The treatment of immature permanent teeth with necrotic pulp is a challenging task for dental professionals. The young pulpless tooth with an open apex frequently has a thin and fragile wall, which increases the risk of tooth fracture and makes it difficult to obtain the necessary apical seal using conventional treatment methods. A novel treatment of pulp revascularization/regeneration is an effective approach for inducing root maturation (Iwaya et al., 2001; Banchs and Trope, 2004). During revascularization/regeneration procedures, the evoked-bleeding step leads to a substantial influx of stem cells from dental apical papilla (SCAPs) into the root canal, which contributes to root development and apical closure (Cao et al., 2015). However, several histological findings reported that most tissues formed in the canal of the revitalized immature permanent tooth lack organized pulp–dentin complex and consist of ectopic bone, fibrous tissues, and cementum (Martin et al., 2012; Saoud et al., 2015; Peng et al., 2017). In some cases, the root canal space became partly obliterated with ectopic bone (Martin et al., 2012; Saoud et al., 2015). It is presumed that stem cells from alveolar bone would migrate into the root canal and initiate ectopic calcification deposition (Cao et al., 2015; Saoud et al., 2015). Mesenchymal stem cells (MSCs) from different dentoalveolar tissues are unique and retain identities from their direct tissue sources. An understanding of the molecular mechanisms underlying the differentiation of odontoblasts and osteoblasts and their secretion of a mineralized matrix is fundamental for developing strategies for the regeneration of hard tissue. It is necessary to find some important factors to regulate the differentiation of stem cells into odontoblasts and osteoblasts.

SCAPs are considered to be the source of primary odontoblasts (Huang et al., 2008; Sonoyama et al., 2008). The interruption of SCAP development results in halted root development (Chueh and Huang, 2006). Stem cell–based approaches and tissue engineering have demonstrated that SCAPs can differentiate into odontoblasts and regenerate a continuous layer of dentin with a uniform thickness on an existing dentinal wall in vivo (Sonoyama et al., 2008; Huang et al., 2010). During bone formation, osteoblasts are derived from bone marrow–derived stem cells (BMSCs). BMSCs can differentiate into osteoblasts after culture in mineral-inducing medium (Dj, 1997; Bianco et al., 2001). When BMSCs are transplanted into the subcutaneous space of immunocompromised mice, they form bone-like complexes, while if the seeding cells are replaced by dental pulp stem cells, they form dentin-like complexes (Gronthos et al., 2000).

Bone and dentin are mineralized tissues mainly composed of calcium hydroxyapatite and type I collage. The process of hydroxyapatite nucleation and collagen mineralization is controlled by proteins that are categorized as non-collagenous proteins (N) (Young et al., 1992). Dentin matrix proteins are a group of NCPs found in the extracellular matrix (ECM) of dentin and bone albeit in different quantities. Dentin phosphophoryn (DPP) and dentin sialoprotein (DSP) are highly specific dentin matrix proteins. The amount of DSP in bone is about 1/400 of that in dentin (Qin et al., 2002). DSP and DPP are proteins encoded by the gene dentin sialophosphoprotein (DSPP). DSPP defects cause dentinogenesis imperfecta and are not associated with bone abnormalities (Zhang et al., 2001; Kim et al., 2005). Dentin matrix protein 1 (DMP1) is one of the early genes expressed during the commitment of neural crest–derived cells into odontoblasts. It is showed that DMP1 is able to bind specifically with the DSPP promoter and activate its transcription (Narayanan et al., 2006). Previous studies have suggested that stem cell–mediated bone and dentin regeneration are regulated by distinct mechanisms, although the differences remain unclear (Batouli et al., 2003). The process by which MSCs differentiate into osteoblasts is regulated by coordinated extracellular signals from autocrine and paracrine loops (Zaidi, 2007; Kim et al., 2013). The secreted ECM provides a series of signals to cells, regulating all aspects of their phenotype from morphology to differentiation.

