Research Article: Human Deciduous Teeth Stem Cells (SHED) Display Neural Crest Signature Characters

Date Published: January 26, 2017

Publisher: Public Library of Science

Author(s): Karlen G. Gazarian, Luis R. Ramírez-García, Gianpaolo Papaccio.


Human dental tissues are sources of neural crest origin multipotent stem cells whose regenerative potential is a focus of extensive studies. Rational programming of clinical applications requires a more detailed knowledge of the characters inherited from neural crest. Investigation of neural crest cells generated from human pluripotent stem cells provided opportunity for their comparison with the postnatal dental cells. The purpose of this study was to investigate the role of the culture conditions in the expression by dental cells of neural crest characters. The results of the study demonstrate that specific neural crest cells requirements, serum-free, active WNT signaling and inactive SMAD 2/3, are needed for the activity of the neural crest characters in dental cells. Specifically, the decreasing concentration of fetal bovine serum (FBS) from regularly used for dental cells 10% to 2% and below, or using serum-free medium, led to emergence of a subset of epithelial-like cells expressing the two key neural crest markers, p75 and HNK-1. Further, the serum-free medium supplemented with neural crest signaling requirements (WNT inducer BIO and TGF-β inhibitor REPSOX), induced epithelial-like phenotype, upregulated the p75, Sox10 and E-Cadherin and downregulated the mesenchymal genes (SNAIL1, ZEB1, TWIST). An expansion medium containing 2% FBS allowed to obtain an epithelial/mesenchymal SHED population showing high proliferation, clonogenic, multi-lineage differentiation capacities. Future experiments will be required to determine the effects of these features on regenerative potential of this novel SHED population.

Partial Text

Stem cells from human exfoliated deciduous teeth (SHED) [1] [2] [3], derived from adult wisdom teeth pulp [4], and periodontal ligament [5] attract wide attention owing to their multipotent stemness and regenerative potentials. Studies on animal embryos documented that dorsal neuroepithelial cells, orchestrated by a gene network [6] delaminate from the border between neural and non-neural ectoderm [7] via a partial epithelial-to-mesenchymal transition (EMT) and migrate as cranial wave migratory cells [8,9] to a plethora of developing tissues [10] [11] [12] [13]]. This ecto-mesenchymal, clonogenic, and multipotent [12] neural crest population was identified by verifying the expression of Sox10, p75, HNK-1, and Ap-2 genes [14] [15] [16] and by the expression of the genes required for their migration, ZEB, SNAIL1, SLUG, FOXD3 and others. Their path could be traced during embryogenesis by specific markers of postnatal sites, mouse pulp [17] and ligament [18]. In human embryogenesis, the same markers–p75, HNK-1, Ap-2–were detectable to the stages earlier than S20 [19] [20] but could not evidently provide evidence on movements of neural crest cells to their destinations. Hence, investigation of the neural crest markers in postnatal dental cells was the approach that was used. Several groups reported on expression of the neural crest marker p75 by a subset of dental cell populations, SHED [21] [22], third molars [23] [24], dental follicle [25] and periodontal ligament [26], as the evidence of their origin from neural crest. Recent studies on in vitro generation of neural crest cells from human pluripotent stem cells, embryonic (ESC) and induced (iPSCs) [27–33], described the specific culture conditions required for the phenotypic and gene expression characters of neural crest cells can be displayed. The studies showed that neural crest culture condition is distinct from that of dental cells. Optimal for neural crest cells is serum-free medium with activated Wnt and inhibited SMADs pathways, whereas optimal for expansion of dental cells is serum-rich (contain regularly 10% foetal bovine serum, FBS) medium without these specific signaling requirements. Respectively, when epithelial-like neural crest cells were transferred from their medium to the dental cell medium, they underwent an epithelial-to-mesenchymal transition losing their attributes [28] [29]. This gives rise to a possibility that epithelial-to-mesenchymal transition is inhibitory for the neural crest identity genes and that such a transition is induced both in neural crest and in dental cells by Tgf-β present in FBS [34] [35] [36]. We describe experiments demonstrating that under the culture conditions adequate for expression of neural crest characters a proportion of SHED undergo mesenchymal-to-epithelial transition and the cells become similar to neural crest cells.

