KLF6-mediated glutamine metabolism governs odontogenic differentiation and matrix mineralization of dental pulp stem cells

Abstract

Background

When a tooth suffers severe injuries, dental pulp stem cells migrate and differentiate into odontoblast-like cells to form restorative dentin. Kruppel-like factor 6 (KLF6) activates the odontoblast differentiation of dental papilla cells during tooth development. However, the mechanisms by which KLF6 regulates the function of odontoblast-like cells differentiated from human dental pulp stem cells (hDPSCs) remain unknown.

Methods

KLF6 was over-expressed or silenced by lentivirus transfection. Transcriptome sequencing and metabolomics were performed to reveal main changes in KLF6 high expressed hDPSCs. Mitochondrial morphology was observed by confocal microscope and cryo-transmission electron microscopy. Metabolic assays and metabolic flux were used to determine changes in cellular metabolic characteristics. Glutamine, glutamate, α-KG, and citrate concentrations were detected in cultured cells. Citrate and Ca concentration were detected in ECM. Adeno-associated virus were used to silence KLF6 in mice. A mouse dental injury model was established to investigate the role of KLF6 and glutamine metabolism in dentin repair in vivo.

Results

RNA sequencing and metabolomics showed a remarkable influence on glutamine metabolism, mitochondrial respiration, and the TCA cycle by KLF6 overexpression. Metabolic assays and mitochondrial morphology observation found KLF6 promoted glutamine metabolism and mitochondrial function, and glutamine metabolism and mitochondrial respiration are enhanced during odontogenic differentiation of hDPSCs. Deprivation of glutamine inhibited mineralization of hDPSCs and restrained deposition of citrate and Ca in ECM. Increased glutamine entry into the tricarboxylic acid (TCA) cycle was both observed in differentiated hDPSCs and KLF6 overexpressed hDPSCs. ChIP-qPCR experiments revealed that KLF6 can directly bind to the promoter sequences of GLS1 and GDH. Supplementation of α-KG rescued suppression of odontogenic differentiation and mineralization induced by KLF6 knockdown. Inhibition of glutamine metabolism and knockdown of KLF6 attenuated tertiary dentin formation in vivo.

Conclusions

Our study shows that KLF6 mediates biomineralization in the newly generated functional odontoblast-like cells differentiated from hDPSCs by altering cell metabolism preferences. KLF6 facilitated glutamine influx into the TCA cycle, leading to increased deposition of citrate in the ECM.These findings may inspire the development of novel strategies for reparative dentin formation.

Background

When dental tissue is attacked by caries, erosion, or abrasion, the defense and repair of the dentin–pulp complex rely on the defensive ability and self-renewing capacity of the dental pulp [1]. Odontoblasts, located at the junction of the dentin and pulp, are the first line of defense against various stimuli [2]. After receiving external stimuli, odontoblasts accelerate Ca salt deposition to seal the dentinal tubules and secrete the dentin matrix to form reactionary dentin [3]. When a tooth suffers severe injuries that lead to odontoblast death, dental pulp stem cells migrate toward the injured sites and differentiate into odontoblast-like cells to form restorative dentin. An important characteristic of odontoblasts is their ability to produce the extracellular matrix (ECM) and mediate Ca salt deposition. However, how human dental pulp cells (hDPSCs) respond to elevated energy and synthetic needs as the cells differentiate is unclear.

Kruppel-like factor 6 (KLF6), a member of the zinc-finger transcription factor family, demonstrates widespread expression in odontoblasts throughout different stages in the murine tooth germ, indicating its important involvement in odontogenesis [4]. KLF6 functions as a “switch” in the regulation of differentiation of various cells. Klf6 knockout mice cannot survive embryonic day 12.5 due to markedly reduced hematopoiesis and disorganized vascularization [5]. Klf6 functions in liver development because Klf6^−/− mouse embryonic stem cells lose their ability to undergo hepatic differentiation [6]. KLF6 serves as a toggle in the differentiation of fibro/adipogenic progenitors in muscle degeneration [7]. Moreover, our previous study verified that KLF6 activates odontoblast differentiation in mouse dental papilla cells (mDPCs) and directly activates the transcription of dentin ECM proteins, including Dmp1 and Dspp [8]. KLF6 also regulates p21 transcription to control the transition of cell proliferation and differentiation of mDPCs, and it is expressed in odontoblast-like cells adjacent to reparative dentin in rat pulp exposure models [4]. However, whether KLF6 has similar impacts on the odontogenic differentiation of hDPSCs and its regulatory role in the function of odontoblast-like cells at the late differentiation stages remain unclear.

Recent studies have shown that KLF6 can be involved in the regulation of cell metabolism. Saiful et al. reported a KLF6-mediated transcriptional regulatory network that was able to regulate lipid homeostasis in clear cell renal cell carcinoma [9]. KLF6 directly promoted LPS-induced HIF1α expression and cellular glycolysis under hypoxic conditions in macrophages [10]. Glutamine metabolism has been reported to be a critical regulator of cell fate [11]. And glutamine is the major source of citrate and a precursor of lipid production [12]. Recently, a study reported that the loss of Klf6 suppresses glutamine-derived ornithine levels in pancreatic cancer [13]. However, no studies have reported an association between KLF6 and glutamine metabolism during stem cell differentiation.

In this study, we demonstrate that KLF6 is a critical regulator of glutamine metabolism, which supports the increased energy and biosynthetic requirements during the odontogenic differentiation of hDPSCs. In particular, KLF6 facilitates glutamine undergoing anaplerotic reactions to generate α-ketoglutarate (α-KG), which enters the tricarboxylic acid (TCA) cycle, leading to increased deposition of citrate in the ECM, thereby promoting matrix mineralization. Our data reveal that KLF6 mediates biomineralization in the newly generated functional odontoblast-like cells differentiated from hDPSCs by altering cell metabolism preferences.

Materials and methods

The work has been reported in line with the ARRIVE guidelines 2.0.

Isolation and culture of hDPSCs

Human DPSCs were isolated from the dental pulp of extracted caries-free third molars obtained with the approval of the Ethics Committee of the Stomatology Hospital, Zhejiang University School of Medicine, China (No. 2023-055 (R)). All teeth used in this study were obtained from healthy young patients (aged 18 to 25 years old) who underwent tooth extraction for impacted teeth or orthodontic reasons at the Department of Oral and Maxillofacial Surgery, Stomatology Hospital, Zhejiang University School of Medicine, China. The freshly extracted teeth were transported in Dulbecco’s modified Eagle medium (DMEM) at 4 ℃. Dental pulp tissues were acquired and cut into pieces (1–2 mm³) under non-sterile conditions. The pieces were placed in culture bottles to obtain cells that outgrew the margins of the pulp fragments. DPSCs were isolated as previously reported [14]. Cells were cultured in DMEM containing 10% FBS and 1% P/S and passaged using 0.25% trypsin at 80–90% confluence. Only the third to fifth passages were used in further experiments.

