Derivation of Mesenchymal Stromal Cells from Pluripotent Stem Cells through a Neural Crest Lineage using Small Molecule Compounds with Defined Media

Neural crest cells (NCCs) are an embryonic migratory cell population with the ability to differentiate into a wide variety of cell types that contribute to the craniofacial skeleton, cornea, peripheral nervous system, and skin pigmentation. This ability suggests the promising role of NCCs as a source for cell-based therapy. Although several methods have been used to induce human NCCs (hNCCs) from human pluripotent stem cells (hPSCs), such as embryonic stem cells (ESCs) and induced pluripotent stem cells (iPSCs), further modifications are required to improve the robustness, efficacy, and simplicity of these methods. Chemically defined medium (CDM) was used as the basal medium in the induction and maintenance steps. By optimizing the culture conditions, the combination of the GSK3β inhibitor and TGFβ inhibitor with a minimum growth factor (insulin) very efficiently induced hNCCs (70–80%) from hPSCs. The induced hNCCs expressed cranial NCC-related genes and stably proliferated in CDM supplemented with EGF and FGF2 up to at least 10 passages without changes being observed in the major gene expression profiles. Differentiation properties were confirmed for peripheral neurons, glia, melanocytes, and corneal endothelial cells. In addition, cells with differentiation characteristics similar to multipotent mesenchymal stromal cells (MSCs) were induced from hNCCs using CDM specific for human MSCs. Our simple and robust induction protocol using small molecule compounds with defined media enabled the generation of hNCCs as an intermediate material producing terminally differentiated cells for cell-based innovative medicine.


Introduction
In order to apply human pluripotent stem cells (hPSCs) to innovative medicine, such as cell therapy, disease modeling, and drug discovery, robust and efficient methods to produce the desired cell types without contaminating undesired cells are indispensable [1]. Since the contamination of hPSCs, in particular, may cause serious adverse effects, careful monitoring, which requires a considerable amount of time and cost, has to be conducted. Therefore, it would be beneficial to have intermediate cells between hPSCs and terminally differentiated cells, which are proved to have no contaminated hPSCs, contain limited but multiple differentiation properties, and stably proliferate without phenotypic changes. One of the promising candidates with such features is the neural crest cell (NCC) [2].
The neural crest emerges at the border of the neural and non-neural ectoderm in gastrula embryos during vertebrate development [3]. Cells in the neural crest, and later in the dorsal part of the neural tube, eventually delaminate and migrate throughout the body while retaining their characteristic phenotype [4]. When they reach their target tissues, NCCs differentiate into specific cell types depending on the location [5]. NCCs give rise to the majority of cranial bone, cartilage, smooth muscle, and pigmented cells in the cranial region, as well as neurons and glia in the peripheral nervous system [3][4][5]. Cardiac NCCs are known to contribute to valves in the heart, while vagal NCCs differentiate into enteric ganglia in the gut [6]. NCCs give rise to neurons and glia in the peripheral nervous system in the trunk region, secretory cells in the endocrine system, and pigmented cells in the skin.
Using a lineage-tracing system, rodent neural crest-derived cells were detected in adult tissues such as bone marrow, and still retained multipotent differentiation properties, which indicated that these cells are one of the cell-of-origin of multipotent mesenchymal stromal cells (MSCs) [7,8]. Therefore, the production of human MSCs (hMSCs) from hPSCs via NCC lineage is a promising approach for the use of hPSCs in innovative medicine [9,10]. A considerable number of studies have been dedicated to establishing robust and efficient induction methods from hPSCs to hNCCs in the past decade [11][12][13]. However, most of these studies used non-human stromal feeder cells or only achieved low induction efficiencies. An ideal method from the standpoint of clinical applications is free from xeno-materials, such as feeder cells or serum, and can be performed using a chemically defined medium (CDM). Two groups have published protocols that are compatible with these requirements [14,15]. The first group employed a twostep approach, in which hPSCs were firstly dissociated into single cells and cultured with CDM for two weeks for the adaptation. Cells were then cultured with CDM that was supplemented with an activator of Wnt signaling and inhibitor of Activin/Nodal/TGFb signaling, but was free from BMP signaling modulation [14]. The other group used MEF-conditioned hESC media for the initial step, and replaced it with knockout serum replacement (KSR)-based medium, which thereafter was gradually replaced with an increasing amount of N2 media. They employed an inhibitor for BMP signaling in addition to an inhibitor for Activin/Nodal/TGFb signaling during the initial 3 days, and then replaced them with an activator of Wnt signaling [15]. Therefore, the requirement for signal modulators, particularly BMP signaling inhibitors, remains controversial. Further modifications are still needed to improve the robustness, efficacy, and simplicity of these methods.
We here developed a robust and efficient induction protocol using CDM containing inhibitors for TGFb signaling and GSK3b, but not for BMP signaling with minimal growth factors. The protocol very efficiently induced hNCCs (70-80%) from hPSCs irrespective of the type (hESCs vs hiPSCs) or generating method (viral-integrated vs plasmid-episomal). Genome-wide analyses revealed that induced hNCCs retained their gene expression profile as NCCs even after 10 passages. As for differentiation properties, induced hNCCs successfully differentiated into peripheral neurons, glia, melanocytes and corneal endothelial cells. In addition, induced hNCCs were able to produce cells comparable to hMSCs, which were free from contaminated hPSCs and could differentiate into osteo-, chondro-, and adipogenic cells. Furthermore, using iPSCs generated and maintained under feeder-free and xeno-free culture systems, we successfully induced hNCCs, hMSCs, and osteogenic cells using chemically defined media.

