A Highly Enriched Niche of Precursor Cells with Neuronal and Glial Potential Within the Hair Follicle Dermal Papilla of Adult Skin

Authors


Abstract

Skin-derived precursor cells (SKPs) are multipotent neural crest-related stem cells that grow as self-renewing spheres and are capable of generating neurons and myelinating glial cells. SKPs are of clinical interest because they are accessible and potentially autologous. However, although spheres can be readily isolated from embryonic and neonatal skin, SKP frequency falls away sharply in adulthood, and primary sphere generation from adult human skin is more problematic. In addition, the culture-initiating cell population is undefined and heterogeneous, limiting experimental studies addressing important aspects of these cells such as the behavior of endogenous precursors in vivo and the molecular mechanisms of neural generation. Using a combined fate-mapping and microdissection approach, we identified and characterized a highly enriched niche of neural crest-derived sphere-forming cells within the dermal papilla of the hair follicle of adult skin. We demonstrated that the dermal papilla of the rodent vibrissal follicle is 1,000-fold enriched for sphere-forming neural crest-derived cells compared with whole facial skin. These “papillaspheres” share a phenotypic and developmental profile similar to that of SKPs, can be readily expanded in vitro, and are able to generate both neuronal and glial cells in response to appropriate cues. We demonstrate that papillaspheres can be efficiently generated and expanded from adult human facial skin by microdissection of a single hair follicle. This strategy of targeting a highly enriched niche of sphere-forming cells provides a novel and efficient method for generating neuronal and glial cells from an accessible adult somatic source that is both defined and minimally invasive.

Disclosure of potential conflicts of interest is found at the end of this article.

Introduction

There is a clinical need for accessible adult sources of precursor/stem cells for the repair of diseases of the nervous system. Skin-derived precursor cells (SKPs) are precursor cells found in rodent and human skin; these cells are capable of generating glia and neurons [1, [2], [3]–4]. SKPs grow as self-renewing spheres, express nestin, and display a number of similarities to embryonic neural crest stem cells [2]. Both SKPs and Schwann cells generated from SKPs are capable of remyelinating central and peripheral demyelinated axons [5], making them a potentially attractive autologous cell source for treatment of demyelinating diseases and spinal cord injury.

However, although these cells are abundant and isolation of them is straightforward during embryogenesis, they are more difficult to isolate from adult skin [2, 6]. This is a particular problem with human skin [7]; although SKPs can be readily generated from human neonatal foreskin [4], the reliable generation of significant numbers of spheres from adult human skin is more problematic [7]. This difficulty in reliably generating large numbers of SKPs from adult human skin hinders biological study and thus clinical translation of these therapeutically promising cells. It is also important to define the minimum volume of adult skin required to reliably generate sufficient numbers of cells for characterization and application to allow future tissue sample collection to be as minimally invasive as possible. Furthermore, there is a need for a defined starting population to address the molecular mechanisms underlying the generation of precursors and their neural progeny, as well as to examine the endogenous role of these cells. SKP culture is heterogeneous, being derived from the whole dermis, and the identity of the culture-initiating cells is not fully understood [3, 8]. In addition, recent work has demonstrated that spheres derived from different anatomical locations display subtle but important differences of origin and potential [8]. Therefore, the ability to generate significant numbers of SKPs from a defined skin region would be of considerable value.

One possible niche for SKPs in the adult is the dermal papilla of the hair follicle [2]. The dermal papilla is a discrete condensation of specialized dermal fibroblasts situated at the base of the hair follicle [9]. The dermal papilla in vivo plays a central role in hair follicle morphogenesis [10, 11] and mesenchymal-epithelial interactions [12]. Evidence from both in vitro and in vivo studies further suggests that dermal papilla derivatives exhibit a significant degree of plasticity with both mesenchymal stem cell-like and hematopoietic stem cell-like activity [13, 14]. There is increasing evidence that the dermal papilla of the facial hair follicle is derived from the neural crest [2, 8, 15, 16].

In this study, we examined whether the dermal papilla represents an enriched niche for multipotent skin-derived precursors of neural crest origin. Using a combined approach of fate-mapping and microdissection, we quantified the high degree of enrichment of the dermal papilla for neural crest-derived sphere-forming cells. We demonstrated that multipotent papillaspheres can be efficiently generated from adult human skin by targeted microdissection of the dermal papilla niche.