Odontogenic induction and alkaline phosphatase (ALP) and Alizarin red S (ARS) staining

Cells were seeded in 6-well plates at 5 × 10⁵ cells/well or in 12-well plates at 1 × 10⁵ cells/well. For the induction of odontogenic differentiation, hDPSCs were cultured in DM (DMEM supplemented with 50 µg/mL ascorbic acid, 10 mM b-glycerophosphate, and 0.1 µM dexamethasone). Alkaline phosphatase (ALP) staining was performed using a BCIP/NBT ALP color development kit (Beyotime, Shanghai, China). ALP activity was quantified using a LabAssay ALP kit (#291-58601, Wako Pure Chemical Industries, Tokyo, Japan). ALP activity was normalized to protein concentration. Cultured cells were stained for 10 min with 2% alizarin red S (ARS; Beyotime).

Lentivirus transfection

The KLF6 expression lentivirus and non-loaded lentivirus were constructed by GenePharma (Suzhou, Jiangsu, China), based on the LV5 shuttle plasmid. Cells were collected and seeded into 6-well plates at 1 × 10⁵ cells/well or 12-well plates at 2 × 10⁴ cells/well. Transfection was conducted at a high multiplicity of infection (MOI = 50). Cells transfected with LV5-KLF6 served as the KLF6 high expression group, whereas cells transfected with LV5-NC served as the negative control group.

For loss-of-function experiments, three KLF6 shRNA and control shRNA lentiviruses were constructed based on the LV3 shuttle plasmid (GenePharma). Cells were transfected in the same manner. The most efficient KLF6 shRNA lentivirus was selected for subsequent experiments. For acquisition of stably transfected cells, a selection process with 2 µg/mL puromycin was carried out for 72 h. For inhibition of glutamine metabolism, cells were treated with 20 µM BPTES (Sigma-Aldrich, St. Louis, MO, USA) or equal volumes of DMSO (Sigma-Aldrich). Rescue experiments were performed with 10 µM cell-permeable dm–α-KG (Sigma-Aldrich).

RT-qPCR

Cellular RNA was extracted using TRIzol reagent (Invitrogen, Waltham, MA, USA). Subsequently, reverse transcription was conducted using SuperScript II Reverse Transcriptase (Invitrogen). RT-qPCR was conducted on an ABI 7500 Real-time PCR system (Thermo Fisher Scientific, Waltham, MA, USA) using SYBR Green (Invitrogen). Each experiment was triplicated in in 384-well plates. The relative changes in gene expression levels were calculated using the formula 2^−ΔΔCT, and the β-actin mRNA level was used as an internal reference control. Primers used are listed in Table S1.

Western blot

Cells were lysed with RIPA lysis buffer containing 1 mM PMSF. The protein concentration was measured with a BCA Protein Assay Kit (Pierce Biotechnology, Rockford, IL, USA). Equal amounts of proteins were subjected to SDS-PAGE and transferred onto a polyvinylidene fluoride membrane. Following the blocking step with 5% skimmed milk powder, the membranes were hybridized overnight with the appropriate primary antibody, such as KLF6 antibody (1:2000, #14716-1-AP, Proteintech, Wuhan, China), glutaminase 1 (GLS1) antibody (1:2000, #49363, Cell Signaling Technology, Danvers, MA, USA), glutamate dehydrogenase (GDH) antibody (1:1000, #12793, Cell Signaling Technology), DMP1 antibody (1:1000, Invitrogen), DSPP antibody (1:2000, #ab216892, Abcam, Waltham, MA, USA), RUNX2 antibody (1:2000, #ET1612-14, HUABIO, Hangzhou, China), COL1A1 antibody (1:1000, #ET1609-68, HUABIO), ALP antibody (1:5000, #ET1601-21, HUABIO), or β-ACTIN antibody (1:2000, #81115-1-RR, Proteintech). The next day, the membranes were incubated with a secondary antibody (1:10000, #SA00001-2, Proteintech) for 1 h. The bands were visualized using an Omni-ECL Femto Light Chemiluminescence Kit (Epizyme Biomedical Technology, Shanghai, China).

Transcriptome sequencing and metabolomics

Passage 3 hDPSCs were collected and seeded into T25 culture bottles at a density of 1 × 10⁶ cells/bottle. Cells were transfected with LV5-KLF6 or LV5-NC at a MOI  of  50 for 48 h. Transfection efficiency was preliminarily checked by observing the GFP-positive cells (> 80%) using a fluorescence microscope. For transcriptome sequencing, RNA was extracted using TRIzol reagent. Libraries of the amplified RNA from each pool were prepared using the Illumina protocol. cDNA library construction and sequencing were conducted using an Illumina Novaseq6000 (Gene Denovo Biotechnology, Guangzhou, China). Principal component analysis (PCA) was performed using the gmodels package. Analysis of differentially expressed genes (DEGs) was conducted using the DESeq2 software. Genes with P adj ≤ 0.05 and absolute fold change ≥ 1.5 were regarded as DEGs. Database annotation of the pathways and functions of the DEGs was conducted using Reactome enrichment analysis. Gene set enrichment analysis (GSEA) was performed using the MSigDB software. Heat maps were generated using the TB tools.

For metabolomics, 1 × 10⁷ cells of each sample were harvested, transferred to centrifuge tubes, and then washed three times with cold PBS. After centrifugation at 4 ℃, 200 × g for 10 min, the supernatants were discarded and the precipitated cells were extracted in 500 µL of methanol–water solution (one part methanol to four parts water, prechilled at -20 ℃). The cell extracts were subjected to ultra-performance liquid chromatography (UPLC; ExionLC™ AD, Sciex Technologies, Framingham, MA, USA) coupled to tandem mass spectrometry (MS/MS) (QTRAP^® 6500 system, Sciex Technologies). Qualitative analyses were performed using the MWDB database. Quantitative analysis was conducted using the multiple reaction monitoring mode of the QTRAP 6500 + LC-MS/MS System (Sciex Technologies). Orthogonal least partial squares discriminant analysis (OPLS-DA) was performed using MetaboAnalyst 5.0. Metabolites with P adj ≤ 0.05 and absolute fold change ≥ 1.5 and VIP scores ≥ 1 were identified as differential metabolites. Pathway enrichment analysis was performed based on the KEGG database.

The detailed experimental procedures of transcriptome sequencing and metabolomics are provided in the supplementary materials.

Mitochondria staining and analysis

DPSCs were seeded in 6-well plates with cell-climbing slices at 5 × 10⁴ cells/well. The attached cells were stained with Mito-Tracker Red CMXRos working solution (100 nM, Beyotime) at 37 ℃ in the dark for 15 min. Following three washes with PBS, the nuclei were stained with Hoechst (Beyotime) for 10 min. The climbing slices were washed three times with PBS and then incubated with ProLong Live Antifade reagent (#P36975, Invitrogen) for 30 min to minimize fluorescence signal loss. Images were captured using a Leica SP8 confocal microscope (Leica Microsystems GmbH, Wetzlar, Germany) under identical imaging settings. Mitochondrial morphology was analyzed using ImageJ and the Mitochondrial Network Analysis Tool.

