Development of a high-efficiency induction system for embryonic oral epithelium from human pluripotent stem cells

Summary

Embryonic oral epithelium is a multipotent, ectodermal tissue that gives rise to various organs, including salivary glands, tooth germs, taste buds, and anterior pituitary glands. Although oral epithelium is contiguous posteriorly with endodermal epithelium and anteriorly with surface ectoderm, it exhibits distinct gene expression profiles during development. However, the molecular mechanisms that govern fate specification between oral epithelium and surface ectoderm remain poorly understood. Here, we present a highly efficient protocol for the PITX2-positive oral epithelium induction from human induced pluripotent stem cells (hiPSCs) using a reporter system. Sonic Hedgehog (SHH) signaling activation is essential for the efficient PITX2-positive epithelial cell generation. The induced cells exhibited gene expression profiles resembling those of embryonic oral epithelium and formed epithelial spheres that replicated features of human oral epithelium. This study established a robust platform for investigating human oral epithelial development and provided a valuable foundation for organoid-based research on oral organs.

Graphical abstract

Highlights

• •Efficient induction of embryonic oral epithelium from human pluripotent stem cells
• •Sonic Hedgehog signaling activation enhances oral epithelial differentiation
• •Induced PITX2⁺ cells recapitulate embryonic oral epithelial gene signatures
• •Oral epithelial spheres provide a platform for oral organoid development

Nakashima et al. establish a highly efficient protocol to generate PITX2⁺ embryonic oral epithelium from human pluripotent stem cells. Activation of Sonic Hedgehog signaling enhances induction efficiency, and the resulting cells reproduce embryonic oral epithelial gene profiles. This system provides a robust platform for studying human oral development and modeling craniofacial disorders.

Introduction

During embryonic development, the posterior region of the oral epithelium is derived from the endodermal epithelium, while the anterior region originates from the ectodermal epithelium, with the boundary located at the base of the tongue (Ono-Minagi et al., 2023; Rothova et al., 2012). The ectoderm-derived oral epithelium is a multipotent tissue that gives rise to various structures, including tooth germs, taste buds, salivary glands, and the anterior pituitary gland. In mice, single-cell RNA sequencing (scRNA-seq) analyses have demonstrated that embryonic oral epithelium exhibits transcriptional profiles distinct from those of the surface ectoderm, with region-specific gene expression emerging prior to organogenesis (Ye et al., 2022). In vitro differentiation systems that enable functional analyses are essential for understanding the spatiotemporal formation of oral structures. Moreover, oral anatomy varies substantially among species owing to evolutionary differences in feeding mechanisms. Thus, establishing oral epithelial induction models using human pluripotent stem cells (PSCs) is critical for elucidating human-specific oral development.

We previously reported the successful differentiation of human induced PSCs (hiPSCs) into salivary gland organoids (Tanaka et al., 2022). Other groups have developed protocols to generate oral organoids, such as ameloblast organoids and anterior pituitary organoids (Alghadeer et al., 2023; Ozone et al., 2016). All of these strategies involve recapitulating embryonic development via the oral epithelium. However, oral epithelial induction remains inefficient, and the formation of target tissues frequently occurs only in parts of the organoids or requires additional purification steps. Therefore, improving the efficiency of oral epithelial cell induction is a common challenge in oral organoid research.

During early development, the oral epithelium originates from the surface ectoderm, and it is well established that BMP signaling is required for differentiation toward surface ectoderm (Britton et al., 2019; Tchieu et al., 2017; Teague et al., 2024; Xie et al., 2020). However, the mechanism by which oral epithelium diverges from the surface ectoderm, which typically gives rise to the epidermis, remains unclear. The oral epithelium plays a central role in craniofacial morphogenesis. As cranial neural crest cells migrate into the mandibular arch, they are exposed to epithelial signals such as FGF8 and BMP4 secreted from the oral epithelium (Neubüser et al., 1997). These cues are essential for neural crest cell survival and proliferation, ultimately contributing to mandibular bone and dental mesenchyme formation (Creuzet et al., 2004; Trumpp et al., 1999). Therefore, embryonic oral epithelium not only orchestrates organogenesis but also underlies craniofacial development. Dissecting the mechanisms driving oral epithelial differentiation may provide insights into the etiology of congenital craniofacial anomalies.

