An improved method of human keratinocyte culture from skin explants: cell expansion is linked to markers of activated progenitor cells


Colin Jahoda, PhD, Department of Biological Sciences, University of Durham, South Road, Durham, DH1 3LE, UK, Tel.: 0191 334 1338, Fax: 0191 334 1201, e-mail:


Abstract:  Human keratinocyte primary cultures are commonly established by tissue dissociation and often rely on feeder cell supports and culture medium that is not defined. Further, contamination by unwanted fibroblasts can be problematic. Here, we developed a skin explant method for growing primary keratinocytes that was rapid, simple, and reliably generated keratinocyte cultures free of fibroblast contamination. The process capitalized on the observation that fibroblasts migrate out of adult skin explants later than epidermal cells, allowing the early harvesting of keratinocytes by trypsinization. When grown subsequently in defined medium in the absence of feeder cells, the explant-derived cells grew rapidly and could be cultured for multiple passages. Immunofluorescence microscopy revealed that a high percentage of cells harvested from the explant outgrowths expressed K15, while very few expressed the differentiation marker K10. Cells that were stained while migrating out from explants strongly expressed markers associated with progenitor cells, including p63, K15 and CD133, and displayed intense K6 expression, indicative of activated keratinocytes in wound-healing epidermis. By replenishing the explants with fresh medium after harvesting, further epidermal outgrowths could be obtained, offering the possibility of greatly increased keratinocyte yields for clinical applications.


A reliable source of cultured keratinocytes is essential as a component of skin substitutes to treat burns and wounds and for laboratory testing. The epidermis is one of only a few tissues from which it is possible to propagate its principal cell type (the keratinocyte) and to use these cultured cells to reconstitute a stratified and fully differentiated human tissue in vitro (1–4). As Medawar (5) successfully separated a pure epidermal sheet from human skin by trypsinization, it has been possible to readily obtain epidermal cells for expansion in tissue culture. In the years that have passed, continuing improvements have been made on the original culture method. In particular, Rheinwald and Green (6) used a feeder layer of X-irradiated 3T3 cells to support keratinocytes in culture. In this method, however, the activation of 3T3 cells must be kept under control, and contamination must be avoided. Kitano and Okada (7) modified the process by introducing a milder protease known as dispase to separate the epidermal sheet from the underlying dermis of the skin. Then, Boyce and Ham (8) adopted a serum-free medium for primary keratinocyte culture, in which the 3T3 feeder layer is no longer needed and therefore has benefits for use in clinical applications. Among several varieties of keratinocyte serum-free medium currently available, Epilife™ (Invitrogen Ltd, Paisley, UK) is one commercial form which we employed for this study. This medium contains various human recombinant growth factors and has a low calcium concentration at 0.06 mM.

While the dissociation method of keratinocyte primary culture is well established, attempts to acquire purified adult stem-cell like/progenitor keratinocytes from whole human skin are ongoing in many laboratories. In particular, different techniques are currently used to achieve high purity or homogeneous primary cultures enriched for keratinocyte progenitor/stem cells. These include filtration, density gradient centrifugation, fluorescence-activated cell sorting with cell surface antibodies as well as differential adhesion to enrich for cells that rapidly attach to particular substrates (9). However, despite the efforts to characterize and purify keratinocyte cells from intact human skin, there still is a need for improved reliable techniques to isolate high quality progenitor keratinocytes and propagate them in culture, preferably in the absence of a feeder layer.

Cultures of adult skin explants have been used to model adult skin epidermal growth and behaviour (10–12). A similar culture method has been previously used with cervical uteri explants to test factors affecting serial cultivation of cervical epithelial cells (13). However, as fibroblasts grow out from these explants, it is difficult to separate the fibroblasts from the keratinocytes once the two cell types are mixed. Interestingly, previous work has suggested that the keratinocytes which migrate out from whole skin explants apparently undergo little or no cell division over the first few days in culture (14). Moreover, transmission and scanning electron microscopy studies (15) have shown that these early migrating cells originate from the basal layer of the epidermis.

