N-linked glycan modification on antigen-presenting cells restores an allospecific cytotoxic T cell response

Four Faces of Cell-Surface HLA Class-I: Their Antigenic and Immunogenic Divergence Generating Novel Targets for Vaccines

Leukocyte cell-surface HLA-I molecules, involved in antigen presentation of peptides to CD8+ T-cells, consist of a heavy chain (HC) non-covalently linked to β2-microglobulin (β2m) (Face-1). The HC amino acid composition varies across all six isoforms of HLA-I, while that of β2m remains the same. Each HLA-allele differs in one or more amino acid sequences on the HC α1 and α2 helices, while several sequences among the three helices are conserved. HCs without β2m (Face-2) are also observed on human cells activated by malignancy, viral transformation, and cytokine or chemokine-mediated inflammation. In the absence of β2m, the monomeric Face-2 exposes immunogenic cryptic sequences on these cells as confirmed by HLA-I monoclonal antibodies (LA45, L31, TFL-006, and TFL-007). Furthermore, such exposure enables dimerization between two Face-2 molecules by SH-linkage, salt linkage, H-bonding, and van der Waal forces. In HLA-B27, the linkage between two heavy chains with cysteines at position of 67 of the amino acid residues was documented. Similarly, several alleles of HLA-A, B, C, E, F and G express cysteine at 67, 101, and 164, and additionally, HLA-G expresses cysteine at position 42. Thus, the monomeric HC (Face-2) can dimerize with another HC of its own allele, as homodimers (Face-3), or with a different HC-allele, as heterodimers (Face-4). The presence of Face-4 is well documented in HLA-F. The post-translational HLA-variants devoid of β2m may expose several cryptic linear and non-linear conformationally altered sequences to generate novel epitopes. The objective of this review, while unequivocally confirming the post-translational variants of HLA-I, is to highlight the scientific and clinical importance of the four faces of HLA and to prompt further research to elucidate their functions and their interaction with non-HLA molecules during inflammation, infection, malignancy and transplantation. Indeed, these HLA faces may constitute novel targets for passive and active specific immunotherapy and vaccines.

Playing hide and seek: Tumor cells in control of MHC class I antigen presentation

MHC class I (MHC-I) molecules present a blueprint of the intracellular proteome to T cells allowing them to control infection or malignant transformation. As a response, pathogens and tumor cells often downmodulate MHC-I mediated antigen presentation to escape from immune surveillance. Although the fundamental rules of antigen presentation are known in detail, the players in this system are not saturated and new modules of regulation have recently been uncovered. Here, we update the understanding of antigen presentation by MHC-I molecules and how this can be exploited by tumors to prevent exposure of the intracellular proteome. This knowledge can provide new ways to improve immune responses against tumors and pathogens.

The SPPL3-Defined Glycosphingolipid Repertoire Orchestrates HLA Class I-Mediated Immune Responses

HLA class I (HLA-I) glycoproteins drive immune responses by presenting antigens to cognate CD8+ T cells. This process is often hijacked by tumors and pathogens for immune evasion. Because options for restoring HLA-I antigen presentation are limited, we aimed to identify druggable HLA-I pathway targets. Using iterative genome-wide screens, we uncovered that the cell surface glycosphingolipid (GSL) repertoire determines effective HLA-I antigen presentation. We show that absence of the protease SPPL3 augmented B3GNT5 enzyme activity, resulting in upregulation of surface neolacto-series GSLs. These GSLs sterically impeded antibody and receptor interactions with HLA-I and diminished CD8+ T cell activation. Furthermore, a disturbed SPPL3-B3GNT5 pathway in glioma correlated with decreased patient survival. We show that the immunomodulatory effect could be reversed through GSL synthesis inhibition using clinically approved drugs. Overall, our study identifies a GSL signature that inhibits immune recognition and represents a potential therapeutic target in cancer, infection, and autoimmunity.

