Category: Formulation and Quality
Purpose: The patent expiration of many top-selling biologics has fueled significant interest in developing biosimilars. By definition, biosimilars are highly similar, but not identical, to their respective originator product, and extensive structural characterization is imperative for any successful biosimilar development program1. Selecting the appropriate analytical methods to best detect and compare small structural differences between a reference product and its proposed biosimilar presents a major challenge to companies and regulatory agencies1. One strategy for making existing structural disparities more easily detected is to subject both products to forced degradation (via temperature, shaking, etc.) conditions2,3. In this study, we selected 3 FDA-approved monoclonal antibody drugs and their corresponding biosimilars to subject to in-depth structural characterization using a wide range of state-of-the-art bioanalytical tools both before and after exposure to forced degradation conditions. The aim of this to work is to highlight suitable techniques for comparing biologics and to provide a framework for assessing biosimilarity with regards to structure.
Methods: Antibodies were stressed by incubation at 40ºC with 250 rpm shaking for 4 weeks. Circular dichroism spectra were collected using a Jasco J-815 CD spectrometer. Intrinsic fluorescence was measured using a SpectraMax 3 (Molecular devices) with 280nm excitation and emission at 280-450nm. Ion-mobility mass spectrometry and collision-induced unfolding were performed with a quadrupole-ion mobility-time-of-flight mass spectrometer (Synapt G2 HDMS, Waters). Intact mass spectrometry was performed using a Q Exactive mass spectrometer (ThermoFisher) following separation on a C4 column (X-Bridge BEH C4 2.1 x 50 mm) (Waters). CEX was performed using an Alliance HPLC (Waters) equipped with a PDA detector. SEC was carried out on an Alliance HPLC (Waters) with a TSKgel SuperSW mAb HR column (Tosoh). SDS-PAGE was conducted under reducing and non-reducing conditions using pre-cast NuPAGE® 3-8% tris-acetate gel (Invitrogen). Electrophoresis was carried out at 150V for 1 hour with XCell SureLock Mini-Cell (Invitrogen), and gels were analyzed with Fluorchem M (ProteinSimple).
Results: Circular dichroism and intrinsic fluorescence spectra were nearly identical for each originator-biosimilar pair over the course of the 4-week incubation, indicating highly similar secondary and tertiary structures. Prior to incubation, only the rituximab pair showed differences in collision induced unfolding as indicated by CIU50 values (the voltage at which 50% of the protein has transitioned to an unfolded state). Following incubation, several significant differences in CIU50s immerged which suggests small changes in folding stability after stress (primarily for the trastuzumab antibodies). In general, intact mass analysis showed similar glycoform profiles for each mAb pair. Most differences in intact mass spectra were the result of additional C-terminal lysine residues present on the biosimilars of each pair; however, differences in minor glycoforms ((Man5)2 and (G2F)2) were detected between the trastuzumab originator and biosimilar. Cation-exchange chromatography revealed a significantly higher portion of basic variants present in the biosimilars of each pair both before and after stress. Prior to stress, size-exclusion chromatograms were highly similar for each mAb pair, with only the rituximab biosimilar showing a slightly higher portion of aggregates compared to the originator product. Over the course of the stress period, all 3 pairs showed highly similar levels of aggregation by SEC, and these results were confirmed by SDS-PAGE.
Conclusion: In this study, we used a multitude of analytical techniques to compare 3 originator mAb drug products and their biosimilars. We performed side-by-side comparisons for each pair before and after a 4-week incubation at 40ºC with shaking at 250 rpm in attempt to exacerbate existing differences. Taking into consideration the results from each method, all originator-biosimilar pairs contained a high degree of structural similarity, with minor differences in C-terminal lysine residues, charge variants, and gas-phase folding stability. Overall, this work provides an in-depth example of how a range of bioanalytical tools can be applied to generate structural profiles when gathering evidence to establish biosimilarity.
Troy Halseth– University of Michigan, Ann Arbor, Michigan
Jukyung Kang– Graduate student, University of Michigan, Ann Arbor, Michigan
Daniel Vallejo– Graduate student, University of Michigan, Ann Arbor, Michigan
Zeynab Izadi Najafabadi– Ann Arbor, Michigan
Ilker Sen– Principle Scientist, Protein Metrics Inc., Cupertino, California
Michael Ford– Co-Founder, MS Bioworks, Ann Arbor, Michigan
Brandon Ruotolo– Professor, University of Michigan, Ann Arbor, Michigan
Anna Schwendeman– University of Michigan, Ann Arbor, Michigan