Inhibition of T cell activation in vivo with mixtures of monoclonal antibodies specific for I-A and I-A/E molecules.

(CBA x B6)F1 (Iak x Iab) T cells were activated to sheep erythrocytes in irradiated F1 mice in the presence of various monoclonal anti-Ia reagents and then tested for their capacity to collaborate with B cells from B10.BR (I-Ak, I-Ek) (kk), B10.A(4R) (kb), and B10 (bb) mice. Anti-I-Ak antibodies blocked the generation of help for B10.A(4R) B cells, but not B10.BR or B10 B cells. An anti-I-Ab antibody blocked help for B10 B cells, but not for B10.BR or B10.A(4R) B cells. An antibody (Y-17) specific for I-Ak/Ek and I-Ab/Ek molecules, but not for I-Ak or I-Ab molecules, failed to impair the generation of help for B10.BR, B10.A (4R), or B10 B cells. In marked contrast to injecting each antibody separately, a mixture of anti-I-Ak and anti-I-Ak,b/Ek (Y-17) antibodies virtually abolished the generation of help for B10.BR B cells. A mixture of anti-I-Ak and anti-I-Ab antibodies effectively blocked help for (4R x B10)F1 B cells, i.e., cells expressing hybrid I-A molecules. These two antibodies only marginally impaired help for (CBA x B6)F1 B cells. To block help for (CBA x B6)F1 B cells required selection in the presence of a cocktail of anti-I-Ak, anti-I-Ab, and anti-I-Ak,b/Ek antibodies. The implications of these findings are discussed.

Herpes simplex virus infection of human fibroblasts and keratinocytes inhibits recognition by cloned CD8+ cytotoxic T lymphocytes.

CD8+ cytotoxic T lymphocytes (CTL) clones with specificity for herpes simplex virus (HSV) were derived from two donors with genital HSV-2 infection. These CTL clones specifically lysed HSV-infected autologous B lymphoblastoid cells, but not HSV-infected fibroblasts. Exogenous peptide loading sensitized both cell types to lysis by an HSV-specific CTL clone of known specificity. HSV infection rendered fibroblasts refractory to peptide sensitization. HSV infection also rendered fibroblasts and keratinocytes insensitive to lysis by allospecific CD8+ CTL clones. Lysis of B lymphoblastoid cells in this system was only slightly reduced by HSV infection. Reduction of fibroblast allospecific lysis was dose and time dependent and was blocked by acyclovir, indicating the involvement of a late HSV gene product. HSV caused a reduction of fibroblast cell surface HLA class I antigen, at least in part due to reduction of synthesis of heavy chain-beta 2 microglobulin heterodimers. These results suggest that HSV-induced blockade of antigen presentation by cutaneous cells to CD8+ CTL may be a mechanism by which HSV limits or evades the immune response of the host.

Hematopoietic subpopulations express cross-reactive, lineage-specific molecules detected by monoclonal antibody.

The molecular specificity of a rat anti-mouse monoclonal antibody for cell surface antigens expressed by T- and B-lymphocyte subsets, erythrocytes and polymorphonuclear neutrophils (PMN) was determined. The antibody reacts with B-lymphocyte-associated molecules which migrated as a sharp 48,000 mol. wt band on SDS-PAGE. The antibody reacts with heterogeneous thymocyte and PMN molecules with a predominant mol. wt of 52,000. The same antibody reacts with heat-stable, amphipathic, organic-solvent-soluble erythrocyte molecules of mol. wt 35,000-40,000 present in Folch upper-phase ganglioside fractions, and evidence is presented that the determinant is protein-defined. Thus, a single monoclonal reagent which recognizes distinct, lineage-specific cell surface proteins on erythroid, lymphoid and myeloid elements may be used to probe not only the characteristic patterns of development of these hematopoietic subsets, but also the biochemical functions of the protein antigens themselves. In the case of B- and T-lymphocytes, such functions may extend to involvement in ligand-induced maturation and repertoire selection.

Fine specificity of a monoclonal anti-testicular cell antibody for glycolipids with terminal N-acetyl-D-glucosamine structure.

