Biopharmaceutical drugs have experienced a steady growth since the approval of the first recombinant protein biologic Humulin, a biosynthetic form of human insulin, in 1982. Biopharmaceutical drugs are derived from biological sources and include peptides, recombinant proteins, mono- and polyclonal antibodies, antibody-drug conjugates, cells (e.g., CAR-Ts, stem cells), tissues, toxins, vaccines, antisense oligonucleotides, and gene therapies. Within the biopharmaceutical segment, antibodies comprise the fastest growing category, in part fueled by recent advances in immune checkpoint inhibitors for cancer treatment. Antibodies accounted for 53% of all first-time global approvals from 2014 until July 2018, up from only 27% from 2010 to 2014. They also constituted 65.6% of 2017 global sales, up from 49.9% in 2011.
To keep pace with this burgeoning market, research has also advanced antibody characterization and quality assurance. Protein biotherapeutics are inherently heterogenous and complex and are orders of magnitude larger than small-molecule pharmaceuticals, which makes analysis challenging. Protein structure is multilayered, with a primary amino acid sequence, that folds into secondary motifs, which may further interact to yield a tertiary structure (i.e., two or more polypeptides). Moreover, post-translational modifications (PTMs), such as glycosylation and phosphorylation, further complicate protein structure. And finally, within the biological milieu, proteins may interact with other biomolecules (e.g., proteins, oligonucleotides) to form complexes with important biological functions.
Monoclonal Antibody Workflows Using High-Resolution Q-TOF LCMS and the Protein Metrics Software Suite
Accurate characterization of monoclonal antibodies (mAbs) is essential to the development of biotherapeutics; a thorough understanding of biotherapeutic properties aids in the optimization of bioprocess production, product formulation, and product dosage.
Download this application note to discover how high-resolution LC-MS has become an important tool to resolve complex mAb structures.
Correct characterization, discovery, and development
Mass spectrometry (MS) has proven extremely adept for characterizing protein biologics, including biosimilars, and assuring their quality through quality control (QC). “MS is kind of a huge deal in pharma with regards to protein and antibody characterization, primarily because of its incredible versatility with regards to pharma’s needs at both the product development and release stages,” discussed Peter Liuni, PhD, Scientist at the Proteomics Center for Protein Degradation at the Dana-Farber Cancer Institute. “Liquid chromatography MS (LC-MS) can detect peptides at attomole sensitivity, which can be incredibly important for accurately formulating your products, monitoring stability, and detecting any residual impurities. Fluorescence assays such as ELISA’s typically reach this level of sensitivity, but when an antibody isn’t available, MS can jump in and typically deliver what is needed.”
At the primary structural level, MS can accurately determine polypeptide mass and amino acid sequence. “More advanced MS methods can even quantify the mass and relative conformation of higher-order structures, such as multimers or non-covalent protein-protein and protein-ligand complexes using ion mobility MS (IM-MS) under native-like conditions, which attempts to preserve the biologic’s natural structure,” Dr Liuni explained. “The advantage here is to utilize the mass spectrometer’s speed, high mass selectivity, and sensitivity to gain insight into native-like protein attributes compared to slower, lower resolution biochemical techniques. However, it’s important to note that native-like is the correct terminology because of the on-going debate regarding the validity of comparing gas-phase to solution-phase protein structure. Nevertheless, you can obtain a wealth of information on your therapeutics just by implementing a native MS workflow.”
Numerous MS methods also exist for characterizing other structural protein attributes. “As far as selectivity, LC-MS and other combined techniques such as size-exclusion chromatography MS (SEC-MS) and electrospray ionization MS (ESI-MS) are, in my opinion, unrivalled at this stage. Nothing is as capable as MS for uncovering the shear breadth of proteoforms. Even when the biochemical analysis might guarantee that the product is “pure”, run it on an MS and sure enough you’ll find things like small degradation products, process impurities, and various proteoforms of your desired product,” Dr Liuni continued. For instance, variability in PTMs, such as protein glycosylation, can occur at various protein residues and by different glycans. Variability in glycosylation can arise from the manufacturing process; however, since glycosylation can affect serum half-life and the presence of non-human glycans could elicit immunogenic reactions, it is critical to identify precise glycosylation. Glycosylation profiling may be performed by a variety of MS methods, which ultimately identify both the type of glycan PTMs and their location on the protein. Disulfide bonds are another crucial structural motif that facilitate folding, lend stability, and confer correct functioning. Thus, it is critical to determine disulfide bonding in protein therapeutics, especially for antibodies that possess multiple disulfide bonds, since they are inherent to protein function and therapeutic efficacy. LC-MS and tandem LC-MS (LC-MS/MS), is the premiere method for disulfide bond mapping, which is capable of identifying intra- and inter-chain disulfide bonds in protein structure.
Hydrogen deuterium exchange MS (HDX-MS) is a technique that leverages the exchange of labile protein amide, hydroxyl, or thiol hydrogens with solvent deuterium. “This technique is kind of interesting in that you use the speed, selectivity, and sensitivity of the MS instrument to identify the positions of the exchanged deuteriums. It rests on the premise that exposed protein surfaces will uptake more deuterium, whereas less exposed surfaces uptake lower amounts, giving clues about protein conformational dynamics,” Dr Liuni describes of HDX-MS. “It is incredibly useful if you are interested in determining structural differences in your protein when comparing different candidate ligands. It’s also good at looking at dynamic differences across point mutants. Pharma is very interested in using HDX-MS as a technique to determine the binding surface of their antibody or antigen therapeutics, especially when nuclear magnetic resonance (NMR) or co-crystals are not available. Again the premise behind this is that areas that don’t exchange deuterium are “blocked” by something – and that something is usually the ligand or protein. This helps from a legal perspective for differentiating your therapeutic from a competitor’s in terms of the location of binding, but also for fundamentally understanding the structural aspects of how your therapeutic effectively elicits a response on its target.”
