OR WAIT 15 SECS
Volume 31, Issue 8
More complex biologic samples must be evaluated to ever higher levels of specificity and sensitivity.
Process- and product-related impurities must be evaluated according to various regulatory guidelines during production and to enable final product release. Impurities can arise from the biological samples themselves or from the process of developing biologics, including handling of materials.
Sample-related impurities include residual host cell-derived proteins (HCPs) and nucleic acids, complexes or aggregates of the biologic (high-molecular-weight [HMW] proteins), and clipped species and half molecules (low-molecular-weight [LMW] proteins). Impurities from cell-culture media can include inducers, antibiotics, and media components.
Impurities that come from downstream processing can include microscopic particulates, metals, and any materials that have carried over from the purification process, including resin particles, surfactants, emulsifiers, and viral-inactivation agents. Biological contaminants derived from handling include mycoplasma, bacteria, and virus particles.
Some of these impurities have known structures, while others may be only partially characterized or completely unknown. Post-translational modifications such as glycosylation and phosphorylation, degradation via oxidation or deamidation, and disulfide bridge scrambling (misfolding) can occur during upstream or downstream processing or storage under inappropriate conditions, resulting in large numbers of possible impurities. Disposable equipment and plastic tubing, stoppers, and containers may be sources of leachables. For antibody-drug conjugates, free drug cytotoxins can be problematic.
The decision on whether to monitor these impurities, and to what levels, is generally risk-based, using knowledge from both analytical and biological assays, and any preclinical experience to assess the impacts of each impurity on the safety, efficacy, or stability of the biotherapeutic, according to Scott Berger, senior manager for biopharmaceutical markets at Waters Corporation.
Monitoring biologic production processes and analyzing products for release testing can be challenging for many reasons. For Jean-Francois Boe, scientific director of SGS Life Sciences, the greatest challenge is the vast number of potential impurities that can be formed when all of the possible chemical modifications that can occur are considered. “Tens of millions of combinations of impurities can be formed, many of which have significantly different physical and chemical properties. One unique analytical technique cannot be used. A number of appropriate analytical methods must be used to create as full a picture as possible of the impurities that are present,” he explains.
“Purification of biologics is often a multi-step process, and there is no one-size-fits-all analytical methodology,” adds Tiffani Manolis, director of global pharma segment marketing with Agilent Technologies. “As a result, analysis of residual impurities is often a time-consuming activity.”
Another major challenge when developing methods to evaluate bioprocess residuals is matrix interference, according to Jon S. Kauffman, president of Eurofins Advantar Laboratories, a member of Eurofins BioPharma Product Testing.
“Developing a robust method for certain impurities is always a challenge. For most of the methods that support in-process or release testing of drug substances, both matrix effects and the presence of a high concentration of product are the main factors which can impact the performance of methods,” agrees Jun Lu, director of analytical development at Catalent Biologics.
Matrix interference can be caused by components in the formulation buffer that interfere with the detection of the residual by suppressing the ionization in the mass spectrometer or from the residual binding to the protein, according to Kauffman. “Further,” he says, “we are typically required to monitor these residuals in various steps throughout the bioprocess. The sample matrices from each step can be quite different and each pose a challenge with respect to interferences and sample preparation.”
Complicating the situation is the fact that many product-related impurities need to be monitored down to low-percentage, or fractional-percentage levels, straining traditional optical, ultraviolet (UV)-based peptide mapping assays, according to Berger. “Increasingly, this necessitates the use of liquid chromatography-mass spectrometry (LC-MS) analysis to obtain the additional the selectivity and dynamic range for detection and monitoring of critical impurities. In addition, some impurities such as clips and unfolded variants may require multiple techniques for efficient detection and quantification, because peptide level analysis is often uninformative for these structures,” he observes.
Some impurities, such as surfactants, often exhibit a broad rather than a sharp peak and can interfere with each other, making specificity difficult to obtain. “For example,” notes Kauffman, “it is virtually impossible to detect poloxamer 188 in a drug substance/product that is formulated with polysorbate. In these instances, we are forced to go backward in the manufacturing process to the step prior to addition of polysorbate.”