The specificity of a mouse anti-testicular cell monoclonal antibody, J1, was investigated. Previous studies suggested that N-acetyl-D-glucosamine (GlcNAc) was a constituent of the determinant recognized by J1. When the antibody was tested against a variety of purified glycolipids containing this saccharide in terminal, penultimate or internal positions, J1 reacted only with species expressing terminal GlcNAc. The influence of oligosaccharide chain length, branch substitution, and haptenic valence on J1 binding was examined using glycolipids prepared by a weak acid hydrolysis and exoglycosidase digestion of bovine I-active ganglioside. Degree of binding was inversely proportional to chain length and was proportional to hapten valence. Failure of J1 to bind partially deglycosylated transferrin implied binding preference for GlcNAc beta 1----3Gal over GlcNAc beta 1----2Man. Immunofluorescence analysis of J1 binding to human neutrophils failed to detect lactotriosylceramide on their surface, although this glycolipid has previously been isolated from these cells, suggesting that this structure exists in a cryptic or intracellular form. Binding results were consistent with J1 having low affinity for GlcNAc or GlcNAc beta 1----3Gal on a variety of lacto-series glycolipids.

Lymphotoxin, tumor necrosis factor, and gamma interferon are cytostatic for normal human keratinocytes.

The effects of crude lymphokine-enriched supernatants, purified recombinant lymphotoxin (LT), tumor necrosis factor-alpha (TNF), and gamma interferon (gamma IF) on proliferating human keratinocytes were assessed using two in vitro culture systems. Activated splenocyte supernatants inhibited keratinocyte colony growth on fibroblast feeder layers and arrested basal keratinocyte DNA synthesis within 24 h. Purified recombinant LT, TNF, and gamma IF inhibited cell proliferation in serum-free medium without noticeably affecting viability. Cytostasis was dose-dependent (up to 90% with LT or TNF and 99% with gamma IF) and was maximal within 24-36 h. Specific antibodies neutralized TNF- and gamma IF-mediated cytostasis. Combined treatment with LT (or TNF) and gamma IF increased the degree of cytostasis, particularly at low lymphokine concentrations. Maximum inhibition of DNA synthesis and the duration of exposure required for this inhibition were comparable for LT and TNF and differed for gamma IF. Each of these lymphokines induced cell enlargement, flattening, and vesiculation, with gamma IF apparently more potent in this respect than LT or TNF. Fusiform keratinocytes with diffusely distributed cytokeratin were observed after prolonged treatment with gamma IF alone or gamma IF plus either LT or TNF. Flow cytometric studies of lymphokine-treated keratinocytes indicated that LT, TNF, and gamma IF could enhance beta-2 microglobulin expression 1.5-fold to threefold, whereas only gamma IF induced class II antigens. Staining for class II and beta-2 microglobulin was reduced on cells treated with high concentrations of gamma IF compared with either optimally treated or untreated cells. The potential relevance of these findings to cutaneous immune defense and disease is discussed.

Lysis of keratinocytes by IL-2-activated peripheral blood lymphocytes.

Interleukin 2 (IL-2)-activated peripheral blood mononuclear cells (PBMC) have been reported to lyse tumor cells while essentially sparing normal cells in vitro. This report concerns IL-2-induced anti-keratinocyte (anti-KC) cytotoxic effectors that lyse normal human keratinocytes (KC) in vitro. Effectors were generated by culturing PBMC for 1-8 d in various concentrations of recombinant IL-2 and then assayed against 51Cr-labeled targets. Effectors stimulated with 10(3) U/ml of IL-2 for 8 d readily lysed adherent or trypsinized autologous or allogeneic KC cultured in serum-free medium. Induction of anti-KC effectors was IL-2 dose-dependent, with as little as 12-25 U/ml of IL-2 inducing increased anti-KC activity after 24 h of treatment. Although anti-KC activity was increased after overnight culture in IL-2, maximal effector potency in terms of lytic units (LU) per 10(6) effector cells required 4 d of IL-2 treatment. Maximal effector yield in terms of LU per input PBMC occurred after 8 d of IL-2 treatment. Antibody plus complement depletion studies showed that the anti-KC effectors predominantly have a CD16 --/CD3 --/CD2+ phenotype. A natural killer (NK)-like specificity of the effectors was suggested by two findings: unlabeled K562 cells totally inhibited lysis of 51Cr-KC in cold target competition assays, and interferon gamma (IFN-g) treatment (2.5 U/ml-500 U/ml of recombinant IFN-g for 48-72 h) down-regulated KC susceptibility to lysis by these effectors. Thus, IL-2 treatment of PBMC induces non-T cell, natural killer-like effectors that can lyse both autologous and allogeneic KC. Furthermore, KC resemble other cell types that become resistant to non-MHC-restricted lysis after treatment with IFN-g. Finally, the contrasting effects of IFN-g treatment on KC lysis by these effectors, as opposed to lysis by specific T cells, suggests that IFN-g could promote a shift from non-MHC-restricted to MHC-restricted KC lysis during epidermal immune responses in vivo.

