Generating Functional and Highly Proliferative Melanocytes Derived from Human Pluripotent Stem Cells: A Promising Tool for Biotherapeutic Approaches to Treat Skin Pigmentation Disorders

Melanocytes are essential for skin homeostasis and protection, and their loss or misfunction leads to a wide spectrum of diseases. Cell therapy utilizing autologous melanocytes has been used for years as an adjunct treatment for hypopigmentary disorders such as vitiligo. However, these approaches are hindered by the poor proliferative capacity of melanocytes obtained from skin biopsies. Recent advances in the field of human pluripotent stem cells have fueled the prospect of generating melanocytes. Here, we have developed a well-characterized method to produce a pure and homogenous population of functional and proliferative melanocytes. The genetic stability and potential transformation of melanocytes from pluripotent stem cells have been evaluated over time during the in vitro culture process. Thanks to transcriptomic analysis, the molecular signatures all along the differentiation protocol have been characterized, providing a solid basis for standardizing the protocol. Altogether, our results promise meaningful, broadly applicable, and longer-lasting advances for pigmentation disorders and open perspectives for innovative biotherapies for pigment disorders.


Introduction
Melanocytes are pigment-producing cells derived from the neural crest, a transient embryonic multipotent stem cell population that emerges from the neural plate in vertebrates [1]. Defects in the development or the homeostasis of melanocytes are present at the onset of many pigmentary disorders such as vitiligo, for which therapeutic solutions remain a challenge [2]. Surgical therapy consists of the re-introduction of melanocytes in depigmented lesions. Primary melanocytes can be isolated from pigmented skin of the same person by different procedures (e.g., tissue grafts or cellular grafts) [3]. However, these techniques require surgical biopsy and can induce the depletion of the site donor. In addition, cells derived from primary culture have a limited life span [4]. To overcome these limitations, human pluripotent stem cells (hPSCs) have become a promising alternative as an unlimited source of melanocytes. In recent years, various protocols have been reported to successfully convert hPSCs into melanocytes [5][6][7][8][9][10]. Notwithstanding this progress, the generation of hPSC-derived melanocytes still relies on the enrichment of the desired cell population by cell sorting or mechanical selection. In addition, the proliferative status of hPSC-derived melanocytes and their maintenance during long periods of culture has not been extensively investigated. Here, we develop a multi-step protocol for specifying hPSCs into a homogeneous population of melanocytes without any mechanical or cell-sorting steps. These hPSC-derived melanocytes, generated from the two origins of hPSCs, human embryonic stem cells and human-induced pluripotent stem cells, are suitable for longterm expansion without losing their phenotypes and their functional capacity to integrate 3D in vitro and in vivo skin equivalents when cocultured with primary fibroblasts and primary keratinocytes.

Characterization of a Pure and Homogenous Population of Melanocytes Derived from hPSC
To reproduce the development of the melanocyte chronobiology, we have established a multistep protocol in which human-induced pluripotent stem cells (hiPSCs) and human embryonic stem cells (hESCs) are differentiated into melanocyte precursors by the induction of the neural crest lineage followed by a maturation step ( Figure 1a). First, hPSCs were aggregated in embryoid bodies (EBs). The induction of hPSCs within the neural crest lineage was initiated by the use of dual-SMAD inhibition in combination with Wnt signaling [11,12]. At day 4, and to initiate the differentiation into melanocytes, small molecules including endothelin-3 (EDN3), stem cell factor (SCF), bone morphogenetic 4 (BMP4), and ascorbic acid (AA) were added at specific