Thermo Fisher Scientific Inc. (TMO) Earnings Call Transcript & Summary

Hello, everyone. Welcome to BioProcess International Spring Digital Week, brought to you by the producers of the Face-to-Face BioProcess International Events, visiting Boston in September and Milan in October this year. BioProcess International Europe will be delivered as a 100% virtual conference and exhibition on July 13 through 17, 2020. My name is Jordan Stillman. I'll be your host for today's session titled Cell Culture Media Developed for Perfusion Can Drive Increased Performance in Perfusion. First, I'll cover some quick housekeeping items. [Operator Instructions] In 24 hours, you'll receive a link to watch the recording of this session. You can also download few featured white papers in the resource list box on the right side of your screen. Let's now begin by introducing our speaker from Thermo Fisher Scientific. Christopher Brau is a Staff Scientist in Cell Biology at Thermo Fisher Scientific. Thank you for joining us today, Christopher. Now I'll hand it over to you to begin the presentation.

Hello, and thank you to everyone attending our BPI Digital Week presentation. For those familiar with our previous presentation, there will be similar contents with some new data sets included. With that said, let's get to it. So the overview of our coverage today will be what is perfusion; why do we use it; the different types of perfusion; our perfusion medium design considerations, results and some best practices when pursuing, evaluating media for yourself. Only 2 things are required to turn a traditional suspension cell batch culture into a perfusion cell culture. A constant exchange of medium and the mechanism to retain cells, so they are not lost as spent medium is removed. The rate of exchange medium is usually expressed as the number of operating volumes of media exchanged per day or VVD. So if we were to maintain a 2-liter working volume and perfuse 2 liters of medium, this would be considered 1 VVD. A pretty common medium exchange rate is 2 to 3 VVD in industry. Process needs help determine the best cell retention method. We'll be focused on suspension cell culture for this presentation, so adherent cell perfusion via fixed bed or hollow fiber reactors won't be discussed. For suspension cell culture, there are generally 2 different approaches to choose from: settling based methods and filter-based methods. Filter methods like tangential flow filtration, or TFF, depicted here; or ATF, which stands for alternating tangential flow, employ a hollow fiber filter to cycle medium from the reactor. Cells are retained and brought back and spent medium is pulled through the hollow fiber membrane. These can be single-use friendly, relatively easy to scale, are 100% efficient at retaining cells in the reactor and have the potential to generate fully clarified spent medium from the reactor. This gives the potential to run spent medium directly into your downstream separations. On the flip side, filters tend to be expensive and require a little know-how in process development to ensure following won't be a problem. Methods that operate via settling, some with centrifuge or acoustics to enhance speed, can be trickier to scale and are not perfectly efficient, meaning some cells will be lost in the spent medium being removed. This also means you still need to clarify that medium before sending it further downstream. However, settling methods can save on consumable costs and generally have less following risk. The last thing I'd like to touch upon briefly are cell bleeds. This is the practice of directly removing medium from the reactor to remove cells and prevent excessively high cell density to maintain culture health in the reactor. Cell bleed and the product pulled off with it is often discarded as it may not be cost-effective to clarify the bleed and send it to downstream processing. This can result in decreases in net product yield and profit. For this reason, it's generally desirable to minimize cell bleed, though not all perfusion types require it. So why the interest in perfusion? What do we expect from it? Generally, we can look at that as 5 categories. So we start with the obvious, higher productivity. As I'm sure you've seen or read with perfusion, cell density and product titer can reach much higher levels than what is possible with conventional fed-batch with the ability to alleviate downstream bottlenecking. Next, we have facility flexibility. The drastic increase in titer produced allows equipment footprint in height to shrink, while directly reducing seed-train steps. Potential flexibility in medium exchange rate and in the case of continuous perfusion, operating duration can be adjusted to better accommodate specific process needs. Single-use solutions can be paired with this approach for easy validation and switching products to support a more diverse product portfolio while still maintaining equipment utilization in a smaller workspace. Additionally, a more compact process better facilitates tech transfer and the ability to set up mirrored workflows in separate locations with lower building requirements, allowing you to produce what is needed, when it's needed, easier. Process quality and risk. This is one of the main reasons perfusion was pursued decades ago. Perfusion allows for continuous product removal. This provides great benefit to products that are either toxic or unstable, getting them to a downstream stabilized state faster. This can be further enhanced with continuous perfusion where a steady state can be maintained, allowing for tightly controlled process conditions for better consistency. Additionally, process risk is reduced 4 ways. Perfusion and continuous processing, in general, allow for more impactful automation reducing human error. Fewer batches and fewer process steps directly reduce process risk. Smaller equipment footprint in size leads to a more easily