THE existence of stem cells in the nervous system has become a reality only in the recent years. An extensive body of work is currently devoted to understand the role of neural progenitors in constructing the embryonic brain and spinal cord as well as to characterize the neural stem cells (NSCs) that are present in discrete regions of the adult brain (1). Besides those stem cells in the central nervous system (CNS), other types of neural progenitors/stem cells have been discovered in the peripheral nervous system (PNS), which originate from the neural crest (NC) during embryogenesis. The progenitors and stem cells derived from the NC are, in all vertebrates, the source of a wide variety of cell types beyond the neural cells of the PNS. Multipotent NC cells (NCCs) have been identified at different stages of embryogenesis, and some persist in many adult organs and tissues, in animals and humans. We review here the current knowledge concerning the isolation, specific markers, and differentiation properties of the distinct NC stem cells (NCSCs) and discuss their potential role in development and repair.
A wide array of neural and non-neural cell types arises from the neural crest during vertebrate embryogenesis. The neural crest forms transiently in the dorsal neural primordium to yield migratory cells that will invade nearly all tissues and later, differentiate into bones and cartilages, vascular smooth muscle cells, connective tissues, neurons and glial cells of the peripheral nervous system, endocrine cells, and melanocytes. Due to the amazingly diversified array of cell types they generate, the neural crest cells represent an attractive model in the stem cell field. We review here in vivo and in vitro studies of individual cells, which led to the discovery and characterization of neural crest progenitors endowed with multipotency and stem cell properties. We also present an overview of the diverse types, marker expression, and locations of the neural crest-derived stem cells identified in the vertebrate body, with emphasis on those evidenced recently in mammalian adult tissues. © 2012 International Society for Advancement of Cytometry
THE NEURAL CREST CELLS, DEFINITION AND GENERAL PROPERTIES
The NC is an embryonic structure characteristic of the vertebrates, which is located at the margins of the neural primordium, forming an area intermediate between the presumptive CNS and epidermis (2). The induction of the early NC territory depends largely upon the coordinated action of a set of signaling molecules and transcription factors during gastrulation and neurulation stages. Current knowledge of the regulatory gene network involved in the formation and specification of the NC is reviewed elsewhere (3).
The NC per se is only transient, since its component cells rapidly delaminate from the neural epithelium and migrate extensively to colonize many embryonic tissues, where they eventually undergo differentiation. The emergence of migratory NC cells (NCCs) from the dorsal neuroepithelium involves an epithelial-to-mesenchymal transition (EMT) (4, 5). Recent cell imaging techniques led to addressing the migration behavior and motile properties of the NCCs in vivo (6, 7) and began to uncover the genetic determinants of the EMT in NCCs, which represents a model system for a key step in dissemination of tumor cells towards metastasis (8).
A plethora of neural and non-neural cell types originate from the NCCs (Table 1). The NC gives rise to the PNS neurons and their associated glial cells (in sympathetic, parasympathetic, enteric, and partly sensory ganglia) as well as to the Schwann cells lining the peripheral nerves (Fig. 1). The NCCs are at the origin of the pigment cells in the skin (melanocytes) and of some neuro-endocrine cells (in the carotid body, adrenal medulla, and thyroid gland) (2). In addition, the NCCs that populate the head and neck structures are the source of diverse mesenchymal cell types, which in the trunk instead derive from the mesodermal germ layer. The cranial NCCs thus contribute to craniofacial bones, cartilages, tendons, connective cells, and adipose tissue, and they yield brain pericytes and vascular smooth muscle cells of the aorticopulmonary cardiac septum (2, 5) (Table 1). All these NC derivatives could be identified in avian embryos and hatching birds; the natural and permanent quail nuclear marker used to follow the fate of NCCs after grafting quail NCCs into a chick host embryo (2). The NC origin of cranial mesenchymal tissues thus established in quail-chick chimeras (2) later was confirmed in mammals by genetic fate mapping in the mouse using the Cre-Lox system associated with early NC-specific gene promoters (e.g., Wnt1, P0, and Sox10) (11, 12).
