PAX9 Is Involved in Periodontal Ligament Stem Cell-like Differentiation of Human-Induced Pluripotent Stem Cells by Regulating Extracellular Matrix

Periodontal ligament stem cells (PDLSCs) play central roles in periodontal ligament (PDL) tissue homeostasis, repair, and regeneration. Previously, we established a protocol to differentiate human-induced pluripotent stem cell-derived neural crest-like cells (iNCs) into PDLSC-like cells (iPDLSCs) using human PDL cell-derived extracellular matrix (ECM). However, it remained unclear what factors principally regulate the differentiation of iNCs into iPDLSCs. In this study, we aimed to identify the transcription factor regulating production of human PDL cell-derived ECM, which is responsible for the generation of iPDLSCs. We cultured iNCs on ECMs of two human PDL cell lines (HPDLC-3S and HPDLC-3U) and of human dermal fibroblasts (HDF). iNCs cultured on HPDLC-3U demonstrated higher iPDLSC-associated gene expression and mesenchymal differentiation capacity than cells cultured on HDF or HPDLC-3S. The transcription factor PAX9 was highly expressed in HPDLC-3U compared with HDF and HPDLC-3S. iNCs cultured on siPAX9-transfected HPDLC-3U displayed downregulation of iPDLSC-associated marker expression and adipocytic differentiation capacity relative to controls. Our findings suggest that PAX9 is one of the transcription factors regulating ECM production in human PDL cells, which is responsible for the differentiation of iNCs into iPDLSCs.


Introduction
Periodontal ligament (PDL) is a highly specialized fibrous connective tissue that plays important roles in anchoring the tooth to the socket bone and regulating proper tooth homeostasis, repair, and nutrition [1]. One of the primary causes of PDL destruction is an advanced inflammatory disease of tissues surrounding the tooth structure, periodontitis. Once PDL is lost by the progression of periodontitis, the socket bone becomes detached from the tooth root and can no longer provide support for the tooth. PDL destruction can ultimately result in tooth loss; therefore, researchers have tried to repair and regenerate injured PDL tissue with the aim of preserving teeth with severe periodontitis.

Cap Analysis Gene Expression and Gene Ontology Enrichment Analysis
Cap analysis of gene expression (CAGE) allows genome-wide analysis of gene transcription start sites and quantitative study of RNA transcribed from them by DNAFORM (Yokohama, Kanagawa, Japan). In brief, RNA quality was assessed with a Bioanalyzer (Agilent, Santa Clara, CA, USA) to ensure that the RNA integrity number was over 7.0, and A260/280 and 260/230 ratios were over 1.7. First-strand cDNAs were transcribed to the 5 end of capped RNAs and attached to CAGE "barcode" tags, which upon sequencing were mapped to the mouse mm9 genomes using BWA software (v.0.5.9, SourceForge Headquar-Biomedicines 2022, 10, 2366 5 of 20 ters, San Diego, CA, USA) after discarding ribosomal or non-A/C/G/T base-containing RNAs. For tag clustering, CAGE-tag 5 coordinates were input for CAGEr clustering using 20 bases as a maximal allowed distance between two neighboring tags and a minimum counts per million (CPM) value of 2 [38]. GO enrichment analysis was also performed using BWA and Gene Set Enrichment Analysis (GSEA) software (Broad Institute, Cambridge, MA, USA) [39,40].

Proliferation Assay
iNC-Unt, iNC-siCont, and iNC-siPAX9 were seeded at a density of 3 × 10 3 cells/well into wells of 48-well plates and incubated for up to 7 days. Subsequently, their proliferation was examined using a cell proliferation assay kit (Takara Bio, Shiga, Japan) on 0, 1, 2, 3, 5, and 7 days. After incubation, 25 µL of the WST-1 kit reagent was added to the culture medium of each well. Following 1 h of treatment, 100 µL of supernatant was collected from each well, and the optical density at 450 nm of each well was measured with an iMark microplate reader (Bio-Rad Laboratories, Hercules, CA, USA).