observable process. And lastly, downstream lots can be logically defined. So in the event of an issue upstream, all downstream lot slices that were processed prior can be tested and released normally, giving you a faster product stabilization and reduced error chance. Speed to market and regulatory support are critical considerations in our modern industry. Quicker and easier facility builds paired with smaller scale up jumps, can lead to easier small-scale modeling and faster production targets realized, including being able to generate sufficient clinical testing materials earlier. Logical lot definition with better quality control and automation support makes for quicker and easier product release. And the FDA's emerging technology team and other regulatory bodies have voiced support and guidance for continuous processing. Barring that, hybrid downstream approaches can be utilized to facilitate a more traditional regulated batch release while still capturing much of the benefit of perfusion allowing a company to approach in stages. Put all the benefits together and you get consistent cost savings, lower CapEx and risk upfront and lower total cost of goods per gram of product produced. In upstream, the reduced process steps in physical batches result in reduced consumables. Continuous product removal allows for easier debottlenecking downstream with reduced holding steps and greatly improved resin efficiency. Easier process automation leads to fewer mistakes, reduced labor and easier digital documentation and augmentation of the process. In a recent webinar, Clone Biotech estimated a 40% cost reduction versus fed-batch. In our internal cost of goods analysis, we saw compelling results as well, making perfusion a really good choice for a lot of people. Perfusion processes can be oriented to fit different needs. For practicality, we'll briefly discuss the foremost common for suspension cell culture organized in relative complexity, starting with N-1 perfusion or intensified seed-train perfusion. The goal here is to achieve high sale densities while maintaining log growth. This is a short duration method usually 4 to 7 days that is used to reduce the number of vessels required in a seed-train, allow for higher density reactor seeding or for generating a high-density seed bank that can be directly thawed into a reactor. Medium exchange rates are selected to maintain the culture and log growth. Productivity is usually not a consideration and cell bleed is not used. So long as your media allows growth rate to proceed and your equipment can handle the increased demand, this is a relatively straightforward operation. To give an idea of how far this can be taken, Biogen recently gave a presentation using a 50-liter Thermo Scientific HyPerforma DynaDrive Single-Use Bioreactor pictured here with the G3 light controller. Using a combination of high cell density freeze, 10:1 turndown ratio or being able to start operation at 1/10th the rated working volume in the reactor and N-1 perfusion, they were able to replace the function of nearly the entire seed train and seed directly into their 15,000 liter production reactor. If they present this again at a later date, I highly recommend watching it. Seeing the picture of a 50-liter reactor dwarfed by all the equipment of their entire seed train suite is worth it alone. Another type of perfusion is concentrated fed-batch. In this approach, filter methods such as ATF or TFF must be used as the retention mechanism as both the cells and the product are being returned to the reactor throughout the run, thus allowing for a concentration of product in the vessel and a greatly increased final titer. This is ideal for stable products, but that have very low productivity in batch or fed-batch processes. The concentrated final titer allows for more efficient downstream batch processing and the potential to eliminate a concentrating step. However, as the product is not continuously removed, quality concerns will also be comparable to batch or fed-batch operation. These are typically 14- to 20-day runs and media exchange rates are targeted to generate final titer sufficient for downstream processing. With no bleed and orientation to be used with existing batch downstream processing, this is a relatively easy process to implement. The figure here shows the product titer in grams per liter during a 16-day run with a 42x increase over batch, more on that in the results section later. The next perfusion method is intensified fed-batch. This is similar in duration to concentrated fed-batch and also generates much higher total product. The difference in this case is that the product is continuously removed from the reactor throughout the run, making this a better option when working with less stable products and better producing cell clones as the constant product removal provides options for debottlenecking downstream. It compare well with a semicontinuous or a hybrid downstream as a result of this, though it needs to be planned around and a holding vessel or surge tank is usually considered here, adding some complexity. The process run time is generally a bit longer than a standard fed-batch run in the neighborhood of 16 to 25 days, especially if a bleed is used, and media exchange rates are often targeted to maximize efficiency and optimize quality. The last perfusion type I'll discuss today is continuous perfusion, which is what most people tend to think of when they think of perfusion in the upstream. The goal here is to develop a process that maintains a steady state where productivity and product quality can be sustained over a long term with minimal variability. These processes can extend from 30 to 90 days, and an active bleed is typically employed to maintain a targeted viable percentage or cell density. Similar to intensified fed-batch, medium exchange rates are optimized to maximize efficiency and product quality. This approach can