Tumor-specific responses in lymph nodes draining murine sarcomas are concentrated in cells expressing P-selectin binding sites.

Tumor-draining lymph node (TDLN) cells develop substantial antitumor activity after activation on immobilized alphaCD3 and culture in low-dose IL-2. This study found that the minor subset of TDLN T cells expressing binding sites for the adhesion receptor P-selectin (Plig(high) T cells) produced T lymphoblasts with the most tumor-specific IFN-gamma synthesis in vitro and antitumor activity following adoptive transfer in vivo. The Plig(high) T cells constituted <25% of the cells with the phenotype of recently activated cells including high levels of CD69, CD44, or CD25, and low levels of CD62L. The cultured Plig(high) TDLN were 10- to 20-fold more active against established pulmonary micrometastases than cultured unfractionated TDLN, and >30-fold more active than cultured TDLN cells depleted of the Plig(high) fraction before expansion (Plig(low) cells). Tumor-specific IFN-gamma synthesis in vitro paralleled the antitumor activities of the cultured fractions in vivo, implying that increased Tc1 and Th1 effector functions contributed to the tumor suppression. Neither nonspecific interaction with the P-selectin chimera used for sorting nor endogenous costimulatory activity in the Plig(high) fraction accounted for the marked increase in antitumor activities after culture. The cultured Plig(high) fraction contained a variety of potential effector cells; however, the CD8 and CD4 subsets of alphabeta T cells accounted for 95-97% of its antitumor activity. The authors propose that P-selectin sorting increased antitumor activities by concentrating Tc1 and Th1 pre-effector/effector cells before culture.

Lymphocyte-endothelial cell adhesive interactions in lung immunity: lessons from the murine response to particulate antigen.

The adhesive interaction between lymphocytes and lung endothelial cells presents an attractive arena for the development of novel therapeutic agents to modify pathologic pulmonary immune responses. The conceptual basis for choosing molecular targets to modulate this adhesive interaction derives, in large part, from results of murine experimental model systems of the pulmonary immune response. This article reviews one such model, the response of primed C57BL/6 mice to the particulate antigen sheep erythrocytes. Novel data are presented on the effect of a blocking anti-alpha(4) integrin monoclonal antibody on lung leukocyte and lymphocyte subset accumulation after intratracheal (IT) antigen challenge. Results from this model system have indicated that lymphocytes may use either the endothelial selectins or alpha(4) integrin as independent pathways to initiate recruitment into the lungs.

Alpha(1,3)-fucosyltransferase VII-dependent synthesis of P- and E-selectin ligands on cultured T lymphoblasts.

T lymphocytes up-regulate the synthesis of ligands for E- and P-selectin during proliferative responses in vivo and in vitro. Previous studies from our laboratories indicated that the alpha(1,3)-fucosyltransferase FucT-VII regulates the synthesis of E-selectin ligands and sialylated Lewis(x)-related epitopes (sLe(x)-related epitopes) in human T lymphoblasts. The current report shows that production of both P- and E-selectin ligands is FucT-VII dependent, but peak synthesis of each occurs at different levels of fucosyltransferase activity in intact cells. In brief, FucT-VII mRNA levels were higher in cultured T lymphoblasts expressing sLe(x)-related epitopes and both selectin ligands than in cells expressing P-selectin ligands alone. However, synthesis of the epitopes and both selectin ligands required the FucT-VII enzyme in transfected Molt-4 cells. In contrast, neither constitutive nor transfection-enhanced levels of the FucT-IV enzyme generated active P-selectin ligands in these lines. In addition, targeted deletion of the FucT-VII gene in mice markedly inhibited the synthesis of both P- and E-selectin ligands during blast transformation in vitro. Finally, the optimal synthesis of active P-selectin ligands occurred at lower level of FucT-VII activity than required for synthesis of equally active E-selectin ligands in both cultured T lymphoblasts and FucT-VII transfectants. Consequently, the FucT-VII enzyme is essential for the synthesis of both P- and E-selectin ligands by T lymphoblasts, and its activity determines whether P-selectin ligands are expressed alone or in conjunction with E-selectin ligands and sLe(x)-related epitopes on human T cells.

The role of valence on the high-affinity binding of Griffonia simplicifolia isolectins to type A human erythrocytes.