Variations in MHC Class II Antigen Processing and Presentation in Health and Disease

MHC class II (MHC-II) molecules are critical in the control of many immune responses. They are also involved in most autoimmune diseases and other pathologies. Here, we describe the biology of MHC-II and MHC-II variations that affect immune responses. We discuss the classic cell biology of MHC-II and various perturbations. Proteolysis is a major process in the biology of MHC-II, and we describe the various components forming and controlling this endosomal proteolytic machinery. This process ultimately determines the MHC-II-presented peptidome, including cryptic peptides, modified peptides, and other peptides that are relevant in autoimmune responses. MHC-II also variable in expression, glycosylation, and turnover. We illustrate that MHC-II is variable not only in amino acids (polymorphic) but also in its biology, with consequences for both health and disease.

HLA class I transgenic mice: development, utilisation and improvement

Classical major histocompatibility complex (MHC) class I antigens are trimeric molecules found on the surface of nucleated cells in all jawed vertebrates. MHC I are recognised by two families of receptors: clonotypic T cell receptors expressed on the surface of CD8+ cytotoxic T lymphocytes (CTLs), and monomorphic receptors expressed by both natural killer cells and CTLs. The production of MHC I molecules within the cells is a sequential process performed with the help of interacting proteins: proteases, chaperones, transporters and so on. Although largely homologous in their structure, organisation and function, the human and mouse MHC I antigen processing and presentation machineries show fine differences. Transgenesis and 'knockout' or 'knock-in' technologies permit the addition of relevant human genes or the replacement of mouse genes by their human orthologues in order to produce immunologically humanised mice. Such experimental animals are especially relevant for the comparative evaluation of immunotherapies and for the characterisation of MHC I peptide epitopes. This review presents the similarities and differences between mouse and human MHC I antigen processing machinery, and describes the development and utilisation of improving mouse models of human cytotoxic T cell immunity.

A carbohydrate neoepitope that is up-regulated on human mononuclear leucocytes by neuraminidase treatment or by cellular activation

The expression of cell-surface antigens can delineate specific leucocyte developmental or functional stages. For example, certain membrane glycoproteins are expressed selectively on leucocyte subsets only after activation. Leucocyte activation can also induce changes in carbohydrate epitopes expressed on surface antigens. In the present studies, we report on a novel monoclonal immunoglobulin M antibody (mAb 13.22) that recognizes a unique carbohydrate epitope expressed on human leucocyte membrane proteins. Characterization of mAb 13.22 specificity by immunoblotting showed that it recognized proteins of MW approximately 95 000 and 150 000, including both CD18 and CD11b. The mAb 13.22 epitope was removed by N-glycosidase F but not by endoglycosidase H or fucosidase, demonstrating that it is an N-linked carbohydrate antigen. Interestingly, immunoblot staining was enhanced after neuraminidase treatment, suggesting that the antibody epitope might also be partially masked by sialic acid. In resting leucocytes, the mAb 13.22 antigen was expressed strongly on neutrophils, while dull staining was present on monocytes, and no lymphocyte staining was observed. In marked contrast, treatment of leucocytes with neuraminidase resulted in exposure of a mAb 13.22 neoepitope on a subset of lymphocytes (primarily T lymphocytes and natural killer cells) as well as up-regulated staining more than 18-fold on monocytes. Activation of lymphocytes in culture with phytohaemagglutinin or concanavalin A also unmasked the mAb 13.22 neoepitope on approximately 37% of the CD45RO+ lymphocytes. Furthermore, analysis of leucocytes collected from the synovial fluid of patients with rheumatoid arthritis showed that approximately 18% of the lymphocytes present expressed the mAb 13.22 neoepitope. Taken together, our results suggest that the mAb 13.22 carbohydrate neoepitope could represent a physiologically relevant marker that is up-regulated on leucocyte subsets during the inflammatory response.

N-linked carbohydrate on human leukocyte antigen-C and recognition by natural killer cell inhibitory receptors

The possible role of carbohydrate in the interaction of HLA-C with a human inhibitory natural Killer cell Immunoglobulin-like Receptor with two Ig domains, KIR2DL1, was investigated. Transfectants of 721.221 (a class I MHC-negative human B cell line) expressing only HLA-Cw4 or -Cw6 or their respective non-glycosylated mutants (N86Q, S88A) were made. The binding of a KIR2DL1-Ig fusion protein to the non-glycosylated mutant HLA-Cw4- or -Cw6-expressing cells was markedly decreased compared to the wild type-expressing cells. The ability to induce an inhibitory signal in the NK tumor line YTS transfected with KIR2DL1 was also impaired in the nonglycosylated mutant expressing cells. Furthermore, in a second functional assay, mutant HLA-Cw4 and -Cw6 molecules had impaired ability to induce signal transduction in BW cells expressing a KIR2DL1-CD3 zeta chain chimeric protein. Thus, the deletion of the N-linked glycosylation signal in HLA-Cw4 and -Cw6 greatly reduced recognition by KIR2DL1. Alternative interpretations of the data are discussed.