It has previously been shown that fibroblasts do not grow out from adult human skin explants until several days after the appearance of keratinocytes. We have taken advantage of this time lag between the migration of keratinocytes versus fibroblasts to develop a quick and easy method to produce keratinocyte cultures from skin explants. This method avoids fibroblast contamination, provides rapid cell growth and offers great simplicity and adaptability to specific experimental needs. In addition, we examined the expression of molecular markers in these cells during early migration. We found that cells in the outgrowth, expressed a range of progenitor and wound healing markers, suggesting they are potentially enriched for activated progenitor cells. Explant-derived keratinocytes could be grown rapidly to multiple passages using any of the current methods of culture, and importantly, the original explants could be recycled and used as a continuing source of keratinocytes.


Keratinocyte cell culture from explants

Primary cultures of human keratinocytes were established using discarded healthy skin from patients of both genders and various ages with informed consent and Ethics Committee approval to CABJ. (C. Jahoda, NHS; Ethics submission version 1, 23/9/2005). Skin samples were soaked in minimal essential medium (MEM, M4655; Sigma–Aldrich, Dorset, UK) with double strength antibiotics (1.250 μg/ml amphotericin B, 200 IU/ml penicillin and 200 μg/ml streptomycin) overnight at 4°C. The skin was then flattened in a 60-mm sterile Petri dish with the epidermis facing up in fresh MEM (above). Strips of split thickness skin about 2-mm wide were obtained by using sharp scissors and fine forceps (Fig. 1a). Each strip was then cut into pieces between 1 and 2 mm in length placed onto the bottom of 35-mm diameter culture dishes [(Primaria; Falcon, Becton Dickinson, Oxford, UK) or Nunclon™ Surface, Nalge Nunc International, Roskilde, Denmark] at a density of around 30 pieces/dish (Fig. 1b). At this stage, it was imperative that the explants were oriented with the epidermis facing up. The culture dish with explants were incubated at 37°C in 5% CO2 for 30 min to 1 h (according to the amount of liquid associated with the specimens) to secure attachment of the explants to the culture dish. At this point, the specimens were still moist. About 1.5–2 ml of the MEM containing 20% fetal bovine serum (FBS, F7524; Sigma) and antibiotics (0.625 μg/ml amphotericin B, 100 IU/ml penicillin and 100 μg/ml streptomycin) were then carefully added to cover the explants and the dishes were returned to the incubator. Cultures were checked for cell growth the next day. On day 3 or 4 of culture the cells were subcultured. The culture dishes with explants were washed with calcium- and magnesium-free phosphate-buffered saline (PBS), and then incubated with 0.25% Trypsin–EDTA (Invitrogen Ltd, Paisley, UK) until the sheets of cells surrounding the explants had been dissociated. The trypsin was inactivated with serum-containing medium, and the cell suspension was spun at 1000 rpm for 5 min. The cell pellet was resuspended in Epilife™ growth medium (Invitrogen Ltd, Paisley, UK) supplemented with human keratinocyte growth supplement (Invitrogen Ltd, Paisley, UK, 5 ml/500 ml) and seeded into T25/T75 flasks. Routinely, the explants were re-fed with fresh medium of MEM containing 20% fetal calf serum (FCS) and incubated further to check their further capacity for keratinocyte growth. Keratinocyte cultures established from explant cultures were serially subcultured in Epilife™ medium. To examine whether keratinocytes produced from explants were capable of normal growth under conventional conditions, some cultures were transferred to Rheinwald and Green medium (6) and grown on 3T3 feeder layers.

Figure 1.

