Model-based optimization strategy for intensification in the chromatographic purification of oligonucleotides

Abstract

Oligonucleotides (ONs) are acquiring clinical relevance and their demand is expected to grow. However, the ON production capacity is currently limited by high manufacturing costs. Since the purification of the target ON sequence from molecularly similar variants represents a major bottleneck, this work presents a resource-effective strategy for the optimization of their preparative reversed-phase chromatographic purification. First, a model based on the equilibrium-dispersive theory was introduced to describe the chromatographic operation. Considering a deoxyribose nucleic acid with 20 nucleobases as case study, a genetic algorithm was developed to efficiently determine the adsorption isotherm and mass transfer parameters for the target ON and impurities. After the estimation of these parameters, a strategy for the in-silico optimization of the operation was established. The product collection window, gradient duration, and resin loading were considered as process variables and their influence on yield and productivity was investigated after setting a purity specification of 99.0%. The optimal process parameters identified through this analysis were experimentally verified, confirming the reliability of the model, calibrated with only 5 experimental runs. In addition, this optimal setpoint was exploited to design the multicolumn countercurrent solvent gradient purification (MCSGP) of this ON mixture, which allowed to boost the yield of the process and to work at cyclic steady state, while respecting the purity constraint. This study confirmed the potential of this in-silico optimization strategy in both improving the performance of the traditional single-column operations and in the rapid development of multicolumn processes.