Identification of GLDN+ odontogenic stem cells as crucial for human tooth development and regeneration
Chengcheng Liao, Jinglun Liu, Maojiao Li, Bingqian Yang, Yejia Yu, Jian Yang, Xiaoxia Su, Shixing Ma, Hanchao Li, Jingyi Zhang, Weidong Tian, Li Liao
TL;DR
This study identifies GLDN+ stem cells as crucial for human tooth development and regeneration, offering new insights into dental tissue repair.
Contribution
The study identifies GLDN+ odontogenic stem cells as a novel subpopulation critical for human dental pulp development and regeneration.
Findings
GLDN+ DPSCs exhibit enhanced self-renewal and odontogenic differentiation potential.
GLDN+ DPSCs induce endothelial cell migration and tube formation essential for tooth development.
GLDN+ DPSCs regenerate a vascularized dental pulp structure in vivo.
Abstract
The dental papilla (DP) is essential for the development of dentin and pulp. The extensive cellular heterogeneity within the DP is a critical factor underlying the complex and precise formation of dental structures during odontogenesis. However, the critical cell types within human DP that play essential role in tooth development and regeneration remain largely uncharacterized. In this study, we analyzed the heterogeneity of human DP cells using single-cell sequencing and identified Gliomedin (GLDN)+ DP stem cells (DPSCs) were a group of progenitors at an early stage of tooth development and play a key role in the development of pulp and dentin. GLDN+ DPSCs strategically accumulate in human DP tissue near the interface of the newly formed dentin or pulp. Functional assays demonstrated that GLDN+ DPSCs exhibited enhanced self-renewal, migratory capacity, and odontogenic differentiation…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8- —the National Natural Science Foundation of China (32271365, 32471183, U21A20369), Sichuan Science and Technology program (2023YFS0151, 2023YFS0056), and the Fundamental Research Funds for the Central
- —the National Key Research and Development Program of China (2022YFA1104400)
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
Topicsdental development and anomalies · Mesenchymal stem cell research · Bone and Dental Protein Studies
Introduction
Dental pulp, centrally located within the tooth and encased by dentin, is a highly vascularized connective tissue, crucial for oral health.^1^ Pulpal and periapical diseases are among the most prevalent oral conditions, often necessitating root canal therapy (RCT) in their advanced stages.^2^ Although RCT is a well-established and effective procedure, the loss of pulp tissue can compromise pulp defense and sensory function and may predispose the tooth to fracture.^3^ Therefore, there is an urgent need for clinical strategies aimed at dental pulp regeneration. In the past decade, significant progress has been achieved in dental pulp regeneration research based on tissue engineering strategies.^1,4^ However, the insufficient understanding of the stem cell subpopulations involved in the formation of distinct dental tissues during tooth development considerably restricts the advancement of dental pulp regeneration research.^1^
Tooth development initiates with the formation of tooth germ consisting oral epithelium and dental mesenchyme derived from migrating cranial neural crest cells.^5^ In the tooth germ, the dental lamina ultimately gives rise to enamel, while the mesenchyme forms the dental papilla (DP) and the dental follicle (DF).^5^ Mesenchymal stem cells (MSCs) within the DP undergo proliferation and subsequently differentiate into distinct progenitor cell subtypes, which give rise to the majority of the tooth structure (dental pulp and dentin).^6–9^ DP MSCs (DPSCs) express well-established MSC markers, including Stro-1, CD29, CD73, CD90, CD105, CD106, and CD146.^6^ However, these traditional markers classify DPSCs as a homogeneous population, failing to differentiate between distinct subtypes within the DP. In MSCs research, cellular heterogeneity has become an increasingly important focus.^10^ In our previous study, we identified a distinct population of CD24^+^ pluripotent stem cells in mouse DP using three-dimensional spheroid culture, which play a crucial role in the formation of the dental pulp-dentin complex.^11^ Jamal et al.^12^ identified a perivascular population of NOTCH3^+^ cells in the DP, which was subsequently established as a marker for dental pulp stem cells in mice.^13,14^ Furthermore, NOTCH3^+^ cells have been implicated in the regulation of dental pulp tissue homeostasis and its repair following injury in mice.^13,14^ These findings underscore the substantial heterogeneity of DPSCs, suggesting that distinct subpopulations with diverse characteristics may play unique roles in the development and regeneration of dentin and dental pulp. A more comprehensive understanding of MSCs heterogeneity will help to fully understand tooth development and regeneration and harness the potential of stem cell-based therapies.
In recent years, single-cell RNA sequencing (scRNA-seq) has been applied to map the cell atlas of tooth development in both mice and humans.^15–22^ Single-cell transcriptomic analyses of tooth development^15–20^ have elucidated the heterogeneity of dental MSCs, providing detailed insights into critical cell types, differentiation pathways, and molecular signaling mechanisms involved in odontogenesis.^15–20^ Jing et al.^15^ defined the specific contributions of distinct cellular domains to various molar mesenchymal tissues in murine models. Hu et al.^17^ revealed the direct involvement of Msh Homeobox 1 (Msx1)^+^/SRY-Box Transcription Factor 9 (Sox9)^+^ niche cells in murine molar development and their potential for tooth regeneration. During the bell stage, CD24^+^ cells in the upper DP and Placenta Associated 8 (Plac8)^+^ cells in the pre-odontoblast layer can independently induce non-dental epithelial cells to form tooth-like structures.^18^ Furthermore, the heterogeneity of MSCs within human dental pulp tissue has been well characterized,^21,22^ which is significant for understanding the mechanisms of dental pulp homeostasis and repair. However, these studies do not identify the critical MSCs populations within the human DP or investigate the regenerative potential of human DPSCs subpopulations.
In this study, we identified Transferrin (TF)^+^/WNT5A^+^/Gliomedin (GLDN)^+^ odontogenic stem cells (OSCs) using scRNA-seq data from human DP, which may play a significant role in the development of human dentin and dental pulp. Human GLDN^+^ OSCs demonstrate substantial self-renewal, migratory ability, dentinogenic differentiation potential, and a heightened capacity to induce angiogenesis in vitro, thereby facilitating the regeneration of pulp tissue characterized by a rich odontoblast layer and well-formed vascular networks. We further established that GLDN is indispensable for these cells to maintain their characteristics. Mechanistically, GLDN sustained the expression of BMP5, which activates the P-SMAD1/5/9 downstream signaling pathways in DP MSCs and endothelial cells through autocrine and paracrine mechanisms, thereby promoting tooth development and dental pulp regeneration. Moreover, GLDN^+^ OSCs were present in the DP of underdeveloped teeth and successfully achieved pulp regeneration after transplantation into the treated dentin matrix (TDM), indicating their potential as seed cells for dental pulp regeneration.