be both the most technically demanding type of perfusion and the simplest. That sounds a little confusing, so let me explain. Manual operations of continuous perfusion is both difficult and complex, and the initial process optimization can be time-consuming. The long-term operation requirements also present an increased risk of operation errors. It also requires a cell line with superior production stability to be able to maintain over the long duration. On the other hand, it takes full advantage of automation and in line sensors to achieve superior control with reduced risk of errors relative to product generated, especially when contrasted against the number of seed train and fed-batch production runs that would be required to reach a similar net titer. Additionally, continuous perfusion is attractive as it can provide improved benefit to continuous and hybrid downstream processing, giving a consistent input to your downstream process. When considering generating a catalog perfusion medium, we identified 2 primary design goals, ease of use and cost-effective performance. Towards this end, the target was a chemically defined animal-origin free formulation that emitted proteins and growth factors. The AGT format was chosen to support fast and simple reconstitution and a 12-month shelf life was specified. The formulation needed to function both as a seed-train medium and a production medium and be flexibly capable of supporting high cell densities and productivity at a 1 VVD medium exchange rate, but also be able to run diluted at higher VVD if needed for quality purposes. Lastly, the medium would need to support a cost-effective scale up. We started the development of this formulation by creating a panel of 10 perfusion media. This panel was tested in-house via spin-tube model and by external alpha testers. This early work was important to stress the panel against diverse cell lines, operating conditions and perfusion applications. In the example shown, one of the alpha panel media demonstrated close to a 300% performance increase over formulations designed for fed-batch. In the other figure, an alpha tester was able to sustain high viable cell density as well as maintain cell-specific productivity, leading to a production of 3 grams per liter per day. After selecting a top medium panel candidate, a series of experiments were designed to further optimize the formulation. Instead of using spin tubes, we translated our scale down model to a perfusion like method using the Ambr 15. Each subsequent design of experiment further advanced the medium candidate. In addition to this, OMX works, such as proteomics and metabolomics, was used to supplement component selection and concentrations during the final design of experiment phases. The center graph shows some of the experiments done in the Ambr 15, where data shows a 160% produced improvement over the controls. The graph on the right is an example of a principal component analysis, which provides confidence in the metabolite selected to drive a statistical significance and should be focused on in the formulation. At this point, the formulation was ready for internal beta tests, for initial beta testing and was challenged against an array of cell lines, operating conditions and screening methods. So we'll go over some different results both from beta and in-house, starting with N-1. So this is one of the more unique cases. A beta tester desire to boost their seed train using a mid-process switch to our perfusion medium at 1 VVD. This experiment was carried out on 2 of their different clones that were frozen, thawed and tested in one of our chemically defined fed-batch formulations as a control. The N-1 perfusion seed-train was set up and perfused with the perfusion medium forcing a mid-process direct adaptation, which we generally do not recommend, for the record. We should always adapt things to our cell line before running. But in this case, it worked very well. Both clones transitioned to the medium seamlessly and we're able to reach their desired cell density that was both higher and over a full day faster than expected. This is even more impressive when we consider that glucose feed was not planned for this as the cell densities were not expected to reach that high that quickly. And thus, glucose was nearly depleted around day 3 causing the inflection seen in the figure. So we returned to the concentrated fed-batch example. This was executed at Bielefeld University and they want to explore using a concentrated fed-batch perfusion process to improve the performance of a lower producing cell clone. The 3 conditions shown here are simple fed-batch, SFB, which is glucose only batch operation using their existing condition. The fed-batch condition uses one of our fed-batch prototype media and 2x concentrated efficient feed C plus; and lastly, concentrated fed-batch is our perfusion medium, running a concentrated fed-batch process. The peak medium exchange rate they used here was 1.2 vessel volumes per day with a TFF to retain the cells and product in the reactor. During the concentrated fed-batch run, the beta tester cell line achieved greater than 120 million cells per milliliter compared to the peak VCD at around 20 to 25 million viable cells per ml in the simple batch and fed-batch conditions. The fed-batch condition provided a very strong 5.5x increase in the cell productivity, which resulted in a final titer of around 260 milligrams per liter compared to the simple batch condition generating about 46. The perfusion meeting condition further intensified the productivity of the cell line by nearly double the fed-batch. And when paired with the drastically increased cell density, this resulted in a 42x increase in productivity, reaching a titer of about 1.9 grams per liter, which is a range that can be comfortably processed and batched downstream without requiring an additional concentrating step. So these are kind