Chemical Resolution of an Epitope Recognized by Blood Group Antibodies Capable of Binding Both A and B Red Blood Cells.

The high specificity of human antibodies to blood group A and B antigens is impressive, especially when considering the structural difference between these antigens (tetrasaccharides) is a NHAc versus a OH-group on the terminal monosaccharide residue. It is well established that in addition to anti-A and anti-B there is a third antibody, anti-A,B capable of recognizing both A and B antigens. To analyze this AB specificity, we synthesized a tetrasaccharide, where the NHAc of the A antigen was replaced with NH2. This NH2-group was used to attach the glycan to an affinity resin, creating an AB-epitope (ABep) adsorbent where the critical site for recognition by A and B antibodies was not accessible, while the rest of the (conformationally compact) tetrasaccharide remained accessible. Anti-ABep antibodies were isolated from blood group O donors and found to have expected A,B-specificity against immobilized and red cell bound synthetic antigens, including ABep, and were able to agglutinate both A and B red cells. The amount of these anti-ABep (anti-A,B) antibodies found in the blood of group O donors was comparable to levels of anti-A and anti-B found in group B and A individuals. Using STD-NMR the location for the AB-epitope on the tetrasaccharide was found.

Specificity profile of αGal antibodies in αGalT KO mice as probed with comprehensive printed glycan array: Comparison with human anti-Galili antibodies.

BACKGROUND: The α1,3-galactosyltransferase gene-knockout (GalT KO) mice are able to produce natural anti-αGal antibodies apparently without any specific immunization. GalT KO mice are commonly used as a model immunological system for studying anti-αGal responses to Gal-positive xenografts in human. In this study, we compared the specificity of mouse and human αGal antibodies to realize the adequacy of the murine model. METHODS: Using hapten-specific affinity chromatography antibodies against Galα1-3Galß1-4GlcNAcß epitope were isolated from both human and GalT KO mice blood sera. Specificity of isolated antibodies was determined using a printed glycan array (PGA) containing 400 mammalian glycans and 200 bacterial polysaccharides. RESULTS: The quantity of isolated specific anti-Galα antibodies corresponds to a content of <0.2% of total Ig, which is an order of magnitude lower than that generally assumed for both human and murine peripheral blood immunoglobulin, with a high predominance of IgM over IgG (95% vs 5%). Analysis using a printed glycan array has demonstrated that (a) antibodies from both species bind not only the Galα1-3Galß1-4GlcNAcß epitope, but also unrelated glycans; (b) particularly, for human (but not mouse) antibodies the best binders appear to be bacterial polysaccharides; (c) the profile of mouse antibodies is broader, it is noteworthy that they recognize a variety of human blood group B epitopes and even glycans without the α-galactosyl residue. CONCLUSIONS: We believe that the mouse model should be used cautiously in xenotransplantation experiments when the fine epitope specificity of antibodies is critical.

Human Natural Antibodies Recognizing Glycan Galß1-3GlcNAc (LeC).

The level of human natural antibodies of immunoglobulin M isotype against LeC in patients with breast cancer is lower than in healthy women. The epitope specificity of these antibodies has been characterized using a printed glycan array and enzyme-linked immunosorbent assay (ELISA), the antibodies being isolated from donors' blood using LeC-Sepharose (LeC is Galß1-3GlcNAcß). The isolated antibodies recognize the disaccharide but do not bind to glycans terminated with LeC, which implies the impossibility of binding to regular glycoproteins of non-malignant cells. The avidity (as dissociation constant value) of antibodies probed with a multivalent disaccharide is 10-9 M; the nanomolar level indicates that the concentration is sufficient for physiological binding to the cognate antigen. Testing of several breast cancer cell lines showed the strongest binding to ZR 75-1. Interestingly, only 7% of the cells were positive in a monolayer with a low density, increasing up to 96% at highest density. The enhanced interaction (instead of the expected inhibition) of antibodies with ZR 75-1 cells in the presence of Galß1-3GlcNAcß disaccharide, indicates that the target epitope of anti-LeC antibodies is a molecular pattern with a carbohydrate constituent rather than a glycan.