Results
scRNA-seq reveals the heterogeneity of human DP cells
To distinguish the heterogenous cells populations in the DP, we analyzed scRNA-seq data from the early root development stage (when the root length is less than one-third of its final length) of the human third molar DP, which were obtained from the public database (GSE202476).^23^ Following quality control and normalization of the scRNA-seq data, six major cell categories (11 subclusters) were identified based on the expression of specific markers (Fig. 1a, b): C-C Motif Chemokine Ligand 5 (CCL5)^+^/IL-32^+^/T Cell Receptor Beta Constant 2 (TRBC2)^+^ clusters 3 and 9 as lymphocytes; CD74^+^/Mannose Receptor C-Type 1 (MRC1)^+^/CD68^+^ clusters 5, 8, and 10 as macrophages; Chondroitin Sulfate Proteoglycan 4 (CSPG4)^+^/ACTA2^+^/KRT18^+^ Cluster 6 as pericyte; VWF^+^/CD31^+^/NOTCH4^+^ Cluster 2 as endothelial cells; S100B^+^/Proteolipid Protein 1 (PLP1)^+^/Peripheral Myelin Protein 2 (PMP2)^+^ Clusters 11 as nerve cells and Clusters 0, 1, 4, and 7 specifically express high levels of cranial neural crest mesenchymal cell markers, including Nestin (NES),^24^ SOX9,^17^ and MSX2^25^ (Figs. 1b and S1a–c), which allows us to identify these clusters as DP mesenchymal cells.Fig. 1. Bioinformatics analysis of the heterogeneity of human dental papilla cells. a UMAP representation of cells identified from scRNA-seq data (GSE202476) of human DP tissue. b Bubble plots illustrating the expression of marker genes across major cell types in human DP. The color of the dots indicates the expression levels, while their size represents the percentage of cells expressing the marker genes across different cell types. c Pseudotime trajectory of Clusters 0, 1, 4 and 7. The expression of GLDN (purple), β-catenin (green) (d), and CK14 (red) in the HERS region (e) and coronal region (f) of sagittal sections of human third molar DP tissues at an early developmental stage, where the root development length does not exceed one-third of its final length. RT root, DP dental papilla
We further investigated the heterogeneity and potential functions of four subsets of DP mesenchymal cells. Interestingly, we observed that cells in Cluster 1 exhibit high expression of the enamel matrix protein Ameloblastin (AMBN). Previous studies have also reported AMBN expression in pre-odontoblasts and young odontoblasts,^26^ with missense mutations in AMBN being associated with dentin disorders.^27^ The detection of AMBN expression in DP mesenchymal cells may thus offer important insights into the cellular source of AMBN in dentin. Additionally, the osteoblast-specific membrane protein Interferon-Induced Transmembrane Protein 5 (IFITM5)^28^ (Fig. S2c) is expressed in cells within Cluster 1, with its expression localized to the region adjacent to the dentin-forming area and the underlying tissue (Fig. S2d). Moreover, Dentin Sialophosphoprotein (DSPP)-positive odontoblasts are primarily localized within Cluster 1 (Fig. S2c). Gene Ontology (GO) enrichment analysis indicated that Cluster 1 cells are closely associated with mineralization-associated biological processes (Fig. S4b). Therefore, we propose that Cluster 1 predominantly consists of pre-odontoblasts and odontoblasts.
Cluster 4 cells are characterized by the expression of CD24a, Secreted Frizzled Related Protein 1 (SFRP1), and Wnt Inhibitory Factor 1 (WIF1) (Fig. S3a). GO analysis suggests that Cluster 4 cells are associated with the formation of the extracellular matrix and collagen fibers (Fig. S4c). Additionally, we observed that CD24a^+^ cells are primarily distributed in the central region of the DP, where the cell density is lower compared to the region where GLDN^+^ cells are distributed (Figs. S2b and S3b). Therefore, Cluster 4 cells may be primarily responsible for the formation of pulp matrix components.
MSX1, a key driver of early DP development,^17^ exhibits the highest expression in Cluster 7 (Fig. S1e). Additionally, in contrast to the potential Wnt-inhibited state of Cluster 4 cells (Fig. S4c), Cluster 7 exhibits the highest expression of CTNNB1 (β-catenin) (Fig. S5a), suggesting that this cell population may be responsive to canonical Wnt signaling. Transmembrane Protein 38B (TMEM38B) is expressed in clusters 0, 1, and 4, but is absent in cluster 7 (Fig. S3c). Within the DP tissue, β-catenin expression is widespread, with the highest levels in the coronal region (Fig. S3d). Moreover, β-catenin^+^/TMEM38B^-^ cells are predominantly located at the peripheral coronal region of the DP (Fig. S5c-e), appearing adjacent to the areas of Hertwig’s epithelial root sheath (HERS) (Fig. 1d, e). GO analysis indicated that Cluster 7 cells are associated with the process of “odontogenesis of dentin-containing tooth,” (Fig. S4d) and the pseudotime analysis results further support that Cluster 7 cells are positioned at a developmental starting point (Fig. 1c). Thus, MSX1^+^/CTNNB1^+^/TMEM38B^-^ Cluster 7 cells are likely localized at the peripheral coronal region of the DP (Fig. S5c–e) and may represent a population of developmental progenitor cells.
Similarly, TF^+^/WNT5A^+^/GLDN^+^ Cluster 0 cells (Fig. S2a) are likely primarily localized in coronal region and the adjacent areas surrounding the HERS within the DP (Figs. 1d–f and S2b), with minimal to no expression observed in other regions. Pseudotime analysis further indicates that Cluster 0 cells may be descendants of Cluster 7 cells (Fig. 1c). GO analysis suggests that Cluster 0 cells may have strong mineralization potential (Fig. S4a). Combined with pseudotime analysis, we hypothesize that Cluster 0 cells may differentiate into Cluster 1 cells and participate in dentin development (Fig. 1c). To further investigate the relationship between Cluster 0 cells and Cluster 7 cells, we examined the distribution of GLDN^+^ and CTNNB1^+^ in the DP (Fig. 1d). GLDN^+^ and CTNNB1^+^ cells are both most densely localized around the Cytokeratin 14 (CK14)^+^ HERS cells (Fig. 1e), but the distribution is reduced in the DP crown region (Fig. 1f). Therefore, we hypothesize that Cluster 7 cells may originate from the region surrounding the HERS epithelium, where they are regulated by the epithelium, and rapidly differentiate into Cluster 0 cells. Subsequently, Cluster 7 and Cluster 0 cells migrate towards the crown of the DP, giving rise to Cluster 1 and Cluster 4 cells, thereby contributing to the formation of the pulp and dentin.
