Stem Cell Reviews

, Volume 4, Issue 4, pp 256–260

Epidermal Neural Crest Stem Cells (EPI-NCSC) and Pluripotency


    • Institute of Human Genetics and North East England Stem Cell Institute, Newcastle University, International Centre for Life
    • Department of Cell BiologyNeurobiology and Anatomy, Medical College of Wisconsin
  • Yaofei Hu
    • Department of Cell BiologyNeurobiology and Anatomy, Medical College of Wisconsin

DOI: 10.1007/s12015-008-9042-0

Cite this article as:
Sieber-Blum, M. & Hu, Y. Stem Cell Rev (2008) 4: 256. doi:10.1007/s12015-008-9042-0


This article serves three purposes. We summarize current knowledge of the origin and characteristics of EPI-NCSC, review their application in a mouse model of spinal cord injury, and we present new data that highlight aspects of pluripotency of EPI-NCSC. EPI-NCSC are multipotent stem cells, which are derived from the embryonic neural crest and are located in the bulge of hair follicles. EPI-NCSC can undergo self-renewal and they are able to generate all major neural crest derivatives, including neurons, nerve supporting cells, smooth muscle cells, bone/cartilage cells and melanocytes. Despite their ectodermal origin, neural crest cells can also generate cell types that typically are derived from mesoderm. We were therefore interested in exploring aspects of EPI-NCSC pluripotency. We here show that EPI-NCSC can fuse with adult skeletal muscle fibers and that incorporated EPI-NCSC nuclei are functional. Furthermore, we show that adult skeletal muscle represents an environment conducive to long-term survival of neurogenic EPI-NCSC. Genes used to create induced pluripotent stem (iPS) cells are present in our EPI-NCSC longSAGE gene expression library. Here we have corroborated this notion by real-time PCR. Our results show similarities in the expression of Myc, Klf4, Sox2 and Lin28 genes between EPI-NCSC and embryonic stem cells (ESC). In contrast there were major differences in Nanog and Pou5f1 (Oct-4) expression levels between EPI-NCSC and ESC, possibly explaining why EPI-NCSC are not tumorigenic. Overall, as embryonic remnants in an adult location EPI-NCSC show several attractive characteristics for future cell replacement therapy and/or biomedical engineering: Due to their ability to migrate, EPI-NCSC can be isolated as a highly pure population of multipotent stem cells by minimally-invasive procedures. The cells can be expanded in vitro into millions of stem cells/progenitors and they share some characteristics with pluripotent stem cells without being tumorigenic. Since the patients’ own EPI-NCSC could be used for autologous transplantation, this would avoid graft rejection.


EPI-NCSCNeural crestSkeletal muscleNeuronSox10MycKlf4Sox2Lin28Oct-4Pou5f1NanogiPS cell


EPI-NCSC are derived from the embryonic neural crest. The neural crest represents a population of ectodermal cells that arise from the neural fold, the boundary between somatic ectoderm and neuroectoderm. Embryonic neural crest cells leave the forming neural tube and migrate to different locations, where they give rise to a number of vital and diverse tissues in the adult organism. Neural crest derivatives include the autonomic nervous system, enteric nervous system, dorsal root ganglia, endocrine cells and melanocytes of the skin and internal organs. Neural crest cells also migrate to the cardiac outflow tract where they contribute to its septation into aorta and pulmonary artery and form smooth musculature of the cardiac outflow tract and great vessels. Further, neural crest cells migrate to craniofacial locations and generate craniofacial bone and cartilage, meninges, tooth papillae, stroma of the cornea, and striated muscle of the eye, among other structures [8].

One fascinating aspect of EPI-NCSC is that they represent remnants of an embryonic tissue that persists in the adult organism, the bulge of hair follicles [14, 15]. The bulge is a specialized region within the outer root sheath that serves as a niche for epidermal stem cells, which generate skin, hair and sebaceous gland. Our data showed that the bulge in fact contains two types of stem cell, epidermal stem cells and EPI-NCSC. Neural crest cells, in particular melanocyte progenitors, are expected within the hair follicle because melanocytes give the hair its color by injecting melanosomes into keratinocytes, which in turn form new hair. The surprising facet of our studies was the presence of multipotent neural crest-derived stem cells in the bulge of hair follicles. The hair follicle is not the sole location in which multipotent neural crest derived stem cells/progenitors persist. Multipotent neural crest cells have been described previously in several embryonic and adult locations, the spinal and sympathetic ganglia [1], the ectoderm [12], the cardiac outflow tract [4], the sciatic nerve [5] and the gastrointestinal tract [6]. In view of future cell replacement therapy and biomedical engineering, the location in the adult hair follicle is of particular interest, as the cells are easily accessible by minimally invasive procedures.

