Commitment of embryonic stem cells to an epidermal cell fate and differentiation in vitro



The epidermis develops from a stem cell population in the surface ectoderm that feeds a single vertical terminal differentiation pathway. To date, however, the limited capacity for the isolation or purification of epidermal stem or precursor cells has hampered studies on early commitment and differentiation events. We have developed a two-step culture scheme in which pluripotent mouse embryonic stem (ES) cells are induced first to a surface ectoderm phenotype and then are positively selected for putative epidermal stem cells. We show that the earliest stages of epidermal development follow an ordered sequence that is similar to that observed in vivo (expression of keratin 8, keratin 19, keratin 17, and keratin 14), suggesting that ES cell-derived surface ectoderm-like cells can be induced to follow the epidermal developmental pathway. At a low frequency, keratin 14-positive early epidermal cells progressed to keratin 1-positive and terminally differentiated cells producing a cornified envelope. This culturing protocol provides an invaluable system in which to study both the mechanisms that direct stem cells along the epidermal pathway as well as those that influence their subsequent epidermal differentiation. Developmental Dynamics 232:293–300, 2005. © 2004 Wiley-Liss, Inc.


The mammalian epidermis is derived from ectoderm through a process of cell fate selection and lineage progression that results in the stratified squamous epithelium; it is a continuously renewing tissue that is replenished through the differentiation of basal cells through the terminal differentiation pathway (Wolpert et al., 2001). The events that occur during the later stages of epidermal differentiation, that is the steps leading from mature keratin 5/keratin 14 (K5/K14) -positive basal cells to the terminally differentiated epidermal cells of the cornified layer, have become well understood (Fuchs, 1993; Fuchs and Byrne, 1994; Byrne, 1997; Watt, 2002). However, an important stumbling block has been the absence of an appropriate in vitro model system in which to study the early events leading to the commitment of epidermal stem and progenitor cell populations. The ability to generate differentiated epidermal progeny from a continuously growing stem cell population in vitro, therefore, would provide a unique system for the study of stem cell/very early progenitor potential (Hall and Watt, 1989). It would also make possible a comprehensive analysis of the underlying molecular mechanisms for the onset of embryonic epidermal commitment and differentiation.

Embryonic stem (ES) cells derived from the inner cell mass of mouse blastocysts (embryonic day [E] 3.5; Evans and Kaufman, 1981; Martin, 1981) cultured on a suitable fibroblast feeder layer proliferate and remain totipotent indefinitely (Martin, 1981). However, the transition of ES cells to suspension culture induces cell aggregation and cystic embryoid body (EB) formation morphologically reminiscent of the embryo at the 6- to 8-day egg cylinder stage (Doetschman et al., 1985). Depending on the inducing agents and conditions, differentiation along pathways leading to islet-like cells, hair cells, male gametes, glial cells, osteoblasts, chondrocytes, neutrophils, and hepatocytes (to name a few) has been demonstrated (Lumelsky et al., 2001; Bronson, 2003; Gregg and Weiss, 2003; Hamazaki and Terada, 2003; Li et al., 2003; Lieber et al., 2003; Sui et al., 2003; Geijsen et al., 2004). We therefore set out to develop a similar in vitro culture system to study and delineate the epidermal lineage in vitro by devising a two-step strategy in which ES cells are differentiated along the epidermal lineage. Our results indicate that the culture conditions used are sufficient to support the progression and differentiation of ES cells along the epidermal lineage, as demonstrated by morphology as well as by the temporally regulated expression of epidermal differentiation markers. This in vitro approach will be an invaluable tool for recapitulating the cellular events of the epidermal differentiation pathway from uncommitted ES cells to functional epidermal cells, including the characterization of putative epidermal cells, the delineation of cell fate selection, as well as cell commitment and differentiation in vitro.


