Search

Cell-cell adhesion was examined in the developing chick neural retina. A dorsoventral gradient of adhesive specificity in dissociated cells was detected which exhibits a complementarity such that the highest cell-cell affinities are exhibited between cells derived from the extremes of the gradient. If a nasotemporal gradient exists it must exhibit significantly lower cell-cell affinity. The relevance of these findings to pattern formation in the nervous system is discussed.

T lymphocytes recognize endogenously produced antigenic peptides in association with major histocompatibility complex (MHC)-encoded molecules. Peptides from the extracellular fluid can be displayed in association with class I and class II MHC molecules. Here we report that mature Kb class I MHC molecules bind peptides upon dissociation and reassociation of their light chain. Intact Kb heterodimers, unlike class II MHC molecules, are relatively unreceptive to binding peptides. This property may maintain segregation of class I and class II MHC-restricted peptides and has implications for the use of peptides as vaccines.

Murine thymocytes induce the monokine interleukin 1 upon in vitro coculture with a radioresistant Ia-bearing accessory cell [murine Ia molecule is a class II major histocompatibility complex (MHC) antigen]. The generation of interleukin 1 is critically dependent on the function of I-region gene products on accessory cells. The induced interleukin 1 appears to allow the activation and proliferation of self-MHC-specific thymocytes. Thus, in the absence of added exogenous factors, there is an Ia-dependent thymocyte proliferation. This selective activation of thymocytes is observed with both mature and immature thymic T cells. This in vitro response results in the selective amplification of developing T cells with self-MHC specificity and could be of importance to the in vivo commitment of T cells to MHC determinants that occurs in the thymus.

Any effect of serum on the antigenicity of peptides is potentially relevant to their use as immunogens in vivo. Here we demonstrate that serum contains distinct proteases that can increase or decrease the antigenicity of peptides. By using a functional assay, we show that a serum component other than beta 2-microglobulin enhances the presentation of ovalbumin peptides produced by cyanogen bromide cleavage. Three features of this serum activity implicate proteolysis: it is temperature dependent, it results in increased antigenicity in a low molecular weight peptide fraction, and it is inhibited by the protease inhibitor leupeptin. Conversely, presentation of the synthetic peptide OVA-(257-264) is inhibited by serum. This inhibition is unaffected by leupeptin but is blocked by bestatin, a protease inhibitor with distinct substrate specificities. Implications for peptide-based vaccine design and immunotherapy are discussed.

Evidence from various antigen systems suggests that antigen processing can be one factor that determines the repertoire of immunogenic peptides. Thus, processing events may account for some of the disparity between the available and expressed helper T-cell repertoires. In this report, we demonstrate that the immunodominant T-cell determinant in ovalbumin [p323-339; ovalbumin-(323-339) heptadecapeptide] is processed differently by two genetically identical antigen-presenting cell lines, M12 and A20. The ovalbumin-specific T-cell-T-cell hybridomas, DO-11.10 and 3DO-54.8, were used to detect processed antigen. These T-T hybridomas have different fine specificities for the p323-339 determinant. A20 cells presented native ovalbumin well to both T-T hybridomas, whereas M12 cells presented native ovalbumin well to 3DO-54.8 but very inefficiently to DO-11.10. M12 and A20 cells effectively stimulated both T-T hybridomas with the same concentrations of the immunogenic synthetic peptide p323-339. Therefore, M12 cells and DO-11.10 can interact with each other, and both T-T hybridomas have similar sensitivities for the same immunogenic peptide. We conclude that genetically identical antigen-presenting cells can display heterogeneity in the fine processing of an immunodominant T-cell determinant, and synthetic model peptides that represent the minimal stimulatory sequence of a T-cell determinant are not necessarily identical to the structure of in vivo processed antigen. Heterogeneity in antigen processing by individual antigen-presenting cells would serve to increase the repertoire of immunogenic peptides that are presented to T cells.

