BioTalk

Biological Manufacturing Excellence

(3rd – 4th June 2025, Framingham, US)

Scott H. Altern

Senior Scientist
AbbVie

Scott H. Altern completed his PhD in Chemical Engineering at Rensselaer Polytechnic Institute (RPI) in Professor Steven Cramer's group (2023). His PhD research was centered on advancing the state-of-the-art in mechanistic modeling of preparative column chromatography and high-throughput process development. Currently, his role is at AbbVie in Worcester, MA where he focuses on mechanistic modeling of preparative chromatography, computational fluid dynamics, and molecular modeling to directly aid in process development. Scott is also highly involved in using high-throughput (HT) experimentation to expedite process development and model generation. Overall, Scott aims to transform industrial bioprocess development using automation and multiscale modeling.

Abstract

In this work, we employed HT instrumentation and multiscale modeling to support early-stage process development for two monoclonal antibodies (mAbs). These molecules had nearly the same isoelectric point but were found to behave differently in CEX. Both antibodies were screened for their dynamic binding capacity (DBC) on POROS XS RoboColumns across a wide range of loading pH and conductivity. Curiously, opposing DBC trends with loading pH and conductivity were observed between the mAbs. Generally, increasing loading pH and conductivity, beyond a certain threshold, results in a reduced DBC in CEX systems. This was consistent with the behavior for mAb B, but contradicted the trends seen for mAb A. Molecular modeling was then employed to identify the molecular features responsible for the unusual screening results. Homology models were constructed for each of the antibodies, from which molecular descriptors were calculated. Descriptors such as net charge and CDR positive charge surface area were correlated with DBC values to confirm that the degree of exclusion was directly related to molecular charge properties. Additionally, mechanistic model parameters were regressed from RoboColumn breakthrough curves to provide additional insight into the unique chromatographic behavior. For instance, trends in surface diffusivity with loading conductivity and pH were examined to determine their contribution to CEX loading performance. Further experimental studies were conducted in order to elucidate the molecular origins of the observed trends. Dynamic light scattering (DLS) experiments were carried out to determine the protein-protein interaction constant over a range of pH and conductivity. Interaction strength was found to be correlated with DBC, suggesting that repulsive forces negatively affect DBC and are dependent on solution conditions. In summary, this study used high-throughput screening and multiscale modeling in conjunction to facilitate CEX PD and improve process understanding from a molecular perspective.