4% paraformaldehyde (PFA) and stained using the BCIP/NBT alkaline phosphatase kit (Beyotime, C3206, China).

After 21 days of induction, the cells were fixed with 4% PFA and stained using a 0.5% ARS solution (pH = 6.5) prepared by dissolving ARS powder in ddH2O. Subsequently, the optical density (OD) value of the stained cells was measured by dissolving them in 10% cetylpyridinium chloride (Aladdin, H108697, China) at 562 nm.

Immunofluorescence staining

For immunofluorescence staining, the cells were fixed with 4% PFA and permeabilized with 0.5% Triton X-100. Then, the cells were blocked with 1% BSA for 30 min at room temperature. The primary antibodies anti-ATF4 (1:100, CST, 11815, USA) and anti-KPNB1 (1:200, Abcam, ab2811, UK) were incubated with the cells overnight at 4 °C. The goat-anti-mouse or goat-anti-rabbit secondary antibodies (1:500, Proteintech, SA00013-3 or SA00013-2, China) were added and incubated for 2 h at 37 °C. The images were captured using an inverted microscope (Leica, Japan).

Bioinformatics analysis and quantitative real-time polymerase chain reaction (RT‒qPCR)

RNA-seq data (GSE138179) of DPSCs and DPSCs undergoing odontoblastic differentiation were retrieved from the GEO database. Differentially expressed gene (DEG) analysis was performed using the “edgeR” package with the following criteria: |logFC|> 0.58 and P < 0.05. The results are presented in the form of volcano plots and were visualized using the “ggplot” package. Gene Ontology (GO) analysis for the identified pathways enriched in DEGs was performed using the “clusterProfiler” package.

Total RNA was extracted and subjected to reverse transcription following the protocols provided by the reverse kit (Takara, RR047A, Japan). Real-time quantitative polymerase chain reaction (RT‒qPCR) was carried out utilizing the SYBR Green kit (Qiagen, 208,504, Germany) and detected using the ABI 7500 system (Applied Biosystems, USA). Supplementary File 1 contains the list of primer sequences utilized in the experiments. The relative mRNA levels of each sample were calculated using the 2(−ΔΔCt) method, with GAPDH expression as the normalization control.

Subcellular fractionation and western blots

Nuclear and cytoplasmic proteins of DPSCs were extracted following the instructions provided by the kit from Thermo Fisher, AM1921, USA. For the isolation of mitochondrial proteins, DPSCs mitochondrial protein extraction was conducted according to the protocol supplied by the kit from Beyotime, C3601, China. Total protein was lysed using radioimmunoprecipitation assay (RIPA) lysis buffer supplemented with 1% PMSF.

Western blotting was carried out to assess the relative expression of the proteins, as previously described [30]. The primary antibodies utilized were as follows: ATF4 (1:1000, CST, 11,815, USA), LC3B (1:1000, Abcam, ab51520, USA), P62 (1:1000, CST, 5114, USA), KPNB1 (1:2000, Abcam, ab2811, USA), FLAG (1:1000, CST, 2368, USA), BNIP3 (1:1000, CST, 44060, USA), DMP1 (1:1000, Signalway, 38779, USA), and DSPP (1:200, Santa, SC-73632, USA). GAPDH (1:10000, Proteintech, HRP-60004, China) served as the cytoplasmic internal reference, Histone3 (1:5000, Affinity, BF9211, China) served as the nuclear internal reference, and COX IV (1:10000, Proteintech, 66110–1-Ig, China) served as the mitochondrial internal reference. Subsequently, the membranes were incubated with the appropriate secondary antibodies (1:10000, Proteintech, SA00001-1 or SA00001-2, China). Immunoreactive protein bands were visualized using enhanced chemiluminescence (Tannon, 180-5001, China), and band quantification was conducted within the linear range. The relative expression of the proteins was measured using ImageJ software.

