Uncovering hard-to-detect impurities in monoclonal antibody preparations
For Research Use Only. Not for use in diagnostic procedures.
Comprehensive identification of host cell proteins with mass spectrometry
Manufacturers of monoclonal antibodies (mAbs) can now quickly identify and quantify almost-imperceptible levels of host cell protein (HCP) contaminants—even at concentrations in the sub-ppm range1—in just a single injection. In a sample dominated by the therapeutic protein-of-interest, finding minor HCP contaminants from amongst milligrams of mAb product can be a challenging prospect. SCIEX provides a solution applicable to all stages of biotherapeutic development: comprehensive, unbiased detection over a broad dynamic range coupled with relative quantitation of process-related impurities in a very short analytical run.2-4
Biopharmaceuticals offer a novel and hopeful approach for treating disease; however, the production of these drugs is complex, requiring multiple steps to achieve an overall level of purity that permits their safe use in humans. Typically, biotherapeutics are generated using recombinant DNA technology, and manufacturers must isolate these drugs away from the complex biological matrices in which they were expressed (e.g., yeast, CHO cells). HCP contamination is often unavoidable, and low-level, background impurities remain or co-purify with mAbs even after multiple rounds of chromatography. Therefore, the full characterization of a biotherapeutic product must include a complete inventory of any residual HCPs. In addition, having a robust HCP detection strategy is essential for gauging a drug’s potential to elicit immunogenic reactions (or other serious side effects) that can ultimately lead to product withdrawals or drug shortages.
“We definitely want to make sure that our customers know exactly how well their purification is working,” said Eric Johansen, SCIEX global technical marketing manager for biologics, who developed a generic approach for the comprehensive detection of low-level HCPs. “It is really important that the manufacturer finds out first [what the HCP profile is] and corrects the process to ensure that product quality is maintained.”
Finding out about product impurities through poor clinical trial outcomes, adverse patient reactions, and convoluted scaling-up processes may lead to delays that jeopardize the commercialization of a new therapy. Regulatory approval of a new mAb also depends on a complete understanding and monitoring of process-related impurities, pushing biopharmaceutical developers to seek out ever more rapid, unbiased approaches for the comprehensive analysis of HCPs.
Establishing a method capable of rapidly detecting all sample components over a broad dynamic range is a tall order, but Johansen’s team arrived at a generic strategy based on MS/MSALL and SWATH® Acquisition. Enabled by a hybrid TripleTOF® LC-MS/MS System that combines fast acquisition speeds with high-resolution detectors, researchers were able to capture information on every protein in a purification sample, even low-abundance, minor contaminants.1-4
“We wanted a way to identify everything in a sample with great certainty and provide the sensitivity and dynamic range at high speeds in a chromatographic time scale that every analyst can access,” explained Johansen. In essence, using SWATH® acquisition allows researchers to discover very minute amounts of impurity by acquiring MS/MS information for each peak, and doing this for all sample peaks at high resolution during the time it takes for a digested protein mixture to elute through a chromatography column. Recent advances in chromatography (based on capillary electrophoresis electrospray ionization (CESI) paired with LC separations) now deliver sub-ppm levels of detection in post-filtration mAb purification samples, permitting an exceptionally broad dynamic range for HCP analysis.1
Using a model mAb spiked with varying levels of a commercial protein mixture,* Johansen’s team demonstrated how HCPs can be detected and quantified at low, ppm levels.2-4 Because researchers cannot predict the identity of contaminants, the use of targeted quantitation methods must be limited until after HCP profiles are determined. SWATH® acquisition, however, captures all ions in a sample, providing information on every component, whether the peak is known or unknown (Figure 1). During SWATH® analysis, all peaks within a given mass window undergo fragmentation at the same time, and all resulting fragments travel en masse to the detector, thereby providing the comprehensive and high-resolution mass information needed for the identification and quantitation of each peak (Figure 1). “What you are able to grab is MS/MS chromatograms of every single fragment ion, and unlike an IDA (information-dependent acquisition) experiment where you get only one MS/MS spectrum, you get a lot more information by processing all MS/MS spectra from the whole chromatogram,” said Johansen.
