Process Modeling Tools Find Use in Bio/Pharma R&D and Manufacturing

Jennifer Markarian

Jennifer Markarian is manufacturing editor of Pharmaceutical Technology.

Equipment and Processing Report

Equipment and Processing Report, Equipment and Processing Report-10-16-2019, Volume 12, Issue 10

Systems-based pharmaceutics case studies, challenges, and new uses were discussed at PSE’s Advanced Process Modeling Forum.

Systems-based Pharmaceutics (SbP) uses a systems engineering approach and calibrated, mechanistic models to design optimally formulated pharmaceutical or biopharmaceutical products. These models-sometimes called digital twins-are crucial to digitalization, both for product and process design and for manufacturing. 

Pre-competitive collaboration

The Systems-based Pharmaceutics Alliance (SbPA), a pre-competitive association, was founded in 2013 to develop integrated digital design techniques to improve efficiency by simultaneously designing the product and the process. The group consists of solution provider Process Systems Enterprise (PSE) and five major pharmaceutical companies. The SbPA developed modeling tools based on mechanistic models and verified by experiments, and the tools, such as PSE’s gPROMS Formulated Products modeling platform, are now being applied. As presenters noted, SbPA shows how a precompetitive alliance can achieve more than can be accomplished by an individual company. The group is actively collaborating and working to advance the technology. “Our focus now is establishing workflows, finding gaps that still need solutions, and capturing outcomes,” says Dana Barrasso, PSE’s technical lead for the SbPA. 

In digital design, used for R&D and for technology transfer, models are used to run “what-if” scenarios. Models are also used for a statistical evaluation technique called global systems analysis (GSA). “GSA allows developers to explore the whole design space and see how disturbances of the system affect critical quality attributes as a probability distribution,” explained Barrasso (1). 

Although the focus of models to date has been primarily for digital design, over the next few years, models will be used for digital operations, predicted Barrasso (1). Models will be used, for example, to replace the full-plant experiments that have been needed to create model-predictive control and to test control systems, as well as to train operators.

New ways of working 

One of the SbPA’s goals is to transform the development process to modeling and simulation said Bob Yule, GSK, in a presentation at PSE’s modeling forum (2). The benefits of simulation are reduced experimentation (along with reduced labor and materials) and greater exploration of the process space, he explained.  In turn, this will allow more robustness in manufacturing, enhance finding simpler processes, and eventually permit the use of real-time monitoring and maybe even real-time control. One of the challenges in implementing modeling and simulation is the infrastructure needed. Although the platform (the model) is available, both subject matter experts in modeling and acceptance of this new way of doing development are needed, said Yule. A traditional design of experiments (DoE) model involves a scoping DoE, a factor-screening DoE, and a statistical model, then identification of critical process parameters (CPPs) and ranges, followed by verification in the lab. The new way of process development, however, is to create a flow model and then perform an in silico DoE to identify CPPs. This method cuts process design time in half, said Yule, but to make it the new way of working, industry inertia must be overcome. It also takes time to build a “critical mass” of people who are capable of modeling. Yule said that at GSK, the most progress in modeling has been made in drug substance and projects are underway in drug product. The company is currently evaluating biopharma modeling in its partnership with Innovate UK. 

Modeling biopharma processes

Although SbP is now well established in small-molecule development and production, with several off-the-shelf libraries that can be used in modeling tools, modeling of biopharmaceutical processing is an emerging area. PSE received seed funding in 2018 from Innovate UK to explore how small-molecule SbP could be translated to large-molecule SbP. Initial work showed that existing methods and models are directly applicable to develop bioprocess modeling, and PSE is preparing to release new modeling software, gPROMS FormulatedProducts for Biologics. 

Several posters at the conference highlighted academic and industry work in modeling cell culture. For example, researchers at the University of Massachusetts Lowell worked on a model in collaboration with PSE and the NSF Advanced Mammalian Biomanufacturing Innovation Center (AMBIC), which includes academic members from UMass Lowell, Johns Hopkins University, Clemson University, the University of Maryland, and the University of Delaware. The project developed a model for viral growth and created a flowsheet model for continuous production of viral vaccines that can be used to optimize the continuous process (3). Another project at UMass Lowell used modeling to optimize the media components that impact cell growth (4). 

Researchers at Rutgers, the State University of New Jersey, presented FDA-funded research on using process analytical technology (PAT) to measure cell growth in continuous upstream monoclonal antibody production using Chinese Hamster Ovary mammalian cells (5). The group is also developing a mechanistic model of the bioreactor, which it will validate with the experimental PAT data. The researchers plan to connect the model of the bioreactor with downstream process models using flowsheet models, such as PSE’s gPROMS.

References

1. D. Barrasso, “Applications of digital twins in continuous drug product manufacturing,” Presentation at Advanced Process Modeling Forum USA (Tarrytown, NY, Sept. 10, 2019).

2. B. Yule, “The Challenges of Implementing System Modeling in Pharmaceutical Process Development,” Presentation at Advanced Process Modeling Forum USA (Tarrytown, NY, Sept. 10, 2019).

3. C. Morris, et al., “In-silico Platform Based on Flowsheet Modeling for Continous Viral Vaccine Production,” Poster at Advanced Process Modeling Forum USA (Tarrytown, NY, Sept. 10, 2019).

4. Galbraith, H. Liu, S. Yoon, “Media Formulation Optimization Based on Multi-Scale Modeling of Heterogeneity in Mammalian Cell Culture Process,” Poster at Advanced Process Modeling Forum USA (Tarrytown, NY, Sept. 10, 2019).

5. O. Yang et al., “Continuous Upstream Biopharmaceuticals Manufacturing,” Poster at Advanced Process Modeling Forum USA (Tarrytown, NY, Sept. 10, 2019).