During the last decade, a variety of skin equivalent models have been developed to provide useful models for investigations of skin physiology. Epidermis reconstructs based on fibroblast-populated collagen matrices are widely used especially for toxicological and pharmacological investigations. At present, there is a special focus on the regulation of skin pigmentation by melanogenic stimulators or inhibitors. Since skin pigmentation is ultimately controlled by melanin-producing melanocytes interspersed in the epidermis the most simple model is melanocyte monoculture (Virador et al. 1999; Yoshimura et al. 2001). However, melanin production in melanocytes is influenced by keratinocytes and fibroblasts (Hedley et al. 2002; Lei et al. 2002). Therefore, coculture models composed of keratinocytes and melanocytes (Lei et al. 2002) as well as different stratified three-dimensional skin equivalent models have been developed (Archambault et al. 1995; Bessou et al. 1995, 1997; Regnier et al. 1999; Todd et al. 1993). These skin models show altered pigmentation as a result of melanin transfer from melanocytes to keratinocytes after treatment with regulators of pigmentation (Bessou et al. 1997).

As normal human keratinocytes show limitations concerning their proliferative capacity and donor variability, the use of a keratinocyte cell line would be more suitable. The spontaneously immortalized human keratinocyte cell line HaCaT would overcome these obstacles (Boukamp et al. 1988). This cell line exhibits differentiation markers similar to normal human keratinocytes when used in submerged culture as well as in surface transplants on nude mice (Breitkreutz et al. 1998; Ryle et al. 1989). In organotypic cultures using collagen gels HaCaT cells show a multilayered epithelium with a high degree of differentiation, but without reaching the final stages of terminal differentiation (Boelsma et al. 1999; Kehe et al. 1999; Schoop et al. 1999).

Although HaCaT cells have often been used in reconstructed skin models, there is only one report on the interaction between HaCaT cells and human melanocytes in such organotypic cultures (Lee et al. 2001). Lee et al. have shown in a skin equivalent model that HaCaT cells are able to receive melanosomes from melanocytes. However, in contrast to the in vivo situation the melanocytes are not only located in the basal layer but also in the suprabasal layers.

The aim of our study was to develop a coculture model using HaCaT cells and normal human melanocytes to set up a reliable and standardized in vitro screening test system. The efficacy of regulators of pigmentation in this model was evaluated by melanin content measurement. Furthermore, a three-dimensional HaCaT skin equivalent containing melanocytes was developed which mimics different skin phototypes and enables the immunohistological investigation of the effects of pigmentation regulators.

HaCaT cells were donated by Prof. Dr. N.E. Fusenig (German Cancer Research Center, Heidelberg, Germany). Normal human dermal fibroblasts and epidermal melanocytes were derived from human foreskin obtained during circumcision and represented skin phototypes II–V according to the Fitzpatrick classification (Fitzpatrick 1988). Dermal fibroblasts and HaCaT cells were grown in DMEM containing 10% fetal bovine serum (FBS). Melanocytes were grown in melanocyte growth medium (MGM, PromoCell). Dermal fibroblasts were used at passage 3, melanocytes were used at passage 4, and HaCaT cells were used at passages 40–43. For coculture experiments HaCaT cells were seeded at 2.25×105 cells/well in six-well culture plates and cultured until confluent. Subsequently, melanocytes (skin phototype IV) were seeded on top at a density of 4×104 cells. The next day fresh medium containing 300 μM IBMX (3-isobutyl-1-methylxanthine) or 0.2 mg/ml kojic acid (2-hydroxymethyl-5-hydroxy-γ-pyrone) was added to the cocultures. HaCaT cells without melanocytes were cultured separately to serve as control. Cocultures were photographed 6 days later and were then lysed as described elsewhere with minor modifications (Lei et al. 2002). Briefly, cocultures were washed with phosphate-buffered saline for 1 min and then lysed with 200 μl 1 N NaOH. The crude cell extract was transferred into 96-well plates. The relative melanin content was determined photometrically at 405 nm using a Labsystems iEMS reader MF.

For the three-dimensional skin equivalents the fibroblast-populated collagen matrix was prepared as follows. A collagen solution mixture was prepared by quickly mixing two volumes of rat-tail collagen type 1 solution with one volume of neutralization solution (pH 7.8) containing 50 vol% 5× DMEM, 25 vol% 1 M HEPES, 25 vol% FBS and 25 μg/100 ml chondroitinsulphate. Fibroblasts were resuspended in the neutralization solution and added to the collagen type 1 solution at a final concentration of 1.7×105 cells/ml. On day 1, aliquots (0.6 ml) of the fibroblast/collagen solution were plated into 10-mm polycarbonate cell culture inserts (8.0 μm; NUNC, Wiesbaden, Germany) in 24-well culture dishes and after polymerization 50 μl fibronectin solution (75 μg/ml) was pipetted on top. The collagen gels were cultured submerged in DMEM overnight. On day 2, melanocytes (2×104 cells per insert) were seeded on top of the collagen gels and cultured submerged overnight. On day 3, HaCaT cells (1×105 cells per insert) were seeded on top of melanocyte-populated collagen gels. The skin equivalents were cultured submerged for an additional 4 days and subsequently at the air-liquid interface for 14 days. On day 4 of the air-liquid interface, melanogenic stimulators or inhibitors were applied for 10 days. IBMX, a melanogenic stimulator, was used at 300 μM diluted in the culture medium. Kojic acid, a melanogenic inhibitor, was used at 250 μM applied topically each day (10 μl). The air-lift medium consisted of a 3:1 mixture of DMEM and Ham's nutrient mixture F12 medium supplemented with 10% FBS, 50 μg/ml ascorbate, 0.4 μg/ml hydrocortisone, 5 μg/ml insulin and 1% GlutaMaxI. The cholera toxin commonly applied to produce HaCaT skin equivalents was omitted because of its known stimulatory effect on melanocytes (Abdel-Malek et al. 1992).