In this study, we established a highly efficient protocol for embryonic oral epithelium induction from hiPSCs, thereby providing a robust platform for investigating oral epithelial development and organoid-based modeling of oral structures.

Results

Generation of a PITX2-positive oral epithelial cell induction monitoring system from human PSCs

To investigate differentiation signals toward embryonic oral epithelium, we established PITX2 reporter hiPSCs. PITX2 is known to be expressed in mesoderm-derived cells such as cardiomyocytes and blood vessels; however, among endodermal and non-neuroectodermal epithelium, its expression is specific to the embryonic oral epithelium (Kitamura et al., 1999; Tanaka et al., 2024). To generate reporter induced PSCs (iPSCs), we inserted an HA tag and EGFP downstream of the PITX2 open reading frame via CRISPR-Cas9 knockin (Figure 1A). Heterozygous knockin of neomycin-resistant clones was confirmed by genomic PCR (Figure S1A). Sanger DNA sequencing confirmed that the insertions were in-frame and that the established iPSCs exhibited typical PSC morphology without EGFP leakage signals (Figure S1B).

Figure 1.

Generation of a PITX2::EGFP reporter system for monitoring oral epithelial differentiation from hiPSCs

(A) Schematic of PITX2::EGFP knockin strategy. An HA tag and EGFP were inserted in-frame downstream of PITX2 open reading frame using CRISPR-Cas9. LHA, left homology arm; RHA, right homology arm; PGK, phosphoglycerate kinase promoter; NeoR, neomycin resistance gene.

(B) Protocol for oral epithelial induction. hiPSCs were seeded in StemFit medium on laminin-511-coated plates. From day 2, cells were cultured in oral epithelial induction medium.

(C) Bright-field and fluorescence images of PITX2::EGFP reporter cells cultured in CnT medium. EGFP signals were barely detectable on day 5 but were observed by day 10. Scale bars, 100 μm.

(D) Flow cytometric analysis of PITX2::EGFP expression before (day 0) and after (day 10) differentiation in CnT medium.

(E) RT-qPCR analysis in induced pluripotent stem cells (iPSCs), unsorted cells (unsort), EGFP-negative (GFP_nega), and EGFP-positive (GFP_posi) populations at day 10. Data are presented as relative fold change; error bars represent means ± SD (n = 6 wells from three independent experiments). Statistical significance was determined by one-way ANOVA followed by Tukey’s post hoc test.

(F) Representative fluorescence-activated cell sorting (FACS) plots showing PITX2::EGFP and EPCAM expression in control and SAG-treated cells at day 10. SHH pathway activation via SAG enhanced the proportion of PITX2::EGFP- and EPCAM-double-positive cells.

(G) Quantification of PITX2::EGFP- and EPCAM-double-positive cell populations following treatment with various signaling molecules from day 2. SAG significantly increased the proportion of double-positive cells compared to control and treatment conditions. Data represent means ± SD (n = 8 wells [Ctrl, SAG], 7 wells [BMP4, EGF, FGF2], 10 wells [SB], 9 wells [LDN, CHIR, FGF7, FGF10], and 6 [IWP2], from three independent experiments); statistical significance was determined by one-way ANOVA followed by Tukey’s post hoc test.

