Establishment of a murine epidermal cell line suitable for in vitro and in vivo skin modelling

Keratinocytes were isolated from the back skin of 2-months old C57BL/DBA mice using a protocol suitable for adult skin [19]. They were maintained in culture in a low calcium (0.07 mM) progenitor cell targeted fully defined media, CnT-07 (CELLnTEC, Bern, Switzerland), in an incubator at 37°C and 5% CO₂. At confluency, cells were subcultured by trypsinisation and plated at 4.5-5.5 × 10³ cells per cm² in seeding media (EMEM, 4% FBS pre-treated with Chelex 100 resin, 0.2 mM CaCl₂) for 8-14 hours to allow cell adhesion. Then, culture plates were washed twice with PBS to eliminate calcium, and switched to CnT-07 to promote cell growth and prevent differentiation. FBS-seeding media is not further needed once the cell line has been established (>20 passages). At this point, 8-14 hours incubation in CnT-07 medium containing 0.2 mM CaCl₂ is enough to ensure good cell attachment after trypsinisation.

Conventional in vitro keratinocyte differentiation was initiated when cells were confluent by changing the low calcium (0.07 mM) media to a high calcium (1.2 mM) differentiation media. This elevation in the concentration of calcium induces the cells to differentiate [20‐22]. Bromodeoxyuridine 10 μM (BrdU) (Sigma Aldrich, St. Louis, MO) was added to the culture medium 1 hour before sample fixation.

All the animal work was approved by the Animal Ethical Committee (CEEA) and conducted in compliance with Centro de Investigaciones Energéticas, Medioambietales y Tecnológicas (CIEMAT) guidelines. NMRI nude (nu/nu) immunodeficient 5-week old female mice were obtained from Janvier (Saint-Berthevin, France) and housed in our animal facilities for 1-2 weeks prior to any experimental procedures. Mice were injected intraperitoneally with 100 μg of BrdU (dissolved in 0.09% NaCl) per g of body weight 1 hour before they were sacrificed to analyse proliferation (BrdU incorporation).

To assay tumour-formation ability, 1 × 10⁶ COCA cells in late passages (>70) were resuspended in 100 μl saline buffer and injected subcutaneously on the flanks of NMRI nude mice. Both flanks were used, in one COCA were injected alone, in the other they were combined with 0.5 × 10⁶ dermal fibroblasts. The experiment was repeated three times using 5 mice (x2 flanks) per experiment. Tumour growth was not evident, and animals were sacrificed and studied at 4-6 months to search for any indication of tumour.

To assay the ability of COCA cells to engraft we followed a protocol adapted from Strachan and Ghadially [23]. Briefly, mice were anesthetised and an excision approximately 5 mm in length was performed in the dorsal skin just between the shoulder blades. Silicone chambers with a 7 mm internal diameter (Renner, Dannstadt, Germany) were implanted and sutured in place when required. COCA epidermal keratinocytes (1.5 × 10⁶) in passage 77 were mixed with dermal fibroblasts (0.5 × 10⁶), resuspended in a final volume of 100 μl PBS and injected into the silicon chambers. The cells were allowed 20 min to settle and adhere before waking up the mice. Epidermis formed by grafted COCA + fibroblasts was harvested 4 weeks after chamber implantation. Mice were injected i.p. with 100 μg/g of body weight BrdU 1 h before they were sacrificed to analyse proliferation.

For histological analysis, cells and tissue samples were fixed in 4% paraformaldehyde and embedded in paraffin prior to sectioning. Sections 3-4 μm thick were stained and processed as described [24]. For BrdU detection, prior to blocking with serum, samples were incubated for 30 min in 1 N HCl (cells) or 1 h in 2 N HCl (tissue). Primary antibodies used were rabbit polyclonal anti-K5, anti-K6, Loricrin, Filaggrin (Covance, Emeryville, CA), and anti-BrdU (AbCam, Cambridge, UK); and mouse monoclonal anti-K10 (for cells clone RKSE60, St. Cruz Biotechnology, Santa Cruz, CA; for tissue sections DE-K10, DakoCytomation, Denmark A/S) and anti-PCNA (clone PC10 Thermo Scientific, Fremount, CA). FITC or TexasRed conjugated secondary antibodies were purchased from Jackson ImmunoResearch (West Grove, PA). Control slides were obtained by replacing primary antibodies with PBS or preimmune sera.

Keratinocytes were isolated from the back skin of 2-months old C57BL/DBA mice and maintained in culture in low calcium (0.07 mM) fully defined media (CnT-07). Cells were repeatedly passaged to establish a cell line. We termed this murine epidermal keratinocyte derived cell line COCA, and we have expanded it in culture for more than 80 passages. COCA cell line has been deposited (October 2010) under the reference 10112001 with ECACC (European Collection of Cell Cultures, http://​www.​hpacultures.​org.​uk/​products/​celllines/​generalcell/​detail.​jsp?​refId=​10112001&​collection=​ecacc_​gc).

