IN THIS ISSUE Free

Peptides associate with class II MHC molecules in the endoplasmic reticulum of dendritic cells (DC). Then, the loaded MHC molecules are transported via polarized tubules to the cell surface at the point of contact of the DC with an Ag-specific T cell. Bertho et al. (p. 5689 ) determined the cell surface requirements of both cell types in endosomal tubulation. DC from bone marrow of knockin mice expressing enhanced green fluorescent protein-I-A^b fusion molecules were incubated with naive CD4⁺ T cells from transgenic mice that expressed an I-A^b-restricted TCR specific for OVA (OTII). Maximal tubule formation occurred with DC that had endocytosed 40 μM OVA for 4 h before addition of OTII cells. Pretreatment of OTII cells with mAb to CD2 or LFA-1 or fixation of OTII cells with either glutaraldehyde or paraformaldehyde blocked polarized endosomal tubulation in the DC. Activation of OTII cells by several reagents increased tubulation in OVA-loaded DC compared with naive OTII cells. Induced tubulation correlated with induced expression of CD40L on the T cells; pretreatment of the T cells with an anti-CD40L mAb reduced tubulation, whereas an anti-CD28 mAb had no effect. Endosomal tubular extensions were polarized toward T cells in 80% of DC that had endocytosed intact OVA compared with <10% of fixed DC loaded with synthetic OVA peptide. The authors conclude that rearrangement of membrane proteins on the surface of Ag-specific T cells and the existence of specific microdomains on the membranes of DC induce tubular endosome formation in DC.

Although self-tolerance is established primarily within the thymus, induction of tolerance also occurs in mature CD4⁺ T cells in the periphery by multiple mechanisms. Ali et al. (p. 6290 ) generated mice transgenic for Vβ5⁺ TCR (Vβ5 Tg) on a C57BL/6 (B6) background. Bone marrow cells from these mice were injected into irradiated B6 mice, and CD4⁺ T cell tolerance to an endogenous retroviral Ag that reacts weakly with the Vβ5⁺ TCR was measured. CD4⁺ T cell deletion was measured as a decline in the CD4:CD8 ratio; TCR revision, which involves rearrangement of TCR β-chain genes, was measured as an accumulation of Vβ5⁻ CD4⁺ peripheral T cells. The extent of deletion was the same in the radiation chimeras and in control Tg mice, whereas TCR revision in the radiation chimeras was delayed longer than 40 wk compared with the controls. Deletion of CD4⁺ T cells in Vβ5 Tg mice deficient in Fas was the same as in wild-type Tg mice, but there was an acceleration of TCR revision in the absence of Fas. Expression of a CD8 transgene on CD4⁺ T cells did not affect either deletion or revision. CD4 lineage cells in Vβ5 Tg in a CD4 null background had a deletion pattern comparable to that seen in CD4 wild-type mice; however, TCR revision decreased compared with controls. Vβ5 Tg mice deficient for inducible costimulatory molecule had the same deletion kinetics but showed a 10-mo delay in TCR revision compared with wild-type mice. The data demonstrate differential regulation of deletion and TCR revision in the induction of peripheral tolerance within the same population of CD4⁺ T cells.

Interaction between CD40 and its ligand CD154 is required for in vivo humoral immune responses. However, the expression of these molecules on a variety of cells in the immune system raises the question as to whether their expression on B cells is required for optimal responses. Lee et al. (p. 5707 ) developed four groups of bone marrow (BM) chimeras in lethally irradiated mice genetically unable to produce B cells. Injected cells were: a mixture of host BM and wild-type BM (WT chimeras), CD40^−/− BM (CD40-KO chimeras), a mixture of host BM and CD40^−/− BM (CD40-B cell knockout, or CD40-BKO, chimeras), or a mixture of host BM and CD154^−/− BM (CD154-B cell knockout, or CD154-BKO, chimeras). OVA-specific TCR transgenic CD4⁺ T cells were adoptively transferred into the chimeras and the animals immunized with OVA. By seven days, OVA specific B cells, germinal centers and plasma cells expressing IgM, IgG, IgG1, and IgE were detected in all groups of chimeras but the numbers were lowest in the CD40-KO and CD40-BKO mice. Affinity levels of total IgG were higher in WT and CD154-BKO chimeras compared with CD-40KO and CD40-BKO chimeras. The adoptively transferred transgenic CD4⁺ T cells expanded the least in the OVA-immunized CD40-KO mice compared with the other groups. Germinal center and humoral immune responses after infection of the chimeras with influenza virus were similar to those seen following OVA immunization. The results support the requirement for CD40, but not CD154, on B cells for normal humoral immune responses in vivo.

