Advancing PAT Solutions for the Bioprocessing Industry

The interest in exploring new tools for process analytical technology (PAT) in biopharmaceutical drug development and manufacturing is strong throughout the industry. However, despite the concept being well accepted and supported for almost two decades, the adoption of PAT in manufacturing settings is slower than first anticipated. 

^{This article is posted on our Science Snippets Blog}

The purpose of PATs is to monitor and control a process based on measurable critical process parameters (CPPs) and the product's critical quality attributes (CQAs). The goal is not to have to test the quality of the final product; this should be built into the process through a quality-by-design (QBD) approach that facilitates real-time monitoring and release.  

Their implementation supports increased manufacturing efficiency, reduces waste and cost, speeds up the time to market, and enables intensified and continuous bioprocessing through integrated real-time monitoring and product release.

A complete process understanding where the critical sources of variation are known and controlled will enable the prediction of CQAs, increasing manufacturing efficiency and extending patient safety. In essence, achieving this would enable biological drug manufacturers to decouple the product from the process and advance their production capabilities towards Bioprocessing 4.0.  

To reach this goal, CPPs and CQAs must be identified early on and used to optimize the process and ensure it is well-controlled and repeatable, with minimal batch-to-batch variation.

PAT Sensors

There are different types of PAT operational modes used in biological manufacturing.

• In-line sensors are deployed in a process vessel or process stream and conduct the measurements in situ.  
• On-line sensors are connected to a diverted side stream for regular sampling and analysis, and the fluid is often passed back to the process stream afterward.  
• At-line and off-line analysis involve the collection of samples from the process, with the measurements carried out separately within the process or in a different facility.  

The focus has been to implement in-line sensors, when possible, as these allow for continuous process monitoring and control. However, in-line sensors are targeted towards one parameter such as pH, temperature, dissolved oxygen (DO), and carbon dioxide, or are based on spectroscopic technologies, e.g., ultraviolet-visible (UV-vis), near-infrared (NIR), Raman, and multiangle light scattering (MALS) to monitor single CQAs, such as titer or aggregate levels.  

At-line|on-line PAT sensors, in combination with on-line aseptic sampling devices, will enable the application of more sophisticated PAT technologies, for instance, ultra-high-performance liquid chromatography (UPLC) and its combination with mass spectrometry (MS). The on-line PAT application of UPLC and UPLC-MS significantly broadens the capability to monitor both CPPs and CQAs in a multi-attribute monitoring (MAM) and in-depth fashion.  

However, these technologies will have a delay between sampling and result feedback, impacting their use for real-time process control.

The combination of in-line and complementary on-line PAT sensors and analytics will most likely be needed to fully realize the potential of real-time manufacturing and product release of biologics.

PATs from Drug Development to Manufacturing

A key factor for implementing new PAT technology in bioprocesses is their adoption in early drug development stages. Applying high-performing detection methods for multi-attribute monitoring (such as UPLC-MS) in cell line development, clone selection, and process development will give more insights and an understanding of how to steer the process. It will also significantly reduce time spent in early drug development stages and thereby reduce time to market.

Sartorius and Waters have recently partnered to offer new PAT solutions for clone selection experiments, usually taking up to 10 weeks per batch. However, that timeline can be significantly reduced by using a solution based on Sartorius Ambr^® multi-parallel bioreactors and Waters' small-footprint BioAccord™ LC-MS System to obtain critical parameters for recombinant-protein production and cell culture media conditions within 48 hours.  

Bioprocess engineers can take control of their own cell culture data in near-real time using accessible and streamlined workflows designed for non-specialists to generate high-quality and high-confidence data with minimal training. The obtained data is best utilized in feedback control loops to optimize process conditions and product quality output.  

The experience and knowledge gained in early-stage clone selection and process development (PD) are strategically implemented in the later stages of scale-up and production. With the CPPs and CQAs identified and validated in the process, PATs will adjust and steer the process towards process monitoring within well-established boundaries for both the process and product.  

Ideally, the controlled and validated manufacturing process is easily monitored by in-line and on-line PAT tools to ensure that the CPPs and CQAs are transparent at any given time.

PAT in Upstream and Downstream Bioprocess

Most PAT tools and solutions available today are utilized for upstream bioprocessing applications. This upstream focus is justified, considering that initial product quality and overall process effectiveness are governed during cell culture or fermentation. During cell culture, feedback from in-line PAT sensors is crucial to control and monitor optimal growth conditions in the bioreactor. Many integrated sensors can be used, regardless of modality, as the different host cells require similar conditions or are affected by the same factors such as nutrients, pH, temperature, and DO.  

In upstream bioprocessing, time is also a factor that favors the implementation of more sophisticated on-line and at-line PAT solutions such as UPLC-MS. The comprehensive CQA information obtained by in-depth MS analysis is valuable to design optimal conditions for the subsequent downstream purification steps. There is also significant value in applying accessible MS-based PAT to downstream applications, although it is more challenging.  

A major constraining factor is the time required from sampling to result. Thus, MS-based PAT solutions in downstream manufacturing have probably the greatest use as process monitoring tools. However, downstream bioprocessing is still largely unit-driven, giving sufficient time for rapid on-line PAT analysis.  

The diversity in unit operations, CQAs, and drug modalities require flexible, reusable, multipurpose, or multi-attribute monitoring tools. These are driving factors for implementing on-line UPLC and UPLC-MS PAT monitoring in the manufacturing process, reducing the need for comprehensive MS analysis for final product characterization and release.   

Multivariate Data Analysis and Integrated Control

Data analytics solutions and integration play an essential role in implementing PAT in bioprocessing. Data management systems that support large-volume multivariate data collection and analysis are key factors.  

Multivariate data analysis (MVDA) software such as SIMCA^© is of great advantage for model building, process understanding, and control. Instruments, sensors, and analysis software must fit into the process control hierarchy framework - all integrated automated, and uniformly controlled.  

Collaboration between bioprocess technology providers and analytical solution suppliers, such as Sartorius and Waters, supports the implementation of sophisticated and high-performing PATs in the development and manufacturing of biopharmaceutical drugs.  

Industry-wide collaborative efforts are also driving the adoption of new PAT solutions for in-line monitoring and real-time release. Although the PAT concept has been around for some time, and we now start to see more integrated workflows and solutions, we still have a bioprocessing "ocean" to explore.