So long as living cells serve as tiny drug-making factories, they will, like much larger factories, generate unwanted byproducts. As industrial byproducts go, the cells’ byproducts, which are known as host cell proteins (HCPs), are inconspicuous. They are also highly variable. That is, different HCPs and different concentrations of HCPs may be generated if processing conditions vary.
The variability of HCPs presents an opportunity to bioprocessors: During development, bioprocessors can vary processing conditions to minimize HCP production and/or optimize HCP removal. This opportunity, however, will remain unrealized if bioprocessors allow HCPs to remain inconspicuous. Fortunately, bioprocessors can take advantage of new tools and techniques that are making it easier to detect and quantify HCPs.
Adopting the new tools and techniques requires a fair degree of sophistication. This is unavoidable because the drug output of living cells will always be harder to purify than the output of chemistry-based systems.
In chemistry-based systems, process-related contamination can be prevented if high-quality reagents and ingredients are used and if reaction conditions are carefully controlled. And should contaminants such as endotoxins and heavy metals enter product streams, they can be removed during post-synthesis processing without too much difficulty. The products are small-molecule drugs, which are less sensitive to the rigors of purification than are the protein-based drugs produced by living cells.
Not only are protein-based drugs more sensitive to purification stresses than are small-molecule drugs, but they are also more sensitive to contaminants. That is, the protein-based drugs and the contaminants may interact with each other. For example, enzymatically active HCPs can degrade therapeutic proteins, thereby lowering production yields and reducing shelf lives.
Unfortunately, HCPs are capable of much greater harms. HCPs that make their way into final products can cause adverse reactions or provoke dangerous immune responses in patients. Though inconspicuous, HCPs are far from innocuous.
Knowing your enemy
HCPs are unavoidable because production cell lines also express their own proteins. To help ensure that HCPs won’t make their way into the final product, one may begin by identifying and characterizing the HCPs in bioprocessing fluids.
“The composition and abundance of HCPs are closely associated with the host expression system itself and other factors during manufacturing process,” says Tingting Jiang, PhD, senior scientist, Merck & Co. “So, changing the process—for example, cell lines, cell culture media, and purification steps—can lead to different HCP profiles. Biopharmaceutical companies usually go through bioprocess improvement or optimization to increase titers and yields and enhance process robustness while following best practices to meet regulatory expectations regarding product quality.”
Reconsidering standard assays
ELISAs—enzyme-linked immunosorbent assays—are tried-and-true analytical tools that may be used to determine which HCPs are likely to be present in a biopharmaceutical product (Zhu-Shimoni et al. Biotechnol. Bioeng. 2014; 111: 2367–2379). Creating an ELISA is relatively straightforward: Proteins are collected from the unmodified cell line; the proteins are injected into an animal; the animal responds by producing antibodies to each protein; the antibodies are bound to a surface; the surface has a culture fluid passed over it; and the antibodies on the surface interact with the HCPs in the fluid.
In practice, rather than developing ELISAs in-house, most drug companies use commercial assays designed to work with commonly used production cell lines. However, commercial ELISAs can detect only those proteins that correspond to the assay’s constituent antibodies. Jian warns that if the production process involves a proprietary or modified cell line, or culture conditions that impact its expression profile, commercial HCP ELISAs are unlikely to be fully representative.
“There are multiple challenges and limitations for HCP ELISA methods,” she notes. “Development of orthogonal analytical methods independent of immunoreactivity can help overcome the limitations of HCP ELISAs and obtain a comprehensive analysis of the HCP population.”
Detecting HCPs with LC-MS
One example of an orthogonal approach is liquid chromatography–mass spectrometry (LC-MS). In contrast to ELISAs, LC-MS-based methods can provide specific HCP identifications and illustrate relative abundance by peak area without bias. ELISAs, in contrast, can provide only bulk HCP quantitation and offer no differentiation of individual protein impurities.
Furthermore, developing an LC-MS method takes only weeks rather than the years required by ELISAs. For specific high-risk HCPs, a targeted LC-MS-based multiple reaction monitoring (MRM) quantification approach can be implemented to monitor their clearance during the purification train.