Previous convincing evidence from tracing experiments demonstrated that embryonic migratory neural crest cells are progenitors of mouse dental pulp [17] and periodontal ligament [18] cells. Human neural crest cells cannot be traced to their postnatal destination, hence searching for the signs of the neural crest identity in postnatal dental cells is a possibility that was used. Thus, several groups [22] [23] [26] described p75-marker in 4% to 10% of dental populations. The neurotrophin receptor p75(NTR) is expressed by adult neural cells playing a fundamental role in the development and maintenance of the nervous system [39]. In neural crest cells, p75 is one of the key markers so that an antibody to p75 alone purifies neural crest cells from the culture of ESC-derived neuroepithelal cells [27] and from dental pulp populations ([22,23] and this study). In vitro experiments with neural crest cells derived from human pluripotent stem cells have revealed that their identity characters, including p75, are suppressed in mesenchymal medium containing 10% FBS [28]. We suggested that dental cells, SHED in particular, might be deprived of the neural crest traits in regular medium containing a similar high concentration of serum. The results of the study proved this suggestion showing that the set of the neural crest signature traits, that were observed previously, in vivo, in human embryos [20] and, in vitro, in human pluripotent stem cell-derivative neural crest cells [28], p75, HNK-1, SOX 10, can be activated in SHED. The role of these genes in dental cells remains to be elucidated. One possibility to verify in further experiments is that the activity of neural crest genes, specifically the p75 neurotrophin receptor, is required for inhibition of mesenchymal lineages [22] to ensure the neuronal lineage commitment via activation of SOX2-MASH-1 genes [30,31]. The other finding in this study which merits a deeper investigation is the heterogeneity of SHED regarding the responsiveness to the inducers of phenotypic and gene expression changes. While a small component of SHED population express p75 and HNK-1 genes even in the presence of 10% FBS ([22,23,26] and Table 2), a significant part of the SHED population retained mesenchymal phenotype in serum-free neural crest medium (Fig 3). At low FBS concentrations (1%, 2%), some of the cells underwent transition to the epithelial state and expressed p75 and HNK-1, but upregulation of SOX10 depended on the signaling requirements. The 2% FBS concentration was a cut-off point (threshold) for inter-transition of mesenchymal and epithelial neural crest markers. Repeated transferring of the cells to serum conditions above this threshold led to mesenchymal state and inactivity of the neural crest markers; conversely, serum concentration below this level led to epithelial state and upregulation of the neural crest genes. Both neural crest cells [30] and dental cells with epithelial phenotype co-expressed the neural crest (p75 and HNK-1) and mesenchymal (CD73 and CD105) genes, which may facilitate the MET and EMT transitions. Coexpression of p75 and mesenchymal markers was also observed in the cells from third molars by Alvares et al [23]. Since exogenously activated Wnt signaling, in the context of inactive TGFβ (SMAD 2/3), was required for upregulation of the key neural crest signature gene SOX10 [28], one can suppose that a low level of endogenous Wnt signaling in our SHED upregulated the p75 and HNK-1 genes in low-serum (below 2% FBS) and serum-free medium. The observed phenotypic and gene expression instability reflects the plasticity of the neural crest derivative lineages as demonstrated on Schwann cells [40], in melanocytes [41] and dental cells converting to melanocytes [42]. Studies on regenerative potential in preclinical experiments [43] [44] [45] [46] [47], banking of “clinical grade” dental stem cells [48] and trials [49] are in progress. Future identification of the factor(s) inducing EMT and MET will permit to employ them for modulating neural crest identity genes and test their role in regenerative applications of dental cells.