Automation & Digitisation
Scaling Up

Adherent Cultures: High Efficiency Scale-Up from the Research Bench to Production Scale

4 November 2021
Corning’s Austin Mogen (AM) and Vasiliy Goral (VG) join Phacilitate Editor, Georgi Makin (GM), to share their expertise with adherent cultures. In this interview, they discuss the technicalities of adherent cell culture at large scale, strategies to ensure lot-to-lot consistency and approaches to improving cell harvest.

GM: Hello everyone and welcome to this Q&A with Phacilitate. Today I’m speaking with Austin Mogen, PhD, Senior Field Application Scientist and Vasiliy Goral, PhD, Development Associate both from Corning Life Sciences and today we will be discussing how a high efficiency adherent scale-up workflow can successfully take shape to bring you from the research bench to production scale.

So, let’s start with adherent cell culture, which is well established at small scale for R&D purposes, but how do you use it for scale-up from vials and flasks to commercial scales, and what strategies and technology platforms are required for successful scale-up and cell expansion?

VG: OK, I’ll take this question. So, for decades, adherent platforms have been used for many anchorage-dependent cell types in research and development labs. Also, such platforms are reliable and well defined, they are very labour-intensive and limited in the scale-up production potential by available growth surface area. As a result, the current platforms that are on the market for the large-scale production of cells can be divided into three big groups; the first is a based on a planar approach and it’s essentially a direct scale-up from benchtop to manufacturing facility. These vessels have the same type of 2D surface as a benchtop flask.

First, they were introduced and implemented as multi-layered vessels and later Corning introduced a HYPER technology that can provide the same type of 2D surface but can deliver a much larger surface area to volume ratio. For a typical production load, to harvest enough cells about 100 vessels at 120 layers each will be required, and this will bring the total surface area to about 480 square meters. Running this number of vessels in parallel is actually a big manufacturing challenge, and it will require automation, which is very capital intensive.

So, to alleviate this issue, the second group is a packed bed approach, and it consists of providing packed beds with a very high surface area to volume ratio in a small footprint. However, materials used in these packed beds are not exactly the same as the ones that are used in 2D platforms, therefore special attention has to be paid to make sure that the cells adhere to the substrate and proliferate and function accordingly. For the successful run of a packed bed, it has to be seeded uniformly with cells, it has to be able to provide a uniform axial and the radial flow distribution and nutrient supply, as well as a metabolite removal. As the packed bed is run at such a high cell density per volume, automatic systems have to be in place to control the process parameters, such as pH, oxygen, temperature and nutrients. The biggest challenge of this packed-bed technology is to recover the cells, to harvest the cells at the end of the process and that comes from the fact that the cells are actually entrapped in the packed bed contrary to the attached cells on 2D planar surfaces.

So, this caused the development of yet another technology and approach for large-scale production, which is suspension cell culture. In this cell culture, cells are attached to the microcarriers that provided a 2D surface for cells to grow and microcarriers are suspended in a stirred tank bioreactor. The problem in this case is that very often during the harvesting step, the residual particulates that come from the microcarriers is very difficult to separate from the cells. In addition, cells experience excessive shear stress during the culture because the microcarriers have to be suspended by the impellers in the bioreactors. To summarise the problem points of large-scale, adherent cell manufacturing, we can imagine the ideal system that should deliver uniform cell seeding. It should enable uniform cell perfusion to remove metabolites and deliver nutrients. It should be able to automatically control the pH, oxygen, temperature, and most importantly, this system should enable the end user to recover the cells at the end of the process. To address these key points, Corning developed the Ascent™ FBR System that is able to grow cells in large quantities, and the key advantage of such platforms is that it actually provides a 2D surface for cells to grow in a very efficient manner, in full, with high surface area to volume ratio. It ensures the uniform cell distribution during the seeding and uniform perfusion of the packed bed during the culture. Most importantly, it enables the cell harvest from the packed bed at the end of the process and because of this, it’s very amenable to use in process development through the scale-up into the manufacturing. It can be used to as a seed train platform to enable the smooth transition of the process that is developed in a laboratory into the manufacturing settings.

Anything to add Austin?

AM: Yeah, I think you covered all the major platforms there; the stacked vessels, microcarriers and the packed bed reactors. The only thing I’ll add is just there are pros and cons of each of those platforms and it really depends on what your goals are in terms of scale, in terms of process control; and what stage are you at. Are you in clinical trials or are you moving more towards commercial scale? So, choosing one of those platforms really depends on a variety of different factors as well as the application. There are some cell types that perform better in 2D planar surfaces or a microcarrier that’s pre coded with a biological coding, so it really depends on what your what your application is and what your overall goals are.

