Biologics In Vitro Characterization Advancements to Streamline Development and Approval Timelines

Monoclonal antibodies (mAbs) account for 120+ FDA approved products and are frequently used to treat patients with chronic autoimmune diseases or cancer. Despite being approved for similar indications, not all mAbs share the same structure-function motifs. Given complexities in sizes, structures and manufacturing processes, there are known differences between mAbs of similar classes. One notable difference between mAbs is their post-translational modification (PTM) profile. PTMs are comprised of features native to amino acids such as oxidation, deamidation, methylation, etc. Two PTMs in particular, glycans and shuffled disulfide bonds, are of great interest to the pharmaceutical field given their impact on drug safety and efficacy. Characterizing these two PTMs on mAbs in vitro in order to predict safety/efficacy implications is what drove my research. Chapter II encompasses my work studying anti-TNFa mAbs: Humira, Remicade and Simponi Aria. For this project we determined glycosylation profiles using LC-MS/MS and LC-FLR. Then we performed in vitro functional assays, TNFa binding ELISA, FcyRIIIa AlphaLISA and ADCC, to correlate structure and function. Humira had the fewest unique glycans, 12.1±0.7% of which were afucosylated and mannosylated, and, perhaps consequently, had the highest Fc binding affinity. Humira had a 7.2-fold higher binding affinity to FcyRIIIa than Remicade and 3.3-fold higher than Simponi Aria. Since Humira had significantly higher Fc and Fab binding affinities, it was 15.1% or 19.7% more potent in the ADCC assay when compared with Remicade and Simponi Aria, respectively. Our results confirmed significant differences between the three mAbs, yet recognized that in vivo efficacy may differ due to confounding variables. Chapter III follows my research on shuffled disulfide bonds found in rituximab and bevacizumab innovator/biosimilar pairs. We studied the formation of shuffled disulfide bonds and subsequent degradation via non-reduced digestion followed by LC-MS/MS, SEC and SDS-PAGE. After a 4-week incubation, innovator bevacizumab had an upward trend in shuffled disulfide bonds (0.58±0.08% to 1.46±1.10%) whereas innovator rituximab maintained its shuffled disulfide bond level (0.24±0.21% to 0.51±0.11%). Bevacizumabs started with an average of 70% more shuffled bonds than rituximabs, leading to a higher propensity for aggregation. The bevacizumabs had approximately 6% monomer loss primarily due to aggregation compared to a 1.5% monomer loss due to fragmentation for rituximabs. Our results showcased the importance of monitoring lower abundance PTMs and degradants. Chapter IV covers my research comparing glycosylation analysis methods using NIST mAb as a standard. We performed five glycan analysis techniques – three FLR released glycans kits, protein digestion followed by LC-MS/MS and intact MS. The LC-MS/MS method identified 25.2% more glycans than the FLR kits and 5.5 times more glycans than intact-MS. When applying these methods to Herceptin® and its biosimilars, Kanjinti and Ogivri, we observed that Kanjinti had a relative % mannosylated contribution of 1.01±0.38% while Ogivri was 5.95±0.97%. This translates to a 142.0% difference between mannosylated glycans identified in Kanjinti compared to Ogivri. When comparing mannosylated glycan contributions between biosimilars and innovator, there was only a 100.5% difference between Kanjinti and Herceptin and a 64.4% difference between Ogivri and Herceptin. This work emphasizes the need for method standardization to consistently identify glycan species. Chapter V summarizes each project and shares potential future directions for this research. Overall, this dissertation highlights techniques for glycan and disulfide shuffling analysis, emphasizes the need to standardized methods and reporting, and discusses potential collaborations to streamline PTM impact analyses.PhDPharmaceutical SciencesUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/178001/1/jcoghlan_1.pd