Osteoblastic Differentiation
iNC-HDF, iNC-3S, iNC-3U, iNC-Unt, iNC-siCont, and iNC-siPAX9 were seeded at a density of 2 × 10 4 cells into wells of a 24-well plate and cultured in 10% FBS/α-MEM (control medium; CM). After reaching confluence, the culture medium was changed to an osteoblastic differentiation medium (ODM) composed of CM supplemented with 50 µg/mL ascorbic acid (Nacalai Tesque) and 2 mM β-glycerophosphate (Sigma Aldrich). As a control, cells were cultured in CM. After 3 weeks of culture, cells were fixed with 10% formalin (Wako) for 1 h, washed with distilled water, and stained with 0.3% Alizarin Red S (Invitrogen), as described previously [16]. Images of stained cell were captured with a Keyence BZ-9000 microscope (Osaka, Japan). Nine fields were randomly chosen for quantification of Alizarin Red S-positive area. Measurements were performed using BZ-X Analyzer Software (Keyence, Osaka, Japan).

Adipocytic Differentiation
iNC-HDF, iNC-3S, iNC-3U, iNC-Unt, iNC-siCont, and iNC-siPAX9 were seeded at a density of 2 × 10 4 cells into wells of a 24-well plate and cultured in CM. After reaching confluence, the culture medium was changed to an adipocytic differentiation medium (ADM) composed of CM supplemented with 1% L-glutamine, 0.1 mM L-ascorbic acid (Wako), 1 mM Sodium Pyruvate Solution (100×, Nacalai Tesque), 10 µM hydroxyethyl-piperazinyl ethanesulfonic acid (Nacalai tesque), 60 mM indomethacin (Sigma-Aldrich), and 10 −7 M dexamethasone (Merck Millipore, Darmstadt, Germany). As a control, cells were cultured in CM. After 4 weeks, cells were fixed with 10% formalin, washed with distilled water, and stained with 0.3% Oil Red O (Invitrogen) for lipid detection as described previously [16]. Images of stained cell were captured with a Keyence BZ-9000 microscope. Nine fields were randomly chosen for quantification of the number of Oil Red O-positive cells.

Statistical Analysis
All values are expressed as the mean ± standard deviation. Statistical analysis was performed using one-way ANOVA, followed by the Bonferroni method for comparisons of three or more groups. p values of < 0.05 were considered statistically significant.

Periodontal Ligament Stem Cell Characteristics of iNC-HDF, iNC-3S, and iNC-3U
HDFs, HPDLC-3S cells, and HPDLC-3U cells demonstrated typical fibroblastic morphologies, namely plump spindle or stellate shapes with centrally placed round nuclei ( Figure 1A). HPDLC-3S and HPDLC-3U cells highly expressed PDL-related genes αSMA, OPG, and ALP compared with HDFs ( Figure 1B). These results indicated that HPDLC-3U, which is a heterogeneous cell population, contained more fibroblasts expressing PDLrelated genes than HDF and HPDLC-3S.  iNC-HDF mainly exhibited rounded shapes, while iNC-3S included cells with round or spindle shapes ( Figure 1C). iNC-3U was mainly comprised of spindle-shaped cells ( Figure 1C). iNC-3U and iNC-3S displayed higher expression of iPDLSC-associated genes OPG, POSTN, COL1A1, and PLAP1 compared with HDF ( Figure 1D). Additionally, expression of OPG, COL1A1, and PLAP1 was upregulated in iNC-3U compared with iNC-3S ( Figure 1D). After osteoblastic induction, iNC-HDF formed a small number of Arizarin Red S-positive mineralized deposits, while iNC-3S generated more mineralized deposits than iNC-HDF ( Figure 1E). iNC-3U formed large numbers of mineralized deposits compared with iNC-HDF and iNC-3S ( Figure 1E). Following adipocytic induction, iNC-HDF formed no Oil Red O-positive lipid droplets and iNC-3S generated only a small number of lipid droplets ( Figure 1F). In contrast, iNC-3U formed large numbers of lipid droplets compared with iNC-HDF and iNC-3S ( Figure 1F). These results indicated that ECM derived from HPDLC-3U had higher induction ability to PDLSC compared with ECM derived from HPDLC-3S and HDF.