|Neural crest-derived cell types|
|Neurons and glial cells||Pigment cells||Endocrine cells||Mesenchymal cells|
|Cranial||PNS neurons||Skin melanocytes||Carotid body cells||Craniofacial osteocytes and chondrocytes|
|Sensory cranial ganglia, parasympathetic (ciliary) and enteric ganglia||Innear ear pigment cells||C cells (thyroid)||Smooth muscle cells (vascular and heart conotruncus)|
|Satellite glial cells in PNS ganglia||Adipocytes|
|Schwann cells||Dermal cells|
|Ensheating olfactory cells||Corneal cells (endothelium and stroma)|
|Trunk||PNS neurons||Skin melanocytes||Adrenal-medullary cells||Endoneurial fibroblasts (mouse sciatic nerve)|
|Sensory ganglia (DRG), sympathetic and parasympathetic ganglia|
|Satellite glial cells in PNS ganglia|
Thus, the NC is a highly pluripotent structure which, in this respect, can be compared to the hematopoïetic system generating the large diversity of blood cells. How do the immature NCCs become phenotypically diversified? Do all the NC-derived cell types arise from one single type of multipotent stem cell or is the premigratory NC a heterogenous collection of distinct committed cells? These issues have attracted developmental biologists for long and begun to be answered since methods were devised to trace the fate of single cells.
MULTIPOTENT PROGENITORS IN THE EARLY NEURAL CREST OF AVIAN AND RODENT EMBRYOS
In vivo lineage tracing studies were performed to monitor the fate of early NCCs in the avian embryo from the beginning of their migration from the neural tube. These studies involved intracellular microinjection of fluorescent dextran into single NCCs located in the dorsal neural primordium at the premigratory stage, after which the progeny of labeled cells could be identified in NC derivatives of the developing embryo (13, 14). Analysis of the injected embryos showed that the clonal descendants of trunk NCCs populate different types of NC derivatives producing after 2 days several combinations of neurons (in dorsal root ganglia (DRG) and/or sympathetic ganglia), nerve glial cells, melanocytes, and adrenomedulary cells. Similar dye labeling of the dorsal neural tube cells/premigratory NCCs in other species also revealed the presence of multipotent progenitors in the early NC in vivo. However, recent data supported early determination of premigratory avian trunk NCCs (15). Evaluating the extent and temporal changes of the potentials of early NCCs in vivo will therefore need further investigations and new methods to permanently trace the fate of NCCs.
Following colony-forming unit assays developed to identify progenitors at the origin of hematopoietic cell diversity, a number of in vitro cloning experiments have permitted to analyze the repertoire of phenotypes that a single NCC is able to produce. These experiments, which require previous isolation of the NCCs as a pure cell population, were initially devised by Cohen and Konigsberg (1975) (16), who found that quail NCCs migrate from embryonic neural tube explants cultured in vitro; after removal of the intact neural epithelium explant, the NCCs can be harvested for subsequent replating and clonal assays (Fig. 2). This method was a mean to first demonstrate that trunk NCCs are heterogeneous in their potential and can give rise to unpigmented, pigmented or mixed progeny (16, 17). This isolation technique for early NCCs is still the basis of methodologies recently used to obtain avian and rodent NCSCs in development (see below). Since this pioneer work, improvements in culture media and availability of an increasing number of Abs against cell type-specific markers, allowed more extensive characterization of NC progenitors, isolated at different stages and from different rostro-caudal levels of the avian NC (5, 9). For example, it was shown that unpigmented and mixed (with unpigmented and pigmented cells) clones derived from avian trunk NCCs contained several hundred of sensory and adrenergic neuroblasts (18).
