Comparison of HCP distribution and Cluster analysis of quantified HCP
The relative fold changes of quantified HCPs were calculated by comparing the protein abundance between the perfusion processes to fed-batch and visualized through a heat map shown in Figure 4. Each band represents one HCP species as identified by mass-spec based proteomics. Blue corresponds to fold decrease and red to fold increase in abundance comparing FB to perfusion processes. HCPs were further classified in three clusters according to their concentration profiles during the cultures. Cluster 1 and cluster 3 represents HCPs that have a large fold increase and decrease (>2.5 fold) respectively from the fed-batch cultures. Cluster 2 contains HCPs that show inconsistent fold changes throughout the culture days. Although there were many HCPs that did not show large fold changes from FB to perfusion processes, distinct red and blue represent the HCPs that show a significant increase and decrease in protein abundance regarding the specific culture processes. SS perfusion culture shows large portions of HCPs with reduced abundances. Depending on the manufacturing processes, these prevalent differences in HCP population suggest to develop “process-specific” HCP assays. Commercially available ELISA kits, typically used the reagents derived from FB, can be limited in their utility by poor sensitivity and coverage.
A cutoff of 2.5 fold was used for the absolute fold changes based on the %CV of the dataset. HCPs that have more than 2.5 fold changes were accounted for functional annotation and biological interpretation of proteomics data. To understand the biological roles of the identified HCPs, GO analysis was performed based on relative quantification from fed-batch culture using DAVID (Supplementary figure 1). Most upregulated HCPs in cluster 1 (Figure 4) were associated with transcription (spliceosome) and translation (RNA transport, ribosome, lysosome). Increased level of transcription and translation correlate to increase in the final titer produced in CHO cells. For example, transcription factor BTF3, cellular nucleic acid binding protein (CNBP) and 60s ribosomal protein L36a (RPL36A) were highly upregulated (more than 10 fold) in both perfusion processes. In perfusion culture (cluster 3), HCPs that associated with metabolic pathway were downregulated with a response to changes in concentrations of key metabolites in the cell. NSS perfusion shows downregulation of phagosome and proteasome, which are responses found in starved and growth arrested cells due to the high production of mAb (Vergara et al., 2018). A recent study reported that the primary cellular function of HCPs is an important factor to be considered when preparing feedstocks for ProA purification (Lintern et al., 2016). Cells with a high productivity such as perfusion process expected to produce more mAb per unit of O2 and nutrient consumed (Pan et al., 2017).
Cellular component analysis shows that compared to the fed-batch, proteins located in both the cytoplasm and nucleus are greatly impacted in both perfusion cultures (Supplementary figure 1). For SS perfusion, the percentage of HCPs associated with the extracellular region, extracellular space was significantly higher due to the high viability. However, in the NSS perfusion, HCPs that are associated with cytosol and nucleoplasm emerged as the most impacted proteins. This result suggests that HCPs in the culture supernatants in SS perfusion culture process include a high percentage of extracellular proteins secreted from viable cells, whereas HCPs in NSS perfusion culture process involve a significant portion of intracellular proteins resulting from cell lysis. Those intracellular HCPs included proteins that were not naturally secreted and could only be found if cells lysed. HCPs that are associated with membrane proteins are upregulated in the early days but downregulated in later days of the culture. In earlier days, a previous study suggested that membrane proteins are secreted by exocytosis after being sorted in the trans-Golgi network into transport vehicles(Le Borgne and Hoflack, 1998). In our case, we believe in both perfusion processes, the active secretion is decreased due to cell lysis as the number of culture days increase. Furthermore, the concentration of lactate dehydrogenase (LDH), which is most widely used to quantify the degree of cell lysis, increased toward the end of the cultures. For the NSS, LDH level is significantly increased that demonstrates shear sensitivity on cell. At the end of the culture, LDH level was 1.1x105 pg/cell/day for NSS and 3.6x103 pg/cell/day for SS. Naturally secreted HCPs include proteins present in extracellular matrix and proteins excreted through exosomes. Overall, the numbers of secreted proteins were upregulated in SS perfusion process while intracellular HCPs are the main contributor to the HCP compositions of NSS perfusion process.