CHO cell engineering to develop new chassis for enhanced difficult to express recombinant protein production

Recombinant biotherapeutic protein based therapies have been developed for the treatment of a range of diseases with several therapies successfully obtaining 'blockbuster' status of sales >US $1 Billion pa. Many of these molecules are IgG monoclonal antibodies (mAb) manufactured from in vitro cultured mammalian cell expression systems due to their ability to correctly synthesise, assemble, post-translationally modify and secrete the target molecule. Chinese hamster ovary (CHO) cells are the mammalian expression cell line of choice because of their ability to undertake 'human like' post-translation modifications, grow to high viable cell concentrations under fed-batch conditions with isolated cell lines able to produce high yields of target recombinant protein (up to 10 g/L). The CHO cell platform is mature for the production of well-expressed mAbs, but many mAbs and new format molecules such as fusion proteins and other designs remain difficult to express (DTE) in this, or any other host. This thesis describes approaches to engineer the CHO cell to manipulate glycolytic flux to enhance flux from pyruvate to acetyl-CoA, a metabolite that feeds the TCA cycle and fatty acid metabolism, both metabolic pathways known to impact CHO cell phenotype regarding growth and secretory recombinant protein productivity. The pyruvate dehydrogenase (PDH) complex, and pyruvate dehydrogenase kinases (PDKs) that regulate this, were targeted to investigate whether their manipulation impacted CHO cell growth and recombinant protein productivity using a model biotherapeutic DTE protein, Etanercept. The three PDH subunits, E1α, E1β and E2 were cloned, with and without N- and C-terminal tags and stably over-expressed in two host CHO cell lines, CHOS and CHOK1. Subunits were expressed individually and together. A triple mutant of the E1α subunit was generated where the three Ser residues phosphorylated by PDKs to regulate its activity were mutated to Ala residues. C-terminally tagged subunits were observed by western blot to be stably expressed better than N-terminally tagged. Immunoprecipitation experiments using the tag to determine if exogenous subunits were able to form complexes with other native subunits was inconclusive. Knockout (KO) of 3 of the PDKs that phosphorylate and inactivate the E1α subunit was undertaken using CRISPR/Cas9 paired gDNA gene editing approach. Guides that resulted in apparent KO were identified and then KO pools of the three individually targeted PDKs and a triple KO were generated. Multiple attempts to isolate single cell clones of KO cells did not recover and thus KO pools were generated. Reduction in gDNA and mRNA was observed in pools, particularly for PDK1 and PDK2, however complete KO in pools was not observed and no impact on PDK protein expression was apparent. Minipools that did not recover under these study conditions showed evidence of a greater target PDK KO. During batch culture and when transiently expressing Etanercept, the apparent gDNA amount of the PDK genes targeted for KO changed. This likely reflects the nature of the KO pools and that certain populations within the pool were preferentially advantaged under the culture and recombinant protein load allowing them to outgrow others. PDH over-expression resulted in cell pools achieving higher maximum viable cell concentrations and prolonged culture viability under batch culture conditions, but this did not lead to increased transient Etanercept production. Similarly, the PDK KO pools instead resulted in lower apparent cell specific Etanercept production. In both the PDH subunit and PDK KO pools there was evidence of changes in the glycolytic flux as determined by measuring extracellular glucose and lactate concentrations, but there was also cell and pool specific impacts. In conclusion, CHO cells can tolerate PDH subunit over-expression whilst, at least under the conditions investigated here, seem very susceptible to PDK KO. The evidence from this work suggests it is unlikely that manipulation of these components of energy metabolism using the approaches outlined in this thesis can deliver a new CHO chassis with a large enhancement in cell grow or DTE recombinant protein that cannot be achieved from the diversity that already exists within the CHO cell hosts, however alternative approaches may allow this to be achieved.