Several amino acid copolymers are potent immunogens under the control of major histocompatibility complex (MHC)-encoded Ir genes. We have further characterized their accessory-cell-dependent, MHC-restricted presentation to T lymphocytes. We initially characterized their processing requirements by investigating the ability of paraformaldehyde-fixed antigen-presenting cells (APC) to present these copolymers. Fixed APC can present poly(Glu56Lys35Phe9) and poly(Glu60Ala30Tyr10) provided that they have been incubated with antigen prior to fixation. The inability of these same fixed preparations to present soluble antigen indicates a fixation-sensitive antigen-processing step. In contrast, the antigens poly(Glu55Lys35Leu10) and poly(Glu55Lys35Tyr10) can be presented by APC fixed before antigen exposure. This differential requirement for antigen processing was exploited to analyze the events of antigen presentation in two related systems. First, the ability of isolated APC membranes to process and present antigen was assessed. APC membranes can present the antigens poly(GluLysLeu) and poly(GluLysTyr) in a specific and MHC-restricted manner. However, the isolated membranes fail to present either poly(GluLysPhe) or poly(GluAlaTyr), suggesting that such preparations can present but not process antigen. Second, the distinct properties of the various copolymers were used with fixed APC to test the effects of antigen processing on the phenomenon of antigen competition. APC that had processed poly(GluLysPhe) or poly(GluAlaTyr) were subsequently fixed and used to present antigen in the presence or absence of various antagonists. Under these conditions, poly(GluLysLeu) and poly(Glu50Tyr50) could effect specific inhibition, clearly indicating that antigen competition occurs distal to and does not require antigen processing. In contrast, native antigen with an absolute processing requirement is not capable of competing with preprocessed antigen on fixed APC. Taken together, these results suggest that processing is important for the molecular interactions between the copolymer antigens and the APC cell surface that are relevant to both antigen presentation and competitive inhibition.

Recently we described two murine T-cell membrane proteins, TAP (T-cell-activating protein) and TAPa (TAP-associated protein). Previous experiments suggested that TAP is involved in physiologic T-cell activation. The subject of this report is a genetic analysis of these molecules. TAP and TAPa map to the Ly-6 locus. The relationship of these molecules to other antigens encoded in this locus is examined. Based on tissue distribution, molecular structure, and functional properties, TAP is distinct from any previously described Ly-6 antigen, whereas TAPa is probably identical to the 34-11-3 antigen. TAP and TAPa are coexpressed on all cell types examined so far. Moreover, comparative studies demonstrate a complex developmentally regulated pattern in the expression of molecules encoded in this locus.

The Ly-6 locus encodes several cell surface proteins whose functions are unknown. Although it is hypothesized that these proteins may be receptors, there is no direct evidence that they bind a ligand. Herein we present evidence that Ly-6A.2, a Ly-6 protein expressed on T lymphocytes, binds a ligand expressed on normal thymocytes and splenic B and T cells. We find that transgenic thymocytes that overexpress Ly-6A.2 spontaneously aggregate in culture. This homotypic adhesion requires the overexpression of Ly-6A.2 because it is not observed in cultures of nontransgenic thymocytes. The aggregation of Ly-6A.2 transgenic thymocytes is inhibited by phosphatidylinositol-specific phospholipase C (which removes Ly-6A.2 and other glycosylphosphatidylinositol-anchored proteins from the membrane). Some anti-Ly-6A.2 monoclonal antibodies, including nonactivating ones and Fab' fragments, inhibit this aggregation. In contrast, other anti-Ly-6A.2 monoclonal antibodies increase the aggregation of transgenic but not nontransgenic thymocytes. To further examine whether Ly-6A.2 mediates adhesion (versus inducing another adhesion pathway) reaggregation assays were performed with paraformaldehyde-fixed Tg+ thymocytes. Paraformaldehyde-fixed Tg+ thymocytes reaggregate in culture and this aggregation is also blocked by phosphatidyl-inositol-specific phospholipase C and anti-Ly-6A.2 monoclonal antibodies. These results indicate that the homotypic adhesion of cultured Ly-6A.2 transgenic thymocytes is directly mediated by Ly-6A.2 and, more importantly, strongly suggests that Ly-6A.2 binds a ligand that is expressed on thymocytes. Tg+ thymocytes also bind to nontransgenic thymocytes, B cells, and T cells, indicating that normal cells naturally express the Ly-6A.2 ligand.