concentrations [7]. At day 7, EBs were placed on gelatin and grown in the melanocyte medium (MGM-4) supplemented by the same molecules. In these conditions, cells migrated out of the EBs. From day 21 to day 30, cells acquired a melanocyte-like dendritic cell morphology. The induction of melanocyte maturation was performed at day 30. For this purpose, cells were first dissociated with trypsin and plated in MGM-4 supplemented with EDN3, SCF, and forskolin (FSK), essential for the survival and maturation of melanocytes. After two passages, melanocytes with a typical mature morphology were observed (Figure 1a,b). FSK was added to the medium in addition to the EDN3 and SCF to induce the maturation from the cell dissociation at day 30 corresponding to passage 0 (P0) up to passage 2 (P2). Then, the melanocyte population (Mel-hPSC) was maintained in MGM-4 medium along the amplification phase and exhibited a typical morphology of human melanocytes (Figure 1c). As a positive control, human epithelial melanocytes (HEMs) were used for further characterization. Analysis by immunofluorescence staining for melanocytic markers indicated that most of the Mel-hiPSC and Mel-hESC expressed MITF (microphthalmia-associated transcription factor), which is considered the main transcription factor of melanogenesis as well as its targets, such as TYRP1 (tyrosinase-related protein 1) (Figure 1c). This result was confirmed by flow cytometry analysis revealing that more than 99% of cells were positive for TYRP1 expression (Figure 1d). A transcriptomic analysis was performed to analyze the gene expression profile of Mel-hESC in comparison with HEMs. Hierarchical clustering showed similar gene expression profiling between Mel-hESC and HEMs compared to hESCs (Figure 1e). Almost 70% of genes upregulated in comparison to hESCs were found to be common between Mel-hESC and HEMs (Figure 1f). Gene ontology analysis of these common genes revealed a close genetic signature with biological processes related to "melanocyte differentiation" and the "pigmentation process" (Figures 1f and S1(i)). Moreover, the gene expression levels were highly correlated between patient iMels and HEMs (R = 0.97) (Figure 1g). These results demonstrated the generation of a pure population of melanocytes.

Functional Characterization of hPSC-Derived Melanocytes
The functionality of melanocytes derived from hPSCs was first assessed by the quantification of the pigmentation in cell pellets by using spectrophotometry analysis. The quantification of the melanin content demonstrated the capacity of mel-hiPSC and mel-hESC cells to synthesize melanin (Figure 2a). Electron microscopy analysis also showed the presence of the four maturation stages of melanosomes localized in the cytoplasm of Mel-hPSC as observed in HEMs (Figure 2b). Finally, we evaluated the capacity of Mel-hPSC to integrate into an in vitro and in vivo 3D reconstructed skin. For this purpose, primary human keratinocytes (K) were added to the equivalent dermis with HEMs or Mel-hPSC. They were expanded until confluency and were then maintained at the air-liquid interface in vitro for 14 days or grafted in vivo onto nude mice for two weeks (Figure 2c). Macroscopic analysis revealed a wellpigmented epidermis under in vitro coculture conditions using Mel-hPSC and K comparable to those observed in the HEM and K conditions (Figure 2d). The presence and distribution of melanocytes and keratinocytes were confirmed by immunostaining for involucrin (IVL) in keratinocytes and TYRP1 in melanocytes (Figure 2d). Their localization was also confirmed by additional immunostaining for Keratin K14 (KRT14) and TYROSINASE (TYR) ( Figure S2). In parallel, the 3D reconstructed skin in vivo showed a normal tissue architecture two weeks after grafting. As shown by the Fontana-Masson coloration, Mel-hPSC produced melanin. Immunostaining analysis also revealed the expression of TYRP1 as well as the localization of the mel-hiPSC and mel-hESC in the basal layer of the epidermis (Figure 2e). Altogether, these data demonstrated the high functionality of Mel-hPSC.