Human dendritic cells handling of binding, uptake and degradation of free and IgG-immune complexed dinitrophenylated human serum albumin in vitro

The handling of free and IgG-complexed dinitrophenylated human serum albumin (DNP-HSA) by human dendritic cells (DC) cultured with granulocyte-macrophage colony-stimulating factor (GM-CSF) and interleukin-4 (IL-4) was studied. It has been shown that the amount of uncomplexed or IgG-complexed antigen required by DC to start an immune response is low compared with other antigen-presenting cells. We therefore examined whether such efficient presentation of immune complexes is due to an enhanced Fc gamma RII-mediated endocytosis or to a specialized and efficient antigen handling, i.e., macropinocytosis. The Fc gamma RII expression was found to be heterogeneous on the GM-CSF- and IL-4-cultured DC, i.e. it ranges from low to high expression. The handling of antigen and immune complexes revealed, that the level of binding and uptake of IgG-DNP-HSA complexes by in vitro expanded DC is low compared with free antigen. Uncomplexed DNP-HSA is probably handled either by endocytosis via receptors being more abundant and/or efficient than the Fc gamma RII or via non-receptor-mediated endocytosis. The binding and uptake of IgG-complexed DNP-HSA was blocked by anti-Fc gamma RII antibody, indicating the specificity of the interaction.

Dendritic cells in immune response induction

The study of dendritic cells (DCs) has seen a rapid expansion in recent years, and their importance within the immune system is now widely recognized. Along with B lymphocytes and mononuclear phagocytes, DCs make up what are known as the professional antigen-presenting cells (APCs). These are cells which are capable of highly efficiently presenting antigens to the immune system in the context of both major histocompatibility complex class I and class II molecules. What makes DCs stand out from other professional APCs, however, is their seemingly unique ability to present antigen to T lymphocytes which have had no previous contact with antigen. This gives DCs central role in the initiation of immune responses, and creates possibilities for their exploitation in the development of therapeutic strategies against tumors and other diseases. What are the characteristics of DCs which enable them to carry out their specialized function? This is a question which is currently gaining much interest. While higher expression levels of the antigen-presentation machinery may account for this, there may also be as yet unidentified mechanisms at work. In this review, we will discuss the evidence for DC-mediated priming of both CD4+ and CD8+ naive T cells, both in vitro and in vivo, current ideas on how DCs achieve their potent function and the implications for the design and execution of immunotherapeutic strategies.

The effects of the site-directed removal of N-glycosylation sites from beta-1,4-N-acetylgalactosaminyltransferase on its function

The amino acid sequence deduced from the cloned human cDNA of beta-1,4-N-acetylgalactosaminyltransferase (GalNAc-T; EC 2.4.1.92) gene predicted three potential sites for N-linked glycosylation. Although many glycosyltransferases isolated contain from 2 to 6 N-glycosylation sites, their significance has not been adequately demonstrated. To clarify the roles of N-glycosylation in GalNAc-T function, we generated a series of mutant cDNAs, in which some or all of the glycosylation recognition sites were eliminated by polymerase chain reaction (PCR)-mediated site-directed mutagenesis. Using transcription/translation in vitro, we confirmed that all potential N-glycosylation sites could be used. Although cell lines transfected with mutant cDNAs showed equivalent levels of GalNAc beta 1-->4(NeuAc alpha 2-->3)Gal beta 1-->4Glc-Cer (GM2) to that of the wild-type, the extracts from mutant cDNA transfectants demonstrated lower enzyme activity than in the wild-type. The decrease in enzyme activity was more evident as the number of deglycosylated sites increased, with about 90% decrease in a totally deglycosylated mutant. The enzyme kinetics analysis revealed no significant change of Km among wild-type and mutant cDNA products. The intracellular localization of GalNAc-T expressed in transfectants with wild-type or mutant cDNAs also showed a similar perinuclear pattern (Golgi pattern). These results suggest that N-linked carbohydrates on GalNAc-T are required for regulating the stability of the enzyme structure.