Cryo-transmission electron microscopy

Cells were collected by trypsin digestion and centrifuged at 300 × g for 3 min. The cells were fixed in 2.5% glutaraldehyde at 25 ℃ for 2 h, stored overnight at 4 ℃, washed in 0.1 M PBS for 10 min, fixed in 1% osmium tetroxide for 1 h at 25 ℃, and then incubated in 2% uranyl acetate in H₂O for 30 min. Parameters were as follows: dehydration: 50% EtOH, 15 min; 70% EtOH, 15 min; 90% EtOH, 15 min; 100% EtOH, 20 min, twice; and 100% propanone, 20 min, twice. For infiltration, the samples were allowed to stand in EPON mixed with propanone (1:1) for 2 h at RT, embedded in EPON, and then allowed to polymerize. Ultrathin sections were prepared with a Leica EM UC7 Ultramicrotome. Mitochondrial morphology was observed using a Talos L120C TEM (Thermo Fisher Scientific).

Measurement of glutamine, glutamate, α-KG, and citrate concentrations in cultured cells

Stably transfected hDPSCs (1 × 10⁶) were obtained and washed with 0.9% NaCl. Each sample was extracted with 1 mL aliquots of a methanol–acetonitrile–water mixture (5:3:2, v/v/v). For absolute quantification, standard references were prepared by dissolving glutamine, glutamate, α-KG, and citrate (purchased from Sigma-Aldrich) in the same solvent at 200 µg/mL. LC-MS/MS quantification of the metabolites was performed using a Xevo TQ-XS triple quadrupole mass spectrometer (Waters). The areas of the analyte peaks were calculated. Intracellular glutamate and citrate contents were determined using a Glutamate Assay Kit (Solarbio, Beijing, China) and a Citrate Assay Kit (BioAssay Systems, Hayward, CA, USA), respectively.

Measurement of citrate and Ca in ECM

The citrate content in the mineralized matrix was analyzed as previously described [15]. After 7 days of odontogenic induction, the cell culture medium was replaced with DM containing ¹³C-glutamine for another 7 days of induction. Subsequently, the cells were lysed with 80% methanol in dH₂O by freeze–thawing in liquid nitrogen to release intracellular metabolites. After centrifugation for 10 min at 14 000 × g, the supernatants were discarded, and the remaining precipitates were harvested with 1 M HCl at 60 ℃ for 1 h to free the mineral-bound citrate. Following centrifugation for 10 min at 14 000 × g, half of the supernatant from each sample was deproteinated and mixed with an equal volume of acetonitrile before LC-MS/MS analysis using a Xevo TQ-XS Triple Quadrupole Mass Spectrometer. The remaining half of the supernatant was diluted with dH₂O (1:100, v/v) to determine the Ca concentration using inductively coupled plasma atomic emission spectroscopy (ICP, PerkinElmer, Waltham, MA, USA). The data were normalized to protein concentration.

Enzyme activity assays

For each sample, 1 × 10⁶ cells were broken ultrasonically at 4 ℃. After centrifugation at 12 000 × g, 4 ℃ for 15 min, the supernatant was collected for further testing. GLS1 and GDH activities in hDPSCs were measured using a GLS1 activity assay kit (#BC1450, Solarbio) and a GDH activity assay kit (#BC1460, Solarbio), respectively. GLS1 activity was measured using a spectrophotometer at 630 nm. To detect the enzyme activity of GDH, absorbance at 340 nm was recorded at 20 s intervals.

Metabolic assays

The oxygen consumption rate (OCR) was measured using a Seahorse XF Cell Mito Stress Test Kit (Agilent Technologies) and Seahorse XF96 Analyser (Agilent Technologies). Briefly, cells were seeded in Seahorse cell culture plates at a density of 1000 cells/well. Before starting the assay, cells were precultured in Seahorse DMEM base medium in a non-CO₂ incubator for 60 min at 37 °C. Oligomycin (1.5 µM), FCCP (0.5 µM), and rotenone/antimycin A (0.5 µM) were added sequentially according to the user guide. The data were further normalized based on the protein concentration.

Metabolic flux

The hDPSCs were transfected with LV5-KLF6 or LV5-NC for 48 h in advance or cultured in OM or DM for 7 days (n = 4). The medium was then replaced with fresh medium containing 2 mM ¹³C-glutamine. Metabolic flux tracing with ¹³C-glutamine was performed using a UHPLC system (Agilent 1290 II) coupled to a mass spectrometer (5600 Triple TOF Plus, AB Sciex, Concord, ON, Canada) as previously described [16]. The data were normalized to protein concentration.

CHIP-qPCR

Cells were transfected with pcDNA3.1 or KLF6 overexpression plasmids using Lipo3000 (Themo Fisher). A SimpleChIP^® Enzymatic Chromatin IP Kit(#9003, CST)was used to verify KLF6 binding sites. Cells were crosslinked in 1% formaldehyde at 37 ℃ for 10 min and washed three times with PBS. DNA was digested with Micrococcal Nuclease to an average length of approximately 150–900 bp. The nuclear extract was centrifuged at 9400  × g to collect the supernatant containing crosslinked chromatin. A 2% sample was set aside as input. The equal amounts of samples were diluted ten fold with ChIP dilution buffer. Each sample was incubated with IgG antibody (1 µg/mL) at 4 ℃ with slow rotation to reduce background. After overnight incubation, 30 µL of Protein G Magnetic Beads were added, and the samples were incubated at 4 ℃ for 60 min (with slow rotation) to pull down non-specific binding complexes. The supernatant was carefully collected and then incubated overnight at 4 ℃ with KLF6 antibody (with slow rotation). Subsequently, 30 µL of Protein G Magnetic Beads were incubated with the samples for 2 h at 4 ℃ with rotation. The tubes were placed in a magnetic separator to pellet the Protein G Magnetic Beads. To wash the precipitated Protein G Magnetic Beads, 1 mL of low-salt wash buffer was added, and the mixture was incubated at 4 ℃ for 5 min, repeated three times. The beads were then washed with 1 mL of high-salt wash buffer for 5 min at 4 ℃ The supernatant was carefully removed, and 150 µL of ChIP elution buffer was added. The samples were vortexed for 30 min at 65 ℃ to elute the chromatin from the antibody/Protein G magnetic beads. The tubes were placed in a magnetic separator to pellet the Protein G magnetic beads, and the eluted chromatin supernatant was transferred to a new tube. To reverse crosslinking, 6 µL of 5 M NaCl and 2 µL of Proteinase K were added, and the mixture was incubated at 65 ℃ for 2 h. DNA purification was performed using the MicroElute DNA Clean-Up Kit (Omega Biotek) according to the manufacturer’s instructions. Quantitative PCR was then conducted and the primers were listed in Table. S1.

Mouse molar injury model

Animal use in this study was approved by the Animal Ethics Committee of Zhejiang University (No. ZJU20230259). Twenty-four C57BL/6 male mice aged 6–8 weeks were used for generate mouse molar injury model. The protocol was depicted in our previous study [4]. Mice were randomized into four groups with 6 mice per group.

