Mass spectrometry plays and important role in advancing on biologic drug development and manufacturing, but limitations still exist.
Mass spectrometry (MS) is a powerful and sensitive technique used to detect, identify, and quantify molecules separated by their mass to charge (m/z) ratio in the gas phase. In a recent publication by FDA reviewing MS usage trends over the past 10 years of biologic license application (BLA) filings, the agency determined that MS is now ubiquitous in filings and that the number of ways it is being used is increasing (1).
The demands of life-science applications have led to the improvement of MS technologies and rapid growth of new types of instruments that feature powerful analytical capabilities-- sensitivity, selectivity, resolution, throughput, mass range, and mass accuracy, according to Gang Huang, vice-president of analytical development and regulatory affairs with WuXi Biologics. In the biopharmaceutical industry today, mass spectrometry finds applications at the earliest discovery stages (e.g., imaging of biomarker/protein binding in cells), during process development, and for product characterization throughout the development cycle, including for release testing.
Challenges with sample preparation and data processing remain, however. Biologics manufacturers and analytical laboratories are working closely with instrument suppliers to address these issues. Improvements are continually introduced, and as a result the potential for additional use of this analytical technology for biologic drug development and manufacturing is significant.
Mass spectrometry is widely used for the characterization of protein therapeutics in the development stage. More specifically, MS is a effective tool for intact protein analysis (measurement of the molecular weight of the protein), fragment analysis, peptide mapping, and identification of post-translational modifications (PTMs), according to Tiffani Manolis, segment manager for pharma at Agilent Technologies.
As a characterization tool, MS is used to assign the primary structure of biologic drugs, including amino acid sequencing for peptides/proteins, disulfide bond mapping, N-glycan profiling of monoclonal antibodies (mAbs), and nucleotide sequencing for oligonucleotides, adds Jie Ding, associate director of mass spectrometry services at PPD Laboratories’ GMP laboratory.
Peptide mapping can also be applied for determination of other critical quality attributes (CQAs), such as physicochemical properties of oxidation, glycosylation, deamidation, and isomerization and the presence of N-terminal cyclization as well as confirmation N- and C-terminal groups, according to Hillary Schuessler, an investigator with GlaxoSmithKline (GSK) R&D Platform Technology & Science. She adds that characterization of product variants and higher-order structure (HOS) analysis via hydrogen-deuterium exchange (HDX) are also important applications of MS.
In addition, MS can be used for media and feed characterization, such as quantitative targeted metabolomics for amino acids, vitamins, lipids, central carbon metabolites, sugars and sugar bases, etc. In this case, the absolute quantities of these compounds are measured as they are depleted/excreted from cells. “A challenge is to differentiate extracellular from intracellular metabolites. It is possible, albeit challenging, to measure the flux of metabolites for use in predictive modeling of protein product quality attributes such as glycoform variants and amino acid misincorporation,” says Greg Kilby, manager of biopharm analytical sciences with GSK R&D Platform Technology & Science.
Profiling of process-related impurities such as host-cell proteins (HCPs) is another application for MS. “The identification of the most abundant HCP species provides valuable information to process scientists for developing a tailored process for removal of these critical impurities,” notes Huang. Sanofi uses MS to support process development and manufacturing, including impurity identification and tracking and the collection of detailed, high-level, advanced structural information to confirm that the product is as intended, according to Jianmei Kochling, director of analytical science and technology for the company.
Sequence variant analysis of production cell lines used for biologic production is an important part of process development. The potential presence of sequence variants, which can result from DNA mutations and amino acid misincorporations, is analyzed at the protein level using high resolution MS and data analysis software. From multiple candidate clones, the one without mutations or low level mutations will be chosen as the final clone for further development, according to Huang.
At early development states, MS is used in cell imaging applications, as well as for characterization of biomarkers and determining drug metabolism and pharmacokinetic (DMPK) profiles (clearance/lifetimes).
These applications are just the beginning, however. “The roles for MS are rapidly expanding to more hybrid qualitative/quantitative applications and the examination of higher-order biotherapeutic structure,” notes Scott J. Berger, senior manager of biopharmaceuticals in the Waters Corporation’s pharmaceutical business group.