The Griffonia simplicifolia-I (GS-I) isolectins have been used to probe the effect of lectin valence on their high-affinity binding to human erythrocytes. These tetrameric lectins are composed of A and B subunits and constitute a series of five isolectins (A4, A3B, A2B2, AB3, B4). The A subunit is specific for alpha-D-GalNAc end groups and binds to the blood type A determinant GalNAcalpha1, as well as to terminal alpha-D-Gal groups found on type B cells. The B subunit is specific for alpha-D-Gal end groups, and binds very specifically to type B erythrocytes. This series of isolectins is tetravalent (A4), trivalent (A3B), divalent (A2B2), and monovalent (AB3) for type A erythrocytes; thus, this system provides the opportunity to examine the effect of lectin valency on the association constants of these GS-I isolectins binding to cells. Cell binding experiments carried out using 125I-labeled GS-I isolectins and type A human erythrocytes allowed us to demonstrate that (1) the association constant of the isolectin monovalent for alpha-D-GalNAc (AB3) is virtually identical to its association constant for the haptenic sugar methyl-N-acetyl-alpha-D-galactosaminide, reported previously, and (2) the association constant of the GS-I isolectins for human type A erythrocytes increases with increasing valency of the isolectin. These results indicate that the increased affinity displayed by the GS-I isolectins for human type A erythrocytes is dependent on their multivalency, and not on an extended binding site nor on nonspecific, or noncarbohydrate, interactions of the lectin with the cell surface. These findings should be of general relevance to understanding the high-affinity interactions observed between other multivalent proteins and multivalent ligands (e.g., cell surfaces).

An sLex-deficient variant of HL60 cells exhibits high levels of adhesion to vascular selectins: further evidence that HECA-452 and CSLEX1 monoclonal antibody epitopes are not essential for high avidity binding to vascular selectins.

Selectins are carbohydrate-binding cell adhesion molecules that play a key role in the initiation of inflammatory responses. Several studies have suggested that the sialylated, fucosylated tetrasaccharide sialyl Lewis X (sLex) is an important component of leukocyte ligands for E- and P-selectin. We have identified a stable variant of the HL60 cell line, HL60var, which displays a nearly complete absence of staining with several mAb directed against sLex and/or sLex-related structures. HL60var also exhibits a concomitant increase in reactivity with mAb directed against the unsialylated Lewis X (Lex/CD15) structure. Despite this sLex deficiency, HL60var binds well to both E- and P-selectin. No significant differences in expression of alpha1,3-fucosyltransferases, C2GnT (Core2 transferase), or P-selectin glycoprotein ligand-1 between HL60var and typical sLex(high) HL60 cells were detected. Although the precise molecular basis for the sLex(-/low) phenotype of HL60var remains uncertain, flow cytometric analysis with the sialic acid-specific Limax flavus lectin revealed a sharp reduction in HL60var surface sialylation. Thus, the loss in mAb reactivity may result from a loss of sialic acid residues from the mAb carbohydrate epitope. However, binding of HL60var to E- and P-selectin remains sensitive to neuraminidase treatment. Taken together, these data indicate that high levels of surface sLex and/or related epitopes are not essential for interactions with vascular selectins, implying that as yet unidentified sialylated, fucosylated structures serve as physiologically relevant ligands for E- and P-selectin.

The fucosyltransferase FucT-VII regulates E-selectin ligand synthesis in human T cells.