Adeno-associated virus 6 (AAV6) was used to silence KLF6 in vivo. AAV6-shKLF6 and AAV6-shNC were constructed by GenePharma. The sequences of AAV6-shKLF6.

were as follows: 5’-GGAAGACCTGTGGACCAAATT-3’. The sequences of AAV6-shNC.

were as follows: 5’-ACTACCGTTGTTATAGGTG-3’. For AAV injection, twelve mice were anesthetized with 1% pentobarbital sodium and then accepted injection of 10 µl AAV6-shKLF6 (n = 6) or AAV6-shNC (n = 6) (1 × 10¹² vg/ml) into gum of mandibular first molars every 3 days for a week. After 2 week, the maxillary first molars were drilled open and iRoot-BP Plus (Innovative Bioceramix) was placed into the perforations. Then the cavities were filled with glass ionomer cement (Fuji IX). Another twelve mice were anesthetized and subjected to prepare cavities in the maxillary first molars without exposure of dental pulp. These mice received intraperitoneal injections of 12.5 mg/kg body weight every 3 days of BPTES (n = 6) or vehicle control (n = 6) (10% DMSO in 200 µl of PBS every 3 days). Mice died during operation were excluded. The final number of mice in each group was: BPTES group (n = 5),vehicle control group (n = 6), AAV6-shNC group (n = 5), and AAV6-shKLF6 group (n = 5). All mice were sacrificed by carbon dioxide (CO₂) inhalation at 4-weeks post-surgery and the maxillae were harvested.

µCT and histological analysis

For µCT, the samples were scanned using µCT with Milabs U-CT-XUHR (Utrecht, Netherlands) at scanning parameters of 116 kV, 60 µA, 8 μm. After image acquisition, Milabs software (version 1.4.4) was used to reconstruct the digital data.

For histological analysis, paraffin sections were stained with hematoxylin and eosin. Immunohistochemistry and immunofluorescence staining was performed as previously described [4, 17]. The slides were incubated overnight with the appropriate primary antibodies at 4 ℃, including GLS1 antibody (1:200, #49363, Cell Signalling Technology), GDH antibody (1:200, #12793, Cell Signalling Technology), and DSPP antibody (1:100, #bs-10316R, Bioss, Beijing, China), DMP1 antibody (1:200, Invitrogen). For immunofluorescence staining, the slides were incubated with the secondary antibodies Alexa Flour 488 or 594 for 1 h at 37 ℃ and then mounted with FluoroShield mounting medium containing DAPI (Thermo Fisher Scientific). The slices were observed using an Olympus VS200 Slide Scanner (Olympus Life Science, Waltham, MA, USA).

Statistical analysis

Data are presented as mean ± S.D. Statistical significance was determined using the SPSS software (version 26.0, IBM). The Shapiro–Wilk test was used to analyze normality. Levene’s test was used to analyze variance homogeneity. Student’s t-test between two groups and one-way ANOVA with LSD multiple comparison tests were performed to determine significance. In case of non-normal distribution or unequal variances, Mann–Whitney U and Kruskal–Wallis H tests were applied. Statistical significance is represented as * P < 0.05, ** P < 0.01, and *** P < 0.001 (α = 0.05).

Data availability statement

The data generated in this study are available from the corresponding author upon reasonable request.

Results

RNA sequencing and targeted metabolomics reveal changes in glutamine metabolism

Human DPSCs were isolated and identified using flow cytometry (Fig. S1). The odontogenic differentiation and matrix mineralization capacities of hDPSCs were identified via ALP and ARS staining (Fig. S1B). KLF6 was over-expressed using lentivirus transfection system. As shown by fluorescence microscopy and flow cytometric analysis, the proportions of transfected cells can achieve more than 98% (Fig. S5).

To explore global changes in gene expression after KLF6 overexpression, we used RNA-seq to reveal expression changes in the transcriptome of hDPSCs (Fig. 1A-C). We found 2650 DEGs between the control group and KLF6 overexpression group. Pathway enrichment analysis results showed that 1390 upregulated genes were particularly enriched in “Extracellular matrix organization,” “Metabolism,” and “Collagen formation” (Fig. 1B). ECM secretion and mineralization are two notable events in dentin formation [18]. Odontoblasts are responsible for the synthesis of the ECM, especially type I collagen. This result implies that KLF6 can promote hDPSC differentiation into odontoblasts, which can then perform their functions, and this is consistent with our previous findings.

Fig. 1

RNA-seq and targeted metabolomics reveal changes in glutamine metabolism and mitochondrial respiration. (A) Principal component analysis (PCA) of RNA-seq. (B) Differential gene enrichment analysis with Reactome. (C) Gene set enrichment analysis (GSEA) shows that gene sets related to mitochondrial respiratory and TCA cycle are positively associated with the KLF6 high expression group. (D) PCA of metabolomics data. (E) Top 9 enriched metabolite sets based on KEGG pathway analysis. (F) Top 25 differential metabolites with the OPLS-DA model. VIP scores, variable importance for projection scores

GSEA of RNA-seq data showed that gene sets related to mitochondrial respiration and the TCA cycle were positively associated with the KLF6 high expression group (Fig. 1C). Furthermore, we performed targeted metabolomic sequencing to identify the metabolic changes after KLF6 overexpression(Fig. 1D-F). The differentially upregulated metabolites between the two groups were enriched mainly in glutamine, alanine, aspartate, and glutamate metabolism and in the TCA cycle (Fig. 1E). Moreover, the levels of glutamine, as well as citrate, a key product of the TCA cycle, increased inside the cell (Fig. 1F). Together with the transcriptome analyses, alterations in these metabolites suggest a remarkable influence on glutamine metabolism, mitochondrial respiration, and the TCA cycle by KLF6.

KLF6 is a critical regulator of glutamine metabolism and mitochondrial respiration

The efficacy of KLF6 overexpression was validated using RT-qPCR and western blot (Fig. 2A and B). The expression levels of GLS1 and GDH increased in the KLF6 high expression group (Fig. 2B). GLS1 and GDH activities were also increased in the KLF6 high expression group (Fig. 2C and D). Glutamine is transported into the cytoplasm, metabolized to glutamate by GLS1, and further converted into α-KG by GDH (Fig. 2E). Liquid chromatography-tandem mass spectrometry (LC-MS/MS) results showed that intracellular glutamine content increased in the KLF6 high expression group, indicating that glutamine uptake was enhanced by KLF6 (Fig. 2F). The results obtained using the kit method for detecting glutamine concentration were consistent (Fig. 2G). Intracellular glutamate and α-KG levels increased in the KLF6 high expression group, suggesting that intracellular glutamine metabolism was enhanced (Fig. 2F). Cellular ATP content also increased, suggesting elevated energy production (Fig. 2H). Thus, we speculated that KLF6 promotes glutamine metabolism and energy production.

