Host Cell Proteins: Ignore at Your Own Risk

Host cell proteins (HCPs) are an inevitable byproduct of cell culture- or fermentation-based biomanufacturing. The biggest fear of not removing HCPs from a biopharmaceutical preparation is the potential for HCPs to elicit severe immune responses in patients, particularly with high-dose drugs like monoclonal antibodies. Traditionally, scientists and manufacturers analyze HCPs using immunoassays, typically ELISA. But as noted in a recent introduction to liquid chromatographic purification of HCPs followed by mass spectrometry, ELISA is a highly imperfect assay in this circumstance.

The process by which polyclonal antibodies are raised to HCPs involves exposing animals to “null” host cells, that is, cells lacking the gene for the expressed protein. In this process the antibodies raised may not be representative of HCPs present, either in the null host cells or in the transfected cells. Such preparations will therefore miss the actual HCPs produced during the cell culture.

Bioprocessors are therefore looking for alternatives to immunoassay.

Inevitable consequence

Impurities, side products, and extraneous chemical species are an inevitable consequence of pharmaceutical manufacturing. In biopharmaceuticals, impurities are either product-related or not. Product-related impurities include excess heavy or light chain fragments, misfolded proteins or aggregates, and molecules that have undergone undesirable post-translational modifications.

Non-product related impurities including carryovers from media and feed, plasticizers that leach from hoses and bioreactors, and HCPs, comprise the most significant class of non-product related impurities. Small molecule drug manufacturing usually stipulates a permissible level of impurities, for example “total impurities less than two percent, with no impurity greater than half a percent.”

The situation differs somewhat for biopharmaceuticals, whose allowable impurity levels depend on many factors. “There’s no hard number on how much of an impurity is allowed because some impurities are totally benign, while some are immunogenic, even in minute quantities,” explains John Gebler, Ph.D., biopharma director at Waters.

The presence of HCPs is particularly significant for monoclonal antibodies, which are dosed multiple times, often at hundreds of milligrams per dose. “At those dosage levels, even part-per-million host cell protein contaminants exist in significant concentrations,” Gebler adds.

Luckily, modern analytical methods can detect potentially dangerous HCPs at ppm levels at a stage in development where process scientists can tweak culture and purification parameters to reduce or eliminate these contaminants.

Proteases comprise another class of HCPs to watch out for. Proteases that carry over through downstream purification steps, even at very low concentrations, can affect a drug’s short- and long-term stability.

During a biopharmaceutical’s early development stages, HCPs are typically quantified and characterized by liquid chromatography-mass spectrometry. Thereafter, manufacturers use immunoassays for rapid, reliable assays of HCP levels. “But you don’t always know which HCPs the immunoassay will target,” Gebler explains. “You can have an immunoassay or ELISA that measures HCPs that are not in the sample, or new HCPs can arise for which an ELISA test does not exist. Increasingly, biopharmaceutical developers want to know the identities of HCPs, which is something immunoassays are incapable of providing.”

Regulators are not pushing manufacturers to use LC/MS during production. However, the method is state-of-the-art for identifying and quantifying HCPs and determining their typical concentration range during manufacturing. From there, if the HCP is deemed to be immunogenic, process engineers can devise ways to eliminate it, or to raise antibodies specific to it for later use, during production, in an immunoassay.

Detection with or without filtering

Characterizing HCPs with LC/MS can be accomplished in several ways. The two most common methods involve data-dependent and data-independent acquisition.

In data-dependent mode, the mass spectrometer selects a fixed number of precursor ions based on an initial survey scan. Masses are typically selected by abundance. This is followed by MS/MS analysis of those ions, and reconstruction of HCP sequences by software. “Data-dependent acquisition is a way of obtaining extensive data on particular ions of interest,” Gebler explains.

In data-independent mode, MS acquisition occurs without the filtering or selection step. The spectrometer does a low-energy scan to identify parent ions, followed by high-energy scans to obtain fragments.

Both methods provide good data on HCPs, but data-independent acquisition represents a filter-free, native approach that is most useful during a biopharmaceutical’s discovery phase. “Data-independent acquisition covers a greater dynamic range, and is less likely to miss minor components,” Gebler says. “The data-dependent method can miss species of very low concentration but whose presence can be significant. Data-independent acquisition catches everything.”

Waters has developed an acquisition method, SONAR, a data-independent method that incorporates some aspects of data-dependent acquisition as well. “Under certain situations SONAR provides more data, and more information. It’s especially useful for finding a needle in a haystack, for example when the product peak is overwhelmingly large relative to a HCP impurity,” Gebler adds.

Inevitability

The ubiquity of HCPs presents challenges to developers of biosimilar drugs. Biotherapies are approved as complex products, with substantial heterogeneity for both drug substance and impurities. Sponsors of such drugs are expected to produce product over and over again within established, approved impurity constraints. A similar process, conducted at a different facility, may exhibit a HCP profile that is different from that of the innovator company’s product. This difference may be beneficial or detrimental to the product’s quality or effectiveness. The fact that it is different, however, could prompt regulators to ask for more testing or complete characterization of these HCPs.

Characterizing HCPs involves a series of more-or-less conventional proteomics experiments. Where academic researchers may wish to identify as many HCPs as possible, biopharmaceutical developers are more interested in identifying a specific subset of HCPs through methods that are robust and reproducible.

“If you’re trying to push analytics to their utmost limits, for example with nanoflow LC and exotic MS methods, there’s a good chance that your results will be riddled with errors,” Gebler says. “If you’re developing a potential blockbuster drug there is no reason to operate on the fringes of your lab’s analytical capabilities. You need an analytical method that’s reproducible, robust, and validated.”

For example, for characterizing HCPs, Waters recommends capillary or analytical flow chromatography over nanoflow LC because capillary and analytical flow LC systems are more robust. “MS has become so sensitive, you no longer need to go down to nanoflow to get answers,” Gebler adds. “You’re not sample-limited, so there is no reason to use ultra-low flow rates.”

Although some authors discuss the benefits of using LC-MS for routine HCP determinations, the likelihood that such complex, lengthy assays will ever catch on during production is slim to none. Bioprocessors have always resisted using analytical methods as complex as LC-MS at-line or in-line. Immunoassays will therefore be around for a while still. Where LC-MS can contribute is in the initial detection and characterization of HCPs.