Alpha 2-3Gal Beta Research Articles

Evidence for glycosphingolipid modification of the type 1 IFN receptor.

P1-determinant glycolipids that include two membrane glycosphingolipids, globotriaosylceramide (Gal alpha 1-4Gal beta 1-4Glc-ceramide, Gb3) and galabiosylceramide (Gal alpha 1-4Gal-ceramide, Gb2) are receptors for an Escherichia coli-derived subunit toxin, verotoxin (VT-1). Studies with Daudi cells and glycosphingolipid-deficient Daudi mutants and U937 cells identified that the presence of Gb2/Gb3 correlates with IFN-alpha sensitivity. Comparison of amino acid sequences between VT-1 and the extracellular region of the 63-kDa IFN-alpha beta receptor (IFNAR) peptide reveals regions of identity, specifically in those domains in the VT-1 B subunit, that have been implicated as Gb2/Gb3 binding sites. In direct ligand binding studies, we show that membrane Gb2/Gb3 content affects the binding capacity of cells for IFN-alpha, although IFNAR cell surface expression is unaffected. Binding of IFN-alpha to the receptor leads to kinase-associated phosphorylation of the latent transcription factor, ISGF3, which activates transcription by binding to IFN-stimulated regulatory elements in IFN-sensitive genes. Electrophoretic mobility band shift assays indicated that U-937 and Daudi mutant cells, deficient in Gb2 and Gb3, exhibited reduced nuclear factor binding to the human 2-5A synthetase IFN-stimulated regulatory element when compared with wild-type Daudi cells, after exposure to IFN-alpha. Moreover, when Daudi cells were treated with a ceramide analogue, 1-phenyl-2-(decanoylamino)-3-morpholino-1-propanol, Gb2 and Gb3 synthesis was inhibited and a concomitant reduction in IFN-induced ISGF3 activation was noted. IFNAR cell surface expression was unaffected by 1-phenyl-2-(decanoylamino)-3-morpholino-1-propanol treatment. By using a fusion protein of the extracellular domain of IFNAR linked at the carboxyl terminus to the Fc portion of IgG1, we demonstrate that IFNAR is able to bind preferentially to Gb2. These results suggest that an association of IFNAR with membrane Gal alpha 1-4Gal containing glycolipids facilitates receptor-mediated signaling.

Epithelial glucosphingolipid expression as a determinant of bacterial adherence and cytokine production.

D-Threo-1-phenyl-2-decanoylamino-3-morpholino-1-propanol (PDMP) is a structural analog of ceramide that inhibits glucosylation of this molecule and thus of glucosphingolipid (GSL) expression by living cells. In this study, we used PDMP to slow the synthesis of the globoseries of GSLs (globo-GSLs) (derived from the precursor Gal alpha 1-4Gal beta 1-4Glc-ceramide) by cultured human kidney and large intestinal epithelial cells. The aim was to deplete the cells of receptors for P-fimbriated Escherichia coli and to examine the effects on the bacterially induced cytokine response. The mammalian cells (A498, HT-29, and Caco2) were cultured in the presence of PDMP in order to deplete them of GSLs. The cells were then subjected to GSL analysis or used to test bacterial adherence and cytokine production. The globo-GSLs were identified by thin-layer chromatography. Bacterial adherence was quantitated by microscopy, and interleukin-6 secretion was quantitated by the B9 bioassay. The interaction of bacteria with the globo-GSLs was studied by using E. coli strains and recombinant clones expressing P fimbriae. E. coli strains expressing type 1 fimbriae binding to mannose-containing glycoproteins were used as controls. PDMP treatment was found to reduce the content of the globo-GSLs in mammalian cells and the adherence of P-fimbriated E. coli to these cells. In contrast, PDMP treatment had no effect on the adherence of type 1-fimbriated E. coli or their activation of cytokine production by A498 cells. P-fimbriated E. coli elicited an interleukin-6 response in the A498 cells; this response was reduced after treatment with PDMP. The results emphasize the role of GSLs as receptors for P-fimbriated E. coli and for the cytokine response elicited by attaching bacteria.

Isolation and Characterization of a Rat Testis Glycosphingolipid Based on Gangliotetraosylceramide and Having a Blood Group B Determinant Extended with β3-Linked N-Acetylgalactosamine1

Monoclonal antibody 14.2, directed against the GalNAc beta 3Gal alpha 3Gal beta epitope, was used to monitor the isolation of a novel glycosphingolipid from rat testes. Structural analysis was performed by mass spectrometry, proton NMR spectroscopy, gas chromatography-mass spectrometry after degradation, and exoglycosidase digestion. It was concluded that the structure was [formula: see text] This is the first report on an extended blood group B determinant.

Specific stimulation of Graves' disease thyrocytes by the natural anti-Gal antibody from normal and autologous serum.

Anti-Gal is a natural Ab abundantly produced in humans. It interacts specifically with the carbohydrate epitope Gal alpha 1-3Gal beta 1-4GlcNAc-R (termed the alpha-galactosyl epitope). This epitope is expressed in large amounts on thyrocytes of nonprimate mammals, but not of humans. We have previously found that binding of anti-Gal to alpha-galactosyl epitopes on porcine thyrocytes results in stimulatory effects similar to those exerted by thyroid-stimulating hormone (thyrotropin). In the present study, we tested the hypothesis that anti-Gal may contribute to Graves' disease (GD) pathogenesis by stimulation of the thyrocytes of patients with this autoimmune disorder. Anti-Gal binding and stimulatory effects were assessed in primary thyrocyte cultures. Anti-Gal specifically bound to GD thyrocytes and induced an increase in cAMP synthesis, 125I uptake, and DNA synthesis in these cells. Furthermore, the stimulatory effects of autologous sera on GD thyrocytes were greatly reduced after specific depletion of anti-Gal from these sera. No binding and no stimulatory effects of anti-Gal were observed, however, with normal human thyrocytes and with thyrocytes from thyrotoxic patients who lack thyroid-stimulating Igs or thyrotropin binding inhibiting Igs. These in vitro stimulatory effects of anti-Gal on GD thyrocytes suggest that this natural Ab may contribute to the in vivo continuous stimulation of thyrocytes in GD patients. The possibility that anti-Gal may stimulate GD thyrocytes via interaction with aberrantly expressed alpha-galactosyl epitopes on the thyroid-stimulating hormone receptor is discussed.

Characterization of porcine kidney neutral glycosphingolipids: identification of a carbohydrate antigen recognized by human natural antibodies.

We have investigated the glycosphingolipids of pig kidney with a special interest in identifying compounds which may be involved in the rejection of tissue in xenotransplantation. Nine neutral glycosphingolipids have been characterized in porcine kidney and structurally characterized by a combination of techniques including 1H-NMR, permethylation analysis and thin-layer chromatography (TLC) immunostaining with carbohydrate sequence-specific monoclonal antibodies. The major components are members of the globo family and represent the human pk and P antigens. Three other compounds were found to contain a neolacto core structure; the major neolacto compound carries a nonreducing terminal epitope (Gal alpha 1-3Gal beta 1-4GlcNAc) recognized by the naturally-occurring human antibody, anti-Gal, and a second neolacto compound carries the blood group A trisaccharide (GalNAc alpha 1-3(Fuc alpha 1-2)Gal). These results are discussed with respect to tissue transplantation.

Bi-antennary oligo-(N-acetyllactosamino)glycans of I-type are galactosylated preferentially at the GlcNAc?1-6Gal linked arms by ?1,3-galactosyltransferase of bovine thymus

alpha 1,3-Galactosylation of radiolabelled bi-antennary acceptors Gal beta 1-4GlcNAc beta 1-3(Gal beta 1-4GlcNAc beta 1-6)Gal-R (R = 1-OH, beta 1-4GlcNAc or beta 1-4Glc) with bovine thymus alpha 1,3-galactosyltransferase was studied. At all stages of the reactions the three acceptors reacted faster at the 1-->6 linked arm than at the 1-->3 linked branch. Hence, in addition to the doubly alpha 1,3-galactosylated products, practically pure Gal beta 1-4GlcNAc beta 1-3(Gal alpha 1-3Gal beta 1-4GlcNAc beta 1-6)Gal-R could be obtained from the three acceptors in reactions that had proceeded to near completion. The isomeric mono-alpha 1,3-galactosylated products were identified by using exoglycosidases to remove the branches unprotected by alpha 1,3-galactoses and by subsequently identifying the resulting linear glycans chromatographically.

Differential host-dependent expression of alpha-galactosyl epitopes on viral glycoproteins: a study of eastern equine encephalitis virus as a model.

The carbohydrate epitope Gal alpha 1-3Gal beta 1-4GlcNAc-R (alpha-galactosyl) is abundantly expressed on cells of non-primate mammals, prosimians and New World monkeys, where it is synthesized by the enzyme alpha 1,3-galactosyltransferase (alpha 1,3GT). Old World monkeys, apes and humans lack alpha 1,3GT and hence do not synthesize alpha-galactosyl epitopes. Instead, these species produce a natural antibody, anti-Gal, which interacts specifically with alpha-galactosyl epitopes and which constitutes up to 1% of circulating immunoglobulins in humans. We have used eastern equine encephalitis (EEE) virus as a model to examine the differential expression of alpha-galactosyl epitopes on the glycoproteins of virus propagated in cells that either produce or lack alpha 1,3GT. As predicted, virus propagated in Vero cells (derived from the African green monkey, an Old World monkey) did not express alpha-galactosyl epitopes. In contrast, virus propagated in mouse 3T3 cells (EEE3T3) expressed approximately 80 alpha-galactosyl epitopes per virion on both the E1 and the E2 envelope glycoproteins. Thus, expression of the alpha-galactosyl epitope on virions paralleled that on host cells. The binding of anti-Gal antibody to these epitopes on EEE3T3 virions partially neutralized virus infectivity, raising the possibility that anti-Gal production in hosts may influence the initial infectious stage of viruses expressing alpha-galactosyl epitopes.

Glycosphingolipid receptor function is modified by fatty acid content. Verotoxin 1 and verotoxin 2c preferentially recognize different globotriaosyl ceramide fatty acid homologues

Verotoxins (VT) are a family of Escherichia coli-derived toxins which have been associated with hemolytic uremic syndrome, the leading cause of acute pediatric renal failure, and hemorrhagic colitis. Verotoxins (VT1 and VT2c) both show terminal gal alpha 1-4gal-dependent binding to globotriaosylceramide (Gal alpha 1-4Gal beta 1-4Glc-Cer; Gb3), yet VT2c shows a thousandfold lower specific cytotoxic activity in vitro. Our previous studies have shown this discrepancy is a function of the receptor binding B subunit and that VT1/Gb3 binding in a lipid matrix is affected by heterogeneity in the ceramide fatty acid chain length. The influence of the fatty acid composition of Gb3 on the binding of VT1 and VT2c has now been compared using 14 homogeneous semisynthetic Gb3 molecular species of differing fatty acid chain length and degree of saturation from C12 to C24. The binding of verotoxin was quantitated by Scatchard analysis using a solid-phase binding assay in the presence of auxiliary lipids, which may in some respects approximate to receptor function within the plasma membrane of sensitive cells. Differential binding was observed for several of these species in the lipid matrix, indicating that the fatty acid moiety of Gb3 is important for VT binding under such conditions. The short chain fatty acid containing Gb3 (C12 and C14) showed minimal binding. Middle and long chain fatty acid Gb3 homologues (C16, C18, C20, C22, and C24) were effectively recognized by VTs. The presence of an unsaturated fatty acid in Gb3 significantly increased VT binding in all cases. C20:0 and C22:1 containing Gb3 had the greatest capacity to bind VT1. In contrast, C18:0 and C18:1 homologues showed the greatest capacity for VT2c binding (higher than VT1). These results were, in general, reflected in cell cytotoxicity in that receptor-deficient cells reconstituted with C22:1Gb3 were maximally sensitive to VT1 in vitro whereas cells reconstituted with C18:1Gb3 were maximally sensitive to VT2c. VT2c was an ineffective inhibitor of 125I-VT1 binding to C22:1 Gb3 but in contrast, more effective than VT1 to compete binding to C18:1 Gb3. Similarly, VT1 was less effective than VT2c to compete binding of 125I-VT2c to C18:1 but more effective than VT2c to compete for C22:1 Gb3 binding. These results suggest that VT1 and VT2c bind selectively to different but overlapping carbohydrate epitopes on the Gb3 molecule which are differentially available in these Gb3 fatty acid homologues in a lipid environment.(ABSTRACT TRUNCATED AT 400 WORDS)

Molecular species analysis of glycosphingolipids from small intestine of Japanese quail, Coturnix coturnix Japonica by HPLC/FAB/MS.