Ethics statement
The experimental protocols dealing human subjects were approved by the Ethics Committee of the Department of Medicine and Graduate School of Medicine, Kyoto University. Written informed consent was provided by each donor.

Fluorescence-Activated Cell Sorting (FACS)
FACS was performed by AriaII (BD) according to the manufacturer's protocol. The antibodies used in FACS were listed in Table S1. In all experiments, FACS histograms of isotype controls were similar to those without antibodies; therefore, histograms without antibodies were used as control populations.

Immunocyto-and immunohistochemistry
Prior to performing immunostaining with antibodies, cells on plates were fixed with 4% paraformaldehyde at 4˚C for 15 minutes, washed two times with PBS, and incubated with 0.3% TritonX100 at 4˚C for 30 minutes as the surface-active agent for penetration processing, and any nonspecific binding was blocked with 2% skim milk/PBS at 4˚C for 1 hour. Cornea samples obtained from rabbits euthanized three days after the injection of cells were fixed with 4% paraformaldehyde and incubated in 1% bovine serum albumin (BSA) (Sigma) to block other bindings. DAPI (1:5000; Sigma) was used to counterstain nuclei. The primary antibodies used in this study were summarized in Table S1. The observation and assessment of samples were performed with BZ-9000E (Keyence, Osaka, Japan).

RT-PCR and qPCR
Total RNA was purified with the RNeasy Mini kit (Qiagen, Valencia, CA, USA) and treated with the DNase-one kit (Qiagen) to remove genomic DNA. One microgram of total RNA was reverse transcribed for single-stranded cDNA using a random primer and Superscript III reverse transcriptase (Invitrogen), according to the manufacturer's instructions. PCR was performed with ExTaq (Takara, Shiga, Japan). Quantitative PCR with the Thunderbird SYBR qPCR Mix (TOYOBO, Osaka, Japan) was performed using the StepOne real-time PCR system (Applied Biosystems, Forester City, CA, USA) in duplicate or triplicate. Primer sequences were listed in Table S2.

cDNA microarray
Total RNA was prepared using the RNeasy Mini Kit (Qiagen). cDNA was synthesized using the GeneChip WT (Whole Transcript) Sense Target Labeling and Control Reagents kit as described by the manufacturer (Affymetrix, Santa Clara, CA, USA). Hybridization to the GeneChip Human Gene 1.0 ST expression arrays, washing, and scanning were performed according to the manufacturer's protocol (Affymetrix). Expression values were calculated using the RMA summarization method and the data obtained were analyzed by GeneSpring GX 11.5.6 (Agilent Technologies, Santa Clara, CA, USA) for correlation coefficients, scatter plots, a volcano plot, heat maps, and hierarchical clustering (Distance metrics: Pearson's Centered, Linkage rule: Average). Differentially expressed genes were identified by statistical analyses and fold changes. Statistical analyses were performed using a one-way ANOVA with a Benjamini and Hochberg False Discovery Rate (BH-FDR 50.01) multiple testing correction followed by Tukey HSD post hoc tests (GeneSpring GX). Microarray data have been submitted to the Gene Expression Omnibus (GEO) public database at NCBI, and the accession number is GSE 60313. Data for hBM90, 91, and 94 have already been described [22]. Data from GSE44727 and GSE45223 were used for a comparison analysis in Figure S3.