Recognition of keratinocytes by cytotoxic T cells specific for conventional HLA class-I alloantigen.

This study analyzed whether human keratinocytes (KC) express conventional HLA class-I molecules as detected by class-I-specific cytotoxic T lymphocytes (CTL), and whether exposure of KC to interferon-gamma (IFN-g) is required for CTL recognition. Basal KC grown in serum-free medium and exposed to recombinant IFN-g for 24-96 h were used as targets in 51Cr-release assays. Target-cell susceptibilities to lysis were compared by analyzing the lytic unit (LU) activity of a given CTL population against IFN-g-treated and untreated KC. CTL effectors were cloned from alloantigen-primed cultures by limiting dilution in the presence of antigenic B lymphoblastoid cells (BCLL) and IL-2. These T-cell clones lysed appropriate BCLL and PHA blasts but not third-party BCLL or K562. Lysis of antigenic BCLL was specifically blocked by antibodies against CD3 or class-I antigens. Specificity of the clones for conventional class-I antigen was demonstrated by cytotoxicity tests employing a panel of HLA-typed BCLL. The clones specifically lysed KC syngeneic with the original effector immunogen, and lysis was also blocked by anti-class-I antibodies. The effect of IFN-g treatment was to increase KC susceptibility to lysis by these clones. From 3-25 times more LU were measured against IFN-g-treated KC than against nontreated KC, and the degree of enhancement was similar for KC treated with concentrations of IFN-g ranging from 2.5-200 U/ml. This effect of IFN-g treatment on KC lysis by CTL, which was detected after only 24 h at all doses tested, emphasizes the potential role of IFN-g in enhancing CTL-mediated antiviral epidermal immunity and in exacerbating epidermal disease mediated by specific lytic T cells. In addition, the finding that normal human KC can be recognized by MHC class-I-specific CTL demonstrates that KC do express conventional class-I-antigens and that KC lysis by CTL can occur independently of exogenous cytokines.

Human epidermal keratinocyte expression of sialyl-Lewis X.

Cell-surface oligosaccharides can function as ligands for intercellular adhesion receptors, matrix proteins, and growth factors. We report that human neonatal and adult epidermal keratinocytes (KC) express sialyl Lewis X [s-Le(x); SA alpha 2-3Gal beta 1-4(Fuc alpha 1-3)GlcNAc beta 1-3R], a ligand for endothelial and platelet selectins. Freshly isolated or cultured KC bind FH6 monoclonal antibody (MoAb), which is specific for s-Le(x)-containing oligosaccharides. The relevant epitope is bona fide s-Le(x), because sialidase treatment of KC suspensions abrogates FH6 binding while generating de novo KC reactivity with anti-Le(x). KC stained in ice-cold suspension display a knobby membrane distribution of s-Le(x) detectable by immunofluorescence microscopy. As others have reported, FH6 appeared not to bind KC in perpendicular skin sections. However, basal KC in intact epidermal sheets exhibited obvious "honeycomb" reactivity with FH6 when stained and viewed en face, suggesting that s-Le(x) in intact epidermis may occur in bands that parallel the major tissue axis. FH6 specifically immunoprecipitated proteins of Mr 34 kd, 44 kd, and 56 kd from [35S]-labeled KC, and anti-Le(x) precipitated similar proteins from sialidase-treated KC. The enzymatic basis for KC s-Le(x) expression was studied by analyzing acceptor specificities and other properties of KC fucosyltransferases. Results indicate that KC express both Lewis- and myeloid-type alpha 1-3fucosyltransferases. KC s-Le(x) could be an important element of the epithelial milieu, because both epithelial cells and immune cells that home to epithelia express s-Le(x) and related structures, and because KC s-Le(x) is well positioned for selectin-mediated platelet binding after trans-cutaneous wounding. The apparent distributions of s-Le(x) in epidermis and on isolated KC are compatible with a functional role for s-Le(x) in these intercellular interactions.