Selectin-ligands on T cells contribute to the recruitment of circulating cells into chronic inflammatory lesions in the skin and elsewhere. This report provides the first evidence that a single fucosyltransferase, termed FucT-VII, controls the synthesis of E-selectin ligands in human T-lymphoblasts. The FucT-IV transferase (the ELFT enzyme), in contrast constructs lower avidity E-selectin ligands and requires enzyme levels found only in myeloid cells. Treatment of Jurkat cells with phorbol myristate acetate increased the expression of sialylated Lewis(x)-related sLe(x)related epitopes and induced the synthesis of E-selectin ligands functional at physiologic levels of linear shear-stress. Northern analysis revealed a parallel increase in the steady-state levels FucT-VII mRNA, but there were no increases in the two other leukocyte-associated fucosyltransferases (FucT-IV and VI). The stable transfection of the FucT-VII gene into Jurkat cells induced high levels of the sLe(x)-related epitopes and the synthesis of E-selectin ligands which equal or exceeded the avidity of those on circulating lymphocytes. The growth of T-lymphoblasts under conditions which induced expression of the sLe(x,a) epitopes increased the level of FucT-VII mRNA, the synthesis of sialylated-Lewis(x) structures by cell-free extracts and the synthesis of E-selectin ligands equal in avidity to those on FucT-VII transfectants. In contrast, neither the mRNA levels nor activities of the FucT-IV and VI enzymes increased in association with E-selectin ligand synthesis in T-lymphoblasts. Myeloid cell lines, unlike lymphoblasts, expressed high levels of both the FucT-VII and IV enzymes in conjunction with E-selectin ligands raising the possibility that both enzymes contributed to ligand synthesis. FucT-IV transfected Jurkat cells synthesized low avidity ligands for E-selectin but only in association with CDw65 (VIM-2) carbohydrate epitope. Only blood neutrophils and myeloid cell lines expressed this epitope at the levels associated with E-ligand synthesis in the transfectants. In contrast, native Jurkat cells, blood monocytes, blood lymphocytes, and cultured T-lymphoblasts expressed low levels or none. We conclude that FucT-VII is a principal regulator of E-selectin ligand synthesis in human T-lymphoblasts while both FucT-VII and FucT-IV may direct ligand synthesis in some myeloid cells.

Differential expression of cell surface sialoglycoconjugates on wild-type and cultured Ehrlich tumor cells as revealed by quantitative lectin-gold ultrastructural cytochemistry.

Three variants of the classical Ehrlich ascites tumor (EAT) cell have been studied by quantitative, sialic acid-specific, lectin-gold ultrastructural cytochemistry. Electron microscopic examination revealed pronounced differences in the surface morphology of the three cell variants. The wild-type Ehrlich cells (EAT-wt), grown in the peritoneal cavity of mice, exhibited a smooth surface profile. A variant form selected for growth as monolayer on basement membrane (EAT-c) showed a complex surface profile with numerous microvilli. The third variant (EAT-c/m), the cultured cells reinoculated into mice and passaged 20-25 times as ascites, presented a smooth surface profile similar to the EAT-wt cells. Quantitative single as well as double lectin-gold labeling revealed significant differences in the nature of cell surface sialoglycoproteins. The most significant finding was the presence of cell surface Neu5Ac alpha 2-6Gal residues as detected with the Sambucus nigra lectin on EAT-c and EAT-c/m cells, whereas EAT-wt cells contained little or none of such carbohydrate sequences. On the contrary, labeling by Maackia amurensis lectin, which recognizes the Neu5Ac alpha 2-3Gal beta 1-4GlcNAc sequence, was intense on all three Ehrlich cell variants; it was 20-60 times greater than alpha-2,6-linked sialic acid-containing glycoconjugates. Specific cell surface lectin binding combined with morphologic study appears to have identified a small subpopulation of cells within the ascites tumor that are capable of attaching to and growing on a basement membrane.

Wild-type and cultured Ehrlich ascites tumour cells differ in tumorigenicity, lectin binding patterns and binding to basement membranes.

