Detection of antibodies to erythropoietin-based drugs: is it possible to create the universal test system?

Abstract

Our aim was to compare different immobilized erythropoietin (EPO) preparations for their ability to detect anti-EPO IgG antibodies in blood sera of EPO-treated patients with ELISA technique. 294 serum samples of the patients treated with erythropoietin were analyzed. 127 serum samples of patients who did not receive recombinant human EPO (rhEPO) were studied for comparative analysis. ELISA assay was performed, and different rhEPO drugs were immobilized on the anti-EPO monoclonal antibody-coated plates. Horseradish peroxidase-conjugated mouse monoclonal antibodies to human IgG, IgG1, and IgG4 was used for detection. The following drugs were studied: recombinant human erythropoietin rhEPO-beta (Shandong Kexing Bioproducts), European standard of erythropoietin BRP 3, commercial drugs Aranesp (Amgen Europe B.V.), Mircera (F. Hoffmann-La Roche Ltd.), Eprex (Johnson & Johnson LLC), Eralfon (Pharmaceutical Company Sotex). The sensitivity of the method was expressed as a positivity index (IP). IP calculated as the ratio of OD from tested sera to OD at the cut-off levels. The latter was assumed as a mean OD±SD for serum samples from EPO-naive patients. The results were evaluated as positive with IP > 1.1, and negative at IP < 0.9. Results in the range of 0.9 ≤ IP < 1.1 were considered as unidentified. Among the 294 samples, 32 specimens were evaluated as positive or unidentified for total IgG anti-EPO antibodies. The unidentified samples were detected in 1.0-1.7% of all cases. IgG1 subclass antibodies were found in 50-56.3% of patients and IgG4 subclass antibodies, in 43.850% of the patients. Mann—Whitney test showed a significant difference between the test samples compared to control group for all the ELISA modifications (p = 0.001). The Kruskal—Wallis test did not show significant differences between the IP results obtained with any of five immobilized EPO drugs (p = 0.05). The correlation quotient of IP was in the range of 0.99-0.96 for total IgG and > 0.98 for two subclasses of antibodies. Linear regression coefficients were close to one, thus indicating absence of significant differences in the sensitivity of the compared methods. This study indicate the opportunity of using the similar test systems to determine anti-EPO antibodies in the patients treated with various rhEPO drugs. Therefore, it is possible to develop a universal commercial test system to this purpose.