The relationship of serum IL-6 levels to acute graft-versus-host disease and hepatorenal disease after human bone marrow transplantation.

The potential involvement of cytokines in acute graft-versus-host disease led us to analyze interleukin-6 in serial serum sets from 22 allogeneic marrow recipients who developed either grade 3 or 4 GVHD (n = 10), grade 2 GVHD (n = 6), or grade 1 or no diagnosed GVHD (n = 6). A total of 279 serial serum samples taken three times weekly before day 35 were analyzed. Maximum IL-6 levels were greater than 40 U/ml (range, 40-1536 U/ml), 11-40 U/ml, and less than or equal to 10 U/ml for six, eleven, and five patients, respectively. Serum IL-6 peaks were temporally related to onset of GVHD, onset of a syndrome of hepatorenal dysfunction (HRD), or bilateral lung infiltration. Eight of ten patients who developed grade 3 or 4 GVHD overall had IL-6 maxima of greater than 10 U/ml an average of 1.5 +/- 1.8 days before the clinical onset. Fifteen of 17 patients with peak IL-6 levels greater than 10 U/ml developed symptoms of hepatic and renal dysfunction within three days of the peak, while none of five patients with less than or equal to 10 U/ml of Il-6 developed HRD. Regression analysis demonstrated a linkage between the log magnitudes of the serum IL-6 peaks and onset of either GVHD or HRD within three days (P = 0.001). Furthermore, IL-6 peaks tended to precede GVHD onset for the 10 patients whose GVHD onset and IL-6 peak were within three days of each other (P = 0.02). These results, confirmed by both specific bioassay and by IL-6 ELISA, support the idea that acute GVHD in humans involves a cytokine cascade that includes production of IL-6 in addition to the previously reported involvement of tumor necrosis factor alpha and interferon-gamma.

1H-n.m.r. analysis of type-2 chain lacto-gangliosides. Confirmation of structure of a novel cancer-associated fucoganglioside, alpha-NeuAc-(2----6)- beta-D-Galp-(1----4)-beta-D-GlcpNAc-(1----3)-beta-D-Galp-(1----4)-[alp ha-L- Fucp-(1----3)]-beta-D-GlcpNAc-(1----3)-beta-D-Galp-(1----4)-beta -D-Glc p- (1----1)-Cer (VI6NeuAcIII3FucnLc6Cer).

Neolacto-glycosphingolipids, substituted with alpha-NeuAc-(2----3)- and -(2----6)-linked D-Galp residues were analyzed by one- and two-dimensional 1H-n.m.r. spectroscopy at 500 MHz in 49:1 (v/v) di(2H3)methyl sulfoxide-deuterium oxide solution. For the simplest structures analyzed, nLc4Cer, IV3NeuAcnLc4Cer, and IV6NeuAcnLc4Cer, sialosylation-induced changes in shifts of terminal and subterminal core residues were interpretable in terms of existing conformational models. Chemical shifts for H-3e and H-3a of NeuAc characteristic for the type of linkage, were also determined. In addition, regularly reproducible shifts were seen for H-1 and other resonances of terminal and subterminal core residues of all structures tested. Chemical-shift correlations proved to be useful in elucidating the structure of a unique ganglioside bearing an internal beta-D-Galp-(1----4)-[alpha-L-Fucp-(1----3)]-beta-D-GlcpNAc-(1---- 3) residue ("X-trisaccharide") with an alpha-NeuAc-(2----6)-substituted terminal group.

CDw17: a neutrophil glycolipid antigen regulated by activation.