Histological analysis of the DP from the early stages of root development to near completion revealed a high abundance of GLDN expression in regions adjacent to the developing dentin (Fig. S6). We observed that GLDN^+^ cells in the human DP migrate from the sides of the coronal region (adjacent to dentin) (Fig. S6a) towards the central area of the DP (Fig. S6b, c), which represents the junction between the DP and the developing dental pulp. The number of GLDN^+^ cells initially increased with root elongation (Fig. S6), followed by a gradual decrease as the root approached completion (Fig. S6c–e). As the root canal matures, the GLDN^+^ cells at this junction gradually disappear (Fig. S6d, e). Subsequently, we analyzed the scRNA-seq data from fully developed dental pulp tissue of five individuals aged 18-35 years^22^ (Fig. S7a, b), where a small population of GLDN^+^ cells was identified (Fig. S7c). Additionally, in vitro culture of primary dental pulp cells derived from a 25-year-old patient revealed the presence of 1.49% GLDN^+^ cells (Fig. S7d). These results suggest that the dynamic changes of GLDN^+^ cells during development may be closely related to the formation of pulp and dentin.
Based on the results, we believe that GLDN^+^ Cluster 0 may consists of early multipotent stem cells directly derived from DP progenitors, playing a potential role in dentin and pulp formation. Due to the lack of more specific markers and the relatively low cell number, Cluster 7 cells may not be further purified or expanded in vitro. Therefore, Cluster 0 cells were selected for further study.
GLDN+ DPSCs exhibit robust self-renewal, migration, dentinogenesis, and proangiogenic function
We selected the human third molar DP from developmental stages consistent with the scRNA-seq analysis (tooth root development not exceeding one-third of the final root length, with patient age ranging from 13 to 22 years) as the tissue source for in vitro studies of DP cells (DPCs). Additionally, the membrane protein GLDN^29^ was used to distinguish cluster 0 cells from other clusters. Approximately 10% of the cells in the first generation of DPCs were GLDN^+^ (Fig. S8a). Subsequently, GLDN^+^ DPCs were isolated from the DP using fluorescence-activated cell sorting (FACS) for further analysis, while GLDN^-^ DPCs were selected as the control (Fig. 2a). The isolated GLDN^+^ DPCs exhibit higher levels of GLDN protein than GLDN^-^ DPCs (Fig. 2b, c), and flow cytometry analysis revealed that the GLDN^+^ DPCs retain over 70% positivity at passage three (P3) (Fig. S8c). However, by P4 (Fig. S9a) and P5 (Fig. S9b), the positivity rate of GLDN in GLDN^+^ DPCs decreased to 65.2% and 56.9%, respectively. These results suggest that GLDN^+^ DPCs can be stably cultured in vitro up to P3, and thus, P3 GLDN^+^ DPCs were used for subsequent studies. Additionally, both GLDN^+^ and GLDN^−^ DPCs express MSCs markers CD44, CD90, and CD105 at P3, while lacking expression of CD31, CD34, or CD45 (Fig. S8b, c). Consequently, we conclude that the isolated and cultured GLDN^+^ and GLDN^−^ DPCs are MSCs derived from DP.Fig. 2. Isolation, identification, and comparative characterization of GLDN^−^ and GLDN⁺ DPSCs in vitro. a GLDN^−^ and GLDN^+^ DPSCs were isolated by fluorescence-activated cell sorting (FACS). b, c The gel image shows GLDN expression in GLDN^−^ and GLDN^+^ DPSCs (c), with statistical analysis of band absorbance (d). n = 3. Alkaline phosphatase (ALP) staining (d, e) (n = 6) and alizarin red staining (f, g) (n = 6) were performed to assess the mineralization potential of GLDN^−^ and GLDN^+^ DPSCs. h, i Western blot analysis was performed to assess the expression levels of DSPP, DMP1, and RUNX2 in GLDN^−^ and GLDN^+^ DPSCs on days 3, 7, and 14 of osteogenic induction. n = 3. j, k The vascular formation assay was performed to investigate the differences in tube formation induced by CM derived from GLDN^-^ and GLDN^+^ DPSCs on HUVECs. n = 5. Nb junctions, Nb meshes, and Tol. length represent the total number of junctions formed, the total number of meshes, and the total tube length, respectively. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.000 1; ns no significance
During tooth development, MSCs must undergo self-renewal, migration, and differentiation to contribute to the formation of tissues within distinct microenvironments. As expected, GLDN^+^ DPSCs demonstrate enhanced colony formation (Fig. S10a, b), proliferation (Fig. S10c), migration (Fig.e S10d–g), and mineralization potential (Fig. 2d–g) compared to GLDN^-^ DPSCs. Following osteogenic induction, GLDN^+^ DPSCs exhibit higher expression levels of dentinogenesis-related proteins DSPP, Dentin Matrix Acidic Phosphoprotein 1 (DMP1), and RUNX Family Transcription Factor 2 (RUNX2) (Fig. 2h, i). These results are consistent with bioinformatic analysis (Fig. S4a), highlighting the dominant role of GLDN^+^ DPSCs in odontogenic differentiation.
Vascular formation is a critical event in the development and regeneration of dental pulp. To explore the role of GLDN^+^ DPSCs in angiogenesis during DP development, we collected conditioned media (CM) from GLDN^+^ (GLDN^+^ CM) and GLDN^−^ DPSCs (GLDN^−^ CM), respectively. HUVEC cells cultured with GLDN^+^ CM and GLDN^−^ CM exhibited enhanced migration (Fig. S10h–k), as well as increased numbers of tube junctions (Nb junctions), meshes (Nb meshes), and total tube length (Tol. length) (Fig. 2j, k), compared to those cultured in standard medium. Importantly, GLDN^+^ CM demonstrated significantly stronger promotive effects on both migration and tube formation than GLDN^−^ CM (Figs. 2j, k and S10h–k). Taken together, these findings suggest that GLDN^+^ DPSCs possess a stronger potential for odontoblast differentiation and are closely associated with pulp vascular development.
