Automation enables intensification of downstream processes.
Using data for predictive maintenance
Software applications that connect to more data, along with improved data formatting and insights, will grow in use over the next five years in biopharmaceutical manufacturing, suggests Petter Mörée, managing director for Europe, Middle East, and Africa at Seeq. “These applications significantly improve an organization’s ability to share data, information, and insights across the entire organization in real-time,” he says.
In downstream processing, one use of such data is for predictive maintenance, in which equipment diagnostic and condition data can be monitored in real-time to determine when equipment is trending out of specification and maintenance is needed. These tools can be used, for example, in chromatography, media optimization, and centrifuge equipment, Mörée explains.
Another example is in monitoring the effectiveness of ultrafiltration membranes. In one case, flow rate and pressure data were used to calculate the change of membrane resistance over time, and a model was created to predict membrane failure. Warning and alert limits were then set, so that filters could be changed prior to membrane failure. This predictive maintenance model enabled more consistent batch quality and increased yield, reported Seeq (1).
Biopharmaceutical manufacturers are looking for ways to run downstream processes more efficiently, so that higher volumes can be made in a shorter time without adding more or larger processing equipment. Intensified and continuous methods are needed to keep up with the increased volume from upstream processing, as well as improve cost, time to process, and the area of floorspace needed to operate, says Martin Lobedann, Process Technology consultant at Sartorius. Intensified processes rely on process analytical technology (PAT) to collect real-time data used for automated process control. “Higher automation per unit operation and orchestration across multiple unit operations enable continuous and closed processing,” says Lobedann.
“In processes that are more intense, where variables are changing at a faster rate, the margin for error is decreased,” says Alexei Voloshin, global application development manager for 3M Separation and Purification Sciences Division. “A machine can analyze and adjust much faster than a person can. The ability to control systems, which may be inherently unstable, is better left to automation technologies.”
Automation can enhance the operation of downstream processes, such as purification and filtration.
Intensifying purification
Downstream processing of monoclonal antibodies (mAbs) traditionally uses chromatography capture followed by viral inactivation and two other chromatography polishing steps, with a batch going through each unit operation in sequence. This method of processing one after the other is “costly, non-optimized production,” says Jérôme Chevalier, manager of Product Management, Chromatography Systems at Sartorius. One way to optimize and speed up production is to use several chromatography columns in parallel. A second strategy, says Chevalier, is to use automation to interconnect all the downstream unit operations. This continuous processing method avoids manual intervention and dead time, he explains.
Chevalier says that Sartorius implemented automation strategies to intensify downstream processing at Phase II/III clinical-scale for a major biopharmaceutical manufacturer. “The BioSC hardware concept linked with an automation platform allowed us to reduce processing time from six days to one day; reduce the footprint by approximately 15-fold; and save up to 42% of the cost of goods,” he reports.
A key to user acceptance is ease of programming and flexibility of the recipe editor, Chevalier adds. A possible future step is to integrate other unit operations, such as filtration, concentration, and diafiltration into the continuous, automated workflow.
Intensification of the capture step would have a significant impact on process economics, suggests Lobedann. “Application of current affinity-based resins in the market make this step the most expensive due to a lower productivity (<20 g/L/h) caused by a combination of low binding capacities and linear flow velocities due to diffusion limitations,” he explains.
Automation systems can be further optimized by incorporating process data, collected using PAT sensors and analyzed with a data analytics platform. These technologies can result in better process control and cost reductions, says Artur Arsenio, head of product management for PAT and Automation at Sartorius. For example, a biosimilars manufacturer is using PAT, data analytics tools, and bioprocessing software to enable a platform with continuous upstream and continuous downstream processing. The upstream process uses Sartorius’ intensified seedtrain, and the downstream process uses Sartorius’ multi-column chromatography combined with flow-through filtration.
Automated filtration method
Researchers in the lab of Professor Anurag S. Rathore at the Indian Institute of Technology Delhi have developed a method for automating a dead-end filtration skid for continuous depth-filtration, used for clarification of microbial and mammalian cell-based biopharmaceuticals (1). One of the challenges for continuous filtration is that the feedstream characteristics, such as turbidity or host cell protein content, can vary over the weeks or months of the continuous process, which affects the size of the filters needed, they explain. The researchers proposed a filtration skid with multiple, small-sized filters, with pressure and turbidity sensors that enable the process to automatically switch to a new filter when the sensors indicate a limit has been reached.
Advertisement
Automating filter switching also saves cost because smaller-sized filters can be used. In batch processing, filters are typically sized up by 1.5 to 2 times the calculated size. With an automated system, the filters do not have to be over-sized because the filters will be switched as needed. This method should result in more consistent filtrate quality (1).