The MSCs derived from different tissues have distinct characteristics. In vitro, MSCs retain the capacity to regenerate unique microenvironments similar to those from which they are derived (Gronthos et al., 2000). Key lineage decisions are made during the early stages of mesenchymal differentiation. Given that the initial event in osteogenesis is the determination of MSCs to become osteoprogenitor cells, we profiled the secretomes of SCAPs and BMSCs after inducing their odonto/osteogenic differentiation in vitro and compared the expression of the secreted proteins using isobaric chemical tags (tandem mass tags, TMTs) and high-performance liquid chromatography with tandem mass spectrometry (HPLC-MS/MS).

Cells with characteristics of multipotent progenitors have been identified in stromal tissues. These progenitors are described as MSCs if they are capable of self-renewal and generation of multiple lineages (Gronthos et al., 2000; Bianco et al., 2001; Crisan et al., 2008; Sonoyama et al., 2008). When grown under mineralizing conditions, MSCs are capable of secreting a mineralized matrix that is used as an indication of their potential for “osteogenic” differentiation. Previous studies have confirmed that MSCs derived from individual tissues have distinct features (Gronthos et al., 2000; Seo et al., 2004; Huang et al., 2010). SCAPs are regarded as the source of primary odontoblasts and promote regeneration of the dentin matrix in vivo (Huang et al., 2008; Sonoyama et al., 2008; Huang et al., 2010). The evaluation of odontogenic and osteogenic differentiation of MSCs is commonly carried out by an illustration of the expression of varied related markers. When grown under mineralizing conditions in vitro, SCAPs undergo osteo/odontogenic differentiation; the expressions of osteoblastic markers and odontoblastic markers are upregulated (Ma et al., 2016; Wang et al., 2018; Zhang et al., 2019). To the best of our knowledge, there are no published systemic proteomic analyses that explore the extracellular environment of SCAPs during osteo/odontogenic differentiation.

Most of the proteins that exhibited significant changes during the osteo/odontogenic differentiation of SCAPs also changed during the differentiation of BMSCs. ECM organization proteins, calcium-binding proteins, and proteins negatively regulating cell proliferation increased significantly during differentiation of both SCAPs and MBSCs, while macromolecular complex assembly proteins and growth factors significantly decreased. In this study, BMSCs were isolated from alveolar bone marrow. Tooth and craniofacial bone arise from neural crest cells. The expressions of neural tissue–related proteins such as tubulin beta-3, midkine, and forkhead box protein G1 all decreased after the induction of differentiation; these findings confirm the differentiation of these MSCs.

As SCAPs are the source of odontoblasts and BMSCs give rise to osteoblasts, we focused our study on the genes and proteins that were differentially expressed during the differentiation; our intent was to reveal distinct regulatory mechanisms responsible for odontogenesis and osteogenesis. In the present study, GLRX5 was upregulated during the differentiation of both SCAPs and BMSCs but was secreted in higher concentrations by the SCAPs. GLRX5 is a mitochondrial protein that contributes to the generation of iron–sulfur complex and plays a multifunctional role in iron–sulfur protein synthesis and maturation (Liu et al., 2016); differential secretion of GLRX5 may be associated with varying capacities for supporting biosynthesis of iron/sulfur complexes.

THBS2, ADIRF, and CYTL1 were secreted by SCAPs albeit to a lesser extent than observed in BMSCs; these factors were downregulated during the differentiation of SCAPs and BMSCs. THBS2 is an autocrine inhibitor of MSC proliferation. THBS2-null cells demonstrate an increased rate of proliferation, whereas addition of recombinant THBS2 resulted in a dose-dependent decrease in MSC proliferation (Hankenson and Bornstein, 2002), although the rate of proliferation of SCAPs was higher than that of BMSCs (Yu et al., 2016). When MSCs undergo differentiation, their proliferation is inhibited. ADIRF (also known as APM2 or C10orf116) is also highly expressed in adipose tissue, where it promotes early-stage adipogenic differentiation (Ni et al., 2013). Likewise, CYTL1, a factor originally identified in CD34⁺ bone marrow and cord blood mononuclear cells, was implicated in promoting neoplastic cell growth and metastasis (Liu et al., 2000; Ai et al., 2016; Wang et al., 2016). Recently, it has been found that CYTL1 negatively regulates the osteogenesis of human MSCs; specific inhibition was demonstrated in response to overexpression or exogenous treatment with CYTL1, and proliferation was enhanced by CYTL1 knockdown (Shin et al., 2019).

