Inhibition of interleukin-5 induced false positive anti-drug antibody responses against mepolizumab through the use of a competitive blocking antibody.

Mepolizumab, a humanized IgG1 monoclonal antibody that blocks native homodimeric interleukin-5 (IL-5) from binding to the IL-5 receptor, has recently been approved for treatment of severe eosinophilic asthma. Our initial immunogenicity assay method for phase I and II studies utilized a bridging electrochemiluminescence format with biotin and ruthenium-labelled mepolizumab linked by anti-drug antibodies (ADA). We discovered that IL-5 significantly increased in dosed subjects from a phase II study and that the increased IL-5 was in the form of a drug-bound complex. We demonstrated that the elevated drug-bound IL-5 produced false-positive response in the in vitro ADA assay, in which drug-bound IL-5 dissociated and then bridged mepolizumab conjugates to yield positive signal. To eliminate the IL-5 interference, we compared two strategies: a solid-phase immunodepletion of IL-5 and an in-solution IL-5 immunocompetition. We identified the best competitive antibody for each purpose. We found both methods demonstrated similar effectiveness in reducing the false positive signal in IL-5 spiked samples; however, the in-solution immunocompetition for IL-5 had fewer false positives in study samples. Additionally, the in-solution immunocompetition method was experimentally simpler to execute. We modified the ADA assay by adding a pre-treatment step with a mepolizumab competitive anti- IL-5 antibody. Using this new method, we retested clinical samples from two phase II studies (MEA112997 and MEA114092). The confirmed ADA positive incidence was reduced from 29% and 61% to 1% and 8% with the modified in-solution immune inhibition method. Target interference is a fairly common problem facing immunogenicity testing, and target-induced false positive cannot be distinguished from true ADA response by the commonly used drug competitive confirmation assay. The approach and method used here for resolving target interference in ADA detection will be useful for differentiating between a true ADA response and target induced false positive as well as similar challenges in other programs.

Copyright © 2016 The Authors. Published by Elsevier B.V. All rights reserved.