Long-Term Expansion of hPSC-Derived Melanocytes and Genomic Integrity
Next, we investigated the proliferative potential of Mel-hPSC by measuring the doubling time at early (from passage 3 to passage 15) and late passages (>passage 15) (Figure 3a). At early passages, both Mel-hESC and Mel-hiPSC maintained a stable doubling time between 50 and 80 h. At later passages, an increased doubling time over 100 h was observed. At passage 26, the doubling time reached 200 h and 120 h for Mel-hiPSC and Mel-hESC, respectively ( Figure 3a). By determining the cumulative cell expansion for Mel-hPSC until P15, our results demonstrated the possibility of generating a large-scale bank of approximately 10 14 cells (Figure 3b). Furthermore, we showed that over time in culture, Mel-hPSC maintained their dendritic morphology ( Figure 3c) and expressed all the melanocytic markers, as shown by immunostaining (Figure 3c), flow cytometry ( Figure 3d) and immunoblot analysis (Figure 3e). We next evaluated the genomic integrity of Mel-hPSC maintained for a long period in culture. The karyotyping of Mel-hPSC at late passages showed no chromosomal abnormalities ( Figure S3). We also sought to analyze the appearance of mutations of 50 oncogenic and tumor suppressor genes using the AmpliSeq Cancer Hotspot Library v2 (Thermo Fisher Scientific: Waltham, MA, USA) in Mel-hPSC at early and late passages. We used one melanoma cell line (SK-MEL-28 from ATCC) carrying the pathogenic mutations in EGFR, BRAF, and PTEN (https://maayanlab.cloud/SK-MEL-28, accessed on 10 January 2023) as a positive control. The presence of single nucleotide polymorphisms (SNP) was also analyzed in Mel-hPSC, revealing the appearance of neutral mutations as referred to by the database FATHMM (http://fathmm.biocompute.org.uk, accessed on accessed on 6 March 2017). Altogether, these results indicated the absence of a genomic tumor in the Mel-hPSC at late passages. To complete this genomic analysis, a soft agar colony formation was assessed on Mel-hPSC at early and late passages to evaluate their potential malignant transformation. While large aggregates were observed with the SKMEL28 melanoma cell line, no aggregate was detected with Mel-hPSC and HEMs. These results showed that over time in culture, the Mel-hPSC is not transformed (Figure 3g).

Functional Genomic Analysis of Melanocyte Differentiation and Maturation Processes
We next sought to determine the molecular pathways that drive melanocyte differentiation by using next-genome sequencing (NGS) technology and differential expression of genes (DEG) analysis. The expression pattern of the whole genome was analyzed in a time-dependent manner at day 0, day 7, day 14, day 21, and day 30 of the differentiation process. Principal component analysis (PCA) showed a similar transcriptional profile between replicates of each time point, demonstrating the robustness of the protocol (Figure 4a). This observation was supported by the hierarchical clustering overview of the transcriptomic analysis ( Figure 4b). To decipher the molecular and cellular pathways involved in each stage of the developmental process, we performed a kinetic analysis on differentially expressed genes for every comparison (respectively, day 0 versus day 32, day 7 versus day 32, day 14 versus day 32, and day 21 versus day 32). Genes were regrouped in eight clusters according to the kinetics of their differential expression (Table 1). Five clusters have been analyzed by gene ontology using the ENRICH'R interface ( Figure S4). These five clusters are shown in Figure 4c. Cluster 1 included genes primarily related to the pluripotency, morphogenesis, and development of epithelial tissues. Cluster 1 was gradually repressed from day 0 to day 30. In the second cluster, the expression of genes associated with cell division and the regulation of the cell cycle increased from day 0 to day 7 and then decreased as the cells migrated out of the EBs. Cluster 3 is composed of several genes involved in the Wnt signaling pathway and in the negative regulation of embryonic development and neurogenesis. Their expression gradually