Neutral glycosphingolipids were isolated from quail small intestine and their structures were analysed. They contained: Gal beta 1-4GlcCer(LacCer), Gal alpha 1-4GalCer(Ga2Cer), Gal alpha 1-4Gal beta 1-4GlcCer(Gb3Cer), GlcNAc beta 1-3Gal beta 1-4GlcCer(Lc3Cer), GalNAc beta 1-4Gal beta 1-4GlcCer(Gg3Cer), GalNAc beta 1-4[GalNAc beta 1-3] Gal beta 1-4GlcCer(LcGg4Cer), and GalNAc alpha 1-3GalNAc beta 1-3Gal alpha 1-4Gal beta 1-4GlcCer (Forssman glycolipid) as well as glucosylceramide, galactosylceramide (Nishimura K et al. 1984) Biochim Biophys Acta 796:269-76) and the LeX glycolipid, III3 Fuc alpha-nLc4Cer (Nishimura K et al. (1989) J. Biochem (Tokyo) 101:1315-18). The molecular species compositions of these glycosphingolipids were examined using fast atom bombardment-mass spectrometry linked with reversed-phase high-performance liquid chromatography. By such analysis, we could classify the quail glycosphingolipids into at least three classes: glycolipids rich in species having four hydroxyl groups in the ceramides (GalCer, Gg3Cer, LcGg4Cer and LeX), those rich in the ceramides of N-acyl trihydroxysphinganine with normal fatty acids (Lc3Cer), and glycolipids rich in the ceramides of N-acyl sphingenine with normal fatty acids (LacCer, Gb3Cer and Forssman glycolipid). Immunohistochemical observation implies that the differences in the hydrophobic moieties specified the localization of glycosphingolipids in the tissue.

The oligosaccharide binding specificities of CD22 beta, a sialic acid-specific lectin of B cells.

CD22 beta is a B cell surface glycoprotein involved in cell adhesion and activation. We previously reported that a recombinant soluble form termed CD22 beta Rg is capable of binding alpha 2-6 sialylated complex N-linked oligosaccharides purified from lymphocyte glycoprotein ligands (Powell, L. D., Sgroi, D., Sjoberg, E. R., Stamenkovic, I., and Varki, A. (1993) J. Biol. Chem. 268, 7019-7027). Here, we utilize a number of naturally and enzymatically sialylated oligosaccharides and sialoglycoproteins to further define its lectin specificity and demonstrate that the minimal structure recognized is Neu5Ac alpha 2-6Gal beta 1-4Glc(NAc). Reduction of the glucose residue of Neu5-Ac alpha 2-6Gal beta 1-4Glc diminishes the interaction, while truncation of the sialic acid side chain by mild periodate oxidation abolishes it. Branched oligosaccharides with two alpha 2-6-sialyl residues bind better, regardless of whether they were derived from N- or O-linked oligosaccharides or from gangliosides. alpha 2-3-Sialyl residues have no effect on binding, whereas increasing the number of alpha 2-6-sialyl residues on multiantennary oligosaccharides progressively improves binding. No specific feature of the core region affects binding, although the spacing of the alpha 2-6-sialyl residues on tetraantennary chains appears to have a significant effect. Of several model sialoglycoproteins examined, fetuin and transferrin had an apparent affinity no greater than that observed with free sialylated N-linked oligosaccharides. Some subfractions of these proteins displayed unexpectedly weak binding, suggesting that the protein backbone can exert a negative effect. In contrast, a subfraction of alpha 1-acid glycoprotein was identified as having a substantially higher apparent affinity than free oligosaccharides derived from it, indicating that multiple glycosylation sites may increase the apparent binding affinity. Thus, CD22 beta Rg contains a lectin activity specific for the minimal motif Neu5Ac alpha 2-6Gal beta 1-4Glc(NAc), and branched, multisialylated oligosaccharides are better ligands, regardless of the core sequences. Intact sialoglycoproteins can also interact, although with a variable affinity not directly predictable from the precise structure of their sialylated oligosaccharides chains. These data may help to explain why certain T and B cell surface sialoglycoproteins with the Neu5Ac alpha 2-6Gal beta 1-4Glc(NAc) motif are superior ligands, capable of mediating CD22 beta-mediated adhesion and activation events.

Immunoreactivity and function of oligosaccharides in cobra venom factor.