In order to isolate EPI-NCSC as a highly pure population of stem cells, we take advantage of the migratory ability of neural crest cells. We explant bulges of whisker follicles into collagen-coated culture plates. Within approximately 3 days, neural crest cells start to emigrate from the bulge explant onto the collagen substratum and they multiply rapidly [14, 15]. Explantation of the bulge thus changes the environment of the stem cell niche in a way that activates proliferation and migration of neural crest stem cells.

By in vitro clonal analysis we have shown that EPI-NCSC are multipotent stem cells, as the progeny of one stem cell can generate multiple neural crest derivatives, including neurons, Schwann cells, myofibroblasts/smooth muscle cells, chondrocytes/bone cells, and melanocytes. Serial cell cloning in vitro showed that secondary clones also contained multiple cell types, thus providing proof that EPI-NCSC can undergo self-renewal. All cells that emigrate from bulge explants are neural crest cells. Of those, approximately 83% are multipotent stem cells despite the fact that the culture conditions used in that study favored cell differentiation. The remaining cells are assumed to represent melanocyte precursors and/or other types of neural crest-derived progenitor. In culture, EPI-NCSC quickly outgrow them due to rapid proliferation. Differentiation of EPI-NCSC can be directed into a particular lineage. In the presence of bone morphogenetic protein-2 (BMP-2) for instance, cultured EPI-NCSC preferentially generate bone/cartliage cells, whereas neuregulin-1 cause them to differentiate preferably along the Schwann cell lineage [15].

EPI-NCSC are not limited to the whisker follicle, but are also present in back skin hair follicles [14]. However, culturing of back skin hair according to conventional published methods leads to two types of fast-growing colonies. One type consists of crest-derived cells, the other of non-crest cells. For this reason it is our preference to dissect individual hair follicles and take advantage of the neural crest-specific migratory ability of EPI-NCSC, which leads to a pure population of neural crest-derived cells.

We have characterized EPI-NCSC at the molecular level by long serial analysis of gene expression (longSAGE; [3]). We prepared three longSAGE libraries, using RNA from embryonic neural crest stem cells, from in vitro differentiated progeny, and from bulge-derived EPI-NCSC (, series number GSE4680). We found that, as expected, embryonic and adult neural crest stem cells share a similar gene expression profile. In contrast, the EPI-NCSC gene expression profile differs significantly from epidermal stem cell profiles published by others [11]. By eliminating genes that are also expressed by epidermal stem cells, we defined a panel of 19 genes, which are abundantly expressed by both embryonic neural crest cells and EPI-NCSC, but not by epidermal stem cells. We used this ‘molecular signature’ to compare it with gene expression profiles of other cells, and found that EPI-NCSC are unique among skin-resident stem cells/progenitors [3].

EPI-NCSC Grafts in a Mouse Model of Spinal Cord Injury

In order to conduct a preliminary investigation into the suitability of EPI-NCSC for cell replacement therapy, we have tested the cells in a mouse spinal cord injury model. First we developed a culture medium for in vitro expansion of the cells. We found that we can generate millions of EPI-NCSC, which do not undergo major changes in stem cell gene expression (Hu et al., unpublished) and resemble neural stem cells insofar as all in vitro expanded cells express glial fibrillary acidic protein (GFAP) and early neuronal markers at very low levels of [16].

Acutely grafted EPI-NCSC integrate into the spinal cord tissue and they survive in large numbers for extended periods of time both in the gray matter and the white matter. We have analyzed grafts for up to 6-months post-surgery and have observed that intraspinal EPI-NCSC undergo morphological changes by elaborating long processes. Subsets of grafted cells express markers for neurons, including GABAergic neurons, and for oligodendrocytes, but not astrocytes. Importantly EPI-CSC grafted into the adult spinal cord do not form tumors [16].