Melanocyte differentiation
Cells were plated onto fibronectin-coated dishes in CDM supplemented with 10 mM SB and 1 mM CHIR. Melanocyte induction was performed the next day with CDM supplemented with 1 mM CHIR, 25 ng/ml BMP4, and 100 nM endothelin-3 (American Peptide Company, Sunnyvale, CA, USA) [15,27,28]. The medium was changed every other day. Differentiation was confirmed by induction of the MITF and c-KIT genes on day 7.

Corneal endothelial cell differentiation
Cells were induced to corneal endothelial cells with corneal endothelial cellconditioned CDM. Conditioned CDM was derived by collecting medium from cultured human corneal endothelial cells [29]. The selective ROCK inhibitor Y-27632 (WAKO) was used on the first day of the induction. The medium was changed every two days, and cells were analyzed by immunocytochemistry after twelve days. RT-qPCR was performed 3, 5, and 8 days after the induction.

Induction of hMSCs from hNCCs
Cells were plated onto tissue culture dishes (BD) at a density of 6.5610 4 cell/cm 2 in CDM supplemented with 10 mM SB and 1 mM CHIR. The medium was replaced the next day with aMEM (Nacalai Tesque, Tokyo, Japan) supplemented with 10% fetal bovine serum (FBS) (Nichirei Inc., Tokyo, Japan) [14,26]. The morphology of cells started to change approximately 4 days after the induction. Passages were performed every week using 0.25% trypsin-EDTA (GIBCO) at a density of 1610 4 cells/cm 2 . hMSC markers (CD73, CD44, CD45 and CD105) were analyzed by FACS 14 days after the hMSC induction. We used STK2 (DS Pharma Biomedical, Osaka, Japan) as the MSC medium and tissue culture dishes coated with fibronectin for the hMSC induction under chemically defined media conditions.

Differentiation of hNCC-derived hMSCs
Osteogenic differentiation A total of 2.5610 5 induced hMSCs/well were seeded on 6-well dishes (BD) and cultured in osteogenic induction medium, aMEM, 10% FBS, 0.1 mM dexamethasone, 50 mg/ml ascorbic acid, and 10 mM b-glycerophosphate for 2 weeks for osteogenic differentiation [24]. STK3 (DS Pharma) was used as the osteogenic induction medium instead of the osteogenic induction medium to achieve osteogenic differentiation under chemically defined media conditions. The culture medium was changed every other day for 1 week. Differentiation properties were confirmed by the formation of calcified nodules, as detected with Alizarin Red staining. Briefly, culture wells were washed twice in phosphate-buffered saline (PBS) and fixed for 10 minutes at room temperature in 100% ethyl alcohol. The Alizarin Red solution (40 mM, pH 4.2) was applied to the fixed wells for 10 min at room temperature. Non-specific staining was removed by several washes with water.

Adipogenic differentiation
Cells were seeded onto 6-well tissue culture dishes at a density of 5.0610 5 cells/ well for adipogenic differentiation, and were cultured in aMEM containing 10% FBS, 1 mM dexamethasone, 10 mg/ml insulin, and 0.5 mM isobutylxanthine for 3 weeks [31]. Induced cells were fixed in 10% formalin for 1 hour at room temperature, followed by 20 minutes in 0.3% Oil Red O staining solution (Sigma).