Materials and Methods

Cell Culture: Rodent Skin-Derived Precursors and Papillaspheres

Skin-derived precursor cells were generated from the skin of Sprague-Dawley rats and Wnt1-Cre/R26R mice according to published protocols [1, [2], [3]–4]. All experiments with animal tissue were performed in accordance with U.K. guidelines on the care and use of experimental animals. Whole skin was dissected from the face, washed in Hanks' balanced saline solution (HBSS), cleaned, and cut into 1–2-mm2 pieces and incubated in 0.05%–0.25% trypsin/EDTA for 20–40 minutes. Following the addition of serum, the skin pieces were mechanically dissociated to release single cells and filtered through a 40-μm mesh (BD Biosciences, San Diego, http://www.bdbiosciences.com). The cells were collected, washed, and resuspended in Dulbecco's modified Eagle's medium (DMEM; Gibco, Grand Island, NY, http://www.invitrogen.com) and Ham's F-12 medium (F12) supplement (Gibco) in a 3:1 ratio with 2% B27 (Invitrogen, Carlsbad, CA, http://www.invitrogen.com), 1% penicillin/streptomycin/Fungizone (Bristol-Myers Squibb, New York, http://www.bms.com), 20 ng/ml epidermal growth factor (EGF; R&D Systems Inc., Minneapolis, http://www.rndsystems.com), and 40 ng/ml fibroblast growth factor 2 (FGF2; R&D Systems), hereafter referred to as SKP proliferation medium. Fresh proliferation medium was added to cultures every 4–5 days, and spheres were passaged by mechanical dissociation with a flame-polished glass pipette. For standardized sphere assays of dissociated whole skin at developmental ages ranging from embryonic day 12 to adult, freshly dissociated skin cells were seeded at a density of 25,000 cells per milliliter in 10 ml of SKP proliferation medium in a 25 cm2 tissue culture flask. Sphere numbers were determined at day 14 by counting under a low-power phase microscope. For methylcellulose assays, dissociated skin cells were plated at the same cell density in SKP proliferation medium containing 1% methylcellulose (Methocult; StemCell Technologies, Vancouver, BC, Canada, http://www.stemcell.com).

Microdissection of the vibrissal follicle was performed according to previously described protocols [13]. The whisker pad of the rat was cut free, and the whisker hair follicles were plucked from the pad. The end bulbs of the rat whisker follicle, which contain the dermal papilla, were isolated and washed in HBSS. The epidermis and dermal sheath of the end bulb were inverted over the end of a pair of fine forceps, exposing the dermal papilla. The dermal papilla was then cut free (Fig. 2D). To generate single dermal papilla cells, the papillae were transferred to a small Petri dish and their inner surface exposed under a dissecting microscope using a fine needle. Papillae were then incubated in 0.1% trypsin for 30 minutes. The papillae were then mechanically dissociated with a Pasteur pipette to release single cells, which were counted under a low-power microscope. Fetal calf serum (FCS) was added to inhibit trypsin, and the cells were washed three times and resuspended in 10 ml of SKP proliferation medium at an ultralow density of 10 cells per milliliter. Papillaspheres appeared around day 7 and were passaged at 14–21 days by mechanical dissociation through a flame-polished Pasteur pipette followed by resuspension in SKP proliferation medium. The number of spheres were counted under a low-power microscope at day 14, and the sphere-forming efficiency was expressed as a ratio of spheres to initial single cell numbers. To assess the degree of enrichment of the dermal papilla for sphere-forming cells, identical ultralow-density cultures were obtained from dissociated whole whisker follicles or dissociated whole facial skin.

Wnt1-Cre Compound Transgenic Mice

Wnt1-Cre [17] and ROSA26R mice [18] were intercrossed to produce litters in which the cells derived from the neural crest stained blue following a reaction with 5-bromo-4-chloro-3-indolyl-β-d-galactoside (X-Gal) [19]. Tissue was harvested from mice of at least 4 weeks of age. Facial skin and whisker pad for immunohistochemical analysis were fixed in 4% paraformaldehyde, immersed in 30% sucrose, mounted in OCT embedding medium (Raymond A. Lamb Limited, East Sussex, England, http://www.ralamb.co.uk), frozen, and cut at 14 μm on a cryostat. β-Galactosidase immunohistochemistry is described below. Hair follicle microdissection, SKP, and papillasphere culture were performed as described above. For X-Gal staining, cells and vibrissal follicles were fixed with 0.05% glutaraldehyde and subsequently incubated with 1 mg/ml X-Gal (Imgenex, San Diego, http://www.imgenex.com), 35 mM potassium ferricyanide and 35 mM potassium ferrocyanide with 2 mM magnesium chloride for 4 hours.

Cell Culture: Human Skin-Derived Precursors and Papillaspheres

This study was approved by Addenbrooke's Hospital NHS Trust Local Research Ethics Committee (LREC 05/Q0108/146). Skin was donated by patients undergoing plastic surgery procedures. Beard and scalp samples were obtained from patients undergoing procedures that required the excision of facial skin. Patients donated surplus skin that would have otherwise been discarded as clinical waste. Full informed consent of each patient was gained. Human skin-derived precursor cultures were generated from whole adult human skin samples using previously described methods [3, 7]. Briefly, whole skin samples were washed and cut into small 1–2-mm pieces using a scalpel blade and incubated in 0.25% trypsin/EDTA for 40–60 minutes. Following the addition of 10% FCS, the skin pieces were mechanically dissociated to release single cells and filtered through a 40-μm mesh (BD Biosciences). The cells were collected, washed, and resuspended in SKP proliferation medium described above and seeded at 25,000 cells per milliliter in 10 ml of medium in a T25 flask. Fresh medium was added every 4–5 days. Spheres were passaged by incubation with 0.1% trypsin for 5 minutes followed by mechanical dissociation with a flame-polished Pasteur pipette and resuspended in SKP proliferation medium. Numbers of spheres were counted at 14 days, and sphere expansion was assessed at 1 month. The dermal papilla of facial/scalp hair follicle was microdissected using a procedure similar to that described above, with minor modifications [20]. The skin sample was transected at the level of the dermis to allow easy separation of the hair follicles from the rest of the dermis. The most proximal section of the end bulb of the hair follicle was cut away to reveal the dermal papilla, which was cut away with fine needles and transferred to a clean culture dish. Single dermal papilla cells were generated by allowing cells to grow out of the explanted papilla on a plastic tissue culture dish in medium containing DMEM and 10% FCS for 4–10 days. When single dermal papilla cells were identifiable, the cells were removed from the tissue culture plastic using 0.1% trypsin/EDTA, counted, and resuspended in 10 ml of SKP proliferation medium at a density of 10 cells per milliliter. These cells grew as floating cultures. Spheres appeared around day 7–10 and were passaged and counted as described above.

