Monoclonal anti-type 2 H: an antibody detecting a precursor of the A and B blood group antigens

Expression of human alpha-l-fucosyltransferase gene homologs in monkey kidney COS cells and modification of potential fucosyltransferase acceptor substrates by an endogenous glycosidase

Previous investigations on the monkey kidney COS cell line demonstrated the weak expression of fucosylated cell surface antigens and presence of endogenous fucosyltransferase activities in cell extracts. RT-PCR analyses have now revealed expression of five homologs of human fucosyltransferase genes, FUT1, FUT4, FUT5, FUT7, and FUT8, in COS cell mRNA. The enzyme in COS cell extracts acting on unsialylated Type 2 structures is closely similar in its properties to the alpha1,3-fucosyltransferase encoded by human FUT4 gene and does not resemble the product of the FUT5 gene. Although FUT1 is expressed in the COS cell mRNA, it has not been possible to demonstrate alpha1,2-fucosyltransferase activity in cell extracts but the presence of Le(y) and blood-group A antigenic determinants on the cell surface imply the formation of H-precursor structures at some stage. The most strongly expressed fucosyltransferase in the COS cells is the alpha1,6-enzyme transferring fucose to the innermost N -acetylglucosamine unit in N -glycan chains; this enzyme is similar in its properties to the product of the human FUT8 gene. The enzymes resembling the human FUT4 and FUT8 gene products both had pH optima of 7.0 and were resistant to 10 mM NEM. The incorporation of fucose into asialo-fetuin was optimal at 5.5 and was inhibited by 10 mM NEM. This result initially suggested the presence of a third fucosyltransferase expressed in the COS cells but we have now shown that triantennary N- glycans with terminal nonreducing galactose units, similar to those present in asialo-fetuin, are modified by a weak endogenous beta-galactosidase in the COS cell extracts and thereby rendered suitable substrates for the alpha1,6-fucosyltransferase.

Recognition of the blood group H type 2 trisaccharide epitope by 28 monoclonal antibodies and three lectins

The patterns of cross-reaction of 30 monoclonal antibodies and three lectins were determined by ELISA with 21 ABH, Ii or Lewis related synthetic oligosaccharides coupled to bovine serum albumin. At least seven main groups of cross-reactive patterns were identified among the antibodies, plus several isolated antibodies which had intermediate patterns between two of the main antibody groups. The three lectins had different cross-reaction patterns, Galactia tenuiflora was different from all the antibodies, Ulex europaeus lectin 1 and Lotus tetragonolobus were similar, but not identical to groups III and V of antibodies respectively. The anti-H antibodies cross-reacting with A type 2 gave similar agglutination scores with all the normal ABO erythrocytes, while the anti-H antibodies not cross-reacting with A type 2 reacted with different scores: O > A2 > A2B > B > A1 > A1B > O(h), suggesting that these antibodies react better with the free H epitopes and do not recognize the H in A or B epitopes. Based on the ELISA and agglutination results and the lowest energy conformations of each oligosaccharide obtained by computer modelling, the most probable oligosaccharide surface areas recognized by each antibody main group are illustrated.

Characterization of a monoclonal antibody specific for H type 2 structure of ABO blood group, and its use for measuring H type 2 on human red blood cells

A monoclonal antibody 3A5 specific for the human red blood cells was produced by immunizing BALB/c mouse with human erythrocyte membranes of group O following the immunization protocol of selectively killing the antigen-stimulated lymphocytes. The monoclonal antibody 3A5 we obtained agglutinated red blood cells regardless of individuals of blood group A, B, or O types, but not those from a person with a rare para-Bombay type which is the H-deficient phenotype. The hemagglutination reaction of 3A5 was not inhibited by saliva from either secretor or nonsecretor individuals. The specificity of 3A5 was studied by adsorption with and elution from synthetic oligosaccharide immunoadsorbents including H disaccharide, and H type 1, 2, 3, and 4 structures. Also, the reactivity of 3A5 with synthetic oligosaccharide-BSA complexes for H type 1, H type 2, Le(a), Le(b), A, and B was determined by an immunoassay. We found that 3A5 did not react with any of these synthetic oligosaccharides except H type 2. From these results, the antigenic epitope recognized by 3A5 was demonstrated to be the H type 2 structure. Additionally, the H type 2 substance on erythrocytes was quantitatively analyzed using 3A5.

