Abstract
rVWF expressed in Chinese hamster ovary (CHO) cells can be hardly distinguished from pdVWF. This holds true for the primary amino acid sequence and for posttranslational modifications including N- and O-linked glycans. However, because CHO cells lack a a2,6-sialyltransferase, the enzyme attaching N-acetylneuraminic acid (Neu5Ac) in a a2,6- linkage to the galactose residues of complex N-glycans, Neu5Ac will be found in CHO-cell derived glycoproteins only in a2,3-linkage. We used this minute difference in glycosylation to develop a lectin-based immunoassay that allows the selective measurement of human rVWF in human plasma based on the lack of a2,6-bound Neu5Ac on rVWF. At equal VWF:Ag concentrations, mixtures of rVWF and pdVWF will show lower ratios of a2,6-bound Neu5Ac to VWF:Ag than those in pdVWF alone. The difference in ratio is directly related to the concentration of rVWF in the mixture.
We used the Sambucs nigra agglutinin (SNA) to measure VWF-bound a2,6-linked Neu5Ac. This lectin binds to Neu5Ac on N-glycans only when Neu5Ac is a2,6-linked to galactose. rVWF/pdVWF was captured from the sample by a plate-immobilized polyclonal rabbit anti-VWF antibody. a2,6-bound Neu5Ac was then specifically detected using biotinylated SNA and streptavidin peroxidase. The assay setup resulted in a linear dose-response curve ranging from 0.2 to 2.9 mU pdVWF:Ag/mL in human plasma. In contrast, rVWF showed no binding, even when tested at concentrations about 100-times higher. Subjecting plate-adsorbed pdVWF to enzymatic desialylation with neuraminidase completely abrogated the binding of SNA. The linkage specificity of the SNA binding was shown in an inhibition study using the trisaccharides 3′-sialyllactose and 6′-sialyllactose as hapten sugars. As expected only 6′-sialyllactose proved able to inhibit the binding of SNA to plate-immobilized pdVWF. This result confirmed the linkage specificity of the SNA binding essential for the discrimination of plasma-derived and rVWF. To demonstrate the feasibility of this approach we spiked a sample of a normal human plasma pool with rVWF in several concentrations ranging from 0 to 1.5 IU/ml. In these samples the VWF:Ag concentrations and the levels of VWF-bound a2,6-Neu5Ac with the SNA binding assay were measured. Then, we calculated the ratio of SNA binding to VWF:Ag on the assumption that all VWF was pdVWF, i.e. contained only a2,6-linked Neu5Ac. The difference between the expected and the actually found ratios of SNA binding to VWF:Ag correlated (R2=0.99) with the amount of rVWF present in the mixtures. Thus, we obtained a calibration curve that allowed the specific measurement of rVWF in a range of 0.2 to 1.5 IU/ml in the presence of 1 IU/ml pdVWF. Both assays, the VWF:Ag ELISA and the SNA-based immunoassay for the measurement of VWF-bound a2,6-bound Neu5Ac, showed acceptable precision.
The data show that rVWF in a human plasma sample can readily be differentiated from endogenous pdVWF based on minute, linkage-specific differences of the N-glycan structures. This type of lectin-based immunoassay could be useful in the clinical setting to determine the circulating concentration of therapeutic CHO-cell-derived rVWF and to specifically measure its concentration in the presence of the endogenous glycoprotein.
Mathematical model of the multi-amino acid multi-transporter system predicts uptake flux in CHO cells