Directed Differentiation of Papillaspheres

Neuronal and Glial Differentiation: Rat Papillaspheres.

After initial experiments to optimize differentiation conditions, neuronal differentiation was induced by plating spheres in SKP proliferation medium and 10% FCS overnight on glass coverslips coated with poly-d-lysine and laminin. Cells were subsequently differentiated in neuronal differentiation medium. This consisted of a basal medium of DMEM/F12 (3:1 ratio) and 1% FCS. To the basal medium, 10 ng/ml brain-derived neurotrophic factor (BDNF; R&D Systems), 10 ng/ml Neurotrophin-3 (NT3) NT3 (Peprotech, Rocky Hill, NJ, http://www.peprotech.com), 10 ng/ml nerve growth factor (NGF), and 40 ng/ml FGF2 (R&D Systems) plus 6 ng/ml retinoic acid (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com) were added. Parallel papillasphere cultures were differentiated in DMEM/F12 and 1% FCS alone. Differentiation experiments were performed across three different papillasphere cultures.

For Schwann cell differentiation, papillaspheres were plated on glass coverslips coated with poly-d-lysine and laminin. Spheres were adhered overnight in SKP proliferation medium with the addition of 10% FCS. Following adherence and outgrowth of cells, the medium was replaced with Schwann cell differentiation medium, which consisted of a basal medium of DMEM/F12 (3:1 ratio), 2% B27, 1% FCS. To this basal medium, 4 μM forskolin (Sigma) and 50 ng/ml neuregulin-1β (R&D Systems) were added. Cells were differentiated for 10–21 days, with 50% replacement of differentiation medium every 4 days. Parallel cultures were differentiated in basal medium alone.

Neuronal and Glial Differentiation: Human Papillaspheres.

Differentiation of human papillaspheres was performed using protocols adapted from rodent differentiation protocols described above. For Schwann cell differentiation, human papillaspheres were plated on poly-d-lysine/laminin-coated coverslips in SKP proliferation medium and 10% FCS with the further addition of 50 ng/ml neuregulin-1β and 4 μM forskolin after 4–5 days. For directed neuronal differentiation, papillaspheres were initially plated in DMEM/F12 and 20% FCS until confluence, at which point the medium was changed to the neuronal differentiation medium previously described for 14–21 days. Differentiation medium was replaced every 4 days.

Immunohisto- and Immunocytochemical Analysis

Cells were fixed in fresh 4% paraformaldehyde for 15 minutes and washed in phosphate-buffered saline (PBS). Cells were permeabilized and blocked for 1 hour with PBS, 0.1%–0.3% Triton X-100 and 5% normal goat serum. Primary antibodies were diluted in PBS and incubated overnight at 4°C. Primary antibodies used were as follows: fibronectin (rabbit polyclonal, 1:800; Sigma-Aldrich), vimentin (mouse monoclonal IgG1, MAB1681; Chemicon, Temecula, CA, http://www.chemicon.com), versican (mouse IgG1, 1:4 supernatant; kind gift Dr. Richard Asher), α-smooth muscle actin (mouse IgG2a, 1:200, clone 1A4; Sigma-Aldrich), nestin (rat: mouse IgG1, 1:200, MAB353; human: mouse IgG, 1:500; Chemicon), musashi (rabbit polyclonal, 1:200, AB5977; Chemicon), βIII-tubulin (mouse IgG2b, 1:500; Sigma-Aldrich), microtubule-associated protein (MAP2ab) (1:200; Sigma-Aldrich), glial fibrillary acidic protein (GFAP; 1:250; Dako, Glostrup, Denmark, http://www.dako.com), S100 (rabbit polyclonal, 1:200; Dako), p75 (mouse IgG1monoclonal, 1:50, MAB365R; Chemicon), CNPase (mouse IgG1 monoclonal, 1:100, C5922; Sigma-Aldrich), fatty acid-binding protein (FABP; goat, 1:10, SC006; R&D Systems), β-galactosidase (rabbit polyclonal, 1:100, AB1211; Chemicon), Sox10 (monoclonal mouse IgG1, 1:50, MAB2864; R&D Systems), and Pax3 (goat IgG, 1:100, AF2457; R&D Systems). Secondary antibody (Alexa Fluor 1:1,000; Invitrogen) was applied for 1 hour at 37°C in PBS/Hoechst (1:5,000). Cells were viewed under a Leitz microscope with appropriate filters for cell identification and counting. For immunohistochemical analysis of the rat vibrissal dermal papilla, regions of the whisker pad containing vibrissal follicles were fixed in 4% paraformaldehyde for 2 hours and then immersed in 30% sucrose for 48 hours and embedded in OCT compound (Raymond Lamb) Blocks were sectioned at 14-μm intervals using a Leica cryostat (Heerbrugg, Switzerland, http://www.leica.com). Schwann cell cultures derived from juvenile sciatic nerve and cortical neuronal cultures were established from rat according to standard culture protocols as immunocytochemically and morphologically positive controls. Quantification of β3-tubulin and S100-positive cells was performed using a Leitz microscope. Ten consecutive randomly chosen fields from each experiment were selected, and the number of cells expressing β3-tubulin or S100 together with an appropriate neuronal or Schwann cell morphology was expressed as a percentage of total cells, determined by Hoechst nuclear stain counts. Quantification of both neuronal and Schwann cell differentiation was performed across three cell lines, in duplicate. The proportion of positive cells was expressed as a mean ± SE. Student's t test and Prism 2.0 software (Prism, Alexandria, VA, http://www.prismstandard.org/) were used for statistical analysis.