Recent studies have implicated proteasomes in the generation of the antigenic peptides that are presented on major histocompatibility complex class I molecules to T lymphocytes. Interferon gamma modifies the subunit composition of proteasomes and causes changes in their peptidase activities that should favor the production of peptides with hydrophobic or basic carboxyl termini (i.e., the types found on major histocompatibility complex class I molecules). It has been proposed that these changes in peptidase activity are due to incorporation into proteasomes of the major histocompatibility complex-encoded subunits LMP2 and -7, which are induced by interferon gamma. Here we show by gene transfection into lymphoblasts or HeLa cells that LMP7 increases the capacity (Vmax) of 20S and 26S proteasomes to cleave peptides after hydrophobic and basic residues without affecting hydrolysis after acidic residues. These changes depended on the amount of LMP7 subunits incorporated into proteasomes. Transfection of LMP2 reduced cleavage of peptides after acidic residues, increased hydrolysis after basic residues, and did not affect the hydrophobic activity. Since the activity of the total proteasome population changed after incorporation of only small amounts of LMP2 or -7, these subunits must cause major alterations in peptidase activity. Thus, their expression can account for the changes in proteasome activity induced by inteferon gamma, and these findings lend further support to the proposed roles of LMPs in altering the nature of the peptides generated for antigen presentation.

Antigens in extracellular fluids can be processed and presented with major histocompatibility complex (MHC) class I molecules by a subset of antigen presenting cells (APCs). Chicken egg ovalbumin (Ova) linked to beads was presented with MHC class I molecules by these cells up to 10(4)-fold more efficiently than soluble Ova. This enhanced presentation was observed with covalently or noncovalently linked Ova and with beads of different compositions. A key parameter in the activity of these conjugates was the size of the beads. The APC that is responsible for this form of presentation is a macrophage. These cells internalize the antigen constructs through phagocytosis, since cytochalasin B inhibited presentation. Processing of the antigen and association with MHC class I molecules appears to occur intracellularly as presentation was observed under conditions where there was no detectable release of peptides into the extracellular fluids. When injected in vivo in C57BL/6 mice, Ova-beads, but not soluble Ova, primed CD4- CD8+ cytotoxic T lymphocytes (CTLs). Similar results were obtained in BALB/c mice immunized with beta-galactosidase-beads. The implications of these findings for development of nonliving vaccines that stimulate CTL immunity are discussed.

The T-cell-activating protein TAP is a murine phosphatidylinositol-anchored glycoprotein whose expression is controlled by the Ly-6 locus. Previous studies have suggested an important role for this protein in physiological T-cell activation. Using oligonucleotide probes, we have now isolated a cDNA clone whose predicted sequence would encode a protein with an NH2-terminal sequence identical to that of the TAP molecule. Further analysis of the predicted protein sequence revealed a cysteine-rich protein with a hydrophobic domain at the COOH terminus and without N-linked glycosylation sites--all features consistent with our previous analysis of the TAP protein. In Southern blot analysis, the Ly-6.2 cDNA clone detects a multigene family and a restriction fragment length polymorphism that maps precisely to the Ly-6 locus. Expression of the cDNA clone in COS cells demonstrates that it codes for TAP and clarifies the relationship between the epitopes recognized by various alpha Ly-6 monoclonal antibodies. Finally, we have studied the expression of Ly-6 mRNA in a variety of cell lineages. Ly-6 transcripts were detected in all organs examined, including spleen, kidney, lung, brain, and heart. This demonstrates that the Ly-6 locus is transcriptionally active in a wide range of organs and suggests that the role of TAP or TAP-like proteins might extend to other tissues.

RMA-S murine cells have a mutation that interferes with the assembly of class I major histocompatibility complex (MHC) heterodimers and are deficient in the expression of class I molecules on the cell surface. The mutant phenotype has been reported to be normalized upon incubation of RMA-S cells at 25 degrees C. We find that much of the increased expression of class I heterodimers is dependent on culturing RMA-S cells in bovine serum or with purified bovine beta 2-microglobulin. Furthermore, epitopes that are associated with class I MHC molecules that have bound xenogeneic beta 2-microglobulin are preferentially formed on RMA-S cells cultured at 25 degrees C. These heterologous class I molecules are thermolabile. Increased expression of class I molecules has also been observed on RMA-S cells incubated at 37 degrees C in the presence of class I-restricted peptides. We find that the increased expression of Db molecules induced by influenza virus nucleoprotein residues 365-380 is similarly dependent on culturing RMA-S cells in bovine serum or with purified bovine beta 2-microglobulin.

