Impact of Purification Process Impurities on Preclinical Assessment of Immunogenicity of Biotherapeutics

Abstract

An immune response to biotherapeutics has the potential for a negative impact on patient safety and efficacy. Predicting biotherapeutic immunogenicity at the preclinical stage of drug discovery is important for the selection of the best biotherapeutic candidate. The early discovery protein production process for antibodies typically consists of protein expression in an human embryonic kidney (HEK) 293 cell line, followed by purification using protein A chromatography. Typically protein purification methods at the early discovery stages are not as stringent as during manufacturing stages of the process, so it is likely that biotherapeutic preparations might contain impurities such as host cell protein (HCP), residual DNA (rDNA), and protein A leaching that could impact immunogenicity predictions. Preclinical immunogenicity predictions can be based on an in vitro peripheral blood mononuclear cell (PBMC) assay, which measures the proliferation of CD4 T-cells in peripheral human PBMC cells when in contact with protein biologics. We used bevacizumab, an anti-VEGF monoclonal antibody which has been shown to have a low rate of immunogenicity clinically, as a control in our assay. While the commercial drug product gives a stable and low signal, we observed a batch to batch variation response in the PBMC assay with various preparations of the antibody. We hypothesized that HCP, rDNA, and protein A ligand contaminants introduced in the early discovery protein production process caused the elevated CD4 T-cell proliferation observed in the various protein batches of bevacizumab, and that a secondary purification step could remove a majority of those contaminants leading to a more reliable CD4 T-cell proliferation response. To test this hypothesis the levels of those impurities were measured in the bevacizumab batches which had been previously produced and showed variability. No clear correlation was found between the levels of HEK HCP and in vitro T-cell proliferation response. There was some correlation where the sample with the lowest levels of rDNA led to the lowest percent donor response, and the sample with the highest levels of protein A led to the highest percent donor response, however these results require further testing to fully understand. We reproduced the bevacizumab mAb using either a one-step purification method consisting of protein A chromatography, or a two-step purification method consisting of protein A chromatography followed by Q-membrane chromatography run in flow through mode (Q Flow Through). These batches were tested for impurities (HEK HCP, rDNA, and protein A ligand) and were then analyzed in the PBMC assay. The results showed that using a secondary purification step (Q Flow Through) reduced the levels of impurities such as HEK HCP, rDNA and protein A ligand of bevacizumab. The process impurities were not a clear source of variability in percent donors responding with positive CD4 T-cell proliferation in the PBMC assay, but they appeared to impact the strength of the CD4 T-cell proliferation response of those donors.