Altogether, data supported heterogeneity of the NCCs regarding their proliferation and differentiation potentials. Most of them generated different combinations of two or more cell types (19–22). The analysis of colonies resulted in a hierarchical model for cephalic and trunk NC lineage diversification (5, 10). The cephalic NC model includes recent results showing the existence of multipotent precursors capable to yield cell types belonging to all the major NC lineages, that is, neural, melanocytic, and mesenchymal (23, 24). Even if they do not contribute to the formation of mesenchymal tissues in vivo, avian trunk NCCs are able to differentiate into chondrocytes in vitro (23).
Using technical approaches similar to that devised for the avian NC, clonal assays were performed with the NCCs of rats and mice (25, 26). Stemple and Anderson (1992) were the first to identify mammalian multipotent NCCs in the migratory outgrowth of isolated trunk neural primordium (25). These NCSCs were isolated by FACS through the high expression of neurotrophin receptor p75 in the cell outgrowth produced in vitro by neural tubes from E10 rats and E9 mice (25). Rodent NCSCs generated glia, autonomic neurons, and myofibroblasts, and were able to self-renew when submitted to in vitro serial subcloning (25). In the avian NC, bipotent progenitors that differentiate into glia and melanocytes or into glia and myofibroblasts were also identified as self-renewing stem cells (22).
Efforts have been made to characterize the differentiation and stem properties of early NCSCs (12). While signaling by Wnt and Bone morphogenetic proteins (BMP) promote neurogenesis along sensory and autonomic phenotypes, respectively, together these two pathways play a key role in the self-renewal and the maintenance of multipotency in mammalian trunk NCSCs (27). The transcription factor Sox10, expressed by premigratory and migratory NCCs (Fig. 1), functions in the maintenance of multipotentality and inhibits neuronal differentiation in the early NCCs both in vitro and in vivo (28). The differentiation of multiple NC lineages also depends upon the activity of FoxD3 regulatory gene. After FoxD3 gene ablation in the NC, mice die perinatally with catastrophic loss of NC-derived craniofacial structures, nerves, DRG, and enteric NCCs (29).
NEURAL CREST STEM CELLS AT POSTMIGRATORY DEVELOPMENTAL STAGES
After leaving the neural primordium, the early-migratory NCCs are modified during migration in response to the distinct environmental cues they received during their homing to their final sites of differentiation. Although some become rapidly specified, the migratory NCCs “en” route to, or arriving at, their embryonic target tissues still comprise a subpopulation of progenitors that retain multiple differentiation options (30). Because migrating NCCs are mixed with cells of the surrounding tissues, their identification was greatly improved through the use of markers combined with flow cytometry. Multifated progenitors thus could be evidenced in the cranial NCCs colonizing the branchial arches during early jaw development in E10.5 Wnt1-Cre/R26R mice. After addition of a β-galactosidase substrate, fluorescent NCCs were isolated and tested in clonal culture assays (31). Multipotent NCCs transiently located in the mesentery were also isolated to expression of the tyrosine-kinase receptor Ret, when they are homing to the gut (32).
Anderson and coworkers were the first to demonstrate the persistence of NCSCs in NC derivatives, isolating subpopulations of multipotent NCSCs in the peripheral nerve, DRG, and gut of E14.5 rat embryos (Table 2). FACS experiments led to identify NCSCs in the embryonic rat sciatic nerve, which were positive for p75 and negative for myelin protein P0 (33). These nerve NCSCs exhibited trilineage (neuronal, glial, and smooth muscular) potential in vitro, similar to early NCSCs. Further, they displayed distinct responsiveness to BMP2 and gave rise to few neuronal types, when carefully compared to the NCSCs of the rat early NC (59). Likewise, enriched NCSC populations could be sorted from the rat gut at fetal or postnatal stages, by using p75/α4-integrin or p75 alone, respectively (34, 35) (Fig. 3).