Fig. 2

KLF6 promotes glutamine metabolism and mitochondrial respiration in human dental pulp stem cells (hDPSCs). (A) Relative mRNA expression of KLF6 after transfection with LV5-NC or LV5-KLF6 (n = 3). (B) KLF6, GLS1, and GDH protein expression. β-Actin serves as a loading control. Full-length blots/gels are presented in Supplementary Figure S8. (C and D) Relative enzyme activity of GLS1 and GDH (n = 3). (E) Schematic diagram of glutamine metabolism. (F) Relative glutamine, glutamate, and α-KG contents determined by LC-MS/MS (n = 3). (G) Glutamate content determined by a glutamate assay kit (n = 3). (H) Intracellular ATP contents (n = 6). (I–M) Seahorse assay results for oxygen consumption rates (OCR) (n = 6). (N) Representative cryo-TEM images of human dental pulp stem cells (hDPSCs) transfected with LV5-NC or LV5-KLF6; blue areas indicate mitochondria. (P) Quantitative comparison of mitochondrial numbers, lengths, and occupied areas (n = 15). (O) Representative confocal microscopy images of the mitochondrial network revealed with a MitoTracker probe. (Q) Relative average mitochondrial footprint and summed branch lengths performed on confocal microscopy images (n = 15). Data were analyzed using Student’s t-test (D, F–J,Q–S, andU) and Mann–Whitney U test (C, L–O, and V). * P < 0.05, ** P < 0.01, and *** P < 0.001

Α-KG is the end-product of glutamine metabolism and can enter the mitochondria to generate citrate through the TCA cycle. To test whether mitochondrial energy metabolism was also altered, a Seahorse XF analyser was used to measure the OCR. Increased basal respiration, maximal respiration, spare respiratory capacity, and ATP production were observed in the KLF6 high expression group, indicating that glutamine-supported oxidative phosphorylation may be promoted (Fig. 2I–M). We further examined mitochondrial morphology and number. Using cryo-TEM, we observed relatively more elongated and mature mitochondria in the KLF6 high expression group (Fig. 2N). Moreover, the average lengths of mitochondria, average numbers of mitochondria per cell, and average mitochondrial area/cell area significantly increased by KLF6 overexpression (Fig. 2P). We also observed mitochondrial morphology using a confocal microscope (Fig. 2O). The average mitochondrial footprint and average summed branch lengths also increased by KLF6 overexpression (Fig. 2Q). Collectively, these data indicated that KLF6 may promote glutamine metabolism and mitochondrial function.

Glutamine metabolism and mitochondrial respiration are enhanced during odontogenic differentiation of hDPSCs

Next, we evaluated the change of glutamine metabolism in hDPSCs during odontoblastic differentiation. The mRNA expression levels of GLS1 and GDH increased with odontogenic induction (Fig. S2). Confocal microscopy and cryo-TEM showed that the number, length, and area occupied by the mitochondria increased in differentiating hDPSCs (Fig. S2G–J). Seahorse metabolic flux measurements showed an elevation in basal respiration, maximal respiration, spare respiratory capacity, and ATP production during the odontoblastic differentiation of hDPSCs (Fig. 3A and B).

To further evaluate glutamine utilization by hDPSCs, we traced glutamine metabolism using ¹³C-glutamine (Fig. 3C). In brief, hDPSCs were precultured in ordinary growth medium (OM) or differentiation medium (DM) for 7 days and subsequently cultured in fresh medium containing ¹³C-glutamine for 12 h before determining the incorporation of ¹³C derived from ¹³C-glutamine in downstream metabolites. Intracellular M + 5 glutamine levels increased, indicating that glutamine uptake was enhanced during odontogenic differentiation (Fig. 3D). Moreover, the increased levels of glutamate (M + 5) and α-KG (M + 5) derived from ¹³C-glutamine in differentiated hDPSCs indicated an enhanced requirement for glutamine metabolism (Fig. 3D). Importantly, TCA cycle intermediates were increased, suggesting increased anaplerotic entry into the TCA cycle in differentiated hDPSCs (Fig. 3D). Moreover, the contribution of glutamine to L-proline (M + 5) was increased in hDPSCs (Fig. 3D). Considering that proline is a main amino acid in collagen, we hypothesized that glutamine also contributes to collagen synthesis during the odontogenic differentiation of hDPSCs. Consistent with the function of differentiating hDPSCs, we found that the levels of aspartate (M + 4), alanine (M + 3), and uridine monophosphate (UMP) (M + 3) derived from glutamine increased, suggesting an increased biosynthetic demand in differentiated hDPSCs (Fig. 3D). Taken together, glutamine metabolism is required to support increased energy and biosynthetic requirements during the odontogenic differentiation of hDPSCs.

Fig. 3

Glutamine metabolism and mitochondrial respiration are enhanced during odontogenic differentiation of human dental pulp stem cells (hDPSCs). (A, B) Seahorse assay results for OCRs of differentiating hDPSCs (odontoblastic induction for 0, 5, and 7 days) (n = 6). (C) Schematic summarizing the flux of ¹³C-glutamine into the TCA cycle and other metabolic activities. (D) LC-MS quantification of ¹³C-labeled (M3/M4/M5) metabolites in hDPSCs cultured in ordinary growth medium (OM) or differentiation medium (DM) (n = 3). Results are calculated as relative peak areas normalized with protein concentration. Data were analyzed using the Kruskal–Wallis H tests in (B–E) and analyzed using Student’s t-test (G–Q). * P < 0.05, ** P < 0.01, and *** P < 0.001

Deprivation of glutamine inhibits odontogenic differentiation and mineralization of hDPSCs and restrains deposition of citrate and Ca in ECM

We explored whether glutamine is necessary for odontogenic differentiation and matrix mineralization. The cells were cultured in an odontogenic medium with or without glutamine. ARS staining at 7 and 14 days showed that the formation of calcified nodules significantly reduced in the glutamine-deprived group (Fig. 4A). ALP staining and quantitative assays confirmed that glutamine deprivation decreased ALP activity (Fig. 4B and E). The protein expression of RUNX2, DSPP, DMP1, and COL1A1 decreased after glutamine deprivation (Fig. 4C). RT-qPCR results revealed that the transcription levels of odontogenic differentiation-related genes, including DSPP, DMP1, ALP, OCN, OSX, BSP, RUNX2, COL1A1, and COL1A2, significantly reduced in the glutamine-deprived group (Fig. 4D). Collectively, glutamine is essential for odontogenic differentiation and matrix mineralization in vitro.