In cases where the protein contaminants are entirely unknown, researchers must first understand which signals are impurity-related and which stem from the product by generating libraries of tryptic HPC and mAb peptides using IDA analysis. Null or blank harvest samples go through the same steps of protein purification and represent the potential cohort of background HCP contaminants. For Johansen’s experiments, he simply digested the commercial protein mixture to represent these background HCP signals, creating a reference library of background peptides for use in subsequent experiments.
Purified biotherapeutic fractions are chromatographically separated prior to analysis by mass spectrometry, an approach that requires only one analytical run. The data are aligned with peptides identified from libraries (via extracted ion chromatograms), and peaks corresponding to impurities are quantitated (Figure 2). What Johansen saw was that although there was a 100,000-fold difference between the mAb and contaminants, he was able to obtain highly reproducible quantitative data with low percentages of variation for each model HCP2-4 (Figure 3). Even in the low ppm range (7-30 ppm), measured peptides displayed, on average, CVs of only 7%. The specific protein type did not affect the variability of the SWATH® data (Figure 3), a result that differs from other protein-detection methods—such as ELISA—where protein binding is biased towards particular structural characteristics. “The key thing to point out here is that we can do SWATH® acquisition in 30 minutes with no complex chromatography using a single, easy-to-build method,” remarked Johansen. “Detecting 10-25 ppm levels in a single run will meet a lot of criteria that manufacturers will want in HCP identification.” (For the complete quantitative data on the levels of each model protein contaminant, see Figure 4).
SCIEX’s new rapid, HCP-detection approaches are poised to revolutionize the production of mAbs and other biotherapeutics—in a high-throughput, reliable, and sensitive way. The improved product quality and time savings resulting from the simultaneous cataloguing and quantifying of all residual background proteins will have a profound effect on the overall efficiency of biotherapeutic manufacturing. Going forward, Johansen and his team will continue to enhance applications for low-level HCP analysis, remarking, “This is really, really exciting technology and is something we believe in. We think this is a really powerful system for biologics and that it will really help everyone in the long run.”
*The SWATH® methodology has successfully detected HCPs in several commercial antibody purification processes, but due to confidentiality agreements, the data cannot be shared.
Submitted by Laura Baker, freelance science and technical writer
Figure 1: Comprehensive identification of sample components using MS/MSALL and SWATH® Acquisition.2,4 Data-independent acquisition with SWATH® Acquisition uses a wide Q1 isolation window (20 Da in the host cell protein experiments) and then rapidly steps across the entire m/z range within the available cycle time (top panel). The wide window permits multiple analytes to pass through Q1 concurrently, resulting in very complex MS/MS spectra of high resolution fragment data. After acquisition, extracted ion chromatograms (XICs) for multiple fragments collected from a single analyte give high quality peaks for quantitation, similar to those obtained from multiple reaction monitoring (MRM) methods (bottom panel).
Figure 2: SWATH® Acquisition detection of multiple fragment ions from a serum albumin peptide.3 The peaks above represent extracted ion chromatograms (XICs) for the fragments from a single serum albumin peptide (SLHTLFGDELCK). Multiple peptide fragments (denoted by different colors) were analyzed over time to provide multiple reaction monitoring (MRM)-like data at several levels of spiked protein: A) 413 ppm, B) 207 ppm, C) 104 ppm, D) 52 ppm, E) 26 ppm, and F) a negative control with no peptides.
Figure 3: Levels of host cell protein detected. 2-3 A range of concentrations of each model protein were measured by quantitating a representative peptide in triplicate (Levels 1-5 shown, top). The coefficient of variation (CV) across a range of protein levels (1-5 shown, bottom), was consistently low, and even the lowest concentration range (level 5, 7-30 ppm) displayed CVs under 12%.
Table 1: Levels of added model host cell proteins in a monoclonal antibody preparation.3 Tryptic digests of six model proteins from a commercial mixture were spiked into a tryptic digest of pure IgG (10 μg) over a range of concentrations (levels 1-5). The proteins were detected and quantified using the SWATH® Acquisition data-independent acquisition method, and each sample was acquired in triplicate at each level.