The earlier known TNF family cytokines have fairly wide physiological actions, mainly in inflammation and immune responses. It is now considered feasible to develop these large proteins themselves as therapeutic agents, but in addition, modular organisation of structures of biological proteins, and the identification of localised ligand-receptor interaction sites, allow rational design of smaller, preferably non-peptide molecules which interfere with these protein protein interactions. Neutralising anti-TNF antibodies and soluble TNF receptor preparations were shown to have striking anti-inflammatory activities in clinical studies, particularly in rheumatoid arthritis. As the TNF beta TNFR1 co-crystal structure was the first in the family to be solved, rational drug design based on the ligand receptor interaction sites is more advanced. Ligand mutations and a peptide sequence from TNF-alpha have given useful information regarding ligand-receptor interactions. Small peptide sequences from TNFR I which interact with the ligand have shown some activity in blocking the biological actions of TNF. The physiological activities of several recently-discovered ligands are more limited, giving possibilities for selective treatment of several diseases. For example, TRAIL can induce apoptosis in a wide range of tumour cells with little effects on normal tissues, both in vitro and in vivo. The co-crystal structure of TRAIL with one of its signalling receptors TRAILR 2 has been solved, opening the way to rational small molecule drug design. TRANCE (RANKligand) has modulatory effects on the dendritic cell T cell interaction in immune responses. However, it plays a more major controlling role in the development of osteoclasts and their bone resorbing activity. In this way, it is a very interesting drug development target for the treatment of bone disorders such as osteoporosis. A recombinant secreted inhibitory receptor, osteoprotegerin (OPG), is in Phase 1 clinical trial for the treatment of hyper-resorptive bone diseases. However, OPG also blocks TRAIL and may not be sufficiently specific in long term therapy, but it is hoped that inhibitors of the interaction of TRANCE and its specific signalling receptor, RANK, can be rationally designed.