Infection with Listeria monocytogenes impairs sialic acid addition to host cell glycoproteins

Listeria monocytogenes is a facultative intracellular bacterium that causes severe disease in neonates and immunocompromised adults. Although entry, multiplication, and locomotion of Listeria in the cytosol of infected cells are well described, the impact of such infection on the host cell is unknown. In this report, we investigate the effect of L. monocytogenes infection on MHC class I synthesis, processing, and intracellular trafficking. We show that L. monocytogenes infection interferes with normal processing of N-linked oligosaccharides on the major histocompatibility complex (MHC) class I heavy chain molecule, H-2Kd, resulting in a reduced sialic acid content. The glycosylation defect is more pronounced as the infection progresses and results from interference with the addition of sialic acid rather than its removal by a neuraminidase. The effect is found in two different cell lines and is not limited to MHC class I molecules since CD45, a surface glycoprotein, and LGP120, a lysosomal glycoprotein, are similarly affected by L. monocytogenes infection. The glycosylation defect is specific for infection by L. monocytogenes since neither Trypanosoma cruzi nor Yersinia enterocolitica, two other intracellular pathogens, reproduces the effect. The resultant hyposialylation of H-2Kd does not impair its surface expression in infected cells. Diminished sialic acid content of surface glycoproteins may enhance host-defense by increasing susceptibility to lysis and promoting clearance of Listeria-infected cells.

Efficient presentation of soluble antigen by cultured human dendritic cells is maintained by granulocyte/macrophage colony-stimulating factor plus interleukin 4 and downregulated by tumor necrosis factor alpha

Using granulocyte/macrophage colony-stimulating factor (GM-CSF) and interleukin 4 we have established dendritic cell (DC) lines from blood mononuclear cells that maintain the antigen capturing and processing capacity characteristic of immature dendritic cells in vivo. These cells have typical dendritic morphology, express high levels of major histocompatibility complex (MHC) class I and class II molecules, CD1, Fc gamma RII, CD40, B7, CD44, and ICAM-1, and lack CD14. Cultured DCs are highly stimulatory in mixed leukocyte reaction (MLR) and are also capable of triggering cord blood naive T cells. Most strikingly, these DCs are as efficient as antigen-specific B cells in presenting tetanus toxoid (TT) to specific T cell clones. Their efficiency of antigen presentation can be further enhanced by specific antibodies via FcR-mediated antigen uptake. Incubation of these cultured DCs with tumor necrosis factor alpha (TNF-alpha) or soluble CD40 ligand (CD40L) for 24 h results in an increased surface expression of MHC class I and class II molecules, B7, and ICAM-1 and in the appearance of the CD44 exon 9 splice variant (CD44-v9); by contrast, Fc gamma RII is markedly and sometimes completely downregulated. The functional consequences of the short contact with TNF-alpha are in increased T cell stimulatory capacity in MLR, but a 10-fold decrease in presentation of soluble TT and a 100-fold decrease in presentation of TT-immunoglobulin G complexes.

Mechanisms of induction of primary virus-specific cytotoxic T lymphocyte responses