Using the PITX2 reporter hiPSC line, we performed two-dimensional (2D) induction in four different epithelial culture media (Figure 1B). We selected chemically defined media without serum or serum replacement factors, such as knockout serum replacement, as batch-to-batch variability in these undefined components can compromise the differentiation reproducibility (Blauwkamp et al., 2012; Gadue et al., 2006; Zimmer et al., 2016). By day 10 of differentiation, the cells failed to survive in Epicult-C Basal Medium and MCDB153 Medium, whereas incubation in CnT-PR-EF and Keratinocyte-SFM resulted in marked cell proliferation (Figures 1C and S1C). Furthermore, in cultures with CnT medium, a subset of cells exhibited PITX2::EGFP signaling on day 10 (Figure 1C). Flow cytometry revealed a distinct GFP-positive population among the CnT-induced cells (Figure 1D). These GFP-positive cells were subsequently sorted and analyzed by quantitative reverse-transcription PCR (RT-qPCR) (Figure 1E). As expected, PITX2 expression was significantly enriched in EGFP-positive cells compared to EGFP-negative cells and was also enriched relative to unsorted populations (unsort), indicating that the reporter cells functioned as intended. Furthermore, pluripotency marker OCT4 expression decreased in the induced cells, whereas the epithelial marker EPCAM was highly expressed in EGFP-positive cells, suggesting that the induced PITX2-positive cells were ectodermal epithelial cells rather than mesodermal cells.

Oral epithelial cell differentiation from hiPSCs

To identify the factors promoting oral epithelial differentiation, we screened cytokines and small molecules for their ability to enhance PITX2::EGFP expression. From day 2 of induction onward, candidate additives were continuously added to the culture medium, and the proportion of PITX2- and EPCAM-double-positive cells was evaluated using flow cytometry on day 10. The screening results showed that most additives had little to no effect or inhibited PITX2 induction (Figures 1F, 1G, and S1D). BMP4-induced EPCAM-positive cells exhibited surface ectoderm features, whereas PITX2::EGFP expression was suppressed (Figure S1D). This finding is consistent with previous developmental studies, which reported that BMP4 represses PITX2 expression during in vivo tooth germ development (Ramanathan et al., 2018). Furthermore, reduction in EPCAM-positive cells with fibroblast growth factor (FGF) addition, compared to the control, supports previous findings that FGF signaling promotes endodermal differentiation (Lau et al., 2023). In contrast, smoothened agonist (SAG), an agonist of the Sonic Hedgehog (SHH) signaling pathway, substantially increased the induction of PITX2-positive cells compared to the control (Figures 1F and 1G). Overall, these results indicate that our 2D differentiation system utilizing PITX2 reporter hiPSCs represents a valuable model for investigating human oral epithelial development. Improved mechanistic insights into this system may facilitate the development of more efficient strategies for oral epithelium generation.

To investigate early differentiation events during the induction of oral epithelium from hiPSCs, we analyzed the initial differentiation stages. Given that BMP4 signaling is known to promote surface ectodermal fate in the ectoderm, we examined its role in our system (Figure 2A). On day 4, under the control condition, TFAP2A-positive cells, a marker of non-neural ectoderm, were observed in the center of colonies (Figure 2B). Notably, BMP4 treatment markedly increased the number of TFAP2A-positive cells, whereas BMP signaling inhibition with LDN193189 (LDN) abolished these cells (Figure 2B). In contrast, SAG addition did not affect early non-neural ectoderm induction (Figure 2B). SOX9-positive neural crest-like cells emerged only under BMP inhibition conditions and were rarely observed under other conditions, indicating that the TFAP2A-positive cells were not neural crest cells but rather surface ectodermal cells (Figure S2A). However, at this stage, only a small subset of the cells expressed K8, a surface ectoderm marker, suggesting that these cells were immature surface ectodermal cells (Figure S2A). Next, we analyzed the effects of BMP4 and SHH signaling on oral epithelial differentiation on day 10 (Figure 2C). Consistent with the screening results, SHH signaling activation by SAG enhanced the induction efficiency of PITX2::HA and E-cadherin (E-Cad) double-positive oral epithelial cells (Figure 2D). This enhancement was observed regardless of SAG addition on day 2 or day 5 (Figure 2D). These findings suggest that SHH promotes differentiation of the surface ectoderm into oral epithelium. In contrast, BMP4 supplementation on day 2 reduced the induction efficiency of PITX2::HA-positive oral epithelial cells, and this inhibitory effect was not rescued by concurrent SAG treatment or BMP4 inhibition by LDN added on day 5 (Figure 2D). These results suggested that while exogenous BMP4 promotes surface ectoderm differentiation, once the cells’ fate toward surface ectoderm is committed by BMP4, they no longer undergo differentiation into oral epithelium.

