A walk round the edges of self tolerance.

The immunologists of my generation were brought up on Sir Macfarlane Burnet's clonal selection theory.' Anticipating Margaret Thatcher's view of society, we believed that all that matters in the immune system is the sum of the decisions taken by individual lympho-cytes. Engagement of its receptor by antigen suffices to drive a lymphocyte to respond. For T cells the only exception is the thymus, where engagement of the receptor by self antigen drives the cell into programmed cell death. This process of negative selection explains how the immune system avoids reacting against self. The two parts of the theory fitted neatly together, and nothing more seemed necessary. When cooperation between T and B cells came along,2 it too could be accommodated comfortably within the theory. Later came Nils Jeme's idiotype network, with the proposal that lymphocytes use their receptors to recognise not only antigen but also the receptors of other lymphocytes.3 Opinion among immunologists divided, with heretics arguing that antigens work not by selection but by perturbing a network of interactive cells. The orthodox viewed the network merely as an inescapable but unhelpful consequence of clonal selection, which the immune system does its best to damp down by suppression. Challenge to clonal selection Over the past year or two Irun Cohen, from the Weizmann Institute in Israel, has been mounting a more radical attack on the clonal selection theory.4 He draws partly on older information, such as the well established fact that many proteins prove unable to induce an immune response without the help of an adjuvant. This alone is sufficient to dispose of the idea that engaging a receptor with antigen is all that is required. Attempts to explain adjuvants simply as devices for maintaining the right local concentration of antigen have failed. Another ofhis arguments is the restricted range of autoimmune diseases. Burnet thought that these diseases resulted from mutations in receptor-encoding genes, so that each case should be unique. In reality the number of such diseases is quite small, perhaps less than 30 in all, with many subjects suffering from each one. Perhaps the most telling argument comes from recent studies on autoreactive T cells in tissue culture. It has long been known that normal subjects contain occasional T cells able to react against self B cells and self macro-phages, but it has not proved possible to identify just what antigens, if any, are involved. The position changed …