Cobra venom factor (CVF) is the nontoxic C-activating glycoprotein in cobra venom. It is a structural and functional analogue of complement component C3. Previous work has established that the major oligosaccharide chain of CVF is a symmetric fucosylated biantennary complex-type N-linked chain containing an alpha-galactosylated Le(x) antigenic epitope, Gal alpha 1-3Gal beta 1-4 (Fuc alpha 1-3) GlcNAc beta 1, a novel carbohydrate structure. We show that naturally occurring anti-alpha-Gal Ab present in normal human serum binds to CVF. In view of a possible human application of CVF, we studied the effect of this anti-alpha-Gal Ab on CVF function as well as the effect of oligosaccharide modification or removal on CVF activity and immunoreactivity with the anti-alpha-Gal Ab. The immunoreactivity of CVF with the anti-alpha-Gal Ab is abolished upon de-alpha-galactosylation or complete deglycosylation. De-alpha-galactosylated CVF and deglycosylated CVF were found to be equally active in forming a stable C3/C5 convertase with human factor B and in decomplementing human serum indicating that the oligosaccharide chains of CVF are not required for its C-activating function. Consistent with the unaltered functional activity, deglycosylation of the molecule caused only minor changes in secondary structure as judged by far UV circular dichroism analysis. However, the oligosaccharide chains appear to contribute to the thermal stability of CVF, because deglycosylation caused the molecule to be more sensitive to temperature. Inasmuch as rCVF produced in mammalian cells can be expected to contain sialylated oligosaccharide chains, we also prepared sialylated CVF by enzymatic sialylation of de-alpha-galactosylated and de-alpha-fucosylated CVF with 2,6-sialyltransferase. It was found that the secondary structure and the activity of sialylated CVF were unchanged compared to native CVF.

Verotoxin-binding in human renal sections.

Gastrointestinal infection with verotoxin-producing Escherichia coli (VTEC) has been strongly implicated in the etiology of the hemolytic uremic syndrome (HUS), the leading cause of pediatric acute renal failure. The binding of fluorescein-conjugated VT1 overlaid on to frozen human renal sections has been examined. Sections from biopsies of infants aged < 2 years were compared with those from adult autopsies. VT primarily stained distal convoluted tubules, particularly those adjacent to glomeruli, and collecting ducts. VT-binding was detected within the infant glomerulus but not the adult. Binding of the toxin was removed when the section was pretreated with alpha-galactosidase, confirming the receptor-binding specificity for globotriaosyl ceramide (gal alpha 1-4gal beta 1-4 glucosylceramide), the glycolipid receptor for VT. These studies may suggest that differential localization of this glycolipid in the pediatric renal glomerulus is a risk factor for the development of HUS following infection with VTEC.

Defining the minimal size of catalytically active primate alpha 1,3 galactosyltransferase: structure-function studies on the recombinant truncated enzyme.

The glycosylation enzyme alpha 1,3 galactosyltransferase, which synthesizes the carbohydrate Gal alpha 1-3Gal beta 1-4GlcNAc-R, is active in non-primate mammals, prosimians and New World monkeys, but not in Old World monkeys, apes and humans. In this study, we have cloned and sequenced the enzyme expressed in a New World monkey, determined the exact size of the stem region and assessed the minimal size of catalytically active alpha 1,3 galactosyltransferase (alpha 1,3GT). Various primer sets were used in the polymerase chain reaction to generate cDNAs which coded for forms of alpha 1,3GT with deletions at the N- or C-terminal domains. The cDNA was inserted into the expression vector pPROTA which contains the coding sequence for protein A, and subsequently transfected into COS cells. The soluble chimeric products (truncated enzyme and protein A) were harvested from the cell culture medium using IgG-Sepharose beads and assayed for enzymatic activity. As many as 67 amino acids could be truncated at the amino terminal region of the luminal portion of the enzyme without affecting its catalytic activity. Truncation of 68, 69 and 74 amino acids resulted in a 50, 75 and > 95% loss in the in vitro catalytic actively, respectively. Introduction of a frameshift mutation which is characteristic of apes and human alpha 1,3GT gene resulted in the complete loss of enzyme activity. Moreover, truncation of as few as three amino acids at the carboxyl end of alpha 1,3GT resulted in complete loss of the catalytic activity.(ABSTRACT TRUNCATED AT 250 WORDS)

Structural analysis of O-linked sugar chains in human blood clotting factor IX.

