Host-cell proteins (HCPs) remain one of bioprocessing’s most stubborn analytical headaches. These residual impurities, shed by production cells during manufacturing, can survive purification and threaten both biologic quality and patient safety. Now, Julie Flecheux, a doctoral student in proteomics at the Université Claude Bernard Lyon’s Institut des Sciences Analytiques, and colleagues report a targeted proteomics strategy designed to make HCP monitoring more robust across the shifting matrices encountered in modern biomanufacturing.
The problem is not trivial. Recombinant therapeutic proteins, especially monoclonal antibodies, are commonly made in Chinese hamster ovary (CHO) cells. Even after downstream purification, trace HCPs can remain. Some might trigger immune responses, while others can damage the product itself. Regulators, including the U.S. Food and Drug Administration and the European Medicines Agency, expect manufacturers to control total HCP levels in the final drug product, typically below 100 parts per million.
Industry has long relied on enzyme-linked immunosorbent assays (ELISAs) for total HCP measurement. But ELISA cannot identify which proteins are present. Liquid chromatography coupled to tandem mass spectrometry offers that protein-specific view, and multiple reaction monitoring (MRM) has become an attractive targeted option. But there’s a catch: conventional retention time-scheduled MRM depends heavily on peptides eluting when expected.
Flecheux and her colleagues show just how fragile that assumption can be in real bioprocess samples. Using an assay targeting 97 CHO-derived HCPs through 240 peptides, they found retention time shifts of up to several minutes across different drug substances and process intermediates. The result was truncated chromatographic peaks, missed peptides, and poor transferability from one matrix to another, even with broad monitoring windows.
Their solution is scout-triggered MRM (st-MRM). Instead of relying on fixed retention-time windows, the method uses predefined scout peptides to dynamically trigger groups of target transitions during the run. In practice, that means that all heavy standard peptides were detected across multiple matrices in a single injection, without re-optimizing the method for each sample type.
The workflow also delivered absolute quantification across six orders of magnitude, with HCPs measured down to 2.9 parts per million in purified drug substance. For bioprocess developers, that combination of sensitivity, transferability, and one-run efficiency could be significant.
The authors do not position st-MRM as a replacement for untargeted discovery workflows or ELISAs. Instead, they frame it as a pragmatic complement: a targeted, protein-specific tool for risk-based monitoring of known process-relevant impurities. In an industry where every batch, matrix, and process step can shift the analytical landscape, that flexibility may be exactly what HCP surveillance has been missing.
