Problematic HCP
The leading cause of particle formation following enzymatic degradation pathway is from a group of host cell proteins, which exhibit polysorbate degradation activity. Lipoprotein lipase (LPL) hydrolyze ester bonds within triglycerides to form alcohols and fatty acid molecules. Knockout mutants of LPL reduce polysorbate degradation by 50% (Chiu et al., 2017). In the development of mAb1, we observed elevated sub visible and visible particles after long term storage at 2-8oC and also showed high PS20 degradation over 4 weeks storage at ambient temperature. LPL was detected in the bulk drug substance in parts per billion (ppb) levels (data not shown). In our proteomic studies, LPL, was found in all three culture processes (FB, SS, NSS). After the ProA purification, LPL was only detected in the fed-batch process, suggesting fed-batch culture process may be harboring higher LPL levels (Table 2).
The DDA method was not adequately sensitive to detect the lipases in purified samples, such as ProA pool. Instead, the PRM method was used to further detect low ppm levels of lipases in mAb1 process sample. The data-acquisition method for PRM contains a list of peptides, which are exclusively present only in proteins that are of interest to the researcher, and their elution time window. After the data is acquired and analyzed, the information on the abundance of peptides observed by PRM is then used to estimate the abundance level of their corresponding proteins. PS degradation was also reported during the long-term storage of CHO-derived purified mAbs with detectable quantities of endogenous lysosomal phospholipase A2 isomer X1 (LPLA2) (Hall et al., 2016) and PLBL2 (Dixit et al., 2016). We used PRM method for targeted quantitative analysis of peptide pairs to identify LPL, LPLA2 and PLBL2 (Table 2). Compared from the FB process, both perfusion processes shows less overall lipases and later removed or greatly reduced the quantity after the ProA purification.
Peroxiredoxin 6 (Prdx6) is the uniquely 1-Cys member of the peroxiredoxin family with calcium-independent phospholipase A2 activities (Shanshan et al., 2017). Phospholipases such as LPLA2, Prdx6 and phospholipase A-2 activating protein (PLAA) were detected in fed-batch and NSS perfusion processes but not in SS perfusion process. LPLA2, Prdx6 and PLAA were also detected in fed-batch intermediate purification pool after ProA purification but not in the both perfusion processes.
For long-term storage of mAbs formulated with PS, difficult to remove lipases that were identified need to be further cleared during the downstream purification. In order to evaluate the mAb1 product stability, PS was analyzed using two different methods after 4 weeks of storage at 25oC. Fluorescence micelle assay (FMA) was utilized for PS quantification and 4-methylumbelliferyl decanoate (4-Mud) assay for lipase activity. Table 2 summarized the PS quantification, lipase activity and its relative abundance in different processes samples. In HCCF, PS degraded from 400 ppm at day 0 to approximately 20 ppm after 4 weeks of incubation. This result was expected due to the high HCP concentration in the cell culture fluid. After ProA purification, PS content was measured to above 200 ppm for both perfusion processes but less than 100 ppm for the fed-batch process. Lipase activity also shows great differences between perfusion and fed-batch processes after the ProA purification. Combining all the data collected in this study lead to the fact that fed-batch samples contained more lipases involved in polysorbate degradation and particle formation compared to both perfusion processes. These lipases were not removed by chromatography methods.