Controlling quality, preventing problems
In addition to characterizing the protein biologic’s structural attributes throughout the drug discovery and development phase, MS has emerged as a strong candidate for analyzing quality throughout and at the end of the manufacturing phase as well. Kevin Van Cott, Associate Professor in the Department of Chemical and Biomolecular Engineering at the University of Nebraska-Lincoln, collaborates with industry to help companies develop a MS pipeline for analyzing impurities in their protein biologics. “I work primarily with private companies to troubleshoot production problems of their biologics, mostly in the early stages of development. The majority of my work is in host cell protein (HCP) analysis by MS. HCPs are process contaminants from the host cell used to express the protein biologic, which can trigger immunogenic responses in patients. My interest in HCPs started in 2012 while I was working on a phase III clinical trial project that was halted because some patients developed antibodies against some HCPs in the product. The U.S. Food and Drug Administration (FDA) asked the company sponsoring the clinical trial to identify the immunogenic HCPs, and we worked with them to do so. We also developed a targeted LC-MS/MS multiple reaction monitoring (MRM) method to track these HCPs as the purification process was revised. Since then, we have worked on a number of HCP projects.”
The sensitivity of MS makes it ideal for detecting contaminants. But it is not without difficulties. “The technical challenges of HCP analysis by MS are numerous,” Professor Van Cott says of his work. “The primary technical challenge of HCP analysis by MS is that product protein biologic is present at much higher concentrations than the trace-level HCPs we are trying to detect. For antibody samples, this usually means that HCPs are present at less than 10 ppm. Since MS instruments have a maximum dynamic range of about 105, this means we face hardware limitations for achieving low ppm HCP detection. We can combat this by overloading the column, but this is at the expense of ionization suppression and MS detector saturation. Enriching samples for HCPs by depleting the product protein is a long-term goal in the field and has seen some success. The problem with most enrichment methods is that many HCPs are ‘hitch-hikers’ that interact with the protein biologic. So, depleting the protein biologic from the sample also removes some of the HCPs. The method published by Huang et al. (2017), which is very simple and takes advantage of native monoclonal antibodies’ natural resistance to trypsin digestion, has worked well in our hands and in other labs.”
Another challenge facing MS for HCP analysis is the variation in sample type. “We receive samples from all purification stages, from clarified cell culture harvest through to a vialed drug product,” explained Professor Van Cott. “Thus, we need to deal with many different sample matrices, and often some of the sample components (e.g., excipients such as Tween) are incompatible with MS. Therefore, our sample preparation and analysis methods have to be flexible enough to deal with this variety. Each new project is a new challenge, but there are several options available for sample pre-treatment.”
In addition to the technical experimental challenges, data analysis for HCP identification poses its own set of obstacles. “Most proteomics search engines were not optimally designed to detect trace-level proteins in a sample. The MS/MS spectra from low-abundance HCPs are not always as strong as one would like, and so care must be taken to ensure the search engine correctly identifies the peptide. The biggest problem in HCP data analysis is dealing with false positives. Despite setting stringent search parameters, search engines will still make mistakes, so one can almost never implicitly trust their output. Curating search engine results takes up the majority of my time in an HCP project – looking at peptides’ MS/MS spectra, extracted ion chromatogram (XIC) integrations, and MS isotope profiles to ensure that all are consistent and can be labeled as a ‘true positive’,” Professor Van Cott elaborated on this process.
The conventional methods for HCP analysis include 2D polyacrylamide gel electrophoresis (2D-PAGE) and western blot (WB), and HCP-ELISA, which is the gold-standard. However, an HCP-ELISA requires antibodies to all the potential HCPs in a sample – something that is not always possible. “MS surpasses ELISA in this regard because it is untargeted and provides an unrivaled level of detail,” Professor Van Cott explained. “Knowing the identity and relative abundance of specific HCPs can inform a rational revision of the purification process and generate a risk assessment on the impact of any remaining HCPs on safety to the patient population. I think it is important to keep in mind that the end-goal is patient safety and care when discussing current and new technologies – it’s not just hardware and software, but how technologies will help patients.”
Is there anything MS cannot do better than other technologies for HCP analysis? “One disadvantage of MS analysis is that it sometimes generates ‘too much data’,” Professor Van Cott admitted. “A frequently posed question is – ‘What do we do with all these data?’ HCP-ELISA gives us one single number as the result, which is easy to interpret. Now with MS we have to deal with all these individual HCPs. The technology still has room for improvement; instruments could benefit from a larger dynamic range. The data analysis process would profit from better user-friendly data analysis tools that could be easily understood by non-experts. I can run samples, process data, curate results, and look at my computer screen and tell someone what their HCP problems are – but transferring that information to a report is still yet another bottleneck in the process. However, the pace at which MS has made inroads in biopharmaceutical analysis has been rapid, and I am confident solutions are forthcoming.”