GLDN+ DPSCs regenerate vascularized dental pulp in vivo
To evaluate the potential role of GLDN^+^ DPSCs in human dental pulp development and regeneration in vivo, we established an ectopic pulp regeneration model (Fig. 3a). We prepared TDM from human teeth and subsequently re-seeded GLDN^+^ and GLDN^−^ DPSCs onto type I collagen gels, which were then incorporated into the TDM and transplanted subcutaneously into nude mice (Fig. 3a). After 4 weeks, in the GLDN^+^ DPSCs group, we observed an increased attachment of cells to the surface of the TDM (Fig. 3b, c) compared to the GLDN^-^ DPSCs group and the Blank group (where no cells were seeded onto type I collagen). Additionally, a denser vascular network was observed in the regenerated dental pulp tissue of the GLDN^+^ DPSCs group (Fig. 3d, e). Furthermore, the collagen fiber area in the regenerated dental pulp of the GLDN^+^ DPSCs group was significantly higher than that in the other two groups (Fig. 3d, f). To further clarify the advantages of GLDN^+^ DPSCs in odontoblast differentiation and the induction of angiogenesis in vivo, we labeled odontoblasts on the TDM surface using DSPP and DMP1 (Fig. S11a), and mapped the distribution of mature blood vessels using VEGF and CD31 markers (Fig. S11d). The results indicate that the TDM surface in the GLDN^+^ DPSCs group exhibited a greater distribution of DSPP^+^/DMP1^+^ odontoblasts (Fig. S11a–c), and the area of mature blood vessels formed in the regenerated dental pulp was significantly larger in the GLDN^+^ DPSCs group compared to the other two groups (Fig. 3d, e). Based on the aforementioned findings and supported by the in vitro results (Figs. 2a–k and S11a–k), we propose that GLDN^+^ DPSCs contribute to the formation of dentin and vascularized dental pulp during tooth development and regeneration. Therefore, these cells are designated as GLDN^+^ OSCs.Fig. 3. Ectopic dental pulp regeneration using GLDN^+^ OSCs. a Flowchart illustrating the experimental procedure for ectopic dental pulp regeneration. Following flow cytometric sorting and in vitro culture, GLDN^+^ OSCs and GLDN^−^ DPSCs were mixed with type I collagen gel, after which the cell-laden scaffolds were incorporated into treated dentin matrix (TDM) derived from human teeth and implanted subcutaneously into nude mice for four weeks. As a positive control, type I collagen gel without cells was also embedded in TDM and subjected to the same procedure as the cell-seeded groups. Assessment of the efficacy of ectopic dental pulp regeneration by GLDN^+^ OSCs using H&E staining (b) and Masson staining (d). c The number of odontoblast-like cells adhered to each TDM surface was quantified and subjected to statistical analysis. n = 6. e The number of mature blood vessels in each regenerated dental pulp was quantified and subjected to statistical analysis. n = 6. f Quantification and statistical analysis of the proportion of collagen fiber area (blue) relative to the total pulp area in regenerated dental pulp. n = 6. **P < 0.01, ***P < 0.001, ****P < 0.000 1
GLDN is a phenotype-maintaining factor for GLDN+ OSCs
Previous research has established the critical role of GLDN in the development and maintenance of the peripheral nervous system;^30^ however, its regulatory function in DPSCs and in tooth development remains uninvestigated. To investigate whether GLDN played a regulatory role in the function of GLDN^+^ OSCs, we silenced the expression of GLDN using siRNAs (Fig. 4a, b). Unexpectedly, silencing GLDN expression inhibited the proliferation (Fig. S12a), migration (Fig. S12b–e), and mineralization potential (Fig. 4c–f) of GLDN^+^ OSCs. Additionally, silencing of GLDN significantly reduced the expression of DSPP, DMP1, and RUNX2 in GLDN^+^ OSCs after osteogenic induction (Fig. 4g, h). These findings suggest that GLDN plays a crucial role in maintaining the self-renewal, migration, and differentiation potential of GLDN^+^ OSCs.Fig. 4. Effects of GLDN silencing on the mineralization, and paracrine potential of GLDN^+^ OSCs. Representative Western blot images (a) and statistical analysis (b) of the silencing efficiency of GLDN in GLDN^+^ OSCs. n = 3. Overview and microscopic images (c) of alkaline phosphatase (ALP) staining after 7 days of osteogenic induction post-siRNA transfection in GLDN^+^ OSCs, with statistical analysis of integrated optical density (IOD) (d). n = 6. Overview and microscopic images (e) of Alizarin red staining after 14 days of osteogenic induction following siRNA transfection in GLDN^+^ OSCs, with statistical analysis of IOD (f). n = 6. g, h Effects of GLDN silencing on the expression of DSPP, DMP1, and RUNX2 in GLDN^+^ OSCs during osteogenic induction at days 3, 7, and 14. n = 3. i, j The impact of GLDN silencing in GLDN^+^ OSCs CM on HUVEC tube formation. n = 5. NC negative control; si-1, siGLDN-1; si-3, siGLDN-3. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.000 1; ns no significance
We also investigated the effects of GLDN silencing on the paracrine activity of GLDN^+^ OSCs. Silencing GLDN reduced the capacity of GLDN^+^ OSCs CM to promote proliferation (Fig. S13a), migration (Fig. S13b–e), and tube formation (Fig. 4i, j) in HUVEC cells. These experiments on vasculogenesis suggest that GLDN is crucial for regulating the paracrine function of GLDN^+^ OSCs.
GLDN promotes the secretion of BMP5 protein to sustain the function of GLDN+ OSCs
We further investigated the paracrine regulatory interactions between Cluster 0 cells and Cluster 2 cells (endothelial cells) through CellPhoneDB4 cell-cell communication analysis, based on scRNA-seq data (Fig. 5a). Several key developmental-related cytokines, including BMP5, BMP6, BMP7, Neural cell adhesion molecule 1 (NCAM1), Nerve Growth Factor (NGF), and Angiopoietin 1 (ANGPT1), are enriched in Cluster 0 (Fig. 5a). qRT-PCR analysis confirmed that BMP5 and the neurotrophic factors NCAM1 and NGF are significantly upregulated in GLDN^+^ OSCs compared to GLDN^-^ DPSCs (Fig. 5c). Additionally, Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis of Cluster 0 cells revealed that TGF-β was the most significantly enriched signaling pathway (Fig. 5b). Previous studies have demonstrated that BMP5 can enhance the expression of phosphorylated SMAD (P-SMAD),^31,32^ thereby activating the classical TGF-β signaling pathway. Therefore, we hypothesize that GLDN^+^ OSCs may influence both themselves and endothelial cells through the secretion of BMP5.Fig. 5. Cytokines and signaling pathways enriched in GLDN^+^ OSCs revealed by scRNA-seq analysis. a CellChat results illustrating interactions between Clusters 0 and 2. b KEGG enrichment analysis of Cluster 0 cells. c qRT-PCR analysis of the expression levels of enriched ligands in Cluster 0, based on the CellChat results (a). n = 3 or 4. *P < 0.05, **P < 0.01, ****P < 0.000 1; ns no significance