Derivation of p75 high cells from hPSCs
To transfer hPSCs from feeder to feeder-free culture conditions, colonies were dissociated into small cell clumps (about 10 cells/clumps) by pipetting several times, seeded on matrigel-coated dishes (2-4 clumps/cm 2 ), and cultured with mTeSR1 medium for two days. hNCC induction was then initiated by substituting CDM supplemented with chemicals ( Figure 1A). Cells gradually migrated from colonies and proliferated during the induction ( Figure 1B). Cells were harvested after 7 days of being induced, and were subsequently sorted according to the expression of p75 ( Figure 1C). We detected two peaks in p75-positive populations, designated p75 low and p75 high , and the efficiency of hNCC induction was evaluated based on the fraction of p75 high cells.   The effects of SB, which has been shown to inhibit Activin/Nodal/TGFb signaling and induce neural cells and hNCCs from hPSCs without the help of other chemicals, were firstly evaluated [14]. In accordance with the reported data, CDM supplemented with SB successfully delivered p75 high cells with a PAX6positive neuroectoderm from 201B7 (date not shown), while the induction efficiency of p75 high cells was approximately 35% (0 mM in Figure 1C). The activation of Wnt signaling was previously shown to play a key role in the induction of hNCCs [14,15], and can be achieved using the GSK3b inhibitor BIO or CHIR. Therefore, we attempted to determine the most effective concentration of CHIR with a fixed concentration of SB (10 mM) to induce p75 high cells. The results obtained revealed that CHIR successfully induced p75 high cells in a dosedependent manner up to 1 mM, whereas higher concentrations of CHIR markedly inhibited the production of p75 high cells ( Figure 1C). We finally examined the effects of BMP signaling on this induction. The addition of BMP4 markedly inhibited the production of p75 high cells, and these results were compatible with BMP signal inhibiting neural differentiation ( Figure S1A). However, the treatment with DMH1 (10 mM), a specific inhibitor of SMAD1/5/8 phosphorylation, also reduced the p75 high fraction ( Figure S1B). The inhibitory effect of DMH1 on the induction of p75 was confirmed at different dosages, and other cytoplasmic (LDN193189) or extracellular (Noggin) inhibitors for BMP signaling also decreased the efficiency ( Figure S1C). Therefore, the combination of SB (10 mM) and CHIR (1 mM) most effectively induced p75 high cells from 201B7 hiPSCs. This result was reproduced in other hPSCs such as hESCs (H9, KhES1, and KhES3) and episomal hiPSCs (414C2) ( Figure 1D). Most cells outgrowing from colonies were stained with NCC markers, p75 and TFAP2A, whereas the cells in colonies were positive for PAX6, a marker for the neuroectoderm ( Figure 1E).

p75 high cells expressed early NCC markers
The expression of marker genes were compared between p75 high and p75 low cells (  Figure 2A). Sorted p75 high cells expressed a number of genes in the early stage of NCCs, such as SOX10, TWIST, and TFAP2A genes. In contrast, the expression of these genes was significantly lower in the p75 low fraction than in p75 high fraction. The expression of PAX3, which is a marker both for NCCs and neurons, was high in both the p75 high and p75 low fractions. The expression of PAX6 and SOX1, which are neural markers, was higher in the p75 low fraction, which is consistent with some populations of p75-negative or TFAP2A-negative cells expressing PAX6 ( Figure 1E). These results indicated the relative enrichment of NCC cells in the p75 high cell population.
In an attempt to further characterize p75 high cells, genome-wide expression profiles were compared between sorted p75 high cells and their corresponding hPSCs using a cDNA microarray (Affymetrix Gene 1.0 ST), and we found that the overall profiles of p75 high cells derived from several PSCs were similar to each other ( Figure S2). Based on the previous report [32], 46 genes were selected as markers for distinct NC-subpopulations and the expression level of these genes were compared between hNCCs and parental PSCs ( Figure 2B). hNCCs in this study highly expressed early stage-related genes such as neural plate border specifier (PAX3) and NC specification (SNAI2, NGFR, TFAP2A, SOX9, and SOX10), but also some region-specifying genes (EFNB2 for cranial, PDGFRA for cardiac, and SOX5 for trunk region), suggesting the heterogeneous population of p75 high cells, which were designated hNCCs hereafter. The profiles of the current hNCCs were compared with those of two PSC-derived NCCs, which were induced by different protocols in previous studies [15,32] (Figure S3). Although the three types of PSC-derived NCCs all highly expressed some genes, such as SNAI2, their