Other newer applications include biosimilar development, where MS is an enabling technology to show that an innovator and biosimilar have identical sequence and comparable variant profiles. Similarly, Berger says, the rise of antibody-drug conjugates (ADCs) has required more advanced liquid chromatography (LC)-MS laboratory workflows for challenging separations, mass detection, and data processing to determine the average number of drug conjugates (DAR) on a molecule, their distribution across the many possible sites of reaction, and individual occupancy levels for each of these sites on these hyper-complex molecules.
Mass spectrometry is a preferred analytical technique in many of these applications because it is a more targeted method that provides detailed information about protein structure/conformation, whether for the desired product or impurities like HCPs that are present at low concentrations.
The sensitivity and specificity, high mass accuracy at low part per million levels, and ability to return precise chemical information on the molecule of interest are main drivers for using mass spectrometry, according to Kilby. Additional attractive characteristics include compatibility with most chromatographic methodologies, the ability to gain both qualitative and quantitative [relative and absolute] information, and the lack of any theoretical limitations on the size of proteins that can be analyzed.
“MS is an ideal tool for supporting process development and a quality-by-design approach. For process impurity identification, it is much more specific than ELISA (enzyme-linked immunosorbent assay) testing, which is the conventional method (and is still used for rapid product release),” Kochling says.
Mass spectrometry is also preferred in these applications because it offers superior sensitivity and specificity without the need for a large volume of samples, according to Ding. “Furthermore,” she adds, “there is a plethora of knowledge and well-established procedures for LC-MS/MS analysis of biologic drugs, from sample preparations to software-assisted data interpretation.”
Berger also notes that MS provides more confident qualitative mass analyses and makes it possible to identify a single peptide or modified peptide in the presence of the many other peptides that are generated in the digest of a biotherapeutic protein, even if that peptide is fully or partially co-eluting with other components. “Monitoring of multiple mass channels simultaneously allows for monitoring of several components at the same time, and in many cases without the need for optimizing the separation conditions to get valuable quantitative information,” he says.
In addition, the ability to detect components at levels lower than optical detection-based assays expands the dynamic range of detection of an assay for the peptide mappeak of a peptide or its variant, according to Berger.
Overall, observes Manolis, MS is an indispensable tool for peptide and protein analysis due to its speed, sensitivity, and versatility. “MS is particularly useful for gaining knowledge about the location of disulfide bonds and amino acid sequences,” asserts Kochling. Adds Mario DiPaola, senior scientific director for Charles River Laboratories: “No other analytical technique can provide the extent of information obtained by mass spectrometry, nor the selectivity or sensitivity. Previously it would have taken months to determine or confirm the entire sequence of a protein by first collecting enzymatic digests and then performing Edman degradation, but now the same analysis can be performed in a matter of days with mass spectrometry while using micrograms of product rather than milligrams.”
Several advances in technology have been enabling the wider use of MS. Examples include developments in high-resolution mass spectrometry, such as quadrupole time-of-flight (QTOF) and orbitrap technology, and workflow-driven software development, according to Ding. Ion mobility methods have made MS useful for HOS analysis, cysteine variant determination, and N-Glycan profiling, while top- and middle-down analyses, which have been made possible through the introduction of electron transfer dissociation (ETD) fragmentation, are useful for determination of CQAs and identification of product variants, says Schuessler.
While not very recent, DiPaola points to the introduction of tandem mass spectrometry and the electrospray ionization (ESI) interface as key advancements in the field of mass spectrometry. ESI has allowed for easy coupling of a high-performance LC (HPLC) system to a mass spectrometer, as a key advancement because it permits separation of species followed by direct in-line mass analysis. Hybrid mass spectrometry has made it possible to obtain very detailed information on PTMs and protein primary structures.
In addition to ETD, the introduction of a variety of other alternate methods to collisional induced dissociation (CID), including electron capture dissociation (ECD), higher-energy collisional dissociation (HCD), electron transfer in the higher-energy collisional dissociation (EThcD), ultraviolet photodissociation (UVPD), and surface induced dissociation (SID), among others, have led to significant improvements in fragmentation technology, according to DiPaola. Some of these fragmentation methods can be used independently or in-series to garner as much information about protein analyte as possible.
