Development of weak cation exchange membrane adsorbers for protein capture
Lai, Yung Priscilla
MetadataShow full item record
To reduce therapeutic protein production costs in bioprocessing such as monoclonal antibodies production, the downstream purification step needs to be optimized. Protein A resins for the chromatographic purification of such proteins have long been used but are expensive, diffusion limited, and may leach into the stream due to proteolysis. Weak cation exchange membrane adsorbers are a viable alternative, enabling lower costs and lower mass transfer limitations from diffusion for higher throughput. The proposed synthesis route to develop weak cation exchange membrane adsorbers was to graft poly(acrylic acid) directly from a regenerated cellulose (RC) membrane surface by aqueous activator regenerated by electron transfer atom transfer radical polymerization (ARGET ATRP), a type of reversible deactivation radical polymerization (RDRP) technique. The technique allows for polymerization under limited amounts of oxygen in aqueous media and for controlled polymerization. The initiator, 2-bromoisobutyryl bromide (BiBB), was first immobilized on the RC membrane followed by ARGET ATRP. The first part of the work investigated the immobilization of BiBB on RC membranes in a heterogeneous acylation reaction. The optimal ratio of BiBB to triethylamine (TEA, used in the reaction to neutralize the hydrogen bromide by-product) was 1/0.67. The effect of NaOH treatment on the methanol-washed RC membrane (i.e. no NaOH or 2 M NaOH), BiBB quantity used per membrane disc for immobilization (0.41, 0.74, or 2.67 mmol), and immobilization solvent (N,N-dimethylformamide, DMF, or tetrahydrofuran, THF) on BiBB immobilization was studied. Relative and absolute immobilized BIBB quantities were studied using attenuated total reflectance Fourier transform infrared spectroscopy (ATR-FTIR) peak area ratios and the degree of substitution (DS) of the RC membranes were calculated from gravimetry, respectively. The highest BiBB immobilization was obtained with the higher BiBB quantity used per membrane disc for immobilization, 2 M NaOH-treated, methanol-washed RC membrane, and DMF as immobilization solvent (ANOVA, 95% confidence level). The uniformity of BiBB immobilized across the surface was found to be improved when larger quantities of BiBB were added to the reaction (i.e. 0.74 mmol per membrane disc or greater). For ARGET ATRP, the amine ligand, 2,2′-bpyridine (bpy), was selected at a CuBr2/bpy/ascorbic molar ratio of 1/2/2 based on CuBr2 reduction studies by ascorbic acid. Using bpy, various ARGET ATRP reactions were conducted to determine the polymerization conditions that produced membranes with high and low poly(acrylic acid) contents (PAA, polymer ATR-FTIR peak area ratios of ~0.7 and ~0.48, respectively). The polymerization conditions that produced high PAA contents and low PAA contents were tested for their static protein binding capacity with lysozyme. The lysozyme static protein binding capacities were 235 mg mL-1 and 510 mg mL-1 for the poly(acrylic acid)-grafted RC membranes (PAA-g-RC) with low and high PAA contents, respectively. The second part of the work investigated the effect of RC membrane treatment conditions and BiBB quantity used per membrane disc on immobilized BiBB in a 32 factorial design (i.e. methanol-washed RC membrane, 0nD; methanol-washed RC membrane with DMF storage for two weeks prior to immobilization, 0D; and methanol-washed RC membrane with 0.5 M NaOH treatment along with DMF storage prior to immobilization, 0.5D; with either 0, 0.74, and 2.67 mmol BiBB used per membrane disc). Energy dispersive x-ray spectroscopy (EDX) and thermogravimetric analysis (TGA) complemented ATR-FTIR and gravimetry in confirming the presence of immobilized BiBB. ANOVA analysis (95% confidence level) of the relative BiBB quantities determined from the ATR-FTIR peak area ratios and the absolute initiator quantities expressed as the degree of substitution (DS) determined by gravimetry confirmed the increased amount of immobilized BiBB when 0D and 0.5D treatments on the RC membranes were used over 0nD. Moreover, increasing the BiBB quantity used per membrane disc gave an increased amount of immobilized BiBB. The BiBB-modified 0D membrane using 2.67 mmol BiBB per membrane disc (0D 2.67) was then used for subsequent ARGET ATRP due to the high DS values without the need for the extra NaOH treatment. The final part of the work investigated the characteristics of the grafted poly(acrylic acid) (i.e. conversion, grafting ratio, and theoretical number-average molecular weight) via gravimetry and ATR-FTIR. Low monomer conversions of 1.8-3.4 % were achieved, resulting in oligomeric theoretical number-average molecular weights (682-1052 g mol-1). However, grafting ratios of 109-202 % were obtained and the PAA-g-RC membrane swelled from a 47 mm diameter circle into an ellipse with a 60 mm long major axis and a 50 mm long minor axis. The PAA-g-RC membrane swelled 8 times its own weight in pH 5 acetate buffer. Finally, the dynamic protein binding capacity for human immunoglobinG (IgG) at 10% breakthrough (DBC10%) for the 0D, 0D 2.67, and PAA-g-RC membranes were measured. 0D and 0D 2.67 were statistically similar while PAA-g-RC membranes (47 mm diameter discs) achieved the highest DBC10% (4.4, 5.7, and 30 mg mL-1, respectively, t-test with 95% confidence level). Successful protein capture was therefore achieved with the weak cation exchange membrane adsorbers developed in this work based on RC membrane supports.