The CD95 (APO-1/Fas) receptor mediates programmed cell death in apoptosis sensitive cells upon oligomerization either by CD95 antibody or by its natural ligand, CD95 ligand (CD95 L). Studies on tissue distribution have shown that CD95 is broadly expressed in normal adult tissues. Furthermore, the spectrum of CD95-expressing cells in inducibly enlarged in the context of chronic inflammation. In contrast, the number of cells capable of expressing CD95 L is strikingly limited to small subsets of lymphocytes and histiocytes and to some specialized normal epithelia. This suggests that cell death via this receptor/ligand system, although possible in almost every tissue and cell type, is limited to very special scenarios one of which is chronic lymphohistiocytic inflammation. The lpr/lpr mouse and the gld/gld mouse are well-known models for autoimmune disorders. Both mutants have abnormal B and T lymphocytes and high titers of autoantibodies. Recently, these mice have been discovered to have functional defects in the murine equivalents of human CD95 and CD95 L, respectively. This suggests that the CD95/CD95 L system might act by preventing autoimmune disease, e.g., by preventing emergence of autoreactive T and B lymphocytes. A human homologue of the lpr mutation has been described as autoimmune lymphoproliferative syndrome. We show that, in mice, CD95/ CD95 L might be operative in experimental graft versus host disease. Further, we suggest a role of this system in early steps of ulcerative colitis. Considering our observations against the background of current literature, CD95/CD95 L is likely to play a dual role in induction and maintenance of peripheral tolerance on the one hand and in tissue damage by chronic inflammation on the other.