GM: Thank you. I think that’s quite comprehensive answer there, and if we think specifically about a biologics manufacturing process, consistency between lots is critical for ensuring therapeutic performance but this can obviously be particularly challenging for larger scale production, right? So, could you talk to some of the automation and process control around the adherent platforms that you’ve just described?

AM: So, you’re absolutely right, one of the biggest challenges of bioprocessing is lot-to-lot consistency. Biology is complicated, therefore built-in process controls and automation can certainly reduce the variability of your process, especially at scale. So, the larger your scale gets, the more potential for introducing that variability into your bioprocess. When it comes to automation and process controls, there are different options for those different platforms that Vasiliy just mentioned. Let’s talk first about the 2D vessels – so things like CellSTACK® culture chambers and HYPERStack® culture vessels – those vessels historically have introduced inherent variability because they require many manual handling steps, so you have to position them in a certain way when you’re adding your cells, when you’re harvesting. There are now automated manipulator platforms that are available to ensure the consistency of those movements and timing, and they can be integrated into your process to reduce some of that variability associated with handling those vessels.

When it comes to process control around microcarrier processes, microcarriers are typically paired with a bioreactor, such as a stir tank reactor, and those bioreactors have built-in process controls, including fluid management control over the gases, such as dissolved oxygen, temperature and pH. Talking about the Ascent™ FBR System that Vasiliy mentioned earlier, and other fixed bed reactors, these types of reactors typically have the same process controls; fluid management, mixing parameters, oxygen temperature, pH. In addition to that, some of them have built-in tubing management as well, such as pinch valves and pumps that, paired with the controller system, automatically move fluid between different parts of the reactor to really automate both the process control within your cell culture – so controlling the process parameters – but also controlling the flow of fluids and liquids. So, the Ascent™ FBR System that Vasiliy mentioned earlier really has been focused on both automation and process control and it has automated programs already built in for many of the cell culture steps such as feeding, media exchanges, harvest and so I see really this is where the field is moving towards; automation to really decrease that process variability but also built-in monitoring of metabolites, gases, pH to ensure that you have a robust process and you can monitor that process throughout. Anything to add there Vasiliy?

VG: Uhm, yes. The only thing that I would like to add is with the automation part, the electronics part it’s historically been pretty reliable and easier to monitor and to control and that is helpful and improves the lot-to-lot reproducibility.

GM: Excellent, thank you. While we’re still talking about the idea of large-scale production, I’d like to discuss the opportunity for contamination. With cultures at such large-scale, a contamination event would be particularly impactful, so what measures can users employ to protect their process?

AM: Yeah, so at research scale, scientists often include antibiotics in their cell culture to decrease the risk of contamination. This really isn’t acceptable, usually by regulatory agencies, when producing therapies at large-scale. Therefore, many bioproduction processes turned to single-use components that come sterile out of the box and don’t require any autoclaving and that are ready to use with that consistency of sterility, reducing the risk of introducing contamination into the system.

In addition to that, a process can be designed as a fully closed system, and you can do this by implementing things such as tubing assemblies, bags and bottles for transferring liquid between different vessels. Then there are some considerations around making those connections between different vessels. So, probably the classic method of doing that in a sterile way is sterile tube welding, which utilises heat to essentially connect weldable tubing and then there’s also a variety of different sterile connectors that can be integrated onto those tubing assemblies to allow that sterile transfer of liquid between different process steps.

Ultimately, it’s important to design a process holistically and it can be really beneficial to audit your process and complete a risk assessment for potential contamination, i.e., looking at different weak points of where within your process there’s a possibility of introducing contamination. Really, the Holy Grail is to design something that utilises minimal open process steps. So, minimal steps with any biosafety cabinet where all the fluid flow and cell culture is being completed within a closed system assembly that has already been sterilised and is ready to use out of the box as a single use closed system.

GM: Thank you. So, whether you’re collecting cells as a product for therapy, or as an intermediate in a seed train, one of the challenges of adherent cell culture is achieving harvest efficiency. So, what are the main factors that reduced harvest efficiency and what are some of the technological solutions that can improve adherent cell harvest?

VG: Yes, harvest efficiency is an important parameter that one should consider when choosing the production platform and the most efficient harvest efficiency is demonstrated by planar platforms, such as Corning HYPERStack vessels, for example, where cells are cultured on a 2D flat surface and, in this case, cells are detached from the surface by exposing cells to a dissociation solution and then collecting the suspensions of the cells from the system. Contrary, in packed bed reactors, very often cells are physically entrapped in a packed bed and because of this it is extremely difficult to recover the cells at the end of the process, and this fact significantly reduces the cell productivity.

