Germ-line selection ensures embryonic autoreactivity and a positive discrimination of self mediated by supraclonal mechanisms.

It is necessary to clarify principles and mechanisms of natural tolerance to body tissues, in order to derive appropriate diagnostics, therapeutics and prognostics of autoimmune diseases (AID). I will argue that AIDs result from deficits in autoreactive regulatory T cell generation and/or function, and propose a model that explains why relatively few prototypes of AID exist, as well as their organ-specificity or systemic nature. The model suggests that natural tolerance is achieved through evolutionarily selected developmental genetic programs: (i) for patterns of V-region expression early in life that ensure auto(multi)reactivity at the outset of the system; (ii) for a cellular composition of thymic stroma that 'breeds' and activates regulatory (autoreactive) T cells in early development; (iii) for lymphocyte differentiation and population dynamics, that results in peripheral 'education' of regulatory tissue-specific cells, while allowing for 'unregulated' clonal responses to nonself. In the present model, S/NS discrimination is 'supraclonal' and 'dominant', related to other 'systemic' properties such as the regulation of total lymphocyte numbers, the 'open-endedness' of repertoires, and their differences in health and disease. Dominant tolerance models in general, also solve the paradox that pathogenic autoreactivity is rare, in spite of the extensive V-region degeneracy of lymphocyte recognition and the high frequency of cross-reactivity between S/NS; in short, it is astonishing that we are not autoimmune every time we get infected. As in other areas of biomedical science, time is perhaps ripe to move from component (clonal) analysis to system's biology, as some have proned for years. Copyright 2000 Academic Press.