Three different variants of the Ehrlich ascites tumour (EAT) cell were derived and the lectin surface reactivities, as well as the malignant characteristics of each variant, were studied. Wild-type cells (EAT-wt) were selected for growth on basement membranes and tissue culture plastic to give EAT-c cells. The EAT-c were passaged in mice by i.p. injection, giving rise to a third variant (EAT-c/m). Each of these three cell variants was characterized for: (i) specific lectin agglutinability patterns; (ii) the ability to produce ascites tumours in mice; (iii) the ability to produce solid tumours; and (iv) the attachment to and growth on basement membranes and purified extracellular matrix molecules. Analysis of the total protein and carbohydrate content of each cell line showed that there was an increase in the glycosylation of the EAT-c cells compared to EAT-wt cells. After repeated passage of the EAT-c/m cells in mice, the glycosylation level of the EAT-c/m cells returned to that of the EAT-wt cell line. In addition, the EAT-c cells displayed an increase in the number of terminal non-reducing sugars which could indicate either an increase degree of branching or the presence of additional N- and/or O-linked oligosaccharide chains of the cellular glycoproteins. This phenotype was retained by the EAT-c/m cells which had been passaged repeatedly in mice. The most significant increase was in the content of sialic acid-containing glycoproteins found in the EAT-c cells. The sialic acid-binding lectin Maackia amurensis leukoagglutinin (MAL) agglutinated all three EAT cell variants, while the sialic acid-binding Sambucus nigra (elderberry bark) lectin (SNA) agglutinated only the EAT-c and early-passage EAT-c/m cells. These findings indicate the presence of alpha 2,3-linked sialic acid on all three variants, but only the cultured cells and early-passage EAT-c/m cells possess the Neu5Ac alpha 2,6 linkage. The EAT-c cells attached avidly to wells coated with either laminin or fibronectin, as well as extracellular matrix produced by cultured bovine endothelial cells, but the EAT-wt and EAT-c/m cells did not. Paradoxically, the EAT-c cells were incapable of producing solid tumours when injected into a basement membrane-rich skeletal muscle bed, whereas the EAT-wt and EAT-c/m cells produced rapidly growing tumours when injected into the same environment. Lectin agglutination patterns established that ascitic tumour cells within the peritoneal cavity were derived from injected EAT-c cells.

Binding determinants of the sialic acid-specific lectin from the slug Limax flavus.

The specific structural features of 24 N-acetylneuraminic acid derivatives required for the high affinity interaction of sialoglycoconjugates with the sialic acid-specific lectin from the slug Limax flavus were studied by hapten inhibition of precipitation. These results provide insight regarding the structure of the binding pocket for N-acetylneuraminic acid that exists in L. flavus agglutinin (LFA). The alpha-anomer of sialic acid is a very important factor in binding to the slug lectin. The carboxylic acid group makes only a moderate contribution to binding, since modifications of the carboxylic group decrease binding approximately 5-fold. Modification or removal of the hydroxyl group on carbon 4 does not affect binding. However, when the C4 epimer was tested, there was a dramatic decrease in binding, suggesting that whereas the equatorial hydroxyl at C4 does not contribute to binding, the introduction of an axial hydroxyl group at C4 sterically hinders the binding interaction. The substituent on the 5-amino group occupies an important role in binding of Neu5Ac to LFA as well. When the acetyl is modified by the addition of a hydroxyl group to yield the N-glycolyl derivative, we observed a 20-fold decrease, while the removal of the methyl to form the N-formyl derivative resulted in a 50-fold decrease. The 5-amino derivative was the poorest inhibitor of all compounds examined, indicating a critical role for the N-acetyl group in high affinity binding to LFA. The glyceryl tail also appears to be critical for binding inasmuch as acetylation of the C9 hydroxyl group or periodate cleavage of carbons 8 and 9 resulted in a 20- to 50-fold decrease in binding. The equilibrium constant (K(a)) for binding of Neu5Ac to LFA is 3.8 x 10(4) M-1, with a single binding site (n = 0.85) per monomer.

Monoclonal antibodies that recognize the trisaccharide epitope Gal alpha 1-3Gal beta 1-4GlcNAc present on Ehrlich tumor cell membrane glycoproteins.

Monoclonal antibodies were prepared against the trisaccharide Gal alpha 1-3Gal beta 1-4GlcNAc, a sequence which occurs on the surface of Ehrlich ascites tumor cells as well as in thyroglobulin, laminin and a variety of other proteins. This was accomplished by immunizing BALB/c mice with the fraction of Ehrlich cell membrane glycoproteins obtained by affinity chromatography on a Griffonia simplicifolia I (GS I) column which selectively binds alpha-D-galactosyl-terminated structures. Detection of Gal alpha 1-3Gal beta 1-4GlcNAc-specific antibodies was accomplished by employing glycoproteins containing the trisaccharide sequence; fusion with spleen cells from an immunized mouse was accomplished in the presence of polyethylene glycol (PEG1500). An enzyme-linked immunosorbent assay (ELISA) system was used to identify two clones (2.10G and 6.8E), which recognized the desired trisaccharide conjugate. These clones also recognized a thyroglobulin fraction isolated by GS I affinity chromatography and murine laminin, both of which possess the Gal alpha 1-3Gal beta 1-4GlcNAc sequence. Inhibition of antibody-trisaccharide reactivity, examined employing an ELISA assay, revealed that two trisaccharides, Gal alpha 1-3Gal beta 1-4GlcNAc/Glc, were the best inhibitory haptens; Gal beta 1-4GlcNAc (LacNAc), Gal alpha 1-3Gal and Gal beta 1-4Glc (lactose) were poor inhibitors. Indirect immunofluorescence staining of unfixed Ehrlich cells using the monoclonal antibody at 4 degrees C revealed fluorescence over the entire cell surface. Indirect immunogold labeling of semithin and ultrathin sections of aldehyde fixed and Lowicryl K4M-embedded Ehrlich cells resulted in specific labeling of the cell surface and internal structure.(ABSTRACT TRUNCATED AT 250 WORDS)