|Sciatic nerve||Rat fetal||N.d.||NCSCs||(p75+P0−), α4||(33, 34), Fig. 3|
|Gut||Rat fetal||N.d.||NCSCs||(p75+α4+)||(34), Fig. 3|
|DRG||Rat fetal||N.d.||NCSCs||p75, P0, PMP22||(35, 36)|
|Mouse adult||Sox10, Nestin, p75||(37, 38)|
|Bone marrow||Mouse adult||N.d.||BM-NCSCs||p75, Sox10, Slug, Snail||(38)|
|Carotid body||Mouse adult||N.d.||NCSCs||(HNK1− p75−) (GFAP+)||(39)|
|Cornea||Mouse adult||Corneal stroma||COPs||Sca1, CD34, Sox9||(40)|
|Juvenile||Corneal limbus||MCCs||Twist, Snail||(41)|
|Iris||Mouse adult||Iris stroma||NCSCs||p75, Sox10||(42)|
|Heart||Mouse adult||Myocardium, valves||NCSCs||Nestin, Musashi||(43, 44), Fig.4|
|Palate||Rat adult||Palatal ridges||pNCSCs||p75, Nestin||(45)|
|Oral mucosa||Human||Lamina propria||OMLP-SCs||CD44, CD90, CD105, CD106||(45)|
|OMSCs||CD29, CD73, CD166||(46)|
|Skin||Mouse adult||Face and trunk||NCSCs||p75, Sox10||(38, 47), Fig. 5|
|Facial dermis||SKPs||Nestin||(48, 49)|
|Whisker follicle||Epi-NCSCs||Nestin, Sox10||(50)|
|Human neonatal, adult||HF dermal papillae||SKPs||p75, Sox10||(51)|
|HF bulge (scalp)||HFSCs||Nestin, Sox9||(52)|
|HF bulge||Epi-NCSCs||Nestin, Sox10||(53)|
|Teeth||Mouse adult||Dental pulp||DPSCs||CD44, CD73, p75||(54)|
|Human||Dental pulp||DPSCs, SHED||Nestin, p75||(55, 56)|
|Periodontal ligament||PDLSCs||Nestin, SSEA4||(57)|
NEURAL CREST STEM CELLS INADULT TISSUES
Amazingly, even adult NC-derived organs and other adult tissues appeared to contain some stem cells of NC origin, in addition to their expected differentiated cell types. As shown in Table 2, NCSCs were found in diverse tissue locations in several mammalian species. The NC-origin of these adult stem cells could be proven in transgenic mice using the Cre/Lox system and NC-specific promoters driving expression of reporter genes like green fluorescent protein (GFP), resulting in permanent labeling of the progeny of early NCCs. NC-derived mutipotent progenitors thus persist in adult rodent DRG and bone marrow (37, 38), gut (35), carotid body (39), heart (43, 44) (Fig. 4), and several craniofacial tissues including the cornea (40, 41), iris (42), dental pulp (54), hard palate (45), and oral mucosa (46, 60) (Table 2). After isolation and in vitro culture, these NC-derived progenitors differentiate into classical NC cell lineages like neurons, glial cells, and myofibroblasts, and in some cases, also into mesenchymal cell types.
The function of these NCSC-like cells in vivo remains elusive, except for rodent NCSCs of the carotid body, which offer a unique example clearly demonstrating a physiological role of adult NC progenitors in vivo (39). During adaptation to hypoxia, these cells undergo proliferation and eventually lead to de novo differentiation of dopaminergic chemosensory neurons that activate the brainstem respiratory center to increase ventilation. The carotid body is therefore a niche for adult neurogenic stem cells in the PNS. These stem cells could be isolated by flow cytometry through the expression of GFAP (glial fibrillary acidic protein) and the absence of p75 and human natural killer-1 (HNK1) antigens (39) (Table 2).
The dental pulp is formed by the cranial NC and contains NC-derived precursors that differentiate into odontoblasts and are responsible for dentinal repair in adults. Shi and coworkers (55–57) have isolated highly proliferative clonogenic stem cells in humans from the dental pulp (of adult molar and exfoliated deciduous teeth) and from the periodontal ligament (Table 2). These progenitors can be expanded ex vivo and reconstitute dentin and bone in transplantation models, thus offering promise for mesenchymal tissue engineering and tooth regeneration (61).
