OR WAIT null SECS
Anne Blackwell, PhD, is Bio Columns Product Support Scientist, Agilent Technologies.
The critical quality attributes of biotherapeutics must be monitored to ensure product safety and efficacy.
Biotherapeutics are highly complex molecules that are challenging to produce consistently and uniformly. The fermentation process used to manufacture these proteins inevitably leads to heterogeneity. Critical quality attributes (CQAs) are the characteristics of the biologic that affect safety and efficacy, and therefore, must be carefully monitored. The biotherapeutic product must be purified, and both the product and any remaining impurities have to be characterized and quantified, necessitating a number of tests on the molecule using a wide variety of analytical techniques.
CQAs fall into several broad categories: aggregation, sequence variations, post-translational modifications (PTMs), and host cell proteins and other process impurities. PTMs, in particular, encompass a wide variety of CQAs including oxidation, deamidation, phosphorylation, and glycosylation, just to name a few (see Figure 1).
Figure 1. Potential product-related impurities. Figure is courtesy of the author.
The CQAs monitored to ensure efficacy and safety of biotherapeutics may exist naturally or may be induced at any point in production, purification, formulation, or storage. The diversity of protein characteristics, heterogeneity within those attributes, and the strengths and weaknesses of the numerous available analytical techniques mean that multiple techniques and numerous assays are needed to fully characterize a biotherapeutic and monitor all of the variants and process impurities that may affect the safety or effectiveness of the final product. Approaches to measuring CQAs depend on the attribute to be measured and on the stage of the product lifecycle, but liquid chromatography (LC) techniques dominate throughout. High-end instrumentation such as high-resolution mass spectrometers are much more commonly used for method development and initial characterization to identify chromatographic peaks, whereas liquid chromatography/ultraviolet (LC/UV) instruments are much more prevalent in quality assurance/quality control (QA/QC) environments.
While not directly measuring a CQA, a titer determination often serves as the first quality check on production of a biotherapeutic protein. An abnormal titer may reflect problems with the cell line or media that may lead to heterogeneous or incorrect product and not simply low yield. Protein A affinity capture chromatography with UV detection is the ubiquitous approach taken in biopharma for monoclonal antibodies (mAbs). Many labs opt for genetically modified, recombinant protein A platforms because the recombinant protein is generally more robust, leading to longer column lifetime. The appeal of native (purified) protein A is its tighter binding affinity for some immunoglobulins such as IgG. Protein A products are available as pre-packed columns, monoliths, and loose media that the user packs. Monolithic columns have the advantage of larger pore frits making them less susceptible to clogging, and therefore, more rugged in the face of complicated sample matrices.
An aberrant titer may necessitate analysis of the spent cell-culture media to troubleshoot the cause of low production. Amino acids are a main component of the media, and are readily analyzed by a variety of techniques, with LC/UV of derivatized amino acids being the most common. Derivatization techniques have the advantage of being widely accessible as well as the ability to introduce a degree of specificity through the reaction chemistry, but interest has been increasing in the analysis of underivatized amino acids to minimize time spent on sample preparation. Underivatized amino acids are not UV-active, and therefore, require alternate detectors such as evaporative light scattering (ELSD) or mass spectrometry (MS). The sensitivity of ELSD, however, is often inadequate, reaching only low nanomole levels, while UV detection of derivatized amino acids can reach low picomole levels. MS analytical detection can be even more sensitive, by orders of magnitude. The cost of MS instrumentation has been a barrier to widespread adoption of this approach, but increased interest has led vendors to introduce cost-effective, fit-for-purpose mass spectrometers into the market. As MS becomes more accessible, the task of quickly and efficiently separating these small, polar molecules has surfaced as a challenge to overcome. Ion-pairing with reverse-phase columns can be robust, but require dedicated instrumentation. Historically, hydrophilic interaction chromatography (HILIC) methods have struggled to meet the market’s requirements for robustness and reproducibility; however, recent column introductions have made significant improvements in that respect.