Other challenges include the need to derivatize LMW compounds before analysis, as well as the ability of some residuals to adhere to the surfaces used during sample preparation, and the instability of others. Understanding these possible issues when developing methods is extremely important, according to Kauffman.
As biopharmaceutical production processes evolve, and with the complexity of new process matrices, the detection and tracking of residual impurities is becoming increasingly difficult and may require various orthogonal techniques, says Vincy Abraham, director of biologics, Catalent Biologics.
Liquid chromatography and electrophoresis remain the two main separation techniques, and immunochemical assays remain unavoidable in specific cases for the evaluation of low levels of residual impurities, according to Boe. He notes that while little has changed with these separation technologies, there are many more advanced detection methods available today. UV or visible light and infrared (IR), fluorescence, mass spec, light scattering, and more have improved capabilities.
Other separation methods include gel-permeation, size-exclusion chromatography, ion exchange chromatography, and gas chromatography. For Kauffman, LC-MS/MS performed using a triple-quad mass spectrometer connected to an ultra-high-pressure LC (UHPLC) system is the technique of choice given its sensitivity, specificity, and ability to provide quantitative results. “This instrumentation is required in most cases to be able to quantitate at the ng/mL or even pg/mL range at which residuals must be evaluated,” he says. HPLC and UHPLC are, however, still used with UV, charged aerosol, or evaporative light scattering detectors for compounds of interest in the µg/mL range or higher that do not ionize.
Detection by mass spectrometry is particularly useful for evaluating residual impurities formed due to chemical modification of the biologic drug substance, according to Manolis. Depending on the specific species of interest, MS can be coupled with LC, gas chromatography, matrix-assisted laser desorption ionization (MALDI), and electrospray ionization (ESI).
The most common method for screening biopharmaceutical products and testing for HCPs is enzyme-linked immunosorbent assay (ELISA), a sensitive assay with a low detection limit, high level of reproducibility, and compatibility with high-throughput screening, according to Laura Moriarty, marketing manager for Bio-Rad’s Drug Discovery and Development Group. She notes, though, that because the ELISA technique does not permit identification of antigens when using mixtures of antibodies, but only provides titers, the accuracy and utility of ELISA relies on a thorough prior assessment of the antibodies used. “Accurate evaluation and validation of antibodies reactive against HCP is crucial for detecting and monitoring HCP both during the product development cycle and during manufacture of biologics,” she says.
The predominant method for assessing anti-host HCP antibodies involves 1-D or 2-D electrophoresis followed by western blotting, according to Moriarty. For polypeptides with similar molecular masses in complex mixtures of proteins, 2-D electrophoresis gives much better resolution because it separates proteins in the first dimension by isoelectric point (pI), followed by molecular mass in the second dimension. Once a good purification system has been established, the final product can be routinely screened with an ELISA to make sure that impurities are continually removed from the samples.
For nucleic acid screening, quantitative polymerase chain reaction (qPCR) and droplet digital PCR are used to detect and signal the presence of nucleic acids in a sample. Mycoplasma can also be detected using PCR, as well as colometric enzyme assays. Bacteria can be detected using endotoxin testing via the limulus amebocyte lysate assay, the United States Pharmacopeia (USP) chromogenic method, and the gel-clot method. The types of viral strains to be tested are specific to the method used to manufacture a therapeutic or biological.
Biologic aggregates are typically detected using sedimentation velocity analytical ultracentrifugation (SV-AUC), size-exclusion chromatography coupled to multi-angle light scattering (SEC-MALS), or dynamic light scattering (DLS) for the analysis of quaternary structures. DLS, as well as resonant mass measurement (RMM), can also be used to detect microscopic particulates, according to Moriarty.
Because there are so many different types of manufacturing processes and residual impurities from low to high molecular weight with varying chemical and physical properties, identifying multifunctional methods that can separate and detect more than one type of impurity is essential for developing optimized methods. “Mass spectrometry is becoming attractive in part for this reason; a mass spectrometer can be used for the detection of numerous different impurities well chromatographically separated or co-eluted in a single chromatographic run,” Boe states.