 Preparation of explants of human skin for keratinocyte culture and keratinocyte outgrowth from skin explants. Strips of skin were obtained by catching the tissue with fine forceps and then cutting with sharp-angled scissors. The depth of cut was below the basal layer of epidermis but relatively shallow in relation to the dermis (a). Each strip was cut into pieces 1–2 mm in length and placed in the bottom of a 35 mm diameter culture dishes at a density of 20–30 pieces per dish (b). Cells started growing out of the explants between 24 to 36 h (c). Outgrowths resembled a ring and continuously expanded surrounding the explants by day 3–4 (d). Cells with fibroblast or melanocyte morphology were first observed beyond the fringe of the keratinocyte outgrowths on day 5 (e). Early growth comparison between keratinocytes harvested from explant culture. Light microscopic views of cells subcultured from explants (after 4 days) and grown for a further 1 day (f), 2 days (g) and 6 days (h). Expression of K5, K10 and K15 in passage 1 cells derived from explant culture. Dual immunofluorescent labeling was performed for K5 (green) and K10 (red) in (i) and for K5 (green) and K15 (red) in (j). Thus, yellow colour indicates areas of K5/K10 or K5/K15 colocalization. 4′-6-Diamidino-2-phenylindole (DAPI) was labeled as in blue. Repeated keratinocyte outgrowth from skin explants. Continuous keratinocyte outgrowth was observed after resupplying the explants with medium and returning dishes into culture under the same conditions as before. Cell outgrowth at 2 days from a recycled explant has a similar but faster growth pattern to the initial culture. Both the inner (k) and outer (l) regions of a keratinocyte outgrowth are free of fibroblasts. Keratinocyte cells harvested for a third time from the same explant growing in Epilife™ 3 days after subculture (m). Scale bars: 60 μm.

Antibodies and immunofluorescent microscopy

To investigate the expression profiles of keratinocyte cells, the cells were subcultured and grown for a 3-day period in Epilife in 35-mm culture dishes. To study patterns of expression during initial explant outgrowth, the cultures of skin explants from breast and abdominal sites were stained after 1 and after 3 days in culture in MEM containing 20% FBS. Briefly, the culture dish with cells/explants were washed with PBS and then fixed with 95% ethanol and 5% acetone for 10 min, and again washed in PBS before blocking with10% donkey serum (D9663, Sigma) in PBS. Cells were then incubated with primary antibodies overnight at 4°C followed by Alexa-Red or fluorescein isothiocyanate-conjugated secondary antibodies with 4′-6-diamidino-2-phenylindole at room temperature for 2 h. The dishes were then washed and mounted.

The primary antibodies included: human basal epidermal marker K5 (goat anti-cytokeratin 5, Santa Cruz Biotechnology, Inc, Heidelberg, Germany); suprabasal and differentiating keratinocyte marker K10 (mouse anti-cytokeratin 10 monoclonal antibody; Chemicon International, Herts, UK); α6β1integrin (mouse anti-rat monoclonal antibody; TCS Biologicals Ltd, Buckingham, UK); cytokeratin15 (mouse monoclonal keratin 15 antibody; Labvision, Cheshire, UK); CD34 (mouse monoclonal to CD34; Abcam, Cambridge, UK); CD133 (purified mouse monoclonal; Abgent, Abingdon, UK); K6 (mouse monoclonal to cytokeratin 6; Abcam, Cambridge, UK); P63 (mouse monoclonal, clone 4A4; Labvision, Cheshire, UK); DSG3 (mouse anti-human monoclonal to desmoglein 3, gifted by Dr Jim Wahl, University of Nebraska, Nebraska, USA). The secondary antibodies used were Alexa Fluor 594 donkey anti-mouse IgG and Alexa Fluor 488 donkey anti-goat IgG (both from Invitrogen). Cells and tissue samples were then examined and images obtained using a Zeiss Axio Imager.M1 (Zeiss, Oberkochen, Germany) fluorescence microscope (or Zeiss LSM 510 Confocal microscope from Carl Zeiss, Germany). To investigate expression in explant-derived cells after subculture, cells were stained with K5, K10 and K15 antibodies as described above and images were captured at low (10× lens) magnification. The percentage of positively expressing cells in each of four randomly chosen fields of view was then counted, and a mean percentage figure obtained.