increased from day 0 to a peak on day 14. The molecular signaling pathway involved in melanogenesis is grouped into two clusters (4 and 5) and included 338 genes. These genes showed a differential expression between day 21 and day 30. Functional enrichment analysis on cluster 5 revealed 19 genes directly connected to MITF, the master regulator of melanogenesis, among which 6 genes including SOX10, DCT, TYR, and RAB27 (yellow circles in Figure 4d) are defined as being involved in "developmental pigmentation" (Figure 4d). To confirm our data, the expression profiles of these markers were validated by quantitative real-time PCR (Figure 4e). While all of this molecular characterization was performed during the differentiation processes, we next sought to molecularly define the maturation step. We thus compared gene expression profiles between P0 and P4, which are immature and mature Mel-hESC, respectively. These transcriptomic analyses showed a differential gene expression pattern at P4 when compared to P0 (Figure 5a). Many differentially expressed genes (DEGs) were identified (p-value adjusted ≤ 0.5; fold change ≥ 1.5; minReads100) with 1309 upregulated and 1719 downregulated DEGs. Gene ontology analysis of the upregulated DEGs revealed biological processes associated with "lytic vacuole", "lysosome", and "melanosome" (Figure 5b). Gene expression analysis for MITF and its downstream target genes (TYPR1, TYR, and PMEL17) in immature Mel-hESC at P0 to the mature P6 stage revealed an increase in melanogenesis genes (Figure 5c). Flow cytometry analysis also showed an increase in the level of TYRP1 expression with the passages of Mel-hESC (Figure 5d). Similarly, Western blot analysis indicated that the major tyrosinase enzyme involved in melanin synthesis was not expressed at passage 0 and became progressively expressed after several passages ( Figure 5e). Accordingly, a progressive increase in melanin synthesis was observed from passage 2 to 6 ( Figure 5f). Thus, our data indicated the acquisition of a mature phenotype of Mel-hESC with the passages.   The global network obtained is sent to Cytoscape to identify a gene sub-network directly linked to MITF. A score is attributed to edges in terms of connectivity with MITF [13]. The size of gene circles is correlated with these scores. Genes circles colored in yellow are members of the "developmental pigmentation" gene ontology biological process term. identify a gene sub-network directly linked to MITF. A score is attributed to edges in terms of connectivity with MITF [13]. The size of gene circles is correlated with these scores. Genes circles colored in yellow are members of the "developmental pigmentation" gene ontology biological process term. (Lower) Top gene ontology biological process terms from enrichment analysis of the global STRINGdb network. Quantitative PCR of SOX10, MITF, DCT, TYR, and RAB27A during hESC differentiation from D0 to D30. The data are normalized to 18S and expressed as relative expressions of undifferentiated hESCs at D0. (e) Cluster 5 validation. Quantitative PCR of SOX10, TYR, MITF, DCT and RAB27 for Mel-hESC from D0 to D30. The data are normalized to 18S and expressed as relative expressions of D0. ** p < 0.01.  LGALS3BP SEMA4D  PTGES3  WNK3  DNAJB1  HERPUD1  CPEB4  CYP27A1  MARVELD1  DPP7  GSDMD  ARID2  LMNB1  HIST1H4I  WASF3  BRP44L  ROPN1B  ARMC9  GOLGA7B  LRRN4CL  LIG1  HIST1H3E  UXS1  HIST1H4L  NAV1  IDH3G  PLSCR4  LMO7  CD109  GBA  SIPA1L3  SIPA1L1  ARID3A  SMOC1  SOX3  RAB27A  STAM2  SETDB2  FBXO32  C7orf41  UNC5B  ZNF770  EPHA4  LGR4  LGR5  WIPI1  KCNJ13  GPNMB  FZD1  MAB21L1  SLC9A3R1  WDHD1  C1orf109  CCDC85C  ATBF1  ST6GALNAC2 MLPH  NDUFB5  C9orf91  GREM2  PVRL1  DAXX  ARID1A  SMN2  BOC  SEL1L  PLCL1  ABI1  TNFRSF14  SLC9A9  SLC38A1  EIF5AL1  KDM3B  NEURL1B  SFRP2  CYBRD1  SLC1A4  STXBP1  TRIM63  AP1S2  NUP155  LYPLA1  HIST1H2AM SCUBE3  CDH19  TRAK2  NOV  CYB5R1  NTM  GALNT3  NCAPH  KIFC1  CDKN2B  SLC6A15  SCAMP3  TFAP2A  SPON2  TMEM119  MSH6  ZWINT  CRB2  PDE8A  GBP1  IRF4  S100B  SOCS3  HDAC1  CIT  AP000654.2  PTGS2  RAB32  CPEB2  LMNA  TMEM173  OLFML3  PPAT  TMEM132D  TYR  KLF9  