Reverse Transcription Polymerase Chain Reaction

RNA was extracted from rat vibrissal papillaspheres using Trizol (Invitrogen). cDNA was synthesized from 2.5 μg of RNA using Moloney murine leukemia virus reverse transcriptase (Invitrogen) and oligo(dT) primers according to the manufacturer's instructions. Polymerase chain reaction (PCR) was carried out using Taq polymerase (Invitrogen). PCR products were separated on a 1% agarose gel and visualized with SYBR Green (Invitrogen). For all PCRs, a positive neural crest control (cDNA from rat embryonic day 12 neural tube) was used. PCR primers used were as follows: Pax3, forward, ggaggcggatctagaaaggaagga; reverse, cccccggaatgagatggttgaa; Slug: forward, aaccagagatcctcacctcagg; reverse, ttgcagacacaaggcaacg; melting temperature, 58°C; size, 433 base pairs.

Results

Influence of Developmental Stage on Frequency of Skin-Derived Precursor Cells

To examine the influence of developmental stage on SKP frequency, we cultured rat SKPs at varying developmental time points from embryonic day 12 through to adulthood, using a standardized sphere-forming assay. We were able to generate nestin-positive spheres from dissociated whole skin between embryonic day 12 through to adulthood (Fig. 1A). SKP frequency peaks at embryonic day 16 (1.06% ± 0.11% of cells; n = 6) and falls away sharply in the postnatal period (0.21% ± 0.08% of cells; n = 4), reducing further in adulthood (0.015% ± 0.005% of cells; n = 8). This substantial and significant reduction in SKP frequency in adulthood (p < .001, t test) was confirmed using methylcellulose-containing medium to prevent aggregation of cells.

Figure Figure 1..

Identification of a niche of neural crest-derived cells in the hair follicle DP of adult skin. (A): The sphere-forming potential of dissociated rat facial skin is dependent on developmental stage and falls away sharply in adulthood, with a 70-fold reduction in percentage of dissociated adult skin cells forming spheres compared with the peak observed at E16 (p < .001). (B): All adult facial skin spheres derived from Wnt1-Cre/R26R neural crest reporter mice expressed β-Gal as determined by XGal reaction (blue) compared with wild-type littermates, suggesting a neural crest origin for these cells. (C): Within adult facial skin, cells expressing β-Gal are concentrated within hair follicles (13.5% ± 1.36%). Within the hair follicle, the DP is highly enriched for β-Gal-positive cells (88.9% ± 4.6%, whisker follicles; 90.9% ± 3.5%, pelage follicles). (D–F): Microdissection of the whisker follicle of the Wnt1-Cre/R26R mouse revealed that the DP (black arrow) displayed an intense XGal reaction compared with adjacent hair follicle structures. (G, H): Immunohistochemical analysis (β-Galactosidase, red, Hoescht, blue) of both pelage and whisker hair follicles from the Wnt1-Cre/R26R mouse detects strong β-Gal activity in facial and whisker DP cells (white arrows). Scale bar = 10 μm (G), 100 μm (H). Abbreviations: B-Gal, β-galactosidase; DP, dermal papilla; dp, dermal papilla; E, embryonic day; XGal, 5-bromo-4-chloro-3- indolyl-β-d-galactoside.

Neural Crest Origins of Skin-Derived Precursor Cells

To investigate the developmental origin of SKPs, we generated spheres from the facial skin of Wnt1-Cre/R26R neural crest reporter mice. In this compound transgenic mouse, cells of neural crest origin that transiently express Wnt1 early in development are permanently labeled with a β-galactosidase reporter, together with their progeny [17, [18]–19]. All SKPs across three separate experiments expressed β-galactosidase, as detected by intense blue X-Gal reaction, compared with spheres derived from wild-type littermates (Fig. 1B). These results are consistent with a neural crest origin of SKPs.