Keratinocyte cell outgrowth from skin explants – growth, expression and maintenance after passaging

Keratinocytes were first observed growing out of the explants in a continuous sheet between 24 and 36 h. By day 2, the outgrowth resembled a ring surrounding the explants (Fig. 1c) and over day 3 and 4 (Fig. 1d) this outgrowth continued to expand. After 5 days in culture, cells with spindle-shaped fibroblast or melanocyte morphology were visible outside of the keratinocyte ring (Fig. 1e). Sufficient numbers of cells for harvesting were produced by 30 explants in 3–4 days. Comparison of numbers of cells obtained from six individual 35-mm dishes derived from skin explants from a single donor revealed a mean of 1.6 ± 0.36 × 105 cells per dish. As an estimate, therefore, 1cm2 of skin would yield approximately 5.45 × 105 cells on the first harvest. When subcultured in Epilife™, these cells exhibited exemplary attachment and growth capabilities. After one day, cells were attached and were growing (Fig. 1f), and between 2 and 3 days, colonies larger than 32 cells were common (Fig. 1g). Importantly, no fibroblast contamination was observed in these cultures. Confluent keratinocyte monolayers were generated 5 or 6 days after passaging (Fig. 1h). We next characterized explant-derived keratinocytes with specific markers after subculture. Cultures had high cytokeratin 5 expression, but less than 5% of cells expressed the differentiation marker keratin 10 (Fig. 1i), while a much higher percentage (43%) expressed the basal marker keratin 15 (Fig. 1j). Levels of K10 staining increased slightly after passaging but remained low (data not shown). Explant-derived cultures have been obtained from adult specimens of different ages and sex and various body sites, including haired regions incorporating terminal hair follicles. Cells initiated from explants and cultured in Epilife™ were routinely grown to passage 5 and beyond, and one cell strain was grown up to passage 10 (Table 1) without evidence of differentiation. Several of the cultures were also frozen and successfully recovered. The K5, K10 and K15 expression profile described above remained stable when cells were stained after four passages, following cryopreservation and recovery (data not shown). We also tested if explant-derived keratinocytes could be switched to conventional Rheinwald and Green (6) growth medium with 3T3 support layers. With a mitomycin-C-treated 3T3 feeder layer, the keratinocytes exhibited a similar growth pattern as in Epilife™. Cells settled after 1 day produced colonies by 3 days, reached confluence at around 9 days and could be subcultured successfully (data not shown).

Table 1.   Details of serial cultivation of human adult skin specimens after explant culture
  1. p, passage number; C, cryofrozen; R, recovered and cultured further; E, used for experimental analysis; O, ongoing culture.

Age (years)68Na436755272543444244
Donor siteBeardBeardBeardLegBreastBreastBreastBreastBreastAbdomenScar tissue
Longevityp10C at p3
R to p6
p7 Op 9C at p4
R to p7
C at p2E for stainingC at p2C at p1E for stainingp6 O

We then investigated the potential of the original to produce more keratinocytes. Interestingly, those explant cultures that were resupplied with medium in the same dish after enzymatic removal of the initial keratinocyte outgrowth produced a second, robust outgrowth of epithelial cells. These cells migrated out in a similar manner and even more rapidly than the first outgrowth, and could also be harvested by trypsinization after 2–3 days, at which point the cultures were still free of fibroblast contamination (Fig. 1k–m and Video Clip S1). The keratinocytes obtained could be subcultured and showed a similar growth pattern to those from the first outgrowths. We have been able to repeat this process up to five times from the same original explants, although some fibroblast contamination was observed on the third and subsequent outgrowths (data not shown).

Keratinocytes from initial explant outgrowths express stem cell and wound healing markers

We then examined what markers cells were expressing as they grew out from the explants. After 24 to 36 h, dual staining revealed that the cells were K5 positive, and co-expressed K15 throughout. In contrast, cells co-expressing K5 and K10 were restricted to the innermost cells of the outgrowth, closest to the explants, while no α6β1 integrin labeling was observed (data not shown).