The plasma membrane and intracellular granules of human polymorphonuclear neutrophils (PMN) contain large amounts of the glycolipid, lactosylceramide (LacCer; Gal beta 1----4Glc beta 1----1Cer). Despite its abundance, novel subcellular distribution, and lineage-restricted expression, nothing of PMN LacCer function is known. We examined the relationship between LacCer and PMN activation by assessing binding of anti-LacCer mAb (T5A7; anti-CDw17) to PMN during and after cell stimulation. CDw17 expression markedly decreased after treatment with PMA, dioctanoylglycerol, calcium ionophore, FMLP (with or without cytochalasin B or added Ca2+), TNF-alpha, or lymphotoxin. Depending on the stimulus, CDw17 declined to levels ranging from 70% (TNF, lymphotoxin) to less than 5% (phorbol ester, dioctanoylglycerol) of levels detected on untreated PMN. Loss of CDw17 from PMA-treated PMN followed dose- and temperature-dependent kinetics, with loss being detected after PMA treatment for 1 min. Membrane internalization explained PMA-induced loss of CDw17, as cell-associated 125I-anti-CDw17 became inaccessible to fluorescent anti-Ig after PMA treatment. CDw17 on PMN cytoplasts or retinoic acid-induced HL-60 cells was only slightly affected by stimulation, suggesting that down-regulation of the epitope is associated with granule exocytosis rather than superoxide production. Results with PMN from a patient with chronic granulomatous disease confirmed that normal superoxide production is not required for CDw17 loss induced by PMA or FMLP treatment. The data collectively demonstrate that reduced levels of cell-surface CDw17 are associated with granule exocytosis after PMN activation.

A two-signal mechanism for the induction of cytotoxic T lymphocytes.

Depletion of adherent cells from stimulator and responder lymphocytes by a single filtration through nylon wool columns led to complete abrogation of the cytotoxic response to the stimulating alloantigen. Cytotoxic responses were restored by adding anti-Thy-1 + complement-greated normal peritoneal exudate cells (PEC) syngeneic or allogeneic to the responding population. Alternatively, the response could be reconstituted with costimulator, a lymphokine obtained by stimulation of spleen cells with concanavalin A. Costimulator was not itself cytotoxic and induced few or no cytotoxic T lymphocytes (CL) in the absence of stimulator cells. Costimulator was also more efficient than allogeneic PEC, which in turn were more efficient than syngeneic PEC, in reconstituting the cytotoxic response. The number of CL produced to the activating alloantigen was shown to increase with increasing concentration of costimulator. More interestingly, in the presence of a relatively high concentration of costimulator, CL were also activated to target cells that differ in H-2 haplotype from the stimulating alloantigen. Lysis of the third-party target cells could not be inhibited by cold targets syngeneic to the activating alloantigen. A clonal assay for cytotoxic precursors was used to confirm that CL for the activating alloantigen and CL for the third-party H-2 antigens were derived from different progenitors. Only about 37% of the cytotoxic clones produced were specific for the activating alloantigen. These observations are explained in terms of a two-signal model of CL activation.

Lysis of human keratinocytes by allogeneic HLA class I-specific cytotoxic T cells. Keratinocyte ICAM-1 (CD54) and T cell LFA-1 (CD11a/CD18) mediate enhanced lysis of IFN-gamma-treated keratinocytes.

Exposure of human KC to IFN-gamma increases their susceptibility to lysis by CTL. The mechanism of this enhanced lysis was investigated by analyzing interactions of IFN-gamma-treated and nontreated cultured KC with allogeneic class I-specific CTL clones. rIFN-gamma treatment augmented KC lysis in a time- and dose-dependent manner. Increased lysis of IFN-KC was detected after only 2 h of IFN-gamma treatment and was maximal by 12 h. Enhanced lysis of IFN-KC was Ag-specific, inasmuch as nonantigenic IFN-KC were not lysed either directly or as bystanders during the lysis of antigenic KC. Parallel immunofluorescence and cytotoxicity assays of KC treated with IFN-gamma for various intervals revealed a direct correlation between the degree of increased KC lysis and levels of cell surface ICAM-1 (CD54), but not of specific alloantigen or beta 2-microglobulin. Lysis of nontreated KC was blocked by mAb against class I or CD3, but not by mAb against ICAM-1 or LFA-1. In contrast, lysis of IFN-KC was partially inhibited by anti-ICAM-1 or anti-LFA-1 mAb, but resisted inhibition by anti-class I mAb except in the presence of anti-ICAM-1. These results indicate that both ICAM-1/LFA-1 and Ag/CD3-TcR interactions are important for Ag-specific lysis of IFN-KC, whereas lysis of nontreated KC depends on Ag/CD3-TcR but not ICAM-1/LFA-1 interactions. Equivalent inhibition of IFN-KC lysis by mAb against ICAM-1 or LFA-1 suggests that ICAM-1 is the only LFA-1 ligand involved in enhanced IFN-KC lysis. Furthermore, enhanced CTL lysis of KC after short-term IFN-gamma treatment can be explained solely on the basis of ICAM-1 induction, because all of the increase in specific lysis associated with IFN-gamma treatment could be blocked by mAb that block ICAM-1/LFA-1 interactions.