Type and structural analysis of O-linked sugar chains in human blood clotting factor IX was performed by the pyridylamination method developed for O-linked sugar chains [Kuraya, N. & Hase, S. (1992) J. Biochem. 112, 122-126]. O- and N-linked sugar chains were released with hydrazine, and then N-acetylated, followed by pyridylamination. The type of sugar chain was determined by reducing-end analysis of the pyridylaminated (PA-) sugar chains. Sugar chains with PA-GalNAc at the reducing terminal and that with PA-Fuc [Nishimura, H. et al. (1992) J. Biol. Chem. 267, 17520-17525] were obtained besides known sugar chains with PA-Glc from the Xyl-Glc-Ser type and those with PA-GlcNAc from asparagine-linked sugar chains. The sugar chains with PA-GalNAc were identified as mono- and disialyl Gal beta 1-3GalNAc by two-dimensional HPLC mapping. The structure of the sugar chain with PA-Fuc was Neu5Ac alpha 2-6Gal beta 1-4GlcNAc beta 1-3Fuc, as determined by exoglycosidase digestion, methylation analysis, and Smith degradation.

One Percent of Human Circulating B Lymphocytes Are Capable of Producing the Natural Anti-Gal Antibody

The natural anti-Gal antibody constitutes 1% of circulating IgG in humans and interacts specifically with the carbohydrate epitope Gal alpha 1-3Gal beta 1-4GlcNAc-R (the alpha-galactosyl epitope). In view of the unusually large amounts of this antibody in the serum, it was of interest to determine the proportion of circulating B lymphocytes capable of synthesizing anti-Gal. For this purpose, blood B lymphocytes were incubated with Epstein-Barr virus (EBV) and plated in microtiter wells. Proliferation of the EBV transformed B lymphocytes was readily visible after 3 weeks of incubation. The supernatants from wells containing proliferating B-lymphoid clones were assayed for anti-Gal by an agglutination assay with rabbit red blood cells and the specificity of the agglutinating antibodies was further confirmed by their interaction with synthetic oligosaccharides and by enzyme-linked immunosorbent assay with glycoproteins. Approximately 5% of the wells contained anti-Gal antibodies. Limiting dilution studies and IgH gene rearrangement patterns suggested that each well contained an average of five proliferating B-lymphoid clones. Thus, it is concluded that approximately 1% of circulating B lymphocytes are capable of producing anti-Gal. The proportion of anti-Gal--producing lymphoid clones exceeds by fourfold that of clones producing anti-blood group A or anti-blood group B antibodies. Individual anti-Gal clones display fine variations in their combining site, as indicated by their differential interaction with alpha-galactosyl epitopes on glycolipids and on N-linked carbohydrate chains of glycoproteins. The high frequency of precursor B lymphocytes capable of producing anti-Gal, found in every individual and the restricted specificity of this antibody to alpha-galactosyl epitopes, potentially makes anti-Gal--producing lymphocytes an effective system for studying human Ig genes involved in the natural immune response to structurally defined haptens.

Xenogeneic thyroid-stimulating hormone-like activity of the human natural anti-Gal antibody. Interaction of anti-Gal with porcine thyrocytes and with recombinant human thyroid-stimulating hormone receptors expressed on mouse cells.

Anti-Gal is a natural polyclonal antibody that constitutes 1% of circulating IgG in all humans and that interacts specifically with the mammalian carbohydrate epitope Gal alpha 1-3Gal beta 1-4GlcNAc-R (termed the alpha-galactosyl epitope). This epitope is abundant on thyrocytes, as well as, on other cells of nonprimate mammals, prosimians and New World monkeys, but its expression is diminished in Old World monkey, ape, and human tissues. We hypothesized that anti-Gal may bind in vitro to alpha-galactosyl epitopes on xenogeneic TSH receptors (TSHR) and mimic the effect of TSH on xenogeneic thyrocytes. Assays performed with porcine thyrocytes have indicated that anti-Gal can mimic in vitro TSH effects in stimulation for cAMP synthesis, 125I uptake, and cell proliferation. Furthermore, depletion of anti-Gal from serum of patients with Graves' disease resulted in elimination of a large proportion of the thyroid stimulating immunoglobulin activity and half of the thyroglobulin binding inhibiting Ig activity, when the sera were assayed with porcine thyrocytes. The effect of anti-Gal binding to alpha-galactosyl epitopes on TSHR was further demonstrated by the antibody-mediated stimulation for cAMP synthesis in mouse 3T3 cells (cells expressing alpha-galactosyl epitopes), which were transfected with recombinant human TSHR. CHO cells (cells lacking alpha-galactosyl epitopes), transfected with recombinant human TSHR were not stimulated by anti-Gal. It is, therefore, suggested that in studies on antibodies in Graves' disease sera, the effect of anti-Gal may be excluded by using target cells that are devoid of alpha-galactosyl epitopes. Alternatively, anti-Gal could be removed from the tested sera, before the assay with xenogeneic thyrocytes.