Induction of ES Cells to Ectodermal Fate and Enrichment for Early Epithelial Cells on Basement Membrane Matrix

ES cells maintained in conditions noninductive to differentiation, such as on mitomycin-C–treated feeder layers (Nagy et al., 1993), remain undifferentiated, as evidenced morphologically by the tightly packed colonies comprising cells without distinct cell borders or cytoplasm as well as high levels of alkaline phosphatase activity (Fig. 1A). To induce epidermal cell development, we first dissociated undifferentiated ES cells and cultured them in suspension on bacterial grade Petri dishes, thereby promoting the formation of floating EBs (Fig. 1B). We previously had observed that culturing EBs at a high plating density (2 × 106 cells/10 cm Petri dish) in suspension for 6 days resulted in the expression of K8 (an ectodermal or early epidermal marker), a process highly dependent upon bone morphogenic protein (BMP) signaling (Turksen and Troy, 2001). By using this density, we found by day 6, large spherical cystic EBs formed with a distinctly cuboidal surface ectoderm-like monolayer around a liquid filled cavity (Fig. 1B). Longer culturing led to larger but increasingly necrotic EBs with poor differentiation progression (Turksen and Troy, 2001).

Figure 1.

Embryonic stem (ES) cells, embryoid bodies (EBs), and differentiating EBs (dEBs) in culture. A: Tightly packed undifferentiated ES cell colonies maintained on mitomycin-C–treated feeder layers express high levels of alkaline phosphatase activity after 2 days in culture. B: Disassociated ES cells were cultured at high density in suspension for EB formation. After 24 hr, small aggregates were visible with irregular outlines that progressed to form spherical structures after 4 days in culture. By 6 days in suspension, EBs were cystic with conspicuous cuboidal surface ectoderm-like cell morphology. C: EB aggregates seeded onto plastic, type I collagen, and BM matrix were assessed for their ability to progress along the epithelial lineage. After 6 days, EBs on plastic exhibited inefficient migration of differentiating cells, whereas dEBs on collagen and BM matrix adhered better with better migration of epithelial-like cells. On BM matrix, epithelialization was most effective with three distinct cell zones evident. Toward the center were cuboidal early epidermal cells followed by an area of tightly packed epithelial cells with flat, larger cells along the edge of the differentiating EB. D: After 4 days, dEBs expressed an early epidermal marker (K8), an early epidermal progenitor marker (K17) as well as a mature epidermal marker (K14), indicating that the migrating sheets of cells of dEB cultures were progressing along the epithelial lineage. After 4 days, migrating epithelial dEB sheets were disassociated and replated on BM matrix. E: After 2 more days, epithelial progenitor cells (EPC) cultures were composed of morphologically recognizable epithelial cells.

Cystic EBs were reseeded onto uncoated tissue culture grade dishes or dishes coated with either collagen type I (∼0.05%) or basement membrane (BM) matrix (growth factor reduced Matrigel, 0.1 mg/ml; Fig. 1C). EBs plated onto plastic displayed the poorest adherence and cell migration (Fig. 1C) as well as the greatest amount of cell death after 6 days. In comparison, EBs seeded onto type I collagen adhered faster, and after 6 days, cells with epithelial morphology were migrating out from the EB center (Fig. 1C); however, cell lysis was still considerable. Conversely, EBs plated onto BM matrix were the quickest to adhere (more than 95% of the plated EBs attached) and large numbers of cells were observed migrating out and generating epithelial sheets after 6 days (Fig. 1C). The efficacy of the BM matrix was evident as early as 2 days after plating. The epithelial-like sheets showed a gradient of smaller progenitor-like epithelial cells at the center of the differentiating EB (dEB) followed by cuboidal cells becoming progressively larger and well spread in their morphology toward the edge. Immunofluorescence analysis confirmed the expression of K8 throughout the spreading area of the highly migratory, early-committed epidermal cells with the existence of smaller K17-positive cells (an early epidermal progenitor marker [McGowan and Coulombe, 1998]) as well as K14-positive cells (a marker associated with the basal layer of the epidermis [Coulombe et al., 1989; Kopan and Fuchs, 1989]) dispersed throughout the migrating epithelial sheet but most often seen at the periphery (Fig. 1D). These observations suggest that these conditions induce a large number of very early epithelial progenitor cells (EPCs) from ES cells.

