After obtaining informed consent, HDFs were isolated from newborn or adult foreskin immediately following circumcision performed at either the Cha Hospital or Goodman Clinic in Seoul, Korea. The epidermal sheet was mechanically removed after overnight incubation with dispase II (Roche, Indianapolis, IN), cut into 1 mm² pieces, and then incubated in an enzyme cocktail consisting of 0.1% collagenase (Worthington, Lakewood, USA), 0.1% hyaluronidase, and 0.075% DNase (Sigma, St. Louis, USA). Isolated HDFs were then cultured in FGM2 medium (Clonetics, San Diego, CA) using 0.1% gelatin-coated culture dishes. When specified, HDFs were purified on MACS columns using anti-fibroblast beads (Miltenyl Biotech, Auburn, CA). Human keratinocytes (MCTT Inc, Seoul, Korea) and MSCs (Clonetics) were cultured according to the manufacturer’s specifications.

After one passage in FGM2 culture medium (Clonetics), HDFs were plated at a density of 10,000 cells/cm² on poly-D-lysine/laminin(Sigma)-coated cover slips, or in culture dishes at 5×10⁴ cells/ml, in identical medium. Cultures were incubated for one day to allow cell attachment, then transferred to one of five different media (9–14): 1) FGM (Clonetics), 2) DMEM/F12 with 10% FBS (M1 medium), 3) serum-free DMEM/F12 supplemented with N2 ((M2 medium), 4) M2 medium supplemented with 10 ng/ml bFGF (R&D Inc, Minneapolis, MN), 1μg/ ml laminin (Sigma) and 2μg/ml heparin (Sigma, M3 medium) for 12 days, and 5) two-step induction medium (i.e., cells were primed with M2 medium for six days, followed by M3 medium for six days). N2 supplements (Gibco, Grand Island, USA) consisted of 5μg/ml insulin, 10μM progesterone, 10μM putrascine, 10μM sodium selenite, and 100μg/ml transferrin. Long-term spheroid expansion was accomplished via cell culture in bacterial dishes containing M3 medium. Cell morphology and spheroid formation were examined under a phase contrast microscope (Olympus, Japan).

Cells cultured on cover slips were fixed in 3.7% form-aldehyde (Sigma), permeabilized with 0.2% Triton X-100 (USB, Cleveland, OH) for 10 minutes, blocked with 20% normal goat serum (Vector Lab, Burlingame, CA) for 1 hour, incubated with primary antibodies for overnight at 4°C followed by FITC- or Texas Red-conjugated IgG (Vector Lab, 1:100) for 1 hour at room temperature, and then stained with 4, 6-diamidino-2-phenylindole (DAPI) for 10 minutes. Sections were mounted with Vectashield and then examined under a Leica confocal microscope (Leica, Germany). Diaminobenzidine (DAB) staining of cells on culture dishes was performed by blocking endogenous peroxidase activity for 30 minutes in the presence of 0.3% H₂O₂ (Sigma) in methanol. The staining area was then cleaned with a cotton swab and demarcated using a paraffin pen and permeabilized with 0.2% Triton X-100 for 10 minutes, blocked with 20% normal goat serum for 1 hour, incubated with primary antibodies for overnight at 4°C. Subsequently, the cells were incubated with a biotinylated secondary antibody for 1 hour and an avidin-biotin peroxidase complex (Vector Lab) for 30 minutes. The final reaction for peroxidase was carried out with the DAB peroxidase substrate kit (Vector Lab). Cells were mounted under a coverslip with Aquatex (Merck, Whitehouse Station, NJ) and examined under a light microscope (Olympus). The spheroids staining were accomplished by impregnating the fixed spheroids with neutralized type I collagen (BD, Bedfold, USA) and embedding in a paraffin block. Paraffin sections (4μm) were deparaffinized, rehydrated, and then boiled in 0.01 M sodium citrate. The remainder of the staining procedure was performed as described above. Primary antibodies consisted of monoclonal antibodies against nestin (Chemicon, Temecula, CA, 1: 200), nestin (Biodesign, Saco, ME, 1:200) and β-III tubulin (Sigma, 1:400), vimentin (DAKO, Glostrup, Demark, 1:50), A2B5 (Chemicon, 1:200), O4 (Chemicon, 1:50), GFAP (DAKO, 1:50), NF-68 (Sigma, 1:400), NF-150 (Chemicon, 1:400), MAP2 (Chemicon, 1: 1,000), Tau (abCAM, Cambridge, MA 1:500) and polyclonal antibodies against p75 (Promega, Madison, WI, 1: 500).

All data are reported as mean±SD. Differences were assessed using the Student’s t-test. Values of P<0.05 were considered statistically significant.

