Does T-cell tolerance require a dedicated antigen-presenting cell?

ALMOST 30 years ago Burnet proposed that the immune system maintained self-tolerance by deleting autoreactive lymphocytes1. Recently it has become clear that for T cells this step occurs in the thymus, where developing T cells first express their antigen-specific receptors2–4. Here a T-cell which encounters its antigen disappears—if it is not dead, it at least stops expressing its receptors. In the periphery by contrast, encounter with antigen leads to activation and proliferation of the responding T-cell. There are two possible explanations for this difference. Either the antigen-presenting cells in the thymus are different from those in the periphery and instead of producing positive signals they directly or indirectly kill the thymocytes5,6; or the T cells themselves are different, and like immature B cells, may die after encounter with antigen7,8. We tested the first possibility and found that dendritic cells from spleen, which are the most potent activators of mature T cells9, are also the most potent inactivators of young developing T cells. Thus it is not the antigen-presenting cell which determines whether a T-cell responds or dies, but the T-cell itself or its thymic environment.