Skin equivalents were used in triplicate for each compound tested. At the end of the culture period the samples were fixed in 5% formalin or snap-frozen in TissueTec and stored at −80°C until immunohistological examination. To visualize melanocytes, cryosections were stained immunohistologically using the StreptABComplex/AP kit (Dako, Hamburg, Germany) and Sigma FAST Red (Sigma, Taufkirchen, Germany). The antibody used was HMB45 (mouse monoclonal against melanoma marker HMB45; Novocastra Laboratories, Newcastle upon Tyne, UK).

Melanocytes proliferated well in DMEM and retained their characteristic dendritic appearance when cocultured with HaCaT cells (Fig. 1a, b). The effect of the applied compounds in regulating pigmentation was evaluated by light microscopy before lysis. After 6 days treatment with the melanogenic stimulator IBMX, melanocytes became darker and more dendritic compared with the untreated control (Fig. 1c, d). In contrast, melanocytes became lighter and slightly less dendritic compared with the untreated control in the presence of the melanogenic inhibitor kojic acid (Fig. 1e, f). Photometric measurement of the lysates revealed a 62% increase in melanin production when stimulated by IBMX compared to the control. Kojic acid led to a 26% decrease in melanin content compared to the control (Fig. 2).

Fig. 1a–f.
figure 1

Phase-contrast (a, c, e) and bright-field (b, d, f) images of cocultures of HaCaT cells and normal human melanocytes exposed for 6 days to 300 μM IBMX (c, d) or 0.2 mg/ml kojic acid (e, f) compared to an untreated control coculture (a, b) (×100)

Full size image
Fig. 2.
figure 2

Melanin content in cocultures of HaCaT cells and melanocytes. Treated cells were lysed with NaOH and the absorbance was measured at 405 nm. The data are presented as the means±SD of three individual determinations each

Full size image

HaCaT cells and normal melanocytes were cocultured in skin equivalents in order to assess the ability of HaCaT cells to mimic different skin phototypes. We used melanocytes from two donors with different skin phototypes (type II/III and type IV/V, respectively). After air exposure, skin equivalents showed different coloration according to the donor skin phototype (Fig. 3).

Fig. 3.
figure 3

Skin colours of skin equivalents composed of HaCaT cells without melanocytes (left), with melanocytes of skin phototype II/III (middle) and with melanocytes of skin phototype IV/V (right) after 14 days at the air-liquid interface

Full size image

The immunohistochemical marker HMB45 showed a clear expression in these skin equivalents as well as in normal human skin (Fig. 4). In contrast to recently reported findings by Lee et al., melanocytes were located mainly in the lower third of the reconstructed epidermis and were not scattered throughout all suprabasal layers (Fig. 4b). HMB45 staining was stronger in the presence of IBMX compared to the untreated control (Fig. 4c). Application of the melanogenic inhibitor kojic acid reduced HMB45 staining (Fig. 4d).

Fig. 4a–d.
figure 4

HMB45-stained cryosections (6 μm) of normal human skin (a) and cryosections (6 μm) of skin equivalents composed of HaCaT cells and melanocytes (b, c, d) treated with IBMX (c) or kojic acid (d) compared to an untreated control skin equivalent (b). There is a slight nonspecific background staining in the fibroblast-populated collagen matrix (b, c, d) (×100)

Full size image

The results reported here show for the first time the ability of HaCaT skin equivalents to mimic different skin phototypes according to the donor melanocytes. The skin equivalents showed a multilayered epidermis consisting of HaCaT cells with melanocytes being mainly located in the basal layer. This basal location of the melanocytes was achieved by a new technical approach during skin equivalent cultivation. It is well known that melanocytes attach strongly to the basal membrane protein fibronectin (Gilchrest et al. 1985). Therefore, fibronectin was applied on top of the collagen matrix before seeding the melanocytes. Furthermore, HaCaT cells were seeded on top not before the next day. This method led to the location of melanocytes in the lower third of the epidermis similar to the pattern in normal skin.

In this study we also showed that regulators of pigmentation exert effects on skin equivalents as shown by HMB45 staining. This might have resulted from melanocyte proliferation (with IBMX) or melanocyte inhibition or destruction (with kojic acid) in the epidermis.

Cocultures based on HaCaT cells and normal human melanocytes represent a good and reliable screening system for the potential efficacy of regulators of pigmentation. The use of the immortalized cell line HaCaT offers a better standardization since interindividual effects of different donors do not influence the experiments. Their human origin (Boukamp et al. 1988) and the use of DMEM medium without the need for special additives are further advantages of HaCaT cells. To our knowledge human immortal melanocytes of all different skin phototypes have not been established. Thus the use of normal human melanocytes is still necessary to obtain a real human coculture system. Once available, their use would enable a highly standardizable coculture system for screening purposes and the limited proliferative capacity of normal human melanocytes could be overcome.

In conclusion, the regulators of pigmentation used in these experiments exerted effects on the melanin content of a coculture model composed of HaCaT cells and normal human melanocytes suggesting that this coculture model is useful for screening experiments. The new method described for the construction of a HaCaT skin equivalent resulted in the location of melanocytes in the lower third of the reconstructed epidermis similar to the pattern in normal skin. Furthermore, regulators of pigmentation also showed an effect on the melanocytes when used in this multilayered three-dimensional skin equivalent.