Expansion and Differentiation of Epidermal Progenitor Cells

We next asked whether EPCs could be expanded through subculturing and induced to further differentiation. dEBs cultured on Matrigel for 4 days (Fig. 1E) were disassociated and reseeded at a cell density of 105 to 106 cells (on BM matrix coated 35-mm dishes) where they quickly adhered and proliferated resulting in colonies with defined cell–cell contacts and a distinct epithelial morphology (Fig. 1E). In addition, members of the Wnt pathway, which is known to be influential in the patterning and differentiation of the epidermis, were expressed in this model (Fig. 2A). Many of the Wnts (including Wnt 1, 3, 4, and 5b) were not different in dEB vs. EPC cultures. However, there was an increase in the expression of other Wnts (Wnt 6, 7A, 7B, and 16) in the more differentiated EPC cultures. Differentiation of cells within developing EPC colonies was followed histologically with time at the population level using Ayoub Shklar (AS) staining (an effective and reliable marker of epithelialization in which undifferentiated epidermal progenitors are blue, differentiated epidermal cells are red, and further differentiation is represented with a colour change from orange to yellow to brown [Ayoub and Shklar, 1963]). AS staining revealed that epidermal differentiation was achieved in a time-dependent manner as cell density increased (note color changes from day 2 to day 16; Fig. 2B).

Figure 2.

Early epithelial gene expression and epithelial progenitor cells (EPC) differentiation. Reverse transcriptase-polymerase chain reaction was used to compare differentiating embryoid bodies (dEBs, 4 days in culture; lane 1) with EPCs (2 days in culture; lane 2) to assess the expression of markers involved in early epidermal lineage commitment. A: Wnt 1, 3, 4, and 5B remained unchanged, while Wnt 6, 7A, 7B, and 16 were upregulated in EPCs. The epithelialization program of EPC cultures was time- and cell density-dependent, as evidenced by Ayoub Shklar staining. B: After 2 days, all EPCs stain blue, consistent with their being undifferentiated epidermal progenitors. On day 4, red-stained EPCs are present in highly packed areas, indicating a commitment to further differentiation. Tightly packed orange-stained cells are present by day 6, while yellow-stained cells (yellow arrow) are present after 12 days in culture. After 14 days, yellow cells are more frequent in dense areas (white arrow) and by day 16, the appearance of brown-stained cells (the most differentiated) is observed (black arrow).

Immunofluorescence staining confirmed the progression of early epidermal differentiation. In early EPC colonies almost 100% of the cells were positive for K8 and the very early epidermal progenitor marker K19 (Michel et al., 1996) and with time in culture, both K19 and K8 remain coexpressed (Fig. 3). In addition, these very early cells also expressed the junctional proteins Claudin-6, Desmoplakin, and E-Cadherin (Fig. 3). That differentiation continued in an ordered progression was evident from double labeling as the number of K17- and K14-positive cells appeared with no coexpression of K8. With time, the number of K17- and K14-positive cells increased concomitant with a clear epidermal sheet-like morphology (Fig. 3); the increase in the levels of protein expression was confirmed through immunoblot analysis (Fig. 4A). Whereas these early stages of epidermal differentiation occurred robustly and reproducibly, further progression of cells to a more mature phenotype, including that consistent with terminal differentiation, occurred at a lower frequency. Nevertheless, as cell density increased, a few K1-positive cells formed (Fig. 3). In support of the immunofluorescence studies, immunoblot analysis indicated there was expression of K1 and later markers of epidermal differentiation (namely Involucrin and Filaggrin; Fig. 4A). In addition, Involucrin, Filaggrin, Loricrin, and terminal differentiation markers (markers in the cross-linking of cornified envelopes: SPRR 1A, 1B, and 2A [Kalinin et al., 2002; Segre, 2003]; Kruppel-like factor 4 [Klf 4; a regulator/transactivator in SPRR expression; Segre et al., 1999; Jaubert et al., 2003]; and distal-less homeobox 3 [Dlx 3; a regulator of Filaggrin; Morasso et al., 1996]) were detected by reverse transcriptase-polymerase chain reaction (RT-PCR) analysis using specific primers (Fig. 4B). A further indication of the progression of terminal differentiation was confirmed through the successful extraction of cornified envelopes from EPC cultures (Fig. 4C).

Figure 3.