Cytogenetic studies using the spectral karyotype (SKY) (Table 1 and Additional file 1) revealed that a translocation between chromosomes 2 and 4 occurred early (already present in passage 8 cells) and is maintained. Murine epidermal keratinocytes in culture have a tendency to became tetraploid [14, 18]; however, COCA showed to be genetically more stable than previously published cell lines [18] and was still diploid at passage 8 (Table 1).

The epidermis is a multilayered epithelium where the process of terminal differentiation takes place as the committed cells in the basal proliferative compartment arrest proliferation and move upwards towards the epidermal surface. This process is accompanied by changes in the expression of keratins and other proteins. Keratin 5 (K5) is expressed in the basal layer, where proliferation occurs, and marks undifferentiated cells. Once the cells start the differentiation program they cease the expression of K5 and start the synthesis of the keratin K10, which is considered as an early marker of differentiation in vivo and in vitro. Terminal differentiation is indicated by the expression of loricrin and filaggrin. The keratin K6 is induced in the interfollicular epidermis as consequence of inflammatory and/or hyperproliferative stimuli, and in many circumstances this is in parallel with reduced K10 expression and expansion of K5 to suprabasal layers.

Keratinocytes can be induced to differentiate in vitro by increasing the concentration of calcium in the culture medium [20‐22]. To evaluate the potential of COCA to differentiate we replaced the low calcium medium (0.07 mM) for a differentiation medium containing a high concentration of calcium (1.2 mM). As early as 24 hours in 1.2 mM calcium differentiation medium, keratinocytes start to flatten (Figure 1A, upper row panels) and to establish cell-cell contacts. Upon calcium elevation, COCA mouse keratinocytes in all passages studied (passage 21, 51 and 76), displayed an arrest in proliferation (Figure 1B) and a similar pattern of differentiation marker expression: K10 expressing cells are detected at 24-48 hours and loricrin positive cells at 48-72 hours in differentiation media (Figure 1A, immunofluorescence panels). Cells from passage 51 and 76 growing in low calcium medium displayed an increased proliferation rate as compared to that of passage 21, this could be due to further adaptation of the cells occurring between passage 21 and 51. Still, cells at all passages tested readily responded with an arrest in proliferation when switched for 24 hours to a high calcium media. Most importantly, after 72 hours in differentiation medium, cells were re-stimulated for 24 hours in the presence of low calcium growth medium and, independently of the passages studied, expression of terminal differentiation markers was achieved and BrdU incorporation was minimal, demonstrating that differentiation is associated to permanent cell cycle withdrawal (Figure 1A, right column panels, and Figure 1B).

Although useful, the above described in vitro differentiation model has important limitations; in this sense, 3D models resemble the in vivo situation much more closely. In these systems, the keratinocytes are organised in a multiple layered structure, with a proliferating basal layer that stratifies as it matures and differentiates. We have followed a 3D epidermal in vitro model that avoids the use of supporting cells, collagen matrix and non-defined cultured media. Cell cultures of COCA in passage 76 were induced to differentiate by creating an air/liquid interface. After three weeks a full epidermis-like structure was formed (Figure 2A). K5 was expressed in basal and suprabasal layer cells while K6 and K10 were only expressed by suprabasal layers (Figure 2B-E). Suprabasal layers express the terminal differentiation markers loricrin and filaggrin (Figure 2D&2E). Proliferating cell nuclear antigen (PCNA) staining shows that the layer of proliferating cells that maintain the architecture of the culture locates at the base (Figure 2F). Limitations to this 3D model affect K10 expression that is patched rather that continuous, as well as that of the terminal differentiation markers loricrin and filaggrin. Also, the normal sequence of expression of these markers is not fully preserved.

To explore whether prolonged culture conditions might have conferred tumorigenic potential to COCA cells we performed in vivo tumorigenicity assays using immunodeficient mice [25‐27]. In this assays, keratinocytes were injected in the flanks of nu/nu mice alone or mixed with dermal fibroblasts to help sustain keratinocyte growth. After 4-6 months there was no external sign of tumour growth and animals were sacrificed and studied to search for any evidence of tumour formation. Experiments using genetically modified murine keratinocytes were done in parallel and yielded 100% tumour growth by 2 months, both in the presence/absence of dermal fibroblasts (data not shown).

To rule out the possibility that subcutaneous injection is an adverse environment for non-genetically modified keratinocytes to thrive we used a different approach, silicone chambers implanted onto the back of immunodeficient mice which might provide a more physiological environment for the keratinocytes to grow. In these grafting assays, keratinocytes are seeded on the subcutaneous fascia of immunodeficient mice with dermal fibroblasts. After 1 month of grafting, COCA keratinocytes generated a fully differentiated epidermis-like structure (Figure 3A). This epidermis displays some hyperplasia due to increased proliferation as demonstrated by the expansion of K5 to suprabasal layers, K6 expression and high rate of BrdU incorporation (Figure 3B, C & 3D). However, expression of early and late differentiation markers is maintained (K10 in Figure 3C-F, loricrin in 3E & filaggrin in 3F). No signs of dysplasia or malignant transformation were evident. When the experiments were done grafting genetically modified keratinocytes, papillomas with evident signs of malignant transformation developed (data not shown).