of a combination of N-1 and intensified fed-batch results. Zurich University was also looking at demonstrating an N-1 application with VVD ramped up to maintain log growth up to a maximum exchange rate of 3 VVD. However, this run was continued to run past log growth to get an initial feel for what intensified fed-batch performance would look like. Unfortunately, O2 transfer became a limiting factor at the peak VCD of about 160 million per ml viable cell density. And O2 fell below 10% air saturation. So the medium exchange was stopped shortly after that. However, for comparative purposes, this same cell line with fed-batch operation reaches a peak VCD of around 18 million per ml viable cells and about 2.2 grams per liter of final product yield after 14 days. At around the 9-day mark when this was stopped, a typical fed-batch run for the cell line reaches about 1 gram per liter. Prior to stopping this process, we were reaching, in this intensified fed-batch model, about 1.5 grams per liter per day, so about 70% of the total IgG produced in the entire 14-day process in a day. As previously discussed, perfusion medium has diverse needs depending on which cell line it is paired with and what process is being used. To better support this, we've worked this formulation to support different concentrations. A 66% dilute concentration is closer to typical fed-batch media. This functions well for cell adaptation and maintaining growth rate at lower viable cell density operations if the cell clone is sensitive to higher concentrated media. A higher concentration oriented for production is designed to counter typical osmolality drop during demanding cell culture perfusion and provides better medium depth to support sustained high cell density and productivity at low-to-moderate medium exchange rates. The continuous perfusion run shown here is then benched through tanked reactors with ATF 2 filters at 1 vessel volume per day medium exchange rate with in-bleed actively adjusted to target a 95% viability. Both conditions start in the 66% diluted concentration of the medium, the bottom left data shows viable cell density in solid lines and percent viability in dash lines. The blue condition, which shows early steady-state behavior is run using 66% dilute medium for the entire process. The yellow condition switches to higher medium concentration twice as osmolality drops, being brought to 80% and again to full concentration. Performing a concentration increase in this matter mitigates the impact to growth rate in a clone that is sensitive to high medium concentration at the low cell density operation. And in this case, shows an increased growth rate in the yellow condition at full concentration. Cell bleed, the dash lines in the upper right plot, are implemented to prevent cell density from reaching excessive cell levels and avoiding strong nutrient imbalances with a risk of sudden viability dropping. The bleed starting point is determined by a sudden change in growth rate as the cell density is rising. In this case, the bleed was started too early and then too high of a flow rate in the yellow production concentration condition. This caused the percent viability to stay too high at around 98% and limited the initial viable cell density. The bleed rate was carefully reduced to bring the process back towards around the 95% viability target and steady state was maintained from around day 30 to 42. This resulted in a steady-state viable cell density of about 90 million cells per milliliter for the 66% dilute medium condition and 120 million cells per milliliter for the full concentration medium. The steady-state productivity for this cell line was excellent. In steady state for this run, it was over 1.7 grams per liter per day in the full concentration and over 1.2 grams per liter per day in the 66% dilute medium. For comparative purposes, in a 14-day simple batch, using this same cell clone and medium, the total productivity after 14 days was about 1.3 grams per liter. When we think about the title of this presentation and what we mean when we say perfusion medium works in perfusion, this is an example of what we're talking about. Very few people would look at that batch production and say, "Hey, that's going to be a really great perfusion medium." It really doesn't shine until you're running it in the application itself. More on that later as we get into considerations for best practices and screening. For this, I want to touch off on one other aspect. When we're running perfusion, it's important to consider the cell size and the cell size change during the operation of the run. We're showing that in the bottom right plot, which takes into account cell size by showing viable cell volume in the solid lines. Viable cell volume is calculated by simply taking the viable cell density, multiplying it by 4-point -- 4/3s x Pi r cubed to get your total volume of cytoplasm that you're supporting during the run. And then the productivity scale per viable cell volume is shown in the dash lines. When we look at the viable cell volume being maintained between the 2 conditions, we can see that the growth rate fall off that was used to determine starting bleed occurred at a similar viable cell mass or a total viable cell volume. This may suggest that signaling was a primary factor in the growth rate change as opposed to any nutrient considerations. Of course, that's speculation. Additional work would need to be executed to prove that out. But additionally, the upper right solid plot lines shows cell productivity based in a traditional method on viable cell density in picograms per cell per day. However, when we look at productivity scaled per viable cell volume in milligrams per cubic centimeter of cytoplasm, we get a very different picture. The productivity is almost perfectly in alignment between both conditions, the entire run. And the reason for this is the cell diameter was different