(436b) CAPACITY, MASS TRANSFER KINETICS, FLOW-PROPERTIES and OPTIMIZATION of A NEW PROTEIN-A ADSORBENT: UNOsphere SUPrA

The paper discusses the experimental determination and modeling of IgG binding capacity and mass transfer kinetics and of the flow properties of a new Protein A adsorbent based on a macroporous resin. The new adsorbent consists of polymeric beads based on hydrophilic acrylamido and vinyl monomers with a pore structure optimized to allow favorable interactions of recombinant Protein A, coupled to the resin, with IgG. The particles are sized with an average diameter of 57 um and a narrow particle size distribution. This combination results in pressures of less than 2 bar at 600 cm in 20 cm long columns. An equilibrium adsorption capacity of 43 mg/mL particle is obtained from batch equilibrium experiments. The mass transfer kinetics for both adsorption and desorption conditions, obtained from transient batch adsorption and microscopic experiments, is consistent with a pore diffusion model with an effective pore diffusivity of 6.0x10-8 cm2/s, essentially independent of protein concentration. Flow properties are determined experimentally in laboratory, preparative, and process scale columns, 1, 20, 30, and 45 cm in diameter. A modified version of the model of Stickel and Fotopoulos [Biotechnol. Progr., 17, 744 (2001)] is developed to describe the packing and pressure drop behavior resulting in a packing compressibility parameter that is approximately independent of column diameter over the range studied. The cyclic adsorption-desorption behavior of IgG in single column systems is thus readily predicted from these measurements allowing the definition of a productivity curve for the optimization of operating conditions. A final component of this work is the extension and use of the new adsorbent in periodic countercurrent systems where multiple columns are used in a merry-go-round sequence to maximize the utilization of the available adsorption capacity. Operation with two or three columns in series is shown to provide substantial productivity improvements, without excessive equipment complexities. A numerical model used to predict the periodic countercurrent behavior is found to be in good agreement with laboratory scale experiments.