While peroxide and leachable metal-induced chemical modifications are among the most important quality attributes in bioprocess development, there is no mainstream characterization method covering all common modifications theoretically possible on therapeutic proteins that also gives consistent and quick results. Here, we describe a method for rapid and consistent global characterization of leachable metal- or peroxide-stressed immunoglobulin (lg) G1 monoclonal antibodies (mAbs). Using two independent protease digestions, data-independent acquisition (DIA) and data-dependent acquisition (DDA) liquid chromatography mass spectrometry (LC-MS), we monitored 55 potential chemical modifications on trastuzumab, a humanized lgG1 mAb.
Processing templates including all observed peptides were developed on Skyline to consistently monitor all modifications throughout the stress conditions for both enzymatic digestions. The Global Characterization Data Processing Site, a universal automated data processing application, was created to batch process data, plot modification trends for peptides, generate sortable and downloadable modification tables, and produce Jmol code for three-dimensional structural models of the analyzed protein. The combination of Skyline templates and the Global Characterization Data Processing Site results in a universally applicable assay allowing users to batch process numerous modifications. Applying this new method to stability studies will promote a broader and deeper understanding of stress modifications on therapeutic proteins.
In this study, fully formulated trastuzumab (a humanized IgG1 mAb) was stressed using H2O2 and common leachable metals and characterized using a bottom-up proteomics approach along with our custom-built Global Characterization Data Processing Site. Briefly, each unstressed and stressed protein sample was split into two aliquots for digestion with either trypsin or chymotrypsin and analyzed using ultra-high-performance liquid chromatography high-resolution mass spectrometry (UHPLC HR-MS) in both data-dependent and data-independent acquisition modes.
Once data was collected, custom Skyline processing templates were applied and fed into the Global Characterization Data Processing Site which automates the generation of modification summary tables, interactive trending plots, and Jmol code for visualizing the percentage of modifications on a three-dimensional molecular model. In the present work, 52 modifications were detected on trastuzumab across the various conditions studied. The site allows for customization of the number or type of digestive enzymes, protein subunits, chemical modifications, and sample replicates. Thus, the site is broadly applicable for modification data processing from any protein or protein complex of interest.
Conclusions
Using our method for rapid global characterization of protein therapeutics, we were able to quickly and reproducibly characterize lgGl following oxidative stress. The Global Characterization Data Processing Site was programmed for fast, accurate, and automatic processing of modification trends, generation of modification summary tables, and production of 3D molecular models. This provides unprecedented visibility into chemical modifications of fully formulated trastuzumab and furnishes the foundations for a new standard in stability and stress studies of biotherapeutics. We have demonstrated its unique ability to monitor 55 degradants on 18 different amino acids and found 53 discrete modification sites on the mAb following oxidative stress. Several modifications were observed in the known binding sites of lgGl Fe to FcgRs, complement protein Clq, and FcRn, which could potentially affect its mechanism of action. Future experiments should target the potential consequence of these degradations on their ability to interact with corresponding binding proteins.
Acknowledgements
The authors would like to thank the cell culture and purification team (Catalent) who supplied the trastuzumab used in this work, Robert Vaughan, PhD (Catalent), for preparing this poster presentation, and both Brendan Maclean and Nicholas Shulman (University of Washington) for upgrading the Skyline maximum number of FASTA imported peptides to facilitate data processing.
*This is an adapted version of a poster that appeared in mAbs, 11:6, 1089-1100, DOI 101080/19420862.2019.1625676.
Yao Chen, PhD, is senior scientist, Analytical Development, Emma Doud, PhD, is senior scientist, Process Development, Todd Stone, PhD, is senior manager, Analytical & Formulation Development, Lun Xin is scientist, Analytical Development, Wei Hong, is scientist, Analytical Development, and Yunsong Li, PhD, ([email protected]) is senior director, Product Development, Catalent Biologics.