Evaluation of Bottom-Up and Middle-Up Liquid Chromatography Mass Spectrometry Methods for Forced Degradation Studies applied in Biosimilar Development

Abstract

The characterization of post-translational modifications (PTMs) is an important part of therapeutic monoclonal antibody (mAb) quality assurance. Modification by oxidation is one of the most commonly observed degradation pathways in mAbs. The present work utilizes two liquid chromatography-mass spectrometry (LC-MS) based methods for investigation of PTMs with a focus on oxidation. The two approaches are compared to understand to which extent the faster and less artefact-prone middle-up analysis can replace the elaborate bottom-up analysis. The reference medicinal products (RMPs) and biosimilar candidates of the two mAbs bevacizumab and infliximab were subjected to forced oxidation by H2O2 to investigate their susceptibility to oxidation. In a first step, middle-up analysis was established. PTM analysis of unstressed samples, including major glycoforms, demonstrated high similarity between the biosimilar candidates and the RMPs of bevacizumab and infliximab. Treatment with H2O2 generated different species of the subunits with different numbers of oxidations. The most abundant species caused by forced oxidation could be assigned to single oxidations of methionine residues (Met, M). Extensive oxidation of the Fc/2 could be attributed to M258, M364 and M434 in bevacizumab and to M255 and M431 in infliximab. Extensive oxidation was also observed in the light chain (LC) and in the Fd’ subunit of infliximab, but not in bevacizumab. By comparison of the most abundant species’ number of oxidations and the positions of Met between the two mAbs, the oxidations in infliximab were assigned to M55 in the LC and M18 in Fd’. However, less abundant species of Fc/2, LC and Fd’ with higher oxidation numbers could not be explained. Despite high similarity of unstressed samples, forced oxidation demonstrated a higher susceptibility of the bevacizumab biosimilar candidate compared with the RMP. In a second step, bottom-up analysis was established and used to confirm the exact location of oxidations. In addition to singly oxidized Met, singly oxidized tryptophan (Trp) and triply oxidized cysteine (Cys) residues were found. The middle-up assignments of the most abundant species to oxidations on Met residues were confirmed by the bottom-up results, demonstrating that those Met that had been assigned by middle-up were completely oxidized. Bottom-up analysis could explain the unexplained less abundant species of Fc/2, LC and Fd’ with higher oxidation numbers by the simultaneous occurrence of completely oxidized Met together with other Met that were only partially oxidized and partially oxidized Cys and Trp. In contrast to the middle-up results, no notable difference was observed in the oxidation susceptibility of the bevacizumab biosimilar candidate compared with the RMP. This discrepancy between middle-up and bottom-up results should be addressed in further studies. The bottom-up experiments were more time-consuming, sample-consuming, and expensive than the middle-up analyses. In conclusion, it is recommended that consistency of the results from the two approaches should be confirmed. If consistency can be established, the middle-up approach can be employed in forced degradation biosimilar studies, with the benefits of lower sample-consumption, faster analysis and lower expenses compared to bottom-up.