Identification of Transcription Factor PAX9 Responsible for the Induction of iNCs to iPDLSCs
Expression levels of iPDLSC-related genes and their ability to differentiate into osteoblasts and adipocytes varied between iNC-HDF, iNC-3S, and iNC-3U. Therefore, we performed CAGE to compare gene expression among HDF, HPDLS-3S, and HPDLC-3U. A total of 13728 genes were screened out (Figure 2A), of which, 1986 were upregulated in HPDLC-3S by >2-fold relative to HDF. GO analysis indicate that the term "DNA binding transcriptional activator activity" was enriched in HPDLC-3S compared with HDF, and 15 genes were included in this term ( Figure 2B). Among them, expression of five genes (FOXF2, SIX2, DLX5, PAX9, and TFAP4) was upregulated in HPDLC-3U relative to HPDLC-3S. GSEA confirmed that these genes played various roles not only in regulation of transcription, but also compound metabolic processes, cellular biosynthetic processes, and animal organ morphogenesis ( Figure 2C). Quantitative RT-PCR analysis performed to investigate their expression in HDF, HPDLC-3S, and HPDLC-3U indicated that expression of all fives genes was significantly higher in HPDLC-3S than in HDF ( Figure 2D). Expression of these genes was also significantly higher in HPDLC-3U than in HDF; however, expression of SIX2, DLX5, and TFAP4 demonstrated no difference between HPDLC-3S and HPDLC-3U ( Figure 2D). Moreover, FOXF2 and PAX9 expression was significantly higher in HPDLC-3U than in HPDLC-3S ( Figure 2D). Further comparison of their expression levels in HPDLC-3S and HPDLC-3U revealed that the rate of increase was higher for PAX9 than FOXF2 ( Figure 2D). Because of these analyses, we focused on PAX9 in this study.

iPDLSC-Associated Marker Expression in iNC-Unt, iNC-siCont, and iNC-siPAX9
Next, we generated iNC-Unt, iNC-siCont, and iNC-siPAX9 by culturing iNCs on Unt-3U, 3U + siCont, and 3U + siPAX9 for 2 weeks, respectively ( Figure 3C). iNC-Unt, iNC-siCont, and iNC-siPAX9 exhibited spindle shapes and did not include cells exhibiting rounded shapes ( Figure 3D). Expression levels of OPG, POSTN, COL1A1, and PLAP1 were almost identical between iNC-Unt and iNC-siCont ( Figure 3E). However, expression of these genes was significantly downregulated in iNC-siPAX9 compared with iNC-Unt and iNC-siCont ( Figure 3E). Moreover, many cells were positive for OPG and POSTN in iNC-Unt and iNC-siCont, while few cells demonstrated positive reactions in iNC-siPAX9 ( Figure 3F). No staining was observed in iNC-Unt stained with control IgG ( Figure 3F). These results suggested that PAX9 was involved in the differentiation of iNC into iPDLSC.

Proliferation and Cell Surface Marker Expression in iNC-Unt, iNC-siCont, and iNC-siPAX9
iNC-Unt, iNC-siCont, and iNC-siPAX9 exhibited a time-dependent increase in proliferation ( Figure 4A). iNC-Unt exhibited higher levels of proliferation than iNC-siCont and iNC-siPAX9 on days 2, 3, 5, and 7; however, there was no statistically significant difference in proliferation between iNC-Unt, iNC-siCont, or iNC-siPAX9 from days 1-7 ( Figure 4A). Cells highly expressed the MSC-associated surface markers CD90, CD105, and CD166, and slightly expressed the hematopoietic cell-associated markers CD34 and CD45 ( Figure 4B). Moreover, expression levels of MSC-and hematopoietic-associated markers were almost identical between iNC-Unt, iNC-siCont, and iNC-siPAX9 ( Figure 4B). These data indicated that PAX9 did not affect proliferative potential or expression of surface antigen markers in differentiation into iPDLSCs.