At 3 days cultured explants, cells on the leading edge strongly expressed K15, while those closer to the explant were weakly labelled (Fig. 2a). Meanwhile, K10 expression was present but was restricted to a relatively small circle of cells closest to the explant (Fig. 2b). α6β1 Integrin was expressed by cells throughout the outgrowth, but a distinctive inner ring of very brightly labelled cells was visible close to the explant (Fig. 2c). All the remaining markers labeled the cell outgrowths positively but showed different patterns of expression. When stained with the p63 antibody, cells throughout the outgrowth showed distinct nuclear staining, with the strongest labeling located towards the leading edge (Fig. 2d). Desmoglein 3 and CD34 were both expressed strongly in cells through most of the outgrowths apart from cells at the leading edge (Fig. 3a and c). CD133 labeling was strongest in a band closest to the explant (Fig. 3b) in a pattern similar to that shown by α6β1 integrin (Fig. 2c). In contrast, fluorescence labeling with the antibody to K6 was also widespread but strongest towards the outer edge of the outgrowths (Fig. 3d) more akin to the expression seen with K15 (Fig. 2a). A schematic diagram summarizing the localization and strength of marker expression in keratinocyte outgrowths at 3 days is shown in Fig. 3e.

Figure 2.

 Expression of K5, K10 and K15, α6β1 integrin and P63 markers by explant outgrowth cells after 3 days in culture. Dual immunofluorescent labeling was performed for K5 (green) and K15 (red) in a; for K5(green) and K10(red) in b; K5 (green) and α6β1 integrin (red) in c; for K5 (green) and P63 (red) in d. Thus, yellow colour indicates areas of K5/K10; K5/K15 co-localization; K5/α6β1 integrin and K5/P63 co-localization. 4′-6-Diamidino-2-phenylindole was labeled as in blue. The purple arrows show the direction of outgrowth, and E indicates the location of the explants. Scale bar: 30 μm except c2 (15 μm).

Figure 3.

 Expression of K5, K6, CD133, CD34 and DSG 3 markers by explant outgrowth cells after 3 days in culture. Dual immunofluorescent labeling was performed for K5 (green) and DSG3 (red) in a; for K5 (green) and CD133(red) in b; K5 (green) and CD34 (red) in c; for K5 (green) and K6 (red) in d. Thus, yellow colour indicates areas of K5/K6; K5/CD133; K5/DSG3 or K5/CD34 co-localization. 4′-6-Diamidino-2-phenylindole (DAPI) was labeled as in blue. The purple arrows show the direction of outgrowth and E indicates the location of the explants. Scale bar: 60 μm.


Culturing epidermal cells from both animal and human skin has traditionally involved one of two approaches: splitting of the epidermis from the dermis followed by epidermal cell dissociation or explant culture. As Rheinwald and Green (6) first described a protocol to achieve single keratinocytes from skin epidermal sheets, this strategy has become the method of choice by most practitioners. Much human work has been performed on new born foreskin, whose cells have a very different replicative profile to keratinocytes from older skin (16). Here, we developed a skin explant method for primary keratinocyte culture from adult human skin that was quick, simple and reliably generated keratinocytes without fibroblast contamination. The cells we produced by this means had robust growth characteristics. This led us to further define the nature of explant outgrowth, in which we observed the presence of putative stem cell and wound healing markers in the cell outgrowths.

Explant culture wherein a small piece of skin will settle on a culture dish and produce a sizeable outgrowth of cells has long been employed as a model of wound healing or adult skin epidermal outgrowth rather than a source of keratinocytes for clinical or experimental purposes (10–12,17–20). The major limitation is due to the fact that fibroblasts also grow out from the same explants, and will eventually outgrow the keratinocytes. In our study, keratinocytes emanated from tissue explants after approximately 24 h, which is consistent with some previous reports (17,21). However, rather than let them continue until fibroblasts began to appear at about 5 days, we removed the epidermal cells from the dish after 3–4 days as an alternative method of obtaining keratinocyte primary cultures free of fibroblasts. As no fibroblast outgrowth was observed until at least 5 days, it was possible to obtain fibroblast-free populations of keratinocytes.