Immunochemical characterization of anti-H monoclonal antibodies obtained from a mouse immunized with human saliva

Three IgM class anti-H monoclonal antibodies (1E3, 1E5 and 3H1) were obtained from a BALB/c mouse immunized with human O type saliva. These antibodies were found to agglutinate red cells from O group and A and B subgroups but not from Bombay and para-Bombay individuals whose H antigen was barely detected by anti-H reagents. The agglutination reactions of these antibodies were inhibited by H antigens from human tissues. It was also demonstrated that both 1E3 and 3H1 reacted with H disaccharide (Fuc alpha 1-->2Gal beta), H type 1 (Fuc alpha 1-->2Gal beta 1-->3GlcNAc beta), H type 2 (Fuc alpha 1-->2Gal beta 1-->4GlcNAc beta), H type 3 (Fuc alpha 1-->2Gal beta 1-->3GalNAc alpha) and H type 4 (Fuc alpha 1-->2Gal beta 1-->3GalNAc beta) but not with Lea (Gal beta 1-->3[Fuc alpha 1-->4]GlcNAc beta), Leb (Fuc alpha 1-->2Gal beta 1-->3[Fuc alpha 1-->4]GlcNAc beta), X (Gal beta 1-->4[Fuc alpha-->3]GlcNAc beta) or Y (Fuc alpha 1-->2Gal beta 1-->4[Fuc alpha 1-->3]GlcNAc beta). On the other hand, 1E5 was found to react with H type 1, H type 2, Leb and Y. Because of the unique reactivities against various fucosyl linkages these monoclonal antibodies could be useful not only as anti-H reagents but also as reagents for the structural analysis of fucosylated glycoconjugates.

Isolation and characterization of anti-H antibody from egg yolk or immunized hens

An anti-H antibody was demonstrated to be produced in egg yolk as well as in serum of hens which were immunized with human type O red blood cells. The antibody in egg yolk was isolated with polyethylene glycol and ethanol and was purified by affinity chromatography using immunoadsorbant beads immobilized with H type 2 hapten (Fuca1----2Gal beta 1----4GlcNAc beta). Hemagglutination reaction of the antibody from egg yolk was inhibited by human saliva samples from secretor types irrespective of their ABO blood types, and by immunoadsorbant beads which contain Fuca1----2Gal beta structures.

Analysis of the fine specificities of 11 mouse monoclonal antibodies reactive with type 2 blood group determinants

The specificity of 11 mouse monoclonal antibodies reacting selectively with type 2 blood group structures was analyzed in detail by studying their reactivities with a panel of standard glycolipids, glycolipids from erythrocytes and blood group glycoproteins. The antibodies reacted with monofucosyl type 2 H, difucosyl type 2 structures (Le gamma) or both; none of the antibodies reacted with type 1 (H, Lea, or Leb) structures. Only a small proportion of the antibodies were completely specific for either type 2H or Le gamma structures. None of the antibodies had identical patterns of reactivity and their specificities were individually distinct. Seven antibodies preferentially agglutinated O and A2 erythrocytes. Anti-Le gamma-specific antibodies, except mAb101, did not agglutinate erythrocytes or react with glycolipids from erythrocytes, indicating the absence of Le gamma structures in erythrocyte glycolipids. The ability of some antibodies to react with A erythrocytes was shown to be due to cross-reactivity of the antibodies with type 3 (repetitive) A structures. The study demonstrates that monoclonal anti-carbohydrate antibodies tend to react with a range of related, and even distantly related, structure in a pattern characteristic of each antibody and that very few antibodies have extremely restricted specificities.

LKE red cell antigen and its relationship to P1 and Pk: serological study of a large family

The fourth example of human anti-LKE was identified in the serum of an antenatal patient. Study of the red cells of the proband and her family confirmed the recessive inheritance of the LKE- phenotype. The blood groups of the family confirmed that Pk expression is greater on cells from LKE- members than on those from LKE+ members. In this family, the expression of LKE varied with the P1 phenotype. LKE- individuals occurred with an incidence of 0.0017 in the donor population of the Glasgow and West of Scotland Region.