Figure 2.

Stepwise optimization of oral epithelium induction from hiPSCs via modulation of developmental signaling pathways

(A) Schematic of experimental design and epithelial lineage differentiation pathway. hiPSCs were seeded at 10,000 cells per well and induced under epithelial differentiation conditions from day 2, with or without BMP4 and the BMP inhibitor LDN. The developmental trajectory of ectodermal derivatives is shown below, including ectoderm (Ecto), neuroectoderm (NE), surface ectoderm (SE), neural crest (NC), and oral epithelium (OE), along with representative markers.

(B) Immunofluorescence of TFAP2A and E-cadherin (E-Cad) expression at day 4 under control (Ctrl), BMP4, LDN, or SAG treatment. BMP4 enhanced TFAP2A-positive surface ectoderm-like differentiation, whereas LDN suppressed it. Scale bars, 200 μm.

(C) Schematic of the induction protocol using BMP and SHH signaling. hiPSCs were seeded at 10,000 cells per well and cultured under epithelial differentiation conditions from day 2. SAG or BMP4 was added from day 2, and LDN was added on day 5.

(D) Immunofluorescence analysis of PITX2::HA and E-Cad expression at day 10 under the following conditions: control (Ctrl); SAG from day 2 (2SAG) or day 5 (5SAG); BMP from day 2 (BMP); combined BMP from day 2 and SAG from day 5 (BMP 5SAG); BMP from day 2 plus SAG and LDN from day 5 (BMP 5SAG LDN); and SAG and LDN from day 5 (5SAG LDN). PITX2::EGFP-positive cell numbers increased with SAG treatment regardless of timing, whereas BMP supplementation significantly decreased their numbers. Scale bars, 200 μm.

(E) Proportion of PITX2::EGFP-positive cells at day 10 under different initial seeding densities (5k, 10k, 25k, 50k cells per well in 24-well plate). Higher cell density enhanced differentiation efficiency. Data are shown as mean ± SD (n = 10 wells from four independent experiments). Statistical significance was determined by one-way ANOVA followed by Tukey’s post hoc test.

(F) Time-course analysis of PITX2::EGFP-positive cells under optimized SAG conditions. EGFP-positive cells appeared as early as day 4 and plateaued by day 10. Data are shown as means ± SD (n = 9 wells from three independent experiments). Statistical significance was determined by one-way ANOVA followed by Tukey’s post hoc test.

(G) Representative FACS plot of cells at day 10 showing efficient induction of PITX2::EGFP- and EPCAM-double-positive cells under SHH pathway activation.

To further improve induction efficiency, we investigated the effect of cell density during differentiation, as previous studies have reported that endogenous BMP4 expression by hiPSCs is influenced by cell density (Xie et al., 2020). Although no major differences in induction efficiency were observed at lower seeding densities, high-density culture conditions significantly increased the proportion of EPCAM- and PITX2-double-positive cells on day 10 (Figure 2E). Under these high-density culture conditions, PITX2-positive epithelial cells began to emerge by day 6 and reached a plateau by day 10 (Figure 2F). Flow cytometric analysis revealed that most PITX2-positive cells were epithelial, as indicated by their EPCAM expression (Figure 2G). These results collectively established that high-efficiency induction of the oral epithelium can be achieved by regulating endogenous BMP signaling through high-density cell seeding and by activating SHH signaling. Next, we examined whether this protocol could also be applied to other hiPSC lines. Using the same protocol, RT-qPCR was performed on day 10 of differentiation. Under SAG conditions, PITX2 expression was highest in the human iPSC lines 1383D6 and 1231A3, similar to 201B7 (Figure S2B).

