Derivation of neural crest stem cells from human epidermal keratinocytes requires FGF‐2, IGF‐1, and inhibition of TGF‐β1

Abstract Neural crest (NC) cells play a central role in forming the peripheral nervous system, the craniofacial skeleton, and the pigmentation of the skin during development due to their broad multilineage differentiation potential into neurons, Schwann cells, melanocytes, and mesenchymal stem cells. Recently, we identified an easily accessible source of pluripotent NC stem cells from human inter‐follicular keratinocyte (KC) cultures (KC‐NC). In this work, we examined specific conditions for the derivation of NC from KC cultures. More specifically, we examined the role of two growth factors, FGF2 and IGF1, in NC proliferation and in expression of two potent NC transcription factors, Sox10 and FoxD3. Using specific chemical inhibitors, we uncovered that the downstream regulatory pathways AKT/PI3K, MEK/ERK, and JNK/cJun may be critical in Sox10 and FoxD3 regulation in KC‐NC. The TGF‐β1 pathway was also implicated in suppressing Sox10 expression and NC proliferation. In summary, our study shed light into the role of FGF2, IGF1, and TGF‐β1 on the induction of NC from KC cultures and the pathways that regulate Sox10 and FoxD3. We also established culture conditions for sustaining KC‐NC multipotency and, therefore, the potential of these cells for regenerative medicine and cellular therapies.

during NC induction: FGF, Wnt, and BMP signaling. 4 These pathways lead to the expression of NC specifiers that define the NC state, albeit transiently, from the beginning to the end of induction. Some of these specifiers include among others, the SoxE family of factors as well as Pax3 and Zic1. [5][6][7][8] While the importance of FGF signaling in the NC induction has been postulated in a number of studies, mainly focusing on the Xenopus embryo model, [9][10][11][12] further investigation is needed to expand these findings in NC stem cells isolated from adult humans.
Genetic mutations can result in dysregulated NC development leading to many congenital human diseases, such as cardiovascular defects and craniofacial abnormities, collectively known as neurocristopathies, 13 myelopathies, neural degenerative diseases, and so forth. Therefore, cultures of human NC cells can provide a model to study human disease and a source of stem cells for treatment of neurodegenerative diseases that may be currently hindered by the lack of an easily accessible and autologous cell source. Interestingly, recent studies have successfully isolated NC cells from different tissues in the adult body, including the adult hair follicle, craniofacial sources such as the palate and the oral mucosa. [14][15][16][17] Recently, our laboratory showed that NC could be derived from cultures of epidermal KCs isolated from glabrous neonatal foreskin. KC-derived NC could be coaxed to differentiate into functional neurons, Schwann cells, melanocytes, osteocytes, chondrocytes, adipocytes and smooth muscle cells, in vitro and in vivo, in lineage tracing experiments in chick embryos. 17 Given the accessibility of human skin, KC-derived NC may provide a valuable source of multipotent stem cells for treatment of myelopathies and other debilitating neurodegenerative diseases. Therefore, it is critical to understand the factors affecting NC derivation, including expansion and maintenance of the NC phenotype and multilineage differentiation potential.
In this study, we focused on the role of growth factors and downstream signaling pathways that may be important in derivation of NC from human KC and identified the culture conditions that may be optimal for NC proliferation and expression of key transcription factors, Sox10 and FoxD3, which have been shown to be critical for maintenance of the NC phenotype and the NC multilineage differentiation potential.

| Isolation of epidermal cells
Glabrous (lacking hair follicles) foreskin from 1-to 3-day-old neonates was procured from John R. Oishei Children's Hospital, Buffalo. Skin

| Immunocytochemistry
Cells were washed with cold PBS (4 C) and permeabilized with 4% (vol/vol) paraformaldehyde (10 min, room temperature; Sigma). Permeabilization (10 min, room temperature) was performed with 0.1% (vol/vol) triton X-100, (Sigma) in PBS and samples were blocked with 5% (vol/vol) normal goat serum (Life Technologies) in PBS. The cells were incubated with primary antibodies overnight (4 C) (Supporting Information Table S1) followed by incubation with appropriate secondary antibodies (1 hr, room temperature) conjugated with Alexa 488 or Alexa 594 (Life Technologies). Hoechst 33342 (Thermo Fisher Scientific, Grand Island, NY) was used for nuclear staining. Cells that were incubated with only secondary antibody served as controls.