We observed that BMP5 was specifically expressed in Cluster 0 (Fig. 6a), and its protein co-localized significantly with GLDN in human DP tissue (Fig. 6b). In vitro, compared to GLDN^-^ DPSCs, GLDN^+^ OSCs also express higher levels of BMP5 (Fig. 6c, d) and phosphorylated SAMD1/5/9 (P-SMAD1/5/9) proteins (Fig. 6c, e). Using ELISA, we quantified BMP5 protein levels exceeding 400 pg/mL in the CM of GLDN^+^ OSCs, while the BMP5 concentration in the CM of GLDN^-^ DPSCs was below 200 pg/mL (Fig. 6f). HUVEC cells treated with GLDN^+^ OSCs CM exhibited increased levels of P-SMAD1/5/9 than GLDN^-^ DPSCs CM (Fig. S14a, b). Additionally, silencing GLDN reduced the expression of BMP5 and P-SMAD1/5/9 within GLDN^+^ OSCs (Fig. 6g–i), as well as BMP5 secretion in their CM (Fig. 6j). Furthermore, silencing GLDN was able to reverse the promoting effect of GLDN^+^ OSCs CM on P-SMAD1/5/9 expression in HUVECs (Fig. 6k, l). These results demonstrate that GLDN regulates the expression and secretion of BMP5 in GLDN^+^ OSCs.Fig. 6GLDN regulates BMP5/P-SMAD1/5/9 signaling in GLDN^+^ OSCs. a uMAP representation of BMP5. b IF co-staining of GLDN (green) and BMP5 (red) in sagittal sections of human developing DP tissues at an early root development stage (with root development not exceeding one-third of its final length), with DAPI (blue) indicating cell nuclei. Gel images (c) and grayscale value analysis of BMP5 (d) and P-SMAD1/5/9 (e) expression in GLDN-negative cells and GLDN^+^ OSCs. n = 3. f ELISA analysis of BMP5 expression levels in CM from GLDN^-^ DPSCs and GLDN^+^ OSCs. n = 4. g–i Effects of GLDN silencing on BMP5 and P-SMAD1/5/9 expression in GLDN^+^ OSCs. n = 3. j Expression of BMP5 in GLDN^+^ OSCs-derived CM after GLDN silencing was assessed by ELISA. n = 4. k, l Effects of CM from GLDN^+^ OSCs with GLDN silencing on P-SMAD1/5/9 expression in HUVEC cells. n = 3. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.000 1; ns no significance
The role of BMP5 in tooth development remains largely underexplored.^33^ We silenced BMP5 expression in GLDN^+^ OSCs using siRNA (Fig. 7a, b), resulting in further inhibited P-SMAD1/5/9 expression (Fig. 7a, c). In contrast, treatment with exogenous BMP5 recombinant protein upregulated P-SMAD1/5/9 expression in GLDN^-^ DPSCs (Fig. S15a, b). Consistent with the changes in P-SMAD1/5/9 expression, reduced BMP5 expression inhibited the proliferation (Fig. S15c), migration (Fig. S15d–g), and mineralization potential (Fig. 7d–g) of GLDN^+^ OSCs. Furthermore, silencing BMP5 reduced the expression of DSPP, DMP1, and RUNX2 in GLDN^+^ OSCs following osteogenic induction (Fig. 7h, i). In contrast, the addition of BMP5 recombinant protein promoted proliferation (Fig. S16a, b), migration (Fig. S16c–f), and mineralization potential (Fig. 7j-m) of GLDN^-^ DPSCs. These results underscore the critical role of the BMP5/P-SMAD1/5/9 signaling pathway in maintaining the self-renewal, migration, and odontogenic differentiation of GLDN^+^ OSCs.Fig. 7. The role of BMP5/P-SMAD1/5/9 signaling in regulating the mineralization potential of GLDN^+^ OSCs. Validation of the silencing efficiency of BMP5 (a, b) and its effects on the expression of P-SMAD1/5/9 (a, c) in GLDN^+^ OSCs. n = 3. Alkaline phosphatase (ALP) (d, e) (n = 6) and Alizarin red staining (f, g) (n = 6) were utilized to assess the effects of BMP5 silencing on the mineralization potential of GLDN^+^ OSCs. h, i Effects of BMP5 silencing on the expression of DSPP, DMP1, and RUNX2 in GLDN^+^ OSCs during osteogenic induction on days 3, 7, and 14 are reported. n = 3. ALP staining (j, k; n = 6) and Alizarin Red staining (l, m; n = 6) to assess the effect of different concentrations of BMP5 on the mineralization potential of GLDN^−^ DPSCs. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.000 1; ns no significance
To investigate the effect of BMP5 secreted by GLDN^+^ OSCs on endothelial cells, we collected the GLDN^+^ OSCs CM after BMP5 silencing. Silencing BMP5 leads to decreased BMP5 expression in GLDN^+^ OSCs CM (Fig. 8a) and diminishes the promotion of P-SMAD1/5/9 expression in HUVEC cells mediated by this CM (Fig. 8b, c). Moreover, silencing BMP5 further diminished the ability of GLDN^+^ OSCs CM to promote HUVEC cell proliferation (Fig. S17a), migration (Fig. S17b–e), and tube formation (Fig. 8d, f–h). Supplementation with recombinant BMP5 protein (10 and 50 ng/mL) significantly augmented the expression of P-SMAD1/5/9 in HUVEC cells (Fig. 8e, i). Furthermore, BMP5 recombinant protein also enhanced the proliferation (Fig. S18a, b), migration (Fig. S18c–f), and tube formation (Fig. 8j–m) of HUVEC cells. Notably, the enhancement of HUVEC proliferation, migration, and tube formation by BMP5 is not strictly concentration-dependent (Figs. 8j–m and S18a–f); higher BMP5 concentrations (200 ng/mL) may prove less effective than lower concentrations (Figs. 8j–m and S18a–f). Taken together, these results clarify the promotive role of GLDN-mediated autocrine and paracrine BMP5 secretion by GLDN^+^ OSCs in dentinogenesis and angiogenesis.Fig. 8. Effects of exogenous BMP5/P-SMAD1/5/9 signaling on the tube formation capacities of HUVECs. a Expression of BMP5 in GLDN^+^ OSCs-derived CM after BMP5 silencing. n = 4. b, c Expression of P-SMAD1/5/9 in HUVECs cultured with GLDN^+^ OSCs CM after BMP5 silencing. n = 3. d, f–h The impact of BMP5 silencing in GLDN^+^ OSCs CM on HUVEC tube formation. n = 5. e, i The effects of varying concentrations of BMP5 recombinant protein on the expression of P-SMAD1/5/9 in HUVEC are presented. n = 3. j–m The effect of different concentrations of BMP5 on tube formation in HUVEC. n = 5. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.000 1; ns no significance
Discussion
Current evidence suggests that DP-derived MSCs constitute a heterogeneous population,^15,17^ which poses a major barrier to their clinical applications.^10^ Specific cells within the DP may play unique roles at specific stages or throughout the entire process of tooth development.^11,15,17,34^ For instance, LHX6^+^ cells within the mouse molar DP determine the number of tooth roots.^34^ In this study, we identified a novel population of GLDN-marked OSCs in human developing DP tissue, which are important for dentin and vascularized pulp formation (Fig. S19). GLDN supports the secretion of BMP5 from GLDN^+^ OSCs, binding to MSCs and endothelial cells in the microenvironment, thereby activating the BMP5/P-SMAD1/5/9 axis to regulate dentin and vascular development. GLDN^+^ OSCs exhibit self-renewal, migration, and mineralization capacities, while actively modulating the local microenvironment, positioning them as an ideal candidate for dental pulp regeneration (Fig. S19).