Osteoblastic Differentiation of iNC-Unt, iNC-siCont, and iNC-siPAX9
After 3 weeks of culture in ODM, iNC-Unt, iNC-siCont, and iNC-siPAX9 formed large numbers of Alizarin Red S-positive mineralized deposits ( Figure 5A). Indeed, there was no statistically significant difference in Alizarin Red S-positive area between cultures ( Figure 5B). In contrast, no deposits were observed in iNC-Unt, iNC-siCont, and iNC-siPAX9 treated with CM ( Figure S1A). Expression of osteoblast-related markers OCN, BMP2, and BSP was upregulated in iNC-Unt, iNC-siCont, and iNC-siPAX9 cultured in DM compared with cells cultured in CM ( Figure 5C) and were not statistically different between these groups ( Figure 5C). These data indicated that PAX9 did not affect the osteoblastic differentiation of iPDLSCs.   siPAX9 treated with CM ( Figure S1A). Expression of osteoblast-related markers OCN, BMP2, and BSP was upregulated in iNC-Unt, iNC-siCont, and iNC-siPAX9 cultured in DM compared with cells cultured in CM ( Figure 5C) and were not statistically different between these groups ( Figure 5C). These data indicated that PAX9 did not affect the osteoblastic differentiation of iPDLSCs.

Adipocytic Differentiation of iNC-Unt, iNC-siCont, and iNC-siPAX9
After 4 weeks of culture in ADM, iNC-Unt and iNC-siCont formed large number of lipid droplets, while iNC-siPAX9 generated only a small number of lipid droplets ( Figure  6A). Indeed, there were statistically significant differences in numbers of cells exhibiting lipid droplets between iNC-Unt and iNC-siPAX9, and iNC-siCont and iNC-siPAX9 (Figure 6B). In contrast, no droplets were observed in cells treated with CM ( Figure S1B[I-III]). Expression of adipocyte-related markers LPL, ADIPOQ, and LEP was upregulated in iNC-Unt, iNC-siCont, and iNC-siPAX9 cultured in ADM compared with cells cultured in CM ( Figure 6C). Moreover, their expression was downregulated in iNC-siPAX9 cultured in ADM compared with iNC-Unt and iNC-siCont cultured in ADM ( Figure 6C). Expression of regulator genes for adipogenesis (CEBPα, PPARγ, and KLF15) was also suppressed in iNC-siPAX9 cultured in ADM relative to iNC-Unt and iNC-siCont cultured in ADM ( Figure 6D). These data indicated that PAX9 was involved in the regulation of adipocytic differentiation in iPDLSCs.

Adipocytic Differentiation of iNC-Unt, iNC-siCont, and iNC-siPAX9
After 4 weeks of culture in ADM, iNC-Unt and iNC-siCont formed large number of lipid droplets, while iNC-siPAX9 generated only a small number of lipid droplets ( Figure 6A). Indeed, there were statistically significant differences in numbers of cells exhibiting lipid droplets between iNC-Unt and iNC-siPAX9, and iNC-siCont and iNC-siPAX9 ( Figure 6B). In contrast, no droplets were observed in cells treated with CM ( Figure S1B(I-III)). Expression of adipocyte-related markers LPL, ADIPOQ, and LEP was upregulated in iNC-Unt, iNC-siCont, and iNC-siPAX9 cultured in ADM compared with cells cultured in CM ( Figure 6C). Moreover, their expression was downregulated in iNC-siPAX9 cultured in ADM compared with iNC-Unt and iNC-siCont cultured in ADM ( Figure 6C). Expression of regulator genes for adipogenesis (CEBPα, PPARγ, and KLF15) was also suppressed in iNC-siPAX9 cultured in ADM relative to iNC-Unt and iNC-siCont cultured in ADM ( Figure 6D). These data indicated that PAX9 was involved in the regulation of adipocytic differentiation in iPDLSCs. and LEP (C) and adipogenesis regulator genes CEBPα, PPARγ, and KLF15 (D) in iNC-Unt, iNC-siCont, and iNC-siPAX9 after 4 weeks of culture in CM or ADM. β-act was used as an internal standard. Data are portrayed as the mean ± standard deviation (n = 4). * p < 0.05 ** p < 0.01. ADM, adipocytic differentiation medium; CM, conditioned medium; iNC, induced pluripotent stem cell-derived neural crest-like cells.