during the period of time that it was over-bled and running at an excessively high percent viability. It's a very different picture than what you see when you're focusing purely on the viable cell density and trying to make conclusions about the behavior of the run. So this is an additional set of perfusion results executed by Patheon's St. Louis site. So in this case, again, the process was over-bled resulting in a percent viability maintained a little too high, around 97%, and restricting the viable cell density target somewhat. However, productivity still reached around 1.6 grams per liter per day. And it's worth noting again, when we look at the viable cell density in the black compared to the viable cell volume in the blue, we would potentially be led to believe that we're practically at steady state if we were just looking at the black viable cell density. However, when we look at the viable cell volume, we can see that the process is still changing and settling towards a final operating steady state. So evaluating the perfusion process and media to be used with it provides a lot to consider. The best practices I'm going to bring up here deserve a lot more time and attention than we can afford to give them right now. But I want to cover some of the main highlights. First and foremost, a good test setup gives you results that are meaningful with respect to your process goals. The table on the upper right shows 4 different ways to think of agitation in some of the typical operating regions. Understanding what parameters you need to scale with your model and also where you can be flexible is really important to be able to build off of your process going forward. Separately, you need to know where your quality has to be and what governs it to establish uniform control boundaries for your run. This is particularly critically important for media screening. We just showed you some examples where tiny differences in the maintained percent viability lead to significant differences in the sustained cell density and total production titer. If you're not maintaining your consistent quality criteria that you're operating under, you're not going to be able to get a fair comparison in your screening operation results, and it won't mean as much when you're trying to take your process further. Second, when preparing media conditions for a run, supplements are often added. So some cell lines have complex or specialty components present in the medium they're developed in, such as growth factors or hydrosolates. It is highly advisable to make sure that these are present in all medium conditions tested, so that we're not trying to both wean off of a growth factor that a clone is addicted to and used to while trying to adapt at the same time. Those should always be taken as separate operations. When performing supplementation for perfusion, components that normally accumulate in the medium are being constantly removed. And so you may need to add those back at a higher concentration to mimic your typical accumulation target in a fed-batch operation. Conversely, components that tend to accumulate and then cells themselves should be considered very carefully as the continuous addition of that component can potentially overwhelm the cells. The lower left figure shows results of a greatly changing cell demand for L-glutamine during a run causing considerable control difficulties. Third, always adapt a cell line to new medium conditions before testing. Ideally make sure your growth rate and your productivity are stable. Similarly, if you want to perform a mid- media process change, including concentration or supplement changes, see how it behaves in your flasks first so that you have an idea of what to expect. Flasks are pretty cheap. Reactor runs can get pricey. Fourth, note limits of your cell line, some cell lines lose productivity fast when selective pressure is removed. Some don't behave well at higher VCD, either due to increased collisions and signaling or due to physical operating stress of the more aggressive environment, more aggressive agitation and gassing, particularly heavy gas barging. The plot at the bottom left is from a cell line stress run with ramping agitation up to 600 RPM. The cell line handled this very well as well as other mechanical stress tests. Unfortunately, the production stability wasn't sufficient and made it a poor candidate for our continuous perfusion testing operations in-house. And a word on screening specifically in short versus long screening, screening medium for a continuous process is particularly difficult to model with short screening. Again, based on cell health profile, in this case, we assume both conditions are similar. If we only ran 15 days for the information shown here, medium A would be a clear choice and looks like steady state has been achieved. However, as the medium components got reduced in towards their final steady-state concentrations, medium B productivity jumped upwards and stayed higher for the rest of the run, giving a much higher net titer. By comparison, medium B, got a similar boost, but then lost productivity and got a much weaker overall performance. This cell line also had some lactate cycling oscillations, which further makes it more complicated for hitting a good steady state early. It's words that people don't want to hear, right? We want quick and easy screening, but we need to be careful when we're screening to try to orient it around the process being targeted. And if we say we want to target a sustained steady-state operation, we should at least be checking for our screening process to be representative of that so that we don't get tripped up as we start moving into actual representative vessel runs. So thank you very much for your time. If you have further inquiries, a great place to start is contacting our global product manager, [email protected]. And I'm looking forward to the Q&A section. Thank you.