A new monoclonal antibody (A46-B/B10) highly specific for the blood group H type 2 epitope: generation, epitope analysis, serological and histological evaluation

A monoclonal antibody recognizing the blood group H type 2 antigen has been obtained from a BALB/c mouse immunized with MCF-7 (human mammary carcinoma) cells. The specificity of this antibody (A46-B/B10, IgM, kappa) has been identified by haemagglutination tests, immunohistochemistry, binding inhibition studies, and absorption experiments performed with synthetic oligosaccharides. The antibody is virtually nonreactive with H type 1 antigen or with closely related type 2 structures (e.g., Y antigen). A46-B/B10 strongly agglutinates human erythrocytes according to the amount of H substance expressed and can, therefore, easily discriminate between blood groups A1 and A2 as well as A1B and A2B (A1 and A1B are not or only weakly agglutinated). In immunohistochemistry, this antibody seems to provide a highly specific reagent for a restricted number of carcinomas and epithelial lineages in tissue sections and in vitro.

Recognition of galactosylgloboside by monoclonal antibodies derived from patients with primary lung cancer

Lymph node lymphocytes from patients with primary lung cancer were immortalized with Epstein-Barr virus, and culture supernatants were screened for cell-surface reactivity against allogeneic cancer cell lines. The percentage of wells containing detectable antibodies in initial screening ranged from 1 to 17%, but the vast majority of the cultures lost antibody activity on subsequent expansion. Two antibody-secreting clones, J309 and D579, derived from separate individuals and reactive with anaplastic lung cancer cell lines, were successfully expanded and fused with the NS-1 mouse myeloma cell line. The antibodies produced by these clones exhibited identical restricted serologic reactivity against cultured cell lines and detected a carbohydrate antigen present in the neutral glycolipid fraction of MCF-7 breast cancer cells. Serologic, immunochemical, and chemical analyses revealed that the antigen recognized by antibodies J309 and D579 is galactosylgloboside [Gal(beta 1----3)GalNAc(beta 1----3)Gal(alpha 1----4)Gal(beta 1----4)- GlcCer]. Conclusions regarding the significance of these findings with respect to the biology of lung cancer await further information concerning the distribution of galactosylgloboside in normal and malignant tissues and the frequency of antibodies to this structure in normal and tumor-bearing individuals.

Carbohydrate sequences detected by murine monoclonal antibodies

The carbohydrate sequences of cell surface glycolipids change during differentiation and oncogenic transformation. To detect these structural changes, murine monoclonal antibodies have been produced in many different laboratories. Some of these antibodies are used to distinguish various cell types such as normal and transformed cells, while others are used to analyze developmentally regulated antigens. Recently, the structures of many of these carbohydrate antigens have been determined. The availability of these well-defined monoclonal antibodies will be useful for the study of the regulation and function of glycoconjugates.

Blood-group-related carbohydrate antigens are expressed on human milk galactosyltransferase and are immunogenic in rabbits

Immunochemical evidence is presented for the presence of blood-group-related carbohydrate structures on human milk galactosyltransferase and for the occurrence of the corresponding specificities among rabbit antibodies to this enzyme. Although these carbohydrate specificities constitute minor populations among antisera and affinity-purified antibodies to galactosyltransferase, their presence is important in the immunohistochemical approach to enzyme localization, since they give rise to strong reactivities with epithelial cells of the gastrointestinal tract.

The expression of ABH antigens during in vitro megakaryocyte maturation: origin of heterogeneity of antigen density

An indirect immunofluorescence technique with single and double labelling has been used to examine cultured human megakaryocytes for ABH antigens. This technique demonstrated the presence of these antigens on megakaryocytes and a population of small mononuclear cells that probably represent the differentiated precursors of megakaryocytes. In contrast to the intense homogeneous labelling with human anti-PlA1, the labelling with human anti-A, mouse monoclonal anti-A and anti-type 2H is light and heterogeneous, with many cells staining weakly or not at all. This variability in blood group ABH surface antigen expression appears to occur at the level of cell proliferation within colonies. While cells within an individual colony are homogeneous in fluorescence intensity, there is considerable variation between colonies. The progeny of individual megakaryocytes also appear uniform in ABH antigen expression.