MitoSOX staining and flow cytometry

The levels of mitochondrial reactive oxygen species (mtROS) in DPSCs were assessed utilizing MitoSOX (Yeasen, 40778ES50, China). DPSCs were cultured in a medium containing 3 μM MitoSOX and incubated at 37 °C in the dark for 30 min. Subsequently, the cells were washed twice with PBS, and the relative fluorescence intensity of MitoSOX was measured using an inverted microscope and a flow cytometer.

Chromatin immunoprecipitation PCR (ChIP‒PCR) assays

ChIP assays were conducted using the ChIP assay kit from Beyotime, P2078, China. After inducing DPSCs towards odontoblastic differentiation for 3 days, the cells were cross-linked with 1% PFA to a final concentration. Subsequently, they were lysed with lysis buffer and sonicated on ice, with each sonication cycle lasting 10 s and repeated ten times. The input group was stored at −80 °C, while the immunoprecipitation (IP) of sonicated chromatin solutions was carried out by incubating with anti-ATF4 (2 μg, CST, 11815, USA) or IgG (2 μg, Beyotime, A7016, China) antibodies overnight at 4 °C. Protein A + G agarose beads were then added, followed by sequential washes with low salt buffer, high salt buffer, LiCl buffer, and TE buffer. The purified DNA was resuspended in TE buffer, and the protein‒DNA complexes were eluted with the reversal of cross-linking. DNA samples were purified and subsequently amplified by PCR (Takara, RR001A, Japan). A 2% agarose gel electrophoresis was performed, and the results were visualized through ultraviolet fluoroscopy. The primer sequences for the ATF4 binding sites of BNIP3 are provided in Supplementary File 2.

Luciferase reporter assay

The luciferase reporter assay was performed following the protocol provided in the dual-luciferase reporter gene assay kit from Yeasen (11402ES). In brief, HEK 293 T cells were cotransfected with various plasmids, including ATF4-vector, wild-type (WT) vector of the BNIP3 promoter, mutant 1 (MUT1) vector of the BNIP3 promoter, mutant 2 (MUT2) vector of the BNIP3 promoter, mutant 1 and 2 (MUT1 and 2) vector of the BNIP3 promoter, and negative control plasmids. Additionally, the pGL-HRE-AdML reporter plasmid and the pRL-tk vector served as the internal normalization controls (Taihe Biotechnology, China).

Coimmunoprecipitation (Co-IP) assay and mass spectrometry assay

For Co-IP in DPSCs, the Pierce Crosslink Magnetic IP/Co-IP Kit (Thermo Fisher, 88805, USA) was employed, following the manufacturer's instructions. In brief, primary antibodies or isotype control IgG were covalently cross-linked to 25 μL of protein A/G magnetic beads. Proteins were extracted using IP lysis/wash buffer supplemented with PMSF and a protease and phosphatase inhibitor cocktail. A portion of the sample was saved as an input for later comparison. An equal amount (1 mg) of each protein extract was incubated with protein A/G magnetic beads cross-linked with primary antibodies or isotype control IgG overnight at 4 °C. The beads were then washed with elution buffer and neutralization buffer to obtain the protein. The samples were subsequently analyzed by western blotting, as described in the section Subcellular fractionation and western blots.

For the liquid chromatography tandem mass spectrometry (LC–MS/MS) assay, Co-IP was performed using the Pierce Classic Magnetic IP/Co-IP kit (Thermo Fisher, 88804, USA) under the manufacturer’s instructions. In this case, 1 mg of protein lysate was incubated with 10 mg of primary antibodies or isotype control IgG overnight at 4 °C. Pierce Protein A/G Magnetic Beads were washed three times using Pierce IP Lysis/Wash Buffer for LC–MS/MS analysis. LC–MS/MS sequencing was conducted by BIOPROFILE (Shanghai, China).

