Regulatory mechanisms involving CD8+ T cells (CD8 regulatory T cells (Tregs)) are important in the maintenance of immune homeostasis. However, the inability to generate functional CD8 Treg clones with defined Ag specificity has precluded a direct demonstration of CD8 Treg-mediated regulation. In the present study, we describe the isolation of functional lines and clones representing a novel population of TCRalphabeta+ Tregs that control activated Vbeta8.2+ CD4 T cells mediating experimental autoimmune encephalomyelitis. They express exclusively the CD8alphaalpha homodimer and recognize a peptide from a conserved region of the TCR Vbeta8.2 chain in the context of the Qa-1a (CD8alphaalpha Tregs). They secrete type 1 cytokines but not IL-2. CD8alphaalpha Tregs kill activated Vbeta8.2+ but not Vbeta8.2- or naive T cells. The CD8alphaalpha Tregs prevent autoimmunity upon adoptive transfer or following in vivo activation. These findings reveal an important negative feedback regulatory mechanism targeting activated T cells and have implications in the development of therapeutic strategies for autoimmune diseases and transplantation.

H2-O is a nonpolymorphic class II molecule whose biological role remains to be determined. H2-O modulates H2-M function, and it has been generally believed to be expressed only in B lymphocytes and thymic medullary epithelial cells, but not in dendritic cells (DCs). In this study, we report identification of H2-O expression in primary murine DCs. Similar to B cells, H2-O is associated with H2-M in DCs, and its expression is differentially regulated in DC subsets as well as during cell maturation and activation. Primary bone marrow DCs and plasmacytoid DCs in the spleen and lymph nodes express MHC class II and H2-M, but not the inhibitor H2-O. In contrast, myeloid DCs in secondary lymphoid organs express both H2-M and H2-O. In CD8αα+ DCs, the ratio of H2-O to H2-M is higher than in CD8αα− DCs. In DCs generated from GM-CSF- and IL-4-conditioned bone marrow cultures, H2-O expression is not detected regardless of the maturation status of the cells. Administration of LPS induces in vivo activation of myeloid DCs, and this activation is associated with down-regulation of H2-O expression. Primary splenic DCs from H2-O−/− and H2-O+/+ mice present exogenous protein Ags to T cell hybridomas similarly well, but H2-O−/− DCs induce stronger allogeneic CD4 T cell response than the H2-O+/+ DCs in mixed leukocyte reactions. Our results suggest that H2-O has a broader role than previously appreciated in regulating Ag presentation.

Qa-1, a nonclassical class I histocompatibility molecule expressed in mice, predominantly assembles with a single nonameric peptide, Qdm, derived from the signal sequence of certain class Ia molecules. The Qa-1/Qdm complex is the primary ligand for CD94/NKG2A inhibitory receptors expressed on a major fraction of natural killer (NK) cells. Cells become susceptible to killing by NK cells under conditions where surface expression of the Qa-1/Qdm inhibitory ligand is reduced. The CD94/NKG2 "missing-self" recognition system serves as mechanism for removing cells that have abnormalities in the intracellular machinery required for assembly and expression of class I-peptides complexes, as a consequence of viral infection, for example. Despite its highly focused peptide-binding specificity, Qa-1 also has a capacity to act as an antigen-presentation molecule for CD8+ T cells. It appears that a small subpopulation of these T cells undergoes positive selection by interaction with Qa-1 in the thymus, and they maintain their specificity for Qa-1 after maturation. The role of these unusual T cells in adaptive immune responses remains to be defined.

Naive T cells have the capacity to expand in a lymphopenic environment in a process called homeostatic expansion, where they gain a memory-like phenotype. Homeostatic expansion is dependent on competition for a number of factors, including growth factors and interactions with their selecting self-MHC molecules. In contrast to conventional T cells, it is unclear whether class Ib-restricted CD8+ T cells have a capacity to undergo homeostatic expansion. In this study, we demonstrate that polyclonal MHC Ib-restricted CD8+ T cells can undergo homeostatic expansion and that their peripheral expansion is suppressed by conventional MHC-restricted T cells. The acute depletion of CD4+ T cells in MHC class Ia-deficient Kb−/−Db−/− mice led to the substantial expansion of class Ib-restricted CD8+ T cells. Adoptive transfer of class Ib-restricted CD8+ T cells to congenic lymphopenic recipients revealed their ability to undergo homeostatic expansion in a MHC Ib-dependent manner. To further study the homeostatic expansion of MHC Ib-restricted T cells in the absence of all conventional MHC-restricted T cells, we generated mice that express only MHC Ib molecules by crossing H-2Kb−/−Db−/− with CIITA−/− mice. CD8+ T cells in these mice exhibit all of the hallmarks of naive T cells actively undergoing homeostatic expansion with constitutive memory-like surface and functional phenotype. These findings provide direct evidence that MHC Ib-restricted CD8+ T cells have the capacity to undergo homeostatic expansion. Their peripheral expansion is suppressed under normal conditions by a numerical excess of conventional MHC class Ia- and class II-restricted T cells.

Naive T cells proliferate in response to self MHC molecules after transfer into lymphopenic hosts, a process that has been termed homeostatic proliferation (HP). Previous studies have demonstrated that HP is driven by low level signaling induced by interactions with the same MHC molecules responsible for positive selection in the thymus. Little is known about the homeostatic regulation of T cells specific for class Ib molecules, including Qa-1 and H2-M3, though it has been suggested that their capacity to undergo homeostatic expansion may be inherently limited. In this study, we demonstrate that naive 6C5 TCR transgenic T cells with specificity for Qa-1b have a capacity similar to conventional T cells to undergo HP after transfer into sublethally irradiated mice. Proliferation was largely dependent on the expression of β2-microglobulin, and experiments with congenic recipients expressing Qa-1a instead of Qa-1b demonstrated that HP is specifically driven by Qa-1b and not through cross-recognition of classical class I molecules. Thus, the same MHC molecule that mediates positive selection of 6C5 T cells is also required for HP. Homeostatic expansion, like positive selection, occurs in the absence of a Qa-1 determinant modifier, the dominant self-peptide bound to Qa-1 molecules. However, experiments with TAP−/− recipients demonstrate a clear distinction between the ligand requirements for thymic selection and HP. Positive selection of 6C5 T cells is dependent on TAP function, thus selection is presumably mediated by TAP-dependent peptides. By contrast, HP occurs in TAP−/− recipients, providing an example where the ligand requirements for HP are less stringent than for thymic selection.

TCR transgenic 6C5 T cells recognize an insulin B chain epitope presented by the nonclassical class I MHC molecule, Qa-1b. Positive selection of these T cells was shown previously to require Qa-1b. Despite dedicated specificity for Qa-1b, evidence presented in the current study indicates that 6C5 T cells can cross-recognize a classical class I molecule. Clonal deletion was observed unexpectedly in 6C5.H-2bxq mice, which do not express I-E MHC class II molecules and thus should not be subject to superantigen-mediated negative selection. 6C5 T cells were observed to respond in vivo and in vitro to spleen cells from allogeneic H-2q mice, and specificity was mapped to Dq. Evidence was obtained for direct recognition of Dq, rather than indirect presentation of a Dq-derived peptide presented by Qa-1b. Polyclonal CD8+ T cells from class Ia-deficient KbDb−/− mice reacted in vitro to allogeneic spleen cells with an apparent frequency comparable to conventional class Ia-restricted T cells. Our results provide a clear example of a Qa-1-specific TCR that can cross-react with a class Ia molecule and evidence supporting the idea that this may be a common property of T cells selected by class Ib molecules.