(60t) A Model-Based Framework for Integrated Design and Operation: An Application to Chromatographic Processes

The development of biopharmaceutical processes traditionally depends on heuristics, expert knowledge, and wet-lab experimentation. While this approach can result in processes with high product quality and titers, it often entails lengthy and costly experimentation [1]. Furthermore, such an approach may initially overlook the need to account for process flexibility, resulting in rigid processes that can lead to off-spec batches [2]. This is particularly problematic when screening resins for chromatography separations as the traditional approach would require a large number of experiments to be carried out. To this end, and in line with the Quality by Design initiative [3], we propose a novel model-based design space framework that enables quantification of process flexibility, facilitating a rigorous assessment of resins.

We present a model-based framework for the integrated assessment of process flexibility and performance. The methodology relies on an experimentally validated model to act as a virtual experimentation platform. First, the problem is formulated by defining the design inputs (DIs), key performance indicators (KPIs) of interest, and the constraints of the process. Next, quasi-random Sobol sequence is utilized to generate design input combinations for virtual experiments, facilitating efficient low-discrepancy exploration of the process conditions. Once the simulated data set is obtained, artificial neural networks (ANNs) are trained and used as interpolators for enhancing the resolution of the data set. This enables the identification of smooth design space boundaries without violations using alpha shape. Finally, based on the identified design space, the quantification of process flexibility in terms of acceptable ranges can be carried out, in addition to investigating KPI performance.

We implement this framework for resin screening used in the design of a batch protein A chromatography process in monoclonal antibody manufacturing. We use the mathematical model developed by Grom, et al. [4] that has been experimentally validated for five commercial resins: MabSelect SuRe, MabSelect SuRe LX, CaptivA PriMAB, Eshmuno A, and POROS MabCapture A. The design inputs of interest are the feed composition, feed flow rate, and column loading time, while we monitor three KPIs: yield, productivity, and resin utilization. Design spaces for all five resins are mathematically quantified and compared. Using the framework, we are able to quantify the non-linear effects of changing performance constraints on the identified design space. We also investigated the largest possible acceptable operating region when operating with the different resins. Based on the investigation, the Eshmuno A resin provided the largest flexibility with it having the largest design space. On the other hand, the MabSelect SuRe LX resin were able to achieve higher productivity. The presented framework demonstrates how process development can be accelerated through the use of computer modelling tools, leading to informed decision making and targeted experimentation.

Acknowledgements:

Funding from the UK Engineering & Physical Sciences Research Council (EPSRC) for the i-PREDICT: Integrated adaPtive pRocEss DesIgn and ConTrol (Grant reference: EP/W035006/1) is gratefully acknowledged.

References

[1] H. Narayanan et al., "Bioprocessing in the Digital Age: The Role of Process Models," Biotechnol J, vol. 15, no. 1, p. e1900172, Jan 2020, doi: 10.1002/biot.201900172.

[2] M. M. Nasr et al., "Regulatory Perspectives on Continuous Pharmaceutical Manufacturing: Moving From Theory to Practice: September 26-27, 2016, International Symposium on the Continuous Manufacturing of Pharmaceuticals," J Pharm Sci, vol. 106, no. 11, pp. 3199-3206, Nov 2017, doi: 10.1016/j.xphs.2017.06.015.

[3] A. S. Rathore and H. Winkle, "Quality by design for biopharmaceuticals," Nat Biotechnol, vol. 27, no. 1, pp. 26-34, Jan 2009, doi: 10.1038/nbt0109-26.

[4] M. Grom, M. Kozorog, S. Caserman, A. Pohar, and B. Likozar, "Protein A affinity chromatography of Chinese hamster ovary (CHO) cell culture broths containing biopharmaceutical monoclonal antibody (mAb): Experiments and mechanistic transport, binding and equilibrium modeling," J Chromatogr B Analyt Technol Biomed Life Sci, vol. 1083, pp. 44-56, Apr 15 2018, doi: 10.1016/j.jchromb.2018.02.032.