No abstract

Toxic epidermal necrolysis is an extremely severe drug reaction, manifesting itself by widespread apoptosis of keratinocytes, generally considered to result from Fas/CD95-FasLigand interaction, but of unknown primary mechanism. We looked at the role of cells present in the skin blisters as probable effectors of this immune reaction. In a patient suffering from cotrimoxazole-induced toxic epidermal necrolysis, blister fluid cells were phenotyped by FACS and tested without prior restimulation for cytotoxicity on autologous and allogeneic cells in the presence of the drug. Blister fluid lymphocytes were predominantly CD8+, DR+, CLA+, CD56+ T lymphocytes, perforin positive and expressing preferentially two Vbeta chains of the T cell receptor repertoire. These lymphocytes were cytotoxic only in the presence of the drug towards autologous EBV transformed lymphocytes and towards allogeneic cells sharing HLA-Cw4. Cytotoxicity occurred in the presence of either cotrimoxazole, sulfamethoxazole, or the nitroso metabolite of sulfamethoxazole, but not with the hydroxylamine metabolite of sulfamethoxazole. The lysis was blocked by an anti-MHC class I monoclonal antibody. It was abolished by EGTA and CMA, but neither by anti-fas, brefeldin A, nor by anti-TRAIL receptor monoclonal antibodies, strongly suggesting perforin/granzyme-mediated cytotoxicity, without implication of Fas or TRAIL at this stage. This is direct evidence that T lymphocytes present within the lesions of toxic epidermal necrolysis may exhibit, without any re-stimulation, a drug-specific cytotoxicity against autologous cells. Harboring the markers of classical CTL and MHC class I restriction these lymphocytes reacted against the parent drug and one of its reactive metabolites. These results challenge several current concepts and could support new therapeutic approaches.

Toxic epidermal necrolysis is a rare disease observed as a consequence of adverse reactions to drugs. It results in the widespread apoptosis of epidermal cells and has a high mortality rate. The mechanisms leading to this apoptosis are not yet elucidated. We investigated whether the cytokines present in the blister fluid, which accumulates under necrotic epidermis, originated from T lymphocytes and may play a role in the propagation of keratinocyte apoptosis. Interferon gamma (IFN-gamma), soluble tumor necrosis factor alpha (TNF-alpha), soluble Fas ligand (sFas-L) were present in much higher concentration in the blister fluids of 13 toxic epidermal necrolysis (TEN) patients than in control fluids from burns. The results of RT-PCR studies, however, indicated that only IFN-gamma and to a lesser extent interleukin (IL)-18 were produced by mononuclear cells present in the fluid. That suggests that the other cytokines also present (TNF-alpha, sFas-L, IL-10) rather originated from activated keratinocytes. Fas-L was indeed overexpressed on the membranes of keratinocytes in lesional skin in situ. The Th1 profile of T lymphocyte activation found in the blister fluid of patients with TEN is consistent with a key role for drug-specific cytotoxic T lymphocytes (CTL) as previously reported, the activation of keratinocytes by IFN-gamma making them sensitive to cell-mediated cytolysis. We propose the hypothesis that the production of Fas-L, TNF-alpha, and IL-10 by keratinocytes could be a defense mechanism against CTL rather than a way of propagating apoptosis among epidermal cells.