For Huang, one of the most powerful developments in the evolution of MS technology is the commercialization of hybrid instruments. “Hybrid MS instruments are made by combining two different types of mass analyzers together in tandem; one can choose almost any combination of quadrupole, time-of-flight, or ion-trap hybrids. These hybrid instruments promise the ability of combining the best features from the different components and allow tandem mass spectrometry experiments and unique scanning modes that are not possible on a single instrument,” he explains.
Huang points to Thermo Scientific’s Orbitrap Fusion mass spectrometer as one example. It is designed to include quantification using isobaric tags, low-level PTM analysis, and data-independent acquisition (DIA). The instrument, according to Huang, features enhanced sensitivity resulting in improved analyte detection, characterization, and quantitation in advanced biopharma, proteomics, and metabolomics applications.
Berger adds that the increasing focus on hybrid quantitative/qualitative workflows favors TOF-based platforms that do not suffer the lower-end dynamic-range limitations of automatic gain control (AGC)/orbitrap-type instruments. “Time-of-flight and quadrupole time-of-flight MS technology has been the primary high-resolution MS tool used for biopharmaceutical characterization and monitoring due to its ability to maintain high resolution and sensitivity independent of the mass of a species, and the ability to do so with increasingly rapid LC and CE [capillary electrophoresis] separations on the front end of these analyses,” he explains.
Various bioinformatics tools associated with MS analyses, such as pathway mapping/analysis, network association, and ADC calculators, have had an impact on the use of MS as well, notes Kilby. “Each major instrument vendor has software that works specifically with its instrument and uses complex algorithms to process its proprietary data files. Vendors constantly seek feedback from users to improve software features, such as chemical intelligence, batch data process, automation, report generation, Title 21 Code of Federal Regulations Part 11 compliance, and so on,” agrees Ding.
More specifically, Manolis notes that dedicated data analysis software has been developed for biopharma applications, and the workflow-specific design this software has streamlined the process. “In addition,” she says, “walkup software has been developed to enable MS novice users, such as biologists, to have access to high-end MS instruments. Special consideration has also been given to providing total workflow solutions to address sample preparation all the way to reporting.”
Overall, Berger believes that the expanded use of MS in more targeted biotherapeutic CQA monitoring experiments, even in later (regulated) development and quality control (QC) environments, is related to the increased robustness of the instrumentation, growing usability of these systems, and deployment on informatics platforms designed for regulated environments. “The ability to follow specific product variants in a peptide map, intact mass profile, or released glycan profile enables targeted quantification of the amount of that modification as an organization develops and matures its manufacturing processes or sets specific limits of a variant in a QC release test,” he explains.
The other area of great expansion has been in HOS analysis, according to Berger. The development of ion mobility MS has introduced the ability to measure collisional cross-section (CCS) data for molecules, bringing an added level of separation to MS analysis and generating data based upon gas-phase cross-sectional area and shape, in addition to the traditional mass and charge characteristics measured by a mass spectrometer. Biologic folding interactions and stability can be screened and assayed using this type of ion mobility information. In addition, Berger notes that more resolving HOS information can now routinely be provided by hydrogen deuterium exchange MS (HDX-MS), which measures the accessibility of backbone amide hydrogens to exchange with deuterated water in solution. HDX-MS is being used to define the interaction of a mAb and its target (epitope mapping) for intellectual property (IP) filings and to demonstrate structural comparability between innovator and biosimilar biotherapeutics.
With current MS technology, there is a disconnect between the practical and actual time it takes to complete complex analyses, according to Kochling. “The analysis of proteins is very complicated, and even with current mass spec instrumentation and software, a significant amount of manual labor is required, and in some cases it can take up to one month to complete an analysis, which is not practical in an industrial setting,” she states. “Although mass spec technology has been ever improved in sensitivity and dynamic range, the hardware capability is ultimately limited by the complexity of samples. Sample preparation technology and procedure improvement can compensate for the instrument capability in sensitivity and dynamic range,” she continues.
