Differences in the stemness characteristics and molecular markers of distinct human oral tissue neural crest‐derived multilineage cells

Abstract Objectives Although multilineage cells derived from oral tissues, especially the dental pulp, apical papilla, periodontal ligament, and oral mucosa, have neural crest‐derived stem cell (NCSC)‐like properties, the differences in the characteristics of these progenitor cell compartments remain unknown. The current study aimed to elucidate these differences. Material and methods Sphere‐forming apical papilla‐derived cells (APDCs), periodontal ligament‐derived cells (PDLDCs), and oral mucosa stroma‐derived cells (OMSDCs) from the same individuals were isolated from impacted developing teeth. All sphere‐forming cells were characterized through biological analyses of stem cells. Results All sphere‐forming cells expressed neural crest‐related markers. The expression of certain tissue‐specific markers such as CD24 and CD56 (NCAM1) differed among tissue‐derived cells. Surprisingly, the expression of only CD24 and CD56 could be discriminated in human tissues. Although APDCs and PDLDCs exhibited greater mineralized cell differentiation than OMSDCs, they exhibited poorer differentiation into adipocytes in vitro. In immunocompromised mice, APDCs formed hard tissues better than PDLDCs and OMSDCs. Conclusions Although cells with NCSC‐like properties present the same phenotype, they differ in the expression of certain markers and differentiation abilities. This study is the first to demonstrate the differences in the differentiation ability and molecular markers among multilineage human APDCs, PDLDCs, and OMSDCs obtained from the same patients, and to identify tissue‐specific markers that distinguish tissues in the developing stage of the human tooth with immature apex.


| INTRODUCTION
Oral and maxillofacial tissues are attractive cell sources for regenerative medicine. A sufficient number of stem cells can be obtained from the third molar tissue without causing aesthetic issues or collecting surplus tissue. Dental pulp stem cells (DPSCs), 1-6 periodontal ligament stem cells (PDLSCs), 7 and oral mucosa stromal stem cells (OMSCs) 8,9 are derived from the neural crest (NC). 10,11 NC-derived stem cell (NCSC)-like cells have been isolated from various mice and human tissues, including the skin, [12][13][14] heart, 15 bone marrow, 16 dental pulp, [17][18][19][20] periodontal ligament, 21,22 and oral mucosa, 23,24 using the sphere formation technique, which can enrich stem/progenitor cells. [12][13][14][15][16][17][18][19][20][21][22][23][24] NCSCs express NC-related markers and can differentiate into the cells of mesenchymal lineage, including osteoblasts, chondrocytes, adipocytes, and smooth muscle cells, as well as cells of neural lineage. 25 Although DPSCs are present in the dental pulp tissue of human teeth with completely formed roots, the apical papilla from human teeth with immature root apices is a developing tissue, and the stem cells present play a crucial role in complete root formation. Sonoyama et al. reported that stem cells from the apical papilla have higher proliferative potential, higher telomerase activity, and a greater capacity for hard/mineralized-tissue formation than stem cells from DPSCs. 4 Similarly, it is expected that the periodontal ligament from developing human teeth also contains stem cells capable of forming cementum and periodontal ligaments along with tooth roots. Therefore, we consider that developing human teeth are an attractive source of stem cells for regenerative medicine. 5,6,17,18 Another source of stem cells, the oral mucosa with a very high regenerative ability following injury, and can be obtained even from patients without teeth. 23 Studies have shown that dental pulp, apical papilla, periodontal ligaments, and oral mucosa-derived stem/ progenitor cells have the characteristics of NCSCs, with each type of tissue-derived cell considered to possess identical stem/progenitor cell properties; [17][18][19][20][21][22][23][24]26 however, these differences have not been characterized. Elucidating these cell-specific characteristics may help not only in demonstrating the importance of each cell type with respect to tissue regeneration via differentiation into suitable target lineages, but also in identifying tissue-specific markers and their developmental role in the currently unknown human tooth developmental process.
The apical papilla, periodontal ligament, and oral mucosa can be obtained concomitantly during the extraction of an impacted third molar with an immature apex. To the best of our knowledge, differences in the differentiation ability and expression of molecular markers among human apical papilla-derived cells (APDCs), periodontal ligament-derived cells (PDLDCs), and oral mucosa stroma-derived cells (OMSDCs) obtained from individuals simultaneously and from the same sites have not been reported previously. This study was performed to characterize human NCSC-like cells from tissues obtained through the extraction of impacted developing third molars and to confirm the differences in cellular characteristics and molecular markers of individual tissue-derived stem cells, which have a great clinical scope and are potentially valuable in dental regenerative medicine.