The corecognition of antigen and class II major histocompatibility complex (MHC) molecules (Ia molecules) by the T-cell receptor is a cell surface event. Before antigen is recognized, it must be taken up, processed, and displayed on the surface of an Ia-bearing accessory cell (antigen-presenting cell, APC). The exact nature of antigen processing and the subsequent associations of antigen with the APC plasma membrane, Ia molecules, and/or the T-cell receptor are not well defined. To further analyze these events, we have characterized the processing and presentation of the soluble polypeptide antigen bovine insulin. We found that this antigen requires APC-dependent processing, as evidenced by the inability of metabolically inactivated APCs to present native antigen to antigen plus Ia-specific T-T hybridomas. The ability of the same APCs to present antigen after uptake and processing showed that this antigen subsequently becomes stably associated with the APC plasma membrane. To characterize the basis for this association, we analyzed its sensitivity to enzymatic digestion. APCs exposed to antigen, treated with phospholipase A2, and then immediately fixed lost the ability to stimulate bovine insulin plus I-Ad-specific hybridomas. In contrast, the ability of these same APCs to stimulate I-Ad allospecific hybridomas was unaffected. This effect of phospholipase is not mimicked by the broadly active protease Pronase, nor is there evidence for contaminating proteases in the phospholipase preparation. These results suggest that one consequence of antigen processing may be an antigen-lipid association that contributes to the anchoring of antigen to the APC membrane. The implications of this model are discussed.

T-cell-activating protein (TAP) is an allelic 12-kDa membrane protein that participates in T-cell activation. Soluble anti-TAP monoclonal antibodies can trigger antigen-specific, major histocompatibility complex-restricted T-cell hybridomas to produce interleukin 2 and are mitogenic for normal T cells and thymocytes. TAP is expressed on 10% of thymocytes, which are mainly cortisone-resistant and mature. In the periphery, TAP is expressed on 70% of resting T cells but not on resting B cells. In this report, we analyze in detail the nature of TAP expression on peripheral lymphocyte subsets by immunofluorescence techniques. We show that all inducer (L3T4+) T cells are TAP+. In contrast, only 50% of Lyt-2+ T cells express detectable TAP. Functional studies demonstrated that at least part of the heterogeneity of TAP expression is present in the Lyt-2+ cytolytic T-cell (CTL) subset. Unstimulated CTL precursors are TAP- but are induced to express TAP in the effector state. Furthermore, this reflects actual synthesis of TAP, as TAP is detectable on activated Lyt-2+ CTLs passaged in vitro under conditions where passive acquisition can be ruled out. To extend this observation, we have studied the expression of TAP on activated T and B cells. Upon activation, all T and B cells became TAP+. Furthermore, the TAP molecules on B and T cells are indistinguishable by NaDodSO4/polyacrylamide gel electrophoresis. This suggests that TAP expression defines further heterogeneity of lymphocytes, with activation being one parameter influencing its expression.

A synthetic peptide corresponding to residues 365-380 of the influenza nucleoprotein (NP365-380) has been previously shown to associate with class I major histocompatibility complex-encoded molecules and to stimulate cytotoxic T lymphocytes [Townsend, A. R. M., Rothbard, J., Gotch, F. M., Bahadur, G., Wraith, D. & McMichael, A. J. (1986) Cell 44, 959-968]. We find that intact Db class I heterodimers on the cell surface are unreceptive to binding this antigen. However, NP365-380 readily associates with Db molecules on the plasma membrane in the presence of exogenous beta 2-microglobulin. In addition, there is a second pathway through which this peptide associates with class I molecules that requires energy and de novo protein synthesis. These findings have implications for maintaining the immunological identity of cells and for the use of peptides as vaccines for priming cytolytic T-cell immunity.