The DP, at the base of the developing crown or root, is responsible for early tooth crown formation, root elongation, and maturation. Due to the accessibility of DP following tooth extraction and its rich stem cell population, DP-derived stem cells represent a promising source for regenerative medicine. During tooth mesenchyme development, cells transition from a relatively homogeneous state to a highly heterogeneous state, correlating with the morphogenesis and development of various tooth structures.^15^ In the present investigation, we identified TF^+^/WNT5A^+^/GLDN^+^ as a critical subpopulation of MSCs during early human DP development, highlighting their potential as key seed cells for dental pulp regeneration. We previously explored the ability of postnatal mouse CD24a^+^ DPCs to regenerate dental pulp with pulp-dentin complexes following 3D culture.^11^ Our findings further confirm that CD24a^+^ DPCs may represent a population of early developmental stem cells. However, previous research indicates that CD24a^+^ cells within the DP display diminished self-renewal capacity in 2D culture,^35,36^ potentially due to Wnt inhibition status in this population (Fig. 1e), while their mineralization potential remains inconclusive.^35,36^ These limitations may restrict the applications of CD24a^+^ DPCs. Our findings identify GLDN^+^ OSCs as a subpopulation with enhanced odontoblastic differentiation potential under 2D culture, maintaining this differentiation advantage even after implantation into animal models. Additionally, we highlighted their strong angiogenic induction capacity, as the establishment of a vascular network is essential for the development and regeneration of nearly all vital organs in the human body. We propose that the secreted products of GLDN^+^ OSCs may have promising applications in the repair of vascularized tissues and organ damage.
In vitro, GLDN^+^ OSCs demonstrated superior proliferation, migration, odontoblastic differentiation, and angiogenic potential compared to GLDN^−^ DPSCs. However, in vitro models cannot fully replicate the biological roles of GLDN^+^ OSCs during pulp development and regeneration. The TDM-based ectopic pulp regeneration model provides extended observation periods, improved biomimetic structures, and a more accurate representation of the in vivo microenvironment.^37,38^ Utilizing this model, we further established that GLDN^+^ OSCs can generate a substantial population of odontoblast-like cells at the interface between TDM and regenerated dental pulp, underscoring their potential in pulp-dentin complex formation. Furthermore, the regenerated pulp exhibited a more robust collagen fiber and vascular network. These results not only suggest a pivotal role for GLDN^+^ OSCs in pulp development but also position them as an ideal candidate for dental pulp regeneration, with significant prospects for clinical application (Fig. S19).
Achieving functional dental pulp regeneration requires reconstruction of the pulp-dentin complex and neurovascular regeneration. Unfortunately, no mature nerve fiber ingrowth was observed in our model, likely due to the brief modeling duration. In our study, we observed that GLDN^+^ OSCs exhibit elevated expression of BMP5),^39–41^ NCAM1,^42^ and NGF,^43^ which are implicated in neural development, neuroprotection, or regeneration. Additionally, secreted GLDN plays a critical role in promoting nerve development by facilitating the formation of nodes of Ranvier, which are essential for efficient nerve signal conduction.^44^ GLDN interacts with neuroglial cells, particularly Schwann cells, to mediate the clustering of nodal proteins such as voltage-gated sodium channels and adhesion molecules like NrCAM at the nodes.^45^ These interactions are crucial for the proper myelination and organization of the axon-glial interface, supporting the maturation, stabilization, and regeneration of neural circuits during development. These findings suggest that GLDN^+^ OSCs and their secretome may not only induce vascular formation but may also confer advantages in promoting neural development and regeneration.
Aside from dental pulp regeneration, DPSCs have been widely studied for their roles in bone, neural, vascular regeneration, immune modulation, and other related fields.^8,46^ Considering the robust mineralization and angiogenic potential of GLDN^+^ OSCs, we hypothesize that these cells may offer significant promise for bone and vascular regeneration, which warrants further exploration in future studies. However, the limited in vivo lifespan of human DP presents a significant barrier to the clinical application of GLDN^+^ OSCs. Additionally, we found that the proportion of GLDN^+^ cells diminished as the passage number of GLDN^+^ OSCs increased. Maintaining the stability of this cell population in long-term in vitro culture remains a challenge requiring further investigation. Notably, we have also discovered GLDN^+^ cells in mature human dental pulp. Given the relative ease of access to dental pulp compared to DP, isolating GLDN^+^ cells or other cell subpopulations from dental pulp holds significant clinical potential. This represents a key challenge we aim to address in future studies.
Tooth development and regeneration are orchestrated by an complex network of regulatory signaling pathways. We highlight the involvement of the GLDN/BMP5/P-SMAD1/5/9 axis in tooth development and dental pulp regeneration. Current research primarily links GLDN primarily to neural development,^44^ while its role in other systems remains largely unexplored. Our findings indicate that GLDN influences self-renewal, migration, mineralization, and paracrine potential in GLDN^+^ OSCs by modulating the BMP5/P-SMAD1/5/9 pathway, which aids in understanding the significance of GLDN in MSCs. BMPs are essential for bone and cartilage formation, as well as for maintaining bone function in adulthood.^47^ BMP5 gene mutations in mice can affect the size and shape of various bone and cartilage structures during embryonic development.^48^ A previous study suggested BMP5 may be involved in enamel formation,^33^ and we further confirmed its role in dentin and pulp vasculature formation. BMP5 has been shown to activate P-SMAD1/5/9,^48^ and P-SMAD1/5/9 is known to promote mineralization in MSCs and vasculogenesis in endothelial cells.^49^ The discovery of the GLDN/BMP5/P-SMAD1/5/9 developmental signaling axis in the DP provides a potential therapeutic target for dental pulp tissue regeneration.
In summary, our study demonstrates that GLDN^+^ OSCs promote dentin and vascular formation via the GLDN/BMP5/P-SMAD1/5/9 signaling pathway and may also play a role in inducing dental pulp nerve formation. GLDN^+^ OSCs and their secretome are not only promising for dental pulp regeneration but also offer potential applications in the reconstruction of hard tissues and neurovascular networks in other regions.
Materials and methods
Statistical analysis of scRNA-seq data
The scRNA-seq data of the developing DP (GSE202476) were obtained from a 13-year-old individual with an immature third molar, in the early stages of root development, where the root length is less than one-third of its final length. Meanwhile, the scRNA-seq data for human mature dental pulp (GSE4998457, GSE4998458, GSE4998459, GSE4998460, GSE4998461) were sourced from five individuals aged 18-35 years, all with fully developed molars. All data were subsequently analyzed using the NovelBrain Cloud Analysis Platform (www.novelbrain.com). Fastp was employed with default parameters to filter adaptor sequences and remove low-quality and short reads, resulting in clean data.^50^ CellRanger v6.1.1 was used to align reads to the human genome (GRCh38 Ensembl: version 91) to generate feature-barcode matrices. Subsequent downsampling analysis was conducted using the CellRanger Aggr function to balance the mapped barcoded reads per cell across the sequenced samples, ultimately achieving the aggregated matrix. Cells with more than 200 expressed genes and a tissue-specific mitochondrial UMI rate below 40% passed the cell quality filtering. Mitochondrial genes were then removed from the expression table. The Seurat package (version 4.0.3) was employed for cell normalization and regression based on UMI counts and mitochondrial percentage, resulting in scaled data. Principal component analysis (PCA) was conducted on the scaled data using the top 2 000 highly variable genes. The top 10 principal components were then used for uniform manifold approximation and projection (UMAP) construction.