Carbohydrate-binding protein 35. II. Analysis of the interaction of the recombinant polypeptide with saccharides.

The carbohydrate binding specificity of recombinant carbohydrate-binding protein 35 (rCBP35) has been investigated by quantitative precipitation using a series of glycoproteins and carbohydrate-protein conjugates and by inhibition of precipitation using well defined carbohydrate haptens. Synthetic glycoconjugates and glycoproteins containing terminal nonreducing galactosyl units in beta-linkage were capable of forming a precipitate with rCBP35. If the glycoprotein or glycoconjugate contained terminal Neu5Ac, or galactose in alpha-linkage, precipitate formation was not observed. We also found that murine laminin, which contains polylactosamine structures, reacted more strongly than did bovine fetuin. Using carbohydrate-bovine serum albumin (BSA) glycoconjugates, we found that the tetrasaccharide Gal beta 1, 4GlcNAc beta 1, 3Gal beta 1,4-GlcNAc-BSA reacted more strongly than the disaccharide Gal beta 1, 4GlcNAc-BSA conjugate, suggesting that the binding site accommodates carbohydrate ligands greater in size than a disaccharide. Equilibrium dialysis experiments using [3H]lactose showed that rCBP35 binds 1 mol (n = 0.84) of lactose/30,000 g atoms of protein, with an affinity constant of 2.07 x 10(4) M-1. The binding site on the polypeptide appears to contain four subsites that recognize the sequence Gal beta 1,4GlcNAc beta 1, Gal beta 1,X-. All disaccharides tested that contain a nonreducing beta-galactosyl unit behaved as inhibitors of precipitation at approximately the same concentration, suggesting that the reducing position of the tetrasaccharide does not play an important role in the specific binding to the fourth subsite. The reducing sugar may serve to hold the saccharide in a tunnel like binding pocket since methyl-beta-D-galactoside itself is an extremely poor inhibitor.

Characterization of the carbohydrate binding specificity of the leukoagglutinating lectin from Maackia amurensis. Comparison with other sialic acid-specific lectins.

The sialic acid-specific leukoagglutinating lectin from the seeds of Maackia amurensis (MAL) has been studied by the techniques of quantitative precipitin formation, hapten inhibition of precipitation, hapten inhibition using an enzyme-linked immunosorbent assay, and lectin affinity chromatography. The ability of the immobilized lectin to fractionate oligosaccharides based on their content of sialic acid has also been investigated. Our results indicate that MAL reacts with greatest affinity with the trisaccharide sequence Neu5Ac/Gc alpha 2,3Gal beta 1,4GlcNAc/Glc. The lectin requires three intact sugar units for binding and does not interact when the beta 1,4-linkage is replaced by a beta 1,3-linkage nor when the "reducing sugar" of the trisaccharide is reduced. Results from enzyme-linked immunosorbent assays show that an N-acetyllactosamine repeating sequence is not required; however, the N-acetyllactosamine repeating sequence does appear to enhance the binding of MAL to a series of glycolipids. In addition, the sialic acid may be substituted with either N-acetyl or N-glycolyl groups without reduction in binding. The C-8 and C-9 hydroxyl groups of sialic acid do not play a role in binding as shown by the strong reaction of periodate-treated glycoproteins. Comparison of the specificity of the three sialic acid-binding lectins indicates that Limax flavus agglutinin binds to Neu5Ac in any linkage and in any position in a glycoconjugate, Sambucus nigra lectin requires a disaccharide of the structure Neu5Ac alpha 2,6Gal/GalNAc, and MAL has a binding site complimentary to the trisaccharide Neu5Ac alpha 2,3Gal beta 1,4GlcNAc/Glc, to which sialic acid contributes less to the total binding affinity than for either S. nigra lectin or L. flavus agglutinin.