Sustained expansion of hNCCs with original characteristics
We investigated whether hNCCs could be stably expanded. The growth of hNCCs cultured in the hNCC induction medium (CDM with SB and CHIR) was very slow (data not shown). We employed a cultured condition using CDM supplemented with SB, EGF (20 ng/ml), and FGF2 (20 ng/ml) based on the  findings of previous studies [33], and consequently observed marked improvements in growth and the stable proliferation of hNCCs even after 10 passages (  Figures 3A, B). The expanded hNCCs maintained their original cell morphology and all cells expressing NCC markers, such as TFAP2A ( Figure 3C). The global gene expression profiles of hNCCs after prolonged expansion (PN10) were similar to those of early-passage cells (PN0) ( Figure S4A and S4B, correlation coefficient 50.96 to 0.98) and markedly different from those of original hPSCs (Figures 3D  and S4C).

Modulation of the characteristics of hNCCs by insulin and retinoic acid (RA)
The results of the microarray analyses revealed that induced hNCCs expressed some genes characteristic to cranial NCCs (high for OTX2 and DLX1; low for HOXA2 and HOXA3) (date not shown). A previous study demonstrated that the depletion of insulin from CDM (growth-factor free CDM; hereafter referred to as gfCDM) induced a more anterior neuroectoderm (rostral hypothalamic progenitor-like cells), while retinoic acid (RA) exhibited posteriorizing activity [15]. Therefore, we compared the expression of regional markers in hNCCs cultured with gfCDM, CDM, and CDM with RA (100 nM) ( Figure 4A). As expected, the expression of OTX2, a marker for mesencephalic NCCs ( Figure 4B) [34], was slightly higher under the gfCDM condition than under the CDM condition ( Figure 4C). The DLX1 gene, a marker for first and second branchial arch NCCs ( Figure 4B) [35], was expressed in cells cultured under all conditions, and was the highest in CDM with the RA condition ( Figure 4C). The expression of the HOXA2 and HOXA3 genes, which are markers of the second and third branchial arches, was negligible under the gfCDM and CDM conditions ( Figures 4B, C) [36,37]. Taken together, these results indicated that the regional identities of hNCCs could be modulated by exogenous signals including insulin and RA.

Derivation of peripheral neurons, glia, and melanocytes from hNCCs
We next examined the differentiation potentials of induced hNCCs. Neuronal differentiation was initiated by sphere formation and promoted by culture media containing a mixture of factors (BDNF, GDNF, NGF, and NT-3). Cells expressed b-tubulin and peripherin after 14 days, which indicated differentiation into peripheral neurons ( Figure 5A). Further cultivation under the same conditions (4 to 6 weeks) promoted the glial differentiation of hNCCs ( Figure 5B).
Melanocytes are well-known derivatives of NCCs. Using a previously described method that included CHIR, EDN3, and BMP4 [15,38], induced hNCCs expressed microphthalmia-associated transcription factor (MITF) and c-KIT, markers for melanocytes ( Figure 5C). These differentiation properties were compatible with those of NCCs in vivo.

Derivation of corneal endothelial cells from hNCCs
Cranial NCCs have been shown to exhibit the ability to differentiate into corneal endothelial cells in vivo [39,40]. Therefore, we examined whether hNCCs grown in gfCDM, which preferentially expressed more anterior NCC markers ( Figures 4B, C), could differentiate into cells harboring the characteristics of corneal endothelial cells. When 201B7-derived hNCCs were cultured in the conditioned medium of corneal endothelial cells for twelve days ( Figure 6A), cells changed their morphology into that of polygonal corneal endothelial-like cells ( Figure 6B) and started to express the corneal endothelial marker, ZO-1 ( Figure 6C). Descemet's membrane is known to consist of collagen type 4 and collagen type 8, which are derived from the corneal endothelium [41]. The mRNA expression of the COL4A1 and COL8A1 genes was confirmed in induced endothelial-like cells ( Figure 6D). These results strongly suggested that the hNCCs induced in this study possessed the characteristics of cranial NCCs, which exhibit the potential to differentiate into cranial NCC-derived structures.

