Pig-to-human xenotransplantation of islet cells or of vascularized organs would offer a welcome treatment alternative for the ever-increasing number of patients with end-stage organ failure who are waiting for a suitable allograph. The main hurdle are preexisting antibodies, most of which are specific for 'Linear-B', carbohydrate epitopes terminated by the unbranched Gal-alpha(1,3)Gal disaccharide. These antibodies are responsible for the 'hyper-acute rejection' of the xenograft by complement mediated hemorrhage. For depletion of such antibodies we have developed an artificial injectable antigen, a glycopolymer (GAS914) with a charge neutral poly-lysine backbone (degree of polymerization n = 1000) and 25% of its side chains coupled to Linear-B-trisaccharide. With an average molecular weight of 400 to 500 kD, presenting 250 trisaccharide epitopes per molecule, this multivalent array binds anti-alphaGal antibodies with at least three orders of magnitude higher avidity on a per-saccharide basis than the monomeric epitope. In vivo experiments with non-human primates documented that rather low doses--1 to 5 mg/kg of GAS914 injected i.v.--efficiently reduce the load of anti-Linear-B antibodies quickly by at least 80%. This treatment can be repeated without any sensitization to GAS914. Interestingly, although the antibody levels start raising 12 h after injection, they do not reach pretreatment levels. The polymer is degraded and excreted within hours, with a minute fraction remaining in lymphoid tissue of anti-alphaGal producing animals only, probably binding to and inhibiting antibody-producing B-cells. The results of pig-to-non-human primate xenotransplantations established GAS914 as a relevant therapeutic option for pig-to-human transplantations as well. The synthesis of GAS914 was successfully scaled up to kg amounts needed for first clinical studies. Key was the use of galactosyl transferases and UDP-galactose for the synthesis of the trisaccharide.