The association of major histocompatibility complex (MHC) class I molecules on the surface of cells with synthetic antigenic peptides of eight or nine amino acid residues was examined. Peptides were synthesized that correspond to the antigenic sequences from ovalbumin and influenza nucleoprotein believed to be naturally processed and presented by cells with Kb and Db MHC class I molecules, respectively. Consistent with the results of others, these peptides were 10(3)-10(5) times more active in stimulating specific T cells as compared to peptides of longer sequences. When cells are incubated with these peptides at less than 0.01-0.1 microM, the association of the peptides with class I molecules is dependent on (i) the reassociation of free beta 2-microglobulin from the extracellular fluids, (ii) a process that requires cells to be metabolically active, or (iii) stabilization of class I heterodimers by chemical crosslinking. In contrast, when cells are incubated with these peptides at greater than 0.1-1.0 microM, the peptides associate with class I molecules in the absence of exogenous beta 2-microglobulin, energy, or chemical crosslinking. Antigen competition experiments suggest that the class I molecules that bind peptides offered at high concentration become only transiently receptive to binding peptide. The concentration of peptides required for presentation to T cells under these conditions corresponds to those that stabilize Kb molecules on the surface of RMA-S mutant cells in the absence of exogenous beta 2-microglobulin. These results support the concept that the receptivity of class I molecules on cells is determined by the dissociation of beta 2-microglobulin from MHC class I that lacks bound peptides.

To characterize the basis for the cell surface association of processed antigen with the antigen-presenting cell (APC) we analyzed its sensitivity to enzymatic digestion. Antigen-exposed APC that are treated with phospholipase and then immediately fixed lose their ability to stimulate antigen-plus-Ia-specific T-T hybridomas. This effect is seen with highly purified phospholipase A2 and phospholipase C. In addition it is observed with three distinct antigens--ovalbumin, bovine insulin, and poly(LGlu56LLys35LPhe9) [(GluLysPhe)n]. The effect of phospholipases is highly specific. Identically treated APC are equivalent to controls in their ability to stimulate alloreactive hybridomas specific for precisely the same Ia molecule that is corecognized by antigen-plus-Ia-specific hybrids. Furthermore, the antigen-presenting function of enzyme-treated, fixed APC can be reconstituted by the addition of exogenous in vitro processed or "processing independent" antigens. In parallel studies 125I-labeled avidin was shown to specifically bind to APC that were previously exposed and allowed to process biotin-insulin. Biotin-insulin-exposed APC that are pretreated with phospholipase bind significantly less 125I-labeled avidin than do untreated, exposed APC. Identical enzyme treatment does not reduce the binding of avidin to a biotinylated antibody already bound to class II major histocompatibility complex molecules of APC. At least some of the biotin-insulin surface sites are immunologically relevant, because the presentation of processed biotin-insulin by fixed APC is blocked by avidin. This effect is specific. Avidin binding to biotin-insulin-exposed APC does not inhibit allospecific stimulation nor the presentation of unconjugated insulin. These studies demonstrate that phospholipase effectively removes processed cell surface antigen.

We have characterized the TAP molecule, an Ly-6 linked T-cell-activating glycoprotein. The three TAP bands that are precipitated from metabolically labeled cells display a common migration pattern in isoelectric focusing/NaDodSO4/PAGE gels and have common N-terminal sequences. This sequence is rich in cysteine and is homologous to that previously reported for the Ly-6.1E antigen. We, therefore, compared TAP and Ly-6.1E biochemically and found them to be structurally distinct. Given the role of TAP in T-cell activation, we further studied whether the molecule was phosphorylated. We have not found evidence for phosphorylation of the TAP protein. The carbohydrates present on the TAP molecule are resistant to peptide N-glycosidase F in vitro and tunicamycin in vivo. The upper band of the TAP triplet is susceptible to treatment with trifluoromethanesulfonic acid and thus seems to be of the O-linked rather than of the N-linked variety. The biosynthetic processing of TAP was studied in pulse-chase experiments. The middle band of the TAP triplet appears to be the earliest detectable species. Its conversion to the O-linked high molecular weight species can be blocked by monensin.