Human iPSC-derived PITX2-positive epithelial cells recapitulate embryonic oral epithelium features

The optimized differentiation protocol established in the preceding experiments allowed for the efficient PITX2-positive cell generation with a sheet-like morphology within 10 days (Figures 3A and 3B). To further characterize hiPSC-derived PITX2-positive epithelial cells, we compared the gene expression profiles of PITX2-positive cells with those of PITX2-negative cells on day 10 using bulk RNA-seq. Clustering and principal-component analysis revealed that PITX2-negative and PITX2-positive cells could be segregated into two distinct clusters (Figures 3C, S3A, and S3B). As expected, PITX2 and PITX1 expressions were significantly upregulated in the PITX2-positive population (Figure S3C). In contrast, mesenchymal markers such as VIM were enriched in the PITX2-negative population (Figure S3C). Gene Ontology (GO) analysis of differentially expressed genes (DEGs) showed significant enrichment of terms related to epithelial differentiation in the PITX2-positive cluster (Figure 3D). Next, we examined the expression of individual gene characteristic of embryonic oral epithelium based on previously published studies and scRNA-seq datasets deposited for the embryonic mouse oral epithelium (Ye et al., 2022) (Figures 3E and S3D–S3F). At the individual gene level, the induced PITX2-positive cells showed high expression of PITX1, VGLL2, TP63, and SOX2, which are marker genes enriched in embryonic oral Pitx2⁺ cells, while the expression of the aboral marker gene TFAP2B was low (Figures 3E, S3E, and S3F). In addition, the keratin expression pattern was positive for KRT8 and negative for KRT5 and KRT14, similar to that observed in the E9.5 mouse oral epithelium (Figures S3D and S3E). Importantly, the hiPSC-derived oral epithelium expressed the functional growth factors FGF8, BMP4, and EDN1, which are known to be produced by the oral epithelium during embryonic development. These factors are secreted by the oral epithelium during development of the first pharyngeal arch and regulate the proliferation and osteogenic differentiation of neural crest cells that migrate into the first pharyngeal arch (Creuzet et al., 2004; Motoike et al., 2025; Sato et al., 2008; Trumpp et al., 1999; Zuniga et al., 2011). These gene expression profiles suggest that the induced oral epithelium recapitulates the characteristics of the embryonic oral epithelium and provides valuable insights for identifying human embryonic epithelial markers and transcriptional programs. Immunofluorescence analysis confirmed that the induced PITX2-positive cells were E-Cad-positive epithelial cells expressing the non-neural ectoderm markers TFAP2A and SOX2, which are characteristics of the embryonic oral epithelium (Figure 3F). Additionally, these cells were positive for Ki-67, indicating a proliferative state (Figure 3F). Taken together with the gene expression profiling data, these findings suggested that induced PITX2-positive cells exhibit a molecular signature resembling that of the embryonic oral epithelium.

Figure 3.

Characterization of hiPSC-derived PITX2-positive oral epithelial cells

(A) Schematic of the optimized induction protocol. hiPSCs were seeded at 50,000 cells per well and cultured in 2D conditions with epithelial induction medium supplemented with SAG from day 2.

(B) Bright-field and fluorescence microscopy images at day 10 showing widespread PITX2::EGFP expression and sheet-like epithelial morphology. Scale bars, 200 μm.

(C) Heatmap of differentially expressed genes comparing PITX2::EGFP-positive and -negative populations at day 10, based on RNA-seq. PITX2-positive cells exhibited a distinct transcriptional profile.