The adult skin is another source where NC-derived stem cells can be isolated with minimally invasive procedures. The multipotent NCC populations recently identified in the murine skin are located in several epidermal (47, 50) and facial dermal structures of the hair follicle (48, 49) (Table 2) (Fig. 5). These multipotent undifferentiated NCCs differ from the NC-derived melanocyte stem cell identified in the permanent lower portion (bulge) of the hair follicle, which replenishes melanocytes during the hair cycle, expresses early markers of melanogenesis and has restricted developmental potential (62). NCSCs isolated from the hair bulge do not contribute to the epidermal epithelial cells and thus are different from the epidermal stem cells that ensure continuous renewal of skin keratinocytes (63). In the back skin, trunk NC-derived multipotent cells are present in the hair bulge and share early markers of glial and melanocytic cell lineages (47). These NCSCs isolated in rodent and human skin display neural and mesenchymal differentiation properties and can be long-term maintained in vitro.
Expanded skin-derived NC progenitors recently showed significant efficiency to repair nerve tissue when tested in murine models of spinal cord and nerve injury (63–67). The skeletogenic potential of human skin-derived NC progenitors has also been exploited to repair bone fracture after transplantation in mice (68). The NCSCs of the adult skin are accessible and share some gene profiling characteristics with pluripotent stem cells without being tumorigenic after transplantation in nude mice (69, 70). They constitute a potentially attractive source of adult somatic stem cells for autologous transplantations and future cell replacement therapy in a variety of human tissues.
TOWARDS THE CHARACTERIZATION OF NEURAL CREST PROGENITORS IN HUMANS
The NC is of particular clinical significance, as its aberrant development results in a large group of human pathologies collectively known as neurocristopathies (71). Theses NC-related pathologies encompass congenital heart defects (mainly in the outflow tract), Hirschsprung disease (with abnormal innervation of the posterior gut) and Waardenburg syndromes, labio-palatine clefting and other craniofacial malformations, and several cancers such as melanoma, pheochromocytoma, and neuroblastoma. It is thus of crucial importance to develop experimental approaches suitable for the dissection of the cellular and genetic features of the human NCCs (hNCCs).
Compared to extensive investigations on NCCs in many animal models, almost nothing is known about the development of hNCCs. Histological analysis and spatio-temporal expression patterns of early NC markers (72) indicated that the premigratory/migratory hNCCs develop toward the end of the first month of pregnancy. Recently, hNCC lines have been derived from the NCCs that emigrated in vitro from human embryonic neural tube cultures (73) (Fig. 6). Long-SAGE transcriptome profiling revealed that hNCCs exhibit a NC-specific gene signature (including FOXD3, MSX1, SNAI2, SOX9, SOX10, and TWIST) while sharing expression of many stem cell genes with human embryonic stem cells (hESCs) like SOX2, NANOG, LIN28, and OCT4 (73). These hNCCs could be maintained and expanded in vitro for months, thus offering an extended source for further characterization of hNCCs; however, the in vitro generation of NC derivatives by these cell lines was so far unsuccessful.
To bypass the difficulties linked to working on the human embryo, in vitro strategies have been developed in the recent years to derive NCCs from human pluripotent cells (Table 3). Populations of hNCCs have thus been derived from neuralized human embryonic stem cells (hESCs) according to various induction protocols that produce neural cells of both CNS and NC types in variable amounts (Table 3). Most of these protocols involved the use of stromal feeder-layers (74, 75) and, more recently, the technique of neural rosettes developed by Studer and coworkers (77). These ESC-derived neuroepithelial structures can mimic the embryonic neuroepithelium in vitro and yield both NCCs and CNS progenitors (77). These procedures, which are quite long and of low efficiency, were combined with flow cytometry to get enrichment for hNCCs using CD57 and CD271 (76, 77). Expression of Cadherin-11 and Frizzled-3 was also exploited by Zhou and Snead (2008) to sort cranial NC-like progenitors from preparations of human embryoid bodies (81).