Oxygen consumption rate (OCR)

The OCR was assessed by the Seahorse Cell Mito Stress Test in real-time using the 24-well Extracellular Flux Analyser XF-24 (Agilent, 103,015, USA). Briefly, DPSCs were seeded in each well of the microplate, and the transfected cells were incubated overnight in 200 µl of medium under different conditions. The microplates were then equilibrated without CO2 for 1 h before measurement. Subsequently, the ATPase inhibitor oligomycin (Oligo, 1 µM), the uncoupling reagent carbonyl cyanide-P-trifluoromethoxyphenylhydrazone (FCCP, 1 µM), and inhibitors of the electron transport chain rotenone/antimycin (R/A, 2 µM) were sequentially injected during real-time measurements of the OCR, whereas respiratory parameters were measured after each injection. The following respiratory parametes were assessed: basal cellular respiration: difference between the last rate measurement before the oligo injection and non-mitochondrial respiration (minimum rate measurement after the R/A injection); maximal respiration: differences between the maximum rate measurement after FCCP injection and the non-mitochondrial respiration; spare respiratory capacity: differences between the maximum respiration and basal respiration; ATP production: difference between the final rate measurement before the oligo injection and the minimum rate measurement after the oligo injection; and proton leak: difference between minimum rate measurement after oligomycin injection and nonmitochondrial respiration.

Animal experiments

To investigate the impact of BNIP3 on DPSCs odontoblastic differentiation in vivo, we used a tooth fragment in vivo mouse model. All animal procedures were approved and conducted following the guidelines of the Institutional Animal Care Committee of China Medical University (CMU2022014) for the odontoblastic differentiation animal model. Single-root human teeth were prepared as follows: The tooth roots were divided into short segments (~ 5 mm), followed by cleaned and shaped using rotary instruments. All fragments were treated with 17% ethylenediaminetetraacetic acid (EDTA) for 10 min and 5.25% NaClO for 15 min to eliminate the organic component, debris, and microorganisms. All segments were incubated in a culture medium at 37 °C before use. DPSCs transfected with non-targeting control (nc-oe-BNIP3), BNIP3 overexpression (oe-BNIP3), nontargeting control (nc-kd-BNIP3), or BNIP3 knockdown (kd-BNIP3) were encapsulated in 6% EFL-GM-PR hydrogel (EFL, EFL-GM-PR, China) and injected into the root canals. After overnight incubation with culture medium, four root segments (nc-oe-BNIP3; oe-BNIP3; nc-kd-BNIP3; kd-BNIP3) were transplanted subcutaneously into each dorsum of the male nude mouse (a total of 6 nude BALB/c mice, 8 weeks old, weighing 18–22 g, male). After 8 weeks, the samples were retrieved, fixed with 4% PFA for 24 h, and demineralized with EDTA for 8 weeks. Haematoxylin–eosin (HE), immunofluorescence (IF), and immunohistochemistry (IHC) analyses were performed on 4-μm thick paraffin sections. HE staining assay was undergone according to the manufacturer's instructions (Solarbio, G1120, China). IHC staining was performed using an IHC kit (Gene Tech, RR001A, China) procedure to detect DSPP and DMP1 in paraffin-embedded tooth fragment sections. The primary antibodies were DSPP (1:100, Santa, SC-73632, USA) and DMP1 (1:200, Signal way, 38779, USA) followed by secondary antibodies, and counterstained with hematoxylin. Images were acquired and quantified immunohistochemistry was determined by the relative intensity of the positive area using Image-J. IF staining was similarly to IHC, the section was incubated with primary antibody: BNIP3 (1:100, CST, 44060, USA), ATF4 (1:100, CST, 11815, USA), DSPP (1:100, Santa, SC-73632, USA), or DMP1 (1:250, Santa, SC-81249, USA) followed by fluorescent secondary antibody, and DAPI-labeled nucleus (Beyotime, P0131, China).

Statistical analysis

Statistical analyses were conducted using R language (version 3.5) and GraphPad Prism 7 software. The data are presented as the means ± standard deviation (SD). Two-group comparisons were conducted using unpaired two-tailed Student’s t-tests. Multiple-group comparisons were analysed using one-way ANOVA or two-way ANOVA, followed by Dunnett's multiple comparisons, Tukey’s multiple comparisons tests, or Šidák’s multiple comparisons tests. Nonparametric tests were applied if the data showed a skewed distribution. Each assay was independently performed at least three times. A significance level of P < 0.05 was considered statistically significant.