Derivation of hMSCs from hNCCs
Cranial NCCs also have differentiation properties toward mesenchymal cells, which construct the cranio-facial skeleton, and may be referred as MSCs [3]. In order to derivate hMSCs from hNCCs, the culture medium was changed from that for hNCC to aMEM with 10% FBS (Figure 7A), which we used previously for human bone marrow-derived MSCs (hBM-MSCs) [24]. Through the induction of hMSCs, the expression of NGFR and SOX10 reduced rapidly within 48 hours (PN0) of the medium change, while that of PAX3 and TFAP2A reduced gradually until passage 3 ( Figure S5A). Conversely, the expression of MSC markers (CD73, CD105, and CD44) increased rapidly within 48 hours, reached a maximum by passage number 2, and maintained their expression at a level comparable to that in BMMSCs ( Figure S5B). These results indicated that the transition from NCCs to MSCs was gradual during passage number three. Cells passaged three times in the medium showed a typical fibroblastic morphology similar to that of hMSCs ( Figure 7B), and expressed surface markers for hMSCs (positive for CD73, CD105, and CD44, and negative for CD45) ( Figure 7C). Microarray analyses revealed that hNCC-derived MSCs had a global expression pattern similar to that of primary hBM-MSCs ( Figure 7D). Differentiation properties toward osteogenic, chondrogenic, and adipogenic lineages are one of the criteria required for MSCs [42], which were clearly confirmed in hNCC-derived MSCs ( Figure 7E). FACS analysis showed that there was no SSEA4-positive cells ( Figure 7F) and the expression of PSC marker genes was below detectable levels ( Figure 7G).

Derivation of osteogenic cells from hiPSCs under defined culture conditions
We determined the feasibility of inducing terminally differentiated cells from iPSCs under defined culture conditions (Figure 8). 987A3 hiPSCs were used as the initial material, which have been generated and maintained under feeder-free and xeno-free conditions [21]. Cells were dissociated into single cells, seeded on iMatrix-coated dishes (0.83-1.35 cells/cm 2 ), and cultured with StemFit medium for five days. hNCCs were then induced for seven to ten days ( Figure 8A). The efficiency of hNCC induction under these conditions was 40.9¡5.5% (¡ SD. N53, biological triplicate). The induction of hMSCs was performed using CDM for MSCs (STK2) instead of aMEM/10% FBS ( Figure 8A). After several passages of hNCCs in STK2, the morphology of cells changed from cuboidal to fibroblastic, similar to that of hBM-MSCs ( Figure 8B). The expression patterns of surface markers were compatible with those of hMSCs (positive for CD73, CD105, and CD44, and negative for CD45) ( Figure 8C) and the differentiation properties for osteogenic, chondrogenic, and adipogenic lineages were confirmed when the standard FBS-containing induction medium was used ( Figure S6). Osteogenic differentiation was also confirmed using the chemically-defined osteogenic medium (STK3) ( Figure 8D). These results indicated that all steps from iPSC to osteogenic cells could be performed under defined culture conditions.