| MATERIALS AND METHODS
Detailed materials and methods are described in the supplementary material (Extended Materials and Methods).

| Patients
Human apical papilla, periodontal ligaments, and oral mucosa tissues were obtained for the extraction of impacted developing third molars.
Written informed consent was obtained from all donors.

| Immunohistochemistry
Immunohistochemistry was performed as previously described. 23 Slides were incubated at room temperature for 1 h with primary

| RNA extraction
Total RNA was extracted from the cells using the RNeasy Mini Kit (Qiagen, Hilden, Germany), according to manufacturer instructions.

| Semi-quantitative and quantitative reverse transcription PCR (RT-PCR)
Semi-quantitative RT-PCR was performed using the PrimeScript One- Step RT-PCR Kit Ver.2 (Takara, Kusatsu, Japan For neural differentiation, cells were cultured in MSC neurogenic differentiation medium (ready-to-use; PromoCell).

| Culturing of APDCs, PDLDCs, and OMSDCs
Primary cells were cultured using the outgrowth culture system. 5

| Expression of NCSC and MSC markers in APDCs, PDLDCs, and OMSDCs
Expanded APDCs, PDLDCs, and OMSDCs were positive for nestin and CD44 ( Figure 2F, G). The expression of these markers did not vary significantly among the cell types ( Figure 2H). Furthermore, these cells were positive for nearly all MSC markers (CD29, CD73, and CD90); however, they were negative for CD34 and CD45, indicating that they were not of hematopoietic stem/progenitor cell origin ( Figure 2I).

| Characteristics of spheres derived from APDCs, PDLDCs, and OMSDCs
To enrich NCSC-like cells, APDCs, PDLDCs, and OMSDCs were cultured using the sphere technique ( Figure 3A). Sphere-forming ability did not differ significantly between cells derived from different tissues ( Figure 3B). Measurement of sphere diameter revealed the same ten-  Figure 4B, Table S1). The quantitative RT-PCR results confirmed the microarray analysis results ( Figure 4C). Only SUSD2 expression was upregulated in the oral mucosa of all samples, although not significantly ( Figure 4C). CD24 expression was observed in 2.93% ± 1.10% of APDCs but not in PDLDCs or OMSDCs ( Figure 4D). CD56 (NCAM1) expression was observed in 9.59% ± 2.16% of APDCs and 3.98% ± 0.69% of PDLDCs but not in OMSDCs ( Figure 4D). Surprisingly, immunohistochemical analyses of CD24 and CD56 (NCAM1) in human tissues revealed that CD24 was expressed in the apical papilla, whereas CD56 (NCAM1) was expressed in the apical papilla and periodontal ligament. In the oral mucosa, the expression of CD24 was observed only in the oral epithelium, whereas that of CD24 and CD56 (NCAM1) was not detected in the lamina propria ( Figure 4E). These results indicated that the expression patterns of CD24 and CD56 (NCAM1) varied with tissues and that they may serve as candidate markers for tissue specificity. and α-SMA (protein) expression were confirmed in all cells ( Figure 5D [ii, iii]) and found to not vary significantly between tissues. During neuronal differentiation, all tissue-derived cells could be recognized as multipolar morphological cells ( Figure 5E[i]) and differentiated into β3-tubulin-positive neural-like cells ( Figure 5E[ii]). TUBB3 (mRNA) and

| Differentiation into NC lineage cells
β3-tubulin (protein) expression were confirmed in all cells ( Figure 5E [iii, iv]) and did not vary significantly between tissues. Furthermore, to evaluate function of neural-like cells, we performed intracellular calcium (Ca 2+ ) imaging. We found that all kinds of cell cultures after