Status of major red cell blood group antigens on neutrophils, lymphocytes and monocytes

Previous investigators have reported that a number of red cell (RBC) antigens are found on neutrophils (PMN) and lymphocytes. However, there is a lack of consensus in the literature. Furthermore, few data are available concerning the occurrence of RBC antigens on monocytes. To address this problem, leucocyte fractions prepared from donors of known RBC antigen phenotype were analysed by fluorescence flow cytometry. Based on recent reports that PMN do not express ABH antigens and on the data presented here, the occurrence of major RBC antigens on leucocytes may be defined as follows: (1) lymphocytes express A, B. Lea and Leb antigens, depending on the ABH secretor (Se) status of the donor; (2) ABH and Lewis antigens cannot be detected on monocytes and PMN regardless of Se status; (3) D, E, e, C, c, Fya, Fyb, Fy5, Jka, Jkb, Jk, K, k, M, N, S, s, U, Vel, Coa, Lan, Jk3Yta, Dib, Ge, Sc:1 or Lub antigens were not detected on lymphocytes, monocytes and PMN; (4) lymphocytes, monocytes and PMN all express I, i, P and P1 antigens. The absence of selected Rhesus, Duffy, Kell, Kidd and other antigens on PMN is at variance with some previous reports. Furthermore, the distribution of RBC antigens on lymphocytes and monocytes has not been previously characterized using immunofluorescence flow cytometry.

Heterogeneous distribution of antigens on human platelets demonstrated by fluorescence flow cytometry

We have used fluorescence flow cytometry to analyse cell-to-cell variability in the density of platelet ABH, Ii, Lewis, P, P1A1, Bak,a and HLA class I antigens. Human IgG and IgM antibodies were used in a two-stage assay with goat FITC-conjugated antihuman IgG (H&L) antibody as the label, followed by single cell analysis of 10 000 platelets per sample using a 256-channel fluorescence flow cytometer (Becton-Dickinson FACS Analyser). Computer analysis of fluorescence intensity histograms for mean and peak channel and coefficient of variation shows that the degree of heterogeneity in platelet antigen density varies with each particular blood group. The broad fluorescence distribution curves with oligosaccharide antigens (CVs: A = 53, B = 40, I = 44, Lea = 40, P = 40) indicate that these antigens possess a greater variability in the number of sites per cell compared to the more homogeneous distribution of P1,A1 BaK,a and HLA (CVs: P1A1 = 24, HLA = 30). These findings may partly account for the mechanism by which transfusion of ABO-incompatible platelets results in a biphasic survival curve, with a period of early rapid removal of those platelets with a high density of antigen sites, followed by a relatively normal survival curve for those platelets that possess only a few or no antigen sites. In contrast, P1A1 and HLA sites are less variable in number from one platelet to another in a given donor, and immune-mediated removal would be more likely to approximate a single exponential curve.

Absence of ABH antigens on neutrophils

Many investigators have concluded that polymorphonuclear leucocytes (PMN) express ABH antigens in parallel to red cells (RBC). We have examined human PMN for ABH antigens using human isoantibodies and mouse monoclonal antibodies with three highly sensitive and specific two-stage assay systems: fluorescence flow cytometry, immunofluorescence microscopy, and avidin-biotin immunoperoxidase microscopy. In all three assays the ABH antigens could not be detected on the surface of PMN. Previous reports alleging that ABH antigens occur on PMN probably represent false positive reactions due to inherent technical problems.

An antibody to human thymic Hassall's body epithelium recognizes a subset of blood group A antigens

TE-19, a mouse monoclonal antibody (mAb) against thymic Hassall's body epithelium, was investigated because of its cross-reactivity with human erythrocytes. Antibody TE-19 was found to react only with group A erythrocytes, though it reacted with Hassall's body epithelium from donors of all blood group phenotypes. TE-19 antibody bound preferentially to cells of subgroup A1, and reacted only weakly with A2 and Aint erythrocytes. Although it apparently bound to A antigen with both glycolipid and glycoprotein backbones, TE-19 antibody only reacted with A antigenic moieties containing long or branched chain structures. Analysis of TE-19 reactivity with erythrocyte membrane components using electroblotting and immunostaining techniques showed antibody reactivity with components migrating in the areas of band 4 X 5 and just ahead of the dye front, consistent with membrane glycolipid. It is possible that all human thymic Hassall's bodies contain epithelial cells which, irrespective of the donor's erythrocyte blood group, bear a carbohydrate antigen similar to A antigen.