Discussion
iNCs require culture on ECM derived from human PDL cells to differentiate into iPDLSCs [16]. It has been reported that the ECM is involved in determining the fate of stem cells such as epidermal stem cells, neural stem cells, and bone marrow stem cells [42][43][44]. Therefore, human PDL cell-derived ECM plays important roles in the generation of iPDLSCs. The morphology, iPDLSC-associated gene expression, and mesenchymal lineage cell differentiation of iNCs cultured on ECM derived from HPDLC-3S varied from cells cultured on ECM derived from HPDLC-3U, although both HPDLC-3S and HPDLC-3U were human PDL cells. HPDLCs were isolated from different individuals (HPDLC-3S: 23-year-old male, HPDLC-3U: 25-year-old female) and some factors, such as sex and age, can reportedly affect the composition of ECM [23]. Recent studies have demonstrated that mechanical stress regulates the ECM expression [45,46]. PDL tissue is located around tooth roots, therefore it continues to receive the occlusal force. Additionally, there are individual differences in the occlusal force [47]. Therefore, it was suggested that there are individual differences in the ECM components in the periodontal ligament tissues and LEP (C) and adipogenesis regulator genes CEBPα, PPARγ, and KLF15 (D) in iNC-Unt, iNC-siCont, and iNC-siPAX9 after 4 weeks of culture in CM or ADM. β-act was used as an internal standard. Data are portrayed as the mean ± standard deviation (n = 4). * p < 0.05 ** p < 0.01. ADM, adipocytic differentiation medium; CM, conditioned medium; iNC, induced pluripotent stem cell-derived neural crest-like cells.

