Host Cell Proteins: Needs Evolving Faster than Capabilities

Dealing with host cell proteins (HCPs) remains challenging for cell culture-based therapeutic protein manufacturing. Although analytical methodology continues to improve, the complexity of today’s “typical” biologic drugs and processes appears to be evolving even faster.

The main concerns with incomplete understanding of HCPs are patient safety (arising from immunogenic responses to foreign proteins) and the related potential for HCPs to accelerate degradation of the drug substance.

HCPs exist as both a consequence of normal cellular activity and in response to genetic engineering, process-related stresses, and the presence of high titers of recombinant proteins. HCPs with enzymatic or catalytic activity are capable of degrading the drug substance, thereby reducing its effective concentration (and effectiveness) and introducing protein fragments which, like the HCPs themselves, are not recognized by the human immune system.

“All such proteins are potentially immunogenic because they are foreign to humans,” says Sayuri Otaki, Segment Lead for Protein Separations at MilliporeSigma. “Leveraging multiple, complementary methods is critical to minimizing the risk of HCP contamination.”

No shortage of HCP-detecting methods

Pharmacopoeias in the United States, Europe, and Japan have issued guidelines for developing and validating HCP assays for biomanufacturing, and these are typically specified in terms of analytical methods.

“The workhorse HCP monitoring and quantitation method involves immunoassays such as sandwich ELISA using anti-HCP polyclonal antibodies. For characterization, the standard is 2-D gel electrophoresis followed by western blot analysis, or immunoaffinity purification followed by 2-D gel electrophoresis,” Otaki says.

Orthogonal analytical methods are also used to support assay development and validation—for example gel electrophoresis, western blot, and mass spectrometry. While these methods meet current needs they can stand improvement, according to Otaki.

For example ELISA is an accessible, high-throughput, highly sensitive method that only works for previously known HCPs.

“That is why evaluating the coverage of ELISA anti-HCP antibodies is critical,” Otaki says. “To evaluate these anti-HCP antibodies, 2-D gel electrophoresis is used with either western blot or immunoaffinity chromatography.” LC-MS, another standard HCP analysis modality, is highly sensitive and capable of identifying individual HCPs. “But it requires significant sample preparation, skilled staff, and sophisticated instrumentation, and the assays require extensive validation.”

According to Otaki a “combination of multiple techniques” on an automated or semi-automated platform has the best chance of arriving at “an optimal outcome.” For example, the automated Millipore^® Auto2D^® 2-D Electrophoresis Device provides the usual benefits of automation, including up to a 90% time saving and greater consistency, according to the company.

Current methods fall short

So, despite a plethora of analytics, no one analysis tool or combination satisfies all the requirements for HCP characterization.

According to Hongjin Huang, Ph.D., Senior Director at Applied Biomics, ELISA detects many HCPs quickly and at low cost. “However, they cannot detect the entire range of HCPs, plus modifications and degradation products. Plus, the consistency and reproducibility of ELISA is lower compared with other methods.”

Even commercial HCP detection standards, for example, cover at best just 95% of known HCPs for a given expression system. The nature of cell-based biomanufacturing is such that even if concentrations of the remaining 5% are deemed too low to include in a kit they may nevertheless show up—not to mention unknown or unanticipated HCPs.

1D western blot detects HCPs with more sensitivity than ELISA, with certain staining methods (e.g., silver staining) offering high sensitivity. “However, 1D western blot only separates proteins by molecular weight, which compromises the detections so consistency and reproducibility are low,” Huang says.

By contrast 2D western blot allows comprehensive HCP profiling by adding isoelectric point to molecular weight as the discriminating characteristic. “But this method requires analysts to run two gels: one for the HCP and one for the western blot,” Huang explains. “Due to gel-to-gel and gel-to-membrane variability, consistency and reproducibility are low. In addition, protein spots are counted from the gel rather than the membrane, and the spots in the two gel images often do not line up exactly.”

Nor does Huang spare criticism of mass spectrometry-based analysis methods. “LC-MS/MS is a comprehensive method which, depending on the instrument, identifies all proteins in the sample with high sensitivity.” LC-MS/MS can detect post-translational modifications but fails to pick up degradation products.

“Because it identifies everything, it also picks up large amounts of keratin and albumin, which can interfere with the data analysis,” Huang explains. “In addition, LC-MS/MS cannot visualize the protein profile after separation by LC, and it only separates peptides by one dimension—hydrophobicity.”

To tackle today’s tough HCP issues Huang suggests 2D DIGE western blot using fluorescent dyes with greater sensitivity than silver staining. In this single-gel method, which Applied Biomics offers as a service, HCPs are first analyzed on a 2D gel followed by scanning.

The same gel is transferred to a membrane and incubated with the antibody, then the membrane is scanned. In the overlay image, HCP spots from gel align perfectly with antibody spots from the membrane.

“This approach is highly accurate and reproducible, eliminating the gel-to-gel and gel-to-membrane variation of a standard 2D western blot. “It allows visualization of HCPs and antibody profiles on the same gel, quantifies HCP antibody coverage, and allows subsequent experimentation for HCP identification.”

For LC-MS/MS analysis, Applied Biomics uses the Ultimate 3000 and Orbitrap Exploris, both top-line instruments with high sensitivity. Using the latest software (such as Proteome Discoverer 2.5) the company claims the ability to detect “thousands of HCPs at very low abundance.”