The recent discovery of multilineage donor leukocyte microchimerism in allograft recipients up to three decades after organ transplantation implies the migration and survival of donor stem cells within the host. It has been postulated that in chimeric graft recipients, reciprocal modulation of immune responsiveness between donor and recipient leukocytes may lead, eventually, to the induction of mutual immunologic nonreactivity (tolerance). A prominent donor leukocyte, both in human organ transplant recipients and in animals, has invariably been the bone marrow-derived dendritic cell (DC). These cells have been classically perceived as the most potent antigen-presenting cells but evidence also exists for their tolerogenicity. The liver, despite its comparatively heavy leukocyte content, is the whole organ that is most capable of inducing tolerance. We have observed that DC progenitors propagated from normal mouse liver in response to GM-CSF express only low levels of major histocompatibility complex (MHC) class II antigen and little or no cell surface B7 family T cell costimulatory molecules. They fail to activate resting naive allogeneic T cells. When injected into normal allogeneic recipients, these DC progenitors migrate to T-dependent areas of host lymphoid tissue, where some at least upregulate cell surface MHC class II. These donor-derived cells persist indefinitely, recapitulating the behavior pattern of donor leukocytes after the successful transplantation of all whole organs, but most dramatically after the orthotopic (replacement) engraftment of the liver. A key finding is that in mice, progeny of these donor-derived DC progenitors can be propagated ex vivo from the bone marrow and other lymphoid tissues of nonimmunosuppressed spontaneously tolerant liver allograft recipients. In humans, donor DC can also be grown from the blood of organ allograft recipients whose organ-source chimerism is augmented with donor bone marrow infusion. DC progenitors cannot, however, be propagated from the lymphoid tissue of nonimmunosuppressed cardiac-allografted mice that reject their grafts. These findings are congruent with the possibility that bidirectional leukocyte migration and donor cell chimerism play key roles in acquired transplantation tolerance. Although the cell interactions are undoubtedly complex, a discrete role can be identified for DC under well-defined experimental conditions. Bone marrow-derived DC progenitors (MHC class II+, B7-1dim, B7-2-) induce alloantigen-specific hyporesponsiveness (anergy) in naive T cells in vitro. Moreover, costimulatory molecule-deficient DC progenitors administered systemically prolong the survival of mouse heart or pancreatic islet allografts. How the regulation of donor DC phenotype and function relates to the balance between the immunogenicity and tolerogenicity of organ allografts remains to be determined.