Human cytomegalovirus (HCMV) encodes several proteins that inhibit major histocompatibility complex (MHC) class I-dependent antigen presentation. The HCMV products US2 and US11 are each sufficient for causing the dislocation of human and murine MHC class I heavy chains from the lumen of the endoplasmic reticulum to the cytosol, where the heavy chains are readily degraded. The apparent redundancy of US2 and US11 has been probed predominantly in cultured cell lines, where differences in their specificities were shown for murine and human MHC class I locus products. Here, we expressed US11 and US2 via adenovirus vectors and show that US11 exhibits a superior ability to degrade MHC class I molecules in primary human dendritic cells. MHC class II complexes are unaffected by US2-and US11-mediated attack. We suggest that multiple HCMV-encoded immunoevasions have evolved complementary functions in response to diverse host cell types and tissues. The CD8ϩ T-cell response plays a key role in containing viral infections. Induction of a T-cell response requires the display of virus-derived peptides by host major histocompatibility complex (MHC) class I molecules. The ability to interfere with antigen presentation has been documented for a number of viruses, including human cytomegalovirus (HCMV) (21). HCMV evades detection by the immune system by a variety of functions encoded by the genomic US and UL regions. Two of these gene products, US2 and US11, target class I heavy chains (HCs) for dislocation from the endoplasmic reticulum (ER) to the cytosol (23, 24). The HCs are then deglycosylated by Nglycanase and subsequently degraded by the proteasome. US2 and US11 are both ER-resident membrane glycoproteins, and they are expressed concomitantly during HCMV infection. Their general modes of action appear to be similar, but in mouse cells they differ in their abilities to attack allelic class I HC products (15). Perhaps in the human system, US2 and US11 also differ in specificity of interaction with class I molecules. Indeed, locus-specific preferences have been described: US2 and US11 mediate degradation of HLA-A and -B locus products but not HLA-C and -G locus products (18). Data obtained for a soluble, recombinant fragment of US2 likewise suggest differences in interaction with HLA-A and HLA-B locus products (9). Interestingly, a possible interaction between US2 and MHC class II products that results in downregulation of DM-␣ and DR-␣ has also been reported (20), although these observations were made with cells that do not normally express MHC class II molecules.In vivo, HCMV infection can enhance MHC class I expression on bystander cells by induction and release of beta interferon. Therefore, a second possibility is that cooperation of US2 and US11 is necessary in vivo in secondary rounds of infection to overcome the increased MHC class I expression in newly infected cells (14).In principle, multiple cell types can present antigen. HCMV can establish latency in hematopoietic progenitor cells and macrophages, while infection...

Dendritic cells (DC) either induce T cell tolerance or contribute to the initiation and modulation of T and B cell responses. Since many of the variables determining the thresholds of naive T cell priming were defined in vitro using a homogeneously matured DC population, we here focused on partially mature DC which might reflect the occurrence of tumor-infiltrating and thymic DC. To predict how those DC regulate the induction of antigen-specific T cell proliferation and T cell tolerance, we co-cultured ovalbumin-pulsed murine DC at different ratios with antigen-specific DO11.10 transgenic T cells. Whereas partially mature DC at a DC/T cell ratio of 1 : 10 supported proliferation, a DC/T cell ratio of 1 : 2 induced proliferation arrest in naive CD4 + T cells. The acquisition of the NK cell inhibitory markers NK1.1 and KLRG on T cells exposed to high numbers of DC suggests a role for these molecules in the protection of antigen-responsive T cells from exhaustion by overstimulation. Mechanistically, abortive T cell proliferation upon encounter of high numbers of partially mature DC is caused by an apoptosis-related pathway, suggesting that excessive antigen density without sufficient costimulation results in activation-induced cell death.

Background: Tumor-associated antigens are appreciated as diagnostic markers, but they have also prompted tremendous efforts to develop tumor-specific immunotherapy. A previously cloned tumor-associated antigen, EBAG9, was initially defined by reactivity with the monoclonal antibody 22-1-1. Functionally, the EBAG9-encoded gene-product was believed to induce apoptosis in activated immune cells. However, using a cell-biological approach we identified EBAG9 as a Golgiresident modulator of O-linked glycan expression, the latter product was then recognized by the 22-1-1 antibody. Secondly, EBAG9 expression was found physiologically in all murine tissues examined. This raised the question if EBAG9 is tumor-specific and mediates apoptosis itself or through O-linked glycans generated, among them the cognate 22-1-1 antigen Tn.