SULF1  C19orf28  NDUFA4L2  TTF2  SMO  WNT7A  GPR137B  ITGB1BP1  THBS1  JOSD2  SHC4  RCAN3  CNN1  DPYSL5  FAP  NR4A3  IFI44  GRASP  KCNQ5  PLXDC2  NUP210  KLHDC8A  CEACAM1  RHOQ  ANXA7  NFIA  IRS2  PLS1  PIK3C2B  PTX3  STX7  IFIT3  ARHGAP24  TM2D1  THBS2  HNRNPA2B1  CDCA8  MAMDC2  DCT  MLANA  CPNE8 LHX8 C6orf1 LGALS1 LRRC39  NFIC  SHH  LRRK2  HRSP12  POU2F1  FAT4  SOX10  INHBA  TMEM50B  BRI3  APOD  NASP  SOX13  VEPH1  TIMP3  SRGN  EMP3  CTSB  TDRD7  SYT6  SKP2  HOXC6  TSPO  BICC1  ECM1  NOS3  MME  DSC2  LIX1  ELOVL2  ATP6V1D  BHLHE41  CTSK  CRYL1  CASP4  HEY2  ADAMTS19  DLL1  CINP  ZC3H13  DUSP10  C10orf10  ANXA4  USP44  ZMYM4  POU3F3  CTSZ  RAB38  LYST  MFAP4  METTL9  LECT1  NR6A1  BIRC7  F13A1  ARL8A  NNMT  ANXA6  SORL1  QTRTD1  TRIB3  SERPINB6  MANF  SLFN5  SLC6A17  FRAS1  UTRN  PLP2  SBDS  EPHA5  STAT6  LRP10  PTPRF  SKA1  RENBP  SIX1  SNCA  AKTIP  UAP1L1  MCL1  SFRS15  ASB9  MASP1  DDIT4L  SLC3A2  DACT3  CDCA5  PPP1R9A  ACP5  ZFP36  OSMR  LTBP3  GSTK1  LPCAT1  IGF2BP1  PLA2G15  VGF  CXCL14 LGI3 C6orf192  IL1RAPL1  SHISA4  PDPN  USP49  PCOLCE  NDFIP1  SDK1  SLC33A1  TGM2  PKDCC  PLEKHA7  AHR  UBE2D2  TACC1  IL8  CARD16  PQLC3  HIC2  AEBP1  TMEM204  EBAG9  HMP19  CFI  ZNF589  IRX2  SFXN3 LGALS3 NFIB CRTAP CAMKV PPID DNAJC12 MATN3 GSN SEMA3E

Discussion
In this study, we provide a multi-stage protocol for generating a pure and homogenous population of functional melanocytes derived from human pluripotent stem cells associated with a description of the molecular signatures during the differentiation and maturation processes.
Our method is based on a multi-step protocol which includes a neural crest induction phase followed by sequential phases of differentiation, maturation, and amplification. This temporality led to the generation of mature and functional melanocytes from both the human embryonic stem cell and human-induced pluripotent stem cell lines, with equal efficacy. We based our approach on the temporal regulation of a combination of cytokines and growth factors at specific concentrations. To obtain an efficient population of neural crest cells in 4 days, we used dual-SMAD inhibition in combination with a Wnt signaling activator in a neural medium. To direct the neural crest cells into melanocytes, we used SCF, EDN-3, and Chir99021, a Wnt activator. This combination of molecules was described as necessary for melanocyte development and survival [14,15]. In addition to these molecules, and based on our previous works [7,16], BMP4 and AA at specific concentrations were added to generate melanocytes after 30 days of differentiation. A major difference in our approach compared to previously published protocols lies in the sequential use of one of the most potent cAMP inducers, the forskolin. In our method, cells were treated with forskolin only during the maturation phase. This activation leads to the transcription of MITF, which in turn regulates the principal enzyme of melanin production [17,18]. In 7 to 10 days, maturation of the unpigmented melanocytes was observed by the acquisition of a dendritic morphology, which is characteristic of the primary human melanocytes as well as by their ability to synthesize melanin. These results suggest that forskolin may act as a maturation agent. Another important difference resides in the fact that our approach is not based on manual selection or FACS, as is the case in a number of previously published studies [7,10,19]. These techniques are time consuming, experimenter dependent, and can lead to high variability. With the sequential addition of specific cytokines mimicking the chronobiology of melanocyte development, we obtained, from day 30, a pure population of unpigmented cells stained positive for the melanocytic enzyme TYRP1 at more than 90% but negative for Tyrosinase, the main enzyme of the melanin synthesis. Tyrosinase expression is observed after 7-10 days in the presence of forskolin, correlating with melanin production. Altogether, our results demonstrated the possibility to generate, without any selection, a pure population of a precursor of melanocytes capable of maturing in the presence of cAMP inducers.