Discussion
In the present study, we developed a simple and efficient induction method for hNCCs from hPSCs. The induction efficiency of this method was high (70-80%) irrespective with the type of hPSC. The induced hNCCs exhibited the cranial NCC characters under maintenance culture conditions, while further treatment with insulin and RA marginally posteriorized hNCCs. Consistent with the expression of cranial NCC markers, induced hNCCs could differentiate into corneal endothelial cells, which is a characteristic of cranial NCCs.
Our protocol was independent of the BMP signal. In our protocol, DMH1, a specific BMP inhibitor, clearly attenuated the induction efficiency of the p75 high fraction ( Figure S1). This result clearly contradicted the findings of previous studies (no effect [14] or increased efficiency [15]). The marked differences in the findings of these studies may be attributed to the seeding density used at the beginning of induction. The seeding density of our protocol was approximately 2-4 clumps/cm 2 (approximately 20 cells/cm 2 ), while other studies used 1610 4 cells/cm 2 [26]. Both CNS and neural crest fates were previously observed when cells were seeded at a low density, while CNS cells primarily formed at a high density [43]. In accordance with these findings, the efficiency of the NCC induction was markedly decreased if clumps were seeded at a higher density (data not shown). The high density of hNCCs may have exaggerated local BMP signaling secreted from the hNCCs themselves. Therefore, we combined high density seeding with the BMP inhibitor treatment; however, the efficiency was still low (data not shown). Based on these results, we could not account for the differences between our protocol and those of previous studies.
In order to compare the hNCCs in this study with those in previous studies, we analyzed gene expression profiles of hNCCs published previously. The comparison of the relative induction levels of NCC specific genes revealed that hNCCs differentiated by our protocol and previous studies showed similarities in some aspects, but overall profiles were different from each other ( Figure S3). These results indicated that induction protocols reported in this study and in the previous studies induced different subset of hNCCs.
Induced hNCCs exhibited differentiation properties for multiple cell lineages including peripheral neurons, glial cells, melanocytes, and corneal endothelial cells, and also delivered hMSCs that further differentiated into osteogenic, chondrogenic, and adipogenic cells. These properties are compatible with NCCs being multipotent stem cells [3]. However, clonal analyses are indispensable for confirming the stemness of induced hNCCs. Previous clonal analyses revealed that 63-65% of the hNCC clones could differentiated into multi-lineage cells positive for markers of neurons, glial cells, and smooth muscle cells [43,44], suggesting that hNCCs induced from hPSCs were multipotent on the clonal level. Although stemness has yet to be investigated in this study, induced hNCCs in this protocol will be a promising cell source for various types of research.
Human diseases that have been related to the development of hNCCs include Hirschsprung's disease, DiGeorge syndrome, Waardenburg syndrome, Charcot-Marie-tooth disease, Hermansky-Pudlak syndrome, familial dysautonomia, Chediak-Higashi syndrome, and CHARGE syndrome [45,46]. hNCCs containing the mutations responsible for these diseases can be induced from hiPSCs established from the respective patients; therefore, this will be a powerful tool for creating in vitro disease models that can contribute to a more detailed understanding of the pathogenesis of NCC disorders and also to the development of novel therapeutic modalities [15]. In addition, hNCCs have been shown to be the cell-of-origin of some cancers such as neuroblastoma [47], which indicates that hNCCs can be used in in vitro transformation experiments. We have already confirmed that the survival rate of freeze-stocked hNCCs was satisfactory and the freeze and thaw process had no impact on the growth and differentiation properties of these cells (data not shown). These are favorable features for a material in research because it is important to use cells of the same quality in order to evaluate reproducibility.
Induced hNCCs-deridatives can also be used for cell therapy. In this regard, hNCC-derived hMSCs will be a very useful material. hMSCs have been used in a wide range of regenerative medicines, and promising results have been reported in some cases [48,49]. In contrast with the advances reported in clinical applications, many issues related to the biology of hMSCs have yet to be investigated, one of which is the cell-of-origin of hMSCs. hNCCs may be the precursors of hMSCs based on the finding that craniofacial skeletal tissues are derived from NCCs [50]. This has also been supported in lineage tracing experiments using P0-cre mice [51,52]. Current sources of hMSCs include bone marrow, fat tissue, synovium, and umbilical cord; however, it remains unclear whether NCC-derived cells exist in all of these adult tissues and serve as the source of hMSCs. A comparison between hNCC-derived MSCs and somatic tissue-derived hMSCs may provide more information related to this issue.
One of the limitations of current hMSCs is their limited proliferative activity, which may pose problems in their application to conditions requiring a large amount of cells. This can be overcome if hNCC-derived MSCs are used because hNCCs can be induced from hPSCs, which have unlimited proliferative activity. Two issues are important for this application. One is to be free from infectious substances that may be derived from animal materials. Using iPSCs generated and maintained under feeder-free and xeno-free conditions, we successfully induced hNCCs and hMSCs with minimum animal material (BSA in CDM) ( Figure 8A). Furthermore, we generated terminally differentiated cells (osteogenic cells) from these MSCs under chemically defined media. To the best of our knowledge, this is the first study to demonstrate the induction of osteogenic cells under feeder-free and serum-free conditions from PSCs. The other concern relates to the contamination of undifferentiated cells, particularly parental hPSCs, which may lead to serious conditions such as the formation of malignant tumors [53]. We confirmed that hNCC-derived hMSCs were free from SSEA4-expressing cells and the expression of PSC-marker genes was below detectable levels ( Figures 7F, G). Although more precise and meticulous analyses are required to prove the safety of these cells, the results of the present study have provided evidence to promote the use of hNCC-derived hMSCs for cell therapy.