Thank you, Christopher, for an excellent presentation. We've received a few questions already, but we'll give the rest of you a moment to enter your questions in a Q&A box to the left of the slides. Before we begin the Q&A, I'll run through some brief announcements. First, I'd like to thank Thermo Fisher scientific for sponsoring this digital week. Next, I'd like to quickly draw your attention to our Face-to-Face BioProcess International Events, visiting Boston in September and Milan in October this year. Additionally, BioProcess International Europe will be delivered as a 100% virtual conference and exhibition on July 13 through 17, 2020. Also be sure to check out the resources list to the right of your screen where you can download a few featured white papers. Now back to Christopher to begin the Q&A. The first question we have here is what clone was the medium tested in?

That's a great question. So in both beta and in-house testing results, we got a favorable performance in all the major CHO cell lines. So K1, DG44, CHO-S, CHO-GS. Additionally, we still have ongoing testing that has understandably been somewhat disrupted and slowed down with the COVID-19 situation, but we are waiting on getting results back in for some of the more exotic CHO cell lines, such as the U.K. and also additional parental cell lines such as HEK and GEx. So it will be very interesting to see that data as it rolls in. Hopefully, that answers the question.

All right. So our next question would be, how do you define a batch?

Right. So I'm going to assume that, that's in reference to continuous processing. So good question. Batches can be defined either by volume or by time in the downstream handling of it and can then be qualified in individual pieces on that. Again, that's a nice risk mitigation consideration. It is a change to a regulatory approach with downstream. If you're not ready to make that jump in the regulatory approach, you can consider doing a hybrid downstream model to still gain some resin cost efficiency benefits, but still have more of a typical batch release process that's defined by the steps and stages of the noncontinuous stages of your downstream function.

Great. Next, we have why would you pick filter-based versus nonfilter-based cell retention?