Specificity of human natural antibodies referred to as anti-Tn.

To understand the role of human natural IgM known as antibodies against the carbohydrate epitope Tn, the antibodies were isolated using GalNAcα-Sepharose affinity chromatography, and their specificity was profiled using microarrays (a glycan array printed with oligosaccharides and bacterial polysaccharides, as well as a glycopeptide array), flow cytometry, and inhibition ELISA. The antibodies bound a restricted number of GalNAcα-terminated oligosaccharides better than the parent monosaccharide, e.g., 6-O-Su-GalNAcα and GalNAcα1-3Galß1-3(4)GlcNAcß. The binding with several bacterial polysaccharides that have no structural resemblance to the affinity ligand GalNAcα was quite unexpected. Given that GalNAcα is considered the key fragment of the Tn antigen, it is surprising that these antibodies bind weakly GalNAcα-OSer and do not bind a wide variety of GalNAcα-OSer/Thr-containing mucin glycopeptides. At the same time, we have observed specific binding to cells having Tn-positive glycoproteins containing similar glycopeptide motifs in a conformationally rigid macromolecule. Thus, specific recognition of the Tn antigen apparently requires that the naturally occurring "anti-Tn" IgM recognize a complex epitope comprising the GalNAcα as an essential component and a fairly long amino acid sequence where the amino acids adjacent to GalNAcα do not contact the antibody paratope; i.e., the antibodies recognize a spatial epitope or a molecular pattern rather than a classical continuous sequence. In addition, we have not found any increase in the binding of natural antibodies when GalNAcα residues were clustered. These results may help in further development of anticancer vaccines based on synthetic Tn constructs.

Are there specific antibodies against Neu5Gc epitopes in the blood of healthy individuals?

Strong discrepancies in published data on the levels and epitope specificities of antibodies against the xenogenic N-glycolyl forms of sialoglycans (Hanganutziu-Deicher Neu5Gcɑ2-3Galß1-4Glc and related antigens) in healthy donors prompted us to carry out a systematic study in this area using the printed glycan array and other methods. This article summarizes and discusses our published and previously unpublished data, as well as publicly available data from the Consortium for Functional Glycomics. As a result, we conclude that (1) the level of antibodies referred to as anti-Neu5Gc in healthy individuals is low; (2) there are antibodies that seem to interact with Neu5Gc-containing epitopes, but in fact they recognize internal fragments of Neu5Gc-containing glycans (without sialic acids), which served as antigens in the assays used and; (3) a population capable of interacting specifically with Neu5Gc (it does not bind the corresponding NAc analogs) does exist, but it binds the monosaccharide Neu5Gc better than the entire glycans containing it. In other words, in healthy donors, there are populations of antibodies capable of binding the Neu5Gc monosaccharide or the inner core -Galß1-4Glc, but very few true anti-Neu5Gcɑ2-3Galß1-4Glc antibodies, i.e., antibodies capable of specifically recognizing the entire trisaccharide.

Why human anti-Galα1-4Galß1-4Glc natural antibodies do not recognize the trisaccharide on erythrocyte membrane? Molecular dynamics and immunochemical investigation.

