Date of Award
Doctor of Philosophy (PhD)
Chemical and Biomolecular Engineering
Dr. Scott M. Husson, Committee Chair
Dr. Christopher L. Kitchens
Dr. Mark A. Blenner
Dr. Eleanor W. Jenkins
This dissertation presents work on the design and synthesis of a new membrane chromatography material, the description of its protein binding behavior using a thermodynamic adsorption isotherm model, and the application of the new membrane material in biologics downstream recovery and purification processes. As protein titers continue to increase dramatically in upstream biomanufacturing, innovations in downstream purification are not keeping pace, resulting in manufacturing capacity constraints and high production costs. Chromatography is the key unit operation used in several steps of the downstream purification platform. Traditional resin bead chromatography, while effective and reliable for isolation and purification of proteins, limits the process productivity and affects product quality. In the case of ion-exchange chromatography steps, traditional materials have a limited operating window, which requires the implementation of buffer exchange steps between chromatography steps to condition the feed for optimal performance in each step. Innovations in purification technologies that can dramatically increase the productivity of existing facilities and simultaneously lower the manufacturing cost are needed. In this dissertation, a new multimodal membrane chromatography material is introduced that could greatly improve the process productivity and product quality.
Chapter 2 describes my work to develop the first cation-exchange multimodal membrane (MMM) adsorber in a two-step synthesis. Surface-initiated atom transfer radical polymerization was used to graft polymer chains containing epoxy side groups from the surface of a commercial macroporous regenerated cellulose membrane. Then, the multimodal functional groups were introduced through an epoxide ring opening reaction by 4-mercaptobenzoic acid. Permeability and protein (IgG) binding capacity measurements showed that polymerization time can be used to achieve high binding capacity (up to 180 mg IgG/mL) while maintaining adequate permeability of the membrane. Kinetic studies with a model cellulose nanolayer suggest that the degree of polymer grafting directly affects the static binding capacity of the multimodal membrane. Measured equilibrium IgG binding capacities using protein solutions at different pH values and ionic strength values demonstrated that both Coulombic and hydrophobic interactions occur between the protein and the membrane. Characteristic of multimodal adsorbers, the multimodal membranes maintained significant binding capacities in excess of 90 mg IgG/mL at ionic strength values that are typical for elution buffers used in multi-stage bioseparation processes. For sodium citrate, a conventional salt used in elution buffers of Protein A columns, increasing ionic strength had only a minor effect on the IgG binding capacity. These results indicate that the newly developed multimodal membrane has great potential to compete with more traditional cation-exchange materials following the Protein A purification step in the downstream processing of antibody products. In addition to work with macroporous membrane supports, a new method was developed to coat cellulose nanolayer on silicon wafer to mimic the morphology of cellulose membrane surface. A kinetic study with the model cellulose nanolayer showed that the polymer thickness is proportion to the static binding capacity of the multimodal membrane. This model substrate could be useful for future membrane design efforts.
Chapter 3 describes my work to evaluate the effects of different salt types (kosmotropic, neutral, chaotropic salts) and ionic strength on IgG binding. Dynamic binding capacity measurements were performed over a range of flow rates. A thermodynamic model was used to provide insights on the nature of protein-MMM interactions and to predict binding capacities under non-test conditions, which is important for limiting the number of experiments needed for process development. It was determined that the rate limiting step of IgG adsorption on the MMM is the reaction rate of IgG binding with the multimodal ligands, rather than the mass transport of protein molecules. Thus, while high load productivities were achieved, improvements in membrane design leading to faster adsorption kinetics would enable still higher productivities. The results of this part of the study indicate that multimodal membrane bind-and-elute chromatography can be a highly productive and scalable process. The ability to work at high salt concentrations may reduce the number of steps in the protein purification train, improving product quality, enhancing manufacturing capacity in existing facilities, and reducing the cost of downstream purification.
Chapter 4 describes my work to purify monoclonal antibodies from Chinese hamster ovary (CHO) cell culture supernatant using the newly designed multimodal membranes. When used after a size exclusion desalting step, the MMM column was effective for recovery of human IgG1 from CHO cell culture supernatant, and neutral pH elution yielded a product pool with purity (>98%) and HCP level (n.d.) equivalent to what could be achieved by Protein A chromatography. Dynamic capacities at 1 CV/min were higher for the MMM column than the commercial Protein A resin column, which is important for reducing the number of cycles needed for purification of a batch and thereby increasing process throughput. Whereas it is unlikely that Protein A chromatography will be replaced anytime soon for mAb capture step purification, this part of my work showed that MMM chromatography following a simple desalting step appears to be an excellent option for capture step purification of proteins when Protein A cannot be used, e.g., for pH sensitive mAbs or biologics lacking the Fc binding domain.
Overall, this dissertation demonstrates the potential of multimodal cation exchange membranes for the effective and high-productivity purification of proteins from cell culture supernatant, either in bind-and-elute mode following a desalting step, or in polishing step mode following a Protein A capture step. Its ability to operate over a wide range of conditions may reduce the number of steps needed to purify proteins, which would increase the overall process productivity and also improve the product quality.
Wang, Juan, "Development of a New Multimodal Membrane Adsorber and Its Application in Chromatographic Bioseparations" (2016). All Dissertations. 2294.