Whenever cells are used to produce biotherapeutics such as antibodies, process impurities known as host cell proteins (HCPs) find their way into downstream product flows. Some HCPs are benign. Others are harmful. But which are which? There’s one way to know for certain: detect and assess all HCPs. Once the harmful HCPs are identified, they can be removed by introducing the appropriate purification measures. Typically, these measures are called upon to remove the vast majority of HCPs.
Historically, most HCP monitoring served to affirm that bioprocesses remained stable, run after run. According to Eric Bishop, vice president of research and development, Cygnus Technologies, HCP clearance results just had to be consistent to be reassuring. Then, in the early 2000s, attitudes began to change. “We realized that some of the adverse events we were seeing in the clinic were actually due to HCPs,” Bishop recalls. “Regulatory agencies became much stricter about monitoring HCPs and making sure you had the right assay.”
Today, HCPs are considered a critical quality attribute (CQA). “HCPs need to be minimized and closely monitored,” notes Parul Angrish, PhD, director, biopharma/pharma market, Agilent Technologies. “HCP identification and quantification are needed for regulatory submissions.”
Building on the gold standard
“The HCP ELISA,” Bishop relates, “is the gold standard method for monitoring HCP clearance throughout your purification process and also as your lot release test.”
“The plate-based ELISA for HCP analysis can be used throughout the entire process. However, there are other promising future alternatives as well, such as real-time and label-free biosensor technologies,” says Fredrik Sundberg, global director of strategic customer relations, Cytiva. “For example, surface plasmon resonance detection holds a great promise as process analytical tool for rapid in-line monitoring of both drug substance and impurities.”
The polyclonal HCP ELISA is not an assay for measuring a single analyte. Instead, a single polyclonal assay is used to measure potentially thousands of HCPs. It’s really important that the selected assay be appropriate for monitoring the process, Bishop observes. In other words, the polyclonal antibodies must be broadly reactive to the HCPs that are found in the process and in the drug substance. Ensuring that this is the case can be very challenging because some HCPs are very immunogenic and elicit a strong immune response very easily, whereas other HCPs, such as housekeeping proteins and other highly conserved antigens, are weakly immunogenic. Once generated, the reagents are typically affinity purified to get rid of any nonspecific immunoglobulin G antibodies.
The next step is to demonstrate that the antibodies used in the ELISA are broadly reactive to the HCPs. It is necessary, Bishop points out, to ask, “How much of the total HCP that is in your starting material do your antibodies react to?” Answers to this question are found through “coverage analysis.”
A traditional coverage analysis technique involves running the harvest material—without the product—on a large-format 2D western blot and on a large-format 2D silver-stained gel, and comparing the spots on the two. The silver stain should contain all the HCPs that can be there, while the western blot should contain the subset of immunoreactive proteins. Differential in-blot gel electrophoresis (DIBE) is an alternative to the classical western blot.
Bishop adds, however, that the large-format 2D western blot presents “a lot of limitations,” including transfer inefficiencies and the need to denature the sample. The latter limitation occasions a concomitant loss of conformational epitopes.
In 2013, Cygnus devised an alternative technology—antibody affinity extraction (AAE). It is an immunoaffinity method in which HCP antibodies are immobilized on a chromatography support. “We pass the harvest material over that column so that all the antibody-antigen reactions are happening in a more natural environment,” Bishop explains. At the end, there are two fractions: the starting HCP from the clarified harvest, and the antibody elution fraction. Bishop indicates that the first fraction “represents all the HCPs that are in the process and could potentially find their way into your product,” and the second fraction represents “all of the HCPs that the antibodies specifically removed from the sample.”
As powerful as the traditional HCP ELISA and its elaborations are, “ELISA lacks the specificity and coverage to identify and quantify individual HCPs,” Angrish notes. To get around these limitations, it is necessary to use alternative HCP analysis technologies. One of them is liquid chromatography–mass spectrometry (LC-MS).
The main challenge for LC-MS analysis is the coelution of low-abundance HCP peptides along with high-abundance peptide drug products. Angrish says that detecting these “hitchhiker proteins” requires better separation of peptides and an LC-MS system that has a broad dynamic range.
“The benefit of physical methods, as opposed to immunological methods, is that you don’t need a specific antibody,” Sundberg observes. MS, for example, may be used to create impurity profiles that can inform purification efforts. However, MS may have issues of sensitivity and also interference from the matrix and the drug product itself.
Impurity analysis diagram
Impurity analysis with conventional ELISA technology typically involves labor-intensive protocols, low-throughput operations, and narrow dynamic ranges. To overcome these limitations, Gyros Protein Technologies has developed technology that miniaturizes and automates impurity analysis. The company says that its Gyrolab CD enables the parallel processing of 96 or 112 microstructures coupled with automated laser-induced fluorescence detection. The flow-through affinity microcolumn can eliminate incubations and reduce matrix interference, and the nanoliter-scale format can lower reagent and sample volumes. The technology, Gyros declares, is cost effective and gives highly reproducible data over broad dynamic ranges.
LC-MS and other physical measures should serve as orthogonal assays, identifying and quantifying HCPs in a process step or even the final product, but they should not be used as replacements for ELISA. “The classical immunoassay is still the gold standard for HCPs,” Sundberg declares. “And it’s actually expected by the regulatory agencies.” He adds that ELISA has the sensitivity to quantify levels of protein in parts per million.
“With ELISA, we could see past the drug substance and get a reasonable idea of the HCP quantities, but we really had no idea of the HCP identities,” Bishop says. “But now with our AAE method, the chromatography column serves as an affinity column, selectively enriching HCP and depleting drug substance,” as well as setting the stage for MS analysis.
Unless AAE or some other sample preparation technique is used, MS can be pretty useless in the HCP analysis of drug substance samples. A typical sample might contain a high concentration of normal monoclonal antibody (100 mg/mL) and low concentrations of HCPs (ng/mL). So, it would be next to impossible to “see” those HCPs due to the drug substance saturating the detector.
Now that you know
Bishop asserts that it has become possible to determine the “wholesale proteins” that a drug substance contains. “We’re working toward making it so that companies can do realistic risk assessments based on that information,” he continues, “but right now, it’s still kind of challenging because not a lot of companies are sharing data.” At present, it’s usually unclear which HCPs or what HCP levels in a formulation may lead to different kinds of problems—problems such as toxicity, immunogenicity, and interference.