Presented at The Bioprocessing Summit 2018, Boston, MA, USA
Mass spectrometry (MS) is generally regarded as the most useful technique available for supplementing conventional immunoassay-based methods for HCP analysis. We used the instrument-independent software platform Genedata Expressionist® (Genedata Inc., Lexington, MA, USA) to develop highly flexible approaches to the identification, quantification, and routine monitoring of HCPs using MS.
HCP contaminants are often present in trace amounts and therefore efficient noise-reduction methods are required to optimize peak detection for low-abundance species. A number of data noise-reduction algorithms were applied individually or in combination to efficiently remove artefactual signals without losing genuine signals from low-abundance species. Replicate data sets containing spiked proteins could be aligned to provide “prior knowledge” of the location of HCP species. This knowledge was then used to efficiently detect trace amounts of HCP in test samples.
We developed a two-stage approach for HCP analysis. In the initial step, all peptide species from the target biopharmaceutical were identified by conventional peptide-spectrum–match searches. Subsequently, a second round of searches was performed using the data set from samples containing the biopharmaceutical and any potential HCP contaminants. In this second search, peaks known to arise from the target protein were excluded, meaning that any identified signals could be assumed to originate from HCPs. This approach significantly reduced the probability that low-abundance signals from the biopharmaceutical were identified as HCP species (false-positives).
In another approach, m/z and expected RT values of known HCP contaminants were recorded in a knowledge base and used to identify and quantify HCPs in a fully automated quality control procedure. Such libraries can be used to annotate and identify signals in downstream samples, even in cases where no fragmentation data is available.