Evidence for phosphatidylinositol-linked forms of human and avian fibronectin receptors.

Phosphatidylinositol (PI)-linked forms of surface molecules have been hypothesized to mediate the initial stages of cell adhesion or signal transduction. We report evidence for the occurrence of a functional PI-linked subset of cell surface fibronectin receptors (FNR). Treatment of human MG63 osteosarcoma cells or primary chicken embryo fibroblasts (CEF) with PI-specific phospholipase C (PI-PLC) reduced cell surface FNR expression by 30% as detected by immunofluorescence. PI-PLC treatment of cell membranes purified from [35S]methionine-labeled CEF or MG63 cells led to a similar loss of membrane-associated immunoprecipitable FNR from the pelleted membranes, while such treatment led to the appearance of FNR in the supernatant of treated MG63 membranes. Biosynthetic labeling of CEF FNR with [3H]palmitate and [3H]ethanolamine demonstrated the acylation and putative PI linkage of avian FNR subunits. PI-PLC treatment of CEF and MG63 cells also reduced fibronectin-specific adhesion in a short-term in vitro assay, suggesting that the avian and human FNR occur in PI-linked isoforms which appear to contribute to cell adhesion to fibronectin.

Biochemical characterization of Ia antigens encoded by the RT1.B and RT1.D loci in the rat MHC.

Ia antigens in rats are genetically associated with typical MHC-linked immune response (Ir) genes. One rat Ir gene (IR-GLT) has recently been mapped to the RT1.D locus in a rat MHC recombinant, WRC. We have studied the expression of Ia antigenic determinants in WRC and its parental strains BN and WRA using a panel of monoclonal antibodies. Our results suggest that the inheritance of IR-GLT corresponds with the inheritance of I-E-like, but not I-A-like, antigens in the WRC rat. These observations were confirmed when WRC I-E-like antigens were analyzed by two-dimensional gel electrophoresis. We propose that RT1.D shares both functional and antigenic homologies with the mouse I-E subregion.

Glycolipid antigens of human polymorphonuclear neutrophils and the inducible HL-60 myeloid leukemia line.

Glycolipid and cell surface carbohydrate antigens of human polymorphonuclear neutrophils (PMN) and of HL-60 myeloid leukemia cells were analyzed with a panel of defined, monoclonal anti-carbohydrate antibodies. Antigenicities of intact PMN, HL-60, and retinoic acid-induced HL-60 (r.a.-HL-60) were studied by flow cytofluorometry. These three cell populations displayed quantitative differences, some of which were induction dependent, in their expression of lactosyl, N-acetyllactosaminyl, Y-hapten (Fuc alpha 1----2Gal beta 1----4(Fuc alpha 1----3)GlcNAc beta 1----R), and sialosyl-X-hapten (SA alpha 2----3Gal beta 1----4(Fuc alpha 1----3)GlcNAc beta 1----R) specificities. Structures reactive with antibodies specific for long-chain mono-, and di- or tri- alpha 1----3 fucosylated lacto-series glycolipids were also detected. Glycosphingolipids purified from organic extracts of these cells were analyzed to seek information concerning the chemical basis for these surface antigenic differences, to assess the structural and antigenic diversity of PMN and HL-60 glycolipids, and to quantitate chemically and antigenically prominent glycolipids. Binding of monoclonal antibodies to thin-layer chromatograms demonstrated that each of the specificities on intact cells was carried by one or more distinct glycolipids. The abundance of immunoreactive glycolipids in the extracts paralleled the relative staining intensities of the intact cell populations. Several "cryptic" glycolipid antigens, including alpha 2----6 sialosylated structures enriched five- to 10-fold in PMN extracts, were not detected on intact cells. Lactosylceramide accounted for two-thirds of the approximately 1.5 X 10(9) glycolipid molecules contained in each PMN. The remaining glycolipid antigens appeared to include structurally diverse fucolipids, fucogangliosides, and neutral and sialosylated glycolipids with Gal beta 1----4GlcNAc beta 1----R terminal core structure. The abundance, diversity, and induction-dependent expression of these structures suggest that they may participate in PMN maturation and function.

