Determination of modulation of ligand properties of synthetic complex-type biantennary N-glycans by introduction of bisecting GlcNAcin silico,in vitroandin vivo : N-glycans with bisecting GlcNAc as ligands

The beta 1-6 structure of N-linked oligosaccharides, formed by beta-1,6-N-acetylglucosaminyltransferase (GnT-V), is associated with metastatic potential. We established a highly metastatic subclone, B16-hm, from low metastatic B16-F1 murine melanoma cells. The gene encoding beta-1,4-N-acetylglucosaminyltransferase (GnT-III) was introduced into the B16-hm cells, and three clones that stably expressed high GnT-III activity were obtained. In these transfectants, the affinity to leukoagglutinating phytohemagglutinin was reduced, whereas the binding to erythroagglutinating phytohemagglutinin was increased, indicating that the level of beta 1-6 structure was decreased due to competition for substrate between intrinsic GnT-V and ectopically expressed GnT-III. Lung metastasis after intravenous injection of the transfectants into syngeneic and nude mice was significantly suppressed, suggesting that the decrease in beta 1-6 structure suppressed metastasis via a mechanism independent of the murine system. These transfectants also displayed decreased invasiveness into Matrigel and inhibited cell attachment to collagen and laminin. Cell growth was not affected. Our results demonstrate a causative role for beta 1-6 branches in invasion and cell attachment in the extravasation stage of metastasis.

The cell density-dependent growth inhibition of human SK-N-MC neuroblastoma cells is initiated by increased ganglioside sialidase activity leading to elevated cell surface presentation of ganglioside GM1, a ligand of galectin-1. We herein show that the extent of the cell surface expression of the galectin coincides with marked increases of the sialidase activity. Reverse transcriptase-polymerase chain reaction analysis excludes a regulation at the transcriptional level. Exposure of cells to purified galectin-1 reveals its carbohydrate-dependent activity to reduce cell proliferation. Assays to detect DNA fragmentation biochemically and cytometrically and to block caspases render it unlikely that galectin-1 acts as a classical proapoptotic factor on these cells. Because the chimeric galectin-3 shares binding sites and binding parameters with galectin-1 for these cells, we tested whether this galectin will elicit the same response as the homodimeric cross-linking galectin-1. Evidently, galectin-3 fails to affect cell growth by itself but interferes with galectin-1 upon coincubation. Its proteolytically truncated variant, the C-terminal lectin domain with impaired capacity to form aggregates when surface bound, has only weak binding properties. Thus, the way in which the galectin-1 interacts topologically with an apparently common set of ligands relative to galectin-3 is crucial for eliciting post-binding events. We conclude that galectin-1 is a probable effector in the sialidase-dependent growth control in this system. Moreover, the experiments with galectin-3 reveal functional divergence, most probably based on different topologies of presentation of homologous carbohydrate-binding sites.