“LC-MS comparison of HCP process clearance and populations can demonstrate the comparability of HCPs before and after process changes and ensure a robust commercial control strategy,” Jiang explains. “Incorporation of LC-MS-based HCP identification leads to greater process understanding, ensures the development of robust and well-controlled bioprocesses, and helps mitigate safety risks when developing and commercializing new therapeutic drugs.”
The value of LC-MS for HCP analysis is also recognized by Midori Greenwood-Goodwin, technical development scientist, Genentech (Walker et al. mAbs 2017; 9: 654–663). “LC-MS/MS methods are used to positively identify and monitor up to thousands of individual HCPs present in different process development samples,” she says. “LC-MS/MS methods can be qualitative (identification only) or quantitative.”
The idea, she says, is to combine the approach with traditional ELISAs: “Together, these methods provide a more complete picture of the HCPs that may be present in our final product.” That is, HCP identities and quantities may be determined.
The key benefit is that the approach reduces the number of preparatory steps required compared with ELISAs alone. “The main advantage in using LC-MS/MS methods is that they are agnostic,” she explains. “We don’t need to know in advance what HCPs might be present to make positive identifications. As such, LC-MS/MS methods have been essential for the discovery of proteins that are difficult to remove during purification.
“These methods can also help us establish better controls for our products and set specific limits around HCPs without the need for costly ELISA development. Overall, MS methods are faster and cheaper to develop, but more costly per a sample test.”
Another advantage of the approach is that it does not require specialist systems. “There are multiple LC and MS systems that can be used for this work,” Greenwood-Goodwin points out. “We can modify the flow rate, mass-to-charge windows, and other instrument parameters to achieve robust HCP identification or quantitation.”
Exploring enrichment strategies
The need to detect HCPs present in low quantities is another active area of process development innovation. One increasingly common approach is to discard proteins present in high concentrations and enrich those present in small quantities. This approach, which is called enrichment, was described in a paper published last year (Chen et al. Anal. Biochem. 2020; 610: 113972). One of the paper’s co-authors, Hui Xiao, PhD, a senior staff scientist at Regeneron Pharmaceuticals, shared details with GEN. According to Xiao, MS-based methods that incorporate enrichment can detect low-abundance HCP impurities despite the massive dynamic range of protein species present in antibody products.
In the paper, Xiao and colleagues described how HCPs in antibody drug substances could be identified by “applying ProteoMiner enrichment with optimized conditions followed by shotgun proteomic analysis.” ProteoMiner, a platform from Bio-Rad Laboratories, can be used to monitor individual HCP levels for in-process drug products.
Enrichment, Xiao asserts, “is a more sensitive method to detect low-abundance HCPs.” And there are other benefits. “Cost is low for each sample—less than $100 for each sample preparation,” Xiao notes. “We regularly use it for detecting low-abundance HCPs.”
Facing cell and gene therapy challenges
The emergence of commercial cell therapies in recent years has opened up the possibility of treating diseases at their root cause. However, from a process development perspective, cell therapies come with new challenges, particularly in terms of HCP analysis, detection, and removal.
“Cell therapy drug substances usually contain far more HCPs compared with fully purified antibody therapies,” says Yiwei Zhao, PhD, an associate scientific fellow at Takeda Pharmaceuticals. He adds that scientists tasked with keeping cell therapies need to assess other potential sources of protein contamination, such as those from viral vectors or media components used in the process.
To address these challenges Zhao and colleagues at Takeda are developing what he calls a nano LC-MS/MS shotgun proteomics approach to improve HCP coverage. “Peptides derived from endopeptidase trypsin digestion were chromatographically separated on a Thermo Fisher Scientific Easy-Spray nanoLC and detected using data-dependent acquisition (DDA) mode on a Thermo Orbitrap Fusion Lumos mass spectrometer,” he details. “DDA raw data were searched for HCP identities using Byos software from Protein Metrics, Proteome Discoverer from Thermo, or MaxQuant open source software. We are also testing different extraction, enrichment, fractionation, and separation methods to solve the complexity of cell therapy HCPs.”