Therapeutic antibody glycosylation impacts antigen recognition and immunogenicity.

In this study we show that glycosylation is relevant for immune recognition of therapeutic antibodies, and that defined glycan structures can modulate immunogenicity. Concerns regarding immunogenicity arise from the high heterogeneity in glycosylation that is difficult to control and can deviate from human glycosylation if produced in non-human cell lines. While non-human glycosylation is thought to cause hypersensitivity reactions and immunogenicity, less is known about effects of Fc-associated glycan structures on immune cell responses. We postulated that glycosylation influences antigen recognition and subsequently humoral responses to therapeutic antibodies by modulating 1) recognition and uptake by dendritic cells (DCs), and 2) antigen routing, processing and presentation. Here, we compared different glycosylation variants of the antibody rituximab (RTX) in in vitro assays using human DCs and T cells as well as in in vivo studies. We found that human DCs bind and internalize unmodified RTX stronger compared to its aglycosylated form suggesting that glycosylation mediates uptake after recognition by glycan-specific receptors. Furthermore, we show that DC-uptake of RTX increases or decreases if glycosylation is selectively modified to recognize activating (by mannosylation) or inhibitory lectin receptors (by sialylation). Moreover, glycosylation seems to influence antigen presentation by DCs because specific glycovariants tend to induce either stronger or weaker T cell activation. Finally, we demonstrate that antibody glycosylation impacts anti-drug antibody (ADA) responses to RTX in vivo. Hence, defined glycan structures can modulate immune recognition and alter ADA responses. Glyco-engineering may help to decrease clinical immunogenicity and ADA-associated adverse events such as hypersensitivity reactions.

Development of a flow-injection analysis (FIA) enzyme sensor for fructosyl amine monitoring.

An enzyme-sensor system with flow-injection analysis (FIA) has been developed for the detection of fructosyl amine compounds; the sensor utilizes fructosyl amine oxidase isolated from the marine yeast Pichia sp. N1-1 strain. With this FIA system 0.2 to 10 mmol L(-1) fructosyl valine can be determined. The sensor is approximately five times more sensitive to fructosyl valine, a model compound for glycated hemoglobin HbA1c, than to N(epsilon)-fructosyl lysine, a model compound for glycated albumin. This FIA system can also be used to detect fructosyl dipeptides. The operational stability of the sensor enabled more than 120 consecutive sample injections over a period of approximately 20 h.

Activation of cellulose membranes with 1,1'-carbonyldiimidazole or 1-cyano-4-dimethylaminopyridinium tetrafluoroborate as a basis for the development of immunosensors.

Methods for the activation of a cellulose dialysis membrane for immunosensor applications have been developed. For activation two reagents, 1,1'-carbonyldiimidazole (CDI) and 1-cyano-4-dimethylaminopyridinium tetrafluoroborate (CDAP), were compared with respect to the coupling efficiency for glucose oxidase (GOx) and 1,8-diamino-2,6-dioxaoctane. The maximum level of activation was 2.4 micromol cm(-2) for CDI and 0.2 micromol cm(-2) for CDAP activation. We observed 1.5 microg cm(-2) and 0.4 x 10(-4) U cm(-2) GOx with CDI-activated membranes whereas 1.7 microg cm(-2) and 7.2 x 10(-4) U cm(-2) GOx were observed with CDAP-activated membranes. With 1,8-diamino-2,6-dioxaoctane amino group densities of 0.165 and 0.09 micromol cm(-2) were observed via CDI and CDAP activation, respectively. An amino-modified membrane was used for coupling a ligand (pentapeptide) and an immunoenzymometric assay for hemoglobin A1c was carried out.