Keratin and junctional marker expression analysis in EPCs. EPCs coexpress K8 and a very early epidermal progenitor marker (K19) within the first 2 days in culture and coexpression persisted throughout our culturing conditions. In addition, colonies of EPCs also express Claudin-6, Desmoplakin, and E-Cadherin at their cell–cell borders, indicative of a bona fide epithelial morphology. EPCs become K17- and K14-positive (without coexpressing K8), indicating that EPC cultures comprise early epidermal progenitors with the capability of progressing along the epithelial lineage in vitro. With time in culture, the number of positive K14 and K17 cells increases to form sheets; with K17 representing a subpopulation of the K14-positive cells. Also, a few cells expressing K1 are detectable.

Figure 4.

Epithelial progenitor cell (EPC) cultures commit to terminal differentiation in vitro. Intermediate filament proteins were extracted from EPC cultures for immunoblot analysis. A: Levels of K17 and K14 increased gradually with time from day 2 to day 10 and in further maturing cultures the expression of K1, Involucrin, and Filaggrin was evident. B: Reverse transcriptase-polymerase chain reaction analysis of RNA extracted from maturing EPC cultures indicated that later markers of epidermal differentiation, including Involucrin, Filaggrin, and Loricrin were expressed as were markers associated with terminal differentiation including SPRR 1A, 1B, and 2A as well as Klf 4 and Dlx 3. C: The production of insoluble cornified envelopes, although rather inefficient, was also apparent, corresponding with further differentiation in EPC cultures.


We report here the development of a two-step culture system in which the differentiation of totipotent ES cells toward the epithelial lineage in vitro occurs in a robust and reproducible manner. The model results first in the generation of large numbers of EPCs expressing K8, K19, and K17 markers. These cells can be expanded and induced with high efficiency to differentiate into K14-positive cells, which at a lower frequency are capable of further progression to terminal differentiation. In vivo, early developmental cell fate selection takes place in the surface ectoderm; therefore, the expression of K8 and various Wnt molecules known to be associated with the early developing epidermis (Wilson et al., 2001; Koster et al., 2002) in the transition from dEB to EPC cultures is supportive of the notion that the cultures contain ectodermal cells fated to the epidermal lineage. Furthermore, the high proliferative capacity of EPCs supports the notion that the cultures likely contain putative stem cells, although unambiguous identification and characterization of these cells remains to be done. Our data suggest, however, that the clusters of K17- and K14-positive cells may be similar to those reported in the developing epidermis by McGowan and Coulombe (1998) as putative bipotential epidermal stem cells/very early progenitors. Together, they also suggest that the culture conditions support primarily early epidermal differentiation stages, including the generation of large numbers of K14-positive cells that are less mature than the K14-positive cells isolated from newborn backskin. Several in vitro systems consisting of organ cultures or epidermal cells isolated from newborn epidermis have been used previously to investigate the process of epidermogenesis (Fusenig et al., 1994; Kamimura et al., 1997; Li et al., 2004). However, the culture system described here is unique and offers the advantage that the starting population of totipotent ES cells progress along the epidermal lineage in a predictable sequence and through a series of different cell types recognizable by both morphology as well as marker (keratin and signaling molecule) expression profiles reminiscent of the in vivo epidermal differentiation program. Remarkably, the timing of the first appearance of various epidermal markers closely resemble that reported in the developing mouse epidermis (Byrne et al., 1994). For instance, the early expression of Wnts in these cultures presumably reflects the earliest stages of epidermal specification and the presence of epidermal precursors in the ES cell-derived cultures. Furthermore, it is understood that in vivo the emergence of EPCs in the developing epidermis depends on ECM in addition to perhaps a microenvironment produced by the dermis. Therefore, another important similarity is the apparent requirement in vitro for BM matrix for efficient epithelial differentiation: without it, we saw only sporadic differentiation. The model system described should be invaluable for studying the various stages of epidermal development especially the evaluation and characterization of stem/very early progenitors in vitro. However, further work will be required to enhance late-stage and terminal-stage differentiation in this model, which occurred at lower frequency and only in culture regions of very high cell density, pointing toward the need for a specialized microenvironment that is not yet adequately mimicked in our model. Nevertheless, as the availability and use of human ES cells become more routine, similar culture conditions may be useful for assessing their utility for generating large numbers of EPCs for studies on potential therapies and drug screening, for a myriad of diseases associated with epithelial differentiation such as skin blistering diseases and defective epidermal permeability barrier disorders.


