Modern biopharmaceutical products, but also most plasma proteins, exhibit extensive structural micro-heterogeneity due to co-occurring post-translational modifications. These modifications affect their functionality and thus need to be characterized in detail. We use high-resolution native mass spectrometry, often combined with glycoproteomics and top- and middle-down proteomics, to analyze this micro-heterogeneity. Taking mAbs, erythropoietin and plasma proteins as model systems, we demonstrate an all-inclusive quantitative profiling of glycoproteins. We demonstrate the usage of a biosimilarity score to quantitatively assess structural similarities.

Moreover, by extending the mass range of Orbitrap mass analyzers we have expanded the reach of native mass spectrometry to higher molecular weight compounds, such as IgG hexamers, endogenous viruses and gene delivery vectors such as the adenovirus. I will describe the latest developments, including Orbitrap based single particle charge detection mass spectrometry, and how that can be used to assess the quality and purity of this next generation of biopharmaceutical products.

Antibody behavior as related to process development, manufacturing, and in vivo biology is diverse, interdependent, and multi-dimensional, with many properties following multiple pathways to exhibit similar behaviors. Biophysical characterization methods are quite useful in their ability to explore a wide range of behaviors, both conformational and colloidal, in a high-throughput in vitro environment. The Multiple Attribute Methodology (MAM) is also useful as a fairly high-throughput method to explore the post-translational modification space of antibody structures. However, the amount of data required to properly predict and modify these properties is too large for standard prediction methods. To attempt to overcome this limitation we have applied generative adversarial network (GAN) technology toward the training of an Antibody-GAN to generate sequences biased toward specific properties, thus enabling the targeted exploration of antibody biology as well as the realization of a humanoid discovery platform. The resulting data is useful for GAN transfer learning, ML-based predictors of properties, and ML-based optimization of antibodies toward desired properties.

Biotransformation of administered protein therapeutics; including deamidation, oxidation and proteolysis, can lead to decreased efficacy and undesirable pharmacokinetics and immunogenicity. Since it is impractical to characterize stability from all candidates in vivo for screening, characterization from in vitro stress conditions are often used as a predictive strategy to inform lead candidate optimization and selection as well as for understanding species to species correlations. Unfortunately, because of technical challenges, there is little work to understand whether such predictive strategy translate to endogenous degradations. We describe recent advancements in sample preparation, front end separations and mass spectrometry methods geared for biopharmaceutics to allow us to capture a comprehensive dataset highlighting the disconnects and translatability between in vitro and in vivo assays for various biotransformations. In addition, translatability of biotransformation between various species and consideration for healthy vs diseased models will also be covered.

As the pharmacological efficiency of novel biologics depends largely on their structural integrity, there is a critical need to better assess and understand their stability in relation to their efficacy upon administration. At Merck PPDM we are trying to define a strategy for the biotransformation of biologics and bioconjugates in both the discovery and development stage. Our goal is to provide early data in the discovery space to drive decisions for the optimal therapeutic design and selection, and facilitate safety and clinical study design, as well as help define critical quality attributes (CQA) / product quality attributes (PQA) for clinical/commercial supply.

Protein sequence liabilities such as deamidation, oxidation, sialylation…, and amino acid sequence clipping, are often identified early in the program in stress studies in buffers. One key question is whether those sequence liabilities translate in vivo and affect efficacy. Here we share case studies we have encountered through the years on biotransformation of biologics and how sequence liabilities translate between in buffer developability and in vivo biotransformation. We also share our new PRM based peptide quant platform capable of identifying previously unknown sequence liabilities. Based on our experience, deamidation have a higher likelihood to translate in vivo while oxidation level usually remains unchanged in vivo. In vivo clipping of amino acid sequence may not all be captured by stress study due to different mechanisms thus special investigation of molecular integrity may be needed for engineered constructs like bi-specific antibodies or fusion proteins.

Drug development and regulatory filing for the mRNA vaccines is supported by the development of analytical methods for the characterization of mRNA and the drug product. The purpose of this talk is to provide a brief overview on the applications/workflows specific to liquid chromatography- mass spectrometry (LC-MS) for the characterization of critical quality attributes (cap, tail and sequence integrity) of mRNA. Oligonucleotide mapping and fingerprinting approaches are discussed in detail. In addition, as a case study, the ability of LC-MS to characterize previously unreported impurity formed in mRNA LNP’s through a novel class of lipid:mRNA reactions is briefly introduced.

Multi-attribute method (MAM) approaches to characterizing and monitoring the production of biopharmaceuticals offer the ability to replace multiple analytical technologies with a single mass spectrometry (MS) analysis. Assessing multiple critical quality attributes (CQAs) at the molecular level delivers a comprehensive understanding of the end product; ultimately enabling a true quality-by-design approach to biotherapeutic development.

Routine use of LC-MS MAM requires overcoming many scientific, technological and methodological challenges:

• Managing large amounts of information-rich MS data generated by highly robust methods,

• Producing unbiased and fully audited results that are generated using standardized procedures,

• Meeting the demands of process validation Stage 1, which requires assessment of method reliability and performances.

Monitoring CQAs through MAM requires flexible solutions (i.e. robotization for samples handling, agile software, etc.) that cope with these challenges, and also enable reporting on the system suitability and Post-Translational Modifications (PTM) levels, while ensuring high level of compliance and data integrity.

This talk will give an overview of challenges met during MAM development for a therapeutic protein in the context of process characterization.