Pigmented melanocytes can then be easily amplifiable in a medium suitable for the long-term culture of melanocytes. At the functional level, our results demonstrated that mel-hESC and mel-hiPSC are capable of producing melanin, synthesizing melanosomes, and integrating into an in vitro reconstituted pluristratified epidermis. In addition, in vivo experiments have demonstrated their ability to move to the basal layer of the epidermis when incorporated into immunodeficient mice. Overall, these experiments demonstrate the production of highly functional mature melanocytes.
To date, most of our knowledge about the molecular mechanisms involved in melanogenesis has been derived from experiments in mice [20]. Due to architectural and functional differences between human and mouse skin, extrapolation of results from mouse to human remains limited. Thanks to a transcriptomic approach, five different clusters of genes have been highlighted as representative of the developmental stages of human melanocytes. Several clusters of genes have been identified, including the important canonical Wnt signaling pathway, that play a role in neural crest induction [21], and then in melanocytic lineage induction [1]. Furthermore, the differentiation process in melanocytes was associated with a signature of genes related to the melanogenesis process, which demonstrated the efficiency of the protocol [22]. In addition, to characterize the maturation process, the gene expression profile of mature (pigmented) melanocytes was compared with immature (non-pigmented) melanocytes. We identified well-known genes upregulated in mature melanocytes such as Tyrosinase and TYRP1, the main melanogenesis enzymes [23]. Our analysis revealed additional genes involved in the biological processes of melanosomes, lysosomes, and pigment granules. Interestingly, several of these genes encode for different isoforms of the RAB family members. Functional impairment of RAB proteins such as Rab27A has been reported to cause pigmentation defects [24]. Our analysis revealed the differential expression of RAB isoforms in mature melanocytes compared to the immature population. Although speculative at this stage, further experiments could be envisaged to evaluate their defects in melanosome biogenesis and transport. The molecular characterization of the maturation step provides new tools to study different stages of human melanocytes. Altogether, our results showed a well-characterized sequential protocol that allowed the generation of melanocytes from hPSCs through the neural crest induction followed by the generation of melanocyte precursors and a maturation step into functional melanocytes. Apparition of specific markers at a specific time point led to a molecular signature, allowing the molecular qualification of the defined steps from the neural crest induction to the melanocyte speciation. These findings will help to standardize the protocol.
Another advantage of the method described in this study resides in the long-term stability of produced melanocytes without the loss of their phenotype. This work opens new perspectives in terms of cell therapy for patients with pigmentation defects. The most widespread acquired depigmentation disorder is vitiligo. No known treatment can consistently induce repigmentation in all patients [25], and for severe vitiligo, the only option is surgery [26]. The efficacy of surgical methods has always been variable, and this technique can cause the depletion of donor sites. The life span of adult melanocytes after transplantation also seems to be limited to less than one year, which limits the treatment possibilities both at the surface and in the number of re-applications. Thus, the therapeutic potential of hiPSCs in autologous transplantation appears to be a promising option. To date, the capacity of melanocytes derived from human pluripotent stem cells to be maintained long term in culture has not been fully investigated. It has been previously reported that the cell proliferation capacity of Mel-hPSC decreases with passages, and after several generations in culture, cells begin to slowly proliferate and differentiate [4]. In our previous published protocol, we isolated melanocytes and amplified them for up to 12 passages [7]. More recently, Cohen et al. demonstrated the proliferative activity of melanocytes derived from different donors during sub-culture from passage 6 to passage 12. In this study, we increased the purity of differentiated populations, which is important to warrant safety prior to clinical cell therapies and show the capacity to generate a highly effective functional melanocyte. We characterized the genomic stability, and potential transformation during the in vitro process of melanocytes at low and late passages was evaluated. The newly melanocytes were able to be amplify for up to 26 passages (more than 250 days).