Phorbol ester induces increased expression, altered glycosylation, and reduced adhesion of K562 erythroleukemia cell fibronectin receptors.

The human multipotential hematopoietic cell line K562 expresses fibronectin receptor (FNR) subunits of 160 kDa (alpha chain) and 120 kDa (beta chain). Treatment with 12-O-tetradecanoylphorbol 13-acetate (TPA) led to reduced binding of K562 to immobilized fibronectin (FN), although treated cells expressed 10-fold more cell surface FNR than untreated cells. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis confirmed this and showed altered electrophoretic mobilities of FNR subunits from TPA-treated cells. TPA treatment affected N-linked glycosylation, as tunicamycin treatment of K562 cells abolished differences in FNR mobility. Sialidase treatment of FNR immunoprecipitates minimized and sialidase treatment of intact cells eliminated these mobility differences between subunits from control and TPA-treated cells. Reduced sialylation of FNR from TPA-treated cells was further demonstrated by chromatography with bead-coupled lectins and by the greater negative charge of untreated K562 FNR subunits in two-dimensional isoelectric focusing-polyacrylamide gel electrophoresis. A relationship between reduced FNR sialylation and reduced FN binding was suggested by adhesion assays of sialidase-treated K562 which showed that desialylation of cell surface FNR was associated with decreased cell adhesion. Thus, TPA treatment reduces the function, increases the expression, and alters the structure of K562 FNR, and these changes appear to involve FNR sialylation.

Expression and function of B7 on human epidermal Langerhans cells.

The B7 molecule is expressed by APC that can costimulate T cells by binding the T cell surface receptors CD28 and CTLA-4. The human epidermal Langerhans cell (LC) is one of the most potent APC, yet B7 expression by this cell type has not previously been assessed. We used a CTLA4-Ig fusion protein that binds B7 with high avidity to probe cell surface expression of B7 by cultured and noncultured LC. LC cultured for 1 or more days were specifically stained with biotinylated CTLA4-Ig and fluorescent streptavidin. In contrast, binding of CTLA4-Ig to freshly isolated LC was not detected. The cell surface distributions of B7 and of HLA-DR on cultured LC differed, as CTLA4-Ig binding was localized to discrete foci, whereas anti-DR mAb uniformly stained the LC plasma membrane. Analyses of epidermal cell (EC) mRNA indicated that the B7 gene is expressed by these cells. Thus, B7 gene probes specifically hybridized to polymerase chain reaction-amplified B7 mRNA isolated from cultured and noncultured EC. As LC are the only normal epidermal cell type that induces proliferation of allogeneic T cells, the role of B7 in this LC function was studied by coculturing highly purified resting CD4+ T cells and allogeneic EC in the presence of CTLA4-Ig, anti-CD54 (RR/1, anti-intercellular adhesion molecule-1) mAb, or both. CTLA4-Ig and RR/1 each inhibited CD4+ T cell responses to freshly isolated allogeneic EC, and cooperative inhibition of more than 90% was observed in cultures treated with both CTLA4-Ig and RR/1 at 5 micrograms/ml. CTLA4-Ig inhibited stimulation by either fresh EC or cultured EC, suggesting that the increased potency of cultured LC vs noncultured LC may reflect the time needed for noncultured LC to express cell surface B7 in vitro. These studies indicate that B7 is expressed on the cell surface of cultured LC, and that LC B7 costimulates the proliferation of resting allogeneic CD4+ T cells.