Differentially expressed genes (DEGs) in cell subgroups were identified using the FindMarkers function in Seurat, with cut-off criteria of |logFC| > 2 and adjusted P value < 0.05. KEGG (www.genome.jp/kegg) and GO (geneontology.org) were utilized for pathway and gene ontology analyses by retrieving and mapping gene sets. Furthermore, cell-cell communications were analyzed using CellPhoneDB.^51^ Based on the interactions and the normalized cell matrix obtained through Seurat normalization, the mean significance and cell communication significance (P < 0.05) were computed.
Isolation and cultivation of DP or dental pulp Cells
DPCs were isolated from the DP tissue of incompletely developed third molars (the root length does not exceed one-third of the final root length) from patients aged 13–22 years. Dental pulp cells were derived from third molars (root development was completed) extracted from individuals aged 18–25 years without local inflammation. The harvesting, utilization, and disposal of human dental tissues were approved by the Ethics Committee of West China Hospital of Stomatology, Sichuan University (Approval No. WCHSIRB-D-2024-283), and informed consent was obtained from the patients or their parents/guardians in the case of minor patients. The principles of the Helsinki Declaration were strictly followed in this study. The DP or dental pulp tissue was processed in accordance with our previous protocols. In brief, the DP/pulp tissue was isolated, sectioned into small tissue blocks, and digested with 3 mg/mL type I collagenase (#17100017; Gibco) at 37 °C for 30 min. DPCs or dental pulp cells were maintained in α-MEM medium (#11095080; Gibco) containing 10% fetal bovine serum (FBS) (#10099158; Gibco) and 1% penicillin-streptomycin (#1514022; Gibco) at 37 °C and 5% CO_2_.
Flow cytometry
First-generation DPCs and third-generation sorted cells were utilized for fluorescence-activated cell sorter (FACS) and cell surface protein flow cytometric analysis, respectively. Briefly, adherent cells were dissociated into a single-cell suspension using a 0.02% EDTA cell dissociation solution (#G4050; Servicebio). Cells were incubated with fluorescently labeled antibodies at 4 °C for 1 h. After thorough washing with PBS, cell strainers (#352340; FALCON) were used to eliminate clumped cells. FACS and flow cytometric analysis were performed using a CytoFLEX SRT and BD Accuri C6 Plus, respectively.
Cell proliferation and colony formation assay
The CCK-8 assay kit (#KGA317; KeyGen BioTECH) was employed to evaluate cellular proliferation capacity. The working solution was prepared following the manufacturer’s instructions and incubated with cells at 37 °C for 70 min. The OD_450_ values were collected using a Synergy™ LX multi-mode reader as a measure of proliferation capability.
Single-cell suspensions were inoculated into wells of a 6-well cell culture plate at a density of 500 cells per well. After a 10-day incubation period, cells were fixed with paraformaldehyde (#P1110; Solarbio) and subsequently stained with crystal violet (#G1062; Solarbio). The accumulation of over 50 cells undergoing growth is considered to constitute a cellular colony.
Cell migration
The scratch assay and Transwell migration assay were employed to evaluate the migratory capacity of the cells. Cells were cultured in a 24-well plate until they reached optimal density, after which a cell-free zone was mechanically created. The migration of cells toward the cell-free zone was observed over a 24-h period. To minimize the potential interference of cell proliferation on experimental outcomes, the cells were cultured in α-MEM medium without FBS. The migratory capacity of the cells was quantified by calculating the percentage reduction in the area of the cell-free zone over a 24-h period using ImageJ software.
The experimental procedure for the Transwell migration assay is as follows: In the upper chamber of Transwell inserts (#3422; Corning), cell suspensions in α-MEM medium (5 × 10^5^ cells per mL) without FBS were added, while the lower chamber contained α-MEM medium supplemented with 10% FBS. After 24 h of incubation, the cells that had migrated passed through the Transwell membrane. Following fixation with paraformaldehyde, removal of the upper chamber cells, and crystal violet staining, the number of migrated cells was counted to evaluate cell migration ability.
Alkaline phosphatase and Alizarin Red staining
MSCs were induced to differentiate into the osteogenic lineage according to the guidelines provided by the OriCell® Human MSC Osteogenic Differentiation Kit (#HUXXC-90021; Oricell). After 7 days of osteogenic induction, the cells were fixed with paraformaldehyde and stained using the BCIP/NBT Alkaline Phosphatase Color Development Kit (#C3206; Beyotime) according to the manufacturer’s instructions. After 14 days of induction, the cells were fixed with paraformaldehyde and stained with Alizarin Red S to assess the mineralized matrix. The collected images were analyzed using ImageJ software to quantify and assess the mineralization potential of the cells.
qRT-PCR
Total RNA was extracted from the cells using the FastPure Cell/Tissue Total RNA Isolation Kit V2 (#7E730G3; Vazyme). mRNA was reverse transcribed into cDNA using the HiScript III RT SuperMix for qPCR (+gDNA wiper) kit (#R323-01; Vazyme). Relative quantitative PCR analysis was conducted on the QuantStudio 6 Flex Real-Time PCR System (Thermo Fisher Scientific, USA) using the Taq Pro Universal SYBR qPCR Master Mix kit (#Q712-02; Vazyme). Relative mRNA levels were quantified using the 2^−ΔΔCT^ method, with GAPDH as the internal control. The qRT-PCR primer sequences used in this study are listed in Table S1.
Conditioned medium (CM) preparation
When the cells reached passage 3, the complete culture medium was replaced with serum-free medium, and incubation continued until 85% confluency was achieved. The conditioned medium (CM) was then collected, centrifuged to remove debris and dead cells, and filtered through a 0.22 μm filter. The collected CM was supplemented with 10% FBS and 1% penicillin-streptomycin and used to culture the experimental group. Cells in the control group were cultured in complete culture medium.
To prepare the osteogenic induction CM, the osteogenic induction medium was mixed with CM in a 1:1 ratio. In the control group, a 1:1 ratio of osteogenic induction medium to α-MEM was used. The frequency and volume of osteogenic induction CM usage were maintained in accordance with the standard osteogenic induction protocol.
ELISA assay
The human bone morphogenetic protein 5 (BMP5) ELISA Kit (#JL13578; Jonlo) was used to quantify BMP5 expression levels in the CM. The final optical density was measured using a microplate reader at a wavelength of 450 nm. BMP5 levels in each group were determined by comparing the optical density readings with a standard curve.
Evaluation of angiogenic capacity in HUVECs
Ten microliters of Matrigel (#356234; Corning) was pre-coated on angiogenesis slides (#91506; ibidi) and incubated at 37 °C for 30 min. HUVECs (2 × 10^4^ cells per well), pre-treated with CM for 24 h, were then seeded on the Matrigel-coated angiogenesis slides. After 4 h of incubation at 37 °C, images were collected and analyzed using ImageJ software to assess the tube formation ability of HUVECs in each group.