Contribution of EPI-NCSC to Skeletal Muscle

By nature, neural crest stem cells have aspects of pluripotency, as they have the physiological capability of generating not only ectoderm-derived cell types, but also cell types that are typical for mesoderm-derived tissues. One mesoderm-derived cell type is striated muscle. Interestingly, neural crest cells have the innate ability to generate striated muscle as they make small contributions to striated muscles of the eye [8]. We were thus interested in the behaviour of EPI-NCSC grafts in adult skeletal muscle. We injected 150,000 enhanced green fluorescent protein (EGFP)—expressing neural crest cells into the adult mouse biceps femoris muscle and analyzed the transplant 4 months later. We obtained three types of result. First, grafted EPI-NCSC can survive in adult muscle.

Second, entire muscle fibers became green-fluorescent (Fig. 1a–d). Green fluorescent EPI-NCSC were in intimate contact with green-fluorescent muscle fibers (Fig. 1b–d). This observation suggests that at least some green-fluorescent EPI-NCSC can fuse with adult striated muscle and that internalized nuclei express EGFP within the muscle fiber.
Fig. 1

EPI-NCSC contribute to adult skeletal muscle and persist as neurogenic cells within muscle. ad Grafted EGFP-expressing EPI-NCSC contribute to adult skeletal muscle 4-month post-grafting. a Phase contrast image of muscle section. b EGFP fluorescence of same section. Three muscle fibers are entirely green fluorescent, whereas surrounding muscle fibers are not. One grafted cell, recognizable by EGFP fluorescence, is in close physical contact with the muscle fiber (arrow, and insertb’). c DAPI nuclear stain of same section; arrow points to nucleus of same cells. d and d′, green-blue merged images. ef Neuron-specific beta-III tubulin immunoreactive grafted EPI-NCSC in adult skeletal muscle 4-months post-grafting (arrow). e EGFP fluorescence marking grafted EPI-NCSC (arrow); f Cy3 fluorescence of same cell stained with neuron specific beta-III tubulin; g red-blue merged images

Third, immunocytochemistry with antibodies against neuron-specific beta-III tubulin showed that a subset of grafted EPI-NCSC become neuronal within the muscle (Fig. 1e–g). Interestingly beta-III tubulin immunoreactive cells were not in close physical contact with muscle fibers, but were aligned between muscle fibers and they elaborated long processes. No muscle fibers were observed that expressed beta-III tubulin. The latter observation suggests that either neurogenic EPI-NCSC do not fuse with host muscle fibers, or, if they fuse, intramuscular EPI-NCSC-derived nuclei do not express beta-III tubulin. The former possibility, i.e. that neurogenic EPI-NCSC do not fuse with muscle fibers, is supported by our observation that beta-III tubulin-expressing cells were not observed in the vicinity of green-fluorescent muscle fibers. Before grafting, the cells had been expanded in culture. Our previous results indicated that the cell population is homogeneous with regard to neural stem cell-related gene expression [16], potentially explaining neurogenic cells within the graft. Nevertheless, this does not preclude the emergence of subsets of cells that differ by other, yet to determined molecular criteria. Alternatively, different areas within the muscle may differentially affect EPI-NCSC differentiation.

Taken together, the significance of these data is that cells in EPI-NCSC grafts not only can contribute to adult muscle, but that they can also generate neuronal cells that survive within adult muscle for extended periods of time. These observations may have implication for degenerative muscle disease, such as amyotrophic lateral sclerosis, and for peripheral neuropathies.

EPI-NCSC and iPS Cell-Producing Genes

Here we show that EPI-NCSC represent a physiological population of iPS-like stem cells that are, however, not tumorigenic. iPS cells are pluripotent stem cells that can be generated by retroviral delivery into somatic fibroblasts of four genes, Oct3/4 (Pou5f1), Sox2, c-Myc and Klf4. These cells were then selected according to Fbxo15 expression and found to have similarities to ESC, including teratoma formation. However, gene expression and gene methylation patterns differed from that of ESC. More recently, it has been shown that selection of iPS cells according to Nanog expression provides a population of cells from which one clone was transmitted through the germ line to the next generation of mice. Approximately 20% of offspring developed tumors, which was attributed to re-expression of c-myc [see e.g., 9, 17, 18, 19].