We found that FHL1 and TGFβ2 were secreted by SCAPs and, to a lesser extent, by BMSCs; both proteins were upregulated during the differentiation of SCAPs and downregulated during the differentiation of BMSCs. It is interesting to consider the possibility that FHL1 may promote osteogenesis; siRNA knockdown of FHL1 resulted in a reduction in mineralized deposits and downregulated the expressions of RUNX2, OPN, and OCN at day 14 (Wu et al., 2015; Hemming et al., 2016). The gene encoding FHL1 contains TCF/LEF transcription factor binding sites and is a downstream target of Wnt signaling in differentiating muscle cells (Lee et al., 2012). However, others have shown that the canonical Wnt signaling pathway is a crucial regulator of osteoprogenitor proliferation and suppresses final osteoblast maturation (Boland et al., 2004; De Boer et al., 2004). In addition, FHL1 regulates calcium homeostasis (Pillar et al., 2017). We suggest that FHL1 plays a complex, multifaceted role in both odontogenesis and osteogenesis.

According to GO analysis, TGFβ2 is one of the most important proteins in the interaction network of ECM proteins. TGFβs are secreted proteins that are recognized as key regulators of stem cell renewal and differentiation (Kumar and Sun, 2005; Mishra et al., 2005; Zhou, 2011). Three mammalian isoforms (TGFβ1, β2, and β3) have been identified in bone and dentin. TGFβ1 predominates among the TGFβ isoforms that have been identified in dentin and bone (Seyedin et al., 1985; Cassidy et al., 1997). TGFβ1 is expressed in dental epithelia and in the mesenchyme at the bud and cap stages and has also been detected in developing teeth from the initiation stage through adulthood (Benedete et al., 2008; Huang and Chai, 2010; Li and Pan, 2017). TGFβ2 expression has been detected in the dental papilla and preodontoblasts but not in the dental epithelium (Benedete et al., 2008; Huang and Chai, 2010). In the tooth germ of a 6-month-old permanent incisor in a porcine species, TGFβ1 mRNA was predominantly detected in odontoblasts, whereas TGFβ2 mRNA was identified at high levels in the dental pulp tip (Niwa et al., 2018).

In mice overexpressing TGFβ1 in teeth, significant reduction in tooth mineralization, defective dentin formation, increased levels of calcium, and significantly diminished expression of DSPP have been reported (Thyagarajan et al., 2001). A previous study of TGFβ2-overexpressing mice documented increased dentin formation and bone mineral apposition rate, without any significant change in dentin microhardness and a more porous, osteoporotic phenotype in the bone (Denbesten et al., 2001). Likewise, in TGFβ receptor 2 gene–deleted mice, a substantial delay in odontoblast differentiation, decreased dentin thickness, and absent dentinal tubules in tooth germs were reported, whereas exogenous TGFβ2 induced nestin and DSPP expression in dental pulp cells (Oka et al., 2007). From these studies, we hypothesize that the roles of TGFβ1 and TGFβ2 in promoting odontogenesis differ markedly from one another.