Antibody-drug conjugates (ADCs) are comprised of a monoclonal antibody conjugated to a cytotoxic drug-linker, both of which can undergo chemical modification that may have impacts on efficacy, toxicity, and patient safety. Capillary electrophoresis (CE) based assays such as icIEF are used to monitor these modifications by assessment of charge variants as part of release and stability testing. Unfortunately, CE assays do not lend themselves to purification and enrichment of sufficient material for detailed downstream direct characterization of the separated variants. These challenges are even greater if the conjugated drug-linkers themselves carry a net charge. To address these difficulties, we propose a novel strategy for characterization of ADC charge variants leveraging stability samples, mass spectrometry, and an in-house probabilistic charge variant distribution model.

Here, we describe our probabilistic in silico model of charge variant separation that combines orthogonal data sources including CE-SDS, SEC-MS, intact mass analysis and peptide mapping. Individual chemical modifications to the antibody backbone and conjugated drug-linkers were quantitated and combined into a binomial distribution generating a full accounting of theoretical species and their charge. We used a novel, average 4-load, antibody-drug conjugate to evaluate this strategy for both neutral and charged drug-linkers and our modeled distributions were validated against an icIEF separation using a forced-degradation study. Based on initial results, the model was refined to include glycation analysis, chymotryptic peptide mapping to improve coverage of posttranslational modification, and updated methods to minimize artifactual modification during peptide mapping sample preparation. The agreement between our modeled profile and the icIEF allows us to infer species identity within observed charge variant peaks and provides insight into the criticality of observed changes.

Disulfide bonds (SS) play a critical stabilizing role in therapeutic mAb tertiary and high order structures by forming cross-links between different regions of polypeptide chains. Therefore, detailed information on native and scrambled disulfide connectivity is essential for biotherapeutics structure integrity assessment.

This work presents a novel method for identifying and then creating a mass spectral library for disulfide-linked peptides originating from the NISTmAb, NIST Reference Material 8671. Analyses involved both partially reduced and non-reduced samples under neutral and weakly basic conditions followed by LC–MS/MS. Disulfide-linked peptides are identified by both MS1 ion and MS2 fragment ion data in order to completely map all the disulfide linkages in the NISTmAb. This led to the detection of 383 distinct disulfide-linked peptide ions, nine native disulfide bonds, as well as 86 additional disulfide linkages arising from disulfide bond shuffling. Fragmentation features of disulfide bonds under low-energy collision dissociation were examined. These include (1) peptide bond cleavage leaving disulfide bonds intact; (2) disulfide bond cleavage; and (3) double cleavage products resulting from breakages of two peptide bonds or both peptide and disulfide bonds. A fully annotated spectral library was created from peptide spectra obtained in this analysis and is publicly available for free download at https://chemdata.nist.gov/dokuwiki/doku.php?id=peptidew:lib:disulfidepeptides.

This work highlights the unique difficulties in a comprehensive analysis of disulfide-linked peptides especially where reduction is incomplete, either by design or not, and where fragmentation lacks readily identifiable diagnostic fragment ions. Transforming these results into a spectral library will enable others to efficiently identify both native and scrambled disulfide peptides in routine proteomics analyses containing IgG1 antibodies. Moreover, we show that one may identify such peptides originating from IgG1 proteins in human serum, thereby serving as a means of monitoring the completeness of protein reduction in proteomics studies.

The next generation of medicines will rely heavily upon our ability to quickly assess the structures and stabilities of such complex macromolecular machines, as well as the influence of large libraries of conformationally-selective small molecule binders and protein-based biotherapeutics. Such endeavors are nearly insurmountable with current tools. In this presentation, I discuss recent developments surrounding collision induced unfolding (CIU) methods that aim to bridge this technology gap. CIU uses ion mobility-mass spectrometry (IM-MS) to measure the stability and unfolding pathways of gas-phase proteins, without the need for covalent labels or tagging, and consuming 10-100 times less sample than almost any other label-free technology. Recent developments in high-throughput CIU screening methods, their ability to track alterations in monoclonal antibody structure as a function of stress, and software developments that seek to enhance CIU information content will be discussed.

The high molecular weight (HMW) size variants present in therapeutic monoclonal antibody (mAb) samples need to be closely monitored and characterized due to their impact on product safety and efficacy. Because of the complexity and often low abundances in final drug substance (DS) samples, characterization of such HMW species is challenging and traditionally requires offline enrichment of the HMW species followed by analysis using various analytical tools. Recent advances in native SEC-MS technique make direct characterization of HMW variants in unfractionated samples possible. However, application of this method alone still cannot obtain a complete picture of the mAb HMW profile and often lead to ambiguous assignment for complex systems. To overcome these challenges, we developed a novel post-column denaturation-assisted nSEC-MS method (PCD-assisted nSEC-MS) that is optimized to dissociate SEC-resolved, non-covalent mAb HMW species into constituent components for subsequent MS detection. As a result, this new approach enables simultaneous detection of both non-covalent and non-dissociable HMW species under identical SEC separation conditions. In addition, this strategy improves the identification of heterogeneous HMW species by 1) confirming the identities of the constituent subunits dissociated from the non-covalent HMW complexes; and 2) achieving more accurate mass measurement of non-dissociable HMW species by removing interference from co-eluting, non-covalent species. Furthermore, by incorporating a limited enzymatic digestion step, the PCD-assisted nSEC-MS method can readily reveal both the interaction nature and interaction interfaces of mAb aggregates at subdomain levels. Notably, as the developed method is highly sensitive, it can provide comprehensive HMW characterization using unfractionated samples, making it a desirable assay to support various tasks during the development of therapeutic mAbs. The utility of this method is demonstrated in different case studies, ranging from enriched HMW characterization at late-stage development, comparability assessment due to process changes, and forced degradation study of co-formulated mAbs.

This talk will focus on CMC strategies used to expedite development of neutralizing monoclonal antibodies against COVID-19. It will also present guiding principles for CMC expectations for Emergency Use Authorization of these products and potential CMC strategies to address emerging COVID-19 variants. Lessons learned from this experience will also be discussed