In silico and in vitro immunogenicity risk assessment of process- and product-related protein impurities reported in adenoviral-vectored COVID-19 vaccines.

Process-related protein impurities (PRPIs) such as host cell proteins (HCPs) in biologic products have been associated with unexpected immune responses, adverse events, and the discontinuation of clinical programs. While an HCP threshold of 100 ng/mg (PPM) is considered to be 'acceptable', that threshold may be too high if individual impurities have strong intrinsic immunogenicity or, in contrast, may be overly cautious if many of the HCP impurities have low immunogenicity. In silico methods for predicting the immunogenicity of PRI and HCP may help triage higher risk from lower risk entities. We explored this hypothesis by selecting candidate impurities from a published list of over 1,800 HEK-derived HCPs and adenoviral Product Related Impurities (PRIs) that were detected in high abundance in Chimp Adenovirus (AdV)-vectored COVID vaccine lots, for in silico and in vitro evaluation. The results of human donor peripheral blood mononuclear cells (PBMC)-based assays were well aligned with in silico estimates. These orthogonal approaches, when combined, have the potential to identify process impurities that should be removed to remove prior to submission of a drug product for approval regulatory authorities. In addition, the in silico and in vitro approach described here provides a framework for improving the overall safety of biologic products.