We examined the role of cytokines in the cutaneous response to the application of trinitrochlorobenzene (TNCB) in both nonsensitized and sensitized mice, i.e., in the irritant reaction (IR) and contact hypersensitivity reactions (CH). When administered immediately before challenge, anti-tumor necrosis factor (TNF) antibody abrogated the ear swelling response in CH; antibody directed against interferon gamma or antibodies to both granulocyte/macrophage colony-stimulating factor and interleukin 3 (IL-3) had a partial inhibitory effect; anti-IL-2 receptor antibody had no effect. Anti-TNF prevented the various features of the CH, as seen on histological sections, e.g., leukocyte infiltration and hemorrhages within the dermis and keratinocytes necrosis. Anti-TNF antibody also prevented the IR. The presence of TNF mRNA was evaluated on Northern blots; TNF-alpha mRNA was detectable in an untreated ear, increased after the application of TNCB in nonsensitized mice, and was highest in sensitized mice. TNF mRNA accumulation, which was evident 0.5 h after hapten application and lasted greater than 72 h, was abolished by treatment with anti-TNF antibody, thus suggesting an auto-amplification of TNF production. The cellular origin of TNF mRNA was explored by in situ hybridization; basal keratinocytes showed the highest labeling, but TNF mRNA was also detectable in cells of the dermal infiltrate. After hapten (TNCB) application at sites susceptible (the ear) or resistant (the foot pad) to CH or IR, a close correlation was observed between TNF mRNA accumulation and the intensity of the inflammatory reaction. The major role played by TNF in both the CH and the IR explains the histologically similar aspects of these reactions and the extreme variability of these reactions at various anatomical sites.