Structure of the major concanavalin A reactive oligosaccharides of the extracellular matrix component laminin.

Laminin, a high molecular weight (1,000,000) glycoprotein component of basement membranes, was isolated from the EHS murine tumor as a noncovalent complex with entactin by lectin affinity chromatography using the alpha-D-galactosyl binding lectin Griffonia simplicifolia I (GS I). Entactin was removed from this complex by passage over Sephacryl S-1000 in the presence of SDS. Compositional analysis showed that the affinity-purified laminin contained 25-30% carbohydrate by weight. Methylation analysis revealed that the oligosaccharides of laminin contained bi- and triantennary chains, the blood group I structure, and repeating sequences of 3Gal beta 1,4GlcNAc beta 1 units. Free oligosaccharides were derived from the asparagine-linked glycans of affinity-purified laminin by hydrazinolysis, re-N-acetylation, and reduction with NaB3H4. When fractionated by affinity chromatography on concanavalin A (Con A)-Sepharose, 80% of the oligosaccharides passed through the column unretarded and a single peak corresponding to 20% of the oligosaccharides was adsorbed and specifically eluted with a linear gradient of 0-30 mM methyl alpha-D-glucopyranoside. Further fractionation of the Con A reactive oligosaccharides on GS I-Sepharose demonstrated that 70% of these oligosaccharides possess at least one terminal nonreducing alpha-D-galactopyranosyl unit. The Con A reactive oligosaccharides were subjected to sequential digestion with endo- and exoglycosidases, and the reaction products were analyzed by gel filtration chromatography on a column of Bio-Gel P4. We thereby obtained evidence for a variety of structures not previously reported to exist on murine laminin including hybrid biantennary complex and biantennary complex structures containing poly(lactosaminyl) repeating units. The poly(lactosaminyl) units occur either on one or on both branches of the biantennary chains, as well as in more highly branched blood group I poly(lactosamine) structures. All sialic acid is present as N-acetylneuraminic acid linked alpha 2,3 to galactose.

Characterization of laminin-stimulated adherence and motility in tumor cells.

Laminin was shown to stimulate attachment, spreading and motility in a line of murine tumor cells. The stimulated cells began to attach and spread within 1 h and remained attached and spread throughout a 48-hour observation period. Pretreatment of the cells with laminin for 24 h beforehand did not interfere with their ability to bind laminin in a second treatment. Laminin-stimulated attachment and spreading occurred in the absence of extracellular Ca2+ or Mg2+, but not in the absence of both divalent cations. A number of cell function inhibitors blocked the adherence response, but were only partially effective even at high concentrations. These characteristics of laminin-stimulated adherence are different from those of the adherence response stimulated by peptide chemotactic factors and phorbol esters.

Attachment, spreading and growth in vitro of highly malignant and low malignant murine fibrosarcoma cells.

Highly malignant cell lines and low-malignant cell lines isolated from three different methylcholanthrene-induced murine fibrosarcomas were examined for their ability to attach to plastic dishes and collagen-coated dishes under serum-free conditions and in the presence of serum. Most of the cells from the three highly malignant lines attached and spread under all conditions. By 72 h, there was a significant increase in the number of cells indicating that at least some of the cells had undergone division (even in the absence of serum). In contrast, fewer of the cells from the three low-malignant lines attached and spread on the plastic or collagen substrates in the absence of serum or in the presence of 0.1 per cent serum. However, when 15 micrograms laminin per dish was added along with the low-malignant cells, they then attached and spread on the plastic and collagen-coated dishes. Previous studies have indicated that the highly malignant lines express cell surface antigens that cross-react with laminin while the low-malignant cell lines do not. We speculate that the differences between the high- and low-malignant cells in the expression of cell surface laminin-like antigens contribute to the dissimilarities in attachment and spreading capacity. These differences may also contribute to the dissimilarity between these cells in malignant potential.