Pulling Out Problem Proteins


On March 17, 2017, a search of ClinicalTrials.gov for “monoclonal antibody” returned 3,162 studies. To make the monoclonal antibodies (mAbs) for these clinical trials and most other applications, scientists typically use a host-cell system. To test and manufacture a safe, effective and consistent biotherapeutic, the mAbs must be as pure as possible. In particular, host-cell proteins (HCPs) must be removed. To do that, mAbs preparations must be monitored for HCPs, and then researchers and manufacturers need methods to remove the impurities.

Part of the problem comes from the mAbs and HCPs sticking together during purification steps, such as protein A affinity chromatography. Asish Chakraborty, senior biopharmaceutical business development manager at Waters, explains: “Although Protein A chromatography provides a high degree of clearance for HCPs, co-purification of a subset of HCPs with the mAb is common due to interactions of HCPs, primarily with the mAb itself.”

If a mAb therapeutic includes HCPs, they can trigger an immune response in a patient.

This, says Chakraborty, can cause “adverse clinical effects, or have biological activities that reduce drug stability or effectiveness.”

To effectively remove HCPs from a product, scientists must know what they are, but that’s not so easy. “‘Host-cell proteins’ is a very diverse term with several thousand possible gene expressions of proteins,” says Tomas Björkman, senior scientist, GE Healthcare Life Sciences. “These proteins vary greatly in size, hydrophobicity and in net charge.” The diversity gets even more complicated, because HCPs “have different post-translational modifications, glycosylation, splicing, etc.,” Björkman notes. “They may also undergo several other modifications, including deamidation, oxidation and clips.”

Fortunately, the HCPs can be identified and quantified with liquid chromatography-mass spectrometry (LC-MS). “The development of LC-MS techniques enables scientists to identify and quantify individual HCPs, facilitating early identification of potential high-risk HCPs and allowing scientists to design rational process optimization to begin to eliminate or minimize the levels of these HCPs,” Chakraborty explains. Also, using LC-MS—rather than immune-based methods, such as an ELISA—provides more universal HCP detection, “including those HCPs that do not elicit immune responses in animals used to create those immunoreagents used for traditional assays,” Chakraborty points out.

Other chemical characteristics also come into play. For example, a mAb can have a large charge distribution, which can overlap with some of the HCPs. That can make it hard to separate the mAbs from HCPs with some techniques, such as using an anion exchanger.

Monitoring the mixture

To monitor HCPs in these products, scientists face several challenges, including the diversity of HCPs and the high concentration of mAbs. “The high mAb concentration makes it convenient to use a sandwich immunoassay—for example, an ELISA—to enrich the HCP and at the same time wash away monoclonal antibody,” Björkman says. The washed-away mAb becomes the product.

That doesn’t always work. “It is, for example, difficult—if not impossible—to get full coverage for the HCPs, and there might not be enough antibodies to quantitatively measure a certain protein,” Björkman explains. Beyond that, the HCPs vary based on how the mAb product is made. “Varieties in cell lines may require cell-line or even product-specific assays,” Björkman explains.

The chromatography platform must also be monitored. Katherine Lintern, a post-doctoral research associate at University College London, and her colleagues explored a common problem: the reduced capacity of protein A affinity chromatography in longer production runs. Using LC-MS/MS, these scientists found that HCPs on the chromatography resin can increase dramatically, such as increasing 10-fold as the process goes from 50 to 100 cycles [1]. Nonetheless, that increase of HCPs in the resin did not reduce the efficiency of the process.

As another challenge, the level of HCPs varies extensively, even though the levels are low. As Chakraborty points out, the HCP levels are “typically parts-per-millions, ppm, concentrations relative to the biotherapeutic, which requires a 105 [to] 106 dynamic range for detecting individual HCPs present in the biotherapeutic product.” That’s a lot of range!

To take on that problem, scientists at Waters and their collaborators developed a two-dimensional LC-MS platform that provides single-digit ppm detection limits to identify and quantify HCPs in mAb preparations. “Working with our customers, we realized that they also needed a higher-throughput method for monitoring/tracking HCPs during product development,” Chakraborty adds. “To fulfill this need, Waters developed a fast and easily deployable, but less sensitive, one-dimensional LC-MS method for monitoring HCPs during the process development. In this 1D approach, HCPs are identified at higher levels from earlier purification steps, and tracked using signature peptide ions below the levels at which they could be directly identified using the 1D method.”

Better backbones and more

To improve some of these processes, scientists can take various approaches. As an example, Björkman points out “the development of hydrophilic or more inert chromatography-resin backbones with minimal nonspecific binding of proteins.” That reduces the problem of HCPs going along with mAbs during purification. That’s just one improvement in resins. Others include higher resolution from smaller particles, and additives that reduce binding of HCPs. Instead of taking out the HCPs, some advances aim to reduce their inclusion in mAb preparations from the start. As an example of that, Björkman mentions “cell lines developed to lower expression of certain problematic HCPs.”

The wide range of HCPs demands bioinformatics tools, as well. For example, “Waters’ LC-MS methodology for HCP analysis makes use of MSE, a data-independent acquisition approach, to detect, identify and—equally importantly—quantify HCP peptides over a larger dynamic range, enabling biotherapeutic organizations to detect HCPs down to single-digit ppm using the 2D LC method,” Chakraborty says.

Today’s tools provide just what some scientists need. “As far as I’m concerned,” says Phillip Wright, pro-vice-chancellor for the Faculty of Science, Agriculture and Engineering at UK-based Newcastle University, “modern mass-spectrometry tools using LC-MS/MS give information-rich analysis.” He adds, “Not only do they tell you what is there, but also you can get relative and/or absolute quantitation of the host-cell proteins from the cell-culture vessel and through the downstream process train.” Wright also appreciates the ability to learn even more about a mAb product. As he says, “If the product has modifications, such as glycosylation, you can identify this and the sites and quantify the product heterogeneity, as well.”

Wright’s overview shows just how powerful today’s analytical tools are in developing and producing mAb-based treatments.


[1] Lintern, K, et al., “Residual on column host cell protein analysis during lifetime studies of protein A chromatography.
J Chromatography A, 1461:70-77, 2016. [PMID: 27473513]