BACKGROUND: Human blood contains a big variety of natural antibodies, circulating throughout life at constant concentration. Previously, we have found natural antibodies capable of binding to trisaccharide Galα1-4Galß1-4Glc (Pk) practically in all humans. Intriguingly, the same trisaccharide is a key fragment of glycosphingolipid globotriaosylceramide (Gb3Cer) - normal component of erythrocyte and endothelial cell membrane, i.e. the antibodies and their cognate antigen coexist without any immunological reaction. AIM: To explain the inertness of human anti-Pk antibodies towards own cells. MATERIALS AND METHODS: We used a combination of immunochemical and molecular dynamics (MD) experiments. Antibodies were isolated using affinity media with Pk trisaccharide, their epitope specificity was characterized using ELISA (enzyme-linked immunosorbent assay) with a set of synthetic glycans related to Pk synthetic glycans and FACS (Fluorescence-Activated Cell Sorting) analysis of cells with inserted natural Gb3Cer and its synthetic analogue. Conformations and clustering of glycolipids immersed into a lipid bilayer were studied using MD simulations. RESULTS: Isolated specific antibodies were completely unable to bind natural Gb3Cer both inserted into cells and in artificial membrane, whereas strong interaction took place with synthetic analogue differing by the presence of a spacer between trisaccharide and lipid part. MD simulations revealed: i) although membrane-bound glycans do not form stable long-living aggregates, their transient packing is more compact in natural Gb3 as compared with the synthetic analog, ii) similar conformation of Pk glycan in composition of the glycolipids, iii) no effect on the mentioned above results when cholesterol was inserted into membrane, and iv) better accessibility of the synthetic version for interaction with proteins. CONCLUSIONS: Both immunochemical and molecular dynamics data argue that the reason of the "tolerance" of natural anti-Pk antibodies towards cell-bound Gb3Cer is the spatial inaccessibility of Pk glycotope for interaction. We can conclude that the antibodies are not related to the blood group P system.

Natural Antibodies Against Sialoglycans.

Natural antibodies, part of the innate immunity system, are produced at strictly regulated levels in normal sera without immunization and thus are part of the innate immune system. The best studied natural antibodies are those directed against blood group antigens A and B and xeno-antigens including glycolylneuraminic acid containing Hanganutziu-Deicher (HD) glycolipid. Abnormal levels of anti-glycan antibodies were found in a number of pathologies. In many cases pathological antibodies are known to bind gangliosides. The genesis of anti-glycan antibodies in healthy humans and the reasons for their changes in pathologies are poorly understood. With a growing interest in their diagnostic applications, it is important to determine the carbohydrate structures that are recognized by antibodies present in the circulation of healthy individuals. We tested a large number of healthy donors using a printed glycan array (PGA) in a microchip format. The PGA contained ~300 glycans, representing mostly normal mammalian structures of glycoproteins and glycolipids, and many of the structures presented are biologically relevant sialylated motifs. As revealed by PGA, the sera interacted with at least 70 normal human glycans. With only few exceptions, antibodies recognizing sialosides have not been identified. Moderate levels of antibodies and moderate variability were observed in the case of SiaT n and its glycolyl variant. Unexpectedly, we found minimal antibody titer directed against Neu5Gcα and the trisaccharide Neu5Gcα2-6Galß1-4GlcNAc, although this form of neuraminic acid does not occur naturally in humans. Antibodies recognizing sialosides in unnatural ß-configuration have been detected and confirmed Springer's paradigm that circulating antibodies represent a reaction against bacteria. Gram-negative bacteria contain LPS with ßKDN and/or ßKDO which are very close analogs of Neu5Ac that are found in ß-connected form. Antibodies against the biantennary N-glycan chain, (Neu5Acα2-6Galß1-4GlcNAcß1-2Manα)2-3,6-Manß1-4GlcNAcß1-4GlcNAc were never observed and similarly we never saw antibodies directed against the SiaLe(a)/SiaLe (x) motifs. Anti-sialoglycan antibodies can be masked with gangliosides: for example, we observe about a five times higher level of anti-GD3 in purified total IgG compared to the same concentration of total Ig in the composition of native serum. For several antibodies we observed anomalous binding in diluted sera, namely, the signals towards sialylated glycans were increased in the PGA if diluted sera were used.

Repertoire of human natural anti-glycan immunoglobulins. Do we have auto-antibodies?

BACKGROUND: Profiling of donor's antibodies using glycan arrays demonstrated presence of antibodies capable of binding to >100 mammalian glycans or their fragments. For example, relatively high binding to Galα1-4Galß1-4GlcNAc (P(1)), Galα1-4Galß1-4Glc (P(k)), Galß1-3GlcNAc (Le(c)), 4-O-SuGalß1-4GlcNAc, and GalNAcα1-3GalNAc (Fs) was found in all tested individuals. Affinity isolation using hapten-specific chromatography in combination with epitope mapping revealed their glycotopes. Notably, a significant part of the antibodies was capable of recognizing a fragment of larger glycans, for example, -Galß1-4Glc of glycolipids, or Fucα1-3GlcNAc motif of Le(X)/Le(Y) antigens. Their epitope specificity did not vary between different healthy individuals. Nominally, all the mentioned immunoglobulins could be classified as auto-antibodies. METHODS: In this work we re-evaluated results published earlier and analyzed new data to address the question why autologous antibodies found in healthy individuals do not cause severe auto-immune reactions. RESULTS: In all cases the presumably "auto" antibodies were found to bind short fragments "subtracted" from larger glycans whereas recognition of the same fragment in the context of the whole natural chain was completely abolished. Thus, in spite of numerous formally positive signals observed on the printed glycan array, we are yet unable to identify in blood serum of healthy individuals true auto-antibodies capable of binding carbohydrate chains in their naturally occurring form. GENERAL SIGNIFICANCE: The identified natural anti-glycan antibodies were found to be specific, high-titer and population conservative immunoglobulins - all of this suggesting as yet unknown biological role(s) of the studied proteins. This article is part of a Special Issue entitled Glycoproteomics.

Natural anti-A and anti-B of the ABO system: allo- and autoantibodies have different epitope specificity.

BACKGROUND: According to Landsteiner's law, alloantibodies are prevalent and autoantibodies are absent in the ABO blood group system. However, one study (Spalter et al., Blood 1999;93:4418-24) has suggested that low-affinity ABO autoantibodies, mitigated by anti-idiotypic immunoglobulins are also prevalent, while another publication (Rieben et al., Eur J Immunol 1992;22:2713-7) shows that humans do not have B-lymphocytes capable of producing immunoglobulin G ABO autoantibodies. STUDY DESIGN AND METHODS: We used hapten-specific chromatography to isolate allo- and autoantibodies from pools of A or B serum and then characterized the resultant antibodies against a wide range of ABO and related glycoconjugates. RESULTS: We found that the apparent autoantibodies are directed against blood group A or B disaccharides, without consideration for the presence of fucose, but requiring the absence of elongating sugar X in composition of Gal(NAc)α1-3(Fucα1-2)Galß1-X-terminated carbohydrate chain. In contrast, ABO alloantibodies required a minimum trisaccharide Gal(NAc)α1-3(Fucα1-2)Gal epitope and recognize the elongated type-specific tetrasaccharides. Furthermore, alloantibodies appear to be a small set of specific yet crossreactive antibodies that detect all backbone types of A or B antigens, rather than being a collection of specific antibodies, each of which detects a different type of A or B antigen. CONCLUSION: Apparent ABO autoantibodies appear to have no natural human target.

Profiling of serum antibodies with printed glycan array: room for data misinterpretation.

Using an example of Galß1-3GlcNAc (Le(C)) related glycans, we here demonstrate a risk of data misinterpretation when polyclonal antibodies are probed for their glycan-binding specificities with help of a printed glycan array (PGA). Affinity isolation of antibodies from human serum using Le(C)-Sepharose or 3'-O-SuLe(C)-Sepharose in conditions of excess of the adsorbents generated identical material regardless of the affinity ligand, with the antibodies equally capable of binding to Le(C) and to 3'-O-SuLe(C) disaccharides, as well as to 3'-O-SiaLe(C) trisaccharide. More detailed profiling has shown that the isolated antibodies bind to the inner part of Galß1-3GlcNAc disaccharide. We therefore conclude that serum does not contain different subsets of antibodies specific either to Le(C) or to 3'-O-SuLe(C), despite their visibly different binding signals to these glycans on PGA.