Mass spectrometry has become the primary analytical technology applied to multiplexed analyte detection within complex samples, agrees Berger. “The additional selectivity of the mass dimension enables detection and higher dynamic range quantification of analytes, even in the presence of co-eluting species. This methodology is now starting to be applied within biopharmaceutical development against a list of targeted product or process impurities,” he observes.
Recently there has been a lot of work done using LC–MS for multi-attribute monitoring method (MAM), which is designed, according to Manolis, to provide simultaneous detection, specificity, identification, quantitation, and monitoring of attributes that are relevant to safety, efficacy, and the overall quality of drug. “MAM provides residue-specific identification, quantitation, and better understanding of any post-translation modification when compared to traditional methods of analysis, thus improving overall operational efficiency, resource consumption, and time required,” she comments.
As long as the transitions monitored are distinct for each compound with little to no cross talk, Kauffman agrees that newer LC–MS/MS systems and software suites allow the detection of multiple impurities at once. “The challenge in these situations is the sample prep. Often times when you optimize a method for multiple analytes, it works really well for some analytes but not for others. Finding the right sample prep that extracts all the analytes of interest can be quite challenging. Methods for sample cleanup often work for one sample matrix but not another. As a result, the rule of one analyte per method is still the preferred approach so that the method can be optimized for the analysis of that particular analyte,” he says.
“With ever-increasing regulatory and compendial stringency to identify, quantify, and monitor impurities, a greater emphasis is being placed on their characterization and analysis at trace levels,” asserts Abraham. “Fortunately,” she continues, “there have been parallel advancement in technologies that allow rapid characterization of impurities at levels of approximately 0.1%.”
To alleviate some of the limitations with ELISA, for instance, Abraham notes that several technologies for quantitation exemplified by Gyrolab, AlphaLISA, and Octet have emerged in the past decade as viable alternatives for HCP. Each represents a different strategy for HCP quantitation.
Bio-Rad recently introduced droplet digital PCR (ddPCR) as a sensitive (picogram range sensitivity in milligrams of recombinant vaccines) and quantitative method for quantification of residual host-cell DNA, according to Madhuri Ganta, senior global product manager in Bio-Rad’s Digital Biology Group. With ddPCR, a sample is partitioned into 20,000 nanoliter-sized droplets, which makes the PCR reaction less susceptible to inhibitory substances. Unlike with qPCR, extraction of total DNA from the protein drug sample is not required; intermediates can be processed directly, and absolute quantification is possible without the need to establish a standard curve, according to Ganta.
While optical-based LC assays are still highly desirable due to the lower system cost and broader organizational accessibility of this technology, Berger observes that the increasing complexity of modern biopharmaceuticals has pushed laboratories to adopt more resolving and sensitive UPLC- and UHPLC-based separations platforms for these newer products. He adds that the additional adoption of mass detection to increase selectivity and dynamic range of these assays has been growing within regulated development and is now starting to appear in quality control for targeted monitoring of product and process attributes and impurities.
The use of mass spectrometry for the characterization and quantification of HCPs is an active area, according to Yunsong (Frank) Li, director of biologics process development at Catalent Biologics. “MS can detect the HCPs not covered by anti-HCP reagents and provide additional information such as molecular weight, theoretical isoelectric point (pI), and immunogenicity potential,” he explains. ProteinSEQ technology (Thermo Fisher Scientific) has also recently been demonstrated to quantify HCPs in a much wider dynamic range than ELISA, according to Li. The combination of ion exchange (IEX)-HPLC and high-throughput western blot is also under development for quantification of low immunoreactive HCPs.
For detection of aggregates, Li adds that nanoparticle tracking analysis (NTA) can track nano-sized particles via particle-scattered light from a focused laser beam. “The system can track many individual particles and therefore count the number of particles. From the rate of the particles’ Brownian movement, the size can also be calculated,” he says. Flow cytometry, traditionally used for cell counting, has also been developed to count the protein aggregation particle size as low as 0.2 µm.
