Athymic mice grafted at birth with allogeneic thymic epithelium (TE) from day 10 embryos before hematopoietic cell colonization reconstitute normal numbers of T cells and exhibit full life-long tolerance to skin grafts of the TE haplotype. Intravenous transfers of splenic cells, from these animals to adult syngeneic athymic recipients, reconstitute T-cell compartments and the ability to reject third-party skin grafts. The transfer of specific tolerance to skin grafts of the TE donor strain, however, is not observed in all reconstituted recipients, and the fraction of nontolerant recipients increases with decreasing numbers of cells transferred. Furthermore

Grafts of thymic epithelium (TE) rudiments restore T cell development and function in allogeneic athymic mice. These TE chimeras are specifically tolerant to grafts of peripheral tissues (e.g. skin and heart) from the TE donor strain, although they harbor peripheral immunocompetent T cells capable of rejecting those grafts. Initial analysis has shown that TE chimeras also harbor TE-selected CD4 T lymphocytes that inhibit graft rejection by tissue-reactive T cells in immunocompetent recipients. Peripheral tolerance in TE chimeras is thus maintained by dominant mechanisms dependent on regulatory CD4 T lymphocytes. Here we show that TE-selected regulatory T cells recruit nontolerant tissue-reactive CD4 and CD8 T cells to express similar regulatory functions. Only recent thymic emigrants, but not peripheral resident mature T cells are susceptible to this process of functional education, which also requires exposure to specific antigens and occurs entirely in the periphery. We propose that these mechanisms play a major role in establishing and maintaining natural self tolerance to tissue-specific antigens.

The avian embryo has provided an appropriate model to study the ontogeny of the primary lymphoid organs, thymus and bursa of Fabricius. By using the quail-chick marker system the embryonic origin of the highly intricate cell components which form these organs could be traced back to the initial endodermal, mesodermal and ectodermal germ layers. The timing and dynamics of the incoming and outcoming flows of hemopoietic cells which characterize their lymphopoietic activity could be revealed in both quail and chick embryos. This knowledge served as a basis for an investigation on the role of the epithelial component of the thymus (derived from the pharyngeal endoderm) on tolerance to tissue graft and, by extension, tolerance to self. When this work was undertaken, the prevailing view was that exposure of the developing immune system to foreign antigens in the embryo allows them to be assimilated to self components in the mature animal. In fact, this was found to be true for allogeneic grafts between MHC-distinct chickens, of certain tissues, such as for instance wing tissues. However, in heterospecific transplantations, i.e. when a limb bud was grafted from quail to chick embryos, the chick host acutely rejected the foreign limb soon after birth. In contrast, grafts of the quail thymic epithelial (TE) rudiment resulted in the development of a chimeric thymus in which the foreign epithelial component was not only tolerated but able to induce full tolerance of the grafted wing from the same donor. By monitoring the amount of quail TE implanted we showed in addition that only part of the peripheral T-cell population had to differentiate in the context of the quail epithelial cells to induce tolerance to quail tissues. This pointed to the generation in the thymus of regulatory T cells, coexisting with specific anti-quail reactive T cells, but able to inhibit them from reacting against the quail wing antigenic determinants. A mammalian model was then devised to further study this mechanism of tolerance that we have qualified as "dominant" by opposition to the current model based on either clonal elimination or anergy which can be considered as recessive or passive. Nude mice of MHC type A were grafted with TE of E10 type B embryos. They became reconstituted for T-cell function but tolerant for B skin allografts. Spleen cells from such tolerant animals injected to naive A nude mice reconstituted T cell function in the recipient and transferred the tolerance to B skin grafts. Reducing the number of donor cells resulted in the segregation of the two phenomena. For low numbers the recipients were restored but not tolerant, thus showing the coexistence in the tolerant donor of anti-B reactive T cells together with regulatory cells able to abolish their reactivity against B determinants. Other experiments demonstrated that TE-induced tolerance does not rely on clonal deletion or anergy. This was shown on systems where elimination of cells directed toward superantigens was screened. It turned out that tolerance to skin grafts a...

Current models of tolerance to peripheral, tissue-specific antigens contain some major caveats. First, they consider peripheral tolerance independently from intrathymic T cell selection, a dichotomy that is challenged by observations on TE-induced tolerance. Second, they do not account for the fact that vertebrates are more readily tolerised in development than in adult life. Third, they do not explain the fact that embryonic/neonatal tolerance to foreign tissues can only be induced by HC or TE. A model of thymic selection and peripheral tolerance is developed here that resolves those problems, by assuming two classes of T cell effector functions, one being regulatory and the other aggressive. Three postulates are required: (1) both epithelial and hemopoietic cellular compartments of the thymic stroma can support both positive and negative selection of T cells, but with vastly different avidity requirements and efficiency; (2) positively selected T cells with the highest avidity that escape deletion are activated intrathymically and irreversibly committed for regulatory effector functions; (3) the functional phenotype of all other thymic emigrants is determined in the periphery upon encounter with antigen. Functional commitment in the periphery depends on the maturity stage (RTE or PMR) of the immunocompetent cell, on the nature of the antigen-presenting cells, and on the effector classes of other T lymphocytes interacting on the same presenting cell. This model explains a number of observations on experimental autoimmune disease and transplantation tolerance, and it contains several readily testable predictions.