Identification of a Niche of Neural Crest-Derived Sphere-Forming Cells in the Hair Follicle Dermal Papilla of Adult Skin

We analyzed facial skin specimens derived from Wnt1-Cre R26R mice to identify discrete skin regions that are enriched for cells of neural crest origin. Consistent with previous observations, β-galactosidase (β-Gal)-positive cells were enriched in hair follicles, as detected by both X-Gal staining (Fig. 1D, representative whisker hair follicle) and β-Gal immunoreactivity (13.5% ± 1.36%; Fig. 1G, 1H). Within hair follicles, there was further concentration of β-Gal-positive cells within the dermal papilla of both small pelage hair follicles (90.9% ± 3.5%; Fig. 1G) and whisker hair follicles (88.9% ± 4.6%; Fig. 1H). Thus, the dermal papillae of both pelage and whisker follicles are almost entirely of neural crest origin.

To investigate whether the neural crest-derived population we identified in the dermal papilla could represent an enriched source of SKPs, we asked whether individual cells isolated from the dermal papilla could proliferate as spheres in SKP proliferation medium. To generate single dermal papilla cells, microdissection of the adult rat vibrissal follicle was undertaken according to an established technique (Fig. 1D–1F), and the vibrissal dermal papilla was cleanly isolated (Fig. 2D). Initial studies revealed that isolated whisker dermal papillae were not broken down by standard enzymatic dissociation using trypsin. Therefore, single dermal papilla cells were generated by microdissection of the papilla followed by incision and exposure of the inner surface of the papilla using fine needles. The papillae were subsequently enzymatically and mechanically dissociated to a single cell suspension. Papilla cells were cultured in SKP proliferation medium at ultralow density (10 cells per milliliter). Free-floating spheres (hereafter referred to as “papillaspheres”) appeared between day 5 and day 14 (Fig. 2G). Papillasphere generation was reliable (8 of 10 consecutive cultures). Clonal papillaspheres could be generated by plating single papilla cells in a 96-well plate. Papillaspheres were propagated and expanded in vitro by repeated passaging for at least 6 months. The doubling time for early passage spheres was approximately 2–3 days and was higher than whole skin SKP cultures derived from the same rat. When cultured at ultralow density as previously described, 24.4% ± 4.3% of dissociated single dermal papilla cells generated spheres (n = 6). To determine the degree of enrichment of the adult whisker papilla niche for sphere-forming cells, we cultured dissociated cells from whole rat vibrissal follicles and whole rat facial skin (without vibrissal follicles) in identical conditions (Fig. 2E), whereupon 2.8% ± 0.6% (n = 12) of dissociated vibrissal follicle cells generated spheres and 0.029% ± 0.028% (n = 36) of dissociated whole facial skin generated spheres. These findings are consistent with an approximate 1,000-fold enrichment of sphere-forming cells in the adult dermal papilla compared with adult whole facial skin (p < .001, unpaired t-test).

Figure Figure 2..

The dermal papilla is highly enriched for sphere-forming cells of neural crest origin. (A–C): Immunohistochemical analysis of the end bulb of the rat vibrissal follicle (green, fibronectin; blue, Hoechst nuclear stain; red, nestin). (D): The dermal papilla was cleanly microdissected prior to dissociation to single cells (described in text). Scale bar = 100 μm. (E): The microdissected adult dermal papilla was highly enriched for sphere-forming cells; 24.4% ± 4.3% of dermal papilla cells generated spheres when dissociated and cultured in SKP proliferation medium. This compares with 2.8% ± 0.6% of dissociated whisker follicle cells and 0.029% ± 0.028% of dissociated whole facial skin. (F): Papillaspheres express the neural crest transcription factors, Pax3 and Slug (embryonic day 12 neural tube = neural crest control). (G–N): Spheres generated from the rat vibrissa dermal papilla had an immunocytochemical profile similar to that of SKPs isolated from whole facial skin. Papillaspheres expressed nestin (red [I, M]) and fibronectin (green [J, N]). Papillaspheres and SKPs derived from the Wnt1-Cre/R26R mouse both expressed β-galactosidase, detected by X-Gal staining (H, L). Abbreviations: GAPDH, glyceraldehyde-3-phosphate dehydrogenase; SKP, skin-derived precursor cell; X-Gal, 5-bromo-4- chloro-3-indolyl-β-d-galactoside.