For the purpose of cell therapy, a number of tests verifying the level of tumorigenicity, distribution, and immunogenicity will be necessary [27]. Recent studies showed that autologous transplantations of iPSC-derived cells were performed with no serious adverse events noted [27,28]. In this study, we evaluated whether over time in culture, mel-CSP may lead to genomic mutation and tumor formation. To accomplish this, we tested, at early and late passages, the most common genes, including HRAS, NRAS, and BRA, known to induce early-stage melanoma transformation and which are not mutated in mel-CSP. There was also no transformation observed over time in culture. For the purpose of cell therapy, the risk of melanomagenesis after transplantation must be further evaluated by reglementary toxicology studies in animal models. Liu et al. showed the absence of tumorigenicity 7 weeks post transplantation of mel-iPSC. In their model, the mel-iPSC from vitiligo patients was injected into the back skin of nude mice and mixed with dermal fibroblasts and epidermal cells isolated from the skin of neonatal mice [29]. However, one of the most important factors known to activate melanomagenesis is ultraviolet light [30,31]. Although narrow-band UV-B phototherapy is used in the context of autologous melanocyte cell transplantation to enhance pigmentation with no side effects reported [3,32,33], additional UV exposure experiments of mel-CSP must be assessed before grafting.
In addition, an important mechanism of vitiligo pathogenesis is a CD8 T cell-mediated autoimmune disease hypothesis. Antigen-specific T-cell reactivity to HLA-A2-restricted melanocyte epitopes, which includes gp100, tyrosinase, and melanA/MART, was detected [30,31,34,35]. While the HLA-A subtype is expressed in 35 to 45% of the population [31], an increase in HLA-A2-positive results has been observed in most vitiligo patients [36,37]. Furthermore, a positive correlation between vitiligo disease activity and reactivity to the melanocyte antigen gp100 has been described [38]. One possible way to avoid the destruction of melanocytes derived from iPSC transplantation could be to combine cell-based therapy with antagonist peptide ligands that might block specific T-cell responses.
Overall, the development of a well-characterized system to generate melanocytes from hPSC will serve to improve the knowledge of the mechanisms involved in the onset of pigmentation disorders and may open up new possibilities for cell-based therapeutic applications.

Materials and Methods
Cell culture. hESCs from one cell line, SA-01 (Cellartis, Götenborg, Sweden), and hiPSCs from one cell line 207c02 were reprogrammed from dermic fibroblasts and grown as previously described [16]. For the differentiation, embryonic bodies (EBs) were formed from hESCs and hiPSCs. Briefly, after a first trypsinization with trypsine/EDTA 0.05% (Ther-moFisher, Waltham, MA, USA) for 2-3 min to eliminate the feeder cells, Stempro Accutase cell dissociation reagent (ThermoFisher) was applied for 1-2 min to detach hPSCs in small clumps. Cells were subjected to spinning for 3 min at 110× g and delicately resuspended to avoid the formation of single cells. Cells were grown on low-attachment dishes in a neural medium composed of neurobasal and Ham's F12 (ratio 1:1) complemented with 2% of B-27 without vitamin A and with 1% of N-2 (all from ThermoFisher). Neural crest induction was performed by using CHIR-99021 Quantitative RT-PCR (QRT-PCR). Total RNA was isolated from hESCs, hiPSCs, and HEMs using RNeasy plus Mini extraction kit on Qiacube (Qiagen, Hilden, Germany) including an on-column DNase digestion step according to the manufacturer's protocol. The total RNA was isolated and quantified using a Nanodrop 2000 spectrophotometer (ThermoFisher, Waltham, MA, USA). A measure of 500 ng of total RNA was used for reverse transcription using the Superscript III reverse transcription kit (ThermoFisher). QRT-PCR analysis was performed using a QuantStudio™ 12 Flex instrument device and Luminaris Color HiGreen QRT-PCR Master Mixes Low Rox (ThermoFisher) following the manufacturer's instructions. The PCR amplification process comprised 40 cycles of denaturation at 95 • C for 10 s, annealing at 55 • C for 30 s, and extension at 95 • C for 5 s. The quantification of gene expression was based on the ∆Ct Method and normalized on 18S expression. QRT-PCR was performed using the primers described in Table 2. Table 2. Primers used in qPCR analysis.
Histology. Paraffin sections were stained with hematoxylin and eosin (H&E) using standard protocols.
Fontana Masson staining. Paraffin-embedded sections were deparaffined using Ventana BenchMark XT according to the manufacturing datasheets and stained using a Melanin Staining Kit (Abcam; ab150669, Cambridge, UK) according to the manufacturer's protocol.
