High-throughput carbohydrate microarray profiling of 27 antibodies demonstrates widespread specificity problems.

Progress toward understanding the biological roles of carbohydrates has been remarkably slow, and efforts to exploit this class of biopolymers as diagnostic and therapeutic targets have proven extremely challenging. Both basic and clinical research rely heavily on identifying and monitoring expression levels of carbohydrates. Over the last 30 years, the majority of expression information has been derived from antibody- and lectin-binding studies. Using a carbohydrate microarray containing 80 different glycans and glycoproteins, the specificities of 27 antiglycan antibodies were evaluated, including antibodies to histo-blood group A, B, and H antigens (81FR2.2, CLCP-19B, B389, 92FR-A2, B480, B460, B376, and B393), Lewis antigens (7LE, 15C02, 28, ZC-18C, 121SLE, CA199.02, PR.5C5, 2-25LE, BR55, T174, T218, F3, A70-C/C8, FR4A5, and K21), and other tumor-associated antigens (B389, 1A4, B1.1, and 5B5). In total, evaluation of over 2000 individual carbohydrate-protein interactions was carried out. More than half of the antibodies considered to be specific for their designated antigen were found to cross-react with other glycans. The cross-reactive glycans could be mistaken for the designated antigen in biopsy samples or other biological samples, leading to inaccurate conclusions.