Recombinant BMP5 protein
In the cell proliferation, migration, tube formation, and osteogenic induction assays, human BMP5 recombinant protein (HEK293, hFc) (#HY-P75472; MCE) was added to the culture medium or osteogenic induction medium at the onset of the experiment and maintained throughout. After 24 h of treatment with human BMP5 recombinant protein, total cellular proteins were extracted for subsequent Western blot analysis.
GLDN and BMP5 siRNA
The human BMP5 siRNA was acquired from HanBio (Shanghai) Co., Ltd. (Contract No.: HH20240705CDNWH-S102). The transfection of siRNA was performed using PepMute™ Transfection Buffer (5×) (#SL100575; SignaGen) and PepMute™ siRNA Transfection Reagent (#SL100566; SignaGen), in accordance with the manufacturer’s instructions. The siRNAs targeting GLDN were synthesized using the following sequences: hs-GLDN-si1: (5′-GCAUUUAUAUUCAUGGGAATT-3′), hs-GLDN-si2: (5′-CAGGAAACUUGUUUUUUUATT-3′), hs-GLDN-si3: (5′-CAAAGAGUCUCAUGUUCAATT-3’). The siRNAs targeting BMP5 were synthesized using the following sequences: hs-BMP5-si1: (5′-CGAAAGACGGGAAAUACAAAG-3′), hs-BMP5-si2: (5′-GAGUCGGAGUACUCAGUAAGGTT-3′), hs-BMP5-si3: (5′-GCUGAAUUCCGGAUAUACAAG-3′) and si-NC: (5′-UUCUCCGAACGUGUCACGUTT-3′).
Ectopic pulp regeneration model in nude mouse
All procedures involving animal experiments were performed in compliance with ethical standards and received approval from the Ethics Committee of the West China School of Stomatology at Sichuan University (Approval No. WCHSIRB-D-2024-484). The experimental procedures were conducted based on our previous experience.^37^ First, the roots of the extracted third molars were truncated and ground into dentin tubes approximately 6–8 mm in length. These tubes were then ultrasonically cleaned three times using sterile PBS for 10 min each time. Subsequently, the tubes were treated with 17%, 10%, and 5% ethylenediaminetetraacetic acid (EDTA) (#E8040; Solarbio) for 10 min each, followed by immersion in a 1% penicillin/streptomycin solution at 4 °C for a minimum of 3 days. According to the manufacturer’s instructions, 3 mg/mL collagen hydrogel (#230308; Cellmatrix) was subsequently prepared. The hydrogel, either containing (1 × 10^6^ cells per mL) or devoid of cells, was injected into the dentin tubes and incubated at 37 °C for 30 min. The TDM tubes loaded with hydrogel were implanted into the subcutaneous space on the backs of male Balb/c nude mice aged 4 weeks (GemPharmatech Co. Ltd). Each nude mouse was implanted with one TDM tube from each of the three experimental groups to minimize inter-animal variability, with a total of six mice used for the study. After 4 weeks, the TDM tubes were removed, fixed with 4% paraformaldehyde, and demineralized using 17% EDTA prior to histological analysis.
Tissue paraffin sections, HE and Masson staining
After decalcification of the dental tissue and TDM tubes with 17% EDTA, the samples were subjected to a series of graded alcohol, xylene, paraffin, and embedding procedures. Finally, the samples were sectioned into 6 μm-thick paraffin sections. Subsequently, the paraffin sections were deparaffinized with xylene and rehydrated through a graded series of alcohol solutions. The sections were then stained using the Hematoxylin and Eosin (HE) Staining Kit (#G1120; Solarbio) and the Modified Masson’s Trichrome Staining Kit (#G1346; Solarbio) for histological examination. The proportion of collagen fiber area to total pulp area in regenerated dental pulp tissue was quantified using ImageJ software following Masson’s trichrome staining.
Immunofluorescence analysis
To enhance membrane permeabilization, tissue sections, following rehydration, or cell spheres fixed in 4% paraformaldehyde were treated with 0.5% Triton X-100. Subsequently, antigen retrieval was conducted using pepsin antigen retrieval solution (#G0142; Servicebio) at 37 °C for 30 min, followed by blocking with 5% bovine serum albumin (BSA) (#4240GR100; BioFroxx). The primary antibody was incubated overnight at 4 °C. The secondary antibody and 4′,6-diamidino-2-phenylindole (DAPI) (#D8200, Solarbio) were incubated for 30 and 5 min, respectively, at room temperature. The antibodies used for immunofluorescence are listed in Table S2. Images were acquired using an Olympus confocal laser scanning microscope (VS200 or FV1200) for subsequent analysis. Fluorescence intensity was quantified using ImageJ software.
Western blot
RIPA lysis buffer (#KGB5203-100; KeyGEN BioTECH), supplemented with 1 mmol/L PMSF (#KGB5105-10; KeyGEN BioTECH), was used to lyse the cell samples. The cell lysates were thoroughly centrifuged at 4 °C, and the resulting supernatants were collected. Protein concentration was measured using the BCA Protein Assay Kit (#KGB2101-100; KeyGEN BioTECH), and the concentrations were subsequently normalized across all groups. After the addition of 4X SDS-PAGE loading buffer (with β-Mercaptoethanol) (#P1016; Solarbio) to the protein supernatant, the mixture was incubated at 100 °C for 5 min.
SDS-PAGE gels were prepared using the PAGE Gel Quick Preparation Kit (#PG112; Epizyme). The 180 kD Prestained Protein Marker (#MP102-01; Vazyme) served as a protein distribution indicator. The proteins were transferred from the SDS-PAGE gel to 0.45 μm PVDF membranes (#ISEQ00010; Millipore) using the wet transfer method. Subsequently, the PVDF membranes were blocked with 5% skim milk (#1172GR500; BioFroxx) for 2–4 h. The primary antibody was then incubated with the PVDF membranes overnight at 4 °C, followed by incubation with the HRP-labeled secondary antibody at room temperature for 1–2 h the following day. The antibodies used for western blotting are listed in Table S2. ECL chemiluminescent substrate (#180-5001; Tanon) was used for visualization. The PVDF membranes with proteins were thoroughly washed between each step with TBS buffer containing 0.1% Tween-20 (#9005-64-5; Solarbio).
The Western blot antibody stripping buffer (#PS107; Epizyme) was applied to the PVDF membrane for 15 min to strip the previously detected protein antibodies. After reblocking, the membrane was incubated with the next antibody to enable the detection of multiple protein expressions on the same PVDF membrane.
Quantification and statistical analysis
Data analysis and statistical graphing were performed using GraphPad Prism 9.5 software, where statistical data are presented as individual data points and as mean ± SD. The difference between two groups was analyzed using an unpaired two-tailed Student’s t test. For comparisons among multiple groups, one-way analysis of variance (ANOVA) with Tukey’s post hoc test was performed. A P-value of <0.05 was conducted statistically significant.
Supplementary information
Supplemental Materials Supplemental Tables