(D) Gene Ontology (GO) term enrichment analysis of genes upregulated in PITX2-positive cells.

(E) Expression levels of representative genes in PITX2-positive (blue) and -negative (red) populations based on transcripts per million (TPM) from RNA-seq. Data are presented as means ± SD (n = 3 independent experiments). Statistical analyses were performed applying a two-tailed unpaired Student’s t test assuming unequal variances. ^∗p < 0.05, ^∗∗p < 0.005.

(F) Immunofluorescence analysis of PITX2::HA-expressing epithelial cells at day 10. PITX2::HA was co-localized with epithelial markers E-Cad, TFAP2A, and SOX2. Proliferative activity was confirmed by Ki-67 staining. Nuclei were counterstained with DAPI. Scale bars, 100 μm.

Generation of oral epithelial spheres as organoid precursors

As the embryonic oral epithelium possesses multilineage differentiation potential toward salivary glands, tooth germs, taste buds, and the anterior pituitary, we investigated whether our differentiated oral epithelium could be maintained in three-dimensional (3D) culture as a platform for induction of these organoids. We investigated whether spheroids could be generated from iPSC-derived oral epithelium under monolayer culture conditions. The induced oral epithelial cells were directly applied to 3D cultures without any purification steps, such as cell sorting, and 2% Matrigel was added to the culture medium as a basement membrane component (Figure 4A). The cells did not efficiently form spheres in 96-well U-bottom plates but successfully formed uniform spheres when cultured in microwell sphere plates (Figure 4B). Analysis on day 3 post-sphere formation revealed that the spheres were composed almost entirely of E-Cad-positive epithelial cells (Figure 4C). These epithelial cells retained the expression profile characteristic of the oral epithelium, including positive expression of PITX2 and SOX2 (Figure 4C). However, similar to monolayer culture conditions, K5 and K14 expression was not observed at this stage (Figure 4C). Mesenchymal cell proliferation was not observed on day 11 post-sphere formation, and the epithelial architecture of the spheroids was preserved (Figure 4D). The spheroids were composed of PITX2- and SOX2-positive oral epithelial cells with K14-positive basal-like cells in the outer layer (Figure 4D). This observation suggests that peripheral cells may have undergone differentiation in response to basement membrane components present in the culture medium. The observed K14 expression pattern was similar to that in lingual organoids derived from tissue stem cells (Hisha et al., 2013). RT-qPCR analysis was performed to evaluate whether the floating-cultured oral epithelial spheres differentiated into salivary glands or other lineages. The results showed that the spheres maintained PITX2 expression, whereas the expression levels of salivary gland markers such as SOX9, SOX10, and AQP5 were significantly lower than those in salivary gland organoids (Figure S4). The expression of KRT8, which is highly expressed in taste buds, was also not significantly increased (Figure S4). Based on these findings and the absence of apparent histological changes, the spheres were suggested to retain the characteristics of oral epithelium, indicating that these oral epithelial spheres could serve as a fundamental platform for developing differentiation protocols toward specific organs derived from the oral epithelium.

Figure 4.

Generation of oral epithelial spheres as organoid precursors

(A) Schematic of sphere induction protocol. Oral epithelial cells derived from hiPSC induced in 2D culture for 10 days were used for sphere formation. From day 10, cells were transferred to sphere formation plates and cultured in epithelial medium supplemented with 2% Matrigel until day 21.

(B) Bright-field and fluorescence images of oral epithelial spheres formed in U-bottomed plates and micro-well plates. PITX2::EGFP-positive spheres were more uniformly formed in micro-well plates. Scale bars, 200 μm.

(C) Immunofluorescence analysis of oral epithelial spheres at day 13. Scale bars, 50 μm.

(D) Immunofluorescence analysis of oral epithelial spheres at day 21. Scale bars, 50 μm.