SKPs were initially identified based upon nestin expression and sphere-forming capacity in scalp skin after long-term culture. These cells are reported to originate from residual embryonic neural crest-related precursor cells in the adult skin (4–8). Therefore, to determine whether nestin-expressing cells are endogenous to the newborn foreskin, or if they are induced by specific culture environments, formalin-fixed or frozen sections of newborn foreskin or adult scalp were stained with two different antibodies capable of recognizing the nestin protein in neuroblastoma SH-SY5Y cells (15–17). Sections were also stained using antibodies against vimentin as a positive control for HDFs. Nestin-expressing cells were not detected in the dermis of either newborn or adult foreskin (data not shown). However, vimentin-expressing HDFs were clearly observed (Fig. 1A).

To further explore the possibility of epigenetic trans-differentiation among vimentin-expressing HDFs, the cells were isolated, divided into five groups, and cultured as described in Materials and methods. DAPI-positive cells expressed vimentin under all five culture conditions, while nestin expression differed among the groups (Fig. 1B). Nestin-expressing cells were detected with weak intensity in FGM, M1, and M2 media, while their numbers increased in M3 medium. However, cells grown in the two-step induction medium strongly expressed nestin, as shown in Fig. 1C. Approximately 80% of the HDFs cultured in the two-step induction medium became nestin-expressing cells within 12 days. HDFs, which do not express neural precursor markers such as nestin in vivo, rapidly began to express nestin under specific culture conditions (e.g., growth in M3 medium). Absence of nestin-immunoreactivity in foreskin tissue sections, combined with the specific induction of nestin expression in two-step cultures, strongly suggest that nestin-expressing cells are induced by components of the M3 medium. It also suggests that this induction can be strongly stimulated by priming the cultures in serum-free M2 medium.

To further characterize the nestin-expressing cells generated using our two-step induction method, cells were stained via double-immunofluorescence labeling using antibodies against either nestin and fibronectin, or nestin and β-III tubulin (Fig. 1D). Following two-step induction, nestin-expressing cells strongly expressed FN and β-III tubulin, both of which are markers for immature neuronal cells. Therefore, two-step induction of HDFs stimulated the expression of markers specific for neural pre-cursor cells, as well as immature neurons such as nestin, fibronectin, and β-III tubulin.

To determine if nestin-expressing cells are capable of forming neurospheroid-like structures, cells primed in M2 medium were re-plated on poly-D-lysine/laminin-coated dishes in M3 medium (Fig. 2A). Spheroids did not form in M2 medium. However, cells exposed to M3 medium began to converge centrally after three days and formed spheroids after six days. Spheroids continued to grow for up to 40 days and either floated in suspension, or adhered to the bacterial culture dishes (Fig. 2B). Split spheroids reformed to generate spheroids (data not shown).

To characterize dermal cells that either attached to the surface of the culture dish or floated as spheroids, cells were fixed and stained using antibodies against nestin and vimentin (neuronal precursor cells), β-III tubulin (immature neuron cells), A2B5 (oligodendrocyte progenitor cells), p75 (neural crest stem cells), O4 (pro-oligodendrocytes), GFAP (astrocytes), and NF-68, NF-150, MAP2, and Tau (mature neurons) (Fig. 3A, B). Both attached cells and spheroids expressed p75, β-III tubulin, and A2B5, but did not express NF-68, GFAP, or O4. Therefore, spheroids induced via the two-step method satisfied most of the previously reported characteristics of SKPs (4–6).

To identify which factors in M3 medium are responsible for nestin and spheroid formation, cells were incubated in M3, both with and without bFGF or laminin, after priming the cultures in serum-free M2 medium. Following incubation, spheroid formation and nestin expression were examined (Fig. 4A, B). Spheroids were successfully formed in M3 medium supplemented with bFGF, regardless of the presence of laminin. However, spheroids were not formed in the absence of bFGF, even when laminin was supplied (Fig. 4A). Similarly, nestin expression also required bFGF (Fig. 4B). Therefore, these results suggest that bFGF induces nestin and spheroid formation in HDFs.

Purified HDFs can be induced to form spheroids that express markers for immature neurons and neuronal precursor cells by first priming in serum- and bFGF-free, followed by incubation in bFGF-containing medium. Previous reports suggest that MSCs have the potential to transdifferentiate into neuronal types capable of aiding in recovery after injury to the central nervous system (3). Therefore, we used our two-step induction method on both human MSCs and keratinocytes. This method did not induce nestin expression or spheroid formation in either of these two primary cultured cell types (Fig. 5). Thus, two-step induction of nestin-expressing spheroids seems possible only for HDFs in the presence of bFGF.