Design, modeling, and optimization of a multicomponent separation process for insulin purification using reversed phase chromatography

Pei-Lun Chung, Purdue University

Abstract

Insulin is one of the top five recombinant therapeutic proteins in global sales. It is the only recombinant protein produced at the tons per year scale. The production of insulin requires many chromatographic steps. A key downstream polishing step, the reversed phase chromatography, RPC, removes the structurally similar impurities before size exclusion chromatography and crystallization. The design and optimization of RPC for insulin purification process has been challenging in the preparative scale. The lack of systematic approach in protein purification results in empirical, time consuming, and expensive screening of various loading and operating conditions. The goal of this work is to develop strategies for designing new insulin purification processes with high purity, high yield, and high productivity. A model-based approach combining experiments, theoretical analysis, and mathematical modeling is used to achieve the research goal. This work is composed of two parts: model development and process design. The modified reversed phase modulator isotherm and the general rate model were tested to describe lispro insulin and its impurities. To estimate the intrinsic model parameters, this study developed a systematic and efficient method that only requires small pulse linear gradient elution runs with two gradient slopes and frontal experiments with three organic fractions. With the same set of intrinsic parameters, the model predicted closely with the experimental data over a wide range of loading and operating conditions: loading volume up to 4 column volumes, acetonitrile fraction from 0.19 to 0.43, linear velocity up to 6 cm/min, and in various elution modes such as frontal, linear gradient, isocratic, and stepwise elution. The second part of this study applied the verified model and parameters to elucidate the competitive adsorption and the competing mass transfer mechanisms in lispro insulin purification. A method combining the dimensionless group analysis and the model-based parameter studies is developed in this study for designing new lispro insulin purification process with high purity, high yield, and high productivity. Compared to linear gradient elution processes, the new stepwise elution processes can achieve close to 99% yield, double the productivity, increase the product concentration, reduce the solvent consumption by half, and maintain the high purity of 99.6%.

Degree

Ph.D.

Advisors

Wang, Purdue University.

Subject Area

Chemical engineering

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS