| Purified viruses can serve as gene delivery vectors and are the active ingredients in many vaccines. Additionally, for purification of therapeutic proteins, not only must the separation steps be capable of removing viruses, the regulatory agencies (FDA, EMEA, etc.) also require understanding of the removal mechanisms to ensure robustness. Ion-exchange chromatography (IEC) is the most widely used method for protein purification and it has been demonstrated to be more efficient for virus purification than density gradient ultracentrifugation, the traditional method for virus purification. Chromatography of viruses has been considered to be a black box and designing IEC processes is empirical and time-consuming. The goal of this work was to improve the understanding and process development of virus IEC.;We developed light scattering and batch binding techniques for rapid optimization of pH and ionic strength. The model virus was human adenovirus type 5 (Ad5), the most commonly used virus in gene therapy. Ad5 was found to aggregate within about 1.5-2 pH units of its isoelectric point. Due to strong repulsion between virus particles at low ionic strengths, binding of Ad5 to a model stationary phase attains a maximum with increasing ionic strength. Besides pH and ionic strength, choice of a stationary phase is another key variable in IEC. In addition to typical resins used for protein purification, which have pores that are smaller than most viruses, we evaluated perfusive resins and monoliths by single-component pulse and step response studies. The evaluation was performed with small molecules, proteins, and the virus to examine how increasing bioparticle size affects chromatographic behavior, and the results were related to stationary phase structural parameters. Binding capacities on a model monolith were predicted from the pore size distribution. Modeling of breakthrough curves on a monolith showed that the experimentally observed slow approach to full saturation is due to the distribution of pore sizes rather than diffusive limitations. For larger bioparticles, retention is generally greater and the transition from a non-eluting to a non-binding state with increasing ionic strength is more abrupt. The drag imposed on adsorbed bioparticles by the moving liquid, which may contribute to desorption, is calculated to be more important for larger analytes. With increasing retention, the height equivalent to a theoretical plate increased significantly more for larger bioparticles. For proper analysis of peak data, an algorithm was developed to account for temporal distortion and extra-column effects. The skew caused by temporal distortion was found to be inversely related to the number of theoretical plates and, therefore, bed height.;Under loading conditions, Ad5 is unable to enter the internal pore space of perfusive resins because virus particles that bind near the external surface prevent others from entering the pores. A more significant issue was observed for both perfusive resins and monoliths: even under non-binding conditions, with increasing flow rate, a progressively larger fraction of virus particles injected onto the columns does not elute. Calculations show that the cause is convective entrapment of virus particles in pores with constrictions. As expected from the developed theory, the entrapment is reversible upon flow rate reduction and does not occur on small-pore resins, nor for proteins. This finding represents a new retention mechanism in chromatography. The theory has multiple implications for chromatography of large biologics on perfusive resins, monoliths, and membranes in terms of product loss, underestimation of analyte concentration, and column cleaning. Entrapment can be controlled by adjusting the flow rate. |