Fig. 4

Deprivation of glutamine inhibits odontogenic differentiation of human dental pulp stem cells (hDPSCs) and matrix mineralization. (A) Alizarin red S (ARS) staining of hDPSCs cultured in odontogenic medium (DM) with or without glutamine for 14 and 21 days. (B and E) Alkaline phosphatase (ALP) staining and quantifying ALP activity of hDPSCs (n = 6) after 7 days of differentiation induction in DM with or without glutamine. (C) Protein expression levels of DMP1, RUNX2, DSPP, COL1A1, and ACTIN in hDPSCs after 7 days of induction. Full-length blots/gels are presented in Supplementary Figure S8. (D) Relative mRNA expression of genes related with odontoblastic differentiation, including ALP, DMP1, OCN, OSX, COL1A1, COL1A2, RUNX2, DSPP, and BSP, in hDPSCs after 7 days of differentiation induction in DM with or without glutamine (n = 3). (F) Relative citrate content in the cytoplasm determined by a citrate detection kit after 7 days of differentiation induction (n = 5). (G) Relative citrate content in the extracellular mineralized matrix of hDPSCs after culturing in DM with or without glutamine for 21 days (determined by LC-MS/MS) (n = 3). (H) Relative Ca content in the extracellular mineralized matrix of hDPSCs after culturing in DM with or without glutamine for 21 days (determined by ICP-MS) (n = 3). Data were analyzed using Student’s t-test (C, D, F, and H) and Mann–Whitney U test (G). * P < 0.05, ** P < 0.01, and *** P < 0.001

Using a citrate content assay kit, we observed a marked decrease in intracellular citrate content under glutamine-deprived conditions, indicating that glutamine deprivation restrains citrate production via the TCA cycle (Fig. 4F). This result was consistent with the observations from the ¹³C-glutamine tracing experiment. Citrate is abundant in bones and teeth and plays a critical role in stabilizing apatite nanocrystals [19]. Recent studies have suggested that citrate exists in dentin as a tightly bound component of a “Ca/citrate/collagen I” complex, similar to that found in bones [20, 21]. Osteoblasts and odontoblasts work as specialized citrate-producing cells to ensure the deposition of citrate into mineral hydroxyapatite [15]. Hence, we determined the citrate and Ca contents in the ECM using LC-MS/MS and ICP-MS. Interestingly, glutamine deprivation significantly decreased the citrate and Ca contents in the ECM (Fig. 4G and H). Thus, we propose that KLF6-mediated glutamine metabolism promotes the matrix mineralization of functional odontoblast-like cells differentiated from hDPSCs by increasing citrate and collagen deposition.

KLF6 affects glutamine metabolism by directly regulating GLS1 and GDH

To investigate whether KLF6 mediates odontogenic differentiation and mineralization by directly regulating glutamine metabolism, KLF6 knockdown was mediated by a lentivirus, and the efficiency of the knockdown was determined by RT-qPCR and western blot (Fig. 5A and B). As expected, the expression levels of GLS1 and GDH were significantly downregulated (Fig. 5B). The enzyme activity of GLS1 was downregulated by KLF6 knockdown (Fig. 5C). KLF6 knockdown decreased intracellular glutamate and α-KG levels, suggesting that glutamine metabolism was attenuated (Fig. 5D and E). It also decreased intracellular ATP content (Fig. 5F). To elucidate the regulatory role of KLF6 on GLS1 and GDH, we employed ChIP-qPCR experiments, which revealed that KLF6 can directly bind to the promoter sequences of GLS1 and GDH, thereby activating their expression (Fig. 5G and H).

Fig. 5

KLF6 directly regulates GLS1 and GDH expression to affect glutamine metabolism. (A) Relative mRNA expression of KLF6 after KLF6 knockdown (n = 3). (B) KLF6, GLS1, and GDH protein expression. β-Actin serves as a loading control. Full-length blots/gels are presented in Supplementary Figure S8. (C) Relative enzyme activity of GLS1 and GDH (n = 3). (D) Relative glutamine, glutamate and α-KG content determined by LC-MS/MS (n = 3). (E) Glutamate content determined by a glutamate assay kit (n = 3). (F) Intracellular ATP content was reduced by KLF6 knockdown (n = 6). (G) ChIP-qPCR showed KLF6 primarily binds to the Site 1 of the GLS1 promoter region (n = 3). (F) ChIP-qPCR showed KLF6 primarily binds to the Site 1 on the promoter region of GDH (n = 3). (I) LC-MS/MS quantification of ¹³C-labeled (M3/M4/M5) metabolites in hDPSCs transfected with LV5-NC or LV5-KLF6 and the results were calculated as relative peak areas normalized with protein concentration (n = 3). (J) Schematic summarizing the flux of ¹³C-glutamine into the TCA cycle and other metabolic activities. (K) Relative ¹³C-labeled citrate contents in the mineralized matrix were determined by LC-MS/MS (n = 3). (L) Relative Ca contents in the mineralized matrix were determined by ICP-MS (n = 3). Data were analyzed using Student’s t-test (A, C–I, and K–L). * P < 0.05, ** P < 0.01, and *** P < 0.001

In addition, to test whether KLF6 affects glutamine flux in hDPSCs, a ¹³C-glutamine tracing experiment was used again. Intracellular glutamine (M + 5), glutamate (M + 5), and α-KG (M + 5) levels significantly increased in the KLF6 high expression group, suggesting that KLF6 promoted glutamine uptake and glutaminolysis (Fig. 5I). Similarly, the levels of ¹³C-labelling TCA cycle intermediates increased, thus corroborating our assumption that KLF6 promotes the conversion of glutamine into α-KG to fuel the TCA cycle (Fig. 5I). Glutamine-derived aspartate (M + 4) levels were increased by KLF6 overexpression, while asparagine (M + 4) and UMP (M + 3) level showed no significant change (Fig. 5I). Importantly, the contents of M + 4 citrate and M + 5 proline significantly increased in the KLF6 high expression group, implying an enhanced accumulation of citrate and collagen (Fig. 5I and J). Thus, these results suggest that KLF6 affects glutamine metabolism by directly regulating GLS1 and GDH.

KLF6 mediates ECM mineralization by promoting collagen and citrate deposition

To confirm whether citrate derived from glutamine is transported outside the cell to participate in matrix mineralization, we measured the contents of mineral-bound citrate with ¹³C labels, as well as Ca, in the ECM in the KLF6 high-expression and control groups. The ¹³C labeled citrate content (derived from ¹³C-glutamine) in the mineralized matrix increased in the KLF6 high expression group, and Ca content in the ECM also increased (Fig. 5K and L). These observations confirm our hypothesis that KLF6-mediated glutamine metabolism promotes matrix mineralization by increasing citrate deposition in the ECM.

We also examined whether KLF6 promotes type I collagen formation and secretion. As shown in Fig. 1E, the upregulated genes in the KLF6 high expression group were particularly enriched in “ECM organization,” “Collagen formation,” and “Metabolism.” Furthermore, GSEA demonstrated that gene sets related to “Secretion of collagens,” “Collagen formation,” “Assembly of collagen fibrils and other multimeric structures,” “Collagen biosynthesis and modifying enzymes,” and “ECM organization” were positively associated with the KLF6 high expression group (Fig. S3A-E). A heatmap of collagen formation-related genes is shown in Fig. S3F. Consistently, KLF6 overexpression increased the expression of type I collagen, whereas KLF6 knockdown significantly inhibited this expression (Fig. S3G, H, J, and K). Collagen deposition in the ECM was quantified by Sirius Red staining. As expected, KLF6 overexpression promoted collagen deposition in the ECM, which was inhibited by KLF6 knockdown (Fig. S3I, L). Taken together, our data indicate that KLF6 promotes collagen formation in hDPSCs.

To investigate whether KLF6-mediated glutamine metabolism is essential for matrix mineralization in differentiated hDPSCs, we used BPTES (an inhibitor of GLS1) to block glutamine metabolism. BPTES attenuated odontogenic differentiation and mineralization induced by KLF6 overexpression (Fig. S4A and B). Inhibition of glutamine metabolism by BPTES significantly suppressed the increase in cytoplasmic citrate content induced by KLF6 overexpression (Fig. S4C). The above findings suggest that KLF6 promotes the conversion of glutamine to generate α-KG, which enters the TCA cycle to promote citrate deposition in the ECM, thereby inducing matrix mineralization.