Anti-carbohydrate antibodies of normal sera: findings, surprises and challenges.

We have used microchip format glycan array to characterize the individual carbohydrate recognition patterns by antibodies (Ab) in sera of 106 healthy donors. The glycan library included blood group antigens and other most frequent terminal oligosaccharides and their cores of mammalian N- and O-linked glycoproteins and glycolipids, tumor-associated carbohydrate antigens, and common components of bacterial/pathogenic polysaccharides and lipopolysaccharides, totally 205 glycans. The serum Ab interacted with at least 50 normal human glyco-motifs. Apart from expected blood group-, xeno- (heterophil) and infection-related binding activities, we observed a number of new and unexpected features. The surprising, relatively high antibody binding was found to the blood group P(1) and P(k) trisaccharides and H(type 2) trisaccharide. Novel and very high binding activities have been observed towards Galbeta1-3GlcNAc (Le(C)) related glycans, especially 3'-O-Su-Le(C), and towards 4'-O-sulfated lactosamine. Relatively high and uniform Ab binding to GalNAcalpha1-3Gal disaccharide demonstrated absence of correlation with fucosylated blood group A GalNAcalpha1-3(Fucalpha1-2)Gal antigen-similarly to well known relationship between Galalpha1-3Gal and true, fucosylated blood group B Galalpha1-3(Fucalpha1-2)Gal antigen. The binding intensity to Galalpha1-3Galbeta1-4GlcNAc xenoantigen was shown to be rather modest. Absence or very low Ab binding was found against oligosialic acid, sialooligosaccharides except SiaT(n), type 2 backbone glycans such as Le(y), and biantennary N-chain as well as its truncated forms, i.e. without terminal Sia, SiaGal, and SiaGalGlcNAc motifs. We have also found that Ab are capable of recognizing the short inner core typical for glycolipids (-Galbeta1-4Glc) and glycoproteins (-GalNAcalpha) as a fragment of bigger glycans.

Normal human serum contains high levels of anti-Gal alpha 1-4GlcNAc antibodies.

BACKGROUND: Natural xenoreactive antibodies (Abs) directed against the Bdi-epitope (Gal alpha 1-3Gal beta) on the cells of non-primate mammals take part in hyperacute rejection of xenotransplanted organs. We found that some Abs, which were one-step affinity purified on Bdi-Sepharose, cross-reacted with the disaccharide Gal alpha 1-4GlcNAc beta. The epitope Gal alpha 1-4GlcNAc has not been identified on mammals or bacterial polysaccharides yet. METHODS: To isolate the antibodies of the corresponding specificity the disaccharide was immobilized on Sepharose and antibodies were affinity purified from pooled serum of blood group O individuals. RESULTS: These one-step purified Abs cross-reacted with Bdi, but after a prior absorption step on Bdi-Sepharose no cross-reactivity with Bdi was observed any longer. Surprisingly, the quantity of anti-Gal alpha 1-4GlcNAc isolated from the same serum pool, 4-7 microg/ml, was equal to that of anti-Bdi or more. Independently of ABO blood groups all the tested healthy donors had anti-Gal alpha 1-4GlcNAc Abs at a similar level. Monospecific anti-Gal alpha 1-4GlcNAc Abs were not cytotoxic towards porcine cells. CONCLUSIONS: 1. The actual concentration of monospecific, xenoreactive Gal alpha 1-3Gal beta Abs in blood may be considerably lower than the value referred to in the literature for 'anti-alpha Gal' or 'anti-Galili' antibodies. 2. Anti-Gal alpha 1-4GlcNAc Abs seem not to be important for xenotransplantation.