Multidimensional glycan arrays for enhanced antibody profiling

Abstract

Carbohydrate-binding antibodies play a critical role in basic and clinical research. Monoclonal antibodies that bind glycans are used to measure carbohydrate expression, and serum antibodies to glycans can be important elements of the immune response to pathogens and vaccines. Carbohydrate antigen arrays, or glycan arrays, have emerged as powerful tools for the high-throughput analysis of carbohydrate-protein interactions. Our group has focused on the development and application of neoglycoprotein arrays, a unique array format wherein carbohydrates are covalently attached to a carrier protein prior to immobilization on the surface. The neoglycoprotein format permits variations of glycan structure, glycan density, and neoglycoprotein density on a single array. The focus of this study was on the effects of neoglycoprotein density on antibody binding. First, we evaluated binding of five monoclonal antibodies (81FR2.2, HE-195, HE-193, B480, and Z2A) to the blood group A antigen and found that neoglycoprotein density had a substantial effect on recognition. Next, we profiled serum antibodies in 15 healthy individuals and showed that inclusion of multiple neoglycoprotein densities helps distinguish different subpopulations of antibodies. Finally, we evaluated immune responses induced by a prostate cancer vaccine and showed that variations in neoglycoprotein density enable one to detect antibody responses that could not be detected otherwise. Neoglycoprotein density is a useful element of diversity for evaluating antibody recognition and, when combined with variations in glycan structure and glycan density, provides multidimensional glycan arrays with enhanced performance for monoclonal antibody development, biomarker discovery, and vaccine optimization.

Figure 1

Depiction of High and Low Neoglycoprotein Density on the Array Surface. Neoglycoproteins are combined with varying amounts of unmodified bovine serum albumin (BSA) and then printed on the array surface. A) The highest neoglycoprotein density contains no added BSA. B) As the proportion of BSA increases, the neoglycoproteins are spaced farther apart on the surface.

Figure 2

Structures of Selected Glycans.

Figure 3

Serum Antibody Binding to Selected Glycans. Serum antibodies were profiled in 15 healthy subjects (see Experimental for details). A) IgM signals to the glycopeptide, AcV-Tn(Thr)-S-G-19 (average of 19 glycopeptides per molecule of BSA) at high neoglycoprotein density (top) and low (bottom). B) IgG signals to the P1 antigen at high neoglycoprotein density (top) and low (bottom). Error bars represent the standard deviations over 4 spots on 2 different arrays.

Figure 4

Principal Component Analysis (PCA) of Sera from Healthy Subjects. PCA provides a graphical overview of similarities and differences among serum antibody profiles of 15 healthy subjects. Analyses of IgG antibodies on high and low density antigen arrays are shown in panels A and B, respectively. Similarly, binding of IgM antibodies was also measured on high (C) and low (D) density antigen arrays. Graphs were generated by projecting array data onto the top three principal components (PC) that account for the most variation between samples for each experiment. Percentage values above each graph indicate the amount of variation between samples represented by the top three PCs.

Figure 5

IgG Responses to the Forssman Disaccharide. Antibody levels were measured prior to and 3 months after vaccination with a prostate cancer vaccine. For the 6 patients, changes were graphed for a) the highest neoglycoprotein density, and b) all of the neoglycoprotein densities for Forssman Di – 04. Error bars represent the standard deviations over 4 spots on 2 different arrays.