| Imaging and image analysis
Immunocytochemistry images were acquired using a Zeiss Axio Observer Z1 inverted microscope with an ORCA-ER CCD camera (Hamamatsu, Japan). The images were acquired using fixed exposure time (NES: 200 ms, SOX10: 400 ms, FOXD3: 500 ms). Cell number quantification was performed using NIH ImageJ. The images were converted to eight-bit. Manual marking and cell counting were performed for NES+, SOX10+, and FOXD3+ cells using the Cell Counter plugin. For each condition, n = 3 separate wells were counted. Statistical significance between the groups was analyzed through Student's t-test (paired, two-tailed) as described previously 18 and a confidence interval of 95% was chosen.

| NC stem cells derived from keratinocyte cultures
An adult NC population has been found in hair follicle's bulge region. 19 To avoid possible contamination from the NC population

| Dynamics of NC induction
To examine the dynamics of NC induction, we followed the expres-    Our results clearly showed that FGF2 was sufficient to induce Sox10 expression even in the absence of serum, to the same extent as the complete medium. FGF2 was also necessary, as blocking FGFR signaling completely eliminated Sox10 and NES expression as well as NC induction altogether. These results are in agreement with studies that used FGF2 to expand NC cells from hair follicles 25 or embryoid bodies. 26 Others implicated FGFR1 in NC migration from the neural plate border area of Xenopus embryos 27 and more recently FGFR4 was also implicated in NC development. 28 However, IGF1 alone was unable to induce Sox10 expression, especially in the absence of serum. However, in combination with FGF2, it increased the percentage of Sox10+ cells and the total number of NES+ cells, suggesting that IGF1 may have promoted proliferation of NC cells. In addition, IGF1 and FGF2 contributed almost to the same extent in FoxD3 expression. Interestingly, their combined action yielded higher percentage of FoxD3+/NES+ cells than the sum of each growth factor alone. This was especially true in the absence of serum, where each factor was unable to induce FoxD3 expression but when added together they resulted in FoxD3 expression in almost all NES+ cells (Figure 8). This is an interesting result that may suggest that the combination of FGF2 and IGF1 may act cooperatively to induce FoxD3 expression, possibly by activating more than one pathway.
Indeed, we identified that at least two such pathways, PI3K/Akt and MAPK/Erk1/2, were necessary for NC induction. Blocking either Furthermore, WB showed that neither NES nor Sox10 was expressed when PI3K/Akt or MAPK/Erk1/2 pathway was blocked. The effect of Akt pathway in pluripotency has been recently discussed in the context of embryonic and progenitor cell differentiation patterns. 29 Furthermore, activating MEK was shown to prolong the expression of pluripotency marker Sox3, 28 a Sox family member that is also regulated by FGF signaling. 30 On the other hand, inhibition of Rac had no effect on the number of NES+ cells after 8 days of induction. It will be interesting to examine whether the PI3K/Akt and MAPK/Erk1/2 pathways are also important in the maintenance of the NC phenotype and/or the differentiation of NC stem cells to NC-specific lineages such as neurons or Schwann cells. Finally, inhibition of TGF-β1 pathway appeared to have a positive effect on the proliferation of NES+ cells and the fraction of Sox10+ cells, suggesting that TGF-β1 might be suppressing Sox10 expression and NES+ cell proliferation.
Previously, we showed that KC-NC are multipotent stem cells that can be coaxed to differentiate into neurons, Schwann cells, melanocytes, and mesenchymal stem cell derivatives (osteocytes, chondrocytes, adipocytes, and smooth muscle cells). Most notably, upon transplantation into chick embryos, KC-NC migrated along stereotypical pathways and gave rise to multiple NC derivatives, providing strong support of their