Another important factor as to why cell harvesting efficiency matters is a seed train. In current packed bed reactor systems, seed train is performed in planar 2D vessels. Cells are harvested from such vessels and then introduced into a final production-scale packed bed reactor. In this case, there is a drastic change in the adherent surface properties in the cell’s environment, which is configured from a 2D – where cells were exposed in the seed train – to the 3D, where cells are exposed in the packed bed reactor. This may impact cell growth and the final product yields.

Corning addressed these problems in the Ascent™ FBR system. The key component of the Ascent™ system is the packed bed reactor that can provide a 2D surface for cells to attach, which is very similar to the planar system. Because of this, cells can be easily recovered from the reactor, increasing the productivity of the system and enabling us to use this platform as a seed train to start the large, manufacturing-scale reactors, so that the cells can be grown in a smaller process development unit (which is one meter square for example) and then harvested and transferred to the pilot unit, around 50 sqm and finally transferred into the large manufacturing 500 sqm reactor. This simplifies the overall transfer of the developed process from the laboratory-scale into the manufacturing settings.

AM: So I’ll just add that each platform has its own considerations around cell harvest, and sometimes cells behave differently in different platforms, both in terms of the amount of extracellular matrix that they secrete and how tightly they bind to the cell culture surface. So, for both the platform, and the cell type being cultured in that platform, there may be different aspects of the harvest process that needs to be optimised. Such as the protease being used, the contact time of that protease, introducing some kind of mechanical agitation into the system, whether that’s manually handling the vessel, such as a CellSTACK or HYPERStack or introducing flow or pressure like you may do in a fixed bed reactor.

Lastly, for microcarriers, there are some that are dissolvable, Corning offers a dissolvable microcarrier that essentially allows you to dissolve the cell culture surface – the bead – and then harvest your cells from that from that cell suspension. Like I said, each platform has its own considerations and sometimes the parameters associated with the harvest need to be optimised specifically for the cell type in culture.

GM: Excellent, thank you. So, for our final question I wanted to ask, for those who are new to these types of technologies, what are the key considerations for the development and implementation of a new process?

AM: Yeah, so there are multiple considerations when you’re developing a new process, particularly when you’re thinking about scale. So, the first really is scale; what is the scale of your process going to be based on your desired cell number and therapeutic dose? Once you’ve figured that out, other considerations are associated with the application or could be cell-type specific. For example, do your cells require a biologically relevant surface coating? This is a particularly important consideration for stem cells or cultured meat.

Applications…another consideration is around time to market. There are some tried and true platforms, such as the stacked vessels, like CellSTACK and HYPERStacks, that typically allow you to get to market faster because the optimisation and modifications to your protocol are minimal compared to a smaller-scale process. But if you’re looking at something that needs to be at a much larger commercial-scale, then you’re talking about a bioreactor, but the downside of that is it often takes longer to develop that process.

There are, of course, equipment and space considerations. So, how does the platform that you choose fit within your facility? We already talked about designing a closed system process but it’s important to understand how the closed system components fit into your overall process to move liquid between different parts of your process, integration of online and inline monitoring, as well as automation to decrease that process variability like we talked about. Cost, of course, is a consideration, so you need to determine what you’re willing to spend in terms of investment in capital equipment versus what you would spend on consumables.

Then, last but certainly not least, it’s really important to define your critical quality attributes up front for the bioprocess. That’s going to be different depending on what your process is. If it’s a stem cell-based therapy, is that going to be the differentiation potential after you’ve harvested your cells? Did they express the same markers? Do they have the same functionality? If it is a viral production process, do you have the target number of infectious particles? Do they contain the gene of interest? So, those critical to quality attributes very much depend on the specific product that you’re producing, but they’re absolutely critical to define up front and then have assays and tests in place to be able to test that throughout your process, because as you change things and scale up, you need to ensure that’s not impacting those quality attributes. 

GM: Vasiliy, is there anything else you’d like to add at this point? Any final comments?

VG: Austin covered pretty much everything. The only thing I would like to add is that when choosing the platform for the manufacturing production one would really have to pay close attention to the scalability because for any process that you develop on a small-scale, you want to choose the right platform to deliver the manufacturing scale so that you don’t have to reoptimise the process. Any changes done on a large-scale are extremely expensive. Processes are finalised on a small-scale and then it must be smoothly transferred to a large-scale, and very often people will mistakenly optimise everything in a 2D format on a flat surface and then suddenly decide to produce the products in packed bed. Transferring the cells from a 2D system to the packed bed will impact a lot of process parameters. So, you want to start your process early in a packed bed and choose the systems that can deliver packed bed on a small-scale, one meter square, to large-scale, 500 meters square.

AM: Very good point Vasiliy.

GM: Excellent, well thank you both for joining me today!

This feature was developed in partnership with Corning.