Secretory Pathway Engineering of CHO Cells to Improve Cytokine Production

Recombinant therapeutic proteins have been widely produced to treat diseases like cancer, autoimmune disorders, and infections. Cytokines, a family of small proteins with immunomodulatory properties, are crucial molecular messengers that facilitate efficient and multifaceted immune responses. They have garnered significant interest in therapeutic applications since the 1980s. Following the marketing authorization of the first recombinant cytokine products—Interleukin-2 (IL-2) and Interferon-alpha (IFN-α)—numerous cytokine-based therapies have been investigated in clinical studies for their therapeutic potential. However, the clinical application of these therapies is often limited by severe side effects and the short half-life of cytokines in the bloodstream.

To overcome these challenges, scientists have developed novel cytokine formats using protein engineering and glycoengineering approaches to enhance therapeutic efficacy by achieving more targeted modes of action without off-target effects. While Escherichia coli has been employed as a production host for some cytokines, mammalian expression systems offer advantages, particularly for cytokines that require post-translational modifications. These systems help mitigate the risk of bacterial endotoxins, which can trigger unwanted immune responses and make purifying cytokines challenging. The complex structure and glycosylation requirements of many new cytokine formats also necessitate advanced expression systems.

Chinese hamster ovary (CHO) cells are the most widely used expression system due to the advancements in cell line engineering. However, certain biologics, such as cytokines, cannot be produced efficiently using standard biomanufacturing approaches. The secretory pathway has been identified as a crucial factor impacting the quality and quantity of the produced protein. While some studies have enhanced some protein titers, it has become evident that each protein may require unique optimization strategies. Recent advances in omics have enabled researchers to characterize cells with different phenotypes, such as varying protein yields, at multiple levels, including the genome, transcriptome, and proteome. These analyses have provided insights into the underlying mechanisms that govern high protein secretion and identified the transcriptome as a major predictor of protein production. Comparative transcriptomics studies of high- and low-producer cells have identified candidate genes that, when modulated, can enhance recombinant protein yields.

This work proposes a meta-analysis of transcriptomics analysis from human cells to identify candidate genes co-expressed with cytokines in high cytokine-producing cells, leading to hypotheses about cell modification strategies that may help improve cytokine quantity and quality. The next phase involves experimental work, where we verified the candidate genes that co varied with cytokine expression across cell types naturally expressing the cytokines of interest. These candidates included proteins identified here that may impact protein titer and glycosyltransferases, which are expected to affect the glycan structures of each cytokine. Then, we screened these candidate genes by transiently expressing them in stable cytokine-producing CHO cells. Finally, we investigated the effects of the glycosyltransferases, which were identified by assessing their varying expression levels and evaluating the impact of the glycans on granulocyte-macrophage colony-stimulating factor (GM-CSF) bioactivity by quantifying GM-CSF activity among six different glycoforms.

In summary, this thesis represents a method for utilizing omics data to enhance the production of challenging cytokines. It also demonstrates that glycoengineering can optimize the quality of biotherapeutics, thereby making cytokines more accessible to patients.