Expression and characterization of a therapeutic monoclonal antibody in mammalian cells

Abstract

The advent of therapeutic recombinant proteins has revolutionized modern medicine. Since the approval of recombinant insulin in 1982 to treat diabetes, many other recombinant proteins have emerged for a diversity of previously incurable conditions. In these years, the manufacturing processes have greatly evolved, but have also often disregarded product quality, an issue only recently addressed and currently a major challenge in biopharmaceutical industry. Regulatory agencies require now a tight control of product quality during production, which typically involves characterization of the protein’s glycosylation, known to affect properties like serum half-life, biological activity and immunogenicity. However, control of this property is not simple due to its intrinsic high variability and its high sensitivity to small changes in the manufacturing process, in ways that are far from being understood. Therefore, the development of products of consistent high-quality requires a deeper understanding on the effect of the production parameters/ processes on glycosylation. In this context, the work described in this thesis intended to contribute to this field of knowledge by evaluating changes on glycosylation of a monoclonal antibody (mAb) produced by Chinese hamster ovary (CHO) cells during different stages of process development, namely cell line transfection, serum-free media adaptation and culture. To accomplish the main goal, a mAb-producing CHO-K1 cell line was initially established, through a novel expression system (OSCARTM), potentially more expeditious than the traditional methods. OSCARTM was indeed considerably faster, requiring about two months to obtain producer cells compared to the minimum six months usually required. OSCARTM did not affect cell growth and reached high levels of productivity (10 pg/cell/day) without requiring the typical processes of amplification. The levels of productivity were initially difficult to stabilize, with a 10- fold decay observed in the first weeks of culture, but remained stable thereafter for at least two years. Furthermore, minigene selection was critical and seemed to be cell/product-specific. This work also evaluated two methods for the initial selection of the highest producer clones obtained after transfection, based on: absorbance values of supernatants; or mAb productivity calculated using mAb yield determined by enzyme-linked immunosorbent assay and cell concentration obtained from cell counting with a haemocytometer. It was shown that the methodology chosen is highly influential on the product yield achievable in the final process of production, and that the productivity-based method, albeit more laborious, is much more reliable than the commonly used absorbance method. Additionally, the highest-producer clones were evaluated for mAb glycosylation by high-performance liquid chromatography, with no differences detected. Glycosylation was also assessed during the adaptation of mAb-producing cells to serum-free medium using a gradual methodology and supplementation with trace elements. This process was long and highlighted the importance of using proper media supplements, avoiding aggressive procedures (centrifugation, enzymes), and giving cells enough time to adapt at each stage. More importantly, it was shown that adaptation alters glycosylation: in the middle stages, fucosylation decreased and galactosylation, sialylation and glycoform heterogeneity increased, with the first two being positive and the last two undesirable for efficacy; in the final stages and after full adaptation, fucosylation returned to the initial (serum-supplemented) levels, while galactosylation, fucosylation and heterogeneity decreased, with the last two being positive. Divergences between the stages were related to lower cell concentrations and viabilities in the middle stages of adaptation, and to a shift of the growth mode of cells from adherent to suspended. The impact of a microporous microcarrier culture on mAb glycosylation was evaluated and compared to common T-flask cultures. The influence of several culture conditions was assessed, including initial culture volume and cell concentration, rocking mechanism and speed, and culture vessel. The microcarrier cultures led to a different mAb glycosylation compared to that obtained in the T-flask culture, attributed to different mAb productivities, as well as to the use of rocking and the generation of microenvironments (pH, accumulation of extracellular enzymes) in the microcarrier cultures. Specifically, higher galactosylation and decreased fucosylation, both beneficial changes, and a variable sialylation were found in the microcarrier cultures. Sialylation was more sensitive to the culture parameters, particularly the type of culture vessel, being almost absent in shake flask cultures. In addition to this advantageous modification, shake flasks also led to a more homogeneous glycosylation, potentially due to improved cell densities. In summary, the work described in this thesis contributes to a better understanding of how glycosylation is affected by technologies used in process development, highlighting the need to implement quality control at early stages. By increasing knowledge on this subject, it will be possible to control and improve the quality and efficacy of therapeutic proteins, which will ultimately lead to their administration at lower doses and frequency