We noted that the genes that were delivered into somatic fibroblasts, as well as Fbxo15, are present in our EPI-NCSC longSAGE library (ref 3;, series number GSE4680). We therefore sought to substantiate the notion that EPI-NCSC physiologically express pertinent iPS cell genes. We performed real-time PCR (iCycler, BIO-RAD) from RNA of EPI-NCSC in day 3 primary explants as described previously [3]. Primer pairs were purchased from SuperArray; catalog numbers are listed in parentheses. Data are expressed as fold-increase or fold-decrease relative to expression levels in mouse ESC (ES-R1 line). Here we show that EPI-NCSC expressed Sox2 (PPM04762A), c-Myc (PPM02924E) and Klf4 (PPM25088A) at levels similar to those in ESC, whereas Oct-4 (Pou5f1; PPM04726A) was expressed at lower levels. Sox10 (PPM04723A) was used as a reference, as it is considered a neural crest marker. As expected, Sox10 was greater than 24 times more abundant in EPI-NCSC than in ESC (Fig. 2). Myc expression was approximately sixfold more abundant in EPI-NCSC than in ESC, whereas Klf4 was about twofold higher in EPI-NCS than in ESC. Sox2 expression was comparable to that of ESC, i.e. 1.8-fold lower in EPI-NCSC than in ESC. Lin28 (PPM39240A) was expressed as well, but at levels that were almost 30-times lower than in ESC. Large differences were observed in Oct-4 and Nanog (PPM25326A) expression. Oct-4 was about 500-times and Nanog 724-times less abundant in EPI-NCSC than in ESC (Fig. 2).
Fig. 2

iPS cell genes expressed in EPI-NCSC. Gene expression levels in EPI-NCSC and in ESC were determined by real-time-PCR. Data are expressed as fold increase or decrease compared to ESC. Sox10 was evaluated as a marker for neural crest cells. Sox10 was expressed at 24.3-fold higher abundance compared to ESC, as expected, because it is considered a neural crest stem cell marker. Myc was 6.3 times and Klf4 2.2 times higher expressed in EPI-NCSC than is ESC. Sox2 was 1.8-fold lower and Linf28 29.9-fold lower. Large differences were observed with regard to Oct-4 and Nanog expression relative to expression levels in ESC. Oct-4 was less abundant by 524-fold whereas Nanog was less abundant by 724-fold in EPI-NCSC compared to ESC, respectively


EPI-NCSC are a fascinating type of stem cell because they are multipotent somatic stem cells of embryonic origin that can be isolated as a highly pure population of stem cells by minimally-invasive procedures, and because they can be expanded in vitro into millions of stem cells/progenitors. In a mouse model of spinal cord injury, EPI-NCSC grafts show desirable traits, which include survival, neurogenesis, expression of oligodendrocyte markers and lack of tumorigenicity. Developmental potentials of neural crest cells are not limited to epidermal derivatives, but also include cell types that are typically generated by the mesoderm, including striated muscle, bone and cartilage. Here we showed that EPI-NCSC could contribute also to adult skeletal muscle. Furthermore, we showed that EPI-NCSC develop into neuronal cells when grafted into skeletal muscle and can persist for a prolonged period of time. Finally, we showed that EPI-NCSC partially share the pertinent stem cell gene expression pattern of iPS cells. An important difference between iPS cells and EPI-NCSC is that latter are not tumorigenic. Since Myc is expressed at higher levels in EPI-NCSC than in ESC cells, tumorigenicity cannot, or not exclusively, be a consequence of Myc expression. This is an important question that demands further investigation. It is of interest to note that we detected Oct-4 and Nanog in EPI-NCSC by real-time PCR, albeit at significanlty lower levels than in ESC. There is evidence that Oct-4 and/or Nanog are involved in tumor formation [see e.g., 2, 7, 10, 13]. Thus the low levels of Oct-4 and Nanog expression likely explain the fact that EPI-NCSC are not tumorigenic.


This work was supported by The Plunkett Family Foundation, Milwaukee, Wisconsin, USA, and the North East England Stem Cell Institute at Newcastle University, Newcastle upon Tyne, UK.

Copyright information

© Humana Press 2008