That's a tricky question. So for, obviously, in the case of a concentrated fed-batch filter is your only option. You're selecting a filter with an intentionally gated poor size to retain your product and concentrate in the reactor. For nonconcentrated fed-batch for the other perfusion type processes, you've got to weigh the 100% efficiency of your filters versus the cost and the scale-up targets. Your nonfilter-based separating methods tend to be a little more finicky in the flow rate needs, which can make them a little trickier to scale. And they do have a constant cell loss. For acoustic method, I believe it's typically around 10%. In an intensified fed-batch or a continuous perfusion that may not matter as much. But if you were looking at, say, N-1, for example, any cell loss is directly countering what you're trying to do in your N-1 and intensified seed-train operations. So I think filter would be kind of an obvious choice there. But really, you've got to make the decision based on the process and how you're applying things. Hopefully, that was useful.

Great. So our next question is, in my work with perfusion, I've had to you different media for the different stages of the workflow. How do you recommend optimizing the medium?

That is a great question. So we -- that is one of the reasons why we oriented and targeted a media design that provides for flexible reconstitution at different concentrations so that you can have one SKU to work with the different process needs that occur during the result of a process run. So as noted before, the dilute concentration, if you've got a cell line that's sensitive and has slower growth rates at higher concentration, you can run that dilute reconstitution for those stages of the process to facilitate it. And then in regard to your production stage, obviously, you can run to a much higher concentration with our media to better support those needs, switching different media mid-process we generally frown upon it because you're asking the cell line to adapt to a totally different formulation during the run, that can be a tall order. We saw an example where it worked very fine in this presentation, but I would consider that more of the exception than the rule. So I generally would not recommend using completely different media SKUs at different points of the process unless you can verify in your flask work that the cells can adapt seamlessly between them.

Great. All right. So here's another question. What supplements do you recommend for use with this medium?

Good question. So a big factor there is going to be the selection process that your clone is based on. For a GS type clone, the only thing that I would consider adding would be our anti-clump agent as needed, if you are having any kind of aggregation issues with your cell line. If it's a non-GS-type cell lines, so a DHFR or what have you, then we would generally recommend also supplementing, in addition to anti-clump agent, if needed, for your clone. We would also recommend supplementing around 4 millimolar of L-glutamine. It gives a nice kick to the initial growth of the cells without overwhelming the cells as you get towards your perfusion steady states and higher viable cell density. Other than that maintaining a glucose feed to keep glucose at a reasonable range for your process. Internally, we usually target around 3 grams per liter. We'll consider going lower if a cell clone is exhibiting some runaway lactate. But other than that, it's that simple. I mean the only other supplement that you would put in would be based on if the clone that you were starting with was used to growth factors or hydrosolate. Although I suppose I can put in a very obvious plug here that we did acquire some very, very nice, highest-quality hydrosolates in the world, high-quality peptones. And in the case of perfusion, where the cell death rate is a huge factor on what you can sustain, especially in a continuous process. I'm actually very interested myself in trying some of our peptones and seeing how effectively that can optimize the process. So hopefully, when the COVID-19 ends and I can spend more time in the lab, that'll be some of the fun results we can play with and share in the future.

Okay. That looks like that's all the time we have for questions today. Thank you, Christopher, for a great session. If anyone has submitted a question that wasn't addressed, please keep in mind that the speaker will reach out to you directly. And if you could, Thermo Fisher would appreciate it if you could select one of the following: I have additional questions. Please have a representative contact me; I am interested in a quote, no follow-up necessary. And I'll just give you a couple of seconds to just fill out that survey there. Again, just select one or more of the options: I have additional questions. Please have a representative contact me; I am interested in a quote, no follow-up necessary. And please be sure to hit submit at the bottom there of the little screen that pops up to make sure that your selection goes through. Okay. So just so everyone knows, this session was recorded. You will receive a notification in 24 hours when the on-demand session is available for viewing. Before you log off, please take a moment to complete the feedback form so we can continue to improve Digital Week Experience. On behalf of Informa Connect Life Sciences, have a great day.