Melanin quantification. Mel-CSP was cultured at 1 × 10 5 cells per well in a 6-well plate for 48 h. After centrifugation, the cells pellets were dissolved in 1N NaOH (Sigma-Aldrich) for 1 h at 65 • C. After centrifugation at 12,000× g for 10 min, the supernatants were transferred to 96-well plates with a standard curve (0-50 µg/mL) prepared from synthetic melanin diluted in 1N NaOH. The melanin content was measured by absorbance at 405 nm on the microplate reader Clariostar (BMGLabtech, Ortenberg, Germany) with its softwares (Clariostar ® v5.70 R3 and MARS®v4.01 R2). Melanin contents were analyzed according to a linear regression line obtained with a graduated concentration of synthetic melanin.
AmpliSeq Cancer Hotspot Panel v2. The cDNA library was constructed using Ion AmpliSeq Cancer Hotspot Panel v2 and Ion AmpliSeq Library kit v2 and barcoded using an Ion Xpress Barcode Adapter (ThermoFisher, Waltham, MA, USA). The samples were quantified using an Agilent High-Sensitivity DNA kit and sequenced on an Ion Proton platform using an Ion PI Hi-Q sequencing 200 kit chemistry. The sequencing results were aligned, and variant caller analysis was generated using Ion Torrent suite software and annotated on the COSMIC database or FATHMM database.
AmpliSeq sequencing. The cDNA library was constructed using the Ion AmpliSeq Transcriptome Human Gene Expression kit and barcoded using an Ion Xpress Barcode Adapter (ThermoFisher). The samples were quantified using an Agilent High-Sensitivity DNA kit and sequenced on an Ion Proton platform using an Ion PI Hi-Q sequencing 200 chemistry kit. The sequencing results were aligned on an hg19 and analyzed with Partekflow (v6 Partek Inc., Chesterfield, MO, USA) and Partek Genomic Suite.
Soft agar assay for colony formation. A base 0.5% agar (Sigma-Aldrich) was prepared under a top 0.3% agar containing cells seeded at 25,000/cm 2 . Plates were incubated at 37 • C in a humidified incubator for 30 days. Cells were fed 1-2 times per week with cell culture media (DMEM + glutamax + 10% FBS, all from ThermoFisher. At day 30, plates were stained with 0.5 mL of 0.005% Crystal Violet (Sigma-Aldrich) for 1.5 h. Colonies were observed and counted under a dissecting microscope.
Karyotype (G-banding). A total of 10 6 cells were blocked in metaphase by adding colchicine (Eurobio, FR) at a final concentration of 1 mg/L for 90 min and then washed twice with PBS. They were detached from the culture dish using an incubation of 5 min with 0.05% trypsin/EDTA, and the resulting cell suspension was transferred into a 15 mL falcon tube and centrifuged. The supernatant was discarded, and the pellet was suspended in 8 mL of hypotonic solution of KCl (5.6 mg/mL) (Sigma-Aldrich) for 25 min at 37 • C and fixed with Carnoy solution (3v methanol/1v acetic acid, all from Sigma-Aldrich. Drops of cell suspensions were spread on several Superfrost histological slides and allowed to dry at room temperature overnight. For G banding, slides were placed directly in a trypsin solution 1× for 25 s, rinsed quickly in two baths of PBS, and then stained with Giemsa (Sigma-Aldrich) for one minute and rinsed under running water. We used our Metasystem platform to identify the metaphases via the Metafer 4 program (version v3.11.8WK). Metaphases were photographed with an AxioImager Zeiss Z2 microscope combined with a camera cool cube and 10× and 63× objectives. For G banding, 50 metaphases were analyzed with Ikaros software (version v5.7.8 WK) and chromosomes were classified into 6 metaphases. For mFISH, hybridization was performed according to the supplier's protocols (MetaSystems, Altlussheim Germany) and slides were incubated overnight with 24× Cyte Human Multicolor FISH Probes. Images were captured with MetaSystems platforms and were analyzed with Isis software (version v5.7.8 WK, MetaSystems).

Data Availability Statement:
The raw and analyzed datasets generated during the study are available for research purposes from the corresponding authors. Transcriptomic datasets related to this article can be found at https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE222639 hosted by the Gene Expression Omnibus database (accession number GSE222639).