We have investigated the ability of various antigen-presenting cell (APC) types to induce primary anti-viral cytotoxic T lymphocyte (CTL) responses by single in vitro stimulation. Of these APC types, only dendritic cells (DC) and RMA-S lymphoma cells could induce primary CTL responses, but by divergent mechanisms. DC were capable of generating primary virus-specific CTL, either by presenting viral peptide or processed infectious virus. In contrast, RMA-S cells could not present endogenous antigen, e.g. after virus infection, but this cell line very efficiently presented exogenous viral peptides to induce primary virus-specific CTL in vitro. Spleen cells, lipopolysaccharide-induced B cell blasts or the non-mutated RMA cells did not have the ability to trigger unprimed T cells by single in vitro stimulation. We have investigated several characteristics important for primary CTL response induction by DC and RMA-S cells (summarized in Fig. 6). Primary CTL response induction by DC or RMA-S cells was blocked by anti-LFA-1 or anti-CD8 monoclonal antibodies (mAb). DC rapidly aggregated with unprimed T cells, which was independent of LFA-1 and CD8 molecules. RMA-S cells did not form conjugates with unprimed T cells. Despite their abundant major histocompatibility complex (MHC) class I cell-surface expression, DC did not bind much exogenously added viral peptide. In contrast, the MHC class I molecules on RMA-S cells bound a large quantity of exogenously administered peptide. Powerful adhesion by DC and high expression of relevant MHC/peptide complexes on RMA-S cells are important features in the initial contact with unprimed T lymphocytes. In a later stage of contact, both DC and RMA-S cells activate LFA-1 (and CD8) molecules at the T cell surface to strengthen and maintain the contact between T cell and APC.

Processing and presentation of intact hen egg-white lysozyme by dendritic cells

Dendritic cells in lymphoid tissues are of key importance as highly specialized antigen-presenting cells for the induction of T lymphocyte responses. Conflicting results have been published regarding antigen processing of intact proteins by dendritic cells. We now report that highly purified dendritic cells isolated from H-2k mouse spleens very efficiently generated immunogenic fragments of intact hen egg-white lysozyme (HEL) protein to present to an I-Ak-restricted T hybridoma cell line, specific for HEL peptide 46-61. Dendritic cells required 100 times less HEL protein than lipopolysaccharide-induced B cell blasts for effective presentation. Uptake of 125I-labeled HEL protein by dendritic cells and inhibition of presentation of HEL protein by chloroquine treatment was observed. This indicates an endocytotic process and the involvement of acidified compartments. Since the supernatant of dendritic cells, that were incubated with intact HEL protein, contained immunogenic fragments, further evidence for processing of HEL protein by dendritic cells was obtained. When HEL protein was covalently coupled to beads, dendritic cells were not able to ingest these beads, but could still process HEL protein for presentation. This suggests cell surface processing of HEL protein, although internalization of HEL protein released from the beads cannot be excluded. Taken together, these data show that H-2k dendritic cells are capable of processing and presenting intact HEL protein.

The ligand recognized by ELAM-1 on HL60 cells is not carried by N-linked oligosaccharides

The sialyl Lewis-x determinant is a ligand for ELAM-1, a major adhesion molecule for HL60 cells and neutrophils. ELAM-1 expression can selectively be induced on human umbilical vein endothelial cells (HUVEC) by tumor necrosis factor, interleukin 1 and lipopolysaccharide. The determinant sialyl Lewis-x is found on both glycolipids as well as on glycoproteins. Using specific inhibitors of the biosynthesis of N-linked glycosylated glycoproteins, we investigated whether N-linked glycans or their modifications are involved in ELAM-1-dependent adhesion of HL60 cells to activated HUVEC. The inhibitors of glycoprotein processing N-methyl-deoxynojirimycin, 1-deoxymannojirimycin and swainsonine did not affect ELAM-1-dependent adhesion. Complex-type N-linked glycans are not required for ELAM-1 mediated adhesion, and therefore the ligand for ELAM-1 is most likely a glycolipid, or a glycoprotein carrying O-linked oligosaccharides.

Positive correlation between oligonucleotide typing and T-cell recognition of HLA-DP molecules