What is the biological basis underpinning the robust growth characteristics and longevity of explant-derived keratinocytes? Previous work has suggested that outgrowing epidermal cells have relatively little thymidine incorporation during the first few days in explant culture. This was interpreted as meaning that migration, rather than cell division, is the main mechanism of outgrowth in early explant cultures (15). Importantly, transmission and scanning electron microscopy by the same group determined that the outgrowing cells originated from the basal layer of the epidermis which conventionally is believed to contain the stem and transient amplifying cell populations. The basal epidermis is conventionally believed to contain at least two distinct subpopulations of keratinocyte progenitors: keratinocyte stem cells, which constitute a minor subpopulation of relatively quiescent or slow-cycling cells with great proliferative potential and an unlimited capacity for self-renewal, and transit amplifying (TA) cells which are the progeny of the stem cells and are believed to have a limited proliferative capacity (22–24). In mature epidermis, the undifferentiated stem cells and proliferative TA cells are both contained in the basal layer. However, recent appreciation of plasticity/multipotency in a number of tissues has lead to a blurring of the distinction between TA and stem cells, and tissues of the quiescent nature of stem cells (25–28). This concept probably applies even more in the context of wound healing epidermis where the environment is transformed. As the explant cultures initially migrate out maintaining contact with each other and with the underlying basement membrane, we postulate that the outgrowing cells are undergoing an active process of self-selection of cells with high replicative potential. By preserving the basement membrane, it is likely that the keratinocytes recruited in the outgrowths emanate from both the stem cell and TA cell compartments, and are mobilized in a way that mimics ‘activated keratinocytes’ (29) during the wound healing process in vivo.

This was reflected in our short-term cultures showing that outgrowing keratinocytes expressed not only stem cell-associated markers, including p63 (30) and K15 (31) but also expressed K6, an early marker of wound healing epidermis (32,33). Strong expression of α6 (34) and β1 (35) integrins have separately been identified as putative stem cell markers. For the current state of the art, high α6 but not β1 is a good indicator for stemness. However, the sublocalization of α6β1 expressing cells within the explant outgrowths may also be a reflection of differences in their motility status as the brightly labelled cells were those close to the explant and not the migratory cells at the periphery. Downregulation of bright desmoglein 3 expression, which is also a putative marker of keratinocyte stemness (36,37) was not observed. Quite surprisingly, the keratinocytes were also labelled by CD34 and CD133 antibodies neither of which has been observed in this context. The CD34 antibody clone that we used has previously been shown to delineate cells in the outer root sheath of the human hair follicle (38) but not interfollicular epidermis. CD133 labels epithelial stem cells in the prostate (39) but only one isoform AC133-2 has been shown to be expressed in cultured keratinocytes (40). After subculture, the explant-derived cultures continued to have few K10 positively expressing cells, and expressed high levels of K15 – which may be an indicative marker of progenitor cells (41).

In conclusion, we show that keratinocyte cells obtained from explant culture display progenitor markers and grow for multiple passages when transferred to serum-free culture. From the clinical and practical standpoint, this method provides a simple means of growing large numbers of keratinocytes from only a small biopsy quickly and in the absence of feeder layers, which could be important in relation to patients with limited donor skin availability. Elegant strategies can be used to isolate and grow enriched stem and TA cells from dissociated epidermis, however, these usually involve cell sorting (34,42), and are thus more time consuming and require equipment that might not be readily available in all laboratories, or in clinical settings where simple, rapid isolation of cells with robust growth characteristics is a priority. Although the method reported here currently involves FBS for a few days, we have confirmed the possibility of replacing FCS with human serum if required in a clinical context.

The finding that serial outgrowths of keratinocytes could be harvested from the same explants further enhances the utility of this method as it provides the possibility of obtaining much larger numbers of keratinocytes from a single source. It also has parallels in split thickness grafting where a donor site can be repeatedly used, highlighting the capacity of the wound stimulated epidermis activity to serially renew itself. Finally, this observation is significant in relation to dermal–epidermal interactions in the skin explants. The absence of fibroblasts during the second or subsequent keratinocyte outgrowth suggests that fibroblast activation or motility is being inhibited by the epidermal cells by a process worthy of future investigation.


We are very grateful to the British Skin Foundation and the MRC (grant number G0300353 to CABJ) for support. We are very thankful to Mr Clifford Lawrence, Dr Reika Taghizadeh, Mr Andrew Owens and Mr Martin Coady for their invaluable help with the provision of skin samples; Dr Trevor Booth for assistance with some confocal microscopy; and Dr James Waters for his helpful suggestions on the manuscript.