Evaluation of a peptide ELISA for the detection of rituximab in serum

Pharmacokinetic studies of therapeutic monoclonal antibodies necessitate the measurement of their biologically active fraction. The aim of this work was to develop an enzyme-linked immunosorbent assay (ELISA) for rituximab, a chimeric anti-CD20 monoclonal antibody, based on its binding to a 20-mer peptide (P20) derived from the extracellular loop of human CD20 (residues 165-184). Derivatives of P20 were prepared by conjugation to bovine serum albumin (BSA-P20ACM) or biotin (Biot-P20ACM). Interactions of P20 and its derived peptides with rituximab were analyzed by surface plasmon resonance (SPR) and by ELISA. A monoclonal anti-idiotype antibody (MB2A4) was used as the reference in each case. SPR analysis showed that P20 (conjugated or unconjugated) had a lower affinity for rituximab than MB2A4. ELISA methods based on P20 or MB2A4 were both highly accurate and reproducible for rituximab measurement in spiked samples, but the MB2A4-based assay had a lower limit of quantification. Interestingly, discrepant results were obtained with the two ELISA methods when analyzing pharmacokinetic samples, with the rituximab concentrations obtained with the MB2A4-based method being systematically higher than those determined by the P20-based method. Possible interference of circulating CD20 with the P20-based method was supported by competition experiments. Rituximab aggregation in the bloodstream may also account for the bias observed in samples from healthy mice. The P20-based ELISA is far less sensitive than the MB2A-based ELISA, thus limiting its utility for pharmacokinetic studies. However, the discrepancy observed between two different approaches for rituximab measurement indicates that data from different studies should be interpreted with care.