Immunohistochemical analysis of the intact adult rat vibrissal follicle demonstrated that in vivo cells of the dermal papilla express fibronectin but do not express nestin (Fig. 2A–2C), two key markers of cultured SKPs [1, [2], [3]–4]. Immunocytochemical analysis of whole spheres demonstrated that papillaspheres expressed nestin, an intermediate filament protein expressed in multipotent stem cells and immature neuroepithelial cells [1, 2] (Fig. 2I). Papillaspheres expressed the mesodermal markers fibronectin (Fig. 2J) and vimentin and also expressed α-smooth muscle actin and versican, markers of cultured dermal papilla cells [21]. Expression of musashi, an RNA-binding protein expressed in a number of precursor cells, including neural and neural crest precursor cells, was also detected (not shown) [22]. Papillaspheres did not express Oct 4 or SSEA, markers of pluripotency (not shown) [23]. To compare the immunocytochemical profile of papillaspheres with that of whole facial skin-derived precursor cells, we next generated SKPs from dissociated whole adult facial skin. Both papillaspheres and whole skin-derived SKPs expressed nestin and fibronectin (Fig. 2I–2M), as well as vimentin and musashi. Expression of SMA and versican was not detected in SKPs by immunocytochemistry. Consistent with a shared neural crest origin, both papillaspheres and SKPs derived from Wnt1-Cre/R26R mice expressed β-Gal (Fig. 2H, 2L). To confirm the neural crest origin of papillaspheres, we demonstrated using reverse transcription PCR that papillaspheres expressed the neural crest transcription factors Pax3 and slug (Fig. 2F).

Directed Differentiation of Neuronal and Glial Cells from Papillaspheres

To examine the phenotypic potential of papillaspheres, spheres were plated on poly-d-lysine/laminin-coated coverslips and initially differentiated in medium containing DMEM and 10% fetal calf serum, whereupon almost all cells displayed a fibroblast-like morphology and expressed fibronectin (Fig. 3F). A small number of cells (<2%) expressed FABP and exhibited a characteristic morphology consistent with adipocytes (Fig. 3G). No cells with a neuronal or glial morphology were identified when papillaspheres were differentiated in serum alone (Fig. 3A, 3J).

Figure Figure 3..

Generation of neuronal and glial cells from papillaspheres. (A–E): Papillaspheres differentiated in neuronal Diff medium containing neurotrophic factors (NT3, brain-derived neurotrophic factor, and NGF), fibroblast growth factor 2, and retinoic acid generated cells with a neuronal morphology that expresses β3 tubulin, MAP2ab, and p75/NGF-R. (F, G): Papillaspheres differentiated in 10% serum predominately generated fibronectin-expressing cells (fibronectin, red [F]) with the morphology of dermal fibroblasts and occasionally generated cells (<2%) with the characteristic morphology of adipocytes, which express FABP (green [G]). (H): +NTs and retinoic acid significantly increased generation of β3-tubulin-positive cells from papillaspheres compared with basal medium alone (-NTs) (8.8% ± 3.2% vs. 0.14% ± 0.14%; p < .001). (I): Forskolin and neuregulin significantly increased the generation of Schwann cells from papillaspheres compared with basal medium alone (21.2% ± 4.4% vs. 1.22% ± 0.32%; p < .005). (J–N): Papillaspheres differentiated in glial Diff medium containing forskolin and neuregulin-1β generated cells with an elongated bipolar morphology that coexpress S100 (green [L]), p75 (red [M]), glial fibrillary acidic protein, and CNPase, consistent with Schwann cells. Abbreviations: B3-Tubulin, β3-tubulin; Diff, differentiation; FABP, fatty acid-binding protein; NGF, nerve growth factor; NGF-R, nerve growth factor receptor; +NT, neurotrophic factors.

Directed differentiation of papillaspheres to neuronal cells and Schwann cells was next examined using differentiation protocols adapted from established SKP culture conditions [3, [4], [5]–6]. To induce neuronal differentiation, papillaspheres were differentiated on poly-d-lysine/laminin-coated coverslips in neuronal differentiation medium containing the neurotrophic factors BDNF, NGF, NT3 plus retinoic acid, and FGF2 [24]. Cells with a characteristic neuronal morphology were evident from day 10 of the differentiation protocol. Immunocytochemical analysis of these cells showed that these cells coexpressed β3-tubulin, p75, and MAP2ab (Fig. 3B–3E). The addition of neurotrophic factors, retinoic acid, and FGF to a basal medium containing 1% fetal calf serum significantly increased the generation of β3-tubulin-positive neuronal cells compared with basal medium alone (SEM, 8.8% ± 3.2% vs. 0.14% ± 0.14%; p < .001, t test; Fig. 3H).

Directed differentiation of papillaspheres to a glial lineage was performed using cues that are known to induce Schwann cell differentiation from skin-derived precursor cells and neural crest stem cells. Following differentiation in Schwann cell differentiation medium containing neuregulin and forskolin, elongated cells with a bipolar morphology and phase-bright cell body were visible from 10 days in vitro. These cells coexpressed S100, p75NGFR, GFAP, and CNPase, consistent with a Schwann cell phenotype (Fig. 3K–3N). Addition of forskolin and neuregulin to a basal medium containing 1% serum significantly increased the generation of Schwann cells from papillaspheres compared with basal medium alone (21.2% ± 4.4% vs. 1.22% ± 0.32%; p < .005, t test; Fig. 3I).