Discussion
iNCs require culture on ECM derived from human PDL cells to differentiate into iPDLSCs [16]. It has been reported that the ECM is involved in determining the fate of stem cells such as epidermal stem cells, neural stem cells, and bone marrow stem cells [42][43][44]. Therefore, human PDL cell-derived ECM plays important roles in the generation of iPDLSCs. The morphology, iPDLSC-associated gene expression, and mesenchymal lineage cell differentiation of iNCs cultured on ECM derived from HPDLC-3S varied from cells cultured on ECM derived from HPDLC-3U, although both HPDLC-3S and HPDLC-3U were human PDL cells. HPDLCs were isolated from different individuals (HPDLC-3S: 23-year-old male, HPDLC-3U: 25-year-old female) and some factors, such as sex and age, can reportedly affect the composition of ECM [23]. Recent studies have demonstrated that mechanical stress regulates the ECM expression [45,46]. PDL tissue is located around tooth roots, therefore it continues to receive the occlusal force. Additionally, there are individual differences in the occlusal force [47]. Therefore, it was suggested that there are individual differences in the ECM components in the periodontal ligament tissues exposed to mechanical loading. CAGE results also demonstrated that gene expression patterns of HPDLC-3S and HPDLC-3U were different to some extent; specifically, 693 genes were upregulated and 675 genes were downregulated in HPDLC-3U by >2-fold relative to HPDLC-3S. These results indicate that differences in the composition of ECM between HPDLC-3S and HPDLC-3U likely cause of the observed difference in their ability to induce iPDLSCs from iNCs.
Our previous study demonstrated that iNCs cultured with the major components of ECM in human PDL cells, COL1 and POSTN, displayed lower expression of iPDLSCassociated genes relative to iPDLSCs. Moreover, a previous report revealed that artificially isolated single ECM components were insufficient to mimic the complex structure and complete function of natural ECM [48]. Therefore, we aimed to identify the transcription factor regulating the production of ECM in human PDL cells, which is responsible for the induction of iNCs to iPDLSCs. We focused on the transcription factor PAX9 based on results from CAGE and GSEA technologies. PAX9 belongs to the paired box family that encodes a group of growth-and development-regulation-related transcription factors [49]. GSEA results also indicated the involvement of PAX9 in head morphogenesis, body morphogenesis, face development, tissue development, and regulation of animal organ morphogenesis. PAX9 expression was previously confirmed in various tissues such as thymus, parathyroid, tonsil, vagina, and cervix [50,51]. Its expression was also identified in oral tissues including salivary glands, taste papilla of the tongue, lip, and developing palatal shelves [52]. Moreover, Pax9 is widely expressed in dental mesenchyme of developing tooth germ, whereby defects in Pax9 were associated with a lack of tooth buds and hypodontia [53]. Dental mesenchyme cells, which originate from neural crest cells, produce PDLSCs that ultimately form PDL tissue [3]; therefore, PAX9 is suggested to be a crucial factor regulating the differentiation of neural crest cells into PDLSCs. PAX9 reportedly regulates some ECM genes; for example, exogenous PAX9 increases expression of dentin matrix protein 1 in human iNCs [54], while endogenous Pax9 positively regulates Col2a1 and Acan in mouse prechondrogenic mesenchymal cells of the intervertebral discs [55]. Additionally, several studies demonstrated the effects of ECM on multipotency of stem cells. Kearns et al. suggested that the ECM environment of the central nervous system is crucial for the maintenance of multipotency of neural stem cells [56], while Antoon et al. revealed that bladder-derived ECM was of great importance to maintain the multipotency in MSCs [57]. Interestingly, tendon stem/progenitor cells contacting ECM derived from the tendon of Bgn −/0 Fmod −/− mice exhibited lower potential to differentiate into tenocytes compared with cells contacting ECM derived from wild-type mice [58]. These results strongly support our finding that ECM derived from 3U-siPAX9 impaired expression of iPDLSC-associated markers and the differentiation of adipocytes from iNCs. However, cell morphology, proliferation, surface marker expression, and osteoblastic differentiation did not vary between iNC-Unt3U, iNC-siCont, and iNC-siPAX9, suggesting that PAX9 partially regulates ECM production in human PDL cells, although other transcription factors are likely involved. As with PAX9, FOXF2 expression was significantly higher in HPDLC-3U compared with HDF and HPDLC-3S. GENE MANIA analysis demonstrated co-expression of PAX9 with FOXF2 ( Figure S2A). FOXF2 is a specific mesenchymal transcription factor expressed in mesenchymal cells adjacent to the epithelium [59]. Ormestad et al. demonstrated that Foxf2 promotes ECM synthesis in fibroblasts to support mouse gut development [60]. Foxf2 −/− mice develop a cleft palate because of defects in ECM synthesis of the secondary palate [61]; intriguingly, this phenotype is consistent with that of PAX9-deficient mice exhibiting a cleft secondary palate at birth [52]. Collectively, these results indicate that FOXF2 may also be involved in the differentiation of iNCs into iPDLSCs via regulation of ECM synthesis in human PDL cells. However, the relationship between tooth development and FOXF2 was not well investigated; only one report demonstrated that the mutation of FOXF2 was associated with the loss of early tooth markers [62]. Conversely, PAX9 was identified as a marker for prospective tooth mesenchyme in 1997 [63]. Following this report, various studies portrayed the involvement of PAX9 in tooth development. Based on these results, we considered that PAX9 would be more involved in tooth development than FOXF2 and we focused on PAX9 in this study. Additionally, GENE MANIA analysis indicated physical interactions, co-expression, and co-localization of PAX9 with transcription factor genes such as MSX1, PAX2, PAX3, PAX4, PAX6, and PAX7 ( Figure S2B). Among them, only the interaction between PAX9 and MSX1 has been clarified. Vieira et al. tested for possible PAX9-MSX1 interactions by observing the transmission of marker alleles from parents, suggesting that PAX9 interacts with MSX to cause tooth agenesis in humans [64]. Ogawa et al. further demonstrated that their interaction regulated Bmp4 expression to determine the fate of the transition from bud stage to cap stage during tooth development [65]. CAGE results revealed slightly increased MSX1 expression in HPDLC-3U relative to HDF (1.72-fold) and HPDLC-3S (1.12-fold). Therefore, further studies are essential to identify the interaction between PAX9 and MSX1, and their involvement in regulation of ECM synthesis in PDL cells.
In summary, we compared the ability of ECM derived from HDF, HPDLC-3S, and HPDLC-3U to induce differentiation of iNCs into iPDLSCs. iNCs cultured on ECM derived from HPDLC-3U displayed PDLSC-like phenotypes ( Figure 7) and strong expression of PAX9 compared with HDF and HPDLC-3S. siPAX9 transfection successfully downregulated PAX9 at gene levels in HPDLC-3U. iNCs cultured on ECM derived from siPAX9-transfected HPDLC-3U exhibited decreased expression of iPDLSC-associated markers and adipocyte differentiation ability (Figure 7). We report the involvement of PAX9 in regulation of ECM production in human PDL cells, which plays important roles in the differentiation of iNCs into iPDLSCs. PAX9 may be an effective marker for selecting human PDL cells that produce ECM responsible for the generation of iPDLSCs. Moreover, human PDL cells highly expressing PAX9 may contribute to the progress of regenerative medicine for PDL tissues via efficient induction of iNCs to iPDLSCs by their ECM. tooth mesenchyme in 1997 [63]. Following this report, various studies portrayed the involvement of PAX9 in tooth development. Based on these results, we considered that PAX9 would be more involved in tooth development than FOXF2 and we focused on PAX9 in this study. Additionally, GENE MANIA analysis indicated physical interactions, co-expression, and co-localization of PAX9 with transcription factor genes such as MSX1, PAX2, PAX3, PAX4, PAX6, and PAX7 ( Figure S2B). Among them, only the interaction between PAX9 and MSX1 has been clarified. Vieira [65]. CAGE results revealed slightly increased MSX1 expression in HPDLC-3U relative to HDF (1.72-fold) and HPDLC-3S (1.12-fold). Therefore, further studies are essential to identify the interaction between PAX9 and MSX1, and their involvement in regulation of ECM synthesis in PDL cells.
In summary, we compared the ability of ECM derived from HDF, HPDLC-3S, and HPDLC-3U to induce differentiation of iNCs into iPDLSCs. iNCs cultured on ECM derived from HPDLC-3U displayed PDLSC-like phenotypes ( Figure 7) and strong expression of PAX9 compared with HDF and HPDLC-3S. siPAX9 transfection successfully downregulated PAX9 at gene levels in HPDLC-3U. iNCs cultured on ECM derived from siPAX9-transfected HPDLC-3U exhibited decreased expression of iPDLSC-associated markers and adipocyte differentiation ability (Figure 7). We report the involvement of PAX9 in regulation of ECM production in human PDL cells, which plays important roles in the differentiation of iNCs into iPDLSCs. PAX9 may be an effective marker for selecting human PDL cells that produce ECM responsible for the generation of iPDLSCs. Moreover, human PDL cells highly expressing PAX9 may contribute to the progress of regenerative medicine for PDL tissues via efficient induction of iNCs to iPDLSCs by their ECM. Figure 7. A schematic of the effects of PAX9 downregulation in HPDLC-3U HPDLC-3U cells highly express PAX9, while iNCs cultured on ECM derived from HPDLC-3U differentiate into iPDLSCs. siPAX9-transfected HPDLC-3U downregulates PAX9 gene expression. iNCs cultured on ECM derived from siPAX9-transfected HPDLC-3U decreased the expression of iPDLSC-associated markers and the ability to differentiate into adipocytes. ECM, extracellular matrix; iNC, induced pluripotent stem cell-derived neural crest-like cells.

