Prediction of Chemical Respiratory and Contact Sensitizers by OX40L Expression in Dendritic Cells Using a Novel 3D Coculture System

The use of animal models in chemical safety testing will be significantly limited due to the recent introduction of the 3Rs principle of animal experimentation in research. Although several in vitro assays to predict the sensitizing potential of chemicals have been developed, these methods cannot distinguish chemical respiratory sensitizers and skin sensitizers. In the present study, we describe a novel in vitro assay that can discriminate respiratory sensitizers from chemical skin sensitizers by taking advantage of the fundamental difference between their modes of action, namely the development of the T helper 2 immune response, which is critically important for respiratory sensitization. First, we established a novel three-dimensional (3D) coculture system of human upper airway epithelium using a commercially available scaffold. It consists of human airway epithelial cell line BEAS-2B, immature dendritic cells (DCs) derived from human peripheral blood CD14+ monocytes, and human lung fibroblast cell line MRC-5. Respective cells were first cultured in individual scaffolds and subsequently assembled into a 3D multi-cell tissue model to more closely mimic the in vivo situation. Then, three typical chemicals that are known respiratory sensitizers (ortho-phthaldialdehyde, hexamethylene diisocyanate, and trimellitic anhydride) and skin sensitizers (oxazolone, formaldehyde, and dinitrochlorobenzene) were added individually to the 3D coculture system. Immunohistochemical analysis revealed that DCs do not migrate into other scaffolds under the experimental conditions. Therefore, the 3D structure was disassembled and real-time reverse transcriptase-PCR analysis was performed in individual scaffolds to analyze the expression levels of molecules critical for Th2 differentiation such as OX40 ligand (OX40L), interleukin (IL)-4, IL-10, IL-33, and thymic stromal lymphopoietin. Both sensitizers showed similarly augmented expression of DC maturation markers (e.g., CD86), but among these molecules, OX40L expression in DCs was most consistently and significantly enhanced by respiratory sensitizers as compared to that by skin sensitizers. Thus, we have established a 3D coculture system mimicking the airway upper epithelium that may be successfully applied to discriminate chemical respiratory sensitizers from skin sensitizers by measuring the critical molecule for Th2 differentiation, OX40L, in DCs.