Targeted Microdissection of the Dermal Papilla of Adult Human Skin Samples Leads to Efficient Generation of Sphere-Forming Cells

Having demonstrated the high degree of enrichment of the rat dermal papilla for neurogenic sphere-forming cells, we asked whether targeting this niche in human skin could provide an efficient and reliable method of generating spheres from a defined region of adult human skin. Consistent with previously reported findings [7], initial studies confirmed the difficulties in generating reliable and expandable sphere cultures from dissociated whole adult human skin, with 5 of 10 surgically derived adult human skin samples generating primary spheres, all of which underwent only limited expansion.

Four adult human skin samples were obtained from hair-bearing facial and scalp skin from both male and female subjects (Table 1). From these samples (Fig. 4A, 4B), the dermal papillae of facial hair follicles were microdissected using a standard technique modified from the vibrissal dermal papilla dissection [20]. Human dermal papillae were resistant to dissociation to single cells by enzymatic dissociation with trypsin and collagenase followed by mechanical dissociation. This suggests that cells from this hair follicle niche are not exposed or dissociated to single cells by the standard techniques of SKP culture. Single dermal papilla cells were therefore generated by allowing the cells to grow out from the explanted papilla on an uncoated tissue culture plastic surface for 4–10 days [13]. Cells were subsequently enzymatically detached from the plastic with trypsin, resuspended as single cells in SKP proliferation medium, and cultured as previously described. Small spheres were seen in culture at days 4–5, with the appearance of larger spheres around day 14 (Fig. 4C). These spheres could be repeatedly passaged and rapidly expanded as floating spheres. Spheres adhered and differentiated upon plating onto PDL/laminin-coated coverslips. Human papillaspheres could be reliably generated from a single hair follicle dermal papilla (four consecutive samples). To demonstrate the increased sphere-generating efficiency of this strategy of microdissecting an enriched niche of sphere-forming cells, sister sphere cultures of whole dissociated skin were generated using standard SKP culture protocols [1, 3, 5]. In all four facial skin samples, despite much lower starting cell numbers, sphere cultures established by microdissection of the hair follicle dermal papilla resulted in both more reliable generation and better expansion of spheres compared with sphere culture of paired whole dissociated skin (Table 1).

Table Table 1.. Efficient generation of spheres from human skin samples by microdissection of the hair follicle dermal papilla niche
original image
Figure Figure 4..

Generation of human papillaspheres by microdissection of facial hair follicles. (A): H&E stain of human facial skin showing epidermis, dermis, and location of the end bulb of the hair follicle within the skin (boxed area). (B): Higher-power image of the same section showing the outline of the dp, which can be microdissected with minimal contamination using a method similar to that outlined in Figure 1. Scale bar = 50 μm. (C): Cells generated from a single human hair follicle dp grew as spheres in SKP proliferation medium. (D–H): The majority (78.6% ± 12.4%) of plated human papillasphere cells expressed the neural crest transcription factor Sox10. A subpopulation of cells coexpressed Pax3 (white arrow). Abbreviation: dp, dermal papilla.

Directed Differentiation of Neuronal and Glial Cells from Human Papillaspheres

We next examined the immunocytochemical profile of spheres derived from the adult human dermal papilla. Immunocytochemical analysis of whole spheres demonstrated that human papillaspheres express nestin and fibronectin (Fig. 5A–5C). Human papillaspheres also express musashi and vimentin (not shown). Human papillaspheres therefore express the same key markers as rodent and human SKPs [2, [3]–4, 7].

Figure Figure 5..

Human papillaspheres express nestin and generate neural progeny. (A–C): Human papillaspheres, like SKPs and rat papillaspheres, express nestin (green [A]) and fibronectin (red [B]). (D–F): Directed differentiation of human papillaspheres in Schwann cell differentiation medium containing forskolin and neuregulin1 leads to the generation of cells with a Schwann cell morphology that coexpressed S100 (green [D]; 36.7% ± 5.06%) and CNPase (red [E]) (large white arrows). (G–I): Human papillaspheres differentiated in neuronal differentiation medium containing neurotrophic factors and retinoic acid generated cells with a neuronal morphology that coexpressed β3-tubulin (red [G]; 7.53% ± 3.34%) and MAP2ab (green [H]).

The majority of cells (78.6% ± 12.3%) derived from papillaspheres express Sox10, a transcription factor that plays a key role in neural crest stem cell maintenance and differentiation [25]. An additional subpopulation of Sox10-positive cells coexpresses the neural crest transcription factor Pax3 (Fig. 4E–4G).

Directed differentiation of human papillaspheres to neuronal cells and Schwann cells was examined using differentiation protocols adapted from vibrissal papillasphere experiments. As observed with rat papillaspheres, initial experiments demonstrated that differentiation of human papillaspheres in 10% FCS uniformly generated cells with the flattened morphology of dermal fibroblasts that expressed fibronectin. In these conditions, occasional cells with the characteristic morphology of adipocytes were seen, but no cells with a neuronal or glial morphology were seen. Following differentiation in a modified Schwann cell differentiation medium containing neuregulin-1 and forskolin, cells with a Schwann cells morphology that expressed CNPase, GFAP, and S100 were seen (36.7% ± 5.1%; Fig. 5D–5F). Human papillaspheres also generated cells with a neuronal morphology when differentiated in neuronal differentiation medium containing the neurotrophic factors BDNF, NT3, and NGF (7.5% ± 3.3%; Fig. 5G–5I). Taken together, these results suggest that targeted microdissection of the human dermal papilla allows the efficient generation of precursors cells with neuronal and glial potential from adult human skin.