Aggregation is an essential attribute to monitor closely. It is a common stress response of proteins and can trigger an undesirable immune response. When assessing risk, high molecular weight aggregates are often assigned a high-risk priority number (RPN) because aggregation is both common and detrimental. Size-exclusion chromatography (SEC) with UV detection is the gold standard for determination of monomer versus aggregate. It is a native separation that preserves the non-covalent aggregation states and is relatively simple to operate and interpret. An example SEC–UV separation of mAb monomer, dimer, and higher order aggregates is shown in Figure 2. SEC is a separation based on the solution-size of the protein and it often serves as an approximate measure of molecular weight. Standard proteins of known molecular weight can be used to generate a curve from which the molecular weight of samples can be estimated and from there, the aggregation state inferred. Dynamic light scattering (DLS) detection allows more precise molecular weight measurements to be made. These measurements are not as precise as MS, but DLS is much simpler to couple to SEC than MS because the buffered mobile phases typically used for SEC do not interfere. Subvisible and visible particles can be excluded from the pores, and alternatives such as analytical ultracentrifugation (AUC), field flow fractionation (FFF), light scattering, or light obscuration are more appropriate.
Fragments of mAbs are also commonly analyzed by SEC, although these are typically more difficult to resolve than aggregates. Generally, a two-fold increase can be readily resolved by SEC, but resolving an intact mAb from a species that has lost a single light chain (~150 kDa versus ~125 kDa) is considerably more difficult. Pore sizes that resolve aggregates well often do not resolve fragments as well, therefore, more than one method is sometimes necessary. Smaller particle SEC columns impart higher back pressure and are more prone to clogging, but do offer higher resolution that is desirable for these fragment separations. United States Pharmacopeia methods for mAb analysis recommends capillary electrophoresis sodium dodecyl sulfate (CE-SDS) as the best-suited approach to quantifying low molecular weight species (1). Other approaches to increase the information output from an SEC separation include combining columns of different pore size in series, or working with smaller diameter columns and volatile mobiles phases to couple SEC to MS detection.
Intact protein, charge variant, and peptide mapping separations all deliver crucial information on the purity and homogeneity (or lack thereof) of a biotherapeutic product. While there is some overlap, each type of separation reveals information unique to that approach as well as practical reasons driving use of one technique or another.
Analysis of intact proteins is a means of observing the purity of the sample, both in terms of other proteins and protein fragments that may be in the sample, as well as PTMs. Reverse-phase (RP) separations with UV detection are most common because reasonable separations can be obtained between protein species. PTMs are harder to separate at the intact level, and changes from one sample to another may be subtle. RP separations readily couple to MS, where accurate mass can be measured, including the identification of a variety of PTMs. MS/MS technology has not yet reached the point of being routinely informative at the intact protein level. Electron-based dissociation techniques have offered some progress; however, top-down fragmentation of proteins remains largely inefficient. As with other CQAs discussed as follows, information gathered on the intact sample is not site-specific. The sample must be digested into smaller components for PTMs to be localized.
Hydrophobic interaction chromatography (HIC) is a separation technique that preserves the native structure of the protein as an alternative to RP protein separations that denature the protein. HIC is attracting more interest of late for its potential to separate protein oxidation as well as for the determination of drug-to-antibody ratios (DAR) in antibody-drug conjugates (ADCs). Although HIC has great potential, it has yet to see widespread adoption into biopharmaceutical labs due to the extremely high salt levels required in the mobile phase and reproducibility challenges of columns currently on the market.
Ion-exchange chromatography (IEX) with UV detection is commonly used to separate charge variants caused by PTMs such as lysine truncation, deamidation, or sialylation. This analysis is also typically done at the intact protein level, as such change can be detected, but not specifically identified and localized. IEX is most often done with salt gradients; these high concentrations of non-volatile salts are not MS compatible, hence, an emerging interest is in using pH gradients rather than salt gradients, enabling use of mobile phases buffered with lower concentrations of salts that are sufficiently volatile to be used with MS. IEX requires that samples have the opposite polarity charge from the stationary phase in order to be retained. Salt gradient IEX uses high salt concentrations to disrupt these ionic interactions and elute analytes. A pH gradient must span the isoelectric point (pI) of the analyte so that the protein will elute when its charge is net neutral. pH gradients can focus analytes into narrower bands for higher resolution than salt gradients, although linear pH gradients are difficult to generate reproducibly. Robust IEX methods can be challenging to develop because the mobile phase pH and ionic strength and gradient composition must be precisely controlled and appropriately selected. A significant amount of method development is often required, but can be facilitated with software to screen gradients with composite buffer systems made from only a handful of stock solutions.
Capillary isoelectric focusing (cIEF) is also commonly used for charge variant analysis. Similarly, to pH gradient IEX, protein variants are separated based on their pI, making cIEF a popular technique to verify IEX results.