The skid has been successfully tested in the lab and is being considered by major manufacturers, says Rathore, who is coordinator of the Center of Excellence for Biopharmaceutical Technology and professor in the department of Chemical Engineering at the Indian Institute of Technology Delhi. He notes that the skid is available to manufacturers as open-source technology and is part of the continuous process technology being developed at the Center.
Intensifying buffer prep
In addition to capture and filtration, polishing and buffer exchange steps can benefit from intensification. “Buffer management—preparation, distribution, and hold—is often overlooked,” says Lobedann. “Intensifying here would greatly reduce the auxiliary footprint, because one would need smaller storage tanks with concentrated buffers, which can then be diluted at point of use.”
In-line dilution of concentrated buffer solutions or in-line blending and conditioning of component stock solutions has been demonstrated to reduce footprint and cost, and automated systems for “on demand” buffers are commercially available; there is some hesitancy by processors, however, to implement this new concept (2). There is room for new technology and methods that could ease implementation.
An open-source system being developed by two precompetitive consortiums, the National Institute for Innovation in Manufacturing Biopharmaceuticals (NIIMBL) and BioPhorum, is the NIIMBL-BioPhorum Buffer Stock Blending (BSB) System, which was described in a December 2019 report (3). The prototype was displayed at INTERPHEX in October 2021, and more details on the concept are planned for publication in 2022. The consortiums’ vision for the open-source system is to accelerate adoption; allow companies to evaluate the technology and dataset; and allow companies to leverage the approach to deliver buffer preparation with a reduced footprint, in less time, with reduced labor (4).
Making connections
Interoperability between different systems in a manufacturing facility is crucial. One challenge in downstream processing is that different pieces of equipment often come from different vendors; each skid from a different supplier may use a different control system platform, resulting in a higher level of complexity required to integrate the skids, explains Yvonne Duckworth, director of Digital Technology and senior automation engineer at CRB. “Making data available via communication protocols can be complex with multiple control system platforms, but [it] is crucial to provide the level of data connectivity that is key to smart manufacturing and digitalization in accordance with Pharma 4.0,” she explains. “A vendor-neutral protocol that has been gaining popularity is the OPC Foundation’s OPC Unified Architecture (UA). OPC UA is an easy way for third-party platforms to communicate, providing enhanced interoperability, while adding more robust security for access control, authentication, and encryption.”
Integration and standardization are important for control systems to handle the growing need for flexible production systems used to make multiple products. “This shift towards a more flexible but also more integrated approach has generated a desire for technologies that make it fast and easy to transfer recipes and integrate new equipment into the control system,” says Michalle Adkins, director of Life Sciences Consulting at Emerson. “If you need to switch to a new product quickly, you can’t get caught up spending weeks or months updating recipes, writing custom code to integrate essential equipment, and testing these changes. To meet this need, the most flexible control systems are incorporating integration of skids with both tighter native controller integration and standardized integration of the equipment controls and graphics using NAMUR’s [User Association of Automation Technology in Process Industries] Module Type Package (MTP). MTP provides a framework for standardized connectivity between equipment to streamline interoperability, making equipment nearly plug and play.”
Standardizing interfaces for technology transfer
Technology transfer is another area that benefits from standardized, automated tools. “To break down the silos that create tech transfer delays and inhibit speed to market, manufacturers are leveraging process knowledge management (PKM) software to build a bridge between development and manufacturing,” says Adkins. “PKM solutions are becoming more important for process knowledge capture and transfer.” She explains that these tools can be used both for developing new products and for managing changeover from one product to another. “Parameter-driven recipe management and the ability to transfer recipe parameters from a higher-level system to the production execution systems has become a focus,” Adkins explains. “By standardizing data, interfaces, and usability across the entire production pipeline, PKM tools make it easier to evaluate sites against a master recipe, transfer master recipes with process details, and push parameters to downstream systems to determine which sites are the best fit and [to] transfer the details for how to manufacture specific products. Emerson has found that under the right circumstances, a more automated technology transfer process can help reduce time to market from 10 years to less than three years.”
Lights-out, continuous manufacturing in the future
Automation technologies that minimize human involvement inside upstream, downstream, and fill/finish manufacturing suites provide the dual benefits of reducing contamination risk and helping address personnel shortages. “The ultimate long-term vision [for many life sciences organizations] is to enable lights-out manufacturing, creating continuous production and packaging lines that operate with minimal human intervention,” says Adkins. “Automation technologies [such as] simulation and predictive modeling, remote monitoring, and even artificial intelligence and machine learning will be key to the success of a future with lights-out manufacturing initiatives.”
References
1. G. Thakur et al., Front. Bioeng. Biotechnol. online, DOI:10.3389/fbioe.2020.00758 (July 3, 2020).
2. F. Mirasol, BioPharm Intl., 33 (12) 36-37,50 (2020).
3. BioPhorum Operations Group, “NIIMBL-BioPhorum Buffer Stock Blending Systems: A More Advanced Concept for Buffer Manufacturing,” December 2019.