| Generation of hard tissue in vivo
The crucial factor for clinical application of stem cells from oral tissues is hard tissue regeneration. Hence, we investigated the regeneration capacity of sphere-forming APDCs, PDLDCs, and OMSDCs related to the formation of hard tissue in vivo ( Figure 6). To investigate cell adhesion to the multiporous HA scaffolds, the sphere-forming cells cultured on the scaffolds were examined by SEM 10 days after seeding ( Figure 6A). The sphere-forming cells from all tissues exhibited vigorous cell sheet-like growth in the multiporous scaffold ( Figure 6A).
Twelve weeks after implantation, APDCs, PDLDCs, and OMSDCs had formed ectopic mature or immature hard tissue ( Figure 6B). The regenerated hard tissues from APDCs were osteodentin-like, tissues from PDLDCs were thin and cementum-like, and those from OMSDCs were thin, immature, and osteoid-like ( Figure 6B). Masson's trichrome staining revealed that these regenerated hard tissues comprised collagen fibers ( Figure 6B). The generated hard/mineralized-tissue stained positively for anti-human OCN ( Figure 6B).
These data suggested that the regenerated tissue was hard tissue.
Subsequently, the ability to form hard tissue was quantified using the hard tissue area stained with H&E ( Figure 6C) and the OCNpositive area ( Figure 6D). APDCs exhibited a significantly larger area of hard tissue formation (p = 0.004, p = 0.003 respectively) and OCN expression (p = 0.007, p = 0.004 respectively) than PDLDCs and OMSDCs ( Figure 6C, D).
reported that LRRC17, a member of the LRR superfamily, acts as a negative regulator of RANKL-induced osteoclast differentiation and is highly expressed in osteoblasts. 39 This is consistent with our finding of high LRRC17 expression levels in APDCs and PDLDCs, which tended to differentiate more into mineralized cells in vitro than did OMSDCs ( Figure 5A). Although it was difficult to identify stem cellspecific markers based only on the characteristic markers identified in this study, we successfully determined the specific markers for each tissue and tissue-derived cell.
Hence, the mineralized cells differentiated from PDLDCs may require BMP-2 and might have formed hard tissue only when it was supplied as a stimulus. 41 The difference in the ability of each tissue-derived cell to form hard tissue in vivo is closely associated with biological roles. For example, APDCs have a high ability to form hard tissues, such as tooth roots. In contrast, in the periodontal ligament, the ability to form cementum remains unaltered under application of occlusal force. In addition, ectopic bone formation in the oral mucosa is extremely rare.
Previous reports have suggested that human "NCSC-like cells" are defined based on their sphere-forming capacity, expression of NCSC-related markers, and in vitro multipotential phenotype.
Recently, Chan et al. identified the skeletal stem cells from human adult bone that also have a hierarchy and have the highest selfrenewal and differentiation potential, even in cells isolated according to the conventional MSCs definition. 43,44 It is suggested that there may be common neural crest stem cells with high stem cell property in each tissue used in this study.
This study, however, showed that the differentiation potential and expression of markers differed depending on the tissue from which the cells were obtained, suggesting that these tissue-specific characteristics should be considered during application in regenerative medicine. 13,18,21,23,24,35 In the future, these distinct tissue-specific markers may play a crucial role in tooth regenerative medicine as they can clearly distinguish each tissue in the tooth organoids generated using induced pluripotent stem cells and tissue-derived stem/progenitor cells.

| CONCLUSIONS
We demonstrated that multilineage sphere-forming APDCs, PDLDCs, and OMSDCs share the same phenotypes as other stem/progenitor cells, although expression of certain tissue-specific markers and differentiating abilities vary based on tissue source. We showed for the first time that APDCs, PDLDCs, and OMSDCs obtained from the same patients and, concomitantly, the same sites, can be used individually in regenerative medicine-based therapy. These differences in the differentiation capacity between each type of tissue-derived cells should be taken into consideration when administering stem cell-based therapy in clinical settings. In addition, we identified human tissue-specific markers in the currently unknown human developing tooth with immature apex. Our study identified important tissue-specific markers that distinguish between apical papilla, periodontal ligament, and oral mucosa in developing stage of human third molar, as well as serves as a basis for future regenerative medicine research.

AUTHOR CONTRIBUTIONS
SA contributed to study conception and design, methodology, investigation, data analysis and interpretation, and manuscript writing. AK, KK, and KN contributed to study conception and design, methodology, investigation, and data analysis and interpretation. NY and YK contributed to study conception and design, methodology, and investigation. MM, TM, CH, HK, and TY contributed to study conception and design and methodology. All authors read and approved the final manuscript.