Discussion

This study established a highly efficient protocol for the induction of PITX2-positive oral epithelium from hiPSCs using a 2D culture system. The induced cells reproduced the gene expression profile of the embryonic oral epithelium, expressing the embryonic oral epithelial genes PITX1, PITX2, SOX2, and VGLL2, as well as the craniofacial growth factors FGF8 and EDN1. These cells were obtained within a 10-day differentiation period and were capable of forming oral epithelial spheres in 3D culture without the need for purification.

The cytokine screening results in this study reflect those of previous mouse developmental studies. BMP4 promoted the induction of TFAP2A-positive non-neural ectodermal cells but inhibited the induction of PITX2-positive oral epithelial cells. These results align with previous reports indicating that BMP4 is important for inducing differentiation into the epidermal ectoderm, while also suppressing PITX2 expression in the enamel knot and odontogenic epithelium of murine tooth embryos at late developmental stages (Britton et al., 2019; Hardcastle et al., 1998). A central finding of this study was that SHH signaling plays a critical role in promoting oral epithelial differentiation from hiPSC-derived surface ectoderm. Studies in chickens and knockout mouse models have largely demonstrated that it plays an important role in the development of various craniofacial organs, including tooth embryos, taste buds, and the anterior pituitary gland (Cobourne and Sharpe, 2005; Treier et al., 2001; Wang et al., 2010). However, these findings primarily pertain to craniofacial morphogenesis and organogenesis after formation of the oral epithelium, and involvement of SHH signaling in differentiation of the oral epithelium from the surface ectoderm has not been demonstrated. In addition, several differentiation protocols for anterior pituitary or ameloblast lineages from hiPSCs have employed SAG (Alghadeer et al., 2023; Ozone et al., 2016). These studies demonstrate that SAG supplementation enhances the efficiency of target organ induction. Together with our findings, these studies suggest that SAG may also function in these induction systems during the early induction phase of the oral epithelium. On the other hand, it remains unclear which cell population provides SHH during the in vivo development of the oral epithelium. At the developmental stage of the oral epithelium, the endodermal epithelium adjacent to the oral epithelium exhibits high SHH expression (Jacobs et al., 2012; Litingtung et al., 1998). SHH secreted from the endodermal epithelium may act on the neighboring oral epithelium to induce its differentiation, whereas the surface ectoderm located farther from the endodermal epithelium might instead be directed toward an epidermal fate. However, testing this hypothesis would require in vivo experiments using lineage-specific knockout mouse models.

Although the induced cells expressed PITX2 and major oral epithelial markers and successfully formed epithelial spheres, further studies are needed to determine whether these cells possess multipotent differentiation potential. Our previously reported salivary gland organoid induction protocol using FGF7 and FGF10, as well as pituitary organoid induction protocols developed by other groups, has provided important insights into the development of organ-specific differentiation methods (Ozone et al., 2016; Tanaka et al., 2022). Importantly, our system recapitulates the early stages of human oral epithelial development, which are otherwise difficult to study due to the limited availability of embryonic human tissues. Therefore, this platform could be used to investigate the mechanisms underlying congenital craniofacial disorders, particularly those involving epithelial dysregulation. Moreover, the ability to generate oral epithelium in a robust and reproducible manner provides a foundation for the development of oral organoid systems.

In conclusion, we have established a robust and efficient differentiation system for generating PITX2-positive oral epithelium from hiPSCs and identified SHH signaling as a critical modulator of this process. This platform opens new avenues for dissecting human oral epithelial development, modeling congenital oral disorders, and developing oral regenerative strategies.