Activated microglia both protect and destroy the CNS through the secretion of inflammatory cytokines. Apoptosis of activated microglia might protect against neurodegenerative diseases. Lee et al. (p. 5802 ) looked for a mediator in the increased sensitivity of LPS/IFN-γ-treated microglia to NO-induced apoptosis. Protein and mRNA expression levels of B cell translocation gene 1 (BTG1), identified by subtractive hybridization of activated BV-2 mouse microglial cell cDNA with unactivated BV-2 cDNA, were increased by LPS/IFN-γ treatment but not by LPS or IFN-γ treatment alone. This increased expression of BTG1 in BV-2 cells correlated with increased cell death and decreased proliferation in the presence of an inhibitor of NO synthase. The induction of BTG1 expression in the presence of nontoxic levels of NO increased the sensitivity of microglial cells to apoptosis. There was no difference in the levels of inflammatory mediators expressed by LPS/IFN-γ-treated or untreated cells. Transfection of a dominant negative mutant of STAT1 into BV-2 cells partially abolished the sensitizing effect of the LPS/IFN-γ treatment on NO-induced apoptosis. Moreover, an inhibitor of Janus kinase decreased both LPS/IFN-γ-induced BTG1 expression and NO production and reduced cell death. BTG1 expression in wild-type and IRF-1 gene deficient microglia was enhanced to the same level by LPS/IFN-γ. The authors conclude that BTG1 is a mediator of apoptosis induced by LPS/IFN-γ and suggest that understanding BTG1 action could lead to new treatments for neurodegenerative diseases.

Macrophages and immature dendritic cells recognize and remove apoptotic cells. However, the molecules expressed on the phagocytes and recognized on the engulfed cells are not known. Tada et al. (p. 5718 ) developed two mAbs against BAM3, a mouse macrophage cell line established by transformation of peritoneal macrophages with SV40, that inhibited phagocytosis of apoptotic thymocytes by 91%. One mAb immunoprecipitated the long and short forms of SH2-domain bearing protein tyrosine phosphatase substrate-1 (SHPS-1). Expression of SHPS-1 mRNA was detected in macrophages but not in lymphoid, myeloid, and fibroblast cell lines. A soluble form of SHPS-1 inhibited phagocytosis of apoptotic thymocytes by BAM3. SHPS-1 ligand, CD47, was detected on normal and apoptotic thymocytes. Lymphoma cells transfected with an expression vector for CD47 and treated with FasL were engulfed by BAM3 cells, whereas nonapoptotic cells were not engulfed. Anti-SHPS-1 mAb inhibited the engulfment. Fluorescein-labeled lymphoma cells transfected with CD47 and treated with FasL were injected into mice. The injected cells were found in splenic dendritic cells to a greater percentage than FasL-treated, fluorescein-labeled nontransfected lymphoma cells. Phagocytosis of apoptotic cells by BAM3 was inhibited in the presence of a molecule that bound phosphatidylserine on apoptotic cells. The authors conclude that CD47-SHPS-1 acts as a tethering bridge in an engulfment process requiring phosphatidylserine.

Dendritic cells (DC) are used in tumor therapy to deliver tumor associated Ags and to activate naive CD4⁺ and CD8⁺ T cells. But the resulting adaptive and innate anti-tumor immune responses are not well characterized. On p. 5842 , van den Broeke et al. report their surprising finding that unpulsed DC protected mice against lung metastases induced by CT26 rhabdomyosarcoma cells compared with untreated or spleen cell-inoculated control animals. All routes of DC injection were equally protective, and the protection lasted at least 14 mo. Treatment of mice with anti-asialo-GM1 Ab, but not with anti-CD8 Ab, before tumor challenge decreased the protection. DC injection into mice severely deficient in NK cells did not protect against tumor challenge. Mature NK cells were detected immunohistochemically in spleen and lung sections of DC-inoculated wild-type mice. Injection of DC from mice deficient in CD40 reduced the level of protection somewhat; injection of DC from mice deficient in CD80/86 resulted in loss of protection. Injection of DC from mice deficient in IL-12 or IL-15 had no effect. Maturation of splenic DC with CD40L was required for the protective activity. No protection against tumor challenge was seen after injection of mature DC into strains of mice lacking T cells, and treatment of mice with anti-CD4 Ab before DC injection and tumor challenge eliminated the protective effect. The authors conclude that CD4⁺ T cells act as a bridge in the DC activation of NK cell-mediated tumor protection.

Compromised epithelial barrier function occurs in the intestinal mucosa of patients with inflammatory bowel disease (IBD). Elevated levels of cytokines are found in IBD mucosa but their contributions to paracellular permeability and apoptosis of colonic crypt cells are not defined. Bruewer et al. (p. 6164 ) found an increase in flux of a paracellular solute in monolayers of T84 epithelial cells treated for 72 h with IFN-γ, with or without TNF-α; no change in flux was seen after treatment with TNF-α alone vs an untreated control. Cells exposed to IFN-γ, with or without TNF-α, had increased levels of a caspase-3 cleavage product compared with TNF-α-treated and control cultures. However, a caspase inhibitor was unable to inhibit the cytokine-induced increase in paracellular permeability. Incubation of monolayers for 48–72 h with IFN-γ, with or without TNF-α, resulted in a redistribution of several transmembrane proteins away from tight junctions (TJ), whereas a TJ cytoplasmic plaque protein was minimally affected. Combined TNF-α and IFN-γ treatment decreased the level of one protein in the subjacent adherens junction (AJ). A different TJ protein and the same AJ protein showed a decrease in association with cytoskeletal elements and with detergent-insoluble glycolipid rafts after TNF-α and IFN-γ combined treatment; other TJ and AJ protein interactions were not affected. The authors conclude that increased paracellular permeability mediated by the two cytokines are separable from their pro-apoptotic effects and are associated with TJ and AJ restructuring in IBD.

Summaries written by Dorothy L. Buchhagen, Ph.D.