BSP is a major ECM component of the bone. BSP is produced mainly by osteoblasts and is considered to be a marker for osteoblastic phenotypes (Mizuno et al., 2000). In contrast, DSPP is differentially expressed in the dentin tissue. In this study, we found that administration of TGFβ2 resulted in enhanced expression of the odontogenic-related markers DSPP and DMP1 in both SCAPs and BMSCs, and downregulated the expression of osteogenic-related markers BSP in SCAPs. It is speculated that TGFβ2 can promote the odontogenic differentiation of MSCs. Meanwhile, TGFβ1 enhanced the BSP expression and inhibited DSPP and DMP1 expression during early odonto/osteogenic differentiation of SCAPs but upregulated the expression of DSPP in SCAPs at the later stage. The BSP and DSPP expression levels did not differ between the OM with TGFβ1 group and OM alone group at the early time points of odonto/osteogenic differentiation of BMSCs, and TGFβ1 upregulated the expression of DSPP in BMSCs at the later stage. Hence, TGFβ1 may play a positive role in the mature stage of odontogenic differentiation of MSCs.

In human bone samples, levels of TGFβ1 protein ranged between 27 and 580 ng/g bone and for TGFβ2 between 7.2 and 35 ng/g bone (Hering et al., 2001). In contrast to TGFβ1, TGFβ2 mRNA expression was significantly enhanced in osteoarthritic bone compared to unaffected bone (Hering et al., 2001). TGFβ2 reduces bone and cartilage formation in healing fractures in rabbits (Critchlow et al., 1995). Previous research showed that the osteogenic differentiation of inflamed dental follicle stem cells (DFSCs) was suppressed, and the cells displayed low levels of TGFβ1 and high levels of TGFβ2 (Um et al., 2018). DFSCs treated with TGFβ2 inhibitors showed significant increases in alizarin red staining and ALP activity (Um et al., 2018). TGFβ1 expression was also increased after inhibition of TGFβ2 (Um et al., 2018). It is speculated that TGFβ2 may inhibit bone formation.

Although TGFβ1 and TGFβ2 are highly homologous at the level of sequence, analysis of their in vivo function by gene knockouts revealed striking differences, suggesting no significant functional redundancy between TGFβ1 and TGFβ2 (Sanford et al., 1997; Denbesten et al., 2001; Thyagarajan et al., 2001). Previous research demonstrated that TGFβ2 operates through a distinct mechanism in which the type III receptor functions as a co-receptor for efficient binding, in contrast to TGFβ1/3, which acts by direct binding to type I and II receptors (Zuniga et al., 2005). It is reported that TβRII-B, an alternatively spliced variant of the TGFβ type II receptor, is a TGFβ2 binding receptor, which mediates signaling via the Smad pathway in the absence of any TGFβ type III receptor (Rotzer et al., 2001). The expression of TβRII-B is restricted to cells originating from tissues such as bone where TGF2 has a predominant role (Rotzer et al., 2001).

In conclusion, we evaluated the secretomes of differentiating SCAPs and BMSCs to understand the molecular events underlying the early steps in dentin matrix formation and to elucidate distinct mechanisms in early odontogenesis and osteogenesis. Our findings identified a set of proteins that are differentially secreted during early odontogenesis and osteogenesis and suggest that the differentiation of odontoblasts and osteoblasts and secretion of a mineralized matrix are controlled by complex regulatory mechanisms. Among our findings, TGFβ2 enhanced the odontogenic differentiation and attenuated the osteogenic differentiation of SCAPs, while capable of enhancing odontogenic but not osteogenic differentiation of BMSCs. Taken together, our results suggest that TGFβ2 may induce pulp regeneration to promote dentin regeneration. Overall, we have identified TGFβ2 as a potentially important molecule with distinct functions promoting the regulation of MSC differentiation.

All datasets generated for this study are included in the article/Supplementary Material.

The studies involving human participants were reviewed and approved by the Ethics Committee of the Health Science Center of Peking University (Beijing, China; IRB00001052-11060 and PKUSSIRB-201734036). Written informed consent to participate in this study was provided by the participants’ legal guardian/next of kin.

SY contributed the conception and design of the study. SY and JL organized the database, performed the statistical analysis, and wrote the manuscript. All authors contributed to manuscript revision, read, and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.