Medullary Thymic Epithelium: A Mosaic of Epithelial “Self”?

T he thymus is organized into discrete subcapsular, corti-cal, and medullary environments that have been defined by the distribution of morphologically and phenotyp-ically distinct populations of epithelial cells and other stromal cell constituents (1–5). T cell progenitor cells entering the thymus move in a centripetal fashion from the subcapsular region through the cortex to the medullary area during their maturation process. Proximal stages of thymocyte development thought to occur in the subcapsu-lar and cortical compartments include expansion of the progenitor pool and expression of pre-TCR, which in turn selects thymocytes with a functional TCR-␤ chain to undergo further maturation and to become CD4 ϩ 8 ϩ (double positive, DP) thymocytes expressing low levels of TCR-␣ր␤. Further development of DP thymocytes is dependent on recognition of MHC–peptide ligands expressed by cortical thymic epithelial cells. DP thymocytes expressing receptors of appropriate specificity for these ligands exhibit some phenotypic characteristics of activation and downregulate expression of either CD4 or CD8 to yield mature CD4 ϩ or CD8 ϩ single positive thymocytes (6). Concurrent with this process of positive selection, self-reactive thymocytes are eliminated. The anatomic setting for this process within the thymus is variable and appears to be dependent on the qualitative and quantitative parameters of ligand expression. Extensive cortical deletion has been observed in some TCR transgenic mouse models, whereas in others deletion was more evident in corticomed-ullary and medullary compartments (7). Hematogenously derived dendritic cells localized to the inner cortex and corticomedullary interface represent one major cell type able to mediate clonal deletion in the thymus. Several transgenic model systems have also pointed to a role for medullary thymic epithelium in this process (8–10), although the extent of functional overlap of these different cell types in their ability to mediate negative selection remains unclear. A significant constraint on the process of thymic negative selection for a given self-protein is a requirement for expression of the cognate peptide–MHC complex by thy-mic APCs, either as a consequence of expression of this particular protein by the thymic APCs or due to adequate levels in the circulation. Although such a requirement would be compatible with negative selection of T cells specific for some self molecules, it poses a challenge for thymic elimination of T cells specific for self-proteins expressed in developmentally, temporally, or spatially regulated patterns, such as inducible tissue–specific proteins, hormones that exhibit oscillating levels in the circulation, or proteins whose pattern of expression …