Evolution and pathophysiology of the human natural anti-alpha-galactosyl IgG (anti-Gal) antibody.

Anti-Gal is a human natural antibody which interacts specifically with the mammalian carbohydrate structure Gal alpha 1-3Gal beta 1-4GlcNAc-R, termed, the alpha-galactosyl epitope. This antibody constitutes approximately 1% of circulating IgG in human serum and is produced, upon stimulation, by 1% of circulating B lymphocytes. Anti-Gal is also present as IgA antibodies in body secretions such as saliva, milk and colostrum. The antigenic source for the constant production of anti-Gal seems to be the alpha-galactosyl-like epitopes found on many bacteria of the gastrointestinal flora. Whereas anti-Gal is abundant in humans, apes and Old World monkeys, it is absent from New World monkeys, prosimians and nonprimate mammals. The latter group of species produces, however, large amounts of alpha-galactosyl epitopes (> 10(6) epitopes per cell). It is estimated that anti-Gal appeared in ancestral Old World primates less than 28 million years ago, possibly as a result of an evolutionary event which exerted a selective pressure for the suppression of alpha-galactosyl epitopes expression by inactivation of the gene for the enzyme alpha 1,3 galactosyltransferase. This also resulted in the loss of immune tolerance to the alpha-galactosyl epitope and the production of anti-Gal. The physiologic role of this antibody is not clear as yet. It may participate in the protection against gastrointestinal bacteria. In addition it seems to contribute to the removal of normal and pathologically senescent red cells by interacting with the few hundred cryptic alpha-galactosyl epitopes which are exposed de novo in the course of red cell aging, thereby opsonizing these cells for phagocytosis by reticuloendothelial macrophages. The alpha-galactosyl epitope has been found to be aberrantly expressed on human cells and the interaction of anti-Gal with such epitopes may result in autoimmune disease. Preliminary data suggest such a mechanism in Graves' disease. Anti-Gal has been found to interact with therapeutic recombinant proteins expressing alpha-galactosyl epitopes, but so far there is no indication that it affects the half-life in the circulation and the biologic activity. Detection of anti-Gal in the seminal fluid and in the cerebrospinal fluid may serve as a simple means for assessment of damage to the blood-genital tract barrier or the blood-brain barrier. Studies on the interaction of anti-Gal with aberrantly expressed alpha-galactosyl epitopes on human cells may elucidate the possible role of anti-Gal in human autoimmune diseases.

E-selectin involvement in in vitro adhesion of blood dendritic cells to human umbilical cord endothelial cells.

Peripheral blood dendritic cells (BDC) are potent antigen-presenting lymphoid cells. In the present study, we have examined the in vitro adhesion of BDC to human umbilical cord venous endothelial cells (HUVEC) and studied the expression of CD molecules and oligosaccharide haptens on BDC and endothelial cells. Immunohistochemistry showed that BDC were strongly positive for antibodies against HLA-DR, CD11c, CD18, CD44 and CD54, and moderately positive for anti-CD11a, CD31, CD43 and CD58. In addition, BDC were moderately positive for anti-Sialyl Lewis a and strongly positive for anti-Sialyl Lewis x and CD77 (Gal alpha 1-4Gal beta 1-4Glc) Non-stimulated HUVEC were positive for anti-CD29, CD31 and CD77. An in vitro adhesion assay showed that only a small percentage of radiolabelled BDC bound to non-stimulated HUVEC (16.9 +/- 5.9%, mean +/- SD). Stimulation of the HUVEC with IL-1 for 4 h produced a significant increase (P < 0.002) in the percentage of radiolabelled BDC that bound to HUVEC (42.3 +/- 7.1%). Preincubation of HUVEC with antibodies against E-selectin (10 micrograms/ml) significantly inhibited (P < 0.02) the binding of radiolabelled BDC to activated HUVEC (32.2 +/- 1.3%) whereas preincubation of BDC with antibodies against CD54, CD18, CD11b, CD11c and Sialyl Lewis x did not produce any significant inhibition. Preincubation of BDC with Sialyl Lewis a antibody and with isotype-matched control antibodies did not affect the increased binding of BDC to IL-1-activated HUVEC. Thus, E-selectin seems to be involved in adhesion of BDC to IL-1-stimulated HUVEC.