Pluripotent ES Cell Cultures

The R1 ES cell line (Nagy et al., 1993) was generously provided by Dr. Janet Rossant and Dr. Andras Nagy from the University of Toronto (Toronto, Canada). ES cells were maintained as described (Nagy et al., 1993) on a feeder layer of mitomycin C (Roche Applied Science, Laval, Canada) treated mouse embryonic fibroblasts (0.01 mg/ml, for 2–3 hr) in Dulbecco's modified Eagle medium (DMEM, high glucose without L-glutamine) supplemented with 1% nonessential amino acids, 1% sodium pyruvate, 1% penicillin –streptomycin (all from Invitrogen, Burlington, Canada), and 15% fetal calf serum (characterized and screened for ES cell cultures; HyClone, Logan, Utah) at 37°C in 5% CO2. To monitor the stem cell characteristics of the R1 line, alkaline phosphatase staining was routinely used to ensure their undifferentiated/pluripotent state. Furthermore, the experiments described in this study were performed within six passages of thawing the ES stocks. As with all phase photography reported here (morphology, alkaline phosphatase, and AS staining), images were captured by using an Olympus CK2 inverted microscope (Olympus, Melville, NY) equipped with a Nikon COOLPIX 4500 digital camera (Nikon Canada, Mississauga, Canada), and image processing was achieved with Adobe Photoshop 7.0 software (Adobe Systems, Inc., San Jose, CA).

Alkaline Phosphatase Histochemistry

ES cell cultures described above were rinsed with cold 1× phosphate buffered saline (PBS) and fixed at room temperature with cold formalin for 15 min followed by a rinse in H2O and a second incubation in H2O for 15 min. Staining solution (25 ml) was prepared by combining 2.5 mg of naphthol AS-MX phosphate, 100 μl of N,N-dimethyl formamide, 12.5 ml of 0.2 M Tris-HCl pH 8.3, 12.5 ml of H2O, and 15 mg Fast Red Violet LB salt. Staining solution was filtered directly onto the cells after the second H2O wash and incubated for 45 min at room temperature. Plates were rinsed with H2O before observation.

In Vitro Differentiation of ES Cells

The differentiation of ES cells was accomplished in a two-step process: (1) embryoid body (EB) formation was obtained by using the cell suspension assay described by Keller et al. (1993). Subconfluent, undifferentiated ES cells were trypsinized to obtain a single cell suspension that was plated in Petri dishes with supplemented DMEM (as described above) at a density of 2.5 × 106 cells/ml and cultured at 37°C in 5% CO2. Cystic EBs cultured in suspension for 6 days were seeded at low density onto uncoated tissue culture dishes or plates coated with either growth factor reduced Matrigel (BM matrix, 0.1 mg/ml; BD Biosciences, Mississauga, Canada) or type I collagen (∼0.05%; Sigma-Aldrich, Oakville, Canada) to initiate dEB formation at 37°C and 5% CO2. For immunofluorescence analysis, EBs were plated on BM matrix-coated coverslips and cultured for various time periods. For RT-PCR analysis EBs were plated on BM coated tissue culture dishes and RNA was extracted at different time points. (2) After 4 days in culture, dEBs were disassociated and replated at high density on BM matrix coated tissue culture plates to expand the EPCs. For the analysis of keratin expression and differentiation, cells were cultured on BM matrix-coated coverslips (105−106 cells/35 mm) and for AS staining, EPCs were seeded on BM matrix-coated 60-mm tissue culture dishes (2 × 106 cells/60 mm). For RT-PCR as well as for protein analysis, EPCs were seeded onto BM matrix-coated 100-mm tissue culture dishes (5.5 × 106 cells/100 mm) for RNA/protein extraction at various time points.

AS Staining as a Marker for Differentiation

EPC cultures were rinsed three times with 1× PBS followed by fixation for 30 min in cold formalin solution at room temperature. Followed by a brief rinse in 1× PBS, cultures were stained for 3 min with 5% Acid Fuschin. Excess stain was removed and cells were then incubated with Aniline Blue–Orange G solution for 45 min at room temperature. Finally, plates were rinsed with 95% ethanol and were inverted to dry before microscopic evaluation and photography.