Conclusion
We report the involvement of PAX9 in regulation of ECM production in human PDL cells, which plays important roles in the differentiation of iNCs into iPDLSCs. PAX9 may be an effective marker for selecting human PDL cells that produce ECM responsible for Figure 7. A schematic of the effects of PAX9 downregulation in HPDLC-3U HPDLC-3U cells highly express PAX9, while iNCs cultured on ECM derived from HPDLC-3U differentiate into iPDLSCs. siPAX9-transfected HPDLC-3U downregulates PAX9 gene expression. iNCs cultured on ECM derived from siPAX9-transfected HPDLC-3U decreased the expression of iPDLSC-associated markers and the ability to differentiate into adipocytes. ECM, extracellular matrix; iNC, induced pluripotent stem cell-derived neural crest-like cells.

Conclusions
We report the involvement of PAX9 in regulation of ECM production in human PDL cells, which plays important roles in the differentiation of iNCs into iPDLSCs. PAX9 may be an effective marker for selecting human PDL cells that produce ECM responsible for the generation of iPDLSCs. Moreover, human PDL cells highly expressing PAX9 may contribute to the progress of regenerative medicine for PDL tissues via efficient induc-tion of iNCs to iPDLSCs by their ECM.

Supplementary Materials:
The following supporting information can be downloaded at: https://www. mdpi.com/article/10.3390/biomedicines10102366/s1, Figure S1: Culturing iNC-Unt, iNC-siCont, and iNC-siPAX9 in control medium; Figure   Data Availability Statement: CASE analysis data generated and/or analyzed during the current study are available in the Gene Expression Omnibus (GEO) repository, Accession Number GSE208250. The other datasets either generated and analyzed or just analyzed in this study are available from the corresponding author upon reasonable request.