Roles of the pap- and prs-encoded adhesins in Escherichia coli adherence to human uroepithelial cells.

In this study, we reexamined the structural prerequisites for the attachment of P-fimbriated Escherichia coli to human urinary tract epithelial cells. The epithelial cells were obtained from A1P1 nonsecretor individuals, who express the globoseries of glycolipids without the ABH blood group determinants, and from A1P1 secretor individuals, who in addition express globo-A, a receptor for the prsJ96 adhesin. The wild-type E. coli strains J96, AD110, and IA2 and the recombinant clones HB101 papJ96, HB101 prsJ96, HB101 papIA2, and HB101 papAD110 were tested for binding. They expressed P fimbriae, as defined by P blood group-dependent agglutination of human erythrocytes of the globoseries, but differed in reactivity with galactose alpha 1-4galactose beta (Gal alpha 1-4Gal beta)-latex beads, isolated glycolipids of the globoseries, sheep erythrocytes, and uroepithelial cells. Three different patterns of binding were represented among the recombinant clones. HB101 papIA2 and HB101 papAD110 agglutinated sheep erythrocytes and Gal alpha 1-4Gal beta-latex beads and attached to both secretor and nonsecretor epithelial cells. HB101 prsJ96 agglutinated sheep erythrocytes, reacted poorly with Gal alpha 1-4Gal beta-latex beads, and attached to A1 secretor but not to A1 nonsecretor epithelial cells. HB101 papJ96 agglutinated Gal alpha 1-4Gal beta-latex beads but not sheep erythrocytes and attached poorly to human uroepithelial cells. The receptors relevant for adhesion were analyzed by inhibition with glycolipids in suspension. The sheep erythrocyte agglutination and attachment to secretor and nonsecretor epithelial cells of HB101 papIA2 and HB101 papAD110 were inhibited by globotetraosylceramide, while the Forssman glycolipid had no effect. The sheep erythrocyte reactivity and attachment to secretor epithelial cells of HB101 prsJ96 were inhibited by the Forssman glycolipid. These results permitted three conclusions. First, the expression of functionally active Gal alpha 1-4Gal beta-specific adhesins, as in HB101 papJ96, was not sufficient to make E. coli competent to attach to human uroepithelial cells. Attachment required P fimbriae of the papIA2 or papAD110 type. Second, the sheep erythrocyte reactivity of P-fimbriated strains could not be attributed solely to recognition of the Forssman glycolipid and may not be used to define the prsJ96-encoded phenotype. Third, the P-fimbrial adhesins which mediate secretor state-independent attachment to human uroepithelial cells recognized receptor epitopes provided by globotetraosylceramide.

Salivary receptors for GalNAc beta-sensitive adherence of Actinomyces spp.: evidence for heterogeneous GalNAc beta and proline-rich protein receptor properties.

The receptors for GalNAc beta 1-3Gal alpha Oethyl (GalNAc beta)-sensitive adherence of Actinomyces strains to salivary pellicles were investigated. Parotid and submaxillary saliva from one individual was size fractionated and utilized in hydroxyapatite adherence assays with Actinomyces naeslundii 12104 and A. viscosus 19246 and LY7 with and without GalNAc beta. Three parotid salivary fractions, the high-molecular-weight, acidic proline-rich protein (PRP), and statherin fractions, promote GalNAc beta-sensitive adherence of strain 12104, whereas only the high-molecular-weight fraction of submaxillary saliva promotes such adherence. In contrast, strain LY7, possessing a variant GalNAc beta specificity, shows GalNAc beta-sensitive adherence to the leading and trailing regions of the submaxillary PRP fractions but less distinct adherence to the parotid and submaxillary high-molecular-weight fractions. In addition, the PRP and statherin fractions promote adherence of strains LY7 and 19246 that is not inhibited by GalNAc beta. However, whereas strain LY7 binds more strongly to the PRP fraction than to the statherin fraction, strain 19246 binds preferentially to the statherin fractions of parotid and submaxillary saliva. These salivary protein fractions were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and immunostained to detect glycosylated proteins. The different salivary receptor properties are paralleled by different glycosylation patterns. The variable GalNAc beta specificities may have evolved to match different salivary glycosylation patterns, and PRP and statherin binding properties seem to be heterogeneous among the Actinomyces strains.