An Inflection Point for Biopharmaceutical Host Cell Protein Analysis

Has LC-MS Become a Generally Accepted Tool for Complementing Immunoassay-based HCP Analysis?

My colleague Catalin Doneanu and I just returned from Lisbon, Portugal, where we attended the BEBPA Conference on Host Cell Protein (HCP) analysis. Along with the other 120 attendees, we participated in open and thoughtful discussions on the best practices for HCP analysis using conventional immunoassay/ELISA technologies. Additionally, we discussed the use of complementary technologies – primarily LC-MS methodologies – to validate and extend the results of these more traditional approaches to HCP analysis and control.

A few things stood out in my mind as key indicators of where the biopharmaceutical industry is heading in this challenging area of analysis.

Waning confidence in both 2D gels and Western blotting as accurate HCP assay validation tools

At the meeting, there were discussions and presentations expressing concerns that commonplace methods of validating immunoreagents used in host cell protein analysis are insufficient for the task.

In particular, several speakers expressed their mistrust of 2D gels and Western blotting as tools for demonstrating HCP coverage in complex samples, and for establishing that such reagents have the broad specificity to profile HCPs quantitatively. Challenges such as poor specificity, limited dynamic range, and recognition of linear versus 3D epitopes were among the reasons cited. But the most fundamental issue still seems to be that many proteins are not recognized by animal immune systems, which generate the polyclonal sera used as reagents in these assays.

The industry seemed to feel comfortable that – with significant time, effort, and cost – they can create reasonable immunoreagents for the few HCPs remaining at significant levels in the final product. However, they’re challenged to create assays that enable process science groups to gain quick feedback during process development, and to support increasing upstream and downstream process QbD studies.

In multiple presentations, there were suggestions that using proteomics-type workflows to study pre- and post- affinity purified fractions could provide better insights into what proteins are recognized by their reagents. In addition, this approach would provide the ability to use data-independent analysis workflows such as LC-MS^E and Top3 quantification. These workflows generate a more quantitative overview of the HCP composition of cell culture supernatant and downstream purification steps of their products.

How low do we need to go?

Low-level HCPs can impact product quality. Discussions of phospholipase and protease enzymes that have been noted to copurify with biotherapeutic proteins reinforced the position that even lower level HCPs can have significant product quality implications.

One talk from Eli Lilly indicated that a phospholipase spiked at single digit ppm levels could degrade the low-level polysorbates used in biotherapeutic formulations in a rapid fashion – thereby raising the risk of product instability and generating efficacy/safety risks. This once again highlighted the need to go beyond a quantitative ELISA result and to know your HCP identities and abundances in order to minimize risks.

On the horizon: HCP identification for vaccines

HCPs are also a significant concern in vaccine products – particularly, the quickest growing class of subunit vaccines that are produced and purified like the “well characterized” recombinant protein therapeutics. Immunogenicity (desirable in a vaccine!) is not the only concern about HCPs. During application reviews, regulators have apparently asked for HCP identification data for vaccine products.

Is LC-MS the right answer to today’s host cell protein challenges?

Where does LC-MS fit today, and what are the expectations going forward? The big shift I saw at this meeting is that only a handful of the attendees challenged the need for LC-MS technologies to complement the traditional immunoassays/ELISA assays and gel-based verification approaches. On the other end of the spectrum, there were only a few attendees arguing that MS can currently displace immune-based assays as release tests used for product specification, or as the only technology required to demonstrate HCP clearance validation within a regulatory filing.

Overwhelmingly, the consensus seems to favor LC-MS based approaches being adopted to complement generic or platform HCP assays in product development. This is largely on account of their ability to ID and quantify HCPs that can be targeted for removal by process improvements, and because of their reliability as a tool to validate the immune reagents used in process-specific HCP clearance assays of a drug substance.

Poster: HCP assay reproducibility and HCP comparability of a biosimilar

At the meeting, Catalin and I presented a poster on multi-laboratory HCP analysis of the NIST Reference mAb and an innovator/biosimilar HCP comparability study. The presentation demonstrated that multiple laboratories can obtain a reproducible list of HCP impurities and their abundances from purified mAb samples.

In addition, the work showed that it is possible for two companies with unique mAb production processes to produce a common biotherapeutic mAb containing a common set of HCPs, although present at varying levels.

We also launched a new resource – called the HCP Toolbox – which has been requested by researchers at past meetings who have seen a similar tool from Waters for HDX-MS analysis. The toolbox provides a presentation with speaker notes, and access to key literature that will help researchers discuss the value of LC-MS for biopharmaceutical HCP analysis. The toolbox can be accessed at waters.com/HCPToolBox.

A great week in Lisbon

Overall, it was a pleasure to visit Lisbon during a sunny week in May, and to attend this focused conference, where honest and open discussions on the challenges and strategies for control and monitoring of Host Cell Proteins permeated the formal program sessions and during the breaks and social events.  Next year’s meeting in San Francisco should continue these great discussions, and I look forward to attending again with my Waters colleagues.