Current mass spectrometers were designed and optimized for the analysis of smaller molecules, and as a result large protein molecules often suffer from poor data quality resulting from limited resolution, sensitivity, and mass range, adds Manolis. She notes that additional alternative and optimized fragmentation methods for large molecules for top-down and middle-down analysis are desired. On the other hand, Berger believes that MS systems are designed to be general-purpose platforms for both large- and small-molecule studies, with compromises on some specific performance attributes for large molecules to accommodate a wider range of applications in the lab.
At GSK R&D Platform Technology & Science, poor parallelization is an issue; currently it is not possible to highly multiplex MS analyses without buying extra mass spectrometers (compared with genomic and microarray technology, for example). In addition, as resolution and scan speed increase, files sizes are getting very large such that current informatics suites struggle with data analysis, especially for large experiments, according to Schuessler. “With the enormous volume of data being generated, data processing and analysis become increasingly important and remain a bottleneck,” agrees Huang.
Furthermore, according to DiPaola, the analysis of these files requires sophisticated and expensive software, as well as highly skilled and knowledgeable users. “Both of these scenarios present recruiting and financial challenges for laboratories and companies,” he asserts.
Kilby also points to the incompatibility of MS with commonly used biologic matrices/buffers/detergents, issues with samples that have extended dynamic ranges (HCPs, serum proteomics, etc.), integration in process analytical technology and continuous flow manufacturing, and the fact that response factors are not universal (unlike for UV, charged-aerosol, and evaporative-light-scattering detectors).
The need to label or spike analytes with an appropriate isotopically labeled species to obtain highly quantitative results due to the ionization variability of species is one issue for DiPaola. Another is the difficulty of detecting low-level impurities (<1-2% in abundance) in biological samples without some prior knowledge of their type. A third issue for DiPaola is the need to confirm isobaric amino acids when conducting sequencing by MS/MS using Edman degradation, which becomes a bottleneck because the peptides containing such isobaric amino acids must be collected and individually sequenced.
Finally, DiPaola notes that the high cost of mass spectrometers presents another obstacle, especially when dealing with high-resolution, tandem systems. “Affordability is a sizable issue for small laboratories and companies with limited financial capacity. Additionally, these systems are rather complex and delicate, so that any repairs can only be performed by skilled engineers, adding additional costs,” he explains.
Berger asserts that many of these limitations with mass spectrometry today do not involve the instrument itself, but other elements of the analytical workflow that interact with the system. “Faster and more robust sample preparation is needed to match the improvement in back-end data processing throughput for many applications of MS in the biopharmaceutical industry,” he notes.
For instance, Berger points to clone screening and other early development activities, where reproducible microscale and nanoscale separations remain a challenge to many MS users. “Waters has developed a chip-based microfluidic platform, the ionKey/MS System, that simplifies the process of microscale LC-MS of proteins and peptides, but the industry is looking for further innovations that raise usability of systems at the microscale to that of analytical scale LC-MS,” he observes.
Berger also notes that innovative software products such as those from Waters for automating biotherapeutic peptide mapping and intact mass analysis, intact mass, peptide mapping, and released glycan analysis characterization and monitoring have been key to advancing the productivity of labs using LC–MS instrumentation. He agrees, though, that “analysts are looking for smarter software, automated processing, custom calculations for critical results, and efficient reporting tools that simplify communication of the results, and they now want to be able to use this software in regulated environments.”
To address throughput issues, there is, according to DiPaola, an effort to transition from serial data acquisition to parallel acquisition. “In combination with novel data interpretive software tools, this move should make possible the multiplexing of data acquisition and analysis for peptides, intact proteins, and whole protein complexes,” he says. He also expects new data formats, compression methods, and storage architectures to be introduced to address file size, storage space, retrieval, and speed of data analysis issues.
Researchers at GSK R&D Platform Technology & Science are excited about a number of new developments in MS technology, including platform-neutral, multi-vendor solutions such as the software suite for biopharma applications from Protein Metrics; integrated analyzers for process monitoring, such as the Waters ACQUITY QDa Mass Detector; commercial solutions for ion mobility, HDX, and ETD fragmentation; and capillary electrophoresis devices that can be attached directly to the MS source such as 908 Devices’ ZipChip, which are fast and sensitive and offer the potential for rapid CE-MS analysis of peptide maps, intact proteins, charge variants, and metabolites, according Kilby.
DiPaola also expects high-resolution, top-down or bottom-up systems combined with novel fragmentation techniques and improved analysis software to be available in a year or two that will overcome the isobaric amino acid issue for de-novo protein sequencing by mass spectrometry. He also notes that new dissociation methods such as CID, HCD, ETD, and infrared multi-photon dissociation (IRMPD) are generating new acquisition paradigms that require more systematic data collection. “New software tools and algorithms with compliance to 21 CFR Part 11 requirements are in development to address these new data acquisition paradigms,” he asserts.
In essence, major developments efforts can be divided into two disciplines, according to Manolis. “From a physics perspective, significant advances in instrument design are very close to release that will greatly improve not just the capability to analyze large biomolecules, but also the reproducibility of such analyses. On the biology front, advancements in data analysis software that can make clear biological sense of the rich data collected by mass spectrometry is critical for multiple applications. Combining these efforts with the latest capabilities in automated sample preparation and separation techniques allows for a big step forward in the total confidence in using mass spectrometry for biopharmaceuticals,” she states.
Despite the advances in MS technology achieved to date, those in the industry expect further notable improvements in the long term. “Research in electron transfer dissociation (ETD), top-down and bottom-up proteomics, high-resolution mass spectrometry (HRMS), and software-assisted structural elucidation will have lasting impact on biologics analysis,” Ding observes.
Huang also believes that multi-attribute methods that utilize LC/MS for separation, identification, and quantification of post-translational modifications will have potential to replace multiple purity methods currently used for protein product release testing. “With this approach, there will be fewer methods for product release testing but richer product critical quality attribute information obtained,” he comments.
While current MS technology has been primarily limited to the analysis of in-vitro proteins, proteoforms, and multi-protein complexes, the future drive in mass spectrometry is to study the same entities in vivo to determine native biological structures and their dynamics within biological processes, according to DiPaola. “New MALDI [matrix-assisted laser desorption ionization] imaging spectrometers are available and more advanced models will be soon introduced for spatial mapping of entities of interest, including proteins, signaling molecules, etc., in tissues and eventually in single cells,” he says. Data files for imaging experiments are starting to approach “Big data” definitions, though, and thus challenge most existing software tools, according to Kilby.
In a similar vein, Kilby notes that the application of MS to quantitative whole body imaging, which is currently very expensive and requires the use of radiolabeled compounds, would allow analysis of all compounds present in a given location, not just the radiolabeled ones, without the need for radiation. “There is a long way to go, however, to extend dynamic ranges and improve compatibility with salts and other materials present in untreated biologic samples, and it is not yet known whether it will be possible to return quantitative results,” he says.
One of the biggest challenges Waters is taking on, according to Berger, is to enable more people in biopharmaceutical organizations to leverage the resolution, sensitivity, and dynamic range offered by mass spectrometry. “We also want to raise the dependability of higher-performance instruments to match that of optical detectors that are routinely used in biopharmaceutical labs today. Photodiode array UV/Vis detectors were once viewed as too complicated for regulated environments, but today we rarely give them a second thought about their applicability for the analysis of pharmaceuticals. My hope is that high-performance MS analysis will soon be seen the same way,” he says.
The goal on the horizon, according to Manolis, is to extract the most information for the least amount of effort. “Can we get to a point where we can directly analyze a biopharmaceutical sample without any time-consuming sample preparation and get all the information about the sample we need?” she asks. “To do so will require improved fragmentation techniques, greater dynamic range, and software tools that can pull more information out from a single sample, as well as aggregate data across the development cycle,” Manolis asserts.
Kochling certainly expects greater applicability for mass spectrometry in the biopharmaceutical industry. “MS is already providing great benefits to biologics development and manufacturing. Although improvements in data processing software for throughout enhancement are still needed, mass spec technology has advanced dramatically in the past 25 years. It is exciting to think about how its capabilities will be expanded over the next 25.”
1. S. Rogstad, A. Faustino, A. Ruth, and J. Park, J. Amer. Soc. Mass. Spectrom. (2016).