Development of mammalian production cell lines expressing CNTO736, a glucagon like peptide‐1‐MIMETIBODY^TM: Factors that influence productivity and product quality

Abstract

In an attempt to develop a high producing mammalian cell line expressing CNTO736, a Glucagon like peptide-1-antibody fusion protein (also known as a Glucagon like peptide-1 MIMETIBODY), we have noted that the N-terminal GLP-1 portion of the MIMETIBODY was susceptible to proteolytic degradation during cell culture, which resulted in an inactive product. Therefore, a number of parameters that had an effect on productivity as well as product quality were examined. Results suggest that the choice of the host cell line had a significant effect on the overall product quality. Product expressed in mouse myeloma host cell lines had a lesser degree of proteolytic degradation and variability in O-linked glycosylation as compared to that expressed in CHO host cell lines. The choice of a specific CHOK1SV derived clone also had an effect on the product quality. In general, molecules that exhibited minimal N-terminal clipping had increased level of O-linked glycosylation in the linker region, giving credence to the hypothesis that O-linked glycosylation acts to protect against proteolytic degradation. Moreover, products with reduced potential for N-terminal clipping had longer in vivo serum half-life. These findings suggest that early monitoring of product quality should be an essential part of production cell line development and therefore, has been incorporated in our process of cell line development for this class of molecules.

Proteomic Analysis of Host Cell Protein Dynamics in the Culture Supernatants of Antibody-Producing CHO Cells

“…The relative values of protein abundance were estimated by spectral counting method, followed by normalized spectral abundance factor (NSAF) method for reliable normalization20. Since large proteins usually tend to contribute more to the ‘peptide/spectra’ than small ones, the NSAF method is useful to minimize the variations originated from protein length.…”

“…The fragmentation of therapeutic proteins by HCPs in the culture supernatants has been observed in rCHO cells producing mAbs9, Fc-fusion proteins71920, interferon γ21, factor VIII8, and erythropoietin (EPO)22. Cathepsins in the culture supernatants have been reported as enzymes that break down proteins23242526.…”

Characterization of the proteases involved in the N‐terminal clipping of glucagon‐like‐peptide‐1‐antibody fusion proteins

“…Unlike the intact antibody (Ab) molecules that are remarkably resistant to proteolytic clipping, glycoproteins, due to their relatively exposed three‐dimensional structure, can be exquisitely sensitive to proteases and other degradative activities (Varki et al,1999). We and others have previously demonstrated that there is a correlation in the amount of glycosylation to the degree of susceptibility to proteases (Dorai et al,2009; Rutledge and Enns,1996). In fact, this knowledge is used in designing a longer lasting Aranesp (Elliott S et al,2003) or Enbrel with extended circulating half‐life (Gottlieb and Bos,2002).…”

“…Table 2 shows that the amount of serine protease present in the spent medium ranged from less than 0.2 pg/mL to more than 380 pg/mL but there was no correlation of the amount of N‐terminal clipping to the level of serine proteases present. For example, very little if any clipping could be detected in CNTO736 expressed in myeloma cell lines NS0 and Sp2/0 (Dorai et al,2009), but these host cell lines accumulate serine proteases that are comparable with that of CHOK1SV derived clones. Moreover, three different clones of CHOK1SV expressing CNTO736, namely, A37 (C1672C), B3, and C64 (C1352C), which had significantly different amounts of N‐terminal clipping, had comparable levels of serine proteases (Figure 2B and Table 2).…”

“…Because the products generated in different hosts have different levels of N‐terminal clipping, we hypothesized that incomplete hinge‐region glycosylation makes the fusion proteins susceptible to N‐terminal clipping. Previously, we had shown that there was a correlation in the amount of O‐linked glycosylation to the degree of clipping of the fusion protein derived from CNTO736 (Dorai et al,2009). Because the location of the major clip (amino acid 30 from the N‐terminal end, between arginine and glycine, Figure 1B) is in close proximity of the site of O‐linked glycosylation near the hinge section, it is possible that the O‐linked glycosylation causes a steric hindrance such that proteolytic enzymes cannot access the proteolytically susceptible site.…”