|Stem cell source||Derivation protocol||Surface markers||References|
|hESCs||Stromal feeder-layers||(74, 75)|
|Neural rosettes||p75/CD271, HNK1 /CD57||(77)|
|Dual Smad inhibitors||(79)|
|Embryoid bodies||Cadherin-11 Frizzled-3||(81)|
|hiPSCs||Fibroblast reprogramming||p75/CD271, HNK1 /CD57||(73, 79, 82)|
Recently, methods could be devised for deriving NCCs from hESCs under defined culture medium conditions (79, 80). One of these culture conditions is based on the action of two inhibitors of Smad, which promote NC fate in human neural rosettes by counteracting signaling by BMPs and transforming growth factors (TGFs)/activin (79). Another approach uses the generation of neurospheres in defined medium supplemented with neural inducers [fibroblast growth factor 2 (FGF2) and epidermal growth factor (EGF)]. In this assay, NCCs are produced by migration from adherent neurospheres in less than 10 days (80).
The availability of unlimited numbers of hNCCs in fully defined conditions may offer a valuable tool for investigating features of human NC development and for modeling NC-related diseases. Further, the role of the chromatin remodeler CHD7 in the generation of premigratory hNCCs and CHARGE syndrome was recently described using neurosphere-mediated derivation of hNCCs from hESCs (84). Using similar approaches, Cimadamore et al. (85) defined the role of the transcription factor SOX2 in hNCC migration and sensory neurogenesis.
Strategies to model NC-related disorders are likely to emerge also from experiments using human induced pluripotent stem cells (hiPSCs) for the production of early NCCs and their various differentiated derivatives (82, 83).
A remarkable diversity of vertebrate cell types arises from the NC in the developing embryo. The NCCs are very invasive; they not only give rise to multiple NC derivatives but also colonize the whole body when they differentiate into the Schwann cells accompanying all the peripheral nerves or into the pigment cells that are present on the entire surface of the skin. Because NCCs contribute to many mesenchymal tissues in the head, it was proposed that the NC might be a fourth germ layer, which produces cell types belonging to derivatives of both mesoderm and ectoderm (86).
Several lines of evidence demonstrated the multipotency and self-renewal capacity of various subsets of NC progenitors in avian and mammalian embryonic development. A significant advance towards deciphering the lineage mechanisms in the early NC was to show that the neural-melanocytic and mesenchymal NC derivatives could originate in vitro from common multipotent progenitors in the cephalic NC. These conclusions were mostly based upon the analysis of the developmental repertoire that NCCs displayed in vitro after single cell culture in adherent or sphere-forming conditions.
The characterization of NCSCs is still in its infancy. Although several genes are known expressed by early NCCs, definitive NCSC-specific markers are lacking, which compromises the identification and isolation of NCSCs. However, a few surface antigens could be successfully used in FACS experiments to prospectively isolate multipotent NC progenitors from the surrounding cells at NC migratory and postmigratory stages during development. The identification of NCSCs is even more challenging in the adult, where large subsets of NC-derived cells have invaded diverse tissues and mingled with many other cell types of non-NC origin. Genetic fate mapping using the Cre/Lox system combined with LacZ or GFP reporters provided a mean to mark NC-derived cells permanently in the mouse. This method thus offered the possibility to identify and sort NCCs from a variety of tissues and organs in order to investigate their stem cell properties in vitro.
The growing evidence for NC-related multipotent cells located in easily accessible adult body sites like skin and teeth attracted a recent interest in NCSCs potential for cell replacement therapy. Much remains, however, to be known with respect to the molecular regulation of stem cell properties and differentiation of the NC progenitors present in diverse tissues. Developmental (and therapeutic) potentials of adult NCSCs are presumably influenced both by intrinsic characteristics imposed by origin and by spatiotemporal environmental cues. These issues are of clinical importance to understand the alterations in NC cell number and function that occur in human neurocristopathies.
We thank Dr. Etchevers (INSERM) for providing Figure 6.