Mechanistic Modeling Based PAT Implementation for Ion‐Exchange Process Chromatography of Charge Variants of Monoclonal Antibody Products

Abstract

Process chromatography is typically used to remove product related impurities and variants that have very similar physicochemical properties to the product. Baseline separation may not be achieved in most cases due to high protein loading and thus, pooling of the elution peak can be challenging for maximizing yield and achieving consistency in product quality. Batch-to-batch variability in quality of the feed material also occurs in commercial manufacturing. Mechanistic modeling of process chromatography, though non-trivial, can be an enabler for implementation of Process Analytical Technology. This paper presents one such application involving prediction of the impact of variability in feed quality and in gradient shape on separation of charge variants by cation exchange process chromatography and thereby facilitating feed forward control. Five batches having different compositions of charge variants have been used to demonstrate the proposed pooling strategy based on simulated chromatograms and the outcome has been compared to offline pooling based on fractionation. For all the conditions examined and for the desired target of main product (67%), the proposed approach resulted in remarkable consistency in product quality (67 ± 2%) while delivering a yield of greater than 90%.

Control of surge tanks for continuous manufacturing of monoclonal antibodies

“…This is because most PAT applications in recent literature require flexibility in unit operation flowrates, and can only be implemented if surge tanks are present to decouple the process flow streams. Examples of such PAT approaches include variable loading in Protein A chromatography to handle upstream titer variations (Rathore et al, 2019; Rüdt et al, 2016; Thakur, Hebbi, & Rathore, 2020), variable pooling in polishing chromatography in case of changes to the product charge variant profile (Kumar & Rathore, 2017; Tiwari et al, 2018), or variable flowrate and pressure in ultrafiltration to handle membrane fouling and concentration deviations (Thakur, Thori, et al, 2020; van Reis et al, 1997).…”

Advanced control strategies for bioprocess chromatography: Challenges and opportunities for intensified processes and next generation products

“…For well-predicted systems, Kumar and Rathore [90] demonstrated that mechanistic model simulations conducted prior to running the separation can be used to dictate fractionation. This feedforward control strategy was dependent on the availability of feed composition data, which in this case was obtained using UPLC.…”

“…In an industrial setting however, feed data may be readily accessible from the upstream operation. A more computationally efficient fractionation method using mechanistic model simulations of the product profile only and an in-line UV signal was also demonstrated [90]. The difference between the overall UV signal and the mechanistic model prediction of the product profile was used as a measure of the impurity content.…”

Model‐based monitoring of industrial reversed phase chromatography to predict insulin variants

“…In addition to the capability of mechanistic models in downstream process development, they have also been used for monitoring of biopharmaceuticals and their impurities. For example, monoclonal antibodies and their variants have been modeled in size exclusion chromatography, cation exchange chromatography, and hydrophobic interaction chromatography . The agreement of these mechanistic models with experimental data demonstrates their potential as monitoring tools.…”

“…For example, monoclonal antibodies and their variants have been modeled in size exclusion chromatography, 20 cation exchange chromatography, [21][22][23][24][25][26] and hydrophobic interaction chromatography. 27,28 The agreement of these mechanistic models with experimental data demonstrates their potential as monitoring tools.…”