Compared to the previously discussed means of detecting PTMs, peptide mapping is the only approach that can specifically identify and localize the modifications through LC/MS/MS. Peptide mapping primarily serves to detect sequence variants of the target protein, but is increasingly used to simultaneously quantify PTMs such as oxidation, deamidation, glycosylation, and isomerization as part of multi attribute methods (MAM). Figure 3 shows an example where peptide mapping reveals differences between an innovator and biosimilar mAb. MS/MS experiments indicate that the difference is due to a C-terminal lysine truncation. While the sample preparation to reduce, digest, and clean up a protein sample is extensive, peptide mapping can arguably offer the most information on multiple CQAs from a single experiment. Peptide mapping relies heavily on MS detection in the protein characterization stage prior to transfer to LC/UV for QA/QC. With only UV detection, it is impossible to be confident that a complete peptide map has been established. Accurate mass measurements, with MS/MS for sequence confirmation and PTM localization, are necessary to truly characterize the protein and identify CQAs. Limitations of peptide mapping include relatively low throughput because LC methods are frequently an hour or longer; selecting column chemistries for maximum chromatographic resolution while maintaining MS analytical sensitivity; and the wide dynamic range and chemical diversity of the modified and unmodified peptides.
Glycans are a unique PTM in the heterogeneity that can exist within the modification. As glycans play a significant role in cellular signaling and can influence protein conformation, variations in glycan profiles can lead to changes in efficacy and safety. Both the glycosylation sites on a protein and the glycan structures themselves can be important to characterize, leading to analysis of multiple sample types-intact protein, glycopeptides, and released glycans.
Because glycans comprise a relatively small portion of an intact protein, chromatographic separations typically reveal little to no information about the glycosylation state of an intact protein. The most significant exception to this is the measurement of sialic acid glycans using ion exchange. Mass spectrometry, however, can accurately measure glycosylation at a high level, and relative quantitation is possible. When coupled with a reverse-phase separation, both protein purity and glycosylation state can be assessed. The caveat to any of these approaches at the intact level is that site-specific modifications cannot be determined.
Released glycan analysis is typically performed using a HILIC separation of labeled glycans with fluorescence detection. Often, method development will be carried out with MS detection to confirm peak identity before the method is transferred to LC/fluorescence. While MS certainly offers more specific information than optical detection, structural characterization of glycans is still immensely problematic. MS/MS technology innovations have contributed significantly to glycan analysis, with electron-based techniques such as electron transfer dissociation (ETD) yielding more informative cross-ring cleavages than the fragmentation patterns typically observed with the more established collision induced dissociation (CID).
The analysis of glycopeptides also relies heavily on MS/MS, but glycopeptides straddle the line between being most amenable to HILIC or reverse-phase separations. Glycopeptides are more hydrophilic than most non-glycosylated peptides, making their retention and separation on the reverse-phase columns used for peptide mapping challenging. However, the peptide moieties of glycopeptides often make them difficult to retain and separate by HILIC. Mixed-mode chromatographies and two-dimensional LC combinations of separation modalities are research tools available for the characterization of glycopeptides. Analogous to the separations issues posed by glycopeptides, peptides fragment well and predictably by CID, while, as mentioned previously, ETD gives more helpful glycan fragmentation. Hybrid ETD/CID techniques are at the leading edge of unknown glycopeptide characterization.
The inherently aqueous nature of biology has caused solution-phase techniques to dominate CQA analysis of biotherapeutics. LC/UV has been the cornerstone of CQA analysis, and the technique is not expected to diminish anytime soon because of its accessibility in terms of cost and the required user expertise. Nonetheless, as regulatory demands become more stringent and biotherapeutics become more complex, techniques such as light scattering and MS that offer more information and higher confidence are gaining traction. Once only found in early stage research and characterization settings of biopharma companies, they are gradually finding their way into downstream QA/QC settings as the ratio of benefit to cost and required skill increases.
MAM are an exciting direction for CQA monitoring-up to six assays may be multiplexed into a single LC/MS/MS method. In addition to the PTMs mentioned in the peptide mapping discussion, it is also possible to measure process impurities such as host cell proteins. While the required investment in a high-resolution mass spectrometer and considerable expertise is a possible drawback, the potential time and cost savings of a single assay that confirms protein identity, measures sequence variants, clips, charge variants, glycans, other PTMs, and process impurities is an opportunity that warrants attention.
1. USP, <129> Analytical Procedures for Recombinant Therapeutic Monoclonal Antibodies, USP 40–NF 35 (USP, 2017).
Note: This article is for research use only and not for use in diagnostic procedures.
Vol. 31, No. 2
When referring to this article, please cite it as A. Blackwell, “The Ins and Outs of LC-Based Analytical Tools and Techniques,” BioPharm International 31 (2) 2018.