Discussion

We have used a fate-mapping approach combined with microdissection to identify and characterize a highly enriched niche of neural crest-derived cells in the adult hair follicle dermal papilla; these cells are able to generate multipotent spheres, similar to skin-derived precursor cells. The enrichment of this niche for sphere-forming cells is striking and is approximately three orders of magnitude higher than primary dissociated whole skin cells from a similar skin region. The spheres generated from this niche are of neural crest origin, express markers of multipotentiality, and are able to generate neuronal cells and glial cells in addition to mesenchymal derivatives. We demonstrated that comparable precursors, which express neural crest transcription factors, can be readily generated from a single human hair follicle dermal papilla. In our hands, the dermal papilla niche was not broken down by the standard mechanical dissection and enzymatic dissociation used to generate SKPs from whole skin, and we demonstrated that targeted microdissection of the human dermal papilla provides a method of efficiently generating neurogenic multipotent cells from a defined adult skin niche.

The skin contains many different cell types and a number of stem and precursor cell populations in the dermis and the follicular and interfollicular epidermis [26]. Other groups have also demonstrated neural crest origins and neural potential of other rodent hair follicle niches, in particular the bulge region [13, 14, 27, 28], another promising reservoir of clinically useful cells. Recent work demonstrates that neural crest-derived sphere-forming cells in adult skin can be traced to multiple lineages [8], highlighting the problem of heterogeneity of spheres derived from different anatomical locations and emphasizing the need to demonstrate specific properties of individual niches. Studies of human neonatal foreskin, which does not contain hair follicles, provide evidence for an extrafollicular SKP niche [3]. However, the difficulty of reliably isolating and expanding comparable populations of SKPs from both whole hairless and hairy adult human skin [7] suggests that this extrafollicular niche may be significantly limited after the neonatal period in humans. The dermal papilla offers a significant practical and experimental advantage in comparison with other niches in adult skin; microdissection is straightforward and more discrete, thus reducing contamination (Fig. 2D).

Our data support the technique of targeted dermal papilla microdissection as a reliable and efficient method of generating defined SKP-like cells from single hair follicles. In addition, we suggest that the cellular differentiation system described here offers a number of advantages for the study of adult multipotent skin-derived neural crest cells. In particular, the system we describe is both anatomically and developmentally defined and derived from a homogeneous starting population. This may enable various issues to be addressed, such as the in vivo role of endogenous precursors and molecular mechanisms by which Schwann cells and neuronal cells can be generated from an accessible autologous adult source. Furthermore, we define a minimum volume of skin that would need to be sampled to study the behavior of neural crest precursor/stem cells, the dysfunction of which has been implicated in a variety of human diseases [29].

We have previously described a method for generating multipotent spheres from adult human skin using a method that involves expansion of small numbers of precursors in serum, with subsequent re-exposure to EGF and FGF2 [7]. These spheres generate electrophysiologically active GABAergic and glutamatergic neurons when differentiated in the presence of hippocampal astrocyte-derived signals. It is not possible to directly compare our findings from a defined dermal niche with those from the whole dermis [7], but it would appear that the resulting neural derivatives may be comparable. It is also of considerable interest to note that certain neural precursor cells are capable of generating both central and peripheral neural cells, depending on the microenvironment, and this may provide an important avenue for further investigation [30].

This study is restricted to facial dermis. The vibrissal follicle of the rat, although attractive for experimental study, may not fully reflect the behavior of the hair follicles on other areas of skin, in particular the trunk, where the neural crest developmental contribution is less well-characterized [8, 31]. It has recently been shown in the mouse that neural crest-derived spheres isolated from trunk skin possess properties different from those of spheres derived from the face [8]. It will be important in future studies to compare the behavior of human papillaspheres generated from nonfacial areas, especially in cosmetically discrete regions such as the axilla.

Summary

We describe the high degree of enrichment of multipotent neural crest-derived sphere-forming cells within a defined adult skin region amenable to microdissection. In the human, targeted microdissection of the hair follicle dermal papilla niche allows the efficient generation of precursors with neurogenic potential. The reported system is both defined and minimally invasive and therefore offers a number of advantages for translational study of the generation of neuronal and glial cells from adult skin.

Disclosure of Potential Conflicts of Interest

The authors indicate no potential conflicts of interest.

Acknowledgements

We are grateful to Dr. Ruma Raha-Chowdhury for advice and assistance. D.P.J.H. is supported by a Wellcome Trust Research Training Fellowship and Sackler studentship. Wnt1-Cre transgenic mice were kindly supplied by Drs. Deborah Henderson and Jon Peat, Institute of Human Genetics, University of Newcastle, Newcastle, U.K. C.A.J. supported by a BBRSC stem cell initiative grant.

Ancillary