Supplementation of α-KG rescues suppression of odontogenic differentiation and mineralization induced by KLF6 knockdown

Considering that α-KG is a bridge between glutamine metabolism and the TCA cycle, we used dimethyl–α-KG (dm–α-KG), a cytomembrane-permeable form of α-KG, to rescue the inhibitory effect of KLF6 knockdown on odontogenic differentiation and mineralization. Dm–α-KG restored the expression levels of DSPP, DMP1, BSP, and COL1A1 inhibited by KLF6 knockdown (Fig. 6A–D). ALP and RUNX2 protein expression levels were restored by dm–α-KG (Fig. 6E). ARS staining revealed that KLF6 knockdown significantly inhibited mineralized nodule formation, whereas supplementation with dm–α-KG rescued that repression. ALP staining showed a similar result; only the KLF6 knockdown group showed an obvious downregulation of ALP activity (Fig. 6F). Moreover, supplementation with dm–α-KG restored the basal respiration, maximal respiration, spare respiratory capacity, and ATP production capacity of the KLF6-knockdown hDPSCs to the levels of the negative control (Fig. 6G–K). Collectively, these results suggest that dm–α-KG can rescue the suppression of odontogenic differentiation and mineralization induced by KLF6 knockdown. Thus, KLF6 may mediate odontogenic differentiation and mineralization by directly regulating glutamine metabolism.

Fig. 6

Supplementation of α-KG rescued suppression of odontogenic differentiation and mineralization induced by KLF6 knockdown. (A–D) Relative mRNA expression odontogenic differentiation-related genes, including DSPP, DMP1, BSP, and COL1A1 (n = 3). (E) ALP and RUNX2 protein expression was rescued by supplementation of α-KG. Full-length blots/gels are presented in Supplementary Figure S8. (F) ARS and ALP staining showed that supplementation of α-KG rescued the suppression of matrix mineralization induced by KLF6 knockdown. (G–K) Seahorse assay results for OCRs (n = 8). (L) Relative citrate contents in the cytoplasm (n = 3). Data were analyzed with one-way ANOVA with LSD multiple comparison tests in (A–D, and H–L). * P < 0.05, ** P < 0.01, and *** P < 0.001

In addition, supplementation of dm–α-KG rescued the citrate content in the cytoplasm decreased by KLF6 knockdown (Fig. 6L). These results further reinforced the hypothesis that KLF6 promotes glutamine to form α-KG, which enters the TCA cycle to accelerate the deposition of citrate in ECM, thereby promoting matrix mineralization.

KLF6 regulates tertiary dentine production in vivo

To investigate the function of glutamine metabolism and KLF6 in tertiary dentine production, we first constructed a mouse dental injury model without exposure of dental pulp. In the BPTES group, glutamine metabolism was inhibited. As shown in Fig. 7A and B, the formation of reactive dentin were inhibited in the BPTES group. The ratio of coronal pulp in the BPTES group were significant higher than which in the control group (Fig. 7C). And IHC staining showed ratios of positive staining area of DSPP and DMP1 were lower in the BPTES group (Fig. 7D). These results suggest that inhibition of glutamine metabolism attenuated reactive dentin formation. Then AAV6 was used to silence KLF6 in a direct capping mouse model. As shown in Fig. 7E and F, the formation of restorative dentin was inhibited in the AAV6-shKLF6 group. The thickness of dental bridges and ratios of positive staining area of DSPP and DMP1 were significantly lower in the AAV6-shKLF6 group (Fig. 7G-H). Thus, knockdown of KLF6 in vivo attenuated restorative dentin formation. Taken together, we suggest KLF6 and glutamine metabolism are important for odontogenic differentiation and mineralization in vivo.

Fig. 7

Inhibition of glutamine metabolism and KLF6 attenuated tertiary dentin formation. (A) Representative µCT images of maxillary first molars from mice received intraperitoneal injections of BPTES or vehicle control. (B) Representative HE and IHC staining images of DSPP and DMP1. (C) The ratio of coronal pulp based on µCT images (control group: n = 11, BPTES group: n = 10). (D) The ratios of DSPP and DMP1 positive staining areas based on IHC staining images (control group: n = 5, BPTES group: n = 6). (E) Representative µCT images of maxillary first molars from mice received injection of AAV6-shKLF6 or AAV6-shNC. (F) Representative HE and IHC staining images of DSPP and DMP1. (G) The thickness of dental bridges based on HE staining images (n = 10). (H) The ratios of DSPP and DMP1 positive staining areas based on IHC staining images (n = 6). Data were analyzed using Student’s t-test in (C–D, and G–H). * P < 0.05, ** P < 0.01, and *** P < 0.001

Moreover, to explore whether KLF6 promotes odontogenic differentiation and biomineralization of hDPSCs in vivo, we used a semi-orthotopic regeneration animal model. Tooth RSs with cell-laden or acellular GelMA hydrogels were implanted subcutaneously on the dorsal sides of male nude mice for 6 weeks before harvest. Cytoskeleton staining showed that hDPSCs were distributed evenly in the 3D hydrogel (Fig. S6A). Live/dead staining confirmed the viability of hDPSCs within the hydrogels (Fig. S6B). In the blank group, only a small number of cells attached to undegraded GelMA were observed. Immunofluorescence staining revealed ordered DSPP-expressing cells attached to dentinal walls in the LV5-KLF6 group (Fig. S7D). More GLS1-positive and GDH-positive cells were found in the LV5-KLF6 group than in the other groups (Fig. S7E–J). Furthermore, semi-quantitative analysis showed that the average fluorescent intensities of DSPP, GLS1, and GDH were higher in the LV5-KLF6 group than in the other groups (Fig. S7K–M). These results suggest that KLF6 promotes the odontogenic differentiation and biomineralization of hDPSCs in vivo.

Discussion

In this study, we identified a vital factor, KLF6, critical for reprogramming the cell metabolism of hDPSCs and promoting odontogenic differentiation and dentinal matrix mineralization. We discovered a novel mechanism whereby KLF6 promotes matrix mineralization by directly regulating glutamine metabolism. Accordingly, targeting KLF6 or glutamine metabolism may be a potential strategy for pulp dentin regeneration and reparative dentin formation.

