Mutations in the filaggrin (FLG) gene are strongly associated with common dermatological disorders such as atopic dermatitis. However, the exact underlying pathomechanism is still ambiguous. Here, we investigated the impact of FLG on skin lipid composition, organization, and skin acidification using a FLG knockdown (FLG-) skin construct. Initially, sodium/hydrogen antiporter (NHE-1) activity was sufficient to maintain the acidic pH (5.5) of the reconstructed skin. At day 7, the FLG degradation products urocanic (UCA) and pyrrolidone-5-carboxylic acid (PCA) were significantly decreased in FLG- constructs, but the skin surface pH was still physiological owing to an upregulation of NHE-1. At day 14, secretory phospholipase A2 (sPLA2) IIA, which converts phospholipids to fatty acids, was significantly more activated in FLG- than in FLG+. Although NHE-1 and sPLA2 were able to compensate the FLG deficiency, maintain the skin surface pH, and ensured ceramide processing (no differences detected), an accumulation of free fatty acids (2-fold increase) led to less ordered intercellular lipid lamellae and higher permeability of the FLG- constructs. The interplay of the UCA/PCA and the sPLA2/NHE-1 acidification pathways of the skin and the impact of FLG insufficiency on skin lipid composition and organization in reconstructed skin are described.

The development and validation of in vitro systems for skin absorption testing is inevitable. More research efforts are required for the development of reconstructed disease models. Reconstructed skin models need to be improved, especially in terms of complexity to mimic the in vivo situation better. It should not, however, be the main goal to imitate the in vivo situation exactly, but to establish reliable systems that ensure predictive and reliable data.

Reconstructed skin models are suitable test systems for toxicity testing and for basic investigations on (patho-)physiological aspects of human skin. Reconstructed human skin, however, has clear limitations such as the lack of immune cells and a significantly weaker skin barrier function compared to native human skin. Potential reasons for the latter might be the lack of mechanical forces during skin model cultivation which is performed classically in static well-plate setups. Mechanical forces and shear stress have a major impact on tissue formation and, hence, tissue engineering. In the present work, a perfusion platform was developed allowing dynamic cultivation of in vitro skin models. The platform was designed to cultivate reconstructed skin at the air-liquid interface with a laminar and continuous medium flow below the dermis equivalent. Histological investigations confirmed the formation of a significantly thicker stratum corneum compared to the control cultivated under static conditions. Moreover, the skin differentiation markers involucrin and filaggrin as well as the tight junction proteins claudin 1 and occludin showed increased expression in the dynamically cultured skin models. Unexpectedly, despite improved differentiation, the skin barrier function of the dynamically cultivated skin models was not enhanced compared with the skin models cultivated under static conditions.