The identification of 19 different HLA-DPB1 sequences implicates the existence of more DP specificities than can be typed for with cellular methods. How many of the DP beta sequences can be specifically recognized by T cells, and which of the polymorphic regions can contribute to the specificity of allorecognition, is not known. In order to investigate the distribution and the immunological relevance of recently described DPB1 alleles, we have typed a panel of 98 randomly selected Dutch Caucasoid donors for the HLA-DPB1 locus by oligonucleotide typing. Comparison of the typing results with primed lymphocyte typing (PLT) defined DP specificities shows an extremely good correlation. Moreover, additional alleles could be defined by oligonucleotide typing reducing the number of DP blanks in the panel. By selecting the appropriate responder stimulator combinations we were able to show that distinctive PLT reagents against oligonucleotide defined specificities DPB1*0401, DPB1*0402, DPB1*0901, and DPB1*1301 can be generated. To investigate in more detail which part of the DP molecule is responsible for the specificity of T-cell recognition, T-cell clones were generated against HLA-DPw3. The clones were tested for the recognition of stimulators carrying DPB1 alleles which had been defined by oligonucleotide typing and sequence analyses and which differed in a variable degree from DPB1*0301. The recognition patterns demonstrated that differences of one amino acid in polymorphic regions situated either in the beta sheets or alpha helix of the hypothetical model of the HLA class II molecule can eliminate T-cell recognition. Furthermore, sequence analyses revealed a new DPB1 allele designated DPB1*Oos.

Class II major histocompatibility complex molecules of murine dendritic cells: synthesis, sialylation of invariant chain, and antigen processing capacity are down-regulated upon culture

Dendritic cells (DCs), such as Langerhans cells (LCs) of the epidermis and the DCs of lymphoid organs such as spleen, are potent antigen presenting cells. DCs express high levels of major histocompatibility complex (MHC) class II molecules, but, partly because of the low numbers of primary DCs in any tissue, there has been no detailed study of the biochemistry of their class II molecules. This information may be needed to help explain recent findings that DCs process native protein antigens when freshly isolated from epidermis and spleen. Processing ceases during culture, yet a strong accessory function for activating resting T cells develops. We studied immunoprecipitates of DC class II and invariant chain (Ii) molecules by two-dimensional gel electrophoresis. We found that (i) freshly isolated LCs synthesize large amounts of class II and Ii polypeptides; (ii) Ii molecules that are known to be involved in antigen processing display an unusually large number of sialic acids in fresh LCs; (iii) with culture, class II and Ii synthesis decreases dramatically and has virtually ceased at 3 days; and (iv) the turnover of class II in pulse/chase experiments is slow, being undetectable over a 12- to 32-hr culture period, whereas the turnover of Ii is rapid. We conclude that MHC class II molecules of DCs do not seem to be qualitatively unique. However, the regulation of class II and Ii expression is distinctive in that biosynthesis proceeds vigorously for a short period of time and the newly synthesized class II remains stably on the cell surface, whereas Ii turns over rapidly. This may enable DCs to process and retain antigens in the peripheral tissues such as skin and migrate to the lymphoid organs to activate T cells there.

Carbohydrate synthesis inhibitors decrease interleukin 1-stimulated lymphocyte binding to endothelial cells

Lymphocyte extravasation (homing) is initiated when lymphocyte adheres to endothelial cells. All know protein structures involved in the lymphocyte binding located on the endothelial surface are heavily glycosylated. We asked therefore whether these carbohydrate motifs had a role in the lymphocyte homing. The relative importance of the N-linked chains on biological effects mediated by glycoproteins can be studied with specific inhibitors of carbohydrate processing, i.e. 1-deoxynojirimycin (DN), castonospermine (CAST), 1-deoxymannojirimycin (DMN) and swainsonine (SW), which produces different kinds of blocked carbohydrate chains. N-linked carbohydrate chains are modified in the Golgi apparatus and in the final glycoprotein they can be either of high-mannose-, hybrid- or complex-type motifs, having the same core structure but different terminal structures. We show here that when all N-linked carbohydrates were cleaved off from the cell surface glycoproteins by treating endothelial cells with N-glycosidase F interleukin 1-induced lymphocyte binding was reduced almost to non-stimulated control values. Treatment of endothelial cells with CAST led to generation of glycoproteins carrying high-mannose-type oligosaccharides, which are glucose capped with three glucose molecules on the chain. CAST treatment resulted in an 85% decrease in lymphocyte binding compared to interleukin 1-induced levels. DMN treatment, resulting in accumulation of high-mannose type oligosaccharides without any terminal glucoses on the cell surface, caused a similar inhibition of lymphocyte binding to that induced by CAST treatment. SW treatment, leading to accumulation of hybrid-type glycoproteins, decreased only slightly the lymphocyte binding. These results suggest that carbohydrates indeed have a role in lymphocyte binding to endothelial cells.