Materials and methods

Oral epithelium induction by 2D culture

On day 0, PITX2 reporter hiPSCs were seeded into 24-well plates at a density of 50,000 cells per well in StemFit AK02N medium supplemented with 10 μM Y-27632. One day after seeding, the medium was replaced with StemFit AK02N medium without Y-27632. On day 2, the medium was replaced with epithelial base medium composed of a 1:1 mixture of CnT-Prime Epithelial Culture Medium, EGF/FGF-free (CnT-PR-EF; CELLnTEC, #CEL-1100-92), and RPMI Medium 1640 (Thermo Fisher Scientific, #11875-093), supplemented with 400 nM smoothened agonist (SAG; Selleckchem, #S6384). The medium was changed daily until day 10. This procedure represents the optimized induction protocol. For mini-screening experiments, cells were seeded at a density of 10,000 cells per well in 24-well plates. From day 2 onward, various signaling molecules were individually added daily at the time of medium change, including 10 ng/mL BMP-4 (PeproTech, #120-05), 100 nM LDN (STEMCELL Technologies, #72147), 5 μM CHIR (Reprocell, #04-0004), 500 nM IWP2 (Reprocell, #04-0034), 5 μM SB-431542 (Stemgent, #04-0010-05), 400 nM SAG (Selleckchem, #S6384), 10 ng/mL EGF (PeproTech, #AF-100-15), 5 ng/mL FGF2 (R&D Systems, #233-FB), 100 ng/mL FGF7 (R&D Systems, #251-KG), and 200 ng/mL FGF10 (Qkine, #QK003-0500). The culture medium used was the 1:1 mixture of CnT-PR-EF and RPMI.

For the evaluation of epithelial media, additional culture media tested included Keratinocyte-SFM (Thermo Fisher Scientific, #10724-011), MCDB153 Medium (Cosmo Bio, #CK015), and EpiCult-C Basal Medium (STEMCELL Technologies, #05631), in addition to CnT-PR-EF.

Sphere culture

Cells were cultured under the optimized induction protocol for 10 days. The induced cells were dissociated with TrypLE at 37°C for 10 min. Cells were resuspended in epithelial medium consisting of CnT-PR-EF with RPMI. Cells were seeded at a density of 5,000 cells per well in 96-well U-bottom low-cell-adhesion plates (Thermo Fisher Scientific) or density of 1,000 cells per micro well in EZSPHERE SP MICROPLATE 96 Well with Lid (REPROCELL; 4860-900SP) using epithelial medium containing 20 μM Y-27632 and 2% growth factor-reduced Matrigel (BD Biosciences). In the case of U-bottom conditions, the plate was centrifuged at 150 × g for 10 min to form cell aggregates and incubated at 37°C with 5% CO₂. The medium was changed every other day using a P200 pipette; 50 μL was gently and replaced with 50 μL of fresh medium, taking care not to disturb the spheres.

Quantification and statistical analysis

Statistical analyses were performed using one-way ANOVA followed by Tukey’s post hoc test for multiple comparisons. All analyses were conducted using RStudio (version 4.4.1). Immunofluorescence experiments were independently repeated at least three times with consistent results. No statistical method was used to predetermine sample size, and the experiments were not randomized.

Resource availability

Lead contact

Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Junichi Tanaka (jtanaka@dent.showa-u.ac.jp).

Materials availability

Materials used in this study are available from the lead contact, Junichi Tanaka, upon request.

Data and code availability

The RNA-seq data that support the findings of this study have been deposited in the DDBJ. The accession number for the RNA-seq data reported in this paper is DDBJ: PRJDB39929.

Acknowledgments

This research was supported by JSPS KAKENHI grant nos. 23K27780 and 25K22697 to J.T. and 22H03264 to K.M., AMED grant no. JP24bm1123060 to J.T., and JST FOREST Program, grant no. JPMJFR245A to J.T.

Author contributions

Conceptualization, J.T. and K.M.; methodology, K.N. and J.T.; investigation, K.N., J.T., E.M., Y.W., and S. Ohnuma.; writing – original draft, K.N., J.T., and K.M.; writing – review and editing, K.N., J.T., and K.M.; supervision, T.S., S. Ohba, and K.M.; project administration, J.T. and K.M.; funding acquisition, J.T. and K.M.

Declaration of interests

The authors declare no competing interests.

Published: February 26, 2026