Identification of the limitations on recombinant gene expression in CHO cell lines with varying luciferase production rates

Abstract

Mammalian cell lines are currently employed as one of the main cellular factories for the expression of recombinant protein-based drugs. The establishment of high-producing cell lines typically begins with a heterogeneous starter population of cells, from which the highest producing cells are selected via empirical approaches. This approach is time consuming, and is likely to encounter natural upper limits imposed by the inherent biology of the cell lines in question. In an attempt to understand both the nature of the variability in populations of cells transfected with recombinant protein encoding DNA and the natural mechanisms of productivity limitation, we developed protocols for the detailed investigation of gene expression pathways in such cell lines. This novel approach was then applied to a set of clonal CHOK1 cell lines producing recombinant luciferase with varying productivities. Our results show that the initial limitation in these cell lines is at the transcriptional level, however in the highest producing cell line post-translational mechanisms affecting both protein turnover and protein folding become severely limiting. The implications for the development of strategies to engineer cells for enhanced recombinant protein production levels are discussed.

mTORC1 signalling and eIF4E/4E-BP1 translation initiation factor stoichiometry influence recombinant protein productivity from GS-CHOK1 cells

“…For example, the viable cell number in the CHO52 cell line declined from day 8 to day 9 much more than the other cell lines. In terms of productivity, Null8 is a non-producing cell line that has been through the same GS selection process as the mAb-producing cell lines, but lacks the heavy and light chain IgG genes, while CHO52 was ranked as a low producer and CHO137 and CHO42 were considered high producers for the present study with their estimated specific production rates (pg/cell/h), having previously been estimated as 0.032, 0.49 and 0.31, respectively [8]. Western blot analysis for the amount of mAb in the cell culture supernatant on different days throughout culture confirmed the relative productivities of these cell lines, with CHO42 being the highest producer and CHO52 the lowest (Figure 1B).…”

“…Previous reports have suggested that the yield of recombinant proteins from cultured mammalian cells is in part attributed to translation efficiency [8,44]. mTORC1 exerts its influence on translation via the mTORC1 effectors 4E-BP1, p70 S6 kinase and eEF2K.…”

“…In this regard, improved growth medium and feeding strategies have resulted in dramatic increases in the maximum viable cell concentration and IVC achievable for mammalian cells in the bioreactor, and this has been associated with extended culture viability and enhanced recombinant protein yields [2]. Indeed, achieving high yields of recombinant protein of a clinically acceptable quality is dependent on a multiplicity of parameters, including the choice of the host expression system, the achieved viable cell mass [3,4], gene copy number [5], site-specific integration [6,7], the cellular processes responsible from gene to protein in the synthesis of the desired product [1,8], the bioreactor environment (e.g. nutrients and oxygen levels) [2,9], the authenticity and homogeneity of the product [10,11] and the yield and success of downstream processing [12,13].…”

Rapid high-throughput characterisation, classification and selection of recombinant mammalian cell line phenotypes using intact cell MALDI-ToF mass spectrometry fingerprinting and PLS-DA modelling

“…In order to address this issue, a number of proteomic, transcriptomic and metabolic based studies have now been undertaken and the subsequent data used to develop models to predict the phenotype of a given cell line (e.g. Clarke et al, 2011;Clarke et al, 2012;Doolan et al, 2013;Jacob et al, 2010;Mead et al, 2009;Mead et al, 2012;Meleady et al, 2011;Sanchez et al, 2014;Selvarasu et al, 2012), however these models are usually developed from cell line data at the end of the cell line development process.…”

“…In order to address this issue, a number of proteomic, transcriptomic and metabolic based studies have now been undertaken and the subsequent data used to develop models to predict the phenotype of a given cell line (e.g. Clarke et al, 2011;Clarke et al, 2012;Doolan et al, 2013;Jacob et al, 2010;Mead et al, 2009;Mead et al, 2012;Meleady et al, 2011;Sanchez et al, 2014;Selvarasu et al, 2012), however these models are usually developed from cell line data at the end of the cell line development process.The goal of this work was to develop a screening system that would allow the selection of highly productive cell lines for monoclonal antibody (mAb) production early in the cell line development process that would use substantially less resource to achieve the same or better success rate as current methods. The vision was to be able to select a small number of cell lines based upon the analysis of data generated in multi-well plates, and take these straight to a lab-scale bioreactorevaluation stage (10 L) with a high probability that the selected cell lines were highly productive.…”