The immortal human keratinocyte line HaCaT has been employed in many studies as paradigm for epidermal keratinocytes. In order to demonstrate its potential to form stable epidermal structures in response to connective tissue, this was challenged in surface transplants on nude mice, where normal keratinocytes rebuild a typical epidermis within two weeks. During the initial regeneration phase (day 1-4) multilayered but poorly organized epithelia formed with proliferating cells in all layers in analogy to normal keratinocytes. Similarly, with tissue consolidation (around day 7) proliferation was reduced and restricted to cells in basal position marked by keratin K14 and beta1-integrin immunostaining. The strong suprabasal reaction for K1 and K10, the appearance of the late markers K2e, filaggrin and loricrin as well as the polarized distribution of alpha2beta1 and alpha3beta1 indicated advancing tissue normalization (day 14). Keratinization further improved at around three weeks switching from the initial parakeratotic to the regular orthokeratotic type which was prominent at six weeks. Accordingly, most ultrastructural features typical for epidermis or normal keratinocyte grafts were detectable including a complete basement membrane (BM) with regular attachment structures. Matrix- and BM-components appeared sequentially with marked linear deposition of laminin-5 (day 4) followed by accumulation of collagen-IV and 'classical' BM-laminin between one and two weeks. With the general codistribution of integrin alpha6beta4 and BM-molecules (day 14) collagen-VII lining of BM became prominent, while epithelium and host connective tissue were still separated by the collagen matrix. In accordance with the delayed orthokeratinization, wound-matrix molecules (fibronectin, tenascin) persisted longer than in normal keratinocyte transplants. Finally, grafts of long-term passaged (no. 310) cells demonstrated a remarkable stability in the expression of epidermal markers. Thus, the immortalized HaCaT cells reveal a generally high competence to realize an epidermal phenotype in a natural environment and appear therefore qualified for in vitro studies on structural and regulatory aspects of keratinocyte physiology and pathology