Odontoblast differentiation and dentinal matrix mineralization have attracted considerable attention for many decades [22]. Pulp repair through mineralization by functional odontoblast-like cells differentiated from hDPSCs is indispensable to support pulp dentin regeneration or reparative dentin formation. However, modulating the odontogenic differentiation of stem cells is extremely difficult without a deeper comprehension of the mechanisms governing odontogenic differentiation and matrix mineralization. KLF6 acts as a crucial regulator of odontoblast differentiation of mDPCs and is expressed not only in odontoblasts in the murine tooth germ but also in odontoblast-like cells around the reparative dentin in pulp exposure models [4, 8]. However, whether KLF6 has similar effects on the odontogenic differentiation of hDPSCs and how it regulates the function of odontoblast-like cells remains unclear. In line with the findings of this study, we propose that KLF6 regulates the function and fate of hDPSCs mainly by altering cell metabolism, especially glutamine metabolism and the TCA cycle. Using RNA-seq and pathway enrichment analyses, we initially found that genes involved in ECM secretion and mineralization were enriched in the transcriptome of KLF6 high-expression hDPSCs, suggesting the differentiation of these cells into odontoblasts. Furthermore, KEGG enrichment analyses of our metabolomics data revealed differentially upregulated metabolites involved in glutamine metabolism, mitochondrial respiration, and the TCA cycle by KLF6, indicating the role of KLF6 in regulating these processes. By tracing glutamine metabolism using ¹³C-glutamine, we later observed that glutamine was consumed to fuel the TCA cycle more efficiently, leading to higher citrate content in differentiated hDPSCs than in undifferentiated cells. Motivated by significant advances made in investigations of the role of energy metabolism in shaping stem cell function and fate, our study takes a fresh look at how KLF6 works in regulating the differentiation and function of odontoblast-like cells. We suggest that KLF6 not only enhances glutamine metabolism to support the increased energy demands and biosynthesis required by differentiated hDPSCs but also mediates glutamine undergoing anaplerotic reactions to form α-KG that enters the TCA cycle to promote citrate deposition in the ECM, which in turn induces matrix mineralization.

Recently, great interest has arisen regarding the involvement of glutamine metabolism in regulating stem cell fate because alternative carbon sources are urgently needed to fulfill the bioenergetic demands for proliferation and differentiation, and metabolic pathways can influence epigenetic changes [23, 24]. A major finding of this study is that KLF6 promotes glutamine metabolism in hDPSCs. Similarly, Lee et al. reported that the loss of Klf6 suppresses glutamine-derived ornithine levels in pancreatic cancer [13]. Glutamine metabolism fulfills many roles, particularly in supplementing the TCA cycle, supporting the biosynthesis of other amino acids and nucleotides, and the generation of energy [25]. Moreover, it plays an essential role in embryonic development and bone anabolism [26, 27]. Recently, glutamine metabolism has been reported to be a critical regulator of cell fate. Glutamine metabolism might coordinate the differentiation of three germ layers and is assumed to be preferred by the mesoderm [28, 29]. Glutamine metabolism contributes to maintaining hair follicle stem cell reversibility [11]. During intestinal development, glutamine plays a crucial role in the stem cell-mediated regeneration of small intestinal epithelial cells by targeting Wnt signaling [9]. GLS1 catalyzes the conversion of glutamine to glutamate, and GDH further transforms it into α-KG. In the current study, KLF6 overexpression enhanced the expression and activity of GLS1 and GDH, thereby increasing the levels of downstream metabolites of glutamine metabolism. The opposite results were obtained following KLF6 knockdown. These findings emphasize the regulatory function of KLF6 in glutamine metabolism in hDPSCs. KLF6’s dual role in regulating both GLS1 and GDH highlights its importance in maintaining metabolic homeostasis. By coordinating the expression of these two key enzymes, KLF6 ensures a balanced flow of metabolites through the glutamine pathway. A recent study reported that GLS1 and GDH dual inhibition prevents TGF-β1–induced collagen secretion in pulmonary fibrosis [30]. There is a physiologically relevant compensation mechanism between GLS1 and GDH. Modulating KLF6 activity might offer a way to simultaneously influence multiple points in the glutamine metabolism pathway, potentially providing more comprehensive treatments.

Moreover, glutamine metabolism is closely related to mineralized tissue formation, particularly bone formation and osteoblast differentiation. Glutamine metabolism is a major determinant of the proliferation and differentiation of skeleton stem cells [10]. Mechanistically, mesenchymal stem cell commitment to the osteogenic lineage rather than the adipogenic lineage can be ascribed to glutamine metabolism because glutamine withdrawal or GLS knockout can promote adipogenic potency [31]. In addition, parathyroid hormone-induced bone formation is related to glutamine metabolism [32]. Notably, skeletal unloading-induced disuse osteoporosis is related to glutamine metabolism, and glutamine supplementation can rescue bone mechanical sensitivity and promote bone regeneration [33]. Kim et al. reported that glutamine is essential for the growth, migration, and differentiation of hDPSCs [34]. However, they did not detect changes in glutamine metabolism during odontogenic differentiation or explore the underlying mechanisms. Our study using ¹³C-glutamine tracking revealed that glutamine is consumed mainly to fuel the TCA cycle in differentiated hDPSCs.

Glutamine serves as a crucial carbon source for the TCA cycle to support the cellular energy requirements [35]. It can generate α-KG that enters the TCA cycle. In the present study, mitochondrial respiration was also enhanced during the odontogenic differentiation of hDPSCs, and KLF6, which functions as a promoter of odontogenic differentiation, induced mitochondrial respiration in hDPSCs. These results are consistent with the consensus on the common metabolic features observed during stem cell differentiation that further differentiation of stem cells is often accompanied with increased mitochondrial oxidation [36].

Type I collagen is the main collagen fiber of dentin, whereas teeth and bones are highly mineralized and hierarchically structured organic–inorganic composites, formed by mechanisms that largely remain to be elucidated. Citrate, a major intermediate of the mitochondrial TCA cycle, may play a key role in this process. Citrate can be transported out of the mitochondria by SLC25A1 [37]. In 1941, Dickens first stated that mineralized tissues store approximately 90% of the total in vivo citrate, which has a high binding affinity to Ca and regulates bone metabolic and structural integrity [38]. Citrate can facilitate the crystallization of amorphous Ca phosphate in collagen fibers, resulting in interactions between the organic matrix and inorganic minerals [23]. Furthermore, citrate can stabilize two precursors of hydroxyapatite, expediting their infiltration into the cavities of collagen fibrils and enhancing collagen mineralization [39, 40]. Bones and teeth have a large citrate storage [41]. Fu et al. have reported that citrate participates in ECM mineralization in osteoblasts [42]. Dirckx et al. found that the level of serum-deprived citrate deposition into mineral hydroxyapatite affects bone strength, but they did not explore whether endogenous citrate production by the matrix-secreting cells themselves also influences matrix mineralization [15]. Similarly, in the present study, glutamine deprivation significantly restrained the deposition of citrate and Ca in the ECM, and ¹³C-glutamine tracking assays confirmed that KLF6-mediated glutamine metabolism promoted matrix mineralization by increasing the deposition of citrate and collagen in the ECM. Our results suggest that endogenous citrate production in differentiated functional odontoblast-like cells also affects dentinal matrix mineralization.

Conclusions

This study revealed that KLF6 not only regulates glutamine metabolism, which supports the energy demands and biosynthesis required by differentiated hDPSCs, but also mediates glutamine undergoing anaplerotic reactions to form α-KG, which enters the TCA cycle. This promotes the deposition of citrate in the ECM, which consequently promotes matrix mineralization. We expect these findings to inspire the development of novel strategies for pulp dentin regeneration and reparative dentin formation.