In other areas, traditional sodium sodecyl sulfate-polyacrylamide gel electrophoresis (SDS–PAGE) is being replaced by capillary electrophoresis (CE-SDS) because it provides superior detection, reproducibility, and robustness, according to Manolis.
Another development, according to Abraham, involves a shift from the conventional protocol of isolation and spectral analysis to online analysis using sophisticated modern hyphenated tools, such as GC-MS, LC-MS, CE-MS, supercritical fluid chromatography-MS (SFC-MS), LC-nuclear magnetic resonance (LC-NMR), CE-NMR, and LC-Fourier-transform infrared spectrometry (LC-FTIR).
Separately, Berger points out that the use of automation for sample preparation is greatly increasing within development and quality control organizations. “In development, this automation often supports higher-throughput clone selection and quality-by-design (QbD) studies, but increasingly the reason for adopting automated sample preparation is the improved consistency of sample generation versus manual workflows. The need for a mid-tier scale of automation has become apparent,” he says.
Indeed, improving the efficiency and reducing the costs associated with residual impurity analysis, which is essential to improving the overall efficiency drug development and manufacturing, requires that workflows be amenable to automation and high-throughput protocols, agrees Moriarty.
Eurofins is typically required to resolve three primary problems that are interconnected: quantitation limits, interfering compounds, and extraction of analytes of interest. “Interfering compounds and poor extraction of the compounds of interest directly affect the quantitation limits of the methods. Mass spectrometry for the most part eliminates co-eluting peaks because we can focus in on a mass transition for the compound of interest, but there are still times when compounds share the same transitions or have cross talk with transitions from other compounds. Extraction techniques have evolved over time especially with the addition of molecular weight cut-off filters and solid-phase extraction cartridges, but the more you manipulate the samples, the more chance you have to introduce error and contamination,” explains Kauffman.
One challenge is the high variability in the process and sample matrix, which can contribute to out-of-specification/out-of-events, which are often time-consuming and costly, according to Manolis. Standardization of specifications for critical reagents and simplified and reproducible processes for sample digestion are also needed. For multi-impurity detection methods such as MAM, Manolis notes that improvements in systems for data processing, handling, and interpretation are needed.
Boe points to the current gap in the ability to accurately characterize and mostly quantify particles (aggregates) that are between several hundred nanometers up to 1 micrometer in size. For HCPs, he notes that the need to switch from commercial kits for HCP analysis to custom-developed methods once a candidate reaches Phase III trials is time consuming.
Currently, the greatest limitation for process-related impurity is analytical technology for HCP analysis, with the major challenge in coverage from existing anti-HCP polyclonal antisera standards, according to Li. The current approach is to develop product- and process-specific assays, which often require long lead times of at least 18 months, or combine multiple existing anti-HCP polyclonal antisera standards.
A general key challenge has been increasing the usability of more informative and complex modern analytical technologies to enable non-specialists to continue to perform these analyses, according to Berger. “While those charged with product characterization are always welcoming greater performance envelopes of their instruments, those charged with product monitoring and release now tend to be focused on minimizing user interactions with their systems and maximizing quality and reproducibility of the results,” he says.
In addition to establishing methods that meet requirements for sensitivity and specificity, there are other factors that are important to consider. “It is essential to first determine the appropriate acceptance criteria and then ensure that methods can be readily transferred from R&D to commercial production. They should be robust, accurate, and precise, as well as easy to implement on equipment that will be available at the manufacturing plant,” Boe asserts.
A validation process that makes sense is also important, as is the need to consider the biological activity of product-related impurities. “Some impurities that are closely related to the product may have the same biologic activity as the drug substance, and therefore may not impact the safety and efficacy of the product. It may be reasonable to classify these compounds as related substances, rather than residual impurities,” Boe explains.
Vol. 31, No. 8
When referring to this article, please cite it as C. Challener, "Expectations for Residual Impurity Analysis Continue to Rise" BioPharm International 31 (8) 2018.