Differentiating EBs and EPCs cultured on BM matrix-coated coverslips were processed for immunofluorescence as previously described (Turksen and Aubin, 1991; Troy and Turksen, 1999; Turksen and Troy, 2002, 2003). Briefly, cells were fixed in methanol at −20°C for 10 min and washed in 1× PBS then incubated with appropriately diluted primary antibody for 30 min in a humidified chamber at room temperature. The following antibodies were evaluated in these cultures: K8 (Troma-1) (undiluted, Hybridoma Bank, Iowa City, IA), K19 (1:5, Amersham Biosciences, Baie d'Urfé, Quebec), K17 (1:500, a gift from Dr. Pierre Coulombe), K14 (1:100, a gift from Dr. Elaine Fuchs), and K1 (1:100, a gift from Dr. Stuart Yuspa) as well as Claudin-6 (1:100), Desmoplakin (1:10; a gift from Dr. Panos Kouklis), and E-Cadherin (1:2,000; Sigma-Aldrich). Coverslips were then rinsed in 1× PBS and incubated for 30 min in a humidified chamber at room temperature with fluorescein isothiocyanate- or Texas Red-conjugated secondary antibodies (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA) against rabbit, mouse, or rat as required at a 1:50 dilution. After secondary antibody incubation, cells were treated with Hoechst 33258 (1:100, Sigma-Aldrich) for 5 min and mounted in Mowiol 4-88 (Polysciences, Inc., Warrington, PA). Images were captured with a Zeiss microscope (Carl Zeiss, Mississauga, Canada) equipped with epifluorescence optics as well a Zeiss digital AxioCam camera using Axio Vision 2.05 software. Acquired images were processed with Adobe Photoshop 7.0 (Adobe Systems, Inc.).

Immunoblot Analysis

EPC cultures were collected at various time points in intermediate filament protein extraction buffer (20 mM Tris-HCl pH 7.4, 0.6 M KCl, 1% Triton X-100, and 0.3 mg/ml phenylmethyl sulfonyl fluoride [PMSF]). Samples were centrifuged at high speed and resuspended in urea/Tris buffer (9 M urea, 20 mM Tris-HCl pH 7.5, 10% β-mercaptoethanol, and 0.3 mg/ml PMSF). A total of 20μg of protein was separated on 7.5% sodium dodecyl sulfate-polyacrylamide gel electrophoresis gels, transferred to nitrocellulose, and processed for immunoblot analysis as previously described (Troy and Turksen, 1999; Turksen and Troy, 2002). Blots were incubated with antibodies against K17 and K14 (diluted 1:10,000), K1 (1:30,000), Involucrin (1:20,000; Covance, Denver, PA), and Filaggrin (1:20,000; a gift from Dr. Beverly Dale) followed by horseradish peroxidase-conjugated secondary antibodies against rabbit (diluted 1:10,000–1:50,000; Amersham). Detection was achieved by using the ECL+ Plus kit (Amersham) according to the manufacturer's instructions and was visualized on Kodak autoradiography film (Kodak, Rochester, NY).

RNA Isolation and RT-PCR

RNA was extracted from cultured cells by using TRIzol reagent (Invitrogen) according to the manufacturer's instructions followed by treatment with DNaseI (Invitrogen) and first-strand cDNAs were synthesized as previously described (Turksen and Troy, 2002, 2003). PCR was performed (Turksen and Troy, 2002, 2003) with specific primers (Table 1), and products were visualized on agarose gels in the presence of ethidium bromide.

Table 1. RT-PCR Primersa
GeneForward primer 5′-3′Reverse primer 5′-3′Size (bp)Reference
  • a

    RT-PCR, reverse transcriptase polymerase chain reaction.



We thank Dr. Janet Rossant and Dr. Andras Nagy for providing us with the R1 ES cell line. In addition, Dr. Elaine Fuchs kindly provided us with antibodies against K14, Dr. Panos Kouklis with Desmoplakin antibodies, Dr. Stuart Yuspa with K1 antibodies, and Dr. Beverly Dale with Filaggrin antibodies. We also thank Dr. Pierre Coulombe for antibodies against K17 as well as for engaging us in many exciting lineage discussions. We also greatly appreciate Dr. Paolo Dotto for his encouragement in support of our studies. Last but certainly not least, we acknowledge Dr. Jane Aubin for her continuing support and enthusiasm for our research.