State of the Art in Host Cell Protein Assays
Presented by Martin Vanderlaan, Genentech, a Member of the Roche Group
Enzyme linked immunosorbent assay (ELISA) remains the “work horse” method in host cell protein (HCP) detection in biopharmaceutical industry. This format provides the sensitivity, precision, speed, and high throughput needed to support process development and GMP lot release. Typically some form of multi-analyte polyclonal antisera is used that recognize a broad population of HCPs. Demonstration of coverage (fraction of the HCP population recognized by the antibodies) remains an important means of determining if the antibodies cover a “diverse enough” population of the potential HCP impurities. Examples will be given to show that platform assays can be better suited than process specific assay in detecting changes in HCP levels that arise from process changes. While the process specific assay gave higher HCP readings in early pools in the process (those steps closest to the cell culture), the final products often had higher readings in the platform assay. The implication is that what shows up in the final product is often most influenced by what sticks to the product and is carried along through the recovery steps.
One important observation about HCP immunoassays is that they occasionally give dilution-dependent results. Dilution dependence arises because the HCPs in the sample (perhaps one or a few HCPs) do not dilute in parallel with those in the standard (a mixture of 1000s of proteins). Limiting amounts of reagents can come in the form of coat, conjugate antibodies, or even Streptavidin-HRP that are associated with “antigen excess” and non-linear dilution behavior. It is advised to check all reagents to minimize impacts of reagent limitation sample dilution linearity. From a product purity perspective, however, sample dilution non-linearity is not an “assay artifact” but rather could be an important indication that a large amount of a particular HCP has co-purified with product. Phospholipase B-like 2 (PLBL2) was found to co-purify with multiple CHO-derived Mab products. It may be that this is a “common” impurity in Mabs produced across the biotech industry since it was found to have co-purified with products independently produced from different CHO-K1 subclones. Not all platform or commercially available CHOP ELISAs detect this protein. Roche and Genentech have now screened all products in clinical development for PLBL2 and instituted process changes when needed to reduce PLBL2 in final product.
Proteomic approaches using liquid chromatography and mass spectrometry to identify and quantify individual HCP impurities have been used in orthogonal fashion to supplement the immunoassays. These analyses have the advantages that they detect HCPs regardless if they are immunogenic and do not require the time needed to develop antisera. It is excellent at assuring that a non-immunogenic protein present at significant level (>100 ng/mg) is detected, even though it otherwise is likely missed by the immunoassay. With instrument capability rapidly improving, mass spectrometry is likely to find greater application in HCP analysis in the future.