Analysis of the expression of H, Lewis, X, Y and precursor blood group determinants in saliva and red cells using a panel of mouse monoclonal antibodies

Salivary glycoproteins from 33 normal individuals were analyzed with a panel of mouse monoclonal antibodies to H-1, H-2, Lea, Leb, X, Y and precursor blood group determinants. Samples from 19/33 individuals co-expressed Leb and Y-determinants (secretors) and 6/33 co-expressed Lea and X-determinants (non-secretors). Erythrocytes of these individuals were typed Le (a-b+) and Le (a + b-), respectively. In seven other salivas, only one specificity, either Lea, Leb, X or Y, was expressed and in one sample none of these determinants could be detected. Only one saliva sample expressed H-1 specificity and none expressed H-2 or type 1 precursor determinants. The absence of H-1 and H-2 structures in secretors and the resulting expression of difucosylated Leb and Y-structures is probably a tissue-specific trait of salivary gland secretions. The strict co-expression of Leb with Y and Lea with X supports the conclusion that only one 2-O-fucosyl-galactose transferase, which can fucosylate both type 1 and type 2 chains, exists in salivary glands. The finding that a number of individuals expressed neither X- nor Y-specificities was unexpected in view of previous work showing that the 3-O-fucosyl N-acetylglucosamine transferase involved in forming this structure is a ubiquitous enzyme. The individualistic expression of blood group phenotypes in tissues should be considered when the altered expression of blood groups in malignancy and other diseases is studied.

Lewis blood group antigens defined by monoclonal anti-colon carcinoma antibodies

Monoclonal antibodies directed against human cancer cells were prepared by the murine hybridoma technique. These antibodies detect Lewis blood group antigens as determined by indirect solid-phase radioimmunoassay, hapten inhibition studies, and chromatogram binding assay. One monoclonal antibody is specific for the Lea terminal carbohydrate of Gal beta 1----3Glc NAc(4----1 alpha Fuc) beta 1----3LacCer. Five monoclonal antibodies react with the Leb terminal carbohydrate sequence of Fuc alpha 1----2Gal beta 1----3GlcNAc(4----1 alpha Fuc) beta 1----3LacCer, and four of these antibodies are highly specific for this glycolipid and do not react with other similar di- and monofucosylated glycolipids. One of the anti-Leb antibodies cross-reacts with blood group H glycolipid and has binding properties similar to those of the previously described antibody NS-10-17 [M. Brockhaus, J. L. Magnani, M. Blaszczyk, Z. Steplewski, H. Koprowski, K.-A. Karlsson, G. Larson, and V. Ginsburg (1981) J. Biol. Chem. 256, 13223-13225]. Two antibodies react with both the Lea and Leb antigens, though both bind preferentially to Leb.

Two mouse hybridoma antibodies against human milk-fat globules recognise the I(Ma) antigenic determinant beta-D-Galp-(1 leads to 4)-beta-D-GlcpNAc-(1 leads to 6)

Two mouse hybridoma antibodies (LICR-LON-M39 and LICR-LON-M18) against the human-milk-fat globules were found to resemble human autoantibodies of anti-I type in their cold agglutinating property and their preferential reactions with erythrocytes of I- rather than i-type. From inhibition of binding assays with glycoproteins having known A, B, H, Lea, Leb, I, and i activities, and oligosaccharides of the Type 1 and Type 2 lacto-N-glycosyl series, it was established that these antibodies are directed at Type 2 structures, and that the I(Ma) determinant, beta-D-Galp-(1 leads to 4)-beta-D-GlcpNAc-(1 leads to 6), which is usually found on branched oligosaccharides, is the preferred sequence. The hybridoma antibodies as well as anti-I Ma were shown to react well with the beta-D-Galp-(1 leads to 4)-beta-D-GlcpNAc-(1 leads to 6)-D-Gal or -D-Man sequence. Studies of the reactions of these antibodies with glycolipids on thin-layer plates showed that the two hybridoma antibodies differ from anti-I Ma in reacting weakly with the unbranched i-type sequence beta-D-Galp-(1 leads to 4)-beta-D-GlcpNAc-(1 leads to 3)-beta-D-Galp-(1 leads to 4)-beta-D-GlcpNAc-(1 leads to 3)-beta-D-galp-(1 leads to 4) as found on lacto-N-norhexasylceramide. Furthermore, they differ from anti-I Ma but resemble anti-I Woj and Sti, and a hybridoma antibody 1B2 in their failure to react with their determinant in the presence of alpha-D-(1 leads to 3)-linked galactosyl groups. From their lack of reactions with blood-group-A and -H active glycoproteins, and their reactions with neuraminidase-treated erythrocytes, it was deduced that the determinants recognised by the two hybridoma antibodies are also masked in the presence of alpha-L-(1 leads to 2)-linked fucosyl groups and sialic acid.